title
stringlengths 1
827
⌀ | uuid
stringlengths 36
36
| pmc_id
stringlengths 5
8
| search_term
stringclasses 44
values | text
stringlengths 8
8.58M
|
|---|---|---|---|---|
AAPM Medical Physics Practice Guideline 12.a: Fluoroscopy dose management | e3359839-e804-47e2-bf7d-25796b6875ce | 8906204 | Internal Medicine[mh] | INTRODUCTION Fluoroscopy equipment is used to observe or guide moving objects such as internal organs, contrast agents, catheters, and guidewires within the body to diagnose and treat disease. Procedure times range from several seconds to multiple hours, and fluoroscopes range from small, mobile C‐arms used to image extremities, to complex single‐ or bi‐plane angiography systems. These more complex fluoroscopes are used to guide performance of fluoroscopically guided interventional (FGI) procedures, and help to provide lifesaving diagnostic and therapeutic services for patients. However, unlike simpler procedures commonly accomplished using general or mobile C‐arm fluoroscopes, long and complex FGI procedures can exceed radiation thresholds for tissue reactions. Proper identification, follow‐up, and management of patients receiving high doses from FGI procedures are essential parts of patient care due to the slowly developing nature of radiation‐induced tissue reactions. Recent standards and requirements from accrediting bodies such as The Joint Commission (TJC) and state regulatory agencies have brought focus to this issue, requiring hospitals to record patient fluoroscopy exam dose indices and to establish committees, policies, and procedures for reviewing those data and providing patient follow‐up as appropriate. These standards are in addition to TJC's updated fluoroscopy sentinel event standard, which requires identification and investigation of severe tissue effects. Many organizations and societies have provided guidance and resources for managing patient dose, including the National Council on Radiation Protection and Measurements (NCRP), the Conference of Radiation Control Program Directors (CRCPD), the Department of Veterans Affairs, the Society of Interventional Radiology (SIR), and multiple cardiology societies under the umbrella of the American College of Cardiology Task Force on Expert Consensus Decision Pathways. 1–5 This AAPM practice guideline aims to outline the role of the diagnostic qualified medical physicist (QMP), as defined by AAPM Policy Number PP 1‐J “Definition of A Qualified Medical Physicist,” in practical patient dose management for FGI procedures. This role includes helping facilities set up policies related to dose management, including pre‐procedure patient consent, intra‐procedure dose index level notification, and post‐procedure follow‐up for potential tissue reactions. Suggestions for methods of complying with TJC standards and various state regulatory requirements for tracking radiation use, setting dose index thresholds, and analyzing dose index data are provided, along with a discussion of the challenges posed by these requirements. The QMP's role in helping facilities comply with TJC's updated “radiation overdose” sentinel event standards by investigating severe tissue reactions is also discussed. Related fluoroscopy topics that may fall under the QMP's oversight, such as operator credentialing and occupational radiation exposure monitoring are briefly discussed.
TISSUE REACTIONS Tissue reactions, also known as deterministic effects, are due to radiation effects on populations of cells and are characterized by a threshold dose and an increase in the severity of the reaction with increasing dose. These reactions are the primary concern for patients undergoing FGI procedures, and will be the focus of this report, with stochastic risks not being addressed. Tissue reactions do not occur at doses below a threshold dose, which the International Commission on Radiological Protection (ICRP) defines as the dose estimated to result in a 1% incidence of the tissue reaction. Tissue reactions in patients undergoing FGI procedures may involve skin, hair, subcutaneous fat, muscle, the lens of the eye, and sometimes bone. , The generally accepted minimum threshold dose for transient skin effects is an absorbed skin dose of approximately 2 Gy, and permanent effects are unlikely below an absorbed skin dose of 5 Gy. , Risks for tissue reaction can conservatively be assumed as cumulative when the same skin area has been irradiated for other procedures. Repair of sublethal radiation injury to skin is typically complete within a day of exposure; repopulation of cells can require months. Tissue reactions may be expressed days to years after exposure, depending on the radiation dose and the tissue affected. Early reactions may be due to inflammation, and may not be noticed by the patient, whereas late reactions are typically due to cell loss. Tissue reactions in the skin range in severity from erythema and transient epilation to dermal necrosis, which can require surgical intervention. Because of individual variability in radiosensitivity, the radiation dose necessary to produce a specific effect and the time course of the tissue reaction are best thought of as ranges, rather than specific values, as shown in Table , reprinted from Balter et al. Additionally, it should be noted that previously irradiated skin is at a higher risk for developing tissue reactions than areas that have had no prior exposure.
AVAILABLE DOSE INDICES Predicting the likelihood of radiation‐induced effects from FGI procedures requires an estimation of the patient's radiation dose. Four measurable radiation dose indices exist to assist with this estimation: fluoroscopy time, cumulative air kerma ( K a,r or CAK), air kerma–area product ( P KA , also commonly written as KAP or dose–area product/DAP), and peak skin dose (PSD). Effective dose is not suitable for assessing the likelihood of tissue effects. The availability and displayed units of each dose index vary depending on equipment type, manufacturer, and equipment age, and each differs in clinical utility and application. Fluoroscopy time is the most widely available index; however, it is also the least useful in terms of predicting potential tissue effects. Fluoroscopy time alone is inadequate to estimate patient dose. It does not consider fluoroscopic dose rate, and dose estimates that rely solely on fluoroscopy time can vary widely, as acquisitions (e.g., cine, Digital Subtraction Angiography (DSA)), which can contribute substantial dose, are not included in the measurement. If additional dose rate and dose per image data are available for the specific mode of operation used, using the number of fluorographic images from a procedure, along with fluoroscopy time, can improve estimations of patient dose. However, while fluoroscopy time is suboptimal for assessing radiation dose, it can be useful for other purposes (e.g., a surrogate for procedure complexity and comparing practice among operators). CAK ( K a,r ) is required to be displayed on all International Electrotechnical Commission (IEC) compliant interventional fluoroscopes and all fluoroscopes sold in the United States since June 2006. , , K a,r is the cumulative kerma for a fluoroscopic procedure, including fluoroscopic and acquisition modes of operation, measured in air at a specific reference point relative to the X‐ray source. For isocentric C‐arms, the IEC definition for the reference point is along the central ray of the X‐ray beam, 15 cm from the isocenter toward the X‐ray tube, though manufacturers can use a different reference point if they choose. The specific reference point used by a piece of equipment is defined in the operator's manual. K a,r is often used as a surrogate for the patient's PSD; however, potentially labor‐intensive corrections and calculations are required if a more accurate estimate of PSD is needed. These corrections include backscatter, table and pad attenuation, displayed dose index accuracy, and tissue‐to‐air ratio. Additionally, due to beam geometry, gantry angulation, table height, and patient size, the reference point may not coincide with the entrance skin surface. This results in a tendency for K a,r to overestimate PSD. Despite these shortcomings, K a,r is generally considered a practical surrogate for skin dose. K a,r is almost universally reported in units of milligray (mGy) in modern fluoroscopes. KAP ( P KA ), sometimes called DAP, is the product of air kerma and the geometric area irradiated in the same plane orthogonal to the propagation of the X‐ray beam. P KA displays are widely available on interventional fluoroscopes and commonly available on many modern mobile C‐arms and conventional diagnostic fluoroscopy equipment. Unlike K a,r , P KA is independent of distance from the focal spot, because the irradiated area increases proportionally to the decreased radiation intensity as distance increases. Therefore, small doses to a large area and large doses to a small area could give equal P KA values. For this reason, P KA is considered a poor indicator of skin dose and radiation‐induced tissue effects. Because P KA represents the total energy deposited in the patient, it is better correlated to stochastic risk as compared with K a,r . Additionally, a lack of standardized units for displayed P KA values on fluoroscopic equipment can make the practical implementation of clinical thresholds difficult. P KA displayed units of uGy × m 2 , mGy × cm 2 , cGy × cm 2 , and Gy × cm 2 are all in use. PSD indicators, with real‐time dose mapping displays, are the least common dose index available at present but are becoming increasingly common on modern FGI equipment. They allow the operator to visualize the three‐dimensional skin dose distribution, potentially preventing tissue reactions. Freestanding software, independent of the FGI equipment, is also available that can estimate PSD based on data provided in the radiation dose structured report (RDSR). Real‐time PSD information provides the greatest clinical utility for predicting the likelihood of tissue reactions because it provides estimates of the highest skin dose and its location using information on patient position, X‐ray field size, and beam angulation during a procedure. Freestanding software that reconstructs PSD from RDSR information also has clinical utility but lacks direct feedback to the operator during a procedure. It is important to be aware of manufacturer‐specific approaches to PSD estimates and their level of sophistication with regard to inclusion of correction factors such as backscatter, table and table pad attenuation, and patient anatomical representation. Because of these differences in approach, it is possible that fluoroscopes from different manufacturers, or even different versions of the same manufacturer's software, could provide different PSD values given identical RDSRs. A review of various PSD software options was presented by Malchair et al. At present, real‐time PSD estimates are limited to FGI fluoroscopes and are not found on mobile C‐arms or general fluoroscopic equipment. PSD is commonly reported in units of milligray (mGy). In setting up a patient fluoroscopy dose management program, the QMP will need to survey the dose indices available on the imaging equipment in a facility or larger healthcare system. The availability of these indices will determine the dose index used to set thresholds for further action. Some dose indices may be available but require equipment configuration in order to be displayed. For older equipment that does not display K a,r or P KA , aftermarket meters can provide this capability. Ideally, one would use PSD, the dose index that best correlates with potential tissue injury, but its limited availability and the variability in assumptions and corrections made by PSD algorithms complicate its use at the present time. Due to its standard definition, ubiquitous implementation, and reasonable correlation with PSD, it is the recommendation of this group that CAK ( K a,r ) be the primary dose index used in setting fluoroscopic threshold dose levels for notification and patient follow‐up.
FGI POLICY The following sections broadly describe the policies that facilities must have in place, regardless of regulatory or accreditation requirements, to increase stakeholder awareness and to help prevent, identify, and properly care for patients with tissue reactions due to high‐dose FGI procedures. These policies are broken into three sections: Pre‐procedure screening and consent of the patient. Intra‐procedure monitoring and notification of patient dose index level to the team. Post‐procedural patient follow‐up above threshold levels. While QMPs are not typically involved in the day‐to‐day implementation of these policies and procedures, their expertise is critical in development of the policies. Additionally, the QMP can be consulted when specific questions arise and may be called upon for PSD or other dose estimates. 5.1 Pre‐procedure consent for risk of skin injury Each healthcare facility must create a policy for obtaining a radiation‐specific consent from patients prior to FGI procedures where patients are given information regarding the risks of radiogenic tissue effects. Whenever possible, this policy should be standardized across all departments utilizing FGI equipment across the entire organization. Local laws may dictate whether consent can be verbal or must be written. A facility may choose to obtain this consent before all procedures performed in FGI suites or only before a subset of procedures classified as potentially high dose. The QMP can aid in reviewing facility data to determine which procedures require consent. NCRP Report No. 168 suggests classifying procedures as “potentially high radiation dose procedures” if more than 5% of procedures result in K a,r exceeding 3 Gy. Obtaining a radiation consent for only a subset of procedures removes a clinical step prior to procedures where tissue effects are highly unlikely. However, it may complicate policy and process by requiring someone to create and keep updated a list of procedures in radiology, cardiology, surgery, and other specialties requiring radiation consent. An example of simple radiation consent language, adapted from the SIR's guidelines for patient dose management, can be found in Appendix A. Additional example consent language can be found in NCRP Report Nos. 168 and 185. , Any such document should be reviewed by appropriate clinical and legal teams prior to implementation. Key elements of consent language should include a description of the use of X‐rays in fluoroscopy, the possible tissue effects resulting from prolonged exposure to X‐rays, the typical time delays for these effects to occur, and the proposed action(s) for the patient and/or caregiver if any effect is observed. 5.2 Pre‐procedure screening In addition to screening for potential medical issues or pregnancy, patients who will undergo potentially high‐dose FGI procedures should be screened to determine if they are at higher risk for tissue effects. The result of pre‐procedure patient screening is then conveyed to the performing physician. Patients at higher risk include those who have had recent high‐dose fluoroscopy procedures or a history of radiation therapy to the same skin area, collagen vascular disease, or certain genetic disorders that affect DNA repair, which are further detailed in Ref. High body mass index (BMI) is also a risk factor since the greater amount of tissue increases the amount of radiation needed to yield an adequate image and can result in the skin being closer to the X‐ray source. 5.3 Intra‐procedure monitoring and notifications Intra‐procedural notifications regarding radiation dose levels allow the performing physician to gauge the benefit–risk ratio at each stage of an FGI procedure. All FGI procedures should be justified, that is, offer a clinical benefit to the patient greater than the potential risks, which include radiation tissue reactions. This benefit–risk ratio is considered by the physician when initially deciding whether to perform a procedure, and later, while the procedure is in progress. For the benefit–risk ratio to be meaningful, the physician needs to have an accurate understanding of radiation risk, the likelihood of tissue injury, and the associated dose–response relationship. The QMP, as the subject matter expert in this area, can provide radiation protection planning knowledge to the team to help ensure that the benefit–risk ratio is formulated correctly. The QMP must also understand the magnitude of radiation risk compared to other procedural risks, which are often much greater than the risk of tissue reaction. The risk of a radiation effect is typically much smaller than other procedural risks, such as bleeding, infection, and organ damage. However, the benefit–risk ratio is not static, and may change during the procedure. This analysis should be performed by a well‐informed operator and should be evaluated throughout the procedure. Procedures should rarely, if ever, be stopped solely due to radiation dose. If all or most progress made during the procedure is lost if the procedure is stopped (e.g., navigation of catheter to a very difficult site, or risk of developing collateral vasculature in the interval), any risks already incurred will have been for no benefit to the patient. When appropriate clinically, very complex procedures may be planned in a staged fashion, with multiple sessions separated by 8–10 weeks, to fractionate the dose to the skin and reduce the likelihood of tissue reactions. The concept and implementation of radiation dose notification levels is simple. The same concept has been applied to other potentially toxic agents, such as iodine contrast or medications. The implementation of notification levels requires the entire procedure team to work together. For example, the radiologic technologist, who is a local expert on the operation of fluoroscopy equipment, may be the one who calls out when the notification level is reached and documents this action in the procedure record. The nurse may then document the same information in the electronic medical record, while the operator or performing physician pauses for a moment to consider their radiation management strategy and the current benefit–risk pace of the procedure. The essential elements of an intra‐procedural dose notification are: Verbal notification to the operator regarding the current notification number and the magnitude of the dose index, for example, “This is the third notification. The current reference air kerma is 5000 mGy.” Consideration of a procedural pause to evaluate setup and the benefit–risk pace of the procedure, if such a pause will not interfere with the conduct of the procedure. Documentation of the notification level, that the operator was notified, that the setup and benefit–risk ratio of the procedure were evaluated, and any specific actions taken by the operator. The notification should be delivered during a natural pause during the procedure, if possible, and should never interrupt a critical phase of the procedure. Re‐consideration of radiation dose management strategy after a notification may include specific actions including, but not limited to: Adjustment of the patient table height. Sometimes it is set within the optimal range at the beginning of the procedure, but may be lowered during the procedure, for example, to isocenter the patient for volumetric imaging, after which the table is not moved back to the original height. Elimination of a large air gap by lowering the image receptor. Evaluation of X‐ray beam collimation and gantry angle. Evaluation of currently selected organ program (aka imaging protocol) for aspects such as dose rate setting, pulse rates, and DSA frame rate. Traditional radiation management practice often suggests that the angle of the gantry be varied during a procedure to “spread” radiation dose across the skin. While this strategy can be useful in specific circumstances as a prophylactic technique, it can be detrimental to radiation management and greatly increase both the skin dose rate and the PSD, especially in cases where larger oblique angles both increase radiation output and put the patient's skin surface closer to the tube. , , If used as a radiation management strategy, tight collimation of the X‐ray beam increases the benefit of this technique. Onboard, real‐time PSD mapping capabilities, if available, can help determine the benefit of these strategies. Until a time at which PSD is widely available as a real‐time displayed dose index, K a,r is the preferred dose index for dose notifications in fluoroscopy. However, as PSD is more closely correlated with the likelihood of tissue reaction, it should be used for intra‐procedural notifications when available. Conservatively, values for both planes in biplane procedures should be summed for the purpose of intra‐procedural notifications. Multiple publications exist with a consensus recommendation for setting notification levels with the first notification at 3 Gy K a,r and subsequent notifications every 1 Gy K a,r . , , , , However, there are practice‐specific considerations for setting notification levels, as well as patient‐specific considerations, such as the presence of sensitizing factors, that should be used to modify notification values when appropriate. Practice‐specific factors to consider when setting notification levels include: Types of procedures performed : Depending on the procedure type (e.g., neuro, body, cardiac) the fluoroscopy dose notification level could be adjusted in consideration of the relationship between K a,r and PSD. The geometry of the procedure affects the ratio of PSD to K a,r , denoted as the “dose index” by Miller et al. To avoid confusion, the “dose index” as described by Miller, will be referred to as the “PSD factor” in this report. For non‐isocentric procedures, such as those performed in vascular and interventional radiology, the PSD is often similar to the reported K a,r (PSD factor ∼ 1.0). For isocentric procedures, where the relevant anatomy is placed at isocenter to facilitate the use of multiple gantry angles, the skin dose will be higher than the K a,r if only a single projection, or a narrow range of projections, is used (PSD factor > 1.0). However, isocentric procedures often require the use of many different gantry angles, which may result in multiple distinct “fields” on the skin of the patient. This tends to reduce the dose index, and the PSD would be less than the K a,r (PSD factor < 1.0). Equipment capabilities : Many interventional fluoroscopes now offer the capability to program one or more fluoroscopy dose notification levels into the system. Upon reaching that notification level, an alarm sounds and/or a visual indication is displayed at the position of the operator. If, for example, a fluoroscope offers only three such levels at fixed values, this is a consideration when selecting the first, second, and third notification levels in policy. Different notification levels may be implemented for different services based on the type of procedures performed or the available dose indices. The QMP should ensure that any programmed dose notification levels are harmonized, though this can be complicated by differences in manufacturer's specific alert configurations. Alarm fatigue : The 5‐min fluoroscopy timer is often ignored in interventional fluoroscopy because it sounds numerous times during most procedures. Avoiding similar alarm fatigue with dose notification levels is an important consideration, both for the value of the first notification level and the intervals between notification levels. One other important intra‐procedural notification is the rapid recognition and correction of unsafe conditions when they exist. Examples of unsafe conditions include the presence of the patient's arm in the field of view when the arm is not the area of interest, and direct irradiation of the eye lens or female breast tissue when unnecessary. Every member of the procedural team should be confident in identifying such conditions when they exist and reporting them to the operator and responsible physician. Diagnostic medical physicists can provide maximum value to their healthcare organizations only when they understand the clinical aspects of fluoroscopically guided procedures. Fluoroscopy dose indices can only be understood completely when accompanied by the clinical context of the procedure, and imaging protocols for fluoroscopy can be optimized only when the tasks involved in completing the procedure are understood. For these reasons, it is important that a QMP who supports a clinical FGI practice be provided clinical instruction and time to observe clinical procedures, including how the equipment and protocols are used by physician operators. It is likely that this observation time can also be used to optimize protocols. 5.4 Establishing the SRDL and post‐procedure actions The substantial radiation dose level (SRDL) is an operational threshold for radiation dose above which additional post‐procedure actions for patient care should be taken due to the potential for biological harm. The SRDL should be set to a level such that a radiation dose below the SRDL is unlikely to result in a tissue injury for an average patient of normal radiation sensitivity. However, “there is no implication that a radiation level below an SRDL is absolutely safe or that a radiation level above an SRDL will always cause an injury,” and patient‐specific factors such as underlying conditions or medications could influence the threshold to induce a tissue reaction. Suggested values for the SRDL are 5 Gy K a,r or 3 Gy PSD. , , Since real‐time PSD estimates are not typically available on most equipment, use of K a,r is generally recommended to establish thresholds for post‐procedure follow‐up. When setting the K a,r SRDL value, the previously discussed PSD factor concept should be considered. If it is known that the entrance skin point is likely to be unchanged during a procedure or the skin is closer to the X‐ray source than the interventional reference point (IRP)/patient entrance reference point (PERP), a lower SRDL may be warranted. Similarly, for biplane procedures, the dose received from each plane should be added for SRDL purposes unless it is known that the fields do not overlap. Each facility must have a policy for identifying patients who receive a radiation dose exceeding the SRDL and for providing patient management and follow‐up, including provision of post‐procedure information written in simple language. This document (see Appendix B, adapted from NCRP Report No. 168 language, for an example) should provide educational information about the procedure, follow‐up information, and provider contact information for questions or concerns. This information can be provided in discharge instructions, in the electronic medical record, or both. Determination of the individual responsible for distributing these instructions is vital and should be part of the facility's policy. Technologists, nurses, physicians, or physician assistants can be assigned this duty. Further, the facility must have an established system for patient follow‐up and must document any such communication in the medical record. Follow‐up may be in‐person or via telemedicine and may reasonably be provided by any trained clinical team member, under the direction of the performing physician. All tissue effects should be assumed radiogenic until proven otherwise. If a severe or prolonged radiogenic tissue reaction is observed, the patient should be seen in person by the performing physician whenever possible and referred to dermatology, wound care, radiation oncology, or another appropriate specialty for further care, with all appropriate information included in the patient's medical record. PSD estimates may be useful for patient management in some cases. Facilities should have a defined process for requesting a PSD estimate performed by or under the supervision of a QMP. PSD estimates required by policy, regulation or accreditation standard, or requested from a licensed provider, must be documented in the patient medical record. The details of estimating PSD from dose indices are outside the scope of this document and have been described elsewhere. , Due to the various assumptions and inherent uncertainties involved, PSD estimates are unlikely to be better than ±50% accurate. This uncertainty should be included in any estimate and QMPs should always include a range of possible values in addition to the likely PSD value. For example, depending on various assumptions made regarding variable procedure aspects such as table height, collimation, and beam angulation, a PSD estimate could be documented as “likely 13 Gy, but with a probable range of 8–18 Gy.” As of January 2021, the American Medical Association (AMA) added a billable Current Procedural Terminology (CPT) code that refers to PSD estimates; code 76145 applies to evaluation of radiation exposure that exceeds institutional thresholds. 5.5 Cumulative doses and risk of tissue reactions There may be instances where a patient has multiple procedures in a short span of time where no single procedure reaches the threshold for patient follow‐up, but the summed doses warrant action. Setting an appropriate time window for a meaningful summation of doses is challenging. As previously mentioned, repair of sublethal radiation injury to skin is typically complete within a day of exposure, so doses delivered within a 24‐h period should invariably be summed. However, repopulation of damaged cells can require months. The SIR's guidelines for patient dose management suggest summing doses over a 60‐day period, while prior TJC sentinel event standards required summing doses over 6–12 months. , , This report recommends 60 days as the most biologically meaningful time window for summing patient skin dose. The same thresholds for follow‐up should be used for single or multiple cumulative procedures that exceed the SRDL. Identifying patients with high doses from multiple procedures, and their potential risk, is technically challenging. The patient's history of radiation exposures may be incomplete if some exposures occurred at outside facilities, or even in a different department within the same facility, if procedure information is not shared. Even with sophisticated dose management software that can track patient dose, there can be a time lag in identifying these patients, who may be discharged before they are identified as having exceeded SRDLs. In these cases, facility policy should identify how the patient will be contacted and who will be responsible for follow‐up. The treating physician is ultimately responsible for ensuring follow‐up is performed.
Pre‐procedure consent for risk of skin injury Each healthcare facility must create a policy for obtaining a radiation‐specific consent from patients prior to FGI procedures where patients are given information regarding the risks of radiogenic tissue effects. Whenever possible, this policy should be standardized across all departments utilizing FGI equipment across the entire organization. Local laws may dictate whether consent can be verbal or must be written. A facility may choose to obtain this consent before all procedures performed in FGI suites or only before a subset of procedures classified as potentially high dose. The QMP can aid in reviewing facility data to determine which procedures require consent. NCRP Report No. 168 suggests classifying procedures as “potentially high radiation dose procedures” if more than 5% of procedures result in K a,r exceeding 3 Gy. Obtaining a radiation consent for only a subset of procedures removes a clinical step prior to procedures where tissue effects are highly unlikely. However, it may complicate policy and process by requiring someone to create and keep updated a list of procedures in radiology, cardiology, surgery, and other specialties requiring radiation consent. An example of simple radiation consent language, adapted from the SIR's guidelines for patient dose management, can be found in Appendix A. Additional example consent language can be found in NCRP Report Nos. 168 and 185. , Any such document should be reviewed by appropriate clinical and legal teams prior to implementation. Key elements of consent language should include a description of the use of X‐rays in fluoroscopy, the possible tissue effects resulting from prolonged exposure to X‐rays, the typical time delays for these effects to occur, and the proposed action(s) for the patient and/or caregiver if any effect is observed.
Pre‐procedure screening In addition to screening for potential medical issues or pregnancy, patients who will undergo potentially high‐dose FGI procedures should be screened to determine if they are at higher risk for tissue effects. The result of pre‐procedure patient screening is then conveyed to the performing physician. Patients at higher risk include those who have had recent high‐dose fluoroscopy procedures or a history of radiation therapy to the same skin area, collagen vascular disease, or certain genetic disorders that affect DNA repair, which are further detailed in Ref. High body mass index (BMI) is also a risk factor since the greater amount of tissue increases the amount of radiation needed to yield an adequate image and can result in the skin being closer to the X‐ray source.
Intra‐procedure monitoring and notifications Intra‐procedural notifications regarding radiation dose levels allow the performing physician to gauge the benefit–risk ratio at each stage of an FGI procedure. All FGI procedures should be justified, that is, offer a clinical benefit to the patient greater than the potential risks, which include radiation tissue reactions. This benefit–risk ratio is considered by the physician when initially deciding whether to perform a procedure, and later, while the procedure is in progress. For the benefit–risk ratio to be meaningful, the physician needs to have an accurate understanding of radiation risk, the likelihood of tissue injury, and the associated dose–response relationship. The QMP, as the subject matter expert in this area, can provide radiation protection planning knowledge to the team to help ensure that the benefit–risk ratio is formulated correctly. The QMP must also understand the magnitude of radiation risk compared to other procedural risks, which are often much greater than the risk of tissue reaction. The risk of a radiation effect is typically much smaller than other procedural risks, such as bleeding, infection, and organ damage. However, the benefit–risk ratio is not static, and may change during the procedure. This analysis should be performed by a well‐informed operator and should be evaluated throughout the procedure. Procedures should rarely, if ever, be stopped solely due to radiation dose. If all or most progress made during the procedure is lost if the procedure is stopped (e.g., navigation of catheter to a very difficult site, or risk of developing collateral vasculature in the interval), any risks already incurred will have been for no benefit to the patient. When appropriate clinically, very complex procedures may be planned in a staged fashion, with multiple sessions separated by 8–10 weeks, to fractionate the dose to the skin and reduce the likelihood of tissue reactions. The concept and implementation of radiation dose notification levels is simple. The same concept has been applied to other potentially toxic agents, such as iodine contrast or medications. The implementation of notification levels requires the entire procedure team to work together. For example, the radiologic technologist, who is a local expert on the operation of fluoroscopy equipment, may be the one who calls out when the notification level is reached and documents this action in the procedure record. The nurse may then document the same information in the electronic medical record, while the operator or performing physician pauses for a moment to consider their radiation management strategy and the current benefit–risk pace of the procedure. The essential elements of an intra‐procedural dose notification are: Verbal notification to the operator regarding the current notification number and the magnitude of the dose index, for example, “This is the third notification. The current reference air kerma is 5000 mGy.” Consideration of a procedural pause to evaluate setup and the benefit–risk pace of the procedure, if such a pause will not interfere with the conduct of the procedure. Documentation of the notification level, that the operator was notified, that the setup and benefit–risk ratio of the procedure were evaluated, and any specific actions taken by the operator. The notification should be delivered during a natural pause during the procedure, if possible, and should never interrupt a critical phase of the procedure. Re‐consideration of radiation dose management strategy after a notification may include specific actions including, but not limited to: Adjustment of the patient table height. Sometimes it is set within the optimal range at the beginning of the procedure, but may be lowered during the procedure, for example, to isocenter the patient for volumetric imaging, after which the table is not moved back to the original height. Elimination of a large air gap by lowering the image receptor. Evaluation of X‐ray beam collimation and gantry angle. Evaluation of currently selected organ program (aka imaging protocol) for aspects such as dose rate setting, pulse rates, and DSA frame rate. Traditional radiation management practice often suggests that the angle of the gantry be varied during a procedure to “spread” radiation dose across the skin. While this strategy can be useful in specific circumstances as a prophylactic technique, it can be detrimental to radiation management and greatly increase both the skin dose rate and the PSD, especially in cases where larger oblique angles both increase radiation output and put the patient's skin surface closer to the tube. , , If used as a radiation management strategy, tight collimation of the X‐ray beam increases the benefit of this technique. Onboard, real‐time PSD mapping capabilities, if available, can help determine the benefit of these strategies. Until a time at which PSD is widely available as a real‐time displayed dose index, K a,r is the preferred dose index for dose notifications in fluoroscopy. However, as PSD is more closely correlated with the likelihood of tissue reaction, it should be used for intra‐procedural notifications when available. Conservatively, values for both planes in biplane procedures should be summed for the purpose of intra‐procedural notifications. Multiple publications exist with a consensus recommendation for setting notification levels with the first notification at 3 Gy K a,r and subsequent notifications every 1 Gy K a,r . , , , , However, there are practice‐specific considerations for setting notification levels, as well as patient‐specific considerations, such as the presence of sensitizing factors, that should be used to modify notification values when appropriate. Practice‐specific factors to consider when setting notification levels include: Types of procedures performed : Depending on the procedure type (e.g., neuro, body, cardiac) the fluoroscopy dose notification level could be adjusted in consideration of the relationship between K a,r and PSD. The geometry of the procedure affects the ratio of PSD to K a,r , denoted as the “dose index” by Miller et al. To avoid confusion, the “dose index” as described by Miller, will be referred to as the “PSD factor” in this report. For non‐isocentric procedures, such as those performed in vascular and interventional radiology, the PSD is often similar to the reported K a,r (PSD factor ∼ 1.0). For isocentric procedures, where the relevant anatomy is placed at isocenter to facilitate the use of multiple gantry angles, the skin dose will be higher than the K a,r if only a single projection, or a narrow range of projections, is used (PSD factor > 1.0). However, isocentric procedures often require the use of many different gantry angles, which may result in multiple distinct “fields” on the skin of the patient. This tends to reduce the dose index, and the PSD would be less than the K a,r (PSD factor < 1.0). Equipment capabilities : Many interventional fluoroscopes now offer the capability to program one or more fluoroscopy dose notification levels into the system. Upon reaching that notification level, an alarm sounds and/or a visual indication is displayed at the position of the operator. If, for example, a fluoroscope offers only three such levels at fixed values, this is a consideration when selecting the first, second, and third notification levels in policy. Different notification levels may be implemented for different services based on the type of procedures performed or the available dose indices. The QMP should ensure that any programmed dose notification levels are harmonized, though this can be complicated by differences in manufacturer's specific alert configurations. Alarm fatigue : The 5‐min fluoroscopy timer is often ignored in interventional fluoroscopy because it sounds numerous times during most procedures. Avoiding similar alarm fatigue with dose notification levels is an important consideration, both for the value of the first notification level and the intervals between notification levels. One other important intra‐procedural notification is the rapid recognition and correction of unsafe conditions when they exist. Examples of unsafe conditions include the presence of the patient's arm in the field of view when the arm is not the area of interest, and direct irradiation of the eye lens or female breast tissue when unnecessary. Every member of the procedural team should be confident in identifying such conditions when they exist and reporting them to the operator and responsible physician. Diagnostic medical physicists can provide maximum value to their healthcare organizations only when they understand the clinical aspects of fluoroscopically guided procedures. Fluoroscopy dose indices can only be understood completely when accompanied by the clinical context of the procedure, and imaging protocols for fluoroscopy can be optimized only when the tasks involved in completing the procedure are understood. For these reasons, it is important that a QMP who supports a clinical FGI practice be provided clinical instruction and time to observe clinical procedures, including how the equipment and protocols are used by physician operators. It is likely that this observation time can also be used to optimize protocols.
Establishing the SRDL and post‐procedure actions The substantial radiation dose level (SRDL) is an operational threshold for radiation dose above which additional post‐procedure actions for patient care should be taken due to the potential for biological harm. The SRDL should be set to a level such that a radiation dose below the SRDL is unlikely to result in a tissue injury for an average patient of normal radiation sensitivity. However, “there is no implication that a radiation level below an SRDL is absolutely safe or that a radiation level above an SRDL will always cause an injury,” and patient‐specific factors such as underlying conditions or medications could influence the threshold to induce a tissue reaction. Suggested values for the SRDL are 5 Gy K a,r or 3 Gy PSD. , , Since real‐time PSD estimates are not typically available on most equipment, use of K a,r is generally recommended to establish thresholds for post‐procedure follow‐up. When setting the K a,r SRDL value, the previously discussed PSD factor concept should be considered. If it is known that the entrance skin point is likely to be unchanged during a procedure or the skin is closer to the X‐ray source than the interventional reference point (IRP)/patient entrance reference point (PERP), a lower SRDL may be warranted. Similarly, for biplane procedures, the dose received from each plane should be added for SRDL purposes unless it is known that the fields do not overlap. Each facility must have a policy for identifying patients who receive a radiation dose exceeding the SRDL and for providing patient management and follow‐up, including provision of post‐procedure information written in simple language. This document (see Appendix B, adapted from NCRP Report No. 168 language, for an example) should provide educational information about the procedure, follow‐up information, and provider contact information for questions or concerns. This information can be provided in discharge instructions, in the electronic medical record, or both. Determination of the individual responsible for distributing these instructions is vital and should be part of the facility's policy. Technologists, nurses, physicians, or physician assistants can be assigned this duty. Further, the facility must have an established system for patient follow‐up and must document any such communication in the medical record. Follow‐up may be in‐person or via telemedicine and may reasonably be provided by any trained clinical team member, under the direction of the performing physician. All tissue effects should be assumed radiogenic until proven otherwise. If a severe or prolonged radiogenic tissue reaction is observed, the patient should be seen in person by the performing physician whenever possible and referred to dermatology, wound care, radiation oncology, or another appropriate specialty for further care, with all appropriate information included in the patient's medical record. PSD estimates may be useful for patient management in some cases. Facilities should have a defined process for requesting a PSD estimate performed by or under the supervision of a QMP. PSD estimates required by policy, regulation or accreditation standard, or requested from a licensed provider, must be documented in the patient medical record. The details of estimating PSD from dose indices are outside the scope of this document and have been described elsewhere. , Due to the various assumptions and inherent uncertainties involved, PSD estimates are unlikely to be better than ±50% accurate. This uncertainty should be included in any estimate and QMPs should always include a range of possible values in addition to the likely PSD value. For example, depending on various assumptions made regarding variable procedure aspects such as table height, collimation, and beam angulation, a PSD estimate could be documented as “likely 13 Gy, but with a probable range of 8–18 Gy.” As of January 2021, the American Medical Association (AMA) added a billable Current Procedural Terminology (CPT) code that refers to PSD estimates; code 76145 applies to evaluation of radiation exposure that exceeds institutional thresholds.
Cumulative doses and risk of tissue reactions There may be instances where a patient has multiple procedures in a short span of time where no single procedure reaches the threshold for patient follow‐up, but the summed doses warrant action. Setting an appropriate time window for a meaningful summation of doses is challenging. As previously mentioned, repair of sublethal radiation injury to skin is typically complete within a day of exposure, so doses delivered within a 24‐h period should invariably be summed. However, repopulation of damaged cells can require months. The SIR's guidelines for patient dose management suggest summing doses over a 60‐day period, while prior TJC sentinel event standards required summing doses over 6–12 months. , , This report recommends 60 days as the most biologically meaningful time window for summing patient skin dose. The same thresholds for follow‐up should be used for single or multiple cumulative procedures that exceed the SRDL. Identifying patients with high doses from multiple procedures, and their potential risk, is technically challenging. The patient's history of radiation exposures may be incomplete if some exposures occurred at outside facilities, or even in a different department within the same facility, if procedure information is not shared. Even with sophisticated dose management software that can track patient dose, there can be a time lag in identifying these patients, who may be discharged before they are identified as having exceeded SRDLs. In these cases, facility policy should identify how the patient will be contacted and who will be responsible for follow‐up. The treating physician is ultimately responsible for ensuring follow‐up is performed.
SPECIAL CONSIDERATION FOR PEDIATRIC AND PREGNANT PATIENTS—QMP INVOLVEMENT IN RISK ASSESSMENT As previously described, QMPs are generally the most knowledgeable individuals on radiation risk and play an important role in the fluoroscopic dose management process. This knowledge is critical in cases of pediatric or pregnant patients undergoing fluoroscopic procedures, as both can be at higher risk of radiation effects. On average, both the pediatric patient and the fetus are at higher risk for stochastic effects from radiation exposure due to more rapid cell growth and a comparatively greater number of undifferentiated cells compared with adults, as well as their longer remaining expected lifespan. In addition to these risks, each has the possibility of unique tissue reactions. Pediatric patients have shown a lower threshold for cataract development. High fetal doses during particular stages of development can result in loss of pregnancy, fetal organ malformation, or intellectual disability. In addition to the dose estimation roles discussed below, for both pediatric and pregnant patients, the QMP can play a role by providing direct counseling to the patient or guardian in order to provide information and address any potential radiation concerns. The principal role of the QMP in pediatric fluoroscopy should be dose management in patients, ensuring that the radiation dose is no greater than necessary to maintain image quality adequate for the clinical task. QMPs should ensure that clinical staff are aware of the magnitude of the radiation risks involved and any thresholds for effects. The QMP must consult with clinical teams to ensure that appropriate fluoroscopic equipment is used and protocols are optimized for pediatric patients. This is particularly important, as patients can range in size from premature infants weighing less than 1 pound to adolescents who exceed normal adult dimensions. Examples of protocol elements that the QMP should consider, when appropriate, for pediatric fluoroscopy are: Reduced pulse rate or width. Reduced tube current. Use of small focal spots for higher resolution imaging. Removal of the anti‐scatter grid in smaller patients. Selection of appropriate image‐processing parameters. Selection of any other appropriate patient dose reduction controls. Operators should be mindful of routine methods for dose reduction, such as proper attention to collimation, magnification, pulse rate, and source‐to‐skin and source‐to‐image‐receptor distance. Regular use of pre‐procedure checklists, such as the one provided by Image Gently, can help the QMP operationalize these behaviors. For fluoroscopy of a pregnant patient and fetus, the QMP plays a similar role. There may also be instances in which FGI procedures are performed on a woman with an unknown pregnancy. In these cases, the QMP will often be involved in estimating fetal dose. Details of fetal dose estimate methods can be found in Wagner et al. When fluoroscopic imaging is necessary on a known pregnant patient, clinicians should consult with the QMP to estimate the radiation dose and potential risk to the fetus. Procedure planning should include a consideration of imaging protocol optimization, collimation, and the relationship between fetal dose and displayed dose indices. Additionally, the QMP may be asked to be available during the procedure for consultation and to make sure adequate information is obtained for a fetal dose estimate. More specific fetal dose estimates can be obtained with the use of thermoluminescent dosimeter (TLDs) or Optically stimulated luminescence dosimeter (OSLDs), as described in Dauer et al. With appropriate preparation, complex interventional procedures in the abdomen and pelvis, such as renal or trauma embolizations, can often be completed with relatively low fetal doses.
FLUOROSCOPIC EQUIPMENT EVALUATIONS Certain aspects of acceptance testing, commissioning, and periodic acceptability testing of fluoroscopy systems deal with patient and staff radiation dose management. Such aspects include, but are not limited to, scatter survey and comparison to manufacturer‐provided isokerma plots, measurement of table and pad transmission factors, measurement of typical and maximum air kerma rates for fluoroscopy and acquisition modes of operation, imaging protocol design and review, and measurement of the accuracy of machine‐reported dose indices. Accreditation and regulatory requirements may require additional tests and establish minimum standards and frequencies for such evaluations. AAPM Task Group reports, Medical Physics Practice Guidelines (MPPG), and American College of Radiology (ACR) Practice Parameters may provide guidance on test methods, expected values, and the role of the QMP and other personnel in performing these evaluations. , , , , It is also important to establish performance baselines during acceptance testing, compare these baselines to manufacturer specifications, where applicable, and to compare future measurement values to established baselines.
THE JOINT COMMISSION FLUOROSCOPY REQUIREMENTS TJC, a healthcare accrediting body in the United States, has instituted standards on the use of fluoroscopy and on the management of radiation exposures to the patients of accredited facilities. Effective 1 January 2019, healthcare organizations accredited by TJC are required to meet new “elements of performance” related to fluoroscopy, including annual equipment performance evaluations, documentation of procedure dose indices, setting SRDLs, and review of procedures exceeding the SRDL. , , , , These requirements are in addition to TJC's updated fluoroscopy sentinel event, which requires root cause analysis review if permanent tissue injury results from improperly performed procedures.. The QMP is an integral part of the healthcare team tasked with meeting these requirements. TJC rescinded pre‐publication standards that specifically required annual radiation dose optimization training for fluoroscopy operators. 8.1 Annual equipment evaluation TJC standards (Element of Performance No. 34, EC.02.04.03) state that at least annually a diagnostic medical physicist must conduct a performance evaluation of all fluoroscopic imaging equipment. This requirement excludes equipment used for therapeutic radiation treatment planning or delivery, but includes FGI suites, over‐ and under‐table fluoroscopy rooms, and mobile and mini C‐arm fluoroscopes. AAPM MPPG 10a suggests that all fluoroscopic performance evaluations be performed by a QMP, though according to TJC, they may be assisted by individuals with the required skills, as determined by the QMP. , 8.2 Dose index documentation TJC standard (Element of Performance #30, PC.02.01.01) states, “The reference‐air kerma, cumulative‐air kerma, or kerma‐area product are [sic] documented in a retrievable format … such as a picture archiving and communication system.” (Note that “reference air kerma” and “CAK” are two terms for the same dose index: K a,r .) If a system does not display the CAK or KAP, the fluoroscopy time, mode of operation, and number of images should be documented instead, in a retrievable format. According to a clarification from TJC in January 2019, this element of performance does not apply to fluoroscopy equipment used for therapeutic radiation treatment planning or fluoroscopy equipment classified as a mini C‐arm. Documentation is still required for other non‐FGI fluoroscopes such as full‐sized mobile C‐arms, remote, and tableside fluoroscopy systems, despite the fact that there can be considerable cost involved in implementing such a system and questionable benefit as the entrance skin dose resulting from these systems are typically very low. A facility's choice of method for documenting fluoroscopy dose indices depends on the equipment in use, the size of the healthcare system, and the resources and technology available. Possible solutions include: Archival of fluoroscope produced dose information into picture archiving and communication system (PACS). Manual logs. Manual entry into permanent patient records, such as patient electronic medical records (EMR), the PACS, and hospital or radiology information systems (HIS and RIS). Automatic radiation dose index monitoring software. The QMP, as a subject matter expert, is often involved in helping facilities comply with these requirements. These solutions require a collaborative approach between medical physics and hospital Information Technology (IT) and/or PACS administrators, and each facility needs to determine which approach is best, based on the capabilities of their equipment and available technology. One solution to documenting dose index data is to rely on RDSRs and dose summary pages produced by the modality, which are often sent to PACS along with any procedure images. While this approach is appealing in that it is simple and requires no additional software systems or manual recording of data, it has several potential disadvantages. Procedure dose summaries and RDSRs are common on modern FGI equipment, but legacy, general, or mobile fluoroscopes might not produce them, necessitating additional documentation methods. The dose summary approach can also result in fragmented storage locations, as dose data reside within individual patient PACS records instead of a central location. Often, various hospital services utilize entirely separate PACS systems, which can further complicate data retrieval. Storage of dose index data in multiple locations can hinder the ability to aggregate dose indices from multiple procedures or to do facility dataset review for quality improvement purposes as discussed later in this report. Paper logs are the simplest and least expensive to implement and meet the recommendations but present possible Health Insurance Portability and Accountability Act (HIPAA) issues due to protected health information (PHI) required for retrieval, as well as potential accuracy and legibility issues common to manual records. Additionally, paper records, if used, do not lend themselves easily to data analysis. Having dose index, equipment, and operator information in a digital format makes it possible to better evaluate dose indices and equipment use, in addition to allowing easier auditing of operators’ habits. For these reasons, this MPPG recommends paper logs be used for fluoroscopy dose index tracking only when digital formats are not possible. Manual entry into digital formats, such as spreadsheets, can avoid some potential HIPAA issues and can be better used for data analysis, but may still suffer accuracy issues from manual entry. Fluoroscopy procedure information entered into digital formats has the advantage of being inherently electronic and can often be directly linked to sources such as HIS, RIS, or HL7 feeds, which can automatically populate patient information with high fidelity. This method also avoids potential HIPAA issues, as these systems are generally much more secure than physical notebooks or spreadsheets stored on network drives. Reports including patient and procedure information, equipment operator, and dose indices can be easily generated from these data, which can be useful for auditing and analysis, though manual entry can lead to data errors and false alarms. Specific methods of data entry can vary. Since fluoroscopy equipment is often operated outside of radiology departments, access to software systems should be considered. Data entry into systems such as a RIS may not be possible in hospital departments outside of Radiology, possibly necessitating multiple avenues of data entry for a centralized dose index database. The QMP will likely need to work closely with hospital IT and/or PACS administrators in order to set up such a system. Automatic radiation dose index monitoring systems can eliminate most manual entry and automatically populate a database with patient and dose index information, greatly increasing data fidelity. This functionality is dependent on the fluoroscopy equipment in use. While most new FGI equipment can produce RDSRs that are sent directly from the modality, much of the fluoroscopy equipment used today does not support RDSR functionality and may not be integrated easily into commercial radiation dose index monitoring systems. Specifically, it may be difficult to get dose information from legacy equipment, R/F rooms, and mobile or mini C‐arms into these systems, necessitating a separate documentation system for those data. An additional challenge is connectivity between dose index monitoring systems and patient electronic medical records, leading to lack of access to patient dose data for referring physicians or interventionalists who perform FGI procedures. Better integration of these systems would improve patient care by providing immediately available radiation dose information to physicians. Regardless of the specific method for storage, regular monitoring and maintenance of the radiation dose index database is critical to its long‐term success. New staff, new devices, and software upgrades have the potential to disrupt the methods used to document radiation dose indices. Even without changes to the hospital workflow, missing or incomplete data are possible. Regular system monitoring and maintenance allows for adjustments as needed to ensure that the quality of radiation dose data is maintained and documented. Again, the QMP is a likely candidate for this task, though time should be allocated for this work, which can take considerable effort. Most facilities will adopt an iterative approach and continue to improve their data capture process over time. 8.3 Identification of radiation exposure thresholds TJC standard (Element of Performance #30, PC.02.01.01) requires a facility to “identify radiation exposure and skin dose threshold levels, that if exceeded, trigger further review and/or patient evaluation to assess for adverse radiation effects.” TJC does not provide any specific recommendations for threshold levels but does refer to NCRP Report No. 168. Prior discussion of SRDLs in this document addresses recommendations for complying with this standard. The QMP should be involved in setting these threshold levels and drafting policies. 8.4 Reviewing and analyzing procedures over threshold TJC standard (Element of Performance #20, PI.02.01.01) requires that the organization providing fluoroscopy services “review and analyze instances where the radiation exposure and skin dose threshold levels identified by the organization are exceeded.” TJC does not limit this threshold review to FGI procedures but includes all fluoroscopic services. TJC does not specify how this review and analysis is to be accomplished, only that it be done. Previous discussion in this document regarding appropriate patient follow‐up for procedures above the SRDL is presumed to be sufficient. For policy purposes, the prior recommendations in this document may be followed with the FGI SRDL applied to lower dose fluoroscopic equipment as well (mobile C‐arm fluoroscopic units for use in operating rooms, general fluoroscopic units used for low‐dose diagnostic studies, etc.). Procedures completed with this equipment typically have very low K a,r and it is unlikely that any non‐interventional procedure will have a K a,r above the SRDL. 8.5 Sentinel event requirement TJC adopted a sentinel event policy in 1996 for monitoring patient safety events that lead to death, permanent harm, or severe temporary harm, and that are not related to the natural course of a patient's underlying illness or condition. In 2005, TJC added “radiation overdose” as a reviewable sentinel event, which, in addition to the delivery of radiotherapy to the wrong region or >25% above the planned dose, included “Prolonged fluoroscopy with cumulative dose >1500 rads [15 Gy] to a single field.” In a subsequent publication, the TJC clarified that monitoring cumulative fluoroscopy PSD over a period of 6 months to 1 year would be reasonable, ultimately leaving the decision of determining the cumulative dose monitoring time window to the accredited institution. If a sentinel event occurs, TJC requires conducting a comprehensive systematic analysis, identifying causal and contributory factors, and documenting a corrective action plan, as well as a recommendation (but not requirement) to report the event to TJC. The most common approach to this analysis is a Root Cause Analysis (RCA), which needs to be completed within 45 days of becoming aware of the event. Additionally, in order to be considered a credible analysis, the RCA is required to include senior healthcare organization leadership. Practical implementation of the “radiation overdose” sentinel event has been difficult. Compliance required knowledge of PSD, an index not widely available on fluoroscopic equipment. For the majority of fluoroscopy systems that do not report PSD, estimations of patient and procedure‐specific PSD are laborious and prone to uncertainty. Even with all information available, skin dosimetry estimates are unlikely to be more accurate than ±50%. Due to PSD uncertainty and variation in individual radiosensitivity, it is possible for patients to develop severe tissue reactions at estimated PSDs below the 15 Gy sentinel event definition. Conversely, it is possible for patients with estimated PSD greater than 15 Gy to experience no tissue effects. The “radiation overdose” fluoroscopy sentinel event was unique among TJC sentinel events. Other sentinel events include patient suicide, abduction, or elopement leading to death or serious harm, unanticipated death of an infant, discharge of an infant to the wrong family, wrong site or patient surgeries or radiotherapy, or assault or homicide of patients or staff. All of these can reasonably be described as preventable events that should never occur in the normal provision of healthcare services. While high radiation tissue doses can result in temporary or permanent harm, a PSD exceeding 15 Gy does not necessarily indicate that standards of care were not upheld. Some complicated FGI procedures require very large radiation skin doses to complete, even when justified and optimized. There is also wide variation in patient size and lesion characteristics for some FGI procedures, which may result in wide ranges of PSD. This is not to say that high skin doses should not be investigated, or that it is not possible that severe tissue reactions could, in some cases, have been avoided by better practice. However, investigating a “radiation overdose” sentinel event often led to large expenditures of time and resources in pursuit of RCA of properly performed procedures, and potentially excluded investigation of serious tissue reactions that occurred at a PSD below the rigidly defined PSD threshold. Due to these limitations, TJC made changes to the fluoroscopy sentinel event requirements in 2021, with changes going into effect in January 2022. The updated requirement now defines a sentinel event as “ Fluoroscopy resulting in permanent tissue injury when clinical and technical optimization were not implemented and/or recognized practice parameters were not followed.”
Annual equipment evaluation TJC standards (Element of Performance No. 34, EC.02.04.03) state that at least annually a diagnostic medical physicist must conduct a performance evaluation of all fluoroscopic imaging equipment. This requirement excludes equipment used for therapeutic radiation treatment planning or delivery, but includes FGI suites, over‐ and under‐table fluoroscopy rooms, and mobile and mini C‐arm fluoroscopes. AAPM MPPG 10a suggests that all fluoroscopic performance evaluations be performed by a QMP, though according to TJC, they may be assisted by individuals with the required skills, as determined by the QMP. ,
Dose index documentation TJC standard (Element of Performance #30, PC.02.01.01) states, “The reference‐air kerma, cumulative‐air kerma, or kerma‐area product are [sic] documented in a retrievable format … such as a picture archiving and communication system.” (Note that “reference air kerma” and “CAK” are two terms for the same dose index: K a,r .) If a system does not display the CAK or KAP, the fluoroscopy time, mode of operation, and number of images should be documented instead, in a retrievable format. According to a clarification from TJC in January 2019, this element of performance does not apply to fluoroscopy equipment used for therapeutic radiation treatment planning or fluoroscopy equipment classified as a mini C‐arm. Documentation is still required for other non‐FGI fluoroscopes such as full‐sized mobile C‐arms, remote, and tableside fluoroscopy systems, despite the fact that there can be considerable cost involved in implementing such a system and questionable benefit as the entrance skin dose resulting from these systems are typically very low. A facility's choice of method for documenting fluoroscopy dose indices depends on the equipment in use, the size of the healthcare system, and the resources and technology available. Possible solutions include: Archival of fluoroscope produced dose information into picture archiving and communication system (PACS). Manual logs. Manual entry into permanent patient records, such as patient electronic medical records (EMR), the PACS, and hospital or radiology information systems (HIS and RIS). Automatic radiation dose index monitoring software. The QMP, as a subject matter expert, is often involved in helping facilities comply with these requirements. These solutions require a collaborative approach between medical physics and hospital Information Technology (IT) and/or PACS administrators, and each facility needs to determine which approach is best, based on the capabilities of their equipment and available technology. One solution to documenting dose index data is to rely on RDSRs and dose summary pages produced by the modality, which are often sent to PACS along with any procedure images. While this approach is appealing in that it is simple and requires no additional software systems or manual recording of data, it has several potential disadvantages. Procedure dose summaries and RDSRs are common on modern FGI equipment, but legacy, general, or mobile fluoroscopes might not produce them, necessitating additional documentation methods. The dose summary approach can also result in fragmented storage locations, as dose data reside within individual patient PACS records instead of a central location. Often, various hospital services utilize entirely separate PACS systems, which can further complicate data retrieval. Storage of dose index data in multiple locations can hinder the ability to aggregate dose indices from multiple procedures or to do facility dataset review for quality improvement purposes as discussed later in this report. Paper logs are the simplest and least expensive to implement and meet the recommendations but present possible Health Insurance Portability and Accountability Act (HIPAA) issues due to protected health information (PHI) required for retrieval, as well as potential accuracy and legibility issues common to manual records. Additionally, paper records, if used, do not lend themselves easily to data analysis. Having dose index, equipment, and operator information in a digital format makes it possible to better evaluate dose indices and equipment use, in addition to allowing easier auditing of operators’ habits. For these reasons, this MPPG recommends paper logs be used for fluoroscopy dose index tracking only when digital formats are not possible. Manual entry into digital formats, such as spreadsheets, can avoid some potential HIPAA issues and can be better used for data analysis, but may still suffer accuracy issues from manual entry. Fluoroscopy procedure information entered into digital formats has the advantage of being inherently electronic and can often be directly linked to sources such as HIS, RIS, or HL7 feeds, which can automatically populate patient information with high fidelity. This method also avoids potential HIPAA issues, as these systems are generally much more secure than physical notebooks or spreadsheets stored on network drives. Reports including patient and procedure information, equipment operator, and dose indices can be easily generated from these data, which can be useful for auditing and analysis, though manual entry can lead to data errors and false alarms. Specific methods of data entry can vary. Since fluoroscopy equipment is often operated outside of radiology departments, access to software systems should be considered. Data entry into systems such as a RIS may not be possible in hospital departments outside of Radiology, possibly necessitating multiple avenues of data entry for a centralized dose index database. The QMP will likely need to work closely with hospital IT and/or PACS administrators in order to set up such a system. Automatic radiation dose index monitoring systems can eliminate most manual entry and automatically populate a database with patient and dose index information, greatly increasing data fidelity. This functionality is dependent on the fluoroscopy equipment in use. While most new FGI equipment can produce RDSRs that are sent directly from the modality, much of the fluoroscopy equipment used today does not support RDSR functionality and may not be integrated easily into commercial radiation dose index monitoring systems. Specifically, it may be difficult to get dose information from legacy equipment, R/F rooms, and mobile or mini C‐arms into these systems, necessitating a separate documentation system for those data. An additional challenge is connectivity between dose index monitoring systems and patient electronic medical records, leading to lack of access to patient dose data for referring physicians or interventionalists who perform FGI procedures. Better integration of these systems would improve patient care by providing immediately available radiation dose information to physicians. Regardless of the specific method for storage, regular monitoring and maintenance of the radiation dose index database is critical to its long‐term success. New staff, new devices, and software upgrades have the potential to disrupt the methods used to document radiation dose indices. Even without changes to the hospital workflow, missing or incomplete data are possible. Regular system monitoring and maintenance allows for adjustments as needed to ensure that the quality of radiation dose data is maintained and documented. Again, the QMP is a likely candidate for this task, though time should be allocated for this work, which can take considerable effort. Most facilities will adopt an iterative approach and continue to improve their data capture process over time.
Identification of radiation exposure thresholds TJC standard (Element of Performance #30, PC.02.01.01) requires a facility to “identify radiation exposure and skin dose threshold levels, that if exceeded, trigger further review and/or patient evaluation to assess for adverse radiation effects.” TJC does not provide any specific recommendations for threshold levels but does refer to NCRP Report No. 168. Prior discussion of SRDLs in this document addresses recommendations for complying with this standard. The QMP should be involved in setting these threshold levels and drafting policies.
Reviewing and analyzing procedures over threshold TJC standard (Element of Performance #20, PI.02.01.01) requires that the organization providing fluoroscopy services “review and analyze instances where the radiation exposure and skin dose threshold levels identified by the organization are exceeded.” TJC does not limit this threshold review to FGI procedures but includes all fluoroscopic services. TJC does not specify how this review and analysis is to be accomplished, only that it be done. Previous discussion in this document regarding appropriate patient follow‐up for procedures above the SRDL is presumed to be sufficient. For policy purposes, the prior recommendations in this document may be followed with the FGI SRDL applied to lower dose fluoroscopic equipment as well (mobile C‐arm fluoroscopic units for use in operating rooms, general fluoroscopic units used for low‐dose diagnostic studies, etc.). Procedures completed with this equipment typically have very low K a,r and it is unlikely that any non‐interventional procedure will have a K a,r above the SRDL.
Sentinel event requirement TJC adopted a sentinel event policy in 1996 for monitoring patient safety events that lead to death, permanent harm, or severe temporary harm, and that are not related to the natural course of a patient's underlying illness or condition. In 2005, TJC added “radiation overdose” as a reviewable sentinel event, which, in addition to the delivery of radiotherapy to the wrong region or >25% above the planned dose, included “Prolonged fluoroscopy with cumulative dose >1500 rads [15 Gy] to a single field.” In a subsequent publication, the TJC clarified that monitoring cumulative fluoroscopy PSD over a period of 6 months to 1 year would be reasonable, ultimately leaving the decision of determining the cumulative dose monitoring time window to the accredited institution. If a sentinel event occurs, TJC requires conducting a comprehensive systematic analysis, identifying causal and contributory factors, and documenting a corrective action plan, as well as a recommendation (but not requirement) to report the event to TJC. The most common approach to this analysis is a Root Cause Analysis (RCA), which needs to be completed within 45 days of becoming aware of the event. Additionally, in order to be considered a credible analysis, the RCA is required to include senior healthcare organization leadership. Practical implementation of the “radiation overdose” sentinel event has been difficult. Compliance required knowledge of PSD, an index not widely available on fluoroscopic equipment. For the majority of fluoroscopy systems that do not report PSD, estimations of patient and procedure‐specific PSD are laborious and prone to uncertainty. Even with all information available, skin dosimetry estimates are unlikely to be more accurate than ±50%. Due to PSD uncertainty and variation in individual radiosensitivity, it is possible for patients to develop severe tissue reactions at estimated PSDs below the 15 Gy sentinel event definition. Conversely, it is possible for patients with estimated PSD greater than 15 Gy to experience no tissue effects. The “radiation overdose” fluoroscopy sentinel event was unique among TJC sentinel events. Other sentinel events include patient suicide, abduction, or elopement leading to death or serious harm, unanticipated death of an infant, discharge of an infant to the wrong family, wrong site or patient surgeries or radiotherapy, or assault or homicide of patients or staff. All of these can reasonably be described as preventable events that should never occur in the normal provision of healthcare services. While high radiation tissue doses can result in temporary or permanent harm, a PSD exceeding 15 Gy does not necessarily indicate that standards of care were not upheld. Some complicated FGI procedures require very large radiation skin doses to complete, even when justified and optimized. There is also wide variation in patient size and lesion characteristics for some FGI procedures, which may result in wide ranges of PSD. This is not to say that high skin doses should not be investigated, or that it is not possible that severe tissue reactions could, in some cases, have been avoided by better practice. However, investigating a “radiation overdose” sentinel event often led to large expenditures of time and resources in pursuit of RCA of properly performed procedures, and potentially excluded investigation of serious tissue reactions that occurred at a PSD below the rigidly defined PSD threshold. Due to these limitations, TJC made changes to the fluoroscopy sentinel event requirements in 2021, with changes going into effect in January 2022. The updated requirement now defines a sentinel event as “ Fluoroscopy resulting in permanent tissue injury when clinical and technical optimization were not implemented and/or recognized practice parameters were not followed.”
THE NCRP STATEMENT 11 PROCESS NCRP Statement No. 11 provides an administrative process for facilities to manage certain adverse events from FGI procedures in line with the new TJC sentinel event definition, which triggers a review based on identified tissue reactions rather than a PSD threshold. Since tissue reactions can occur at PSD below the previous 15 Gy threshold, the Statement No. 11 process factors in an individual patient's radiosensitivity and applies to procedures that would not previously rise to the level of a sentinel event. By focusing on patient outcomes rather than PSD thresholds, the NCRP process also avoids the ambiguous time frame for multi‐procedure dose accumulation in the TJC's previous sentinel event definition. It also eliminates any ambiguity as to whether an estimated PSD value with large uncertainty requires a full RCA. NCRP Statement No. 11 recommends that facilities develop a quality assurance and peer‐review (QA‐PR) program. This program promotes radiation management by tracking and reviewing available radiation dose indices periodically and triggering patient follow‐up when specified threshold values are exceeded, as previously discussed in this document. The program includes a QA‐PR committee, which evaluates radiation management for FGI procedures based on clinical and dosimetric data. This committee is composed of a QMP and professional practitioners so that it is competent to evaluate the clinical appropriateness and relevance of quality and safety matters. This committee would then suggest and implement corrective actions as needed. In instances where patient follow‐up results in a suspected clinically important radiogenic tissue reaction, this QA‐PR committee would evaluate the procedure, including clinical and dosimetric data, to determine if it met recognized practice parameters, using the following criteria, where applicable: Clinical justification of the procedure. Proper pre‐procedural review and evaluation of the patient's past FGI encounters for skin injury. Proper discussion of the potential for tissue reactions during informed consent process. Appropriate use of radiation during the procedure. Appropriate post‐procedure patient follow‐up. According to the NCRP, the possible outcomes of the QA‐PR evaluation are: The tissue reaction was detected through follow‐up and likely unavoidable. No action required. Clinical or technical optimization might have reduced the severity or improved in the detection of the reaction, but overall practice criteria were met. Methods for optimization should be implemented. Radiation use did not meet recognized practice parameters. A clinically important tissue reaction was potentially avoidable, its severity could have been minimized, or it was not detected. Corrective action is required. Statement No. 11 recommends that an RCA be undertaken only in the last case, if one or more practice parameter criteria were not met. This initial QA‐PR review eliminates the need for a full RCA with executive administration involvement for procedures where appropriate clinical care was provided. The initial goals of TJC's “radiation overdose” sentinel event were proper identification and follow‐up of patients with the potential for tissue effects, and the identification of instances where high tissue doses were not justified. It is the opinion of this group that these goals are better accomplished under the new sentinel event requirement and by following the recommendations of NCRP Statement No. 11, which makes better use of clinical resources to address TJC's concern of undiagnosed radiation‐induced tissue effects. Under this system, a PSD calculation by a QMP is only needed to direct proper dermatological care in cases of a known severe reaction, or as required by the facility's established process for PSD evaluation. Radiation fields should be assumed to be overlapping unless evidence suggests otherwise. Readers are again referred to Jones and Pasciak and AAPM TG 357 for more detailed discussion of performing PSD estimates. In cases where radiation use did not meet recognized practice parameters, an RCA, including the QMP and members of hospital executive administration, must be performed as described by TJC.
SUGGESTED STATE REGULATIONS The CRCPD maintains guidance that states may use when drafting regulations regarding the safe use of ionizing radiation through its Suggested State Regulations for Control of Radiation (SSRCR, or more commonly, SSR). The SSRs are adopted without modification by some states as regulations. The following discussion is based on Part F of the SSRs “Medical Diagnostic and Interventional X‐ray and Imaging Systems.” 10.1 Radiation Protocol Committee The SSRs recommend the creation of a Radiation Protocol Committee (RPC) responsible for ensuring “that exams being performed achieve the desired diagnostic image quality at the lowest radiation dose possible while properly exploiting the capabilities of the equipment being used.” The SSRs state that the RPC includes, at a minimum, a supervising physician, a QMP, and a lead technologist. This MPPG recommends the following specific responsibilities for each of these individuals: Supervising physician: The supervising physician, a physician who performs FGI procedures, has the responsibility for overseeing the activities of the RPC. This physician also works with the lead technologist to develop and maintain imaging protocols. When necessary, this physician serves as the liaison to other physicians for issues related to patient safety, such as long fluoroscopy times compared to peers. QMP: The QMP brings a knowledge of radiation safety and dosimetry to the RPC. The QMP is responsible for testing, or overseeing testing, of the FGI equipment to ensure it operates safely. The QMP also provides PSD estimates when necessary and advises on the impact of the administered radiation dose on the patient. Lead technologist: The lead technologist is responsible for maintaining imaging protocols under the direction of the supervising physician and QMP. This technologist advises the other RPC members of equipment issues or other observations that may affect patient dose or image quality. When PSD estimates are required, the lead technologist assists the QMP in gathering the necessary data. The organizational structure of the healthcare system will dictate the optimal committee arrangement. For a stand‐alone facility, if different medical specialties utilize FGI equipment (e.g., interventional cardiology, interventional radiology, vascular surgery), the facility should consider including physicians and technologists from each specialty. Other possible committee members are department managers, health system risk managers, and the radiation safety officer (RSO). Other individuals may be added to the committee as deemed necessary. If a larger healthcare system has more than one facility, a system‐wide committee may be established to ensure consistency among facilities. In this structure, one RPC may be formed, provided each facility within the enterprise has appropriate representation on the committee. Another option is to add the scope of the RPC into the responsibilities of an already established Radiation Safety Committee, as long as the recommended RPC membership is met. Per the SSRs, the RPC is responsible for establishing procedures and protocols to be followed before, during, and after FGI procedures. The SSRs state that protocols addressing the following are reviewed by the committee at least annually: Authorized users of FGI equipment. Intra‐procedure patient radiation dose monitoring. Dose notification levels. Establishment of SRDL values. Actions to be taken when a SRDL is exceeded. These protocols are discussed elsewhere in this report. The SSRs only require the RPC to create and implement policies and does not require committee oversight of clinical data. However, given the membership, it is the recommendation of this MPPG that such review fall under RPC oversight. This review includes evaluation of clinical and dosimetric data from any procedure in which the SRDL is exceeded or a tissue reaction has occurred, as well as analysis of facility or system case volume and dose index datasets, described in a later section. These datasets can include case volume and comparative dosimetry. The duties of the RPC could also be extended to include the previously discussed QA‐PR functions. In doing so, the committee reviews procedure justification, patient‐specific factors, radiation dose optimization, the time course over which radiation doses were administered, disease severity, and procedure complexity. Depending upon the size of the healthcare system, initial oversight might better be accomplished by department‐level committees. Under such a framework, each individual department (e.g., interventional cardiology, interventional radiology, vascular surgery), would have its own committee to oversee clinical and dosimetric data, and to perform QA‐PR review of the clinical appropriateness of procedures resulting in tissue effects. The results of this department‐level review would then be sent to the facility or system‐wide RPC for final review and approval. The healthcare system determines the frequency of RPC meetings. The SSRs suggest that the committee meet at least annually. More frequent meetings may allow for more rapid identification of potential problems and swifter implementation of subsequent changes in practice. Considerations that affect meeting frequency include the duties of the committee, the volume and complexity of FGI procedures performed, the number of physicians and medical staff involved, and the capabilities of the imaging equipment. Of these, procedure volume and complexity may be the most important, as both play a role in the number of potential procedures that may require committee review, if such review is tasked to the RPC. Some events may prompt the need for ad hoc meetings, such as sentinel events or serious patient injury review, per the previous NCRP Statement 11 QA‐PR discussion. The SSRs suggest that the RPC provide an annual report to the radiation safety committee, or, if the facility does not have a radiation safety committee, to the RSO. Although not addressed in the SSRs, this MPPG recommends that the annual report contain the following: A list of individuals authorized to use fluoroscopic equipment for FGI procedures. The total number of procedures that exceeded the SRDL and, if available, percentage of total procedures. A review of all policies and procedures and any significant changes. 10.2 Procedures for maintaining records The SSRs also recommend that all available radiation dose indices be recorded to facilitate skin dose estimation, if needed. These recommendations are similar to TJC requirements previously discussed, and the role of the QMP is to provide guidance on what specific information must be recorded for skin dose estimates. If RDSRs are produced by a fluoroscope, they should be archived, if possible, and the QMP must have access to them when needed. Where the RSDR is not available, the QMP may request that additional information be recorded, such as gantry angles, fluoroscopic technique factors, table height, patient dimensions, etc. Appropriate patient follow‐up procedures, including skin dose estimation, are performed according to policies set forth by the committee. 10.3 Facility dataset review for quality improvement Beyond CRCPD recommendations and TJC standards to review and analyze procedures over threshold values, collection of fluoroscopic dose index data provides facilities with quality improvement opportunities, as these data can be used to evaluate clinical practice. This type of analysis is useful because dose to the patient in fluoroscopy is highly dependent on the operator of the equipment. Since the QMP often oversees the collection of fluoroscopic data, they are well positioned to handle the task of data analysis. As detailed in NCRP Report 168, these data can be used to build “facility datasets” for each type of exam performed in a practice that can be compared among sites or to normative datasets, such as the American College of Radiology Fluoroscopy Dose Index Registry (DIR). These analyses can provide valuable insight into how the equipment is being used clinically and identify deviations from expected practice. NCRP Report No. 168 also recommends calculating the percentages of each procedure that exceed the SRDL set by the facility. This process can help determine those procedures that are more likely to result in high doses to patients. This information has the potential to affect the pre‐procedure consent process and room assignment decisions. Fluoroscopy dose index data may also be used to set procedure‐specific dose index review levels for evaluating clinical practice. These are not to be confused with the SRDL, which is a biologically based threshold related to the potential for tissue effects. Procedure‐specific review levels represent target achievable dose indices for specific procedures and can be useful for comparing clinical practice. Beyond reported indices, collected data can be used to derive additional metrics for practice insight. Dividing K a,r by total fluoroscopy time yields a measure of radiation utilization efficiency for a particular procedure, assuming fluoroscopy time is in some way related to the complexity of the procedure. For two procedures of similar complexity and fluoroscopy time, a procedure with higher utilization of acquisition or high‐level control modes would yield a higher value. Additionally, total P KA and K a,r data can be used to calculate the average field size ( P KA / K a,r ) at the IRP/PERP for a given procedure, thus permitting comparison of an operator's use of collimation. Beyond individual procedures, these derived metrics can be calculated for a fluoroscope or operator over a given month or quarter to assess and compare longer term operational trends. Values of any metric found to be substantially different from facility benchmarks can be investigated for potential protocol optimization or practice review. QMPs need to be mindful of the units used for a given dose index when comparing data. While K a,r is often reported in units of mGy, displayed P KA values vary by manufacturer and even by model, necessitating unit conversion for valid comparisons. However, differences in procedural dose index data do not necessarily indicate improper practice. It is likely that more experienced physicians perform a larger share of more complex procedures, which would affect their data. A similar situation can arise with equipment, where a newer Interventional Radiology (IR) suite with more features is rightfully utilized more often for complex procedures than older equipment. Additionally, patient body habitus and anatomy can drastically affect dose indices. Individual procedures may exceed SRDLs not because they were performed improperly, but because larger patients or complicated vascular anatomy require higher tube outputs or longer procedure time. The results of this type of analysis can be presented to operators or department chairs on a regular basis, either through a fluoroscopic RPC or dose review committee, or through other means such as existing staff meetings, radiation safety committee meetings, QA‐PR committee, or morbidity and mortality (M&M) conferences. This feedback can be an important part of practice improvement. Beyond comparing data within a practice, comparison to outside practices may be possible through programs such as the ACR Fluoroscopy DIR. The ACR Fluoroscopy DIR is the latest addition to the National Radiology Data Registry, and the first modality to join computed tomography (CT) in the DIR. The Fluoroscopy DIR provides a continuously updated normative dataset for fluoroscopically guided procedures, which can be used as an advisory dataset to which participating sites can compare their facility datasets of fluoroscopy dose indices. The ability of a hospital to perform this kind of analysis will be heavily dependent on the resources available. If full procedure records, including the exam type, fluoroscope used for the procedure, operator, and available dose indices are electronically captured, datasets can be created and evaluated with relative ease. However, this level of investigation is not practical if paper logs are used to meet documentation requirements. If resources are limited, analysis should focus on higher dose FGI procedures, as they present the greatest radiation risk. However, if fluoroscopy data for all equipment are being collected electronically, further evaluations for basic fluoroscopy rooms and mobile C‐arms and their operators can be performed without significant additional effort.
Radiation Protocol Committee The SSRs recommend the creation of a Radiation Protocol Committee (RPC) responsible for ensuring “that exams being performed achieve the desired diagnostic image quality at the lowest radiation dose possible while properly exploiting the capabilities of the equipment being used.” The SSRs state that the RPC includes, at a minimum, a supervising physician, a QMP, and a lead technologist. This MPPG recommends the following specific responsibilities for each of these individuals: Supervising physician: The supervising physician, a physician who performs FGI procedures, has the responsibility for overseeing the activities of the RPC. This physician also works with the lead technologist to develop and maintain imaging protocols. When necessary, this physician serves as the liaison to other physicians for issues related to patient safety, such as long fluoroscopy times compared to peers. QMP: The QMP brings a knowledge of radiation safety and dosimetry to the RPC. The QMP is responsible for testing, or overseeing testing, of the FGI equipment to ensure it operates safely. The QMP also provides PSD estimates when necessary and advises on the impact of the administered radiation dose on the patient. Lead technologist: The lead technologist is responsible for maintaining imaging protocols under the direction of the supervising physician and QMP. This technologist advises the other RPC members of equipment issues or other observations that may affect patient dose or image quality. When PSD estimates are required, the lead technologist assists the QMP in gathering the necessary data. The organizational structure of the healthcare system will dictate the optimal committee arrangement. For a stand‐alone facility, if different medical specialties utilize FGI equipment (e.g., interventional cardiology, interventional radiology, vascular surgery), the facility should consider including physicians and technologists from each specialty. Other possible committee members are department managers, health system risk managers, and the radiation safety officer (RSO). Other individuals may be added to the committee as deemed necessary. If a larger healthcare system has more than one facility, a system‐wide committee may be established to ensure consistency among facilities. In this structure, one RPC may be formed, provided each facility within the enterprise has appropriate representation on the committee. Another option is to add the scope of the RPC into the responsibilities of an already established Radiation Safety Committee, as long as the recommended RPC membership is met. Per the SSRs, the RPC is responsible for establishing procedures and protocols to be followed before, during, and after FGI procedures. The SSRs state that protocols addressing the following are reviewed by the committee at least annually: Authorized users of FGI equipment. Intra‐procedure patient radiation dose monitoring. Dose notification levels. Establishment of SRDL values. Actions to be taken when a SRDL is exceeded. These protocols are discussed elsewhere in this report. The SSRs only require the RPC to create and implement policies and does not require committee oversight of clinical data. However, given the membership, it is the recommendation of this MPPG that such review fall under RPC oversight. This review includes evaluation of clinical and dosimetric data from any procedure in which the SRDL is exceeded or a tissue reaction has occurred, as well as analysis of facility or system case volume and dose index datasets, described in a later section. These datasets can include case volume and comparative dosimetry. The duties of the RPC could also be extended to include the previously discussed QA‐PR functions. In doing so, the committee reviews procedure justification, patient‐specific factors, radiation dose optimization, the time course over which radiation doses were administered, disease severity, and procedure complexity. Depending upon the size of the healthcare system, initial oversight might better be accomplished by department‐level committees. Under such a framework, each individual department (e.g., interventional cardiology, interventional radiology, vascular surgery), would have its own committee to oversee clinical and dosimetric data, and to perform QA‐PR review of the clinical appropriateness of procedures resulting in tissue effects. The results of this department‐level review would then be sent to the facility or system‐wide RPC for final review and approval. The healthcare system determines the frequency of RPC meetings. The SSRs suggest that the committee meet at least annually. More frequent meetings may allow for more rapid identification of potential problems and swifter implementation of subsequent changes in practice. Considerations that affect meeting frequency include the duties of the committee, the volume and complexity of FGI procedures performed, the number of physicians and medical staff involved, and the capabilities of the imaging equipment. Of these, procedure volume and complexity may be the most important, as both play a role in the number of potential procedures that may require committee review, if such review is tasked to the RPC. Some events may prompt the need for ad hoc meetings, such as sentinel events or serious patient injury review, per the previous NCRP Statement 11 QA‐PR discussion. The SSRs suggest that the RPC provide an annual report to the radiation safety committee, or, if the facility does not have a radiation safety committee, to the RSO. Although not addressed in the SSRs, this MPPG recommends that the annual report contain the following: A list of individuals authorized to use fluoroscopic equipment for FGI procedures. The total number of procedures that exceeded the SRDL and, if available, percentage of total procedures. A review of all policies and procedures and any significant changes.
Procedures for maintaining records The SSRs also recommend that all available radiation dose indices be recorded to facilitate skin dose estimation, if needed. These recommendations are similar to TJC requirements previously discussed, and the role of the QMP is to provide guidance on what specific information must be recorded for skin dose estimates. If RDSRs are produced by a fluoroscope, they should be archived, if possible, and the QMP must have access to them when needed. Where the RSDR is not available, the QMP may request that additional information be recorded, such as gantry angles, fluoroscopic technique factors, table height, patient dimensions, etc. Appropriate patient follow‐up procedures, including skin dose estimation, are performed according to policies set forth by the committee.
Facility dataset review for quality improvement Beyond CRCPD recommendations and TJC standards to review and analyze procedures over threshold values, collection of fluoroscopic dose index data provides facilities with quality improvement opportunities, as these data can be used to evaluate clinical practice. This type of analysis is useful because dose to the patient in fluoroscopy is highly dependent on the operator of the equipment. Since the QMP often oversees the collection of fluoroscopic data, they are well positioned to handle the task of data analysis. As detailed in NCRP Report 168, these data can be used to build “facility datasets” for each type of exam performed in a practice that can be compared among sites or to normative datasets, such as the American College of Radiology Fluoroscopy Dose Index Registry (DIR). These analyses can provide valuable insight into how the equipment is being used clinically and identify deviations from expected practice. NCRP Report No. 168 also recommends calculating the percentages of each procedure that exceed the SRDL set by the facility. This process can help determine those procedures that are more likely to result in high doses to patients. This information has the potential to affect the pre‐procedure consent process and room assignment decisions. Fluoroscopy dose index data may also be used to set procedure‐specific dose index review levels for evaluating clinical practice. These are not to be confused with the SRDL, which is a biologically based threshold related to the potential for tissue effects. Procedure‐specific review levels represent target achievable dose indices for specific procedures and can be useful for comparing clinical practice. Beyond reported indices, collected data can be used to derive additional metrics for practice insight. Dividing K a,r by total fluoroscopy time yields a measure of radiation utilization efficiency for a particular procedure, assuming fluoroscopy time is in some way related to the complexity of the procedure. For two procedures of similar complexity and fluoroscopy time, a procedure with higher utilization of acquisition or high‐level control modes would yield a higher value. Additionally, total P KA and K a,r data can be used to calculate the average field size ( P KA / K a,r ) at the IRP/PERP for a given procedure, thus permitting comparison of an operator's use of collimation. Beyond individual procedures, these derived metrics can be calculated for a fluoroscope or operator over a given month or quarter to assess and compare longer term operational trends. Values of any metric found to be substantially different from facility benchmarks can be investigated for potential protocol optimization or practice review. QMPs need to be mindful of the units used for a given dose index when comparing data. While K a,r is often reported in units of mGy, displayed P KA values vary by manufacturer and even by model, necessitating unit conversion for valid comparisons. However, differences in procedural dose index data do not necessarily indicate improper practice. It is likely that more experienced physicians perform a larger share of more complex procedures, which would affect their data. A similar situation can arise with equipment, where a newer Interventional Radiology (IR) suite with more features is rightfully utilized more often for complex procedures than older equipment. Additionally, patient body habitus and anatomy can drastically affect dose indices. Individual procedures may exceed SRDLs not because they were performed improperly, but because larger patients or complicated vascular anatomy require higher tube outputs or longer procedure time. The results of this type of analysis can be presented to operators or department chairs on a regular basis, either through a fluoroscopic RPC or dose review committee, or through other means such as existing staff meetings, radiation safety committee meetings, QA‐PR committee, or morbidity and mortality (M&M) conferences. This feedback can be an important part of practice improvement. Beyond comparing data within a practice, comparison to outside practices may be possible through programs such as the ACR Fluoroscopy DIR. The ACR Fluoroscopy DIR is the latest addition to the National Radiology Data Registry, and the first modality to join computed tomography (CT) in the DIR. The Fluoroscopy DIR provides a continuously updated normative dataset for fluoroscopically guided procedures, which can be used as an advisory dataset to which participating sites can compare their facility datasets of fluoroscopy dose indices. The ability of a hospital to perform this kind of analysis will be heavily dependent on the resources available. If full procedure records, including the exam type, fluoroscope used for the procedure, operator, and available dose indices are electronically captured, datasets can be created and evaluated with relative ease. However, this level of investigation is not practical if paper logs are used to meet documentation requirements. If resources are limited, analysis should focus on higher dose FGI procedures, as they present the greatest radiation risk. However, if fluoroscopy data for all equipment are being collected electronically, further evaluations for basic fluoroscopy rooms and mobile C‐arms and their operators can be performed without significant additional effort.
ADDITIONAL FGI RELATED QMP DUTIES The following sections briefly discuss QMP fluoroscopy duties beyond patient dose management and serve to direct the reader to related resources. 11.1 Training and privileging of fluoroscopy users Training in the safe use of fluoroscopy, credentialing, and privileging of fluoroscopy users are important aspects of patient and staff radiation safety. Privileges identify which medical procedures a staff member may perform, while credentialing involves gathering relevant data for privileging an applicant. The ACR and AAPM recommend that each facility have a policy for granting privileges for fluoroscopy use, and CRCPD SSRs suggest this policy be developed by the RPC. , The QMP should be involved in the following tasks as they apply to the training and credentialing of fluoroscopy users: determining the need for a credentialing program, developing a fluoroscopy privileging policy, verifying compliance with requirements, developing and maintaining didactic content, and providing in‐person training on the safe use of fluoroscopy equipment. The nature of the training may be dictated by the specific scenario (e.g., new FGI operator vs. renewal, or FGI training vs. mini C‐arm training) as well as state regulations. Specific details of the QMP's role are beyond the scope of this practice guideline, but a comprehensive review is provided in AAPM Report No. 124. 11.2 Occupational dose monitoring and badging Personnel involved in interventional fluoroscopy procedures often require occupational radiation monitoring according to state regulations. For a given healthcare system, oversight of occupational dosimetry may be the responsibility of the QMP or a separate RSO. Even if not a direct job duty, the QMP may have better insights into clinical FGI practice and can aid in setting up a dosimetry program and can help verify the following: 11.2.1 Appropriate personnel are monitored Most state regulations mirror the SSR language that require monitoring of anyone likely to exceed 10% of the annual occupational dose limits. In an FGI environment, operators, be they physicians or technologists, generally require monitoring as they are likely to exceed the 5 mSv (500 mrem) effective dose equivalent (EDE) threshold, which is 10% of the annual (EDE) limit of 50 mSv (5000 mrem). However, depending on local practice and positioning during procedures, ancillary personnel such as nurses or anesthetists often record annual doses well below the 5 mSv threshold for monitoring. Ancillary personnel are often positioned farther from the patient during fluoroscopy and can be behind additional protective measures such as rolling shields. Previous dosimetry histories must be investigated in order to determine which personnel groups require monitoring. Monitoring policies may need to be reevaluated when practices or volumes change. 11.2.2 Monitoring devices are appropriate Dosimetry service providers must be accredited by the National Voluntary Laboratory Accreditation Program (NVLAP) for the type of radiation for which monitoring is performed. A one‐ or two‐badge protocol can be used to monitor occupational dose. Single monitoring devices are worn at the neck/collar outside any radiation personal protective equipment (PPE). The readings from this badge are directly used to calculate the various personal dose equivalents. Some states allow modification of measured deep dose equivalent according to equations in the SSRs and in NCRP Report Nos. 122 and 168 to account for the presence of PPE to derive a more accurate EDE ( H E ). This modification results in lower annual H E readings compared to an un‐modified deep dose, which can be advantageous for workers approaching annual limits. The QMP must verify if this correction is allowed, as state regulations vary, and sometimes the correction can only be used in cases where the reported dose exceeds 25% of annual limits. With a two‐badge protocol, a second monitoring device is issued and worn underneath the protective garment at the waist. A two‐badge protocol best represents the occupational dose to the worker and will result in lower H E values than a corrected single‐badge protocol, though again, this may not be allowed by local regulation. Even when allowed, the two‐badge protocol has the disadvantage of requiring twice as many badges. In addition to greater cost and administrative burden, the two‐badge protocol increases the risk of lost or improperly used badges. Wearers often inappropriately mix up the collar and waist badges, leading to faulty data. A recent large scale review of occupational dose data from FGI areas noted that over a third of individual entries from two‐badge wearers were invalid due to either improper badge use, or failure to return both badges. Due to these likely operational problems with double badging, this MPPG recommends that a single‐badge protocol be used in FGI environments, with modifications to badge readings applied to account for the presence of leaded PPE whenever possible. Doses to the extremities and skin (shallow dose equivalent) and lens of the eye (lens dose equivalent) are estimated by the dosimetry service provider based on the dose received by the monitoring device worn at the neck. Some dosimetry providers allow for corrections to lens dose to account for the use of leaded eyewear. Some also provide separate monitors that attach to eyeglass. Lens dose should be assessed, as evidence has shown a lower threshold for radiation‐induced cataract than previously thought. , The ICRP has lowered the annual dose limit to the lens to 20 mSv/year, compared to the 150 mSv limit in place in the United States. The NCRP has recommended a 50 mSv annual limit. The previous large scale study found that over 15% of workers exceeded the ICRP limit to the lens. The study found no difference in reported eye dose between a one‐ or two‐badge protocol. 11.2.3 Monitoring frequency is appropriate Any state regulatory requirements regarding monitoring frequency must be followed. Monthly monitoring is recommended as it permits more timely identification and investigation of high doses than does bimonthly or quarterly monitoring. However, less frequent monitoring may be appropriate for groups with lower occupational doses to reduce costs and administrative burden. 11.2.4 Monitoring devices are used correctly by staff Each individual should wear their assigned monitoring device(s) during procedures, at the correct location on their body and in the correct position with respect to the protective apron. When not in use, monitoring devices are stored in a location where they are not exposed to radiation above background levels. Control badges should not be exposed to radiation above background levels. Monitoring devices are exchanged, or data are retrieved from electronic monitoring devices in a timely manner. During the previously described observations of clinical FGI procedures, the QMP can observe occupational dosimetry usage and offer suggestions for improvement. 11.2.5 Monitoring results are reviewed Doses received by monitored employees should be reviewed at least quarterly to identify exposures above the facility's As Low as Reasonably Achievable (ALARA) investigational levels and regulatory limits. Local regulations and accreditation body requirements must be followed. Regular dose review can identify trends, opportunities for dose reduction, and possible non‐compliance issues relating to badge misuse or nonuse. Feedback to users, department chairs, and the RPC on the results of dosimetry review are an important part of quality improvement. The review may be done by the QMP, the RSO, or an individual delegated this responsibility by the QMP or RSO. Even in circumstances where an RSO oversees monitoring results, the QMP can provide additional radiation protection guidance to FGI participants though didactic or hands on demonstrations. 11.2.6 Occupational radiation exposures meet regulatory requirements Results should be compared to limits in state regulations, the US Nuclear Regulatory Commission (NRC) regulations, or, if applicable, United States Occupational Safety and Health Administration (OSHA) limits. , Mandatory reporting to applicable regulatory bodies must be performed if dose limits are exceeded. 11.3 Institutional review board QMPs in academic or research settings may be tasked with participation in hospital institutional review boards (IRB), which oversee human research studies and ensure ethical treatment of participants. Most often, the QMP serves to provide dose and risk estimates for imaging research studies that differ from standard of care. NCRP Report No. 185, “Evaluating and Communicating Radiation Risks for Studies Involving Human Subjects: Guidance for Researchers and Institutional Review Board” provides useful information for QMPs serving this capacity.
Training and privileging of fluoroscopy users Training in the safe use of fluoroscopy, credentialing, and privileging of fluoroscopy users are important aspects of patient and staff radiation safety. Privileges identify which medical procedures a staff member may perform, while credentialing involves gathering relevant data for privileging an applicant. The ACR and AAPM recommend that each facility have a policy for granting privileges for fluoroscopy use, and CRCPD SSRs suggest this policy be developed by the RPC. , The QMP should be involved in the following tasks as they apply to the training and credentialing of fluoroscopy users: determining the need for a credentialing program, developing a fluoroscopy privileging policy, verifying compliance with requirements, developing and maintaining didactic content, and providing in‐person training on the safe use of fluoroscopy equipment. The nature of the training may be dictated by the specific scenario (e.g., new FGI operator vs. renewal, or FGI training vs. mini C‐arm training) as well as state regulations. Specific details of the QMP's role are beyond the scope of this practice guideline, but a comprehensive review is provided in AAPM Report No. 124.
Occupational dose monitoring and badging Personnel involved in interventional fluoroscopy procedures often require occupational radiation monitoring according to state regulations. For a given healthcare system, oversight of occupational dosimetry may be the responsibility of the QMP or a separate RSO. Even if not a direct job duty, the QMP may have better insights into clinical FGI practice and can aid in setting up a dosimetry program and can help verify the following: 11.2.1 Appropriate personnel are monitored Most state regulations mirror the SSR language that require monitoring of anyone likely to exceed 10% of the annual occupational dose limits. In an FGI environment, operators, be they physicians or technologists, generally require monitoring as they are likely to exceed the 5 mSv (500 mrem) effective dose equivalent (EDE) threshold, which is 10% of the annual (EDE) limit of 50 mSv (5000 mrem). However, depending on local practice and positioning during procedures, ancillary personnel such as nurses or anesthetists often record annual doses well below the 5 mSv threshold for monitoring. Ancillary personnel are often positioned farther from the patient during fluoroscopy and can be behind additional protective measures such as rolling shields. Previous dosimetry histories must be investigated in order to determine which personnel groups require monitoring. Monitoring policies may need to be reevaluated when practices or volumes change. 11.2.2 Monitoring devices are appropriate Dosimetry service providers must be accredited by the National Voluntary Laboratory Accreditation Program (NVLAP) for the type of radiation for which monitoring is performed. A one‐ or two‐badge protocol can be used to monitor occupational dose. Single monitoring devices are worn at the neck/collar outside any radiation personal protective equipment (PPE). The readings from this badge are directly used to calculate the various personal dose equivalents. Some states allow modification of measured deep dose equivalent according to equations in the SSRs and in NCRP Report Nos. 122 and 168 to account for the presence of PPE to derive a more accurate EDE ( H E ). This modification results in lower annual H E readings compared to an un‐modified deep dose, which can be advantageous for workers approaching annual limits. The QMP must verify if this correction is allowed, as state regulations vary, and sometimes the correction can only be used in cases where the reported dose exceeds 25% of annual limits. With a two‐badge protocol, a second monitoring device is issued and worn underneath the protective garment at the waist. A two‐badge protocol best represents the occupational dose to the worker and will result in lower H E values than a corrected single‐badge protocol, though again, this may not be allowed by local regulation. Even when allowed, the two‐badge protocol has the disadvantage of requiring twice as many badges. In addition to greater cost and administrative burden, the two‐badge protocol increases the risk of lost or improperly used badges. Wearers often inappropriately mix up the collar and waist badges, leading to faulty data. A recent large scale review of occupational dose data from FGI areas noted that over a third of individual entries from two‐badge wearers were invalid due to either improper badge use, or failure to return both badges. Due to these likely operational problems with double badging, this MPPG recommends that a single‐badge protocol be used in FGI environments, with modifications to badge readings applied to account for the presence of leaded PPE whenever possible. Doses to the extremities and skin (shallow dose equivalent) and lens of the eye (lens dose equivalent) are estimated by the dosimetry service provider based on the dose received by the monitoring device worn at the neck. Some dosimetry providers allow for corrections to lens dose to account for the use of leaded eyewear. Some also provide separate monitors that attach to eyeglass. Lens dose should be assessed, as evidence has shown a lower threshold for radiation‐induced cataract than previously thought. , The ICRP has lowered the annual dose limit to the lens to 20 mSv/year, compared to the 150 mSv limit in place in the United States. The NCRP has recommended a 50 mSv annual limit. The previous large scale study found that over 15% of workers exceeded the ICRP limit to the lens. The study found no difference in reported eye dose between a one‐ or two‐badge protocol. 11.2.3 Monitoring frequency is appropriate Any state regulatory requirements regarding monitoring frequency must be followed. Monthly monitoring is recommended as it permits more timely identification and investigation of high doses than does bimonthly or quarterly monitoring. However, less frequent monitoring may be appropriate for groups with lower occupational doses to reduce costs and administrative burden. 11.2.4 Monitoring devices are used correctly by staff Each individual should wear their assigned monitoring device(s) during procedures, at the correct location on their body and in the correct position with respect to the protective apron. When not in use, monitoring devices are stored in a location where they are not exposed to radiation above background levels. Control badges should not be exposed to radiation above background levels. Monitoring devices are exchanged, or data are retrieved from electronic monitoring devices in a timely manner. During the previously described observations of clinical FGI procedures, the QMP can observe occupational dosimetry usage and offer suggestions for improvement. 11.2.5 Monitoring results are reviewed Doses received by monitored employees should be reviewed at least quarterly to identify exposures above the facility's As Low as Reasonably Achievable (ALARA) investigational levels and regulatory limits. Local regulations and accreditation body requirements must be followed. Regular dose review can identify trends, opportunities for dose reduction, and possible non‐compliance issues relating to badge misuse or nonuse. Feedback to users, department chairs, and the RPC on the results of dosimetry review are an important part of quality improvement. The review may be done by the QMP, the RSO, or an individual delegated this responsibility by the QMP or RSO. Even in circumstances where an RSO oversees monitoring results, the QMP can provide additional radiation protection guidance to FGI participants though didactic or hands on demonstrations. 11.2.6 Occupational radiation exposures meet regulatory requirements Results should be compared to limits in state regulations, the US Nuclear Regulatory Commission (NRC) regulations, or, if applicable, United States Occupational Safety and Health Administration (OSHA) limits. , Mandatory reporting to applicable regulatory bodies must be performed if dose limits are exceeded.
Appropriate personnel are monitored Most state regulations mirror the SSR language that require monitoring of anyone likely to exceed 10% of the annual occupational dose limits. In an FGI environment, operators, be they physicians or technologists, generally require monitoring as they are likely to exceed the 5 mSv (500 mrem) effective dose equivalent (EDE) threshold, which is 10% of the annual (EDE) limit of 50 mSv (5000 mrem). However, depending on local practice and positioning during procedures, ancillary personnel such as nurses or anesthetists often record annual doses well below the 5 mSv threshold for monitoring. Ancillary personnel are often positioned farther from the patient during fluoroscopy and can be behind additional protective measures such as rolling shields. Previous dosimetry histories must be investigated in order to determine which personnel groups require monitoring. Monitoring policies may need to be reevaluated when practices or volumes change.
Monitoring devices are appropriate Dosimetry service providers must be accredited by the National Voluntary Laboratory Accreditation Program (NVLAP) for the type of radiation for which monitoring is performed. A one‐ or two‐badge protocol can be used to monitor occupational dose. Single monitoring devices are worn at the neck/collar outside any radiation personal protective equipment (PPE). The readings from this badge are directly used to calculate the various personal dose equivalents. Some states allow modification of measured deep dose equivalent according to equations in the SSRs and in NCRP Report Nos. 122 and 168 to account for the presence of PPE to derive a more accurate EDE ( H E ). This modification results in lower annual H E readings compared to an un‐modified deep dose, which can be advantageous for workers approaching annual limits. The QMP must verify if this correction is allowed, as state regulations vary, and sometimes the correction can only be used in cases where the reported dose exceeds 25% of annual limits. With a two‐badge protocol, a second monitoring device is issued and worn underneath the protective garment at the waist. A two‐badge protocol best represents the occupational dose to the worker and will result in lower H E values than a corrected single‐badge protocol, though again, this may not be allowed by local regulation. Even when allowed, the two‐badge protocol has the disadvantage of requiring twice as many badges. In addition to greater cost and administrative burden, the two‐badge protocol increases the risk of lost or improperly used badges. Wearers often inappropriately mix up the collar and waist badges, leading to faulty data. A recent large scale review of occupational dose data from FGI areas noted that over a third of individual entries from two‐badge wearers were invalid due to either improper badge use, or failure to return both badges. Due to these likely operational problems with double badging, this MPPG recommends that a single‐badge protocol be used in FGI environments, with modifications to badge readings applied to account for the presence of leaded PPE whenever possible. Doses to the extremities and skin (shallow dose equivalent) and lens of the eye (lens dose equivalent) are estimated by the dosimetry service provider based on the dose received by the monitoring device worn at the neck. Some dosimetry providers allow for corrections to lens dose to account for the use of leaded eyewear. Some also provide separate monitors that attach to eyeglass. Lens dose should be assessed, as evidence has shown a lower threshold for radiation‐induced cataract than previously thought. , The ICRP has lowered the annual dose limit to the lens to 20 mSv/year, compared to the 150 mSv limit in place in the United States. The NCRP has recommended a 50 mSv annual limit. The previous large scale study found that over 15% of workers exceeded the ICRP limit to the lens. The study found no difference in reported eye dose between a one‐ or two‐badge protocol.
Monitoring frequency is appropriate Any state regulatory requirements regarding monitoring frequency must be followed. Monthly monitoring is recommended as it permits more timely identification and investigation of high doses than does bimonthly or quarterly monitoring. However, less frequent monitoring may be appropriate for groups with lower occupational doses to reduce costs and administrative burden.
Monitoring devices are used correctly by staff Each individual should wear their assigned monitoring device(s) during procedures, at the correct location on their body and in the correct position with respect to the protective apron. When not in use, monitoring devices are stored in a location where they are not exposed to radiation above background levels. Control badges should not be exposed to radiation above background levels. Monitoring devices are exchanged, or data are retrieved from electronic monitoring devices in a timely manner. During the previously described observations of clinical FGI procedures, the QMP can observe occupational dosimetry usage and offer suggestions for improvement.
Monitoring results are reviewed Doses received by monitored employees should be reviewed at least quarterly to identify exposures above the facility's As Low as Reasonably Achievable (ALARA) investigational levels and regulatory limits. Local regulations and accreditation body requirements must be followed. Regular dose review can identify trends, opportunities for dose reduction, and possible non‐compliance issues relating to badge misuse or nonuse. Feedback to users, department chairs, and the RPC on the results of dosimetry review are an important part of quality improvement. The review may be done by the QMP, the RSO, or an individual delegated this responsibility by the QMP or RSO. Even in circumstances where an RSO oversees monitoring results, the QMP can provide additional radiation protection guidance to FGI participants though didactic or hands on demonstrations.
Occupational radiation exposures meet regulatory requirements Results should be compared to limits in state regulations, the US Nuclear Regulatory Commission (NRC) regulations, or, if applicable, United States Occupational Safety and Health Administration (OSHA) limits. , Mandatory reporting to applicable regulatory bodies must be performed if dose limits are exceeded.
Institutional review board QMPs in academic or research settings may be tasked with participation in hospital institutional review boards (IRB), which oversee human research studies and ensure ethical treatment of participants. Most often, the QMP serves to provide dose and risk estimates for imaging research studies that differ from standard of care. NCRP Report No. 185, “Evaluating and Communicating Radiation Risks for Studies Involving Human Subjects: Guidance for Researchers and Institutional Review Board” provides useful information for QMPs serving this capacity.
CONCLUSION QMPs have a major role in helping healthcare institutions manage modern FGI practices. Their expertise related to the equipment, radiation biology, and regulatory environment surrounding fluoroscopy makes them an invaluable member of clinical team. QMPs can help in the development and implementation of appropriate protocols for managing patient radiation dose before, during, and after potentially high‐dose FGI procedures, as well as oversee analysis of clinical dose index data to further improve practice. Through educational efforts, QMPs can ensure that all clinical staff are appropriately informed of the potential radiation risks from fluoroscopic procedures so that appropriate measures can be taken to minimize risks and maximize benefits for patients and staff.
This guideline was reviewed and updated by the Medical Physics Practice Guideline Task Group 343 of the Professional Council of the AAPM. Each author reviewed recent literature on the topic and offered opinions on and language for the guideline. They also reviewed and applied comments from the full AAPM membership to the document.
Jaydev K. Dave, PhD has received support from Philips Healthcare, GE Healthcare, and Lantheus Medical Imaging Inc. for research unrelated to this project. Philips Healthcare and GE Healthcare manufacture fluoroscopy equipment. Jaydev K. Dave has also provided imaging physics consulting services as a part of Rayscan Inc. A. Kyle Jones, PhD, FAAPM is President of FluoroSafety, a company that produces CME on quality and safety in medical imaging. Dr. Jones recused himself from all discussions related to personnel education and training during the development and review of this MPPG. The remaining authors have no conflict of interest.
|
Research note: Proteomics profiling reveal key proteins in egg white emulsification | cfba21be-af27-4e34-9867-5bb8c9ede7c0 | 11741918 | Biochemistry[mh] | Emulsions represent fundamental systems in food processing, with their stability and properties directly impacting product quality and functionality. The amphiphilic nature of egg white (EW) proteins enables them to adsorb and form a physical barrier at the oil-water interface, which not only reduces the interfacial tension, but also enhances the stability of the emulsion and prevents phase separation . EW is widely used in the food industry as a natural and multifunctional food ingredient that is rich in protein. Clarifying which proteins are involved in the emulsification properties of EW will provide a scientific basis for efficient and precise processing. The oil-water interface is a critical and complex interface, and the adsorption kinetics of proteins at this interface and the structure of the protein film formed are critical to the formation and stability of emulsions . Compared to other proteins such as whey and soy proteins, the emulsification properties of EW are generally considered to be poor, limiting the application of EW in emulsion systems. In recent years, with the development of new emulsification processes such as pH treatment, high-pressure homogenization and ultrasound, researchers have found that these operations can significantly affect the structural and interfacial properties of proteins . For instance, high-pressure homogenization induces protein defolding and enhances protein-protein interactions, resulting in a tighter protein layer and smaller droplet size . Although a large number of studies have focused on the emulsification properties of the major proteins of EW , e.g., ovalbumin, ovotransferrin, lysozyme, there is still a lack of systematic understanding of the emulsification mechanism under the synergistic effect of multiple proteins in natural EW, since the emulsion behaviors of purified proteins are quite different from the emulsification characteristics of natural complex protein systems. In particular, a number of proteins in EW that are less abundant but may play an important role in the emulsification process are often overlooked, choosing EW as a research subject is more in line with production reality. To address these knowledge gaps, utilizing 4D-SmartDIA, which can significantly enhance the depth of proteomics detection with the advantages of good quantitative reproducibility and reliability, we systematically identify and analyze the key proteins in the emulsification process of EW, and to comprehensively discuss the structural characteristics of the major high-abundance proteins, differential proteins, and the behavior of the oil-water interface. By using 4D-SmartDIA quantitative proteomics technology, the dynamic changes of the proteome during the emulsification process of natural EW were comprehensively revealed, providing new insights into the emulsification mechanism of EW.
Preparation of EW and samples after emulsification Fifteen fresh eggs (egg weight = 60.0 ± 3.0 g; Sichuan Sundaily Village Ecological Foods, Mianyang, China) were taken; manually cracked, the EW were carefully separated, and the impurities were removed with tweezers, then the EW were stirred homogeneously for 30 min at 4°C using a magnetic stirrer at 300 rpm. Its protein concentration was adjusted to 1% using distilled water. Protein quantification by bicinchoninic acid (BCA) assay. Egg white emulsion (EWE) was prepared by adding oil according to EW solution: corn oil = 9:1 (v:v), and oil-in-water emulsions were prepared by mixing with a high-speed disperser (XHF-DY, NingBo Scientz Biotechnology Co.,Ltd, Zhejiang, China) at 11,000 rpm for 2 min and then subjected to homogenization treatments (SCIENTZ-207A, NingBo Scientz Biotechnology Co.,Ltd, Zhejiang, China) at 50 Mpa for three times. Quantitative proteomics analysis As shown in A, EWE (35 mL) was placed in a 50 mL centrifuge tube and centrifuged at 14,000 × g for 45 min at 4°C, the aqueous phase solution was collected and repeat the centrifugation 3 times, and the aqueous phase solution obtained from the last centrifugation was collected aqueous phases of egg white emulsions (EWE-W). During emulsification, a portion of the EW proteins are involved in the oil-water interface formation and are retained in the oil-phase layer during centrifugation, while proteins not participated in the oil-water interface formation are maintained in the EWE-W. The preparation of EWE-W was repeated three times. Equal amounts of protein were taken from EW and EWE-W, then trypsinized and quantified by 4D-unlabeling quantitative proteomic analysis . The tryptic peptides were separated using a NanoElute UHPLC system (Bruker Daltonics). The mobile phase consisted of solvent A (0.1% formic acid, 2% acetonitrile/in water) and solvent B (0.1% formic acid in acetonitrile), and the gradient increased from 6% to 24% at 0∼9 min, to 35% at 9∼11 min, to 80% at 11∼13 min, and was maintained at 80% at a flow rate of 500 nL/min. The peptides were subjected to capillary source followed by the timsTOF Pro mass spectrometry. The electrospray voltage applied was 1.75 kV. Precursors and fragments were analyzed at the TOF detector. The timsTOF Pro was operated in data independent parallel accumulation serial fragmentation (dia-PASEF) mode. The full MS scan was set as 300-1500 (MS/MS scan range) and 20PASEF (MS/MS mode) -MS/MS scans were acquired per cycle. The MS/MS scan range was set as 400 and isolation window was set as 7 m/z. The DIA data were processed using DIA-NN search engine (v.1.8). Tandem mass spectra were searched against Gallus _ gallus _9031_PR_20231121.fasta (42741 entries) concatenated with reverse decoy database. Trypsin/P was specified as cleavage enzyme allowing up to 1 missing cleavages. The resolution of both mass spectral scans was 30,000, with a mass tolerance of 23 ppm for primary precursor ions and 11 ppm for secondary fragment ions. Excision on N-term Met and carbamidomethyl on Cys were specified as fixed modification. FDR was adjusted to < 1%. Statistical analysis All data were the mean of four replicate samples. Statistical analyses were performed using Prism 9.50 (GraphPad, La Jolla, CA, USA), and analyze significant differences at the P < 0.05 and P < 0.01 levels, respectively. Access to protein information through the UniProt online platform ( https://www.uniprot.org/ ), and the grand average of hydropathicity the target proteins and the hydrophobicity values of each amino acid site were obtained using the Expasy online platform ( https://web.expasy.org/ ).
Fifteen fresh eggs (egg weight = 60.0 ± 3.0 g; Sichuan Sundaily Village Ecological Foods, Mianyang, China) were taken; manually cracked, the EW were carefully separated, and the impurities were removed with tweezers, then the EW were stirred homogeneously for 30 min at 4°C using a magnetic stirrer at 300 rpm. Its protein concentration was adjusted to 1% using distilled water. Protein quantification by bicinchoninic acid (BCA) assay. Egg white emulsion (EWE) was prepared by adding oil according to EW solution: corn oil = 9:1 (v:v), and oil-in-water emulsions were prepared by mixing with a high-speed disperser (XHF-DY, NingBo Scientz Biotechnology Co.,Ltd, Zhejiang, China) at 11,000 rpm for 2 min and then subjected to homogenization treatments (SCIENTZ-207A, NingBo Scientz Biotechnology Co.,Ltd, Zhejiang, China) at 50 Mpa for three times.
As shown in A, EWE (35 mL) was placed in a 50 mL centrifuge tube and centrifuged at 14,000 × g for 45 min at 4°C, the aqueous phase solution was collected and repeat the centrifugation 3 times, and the aqueous phase solution obtained from the last centrifugation was collected aqueous phases of egg white emulsions (EWE-W). During emulsification, a portion of the EW proteins are involved in the oil-water interface formation and are retained in the oil-phase layer during centrifugation, while proteins not participated in the oil-water interface formation are maintained in the EWE-W. The preparation of EWE-W was repeated three times. Equal amounts of protein were taken from EW and EWE-W, then trypsinized and quantified by 4D-unlabeling quantitative proteomic analysis . The tryptic peptides were separated using a NanoElute UHPLC system (Bruker Daltonics). The mobile phase consisted of solvent A (0.1% formic acid, 2% acetonitrile/in water) and solvent B (0.1% formic acid in acetonitrile), and the gradient increased from 6% to 24% at 0∼9 min, to 35% at 9∼11 min, to 80% at 11∼13 min, and was maintained at 80% at a flow rate of 500 nL/min. The peptides were subjected to capillary source followed by the timsTOF Pro mass spectrometry. The electrospray voltage applied was 1.75 kV. Precursors and fragments were analyzed at the TOF detector. The timsTOF Pro was operated in data independent parallel accumulation serial fragmentation (dia-PASEF) mode. The full MS scan was set as 300-1500 (MS/MS scan range) and 20PASEF (MS/MS mode) -MS/MS scans were acquired per cycle. The MS/MS scan range was set as 400 and isolation window was set as 7 m/z. The DIA data were processed using DIA-NN search engine (v.1.8). Tandem mass spectra were searched against Gallus _ gallus _9031_PR_20231121.fasta (42741 entries) concatenated with reverse decoy database. Trypsin/P was specified as cleavage enzyme allowing up to 1 missing cleavages. The resolution of both mass spectral scans was 30,000, with a mass tolerance of 23 ppm for primary precursor ions and 11 ppm for secondary fragment ions. Excision on N-term Met and carbamidomethyl on Cys were specified as fixed modification. FDR was adjusted to < 1%.
All data were the mean of four replicate samples. Statistical analyses were performed using Prism 9.50 (GraphPad, La Jolla, CA, USA), and analyze significant differences at the P < 0.05 and P < 0.01 levels, respectively. Access to protein information through the UniProt online platform ( https://www.uniprot.org/ ), and the grand average of hydropathicity the target proteins and the hydrophobicity values of each amino acid site were obtained using the Expasy online platform ( https://web.expasy.org/ ).
Proteome identification and quantification analysis EW and EWE-W were analyzed using 4D-SmartDIA quantitative proteomics technology, and a total of 96 proteins were identified. Principal component analysis (PCA) revealed distinct clustering patterns between EW and EWE-W groups with excellent reproducibility within groups ( B). PC1 and PC2 accounted for 72.8% and 7.9% of the total variance, respectively, demonstrating substantial alterations in protein composition during the emulsification process. To identify proteins crucial for emulsification, differentially abundant proteins were screened using stringent criteria (fold change > 1.50 or < 0.67, P < 0.05). In EWE-W, a total of 64 proteins were found to be significantly associated with EW emulsification capacity, of which 20 were increased in and 44 were decreased in abundance ( C). The predominant decrease in protein abundance in EWE-W suggests that the majority of emulsification-active proteins were preferentially recruited to the oil-water interface. These differential proteins may be key factors influencing the emulsification properties of EW, and their behavior at the oil-water interface deserves further exploration. Differential protein analysis in EW emulsification Quantitative proteomic analysis revealed distinct patterns of protein abundance changes between EW and EWE-W systems, providing insights into the molecular mechanisms of EW emulsification. Several proteins exhibited significantly increased abundance in EWE-W, including Ig-like domain-containing protein (8.65-fold, P < 0.01), PIT54 protein (3.40-fold, P < 0.01) and Mucin-6 (Fragment) (2.28-fold, P < 0.05), etc. ( A). The enhanced abundance of these proteins in the aqueous phase indicated their limited participation in the oil-water interface, possibly due to structural characteristics unfavorable for oil-water interface stabilization. More notably, a larger subset of proteins showed significantly decreased abundance in EWE-W, suggesting that they may be actively involved in the formation of the oil-water interface. Among them, the Protein TENP showed the most significant change in abundance (0.07-fold, P < 0.05) among all abundance-reducing proteins ( B), suggesting its strong affinity for the oil-water interface. Although the Protein TENP is a non-dominant protein in EW, it plays a crucial role in the emulsification process. In addition, proteins with significantly decreased abundance include lipid-binding serum glycoprotein C-terminal domain-containing protein (0.11-fold, P < 0.05), Glutathione peroxidase (0.12-fold, P < 0.05), acyloxyacyl hydrolase (0.12-fold, P < 0.05) and SRCR domain-containing protein (0.14-fold, P < 0.05), etc. The similar magnitude of reduction among these proteins suggests a coordinated mechanism in interface formation, where multiple proteins contribute synergistically to emulsion stabilization. Considering that ovalbumin, ovotransferrin, lysozyme and ovomucin are generally considered to be the main bearers of functional properties of EW, the next section focuses on the relative content, molecular structure, and protein-protein interactions of these proteins, as well as the effect of the emulsification operation process on the interfacial behavior of the proteins. Due to the large difference in polarity between the two sides of the oil-water interface, protein hydrophobicity may be a key factor affecting its emulsification properties, we also predicted and analyzed the hydrophobicity of each amino acid site of key proteins. Additionally, the influence of the arrangement of hydrophilic and hydrophobic structures in proteins on its interfacial behavior cannot be ignored, since proteins with different original structures may have different adsorption states at the interface . Structure-function analysis of key emulsification proteins Ovalbumin and ovalbumin-related proteins As the predominant protein in EW (∼54%), ovalbumin (OVA) exhibited unexpected behavior during emulsification. The abundance of OVA in EWE-W was higher than that of EW (1.67-fold, P < 0.05), indicating limited participation in interfacial film formation, and the results are in line with the expected outcomes based on existing research. As shown in C, positive values indicated hydrophobic intervals, and the opposite indicated the hydrophilic interval, hydrophobicity analysis of OVA reveals the presence of hydrophobic fragments at both ends of the chain, with the hydrophilic region relatively concentrated in the middle, whereas from the protein structure, as a globular structured protein, the hydrophobic structure of OVA tends to be embedded within the molecule . This structural arrangement may explain its reduced emulsification capacity under our experimental conditions. Within the OVA protein family, divergent behaviors were observed. Ovalbumin-related protein Y (OVA-Y) indicates low oil-water interface involvement also (1.43-fold, P < 0.01), with an estimated concentration of 13% of OVA in EW, belongs to the family of OVA serine protease inhibitors along with OVA and ovalbumin-related protein X (OVA-X) . Despite sharing similar amino acid sequence with OVA, OVA-X (43 kDa) was significantly less abundant in the EWE-W system (0.45-fold, P < 0.01) and belongs to the key proteins involved in emulsification. OVA-Y has a secondary structure similar to that of OVA, but its three-dimensional modeling reveals the presence of exposed positively charged clusters with a stronger affinity for heparin, suggesting that although OVA-X and OVA are highly similar in sequence, there may be specific sequences or structures that may affect their respective functions , which may be an important reason for the differences in emulsification properties between them. The results indicate that OVA, OVA-Y, and OVA-X have different behaviors at the oil-water interface, and that the modified structures as well as the structural domains may be the key factors affecting the emulsification properties of the proteins. The current structural knowledge about them is insufficient and more in-depth studies are needed. An appropriate ratio of hydrophobic and hydrophilic structural domains might be beneficial to the emulsification properties of protein molecules. Ovotransferrin As one of the major proteins in EW, ovotransferrin has been studied more on its bioactivity and less related to its emulsification properties. Quantitative experiments showed that the abundance of ovotransferrin in EWE-W (1.71-fold, P < 0.05) was significantly higher than that in EW, it did not exhibit good emulsifying activity. This may be due to the fact that no additional treatment of the EW system was performed in our experiments other than the necessary steps of the emulsification operation process. Most of the hydrophobic structural domains of ovotransferrin are hidden within the molecule , and the hydrophobic amino acid residues are difficult to expose under neutral or alkaline conditions. It has been shown that acid and acid-heat treatments are effective in improving the emulsification properties of ovotransferrin , which is attributed to the fact that globular proteins partially unfold the structure of the protein molecule under extremely acidic conditions to form a molten globular conformation, which leads to an increase in the flexibility of the molecule. Currently, some studies have attempted to prepare Pickering emulsions using ovotransferrin, which provides a new application for proteins with poor emulsification properties in food systems. Lysozyme Many studies have emphasized the antimicrobial role of lysozyme in food systems, but its effect on processing characteristics is of equal interest. Lysozyme (14.3 kDa) was the only positively charged protein in EW, and its abundance in EWE-W was 33.8% of EW ( P < 0.01), suggesting the involvement of lysozyme during the emulsification process. Hydrophobicity analyses showed significant hydrophobicity differences between the ends of the molecular chains of lysozyme proteins ( C) . Lysozyme has a rigid molecular structure with many intramolecular disulfide bonds and retains much of its natural conformation when adsorbed at the interface, which is not conducive to its adsorption at the oil-water interface. High-pressure homogenization leads to conformational changes in lysozyme, as evidenced by an increase in volume, surface area, and solvent-accessible surface area . Inducing the conversion of ordered α-helix structure to disordered structure in lysozyme molecule by emulsification process operation , will loosen the protein thus exposing more hydrophobic groups and improving the emulsification capacity. Small molecular weight proteins can adsorb to the oil-water interface through protein-protein interactions, and lysozyme is a typical small molecular weight protein in EW; therefore, a portion of lysozyme may be involved in the formation of the interfacial membrane by means of intermolecular forces. Overall, lysozyme emulsification activity is not only related to the nature of the protein molecule itself, but also to the effect of the processing method on the molecular structure. Ovomucin Ovomucin is thought to be the main reason for the thick gelatinous properties of liquid EW, accounting for about 3.5% of total EW protein. We found that the ovomucin α-subunit (mucin-5B) was only 28.9% as abundant in EWE-W as it was in EW ( P < 0.01). As highly glycosylated proteins, the emulsification properties of ovomucin are significantly influenced by their unique structure: in general, the hydrophobic region of the protein is responsible for anchoring to the oil-water interface and forming a viscoelastic layer, whereas the glycanchains, as hydrophilic structures, provide more spatial stability , and the two factors work together to influence the emulsifying properties of ovalbumin. In addition to adhesion and hydrophobicity, flexibility is another factor that may affects the emulsifying properties of ovomucin. Being a large molecular weight protein with low solubility in water, ovomucin still showed better emulsification performance in this study, perhaps due to a more flexible molecule, allowing it to undergo structural shifts adapted to the oil-water interfacial force field during the emulsification process, thus exposing more hydrophobic structures. In addition, under high-pressure homogenization, the droplet size is small, and it is unknown whether the flexible particle structure of ovomucin can participate in the formation of the oil-water interface by encapsulating the droplets. It is noteworthy that lysozyme-ovomucin in EW is generally present in a complex form . These complexes are sustained by a variety of intermolecular forces such as hydrogen bonding, disulfide bonding, electrostatic interactions and surface hydrophobic interactions. High-pressure homogenization promotes protein-protein interactions, and the lysozyme molecule can promotes the rearrangement of specific amino acid molecules of ovomucin and thus the formation of conformations conducive to interfacial adsorption . In addition, various protein complexes exist in addition to lysozyme-ovomucin complex, such as ova-ovomucin complex and ovotransferrin-ovomucin complex. The effect of their protein-protein interactions on the oil-water interface remains to be investigated. Protein TENP Protein TENP is a protein with a theoretical molecular weight of 47-50 kDa, a member of the BPI (bactericidal permeabilityincreasing protein) fold-containing family B (BPIFB) , and classified as BPIFB7 (member 7). It accounts for 0.1∼0.5% of total EW protein. Protein TENP was significantly less abundant in EWE-W (0.07-fold, P < 0.05), suggesting that it adsorbs faster to the interface during emulsification operations. Meanwhile, Protein TENP is also the protein with the largest change in abundance value reduction among all the proteins in the EW system, indicating that it plays a very important role during EW emulsification process. According to protein hydrophobicity analysis ( C), the presence of a large number of hydrophobic fragments favors the adsorption of Protein TENP at the oil-water interface. The results showed that despite the fact that Protein TENP is a minor protein in EW, its influence on the functionality of EW. In a quantitative proteomics study on the mechanism of enhancing the foaming properties of EW proteins by pH treatment, lysozyme, ovalbumin, and Protein TENP were the key proteins for the formation of excellent foaming at pH 2, pH 12, and pH 12-7 , and in one of our previous studies on the potential mechanisms underlying the differences in thermogenic gel properties between thick and thin proteins, Protein TENP was found to be more abundant (26.6%) in thick EW than in thin EW, which was associated with EW gel properties . Through comprehensive proteomic analysis of EW emulsification process, this study reveals several critical insights into protein-interface interactions in complex food systems. Most significantly, we discovered that minor proteins, particularly Protein TENP, play pivotal roles in emulsion formation and stability, challenging the traditional focus on abundant proteins. Our findings demonstrate that interfacial activity is primarily determined by protein structural characteristics and molecular flexibility rather than initial abundance, with high-pressure homogenization enhancing functionality through conformational changes and protein-protein interactions. The complex interplay between proteins, especially Lysozyme-Ovomucin interactions, contributes substantially to interface stability through multiple molecular mechanisms. These insights not only advance our understanding of natural protein emulsification systems but also provide new directions for optimizing EW functionality in food applications.
EW and EWE-W were analyzed using 4D-SmartDIA quantitative proteomics technology, and a total of 96 proteins were identified. Principal component analysis (PCA) revealed distinct clustering patterns between EW and EWE-W groups with excellent reproducibility within groups ( B). PC1 and PC2 accounted for 72.8% and 7.9% of the total variance, respectively, demonstrating substantial alterations in protein composition during the emulsification process. To identify proteins crucial for emulsification, differentially abundant proteins were screened using stringent criteria (fold change > 1.50 or < 0.67, P < 0.05). In EWE-W, a total of 64 proteins were found to be significantly associated with EW emulsification capacity, of which 20 were increased in and 44 were decreased in abundance ( C). The predominant decrease in protein abundance in EWE-W suggests that the majority of emulsification-active proteins were preferentially recruited to the oil-water interface. These differential proteins may be key factors influencing the emulsification properties of EW, and their behavior at the oil-water interface deserves further exploration.
Quantitative proteomic analysis revealed distinct patterns of protein abundance changes between EW and EWE-W systems, providing insights into the molecular mechanisms of EW emulsification. Several proteins exhibited significantly increased abundance in EWE-W, including Ig-like domain-containing protein (8.65-fold, P < 0.01), PIT54 protein (3.40-fold, P < 0.01) and Mucin-6 (Fragment) (2.28-fold, P < 0.05), etc. ( A). The enhanced abundance of these proteins in the aqueous phase indicated their limited participation in the oil-water interface, possibly due to structural characteristics unfavorable for oil-water interface stabilization. More notably, a larger subset of proteins showed significantly decreased abundance in EWE-W, suggesting that they may be actively involved in the formation of the oil-water interface. Among them, the Protein TENP showed the most significant change in abundance (0.07-fold, P < 0.05) among all abundance-reducing proteins ( B), suggesting its strong affinity for the oil-water interface. Although the Protein TENP is a non-dominant protein in EW, it plays a crucial role in the emulsification process. In addition, proteins with significantly decreased abundance include lipid-binding serum glycoprotein C-terminal domain-containing protein (0.11-fold, P < 0.05), Glutathione peroxidase (0.12-fold, P < 0.05), acyloxyacyl hydrolase (0.12-fold, P < 0.05) and SRCR domain-containing protein (0.14-fold, P < 0.05), etc. The similar magnitude of reduction among these proteins suggests a coordinated mechanism in interface formation, where multiple proteins contribute synergistically to emulsion stabilization. Considering that ovalbumin, ovotransferrin, lysozyme and ovomucin are generally considered to be the main bearers of functional properties of EW, the next section focuses on the relative content, molecular structure, and protein-protein interactions of these proteins, as well as the effect of the emulsification operation process on the interfacial behavior of the proteins. Due to the large difference in polarity between the two sides of the oil-water interface, protein hydrophobicity may be a key factor affecting its emulsification properties, we also predicted and analyzed the hydrophobicity of each amino acid site of key proteins. Additionally, the influence of the arrangement of hydrophilic and hydrophobic structures in proteins on its interfacial behavior cannot be ignored, since proteins with different original structures may have different adsorption states at the interface .
Ovalbumin and ovalbumin-related proteins As the predominant protein in EW (∼54%), ovalbumin (OVA) exhibited unexpected behavior during emulsification. The abundance of OVA in EWE-W was higher than that of EW (1.67-fold, P < 0.05), indicating limited participation in interfacial film formation, and the results are in line with the expected outcomes based on existing research. As shown in C, positive values indicated hydrophobic intervals, and the opposite indicated the hydrophilic interval, hydrophobicity analysis of OVA reveals the presence of hydrophobic fragments at both ends of the chain, with the hydrophilic region relatively concentrated in the middle, whereas from the protein structure, as a globular structured protein, the hydrophobic structure of OVA tends to be embedded within the molecule . This structural arrangement may explain its reduced emulsification capacity under our experimental conditions. Within the OVA protein family, divergent behaviors were observed. Ovalbumin-related protein Y (OVA-Y) indicates low oil-water interface involvement also (1.43-fold, P < 0.01), with an estimated concentration of 13% of OVA in EW, belongs to the family of OVA serine protease inhibitors along with OVA and ovalbumin-related protein X (OVA-X) . Despite sharing similar amino acid sequence with OVA, OVA-X (43 kDa) was significantly less abundant in the EWE-W system (0.45-fold, P < 0.01) and belongs to the key proteins involved in emulsification. OVA-Y has a secondary structure similar to that of OVA, but its three-dimensional modeling reveals the presence of exposed positively charged clusters with a stronger affinity for heparin, suggesting that although OVA-X and OVA are highly similar in sequence, there may be specific sequences or structures that may affect their respective functions , which may be an important reason for the differences in emulsification properties between them. The results indicate that OVA, OVA-Y, and OVA-X have different behaviors at the oil-water interface, and that the modified structures as well as the structural domains may be the key factors affecting the emulsification properties of the proteins. The current structural knowledge about them is insufficient and more in-depth studies are needed. An appropriate ratio of hydrophobic and hydrophilic structural domains might be beneficial to the emulsification properties of protein molecules.
As the predominant protein in EW (∼54%), ovalbumin (OVA) exhibited unexpected behavior during emulsification. The abundance of OVA in EWE-W was higher than that of EW (1.67-fold, P < 0.05), indicating limited participation in interfacial film formation, and the results are in line with the expected outcomes based on existing research. As shown in C, positive values indicated hydrophobic intervals, and the opposite indicated the hydrophilic interval, hydrophobicity analysis of OVA reveals the presence of hydrophobic fragments at both ends of the chain, with the hydrophilic region relatively concentrated in the middle, whereas from the protein structure, as a globular structured protein, the hydrophobic structure of OVA tends to be embedded within the molecule . This structural arrangement may explain its reduced emulsification capacity under our experimental conditions. Within the OVA protein family, divergent behaviors were observed. Ovalbumin-related protein Y (OVA-Y) indicates low oil-water interface involvement also (1.43-fold, P < 0.01), with an estimated concentration of 13% of OVA in EW, belongs to the family of OVA serine protease inhibitors along with OVA and ovalbumin-related protein X (OVA-X) . Despite sharing similar amino acid sequence with OVA, OVA-X (43 kDa) was significantly less abundant in the EWE-W system (0.45-fold, P < 0.01) and belongs to the key proteins involved in emulsification. OVA-Y has a secondary structure similar to that of OVA, but its three-dimensional modeling reveals the presence of exposed positively charged clusters with a stronger affinity for heparin, suggesting that although OVA-X and OVA are highly similar in sequence, there may be specific sequences or structures that may affect their respective functions , which may be an important reason for the differences in emulsification properties between them. The results indicate that OVA, OVA-Y, and OVA-X have different behaviors at the oil-water interface, and that the modified structures as well as the structural domains may be the key factors affecting the emulsification properties of the proteins. The current structural knowledge about them is insufficient and more in-depth studies are needed. An appropriate ratio of hydrophobic and hydrophilic structural domains might be beneficial to the emulsification properties of protein molecules.
As one of the major proteins in EW, ovotransferrin has been studied more on its bioactivity and less related to its emulsification properties. Quantitative experiments showed that the abundance of ovotransferrin in EWE-W (1.71-fold, P < 0.05) was significantly higher than that in EW, it did not exhibit good emulsifying activity. This may be due to the fact that no additional treatment of the EW system was performed in our experiments other than the necessary steps of the emulsification operation process. Most of the hydrophobic structural domains of ovotransferrin are hidden within the molecule , and the hydrophobic amino acid residues are difficult to expose under neutral or alkaline conditions. It has been shown that acid and acid-heat treatments are effective in improving the emulsification properties of ovotransferrin , which is attributed to the fact that globular proteins partially unfold the structure of the protein molecule under extremely acidic conditions to form a molten globular conformation, which leads to an increase in the flexibility of the molecule. Currently, some studies have attempted to prepare Pickering emulsions using ovotransferrin, which provides a new application for proteins with poor emulsification properties in food systems.
Many studies have emphasized the antimicrobial role of lysozyme in food systems, but its effect on processing characteristics is of equal interest. Lysozyme (14.3 kDa) was the only positively charged protein in EW, and its abundance in EWE-W was 33.8% of EW ( P < 0.01), suggesting the involvement of lysozyme during the emulsification process. Hydrophobicity analyses showed significant hydrophobicity differences between the ends of the molecular chains of lysozyme proteins ( C) . Lysozyme has a rigid molecular structure with many intramolecular disulfide bonds and retains much of its natural conformation when adsorbed at the interface, which is not conducive to its adsorption at the oil-water interface. High-pressure homogenization leads to conformational changes in lysozyme, as evidenced by an increase in volume, surface area, and solvent-accessible surface area . Inducing the conversion of ordered α-helix structure to disordered structure in lysozyme molecule by emulsification process operation , will loosen the protein thus exposing more hydrophobic groups and improving the emulsification capacity. Small molecular weight proteins can adsorb to the oil-water interface through protein-protein interactions, and lysozyme is a typical small molecular weight protein in EW; therefore, a portion of lysozyme may be involved in the formation of the interfacial membrane by means of intermolecular forces. Overall, lysozyme emulsification activity is not only related to the nature of the protein molecule itself, but also to the effect of the processing method on the molecular structure.
Ovomucin is thought to be the main reason for the thick gelatinous properties of liquid EW, accounting for about 3.5% of total EW protein. We found that the ovomucin α-subunit (mucin-5B) was only 28.9% as abundant in EWE-W as it was in EW ( P < 0.01). As highly glycosylated proteins, the emulsification properties of ovomucin are significantly influenced by their unique structure: in general, the hydrophobic region of the protein is responsible for anchoring to the oil-water interface and forming a viscoelastic layer, whereas the glycanchains, as hydrophilic structures, provide more spatial stability , and the two factors work together to influence the emulsifying properties of ovalbumin. In addition to adhesion and hydrophobicity, flexibility is another factor that may affects the emulsifying properties of ovomucin. Being a large molecular weight protein with low solubility in water, ovomucin still showed better emulsification performance in this study, perhaps due to a more flexible molecule, allowing it to undergo structural shifts adapted to the oil-water interfacial force field during the emulsification process, thus exposing more hydrophobic structures. In addition, under high-pressure homogenization, the droplet size is small, and it is unknown whether the flexible particle structure of ovomucin can participate in the formation of the oil-water interface by encapsulating the droplets. It is noteworthy that lysozyme-ovomucin in EW is generally present in a complex form . These complexes are sustained by a variety of intermolecular forces such as hydrogen bonding, disulfide bonding, electrostatic interactions and surface hydrophobic interactions. High-pressure homogenization promotes protein-protein interactions, and the lysozyme molecule can promotes the rearrangement of specific amino acid molecules of ovomucin and thus the formation of conformations conducive to interfacial adsorption . In addition, various protein complexes exist in addition to lysozyme-ovomucin complex, such as ova-ovomucin complex and ovotransferrin-ovomucin complex. The effect of their protein-protein interactions on the oil-water interface remains to be investigated.
Protein TENP is a protein with a theoretical molecular weight of 47-50 kDa, a member of the BPI (bactericidal permeabilityincreasing protein) fold-containing family B (BPIFB) , and classified as BPIFB7 (member 7). It accounts for 0.1∼0.5% of total EW protein. Protein TENP was significantly less abundant in EWE-W (0.07-fold, P < 0.05), suggesting that it adsorbs faster to the interface during emulsification operations. Meanwhile, Protein TENP is also the protein with the largest change in abundance value reduction among all the proteins in the EW system, indicating that it plays a very important role during EW emulsification process. According to protein hydrophobicity analysis ( C), the presence of a large number of hydrophobic fragments favors the adsorption of Protein TENP at the oil-water interface. The results showed that despite the fact that Protein TENP is a minor protein in EW, its influence on the functionality of EW. In a quantitative proteomics study on the mechanism of enhancing the foaming properties of EW proteins by pH treatment, lysozyme, ovalbumin, and Protein TENP were the key proteins for the formation of excellent foaming at pH 2, pH 12, and pH 12-7 , and in one of our previous studies on the potential mechanisms underlying the differences in thermogenic gel properties between thick and thin proteins, Protein TENP was found to be more abundant (26.6%) in thick EW than in thin EW, which was associated with EW gel properties . Through comprehensive proteomic analysis of EW emulsification process, this study reveals several critical insights into protein-interface interactions in complex food systems. Most significantly, we discovered that minor proteins, particularly Protein TENP, play pivotal roles in emulsion formation and stability, challenging the traditional focus on abundant proteins. Our findings demonstrate that interfacial activity is primarily determined by protein structural characteristics and molecular flexibility rather than initial abundance, with high-pressure homogenization enhancing functionality through conformational changes and protein-protein interactions. The complex interplay between proteins, especially Lysozyme-Ovomucin interactions, contributes substantially to interface stability through multiple molecular mechanisms. These insights not only advance our understanding of natural protein emulsification systems but also provide new directions for optimizing EW functionality in food applications.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
|
COVID-19 Marked a Change in the Scope of Occupational Medicine from Occupational to Work-Related Diseases and Total Worker Health | 7f4561c5-ead3-4cfd-99da-f8523180f7b5 | 10731568 | Preventive Medicine[mh] | Occupational Health has witnessed a significant change in paradigm over the years, shifting from a focus on occupational medicine to a comprehensive approach known as Total Worker Health ® (TWH). This transition has been driven by recognizing that work-related diseases gradually replaced occupational diseases in the discipline’s scope. This has led to the unprecedented inclusion of COVID-19 as a significant concern for occupational health . Amid a new and potentially lethal infectious disease affecting all strata of the general population, and the first report covering the 1 st semester of 2020 reporting an 11.1% excess mortality in Italy and an almost 50% excess in Lombardy, the most affected region, we decided to publish a series of papers monitoring mortality because of its relevant implications for controlling the time-course of the COVID-19 pandemic . Social distancing, facial masks, and other measures that aimed at preventing COVID-19 also prevented the usual influenza epidemics in 2021; nevertheless, a 7.9% excess mortality was observed; of these deaths, 3,667 occurred among individuals of working age (25-64 years) . In the 2020-2022 triennium, 225,965 deaths exceeded expected rates, and 16,017 of these occurred in working age . Data on total mortality for the first half of 2023 suggested a rebound due to harvesting in previous years, 6,947 and 1,879 lower than expected in the general population and working age, respectively . Increased mortality during the COVID-19 pandemic period could be due to delayed or missing access to treatment of highly prevalent chronic diseases, such as cardiovascular diseases, cancer, and diabetes, as well as to other factors such as fear of facing contact with other people and attending crowded places, and social anxiety . Occupational Medicine is critical in identifying vulnerable workers during pandemics, as it actually did during Sars-CoV-2 pandemics. A series of articles have highlighted specific occupational groups that face an increased risk of infection due to their proximity to infected individuals or their involvement in essential services. Individuals working in crowded and confined environments, such as taxi and bus drivers, have shown to be more susceptible to exposure. Healthcare workers, first responders, and frontline workers have also been identified as groups at higher risk, which justifies labeling COVID-19 as a work-related disease . The accurate and timely diagnosis of COVID-19 is paramount for effective disease management and prevention of further transmission. Various diagnostic methods, such as PCR antigen and antibody testing for diagnosing COVID-19 cases, have been discussed to improve their use and to choose the best one among available technologies, depending on the application context, to help occupational health professionals safeguard workers’ health . Preventing the spread of COVID-19 within workplaces is crucial to protect workers and maintain business continuity. Efforts to implement preventive measures such as personal protective equipment (PPE), social distancing, improved ventilation systems, and sanitization protocols emphasized the importance of comprehensive infection control strategies within occupational settings. Understanding the modality of transmission and ruling out seemingly obvious pathways has been essential to focus prevention measures . As the pandemic evolves, new variants of SARS-CoV-2 keep emerging, with potentially different levels of virulence and lethality. Articles published in this journal have addressed these variations within the pandemic scenario, providing insights into their impact on workers’ health. Occupational medicine must promptly adapt to evolving strains to ensure adequate risk assessment and management strategies. The COVID-19 pandemic has posed an unprecedented challenge to global health, economies, and societies. However, amidst the chaos, the rapid development and deployment of effective vaccines have emerged as a game-changer strategy in our fight against the virus. Vaccines have been pivotal in mitigating the impact of the pandemic, promoting population’s immunity, and favoring the diffusion of more contagious but less lethal variants of SARS-CoV-2 . Occupational health practitioners played a relevant role in promoting vaccination acceptance and implementation among workers. This contributed to minimizing workplace transmission risks, owing to vaccine efficacy, safety, and effective distribution strategies . While the initial focus has been on acute infection and mortality rates, understanding the long-term effects of COVID-19 is essential for managing occupational health. Persistent symptoms experienced by individuals who have recovered from the acute phase emphasize the need for long-term monitoring of affected workers, especially during the back-to-work phase . Psychosocial factors associated with COVID-19 at the workplace or distance working can significantly impact employees’ well-being and mental health. The sudden shift to remote work or changes in responsibilities and procedures due to the pandemic often led to increased workload and job demands. This has been able to cause stress (including technostress), burnout, and decreased job satisfaction . Balancing work responsibilities with personal life has often been challenging during the pandemic, especially when working from home. The boundaries between work and personal life has sometimes become blurred, leading to increased stress and difficulty in disconnecting from work. Remote work can result in social isolation and reduced opportunities for social interaction with colleagues. Lack of social support and limited communication can impact mental well-being and increase feelings of loneliness. Remote work may present challenges in effective communication, collaboration, and teamwork. Miscommunication or difficulties in getting timely responses from supervisors or colleagues can hinder productivity and create frustration among employees . Dependence on technology for remote work can lead to technical difficulties, connectivity issues, or inadequate infrastructure support, including issues related to the environment in which work occurs, originally not designed for that purpose. These challenges, particularly difficult to manage by occupational physicians, could add further strain on employees’ mental well-being . COVID-19 has completed the transition of the scope of Occupational Medicine from primarily addressing occupational diseases to encompassing a much wider veriety of work-related illnesses. Articles published in our journal provided valuable insights into various aspects of this challenging disease. From epidemiological studies to high-risk groups identification, effective diagnostic methods, preventive measures, vaccination efforts, long-term effects monitoring, and considerations regarding changes in virulence and lethality, but above all the duration of SARS-CoV-2 shedding and infectivity in working populations , these articles underscore the critical role that occupational medicine plays during a global pandemic like COVID-19. Where applied, the traditional principles of public or occupational health already applied to protect workers from air pollution in several working settings, e.g., in mines, metallurgies, chemical industries – otherwise flawed and often overlooked – proved to be effective in controlling infection spreading . Therefore, it is important implementing these principles to ensure clean air in workplaces and other settings, leaving facial masks as a last resort to protect against infections. The rapid development of effective COVID-19 vaccines is a testament to human resilience, scientific advancements, and global collaboration. It has provided us with a powerful tool to combat the struggles imposed by this deadly virus. Through accelerated vaccine development processes, stringent clinical trials, global collaboration efforts, and successful vaccination campaigns, we are now on the path toward recovery. However, it is essential to continue addressing vaccine hesitancy, ensuring equitable distribution worldwide, monitoring virus variants vigilantly, and adapting vaccination strategies accordingly. Occupational physicians are playing a new Public Health role. The paradigm shifts from occupational diseases to TWH ® – embracing work-related disorders as an intermediate step to broaden the scope of Occupational Medicine – and this is a crucial milestone for Occupational Health . The inclusion of COVID-19 as a work-related disease has also highlighted the need for a comprehensive approach that addresses not only physical health but also mental and social well-being. By embracing health promotion in the framework of TWH ® , occupational health professionals can create safer and healthier workplaces, ultimately benefiting employees and organizations. |
Multifaceted impact of COVID-19 on dental practice | e52a941d-4d94-4e7c-a95b-b698ae5d6ea0 | 8520822 | Dental[mh] | We engaged dental care practitioners who are members of Texas Dental Association (TDA), the San Antonio District Dental Society (SADDS), and the South Texas Oral Health Network (STOHN), a Texas dental practice-based research network, with an observational survey study approved by The University of Texas Health Science Center at San Antonio Institutional Review Board (protocol HSC20200374E). We piloted the survey, which consisted of 25 items, among 3 through 5 dental care practitioners to test its accessibility and usability, as well as the time for completion. We delivered the survey study online using the web-based survey platform Qualtrics XM (Version August 2020) and distributed it via an anonymous survey link to members of the professional organizations. The Qualtrics survey was protected from ballot box stuffing to prevent participants from taking the survey more than once, given that a member of 1 participating organization may also have been a member of another participating organization. Study recruitment took place from August 14, 2020, through October 19, 2020. Responses were voluntary and anonymous, and participants were not offered any incentive to complete the survey. Dental care practitioners received an email invitation (that also encouraged participation of their dental staff members) from the representative of the participating organizations explaining the study and inviting them to participate voluntarily using a web link to the questionnaire. Participants were licensed in the United States and members of dental associations or research networks and maintained an active email address at which they could be contacted. Variables assessed in the survey included participant demographics, sources of information related to COVID-19, resources and working satisfaction, concerns, challenges, and the impact of COVID-19. Participants were also provided with a section to leave comments about their experience during the COVID-19 pandemic. A subset of participants, those who are involved in clinical dental practice or direct patient care, were asked additional clinically related questions about challenges, concerns, treatment changes, safety precautions, and preparedness while resuming dental practice, as well as views on dentistry as a career that has been affected during the pandemic. The survey is presented in , available online at the end of this article. Scales for questions and methodology used to compile responses into positive versus negative responses are described in the results section. Statistical analysis We analyzed the data obtained from the survey using R statistical computing software (R Foundation for Statistical Computing). We conducted statistical comparisons of responses by the following demographic groups: sex, age, and years of experience in dental practice. The reported P values are from the χ 2 test. In addition to P values, we report 95% CIs for the proportions in each group, along with odds ratios (ORs) comparing the group with the higher rate with the group with the lower rate. For sentiment analysis of the participants’ comments, we used Wilcoxon rank sum test (nonparametric test) to compare whether the overall sentiment of the group deviated from neutral (alternative hypothesis: true location shift is not equal to 0). Hypotheses related to the survey were that dental care professionals were concerned about the impact of COVID-19 on their dental practices and their own well-being but still were comfortable with their dental career choices and confidently prepared to resume dental practice despite concerns about challenges, long-term impacts, financial resources, and efficiency. We used a level of significance of P equaling .05 for all statistical tests.
We analyzed the data obtained from the survey using R statistical computing software (R Foundation for Statistical Computing). We conducted statistical comparisons of responses by the following demographic groups: sex, age, and years of experience in dental practice. The reported P values are from the χ 2 test. In addition to P values, we report 95% CIs for the proportions in each group, along with odds ratios (ORs) comparing the group with the higher rate with the group with the lower rate. For sentiment analysis of the participants’ comments, we used Wilcoxon rank sum test (nonparametric test) to compare whether the overall sentiment of the group deviated from neutral (alternative hypothesis: true location shift is not equal to 0). Hypotheses related to the survey were that dental care professionals were concerned about the impact of COVID-19 on their dental practices and their own well-being but still were comfortable with their dental career choices and confidently prepared to resume dental practice despite concerns about challenges, long-term impacts, financial resources, and efficiency. We used a level of significance of P equaling .05 for all statistical tests.
We implemented the survey study via use of organizational distribution avenues: TDA, SADDS, and STOHN. We invited 7,805 dental care practitioners to participate in this survey study. Of 7,805 invited members, 622 participated in the survey (different response rates from the participating organizations: 73% response rate from STOHN [120 members], 37% from SADDS [920 members], and a combined response rate of 8% for all participating organizations when combined with the response rate from TDA, which has thousands of members). A total of 572 participants completed the first part of the survey (general questions). Of that group, 514 (90%) completed the second clinical practice–related portion. At the time of the survey, 42 (7%) of total participants reported that they or a household member had tested positive for COVID-19. shows demographics as reported by participants. shows that 71% of participants were general dentists, 24% were specialists, and 5% were administration or allied dental staff members (because most of the members of participating organizations were dentists). Four of 10 participants in the first part of the survey identified as female, 44% were 55 years or older, 79% had been in dental practice at least 10 years, and 95% were dentists. Participants of the clinical practice–related second part had nearly identical demographics as the participants of the first part. General challenges during the pandemic A series of questions asked participants to provide their opinions and beliefs using a 5-point scale. We used a common standard methodology for surveys to further compile the positive responses into 1 and the negative responses into another: a positive response included words of agreement (might, slightly, somewhat, moderately very, strongly agree, yes), and a negative response included words of disagreement (neither, might not disagree, not, no). The results of this part of the survey are shown in . More than one-half the participants found working during the pandemic difficult (≈ 60%). Participants were concerned about the impact of COVID-19 on their general health, safety, and well-being (77%) and even more so about the impact to their dental practice (90%). However, 95% of participants were satisfied with the resources their dental office provided to help support them during the pandemic. There was a statistically significant difference between men and women regarding concern about general health, with 83% of women being concerned (95% CI, 78% to 88%) compared with 72% of men being concerned (95% CI, 67% to 77%) (OR, 1.86; P = .003). There was also a statistically significant difference by age, with participants in the highest age group (> 55 years) being more likely to respond that work was difficult during the COVID-19 pandemic (66% [95% CI, 60% to 72%] in oldest age group versus 56% [95% CI, 51% to 61%] in other age groups [OR, 1.55; P = .043]). This possibly was because older adults (with preexisting health conditions) seem to be at a higher risk of developing more serious complications from COVID-19. There were no statistically significant differences by job function or years of experience. Another question asked participants to rank the importance of their main sources of information to keep updated regarding COVID-19 ( ). We selected 2 methods to evaluate the responses. The first was average importance, in which a value of 1 was assigned to the most important resource, 2 to the second most important, and so on, after which a mean value of the importance score across all participants was calculated. The second was favorability rating, which reported the percentage of participants who ranked the resource in the top or bottom 2. The participants rated the COVID-19 websites and internet search engines in the top 2 of importance more often (84%) than any other resource and had the fewest number of members rate it at the bottom. Journals and publications, despite the longer lead time to publish, were rated in the top 2 by 58% of the participants. Television channels and news were rated at the bottom by more than one-half of the participants. Social media sites had the lowest favorability rating, possibly a reflection of the generally older age of the participants. Challenges related to resuming dental practice Participants in the clinical practice–related part of the survey were asked to respond using a 5-point scale to 8 statements and questions regarding their opinions and beliefs on challenges related to dental practice during the pandemic. Responses are presented in , grouped into 2 response levels as described above. Results of importance ▪ Nearly all participants (98%) felt at least slightly prepared to resume dental practice. There was no statistical difference by age, sex, or years of experience. ▪ More than 75% of participants agreed (somewhat or strongly) that there were challenges and long-term impacts on dentistry that needed to be considered. There was a statistically significant split in the strength of those opinions by sex; challenges were endorsed by 83% (95% CI, 78% to 88%) of women versus 72% (95% CI, 67% to 77%) of men (OR, 1.91; P = .006), and long-term impacts were identified by 92% (95% CI, 88% to 96%) of women versus 76% (95% CI, 71% to 81%) of men (OR, 3.56; P < .001). ▪ More than 95% of all participants were slightly or more confident of the safety precautions and PPE provided by their dental practice to protect them while performing aerosol-generating dental procedures. There were no significant differences by any demographic group. ▪ Only 21% (95% CI, 17% to 24%) of dentists felt that COVID-19 changed types of dental treatments offered to patients, perhaps because of the strict safety precautions and infection control measures already practiced by dentists who are, in general, at higher risk of developing infections and, thus, are used to handling risks of developing infections properly. Women were also more likely than men to endorse changes in dental treatments (30% of women [95% CI, 24% to 37%] versus 18% of men [95% CI, 14% to 22%] [OR, 1.97, P = .002]). At least two-thirds of all participants agreed that the precautions would influence their efficiency adversely, with no statistical evidence of difference by demographic group. ▪ Approximately 70% would probably or definitely not reconsider their dental career choices and would still recommend studying dentistry to others. Women were more ambivalent about their career choices, with 36% (95% CI, 29% to 42%) unsure or thinking of reconsidering compared with 25% (95% CI, 20% to 30%) of men (OR, 1.66; P = .013). Participants were also asked to rank the importance of 9 challenges while resuming dental practice during the pandemic ( ). More than 60% of all participants rated PPE availability and having patients willing to come in for dental treatment in the top 3 of importance, far more often than any other challenge. The process of testing dental staff members was rated as of least concern, with two-thirds of all participants rating this in the bottom 3. Substantial differences by demographic group Grouping answers by top 3 and bottom 3 importance is a useful tool for comparing the challenges, but there has not been a statistical method established for this particular type of analysis. However, it is appropriate to identify challenges with a substantial difference in opinion between demographic groups. For groupings that were relatively balanced (sex, age, years of experience), we decided to use a 20 percentage point difference as evidence of a substantial difference. ▪ There were no substantial differences in opinion by sex or years of experience. ▪ There were 2 substantial differences by age group. Participants older than 55 years were more likely than those 18 through 34 years to rate patients willing to come in for treatment as a top 3 challenge than a bottom 3 challenge (69% versus 47%). The youngest age group was less likely to identify financial losses of the dental practice as a top 3 challenge (36% compared with 56% among all the other participants). ▪ Only one-fourth of dentists rated concern about outbreaks and spikes as a top 3 concern. ▪ Testing patients for COVID-19 infection was a low concern for dentists, with nearly two-thirds rating it in the bottom 3. ▪ 55% of dentists were worried about the financial status of the dental practice and rated it in the top 3, and 23% of dentists rated it in the bottom 3. ▪ 43% of dentists rated concern about the resources of patients in the bottom 3. Participants’ feedback and sentiment analysis The final question of the survey asked participants whether they had any other comments regarding their experiences during the pandemic. We performed sentiment analysis of the comments on the basis of the participants’ words; we extracted the individual words from the comments and identified words that indicate positive or negative sentiment to perform the sentiment analysis ( and ). Sentiment comparison The overall sentiment for each provider based on the number of positive or negative words used in their comment ranged from –8 through 8 with the median being neutral. The overall mean was –0.19 with a standard error of 2.4. We used Wilcoxon rank sum test to compare whether the overall sentiment of the group deviated from neutral. We found that it did not. We found no statistically significant overall negative sentiment for general dentists compared with specialists.
A series of questions asked participants to provide their opinions and beliefs using a 5-point scale. We used a common standard methodology for surveys to further compile the positive responses into 1 and the negative responses into another: a positive response included words of agreement (might, slightly, somewhat, moderately very, strongly agree, yes), and a negative response included words of disagreement (neither, might not disagree, not, no). The results of this part of the survey are shown in . More than one-half the participants found working during the pandemic difficult (≈ 60%). Participants were concerned about the impact of COVID-19 on their general health, safety, and well-being (77%) and even more so about the impact to their dental practice (90%). However, 95% of participants were satisfied with the resources their dental office provided to help support them during the pandemic. There was a statistically significant difference between men and women regarding concern about general health, with 83% of women being concerned (95% CI, 78% to 88%) compared with 72% of men being concerned (95% CI, 67% to 77%) (OR, 1.86; P = .003). There was also a statistically significant difference by age, with participants in the highest age group (> 55 years) being more likely to respond that work was difficult during the COVID-19 pandemic (66% [95% CI, 60% to 72%] in oldest age group versus 56% [95% CI, 51% to 61%] in other age groups [OR, 1.55; P = .043]). This possibly was because older adults (with preexisting health conditions) seem to be at a higher risk of developing more serious complications from COVID-19. There were no statistically significant differences by job function or years of experience. Another question asked participants to rank the importance of their main sources of information to keep updated regarding COVID-19 ( ). We selected 2 methods to evaluate the responses. The first was average importance, in which a value of 1 was assigned to the most important resource, 2 to the second most important, and so on, after which a mean value of the importance score across all participants was calculated. The second was favorability rating, which reported the percentage of participants who ranked the resource in the top or bottom 2. The participants rated the COVID-19 websites and internet search engines in the top 2 of importance more often (84%) than any other resource and had the fewest number of members rate it at the bottom. Journals and publications, despite the longer lead time to publish, were rated in the top 2 by 58% of the participants. Television channels and news were rated at the bottom by more than one-half of the participants. Social media sites had the lowest favorability rating, possibly a reflection of the generally older age of the participants.
Participants in the clinical practice–related part of the survey were asked to respond using a 5-point scale to 8 statements and questions regarding their opinions and beliefs on challenges related to dental practice during the pandemic. Responses are presented in , grouped into 2 response levels as described above.
▪ Nearly all participants (98%) felt at least slightly prepared to resume dental practice. There was no statistical difference by age, sex, or years of experience. ▪ More than 75% of participants agreed (somewhat or strongly) that there were challenges and long-term impacts on dentistry that needed to be considered. There was a statistically significant split in the strength of those opinions by sex; challenges were endorsed by 83% (95% CI, 78% to 88%) of women versus 72% (95% CI, 67% to 77%) of men (OR, 1.91; P = .006), and long-term impacts were identified by 92% (95% CI, 88% to 96%) of women versus 76% (95% CI, 71% to 81%) of men (OR, 3.56; P < .001). ▪ More than 95% of all participants were slightly or more confident of the safety precautions and PPE provided by their dental practice to protect them while performing aerosol-generating dental procedures. There were no significant differences by any demographic group. ▪ Only 21% (95% CI, 17% to 24%) of dentists felt that COVID-19 changed types of dental treatments offered to patients, perhaps because of the strict safety precautions and infection control measures already practiced by dentists who are, in general, at higher risk of developing infections and, thus, are used to handling risks of developing infections properly. Women were also more likely than men to endorse changes in dental treatments (30% of women [95% CI, 24% to 37%] versus 18% of men [95% CI, 14% to 22%] [OR, 1.97, P = .002]). At least two-thirds of all participants agreed that the precautions would influence their efficiency adversely, with no statistical evidence of difference by demographic group. ▪ Approximately 70% would probably or definitely not reconsider their dental career choices and would still recommend studying dentistry to others. Women were more ambivalent about their career choices, with 36% (95% CI, 29% to 42%) unsure or thinking of reconsidering compared with 25% (95% CI, 20% to 30%) of men (OR, 1.66; P = .013). Participants were also asked to rank the importance of 9 challenges while resuming dental practice during the pandemic ( ). More than 60% of all participants rated PPE availability and having patients willing to come in for dental treatment in the top 3 of importance, far more often than any other challenge. The process of testing dental staff members was rated as of least concern, with two-thirds of all participants rating this in the bottom 3.
Grouping answers by top 3 and bottom 3 importance is a useful tool for comparing the challenges, but there has not been a statistical method established for this particular type of analysis. However, it is appropriate to identify challenges with a substantial difference in opinion between demographic groups. For groupings that were relatively balanced (sex, age, years of experience), we decided to use a 20 percentage point difference as evidence of a substantial difference. ▪ There were no substantial differences in opinion by sex or years of experience. ▪ There were 2 substantial differences by age group. Participants older than 55 years were more likely than those 18 through 34 years to rate patients willing to come in for treatment as a top 3 challenge than a bottom 3 challenge (69% versus 47%). The youngest age group was less likely to identify financial losses of the dental practice as a top 3 challenge (36% compared with 56% among all the other participants). ▪ Only one-fourth of dentists rated concern about outbreaks and spikes as a top 3 concern. ▪ Testing patients for COVID-19 infection was a low concern for dentists, with nearly two-thirds rating it in the bottom 3. ▪ 55% of dentists were worried about the financial status of the dental practice and rated it in the top 3, and 23% of dentists rated it in the bottom 3. ▪ 43% of dentists rated concern about the resources of patients in the bottom 3.
The final question of the survey asked participants whether they had any other comments regarding their experiences during the pandemic. We performed sentiment analysis of the comments on the basis of the participants’ words; we extracted the individual words from the comments and identified words that indicate positive or negative sentiment to perform the sentiment analysis ( and ).
The overall sentiment for each provider based on the number of positive or negative words used in their comment ranged from –8 through 8 with the median being neutral. The overall mean was –0.19 with a standard error of 2.4. We used Wilcoxon rank sum test to compare whether the overall sentiment of the group deviated from neutral. We found that it did not. We found no statistically significant overall negative sentiment for general dentists compared with specialists.
Owing to a high potential risk of transmitting COVID-19 in dental clinics, a pause on dental practice and strict infection control protocols were implemented early in the pandemic to ensure a healthy, safe environment and mitigate the spread of the virus. Dental care practitioners are at high risk of contracting and transmitting COVID-19 owing to the peculiar nature of dental procedures and direct face-to-face communication with patients. , , , COVID-19 has affected dental practice substantially, with high levels of stress and anxiety reported among dental care practitioners and with anticipated long-term effects on dentistry. , , Worldwide, many countries have evaluated the psychological impact of COVID-19 and reported high levels of stress, fear, and anxiety. , , In agreement with these reports, most participants in our study were concerned about the impact of COVID-19 on their dental practice, as well as on their general health, safety, and well-being, with women being more concerned than men. Also, most respondents agreed that there were challenges to and long-term impacts on dentistry. More than one-half found working to be difficult during the pandemic, especially in the oldest groups. Owing to intervention to contain the spread of the virus via limiting clinical activities during the pandemic, the economy of the dental sector has been profoundly affected. In concordance with our findings, previous studies have found that COVID-19 has affected the economy of the dental practices. , A study evaluating Turkish dentists found they received COVID-19 information primarily from websites of official organizations and from social media. In concordance, our study found that most participants rated COVID-19 websites and internet search engines as their main sources of information to keep updated regarding COVID-19. Furthermore, a previous study at the School of Dentistry at The University of Texas Health San Antonio found that its COVID-19 website and virtual faculty-staff meetings were the main sources of information. Nearly all participants in our study felt prepared to resume dental practice and were confident of safety precautions and PPE their dental practice provided to protect them while performing aerosol-generating procedures. Only 21% of dentists felt COVID pandemic changed dental treatment protocols, with at least two-thirds agreeing that the precautions would influence their efficiency adversely. These findings are in agreement with those of a previous study conducted at the School of Dentistry at The University of Texas Health San Antonio regarding sources of information, feeling prepared for and confident about safety precautions while resuming dental practice, and precautions adversely influencing their efficiency; however, 60% of the participants in that study felt the pandemic changed dental treatment protocols compared with 21% reported in our study. In comparison, another study reported COVID-19 infection was considered to be highly dangerous by Italian dentists, who did not feel confident in safely resuming dental practice, felt uncertain about infection control measures and PPE, and were apprehensive about health and economic impacts. Other studies have reported concerns about availability of PPE and recent treatment protocol changes. , , , In comparison with the previously discussed the School of Dentistry at The University of Texas Health San Antonio study that evaluated some similar factors but was conducted early in the pandemic, our study, which evaluated a much broader dental community and was conducted during the COVID-19 peak period in Texas, found that the responses, overall, reflected fewer concerns, more satisfaction with resources provided for support, and much higher levels of preparedness to resume dental practice and confidence in safety precautions, with a significantly lower percentage of dentists feeling the pandemic changed their dental treatment protocols. Most participants rated PPE availability and having patients willing to come in for dental treatment as main challenges while resuming dental practice but reported they would not reconsider their dental career choices and would still recommend studying dentistry to others. The respondents in our study were mostly men compared with the dental school study in which most of the respondents were women; as reported in the literature, female sex is associated with higher levels of stress, anxiety, and depression and a greater psychological impact from COVID-19. , The findings of our study are consistent with the hypotheses related to this study that dental care professionals were confidently prepared to resume dental practice despite being concerned about the impact of COVID-19 on their dental practices and their own well-being, as well as challenges, long-term impacts, financial resources, and efficiency during the pandemic. Although this study was distributed during a peak period of COVID-19 in Texas, the second largest state in the United States with a diverse population and many dental care practitioners who went to dental school outside Texas, people who were more interested in participating may not reflect the rest of the dentist population. Future research with larger, nationwide samples should evaluate the long-term effects of the pandemic on dentistry and oral health of the population.
Almost all participants (98%) in this study felt prepared to resume dental practice and were confident of safety precautions their dental practices provided to protect them while performing dental procedures. Only one-fifth of dentists reported that they felt that the COVID-19 pandemic changed their dental treatment protocols, with more than two-thirds agreeing that the precautions would influence their efficiency adversely. Most respondents were satisfied with the resources their dental practices provided for support during the pandemic; were concerned about the impact of the pandemic to their dental practice and on their general health, safety and well-being; and agreed that there were challenges and long-term impacts on the dental profession. Most respondents rated COVID-19 websites and internet search engines as their main sources of information to keep updated. Although most rated PPE availability and patients willing to come in for treatment as top challenges while resuming dental practice, they would still recommend studying dentistry and would not reconsider their dental career choices.
|
Spinal cord injury as an indicator of abuse in forensic assessment of abusive head trauma (AHT) | 96ec3801-a85f-43f0-9d62-810900c046fa | 8205921 | Pathology[mh] | Abusive head trauma (AHT) in children including “Shaken Baby Syndrome” refers to intracranial traumatic brain injury in child victims of abuse. Shaking (with or without impact) has been identified as the leading mechanism resulting in the common AHT features, namely, the “triad” of intracranial subdural haematoma, cerebral oedema with hypoxic-ischaemic changes and retinal haemorrhages . The term shaken baby syndrome (SBS) has been used for decades, since Caffey introduced the term “whiplash shaken infant syndrome” in 1974 to describe a combination of intracranial and extracranial findings in abuse, in the absence of external signs of violence . AHT is a well-known cause of mortality, morbidity and disability, usually but not exclusively, in children under 1 year of age. Infants have high weight heads and weak neck muscles, and when a very young baby is shaken, the head moves repeatedly and excessively forwards and backwards causing excessive acceleration and deceleration of the brain. The precise mechanism and pathophysiology are usually challenging to determine, although many injuries are caused by violent hyperflexion and hyperextension, resulting in a condition similar to that now called shaken baby syndrome (SBS), but others are caused by blunt trauma (impact) as well as by the combination of the two mechanisms. The incidence of AHT has been reported to be as high as 15–29 per 100,000 infants but is probably underestimated . The pathological features in the brain have been described in many articles as a constellation of pathological features, centred on demonstration of the triad of subdual haematoma and hypoxic encephalopathy together with retinal haemorrhages . However, these have been determined to not be sufficient to confirm abuse by a systematic review from the Swedish Agency for Health and Technology Assessment . Several efforts to find additional pathological features associated with AHT have focused mainly on cervical spinal cord, due to the mechanism of shaking and the violent head and neck movements. Although pathological changes at this level have been reported, to date, they are not firmly embedded with the triad as solid diagnostic criteria. The present article is a review of the current scientific knowledge regarding spinal cord injuries in children suffering from AHT in order to better understand the role of spinal cord examination in forensic assessment of child abuse. Intracranial AHT pathological findings Since Guthkelck suggested the leading role of shaking in the development of brain subdural collections usually observed in battered babies in 1971, the detection of intracranial subdural haematoma (SDH) has become the cardinal point for the diagnosis of shaken baby syndrome . When Caffey (1974) coined the term “whiplash shaken infant syndrome” for the first time, the common clinical manifestation consisted of SDH in association with extracranial findings such as intraocular bleeding . The observation was subsequently confirmed by a wide range of studies on AHT children . The association of extra-axial bleeding with inflicted head injuries was statistically proven by Dashti et al. (1999) who studied 32 AHT children under the age of 2 years, in comparison with a group of 68 accidentally injured babies . In this study, the SDH appeared drastically more frequent in the AHT cohort (69% vs 7%, p < 0.001). The result was confirmed by Vinchon et al. (2010) who reported the percentage of children with SDH to be as high as 82.2% in those with AHT compared with only 43.6% in the control group ( p < 0.001) . Moreover, in a comprehensive systematic review by Kandom et al. (2014), the presence of subdural collections on neuroimaging was related to abuse by an odds ratio of 8.2 (95% CI) . Intracranial subdual haematoma is usually observed at imaging examination as a small amount of SDH with compressed cerebral sulci, displaced corticodural veins, sometimes in association with subdural membranes . Despite the initial description of SDH in AHT as chronic, the most common observation of SDH has been that of blood collection of recent onset , and a review of all the English-language literature on shaken baby syndrome over a 32-year period demonstrated the higher prevalence of acute blood collections rather than chronic . In agreement with the description that “interhemispheric haemorrhages and spinal SDH in multiple sites or of different densities were almost exclusively seen in AHT” , intracranial SDH collection is nowadays well known to have a different densities aspect, and the association of abusive head trauma with mixed intensities SDH on CT scan has been statistically proven ( p < 0.001) . Furthermore, paediatric patients with SDH of different intensities were found to be more likely to suffer from abusive head trauma (OR 6.39, 95% CI) . The distribution of SDH has been more substantially assessed, and it usually appears as a unilateral or bilateral thin film of subdural collection over the convexities, with a particular predisposition for the interhemispheric fissure . Adamsbaum et al. (2010) reported SDH incidence up to 95.5% in the interhemispheric fissure, 86% in the tentorium cerebelli and 100% in the right or left lateral spaces . The predominantly supratentorial locations have been proposed to be the hallmark of AHT and the main tool to distinguish AHT-associated haematomas from those related to birth . In a study where the intracranial imaging result was compared with the mechanism of trauma, 73% of interhemispheric SDH resulted from intentional injuries as well as 72% of SDH over the convexity . The significant association between the interhemispheric location of blood collections and AHT has been also statistically proven (OR 9.5, 95% CI) . Surprisingly, when Barlow et al. (1999) studied 12 children admitted to hospital with a diagnosis of abusive head injury on MRI, the most common site of SDH appeared to be the subtemporal area, but the result should be read with the limitation of the weakness of CT to investigate this area and thus the possibility of overlooking them in many cases . Subdural collections in the brain of AHT children has been confirmed at post mortem examination . The results from 53 AHT autopsied cases allowed Geddes et al. to recognize an age-related pattern of injuries . Infants less than 1 year old usually presented with a bilateral, thin film of SDH as opposed to older children who showed large and localized subdural haematomas. The older group presented frequently with axonal damage in the hemispheric matter and with extracranial injuries. The younger group, on the other hand, was more prone to the presence of axonal damage at the craniocervical junction and to skull fractures. Subdural collection usually occurs in association with other intracranial findings, such as subarachnoid haemorrhage and cerebral oedema. Cerebral oedema and hypoxic-ischaemic changes are features commonly associated with abuse . Wells et al. (2002) and Keenan et al. (2004) demonstrated that cerebral oedema is more common in abused babies than in those accidentally injured (78% vs 13% and 31% vs 13%) . The relationship between abuse and cerebral oedema has been statistically proven ( p < 0.002) (OR 2.2, 95% CI) as well as that between abuse and hypoxic-ischaemic injuries (OR 3.7, 95% CI) . Subarachnoid haemorrhage (SAH) is frequently reported in AHT children with a higher incidence rate in autopsied children (50% Geddes et al. 2001, 92% Brennan et al. 2009) than those investigated through imaging (18% Dashti et al. 1999 on CT and MRI) . When AHT children are compared with accidentally injured babies, SAH is seen more frequently in the accidental group (33% vs 61% respectively in Wells et al.’s study (2002) and 11.3% vs 22.7% in Keenan et al.’s study (2004) . Intraparenchymal changes are sometimes reported in association with the above-discussed intracranial injuries. Although Brennan et al. (2009) found a peculiarly high incidence of them between AHT babies (66% AHT had intracerebral bleeding, 65% had superficial cerebral contusion and lacerations, 65% had deep cerebral contusion and lacerations), intraparenchymal injuries are usually referred to as a sporadic finding . Skull fractures in young babies are seen more commonly in association with extradural haematoma (EDH) which is, in its turn, an uncommon finding in abused babies . Children suffering from abuse presented less frequently with skull fractures than those accidentally injured (57% vs 30.4%) as opposed to multiple skull fractures which are more frequently observed among those abused (14.2% vs 2.7%) . Spinal cord in AHT: evidence from neuroradiology Spinal blood collections Evidence of spinal cord involvement in suspect of abuse comes mostly from radiological investigation performed at the moment of hospital admission when brain and spinal cord are imaged through CT scan or MRI. In 1994, Diamond et al. published a case report of a 12-month-old female admitted to the hospital with T12 over L1 anterior spondylolisthesis . The MRI scan showed a T12-L3 pre-spinal mass possibly of haemorrhagic nature and tethered cord. A court confirmed the diagnosis of child abuse, but no further information on the mechanism of trauma was given. Three years later, Feldman et al. (1997) analysed 12 AHT children at cervical level in an attempt to assess the convenience of MRI in detecting the AHT cases (Table ). As opposed to the positive results from post mortem examinations in five deceased children which successfully managed to detect spinal blood collections (subdural haematoma on the upper cervical cord in one child along with subarachnoid collections between the remaining three children), the MRI failed to detect any signs of blood collections. The little sensibility of the MRI methodology at the time the study was performed could possibly explain the poor data. When Koumellis et al. (2009) analysed the spinal MRI findings between 18 AHT children (mean age 3 months) admitted to a tertiary neuroscience centre over a 7-year period in 2009, they showed completely different outcomes . The examination was performed on the whole spinal column, and almost half (44%) of the entire study cohort were positive for spinal subdural haematoma. All the six “large” collections spread out from the lower spinal canal point (the sacral thecal cul-de-sac) to variable upper levels. Only two were seen to reach the cervical spine, along with one of the so-called “small collections” who was detected exclusively at this level. It followed that the higher majority of blood subdural collection involved the thoracolumbar portion of the spine rather than the cervical region. Tearing and laceration of blood vessels located in the spinal canal travelling along with the spinal nerves and ventral and dorsal nerve roots were supposed to be the primary source of spinal blood collection. Gruber et al. described a case of a 4-month-old boy brought to a trauma centre in respiratory arrest after being repeatedly shaken in 2008 . After a T10-L1 subdural haematoma was seen on MRI, the source of the haemorrhage was intraoperatively identified within a lacerated radicular vein dorsal to the conus medullaris, which was coagulated and the bleeding stopped. The authors hypothesised that the connection between the thoracic and lumbar column is the pivot point in the shaking backward and forward body movements in the same manner of the cervical spine, and this can give a valuable explanation as to the thoracolumbar location of the haematoma. Along with MRI and CT scan, ultrasound (US) examination was used by Edelbauer et al. (2012) to investigate the spinal cord in six AHT infants (mean age 3.3 ± 1.5 months) . Spinal SDH was successfully seen, topographically extended from the cervical spine to the cauda equine as opposed to none of the 12 control. In 2012, Choudhary et al. focused on the incidence of spinal subdural haemorrhage on imaging examination (MRI and CT) between 67 AHT babies, in comparison to 70 cases of accidental head trauma . The cases were collected from an abusive head trauma registry, and no further information on the abuse assessment was given. In the AHT group spinal SDH accounted for about half of the cases (46%) as opposed to just above zero in those accidentally injured (1%). Cervical SDH was seen in 34% compared with subdural bleeding at thoracolumbar level in 63%. It was shown that abusive head trauma is statistically associated with subdural haematoma in spinal cord ( p < 0.001) and is more frequently seen at the thoracolumbar region rather than the cervical level. Furthermore, Choudhary et al.’s work from 2014 confirmed the high proportion of SDH in 67 AHT cases (48%) as opposed to just 2% in the accidentally injured group in a comprehensive study of 46 babies (mean age 4 months) . None of the 70 non-traumatic cases showed SDH. Many studies focused on the MRI examination of the cervical level only. When Kadom et al. (2014) assessed the cervical MRI results from 38 AHT cases, only one had blood collection at subdural levels . Similarly, Jacob et al. (2016) collected the cervical spinal cord MRI findings in 89 AHT infants (mean age 9.1 months) detecting an overall amount of SDH collections as low as 18% . Finally, 53 AHT cases studied by Baerg et al. (2017) were all negative for spinal blood collection on MRI examination . The study from Oh et al. (2017) analysed the overall results of imaging of the cervical spine in 503 abused children under the age of 9 years old. MRI was performed on 91 patients, and only two were positive for subdural blood collections . In the study from Henry et al. (2018) on 74 AHT and 14 accidental injury head trauma children under the age of 2, who underwent cervical MRI or CT for causes other than motor vehicle crash, spinal extra-axial haemorrhage was detected in up to 23% in those with AHT as opposed to only 1.3% in those accidentally injured . SDH in the spinal cord is commonly detected, when imaging investigation in AHT cases is performed on the whole length rather than part of the spinal cord. Agarwal et al. (2016) described a 6-month-old girl with intracranial bilateral SDH and retinal haemorrhages. MRI showed spinal haematoma extending from the thoracolumbar junction to the sacrum with a mass effect . In 2019 Hong et al. reported a case of a 5-month-old boy suspect for abuse with bilateral intracranial SDH, and a subdural haematoma from T4 to L5 was seen at MRI examination . As showed by Rabbitt et al.’s study from 2020 on 76 children who received spine MRI for identification of abuse, children with whole spine imaging were more likely to have spinal SDH ( p = 0.03) compared with those with spinal cervical assessment only . Unfortunately, 93% of abused babies were imaged at cervical and upper thoracic levels only. Although association between spine injury and abuse was not found, spinal subdural haemorrhage was the only finding associated with a combination of retinal haemorrhage ( p = 0.01), non-contact head injury ( p = 0.008) and a diagnosis of AHT ( p < 0.05). Finally, when intracranial haemorrhage was analysed, it was shown to not be statistically associated with spinal SDH ( p = 0.28). Spinal ligamentous injuries When spinal cord is studied through CT scan or MRI, one of the most recurrent features are changes in the soft-tissue apparatus, specifically in the ligamentous structure of the cervical column. Ghatan et al. reported a case of a 24-day-old female victim of AHT . MRI at cervical spinal cord showed ligamentous injury at occipitocervical junction, with atlantoaxial subluxation and narrowing of the spinal canal in 2002. Another case report from Bode et al. (2007) of a 8-month-old boy showed spinal ligamentous injuries at a lower level (disruption of the posterior ligament structure and cord contusion at T11–T12) . In both the reports, it is not specified how the abuse was assessed. Kemp et al. (2011) published a comprehensive systematic review of 19 previous studies, and a total of 25 children (between 1 and 48 months of age) with the aim of identifying the clinical and radiological spinal cord features of abuse and all the children with a highly assured diagnosis of AHT who underwent spinal radiological examination (RMI, CT and RX) were included . They found that the number with cervical lesions was as high as those with thoracolumbar lesions, accounting for 12 cases each. Both the cervical and the thoracolumbar injuries were mainly musculoskeletal, frequently in association with spinal cord involvement. In the cervical-lesion group 10 out of 12 had musculoskeletal injury, six of them with spinal cord compressions, transections, lacerations, stroke and parenchymal injury, while in the thoracolumbar-lesion group, the musculoskeletal injury accounted for 11/12, six of them with a spinal cord involvement (compression, contusion and tethering). Although the two groups had many aspects in common, those with lesions at the cervical level appeared younger as the majority of children were under 1 year of age as opposed to the thoracolumbar-injury group where the mean age was 13.5 months. In contrast with Feldman et al.’s statement that MRI should be performed only in the presence of spinal cord signs , the article from Kemp highlighted the mandatory role of MRI in order to prevent delayed recognition of spinal injuries. When Choudhary et al.’s study from 2014 compared 67 AHT babies (mean age 4 months) to 46 accidental-injury and 70 additional cases who underwent MRI for causes other than trauma (mean age 15 and 14 months, respectively), spinal ligamentous injury appeared related to the abuse mechanism of trauma . In those with AHT, ligamentous injuries accounted for 78% of the cases, compared with 46% in those accidental injuries and just over 0% in those with non-traumatic causes. In 2014, Kadom et al. published a study on 74 children, 38 of them with abusive, 26 with accidental head trauma and 10 so-called “undefined-head trauma” who underwent brain and cervical MRI (mean age 5.5, 0.6 and 22.6 months, respectively). The AHT cohort was assessed through modified Duhaime criteria, an algorithm including injury type, history and associated findings used to classify each injury as inflicted or accidental . Overall, 27/74 had cervical soft-tissue injuries, but data on single categories were not given. The author stated the absence of a significant relationship between cervical spinal injuries and abusive head trauma and therefore suggests that MRI lacks the ability to discriminate between accidental and abusive head traumas. However, the precise rate of AHT and accidentally injured children suffering from cervical injury was not given. In more recent years, Jacob (2016) published a retrospective review on 89 AHT children under the age of 5 years (mean age 9.1 months) to identify the features of cervical spine on MRI . Cervical spine injury was reported to be as high as 69%, mainly based on ligamentous alterations (67%) and vertebral joint swelling. Furthermore, Baerg et al. (2017) analysed MRI data from cervical spinal cord of 53 AHT children under the age of 36 months (mean age 5 months) . The percentage with cervical spine injury was reported to be 8/56 while ligamentous injuries were seen in 2/8 (25%). Finally, when Oh et al. (2017) studied the results form 91 abused patients (under the age of 9 years old) with cervical MRI, 13 (14%) were positive for ligamentous injuries , according to Henry et al.’s study from 2018, where ligamentous injuries were up to 9% in AHT and 6% in those accidentally injured when cervical spinal cord is imaged by MRI or CT . In conclusion, the incidence of spinal ligamentous injuries in AHT varies from 9 to 78%. As opposed to spinal subdural blood collections (mainly seen at thoracolumbar level), cervical level seems to be the ideal topographic location for detecting ligamentous injuries due to abuse trauma. However, the correlation between abuse and changes in ligamentous structures appeared to be not statistically proven, and even if soft-tissue lesions can strengthen the suspect of abuse, the finding alone is not sufficient to lead the diagnosis of abusive head trauma. Additional radiological findings According to the different modalities, abuse can happen as it is easy to suppose spinal structure is involved in many ways showing a wide range of additional features. Here below are reported those available from the current scientific knowledge i.e. cord parenchyma and spinal bone structure injuries. Vertebral fractures were found in 2/18 cases in Koumellis et al.’s study from 2009 at the level of the thoracic spine imaged by plan radiography . In Kemp et al.’s study (2011), those with cervical lesions (8/12) had spinal cord involvement (central cord injury, spinal cord compression and transection) , and just one case had vertebral arterial obstruction and stroke. In the group of musculoskeletal lesions (10/12), skeletal injuries varied between Hangman’s fracture at C2/C3, anterolisthesis, compression fracture of vertebral body and bilateral pedicle fractures. Between those with lesions at thoracolumbar level, 6/12 had spinal cord involvement with compression, contusion and tethering, and nine out of 12 had fracture dislocations, and three had compression of the vertebral body. Joint swelling in 32% of AHT cases was reported by Jacob et al. (2016) . They also highlighted that bone marrow oedema is usually seen in older children (mean age 14.9 months, p = 0.028) while capsular injury is commonly seen in younger children (mean age 5.5 months, p = 0.006). A spinal cord transection was detected by CT at T4 level in association with a distraction fracture of the spine on MRI in Brink et al.’s case (2017) of a 5-week-old boy ; the mother confessed to have grabbed him from the ankles and hit his back against a solid surface. Probably the most comprehensive study was the one from Jauregui et al. (2019) who retrospectively reviewed 22,192 children with spinal column fractures or spinal cord injuries . Patients were identified from Kids’ Inpatients Database (KID) using ICD-9-CM diagnosis (cervical, thoracic, lumbar vertebral fracture and spinal cord injury). One hundred and sixteen cases had a documented diagnosis of abuse and were shown to be at higher risk of thoracic (OR = 2.57) and lumbar (OR = 1.67) vertebral fractures as compared with non-abused patients. Additionally, abused patients were significantly less likely to be admitted with cervical column fractures than non-abused patients (OR = 0.51). Overall, no increased risk of spinal cord injury in abused compared with non-abused cases was seen. In conclusion, although the findings are commonly seen in abused babies, they appeared to be highly specific for the mechanism of trauma and so not indicative for abuse or accidental mechanism of trauma . Spinal cord AHT: evidence from neuropathology Spinal blood collections The first recorded study examining the neuropathology of spinal cord injury related to AHT was carried out in 1989 by Hadley et al. when they studied 13 infants (mean age 3 months) who died of confessed shaking without evidence of head impact trauma (Table ). Six of them underwent post mortem examination, and spinal cords were examined. All except one of the autopsied children showed injuries in the spinal cord, five had epidural haematoma and four had subdural haematoma at the cervicomedullary junction along with contusions of the ventral high cervical levels. All the autopsied children who had SDH presented with cerebral contusions, swelling and herniations. Hadley concluded that spinal injury at the high cervical cord level can contribute to the dramatic outcome of shaking without direct cranial impact. He also underlined the very young age of the studied babies, suggesting infants are more susceptible to injury from shaking. Eight years later, Feldman et al. (1997) focused on spinal cord injuries in five AHT autopsied children (mean age 5.8 months) enrolled through the child protection team . The diagnosis of inflicted head injury was then corroborated by the infant’s attending physician. In only one case, subdural blood collection was seen in the cervical spinal cord, while 3/5 showed subarachnoid bleeding. Both the subdural and the subarachnoid haemorrhages were seen in association with similar intracranial findings, and subdural haematoma was in clear continuity with the spinal one. In the three autopsied children reported by Saternus et al. (2000) (mean age 16 months), the AHT diagnoses were assumed from the history taken in the police notes in association with the intracranial subdural haematoma and the absence of head impact signs . One had epidural haemorrhage at the cervical level, and none of the cases had subdural blood collections in the spinal cord. The cervical spine showed intervertebral disc rupture and blood collections in the soft tissue in 2/3 patients. Similarly, when Geddes et al. (2001) performed a comprehensive retrospective study in order to identify the neuropathological changes in AHT children, the only spinal blood collection was the epidural haematoma seen in three cases . In an attempt to determine specifically the neuropathological findings in the cervical spinal cord of AHT, Brennan et al. (2009) reported the outcomes from 41 children who underwent cervical examination at post mortem and assessed to have died from AHT by the chief medical examiner . A very high proportion of them showed meningeal haemorrhages (83%) (epidural, intradural, subdural and/or subarachnoid, in an unspecified proportion) and parenchymal lesions such as contusions, lacerations and transections (72%). Additionally, nerve root avulsions and dorsal root ganglion were seen in slightly more than a half (55%). None of the children had vertebral fractures, and only one fifth (21%) had soft-tissue injuries in the neck. All the above-mentioned studies focused on cervical level spinal cord. Following the evidence from neuroradiological investigation of spinal cord, showing thoracolumbar level is the common location for spinal blood collection, is it possible that spinal cord injuries were overlooked and that the real incidence of them from neuropathology investigation is underestimated? For instance when homicide victims from physical abuse under the age of 3 years old were studied by Serenelli et al. in 2017, the majority of spinal cord lesions were at spinal level lower than cervical . In this cohort of 51 children (42 AHT), the most common finding was SDH across the spinal cord. Spinal cord injuries at a thoracolumbar location accounted for the majority of the cases (33.3%) as compared with the lumbosacral area (27.5%) and the cervical level (15.5%). Thoracic location appeared more frequent in infants, and the correlation was statistically proven ( p = 0.048). What is the source of spinal SDH? The anatomical structure of the spinal cord was proposed as the reason for the thoracolumbar distribution of subdural collections . The blood would flow from the posterior fossa into the spinal canal, collecting at the thoracolumbar region, where “a natural convexity in the supine position is present”. In our review, all the cases in the radiological and pathological investigations showed spinal SDH in association with intracranial SDH which was an inclusion criteria in only 5/9 and 1/9 articles (Tables and ) (Figs. , and ). On the other hand, as recently shown by Rabbitt et al. (2019), spinal SDH in children evaluated for abusive head trauma is not commonly associated with intracranial haemorrhage ( p = 0.28) . The small size of the intracranial bleeding and the lack of SDH in the posterior fossa compartment are not consistent with the intracranial source of spinal blood subdual collections. Given the close contact of the subdural and subarachnoid sheets is easy to suppose that a small quantity of blood leaking from upper level is not enough to separate the two sheets and to collect downwards at the thoracolumbar level. In his case report, Gruber et al. (2008) described the origin of bleeding as primary in the spinal cord, intraoperatively seen as a lacerated radicular vein . According to the finding, the author suggested the blood vessels travelling along with the spinal nerve roots are the source of the spinal subdural haemorrhages. Anatomically, the spinal cord is surrounded by the meninges, in the same way as the brain but with some differences. The dura mater is composed of only one sheet which is the direct continuation of the inner meningeal layer of the cranial dura when the outer layer of intracranial dura mater ceases at the foramen magnum. The spinal dura mater is firmly attached to the circumference of the foramen magnum, to the second and third cervical vertebrae and with the posterior longitudinal ligaments as well. The sheath of dura mater is much larger than is necessary for the accommodation of its contents, and its size is greater in the cervical and lumbar regions than in the thoracic. The epidural space contains a plexus of veins, while the subdural cavity is not actual, but it is a virtual space as the dura is in close contact with the arachnoid. Therefore, it is possible that a large intracranial subdural haematoma may have enough volume and weight to force itself through the virtual subdural space and present as a spinal cord subdural haematoma. But it is less likely that a small mainly intracranial SDH (commonly seen in AHT) can travel through different compartments to reach the spinal cord. Both the dura and the arachnoid surround the spinal nerves at the level of their entrance in the spinal cord. The pia and arachnoid membranes continue along with the spinal nerve roots as they leave the spinal cord and exit through the intervertebral foramina, where they blend with the perineurium of the spinal nerves . Arterial and venous vessels run along the surface of the spinal cord, between the arachnoid and pia mater. The latter is composed of collagen and reticular fibres which wraps the surface of spinal cord, while collagen fibres are external and form bundles with the above arachnoid; it is in this virtual space where vessels are found . One possible source of the spinal cord SDH is the radicular veins which run along nerve roots, and then the spinal cord surface needs to penetrate the arachnoid membrane in order to reach the subarachnoid location, resulting in an area of weakness when exposed to a high energy trauma such as that in shaking. The laceration of radicular veins at the point of passage from subdural to subarachnoid space could explain the blood collection in the spinal subdural space. The hypothesis of primary damage in the spinal nerve roots is supported by further evidence, and damage in the nerve roots especially the dorsal nerve roots has been detected . Brennan et al. reported the frequency of nerve roots avulsion and dorsal root ganglion haemorrhages as up to 55% between AHT children . Likewise, the βAPP immunohistochemistry examinations were positive in the nerve roots of abused babies as compared with the control group where no expression was detectable . It is well known that spinal nerve roots are the site of CSF absorption, and so they are surrounded by a high density vein vessel mesh and therefore are prone to bleeding . An association of high intracranial pressure and vessel damage due to hypoxic endothelial damage has been suggested as the cause of SDH. It is possible that the same mechanisms play a role in spinal cord bleeding . The primary spinal source of blood collection is also supported by the increasing evidence that the thoracolumbar level is frequently involved in cases of spinal trauma due to abuse, such as Jauregui et al.’s observation of increased risk of thoracic (OR = 2.75) and lumbar (OR = 1.67) vertebral fractures in his cohort of 116 abused children . In a recent study on 51 homicide victims, the frequency of thoracic and lumbar spinal injuries was reported around 30% as compared with just 15.5% at cervical level . Consequently, the side of maximum forces could be thoracolumbar rather than the cervical as the thoracic spine with the ribcage provides another valuable pivot point at the level of its articulation with the lumbar spine. Detecting the primary source of blood collection is of great interest to clarify where the trauma forces acted and therefore to better understand the trauma mechanism. Further neuropathological studies looking at the whole spinal cord and spinal nerve roots are needed to solve the issue. Spinal cord (parenchymal) injuries When Hadley et al. (1989) studied six AHT cases at post mortem, four had spinal contusions . Twenty years later, Brennan et al. (2009) confirmed the recurrent involvement of spinal cord injury in abuse following the observation of parenchymal lesions in up to 72% of his cohort of 41 abused children . Parenchymal lesions have been studied mainly through histology (Figs. and ). The first histological documentation of spinal injury in shaken babies was from Shannon et al. (1998) when fourteen cases (mean age 5 months) of witnessed, confessed or corroborated shaking without skull fractures underwent CD68 and βAPP immunohistochemistry . Cervical spinal cord showed βAPP-positive axons in 7 out of 11 cases, along with the spinal nerve roots (especially in the glial head). On the contrary, none of the control group cases (death from hypoxic-ischaemic encephalopathy and asphyxia) showed positivity for βAPP staining in the same structure, and the author hypothesised that βAPP positivity in the white matter tracts of cervical spinal cord may be associated with shaking. Shannon also studied the medulla, midbrain and cerebral white matters, where no differences were seen between the two groups. When Geddes et al. (2001) performed a comprehensive retrospective study in order to identify the neuropathological changes in AHT children, the findings were remarkable . The study cohort was comprised of 53 AHT children (37 infants and 16 toddlers) retrospectively collected and corroborated according to diagnostic criteria proposed by the author based on perpetrator’s confession/conviction and extra-cranial injuries. On histology, 8 out of the 53 showed localized βAPP-axonal positivity in the corticospinal tracts of the lower brainstem (lower pons and medulla) as well as in cervical cord roots in three additional cases. The findings were explained by the author as the result of stretching forces acting on the corticospinal tract in the lower brainstem which may lead to apnoea and hypoxic-ischaemic damage. In the same year, Geddes et al. published another article where the same results from the 37 infants were compared with a control group of 14 infants who died from causes other than traumatic . Three out of the 28 AHT cases immunologically stained were positive for βAPP in cervical cord and/or dorsal nerve roots, while eight showed βAPP positivity in the lower pons and medulla. None of the controls was positive for βAPP in the brain stem and spinal cord. Although the current knowledge on spinal cord changes in association to abuse need to be better investigated, is it possible for the time being to suppose that parenchymal injuries are not specific for the mechanism of trauma but possibly useful to understand the pathological mechanisms following abusive traumas? Effect of age on spinal cord injuries in AHT It is well known that a disproportionately larger head in children is supported by weaker and more lax cervical muscle and ligaments than in adults . What appears peculiar is the topical distribution of spinal injuries, which seems to be age-related. Cervical spinal injuries were seen more frequently in younger infants (age range 1–48 months, median age 5 months), while thoracolumbar was more frequent in older infants (age range 6–16 months, median age 13.5 months) . According to the current literature, young children tend to injure the upper cervical spine at the craniocervical junction to the C3 spinal level . As previously reported, infants have larger heads in proportion to their body and a more horizontal vertebral facet, allowing a higher degree of freedom in motion. When children grow, injuries are seen mainly in the lower level reaching the adult proportion at about 8–10 years. It has been proven that upper cervical spine injury (C1–C4), cervical fracture and spinal cord injury, spinal cord injury without radiographic abnormality (SCIWORA) and dislocation show a downward trend with increasing age . As such, the pathophysiology of trauma could also be different. If younger children have craniocervical junction injury, which leads to hypoxic-ischaemic brain injury, older babies can a have major incidence of extracranial injury points such as at thoracolumbar spinal cord levels. It follows that radiological examination of the spinal cord should be routinely performed at all levels, particularly when the suspected abuse involves children older than 1 year of age. Intracranial subdual haematoma is a common finding in neonates. When 101 term neonates were imaged by MRI within the first 72 h of life, the incidence of SDH was as high as 46%, the topical distribution involving both supra and infra tentorial compartments . It has been suggested that SDH in a child over the age of 1 month should not be attributed completely to birth trauma . Spinal cord haemorrhage in newborns was studied by Vlasyuk et al. in 2013 when he analysed 14 premature still-born babies with intraventricular haemorrhages. All cases with grade III were accompanied by drops of fluid in the subarachnoid space of the cervical dorsal and lumbar parts of the spinal cord . No other cases of spinal blood in newborns are reported in the literature. Similarly, intracranial haemorrhage can be seen in newborns lacking vitamin K in association with cerebral oedema and can mimic abuse. Brain infection and spontaneous intracranial haemorrhages linked with diastasis may also mimic AHT as oedema, and SDH can be present on neuroimaging in cases of encephalopathy. The clinical manifestation is also age-related. Children under 1 year of age present with impaired consciousness and respiratory distress while older patients have spinal deformity and focal neurological signs . Infants have thin bilateral SDH and higher incidence of skull fracture, while children older than 1 year of age had larger intracranial SDH collections and a higher incidence of extracranial damage, suggesting that the head and craniocervical junction are the main injury points in infants, but not in older children . It could be supposed that shaking acts on different levels depending on the age of the victim. Since Guthkelck suggested the leading role of shaking in the development of brain subdural collections usually observed in battered babies in 1971, the detection of intracranial subdural haematoma (SDH) has become the cardinal point for the diagnosis of shaken baby syndrome . When Caffey (1974) coined the term “whiplash shaken infant syndrome” for the first time, the common clinical manifestation consisted of SDH in association with extracranial findings such as intraocular bleeding . The observation was subsequently confirmed by a wide range of studies on AHT children . The association of extra-axial bleeding with inflicted head injuries was statistically proven by Dashti et al. (1999) who studied 32 AHT children under the age of 2 years, in comparison with a group of 68 accidentally injured babies . In this study, the SDH appeared drastically more frequent in the AHT cohort (69% vs 7%, p < 0.001). The result was confirmed by Vinchon et al. (2010) who reported the percentage of children with SDH to be as high as 82.2% in those with AHT compared with only 43.6% in the control group ( p < 0.001) . Moreover, in a comprehensive systematic review by Kandom et al. (2014), the presence of subdural collections on neuroimaging was related to abuse by an odds ratio of 8.2 (95% CI) . Intracranial subdual haematoma is usually observed at imaging examination as a small amount of SDH with compressed cerebral sulci, displaced corticodural veins, sometimes in association with subdural membranes . Despite the initial description of SDH in AHT as chronic, the most common observation of SDH has been that of blood collection of recent onset , and a review of all the English-language literature on shaken baby syndrome over a 32-year period demonstrated the higher prevalence of acute blood collections rather than chronic . In agreement with the description that “interhemispheric haemorrhages and spinal SDH in multiple sites or of different densities were almost exclusively seen in AHT” , intracranial SDH collection is nowadays well known to have a different densities aspect, and the association of abusive head trauma with mixed intensities SDH on CT scan has been statistically proven ( p < 0.001) . Furthermore, paediatric patients with SDH of different intensities were found to be more likely to suffer from abusive head trauma (OR 6.39, 95% CI) . The distribution of SDH has been more substantially assessed, and it usually appears as a unilateral or bilateral thin film of subdural collection over the convexities, with a particular predisposition for the interhemispheric fissure . Adamsbaum et al. (2010) reported SDH incidence up to 95.5% in the interhemispheric fissure, 86% in the tentorium cerebelli and 100% in the right or left lateral spaces . The predominantly supratentorial locations have been proposed to be the hallmark of AHT and the main tool to distinguish AHT-associated haematomas from those related to birth . In a study where the intracranial imaging result was compared with the mechanism of trauma, 73% of interhemispheric SDH resulted from intentional injuries as well as 72% of SDH over the convexity . The significant association between the interhemispheric location of blood collections and AHT has been also statistically proven (OR 9.5, 95% CI) . Surprisingly, when Barlow et al. (1999) studied 12 children admitted to hospital with a diagnosis of abusive head injury on MRI, the most common site of SDH appeared to be the subtemporal area, but the result should be read with the limitation of the weakness of CT to investigate this area and thus the possibility of overlooking them in many cases . Subdural collections in the brain of AHT children has been confirmed at post mortem examination . The results from 53 AHT autopsied cases allowed Geddes et al. to recognize an age-related pattern of injuries . Infants less than 1 year old usually presented with a bilateral, thin film of SDH as opposed to older children who showed large and localized subdural haematomas. The older group presented frequently with axonal damage in the hemispheric matter and with extracranial injuries. The younger group, on the other hand, was more prone to the presence of axonal damage at the craniocervical junction and to skull fractures. Subdural collection usually occurs in association with other intracranial findings, such as subarachnoid haemorrhage and cerebral oedema. Cerebral oedema and hypoxic-ischaemic changes are features commonly associated with abuse . Wells et al. (2002) and Keenan et al. (2004) demonstrated that cerebral oedema is more common in abused babies than in those accidentally injured (78% vs 13% and 31% vs 13%) . The relationship between abuse and cerebral oedema has been statistically proven ( p < 0.002) (OR 2.2, 95% CI) as well as that between abuse and hypoxic-ischaemic injuries (OR 3.7, 95% CI) . Subarachnoid haemorrhage (SAH) is frequently reported in AHT children with a higher incidence rate in autopsied children (50% Geddes et al. 2001, 92% Brennan et al. 2009) than those investigated through imaging (18% Dashti et al. 1999 on CT and MRI) . When AHT children are compared with accidentally injured babies, SAH is seen more frequently in the accidental group (33% vs 61% respectively in Wells et al.’s study (2002) and 11.3% vs 22.7% in Keenan et al.’s study (2004) . Intraparenchymal changes are sometimes reported in association with the above-discussed intracranial injuries. Although Brennan et al. (2009) found a peculiarly high incidence of them between AHT babies (66% AHT had intracerebral bleeding, 65% had superficial cerebral contusion and lacerations, 65% had deep cerebral contusion and lacerations), intraparenchymal injuries are usually referred to as a sporadic finding . Skull fractures in young babies are seen more commonly in association with extradural haematoma (EDH) which is, in its turn, an uncommon finding in abused babies . Children suffering from abuse presented less frequently with skull fractures than those accidentally injured (57% vs 30.4%) as opposed to multiple skull fractures which are more frequently observed among those abused (14.2% vs 2.7%) . Spinal blood collections Evidence of spinal cord involvement in suspect of abuse comes mostly from radiological investigation performed at the moment of hospital admission when brain and spinal cord are imaged through CT scan or MRI. In 1994, Diamond et al. published a case report of a 12-month-old female admitted to the hospital with T12 over L1 anterior spondylolisthesis . The MRI scan showed a T12-L3 pre-spinal mass possibly of haemorrhagic nature and tethered cord. A court confirmed the diagnosis of child abuse, but no further information on the mechanism of trauma was given. Three years later, Feldman et al. (1997) analysed 12 AHT children at cervical level in an attempt to assess the convenience of MRI in detecting the AHT cases (Table ). As opposed to the positive results from post mortem examinations in five deceased children which successfully managed to detect spinal blood collections (subdural haematoma on the upper cervical cord in one child along with subarachnoid collections between the remaining three children), the MRI failed to detect any signs of blood collections. The little sensibility of the MRI methodology at the time the study was performed could possibly explain the poor data. When Koumellis et al. (2009) analysed the spinal MRI findings between 18 AHT children (mean age 3 months) admitted to a tertiary neuroscience centre over a 7-year period in 2009, they showed completely different outcomes . The examination was performed on the whole spinal column, and almost half (44%) of the entire study cohort were positive for spinal subdural haematoma. All the six “large” collections spread out from the lower spinal canal point (the sacral thecal cul-de-sac) to variable upper levels. Only two were seen to reach the cervical spine, along with one of the so-called “small collections” who was detected exclusively at this level. It followed that the higher majority of blood subdural collection involved the thoracolumbar portion of the spine rather than the cervical region. Tearing and laceration of blood vessels located in the spinal canal travelling along with the spinal nerves and ventral and dorsal nerve roots were supposed to be the primary source of spinal blood collection. Gruber et al. described a case of a 4-month-old boy brought to a trauma centre in respiratory arrest after being repeatedly shaken in 2008 . After a T10-L1 subdural haematoma was seen on MRI, the source of the haemorrhage was intraoperatively identified within a lacerated radicular vein dorsal to the conus medullaris, which was coagulated and the bleeding stopped. The authors hypothesised that the connection between the thoracic and lumbar column is the pivot point in the shaking backward and forward body movements in the same manner of the cervical spine, and this can give a valuable explanation as to the thoracolumbar location of the haematoma. Along with MRI and CT scan, ultrasound (US) examination was used by Edelbauer et al. (2012) to investigate the spinal cord in six AHT infants (mean age 3.3 ± 1.5 months) . Spinal SDH was successfully seen, topographically extended from the cervical spine to the cauda equine as opposed to none of the 12 control. In 2012, Choudhary et al. focused on the incidence of spinal subdural haemorrhage on imaging examination (MRI and CT) between 67 AHT babies, in comparison to 70 cases of accidental head trauma . The cases were collected from an abusive head trauma registry, and no further information on the abuse assessment was given. In the AHT group spinal SDH accounted for about half of the cases (46%) as opposed to just above zero in those accidentally injured (1%). Cervical SDH was seen in 34% compared with subdural bleeding at thoracolumbar level in 63%. It was shown that abusive head trauma is statistically associated with subdural haematoma in spinal cord ( p < 0.001) and is more frequently seen at the thoracolumbar region rather than the cervical level. Furthermore, Choudhary et al.’s work from 2014 confirmed the high proportion of SDH in 67 AHT cases (48%) as opposed to just 2% in the accidentally injured group in a comprehensive study of 46 babies (mean age 4 months) . None of the 70 non-traumatic cases showed SDH. Many studies focused on the MRI examination of the cervical level only. When Kadom et al. (2014) assessed the cervical MRI results from 38 AHT cases, only one had blood collection at subdural levels . Similarly, Jacob et al. (2016) collected the cervical spinal cord MRI findings in 89 AHT infants (mean age 9.1 months) detecting an overall amount of SDH collections as low as 18% . Finally, 53 AHT cases studied by Baerg et al. (2017) were all negative for spinal blood collection on MRI examination . The study from Oh et al. (2017) analysed the overall results of imaging of the cervical spine in 503 abused children under the age of 9 years old. MRI was performed on 91 patients, and only two were positive for subdural blood collections . In the study from Henry et al. (2018) on 74 AHT and 14 accidental injury head trauma children under the age of 2, who underwent cervical MRI or CT for causes other than motor vehicle crash, spinal extra-axial haemorrhage was detected in up to 23% in those with AHT as opposed to only 1.3% in those accidentally injured . SDH in the spinal cord is commonly detected, when imaging investigation in AHT cases is performed on the whole length rather than part of the spinal cord. Agarwal et al. (2016) described a 6-month-old girl with intracranial bilateral SDH and retinal haemorrhages. MRI showed spinal haematoma extending from the thoracolumbar junction to the sacrum with a mass effect . In 2019 Hong et al. reported a case of a 5-month-old boy suspect for abuse with bilateral intracranial SDH, and a subdural haematoma from T4 to L5 was seen at MRI examination . As showed by Rabbitt et al.’s study from 2020 on 76 children who received spine MRI for identification of abuse, children with whole spine imaging were more likely to have spinal SDH ( p = 0.03) compared with those with spinal cervical assessment only . Unfortunately, 93% of abused babies were imaged at cervical and upper thoracic levels only. Although association between spine injury and abuse was not found, spinal subdural haemorrhage was the only finding associated with a combination of retinal haemorrhage ( p = 0.01), non-contact head injury ( p = 0.008) and a diagnosis of AHT ( p < 0.05). Finally, when intracranial haemorrhage was analysed, it was shown to not be statistically associated with spinal SDH ( p = 0.28). Spinal ligamentous injuries When spinal cord is studied through CT scan or MRI, one of the most recurrent features are changes in the soft-tissue apparatus, specifically in the ligamentous structure of the cervical column. Ghatan et al. reported a case of a 24-day-old female victim of AHT . MRI at cervical spinal cord showed ligamentous injury at occipitocervical junction, with atlantoaxial subluxation and narrowing of the spinal canal in 2002. Another case report from Bode et al. (2007) of a 8-month-old boy showed spinal ligamentous injuries at a lower level (disruption of the posterior ligament structure and cord contusion at T11–T12) . In both the reports, it is not specified how the abuse was assessed. Kemp et al. (2011) published a comprehensive systematic review of 19 previous studies, and a total of 25 children (between 1 and 48 months of age) with the aim of identifying the clinical and radiological spinal cord features of abuse and all the children with a highly assured diagnosis of AHT who underwent spinal radiological examination (RMI, CT and RX) were included . They found that the number with cervical lesions was as high as those with thoracolumbar lesions, accounting for 12 cases each. Both the cervical and the thoracolumbar injuries were mainly musculoskeletal, frequently in association with spinal cord involvement. In the cervical-lesion group 10 out of 12 had musculoskeletal injury, six of them with spinal cord compressions, transections, lacerations, stroke and parenchymal injury, while in the thoracolumbar-lesion group, the musculoskeletal injury accounted for 11/12, six of them with a spinal cord involvement (compression, contusion and tethering). Although the two groups had many aspects in common, those with lesions at the cervical level appeared younger as the majority of children were under 1 year of age as opposed to the thoracolumbar-injury group where the mean age was 13.5 months. In contrast with Feldman et al.’s statement that MRI should be performed only in the presence of spinal cord signs , the article from Kemp highlighted the mandatory role of MRI in order to prevent delayed recognition of spinal injuries. When Choudhary et al.’s study from 2014 compared 67 AHT babies (mean age 4 months) to 46 accidental-injury and 70 additional cases who underwent MRI for causes other than trauma (mean age 15 and 14 months, respectively), spinal ligamentous injury appeared related to the abuse mechanism of trauma . In those with AHT, ligamentous injuries accounted for 78% of the cases, compared with 46% in those accidental injuries and just over 0% in those with non-traumatic causes. In 2014, Kadom et al. published a study on 74 children, 38 of them with abusive, 26 with accidental head trauma and 10 so-called “undefined-head trauma” who underwent brain and cervical MRI (mean age 5.5, 0.6 and 22.6 months, respectively). The AHT cohort was assessed through modified Duhaime criteria, an algorithm including injury type, history and associated findings used to classify each injury as inflicted or accidental . Overall, 27/74 had cervical soft-tissue injuries, but data on single categories were not given. The author stated the absence of a significant relationship between cervical spinal injuries and abusive head trauma and therefore suggests that MRI lacks the ability to discriminate between accidental and abusive head traumas. However, the precise rate of AHT and accidentally injured children suffering from cervical injury was not given. In more recent years, Jacob (2016) published a retrospective review on 89 AHT children under the age of 5 years (mean age 9.1 months) to identify the features of cervical spine on MRI . Cervical spine injury was reported to be as high as 69%, mainly based on ligamentous alterations (67%) and vertebral joint swelling. Furthermore, Baerg et al. (2017) analysed MRI data from cervical spinal cord of 53 AHT children under the age of 36 months (mean age 5 months) . The percentage with cervical spine injury was reported to be 8/56 while ligamentous injuries were seen in 2/8 (25%). Finally, when Oh et al. (2017) studied the results form 91 abused patients (under the age of 9 years old) with cervical MRI, 13 (14%) were positive for ligamentous injuries , according to Henry et al.’s study from 2018, where ligamentous injuries were up to 9% in AHT and 6% in those accidentally injured when cervical spinal cord is imaged by MRI or CT . In conclusion, the incidence of spinal ligamentous injuries in AHT varies from 9 to 78%. As opposed to spinal subdural blood collections (mainly seen at thoracolumbar level), cervical level seems to be the ideal topographic location for detecting ligamentous injuries due to abuse trauma. However, the correlation between abuse and changes in ligamentous structures appeared to be not statistically proven, and even if soft-tissue lesions can strengthen the suspect of abuse, the finding alone is not sufficient to lead the diagnosis of abusive head trauma. Additional radiological findings According to the different modalities, abuse can happen as it is easy to suppose spinal structure is involved in many ways showing a wide range of additional features. Here below are reported those available from the current scientific knowledge i.e. cord parenchyma and spinal bone structure injuries. Vertebral fractures were found in 2/18 cases in Koumellis et al.’s study from 2009 at the level of the thoracic spine imaged by plan radiography . In Kemp et al.’s study (2011), those with cervical lesions (8/12) had spinal cord involvement (central cord injury, spinal cord compression and transection) , and just one case had vertebral arterial obstruction and stroke. In the group of musculoskeletal lesions (10/12), skeletal injuries varied between Hangman’s fracture at C2/C3, anterolisthesis, compression fracture of vertebral body and bilateral pedicle fractures. Between those with lesions at thoracolumbar level, 6/12 had spinal cord involvement with compression, contusion and tethering, and nine out of 12 had fracture dislocations, and three had compression of the vertebral body. Joint swelling in 32% of AHT cases was reported by Jacob et al. (2016) . They also highlighted that bone marrow oedema is usually seen in older children (mean age 14.9 months, p = 0.028) while capsular injury is commonly seen in younger children (mean age 5.5 months, p = 0.006). A spinal cord transection was detected by CT at T4 level in association with a distraction fracture of the spine on MRI in Brink et al.’s case (2017) of a 5-week-old boy ; the mother confessed to have grabbed him from the ankles and hit his back against a solid surface. Probably the most comprehensive study was the one from Jauregui et al. (2019) who retrospectively reviewed 22,192 children with spinal column fractures or spinal cord injuries . Patients were identified from Kids’ Inpatients Database (KID) using ICD-9-CM diagnosis (cervical, thoracic, lumbar vertebral fracture and spinal cord injury). One hundred and sixteen cases had a documented diagnosis of abuse and were shown to be at higher risk of thoracic (OR = 2.57) and lumbar (OR = 1.67) vertebral fractures as compared with non-abused patients. Additionally, abused patients were significantly less likely to be admitted with cervical column fractures than non-abused patients (OR = 0.51). Overall, no increased risk of spinal cord injury in abused compared with non-abused cases was seen. In conclusion, although the findings are commonly seen in abused babies, they appeared to be highly specific for the mechanism of trauma and so not indicative for abuse or accidental mechanism of trauma . Evidence of spinal cord involvement in suspect of abuse comes mostly from radiological investigation performed at the moment of hospital admission when brain and spinal cord are imaged through CT scan or MRI. In 1994, Diamond et al. published a case report of a 12-month-old female admitted to the hospital with T12 over L1 anterior spondylolisthesis . The MRI scan showed a T12-L3 pre-spinal mass possibly of haemorrhagic nature and tethered cord. A court confirmed the diagnosis of child abuse, but no further information on the mechanism of trauma was given. Three years later, Feldman et al. (1997) analysed 12 AHT children at cervical level in an attempt to assess the convenience of MRI in detecting the AHT cases (Table ). As opposed to the positive results from post mortem examinations in five deceased children which successfully managed to detect spinal blood collections (subdural haematoma on the upper cervical cord in one child along with subarachnoid collections between the remaining three children), the MRI failed to detect any signs of blood collections. The little sensibility of the MRI methodology at the time the study was performed could possibly explain the poor data. When Koumellis et al. (2009) analysed the spinal MRI findings between 18 AHT children (mean age 3 months) admitted to a tertiary neuroscience centre over a 7-year period in 2009, they showed completely different outcomes . The examination was performed on the whole spinal column, and almost half (44%) of the entire study cohort were positive for spinal subdural haematoma. All the six “large” collections spread out from the lower spinal canal point (the sacral thecal cul-de-sac) to variable upper levels. Only two were seen to reach the cervical spine, along with one of the so-called “small collections” who was detected exclusively at this level. It followed that the higher majority of blood subdural collection involved the thoracolumbar portion of the spine rather than the cervical region. Tearing and laceration of blood vessels located in the spinal canal travelling along with the spinal nerves and ventral and dorsal nerve roots were supposed to be the primary source of spinal blood collection. Gruber et al. described a case of a 4-month-old boy brought to a trauma centre in respiratory arrest after being repeatedly shaken in 2008 . After a T10-L1 subdural haematoma was seen on MRI, the source of the haemorrhage was intraoperatively identified within a lacerated radicular vein dorsal to the conus medullaris, which was coagulated and the bleeding stopped. The authors hypothesised that the connection between the thoracic and lumbar column is the pivot point in the shaking backward and forward body movements in the same manner of the cervical spine, and this can give a valuable explanation as to the thoracolumbar location of the haematoma. Along with MRI and CT scan, ultrasound (US) examination was used by Edelbauer et al. (2012) to investigate the spinal cord in six AHT infants (mean age 3.3 ± 1.5 months) . Spinal SDH was successfully seen, topographically extended from the cervical spine to the cauda equine as opposed to none of the 12 control. In 2012, Choudhary et al. focused on the incidence of spinal subdural haemorrhage on imaging examination (MRI and CT) between 67 AHT babies, in comparison to 70 cases of accidental head trauma . The cases were collected from an abusive head trauma registry, and no further information on the abuse assessment was given. In the AHT group spinal SDH accounted for about half of the cases (46%) as opposed to just above zero in those accidentally injured (1%). Cervical SDH was seen in 34% compared with subdural bleeding at thoracolumbar level in 63%. It was shown that abusive head trauma is statistically associated with subdural haematoma in spinal cord ( p < 0.001) and is more frequently seen at the thoracolumbar region rather than the cervical level. Furthermore, Choudhary et al.’s work from 2014 confirmed the high proportion of SDH in 67 AHT cases (48%) as opposed to just 2% in the accidentally injured group in a comprehensive study of 46 babies (mean age 4 months) . None of the 70 non-traumatic cases showed SDH. Many studies focused on the MRI examination of the cervical level only. When Kadom et al. (2014) assessed the cervical MRI results from 38 AHT cases, only one had blood collection at subdural levels . Similarly, Jacob et al. (2016) collected the cervical spinal cord MRI findings in 89 AHT infants (mean age 9.1 months) detecting an overall amount of SDH collections as low as 18% . Finally, 53 AHT cases studied by Baerg et al. (2017) were all negative for spinal blood collection on MRI examination . The study from Oh et al. (2017) analysed the overall results of imaging of the cervical spine in 503 abused children under the age of 9 years old. MRI was performed on 91 patients, and only two were positive for subdural blood collections . In the study from Henry et al. (2018) on 74 AHT and 14 accidental injury head trauma children under the age of 2, who underwent cervical MRI or CT for causes other than motor vehicle crash, spinal extra-axial haemorrhage was detected in up to 23% in those with AHT as opposed to only 1.3% in those accidentally injured . SDH in the spinal cord is commonly detected, when imaging investigation in AHT cases is performed on the whole length rather than part of the spinal cord. Agarwal et al. (2016) described a 6-month-old girl with intracranial bilateral SDH and retinal haemorrhages. MRI showed spinal haematoma extending from the thoracolumbar junction to the sacrum with a mass effect . In 2019 Hong et al. reported a case of a 5-month-old boy suspect for abuse with bilateral intracranial SDH, and a subdural haematoma from T4 to L5 was seen at MRI examination . As showed by Rabbitt et al.’s study from 2020 on 76 children who received spine MRI for identification of abuse, children with whole spine imaging were more likely to have spinal SDH ( p = 0.03) compared with those with spinal cervical assessment only . Unfortunately, 93% of abused babies were imaged at cervical and upper thoracic levels only. Although association between spine injury and abuse was not found, spinal subdural haemorrhage was the only finding associated with a combination of retinal haemorrhage ( p = 0.01), non-contact head injury ( p = 0.008) and a diagnosis of AHT ( p < 0.05). Finally, when intracranial haemorrhage was analysed, it was shown to not be statistically associated with spinal SDH ( p = 0.28). When spinal cord is studied through CT scan or MRI, one of the most recurrent features are changes in the soft-tissue apparatus, specifically in the ligamentous structure of the cervical column. Ghatan et al. reported a case of a 24-day-old female victim of AHT . MRI at cervical spinal cord showed ligamentous injury at occipitocervical junction, with atlantoaxial subluxation and narrowing of the spinal canal in 2002. Another case report from Bode et al. (2007) of a 8-month-old boy showed spinal ligamentous injuries at a lower level (disruption of the posterior ligament structure and cord contusion at T11–T12) . In both the reports, it is not specified how the abuse was assessed. Kemp et al. (2011) published a comprehensive systematic review of 19 previous studies, and a total of 25 children (between 1 and 48 months of age) with the aim of identifying the clinical and radiological spinal cord features of abuse and all the children with a highly assured diagnosis of AHT who underwent spinal radiological examination (RMI, CT and RX) were included . They found that the number with cervical lesions was as high as those with thoracolumbar lesions, accounting for 12 cases each. Both the cervical and the thoracolumbar injuries were mainly musculoskeletal, frequently in association with spinal cord involvement. In the cervical-lesion group 10 out of 12 had musculoskeletal injury, six of them with spinal cord compressions, transections, lacerations, stroke and parenchymal injury, while in the thoracolumbar-lesion group, the musculoskeletal injury accounted for 11/12, six of them with a spinal cord involvement (compression, contusion and tethering). Although the two groups had many aspects in common, those with lesions at the cervical level appeared younger as the majority of children were under 1 year of age as opposed to the thoracolumbar-injury group where the mean age was 13.5 months. In contrast with Feldman et al.’s statement that MRI should be performed only in the presence of spinal cord signs , the article from Kemp highlighted the mandatory role of MRI in order to prevent delayed recognition of spinal injuries. When Choudhary et al.’s study from 2014 compared 67 AHT babies (mean age 4 months) to 46 accidental-injury and 70 additional cases who underwent MRI for causes other than trauma (mean age 15 and 14 months, respectively), spinal ligamentous injury appeared related to the abuse mechanism of trauma . In those with AHT, ligamentous injuries accounted for 78% of the cases, compared with 46% in those accidental injuries and just over 0% in those with non-traumatic causes. In 2014, Kadom et al. published a study on 74 children, 38 of them with abusive, 26 with accidental head trauma and 10 so-called “undefined-head trauma” who underwent brain and cervical MRI (mean age 5.5, 0.6 and 22.6 months, respectively). The AHT cohort was assessed through modified Duhaime criteria, an algorithm including injury type, history and associated findings used to classify each injury as inflicted or accidental . Overall, 27/74 had cervical soft-tissue injuries, but data on single categories were not given. The author stated the absence of a significant relationship between cervical spinal injuries and abusive head trauma and therefore suggests that MRI lacks the ability to discriminate between accidental and abusive head traumas. However, the precise rate of AHT and accidentally injured children suffering from cervical injury was not given. In more recent years, Jacob (2016) published a retrospective review on 89 AHT children under the age of 5 years (mean age 9.1 months) to identify the features of cervical spine on MRI . Cervical spine injury was reported to be as high as 69%, mainly based on ligamentous alterations (67%) and vertebral joint swelling. Furthermore, Baerg et al. (2017) analysed MRI data from cervical spinal cord of 53 AHT children under the age of 36 months (mean age 5 months) . The percentage with cervical spine injury was reported to be 8/56 while ligamentous injuries were seen in 2/8 (25%). Finally, when Oh et al. (2017) studied the results form 91 abused patients (under the age of 9 years old) with cervical MRI, 13 (14%) were positive for ligamentous injuries , according to Henry et al.’s study from 2018, where ligamentous injuries were up to 9% in AHT and 6% in those accidentally injured when cervical spinal cord is imaged by MRI or CT . In conclusion, the incidence of spinal ligamentous injuries in AHT varies from 9 to 78%. As opposed to spinal subdural blood collections (mainly seen at thoracolumbar level), cervical level seems to be the ideal topographic location for detecting ligamentous injuries due to abuse trauma. However, the correlation between abuse and changes in ligamentous structures appeared to be not statistically proven, and even if soft-tissue lesions can strengthen the suspect of abuse, the finding alone is not sufficient to lead the diagnosis of abusive head trauma. According to the different modalities, abuse can happen as it is easy to suppose spinal structure is involved in many ways showing a wide range of additional features. Here below are reported those available from the current scientific knowledge i.e. cord parenchyma and spinal bone structure injuries. Vertebral fractures were found in 2/18 cases in Koumellis et al.’s study from 2009 at the level of the thoracic spine imaged by plan radiography . In Kemp et al.’s study (2011), those with cervical lesions (8/12) had spinal cord involvement (central cord injury, spinal cord compression and transection) , and just one case had vertebral arterial obstruction and stroke. In the group of musculoskeletal lesions (10/12), skeletal injuries varied between Hangman’s fracture at C2/C3, anterolisthesis, compression fracture of vertebral body and bilateral pedicle fractures. Between those with lesions at thoracolumbar level, 6/12 had spinal cord involvement with compression, contusion and tethering, and nine out of 12 had fracture dislocations, and three had compression of the vertebral body. Joint swelling in 32% of AHT cases was reported by Jacob et al. (2016) . They also highlighted that bone marrow oedema is usually seen in older children (mean age 14.9 months, p = 0.028) while capsular injury is commonly seen in younger children (mean age 5.5 months, p = 0.006). A spinal cord transection was detected by CT at T4 level in association with a distraction fracture of the spine on MRI in Brink et al.’s case (2017) of a 5-week-old boy ; the mother confessed to have grabbed him from the ankles and hit his back against a solid surface. Probably the most comprehensive study was the one from Jauregui et al. (2019) who retrospectively reviewed 22,192 children with spinal column fractures or spinal cord injuries . Patients were identified from Kids’ Inpatients Database (KID) using ICD-9-CM diagnosis (cervical, thoracic, lumbar vertebral fracture and spinal cord injury). One hundred and sixteen cases had a documented diagnosis of abuse and were shown to be at higher risk of thoracic (OR = 2.57) and lumbar (OR = 1.67) vertebral fractures as compared with non-abused patients. Additionally, abused patients were significantly less likely to be admitted with cervical column fractures than non-abused patients (OR = 0.51). Overall, no increased risk of spinal cord injury in abused compared with non-abused cases was seen. In conclusion, although the findings are commonly seen in abused babies, they appeared to be highly specific for the mechanism of trauma and so not indicative for abuse or accidental mechanism of trauma . Spinal blood collections The first recorded study examining the neuropathology of spinal cord injury related to AHT was carried out in 1989 by Hadley et al. when they studied 13 infants (mean age 3 months) who died of confessed shaking without evidence of head impact trauma (Table ). Six of them underwent post mortem examination, and spinal cords were examined. All except one of the autopsied children showed injuries in the spinal cord, five had epidural haematoma and four had subdural haematoma at the cervicomedullary junction along with contusions of the ventral high cervical levels. All the autopsied children who had SDH presented with cerebral contusions, swelling and herniations. Hadley concluded that spinal injury at the high cervical cord level can contribute to the dramatic outcome of shaking without direct cranial impact. He also underlined the very young age of the studied babies, suggesting infants are more susceptible to injury from shaking. Eight years later, Feldman et al. (1997) focused on spinal cord injuries in five AHT autopsied children (mean age 5.8 months) enrolled through the child protection team . The diagnosis of inflicted head injury was then corroborated by the infant’s attending physician. In only one case, subdural blood collection was seen in the cervical spinal cord, while 3/5 showed subarachnoid bleeding. Both the subdural and the subarachnoid haemorrhages were seen in association with similar intracranial findings, and subdural haematoma was in clear continuity with the spinal one. In the three autopsied children reported by Saternus et al. (2000) (mean age 16 months), the AHT diagnoses were assumed from the history taken in the police notes in association with the intracranial subdural haematoma and the absence of head impact signs . One had epidural haemorrhage at the cervical level, and none of the cases had subdural blood collections in the spinal cord. The cervical spine showed intervertebral disc rupture and blood collections in the soft tissue in 2/3 patients. Similarly, when Geddes et al. (2001) performed a comprehensive retrospective study in order to identify the neuropathological changes in AHT children, the only spinal blood collection was the epidural haematoma seen in three cases . In an attempt to determine specifically the neuropathological findings in the cervical spinal cord of AHT, Brennan et al. (2009) reported the outcomes from 41 children who underwent cervical examination at post mortem and assessed to have died from AHT by the chief medical examiner . A very high proportion of them showed meningeal haemorrhages (83%) (epidural, intradural, subdural and/or subarachnoid, in an unspecified proportion) and parenchymal lesions such as contusions, lacerations and transections (72%). Additionally, nerve root avulsions and dorsal root ganglion were seen in slightly more than a half (55%). None of the children had vertebral fractures, and only one fifth (21%) had soft-tissue injuries in the neck. All the above-mentioned studies focused on cervical level spinal cord. Following the evidence from neuroradiological investigation of spinal cord, showing thoracolumbar level is the common location for spinal blood collection, is it possible that spinal cord injuries were overlooked and that the real incidence of them from neuropathology investigation is underestimated? For instance when homicide victims from physical abuse under the age of 3 years old were studied by Serenelli et al. in 2017, the majority of spinal cord lesions were at spinal level lower than cervical . In this cohort of 51 children (42 AHT), the most common finding was SDH across the spinal cord. Spinal cord injuries at a thoracolumbar location accounted for the majority of the cases (33.3%) as compared with the lumbosacral area (27.5%) and the cervical level (15.5%). Thoracic location appeared more frequent in infants, and the correlation was statistically proven ( p = 0.048). The first recorded study examining the neuropathology of spinal cord injury related to AHT was carried out in 1989 by Hadley et al. when they studied 13 infants (mean age 3 months) who died of confessed shaking without evidence of head impact trauma (Table ). Six of them underwent post mortem examination, and spinal cords were examined. All except one of the autopsied children showed injuries in the spinal cord, five had epidural haematoma and four had subdural haematoma at the cervicomedullary junction along with contusions of the ventral high cervical levels. All the autopsied children who had SDH presented with cerebral contusions, swelling and herniations. Hadley concluded that spinal injury at the high cervical cord level can contribute to the dramatic outcome of shaking without direct cranial impact. He also underlined the very young age of the studied babies, suggesting infants are more susceptible to injury from shaking. Eight years later, Feldman et al. (1997) focused on spinal cord injuries in five AHT autopsied children (mean age 5.8 months) enrolled through the child protection team . The diagnosis of inflicted head injury was then corroborated by the infant’s attending physician. In only one case, subdural blood collection was seen in the cervical spinal cord, while 3/5 showed subarachnoid bleeding. Both the subdural and the subarachnoid haemorrhages were seen in association with similar intracranial findings, and subdural haematoma was in clear continuity with the spinal one. In the three autopsied children reported by Saternus et al. (2000) (mean age 16 months), the AHT diagnoses were assumed from the history taken in the police notes in association with the intracranial subdural haematoma and the absence of head impact signs . One had epidural haemorrhage at the cervical level, and none of the cases had subdural blood collections in the spinal cord. The cervical spine showed intervertebral disc rupture and blood collections in the soft tissue in 2/3 patients. Similarly, when Geddes et al. (2001) performed a comprehensive retrospective study in order to identify the neuropathological changes in AHT children, the only spinal blood collection was the epidural haematoma seen in three cases . In an attempt to determine specifically the neuropathological findings in the cervical spinal cord of AHT, Brennan et al. (2009) reported the outcomes from 41 children who underwent cervical examination at post mortem and assessed to have died from AHT by the chief medical examiner . A very high proportion of them showed meningeal haemorrhages (83%) (epidural, intradural, subdural and/or subarachnoid, in an unspecified proportion) and parenchymal lesions such as contusions, lacerations and transections (72%). Additionally, nerve root avulsions and dorsal root ganglion were seen in slightly more than a half (55%). None of the children had vertebral fractures, and only one fifth (21%) had soft-tissue injuries in the neck. All the above-mentioned studies focused on cervical level spinal cord. Following the evidence from neuroradiological investigation of spinal cord, showing thoracolumbar level is the common location for spinal blood collection, is it possible that spinal cord injuries were overlooked and that the real incidence of them from neuropathology investigation is underestimated? For instance when homicide victims from physical abuse under the age of 3 years old were studied by Serenelli et al. in 2017, the majority of spinal cord lesions were at spinal level lower than cervical . In this cohort of 51 children (42 AHT), the most common finding was SDH across the spinal cord. Spinal cord injuries at a thoracolumbar location accounted for the majority of the cases (33.3%) as compared with the lumbosacral area (27.5%) and the cervical level (15.5%). Thoracic location appeared more frequent in infants, and the correlation was statistically proven ( p = 0.048). The anatomical structure of the spinal cord was proposed as the reason for the thoracolumbar distribution of subdural collections . The blood would flow from the posterior fossa into the spinal canal, collecting at the thoracolumbar region, where “a natural convexity in the supine position is present”. In our review, all the cases in the radiological and pathological investigations showed spinal SDH in association with intracranial SDH which was an inclusion criteria in only 5/9 and 1/9 articles (Tables and ) (Figs. , and ). On the other hand, as recently shown by Rabbitt et al. (2019), spinal SDH in children evaluated for abusive head trauma is not commonly associated with intracranial haemorrhage ( p = 0.28) . The small size of the intracranial bleeding and the lack of SDH in the posterior fossa compartment are not consistent with the intracranial source of spinal blood subdual collections. Given the close contact of the subdural and subarachnoid sheets is easy to suppose that a small quantity of blood leaking from upper level is not enough to separate the two sheets and to collect downwards at the thoracolumbar level. In his case report, Gruber et al. (2008) described the origin of bleeding as primary in the spinal cord, intraoperatively seen as a lacerated radicular vein . According to the finding, the author suggested the blood vessels travelling along with the spinal nerve roots are the source of the spinal subdural haemorrhages. Anatomically, the spinal cord is surrounded by the meninges, in the same way as the brain but with some differences. The dura mater is composed of only one sheet which is the direct continuation of the inner meningeal layer of the cranial dura when the outer layer of intracranial dura mater ceases at the foramen magnum. The spinal dura mater is firmly attached to the circumference of the foramen magnum, to the second and third cervical vertebrae and with the posterior longitudinal ligaments as well. The sheath of dura mater is much larger than is necessary for the accommodation of its contents, and its size is greater in the cervical and lumbar regions than in the thoracic. The epidural space contains a plexus of veins, while the subdural cavity is not actual, but it is a virtual space as the dura is in close contact with the arachnoid. Therefore, it is possible that a large intracranial subdural haematoma may have enough volume and weight to force itself through the virtual subdural space and present as a spinal cord subdural haematoma. But it is less likely that a small mainly intracranial SDH (commonly seen in AHT) can travel through different compartments to reach the spinal cord. Both the dura and the arachnoid surround the spinal nerves at the level of their entrance in the spinal cord. The pia and arachnoid membranes continue along with the spinal nerve roots as they leave the spinal cord and exit through the intervertebral foramina, where they blend with the perineurium of the spinal nerves . Arterial and venous vessels run along the surface of the spinal cord, between the arachnoid and pia mater. The latter is composed of collagen and reticular fibres which wraps the surface of spinal cord, while collagen fibres are external and form bundles with the above arachnoid; it is in this virtual space where vessels are found . One possible source of the spinal cord SDH is the radicular veins which run along nerve roots, and then the spinal cord surface needs to penetrate the arachnoid membrane in order to reach the subarachnoid location, resulting in an area of weakness when exposed to a high energy trauma such as that in shaking. The laceration of radicular veins at the point of passage from subdural to subarachnoid space could explain the blood collection in the spinal subdural space. The hypothesis of primary damage in the spinal nerve roots is supported by further evidence, and damage in the nerve roots especially the dorsal nerve roots has been detected . Brennan et al. reported the frequency of nerve roots avulsion and dorsal root ganglion haemorrhages as up to 55% between AHT children . Likewise, the βAPP immunohistochemistry examinations were positive in the nerve roots of abused babies as compared with the control group where no expression was detectable . It is well known that spinal nerve roots are the site of CSF absorption, and so they are surrounded by a high density vein vessel mesh and therefore are prone to bleeding . An association of high intracranial pressure and vessel damage due to hypoxic endothelial damage has been suggested as the cause of SDH. It is possible that the same mechanisms play a role in spinal cord bleeding . The primary spinal source of blood collection is also supported by the increasing evidence that the thoracolumbar level is frequently involved in cases of spinal trauma due to abuse, such as Jauregui et al.’s observation of increased risk of thoracic (OR = 2.75) and lumbar (OR = 1.67) vertebral fractures in his cohort of 116 abused children . In a recent study on 51 homicide victims, the frequency of thoracic and lumbar spinal injuries was reported around 30% as compared with just 15.5% at cervical level . Consequently, the side of maximum forces could be thoracolumbar rather than the cervical as the thoracic spine with the ribcage provides another valuable pivot point at the level of its articulation with the lumbar spine. Detecting the primary source of blood collection is of great interest to clarify where the trauma forces acted and therefore to better understand the trauma mechanism. Further neuropathological studies looking at the whole spinal cord and spinal nerve roots are needed to solve the issue. Spinal cord (parenchymal) injuries When Hadley et al. (1989) studied six AHT cases at post mortem, four had spinal contusions . Twenty years later, Brennan et al. (2009) confirmed the recurrent involvement of spinal cord injury in abuse following the observation of parenchymal lesions in up to 72% of his cohort of 41 abused children . Parenchymal lesions have been studied mainly through histology (Figs. and ). The first histological documentation of spinal injury in shaken babies was from Shannon et al. (1998) when fourteen cases (mean age 5 months) of witnessed, confessed or corroborated shaking without skull fractures underwent CD68 and βAPP immunohistochemistry . Cervical spinal cord showed βAPP-positive axons in 7 out of 11 cases, along with the spinal nerve roots (especially in the glial head). On the contrary, none of the control group cases (death from hypoxic-ischaemic encephalopathy and asphyxia) showed positivity for βAPP staining in the same structure, and the author hypothesised that βAPP positivity in the white matter tracts of cervical spinal cord may be associated with shaking. Shannon also studied the medulla, midbrain and cerebral white matters, where no differences were seen between the two groups. When Geddes et al. (2001) performed a comprehensive retrospective study in order to identify the neuropathological changes in AHT children, the findings were remarkable . The study cohort was comprised of 53 AHT children (37 infants and 16 toddlers) retrospectively collected and corroborated according to diagnostic criteria proposed by the author based on perpetrator’s confession/conviction and extra-cranial injuries. On histology, 8 out of the 53 showed localized βAPP-axonal positivity in the corticospinal tracts of the lower brainstem (lower pons and medulla) as well as in cervical cord roots in three additional cases. The findings were explained by the author as the result of stretching forces acting on the corticospinal tract in the lower brainstem which may lead to apnoea and hypoxic-ischaemic damage. In the same year, Geddes et al. published another article where the same results from the 37 infants were compared with a control group of 14 infants who died from causes other than traumatic . Three out of the 28 AHT cases immunologically stained were positive for βAPP in cervical cord and/or dorsal nerve roots, while eight showed βAPP positivity in the lower pons and medulla. None of the controls was positive for βAPP in the brain stem and spinal cord. Although the current knowledge on spinal cord changes in association to abuse need to be better investigated, is it possible for the time being to suppose that parenchymal injuries are not specific for the mechanism of trauma but possibly useful to understand the pathological mechanisms following abusive traumas? When Hadley et al. (1989) studied six AHT cases at post mortem, four had spinal contusions . Twenty years later, Brennan et al. (2009) confirmed the recurrent involvement of spinal cord injury in abuse following the observation of parenchymal lesions in up to 72% of his cohort of 41 abused children . Parenchymal lesions have been studied mainly through histology (Figs. and ). The first histological documentation of spinal injury in shaken babies was from Shannon et al. (1998) when fourteen cases (mean age 5 months) of witnessed, confessed or corroborated shaking without skull fractures underwent CD68 and βAPP immunohistochemistry . Cervical spinal cord showed βAPP-positive axons in 7 out of 11 cases, along with the spinal nerve roots (especially in the glial head). On the contrary, none of the control group cases (death from hypoxic-ischaemic encephalopathy and asphyxia) showed positivity for βAPP staining in the same structure, and the author hypothesised that βAPP positivity in the white matter tracts of cervical spinal cord may be associated with shaking. Shannon also studied the medulla, midbrain and cerebral white matters, where no differences were seen between the two groups. When Geddes et al. (2001) performed a comprehensive retrospective study in order to identify the neuropathological changes in AHT children, the findings were remarkable . The study cohort was comprised of 53 AHT children (37 infants and 16 toddlers) retrospectively collected and corroborated according to diagnostic criteria proposed by the author based on perpetrator’s confession/conviction and extra-cranial injuries. On histology, 8 out of the 53 showed localized βAPP-axonal positivity in the corticospinal tracts of the lower brainstem (lower pons and medulla) as well as in cervical cord roots in three additional cases. The findings were explained by the author as the result of stretching forces acting on the corticospinal tract in the lower brainstem which may lead to apnoea and hypoxic-ischaemic damage. In the same year, Geddes et al. published another article where the same results from the 37 infants were compared with a control group of 14 infants who died from causes other than traumatic . Three out of the 28 AHT cases immunologically stained were positive for βAPP in cervical cord and/or dorsal nerve roots, while eight showed βAPP positivity in the lower pons and medulla. None of the controls was positive for βAPP in the brain stem and spinal cord. Although the current knowledge on spinal cord changes in association to abuse need to be better investigated, is it possible for the time being to suppose that parenchymal injuries are not specific for the mechanism of trauma but possibly useful to understand the pathological mechanisms following abusive traumas? It is well known that a disproportionately larger head in children is supported by weaker and more lax cervical muscle and ligaments than in adults . What appears peculiar is the topical distribution of spinal injuries, which seems to be age-related. Cervical spinal injuries were seen more frequently in younger infants (age range 1–48 months, median age 5 months), while thoracolumbar was more frequent in older infants (age range 6–16 months, median age 13.5 months) . According to the current literature, young children tend to injure the upper cervical spine at the craniocervical junction to the C3 spinal level . As previously reported, infants have larger heads in proportion to their body and a more horizontal vertebral facet, allowing a higher degree of freedom in motion. When children grow, injuries are seen mainly in the lower level reaching the adult proportion at about 8–10 years. It has been proven that upper cervical spine injury (C1–C4), cervical fracture and spinal cord injury, spinal cord injury without radiographic abnormality (SCIWORA) and dislocation show a downward trend with increasing age . As such, the pathophysiology of trauma could also be different. If younger children have craniocervical junction injury, which leads to hypoxic-ischaemic brain injury, older babies can a have major incidence of extracranial injury points such as at thoracolumbar spinal cord levels. It follows that radiological examination of the spinal cord should be routinely performed at all levels, particularly when the suspected abuse involves children older than 1 year of age. Intracranial subdual haematoma is a common finding in neonates. When 101 term neonates were imaged by MRI within the first 72 h of life, the incidence of SDH was as high as 46%, the topical distribution involving both supra and infra tentorial compartments . It has been suggested that SDH in a child over the age of 1 month should not be attributed completely to birth trauma . Spinal cord haemorrhage in newborns was studied by Vlasyuk et al. in 2013 when he analysed 14 premature still-born babies with intraventricular haemorrhages. All cases with grade III were accompanied by drops of fluid in the subarachnoid space of the cervical dorsal and lumbar parts of the spinal cord . No other cases of spinal blood in newborns are reported in the literature. Similarly, intracranial haemorrhage can be seen in newborns lacking vitamin K in association with cerebral oedema and can mimic abuse. Brain infection and spontaneous intracranial haemorrhages linked with diastasis may also mimic AHT as oedema, and SDH can be present on neuroimaging in cases of encephalopathy. The clinical manifestation is also age-related. Children under 1 year of age present with impaired consciousness and respiratory distress while older patients have spinal deformity and focal neurological signs . Infants have thin bilateral SDH and higher incidence of skull fracture, while children older than 1 year of age had larger intracranial SDH collections and a higher incidence of extracranial damage, suggesting that the head and craniocervical junction are the main injury points in infants, but not in older children . It could be supposed that shaking acts on different levels depending on the age of the victim. Forensic pathologist is usually asked to give a reason for the cause and mechanism of injuries detected through a wide-ranging examination. Then, abusive head trauma has been a source of interest for many years, and much effort has been spent in order to identify the characteristic features associated with it. Being a medical diagnosis AHT needs to be integrated with non-medical evidence to answer legal questions of actus reus (guilty act) or mens rea (guilty mind), and scientific evidence has the main role in providing an answer to the medical question of “whether an infant’s injuries were most likely caused by abuse or could they be plausibly explained by a hypothetical alternative” . Spinal cord injuries could play a role in supporting the challenging AHT diagnosis when investigated both radiologically at time of admission and neuropathologically at post mortem examination. According to a recently published consensus statement supported by multiple societies, spinal cord MRI assessment of all spinal levels is now recommended in cases of suspected child abuse . The vast majority of spinal cord assessment in cases of abuse has been obtained by MRI in cases of AHT variable incidence from 4.3 to 84.3% according to the different institutions . MRI has been shown to be an effective method of diagnosis in the paediatric population, because of the anatomical conformation of the spinal cord (made of cartilage in many vertebral components) and preferable to CT because of its high sensitivity for ligament injury . One of the common features associated to abusive head trauma is subdural haemorrhage in the spinal cord which is poorly investigated compared with widely studied intracranial subdural haemorrhage. This could be due to limited literature on the incidence of spinal findings in AHT cases, non-specificity of clinical signs predicting the risk of spinal peculiar injury, logistical issues, high cost of imaging the entire spine, limited spinal stability and limited understanding of forensic markers . Peculiarly, our results showed the more frequent location is thoracolumbar level rather than cervical. This data has major implication for the forensic pathologist as the suspect of abuse could not be completely ruled out without an investigation on the whole length of the spinal cord. The origin of spinal blood collections is most likely the traumatic damage to the radicular veins which run close to the spinal nerve roots; both of which are probably damaged in movement of spinal cord within the spinal column following excessive forward and backward movements of the head and neck commonly encountered in AHT. Another important feature associated to abuse is spinal ligamentous injury. Although it has been reported an incidence up to 78% in abused babies , association with abuse trauma was not statistically proven, and the role the finding plays in the diagnose of abusive head trauma has to be carefully considered on the singular case. The data from radiological or pathological investigation suggest that cervical injuries are mainly seen in infants while children older than 1 year of age frequently presented with spinal lesions at a lower level, a view that probably indicates that the same mechanism of trauma can lead to different spinal changes according to the age of the victim. In conclusion, spinal cord investigation is of great interest in forensic assessment of head trauma from abuse with enormous impact on the approach to child autopsies. According to the evidence shown here, forensic pathologists should routinely include the examination of the spinal cord in a systematic approach to AHT cases. Finding the typical intracranial changes of the triad together with spinal lesions will support the diagnosis of abusive head trauma with a higher level of certainty. On the contrary, overlooking the spinal changes could lead to harmful misinterpretation since the triad is nowadays considered not sufficient to support the diagnosis of abuse. Indeed, the view that spinal subdural blood collection is highly suggestive of abuse is in agreement with the clinical practice guidelines on shaken baby syndrome produced by the French Haute Autorité de Santé which underlines the importance of spinal SDH in differentiating abusive from accidental aetiology . In the guidelines, spinal cord lesions in association with unifocal intracranial SDH is considered sufficient to support the diagnosis of AHT within infants in whom differential diagnoses have been ruled out. In the forensic setting, where objectivity is the leading rule to follow, new objective anatomopathological data are of great interest to help the accurate evaluation of the diagnosis and to allow the court to reach the right verdict. The present work underlined the common finding of spinal blood collection at lower level than previously shown. Consequently, further studies need to be performed in order to better investigate the incidence of spinal blood collection at thoracolumbar level and the origin of it in association to abusive trauma, providing more evidence of the importance of spinal cord injuries in the recognition of abuse. |
Development of an Anthropomorphic Heterogeneous Female Pelvic Phantom and Its Comparison with a Homogeneous Phantom in Advance Radiation Therapy: Dosimetry Analysis | 9f8257e0-4dc8-4079-a9c4-724ee605cd4d | 10535781 | Internal Medicine[mh] | Cervical cancer is a significant global health concern and a leading cause of cancer-related deaths among women in many developing countries. Treatment modalities for cervical cancer depend on the stage and extent of the disease. They may include surgery, chemotherapy, immunotherapy, and radiation therapy. Radiotherapy plays a crucial role in the management of cervical cancer, especially in its early and locally advanced stages. It can be used as a primary treatment, combined with surgery or chemotherapy, or as a postoperative adjuvant therapy to prevent disease recurrence. The role of radiotherapy in cervical cancer treatment is to deliver a targeted dose of ionizing radiation to the tumor, destroying cancerous cells and shrinking tumors while sparing adjacent normal tissues. It is an effective and non-invasive treatment option that can achieve high cure rates and preserve fertility in early-stage cases, making it an indispensable component of comprehensive cervical cancer management . The paramount role of physics in radiation therapy is to continually enhance the precision and accuracy of delivering the radiation dose to the target volume. The historical evolution of radiation therapy witnessed a paradigm shift from a two-dimensional (2D) approach prevailing from the 1950s to the late 1980s. In this era, radiation plans were manually designed, and a single radiation beam was delivered from one to four directions using specialized shielding blocks for beam collimation . The advent of advanced imaging technologies, such as ultrasound (US), computed tomography (CT), and magnetic resonance imaging (MRI), led to a transformative transition to three-dimensional conformal therapy (3DCT). This revolutionary approach enables precise treatment field shaping to the target volume, ensuring uniform-intensity delivery to the tumor while sparing surrounding healthy tissues. The innovative multileaf collimator (MLC) system replaced traditional shielding blocks, which significantly improved treatment accuracy . In the 1990s and early 2000s, intensity-modulated radiation therapy (IMRT) emerged as a major advancement in cancer treatment. It provided more precise and conformal dose distributions compared to the traditional 3D conformal radiation therapy (3DCRT) technique . This led to further developments, such as intensity-modulated arc therapy (IMAT), which used dynamic manipulation of the multileaf collimator (MLC) while rotating the treatment machine’s gantry in an arc. IMAT then paved the way for volumetric modulated arc therapy (VMAT) in 2007, which incorporated variable gantry rotation speeds and dose rates, making treatments even more accurate and efficient. These advanced techniques have revolutionized modern radiation therapy, improving treatment outcomes and quality of life for cancer patients while reducing radiation exposure to healthy tissues . Also, these high-end radiation therapy techniques require accurate pretreatment and patient-specific quality assurance (PSQA) prior to the start of patient treatment . Two important factors must be considered in the evaluation of any radiotherapy plans or treatment procedures: a realistic environment that mimics how radiation interacts with real biological tissue, and a precise pretreatment plan verification system . Phantoms, which have been in use since the inspection of radiotherapy, are substitutes that conform to the real-body scenario. Although the majority of the human body consists of water, physical phantoms that are made of water or solid water-equivalent materials have mostly been used for PSQA . These phantoms were used because of their cost effectiveness, universal availability, uniform density of 1 g/cc, and simpler designs. However, we know that in addition to water, the human body consists of bones, soft tissues, air cavities, etc., of varying densities. So, there is a need to develop an anthropomorphic heterogeneous pelvic phantom that should exactly represent the actual human body. Previous studies must be consulted to ensure that the phantom would be constructed to be realistic in size and shape . The phantom would also be suitable for the assessment of accurate delivery of treatment doses, and improve dosimetry in clinical fields. By using this phantom, uncertainties during patient positioning and dosimetry can be reduced, making it an important tool for practical tests in India.
2.1. Phantom Design To design an anthropomorphic heterogeneous female pelvic (AHFP) phantom, the average pelvic dimensions of 50 adult female patients were utilized, as shown in A. To accurately replicate the radiological characteristics of the involved tissues, a combination of materials was chosen, including paraffin wax, water, gauze (cotton), polyvinyl chloride (PVC), and polymerized siloxanes, as seen in . For the uterine part, we mixed 150 g of polymerized siloxanes with 50 g of regular wax. This mixture helped us make a structure that feels like a real uterus. The rectum simulation involved a combination of materials: a PVC hollow pipe, paraffin wax, and a thin gauze piece. The foundation was the 14 cm long, 1.5 cm diameter, and 1 mm thick PVC pipe, housing 10 g of paraffin wax for soft tissue emulation. A 10 cm long, 1 mm thick gauze piece was inserted within the pipe to simulate fecal matter, making the rectal part more realistic. For the bladder, we used a balloon and filled it with 220 mL of water. This made the balloon act like a real bladder filled with urine. We combined paraffin wax with sodium chloride (salt) to replicate the characteristics of fat and skin. 2.2. Fabrication of Phantom Initially, a female pelvic dummy was meticulously crafted employing thermoplastic sheets and cloth tape. Subsequently, internal organ models were precisely situated to mirror the density of human pelvic bone, and their placement was secured through the application of gypsum bandages. These internal organs, along with the pelvic bones, were harmoniously integrated into the pelvic dummy, ensuring anatomical alignment. We then poured liquid paraffin wax for surface molding, subsequently allowing it to cool and stabilize. A cavity was prepared approximately at the uterus area in the phantom, and for that purpose, a 0.60 cc ion chamber (PTW, Freiburg, Germany) was kept at the same position to make sure that the cavity’s dimensions were equal to the ion chamber. Additionally, three reference points were created using fiducial lead markers placed on two bilateral points and one anterior point on the phantom’s surface within the same cross-sectional plane. The phantom’s physical measurements are 25.5 cm in terms of anterior–posterior separation, 32.7 cm in terms of lateral separation, and around 31.8 cm in the vertical dimension, with the extent being from the lower abdomen to the upper thigh region. The phantom weighs approximately 14.8 kg. 2.3. Comparison of the Hounsfield Units and the Relative Electron Densities of the Organs To determine how accurately the finished phantom product represents a real patient, the AHFP phantom was scanned with a CT scanner (Toshiba Alexion 16 multi-slice CT scanner) at 120 kVp and 250 mAs with a slice thickness of 2 mm. The CT images were transferred to the Eclipse treatment planning system (version 11.0.31) (Varian Medical Systems, Palo Alto, CA, USA). The CT images of the phantom were compared to CT images of randomly selected cervical cancer patients with similar scanning parameters (120 kVp, 250 mAs, and 2 mm slice thickness), which are shown in A,B. shows the mean and standard deviation of the CT number in Hounsfield units (HU) for patient and phantom CT images. The relative electron density of each of the materials was calculated by the given Formulas in (1) and (2) (Thomas, 2014) . P e = HU /1000 + 1 HU < 100 (1) P e = HU /1950 + 1 HU ≥ 100 (2) where p e is the relative electron density of the material. 2.4. Anatomical and Measuring Point Identification The anatomical structures, including the external body, skeletal or bone structure, bladder, rectum, uterus, femoral heads, and pelvic bone, were meticulously delineated using the Eclipse contouring station (Version 11.0.31). The delineated structures and contours were then exported to the Eclipse planning system, allowing for 3D visualization of each structure. To facilitate accurate dosimetry measurements, an ionization chamber was strategically positioned at the clinic’s areas of interest, particularly near critical structures. This allowed for precise monitoring and recording of the absorbed dose at these specific locations during the course of the treatment. 2.5. Pretreatment Plan Verification 2.5.1. Patient-Specific Absolute Dosimetry Two kinds of phantoms were chosen for the patient-specific absolute dosimetry of the completed RapidArc treatment plans. The first one was a homogeneous “water-equivalent RW3 solid phantom” (PTW Freiburg, Freiburg, Germany), as shown in B, each slab of which was made of polystyrene with the effective atomic number 5.74. The second phantom was the AHFP phantom, as shown in A. The density of the internal organs of this AHFP phantom was equivalent to that of the human pelvis. The CT scanning of the phantoms was conducted on a Toshiba Alexion 16 multi-slice CT scanner, with a slice thickness of 2 mm for planning purposes. The CT images were imported into the Eclipse (version 11.0.31) TPS (Varian Medical Systems, Palo Alto, CA, USA), and RapidArc plans already conducted for patient treatment were exported into both phantoms, which can be seen in A,B. Thirty cervical cancer patients who underwent RapidArc therapy, ranging in age from 37 to 70 years (average 53.5 years), were selected randomly for the study. Dual arcs were used for all the RapidArc plans since dual arcs can improve PTV coverage, enhance the modulation factor during optimization, and spare the OARs compared to single arcs. The first arc was a clockwise rotation with a gantry angle of 181° to 179° and a collimation angle of 30°. The second arc had a collimation angle of 330° and an anticlockwise rotation with gantry angles of 179° to 181°. All the selected plans were performed with a 6 MV photon beam, and field arrangement was conducted in such a way that all fields were coplanar with a couch angle of 0°. A dose volume optimizer (DVO) was used for plan optimization, and an anisotropic analytical algorithm (AAA) (version 11.30.1) with a grid size of 0.25 cm was used for dose calculation. All the plans were delivered, and the dose for each plan was measured using a PTW UNIDOSE electrometer connected with a 0.6 cc ionization chamber (IBA Dosimetry Germany), which was fixed in phantoms. The percentage variation between the measured dose of the linear accelerator and the planned dose of the TPS was calculated by the following formula: Percentage of variation = (measured dose of linac − TPS planned dose)/TPS planned dose × 100 The planned doses of the TPS and the measured doses from the machines of the homogeneous phantom and the AHFP phantom are compared and represented in and , respectively. 2.5.2. Relative Dosimetry To evaluate the effectiveness of the AHFP phantom as a quality assurance (QA) tool, planning target volumes (PTVs) were generated for the phantom. To assess the dose received by healthy organs during radiation therapy, organs at risk (OARs), like the bladder and rectum, were also considered. Both RapidArc and intensity-modulated radiation therapy (IMRT) plans were created on a treatment planning system (TPS), and the anisotropic analytical algorithm (AAA) (version 11.0.31) was used to calculate the dose. The 2D fluence generated by the TPS on the electronic portal imaging device (EPID) was sent to the linear accelerator (linac) for further analysis, which is represented in . Most modern linacs are equipped with flat-panel detectors based on amorphous silicon (aS1000 model) for megavoltage imaging. Various methods have been developed to utilize EPIDs in IMRT/RapidArc patient-specific quality assurance (PSQA). All measurements were carried out using an EPID detector calibrated for a 100 cm source-to-imager distance (SID). Data collection was performed with the same gantry and collimator positions specified in the treatment plan. The imaging system software was employed to compare and analyze the 2D fluence imaging obtained from the treatment planning system (TPS). For plan evaluation, pixel-based passing criteria were utilized. A pass condition was set at an average gamma value (g) of ≤1, indicating successful plan agreement, while a failure condition was defined as g > 1, indicating discrepancies. In our assessment, we used specific tolerance levels as acceptance standards. These included a distance-to-agreement (DTA) of 3 mm, representing the maximum allowed spatial difference between the measured and expected dose distributions. Additionally, a dose difference (DD) of 3% was considered, signifying the permissible variation between the measured and expected doses. 2.6. Statistical Analysis The statistical analysis in this study involved a paired two-tailed Student’s t -test using SPSS ® v.13.0 (SPSS Inc., Chicago, IL, USA) to compare the differences between the homogeneous RW3 phantom and the AHFP phantom. A significance level of p < 0.05 was considered as statistically significant to determine the presence of significant differences.
To design an anthropomorphic heterogeneous female pelvic (AHFP) phantom, the average pelvic dimensions of 50 adult female patients were utilized, as shown in A. To accurately replicate the radiological characteristics of the involved tissues, a combination of materials was chosen, including paraffin wax, water, gauze (cotton), polyvinyl chloride (PVC), and polymerized siloxanes, as seen in . For the uterine part, we mixed 150 g of polymerized siloxanes with 50 g of regular wax. This mixture helped us make a structure that feels like a real uterus. The rectum simulation involved a combination of materials: a PVC hollow pipe, paraffin wax, and a thin gauze piece. The foundation was the 14 cm long, 1.5 cm diameter, and 1 mm thick PVC pipe, housing 10 g of paraffin wax for soft tissue emulation. A 10 cm long, 1 mm thick gauze piece was inserted within the pipe to simulate fecal matter, making the rectal part more realistic. For the bladder, we used a balloon and filled it with 220 mL of water. This made the balloon act like a real bladder filled with urine. We combined paraffin wax with sodium chloride (salt) to replicate the characteristics of fat and skin.
Initially, a female pelvic dummy was meticulously crafted employing thermoplastic sheets and cloth tape. Subsequently, internal organ models were precisely situated to mirror the density of human pelvic bone, and their placement was secured through the application of gypsum bandages. These internal organs, along with the pelvic bones, were harmoniously integrated into the pelvic dummy, ensuring anatomical alignment. We then poured liquid paraffin wax for surface molding, subsequently allowing it to cool and stabilize. A cavity was prepared approximately at the uterus area in the phantom, and for that purpose, a 0.60 cc ion chamber (PTW, Freiburg, Germany) was kept at the same position to make sure that the cavity’s dimensions were equal to the ion chamber. Additionally, three reference points were created using fiducial lead markers placed on two bilateral points and one anterior point on the phantom’s surface within the same cross-sectional plane. The phantom’s physical measurements are 25.5 cm in terms of anterior–posterior separation, 32.7 cm in terms of lateral separation, and around 31.8 cm in the vertical dimension, with the extent being from the lower abdomen to the upper thigh region. The phantom weighs approximately 14.8 kg.
To determine how accurately the finished phantom product represents a real patient, the AHFP phantom was scanned with a CT scanner (Toshiba Alexion 16 multi-slice CT scanner) at 120 kVp and 250 mAs with a slice thickness of 2 mm. The CT images were transferred to the Eclipse treatment planning system (version 11.0.31) (Varian Medical Systems, Palo Alto, CA, USA). The CT images of the phantom were compared to CT images of randomly selected cervical cancer patients with similar scanning parameters (120 kVp, 250 mAs, and 2 mm slice thickness), which are shown in A,B. shows the mean and standard deviation of the CT number in Hounsfield units (HU) for patient and phantom CT images. The relative electron density of each of the materials was calculated by the given Formulas in (1) and (2) (Thomas, 2014) . P e = HU /1000 + 1 HU < 100 (1) P e = HU /1950 + 1 HU ≥ 100 (2) where p e is the relative electron density of the material.
The anatomical structures, including the external body, skeletal or bone structure, bladder, rectum, uterus, femoral heads, and pelvic bone, were meticulously delineated using the Eclipse contouring station (Version 11.0.31). The delineated structures and contours were then exported to the Eclipse planning system, allowing for 3D visualization of each structure. To facilitate accurate dosimetry measurements, an ionization chamber was strategically positioned at the clinic’s areas of interest, particularly near critical structures. This allowed for precise monitoring and recording of the absorbed dose at these specific locations during the course of the treatment.
2.5.1. Patient-Specific Absolute Dosimetry Two kinds of phantoms were chosen for the patient-specific absolute dosimetry of the completed RapidArc treatment plans. The first one was a homogeneous “water-equivalent RW3 solid phantom” (PTW Freiburg, Freiburg, Germany), as shown in B, each slab of which was made of polystyrene with the effective atomic number 5.74. The second phantom was the AHFP phantom, as shown in A. The density of the internal organs of this AHFP phantom was equivalent to that of the human pelvis. The CT scanning of the phantoms was conducted on a Toshiba Alexion 16 multi-slice CT scanner, with a slice thickness of 2 mm for planning purposes. The CT images were imported into the Eclipse (version 11.0.31) TPS (Varian Medical Systems, Palo Alto, CA, USA), and RapidArc plans already conducted for patient treatment were exported into both phantoms, which can be seen in A,B. Thirty cervical cancer patients who underwent RapidArc therapy, ranging in age from 37 to 70 years (average 53.5 years), were selected randomly for the study. Dual arcs were used for all the RapidArc plans since dual arcs can improve PTV coverage, enhance the modulation factor during optimization, and spare the OARs compared to single arcs. The first arc was a clockwise rotation with a gantry angle of 181° to 179° and a collimation angle of 30°. The second arc had a collimation angle of 330° and an anticlockwise rotation with gantry angles of 179° to 181°. All the selected plans were performed with a 6 MV photon beam, and field arrangement was conducted in such a way that all fields were coplanar with a couch angle of 0°. A dose volume optimizer (DVO) was used for plan optimization, and an anisotropic analytical algorithm (AAA) (version 11.30.1) with a grid size of 0.25 cm was used for dose calculation. All the plans were delivered, and the dose for each plan was measured using a PTW UNIDOSE electrometer connected with a 0.6 cc ionization chamber (IBA Dosimetry Germany), which was fixed in phantoms. The percentage variation between the measured dose of the linear accelerator and the planned dose of the TPS was calculated by the following formula: Percentage of variation = (measured dose of linac − TPS planned dose)/TPS planned dose × 100 The planned doses of the TPS and the measured doses from the machines of the homogeneous phantom and the AHFP phantom are compared and represented in and , respectively. 2.5.2. Relative Dosimetry To evaluate the effectiveness of the AHFP phantom as a quality assurance (QA) tool, planning target volumes (PTVs) were generated for the phantom. To assess the dose received by healthy organs during radiation therapy, organs at risk (OARs), like the bladder and rectum, were also considered. Both RapidArc and intensity-modulated radiation therapy (IMRT) plans were created on a treatment planning system (TPS), and the anisotropic analytical algorithm (AAA) (version 11.0.31) was used to calculate the dose. The 2D fluence generated by the TPS on the electronic portal imaging device (EPID) was sent to the linear accelerator (linac) for further analysis, which is represented in . Most modern linacs are equipped with flat-panel detectors based on amorphous silicon (aS1000 model) for megavoltage imaging. Various methods have been developed to utilize EPIDs in IMRT/RapidArc patient-specific quality assurance (PSQA). All measurements were carried out using an EPID detector calibrated for a 100 cm source-to-imager distance (SID). Data collection was performed with the same gantry and collimator positions specified in the treatment plan. The imaging system software was employed to compare and analyze the 2D fluence imaging obtained from the treatment planning system (TPS). For plan evaluation, pixel-based passing criteria were utilized. A pass condition was set at an average gamma value (g) of ≤1, indicating successful plan agreement, while a failure condition was defined as g > 1, indicating discrepancies. In our assessment, we used specific tolerance levels as acceptance standards. These included a distance-to-agreement (DTA) of 3 mm, representing the maximum allowed spatial difference between the measured and expected dose distributions. Additionally, a dose difference (DD) of 3% was considered, signifying the permissible variation between the measured and expected doses.
Two kinds of phantoms were chosen for the patient-specific absolute dosimetry of the completed RapidArc treatment plans. The first one was a homogeneous “water-equivalent RW3 solid phantom” (PTW Freiburg, Freiburg, Germany), as shown in B, each slab of which was made of polystyrene with the effective atomic number 5.74. The second phantom was the AHFP phantom, as shown in A. The density of the internal organs of this AHFP phantom was equivalent to that of the human pelvis. The CT scanning of the phantoms was conducted on a Toshiba Alexion 16 multi-slice CT scanner, with a slice thickness of 2 mm for planning purposes. The CT images were imported into the Eclipse (version 11.0.31) TPS (Varian Medical Systems, Palo Alto, CA, USA), and RapidArc plans already conducted for patient treatment were exported into both phantoms, which can be seen in A,B. Thirty cervical cancer patients who underwent RapidArc therapy, ranging in age from 37 to 70 years (average 53.5 years), were selected randomly for the study. Dual arcs were used for all the RapidArc plans since dual arcs can improve PTV coverage, enhance the modulation factor during optimization, and spare the OARs compared to single arcs. The first arc was a clockwise rotation with a gantry angle of 181° to 179° and a collimation angle of 30°. The second arc had a collimation angle of 330° and an anticlockwise rotation with gantry angles of 179° to 181°. All the selected plans were performed with a 6 MV photon beam, and field arrangement was conducted in such a way that all fields were coplanar with a couch angle of 0°. A dose volume optimizer (DVO) was used for plan optimization, and an anisotropic analytical algorithm (AAA) (version 11.30.1) with a grid size of 0.25 cm was used for dose calculation. All the plans were delivered, and the dose for each plan was measured using a PTW UNIDOSE electrometer connected with a 0.6 cc ionization chamber (IBA Dosimetry Germany), which was fixed in phantoms. The percentage variation between the measured dose of the linear accelerator and the planned dose of the TPS was calculated by the following formula: Percentage of variation = (measured dose of linac − TPS planned dose)/TPS planned dose × 100 The planned doses of the TPS and the measured doses from the machines of the homogeneous phantom and the AHFP phantom are compared and represented in and , respectively.
To evaluate the effectiveness of the AHFP phantom as a quality assurance (QA) tool, planning target volumes (PTVs) were generated for the phantom. To assess the dose received by healthy organs during radiation therapy, organs at risk (OARs), like the bladder and rectum, were also considered. Both RapidArc and intensity-modulated radiation therapy (IMRT) plans were created on a treatment planning system (TPS), and the anisotropic analytical algorithm (AAA) (version 11.0.31) was used to calculate the dose. The 2D fluence generated by the TPS on the electronic portal imaging device (EPID) was sent to the linear accelerator (linac) for further analysis, which is represented in . Most modern linacs are equipped with flat-panel detectors based on amorphous silicon (aS1000 model) for megavoltage imaging. Various methods have been developed to utilize EPIDs in IMRT/RapidArc patient-specific quality assurance (PSQA). All measurements were carried out using an EPID detector calibrated for a 100 cm source-to-imager distance (SID). Data collection was performed with the same gantry and collimator positions specified in the treatment plan. The imaging system software was employed to compare and analyze the 2D fluence imaging obtained from the treatment planning system (TPS). For plan evaluation, pixel-based passing criteria were utilized. A pass condition was set at an average gamma value (g) of ≤1, indicating successful plan agreement, while a failure condition was defined as g > 1, indicating discrepancies. In our assessment, we used specific tolerance levels as acceptance standards. These included a distance-to-agreement (DTA) of 3 mm, representing the maximum allowed spatial difference between the measured and expected dose distributions. Additionally, a dose difference (DD) of 3% was considered, signifying the permissible variation between the measured and expected doses.
The statistical analysis in this study involved a paired two-tailed Student’s t -test using SPSS ® v.13.0 (SPSS Inc., Chicago, IL, USA) to compare the differences between the homogeneous RW3 phantom and the AHFP phantom. A significance level of p < 0.05 was considered as statistically significant to determine the presence of significant differences.
Overall, there is good agreement between the measured CT number (HU) and relative electron density (RED) of the AHFP phantom and the patient groups. displays the findings of the comparison between measured CT numbers from a sample of patients from our institution who were selected at random and the CT numbers of the phantom. Hence, it was observed that the AHFP fabricated for this study matched both the qualitative and quantitative aspects of the CT evaluation. In the case of the homogeneous phantom, the mean percentage variations between planned and measured doses of all rapid arc QA plans were 1.4299, and the standard deviation was 0.768 (t = 0.00508, ρ = 0.497982). The result is not significant at p < 0.05, as shown in . For the AHFP phantom, the mean percentage variations between planned and measured doses of all rapid arc QA plans were 6.890, and the standard deviation was 2.565 (t = 3.21604, ρ = 0.001063 < 0.05). The outcome is significant, as shown in . The comparative study of the percentage of variation between the homogeneous slab phantom and AHFP phantom is given in . The t-value is −11.17016 and the p -value is <0.00001. The result is significant at p < 0.05, and their graphical representation is shown in . The results obtained from relative dosimetry are tabulated in . The average gamma values of the RapidArc plans are 0.29, 0.32, and 0.35 (g ≤ 1), and these values for the IMRT plans are 0.45, 0.44, and 0.42 (g ≤ 1) for targets 1, 2, and 3, respectively.
The results of our research, as presented in , demonstrate a close similarity between the Hounsfield unit (HU) and relative electron density (RED) values of the locally manufactured AHFP phantom and those of a human female pelvis. This finding aligns with previous research that emphasizes the importance of phantom design for dose measurement accuracy Johns & Cunningham . Our study emphasizes the significance of tissue-equivalent phantoms, such as the AHFP, in achieving clinically relevant and precise dose measurements, as discussed by Almond et al. . Numerous studies have contributed valuable data on the Hounsfield units and relative electron densities of human tissues, supporting the consistency and accuracy of these measurements. Winslow et al. determined the Hounsfield units for human muscles, and established the soft tissue equivalent range as −55 to −155, with the bone tissue equivalent range being 660. This aligns with the research by Trujillo-Bastidas et al. and Kanematsu , which reported relative electron densities for adipose, muscle, and bone tissues as 0.97, 1.05, and 1.4 and 0.96, 1.06, and 1.12, respectively. Similarly, the research conducted by Shrotriya et al. and S. Singh et al. revealed relative electron density values for bladder, rectum, fat, and bone tissues that closely align with our study’s findings (1.015, 1.069, 0.909, and 1.628, respectively). Shrotriya et al. (2018) reported relative electron densities of 1.31, 1.025, 0.91, and 1.6, respectively, while S. Singh et al. (2020) observed values of 1.04, 1.05, 0.89, and 1.63 for the same tissues. These consistent results across studies contribute to the overall understanding and validation of relative electron density values for different tissues, adding to the reliability of dosimetry calculations in radiation therapy planning. A variety of techniques have been created to compare sets of planned and measured radiation dose distributions in radiotherapy dosimetry. Here, we compare the homogeneous phantom’s and the AHFP phantom’s measured dose of the linac (Clinac iX medical linear accelerator, Varian Medical System, Palo Alto, CA, USA) and the planned dose of the Eclipse planning system (Version 11.0.31) (Varian Medical System, Palo Alto, CA, USA). In the case of the homogenous phantom, there is a less than 3% difference in the percentage between planned and measured doses, with a standard deviation of 0.7682 (t = 0.00508, p = 0.497982. The result is not significant at p < 0.05. The deviations of the planned and measured values of the dose of the AHFP phantom were found to be 10.67% (maximum value), 2.31% (minimum value), and 6.89% (average value), with a standard deviation of 2.565 (t = 3.21604, p = 0.001063). The result is significant at p < 0.05. Also, for the percentage of variation between homogeneous phantoms and AHFP phantoms, the t-value is −11.17016 and the p -value is <0.00001. The result is therefore significant at p < 0.05. In , we can see that the outcomes differ significantly due to the influence of heterogeneous media. The observed deviations in dose measurements between planned and measured values in the AHFP phantom could have potential clinical implications. Higher deviations could lead to suboptimal treatment delivery, affecting treatment outcomes and patient safety. Similar concerns have been raised in previous research, indicating the importance of minimizing dose calculation errors to improve treatment quality and patient safety (ICRU Report 50 ). The human body is made up of various densities, such as fat, bones, air cavities, and tissue. The amount of radiation dose deposited at the interface of two mediums varies significantly due to the difference in electron densities between the two media. Because bones have a larger density than soft tissue, they produce more secondary electrons . As a result, the dosage at the bone–soft tissue interface is higher. A similar phenomenon occurs at the interface of all two metals with different densities. Heterogeneity is one of the hardest problems that dose calculation algorithms must solve. The TPS currently uses newer and more precise algorithms that, like AAA, apply the heterogeneity adjustment factor when calculating dose . The patient-specific absolute dosimetry should be carried out using a heterogeneous phantom that mimics the density of the human body to confirm the correctness of the dose computed by these algorithms in the instance of each patient. O. Gurjar et al. conducted a study on radiation dosimetry for a contemporary radiotherapy approach, employing a real tissue phantom . With IMRT (head phantom) and IMRT (tissue phantom), the mean percentage deviation between planned and measured doses was found to be 2.36 (SD: 0.77) and 3.31 (SD: 0.78), respectively. Although the percentage variation in the case of the head phantom was within the tolerance limit (3%), it was nonetheless larger than the outcomes obtained utilizing phantoms that were readily available in the marketplace . And the majority of tissue phantom cases had percentage variations that exceeded the tolerance level. Chen et al. and Lee et al. conducted dosimetric validation and accuracy assessment of an in-house-developed anthropomorphic heterogeneous female pelvic phantom. Their findings showcased the suitability of the phantom for radiotherapy quality assurance, confirming its ability to accurately simulate patient anatomy and tissue heterogeneity. This study emphasizes the importance of utilizing reliable phantoms to ensure precise dose delivery and patient safety during treatment. Collectively, these studies highlight the importance of utilizing anthropomorphic heterogeneous female pelvic phantoms for accurate dosimetric verification in radiotherapy. The use of such phantoms improves confidence in treatment planning processes and ensures optimal dose delivery to target volumes while sparing healthy tissues. Further research and validation of these phantoms on a larger scale will likely enhance their clinical applicability and contribute to improved patient outcomes in radiation therapy. In , we present a comprehensive assessment of all planned target locations on the AHFP phantom, utilizing both RapidArc and IMRT treatment techniques. The table provides essential parameters, including area gamma, maximum dose difference, and average dose difference, for each target location. The results highlight the remarkable agreement between the measured and expected dose distributions, as indicated by the area gamma values ranging from 97.9% to 99.8% across all target locations. This excellent level of concurrence underscores the AHFP phantom’s ability to faithfully replicate radiation interactions and accurately deliver doses to the intended target volumes. Examining the maximum dose difference, we observe variations ranging from 18.7% to 39.4% for the different target locations. Additionally, the average dose difference ranges from 1.1% to 2.5%. These values, although demonstrating some variability, remain well within acceptable tolerance limits for relative dosimetric purposes. Our research findings are in line with the study conducted by Smith et al. . By utilizing the gamma index approach, both studies assessed the agreement between calculated and measured dose distributions, which provided crucial insights into the accuracy of treatment planning. The study by Smith et al. reinforces the significance of gamma index analysis as a valuable and comprehensive tool for dosimetric evaluation, further validating its importance in radiation therapy quality assurance and treatment optimization. Overall, these findings underscore the AHFP phantom’s effectiveness as a reliable quality assurance (QA) tool in radiotherapy. Its capability to mimic human anatomy and accurately simulate radiation dose delivery allows it to play a crucial role in verifying treatment plans and ensuring precise dose administration to target areas. The AHFP phantom’s performance, as evidenced by these results, reinforces its significance in advancing the safety and efficacy of radiotherapy procedures.
This study highlights the importance of precise dose delivery in radiotherapy for cervical cancer treatment. The anthropomorphic heterogeneous female pelvic (AHFP) phantom successfully replicates the radiological characteristics of human tissues, providing a realistic environment for accurate dose measurements. The comparison between the AHFP phantom and patient CT images demonstrates close similarity, confirming the phantom’s suitability for patient-specific quality assurance. The AHFP phantom’s measured dose variations using RapidArc plans indicate significant discrepancies compared to homogeneous phantoms, emphasizing the impact of heterogeneous media on dose calculations. However, the AHFP phantom’s performance remains within acceptable limits for relative dosimetric purposes. Furthermore, the AHFP phantom proves effective in plan verification, with high agreement between measured and expected dose distributions. The phantom’s ability to simulate human anatomy and accurately deliver doses to target volumes reinforces its role as a reliable QA tool in radiotherapy. Overall, the AHFP phantom’s development and evaluation contribute to advancing the precision and efficacy of radiotherapy for cervical cancer, ultimately improving patient outcomes, and ensuring safer and more effective treatment strategies.
|
Pharmacogenomics of clinical response to Natalizumab in multiple sclerosis: a genome-wide multi-centric association study | 82fe8755-60ea-44df-bbae-18583d718d6f | 11561017 | Pharmacology[mh] | Multiple Sclerosis (MS; MIM 126200) is a disease of the central nervous system (CNS) characterized by chronic inflammation, demyelination, and axonal loss . It is a complex multifactorial disorder, with both genetic and environmental components playing a role in disease susceptibility. Genome-wide association studies (GWAS) greatly helped to elucidate genetic susceptibility for MS, revealing a highly polygenic architecture, with an ever-increasing number of common SNPs associated with risk . Despite the expanded availability of multiple disease-modifying therapies (DMT), to date, no single drug has proven to be effective in controlling or delaying disease progression in the vast majority of patients. Further, the disease is highly heterogeneous and unpredictable in its expression and marked inter-individual differences have been observed in response to different treatments: currently, there is in fact paucity of biological markers that can help identifying responders and non-responders before starting a new drug . The identification of such biomarkers can help the neurologist to optimize treatment strategies, thus performing treatment decisions in clinical practice on an individual or stratum basis. This is a critical need also for Natalizumab (NTZ), a second-line DMT approved in 2004 for relapsing–remitting course of MS, with established clinical efficacy in reducing the rate of clinical relapses, risk of sustained disability progression, and the number of new or enlarging brain lesions on Magnetic Resonance Imaging (MRI) . The drug is a humanized monoclonal antibody that selectively inhibits α4β1 and α4β7 integrins, expressed on the surface of lymphocytes, hindering their binding to vascular endothelial adhesion molecules and their migration in the CNS across the blood–brain barrier (BBB), with the result of diminishing inflammation . Despite the high efficacy, a subset of patients treated with NTZ, estimated around 25% , do not respond or respond sub-optimally to the drug. To date, only a few candidate gene studies have been pursued to identify the factors that can genetically influence the response to NTZ, mostly relying on the putative mechanism of action of the drug. The difficulty in collecting large cohorts of well-phenotyped patients has hampered sufficiently powered pharmacogenetic studies. In this multi-center study, we report the results from a meta-analysis of genome-wide screens of common variants with response to NTZ in cohorts of MS patients from Italy, Germany and Sweden followed up for 4 years. Analyses were conducted at variant and pathway level, followed by a network approach to investigate joint association signals and to facilitate elucidation of mechanisms underlying response to drug. We then tested association of two genes emerging from a previous candidate study of response to NTZ , focusing on detoxification enzymes that counteract toxic compounds of oxidative stress (OS): a variety of reactive oxygen species are in fact produced in MS pathogenesis, enhancing mitochondrial injury, energy failure, and consequent oligodendrocyte apoptosis and putative role in response may be played by detoxification enzymes in the context of OS. Study population The study included patients enrolled at three centers, from Italy, Germany, and Sweden. Given the diversity in clinical data collection, a harmonization effort was put in place using data dictionary, which defined the variables that were then utilized in the common harmonized database. Data dictionary also defined the unit types that the data would be transformed to, and the final values for enumeration type variables. The data from the three centers were then mapped into the data dictionary variables by providing file name, column name, and unit type for each, and in case of enumeration types, the values were mapped to the harmonized values. For each center, we constructed the study cohort including patients for which there were imputed data and complete availability of baseline variables like age and disease duration and the number of relapses 2 years before starting NTZ. We excluded: (i) patients with age at treatment start <18 years and >55 years, (ii) patients who were on progressive courses (primary or secondary) at treatment start, and (iii) patients with Expanded Disability Status Scale (EDSS) >4 at treatment start, given that these patients were likely in secondary progressive phase and thus different from the rest of the cohort. The resulting values from harmonization phase were generated in longitudinal format, where each row corresponded to one treatment exposure for the patient. If a patient had multiple exposures to NTZ, the first observation with exposure >12 months was considered eligible. If none of the observations had exposure larger than this threshold, the observation with the largest exposure was considered independently of the order. In case of a previous short exposure to NTZ, the patient was included in the analysis only if the interval between the two exposures was longer than 12 months, to avoid any reactivation/rebound activity during the second NTZ treatment related to the withdrawal of the previous exposure. In the same way, patients previously treated with fingolimod were not included in the analysis, unless the interval between fingolimod withdrawal and NTZ start was longer than 12 months. The study was approved by the local ethical committees. Response to therapy We assessed response to NTZ with a dichotomous outcome, designating as responders patients who were relapse-free in the 4 years follow-up and non-responders those who experienced at least one relapse. Relapses were defined as new symptoms or exacerbation of existing symptoms persisting for ≥24 h, in the absence of concurrent illness/fever, and occurring ≥30 days after a previous relapse. Quality control Our study cohort was derived as a subset of a larger multi-centric dataset, that constituted the replication cohort for study on MS severity, performed in the context of the International Multiple Sclerosis Genetic Consortium (IMSGC). Details on pre-imputation quality control, phasing and imputation steps are thus described therein . Quality control steps on imputed data were performed within data from each center. For the two centers that had multiple distinct genotyping platforms (Italy and Sweden), we performed post-imputation quality control after obtaining a unique merged dataset. On a ‘per-marker’ basis, we excluded variants that: (1) had a call rate less than 95%; (2) had minor allele frequency (MAF) below 5%; (3) deviated from Hardy–Weinberg Equilibrium exact test at p < 10 -5 . On a ‘per-individual’ basis we excluded subjects who had high rates of genotype missingness (>5%) and one member of each pair of samples that showed, across platforms and within centers, high degree of recent shared ancestry (up to the second degree of kinship) inferred by robust estimation of their kinship coefficient . We finally used Principal Component Analysis (PCA) pruning from the data variants with a call rate less than 99% and regions of extended linkage disequilibrium, to control for population stratification and to discard individuals with outlying values in ancestry. We considered as outliers those samples being more than 4 standard deviations away from the mean of the first two PCs. All the quality control steps were performed with PLINK 2.0 . Association analysis The workflow of the study is depicted in Fig. . We performed single-SNP association analysis fitting logistic regression models as implemented in PLINK 2.0, assuming additive effects of imputed continuous dosages of minor alleles. Models were adjusted for age and disease duration at treatment start, sex, the number of relapses in the 2 years preceding NTZ therapy, and the first five eigenvectors from PCA to account for population substructure. Summary statistics were aggregated using fixed-effect meta-analysis with inverse-variance weighting of log(odds-ratios), as implemented in PLINK 2.0. Variants were annotated with ANNOVAR and visualization of the top-associated locus was generated via regional plot with LocusZoom . Gene-based analysis was conducted by means of Multi-marker Analysis of GenoMic Annotation (MAGMA, ) method v1.10, adjusting for the same set of covariates. The tool accounts for linkage disequilibrium and confounders like gene size and density. We used the multi option, which combines evidence from three models (principal components regression, mean of SNP squared Z-scores, and top SNP association). A critical choice in gene-based and gene-set analysis is the assignment of SNPs to genes, since inclusion of noisy variants can be detrimental to association analysis. We assigned SNPs to the target gene with a “proximity rule” using a flanking window of 5 kb, to minimize overlap between nearby genes. Second, we applied a “functional rule” by: integrating in target genes cis-eQTL SNPs, based on significant SNP–gene associations in immune cells (FDR < 5%), as available in DICE repository , that identified common genetic variants that are associated with the expression of > 12,000 genes in 13 human immune cell types. integrating in target genes those variants significantly affecting splicing regulation (FDR<5%), using a catalogue of cis-splicing QTLs (sQTL, ) processed on transcriptome data in blood tissue from the Genotype-Tissue Expression Consortium . Gene-wise statistics were then meta-analysed with weighted Stouffer’s procedure, which combines the Z-scores for each strata with weights set to the square root of the sample size. Approximately independent signals were identified upon application of a clumping procedure (primary p1 < 5 × 10 -5 , secondary p 2 < 0.01, r 2 > 0.6, maximum distance = 250 kb). Gene-set analysis We used the gene-wise meta-analytic p values as input for gene-set analysis of association using as reference Gene Ontology (GO) Biological Processes, retrieved from Human Molecular Signature Database (MSigDB v2023.1, ). We filtered out GO terms with less than 10 and more than 500 annotated genes, finally testing 5376 gene-sets. Each GO term was tested under the competitive null hypothesis, which states that aggregated variation in genes annotated to the gene set is no more associated with the outcome than that in all other genes in the genome. To accomplish this, we used as background signal the whole set of genes used in meta-analysis ( n = 24,110). Network analysis We performed a subnetwork detection analysis, projecting meta-analysed statistics onto STRING v11.5 reference interactome . We retained only those links with high interaction evidence (score > 0.7) in one of the three evidence domains: (i) protein–protein interaction, derived from multiple interactomes, such as IntAct, BioGrid, MINT, and others; (ii) co-expression, which leverages gene expression data from multiple sources; (iii) databases, which collects evidence of interaction from curated pathway resources. We then used dmGWAS tool , which applies a greedy search algorithm of dense modules within the node-weighted interactome, to detect association signals that aggregate in subnetworks. This procedure scores each module by a Z-score corresponding to the association level of the gene: the module score is obtained dividing the sum of the nodes scores by the square root of each module size. Starting from each seed, the procedure examines first-order neighbours and identifies those that generate the maximum increment of module score. We selected the top 1% of the top-scoring modules and merged them in a final subnetwork. The top-scoring subnetwork was imported into Cytoscape v3.8 environment for visualization, manipulation, and extraction of topologically relevant nodes (hubs, bottlenecks) with CentiScaPe plugin . We computed distributions of graph centrality metrics like degree, betweenness and eigenvector centrality for the detected module and selected nodes residing in the top 5% of at least one of the four distributions: these metrics should measure the functional importance of genes in the module. ClusterProfiler R package was used to perform gene-set over-representation analysis with hypergeometric test of the genes annotated to the extracted module and the detected communities, using GO Biological Process domain as reference database; the list of genes from meta-analysis that were present on the filtered STRING interactome was used as background universe. Benjamini–Hochberg adjusted p values < 0.05 were used to nominate significant gene-sets. The study included patients enrolled at three centers, from Italy, Germany, and Sweden. Given the diversity in clinical data collection, a harmonization effort was put in place using data dictionary, which defined the variables that were then utilized in the common harmonized database. Data dictionary also defined the unit types that the data would be transformed to, and the final values for enumeration type variables. The data from the three centers were then mapped into the data dictionary variables by providing file name, column name, and unit type for each, and in case of enumeration types, the values were mapped to the harmonized values. For each center, we constructed the study cohort including patients for which there were imputed data and complete availability of baseline variables like age and disease duration and the number of relapses 2 years before starting NTZ. We excluded: (i) patients with age at treatment start <18 years and >55 years, (ii) patients who were on progressive courses (primary or secondary) at treatment start, and (iii) patients with Expanded Disability Status Scale (EDSS) >4 at treatment start, given that these patients were likely in secondary progressive phase and thus different from the rest of the cohort. The resulting values from harmonization phase were generated in longitudinal format, where each row corresponded to one treatment exposure for the patient. If a patient had multiple exposures to NTZ, the first observation with exposure >12 months was considered eligible. If none of the observations had exposure larger than this threshold, the observation with the largest exposure was considered independently of the order. In case of a previous short exposure to NTZ, the patient was included in the analysis only if the interval between the two exposures was longer than 12 months, to avoid any reactivation/rebound activity during the second NTZ treatment related to the withdrawal of the previous exposure. In the same way, patients previously treated with fingolimod were not included in the analysis, unless the interval between fingolimod withdrawal and NTZ start was longer than 12 months. The study was approved by the local ethical committees. We assessed response to NTZ with a dichotomous outcome, designating as responders patients who were relapse-free in the 4 years follow-up and non-responders those who experienced at least one relapse. Relapses were defined as new symptoms or exacerbation of existing symptoms persisting for ≥24 h, in the absence of concurrent illness/fever, and occurring ≥30 days after a previous relapse. Our study cohort was derived as a subset of a larger multi-centric dataset, that constituted the replication cohort for study on MS severity, performed in the context of the International Multiple Sclerosis Genetic Consortium (IMSGC). Details on pre-imputation quality control, phasing and imputation steps are thus described therein . Quality control steps on imputed data were performed within data from each center. For the two centers that had multiple distinct genotyping platforms (Italy and Sweden), we performed post-imputation quality control after obtaining a unique merged dataset. On a ‘per-marker’ basis, we excluded variants that: (1) had a call rate less than 95%; (2) had minor allele frequency (MAF) below 5%; (3) deviated from Hardy–Weinberg Equilibrium exact test at p < 10 -5 . On a ‘per-individual’ basis we excluded subjects who had high rates of genotype missingness (>5%) and one member of each pair of samples that showed, across platforms and within centers, high degree of recent shared ancestry (up to the second degree of kinship) inferred by robust estimation of their kinship coefficient . We finally used Principal Component Analysis (PCA) pruning from the data variants with a call rate less than 99% and regions of extended linkage disequilibrium, to control for population stratification and to discard individuals with outlying values in ancestry. We considered as outliers those samples being more than 4 standard deviations away from the mean of the first two PCs. All the quality control steps were performed with PLINK 2.0 . The workflow of the study is depicted in Fig. . We performed single-SNP association analysis fitting logistic regression models as implemented in PLINK 2.0, assuming additive effects of imputed continuous dosages of minor alleles. Models were adjusted for age and disease duration at treatment start, sex, the number of relapses in the 2 years preceding NTZ therapy, and the first five eigenvectors from PCA to account for population substructure. Summary statistics were aggregated using fixed-effect meta-analysis with inverse-variance weighting of log(odds-ratios), as implemented in PLINK 2.0. Variants were annotated with ANNOVAR and visualization of the top-associated locus was generated via regional plot with LocusZoom . Gene-based analysis was conducted by means of Multi-marker Analysis of GenoMic Annotation (MAGMA, ) method v1.10, adjusting for the same set of covariates. The tool accounts for linkage disequilibrium and confounders like gene size and density. We used the multi option, which combines evidence from three models (principal components regression, mean of SNP squared Z-scores, and top SNP association). A critical choice in gene-based and gene-set analysis is the assignment of SNPs to genes, since inclusion of noisy variants can be detrimental to association analysis. We assigned SNPs to the target gene with a “proximity rule” using a flanking window of 5 kb, to minimize overlap between nearby genes. Second, we applied a “functional rule” by: integrating in target genes cis-eQTL SNPs, based on significant SNP–gene associations in immune cells (FDR < 5%), as available in DICE repository , that identified common genetic variants that are associated with the expression of > 12,000 genes in 13 human immune cell types. integrating in target genes those variants significantly affecting splicing regulation (FDR<5%), using a catalogue of cis-splicing QTLs (sQTL, ) processed on transcriptome data in blood tissue from the Genotype-Tissue Expression Consortium . Gene-wise statistics were then meta-analysed with weighted Stouffer’s procedure, which combines the Z-scores for each strata with weights set to the square root of the sample size. Approximately independent signals were identified upon application of a clumping procedure (primary p1 < 5 × 10 -5 , secondary p 2 < 0.01, r 2 > 0.6, maximum distance = 250 kb). We used the gene-wise meta-analytic p values as input for gene-set analysis of association using as reference Gene Ontology (GO) Biological Processes, retrieved from Human Molecular Signature Database (MSigDB v2023.1, ). We filtered out GO terms with less than 10 and more than 500 annotated genes, finally testing 5376 gene-sets. Each GO term was tested under the competitive null hypothesis, which states that aggregated variation in genes annotated to the gene set is no more associated with the outcome than that in all other genes in the genome. To accomplish this, we used as background signal the whole set of genes used in meta-analysis ( n = 24,110). We performed a subnetwork detection analysis, projecting meta-analysed statistics onto STRING v11.5 reference interactome . We retained only those links with high interaction evidence (score > 0.7) in one of the three evidence domains: (i) protein–protein interaction, derived from multiple interactomes, such as IntAct, BioGrid, MINT, and others; (ii) co-expression, which leverages gene expression data from multiple sources; (iii) databases, which collects evidence of interaction from curated pathway resources. We then used dmGWAS tool , which applies a greedy search algorithm of dense modules within the node-weighted interactome, to detect association signals that aggregate in subnetworks. This procedure scores each module by a Z-score corresponding to the association level of the gene: the module score is obtained dividing the sum of the nodes scores by the square root of each module size. Starting from each seed, the procedure examines first-order neighbours and identifies those that generate the maximum increment of module score. We selected the top 1% of the top-scoring modules and merged them in a final subnetwork. The top-scoring subnetwork was imported into Cytoscape v3.8 environment for visualization, manipulation, and extraction of topologically relevant nodes (hubs, bottlenecks) with CentiScaPe plugin . We computed distributions of graph centrality metrics like degree, betweenness and eigenvector centrality for the detected module and selected nodes residing in the top 5% of at least one of the four distributions: these metrics should measure the functional importance of genes in the module. ClusterProfiler R package was used to perform gene-set over-representation analysis with hypergeometric test of the genes annotated to the extracted module and the detected communities, using GO Biological Process domain as reference database; the list of genes from meta-analysis that were present on the filtered STRING interactome was used as background universe. Benjamini–Hochberg adjusted p values < 0.05 were used to nominate significant gene-sets. Association analysis Clinico-demographic variables are reported for the three cohorts of the study population in Table . It can be observed that the sample size of the Swedish cohort (SWE) was much larger ( N = 1634) as compared to the other two cohorts [ N = 119 and N = 81 from Italy (ITA) and Germany (GER), respectively]. While there was some degree of heterogeneity at baseline variables, the proportion of R/NR did not significantly differ across the three cohorts (Table ). After quality control, a total of ~ 4.7 M variants were retained for downstream association analyses (SWE: 4768680, ITA: 4612675, and GER: 4716021). Fixed-effect inverse-variance weighted meta-analysis of effect sizes across the three strata was finally performed on a set of 4747971 variants shared between at least two of the three cohorts. QQ plot reported in Supplementary Fig. 1 did not show effect of genomic inflation due to population substructure. No variants showed significant association at genome-wide level ( p < 5 × 10 −8 ). Overall pattern of association is reported on the Manhattan plot in Fig. , whereas independent clumped variants associated at suggestive level ( p < 5 × 10 –5 ) are reported in Table . The most significant signal of association was detected at rs11132400 T , located in the intronic region of the F11-AS1 gene, coding for an antisense RNA, on chromosome 4 ( p = 1.3 × 10 –6 , OR = 0.58, Fig. ). For this variant, at least nominally significant eQTL effects in multiple tissues were detected using QTLbase : In particular, eQTLs were identified in Induced Pluripotent Stem Cell for genes KLKB1 , CYP4V2 , F11 and in Blood-Macrophage for FAT1 gene . Other top-associated variants were rs12885261 T , located in the intergenic region between genes PIGH and ARG2 on chromosome 14 ( p = 1.67 × 10 –6 , OR = 1.53) and rs1323374 T , located in the intergenic region between genes KLF4 and ACTL7B ( p = 2.79 × 10 –6 , OR = 0.59) (Fig. ). In QTLbase, rs12885261 T was found to exert eQTL effect in blood B cells on ARG2 gene (T allele, beta = 0.31, p = 7.45 × 10 –8 ). The three mentioned variants exhibited an I 2 heterogeneity index that was low (0% or 12.9%), reflecting concordance in effect sizes across the three cohorts, which can be observed in Supplementary Table 1, reporting the single-stratum effects of the SNPs in Table . We further report in Supplementary Table 2 the top ten variants with highest and lowest meta-analytic odds ratios. Upon assignment of SNPs to genes according to proximity and function rules, a total of 24,110 autosomal genes had at least one assigned variant and 2,536,188 variants (63%) mapped for proximity, eQTL or sQTL effect to at least one gene. The complete list of meta-analysed genes from gene-based associations at p < 0.05 is reported in Supplementary Table 3. In addition, we used gene-based statistics to test association for two genes, NQO1 and GSTP1 on chromosome 16 and 11, respectively, encoding for detoxification enzymes, whose nonsynonymous polymorphisms have been identified as associated with the response to NTZ in a candidate study . Although we could not replicate findings at single-variant level for the two polymorphisms (rs1800566 in NQO1 and rs1695 in GSTP1 ), our data from gene-based meta-analysis revealed nominal association for both genes ( p NQO1 = 0.056, p GSTP1 = 0.029), indicating a possible role for them in the response to NTZ. Gene-set analysis Gene-set analysis from meta-analysed genes under the competitive hypothesis did not yield significant results after multiple testing correction. Nevertheless, several GO terms that point to immune-related processes, in particular T helper cell differentiation, in response to NTZ were observed, like Regulation of CD4 positive alpha beta T-cell differentiation ( p = 0.0009), T helper 17 cell lineage commitment ( p = 0.0033749), Positive regulation of adaptive immune response ( p = 0.00399), Regulation of alpha beta T-cell differentiation ( p = 0.00435), Negative regulation of type 2 immune response ( p = 0.0079), and Regulation of T helper cell differentiation ( p = 0.00848). The complete set of nominally associated GO terms is reported in Supplementary Table 4. Network analysis We searched for subnetworks with enriched genetic signals of response to NTZ, conducting dense module searching on the STRING high-confidence reference interactome, which consisted of 10,698 nodes, matched to meta-analysed genes, connected by 121,565 edges. The nodes were weighted by gene-based summary statistics from GWAS meta-analysis ( z -score), aggregating the scores at the module level (see Methods). The algorithm identified 6,766 modules, and we prioritized those residing in the top-1% of the distribution of z -scores ( N = 68), which exhibited extensive overlap. The minimum and maximum size of the top-ranking modules was 6 and 10 nodes, and the largest connected component obtained by merging them was a subnetwork of 135 nodes and 290 edges. The top-associated genes in the module were TH ( p = 1.31 × 10 –3 ) and SP100 ( p = 1.33 × 10 –3 ): by construction, not all genes which are part of the merged module were associated with response to NTZ: nevertheless, there were many of them which are directly or indirectly connected with associated genes (Fig. ). From the detected module, we produced the most topologically important nodes prioritizing genes with values in the top 5% of the distribution of three node centrality metrics (Supplementary Table 5). The top-ranked genes like PPP2CB and PPP4R2 , encode for part of protein complexes and thus shared high values of graph centrality metrics (degree, betweenness, and eigenvector centrality), Among other genes with topological relevance that were also significantly associated with response to NTZ, we identified two genes like LRP6 ( p = 0.045) and GRB2 ( p = 0.023) which have been already implicated in MS (see Discussion). Regional plots illustrating the overall pattern of association for the two genes are reported in Supplementary Fig. 2. Functional gene-set over-representation analysis was performed to yield possible biological mechanisms of module interacting genes. From GO BP terms, 75 terms were significant at FDR < 5% (Fig. a). Many of these terms were semantically related due to GO hierarchical structure. Among them, according to semantic similarity estimated with Jaccard similarity coefficient, four themes emerged (Fig. b): Canonical WNT signaling pathways ( p adjust = 7.08 × 10 –6 ); Protein dephosphorilation ( p adjust = 1.42 × 10 –3 ); mRNA stabilization ( p adjust = 0.0144); Regulation of calcium ion transmembrane activity ( p adjust = 0.0161). The complete list of over-represented GO BP terms is reported in Supplementary Table 6, together with the annotated module genes. Clinico-demographic variables are reported for the three cohorts of the study population in Table . It can be observed that the sample size of the Swedish cohort (SWE) was much larger ( N = 1634) as compared to the other two cohorts [ N = 119 and N = 81 from Italy (ITA) and Germany (GER), respectively]. While there was some degree of heterogeneity at baseline variables, the proportion of R/NR did not significantly differ across the three cohorts (Table ). After quality control, a total of ~ 4.7 M variants were retained for downstream association analyses (SWE: 4768680, ITA: 4612675, and GER: 4716021). Fixed-effect inverse-variance weighted meta-analysis of effect sizes across the three strata was finally performed on a set of 4747971 variants shared between at least two of the three cohorts. QQ plot reported in Supplementary Fig. 1 did not show effect of genomic inflation due to population substructure. No variants showed significant association at genome-wide level ( p < 5 × 10 −8 ). Overall pattern of association is reported on the Manhattan plot in Fig. , whereas independent clumped variants associated at suggestive level ( p < 5 × 10 –5 ) are reported in Table . The most significant signal of association was detected at rs11132400 T , located in the intronic region of the F11-AS1 gene, coding for an antisense RNA, on chromosome 4 ( p = 1.3 × 10 –6 , OR = 0.58, Fig. ). For this variant, at least nominally significant eQTL effects in multiple tissues were detected using QTLbase : In particular, eQTLs were identified in Induced Pluripotent Stem Cell for genes KLKB1 , CYP4V2 , F11 and in Blood-Macrophage for FAT1 gene . Other top-associated variants were rs12885261 T , located in the intergenic region between genes PIGH and ARG2 on chromosome 14 ( p = 1.67 × 10 –6 , OR = 1.53) and rs1323374 T , located in the intergenic region between genes KLF4 and ACTL7B ( p = 2.79 × 10 –6 , OR = 0.59) (Fig. ). In QTLbase, rs12885261 T was found to exert eQTL effect in blood B cells on ARG2 gene (T allele, beta = 0.31, p = 7.45 × 10 –8 ). The three mentioned variants exhibited an I 2 heterogeneity index that was low (0% or 12.9%), reflecting concordance in effect sizes across the three cohorts, which can be observed in Supplementary Table 1, reporting the single-stratum effects of the SNPs in Table . We further report in Supplementary Table 2 the top ten variants with highest and lowest meta-analytic odds ratios. Upon assignment of SNPs to genes according to proximity and function rules, a total of 24,110 autosomal genes had at least one assigned variant and 2,536,188 variants (63%) mapped for proximity, eQTL or sQTL effect to at least one gene. The complete list of meta-analysed genes from gene-based associations at p < 0.05 is reported in Supplementary Table 3. In addition, we used gene-based statistics to test association for two genes, NQO1 and GSTP1 on chromosome 16 and 11, respectively, encoding for detoxification enzymes, whose nonsynonymous polymorphisms have been identified as associated with the response to NTZ in a candidate study . Although we could not replicate findings at single-variant level for the two polymorphisms (rs1800566 in NQO1 and rs1695 in GSTP1 ), our data from gene-based meta-analysis revealed nominal association for both genes ( p NQO1 = 0.056, p GSTP1 = 0.029), indicating a possible role for them in the response to NTZ. Gene-set analysis from meta-analysed genes under the competitive hypothesis did not yield significant results after multiple testing correction. Nevertheless, several GO terms that point to immune-related processes, in particular T helper cell differentiation, in response to NTZ were observed, like Regulation of CD4 positive alpha beta T-cell differentiation ( p = 0.0009), T helper 17 cell lineage commitment ( p = 0.0033749), Positive regulation of adaptive immune response ( p = 0.00399), Regulation of alpha beta T-cell differentiation ( p = 0.00435), Negative regulation of type 2 immune response ( p = 0.0079), and Regulation of T helper cell differentiation ( p = 0.00848). The complete set of nominally associated GO terms is reported in Supplementary Table 4. We searched for subnetworks with enriched genetic signals of response to NTZ, conducting dense module searching on the STRING high-confidence reference interactome, which consisted of 10,698 nodes, matched to meta-analysed genes, connected by 121,565 edges. The nodes were weighted by gene-based summary statistics from GWAS meta-analysis ( z -score), aggregating the scores at the module level (see Methods). The algorithm identified 6,766 modules, and we prioritized those residing in the top-1% of the distribution of z -scores ( N = 68), which exhibited extensive overlap. The minimum and maximum size of the top-ranking modules was 6 and 10 nodes, and the largest connected component obtained by merging them was a subnetwork of 135 nodes and 290 edges. The top-associated genes in the module were TH ( p = 1.31 × 10 –3 ) and SP100 ( p = 1.33 × 10 –3 ): by construction, not all genes which are part of the merged module were associated with response to NTZ: nevertheless, there were many of them which are directly or indirectly connected with associated genes (Fig. ). From the detected module, we produced the most topologically important nodes prioritizing genes with values in the top 5% of the distribution of three node centrality metrics (Supplementary Table 5). The top-ranked genes like PPP2CB and PPP4R2 , encode for part of protein complexes and thus shared high values of graph centrality metrics (degree, betweenness, and eigenvector centrality), Among other genes with topological relevance that were also significantly associated with response to NTZ, we identified two genes like LRP6 ( p = 0.045) and GRB2 ( p = 0.023) which have been already implicated in MS (see Discussion). Regional plots illustrating the overall pattern of association for the two genes are reported in Supplementary Fig. 2. Functional gene-set over-representation analysis was performed to yield possible biological mechanisms of module interacting genes. From GO BP terms, 75 terms were significant at FDR < 5% (Fig. a). Many of these terms were semantically related due to GO hierarchical structure. Among them, according to semantic similarity estimated with Jaccard similarity coefficient, four themes emerged (Fig. b): Canonical WNT signaling pathways ( p adjust = 7.08 × 10 –6 ); Protein dephosphorilation ( p adjust = 1.42 × 10 –3 ); mRNA stabilization ( p adjust = 0.0144); Regulation of calcium ion transmembrane activity ( p adjust = 0.0161). The complete list of over-represented GO BP terms is reported in Supplementary Table 6, together with the annotated module genes. Identification of genetic markers, together with other biomarkers, that associate with response to DMTs is a crucial clinical need for MS patients’ stratification and their tailored management. In the case of a highly effective treatment such as Natalizumab, to date, only a few candidate gene studies have been performed to elicit such markers , mainly due to reduced sample size caused by the relatively low number of non-responders to the drug. Here, we conducted a multi-centric GWAS of response to NTZ, to our knowledge the largest in pharmacogenomics of this DMT, that we pursued at multiple analytical levels. Our study could not identify any locus at genome-wide significance: nevertheless, the top-associated SNP rs11132400 T , an intronic variant in F11-AS1 gene , was found to have eQTL effects on multiple genes with biological plausibility, such as KLKB1 , F11 , and FAT1 . The gene KLKB1 encodes prekallikrein, a protein which modulates the integrity of BBB, whereas F11 encodes the coagulation factor XI. Notably, both proteins can act as important mediators of the adaptive immune response during neuroinflammation. Specifically, in the contact activation pathway, three proenzymes in blood (plasma factor XII “FXII”, factor XI “FXI”, prekallikrein “PK”, and high-molecular-weight kininogen “HK”) bind to a surface and cause blood coagulation and inflammation by activating their respective enzymes (FXIIa, FXIa and α-kallikrein). Several lines of evidence show that F11 and KLKB1 are also implied in MS aetiology. Indeed, targeting of factor FXI improves neurological function and attenuates CNS damage in Experimental Autoimmune Encephalomyelitis (EAE), the animal model of MS . Moreover, a deficiency of plasma prekallikrein, the precursor of kallikrein which is found to be upregulated in EAE, leads to decreased immune cell trafficking in the course of neuroinflammation rendering mice less susceptible to the disease . As of FAT1 gene, its product functions as an adhesion molecule and as signaling receptor, and its importance in developmental processes and cell communication is well assessed. Several lines of evidence show that FAT1 activates a variety of signaling pathways through protein–protein interactions, including the Wnt/β-catenin and MAPK/ERK signaling pathways, which affect cell proliferation, migration, and invasion . Interestingly, other intronic variants in F11-AS1 have been identified as associated with neuroimaging measurements, such as brain morphology, subcortical volume, cortical surface area, and cortical thickness . The second most associated variant from our meta-analysis (rs12885261 T ) exerted eQTL effect on ARG2 gene, with subjects carrying T allele having higher expression level of the gene, as of QTLbase resource. ARG2 encodes for an enzyme ubiquitously expressed at low level within the mitochondria, having arginine as substrate. This arginase isoform appears to play important roles in regulation of inflammation and pathogenesis of immune-mediated diseases, thus inducing changes in intracellular levels of arginine, whose metabolism is a critical regulator of innate and adaptive immune responses . A recent study showed the beneficial effect of ARG2 deletion in suppressing retinal neurodegeneration and inflammation in an experimental model of MS . In another study, there was evidence of a significant reduction of Th17 cells and IL-23 + cells in relevant draining lymph nodes associated with Arg II knockout in murine model . This is in line with our findings, which show that patients carrying T allele, possibly having higher transcriptional level of ARG2 , also have a higher risk of relapsing and being non-responders to NTZ. Given the increasing awareness of the importance of molecular interactions in shaping complex traits , we then integrated human interactome data with our gene-based association statistics using a module search algorithm, with further investigation of topological properties of nodes/genes. The network-based approach, as a complementary strategy, can enhance understanding of molecular mechanisms: the module detected from overlapping our meta-analyzed statistics onto STRING interactome pointed to multiple enriched GO terms. Among these, we found a significant enrichment of terms semantically related to Wnt/β-catenin signaling. It is known that this pathway plays important roles in oligodendrocyte development and myelin formation and its dysregulation may hamper BBB formation. Once the barrier is fully formed, this pathway is also essential to maintain its properties in the adult CNS. Furthermore, it was found that inducible inhibition of this pathway in endothelial cells resulted in clinically exacerbated EAE, thus suggesting that reactivation of Wnt/β-catenin signaling might be beneficial to limit BBB leakage and immune cell infiltration into the CNS . The network approach could then also highlight key players involved in response to treatment: central genes in the network, which would go undetected due to their milder association level, can in fact gain relevance because of their sharing many functional links with other response-associated genes. We focused our attention on two genes, GRB2 and LRP6 , which were topologically relevant nodes within our detected module, while being also associated with response and that already showed prior evidence of association with MS from multiple studies. The LRP6 gene (lipoprotein receptor-related protein 6) encodes a transmembrane cell surface protein. It plays a key role in the Wnt/β-catenin signaling pathway, being a member of the transmembrane receptor complex to which the Wnt ligand binds, allowing cytosolic β-catenin accumulation and translocation to the nucleus through transcription and regulations of target genes. β-catenin mediates negative effect on differentiation of oligodendrocytes progenitor cells, thus affecting the process of myelin sheath formation: this was confirmed by experimental studies in which expression levels of LRP6 were markedly increased in RRMS patients and in cuprizone-induced demyelination mice . Furthermore, additional evidence inferred from murine model indicates that enhanced β-catenin expression in T cells leads to aberrant and Th1-biased T-cell activation, infiltration of activated T cells into the spinal cord, and enhanced expression of integrin α4β1 through regulation of Itgb1 and Itga4 genes that encode for α4β1/VLA-4 subunits β1 and α4, α4β1/VLA-4 being one of the main two targets of Natalizumab, preventing migration of autoreactive leukocytes through the blood-brain barrier and preventing inflammation . The GRB2 gene, which encodes for the growth factor receptor bound protein, was found associated in the gene-based meta-analysis and detected as a central gene in the top-ranking module (top 5% percentile in betweenness). The gene is ubiquitously expressed and encodes an adaptor protein, which facilitates the formation of complexes to integrate signals from a wide array of binding partners to inner signaling pathways . The gene acts as a modifier of Wnt/β-catenin signaling, synergizing with multiple components of this pathway, including LRP6, to amplify β-catenin dependent transcription. Both in silico and in vivo evidence demonstrate that GRB2 operates either downstream of, or in parallel with, β-catenin to drive LEF/TCF-mediated transcription of specific genes, including ITGB1 and ITGA4 . GRB2 itself acts downstream of external growth factor receptors and integrins thus providing a way for cells to fine-tune Wnt/β-catenin signaling depending on the extracellular context . Moreover, in mouse, Grb2 -deficient T cells are impaired in their development and maturation and were found to favor the induction of EAE . Notably, the gene has already been reported as one of the most topologically relevant genes in another network-based study, which jointly investigated two MS GWAS susceptibility cohorts . Further, the intronic variant rs9900529 in GRB2 was one of the 200 non-MHC loci identified in the to date largest multi-centric study of MS genetic risk from the IMSGC and it has been identified as associated with the response to interferon-beta in MS . More generally, we did not identify significant association of the 200 non-MHC loci from IMSGC study of susceptibility, after Bonferroni correction , nor of the genome-wide significant variant rs10191329 in the DYSF–ZNF638 locus, emerged from the IMSGC study on progression (data not shown). There are some limitations that must be acknowledged regarding our study. The first concerns the fact that it is under-powered for a genome-wide scan. This is particularly true for the two smaller Italian and German cohort, for which effect estimates exhibited as expected high standard errors with wide confidence intervals. This of course impacted in the fixed-effects meta-analysis, in which contribution of estimates from the two smaller cohorts was down-weighted given their lower precision. We sought to partially mitigate this issue by complementing GWAS with pathway and network level of analysis. In doing so, given the importance of regulatory information demonstrated by the enrichment of GWAS signals in eQTL loci , we also tried to boost signals integrating with SNP–gene assignment information derived from robustly established eQTLs and sQTLs from tissues that are relevant for MS. Another limitation, which is typical of network-based studies, refers to the fragmented interactome information, since the current knowledge of protein and gene interactions is incomplete and static. We decided, however, to only retain high-confidence links, drawn from the most reliable sources of evidence of STRING repository, such as PPI, co-expression, and functional databases. Finally, we are aware of the limited sensitivity of relapses, compared to MRI parameters, for the assessment of response to Natalizumab. We considered relapses as clinical outcome of response to maximize the number of patients that could be included in the study. To increase the chance for detecting clinical relapses, we used a period of observation up to 4 years, to obtain data on a medium-term follow-up. In conclusion, by investigating a multi-centric cohort of MS patients treated with NTZ, we were able to highlight a variant with a putative role in response to drug, rs11132400, and two genes already implicated in MS pathogenesis, GRB2 and LRP6 . In addition, from the network module perspective, we report an enrichment of Wnt/β-catenin signaling pathway, which is an essential component for BBB formation and maintenance. A replication study of these findings in an independent cohort would be desirable to support future clinical applications. Below is the link to the electronic supplementary material. Supplementary file1 (PDF 280 KB) Supplementary file2 (PDF 161 KB) Supplementary file3 (XLS 17 KB) Supplementary file4 (XLS 13 KB) Supplementary file5 (XLS 163 KB) Supplementary file6 (XLS 12 KB) Supplementary file7 (XLS 10 KB) Supplementary file8 (XLS 25 KB) |
Calibration of confidence and assessed clinical skills competence in undergraduate paediatric OSCE scenarios: a mixed methods study | c05c6882-ffc6-41ae-9bd5-76b45967a398 | 6142704 | Pediatrics[mh] | The relationship between a practitioner’s confidence in performing a skill and subsequent performance is intriguing and important. Confidence in performing a task influences the willingness and ultimate decision to undertake the task . However, if the confidence does not match the subsequent performance, the practitioner may fall short and potentially risk an adverse incident. Moreover, level of experience correlates well with confidence but not with assessed performance . A recent study of medical student prescribing skills found a poor correlation between reported confidence and actual competence and concluded that students lack insight into their strengths and weaknesses in prescribing . It is therefore important that confidence and competence are not used synonymously . Tomorrow’s Doctors (2009) has been used as the roadmap for undergraduate medical training in the United Kingdom . This General Medical Council (GMC) document outlines the expected competencies of a newly qualified doctor. Among these specific competencies are drug and intravenous fluid (IVF) prescribing and plotting of growth charts. These have been completed poorly in Queens University Belfast (QUB) undergraduate paediatric Objective Structured Clinical Examinations (OSCEs) over the last five years (personal communication with QUB examiners). In the Undergraduate medical course at QUB students start adult clinical skills training in their second year. Their exposure to paediatric medicine does not occur until the ‘Healthcare of Children’ module which is a six week course in the fourth year of their degree (the module runs six times through the academic year with a total of approximately two hundred and forty students per year). A week of core teaching (lectures, practical sessions and online learning) precedes a five week clinical attachment in a paediatric ward. Learning outcomes are mapped to a paediatric syllabus. The syllabus lists clinical skills that students are expected to have completed by the end of the module. Students have access to online lectures on fluid prescribing and completion of growth charts. It is a mandatory requirement that all students complete the online British Medical Journal (BMJ) module on fluid prescribing . During the clinical attachment students also attend a drug prescribing workshop where they practice completing drug prescribing records. Due to their limited exposure to paediatrics compared with adult medicine, student confidence in the performance of paediatric clinical skills may be expected to be at a low level. The relationship between medical student confidence and competence in the performance of core paediatric clinical skills has not previously been explored. The aim of this study is to assess undergraduate paediatric student confidence in performing three common and important clinical paediatric skills, compare this with actual performance and determine any barriers to successful skill completion.
Study design A mixed methods study of fourth year medical students was conducted. Medical student pre-test confidence in completing common paediatric skills was determined using a questionnaire. This was compared with skills competence using a skills assessment. Focus groups were conducted to explore the relationship between these levels of confidence and assessed competence. Setting A convenience sample of Fourth year medical students at QUB in the first two modules of the academic year 2014–15 were invited to participate in the study in seven paediatric units during the final week of the ‘Healthcare of Children’ module. Students were informed that this was not part of the module and was voluntary, anonymous and inconsequential. Written consent from the students was obtained. Ethical approval was granted by the QUB School of Medicine and Dentistry Research Ethics Committee. Quantitative methods Each student completed a two-part comparative evaluation which included a paediatric skills questionnaire and a subsequent assessment of performed skills. The questionnaire was constructed by the researchers to determine levels of confidence in performing mandatory skills listed in the syllabus. These skills need to have been completed for successful completion of the module. On a four point Likert scale (1 = Excellent, 2 = Good, 3 = Pass 4 = Fail) students documented their anticipated skills performance. The skills included ‘prescribing common paediatric medication’, ‘prescribing paediatric IVF’ and ‘plotting growth parameters on an appropriate growth chart’. The questions were reviewed and refined by an iterative process involving all the researchers ensuring content validity. Skills-based scenarios were developed by the researchers based on previously standardised paediatric OSCE stations. These had been standard-set using the Angoff method. The scenarios assessed the three skills from the questionnaire - medication prescribing, plotting of growth parameters on growth charts and IVF prescribing. They were completed on the regional Paediatric Medicine Prescription and Administration Record, the UK-WHO Growth Chart 0–4 years and the regional Paediatric Daily Fluid Prescription and Balance Chart . Students were awarded a global score based on QUB OSCE assessment guidance . This assessment stage was initiated after all survey forms had been collected. Students were given ten minutes to complete each stage and were unable to modify any completed documentation after submission. All charts were original and numbered in advance for anonymised identification. Students were provided with a British National Formulary for Children (BNFC) to assist prescribing . The skills were assessed by two examiners (GD and LH) using the modified global scoring system as being: 1 = Excellent, 2 = Good, 3 = Pass 4 = Fail . Data were analysed using IBM SPSS Statistics version 20.0 (SPSS Inc. Armonk, New York). The Internal consistency of the questionnaire was evaluated using the Crohnbach’s alpha coefficient and the Kuder- Richardson formula 20 (KR 20 ) was used to determine internal consistency in relation to passing or failing the skills assessments. Paired confidence and assessed skills competence results were compared using Wilcoxon signed ranks exact probability tests and agreement between confidence and assessed competence was analysed using the linear- weighted Kappa test. The kappa statistic is a measure of agreement standardised to lie between − 1 and 1 where 1 indicates perfect agreement, 0 is what would be expected by chance and − 1 indicates potential disagreement. Statistical significance was assumed for p values < 0.05. Qualitative methods Focus groups were used to gain a deeper understanding of student skills performance. An e-mail inviting participation was sent to students who had completed the skills assessment. Written consent was obtained from volunteers. The focus groups were moderated (DOD) to encourage guided discussion. To facilitate moderation, key themes identified from the skills assessment were used to create a focus group guide. These themes were extracted following independent review of the skills data by three authors (DOD, AT and BM). Focus group discussions were recorded with two devices and transcribed verbatim. The authors reviewed the transcripts independently and identified preliminary codes. A framework analysis technique was used to identify emerging themes for each skill . The number of focus groups to be held was determined by the stage at which this iterative process had reached thematic saturation with a minimum number of two focus groups.
A mixed methods study of fourth year medical students was conducted. Medical student pre-test confidence in completing common paediatric skills was determined using a questionnaire. This was compared with skills competence using a skills assessment. Focus groups were conducted to explore the relationship between these levels of confidence and assessed competence.
A convenience sample of Fourth year medical students at QUB in the first two modules of the academic year 2014–15 were invited to participate in the study in seven paediatric units during the final week of the ‘Healthcare of Children’ module. Students were informed that this was not part of the module and was voluntary, anonymous and inconsequential. Written consent from the students was obtained. Ethical approval was granted by the QUB School of Medicine and Dentistry Research Ethics Committee.
Each student completed a two-part comparative evaluation which included a paediatric skills questionnaire and a subsequent assessment of performed skills. The questionnaire was constructed by the researchers to determine levels of confidence in performing mandatory skills listed in the syllabus. These skills need to have been completed for successful completion of the module. On a four point Likert scale (1 = Excellent, 2 = Good, 3 = Pass 4 = Fail) students documented their anticipated skills performance. The skills included ‘prescribing common paediatric medication’, ‘prescribing paediatric IVF’ and ‘plotting growth parameters on an appropriate growth chart’. The questions were reviewed and refined by an iterative process involving all the researchers ensuring content validity. Skills-based scenarios were developed by the researchers based on previously standardised paediatric OSCE stations. These had been standard-set using the Angoff method. The scenarios assessed the three skills from the questionnaire - medication prescribing, plotting of growth parameters on growth charts and IVF prescribing. They were completed on the regional Paediatric Medicine Prescription and Administration Record, the UK-WHO Growth Chart 0–4 years and the regional Paediatric Daily Fluid Prescription and Balance Chart . Students were awarded a global score based on QUB OSCE assessment guidance . This assessment stage was initiated after all survey forms had been collected. Students were given ten minutes to complete each stage and were unable to modify any completed documentation after submission. All charts were original and numbered in advance for anonymised identification. Students were provided with a British National Formulary for Children (BNFC) to assist prescribing . The skills were assessed by two examiners (GD and LH) using the modified global scoring system as being: 1 = Excellent, 2 = Good, 3 = Pass 4 = Fail . Data were analysed using IBM SPSS Statistics version 20.0 (SPSS Inc. Armonk, New York). The Internal consistency of the questionnaire was evaluated using the Crohnbach’s alpha coefficient and the Kuder- Richardson formula 20 (KR 20 ) was used to determine internal consistency in relation to passing or failing the skills assessments. Paired confidence and assessed skills competence results were compared using Wilcoxon signed ranks exact probability tests and agreement between confidence and assessed competence was analysed using the linear- weighted Kappa test. The kappa statistic is a measure of agreement standardised to lie between − 1 and 1 where 1 indicates perfect agreement, 0 is what would be expected by chance and − 1 indicates potential disagreement. Statistical significance was assumed for p values < 0.05.
Focus groups were used to gain a deeper understanding of student skills performance. An e-mail inviting participation was sent to students who had completed the skills assessment. Written consent was obtained from volunteers. The focus groups were moderated (DOD) to encourage guided discussion. To facilitate moderation, key themes identified from the skills assessment were used to create a focus group guide. These themes were extracted following independent review of the skills data by three authors (DOD, AT and BM). Focus group discussions were recorded with two devices and transcribed verbatim. The authors reviewed the transcripts independently and identified preliminary codes. A framework analysis technique was used to identify emerging themes for each skill . The number of focus groups to be held was determined by the stage at which this iterative process had reached thematic saturation with a minimum number of two focus groups.
The student consent rate to participate in the questionnaire and skills assessment study was 100% (85 students out of 85). All students completed both the questionnaire and three skills assessments. The Crohnbach’s alpha coefficient for the questionnaire data was 0.81 suggesting high levels of internal consistency. The skills assessments also demonstrate high levels of internal consistency (KR 20 values for drug prescribing, growth chart plotting, and IV fluid prescribing were 0.70, 0.89 and 0.89 respectively). The cross-tabulations for student confidence ratings and skills assessment scores for each skill are shown in Table . For each of the assessed skills tasks, there was little or no agreement between pre-test confidence and assessed competence (Tables and ). The pre-test skills confidence and assessed skills assessment scores are represented graphically in Fig. and Fig. respectively. The students’ confidence in their ability to successfully complete the task was not matched by their task competence with students believing that they would perform to a higher level than they actually did (Wilcoxon signed ranks exact probability tests for paired comparisons of confidence and competence for each of the three skills - p -value < 0.001). Focus groups Twenty six students volunteered to take part in focus groups and students were assigned to focus groups in the order that they volunteered. There were six students in each focus group and thematic saturation occurred following completion of three focus groups. Analysis of focus group transcripts identified five distinct overarching themes. The themes were similar for the three tasks (Table ). Five themes were identified and a summary of sub-themes with illustrative student quotes within each is provided below: Conflicting teaching This theme only arose for fluid prescribing. Teaching on IVF prescribing varied depending on which hospital the student had been based in. The advice on when and when not to prescribe potassium in maintenance fluids created the most difficulty: ‘In Hospital A I had been taught not to prescribe potassium chloride for maintenance fluids without an electrolyte profile being available but now I realise this is completely different to what students has been told in Hospital B.’ Prescribing potassium in IV fluids for a child whose blood electrolyte profile is not known is potentially dangerous as the potassium levels can rise and potentially result in an abnormality of the heart rhythm. Therefore such conflicting advice may lead to a serious adverse incident. Complexity It became apparent that students had limited experience of drug or IVF prescribing in previous adult-medicine attachments. In addition, students felt that incorporation of a weight calculation in paediatric drug and IVF prescribing significantly increased the task complexity compared to completion of the same task in adult medicine. Many students described how the completion of these paediatric skills was more challenging because of this. Correcting for prematurity in plotting growth was unfamiliar and confusing to some students: ‘It’s ok plotting a baby’s height and weight but plotting a premature baby is confusing because I don’t know whether to correct to 38 weeks or 40 weeks. Correctly plotting and tracking a premature baby’s growth parameters accurately by taking account of the degree of prematurity (rather than simply plotting their chronological age) is extremely important in deciding whether a baby is achieving appropriate weight gain and brain growth for their age. Prior experience There was variation in the experience and opportunities that students had in skills completion. All students had completed a drug prescribing workshop. They had also completed the required number of patient admissions. As part of the documentation for admitting patients to a ward students are asked to plot each patient’s height and weight on an appropriate growth chart. It is assumed that completion of a growth chart is done at this time. However students admitted that they occasionally forgot to do this. Students had viewed an online lecture on fluid prescribing and all had completed an online BMJ module on paediatric fluid prescribing but some had not previously completed an actual fluid prescription: ‘I had seen the online lecture that takes us through fluids step-by step so I was happy that I knew what to do. When I was faced with the fluid chart and had to do it then it was a different matter’. This quotation illustrates the point that, for some students, there was a clear gap between theory and application. They appeared to assume that, armed with the relevant background theoretical knowledge, the clinical skill would be straightforward. The skills assessment demonstrated that this is not necessarily the case. Prior supervised experience Students differentiated between skills experience and supervised skills experience. Many students had completed fluid prescriptions but rarely received feedback on this to allow them to calibrate their skills. Most students had plotted growth charts when admitting patients to a ward but these were not always appraised by a supervising doctor: ‘I try to complete the growth chart for every child I see as told. The doctors that I present my cases to go through the diagnosis and tests but don’t look over my notes or the photocopied growth chart I have filled in.’ The students appeared concerned that they do not always receive optimal supervision or specific feedback tailored to their learning needs. Chart familiarity A number of students had prescribed a medication correctly but in the wrong section of the drug prescribing record e.g. an oral antibiotic in the intravenous section and they appeared to be unfamiliar with navigating the prescription record. A few students also failed to complete the record of drug allergies: ‘I was focusing on getting the right dose written in the right place. I hadn’t filled in drug allergy boxes previously and I didn’t notice it when looking where to write the prescription’. It became apparent that there were slight variations in the drug prescribing charts between the paediatric units where students were based for their paediatric attachment. To standardise prescribing practice a national paediatric drug prescribing record has subsequently been developed and is now used in all paediatric units in Northern Ireland.
Twenty six students volunteered to take part in focus groups and students were assigned to focus groups in the order that they volunteered. There were six students in each focus group and thematic saturation occurred following completion of three focus groups. Analysis of focus group transcripts identified five distinct overarching themes. The themes were similar for the three tasks (Table ). Five themes were identified and a summary of sub-themes with illustrative student quotes within each is provided below: Conflicting teaching This theme only arose for fluid prescribing. Teaching on IVF prescribing varied depending on which hospital the student had been based in. The advice on when and when not to prescribe potassium in maintenance fluids created the most difficulty: ‘In Hospital A I had been taught not to prescribe potassium chloride for maintenance fluids without an electrolyte profile being available but now I realise this is completely different to what students has been told in Hospital B.’ Prescribing potassium in IV fluids for a child whose blood electrolyte profile is not known is potentially dangerous as the potassium levels can rise and potentially result in an abnormality of the heart rhythm. Therefore such conflicting advice may lead to a serious adverse incident. Complexity It became apparent that students had limited experience of drug or IVF prescribing in previous adult-medicine attachments. In addition, students felt that incorporation of a weight calculation in paediatric drug and IVF prescribing significantly increased the task complexity compared to completion of the same task in adult medicine. Many students described how the completion of these paediatric skills was more challenging because of this. Correcting for prematurity in plotting growth was unfamiliar and confusing to some students: ‘It’s ok plotting a baby’s height and weight but plotting a premature baby is confusing because I don’t know whether to correct to 38 weeks or 40 weeks. Correctly plotting and tracking a premature baby’s growth parameters accurately by taking account of the degree of prematurity (rather than simply plotting their chronological age) is extremely important in deciding whether a baby is achieving appropriate weight gain and brain growth for their age. Prior experience There was variation in the experience and opportunities that students had in skills completion. All students had completed a drug prescribing workshop. They had also completed the required number of patient admissions. As part of the documentation for admitting patients to a ward students are asked to plot each patient’s height and weight on an appropriate growth chart. It is assumed that completion of a growth chart is done at this time. However students admitted that they occasionally forgot to do this. Students had viewed an online lecture on fluid prescribing and all had completed an online BMJ module on paediatric fluid prescribing but some had not previously completed an actual fluid prescription: ‘I had seen the online lecture that takes us through fluids step-by step so I was happy that I knew what to do. When I was faced with the fluid chart and had to do it then it was a different matter’. This quotation illustrates the point that, for some students, there was a clear gap between theory and application. They appeared to assume that, armed with the relevant background theoretical knowledge, the clinical skill would be straightforward. The skills assessment demonstrated that this is not necessarily the case. Prior supervised experience Students differentiated between skills experience and supervised skills experience. Many students had completed fluid prescriptions but rarely received feedback on this to allow them to calibrate their skills. Most students had plotted growth charts when admitting patients to a ward but these were not always appraised by a supervising doctor: ‘I try to complete the growth chart for every child I see as told. The doctors that I present my cases to go through the diagnosis and tests but don’t look over my notes or the photocopied growth chart I have filled in.’ The students appeared concerned that they do not always receive optimal supervision or specific feedback tailored to their learning needs. Chart familiarity A number of students had prescribed a medication correctly but in the wrong section of the drug prescribing record e.g. an oral antibiotic in the intravenous section and they appeared to be unfamiliar with navigating the prescription record. A few students also failed to complete the record of drug allergies: ‘I was focusing on getting the right dose written in the right place. I hadn’t filled in drug allergy boxes previously and I didn’t notice it when looking where to write the prescription’. It became apparent that there were slight variations in the drug prescribing charts between the paediatric units where students were based for their paediatric attachment. To standardise prescribing practice a national paediatric drug prescribing record has subsequently been developed and is now used in all paediatric units in Northern Ireland.
This theme only arose for fluid prescribing. Teaching on IVF prescribing varied depending on which hospital the student had been based in. The advice on when and when not to prescribe potassium in maintenance fluids created the most difficulty: ‘In Hospital A I had been taught not to prescribe potassium chloride for maintenance fluids without an electrolyte profile being available but now I realise this is completely different to what students has been told in Hospital B.’ Prescribing potassium in IV fluids for a child whose blood electrolyte profile is not known is potentially dangerous as the potassium levels can rise and potentially result in an abnormality of the heart rhythm. Therefore such conflicting advice may lead to a serious adverse incident.
It became apparent that students had limited experience of drug or IVF prescribing in previous adult-medicine attachments. In addition, students felt that incorporation of a weight calculation in paediatric drug and IVF prescribing significantly increased the task complexity compared to completion of the same task in adult medicine. Many students described how the completion of these paediatric skills was more challenging because of this. Correcting for prematurity in plotting growth was unfamiliar and confusing to some students: ‘It’s ok plotting a baby’s height and weight but plotting a premature baby is confusing because I don’t know whether to correct to 38 weeks or 40 weeks. Correctly plotting and tracking a premature baby’s growth parameters accurately by taking account of the degree of prematurity (rather than simply plotting their chronological age) is extremely important in deciding whether a baby is achieving appropriate weight gain and brain growth for their age.
There was variation in the experience and opportunities that students had in skills completion. All students had completed a drug prescribing workshop. They had also completed the required number of patient admissions. As part of the documentation for admitting patients to a ward students are asked to plot each patient’s height and weight on an appropriate growth chart. It is assumed that completion of a growth chart is done at this time. However students admitted that they occasionally forgot to do this. Students had viewed an online lecture on fluid prescribing and all had completed an online BMJ module on paediatric fluid prescribing but some had not previously completed an actual fluid prescription: ‘I had seen the online lecture that takes us through fluids step-by step so I was happy that I knew what to do. When I was faced with the fluid chart and had to do it then it was a different matter’. This quotation illustrates the point that, for some students, there was a clear gap between theory and application. They appeared to assume that, armed with the relevant background theoretical knowledge, the clinical skill would be straightforward. The skills assessment demonstrated that this is not necessarily the case.
Students differentiated between skills experience and supervised skills experience. Many students had completed fluid prescriptions but rarely received feedback on this to allow them to calibrate their skills. Most students had plotted growth charts when admitting patients to a ward but these were not always appraised by a supervising doctor: ‘I try to complete the growth chart for every child I see as told. The doctors that I present my cases to go through the diagnosis and tests but don’t look over my notes or the photocopied growth chart I have filled in.’ The students appeared concerned that they do not always receive optimal supervision or specific feedback tailored to their learning needs.
A number of students had prescribed a medication correctly but in the wrong section of the drug prescribing record e.g. an oral antibiotic in the intravenous section and they appeared to be unfamiliar with navigating the prescription record. A few students also failed to complete the record of drug allergies: ‘I was focusing on getting the right dose written in the right place. I hadn’t filled in drug allergy boxes previously and I didn’t notice it when looking where to write the prescription’. It became apparent that there were slight variations in the drug prescribing charts between the paediatric units where students were based for their paediatric attachment. To standardise prescribing practice a national paediatric drug prescribing record has subsequently been developed and is now used in all paediatric units in Northern Ireland.
This study shows relatively poor skills performance for all three assessed tasks. As all students had received instruction and had experience in completing most or all of these tasks, the results are concerning. There is no agreement between confidence and subsequent performance of a paediatric skill. These results align with those of a postgraduate study of newly-qualified doctors that found poor correlation between self-reported confidence and assessed competence in the completion of seven clinical skills . The focus group data identified a number of potentially modifiable factors related to skills performance. This is the first study to explore this relationship in paediatric clinical OSCE scenarios and to triangulate confidence and competence data using qualitative methodology. The 4th year paediatric attachment is likely to be the last contact with paediatric medicine until a student becomes a Foundation Year 2 (FY2) doctor in paediatrics. Therefore it is important that this opportunity is optimised. Monitoring of growth is an essential component of childhood healthcare and the Department of Health (DOH) and Royal College of Paediatrics and Child Health (RCPCH) recommend that professionals who plot or interpret UK-WHO growth charts should receive suitable training or supervision . However, growth charts are frequently missing from clinical notes or are incomplete [ – ]. A case review of the clinical notes of fifty hospitalised infants found that nearly 30 % of points plotted on infant growth charts were plotted in error . The main sources of error identified were in plotting age and correcting for prematurity. In a further study, up to 20 % of health-care professionals were unable to interpret the UK-WHO growth charts with many receiving no formal teaching in their use . In the present study performance was better for completion of growth charts than for the prescribing tasks. This may be as a result of having relatively more supervised practice in plotting, as students are expected to complete and present fifteen patient records. A completed growth chart should be included as part of this. However this was occasionally omitted and the focus groups revealed that, even when completed, some students reported that they were not always reviewed by a supervising doctor. The focus groups also echoed the results of a previous study in finding that completion of growth charts for ex-preterm infants was particularly difficult . Correcting for prematurity was problematic and many students stated that they had not specifically done this before. In a systematic literature review of the level of clinical preparedness in acute care, United Kingdom graduates feel poorly prepared in the area of medication prescribing . Prescribing errors have been found in 8.4% and 10.3% of prescriptions completed by Foundation Year 1 and 2 doctors respectively . Despite this, much of the teaching of paediatric prescribing in the postgraduate domain has been by lectures with little assessment of competency . Interestingly, a recent review of strategies to ensure new graduates are safe prescribers suggests that prescribing teaching should focus on the development of expertise rather than competency . The development of a ‘theoretical framework of knowledge application’ advocates that prescribing should be contextualised and embedded within the workplace setting rather than being studied in isolation. It has been shown that practical prescribing courses delivered to undergraduate medical students improve postgraduate prescribing . Safe and competent prescribing should be an aim of educators charged with developing the undergraduate curriculum. Prescribing intravenous fluids is an important skill for all hospital doctors. Paediatric fluid prescribing is complicated by the physiological and anatomical differences between adults and children, thus making it a greater challenge in this population. The National Institute of Clinical Excellence (NICE) has recently produced guidelines specific to children . Prescribing the wrong type of fluid to children can result in significant morbidity and in some cases death . A recent high-profile public inquiry into childhood deaths related to hyponatraemia in Northern Ireland has highlighted the importance of fluid prescribing . Consequently the government released guidance on fluid prescribing in children . Out of the three simulated tasks, students were very confident about successfully completing this task but actually performed least well overall. It became apparent in the focus groups that some students had received conflicting teaching during their clinical attachments. This has also been shown in a recent study examining the challenges of fluid prescribing . As a result of this study medical students have developed fluid prescribing vignettes that have been disseminated regionally to ensure uniformity in teaching. In the focus groups students specifically describe the limited skills supervision received in fluid prescribing. The Regional Hyponatraemia Competence Framework document advocates ‘demonstrated competency’ in the safe prescription of intravenous fluids to avoid the risk of hyponatraemia . This involves supervised prescribing and completion of case studies ratified by a senior member of medical staff. Extrapolating this advice to student intravenous fluid prescribing would suggest that students should also demonstrate competency and this may help them to calibrate their progress in skills development. This ratification could be by a senior doctor although a recent study suggests that medical students prefer ‘near-peer’ education led by supervised junior medical staff . In addition, the use of peer-assisted learning may help to develop paediatric knowledge and skills . Although the benefit of timely and specific feedback is well recognised, and students are keen to receive it, a recent study exploring feedback given to medical students on their communications skills suggests that it is uncommon for students to be directly observed practicing this skill and even less common to receive feedback on it . It is recognised that newly qualified doctors have deficiencies in clinical skills . Preparedness for practice and induction courses have been introduced for Foundation doctors to aid this transition and help with skills acquisition. However, Foundation Year 1 (FY1) doctors do not work in Paediatric wards and the paediatric content of these Foundation courses is minimal . Paediatrics is a competency-based specialty and in the UK the curriculum for postgraduate paediatric training, published by the RCPCH, is modelled on competencies that need to be achieved for successful progression . It is anticipated that completion of these competencies will ultimately lead to better patient outcomes. The RCPCH has recently published a competency-based undergraduate curriculum that specifies essential paediatric skills required of a medical student . The curriculum includes the three skills examined in this study. We propose that acquisition of the skills listed in the curriculum is ratified by demonstration of competence in the clinical setting analogous to Directly Observed Procedural Skills (DOPs) in the postgraduate domain . Entrustable Professional Activities (EPAs) are units of professional activity that can be entrusted to a competent learner and comprise of a number of core competencies. They have been adopted in postgraduate competency-based frameworks and incorporate work-based learning, trust and transparency. They are increasingly being used in undergraduate training to monitor and ratify skills completion and core EPAs for entering a residency programme were proposed and piloted by the Association of American Colleges (AAMC) in 2014 . EPAs will likely help with the integration of postgraduate and undergraduate training and have a positive impact on the development of undergraduate medical skills training. Although there are important messages to drawn from this study, there are also a number of limitations. It is a single-centre study involving a convenience sample of students so it is not known if the results are applicable to other centres. The examiners conducting the skills assessments made an agreed assessment following a period of discussion. The examiners were experienced in examining OSCE scenarios but it is recognised that having agreed decisions rather than analysing independent assessments may contribute to differences. It was fortunate that the study was able to get full participation from all students approached. There were no penalties for opting out of the study. It was clear from focus group discussions that students were keen to get examination practice and that this was likely the reason for full participation.
Skills proficiency can only be acquired through practice. However, practice without appropriate feedback may result in high levels of confidence without resulting in high levels of skills competence. This study has shown that, for completion of three important paediatric tasks, student confidence does not align with competence. These findings are similar to studies comparing confidence and skills competence in adult medicine. An overly confident practitioner who is unable to successfully complete these fundamental paediatric tasks is potentially unsafe. Regarding confidence in skills completion, the focus should move towards the development of trainee task-oriented confidence or self-efficacy rather than the isolated non-specific trait of confidence alone. The study identified modifiable factors that should be addressed to improve skills acquisition and these may also help in the development of self-efficacy. Standardisation of teaching is essential. Demonstrated and ratified competence is required to ensure that students are capable of performing clinical skills. This allows students to apply knowledge, skills and attitudes in a safe, supportive environment. This process may be integrated into the workplace to allow students to calibrate their clinical skills development. The implementation of a comprehensive framework of supervised clinical skills taught and assessed in undergraduate paediatric training will be key to the development of safe and competent medical practitioners.
|
Parsimonious versus extensive bleeding score: can we simplify risk stratification after percutaneous coronary intervention and reduce bleeding events by de-escalation of the antiplatelet strategy? | ba949b08-ef2a-4e13-88c2-63a49d411845 | 11781092 | Surgical Procedures, Operative[mh] | Since the introduction of dual antiplatelet therapy (DAPT), numerous studies have explored varying durations in search of an optimal balance between ischaemic and bleeding events, which resulted in multiple risk models based on patient and PCI characteristics. Due to the multitude of risk factors and risk models, personalised DAPT guidance is complex and rarely implemented in routine practice.
The simplified Zuidoost Nederland Hart Registratie (ZON-HR) classification showed a consistent accuracy compared with the acknowledged PREdicting bleeding Complications In patients undergoing Stent implantation and subsEquent DAPT score. The benefit of aspirin omission was the largest in acute coronary syndrome patients without high bleeding risk, although this finding should be interpreted with care due to the large variation in group sizes.
By increasing the feasibility of bleeding risk stratification after percutaneous coronary intervention (PCI), the ZON-HR classification may augment the use of a patient-tailored antiplatelet strategy after PCI.
In the prevention of thromboembolic complications after percutaneous coronary intervention (PCI), dual antiplatelet therapy (DAPT) has for many years been the unequivocal recommendation in the guidelines. Since the introduction of DAPT, numerous studies have explored varying durations in search of an optimal balance between ischaemic and bleeding events. This ongoing pursuit has yielded multiple risk scores that estimate bleeding- and ischaemic event risks post PCI based on patient and PCI characteristics. While these risk models have proven effective and are integrated into current guidelines, their practical application is questionable. Due to the multitude of risk factors outlined in the guidelines, personalised DAPT guidance is complex and rarely implemented in routine practice. To facilitate the adoption and use of risk stratification among clinicians, a consortium in the South East of the Netherlands, the ‘Zuidoost Nederland Hart Registratie’ (ZON-HR), has developed a simplified classification which includes a limited selection of risk factors provided by previous studies. In line with the guidelines of the European Society of Cardiology (ESC), this classification corresponds with the major Academic Research Consortium for High Bleeding Risk (ARC-HBR) criteria. The selection of criteria incorporated in this classification was based on their known impact on bleeding risk (as demonstrated in the previously mentioned studies), the expected availability at the time of the procedure and excludes rare diseases. This way it can seamlessly integrate into electronic patient records for immediate stratification post PCI. This study aimed to externally validate the ZON-HR classification for predicting high bleeding risk (HBR), focusing on patients without oral anticoagulants (OAC). As per the ARC-HBR consensus, patients on OAC are inherently considered HBR, irrespective of other risk factors. The analysis used the GLOBAL LEADERS trial population and compared the ZON-HR classification with the PREdicting bleeding Complications In patients undergoing Stent implantation and subsEquent DAPT (PRECISE-DAPT) score, endorsed by current ESC guidelines and previously validated in the GLOBAL LEADERS population. Additionally, the study explored the impact of a de-escalation strategy by comparing P2Y12 inhibitor monotherapy after 1 month of DAPT with 12-month DAPT in patients with or without HBR, leveraging the ZON-HR classification for risk stratification.
Study design and participants In a post hoc analysis of the GLOBAL LEADERS trial (NCT01813435), a multicentre, open-label randomised trial comparing ticagrelor monotherapy following 1 month of DAPT with standard DAPT, we sought to examine the outcomes. The trial encompassed 15 968 patients undergoing PCI for chronic coronary syndrome (CCS) or acute coronary syndrome (ACS), with exclusion criteria applying to those on OAC. Randomisation placed patients into either the experimental group (ticagrelor monotherapy for 24 months after 1 month of ticagrelor plus aspirin) or the control group (DAPT per guidelines: clopidogrel plus aspirin for CCS and ticagrelor plus aspirin for ACS). The follow-up duration was 2 years, with the control group transitioning to aspirin monotherapy after 1 year, aligning with standard practice. Comprehensive study design, protocol, outcome details and information on data sharing are available elsewhere. As this is a post hoc analysis, patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research. For this analysis, we focused on 1-year follow-up data, excluding patients with missing variables essential for the PRECISE-DAPT score or the ZON-HR classification. The Standards for Reporting of Diagnostic Accuracy Studies guidelines were followed in this analysis. Definitions The parameters for the ZON-HR classification and PRECISE-DAPT score were derived from the patients’ medical history and clinical characteristics at time of enrolment. In the ZON-HR classification, HBR is stratified according to either: a history of intracranial haemorrhage; a previous spontaneous bleeding in the past year with a Bleeding Academic Research Consortium (BARC) Score of at least 2; a haemoglobin level less than 11 g/dL at baseline or an estimated glomerular filtration rate (eGFR) below 30 mL/min at baseline. In the GLOBAL LEADERS trial, previous bleeding was not restricted to bleeding events in the last 12 months before PCI, therefore we replaced this ZON-HR parameter with all prior bleeding events. Furthermore, a history of intracranial haemorrhage was an exclusion criterion for the GLOBAL LEADERS trial. Therefore, no patients met this ZON-HR criterion for HBR. The PRECISE-DAPT score was calculated for each patient based on age, creatinine clearance, haemoglobin concentration, white blood cell count and previous spontaneous bleeding. Study endpoints We stratified the patients for HBR according to the ZON-HR classification or a PRECISE-DAPT score ≥25 and assessed the predictive performance and prognostic value of both risk stratification methods for site reported minor and major bleeding events according to the BARC classification and for major adverse cardiac and cerebral events (MACCE), consisting of site reported myocardial infarction, stroke and cardiac death at 1-year follow-up. The analyses were performed in the overall population and in subgroups for PCI indication (ACS and CCS). Furthermore, we assessed whether bleeding risk stratification according to ZON-HR is effective in patients who are randomised to P2Y12 inhibitor monotherapy after 1 month of DAPT instead of 12 months DAPT, and we determined the interaction of HBR according to ZON-HR on the treatment effect. Statistical analysis We summarised continuous variables as mean with SD or median with IQR. Categorical variables are summarised as counts and percentages. The predictive value of both methods are evaluated using positive predictive values (PPVs), negative predictive values (NPVs), sensitivity, specificity, accuracy and the corresponding CIs of these metrics. To make a comparison of both classification methods to the c-statistic of the continuous PRECISE-DAPT score, the sensitivity and specificity of the risk stratification according to the ZON-HR classification and according to a PRECISE-DAPT score ≥25 were depicted on the receiver operating characteristic (ROC) curve of the continuous PRECISE-DAPT score. Cohen’s kappa coefficient was estimated to evaluate the level of agreement between the two methods. Cox regression analysis was performed to estimate the HRs and 95% CIs of bleeding events and of MACCE at 1-year follow-up according to bleeding risk. The treatment effects of P2Y12 inhibitor monotherapy versus DAPT were tested for interaction between the HBR and non-HBR subgroups. All data are processed using R V.4.1.3. (R Foundation for Statistical Computing, Vienna, Austria).
In a post hoc analysis of the GLOBAL LEADERS trial (NCT01813435), a multicentre, open-label randomised trial comparing ticagrelor monotherapy following 1 month of DAPT with standard DAPT, we sought to examine the outcomes. The trial encompassed 15 968 patients undergoing PCI for chronic coronary syndrome (CCS) or acute coronary syndrome (ACS), with exclusion criteria applying to those on OAC. Randomisation placed patients into either the experimental group (ticagrelor monotherapy for 24 months after 1 month of ticagrelor plus aspirin) or the control group (DAPT per guidelines: clopidogrel plus aspirin for CCS and ticagrelor plus aspirin for ACS). The follow-up duration was 2 years, with the control group transitioning to aspirin monotherapy after 1 year, aligning with standard practice. Comprehensive study design, protocol, outcome details and information on data sharing are available elsewhere. As this is a post hoc analysis, patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research. For this analysis, we focused on 1-year follow-up data, excluding patients with missing variables essential for the PRECISE-DAPT score or the ZON-HR classification. The Standards for Reporting of Diagnostic Accuracy Studies guidelines were followed in this analysis.
The parameters for the ZON-HR classification and PRECISE-DAPT score were derived from the patients’ medical history and clinical characteristics at time of enrolment. In the ZON-HR classification, HBR is stratified according to either: a history of intracranial haemorrhage; a previous spontaneous bleeding in the past year with a Bleeding Academic Research Consortium (BARC) Score of at least 2; a haemoglobin level less than 11 g/dL at baseline or an estimated glomerular filtration rate (eGFR) below 30 mL/min at baseline. In the GLOBAL LEADERS trial, previous bleeding was not restricted to bleeding events in the last 12 months before PCI, therefore we replaced this ZON-HR parameter with all prior bleeding events. Furthermore, a history of intracranial haemorrhage was an exclusion criterion for the GLOBAL LEADERS trial. Therefore, no patients met this ZON-HR criterion for HBR. The PRECISE-DAPT score was calculated for each patient based on age, creatinine clearance, haemoglobin concentration, white blood cell count and previous spontaneous bleeding.
We stratified the patients for HBR according to the ZON-HR classification or a PRECISE-DAPT score ≥25 and assessed the predictive performance and prognostic value of both risk stratification methods for site reported minor and major bleeding events according to the BARC classification and for major adverse cardiac and cerebral events (MACCE), consisting of site reported myocardial infarction, stroke and cardiac death at 1-year follow-up. The analyses were performed in the overall population and in subgroups for PCI indication (ACS and CCS). Furthermore, we assessed whether bleeding risk stratification according to ZON-HR is effective in patients who are randomised to P2Y12 inhibitor monotherapy after 1 month of DAPT instead of 12 months DAPT, and we determined the interaction of HBR according to ZON-HR on the treatment effect.
We summarised continuous variables as mean with SD or median with IQR. Categorical variables are summarised as counts and percentages. The predictive value of both methods are evaluated using positive predictive values (PPVs), negative predictive values (NPVs), sensitivity, specificity, accuracy and the corresponding CIs of these metrics. To make a comparison of both classification methods to the c-statistic of the continuous PRECISE-DAPT score, the sensitivity and specificity of the risk stratification according to the ZON-HR classification and according to a PRECISE-DAPT score ≥25 were depicted on the receiver operating characteristic (ROC) curve of the continuous PRECISE-DAPT score. Cohen’s kappa coefficient was estimated to evaluate the level of agreement between the two methods. Cox regression analysis was performed to estimate the HRs and 95% CIs of bleeding events and of MACCE at 1-year follow-up according to bleeding risk. The treatment effects of P2Y12 inhibitor monotherapy versus DAPT were tested for interaction between the HBR and non-HBR subgroups. All data are processed using R V.4.1.3. (R Foundation for Statistical Computing, Vienna, Austria).
Patient population Of the 15 968 patients included in the GLOBAL LEADERS trial, the parameters for the ZON-HR classification were available in 15 947 (99.9%) patients and the PRECISE-DAPT score could be calculated in 14 928 (93%) patients. This study included only patients with available parameters for both risk stratification methods, which was a total of 14 909 (93%). A flowchart of included patients is presented in . Baseline differences between the included and excluded patients are presented in . Baseline characteristics of the patient population are presented for the included population and stratified for HBR according to either PRECISE-DAPT score ≥25 or the ZON-HR classification . A total of 2467 (16.6%) patients had HBR according to a PRECISE-DAPT score ≥25 and 555 (3.7%) patients according to the ZON-HR classification, of which 485 were also deemed HBR according to a PRECISE-DAPT score ≥25. Of these 555 patients, 20.9% had an eGFR below 30 mL/min/1.73 m 2 , 16.2% had a previous bleeding and 70.5% had an anaemia (7.6% of patients had two risk factors). In addition to the differences in bleeding risk factors, shows that patients with HBR according to either classification also have more ischaemic risk factors. Allocated treatment with DAPT for 12 months or ticagrelor for 11 months after 1 month of DAPT was comparable between patients with or without HBR according to either classification. Predictive value shows the PPV, NPV, positive likelihood ratio (PLR), negative likelihood ratio (NLR) and accuracy of both risk stratifications. Although the incidence of HBR is low according to ZON-HR, these patients more often have BARC 2, 3 and 5 bleeding events (13%) compared with patients with HBR according to the PRECISE-DAPT score (8%). This results in a higher PPV, PLR, NLR and accuracy for the ZON-HR classification. shows the ROC curves for the prediction of bleeding events of the continuous PRECISE-DAPT score. The sensitivity and specificity of the risk stratification according to ZON-HR and according to a PRECISE-DAPT score ≥25 are depicted on the curves. CCS and ACS subgroups are presented in . The continuous PRECISE-DAPT score showed to have poor predictability in the GLOBAL LEADERS population with an area under the curve (AUC) of 0.60 (95% CI: 0.58 to 0.62) for BARC 2, 3 and 5 bleeding events (panel A) and an AUC of 0.65 (95% CI: 0.61 to 0.68) for BARC 3 and 5 bleedings (panel B). The sensitivity and specificity of the ZON-HR classification for predicting BARC 2, 3 or 5 bleedings lie on the ROC curve with a sensitivity of 0.09 (95% CI: 0.07 to 0.11) and specificity of 0.97 (95% CI: 0.96 to 0.97) compared with 0.26 (95% CI: 0.23 to 0.30) and 0.84 (95% CI: 0.83 to 0.85), respectively, for a PRECISE-DAPT score ≥25. Both methods showed slight better predictive performance for BARC 3 and 5 bleeding events. Cohen’s kappa coefficient showed a fair level of agreement between the two methods (Κ=0.28). Prognostic value for trial endpoints demonstrates the HRs with corresponding CIs for bleeding events and MACCE in patients with HBR versus no HBR according to the PRECISE-DAPT score and according to ZON-HR. In the overall population, patients with HBR according to either classification showed significant more BARC 2, 3 and 5 and BARC 3 and 5 bleeding events and more MACCE compared with patients without HBR. Subgroup analyses for patients with CCS and ACS showed similar results. The estimated association of the ZON-HR classification with bleeding risk was consistent for patients receiving monotherapy group with an HR of 3.47 (95% CI: 2.53 to 4.75). The ischaemic and bleeding risks are graphically presented in the for CCS and ACS subgroups stratified for bleeding risk. Treatment effect stratified by bleeding risk The effect of P2Y12 inhibitor monotherapy after 1 month of DAPT compared with DAPT for 1 year on bleeding events stratified by bleeding risk according to ZON-HR is demonstrated separately for CCS and ACS in . In the CCS population, clopidogrel plus aspirin was compared with ticagrelor monotherapy after 1 month of ticagrelor plus aspirin. There was no difference between treatment strategies regarding BARC 2, 3 and 5 bleedings in patients with or without HBR. However, patients with HBR in the experimental group showed a numerical increase of BARC 2, 3 and 5 bleeding events compared with patients with HBR receiving standard treatment with a non-significant HR of 1.57 (95% CI: 0.80 to 3.06) . The cumulative incidence curve of the experimental group shows the strongest increase of bleeding events in the first month after PCI during which the patients received ticagrelor plus aspirin. In the ACS population, ticagrelor plus aspirin was compared with ticagrelor monotherapy after 1 month of ticagrelor plus aspirin. There was a significant reduction of bleeding events in the experimental group of non-HBR patients which was not observed in patients with HBR, with a significant interaction of the bleeding risk on the treatment effect .
Of the 15 968 patients included in the GLOBAL LEADERS trial, the parameters for the ZON-HR classification were available in 15 947 (99.9%) patients and the PRECISE-DAPT score could be calculated in 14 928 (93%) patients. This study included only patients with available parameters for both risk stratification methods, which was a total of 14 909 (93%). A flowchart of included patients is presented in . Baseline differences between the included and excluded patients are presented in . Baseline characteristics of the patient population are presented for the included population and stratified for HBR according to either PRECISE-DAPT score ≥25 or the ZON-HR classification . A total of 2467 (16.6%) patients had HBR according to a PRECISE-DAPT score ≥25 and 555 (3.7%) patients according to the ZON-HR classification, of which 485 were also deemed HBR according to a PRECISE-DAPT score ≥25. Of these 555 patients, 20.9% had an eGFR below 30 mL/min/1.73 m 2 , 16.2% had a previous bleeding and 70.5% had an anaemia (7.6% of patients had two risk factors). In addition to the differences in bleeding risk factors, shows that patients with HBR according to either classification also have more ischaemic risk factors. Allocated treatment with DAPT for 12 months or ticagrelor for 11 months after 1 month of DAPT was comparable between patients with or without HBR according to either classification.
shows the PPV, NPV, positive likelihood ratio (PLR), negative likelihood ratio (NLR) and accuracy of both risk stratifications. Although the incidence of HBR is low according to ZON-HR, these patients more often have BARC 2, 3 and 5 bleeding events (13%) compared with patients with HBR according to the PRECISE-DAPT score (8%). This results in a higher PPV, PLR, NLR and accuracy for the ZON-HR classification. shows the ROC curves for the prediction of bleeding events of the continuous PRECISE-DAPT score. The sensitivity and specificity of the risk stratification according to ZON-HR and according to a PRECISE-DAPT score ≥25 are depicted on the curves. CCS and ACS subgroups are presented in . The continuous PRECISE-DAPT score showed to have poor predictability in the GLOBAL LEADERS population with an area under the curve (AUC) of 0.60 (95% CI: 0.58 to 0.62) for BARC 2, 3 and 5 bleeding events (panel A) and an AUC of 0.65 (95% CI: 0.61 to 0.68) for BARC 3 and 5 bleedings (panel B). The sensitivity and specificity of the ZON-HR classification for predicting BARC 2, 3 or 5 bleedings lie on the ROC curve with a sensitivity of 0.09 (95% CI: 0.07 to 0.11) and specificity of 0.97 (95% CI: 0.96 to 0.97) compared with 0.26 (95% CI: 0.23 to 0.30) and 0.84 (95% CI: 0.83 to 0.85), respectively, for a PRECISE-DAPT score ≥25. Both methods showed slight better predictive performance for BARC 3 and 5 bleeding events. Cohen’s kappa coefficient showed a fair level of agreement between the two methods (Κ=0.28).
demonstrates the HRs with corresponding CIs for bleeding events and MACCE in patients with HBR versus no HBR according to the PRECISE-DAPT score and according to ZON-HR. In the overall population, patients with HBR according to either classification showed significant more BARC 2, 3 and 5 and BARC 3 and 5 bleeding events and more MACCE compared with patients without HBR. Subgroup analyses for patients with CCS and ACS showed similar results. The estimated association of the ZON-HR classification with bleeding risk was consistent for patients receiving monotherapy group with an HR of 3.47 (95% CI: 2.53 to 4.75). The ischaemic and bleeding risks are graphically presented in the for CCS and ACS subgroups stratified for bleeding risk.
The effect of P2Y12 inhibitor monotherapy after 1 month of DAPT compared with DAPT for 1 year on bleeding events stratified by bleeding risk according to ZON-HR is demonstrated separately for CCS and ACS in . In the CCS population, clopidogrel plus aspirin was compared with ticagrelor monotherapy after 1 month of ticagrelor plus aspirin. There was no difference between treatment strategies regarding BARC 2, 3 and 5 bleedings in patients with or without HBR. However, patients with HBR in the experimental group showed a numerical increase of BARC 2, 3 and 5 bleeding events compared with patients with HBR receiving standard treatment with a non-significant HR of 1.57 (95% CI: 0.80 to 3.06) . The cumulative incidence curve of the experimental group shows the strongest increase of bleeding events in the first month after PCI during which the patients received ticagrelor plus aspirin. In the ACS population, ticagrelor plus aspirin was compared with ticagrelor monotherapy after 1 month of ticagrelor plus aspirin. There was a significant reduction of bleeding events in the experimental group of non-HBR patients which was not observed in patients with HBR, with a significant interaction of the bleeding risk on the treatment effect .
Parsimonious versus extensive score This external validation of the PRECISE-DAPT score and the more concise ZON-HR classification for bleeding risk stratification resulted in several important and clinically relevant findings: In comparison to the PRECISE-DAPT score, risk stratification was more frequently possible with the ZON-HR classification. Due to the small number of commonly available risk factors, only 0.1% of the included patients in the GLOBAL LEADERS trial lacked one or more parameters required by the ZON-HR classification. This finding emphasises the main advantage of a parsimonious classification compared with a more extensive risk score. Using only variables that are commonly available periprocedural, it is more feasible for the treating physician to provide a patient-tailored antiplatelet strategy based on the bleeding risk, as is supported by the guidelines. The feasibility of the currently available risk scores seems to be a pitfall as practical use is limited and the predictive capacities are poor or moderate. The ZON-HR classification was relatively conservative in the prediction of bleeding risk compared with the more liberal PRECISE-DAPT score, as only 3.7% of the included population had HBR according to ZON-HR compared with 16.6% according to a PRECISE-DAPT score ≥25. This conservative character of the ZON-HR classification showed several advantages and disadvantages compared with the PRECISE-DAPT score. The PPV and the specificity of the ZON-HR classification were higher, indicating that patients with low bleeding risk are less often unjustly stratified to HBR. This creates the possibility to even extend antiplatelet therapy based on thromboembolic risk in these patients, which would otherwise be advised against based on the supposed bleeding risk. However, the ZON-HR classification may be too conservative, as the sensitivity showed to be very low indicating that a high number of patients with bleeding events were stratified to non-HBR. As the ZON-HR classification is binary, no AUC could be calculated. In order to make a comparison to the PRECISE-DAPT score, the sensitivity and specificity of both the ZON-HR classification and a PRECISE-DAPT score ≥25 were depicted on the ROC curve of the continuous PRECISE-DAPT score. The ZON-HR classification showed to lie on the ROC curve of the PRECISE-DAPT score, this suggests a comparable accuracy of the two methods in the prediction of bleeding risk. Cox regression analysis of the two methods showed that both are capable to discriminate for the risk of a bleeding as bleeding events were significantly more common in patients with HBR according to either method. This was similar for all bleedings (BARC 2, 3 and 5) and for major bleedings (BARC 3 and 5) only. The ZON-HR classification showed slightly higher HRs for bleeding events compared with the PRECISE-DAPT score. In combination with the sensitivity and specificity, this may indicate that the ZON-HR classification mainly identifies patients with a very high risk of bleeding but disregards patients with less high risk of bleeding. Although the two stratification methods partly use the same risk factors, a difference in performance can be explained by the cumulative versus separate contribution of the risk factors to the bleeding risk. According to the ZON-HR method, the separate risk factors are regarded as predictor for bleeding risk, as opposed to the cumulative contribution of the risk factors in the PRECISE-DAPT score. Our findings are in line with a recent study that compared a simplified clinical evaluation tool in elderly patients, consisting of three major ARC-HBR criteria. Although the selection of criteria was slightly different in this study, it also proved effective in the prediction of major bleedings in elderly patients when compared with the PRECISE-DAPT score and the ARC-HBR criteria. This further supports the concept of a small selection of common risk factors as a feasible tool for the prediction of bleeding risk. Treatment effect The cumulative event curves of treatment effect showed that P2Y12 inhibitor monotherapy only reduced bleeding events in ACS patients without HBR. This is consistent with previous findings within the GLOBAL LEADERS population. Apart from clinical presentation, these results could be attributed to differences in treatment in the control group between patients with CCS and ACS. In patients with CCS, clopidogrel plus aspirin was compared with ticagrelor monotherapy after 1 month of ticagrelor plus aspirin which may have resulted in comparable numbers of bleeding between treatment groups. Compared with clopidogrel and aspirin, CCS patients with HBR according to ZON-HR treated with ticagrelor showed a non-significant higher number of bleeding events which suggests that patients with (very) high risk of bleeding may experience more harm of ticagrelor (monotherapy) compared with clopidogrel plus aspirin. This effect was the strongest in the first days after PCI and indicates that the risk of bleeding is the highest in this period which corresponds to previous findings. In patients with ACS, the control group received ticagrelor plus aspirin, which provided a fairer comparison to ticagrelor monotherapy. Our results showed that ACS patients with HBR did not benefit from ticagrelor monotherapy. In this population, the (non-significant) higher number of bleeding events in the HBR subgroup cannot be explained by antiplatelet regime as both treatment groups received ticagrelor. However, the fact that treatment with ticagrelor monotherapy did not result in a reduction of bleeding events in patients with HBR as opposed to non-HBR patients might be attributed to the potency of ticagrelor. The omission of aspirin may not be enough de-escalation in patients with HBR when co-treated with ticagrelor to prevent a bleeding event. Previous studies showed that the omission of aspirin after 1 month reduces bleeding events in patients with HBR when co-treatment (mainly) consists of clopidogrel. A substudy of the TWILIGHT trial showed similar results when comparing ticagrelor monotherapy after 3 months of DAPT to ticagrelor plus aspirin in patients with HBR. The difference in outcomes between the TWILIGHT substudy and the results found in the GLOBAL LEADERS population may be attributed to the difference in DAPT duration in the experimental groups and the differences in risk stratification, which was performed according to the ARC-HBR criteria in TWILIGHT-HBR, or by differences between inclusion and exclusion criteria. Furthermore, the relative small group size of patients with HBR may have influenced the outcomes. As our study design and outcomes are different compared with previous studies, the results regarding treatment effect should be regarded purely as hypothesis generating and require further investigation. The differences between study outcomes and the combined study effect of ticagrelor monotherapy after a short period of DAPT compared with standard DAPT in patients with or without HBR has previously been investigated and showed a comparable treatment effect. The ZON-HR bleeding risk classification has been implemented within the participating centres of the ZON-HR consortium in which patients with HBR receive a shortened DAPT duration. The clinical effect of shortened DAPT in this population will be determined in future analyses. Limitations This external validation has several limitations. Most importantly, the inclusion and exclusion criteria of the GLOBAL LEADERS trial caused selection bias of the patient population. This may explain the low number of patients with HBR as the study population may have been of relatively low risk. This could also explain the relatively low c-statistic of 0.60 of the continuous PRECISE-DAPT score when compared with the derivation cohort in which the AUC showed to be 0.70 (95% CI: 0.65 to 0.74) and to a previous external validation in an all-comers PCI registration showing an AUC of 0.66 (95% CI: 0.61 to 0.71). Therefore, the results of our analyses cannot be directly translated to an all-comers PCI population. Also, the GLOBAL LEADERS mainly consists of a western population and included less than 1% Asian patients. East Asian patients are known to have a higher risk of bleeding events and therefore may require an adjusted risk stratification such as the Japanese HBR criteria. However, it has recently been demonstrated that these criteria have a similar discriminative ability compared with the ARC-HBR criteria and the PRECISE-DAPT score in a Japanese population. Furthermore, a history of intracranial haemorrhage was an exclusion criterion in the GLOBAL LEADERS trial. This is one of the four criteria of the ZON-HR classification. As no patient fulfilled this criterion in this study, the effect of a previous intracranial haemorrhage on bleeding events could not be measured. Lastly, the analyses relating to various subgroups in the study were not part of the original study design, and it is possible that power was insufficient to detect important differences between the risk stratification methods and treatment strategies.
This external validation of the PRECISE-DAPT score and the more concise ZON-HR classification for bleeding risk stratification resulted in several important and clinically relevant findings: In comparison to the PRECISE-DAPT score, risk stratification was more frequently possible with the ZON-HR classification. Due to the small number of commonly available risk factors, only 0.1% of the included patients in the GLOBAL LEADERS trial lacked one or more parameters required by the ZON-HR classification. This finding emphasises the main advantage of a parsimonious classification compared with a more extensive risk score. Using only variables that are commonly available periprocedural, it is more feasible for the treating physician to provide a patient-tailored antiplatelet strategy based on the bleeding risk, as is supported by the guidelines. The feasibility of the currently available risk scores seems to be a pitfall as practical use is limited and the predictive capacities are poor or moderate. The ZON-HR classification was relatively conservative in the prediction of bleeding risk compared with the more liberal PRECISE-DAPT score, as only 3.7% of the included population had HBR according to ZON-HR compared with 16.6% according to a PRECISE-DAPT score ≥25. This conservative character of the ZON-HR classification showed several advantages and disadvantages compared with the PRECISE-DAPT score. The PPV and the specificity of the ZON-HR classification were higher, indicating that patients with low bleeding risk are less often unjustly stratified to HBR. This creates the possibility to even extend antiplatelet therapy based on thromboembolic risk in these patients, which would otherwise be advised against based on the supposed bleeding risk. However, the ZON-HR classification may be too conservative, as the sensitivity showed to be very low indicating that a high number of patients with bleeding events were stratified to non-HBR. As the ZON-HR classification is binary, no AUC could be calculated. In order to make a comparison to the PRECISE-DAPT score, the sensitivity and specificity of both the ZON-HR classification and a PRECISE-DAPT score ≥25 were depicted on the ROC curve of the continuous PRECISE-DAPT score. The ZON-HR classification showed to lie on the ROC curve of the PRECISE-DAPT score, this suggests a comparable accuracy of the two methods in the prediction of bleeding risk. Cox regression analysis of the two methods showed that both are capable to discriminate for the risk of a bleeding as bleeding events were significantly more common in patients with HBR according to either method. This was similar for all bleedings (BARC 2, 3 and 5) and for major bleedings (BARC 3 and 5) only. The ZON-HR classification showed slightly higher HRs for bleeding events compared with the PRECISE-DAPT score. In combination with the sensitivity and specificity, this may indicate that the ZON-HR classification mainly identifies patients with a very high risk of bleeding but disregards patients with less high risk of bleeding. Although the two stratification methods partly use the same risk factors, a difference in performance can be explained by the cumulative versus separate contribution of the risk factors to the bleeding risk. According to the ZON-HR method, the separate risk factors are regarded as predictor for bleeding risk, as opposed to the cumulative contribution of the risk factors in the PRECISE-DAPT score. Our findings are in line with a recent study that compared a simplified clinical evaluation tool in elderly patients, consisting of three major ARC-HBR criteria. Although the selection of criteria was slightly different in this study, it also proved effective in the prediction of major bleedings in elderly patients when compared with the PRECISE-DAPT score and the ARC-HBR criteria. This further supports the concept of a small selection of common risk factors as a feasible tool for the prediction of bleeding risk.
The cumulative event curves of treatment effect showed that P2Y12 inhibitor monotherapy only reduced bleeding events in ACS patients without HBR. This is consistent with previous findings within the GLOBAL LEADERS population. Apart from clinical presentation, these results could be attributed to differences in treatment in the control group between patients with CCS and ACS. In patients with CCS, clopidogrel plus aspirin was compared with ticagrelor monotherapy after 1 month of ticagrelor plus aspirin which may have resulted in comparable numbers of bleeding between treatment groups. Compared with clopidogrel and aspirin, CCS patients with HBR according to ZON-HR treated with ticagrelor showed a non-significant higher number of bleeding events which suggests that patients with (very) high risk of bleeding may experience more harm of ticagrelor (monotherapy) compared with clopidogrel plus aspirin. This effect was the strongest in the first days after PCI and indicates that the risk of bleeding is the highest in this period which corresponds to previous findings. In patients with ACS, the control group received ticagrelor plus aspirin, which provided a fairer comparison to ticagrelor monotherapy. Our results showed that ACS patients with HBR did not benefit from ticagrelor monotherapy. In this population, the (non-significant) higher number of bleeding events in the HBR subgroup cannot be explained by antiplatelet regime as both treatment groups received ticagrelor. However, the fact that treatment with ticagrelor monotherapy did not result in a reduction of bleeding events in patients with HBR as opposed to non-HBR patients might be attributed to the potency of ticagrelor. The omission of aspirin may not be enough de-escalation in patients with HBR when co-treated with ticagrelor to prevent a bleeding event. Previous studies showed that the omission of aspirin after 1 month reduces bleeding events in patients with HBR when co-treatment (mainly) consists of clopidogrel. A substudy of the TWILIGHT trial showed similar results when comparing ticagrelor monotherapy after 3 months of DAPT to ticagrelor plus aspirin in patients with HBR. The difference in outcomes between the TWILIGHT substudy and the results found in the GLOBAL LEADERS population may be attributed to the difference in DAPT duration in the experimental groups and the differences in risk stratification, which was performed according to the ARC-HBR criteria in TWILIGHT-HBR, or by differences between inclusion and exclusion criteria. Furthermore, the relative small group size of patients with HBR may have influenced the outcomes. As our study design and outcomes are different compared with previous studies, the results regarding treatment effect should be regarded purely as hypothesis generating and require further investigation. The differences between study outcomes and the combined study effect of ticagrelor monotherapy after a short period of DAPT compared with standard DAPT in patients with or without HBR has previously been investigated and showed a comparable treatment effect. The ZON-HR bleeding risk classification has been implemented within the participating centres of the ZON-HR consortium in which patients with HBR receive a shortened DAPT duration. The clinical effect of shortened DAPT in this population will be determined in future analyses.
This external validation has several limitations. Most importantly, the inclusion and exclusion criteria of the GLOBAL LEADERS trial caused selection bias of the patient population. This may explain the low number of patients with HBR as the study population may have been of relatively low risk. This could also explain the relatively low c-statistic of 0.60 of the continuous PRECISE-DAPT score when compared with the derivation cohort in which the AUC showed to be 0.70 (95% CI: 0.65 to 0.74) and to a previous external validation in an all-comers PCI registration showing an AUC of 0.66 (95% CI: 0.61 to 0.71). Therefore, the results of our analyses cannot be directly translated to an all-comers PCI population. Also, the GLOBAL LEADERS mainly consists of a western population and included less than 1% Asian patients. East Asian patients are known to have a higher risk of bleeding events and therefore may require an adjusted risk stratification such as the Japanese HBR criteria. However, it has recently been demonstrated that these criteria have a similar discriminative ability compared with the ARC-HBR criteria and the PRECISE-DAPT score in a Japanese population. Furthermore, a history of intracranial haemorrhage was an exclusion criterion in the GLOBAL LEADERS trial. This is one of the four criteria of the ZON-HR classification. As no patient fulfilled this criterion in this study, the effect of a previous intracranial haemorrhage on bleeding events could not be measured. Lastly, the analyses relating to various subgroups in the study were not part of the original study design, and it is possible that power was insufficient to detect important differences between the risk stratification methods and treatment strategies.
The ZON-HR classification demonstrated higher feasibility compared with the PRECISE-DAPT score, properly identifying bleeding risk with a selection of a limited set of commonly available ARC-HBR criteria. While conservative, it provided accuracy which was consistent with the PRECISE-DAPT Score. Notably, aspirin omission reduced bleeding events solely in ACS patients without HBR, emphasising the need for a robust de-escalation strategy in this subgroup. However, limitations in patient selection, demographic diversity and group sizes warrant caution in generalising these findings.
10.1136/openhrt-2024-003083 online supplemental file 1 10.1136/openhrt-2024-003083 online supplemental file 2
|
Evaluación del cumplimiento de las recomendaciones de "no hacer" de la Sociedad Española de Medicina Preventiva y Salud Pública | 79e24616-d496-4b67-aa65-33ef17ad137c | 11582859 | Preventive Medicine[mh] | Hasta el momento, casi todo el conocimiento científico se ha dirigido a analizar y a evaluar intervenciones sanitarias que se deben realizar en el paciente. Sin embargo, existen evidencias de que determinadas prácticas diagnósticas, terapéuticas y perfiles de cuidados sanitarios son ineficientes, inseguros o innecesarios y no aportan un valor añadido relevante para el paciente ). La reducción de estas prácticas es una medida de eficiencia y la toma de decisiones clínicas eficientes es un compromiso ético reflejado en diversos códigos profesionales . Durante la última década, las sociedades científicas y profesionales han mostrado interés en la mejora de la atención sanitaria y, para ello, se han desarrollado distintos proyectos. En el año 2009 se elaboran las iniciativas institucionales " Choosing Wisely " de la Alianza Nacional de Médicos ( National Physicians Alliance ) , a través del American Board of Internal Medicine , y " Less is more " de la American Medical Association , donde las sociedades científicas deben proponer cinco recomendaciones principales sobre "no hacer", facilitando la toma de decisiones compartidas en la práctica clínica y promoviendo la eficiencia. Del mismo modo y de forma simultánea, el National Institute for Clinical Excellence (NICE) determina desde 2007 algunas prácticas clínicas que recomienda no hacer (" Do not do ") porque no aportan beneficio, la relación riesgo-beneficio no está clara o no existe evidencia suficiente . En 2013 surge en España el proyecto " Compromiso por la Calidad de las Sociedades Científicas en España " también conocido como " No hacer ", iniciativa de la Sociedad Española de Medicina Interna . El Ministerio de Sanidad, Consumo y Bienestar Social pone en marcha el proyecto con el fin de acordar con las diferentes sociedades científicas diversas recomendaciones de "no hacer" basadas en la evidencia científica . En cada sociedad se nombra un panel de expertos, para elegir después las 5 recomendaciones utilizando el método Delphi. Con ello, además de disminuir el uso de intervenciones médicas innecesarias y la iatrogenia, se pretende la reducción de la variabilidad y la promoción de la seguridad en la práctica clínica, así como la difusión de las recomendaciones para orientar en la toma de decisiones . En 2018, la Sociedad Española de Medicina Preventiva, Salud Pública e Higiene (SEMPSPH), adherida al proyecto, presenta la propuesta de estas cinco recomendaciones: "No eliminar el vello de forma sistemática para reducir el riesgo de infección de sitio quirúrgico." Si fuera necesario, usar cortadoras de pelo adecuadas (maquinillas eléctricas, cortadoras de pelo, depilación química). "No continuar con antibióticos más de 24-48 horas en pacientes hospitalizados", a menos que haya evidencia clara de infección. "No se recomienda el análisis de la toxina Clostridium difficile en pacientes asintomáticos." "No utilice la descontaminación nasal con agentes antimicrobianos tópicos destinados a eliminar el Staphylococcus aureus rutinariamente, para reducir el riesgo de infección del sitio quirúrgico, ni ante procedimientos cardíacos ni ortopédicos." "No se recomienda el reemplazo rutinario de catéteres venosos periféricos cada 72-96 horas." El objetivo del presente trabajo fue evaluar el cumplimiento de las cinco recomendaciones "No hacer" propuestas por la SEMPSPH en los pacientes atendidos en el Hospital Universitario de La Princesa de Madrid. Se diseñó un estudio de la evaluación de la calidad asistencial siguiendo la metodología de Palmer o ciclo de evaluación PDCA, que incluye cuatro fases: - Planificación o definición de objetivos ( Plan ). - Diseño del estudio y recogida de información ( Do ). - Análisis de datos ( Check ). - Implantación de medidas correctoras o de mejora ( Act ) . En primer lugar, se definió la dimensión de la calidad evaluada, siendo esta la calidad científico-técnica. Después, se elaboraron los criterios, que son el instrumento de medida que se utilizó para evaluar la calidad. Se establecieron como criterios las 5 recomendaciones propuestas por la SEMPSPH para el proyecto "No hacer". Estas recomendaciones fueron establecidas por etapas por un panel de 25 expertos designados por la Sociedad Científica, empleando el método Delphi. El panel estableció un listado preliminar de 15 recomendaciones de "no hacer" basadas en la evidencia científica, obtenidas de las Guías de Práctica Clínica como fuente principal. Posteriormente, se realizó una técnica Delphi en la que los panelistas valoraron cada recomendación con una escala de puntuación y jerarquización, obteniéndose el consenso por un procedimiento matemático de agregación de juicios individuales, utilizando la mediana y el rango intercuartílico . Se definieron las excepciones, es decir, aquellas circunstancias en las que no era exigible que se cumpliera el criterio, y se empleó como indicador para medir el criterio y permitir su interpretación el Índice Global de Calidad (IC) expresado en forma de porcentajes. Así, para la Recomendación 1 (" No eliminar el vello de forma sistemática para reducir el riesgo de infección de sitio quirúrgico ") se definió como criterio el no rasurado del sitio quirúrgico, con excepción de rasurado del sitio quirúrgico por indicación médica. Como indicador se estableció el Índice Global de Calidad (IC=Pacientes no rasurados/Total pacientes intervenidos) expresado en porcentaje. Para la Recomendación 2 ("No continuar con antibióticos más de 24-48 horas en pacientes hospitalizados, a menos que haya evidencia clara de infección") se definió como criterio la continuación del tratamiento antibiótico durante un periodo inferior a 24-48 horas en ausencia de infección, con excepción de continuar el tratamiento antibiótico durante un periodo mayor a 48 horas en presencia de signos claros de infección. Como indicador se estableció el Índice Global de Calidad (IC=Pacientes que no continúen tratamiento antibiótico/Total pacientes intervenidos) expresado en porcentaje. Para la Recomendación 3 ("No se recomienda el análisis de la toxina Clostridium diffícile en pacientes asintomáticos") se definió como criterio el análisis de la toxina Clostridium diffícile en pacientes sintomáticos. Como indicador se estableció el Índice Global de Calidad (IC=Pacientes sintomáticos/Total peticiones de análisis de la toxina Clostridium diffícile ) expresado en porcentaje. Para la Recomendación 4 (" No utilice la descontaminación nasal con agentes antimicrobianos tópicos destinados a eliminar el Staphylococcus aureus rutinariamente, para reducir el riesgo de infección del sitio quirúrgico, ante procedimientos cardíacos ni ortopédicos ") se definió como criterio la no descontaminación nasal con agentes antimicrobianos tópicos frente al Staphylococcus aureus . Como indicador se estableció el Índice Global de Calidad (IC=Pacientes no descontaminados/Total pacientes intervenidos de procedimientos cardiacos y ortopédicos) expresado en porcentaje. Para la Recomendación 5 (" No se recomienda el reemplazo rutinario de catéteres venosos periféricos cada 72-96 horas ") se definieron como criterios el ser paciente portador de catéter venoso periférico y el no reemplazo rutinario del catéter en un periodo inferior a 72-96 horas, con excepción del recambio del mismo cada 72-96 horas por indicación médica. Como indicador se estableció el Índice Global de Calidad (IC=Pacientes sin recambio de catéter/Total pacientes portadores de catéteres venosos periféricos) expresado en porcentaje. Se estableció que el cumplimiento de las muestras debía ser elevado, idealmente cercano al 100%, por lo que se determinó que el estándar o valor que se consideraría como aceptable fuese del 100%. Se diseñó un estudio observacional, prospectivo y descriptivo para la evaluación del cumplimiento de cada recomendación. El estudio se realizó desde el 1 de diciembre de 2018 hasta el 31 de enero de 2019. Los pacientes que se estudiaron fueron incluidos según unos criterios diferentes para cada recomendación. Recomendación 1: "No eliminar el vello de forma sistemática para reducir el riesgo de infección de sitio quirúrgico". Se incluyeron los pacientes hospitalizados que fueron sometidos a intervención quirúrgica en el Hospital Universitario de La Princesa durante ese periodo. Se excluyeron los pacientes no hospitalizados y sometidos a cirugía mayor ambulatoria. De los 580 pacientes, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 231 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. Se emplearon el registro de intervenciones quirúrgicas, la historia clínica informatizada y el listado de preparación quirúrgica del paciente como fuentes de datos. Se elaboró una hoja de recogida de datos, con variables demográficas (edad y sexo) y variables del proceso (servicio, rasurado del sitio quirúrgico y procedimiento quirúrgico realizado). Recomendación 2: "No continuar con antibióticos más de 24-48 horas en pacientes hospitalizados, a menos que haya evidencia clara de infección". Se incluyeron los pacientes hospitalizados que fueron intervenidos en el hospital y que recibieron tratamiento antibiótico profiláctico. Se excluyeron los pacientes no hospitalizados que fueron sometidos a cirugía mayor ambulatoria y aquellos que no recibieron tratamiento antibiótico profiláctico. De los 421 pacientes, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 201 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. Se empleó la historia clínica informatizada como fuente de datos. Se realizó un seguimiento de los pacientes durante su ingreso hospitalario. Se elaboró una hoja de recogida de datos con variables demográficas (edad y sexo) y variables del proceso (fecha de intervención, procedimiento quirúrgico, clasificación de la cirugía -limpia, limpia/contaminada, contaminada, sucia/infectada-, profilaxis antibiótica empleada, duración de la profilaxis y signos de infección). Recomendación 3: "No se recomienda el análisis de la toxina Clostridium diffícile en pacientes asintomáticos" Se incluyeron las peticiones de análisis de la toxina Clostridium diffícile de pacientes pertenecientes al hospital y de centros de salud adscritos al laboratorio de Microbiología del hospital durante ese periodo. Se excluyeron los pacientes no sometidos a análisis de la toxina Clostridium diffícile . De 401 peticiones, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 196 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. La fuente de datos empleada fue el listado de los análisis de la toxina Clostridium diffícile realizados a los pacientes en el laboratorio de Microbiología durante el periodo establecido. Se elaboró una hoja de recogida de datos con variable demográficas (edad y sexo) y variables del proceso (centro de procedencia de la petición de análisis, servicio peticionario, fecha de petición y resultado del análisis de la toxina Clostridium difficile ). Recomendación 4: "No utilice la descontaminación nasal con agentes antimicrobianos tópicos destinados a eliminar el Staphylococcus aureus rutinariamente, para reducir el riesgo de infección del sitio quirúrgico, ni ante procedimientos cardíacos ni ortopédicos". Se incluyó a los pacientes sometidos a procedimientos cardíacos en el hospital y a pacientes sometidos a procedimientos ortopédicos en el hospital. Se excluyeron los pacientes sometidos a otro tipo de intervenciones quirúrgicas. De los 295 pacientes, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 167 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. Se empleó la historia clínica informatizada y el listado de preparación quirúrgica del paciente como fuentes de datos. Se elaboró una hoja de recogida de datos, con variables demográficas (edad y sexo) y variables del proceso (servicio y descontaminación nasal con agentes antimicrobianos). Recomendación 5: "No se recomienda el reemplazo rutinario de catéteres venosos periféricos cada 72-96 horas". Se incluyó a los pacientes hospitalizados en el hospital durante un periodo igual o superior a 72 horas que fuesen portadores de catéter venoso periférico. Se excluyeron los pacientes no hospitalizados, los pacientes hospitalizados durante un periodo inferior a 72 horas y aquellos no portadores de catéter venoso periférico. De los 255 pacientes, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 153 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. Se empleó la historia clínica informatizada como fuente de datos. Se realizó un seguimiento de los cuidados de los pacientes durante su ingreso hospitalario. Se elaboró una hoja de recogida de datos, con variables demográficas (edad y sexo) y variables del proceso (servicio hospitalario donde ingresa, periodo de ingreso, cuidados de enfermería de catéter venoso periférico, reemplazo del mismo e intervalo de reemplazo). Los datos obtenidos de cada paciente se tabularon en un archivo Excel y se elaboraron cinco bases de datos, una para la evaluación del cumplimiento de cada recomendación. El análisis de los datos se realizó con Excel 2011 y el programa estadístico Epi Info 7. Las variables cuantitativas se describieron mediante media y desviación estándar, y en el caso de las variables cualitativas se definieron por el número de casos y el porcentaje. Se consideró nivel de significación estadística valores de p<0,05. Con este método se obtuvo un Índice Global de Calidad (IC=Pacientes que cumplen criterio/Total de pacientes) para la unidad de estudio seleccionada, que permitió comprobar si se alcanzaron los estándares previamente definidos, es decir, el cumplimiento de las recomendaciones y, en caso de no ser así, analizar los posibles factores causales. Durante el desarrollo del estudio se respetaron los principios éticos básicos, enunciados en el Informe Belmont: de respeto por las personas, beneficencia, no maleficencia y justicia distributiva. Así mismo, se cumplieron los preceptos legales aplicables, contenidos en las siguientes leyes: Ley 41/2002, de 14 de noviembre, básica reguladora de la autonomía del paciente y de derechos y obligaciones en materia de información y documentación clínica; Ley 14/2007, de 3 julio, de Investigación biomédica; y Ley Orgánica 3/2018, de 5 de diciembre, de Protección de Datos Personales y garantías de los derechos digitales. El acceso a la información fue autorizado por el responsable del fichero y no se consideró necesario el obtener el consentimiento informado de cada paciente por tratarse de un estudio para valorar la calidad asistencial. Todo ello fue valorado por el Comité de Ética de la Investigación del Hospital Universitario de La Princesa, emitiendo un informe favorable al proyecto antes de iniciar su realización. Se incluyeron los pacientes hospitalizados que fueron sometidos a intervención quirúrgica en el Hospital Universitario de La Princesa durante ese periodo. Se excluyeron los pacientes no hospitalizados y sometidos a cirugía mayor ambulatoria. De los 580 pacientes, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 231 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. Se emplearon el registro de intervenciones quirúrgicas, la historia clínica informatizada y el listado de preparación quirúrgica del paciente como fuentes de datos. Se elaboró una hoja de recogida de datos, con variables demográficas (edad y sexo) y variables del proceso (servicio, rasurado del sitio quirúrgico y procedimiento quirúrgico realizado). Se incluyeron los pacientes hospitalizados que fueron intervenidos en el hospital y que recibieron tratamiento antibiótico profiláctico. Se excluyeron los pacientes no hospitalizados que fueron sometidos a cirugía mayor ambulatoria y aquellos que no recibieron tratamiento antibiótico profiláctico. De los 421 pacientes, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 201 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. Se empleó la historia clínica informatizada como fuente de datos. Se realizó un seguimiento de los pacientes durante su ingreso hospitalario. Se elaboró una hoja de recogida de datos con variables demográficas (edad y sexo) y variables del proceso (fecha de intervención, procedimiento quirúrgico, clasificación de la cirugía -limpia, limpia/contaminada, contaminada, sucia/infectada-, profilaxis antibiótica empleada, duración de la profilaxis y signos de infección). Clostridium diffícile en pacientes asintomáticos" Se incluyeron las peticiones de análisis de la toxina Clostridium diffícile de pacientes pertenecientes al hospital y de centros de salud adscritos al laboratorio de Microbiología del hospital durante ese periodo. Se excluyeron los pacientes no sometidos a análisis de la toxina Clostridium diffícile . De 401 peticiones, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 196 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. La fuente de datos empleada fue el listado de los análisis de la toxina Clostridium diffícile realizados a los pacientes en el laboratorio de Microbiología durante el periodo establecido. Se elaboró una hoja de recogida de datos con variable demográficas (edad y sexo) y variables del proceso (centro de procedencia de la petición de análisis, servicio peticionario, fecha de petición y resultado del análisis de la toxina Clostridium difficile ). Staphylococcus aureus rutinariamente, para reducir el riesgo de infección del sitio quirúrgico, ni ante procedimientos cardíacos ni ortopédicos". Se incluyó a los pacientes sometidos a procedimientos cardíacos en el hospital y a pacientes sometidos a procedimientos ortopédicos en el hospital. Se excluyeron los pacientes sometidos a otro tipo de intervenciones quirúrgicas. De los 295 pacientes, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 167 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. Se empleó la historia clínica informatizada y el listado de preparación quirúrgica del paciente como fuentes de datos. Se elaboró una hoja de recogida de datos, con variables demográficas (edad y sexo) y variables del proceso (servicio y descontaminación nasal con agentes antimicrobianos). Se incluyó a los pacientes hospitalizados en el hospital durante un periodo igual o superior a 72 horas que fuesen portadores de catéter venoso periférico. Se excluyeron los pacientes no hospitalizados, los pacientes hospitalizados durante un periodo inferior a 72 horas y aquellos no portadores de catéter venoso periférico. De los 255 pacientes, se hizo una estimación del tamaño muestral para garantizar un nivel de confianza del 95%, con una precisión del 5% y una incidencia esperada de 50%, siendo necesario el estudio de 153 intervenciones. Se obtuvo la muestra mediante muestreo aleatorio simple entre los pacientes incluidos. Se empleó la historia clínica informatizada como fuente de datos. Se realizó un seguimiento de los cuidados de los pacientes durante su ingreso hospitalario. Se elaboró una hoja de recogida de datos, con variables demográficas (edad y sexo) y variables del proceso (servicio hospitalario donde ingresa, periodo de ingreso, cuidados de enfermería de catéter venoso periférico, reemplazo del mismo e intervalo de reemplazo). Los datos obtenidos de cada paciente se tabularon en un archivo Excel y se elaboraron cinco bases de datos, una para la evaluación del cumplimiento de cada recomendación. El análisis de los datos se realizó con Excel 2011 y el programa estadístico Epi Info 7. Las variables cuantitativas se describieron mediante media y desviación estándar, y en el caso de las variables cualitativas se definieron por el número de casos y el porcentaje. Se consideró nivel de significación estadística valores de p<0,05. Con este método se obtuvo un Índice Global de Calidad (IC=Pacientes que cumplen criterio/Total de pacientes) para la unidad de estudio seleccionada, que permitió comprobar si se alcanzaron los estándares previamente definidos, es decir, el cumplimiento de las recomendaciones y, en caso de no ser así, analizar los posibles factores causales. Durante el desarrollo del estudio se respetaron los principios éticos básicos, enunciados en el Informe Belmont: de respeto por las personas, beneficencia, no maleficencia y justicia distributiva. Así mismo, se cumplieron los preceptos legales aplicables, contenidos en las siguientes leyes: Ley 41/2002, de 14 de noviembre, básica reguladora de la autonomía del paciente y de derechos y obligaciones en materia de información y documentación clínica; Ley 14/2007, de 3 julio, de Investigación biomédica; y Ley Orgánica 3/2018, de 5 de diciembre, de Protección de Datos Personales y garantías de los derechos digitales. El acceso a la información fue autorizado por el responsable del fichero y no se consideró necesario el obtener el consentimiento informado de cada paciente por tratarse de un estudio para valorar la calidad asistencial. Todo ello fue valorado por el Comité de Ética de la Investigación del Hospital Universitario de La Princesa, emitiendo un informe favorable al proyecto antes de iniciar su realización. Recomendación 1 Se incluyeron un total de 231 pacientes. El 63,6% de los pacientes intervenidos fueron hombres y el 36,4% mujeres. La edad media de los pacientes fue de 64,7 años (DE 16 años). La incidencia observada de pacientes que no fueron rasurados y que, por tanto, cumplían el criterio, fue del 83,55% (IC95%: 78,77-88,33%; 193 casos). En cuanto a los pacientes que sí fueron rasurados, se obtuvo una incidencia del 16,45% (IC95%: 11,67-20,46%; 38 casos), todos ellos varones. Estos pacientes fueron rasurados por indicación médica debido a que el vello interfería con la incisión quirúrgica, y en todos los casos se empleó el material adecuado. Por ello, se consideró que en estos casos no era exigible que se cumpliese el criterio. La mayor incidencia de pacientes rasurados se observó en el Servicio de Cardiología Hemodinámica (10,39% del total), con un total de 24 casos (63,2%). De ellos, 22 pacientes fueron sometidos a cateterismo cardíaco mientras que a 2 casos se les realizó una inserción de marcapasos permanente. En la se muestra la incidencia de la eliminación del vello en los servicios que realizaron intervenciones quirúrgicas durante el periodo estudiado. Se obtuvo un Índice Global de Calidad mediante la proporción de pacientes que cumplían el criterio previamente establecido de no eliminación del vello (193), incluidas las excepciones permitidas (38), con respecto al número total de pacientes incluidos, con un resultado de 100% (IC95%: 98,27-100%; 231 casos) . Por tanto, se comprobó el cumplimiento del estándar previamente definido y de la recomendación. Recomendación 2 En el estudio se incluyeron un total de 201 pacientes con una edad media de 63,7 años (DE 17,1 años). El 53,2% de los pacientes fueron hombres y el 46,8% mujeres. La incidencia observada de pacientes que no continuaron con tratamiento antibiótico más de 24-48 horas y que, por tanto, cumplían el criterio, fue del 81,59% (IC95%: 76,23-86,95%; 164 casos). De ellos, un total de 77 casos (un 46,95% de los pacientes que cumplían el criterio y un 38,31% del total) finalizaron el tratamiento en un periodo inferior a 24 horas tras la intervención quirúrgica, siendo 87 los pacientes que concluyeron la profilaxis antibiótica en un periodo inferior a 48 horas, lo que supone un 53,04% de los pacientes que cumplían el criterio y un 43,28% del total de pacientes incluidos. El porcentaje de pacientes que continuaron la profilaxis antibiótica por un periodo superior a 48 horas fue del 18,41%. Se observó una incidencia de infección del 4,98% (IC95%: 1,97-7,98%5; 10 casos), siendo el Servicio de Cirugía General y Digestivo donde se observó mayor incidencia, con un total de 4 casos, siendo dos de ellos intervenciones en el colon e ileostomías. Un 1,49% de los casos incluidos (3 pacientes) se encontraban en tratamiento antibiótico en el momento de la intervención quirúrgica y 11 pacientes (un 5,47%) fueron sometidos a cirugía maxilofacial y otorrinolaringológica, que fueron clasificadas como intervenciones sucias. Estos 24 casos, un 11,94% de los pacientes incluidos, fueron considerados excepciones en los que no era exigible el cumplimiento del criterio. Se obtuvo una incidencia de 6,47% (IC95%: 3,07-9,87%; 13 casos) de pacientes que no cumplían el criterio. En ellos, la profilaxis fue administrada por un periodo superior a 48 horas tras la finalización de la intervención quirúrgica, sin evidencia clara de infección ni otra causa que determinase la inadecuación de la duración de la profilaxis antibiótica. La mayor incidencia de incumplimiento del criterio se observó en el Servicio de Neurocirugía, con un 3,48% de los pacientes incluidos, resultando el 53,81% de los pacientes que incumplían el criterio. De ellos, los principales procedimientos realizados fueron dos craneotomías y dos abordajes transesfenoidales. Se determinó un Índice Global de Calidad del 93,53% (IC95%: 90,09-96,91; 188 casos). Se obtuvo de la proporción de pacientes que cumplían el criterio (164), incluyendo las excepciones aceptadas (24), en relación al número de pacientes incluidos en el estudio. Se comprobó un cumplimiento inferior al estándar previamente definido y, por tanto, una oportunidad de mejora. Recomendación 3 Se seleccionó una muestra de 200 peticiones de análisis de la toxina Clostridium diffícile a pacientes, donde el 50% fueron hombres y el 50% mujeres. La edad media de los pacientes incluidos en el estudio fue de 67,7 años (DE 19,9 años). El 79.50% de las peticiones fueron realizadas por el Hospital Universitario de La Princesa, siendo el Servicio de Urgencias el servicio que más peticiones realizó (33), seguido del Servicio de Medicina Interna (25). Un 12% de las peticiones procedían de las consultas de Atención Primaria. Finalmente, las consultas del hospital realizaron un 8,5% de las peticiones, procedentes en un 3,5% (7 peticiones) de las consultas externas del Servicio de Digestivo. La petición de análisis de la toxina se solicitó en 187 pacientes (93,5%), en los que estaba indicada por presentar sintomatología (heces no formes). En el 3,5% de las peticiones indicadas se obtuvo un resultado positivo (7 casos), siendo negativo en el 90% de las mismas (180 casos). Se obtuvo una incidencia de peticiones que incumplían el criterio por no estar indicadas del 6,5% (IC95%: 3,08-9,92%; 13 peticiones). En dos casos no estaba indicada la petición por heces formes (1%), peticiones realizadas por las consultas externas de Hematología (1) y Urgencias (1). Un 4,5% de las peticiones realizadas (9 casos) se recibieron de forma duplicada y en un 1% de las peticiones (2 casos) se produjo un error en la recepción de la muestra . El Índice Global de Calidad dio como resultado un 93,5% (IC95%: 90,08-96,92%; 187 casos). Se determinó de la proporción de pacientes que cumplían el criterio de petición del análisis de la toxina en pacientes indicados (187), en relación al total de peticiones realizadas. Se comprobó un cumplimiento inferior al estándar previamente definido y, por tanto, una oportunidad de mejora. Recomendación 4 Se incluyeron 167 pacientes. En el protocolo de preparación quirúrgica del Hospital Universitario de La Princesa no constaba la descontaminación nasal como medida preoperatoria, y en la historia clínica informatizada no se obtuvo ningún dato al respecto. Se consultó a loas supervisoras de quirófano y al servicio de farmacia. Confirmaron que no se utilizaba esta profilaxis tópica. Por ello, podría asumirse un cumplimiento del 100% (IC95%: 97,6-100%). Recomendación 5 Se estudiaron un total de 153 pacientes. El 41,2% de los pacientes incluidos fueron mujeres y el 58,8% hombres. La edad media fue de 67,1 años (DE 16,5 años). La incidencia observada de pacientes que portaban catéter venoso periférico a los que no se le reemplazó de forma rutinaria en un periodo inferior a 72 horas y que, por tanto, cumplían el criterio, fue del 83% (127 pacientes). En un 66% de los pacientes (101) no se les realizó ningún reemplazo del catéter que portaban, observándose la mayor incidencia en el Servicio de Traumatología (16,34%). Se obtuvo una incidencia de reemplazo en un periodo superior a 72-96 horas del 17%. Las principales causas fueron extravasación (7,84%) y flebitis (3,92%). El reemplazo de catéteres se realizó en un periodo inferior a 72 horas en 23 pacientes (15,03%) por indicación médica. La principal causa fue la extravasación, con un 11,11% de los casos (17 pacientes). Estos fueron considerados como excepciones. El porcentaje de pacientes que incumplieron el criterio fue del 1,96% (3), a los que se realizó reemplazo del catéter de forma rutinaria cada aproximadamente 72 horas, sin precisar indicación ni constar motivo del reemplazo en la historia clínica. Los servicios donde se observó fueron el Servicio de Traumatología (1,31%) y el Servicio de Angiología y Cirugía Vascular (0,65%). Se obtuvo un Índice Global de Calidad mediante la proporción de pacientes que cumplían el criterio (127) y los pacientes considerados como excepciones permitidas (23), con respecto al número total de pacientes incluidos, con un resultado del 98,04% (IC95%: 94,12-99,35%). Se comprobó un cumplimiento inferior al estándar previamente definido y, por tanto, una oportunidad de mejora. Se incluyeron un total de 231 pacientes. El 63,6% de los pacientes intervenidos fueron hombres y el 36,4% mujeres. La edad media de los pacientes fue de 64,7 años (DE 16 años). La incidencia observada de pacientes que no fueron rasurados y que, por tanto, cumplían el criterio, fue del 83,55% (IC95%: 78,77-88,33%; 193 casos). En cuanto a los pacientes que sí fueron rasurados, se obtuvo una incidencia del 16,45% (IC95%: 11,67-20,46%; 38 casos), todos ellos varones. Estos pacientes fueron rasurados por indicación médica debido a que el vello interfería con la incisión quirúrgica, y en todos los casos se empleó el material adecuado. Por ello, se consideró que en estos casos no era exigible que se cumpliese el criterio. La mayor incidencia de pacientes rasurados se observó en el Servicio de Cardiología Hemodinámica (10,39% del total), con un total de 24 casos (63,2%). De ellos, 22 pacientes fueron sometidos a cateterismo cardíaco mientras que a 2 casos se les realizó una inserción de marcapasos permanente. En la se muestra la incidencia de la eliminación del vello en los servicios que realizaron intervenciones quirúrgicas durante el periodo estudiado. Se obtuvo un Índice Global de Calidad mediante la proporción de pacientes que cumplían el criterio previamente establecido de no eliminación del vello (193), incluidas las excepciones permitidas (38), con respecto al número total de pacientes incluidos, con un resultado de 100% (IC95%: 98,27-100%; 231 casos) . Por tanto, se comprobó el cumplimiento del estándar previamente definido y de la recomendación. En el estudio se incluyeron un total de 201 pacientes con una edad media de 63,7 años (DE 17,1 años). El 53,2% de los pacientes fueron hombres y el 46,8% mujeres. La incidencia observada de pacientes que no continuaron con tratamiento antibiótico más de 24-48 horas y que, por tanto, cumplían el criterio, fue del 81,59% (IC95%: 76,23-86,95%; 164 casos). De ellos, un total de 77 casos (un 46,95% de los pacientes que cumplían el criterio y un 38,31% del total) finalizaron el tratamiento en un periodo inferior a 24 horas tras la intervención quirúrgica, siendo 87 los pacientes que concluyeron la profilaxis antibiótica en un periodo inferior a 48 horas, lo que supone un 53,04% de los pacientes que cumplían el criterio y un 43,28% del total de pacientes incluidos. El porcentaje de pacientes que continuaron la profilaxis antibiótica por un periodo superior a 48 horas fue del 18,41%. Se observó una incidencia de infección del 4,98% (IC95%: 1,97-7,98%5; 10 casos), siendo el Servicio de Cirugía General y Digestivo donde se observó mayor incidencia, con un total de 4 casos, siendo dos de ellos intervenciones en el colon e ileostomías. Un 1,49% de los casos incluidos (3 pacientes) se encontraban en tratamiento antibiótico en el momento de la intervención quirúrgica y 11 pacientes (un 5,47%) fueron sometidos a cirugía maxilofacial y otorrinolaringológica, que fueron clasificadas como intervenciones sucias. Estos 24 casos, un 11,94% de los pacientes incluidos, fueron considerados excepciones en los que no era exigible el cumplimiento del criterio. Se obtuvo una incidencia de 6,47% (IC95%: 3,07-9,87%; 13 casos) de pacientes que no cumplían el criterio. En ellos, la profilaxis fue administrada por un periodo superior a 48 horas tras la finalización de la intervención quirúrgica, sin evidencia clara de infección ni otra causa que determinase la inadecuación de la duración de la profilaxis antibiótica. La mayor incidencia de incumplimiento del criterio se observó en el Servicio de Neurocirugía, con un 3,48% de los pacientes incluidos, resultando el 53,81% de los pacientes que incumplían el criterio. De ellos, los principales procedimientos realizados fueron dos craneotomías y dos abordajes transesfenoidales. Se determinó un Índice Global de Calidad del 93,53% (IC95%: 90,09-96,91; 188 casos). Se obtuvo de la proporción de pacientes que cumplían el criterio (164), incluyendo las excepciones aceptadas (24), en relación al número de pacientes incluidos en el estudio. Se comprobó un cumplimiento inferior al estándar previamente definido y, por tanto, una oportunidad de mejora. Se seleccionó una muestra de 200 peticiones de análisis de la toxina Clostridium diffícile a pacientes, donde el 50% fueron hombres y el 50% mujeres. La edad media de los pacientes incluidos en el estudio fue de 67,7 años (DE 19,9 años). El 79.50% de las peticiones fueron realizadas por el Hospital Universitario de La Princesa, siendo el Servicio de Urgencias el servicio que más peticiones realizó (33), seguido del Servicio de Medicina Interna (25). Un 12% de las peticiones procedían de las consultas de Atención Primaria. Finalmente, las consultas del hospital realizaron un 8,5% de las peticiones, procedentes en un 3,5% (7 peticiones) de las consultas externas del Servicio de Digestivo. La petición de análisis de la toxina se solicitó en 187 pacientes (93,5%), en los que estaba indicada por presentar sintomatología (heces no formes). En el 3,5% de las peticiones indicadas se obtuvo un resultado positivo (7 casos), siendo negativo en el 90% de las mismas (180 casos). Se obtuvo una incidencia de peticiones que incumplían el criterio por no estar indicadas del 6,5% (IC95%: 3,08-9,92%; 13 peticiones). En dos casos no estaba indicada la petición por heces formes (1%), peticiones realizadas por las consultas externas de Hematología (1) y Urgencias (1). Un 4,5% de las peticiones realizadas (9 casos) se recibieron de forma duplicada y en un 1% de las peticiones (2 casos) se produjo un error en la recepción de la muestra . El Índice Global de Calidad dio como resultado un 93,5% (IC95%: 90,08-96,92%; 187 casos). Se determinó de la proporción de pacientes que cumplían el criterio de petición del análisis de la toxina en pacientes indicados (187), en relación al total de peticiones realizadas. Se comprobó un cumplimiento inferior al estándar previamente definido y, por tanto, una oportunidad de mejora. Se incluyeron 167 pacientes. En el protocolo de preparación quirúrgica del Hospital Universitario de La Princesa no constaba la descontaminación nasal como medida preoperatoria, y en la historia clínica informatizada no se obtuvo ningún dato al respecto. Se consultó a loas supervisoras de quirófano y al servicio de farmacia. Confirmaron que no se utilizaba esta profilaxis tópica. Por ello, podría asumirse un cumplimiento del 100% (IC95%: 97,6-100%). Se estudiaron un total de 153 pacientes. El 41,2% de los pacientes incluidos fueron mujeres y el 58,8% hombres. La edad media fue de 67,1 años (DE 16,5 años). La incidencia observada de pacientes que portaban catéter venoso periférico a los que no se le reemplazó de forma rutinaria en un periodo inferior a 72 horas y que, por tanto, cumplían el criterio, fue del 83% (127 pacientes). En un 66% de los pacientes (101) no se les realizó ningún reemplazo del catéter que portaban, observándose la mayor incidencia en el Servicio de Traumatología (16,34%). Se obtuvo una incidencia de reemplazo en un periodo superior a 72-96 horas del 17%. Las principales causas fueron extravasación (7,84%) y flebitis (3,92%). El reemplazo de catéteres se realizó en un periodo inferior a 72 horas en 23 pacientes (15,03%) por indicación médica. La principal causa fue la extravasación, con un 11,11% de los casos (17 pacientes). Estos fueron considerados como excepciones. El porcentaje de pacientes que incumplieron el criterio fue del 1,96% (3), a los que se realizó reemplazo del catéter de forma rutinaria cada aproximadamente 72 horas, sin precisar indicación ni constar motivo del reemplazo en la historia clínica. Los servicios donde se observó fueron el Servicio de Traumatología (1,31%) y el Servicio de Angiología y Cirugía Vascular (0,65%). Se obtuvo un Índice Global de Calidad mediante la proporción de pacientes que cumplían el criterio (127) y los pacientes considerados como excepciones permitidas (23), con respecto al número total de pacientes incluidos, con un resultado del 98,04% (IC95%: 94,12-99,35%). Se comprobó un cumplimiento inferior al estándar previamente definido y, por tanto, una oportunidad de mejora. El proyecto "No hacer" propone reducir las intervenciones sanitarias no coste-efectivas, así como las de dudosa o nula eficacia y efectividad, promoviendo la colaboración de las Sociedades Científicas para conseguir una mejora continua en la calidad asistencial . La SEMPSPH se adhirió en 2017 al proyecto de "No hacer" y presentó en 2018 las 5 medidas. Por el corto período transcurrido entre la presentación de las medidas por la Sociedad y este estudio que las evalúa, no se han encontrado datos de referencia para comparar los resultados obtenidos en España. Así, con respecto a la eliminación del vello, que hasta el momento ha sido empleada como medida de preparación prequirúrgica, se demuestra no solo la ausencia de beneficio de esta medida en la prevención de infección de localización quirúrgica (ILQ), sino un incremento del riesgo de ILQ debido a las erosiones producidas por el rasurado . No se encuentran diferencias significativas en la incidencia de ILQ entre la ausencia de rasurado y el empleo de maquinilla eléctrica, cortadora de pelo o depilación química , por lo que estas técnicas se pueden seleccionar en caso de considerarse imprescindible la eliminación del vello por interferencia con la localización quirúrgica (NICE, 2008). Esta evidencia queda recogida en la recomendación 1, en las guías de buenas prácticas de la Comunidad de Madrid y, así mismo, en el Protocolo de Preparación Quirúrgica del Hospital Universitario de La Princesa. El cumplimiento de la recomendación al 100% implica que en ningún caso se ha eliminado el vello de forma innecesaria, reduciendo la potencial iatrogenia. Esto es reflejo de una adecuada difusión del protocolo y conduce a una mejora de la calidad asistencial y seguridad del paciente. Por otro lado, existen evidencias de que los pacientes que son sometidos de forma innecesaria o inapropiada a antibióticos presentan riesgo de efectos secundarios graves sin beneficio clínico alguno . Su uso incorrecto contribuye al aumento de resistencias a los antibióticos, un serio problema para la salud pública que se ha incrementado durante las últimas décadas . En el estudio, el cumplimiento de la Recomendación 2, aunque cercano al 100%, es incompleto (93,53%). Esto supone unos pocos pacientes (13 casos) en los que se incumple la recomendación en un periodo de dos meses. Sin embargo, este número es mayor en las 5.467 intervenciones realizadas durante todo el año 2018. Entre los posibles factores causales se puede encontrar la incorrecta difusión de la evidencia científica o la variabilidad existente en la práctica clínica entre los diferentes servicios. Se demuestra que los programas hospitalarios dedicados a promover el uso correcto de los antibióticos optimizan tanto el tratamiento de las infecciones como reducen los efectos adversos, mejorando la calidad asistencial . La infección por Clostridium diffícile parece estar cambiando, siendo mayor la incidencia y la virulencia de los casos. El diagnóstico preciso es crítico y, sin embargo, existen intervenciones que conducen a un diagnóstico erróneo . Una de ellas, recogida en la Recomendación 3, es el análisis de la toxina Clostridium diffícile en pacientes asintomáticos. Los resultados pueden ser falsamente positivos, lo que supone sobrediagnóstico y sobretratamiento . El cumplimiento en el estudio fue del 93,5%. Sin embargo, a pesar de no ser completo, el laboratorio de Microbiología no realizó el análisis en aquellas muestras en que no estaba indicado y limitó intervenciones sanitarias innecesarias. Esto supone un diagnóstico más preciso, con la reducción de peticiones en pacientes sin diarrea o con solo un episodio, y otorgando más importancia a elementos clave en la historia clínica del paciente . Resaltar que en el 90% de las peticiones indicadas, el resultado es negativo. Por otra parte, en 200 peticiones, lo que supone una mínima cantidad de peticiones inadecuadas (13), se incrementa en relación al total de peticiones realizadas en el año, alcanzando un número considerable de errores. Las causas principales de incumplimiento son las muestras duplicadas, lo que refleja un consumo innecesario de los recursos sanitarios. El principal factor causal podría ser la falta de coordinación entre los miembros del equipo médico. En relación a la Recomendación 4, las guías NICE determinan el empleo de mupirocina nasal en aquellos casos en los que sea causa de ILQ, teniendo en cuenta el tipo de procedimiento, los factores de riesgo del paciente y el potencial impacto de la infección, vigilando la resistencia antimicrobiana asociada al empleo de mupirocina . En el Hospital Universitario de La Princesa, entre las medidas incluidas en el Protocolo de Preparación Quirúrgica no se encuentra el empleo de mupirocina nasal. Esto se debe a su actualización en base a la evidencia científica disponible realizada por varios servicios del hospital. Supondría una acción de mejora especificar en el protocolo la no realización de esta medida, con el fin de facilitar el cumplimiento y la recogida de información en estudios posteriores. Por otra parte, son muchos los pacientes que durante su ingreso hospitalario reciben medicación, fluidos o nutrientes vía endovenosa por medio de catéter venoso periférico. Estos catéteres son a menudo reemplazados cada 72 o 96 horas para prevenir infecciones o molestias. Sin embargo, esta práctica incrementa el coste sanitario y supone someter a los pacientes de forma repetida a un procedimiento invasivo . Varios análisis de coste-efectividad realizados concluyen que el reemplazo de catéter por indicación médica reduce el coste comparado con el reemplazo rutinario . Por ello, aunque el cumplimiento ideal de la Recomendación 5 fuese del 100%, un cumplimiento cercano del 98% mediante el reemplazo en caso de indicación médica supone una medida de eficiencia clínica que no repercute en los resultados de salud obtenidos y sí en la calidad de la asistencia, ya que contribuye a la sostenibilidad y mejora del sistema sanitario, así como del hospital. Globalmente, un cumplimiento cercano al 100% de las recomendaciones evaluadas supone una adecuada difusión de la evidencia científica y el compromiso con el uso apropiado de los recursos sanitarios en el Hospital Universitario de La Princesa. Se cumple el principal objetivo del proyecto "No hacer", limitando así las intervenciones sanitarias innecesarias, lo que promueve la seguridad clínica y la calidad asistencial. Los buenos resultados en las Recomendaciones 1 y 2 están asociados a la actualización y difusión del Protocolo de Preparación Quirúrgica del hospital. Sin embargo, aún hay margen de mejora. La difusión de la evidencia y de los resultados obtenidos en el hospital supone una medida correctora que facilita su cumplimiento y permite sensibilizar a los profesionales sanitarios en el compromiso con la calidad en la asistencia sanitaria y en el uso eficiente de los recursos. Las limitaciones de este estudio son las propias de los estudios observacionales, y también están relacionadas con el corto tiempo analizado. En cuanto a la recomendación 4, la ausencia de información al respecto no permitió realizar un análisis de la misma. De la información y resultados del estudio se puede concluir que: - La realización de estudios de calidad asistencial y la metodología empleada para su evaluación son útiles y aportan información relevante. - Hay recomendaciones (R1 y R4) en las que el cumplimiento de "No hacer" es del 100%. - En las otras recomendaciones (R2, R3 y R5), el cumplimiento es elevado pero no alcanza el 100%, siendo por tanto áreas a mejorar. En conjunto, se obtiene un cumplimiento elevado de las recomendaciones "No hacer" en el Hospital Universitario de La Princesa. No obstante, no se pueden comparar los resultados por falta de publicación de estudios similares. Es aconsejable la realización periódica de este tipo de estudios, así como la difusión de sus resultados, para facilitar el cumplimiento. El fin es mejorar en aquellas recomendaciones que no alcanzan el 100% del mismo y mantenerlo a un nivel adecuado en aquellas que sí lo hacen. |
A Phase I Clinical Trial of Intrahepatic Artery Delivery of TG6002 in Combination with Oral 5-Fluorocytosine in Patients with Liver-Dominant Metastatic Colorectal Cancer | 953bd110-da5d-40ee-b9d3-f01d0eb9a24d | 11959272 | Neoplasms[mh] | Colorectal cancer (CRC) is a leading cause of cancer-associated deaths in Western populations and the third most frequent cause of cancer-related deaths worldwide ( ). The 5-year survival rate for localized disease is approximately 91%; however, around 25% to 30% of patients with colorectal cancer develop liver metastases ( ), which is associated with a 5-year survival rate of only 13% ( ). For these patients, systemic anticancer therapy is the mainstay of treatment, with 5-fluorouracil (5-FU) being commonly employed either as monotherapy or in combination with other cytotoxics. However, 5-FU has limitations including intravenous (i.v.) administration, short half-life, significant systemic toxicity, and drug resistance ( ). For patients with liver-dominant metastatic colorectal cancer (mCRC), locoregional therapies offer the prospect of effective treatment while limiting systemic toxicity. Oncolytic viruses (OVs) are principally immunotherapeutic agents that preferentially replicate in malignant cells, ultimately inducing immunogenic cell death (ICD). OVs can be engineered to express transgenes with immune-stimulating functions or highly specific downstream targets ( ). Many engineered OVs have been evaluated in randomized trials, with three currently licensed as standard care ( ). One virus that has been tested extensively in the clinical setting is pexastimogene devacirepvec ( Pexa- Vec ; JX-594, TG6006), an engineered Wyeth-strain vaccinia virus ( ) developed by Transgene and Sillagen. Clinical efficacy as a single agent by intratumoral (i.t.) injection was observed in a dose comparison, randomized study in patients with hepatocellular carcinoma (HCC), in which overall survival was significantly longer for patients in the high-dose group ( ). Furthermore, i.v. delivery to the tumor is also possible at a dose of 1 × 10 9 plaque-forming units (pfu; ref. ). TG6002 was developed by engineering the highly oncolytic Copenhagen vaccinia strain ( , ), incorporating gene modifications to enhance its antitumor activity and clinical impact. Thymidine kinase and ribonucleotide reductase genes are deleted in TG6002, enhancing selective replication in cancer cells ( ). In addition, the insertion of the chimeric yeast FCU1 gene enables the selective conversion of the prodrug 5-fluorocytosine (5-FC) into the cytotoxic 5-FU and 5-fluorouridine monophosphate ( ), bypassing the natural resistance of tumor cells to 5-FU alone and reducing systemic toxicity ( ). Moreover, TG6002 induces an antitumor immune response involving CD8 T cells and tumor-infiltrating lymphocytes and myeloid cells ( ). TG6002 with 5-FC is a promising combination therapy for cancers that are sensitive to 5-FU. An open-label phase I dose-escalation trial of i.v. TG6002 plus 5-FC was initiated in 2018 (TG6002.02; NCT03724071) in patients with advanced gastrointestinal malignancies. Overall, the combination was well tolerated, and no maximum-tolerated dose (MTD) was observed. Preliminary results indicate effective biodistribution of TG6002 in tumor cells, associated with localized FCU1 activity ( ). Critical to successful OV therapy is the delivery of the virus to the tumor site. Various routes of administration have been investigated, predominantly i.t. and i.v. Intravenous administration is simple and relatively noninvasive and can achieve systemic delivery of the virus to all vascularized tumors although very high doses are required to achieve sufficient concentration at the tumor site, as the majority is redistributed throughout normal body organs. Intratumoral administration delivers the virus directly to the tumor; however, i.t. injection is limited to radiologically detectable, anatomically and technically injectable lesions although abscopal effects have been reported at distant sites. For patients with liver-dominant cancers, an alternative route of delivery is via selective catheterization of the hepatic artery; indeed, locoregional delivery of chemotherapy via the hepatic artery has been extensively studied ( , ) and was initially considered for administration of OVs in the early 2000s ( ). Intrahepatic artery (IHA) infusion of OVs has the potential to enhance delivery and distribution to multiple tumors across the liver while limiting systemic chemotherapy toxicity. We describe the clinical and translational data from a dose-escalation study of TG6002 via IHA administration plus oral 5-FC. The results show that IHA administration of an OV is clinically achievable and results in the delivery of replication-competent virus to the tumor, expression and clinically relevant activity of the FCU1 transgene, peripheral activation of the immune system, and potential ICD.
Study design TG6002.03 was an open-label, dose-escalation, 3 + 3 design, phase I study (Eudra-CT 2018-004103-39) conducted in three sites across the UK and France in patients with unresectable colorectal cancer with liver metastases having progressed on or after standard chemotherapy, including at least a fluoropyrimidine, oxaliplatin, and irinotecan or, in the UK only, entering a period of clinical observation following discontinuation of chemotherapy. Patients received up to two cycles of TG6002 combined with oral 5-FC ( ). TG6002 was administered via the main hepatic artery, through a catheter inserted into the femoral artery under angiographic assessment, for more than 30 minutes at doses of 1 × 10 6 , 1 × 10 7 , 1 × 10 8 , and 1 × 10 9 pfu. 5-FC was taken orally from days 5 to 14 at a dose of 50 mg/kg four times daily. A second treatment cycle was to be administered from day 43 in the absence of disease progression or unacceptable toxicity. Dose escalation proceeded after a review of safety data from each cohort by an independent safety review committee. The clinical trial protocol was approved by institutional ethics committees and conducted in accordance with the Declaration of Helsinki. All patients signed an informed consent document prior to study participation. Patient samples Blood and tissue samples were collected and processed using the Translational Cancer Immunotherapy Team quality-assured lab manual, which included standard operating procedures to regulate all processes. Peripheral blood was collected into tripotassium ethylene-diaminetetraacetic acid (K3EDTA) or serum clot-activator vacutainer tubes (both Scientific Laboratory Supplies) and processed within 2 hours of venepuncture, or as soon as possible thereafter. Tumor biopsies collected on day 4 or 8 were placed in formalin or RNAlater (both ThermoFisher) for IHC or PCR analyses, respectively. All sample collection time points are shown in . Isolation of PBMCs, plasma, and serum from whole blood Serum clot-activator tubes were left for a minimum of 30 minutes after venepuncture. All blood collection tubes were centrifuged for 10 minutes at 2,000 g. Plasma and serum aliquots from the upper layers were stored at −80°C. Peripheral blood mononuclear cells (PBMCs) were isolated by density-gradient centrifugation over Lymphoprep (STEMCELL Technologies) as per the manufacturer’s instructions. Cells were frozen at 1 × 10 7 /mL in 40% (v/v) Roswell Park Memorial Institute medium containing 5 mmol/L L-glutamine and 1 mmol/L sodium pyruvate (all Sigma), plus 50% (v/v) pooled human serum (SeraLab) and 10% (v/v) dimethyl sulfoxide (Sigma). PBMCs were stored in liquid nitrogen. Detection of virus in tumor biopsies and plasma qPCR DNA was extracted from tumor biopsies using DNeasy Blood and Tissue Kits (QIAGEN). Primers (forward 5′-CGATGATGGAGTAATAAGTGGTAGGA-3′ and reverse 5′-CACCGACCGATGATAAGATTTG-3′; Integrated DNA Technologies) were used to detect the presence of TG6002. qRT-PCR qRT-PCR was performed on tumor biopsies and plasma for the detection of viral early D7R , viral late A10L , and FCU1 transcripts. RNA was extracted using RNeasy Plus Mini Kits (QIAGEN). The remaining viral and cellular DNA in samples was digested with TURBO DNase (ThermoFisher). Primers and probes for D7R (forward TTTAGCGATTCAAAGTACTGCTTTTT, reverse GCAGTGACTTCGCTGCCATT, and probe FAM-CGAAATGGTAATGCGTATGA), A10L (forward CTTCATACTCGCGATCCTCAAA, reverse TCGCCAACAGGTTAAAGAAATTAA, and probe ABY-TGGCGCTTCCAAACGTGCAATTT), and FCU1 (forward TCGTGGTCACAACATGAGATTTC, reverse TCTAATCTCCCACAGTTTTCCAAAG, and probe ABY-TCCGCCACACTACATGGTGAGATCTCC) were used. Detection of the TG6002 viral genome in plasma was performed by Charles River Laboratories, Evreux, using in-house methods. All PCR data were acquired on Applied Biosystems QuantStudio™ 5 Real-Time PCR Systems (ThermoFisher; RRID:SCR_020240) and analyzed using QuantStudio 3D AnalysisSuite Cloud (ThermoFisher; RRID:SCR_020238). The presence of the virus in tumor biopsies was considered positive by RT-qPCR if at least one of the viral mRNAs ( D7R , A10L , or FCU1 ) was detected. RNase-free water was used as a negative control. Plaque assays were performed by Transgene, France, using in-house methods. Briefly, tissue biopsies were sonicated for 15 seconds at room temperature before incubation with a permissive cell line, Vero (CCL-81; ATCC; RRID:CVCL_0059). Veros (passage number 128 at thawing) were cultured in Dulbecco's modified Eagle's medium (DMEM; Sigma) supplemented with 10% (v/v) fetal calf serum (FCS; ThermoFisher) and 40 mg/L gentamicin (Sigma) for two passages prior to use in plaques assays. Confluent monolayers were incubated with tumor samples for 30 minutes prior to incubation at 37°C under 1% (w/v) agarose for 3 days. Positive infection was determined by the presence of viral plaques. Veros were authenticated in 2014 (by qPCR and epifluorescence microscopy) by Clean Cells and tested negative for Mycoplasma infection. IHC for viral protein was performed by Cerba Research. A polyclonal anti-vaccinia virus antibody (Meridian Life Science; RRID:AB_153134) was used to detect virions; negative control was secondary antibody alone. 3, 3-diaminobenzidine-horseradish peroxidase (DAB-HRP) was used to visualize virus-positive cells. Data from all methods are expressed as positive or negative/below the level of detection. Biopsies from two patients (13 and 15) were not available. 5-FU concentrations in serum and tumor tissue Quantification of 5-FC, 5-FU, and 5-fluoro-β-alanine (F-BAL) levels was performed using liquid chromatography coupled with high-resolution mass spectrometry (Hospices Civils de Lyon) as described previously ( ). Serum was evaluated on day 8 after TG6002. Additionally, 5-FU concentrations were measured in available tumor biopsies on day 8 post-TG6002. F-BAL was not measured in cohort 1 serum samples or tumor biopsies due to sample insufficiency. Calreticulin ELISA Patient plasma was analyzed for calreticulin (CRT) by ELISA (ThermoFisher) as per the manufacturer’s instructions. Data are expressed as mean plasma concentration (ng/mL) ± SEM, calculated using a standard curve. Statistical significance was determined using paired two-tailed t tests (GraphPad Prism; RRID:SCR_002798) between sample time points (* P < 0.05; n = 14 patients), dependent on sample availability. mRNA expression analysis of patient PBMCs mRNA sequencing was performed by Novogene as per validated methods. After PCR, the gene expression level was calculated by the number of mapped reads. Statistically significant differentially expressed genes (ssDEG) were defined as ± >2 log 2 fold change of post-treatment samples compared with baseline with an associated P adj < 0.05. Data from six patients, across three cohorts, are shown for ssDEGs for C1D2 and C1D15. P adj values were transformed into −log 10 ( P adj) values, which were plotted against log 2 fold change values in volcano plots. Volcano plots depict all DEGs, not just ssDEGs, which are upregulated, downregulated, or unchanged for three patients at C1D2 and C1D15 compared with BS. ssDEG lists were analyzed using the Search Tool for Retrieval of Interacting Genes/Proteins (STRING) database ( http://string-db.org ; RRID:SCR_005223) to identify potential interactions between the genes and their reported biological function(s). Interaction confidence scores (ICS) were assigned to each protein association and ranked from 0 to 1, in which 1 is most likely to be accurate and 0 is least likely to be correct. An ICS of 0.5 indicates that every second interaction may be a false positive; therefore, only ssDEGs with an ICS of >0.5 were used for visualization and analysis. Gene Ontology (GO) biological functions were assessed at the ssDEG level, in which common genes were detected across multiple patients. These nine commonly expressed ssDEGs were analyzed independently in the STRING database to identify highly responsive signaling pathways following treatment. Immunophenotyping PBMCs were analyzed for specific activation/immune checkpoint molecules. Briefly, PBMCs were stained for CD3 (HIT3a/FITC; RRID:AB_395745), CD4 (RPA-T4/APC-H7; RRID:AB_1645478), CD8 (RPA-T8/Alexa 700; RRID:AB_396953), CD56 (B159/PE-Cy7; RRID:AB_396853), CD19 (SJ25C1/APC-H7; RRID:AB_1645470), CD14 (M5E2/FITC; RRID:AB_395798), CD69 (FN50/APC; RRID:AB_398602), programmed cell death-ligand 1 (PD-L1; MIH1/PE-CF594; RRID:AB_2738400), programmed cell death protein 1 (PD-1; MIH4/PE; RRID:AB_647199), T-cell immunoglobulin and mucin-domain containing-3 (TIM-3; 7D3/BV786; RRID:AB_2741100), OX40 (L106/BV421; RRID:AB_2742558), CD40 (5C3/BUV395; RRID:AB_2739110), and CD25 (M-A251/BB700; RRID:AB_2744335; all BD Biosciences) plus CD40 ligand (L) (24-31/BV786; RRID:AB_2572187) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4; BNI3/BV605; RRID:AB_2632779; both BioLegend). Fluorescence minus one (FMO) was used as a negative control. Data were acquired on a CytoFLEX LX and analyzed using CytExpert software (RRID:SCR_017217) (both Beckman Coulter). Positive expression of markers was used to calculate fold-change differences ± SEM in expression from baseline samples. Paired two-tailed t tests (GraphPad Prism) were used to determine statistical significance between samples (* P < 0.05; n = 9 patients), dependent on sample availability. IHC for PD-L1 Formalin-fixed, paraffin-embedded tissue biopsies were stained with rabbit anti-human PD-L1 antibody (1:500; Abcam; RRID:AB_2884993) and ImmPRESS HRP anti-rabbit IgG (peroxidase) secondary antibody (Vector Laboratories; RRID:AB_2336529). Positive staining was visualized using ImmPACT DAB Peroxidase (HRP) Substrate kit (Vector Laboratories). Control sections contained no primary antibody. Digital images were acquired at 20× magnification and quantified using QuPath software (RRID:SCR_018257). Data are expressed as cells positive for PD-L1 per mm 2 for n = 7 patients. Enzyme-linked immunosorbent spot (ELISpot) Briefly, triplicates of 1 × 10 5 /well PBMCs were incubated in the presence of either 2 μg/mL of overlapping peptide pools for carcinoembryonic antigen (CEA) or cytomegalovirus/Epstein-Barr virus/influenza (CEF; positive control) (both Cambridge Biosciences). Negative control was media alone; 10 pfu/cell TG6002 was used to assess response to treatment. IFNγ secretion from activated T cells was detected using a matched paired antibody kit (Mabtech). Spot-forming units (SFU) were visualized using 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium substrate (MabTech). Images were captured and quantified using an S6 FluoroSpot analyzer (Cellular Technology Limited). Data are presented as mean fold-change SFU per well ± SEM for post-treatment samples compared with baseline, for n = 6 patients, dependent on sample availability. TCR-β sequencing T-cell receptor (TCR) β sequencing was performed by Adaptive Technologies using a “survey” resolution to generate data from productive rearrangements only, which were exported from the immunoSEQ Analyzer (Adaptive Technologies) for further analysis. Complementary-determining region 3 (CDR3) sequences were input into the McPAS-TCR database (RRID:SCR_026024) and matched to known human TCR sequences. TCRs in each patient sample that matched known cancer antigen or neoantigen epitopes were identified; these were counted and their total productive frequency was calculated. Data are expressed as the number of TCRs matching cancer antigens/neoantigens ( x -axis) against the productive frequency of TCRs matching cancer antigens/neoantigens ( y -axis) for all available PBMC samples. Data availability The mRNA sequencing FASTQ files are available in the Sequence Read Archive database (accession number PRJNA1192197). TCR sequencing data can be accessed via the immuneACCESS server using the following link: https://doi.org/10.21417/EW2024S . All data generated in this study are available upon request from the corresponding author.
TG6002.03 was an open-label, dose-escalation, 3 + 3 design, phase I study (Eudra-CT 2018-004103-39) conducted in three sites across the UK and France in patients with unresectable colorectal cancer with liver metastases having progressed on or after standard chemotherapy, including at least a fluoropyrimidine, oxaliplatin, and irinotecan or, in the UK only, entering a period of clinical observation following discontinuation of chemotherapy. Patients received up to two cycles of TG6002 combined with oral 5-FC ( ). TG6002 was administered via the main hepatic artery, through a catheter inserted into the femoral artery under angiographic assessment, for more than 30 minutes at doses of 1 × 10 6 , 1 × 10 7 , 1 × 10 8 , and 1 × 10 9 pfu. 5-FC was taken orally from days 5 to 14 at a dose of 50 mg/kg four times daily. A second treatment cycle was to be administered from day 43 in the absence of disease progression or unacceptable toxicity. Dose escalation proceeded after a review of safety data from each cohort by an independent safety review committee. The clinical trial protocol was approved by institutional ethics committees and conducted in accordance with the Declaration of Helsinki. All patients signed an informed consent document prior to study participation.
Blood and tissue samples were collected and processed using the Translational Cancer Immunotherapy Team quality-assured lab manual, which included standard operating procedures to regulate all processes. Peripheral blood was collected into tripotassium ethylene-diaminetetraacetic acid (K3EDTA) or serum clot-activator vacutainer tubes (both Scientific Laboratory Supplies) and processed within 2 hours of venepuncture, or as soon as possible thereafter. Tumor biopsies collected on day 4 or 8 were placed in formalin or RNAlater (both ThermoFisher) for IHC or PCR analyses, respectively. All sample collection time points are shown in .
Serum clot-activator tubes were left for a minimum of 30 minutes after venepuncture. All blood collection tubes were centrifuged for 10 minutes at 2,000 g. Plasma and serum aliquots from the upper layers were stored at −80°C. Peripheral blood mononuclear cells (PBMCs) were isolated by density-gradient centrifugation over Lymphoprep (STEMCELL Technologies) as per the manufacturer’s instructions. Cells were frozen at 1 × 10 7 /mL in 40% (v/v) Roswell Park Memorial Institute medium containing 5 mmol/L L-glutamine and 1 mmol/L sodium pyruvate (all Sigma), plus 50% (v/v) pooled human serum (SeraLab) and 10% (v/v) dimethyl sulfoxide (Sigma). PBMCs were stored in liquid nitrogen.
qPCR DNA was extracted from tumor biopsies using DNeasy Blood and Tissue Kits (QIAGEN). Primers (forward 5′-CGATGATGGAGTAATAAGTGGTAGGA-3′ and reverse 5′-CACCGACCGATGATAAGATTTG-3′; Integrated DNA Technologies) were used to detect the presence of TG6002. qRT-PCR qRT-PCR was performed on tumor biopsies and plasma for the detection of viral early D7R , viral late A10L , and FCU1 transcripts. RNA was extracted using RNeasy Plus Mini Kits (QIAGEN). The remaining viral and cellular DNA in samples was digested with TURBO DNase (ThermoFisher). Primers and probes for D7R (forward TTTAGCGATTCAAAGTACTGCTTTTT, reverse GCAGTGACTTCGCTGCCATT, and probe FAM-CGAAATGGTAATGCGTATGA), A10L (forward CTTCATACTCGCGATCCTCAAA, reverse TCGCCAACAGGTTAAAGAAATTAA, and probe ABY-TGGCGCTTCCAAACGTGCAATTT), and FCU1 (forward TCGTGGTCACAACATGAGATTTC, reverse TCTAATCTCCCACAGTTTTCCAAAG, and probe ABY-TCCGCCACACTACATGGTGAGATCTCC) were used. Detection of the TG6002 viral genome in plasma was performed by Charles River Laboratories, Evreux, using in-house methods. All PCR data were acquired on Applied Biosystems QuantStudio™ 5 Real-Time PCR Systems (ThermoFisher; RRID:SCR_020240) and analyzed using QuantStudio 3D AnalysisSuite Cloud (ThermoFisher; RRID:SCR_020238). The presence of the virus in tumor biopsies was considered positive by RT-qPCR if at least one of the viral mRNAs ( D7R , A10L , or FCU1 ) was detected. RNase-free water was used as a negative control. Plaque assays were performed by Transgene, France, using in-house methods. Briefly, tissue biopsies were sonicated for 15 seconds at room temperature before incubation with a permissive cell line, Vero (CCL-81; ATCC; RRID:CVCL_0059). Veros (passage number 128 at thawing) were cultured in Dulbecco's modified Eagle's medium (DMEM; Sigma) supplemented with 10% (v/v) fetal calf serum (FCS; ThermoFisher) and 40 mg/L gentamicin (Sigma) for two passages prior to use in plaques assays. Confluent monolayers were incubated with tumor samples for 30 minutes prior to incubation at 37°C under 1% (w/v) agarose for 3 days. Positive infection was determined by the presence of viral plaques. Veros were authenticated in 2014 (by qPCR and epifluorescence microscopy) by Clean Cells and tested negative for Mycoplasma infection. IHC for viral protein was performed by Cerba Research. A polyclonal anti-vaccinia virus antibody (Meridian Life Science; RRID:AB_153134) was used to detect virions; negative control was secondary antibody alone. 3, 3-diaminobenzidine-horseradish peroxidase (DAB-HRP) was used to visualize virus-positive cells. Data from all methods are expressed as positive or negative/below the level of detection. Biopsies from two patients (13 and 15) were not available.
DNA was extracted from tumor biopsies using DNeasy Blood and Tissue Kits (QIAGEN). Primers (forward 5′-CGATGATGGAGTAATAAGTGGTAGGA-3′ and reverse 5′-CACCGACCGATGATAAGATTTG-3′; Integrated DNA Technologies) were used to detect the presence of TG6002.
qRT-PCR was performed on tumor biopsies and plasma for the detection of viral early D7R , viral late A10L , and FCU1 transcripts. RNA was extracted using RNeasy Plus Mini Kits (QIAGEN). The remaining viral and cellular DNA in samples was digested with TURBO DNase (ThermoFisher). Primers and probes for D7R (forward TTTAGCGATTCAAAGTACTGCTTTTT, reverse GCAGTGACTTCGCTGCCATT, and probe FAM-CGAAATGGTAATGCGTATGA), A10L (forward CTTCATACTCGCGATCCTCAAA, reverse TCGCCAACAGGTTAAAGAAATTAA, and probe ABY-TGGCGCTTCCAAACGTGCAATTT), and FCU1 (forward TCGTGGTCACAACATGAGATTTC, reverse TCTAATCTCCCACAGTTTTCCAAAG, and probe ABY-TCCGCCACACTACATGGTGAGATCTCC) were used. Detection of the TG6002 viral genome in plasma was performed by Charles River Laboratories, Evreux, using in-house methods. All PCR data were acquired on Applied Biosystems QuantStudio™ 5 Real-Time PCR Systems (ThermoFisher; RRID:SCR_020240) and analyzed using QuantStudio 3D AnalysisSuite Cloud (ThermoFisher; RRID:SCR_020238). The presence of the virus in tumor biopsies was considered positive by RT-qPCR if at least one of the viral mRNAs ( D7R , A10L , or FCU1 ) was detected. RNase-free water was used as a negative control. Plaque assays were performed by Transgene, France, using in-house methods. Briefly, tissue biopsies were sonicated for 15 seconds at room temperature before incubation with a permissive cell line, Vero (CCL-81; ATCC; RRID:CVCL_0059). Veros (passage number 128 at thawing) were cultured in Dulbecco's modified Eagle's medium (DMEM; Sigma) supplemented with 10% (v/v) fetal calf serum (FCS; ThermoFisher) and 40 mg/L gentamicin (Sigma) for two passages prior to use in plaques assays. Confluent monolayers were incubated with tumor samples for 30 minutes prior to incubation at 37°C under 1% (w/v) agarose for 3 days. Positive infection was determined by the presence of viral plaques. Veros were authenticated in 2014 (by qPCR and epifluorescence microscopy) by Clean Cells and tested negative for Mycoplasma infection. IHC for viral protein was performed by Cerba Research. A polyclonal anti-vaccinia virus antibody (Meridian Life Science; RRID:AB_153134) was used to detect virions; negative control was secondary antibody alone. 3, 3-diaminobenzidine-horseradish peroxidase (DAB-HRP) was used to visualize virus-positive cells. Data from all methods are expressed as positive or negative/below the level of detection. Biopsies from two patients (13 and 15) were not available.
Quantification of 5-FC, 5-FU, and 5-fluoro-β-alanine (F-BAL) levels was performed using liquid chromatography coupled with high-resolution mass spectrometry (Hospices Civils de Lyon) as described previously ( ). Serum was evaluated on day 8 after TG6002. Additionally, 5-FU concentrations were measured in available tumor biopsies on day 8 post-TG6002. F-BAL was not measured in cohort 1 serum samples or tumor biopsies due to sample insufficiency.
Patient plasma was analyzed for calreticulin (CRT) by ELISA (ThermoFisher) as per the manufacturer’s instructions. Data are expressed as mean plasma concentration (ng/mL) ± SEM, calculated using a standard curve. Statistical significance was determined using paired two-tailed t tests (GraphPad Prism; RRID:SCR_002798) between sample time points (* P < 0.05; n = 14 patients), dependent on sample availability.
mRNA sequencing was performed by Novogene as per validated methods. After PCR, the gene expression level was calculated by the number of mapped reads. Statistically significant differentially expressed genes (ssDEG) were defined as ± >2 log 2 fold change of post-treatment samples compared with baseline with an associated P adj < 0.05. Data from six patients, across three cohorts, are shown for ssDEGs for C1D2 and C1D15. P adj values were transformed into −log 10 ( P adj) values, which were plotted against log 2 fold change values in volcano plots. Volcano plots depict all DEGs, not just ssDEGs, which are upregulated, downregulated, or unchanged for three patients at C1D2 and C1D15 compared with BS. ssDEG lists were analyzed using the Search Tool for Retrieval of Interacting Genes/Proteins (STRING) database ( http://string-db.org ; RRID:SCR_005223) to identify potential interactions between the genes and their reported biological function(s). Interaction confidence scores (ICS) were assigned to each protein association and ranked from 0 to 1, in which 1 is most likely to be accurate and 0 is least likely to be correct. An ICS of 0.5 indicates that every second interaction may be a false positive; therefore, only ssDEGs with an ICS of >0.5 were used for visualization and analysis. Gene Ontology (GO) biological functions were assessed at the ssDEG level, in which common genes were detected across multiple patients. These nine commonly expressed ssDEGs were analyzed independently in the STRING database to identify highly responsive signaling pathways following treatment.
PBMCs were analyzed for specific activation/immune checkpoint molecules. Briefly, PBMCs were stained for CD3 (HIT3a/FITC; RRID:AB_395745), CD4 (RPA-T4/APC-H7; RRID:AB_1645478), CD8 (RPA-T8/Alexa 700; RRID:AB_396953), CD56 (B159/PE-Cy7; RRID:AB_396853), CD19 (SJ25C1/APC-H7; RRID:AB_1645470), CD14 (M5E2/FITC; RRID:AB_395798), CD69 (FN50/APC; RRID:AB_398602), programmed cell death-ligand 1 (PD-L1; MIH1/PE-CF594; RRID:AB_2738400), programmed cell death protein 1 (PD-1; MIH4/PE; RRID:AB_647199), T-cell immunoglobulin and mucin-domain containing-3 (TIM-3; 7D3/BV786; RRID:AB_2741100), OX40 (L106/BV421; RRID:AB_2742558), CD40 (5C3/BUV395; RRID:AB_2739110), and CD25 (M-A251/BB700; RRID:AB_2744335; all BD Biosciences) plus CD40 ligand (L) (24-31/BV786; RRID:AB_2572187) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4; BNI3/BV605; RRID:AB_2632779; both BioLegend). Fluorescence minus one (FMO) was used as a negative control. Data were acquired on a CytoFLEX LX and analyzed using CytExpert software (RRID:SCR_017217) (both Beckman Coulter). Positive expression of markers was used to calculate fold-change differences ± SEM in expression from baseline samples. Paired two-tailed t tests (GraphPad Prism) were used to determine statistical significance between samples (* P < 0.05; n = 9 patients), dependent on sample availability.
Formalin-fixed, paraffin-embedded tissue biopsies were stained with rabbit anti-human PD-L1 antibody (1:500; Abcam; RRID:AB_2884993) and ImmPRESS HRP anti-rabbit IgG (peroxidase) secondary antibody (Vector Laboratories; RRID:AB_2336529). Positive staining was visualized using ImmPACT DAB Peroxidase (HRP) Substrate kit (Vector Laboratories). Control sections contained no primary antibody. Digital images were acquired at 20× magnification and quantified using QuPath software (RRID:SCR_018257). Data are expressed as cells positive for PD-L1 per mm 2 for n = 7 patients.
Briefly, triplicates of 1 × 10 5 /well PBMCs were incubated in the presence of either 2 μg/mL of overlapping peptide pools for carcinoembryonic antigen (CEA) or cytomegalovirus/Epstein-Barr virus/influenza (CEF; positive control) (both Cambridge Biosciences). Negative control was media alone; 10 pfu/cell TG6002 was used to assess response to treatment. IFNγ secretion from activated T cells was detected using a matched paired antibody kit (Mabtech). Spot-forming units (SFU) were visualized using 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium substrate (MabTech). Images were captured and quantified using an S6 FluoroSpot analyzer (Cellular Technology Limited). Data are presented as mean fold-change SFU per well ± SEM for post-treatment samples compared with baseline, for n = 6 patients, dependent on sample availability.
T-cell receptor (TCR) β sequencing was performed by Adaptive Technologies using a “survey” resolution to generate data from productive rearrangements only, which were exported from the immunoSEQ Analyzer (Adaptive Technologies) for further analysis. Complementary-determining region 3 (CDR3) sequences were input into the McPAS-TCR database (RRID:SCR_026024) and matched to known human TCR sequences. TCRs in each patient sample that matched known cancer antigen or neoantigen epitopes were identified; these were counted and their total productive frequency was calculated. Data are expressed as the number of TCRs matching cancer antigens/neoantigens ( x -axis) against the productive frequency of TCRs matching cancer antigens/neoantigens ( y -axis) for all available PBMC samples.
The mRNA sequencing FASTQ files are available in the Sequence Read Archive database (accession number PRJNA1192197). TCR sequencing data can be accessed via the immuneACCESS server using the following link: https://doi.org/10.21417/EW2024S . All data generated in this study are available upon request from the corresponding author.
Patient characteristics In total, 20 patients were screened; of these, 15 patients were entered into the study across three sites (Supplementary Table S1) and received at least one infusion of TG6002. The mean age was 61 years (range, 37–78 years), comprising 11 males and four females, which is a slightly younger age range and a higher male:female ratio compared with global patient demographics (Supplementary Table S2). Patients had mismatch repair (MMR) proficient cancers. Included patients had either progressed or were intolerant to both oxaliplatin- and irinotecan-based chemotherapy or were undergoing a period of observation following a course of chemotherapy. The primary tumor was in the colon in 11 patients and the rectum in four patients. The mean time from initial diagnosis to trial entry was 36.5 months (range, 8.1–89.8 months), and the patients had received a mean of 3.3 prior lines of antineoplastic therapy (range, 1–7), including adjuvant lines. Thirteen patients completed the trial; two patients withdrew prematurely, one with dose-limiting toxicity (DLT) and one because of palliative best supportive care. No patient deaths were related to TG6002 and/or 5-FC; of the 13 patients completing the trial, the cause of death for 12 patients was a progression of the underlying disease and one case of non-study treatment-related pneumonia occurred 22.3 months after inclusion. Patient exposure Of the 15 patients entered, three were treated in each of the 3 + 3 design cohorts (1, 2, and 3), with six treated in cohort 4 (Supplementary Table S3), as planned. Fourteen patients received only a single cycle of treatment due to the progression of their disease, and one received two cycles; each cycle being a single dose of TG6002 via IHA infusion on day 1 followed by 10 days of oral 5-FC on days 5 to 14 ( ). All infusions were fully administered. In 13 patients, the whole liver was perfused, whereas two patients had partial liver perfusion because of anatomical considerations. One patient did not receive 5-FC, having withdrawn from the trial on day 1 after the TG6002 infusion and before receiving 5-FC. Thirteen of the remaining 14 patients received their complete 10-day course of 5-FC, but one patient discontinued 5-FC after 9 days. Safety data Overall, 14 patients (93.3%) experienced at least one study treatment–related adverse event (AE), of whom 13 (86.7%) experienced at least one AE related to TG6002 and nine (60.0%) experienced at least one AE related to 5-FC ( and ). Eight grade 3 AEs were observed in five patients, including myocardial infarction (MI), diarrhea, vomiting, pyrexia, and increased aspartate aminotransferase related to study treatment ( and ), plus hypertension and anemia unrelated to treatment. Overall, no AEs greater than grade 3 were reported. Grade 3 TG6002-related AEs included pyrexia and MI in one cohort 4 patient, which constituted a DLT ( and ). This patient acquired an asymptomatic COVID-19 infection near the time of TG6002 infusion, precipitating a supraventricular tachycardia. Four hours following the TG6002 infusion, the patient had a fever and tachycardia that peaked at 40.1°C and 148 bpm. On day 2, an electrocardiogram showed negative T waves, and the troponin I level was increased at 1,555 pg/L (normal level < 45). A coronary angiography was performed on day 3 showing a monotruncular lesion of the anterior interventricular artery leading to coronary stent insertion and successful revascularization. No cases of vesicular or pustular skin or mucosal lesions were reported following TG6002 infusion. Grade 3 AEs related to 5-FC included diarrhea and vomiting in one cohort 4 patient. Aspartate aminotransferase increase occurred in three patients: grade 1 in two patients, with one case related to TG6002, and grade 2 in one patient. Hypertension was reported for three patients: grade 2 in a cohort 2 patient and grade 3 in a patient each in cohorts 1 and 4. Despite being assessed as not related to study treatment, this resulted from antihypertensives being withdrawn during TG6002 administration. Trial endpoints and objective efficacy data No patients had an objective response based on a 10-week disease control rate according to RECIST version 1.1. The primary objective of the maximum feasible dose was 1 × 10 9 ; MTD was not reached. Secondary objectives of safety and tolerability were achieved; in addition, viral shedding was not evident in saliva, urine, or feces. Median progression-free survival was 1.05 months, with a range of 0.0 to 2.3 (in which 0.0 relates to the MI reported previously), which is very short because of the timing of the CT scan 4 weeks after TG6002 infusion ahead of the planned second TG6002 infusion. Median overall survival was 5.4 months (range, 1.6–22.3). Although a minority of patients had progressive disease localized within the liver, the majority had indications of both intra- and extrahepatic progressive disease; elevations in circulating CEA levels, compared with baseline levels, were also observed at later time points. Detection of TG6002 and FCU1 transgene activity in tumor biopsy Blood and tissue samples for translational analyses were taken as outlined in the trial schedule ( ). As only one patient received a second cycle of treatment, the translational assays for all patients were performed on samples obtained during the first cycle only. Tumor biopsies obtained at screening and after treatment (day 4 or 8) were examined for the presence of the TG6002 virus or for the activity of the viral FCU1 transgene ( ). Viral DNA by qPCR was detected in five of 13 evaluable tumor samples, predominantly in biopsies from patients who received a higher viral dose than patients in earlier cohorts, both at days 4 and 8 ( ). Furthermore, the virus was detected by qRT-PCR in four of nine patients with suitable biospecimens. Plaque assays indicated live replicating TG6002 in two of nine tumors. Viral protein was detected by IHC in three of six evaluable biopsies: one on day 4 and two on day 8, with representative examples shown in . The active conversion of 5-FC to its metabolite, 5-FU, was detected in three of six evaluable post-treatment biopsies, predominantly in tumors of patients in later cohorts, suggesting that a higher virus dose is required for TG6002 activity within tumor. One patient in cohort 2 (one out of two available biopsies) had 16.2 pg of 5-FU/mg of tumor tissue, whereas two patients in cohort 4 (two out of three available biopsies) had 35 and 29 pg of 5-FU/mg of tumor tissue. Overall, there was evidence of virus infection and viral replication in 10 of 13 patients’ tumors. Detection of TG6002 and FCU1 transgene activity in plasma Despite positive detection in tumors, TG6002 was not found in the vast majority of plasma samples from cohorts 1 to 3, with the exception of one patient in cohort 2 and one in cohort 3, both on day 8, indicating active virus replication ( ). In contrast, in the highest dose cohort, TG6002 was detected in plasma 30 minutes post-infusion in five of six patients, followed by undetectable levels indicating rapid clearance. A rebound of circulating TG6002 was observed in one patient on day 4 and three patients on day 8, again indicating active virus replication. Titers of neutralizing antibody (nAb) against TG6002 significantly increased following treatment in all patients ( P < 0.05), with a trend for higher titers in the highest dose cohort ( ). Serum levels of 5-FC, 5-FU, and the catabolite F-BAL were measured 8 days after exposure to TG6002. Although serum 5-FC concentrations were comparable across all cohorts, patients who received higher virus doses had higher levels of circulating 5-FU than those in cohort 1, indicating replication of TG6002 ( ). F-BAL was detected in plasma from all patients from cohorts 2 to 4, where samples were available. Similar to the tumor data, there was evidence of viral presence/replication in serum in all evaluable patients. Host response to TG6002/5-FC The peripheral immune response to IHA infusion of TG6002 was assessed using serial blood samples from patients. CRT was investigated as an indicator of ICD following treatment. CRT concentration in patient plasma significantly increased following TG6002 infusion ( ); a peak was detected at 6 hours post-treatment ( P < 0.05), which remained higher than pre-treatment levels on day 2 ( P < 0.05), potentially indicating a peak in ICD. Immunophenotyping of PBMCs revealed CD69 upregulation, an early activation marker, 6 to 24 hours post-infusion on cell populations including CD4 + and CD8 + T cells, natural killer (NK) cells, natural killer T (NKT) cells, and B cells ( ). An increase in PD-L1 expression was also observed, as exemplified on NK cells ( ) alongside other immune checkpoint molecules, such as PD-1, TIM-3, and OX40 (Supplementary Fig. S1A). In addition, elevation in CD40L on T cells and NK(T) cells (Supplementary Fig. S1B) and an associated increase in its receptor (CD40) on both monocytes and B cells (Supplementary Fig. S1C) were found, indicating an enhanced capacity for the maturation of antigen-presenting cells. In contrast, a decrease in both CD25 and CTLA-4 on the surface of T cells was apparent, which appeared to be prolonged over time, indicating reduced regulatory T cell functions (Supplementary Fig. S1D). IHC on tumor biopsies ( ) sampled before and after infusion showed low-level PD-L1 expression at baseline and a small reduction in the level of PD-L1 in the tumor by day 8 in patients receiving a lower virus dose (cohorts 1–3). However, with the highest dose (cohort 4), there was a substantial increase in the expression of PD-L1 following virus infusion, reflecting the PBMC data. Representative examples for patients in cohorts 1 and 4 are shown, indicating the extent of cells positive for PD-L1 in patients who received the higher viral load ( ). mRNA sequencing was used to characterize the effects of TG6002/5-FC treatment at the transcriptional level ( ). A greater number of ssDEGs were increased at day 2 in PBMCs from patients in cohorts 2 and 3 compared with patients in cohort 1 ( ). Patients from cohort 1, who did not show elevated ssDEGs by day 2, had a greater number of ssDEGs on day 15, potentially indicating a delayed response to replicating virus or to 5-FC and its metabolites, including 5-FU. Volcano plots ( ) show the pattern of all DEGs in three patients. Nine commonly expressed genes were significantly upregulated in response to TG6002 in three patients, namely, CXCL10 , IFIT1 , IFIT3 , IFI27 , IFI44L , IFITM3 , IFI6 , RSAD2 , and SERPING1 , all of which are involved in immune-related signaling pathways. Although these specific ssDEGs were evident by day 2 in Pt-07 (cohort 3), an increase in expression was only apparent in patients from cohort 1 (Pt-01 and Pt-03) by day 15. GO analysis of all upregulated ssDEGs generated cluster plots depicting pathways in which the DEGs are highly involved ( ). Cluster analysis of the nine commonly expressed ssDEGs in the three specified patients ( ) revealed a number of pathways highly relevant to immune responses, type I interferon signaling, and, more specifically, response to the virus ( ). Each pathway identified had a significant proportion of the nine ssDEGs involved, as indicated. More widely, the predominant clustering of all ssDEGs highlights several immune-related signaling pathways. Pathway clustering was apparent by day 2 in the cohort 3 patient in comparison with day 15 in cohort 1 patients ( ). A greater extent of clustering mirrors both the enhanced number of ssDEGs and their earlier appearance following TG6002 infusion, as previously observed ( ). The adaptive T-cell response to virus infusion was examined by ELISpot assay against a tumor-associated antigen (TAA; CEA) and TG6002 ( and ). CEA-specific T-cell responses were detected by day 4, likely indicating enhanced activation of pre-existing CEA-specific T-cell clones. In contrast, the appearance of TG6002-specific T cells occurred later, at day 15 ( ). Furthermore, only patients in the later cohorts (2 and 3) elicited TG6002-specific T-cell responses, presumably due to a higher virus load. Representative wells showing IFNγ responses to CEA and TG6002 are depicted for two patients ( ). TCRβ sequencing of patient tumor and PBMCs at all available time points was also performed. T-cell clonal response to treatment revealed CDR3 sequences matched to cancer antigens ( ) and specifically to neoantigens ( ), which showed greater frequencies in later cohorts treated with higher viral doses than in earlier lower-dose cohorts.
In total, 20 patients were screened; of these, 15 patients were entered into the study across three sites (Supplementary Table S1) and received at least one infusion of TG6002. The mean age was 61 years (range, 37–78 years), comprising 11 males and four females, which is a slightly younger age range and a higher male:female ratio compared with global patient demographics (Supplementary Table S2). Patients had mismatch repair (MMR) proficient cancers. Included patients had either progressed or were intolerant to both oxaliplatin- and irinotecan-based chemotherapy or were undergoing a period of observation following a course of chemotherapy. The primary tumor was in the colon in 11 patients and the rectum in four patients. The mean time from initial diagnosis to trial entry was 36.5 months (range, 8.1–89.8 months), and the patients had received a mean of 3.3 prior lines of antineoplastic therapy (range, 1–7), including adjuvant lines. Thirteen patients completed the trial; two patients withdrew prematurely, one with dose-limiting toxicity (DLT) and one because of palliative best supportive care. No patient deaths were related to TG6002 and/or 5-FC; of the 13 patients completing the trial, the cause of death for 12 patients was a progression of the underlying disease and one case of non-study treatment-related pneumonia occurred 22.3 months after inclusion.
Of the 15 patients entered, three were treated in each of the 3 + 3 design cohorts (1, 2, and 3), with six treated in cohort 4 (Supplementary Table S3), as planned. Fourteen patients received only a single cycle of treatment due to the progression of their disease, and one received two cycles; each cycle being a single dose of TG6002 via IHA infusion on day 1 followed by 10 days of oral 5-FC on days 5 to 14 ( ). All infusions were fully administered. In 13 patients, the whole liver was perfused, whereas two patients had partial liver perfusion because of anatomical considerations. One patient did not receive 5-FC, having withdrawn from the trial on day 1 after the TG6002 infusion and before receiving 5-FC. Thirteen of the remaining 14 patients received their complete 10-day course of 5-FC, but one patient discontinued 5-FC after 9 days.
Overall, 14 patients (93.3%) experienced at least one study treatment–related adverse event (AE), of whom 13 (86.7%) experienced at least one AE related to TG6002 and nine (60.0%) experienced at least one AE related to 5-FC ( and ). Eight grade 3 AEs were observed in five patients, including myocardial infarction (MI), diarrhea, vomiting, pyrexia, and increased aspartate aminotransferase related to study treatment ( and ), plus hypertension and anemia unrelated to treatment. Overall, no AEs greater than grade 3 were reported. Grade 3 TG6002-related AEs included pyrexia and MI in one cohort 4 patient, which constituted a DLT ( and ). This patient acquired an asymptomatic COVID-19 infection near the time of TG6002 infusion, precipitating a supraventricular tachycardia. Four hours following the TG6002 infusion, the patient had a fever and tachycardia that peaked at 40.1°C and 148 bpm. On day 2, an electrocardiogram showed negative T waves, and the troponin I level was increased at 1,555 pg/L (normal level < 45). A coronary angiography was performed on day 3 showing a monotruncular lesion of the anterior interventricular artery leading to coronary stent insertion and successful revascularization. No cases of vesicular or pustular skin or mucosal lesions were reported following TG6002 infusion. Grade 3 AEs related to 5-FC included diarrhea and vomiting in one cohort 4 patient. Aspartate aminotransferase increase occurred in three patients: grade 1 in two patients, with one case related to TG6002, and grade 2 in one patient. Hypertension was reported for three patients: grade 2 in a cohort 2 patient and grade 3 in a patient each in cohorts 1 and 4. Despite being assessed as not related to study treatment, this resulted from antihypertensives being withdrawn during TG6002 administration.
No patients had an objective response based on a 10-week disease control rate according to RECIST version 1.1. The primary objective of the maximum feasible dose was 1 × 10 9 ; MTD was not reached. Secondary objectives of safety and tolerability were achieved; in addition, viral shedding was not evident in saliva, urine, or feces. Median progression-free survival was 1.05 months, with a range of 0.0 to 2.3 (in which 0.0 relates to the MI reported previously), which is very short because of the timing of the CT scan 4 weeks after TG6002 infusion ahead of the planned second TG6002 infusion. Median overall survival was 5.4 months (range, 1.6–22.3). Although a minority of patients had progressive disease localized within the liver, the majority had indications of both intra- and extrahepatic progressive disease; elevations in circulating CEA levels, compared with baseline levels, were also observed at later time points.
FCU1 transgene activity in tumor biopsy Blood and tissue samples for translational analyses were taken as outlined in the trial schedule ( ). As only one patient received a second cycle of treatment, the translational assays for all patients were performed on samples obtained during the first cycle only. Tumor biopsies obtained at screening and after treatment (day 4 or 8) were examined for the presence of the TG6002 virus or for the activity of the viral FCU1 transgene ( ). Viral DNA by qPCR was detected in five of 13 evaluable tumor samples, predominantly in biopsies from patients who received a higher viral dose than patients in earlier cohorts, both at days 4 and 8 ( ). Furthermore, the virus was detected by qRT-PCR in four of nine patients with suitable biospecimens. Plaque assays indicated live replicating TG6002 in two of nine tumors. Viral protein was detected by IHC in three of six evaluable biopsies: one on day 4 and two on day 8, with representative examples shown in . The active conversion of 5-FC to its metabolite, 5-FU, was detected in three of six evaluable post-treatment biopsies, predominantly in tumors of patients in later cohorts, suggesting that a higher virus dose is required for TG6002 activity within tumor. One patient in cohort 2 (one out of two available biopsies) had 16.2 pg of 5-FU/mg of tumor tissue, whereas two patients in cohort 4 (two out of three available biopsies) had 35 and 29 pg of 5-FU/mg of tumor tissue. Overall, there was evidence of virus infection and viral replication in 10 of 13 patients’ tumors.
FCU1 transgene activity in plasma Despite positive detection in tumors, TG6002 was not found in the vast majority of plasma samples from cohorts 1 to 3, with the exception of one patient in cohort 2 and one in cohort 3, both on day 8, indicating active virus replication ( ). In contrast, in the highest dose cohort, TG6002 was detected in plasma 30 minutes post-infusion in five of six patients, followed by undetectable levels indicating rapid clearance. A rebound of circulating TG6002 was observed in one patient on day 4 and three patients on day 8, again indicating active virus replication. Titers of neutralizing antibody (nAb) against TG6002 significantly increased following treatment in all patients ( P < 0.05), with a trend for higher titers in the highest dose cohort ( ). Serum levels of 5-FC, 5-FU, and the catabolite F-BAL were measured 8 days after exposure to TG6002. Although serum 5-FC concentrations were comparable across all cohorts, patients who received higher virus doses had higher levels of circulating 5-FU than those in cohort 1, indicating replication of TG6002 ( ). F-BAL was detected in plasma from all patients from cohorts 2 to 4, where samples were available. Similar to the tumor data, there was evidence of viral presence/replication in serum in all evaluable patients.
The peripheral immune response to IHA infusion of TG6002 was assessed using serial blood samples from patients. CRT was investigated as an indicator of ICD following treatment. CRT concentration in patient plasma significantly increased following TG6002 infusion ( ); a peak was detected at 6 hours post-treatment ( P < 0.05), which remained higher than pre-treatment levels on day 2 ( P < 0.05), potentially indicating a peak in ICD. Immunophenotyping of PBMCs revealed CD69 upregulation, an early activation marker, 6 to 24 hours post-infusion on cell populations including CD4 + and CD8 + T cells, natural killer (NK) cells, natural killer T (NKT) cells, and B cells ( ). An increase in PD-L1 expression was also observed, as exemplified on NK cells ( ) alongside other immune checkpoint molecules, such as PD-1, TIM-3, and OX40 (Supplementary Fig. S1A). In addition, elevation in CD40L on T cells and NK(T) cells (Supplementary Fig. S1B) and an associated increase in its receptor (CD40) on both monocytes and B cells (Supplementary Fig. S1C) were found, indicating an enhanced capacity for the maturation of antigen-presenting cells. In contrast, a decrease in both CD25 and CTLA-4 on the surface of T cells was apparent, which appeared to be prolonged over time, indicating reduced regulatory T cell functions (Supplementary Fig. S1D). IHC on tumor biopsies ( ) sampled before and after infusion showed low-level PD-L1 expression at baseline and a small reduction in the level of PD-L1 in the tumor by day 8 in patients receiving a lower virus dose (cohorts 1–3). However, with the highest dose (cohort 4), there was a substantial increase in the expression of PD-L1 following virus infusion, reflecting the PBMC data. Representative examples for patients in cohorts 1 and 4 are shown, indicating the extent of cells positive for PD-L1 in patients who received the higher viral load ( ). mRNA sequencing was used to characterize the effects of TG6002/5-FC treatment at the transcriptional level ( ). A greater number of ssDEGs were increased at day 2 in PBMCs from patients in cohorts 2 and 3 compared with patients in cohort 1 ( ). Patients from cohort 1, who did not show elevated ssDEGs by day 2, had a greater number of ssDEGs on day 15, potentially indicating a delayed response to replicating virus or to 5-FC and its metabolites, including 5-FU. Volcano plots ( ) show the pattern of all DEGs in three patients. Nine commonly expressed genes were significantly upregulated in response to TG6002 in three patients, namely, CXCL10 , IFIT1 , IFIT3 , IFI27 , IFI44L , IFITM3 , IFI6 , RSAD2 , and SERPING1 , all of which are involved in immune-related signaling pathways. Although these specific ssDEGs were evident by day 2 in Pt-07 (cohort 3), an increase in expression was only apparent in patients from cohort 1 (Pt-01 and Pt-03) by day 15. GO analysis of all upregulated ssDEGs generated cluster plots depicting pathways in which the DEGs are highly involved ( ). Cluster analysis of the nine commonly expressed ssDEGs in the three specified patients ( ) revealed a number of pathways highly relevant to immune responses, type I interferon signaling, and, more specifically, response to the virus ( ). Each pathway identified had a significant proportion of the nine ssDEGs involved, as indicated. More widely, the predominant clustering of all ssDEGs highlights several immune-related signaling pathways. Pathway clustering was apparent by day 2 in the cohort 3 patient in comparison with day 15 in cohort 1 patients ( ). A greater extent of clustering mirrors both the enhanced number of ssDEGs and their earlier appearance following TG6002 infusion, as previously observed ( ). The adaptive T-cell response to virus infusion was examined by ELISpot assay against a tumor-associated antigen (TAA; CEA) and TG6002 ( and ). CEA-specific T-cell responses were detected by day 4, likely indicating enhanced activation of pre-existing CEA-specific T-cell clones. In contrast, the appearance of TG6002-specific T cells occurred later, at day 15 ( ). Furthermore, only patients in the later cohorts (2 and 3) elicited TG6002-specific T-cell responses, presumably due to a higher virus load. Representative wells showing IFNγ responses to CEA and TG6002 are depicted for two patients ( ). TCRβ sequencing of patient tumor and PBMCs at all available time points was also performed. T-cell clonal response to treatment revealed CDR3 sequences matched to cancer antigens ( ) and specifically to neoantigens ( ), which showed greater frequencies in later cohorts treated with higher viral doses than in earlier lower-dose cohorts.
In total, 15 patients received an IHA infusion of TG6002 plus oral 5-FC as part of a dose-escalation phase I study highlighting the feasibility of this locoregional route of OV delivery in the treatment of liver tumors. IHA delivery of TG6002 was clinically feasible and safe, with the MTD not reached. Dose escalation proceeded as per the protocol, with six patients receiving the highest intended dose of 1 × 10 9 pfu. Only one patient, in cohort 4, had a DLT that consisted of the aforementioned MI. Disappointingly, no patients experienced clinical or radiological tumor responses, with almost all patients showing continued disease progression one month-post TG6002 delivery. The reasons for this could include the heavily pre-treated population; patients in this trial had all exhausted standard chemotherapy options, meaning that their cancers are likely 5-FU-resistant. Patients in this trial were not selected on the basis of MMR tumor status, and no tumors were known to be MMR-deficient within the recruited cohort. It is likely that all or the majority of recruited patient tumors were MMR-proficient and relatively resistant to immunotherapy so less likely to benefit from OV therapy. Successful delivery of TG6002 to tumor lesions was achieved via the IHA route. Viral persistence in tumor biopsies sampled post-treatment was evident from a number of analyses, including qPCR, RT-qPCR, plaque assay, IHC, and transgene activity, with the vast majority of patients exhibiting positive detection by one or more methods, despite only limited tissue from a core biopsy being available for analysis. Neutralizing antibodies against TG6002 developed at low levels following infusion, reaching peak titers by day 15 or 29, indicating a humoral immune response to IHA delivery. It is unknown how the presence of low-level nAb might affect the repeated delivery of TG6002 via IHA; further investigation of locoregional oncolytic vaccinia virus therapy for immunotherapy-sensitive tumors is merited. Functional transcription of the FCU1 transgene, indicative of a replicating virus, was evident from pharmacokinetic analyses. Serum 5-FU concentrations ranged from 1 to 1,072 ng/mL across all cohorts with significantly elevated levels at higher dose cohorts. Tumor 5-FU titers were detectable in the higher treatment doses, with two patients who received the highest dose of virus having concentrations exceeding 25 pg/mg of tissue. The range of 5-FU concentrations from 16 to 35 pg/mg of tissue compared favorably with that of 5.9 ± 0.9 pg/mg reported in tumor tissue of patients with HCC treated with an oral prodrug of 5-FU ( ) and were close to the mean 5-FU concentration of 56.6 pg/mg after i.t. injection of a nonpropagative vaccinia virus expressing FCU1 in combination with oral 5-FC ( ). Of interest, i.v. administered 5-FU can result in higher serum levels of 5-FU, with targeted serum concentrations of 2,500 to 3,000 ng/mL ( , ), compared with a median of 82 ng/mL detected in our patients receiving oral 5-FC. Therefore, maximizing 5-FU concentrations in the tumor tissue using TG6002/oral 5-FC combination allows direct targeting of malignant cells while minimizing systemic toxicity. Although 5-FU might diffuse from the higher concentrations produced within the tumor microenvironment, the levels should remain higher where a therapeutic effect is desirable. Higher serum concentrations of 5-FU, as experienced during standard i.v. delivery can be problematic, as side effects can be very significant, and many patients are unable to tolerate repeated cycles. Moreover, there is frequently rapid development of resistance to 5-FU alone when administered via an i.v. route. Despite an apparent lack of clinical efficacy, early peripheral blood immune cell responses indicated immune activity resulting from the combination therapy; in addition to promoting an antitumor response, this may also represent a virally driven immune response to pave the way for both direct tumor lysis and abscopal effects through immune modulation and 5-FU activation. CRT plasma concentrations increased shortly after TG6002 infusion, peaking 6 hours post-treatment and remaining elevated up to 24 hours. CRT is an endoplasmic reticulum–associated chaperone protein ubiquitously expressed intracellularly but also released from cells undergoing ICD ( , ) and is one of the main hallmarks of ICD in malignant disease. ICD is a unique class of regulated cell death that elicits antigen-specific adaptive immune processes via the release of damage-associated molecular patterns (DAMPs), of which CRT is a key component. The overall role of ICD and DAMP release is the recruitment of antigen-presenting cells to the site of dying tumor cells, in order to promote antigen uptake and processing, prior to cross-presentation to T cells to initiate a tumor-specific immune response. Associated with the occurrence of ICD, mRNA sequencing revealed a significant response of immune cells at the transcriptional level. A considerable elevation in the number of ssDEGs was apparent in patients who received the highest doses of TG6002. These ssDEGs formed clusters representing immune-related pathways in patients receiving higher virus doses at earlier time points than patients receiving lower doses, suggesting greater immune activation at higher doses of the virus. GO annotations revealed signaling pathways associated with response to virus involving IFN-stimulated genes and, subsequently, immune activation. Specifically, nine ssDEGs were predominantly upregulated in multiple patients and demonstrated to be highly relevant in the aforementioned pathways. Immunophenotyping of patient PBMCs evidenced immune cell activation across multiple cell populations. CD69, an early activation marker, was elevated shortly after TG6002 infusion, as was the immune checkpoint ligand PD-L1, a common marker for immune activation and exhaustion following OV therapy ( ). PD-L1 expression following virus therapy is associated with an anti-tumor immune response driven by IFNs and other inflammatory molecules ( ). Furthermore, increased PD-L1 expression by both peripheral and tumor-infiltrating T cells has been associated with a better prognosis for immune checkpoint blockade ( , ). Likewise, PD-1 and TIM-3 expression are indicative of inflammatory cytokine signaling, including IL-12, IL-15, IL-18, and IFNγ ( , ). Elevated expression of OX40 on NK cells is broadly associated with activation, in line with previous data ( ). Engagement of the CD40/CD40L complex is a secondary activation signal required for T-cell activation and is associated with enhanced cytokine production by dendritic cells, coupled with enhanced cross-presentation capacity ( ). This interaction has beneficial effects across the immune system as a whole and is, therefore, regarded as instrumental in an inflammatory response. Furthermore, the downregulation of CD25 and CTLA-4 on T cells also indicates the immunological switch from suppressive to activation, in response to TG6002/5-FC therapy. Functional T-cell responses to treatment were evident against TG6002 but, more importantly, against TAAs, suggesting an antigen-targeted cytotoxic effect against tumors. TCRβ sequencing revealed that a greater number of T-cell clones targeting both cancer antigens in general and specifically neoantigens, emerged within patients receiving higher doses of virus. Anti-cancer clonal T-cell expansion is associated with improved anti-cancer activity ( ), where increased CD8 T-cell tumor infiltration was also observed in the highest-dose cohort upon treatment. As such, whether a T cell is activated toward a TAA/neoantigen or the virus itself may ultimately result in a parallel anti-tumor response against virus-infected cells within the tumor microenvironment. These data, in summary, represent the first-in-human dose-escalation study using IHA delivery of a vaccinia virus. This treatment strategy directly targets tumor tissue and is associated with effective viral replication, expression and activity of the FCU1 transgene, immune activation, and evidence for antitumor immune activity. Further assessment of IHA delivery of OV in immunotherapy-sensitive cancers is merited.
Supplementary Figure S1 Immunophenotyping patient PBMCs Supplementary Table S1 Patient demographics and disease summary Supplementary Table S2 Representativeness of study participants Supplementary Table S3 Treatment cohorts with patient trial/manuscript ID designation
|
Antimicrobial Activity of Antibacterial Sutures in Oral Surgery: A Scoping Review | 7bb8271f-93ea-4537-b4aa-b76aa1589350 | 11287137 | Suturing[mh] | Sutures in surgery, especially in oral surgery, are pivotal for the healing of surgical sites. However, sutures tend to attract bacteria to their surface , due to intrinsic characteristics (eg, braided vs monofilament) and external factors such as the deposition of fibrinogen and fibronectin during the healing process, which can favor bacterial colonisation. , , , , The bacterial retention within sutures should be minimised to reduce postoperative infection risk. As it is reliable, easy to use, and stable, the black braided silk suture is one of the most common sutures used in oral surgery. The prolonged and delayed healing associated with black braided silk sutures has led to the exploration of alternative suture materials. Antibacterial suture materials are coated or treated with antimicrobial agents to promote healing and prevent postsurgical infection. Antibacterial suture materials currently available include polymers or absorbable materials (ie, polyglactin 910, polyglecaprone 25, polydioxanone) coated with antibacterial substances (ie, triclosan or chlorhexidine). , , , , Triclosan is a phenolic derivative antibacterial agent with antiseptic and bacteriostatic activity both against Gram-positive bacteria and to a lesser extent against Gram-negative bacteria. Since 2003, in the US, the Vicryl Plus suture has been covered by triclosan, a substance already used for about 20 years as an additive to oral hygiene products such as toothpastes and mouthwashes , with documented biocompatibility. , Chlorhexidine exhibits a broad antibacterial spectrum, bacteriostatic and bactericidal effects, and high biocompatibility indexes. , It is also effective against most Gram-negative and Gram-positive bacteria. In the literature, results on antibacterial efficacy of the antimicrobial-coated sutures in the human body are conflicting. A remarkable decline in the number of microorganisms on the surface of these sutures has been reported, , , although most referred to skin bacteria. , Due to saliva, specific microorganisms, quality of tissue involved, and local factors, sutures placed in the oral cavity behave differently than those placed outside it. , , , The antimicrobial effects of new antibacterial-coated sutures in oral surgery are still unclear. Thus, the aim of this scoping review was to assess the antimicrobial activity of antibacterial suture materials in oral surgery in order to provide an updated synthesis of the current studies and provide useful insights for future research.
This scoping review was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Extension for Scoping Reviews and focussed on the following research question: What is the antimicrobial activity of antibacterial-coated suture materials employed in oral surgery? Search strategy A literature search was conducted in PubMed and Scopus databases to identify articles published from January 2013 to May 2023 that investigated the antimicrobial activity of antibacterial sutures used in oral surgery. The following strings were adopted for each database: (antibacterial suture oral surgery) OR (antimicrobial suture oral surgery) OR (antimicrobial suture oral surgical site) OR (antibacterial suture oral surgical site) OR (anti-bacterial suture oral surgery) OR (anti-microbial suture oral surgery) OR (anti-microbial suture oral surgical site) OR (anti-bacterial suture oral surgical site) OR (antimicrobial suture periodontal surgery) OR (anti-bacterial sutures periodontal surgery) OR (anti-microbial sutures periodontal surgery) OR (oral surgery vicryl plus). No language restriction was used. Reference lists of the included studies were further screened for other pertinent studies. Principal peer-reviewed scientific journals in oral surgery and miscellaneous were also hand-searched ( Journal of Clinical Periodontology, Periodontology 2000, Journal of Periodontology, Journal of Dental Research, Clinical Oral Implants Research, Clinical Oral Investigations, International Journal of Oral & Maxillofacial Implants, International Journal of Oral and Maxillofacial Surgery, BMC Oral Health, Odontology, Oral Surgery Oral Medicine Oral Pathology Oral Radiology, International Journal of Implant Dentistry, Journal of Cranio-Maxillofacial Surgery, Journal of Oral and Maxillofacial Surgery ). Eligibility criteria All human clinical studies (ie, cohort studies, randomised clinical trials, quasi-experimental studies, case reports, and case series) exploring the antimicrobial activity of antibacterial-coated sutures used in any field of oral surgery were considered suitable for inclusion. The exclusion criteria regarded the study design (ie, in vitro and ex vivo studies and animal studies), article type (ie, editorials, commentaries, short communication, and reviews), peer revision (ie, abstracts and preprint articles), and language (ie, studies without an English abstract). Studies selection Two authors independently reviewed studies and decided which studies to include. Disagreement was solved through discussion or by the decision of a third expert reviewer. The references were managed using EndNote software (EndNote X9; Thomson Reuters). Data extraction For each study, the following items were tabulated: author, year, and country; study design; sample size; population; intervention; control; type of surgery; suture removal time; methodology; main results; and additional information. Data were extracted independently by 2 reviewers. Any discrepancies were solved by discussion or intervention of a third reviewer.
A literature search was conducted in PubMed and Scopus databases to identify articles published from January 2013 to May 2023 that investigated the antimicrobial activity of antibacterial sutures used in oral surgery. The following strings were adopted for each database: (antibacterial suture oral surgery) OR (antimicrobial suture oral surgery) OR (antimicrobial suture oral surgical site) OR (antibacterial suture oral surgical site) OR (anti-bacterial suture oral surgery) OR (anti-microbial suture oral surgery) OR (anti-microbial suture oral surgical site) OR (anti-bacterial suture oral surgical site) OR (antimicrobial suture periodontal surgery) OR (anti-bacterial sutures periodontal surgery) OR (anti-microbial sutures periodontal surgery) OR (oral surgery vicryl plus). No language restriction was used. Reference lists of the included studies were further screened for other pertinent studies. Principal peer-reviewed scientific journals in oral surgery and miscellaneous were also hand-searched ( Journal of Clinical Periodontology, Periodontology 2000, Journal of Periodontology, Journal of Dental Research, Clinical Oral Implants Research, Clinical Oral Investigations, International Journal of Oral & Maxillofacial Implants, International Journal of Oral and Maxillofacial Surgery, BMC Oral Health, Odontology, Oral Surgery Oral Medicine Oral Pathology Oral Radiology, International Journal of Implant Dentistry, Journal of Cranio-Maxillofacial Surgery, Journal of Oral and Maxillofacial Surgery ).
All human clinical studies (ie, cohort studies, randomised clinical trials, quasi-experimental studies, case reports, and case series) exploring the antimicrobial activity of antibacterial-coated sutures used in any field of oral surgery were considered suitable for inclusion. The exclusion criteria regarded the study design (ie, in vitro and ex vivo studies and animal studies), article type (ie, editorials, commentaries, short communication, and reviews), peer revision (ie, abstracts and preprint articles), and language (ie, studies without an English abstract).
Two authors independently reviewed studies and decided which studies to include. Disagreement was solved through discussion or by the decision of a third expert reviewer. The references were managed using EndNote software (EndNote X9; Thomson Reuters).
For each study, the following items were tabulated: author, year, and country; study design; sample size; population; intervention; control; type of surgery; suture removal time; methodology; main results; and additional information. Data were extracted independently by 2 reviewers. Any discrepancies were solved by discussion or intervention of a third reviewer.
Search findings The electronic search resulted in 150 articles. After exclusion of duplicates, 129 abstracts were reviewed and the full texts of 10 studies were screened. Finally, 5 studies were included for qualitative analysis , , , , ( ). The characteristics of the included studies are shown in the . General characteristics of the included studies The selected studies were published from 2014 to 2019; 3 were conducted in India, , , one in Germany, and one in Spain. Three studies were randomised clinical trials, , , one of which was a split-mouth study. The other studies did not report any information on randomisation. , Blinding of operators with regard to type of suture was reported in 3 studies. , , The sample size ranged from 10 to 40 patients with ages between 18 and 60 years. Sex of participants was specified in 3 studies, with a slight preponderance of females. , , Almost all studies enrolled healthy patients, except for the one by Pelz et al, which did not report information on the health status of patients enrolled. Types of surgery were 3 dental extractions , , —2 of which included third molars , —and 2 periodontal surgeries. , All antimicrobial-coated sutures were braided (ie, multifilament) except for polyglecaprone 25 + triclosan, which was monofilament. Antimicrobial-coated sutures used were absorbable polyglactin 910 + triclosan, , resorbable polyglycolic + triclosan, polyglecaprone 25 + triclosan, polyglactin 910 + chlorhexidine, and polyglicolic + chlorhexidine. Control groups used absorbable polyglactin 910, , , braided natural black silk, and resorbable polyglycolic sture materials. Suture removal time ranged from a minimum of 3 days to a maximum of 8 days. Postoperative rinses were suggested in 2 studies, , one of which used warm water and the other which used saline solution. No postoperative antibiotic therapy was administered in one study, and another study reported antibiotic therapy until 7 days after surgery ; 3 studies did not clarify whether antibiotics were used , or the time frame for which they were prescribed. In all studies, the method to assess antimicrobial activity was the evaluation of the number of bacterial colonies detected on the sutures after their removal, with morphologic bacteria characterisation also carried out. , , , , Additionally, the examination of suture materials was performed with scanning electron microscopy (SEM) or confocal laser scanning microscopy (CLSM) to analyse the biofilm. Biochemical analysis and in vitro susceptibility tests were used to characterise the isolated aerobic bacteria and to determine the antibacterial effect of the coated suture, respectively. Summary of findings Overall, significantly reduced bacterial count was reported for antimicrobial-coated sutures compared with noncoated sutures in 3 studies, , , whilst 2 studies did not show significant differences. , Specifically, chlorhexidine-coated sutures were reported to be effective in reducing bacterial adherence in one study and not significantly different from noncoated sutures in another one. Triclosan-coated sutures showed greater antimicrobial activity compared with noncoated sutures in 3 studies , , and did not result in any significant differences in the study by Pelz et al. In addition, significantly less growth of aerobe and anaerobe species with antimicrobial-coated sutures compared with noncoated sutures was reported in one study, whilst no significant difference emerged in 2 other studies. , Kruthi et al reported a significant reduction in anaerobic bacteria adherence with triclosan-coated compared with noncoated sutures, with no significant difference in aerobes. Lower concentrations of aerobes and anaerobes were shown for triclosan-coated sutures compared with chlorhexidine-coated and noncoated sutures in the study by Karde et al, yet no significant difference was detected between chlorhexidine-coated and noncoated sutures.
The electronic search resulted in 150 articles. After exclusion of duplicates, 129 abstracts were reviewed and the full texts of 10 studies were screened. Finally, 5 studies were included for qualitative analysis , , , , ( ). The characteristics of the included studies are shown in the .
The selected studies were published from 2014 to 2019; 3 were conducted in India, , , one in Germany, and one in Spain. Three studies were randomised clinical trials, , , one of which was a split-mouth study. The other studies did not report any information on randomisation. , Blinding of operators with regard to type of suture was reported in 3 studies. , , The sample size ranged from 10 to 40 patients with ages between 18 and 60 years. Sex of participants was specified in 3 studies, with a slight preponderance of females. , , Almost all studies enrolled healthy patients, except for the one by Pelz et al, which did not report information on the health status of patients enrolled. Types of surgery were 3 dental extractions , , —2 of which included third molars , —and 2 periodontal surgeries. , All antimicrobial-coated sutures were braided (ie, multifilament) except for polyglecaprone 25 + triclosan, which was monofilament. Antimicrobial-coated sutures used were absorbable polyglactin 910 + triclosan, , resorbable polyglycolic + triclosan, polyglecaprone 25 + triclosan, polyglactin 910 + chlorhexidine, and polyglicolic + chlorhexidine. Control groups used absorbable polyglactin 910, , , braided natural black silk, and resorbable polyglycolic sture materials. Suture removal time ranged from a minimum of 3 days to a maximum of 8 days. Postoperative rinses were suggested in 2 studies, , one of which used warm water and the other which used saline solution. No postoperative antibiotic therapy was administered in one study, and another study reported antibiotic therapy until 7 days after surgery ; 3 studies did not clarify whether antibiotics were used , or the time frame for which they were prescribed. In all studies, the method to assess antimicrobial activity was the evaluation of the number of bacterial colonies detected on the sutures after their removal, with morphologic bacteria characterisation also carried out. , , , , Additionally, the examination of suture materials was performed with scanning electron microscopy (SEM) or confocal laser scanning microscopy (CLSM) to analyse the biofilm. Biochemical analysis and in vitro susceptibility tests were used to characterise the isolated aerobic bacteria and to determine the antibacterial effect of the coated suture, respectively.
Overall, significantly reduced bacterial count was reported for antimicrobial-coated sutures compared with noncoated sutures in 3 studies, , , whilst 2 studies did not show significant differences. , Specifically, chlorhexidine-coated sutures were reported to be effective in reducing bacterial adherence in one study and not significantly different from noncoated sutures in another one. Triclosan-coated sutures showed greater antimicrobial activity compared with noncoated sutures in 3 studies , , and did not result in any significant differences in the study by Pelz et al. In addition, significantly less growth of aerobe and anaerobe species with antimicrobial-coated sutures compared with noncoated sutures was reported in one study, whilst no significant difference emerged in 2 other studies. , Kruthi et al reported a significant reduction in anaerobic bacteria adherence with triclosan-coated compared with noncoated sutures, with no significant difference in aerobes. Lower concentrations of aerobes and anaerobes were shown for triclosan-coated sutures compared with chlorhexidine-coated and noncoated sutures in the study by Karde et al, yet no significant difference was detected between chlorhexidine-coated and noncoated sutures.
Main findings In terms of surgical complications, postoperative wound infections remain the second most prevalent issue. The majority of cases of surgical site infections occur in the area surrounding the incision, especially when suture material is present. Given the considerable infection risk, contemporary research has focussed on preventing bacterial colonisation of medical materials by using antibacterial coatings. The purpose of this scoping review was to evaluate the antimicrobial activity of antimicrobial-coated sutures in oral surgery. A scoping review is a methodological approach to define the current boundaries of a large and recent topic and provide useful indications for future research. , , The antibacterial efficacy of sutures coated with antibacterial material is a topic of clinical interest in dentistry. Accordingly, the primary purpose of this scoping review was to identify the currently available studies assessing the antimicrobial activity of antimicrobial-coated sutures in oral surgery and the major limitations of the current research. For this reason, all available human clinical trials were included, regardless of study design and their reported quality. A 10-year time limit has been set in database research in order to ensure that the most recent information was selected and provided to researchers and clinicians. There is still controversy regarding the impact of the physical and chemical characteristics of sutures on the risk of surgical site infection. , In the included studies, antimicrobial-coated sutures were almost all braided, with the exception of the polyglecaprone 25 + triclosan suture. Antibacterial coatings used triclosan , , , and chlorhexidine. , Triclosan is a broad-spectrum antimicrobial agent active against Gram-positive and Gram-negative bacteria. Although it is most often used for antisepsis of the skin and other surfaces, incorporation of triclosan into medical devices and dentifrices has well confirmed its effective intraoral use too. , Chlorhexidine is another commonly used antimicrobial agent useful for management of oral diseases. At low concentrations chlorhexidine is bacteriostatic, and at high concentrations it acts in a bactericidal manner, causing cell death by cytolysis. , Both agents, therefore, exhibit good antibacterial properties and safety, , making them suitable as antibacterial coatings for oral sutures. Yet, the results regarding their effectiveness in reducing bacteria counts are controversial. The included studies showed remarkable methodological differences, making difficult a comparison amongst them. Two of the 5 included studies did not report details on procedures of patients’ enrollment and randomisation. , Randomised trials represent the best methodological condition to make 2 groups comparable and, consequently, this study design is preferred when the effect of one or more interventions is investigated. Furthermore, sample size calculation was reported in only 2 studies. , The number of patients included should be suitable for the objective to be evaluated; a low number of patients limits the generalisability of the results obtained. Notably, some of the included studies did not specify sex of the enrolled participants. , This could introduce a selection bias, altering the results obtained. Indeed, sex may be considered a confounding factor, as hormonal differences may alter the results by modifications in immune response and blood flow. , In order to minimise bias, immunocompromised patients were in general excluded from these kinds of studies because immunosuppression is known to increase the risk of infections and bacterial adherence when compared to healthy participants. All patients should be treated in clean conditions through preventive and appropriate oral hygiene sessions to control the environmental setting as well as possible. Two studies reported that intervention was performed after oral prophylaxis procedures. , Furthermore, the potential mouth rinses after the surgical intervention may also impact the results. With regards to this, 2 studies specified that recruited patients were invited to perform mouth rinses after surgery with warm water and saline solutions. Warm saline rinses have a bacteriostatic action, promote healing, and limit the development of postextraction alveolitis. , Warm saline is preferred over chlorhexidine mouthwash because of chlorhexidine's antibacterial action , and the consequent possible influence on the number of colonies on the sutures’ surface. The possible risk factors that interfere with tissue healing and thus with bacteria growth and accumulation include smoking and alcohol use as well as drug abuse and previous radiotherapy/chemotherapy. Smoking impairs circulation by constricting blood vessels and reducing blood flow, affecting the outcome of surgical therapy. Alcohol intake also compromises the healing of oral wounds ; for this reason, all patients who use alcohol should be excluded. A total of 3 studies mentioned some of the above risk factors as exclusion criteria during participants’ recruitment, in particular smoking , and alcohol/drugs intake. Antibiotic therapy could act as a confounder and increase the risk for bacterial resistance phenomena. For these reasons, pre- and postoperative antibiotic therapy should not be prescribed for patients enrolled in these kind of trials in order to determine the effective impact of the antibacterial coating without confounders and to avoid bacterial resistance phenomena. , Yet, Asher et al reported that postsurgical antibiotic treatment had only a minor effect on bacterial accumulation on the sutures tested. Of note, no antimicrobial-coated sutures were employed in that study; thus, further studies are needed to clarify this aspect. Types of surgery included extractions and periodontal flap surgery. According to Asher et al, type of surgery seems not to significantly influence the bacterial accumulation on sutures’ surface. However, Asher et al considered periodontal surgery, implant surgery, guided bone regeneration, and second-stage implant surgery. Postoperative infections are more common in mandibular tooth extractions due to the difficulty of managing oral hygiene and the consequent increased bacteria adhesion. Additionally, the removal time of sutures varied from a minimum of 3 days to a maximum of 8 days. It is noteworthy that an extended suture persistence can promote bacterial growth. Typically, depending on the type of procedure, sutures are removed within 5 to 7 days, with the highest colonisation reported at 3 days. Operators were blinded with regard to the type of suture used in 3 studies. , , Operator blinding is a fundamental procedure for preventing observer bias in clinical trials. Yet, considering that the outcome measurement is objective and is based on the measurement of bacterial counts, it is reasonable to consider the influence of this bias to be negligible. Overall, most of the studies reported a decrease in bacterial load in coated sutures compared with the noncoated ones, whilst a significant advantage was not reported by Pelz et al in terms of reduction in microbial numbers. Of note, in the study by Sharma et al, chlorhexidine-coated sutures showed potential in preventing microbial colonisation even if the results were not statistically significant. The contrasting results are probably due to the reduced number of patients enrolled, which did not ensure adequate statistical power; to the properties and type of suture (eg, antimicrobial agent concentration in the suture); as well as to the patient's postoperative management. Odontogenic infections are caused by a wide variety of bacteria. Overall, a higher number of aerobic microorganisms compared to anaerobes was found. In addition, commensal species (eg, Streptococcus groups) were more common than pathogenic ones (eg, Prevotella and Fusobacterium spp ). This could explain the relatively low incidence of infectious complications reported. For the evaluation of the sutures, this aspect is pivotal considering the well-documented role anaerobic bacteria play in oral infections and inflammation. , Limitations Despite the rigorous methodology and the reporting according to PRISMA requirements, some limitations must be underlined. First, the great variety in the methodology of included studies made a direct comparison amongst them difficult. In addition, antimicrobial activity is a complex outcome influenced by different confounding factors that may impact results, including differences amongst participants in microbiota, inflammation status, and characteristics of sutures tested. Clinical implications and future perspectives The reduction in bacterial adherence promoted by antimicrobial-coated agents would translate into a decrease in surgical site infection and related symptoms, including swelling, postoperative pain, and oedema, especially for high-risk patients undergoing surgical treatments. The clinical impact of antimicrobial-coated sutures has not been evaluated in the current scoping review because it was out of scope; future primary and secondary studies should address this. In addition, the properties of antimicrobial-coated sutures, such as strength, knot-holding capacity, and ease of handling, as well as the precise concentration of the antibacterial agent and its drug-release profile must also be further investigated. Whilst animal and in vitro studies are fundamental to the development of clinical research, their findings may not directly translate to human outcomes. , Therefore, to ensure the relevance and applicability of the research findings to humans, these studies were excluded from the review, reflecting a cautious approach to translating preclinical results to clinical outcomes. Implications for research An optimal condition to explore the antimicrobial efficacy of coated sutures could be represented by the extraction in the same intervention (eg, as for orthodontic reasons) of the lower third molars in total osteomucosal inclusion and free of local infections. Furthermore, performing the extractions on the same day would minimise the confounding effect of the possible change in the oral microbiota following tooth extractions. , Sutures to be compared should be of the same diameter and type, monofilament or multifilament. Braided sutures, which retain more plaque and are less easily cleanable, are preferred in clinical research to adequately test antibacterial activity. Furthermore, sutures should be ideally tested after 3 days to find a greater number of bacterial colonies and thus compare the 2 materials when the microbial colonisation is greater.
In terms of surgical complications, postoperative wound infections remain the second most prevalent issue. The majority of cases of surgical site infections occur in the area surrounding the incision, especially when suture material is present. Given the considerable infection risk, contemporary research has focussed on preventing bacterial colonisation of medical materials by using antibacterial coatings. The purpose of this scoping review was to evaluate the antimicrobial activity of antimicrobial-coated sutures in oral surgery. A scoping review is a methodological approach to define the current boundaries of a large and recent topic and provide useful indications for future research. , , The antibacterial efficacy of sutures coated with antibacterial material is a topic of clinical interest in dentistry. Accordingly, the primary purpose of this scoping review was to identify the currently available studies assessing the antimicrobial activity of antimicrobial-coated sutures in oral surgery and the major limitations of the current research. For this reason, all available human clinical trials were included, regardless of study design and their reported quality. A 10-year time limit has been set in database research in order to ensure that the most recent information was selected and provided to researchers and clinicians. There is still controversy regarding the impact of the physical and chemical characteristics of sutures on the risk of surgical site infection. , In the included studies, antimicrobial-coated sutures were almost all braided, with the exception of the polyglecaprone 25 + triclosan suture. Antibacterial coatings used triclosan , , , and chlorhexidine. , Triclosan is a broad-spectrum antimicrobial agent active against Gram-positive and Gram-negative bacteria. Although it is most often used for antisepsis of the skin and other surfaces, incorporation of triclosan into medical devices and dentifrices has well confirmed its effective intraoral use too. , Chlorhexidine is another commonly used antimicrobial agent useful for management of oral diseases. At low concentrations chlorhexidine is bacteriostatic, and at high concentrations it acts in a bactericidal manner, causing cell death by cytolysis. , Both agents, therefore, exhibit good antibacterial properties and safety, , making them suitable as antibacterial coatings for oral sutures. Yet, the results regarding their effectiveness in reducing bacteria counts are controversial. The included studies showed remarkable methodological differences, making difficult a comparison amongst them. Two of the 5 included studies did not report details on procedures of patients’ enrollment and randomisation. , Randomised trials represent the best methodological condition to make 2 groups comparable and, consequently, this study design is preferred when the effect of one or more interventions is investigated. Furthermore, sample size calculation was reported in only 2 studies. , The number of patients included should be suitable for the objective to be evaluated; a low number of patients limits the generalisability of the results obtained. Notably, some of the included studies did not specify sex of the enrolled participants. , This could introduce a selection bias, altering the results obtained. Indeed, sex may be considered a confounding factor, as hormonal differences may alter the results by modifications in immune response and blood flow. , In order to minimise bias, immunocompromised patients were in general excluded from these kinds of studies because immunosuppression is known to increase the risk of infections and bacterial adherence when compared to healthy participants. All patients should be treated in clean conditions through preventive and appropriate oral hygiene sessions to control the environmental setting as well as possible. Two studies reported that intervention was performed after oral prophylaxis procedures. , Furthermore, the potential mouth rinses after the surgical intervention may also impact the results. With regards to this, 2 studies specified that recruited patients were invited to perform mouth rinses after surgery with warm water and saline solutions. Warm saline rinses have a bacteriostatic action, promote healing, and limit the development of postextraction alveolitis. , Warm saline is preferred over chlorhexidine mouthwash because of chlorhexidine's antibacterial action , and the consequent possible influence on the number of colonies on the sutures’ surface. The possible risk factors that interfere with tissue healing and thus with bacteria growth and accumulation include smoking and alcohol use as well as drug abuse and previous radiotherapy/chemotherapy. Smoking impairs circulation by constricting blood vessels and reducing blood flow, affecting the outcome of surgical therapy. Alcohol intake also compromises the healing of oral wounds ; for this reason, all patients who use alcohol should be excluded. A total of 3 studies mentioned some of the above risk factors as exclusion criteria during participants’ recruitment, in particular smoking , and alcohol/drugs intake. Antibiotic therapy could act as a confounder and increase the risk for bacterial resistance phenomena. For these reasons, pre- and postoperative antibiotic therapy should not be prescribed for patients enrolled in these kind of trials in order to determine the effective impact of the antibacterial coating without confounders and to avoid bacterial resistance phenomena. , Yet, Asher et al reported that postsurgical antibiotic treatment had only a minor effect on bacterial accumulation on the sutures tested. Of note, no antimicrobial-coated sutures were employed in that study; thus, further studies are needed to clarify this aspect. Types of surgery included extractions and periodontal flap surgery. According to Asher et al, type of surgery seems not to significantly influence the bacterial accumulation on sutures’ surface. However, Asher et al considered periodontal surgery, implant surgery, guided bone regeneration, and second-stage implant surgery. Postoperative infections are more common in mandibular tooth extractions due to the difficulty of managing oral hygiene and the consequent increased bacteria adhesion. Additionally, the removal time of sutures varied from a minimum of 3 days to a maximum of 8 days. It is noteworthy that an extended suture persistence can promote bacterial growth. Typically, depending on the type of procedure, sutures are removed within 5 to 7 days, with the highest colonisation reported at 3 days. Operators were blinded with regard to the type of suture used in 3 studies. , , Operator blinding is a fundamental procedure for preventing observer bias in clinical trials. Yet, considering that the outcome measurement is objective and is based on the measurement of bacterial counts, it is reasonable to consider the influence of this bias to be negligible. Overall, most of the studies reported a decrease in bacterial load in coated sutures compared with the noncoated ones, whilst a significant advantage was not reported by Pelz et al in terms of reduction in microbial numbers. Of note, in the study by Sharma et al, chlorhexidine-coated sutures showed potential in preventing microbial colonisation even if the results were not statistically significant. The contrasting results are probably due to the reduced number of patients enrolled, which did not ensure adequate statistical power; to the properties and type of suture (eg, antimicrobial agent concentration in the suture); as well as to the patient's postoperative management. Odontogenic infections are caused by a wide variety of bacteria. Overall, a higher number of aerobic microorganisms compared to anaerobes was found. In addition, commensal species (eg, Streptococcus groups) were more common than pathogenic ones (eg, Prevotella and Fusobacterium spp ). This could explain the relatively low incidence of infectious complications reported. For the evaluation of the sutures, this aspect is pivotal considering the well-documented role anaerobic bacteria play in oral infections and inflammation. ,
Despite the rigorous methodology and the reporting according to PRISMA requirements, some limitations must be underlined. First, the great variety in the methodology of included studies made a direct comparison amongst them difficult. In addition, antimicrobial activity is a complex outcome influenced by different confounding factors that may impact results, including differences amongst participants in microbiota, inflammation status, and characteristics of sutures tested.
The reduction in bacterial adherence promoted by antimicrobial-coated agents would translate into a decrease in surgical site infection and related symptoms, including swelling, postoperative pain, and oedema, especially for high-risk patients undergoing surgical treatments. The clinical impact of antimicrobial-coated sutures has not been evaluated in the current scoping review because it was out of scope; future primary and secondary studies should address this. In addition, the properties of antimicrobial-coated sutures, such as strength, knot-holding capacity, and ease of handling, as well as the precise concentration of the antibacterial agent and its drug-release profile must also be further investigated. Whilst animal and in vitro studies are fundamental to the development of clinical research, their findings may not directly translate to human outcomes. , Therefore, to ensure the relevance and applicability of the research findings to humans, these studies were excluded from the review, reflecting a cautious approach to translating preclinical results to clinical outcomes.
An optimal condition to explore the antimicrobial efficacy of coated sutures could be represented by the extraction in the same intervention (eg, as for orthodontic reasons) of the lower third molars in total osteomucosal inclusion and free of local infections. Furthermore, performing the extractions on the same day would minimise the confounding effect of the possible change in the oral microbiota following tooth extractions. , Sutures to be compared should be of the same diameter and type, monofilament or multifilament. Braided sutures, which retain more plaque and are less easily cleanable, are preferred in clinical research to adequately test antibacterial activity. Furthermore, sutures should be ideally tested after 3 days to find a greater number of bacterial colonies and thus compare the 2 materials when the microbial colonisation is greater.
Within the limits of this scoping review, antibacterial-coated sutures used in oral surgery showed less bacterial retention compared to noncoated sutures. However, the notable methodological variability, the available studies with small sizes, and the numerous confounding factors identified limit the generalisability and reliability of these results. Thus, high-quality randomised clinical studies with large sample sizes are required to draw firm conclusions.
Giusy Rita Maria La Rosa: Conceptualization, Formal analysis, Methodology, Software, Writing – original draft. Simone Scapellato: Conceptualization, Formal analysis, Investigation, Methodology, Software, Writing – original draft. Marco Cicciù: Data curation, Supervision, Validation, Visualization, Writing – review & editing. Eugenio Pedullà: Data curation, Formal analysis, Project administration, Resources, Supervision, Writing – review & editing.
None disclosed.
|
Biogeographic survey of soil bacterial communities across Antarctica | 9042dc06-8f4a-4086-afe6-c5e3b5bec6ba | 10785390 | Microbiology[mh] | Antarctica is the coldest, windiest, and driest continent, posing some of the harshest and most challenging conditions for life . The vast majority of the continent is permanently covered by ice, with less than 0.3% of its continental area being ice-free . However, these areas include a wide range of geographical/geological features and a total area of over 20,000 km 2 [ – ]. Ice-free areas are typically small exposed islands of land surrounded by ice or ocean, with those hosting most biodiversity mainly located in proximity to the coast along the Antarctic Peninsula and its associated Scotia Arc archipelagoes, and in coastal oases around the edge of the Antarctic continent. They also include the hyperarid deserts of the Victoria Land Dry Valleys (individually the largest ice-free areas in Antarctica), large inland mountain ranges such as the Transantarctic Mountains, and Ellsworth Mountains, and smaller ranges and isolated inland nunataks [ – ]. These systems are generally characterized by exposure to a combination of extreme environmental stresses, including high UV radiation, low water availability, high salinity, low temperatures, and low nutrient availability, posing multiple challenges to life . The Antarctic continent is commonly divided into maritime and continental zones in descriptions of its biodiversity [ , , ]. Under this definition, maritime Antarctica comprises the western coastal regions and offshore islands of the Antarctic Peninsula and the Scotia Arc archipelagoes of the South Shetland, South Orkney, and South Sandwich Islands, plus the isolated oceanic islands of Bouvet and Peter I . The much larger area of continental Antarctica comprises the geological regions of East and West Antarctica and their offshore islands, plus the eastern coastal regions of the Antarctic Peninsula. However, the latter was historically included under the “continental Antarctica” definition primarily due to the comparability of its climatic conditions, as virtually no biological survey and diversity data exist for the eastern Peninsula. Beyond the maritime and continental Antarctic regions, the sub-Antarctic region lies at mid-latitudes in the Southern Ocean and consists of a range of isolated islands generally in proximity to the Antarctic Polar Front . Beyond the “core” sub-Antarctic, further peri-Antarctic islands are recognized at lower latitudes, whose biodiversity overlaps with that of the formally defined sub-Antarctic Islands . The maritime Antarctica is characterized by slightly higher summer temperatures and greater precipitation compared to continental Antarctica, while the sub-Antarctic islands are generally cool with only limited seasonal variation in environmental conditions and a considerable proportion of precipitation as rain . In the continental Antarctica, precipitation occurs almost entirely in the form of snow and is often extremely limited. In some parts of the continent where low atmospheric humidity is typical, much of this snow undergoes sublimation and is lost from terrestrial ecosystems before wetting of the underlying soil can occur. Common features of continental Antarctic soils are low moisture and nutrient concentrations, subzero temperatures for an extended period of each annual cycle, and frequent soil freeze–thaw cycles during the austral summer . Antarctic soils are also characterized by different formation histories and environmental factors (e.g., humidity, temperature, UV radiation, salinity, pH and carbon, and nutrient availability) [ , , , – ]. These diverse climate characteristics and soil properties influence macro- and micro-biological diversity patterns [ , , ]. The harsh environmental conditions mean that the continent hosts only two native species of vascular plants, present exclusively on the Antarctic Peninsula and maritime Antarctic archipelagoes, where there is a much higher diversity and wider distribution of bryophytes and lichens compared to vascular plants [ , – ]. Marine vertebrate breeding colonies, and haul-out and molting areas, are present around the continent, with the vast majority in close proximity to the coast, where soil nutrients are locally enriched in marine-derived nutrients introduced by animal activity [ – ]. Invertebrate groups, such as springtails, mites, tardigrades, rotifers and nematodes, are often common but with varying patterns of species distributions across the continent [ , – ]. Some groups that are otherwise widely regarded to be cosmopolitan, such as nematodes, appear to be completely absent in certain regions even though other micro-invertebrates are present , and many show high degrees of species-level endemism at various spatial scales [ , , – ]. Compared to plants and animals, microorganisms (e.g., bacteria, archaea, fungi) are numerically and phylogenetically dominant across the Antarctic continent [ , – ]. The unique biodiversity hosted by this largely pristine continent is currently threatened by a number of factors , including climate change where new climatic conditions could destabilize biotic equilibria , the introduction of new non-native species , and direct physical human impacts [ – ]. The unique and largely pristine nature of terrestrial Antarctica (south of 60°) has been subject to Antarctic Treaty governance since 1961 and is currently protected under the Protocol on Environmental Protection to the Antarctic Treaty, whose requirements have led to several international research initiatives focusing on the current and future protection and conservation of the region [ – ]. Drawing on this initiative, the cataloged distribution patterns of micro- and macro-invertebrates, plants, algae, and some microorganisms have been used to define a series of Antarctic Conservation Biogeographic Regions (ACBRs) . The current definition of 16 ACBRs was based on the distribution of 1823 taxa derived from 38,854 biological records and expert-defined bioregions. However, it was also recognized that there was uneven distribution of data across both taxa and regionally within Antarctica. The available data represented vascular and non-vascular plants (4 and 258 taxa, respectively), metazoa (153 taxa), multi-cellular algae (182 taxa), eukaryotic microalgae and protists (283 taxa), fungi (760 taxa), and bacteria (183 taxa). It is notable that only 10% of the available taxa and 3% of the records belonged to the bacterial domain, and only to the phyla Actinomycetota and Cyanobacteriota . A good fit into the ACBR frame has been reported for diatoms in Antarctic lakes , but no previous studies have been reported that test the distribution of much broader taxonomic groups (i.e., total microbiomes) within the context of the ACRB framework. Further, to date, no studies analyzing Antarctic bacterial community patterns at a continental scale have been conducted using data generated by high-throughput sequencing technology, despite the general upsurge in the application of these modern molecular biological approaches. Regional and local biogeographic surveys suggest that prokaryotic distribution patterns are complex, varying in relation to the studied area and scale, driven by both abiotic and biotic factors, and potentially impacted by different dispersal mechanisms [ , , , , – ]. High soil microbial community spatial heterogeneity, arising from variability in edaphic factors, has been observed in some Antarctic areas such as the continental Antarctic Ross Sea region [ , , , ] and maritime Antarctic Signy Island . Some prokaryotic taxa are potentially endemic to the Antarctic continent or specific Antarctic areas [ , – ], while others show cosmopolitan distributions, possibly indicative of different past and ongoing dispersal strategies . Despite the existence of multiple prokaryote phylogenetic datasets from many Antarctic ice-free areas, very few of these data have yet contributed to the development of the ACBR classifications. In this study, we set out to conduct the first continental-scale biogeographic survey of soil bacterial communities across Antarctica. We used 16S rRNA gene amplicon datasets to investigate bacterial phylogenetic patterns at a continental scale (i.e., distance-decay analysis) and assess major environmental drivers likely to contribute to these patterns (e.g., temperature, precipitation). Our aim was to assess whether any bacterial distribution patterns identified were consistent with and/or could provide a valuable addition to the current ACBR classification system. Furthermore, we compared how bacterial communities from mainland and island samples differed and investigated how these communities were shaped by bioclimatic variables such as air temperature and precipitation.
Dataset creation Our dataset comprised 1164 samples from 17 16S rRNA gene sequencing datasets (Table S ) derived from shallow (0–10 cm depth) soils. Soil storage conditions and DNA extraction methods are reported in Table S . These 17 datasets were all derived from the 16S rRNA gene but spanned five distinct 16S rRNA regions, viz . , V1–V3, V3–V4, V4, V4–V5, and V8–V9. The primer pairs used for amplification were 27F–519R and pA–BKL1 for V1–V3, 341F–805R and 341F–806R for V3–V4, 515F–806R for V4, 515F–926R for V4–V5, and 926F–1392wR for V8–V9 (Table S ). All datasets were collated from online repositories and collaborators except for dataset 10, which was obtained by resequencing samples collected from the Prince Charles Mountains and coastal areas in Eastern Antarctica ACBR [ – ]. 16S rDNA amplicon libraries were prepared using the KAPA HiFi PCR kit (Roche) and sequenced using Illumina MiSeq technology (paired-end, 300 cycles) by Omega Bioservices (Norcross, USA). Data availability for all datasets is reported in Table S . Bioclimatic variables and metadata Bioclimatic variables (1981–2010) were extracted from the CHELSA database v 2.1 using the R package terra v 1.6–47 in the R environment v 4.1.3 . The extracted bioclimatic variables were BIO1 (mean annual air temperature, °C), BIO2 (mean diurnal air temperature range, °C), BIO4 (temperature seasonality, °C/100), BIO5 (mean daily maximum air temperature of the warmest month, °C), BIO10 (mean daily mean air temperatures of the warmest quarter, °C), BIO12 (annual precipitation, kg m −2 ), BIO14 (precipitation in the driest month, kg m −2 ), BIO15 (precipitation seasonality, %), BIO17 (mean monthly precipitation in the driest quarter, kg m −2 ), BIO18 (mean monthly precipitation in the warmest quarter, kg m −2 ), and SWE (snow water equivalent, kg m −2 ) (Table S ). Elevation values were extracted from the REMA digital elevation model (100 m DEMs) in QGIS Desktop 3.28.2 . Distance from coast and ocean for each sample point (the latter relevant in the presence of floating ice shelves) was obtained using Bedmap2 raster files in QGIS (Table S ). Elevation and distance from coast/ocean data were used only for analyses focused on mainland samples. All maps reported in this work were created using QGIS Desktop 3.28.2 in Quantarctica . Sample classification into ACRBs Classically, terrestrial Antarctica is considered to include three broad biogeographic regions: the sub-, maritime, and continental Antarctic . Chown and Convey (2007) redefined the boundary between the latter two regions by their definition of the “Gressitt Line” as an important biogeographic boundary at the base of the Antarctic Peninsula . More recently, the regions have been further divided into 16 distinct “Antarctic Conservation Biogeographic Regions” (ACBRs) , a classification that applies specifically to the area of Antarctic Treaty governance south of the 60° latitude parallel. The samples in the 17 datasets represented here were obtained from 10 of the 16 currently recognized ACBRs: 1, 3, 4, 6–10, 12, and 16, as well as several of the sub- and peri-Antarctic islands (Fig. ; Table S ) [ , , – ]. Our dataset includes 846 samples from the Antarctic mainland, 129 samples from islands and associated archipelagos of the Antarctic mainland, and 13 samples from the sub- and peri-Antarctic islands. Excluding samples from Bouvetøya, Peter I Øya, and Scott Island (accounting for 11 samples), all samples from the Antarctic mainland and its associated islands and archipelagos are included in the ACBR classification. The 118 samples obtained from islands and archipelagos associated with mainland Antarctica and included in the ACBR classification (ACBRs 1, 3, 4, 6, 7, 9, 12, and 16) are shown in Figure S and represent: James Ross Island (ACBR 1); South Shetland Islands (e.g., Livingston Island), Palmer Archipelago (e.g., Anvers Island), Adelaide Island and islands in its proximity, and Alectoria Island (ACBR 3); Alexander Island (ACBR 4); an un-named island in the Jelbart ice shelf (ACBR 6); Herring Island (Windmill Islands) and Hop Island (ACBR 7); Ross Island (ACBR 9); Siple, Lauff and Maher Islands (ACBR 12); and Samson Island (ACBR 16). The sub- and peri-Antarctic islands sampled included Marion, Possession, Kerguelen, Bartolomé, and South Georgia; sub- and peri-Antarctic islands are not included in the ACBR classification (Figure S ). Samples from these islands were considered together with those from Bouvetøya (formally included in the maritime Antarctic), Peter I Øya, and Scott Island (islands in the area of Treaty governance but not included in the ACBR classification) in the analyses described below and are referred to below as “ACBR unclassified islands” (AUI) for brevity. The AUI group does not aim to group islands on a bioregional basis; instead, it simply groups islands that did not fall into any ACBR to facilitate further analyses. Data analyses All the 16S rRNA gene datasets except datasets 6 and 7 comprised raw forward and reverse reads. Data for datasets 6 and 7 (i.e., Victoria Land, Antarctic Peninsula and sub- and peri Antarctic island samples) were retrieved as pre-merged forward and reverse reads. For all other datasets, adapters were trimmed using Trimmomatic v 0.39 . Each dataset was then separately analyzed in the R environment v 4.1.3 . Sequences were processed with the same pipeline using dada2 v 1.22 . Parameters “trimRight”, “trimLeft”, and “maxEE” from the function filterAndTrim() were set accordingly for the read length and error profile specific to each library (Table S ). Taxonomy was assigned using the SILVA database v 138.1 . Blanks were present in datasets 1, 8, 10, and 16 and were removed using decontam v 1.14.0 . All datasets were combined using the R package phyloseq v 1.38 . Only sequences assigned to Bacteria, excluding mitochondrial and chloroplast DNA, were retained. Sequences assigned to the domain Archaea were excluded because previous tests demonstrated that the primer sets used in this study differentially targeted archaeal organisms . Only samples with a sequence count of > 5000 were retained . Due to the high disparity in sample read depth ( n = 5247–988,658 reads) (Table S ), the dataset was normalized using the R package SRS v 0.2.3 using the sample with the lowest amplicon counts ( n = 5247) as the reference. Two samples in which all reads were assigned to unknown taxa were removed from the dataset after exploring their taxonomy as 100% of the reads were assigned to unknown taxa in those samples. Therefore, of the initial 1164 samples, 988 were used in further analyses. All of the following analyses were performed at the genus level because this composite dataset, comprising of several amplicon datasets spanning different 16S rRNA regions, did not allow accurate taxonomic analysis at the amplicon sequence variant (ASV) level . Alpha diversity indices were generated using vegan v 2.6–4 to calculate richness and Shannon indices on the genus-level taxonomic dataset. To test whether diversity indices differed across the defined regions, analysis of variance (ANOVA) was performed using the function aov followed, when significant, by Tukey’s honest significant difference (HSD) statistical tests to obtain pairwise comparisons. A conservative significance level of p < 0.01 was considered significant. ANOVA was only performed on bioclimatic variables from ACBRs 3, 6–10, and 16, and the ACBR unclassified islands (AUI) which were represented by at least 20 samples. Permutational multivariate ANOVA (PERMANOVA) was performed using adonis2 (vegan), again using data only from ACBRs 3, 6–10, and 16, and the AUI which were represented by at least 20 samples. Pairwise comparison statistics were obtained using pairwiseAdonis v 0.4 , and p values were adjusted using the false discovery rate method (FDR) . Both PERMANOVA and pairwise comparisons were calculated using 1000 permutations and calculated on Bray–Curtis dissimilarity matrices obtained from taxonomic datasets at the genus level, excluding reads that could not be classified at this level. Bioclimatic variables were first standardized using the function decostand (vegan) and then checked for collinearity with vif.cca, and only variables with VIF < 20 were retained (maximum VIF retained = 12.33). The function ordiR2step was run to perform stepwise model building for distance-based redundancy analysis (dbRDA). The significance of the tested variables was obtained using anova.cca. Variation partitioning was performed with varpart on the same variables as selected from dbRDA. When variation partitioning was applied at the single ACBR level, only bioclimatic variables highlighted by dbRDA analysis performed on the entire dataset (BIO2, BIO4, BIO10, BIO15, and BIO18) were used. Geographical coordinates were transformed calculating principal coordinates of neighbor matrices using the function pcnm. Principal Coordinate Analysis (PCoA) was performed using the pcoa function from the R library ape v 5.6–2 . The function envfit (vegan) was then used to calculate ACBR and bioclimatic variable multiple regressions with the PCoA ordination axes. Only bioclimatic regions identified as significant by the function anova.cca were used in the envfit analysis. Correlations between the genus-level taxonomic dataset and geographical distance (distance-decay) and bioclimatic dataset were assessed by calculating distance matrices using the vegan function vegdist for the community and bioclimatic datasets (Bray–Curtis and Euclidean distances, respectively) and distm for the geographical coordinates using the R library geosphere v 1.5–18 . Mantel tests were then calculated using Spearman’s correlation and 1000 permutations. Dendrograms were created by calculating distance matrices as previously described and then by running the functions hclust and as.dendogram in the R package dendextend v 1.16.0 . Dendrograms were combined with the function tanglegram. Detailed taxonomic analyses were performed only on taxonomically consistent datasets derived from the amplification of the V3–V4 and V4 regions of the 16S rRNA gene (503 samples) . This dataset subset represented 7 out of the 10 analyzed ACBRs (plus AUI) and contained samples from the Antarctic Peninsula, the Transantarctic Mountains, and the Victoria Land. No samples from Eastern Antarctica or Dronning Maud Land were included. The dominant community was defined as including all genera with a relative abundance of > 1% in at least one sample that was present in at least 10% of samples. The 10% threshold was chosen to include genera present exclusively on the Antarctic Peninsula. Random forest analysis was performed using the R package randomForest v 4.7 on the genus-level dominant community dataset, which was transformed to relative abundance. Genera with “explained variance” higher than 30% using the randomForest algorithm were then used for Spearman’s semi-partial correlation with bioclimatic variables using the R packages ppcor v 1.1 and rfPermute v 2.5.1 . Correlations with p < 0.01 were considered significant. Only bioclimatic variables highlighted in dbRDA were used in these analyses to exclude collinear variables and therefore avoid redundancy (SWE, BIO2, BIO4, BIO10, BIO5, and BIO18). Genus relative abundances in each ACBR and the AUI were obtained by summing the reads of samples from the same ACBRs and then calculating relative abundances for each of them. Linear discriminant analysis effect size (LEfSe) analysis based on Kruskal–Wallis tests ( p < 0.01) was performed on this data subset using the R package microeco v 1.1.0 . Holm correction was used to adjust p values . Figures were plotted using the R packages ggplot2 v 3.3.5 , gplots v 3.1.3 , gridExtra v 2.3 , and Cairo v 1.5.12.2 . Data manipulation was carried out using default R packages and usedist v 0.4.0.9000 . The R scripts used for the analysis of the sequencing data can be found on the GitHub page https://github.com/gvMicroarctic/AntarcticBiogeographyPaper .
Our dataset comprised 1164 samples from 17 16S rRNA gene sequencing datasets (Table S ) derived from shallow (0–10 cm depth) soils. Soil storage conditions and DNA extraction methods are reported in Table S . These 17 datasets were all derived from the 16S rRNA gene but spanned five distinct 16S rRNA regions, viz . , V1–V3, V3–V4, V4, V4–V5, and V8–V9. The primer pairs used for amplification were 27F–519R and pA–BKL1 for V1–V3, 341F–805R and 341F–806R for V3–V4, 515F–806R for V4, 515F–926R for V4–V5, and 926F–1392wR for V8–V9 (Table S ). All datasets were collated from online repositories and collaborators except for dataset 10, which was obtained by resequencing samples collected from the Prince Charles Mountains and coastal areas in Eastern Antarctica ACBR [ – ]. 16S rDNA amplicon libraries were prepared using the KAPA HiFi PCR kit (Roche) and sequenced using Illumina MiSeq technology (paired-end, 300 cycles) by Omega Bioservices (Norcross, USA). Data availability for all datasets is reported in Table S .
Bioclimatic variables (1981–2010) were extracted from the CHELSA database v 2.1 using the R package terra v 1.6–47 in the R environment v 4.1.3 . The extracted bioclimatic variables were BIO1 (mean annual air temperature, °C), BIO2 (mean diurnal air temperature range, °C), BIO4 (temperature seasonality, °C/100), BIO5 (mean daily maximum air temperature of the warmest month, °C), BIO10 (mean daily mean air temperatures of the warmest quarter, °C), BIO12 (annual precipitation, kg m −2 ), BIO14 (precipitation in the driest month, kg m −2 ), BIO15 (precipitation seasonality, %), BIO17 (mean monthly precipitation in the driest quarter, kg m −2 ), BIO18 (mean monthly precipitation in the warmest quarter, kg m −2 ), and SWE (snow water equivalent, kg m −2 ) (Table S ). Elevation values were extracted from the REMA digital elevation model (100 m DEMs) in QGIS Desktop 3.28.2 . Distance from coast and ocean for each sample point (the latter relevant in the presence of floating ice shelves) was obtained using Bedmap2 raster files in QGIS (Table S ). Elevation and distance from coast/ocean data were used only for analyses focused on mainland samples. All maps reported in this work were created using QGIS Desktop 3.28.2 in Quantarctica .
Classically, terrestrial Antarctica is considered to include three broad biogeographic regions: the sub-, maritime, and continental Antarctic . Chown and Convey (2007) redefined the boundary between the latter two regions by their definition of the “Gressitt Line” as an important biogeographic boundary at the base of the Antarctic Peninsula . More recently, the regions have been further divided into 16 distinct “Antarctic Conservation Biogeographic Regions” (ACBRs) , a classification that applies specifically to the area of Antarctic Treaty governance south of the 60° latitude parallel. The samples in the 17 datasets represented here were obtained from 10 of the 16 currently recognized ACBRs: 1, 3, 4, 6–10, 12, and 16, as well as several of the sub- and peri-Antarctic islands (Fig. ; Table S ) [ , , – ]. Our dataset includes 846 samples from the Antarctic mainland, 129 samples from islands and associated archipelagos of the Antarctic mainland, and 13 samples from the sub- and peri-Antarctic islands. Excluding samples from Bouvetøya, Peter I Øya, and Scott Island (accounting for 11 samples), all samples from the Antarctic mainland and its associated islands and archipelagos are included in the ACBR classification. The 118 samples obtained from islands and archipelagos associated with mainland Antarctica and included in the ACBR classification (ACBRs 1, 3, 4, 6, 7, 9, 12, and 16) are shown in Figure S and represent: James Ross Island (ACBR 1); South Shetland Islands (e.g., Livingston Island), Palmer Archipelago (e.g., Anvers Island), Adelaide Island and islands in its proximity, and Alectoria Island (ACBR 3); Alexander Island (ACBR 4); an un-named island in the Jelbart ice shelf (ACBR 6); Herring Island (Windmill Islands) and Hop Island (ACBR 7); Ross Island (ACBR 9); Siple, Lauff and Maher Islands (ACBR 12); and Samson Island (ACBR 16). The sub- and peri-Antarctic islands sampled included Marion, Possession, Kerguelen, Bartolomé, and South Georgia; sub- and peri-Antarctic islands are not included in the ACBR classification (Figure S ). Samples from these islands were considered together with those from Bouvetøya (formally included in the maritime Antarctic), Peter I Øya, and Scott Island (islands in the area of Treaty governance but not included in the ACBR classification) in the analyses described below and are referred to below as “ACBR unclassified islands” (AUI) for brevity. The AUI group does not aim to group islands on a bioregional basis; instead, it simply groups islands that did not fall into any ACBR to facilitate further analyses.
All the 16S rRNA gene datasets except datasets 6 and 7 comprised raw forward and reverse reads. Data for datasets 6 and 7 (i.e., Victoria Land, Antarctic Peninsula and sub- and peri Antarctic island samples) were retrieved as pre-merged forward and reverse reads. For all other datasets, adapters were trimmed using Trimmomatic v 0.39 . Each dataset was then separately analyzed in the R environment v 4.1.3 . Sequences were processed with the same pipeline using dada2 v 1.22 . Parameters “trimRight”, “trimLeft”, and “maxEE” from the function filterAndTrim() were set accordingly for the read length and error profile specific to each library (Table S ). Taxonomy was assigned using the SILVA database v 138.1 . Blanks were present in datasets 1, 8, 10, and 16 and were removed using decontam v 1.14.0 . All datasets were combined using the R package phyloseq v 1.38 . Only sequences assigned to Bacteria, excluding mitochondrial and chloroplast DNA, were retained. Sequences assigned to the domain Archaea were excluded because previous tests demonstrated that the primer sets used in this study differentially targeted archaeal organisms . Only samples with a sequence count of > 5000 were retained . Due to the high disparity in sample read depth ( n = 5247–988,658 reads) (Table S ), the dataset was normalized using the R package SRS v 0.2.3 using the sample with the lowest amplicon counts ( n = 5247) as the reference. Two samples in which all reads were assigned to unknown taxa were removed from the dataset after exploring their taxonomy as 100% of the reads were assigned to unknown taxa in those samples. Therefore, of the initial 1164 samples, 988 were used in further analyses. All of the following analyses were performed at the genus level because this composite dataset, comprising of several amplicon datasets spanning different 16S rRNA regions, did not allow accurate taxonomic analysis at the amplicon sequence variant (ASV) level . Alpha diversity indices were generated using vegan v 2.6–4 to calculate richness and Shannon indices on the genus-level taxonomic dataset. To test whether diversity indices differed across the defined regions, analysis of variance (ANOVA) was performed using the function aov followed, when significant, by Tukey’s honest significant difference (HSD) statistical tests to obtain pairwise comparisons. A conservative significance level of p < 0.01 was considered significant. ANOVA was only performed on bioclimatic variables from ACBRs 3, 6–10, and 16, and the ACBR unclassified islands (AUI) which were represented by at least 20 samples. Permutational multivariate ANOVA (PERMANOVA) was performed using adonis2 (vegan), again using data only from ACBRs 3, 6–10, and 16, and the AUI which were represented by at least 20 samples. Pairwise comparison statistics were obtained using pairwiseAdonis v 0.4 , and p values were adjusted using the false discovery rate method (FDR) . Both PERMANOVA and pairwise comparisons were calculated using 1000 permutations and calculated on Bray–Curtis dissimilarity matrices obtained from taxonomic datasets at the genus level, excluding reads that could not be classified at this level. Bioclimatic variables were first standardized using the function decostand (vegan) and then checked for collinearity with vif.cca, and only variables with VIF < 20 were retained (maximum VIF retained = 12.33). The function ordiR2step was run to perform stepwise model building for distance-based redundancy analysis (dbRDA). The significance of the tested variables was obtained using anova.cca. Variation partitioning was performed with varpart on the same variables as selected from dbRDA. When variation partitioning was applied at the single ACBR level, only bioclimatic variables highlighted by dbRDA analysis performed on the entire dataset (BIO2, BIO4, BIO10, BIO15, and BIO18) were used. Geographical coordinates were transformed calculating principal coordinates of neighbor matrices using the function pcnm. Principal Coordinate Analysis (PCoA) was performed using the pcoa function from the R library ape v 5.6–2 . The function envfit (vegan) was then used to calculate ACBR and bioclimatic variable multiple regressions with the PCoA ordination axes. Only bioclimatic regions identified as significant by the function anova.cca were used in the envfit analysis. Correlations between the genus-level taxonomic dataset and geographical distance (distance-decay) and bioclimatic dataset were assessed by calculating distance matrices using the vegan function vegdist for the community and bioclimatic datasets (Bray–Curtis and Euclidean distances, respectively) and distm for the geographical coordinates using the R library geosphere v 1.5–18 . Mantel tests were then calculated using Spearman’s correlation and 1000 permutations. Dendrograms were created by calculating distance matrices as previously described and then by running the functions hclust and as.dendogram in the R package dendextend v 1.16.0 . Dendrograms were combined with the function tanglegram. Detailed taxonomic analyses were performed only on taxonomically consistent datasets derived from the amplification of the V3–V4 and V4 regions of the 16S rRNA gene (503 samples) . This dataset subset represented 7 out of the 10 analyzed ACBRs (plus AUI) and contained samples from the Antarctic Peninsula, the Transantarctic Mountains, and the Victoria Land. No samples from Eastern Antarctica or Dronning Maud Land were included. The dominant community was defined as including all genera with a relative abundance of > 1% in at least one sample that was present in at least 10% of samples. The 10% threshold was chosen to include genera present exclusively on the Antarctic Peninsula. Random forest analysis was performed using the R package randomForest v 4.7 on the genus-level dominant community dataset, which was transformed to relative abundance. Genera with “explained variance” higher than 30% using the randomForest algorithm were then used for Spearman’s semi-partial correlation with bioclimatic variables using the R packages ppcor v 1.1 and rfPermute v 2.5.1 . Correlations with p < 0.01 were considered significant. Only bioclimatic variables highlighted in dbRDA were used in these analyses to exclude collinear variables and therefore avoid redundancy (SWE, BIO2, BIO4, BIO10, BIO5, and BIO18). Genus relative abundances in each ACBR and the AUI were obtained by summing the reads of samples from the same ACBRs and then calculating relative abundances for each of them. Linear discriminant analysis effect size (LEfSe) analysis based on Kruskal–Wallis tests ( p < 0.01) was performed on this data subset using the R package microeco v 1.1.0 . Holm correction was used to adjust p values . Figures were plotted using the R packages ggplot2 v 3.3.5 , gplots v 3.1.3 , gridExtra v 2.3 , and Cairo v 1.5.12.2 . Data manipulation was carried out using default R packages and usedist v 0.4.0.9000 . The R scripts used for the analysis of the sequencing data can be found on the GitHub page https://github.com/gvMicroarctic/AntarcticBiogeographyPaper .
Alpha diversity and unclassified reads across ACBRs A total of 52 bacterial phyla were identified from the Antarctic soils (Table S ). The most abundant phylum was Actinomycetota, accounting for 30.6% of reads, followed by Bacteroidota (13.9%), Pseudomonadota (13.5%), Chloroflexota (9.9%), Acidobacteriota (8.5%), Cyanobacteriota (5.1%), Verrucomicrobiota (3.4%), Gemmatimonadota (3.0%), Bacillota (2.7%), and Deinococcota (1.3%). One percent of reads were unclassified at the phylum level. The total number of genera included in the dataset was 1445, ranging from 3 to 278 across individual samples (Table S ). The highest numbers of genera were recorded from the AUI and in ACBRs 1, 3, 12, and 16, while ACBRs 4 and 6–10 had the lowest numbers of genera (Fig. A). The numbers of genera differed significantly between ACBRs (ANOVA F = 56.84, df = 10, p < 2e − 16). High numbers of Tukey’s pairwise comparisons (from 7 to 9) were significant ( p < 0.01) for ACBRs 3, 12, and 16, while ACBRs 1, 4, and 8 showed the lowest numbers of significant pairwise differences (from 1 to 4) (Fig. B). Similar trends were observed for Shannon diversity (ANOVA F = 66.71, df = 10, p < 2e − 16), although fewer significant pairwise correlations were obtained, suggesting less divergent Shannon diversity across ACBRs (Figure S ). Because alpha diversity indices were calculated at the genus level, low richness and Shannon diversity values for some ACBRs could result from samples with high “unknown” counts (Table S ). The percentage of reads assigned to unknown genera varied across samples, ranging from 0.0 to 84.3% (Fig. C). This potential bias was confirmed by a significant Pearson’s correlation between the total number of genera and the percentage of unknown genera ( r = − 0.1624, p = 2.858e − 07). Alpha diversity showed significant correlations with bioclimatic variables, with significant positive correlations for BIO1, BIO5, BIO10, BIO12, BIO14, BIO17, BIO18, and SWE, and negative correlations for BIO2, BIO4, BIO15, and elevation, for both Shannon diversity and richness (Figures S –S ). Distance to the ocean was significantly correlated only with Shannon diversity (Figures S –S ). ACBR clustering Soil bacterial communities from AUI and ACBRs 3 and 12 clustered together in the PCoA plot (Fig. A). Samples from ACBR 8 showed a distinct clustering, suggesting consistent bacterial community structures within that ACBR (Fig. A). Samples from all other ACBRs did not show clear clustering according to region. ACBRs 1, 3, and 4 differed geographically from other mainland ACBRs, as their samples were mainly represented by soils collected from islands close to the Antarctic Peninsula. In terms of soil bacterial community structure, ACBR 3 samples were more closely related to those from sub- and peri-Antarctic islands (i.e., part of AUI; Fig. B). Despite the high level of overlap between soil bacterial communities, different ACBRs could be significantly distinguished using a Bray–Curtis dissimilarity matrix (PERMANOVA R 2 = 0.19, p < 0.01). Pairwise comparisons showed that soil bacterial communities from ACBRs 3 and 8 were the most distinct (Fig. C). AUI showed the lowest degree of dissimilarity to ACBRs 3, 7, and 9, all of which themselves included soil samples collected from islands. dbRDA only explained 12.0% (adjusted R 2 ) of the observed variance in community structure across the dataset (Fig. ). Bioclimatic variables related to temperature (BIO10, mean daily mean air temperatures of the warmest quarter) and precipitation (BIO18, mean monthly precipitation amount of the warmest quarter) were drivers for the bacterial community composition in ACBRs 1, 3, 12, and AUI; most of these samples were obtained from Antarctic, sub-Antarctic, and peri-Antarctic islands. Bioclimatic variables related to precipitation and temperature seasonality and daily ranges (BIO2, BIO4, and BIO5) were drivers for the bacterial communities in ACBRs 9 and 10; most samples from these ACBRs were sampled from mainland Antarctica. Sample distributions partially followed bioclimatic conditions in the different ACBRs. ACBRs 1, 3, 4, 7, and 12 have higher mean daily air temperatures in the warmest quarter (BIO10) and mean monthly precipitation amounts in the warmest quarter (BIO18) compared to ACBRs 6, 8–10, and 16 (Figure S ; Table S ). Mean diurnal air temperature range, temperature, and precipitation seasonality (BIO2, BIO4, and BIO15, respectively) were lowest in ACBRs 1, 3, and 12 compared to the other ACBRs (Figure S ). All bioclimatic variables showed statistically significant differences ( p < 0.01) between ACBRs; this was also the case for elevation, “distance from the ocean” and “distance from the coast” (Table S ). ACBR 4 showed mixed characteristics with high temperatures but also high ranges and seasonality values (Figure S ); this was reflected in the dbRDA, in which ACBR 4 samples showed a wide clustering in the plot (Fig. ). Distance-decay analyses showed that geographic distance between samples exhibited a low, but significant, correlation ( r = 0.1499, p = 0.0009) with bacterial composition dissimilarity scores between samples (Figure S A). Variation partitioning also showed that 13.2% of bacterial community variation was explained by geographical distance between samples alone, while only 2.7% of the variation was explained by the bioclimatic data, with 6.6% of variance being explained by their interaction (Figure S A) at a significance threshold of p < 0.01 (Table S ). Bacterial community structure explained by ACBRs The most consistent clustering of samples was detected for ACBR 3 (Fig. A and S ), where samples largely clustered together for both abiotic variables and soil bacterial community structures (Fig. A). Whereas samples collected from islands showed similar bacterial communities for ACBRs 3, 12, and AUI, other island samples clustered more closely to soils taken from the same ACBR compared to samples collected from other islands, as in the cases of ACBR 7 and 16 (Fig. A). Samples from other ACBRs followed different trends (Fig. B–G). For example, we observed a consistent clustering of samples from ACBR 6 according to bacterial composition and geographical trends, whereas the bioclimatic data did not show a detectable clustering (Fig. B). By comparison, relatively consistent geography was shown for samples from ACBRs 8–10 and 16, but samples were characterized by variable bioclimatic data, which was reflected by a wider bacterial community distribution (Fig. D–G). Samples from ACBR 7 were consistent both in geography and bioclimatic data but the bacterial community formed separate clusters (Fig. C). These clusters partially corresponded to samples from the Windmill Islands and Vestfold Hills regions (Figure S ) and PERMANOVA performed using this grouping generated a significant outcome ( p < 0.01) with R 2 = 0.14. Samples from the Vestfold Hills clustered more closely to those collected from ACBRs 9 and 10, whereas samples from the Windmill Islands clustered more closely to those from ACBR 3 and AUI. AUI and ACBRs 8 and 10 showed higher variance explained by bioclimatic data, while ACBRs 3, 6, 7, 9 and 11 showed higher variance explained by geography (Figure S B–I). Variation partitioning performed on AUI and ACBR 8 showed a lower residual variance not explained by bioclimatic variables and geography compared to the other ACBRs (0.42 and 0.362, respectively). The ACBRs with the highest residual component of variance were ACBRs 9 and 16 (0.816 and 0.879, respectively). The other ACBRs showed residual variance between 0.662 and 0.764. Islands vs mainland PERMANOVA performed on the dissimilarity matrices of soil bacterial community structures of island samples vs mainland samples was significant, although significance scores were lower than for ACBRs ( p = 0.0009, R 2 = 0.0361). This indicated that only a low percentage of observed variance was explained by whether a sample was collected from the Antarctic mainland or associated islands and is consistent with our previous observations that island soil microbiomes were more similar to geographically less distant soils than to soils collected from other islands (Fig. B and A). We also identified a division between islands from ACBRs 1, 3, and 4 (islands from the Antarctic Peninsula) and all the other islands (Fig. A). Islands that were not grouped into the ACBR classification (i.e., the AUI) also showed distinct clustering: soil bacterial communities from the islands of South Georgia, Bouvet, and Marion islands all clustered with those from Antarctic Peninsula islands, whereas Peter Øya, Scott, and Bartolomé islands grouped closer to islands from ACBRs 7, 9, 12, and 16 (except for two samples collected from Possession and Kerguelen islands, which clustered close to Antarctic Peninsula samples) (Fig. A). Samples from the Antarctic Peninsula islands (ACBRs 1, 3, and 4) correlated significantly ( p < 0.01) with all temperature- and precipitation-related bioclimatic variables (BIO10, BIO18, and SWE), while the other islands correlated with seasonal or diurnal differences in precipitation or temperature (BIO2, BIO4, and BIO15) (Fig. B and S A). It is appropriate to note that ACBR 3 contained the highest number of island samples in our dataset, and these showed a strong clustering of bacterial composition according to the island from which the samples were collected (Fig. C). This consistent clustering by island was confirmed by PERMANOVA with just the island dataset ( p = 0.0009, R 2 = 0.53) when using the island as the explanatory factor for differences in bacterial composition between samples. By comparison, the R 2 value was 0.04 when ACBR was used as an explanatory factor for the island dataset. Grouping mainland samples by factor “ACBR” also resulted in a weak ( R 2 = 0.025), albeit statistically significant correlation ( p < 0.01). Stronger correlations were identified when Mantel tests were used to compare bacterial community dissimilarities and geography (distance-decay, with r = 0.1611, p = 0.0009), and bacterial community dissimilarities and environmental variables ( r = 0.1102, p = 0.0009) (Figure S E–F). By comparison, Mantel tests for differences between bacterial community dissimilarities and geography for the island dataset yielded r = 0.37776 ( p = 0.0009), and r = 0.3001 ( p = 0.0009) for bacterial community dissimilarities and environmental variables (Figure S C–D). When Mantel tests were used to compare mainland bacterial community and elevation, distance from the coast, or distance from the ocean, only elevation correlated significantly ( r = 0.0780, p = 0.0009) (Figure S G–I). The separate PCoA clustering observed for ACBR 8 might be due to negative correlations with elevation and distances from the coast/ocean of ACBR 8 samples (Fig. D–E and S B). When only the island dataset was analyzed, 19.1% of the variance was explained by geography, 9.2% by bioclimatic data, 4.6% by their interaction, and 67.1% was residual (Figure S J). When only mainland samples were analyzed, variation partitioning analyses showed 78.9% residual variance, with geographic distances explaining the most variance in bacterial community structure between samples (geography = 9.8%, bioclimatic data = 4.0%, elevation/distance from coast/ocean = 1.3%) (Figure S K). Climatic drivers of the dominant soil bacteria and indicator taxa across Antarctica In order to understand how climatic variables impacted the more abundant bacterial taxa in Antarctic soil communities, we performed a correlation analysis between the bioclimatic variables and the relative abundance of the dominant genera across the sample sets. A total of 149 genera were identified as dominant (i.e., present in more than 10% of samples and with a relative abundance higher than 1% in more than one sample). Of these, 51 genera showed habitat preferences linked to the tested bioclimatic variables based on random forest predictions (Figure S ). Semi-partial correlations performed on the 51 genera and the six bioclimatic variables showed that most of the significant correlations were linked to mean daily mean air temperatures of the warmest quarter (BIO10, 35 correlations), followed by precipitation seasonality (BIO15, 20 correlations), temperature seasonality (BIO4, eight correlations) and mean diurnal air temperature range (BIO2, three correlations). No significant correlations ( p ≥ 0.01) were found for snow water equivalent (SWE) and mean monthly precipitation amount of the warmest quarter (BIO18) (Fig. A). All of the dominant genera presented in Fig. showed habitat preferences correlated with BIO10, except for Crossiella, Gaiella , Thermobacum , Chthoniobacter , Bryobacter , Iamia , and Candidatus Udaeobacter, which showed negative correlations with BIO15. All correlations between genera and BIO10 were positive, except for Conexibacter which showed a negative correlation with BIO10. Thirteen of the genera that significantly correlated with BIO10 also correlated with BIO15. Of the eight genera correlating with BIO4, seven showed a negative correlation ( Persicitalea, Algoriphagus, Simplicispira, Rhodoferax, Polaromonas, Dokdonella, and Thermomonas ) and one showed a positive correlation ( Crossiella ). Three genera showed positive correlations with BIO2 ( Persicitalea, Qipengyuania, and Gaiella ) (Fig. A). All of the genera identified in the random forest analysis with at least one positive or negative correlation to the tested bioclimatic variables were also identified as indicator taxa for the different ACBRs using LEfSe analysis. The highest numbers of indicator taxa which also correlated to bioclimatic variables were in ACBR 4 (12 indicators), ACBR 12 (10), ACBR 3 (9), AUI (4), and then ACBRs 1, 8, 9, and 10 reporting only between 1 and 3 indicator taxa (Fig. B). The highest number of bacterial indicators was therefore ascribed to maritime Antarctica, corresponding to the areas where the highest diversity of genera was observed (Fig. ). The most abundant genera that showed correlations to climatic parameters and were also identified as indicator taxa were Conexibacter (0.1–8.7%), Gaiella (0.0–4.2%), Nocardioides (0.5–3.8%), Candidatus Udaeobacter (0.0–3.2%), and Dokdonella (0.0–3.2%) (Fig. C; Table S ). The most abundant genera that were not selected by random forest modeling (variance explained < 30%) were Nostoc (0.0–7.0% in the different ACBRs), Tychonema (0.0–6.3%), Blastocatella (0.8–5.6%), Pedobacter (0.6–5.1%), and Sphingomonas (0.8–3.4%) (Figure S ).
A total of 52 bacterial phyla were identified from the Antarctic soils (Table S ). The most abundant phylum was Actinomycetota, accounting for 30.6% of reads, followed by Bacteroidota (13.9%), Pseudomonadota (13.5%), Chloroflexota (9.9%), Acidobacteriota (8.5%), Cyanobacteriota (5.1%), Verrucomicrobiota (3.4%), Gemmatimonadota (3.0%), Bacillota (2.7%), and Deinococcota (1.3%). One percent of reads were unclassified at the phylum level. The total number of genera included in the dataset was 1445, ranging from 3 to 278 across individual samples (Table S ). The highest numbers of genera were recorded from the AUI and in ACBRs 1, 3, 12, and 16, while ACBRs 4 and 6–10 had the lowest numbers of genera (Fig. A). The numbers of genera differed significantly between ACBRs (ANOVA F = 56.84, df = 10, p < 2e − 16). High numbers of Tukey’s pairwise comparisons (from 7 to 9) were significant ( p < 0.01) for ACBRs 3, 12, and 16, while ACBRs 1, 4, and 8 showed the lowest numbers of significant pairwise differences (from 1 to 4) (Fig. B). Similar trends were observed for Shannon diversity (ANOVA F = 66.71, df = 10, p < 2e − 16), although fewer significant pairwise correlations were obtained, suggesting less divergent Shannon diversity across ACBRs (Figure S ). Because alpha diversity indices were calculated at the genus level, low richness and Shannon diversity values for some ACBRs could result from samples with high “unknown” counts (Table S ). The percentage of reads assigned to unknown genera varied across samples, ranging from 0.0 to 84.3% (Fig. C). This potential bias was confirmed by a significant Pearson’s correlation between the total number of genera and the percentage of unknown genera ( r = − 0.1624, p = 2.858e − 07). Alpha diversity showed significant correlations with bioclimatic variables, with significant positive correlations for BIO1, BIO5, BIO10, BIO12, BIO14, BIO17, BIO18, and SWE, and negative correlations for BIO2, BIO4, BIO15, and elevation, for both Shannon diversity and richness (Figures S –S ). Distance to the ocean was significantly correlated only with Shannon diversity (Figures S –S ).
Soil bacterial communities from AUI and ACBRs 3 and 12 clustered together in the PCoA plot (Fig. A). Samples from ACBR 8 showed a distinct clustering, suggesting consistent bacterial community structures within that ACBR (Fig. A). Samples from all other ACBRs did not show clear clustering according to region. ACBRs 1, 3, and 4 differed geographically from other mainland ACBRs, as their samples were mainly represented by soils collected from islands close to the Antarctic Peninsula. In terms of soil bacterial community structure, ACBR 3 samples were more closely related to those from sub- and peri-Antarctic islands (i.e., part of AUI; Fig. B). Despite the high level of overlap between soil bacterial communities, different ACBRs could be significantly distinguished using a Bray–Curtis dissimilarity matrix (PERMANOVA R 2 = 0.19, p < 0.01). Pairwise comparisons showed that soil bacterial communities from ACBRs 3 and 8 were the most distinct (Fig. C). AUI showed the lowest degree of dissimilarity to ACBRs 3, 7, and 9, all of which themselves included soil samples collected from islands. dbRDA only explained 12.0% (adjusted R 2 ) of the observed variance in community structure across the dataset (Fig. ). Bioclimatic variables related to temperature (BIO10, mean daily mean air temperatures of the warmest quarter) and precipitation (BIO18, mean monthly precipitation amount of the warmest quarter) were drivers for the bacterial community composition in ACBRs 1, 3, 12, and AUI; most of these samples were obtained from Antarctic, sub-Antarctic, and peri-Antarctic islands. Bioclimatic variables related to precipitation and temperature seasonality and daily ranges (BIO2, BIO4, and BIO5) were drivers for the bacterial communities in ACBRs 9 and 10; most samples from these ACBRs were sampled from mainland Antarctica. Sample distributions partially followed bioclimatic conditions in the different ACBRs. ACBRs 1, 3, 4, 7, and 12 have higher mean daily air temperatures in the warmest quarter (BIO10) and mean monthly precipitation amounts in the warmest quarter (BIO18) compared to ACBRs 6, 8–10, and 16 (Figure S ; Table S ). Mean diurnal air temperature range, temperature, and precipitation seasonality (BIO2, BIO4, and BIO15, respectively) were lowest in ACBRs 1, 3, and 12 compared to the other ACBRs (Figure S ). All bioclimatic variables showed statistically significant differences ( p < 0.01) between ACBRs; this was also the case for elevation, “distance from the ocean” and “distance from the coast” (Table S ). ACBR 4 showed mixed characteristics with high temperatures but also high ranges and seasonality values (Figure S ); this was reflected in the dbRDA, in which ACBR 4 samples showed a wide clustering in the plot (Fig. ). Distance-decay analyses showed that geographic distance between samples exhibited a low, but significant, correlation ( r = 0.1499, p = 0.0009) with bacterial composition dissimilarity scores between samples (Figure S A). Variation partitioning also showed that 13.2% of bacterial community variation was explained by geographical distance between samples alone, while only 2.7% of the variation was explained by the bioclimatic data, with 6.6% of variance being explained by their interaction (Figure S A) at a significance threshold of p < 0.01 (Table S ).
The most consistent clustering of samples was detected for ACBR 3 (Fig. A and S ), where samples largely clustered together for both abiotic variables and soil bacterial community structures (Fig. A). Whereas samples collected from islands showed similar bacterial communities for ACBRs 3, 12, and AUI, other island samples clustered more closely to soils taken from the same ACBR compared to samples collected from other islands, as in the cases of ACBR 7 and 16 (Fig. A). Samples from other ACBRs followed different trends (Fig. B–G). For example, we observed a consistent clustering of samples from ACBR 6 according to bacterial composition and geographical trends, whereas the bioclimatic data did not show a detectable clustering (Fig. B). By comparison, relatively consistent geography was shown for samples from ACBRs 8–10 and 16, but samples were characterized by variable bioclimatic data, which was reflected by a wider bacterial community distribution (Fig. D–G). Samples from ACBR 7 were consistent both in geography and bioclimatic data but the bacterial community formed separate clusters (Fig. C). These clusters partially corresponded to samples from the Windmill Islands and Vestfold Hills regions (Figure S ) and PERMANOVA performed using this grouping generated a significant outcome ( p < 0.01) with R 2 = 0.14. Samples from the Vestfold Hills clustered more closely to those collected from ACBRs 9 and 10, whereas samples from the Windmill Islands clustered more closely to those from ACBR 3 and AUI. AUI and ACBRs 8 and 10 showed higher variance explained by bioclimatic data, while ACBRs 3, 6, 7, 9 and 11 showed higher variance explained by geography (Figure S B–I). Variation partitioning performed on AUI and ACBR 8 showed a lower residual variance not explained by bioclimatic variables and geography compared to the other ACBRs (0.42 and 0.362, respectively). The ACBRs with the highest residual component of variance were ACBRs 9 and 16 (0.816 and 0.879, respectively). The other ACBRs showed residual variance between 0.662 and 0.764.
PERMANOVA performed on the dissimilarity matrices of soil bacterial community structures of island samples vs mainland samples was significant, although significance scores were lower than for ACBRs ( p = 0.0009, R 2 = 0.0361). This indicated that only a low percentage of observed variance was explained by whether a sample was collected from the Antarctic mainland or associated islands and is consistent with our previous observations that island soil microbiomes were more similar to geographically less distant soils than to soils collected from other islands (Fig. B and A). We also identified a division between islands from ACBRs 1, 3, and 4 (islands from the Antarctic Peninsula) and all the other islands (Fig. A). Islands that were not grouped into the ACBR classification (i.e., the AUI) also showed distinct clustering: soil bacterial communities from the islands of South Georgia, Bouvet, and Marion islands all clustered with those from Antarctic Peninsula islands, whereas Peter Øya, Scott, and Bartolomé islands grouped closer to islands from ACBRs 7, 9, 12, and 16 (except for two samples collected from Possession and Kerguelen islands, which clustered close to Antarctic Peninsula samples) (Fig. A). Samples from the Antarctic Peninsula islands (ACBRs 1, 3, and 4) correlated significantly ( p < 0.01) with all temperature- and precipitation-related bioclimatic variables (BIO10, BIO18, and SWE), while the other islands correlated with seasonal or diurnal differences in precipitation or temperature (BIO2, BIO4, and BIO15) (Fig. B and S A). It is appropriate to note that ACBR 3 contained the highest number of island samples in our dataset, and these showed a strong clustering of bacterial composition according to the island from which the samples were collected (Fig. C). This consistent clustering by island was confirmed by PERMANOVA with just the island dataset ( p = 0.0009, R 2 = 0.53) when using the island as the explanatory factor for differences in bacterial composition between samples. By comparison, the R 2 value was 0.04 when ACBR was used as an explanatory factor for the island dataset. Grouping mainland samples by factor “ACBR” also resulted in a weak ( R 2 = 0.025), albeit statistically significant correlation ( p < 0.01). Stronger correlations were identified when Mantel tests were used to compare bacterial community dissimilarities and geography (distance-decay, with r = 0.1611, p = 0.0009), and bacterial community dissimilarities and environmental variables ( r = 0.1102, p = 0.0009) (Figure S E–F). By comparison, Mantel tests for differences between bacterial community dissimilarities and geography for the island dataset yielded r = 0.37776 ( p = 0.0009), and r = 0.3001 ( p = 0.0009) for bacterial community dissimilarities and environmental variables (Figure S C–D). When Mantel tests were used to compare mainland bacterial community and elevation, distance from the coast, or distance from the ocean, only elevation correlated significantly ( r = 0.0780, p = 0.0009) (Figure S G–I). The separate PCoA clustering observed for ACBR 8 might be due to negative correlations with elevation and distances from the coast/ocean of ACBR 8 samples (Fig. D–E and S B). When only the island dataset was analyzed, 19.1% of the variance was explained by geography, 9.2% by bioclimatic data, 4.6% by their interaction, and 67.1% was residual (Figure S J). When only mainland samples were analyzed, variation partitioning analyses showed 78.9% residual variance, with geographic distances explaining the most variance in bacterial community structure between samples (geography = 9.8%, bioclimatic data = 4.0%, elevation/distance from coast/ocean = 1.3%) (Figure S K).
In order to understand how climatic variables impacted the more abundant bacterial taxa in Antarctic soil communities, we performed a correlation analysis between the bioclimatic variables and the relative abundance of the dominant genera across the sample sets. A total of 149 genera were identified as dominant (i.e., present in more than 10% of samples and with a relative abundance higher than 1% in more than one sample). Of these, 51 genera showed habitat preferences linked to the tested bioclimatic variables based on random forest predictions (Figure S ). Semi-partial correlations performed on the 51 genera and the six bioclimatic variables showed that most of the significant correlations were linked to mean daily mean air temperatures of the warmest quarter (BIO10, 35 correlations), followed by precipitation seasonality (BIO15, 20 correlations), temperature seasonality (BIO4, eight correlations) and mean diurnal air temperature range (BIO2, three correlations). No significant correlations ( p ≥ 0.01) were found for snow water equivalent (SWE) and mean monthly precipitation amount of the warmest quarter (BIO18) (Fig. A). All of the dominant genera presented in Fig. showed habitat preferences correlated with BIO10, except for Crossiella, Gaiella , Thermobacum , Chthoniobacter , Bryobacter , Iamia , and Candidatus Udaeobacter, which showed negative correlations with BIO15. All correlations between genera and BIO10 were positive, except for Conexibacter which showed a negative correlation with BIO10. Thirteen of the genera that significantly correlated with BIO10 also correlated with BIO15. Of the eight genera correlating with BIO4, seven showed a negative correlation ( Persicitalea, Algoriphagus, Simplicispira, Rhodoferax, Polaromonas, Dokdonella, and Thermomonas ) and one showed a positive correlation ( Crossiella ). Three genera showed positive correlations with BIO2 ( Persicitalea, Qipengyuania, and Gaiella ) (Fig. A). All of the genera identified in the random forest analysis with at least one positive or negative correlation to the tested bioclimatic variables were also identified as indicator taxa for the different ACBRs using LEfSe analysis. The highest numbers of indicator taxa which also correlated to bioclimatic variables were in ACBR 4 (12 indicators), ACBR 12 (10), ACBR 3 (9), AUI (4), and then ACBRs 1, 8, 9, and 10 reporting only between 1 and 3 indicator taxa (Fig. B). The highest number of bacterial indicators was therefore ascribed to maritime Antarctica, corresponding to the areas where the highest diversity of genera was observed (Fig. ). The most abundant genera that showed correlations to climatic parameters and were also identified as indicator taxa were Conexibacter (0.1–8.7%), Gaiella (0.0–4.2%), Nocardioides (0.5–3.8%), Candidatus Udaeobacter (0.0–3.2%), and Dokdonella (0.0–3.2%) (Fig. C; Table S ). The most abundant genera that were not selected by random forest modeling (variance explained < 30%) were Nostoc (0.0–7.0% in the different ACBRs), Tychonema (0.0–6.3%), Blastocatella (0.8–5.6%), Pedobacter (0.6–5.1%), and Sphingomonas (0.8–3.4%) (Figure S ).
Our data reveal that bacterial communities in Antarctic soils do not align closely with the ACBR characterization proposed by Terauds et al. and Terauds and Lee . Only 19% of the variability was explained by ACBR as a discriminating factor, with high unexplained residual variability. This percentage further decreased (to 2.5%) when Antarctic-associated island soil samples were removed from the dataset (although this also had the confounding effect of removing most samples obtained from maritime Antarctica), supporting the conclusion that Antarctic mainland soil microbial community structures poorly reflect the ACBR classification, which itself is largely based on diversity patterns of macroscopic organisms. The most significant clustering in our dataset was between samples collected from either the continental Antarctica or the maritime Antarctica (the Antarctic Peninsula and the large off-shore archipelagoes of the South Shetland Islands and South Orkney Islands). Within these two areas, microbial communities showed high structural homogeneity, therefore not closely reflecting the ACBR classification . Bacterial communities from North Victoria Land (ACBR8) represented an exception, as they showed the highest microbial compositional differences compared to all the other ACBRs in continental Antarctica, possibly due to the fact that ACBR 8 experiences higher temperatures and precipitation rates compared to the remainder of the Victoria Land region (Figure S ). The observed clustering was principally explained by bioclimatic conditions, where samples from AUI and ACBRs 1, 3, 4, and 12 showed higher precipitation and temperature values (BIO10 and BIO18), and lower daily and seasonal precipitation and temperature ranges (BIO2, BIO4 and BIO15). It is probable that soil bacterial communities in these ACBRs are subject to more favorable environmental conditions than those in ACBRs 6–10 and 16 . More challenging environmental conditions impose increased selection for resistant and/or resilient microbial communities . These observations are consistent with the observed reductions in Antarctic soil bacterial diversity at higher latitudes , which are largely explained by reductions in air temperature and water availability . This was also supported by the observed correlations between air temperature, and other bioclimatic variables, with bacterial diversity: soils subjected to higher air temperatures and precipitation showed higher bacterial diversity, whereas those sampled from environments with wider seasonal and daily changes in temperature and precipitation showed lower diversity (Figure S –S ). This suggests that bacterial diversity is higher under more favorable environmental conditions (higher water availability, higher temperatures, and more stable environmental conditions) and lower in more life-challenging conditions . As temperature and precipitation regimes influence water and nutrient bioavailability , microbiome functionality , and diversity in soil, they are logical explanatory factors (SWE, BIO2, BIO4, BIO10, BIO15, and BIO18) of observed Antarctic soil bacterial community patterns (Figs. and and Figure S ). Nevertheless, only a low percentage of variability overall was explained by bioclimatic variables (Figs. and , Figures S and S ). A high percentage of unexplained variance is observed both when ACBR classification and bioclimatic variables are taken into consideration as explanatory factors. This is not surprising as microbial communities on the Antarctic continent are distributed and structured by a multiplicity of environmental factors and dynamics. Distance-decay relationships between geographical distances and microbial diversity vary in relation to environmental connectivity , where the Antarctic continent includes both connected and fragmented habitats , and therefore presents a diversity of factors that shape microbial distributions in different Antarctic areas. Notably, it is currently thought that eolian microbial dispersal in continental Antarctica is limited and that microbial community distribution patterns in Antarctica are more influenced by the existence of suitable glacial refugia (habitable areas that persisted during glaciation cycles) and therefore by soil histories and climate legacies . A regional study conducted in the McMurdo Dry Valleys showed that soil microbial composition was impacted by geographical distance, in part due to variations of the geochemical variables across the studied region . In the Transantarctic Mountains, microbial diversity in soil ecosystems significantly varies with terrain age, illustrating that biotic communities may vary with both abiotic spatial heterogeneity and geological history . At continental scales, geographical distances can shape microbial distributions due to dispersal patterns, climatic characteristics, and continental formation history . The lack of available consistent geochemical data for our datasets probably also contributed to the observed high percentage of unexplained variance. It has frequently been observed that Antarctic soil microbial distribution varies in relation to edaphic characteristics such as pH, electrical conductivity, soil moisture, soil temperature, and nutrient content [ , , , , , ]. However, climate influences edaphic characteristics, therefore indirectly influencing soil microbial communities . Our results suggest that the geographically large ACBRs may be too broad to capture microbial biodiversity patterns as they encompass regions that have very different environmental conditions and, thus, soil bacterial communities. A pertinent example in the current study is that of ACBR 7, in which distinct clustering separating the Windmill Islands and the Vestfold Hills was observed. These regions represent very different environments , with the Vestfold Hills being more extreme and dominated by Actinobacteriota, and the near-coast Windmill Islands having unique microbial communities with some sites dominated by Candidatus Dormibacteraeota and Eremiobacteriota . Notably, the bacterial communities of the Vestfold Hills clustered more closely with samples collected from ACBRs 9 and 10, also characterized by extreme climatic conditions, compared to ACBR 3 and the sub- and peri-Antarctic islands which clustered with samples collected from the Windmill Islands region (Figures S and S ). We also identified a strong divergence in the drivers of soil bacterial community composition when considering specifically the Antarctic/sub-Antarctic islands and the continental mainland. For instance, bacterial communities clustered consistently within each island/island group examined, but did not show any clear geographical differentiation on the Antarctic mainland. This might be because the island and mainland soils have different formational histories . Island bacterial communities were generally distinct from each other, highlighting the strong effect of island isolation on bacterial community development . A previous study reported changes between inland and coastal soils in terms of fungal diversity due to differences in soil geochemistry and environmental conditions . Because microbial community distributional patterns are partially shaped by bioclimatic conditions, and because observations and model predictions highlight that maritime Antarctica is most strongly affected by global warming [ – ], microbial community distribution and diversity in this region may be highly impacted in the near future . Some of the sampled areas in our dataset are not included in the ACBR classification. These included remote oceanic islands included within the maritime Antarctic (Bouvetøya, Peter I Øya) and the continental Antarctic near-coast (Scott Island), and the sub- and peri-Antarctic islands (South Georgia and Marion, Possession, Kerguelen and Bartolomé islands). Soil bacterial communities from these islands clustered more closely with geographically associated islands/groups within the ACBR system, and soils from South Georgia Island, Marion Island, and Bouvetøya clustered closely with those from the maritime Antarctic ACBRs 1, 3, and 4 (Group 1), whereas soils from Scott Island, Peter I Øya, and Bartolomé Island clustered closer to the continental Antarctic islands (from ACBRs 7, 9, and 12) (Group 2). Group 2 might indicate the influence of the Pacific Ocean and currents around the entire continental Antarctic coastline. Furthermore, the Antarctic Coastal current flows south along the western Antarctic Peninsula (Group 1), potentially influencing soil microbiomes on the South Shetlands and maritime Antarctic coastline (also Group 1). A similar argument was previously advanced by Pugh and Convey (2008) , but more studies are needed to test this hypothesis. Lebre et al. (2023) have proposed a differentiation of the sub- and peri-Antarctic islands presented in this study into the “classical” maritime, continental, and sub-Antarctic regions, based on soil microbial community compositions . However, the addition of maritime Antarctic Islands from the Antarctic Peninsula region in the current study is not consistent with this proposed island clustering (c.f., ), further highlighting the importance of spatial scales when studying patterns of soil microbial ecology . Of the dominant genera, 28% showed habitat preferences connected to at least one of the bioclimatic variables, particularly for habitats with higher temperatures (BIO10) and precipitation seasonality (BIO15) (Fig. A). Some of the genera showing habitat preferences, such as Nitrospira and Thiobacillus , have important ecosystem functions. Members of Nitrospira and Thiobacillus are chemolithotrophic nitrite- and sulphur-oxidizers and probably play a key role in enriching these soils with bioavailable nitrate and sulfate [ – ]. Other organisms potentially contributing to inputs of bioavailable elements (e.g., organic carbon) to Antarctic soils are the photoautotrophic genus Rhodoferax . All of these genera showed preferences for warmer habitats and probably have higher metabolic efficiencies at higher temperatures . Only Conexibacter showed a preference for lower temperature habitats. This genus has, in fact, been reported as widely present and dominant in Antarctic soils [ – ]. On the contrary, known cold-active taxa such as Polaromonas and Aequorivita showed a preference for higher temperature habitats (Fig. ) . Acidophilic and acidotolerant genera ( Acidibacter and Bryobacter ) also showed preferences for habitats with warmer air temperatures: these genera were more abundant in vegetated sub-Antarctic island soils, which also typically exhibit lower pH values [ – ]. Taxa known to be extremely competitive in microbial communities, such as members of the genus Lysobacter, also showed higher temperature habitat preferences . We suggest that taxa that exhibit habitat preferences have undergone some degree of habitat filtering relating to the inherent climate conditions. It is therefore reasonable to hypothesize that these taxa are more likely to respond, either positively or negatively, to changes in climate than those taxa with more homogeneous distributions. Given that these taxa are dominant members of the community, it is likely that climate change-induced changes in their population numbers and/or function will impact the structure and function of the entire communities, and potentially also the soil geochemistry. Furthermore, all taxa that showed climatic preferences were also statistically identified as indicator genera of different ACBRs by LEfSe, showing that altered climatic conditions could also significantly affect bacterial communities, particularly the dominant taxa associated with different ACBRs. Our study reports a comprehensive dataset from published and de novo Antarctic soil bacterial community datasets, where the combination of datasets from different sources has provided a viable means of circumventing the limitations imposed by the remoteness and sampling challenges of the Antarctic region. Despite the use of different 16S rRNA hyper-variable regions in generating the separate datasets used in this study, we argue that this is a valid approach, supported by the recently published study of Varliero et al., in which a subset of Antarctic soil samples was amplified with multiple 16S primer sets . Analysis of the sequence data obtained indicated that the principal similarity/dissimilarity trends in bacterial community composition were effectively preserved, irrespective of the hyper-variable region amplified .
We conclude that Antarctic soil bacterial diversity patterns and community structure identified in this study do not conform closely to the current ACBR classification, which is based primarily on data representing eukaryotic organisms, often from limited taxonomic groups. This might be due to the fact that prokaryotic soil communities, even in a single sample, are hugely diverse, represent a wide range of physiologies and functions, and are likely to be influenced by different distributional and dispersal drivers compared to eukaryotes. Soil prokaryotes may also exhibit higher levels of resistance and resilience to environmental stressors than higher organisms, which dominate the ACBR categories [ , , ]. Furthermore, bacterial distributions are also based on soil heterogeneity and the presence of favorable microenvironments [ , , , ]. Our results further suggest that Antarctic bacterial communities, and particularly the identified indicator taxa, might be impacted by climatic and other environmental changes in the different ACBRs . The combined dataset used in this study represents a comprehensive baseline upon which future studies of Antarctic microbial ecology can be developed, and the outcomes of this study can also be used to bolster biodiversity conservation efforts on the continent and its associated islands. However, considering how soil geochemistry can vary at the micro-scale and how these microbial microenvironments and refugia are important in shaping microbial communities , future increased coverage of soil sampling design across Antarctica would allow for the development of a more robust ACBR-style classification including information derived from prokaryotic communities and also extending to islands that lie beyond the region of Antarctic Treaty governance. Therefore, further studies are clearly required and we emphasize the need for more extensive campaigns to systematically sample and better characterize Antarctic soil microbial communities. In particular, more even and representative mapping of Antarctic microbial distributions is required with, at present, some areas being relatively well studied (e.g., the Antarctic Peninsula and parts of Victoria Land) and others being heavily underrepresented [ , , ]. Similarly, extensive microbial diversity and community characterization should also be extended to other microbial habitats such as rocks, freshwater lakes, and ice sheet environments [ , – ].
Additional file 1: Figure S1. Analysed islands part of the Antarctic Conservation Biogeographic Regions (ACBRs). Geographic positions of islands included in the ACBR classification proposed by Terauds and Lee (2016) for ACBR1 and ACBR3 (A), ACBR 4 (B), ACBR 12 (C), ACBR 9 (D), ACBR 7 and ACBR16 (E-F), and ACBR 6 (G). Figure S2. Antarctic Conservation Biogeographic Regions (ACBR) unclassified islands (AUI). Sample locations are indicated by white dots. Figure S3. Bacterial Shannon diversity trends across ACBRs. Shannon diversity calculated using the genus dataset (A). Significant Tukey's statistical tests ( p < 0.01) for Shannon diversity calculated at the genus-level (B). White dots correspond to non-significant Tukey’s statistical tests ( p ≥ 0.01). ACBRs from 1 to 16 correspond to those described from Terauds and Lee (2016). ACBR 1: North-east Antarctic Peninsula; ACBR 3: North-west Antarctic Peninsula; ACBR 4: Central South Antarctic Peninsula; ACBR 6: Dronning Maud Land; ACBR 7: East Antarctica; ACBR 8: North Victoria Land; ACBR 9: South Victoria Land; ACBR 10: Transantarctic Mountains; ACBR 12: Marie Byrd Land; ACBR 16: Prince Charles Mountains. AUI: ACBR unclassified islands. Figure S4. Correlations between number of genera and bioclimatic variables. Pearson’s correlations between number of genera (i.e., richness) and BIO1 (A), BIO2 (B), BIO4 (C), BIO5 (D), BIO10 (E), BIO12 (F), BIO14 (G), BIO15 (H), BIO17 (I), BIO18 (J), SWE (K), distance to ocean (L) and elevation (M). BIO1: mean annual air temperature, °C; BIO2: mean diurnal air temperature range, °C; BIO4: temperature seasonality,°C/100; BIO5: mean daily maximum air temperature of the warmest month, °C; BIO10: mean daily mean air temperatures of the warmest quarter, °C; BIO12: annual precipitation, kg m -2 ; BIO14: precipitation in the driest month, kg m -2 ; BIO15: precipitation seasonality, %; BIO17: mean monthly precipitation in the driest quarter, kg m -2 ; BIO18: mean monthly precipitation in the warmest quarter, kg m -2 ; SWE: snow water equivalent, kg m -2 ; Distance to ocean: km; Elevation: m. Figure S5. Correlations between Shannon diversity and bioclimatic variables. Pearson’s correlations between Shannon diversity and BIO1 (A), BIO2 (B), BIO4 (C), BIO5 (D), BIO10 (E), BIO12 (F), BIO14 (G), BIO15 (H), BIO17 (I), BIO18 (J), SWE (K), distance to ocean (L) and elevation (M). BIO1: mean annual air temperature, °C; BIO2: mean diurnal air temperature range, °C; BIO4: temperature seasonality, °C/100; BIO5: mean daily maximum air temperature of the warmest month, °C; BIO10: mean daily mean air temperatures of the warmest quarter, °C; BIO12: annual precipitation, kg m -2 ; BIO14: precipitation in the driest month, kg m -2 ; BIO15: precipitation seasonality, %; BIO17: mean monthly precipitation in the driest quarter, kg m -2 ; BIO18: mean monthly precipitation in the warmest quarter, kg m -2 ; SWE: snow water equivalent, kg m -2 ; Distance to ocean: km; Elevation: m. Figure S6. Bioclimatic variables. Selected bioclimatic variables and characteristics for each ACBR and for ACBR unclassified islands (AUI): BIO1 (mean annual air temperature) (A), BIO2 (mean diurnal air temperature range) (B), BIO4 (temperature seasonality) (C), BIO10 (mean daily mean air temperatures of the warmest quarter) (D), BIO12 (annual precipitation amount) (E), BIO15 (precipitation seasonality) (F), BIO18 (mean monthly precipitation amount of the warmest quarter) (G), SWE (snow water equivalent) (H), elevation (I), distance from coast (J) and distance from ocean (K). All bioregions were reported except from H-J where AUI was excluded by data representation. ACBR 1: North-east Antarctic Peninsula; ACBR 3: North-west Antarctic Peninsula; ACBR 4: Central South Antarctic Peninsula; ACBR 6: Dronning Maud Land; ACBR 7: East Antarctica; ACBR 8: North Victoria Land; ACBR 9: South Victoria Land; ACBR 10: Transantarctic Mountains; ACBR 12: Marie Byrd Land; ACBR 16: Prince Charles Mountains. AUI: ACBR unclassified islands. Figure S7. Correlations between bacterial community composition and geographic distance, bioclimatic data, elevation, distance to coast and ocean. Relation between Bray-Curtis dissimilarity matrix performed on genus dataset and Euclidean distance matrix calculated for the entire dataset on geographic sample location (A) and on bioclimatic data (B), for the island dataset on geographic sample location (C) and on bioclimatic data (D), for the mainland dataset on geographic sample location (E), on bioclimatic data (F), on elevation (G), on distance to coast (H) and on distance to ocean (I). Bioclimatic data: BIO1, BIO2, BIO4, BIO5, BIO10, BIO12, BIO14, BIO15, BIO17, BIO18 and SWE associated to each sample. Figure S8. Variation partitioning performed for entire dataset and single ACBRs. Variation partitioning analyses performed on geography (distance) and bioclimatic variables for the entire dataset (A), AUI (B), ACBR 3 (C), ACBR 6 (D), ACBR 7 (E), ACBR 8 (F), ACBR 9 (G), ACBR 10 (H), ACBR 16 (I), only island samples (J), and only mainland samples (K). In addition to geography (distance) and bioclimatic variable, elevation and distances from coast and ocean were taken in consideration for variation partitioning performed only on mainland samples. ACBR 1: North-east Antarctic Peninsula; ACBR 3: North-west Antarctic Peninsula; ACBR 4: Central South Antarctic Peninsula; ACBR 6: Dronning Maud Land; ACBR 7: East Antarctica; ACBR 8: North Victoria Land; ACBR 9: South Victoria Land; ACBR 10: Transantarctic Mountains; ACBR 12: Marie Byrd Land; ACBR 16: Prince Charles Mountains. AUI: ACBR unclassified islands. Figure S9. Sample clustering at bioclimatic, bacterial community and geographic level. Tanglegram performed between dendrograms created using geography and bacterial community datasets (A) and bioclimatic and bacterial community datasets (B). Geography: geographical distances between samples in the form of latitude and longitude information; Bacterial community: Hellinger-transformed community at genus-level; Bioclimatic data: BIO1, BIO2, BIO4, BIO5, BIO10, BIO12, BIO14, BIO15, BIO17, BIO18 and SWE associated to each sample. ACBR 1: North-east Antarctic Peninsula; ACBR 3: North-west Antarctic Peninsula; ACBR 4: Central South Antarctic Peninsula; ACBR 6: Dronning Maud Land; ACBR 7: East Antarctica; ACBR 8: North Victoria Land; ACBR 9: South Victoria Land; ACBR 10: Transantarctic Mountains; ACBR 12: Marie Byrd Land; ACBR 16: Prince Charles Mountains. AUI: ACBR unclassified islands. Figure S10. ACBR 7 bacterial composition. PCoA where only samples collected from ACBR 7 were collected and are colored in blue if from Vestfold hill region, and in red if from Windmill island region. Figure S11. dbRDA performed on only island samples or mainland samples. Distance-based redundancy analysis (dbRDA) performed on Hellinger transformed genus dataset and standardized bioclimatic variable dataset for only island samples ( n = 142) (A) and only mainland samples ( n = 846) (B). BIO2: mean diurnal air temperature range; BIO4: temperature seasonality; BIO10: mean daily mean air temperatures of the warmest quarter; BIO15: precipitation seasonality; BIO18: mean monthly precipitation amount of the warmest quarter; SWE: Snow water equivalent. Figure S12. Predictors of the dominant community distribution across Antarctica. Mean decrease accuracy associated to each bioclimatic variable (A). Number of taxa associated to the best predictor for each taxon distribution (predictor with highest %lncMSE) related to random forest analysis (B). BIO2: mean diurnal air temperature range; BIO4: temperature seasonality; BIO10: mean daily mean air temperatures of the warmest quarter; BIO15: precipitation seasonality; BIO18: mean monthly precipitation amount of the warmest quarter; SWE: Snow water equivalent. Figure S13. Relative abundance of dominant genera that were not selected by random forest model (variance explained < 30%). Only samples sequenced with V3-V4 and V4 16S rRNA primers were used for this analysis to ensure the best taxonomic consistency between samples (Varliero et al., 2023). Dominant genera were defined as those with a relative abundance of > 1% in at least one sample that were present in at least 10% of samples. Correspondingly, this approach included samples from AUI and ACBRs 1, 3, 4, 8, 9, 10 and 12. BIO2: mean diurnal air temperature range; BIO4: temperature seasonality; BIO10: mean daily mean air temperatures of the warmest quarter; BIO15: precipitation seasonality; BIO18: mean monthly precipitation amount of the warmest quarter; SWE: Snow water equivalent. Additional file 2: Table S1. Specifics for all analysed datasets. The total number of samples was 1164, whereas the number of samples passing all the quality steps was 988. *ACBRs from 1 to 16 correspond to those described from Terauds and Lee (2016), "AUI" stands for “ACBR unclassified islands” and represents islands associated with the Antarctic mainland not included in the ACBR classification, and sub- and peri-Antarctic islands. **years correspond to Austral summers except from when specified otherwise. ***number of samples after a cutoff of 5000 reads per sample was applied. Table S2. Sample specifics. BIO1: mean annual air temperature, °C; BIO2: mean diurnal air temperature range, °C; BIO4: temperature seasonality, °C/100; BIO5: mean daily maximum air temperature of the warmest month, °C; BIO10: mean daily mean air temperatures of the warmest quarter, °C; BIO12: annual precipitation, kg m -2 ; BIO14: precipitation in the driest month, kg m -2 ; BIO15: precipitation seasonality, %; BIO17: mean monthly precipitation in the driest quarter, kg m -2 ; BIO18: mean monthly precipitation in the warmest quarter, kg m -2 ; SWE: snow water equivalent, kg m -2 ; Distance to coast: km; Distance to ocean: km; Elevation: m. ACBRs from 1 to 16 correspond to those described from Terauds and Lee (2016). ACBR 1: North-east Antarctic Peninsula; ACBR 3: North-west Antarctic Peninsula; ACBR 4: Central South Antarctic Peninsula; ACBR 6: Dronning Maud Land; ACBR 7: East Antarctica; ACBR 8: North Victoria Land; ACBR 9: South Victoria Land; ACBR 10: Transantarctic Mountains; ACBR 12: Marie Byrd Land; ACBR 16: Prince Charles Mountains. AUI: ACBR unclassified islands. Table S3. Paramenters used in the dada2 function filterAndTrim() in each dataset. All the other options were set to default except for truncQ which was set to 0. Table S4. Number of reads at each step of the 16S rRNA gene processing pipeline for all datasets. *counts reported as read pairs. Table S5. Taxonomic relative abundance at phylum- (A), class- (B), order- (C) and family-level (D). Table S6. Relative abundance and taxonomy associated to genus dataset. Table S7. Analyses of variance (ANOVA) performed on bioclimatic variables, elevation and sample distance from coast/ocean. BIO1 (mean annual air temperature, °C), BIO2 (mean diurnal air temperature range, °C), BIO4 (temperature seasonality,°C), BIO5 (mean daily maximum air temperature of the warmest month, °C), BIO10 (mean daily mean air temperatures of the warmest quarter, °C), BIO12 (annual precipitation amount, kg m -2 ), BIO14 (precipitation amount of the driest month, kg m -2 ), BIO15 (precipitation seasonality, kg m -2 ), BIO17 (mean monthly precipitation amount of the driest quarter, kg m -2 ), BIO18 (mean monthly precipitation amount of the warmest quarter, kg m -2 ) and SWE (snow water equivalent, kg m -2 ). Table S8. Statistics from dbRDA (A-B) and variation partitioning (C-G). A and B only performed on bioclimatc varaibles selected by interactive dbRDA selection. Statistics from function varpart() with X1 as bioclimac dataset and X2 as geography (C) and individual statistical tests using anova.cca(): geography without controlling for environmental variables (D), environmental variables without controlling geography (E, geography alone (F) and environmental variables alone (G). Table S9. Indicator taxa across ACBRs and AUI at genus-level (LEfSe analysis based on Kruskal–Wallis p < 0.01).
|
Vacuum Sealing Drainage for Primary Thoracolumbar Spondylodiscitis: A Technical Note | 78f3cf16-bb9d-4da4-b901-71de354e65a7 | 9381288 | Debridement[mh] | Spinal infection is a disease with some descriptions accompanying human evolution . There are several types of spinal infections, and when the infection affects only the intervertebral discs, the term used to describe the condition is usually discitis. If the infection invades the endplates of the vertebral body, the infection is more correctly designated as vertebral osteomyelitis or spondylitis. However, in many cases, at the time of diagnosis, the infection has damaged both structures; therefore, this condition is often diagnosed as spondylodiscitis . The literature reports that the incidence of spinal infection in the general population is 2.4/100000, and the incidence rate increases significantly with age, and the incidence of spinal infection in people over 50 years old increases to 6.5/100000, mainly due to reduced immunity . At the same time, comorbidities including diabetes, uremia, urinary tract infection, pulmonary infection, and body surface infection are also important reasons for the increased incidence . The conservative treatment of spinal infections is challenging. Although antibiotic therapy is crucial and necessary in treating spinal infections, acquiring pathogenic microorganisms in spinal infections is more challenging than other bone infections . As a result, most spinal infections are treated with antibiotics based on clinical experience alone . In addition, the inadequate blood supply to the disc tissue renders antibiotic therapy ineffective . Chandra et al. reported that conservative treatment of spinal infections with comorbidities is inefficient and requires surgical treatment, including neurological symptoms, lumbar instability, kyphosis, spinal abscess, infection involving more than 4 vertebrates, and infection involving the intervertebral disc . According to literature, conservative treatment is frequently ineffective for spinal infections, and approximately 50% of patients require surgery [ – ]. Presently, the surgical treatment of spinal infection consists mainly of the classic approach of lesion excision combined with internal fixation, which is more traumatic and cannot be tolerated by patients with a spinal infection due to their physical state. Since the 1990s, it has been universally accepted that vacuum sealing drainage (VSD), also known as Negative Pressure Wound Therapy (NPWT), provides therapeutic effects for soft tissue infections, bone infections of the extremities, and chronic refractory wounds. Nonetheless, the clinical impact of VSD on spontaneous spondylodiscitis has not been investigated. This study presents an all-new surgical paradigm for primary thoracolumbar infection patients with severe comorbidities. In this surgical paradigm, we, for the first time, apply VSD to treat primary spinal infection with intervertebral pus. The VSD treatment has proven feasible and effective for serious spondylodiscitis based on the short- and medium-term follow-up outcomes.
2.1. Patient Selection The inclusion criteria were as follows: clinically diagnosed with lumbar spine infection with no spinal cord injury, with severe comorbidities, bedridden for more than three months, and accepted VSD treatment. From June 30, 2018, to August 31, 2019, 6 patients were enrolled in this study, including 1 male and 5 females, aged 57.7 ± 7.83. All 6 patients had severe comorbidities and were incapacitated with bedridden for 5.5 ± 6.17 months, including one with renal abscess, one with cervical spondylotic myelopathy and incomplete paralysis, one with renal failure, two with renal failure with rheumatoid arthritis (stage 4), and one with fibula total nerve damage. All 6 patients were diagnosed with spinal infection and received antibiotics for 2 ± 0.67 months before admission. All 6 cases had low back pain symptoms at the consultation time, and 3 cases were accompanied by fever. As for the pathogenic microorganisms, 1 case of hospital blood culture was Escherichia coli, 1 case had a history of renal abscess due to Escherichia coli infection 4 months ago, and the remaining cases were unknown. summarizes the patient features. 2.2. Surgical Procedure Percutaneous pedicle screws and rods were placed on healthy vertebrae spanning the level of infection under C-arm guidance. Then, part of the lateral facet joint was excised, and the superior border of the lower pedicle was exposed through the intervertebral foramina approach through the working channel. Subsequently, the lumbar annulus was dissected, and the disc infection was removed entirely. Rinse repeatedly with hydrogen peroxide, bromine, and saline solution. The VSD sponge is placed in the intervertebral disc space. The wound was sealed with negative pressure ( ). About 7 days after placing the VSD sponge, the VSD sponge was removed entirely, and the intervertebral space was scraped with a spatula and a curette and repeatedly rinsed for debridement. The formation of granulation tissue on the wound surface was closely observed. Then, place a new VSD sponge in the intervertebral space. The VSD changes are performed weekly under general anaesthesia in the operating room or under local anaesthesia at the bedside. When fresh granulation grows on the intervertebral space endplate, it is time for bone grafting. Use a particular iliac bone extraction instrument with a minimally invasive incision, and take an appropriate amount of iliac bone according to the bone defect. After the VSD sponge was removed, the intervertebral space was scratched, and the iliac bone was implanted after irrigation. The wound was sutured and sterile bandaged. 2.3. Postoperative Management and Follow-Up All patients were administered intravenous susceptibility (based on the findings of drug susceptibility testing) or broad-spectrum antibiotics (cefuroxime sodium, 1.5 g, q8 h) until C-reactive protein and ESR readings returned to normal levels. Then, continue intravenous or oral antibiotics for 8 weeks . Postoperative computed tomography (CT) and C-reactive protein were performed to evaluate the spinal fusion and infection. Follow-up was conducted 12 months postoperatively. The normal value of C-reactive protein and the new bone formation confirmed by CT in the intervertebral space were evaluated as a clinical cure. The JOA scores were measured in all cases before and 3 months after surgery to evaluate the changes in the neurological status of patients before and after surgery.
The inclusion criteria were as follows: clinically diagnosed with lumbar spine infection with no spinal cord injury, with severe comorbidities, bedridden for more than three months, and accepted VSD treatment. From June 30, 2018, to August 31, 2019, 6 patients were enrolled in this study, including 1 male and 5 females, aged 57.7 ± 7.83. All 6 patients had severe comorbidities and were incapacitated with bedridden for 5.5 ± 6.17 months, including one with renal abscess, one with cervical spondylotic myelopathy and incomplete paralysis, one with renal failure, two with renal failure with rheumatoid arthritis (stage 4), and one with fibula total nerve damage. All 6 patients were diagnosed with spinal infection and received antibiotics for 2 ± 0.67 months before admission. All 6 cases had low back pain symptoms at the consultation time, and 3 cases were accompanied by fever. As for the pathogenic microorganisms, 1 case of hospital blood culture was Escherichia coli, 1 case had a history of renal abscess due to Escherichia coli infection 4 months ago, and the remaining cases were unknown. summarizes the patient features.
Percutaneous pedicle screws and rods were placed on healthy vertebrae spanning the level of infection under C-arm guidance. Then, part of the lateral facet joint was excised, and the superior border of the lower pedicle was exposed through the intervertebral foramina approach through the working channel. Subsequently, the lumbar annulus was dissected, and the disc infection was removed entirely. Rinse repeatedly with hydrogen peroxide, bromine, and saline solution. The VSD sponge is placed in the intervertebral disc space. The wound was sealed with negative pressure ( ). About 7 days after placing the VSD sponge, the VSD sponge was removed entirely, and the intervertebral space was scraped with a spatula and a curette and repeatedly rinsed for debridement. The formation of granulation tissue on the wound surface was closely observed. Then, place a new VSD sponge in the intervertebral space. The VSD changes are performed weekly under general anaesthesia in the operating room or under local anaesthesia at the bedside. When fresh granulation grows on the intervertebral space endplate, it is time for bone grafting. Use a particular iliac bone extraction instrument with a minimally invasive incision, and take an appropriate amount of iliac bone according to the bone defect. After the VSD sponge was removed, the intervertebral space was scratched, and the iliac bone was implanted after irrigation. The wound was sutured and sterile bandaged.
All patients were administered intravenous susceptibility (based on the findings of drug susceptibility testing) or broad-spectrum antibiotics (cefuroxime sodium, 1.5 g, q8 h) until C-reactive protein and ESR readings returned to normal levels. Then, continue intravenous or oral antibiotics for 8 weeks . Postoperative computed tomography (CT) and C-reactive protein were performed to evaluate the spinal fusion and infection. Follow-up was conducted 12 months postoperatively. The normal value of C-reactive protein and the new bone formation confirmed by CT in the intervertebral space were evaluated as a clinical cure. The JOA scores were measured in all cases before and 3 months after surgery to evaluate the changes in the neurological status of patients before and after surgery.
summarizes the surgical procedure and results. This series of patients includes 1 male and 5 females. The hospital stay was 44.3 ± 16.44 days. All 6 cases of thoracolumbar infection achieved clinical cure at 3-month follow-up, and no surgical-related mortalities occurred in our series. The total operation time was 283 ± 53 min, and the total blood loss was 240.8 ± 29.44 mL. All 6 patients completed the 12-month follow-up except for 1 patient who died of acute cerebral infarction 5 months after surgery due to bedridden and noncompliance with antithrombotic therapy after discharge from hospital. Among the 6 patients, 1 suffered anaphylactic shock from plasma infusion during hospitalisation and recovered after rescue; 1 suffered from stress gastritis and recovered after symptomatic treatment for 6 days. Furthermore, JOA scores improved significantly in all 6 patients at 3-month follow-up, demonstrating the effectiveness of this surgical paradigm ( ).
The challenge of spinal infection is that it is difficult to achieve complete debridement and adequate drainage of paravertebral abscesses with traditional surgery, requiring postoperative antibiotic treatment. However, it is difficult for antibiotics to reach sequestrum and bloodless tissues due to the inadequate blood supply. Notably, the intervertebral disc, a structure frequently linked with spinal infections, is supplied by endplate arterioles in adolescence but develops avascular in age . Low blood antibiotic levels in dead, pus-filled, and avascular tissues result in the formation of drug-resistant bacteria and bacterial biofilms, which are essential for the recurrence of postoperative infections . Literature indicates that the recurrence rate will be significantly reduced if all infected lesions are entirely eliminated . Evidence shows that complete debridement substantially reduces the rate of infection recurrence. However, due to the specificity of the spine structure, extensive debridement of extremity bone infections is contraindicated. Although VSD has been reported in the literature for other sites and spinal SSI infections, in this study, for the first time, we applied VSD to the treatment of primary spinal infections with intervertebral pus. By removing exudate, necrotic tissue, and bacteria using VSD, a microenvironment favorable to bacterial development is destroyed . Additionally, VSD increases the formation of granulation tissue, which is highly antimicrobial and good for wound healing . Simultaneously, VSD fills the postdebridement void and avoids hematoma development, facilitating autologous iliac bone grafting. Although spine surgeons have a general consensus about the surgical indications for spinal infections, not all infected patients are tolerant of conventional surgical treatments. First, elderly patients with comorbidities have a much higher incidence of severe postoperative complications than general patients . On the other hand, most patients with spinal infections require debridement, spinal internal fixation, and conventional surgery with autologous iliac bone grafting. The extended operation time and significant blood loss of this traditional surgery will significantly reduce systemic immunity, which is not conducive to postoperative recovery and infection control. This surgical approach usually results in higher complication and recurrence rates. We summarize the traditional surgical modalities reported in the literature in . The results showed that minimally invasive implantation of VSDs significantly reduces operative time, blood loss, and complications. Therefore, minimally invasive VSD is a feasible approach for patients with clear surgical indications but severe comorbidities who cannot afford surgery . Antibiotic treatment is required for all spinal infections. However, identifying pathogenic microorganisms is necessary to determine the most effective antibiotic. After repeated collection of blood, intraoperative tissue, abscess, and postoperative pathogen drainage, it was challenging to get pathogenic microorganisms from 4 patients in this series. These negative results may be caused by prehospital antibiotic administration or blood and specimen collection methods . It has been claimed that metagenomics is utilised to promptly and accurately discover harmful microbes, although this use must be proven in clinics. The benefits of VSD for patients with poor surgical tolerance are as follows. First, the intervertebral space installation of the VSD sponge is minimally invasive. This method is less invasive, causes less bleeding, and is suitable for patients with severe comorbidities . Second, this is a staged paradigm of precise individualised treatment, which provides a buffer opportunity for patients with severe comorbidities and determines whether further surgical treatment is required based on the treatment effect. Compared with the conventional surgery in , this new surgical paradigm significantly reduced the total blood loss and resulted in a faster postoperative recovery without increasing the total operative time [ , – ]. In conclusion, the VSD is safe and effectively treats spinal infections with severe comorbidities. Short- and medium-term follow-up demonstrated its efficacy. To our knowledge, this is the first time that VSD has been applied to treating the primary spinal infection with intervertebral pus. In long-term follow-up, complications and recurrence need to be further studied. Furthermore, this surgical paradigm requires further prospective controlled studies.
|
Metabolomics Combined with Transcriptomics Analysis Reveals the Regulation of Flavonoids in the Leaf Color Change of | 5e5495cc-5f3e-4f51-87b2-cb97e5eb1d4b | 11678339 | Biochemistry[mh] | Acer truncatum (Sapindaceae) is one of the most important deciduous maple trees and is often used for its ornamental autumn leaf color in China . More and more studies are focusing on the mechanism of leaf discoloration in which the leaf color of trees often changes from green to yellow and then to red in autumn . Some studies found that the chloroplast ultrastructure of Forsythia suspensa (thunb.) vahl had a looser structure than green leaves, and V-myb avian myeloblastosis viral oncogene homolog ( MYB ), basic helix–loop–helix ( bHLH), NAM, ATAF and CUC ( NAC ) were associated with the leaf pigment compounds of Fraxinus angustifolia vahl . Some studies on maple leaf color changes found that anthocyanin synthase ( ANS ) and bronze-1 ( BZ1 ) in A. mandshuricum Maxim resulted in the accumulation of cyanidin 3-O-glucoside, which causes a significant reddening of the leaf blade . However, little effort has been made to elucidate the mechanism underlying the variation in color between red, yellow and green leaves of A. truncatum . In general, the color variation of leaves is more complex than that of corollas . Green leaves are mainly the result of the dominance of chlorophyll content among all pigments . By contrast, the formation of yellow leaves is mainly due to the gradual degradation of chlorophyll, resulting in the dominance of carotenoids in the leaves . However, flavonoid biosynthesis is the determining factor for the change of leaf color to a non-yellow color . Flavones, flavonols, flavanones, flavanonols, isoflavones, catechins, anthocyanins and proanthocyanidins belong to subgroups of flavonoids . Anthocyanins are particularly the main reason for the red color of the leaves, and this pigment also helps the plant resist various biotic and abiotic stresses . In line with the previous analyses, there are two main functional functions of naturally occurring anthocyanins : one is the resistance to external stresses and the other is using different color strategies to complete the life cycle . Consequently, the accurate identification of flavonoids and even anthocyanin compounds in maple is very important for the effective use of forest resources. Plant metabolomics is the qualitative and quantitative study of small-molecule metabolites in plants that helps researchers better understand patterns of metabolite synthesis and accumulation . Currently, research on plant metabolites is mainly concerned with crop improvement, assisted breeding, discovering biomarkers, the assessment of nutrients, and biotic and abiotic stress studies . Metabolomics for the analysis of flavonoid metabolites is commonly used to analyze the mechanism of plant color formation . For example, rice during yellowing is induced by the metabolism of flavones, flavonols, isoflavones, and anthocyanidins . Furthermore, a significant correlation was found between the accumulation of malvidin 3-O-glucoside and pelargonidin 3-O-glucoside and the change in leaf color from green to red in Padus virginiana . These results indicate that metabolomics is an important and effective method for analyzing the mechanisms of plant color formation. In order to comprehensively analyze the relationship between leaf color and flavonoids and anthocyanins, leaves from the same plant at three developmental stages, namely red, medium and green leaves, were selected as study materials. Metabolomics was used to analyze the changing pattern of flavonoid metabolites and key metabolites in the process of the leaf color change from red to green, and the mechanism of the leaf color difference was explored using transcriptomics analysis. Candidate genes and metabolic pathways for leaf color variation were further demonstrated. Our results provide a new perspective to understand the flavonoid metabolism of A. truncatum at different developmental stages, which is conducive to the utilization of its leaf resources.
2.1. Variations in Phenotypes and Analysis of Colour Parameters The leaves of A. truncatum showed colors at different stages of fall leaf coloration. With the development of the leaves, the leaves gradually changed from green to middle, and finally the mature leaves developed into red, as seen in A. During the change of leaf color from green to red, observations showed that the trend of L* and b* values followed the same patterns, with an initial increase and then a decrease, and the overall fluctuation range was large ( B). However, L* and b* were not significantly different between red and green leaves. Notably, a* showed a continuous upward trend, which was significantly different among all groups. 2.2. Analysis of Bioactive Flavonoids In the OPLS-DA model, the Q2 values for pairwise comparisons exceeded 0.82 and the Q2 values exceeded 0.9 in the two pairwise groups compared with green leaves ( A). The PCA showed that the composition of the metabolites for the leaf colors of the three different developmental phases differed considerably ( B). In the clustered heat map containing all samples, leaf metabolites in the red and middle phases had similar expression patterns. The metabolites of the green phase were distinctly different from the metabolites of the other phases and they were assigned between the two branches of the cluster ( C). The flavonoids were analyzed in A. truncatum ( and ). In the flavonoid category, o-methylated flavonoids, flavans, flavonoid glycosides, flavones, and biflavonoids and polyflavonoids were identified. Among the o-methylated flavonoids, hesperetin was identified. Among the flavonoid glycosides, cyanidin 3-glucoside, isoquercitrin, kaempferitrin, peonidin-3-glucoside, pelargonidin 3-sophoroside, myricitrin, quercitrin and astragalin were found. In the flavans class, epicatechin, catechin, naringenin, leucopelargonidin, -epigallocatechin, epigallocatechin gallate and eriodictyol were identified. Among the biflavonoids and polyflavonoids, procyanidin B2 was found. The last remaining six belonged to the flavans. 2.3. Analysis of Differentially Expressed Flavonoid Metabolites and KEGG Classification Pairwise comparisons were made between the three groups of materials with different periods of leaf color change ( A and ). A total of 134 significant differentially expressed metabolites (DEMs) were identified between red and middle leaves, of which 98 increased and 36 decreased. In addition, 243 significant DEMs were identified in the comparison between red and green leaves, of which 158 were increasing and 85 were decreasing. Meanwhile, 237 significant DEMs were identified in middle leaves compared with green leaves, of which 143 increased and 94 decreased. The KEGG classification showed that the significant DEMs of red and middle color were mainly involved in aminobenzoate degradation, the biosynthesis of phenylpropanoids, central carbon metabolism in cancer, tyrosine metabolism and styrene degradation pathways ( B). Red and green were mainly related to the biosynthesis of biosynthesis of phenylpropanoids, flavonoid biosynthesis, phenylpropanoid biosynthesis, phenylalanine metabolism and ABC transporters having the smallest p -value. The significant DEMs of middle and green were mainly enriched in the biosynthesis of phenylpropanoids, flavonoid biosynthesis, phenylpropanoid biosynthesis, ABC transporters and the biosynthesis of plant secondary metabolites. The significant DEMs of flavonoid biosynthesis in red vs. green and middle vs. green were subsequently analyzed ( C). In red compared with green, there were 13 flavonoid metabolites of which 5 were significantly up-regulated. In addition, the expression of seven flavonoid metabolites was increased in middle compared to green. Notably, most flavonoid metabolites, especially naringenin, chlorogenic acid, apigenin, taxifolin, dihydromyricetin, 4-coumaroylshikimate, leucopelargonidin, 5,7-dihydroxyflavone, dattelic acid and -gallocatechin, were expressed in both comparison groups. 2.4. Analysis of Transcriptome Results and Functional Annotation In the clustered heat map containing all samples, genes in the red and middle phases had similar expression patterns ( A). The genes of the green phase were distinctly different from the metabolites of the other phases and they were assigned between the two branches of the cluster. The differentially expressed genes (DEGs) of red, middle and green leaves in different developmental stages were analyzed ( B). A total of 2101 DEGs were obtained from the red vs. middle comparison, and the higher number of 1051 DEGs was upregulated. In addition, the highest number of DEGs was 9627 in red vs. green, among which 5369 genes showed upregulated expression, and 4258 genes showed downregulated expression. Simultaneously, the number of 7102 DEGs was obtained in middle vs. green, in which the genes with downregulated expression accounted for 41.98% of all DEGs. The gene ontology (GO) enrichment results showed that the DEGs of the red vs. middle group were significantly enriched in biological processes such as response to starvation, cellular response to phosphate starvation, response to extracellular stimulus, cellular response to starvation and cellular response to external stimulus ( C). The DEGs of red vs. green were enriched in plastid, chloroplast, structural constituent of ribosome, ribosome and thylakoid. The DEGs of the red vs. green group were enriched in chloroplast. In addition, 37 of these genes were annotated to the porphyrin metabolism pathway ( and ). The Kyoto encyclopedia of genes and genomes (KEGG) enrichment results showed that the red vs. middle group was annotated into 116 metabolic pathways ( D), where alanine, aspartate and glutamate metabolism, starch and sucrose metabolism, glycerolipid metabolism, ether lipid metabolism and steroid biosynthesis pathways were significantly enriched. The red vs. green were enriched in 130 metabolic pathways and the middle vs. green were enriched in 128 metabolic pathways. 2.5. Flavonoid Biosynthesis in Relation to Genes and Metabolites The nine-quadrant plot based on correlation analysis shows that metabolite and gene expression patterns are consistent in quadrants three and seven ( A). We found that kaempferin, proanthocyanidin B2, 5,7-dihydroxyflavone, apigenin, quercetin, epigallocatechin gallate and astragalin were identified as being related the six genes F3H , FLS , ANS , LAR , DFR, CHS and CYP75B1 . The epicatechin, naringenin, leucopelargonidin and UGT75C1 genes were identified as related. To further confirm the reliability of the RNA-Seq results, these eight candidate genes were selected for verification. The qRT-PCR analysis of the genes encoding these enzymes showed that all genes were significantly down-regulated ( B). A comprehensive analysis of enzyme activity and the corresponding gene expression pattern showed that the changing trend of enzyme activity of ANS , CHS , DFR , FLS and LAR was consistent with that of the gene expression pattern, while that of the other two enzymes was different from that of the gene expression pattern. 2.6. Analysis of Flavonoid and Anthocyanin Biosynthesis The combination analysis indicated that the expression of genes related to flavonoid synthesis in red leaves was higher than that in green leaves in the same developmental stage . Under the action of DFR , dihydroquercetin (DHQ), dihydrokempferol (DHK) and dihydromyricetin (DHM) were transformed into leucocyanidin, leucopelargonidin and eucodelphinidin. They were generated as anthocyanins under the action of ANS , so anthocyanins accumulate in red and middle leaves. With the development of leaf color, the CHS , F3H , DFR and ANS genes were continuously upregulated in the subsequent developmental stages. By contrast, LAR showed continuous downregulation, which led to the reduced conversion of leucocyanidin into catechin. Five relevant metabolites, namely cyanidin 3-O-β-D-sambubioside, cyanidin 3-O rutinoside, pelargonidin 3-O-3″,6″-O-dimalonylglucoside, delphinidin 3,7-di-O-β-D-glucoside and 3-O-β-D-sambubioside, showed differential changes in the process. In addition, compared with green leaves, cyanidin 3-O-beta-D-sambubioside and cyanidin 3-O rutinoside showed the most significant increases in red leaves during the two stages. 2.7. Analysis of Proteins Encoding Genes for Flavonoids and Anthocyanin Metabolites The predicted secondary structures of the eight proteins encoding genes showed that all the proteins consisted of four parts: an alpha helix, an extended chain, a beta turn, and a random coil . For another, these proteins had the higher proportion of random coil or α-helix, followed by an alpha helix and an extended chain, and the secondary structure of CYP75B1 had the lowest proportion of β-turn. Second, their encoded proteins’ tertiary structure includes alpha helices, extended chains, and random coils . The tertiary structure shows that the secondary structure further folds in a more regular manner, with similar structures formed by different proteins, indicating that their functions are different and further demonstrating the diversity of functions of its members. These genes encode a minimum of 265 amino acids and a maximum of 517 amino acids (CYP75B1) . Meanwhile, three identical structural domains with two identical motifs were found in F3H, FLS and ANS. The three proteins were predicted to belong to the protein family Plant 2OG-oxidoreductases (2ODOs). CHS possesses the active site of the enzyme chalcone/stilbene synthase. Furthermore, the protein encoded by CYP75B1 belongs to the cytochrome P450. In addition, the structural genes were usually controlled by transcription factors in leaf color-related biosynthetic pathways, such as MYB , bHLH , NAC and WRKY . Our study identified 36 MYBs , 126 bHLHs , 78 NACs and 41 WRKYs that were up-regulated in the group of red vs. green. These up-regulated transcription factors were positively correlated with anthocyanin metabolism in leaves. One NAC was annotated to the flavonoid synthesis pathway, while two MYBs were localized to the porphyrin metabolism pathway .
The leaves of A. truncatum showed colors at different stages of fall leaf coloration. With the development of the leaves, the leaves gradually changed from green to middle, and finally the mature leaves developed into red, as seen in A. During the change of leaf color from green to red, observations showed that the trend of L* and b* values followed the same patterns, with an initial increase and then a decrease, and the overall fluctuation range was large ( B). However, L* and b* were not significantly different between red and green leaves. Notably, a* showed a continuous upward trend, which was significantly different among all groups.
In the OPLS-DA model, the Q2 values for pairwise comparisons exceeded 0.82 and the Q2 values exceeded 0.9 in the two pairwise groups compared with green leaves ( A). The PCA showed that the composition of the metabolites for the leaf colors of the three different developmental phases differed considerably ( B). In the clustered heat map containing all samples, leaf metabolites in the red and middle phases had similar expression patterns. The metabolites of the green phase were distinctly different from the metabolites of the other phases and they were assigned between the two branches of the cluster ( C). The flavonoids were analyzed in A. truncatum ( and ). In the flavonoid category, o-methylated flavonoids, flavans, flavonoid glycosides, flavones, and biflavonoids and polyflavonoids were identified. Among the o-methylated flavonoids, hesperetin was identified. Among the flavonoid glycosides, cyanidin 3-glucoside, isoquercitrin, kaempferitrin, peonidin-3-glucoside, pelargonidin 3-sophoroside, myricitrin, quercitrin and astragalin were found. In the flavans class, epicatechin, catechin, naringenin, leucopelargonidin, -epigallocatechin, epigallocatechin gallate and eriodictyol were identified. Among the biflavonoids and polyflavonoids, procyanidin B2 was found. The last remaining six belonged to the flavans.
Pairwise comparisons were made between the three groups of materials with different periods of leaf color change ( A and ). A total of 134 significant differentially expressed metabolites (DEMs) were identified between red and middle leaves, of which 98 increased and 36 decreased. In addition, 243 significant DEMs were identified in the comparison between red and green leaves, of which 158 were increasing and 85 were decreasing. Meanwhile, 237 significant DEMs were identified in middle leaves compared with green leaves, of which 143 increased and 94 decreased. The KEGG classification showed that the significant DEMs of red and middle color were mainly involved in aminobenzoate degradation, the biosynthesis of phenylpropanoids, central carbon metabolism in cancer, tyrosine metabolism and styrene degradation pathways ( B). Red and green were mainly related to the biosynthesis of biosynthesis of phenylpropanoids, flavonoid biosynthesis, phenylpropanoid biosynthesis, phenylalanine metabolism and ABC transporters having the smallest p -value. The significant DEMs of middle and green were mainly enriched in the biosynthesis of phenylpropanoids, flavonoid biosynthesis, phenylpropanoid biosynthesis, ABC transporters and the biosynthesis of plant secondary metabolites. The significant DEMs of flavonoid biosynthesis in red vs. green and middle vs. green were subsequently analyzed ( C). In red compared with green, there were 13 flavonoid metabolites of which 5 were significantly up-regulated. In addition, the expression of seven flavonoid metabolites was increased in middle compared to green. Notably, most flavonoid metabolites, especially naringenin, chlorogenic acid, apigenin, taxifolin, dihydromyricetin, 4-coumaroylshikimate, leucopelargonidin, 5,7-dihydroxyflavone, dattelic acid and -gallocatechin, were expressed in both comparison groups.
In the clustered heat map containing all samples, genes in the red and middle phases had similar expression patterns ( A). The genes of the green phase were distinctly different from the metabolites of the other phases and they were assigned between the two branches of the cluster. The differentially expressed genes (DEGs) of red, middle and green leaves in different developmental stages were analyzed ( B). A total of 2101 DEGs were obtained from the red vs. middle comparison, and the higher number of 1051 DEGs was upregulated. In addition, the highest number of DEGs was 9627 in red vs. green, among which 5369 genes showed upregulated expression, and 4258 genes showed downregulated expression. Simultaneously, the number of 7102 DEGs was obtained in middle vs. green, in which the genes with downregulated expression accounted for 41.98% of all DEGs. The gene ontology (GO) enrichment results showed that the DEGs of the red vs. middle group were significantly enriched in biological processes such as response to starvation, cellular response to phosphate starvation, response to extracellular stimulus, cellular response to starvation and cellular response to external stimulus ( C). The DEGs of red vs. green were enriched in plastid, chloroplast, structural constituent of ribosome, ribosome and thylakoid. The DEGs of the red vs. green group were enriched in chloroplast. In addition, 37 of these genes were annotated to the porphyrin metabolism pathway ( and ). The Kyoto encyclopedia of genes and genomes (KEGG) enrichment results showed that the red vs. middle group was annotated into 116 metabolic pathways ( D), where alanine, aspartate and glutamate metabolism, starch and sucrose metabolism, glycerolipid metabolism, ether lipid metabolism and steroid biosynthesis pathways were significantly enriched. The red vs. green were enriched in 130 metabolic pathways and the middle vs. green were enriched in 128 metabolic pathways.
The nine-quadrant plot based on correlation analysis shows that metabolite and gene expression patterns are consistent in quadrants three and seven ( A). We found that kaempferin, proanthocyanidin B2, 5,7-dihydroxyflavone, apigenin, quercetin, epigallocatechin gallate and astragalin were identified as being related the six genes F3H , FLS , ANS , LAR , DFR, CHS and CYP75B1 . The epicatechin, naringenin, leucopelargonidin and UGT75C1 genes were identified as related. To further confirm the reliability of the RNA-Seq results, these eight candidate genes were selected for verification. The qRT-PCR analysis of the genes encoding these enzymes showed that all genes were significantly down-regulated ( B). A comprehensive analysis of enzyme activity and the corresponding gene expression pattern showed that the changing trend of enzyme activity of ANS , CHS , DFR , FLS and LAR was consistent with that of the gene expression pattern, while that of the other two enzymes was different from that of the gene expression pattern.
The combination analysis indicated that the expression of genes related to flavonoid synthesis in red leaves was higher than that in green leaves in the same developmental stage . Under the action of DFR , dihydroquercetin (DHQ), dihydrokempferol (DHK) and dihydromyricetin (DHM) were transformed into leucocyanidin, leucopelargonidin and eucodelphinidin. They were generated as anthocyanins under the action of ANS , so anthocyanins accumulate in red and middle leaves. With the development of leaf color, the CHS , F3H , DFR and ANS genes were continuously upregulated in the subsequent developmental stages. By contrast, LAR showed continuous downregulation, which led to the reduced conversion of leucocyanidin into catechin. Five relevant metabolites, namely cyanidin 3-O-β-D-sambubioside, cyanidin 3-O rutinoside, pelargonidin 3-O-3″,6″-O-dimalonylglucoside, delphinidin 3,7-di-O-β-D-glucoside and 3-O-β-D-sambubioside, showed differential changes in the process. In addition, compared with green leaves, cyanidin 3-O-beta-D-sambubioside and cyanidin 3-O rutinoside showed the most significant increases in red leaves during the two stages.
The predicted secondary structures of the eight proteins encoding genes showed that all the proteins consisted of four parts: an alpha helix, an extended chain, a beta turn, and a random coil . For another, these proteins had the higher proportion of random coil or α-helix, followed by an alpha helix and an extended chain, and the secondary structure of CYP75B1 had the lowest proportion of β-turn. Second, their encoded proteins’ tertiary structure includes alpha helices, extended chains, and random coils . The tertiary structure shows that the secondary structure further folds in a more regular manner, with similar structures formed by different proteins, indicating that their functions are different and further demonstrating the diversity of functions of its members. These genes encode a minimum of 265 amino acids and a maximum of 517 amino acids (CYP75B1) . Meanwhile, three identical structural domains with two identical motifs were found in F3H, FLS and ANS. The three proteins were predicted to belong to the protein family Plant 2OG-oxidoreductases (2ODOs). CHS possesses the active site of the enzyme chalcone/stilbene synthase. Furthermore, the protein encoded by CYP75B1 belongs to the cytochrome P450. In addition, the structural genes were usually controlled by transcription factors in leaf color-related biosynthetic pathways, such as MYB , bHLH , NAC and WRKY . Our study identified 36 MYBs , 126 bHLHs , 78 NACs and 41 WRKYs that were up-regulated in the group of red vs. green. These up-regulated transcription factors were positively correlated with anthocyanin metabolism in leaves. One NAC was annotated to the flavonoid synthesis pathway, while two MYBs were localized to the porphyrin metabolism pathway .
The regulation of plant leaf color is a complex process . Some results showed that anthocyanins were a major class of compounds that cause changes in leaf color in plants . The color of the leaves was in turn related to the content of the mostly studied phytochemicals, which include flavonoids and phenolic acids . Furthermore, fluctuations in plant leaf chemical levels could trigger leaves to take on different colors. The other results showed that leaf pigment content affects leaf color differences and its content was negatively correlated with L* . Furthermore, when the anthocyanin content of Lycium barbarum L. accumulated, it was reflected in the color by the process of turning from green to red and continuing to deepen . In this study, L* was assigned the highest value in the middle color, which indicated that the chlorophyll content might reach its lowest at this point. When the leaf color changed from green to red, the chlorophyll content had a tendency to decrease and then increase. It is known that the a* value is positively correlated with anthocyanin; when a* gradually increased, the leaf undergoes the process of anthocyanin accumulation. Therefore, we assumed that the content of both flavonoids and anthocyanins increase gradually in the process of coloration of the leaves to red. The molecular basis of flavonoid biosynthesis is increasingly important today . For instance, the metabolite analysis of Ficus carica L. detected 15 different flavonoid-related metabolites, which included the very significant accumulation of the colorless flavonoids procyanidin B1, luteolin-3′,7-di-O-glucoside, epicatechin and quercetin-3-O-rhamnoside in the mature purple peel . A total of 40 flavonoid metabolites were identified through the metabolite extraction and characterization of Cucumis melo . In these metabolites, flavonoids, flavanones, isoflavones and anthocyanins were the substances that mainly affected fruit color . The quantitative analysis revealed that the four varieties in question contained a combination of 125 distinct flavonoids in Actinidia arguta , with only delphinidin 3-O-glucoside, cyanidin O-octanoic acid and pelargonidin 3-O-β-D-glucoside being detected in the red and purple fruits . An aggregate of 23 flavonoid-related metabolites were detected in this article. These belong to o-methylated flavonoids, flavones, flavans, flavonoid glycosides and biflavonoids and polyflavonoids. However, in this study, only five anthocyanins had significant discrepancies among the different stages: cyanidin 3-O-β-D-sambubioside, cyanidin 3-O rutinoside, pelargonidin 3-O-3″,6″-O-dimalonylglucoside, delphinidin 3,7-di-O-β-D-glucoside and 3-O-beta-D-sambubioside. Interestingly, more anthocyanins associated with the leaf color difference were detected. Additionally, there were similarities and differences in the types of flavonoid metabolites between the above plants and A. truncatum . Both it and Actinidia arguta also contained delphinidin, which was not present in the other plants, suggesting differences in flavonoid composition between species. Phenylalanine, an upstream reaction of flavonoids and anthocyanins, was first converted to p-coumaroyl-CoA coenzyme a catalyzed by PAL , C4H and 4CL . In the presence of CHS and CHI , p-coumaroyl-CoA was converted to naringenin chalcone and then to naringenin. The expression of CHS was up-regulated in the process of the leaf color change, which was consistent with the qRT-PCR results. Meanwhile, CHS was the first key enzyme in flavonoid synthesis and its activity determines the formation of related metabolites . The expression of CHS in the red leaf stage was highly significant for the green leaf stage, suggesting that CHS may be involved in the process of leaf color changes. Then, naringenin was catalyzed by F3H to form DHK, which then continued to be catalyzed by F3H to form DHQ and DHM, respectively. Under the catalytic action of DFR and ANS , DHK, DHQ and DHM formed unstable anthocyanins such as cyanidin, pelargonidin and delphinidin, respectively. DHK, DHQ and DHM catalyzed the DFR to produce different ones in different plants and the ANS was necessary for the accumulation of different anthocyanins . The unstable anthocyanins produced eventually formed stable anthocyanins in the presence of UGT75C1 . To summarize, the five stabilities of anthocyanins that were described in detail in the previous paragraph were predominantly found in A. truncatum . 2ODDs are a family of proteins with both DIOX_N and 2OG-FeII_Oxy conserved structures . The related article pointed to them as the second-largest family of oxidative enzymes in plants, involved in various oxidative reactions . They are widely involved in secondary metabolic processes in plants, such as the biosynthesis of flavonoids, alkaloids and terpenoids . We expected that 2ODD could catalyze the conversion of naringenin into dihydrokaempferol, indicating that the enzyme was a typical F3H . This was demonstrated in the investigation that we have carried out. In our study, F3H, FLS and ANS belonged to this family. They possessed the same conserved structural domains DIOX_N and 2OG-FeII_Oxy described above, which could be involved in flavonoid formation and have an impact on anthocyanin biosynthesis. Correspondingly, we need to focus on the cytochrome P450 family. The P450s are thioredoxin proteins involved in the oxidative degradation of various compounds. P450 was imprinted in Scutellaria baicalensis for anthocyanin modification . The characterization and analysis of P450 from grapes revealed that its subfamily CYP75 is involved in anthocyanin production in a similar way . CYP75B1, which undoubtedly possesses the conserved structural domain of P450, affects the production of DHQ, DHK and DHM from naringenin. Increased gene activity of these members from different families triggers the accumulation of flavonoids and anthocyanins in red leaves.
4.1. Plant Materials and Sampling A. truncatum growing in the wild on Jilin Agricultural University campus (43°05′–45°15′ N, 124°18′–127°05′ E) was used. Based on the color pattern of the leaves, three different colored leaves of the same plant were selected as study material (red, middle and green leaves). Ten leaves were collected from each group and leaves with a similar location and color on the branches were selected. The test materials were divided into two parts: one half of the samples was quickly scanned for leaf color parameters, and the other half of the samples was quickly fixed with liquid nitrogen and later moved to storage at −80 °C. 4.2. Determination of Leaf Colour Parameters Leaf color was quantified in accordance with the International Commission on Illumination (CII) color standard. Luminosity (L*), a*, and b* values were obtained through the use of a CR30 colorimeter (CHNSpec, Hanghzhou China). After calibration using the colorimetric plate, five points were randomly selected and averaged on each blade, with the objective of avoiding the leaf edges and radial main veins . The exercise was repeated five times for each leaf color. The meanings of the three parameters are as follows. L* indicates the brightness of the color, where a positive number means whitish and a negative number means blackish. Additionally, a* indicates the red–green value, where a positive value means the color is redder and a negative value means it is greener. Finally, b* indicates the yellow–blue value, where positive values are yellowish and negative values are bluish. 4.3. Extraction of Total Metabolites First, the sample was weighed accurately in a 2 mL centrifuge tube and 600 µL of MeOH containing 2-Amino-3-(2-chloro-phenyl)-propionic acid (4 ppm) was added and vortexed for 30 s. Second, The sample was then placed in a tissue grinder and ground at 55 Hz for 60 s, followed by sonication at room temperature for 15 min. Finally, the sample was centrifuged at 12,000 rpm at 4 °C for 10 min on a H1850-R refrigerated centrifuge (Cence, Changsha, China), and the supernatant was taken through a 0.22 μm membrane and added into the detection vial for LC-MS detection . 4.4. Metabolomics Analysis The LC analysis was performed on a Vanquish UHPLC System (Thermo Fisher Scientific, Waltham, MA, USA). Chromatography was carried out with an ACQUITY UPLC ® HSS T3 (2.1 × 100 mm, 1.8 µm) (Waters, Milford, MA, USA). The column was maintained at 40 °C. The flow rate and injection volume were set at 0.3 mL/min and 2 μL, respectively. Mass spectrometric detection of metabolites was performed on Q Exactive (Thermo Fisher Scientific, Waltham, MA, USA) with an ESI ion source. Simultaneous MS1 and MS/MS (full MS-ddMS2 mode, data-dependent MS/MS) acquisition was used. The parameters were as follows: capillary temperature, 325 °C; MS1 range, m / z 100–1000; MS1 resolving power, 70,000 FWHM; number of data-dependent scans per cycle, 10; MS/MS resolving power, FWHM; normalized collision energy, 30 eV; dynamic exclusion time, automatic. The raw mass spectrometry downcomer files were converted into the mzXML file format using the MSConvert tool in the Proteowizard software package (v3.0.8789). Peak detection, peak filtering and peak alignment were carried out using the R XCMS software package (v3.12.0), and a list of substances for quantification was obtained. Substances with coefficients of variation smaller than 30% in the QC samples were then retained for subsequent analysis. The molecular weight of the metabolite was ascertained by means of the mass-to-charge ratio of the parent ion present in the primary mass spectrum. Furthermore, the molecular formula was deduced on the basis of the mass number deviation, as well as the information provided by the additional ions. Following this, the metabolite was matched with a previously existing record in a database, thereby achieving its preliminary identification. Meanwhile, the metabolites detected in the secondary spectrum were subjected to a process of secondary identification. This involved the matching of the metabolite data with the information contained in the database, including the fragment ions of each metabolite. 4.5. Total RNA Extraction and Transcriptome Sequencing The mRNA with polyA structure in the total RNA was enriched by Oligo(dT) magnetic beads. Subsequently, the RNA was subjected to ionic interruption, which fragmented it into fragments of approximately 300 base pairs in length. The initial cDNA strand was synthesized using RNA as a template with a 6-base random primer and reverse transcriptase. The second cDNA strand was then synthesized with the initial cDNA strand serving as the template. Following the construction of the library, the library fragments were amplified by PCR. The library size was 450 bp, and the total and effective concentrations were subsequently determined by an Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA). Subsequently, the mixed libraries were uniformly diluted to 2 nM and denatured by alkaline denaturation, thereby forming single-stranded libraries. Following the extraction, purification and construction of the libraries, the libraries were subjected to paired-end sequencing using next-generation sequencing (NCBI project: PRJNA1196921). 4.6. Identification and Analysis of DEGs FPKM was used to normalize the raw gene expression calculation. In order to identify genes that were differentially expressed among the three different groups, we identified genes with |log 2 foldchang| > 1 and a p -value < 0.05 as DEGs. GO functional enrichment analysis and KEGG pathway analysis were performed on the confirmed DEGs. 4.7. qRT-PCR Analysis Primers specific to structural genes involved in flavonoid biosynthesis were designed for qRT-PCR analysis using Primer Premier 5 software . All samples were subjected to three replicates, as were three technical replicates. The internal control genes employed were actin and β-tubulin . The relative expression levels of the target genes were calculated employing the 2 −∆∆Ct methodology. Three experimental replicates were performed for each sample. 4.8. Protein Biology Analysis The secondary structure of the protein encoding the key enzyme gene for flavonoids and anthocyanin synthesis was predicted using the SOPMA online tool ( https://npsa.lyon.inserm.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_sopma.html , accessed on 4 October 2024) and modeled for the tertiary structure using the Phyre2 online tool ( http://www.sbg.bio.ic.ac.uk/phyre2/ , accessed on 4 October 2024) for tertiary structure modeling. The proteins were also functionally analyzed using the website InterPro ( https://www.ebi.ac.uk/interpro/ , accessed on 5 October 2024).
A. truncatum growing in the wild on Jilin Agricultural University campus (43°05′–45°15′ N, 124°18′–127°05′ E) was used. Based on the color pattern of the leaves, three different colored leaves of the same plant were selected as study material (red, middle and green leaves). Ten leaves were collected from each group and leaves with a similar location and color on the branches were selected. The test materials were divided into two parts: one half of the samples was quickly scanned for leaf color parameters, and the other half of the samples was quickly fixed with liquid nitrogen and later moved to storage at −80 °C.
Leaf color was quantified in accordance with the International Commission on Illumination (CII) color standard. Luminosity (L*), a*, and b* values were obtained through the use of a CR30 colorimeter (CHNSpec, Hanghzhou China). After calibration using the colorimetric plate, five points were randomly selected and averaged on each blade, with the objective of avoiding the leaf edges and radial main veins . The exercise was repeated five times for each leaf color. The meanings of the three parameters are as follows. L* indicates the brightness of the color, where a positive number means whitish and a negative number means blackish. Additionally, a* indicates the red–green value, where a positive value means the color is redder and a negative value means it is greener. Finally, b* indicates the yellow–blue value, where positive values are yellowish and negative values are bluish.
First, the sample was weighed accurately in a 2 mL centrifuge tube and 600 µL of MeOH containing 2-Amino-3-(2-chloro-phenyl)-propionic acid (4 ppm) was added and vortexed for 30 s. Second, The sample was then placed in a tissue grinder and ground at 55 Hz for 60 s, followed by sonication at room temperature for 15 min. Finally, the sample was centrifuged at 12,000 rpm at 4 °C for 10 min on a H1850-R refrigerated centrifuge (Cence, Changsha, China), and the supernatant was taken through a 0.22 μm membrane and added into the detection vial for LC-MS detection .
The LC analysis was performed on a Vanquish UHPLC System (Thermo Fisher Scientific, Waltham, MA, USA). Chromatography was carried out with an ACQUITY UPLC ® HSS T3 (2.1 × 100 mm, 1.8 µm) (Waters, Milford, MA, USA). The column was maintained at 40 °C. The flow rate and injection volume were set at 0.3 mL/min and 2 μL, respectively. Mass spectrometric detection of metabolites was performed on Q Exactive (Thermo Fisher Scientific, Waltham, MA, USA) with an ESI ion source. Simultaneous MS1 and MS/MS (full MS-ddMS2 mode, data-dependent MS/MS) acquisition was used. The parameters were as follows: capillary temperature, 325 °C; MS1 range, m / z 100–1000; MS1 resolving power, 70,000 FWHM; number of data-dependent scans per cycle, 10; MS/MS resolving power, FWHM; normalized collision energy, 30 eV; dynamic exclusion time, automatic. The raw mass spectrometry downcomer files were converted into the mzXML file format using the MSConvert tool in the Proteowizard software package (v3.0.8789). Peak detection, peak filtering and peak alignment were carried out using the R XCMS software package (v3.12.0), and a list of substances for quantification was obtained. Substances with coefficients of variation smaller than 30% in the QC samples were then retained for subsequent analysis. The molecular weight of the metabolite was ascertained by means of the mass-to-charge ratio of the parent ion present in the primary mass spectrum. Furthermore, the molecular formula was deduced on the basis of the mass number deviation, as well as the information provided by the additional ions. Following this, the metabolite was matched with a previously existing record in a database, thereby achieving its preliminary identification. Meanwhile, the metabolites detected in the secondary spectrum were subjected to a process of secondary identification. This involved the matching of the metabolite data with the information contained in the database, including the fragment ions of each metabolite.
The mRNA with polyA structure in the total RNA was enriched by Oligo(dT) magnetic beads. Subsequently, the RNA was subjected to ionic interruption, which fragmented it into fragments of approximately 300 base pairs in length. The initial cDNA strand was synthesized using RNA as a template with a 6-base random primer and reverse transcriptase. The second cDNA strand was then synthesized with the initial cDNA strand serving as the template. Following the construction of the library, the library fragments were amplified by PCR. The library size was 450 bp, and the total and effective concentrations were subsequently determined by an Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA). Subsequently, the mixed libraries were uniformly diluted to 2 nM and denatured by alkaline denaturation, thereby forming single-stranded libraries. Following the extraction, purification and construction of the libraries, the libraries were subjected to paired-end sequencing using next-generation sequencing (NCBI project: PRJNA1196921).
FPKM was used to normalize the raw gene expression calculation. In order to identify genes that were differentially expressed among the three different groups, we identified genes with |log 2 foldchang| > 1 and a p -value < 0.05 as DEGs. GO functional enrichment analysis and KEGG pathway analysis were performed on the confirmed DEGs.
Primers specific to structural genes involved in flavonoid biosynthesis were designed for qRT-PCR analysis using Primer Premier 5 software . All samples were subjected to three replicates, as were three technical replicates. The internal control genes employed were actin and β-tubulin . The relative expression levels of the target genes were calculated employing the 2 −∆∆Ct methodology. Three experimental replicates were performed for each sample.
The secondary structure of the protein encoding the key enzyme gene for flavonoids and anthocyanin synthesis was predicted using the SOPMA online tool ( https://npsa.lyon.inserm.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_sopma.html , accessed on 4 October 2024) and modeled for the tertiary structure using the Phyre2 online tool ( http://www.sbg.bio.ic.ac.uk/phyre2/ , accessed on 4 October 2024) for tertiary structure modeling. The proteins were also functionally analyzed using the website InterPro ( https://www.ebi.ac.uk/interpro/ , accessed on 5 October 2024).
In this study, we focused on the diversity of flavonoid compounds in the red, middle and green leaves of A. truncatum to explore the molecular mechanisms of leaf color formation. The visual diversity of different leaf colors was first described through leaf color parameters, using a digital method. Moreover, a total of 23 different modified flavonoids were detected by metabolomics. In particular, cyanidin 3-O-β-D-sambubioside, cyanidin 3-O rutinoside, pelargonidin 3-O-3″,6″-O-dimalonylglucoside, delphinidin 3,7-di-O-β-D-glucoside and 3-O-β-D-sambubioside could be responsible for the differences between green and red leaves. Furthermore, RNA-seq analysis showed that the up-regulation of CHS , DFR and ANS expression led to an increase in the corresponding anthocyanin red color, which resulted in the reddening of the leaves. Additionally, UGT75C1 was correspondingly essential as a downstream gene for the synthesis of anthocyanin. By modeling the proteins encoded by these genes and analyzing the conserved structural domains, the results corroborated the reliability of the metabolomic and transcriptomic data. Overall, the study offers valuable insights into the flavonoid-related metabolite composition in A. truncatum, providing essential reference points for breeders seeking to enhance the pigmentation of this species.
|
Petri-plate, bacteria, and laser optical scattering sensor | abbbf0cb-8495-4ac3-9f9e-4cbd91d798f9 | 9813400 | Microbiology[mh] | Possibly, the single most important invention in the field of microbiology was the Petri plate, which enabled the separation and isolation of microbes from a complex mixture ( ). At the time he invented the Petri plate, Julius Richard Petri was a military physician working in Robert Koch’s lab in Germany during the 1880s. Today it is a quintessential tool in microbiology laboratories allowing bacterial isolation, enumeration, mutagenesis, genetic manipulation, antibiotic sensitivity/resistance testing, enzymatic activity, hemolytic activity assessment, and many more ( ). Before Petri’s invention, scientists used sliced potatoes’ surfaces as a means for the separation and isolation of microbes, but the approach did not reliably yield a sterile environment. The Petri plate, with a sterile media and a lid, can also be sterilized, providing a pristine environment in which microbes can be cultured, isolated, identified, and studied based on individual colony formation. A single bacterial cell divides through binary fission, first in two-dimensional space and subsequently in three-dimensional space ( ), forming a colony on the agar surface. Depending on the organism, colony architecture, shape, size, and chromogen production become characteristic features of a given bacterium for visual identification. A bacterial colony is viewed as a self-engineered multicellular organism exhibiting intricate communication skills and social intelligence ( ; ), in response to the adjoining environment ( ; ; ). Genotypic variations in an individual bacterium amplify in the colony, by a million or billion-fold, which affects the colony morphotype ( ; ). Modern-day nano-bio sensors have begun to revolutionize the detection of a single molecule or a single cell with high precision, but most depend on probes (labels) or signature tags for the identification of targets ( ; ; ; ; ; ). Very few biosensors operate independently of molecular probes, but are reliant on a database, and are categorized as label-free. We have discovered that a red diode laser (635 nm) upon shining on the center of a bacterial colony generates a unique scatter signature, allowing direct and instantaneous identification of bacteria on the Petri plate without disturbing the colony integrity ( ) ( ). The device is called BARDOT (Bacterial Rapid Detection using Optical scattering Technology), which exploits genetic (metabolic) and phenotypic (functional) differences in microbes for identification and can aid in screening the most desirable isolates for further study including gene sequencing, pathogenicity, antibiotic susceptibility, vaccine development, and for characterizing industrially beneficial traits ( ; ; ). Recent emphasis on culturomics ( ) lead to explosive progress in the development of specialized growth media for the isolation of microbial pathogens on Petri-plate. Integration of culturomics with the biophysical identification tool such as BARDOT would not only help in making an informed decision but also would be an indispensable tool for early diagnosis and detection of microbes in human and veterinary medicine, food hygiene, and agriculture (plant and fish pathogens). Petri plate together with the BARDOT device would aid in high throughput screening of a large number of colonies for identification and facilitate polymerase chain reaction (PCR)-based confirmation, whole-genome sequencing, and mass spectrometry ( ). BARDOT can also be used to study pathogens and spoilage microbes and the microbial community and their shift in response to environmental cues. In this review, the historical perspectives of the Petri plate and its contribution to the modern-day microbial study using laser optical sensors, including BARDOT, Raman, and hyperspectral imaging systems are emphasized.
Petri plate “We have here the first description of the Petri dish, a simple yet effective device for culturing microorganisms on solid media” – R.J. ( ). Julius Richard Petri, a military physician working in Robert Koch’s lab in Germany during the 1880s invented the Petri plate. In 1887, Petri was trying to find a solution to the problem of culture contamination. Before the development of the Petri plate, scientists used various growth media that would allow their cultures to grow, such as sliced potato, cooked egg whites, and gelatin. These were all attempts at creating a solid growth medium that would facilitate the growth of microorganisms. To make the sample contaminant-free, a bell jar was placed over the culture. Petri’s innovation created a simple tool that effectively provides a small, isolated environment for a culture to grow free of contamination. The modern Petri plate was the culmination of several innovations that created a sterile environment for bacterial growth. In 1882, Frannie Hesse, the wife and lab technician of Walter Hesse, began substituting agar for the gelatin that she was putting into his solid media in tubes. By replacing gelatin with agar, scientists were able to have a stable solid media that could be incubated, could not easily be degraded by microbes, and was transparent. The invention of the autoclave by Charles Chamerbland in 1884 allowed for the complete sterilization of equipment and media. In 1887, when Julius Petri developed a glass plate that had another, larger plate as its lid, he created an intuitive tool that was compact, reusable, and able to house any solid media. With the combination of these three technologies, the modern Petri plate was born. Scientists could now be certain that their cultures were stable due to their agar base, sterile because of the use of the autoclave, and safe from contamination. This simple solution was fundamental to the primary use of solid media for the isolation and separation of culturable bacteria. His plates have since allowed researchers to easily isolate, observe, study, and manipulate the microorganisms. The primary function of routinely separating and isolating microbes on solid media can not be trivialized for its impact as a routine and integral part of traditional or modern microbiology laboratories. Its use has affected and still impacts all aspects of our lives. Some primary examples can be seen in food processing for evaluating the efficacy of microbial inactivation methods to eliminate pathogens and spoilage organisms and biotechnology for isolating recombinant organisms expressing foreign proteins for medical and commercial applications. An example of the Petri plate’s importance in research can be illustrated in its use in the transformative publication “Studies on the chemical nature of the substance inducing transformation of pneumococcal types” ( ). The Petri plate was the tool used to determine that DNA was the genetic material by examining the rough and smooth phenotypes of Streptococcus pneumoniae after growth on plates. The differences in phenotypes in Avery et al.’s work ( ) were obvious and visualized with the naked eye. However, if a more detailed nondestructive analysis of bacterial cultures on plates could be obtained to detect differences in genus, species, and even strains of bacteria it would be of great value to the microbiology community. In 2007, our lab at Purdue University reported a prototype laser optical sensor, called BARDOT (bacterial rapid detection using optical scattering technology) that is capable of bacterial colony detection and characterization directly on the agar surface of the Petri plate using a red-diode laser beam ( ; ) culminated from the earlier groundbreaking work using a laser to physically map bacterial cell morphology on a Petri plate ( ).
“We have here the first description of the Petri dish, a simple yet effective device for culturing microorganisms on solid media” – R.J. ( ). Julius Richard Petri, a military physician working in Robert Koch’s lab in Germany during the 1880s invented the Petri plate. In 1887, Petri was trying to find a solution to the problem of culture contamination. Before the development of the Petri plate, scientists used various growth media that would allow their cultures to grow, such as sliced potato, cooked egg whites, and gelatin. These were all attempts at creating a solid growth medium that would facilitate the growth of microorganisms. To make the sample contaminant-free, a bell jar was placed over the culture. Petri’s innovation created a simple tool that effectively provides a small, isolated environment for a culture to grow free of contamination. The modern Petri plate was the culmination of several innovations that created a sterile environment for bacterial growth. In 1882, Frannie Hesse, the wife and lab technician of Walter Hesse, began substituting agar for the gelatin that she was putting into his solid media in tubes. By replacing gelatin with agar, scientists were able to have a stable solid media that could be incubated, could not easily be degraded by microbes, and was transparent. The invention of the autoclave by Charles Chamerbland in 1884 allowed for the complete sterilization of equipment and media. In 1887, when Julius Petri developed a glass plate that had another, larger plate as its lid, he created an intuitive tool that was compact, reusable, and able to house any solid media. With the combination of these three technologies, the modern Petri plate was born. Scientists could now be certain that their cultures were stable due to their agar base, sterile because of the use of the autoclave, and safe from contamination. This simple solution was fundamental to the primary use of solid media for the isolation and separation of culturable bacteria. His plates have since allowed researchers to easily isolate, observe, study, and manipulate the microorganisms. The primary function of routinely separating and isolating microbes on solid media can not be trivialized for its impact as a routine and integral part of traditional or modern microbiology laboratories. Its use has affected and still impacts all aspects of our lives. Some primary examples can be seen in food processing for evaluating the efficacy of microbial inactivation methods to eliminate pathogens and spoilage organisms and biotechnology for isolating recombinant organisms expressing foreign proteins for medical and commercial applications. An example of the Petri plate’s importance in research can be illustrated in its use in the transformative publication “Studies on the chemical nature of the substance inducing transformation of pneumococcal types” ( ). The Petri plate was the tool used to determine that DNA was the genetic material by examining the rough and smooth phenotypes of Streptococcus pneumoniae after growth on plates. The differences in phenotypes in Avery et al.’s work ( ) were obvious and visualized with the naked eye. However, if a more detailed nondestructive analysis of bacterial cultures on plates could be obtained to detect differences in genus, species, and even strains of bacteria it would be of great value to the microbiology community. In 2007, our lab at Purdue University reported a prototype laser optical sensor, called BARDOT (bacterial rapid detection using optical scattering technology) that is capable of bacterial colony detection and characterization directly on the agar surface of the Petri plate using a red-diode laser beam ( ; ) culminated from the earlier groundbreaking work using a laser to physically map bacterial cell morphology on a Petri plate ( ).
BARDOT Laser-based interrogation of individual bacterial cells in liquid suspension has been previously attempted by P.J. Wyatt and his team in the late ‘60s ( ; ). Measuring the intensities of the full 4-π radian of scattered light from single cells was the detection principle where the light scattering phenomenon occurs in a single scattering regime. The system requires the suspension of one type of organism (purified target culture) to avoid the generation of multiple overlapping scatter signatures from mixed cultures since multiple scattering events may interfere with the specific detection of the target ( ). Therefore, pure cultures are essential, which could be obtained from isolated colonies from a Petri plate. Furthermore, such a detection approach also requires a very low cell density (about 100 cells/ml) to ensure capturing of the single scattering event and avoiding interferences from other scatterers. Our approach in using elastic light scattering for bacterial interrogation was focused on bacterial microcolonies (0.7 mm to 1.2 mm diameter). Single bacterial cells, through binary fission, give rise to a colony on an agar surface, providing masses of cells that can provide a volume of scattered light for the interrogation of those bacteria with relative ease. In BARDOT, a laser beam with 635 nm (1 mm diameter) wavelength and 1 mW power is passed through the center of spatially located well-separated bacterial colonies (about 1 mm diameter) on a Petri dish and generates a unique scatter signature with circular boundaries with concentric rings, radial spokes, wavy lines, or speckles as a fingerprint. Each scatter pattern is unique for a bacterial culture at the species and serovar level on an agar plate containing specific growth media at a specified time of growth ( ; ). Colony scatter patterns changes with time; therefore, it is also critical to find a time window when a scatter signature with multiple features can be reliably used for bacterial identification ( ; ). The physics behind the forward scattering is well understood ( ): An incoming wavefront is capable of interacting with the micro/macro structures of a colony which ‘imprints’ its signature on the outgoing wavefront. This is further propagated to the detector and decodes the characteristics of scattered light intensity ( ; ; ). The BARDOT unit is designed with two subcomponents: a microbial colony locator and a forward scatterometer. The first component is responsible for counting and locating the center coordinates of the individual colony and excluding those that don’t match the detection criteria (doublets or diameter outside of the detection range). The colony locator consists of a ring-type light-emitting diode (LED) array for illumination purposes along with the plate diffuser to provide equal illumination across the plate. A monochromatic CMOS (complementary metal-oxide-semiconductor) camera with 1024 x1280 pixels is located on the top of the plate along with an imaging lens with a viewing angle of 34° x 25.6°. Once the candidate colonies for detection are determined, the list of center locations for those colonies is sent to the forward scatterometer. This subcomponent is responsible for capturing the forward scatter patterns from each colony and the scatter features are captured with a second CCD (charge-coupled device) camera placed on the bottom of the Petri dish. To ensure the capture of a quality scatter pattern, a centering algorithm that minimizes the distance between the center of the laser and the colony has been implemented by calculating the difference in geometric moments ( ). A four-quadrant balancing algorithm was used for quantitatively aligning the laser with the colony while a “traveling salesman algorithm” was implemented to minimize the traveling time between two colonies ( ). At the same time, automated image processing and classification software were also integrated for seamless analysis and identification of microbes for high throughput screening ( ). Using the same optical scattering principle, Buzalewicz et al. ( ) reported the development of BISLD (Bacteria Identification System by Light Diffraction) for the detection of variable-size colonies at a fixed incubation period by adjusting laser beam diameter. A similar fixed incubation approach using the laser scattering method was also employed by others to interrogate variable-size bacterial colonies on a Petri dish ( ; ). Such an approach can overcome the major limitation of BARDOT which uses a fixed-diameter laser beam targeting colonies of a specific size range (0.7 mm – 1.2 mm) while ignoring the colonies outside this range. BARDOT generated scatter image analysis and pathogen identification Once the scattering patterns are captured, they are stored in the database as a fingerprint library for future detection and presumptive identification of the bacteria using advanced classification algorithms. The captured scatter patterns of the bacterial colony are automatically analyzed by the quantitative image processing software. Two major features are used for image analysis: the rotation-invariant feature (circularly symmetric patterns) and texture features (random and speckle patterns). The performance of the classifier is estimated using cross-validation ( ). The importance of quantitative classification software is needed to reduce human errors, which in turn can provide higher sensitivity and specificity than visual observation. Each 2-D scatter pattern is analyzed via Zernike moments and Haralick textures ( ). The former extracts features of circularly symmetric features from the scatter patterns, while the Haralick describes the texture of the scatter patterns. The combination of these two features results in hundreds of signature attributes from a single 2-D image which can be used as an orthogonal basis for the fingerprint library in the classification of scattering patterns of bacterial test samples. Once the training library is built, the sample under investigation can be compared against the fingerprint library that is already built and trained. The results are then reported in a matrix format. The diagonal numbers represent the expected correct classification rate (true positive and true negative) while the off-diagonal numbers show the missed classification (false positive and false negative). More details about the CV matrix are discussed in our relevant publications ( ; ). Raman spectroscopy Raman spectroscopy uses a laser to record the vibrational and rotational properties of molecules yielding a scattering signature referred to inelastic scattering ( ; ). Earlier attempts to detect microcolonies of clinically relevant bacterial pathogens of Staphylococcus aureus , Staphylococcus epidermidis , Escherichia coli and Enterococcus faecium directly from agar plate using a Raman microspectrometer equipped with an 830-nm titanium−sapphire laser at wavenumber 250 to 2150 cm -1 was moderately successful ( ). Later, Rosch et al. ( ) used lasers with three different wavelengths (785 nm; 633, and 514 nm) to differentiate colonies of test organisms ( Micrococcus luteus , Bacillus subtilis , and Pseudomonas fluorescens ) where chromophores produced by these organisms aided in spectral classification. However, bacterial viability was lost due to the destruction of bacterial cells during laser exposure, a major impediment to the isolation of viable cells after Raman spectroscopy ( ). Raman spectroscopy was also successfully used for the detection of colonies of clinically relevant Staphylococcus epidermidis, S. aureus and Escherichia coli strains on blood agar or Mueller-Hinton agar plates ( ; ). Most recently, Shen et al. ( ) reported a rapid fiber probe-based Raman (785 nm diode laser) technique for the classification and identification of 33 strains of 8 different species including Candida albicans , Staphylococcus epidermidis , S. aureus , Klebsiella pneumoniae , K. oxytoca , Escherichia coli , Enterococcus faecalis , E. faecium , and Acinetobacter baumannii on Luria–Bertani (LB) agar plates. Nevertheless, Raman spectroscopy continued to be an attractive on-plate microbial pathogen detection tool for foodborne ( ), and clinical relevant pathogens ( ; ). However, bacterial physiological state, growth phase and growth media can affect the spectral fingerprints thus these parameters should be controlled with care ( ; ). Hyperspectral imaging Hyperspectral imaging (HSI) technology combines spectroscopy and imaging as a reliable nondestructive technique for bacterial colony counting, and detection and identification on various food, inert surfaces, or clinical specimens ( ; ; ; ; ). Lights with small wavelength bandwidth from visible to near-infrared (Vis-NIR) are often used to generate a complete spatiospectral map of a colony for pathogen detection and identification. Scientists at the US Department of Agriculture developed a hyperspectral imaging system with a spectral range from 400 to 1000 nm to detect and differentiate serovars of Shiga-toxin-producing Escherichia coli (STEC) pathogens on Rainbow agar with very high accuracy ( ; ). They also successfully used this platform to detect Campylobacter species ( ). Direct identification of colonies on culture plates is also highly important for clinical diagnostic applications. Arrigoni et al. ( ) applied HSI coupled with a classification algorithm to identify pathogens that are responsible for urinary tract infection, including Escherichia coli , Enterococcus faecalis , Staphylococcus aureus , Proteus mirabilis , Proteus vulgaris , Klebsiella pneumoniae , and Pseudomonas aeruginosa on blood agar plates. HSI was also used for the discrimination of colonies of three different bacterial cultures including E. coli , Listeria monocytogenes and Staphylococcus aureus for application in food safety ( ). This method employed a non-selective agar plate (tryptic soy agar, TSA) for colony identification. Likewise, using HSI and chemometric classification algorithms, Gu et al. ( ) differentially distinguished colonies of Escherichia coli, Staphylococcus aureus , and Salmonella enterica on three different nonselective agar plates. However, the drawback of using non-selective agar plates for HSI application is that the growth of commensal bacteria can interfere with target organism identification when testing with food samples, thus conventional broth culturing techniques using selective antimicrobial agents must be employed before testing food samples on agar plates. Near-infrared (NIR) HSI with multivariate data analysis was shown to be useful for the discrimination of colonies of Bacillus cereus , Escherichia coli , Salmonella enterica serovar Enteritidis, Staphylococcus aureus and S. epidermidis ( ). The application of HSI in a reflectance mode is highly useful for differentially distinguishing bacterial colonies from particulate foods on agar plate surfaces for food safety analysis ( ).
Laser-based interrogation of individual bacterial cells in liquid suspension has been previously attempted by P.J. Wyatt and his team in the late ‘60s ( ; ). Measuring the intensities of the full 4-π radian of scattered light from single cells was the detection principle where the light scattering phenomenon occurs in a single scattering regime. The system requires the suspension of one type of organism (purified target culture) to avoid the generation of multiple overlapping scatter signatures from mixed cultures since multiple scattering events may interfere with the specific detection of the target ( ). Therefore, pure cultures are essential, which could be obtained from isolated colonies from a Petri plate. Furthermore, such a detection approach also requires a very low cell density (about 100 cells/ml) to ensure capturing of the single scattering event and avoiding interferences from other scatterers. Our approach in using elastic light scattering for bacterial interrogation was focused on bacterial microcolonies (0.7 mm to 1.2 mm diameter). Single bacterial cells, through binary fission, give rise to a colony on an agar surface, providing masses of cells that can provide a volume of scattered light for the interrogation of those bacteria with relative ease. In BARDOT, a laser beam with 635 nm (1 mm diameter) wavelength and 1 mW power is passed through the center of spatially located well-separated bacterial colonies (about 1 mm diameter) on a Petri dish and generates a unique scatter signature with circular boundaries with concentric rings, radial spokes, wavy lines, or speckles as a fingerprint. Each scatter pattern is unique for a bacterial culture at the species and serovar level on an agar plate containing specific growth media at a specified time of growth ( ; ). Colony scatter patterns changes with time; therefore, it is also critical to find a time window when a scatter signature with multiple features can be reliably used for bacterial identification ( ; ). The physics behind the forward scattering is well understood ( ): An incoming wavefront is capable of interacting with the micro/macro structures of a colony which ‘imprints’ its signature on the outgoing wavefront. This is further propagated to the detector and decodes the characteristics of scattered light intensity ( ; ; ). The BARDOT unit is designed with two subcomponents: a microbial colony locator and a forward scatterometer. The first component is responsible for counting and locating the center coordinates of the individual colony and excluding those that don’t match the detection criteria (doublets or diameter outside of the detection range). The colony locator consists of a ring-type light-emitting diode (LED) array for illumination purposes along with the plate diffuser to provide equal illumination across the plate. A monochromatic CMOS (complementary metal-oxide-semiconductor) camera with 1024 x1280 pixels is located on the top of the plate along with an imaging lens with a viewing angle of 34° x 25.6°. Once the candidate colonies for detection are determined, the list of center locations for those colonies is sent to the forward scatterometer. This subcomponent is responsible for capturing the forward scatter patterns from each colony and the scatter features are captured with a second CCD (charge-coupled device) camera placed on the bottom of the Petri dish. To ensure the capture of a quality scatter pattern, a centering algorithm that minimizes the distance between the center of the laser and the colony has been implemented by calculating the difference in geometric moments ( ). A four-quadrant balancing algorithm was used for quantitatively aligning the laser with the colony while a “traveling salesman algorithm” was implemented to minimize the traveling time between two colonies ( ). At the same time, automated image processing and classification software were also integrated for seamless analysis and identification of microbes for high throughput screening ( ). Using the same optical scattering principle, Buzalewicz et al. ( ) reported the development of BISLD (Bacteria Identification System by Light Diffraction) for the detection of variable-size colonies at a fixed incubation period by adjusting laser beam diameter. A similar fixed incubation approach using the laser scattering method was also employed by others to interrogate variable-size bacterial colonies on a Petri dish ( ; ). Such an approach can overcome the major limitation of BARDOT which uses a fixed-diameter laser beam targeting colonies of a specific size range (0.7 mm – 1.2 mm) while ignoring the colonies outside this range.
Once the scattering patterns are captured, they are stored in the database as a fingerprint library for future detection and presumptive identification of the bacteria using advanced classification algorithms. The captured scatter patterns of the bacterial colony are automatically analyzed by the quantitative image processing software. Two major features are used for image analysis: the rotation-invariant feature (circularly symmetric patterns) and texture features (random and speckle patterns). The performance of the classifier is estimated using cross-validation ( ). The importance of quantitative classification software is needed to reduce human errors, which in turn can provide higher sensitivity and specificity than visual observation. Each 2-D scatter pattern is analyzed via Zernike moments and Haralick textures ( ). The former extracts features of circularly symmetric features from the scatter patterns, while the Haralick describes the texture of the scatter patterns. The combination of these two features results in hundreds of signature attributes from a single 2-D image which can be used as an orthogonal basis for the fingerprint library in the classification of scattering patterns of bacterial test samples. Once the training library is built, the sample under investigation can be compared against the fingerprint library that is already built and trained. The results are then reported in a matrix format. The diagonal numbers represent the expected correct classification rate (true positive and true negative) while the off-diagonal numbers show the missed classification (false positive and false negative). More details about the CV matrix are discussed in our relevant publications ( ; ).
Raman spectroscopy uses a laser to record the vibrational and rotational properties of molecules yielding a scattering signature referred to inelastic scattering ( ; ). Earlier attempts to detect microcolonies of clinically relevant bacterial pathogens of Staphylococcus aureus , Staphylococcus epidermidis , Escherichia coli and Enterococcus faecium directly from agar plate using a Raman microspectrometer equipped with an 830-nm titanium−sapphire laser at wavenumber 250 to 2150 cm -1 was moderately successful ( ). Later, Rosch et al. ( ) used lasers with three different wavelengths (785 nm; 633, and 514 nm) to differentiate colonies of test organisms ( Micrococcus luteus , Bacillus subtilis , and Pseudomonas fluorescens ) where chromophores produced by these organisms aided in spectral classification. However, bacterial viability was lost due to the destruction of bacterial cells during laser exposure, a major impediment to the isolation of viable cells after Raman spectroscopy ( ). Raman spectroscopy was also successfully used for the detection of colonies of clinically relevant Staphylococcus epidermidis, S. aureus and Escherichia coli strains on blood agar or Mueller-Hinton agar plates ( ; ). Most recently, Shen et al. ( ) reported a rapid fiber probe-based Raman (785 nm diode laser) technique for the classification and identification of 33 strains of 8 different species including Candida albicans , Staphylococcus epidermidis , S. aureus , Klebsiella pneumoniae , K. oxytoca , Escherichia coli , Enterococcus faecalis , E. faecium , and Acinetobacter baumannii on Luria–Bertani (LB) agar plates. Nevertheless, Raman spectroscopy continued to be an attractive on-plate microbial pathogen detection tool for foodborne ( ), and clinical relevant pathogens ( ; ). However, bacterial physiological state, growth phase and growth media can affect the spectral fingerprints thus these parameters should be controlled with care ( ; ).
Hyperspectral imaging (HSI) technology combines spectroscopy and imaging as a reliable nondestructive technique for bacterial colony counting, and detection and identification on various food, inert surfaces, or clinical specimens ( ; ; ; ; ). Lights with small wavelength bandwidth from visible to near-infrared (Vis-NIR) are often used to generate a complete spatiospectral map of a colony for pathogen detection and identification. Scientists at the US Department of Agriculture developed a hyperspectral imaging system with a spectral range from 400 to 1000 nm to detect and differentiate serovars of Shiga-toxin-producing Escherichia coli (STEC) pathogens on Rainbow agar with very high accuracy ( ; ). They also successfully used this platform to detect Campylobacter species ( ). Direct identification of colonies on culture plates is also highly important for clinical diagnostic applications. Arrigoni et al. ( ) applied HSI coupled with a classification algorithm to identify pathogens that are responsible for urinary tract infection, including Escherichia coli , Enterococcus faecalis , Staphylococcus aureus , Proteus mirabilis , Proteus vulgaris , Klebsiella pneumoniae , and Pseudomonas aeruginosa on blood agar plates. HSI was also used for the discrimination of colonies of three different bacterial cultures including E. coli , Listeria monocytogenes and Staphylococcus aureus for application in food safety ( ). This method employed a non-selective agar plate (tryptic soy agar, TSA) for colony identification. Likewise, using HSI and chemometric classification algorithms, Gu et al. ( ) differentially distinguished colonies of Escherichia coli, Staphylococcus aureus , and Salmonella enterica on three different nonselective agar plates. However, the drawback of using non-selective agar plates for HSI application is that the growth of commensal bacteria can interfere with target organism identification when testing with food samples, thus conventional broth culturing techniques using selective antimicrobial agents must be employed before testing food samples on agar plates. Near-infrared (NIR) HSI with multivariate data analysis was shown to be useful for the discrimination of colonies of Bacillus cereus , Escherichia coli , Salmonella enterica serovar Enteritidis, Staphylococcus aureus and S. epidermidis ( ). The application of HSI in a reflectance mode is highly useful for differentially distinguishing bacterial colonies from particulate foods on agar plate surfaces for food safety analysis ( ).
Mass-spectrometry As a routine microbiological laboratory practice, bacterial cells collected from a well-separated colony (to assure pure culture) from a Petri dish, are tested for their unique sugar or amino acid utilization patterns as an identifying tool ( ; ). Likewise, bacterial cells from colonies are also tested by using the matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry. A laser beam ionizes the sample matrix creating single protonated ions from analytes in the sample. Using acceleration at a stable potential, protonated ions are separated based on their mass-to-charge ratio. The time of flight measures the mass-to-charge ratio by the time it takes the ion to travel the length of the flight tube ( ). The spectral signatures are matched with the database for identification. This method is reliably used for pathogen detection from food and clinical samples ( ; ). Spectroscopy Inelastic scattering technologies such as near-infrared (NIR) ( ), Fourier transform Infrared (FT-IR) ( ), Raman ( ), and hyperspectral imaging ( ) have been used to identify bacterial pathogens that are obtained from isolated colonies from a Petri-dish. Often colony isolated bacterial cell suspensions are dispersed on appropriate substrates (for example, silicon wafer, CaF 2 ) and applied to above mentioned inelastic/vibrational spectroscopy for pathogen identification. Michael et al. ( ) applied HSI to identify several bacterial pathogens including Cronobacter sakazakii , Salmonella spp., Escherichia coli , Listeria monocytogenes and Staphylococcus aureus smeared on glass slides obtained from an isolated colony from agar plates. Likewise, Raman ( ; ; ; ; ; ) and FT-IR ( ; ; ; ) spectrocopies have been shown to be very promising diagnostic tools for detection, identification or antibiotic susceptibility testing of various pathogens obtained from isolated colonies. In addition, laser-induced breakdown spectroscopy (LIBS) was developed in response to the rapid identification of biothreat agents including pathogens or toxic gas. In this technique, the breakdown of the target analyte by a laser shot (1 ms) reaching a temperature of >10,000 K can generate plasma composed of ionic and atomic species ( ). Quantitative spectrochemical analyses of plasma allow rapid identification of a target analyte. LIBS have been used for the differentiation and classification of foodborne and clinically relevant microbial pathogens obtained from isolated colonies ( ; ; ). Molecular methods Molecular methods are also increasingly becoming integral to the pathogen detection regimen in agriculture, food, and medicine. Historically, the colony hybridization technique has been a quintessential tool for verification of the acquisition of a target gene(s) by host microbes, where colonies are transferred from a Petri dish to a membrane for hybridization with a pre-labeled nucleic acid probe ( ). Membrane-transferred colonies are also probed with antibodies (colony immunoblot) for the detection and identification of many pathogens including E. coli O157:H7 ( ), L. monocytogenes ( ; ), Helicobacter pylori ( ) and Campylobacter species ( ). Polymerase chain reaction (PCR) assay has been also applied to colonies (Colony PCR) for the detection of various bacterial ( ; ) and fungal ( ; ) pathogens. Recently, whole genome sequencing (WGS) of isolated colonies is also finding widespread application in food safety and clinical medicine ( ; ; ; ).
As a routine microbiological laboratory practice, bacterial cells collected from a well-separated colony (to assure pure culture) from a Petri dish, are tested for their unique sugar or amino acid utilization patterns as an identifying tool ( ; ). Likewise, bacterial cells from colonies are also tested by using the matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry. A laser beam ionizes the sample matrix creating single protonated ions from analytes in the sample. Using acceleration at a stable potential, protonated ions are separated based on their mass-to-charge ratio. The time of flight measures the mass-to-charge ratio by the time it takes the ion to travel the length of the flight tube ( ). The spectral signatures are matched with the database for identification. This method is reliably used for pathogen detection from food and clinical samples ( ; ).
Inelastic scattering technologies such as near-infrared (NIR) ( ), Fourier transform Infrared (FT-IR) ( ), Raman ( ), and hyperspectral imaging ( ) have been used to identify bacterial pathogens that are obtained from isolated colonies from a Petri-dish. Often colony isolated bacterial cell suspensions are dispersed on appropriate substrates (for example, silicon wafer, CaF 2 ) and applied to above mentioned inelastic/vibrational spectroscopy for pathogen identification. Michael et al. ( ) applied HSI to identify several bacterial pathogens including Cronobacter sakazakii , Salmonella spp., Escherichia coli , Listeria monocytogenes and Staphylococcus aureus smeared on glass slides obtained from an isolated colony from agar plates. Likewise, Raman ( ; ; ; ; ; ) and FT-IR ( ; ; ; ) spectrocopies have been shown to be very promising diagnostic tools for detection, identification or antibiotic susceptibility testing of various pathogens obtained from isolated colonies. In addition, laser-induced breakdown spectroscopy (LIBS) was developed in response to the rapid identification of biothreat agents including pathogens or toxic gas. In this technique, the breakdown of the target analyte by a laser shot (1 ms) reaching a temperature of >10,000 K can generate plasma composed of ionic and atomic species ( ). Quantitative spectrochemical analyses of plasma allow rapid identification of a target analyte. LIBS have been used for the differentiation and classification of foodborne and clinically relevant microbial pathogens obtained from isolated colonies ( ; ; ).
Molecular methods are also increasingly becoming integral to the pathogen detection regimen in agriculture, food, and medicine. Historically, the colony hybridization technique has been a quintessential tool for verification of the acquisition of a target gene(s) by host microbes, where colonies are transferred from a Petri dish to a membrane for hybridization with a pre-labeled nucleic acid probe ( ). Membrane-transferred colonies are also probed with antibodies (colony immunoblot) for the detection and identification of many pathogens including E. coli O157:H7 ( ), L. monocytogenes ( ; ), Helicobacter pylori ( ) and Campylobacter species ( ). Polymerase chain reaction (PCR) assay has been also applied to colonies (Colony PCR) for the detection of various bacterial ( ; ) and fungal ( ; ) pathogens. Recently, whole genome sequencing (WGS) of isolated colonies is also finding widespread application in food safety and clinical medicine ( ; ; ; ).
According to the official detection scheme for a pathogen from food or environmental samples, initial liquid culturing in primary and/or secondary enrichment broths followed by plating of the enriched samples on various selective agar plates for isolation of individual colonies are practiced for presumptive identification ( ). Enrichment broths help resuscitate stressed or injured cells to expand and antimicrobial selective agents in broths help reduce background microflora ( ). To expedite the diagnostic workflow for faster results (less than 24 h), enrichment broths are routinely tested by a pathogen-specific PCR assay or antibody-based lateral flow immunochromatographic assays ( ; ; ). Bacteriophage-based pathogen detection is also gaining significant interest among microbiologists ( ). However; most clinical laboratory diagnostic approaches rely on isolated individual colonies, which are achieved by direct plating of clinical samples on Petri plates. Depending on the species of the bacteria to be isolated, differential or selective agar media are used to obtain a colony with typical phenotypic characteristics (color, texture, diameter). The bacterial growth rate on the Petri plate varies widely and may typically require 12 to 48 h to obtain a colony with a 1-2 mm diameter depending on the genus/species, the selective agents used in the agar media, and the physiological state of the bacteria. While extremely slow-growing organism (ex. Mycobacterium species) requires several days. Application of BARDOT can significantly shorten the liquid enrichment and on-plate growth time yielding results much earlier than the conventional culturing method ( ). As mentioned above, colony isolation is essential for performing more comprehensive tests such as mass-spec analysis, whole-genome sequencing, pathogenicity testing, antibiotic susceptibility analysis, sensitivity to various food preservatives or chemical or biological sanitizers, and other physiological parameters ( ). On the other hand, BARDOT can be applied directly to the colonies growing on the Petri dishes for interrogation without any physical contact with the colony thus preserving colony integrity and cell viability. The cellular organization, extracellular matrix, phenotype variation, refractive indices, and size of cells within the confinement of a colony are attributed to producing differential scatter signatures ( ; ). Changes in media formulations can also alter scatter signature patterns implying nutrient utilization and metabolic activity are directly linked to a bacterial phenotype which can be used for further validation of cultural identity ( ; ) ( ). Further molecular, immunological or biochemical testing of colonies can be done to validate BARDOT results ( ). Therefore, the Petri plate is considered to be an irreplaceable tool for both clinical diagnosis and food testing. We will review how BARDOT is utilized for the detection and identification of major pathogens of interest in public health and food safety. BARDOT has been used to interrogate colonies on Petri plates of various foodborne and clinically relevant bacterial pathogens, such as Listeria monocytogenes ( ; ; ), Vibrio species ( ), Escherichia coli O157:H7 ( ; ), Salmonella enterica ( ; ; ), Bacillus spp. ( ), Staphylococcus spp. ( ), and Campylobacter spp. ( ) and the members of the Enterobacteriaceae family ( ). Hence, BARDOT has been considered a non-invasive, non-destructive, and reagent-free detection platform for pathogens of food and clinical relevance. In addition, BARDOT showed utility in differentiating mutant strains deficient in virulence-gene in L. monocytogenes ( ) or antibiotics-induced stress response by bacterial pathogens ( ; ). The optical forward scattering technique was also evaluated for application in clinical microbiology for the detection of colonies of E. coli , S. aureus , Proteus mirabilis , Yersinia enterocolitica , and Salmonella Typhimurium in an automated pathogen identification platform with the variable as well as fixed incubation time ( ). Elastic light scattering device has also been reported by Kitaoka et al. ( ) demonstrating the capacity to detect and identify various microorganisms, including Bacillus subtilis , S. aureus , and Saccharomyces cerevisiae . Using the BISLD system with improved Fresnel diffraction pattern analysis, Buzalewicz et al. ( ) successfully detected Candida albicans and several clinically relevant bacterial species including Citrobacter freundii, Enterobacter cloacae, Enterococcus faecalis, E. faecium, Escherichia coli, Klebsiella oxytoca, K. pneumoniae, Pseudomonas aeruginosa, P. putida, Serratia marcescens, and Staphylococcus aureus with 97-100% accuracy. Listeria species Listeria monocytogenes is an opportunistic invasive pathogen and is the primary pathogenic member of the genus which has over 27 species. Food is the primary vehicle for transmission. L. monocytogenes infects immunocompromised individuals such as the elderly, neonates, and pregnant women resulting in premature birth or stillbirth and the case fatality rate is about 19%. Depending on the official scheme as outlined by USDA-FSIS, FDA, or ISO methods ( ) for Listeria from food or environmental samples, detection involves culturing in primary (ex. UVM, University of Vermont Medium) and/or secondary enrichment broths (ex. Fraser broth) followed by plating on various selective agar plates (ex. Modified Oxford agar, MOX) for isolation of individual colonies exhibiting typical colony morphology that aid in presumptive identification. PCR and lateral flow immunoassays are used for the detection of Listeria from enrichment broths for faster results and to bypass the lengthy plating and colony isolation steps ( ; ; ; ). For the application of BARDOT in Listeria detection, the laser beam was directly applied to the microcolonies (about 1 mm diameter) on agar plates (MOX prepared without ferric ammonium citrate to prevent the black precipitate formation at the center of the colony or BHI plates) to demonstrate the feasibility of differentiating species of Listeria within the genus based on their scatter patterns ( ; ). Image analysis software using Zernike moment invariants and principal component analysis subjected 91–100% accuracy in detecting different species of Listeria . Diffraction theory was used to model the scattering patterns to explain the appearance of radial spokes and the rings seen in the scattering images of L. monocytogenes ( and ). In another study, Kim et al. ( ) employed BARDOT as a complementary tool to differentiate between certain species of Listeria sensu stricto and Listeria sensu lato ( ) using Listeria species-specific PCR assays. PCR assay amplified a housekeeping gene ( lmo1634 ) encoding acetaldehyde alcohol dehydrogenase (AdhE), also known as Listeria adhesion protein (LAP) ( ). Both PCR and BARDOT were complementary in their abilities to detect Listeria from inoculated food samples that contained mixed Listeria cultures with a detection limit of about 10 4 CFU/mL. In two separate studies, Koo et al. ( ) and Mendonca et al. ( ) used BARDOT to confirm the presence of L. monocytogenes in food samples. They used antibody- or receptor-coated magnetic beads to capture L. monocytogenes from enriched food samples before plating them onto selective agar plates. Colonies from Petri plates were analyzed by BARDOT for confirmation. More recently, Zhu et al. ( ) used BARDOT to detect L. monocytogenes from inoculated milk samples. Application of BARDOT to detect L. monocytogenes from experimentally infected mouse tissues was also demonstrated ( ) where, L. monocytogenes was successfully detected in liver, spleen, and intestinal chymus demonstrating the feasibility of BARDOT in potential clinical diagnostics. Salmonella enterica Salmonella enterica causes typhoid fever and gastroenteritis. It is estimated that each year in the United States, among 9.4 million foodborne illnesses, gastroenteritis causing non-typhoidal Salmonella (NTS) alone is responsible for 1 million illnesses, 19,581 hospitalizations, and 378 deaths ( ). Salmonella is a robust organism and can survive at low pH, high salt, desiccation, and thermal processing, making it imperative for food companies to develop comprehensive food safety programs to reduce contamination and prevent contaminated products from reaching consumers. BARDOT was applied to investigate its ability to selectively detect and identify NTS from the top twenty frequently reported serovars of Salmonella enterica ( ). The initial study involved the capacity of BARDOT to classify colonies of six Salmonella serovars grown on brain heart infusion (BHI), brilliant green (BG), xylose lysine deoxycholate (XLD), and xylose lysine tergitol 4 (XLT4) agar plates. Cultures on XLT generated highly accurate discriminatory (95.9%) scatter signatures among the S. enterica serovars ( ). Later, BARDOT yielded classification precision of 88-100% when tested with 36 serovars (top 20 plus 16 miscellaneous serovars), which showed a strong correlation with pulsed-field gel electrophoresis (PFGE)-based genetic fingerprints. For testing of food samples for Salmonella using BARDOT, a sequential enrichment in nonselective (buffered peptone water) and selective enrichment (modified Rappaport Vassiliadis) broths for 4 h each followed by growth on XLT4 (~16 h) was used. BARDOT delivered results within 24 h with a detection sensitivity of 1.2×10 2 CFU/30 g, much faster than the USDA-FSIS method, which requires about 72 h. Genetic analysis (16S rRNA gene sequencing and PFGE) also confirmed BARDOT results. In another study, a combination of a fiber optic immunosensor ( ) and BARDOT was used to detect Salmonella from naturally contaminated poultry samples ( ). Poultry samples were sequentially enriched in primary and secondary enrichment broths for 4 h each before plating on selective agar plates (XLT4). The scatter signatures of colonies on XLT4 generated by BARDOT were matched with the image library for identification of Salmonella in less than 24 h. While the fiber optic sensor was applied directly to the broth sample from secondary enrichment thus results were obtained in less than 12 h. Though the fiberoptic sensor provided faster results, the BARDOT-based detection approach provides an opportunity to obtain pure isolated colonies that can be further used for antibiotic sensitivity testing, whole-genome sequencing, source tracking and other studies including pathogenicity assays. Bacillus species The genus Bacillus comprises pathogens, nonpathogens, and industrially relevant beneficial bacteria. Bacilli are spore-forming Gram-positive bacteria. Singh et al. ( ), used BARDOT to screen and differentiate colonies of Bacillus species on Petri plates containing phenol red mannitol (PRM) agar ( ). Colony morphology is highly diverse among the species of bacillus producing a flat surface with rough topography therefore the colony scatter patterns consisting of speckles are unique and do not overlap with scatter patterns from other bacterial species ( ). Initially, a colony scatter image library was created using a total of 265 Bacillus and non- Bacillus isolates from our collection. Cross-validation experiments demonstrated that all Bacillus species ( n = 118) gave a positive predictive value (PPV) above 90% while non- Bacillus spp. showed a PPV of <0.5%. Spiked baby formula and cheese samples, or naturally contaminated bovine unpasteurized milk samples were surface-plated on PRM and the microcolonies were scanned by BARDOT in 7-16 h to capture colony scatter signatures. BARDOT-identified Bacillus cultures were further verified by using PCR and 16S rRNA gene sequencing to provide high accuracy ( ). Shiga toxin-producing E. coli Escherichia coli is one of the ubiquitous Gram-negative bacteria that resides in the intestine of animals and humans. A majority of E. coli are nonpathogenic while a small subset is pathogenic and causes diseases including gastroenteritis, urinary tract infection, kidney disease, and central nervous system infection. Based on the nature of the gastrointestinal infection, E. coli is grouped into 5 major pathotypes; enterotoxigenic E. coli (ETEC), enteroaggregative E. coli (EAgEC), enteroinvasive E. coli (EIEC), enteropathogenic E. coli (EPEC), and enterohaemorrhagic E. coli (EHEC), which is a subset of broadly defined Shiga toxin-producing E. coli (STEC) ( ; ). STEC has emerged as an important foodborne pathogen, among which seven serogroups (O26, O45, O103, O111, O121, O145, O157) are most frequently implicated in human infection. Tang et al. ( ) used BARDOT to differentiate STEC serovars. The goal was to determine if BARDOT can be used to rapidly identify the colonies of STEC serogroups on selective agar plates. Multiple selective/differential agar media on the Petri plate was evaluated that including sorbitol MacConkey (SMAC), Rainbow ® Agar O157, BBL™ CHROMagarO157, and R&F ® E. coli O157:H7, and BHI ( ). Colony scatter signatures obtained after 10-12 h growth on both SMAC and Rainbow produced results that successfully differentiated all seven serovars (O26, O45, O103, O111, O121, O145, O157) of STEC with greater than 90% accuracy. Colonies of E. coli O157 and O26 serovars in a mixed culture and inoculated food samples (lettuce and ground beef) were accurately identified by BARDOT requiring a sample-to-result in less than 24 h ( ). Vibrio species The genus Vibrio consists of three major human pathogens including Vibrio cholerae , V. parahaemolyticus, V. vulnificus that are associated with water- and seafood-related outbreaks worldwide ( ). V. cholerae is responsible for cholera, a severe diarrheal disease. V. parahaemolyticus and V. vulnificus cause gastroenteritis but may also cause fatal septicemic disease and are primarily transmitted via seafood. Their primary habitat is an aquatic and marine environment with a high preference for brackish (salt and fresh water mix) water. These halophiles (salt-loving) organisms also undergo a viable but nonculturable state under stressful conditions thus their culturing becomes very difficult. Culture enrichment in alkaline peptone water can help resuscitate a culturable state. For isolation of individual colonies, the enriched samples can be plated on BHI containing 1% NaCl or on selective Thiosulphate Citrate Bile Salts Sucrose (TCBS) agar. BARDOT was applied for Vibrio detection on Petri plates after 12 h growth at 30°C and it successfully detected V. cholerae , V. parahaemolyticus , and V. vulnificus present in oyster or water samples in 18 h even in the presence of other vibrios or other bacteria, indicating the suitability of the sensor as a powerful screening tool for pathogens on agar plates ( ). Staphylococcus aureus Staphylococcus aureus is a major pathogen responsible for nosocomial infections and foodborne illnesses ( ). Common habitat for S. aureus is nares and skin. BARDOT was used for rapid colony screening and detection of Staphylococcus on an agar plate and to differentiate these colonies from non- Staphylococcus spp. ( ). Phenol red mannitol agar (PRMA) was used for building the Staphylococcus species scatter image libraries since this medium generated distinguishing scatter signatures when compared with other species. The scatter image library for Staphylococcus species gave a high positive predictive value (PPV 87.5–100%) when tested against known laboratory strains of Staphylococcus spp., while the PPV against non- Staphylococcus spp. was 0–38%. BARDOT detected S. aureus with 80–100% PPV from naturally contaminated cow milk and ready-to-eat chicken salad samples and the results were validated with PCR and 16S rRNA gene sequencing ( ). Enterobacteriaceae BARDOT was also used to study colonies formed by the members of the Enterobacteriaceae family, which comprises pathogens and commensals and has a significant impact on food safety, clinical microbiology, and public health. Their presence in foods/water indicates potential contamination with pathogens. Enterobacteriaceae (EB) detection has been used as an indicator for assessing the safety of food products or water. Various selective chromogenic media are used for quantification and analysis of colonies of EB; however, many produce similar chromogenic by-products thus they cannot be accurately visually identified on the Petri plate. Singh and Bhunia ( ) applied BARDOT to screen colonies of the Enterobacteriaceae family including Klebsiella, Enterobacter, Citrobacter, Serratia, Proteus, Morganella , and Providencia cultured on CHROMagar™ Orientation medium ( ). A scatter image library (1683 scatter images) was made that contains colony scatter signatures of 36 isolates representing 12 genera and 15 species. This library helped BARDOT-based detection of colonies of members of Enterobacteriaceae and non- Enterobacteriaceae family ( Pseudomonas aeruginosa , Acinetobacter spp., and Staphylococcus aureus ) with high accuracy (83-100%) in 10-22 h or even before visible production of chromogens. BARDOT as a tool to study pathogenesis, physiology, and community diversity BARDOT was used in the screening of mutant strains that are deficient in several virulence genes essential for pathogenicity ( ). During microbial pathogenesis studies, virulence-encoding genes are routinely disrupted by deletion or insertion to create mutant strains. Screening mutant strains is a laborious process involving plating on growth media containing antibiotics marker, replica plating, colony hybridization, DNA isolation, and PCR or immunoassays. BARDOT was used to screen virulence-gene-associated mutant colonies during microbial pathogenesis, co-infection, and genetic manipulation studies in L. monocytogenes . BARDOT generated differential scatter patterns in L. monocytogenes , deficient in Listeria adhesion protein ( lap - ), Internalin A (Δ inlA ), and an accessory secretory protein (Δ secA2 ). Furthermore, the mutant strains complemented with respective genes were also able to restore the scatter signature to that of the WT ( ). BARDOT was also useful in differential counting of mutant strains in the presence of WT strain in a co-infection experiment. These data demonstrate that BARDOT can be used as a label-free tool to aid researchers in screening virulence-gene-associated mutant colonies during microbial pathogenesis and genetic manipulation studies. In another study, Singh et al. ( ) also investigated the streptomycin-induced stress response in Salmonella enterica serovars with BARDOT. Streptomycin-sensitive or streptomycin-resistant Salmonella serovars were exposed to various levels of streptomycin and grown on Petri plates and the colonies were screened by BARDOT to assess their stress responses and colony scatter signatures. A substantial qualitative and quantitative difference in the scatter signatures was observed for colonies that were grown in the presence of streptomycin than the colonies grown in the absence of antibiotics. Levels of a stress response protein, GroEL, were increased in the colony confirmed by mass-spec, quantitative RT-PCR, and immunoassays that were implicated to contribute to the differential scatter patterns. The study highlights the suitability of the BARDOT to investigate stress response in bacteria in conjunction with molecular or other analytical methods. In another study, Zhu et al. ( ) used BARDOT to study the effect of tunicamycin, a cell wall teichoic acid (WTA) synthesis inhibitor on colony morphology and colony scatter patterns. WTA is a major component of the cell wall of Gram-positive bacteria and plays a significant role in physiology, biofilm formation, and pathogenesis ( ). BARDOT was evaluated for its ability to simultaneously detect colonies of three pathogens ( L. monocytogenes , Salmonella enterica , and E. coli ) from the same test sample, if present together on the same Petri plate ( ). Test samples were first enriched in a multi-pathogen enrichment broth, SEL ( Salmonella , Escherichia , Listeria ) ( ) before plating onto SEL agar for BARDOT-based colony identification. The BARDOT sensor successfully detected Salmonella , Shiga-toxin-producing E. coli , and Listeria on the SEL agar plate with greater than 90% accuracy within 29–40 h demonstrating its simultaneous multi-pathogen detection potential ( ) ( ). BARDOT coupled with 16s RNA sequencing was instrumental in identifying thermostable bacterial colonies in fluid milk subjected to a novel low temperature−short time (LTST) process for pasteurization ( ).
species Listeria monocytogenes is an opportunistic invasive pathogen and is the primary pathogenic member of the genus which has over 27 species. Food is the primary vehicle for transmission. L. monocytogenes infects immunocompromised individuals such as the elderly, neonates, and pregnant women resulting in premature birth or stillbirth and the case fatality rate is about 19%. Depending on the official scheme as outlined by USDA-FSIS, FDA, or ISO methods ( ) for Listeria from food or environmental samples, detection involves culturing in primary (ex. UVM, University of Vermont Medium) and/or secondary enrichment broths (ex. Fraser broth) followed by plating on various selective agar plates (ex. Modified Oxford agar, MOX) for isolation of individual colonies exhibiting typical colony morphology that aid in presumptive identification. PCR and lateral flow immunoassays are used for the detection of Listeria from enrichment broths for faster results and to bypass the lengthy plating and colony isolation steps ( ; ; ; ). For the application of BARDOT in Listeria detection, the laser beam was directly applied to the microcolonies (about 1 mm diameter) on agar plates (MOX prepared without ferric ammonium citrate to prevent the black precipitate formation at the center of the colony or BHI plates) to demonstrate the feasibility of differentiating species of Listeria within the genus based on their scatter patterns ( ; ). Image analysis software using Zernike moment invariants and principal component analysis subjected 91–100% accuracy in detecting different species of Listeria . Diffraction theory was used to model the scattering patterns to explain the appearance of radial spokes and the rings seen in the scattering images of L. monocytogenes ( and ). In another study, Kim et al. ( ) employed BARDOT as a complementary tool to differentiate between certain species of Listeria sensu stricto and Listeria sensu lato ( ) using Listeria species-specific PCR assays. PCR assay amplified a housekeeping gene ( lmo1634 ) encoding acetaldehyde alcohol dehydrogenase (AdhE), also known as Listeria adhesion protein (LAP) ( ). Both PCR and BARDOT were complementary in their abilities to detect Listeria from inoculated food samples that contained mixed Listeria cultures with a detection limit of about 10 4 CFU/mL. In two separate studies, Koo et al. ( ) and Mendonca et al. ( ) used BARDOT to confirm the presence of L. monocytogenes in food samples. They used antibody- or receptor-coated magnetic beads to capture L. monocytogenes from enriched food samples before plating them onto selective agar plates. Colonies from Petri plates were analyzed by BARDOT for confirmation. More recently, Zhu et al. ( ) used BARDOT to detect L. monocytogenes from inoculated milk samples. Application of BARDOT to detect L. monocytogenes from experimentally infected mouse tissues was also demonstrated ( ) where, L. monocytogenes was successfully detected in liver, spleen, and intestinal chymus demonstrating the feasibility of BARDOT in potential clinical diagnostics.
Salmonella enterica causes typhoid fever and gastroenteritis. It is estimated that each year in the United States, among 9.4 million foodborne illnesses, gastroenteritis causing non-typhoidal Salmonella (NTS) alone is responsible for 1 million illnesses, 19,581 hospitalizations, and 378 deaths ( ). Salmonella is a robust organism and can survive at low pH, high salt, desiccation, and thermal processing, making it imperative for food companies to develop comprehensive food safety programs to reduce contamination and prevent contaminated products from reaching consumers. BARDOT was applied to investigate its ability to selectively detect and identify NTS from the top twenty frequently reported serovars of Salmonella enterica ( ). The initial study involved the capacity of BARDOT to classify colonies of six Salmonella serovars grown on brain heart infusion (BHI), brilliant green (BG), xylose lysine deoxycholate (XLD), and xylose lysine tergitol 4 (XLT4) agar plates. Cultures on XLT generated highly accurate discriminatory (95.9%) scatter signatures among the S. enterica serovars ( ). Later, BARDOT yielded classification precision of 88-100% when tested with 36 serovars (top 20 plus 16 miscellaneous serovars), which showed a strong correlation with pulsed-field gel electrophoresis (PFGE)-based genetic fingerprints. For testing of food samples for Salmonella using BARDOT, a sequential enrichment in nonselective (buffered peptone water) and selective enrichment (modified Rappaport Vassiliadis) broths for 4 h each followed by growth on XLT4 (~16 h) was used. BARDOT delivered results within 24 h with a detection sensitivity of 1.2×10 2 CFU/30 g, much faster than the USDA-FSIS method, which requires about 72 h. Genetic analysis (16S rRNA gene sequencing and PFGE) also confirmed BARDOT results. In another study, a combination of a fiber optic immunosensor ( ) and BARDOT was used to detect Salmonella from naturally contaminated poultry samples ( ). Poultry samples were sequentially enriched in primary and secondary enrichment broths for 4 h each before plating on selective agar plates (XLT4). The scatter signatures of colonies on XLT4 generated by BARDOT were matched with the image library for identification of Salmonella in less than 24 h. While the fiber optic sensor was applied directly to the broth sample from secondary enrichment thus results were obtained in less than 12 h. Though the fiberoptic sensor provided faster results, the BARDOT-based detection approach provides an opportunity to obtain pure isolated colonies that can be further used for antibiotic sensitivity testing, whole-genome sequencing, source tracking and other studies including pathogenicity assays.
species The genus Bacillus comprises pathogens, nonpathogens, and industrially relevant beneficial bacteria. Bacilli are spore-forming Gram-positive bacteria. Singh et al. ( ), used BARDOT to screen and differentiate colonies of Bacillus species on Petri plates containing phenol red mannitol (PRM) agar ( ). Colony morphology is highly diverse among the species of bacillus producing a flat surface with rough topography therefore the colony scatter patterns consisting of speckles are unique and do not overlap with scatter patterns from other bacterial species ( ). Initially, a colony scatter image library was created using a total of 265 Bacillus and non- Bacillus isolates from our collection. Cross-validation experiments demonstrated that all Bacillus species ( n = 118) gave a positive predictive value (PPV) above 90% while non- Bacillus spp. showed a PPV of <0.5%. Spiked baby formula and cheese samples, or naturally contaminated bovine unpasteurized milk samples were surface-plated on PRM and the microcolonies were scanned by BARDOT in 7-16 h to capture colony scatter signatures. BARDOT-identified Bacillus cultures were further verified by using PCR and 16S rRNA gene sequencing to provide high accuracy ( ).
E. coli Escherichia coli is one of the ubiquitous Gram-negative bacteria that resides in the intestine of animals and humans. A majority of E. coli are nonpathogenic while a small subset is pathogenic and causes diseases including gastroenteritis, urinary tract infection, kidney disease, and central nervous system infection. Based on the nature of the gastrointestinal infection, E. coli is grouped into 5 major pathotypes; enterotoxigenic E. coli (ETEC), enteroaggregative E. coli (EAgEC), enteroinvasive E. coli (EIEC), enteropathogenic E. coli (EPEC), and enterohaemorrhagic E. coli (EHEC), which is a subset of broadly defined Shiga toxin-producing E. coli (STEC) ( ; ). STEC has emerged as an important foodborne pathogen, among which seven serogroups (O26, O45, O103, O111, O121, O145, O157) are most frequently implicated in human infection. Tang et al. ( ) used BARDOT to differentiate STEC serovars. The goal was to determine if BARDOT can be used to rapidly identify the colonies of STEC serogroups on selective agar plates. Multiple selective/differential agar media on the Petri plate was evaluated that including sorbitol MacConkey (SMAC), Rainbow ® Agar O157, BBL™ CHROMagarO157, and R&F ® E. coli O157:H7, and BHI ( ). Colony scatter signatures obtained after 10-12 h growth on both SMAC and Rainbow produced results that successfully differentiated all seven serovars (O26, O45, O103, O111, O121, O145, O157) of STEC with greater than 90% accuracy. Colonies of E. coli O157 and O26 serovars in a mixed culture and inoculated food samples (lettuce and ground beef) were accurately identified by BARDOT requiring a sample-to-result in less than 24 h ( ).
species The genus Vibrio consists of three major human pathogens including Vibrio cholerae , V. parahaemolyticus, V. vulnificus that are associated with water- and seafood-related outbreaks worldwide ( ). V. cholerae is responsible for cholera, a severe diarrheal disease. V. parahaemolyticus and V. vulnificus cause gastroenteritis but may also cause fatal septicemic disease and are primarily transmitted via seafood. Their primary habitat is an aquatic and marine environment with a high preference for brackish (salt and fresh water mix) water. These halophiles (salt-loving) organisms also undergo a viable but nonculturable state under stressful conditions thus their culturing becomes very difficult. Culture enrichment in alkaline peptone water can help resuscitate a culturable state. For isolation of individual colonies, the enriched samples can be plated on BHI containing 1% NaCl or on selective Thiosulphate Citrate Bile Salts Sucrose (TCBS) agar. BARDOT was applied for Vibrio detection on Petri plates after 12 h growth at 30°C and it successfully detected V. cholerae , V. parahaemolyticus , and V. vulnificus present in oyster or water samples in 18 h even in the presence of other vibrios or other bacteria, indicating the suitability of the sensor as a powerful screening tool for pathogens on agar plates ( ).
Staphylococcus aureus is a major pathogen responsible for nosocomial infections and foodborne illnesses ( ). Common habitat for S. aureus is nares and skin. BARDOT was used for rapid colony screening and detection of Staphylococcus on an agar plate and to differentiate these colonies from non- Staphylococcus spp. ( ). Phenol red mannitol agar (PRMA) was used for building the Staphylococcus species scatter image libraries since this medium generated distinguishing scatter signatures when compared with other species. The scatter image library for Staphylococcus species gave a high positive predictive value (PPV 87.5–100%) when tested against known laboratory strains of Staphylococcus spp., while the PPV against non- Staphylococcus spp. was 0–38%. BARDOT detected S. aureus with 80–100% PPV from naturally contaminated cow milk and ready-to-eat chicken salad samples and the results were validated with PCR and 16S rRNA gene sequencing ( ).
BARDOT was also used to study colonies formed by the members of the Enterobacteriaceae family, which comprises pathogens and commensals and has a significant impact on food safety, clinical microbiology, and public health. Their presence in foods/water indicates potential contamination with pathogens. Enterobacteriaceae (EB) detection has been used as an indicator for assessing the safety of food products or water. Various selective chromogenic media are used for quantification and analysis of colonies of EB; however, many produce similar chromogenic by-products thus they cannot be accurately visually identified on the Petri plate. Singh and Bhunia ( ) applied BARDOT to screen colonies of the Enterobacteriaceae family including Klebsiella, Enterobacter, Citrobacter, Serratia, Proteus, Morganella , and Providencia cultured on CHROMagar™ Orientation medium ( ). A scatter image library (1683 scatter images) was made that contains colony scatter signatures of 36 isolates representing 12 genera and 15 species. This library helped BARDOT-based detection of colonies of members of Enterobacteriaceae and non- Enterobacteriaceae family ( Pseudomonas aeruginosa , Acinetobacter spp., and Staphylococcus aureus ) with high accuracy (83-100%) in 10-22 h or even before visible production of chromogens.
BARDOT was used in the screening of mutant strains that are deficient in several virulence genes essential for pathogenicity ( ). During microbial pathogenesis studies, virulence-encoding genes are routinely disrupted by deletion or insertion to create mutant strains. Screening mutant strains is a laborious process involving plating on growth media containing antibiotics marker, replica plating, colony hybridization, DNA isolation, and PCR or immunoassays. BARDOT was used to screen virulence-gene-associated mutant colonies during microbial pathogenesis, co-infection, and genetic manipulation studies in L. monocytogenes . BARDOT generated differential scatter patterns in L. monocytogenes , deficient in Listeria adhesion protein ( lap - ), Internalin A (Δ inlA ), and an accessory secretory protein (Δ secA2 ). Furthermore, the mutant strains complemented with respective genes were also able to restore the scatter signature to that of the WT ( ). BARDOT was also useful in differential counting of mutant strains in the presence of WT strain in a co-infection experiment. These data demonstrate that BARDOT can be used as a label-free tool to aid researchers in screening virulence-gene-associated mutant colonies during microbial pathogenesis and genetic manipulation studies. In another study, Singh et al. ( ) also investigated the streptomycin-induced stress response in Salmonella enterica serovars with BARDOT. Streptomycin-sensitive or streptomycin-resistant Salmonella serovars were exposed to various levels of streptomycin and grown on Petri plates and the colonies were screened by BARDOT to assess their stress responses and colony scatter signatures. A substantial qualitative and quantitative difference in the scatter signatures was observed for colonies that were grown in the presence of streptomycin than the colonies grown in the absence of antibiotics. Levels of a stress response protein, GroEL, were increased in the colony confirmed by mass-spec, quantitative RT-PCR, and immunoassays that were implicated to contribute to the differential scatter patterns. The study highlights the suitability of the BARDOT to investigate stress response in bacteria in conjunction with molecular or other analytical methods. In another study, Zhu et al. ( ) used BARDOT to study the effect of tunicamycin, a cell wall teichoic acid (WTA) synthesis inhibitor on colony morphology and colony scatter patterns. WTA is a major component of the cell wall of Gram-positive bacteria and plays a significant role in physiology, biofilm formation, and pathogenesis ( ). BARDOT was evaluated for its ability to simultaneously detect colonies of three pathogens ( L. monocytogenes , Salmonella enterica , and E. coli ) from the same test sample, if present together on the same Petri plate ( ). Test samples were first enriched in a multi-pathogen enrichment broth, SEL ( Salmonella , Escherichia , Listeria ) ( ) before plating onto SEL agar for BARDOT-based colony identification. The BARDOT sensor successfully detected Salmonella , Shiga-toxin-producing E. coli , and Listeria on the SEL agar plate with greater than 90% accuracy within 29–40 h demonstrating its simultaneous multi-pathogen detection potential ( ) ( ). BARDOT coupled with 16s RNA sequencing was instrumental in identifying thermostable bacterial colonies in fluid milk subjected to a novel low temperature−short time (LTST) process for pasteurization ( ).
The optical forward scattering system, BARDOT for colony detection/identification described here operates based on the propagation of light through the center of the colony and the solid agar media on the Petri plate. Therefore, microbial colonies must be translucent to allow laser propagation for the generation of forward scattering patterns, which is observed for the most of bacterial pathogens tested using BARDOT. In contrast, colonies produced by yeast and mold form opaque structures and thus are unsuitable for interrogation by BARDOT (unpublished observation). Likewise, agar media that are opaque, such as blood agar plate, Baird Parker agar, and media with certain chromogens do not permit laser penetration thus unsuitable for use ( ; ). In such situations, a laser-based backscattering device, hyperspectral imaging ( ), or Raman scattering ( ) system can be used to overcome the limitations of forward scattering platforms, such as BARDOT. In addition, variation in agar concentration (above or below the recommended concentration of 1.5% w/w), and media formulations can affect the scatter signatures produced by both elastic or inelastic scatterometers, thus these parameters must be controlled to acquire reproducible scatter signatures ( ; ; ; ; ). BARDOT-based detection time varies widely since it depends on the rate of bacterial growth to achieve the desired diameter range (0. 7 mm–1.2 mm) that can be detected. Thus fast-growing bacteria can be detected earlier than slow growers. In most cases, as discussed above, a majority of the pathogens were detected in less than 24 h starting with the test sample, providing a culture-based rapid method. BARDOT has been found useful for the detection of physiologically stressed or injured cells, provided a brief resuscitation/enrichment step is included before plating on selective or differential agar plates ( ; ; ). However, the BARDOT-based method may not be useful for inherently naturally slow growers, such as Mycobacterium species, if one is looking to obtain fast results from colony fingerprints. We also attempted to shorten the BARDOT-based assay time by interrogating microcolonies (0.1–0.2 mm diameter) of three different genera ( Escherichia , Salmonella , and Listeria incubated for 7, 9, and 12 h, respectively to achieve 0.1 – 0.2 mm diameter colonies) produced scatter signatures that could be used for differential diagnosis at the earlier time points ( ). Similarly, Marcoux et al. ( ) reported similar success in discriminating microcolonies of several Gram-negative bacteria after 6 h of incubation. Overall, microcolony (< 0.2 mm) detection by using scattering technology faces many challenges; (i) precise colony location on the plate due to smaller size, (ii) too small to be differentiated from particulate foods or samples on the plate, and most importantly, (iii) incomplete metabolic activity of growing cells within a microcolony may not produce adequate by-products that can yield robust features to differentiate them from closely related genera or species. Since the colony scatter patterns of a bacterium depend on a specific growth media, thus for identification, the pathogen and the media-specific scatter image library must be developed for the identification of the target pathogen. Since there are much pathogen-specific selective media available commercially and many more continued to be developed, one must generate an image library database for each medium making the assay development cumbersome and intensive. Most importantly, an image library must be built using bacterial strains that are typed strains procured from authentic and reliable sources (for example, American Type Culture Collection) or genetically validated. In addition, for improved confidence in BARDOT-based identification, the scatter image of the target pathogen could be cross-validated using a scatter image library of commensals that are likely to grow on the selective/differential media ( ; ). The use of inelastic scattering technologies (FT-IR, Raman) or molecular methods, such as colony PCR, colony immunoblot, and whole genome sequencing can also be used to validate results.
The Petri plate has facilitated some of the most significant discoveries in microbiology. It is important to note that during the beginning of the field of microbiology, many discoveries were made by accident or luck. If Fleming had swept his cultures off the bench into the trash, he may not have seen the zones of inhibition around the Penicillium colonies. If he had been using liquid media instead of modern Petri plates, he may not have even noticed the mold contamination. The rise of differential chromogenic or selective media facilitated the discovery of many organisms because it allowed researchers to isolate organisms based on their metabolic activities. Selective, differential, and enrichment media all had a significant impact on the field of microbiology, and the implementation of these media would have been made much more difficult without the Petri plate. Perhaps the single most important effect that the Petri plate has had is its ability to separate, isolate and protect cultures. Not until the modern age of genomic sequencing was a diverse culture able to be analyzed and characterized, hierarchical testing demanded that a pure culture be maintained and tested to characterize and identify the organism. Studies on microbial community and the search for novel probiotics against varieties of ailments would require improved culturing and isolation and Petri-plate would be at the forefront of the exploration. In 2007, researchers at Purdue University reported a light scattering sensor technology, called BARDOT that could scan a diverse sample that had been dispersed on a Petri plate. This technology made it possible to rapidly identify organisms directly from the Petri plate. By directing a red-diode laser (635 nm) through the center of the bacterial colony, BARDOT was able to identify a bacterial colony by its unique scatter image. This high throughput, label-free technology is capable of revolutionizing the way that industries can detect organisms further aiding in gene sequencing, pathogenicity, antibiotic susceptibility, vaccine development, and characterizing industrially beneficial traits. Likewise, the development of hyperspectral imaging and Raman spectrometry tools is gaining significant interest among scientists as reliable on-plate pathogen detection tools. All of this would not be possible without the simple innovation that Julius Petri made in 1887, which we call the Petri plate.
Conceptualization AB, study design AB, AS, KP, BA, writing and editing AB, AS, KP, BA. All authors contributed to the article and approved the submitted version.
|
BRCA1‐associated‐protein‐1 inactivated melanocytic tumours: characterisation of the clinicopathological spectrum and immunohistochemical expression pattern of preferentially expressed antigen in melanoma | aa92661c-37bb-49b8-bcf2-6836c2332338 | 11649512 | Anatomy[mh] | BAP1‐inactivated melanocytic tumours (BIMT), also referred to as BAP1‐inactivated melanocytomas, present clinically as skin‐coloured, dome‐shaped papules typically effecting trunk, head and neck area and extremities. Histologically, they are dermally based and show an epithelioid cell morphology with varying degrees of cytological atypia. BAP1 stands for BRCA1‐associated protein 1 and is encoded on chromosome 3 (locus 3q21.1). It functions as a tumour suppressor gene that is implicated in DNA damage response, transcriptional regulation and chromatin modulation. Germline mutations in BAP1 result in an autosomal dominant tumour predisposition syndrome that is associated with a high risk of developing various tumours, including BIMT, cutaneous melanoma, uveal melanoma, mesothelioma, renal cell carcinoma, lung adenocarcinoma and meningioma. BIMT are often the first symptom of the tumour predisposition syndrome, and histopathological recognition is therefore important to guide patient management. The presence of varying degrees of cytological atypia in BIMT poses a diagnostic challenge to distinguish the entity from atypical Spitz tumours and melanoma. , PRAME (preferentially expressed antigen in melanoma) is a cancer testis antigen that has demonstrated valuable utility in assisting in differentiating melanoma from benign counterparts, given its high specificity of diffuse expression in melanoma and absence to low expression in benign melanocytic naevi. , The PRAME expression profile has not been comprehensively investigated in BIMT. The aims of this study were to (1) describe the clinical and histopathological spectrum of BIMT in a large patient cohort in Southern Alberta, Canada, (2) study the behaviour of BIMT by providing long‐term follow‐up and (3) study the expression pattern of PRAME in BIMT. Ethical approval was obtained from the health research ethics board of Alberta (HREBA.CC‐19‐0379). Haematoxylin and eosin‐stained sections of 65 BAP1‐inactivated melanocytic tumours were retrieved from the departmental files of the Alberta Precision Laboratories, Calgary, Alberta, Canada. BAP1 inactivation was defined as complete loss of nuclear staining of the immunohistochemical marker BAP1. The histological features were reviewed and the following histopathological criteria were documented: tumour circumscription, junctional and dermal component, pigmentation, stromal fibrosis, cytological atypia (mild, moderate, severe), mitotic activity, necrosis, infiltrative growth, inflammatory infiltrate and existence of a conventional melanocytic in the background. Clinical data, including genetic test results for BAP1 germline mutations and follow‐up, were obtained from patient records. All patient records were reviewed for additional cancer diagnoses and additional skin excisions. Immunohistochemistry with adequate controls was performed for S‐100, Sox10, MelanA, HMB45, p16, BAP1 and Ki‐67, according to the manufacturer's instructions (Table ). In a subset of cases, immunohistochemical staining had been performed for routine work‐up, and in these cases the slides were reviewed without repeating the stains. Immunohistochemistry for PRAME was performed on 4‐μm‐thick formalin‐fixed paraffin‐embedded whole tissue sections following pressure cooker antigen retrieval (Target Retrieval Solution; pH 6.1 citrate buffer; Dako, Carpinteria, CA, USA) using a rabbit anti‐PRAME monoclonal antibody (1:100 dilution; clone EPR20330; Biocare Medical, Pacheco, CA, USA); the Novolink polymer detection system (Leica, Buffalo Grove, IL, USA) was used. Nuclear staining was assessed and given as the percentage of overall tumour cells (0%: 0; 1–25%: 1, 26–50%: 2; 51–75%: 3, 76–100%: 4) and staining intensity (ranging from 0 to 4 as follows: absent: 0; weak: 1; moderate: 2; strong: 3). A combined score was calculated as the sum of the quantity and staining intensity scores. The normal peritumoural tissue served as positive control. The source of the antibodies and their dilutions are listed in Table . Clinical features Our histopathological archives for BAP1‐inactivated tumours were searched from 2010 to 2022. The Calgary health zone includes an average of 1 500 000 people. Sixty‐five BIMT involving 31 female and seven male patients (ratio f:m = 4.4:1) were included in the study. With 38 patients affected by BIMT the estimated prevalence of BIMT in the Calgary health zone is 0.000025. With seven patients carrying a BAP1 germline mutation the prevalence of BAP1 germline mutations presenting with BIMT is approximately 0.0000047 in this cohort. The patient age ranged from 16 to 77 years with a mean of 39.6 years; two patients were younger than 18 years. All tumours were completely excised by primary excision or re‐excision. Seven patients (18.4%) had a BAP1 germline mutation. These patients presented at a younger age (range = 16–66, median = 25 years) without sex predilection (four females, three males). The results of the genetic analyses with documentation of the specific mutations were available for six patients and included the following: c.376‐2A>G, c.376‐2_392del, c.1717delCp.(Leu573TrpfsTer3), c.458_549delCT, c.1358_1359del and c.485_495delCT. The majority of BIMT were located on the trunk ( n = 26, 43%, including 15 BIMT located on the back, three on the shoulder, four on the chest and four on the abdomen) and on the head and neck area ( n = 26, 43%, 11 BIMTs on the face, seven on the ears, five on the neck and three located on the scalp). The remaining tumours were located on the extremities ( n = 13, seven located on the upper extremities, six on the lower extremities including acral sites). The tumours presented with a median size of 0.55 cm (range = 0.2–1.5 cm, available for n = 22). Patients with BAP1 germline mutations presented frequently with multiple BIMT (range of number of BIMT per patient 1–8, mean 6). The anatomical distribution did not differ significantly between germline‐associated and sporadic tumours, nor did the size. No recurrences or metastases of BIMT were noted in the entire cohort (follow‐up period = 4–111 months, mean = 44 months). One male patient with BAP1 germline mutation died of complications of mesothelioma at the age of 69 years, 42 months after the diagnosis of one BIMT. This patient's history is also remarkable for two basal cell carcinomas and an invasive melanoma that the patient developed 3 years prior to his mesothelioma. The melanoma (size 1.5 × 1.4 cm) was amelanotic, showed spindle cell morphology and a maximum tumour thickness of 2.1 cm. None of the remaining patients with BAP1 germline mutations developed a malignant tumour diagnosis during the follow‐up period (follow‐up period range = 8–111 months, mean = 49 months). According to the medical records, 26 patients also had a history of conventional melanocytic naevi unrelated to their BIMT(s). The number of conventional melanocytic naevi ranged from one to 28 per patient. Within the group of patients with sporadic BIMT, the mean number of additional conventional melanocytic naevi was four per patient. Within the group of patients with BAP1 germline mutations, the mean number of additional conventional naevi was 5.5. Within the group of patients with sporadic BIMT (patient n = 31), one patient had a remote history of melanoma (no histopathological data available), two had a history of basal cell carcinoma, one patient had diffuse large B cell lymphoma, one patient revealed a history of Hodgkin lymphoma plus a history of ductal carcinoma in‐situ of the breast, one patient developed invasive ductal carcinoma of the breast and one patient had a remote history of prostatic adenocarcinoma plus a remote history of melanoma. Two patients with sporadic BIMT harboured BRCA germline mutations (one BRCA1 and one BRCA2 mutation). The clinical characteristics of all BIMT are summarised in Table . Histological features All BIMT were well‐circumscribed, nodular tumours located within the superficial and mid‐dermis (Figure ). Twenty‐six tumours (40%) showed a minor junctional component consisting of small melanocytic nests composed of epithelioid melanocytes and few single epithelioid cells (Figure ). The tumours were composed of nests and sheets of non‐pigmented (37 tumours, 57%) or lightly pigmented (28 tumours, 43%) epithelioid cells with amphophilic cytoplasm and round to ovoid nuclei with evenly dispersed chromatin and prominent nucleoli (Figure ). Moderate to severe cytological atypia, including irregular nuclear contours, nuclear pseudoinclusions, bizarre‐formed nuclei, multinucleation and hyperchromasia, was present in 41 tumours (63%) (Figure ). No dermal mitoses were observed in the majority of BIMT (80.2%). Low mitotic activity was observed in seven tumours (10.8%), ranging from one to two mitoses per mm 2 , but no atypical mitotic figures were identified. No foci of tumour necrosis were observed in any tumour. A brisk lymphocytic inflammatory infiltrate was present in seven tumours (10.8%), a mild to moderate inflammatory infiltrate in 35 cases (53.8%) and no inflammation was seen in 23 BIMT (35.4%). Only one germline‐associated tumour showed significant stromal fibrosis; the remainder of cases had no remarkable fibrosis. One germline‐associated BIMT showed angiomatoid features with multiple small, dilated vessels intermingling with the epithelioid melanocytes. A conventional background naevus flanking the BAP1 inactivated proliferation on one or both sides was present in majority of the tumours ( n = 50, 76.9%). The background naevus made up the lesser part of the tumours in all cases; the dominant component was the dermal BAP1‐inactivated proliferation. The background conventional naevus was composed of dermal melanocytic nests only (33 cases), but also revealed compound architecture in 17 cases. Two cases showed intermingling of the banal, smaller naevus cells with the large, epithelioid BAP1‐inactivated tumour cells (Figure ), whereas most tumours had a clear demarcation of conventional naevus and BAP1‐inactivated cells. The diagnosis of the cutaneous invasive melanoma arising in one patient with BAP1 germline mutation was straightforward histopathologically. The tumour was composed of atypical spindle cells arranged in fascicles and sheets within dermis demonstrating an infiltrative growth pattern and a high mitotic rate (20 per mm 2 ). No adjacent conventional naevus nor a BIMT was present in the periphery of the melanoma. Immunohistochemistry All tumours were strongly and diffusely positive for S100 (nuclear and cytoplasmic), Sox10 (nuclear), MelanA (cytoplasmic) (Figure ) and negative for HMB45, except for a junctional component in the conventional background naevi. Nuclear p16 staining (available for 37 cases) was either retained (19 cases, 51.4%) or mosaic (Figure ). PRAME showed focal or patchy, weak nuclear staining in all tumours (Figure ). The overall combined score was low with a mean of 3 (range = 0–80), quantity range = 0–40% of tumour cells, intensity range = 0–2). Ki‐67 staining revealed a low mitotic index (Figure ) in all BAP inactivated tumours. The melanoma arising in the patient with BAP1 germline mutation showed loss of nuclear BAP1 staining but diffusely positive cytoplasmic staining, and PRAME showed focal nuclear expression. Our histopathological archives for BAP1‐inactivated tumours were searched from 2010 to 2022. The Calgary health zone includes an average of 1 500 000 people. Sixty‐five BIMT involving 31 female and seven male patients (ratio f:m = 4.4:1) were included in the study. With 38 patients affected by BIMT the estimated prevalence of BIMT in the Calgary health zone is 0.000025. With seven patients carrying a BAP1 germline mutation the prevalence of BAP1 germline mutations presenting with BIMT is approximately 0.0000047 in this cohort. The patient age ranged from 16 to 77 years with a mean of 39.6 years; two patients were younger than 18 years. All tumours were completely excised by primary excision or re‐excision. Seven patients (18.4%) had a BAP1 germline mutation. These patients presented at a younger age (range = 16–66, median = 25 years) without sex predilection (four females, three males). The results of the genetic analyses with documentation of the specific mutations were available for six patients and included the following: c.376‐2A>G, c.376‐2_392del, c.1717delCp.(Leu573TrpfsTer3), c.458_549delCT, c.1358_1359del and c.485_495delCT. The majority of BIMT were located on the trunk ( n = 26, 43%, including 15 BIMT located on the back, three on the shoulder, four on the chest and four on the abdomen) and on the head and neck area ( n = 26, 43%, 11 BIMTs on the face, seven on the ears, five on the neck and three located on the scalp). The remaining tumours were located on the extremities ( n = 13, seven located on the upper extremities, six on the lower extremities including acral sites). The tumours presented with a median size of 0.55 cm (range = 0.2–1.5 cm, available for n = 22). Patients with BAP1 germline mutations presented frequently with multiple BIMT (range of number of BIMT per patient 1–8, mean 6). The anatomical distribution did not differ significantly between germline‐associated and sporadic tumours, nor did the size. No recurrences or metastases of BIMT were noted in the entire cohort (follow‐up period = 4–111 months, mean = 44 months). One male patient with BAP1 germline mutation died of complications of mesothelioma at the age of 69 years, 42 months after the diagnosis of one BIMT. This patient's history is also remarkable for two basal cell carcinomas and an invasive melanoma that the patient developed 3 years prior to his mesothelioma. The melanoma (size 1.5 × 1.4 cm) was amelanotic, showed spindle cell morphology and a maximum tumour thickness of 2.1 cm. None of the remaining patients with BAP1 germline mutations developed a malignant tumour diagnosis during the follow‐up period (follow‐up period range = 8–111 months, mean = 49 months). According to the medical records, 26 patients also had a history of conventional melanocytic naevi unrelated to their BIMT(s). The number of conventional melanocytic naevi ranged from one to 28 per patient. Within the group of patients with sporadic BIMT, the mean number of additional conventional melanocytic naevi was four per patient. Within the group of patients with BAP1 germline mutations, the mean number of additional conventional naevi was 5.5. Within the group of patients with sporadic BIMT (patient n = 31), one patient had a remote history of melanoma (no histopathological data available), two had a history of basal cell carcinoma, one patient had diffuse large B cell lymphoma, one patient revealed a history of Hodgkin lymphoma plus a history of ductal carcinoma in‐situ of the breast, one patient developed invasive ductal carcinoma of the breast and one patient had a remote history of prostatic adenocarcinoma plus a remote history of melanoma. Two patients with sporadic BIMT harboured BRCA germline mutations (one BRCA1 and one BRCA2 mutation). The clinical characteristics of all BIMT are summarised in Table . All BIMT were well‐circumscribed, nodular tumours located within the superficial and mid‐dermis (Figure ). Twenty‐six tumours (40%) showed a minor junctional component consisting of small melanocytic nests composed of epithelioid melanocytes and few single epithelioid cells (Figure ). The tumours were composed of nests and sheets of non‐pigmented (37 tumours, 57%) or lightly pigmented (28 tumours, 43%) epithelioid cells with amphophilic cytoplasm and round to ovoid nuclei with evenly dispersed chromatin and prominent nucleoli (Figure ). Moderate to severe cytological atypia, including irregular nuclear contours, nuclear pseudoinclusions, bizarre‐formed nuclei, multinucleation and hyperchromasia, was present in 41 tumours (63%) (Figure ). No dermal mitoses were observed in the majority of BIMT (80.2%). Low mitotic activity was observed in seven tumours (10.8%), ranging from one to two mitoses per mm 2 , but no atypical mitotic figures were identified. No foci of tumour necrosis were observed in any tumour. A brisk lymphocytic inflammatory infiltrate was present in seven tumours (10.8%), a mild to moderate inflammatory infiltrate in 35 cases (53.8%) and no inflammation was seen in 23 BIMT (35.4%). Only one germline‐associated tumour showed significant stromal fibrosis; the remainder of cases had no remarkable fibrosis. One germline‐associated BIMT showed angiomatoid features with multiple small, dilated vessels intermingling with the epithelioid melanocytes. A conventional background naevus flanking the BAP1 inactivated proliferation on one or both sides was present in majority of the tumours ( n = 50, 76.9%). The background naevus made up the lesser part of the tumours in all cases; the dominant component was the dermal BAP1‐inactivated proliferation. The background conventional naevus was composed of dermal melanocytic nests only (33 cases), but also revealed compound architecture in 17 cases. Two cases showed intermingling of the banal, smaller naevus cells with the large, epithelioid BAP1‐inactivated tumour cells (Figure ), whereas most tumours had a clear demarcation of conventional naevus and BAP1‐inactivated cells. The diagnosis of the cutaneous invasive melanoma arising in one patient with BAP1 germline mutation was straightforward histopathologically. The tumour was composed of atypical spindle cells arranged in fascicles and sheets within dermis demonstrating an infiltrative growth pattern and a high mitotic rate (20 per mm 2 ). No adjacent conventional naevus nor a BIMT was present in the periphery of the melanoma. All tumours were strongly and diffusely positive for S100 (nuclear and cytoplasmic), Sox10 (nuclear), MelanA (cytoplasmic) (Figure ) and negative for HMB45, except for a junctional component in the conventional background naevi. Nuclear p16 staining (available for 37 cases) was either retained (19 cases, 51.4%) or mosaic (Figure ). PRAME showed focal or patchy, weak nuclear staining in all tumours (Figure ). The overall combined score was low with a mean of 3 (range = 0–80), quantity range = 0–40% of tumour cells, intensity range = 0–2). Ki‐67 staining revealed a low mitotic index (Figure ) in all BAP inactivated tumours. The melanoma arising in the patient with BAP1 germline mutation showed loss of nuclear BAP1 staining but diffusely positive cytoplasmic staining, and PRAME showed focal nuclear expression. Prior to their first description in 2011 by Wiesner et al . BIMT had been classified as epithelioid Spitz tumours or melanomas. Since 2011, more information and details concerning the histopathological features, the pathogenesis and the genetic background of BIMT have been gathered. BIMT show a bi‐allelic inactivation of the BAP1 tumour suppressor gene located on chromosome 3q21, which can be caused by loss‐of‐function mutations or by deletion affecting the BAP1 locus. BIMT typically arises from a conventional naevus with BRAFp.V600E or NRAS mutations or RAF1 fusion. , , The double hit results in a clonal expansion of the BAP1‐inactivated clone with the typical epithelioid phenotype. In the sporadic setting, the double hit is caused by loss‐of‐function mutations altering the BAP1 nucleotide sequence often combined with a chromosomal deletion involving the wild‐type BAP1 locus. In patients with BAP1 germline mutations, the second hit is the inactivation of the remaining wild‐type BAP1 allele. , , Despite this pathogenetic insight, data regarding the prevalence and the behaviour of BIMT have been scarce. Approximately 200 families with BAP1 germline variants have been described to date. In our cohort, seven patients presenting with BIMT carried a BAP1 germline mutation, and we calculated the prevalence to be approximately 0.0000047 in the Calgary health zone. The overall prevalence of BIMT in out cohort was 0.000025, stressing that BIMT occur more commonly in the sporadic setting than in the syndromic setting. Our data confirm that BIMT are often the primary manifestation in patients carrying a BAP1 germline mutation, and that these patients present with multiple BIMT at a young age, commonly within the second decade of life, as reported previously. The diagnosis of a single BIMT in a patient does not imply genetic testing for a BAP1‐germline mutation unless multiple BIMT are seen in the same patient or there is clinical suspicion due to a positive family history or manifestation of other tumours, especially uveal melanoma, cutaneous melanoma, mesothelioma and renal cell carcinoma. Malignant transformation of BIMT has been reported in both sporadic and germline‐associated tumours, , but the majority of BIMT show an indolent behaviour. Our data stress the indolent behaviour of BIMT in both settings, sporadic and syndromic. None of the tumours showed recurrence or aggressive behaviour despite worrisome histopathological features. No malignant transformation was noted in any of the BIMT in this study, which argues against BIMT being a significant melanoma precursor lesion. The spindle cell melanoma arising in one of the patients with BAP1 germline mutation did not show any histopathological resemblance with BIMT, despite loss of nuclear BAP1 by immunohistochemistry, and no conventional naevus or a BIMT was present in the background. The cytoplasmic expression of BAP1 in this melanoma is a finding of unknown significance. Cytoplasmic expression of BAP1 has been described in a subset of uveal melanomas, suggesting a functional role of BAP1 within the cytoplasm that warrants further investigation. Histopathologically, the classic morphology of BIMT is described as a biphasic growth pattern with a nodular or sheet‐like proliferation of epithelioid melanocytes in the background of an adjacent conventional naevus component. The larger epithelioid cells typically display abundant amphophilic cytoplasm with vesicular nuclei and prominent nucleoli, resembling the melanocytes seen in Spitz naevi. Other rare morphological features including rhabdoid, adipocyte metaplasia and nuclear pseudo‐inclusions have also been described. , , In a larger histopathological study conducted by Garfield et al ., a significant association of an extensive junctional component in BIMTs arising in patients with germline BAP1 mutations was observed. In our study, we did not see any significant histopathological differences between BIMT arising in the sporadic versus germline‐associated setting; in particular, a more prominent junctional component was occasionally present in both groups. According to our observations, germline‐associated BIMTs can present purely dermally or with a junctional involvement in the same patient. Pagetoid spread was observed in one single BIMT, occurring in a 16‐year‐old female with BAP1 germline mutation on the upper back. A prominent lymphohistiocytic infiltrate has frequently been described as a typical feature of BIMT. In our study we only saw a brisk lymphohistiocytic infiltrate in 10.8% of tumours, a mild lymphohistiocytic infiltrate in 40% of cases and no inflammatory infiltrate in 40.2%. The review of all patient records in our cohort reveals that patients with BIMT also commonly develop multiple banal conventional naevi without significant atypia, and only one patient had a history of primary cutaneous melanoma. Two patients with sporadic BIMT also had breast carcinomas and harboured BRCA mutations, which is probably a coincidental finding. The same assumption applies to the other malignant diagnoses in patients with sporadic BIMT in this cohort. PRAME immunohistochemistry achieved low combined scores of quantity and intensity in all BIMT in this study. Previous data on PRAME expression in BIMT are limited. Lopez et al . studied five BIMT, none of which had an immunoreactivity score greater than 1+ (staining of 1 to 25% of tumour cells), and all cases demonstrated a weak staining intensity. In another study, Turner et al . reviewed PRAME immunohistochemistry in a small number of BIMT. In their study, diffuse PRAME positivity (defined as at least weak nuclear positivity in > 75% of atypical cells) was present in two of five cases. The remaining three cases showed non‐diffuse or negative PRAME staining. Both studies used a different PRAME antibody and incubation protocol compared to our study. Taken together, the data suggest that diffuse PRAME positivity in BIMT is a rare phenomenon. BIMT are indolent tumours characterised by large dermal epithelioid melanocytes with nuclear loss of BAP1, and present with a variable amount of cytological atypia and low mitotic activity. A conventional background naevus is present in > 75% of cases that should not mislead to a diagnosis of melanoma arising in a naevus. Malignant transformation of BIMT was not noted. Most BIMT occur sporadically as single tumours in patients who may also develop conventional melanocytic naevi without significant atypia. When arising in patients with BAP1 germline mutation, BIMT are often multiple and affect patients in their second decade of life. Cutaneous melanoma arising in patients with BAP1 germline mutation develop de‐novo without precursor in majority of patients. PRAME consistently shows patchy and weak staining in BIMT and serves as a reassuring tool to distinguish BIMT from melanoma. Yitong Xu: data collection, analysis and interpretation of results. Alejandro A Gru: immunohistochemical staining and interpretation of results. Thomas Brenn: study conception and design, analysis and interpretation of results. Katharina Wiedemeyer: study conception and design, interpretation of results and manuscript preparation. The authors have nothing to disclose. |
Circulating Cell-Free DNA Integrity for Breast and Prostate Cancer: What Is the Landscape for Clinical Management of the Most Common Cancers in Women and Men? | f33469e9-1905-445e-8c00-9f58edefb5b6 | 11817310 | Biopsy[mh] | Breast cancer (BC) is a complex and heterogeneous disease , which is one of the most common cancers in women in terms of both incidence and mortality . According to Globocan 2024, 2,308,897 new cases and 665,684 deaths are expected worldwide in 2022 . However, thanks to increasing therapeutic options, the prognosis of BC patients has improved considerably in recent years . Tumour grade, hormone receptor status, and human epidermal growth factor receptor 2 (HER2) are the most important cancer-related factors that influence metastatic potential, response to treatment, and prognosis . Metastatic BC (MBC) remains an incurable disease, but early detection can increase survival rates in the early stages . Nevertheless, in both cases, a tissue biopsy is urgently required for pathological assessment and for the selection of biomarker-based therapeutic approaches . Unfortunately, standard methods are associated with high costs, invasiveness, and low sensitivity and specificity ; therefore, they are not suitable for all patients . Another important limitation is the ability to capture the entire genomic landscape of the tumour, including intratumoural heterogeneity . Prostate cancer (PCa) is a widespread malignant disease in men. It is the second most commonly diagnosed cancer and one of the main causes of cancer-related deaths worldwide. In 2022, 1,466,680 new cases and 396,792 deaths were reported worldwide . The incidence of PCa increases with age, with over 85% of diagnoses occurring in men over the age of 60 . In general, for both BC and PCa patients, conventional diagnostic methods, such as tissue biopsies, are invasive and may not fully capture the heterogeneity of the tumour. Peripheral blood is one of the most important sources of various circulating cell-free nucleic acids, circulating tumour cells (CTCs), and exosomes. One type of nucleic acid in the blood is cell-free DNA (cfDNA), which is derived from DNA fragments released into the peripheral circulation from either dying tumours or normal cells. A special population of cfDNA is the circulating tumour DNA (ctDNA). It is differentially selected on the basis of tumour-specific DNA changes such as gene amplifications, mutations, methylations, and rearrangements . Liquid biopsy, a minimally invasive technique that analyses tumour-derived material such as cfDNA, CTCs, and extracellular vesicles (EVs) in body fluids, has emerged as a promising diagnostic, prognostic, and predictive tool . The use of liquid biopsy has completely changed some clinical practices and improved patient outcomes . Circulating tumour DNA (ctDNA), i.e., the part of cfDNA that enters the bloodstream as a result of tumour cell death , has been shown to be an informative circulating tool in both BC and PCa . For example, early-stage ctDNA clearance has been associated with higher rates of complete pathological response after neoadjuvant treatment and with a low rate of disease relapse in BC . Through longitudinal sampling, ctDNA can monitor response to treatment and individualise optimal treatment in MBC . In localised PCa, liquid biopsy can differentiate between low- and high-grade cancer, helping to decide whether to perform or postpone a tissue biopsy. In advanced stages of disease, it provides prognostic information beyond standard tests such as PSA (prostate-specific antigen) levels and can monitor response to treatment and tumour progression. ctDNA has been used to screen for androgen receptor gene mutations in PCa patients who develop castration-resistant metastatic cancer to achieve prognostic and therapeutic goals or to determine genomic mutational burden in relation to disease aggressiveness and progression . In patients with metastatic castration-resistant disease, specific genomic alterations in ctDNA have been associated with clinical outcomes such as progression-free survival (PFS) and overall survival (OS) . These biomarkers can give information that is either qualitative (e.g., mutation type) or qualitative (i.e., copy numbers of mutations). However, there is a limitation to this method, as not all patients have the same mutations. For example, p53 is mutated in only 26–88% of BC patients. The variation depends on the BC subtype . The same is true for metastatic PCa patients, where the mutation burden involving androgen receptors is 10–20% of metastatic castrate-resistant PCa patients; another tumour-related gene, such as PI3KCA, accounts for 6% of metastatic patients, or germline/somatic mutation of DNA damage repair genes that are found in 15–30% of metastatic patients, half of which is of germline origin . In BC, PIK3CA mutations vary between 25 and 40%, with the highest mutation rate in HR+/HER2− metastasised BC subtypes . In contrast, cfDNA integrity analysis (cfDI) is a potential tool in liquid biopsies that can overcome the limitations of heterogeneity and differential mutation rate in cancer development and progression. The main principle behind cell-free DNA integrity (cfDI) is the fact that normal cells undergo apoptosis, releasing shorter fragments of about 200 bp, whereas tumour cells undergo different death processes, including autophagy or simply necrosis, and release different fragment sizes, often longer than those of the apoptotic process . cfDI can be detected by the longer/shorter DNA fragments ( ). Furthermore, the analysis of cfDI has more targets in the samples and is independent of genetic and epigenetic ctDNA. In this review, we provide an overview of the current applications of cfDI tests in BC and PCa and their potential development in clinical practice. We have focused on these two cancers because they are the most common hormone-dependent cancers in women and men, respectively, for which a cost-effective liquid biomarker representative of tumour heterogeneity is a need in many Western countries to aid diagnosis, prognosis, and prediction of response to therapy in the context of precision medicine. To ensure a complete search in PubMed for DNA integrity, we used the following keywords: “circulating” or “cell-free” “DNA” “integrity” “breast cancer” or “prostate cancer”. From the search results, we selected all original articles without time limitations.
Notwithstanding the great progress that has been made in recent years in the field of BC and PCa, there is still a need for better diagnostic and prognostic tools that can predict the onset, recurrence, or progression of these diseases at an early stage. Liquid biopsy has been proposed as a less invasive and equally valid method compared to needle biopsy. One of the targets in liquid biopsy is known as circulating cell-free DNA integrity (cfDI), which is the fragmentation pattern of cfDNA. Cancer cells can undergo many different types of cell death, including apoptosis, but also necrosis and other types of cell death. Differences in the amount of longer and shorter fragments can be associated with cancer. DNA aberrations in cancer cells also contribute to the difference between longer and shorter fragments that can enter the bloodstream. cfDI was mainly characterised by qPCR using a delta–delta formula based on a threshold cycle (Cp), e(−ΔΔCp × ln(2)), or based on a ratio between the amount of longer fragments and shorter fragments considered representative of the total amount of cfDNA associated with the process of cell death. In fact, DNA fragmentation profiles are different in cancer patients from healthy subjects . 2.1. cfDI as a Biomarker in Breast Cancer The following section summarises the cfDI discoveries published in the literature in chronological order ( ). In the table, the average age of the cohorts was rounded to decimal numbers. Main results refer to significant results; if no significant results are available, a trend is indicated as non-significant. Clinical value refers to the potential clinical application if significant results are demonstrated. For most studies, qPCR determination of target molecules for cfDI was used, and only methods deviating from qPCR are mentioned in the text. The first pioneering work was that of Umetani et al. in 2006, who investigated the prognostic ability of cfDI in serum to predict the progression of BC. For this purpose, the authors analysed ALU 260/111 in 83 BC and 51 healthy control women (HC). The results showed that cfDI levels were significantly higher in AJCC stage II, III, and IV BC than in healthy women. They also demonstrated a linear correlation between tumour stage and size, with the ROC curve allowing differentiation between stage II to IV BC and HC (AUC = 0.79). Their data show that cfDI can predict preoperative lymph node metastases . Deligezer et al. then used ALU 247/115 to investigate whether this target could mark the effects of chemotherapy in BC patients. They found that this cfDI did not differ in patients before and after chemotherapy but was significantly higher in BC than in HC . In 2012, Agostini et al. analysed preoperative plasma cfDI as a biomarker for regional lymph node metastases in BC. For this purpose, the authors analysed ALU 247/115 in a small sample of BC patients and healthy women. The results showed that cfDI levels were higher in BC patients than in healthy controls. In addition, cfDI showed a moderate ability to detect patients with lymph node metastases . However, Lehner et al. demonstrated that cfDI of ALU 247/115 was not related to the response to neoadjuvant treatment in BC patients , but in 2014 Stötzer et al. investigated cfDI as a diagnostic biomarker for BC progression in a large sample population of BC patients. To this end, the authors investigated ALU 247/115 in localised BC, metastatic breast cancer (MBC), benign breast disease (BBD), and healthy female controls (HC). They used two different cfDI determinations: cfDI 1 as the result of the ratio of the qPCR amount of fragments and cfDI 2 as the ratio of qPCR Cp as the ΔΔCp formula. The results showed that cfDI 1 was significantly higher in HC than in BBD and higher in BC and MBC than in BBD. The cfDI 2 confirmed higher values in HC than in BBD and was higher in MBC than in local BC and BBD. However, both cfDI values were not suitable to identify BC compared to HC. On the contrary, it should be noted that the concentrations of ALU fragments allowed a better differentiation between BC and HC (AUC > 95%) and that for MBC, the classical CA 15-3 and CEA had the best diagnostic value . Interestingly, during the same year, Madhavan et al. evaluated cfDI as an early diagnostic biomarker in a large cohort of primary BC (PBC) and MBC patients, using other targets. The authors studied ALU 260/111 and LINE1 266/97 in a large cohort of MBC, PBC, and healthy females (HC). In contrast to the other two groups, the results obtained from this group showed that cfDI was significantly lower in PBC patients than healthy controls. The ROC curves showed a discriminating ability to distinguish HC from PBC (AUC = 0.75), HC from CTC-positive MBC (AUC = 0.93), and HC from CTC-negative MBC (AUC = 0.81). Also, discrimination ability was found for PBC from CTC-negative and CTC-positive MBC (AUC = 0.71 and 0.86, respectively) and for CTC-negative MBC from CTC-positive MBC (AUC = 0.93). In MBC the lowest level was shown in CTC-positive MBC, and the cfDI correlated with worse PFS and OS too . Iqbal et al. evaluated cfDI as a prognostic biomarker for BC prediction. For this purpose, the authors investigated ALU 247/115 in PBC and healthy females. The results contemplated that cfDI was significantly higher in PBC patients than HC. Interestingly, tumour size and cfDI together showed a marked association with disease-free survival (DFS) . Kamel et al. investigated the diagnostic potential of cfDI in BC using β-actin fragments (400 bp and 100 bp) in a large group of BC patients at different stages prior to therapeutic intervention. They found cfDI β-actin 400/100 was significantly higher in BC patients with respect to benign disease patients (BBD) or matched HC. The ROC curves of cfDI showed 95% CI for discriminating BC from BBD or HC. Also, cfDI was shown to increase with the increase in BC stage, although it did not relate with histopathological type or grade . Maltoni et al. investigated the role of cfDI as a prognostic biomarker using different targets such as BCAS 1 266/129, MYC 264/128, and PIK3CA 274/129 in relapsed BC, non-relapsed BC, and HC. The results showed that cfDI of MYC and of PIK3CA were significantly lower in patients than in HC, but no significant results were obtained in distinguishing between relapsed and non-relapsed BC patients . Wang et al. investigated the cfDI of ALU 260/111 in non-metastatic BC compared to patients with benign breast disease. They found that cfDI was significantly lower in BC patients than in benign controls. In addition, they found that the cfDI of ALU 260/111 had a ROC curve with an AUC of 0.67 and was a better biomarker than CTCs, cfDNA concentration, and CEA153. Interestingly, cfDI and CTCs together improve sensitivity and specificity and reduce the false positivity of cfDI alone from 50% to 10.7% . Cheng et al. showed that cfDI was lower in a large group of recurring BCs compared to a non-recurring BC group. To this end, the authors analysed ALU 260/111 and LINE1 266/97 and showed that the cfDI of ALU, LINE1, or both in the ROC had an AUC of 0.71, 0.704, and 0.732, respectively. These data also suggested that cfDI was an independent marker of recurrence in BC . Cheng et al. then investigated cfDI as a biomarker for response to therapy by measuring ALU 260/111 and LINE1 266/97 in a large group of MBC patients. The results showed an increase in cfDI ( p = 1.21 × 10 −7 for ALU and p = 1.87 × 10 −3 for LINE1) after one cycle of therapy in MBC patients. Of note, cfDI was an independent prognostic marker . Then, Tang et al. investigated the function of cfDI as a diagnostic biomarker in BC by ALU 247/115 in small sample groups of BC, BBD, and HC. The results showed that cfDI was significantly higher in BC than in BBD and HC (cfDI did not differ between BBD and HC). The ROC curve showed an AUC of 0.97, and cfDI correlated with lymph node metastasis and tumour stage . Salimi et al. investigated the role of cfDI biomarkers for BC progression in triple-negative breast cancer (TNBC). The authors analysed the ratio of β-actin 394 bp/ β-actin 99 bp in TNBC and non-TNBC patients as well as in healthy women. The results showed that cfDI was significantly higher in TNBC and non-TNBC patients than in healthy controls and was associated with lympho-nodal metastasis and tumour stage. In particular, the ROC curve showed an AUC = 0.997. Finally, the cfDI significantly increases in TNBC compared to non-TNBC, thus discriminating between the two BC types . Wang et al. investigated whether cfDI of ALU 260/111 and LINE1 266/97 can predict response to neoadjuvant chemotherapy (NACT) in a small group of BC patients. The results showed that cfDI was significantly higher in patients after NACT than in patients before NACT. Interestingly, the increase in cfDI in NACT-treated patients was associated with tumour shrinkage and a reduction in Ki67 levels . Miao et al. came to the opposite conclusion by using a cfDI determined by LINE1 259/97 in a study population of young BC patients, most of whom had high TNM and tumour size and were undergoing adjuvant chemotherapy. They found that cfDI was higher in BC patients before chemotherapy than in BBD or HC. Then, they proved that cfDI decreases after chemotherapy, but they collected blood samples 4 weeks after the completion of chemotherapy . Arko-Boham et al. investigated the role of cfDI of ALU247/115 as a biomarker to identify BC and prostate cancer compared to healthy individuals in small populations. The results showed that cfDI levels were lower in BC than in HC but without statistical significance. In contrast, a statistically higher cfDI value was found in stage II BC . Hussein et al. investigated the role of cfDI ALU 247/115 as a diagnostic and prognostic biomarker in a small group of BCs. The results showed that ALU 247/115 was significantly higher in BC patients than in HC with a ROC AUC = 0.825, but it could not be differentiated between metastatic and non-metastatic patients . Park et al. investigated the role of cfDI of ALU 263/58 as a diagnostic biomarker in BC patients and HC. It was found that the ALU-derived cfDI was significantly higher in patients compared to controls with an AUC of 0.724. However, the methylation status of the LINE1 target provided a better test to differentiate cases from healthy controls . Lamminaho et al. analysed cfDI in a large cohort of non-metastatic BC patients at the time of diagnosis. The cfDI was determined by the ratio of cfDNA fragments greater than 1000 bp to those with less than 1000 bp, measured by electrophoretic separation using the ScreenTape D5000 system. The results showed that a higher cfDI was correlated with significantly poorer survival, but only at a follow-up of 10 years. They showed that a high cfDI was an independent prognostic factor for OS and Breast Cancer Specific Survival (BCSS). This was seen in ER + BC patients in the multivariate analysis. Interestingly, a multivariate logistic regression analysis with cfDI, tumour characteristics, and age at diagnosis strongly improved the predictive results . Adusei et al. investigated in a small cohort of BC patients whether cfDI can be used to predict response to chemotherapy using ALU 247/115. The results showed that cfDI increased after the third cycle of chemotherapy (T2) compared to the second cycle (T1), but without statistical significance. No statistical significance was also found in cfDI of BC cases compared to HC . Cirmena et al. used the quantification of electrophoretic fragments of cfDNA from plasma to evaluate cfDI as a biomarker for response to neoadjuvant chemotherapy (NACT). To this end, the authors used the highly sensitive D1000 Screen Tape Station to measure fragments of 321 to 1000 bp versus fragments of 150 to 220 bp in a small sample of BC patients and HC. The results showed that at time 2 after NACT treatment, cfDI was significantly higher in pathological complete responders than in non-complete responders. Of note, cfDI with magnetic resonance imaging data analysed by ROC showed a predictive value for complete response of 0.875 . In an observational study, Elhelaly et al. investigated the role of cfDI ALU 247/115 as a biomarker for the early detection of BC and the discrimination of BBD. In their cohort of age-matched cases and controls, cfDI was significantly higher in BC than in BBD or control, but it did not discriminate between BBD and HC. The ROC showed an AUC of 0.727 . Hafeez et al. investigated cfDI as a prognostic biomarker for BC. To this end, the authors analysed ALU 247/115 in a small population study of BC, BBD, and healthy women. The results showed that cfDI was significantly higher in BC and BBD than in HC. cfDI was higher in early-stage BC and advanced BC than in BBD, but without statistical significance. To distinguish BC from HC, the ROC of cfDI showed an AUC = 0.71 . Then, Nair et al. analysed cfDI of ALU 247/115 as a prognostic factor in 167 BC patients with different molecular subtypes and stages of disease. They found that cfDI was significantly higher in preoperative BC patients than in postoperative patients, and interestingly, higher cfDI was associated with a mild immune infiltrate and thus a poorer prognosis. This evidence was confirmed by the negative correlation with DFS: DFS decreased from 82 months to 58 months in BC patients with high cfDI values. However, the authors emphasised that higher ALU 247 bp levels may be an independent factor for BC prognosis . In contrast to other groups using qPCR, Bortul et al. investigated cfDI with droplet digital PCR (ddPCR) in a large cohort of BC in 2023. For this purpose, they detected ALU 260/111 and LINE-1 266/97 with ddPCR in 106 BC and 103 HC. The results showed that both cfDI were significantly lower in BC compared to HC. In fact, cfDI LINE1 proved to be more accurate in discriminating cases from controls in ROC (AUC = 0.77) . In 2024, Celik et al. analysed for the first time ccfDNA levels, mtDNA, and DNA integrities together in a small sample of BC patients undergoing NACT compared to healthy controls. The authors investigated whether cfDI of ALU 247/ALU115 and mitochondrial nicotinamide dinucleotide adenine dehydrogenase 4 and 1 (ND4/ND1) can be used to predict response to chemotherapy alone or together with other biomarkers. cfDI ALU 247/ALU 115 was not significantly different in patients before treatment compared to controls but was significantly lower in patients after treatment compared to controls. In contrast, the copy number of mtDNA (ND4/ND1) was significantly higher in patients than in controls both before and after treatment . Also, in 2024, Giro et al. investigated cfDI ALU 247/ALU 115 in a small sample of BC patients 15 days after neoadjuvant chemotherapy. They found that cfDI was significantly higher in patients who achieved a pathological complete response (pCR) and correlated significantly with disease-free survival (DFS) . Finally, Gameel et al. investigated the potential of cfDI ALU 247/115 in a small sample population of BC, BBD, and HC as a biomarker for predicting recurrence. Although cfDI was higher in BC patients than in HC, the differences were not statistically significant . 2.2. cfDI as Biomarker in Prostate Cancer The efficiency of cfDI as a tumour biomarker was also investigated in prostate cancer (PCa), but the number of articles was lower than in BC (9 and 25, respectively). Unfortunately, the standard screening method using PSA analysis has too high sensitivity and too low specificity, which poses a diagnostic challenge as it may lead to overdiagnosis and treatment of latent cancer . Therefore, efforts are focused on the evaluation of other biomarkers such as cfDI, the studies of which are summarised in chronological order in this section ( ). Analogously to the BC section, in the table, the average age of the cohorts was rounded to decimal numbers. Main results refer to significant results; if no significant results are available, a trend is indicated as non-significant. Clinical value refers to the potential clinical application if significant results are demonstrated. All studies used qPCR to quantify the longer and shorter fragments for cfDI determination. In 2006, Boddy et al. investigated cfDI for the first time as a biomarker to differentiate PCa patients from patients with benign prostatic hyperplasia (BPH). They quantified the plasma levels of 356 bp and 105 bp sequences of the leptin gene (LEP) in PCa and BPH patients. Leptin is a pleiotropic peptide hormone secreted by adipose tissues and has been associated with various cancers related to obesity . The authors expressed the cfDI as ∆Ct between the average Cts of LEP 356 bp over 105 bp fragments. This ratio value showed no significant difference between PCa and BPH groups . In the same year, Hanley et al. compared the plasma cfDI of PCa patients with those of three different control groups: (1) healthy volunteers (HC) under 40 years of age (Ctr1); (2) patients with radical prostatectomy and low PSA six months after surgery (Ctr2); (3) patients with negative prostate biopsy (Ctr3). The authors used primer sets to identify fragments of 1.3, 1.8, and 2.4 kb from four genomic loci by qPCR. Each of the twelve resulting amplifications (3 fragments for four loci) was assigned a score of 0 (qPCR negativity) or 1 (qPCR positivity). The cfDI was expressed as the sum of the individual scores and ranged from 0 to 10 (2 conditions were excluded as they did not amplify within the 95% error range). The results showed that PCa patients had a significantly higher cfDI compared to Ctr1 and Ctr2, but not compared to Ctr3. A cfDI cut-off of 7 identified 89 of 123 PCa patients, corresponding to a sensitivity of 69.9%. In addition, in another group of 30 PCa patients with negative age-adjusted PSA, this cut-off identified 19 out of 30 PC (63.3%) that were not detected by PSA analysis . To our knowledge, this is the first article to report promising results on cfDI as a biomarker for PCa. In 2013, Feng et al. investigated cfDI as a biomarker to differentiate between PCa patients and BPH patients. In this case, the authors quantified the plasma levels of ALU 247 bp and 115 bp by qPCR and expressed cfDI as the ratio ALU 247/115. The results showed a significantly higher ALU 247/115 ratio in PCa patients than in BPH patients. This significant difference also persisted in PCa and BPH patients with a PSA level of more than 4 ng/mL. In addition, cfDI showed a sensitivity of 81.7% and a specificity of 78.8% in differentiating PC from BPH with PSA ≥ 4 ng/mL (AUC = 0.910) . Casadio et al. investigated the diagnostic value of cfDI by analysing urine samples from PCa patients and HC. The authors amplified sequences longer than 250 bp of the oncogenes c-MYC, BCAS1, and HER2 by qPCR and expressed the cfDI as the sum of the three resulting quantifications. The analysis revealed a significantly higher cfDI value in the urine of PCa patients compared to HC. At a cut-off value of 0.04 ng/µL, the cfDI value in urine showed an AUC of 0.80 (sensitivity: 0.79; specificity: 0.84), indicating a good accuracy of this value in differentiating PCa from HC . These results emphasise that urine is a valuable alternative liquid biopsy for PCa screening compared to plasma and serum. Similarly, Salvi et al. investigated the efficiency of cfDI in urine to differentiate PCa from benign diseases of the urogenital tract (BDUT). The design of the cfDI analysis was the same as Casadio et al. , with the only difference that the oncogenes considered were c-MYC, AR, and HER2. The analysis showed no significant difference between PCa and BDUT. Furthermore, the ROC curve analysis showed a lower AUC for cfDI in urine than for PSA (0.50 vs. 0.84) , suggesting that the latter is the better choice to distinguish PCa from BDUT. Another comprehensive analysis by Fawzy et al. compared cfDI, expressed as the ALU 247/115 ratio, between PCa, BHP, and HC. The qPCR analysis revealed a significantly higher cfDI value for the PCa group than for the other groups. The PCa group was further subdivided into metastatic PCa (MPC) and non-metastatic PCa (nMPCa). In this case, the cfDI value was slightly but significantly lower in MPC than in nMPCa . The diagnostic value of cfDI in differentiating PCa from BHP and HC was further investigated by Khani et al. in an Iranian cohort of PCa and BHP patients and HC subjects. Here, too, the ALU 247/115 ratio was determined as a cfDI parameter. The analysis revealed a higher cfDI in PCa patients compared to BHP patients and HC, confirming the results of Fawzy et al. Interestingly, there was no difference in cfDI between the MPCa patients and the nMPCa patients . Arko-Boham et al. compared the diagnostic value of cfDI in serum in two different hormone-related cancers: PCa and BC. In PCa, the authors analysed the ALU 247/115 ratio in the serum of PCa patients and HC. The results again showed a significantly higher ratio in PCa patients compared to HC. In addition, the cfDI correlated positively with high tumour stage and stage . The most recent article investigating cfDI in PCa was published in 2020 by Condappa et al. , who analysed cfDI in the plasma of PCa and BHP patients. As usual in recent years, cfDI was assessed as an ALU 247/115 ratio. The results showed no difference in cfDI between PCa and BHP patients, regardless of the results of other similar studies . This discrepancy could be explained by the smallest analysed patient pool of the studies discussed so far (11 PCa and 9 BHP).
The following section summarises the cfDI discoveries published in the literature in chronological order ( ). In the table, the average age of the cohorts was rounded to decimal numbers. Main results refer to significant results; if no significant results are available, a trend is indicated as non-significant. Clinical value refers to the potential clinical application if significant results are demonstrated. For most studies, qPCR determination of target molecules for cfDI was used, and only methods deviating from qPCR are mentioned in the text. The first pioneering work was that of Umetani et al. in 2006, who investigated the prognostic ability of cfDI in serum to predict the progression of BC. For this purpose, the authors analysed ALU 260/111 in 83 BC and 51 healthy control women (HC). The results showed that cfDI levels were significantly higher in AJCC stage II, III, and IV BC than in healthy women. They also demonstrated a linear correlation between tumour stage and size, with the ROC curve allowing differentiation between stage II to IV BC and HC (AUC = 0.79). Their data show that cfDI can predict preoperative lymph node metastases . Deligezer et al. then used ALU 247/115 to investigate whether this target could mark the effects of chemotherapy in BC patients. They found that this cfDI did not differ in patients before and after chemotherapy but was significantly higher in BC than in HC . In 2012, Agostini et al. analysed preoperative plasma cfDI as a biomarker for regional lymph node metastases in BC. For this purpose, the authors analysed ALU 247/115 in a small sample of BC patients and healthy women. The results showed that cfDI levels were higher in BC patients than in healthy controls. In addition, cfDI showed a moderate ability to detect patients with lymph node metastases . However, Lehner et al. demonstrated that cfDI of ALU 247/115 was not related to the response to neoadjuvant treatment in BC patients , but in 2014 Stötzer et al. investigated cfDI as a diagnostic biomarker for BC progression in a large sample population of BC patients. To this end, the authors investigated ALU 247/115 in localised BC, metastatic breast cancer (MBC), benign breast disease (BBD), and healthy female controls (HC). They used two different cfDI determinations: cfDI 1 as the result of the ratio of the qPCR amount of fragments and cfDI 2 as the ratio of qPCR Cp as the ΔΔCp formula. The results showed that cfDI 1 was significantly higher in HC than in BBD and higher in BC and MBC than in BBD. The cfDI 2 confirmed higher values in HC than in BBD and was higher in MBC than in local BC and BBD. However, both cfDI values were not suitable to identify BC compared to HC. On the contrary, it should be noted that the concentrations of ALU fragments allowed a better differentiation between BC and HC (AUC > 95%) and that for MBC, the classical CA 15-3 and CEA had the best diagnostic value . Interestingly, during the same year, Madhavan et al. evaluated cfDI as an early diagnostic biomarker in a large cohort of primary BC (PBC) and MBC patients, using other targets. The authors studied ALU 260/111 and LINE1 266/97 in a large cohort of MBC, PBC, and healthy females (HC). In contrast to the other two groups, the results obtained from this group showed that cfDI was significantly lower in PBC patients than healthy controls. The ROC curves showed a discriminating ability to distinguish HC from PBC (AUC = 0.75), HC from CTC-positive MBC (AUC = 0.93), and HC from CTC-negative MBC (AUC = 0.81). Also, discrimination ability was found for PBC from CTC-negative and CTC-positive MBC (AUC = 0.71 and 0.86, respectively) and for CTC-negative MBC from CTC-positive MBC (AUC = 0.93). In MBC the lowest level was shown in CTC-positive MBC, and the cfDI correlated with worse PFS and OS too . Iqbal et al. evaluated cfDI as a prognostic biomarker for BC prediction. For this purpose, the authors investigated ALU 247/115 in PBC and healthy females. The results contemplated that cfDI was significantly higher in PBC patients than HC. Interestingly, tumour size and cfDI together showed a marked association with disease-free survival (DFS) . Kamel et al. investigated the diagnostic potential of cfDI in BC using β-actin fragments (400 bp and 100 bp) in a large group of BC patients at different stages prior to therapeutic intervention. They found cfDI β-actin 400/100 was significantly higher in BC patients with respect to benign disease patients (BBD) or matched HC. The ROC curves of cfDI showed 95% CI for discriminating BC from BBD or HC. Also, cfDI was shown to increase with the increase in BC stage, although it did not relate with histopathological type or grade . Maltoni et al. investigated the role of cfDI as a prognostic biomarker using different targets such as BCAS 1 266/129, MYC 264/128, and PIK3CA 274/129 in relapsed BC, non-relapsed BC, and HC. The results showed that cfDI of MYC and of PIK3CA were significantly lower in patients than in HC, but no significant results were obtained in distinguishing between relapsed and non-relapsed BC patients . Wang et al. investigated the cfDI of ALU 260/111 in non-metastatic BC compared to patients with benign breast disease. They found that cfDI was significantly lower in BC patients than in benign controls. In addition, they found that the cfDI of ALU 260/111 had a ROC curve with an AUC of 0.67 and was a better biomarker than CTCs, cfDNA concentration, and CEA153. Interestingly, cfDI and CTCs together improve sensitivity and specificity and reduce the false positivity of cfDI alone from 50% to 10.7% . Cheng et al. showed that cfDI was lower in a large group of recurring BCs compared to a non-recurring BC group. To this end, the authors analysed ALU 260/111 and LINE1 266/97 and showed that the cfDI of ALU, LINE1, or both in the ROC had an AUC of 0.71, 0.704, and 0.732, respectively. These data also suggested that cfDI was an independent marker of recurrence in BC . Cheng et al. then investigated cfDI as a biomarker for response to therapy by measuring ALU 260/111 and LINE1 266/97 in a large group of MBC patients. The results showed an increase in cfDI ( p = 1.21 × 10 −7 for ALU and p = 1.87 × 10 −3 for LINE1) after one cycle of therapy in MBC patients. Of note, cfDI was an independent prognostic marker . Then, Tang et al. investigated the function of cfDI as a diagnostic biomarker in BC by ALU 247/115 in small sample groups of BC, BBD, and HC. The results showed that cfDI was significantly higher in BC than in BBD and HC (cfDI did not differ between BBD and HC). The ROC curve showed an AUC of 0.97, and cfDI correlated with lymph node metastasis and tumour stage . Salimi et al. investigated the role of cfDI biomarkers for BC progression in triple-negative breast cancer (TNBC). The authors analysed the ratio of β-actin 394 bp/ β-actin 99 bp in TNBC and non-TNBC patients as well as in healthy women. The results showed that cfDI was significantly higher in TNBC and non-TNBC patients than in healthy controls and was associated with lympho-nodal metastasis and tumour stage. In particular, the ROC curve showed an AUC = 0.997. Finally, the cfDI significantly increases in TNBC compared to non-TNBC, thus discriminating between the two BC types . Wang et al. investigated whether cfDI of ALU 260/111 and LINE1 266/97 can predict response to neoadjuvant chemotherapy (NACT) in a small group of BC patients. The results showed that cfDI was significantly higher in patients after NACT than in patients before NACT. Interestingly, the increase in cfDI in NACT-treated patients was associated with tumour shrinkage and a reduction in Ki67 levels . Miao et al. came to the opposite conclusion by using a cfDI determined by LINE1 259/97 in a study population of young BC patients, most of whom had high TNM and tumour size and were undergoing adjuvant chemotherapy. They found that cfDI was higher in BC patients before chemotherapy than in BBD or HC. Then, they proved that cfDI decreases after chemotherapy, but they collected blood samples 4 weeks after the completion of chemotherapy . Arko-Boham et al. investigated the role of cfDI of ALU247/115 as a biomarker to identify BC and prostate cancer compared to healthy individuals in small populations. The results showed that cfDI levels were lower in BC than in HC but without statistical significance. In contrast, a statistically higher cfDI value was found in stage II BC . Hussein et al. investigated the role of cfDI ALU 247/115 as a diagnostic and prognostic biomarker in a small group of BCs. The results showed that ALU 247/115 was significantly higher in BC patients than in HC with a ROC AUC = 0.825, but it could not be differentiated between metastatic and non-metastatic patients . Park et al. investigated the role of cfDI of ALU 263/58 as a diagnostic biomarker in BC patients and HC. It was found that the ALU-derived cfDI was significantly higher in patients compared to controls with an AUC of 0.724. However, the methylation status of the LINE1 target provided a better test to differentiate cases from healthy controls . Lamminaho et al. analysed cfDI in a large cohort of non-metastatic BC patients at the time of diagnosis. The cfDI was determined by the ratio of cfDNA fragments greater than 1000 bp to those with less than 1000 bp, measured by electrophoretic separation using the ScreenTape D5000 system. The results showed that a higher cfDI was correlated with significantly poorer survival, but only at a follow-up of 10 years. They showed that a high cfDI was an independent prognostic factor for OS and Breast Cancer Specific Survival (BCSS). This was seen in ER + BC patients in the multivariate analysis. Interestingly, a multivariate logistic regression analysis with cfDI, tumour characteristics, and age at diagnosis strongly improved the predictive results . Adusei et al. investigated in a small cohort of BC patients whether cfDI can be used to predict response to chemotherapy using ALU 247/115. The results showed that cfDI increased after the third cycle of chemotherapy (T2) compared to the second cycle (T1), but without statistical significance. No statistical significance was also found in cfDI of BC cases compared to HC . Cirmena et al. used the quantification of electrophoretic fragments of cfDNA from plasma to evaluate cfDI as a biomarker for response to neoadjuvant chemotherapy (NACT). To this end, the authors used the highly sensitive D1000 Screen Tape Station to measure fragments of 321 to 1000 bp versus fragments of 150 to 220 bp in a small sample of BC patients and HC. The results showed that at time 2 after NACT treatment, cfDI was significantly higher in pathological complete responders than in non-complete responders. Of note, cfDI with magnetic resonance imaging data analysed by ROC showed a predictive value for complete response of 0.875 . In an observational study, Elhelaly et al. investigated the role of cfDI ALU 247/115 as a biomarker for the early detection of BC and the discrimination of BBD. In their cohort of age-matched cases and controls, cfDI was significantly higher in BC than in BBD or control, but it did not discriminate between BBD and HC. The ROC showed an AUC of 0.727 . Hafeez et al. investigated cfDI as a prognostic biomarker for BC. To this end, the authors analysed ALU 247/115 in a small population study of BC, BBD, and healthy women. The results showed that cfDI was significantly higher in BC and BBD than in HC. cfDI was higher in early-stage BC and advanced BC than in BBD, but without statistical significance. To distinguish BC from HC, the ROC of cfDI showed an AUC = 0.71 . Then, Nair et al. analysed cfDI of ALU 247/115 as a prognostic factor in 167 BC patients with different molecular subtypes and stages of disease. They found that cfDI was significantly higher in preoperative BC patients than in postoperative patients, and interestingly, higher cfDI was associated with a mild immune infiltrate and thus a poorer prognosis. This evidence was confirmed by the negative correlation with DFS: DFS decreased from 82 months to 58 months in BC patients with high cfDI values. However, the authors emphasised that higher ALU 247 bp levels may be an independent factor for BC prognosis . In contrast to other groups using qPCR, Bortul et al. investigated cfDI with droplet digital PCR (ddPCR) in a large cohort of BC in 2023. For this purpose, they detected ALU 260/111 and LINE-1 266/97 with ddPCR in 106 BC and 103 HC. The results showed that both cfDI were significantly lower in BC compared to HC. In fact, cfDI LINE1 proved to be more accurate in discriminating cases from controls in ROC (AUC = 0.77) . In 2024, Celik et al. analysed for the first time ccfDNA levels, mtDNA, and DNA integrities together in a small sample of BC patients undergoing NACT compared to healthy controls. The authors investigated whether cfDI of ALU 247/ALU115 and mitochondrial nicotinamide dinucleotide adenine dehydrogenase 4 and 1 (ND4/ND1) can be used to predict response to chemotherapy alone or together with other biomarkers. cfDI ALU 247/ALU 115 was not significantly different in patients before treatment compared to controls but was significantly lower in patients after treatment compared to controls. In contrast, the copy number of mtDNA (ND4/ND1) was significantly higher in patients than in controls both before and after treatment . Also, in 2024, Giro et al. investigated cfDI ALU 247/ALU 115 in a small sample of BC patients 15 days after neoadjuvant chemotherapy. They found that cfDI was significantly higher in patients who achieved a pathological complete response (pCR) and correlated significantly with disease-free survival (DFS) . Finally, Gameel et al. investigated the potential of cfDI ALU 247/115 in a small sample population of BC, BBD, and HC as a biomarker for predicting recurrence. Although cfDI was higher in BC patients than in HC, the differences were not statistically significant .
The efficiency of cfDI as a tumour biomarker was also investigated in prostate cancer (PCa), but the number of articles was lower than in BC (9 and 25, respectively). Unfortunately, the standard screening method using PSA analysis has too high sensitivity and too low specificity, which poses a diagnostic challenge as it may lead to overdiagnosis and treatment of latent cancer . Therefore, efforts are focused on the evaluation of other biomarkers such as cfDI, the studies of which are summarised in chronological order in this section ( ). Analogously to the BC section, in the table, the average age of the cohorts was rounded to decimal numbers. Main results refer to significant results; if no significant results are available, a trend is indicated as non-significant. Clinical value refers to the potential clinical application if significant results are demonstrated. All studies used qPCR to quantify the longer and shorter fragments for cfDI determination. In 2006, Boddy et al. investigated cfDI for the first time as a biomarker to differentiate PCa patients from patients with benign prostatic hyperplasia (BPH). They quantified the plasma levels of 356 bp and 105 bp sequences of the leptin gene (LEP) in PCa and BPH patients. Leptin is a pleiotropic peptide hormone secreted by adipose tissues and has been associated with various cancers related to obesity . The authors expressed the cfDI as ∆Ct between the average Cts of LEP 356 bp over 105 bp fragments. This ratio value showed no significant difference between PCa and BPH groups . In the same year, Hanley et al. compared the plasma cfDI of PCa patients with those of three different control groups: (1) healthy volunteers (HC) under 40 years of age (Ctr1); (2) patients with radical prostatectomy and low PSA six months after surgery (Ctr2); (3) patients with negative prostate biopsy (Ctr3). The authors used primer sets to identify fragments of 1.3, 1.8, and 2.4 kb from four genomic loci by qPCR. Each of the twelve resulting amplifications (3 fragments for four loci) was assigned a score of 0 (qPCR negativity) or 1 (qPCR positivity). The cfDI was expressed as the sum of the individual scores and ranged from 0 to 10 (2 conditions were excluded as they did not amplify within the 95% error range). The results showed that PCa patients had a significantly higher cfDI compared to Ctr1 and Ctr2, but not compared to Ctr3. A cfDI cut-off of 7 identified 89 of 123 PCa patients, corresponding to a sensitivity of 69.9%. In addition, in another group of 30 PCa patients with negative age-adjusted PSA, this cut-off identified 19 out of 30 PC (63.3%) that were not detected by PSA analysis . To our knowledge, this is the first article to report promising results on cfDI as a biomarker for PCa. In 2013, Feng et al. investigated cfDI as a biomarker to differentiate between PCa patients and BPH patients. In this case, the authors quantified the plasma levels of ALU 247 bp and 115 bp by qPCR and expressed cfDI as the ratio ALU 247/115. The results showed a significantly higher ALU 247/115 ratio in PCa patients than in BPH patients. This significant difference also persisted in PCa and BPH patients with a PSA level of more than 4 ng/mL. In addition, cfDI showed a sensitivity of 81.7% and a specificity of 78.8% in differentiating PC from BPH with PSA ≥ 4 ng/mL (AUC = 0.910) . Casadio et al. investigated the diagnostic value of cfDI by analysing urine samples from PCa patients and HC. The authors amplified sequences longer than 250 bp of the oncogenes c-MYC, BCAS1, and HER2 by qPCR and expressed the cfDI as the sum of the three resulting quantifications. The analysis revealed a significantly higher cfDI value in the urine of PCa patients compared to HC. At a cut-off value of 0.04 ng/µL, the cfDI value in urine showed an AUC of 0.80 (sensitivity: 0.79; specificity: 0.84), indicating a good accuracy of this value in differentiating PCa from HC . These results emphasise that urine is a valuable alternative liquid biopsy for PCa screening compared to plasma and serum. Similarly, Salvi et al. investigated the efficiency of cfDI in urine to differentiate PCa from benign diseases of the urogenital tract (BDUT). The design of the cfDI analysis was the same as Casadio et al. , with the only difference that the oncogenes considered were c-MYC, AR, and HER2. The analysis showed no significant difference between PCa and BDUT. Furthermore, the ROC curve analysis showed a lower AUC for cfDI in urine than for PSA (0.50 vs. 0.84) , suggesting that the latter is the better choice to distinguish PCa from BDUT. Another comprehensive analysis by Fawzy et al. compared cfDI, expressed as the ALU 247/115 ratio, between PCa, BHP, and HC. The qPCR analysis revealed a significantly higher cfDI value for the PCa group than for the other groups. The PCa group was further subdivided into metastatic PCa (MPC) and non-metastatic PCa (nMPCa). In this case, the cfDI value was slightly but significantly lower in MPC than in nMPCa . The diagnostic value of cfDI in differentiating PCa from BHP and HC was further investigated by Khani et al. in an Iranian cohort of PCa and BHP patients and HC subjects. Here, too, the ALU 247/115 ratio was determined as a cfDI parameter. The analysis revealed a higher cfDI in PCa patients compared to BHP patients and HC, confirming the results of Fawzy et al. Interestingly, there was no difference in cfDI between the MPCa patients and the nMPCa patients . Arko-Boham et al. compared the diagnostic value of cfDI in serum in two different hormone-related cancers: PCa and BC. In PCa, the authors analysed the ALU 247/115 ratio in the serum of PCa patients and HC. The results again showed a significantly higher ratio in PCa patients compared to HC. In addition, the cfDI correlated positively with high tumour stage and stage . The most recent article investigating cfDI in PCa was published in 2020 by Condappa et al. , who analysed cfDI in the plasma of PCa and BHP patients. As usual in recent years, cfDI was assessed as an ALU 247/115 ratio. The results showed no difference in cfDI between PCa and BHP patients, regardless of the results of other similar studies . This discrepancy could be explained by the smallest analysed patient pool of the studies discussed so far (11 PCa and 9 BHP).
Liquid biopsy is a valid, non-invasive, or minimally invasive method that uses biological fluids to identify biomarkers of tumour development and growth, clonal formation at the site of metastasis and metastases, prognosis of DSF and OS, and can be helpful in predicting response to therapy. Despite its potential, liquid biopsy faces challenges, including the standardisation of detection methods and the validation of clinical utility. Ongoing research aims to integrate liquid biopsy into the routine treatment of BC and PCa patients to improve early detection, surveillance, and personalised therapy . Various studies have quantified total cfDNA levels with the following target genes in both BC and PCa: β- globin , β- 2 Microglobin , GAPDH , hTERT , ALU, or LINE1. Higher levels of the measured cfDNA could be used to distinguish benign from malignant BC [ , , , , ]. It is worth noting that changes in cfDNA levels can be altered by other pathological conditions such as inflammation and infection, which may influence the results . In addition, diurnal variations in individual cfDNA levels have been demonstrated in both healthy individuals and those with cancer . In contrast, cfDI, which is a ratio between longer and shorter fragments of a target, is in principle less affected by the variability of cfDNA quantity from sample to sample. Therefore, the comparison between groups could be more reliable. Of course, the choice of target also has an important influence. Certainly, with respect to ctDNA mutation targets, the cfDI can overcome limitations related to the mutational rate. In addition, the cfDI could provide an incredible advantage also in terms of sensibility by using repetitive elements as targets, such as ALU and LINE1, that represent about 10 and 17% of the genome, respectively . Here, the current literature on the importance of cfDI in BC and PCa is analysed, and patients are grouped based on the studies examining the nature of the repetitive elements or target genes that determine cfDI. A graphical summary of the literature results in BC can be found in . In BC, the majority of studies addressed the cfDI of repetitive short interspersed nuclear element (ALU) and long interspersed nuclear element (LINE1) sequences, which accounted for 79% of all studies. Regarding the cfDI of ALU 260/111 as a diagnostic biomarker, the first study by Umetani showed an increase in serum cfDI in BC at stages from II to IV, accounting for 51 BC patients compared to 51 HC . On the contrary, two other studies on plasma of a large sample population of 188 primary BC and 185 HC showed opposite results with a cfDI decrease in BC compared to HC, confirmed also in 201 MBC compared to HC . The discrepancy between the work of Umetani et al. and that of Madhavan et al. and Bortul et al. could be due to differences in the type of samples (serum versus plasma) and in the number and clinicopathological conditions of the cohorts. Consistent with the decrease in cfDI in BC compared to HC, recurrent BC patients had a lower cfDI than non-recurrent BC patients . Interestingly, cfDI ALU 260/111 had prognostic value in terms of tumour stage, metastatic disease, PFS, and OS [ , , ]. Finally, a higher cfDI was observed in MBC patients after one cycle of chemotherapy and in BC patients with complete response to NACT compared to no-responder patients as a biomarker for the success of the therapy, which measures the increase in circulating DNA due to the death of cancer cells. Differently from ALU 260/111, the cfDI ALU 247/115 showed concordant results in all studies where it was found higher in 522 BC or MBC than in 381 HC or breast benign disease (BBD) [ , , , , ]. A prognostic role of cfDI ALU 247/115 was found to be in accordance with the diagnostic findings for tumour stage and grade, metastasis, and disease-free survival [ , , , , , , , ] and a predictive role for complete response to NACT or chemotherapy [ , , ]. It is noteworthy that no differences were found in studies using serum or plasma. A study, in which other targets were used in ALU sequences (ALU 263/58), also showed a higher cfDI in BC . LINE1 as a target for cfDI is less studied than ALU, but cfDI LINE1 266/97 showed potential for early detection and screening of BC , recurrence , and response to therapy . Overall, all studies confirmed a lower cfDI in BC at the time of diagnosis compared to HC or at the time of disease recurrence compared to non-recurrence and interestingly showed an increase in cfDI in BC responders to therapy compared to non-responders. The cfDI LINE1 259/97 gave opposite results: The cfDI was higher in BC than in HC or BBD, and consistent with this, it decreased after adjuvant therapy in BC patients who responded to therapy . This raises puzzling questions about these different results when repetitive sequences are used as targets for cfDI. The overall impression is that the target region and the length of the amplified fragment might be the reason for these differences. For example, with more bp in longer fragments, more variability in the results can be attended to because of the susceptibility of circulating DNA to fragmentation and degradation in the pre-analytical phases. In the studies where specific genes were the target for cfDI at diagnosis, there was also heterogeneity in the results: the cfDI of β-actin 400/100 was higher in BC than in HC , and that of β-actin 394/99 was higher in TNBC than in non-TNBC . However, when MYC or PIK3CA targets were used, the cfDI was lower in BC than in HC . Finally, emerging evidence of the cfDNA fragmentomic is confirming cfDI as a prognostic and predictive biomarker . A graphical summary of the literature results in PCa can be found in . In the case of PCa, it is initially noticeable that the number of studies on PCa ( n = 9) is very low compared to BC , some of which were not significant ( n = 3). This is not surprising, as PCa, despite its high incidence in men, is a cancer that affects older people, often as an indolent form, and it is only in the last decade that the interest in early detection of aggressive cancer, active surveillance of localised cancer, and follow-up of patients with castration-resistant forms has increased. Of particular interest are metastatic castration-resistant PCa patients, for whom new targeted therapy options are available to control the disease . Of note, most studies in PCa are based on the analysis of ALU repeat sequences. Interestingly, results of cfDI of ALU 247/115 confirm the same trend found in BC: cfDI was higher in PCa than in HC or benign hyperplasia (BPH) [ , , ], confirmed also by specific gene analysis . In addition, a study using fragmentomic analysis supported a higher cfDI in PCa than in HC . Finally, we would like to point out that in PCa it is possible to analyse cfDI in urine and not in blood, so this test is without any risk .
The growing therapeutic scenario in the clinical management of BC and PCa patients raises the question of the utility of integrating cfDI analysis into clinical practice. The majority of the studies showed that cfDI differs between patients and controls, emphasising the potential usefulness of this biomarker. Compared to NGS technologies, cfDI has the advantage of being a more sensitive and cost-effective method that, in combination with other biomarkers, could help the clinical management of both BC and PCa patients. Data from the literature have shown that the potential clinical value of cfDI varies depending on the target. Therefore, identifying the best target for BC and PCa is a priority. In addition, the lack of standardised pre-analytical and analytical protocols for cfDI determination remains a limitation. This fact and the nature of the different targets analysed in terms of bp could have an impact on the different results regarding the increase or decrease in cfDI between patients and controls. In addition, for repetitive sequences, which are the most studied targets, there is a lack of robust evidence in large cohorts to deepen the power of cfDI in discriminating BC or PCa patients from HC and benign diseases. Further studies on the role of cfDI as a biomarker for disease progression and response to therapy are also needed. Overall, the results suggest that cfDI in BC could be used as one of the parameters together with other liquid biopsy biomarkers (e.g., cfDNA, ctDNA, CTCs, etc.) or classical biochemical, radiological, and/or anthropometric data to improve early diagnosis and screening programmes for specific populations (e.g., women with family history) for prognosis and predictive medicine. Since MBC can be cured as a chronic disease, there is also an interest in exploring the prognostic and predictive value of cfDI in MBC. In this regard, cfDI together with ctDNA mutations could be an important complementary tool to obtain relevant information on tumour progression and response to therapy. Very important is the potential of cfDI to differentiate between TNBC and non-TNBC patients, which emphasises the utility of cfDI. It is noteworthy that five studies in BC have found cfDI to be a dynamic biomarker for tracking response to therapy. They show that cfDI levels change during chemotherapy and weeks after the end of chemotherapy [ , , , , ]. An interesting question to address is whether cfDI could be a biomarker for response to immunotherapy and, in particular, to immune checkpoint inhibitors, as tumour mutational burden in ctDNA has been shown to be useful . Unfortunately, to our knowledge, there are no data on the efficacy of cfDI as a biomarker of response to immunotherapy in BC. A future challenge could be to investigate the usefulness of cfDI for predicting and monitoring response to immunotherapy in BC patients. This is justified by the fact that in colorectal cancer, cfDI measured at the β-actin gene (400 bp over 100 bp fragments) predicted response to immunotherapy in terms of PFS, which was higher in a subset of patients with low cfDI than in those with high cfDI levels . Waki et al. also showed that the cfDI of ALU 247/115 was related to vaccine response to vaccination in non-small cell lung cancer. In this case, a high cfDI value was associated with longer survival before and after vaccination. In contrast, in PCa, there is a paucity of evidence of the potential of cfDI as a diagnostic tool useful for surveillance of high-risk men or for the screening population. However, it should be noted that the majority of the studies analysed examined cfDI ALU 245/115, and the results are consistent with the results of the studies in BC. This encourages further investigations based on BC studies also in PCa patients. Studies on cfDI as a dynamic biomarker in metastatic castration-resistant prostate cancer to monitor the effect of new targeted therapies could be of great interest. Molecular diagnosis prognosis and therapy response by liquid biopsy is an increasing need for the best treatment of BC and PCa patients. Efforts to standardise analysis through precise pre-analytical protocols and the use of high-quality and cost-effective technologies such as ddPCR in conjunction with large multicentre studies are needed to define the potential of cfDI as a biomarker for BC and PCa clinical practice.
|
Feasibility study of AI-assisted multi-parameter MRI diagnosis of prostate cancer | cea97104-7ed9-4f73-af7e-40151dc7480d | 11950164 | Neoplasms[mh] | Prostate cancer (PCa) is the second most common malignancy in men globally, accounting for over one-fifth of male cancer diagnoses and posing significant challenges to healthcare systems , . Early treatment can reduce mortality, prompting widespread prostate cancer screening . Consequently, prostate cancer screening has been widely promoted and implemented over the past decades , . Currently, prostate biopsy is the standard diagnostic method, but its invasive nature raises concerns about detecting small tumors, leading to overtreatment and psychological distress. Thus, optimizing biopsy protocols to improve detection while minimizing unnecessary procedures has become a critical research focus, aiming to alleviate patient burden and conserve resources. Over the past decades, researchers have developed various computer-aided diagnosis (CAD) systems to classify malignant and benign prostate lesions – . However, their effectiveness depends on accurate segmentation and optimal image selection. Until these challenges are addressed, the performance of traditional CAD systems remains questionable. In recent years, rapid advancements in deep learning technology have led to the development of novel CAD systems for prostate cancer diagnosis , . Deep learning methods can simplify the classification of prostate lesion images and automatically learn complex patterns, reducing manual intervention, provided the dataset is sufficiently large , , . Among various deep learning models, convolutional neural networks (CNNs) are the most widely used in medical imaging due to their simplicity and effectiveness , . It has been applied to in other cancer diagnoses to classify benign and malignant cases , , . Compared to the current popular algorithms such as U-Net or DenseNet, the ResNet50 architecture has better performance in alleviating the gradient disappearance problem and achieving deeper network training, which can better capture the detailed features in the image . The U-Net model features a symmetric encoder-decoder structure with downsampling (contracting path) and upsampling (expansive path) parts . It employs skip connection to combine the features of different layers and performs well on small sample datasets, which can effectively capture context information. However, it may consume more memory when processing larger images, making it less suitable for MRI image processing. The DenseNet model uses dense connections, where each layer is connected to all previous layers, resulting in a dense feature map . The DenseNet model is widely used in image classification, object detection, and segmentation due to its ability to reduce parameters, improve gradient flow, and enhance gradient reuse. However, its high complexity makes training time-consuming and less accessible for clinicians without a background in convolutional networks. In contrast, ResNet50, a deep residual network, uses residual learning to address training challenges. By introducing skip connections, it alleviates the vanishing gradient problem, allowing deeper networks (e.g., 50 layers) without compromising performance, particularly in image classification tasks like those on the ImageNet dataset. Additionally, the multi-head attention mechanism enhances feature representation by emphasizing key details and reducing redundancy in object recognition tasks . This technique has rarely been applied in developing deep transfer learning models for medical image classification, particularly for distinguishing benign from malignant prostate lesions in MRI images. We designed a study that integrates a multi-head attention mechanism into the ResNet50 model, leveraging the strengths of both techniques to enhance CAD system performance in classifying prostate MRI lesions. This study aims to validate this hypothesis, with additional details provided in the following sections.
Features were extracted for each modality (T2, DWI, and DCE) using separate pre-trained ResNet50 models. ResNet50 is a deep residual network that addresses vanishing gradients and explosions in deep neural networks through residual blocks. It consists of a 50-layer convolutional architecture with batch normalization and ReLU activation, enabling high-level feature extraction . In this study, we removed the classification layers of ResNet50, retaining only the feature extraction component, which captures rich representations from input images, independent of specific classification tasks. The input data passes through a 7 × 7 convolutional layer, a 3 × 3 max pooling layer, and four residual modules with multiple convolutional layers and skip connections. This architecture enables the extraction of both local and global features, producing a 2048-dimensional feature vector that encapsulates high-level semantic information, including shape, texture, and edge features, which are essential for subsequent classification tasks. The same ResNet50 architecture is consistently used for processing T2, DWI, and DCE images. The multi-head attention mechanism establishes associations between different modal features and captures dependencies to enhance classification performance. While traditional CNNs effectively extract features from single modalities, they may not fully utilize complementary information in multimodal data . This mechanism improves the mining and fusion of information by weighting combinations of modal features. In the specific implementation, the 2048-dimensional feature vector extracted from ResNet50 is input into the multi-head attention mechanism. The multi-head attention mechanism comprises multiple parallel attention heads, each independently calculating attention weights and weighted feature representations. For each modality’s feature vector, the attention mechanism first computes the similarity with the feature vectors of other modalities to generate the attention weight matrix. These weights are then used to compute a weighted sum of the feature vectors, resulting in the weighted feature representation. The parallel computation of multiple attention heads captures feature associations across different dimensions, generating richer and more diverse representations. The outputs of all attention heads are concatenated to form a comprehensive feature vector, containing independent information from each modality while fusing complementary information. This processed feature vector is then concatenated with the original feature vectors from the three modalities, creating a high-dimensional representation that encompasses multimodal information. The parallel computation of multiple attention heads captures feature associations across different dimensions, generating richer representations. The outputs from all heads are concatenated to form a comprehensive feature vector, which contains independent information from each modality while fusing complementary data. This vector is then combined with the original feature vectors from the three modalities, creating a high-dimensional representation that encompasses multimodal information.
This study adheres to the Helsinki Declaration (revised in 2013) and does not require ethical approval as the data we utilized was derived from previous image database of the hospitals. In the light of the study’s retrospective nature, the requirement for informed consent was waived. Datasets description In this study, we collected mp-MRI images corresponding to pathological sections from 106 prostate cancer cases, including T2-weighted (T2WI) and diffusion-weighted imaging. Our inclusion criteria included: (1) undergoing prostate biopsy; (2) Multiparametric magnetic resonance imaging was performed before prostate biopsy. Exclusion criteria: (1) patients who had undergone prostate biopsy within the first 6 weeks of MRI scan; (2) after local prostate surgery, or the patient has received radiotherapy and chemotherapy; (3) Multiple metastasis of prostate cancer in pelvic cavity and obvious destruction of local glands; (4) Patients were pathologically diagnosed with bladder cancer, sarcoma, or other types of malignant tumors that had metastasize to the prostate; (5) Foreign bodies, gas, motion and other artifacts cause poor quality of MRI images, or incomplete sequences. A total of 137 mp-MRI images were gathered, resulting in 274 groups of ROI data from three sequences of prostate mp-MRI images. Of these, 206 sets are designated for training and validation, while 68 sets are reserved for testing. The dataset comprises images from three modalities, each associated with a corresponding ROI (region of interest). The dataset is divided into training, validation, and test sets. A total of 206 samples were used for the training and validation sets, divided in a 6:4 ratio: 124 for training and 82 for validation. The 68 samples in the test set are used to evaluate the model’s generalization performance. Informed written consent was obtained from the patient for the publication of this report and any accompanying images. Datasets-prostate segmentation After denoising the three sequences of prostate MRI images, the regions of interest (ROIs) must be further extracted for subsequent input into the neural parameters. In this study, two senior pathologists reviewed the selected pathological sections and annotated the tumor area, prostate capsule, and transition zone boundary. If the pathologists disagree on the image interpretation, they will discuss their findings to reach a consensus. ITK-Snap software was utilized to open the original three sequences of prostate MRI images from selected patients (exported from DICOM format). MRI images of T2WI, DWI, and DCE sequences were collected at intervals of 5–8 mm from the tip of the prostate to the bladder neck and saved. Corresponding prostate specimens were prepared as axial pathological sections at 5–8 mm intervals from the tip of the prostate to the bladder neck. Two senior pathologists independently marked the tumor area, benign area, and transition zone boundary at each level after reviewing the sections, and these annotations were stored in the computer for later use.Both the imaging and pathological sections contain intrinsic markers (e.g., stones, cysts, nodules) that assist in matching them. Validation enables easy identification of corresponding levels in the pathological and imaging sections. Transparent mapping technology is used to virtually overlay the pathological sections onto the three axial MRI sequences at the same level on the computer, with corresponding benign and tumor tissue regions labeled on each sequence. Details of the experiments The primary goal of the experiment is to enhance the accuracy and robustness of the classification task by integrating multimodal image data. The pre-trained ResNet50 model serves as a feature extractor, capturing high-level features from the images of each modality. The Multihead Attention mechanism fuses the features from different modalities and captures their correlations. Finally, classification is performed using the Fully Connected Layer to produce predictions of benign or malignant lesions. Neural network architecture Neural network architecture diagram (Figs. – ), including model construction, training, validation, and testing. Data preprocessing and loading Data augmentation techniques, including random cropping, rotation, and flipping, are applied to the training data to enhance the model’s robustness and generalization ability. The images are resized to a uniform dimension of 224 × 224 pixels and normalized to ensure their mean and variance align with the input requirements of the pre-trained ResNet50 model. PyTorch’s DataLoader is utilized to batch load the training, validation, and test sets, thereby improving training efficiency. Model training The Binary Cross-Entropy Loss function (BCELoss), suitable for binary classification tasks, measures the difference between predicted results and true labels. The Adam optimizer is selected for its adaptive learning rate adjustment feature and quick convergence, with an initial learning rate set to 0.0001. In each epoch, the model performs forward propagation on the training data, computes the loss, backpropagates, and updates the parameters. Simultaneously, the model’s performance is evaluated on the validation set, calculating the validation loss and accuracy. The training and validation loss, along with the accuracy of each modality, are recorded for each epoch. Model evaluation and preservation At the end of each epoch, the training and validation loss values and accuracy are recorded and saved as CSV files. If the current validation accuracy is equal to or greater than the historical best value, and the accuracy of at least two of the three modalities—T2, DWI, and DCE—has improved, the optimal model is updated. Test steps Load the best model During the testing phase, the best model parameters saved during training are loaded to ensure that the best model is used for testing. Evaluate on the test set The model is evaluated using test set data, calculating the difference between predicted results and true labels. Various evaluation metrics are calculated on the test set, including the ROC curve, PR curve, accuracy, F1 score, and confusion matrix. The evaluation metrics for the three modalities—T2, DWI, and DCE—are calculated separately to analyze each modality’s performance in the classification task. Draw evaluation charts Draw and save ROC curves, PR curves, and confusion matrices. The prediction results of the total model and each modality are plotted and saved separately. Ethical approval This study has been approved by the Ethics Committee of the First Affiliated Hospital of Huzhou Normal University (20,180,017), and all methods were conducted in accordance with relevant guidelines and regulations. This study involved the use of patients’ tissue samples in the model, and all patients signed the specific informed consent forms collected.
In this study, we collected mp-MRI images corresponding to pathological sections from 106 prostate cancer cases, including T2-weighted (T2WI) and diffusion-weighted imaging. Our inclusion criteria included: (1) undergoing prostate biopsy; (2) Multiparametric magnetic resonance imaging was performed before prostate biopsy. Exclusion criteria: (1) patients who had undergone prostate biopsy within the first 6 weeks of MRI scan; (2) after local prostate surgery, or the patient has received radiotherapy and chemotherapy; (3) Multiple metastasis of prostate cancer in pelvic cavity and obvious destruction of local glands; (4) Patients were pathologically diagnosed with bladder cancer, sarcoma, or other types of malignant tumors that had metastasize to the prostate; (5) Foreign bodies, gas, motion and other artifacts cause poor quality of MRI images, or incomplete sequences. A total of 137 mp-MRI images were gathered, resulting in 274 groups of ROI data from three sequences of prostate mp-MRI images. Of these, 206 sets are designated for training and validation, while 68 sets are reserved for testing. The dataset comprises images from three modalities, each associated with a corresponding ROI (region of interest). The dataset is divided into training, validation, and test sets. A total of 206 samples were used for the training and validation sets, divided in a 6:4 ratio: 124 for training and 82 for validation. The 68 samples in the test set are used to evaluate the model’s generalization performance. Informed written consent was obtained from the patient for the publication of this report and any accompanying images.
After denoising the three sequences of prostate MRI images, the regions of interest (ROIs) must be further extracted for subsequent input into the neural parameters. In this study, two senior pathologists reviewed the selected pathological sections and annotated the tumor area, prostate capsule, and transition zone boundary. If the pathologists disagree on the image interpretation, they will discuss their findings to reach a consensus. ITK-Snap software was utilized to open the original three sequences of prostate MRI images from selected patients (exported from DICOM format). MRI images of T2WI, DWI, and DCE sequences were collected at intervals of 5–8 mm from the tip of the prostate to the bladder neck and saved. Corresponding prostate specimens were prepared as axial pathological sections at 5–8 mm intervals from the tip of the prostate to the bladder neck. Two senior pathologists independently marked the tumor area, benign area, and transition zone boundary at each level after reviewing the sections, and these annotations were stored in the computer for later use.Both the imaging and pathological sections contain intrinsic markers (e.g., stones, cysts, nodules) that assist in matching them. Validation enables easy identification of corresponding levels in the pathological and imaging sections. Transparent mapping technology is used to virtually overlay the pathological sections onto the three axial MRI sequences at the same level on the computer, with corresponding benign and tumor tissue regions labeled on each sequence.
The primary goal of the experiment is to enhance the accuracy and robustness of the classification task by integrating multimodal image data. The pre-trained ResNet50 model serves as a feature extractor, capturing high-level features from the images of each modality. The Multihead Attention mechanism fuses the features from different modalities and captures their correlations. Finally, classification is performed using the Fully Connected Layer to produce predictions of benign or malignant lesions.
Neural network architecture diagram (Figs. – ), including model construction, training, validation, and testing.
Data augmentation techniques, including random cropping, rotation, and flipping, are applied to the training data to enhance the model’s robustness and generalization ability. The images are resized to a uniform dimension of 224 × 224 pixels and normalized to ensure their mean and variance align with the input requirements of the pre-trained ResNet50 model. PyTorch’s DataLoader is utilized to batch load the training, validation, and test sets, thereby improving training efficiency.
The Binary Cross-Entropy Loss function (BCELoss), suitable for binary classification tasks, measures the difference between predicted results and true labels. The Adam optimizer is selected for its adaptive learning rate adjustment feature and quick convergence, with an initial learning rate set to 0.0001. In each epoch, the model performs forward propagation on the training data, computes the loss, backpropagates, and updates the parameters. Simultaneously, the model’s performance is evaluated on the validation set, calculating the validation loss and accuracy. The training and validation loss, along with the accuracy of each modality, are recorded for each epoch.
At the end of each epoch, the training and validation loss values and accuracy are recorded and saved as CSV files. If the current validation accuracy is equal to or greater than the historical best value, and the accuracy of at least two of the three modalities—T2, DWI, and DCE—has improved, the optimal model is updated.
Load the best model During the testing phase, the best model parameters saved during training are loaded to ensure that the best model is used for testing. Evaluate on the test set The model is evaluated using test set data, calculating the difference between predicted results and true labels. Various evaluation metrics are calculated on the test set, including the ROC curve, PR curve, accuracy, F1 score, and confusion matrix. The evaluation metrics for the three modalities—T2, DWI, and DCE—are calculated separately to analyze each modality’s performance in the classification task. Draw evaluation charts Draw and save ROC curves, PR curves, and confusion matrices. The prediction results of the total model and each modality are plotted and saved separately.
During the testing phase, the best model parameters saved during training are loaded to ensure that the best model is used for testing.
The model is evaluated using test set data, calculating the difference between predicted results and true labels. Various evaluation metrics are calculated on the test set, including the ROC curve, PR curve, accuracy, F1 score, and confusion matrix. The evaluation metrics for the three modalities—T2, DWI, and DCE—are calculated separately to analyze each modality’s performance in the classification task.
Draw and save ROC curves, PR curves, and confusion matrices. The prediction results of the total model and each modality are plotted and saved separately.
This study has been approved by the Ethics Committee of the First Affiliated Hospital of Huzhou Normal University (20,180,017), and all methods were conducted in accordance with relevant guidelines and regulations. This study involved the use of patients’ tissue samples in the model, and all patients signed the specific informed consent forms collected.
In this section, we analyze the performance of each modality using evaluation metrics such as ROC curve, PR curve, accuracy, confusion matrix, etc. Confusion matrix is an important tool for evaluating the performance of classification models, which shows the relationship between the model prediction results and the actual labels in the form of a matrix. In a clinical setting, the confusion matrix provides insights into the percentages of false positives and false negatives from the model. False positives occur when healthy patients are incorrectly identified as having a condition, leading to unnecessary treatments and increased psychological burden. False negatives happen when sick patients are misclassified as healthy, delaying treatment and potentially worsening their condition. Such misdiagnoses can have serious consequences, including health risks, financial burdens, reduced trust, and legal implications. Looking at the confusion matrix, we can see that our model has trouble identifying benign samples, which may be caused by unbalanced samples. We can improve this by enhancing feature selection or expanding the dataset. The training and validation loss plots exhibit significant fluctuations during the initial training phase, particularly with several higher peaks in the validation loss. This suggests that the model may overfit or underfit the data in certain epochs . Overall, both training and validation losses exhibit a gradual downward trend; however, the fluctuations in validation loss suggest that the model’s generalization ability requires further optimization (Fig. -a). The figure illustrates that the validation accuracy of the DCE modality exhibits significant fluctuations during training. Accuracy improves markedly in the initial stage, followed by noticeable fluctuations and a downward trend. Overall, the validation accuracy of the DCE modality remains between 0.5 and 0.9 throughout the training process. Possible reasons for this include insufficient adaptation of the model to the DCE features or instability during training due to the complexity of the DCE data (Fig. -b). The validation accuracy of the T2 modality also exhibits significant fluctuations during training, particularly in the early and middle stages. Overall, accuracy fluctuates between 0.2 and 0.7, indicating the model’s unstable performance with T2 modal data. Enhancing the model’s adaptation to T2 data may necessitate improved feature extraction and fusion techniques (Fig. -c). The validation accuracy of the DWI modality fluctuates significantly during the initial training phase before gradually stabilizing; however, overall accuracy remains low, primarily fluctuating between 0.4 and 0.55. This suggests that the model faces challenges in handling the modal features of DWI, indicating that further optimization may be required to enhance the classification performance of DWI data(Fig. -d). The ROC curve for the training set nearly completely encompasses the ideal (0, 1) point, achieving an AUC value of 1.00. This suggests that the model demonstrates an exceptionally high classification ability on the training set, allowing it to nearly perfectly distinguish between benign and malignant samples(Fig. -e). The ROC curve for the validation set closely resembles that of the training set, with an AUC value near 1.00, indicating that the model’s classification ability on the validation set is also robust. This demonstrates that the model performs well on both the training and validation sets, reflecting its solid discrimination ability and indicating a degree of generalization capability (Fig. -f). The confusion matrix for the DCE modality indicates that all samples are classified as malignant (label 1), with none correctly identified as benign (label 0). This suggests that the model exhibits poor classification performance for the DCE modality, failing to accurately distinguish between benign and malignant samples (Fig. -a). The PR curves for the DCE modality indicate that while the model achieves high precision at certain recall values, its overall performance is lacking. The AUC value of 0.99 suggests that, despite the model maintaining high accuracy with DCE data, it struggles to effectively distinguish between sample categories (Fig. -b). The ROC curve for the DCE modality displays an AUC value of 0.99, approaching 1.0, which indicates the model’s high classification ability on the test set. However, the results from the confusion matrix reveal that the actual classification performance is suboptimal (Fig. -c). The confusion matrix for the DWI modality indicates that all 68 malignant samples were correctly classified; however, only 3 benign samples were accurately identified, while the remaining 65 were misclassified as malignant. This suggests that the model encounters significant challenges in distinguishing benign samples (Fig. -a). The PR curve for the DWI modality indicates that the model maintains high precision across most recall values, with an AUC value of 0.96, suggesting that it can effectively distinguish between benign and malignant samples in most instances (Fig. -b). The ROC curve for the DWI modality displays an AUC value of 0.95, indicating the model’s high classification ability on the test set. However, the results from the confusion matrix reveal that the model struggles to perform effectively with benign samples (Fig. -c). The confusion matrix for the T2 modality indicates that all samples are classified as malignant (label 1), with none correctly identified as benign (label 0). This resembles the results for the DCE modality, reflecting the model’s very poor classification performance on T2 data (Fig. -a). The PR curve for the T2 modality indicates that the model performs poorly on T2 data, with an AUC value of 0.31, suggesting that it is largely unable to distinguish between sample categories (Fig. -b). The ROC curve for the T2 modality displays an AUC value of 0.01, significantly lower than the ideal value, further confirming the model’s very poor classification performance on T2 data (Fig. -c). The confusion matrix for the overall test reveals: 45 benign samples correctly classified (True Negatives). 23 benign samples misclassified as malignant (False Positives). All 68 malignant samples were correctly classified (True Positives). No malignant samples were misclassified as benign (False Negatives). The confusion matrix indicates that the model performs exceptionally well in distinguishing malignant samples, with no malignant samples misclassified. However, the model struggles to distinguish benign samples, misclassifying 23 as malignant. This suggests that the model has some limitations in handling benign samples and may require further optimization (Fig. -a). The PR curve for the overall test indicates that the model maintains high precision across most recall values, achieving an AUC value of 0.91. This suggests that the model can effectively distinguish between benign and malignant samples in the majority of cases. The decreasing trend of the PR curve suggests that precision declines as recall increases; however, overall precision remains high (Fig. -b). The ROC curve for the overall test displays an AUC value of 0.89, suggesting that the model demonstrates high classification ability on the test set. The ROC curve is positioned close to the top left, indicating that the model effectively distinguishes between positive and negative samples. However, the AUC value is below 1, suggesting that the model may still experience misclassification in certain instances (Fig. -c). According to the description in Table , we can obtain the following information. The overall accuracy is 0.8309, indicating that 83.09% of the tested samples were correctly classified by the model. Overall accuracy is a crucial metric for assessing the model’s classification performance on the entire test set, demonstrating its ability to distinguish between benign and malignant samples in most instances. The overall F1 score, defined as the harmonic mean of precision and recall, is 0.8553. This score comprehensively reflects the model’s ability to classify correctly and recognize samples from different classes, making it particularly suitable for datasets with class imbalance. The model demonstrates strong performance in balancing precision and recall. In summary, the model’s performance on the overall test set is as follows: Classification Ability: The overall ROC curve indicates that the model has high classification ability, with an AUC value of 0.89, demonstrating its effectiveness in distinguishing between positive and negative samples on the test set. Misclassification of Benign Samples: The confusion matrix indicates that the model struggles to distinguish benign samples, misclassifying 23 benign samples as malignant. This could result in a high false positive rate, necessitating further optimization of the model to enhance the recognition accuracy of benign samples. Identification of Malignant Samples: The model demonstrates excellent performance in distinguishing malignant samples, accurately classifying all malignant samples, which indicates a strong ability to identify such cases. Overall Precision and Recall: The PR curve indicates that the model maintains high precision across most recall values, achieving an AUC value of 0.91. This suggests that the model can maintain high accuracy and recall for most samples; however, precision decreases at higher recall levels.
Accurate evaluation of MRI-detected prostate cancer (PCa) is essential for early diagnosis and management, but it remains a challenge in clinical practice. This study employs a pre-trained ResNet50 model as a feature extractor, using a multi-head attention mechanism to fuse features from different modalities and capture their interrelationships. Classification is completed through a fully connected layer to predict benign or malignant lesions. Our study shows that the proposed model, integrating T2-weighted, DCE, and DWI MRI sequences, can identify interpretable and diagnostically relevant changes by analyzing regions of interest. The overall AUC of our AI model is 0.89, indicating strong classification ability. While it effectively identifies malignant samples, its performance in distinguishing benign samples is suboptimal, highlighting the need for further optimization. The findings of this study lay the groundwork for future research on using AI to detect MRI lesion changes and assess clinical relevance, ultimately improving prostate cancer diagnostic rates and reducing misdiagnoses and missed diagnoses, while providing clearer guidance for patients. In recent years, ResNet50 has become a prominent CNN architecture for disease diagnosis. Compared to traditional deep networks, it effectively alleviates the vanishing gradient problem and enables deeper training, allowing for better feature capture in images . Using the ResNet50 model directly to classify benign and malignant lesions in prostate MRI may not be the most effective approach, as prostate tumors are subtler and harder to differentiate than other cancers. To address this challenge and enhance performance, we incorporate a multi-head attention mechanism to optimize the ResNet50 model for classifying prostate cancer lesions. Compared to traditional CNNs, the multi-head attention mechanism effectively handles multi-modal data. By using a weighted combination of modal features, it better extracts and fuses complementary information across modalities . We compare our model with previous AI algorithms that evaluate features from three complete MRI sequences, optimized to identify the best machine learning pipeline. Other models mainly use one or two sequences, such as T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI), whereas we innovatively incorporate DCE-MRI sequences . DCE-MRI sequences can quantitatively and semi-quantitatively evaluate tumor blood flow and energy changes. The density of blood vessels around prostate cancer (PCa) tissue is twice that of normal cells, showing higher permeability and rapid enhancement, indicating an "outflow type." In contrast, early-stage tumors exhibit an “inflow type” enhancement, aiding in distinguishing prostate lesions and improving diagnostic accuracy. Compared to T2 sequences, DCE-MRI better reveals tumor location. Additionally, unlike other studies, we use pathological slices to align imaging data. This verification allows us to easily identify corresponding levels in both pathological and imaging slices, enhancing our model’s accuracy. In terms of cancer detection performance, most reported AUC values range from 0.82 to 0.89, while our model achieves an AUC of 0.89, indicating its superiority in diagnostic accuracy compared to most other models. , – . While we observe encouraging results, we acknowledge several limitations. The dataset used to train and validate the classifier is relatively small, necessitating larger datasets for further validation. Our single-center study lacks ethnic and geographic diversity, which may introduce bias, highlighting the need for nationwide studies to confirm generalizability. Future research will involve generating new samples through transformations to enhance diversity and improve model generalization. Additionally, we employ cross-validation to address small sample size challenges, primarily using multi-parametric MRI for prostate cancer diagnosis, achieving high sensitivity, specificity, and accuracy . However, grading malignancies remains a challenge. Blessin et al. developed machine learning models for automatic grading by combining imaging features from whole slide images with the Ki-67PI marker for brain tumors . Ying et al. used traditional machine learning techniques, specifically SVM, as the classifier . This approach achieved high classification accuracy in grading brain tumors. Our study builds on their method by combining MRI and pathological images to enhance our understanding of tumor characteristics. Kwak et al. digitized pathological images and compared them with standardized MRI images, significantly improving machine learning efficiency and enabling accurate detection of clinically significant prostate cancer (csPCa) that may be missed with sequential MRI features alone . Thirdly, our study relied on expert manual segmentation to extract radiological features from MRI scans, while Mehralivand et al. proposed utilizing the system to achieve similar performance . Nonetheless, current machine learning models have not yet matched the accuracy of senior radiologists, requiring some manual calibration. In the future, we plan to optimize our model using ensemble methods, combining ResNet50 with U-Net, DenseNet, and others to enhance diagnostic accuracy. We will also employ automated methods, such as grid search and Bayesian optimization, to fine-tune hyperparameters like learning rate and batch size. Additionally, we are considering multi-center data collection in collaboration with various medical institutions to gather diverse prostate cancer imaging data, ensuring better dataset diversity and representation to enhance our study.
This paper presents a novel study investigating the feasibility and advantages of a new CAD scheme based on a deep learning ResNet50 model, enhanced by a multi-head attention mechanism for fine-tuning. This approach aims to classify malignant and benign lesions in prostate MRI examinations. Data analysis shows that the lesion classification performance, indicated by the AUC value, significantly exceeds that of other models. Incorporating attention mechanisms or optimization strategies can enhance deep learning model performance in medical imaging.
|
The use of comics to promote health awareness: A template using nonalcoholic fatty liver disease | dd508efc-ffb9-4810-80f5-51b765482ad1 | 9285735 | Health Communication[mh] | INTRODUCTION Health promotion initiatives geared towards the general public, or targeted to specific at‐risk segments of the population, are an important aspect related to biomedical research, notably in its more translational aspects of eliciting health‐conscious behavioural changes. These are important endeavours that must, however, be rationally built and carefully monitored for efficacy and impact, unlike what is the case for most outreach activities normally carried out in this field. This involves the need for appropriate strategies and methodologies, including, for example, appropriate focus groups representing the intended audience, participatory research, structured interviews (evaluated from both qualitative and quantitative standpoints), or structured questionnaires, both to determine the best way to convey a message, and if that message is indeed conveyed and/or elicits change. While these strategies and methodologies may seem foreign to most biomedical researchers, they have been proven robust in different contexts, from sociology to psychology, , and any difficulty merely stresses the importance of interdisciplinary collaborations. In this paper, we describe the reasoning and strategies involved in building a comic‐book narrative tailored to both raise awareness for nonalcoholic fatty liver disease (NAFLD), and elicit health‐promoting lifestyle changes in at‐risk individuals. We believe that comics, in conjunction with other tools, have the potential of providing an excellent platform towards advancing health promotion and effecting behavioural change in different situations.
NAFLD AS A PUBLIC HEALTH THREAT NAFLD constitutes a major threat to public health systems worldwide on account of its widespread prevalence, its increasing incidence, the pervasive nature of its etiological drivers and the serious hepatic and extra‐hepatic consequences it can unfold in both children and adults. , , The clinical construct of NAFLD (Figure ), encompassing the diagnostic entities of nonalcoholic fatty liver (NAFL) and nonalcoholic steatohepatitis (NASH) with and without fibrosis, is estimated to affect 25% of the global adult population and has an incidence of 3.62 million cases per year. , As the more severe form of the disease and with a prevalence of 1%‐3%, NASH constitutes a risk factor for the development of end‐stage liver disease (ESLD), namely, decompensated cirrhosis and hepatocellular carcinoma (HCC), and is expected to become the most common indication for liver transplantation. , , In fact, although cardiovascular diseases (CVDs) are currently the primary cause of death for NAFLD patients, an increase in liver‐related mortality is expected in upcoming years. , , , , Sharing the same underlying aetiology with other behaviourally driven metabolic conditions including obesity and type 2 diabetes mellitus (T2DM), NAFLD is caused by an energy imbalance via hypercaloric diets and sedentary lifestyles, leading to the ectopic accumulation of fat in the liver. The most effective and comprehensive therapy for the management of NAFLD is the implementation of a lifestyle intervention through calorie‐restricted diets (Figure ) and regular physical activity. Although those with comorbidities can profit from the NAFLD‐related benefits of therapeutic options available for their respective pathologies, for instance bariatric surgery in obese patients or medical drugs for the treatment of diabetes and hyperlipidaemia, there are no pharmacological agents approved for the treatment of NAFLD, and although clinical trials testing the effectivity of a myriad of drugs are underway, any therapy will likely still require an adjuvant lifestyle intervention. , Nonetheless, despite its proven efficacy and similar to the case of other cardiometabolic conditions, the success of lifestyle interventions for the treatment of NAFLD is low. Weight loss is difficult to achieve and specially to maintain long‐term. In fact, after one year, only 10% of patients seem to successfully achieve the 7%‐10% weight loss required for histological improvement of NAFLD including steatosis, inflammation and fibrosis. Similarly, more broadly at a population level, health promotion campaigns in the past few decades have also attempted to tackle the metabolic epidemic, but transitioning from an obesogenic lifestyle into calorie‐restricted diets and regular physical activity is not an easy task. , ,
THE DEFICIT IN NAFLD AWARENESS While the preventive and therapeutic approaches of most cardiometabolic conditions are similar, the management of NAFLD faces the additional challenge posed by the generalized lack of disease awareness. Indeed, the scarce literature documenting NAFLD awareness reports that less than 2%‐17% of the general population or cardiometabolic patients are familiar with this form of liver disease. , Awareness of any condition alone is unlikely to result in overt lifestyle modifications, and it could be argued that NAFLD awareness efforts would be equally liable to the constraints responsible for the low effectivity of initiatives targeting better‐known metabolic disorders. , , Undeniably, public awareness of the consequences of overconsumption lifestyle choices needs to accompany policy changes and infrastructure investments for the promotion of healthy habits able to redirect the fate of this epidemic. Yet, according to models of behaviour change, disease unawareness prevents the adoption of precautionary action against the development of the disease, and as such represents a major barrier for NAFLD health promotion, disease prevention and therapeutic adherence. In fact, becoming cognizant of the disease and its health consequences is envisioned as the first step within the trajectorial sequence of an individual's readiness for change. Thus, the provision of NAFLD knowledge to increase public awareness about the disease should set the process of individual behaviour change in motion and promote the adoption of healthier lifestyles. Following this rationale, the specific case of NAFLD requires a customized strategy, with communication tools that take this into account when attempting to address this health threat and its dragging clinical, economic and societal burdens. , The design of NAFLD awareness campaigns or communication tools should be specifically tailored to induce the formation of health‐conductive beliefs and attitudes through the provision of NAFLD‐related knowledge.
ONGOING EFFORTS IN NAFLD AWARENESS The last decade has witnessed the development of formal and informal initiatives seeking to rise NAFLD awareness, ranging from videos on YouTube developed by freelancers to explanatory guides by nonprofit organizations and an increased NAFLD media coverage. Several institutions, such as the Fatty Liver Foundation (FLF), The British Liver Trust (BLT), the Global Liver Institute (GLI) and the American Liver Foundation (ALF), have an education section on their websites offering information in brochures, videos and infographics as well as patient stories and hold periodical outreach campaigns and programmes to raise NAFLD awareness. Similarly, nonprofit organizations such as NASH kNOWledge and the NASH Education Program TM also provide a broad range of education and communication resources including videos, a documentary and brochures for the general public, at‐risk individuals and NAFLD patients. Moreover, since 2018, many of these institutions join forces on June 12th for the annual International NASH day, with online materials, text messaging initiatives, social media awareness campaigns and outreach activities throughout the world. With the support of the EASL, the American Association for the Study of the Liver (AASD), the European Liver Patient's Association (ELPA) and Liver Patients International (LPI) among many other institutions, the 2020 edition included virtual education panels, radio tours, press releases and social media coverage by over 80 partners in 26 countries. These local, national and internationally organized events and advocate institutions do not only target the general public, at‐risk individuals and liver disease patients, but also to the medical community, public health authorities and the media. , Contributing to the advocacy of these initiatives, the FOIE GRAS Project, as a European Training Network focused on the bioenergetic remodelling in the pathophysiology and treatment of nonalcoholic fatty liver disease, has been deeply committed to raise NAFLD awareness and promote healthy lifestyles. Using NAFLD biomedical knowledge as the basis to raise awareness about the disease, the consortium developed a science communication multimodal campaign for the 2018 edition of the European University Games (EUG), the largest university multisport event in Europe. Combining an on‐site outreach booth with a multimodal media campaign including illustrated chronicles, radio clips, flyers, videos and a comic strip, the initiative reached over 100.000 members of the local and the international community. Furthermore, the network developed a comic book designed to convey the more relevant issues of NAFLD to a general population. This was a layered approach that first determined the appropriateness of the comic medium for this purpose and that strove to methodically identify both the themes that would be addressed, as well as the way in which they were addressed. These aspects are briefly described in the following sections, and while specifically tailored to NAFLD, we hope they may serve as a template for similar efforts addressing different clinical conditions.
COMICS AS A TOOL IN SUCCESSFUL HEALTH‐BASED COMMUNICATION 5.1 An introduction to biomedical comics In a quest to optimize the impact of advocated health messages on the knowledge, beliefs, attitudes, intentions and practices carving people's health status and well‐being, health communication research has developed a series of strategies and approaches that maximize the potential effectivity of such efforts. Visual design and narrative storytelling have been consistently appointed as successful tools in the construction of persuasive health communication. As a medium combining visual and narrative attributes, comics embody a particularly suited format to effectively convey biomedical knowledge and exert a health‐promoting influence on covert and overt behaviours of individuals and populations. The visual dimension of comics makes scientific concepts more concrete through their capacity to represent and portray structures and phenomena that lay beyond our sensory experiences. Besides the use of colour to highlight/stress different aspects (see Figure ), the visual properties of comics are able to modify scalar proportions and relational perspectives, providing a perceptual experience for abstract concepts or entities, for instance biological processes. , Moreover, resources such as visual metaphors and anthropomorphism are easily used in comics to make the unfamiliar familiar (see Figures and , bottom panels, showing ‘happy’ and ‘unhappy’ cells related to fat deposits). The attribution of human‐like properties to inert or abstract entities creates a framework to approach new concepts, allowing for the transference of human behaviour knowledge to these concepts and facilitating comprehension. Additionally, unlike other entertainment formats able to promote knowledge acquisition and influence beliefs, attitudes and behaviours, the medium of comics is distinguished by its permanence. Other media, such as TV, film or animations, have a determined speed at which the viewer is expected to acquire the information provided. Conversely, comics have a self‐determined reading pace and its fixed visual component allows the reader to dwell in the details of the imagery and explore the different meanings conveyed by the semiotic interplay offered by the layout of the page. In fact, although some evidence contradicts these findings, this permanence of comics seem to be an advantage over animations in terms of information processing and, furthermore, their nature lends them to adaptation to different formats, including printed leaflets, websites or social networks. , While the visual attributes of comics are important, the persuasive and communicative potential of the medium is largely attributed to its narrative format, conveyed both by its visual and verbal dimensions. , Rather than the close association between words and images, the distinctive feature of comics is the format of sequential art or the capacity to create a story or narrative through the particular semiotic interplay between imagery and text. When used as a tool for health or science communication, and alongside well‐designed characters and biomedical visual representations, the narrative contextualizes the scientific or health‐related knowledge and engages the reader at a personal level. , As graphic narratives, comics have the power of turning the factual, emotional and social contexts of disease and illness more visible. Narrative is a mode of discourse that can intertwine factual or plausible statements with fictional representations of the world and has the potential to portray beliefs, goals and experiences of fictional or nonfictional characters that closely resemble real‐life experiences of the interactor. , In fact, narrative experience is a participatory process in which the interactor is not just a passive agent but engages with the narrative characters in similar fashion to conversational interactions. As the story unfolds, the audience witnesses the characters’ experiences, becomes familiar with their traits and establishes an affective connection with them. Through these phenomena of transportation and identification , the visual and verbal narrative dimensions of comics can support the creation of situational models and reduce cognitive resistance, promoting the extension or modification of mental models and facilitating the adoption of healthier lifestyles through covert and overt changes in health‐related beliefs, attitudes and practices. , , 5.2 Examples of health‐promoting comics Indeed, public health professionals have actively adopted the comics medium and have capitalized on this persuasive potential for health communication purposes throughout the domains of healthcare, disease prevention and health promotion. , , , , Besides its capacity to influence behaviours, comics have been reported to be approachable and promote more engagement and motivation than regular text‐based materials while garnering the same knowledge acquisition, which is particularly advantageous in engaging noninterested and new audiences while demonstrating that entertainment and education are not mutually exclusive. , , , Comics have been used to raise awareness about diseases and their symptoms among the general public, to assist patients and their families better understand their illness, to promote the self‐management of chronic conditions, to improve therapeutic adherence in children, to support informed decision‐making in clinical procedures and therapeutic options, , to facilitate or ameliorate patients’ healthcare experience, to promote surgical procedures and organ donation , and to increase follow‐up rates in primary care provider transitions. , The appropriation of comics in the domains of disease prevention and health promotion began with efforts directed at communicable diseases, mostly targeting children and often as part of multimodal campaigns, alongside other media formats. Seeking to increase awareness about risk of infection and transmission of schistosomiasis, taeniasis, cysticercosis, soil‐transmitted helminth infections, filariasis, tuberculosis, malaria and acquired immunodeficiency syndrome (AIDS), these programmes promoted preventive practices focusing on water sanitation, diarrhoea prevention and immunization. , , , , , , Beginning in the late 1980s and through the 1990s, efforts dedicated to human immunodeficiency virus (HIV)/AIDS prevention programmes in developing countries, recognized the use of participatory research for the development of culturally relevant health promotion media in effectively appealing for behaviour change. , From the 1980s and through the 2000s, the use of public health comics was extended to noncommunicable diseases (NCDs) and researchers began to evaluate the impact of comics on knowledge, beliefs, attitudes, intentions and behaviours. Comic interventions through these decades have demonstrated their capacity to raise awareness on routes of exposure or risk behaviours, reduce stigma and promote self‐efficacy and preventive, detection, adherence and self‐management practices in both infectious and NCDs. Although not always informed by formative research, these efforts targeted environmental and behavioural hazards, , , , , , sexual health education, , rheumatoid disease, cancer, , mental health , , or metabolic disorders. , , In the 2010s, following the rise of the metabolic epidemic and undergoing a transition from hard copies to digital media, many comic interventions, some sponsored by relevant Medical and Scientific Associations, have successfully used the medium to promote healthier diets, increase physical activity and weight loss as preventive and therapeutic practices for obesity, diabetes and cardiovascular diseases, using scientifically rigorous, but understandable, messages. , , , Most recently, the effectivity of public health comics in inducing knowledge acquisition and the promotion of belief, attitude and behaviour change has been further curated through the development of characters and storylines not only informed by formative research with target audiences, but also designed based on principles of behaviour change research. , Crafting character interactions and story events to portray disease severity and susceptibility beliefs through realistic and relatable experiences, these graphic narratives model the overcoming of behavioural barriers and the benefits of taking precautionary action, ultimately constituting a persuasive and informative health communication tool. , , , ,
An introduction to biomedical comics In a quest to optimize the impact of advocated health messages on the knowledge, beliefs, attitudes, intentions and practices carving people's health status and well‐being, health communication research has developed a series of strategies and approaches that maximize the potential effectivity of such efforts. Visual design and narrative storytelling have been consistently appointed as successful tools in the construction of persuasive health communication. As a medium combining visual and narrative attributes, comics embody a particularly suited format to effectively convey biomedical knowledge and exert a health‐promoting influence on covert and overt behaviours of individuals and populations. The visual dimension of comics makes scientific concepts more concrete through their capacity to represent and portray structures and phenomena that lay beyond our sensory experiences. Besides the use of colour to highlight/stress different aspects (see Figure ), the visual properties of comics are able to modify scalar proportions and relational perspectives, providing a perceptual experience for abstract concepts or entities, for instance biological processes. , Moreover, resources such as visual metaphors and anthropomorphism are easily used in comics to make the unfamiliar familiar (see Figures and , bottom panels, showing ‘happy’ and ‘unhappy’ cells related to fat deposits). The attribution of human‐like properties to inert or abstract entities creates a framework to approach new concepts, allowing for the transference of human behaviour knowledge to these concepts and facilitating comprehension. Additionally, unlike other entertainment formats able to promote knowledge acquisition and influence beliefs, attitudes and behaviours, the medium of comics is distinguished by its permanence. Other media, such as TV, film or animations, have a determined speed at which the viewer is expected to acquire the information provided. Conversely, comics have a self‐determined reading pace and its fixed visual component allows the reader to dwell in the details of the imagery and explore the different meanings conveyed by the semiotic interplay offered by the layout of the page. In fact, although some evidence contradicts these findings, this permanence of comics seem to be an advantage over animations in terms of information processing and, furthermore, their nature lends them to adaptation to different formats, including printed leaflets, websites or social networks. , While the visual attributes of comics are important, the persuasive and communicative potential of the medium is largely attributed to its narrative format, conveyed both by its visual and verbal dimensions. , Rather than the close association between words and images, the distinctive feature of comics is the format of sequential art or the capacity to create a story or narrative through the particular semiotic interplay between imagery and text. When used as a tool for health or science communication, and alongside well‐designed characters and biomedical visual representations, the narrative contextualizes the scientific or health‐related knowledge and engages the reader at a personal level. , As graphic narratives, comics have the power of turning the factual, emotional and social contexts of disease and illness more visible. Narrative is a mode of discourse that can intertwine factual or plausible statements with fictional representations of the world and has the potential to portray beliefs, goals and experiences of fictional or nonfictional characters that closely resemble real‐life experiences of the interactor. , In fact, narrative experience is a participatory process in which the interactor is not just a passive agent but engages with the narrative characters in similar fashion to conversational interactions. As the story unfolds, the audience witnesses the characters’ experiences, becomes familiar with their traits and establishes an affective connection with them. Through these phenomena of transportation and identification , the visual and verbal narrative dimensions of comics can support the creation of situational models and reduce cognitive resistance, promoting the extension or modification of mental models and facilitating the adoption of healthier lifestyles through covert and overt changes in health‐related beliefs, attitudes and practices. , ,
Examples of health‐promoting comics Indeed, public health professionals have actively adopted the comics medium and have capitalized on this persuasive potential for health communication purposes throughout the domains of healthcare, disease prevention and health promotion. , , , , Besides its capacity to influence behaviours, comics have been reported to be approachable and promote more engagement and motivation than regular text‐based materials while garnering the same knowledge acquisition, which is particularly advantageous in engaging noninterested and new audiences while demonstrating that entertainment and education are not mutually exclusive. , , , Comics have been used to raise awareness about diseases and their symptoms among the general public, to assist patients and their families better understand their illness, to promote the self‐management of chronic conditions, to improve therapeutic adherence in children, to support informed decision‐making in clinical procedures and therapeutic options, , to facilitate or ameliorate patients’ healthcare experience, to promote surgical procedures and organ donation , and to increase follow‐up rates in primary care provider transitions. , The appropriation of comics in the domains of disease prevention and health promotion began with efforts directed at communicable diseases, mostly targeting children and often as part of multimodal campaigns, alongside other media formats. Seeking to increase awareness about risk of infection and transmission of schistosomiasis, taeniasis, cysticercosis, soil‐transmitted helminth infections, filariasis, tuberculosis, malaria and acquired immunodeficiency syndrome (AIDS), these programmes promoted preventive practices focusing on water sanitation, diarrhoea prevention and immunization. , , , , , , Beginning in the late 1980s and through the 1990s, efforts dedicated to human immunodeficiency virus (HIV)/AIDS prevention programmes in developing countries, recognized the use of participatory research for the development of culturally relevant health promotion media in effectively appealing for behaviour change. , From the 1980s and through the 2000s, the use of public health comics was extended to noncommunicable diseases (NCDs) and researchers began to evaluate the impact of comics on knowledge, beliefs, attitudes, intentions and behaviours. Comic interventions through these decades have demonstrated their capacity to raise awareness on routes of exposure or risk behaviours, reduce stigma and promote self‐efficacy and preventive, detection, adherence and self‐management practices in both infectious and NCDs. Although not always informed by formative research, these efforts targeted environmental and behavioural hazards, , , , , , sexual health education, , rheumatoid disease, cancer, , mental health , , or metabolic disorders. , , In the 2010s, following the rise of the metabolic epidemic and undergoing a transition from hard copies to digital media, many comic interventions, some sponsored by relevant Medical and Scientific Associations, have successfully used the medium to promote healthier diets, increase physical activity and weight loss as preventive and therapeutic practices for obesity, diabetes and cardiovascular diseases, using scientifically rigorous, but understandable, messages. , , , Most recently, the effectivity of public health comics in inducing knowledge acquisition and the promotion of belief, attitude and behaviour change has been further curated through the development of characters and storylines not only informed by formative research with target audiences, but also designed based on principles of behaviour change research. , Crafting character interactions and story events to portray disease severity and susceptibility beliefs through realistic and relatable experiences, these graphic narratives model the overcoming of behavioural barriers and the benefits of taking precautionary action, ultimately constituting a persuasive and informative health communication tool. , , , ,
DEVELOPING COMICS FOR HEALTH COMMUNICATION 6.1 Curating available information and setting the stage The planning of an effective communication tool starts with the main issues of who the comic is designed for, and what previous knowledge should be assumed. In ours, we decided to target potential NAFLD patients with little or no knowledge of the disease or its underlying mechanisms. We sought to inform potential patients on the nature and progression of the disease (from a cellular to a societal level) and to offer concrete, simple, implementable and effective guidelines that might lead to better disease prevention or self‐management (Figure ). Then one has to be aware that there are two distinct dimensions that ultimately determine the structure of the comic: biomedical and narrative. In the first dimension, one must determine which concepts are to be discussed, and how to portray them, including both concepts that the readers should understand, and behaviours that are meant to be instilled. In the second, choices must be made in terms of how this information is conveyed, what characters are to be used and what they represent, as well as the narrative arc that hopefully will send an effective message across. The narrative should thus be relatable to potential patients and caregivers, focusing on their experience, knowledge and behaviours. From a biomedical standpoint, it is therefore always necessary to extensively review the pathophysiology and treatment of any condition to identify adequate biomedical concepts to communicate, building a conceptual map of events from initial to later stages, in order to provide useful factual knowledge to inform and promote pro‐active preventive decision‐making via individual action. These concepts were related to key issues in NAFLD, such as basic metabolism, nutritional interconversion energy imbalance, liver function, lipid storage, inflammation and the lifestyle choices (sedentarism, poor eating habits) that exacerbate them, , , linked to disease progression markers for which graphical representations had then to be defined (Figures and ). It is important to note that, as any other artform, comics have distinct styles and to appropriately convey biomedical information the choice of a clear legible style is preferrable. Ideally, and in order to build a relatable narrative, the biomedical concepts should then be validated/tested with actual patients/caregivers to whom the comic is destined, in order to acquire insights to curate the quantitative and qualitative nature of the factual knowledge to include in the construction of characters and storyline of the narrative approach, fleshing psychosocial contextual details crucial for persuasive intent, and relatedness to graphical representations. This includes not only whatever biomedical knowledge the individuals possess, but also lifestyle habits, how they were diagnosed and how they manage the disease, interactions with a particular socio‐cultural environment so that these issues are taken into account as well, etc This can greatly inform the development of realistic, relatable characters, as well as a credible and engaging narrative. However, although this is an ideal approach, it may not always be possible for all conditions, given issues related to access, specific lack of patient population characterization in a given setting, or the fact that other outreach efforts may already be underway and it may be advantageous to join efforts. In this case, individuals with similar conditions could serve as a useful ersatz group. 6.2 Creating characters and staging the story With this information curated, characters and a narrative must be defined. Character development entails, not just the construction of identities, but also of interpersonal relationships able to elicit sympathy and empathy as well as to facilitate a scaffold for the discussion of biomedical, social and environmental aspects associated. , Furthermore, a realistic representation style can increase the persuasive potential of the comic, acting in combination with the realistic accounts of the plot, especially if they are inspired in real‐life stories. In biomedical comics, characters are often patients, family members and caregivers, with different sorts of relationships, and representing different aspects, from purveyors of information, to roles in providing concrete experience as well as positive (or negative) reinforcements of behaviours that are to be conveyed to the reader. This is true even when the characters are not built with a persuasive intent in mind, for instance in comics produced by patients or caregivers in conditions ranging from cancer, infertility or neurodegenerative disorders. Often classified under the umbrella of ‘narrative medicine’, or ‘graphic medicine’ when comics are used, these narratives focus on personal stories, stressing (to both the general public, other patients, caregivers, clinicians, etc) how similar biomedical issues differently impact people with distinct backgrounds and life histories, thus constituting a humanities‐based approach to precision/personalized medicine. Besides providing these different personal perspectives and allowing the reader to develop empathy, these narratives/comics also convey biomedical information, disease progression advice, possible roadblocks and difficulties, as well as resources that a reader may wish to explore. , , , , , , , In terms of their specific roles Transitional characters, such as those recently diagnosed with a medical condition that must adapt to this novel reality, are particularly effective in promoting health‐related behavioural changes, as throughout the course of the narrative, these characters discover new things, and display a shift in ideas, attitudes and behaviours in the advocated direction and their health is benefited as a result. Another type of character is a positive role model , used as a source of information, for instance a character that, through a longer experience with the same (or similar) condition, has already obtained detailed self‐management knowledge, being able, for example, to provide advice on basic sanitation procedures, adherence to medical advice, proper device management or drug administration, dietary or exercise practices. As a mirror image a negative role model , might also be useful, representing the life‐threatening consequences of the natural course of the disease if no precautionary action is taken, prompting reflexivity on perceptions of severity of and vulnerability to whatever disease threat the transitional character is undergoing. In the development of our comic tailored to NAFLD patients, we decided to include two transitional characters, a main character with a recent NAFLD diagnosis, expected to represent individuals similarly recently diagnosed with the disease and in the process of adopting therapeutic recommendations, and a NAFLD unaware character, acting as a representation for individuals that transition towards NAFLD awareness and the potential adoption of advocated behaviours as a preventive strategy (Figure ). Embodying an initially disease unaware character as a young child is often used in this type of comic as it allows for several additional features: the introduction of biomedical concepts in a didactic manner without condescendence, and the fact that these characters can both serve as an inspiration to their elders who are in a potential transitional role, and in terms of introducing health‐related concepts and behaviours earlier in life, thus serving as a preventive measure. On the other hand, positive and negative role models must be relatable enough so as to not be dismissed (by either transitional characters or readers). In particular, elders of the same family facilitate the formation of outcome expectations regarding disease severity for both transitional characters and the readers. , Indeed, the potential interrelationships between characters can be crucial in the introduction of concepts regarding the association between their defining traits. Finally, construction of the plot should be such that it fosters the adoption of an intended behaviour, best accomplished by complementing disease threat information with appraisals of the efficacy and benefits of taking preventive action as well as the promotion of feelings of self‐efficacy, when comparing with role models or other characters. As such, much as the conflict‐crisis‐resolution arc common in other narrative artforms, the introduction of disease threat beliefs and attitudes as portrayed by the characters should ideally be followed by a narrative arc illustrating the characters’ confidence and mutual encouragement to succeed in their attempts to change behaviour and their eventual success. , Given that there might be quite a bit of information involved in biomedical comics, to ensure narrative continuity an omniscient narrative style, voiced by a character, can be used in interdependent combination with other visual aspects, including biomedical imagery. This interplay allows for an overlap of biographic or personal illness narratives with biomedical constructs, integrating the biomedical conception of a pathology (ie disease) and the lived experience with such pathology (ie illness). , Finally, it is very important to note that any communication tool of this type should, as much as possible, be locally tailored, both in terms of text (expressions and speech habits) and image (local environment, clothing, etc) so that it presents itself as relatable to a particular audience. Building on the work done in the field of public health comics, and using the principles outlined above, we sought to create a health communication tool in the form of a comic with the goal of rising NAFLD awareness. The comic book ‘A Healthy Liver Will Always Deliver!’ (Figures , , ), was intended to convey health threat beliefs through biomedical imagery while modelling the beneficial health outcomes of adopting healthy nutrition and physical activity habits for NAFLD prevention and regression through the experiences of story characters. It is now available at the website www.fattylivercomics.com in 9 languages, including Portuguese, English , Czech, Italian, French, Spanish, Polish and German, the languages of all the MSCA‐ITN FOIE_GRAS Network partners. 6.3 Concluding remarks: tailoring a narrative and monitoring impact It is important to note that translation of any comic‐based narrative involves much more than language. In fact, if the target audience does not identify with the characters, their surroundings or situations it well be less relatable, and hinder the adoption of health promotion behaviours. This is important at many levels, including housing, clothing, character physiognomy, typical habits, expressions and type of language, specific food items and other aspects related to daily life. Thus, both text and drawing may need to be adapted according to socioeconomic and cultural background, ethnicity or age of intended audiences, so as to reach its full potential, especially in terms of reaching underserved and underprivileged segments of the population. This must always include an initial approach (focus groups, interviews), followed by validation with larger groups in appropriate settings (communities, clinical facilities, schools, etc). The goal with the first part (as shown in Figure ) is to determine what the audience already knows, what it may be more amenable to accept as a biomedical behaviour change message, and how it could be motivated to act upon it, as well as to identify types and themes that can serve as models for characters and situations. Finally, impact monitoring is crucial to determine the effect of any such strategy and is rarely carried out in normal biomedical outreach activities, with a few exceptions. , Impact monitoring is typically accomplished with structured and carefully designed questionnaires, filled out before and after reading a comic or encountering another health promotion tool. This allows researchers and clinicians to determine what messages were correctly transmitted, and, more importantly, where the comic was most and least effective. More long‐term follow‐up studies to determine the persistence of the message, as well as behavioural changes that it may impact should also be considered, for example in a context of regular clinical consultations. In what concerns our NAFLD comic, although more data will have to be collected, a preliminary approach, reduced, in the current pandemic realm, to online questionnaires prior to and immediately following reading the comic, suggest that this approach was effective in increasing NAFLD threat perceptions and response efficacy and self‐efficacy beliefs, normative and control beliefs and positive attitudes regarding healthy dietary and lifestyle practices, although this was done, due to pandemic constrictions, mostly using an educated audience with a high level of awareness, and access to information. This health communication tool hopes therefore to bring a comprehensive contribution to NAFLD awareness for the general public, at‐risk individuals and NAFLD patients, suggesting that other efforts following the same rational development of both characters, plot and biomedical concepts may be useful, not only towards NAFLD awareness, but as a possible template to develop similar materials for other conditions where the same principles may apply. In fact, given previous examples showing the efficacy of comics in transmitting scientific and clinically relevant messages, we believe that, together with other tools (other types of media or narrative medicine strategies), they may provide an excellent platform towards advancing health promotion and effecting behavioural change in different situations.
Curating available information and setting the stage The planning of an effective communication tool starts with the main issues of who the comic is designed for, and what previous knowledge should be assumed. In ours, we decided to target potential NAFLD patients with little or no knowledge of the disease or its underlying mechanisms. We sought to inform potential patients on the nature and progression of the disease (from a cellular to a societal level) and to offer concrete, simple, implementable and effective guidelines that might lead to better disease prevention or self‐management (Figure ). Then one has to be aware that there are two distinct dimensions that ultimately determine the structure of the comic: biomedical and narrative. In the first dimension, one must determine which concepts are to be discussed, and how to portray them, including both concepts that the readers should understand, and behaviours that are meant to be instilled. In the second, choices must be made in terms of how this information is conveyed, what characters are to be used and what they represent, as well as the narrative arc that hopefully will send an effective message across. The narrative should thus be relatable to potential patients and caregivers, focusing on their experience, knowledge and behaviours. From a biomedical standpoint, it is therefore always necessary to extensively review the pathophysiology and treatment of any condition to identify adequate biomedical concepts to communicate, building a conceptual map of events from initial to later stages, in order to provide useful factual knowledge to inform and promote pro‐active preventive decision‐making via individual action. These concepts were related to key issues in NAFLD, such as basic metabolism, nutritional interconversion energy imbalance, liver function, lipid storage, inflammation and the lifestyle choices (sedentarism, poor eating habits) that exacerbate them, , , linked to disease progression markers for which graphical representations had then to be defined (Figures and ). It is important to note that, as any other artform, comics have distinct styles and to appropriately convey biomedical information the choice of a clear legible style is preferrable. Ideally, and in order to build a relatable narrative, the biomedical concepts should then be validated/tested with actual patients/caregivers to whom the comic is destined, in order to acquire insights to curate the quantitative and qualitative nature of the factual knowledge to include in the construction of characters and storyline of the narrative approach, fleshing psychosocial contextual details crucial for persuasive intent, and relatedness to graphical representations. This includes not only whatever biomedical knowledge the individuals possess, but also lifestyle habits, how they were diagnosed and how they manage the disease, interactions with a particular socio‐cultural environment so that these issues are taken into account as well, etc This can greatly inform the development of realistic, relatable characters, as well as a credible and engaging narrative. However, although this is an ideal approach, it may not always be possible for all conditions, given issues related to access, specific lack of patient population characterization in a given setting, or the fact that other outreach efforts may already be underway and it may be advantageous to join efforts. In this case, individuals with similar conditions could serve as a useful ersatz group.
Creating characters and staging the story With this information curated, characters and a narrative must be defined. Character development entails, not just the construction of identities, but also of interpersonal relationships able to elicit sympathy and empathy as well as to facilitate a scaffold for the discussion of biomedical, social and environmental aspects associated. , Furthermore, a realistic representation style can increase the persuasive potential of the comic, acting in combination with the realistic accounts of the plot, especially if they are inspired in real‐life stories. In biomedical comics, characters are often patients, family members and caregivers, with different sorts of relationships, and representing different aspects, from purveyors of information, to roles in providing concrete experience as well as positive (or negative) reinforcements of behaviours that are to be conveyed to the reader. This is true even when the characters are not built with a persuasive intent in mind, for instance in comics produced by patients or caregivers in conditions ranging from cancer, infertility or neurodegenerative disorders. Often classified under the umbrella of ‘narrative medicine’, or ‘graphic medicine’ when comics are used, these narratives focus on personal stories, stressing (to both the general public, other patients, caregivers, clinicians, etc) how similar biomedical issues differently impact people with distinct backgrounds and life histories, thus constituting a humanities‐based approach to precision/personalized medicine. Besides providing these different personal perspectives and allowing the reader to develop empathy, these narratives/comics also convey biomedical information, disease progression advice, possible roadblocks and difficulties, as well as resources that a reader may wish to explore. , , , , , , , In terms of their specific roles Transitional characters, such as those recently diagnosed with a medical condition that must adapt to this novel reality, are particularly effective in promoting health‐related behavioural changes, as throughout the course of the narrative, these characters discover new things, and display a shift in ideas, attitudes and behaviours in the advocated direction and their health is benefited as a result. Another type of character is a positive role model , used as a source of information, for instance a character that, through a longer experience with the same (or similar) condition, has already obtained detailed self‐management knowledge, being able, for example, to provide advice on basic sanitation procedures, adherence to medical advice, proper device management or drug administration, dietary or exercise practices. As a mirror image a negative role model , might also be useful, representing the life‐threatening consequences of the natural course of the disease if no precautionary action is taken, prompting reflexivity on perceptions of severity of and vulnerability to whatever disease threat the transitional character is undergoing. In the development of our comic tailored to NAFLD patients, we decided to include two transitional characters, a main character with a recent NAFLD diagnosis, expected to represent individuals similarly recently diagnosed with the disease and in the process of adopting therapeutic recommendations, and a NAFLD unaware character, acting as a representation for individuals that transition towards NAFLD awareness and the potential adoption of advocated behaviours as a preventive strategy (Figure ). Embodying an initially disease unaware character as a young child is often used in this type of comic as it allows for several additional features: the introduction of biomedical concepts in a didactic manner without condescendence, and the fact that these characters can both serve as an inspiration to their elders who are in a potential transitional role, and in terms of introducing health‐related concepts and behaviours earlier in life, thus serving as a preventive measure. On the other hand, positive and negative role models must be relatable enough so as to not be dismissed (by either transitional characters or readers). In particular, elders of the same family facilitate the formation of outcome expectations regarding disease severity for both transitional characters and the readers. , Indeed, the potential interrelationships between characters can be crucial in the introduction of concepts regarding the association between their defining traits. Finally, construction of the plot should be such that it fosters the adoption of an intended behaviour, best accomplished by complementing disease threat information with appraisals of the efficacy and benefits of taking preventive action as well as the promotion of feelings of self‐efficacy, when comparing with role models or other characters. As such, much as the conflict‐crisis‐resolution arc common in other narrative artforms, the introduction of disease threat beliefs and attitudes as portrayed by the characters should ideally be followed by a narrative arc illustrating the characters’ confidence and mutual encouragement to succeed in their attempts to change behaviour and their eventual success. , Given that there might be quite a bit of information involved in biomedical comics, to ensure narrative continuity an omniscient narrative style, voiced by a character, can be used in interdependent combination with other visual aspects, including biomedical imagery. This interplay allows for an overlap of biographic or personal illness narratives with biomedical constructs, integrating the biomedical conception of a pathology (ie disease) and the lived experience with such pathology (ie illness). , Finally, it is very important to note that any communication tool of this type should, as much as possible, be locally tailored, both in terms of text (expressions and speech habits) and image (local environment, clothing, etc) so that it presents itself as relatable to a particular audience. Building on the work done in the field of public health comics, and using the principles outlined above, we sought to create a health communication tool in the form of a comic with the goal of rising NAFLD awareness. The comic book ‘A Healthy Liver Will Always Deliver!’ (Figures , , ), was intended to convey health threat beliefs through biomedical imagery while modelling the beneficial health outcomes of adopting healthy nutrition and physical activity habits for NAFLD prevention and regression through the experiences of story characters. It is now available at the website www.fattylivercomics.com in 9 languages, including Portuguese, English , Czech, Italian, French, Spanish, Polish and German, the languages of all the MSCA‐ITN FOIE_GRAS Network partners.
Concluding remarks: tailoring a narrative and monitoring impact It is important to note that translation of any comic‐based narrative involves much more than language. In fact, if the target audience does not identify with the characters, their surroundings or situations it well be less relatable, and hinder the adoption of health promotion behaviours. This is important at many levels, including housing, clothing, character physiognomy, typical habits, expressions and type of language, specific food items and other aspects related to daily life. Thus, both text and drawing may need to be adapted according to socioeconomic and cultural background, ethnicity or age of intended audiences, so as to reach its full potential, especially in terms of reaching underserved and underprivileged segments of the population. This must always include an initial approach (focus groups, interviews), followed by validation with larger groups in appropriate settings (communities, clinical facilities, schools, etc). The goal with the first part (as shown in Figure ) is to determine what the audience already knows, what it may be more amenable to accept as a biomedical behaviour change message, and how it could be motivated to act upon it, as well as to identify types and themes that can serve as models for characters and situations. Finally, impact monitoring is crucial to determine the effect of any such strategy and is rarely carried out in normal biomedical outreach activities, with a few exceptions. , Impact monitoring is typically accomplished with structured and carefully designed questionnaires, filled out before and after reading a comic or encountering another health promotion tool. This allows researchers and clinicians to determine what messages were correctly transmitted, and, more importantly, where the comic was most and least effective. More long‐term follow‐up studies to determine the persistence of the message, as well as behavioural changes that it may impact should also be considered, for example in a context of regular clinical consultations. In what concerns our NAFLD comic, although more data will have to be collected, a preliminary approach, reduced, in the current pandemic realm, to online questionnaires prior to and immediately following reading the comic, suggest that this approach was effective in increasing NAFLD threat perceptions and response efficacy and self‐efficacy beliefs, normative and control beliefs and positive attitudes regarding healthy dietary and lifestyle practices, although this was done, due to pandemic constrictions, mostly using an educated audience with a high level of awareness, and access to information. This health communication tool hopes therefore to bring a comprehensive contribution to NAFLD awareness for the general public, at‐risk individuals and NAFLD patients, suggesting that other efforts following the same rational development of both characters, plot and biomedical concepts may be useful, not only towards NAFLD awareness, but as a possible template to develop similar materials for other conditions where the same principles may apply. In fact, given previous examples showing the efficacy of comics in transmitting scientific and clinically relevant messages, we believe that, together with other tools (other types of media or narrative medicine strategies), they may provide an excellent platform towards advancing health promotion and effecting behavioural change in different situations.
None of the authors have potential conflicts of interest to be disclosed.
|
Exploration of the metabolomic mechanisms of postmenopausal hypertension induced by low estrogen state | b880a86f-eca0-4a9b-994a-8c5b4fa1e803 | 11737871 | Biochemistry[mh] | Estrogen has a multifaceted impact on women, with postmenopausal hypertension being a representative example, characterized by significant fluctuations in blood pressure, primarily in systolic and pulse pressure. This leads to more damage to target organs, ultimately severely affecting the quality of life of middle-aged and elderly women . Doctors have attempted to improve patients’ conditions through hormone replacement therapy, but the serious risk of breast cancer has greatly limited the clinical application of this treatment . Angiotensin-converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB) in combination with beta-blockers or verapamil are common control strategies for postmenopausal hypertension, yet satisfactory therapeutic effects are still difficult to achieve . Research has long confirmed the influence of estrogen on vascular pressure, for instance, oxidative stress caused by estrogen metabolism changes may result in pulmonary arterial hypertension in obese patients . The detailed mechanism by which the drastic decrease in estrogen post-menopause mediates postmenopausal hypertension remains unclear. In recent years, metabolomics strategies have been widely used to explore potential mechanisms of various diseases, with plasma metabolomics being the most popular . Some researchers have utilized plasma metabolomics to investigate the pathogenesis of thrombotic or idiopathic pulmonary arterial hypertension , while others have described the plasma metabolomic characteristics of high blood pressure induced by high sodium intake . There has been no research exploring the metabolic changes in postmenopausal hypertension yet. Our research team conducted this study to address the aforementioned scientific issues. By simulating the postmenopausal state using an ovariectomized rat model, we monitored the correlation between estrogen and blood pressure changes in rats and revealed the inherent relationship between these changes and postmenopausal hypertension through alterations in the metabolic characteristics of aortic tissues .
Development of animal models with estrogen depletion In the Sham group, normal estrous cycle patterns were observed, with small, round vaginal exfoliated cells present during the nonestrous phase and larger, polygonal cells with abundant cytoplasm seen during the estrous phase. Conversely, in the OVX and OVX + E groups (pre-estrogen supplementation), the estrous cycle ceased, and the vaginal exfoliated cells remained small and round, confirming the effectiveness of bilateral ovariectomy. Serum estrogen levels in the OVX group were significantly lower compared to the Sham group (7.54 ± 1.46pg/mL vs. 36.12 ± 6.07pg/mL, n = 8, p< 0.001), while levels in the OVX + E group were markedly higher than those in the OVX group (36.21 ±4.30pg/mL vs. 7.54 ± 1.46pg/mL, n = 10, p< 0.001). There was no significant difference between the Sham and OVX + E groups, further confirming the validity of the experimental model . BP features of the OVX model The SBP (151.98±2.79 mmHg), DBP (105.10±2.89 mmHg), and PP (46.88±3.78 mmHg) of the OVX group were significantly higher than the other two groups (p<0.001). There was no statistical difference between the Sham group and the OVX + E group in SBP (136.43±2.45 mmHg vs. 137.60±2.03 mmHg), DBP (97.65±2.34 mmHg vs. 98.13±2.16 mmHg), and PP (38.78±3.24 mmHg vs. 39.48±2.92 mmHg) (p>0.05; ). Characteristics of aortic metabolites related to low estrogen levels A total of 184 metabolites in aortic tissues were identified using metabolomics analysis ( : Metabolomics raw data), mainly categorized into 8 major classes, with amino acids accounting for over 1/3 . To normalize the data distribution, both metabolites and samples underwent normalization procedures . To explore the differences in aortic tissue metabolites under varying estrogen levels, a one-way analysis of variance (ANOVA) was utilized to select differentially expressed metabolites among the three groups of aortas. The false discovery rate threshold was set at 0.05. Since the normalized data approximated a normal distribution, hypothesis testing was carried out through two methods: non-parametric tests identified 17 different metabolites ( : Non-parametric tests of metabolites), while Fisher’s Least Significant Difference (LSD) tests revealed 23 different metabolites ( ：Fisher’s LSD tests of metabolites). By taking the intersection of the two methods, a total of 15 different metabolites were identified ( ：Intersection differential metabolites). Further selection of the most promising differentially expressed metabolites among groups was shown in . Using multi-class significance analysis of microarrays (SAM) (with a delta set at 1.1, ), a total of 45 different metabolites were identified ( , ：Differential metabolites identified by SAM), with L-Alpha-aminobutyric acid (L-AABA), Methylpicraquassioside A, Pyroglutamine, and D-Ribose 5-phosphate being the most promising inter-group differentially expressed metabolites. A heatmap of metabolite correlations indicated that all metabolites could be roughly categorized into 4 distinct clusters based on their correlation relationships ( , , ：The correlation coefficient and p-value of the correlation analysis for all metabolites). The Pearson correlation analysis for the three different sample groups clearly demonstrated distinct differences between the OVX group and the other two groups, while the Sham and OVX + E groups with similar estrogen concentrations were challenging to differentiate ( , , ：The correlation coefficient and p-value of the correlation analysis for all samples). The trends highlighted in the hierarchical clustering dendrogram further emphasized these distinctions . The hierarchical clustering heatmap, created based on the top 25 differentially expressed metabolites filtered by ANOVA, depicted two distinct patterns of metabolite expression across different groups. A class of metabolites represented by Inosinic acid showed a significant upregulation in the OVX group, while displaying consistent downregulation in the Sham and OVX + E groups; another class of metabolites primarily displayed downregulation in the OVX group . Further exploration of the expression characteristics of aortic metabolites in the OVX group was performed using dimensionality reduction analyses. Principal Component Analysis (PCA), Partial Least Squares Discriminant Analysis (PLS-DA), and sparse PLS-DA (sPLS-DA) were applied as the three analytical strategies. In the Unsupervised strategy, the first component accounted for 21.5% and the second component for 15%, with a high degree of overlap among the three groups’ data . Implementing the PLS-DA strategy distinguished the differences between the OVX group and the other two groups clearly, although there was still some minor overlap . Although the model could select some differentially expressed metabolites using VIP scores , a fivefold cross-validation (CV) indicated suboptimal values for R 2 , Q 2 , and accuracy indicators ( , ：PLS-DA CV details). The permutation test also suggested a risk of model overfitting . As the predictive performance of both the PCA and PLS-DA models fell short of expectations, the sPLS-DA model was tested, showing optimal performance in the fivefold CV results ( ， ). According to the VIP scores, L-AABA and Methylpicraquassioside A were identified as the top differentially expressed metabolites ( , ：VIP scores for differential metabolites). Similarly, in the random forest tree model, L-AABA was identified as the most important metabolite for classification accuracy evaluation (Mean Decrease accuracy; , ：Mean Metabolic Accuracy of differential metabolites in random forest tree model). Unfortunately, the out-of-bag (OOB) error for the random forest tree model did not reach 0 (0.0417, ). Subgroup analysis of metabolic characteristics Using the aforementioned methods, we obtained overall differences in three sets of aortic metabolites, and further verification is needed to determine whether these differences exhibit the expected trends among different groups. To this end, we comprehensively employed various methods to assess the differential features of metabolites between OVX and the other two groups (t-test, PCA, PLS-DA and Orthogonal PLS-DA (OPLS-DA), Random Forest, and Empirical Bayesian Analysis of Metabolomics (EBAM)). When comparing data of the subgroups, we recalibrated the data . The Fold Change (FC) threshold for the t-test was set to 2, and the p-value was 0.05. Compared to the Sham group, in the OVX group, Adenylsuccinic acid, 4-Hydroxybutyric acid, Cholic acid, O-dodecanoylcarnitine, L-Hexanoylcarnitine, Adenosine 3’-monophosphate, 1-Methylguanine, O-decanoyl-L-carnitine, and Butyrylcarnitine were significantly upregulated, while Uracil, Uridine, Ribothymidine, and L-AABA were significantly downregulated ( , ：Details of t-test for differential metabolites between Sham group and OVX group). Compared to the OVX group, Cysteinylglycine, Adenosine 3’-monophosphate, Adenylsuccinic acid, and 4-Hydroxybutyric acid were significantly downregulated in the OVX + E group, while Glycocholic acid, L-AABA, L-Erythrulose, 4-Guanidinobutanoic acid, p-Cresol sulfate, 1-Methylguanosine, Indoxyl sulfate, and Uridine were significantly upregulated ( , ：Details of t-test for differential metabolites between OVX + E group and OVX group). The hierarchical clustering dendrogram clearly distinguished the samples of different subgroups . The hierarchical clustering heatmap visually displayed the above results . Although PCA analysis as an unsupervised analysis could effectively differentiate between the OVX and OVX + E groups, it faced difficulties in distinguishing differences between the OVX and Sham groups . In contrast, supervised methods were more adept at discerning intergroup differences, but these models also had certain limitations. The PLS-DA model had good intergroup discriminative ability , but showed clear signs of model overfitting (Figure S6E-F). The OPLS-DA model also had advantages in discriminative ability , and the permutations test results indicated that the R 2 Y values of the OPLS-DA models between OVX and the other two groups were 0.981 (p=0.009) and 0.986 (p=0.001), with Q 2 values of 0.848 (p<0.002) and 0.893 (p=0.001; ). The cross-validation results were consistent with the permutations test , demonstrating the high predictive value of the models. Utilizing the VIP scores and S-Plot of this model, we further identified four important differential metabolites, namely L-AABA, 4-Hydroxyproline, O-dodecanoylcarnitine, and Methylpicraquassioside A ( , , ：VIP scores of OPLS-DA models between subgroups). The Random Forest model exhibited strong advantages in determining intergroup differences, with an OOB error of 0 , and the VIP plots based on contribution to classification accuracy provided several promising differential metabolites ( , , ：VIP scores of the random forest model between subgroups), which still included L-AABA. The evaluation results of EBAM and SAM are shown in ( ：The evaluation details of EBAM and SAM model between subgroups). Metabolite expression characteristics Metabolite expression characteristics are another analytical strategy to explore the impact of low estrogen on the aorta. In the subgroup analysis, using the OVX group as a control, representative metabolites associated with the Sham group and OVX + E group are shown in ; , : Details of intergroup differential analysis of low estrogen related metabolites. Among them, the metabolites with a positive correlation coefficient exceeding 0.8 include Methylpicraquassioside A, L-AABA, and D-Ribose 5-phosphate (Sham group), as well as L-AABA, 4-Guanidinobutanoic acid, Methylpicraquassioside A, and L-Erythrulose (OVX +E group). The trend arranged from low to high estrogen concentrations (OVX-Sham-OVX+E) is shown in , ：Details of metabolite related trends from low to high estrogen concentrations. L-AABA exhibits a significant positive correlation feature (correlation coefficient of 0.89), while Inosinic acid is the only metabolite with a negative correlation coefficient exceeding –0.7 (correlation coefficient of –0.7). Therefore, we further explored metabolites with correlated expression to L-AABA ( , : Correlation coefficient of metabolites related to L-AABA expression), where L-Erythrulose shows the most positive correlation (correlation coefficient of 0.87171), and Inosinic acid shows the most negative correlation (correlation coefficient of –0.62298). Identification of biomarkers Univariate receiver operating characteristic (ROC) curve analysis was used to screen promising biomarkers, and L-AABA was found to be the most promising biomarker, with an AUC of 1 in the comparison process between OVX and the other two groups ( , , : The promising differential metabolites AUC results of ROC curves between subgroups). Multivariate ROC curve analysis demonstrated unique value in biomarker selection, with the AUC consistently above 0.93 starting from 5 variables . The error classification of the multivariate ROC curve showed that there were no misclassifications when using the model for samples in the OVX and OVX +E groups, but one sample from the Sham group was incorrectly classified into the OVX group . In this multivariate model, L-AABA still ranked in the top three in terms of average importance . Enrichment analysis of differential metabolites Further enrichment analysis of differential metabolites was conducted to elucidate the underlying mechanisms. Differential metabolites showed a wide range of classifications, with amino acids and peptides, fatty acids and conjugates, and monosaccharides ranking among the top three . Results of metabolic pathway enrichment analysis indicated that the Warburg effect, glycolysis, and gluconeogenesis were the top three enriched metabolic pathways . Enzyme-specific metabolic analysis revealed that deoxyuridine phosphorylase (UPP) and O 2 transport (diffusion) ranked in the top two . To eliminate the impact of species differences on enrichment results, we specifically conducted rat-specific metabolic pathway analysis, which showed significant differences in the Pentose phosphate pathway, Glycerophospholipid metabolism, Arginine and proline metabolism, and Pyrimidine metabolism ( , ：Rat-specific metabolic pathway analysis details).
In the Sham group, normal estrous cycle patterns were observed, with small, round vaginal exfoliated cells present during the nonestrous phase and larger, polygonal cells with abundant cytoplasm seen during the estrous phase. Conversely, in the OVX and OVX + E groups (pre-estrogen supplementation), the estrous cycle ceased, and the vaginal exfoliated cells remained small and round, confirming the effectiveness of bilateral ovariectomy. Serum estrogen levels in the OVX group were significantly lower compared to the Sham group (7.54 ± 1.46pg/mL vs. 36.12 ± 6.07pg/mL, n = 8, p< 0.001), while levels in the OVX + E group were markedly higher than those in the OVX group (36.21 ±4.30pg/mL vs. 7.54 ± 1.46pg/mL, n = 10, p< 0.001). There was no significant difference between the Sham and OVX + E groups, further confirming the validity of the experimental model .
The SBP (151.98±2.79 mmHg), DBP (105.10±2.89 mmHg), and PP (46.88±3.78 mmHg) of the OVX group were significantly higher than the other two groups (p<0.001). There was no statistical difference between the Sham group and the OVX + E group in SBP (136.43±2.45 mmHg vs. 137.60±2.03 mmHg), DBP (97.65±2.34 mmHg vs. 98.13±2.16 mmHg), and PP (38.78±3.24 mmHg vs. 39.48±2.92 mmHg) (p>0.05; ).
A total of 184 metabolites in aortic tissues were identified using metabolomics analysis ( : Metabolomics raw data), mainly categorized into 8 major classes, with amino acids accounting for over 1/3 . To normalize the data distribution, both metabolites and samples underwent normalization procedures . To explore the differences in aortic tissue metabolites under varying estrogen levels, a one-way analysis of variance (ANOVA) was utilized to select differentially expressed metabolites among the three groups of aortas. The false discovery rate threshold was set at 0.05. Since the normalized data approximated a normal distribution, hypothesis testing was carried out through two methods: non-parametric tests identified 17 different metabolites ( : Non-parametric tests of metabolites), while Fisher’s Least Significant Difference (LSD) tests revealed 23 different metabolites ( ：Fisher’s LSD tests of metabolites). By taking the intersection of the two methods, a total of 15 different metabolites were identified ( ：Intersection differential metabolites). Further selection of the most promising differentially expressed metabolites among groups was shown in . Using multi-class significance analysis of microarrays (SAM) (with a delta set at 1.1, ), a total of 45 different metabolites were identified ( , ：Differential metabolites identified by SAM), with L-Alpha-aminobutyric acid (L-AABA), Methylpicraquassioside A, Pyroglutamine, and D-Ribose 5-phosphate being the most promising inter-group differentially expressed metabolites. A heatmap of metabolite correlations indicated that all metabolites could be roughly categorized into 4 distinct clusters based on their correlation relationships ( , , ：The correlation coefficient and p-value of the correlation analysis for all metabolites). The Pearson correlation analysis for the three different sample groups clearly demonstrated distinct differences between the OVX group and the other two groups, while the Sham and OVX + E groups with similar estrogen concentrations were challenging to differentiate ( , , ：The correlation coefficient and p-value of the correlation analysis for all samples). The trends highlighted in the hierarchical clustering dendrogram further emphasized these distinctions . The hierarchical clustering heatmap, created based on the top 25 differentially expressed metabolites filtered by ANOVA, depicted two distinct patterns of metabolite expression across different groups. A class of metabolites represented by Inosinic acid showed a significant upregulation in the OVX group, while displaying consistent downregulation in the Sham and OVX + E groups; another class of metabolites primarily displayed downregulation in the OVX group . Further exploration of the expression characteristics of aortic metabolites in the OVX group was performed using dimensionality reduction analyses. Principal Component Analysis (PCA), Partial Least Squares Discriminant Analysis (PLS-DA), and sparse PLS-DA (sPLS-DA) were applied as the three analytical strategies. In the Unsupervised strategy, the first component accounted for 21.5% and the second component for 15%, with a high degree of overlap among the three groups’ data . Implementing the PLS-DA strategy distinguished the differences between the OVX group and the other two groups clearly, although there was still some minor overlap . Although the model could select some differentially expressed metabolites using VIP scores , a fivefold cross-validation (CV) indicated suboptimal values for R 2 , Q 2 , and accuracy indicators ( , ：PLS-DA CV details). The permutation test also suggested a risk of model overfitting . As the predictive performance of both the PCA and PLS-DA models fell short of expectations, the sPLS-DA model was tested, showing optimal performance in the fivefold CV results ( ， ). According to the VIP scores, L-AABA and Methylpicraquassioside A were identified as the top differentially expressed metabolites ( , ：VIP scores for differential metabolites). Similarly, in the random forest tree model, L-AABA was identified as the most important metabolite for classification accuracy evaluation (Mean Decrease accuracy; , ：Mean Metabolic Accuracy of differential metabolites in random forest tree model). Unfortunately, the out-of-bag (OOB) error for the random forest tree model did not reach 0 (0.0417, ).
Using the aforementioned methods, we obtained overall differences in three sets of aortic metabolites, and further verification is needed to determine whether these differences exhibit the expected trends among different groups. To this end, we comprehensively employed various methods to assess the differential features of metabolites between OVX and the other two groups (t-test, PCA, PLS-DA and Orthogonal PLS-DA (OPLS-DA), Random Forest, and Empirical Bayesian Analysis of Metabolomics (EBAM)). When comparing data of the subgroups, we recalibrated the data . The Fold Change (FC) threshold for the t-test was set to 2, and the p-value was 0.05. Compared to the Sham group, in the OVX group, Adenylsuccinic acid, 4-Hydroxybutyric acid, Cholic acid, O-dodecanoylcarnitine, L-Hexanoylcarnitine, Adenosine 3’-monophosphate, 1-Methylguanine, O-decanoyl-L-carnitine, and Butyrylcarnitine were significantly upregulated, while Uracil, Uridine, Ribothymidine, and L-AABA were significantly downregulated ( , ：Details of t-test for differential metabolites between Sham group and OVX group). Compared to the OVX group, Cysteinylglycine, Adenosine 3’-monophosphate, Adenylsuccinic acid, and 4-Hydroxybutyric acid were significantly downregulated in the OVX + E group, while Glycocholic acid, L-AABA, L-Erythrulose, 4-Guanidinobutanoic acid, p-Cresol sulfate, 1-Methylguanosine, Indoxyl sulfate, and Uridine were significantly upregulated ( , ：Details of t-test for differential metabolites between OVX + E group and OVX group). The hierarchical clustering dendrogram clearly distinguished the samples of different subgroups . The hierarchical clustering heatmap visually displayed the above results . Although PCA analysis as an unsupervised analysis could effectively differentiate between the OVX and OVX + E groups, it faced difficulties in distinguishing differences between the OVX and Sham groups . In contrast, supervised methods were more adept at discerning intergroup differences, but these models also had certain limitations. The PLS-DA model had good intergroup discriminative ability , but showed clear signs of model overfitting (Figure S6E-F). The OPLS-DA model also had advantages in discriminative ability , and the permutations test results indicated that the R 2 Y values of the OPLS-DA models between OVX and the other two groups were 0.981 (p=0.009) and 0.986 (p=0.001), with Q 2 values of 0.848 (p<0.002) and 0.893 (p=0.001; ). The cross-validation results were consistent with the permutations test , demonstrating the high predictive value of the models. Utilizing the VIP scores and S-Plot of this model, we further identified four important differential metabolites, namely L-AABA, 4-Hydroxyproline, O-dodecanoylcarnitine, and Methylpicraquassioside A ( , , ：VIP scores of OPLS-DA models between subgroups). The Random Forest model exhibited strong advantages in determining intergroup differences, with an OOB error of 0 , and the VIP plots based on contribution to classification accuracy provided several promising differential metabolites ( , , ：VIP scores of the random forest model between subgroups), which still included L-AABA. The evaluation results of EBAM and SAM are shown in ( ：The evaluation details of EBAM and SAM model between subgroups).
Metabolite expression characteristics are another analytical strategy to explore the impact of low estrogen on the aorta. In the subgroup analysis, using the OVX group as a control, representative metabolites associated with the Sham group and OVX + E group are shown in ; , : Details of intergroup differential analysis of low estrogen related metabolites. Among them, the metabolites with a positive correlation coefficient exceeding 0.8 include Methylpicraquassioside A, L-AABA, and D-Ribose 5-phosphate (Sham group), as well as L-AABA, 4-Guanidinobutanoic acid, Methylpicraquassioside A, and L-Erythrulose (OVX +E group). The trend arranged from low to high estrogen concentrations (OVX-Sham-OVX+E) is shown in , ：Details of metabolite related trends from low to high estrogen concentrations. L-AABA exhibits a significant positive correlation feature (correlation coefficient of 0.89), while Inosinic acid is the only metabolite with a negative correlation coefficient exceeding –0.7 (correlation coefficient of –0.7). Therefore, we further explored metabolites with correlated expression to L-AABA ( , : Correlation coefficient of metabolites related to L-AABA expression), where L-Erythrulose shows the most positive correlation (correlation coefficient of 0.87171), and Inosinic acid shows the most negative correlation (correlation coefficient of –0.62298).
Univariate receiver operating characteristic (ROC) curve analysis was used to screen promising biomarkers, and L-AABA was found to be the most promising biomarker, with an AUC of 1 in the comparison process between OVX and the other two groups ( , , : The promising differential metabolites AUC results of ROC curves between subgroups). Multivariate ROC curve analysis demonstrated unique value in biomarker selection, with the AUC consistently above 0.93 starting from 5 variables . The error classification of the multivariate ROC curve showed that there were no misclassifications when using the model for samples in the OVX and OVX +E groups, but one sample from the Sham group was incorrectly classified into the OVX group . In this multivariate model, L-AABA still ranked in the top three in terms of average importance .
Further enrichment analysis of differential metabolites was conducted to elucidate the underlying mechanisms. Differential metabolites showed a wide range of classifications, with amino acids and peptides, fatty acids and conjugates, and monosaccharides ranking among the top three . Results of metabolic pathway enrichment analysis indicated that the Warburg effect, glycolysis, and gluconeogenesis were the top three enriched metabolic pathways . Enzyme-specific metabolic analysis revealed that deoxyuridine phosphorylase (UPP) and O 2 transport (diffusion) ranked in the top two . To eliminate the impact of species differences on enrichment results, we specifically conducted rat-specific metabolic pathway analysis, which showed significant differences in the Pentose phosphate pathway, Glycerophospholipid metabolism, Arginine and proline metabolism, and Pyrimidine metabolism ( , ：Rat-specific metabolic pathway analysis details).
The primary objective of this study is to explore the correlation and potential mechanisms between low estrogen status and postmenopausal hypertension through metabolomics analysis. The modeling in this study is consistent with previous research results, where estrogen levels significantly decrease after ovariectomy, and exogenous estrogen supplementation can restore estrogen to physiological levels. Correspondingly, the SBP, DBP, and PP of the OVX group are significantly higher than those of the control group (Sham group), while after estrogen supplementation (OVX + E group), the above blood pressure parameters can return to the level of the control group . The phenotypic changes mentioned above support further exploration of the mechanisms related to estrogen and hypertension. Through various statistical analysis methods, we attempt to identify the most important mediator between estrogen and postmenopausal hypertension from a large number of differential metabolites. After extensive screening, we believe that L-AABA is the most promising biomarker. L-AABA plays a crucial role in certain metabolic pathways in the human body and can be used as a supplement to support overall health . It is an optically active form of α-aminobutyric acid (AABA), a non-essential amino acid primarily derived from the breakdown metabolism of methionine, threonine, and serine . It serves as a biomarker for various diseases such as esophageal variceal bleeding and septic liver injury . Recent studies have linked AABA to female osteoporosis , a common condition in postmenopausal women, similar to postmenopausal hypertension. Our research found a significant downregulation of AABA in the aortic tissues of the OVX group, and estrogen supplementation could reverse the decrease in AABA levels, indicating that AABA may act as a protective factor in postmenopausal hypertension. Based on our knowledge, AABA plays an important role in cardiovascular diseases: elevated levels of AABA were found in the heart tissues of dilated cardiomyopathy Syrian hamster models , significant increase of serum AABA in patients with atrial septal defect compared to healthy volunteers, and the ability of AABA levels to decrease to that of healthy volunteers after ductal occlusion, animal experiments have confirmed that AABA can prevent doxorubicin-induced cardiomyopathy in mice by increasing circulation and myocardial glutathione levels . It is worth noting that there has been no research exploring the correlation and potential mechanisms of AABA with hypertension. Current studies believe that high-intensity interval exercise has a corrective effect on hypertension , and research has found that high-intensity interval exercise can increase plasma AABA concentrations , consistent with the trend of our experiments, suggesting that AABA may be a protective factor for hypertension. Although it is currently unclear whether AABA is involved in the improvement of hypertension, existing research results are enough to provide some hints for future research: vascular remodeling is a classic change in the pathophysiological process of hypertension and early vascular remodeling allows vessels to adapt to transient changes in hemodynamics and play compensatory protective effects. However, when blood pressure continues to increase, vessels cannot continue to compensate and undergo adverse restructuring, ultimately leading to a vicious cycle of lumen narrowing and rising blood pressure . Although the related mechanisms of vascular remodeling in hypertensive patients have not been fully elucidated, current research suggests that the accumulation of macrophages in the vascular lumen and the inflammation resulting from it play an important role in the progression of vascular remodeling . It is worth noting that recent studies have confirmed that AABA can regulate the polarization and function of M1 macrophages by promoting oxidative phosphorylation and inhibiting glycolysis through the Warburg effect and gluconeogenesis . Our rat-specific pathway enrichment results also coincidentally enriched in three important metabolic pathways: Warburg effect, glycolysis, and gluconeogenesis. Based on the above experimental results, we propose the following hypothesis: the low estrogen state caused by menopause leads to a decrease in L-AABA in the aorta, mediates macrophage activation, causes vascular remodeling, and ultimately leads to the occurrence of postmenopausal hypertension. Our team is conducting further research to verify the above points. Enrichment analysis results indicate that the top three differential metabolites are amino acids and peptides, fatty acids and conjugates, and monosaccharides, which is in line with expectations as sugars, fats, and proteins are the three main types of metabolites in the human body. To avoid biases in research results caused by species differences, we further analyzed the specific metabolic pathways in rats, and the results show that the Warburg effect, glycolysis, and gluconeogenesis are the most enriched metabolic pathways. This suggests that sugar metabolism is the most significantly affected aspect of aortic metabolism in a low estrogen state. Current research generally believes that enhanced Warburg effect and glycolysis-mediated pulmonary vasoconstriction are key mechanisms in the development of pulmonary arterial hypertension . Therefore, it is reasonable to speculate that changes in aortic Warburg effect and glycolysis may be involved in the occurrence of Postmenopausal Hypertension. Furthermore, researchers from the United States have found that the main culprit mediating the Warburg effect in tumors is macrophages in tumor tissue . Whether aortic remodeling in postmenopausal women is also mediated by macrophage glucose metabolism reprogramming is a research direction worthy of exploration. Finally, in the exploration of key enzymes, we found deoxyuridine phosphorylase and O 2 transport (diffusion) as the top two candidates. In pancreatic cancer, UPP1 mediates the redox balance, survival, and proliferation of tumor cells under hypoxic conditions , and considering the enrichment of O 2 transport-related enzymes, the UPP-mediated hypoxic adaptation mechanism may also be an important participant in Postmenopausal Hypertension. The current study is an exploratory work on the relationship between low estrogen and Postmenopausal Hypertension. Although some meaningful findings have been obtained, there are still some aspects that need further improvement. Firstly, there are several ways to construct animal models of menopause, including X-ray induced ovarian injury animal model , ovariectomized animal model , and chemical-induced ovarian failure model . Among them, the ovariectomy animal model is the most commonly used at present, but it has limitations in simulating the drastic fluctuations in estrogen levels after ovariectomy. Secondly, the hypothesis that L-AABA regulates macrophage function to mediate blood pressure regulation has not been experimentally confirmed. Lastly, there is a lack of direct evidence of metabolic changes from human samples.
Animal models The experimental animals for this study were purchased from the Department of Laboratory Animal Science at Peking University Health Science Center. The animals’ care and handling followed the guidelines established by the Animal Experimental Control and Supervision Committee, along with the Declaration of Helsinki by the World Medical Association regarding ethical standards for medical research involving animals. Approval for the experimental protocol was obtained from the Laboratory Animal Welfare Ethics Committee (No. LA2018092). Twenty-four 12-week-old female Sprague Dawley rats (Without any genetic modification) of SPF grade were randomly assigned to three groups: the Sham surgery group (Sham), ovariectomy group (OVX), and ovariectomy group treated with estrogen (OVX +E). These animals were housed under controlled conditions, including temperature (22–26°C), humidity (50–60%), and a 12 hr light/12 hr dark cycle, provided with a non-soy diet and ad libitum access to water. Surgery was performed one week after acclimatization. The Sham group underwent skin incision and closure, while the OVX and OVX +E groups underwent ovariectomy. On the 14th day post-surgery, all rats received subcutaneous injections of specific drugs between 9am and 10am daily for four weeks. Rats in the OVX +E group were administered 17β-estradiol (25 μg/kg/day; Sigma, St. Louis, MO, USA) dissolved in ethanol and diluted with sterile sesame oil (10 mg/0.1 mL, 0.25 mL/kg; GLBIO, Montclair, California, USA). The other groups were given an equal volume of sterile sesame oil. Starting from the third day post-operation, vaginal exfoliated cells were smeared daily for 7 days consecutively to confirm the successful establishment of the model. Vaginal cells were collected using a cotton swab dampened in 0.9% saline, stained with hematoxylin and eosin, dehydrated with alcohol, clarified with xylene, and ultimately preserved with resin. Blood pressure measurement After 4 weeks of ovariectomy, blood pressure was measured in 24 rats every night from 22:00 to 24:00 for the next seven days. The first 6 days were used for adaptation training to reduce the impact of the surgery and surroundings on blood pressure. After blood pressure values stabilized, data collection began on the seventh day. The CODA-HT6 non-invasive blood pressure system (Kent Scientific Corporation, CT, USA) was used for the measurements. The cuff was positioned 1 cm from the tail base, attaching the VPR sensor to the tail. The tubing and VPR sensor were gradually inflated until the tail was completely blocked off, and then gradually deflated. Systolic blood pressure (SBP) was measured when blood flow was detected in the artery. The peak slope of the blood pressure change as determined by VPR revealed the diastolic blood pressure (DBP). Pulse pressure (PP) is equal to SBP minus DBP. Every rat was subjected to ten to fifteen measurement cycles; the average of the last five cycles was taken as the final result. Harvesting aorta and blood samples All rats were killed after having their blood pressure measured and given an intraperitoneal injection of 1% pentobarbital sodium (80 mg/kg; Sigma, St. Louis, MO, USA). After drawing blood samples from the heart, they were centrifuged at 4 °C. A cold 0.9% saline solution was utilized to perfuse the heart prior to the collection of aortic tissues. After that, the whole aorta and the serum fraction were kept in storage at –80 °C. Radioimmunoassay With a detection limit of 3 pg/mL, the Rat E2 ELISA kit (RE1649-48T, Bioroyee, Beijing, China) was used for radioimmunoassay to evaluate serum estrogen levels. Samples were separated, centrifuged, and then incubated for analysis in accordance with the instructions. Preparing the aorta for metabolomics Precisely weigh out 20±1 mg of aorta tissue, and then combine it with 500 µL of methanol containing an internal standard of 5 µg/mL 2-chloro-l-phenylalanine. The mixture was homogenized for 90 s using a high-throughput tissue grinder (60 Hz; Tissuelyser-24, Jingxin, Shanghai, China). A 15-min centrifugation at 12,000 rpm and 4 °C resulted in the separation of 100 µL of the supernatant for analysis using metabolomics. Metabolomics measurement The UHPLC-Q-TOF technique was used to analyze the metabolomics of aorta tissue. ACQUITY UPLC HSS T3 columns (1.8 μm, 2.1 mm × 100 mm, Waters, Dublin, Ireland) were used as the chromatographic column in the Agilent 1290 II UPLC-QTOF 5600 PLUS (Sciex) liquid chromatography-mass spectrometry system, which was used in the electric spray ionization (ESI) mode. Curtain gas = 35, ion spray voltage = 5500 V (positive ion mode) and –4500 V (negative ion mode), temperature = 450 °C, ion source gas 1=50, and ion source gas 2=50 were the parameters for liquid chromatography-mass spectrometry. Software for Agilent MassHunter workstations (version B.01.04; Agilent, Lexington, MA, USA) was used to process raw data. By increasing the intensity threshold to 300, noise was filtered out and isotope interference was eliminated. The method of identifying metabolite was by comparison with the publicly available METLIN database ( here , access date: 9 October 2021). Data preprocessing and bioinformatics analysis In the current research, data preprocessing and bioinformatics analysis were conducted using MetaboAnalyst 5.0 ( http://www.metaboanalyst.ca/ ; visited on October 12, 2021). The Sham group served as the reference group while the data were normalized using the group probability quotient normalization technique . Data normalization was accomplished using log transformation (base 10) and the Pareto approach (mean-centered, divided by the square root of each variable’s standard deviation). Several analytical techniques were used for information mining, such as the Debiased Sparse Partial Correlation (DSPC) network summarizes the aforementioned study procedure. Conclusion This study for the first time delineated the metabolic characteristics of the aorta under low estrogen status and explored potential mechanisms of postmenopausal hypertension. Sugar metabolism reprogramming plays an important role in Postmenopausal Hypertension, and AABA may be a key link in the pathogenic mechanism. The aforementioned mechanisms may be the future focus of work on Postmenopausal Hypertension and deserve further in-depth exploration.
The experimental animals for this study were purchased from the Department of Laboratory Animal Science at Peking University Health Science Center. The animals’ care and handling followed the guidelines established by the Animal Experimental Control and Supervision Committee, along with the Declaration of Helsinki by the World Medical Association regarding ethical standards for medical research involving animals. Approval for the experimental protocol was obtained from the Laboratory Animal Welfare Ethics Committee (No. LA2018092). Twenty-four 12-week-old female Sprague Dawley rats (Without any genetic modification) of SPF grade were randomly assigned to three groups: the Sham surgery group (Sham), ovariectomy group (OVX), and ovariectomy group treated with estrogen (OVX +E). These animals were housed under controlled conditions, including temperature (22–26°C), humidity (50–60%), and a 12 hr light/12 hr dark cycle, provided with a non-soy diet and ad libitum access to water. Surgery was performed one week after acclimatization. The Sham group underwent skin incision and closure, while the OVX and OVX +E groups underwent ovariectomy. On the 14th day post-surgery, all rats received subcutaneous injections of specific drugs between 9am and 10am daily for four weeks. Rats in the OVX +E group were administered 17β-estradiol (25 μg/kg/day; Sigma, St. Louis, MO, USA) dissolved in ethanol and diluted with sterile sesame oil (10 mg/0.1 mL, 0.25 mL/kg; GLBIO, Montclair, California, USA). The other groups were given an equal volume of sterile sesame oil. Starting from the third day post-operation, vaginal exfoliated cells were smeared daily for 7 days consecutively to confirm the successful establishment of the model. Vaginal cells were collected using a cotton swab dampened in 0.9% saline, stained with hematoxylin and eosin, dehydrated with alcohol, clarified with xylene, and ultimately preserved with resin.
After 4 weeks of ovariectomy, blood pressure was measured in 24 rats every night from 22:00 to 24:00 for the next seven days. The first 6 days were used for adaptation training to reduce the impact of the surgery and surroundings on blood pressure. After blood pressure values stabilized, data collection began on the seventh day. The CODA-HT6 non-invasive blood pressure system (Kent Scientific Corporation, CT, USA) was used for the measurements. The cuff was positioned 1 cm from the tail base, attaching the VPR sensor to the tail. The tubing and VPR sensor were gradually inflated until the tail was completely blocked off, and then gradually deflated. Systolic blood pressure (SBP) was measured when blood flow was detected in the artery. The peak slope of the blood pressure change as determined by VPR revealed the diastolic blood pressure (DBP). Pulse pressure (PP) is equal to SBP minus DBP. Every rat was subjected to ten to fifteen measurement cycles; the average of the last five cycles was taken as the final result.
All rats were killed after having their blood pressure measured and given an intraperitoneal injection of 1% pentobarbital sodium (80 mg/kg; Sigma, St. Louis, MO, USA). After drawing blood samples from the heart, they were centrifuged at 4 °C. A cold 0.9% saline solution was utilized to perfuse the heart prior to the collection of aortic tissues. After that, the whole aorta and the serum fraction were kept in storage at –80 °C.
With a detection limit of 3 pg/mL, the Rat E2 ELISA kit (RE1649-48T, Bioroyee, Beijing, China) was used for radioimmunoassay to evaluate serum estrogen levels. Samples were separated, centrifuged, and then incubated for analysis in accordance with the instructions.
Precisely weigh out 20±1 mg of aorta tissue, and then combine it with 500 µL of methanol containing an internal standard of 5 µg/mL 2-chloro-l-phenylalanine. The mixture was homogenized for 90 s using a high-throughput tissue grinder (60 Hz; Tissuelyser-24, Jingxin, Shanghai, China). A 15-min centrifugation at 12,000 rpm and 4 °C resulted in the separation of 100 µL of the supernatant for analysis using metabolomics.
The UHPLC-Q-TOF technique was used to analyze the metabolomics of aorta tissue. ACQUITY UPLC HSS T3 columns (1.8 μm, 2.1 mm × 100 mm, Waters, Dublin, Ireland) were used as the chromatographic column in the Agilent 1290 II UPLC-QTOF 5600 PLUS (Sciex) liquid chromatography-mass spectrometry system, which was used in the electric spray ionization (ESI) mode. Curtain gas = 35, ion spray voltage = 5500 V (positive ion mode) and –4500 V (negative ion mode), temperature = 450 °C, ion source gas 1=50, and ion source gas 2=50 were the parameters for liquid chromatography-mass spectrometry. Software for Agilent MassHunter workstations (version B.01.04; Agilent, Lexington, MA, USA) was used to process raw data. By increasing the intensity threshold to 300, noise was filtered out and isotope interference was eliminated. The method of identifying metabolite was by comparison with the publicly available METLIN database ( here , access date: 9 October 2021).
In the current research, data preprocessing and bioinformatics analysis were conducted using MetaboAnalyst 5.0 ( http://www.metaboanalyst.ca/ ; visited on October 12, 2021). The Sham group served as the reference group while the data were normalized using the group probability quotient normalization technique . Data normalization was accomplished using log transformation (base 10) and the Pareto approach (mean-centered, divided by the square root of each variable’s standard deviation). Several analytical techniques were used for information mining, such as the Debiased Sparse Partial Correlation (DSPC) network summarizes the aforementioned study procedure.
This study for the first time delineated the metabolic characteristics of the aorta under low estrogen status and explored potential mechanisms of postmenopausal hypertension. Sugar metabolism reprogramming plays an important role in Postmenopausal Hypertension, and AABA may be a key link in the pathogenic mechanism. The aforementioned mechanisms may be the future focus of work on Postmenopausal Hypertension and deserve further in-depth exploration.
|
The Role of Continuous Monitoring of Venous Drainage Flow and Integrated Oxygen Extraction (ER | 79d11b61-70f1-4bfc-9363-d92f7bf6b317 | 11857250 | Surgical Procedures, Operative[mh] | Aortic arch surgery, particularly procedures involving selective antegrade cerebral perfusion (SACP) under moderate hypothermia, demands careful management to protect neurological function during periods of circulatory arrest . The real-time monitoring of cerebral perfusion has become an essential component of these procedures, with tools like near-infrared spectroscopy (NIRS) providing valuable insights into cerebral oxygenation and metabolic status. NIRS, which measures regional oxygen saturation (rSO 2 ), is non-invasive and widely used, but its ability to precisely reflect the balance between oxygen delivery and consumption remains a topic of investigation . Recent advancements have introduced the continuous monitoring of venous drainage flow and integrated oxygen extraction (ERiO 2 ) as complementary methods to assess cerebral perfusion more comprehensively . Venous drainage flow reflects the adequacy of cerebral outflow, while ERiO 2 quantifies the proportion of oxygen extracted from the delivered supply, offering a direct measure of metabolic activity . The interplay between these parameters and NIRS readings may reveal critical insights into the efficiency of cerebral perfusion strategies. However, the relationship between these variables is not fully understood, and questions remain about the reliability of NIRS as a surrogate marker for cerebral metabolic status in comparison to ERiO 2 and venous drainage measurements . Additionally, factors influencing these relationships, such as patient-specific physiology, cannula positioning, and perfusion flow rates, add complexity to their interpretation . This retrospective study aims to analyze the correlation between NIRS-derived rSO 2 , continuous venous drainage flow, and ERiO 2 during aortic arch surgery utilizing the Kazui technique. By examining data collected intraoperatively from patients undergoing SACP, this study seeks to explore the interplay among these variables and their potential role in guiding perfusion management. Understanding these relationships could enhance the precision of intraoperative monitoring, improve decision making, and ultimately contribute to better neurological outcomes in complex aortic arch procedures.
This retrospective study analyzed data from ten patients who underwent aortic arch surgery using the Kazui technique for selective antegrade cerebral perfusion (SACP) under profound hypothermia. The study period spanned from November 2022 to November 2024 and included four patients diagnosed with type I A aortic dissections (urgent) and six with aortic arch aneurysms. Surgeries were performed following a standardized protocol designed to optimize cerebral perfusion and minimize neurological risks . Approval was obtained from the Internal Institutional Review Board (IRB), and all protocols complied with the principles outlined in the Declaration of Helsinki. The study adhered strictly to international and institutional guidelines for the ethical conduct of research involving human participants. Due to the retrospective design of the study and the use of de-identified patient data, the requirement for individual informed consent was waived. Surgical procedures were conducted with standard cardiopulmonary bypass (CPB) initiated via peripheral cannulation of the femoral artery (Bio-medicus 19 Fr by Medtronic, Minneapolis, MN, USA) and the femoral vein (23–25 Fr by Livanova, UK), and Custoidol ® HTK solution was used for myocardial protection in all procedures. Systemic cooling was performed to achieve a target temperature of 20 °C, monitored through nasopharyngeal and rectal temperature sensors. Upon reaching the target temperature, systemic circulatory arrest was instituted, and bilateral SACP (13 FR manual insufflation cannulas by Medtronic, Minneapolis, MN, USA) was initiated, delivering oxygenated blood at a flow rate calculated as 10 mL/kg based on the patient’s body weight. Perfusion flow was dynamically managed by integrating measured, calculated, and recorded parameters, including right radial arterial pressure, NIRS-derived regional oxygen saturation (rSO 2 ) (Masimo), oxygen extraction ratio (ERiO 2 ), and venous drainage flow. During cerebral perfusion, continuous monitoring was performed using bilateral near-infrared spectroscopy (NIRS) to assess rSO 2 (Root ® Patient Monitoring by Masimo) , while cerebral venous drainage flow and ERiO 2 were measured using the Landing monitoring system (Eurosets, Medolla SRL, Italy) ( and ). These parameters were used collectively to guide perfusion adjustments and ensure optimal delivery and metabolic balance during the procedure. Following systemic circulatory arrest and the cerebral perfusion phase, reperfusion was initiated for 10 min at a nasopharyngeal temperature of 20 °C, with blood temperature matching the nasopharyngeal target to minimize thermal gradients. Subsequently, rewarming was carried out, maintaining a temperature gradient of 10 °C between the blood and rectal temperature, until a target core temperature of 36.8 °C was reached. This controlled gradient ensured gradual rewarming and reduced the risk of ischemic–reperfusion injury or excessive thermal stress. All intraoperative parameters, including CPB time, cross-clamp time, cooling and rewarming phases, perfusion pressures, and monitored variables (rSO2, ERiO2, and venous return flow), were recorded. Mean values and standard deviations were calculated for quantitative variables, and Pearson correlation coefficients were used to assess the relationships among NIRS, ERiO 2 , and venous drainage flow. Data analysis was performed using SPSS Statistics Version 28.0, with statistical significance set at p < 0.05. This analysis aimed to evaluate the interplay between continuous cerebral monitoring parameters and their correlation with optimal cerebral perfusion during SACP, with secondary outcomes, including postoperative neurological status, assessed using standardized evaluation protocols.
This study included a total of 10 patients who underwent aortic arch surgery using the Kazui technique under profound hypothermia. The cohort consisted of four patients with type A aortic dissections and six with aortic arch aneurysms . Intraoperative data revealed that the mean cardiopulmonary bypass (CPB) time was 182 ± 15 min, and the mean cross-clamp time was 98 ± 12 min. The cooling phase to achieve a target temperature of 20 °C lasted 29 ± 3 min, followed by a reperfusion phase at 20 °C for 10 ± 1.5 min. Rewarming was completed in 40 ± 5 min, with a controlled gradient of 10 °C between the blood and rectal temperature, reaching a final core temperature of 36.8 °C. The bilateral selective antegrade cerebral perfusion (SACP) flow rates averaged 620 ± 30 mL/min, while the venous return flow averaged 570 ± 25 mL/min. The cerebral perfusion time during SACP was 40 ± 6 min, and the circulatory arrest time was 42 ± 7 min. Continuous cerebral monitoring showed stable bilateral NIRS-derived regional oxygen saturation (rSO2) values, averaging 65 ± 5% , and a mean oxygen extraction ratio (ERiO2) of 28 ± 4% . During the cooling phase, a reduction in oxygen extraction was observed, while, during the warming phase, the ERiO2 increased significantly. This dynamic behavior highlighted the responsiveness of cerebral oxygenation to temperature modulation . A strong correlation was observed between rSO2 and ERiO2 (r = 0.91, p < 0.01), underscoring the consistency between cerebral oxygenation and metabolic activity during perfusion . Postoperatively, all patients demonstrated favorable neurological outcomes, with no cases of stroke, transient ischemic attacks, or major neurological deficits. The mean duration of mechanical ventilation was 8.5 ± 2.3 h, the mean intensive care unit (ICU) stay was 48 ± 12 h, and the mean hospital stay was 10 ± 3 days. No significant complications related to the perfusion strategy were observed, and the 30-day mortality rate was 0% .
This study highlights the feasibility and effectiveness of integrating near-infrared spectroscopy (NIRS), oxygen extraction ratio (ERiO2), and continuous venous drainage flow monitoring to optimize cerebral perfusion during aortic arch surgery using the Kazui technique . The observed correlation between rSO2 and ERiO2 across the phases of cardiopulmonary bypass (CPB) offers valuable insights into the dynamics of cerebral oxygenation and metabolism under profound hypothermia . By offering a more comprehensive assessment of cerebral perfusion, this approach could be pivotal in minimizing neurological complications and optimizing surgical results. Despite these promising findings, several limitations must be acknowledged when interpreting the results. First, the small sample size of ten patients limits the generalizability of the results, necessitating larger multicenter studies to validate these findings. The single-center nature of the study introduces potential bias, as the results may be influenced by standardized institutional protocols, which might not be representative of broader practices. Furthermore, the retrospective design is inherently limited by the absence of randomization and the inability to control for all potential confounders. Unmeasured variables, such as individual patient comorbidities or slight differences in surgical technique, may have influenced the outcomes. Additionally, this study focuses exclusively on short-term outcomes, such as neurological status, ICU, and hospital stay durations, and does not address long-term follow-up or the durability of these favorable results. Moreover, the absence of a control group or comparative cohort prevents definitive conclusions about the superiority of this monitoring approach over other strategies. The accuracy of NIRS measurements can be affected by extracranial contamination or variations in systemic factors like hemoglobin levels and the accuracy of ERiO2 under varying conditions. Similarly, ERiO2 interpretation depends on multiple factors, including perfusion pressure and flow dynamics, which were not fully explored in this study. While profound hypothermia is widely accepted as a strategy for cerebral protection during aortic arch surgery, it carries inherent risks, such as coagulopathy and prolonged CPB times, which were not specifically evaluated for their impact on the observed outcomes. Future efforts should focus on conducting larger multicenter prospective trials, incorporating control or comparative groups, and performing long-term follow-up to assess the sustainability of these outcomes. Investigating the interplay between NIRS, ERiO2, and systemic parameters, such as blood pressure, hemoglobin levels, and CPB flows, could refine perfusion protocols. Additionally, advanced analytical approaches, such as machine learning, could leverage real-time data from NIRS and ERiO2 to develop predictive models for optimizing cerebral perfusion strategies . This study underscores the importance of integrating NIRS, ERiO2, and venous drainage flow monitoring in managing cerebral perfusion during complex aortic surgeries. While the findings are promising, further research is essential to address the limitations and refine strategies for achieving optimal neurological outcomes.
This preliminary study provides insights into the potential benefits of integrating the continuous monitoring of venous drainage flow, oxygen extraction ratio (ERiO2), and NIRS-derived regional oxygen saturation (rSO2) during aortic arch surgery using the Kazui technique. Our observations indicate a correlation between rSO2 and ERiO2, suggesting that these parameters may be complementary in assessing cerebral oxygenation and metabolic balance. Venous drainage flow also appears to play a crucial role in maintaining cerebral perfusion and oxygen delivery. While these findings are encouraging, it is important to acknowledge that they are based on a limited dataset from a single-center study with a small sample size. Therefore, conclusions regarding the efficacy and safety of the monitoring approach should be viewed with caution. The data do not conclusively prove the superiority of this technique over others, but rather point to potential areas where such integrated monitoring could enhance real-time decision making and surgical outcomes. Future research should focus on validating these preliminary results in a larger and more diverse cohort. Expanding the study to include multicenter data would help in assessing the reproducibility and generalizability of the findings. Moreover, exploring long-term outcomes and integrating advanced predictive models could provide deeper insights into refining cerebral perfusion strategies during complex cardiovascular procedures. This approach would further substantiate the role of continuous monitoring modalities in improving patient care in aortic arch surgeries.
|
Hodgkin Lymphoma Presenting With Spinal Cord Compression: Challenges for Diagnosis and Initial Management | ee663493-efa8-4296-bede-37ce8365bfff | 9109237 | Anatomy[mh] | Hodgkin lymphoma (HL) is primarily a malignancy of the lymph nodes but can present with extra-nodal involvement. Incidence follows a bimodal age distribution, rising sharply during childhood and peaking initially in young adults aged 20-24 years (y), with a second larger peak in incidence during the eighth decade of life. In the UK, 435 children and young adults (<25y of age) are diagnosed with HL each year, with an average incidence rate of 9.4 per 100,000 for those 20-24y. HL is characterised by the presence of binucleated giant cells termed Reed-Sternberg cells or large mononuclear cell variants (lymphocytic and histiocytic cells) on a background of inflammatory cells and is broadly divided into two pathologic classes: classic or nodular lymphocyte predominant disease. , Immunophenotyping is essential to distinguish between the two classes, with classic HL expressing CD30 in nearly all cases and CD15 in 75-85%, whilst typically lacking expression of the B-cell markers (CD19, CD20 and CD79a). , PAX-5 is the only B-cell restricted antigen that is nearly always expressed in classic HL, but staining is often weaker than that seen in reactive B-cells and is a useful diagnostic feature. Classic HL accounts for the majority of childhood, adolescent and young adult cases, , with the nodular sclerosis type being the most common subtype, primarily affecting adolescents and young adults (15-34y) with a female predominance. – Treatment advances have dramatically increased HL survival over the past 40 years across all patient age-groups. Age-standardised 10y survival has increased from 47% in 1971 to 80% in 2011 in the UK. Outcomes are better still for children and young adults, with 5y survival exceeding 95% across much of Europe and the US. , , Current trials in this group are aimed at maintaining overall survival whilst minimising the morbidity associated with the late-effects of treatment, with a move away from radiotherapy as standard treatment. , A prolonged time-to-diagnosis is recognised in patients with cancer, including HL, – which may result in more advanced stage at presentation. Although overall survival is generally not compromised in such cases, advanced stage disease requires more treatment for cure and hence late-effects are likely to be more substantial. Consequently, raising awareness of such extra-nodal cases is important to minimise time-to-diagnosis and late sequelae. In particular, extra-nodal HL only very rarely presents with spinal cord compression (SCC), described in only a few rare case reports. – Hence, it may not be in the list of potential differential diagnoses and management of such patients may be suboptimal as a result. Here, we describe the case of a pediatric, adolescent patient presenting with SCC and discuss the challenges for HL diagnosis and initial management.
A 15-year-old female presented to primary care after 2 months (68 days) of left hip pain and shooting pains in the left shin. She was given non-steroidal anti-inflammatory drugs with some improvement, and referred to pediatric rheumatology after blood testing revealed raised inflammatory markers [high erythrocyte sedimentation rate (ESR) 76mm/hr and C-reactive protein (CRP) 17mg/L]. Full blood count (FBC) at this time revealed hemoglobin 127g/L, white cell count 10.7 × 10 9 /L, and platelets 444 × 10 9 /L. Urea and electrolytes (U&E) and bone (BFT) and liver function tests (LFT) were normal and antinuclear antibody (ANA) and rheumatoid factor were negative. At four months from symptom onset (117 days), she was seen in pediatric rheumatology, where clinical examination was normal, and further investigations revealed persistently elevated CRP (44mg/L) and ESR (75mm/hr); other bloods including repeat FBC, U&E, LFTs, creatine kinase (CK) and immunoglobulins were all normal. Five months after symptoms onset (159 days) left hip pain was worse, and lower back pain and stiffness had developed causing antalgic gait. The MRI of hip, spine and pelvis was delayed due to the COVID-19 pandemic, and undertaken seven months after symptom onset (208 days), showing an epidural soft tissue mass, resulting in referral to pediatric oncology for suspected cancer. By this time, examination findings were bilaterally absent lower limb reflexes only, with no clinical evidence of cauda equina syndrome. No palpable lymphadenopathy was detected. The MRI scan demonstrated widespread lymphadenopathy, with multiple enlarged carotid and submandibular lymph nodes as well as numerous para-aortic lymph nodes extending around the fourth lumbar (L4) vertebral body and via the L3/4 and L4/5 neural foramina bilaterally into a largely anterior epidural mass ( ). The mass extended inferiorly, occupying most of the sacral canal and extended anteriorly through the sacral foramina. The thecal sac was displaced and moderately compressed at the lumbar region below L3 with enhancement of the cauda equina and severe thecal compression in the sacrum ( ). Radiologically, the most likely differential diagnosis was lymphoma. Following multidisciplinary discussion, it was determined that cauda equina syndrome was imminent and therefore surgical debulking was undertaken, both to prevent this complication and establish a diagnosis, with emergency decompression of the theca and left L4 nerve root. At surgery, the tumor was noted to be highly vascular. Frozen section confirmed lesional material. Following surgery, a short course of oral steroids was commenced to reduce any peri-surgical edema. However, the material presented for histopathologic examination was ultimately non-diagnostic, both by conventional pathology and urgent flow cytometry. Samples showed poorly preserved connective tissue with diathermy and freezing artefacts. An excess of eosinophils was seen with some larger cells noted ( ). Immunostaining of these larger, atypical cells showed CD30 and very weak MUM-1 expression, but overall did not confirm a diagnosis of HL as expression of CD15 and PAX-5 was not seen ( ) and morphology was not typical. Flow cytometry was non-contributory. Further immunostains, including negative CD1a (excluding Langerhans Cell Histiocytosis) and ALK (excluding Anaplastic Large Cell Lymphoma) were performed, but with concerns noted over the suitability of the tissue for such immunostaining given the diathermy and freezing artefacts. All available lesional material was used for these analyses but unfortunately a diagnosis was not secured. Consequently, and following further MDT discussion, open surgical excisional biopsy of (non-palpable) cervical lymph nodes was performed five days later as guided by ultrasound findings, along with bone marrow sampling due to the putative HL diagnosis, and insertion of a central venous access device in order to deliver systemic chemotherapy. On histopathology, effacement of the normal lymph node architecture was observed, with prominent bands of fibrosis and a marked infiltrate of eosinophils. A conspicuous population of larger pale cells was present with varying nuclear morphology ( ). These larger, atypical cells stained for CD15 and CD30 with weak PAX-5 positivity ( ). Immunostains for BCL-6, CD79a, CD20, CD45 and ALK were all negative. The findings were consistent with nodular sclerosis classic HL. Bone marrow trephines demonstrated the presence of atypical cells and an excess of eosinophils, confirming the presence of (stage 4) metastatic disease ( ). A staging fluoro-deoxyglucose (FDG) positron-emission-tomography (PET) scan demonstrated high tracer uptake in enlarged lymph nodes on both sides of the diaphragm, spleen, retroperitoneal nodules, paraspinal soft tissue and bone marrow, consistent with stage 4 disease. During the final few days of diagnostic work-up, prior to commencement of treatment, the patient developed ‘B’ symptoms (night sweats) and therefore assigned 4B staging. The patient was enrolled on the EuroNet-PHL-C2 Phase 3 trial for children and young adults (<25y) with classic HL, in treatment level 3 (TL3), and responded well to treatment. Early-response-assessment FDG-PET after two cycles of induction ‘OEPA’ chemotherapy confirmed the need for radiotherapy following completion of chemotherapy. Late-response-assessment FDG-PET after four cycles of consolidation ‘COPDAC’ chemotherapy showed further treatment response but showed some residual increased uptake at the site of disease around the L4 vertebral body, confirming the need for radiotherapy boost to this area.
This case highlights a number of challenges in establishing the diagnosis, and of the early management of spinal cord compression (SCC), in the context of previously undiagnosed Hodgkin lymphoma (HL). Prompt diagnosis of HL is important to minimise the morbidity associated with the late-effects of the more intensive treatment required to cure high-stage disease, and, in the setting of SCC, to optimise functional outcomes. Whilst we are not aware of published data on the time-to-diagnosis of HL in purely pediatric populations, a recent published study found that self-reported time-to-diagnosis was a median of 158 days (interquartile range 84-288 days) for HL in an adult population (>18y). Of this time-to-diagnosis, the median diagnostic interval (help-seeking to diagnosis) was nearly three times greater than the median patient interval (symptom onset to help-seeking), namely 87 days and 30 days, respectively. These findings are consistent with this case, with a protracted diagnostic interval and total time-to-diagnosis of seven months, exacerbated by the COVID-19 pandemic, and similar to other reports. As a result of variability in time-to-diagnosis, over one-quarter of HL patients present with advanced stage (3 or 4) disease. The majority (80%) of patients with classic HL present with painless lymphadenopathy, usually involving the supraclavicular and cervical nodes. , Of note, our patient did not have palpable lymphadenopathy, adding to the diagnostic challenge. A similar proportion of adolescents and young adults will have anterior mediastinal involvement at presentation, often asymptomatic. , Extra-nodal involvement usually arises from hematogenous dissemination. SCC is a rare extra-nodal manifestation of HL, occurring in ∼5% of cases and usually in the context of progressive or advanced disease. , It is the main presenting feature in only ∼0.2% of cases. Epidural lesions in HL may arise from hematogenous dissemination but are more likely to develop as a result of local invasion from retroperitoneal or thoracic lymph nodes. The thoracic spine appears to be the most commonly implicated site, followed by the lumbar spine. Of note, back pain is not an uncommon complaint in the pediatric population, with a prevalence of ∼10% at age 10y and ∼30% at 13y of age. That notwithstanding, chronic back pain has a relatively high yield for serious pathology including mechanical, infectious, inflammatory and neoplastic etiologies. For this reason, it is important to keep a very broad differential when a child or adolescent presents with back pain. For example, back pain in a child causing waking from sleep should result in urgent referral to pediatrics and neurological deficits should also always be considered a ‘red flag’ and urgent imaging arranged. A study of patients presenting with a malignant cause for spinal cord compression found that neuroblastoma (29%), soft-tissue sarcomas (21%), neuroectodermal tumors (17%) and non-Hodgkin lymphoma (13%) made up the majority of cases. Other malignant causes included astrocytoma, Wilms tumor and leukemia. Another study showed that SCC was the presenting feature of a previously undiagnosed malignancy in 75% of cases. Motor deficit was the presenting symptom in all patients, while pain was reported in 60% and sphincter dysfunction in 43%. The UK NICE guidelines for the emergency management of malignant SCC recommend steroids as an essential component and that they should be offered to all patients while definitive treatment is planned. Steroids are, however, contraindicated in those patients with a significant suspicion of lymphoma. HL is highly steroid sensitive, with treatment protocols utilising them as a key backbone of therapy. , Even very short courses of steroids prior to biopsy may reduce the likelihood of reaching a definitive and accurate histologic diagnosis in such cases. Although most cases of classic HL can be diagnosed on morphologic and immunophenotypic features, there are a number of malignant lymphoid proliferations that can display histological features resembling HL. These include grey-zone lymphoma, EBV-positive diffuse large B-cell lymphoma, anaplastic large-cell lymphoma and peripheral T-cell lymphoma with ‘Hodgkin-like’ cells. Despite steroids being withheld in this case for these reasons until after the first biopsy, a diagnosis could not be reached from the spinal biopsy alone. In this clinically emergent setting of spinal cord compression, a rapid and accurate pathological diagnosis is however important. It should be noted that although the morphological appearances of the original epidural biopsy were concerning for HL, and a very small minority of HL cases can be PAX5 negative, the combination of negative staining for both CD15 and PAX-5 in this first biopsy in conjunction with the poorly preserved morphology, meant it was not felt possible to make a formal diagnosis on this tissue. This is likely to be related to poor preservation of tissue due to technical (diathermy/freezing artefact) reasons; immunohistochemistry is sensitive to tissue handling and preservation. However, use of diathermy was required to safely neurosurgically decompress the spinal cord and nerve roots, given the highly vascular nature of the tumor. In this case, as the patient had undergone emergency decompression of the theca and left L4 nerve root and had started steroids, with some clinical improvement, and given the putative HL diagnosis, it was therefore deemed clinically prudent to obtain more tissue to securely establish the exact diagnosis under the same general anaesthetic used for central line placement and staging bone marrow examination five days later. Clearly, such a decision for repeat biopsy or otherwise has to be assessed on a case-by case basis. In certain instances, where repeat biopsy is not feasible, or if repeat biopsy is non-diagnostic due to steroid delivery, a pragmatic and empirical diagnosis may have to be accepted based on the findings of the original biopsy. In summary, we present a 15-year-old with HL presenting with SCC, advanced stage disease and prolonged time-to-diagnosis. The diagnosis was challenging as the first neurosurgical biopsy was non-diagnostic, resulting in a second ultrasound-guided excision biopsy of a non-palpable cervical lymph node five days later, that allowed the diagnosis of nodular sclerosis classic HL to be made, despite the interim use of steroids following the first biopsy. The case highlights that pediatric and adolescent patients presenting with back pain need timely investigation and onward referral if any abnormal signs, symptoms or results, and the need to maintain a broad differential when assessing such patients. Such an approach will minimize diagnostic intervals and result in optimized patient outcomes.
|
Programmed death-ligand 1 expression in carcinoma of unknown primary | 8c5e1333-aa5f-406b-8bf5-848e3ab9e05a | 11155179 | Anatomy[mh] | Carcinoma of unknown primary origin (CUP) is a metastatic carcinoma in which the primary tumor remains elusive even after evaluation of the clinical history, physical examination, radiological findings, laboratory tests and other diagnostic investigations . CUP accounts for approximately 5–15% of malignant tumors , and advances in imaging and molecular testing have reduced this proportion to 1–2% in recent years . Histologically, CUP comprises adenocarcinomas (50–60%) or poorly differentiated carcinomas (30–40%), with other histological types, including squamous cell carcinomas (5–8%) and undifferentiated carcinomas (2–5%) . Although the precise nature of CUP remains uncertain, two main hypotheses have been suggested: the first postulates that CUP represents a true metastatic tumor with a primary focus that is markedly small to be identified; the second suggests that CUP is a distinct entity with independent characteristics due to regression or dormancy of the primary lesion, known as the ‘true’ or ‘true” genuine’ or ‘genuine’ CUP hypothesis . Treatment planning for metastatic carcinoma is generally determined by the type of primary cancer, making the absence of a known primary tumor in CUP a critical treatment challenge. The traditional diagnostic and treatment algorithm for CUP involves identifying favorable subgroups by undertaking a traditional diagnostic work-up and administering tissue origin-specific therapy while administering empirical chemotherapy or tissue origin-specific therapy based on the characteristics of each CUP in unfavorable subgroups . Techniques such as immunohistochemistry (IHC) and molecular tools such as gene expression profiling, miRNA expression, and DNA methylation analysis have been employed to determine the most appropriate tissue-of-origin for a specific CUP . Furthermore, precision medicine concepts based on advances in genomic tools are being applied to CUP to attempt targeted therapy by identifying possible treatment targets . Therefore, identifying an appropriate treatment target for CUP is crucial to ensure proper treatment. Programmed death 1 (PD-1) is an immune checkpoint molecule found on different immune cells, playing a crucial role in immune responses . Conversely, programmed death-ligand 1 (PD-L1) acts as a ligand for PD-1. Tumor cells express PD-L1, which facilitates their evasion of antitumor immune responses by interacting with PD-1 and forming a suppressive pathway . PD-L1 is expressed in 20–70% of tumors, including lung cancer , urinary bladder cancer , malignant melanoma , ovarian cancer , breast cancer , and gastric cancer . In patients with PD-L1-positive tumors, targeted therapy against PD-L1 can be used to induce an antitumor immune response. Notably, PD-L1 inhibitors have been approved as effective treatments for non-small cell lung cancer, urothelial carcinoma, gastric carcinoma, esophageal carcinoma, cervical cancer, and triple-negative breast cancer (TNBC) . In addition, various drugs such as pembrolizumab, atezolizumab, durvalumab, nivolumab, and ipilimumab have been developed as PD-L1 inhibitors . Therefore, it is important to determine whether PD-L1 is expressed in tumor cells prior to targeted therapy. The most common and simple method for detecting PD-L1 expression is IHC using a monoclonal PD-L1 antibody on formalin-fixed paraffin-embedded (FFPE) specimens. Monoclonal PD-L1 antibodies, such as clone 28 − 8 , 22C3 , SP142 , and SP263 are commercially available, and appropriate antibodies and scoring systems have been established as companion diagnostics for different types of cancer. Although several studies have investigated PD-L1 expression in various tumors using various antibodies, PD-L1 expression in CUP has been poorly explored. Therefore, the purpose of the present study was to examine PD-L1 expression in CUP according to different antibodies and scoring systems and to explore its implications. Patient selection and clinicopathologic evaluation In this study, we utilized FFPE tissue samples obtained from patients with Carcinoma of Unknown Primary (CUP) at Severance Hospital. The study adhered to the principles of the Declaration of Helsinki and obtained approval from the Institutional Review Board of Yonsei University Severance Hospital (IRB number: 4-2022-1380). Due to the retrospective nature of the study, patient consent was exempted by the Institutional Review Board of Yonsei University Severance Hospital. The selected patients were diagnosed with metastatic carcinoma by a pathologist between January 1999 and December 2012. In this study, needle biopsies yielding insufficient tissue for TMA construction were excluded, while excisional biopsies suitable for TMA construction were included. Cases that received chemotherapy or targeted therapy before tissue diagnosis were excluded. All available hematoxylin and eosin (H&E)-stained slides were carefully reviewed. Clinicopathological parameters, including patient age, sex, histological type, organ involvement, and patient outcomes, were assessed for each tumor. Based on histological criteria, CUPs were categorized into four distinct groups : adenocarcinomas (ADCs) displayed glandular differentiation, while squamous cell carcinomas (SCCs) exhibited evidence of squamous differentiation. Poorly differentiated carcinomas (PDCs) did not exhibit any specific lineage differentiation, and undifferentiated carcinomas (UDCs) consisted of syncytial tumor cell nests or individual tumor cells closely intertwined with dense lymphoplasmacytic infiltration, resembling the pattern seen in nasopharyngeal undifferentiated carcinomas. Additionally, CUPs were classified into favorable and unfavorable subgroups according to international guidelines . In accordance with international guidelines, the following nine scenarios are defined as the favorable subgroup. In this study, these same nine scenarios were also defined as the favorable subgroup; (1) poorly differentiated neuroendocrine CUP, (2) well-differentiated neuroendocrine tumor of unknown primary, (3) peritoneal adenocarcinomatosis of a serous papillary in females, (4) isolated axillary nodal metastases in females, (5) SCC involving non-supraclavicular cervical lymph nodes, (6) CUP with a colorectal IHC or molecular profile, (7) single metastatic deposit from unknown primary, (8) males with blastic bone metastases or IHC/serum prostate-specific antigen expression, and (9) SCC involving isolated inguinal adenopathy. CUP cases outside the defined favorable subgroup were categorized as the unfavorable subgroup. Tissue microarray Following the assessment of H&E-stained slides, suitable FFPE tumor tissue samples were retrospectively gathered, focusing on the most representative tumor region, which was carefully demarcated. A punch machine was utilized to extract the chosen area, and a 3 mm tissue core was inserted into a 6 × 5 recipient block. For each sample, tissue microarrays were created, with two tissue cores included in each array. IHC Immunohistochemistry (IHC) was conducted on FFPE tissue sections, and the antibodies employed are specified in Supplementary Table . Briefly, 3-µm thick tissue sections were prepared from paraffin blocks and then deparaffinized and rehydrated using xylene and alcohol solution. The IHC procedure was carried out using a Ventana Discovery XT automated stainer (Ventana Medical System, Tucson, AZ, USA). Antigen retrieval was achieved using CC1 buffer (Cell Conditioning 1; citrate buffer, pH 6.0; Ventana Medical System). Immunohistochemical staining was performed, incorporating appropriate positive and negative controls. For the negative control group, the primary antibody incubation step was omitted. Each antibody’s recommended positive control, as specified by the manufacturer, was utilized. Interpretation of immunohistochemical results Immunohistochemical staining of PD-L1 was performed according to the antibody used. PD-L1 22C3 expression was evaluated using tumor cells (TC) (tumor proportion score [TPS]), immune cell score (IC), and combined positive score (CPS). TPS was calculated by dividing the number of PD-L1 staining tumor cells by the number of viable tumor cells and multiplying by 100%. The CPS was calculated by dividing the number of PD-L1 staining cells (including tumor cells, lymphocytes, and histiocytes) by the number of viable tumor cells and multiplying by 100%. PD-L1 28 − 8, SP142, and SP263 were evaluated for TC and IC. TC was defined as the percentage of tumor cells showing any intensity of membranous staining for PD-L1, while IC was defined as the percentage of the tumor area occupied by PD-L1 staining immune cells (including lymphocytes, histiocytes, dendritic cells, and granulocytes). In this study, PD-L1 interpretation was conducted by two pathologists (HM Kim and JS Koo) who participated in the study, using a multi-view microscope. They determined TC, IC, and CPS of PD-L1 for each case while reviewing the TMA slides. For cases near the cut-off value, the two pathologists reached a final decision through consensus. The pathologist (JS Koo) who interpreted the PD-L1 IHC in this study is a board-certified pathologist with over 20 years of experience in the field. Their expertise lies particularly in breast cancer, where they have been routinely interpreting PD-L1 (SP142 and 22C3) for several years in daily practice. Additionally, they have published research papers on PD-L1 . Two different methods were used to analyze the TPS, IC, and CPS. First, the cutoff values established for each PD-L1 clone in other tumor types were used. For PD-L1 22C3, TPS of ≥ 1 and CPS of ≥ 10 were considered positive . For PD-L1 SP142, TC of ≥ 50 and IC of ≥ 10 were considered positive . For PD-L1 28 − 8 and SP263, TC and IC of ≥ 1 were considered positive . Second, to compare the results for each antibody, the criteria for positivity were set as TC(TPS) ≥ 1%, TC(TPS) ≥ 50%, IC ≥ 1%, and IC ≥ 10%. For CK7 and CK20, the cutoff value was set at 10%; cases with < 10% staining were considered negative, whereas those with ≥ 10% staining were considered positive . Statistical analysis Data analysis was performed using SPSS for Windows (version 24.0; IBM Corp., Armonk, NY, USA). Continuous variables were analyzed using Student’s t-test, while categorical variables were assessed using Fisher’s exact tests. The threshold for statistical significance was set at p < 0.05. To evaluate the agreement between any two PD-L1 antibody clones for each scoring method, Cohen’s kappa coefficient was utilized. The interpretation of the kappa coefficient values was as follows: <0 indicated no agreement, 0.0–0.20 represented slight agreement, 0.21–0.40 indicated fair agreement, 0.41–0.60 signified moderate agreement, 0.61–0.80 suggested substantial agreement, and 0.81–1.00 denoted almost perfect agreement . Kaplan-Meier survival curves and log-rank statistics were employed to assess the survival time. Additionally, multivariate regression analysis was conducted using a Cox proportional hazards model. In this study, we utilized FFPE tissue samples obtained from patients with Carcinoma of Unknown Primary (CUP) at Severance Hospital. The study adhered to the principles of the Declaration of Helsinki and obtained approval from the Institutional Review Board of Yonsei University Severance Hospital (IRB number: 4-2022-1380). Due to the retrospective nature of the study, patient consent was exempted by the Institutional Review Board of Yonsei University Severance Hospital. The selected patients were diagnosed with metastatic carcinoma by a pathologist between January 1999 and December 2012. In this study, needle biopsies yielding insufficient tissue for TMA construction were excluded, while excisional biopsies suitable for TMA construction were included. Cases that received chemotherapy or targeted therapy before tissue diagnosis were excluded. All available hematoxylin and eosin (H&E)-stained slides were carefully reviewed. Clinicopathological parameters, including patient age, sex, histological type, organ involvement, and patient outcomes, were assessed for each tumor. Based on histological criteria, CUPs were categorized into four distinct groups : adenocarcinomas (ADCs) displayed glandular differentiation, while squamous cell carcinomas (SCCs) exhibited evidence of squamous differentiation. Poorly differentiated carcinomas (PDCs) did not exhibit any specific lineage differentiation, and undifferentiated carcinomas (UDCs) consisted of syncytial tumor cell nests or individual tumor cells closely intertwined with dense lymphoplasmacytic infiltration, resembling the pattern seen in nasopharyngeal undifferentiated carcinomas. Additionally, CUPs were classified into favorable and unfavorable subgroups according to international guidelines . In accordance with international guidelines, the following nine scenarios are defined as the favorable subgroup. In this study, these same nine scenarios were also defined as the favorable subgroup; (1) poorly differentiated neuroendocrine CUP, (2) well-differentiated neuroendocrine tumor of unknown primary, (3) peritoneal adenocarcinomatosis of a serous papillary in females, (4) isolated axillary nodal metastases in females, (5) SCC involving non-supraclavicular cervical lymph nodes, (6) CUP with a colorectal IHC or molecular profile, (7) single metastatic deposit from unknown primary, (8) males with blastic bone metastases or IHC/serum prostate-specific antigen expression, and (9) SCC involving isolated inguinal adenopathy. CUP cases outside the defined favorable subgroup were categorized as the unfavorable subgroup. Following the assessment of H&E-stained slides, suitable FFPE tumor tissue samples were retrospectively gathered, focusing on the most representative tumor region, which was carefully demarcated. A punch machine was utilized to extract the chosen area, and a 3 mm tissue core was inserted into a 6 × 5 recipient block. For each sample, tissue microarrays were created, with two tissue cores included in each array. Immunohistochemistry (IHC) was conducted on FFPE tissue sections, and the antibodies employed are specified in Supplementary Table . Briefly, 3-µm thick tissue sections were prepared from paraffin blocks and then deparaffinized and rehydrated using xylene and alcohol solution. The IHC procedure was carried out using a Ventana Discovery XT automated stainer (Ventana Medical System, Tucson, AZ, USA). Antigen retrieval was achieved using CC1 buffer (Cell Conditioning 1; citrate buffer, pH 6.0; Ventana Medical System). Immunohistochemical staining was performed, incorporating appropriate positive and negative controls. For the negative control group, the primary antibody incubation step was omitted. Each antibody’s recommended positive control, as specified by the manufacturer, was utilized. Immunohistochemical staining of PD-L1 was performed according to the antibody used. PD-L1 22C3 expression was evaluated using tumor cells (TC) (tumor proportion score [TPS]), immune cell score (IC), and combined positive score (CPS). TPS was calculated by dividing the number of PD-L1 staining tumor cells by the number of viable tumor cells and multiplying by 100%. The CPS was calculated by dividing the number of PD-L1 staining cells (including tumor cells, lymphocytes, and histiocytes) by the number of viable tumor cells and multiplying by 100%. PD-L1 28 − 8, SP142, and SP263 were evaluated for TC and IC. TC was defined as the percentage of tumor cells showing any intensity of membranous staining for PD-L1, while IC was defined as the percentage of the tumor area occupied by PD-L1 staining immune cells (including lymphocytes, histiocytes, dendritic cells, and granulocytes). In this study, PD-L1 interpretation was conducted by two pathologists (HM Kim and JS Koo) who participated in the study, using a multi-view microscope. They determined TC, IC, and CPS of PD-L1 for each case while reviewing the TMA slides. For cases near the cut-off value, the two pathologists reached a final decision through consensus. The pathologist (JS Koo) who interpreted the PD-L1 IHC in this study is a board-certified pathologist with over 20 years of experience in the field. Their expertise lies particularly in breast cancer, where they have been routinely interpreting PD-L1 (SP142 and 22C3) for several years in daily practice. Additionally, they have published research papers on PD-L1 . Two different methods were used to analyze the TPS, IC, and CPS. First, the cutoff values established for each PD-L1 clone in other tumor types were used. For PD-L1 22C3, TPS of ≥ 1 and CPS of ≥ 10 were considered positive . For PD-L1 SP142, TC of ≥ 50 and IC of ≥ 10 were considered positive . For PD-L1 28 − 8 and SP263, TC and IC of ≥ 1 were considered positive . Second, to compare the results for each antibody, the criteria for positivity were set as TC(TPS) ≥ 1%, TC(TPS) ≥ 50%, IC ≥ 1%, and IC ≥ 10%. For CK7 and CK20, the cutoff value was set at 10%; cases with < 10% staining were considered negative, whereas those with ≥ 10% staining were considered positive . Data analysis was performed using SPSS for Windows (version 24.0; IBM Corp., Armonk, NY, USA). Continuous variables were analyzed using Student’s t-test, while categorical variables were assessed using Fisher’s exact tests. The threshold for statistical significance was set at p < 0.05. To evaluate the agreement between any two PD-L1 antibody clones for each scoring method, Cohen’s kappa coefficient was utilized. The interpretation of the kappa coefficient values was as follows: <0 indicated no agreement, 0.0–0.20 represented slight agreement, 0.21–0.40 indicated fair agreement, 0.41–0.60 signified moderate agreement, 0.61–0.80 suggested substantial agreement, and 0.81–1.00 denoted almost perfect agreement . Kaplan-Meier survival curves and log-rank statistics were employed to assess the survival time. Additionally, multivariate regression analysis was conducted using a Cox proportional hazards model. Basal characteristics of patients with CUP according to the histologic and clinical subtypes Supplementary Tables and show the basal characteristics according to histological and clinical subtypes in the 72 CUP cases. Overall, 22 (30.6%) patients had ADC, 15 (20.8%) had PDC, 19 (26.4%) had SCC, and 16 (22.2%) had UDC. The clinical subtype was favorable in 17 (23.6%) and unfavorable in 55 (76.4%) cases. The involved organs were as follows: lymph nodes 49 (68.1%), bone 8 (11.1%), brain 7 (9.7%), and other 8 (11.1%). Moreover, there was a difference in clinical subtype according to the histologic subtype, with ADC and UDC showing a higher proportion of the unfavorable type, while SCC showed a higher proportion of the favorable type ( p = 0.003). Additionally, postoperative treatment differed according to the histologic subtype, with chemotherapy most commonly employed in ADC, chemoradiotherapy in PDC, and surgery only in UDC ( p = 0.007). Among the CUP cases, 37 (51.4%) were CK7 /CK20 , 3 (4.2%) were CK7 /CK20 , 3 (4.2%) were CK7 /CK20 , and 29 (40.3%) were CK7 /CK20 , with no significant difference in histologic subtype ( p = 0.522). PD-L1 expression in CUP In CUP, tumor and immune cells exhibited PD-L1 expression at varying proportions and intensities (Fig. ). PD-L1 expression was examined in CUP using cutoff values as follows: TPS ≥ 1%, CPS ≥ 10 for SP142; TC ≥ 50%, IC ≥ 10% for 22C3; TC and IC ≥ 1% for 28 − 8 and SP263. PD-L1 positivity rates ranged between 5.6 and 48.6%, with the lowest rate of 5.6% observed in PD-L1 SP142 TC and the highest rate of 48.6% in PD-L1 SP263 IC. PD-L1 positivity rates did not show significant differences according to histologic subtype (Table ), clinical subtype (Table ), or CK7/CK20 pattern (Table ) across clones. Difference and concordance of PD-L1 expression in CUP according to PD-L1 antibody clones and Scoring systems We then analyzed differences in PD-L1 expression among the four clones and scoring systems in CUP. For the TC system, PD-L1 positivity ranged between 18.1 and 36.1% for a cutoff value of 1% and between 4.2 and 20.8% for a cutoff value of 50%. Among the examined clones, 22C3 and SP263 showed the lowest and highest positivity rates, respectively. For the IC system, PD-L1 positivity ranged between 26.4 and 48.6% for a cutoff value of 1% and between 9.7 and 38.9% for a cutoff value of 10%. Among the clones, 28 − 8 and SP263 exhibited the lowest and highest positivity rates, respectively (Table ). Next, we examined the concordance of PD-L1 expression among clones according to the scoring system (Table ). For TC ≥ 1%, all clones showed moderate or high agreement, with the highest agreement between 22C3 and SP142 (OA = 93.1%, ≥ =0.772) and the lowest agreement between 22C3 and SP263 (OA = 83.3%, ≥ =0.599). For TC κ 50%, all clones showed fair or higher agreement, with the highest agreement between 28 − 8 and SP263 (OA = 90.3%, κ = 0.664) and the lowest agreement between 22C3 and SP263 (OA = 83.3%, κ = 0.284). For IC ≥ 1%, all clones showed moderate or higher agreement, with the highest agreement between SP263 and SP142 (OA = 90.3%, κ = 0.805) and the lowest agreement between 28 − 8 and SP263 (OA = 77.8%, κ = 0.550). For IC κ 10%, all clones showed fair or high agreement, with the highest agreement between 22C3 and SP142 (OA = 91.7%, κ = 0.578) and the lowest agreement between SP142 and SP263 (OA = 69.4%, κ = 0.261). Impact of clinicopathologic factors and PD-L1 status on prognosis of CUP We subsequently performed univariate analysis to determine the impact of clinicopathological factors and PD-L1 expression on prognosis. We observed that the histological subtype was associated with shorter overall survival (OS) (UDC > SCC > ADC > PDC, p = 0.030), whereas PD-L1 expression was not significantly associated with shorter OS (Table ). In subgroup analysis, PD-L1 SP263 TC positivity ( p = 0.030) and PD-L1 SP263 IC negativity ( p = 0.007) were significantly associated with shorter OS for CUP with ADC histologic subtypes. For CK7 positive CUP, PD-L1 SP263 IC negativity ( p = 0.041) and PD-L1 28 − 8 IC negativity ( p = 0.029) were significantly associated with a shorter OS. For CK7 and CK20 positive CUP and unfavorable clinical type CUP, PD-L1 28 − 8 IC negativity ( p = 0.037 and p = 0.040, respectively) was significantly associated with shorter OS (Fig. ). Supplementary Tables and show the basal characteristics according to histological and clinical subtypes in the 72 CUP cases. Overall, 22 (30.6%) patients had ADC, 15 (20.8%) had PDC, 19 (26.4%) had SCC, and 16 (22.2%) had UDC. The clinical subtype was favorable in 17 (23.6%) and unfavorable in 55 (76.4%) cases. The involved organs were as follows: lymph nodes 49 (68.1%), bone 8 (11.1%), brain 7 (9.7%), and other 8 (11.1%). Moreover, there was a difference in clinical subtype according to the histologic subtype, with ADC and UDC showing a higher proportion of the unfavorable type, while SCC showed a higher proportion of the favorable type ( p = 0.003). Additionally, postoperative treatment differed according to the histologic subtype, with chemotherapy most commonly employed in ADC, chemoradiotherapy in PDC, and surgery only in UDC ( p = 0.007). Among the CUP cases, 37 (51.4%) were CK7 /CK20 , 3 (4.2%) were CK7 /CK20 , 3 (4.2%) were CK7 /CK20 , and 29 (40.3%) were CK7 /CK20 , with no significant difference in histologic subtype ( p = 0.522). In CUP, tumor and immune cells exhibited PD-L1 expression at varying proportions and intensities (Fig. ). PD-L1 expression was examined in CUP using cutoff values as follows: TPS ≥ 1%, CPS ≥ 10 for SP142; TC ≥ 50%, IC ≥ 10% for 22C3; TC and IC ≥ 1% for 28 − 8 and SP263. PD-L1 positivity rates ranged between 5.6 and 48.6%, with the lowest rate of 5.6% observed in PD-L1 SP142 TC and the highest rate of 48.6% in PD-L1 SP263 IC. PD-L1 positivity rates did not show significant differences according to histologic subtype (Table ), clinical subtype (Table ), or CK7/CK20 pattern (Table ) across clones. We then analyzed differences in PD-L1 expression among the four clones and scoring systems in CUP. For the TC system, PD-L1 positivity ranged between 18.1 and 36.1% for a cutoff value of 1% and between 4.2 and 20.8% for a cutoff value of 50%. Among the examined clones, 22C3 and SP263 showed the lowest and highest positivity rates, respectively. For the IC system, PD-L1 positivity ranged between 26.4 and 48.6% for a cutoff value of 1% and between 9.7 and 38.9% for a cutoff value of 10%. Among the clones, 28 − 8 and SP263 exhibited the lowest and highest positivity rates, respectively (Table ). Next, we examined the concordance of PD-L1 expression among clones according to the scoring system (Table ). For TC ≥ 1%, all clones showed moderate or high agreement, with the highest agreement between 22C3 and SP142 (OA = 93.1%, ≥ =0.772) and the lowest agreement between 22C3 and SP263 (OA = 83.3%, ≥ =0.599). For TC κ 50%, all clones showed fair or higher agreement, with the highest agreement between 28 − 8 and SP263 (OA = 90.3%, κ = 0.664) and the lowest agreement between 22C3 and SP263 (OA = 83.3%, κ = 0.284). For IC ≥ 1%, all clones showed moderate or higher agreement, with the highest agreement between SP263 and SP142 (OA = 90.3%, κ = 0.805) and the lowest agreement between 28 − 8 and SP263 (OA = 77.8%, κ = 0.550). For IC κ 10%, all clones showed fair or high agreement, with the highest agreement between 22C3 and SP142 (OA = 91.7%, κ = 0.578) and the lowest agreement between SP142 and SP263 (OA = 69.4%, κ = 0.261). We subsequently performed univariate analysis to determine the impact of clinicopathological factors and PD-L1 expression on prognosis. We observed that the histological subtype was associated with shorter overall survival (OS) (UDC > SCC > ADC > PDC, p = 0.030), whereas PD-L1 expression was not significantly associated with shorter OS (Table ). In subgroup analysis, PD-L1 SP263 TC positivity ( p = 0.030) and PD-L1 SP263 IC negativity ( p = 0.007) were significantly associated with shorter OS for CUP with ADC histologic subtypes. For CK7 positive CUP, PD-L1 SP263 IC negativity ( p = 0.041) and PD-L1 28 − 8 IC negativity ( p = 0.029) were significantly associated with a shorter OS. For CK7 and CK20 positive CUP and unfavorable clinical type CUP, PD-L1 28 − 8 IC negativity ( p = 0.037 and p = 0.040, respectively) was significantly associated with shorter OS (Fig. ). In the present study, we determined the expression of PD-L1 in various CUP clones, detecting various positive rates depending on the antibodies used, the applied scoring system, and cutoff values. Although PD-L1 expression in CUP remains poorly established, a positivity rate of 22% in tumor cells has been reported using the antibody SP142, with the positivity criteria defined as moderate (2+) membranous positivity in at least 5% of tumor cells, indicating that the previous criteria were TC ≥ 5%. In addition, PD-L1 was found to be expressed in tumor cells in 14% of the CUP cases ; the antibody used was 22C3, and the positivity criteria were defined as at least 50% of tumor cells being positive, indicating that the previous criteria were TPS ≥ 50%. In the current study, the positivity rates were 18.1% (for TC ≥ 1%) and 5.6% (for TC ≥ 50%) using SP142, and 19.4% (for TC ≥ 1%) and 4.2% (for TC ≥ 50%) with 22C3. The PD-L1 positivity rate varied depending on the PD-L1 antibody clone, scoring system, and cutoff values, as well as based on the interpretation by the pathologist and sample tissue type. Therefore, a direct comparison can be challenging. Although some studies have examined the expression of PD-L1 using only one PD-L1 antibody, no study has explored PD-L1 expression using multiple PD-L1 antibodies with various scoring systems or cutoff values. As previously mentioned, various factors can impact the results of PD-L1 IHC in tumors, including the PD-L1 antibody clone, scoring system, cutoff value, interpretation pathologist, sample tissue type (biopsy or resection), and primary or metastasis. Accordingly, several studies have investigated the expression and consistency of PD-L1 according to these factors in various types of cancers. PD-L1 expression has been extensively explored in cancers such as non-small cell lung carcinoma (NSCLC), TNBC, melanoma, renal cell carcinoma, bladder cancer, and gastric cancer. The positivity rates of PD-L1 in each cancer type were as follows: NSCLC (TC = 23–86%, IC = 23–68%) , breast TNBC (IC = 23–74%, CPS = 17–81%) , renal cell carcinoma (TPS = 25–60%) , bladder cancer (TC = 12–72%) , and gastric cancer (TC = 15–69%) . In tumors, the main function of PD-L1 is to predict the response to immune checkpoint inhibitors (ICI), and various clinical trials are underway to optimize its function as a predictive factor, depending on the type of tumor. Accordingly, a companion diagnosis has been established in clinical practice for each cancer type, determining the optimal PD-L1 antibody clone, IHC platform, scoring system, cutoff value, and specific ICI. Representative tumors include NSCLC, TNBC, urothelial carcinoma, uterine cervical cancer, and gastric/esophageal cancer. Therefore, additional preclinical and clinical studies are required to determine the optimal PD-L1 conditions for CUP. Although the possibility of an ICI therapy response according to PD-L1 expression status in CUP warrants clinical trials and extensive research, a potential response to ICI therapy according to the PD-L1 expression status can be sufficiently suggested. Currently, the treatment approach in CUP involves site-specific therapy if the tissue-of-origin is determined using an IHC panel and/or molecular tissue-of-origin assay . Given that the efficacy of ICI therapy based on PD-L1 has been confirmed in NSCLC, TNBC, urothelial carcinoma, uterine cervical cancer, and gastric/esophageal cancer, if the tissue origin is determined for CUP using an IHC panel and/or molecular tissue-of-origin assay, ICI therapy could be initiated on assessing PD-L1 expression. However, it is necessary to consider that the currently defined PD-L1 clones, IHC platforms, scoring systems, and cutoff values for each cancer type tend to differ; therefore, additional research is needed to determine whether different PD-L1 evaluation systems should be used according to the tissue origin in CUP. Based on the subgroup analysis of CUP, PD-L1 SP263 TC positivity, PD-L1 SP263 IC negativity, and PD-L1 28 − 8 IC negativity were associated with a poor prognosis. Other tumors, including urothelial carcinoma, NSCLC, head and neck cancer, and liver cholangiocarcinoma, have shown similar results, where PD-L1 expression in tumor cells was associated with poor prognosis, whereas PD-L1 expression in immune cells was associated with better prognosis . In this study, only PD-L1 staining was conducted. However, previous studies in other cancer types have performed double staining such as CD68/PD-L1 to distinguish staining differences between PD-L1 and tumor-associated macrophages (TAMs) and other immune cells, and have presented differences in tumor subtypes and prognosis accordingly . Therefore, dual staining like CD68/PD-L1 can provide important insights into the role of immune cells in the tumor microenvironment and the mechanisms of tumor immune evasion. This could aid in developing treatment strategies and identifying the origin of tumors. Therefore, additional research on dual staining, such as CD68/PD-L1, is deemed necessary to accurately characterize the tumor properties and develop personalized treatment strategies, especially in cases like CUP where the tumor origin is unknown. One limitation of this study is that PD-L1 staining was conducted on a limited amount of tissue using TMA, which may not adequately reflect tumor heterogeneity. Previous studies investigating the differences in PD-L1 expression between biopsy and surgical tissue in various cancer types have shown a concordance rate of 70% or higher in most cases . Additionally, in clinical practice, obtaining small biopsies rather than excising the entire lesion surgically is more common in cases of CUP, suggesting that the results from TMA studies may be more similar to the actual clinical environment. Moreover, in cases where small biopsies are not feasible due to various clinical circumstances, cytological samples may be considered for assessing PD-L1 status in CUP patients. Previous studies have reported moderate or higher concordance rates between cytology and histology samples regarding PD-L1 expression , indicating the need for additional research on PD-L1 expression in cytological samples from CUP patients. In conclusion, PD-L1 expression was observed in CUP, with varying positivity rates depending on the antibody and scoring system employed. There was no difference in PD-L1 expression based on histological or clinical subtypes. Therefore, ICI treatment based on PD-L1 expression in CUP can be an effective treatment strategy. Below is the link to the electronic supplementary material. Supplementary Material 1 |
Cable-asisted bone transport versus circular external fixators-asisted bone transport in the management of bone defects of the Tibia: clinical and imaging results | 5611c6dc-e0a6-4165-9672-7fa8c4e32865 | 11899003 | Musculoskeletal System[mh] | Tibial bone defects are one of the most challenging and complex problems encountered in orthopaedic practice, occurring for various reasons such as high-energy trauma, nonunion, osteomyelitis or tumor resection. Such defects are often accompanied by many comorbid conditions such as soft tissue loss, deformities, limb shortness, infection and joint motion limitations [ – ]. All these factors both prolong the treatment process and negatively affect the final functional outcome . Successful management of bone defects involves not only the restoration of bone integrity but also a complex approach aimed at long-term infection control, correction of deformity and restoration of the patient’s functional independence. Various surgical methods have been developed for the management of bone defects of the tibia, including vascularized free fibula grafts, the Masquelet-induced membrane technique, acute shortening, and segmental bone transport [ – ]. Each of these methods has advantages as well as disadvantages and limitations [ – ]. Although vascularized fibula graft provides biological reconstruction in the treatment of large defects with good union rates, loading before graft hypertrophy occurs carries a fracture risk . The Masquelet technique provides periarticular space advantage, but the increasing need for autograft as the defect size increases will cause donor site problems. Furthmore, it requires a two-stage procedure and long recovery times . Acute shortening and bone transport techniques allow biological reconstruction, and although defects ≤ 3 cm can often be treated with these methods, defects > 4 cm usually require distraction osteogenesis with segmental bone transport methods [ , – ]. Despite advances in treatment modalities, achieving full functional recovery remains a complex and time-consuming process. Segmental bone transport is one of the most effective methods in the management of large bone defects , which can be achieved through the utilisation of a variety of techniques. Circular external fixator-assisted methods and cable-assisted bone methods are among the methods most used. While circular external fixator-asisted bone transport (CEFt) offers advantages in terms of deformity correction and stability, complications such as the numerous transosseous wires penetrating the skin and pin-tract infection have been reported [ – ]. Cable-asisted bone transport (CASt) offers a less invasive alternative, reducing the risk of these complications . However, in cases where the cable system does not provide adequate compression at the docking site, additional surgical interventions may be needed . The aim of this study was to compare CASt and CEFt methods used in the management of bone defects of the tibia and to evaluate the efficacy, complications, and clinical and radiological results of these two methods. The study hypothesis was that the CASt method would have similar effectiveness but with some advantages.
Study design and ethical approval This retrospective comparative study was conducted in accordance with the Declaration of Helsinki with Institutional Ethics Committee approval (Approval Number:3031, Date: 24/11/2020). Informed consent was obtained from all patients for the use of their medical data for research purposes.
This retrospective comparative study was conducted in accordance with the Declaration of Helsinki with Institutional Ethics Committee approval (Approval Number:3031, Date: 24/11/2020). Informed consent was obtained from all patients for the use of their medical data for research purposes.
Patients who underwent segmental bone transport surgery for tibial bone defects between January 2006 and January 2020 were retrospectively analyzed through the institutional hospital patient data registry system. The study included patients who met the following inclusion criteria: Patients with tibial bone defects, Treatment with segmental bone transport method, Completed treatment and adequate follow-up for at least 12 months. The exclusion criteria were defined as: Bone defects < 4 cm, Patients for whom the final treatment was not a segmental bone transport method. A total of 37 patients were identified as having tibial bone defects and undergoing segmental bone transport treatment. A total of 5 patients were excluded from the study; 4 due to insufficient follow-up data or treatment completion issues, and 1 with a bone defect < 4 cm. Consequently, 32 patients met the eligibility criteria and were included in the study.
The selection of cable-assisted bone transport (CASt) or circular external fixator-assisted bone transport (CEFt) was determined by the individual clinical preferences of two experienced surgeons. In their respective practices, these surgeons routinely prefer different segmental bone transport techniques. Patients presenting at their respective clinics were treated with the method preferred by the attending surgeon. A formal sample size calculation was not performed due to the retrospective nature of the study and the limited number of eligible patients available during the study period. However, the study included all eligible patients who met the defined inclusion criteria.
) The defective area was opened, bone debridement was performed and the bone ends were levelled (Fig. ). The bone length and axis were corrected and the LRS type external fixator system was applied to bridge the defect area. The cable was folded in half and advanced intramedullarly into the segment to be transported antegrade or retrograde. When it reached a sufficient level, one steel cannulated screw was applied through the centre of the cable loop and the segment to be transported was connected to the cable (Fig. ). The cable system was advanced through the medulla in the proximal or distal fixed medulla, and was passed around one Schanz screw placed in the anteroposterior direction and removed from an area with little soft tissue such as the medial malleolus in the distal or the tibial condyle in the proximal. This Schanz screw acts as a pulley in the area where the cable will exit the bone and provides stability to the system in the second stage. Thus, the intramedullary cable also acts as a guide for the transported segment and prevents axis deviations. It was applied to the clamps of the LRS type external fixator specially prepared for the cable system, and the system was locked after ensuring the appropriate tension (Fig. ). Osteotomy was performed using a drill at the appropriate level in the segment to be transported and the transport process was controlled with the help of a distractor on the rail. After the distraction process was completed, the steel cannulated screw and cable were removed. The Schanz screw was applied to the transported segment and compression of the docking site was provided (Fig. ).
The defective area was opened, bone debridement was performed and the bone ends were levelled. Proximal and distal bone fragments were fixed to the circular rings with wire and Schanz screws. The bone was osteotomized using a drill. The conventional segmental bone transport protocol was performed. Post-operative management and follow-up protocol Distraction was usually initiated between postoperative days 5 and 7. Distraction was performed at a rate of 1 mm per day in 4 × 0.25 mm increments, in accordance with distraction osteogenesis protocols. Patients were given detailed training on how to perform the distraction procedure and pin-tract care. The fixators were removed once regenerate maturation was achieved and the docking site was fused (Figs. and ).
Distraction was usually initiated between postoperative days 5 and 7. Distraction was performed at a rate of 1 mm per day in 4 × 0.25 mm increments, in accordance with distraction osteogenesis protocols. Patients were given detailed training on how to perform the distraction procedure and pin-tract care. The fixators were removed once regenerate maturation was achieved and the docking site was fused (Figs. and ).
Demographic and preoperative data The demographic characteristics of the patients, including age, gender, etiology of bone loss, amount of bone loss, segmental bone transport method, and bone transport distance, were recorded. Open fractures were evaluated using the Gustilo-Anderson open fracture classification, and bone defects were classified preoperatively using Paley’s tibial pseudoarthrosis classification . Primary outcome measures Radiological parameters including the External Fixator Index (EFI), Radiological Consolidation Time (RCT), and Radiological Consolidation Index (RCI). Functional outcomes were assessed using the ASAMI Bone and Functional Scores. Patient-reported functional independence, assessed using the Lower Extremity Functional Index (LEFI). Secondary outcome measures Pain levels during distraction, evaluated using a Visual Analog Scale (VAS). Complications were assessed according to Paley’s classification and Checketts-Otterburn classification for pin-tract infections. All measurements were conducted using standardized radiographic and clinical assessment protocols. Radiological parameters were evaluated by two independent orthopedic surgeons blinded to the patient treatment groups, and clinical outcomes were assessed using validated scoring systems . Statistical analysis The statistical analysis of this study was performed using NCSS (Number Cruncher Statistical System) (Kaysville, Utah, USA). Continuous variables were tested for normality using the Shapiro-Wilk test. Normally distributed data were compared with the Student’s t-test, and non-normally distributed data with the Mann-Whitney U test. Categorical variables were compared using Pearson’s chi-square test or Fisher’s exact test where appropriate. The level of statistical significance was set at p < 0.05.
The demographic characteristics of the patients, including age, gender, etiology of bone loss, amount of bone loss, segmental bone transport method, and bone transport distance, were recorded. Open fractures were evaluated using the Gustilo-Anderson open fracture classification, and bone defects were classified preoperatively using Paley’s tibial pseudoarthrosis classification .
Radiological parameters including the External Fixator Index (EFI), Radiological Consolidation Time (RCT), and Radiological Consolidation Index (RCI). Functional outcomes were assessed using the ASAMI Bone and Functional Scores. Patient-reported functional independence, assessed using the Lower Extremity Functional Index (LEFI).
Pain levels during distraction, evaluated using a Visual Analog Scale (VAS). Complications were assessed according to Paley’s classification and Checketts-Otterburn classification for pin-tract infections. All measurements were conducted using standardized radiographic and clinical assessment protocols. Radiological parameters were evaluated by two independent orthopedic surgeons blinded to the patient treatment groups, and clinical outcomes were assessed using validated scoring systems .
The statistical analysis of this study was performed using NCSS (Number Cruncher Statistical System) (Kaysville, Utah, USA). Continuous variables were tested for normality using the Shapiro-Wilk test. Normally distributed data were compared with the Student’s t-test, and non-normally distributed data with the Mann-Whitney U test. Categorical variables were compared using Pearson’s chi-square test or Fisher’s exact test where appropriate. The level of statistical significance was set at p < 0.05.
A total of 32 patients were included in the study, of which 16 of were treated with the CASt method (Group 1) and 16 with the CEFt method (Group 2). Demographic and preoperative data All details regarding demographic and preoperative data (Etiology, Gustilo Anderson Classification, Paley Classification, Follow-up) are summarized in Table and no statistical difference was found between the groups ( p < 0.05). Radiological and clinical outcomes The data of the radiological and clinical outcomes are summarized in Table . There were no significant differences between the groups in terms of bone defect size, residual shortness, external fixator duration, external fixator index, time to radiological consolidation, and radiological consolidation index ( p > 0.05) (Fig. ). The pain level during the distraction period, evaluated with VAS, was significantly lower in Group 1 ( p = 0.001) (Fig. ). There was no significant difference in pain level during daily activities ( p = 0.560). Bone and functional outcomes The ASAMI bone and functional results and Lower Extremity Functional Index evaluations are summarized in Table . The ASAMI bone and functional results showed no significant difference between the groups ( p > 0.05). The LEFI results were similar with no statistical difference between the groups ( p > 0.05). Complications The evaluation of complications according to Paley classification by groups is shown in Table and the incidence of pin-tract infections is shown in Table . According to the Paley classification, 23 complications were reported in Group 1 and 30 in Group 2. Pin-tract infection was detected in 50% in Group 1 and 93.8% in Group 2 ( p = 0.013; Table ). According to the Checketts-Otterburn Classification, a total of 8 patients in Group 1 developed pin-tract infections. Among these infections, 5 patients were classified as Grade 1 and 3 patients as Grade 2. In Group 2, a total of 15 patients developed pin-tract infections; 7 were classified as Grade 1, 6 as Grade 2, and 2 as Grade 3 (Fig. ). Management of complications All complications were managed based on their severity, as follows: Pin-tract infections: Treated with pin-tract care, local antiseptic dressings, and oral antibiotics when needed. Loss of joint motion: Physical therapy and rehabilitation programs were used to help patients regain their ability to move. Delayed regenerate maturation, delayed consolidation, and delayed healing at the docking site: Managed with autologous bone marrow-derived mesenchymal stem cell (MSC) injections and bone grafting when necessary. Compression was additionally performed as needed for delayed healing at the docking site. >2.5 cm limb shortening: Patients were followed up and required limb elevating insoles or shoe modification. Re-fracture: Managed using circular external fixators to provide stability and allow for proper healing.
All details regarding demographic and preoperative data (Etiology, Gustilo Anderson Classification, Paley Classification, Follow-up) are summarized in Table and no statistical difference was found between the groups ( p < 0.05).
The data of the radiological and clinical outcomes are summarized in Table . There were no significant differences between the groups in terms of bone defect size, residual shortness, external fixator duration, external fixator index, time to radiological consolidation, and radiological consolidation index ( p > 0.05) (Fig. ). The pain level during the distraction period, evaluated with VAS, was significantly lower in Group 1 ( p = 0.001) (Fig. ). There was no significant difference in pain level during daily activities ( p = 0.560).
The ASAMI bone and functional results and Lower Extremity Functional Index evaluations are summarized in Table . The ASAMI bone and functional results showed no significant difference between the groups ( p > 0.05). The LEFI results were similar with no statistical difference between the groups ( p > 0.05).
The evaluation of complications according to Paley classification by groups is shown in Table and the incidence of pin-tract infections is shown in Table . According to the Paley classification, 23 complications were reported in Group 1 and 30 in Group 2. Pin-tract infection was detected in 50% in Group 1 and 93.8% in Group 2 ( p = 0.013; Table ). According to the Checketts-Otterburn Classification, a total of 8 patients in Group 1 developed pin-tract infections. Among these infections, 5 patients were classified as Grade 1 and 3 patients as Grade 2. In Group 2, a total of 15 patients developed pin-tract infections; 7 were classified as Grade 1, 6 as Grade 2, and 2 as Grade 3 (Fig. ).
All complications were managed based on their severity, as follows: Pin-tract infections: Treated with pin-tract care, local antiseptic dressings, and oral antibiotics when needed. Loss of joint motion: Physical therapy and rehabilitation programs were used to help patients regain their ability to move. Delayed regenerate maturation, delayed consolidation, and delayed healing at the docking site: Managed with autologous bone marrow-derived mesenchymal stem cell (MSC) injections and bone grafting when necessary. Compression was additionally performed as needed for delayed healing at the docking site. >2.5 cm limb shortening: Patients were followed up and required limb elevating insoles or shoe modification. Re-fracture: Managed using circular external fixators to provide stability and allow for proper healing.
The most important conclusion of this study is that both the CASt and CEFt techniques are efficacious in the management of tibial bone defects, and provide similar clinical and radiologic results. However, CASt offers certain advantages over CEFt, including lower pain scores during distraction and a lower incidence of pin-tract infection. These findings support the hypothesis that the CASt method would have similar efficacy while providing certain benefits. Various external fixator systems can be used to treat bone defects using the principles of distraction osteogenesis. The most commonly used methods are circular external fixators and monolateral external fixators. Both methods have similar union and functional scores and have high success rates . It has been shown that monolateral fixators are often preferred due to ease of application and lower patient discomfort, while circular fixators allow better deformity correction and stability . Although the CASt method is known to have a more limited structure in terms of mechanical stability compared to the CEFt method, it is thought that this method may positively affect the consolidation process because it is less invasive and there is less soft tissue damage . It has been documented that regenerated mineralisation is more pronounced in monolateral fixators than in circular fixators . This hypothesis is supported by the observations of this study. Although no significant difference was observed between the two groups in terms of EFI and RCI, the values for these parameters were found to be lower in the CASt group. The use of an external fixator presents a significant challenge for patients. It has been reported that patients who used monolateral fixators for the treatment of tibial bone defect were more satisfied with the postoperative results and their quality of life improved. The use of monolateral fixators in the management of bone defects of the tibia has therefore been recommended . The lengthy nature of the management process can have a detrimental impact on patients’ mental health, both during and after the course of treatment. This negative effect is most evident during the distraction phase and is significantly higher in patients undergoing management with a circular external fixator . This may have an impact on the patients’ perception of pain and comfort during the course of treatment, and may be one of the reasons why the VAS scores in the distraction phase, were significantly lower in the CASt group in this study. The present study demonstrated that pain scores (VAS) during distraction were significantly lower in the CASt group. This may be attributed to the fact that the CASt method minimises skin irritation and soft tissue complications by reducing the utilisation of transosseous pins. The use of Schanz and taut pins in the CEFt method results in the tearing of the skin during segmental sliding, consequently leading to an increase in the pain sensation (Fig. ) . The disappearance of the notable disparity in the distraction phase observed in this study following treatment lends support to this notion. The functional results evaluated with ASAMI scores demonstrated no statistically significant difference between the two groups. Nevertheless, the fact that patients in the CASt group had an earlier return to activities of daily living and fewer problems such as limping and joint stiffness compared to the CEFt group can be considered an indicator of the advantages of this method. This is supported by the slightly higher rate of excellent functional outcome in the CASt group. Previous research has similarly indicated that both techniques are efficacious in facilitating functional independence [ , – ]. The most prevalent complication associated with segmental bone transport is the development of pin-tract infection . The rate is increased by prolonged treatment time, a larger defect, transosseous advancing wires and Schanz pins . Nevertheless, it has been emphasized in the literature that the implementation of appropriate pin-tract care and infection control protocols can mitigate the long-term effects of these complications [ , , ]. In the current study, among the most common complications encountered was pin-tract infection, which exhibited a significantly lower incidence in the CASt method. It was thought that this discrepancy can be attributed to the reduction in the number of Schanz and pins utilised in CASt, which has the effect of preserving the integrity of the surrounding soft tissue. Additional complications associated with segmental bone transport include diminished joint mobility, pain, re-fracture, delayed regeneration, and complications at the docking site. Specifically, patients with large bone defects showed decreased joint mobility and delayed regenerative maturation. The majority of these complications are associated with the dimensions of the bone defect and the prolonged use of external fixation devices, as previously discussed . Delayed regenerated maturation, delayed consolidation, and delayed healing at the docking site were managed with autologous bone marrow-derived mesenchymal stem cell (MSC) injections and bone grafting when necessary, which are the recommended methods. Although meta-analyses have not demonstrated a significant effect, patients were also supported with calcium and vitamin D supplementation . Moreover, adjunctive approaches such as low-intensity pulsed ultrasound (LIPUS), extracorporeal shock wave therapy (ESWT), and pneumatic compression therapy, which are recommended in the literature to enhance bone healing and promote regenerate maturation, were not utilized in the treatment . Three patients with re-fracture were successfully treated with a circular external fixator. The results of this study demonstrate that both CASt and CEFt techniques can be employed effectively in the management of tibial bone defects. Furthermore, the findings indicate that the CASt method is an efficacious management modality that mitigates the adverse effects associated with conventional techniques and enhances patient comfort. It should be noted, however, that CASt is a management method with a relatively high learning curve and therefore requires a significant investment of time to master. Limitation This study had several limitations, primarily that the retrospective design may have posed a methodological limitation due to the lack of timely observation of events. Secondly, due to the relatively small sample size, the results obtained may be insufficient to detect smaller differences between groups. Furthermore, the data used in the study were collected from a single centre, which may make it difficult to generalise to patient groups in different geographical regions and involving different management approaches. Being aware of the limitations, the results obtained were interpreted carefully by taking these factors into consideration.
This study had several limitations, primarily that the retrospective design may have posed a methodological limitation due to the lack of timely observation of events. Secondly, due to the relatively small sample size, the results obtained may be insufficient to detect smaller differences between groups. Furthermore, the data used in the study were collected from a single centre, which may make it difficult to generalise to patient groups in different geographical regions and involving different management approaches. Being aware of the limitations, the results obtained were interpreted carefully by taking these factors into consideration.
The results of this study demonstrated that both CASt and CEFt methods are effective in the management of bone defects of the tibia and provide similar clinical and radiological outcomes. However, the CASt method has the potential to improve patient comfort with lower pain scores and complication rates. Given the less invasive nature of CASt, this may be preferable in patients at higher risk of infection or with a low pain threshold. However, the technical complexity of the method requires experienced surgical application.
|
Factors affecting patency time and semen quality in a single-armed microsurgical vasoepididymostomy | bbccf32e-22e6-4d1c-b4d1-ad1a5aad920f | 11614167 | Microsurgery[mh] | Epididymal obstructive azoospermia (EOA) refers to a condition where the testes produce sperm normally, but the sperm cannot be ejaculated because of a blockage in the epididymal ducts, resulting in azoospermia and accounting for 30%–77% of azoospermic nonvasectomized men. Microsurgical vasoepididymostomy (MVE) is an effective surgical method for treating EOA and is considered the most challenging surgical procedure of all male microsurgeries. MVE techniques have gradually developed and improved over the past few decades. The two-suture double-armed longitudinal intussusception vasoepididymostomy (DA-LIVE) procedure has demonstrated superior outcomes and has now become the established standard of care. However, Monoski et al . showed comparable patency rates to the DA-LIVE procedure using the single-armed MVE method in animal experiments. Zhao et al . first performed a two-suture single-armed longitudinal intussusception vasoepididymostomy (SA-LIVE) for EOA in humans. They believed that this approach offered the advantage of using an easily accessible and cost-effective single-armed suture. Several studies have reported that the patency rates of SA-LIVE range from 55.2% to 83.1%. Despite the established efficacy of SA-LIVE in EOA management, uncertainties remain, particularly in the patency time and semen quality postsurgery. The length of time required for postoperative patency of SA-LIVE, sperm parameters, and related influencing factors are critical issues for both patients and surgeons. Extended patency time and potential suboptimal sperm quality postpatency may not fulfill patients’ and their partners’ desires for early conception. Hence, further studies are essential to elucidate the factors impacting patency time and semen quality after MVE employing SA-LIVE technique. Therefore, this study aimed to investigate the patency time and semen parameters after SA-LIVE. Simultaneously, relevant influencing factors were analyzed to provide effective information for treatment and decision-making in patients with EOA.
Study population and participants We retrospectively collected data from 181 patients diagnosed with EOA who underwent the SA-LIVE procedure at the First Affiliated Hospital of Fujian Medical University (Fuzhou, China) from October 2019 to February 2023. This study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Fujian Medical University [MRCTA, ECFAH of FMU (2019) 213] and was performed in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients. The diagnostic criteria for EOA included azoospermia confirmed by at least two centrifuged semen analyses, and normal serum levels of total testosterone, follicle-stimulating hormone, and inhibin B. No significant genetic abnormalities were detected through chromosome analysis of peripheral blood lymphocytes and Y chromosome microdeletion testing. Color Doppler ultrasound examination (GE LOGIQ Fortis; GE Healthcare, Chicago, IL, USA) showed at least one normal testicular volume and the presence of vas deferens. Magnetic resonance imaging (MRI; Magnetom Prisma; Siemens, Munich, Germany) showed no abnormalities in the seminal vesicles and prostate. This study exclusively included cases of obstructive azoospermia due to epididymal obstruction, excluding cases of intratesticular obstructive azoospermia and those with ejaculatory duct obstruction. SA-LIVE surgical approach The patients were positioned supine and given combined lumbar and epidural or general anesthesia. A surgical microscope with a maximum magnification of 20× (Zeiss S88; Carl Zeiss, Oberkochen, Germany), an inverted microscope (CKX31; Olympus, Tokyo, Japan), a microbipolar electrocautery, and standard microscopic instruments were prepared for the procedure. All surgeries were performed by the same surgeon (SXT). We confirmed the presence of epididymal obstruction by observing the absence of sperm in fluid collected from the testicular end of the vas deferens following its partial transection under a microscope. To ensure the patency of the distal vas deferens, we injected a diluted solution of methylene blue into the vas deferens lumen toward the seminal vesicles and verified the presence of methylene blue in urine. By examining the epididymis microscopically, we determined the locations of blockages within the epididymis and pinpointed an appropriately dilated epididymal duct proximal to the obstruction. This segment typically exhibited dilation, indicating its suitability for anastomosis. The chosen longitudinal epididymal tubule was located as close as possible to the blockage for anastomosis. A precise incision was made into the epididymal tunic, ensuring alignment with the vas deferens’ diameter. The identified epididymal tubule was then carefully dissected under microscopic guidance, preparing for subsequent anastomosis ( ). Subsequently, the vas deferens was completely transected under a microscope, the proximal vas deferens was ligated, and the distal vas deferens was freed to reduce anastomotic tension. A micro-marking pen was used to make four micropoints at the 1-o’clock, 5-o’clock, 7-o’clock, and 11-o’clock positions in the muscular layer of the vas deferens as insertion points ( ). One 8-0 prolene thread (W2777; 6.5 mm, 3/8C; Ethicon, Somerville, MA, USA.) was used to fix the outer membrane of the vas deferens and epididymis in the 6-o’clock position. Two 10-0 single-armed prolene threads (W2790; 3.8 mm, 3/8C; Ethicon.) were used to penetrate the vas deferens lumen from the outside to the inside of the mucosal layer at the marked points at 5-o’clock and 7-o’clock positions ( and ). Two parallel stitches were sewn into the preanastomosed epididymal tube without temporarily pulling it out. A sharp knife was used to gently cut the wall of the epididymal tube longitudinally between the two single-armed prolene threads ( and ). After the epididymal fluid overflowed, a thin tube was used to draw and dilute the fluid, and the presence of sperm was immediately observed under an inverted microscope. After microscopically identifying the presence of sperm, two single-armed 10-0 prolene threads (Ethicon) were delicately pulled out and threaded through the vas deferens lumen from the inside out, precisely at the 1-o’clock and 11-o’clock positions ( and ). The surgical steps described above are similar to those previously reported by Hong et al . However, we made minor adjustments to our operating procedure. We also utilized an 8-0 prolene thread (Ethicon) to suture the vas deferens and epididymal tunic in the 12-o’clock direction. Before proceeding to tie the knot, we chose to first gently tighten and separately tie the two 10-0 prolene threads (Ethicon; and ). This modification had an advantage in that the epididymal tubules were precisely intussuscepted into the lumen of the vas deferens under direct visualization through the microscope. The schematic diagram of the procedure was shown in . Finally, the vas deferens and epididymal tunic were intermittently sutured with 12–14 stitches using 8-0 prolene threads (Ethicon), serving to secure and reduce tension. Postoperative care and follow-up After the surgery, patients were instructed to wear a scrotal support for 6 weeks and refrain from strenuous exercise. Masturbation and sexual intercourse were not allowed for 4 weeks, after that sexual activity was resumed 1–2 times per week. The first two semen analyses were conducted at 1.5 months and 3 months after surgery, followed by subsequent analyses at 4–6 months, 7–9 months, and 10–12 months or later after the procedure. The semen analysis timing, semen parameters, and the spouse’s pregnancy status were recorded either through outpatient consultations or over the phone. The 5 th edition of the World Health Organization standards was employed for semen analysis, with our laboratory’s sperm analysis corresponding with the standardization described in the inspection table published by Björndahl et al . Patency was defined as a sperm concentration ≥1 × 10 ml −1 in the postoperative semen sample. The patency time was defined as the months between surgery and the first semen analysis that confirmed patency. If sperm was initially detected and then not detected again more than two times, re-obstruction was considered. Clinical pregnancy in the patient’s partner indicated successful natural pregnancy. If no sperm was found in the semen examination after more than 12 months postoperatively, assisted reproductive technology was recommended. Statistical analyses SPSS 26.0 software (IBM Corp., Armonk, NY, USA) was used for statistical data processing, and the measurement data were expressed as mean ± standard deviation (s.d.) or median (interquartile range [IQR]). Count data were described using examples or percentages. Data distribution was assessed using the Shapiro–Wilk test, and homogeneity of variance was examined using Levene’s test. Pearson’s Chi-square test was used to compare the rates. The Mann–Whitney U test was employed to compare data that did not conform to the normal distribution. Univariate linear regression analysis was conducted to ascertain the determinants of patency time. Additionally, both univariate and multivariate linear regression analyses were conducted to identify the factors influencing total motile sperm count (TMSC) among patients who achieved patency following MVE. All statistical tests were two-tailed, with statistical significance set at P < 0.05.
We retrospectively collected data from 181 patients diagnosed with EOA who underwent the SA-LIVE procedure at the First Affiliated Hospital of Fujian Medical University (Fuzhou, China) from October 2019 to February 2023. This study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Fujian Medical University [MRCTA, ECFAH of FMU (2019) 213] and was performed in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients. The diagnostic criteria for EOA included azoospermia confirmed by at least two centrifuged semen analyses, and normal serum levels of total testosterone, follicle-stimulating hormone, and inhibin B. No significant genetic abnormalities were detected through chromosome analysis of peripheral blood lymphocytes and Y chromosome microdeletion testing. Color Doppler ultrasound examination (GE LOGIQ Fortis; GE Healthcare, Chicago, IL, USA) showed at least one normal testicular volume and the presence of vas deferens. Magnetic resonance imaging (MRI; Magnetom Prisma; Siemens, Munich, Germany) showed no abnormalities in the seminal vesicles and prostate. This study exclusively included cases of obstructive azoospermia due to epididymal obstruction, excluding cases of intratesticular obstructive azoospermia and those with ejaculatory duct obstruction.
The patients were positioned supine and given combined lumbar and epidural or general anesthesia. A surgical microscope with a maximum magnification of 20× (Zeiss S88; Carl Zeiss, Oberkochen, Germany), an inverted microscope (CKX31; Olympus, Tokyo, Japan), a microbipolar electrocautery, and standard microscopic instruments were prepared for the procedure. All surgeries were performed by the same surgeon (SXT). We confirmed the presence of epididymal obstruction by observing the absence of sperm in fluid collected from the testicular end of the vas deferens following its partial transection under a microscope. To ensure the patency of the distal vas deferens, we injected a diluted solution of methylene blue into the vas deferens lumen toward the seminal vesicles and verified the presence of methylene blue in urine. By examining the epididymis microscopically, we determined the locations of blockages within the epididymis and pinpointed an appropriately dilated epididymal duct proximal to the obstruction. This segment typically exhibited dilation, indicating its suitability for anastomosis. The chosen longitudinal epididymal tubule was located as close as possible to the blockage for anastomosis. A precise incision was made into the epididymal tunic, ensuring alignment with the vas deferens’ diameter. The identified epididymal tubule was then carefully dissected under microscopic guidance, preparing for subsequent anastomosis ( ). Subsequently, the vas deferens was completely transected under a microscope, the proximal vas deferens was ligated, and the distal vas deferens was freed to reduce anastomotic tension. A micro-marking pen was used to make four micropoints at the 1-o’clock, 5-o’clock, 7-o’clock, and 11-o’clock positions in the muscular layer of the vas deferens as insertion points ( ). One 8-0 prolene thread (W2777; 6.5 mm, 3/8C; Ethicon, Somerville, MA, USA.) was used to fix the outer membrane of the vas deferens and epididymis in the 6-o’clock position. Two 10-0 single-armed prolene threads (W2790; 3.8 mm, 3/8C; Ethicon.) were used to penetrate the vas deferens lumen from the outside to the inside of the mucosal layer at the marked points at 5-o’clock and 7-o’clock positions ( and ). Two parallel stitches were sewn into the preanastomosed epididymal tube without temporarily pulling it out. A sharp knife was used to gently cut the wall of the epididymal tube longitudinally between the two single-armed prolene threads ( and ). After the epididymal fluid overflowed, a thin tube was used to draw and dilute the fluid, and the presence of sperm was immediately observed under an inverted microscope. After microscopically identifying the presence of sperm, two single-armed 10-0 prolene threads (Ethicon) were delicately pulled out and threaded through the vas deferens lumen from the inside out, precisely at the 1-o’clock and 11-o’clock positions ( and ). The surgical steps described above are similar to those previously reported by Hong et al . However, we made minor adjustments to our operating procedure. We also utilized an 8-0 prolene thread (Ethicon) to suture the vas deferens and epididymal tunic in the 12-o’clock direction. Before proceeding to tie the knot, we chose to first gently tighten and separately tie the two 10-0 prolene threads (Ethicon; and ). This modification had an advantage in that the epididymal tubules were precisely intussuscepted into the lumen of the vas deferens under direct visualization through the microscope. The schematic diagram of the procedure was shown in . Finally, the vas deferens and epididymal tunic were intermittently sutured with 12–14 stitches using 8-0 prolene threads (Ethicon), serving to secure and reduce tension.
After the surgery, patients were instructed to wear a scrotal support for 6 weeks and refrain from strenuous exercise. Masturbation and sexual intercourse were not allowed for 4 weeks, after that sexual activity was resumed 1–2 times per week. The first two semen analyses were conducted at 1.5 months and 3 months after surgery, followed by subsequent analyses at 4–6 months, 7–9 months, and 10–12 months or later after the procedure. The semen analysis timing, semen parameters, and the spouse’s pregnancy status were recorded either through outpatient consultations or over the phone. The 5 th edition of the World Health Organization standards was employed for semen analysis, with our laboratory’s sperm analysis corresponding with the standardization described in the inspection table published by Björndahl et al . Patency was defined as a sperm concentration ≥1 × 10 ml −1 in the postoperative semen sample. The patency time was defined as the months between surgery and the first semen analysis that confirmed patency. If sperm was initially detected and then not detected again more than two times, re-obstruction was considered. Clinical pregnancy in the patient’s partner indicated successful natural pregnancy. If no sperm was found in the semen examination after more than 12 months postoperatively, assisted reproductive technology was recommended.
SPSS 26.0 software (IBM Corp., Armonk, NY, USA) was used for statistical data processing, and the measurement data were expressed as mean ± standard deviation (s.d.) or median (interquartile range [IQR]). Count data were described using examples or percentages. Data distribution was assessed using the Shapiro–Wilk test, and homogeneity of variance was examined using Levene’s test. Pearson’s Chi-square test was used to compare the rates. The Mann–Whitney U test was employed to compare data that did not conform to the normal distribution. Univariate linear regression analysis was conducted to ascertain the determinants of patency time. Additionally, both univariate and multivariate linear regression analyses were conducted to identify the factors influencing total motile sperm count (TMSC) among patients who achieved patency following MVE. All statistical tests were two-tailed, with statistical significance set at P < 0.05.
In our study, 181 patients successfully underwent at least a one-sided MVE. Of these, 42.0% (76/181) had a history of epididymitis, and 58.0% (105/181) had idiopathic causes for their condition; none were related to vasectomy. We successfully followed up 150 of these patients over 12 months. The follow-up duration (mean ± s.d.) was 15.2 ± 2.5 months, ranging from 12 months to 29 months. Patients’ ages ranged from 20 years to 49 years (mean ± s.d.: 30.2 ± 5.0 years). Of the 150 patients, 122 patients underwent bilateral MVE, while the remaining 28 patients had unilateral MVE. Thirty-five patients underwent anastomosis solely at the caput, while 115 patients underwent anastomosis on at least one side to the corpus or cauda. The patency rate was 75.3% (113/150). However, three initial patent cases later developed re-obstruction, as evidenced by the absence of sperm in subsequent evaluations, resulting in a re-obstruction rate of 2.7% (3/113). We investigated factors associated with successful patency versus nonpatency after the MVE procedure. Our analysis indicated that a significantly higher proportion of patients with successful patency had motile sperm in the intraoperative epididymal fluid compared with those without patency ( P < 0.001; ). We determined the initial patency time across 100 cases, with observed time ranging from 1.5 months to 12 months (mean ± s.d.: 4.9 ± 2.7 months). Analysis revealed that 18 patients achieved patency within 1.5 months, 27 patients between 1.5 months and 3 months, and 41 patients between 4 months and 6 months. A smaller group achieved patency after 6 months; 10 patients between 7 months and 9 months, and four patients between 10 months and 12 months. Notably, 86.0% (86/100) of the patients achieved patency within the 6 months after MVE. Through univariate linear regression analysis, it was found that age, history of epididymitis, surgical duration, side of anastomosis, sperm motility of epididymal fluid, and site of anastomosis were not correlated with the patency time (all P > 0.05; ). Among the 113 patients who underwent MVE and achieved successful patency, the median TMSC was 6.05 × 10 6 (range from 0 to 150.42 × 10 6 ). Through both univariate and multivariate linear regression analyses, the site of anastomosis was identified as a significant independent predictor of TMSC after MVE. Notably, patients who had anastomosis at the epididymal corpus or cauda exhibited a significantly higher TMSC compared with those with anastomosis solely at the caput ( ). The natural pregnancy rate was 37.3% (56/150). Among the subset of 113 patients who demonstrated successful patency postprocedure, three experienced re-obstruction and nine were unmarried, leaving 101 married patients for our analysis. The age of the spouses (mean ± s.d.) in this group was 28.1 ± 3.7 years (range from 22 years to 38 years). The TMSC of patients who had a natural pregnancy postpatency were significantly higher than those of who did not (median [IQR]: 24.85 [6.43, 39.72] vs 0.95 [0.11, 5.55]; P < 0.001).
Our review of 181 cases performed by the same surgeon revealed a patency rate of 75.3% and a natural pregnancy rate of 37.3%. These results, derived from a significant caseload, further substantiate the efficacy of the SA-LIVE technique for patients with EOA. We also noted postoperative delays in patency time. In the preliminary semen analysis at 1.5 months, only 18.0% showed first-time patency. However, this figure rose to 45.0% at 3-month and climbed to 86.0% within 6 months. This finding indicates that most of the patency for SA-LIVE occurs within 6 months, in line with past findings on patency time after DA-LIVE. The causes of this delay remain unclear but may be due to several factors, such as resolution of local tissue inflammation, edema, and blood clots, changes in epididymal physiology, or recovery of spermatogenic efficiency in testes following long-term epididymal obstruction. To explore the factors influencing patency time following SA-LIVE, we conducted linear regression analysis on several variables, including age, history of epididymitis, duration of surgery, side of anastomosis, sperm motility in epididymal fluid, and site of anastomosis. However, no significant associations were observed between these factors and patency time following SA-LIVE, contrasting with previously reported findings associated with patency rates. Prior studies indicated that the presence of motile sperm in the epididymal fluid, as well as bilateral or distal anastomoses, were associated with higher patency rates. These findings indicate that the mechanisms affecting patency time postsurgery could be more complex and multifaceted than considered in our study. To the best of our knowledge, few studies have analyzed these specific variables in relation to patency time after SA-LIVE. Future research is warranted to explore potential factors and to further our understanding of the determinants of patency time postsurgery. To comprehensively assess sperm motility and counts following patency, we employed the TMSC as the postoperative evaluation metric. Univariate and multivariate linear regression analysis revealed that exclusively the site of anastomosis was positively correlated with and independently predictive of the postoperative TMSC. This suggests that the site of anastomosis, especially in the corpus or caudal region of the epididymis during SA-LIVE, is an effective predictor of postpatency TMSC. One possible explanation is the relatively poor sperm motility in the caput region compared with the cauda, as noted in the assessment during MVE by Pal et al . Additionally, MVE surgery may compromise the functional length of the epididymis. This could lead to partial epididymal transit and subsequently reduced semen quality. As a result, we propose that the site of anastomosis may predict optimal postoperative semen quality after the MVE procedure. Although pregnancy post-MVE involves partner-related factors, Peng et al . maintained that sperm concentration and motility could effectively predict natural pregnancy outcomes after patency. Our study also revealed higher levels of TMSC in patients who successfully achieved a natural pregnancy postpatency, compared with those who did not. Hence, the aim of EOA treatment should extend beyond improving patency rates to include enhancing sperm quality postpatency, thereby improving the partner’s conception likelihood. Consequently, it is essential to identify patients who have the potential for favorable sperm quality after MVE. Our study has certain limitations. First, we had to exclude some patients from the study because of inadequate postoperative semen analysis data and could potentially limit the generalizability of our findings. For instance, patients with pregnant spouses may not have completed the semen analysis at each interval. Second, there was a lack of data on sperm morphology, attributed to the low sperm count observed in some patients following the procedures, which may limit the understanding of the procedures’ full impact of sperm quality. In conclusion, our results showed that after SA-LIVE, most patients with EOA experienced patency within 6 months. Furthermore, factors such as patient age, history of epididymitis, duration of surgery, side of anastomosis, sperm motility in epididymal fluid, as well as the site of anastomosis, did not predict patency time. The site of anastomosis was found to be a positively correlated independent factor affecting sperm quality. By incorporating these findings, clinicians can provide patients with EOA more detailed information about expected outcomes, in terms of patency time and sperm quality following SA-LIVE, thus refining fertility planning.
SXT, HX, HXZ, and HLZ made significant contributions to the concept and design. SXT, HX, YLD, PY, HLH, XC, and SZ conducted those clinical studies. SXT, HX, HXZ, and QC took part in the collection, analysis, and interpretation of the data and made crucial changes to the primary knowledge content of the article. SXT, HX, HLZ, and HXZ provided major revisions to critical knowledge content. All authors were involved in revising the manuscript, and read and approved the final manuscript.
All authors declare no competing interests.
|
Regulation of heterogeneous cancer-associated fibroblasts: the molecular pathology of activated signaling pathways | 5a77eb1f-1f44-47fa-8b3f-c66dc631f0d2 | 7296768 | Pathology[mh] | Fibroblasts are spindle-shaped cells that secrete collagen and have a cytoplasm with a predominant rough endoplasmic reticulum. Fibroblasts synthesize the extracellular matrix (ECM) of the connective tissue and play a crucial role in maintaining the structural integrity of most tissues including the skin . The mammalian dermis represents an archetypal mesenchymal tissue that is largely composed of ECM elements, including type I and type III collagens, as well as proteoglycans and elastin . Fibroblast heterogeneity depends on developmental stage and the tissue microenvironment . Fibroblasts exhibit distinct cellular phenotypes according to the surrounding microenvironment. Activated fibroblasts in tumor tissues are defined as cancer-associated fibroblasts (CAFs) . In recent years, extensive research demonstrated that CAFs are the major cellular components of the tumor microenvironment in both primary and metastatic tumors; CAFs contribute to the regulation of a series of steps critical for malignant progression, including cancer initiation, proliferation, invasion, and metastasis, by producing various types of cytokines, chemokines, growth factors, and matrix-degrading enzymes . CAFs are distinguished from their normal counterparts by the differential expression of markers such as α-smooth muscle actin (α-SMA), fibroblast activation protein (FAP), fibroblast-specific protein 1 (FSP1), and platelet-derived growth factor receptor (PDGFR) . In addition to these markers, three proteins including collagen 11-α1, microfibrillar-associated protein 5, and asporin tend to be exclusively expressed in CAFs . The recent identification of proteins whose expression is restricted to CAFs may improve the reliable identification of CAFs and increase their value as candidate biomarkers and therapeutic targets. CAFs and activated fibroblasts play similar roles in wound healing and fibrosis . Fibroblasts are essential for tissue repair after damage and are involved in wound contraction, deposition of granulation tissue, and production and remodeling of the ECM in parallel with recruitment of platelets, neutrophils, and macrophages (Fig. ). Fibroblasts in granulation tissue acquire a myofibroblastic phenotype characterized by α-SMA expression. On the other hand, CAFs alter the microenvironment by directly interacting with cancer cells and regulating paracrine signaling via inflammatory cytokines, control the immune response to neoplasia, deposit diverse ECM components, stimulate angiogenesis, and provide a scaffold for tumor metastasis and invasion . CAFs can be recruited to the tumor from a distant source such as the bone marrow . The trans-differentiation of epithelial cells and pericytes can also give rise to CAF-like populations in response to epithelial-mesenchymal transition (EMT) and endothelial-mesenchymal transition (EndoMT), respectively . To define and identify the origin of CAFs, it is important to consider that CAFs are ‘activated fibroblasts’, which, by striking contrast to non-activated (quiescent) tissue-resident fibroblasts, are an expanding population of cells that either proliferates in situ or is recruited to the tumor . The key features of CAFs that distinguish them from quiescent fibroblasts include metabolic adaptations that support their need for enhanced proliferation and biosynthetic activities, such as the production of ECM components and cytokines, growth factors, and enzymes to remodel the stroma . However, the cellular origin of CAFs and the mechanisms underlying the reprogramming of normal fibroblasts into CAFs remain largely unknown. The heterogeneity and mutual exclusivity of CAF marker expression patterns may be associated with unique functions in different types of malignancy. In breast cancer, CAFs positive for both FAP and podoplanin are immunosuppressive through a nitric oxide (NO)-dependent mechanism . In prostate cancer, CAFs expressing high levels of CD90 play a pivotal role in promoting tumor progression through the upregulation of angiogenic factors, activation of the Hedgehog (Hh) signal, and decreased androgen receptor signaling . In pancreatic ductal adenocarcinoma (PDAC), a specific subpopulation of CAFs was identified that is distinct from myofibroblastic CAFs strongly expressing α-SMA. These inflammatory CAFs express pro-inflammatory cytokines such as interleukin-6 (IL-6) and IL-11, thereby activating the Janus kinase (JAK)/signal transducer and activator of transcription (STAT) signaling pathway . Therefore, this novel review article focuses on critical signal pathways for CAFs to regulate the malignant phenotype, given the crosstalk between tumor cells and CAFs as well as the heterogeneity of CAF population. Increasing evidence strongly suggests that CAFs have diverse functions, implying that tumor-promoting CAF and tumor-suppressing CAFs coexist in the tumor stroma . Current cancer immunotherapy strategies primarily target programmed cell death-1 ligand-1 (PD-L1), chimeric antigen receptors, and cytotoxic T lymphocyte-associated antigen 4 ; however, the effects of CAFs on tumor immunosuppression remain relatively unexplored. FAP-positive CAF populations drive immunosuppression and promote resistance to anti-PD-L1 immunotherapy . Targeting the C-X-C motif chemokine 12 (CXCL12)–C-X-C chemokine receptor type 4 (CXCR-4) axis with AMD3100 (Plerixafor) reverses FAP-positive CAF-mediated immunosuppression and synergizes with anti-PD-L1 immunotherapy in PDAC . Biffi et al. recently identified IL-1 and transforming growth factor-β (TGF-β) as tumor-secreted cytokines that promote CAF heterogeneity . TGF-β signaling inhibits IL-1 receptor 1 (IL-1R1) expression, antagonizes IL-1α responses, and promotes the differentiation of paraneoplastic fibroblasts into myofibroblastic CAFs. By contrast, an IL-1-induced signaling cascade that activates JAK/STAT promotes the generation of inflammatory CAFs . Therefore, IL-1α signaling is a potential therapeutic target against PDAC cells and inflammatory CAFs in the tumor microenvironment. Elyada et al. used single-cell transcriptomics to examine CAF heterogeneity associated with PDAC and identified a novel CAF population characterized by high expression levels of major histocompatibility complex class II . These antigen-presenting CAFs can present antigens to CD4-positive T lymphocytes. Costa et al. used multicolor flow cytometry to identify four subtypes of CAFs (CAF-S1, 2, 3, and 4) associated with breast cancer that show differential expression of CAF marker molecules such as α-SMA, caveolin-1 (Cav-1), FAP, and PDGFRβ. CAF subset maps confirmed that CAF-S2 is enriched in luminal-type breast cancer, whereas CAF-S1 and CAF-S4 accumulate in triple-negative breast cancer. The CAF-S1 type attracts CD4 + CD25+ T lymphocytes and retains them through OX40L, PD-L2, and JAM2. CAF-S1 cells promote the inhibition of T effector proliferation by regulatory T cells (Treg). Mechanistically, Costa et al. identified dipeptidyl peptidase 4, a FAP-dimerization counterpart, as a key player in CAF-S1-mediated Treg activation . Givel et al. identified four subpopulations of CAFs in mesenchymal-type high-grade serous ovarian cancer (HGSOC) according to α-SMA, CD29 (integrin β-1), FAP, and FSP1 expression levels . CAF-S2 and CAF-S3 are defined as non-activated CAFs because they have low levels of α-SMA, whereas CAF-S1 and CAF-S4 are considered activated CAFs with high α-SMA expression levels. FOXP3-positive T lymphocytes are enriched exclusively in the CAF-S1 population, which differs from the CAF-S4 subtype in mesenchymal-type HGSOC. The chemokine CXCL12β is expressed at higher levels in CAF-S1 than in CAF-S4, which is associated with the poor prognosis of HGSOC . The CXCL12β isoform is regulated by microRNA (miR)-141/200a; the miR-200 family members miR-141 and miR-200a are responsible for the downregulation of CXCL12β in CAF-S4, whereas miR141/200a promote the specific accumulation of CACL12β in the CAF-S1 subpopulation, inducing the infiltration of CD4 + CD25+ T lymphocytes . Taken together, these data indicate that the antagonistic effects of CAFs on the malignant phenotype may be related to the existence of subpopulations of CAFs with opposing functions. Recent PDAC studies challenged the concept of tumor-promoting CAFs based on data showing increased tumor growth and aggressiveness following eradication of α-SMA-expressing CAFs and/or targeting of the desmoplastic response induced by the Hh signaling pathway . PDAC lesions with approximately 80% depletion of α-SMA-positive interstitial myofibroblasts show an activated EMT program associated with increased numbers of cancer stem cells (CSCs) and upregulation of EMT-related transcription factors such as Snail, Slug, and Twist. Clinically, lower CAF numbers are correlated with decreased survival in patients with PDAC . Although sonic hedgehog (Shh) ligand and downstream signaling are induced de novo in preneoplastic lesions linked to pancreatic intraepithelial neoplasia and increase significantly during PDAC progression as the stromal compartment enlarges , a Shh-depleted PDAC mouse model showed Slug and Zeb1 upregulation leading to poorly differentiated histology . Mizutani et al. identified Meflin as a functional marker of tumor-retarding CAFs in PDAC . Meflin, a glycosylphosphatidylinositol-anchored protein, is a marker of mesenchymal stem cells (MSCs) and maintains their undifferentiated state . In situ hybridization analysis revealed an inverse correlation between α-SMA and Meflin expression in PDAC-associated CAFs. Kaplan-Meier survival analyses showed that high expression levels of Meflin in surgically resected human PDAC tissues are positively correlated with better prognosis, and Meflin-high PDAC displays a more differentiated pathohistology than the Meflin-low group. This suggests that the phenotype of Meflin-high CAFs is distinct from that of tumor-promoting CAFs with high α-SMA expression levels. The PDAC-associated stroma in Meflin-KO genetically engineered model mice shows straighter and wider collagen structures than those of tumors in Meflin-WT PDAC model mice. Lineage-tracing experiments indicate that Meflin-lineage stromal cells contain α-SMA-positive CAFs, which downregulate Meflin and upregulate α-SMA in response to TGF-β and tumor stiffness. This explains why the stromagenic switch, in which tumor-restricting CAFs with high Meflin expression generate tumor-promoting CAFs, contributes to CAF heterogeneity during tumor progression. Increasing evidence suggests that CAFs contribute to collective cell migration and invasion by remodeling the ECM to create tracks for tumor cell migration and/or by expressing different cadherins that enable cells to retain adhesion while controlling front/rear polarization of the leading cells . Labernadie et al. showed that CAFs increase the invasive potential of tumor cells through N-cadherin upregulation . Intercellular physical force is transmitted between cancer cells and CAFs by a heterophilic adhesion complex involving E-cadherin at the cancer cell membrane and N-cadherin at the CAF membrane. This heterotypic cancer cell-CAF interaction triggers β-catenin recruitment, α-catenin/vinculin interaction, and actin remodeling, allowing CAFs to exert an intercellular physical force on cancer cells and promote cooperative tumor invasion . CAFs contribute to the ‘education’ of carcinoma cells into an invasive and metastatic phenotype . TGF-β and CXCL12 secreted by CAFs enhances the metastatic potential of breast cancer cells undergoing incomplete EMT. Although developmental cells undergo complete EMT during embryogenesis, which is characterized by the cadherin switch, tumor cells express both epithelial and mesenchymal markers [epithelial/mesenchymal (E/M) hybrid phenotype] concurrently, which is defined as “partial EMT” in the process of invasion and distant metastasis . Indeed, circulating tumor cells that survive in the bloodstream show an E/M hybrid phenotype, become resistant to anoikis, and exit the bloodstream more efficiently . CAFs stimulate the invasion of E/M hybrid-type breast cancer cells, which are associated with epithelial-type cancer cell clusters, leading to collective invasion of both epithelial and E/M hybrid tumor cell clusters . Chen et al. recently reported that the epithelial-to-mesenchymal plasticity of lung cancer cells established from a patient-derived xenograft (PDX) is enhanced in the presence of CAFs under three-dimensional culture . CAFs antagonize the oncogenic transcriptional factor SOX2 to restore the formation of luminal structures and promote invasion. Stromal cell-derived factor 1 promotes EMT and increases the stemness of lung squamous cancer cells (LSCCs). Most LSCCs express E-cadherin, and only a small population is positive for vimentin . This finding suggests that spheroids derived from PDX are heterogeneous. The presence of tumor cells positive for both E-cadherin and vimentin suggests that partial EMT occurs in the original tumor, PDX model, and spheroids . CAFs play an important role in the establishment of the omental tumor microenvironment in ovarian cancer. Omental fibroblasts contribute to the creation of a pre-metastatic niche, and influence tropism for the omentum and the metastatic colonization of ovarian tumor cells . Ovarian cancer-derived lysophosphatidic acid and exosomes promote the differentiation of adipose-derived MSCs into CAFs , which are characterized by the expression of α-SMA, FAP, FSP1, and PDGFR, by activating TGF-β-related signaling pathways . Furthermore, ovarian cancer cells reprogram normal omental fibroblasts into CAFs by upregulating miR-155 and downregulating miR-31 and miR-214 . This action promotes tumor proliferation by increasing the secretion of CCL5. Ovarian cancer-derived TGF-β is involved in stimulating the production of various tumor-promoting factors including IL-6, CXCL12, and VEGF-A in the metastatic tumor microenvironment . Omental dissemination induced by this cascade is driven by overexpression of HOXA9 in ovarian cancer cells. CAF-derived TGF-α promotes the metastatic colonization of ovarian cancer cells via the activation of the Akt, epidermal growth factor receptor (EGFR), and extracellular signal-regulated kinase (ERK)-1/2 signaling pathways . Metastasizing ovarian cancer cells can activate p38α MAPK signaling in omental CAFs, and CAF-derived p38α MAPK-regulated cytokines and chemokines, including IL-6, CCL5, and CXCL10, induce glycogen metabolism in cancer cells via glycolysis, which mediates energy production and promotes the aggressiveness of ovarian cancer cells . Furthermore, the differential expression patterns of monocarboxylate transporters (MCT) in cancer cells and CAFs contribute to metabolic symbiosis, in which CAFs depend on aerobic glycolysis and secrete lactate via MCT4 . This “reverse Warburg effect” enables MCT1-positive CSCs to play a fundamental role in maintaining the hierarchy in tumor cellular society unlike MCT4-positive CAFs . In addition, CAFs tend to exhibit robust activity regarding aerobic glycolysis as well as Atg5/7-dependent selective autophagy because of the loss of Cav-1 expression . Such stromal autophagy generates building blocks from recycled free amino acids, fatty acids, and nucleotides, which can be directly utilized by tumor cells to sustain growth and maintain cellular viability. Therefore, CAFs evolve with ovarian cancer cells in the intraperitoneal metastatic microenvironment and govern the metastatic cascade, including the adhesion, proliferation, invasion, and colonization of metastatic sites . Podoplanin-positive CAFs drive tumor progression in a xenograft model, and podoplanin expression in CAFs predicts a poor outcome in patients with lung adenocarcinoma . However, CAFs positive for podoplanin are more frequent in poorly differentiated adenocarcinoma. Clinical cases characterized by the presence of podoplanin-expressing CAFs display a poor response to EGFR tyrosine kinase inhibitors (EGFR-TKIs) in patients with lung adenocarcinoma harboring constitutively active mutations of EGFR . By contrast, knockdown of podoplanin makes CAFs susceptible to EGFR-TKIs . Direct contact between cancer cells and CAFs is necessary for acquired resistance to EGFR-TKIs. The epidermal growth factor receptor (EGFR) belongs to the ErbB family of receptor tyrosine kinases (RTKs) and exhibits critical functions in the epithelial cell physiology . Ligand-dependent activation of EGFR transduces multiple signaling pathways such as PI3K/Akt and Ras/MAPK pathways . Canonical EGFR signaling is essential for several cellular functions including differentiation, proliferation and survival . Notably, increased EGFR expression is positively correlated with reduced recurrence-free and overall survival periods in several kinds of malignancy . Grasset et al. demonstrated that collective invasion of squamous cancer cells (SCCs) is driven by the matrix-dependent mechano-sensitization of EGF signaling (Fig. a). Increasing evidence suggests a connection between mechanotransduction and receptor tyrosine kinase (RTK) signaling pathways. RTKs are activated by dimerization and are involved in integrin-mediated mechanotransduction signaling, which promotes tumor progression . Induction of collagen crosslinking results in stiffness of the ECM, promotes focal adhesion kinase (FAK) expression, increases phosphoinositide 3-kinase activity, and promotes the invasion of oncogene-initiated epithelial cells. By contrast, suppression of integrin signaling inhibits the invasion of a premalignant epithelium into a stiffened, crosslinked stroma. Cell-to-ECM adhesion favors EGFR-dependent cancer proliferation . Because RTKs interact exclusively with active integrins, the composition of the ECM determines the type of RTK/integrin interaction occurring at the cellular membrane. This selectivity may change the intracellular location or conformation of EGFR, thereby changing the accessibility of the receptor intracellular domain to downstream signaling molecules. One of the downstream proteins is FAK, which is targeted to sites of integrin/RTK complex formation and is essential for the transmission of motility signals from EGFR . Furthermore, the EGFR gene is amplified, overexpressed, or mutated in SCCs, such as head and neck squamous cell carcinoma (HNSCC) . In the clinical setting, EGFR amplification predicts sensitivity to gefitinib in HNSCC . EGFR activation and expression levels are positively correlated with poor prognosis of breast cancer and HNSCC independently from anticancer therapeutics . Grasset et al. identified an association between EGFR activity and stromal stiffness during collective cellular migration . The degree of EGFR signaling is positively correlated with collective cell migration (Fig. a). The L-type calcium channel Ca v 1.1 is a critical regulatory element during the collective invasion of squamous cell carcinoma and acts downstream of ECM stiffness and EGFR signaling both in vitro and in vivo. The L-type calcium channel Ca v 1.1 is a critical regulator of SCC collective migration in response to stromal stiffness and EGFR signaling activation (Fig. b), and calcium channel blockers, which are widely used for the treatment of arrhythmia and hypertension, are promising therapeutic agents against SCC invasion and metastasis. EGFR blockage induces EMT and CAF activation in HNSCC , and the calcium channel antagonists verapamil and diltiazem reduce resistance to EGFR-targeting treatments . This is an example of drug repositioning, namely, screening for the anticancer therapeutic effects of conventionally administered medications for non-malignant disorders . Increased rigidity in the tumor stroma favors EGFR activity and results in the calcium-dependent regulation of Cdc42 small GTPase activity in tumor cells. This signaling route is critical for HNSCC cell invasion into the stiffened stroma. Although myosin light chain (MLC) kinase, an important regulator of actomyosin contractility in cancer cells , does not play a role in collective SCC migration, Cdc42 finely regulates the actomyosin-dependent remodeling of the ECM by CAFs . PDX models developed in verapamil- or diltiazem-treated mice show reduced levels of phosphorylated MLC2 and a decrease in the number of filopodia, which regulate tumor cell invasion (Fig. b). More recently, Gao et al. have demonstrated that CAFs associated with HGSOC contribute to the formation of heterotypic spheroids in the malignant ascites . Those CAFs-centered spheroids recruits floating ovarian cancer cells, resulting in the formation of ‘metastatic units’ at early stages of transcoelomic metastasis . Mechanistically, floating ovarian cancer cells drive the production and secretion of EGF by CAFs located at the center region of the spheroids. This consequently promotes ITGA5 (integrin α5) expression in tumor cells, which in turn further enhances the tumor-stromal interaction inside the heterotypic spheroids . That is why EGF is expected to the promising therapeutic target to prevent the peritoneal dissemination of HGSOC. Indeed, it has been shown that a neutralizing anti-EGF antibody can suppress the formation of spheroid in ascites mediated by the attenuated expression of ITGA5 in floating ovarian cancer cells, leading to the prolonged survival period . Signal activation due to the Wnt family of secreted glycolipoproteins is one of the crucial machineries underlying the cellular polarity, proliferation, and cell fate determination during the embryonic development and tissue homeostasis . In the absence of Wnt ligand, cytoplasmic β-catenin is degraded constantly by Axin complex, which is composed of Axin, APC, casein kinase 1 (CK1), and glycogen synthase kinase 3 (GSK3) . CD44, c-Myc, Axin2, and Cyclin D1 are the typical target molecules regulated by the nuclear translocation of β-catenin. Interestingly, reactive oxygen species (ROS) activates canonical β-catenin-dependent Wnt signal pathway , which is responsible for up-regulation of c-Myc at the invasive front enriched in cancer stem-like cells . Ferrari et al. recently reported that Dickkopf (DKK)-3, which activates the canonical Wnt signaling pathway, is highly expressed in CAFs in breast, colon, and ovarian cancer stroma . DKK1 expression is downregulated in fibroblasts in pachydermoperiostosis (PDP), a rare chronic inflammatory disease characterized by unique skin and bone phenotypes associated with loss-of-function mutation of the HPGD gene, thereby increasing the proliferation capacity of PDP-associated fibroblasts . In contrast to DKK1, DKK2, and DKK4, which suppress Wnt/β-catenin signal transduction, DKK3 does not interact with LDL-receptor-related protein (LRP) 5/6 and therefore cannot fulfill the bona fide antagonistic role of the DKK family in the canonical Wnt signaling pathway . Instead, DKK3 decreases the stability of Kremen, a Wnt negative regulator, resulting in increased LRP6 membrane localization, which in turn stabilizes both β-catenin and Yes-associated protein (YAP)/transcriptional coactivator with PDZ-binding motif (TAZ) levels . Of note, heat-shock factor 1 (HSF1) interacts with the enhancer and promoter regions of the Dkk3 locus, and contributes to the upregulation of DKK3 in CAFs. HSF1 promotes tumor growth by inducing cytokines such as CXCL12 and TGF-β . Although β-catenin-mediated Wnt signaling is dispensable for the function of CAFs in remodeling the ECM and promoting tumor cell proliferation and invasion, DKK3-driven YAP activation is necessary to induce a tumor-promoting phenotype . Absence of DKK3 in DKK3-null normal fibroblasts and CAFs is associated with decreased YAP/TAZ and β-catenin activity. By contrast, depletion of DKK3 leads to the concomitant upregulation of Kremen, LRP6 inactivation, and destabilization of both β-catenin and YAP/TAZ in CAFs. This DKK3-mediated localization and stabilization of YAP/TAZ in the nucleus is independent from the Hippo pathway, in which phosphorylated YAP (Ser127) plays a central role . Thus, DKK3 is expected to mediate the crosstalk between Wnt/β-catenin signaling and YAP/TAZ. Periostin, which is abundantly produced and secreted by CAFs in HNSCC, promotes the CSC phenotype via the canonical Wnt/β-catenin signaling pathway . Periostin is highly expressed in the tumor stroma compared with cancer cells and promotes tumor progression and metastasis in HNSCC . Yu et al. showed that periostin secreted by CAFs is a potential ligand for protein tyrosine kinase 7 (PTK7), which is frequently upregulated in HNSCC tissues and is correlated with Wnt/β-catenin pathway activation and poor clinical outcome in HNSCC patients . CAF-derived periostin is highly likely to bind to PTK7 on the cancer cell membrane and transduces signals to disheveled protein through the cell surface receptor LRP6; this induces the phosphorylation of GSK-3β and the hypophosphorylation of β-catenin, which causes β-catenin to translocate into the nucleus, suggesting that the periostin-PTK7 axis activates the canonical Wnt signaling pathway . The periostin-PTK7 axis promotes tumorigenesis, lung metastasis, and chemoresistance mediated by β-catenin expression in HNSCC. Thus, treatment with a PTK7 neutralizing antibody increases the therapeutic efficacy of erlotinib, a small-molecule TKI effective for the treatment of metastatic and/or recurrent HNSCC, by downregulating β-catenin . Hippo signal pathway is an evolutionary well-conserved regulator of organ size, which is first discovered in Drosophila . Central to this signaling is a kinase cascade leading from the tumor suppressor Hippo (Mst1/ Mst2 in mammals) to the oncogenic Yki (YAP/TAZ in mammals), which is a transcriptional coactivator of target genes involved in cell proliferation and survival . The major target transcription factors regulated by YAP and TAZ are the four proteins of the TEA-domain-containing (TEAD) family (TEAD1-TEAD4). While Mst1/2 is downregulated in several kinds of carcinomas, TAZ has been reported to be upregulated in invasive breast cancer . The Hippo pathway is activated by stromal stiffness in solid tumor tissues, and a growing body of evidence suggests that the transcriptional factor YAP is activated in CAFs . YAP/TAZ is activated in response to mechanical stress and perturbation of the actin cytoskeleton . YAP/TAZ activation by mechanical stimuli in cells is influenced by Rho-GTPase, Rho-associated protein kinase (ROCK), and the integrity of the actomyosin cytoskeleton in a manner largely independent from the large tumor suppressor kinase . Pathohistological analysis of normal murine mammary tissues and PyMT-driven breast tumors shows nuclear accumulation of YAP in the stroma of both adenoma and carcinoma lesions . YAP activation in the stroma is further enhanced in the peripheral tumor regions of advanced carcinomas such as breast cancer and squamous cell carcinoma. YAP controls the expression of several cytoskeletal regulators including ANLN, connective tissue growth factor (CTGF), and diaphanous homolog 3 (DIAPH3), and then regulates the expression levels of MYL9/myosin light chain (MLC)-2. Matrix stiffening promotes YAP activation, thereby establishing a positive-feedback loop that helps maintain the CAF phenotype . Increased interstitial fluid pressure (IFP) blocks the delivery of therapeutic agents, whereas reduced tumor IFP improves the uptake of chemotherapeutic drugs . Therefore, lowering tumor interstitial hypertension, which acts as a barrier against tumor transvascular transport, has been proposed as a general strategy to increase tumor uptake as well as the therapeutic effects of anticancer drugs. Blocking the tyrosine kinase PDGFRβ increases the susceptibility to conventional chemotherapy in a xenograft model of anaplastic thyroid carcinoma . YAP activation serves as an independent prognostic marker for the overall survival of PDAC patients and its association with liver metastasis . Hyaluronic acid (HA) is the major determinant of elevated IFP in PDAC. In addition, the presence of α-SMA-positive myofibroblasts increases the number of fibrotic foci and promotes contraction of the interstitial space . The increased number of blood vessels together with increased hydraulic conductivity or the relative ease with which fluid moves across the vessel wall is responsible for an irregular and increased influx of fluid into the tumor stroma. Increased IFP is frequently reported in solid tumors such as breast carcinoma, glioblastoma, and malignant melanoma . Enzymatic degradation of HA results in the rapid reduction of IFP and the appearance of widely patent functioning vessels in the tumor microenvironment . Removing these barriers permits high concentrations of chemotherapy agents to reach PDAC tissues, which improves survival and reveals an unappreciated sensitivity of the disease to conventional cytotoxic agents (Fig. ). In the clinical setting, the combination of gemcitabine and PEGPH20 has attracted attention for the treatment of stage IV PDAC because of its effect on HA degradation in the tumor stroma . TGF-β signal pathway contributes to the maintenance of tissue homeostasis and prevention of incipient malignancy by regulating not only cellular adhesion, differentiation, proliferation and survival, but also the microenvironment . Injured epithelial tissue is gradually repaired by the formation of granulation tissues composed of α-SMA-positive myofibroblasts, macrophages, platelets, newly formed blood vessels and ECM (Fig. ). Pathological forms of TGF-β signal pathway drive tumor growth and invasive phenotype, evasion of immune surveillance, and distant metastasis including cancer cell dissemination . As with the wound-healing process, tumor-derived TGF-β is likely to recruit other stromal cell types characterized by CAFs and osteoclasts, which are enriched at the invasive front and at the bone metastatic disease, respectively. The dependence of myofibroblastic CAFs on autocrine TGF-β signaling remained unclear until a study demonstrated that the establishment of self-sustaining CXCL12 and TGF-β autocrine signaling pathways results in the formation of tumor-promoting CAFs during breast cancer progression . The two autocrine signaling pathways triggered by CXCL12 and TGF-β may promote CAF differentiation associated with increased α-SMA expression levels . TGF-β and CXCL12 secreted by cancer cells upregulate CXCR4 and stabilize the Smad-dependent TGF-β pathway. Loss of Cav-1 in the tumor stroma activates TGF-β signaling in CAFs . Activation of the TGF-β ligand by proteolytic cleavage promotes its interaction with specific receptors. TGF-β binds to TGF-β receptor type II, and promotes the formation of a hetero-oligomeric complex with TGF-β receptor type I, resulting in the activation of the TGF-β receptor kinase. The TGF-β receptor in turn phosphorylates serine/threonine residues in downstream target effectors such as Smad proteins. The activated TGF-β receptor complex initiates several downstream cascades, including the canonical Smad2/3 signaling pathway and non-canonical pathways, such as TGF-β-activated kinase-mediated p38- or JNK-signaling . Activation of TGF-β signaling leads to EMT in cancer cells, which express high levels of matrix metalloproteinase (MMP) . Discoidin domain receptor 2 upregulates MMP2 and MT1-MMP in a manner dependent on the ERK2/SNAIL1 axis in hepatocellular carcinoma . Ligand-initiated activation of the Smad-independent TGF-β pathway triggers GSK-3β phosphorylation by c-Abl and PKC-δ non-receptor kinases. Phosphorylation of GSK-3β at serine 9 (Ser9) causes its inhibition, which in turn allows Snail1 to enter the nucleus. Nuclear accumulation of Snail1 leads to acquisition of the myofibroblastic phenotype with stimulation of α-SMA and type I collagen instead of VE-cadherin. Inhibition of c-Abl activity with imatinib allows GSK-3β to phosphorylate Snail1, which targets it for proteasomal degradation and effectively abolishes the acquisition of the myofibroblastic phenotype and the fibrotic response. Rottlerin and imatinib abrogate EndoMT by inhibiting PKC-δ and c-Abl, respectively (Fig. ). Mammary fibroblasts in which Cav-1 is depleted show CAF phenotypes, such as the conversion of fibroblasts to myofibroblasts and enhanced TGF-β signaling . Cav-1 directly inhibits TGF-β signaling. Mechanistically, Cav-1 interacts with the TGF-β receptor type 1, inducing its degradation, and suppresses TGF-β-dependent Smad2 phosphorylation and nuclear translocation . Activation of the TGF-β signaling pathway is sufficient to downregulate Cav-1 expression , thereby forming a positive-feedback loop involving Cav-1 expression levels and TGF-β signaling activity. Significant downregulation of stromal Cav-1 is responsible for the metabolic reprogramming of CAFs, which is characterized by the induction of aerobic glycolysis (also referred to as “reverse Warburg effect”) and autophagy in the tumor-associated stroma. This results in the stromal production of energy-rich metabolites including L-lactate, pyruvate, and ketone bodies, as well as chemical building blocks such as amino acids (glutamine), nucleotides, and fatty acids . These recycled nutrients are then transferred to adjacent epithelial tumor cells, thereby fueling cancer progression in a paracrine fashion. Because activation of TGF-β signaling attenuates mitochondrial metabolism, and enhances aerobic glycolysis and autophagy (especially mitophagy, in which old dysfunctional mitochondria undergo degradation) , CAFs that secrete TGF-β ligands in an autocrine manner can proliferate independently of angiogenesis. Cancer cell-induced ROS promote the loss of stromal Cav-1 in fibroblasts via autophagy and activate hypoxia-inducible factor α (HIF1-α) under ROS-induced pseudohypoxic conditions . This phenomenon is termed metabolic symbiosis . Enhanced expression levels of MCT4 and BNIP3 in CAFs are responsible for the activation of aerobic glycolysis via metabolic symbiosis and mitophagy, respectively . JAKs are non-receptor tyrosine kinases mediating signal transduction which is involved in cellular proliferation and survival. The seven mammalian STAT family contain the tyrosine residue near the C-terminus which is phosphorylated by JAK family in the presence of growth factors, interleukins and interferons (IFN) . This phosphorylation allows STATs to form the dimer via the interaction with a conserved SH2 domain. Remarkably, there are several cytokines with distinct, and sometimes opposing, functions are likely to activate the same STAT protein . For a typical instance, IL-6, a pro-inflammatory cytokine which utilizes gp130, promotes the activation of STAT3. In contrast, IL-10, which is a potent anti-inflammatory cytokine, does not utilize gp130 but promotes STAT3 phosphorylation. Actomyosin contractility plays a key role in tumor cell migration, affecting both the tumor cells themselves and the remodeling of the ECM by tumor fibroblasts to permit cell migration . CAFs remodel the ECM using contractile forces and proteolytic activity, thereby generating tracks for the migration of tumor cells as collective strands led by a fibroblast . Force-mediated matrix remodeling largely depends on integrins α3 and α5, as well as Rho-mediated regulation of MLC activity in fibroblasts. However, these factors are not required in cancer cells. Instead, tumor cells depend on Cdc42 and myotonic dystrophy kinase-related Cdc42-binding protein kinase (MRCK)-mediated regulation of MLC to follow the tracks generated by fibroblasts in the ECM. Force-mediated matrix remodeling by CAFs depends on actomyosin contractility modulated by the ROCK signaling pathway . Rab21-positive vesicles preferentially localize to the areas of cell contraction, and both integrin α5 and Rab21 are required for MLC phosphorylation . Rab21 delivers integrin α5 to the cellular membrane, where it signals to the contractile machinery. At least three Rab proteins, including 5a, 11b, and 21 subtypes, are needed in CAFs for their ability to promote SCC invasion. Depletion of these Rab proteins does not affect the ability of SCC cells to invade the ECM previously remodeled by CAFs. This can be attributed to the fact that the ‘following’ SCC cells do not remodel the matrix, and matrix remodeling pathways are therefore dispensable in SCC cells. Cytokine signaling through GP130-IL6ST and JAK1 stimulates actomyosin-mediated contractility in cancer cells and in the tumor-associated stroma . GP130-IL6ST signaling to JAK1 drives actomyosin-mediated contractility in CAFs and promotes matrix remodeling. JAK1 signaling regulates actomyosin contractility by regulating the level of phosphorylated-MLC2 in both melanoma cells and CAFs, the latter of which lead SCC invasion. Pro-inflammatory cytokines, such as IL-6 and LIF, are aberrantly expressed by CAFs in the tumor microenvironment and induce chemoresistance as well as EMT . The axis involving the cytokine oncostatin M (OSM) acts through GP130-IL6ST, JAK1, and ROCK to drive actomyosin contractility and matrix remodeling by CAFs for the collective invasion of SCCs . OSM induces fibrotic changes in the lungs and liver, and promotes EMT and the myofibroblastic phenotype via the JAK/STAT axis, thereby predisposing to cancer development . JAK/STAT signaling may involve a Rho guanine nucleotide exchange factor, ARHGEF1, to activate RhoA to the GTP-bound state as in vascular smooth muscle cells. This is supported by data showing that basal RhoA activity in CAFs is sensitive to inhibition of JAK, and OSM activates RhoA in a JAK-dependent manner . Unlike melanoma cells, in which GP130-IL6ST/JAK1-ROCK signaling is required for cancer cell migration, this signaling pathway is not necessary in tumor cells, whereas it is required in CAFs for ECM remodeling leading to the collective invasion of SCCs . Therapeutic agents, including blocking antibodies against cytokines, such as IL-6, or small molecule inhibitors of JAK kinase or STAT activity, could be useful agents to block invasion and metastasis in malignant diseases. IL-6 receptor blockage inhibits lung metastasis of breast cancer cells by suppressing IL-6-induced JAK/STAT signaling . Furthermore, an anti-IL-6 neutralizing antibody named siltuximab inhibits non-small cell lung cancer progression . In an analysis of epigenetic alterations, Albrengues et al. demonstrated that aberrant DNA methylation maintains the phenotype of tumor-promoting CAFs via the JAK/STAT cascade . JAK1/STAT3 signaling is constitutively activated in CAFs, partly because STAT3 acetylation induced by CBP/p300 leads to the epigenetic-dependent loss of expression of Src homology phosphatase-1 (SHP-1), which is a negative regulator of JAK/STAT signaling. SHP-1, also known as tyrosine-protein phosphatase non-receptor type 6, dephosphorylates several tyrosine kinases including JAK1 . Silencing of SHP-1 by DNA methyltransferase 1-mediated promoter hypermethylation leads to the sustained constitutive phosphorylation of JAK1 kinase and the STAT3 transcription factor, which maintain the contractile and pro-invasive properties of CAFs . Pharmacological inhibition with both 5-azacytidine and ruxolitinib results in the long-term abrogation of JAK1/STAT3 phosphorylation and rescues the expression of SHP-1, thereby inhibiting the tumor-promoting invasive phenotype of CAFs. Given that genetic mutations are rare in CAFs, further investigations are warranted to identify epigenetic abnormalities in the cancer-associated stroma . CAFs contribute to the formation and maintenance of the tumor microenvironment in cooperation with tumor cells by activating several signaling cascades including the EGFR, JAK/STAT, TGF-β, and Wnt pathways. In addition, stromal stiffness leads to drug resistance and poor prognosis of cancer patients. Given that α-SMA-positive activated fibroblasts form a senescence-associated secretory phenotype loop in response to treatment with HDAC inhibitors , re-education of the tumor stroma could be a promising therapeutic strategy. Treatment with chemotherapy and/or radiotherapy eradicates responsive diseases. However, survival of CAFs is associated with minimal residual disease. The surviving CAFs acquire innate and adaptive resistance to therapy, which is accompanied by stromal inflammation and increased ECM accumulation, leading to iatrogenic tumor stiffness and the development of chemoresistant tumors . Hirata et al. indicated that CAFs associated with BRAF -mutant malignant melanoma are activated in response to PLX4720, a selective BRAF inhibitor. PLX4720 paradoxically activates ERK/MAPK signaling in residual disease, promotes collagen production and matrix remodeling, and promotes MLC phosphorylation . This iatrogenic activation of CAFs is responsible for the FAK-dependent persistent survival of melanoma cells. The ability of melanoma-associated fibroblasts to confer PLX4720 tolerance largely depends on both FAK and integrin β1 in melanoma cells. Furthermore, stiffness of the fibronectin-rich stroma is sufficient to abrogate the effects of BRAF inhibition. PDX models indicate that dual inhibition of BRAF and FAK inhibits ERK/MAPK re-activation in the tumor stroma, which facilitates the efficient therapeutic control of BRAF -mutant melanoma . As such, the tumor microenvironment mainly composed of CAFs determines the dynamic phenotype and plasticity of cancer cells in cooperation with intrinsic genetic/epigenetic alterations . The degree of cancer cell differentiation may be largely controlled by the “stromagenic switch”, which results in CAF heterogeneity. In addition, α-SMA-negative and PDGFRβ-positive CAF subpopulations contribute to the malignant potential of tumor cells by interacting with integrin α11 . Of note, α-SMA-negative inflammatory CAFs secrete high levels of pro-inflammatory cytokines such as IL-6, IL-11, and LIF, and activate the JAK1/STAT3 cascade . In verity, several molecular machineries underlying invasive/metastatic phenotype and therapy-resistance driven by CAFs have been uncovered. Surprisingly enough, there exist tumor-restricting CAF populations which have been shown to inhibit tumor growth and progression . Accumulating evidence demonstrates that activation of Hh signal pathway in CAFs suppresses the growth of tumors mediated by bone morphogenetic protein (BMP) signaling in cancer cells, which strongly suggests the presence of CAF populations with tumor-suppressive functions. Taken together, the existence of several potential CAF markers suggests that further investigation is warranted to identify the pathophysiological functions of these molecules. |
Impact of COVID-19 lockdown on a tertiary center pediatric otolaryngology emergency department | 7b8c8153-e301-4dcc-af2f-6e5470cd8fa9 | 8425462 | Otolaryngology[mh] | At midday, March 17, 2020, the French government introduced a strict lockdown in response to the pandemic implicating the SARS-CoV-2. This unprecedented measure made necessary by the rapid spread of the virus and lack of evidence-based guidance, which lasted until May 11, 2020 . The health system was obliged to react. Hospital care was reorganized to limit the viral spread and to channel personnel into dedicated COVID-19 units. In pediatric otolaryngology, as in all other specialties, guidelines restricted activity to emergencies [ – ]. Thus, emergency activity was maintained but many departments found a sharp reduction in volume [ – ]. Guidelines also modified the management of many pathologies (for example in the ENT clinical practice: limitation of the use of flexible nasal endoscopy, limitation of surgical drainage for infectious disease, or limitation of airway endoscopies for suspected foreign bodies aspiration) . In our region, which includes the city of Paris, with more than 12 million inhabitants ( www.insee.fr ), our pediatric center continued its mission as a regional-level pediatric otolaryngology emergency department, and during the daytime, moreover, admitted emergencies from other centers in which pediatric activity were restricted by the epidemic. The aim of the present study was to objectively assess the impact of the COVID-19 lockdown on pediatric ENT emergency care in a tertiary center.
A single-center retrospective study was performed in a tertiary center providing pediatric otolaryngology emergency care for the Ile-de-France Region around the capital, Paris. All children managed for otolaryngology emergencies during lockdown (March 7 to May 11, 2020), and during the equivalent periods in 2018 and 2019 were included. For each inclusion period, the following data were collected: number of emergency pediatric otolaryngology consultations: • weekdays (Monday to Friday, 8 am to 6.30 pm) • evenings and weekends (duty) number of emergency procedures type of surgery: • infection: cervical abscess, para- or retro-pharyngeal abscess, peritonsillar abscess, mastoiditis, complicated sinusitis, other • endoscopy: esophageal foreign body, foreign body aspiration, laryngomalacia, other (e.g., subglottic stenosis, vallecular cyst, ingestion of caustic substance) • post-tonsillectomy scar hemorrhage results of systematic COVID-19 screening in operated patients, on RT-PCR of nasopharyngeal swab Data were collected using the department’s coding software and databases. Data were analyzed on Excel® software (Microsoft, Redmond, WA, USA). The periods March 17 to May 11, 2018 and 2019, were taken as reference, and 2020 values were compared against the average of the two. Statistical analysis used XLSTAT software® (Addinsoft, New York, NY, USA). Surgical procedures as a proportion of consultations and types of surgery as proportions of overall surgery were compared between years on chi 2 or Fisher test. Differences were considered significant if p < 0.05.
Emergency pediatric otolaryngology consultations In the 2020 lockdown period, 350 children were taken in charge in the emergency department: 136 (39%) during weekday hours and 214 in duty periods, compared to 761, 297 (39%), and 464, respectively, on average in 2018–2019. In 2020, there was thus a 54% decrease in emergency consultation, while proportions for weekdays and duty periods were unchanged (Table ). Emergency surgical procedures In 2020, 62 children underwent an emergency surgical procedure, compared to an average 93 for 2018–2019, i.e., a decrease of 33%. Only 2 of the 62 children (3%) had positive COVID-19 screening. During the lockdown, rates of surgery for infection of whatever type decreased: 10 procedures in 2020, compared to an average of 31 for 2018–2019 (68%). There were no cases of post-tonsillectomy scar hemorrhage during the lockdown, compared to an average of 4 previously. Endoscopy rates were stable for 2018, 2019, and 2020: respectively, 51, 66, and 52 (Table , Fig. ). The percentage of surgical procedures after consultation was significantly higher in 2020: 18%, versus 12% in 2018–2019 ( p = 0.014). In 2020, the types of emergency surgery procedures performed as proportions of overall surgeries were significantly modified in comparison to 2018–2019: decrease of procedures for infectious diseases and increase of endoscopic airway surgeries. Indeed, 16% of emergency procedures in 2020 were for infection, versus 33% previously ( p = 0.017) (Fig. ). Eighty-four percent of procedures were endoscopic airway surgery, versus 63% previously ( p = 0.006) (Fig. ): 27% for suspected tracheobronchial or esophageal foreign body, compared to 66% previously, and 73% for laryngeal or tracheal pathology, compared to 34% previously ( p < 0.0001) (Fig. ). Emergency endoscopic surgery for laryngeal or tracheal pathology in 2020 is detailed in Table .
In the 2020 lockdown period, 350 children were taken in charge in the emergency department: 136 (39%) during weekday hours and 214 in duty periods, compared to 761, 297 (39%), and 464, respectively, on average in 2018–2019. In 2020, there was thus a 54% decrease in emergency consultation, while proportions for weekdays and duty periods were unchanged (Table ).
In 2020, 62 children underwent an emergency surgical procedure, compared to an average 93 for 2018–2019, i.e., a decrease of 33%. Only 2 of the 62 children (3%) had positive COVID-19 screening. During the lockdown, rates of surgery for infection of whatever type decreased: 10 procedures in 2020, compared to an average of 31 for 2018–2019 (68%). There were no cases of post-tonsillectomy scar hemorrhage during the lockdown, compared to an average of 4 previously. Endoscopy rates were stable for 2018, 2019, and 2020: respectively, 51, 66, and 52 (Table , Fig. ). The percentage of surgical procedures after consultation was significantly higher in 2020: 18%, versus 12% in 2018–2019 ( p = 0.014). In 2020, the types of emergency surgery procedures performed as proportions of overall surgeries were significantly modified in comparison to 2018–2019: decrease of procedures for infectious diseases and increase of endoscopic airway surgeries. Indeed, 16% of emergency procedures in 2020 were for infection, versus 33% previously ( p = 0.017) (Fig. ). Eighty-four percent of procedures were endoscopic airway surgery, versus 63% previously ( p = 0.006) (Fig. ): 27% for suspected tracheobronchial or esophageal foreign body, compared to 66% previously, and 73% for laryngeal or tracheal pathology, compared to 34% previously ( p < 0.0001) (Fig. ). Emergency endoscopic surgery for laryngeal or tracheal pathology in 2020 is detailed in Table .
The present study reports the impact of the COVID-19 lockdown on pediatric otolaryngology emergency activity in a tertiary center. Several articles studied the evolution of ENT surgical emergencies during COVID-19 lockdown, but no previous one focused on the pediatric ENT particularity. Our study is the first to evaluate the effect of COVID-19 lockdown on pediatric ENT emergencies. General emergency activity Over the lockdown period of March 17 to May 11, 2020, our center continued to serve as the on-call pediatric ENT department in the evenings and at weekends for the entire region (population of 12 million people). It also handled daytime emergencies for other centers that were focusing on COVID-19 + adults. We had been expecting a surge of regionally referred emergency cases that could not be performed elsewhere . In fact, like other teams, we experienced a drop [ – , , ]. Emergency activity was much lower than for the equivalent periods in 2018 and 2019: 54% fewer consultations and 33% fewer surgeries. In Italy, Ralli et al. reported a 70% drop in head and neck abscess surgery and 50% in endoscopy in their university ENT department compared with the previous year . Lazzerini et al. reported a 73–88% decrease in pediatric emergencies . Several factors may account for this decrease. First, the population of Paris fell by 27% during lockdown according to the official statistics of INSEE ( www.insee.fr ), obviously leading to fewer emergencies; however, the fall in activity we observed exceeded the decrease in population. Secondly, public awareness of the pressure on the hospital system, and the recommendations formulated by the French Society of Otolaryngology (SFORL) in the form of practical guidance notes for patients, reduced the number of unjustified emergency consultations . Finally, the epidemic also had the psychological effect of increasing public anxiety . This stress, partly due to a fear of contamination, reduced the use of emergency services even for children, even though they are much less at risk than adults of developing severe SARS-CoV-2 infection . Emergency infection surgery There was a sharp fall in the rate of emergency infection surgery: 16% down from 33% ( p = 0.017). Lockdown was intended to reduce SARS-CoV-2 transmission by limiting physical interaction and likely had the positive collateral effect of reducing transmission of other frequent pediatric viral infections. Information to parents and to children concerning hygiene and social distancing doubtless impacted transmission of infection in general. Closure of schools and the physical distancing rules was related to a reduction in the prevalence of otitis media with effusion and fostered the resolution of its chronic forms . However, delayed consultations may explain why the relative rate of surgery was significantly greater during lockdown than in the previous years: 18% vs. 12% ( p = 0.014). Emergency endoscopic procedures The number of endoscopic procedures was stable during 2018, 2019, and 2020. Nevertheless, the proportion of emergency endoscopic procedures greatly increased during the lockdown: 84% of emergency surgery as a whole, versus 63% previously ( p = 0.006). This is mainly due to a proportional decrease in another type of emergencies in our department, such as infectious diseases. Emergency laryngotracheal endoscopic procedures for pathologies other than foreign body increased significantly ( p < 0.0001), showing that these emergencies are not dependent on the pandemic situation. This was due to first-line emergency services being overrun by COVID-19 cases, and systematic resort to a tertiary center in case of life-threatening emergencies. The huge increase of endoscopies for laryngeal or tracheal pathology is explained by the regional reorganization of care secondary to the epidemic. Indeed, our center kept its availability for these diseases while other pediatric centers closed their pediatric and neonatal intensive care units redirecting caregivers towards COVID-19-dedicated units. In addition, our center had facilitated access to the operating room for emergency endoscopies due to the cancellation of all elective surgeries. In contrast, the rate of endoscopy for suspected airway or esophageal foreign body fell to 27% from 66% ( p < 0.0001). Although children were spending more time at home, increased parental vigilance may partly account for this. Moreover, in line with recommendations made by the SFORL and the French Association of Pediatric Otolaryngology (AFOP), in order to reduce indications for head and neck endoscopy, which generates aerosols and thus increases the risk of SARS-CoV-2 transmission, we modified our diagnostic strategy for suspected foreign body inhalation, using thoracic CT as first-line examination . Emergency for post-tonsillectomy hemorrhage The fall in hemostasis procedures for post-tonsillectomy hemorrhage was related to the sharp reduction in the number of tonsillectomies during the lockdown, again in line with SFORL-AFOP recommendations .
Over the lockdown period of March 17 to May 11, 2020, our center continued to serve as the on-call pediatric ENT department in the evenings and at weekends for the entire region (population of 12 million people). It also handled daytime emergencies for other centers that were focusing on COVID-19 + adults. We had been expecting a surge of regionally referred emergency cases that could not be performed elsewhere . In fact, like other teams, we experienced a drop [ – , , ]. Emergency activity was much lower than for the equivalent periods in 2018 and 2019: 54% fewer consultations and 33% fewer surgeries. In Italy, Ralli et al. reported a 70% drop in head and neck abscess surgery and 50% in endoscopy in their university ENT department compared with the previous year . Lazzerini et al. reported a 73–88% decrease in pediatric emergencies . Several factors may account for this decrease. First, the population of Paris fell by 27% during lockdown according to the official statistics of INSEE ( www.insee.fr ), obviously leading to fewer emergencies; however, the fall in activity we observed exceeded the decrease in population. Secondly, public awareness of the pressure on the hospital system, and the recommendations formulated by the French Society of Otolaryngology (SFORL) in the form of practical guidance notes for patients, reduced the number of unjustified emergency consultations . Finally, the epidemic also had the psychological effect of increasing public anxiety . This stress, partly due to a fear of contamination, reduced the use of emergency services even for children, even though they are much less at risk than adults of developing severe SARS-CoV-2 infection .
There was a sharp fall in the rate of emergency infection surgery: 16% down from 33% ( p = 0.017). Lockdown was intended to reduce SARS-CoV-2 transmission by limiting physical interaction and likely had the positive collateral effect of reducing transmission of other frequent pediatric viral infections. Information to parents and to children concerning hygiene and social distancing doubtless impacted transmission of infection in general. Closure of schools and the physical distancing rules was related to a reduction in the prevalence of otitis media with effusion and fostered the resolution of its chronic forms . However, delayed consultations may explain why the relative rate of surgery was significantly greater during lockdown than in the previous years: 18% vs. 12% ( p = 0.014).
The number of endoscopic procedures was stable during 2018, 2019, and 2020. Nevertheless, the proportion of emergency endoscopic procedures greatly increased during the lockdown: 84% of emergency surgery as a whole, versus 63% previously ( p = 0.006). This is mainly due to a proportional decrease in another type of emergencies in our department, such as infectious diseases. Emergency laryngotracheal endoscopic procedures for pathologies other than foreign body increased significantly ( p < 0.0001), showing that these emergencies are not dependent on the pandemic situation. This was due to first-line emergency services being overrun by COVID-19 cases, and systematic resort to a tertiary center in case of life-threatening emergencies. The huge increase of endoscopies for laryngeal or tracheal pathology is explained by the regional reorganization of care secondary to the epidemic. Indeed, our center kept its availability for these diseases while other pediatric centers closed their pediatric and neonatal intensive care units redirecting caregivers towards COVID-19-dedicated units. In addition, our center had facilitated access to the operating room for emergency endoscopies due to the cancellation of all elective surgeries. In contrast, the rate of endoscopy for suspected airway or esophageal foreign body fell to 27% from 66% ( p < 0.0001). Although children were spending more time at home, increased parental vigilance may partly account for this. Moreover, in line with recommendations made by the SFORL and the French Association of Pediatric Otolaryngology (AFOP), in order to reduce indications for head and neck endoscopy, which generates aerosols and thus increases the risk of SARS-CoV-2 transmission, we modified our diagnostic strategy for suspected foreign body inhalation, using thoracic CT as first-line examination .
The fall in hemostasis procedures for post-tonsillectomy hemorrhage was related to the sharp reduction in the number of tonsillectomies during the lockdown, again in line with SFORL-AFOP recommendations .
The lockdown in response to the SARS-CoV-2 pandemic impacted pediatric ENT emergency activity quantitatively and qualitatively. There was a 54% drop in consultations and 33% in surgeries compared to the corresponding period in 2018 and 2019. The proportion of emergency surgeries for infection decreased, while that of endoscopies increased. The behavioral changes imposed by the pandemic highlight the importance of prevention and parental vigilance in the incidence of infections and foreign body aspiration. This pandemic has also imposed a change in our clinical practices that could be prolonged after the pandemic, such as the indication of CT scan as the first-line examination in case of poor suspicion of foreign body aspiration instead of systematic endoscopic procedure, to avoid general anesthesia and its risks.
|
Machine learning validation of the AVAS classification compared to ultrasound mapping in a multicentre study | a5e3364d-64b0-4f9a-9bc3-af3db85e586d | 11756420 | Surgical Procedures, Operative[mh] | Rationale The rising number of patients with chronic and end-stage renal disease (ESRD) leads to increased demand for renal replacement therapy (RRT) . Kidney transplant is generally considered the best option in terms of patient’s survival and quality of life, however, there are still significant challenges associated with this option – . Firstly, not all patients are healthy enough to undergo surgery due to their comorbidities, and as a result, they are not placed on the waiting list. Secondly, a number of patients on the waiting list pass away before a suitable organ becomes available , . Lastly, even after a successful kidney transplant, complications can arise leading to graft loss, and if the recipient lives long enough all transplanted kidney grafts will eventually fail . Consequently, all the patients who may not be considered as adequate candidate for kidney transplant as well as those already needing RRT while on transplant list need dialysis. This can be achieved either through peritoneal dialysis (via an implanted peritoneal dialysis catheter) or haemodialysis (HD), via creation of vascular access (VA). Indeed, HD is the most prevalent method for RRT, with VA comprising 89% of all dialysis procedures and HD accounting for 69% of all renal replacement therapies . Dialysis VA is established either by a central venous catheter inserted into a central vein or by creating arteriovenous access (AVA) usually in the upper extremity. A functional AVA requires connection of a superficial vein to an artery, which causes the vein to dilate and strengthen, making it suitable for sustaining regular dialysis cannulations . This type of VA, where the vein is connected directly to the artery, is called a native arteriovenous fistula (AVF). If the superficial veins are inadequate, a prosthetic graft can also be placed between the artery and the vein, called an arteriovenous graft (AVG). The prosthetic graft positioned beneath the skin is used for needle cannulation during HD. It is well known, that native AVFs offer the best outcomes in terms of patency and complications, followed by AVGs, with central dialysis catheters being the least favourable option , . When treating ESRD patients in need of HD, it is essential to carefully plan the ideal VA type, aiming to provide a functional VA with minimal risk of complications, while also preserving the vascular tree to ensure future options for additional VA creations if needed , . Care of these patients requires a multidisciplinary approach of nephrologists, surgeons, radiologists, dialysis nurses and other health care professionals. Every patient should undergo a thorough physical examination and sonographic vascular mapping in the upper extremity (further called “mapping”) to evaluate individual vessels and determine the most appropriate type of VA based on their unique anatomical conditions. “Mapping” in the clinical practice is a variable process. It typically includes evaluation of arteries and veins, assessing their diameters, lumen content, depth, wall quality, according to some protocols also arterial flow (not involved in AVAS protocol). Additionally, Allen’s test, a clinical evaluation of arterial blood flow to the hand, should be routinely performed, especially when planning distal access. These parameters provide a detailed understanding of vascular status and guide decisions on possibilities for the type of VA creation. Details of the vascular parameters collected for this study are summarized in Table . Given the complexity of vascular anatomy that is often affected by previous VA creations and interventions, complex patient history and findings of ultrasound mapping, it is frequent for the vessel status to be described in lengthy paragraphs. To improve communication, a novel classification system called AVAS (Arteriovenous Access Stage) has been developed. This system provides concise information about the condition of vessels in the upper extremity, indicating what type of VA can be created. According to this scheme, patients may be categorized into three main groups: AVAS 1 has patients with vessels suitable for creation of native AVF. AVAS 2 involves patients with worse vessels, where native AVF cannot be created, but a prosthetic AVG can still be placed. AVAS 3 has patients with poor vessel anatomy, where neither a typical AVF nor AVG can be considered. For AVAS 3, central dialysis catheters or a special type of VA (e.g. VA on lower limb) may be considered . Also, AVAS 1 and AVAS 2 are further subdivided based on suitability of outflow veins and the location where the AVA can be placed. The locations are indicated by letters A, B, C and D. (Supplementary Figure S1-S3) This classification has proven to be clinically useable and was recently validated in a prospective international multicentre study. This study showed 100% applicability in clinical settings and high ability to predict the type of AVA created. When AVAS was followed by clinicians, early failure of VA was also significantly decreased . Objectives In the previous study, AVAS was tested in clinical settings across eleven centres, where it was consecutively applied during preoperative mapping. The study reported an 89% agreement between evaluators, highlighting its reliability and easy use in clinical practice. Although VAs were created independently of AVAS classifications, AVAS outperformed clinicians in predicting the success of the created VA. While clinical validation of the AVAS system was successful and showed excellent implementation across multiple centres, the performance of this classifier has yet to be examined by using machine learning (ML) methods . The use of ML into healthcare has transformed the landscape of clinical decision-making. In the context of dialysis VA, ML models have showed considerable promise in optimizing patient outcomes. Recent applications include using ML algorithms to predict the success of VA maturation, and to assess VA quality, combining clinical, demographic, and imaging data for superior predictive accuracy compared to traditional methods – . These advancements highlight the growing role of ML in addressing challenges associated with VA planning, such as identifying the most suitable access type and minimizing complications . We consider ML a suitable method for validating the predictive performance of the AVAS system, enabling a robust comparison with sonographic mapping. Our hypothesis is that AVAS classification system can perform as effectively as all mapping ultrasound measurements including clinical examination and Allen’s test in deciding the predicted type of VA procedure. While the AVAS offers a simple and standardized approach to VA planning, its predictive ability has not been validated by ML methods in comparison to mapping. Despite its advantages, mapping faces challenges in clinical practice. Its documentation is not standardized, with different centres employing varying protocols, and even specialists may not adhere to the same guidelines. Some clinicians bypass ultrasound mapping altogether, relying solely on clinical examinations. As a result, findings from mapping are often poorly communicated and lack consistency in clinical settings. Furthermore, the lengthy and detailed descriptions can be time-consuming and difficult to interpret, making them challenging to incorporate into clinical practice. This study aims to address the gap between the clinical usability of AVAS and the predictive ability of mapping. By applying ML methodologies, specifically the random forests algorithm, this study evaluates whether AVAS’s predictive performance can be enhanced while maintaining its standardized and straightforward approach. This approach seeks to determine the extent to which AVAS can serve as an alternative to ultrasound mapping. The streamlined nature of AVAS may enable quicker decision-making, improving workflow efficiency in clinical practice. Additionally, its predictive capability could assist in determining the most suitable type of VA and identifying early failures, potentially preserving critical vascular sites and enhancing patient outcomes. By categorizing patients systematically, AVAS could allow fair comparisons across patient populations and healthcare units, allowing the appropriate remuneration for the cost of care, while also supporting research and quality improvement efforts in VA management. These potential benefits support the importance of validating AVAS as a tool for preoperative vascular assessment. A comparison of the ML prediction ability of mapping-based versus AVAS system-based models will be performed. The target (outcome) will be predicted VA (pVA) and created VA (cVA). Prediction ability of models will be assessed with and without additional clinical risk factors and demographic parameters. We included these additional parameters to evaluate whether they influence the decision-making process for VA placement.
The rising number of patients with chronic and end-stage renal disease (ESRD) leads to increased demand for renal replacement therapy (RRT) . Kidney transplant is generally considered the best option in terms of patient’s survival and quality of life, however, there are still significant challenges associated with this option – . Firstly, not all patients are healthy enough to undergo surgery due to their comorbidities, and as a result, they are not placed on the waiting list. Secondly, a number of patients on the waiting list pass away before a suitable organ becomes available , . Lastly, even after a successful kidney transplant, complications can arise leading to graft loss, and if the recipient lives long enough all transplanted kidney grafts will eventually fail . Consequently, all the patients who may not be considered as adequate candidate for kidney transplant as well as those already needing RRT while on transplant list need dialysis. This can be achieved either through peritoneal dialysis (via an implanted peritoneal dialysis catheter) or haemodialysis (HD), via creation of vascular access (VA). Indeed, HD is the most prevalent method for RRT, with VA comprising 89% of all dialysis procedures and HD accounting for 69% of all renal replacement therapies . Dialysis VA is established either by a central venous catheter inserted into a central vein or by creating arteriovenous access (AVA) usually in the upper extremity. A functional AVA requires connection of a superficial vein to an artery, which causes the vein to dilate and strengthen, making it suitable for sustaining regular dialysis cannulations . This type of VA, where the vein is connected directly to the artery, is called a native arteriovenous fistula (AVF). If the superficial veins are inadequate, a prosthetic graft can also be placed between the artery and the vein, called an arteriovenous graft (AVG). The prosthetic graft positioned beneath the skin is used for needle cannulation during HD. It is well known, that native AVFs offer the best outcomes in terms of patency and complications, followed by AVGs, with central dialysis catheters being the least favourable option , . When treating ESRD patients in need of HD, it is essential to carefully plan the ideal VA type, aiming to provide a functional VA with minimal risk of complications, while also preserving the vascular tree to ensure future options for additional VA creations if needed , . Care of these patients requires a multidisciplinary approach of nephrologists, surgeons, radiologists, dialysis nurses and other health care professionals. Every patient should undergo a thorough physical examination and sonographic vascular mapping in the upper extremity (further called “mapping”) to evaluate individual vessels and determine the most appropriate type of VA based on their unique anatomical conditions. “Mapping” in the clinical practice is a variable process. It typically includes evaluation of arteries and veins, assessing their diameters, lumen content, depth, wall quality, according to some protocols also arterial flow (not involved in AVAS protocol). Additionally, Allen’s test, a clinical evaluation of arterial blood flow to the hand, should be routinely performed, especially when planning distal access. These parameters provide a detailed understanding of vascular status and guide decisions on possibilities for the type of VA creation. Details of the vascular parameters collected for this study are summarized in Table . Given the complexity of vascular anatomy that is often affected by previous VA creations and interventions, complex patient history and findings of ultrasound mapping, it is frequent for the vessel status to be described in lengthy paragraphs. To improve communication, a novel classification system called AVAS (Arteriovenous Access Stage) has been developed. This system provides concise information about the condition of vessels in the upper extremity, indicating what type of VA can be created. According to this scheme, patients may be categorized into three main groups: AVAS 1 has patients with vessels suitable for creation of native AVF. AVAS 2 involves patients with worse vessels, where native AVF cannot be created, but a prosthetic AVG can still be placed. AVAS 3 has patients with poor vessel anatomy, where neither a typical AVF nor AVG can be considered. For AVAS 3, central dialysis catheters or a special type of VA (e.g. VA on lower limb) may be considered . Also, AVAS 1 and AVAS 2 are further subdivided based on suitability of outflow veins and the location where the AVA can be placed. The locations are indicated by letters A, B, C and D. (Supplementary Figure S1-S3) This classification has proven to be clinically useable and was recently validated in a prospective international multicentre study. This study showed 100% applicability in clinical settings and high ability to predict the type of AVA created. When AVAS was followed by clinicians, early failure of VA was also significantly decreased .
In the previous study, AVAS was tested in clinical settings across eleven centres, where it was consecutively applied during preoperative mapping. The study reported an 89% agreement between evaluators, highlighting its reliability and easy use in clinical practice. Although VAs were created independently of AVAS classifications, AVAS outperformed clinicians in predicting the success of the created VA. While clinical validation of the AVAS system was successful and showed excellent implementation across multiple centres, the performance of this classifier has yet to be examined by using machine learning (ML) methods . The use of ML into healthcare has transformed the landscape of clinical decision-making. In the context of dialysis VA, ML models have showed considerable promise in optimizing patient outcomes. Recent applications include using ML algorithms to predict the success of VA maturation, and to assess VA quality, combining clinical, demographic, and imaging data for superior predictive accuracy compared to traditional methods – . These advancements highlight the growing role of ML in addressing challenges associated with VA planning, such as identifying the most suitable access type and minimizing complications . We consider ML a suitable method for validating the predictive performance of the AVAS system, enabling a robust comparison with sonographic mapping. Our hypothesis is that AVAS classification system can perform as effectively as all mapping ultrasound measurements including clinical examination and Allen’s test in deciding the predicted type of VA procedure. While the AVAS offers a simple and standardized approach to VA planning, its predictive ability has not been validated by ML methods in comparison to mapping. Despite its advantages, mapping faces challenges in clinical practice. Its documentation is not standardized, with different centres employing varying protocols, and even specialists may not adhere to the same guidelines. Some clinicians bypass ultrasound mapping altogether, relying solely on clinical examinations. As a result, findings from mapping are often poorly communicated and lack consistency in clinical settings. Furthermore, the lengthy and detailed descriptions can be time-consuming and difficult to interpret, making them challenging to incorporate into clinical practice. This study aims to address the gap between the clinical usability of AVAS and the predictive ability of mapping. By applying ML methodologies, specifically the random forests algorithm, this study evaluates whether AVAS’s predictive performance can be enhanced while maintaining its standardized and straightforward approach. This approach seeks to determine the extent to which AVAS can serve as an alternative to ultrasound mapping. The streamlined nature of AVAS may enable quicker decision-making, improving workflow efficiency in clinical practice. Additionally, its predictive capability could assist in determining the most suitable type of VA and identifying early failures, potentially preserving critical vascular sites and enhancing patient outcomes. By categorizing patients systematically, AVAS could allow fair comparisons across patient populations and healthcare units, allowing the appropriate remuneration for the cost of care, while also supporting research and quality improvement efforts in VA management. These potential benefits support the importance of validating AVAS as a tool for preoperative vascular assessment. A comparison of the ML prediction ability of mapping-based versus AVAS system-based models will be performed. The target (outcome) will be predicted VA (pVA) and created VA (cVA). Prediction ability of models will be assessed with and without additional clinical risk factors and demographic parameters. We included these additional parameters to evaluate whether they influence the decision-making process for VA placement.
Study design A prospective multicentre international study, named the VAVASC study (Validation of Arterio Venous Access Stage Classification), has been conducted, involving eleven centres from eight countries. The study has been registered in the clinical trials registry (NCT 04796558) . (See details in Supplementary Table S1). Settings The study commenced in March 2021 and patient recruitment was terminated on the 18 th of July 2024. Some patients are still under observation, with each being followed up for two years from the date of AVA creation. All the participating centres gained their own approvals for patient’s data collection (e.g. ethical and research governance, (EK-VP/06/0/2021). All methods were performed in accordance with the relevant guidelines and regulations. The anonymized data were entered into electronic database. Participants To minimise selection bias, all patients indicated for VA placement were systematically included, regardless of demographics, comorbidities, or AVAS classification. No minors were involved in this study, as the youngest participant was 18 years old. Therefore, informed consent from a parent or legal guardian was not required. The patients were recruited consecutively. The only exclusion criterion beside low age was the patient’s disapproval of their data being included in the study. Informed consent was obtained from all recruited patients or from their legal guardians. The prospective consecutive recruitment was ceased on 31 st of January 2024. Due to low numbers of patients in certain classes, additional recruitment was continued to accomplish sufficient sample size for ML validation. This special focus on rare AVAS classes (AVAS 1BC, 1ACD, 1AD, 1B, AVAS 2 and 3) through targeted extended recruitment aimed to gain adequate representation of these less common groups. Features (predictors) Main features All enrolled patients were classified according to AVAS scheme into one of the described classes . This assessment was performed by the lead researcher(s) at each participating centre and reviewed by the principal investigator (K. Lawrie) . Parameters of upper extremity vessels measured by ultrasound (referred as “vascular mapping”) and Allen’s test (= clinical examination of the patency of arterial palmar arch) were collected. The details are presented in Table . Additional features Basic demographic data and patient’s comorbidities were also collected. The details are also specified in Table . Targets (outcomes) Type of predicted vascular access A pVA is the type of VA that is recommended for the patient by VA specialists who conducted the examination. It is based on the clinical assessment and sonographic vascular mapping, independent of AVAS classification. The list of VA types follows the terminology from the Recommended standards for reports approved by the Committee on Reporting Standards of the Society for Vascular Surgery and the American Association for Vascular Surgery (Supplementary Table S2) . Additionally, endovascular access and proximal radial-cephalic direct access were also included for clinical purposes and to ensure consecutive recruitment. Type of created vascular access A cVA refers the type of VA, that was ultimately created by VA surgeon. It is also independent from AVAS class as it was based on the operating surgeon’s clinical judgement. The list of potential cVAs is identical to the list of pVAs as presented in Supplementary Table S2. Bias Efforts to reduce bias were implemented in the following manners: The study is multicentre and international. It was registered in the trials registry and the study protocol was published to secure its transparency . The data were collected prospectively by VA specialists. Additional recruitment of rare categories involved only patients with high quality data. Study sample Prior to the beginning of the study, we decided on a minimum number of 800 patients to be included in the study, with 640 patients in the training dataset and 160 patients in the validation dataset . This value was chosen without prior knowledge of the distribution of each AVAS and VA class. Recruitment exceeded this amount by 43.9%. Additionally, extra recruitment was conducted due to low number of patients in rare classes. This number of patients was sufficient to detect a difference in the ROC AUC (receiver operating characteristic area under curve) of the models greater than 2%. Statistical analysis R version 4.4.2 with RStudio 2024.09.1 were used for all data analysis , . Tidymodels 1.2.0 were used as ML framework . Exploratory data analysis and missing data analysis was performed for all parameters. Categorical parameters are presented as counts and percentages. Continuous parameters are displayed as either mean and standard deviation (SD) or median with 25 th and 75 th percentiles, based on their distributions. Despite extra recruitment, rare AVAS classes (with fewer than 9 patients, (< 2.5%)) did not yield sufficient number of patients for separate validation of the class. Due to their close clinical relevance, these classes were grouped into larger anatomically corresponding AVAS classes. This resulted into consolidation of 19 categories into 10. (See Supplementary Table S3). Types of pVAs and cVAs were also consolidated into seven classes based on their anatomical relevance. The reason was, again, rare incidence of certain types of VAs. (See Supplementary Table S4). The alternative options were to exclude rare classes, combine them all into a single group without considering anatomical relevance, or create miscellaneous categories. However, these alternative options would lead to clinical inaccuracies and incorrect validation. Initial data split The dataset was divided into training and a validation set with an 80/20 split. The stratification was based on the cVA category to ensure equal representation of cVA categories across both sets. The training dataset was used to develop all models, and the validation dataset was used to evaluate and compare performance of the models. Model development For development of all models, three algorithms were employed: multinomial logistic regression, random forest, and eXtreme Gradient Boosting (XGBoost, eXGB). Each of these algorithms was applied to the eight models described below, ensuring robust comparisons across techniques. Multinomial Logistic Regression The multinomial logistic regression models were developed using the multinom_reg function from the parsnip library within the tidymodels framework. Multinomial logistic regression is a statistical method used for predicting categorical outcomes with more than two categories. It models the probability of each category as a function of the predictors by estimating coefficients that maximize the likelihood of the observed data. The engine used for these models was glmnet . Hyperparameter tuning was performed for the following parameters: Penalty (Regularisation strength): Controls the degree of shrinkage applied to the model coefficients. Higher values enforce greater regularisation. Mixture (Proportion of L1 regularisation): Determines the balance between L1 (lasso) and L2 (ridge) regularisation. A value of 1 applies pure L1 regularisation, while a value of 0 applies pure L2 regularisation. Intermediate values represent a combination (elastic net regularisation). Random Forest The random forest models were implemented using the rand_forest function from the parsnip library within the tidymodels framework. Random forest is an ensemble learning method that constructs multiple decision trees during training and combines their outputs to improve prediction accuracy. It reduces overfitting and increases robustness by averaging the results of the trees, each built on random subsets of data and predictors. The engine used for random forest models was ranger . The following settings and hyperparameters were applied: Number of Trees (trees): Set empirically to 1000. Hyperparameter Tuning: o Minimum number of observations in a terminal node (min_n). o Number of predictors sampled for splitting at each node (mtry). eXtreme Gradient Boosting (XGBoost) The XGBoost models were developed using the boost_tree function from the parsnip library within the tidymodels framework. XGBoost is a powerful algorithm that builds an ensemble of decision trees in a sequential manner, using a gradient boosting framework where each new tree corrects the errors of the previous ones by optimizing a differentiable loss function. The engine used for these models was xgboost . The following configurations were applied: Number of Trees (trees): Set empirically to 100. Hyperparameter Tuning: o Minimum number of data points required in a node for it to be split further (n_min). o Maximum depth of the tree (tree_depth). o Learning Rate: Determines the step size at each iteration (learn_rate). o Sample Size: Proportion of the training data to sample for each tree (sample_size). Defined models Eight models were defined based on the predicted outcomes and the input features (the model abbreviation is in brackets): pVA predicted using mapping (Mapping pVA) pVA predicted using AVAS (AVAS pVA) pVA predicted using mapping and additional features (Mapping pVA param) pVA predicted using AVAS and additional features (AVAS pVA param) cVA predicted using mapping (Mapping cVA) cVA predicted using AVAS (AVAS cVA) cVA predicted using mapping and additional features (Mapping cVA param) cVA predicted using AVAS and additional features (AVAS cVA param) All 24 models (8 models across 3 algorithms) were evaluated using fivefold cross-validation, stratified by the pVA category to ensure equal representation of pVA categories across all folds. The main performance metric was ROC AUC, calculated using the method from Hand and Till (2001) for multiclass scenarios . The same cross-validation folds were used consistently across all algorithms to maintain comparability. Data preprocessing steps The same sequence of preprocessing steps (recipes) was used to create all models (library recipes 1.1.0): . Missing data imputation using a bagged tree algorithm to handle missing values. Categories with low frequency (below a 2.5% threshold) were collapsed. Categorical parameters were converted into dummy variables. Data upsampling was used to balance the occurrence of levels within the cVA and pVA. All models with preprocessing steps were developed in one workflow. Finalising models The performance of the models with tuned hyperparameters was assessed by using the validation dataset. Estimation and comparison of model performance (ROC AUC) was done by using Bayesian generalised linear models . ROC AUCs are expressed as mean with 95% credible intervals (95% CI). Differences between model performance are expressed as mean difference in ROC AUC with 95% CI. Differences with 95% CI that does not reach zero are supposed to be statistically significant. Three groups of model comparisons were performed (12 comparisons in total) for each of the three algorithms (multinomial logistic regression, random forest, and XGBoost): Predictive ability of mapping and AVAS (for both pVA and cVA) with and without additional features (demographic and clinical parameters), resulting in four comparisons. Effect of additional features on prediction of cVA and pVA by mapping and AVAS, resulting in four comparisons. How AVAS predicts pVA and cVA and what are the differences between these two predictions. How mapping predicts pVA and cVA and what are the differences between these two predictions (both with and without additional features), four comparisons. The algorithm with the best predictive performance was selected for the final presentation of results. Permutation-based variable importance was used to identify the additional features responsible for improving model performance (library vip) . To interpret the influence of the most important additional features on the prediction of cVA and pVA from models combining mapping and additional features (Mapping cVA param and Mapping pVA param models), we utilised Partial Dependence profiles (PDPs) and Accumulated Local Effects (ALE) profiles. PDPs were employed to visualise the marginal effect of key features on the model’s predictions. These plots depict the average predicted response as a given feature varies, while all other features are held constant . ALE profiles complemented PDPs by capturing local feature effects and accounting for potential interactions between features. Unlike PDPs, ALEs provide unbiased insights even in the presence of correlated features . Both PDPs and ALE profiles were generated using the DALEX (version 2.4.3) and DALEXtra (version 2.3.0) packages, leveraging the explain_tidymodels and model_profile functions. Visualisation was performed using the plot function within the same framework.
A prospective multicentre international study, named the VAVASC study (Validation of Arterio Venous Access Stage Classification), has been conducted, involving eleven centres from eight countries. The study has been registered in the clinical trials registry (NCT 04796558) . (See details in Supplementary Table S1).
The study commenced in March 2021 and patient recruitment was terminated on the 18 th of July 2024. Some patients are still under observation, with each being followed up for two years from the date of AVA creation. All the participating centres gained their own approvals for patient’s data collection (e.g. ethical and research governance, (EK-VP/06/0/2021). All methods were performed in accordance with the relevant guidelines and regulations. The anonymized data were entered into electronic database.
To minimise selection bias, all patients indicated for VA placement were systematically included, regardless of demographics, comorbidities, or AVAS classification. No minors were involved in this study, as the youngest participant was 18 years old. Therefore, informed consent from a parent or legal guardian was not required. The patients were recruited consecutively. The only exclusion criterion beside low age was the patient’s disapproval of their data being included in the study. Informed consent was obtained from all recruited patients or from their legal guardians. The prospective consecutive recruitment was ceased on 31 st of January 2024. Due to low numbers of patients in certain classes, additional recruitment was continued to accomplish sufficient sample size for ML validation. This special focus on rare AVAS classes (AVAS 1BC, 1ACD, 1AD, 1B, AVAS 2 and 3) through targeted extended recruitment aimed to gain adequate representation of these less common groups.
Main features All enrolled patients were classified according to AVAS scheme into one of the described classes . This assessment was performed by the lead researcher(s) at each participating centre and reviewed by the principal investigator (K. Lawrie) . Parameters of upper extremity vessels measured by ultrasound (referred as “vascular mapping”) and Allen’s test (= clinical examination of the patency of arterial palmar arch) were collected. The details are presented in Table . Additional features Basic demographic data and patient’s comorbidities were also collected. The details are also specified in Table .
All enrolled patients were classified according to AVAS scheme into one of the described classes . This assessment was performed by the lead researcher(s) at each participating centre and reviewed by the principal investigator (K. Lawrie) . Parameters of upper extremity vessels measured by ultrasound (referred as “vascular mapping”) and Allen’s test (= clinical examination of the patency of arterial palmar arch) were collected. The details are presented in Table .
Basic demographic data and patient’s comorbidities were also collected. The details are also specified in Table .
Type of predicted vascular access A pVA is the type of VA that is recommended for the patient by VA specialists who conducted the examination. It is based on the clinical assessment and sonographic vascular mapping, independent of AVAS classification. The list of VA types follows the terminology from the Recommended standards for reports approved by the Committee on Reporting Standards of the Society for Vascular Surgery and the American Association for Vascular Surgery (Supplementary Table S2) . Additionally, endovascular access and proximal radial-cephalic direct access were also included for clinical purposes and to ensure consecutive recruitment. Type of created vascular access A cVA refers the type of VA, that was ultimately created by VA surgeon. It is also independent from AVAS class as it was based on the operating surgeon’s clinical judgement. The list of potential cVAs is identical to the list of pVAs as presented in Supplementary Table S2.
A pVA is the type of VA that is recommended for the patient by VA specialists who conducted the examination. It is based on the clinical assessment and sonographic vascular mapping, independent of AVAS classification. The list of VA types follows the terminology from the Recommended standards for reports approved by the Committee on Reporting Standards of the Society for Vascular Surgery and the American Association for Vascular Surgery (Supplementary Table S2) . Additionally, endovascular access and proximal radial-cephalic direct access were also included for clinical purposes and to ensure consecutive recruitment.
A cVA refers the type of VA, that was ultimately created by VA surgeon. It is also independent from AVAS class as it was based on the operating surgeon’s clinical judgement. The list of potential cVAs is identical to the list of pVAs as presented in Supplementary Table S2.
Efforts to reduce bias were implemented in the following manners: The study is multicentre and international. It was registered in the trials registry and the study protocol was published to secure its transparency . The data were collected prospectively by VA specialists. Additional recruitment of rare categories involved only patients with high quality data.
Prior to the beginning of the study, we decided on a minimum number of 800 patients to be included in the study, with 640 patients in the training dataset and 160 patients in the validation dataset . This value was chosen without prior knowledge of the distribution of each AVAS and VA class. Recruitment exceeded this amount by 43.9%. Additionally, extra recruitment was conducted due to low number of patients in rare classes. This number of patients was sufficient to detect a difference in the ROC AUC (receiver operating characteristic area under curve) of the models greater than 2%.
R version 4.4.2 with RStudio 2024.09.1 were used for all data analysis , . Tidymodels 1.2.0 were used as ML framework . Exploratory data analysis and missing data analysis was performed for all parameters. Categorical parameters are presented as counts and percentages. Continuous parameters are displayed as either mean and standard deviation (SD) or median with 25 th and 75 th percentiles, based on their distributions. Despite extra recruitment, rare AVAS classes (with fewer than 9 patients, (< 2.5%)) did not yield sufficient number of patients for separate validation of the class. Due to their close clinical relevance, these classes were grouped into larger anatomically corresponding AVAS classes. This resulted into consolidation of 19 categories into 10. (See Supplementary Table S3). Types of pVAs and cVAs were also consolidated into seven classes based on their anatomical relevance. The reason was, again, rare incidence of certain types of VAs. (See Supplementary Table S4). The alternative options were to exclude rare classes, combine them all into a single group without considering anatomical relevance, or create miscellaneous categories. However, these alternative options would lead to clinical inaccuracies and incorrect validation. Initial data split The dataset was divided into training and a validation set with an 80/20 split. The stratification was based on the cVA category to ensure equal representation of cVA categories across both sets. The training dataset was used to develop all models, and the validation dataset was used to evaluate and compare performance of the models. Model development For development of all models, three algorithms were employed: multinomial logistic regression, random forest, and eXtreme Gradient Boosting (XGBoost, eXGB). Each of these algorithms was applied to the eight models described below, ensuring robust comparisons across techniques.
The dataset was divided into training and a validation set with an 80/20 split. The stratification was based on the cVA category to ensure equal representation of cVA categories across both sets. The training dataset was used to develop all models, and the validation dataset was used to evaluate and compare performance of the models.
For development of all models, three algorithms were employed: multinomial logistic regression, random forest, and eXtreme Gradient Boosting (XGBoost, eXGB). Each of these algorithms was applied to the eight models described below, ensuring robust comparisons across techniques.
The multinomial logistic regression models were developed using the multinom_reg function from the parsnip library within the tidymodels framework. Multinomial logistic regression is a statistical method used for predicting categorical outcomes with more than two categories. It models the probability of each category as a function of the predictors by estimating coefficients that maximize the likelihood of the observed data. The engine used for these models was glmnet . Hyperparameter tuning was performed for the following parameters: Penalty (Regularisation strength): Controls the degree of shrinkage applied to the model coefficients. Higher values enforce greater regularisation. Mixture (Proportion of L1 regularisation): Determines the balance between L1 (lasso) and L2 (ridge) regularisation. A value of 1 applies pure L1 regularisation, while a value of 0 applies pure L2 regularisation. Intermediate values represent a combination (elastic net regularisation).
The random forest models were implemented using the rand_forest function from the parsnip library within the tidymodels framework. Random forest is an ensemble learning method that constructs multiple decision trees during training and combines their outputs to improve prediction accuracy. It reduces overfitting and increases robustness by averaging the results of the trees, each built on random subsets of data and predictors. The engine used for random forest models was ranger . The following settings and hyperparameters were applied: Number of Trees (trees): Set empirically to 1000. Hyperparameter Tuning: o Minimum number of observations in a terminal node (min_n). o Number of predictors sampled for splitting at each node (mtry).
The XGBoost models were developed using the boost_tree function from the parsnip library within the tidymodels framework. XGBoost is a powerful algorithm that builds an ensemble of decision trees in a sequential manner, using a gradient boosting framework where each new tree corrects the errors of the previous ones by optimizing a differentiable loss function. The engine used for these models was xgboost . The following configurations were applied: Number of Trees (trees): Set empirically to 100. Hyperparameter Tuning: o Minimum number of data points required in a node for it to be split further (n_min). o Maximum depth of the tree (tree_depth). o Learning Rate: Determines the step size at each iteration (learn_rate). o Sample Size: Proportion of the training data to sample for each tree (sample_size).
Eight models were defined based on the predicted outcomes and the input features (the model abbreviation is in brackets): pVA predicted using mapping (Mapping pVA) pVA predicted using AVAS (AVAS pVA) pVA predicted using mapping and additional features (Mapping pVA param) pVA predicted using AVAS and additional features (AVAS pVA param) cVA predicted using mapping (Mapping cVA) cVA predicted using AVAS (AVAS cVA) cVA predicted using mapping and additional features (Mapping cVA param) cVA predicted using AVAS and additional features (AVAS cVA param) All 24 models (8 models across 3 algorithms) were evaluated using fivefold cross-validation, stratified by the pVA category to ensure equal representation of pVA categories across all folds. The main performance metric was ROC AUC, calculated using the method from Hand and Till (2001) for multiclass scenarios . The same cross-validation folds were used consistently across all algorithms to maintain comparability. Data preprocessing steps The same sequence of preprocessing steps (recipes) was used to create all models (library recipes 1.1.0): . Missing data imputation using a bagged tree algorithm to handle missing values. Categories with low frequency (below a 2.5% threshold) were collapsed. Categorical parameters were converted into dummy variables. Data upsampling was used to balance the occurrence of levels within the cVA and pVA. All models with preprocessing steps were developed in one workflow. Finalising models The performance of the models with tuned hyperparameters was assessed by using the validation dataset. Estimation and comparison of model performance (ROC AUC) was done by using Bayesian generalised linear models . ROC AUCs are expressed as mean with 95% credible intervals (95% CI). Differences between model performance are expressed as mean difference in ROC AUC with 95% CI. Differences with 95% CI that does not reach zero are supposed to be statistically significant. Three groups of model comparisons were performed (12 comparisons in total) for each of the three algorithms (multinomial logistic regression, random forest, and XGBoost): Predictive ability of mapping and AVAS (for both pVA and cVA) with and without additional features (demographic and clinical parameters), resulting in four comparisons. Effect of additional features on prediction of cVA and pVA by mapping and AVAS, resulting in four comparisons. How AVAS predicts pVA and cVA and what are the differences between these two predictions. How mapping predicts pVA and cVA and what are the differences between these two predictions (both with and without additional features), four comparisons. The algorithm with the best predictive performance was selected for the final presentation of results. Permutation-based variable importance was used to identify the additional features responsible for improving model performance (library vip) . To interpret the influence of the most important additional features on the prediction of cVA and pVA from models combining mapping and additional features (Mapping cVA param and Mapping pVA param models), we utilised Partial Dependence profiles (PDPs) and Accumulated Local Effects (ALE) profiles. PDPs were employed to visualise the marginal effect of key features on the model’s predictions. These plots depict the average predicted response as a given feature varies, while all other features are held constant . ALE profiles complemented PDPs by capturing local feature effects and accounting for potential interactions between features. Unlike PDPs, ALEs provide unbiased insights even in the presence of correlated features . Both PDPs and ALE profiles were generated using the DALEX (version 2.4.3) and DALEXtra (version 2.3.0) packages, leveraging the explain_tidymodels and model_profile functions. Visualisation was performed using the plot function within the same framework.
The same sequence of preprocessing steps (recipes) was used to create all models (library recipes 1.1.0): . Missing data imputation using a bagged tree algorithm to handle missing values. Categories with low frequency (below a 2.5% threshold) were collapsed. Categorical parameters were converted into dummy variables. Data upsampling was used to balance the occurrence of levels within the cVA and pVA. All models with preprocessing steps were developed in one workflow.
The performance of the models with tuned hyperparameters was assessed by using the validation dataset. Estimation and comparison of model performance (ROC AUC) was done by using Bayesian generalised linear models . ROC AUCs are expressed as mean with 95% credible intervals (95% CI). Differences between model performance are expressed as mean difference in ROC AUC with 95% CI. Differences with 95% CI that does not reach zero are supposed to be statistically significant. Three groups of model comparisons were performed (12 comparisons in total) for each of the three algorithms (multinomial logistic regression, random forest, and XGBoost): Predictive ability of mapping and AVAS (for both pVA and cVA) with and without additional features (demographic and clinical parameters), resulting in four comparisons. Effect of additional features on prediction of cVA and pVA by mapping and AVAS, resulting in four comparisons. How AVAS predicts pVA and cVA and what are the differences between these two predictions. How mapping predicts pVA and cVA and what are the differences between these two predictions (both with and without additional features), four comparisons. The algorithm with the best predictive performance was selected for the final presentation of results. Permutation-based variable importance was used to identify the additional features responsible for improving model performance (library vip) . To interpret the influence of the most important additional features on the prediction of cVA and pVA from models combining mapping and additional features (Mapping cVA param and Mapping pVA param models), we utilised Partial Dependence profiles (PDPs) and Accumulated Local Effects (ALE) profiles. PDPs were employed to visualise the marginal effect of key features on the model’s predictions. These plots depict the average predicted response as a given feature varies, while all other features are held constant . ALE profiles complemented PDPs by capturing local feature effects and accounting for potential interactions between features. Unlike PDPs, ALEs provide unbiased insights even in the presence of correlated features . Both PDPs and ALE profiles were generated using the DALEX (version 2.4.3) and DALEXtra (version 2.3.0) packages, leveraging the explain_tidymodels and model_profile functions. Visualisation was performed using the plot function within the same framework.
Exploratory data analysis Patients’ characteristics In total, 1151 patients were included in the analysis. One thousand and thirty-four patients were prospectively consecutively recruited, and 117 were enrolled as special recruitment. One hundred and seven patients did not have VA created and opted for a different renal replacement therapy option. Details of the demographic data, clinical parameters and vascular mapping are presented in Table . Prevalence of AVAS classes All patients were successfully classified using the AVAS classification system, see details in Fig. A. The largest group was AVAS1 (native fistula suitable) containing 1027 (89.2%) patients, then AVAS2 (prosthetic AVG suitable) involving 91 (7.9%) patients. AVAS3 (not conventional AVF or AVG suitable) had only 33 (2.9%) patients. Corresponding collapsed classes are seen in Fig. B. Predicted vascular access The most frequent type of pVA was autogenous radial-cephalic direct wrist access (402, 34.9%), followed by autogenous brachial-cephalic upper arm direct access (337, 29.3%) and autogenous radial-cephalic proximal access (109, 9.5%). No AVA placement was predicted in 17 (1.5%) patients. See the details in Fig. C. Corresponding collapsed classes are seen in Fig. D. Created vascular access The most frequent type of cVA was autogenous radial-cephalic direct wrist access (388, 33.7%), followed by autogenous brachial-cephalic upper arm direct access (309, 26.8%). One hundred and seven patients (9.3%) had no AVA created. See the details including other types of VAs in Fig. E. Corresponding collapsed classes are seen in Fig. F. Others in pVA are: 1 HeRO graft, 1 Axillar-axillary graft, 1 Autogenous ulnar-cephalic forearm transposition, 1 Autogenous radial-basilic forearm transposition. Others in cVA are: 2 Axillar-axillary grafts, 2 Autogenous radial-basilic forearm transpositions, 1 HeRO graft, 1 Autogenous ulnar-cephalic proximal access, and 1 Ulnar-basilic direct access without transposition. Predictive modelling The predictive performance of all three algorithms (multinomial logistic regression, random forest, and XGBoost), expressed as ROC AUC with 95% CI for all eight models, is graphically presented in Supplementary Figure S4. The comparison of performance between individual models is graphically shown in Supplementary Figure S5. Random forest was selected as the final algorithm for presenting results because it achieved the highest ROC AUC values for most models, although the differences compared to XGBoost were not statistically significant (see Supplementary Figure S5). Mapping versus AVAS in prediction of pVA (random forest algorithm) The mean ROC AUC with 95% CI for pVA prediction by AVAS was 0.79 (0.77;0.82) and by mapping was 0.85 (0.83;0.88), mean difference with 95% CI was 0.06 (0.04;0.08). Adding additional risk and demographic features increased the ROC AUC for AVAS to 0.87 (0.84;0.89)—an increase of 0.07 (0.05;0.09) and for mapping to 0.88 (0.86;0.91)—an increase of 0.03 (0.01;0.05). The difference between AVAS and mapping decreased and became statistically non-significant: 0.01 (-0.01;0.03). Mapping versus AVAS in prediction of cVA (random forest algorithm) The mean ROC AUC with 95% CI for cVA prediction by AVAS was 0.71 (0.69;0.74) and by mapping 0.8 (0.78;0.83), mean difference with 95% CI was 0.09 (0.07;0.11). Adding additional risk and demographic parameters increased the ROC AUC for AVAS to 0.82 (0.79;0.84)—an increase of 0.1 (0.08;0.12) and for mapping to 0.85 (0.83;0.88)—an increase of 0.05 (0.03;0.07). The difference between AVAS and mapping remained statistically significant in favour of mapping but decreased significantly to 0.03 (0.01;0.05). The results are summarized in Supplementary Table S5 and shown graphically in Figs. , and Supplementary Figures S6 and S7. Relationship between VA prediction frequency and VA creation frequency—data from Fig. D and 1F can be seen in Supplementary Figure S8. The native AVFs except autogenous posterior radial branch-cephalic direct access (snuffbox fistula), and AVG are more often predicted than created. The snuffbox fistula and class 3 categories (others, none, lower limb) are more often created than predicted. The biggest difference is in category 3 (predicted in 2.3% and created in 10.6%). The smallest difference is in the proximal radial-cephalic direct access class (predicted in 9.5% and created in 8.9%). Feature importance (random forest algorithm) The strongest parameters in the decision-making algorithm were the centre, age and BMI. Clinical parameters as sex, hypertension, previous and/or current central venous line, diabetes, history of cancer, smoking, and heart failure and ischaemic heart disease did not have a major impact. The details for all the models are displayed in Fig. . Feature explanation (random forest algorithm) The influence of the most significant additional features on the prediction of cVA and pVA was assessed using PDPs and ALE profiles. Supplementary Figure S9 illustrates the PDPs and ALE profiles for cVA prediction based on the Mapping cVA param model, while Supplementary Figure S10 presents analogous visualisations for pVA prediction from the Mapping pVA param model. Both figures focus on the additional features identified as most influential in the Feature Importance section. The most influential feature is centre. Next features are age and BMI with only slight effect. Further additional features such as hypertension and sex are insignificant.
Patients’ characteristics In total, 1151 patients were included in the analysis. One thousand and thirty-four patients were prospectively consecutively recruited, and 117 were enrolled as special recruitment. One hundred and seven patients did not have VA created and opted for a different renal replacement therapy option. Details of the demographic data, clinical parameters and vascular mapping are presented in Table . Prevalence of AVAS classes All patients were successfully classified using the AVAS classification system, see details in Fig. A. The largest group was AVAS1 (native fistula suitable) containing 1027 (89.2%) patients, then AVAS2 (prosthetic AVG suitable) involving 91 (7.9%) patients. AVAS3 (not conventional AVF or AVG suitable) had only 33 (2.9%) patients. Corresponding collapsed classes are seen in Fig. B. Predicted vascular access The most frequent type of pVA was autogenous radial-cephalic direct wrist access (402, 34.9%), followed by autogenous brachial-cephalic upper arm direct access (337, 29.3%) and autogenous radial-cephalic proximal access (109, 9.5%). No AVA placement was predicted in 17 (1.5%) patients. See the details in Fig. C. Corresponding collapsed classes are seen in Fig. D. Created vascular access The most frequent type of cVA was autogenous radial-cephalic direct wrist access (388, 33.7%), followed by autogenous brachial-cephalic upper arm direct access (309, 26.8%). One hundred and seven patients (9.3%) had no AVA created. See the details including other types of VAs in Fig. E. Corresponding collapsed classes are seen in Fig. F. Others in pVA are: 1 HeRO graft, 1 Axillar-axillary graft, 1 Autogenous ulnar-cephalic forearm transposition, 1 Autogenous radial-basilic forearm transposition. Others in cVA are: 2 Axillar-axillary grafts, 2 Autogenous radial-basilic forearm transpositions, 1 HeRO graft, 1 Autogenous ulnar-cephalic proximal access, and 1 Ulnar-basilic direct access without transposition.
In total, 1151 patients were included in the analysis. One thousand and thirty-four patients were prospectively consecutively recruited, and 117 were enrolled as special recruitment. One hundred and seven patients did not have VA created and opted for a different renal replacement therapy option. Details of the demographic data, clinical parameters and vascular mapping are presented in Table .
All patients were successfully classified using the AVAS classification system, see details in Fig. A. The largest group was AVAS1 (native fistula suitable) containing 1027 (89.2%) patients, then AVAS2 (prosthetic AVG suitable) involving 91 (7.9%) patients. AVAS3 (not conventional AVF or AVG suitable) had only 33 (2.9%) patients. Corresponding collapsed classes are seen in Fig. B.
The most frequent type of pVA was autogenous radial-cephalic direct wrist access (402, 34.9%), followed by autogenous brachial-cephalic upper arm direct access (337, 29.3%) and autogenous radial-cephalic proximal access (109, 9.5%). No AVA placement was predicted in 17 (1.5%) patients. See the details in Fig. C. Corresponding collapsed classes are seen in Fig. D.
The most frequent type of cVA was autogenous radial-cephalic direct wrist access (388, 33.7%), followed by autogenous brachial-cephalic upper arm direct access (309, 26.8%). One hundred and seven patients (9.3%) had no AVA created. See the details including other types of VAs in Fig. E. Corresponding collapsed classes are seen in Fig. F. Others in pVA are: 1 HeRO graft, 1 Axillar-axillary graft, 1 Autogenous ulnar-cephalic forearm transposition, 1 Autogenous radial-basilic forearm transposition. Others in cVA are: 2 Axillar-axillary grafts, 2 Autogenous radial-basilic forearm transpositions, 1 HeRO graft, 1 Autogenous ulnar-cephalic proximal access, and 1 Ulnar-basilic direct access without transposition.
The predictive performance of all three algorithms (multinomial logistic regression, random forest, and XGBoost), expressed as ROC AUC with 95% CI for all eight models, is graphically presented in Supplementary Figure S4. The comparison of performance between individual models is graphically shown in Supplementary Figure S5. Random forest was selected as the final algorithm for presenting results because it achieved the highest ROC AUC values for most models, although the differences compared to XGBoost were not statistically significant (see Supplementary Figure S5). Mapping versus AVAS in prediction of pVA (random forest algorithm) The mean ROC AUC with 95% CI for pVA prediction by AVAS was 0.79 (0.77;0.82) and by mapping was 0.85 (0.83;0.88), mean difference with 95% CI was 0.06 (0.04;0.08). Adding additional risk and demographic features increased the ROC AUC for AVAS to 0.87 (0.84;0.89)—an increase of 0.07 (0.05;0.09) and for mapping to 0.88 (0.86;0.91)—an increase of 0.03 (0.01;0.05). The difference between AVAS and mapping decreased and became statistically non-significant: 0.01 (-0.01;0.03). Mapping versus AVAS in prediction of cVA (random forest algorithm) The mean ROC AUC with 95% CI for cVA prediction by AVAS was 0.71 (0.69;0.74) and by mapping 0.8 (0.78;0.83), mean difference with 95% CI was 0.09 (0.07;0.11). Adding additional risk and demographic parameters increased the ROC AUC for AVAS to 0.82 (0.79;0.84)—an increase of 0.1 (0.08;0.12) and for mapping to 0.85 (0.83;0.88)—an increase of 0.05 (0.03;0.07). The difference between AVAS and mapping remained statistically significant in favour of mapping but decreased significantly to 0.03 (0.01;0.05). The results are summarized in Supplementary Table S5 and shown graphically in Figs. , and Supplementary Figures S6 and S7. Relationship between VA prediction frequency and VA creation frequency—data from Fig. D and 1F can be seen in Supplementary Figure S8. The native AVFs except autogenous posterior radial branch-cephalic direct access (snuffbox fistula), and AVG are more often predicted than created. The snuffbox fistula and class 3 categories (others, none, lower limb) are more often created than predicted. The biggest difference is in category 3 (predicted in 2.3% and created in 10.6%). The smallest difference is in the proximal radial-cephalic direct access class (predicted in 9.5% and created in 8.9%). Feature importance (random forest algorithm) The strongest parameters in the decision-making algorithm were the centre, age and BMI. Clinical parameters as sex, hypertension, previous and/or current central venous line, diabetes, history of cancer, smoking, and heart failure and ischaemic heart disease did not have a major impact. The details for all the models are displayed in Fig. . Feature explanation (random forest algorithm) The influence of the most significant additional features on the prediction of cVA and pVA was assessed using PDPs and ALE profiles. Supplementary Figure S9 illustrates the PDPs and ALE profiles for cVA prediction based on the Mapping cVA param model, while Supplementary Figure S10 presents analogous visualisations for pVA prediction from the Mapping pVA param model. Both figures focus on the additional features identified as most influential in the Feature Importance section. The most influential feature is centre. Next features are age and BMI with only slight effect. Further additional features such as hypertension and sex are insignificant.
The mean ROC AUC with 95% CI for pVA prediction by AVAS was 0.79 (0.77;0.82) and by mapping was 0.85 (0.83;0.88), mean difference with 95% CI was 0.06 (0.04;0.08). Adding additional risk and demographic features increased the ROC AUC for AVAS to 0.87 (0.84;0.89)—an increase of 0.07 (0.05;0.09) and for mapping to 0.88 (0.86;0.91)—an increase of 0.03 (0.01;0.05). The difference between AVAS and mapping decreased and became statistically non-significant: 0.01 (-0.01;0.03).
The mean ROC AUC with 95% CI for cVA prediction by AVAS was 0.71 (0.69;0.74) and by mapping 0.8 (0.78;0.83), mean difference with 95% CI was 0.09 (0.07;0.11). Adding additional risk and demographic parameters increased the ROC AUC for AVAS to 0.82 (0.79;0.84)—an increase of 0.1 (0.08;0.12) and for mapping to 0.85 (0.83;0.88)—an increase of 0.05 (0.03;0.07). The difference between AVAS and mapping remained statistically significant in favour of mapping but decreased significantly to 0.03 (0.01;0.05). The results are summarized in Supplementary Table S5 and shown graphically in Figs. , and Supplementary Figures S6 and S7. Relationship between VA prediction frequency and VA creation frequency—data from Fig. D and 1F can be seen in Supplementary Figure S8. The native AVFs except autogenous posterior radial branch-cephalic direct access (snuffbox fistula), and AVG are more often predicted than created. The snuffbox fistula and class 3 categories (others, none, lower limb) are more often created than predicted. The biggest difference is in category 3 (predicted in 2.3% and created in 10.6%). The smallest difference is in the proximal radial-cephalic direct access class (predicted in 9.5% and created in 8.9%).
The strongest parameters in the decision-making algorithm were the centre, age and BMI. Clinical parameters as sex, hypertension, previous and/or current central venous line, diabetes, history of cancer, smoking, and heart failure and ischaemic heart disease did not have a major impact. The details for all the models are displayed in Fig. .
The influence of the most significant additional features on the prediction of cVA and pVA was assessed using PDPs and ALE profiles. Supplementary Figure S9 illustrates the PDPs and ALE profiles for cVA prediction based on the Mapping cVA param model, while Supplementary Figure S10 presents analogous visualisations for pVA prediction from the Mapping pVA param model. Both figures focus on the additional features identified as most influential in the Feature Importance section. The most influential feature is centre. Next features are age and BMI with only slight effect. Further additional features such as hypertension and sex are insignificant.
AVAS classification was designed to ease communication between VA specialists. It classifies patients dependent on HD as each patient has different anatomy which leads to different challenges and varying needs in health care. This validation was conducted to evaluate its predictive performance as we recognise its potential in clinical practice. It presents valuable information without the need for lengthy paragraphs and detailed reports. In a busy clinic with many patients undergoing mapping, it may be more convenient for the clinician to use AVAS rather than consider all the mapping measurements themselves. AVAS also helps to determine the appropriate type of VA centre for patients care as AVAS class provides a clear information on how challenging the case might be. Since AVAS was derived from mapping, we anticipated lower performance due to expected loss of some details. Overall, all model performances had ROC AUC scores above 0.7, which we consider high, especially given the limitations posed by rare classes and their consolidation. The highest performance was noted in mapping for pVA with additional parameters (0.88, (0.86;0.91)). In general, mapping had overall better performance. However, when additional parameters were implemented, the already slim performance gap narrowed significantly. For pVA, the difference was 1.4% statistically not significant, for cVA the difference was 3.3% statistically significant. Prediction of pVA showed consistently higher performance than cVA. One of the reasons is that surgeons sometimes opt for more distal AVA despite less suitable anatomy to spare proximal vessels (and future options for VA). This has been associated with significantly higher risk of VA early failure. AVAS prediction of cVA with additional parameters still had high ROC AUC score 0.82 (0.79;0.84). AVAS was designed to inform the VA specialists about patient’s vascular status, aiding to personalise patients care. However, factors such as specific practices of each centre, as well as the surgeon’s training and experience, cannot be fully accounted for by the classification system. Nevertheless, these factors play a significant role in the decision-making algorithm. The predictive ability of AVAS in predicting cVA was greatly improved by implementing additional parameters (mean difference ROC AUC 0.1 (0.08;0.12)). They also play a very slight role in prediction of pVA by mapping (mean difference ROC AUC 0.03 (0.01;0.05)). This also highlights the impact on surgeon’s decision, but not on the VA prediction by the clinicians, who initially examined the patient. The parameters that impacted predictive ability of the models were in descending order: centre, age, and BMI (Supplementary Figures S9-10). Age seems to have a stronger influence on certain types of VAs. Proximal native fistulae, such as brachial-basilic and brachial-cephalic direct access, show a clear preference in prediction and creation as patients age. This indicates a tendency toward favouring proximal VA in older patients. In contrast, distal radial-cephalic direct access exhibits an opposite trend, with both prediction and creation declining significantly in patients aged 80 years and older, suggesting that distal access is less favourable in this age group. Other types of VAs demonstrate relatively stable predictions across different age groups. BMI appears to influence radial-cephalic VA types as well. Distal radial-cephalic direct access shows a slight upward trend, peaking at a BMI of around 30, but then rapidly declines as BMI approaches 40. In contrast, proximal radial-cephalic direct access demonstrates a steady increase with rising BMI. Other vascular access types show relatively stable predictions across the BMI range. Centre was the most influential factor with permutation feature importance up to 31.4% for cVA by mapping with additional parameters. The analysis reveals that the most common VA types display consistent preferences across all centres, with only slight variations in their relative dominance. Two most frequent VA types—distal radial-cephalic direct access and brachial-cephalic direct access are consistently favoured across most institutions. However, certain centres demonstrate stronger preferences for specific VA types, reflecting the impact of local practices and resources on VA predictions. These findings emphasize the variability introduced by centre-specific factors and the importance of incorporating these differences into VA planning and predictive modelling. Overall, the graphs illustrate the dominant role centres play in shaping VA predictions, driven by local practices, resources, and clinician expertise. (Supplementary Figures S9-10). Based on our results, we believe, that AVAS implementation into clinical practice could benefit clinicians. Possible logistical barriers to implementing AVAS in routine practice can be overcome by following strategies: Firstly, education and training are needed to ensure that clinicians apply AVAS accurately, which can be addressed through educational materials, workshops and visual guides. Secondly, embedding AVAS into electronic health records templates can streamline workflows and ensure seamless documentation. Thirdly, variability in clinical workflows and expertise across centres may challenge AVAS’s uniform implementation. Establishing clear guidelines, incorporating AVAS into standard vascular mapping protocols, and promoting its use through professional societies can help ensure consistency and reliability. Time constraints in busy clinics may be a concern, but AVAS simplifies documentation and improves communication, saving time overall. Furthermore, implementation into research for classifying the haemodialysis population, can support the adoption and implementation of this system. By addressing these barriers with training and system integration AVAS could become a widely adopted tool for improving VA planning. Strengths This study has several key strengths that contribute to its significance in advancing VA management. One of it is the participation of eleven centres from eight countries, which enhances the generalisability of the findings. By capturing a diverse range of patient populations, institutional protocols, and geographical contexts, the study ensures its relevance across various clinical settings. Large dataset of 1151 patients enabled to detect significant differences that would have been missed with smaller sample size. It also allowed us to evaluate the importance of clinical and demographic factors in predictive capability and decision algorithms of mapping and AVAS. The study has a rigorous methodological design. The use of consecutive patient recruitment minimises selection bias, ensuring that the results reflect real-world clinical scenarios. Additionally, the incorporation of ML models adds an objective layer of analysis, bolstering the reliability of the findings. Limitations The primary limitation is the uneven distribution of AVAS classes and types of VA, which nonetheless reflects the clinical reality. Despite 3-year recruitment followed by a half-year extra recruitment of rare categories, some remained with very low numbers. As a result, we had to merge the very rare classes with those most similar from a clinical perspective, which may lead to several forms of bias. Firstly, merging rare AVAS classes leads to a loss of granularity and detail. These changes are maybe subtle, but differences in clinical outcomes between these classes are obscured. Patients in distinct rare classes may present unique vascular characteristics or clinical needs that are not fully captured when these groups are combined. This simplification can reduce the accuracy of predictive models for these rare cases, as their unique patterns are diluted into larger, more complex categories. Secondly, the rare cases may require individualised planning and care. By merging them, the unique challenges they present are less visible in the data, potentially limiting the clinical utility of the models when applied to this subgroup. Further, consolidating rare classes into larger ones, the model tends to emphasise the characteristics of the more frequent or dominant AVAS groups, leading to a skewed prediction bias. For instance, model accuracy may remain high for the combined group overall, but the performance may degrade for patients originally classified into the smaller, rare classes. This bias could underestimate the challenges and outcomes of the rare cases. Nevertheless, we would argue that the ranking of the predictive models remained clinically relevant despite these limitations. From a ML perspective, merging classes improves statistical robustness and avoids issues associated with insufficient sample size. Even though, the interpretability of the model is reduced, as clinicians may question whether predictions for a merged group apply equally to all its constituent subcategories. To address these biases, rare AVAS classes were merged based on clinically and anatomically relevant similarities, ensuring that groupings remained logical and interpretable. This strategy was chosen over alternative approaches, such as excluding rare classes entirely (which would compromise generalisability) or creating a “miscellaneous” category (which lacks clinical relevance). Moreover, the study’s large overall sample size (n = 1151) and international, multicentre design help offset some of these limitations by enhancing the robustness and generalisability of the findings.
This study has several key strengths that contribute to its significance in advancing VA management. One of it is the participation of eleven centres from eight countries, which enhances the generalisability of the findings. By capturing a diverse range of patient populations, institutional protocols, and geographical contexts, the study ensures its relevance across various clinical settings. Large dataset of 1151 patients enabled to detect significant differences that would have been missed with smaller sample size. It also allowed us to evaluate the importance of clinical and demographic factors in predictive capability and decision algorithms of mapping and AVAS. The study has a rigorous methodological design. The use of consecutive patient recruitment minimises selection bias, ensuring that the results reflect real-world clinical scenarios. Additionally, the incorporation of ML models adds an objective layer of analysis, bolstering the reliability of the findings.
The primary limitation is the uneven distribution of AVAS classes and types of VA, which nonetheless reflects the clinical reality. Despite 3-year recruitment followed by a half-year extra recruitment of rare categories, some remained with very low numbers. As a result, we had to merge the very rare classes with those most similar from a clinical perspective, which may lead to several forms of bias. Firstly, merging rare AVAS classes leads to a loss of granularity and detail. These changes are maybe subtle, but differences in clinical outcomes between these classes are obscured. Patients in distinct rare classes may present unique vascular characteristics or clinical needs that are not fully captured when these groups are combined. This simplification can reduce the accuracy of predictive models for these rare cases, as their unique patterns are diluted into larger, more complex categories. Secondly, the rare cases may require individualised planning and care. By merging them, the unique challenges they present are less visible in the data, potentially limiting the clinical utility of the models when applied to this subgroup. Further, consolidating rare classes into larger ones, the model tends to emphasise the characteristics of the more frequent or dominant AVAS groups, leading to a skewed prediction bias. For instance, model accuracy may remain high for the combined group overall, but the performance may degrade for patients originally classified into the smaller, rare classes. This bias could underestimate the challenges and outcomes of the rare cases. Nevertheless, we would argue that the ranking of the predictive models remained clinically relevant despite these limitations. From a ML perspective, merging classes improves statistical robustness and avoids issues associated with insufficient sample size. Even though, the interpretability of the model is reduced, as clinicians may question whether predictions for a merged group apply equally to all its constituent subcategories. To address these biases, rare AVAS classes were merged based on clinically and anatomically relevant similarities, ensuring that groupings remained logical and interpretable. This strategy was chosen over alternative approaches, such as excluding rare classes entirely (which would compromise generalisability) or creating a “miscellaneous” category (which lacks clinical relevance). Moreover, the study’s large overall sample size (n = 1151) and international, multicentre design help offset some of these limitations by enhancing the robustness and generalisability of the findings.
In this study, the AVAS system has been validated by predictive modelling using ML algorithms. Although mapping demonstrated slightly higher performance, the difference, particularly with additional parameters, were not clinically significant. Therefore, we consider AVAS to be well-validated and highly relevant in clinical practice. Our next goal is to present a simpler version of AVAS derived from this dataset, without compromising its effectiveness and maintaining its clinical value.
Supplementary Information.
|
Postmortem interval estimation of time since death: impact of non-histone binding proteins, immunohistochemical, and histopathological changes in vivo | 717e645f-8c35-4923-9612-eeb500801266 | 11611056 | Anatomy[mh] | The human body undergoes complicated changes after death, which are known to be influenced by internal and external variables. Both the environment and the way an individual dies can have a substantial impact on the rate of decomposition. Nevertheless, it has been shown that the decomposition process is predictable, providing the opportunity to determine the postmortem interval (PMI), or the duration since death, based on microscopic and/or gross morphological changes to the body. A precise and accurate PMI estimation can assist medicolegal investigators in identifying the deceased, help establish the chronology of events leading up to death, and confirm or refute other forensic evidence . Determining the PMI is one of the primary objectives and most challenging tasks of forensic pathology. Even though this parameter is unquestionably vital, particularly in criminal investigations, many studies on the subject only address the postmortem parameters' time dependency, which has a minimal bearing on actual forensic practice . Contempt for the emergence of multiple novel methods for PMI assessment in recent years indicates that methods based on the examination of rigor, algor, and livor mortis continue to be the most commonly utilized. However, a variety of individual and environmental factors, such as age, gender, ambient temperature, and physiological and pathological states, can have an impact on these variables. They are primarily the result of physical and chemical processes during the postmortem period. Because of this, traditional methods frequently exhibit limitations in their application, as well as inaccuracy and unreliability . All body tissues undergo a series of extensive biochemical alterations following death due to low oxygen levels, hampered enzymatic reactions, lack of metabolites, and disintegration of cellular components. These biochemical alterations may serve as indicators for a more precise PMI assessment . Researchers have looked into biochemical markers that aid in determining the amount of time that has passed since death. These consist of protein fractions, urea, creatinine, glucose, iron, potassium, calcium, enzymes, myo-albumin fraction, and the amount of strontium-90 calcium analogs . Another one of these is the immunohistochemical detection of insulin in pancreatic β-cells. Medical procedures like measuring rigor mortis, checking the liver, or taking a body temperature can only be used to accurately measure the PMI in the first two or three days after death . There is also a negative correlation between the significance and practical utility of PMI estimation and the amount of research focused on it . One of the most critical areas of research in PMI estimation involves the protein high mobility group box 1 (HMGB1), a nuclear protein with a highly conserved amino acid sequence across species found in many eukaryotic cells. HMGB1 seems to have two distinct functions in cellular systems. First, it acts as an intracellular transcription regulator, essential for maintaining DNA function. Second, during necrosis, triggered by factors such as lipopolysaccharides, tumor necrosis factor-alpha, interleukin-1, and interferon-gamma, HMGB1 translocates from the nucleus to the periphery and is released from macrophages . In this study, gastrocnemius muscle tissue was used because it undergoes autolytic changes and putrefaction more slowly than body fluids and other organs. Despite limited studies on the role of HMGB1 in PMI estimation, its involvement is promising. As necrosis is part of the postmortem process, and necrotic cells release HMGB1, detecting serum HMGB1 from necrotic tissue could provide a potential link to PMI . This research aimed to examine the potential of serum HMGB1 as a postmortem marker using ELISA analysis in 50 male Wistar rats. Additionally, immunohistochemical staining of desmin, combined with histopathological changes in the gastrocnemius muscle, was evaluated to assess the effects of these biomarkers in estimating PMI. Animals All methods were conducted in accordance with ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines and regulations. Male Wistar rats were purchased from King Abdulaziz University in Jeddah, Saudi Arabia. Before beginning the experiment, the animals were given one week to acclimate to the laboratory environment. They were kept in metal cages for the experiment, provided a standard diet, and had unlimited access to water. The rats were kept in a 12-hour light-dark cycle with a controlled temperature of 22–24°C and relative humidity of 50–60% at the Anatomy Department, Faculty of Medicine, Al-Baha University. Experimental animals and animal grouping Fifty male Wistar rats weighing 230 ± 10 g and 12 weeks old were used in this study. The rats were divided into five groups of ten and housed in plastic cages with free access to water and a standard laboratory diet. After light sedation with halothane, the rats were sacrificed by cervical dislocation. Skeletal muscle samples were collected at 0, 24-, 48-, 72-, and 96-hours postmortem. The first group, serving as the control, had samples taken immediately after sacrifice (0 hours). In contrast, samples were taken at 24-, 48-, 72-, and 96-hours postmortem for the second, third, fourth, and fifth groups, respectively. The sacrificed rats were maintained at room temperature throughout the experiment, 22 ± 2ºC during the day and 9 ± 2ºC at night. Rats were chosen as the model for this study due to their physiological and morphological similarities to humans. Additionally, the gastrocnemius muscle is easier to extract in rats compared to mice. The use of rats for experimentation was conducted in strict accordance with ethical standards and the internationally recognized guidelines for the care and use of laboratory animals (Approval Ethical Number: REC/ANT/BU-FM/2024/55). Rat samples for HMGB1 ELISA detection Blood samples were obtained from the heart and the great vessels during the autopsy. The samples were centrifuged at 3,000 rpm for 5 minutes and stored at -80°C for further analysis. Serum HMGB1 concentration was determined with an ELISA kit (ThermoFisher; Sci, Cat no: EEL102) . Immunohistochemical and histological studies Tissue samples were fixed in 10% formal saline for 48 hours, rinsed with tap water, and processed to create paraffin sections for immunohistochemical and histological analysis. The skeletal muscle sections used in the immunohistochemical study were mounted on charged slides, deparaffinized with two fresh xylene changes, and rehydrated in graded ethanol (99%, 95%, and 70%) before being rinsed in PBS for ten minutes. Antigen retrieval was then carried out after the endogenous peroxidase activity was inhibited. Specimens were then kept in an incubator at room temperature. The primary desmin antibodies (Sigma Aldrich Company) were applied to the sections and incubated for the entire night at room temperature in a humidified chamber. Each section received a labeling antibody, and dimethoxybenzidine was used as a chromogen. After that, the sections were mounted in DPX, counterstained, and dehydrated in increasing alcohol grades before being inspected under a regular light microscope . For histological analysis, gastrocnemius muscle samples were embedded in paraffin blocks, cleared with xylene, dehydrated with alcohol, and fixed in buffered formalin. Hematoxylin and eosin stain were applied after the paraffin blocks were cut into 5 µm thickness . Sample size The sample size was determined based on similar previous studies . Ten samples were determined in each group using G*power v.3.1.9.5, which was used to calculate the sample size based on an effect size of 1.6254, a two-tailed test, an α error of 0.05, and a power of 90.0%. Computer-assisted digital image analysis (digital morphometric study) Slides were photographed using an MVV5000CL digital eyepiece installed on a MEIJI MX5200L microscope and Future WinJoe software using a 200X objective. An Intel Core I7-based computer was used to analyze the resulting 20X images for immunohistochemical staining surface area percentage. To achieve this, Fiji ImageJ (version 1.51r; NIH) software was used, along with the color deconvolution 2 plugin (histological dyes digital separation). As a result, three separate digital images were produced (hematoxylin, DAB, and supplementary). The threshold function was used on the DAB channel in immunohistochemistry for intensity calibration, followed by positive area percentage measurement. Data was exported to Excel for further analysis. Five random fields sized 200×200 µm from each slide were analyzed and averaged. In order to determine the histological autolysis score, samples were analyzed and ranked using a semi-quantitative scale in relation to the control samples: 1 denotes minimal autolysis (<5%), 2 mild autolysis (5–10%), 3 moderate autolysis (10–50%), and 4 severe autolysis (>50%). For every slide, a minimum of ten fields (10X) were examined and averaged . Statistical analysis and data interpretation GraphPad Prism 8 (GraphPad Software) was used to analyze the data. The Kolmogorov-Smirnov and Shapiro-Wilk tests were used to assess the normality of numerical data. The surface area percentage data from immunohistochemical staining displayed a normal (parametric) distribution. The mean and standard deviation (SD) values of the data were displayed. The groups were compared using a one-way ANOVA and an ad hoc Tukey's multiple comparison test. The significance level of these results was determined at 0.05. All methods were conducted in accordance with ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines and regulations. Male Wistar rats were purchased from King Abdulaziz University in Jeddah, Saudi Arabia. Before beginning the experiment, the animals were given one week to acclimate to the laboratory environment. They were kept in metal cages for the experiment, provided a standard diet, and had unlimited access to water. The rats were kept in a 12-hour light-dark cycle with a controlled temperature of 22–24°C and relative humidity of 50–60% at the Anatomy Department, Faculty of Medicine, Al-Baha University. Fifty male Wistar rats weighing 230 ± 10 g and 12 weeks old were used in this study. The rats were divided into five groups of ten and housed in plastic cages with free access to water and a standard laboratory diet. After light sedation with halothane, the rats were sacrificed by cervical dislocation. Skeletal muscle samples were collected at 0, 24-, 48-, 72-, and 96-hours postmortem. The first group, serving as the control, had samples taken immediately after sacrifice (0 hours). In contrast, samples were taken at 24-, 48-, 72-, and 96-hours postmortem for the second, third, fourth, and fifth groups, respectively. The sacrificed rats were maintained at room temperature throughout the experiment, 22 ± 2ºC during the day and 9 ± 2ºC at night. Rats were chosen as the model for this study due to their physiological and morphological similarities to humans. Additionally, the gastrocnemius muscle is easier to extract in rats compared to mice. The use of rats for experimentation was conducted in strict accordance with ethical standards and the internationally recognized guidelines for the care and use of laboratory animals (Approval Ethical Number: REC/ANT/BU-FM/2024/55). Blood samples were obtained from the heart and the great vessels during the autopsy. The samples were centrifuged at 3,000 rpm for 5 minutes and stored at -80°C for further analysis. Serum HMGB1 concentration was determined with an ELISA kit (ThermoFisher; Sci, Cat no: EEL102) . Tissue samples were fixed in 10% formal saline for 48 hours, rinsed with tap water, and processed to create paraffin sections for immunohistochemical and histological analysis. The skeletal muscle sections used in the immunohistochemical study were mounted on charged slides, deparaffinized with two fresh xylene changes, and rehydrated in graded ethanol (99%, 95%, and 70%) before being rinsed in PBS for ten minutes. Antigen retrieval was then carried out after the endogenous peroxidase activity was inhibited. Specimens were then kept in an incubator at room temperature. The primary desmin antibodies (Sigma Aldrich Company) were applied to the sections and incubated for the entire night at room temperature in a humidified chamber. Each section received a labeling antibody, and dimethoxybenzidine was used as a chromogen. After that, the sections were mounted in DPX, counterstained, and dehydrated in increasing alcohol grades before being inspected under a regular light microscope . For histological analysis, gastrocnemius muscle samples were embedded in paraffin blocks, cleared with xylene, dehydrated with alcohol, and fixed in buffered formalin. Hematoxylin and eosin stain were applied after the paraffin blocks were cut into 5 µm thickness . The sample size was determined based on similar previous studies . Ten samples were determined in each group using G*power v.3.1.9.5, which was used to calculate the sample size based on an effect size of 1.6254, a two-tailed test, an α error of 0.05, and a power of 90.0%. Slides were photographed using an MVV5000CL digital eyepiece installed on a MEIJI MX5200L microscope and Future WinJoe software using a 200X objective. An Intel Core I7-based computer was used to analyze the resulting 20X images for immunohistochemical staining surface area percentage. To achieve this, Fiji ImageJ (version 1.51r; NIH) software was used, along with the color deconvolution 2 plugin (histological dyes digital separation). As a result, three separate digital images were produced (hematoxylin, DAB, and supplementary). The threshold function was used on the DAB channel in immunohistochemistry for intensity calibration, followed by positive area percentage measurement. Data was exported to Excel for further analysis. Five random fields sized 200×200 µm from each slide were analyzed and averaged. In order to determine the histological autolysis score, samples were analyzed and ranked using a semi-quantitative scale in relation to the control samples: 1 denotes minimal autolysis (<5%), 2 mild autolysis (5–10%), 3 moderate autolysis (10–50%), and 4 severe autolysis (>50%). For every slide, a minimum of ten fields (10X) were examined and averaged . GraphPad Prism 8 (GraphPad Software) was used to analyze the data. The Kolmogorov-Smirnov and Shapiro-Wilk tests were used to assess the normality of numerical data. The surface area percentage data from immunohistochemical staining displayed a normal (parametric) distribution. The mean and standard deviation (SD) values of the data were displayed. The groups were compared using a one-way ANOVA and an ad hoc Tukey's multiple comparison test. The significance level of these results was determined at 0.05. Postmortem change of serum HMGB1 levels Serum HMGB1 levels were measured during the postmortem period using ELISA. The study found that serum HMGB1 levels in the deceased rats at 22 ± 2°C had significant changes over time. HMGB1 levels peaked on the second-day postmortem, followed by a gradual decline over the third and fourth days in a time-dependent manner . Immunohistochemical findings: cleaved desmin expression When the PMI increased, positively stained areas gradually decreased, according to desmin immunohistochemical staining. The greatest increase in desmin expression (nuclear and cytoplasmic) occurred at 0 hours postmortem. After 24 hours, moderate to high cytoplasmic desmin expression was found, followed by moderate expression 48 hours later (72 hours from time of death), with weak expression occurring 96 hours from time of death . Postmortem gastrocnemius muscle histological findings Hematoxylin and eosin-stained sections of the rat gastrocnemius muscle showed typical histological features at the time of death, including flat, peripherally located nuclei and parallel striated muscle fibers . At 24 hours postmortem, skeletal muscle myofibers began to lose striation, and nuclei became eccentrically located . By 48 hours, the gastrocnemius muscle showed eccentric pyknotic nuclei, widely spaced myofibers, and wavy nuclei . At 72 hours postmortem, segmented muscle fibers, loss of intracellular nuclei, and the disappearance of cross striations were evident . By 96 hours, the skeletal muscle tissue exhibited dilated, clogged blood vessels, fragmented muscle fibers with irregular nuclei, and large regions of tissue loss replaced by thin collagenous connective tissue . A quantified comparison of muscle autolysis across different time intervals is shown in . Serum HMGB1 levels were measured during the postmortem period using ELISA. The study found that serum HMGB1 levels in the deceased rats at 22 ± 2°C had significant changes over time. HMGB1 levels peaked on the second-day postmortem, followed by a gradual decline over the third and fourth days in a time-dependent manner . When the PMI increased, positively stained areas gradually decreased, according to desmin immunohistochemical staining. The greatest increase in desmin expression (nuclear and cytoplasmic) occurred at 0 hours postmortem. After 24 hours, moderate to high cytoplasmic desmin expression was found, followed by moderate expression 48 hours later (72 hours from time of death), with weak expression occurring 96 hours from time of death . Hematoxylin and eosin-stained sections of the rat gastrocnemius muscle showed typical histological features at the time of death, including flat, peripherally located nuclei and parallel striated muscle fibers . At 24 hours postmortem, skeletal muscle myofibers began to lose striation, and nuclei became eccentrically located . By 48 hours, the gastrocnemius muscle showed eccentric pyknotic nuclei, widely spaced myofibers, and wavy nuclei . At 72 hours postmortem, segmented muscle fibers, loss of intracellular nuclei, and the disappearance of cross striations were evident . By 96 hours, the skeletal muscle tissue exhibited dilated, clogged blood vessels, fragmented muscle fibers with irregular nuclei, and large regions of tissue loss replaced by thin collagenous connective tissue . A quantified comparison of muscle autolysis across different time intervals is shown in . In any postmortem investigation, one of the most crucial medico-legal questions is determining the time since death, also known as the postmortem interval. PMI refers to the period that has passed following a person's death. If the exact time of death is unknown, various medical and scientific methods are employed to estimate it, including examining blood markers, organs, and bodily fluids after death . It is challenging to evaluate our results because studies have been done using various analytical techniques to determine the time of death, as evidenced by technological advancements. These studies measured different parameters to determine the PMI. The time of death has been predicted using a variety of biochemical, immunohistochemical, and histopathological approaches; however, the usefulness of these approaches is restricted. The current study found that the rats' bodies developed changes in serum HMGB1 levels after death. This research sought to determine whether serum HMGB1 could be used to estimate PMI. Our results revealed that HMGB1 levels increased up to 48 hours after death. This increase was most pronounced at 22 ± 2°C between zero and 48 hours. After peaking at 48 hours, the serum HMGB1 levels decreased over the following days at 72 and 96 hours. Measuring HMGB1 levels to assess the PMI is helpful because organs do not need to be homogenized; only serum samples are needed. ELISA allows for HMGB1 analysis within two days, making it a practical tool for PMI calculation. This approach may provide insights into the biochemical changes after death and offer potential markers for PMI estimation. Our findings align with Kikuchi’s study on Wistar rats, which also demonstrated a time-dependent increase in HMGB1 levels during the first two days postmortem, with a peak on the third day followed by a decline on the fourth, then plateauing at a specific temperature . Similarly, El-Din et al . highlighted the critical role of HMGB1 in PMI estimation within the timeframes examined in our study . In addition, our research revealed a relationship between PMI and the degradation of the desmin protein. Desmin expression, as observed through immunohistochemical analysis, decreased from strong immunoreactivity in nearly all muscle fibers at 0 hours (38.19%) to minimal expression (1.09%) at 96 hours. Desmin is highly sensitive to the activities of calpains, proteasomes, and lysosomes, which are crucial for skeletal muscle cytoskeletal integrity . Additionally, desmin makes up most of the intermediate filament that surrounds the sarcomere's Z-disk and connects it to the subsarcolemmal cytoskeleton, accounting for 0.35% of all proteins found in muscles . Our results are consistent with earlier studies on pigs, which showed that desmin protein persists for a few days after death (up to 224 hours postmortem) . Furthermore, our findings agree with Koohmaraie et al . , Taylor et al . , and Pittner et al . , all of whom documented the degradation of desmin as a reliable indicator of postmortem protein breakdown. Additionally, the products of desmin degradation in humans are comparable to those discovered in research examining the molecular makeup of muscle proteins in a variety of domestic animals, including cattle , pigs , and lambs , which showed regular degradation one to two days after death. Moreover, our study's histological examination of muscle tissues revealed a strong correlation with PMI. Morphological changes in the gastrocnemius muscle could serve as valuable indicators for estimating the time of death. At 0 hours postmortem, the gastrocnemius muscle exhibited normal histological morphology, with parallel striated muscle fibers and flat, peripherally located nuclei. By 24 hours, changes in muscle striations and nuclear alterations were noted. After 48 hours, the muscle displayed cytoplasmic vacuoles and increased nuclear alterations. By 72 and 96 hours, the muscle showed a reduction in striations and nuclear presence, with fragmented muscle fibers and pyknotic nuclei. These findings agree with Guerrero-Urbina et al . , who observed pyknotic nuclei in myofibers at 96-120 hours PMI and were not present at earlier postmortem intervals. The changes in the histology of the human lingual striated muscle make it possible to estimate PMI. Similarly, Yahia et al . demonstrated that histopathological changes in the kidneys, liver, heart, and skeletal muscles could be used to estimate PMI in dogs. The histological changes observed in multiple studies, including our muscle histology results, support the use of histological examination as a reliable tool for PMI estimation. Our findings also align with Mostafa et al . , who found a strong correlation between histological changes in muscle tissue and postmortem intervals. Limitations Postmortem, the gastrocnemius muscle is more stable compared with other bodily tissues. This study has shown that it is easily applicable to PMI investigations. To validate these results, more work is necessary. To assess future chosen biomarkers, a larger cohort of experimental animals would need to undergo targeted analysis in the following phase. A model for estimating PMI using muscle tissue might be developed if the concentrations of certain targeting minerals and multiple proteins were quantified. This would enable statistical analysis of the data. Applying techniques related to molecular levels to muscle will also be crucial. Postmortem, the gastrocnemius muscle is more stable compared with other bodily tissues. This study has shown that it is easily applicable to PMI investigations. To validate these results, more work is necessary. To assess future chosen biomarkers, a larger cohort of experimental animals would need to undergo targeted analysis in the following phase. A model for estimating PMI using muscle tissue might be developed if the concentrations of certain targeting minerals and multiple proteins were quantified. This would enable statistical analysis of the data. Applying techniques related to molecular levels to muscle will also be crucial. This study concluded that postmortem changes occur in a time-dependent manner, with a gradual deterioration of the gastrocnemius muscle’s histological structure as time progresses after death. The findings demonstrated a clear correlation between the PMI and the decline in the HMGB1 biomarker, along with changes in immunoreactivity and histopathological alterations in the muscle tissue. Ultimately, it was determined that the HMGB1 biomarker desmin expression and the observed histological changes in the gastrocnemius muscle could be reliable indicators for accurate PMI estimation. Both HMGB1 and desmin were identified as potential functional biomarkers for PMI in this study, which also emphasizes the need for further in-depth research in this area. The study suggests that these markers hold promise as valuable tools for postmortem analysis in humans and should be explored further in future investigations. |
Neuroendocrine neoplasms of the pancreas: diagnosis and pitfalls | 6fdbe4ab-0d57-44d9-8d6b-be973aba0331 | 8986719 | Anatomy[mh] | Diagnosis and pitfall are like two sides of the same coin. The better the diagnostic criteria, the less the number of diagnostic pitfalls. However, careful processing of diagnostic failures has helped in many cases to improve diagnostic criteria. In this article on pancreatic neuroendocrine neoplasms (PanNENs), we will therefore focus on both, presenting the most important diagnostic criteria and providing clues to avoid main pitfalls. PanNENs belong to the tumors that bear the generic name “neuroendocrine neoplasms.” This name is used as a collective term for two tumor families that share the expression of neuroendocrine markers, such as synaptophysin and chromogranin A, but differ distinctly in their morphological and molecular profiles . In the first group, growth and behavior is slower and individually more different than in the second group, where it is generally faster . Both groups of NENs can arise almost anywhere in the body, even though they all show a strong preference for the gastroenteropancreatic system and the lung, with a varying and interesting site-specific distribution. For historical reasons, the world health organization (WHO) classifications of the NENs of the various organ systems do not follow a uniform terminology . However, the WHO generally follows the principle of distinguishing between well and poorly differentiated NENs and the delimitation of mixed neoplasms [ , , – ]. The PanNENs play a pioneering role in the classification of NENs because they are frequent among the NENs, have a very varied morphology, and may show a multifaceted functionality . Currently, they are represented in two WHO “blue books,” the classifications of tumors of endocrine organs and the classification of digestive system tumors . This review aims to outline the morpho-genetic characteristics of pancreatic NENs and to provide a practical approach to daily routine diagnostics with highlighting of main diagnostic pitfalls and important NEN mimics.
The 2017 and 2019 WHO classifications stratify the PanNENs into well-differentiated NENs (pancreatic neuroendocrine tumors, PanNETs) and poorly differentiated NENs (pancreatic neuroendocrine carcinomas, PanNECs) and presume that all PanNENs have a malignant potential, however, with a different probability to metastasize . PanNET The histological profile of PanNETs is characterized by an organoid growth pattern with an ordered arrangement of cells, a variable amount of fibrotic stroma, and only rare necrotic changes (Figs. and ). Although the organoid pattern is not uniform but very diverse, the histological diversity can probably be traced back to either a solid or trabecular architecture that can be subdivided due to the composition of the stroma and its vascularization into solid-nested, solid-paraganglioma-like, solid-microglandular, and trabecular-reticulated, trabecular-gyriform, and trabecular-cystic patterns [ – ]. Interestingly, some of these patterns seem to have a relationship to the functionality of the tumor cells, because they show a strong association with the expression of certain hormones (see below). The cells of this organized tumor tissue mostly display an eosinophilic cytoplasm and mainly round uniform nuclei with hyperchromatic (pepper and salt) chromatin, small inconspicuous nucleoli, and a rather low mitotic rate. PanNETs with oncocytic, lipid-rich (clear), or hepatoid cells or with pleomorphic nuclei are rare . PanNETs are usually well demarcated from the surrounding parenchyma, when they are small (< 1 cm). When they are larger, they can widely infiltrate into the adjacent acinar tissue, thereby invading vessels and nerves and entrapping preexistent islets or single ducts. Rare PanNETs show a peculiar mixture of solid or trabecular cell clusters with small non-neoplastic ducts often embedded in the sclerotic stroma [ – , , ]. The immunohistochemical profile of PanNETs that is essential to establish the diagnosis includes the expression of cytokeratin, synaptophysin, and chromogranin A and Ki67 (Fig. ) . In the case of a PanNET, G3 staining for p53 and RB1 is highly recommended to distinguish these tumors from PanNECs (Fig. ). The staining of peptide hormones, of the somatostatin receptor 2A (SST2) or the site-specific transcription factor ISLET-1, is recommended where the diagnosis needs it to be complete [ , , , ]. Labeling for cytokeratin proves the epithelial nature of a NEN in cases where a neuroectodermal tumor such as a paraganglioma must be excluded . Diffuse and intense cytoplasmic expression of synaptophysin and chromogranin A and nuclear staining for insulinoma-associated 1 (INSM1) (Fig. ) reveals the tumor’s neuroendocrine differentiation, the common denominator of NENs . The labeling of the nuclei with Ki67 is the best way to accurately determine the proliferative activity of the tumor cells, and this method has largely replaced the counting of mitoses. The exact assessment of the proportion of Ki67-labeled cells as the basis for the calculation of the Ki67 index has emerged as indispensable for the prognostic and therapeutic stratification of PanNETs . The stratification is based on a three-tired grading that separates G1, G2, and G3 PanNETs according to their Ki67 index (Table ) . PanNETs G3, which represent a new category among NENs, have no defined upper mitotic or Ki67 rate limit; however, usually their mitotic rate and Ki67 index do not exceed 20/10 HPF and 50% (Fig. ), respectively. Most PanNETs G3 appear to develop from a low-grade NET, since they often manifest themselves as metastases in patients with a prior history of a G1 or G2 PanNET . PanNETs produce peptide hormones which are orthotopic (insulin, glucagon, somatostatin, pancreatic polypeptide, and serotonin) or ectopic (gastrin, vasoactive intestinal polypeptide; VIP, adrenocorticotropin; ACTH; and others) to the pancreas and can be identified by specific antibodies. Approximately 30% of PanNETs are functioning, meaning that the peptide hormone which is produced and secreted also causes a hormonal syndrome. The functioning NETs of the pancreas include insulinoma, glucagonoma, gastrinoma, VIPoma, GRHoma, ACTHoma, or PanNET with carcinoid syndrome and serotonin expression . Whether PanNETs producing and secreting somatostatin can cause a somatostatin syndrome, as described in 1979, is currently under debate, since a study from 2008 was unable to find any syndrome in somatostatin-positive PanNETs or duodenal NETs, and the evidence given in the most recent study is inconclusive . Rarely, there are PanNETs with hypercalcemia, which may produce calcitonin, but the occurrence of hypercalcemia is not necessarily tied to calcitonin secretion . The most frequent functioning PanNETs are insulinomas, which in 90% are small (< 2 cm) and behave benignly . All the other functioning PanNETs are rare and usually show a malignant behavior, especially the tumors with ACTH production and Cushing syndrome. All PanNETs that produce and also secrete a hormone, but are non-syndromic, fall into the category of non-functioning PanNETs and account for about 70% of all PanNETs. Forty percent of PanNETs are multihormonal and are generally found among the non-functioning tumors . Interestingly, the rare (about 20%) malignant insulinomas seem to start in the pancreas as non-functioning multihormonal tumors, with only single insulin cells, but become syndromic after large liver metastases have developed in which the number of insulin-secreting cells is sufficient to produce a hypoglycemic syndrome (GK, personal observation). It is increasingly noticed that the production of some hormones is associated with certain histological patterns of the PanNETs. Thus, a solid-nested pattern with amyloid is usually associated with the expression of insulin and is found in insulinomas . Tumors with a trabecular-reticulated and often cystic pattern express glucagon . Tumors with a solid paraganglioma-like or solid-microglandular pattern with psammoma-bodies usually contain somatostatin-positive cells , and a trabecular sclerosing pattern of a tumor adjacent to the main duct commonly associates with serotonin positive cells . In some multihormonal PanNETs with a clear separation of a solid from a trabecular pattern, each pattern may have its own hormone production. Membranous SST2 expression on tumor cells is needed to visualize and treat the tumors with radioisotope-labeled somatostatin . It can be detected in most PanNETs, except for insulinomas which express SST2 in only 50% of the tumors and rather express GLP1R than SST2 . If a primary tumor is SST2-positive, it can be assumed that later metastases are also positive and are therefore detectable in the follow-up. Very helpful for the localization of a primary in the pancreas (or duodenum) in case of a liver metastasis with unknown primary is the nuclear expression of the transcription factor ISLET-1 [ , , ]. The molecular profile of PanNET is profoundly different from that of pancreatic ductal adenocarcinomas (PDAC). Key drivers of PanNETs are alterations in MEN1 and ATRX or DAXX [ – ], while abnormalities of KRAS, TP53 , CDKN2A, and SMAD4 are the drivers in PDACs [ – ]. MEN1 is a tumor suppressor gene located on chromosome 11 encoding for the protein Menin, which is an important factor for the regulation of chromatin remodeling [ – ] and seems to play a key role in tumor initiation as MEN1 alterations are already detectable in pancreatic microtumors and as germline mutation in the genetic syndrome multiple endocrine neoplasia type 1 (MEN1, see below). MEN1 is known to interact with genes of chromatin modifications, altered telomere length, DNA damage repair, and mTOR signaling, which are the four main genetic pathways involved in the development of pancreatic NETs . ATRX and DAXX are also genes involved in chromatin remodeling with a high frequency of alterations (40%) in PanNETs . Inactivating mutations of ATRX or DAXX are associated with an alternative lengthening of telomeres (ALT), a telomerase-independent telomere maintenance mechanism . ATRX/DAXX alterations seem to be late events in tumorigenesis as they are only detectable in large fully developed NET but not in microadenomas . Furthermore, NET with ATRX/DAXX mutations appear to be associated with a poor prognosis compared to ATRX/DAXX wildtype tumors [ , , ]. ATRX and DAXX are genes that are strongly involved in different epigenetic mechanisms regulating gene expressions per methylation pattern . MEN1 , ATRX , and DAXX alterations are associated with a pathological protein expression which is detectable by immunohistochemistry The third relevant cluster of commonly mutated genes are alterations in genes belonging to the mTOR pathway, which are mutated in about 15% of the pancreatic NET , mostly affecting PTEN , TSC1 , and TSC2 . Less recurrent mutations detected in pancreatic NET involve ATM , YY1 , and MUTYH [ , , ]. TP53 and RB1 are usually wild types in PanNET, in contrast to PanNEC [ , , ]. A recent study focusing on the DNA-methylation profile distinguished between alpha-like, beta-like, and intermediate PanNET clusters that differed in prognosis . Most NETs are sporadic and solitary tumors. However, approximately 10% of pancreatic NETs develop in association with genetic syndromes and then often manifest as multiple tumors, usually also affecting extrapancreatic organs. The most common syndrome is MEN 1 (see above), followed by the syndromes of Von-Hippel-Lindau, neurofibromatosis type 1, and tuberous sclerosis, with germline mutations in the genes MEN1 , VHL , NF1 , and TSC2 , respectively. Functioning PanNETs occur predominantly in MEN1, in which they account for about 30% of the cases and include mainly insulinomas and duodenal gastrinomas. PanNEC PanNECs are rare high-grade pancreatic neoplasms accounting for a maximum of 10% of PanNENs. They arise as sporadic, solitary, and non-functioning neoplasms, and have not been observed in association with genetic syndromes . An association with smoking can be suspected, has so far however not been established. The histological profile of PanNECs is characterized by diffuse solid sheet-like and/or a more irregularly nested pattern (Fig. ). Common to both patterns are geographical necrosis. NECs with diffuse sheet-like patterns are often composed of highly atypical small- to medium-sized cells that have a scant cytoplasm and hyperchromatic nuclei with inconspicuous nucleoli and focal nuclear molding. NECs with more nested patterns are mostly composed of larger cells whose cytoplasm is rather well-developed and eosinophilic, carrying a polymorphous nucleus with a prominent nucleolus within vesicular chromatin delimited by a delicate nuclear membrane. Mitoses, including atypical mitoses, are common. PanNECs are usually indistinguishable from NECs of other sites by histology alone [ , , , ]. The immunohistochemical profile of PanNECs that is essential to establish the diagnosis includes the expression of cytokeratin (CK), synaptophysin, INSM1 (Fig. ), chromogranin A, and Ki67, as well as the overexpression/loss of p53 (Fig. ) and the loss of nuclear RB1 staining [ , , , ]. PanNECs express CK8 and 18. In small cell type PanNECs, CK labeling may show a punctuate pattern, and in exceptional cases, CK labeling can even be lacking. A few PanNECs also express vimentin. Synaptophysin is typically diffusely but faintly and somewhat patchy expressed, often displaying a dot-like pattern. Chromogranin A is usually focally and scarcely expressed and may even be lacking, since neurosecretory granules, whose membranes contain chromogranin A, are rare in NEC cells. CD56 labels the membranes of PanNECs broadly, but it should be never the only neuroendocrine marker on which the diagnosis of a NEC is based, since it has a high degree of unspecificity. PanNECs are mainly ISLET-1-negative [ – , , , , , , ]. All PanNECs show a Ki67 index greater than 20%, with a mean of 50 to 60% (Table ). The molecular profile of PanNECs is characterized by TP53 and RB1 mutations which are the key drivers of PanNECs as well as of extrapancreatic NECs [ , , , , – ]. Later studies additionally identified KRAS as a third driver in PanNEC , suggesting a potential relationship to PDAC. Next-generation sequencing studies using larger gene panels revealed no further recurrent gene mutations in PanNECs and no clear molecular differences between small and large cell subtypes . However, NECs seem to possess an organ-specific signature, since PanNECs have only KRAS mutations, while colorectal NECs have also APC mutations . TP53 and RB1 mutations are important in the distinction of PanNEC from PanNET, as they are absent in G1/G2 PanNETs and are only occasionally present in PanNETs G3 [ , , ]. Immunohistochemically, almost 70% of PanNECs overexpress p53 that reflects an underlying TP53 alteration, and show loss of RB1 nuclear staining, indicating a RB1 alteration . Unlike well-differentiated PanNETs, PanNECs retain the expression of DAXX/ATRX, since the corresponding genes are not mutated. SST2 expression is negative in 85% of the tumors . These tumors, which are negative on somatostatin radio receptor scintigraphy, are often positive on FDG-PET . Mixed neuroendocrine-non-neuroendocrine neoplasms (MiNEN) PanNENs may contain coexisting high-grade PDAC or acinar cell carcinoma. If one component exceeds 30% (an arbitrarily chosen threshold) of the total tumor cell population, such tumors are called “mixed neuroendocrine-non-neuroendocrine neoplasms (MiNENs).” If the non-neuroendocrine component is an adenocarcinoma and the neuroendocrine component presents as NEC, the old term “mixed adenoneuroendocrine neoplasm (MANEC)” can be retained [ , , ]. In a small series of pancreatic MiNENs, the neuroendocrine as well as the non-neuroendocrine component displayed poor differentiation and either a mosaic or a composite/amphicrine pattern. Single cases of published pancreatic MiNEN revealed a close relationship to PDAC , as it was also found in colorectal MiNEN , and interestingly also to its precursors, as two cases of pancreatic intraductal papillary mucinous neoplasms (IPMN) associated with NEN were reported, in which the NEN component showed GNAS mutations, typical for IPMN. In one case, the NEN component was a NEC , and in the other case, a NET . In mixed acinar-neuroendocine carcinomas, the expression of trypsin and synaptophysin can be so intense and overlapping that an amphicrine pattern can be observed. Genetically and biologically, these neoplasms are closely related to the conventional acinar cell carcinomas .
The histological profile of PanNETs is characterized by an organoid growth pattern with an ordered arrangement of cells, a variable amount of fibrotic stroma, and only rare necrotic changes (Figs. and ). Although the organoid pattern is not uniform but very diverse, the histological diversity can probably be traced back to either a solid or trabecular architecture that can be subdivided due to the composition of the stroma and its vascularization into solid-nested, solid-paraganglioma-like, solid-microglandular, and trabecular-reticulated, trabecular-gyriform, and trabecular-cystic patterns [ – ]. Interestingly, some of these patterns seem to have a relationship to the functionality of the tumor cells, because they show a strong association with the expression of certain hormones (see below). The cells of this organized tumor tissue mostly display an eosinophilic cytoplasm and mainly round uniform nuclei with hyperchromatic (pepper and salt) chromatin, small inconspicuous nucleoli, and a rather low mitotic rate. PanNETs with oncocytic, lipid-rich (clear), or hepatoid cells or with pleomorphic nuclei are rare . PanNETs are usually well demarcated from the surrounding parenchyma, when they are small (< 1 cm). When they are larger, they can widely infiltrate into the adjacent acinar tissue, thereby invading vessels and nerves and entrapping preexistent islets or single ducts. Rare PanNETs show a peculiar mixture of solid or trabecular cell clusters with small non-neoplastic ducts often embedded in the sclerotic stroma [ – , , ]. The immunohistochemical profile of PanNETs that is essential to establish the diagnosis includes the expression of cytokeratin, synaptophysin, and chromogranin A and Ki67 (Fig. ) . In the case of a PanNET, G3 staining for p53 and RB1 is highly recommended to distinguish these tumors from PanNECs (Fig. ). The staining of peptide hormones, of the somatostatin receptor 2A (SST2) or the site-specific transcription factor ISLET-1, is recommended where the diagnosis needs it to be complete [ , , , ]. Labeling for cytokeratin proves the epithelial nature of a NEN in cases where a neuroectodermal tumor such as a paraganglioma must be excluded . Diffuse and intense cytoplasmic expression of synaptophysin and chromogranin A and nuclear staining for insulinoma-associated 1 (INSM1) (Fig. ) reveals the tumor’s neuroendocrine differentiation, the common denominator of NENs . The labeling of the nuclei with Ki67 is the best way to accurately determine the proliferative activity of the tumor cells, and this method has largely replaced the counting of mitoses. The exact assessment of the proportion of Ki67-labeled cells as the basis for the calculation of the Ki67 index has emerged as indispensable for the prognostic and therapeutic stratification of PanNETs . The stratification is based on a three-tired grading that separates G1, G2, and G3 PanNETs according to their Ki67 index (Table ) . PanNETs G3, which represent a new category among NENs, have no defined upper mitotic or Ki67 rate limit; however, usually their mitotic rate and Ki67 index do not exceed 20/10 HPF and 50% (Fig. ), respectively. Most PanNETs G3 appear to develop from a low-grade NET, since they often manifest themselves as metastases in patients with a prior history of a G1 or G2 PanNET . PanNETs produce peptide hormones which are orthotopic (insulin, glucagon, somatostatin, pancreatic polypeptide, and serotonin) or ectopic (gastrin, vasoactive intestinal polypeptide; VIP, adrenocorticotropin; ACTH; and others) to the pancreas and can be identified by specific antibodies. Approximately 30% of PanNETs are functioning, meaning that the peptide hormone which is produced and secreted also causes a hormonal syndrome. The functioning NETs of the pancreas include insulinoma, glucagonoma, gastrinoma, VIPoma, GRHoma, ACTHoma, or PanNET with carcinoid syndrome and serotonin expression . Whether PanNETs producing and secreting somatostatin can cause a somatostatin syndrome, as described in 1979, is currently under debate, since a study from 2008 was unable to find any syndrome in somatostatin-positive PanNETs or duodenal NETs, and the evidence given in the most recent study is inconclusive . Rarely, there are PanNETs with hypercalcemia, which may produce calcitonin, but the occurrence of hypercalcemia is not necessarily tied to calcitonin secretion . The most frequent functioning PanNETs are insulinomas, which in 90% are small (< 2 cm) and behave benignly . All the other functioning PanNETs are rare and usually show a malignant behavior, especially the tumors with ACTH production and Cushing syndrome. All PanNETs that produce and also secrete a hormone, but are non-syndromic, fall into the category of non-functioning PanNETs and account for about 70% of all PanNETs. Forty percent of PanNETs are multihormonal and are generally found among the non-functioning tumors . Interestingly, the rare (about 20%) malignant insulinomas seem to start in the pancreas as non-functioning multihormonal tumors, with only single insulin cells, but become syndromic after large liver metastases have developed in which the number of insulin-secreting cells is sufficient to produce a hypoglycemic syndrome (GK, personal observation). It is increasingly noticed that the production of some hormones is associated with certain histological patterns of the PanNETs. Thus, a solid-nested pattern with amyloid is usually associated with the expression of insulin and is found in insulinomas . Tumors with a trabecular-reticulated and often cystic pattern express glucagon . Tumors with a solid paraganglioma-like or solid-microglandular pattern with psammoma-bodies usually contain somatostatin-positive cells , and a trabecular sclerosing pattern of a tumor adjacent to the main duct commonly associates with serotonin positive cells . In some multihormonal PanNETs with a clear separation of a solid from a trabecular pattern, each pattern may have its own hormone production. Membranous SST2 expression on tumor cells is needed to visualize and treat the tumors with radioisotope-labeled somatostatin . It can be detected in most PanNETs, except for insulinomas which express SST2 in only 50% of the tumors and rather express GLP1R than SST2 . If a primary tumor is SST2-positive, it can be assumed that later metastases are also positive and are therefore detectable in the follow-up. Very helpful for the localization of a primary in the pancreas (or duodenum) in case of a liver metastasis with unknown primary is the nuclear expression of the transcription factor ISLET-1 [ , , ]. The molecular profile of PanNET is profoundly different from that of pancreatic ductal adenocarcinomas (PDAC). Key drivers of PanNETs are alterations in MEN1 and ATRX or DAXX [ – ], while abnormalities of KRAS, TP53 , CDKN2A, and SMAD4 are the drivers in PDACs [ – ]. MEN1 is a tumor suppressor gene located on chromosome 11 encoding for the protein Menin, which is an important factor for the regulation of chromatin remodeling [ – ] and seems to play a key role in tumor initiation as MEN1 alterations are already detectable in pancreatic microtumors and as germline mutation in the genetic syndrome multiple endocrine neoplasia type 1 (MEN1, see below). MEN1 is known to interact with genes of chromatin modifications, altered telomere length, DNA damage repair, and mTOR signaling, which are the four main genetic pathways involved in the development of pancreatic NETs . ATRX and DAXX are also genes involved in chromatin remodeling with a high frequency of alterations (40%) in PanNETs . Inactivating mutations of ATRX or DAXX are associated with an alternative lengthening of telomeres (ALT), a telomerase-independent telomere maintenance mechanism . ATRX/DAXX alterations seem to be late events in tumorigenesis as they are only detectable in large fully developed NET but not in microadenomas . Furthermore, NET with ATRX/DAXX mutations appear to be associated with a poor prognosis compared to ATRX/DAXX wildtype tumors [ , , ]. ATRX and DAXX are genes that are strongly involved in different epigenetic mechanisms regulating gene expressions per methylation pattern . MEN1 , ATRX , and DAXX alterations are associated with a pathological protein expression which is detectable by immunohistochemistry The third relevant cluster of commonly mutated genes are alterations in genes belonging to the mTOR pathway, which are mutated in about 15% of the pancreatic NET , mostly affecting PTEN , TSC1 , and TSC2 . Less recurrent mutations detected in pancreatic NET involve ATM , YY1 , and MUTYH [ , , ]. TP53 and RB1 are usually wild types in PanNET, in contrast to PanNEC [ , , ]. A recent study focusing on the DNA-methylation profile distinguished between alpha-like, beta-like, and intermediate PanNET clusters that differed in prognosis . Most NETs are sporadic and solitary tumors. However, approximately 10% of pancreatic NETs develop in association with genetic syndromes and then often manifest as multiple tumors, usually also affecting extrapancreatic organs. The most common syndrome is MEN 1 (see above), followed by the syndromes of Von-Hippel-Lindau, neurofibromatosis type 1, and tuberous sclerosis, with germline mutations in the genes MEN1 , VHL , NF1 , and TSC2 , respectively. Functioning PanNETs occur predominantly in MEN1, in which they account for about 30% of the cases and include mainly insulinomas and duodenal gastrinomas.
PanNECs are rare high-grade pancreatic neoplasms accounting for a maximum of 10% of PanNENs. They arise as sporadic, solitary, and non-functioning neoplasms, and have not been observed in association with genetic syndromes . An association with smoking can be suspected, has so far however not been established. The histological profile of PanNECs is characterized by diffuse solid sheet-like and/or a more irregularly nested pattern (Fig. ). Common to both patterns are geographical necrosis. NECs with diffuse sheet-like patterns are often composed of highly atypical small- to medium-sized cells that have a scant cytoplasm and hyperchromatic nuclei with inconspicuous nucleoli and focal nuclear molding. NECs with more nested patterns are mostly composed of larger cells whose cytoplasm is rather well-developed and eosinophilic, carrying a polymorphous nucleus with a prominent nucleolus within vesicular chromatin delimited by a delicate nuclear membrane. Mitoses, including atypical mitoses, are common. PanNECs are usually indistinguishable from NECs of other sites by histology alone [ , , , ]. The immunohistochemical profile of PanNECs that is essential to establish the diagnosis includes the expression of cytokeratin (CK), synaptophysin, INSM1 (Fig. ), chromogranin A, and Ki67, as well as the overexpression/loss of p53 (Fig. ) and the loss of nuclear RB1 staining [ , , , ]. PanNECs express CK8 and 18. In small cell type PanNECs, CK labeling may show a punctuate pattern, and in exceptional cases, CK labeling can even be lacking. A few PanNECs also express vimentin. Synaptophysin is typically diffusely but faintly and somewhat patchy expressed, often displaying a dot-like pattern. Chromogranin A is usually focally and scarcely expressed and may even be lacking, since neurosecretory granules, whose membranes contain chromogranin A, are rare in NEC cells. CD56 labels the membranes of PanNECs broadly, but it should be never the only neuroendocrine marker on which the diagnosis of a NEC is based, since it has a high degree of unspecificity. PanNECs are mainly ISLET-1-negative [ – , , , , , , ]. All PanNECs show a Ki67 index greater than 20%, with a mean of 50 to 60% (Table ). The molecular profile of PanNECs is characterized by TP53 and RB1 mutations which are the key drivers of PanNECs as well as of extrapancreatic NECs [ , , , , – ]. Later studies additionally identified KRAS as a third driver in PanNEC , suggesting a potential relationship to PDAC. Next-generation sequencing studies using larger gene panels revealed no further recurrent gene mutations in PanNECs and no clear molecular differences between small and large cell subtypes . However, NECs seem to possess an organ-specific signature, since PanNECs have only KRAS mutations, while colorectal NECs have also APC mutations . TP53 and RB1 mutations are important in the distinction of PanNEC from PanNET, as they are absent in G1/G2 PanNETs and are only occasionally present in PanNETs G3 [ , , ]. Immunohistochemically, almost 70% of PanNECs overexpress p53 that reflects an underlying TP53 alteration, and show loss of RB1 nuclear staining, indicating a RB1 alteration . Unlike well-differentiated PanNETs, PanNECs retain the expression of DAXX/ATRX, since the corresponding genes are not mutated. SST2 expression is negative in 85% of the tumors . These tumors, which are negative on somatostatin radio receptor scintigraphy, are often positive on FDG-PET .
PanNENs may contain coexisting high-grade PDAC or acinar cell carcinoma. If one component exceeds 30% (an arbitrarily chosen threshold) of the total tumor cell population, such tumors are called “mixed neuroendocrine-non-neuroendocrine neoplasms (MiNENs).” If the non-neuroendocrine component is an adenocarcinoma and the neuroendocrine component presents as NEC, the old term “mixed adenoneuroendocrine neoplasm (MANEC)” can be retained [ , , ]. In a small series of pancreatic MiNENs, the neuroendocrine as well as the non-neuroendocrine component displayed poor differentiation and either a mosaic or a composite/amphicrine pattern. Single cases of published pancreatic MiNEN revealed a close relationship to PDAC , as it was also found in colorectal MiNEN , and interestingly also to its precursors, as two cases of pancreatic intraductal papillary mucinous neoplasms (IPMN) associated with NEN were reported, in which the NEN component showed GNAS mutations, typical for IPMN. In one case, the NEN component was a NEC , and in the other case, a NET . In mixed acinar-neuroendocine carcinomas, the expression of trypsin and synaptophysin can be so intense and overlapping that an amphicrine pattern can be observed. Genetically and biologically, these neoplasms are closely related to the conventional acinar cell carcinomas .
Pitfalls in the diagnosis of PanNENs concern mainly the confusion of NECs with NETs G3 and of NENs with a variety of non-NENs such as acinar cell carcinoma, solid pseudopapillary neoplasm, pancreatic paragangliomas, PDACs with neuroendocrine cells, and subsets of non-NENs of epithelial and mesenchymal origin with neuroendocrine features (Fig. ) as well as tumor-like lesions . PanNET G3 versus NEC The delimitation between NETs and NECs is important as their clinical management differs fundamentally . In small biopsies and occasionally in resection specimens, the distinction between NETs G3 (also called high-grade NETs) and NECs, especially of large cell type, can be difficult. Histologically, NETs G3 mostly show an organoid solid or solid-trabecular (rarely pseudoglandular) pattern that is well distinguishable from the usually sheet-like architecture of small cell NECs, but difficult to distinguish from the nested architecture of most large-cell NECs [ , , , ], particularly in biopsies. Hyalinized versus desmoplastic stroma and regular versus random vessel pattern are also criteria for the distinction of NET from NEC, but are usually not helpful in biopsies. Helpful are cytological criteria, with more polymorphous nuclei with conspicuous nucleoli in NECs, and immunohistochemical markers, including SST2, p53, and RB1. The vast majority of NETs express SST2 and show a normal (weak and no more than 20%) expression of p53 and RB1 (no complete loss) [ , , ]. In contrast, only 16% of NEC express SST2 and show an abnormal expression of p53 and/or RB1 in 70% of the cases . If the distinction is still difficult, the testing for the expression of hormones can help, as most NETs, but no NECs, are hormone-producing tumors . Acinar cell carcinoma versus PanNET G3 Acinar cell carcinomas are histologically recognized by their more or less striking acinar pattern, faint PAS positivity, round nuclei with prominent nucleoli, high mitotic rate, and scant fibrous stroma. However, some tumors displaying solid or trabecular patterns are very reminiscent of NETs . In addition, approximately 40% of acinar cell carcinomas express synaptophysin and chromogranin A, and some of these qualify as mixed acinar-neuroendocrine (ductal) carcinomas, with an intimate and amphicrine mixture of the two components. In the latter tumors, the co-expression of trypsin and synaptophysin can reach such an extent in all tumor cells that, if one only stains for synaptophysin, the tumor, which labels diffusely for synaptophysin, is easily misdiagnosed as NETs (especially NETs G3 when the mitotic activity is high) and only correctly recognized as mixed acinar-neuroendocrine carcinoma when trypsin (and/or BCL10) is added to the marker panel (Fig. ) [ – ]. Very helpful in these cases are SST2 and ISLET-1, since both markers are negative in acinar cell carcinomas with neuroendocrine features (authors’ personal observation). Solid pseudopapillary neoplasm versus PanNET Solid pseudopapillary neoplasms may present with a predominantly solid and monomorphous cell pattern and the expression of synaptophysin and cytokeratin. These cases, which very much mimic a NET, however, differ from NETs by a lack of chromogranin A staining and a positive nuclear (and cytoplasmic) labeling for ß-catenin, indicating a CTNNB1 -mutation found in almost all SPN [ – ]. Furthermore, it is helpful that the synaptophysin and cytokeratin staining in SPN is rarely diffuse, but usually patchy . Paraganglioma-like PanNET versus true paraganglioma A small fraction of PanNETs shows a solid, paraganglioma-like histology and may therefore mimic rare paragangliomas occurring in or, more commonly, at the pancreas . While paragangliomas usually show a benign behavior, paraganglioma-like PanNETs do not differ biologically from the other PanNETs. The key markers for the distinction of the two entities are CKs and GATA3 . GATA3 is positive in paragangliomas and negative in PanNETs, while CKs are negative in paragangliomas and positive in PanNETs. Furthermore, paragangliomas do not infiltrate into the pancreatic tissue, while paraganglioma-like PanNETs usually do . Ductal adenocarcinoma with islet cells versus MiNEN The distinction of conventional PDAC intimately associated with islet cells from MiNEN is in most cases not a problem. First, the WHO classification requires that the neuroendocrine component of MiNENs exceeds 30% of the tumor cell population . In PDAC, the number of islet cells that combine with duct-like glands of a PDAC to form ductal-neuroendocrine complexes is less than 30% of the tumor cell population. In addition, they show no Ki67 labeling as a sign of proliferation in contrast to PDAC cells. Moreover, immunostains for pancreatic hormones identify the cells that associated with PDAC structures as islet cell types. Finally, the metastases of these PDACs never contain neuroendocrine cells of the type found in the pancreas . Mesenchymal and non-epithelial neoplasms versus NEC Recently a range of mesenchymal and non-epithelial mimickers of neuroendocrine neoplasms, mainly of the large-cell NECs, have been described in a large consultation series . Of particular interest and rather new among these NEN mimickers are tumors from the Ewing Sarcomas group, desmoplastic small round cell tumors, epithelioid neoplasms with FUS-CREM gene fusions, epithelioid sarcomas, synovial sarcomas, SMARCA4- and SMARCB1-deficient neoplasms (Fig. ), clear cell sarcomas of the gastrointestinal tract, alveolar soft part sarcomas, solitary fibrous tumors, chordomas, melanomas, and sclerosing epithelioid mesenchymal neoplasms. Six of these neoplasms were located in the pancreas and included Ewing sarcoma, SMARCB1(INI1)-deficient neoplasms (Fig. ), melanoma, and sclerosing epithelioid sclerosing neoplasms . To unmask these neoplasms as NEN mimickers, the testing of the key markers CD99, INI1, and S100 is necessary in cases of Ewing sarcomas, SMARCB1-deficient neoplasms, and melanomas, respectively. The sclerosing epithelioid mesenchymal neoplasm of the pancreas is an exceedingly rare pancreatic tumor that has recently been proposed as a new tumor entity with only single reported cases in the literature . The synaptophysin expression in mesenchymal/non-epithelial NEN mimickers is mostly patchy with co-expression of chromogranin A in one-third of the cases . Tumor-like lesions Tumor-like lesions as NEN mimickers are islet cell aggregates in specimens of chronic pancreatitis, particularly of the obstructive type of pancreatitis associated with duct occluding tumors. The islet clusters that are found in fibrotic and/or lipomatous tissue devoid of acinar cells can histologically imitate an infiltrating PanNET. To avoid a diagnostic pitfall, the islet cell nature of the cell aggregates has to be demonstrated. This is easily done by immunostainings for insulin and glucagon, which reveal the normal non-random distribution of the two islet types in the islet cell clusters that profoundly differs from the monohormonal expression in most NETs . The same applies to the PP-rich islet clusters in the posterior-caudal and uncinate lobe of the pancreatic head, when this part of the pancreas is resected with a PDAC. Again, these aggregates which can have a NET-like appearance reveal their normal islet cell type composition when tested for pancreatic polypeptide .
The delimitation between NETs and NECs is important as their clinical management differs fundamentally . In small biopsies and occasionally in resection specimens, the distinction between NETs G3 (also called high-grade NETs) and NECs, especially of large cell type, can be difficult. Histologically, NETs G3 mostly show an organoid solid or solid-trabecular (rarely pseudoglandular) pattern that is well distinguishable from the usually sheet-like architecture of small cell NECs, but difficult to distinguish from the nested architecture of most large-cell NECs [ , , , ], particularly in biopsies. Hyalinized versus desmoplastic stroma and regular versus random vessel pattern are also criteria for the distinction of NET from NEC, but are usually not helpful in biopsies. Helpful are cytological criteria, with more polymorphous nuclei with conspicuous nucleoli in NECs, and immunohistochemical markers, including SST2, p53, and RB1. The vast majority of NETs express SST2 and show a normal (weak and no more than 20%) expression of p53 and RB1 (no complete loss) [ , , ]. In contrast, only 16% of NEC express SST2 and show an abnormal expression of p53 and/or RB1 in 70% of the cases . If the distinction is still difficult, the testing for the expression of hormones can help, as most NETs, but no NECs, are hormone-producing tumors .
Acinar cell carcinomas are histologically recognized by their more or less striking acinar pattern, faint PAS positivity, round nuclei with prominent nucleoli, high mitotic rate, and scant fibrous stroma. However, some tumors displaying solid or trabecular patterns are very reminiscent of NETs . In addition, approximately 40% of acinar cell carcinomas express synaptophysin and chromogranin A, and some of these qualify as mixed acinar-neuroendocrine (ductal) carcinomas, with an intimate and amphicrine mixture of the two components. In the latter tumors, the co-expression of trypsin and synaptophysin can reach such an extent in all tumor cells that, if one only stains for synaptophysin, the tumor, which labels diffusely for synaptophysin, is easily misdiagnosed as NETs (especially NETs G3 when the mitotic activity is high) and only correctly recognized as mixed acinar-neuroendocrine carcinoma when trypsin (and/or BCL10) is added to the marker panel (Fig. ) [ – ]. Very helpful in these cases are SST2 and ISLET-1, since both markers are negative in acinar cell carcinomas with neuroendocrine features (authors’ personal observation).
Solid pseudopapillary neoplasms may present with a predominantly solid and monomorphous cell pattern and the expression of synaptophysin and cytokeratin. These cases, which very much mimic a NET, however, differ from NETs by a lack of chromogranin A staining and a positive nuclear (and cytoplasmic) labeling for ß-catenin, indicating a CTNNB1 -mutation found in almost all SPN [ – ]. Furthermore, it is helpful that the synaptophysin and cytokeratin staining in SPN is rarely diffuse, but usually patchy .
A small fraction of PanNETs shows a solid, paraganglioma-like histology and may therefore mimic rare paragangliomas occurring in or, more commonly, at the pancreas . While paragangliomas usually show a benign behavior, paraganglioma-like PanNETs do not differ biologically from the other PanNETs. The key markers for the distinction of the two entities are CKs and GATA3 . GATA3 is positive in paragangliomas and negative in PanNETs, while CKs are negative in paragangliomas and positive in PanNETs. Furthermore, paragangliomas do not infiltrate into the pancreatic tissue, while paraganglioma-like PanNETs usually do .
The distinction of conventional PDAC intimately associated with islet cells from MiNEN is in most cases not a problem. First, the WHO classification requires that the neuroendocrine component of MiNENs exceeds 30% of the tumor cell population . In PDAC, the number of islet cells that combine with duct-like glands of a PDAC to form ductal-neuroendocrine complexes is less than 30% of the tumor cell population. In addition, they show no Ki67 labeling as a sign of proliferation in contrast to PDAC cells. Moreover, immunostains for pancreatic hormones identify the cells that associated with PDAC structures as islet cell types. Finally, the metastases of these PDACs never contain neuroendocrine cells of the type found in the pancreas .
Recently a range of mesenchymal and non-epithelial mimickers of neuroendocrine neoplasms, mainly of the large-cell NECs, have been described in a large consultation series . Of particular interest and rather new among these NEN mimickers are tumors from the Ewing Sarcomas group, desmoplastic small round cell tumors, epithelioid neoplasms with FUS-CREM gene fusions, epithelioid sarcomas, synovial sarcomas, SMARCA4- and SMARCB1-deficient neoplasms (Fig. ), clear cell sarcomas of the gastrointestinal tract, alveolar soft part sarcomas, solitary fibrous tumors, chordomas, melanomas, and sclerosing epithelioid mesenchymal neoplasms. Six of these neoplasms were located in the pancreas and included Ewing sarcoma, SMARCB1(INI1)-deficient neoplasms (Fig. ), melanoma, and sclerosing epithelioid sclerosing neoplasms . To unmask these neoplasms as NEN mimickers, the testing of the key markers CD99, INI1, and S100 is necessary in cases of Ewing sarcomas, SMARCB1-deficient neoplasms, and melanomas, respectively. The sclerosing epithelioid mesenchymal neoplasm of the pancreas is an exceedingly rare pancreatic tumor that has recently been proposed as a new tumor entity with only single reported cases in the literature . The synaptophysin expression in mesenchymal/non-epithelial NEN mimickers is mostly patchy with co-expression of chromogranin A in one-third of the cases .
Tumor-like lesions as NEN mimickers are islet cell aggregates in specimens of chronic pancreatitis, particularly of the obstructive type of pancreatitis associated with duct occluding tumors. The islet clusters that are found in fibrotic and/or lipomatous tissue devoid of acinar cells can histologically imitate an infiltrating PanNET. To avoid a diagnostic pitfall, the islet cell nature of the cell aggregates has to be demonstrated. This is easily done by immunostainings for insulin and glucagon, which reveal the normal non-random distribution of the two islet types in the islet cell clusters that profoundly differs from the monohormonal expression in most NETs . The same applies to the PP-rich islet clusters in the posterior-caudal and uncinate lobe of the pancreatic head, when this part of the pancreas is resected with a PDAC. Again, these aggregates which can have a NET-like appearance reveal their normal islet cell type composition when tested for pancreatic polypeptide .
The diagnostic and therapeutic management of PanNENs has very much improved during the last two decades in which NET centers have emerged, supported by societies such as the European Neuroendocrine Tumor Society. This development was accompanied and supported by the steady improvement of classifications, diagnostic criteria, and prognostic assessments. The basis of all, the morphology, is still the starting point of diagnosis. The diagnostic criteria for assessing PanNEN histology have been refined and expanded at the same time in an attempt to capture the tumor´s individuality, to which our attention is more and more directed. This path is continued by the use of biomarker immunohistology that reveals further properties of tumor cells which help to improve classification, precisely record proliferative activity, specify therapeutic approaches, and determine function. The final approach to the diagnosis is the genetic evaluation of the tumor which is currently increasingly integrated into our diagnostic pathways. The correct application of all the diagnostic criteria should protect us from misdiagnosis and pitfalls. Nevertheless, it is important to know and study the special pitfalls of NENs to avoid their potential confusion with NEN in general and PanNEN in particular.
|
Emergency Position Recovery Using Forward Kinematics in Robotic Patient Positioning Systems for Radiosurgery | 108f079e-75d0-4413-901f-ed8b40936797 | 11861796 | Surgical Procedures, Operative[mh] | Radiosurgery, particularly in the treatment of brain tumors, has undergone significant evolution, characterized by remarkable advancements in precision and safety. A pivotal moment occurred in the 1960s with the introduction of the Gamma Knife, a non-invasive technique that provided a groundbreaking alternative to traditional neurosurgery. By delivering highly focused radiation with exceptional accuracy, it became especially effective for treating tumors near sensitive brain structures, where surgical techniques pose significant risks . Stereotactic radiosurgery (SRS) is now a cornerstone for intracranial tumor treatment, offering a single-dose, highly precise approach that minimizes damage to surrounding healthy tissue. It is widely used to treat brain metastases, with studies demonstrating its effectiveness as comparable to or better than whole-brain radiation therapy (WBRT) in certain cases. Moreover, SRS alone significantly reduces the risk of memory and cognitive side effects often associated with WBRT . Advanced technologies like the Gamma Knife, CyberKnife, and stereotactic linear accelerators have further refined the accuracy and safety of SRS, making it an indispensable tool in modern neurosurgery . The effectiveness of the Gamma Knife lies in its ability to minimize collateral damage to surrounding healthy tissues, a critical concern in neurosurgery, thereby significantly improving patient outcomes. Over the years, enhancements in imaging techniques—such as higher-resolution MRI and CT scans—and sophisticated treatment planning software have further increased the precision of the Gamma Knife. These advancements have enabled more complex and accurate treatment strategies for previously inoperable brain tumors . However, challenges such as optimizing treatment for dynamic anatomical changes during radiation delivery remain a focus of ongoing research . The origins of radiosurgery can be traced back to 1949, when Lars Leksell, a neurosurgeon at Karolinska University in Stockholm, and Bjorn Larsson, a radiobiologist, pioneered the first procedure. Their initial experiments used cyclotron-derived proton radiation, but they soon adopted cobalt-60 as a radiation source due to its effectiveness and simplicity . By 1968, the Gamma Knife, exclusively designed for intracranial radiosurgery, was completed. The device utilized cobalt-60 isotopes arranged radially to deliver highly focused gamma radiation, significantly reducing the radiation load on surrounding healthy tissues . This foundational design remains largely unchanged, evolving to incorporate advanced imaging and treatment planning systems. A key aspect of Gamma Knife radiosurgery is the use of a rigid stereotactic frame that ensures precise targeting by immobilizing the patient’s head. This approach, combined with three-dimensional (3D) imaging modalities such as CT and MRI, has further enhanced treatment accuracy. These technological advancements have cemented the Gamma Knife as the “gold standard” for intracranial radiosurgery. While other technologies like linear accelerators (LINAC) and CyberKnife systems have emerged, the Gamma Knife remains uniquely suited for the high-precision treatment of brain tumors . LINAC-based systems, for example, allow for broader applications, including body-wide radiosurgery, and incorporate frameless and robotic technologies. However, for intracranial tumors, the Gamma Knife’s unparalleled precision and accumulated clinical experience continue to set it apart . Parallel to these developments in radiosurgery, the field of robotic-assisted surgery has undergone transformative changes. The introduction of robotic systems such as the da Vinci Surgical System in 2000 revolutionized minimally invasive surgery by setting new standards for precision and control . The principles of robotic precision, which have been widely adopted in various surgical disciplines, have gradually extended to radiosurgery as well. These innovations have influenced both the design and functionality of modern radiosurgical systems, providing enhanced capabilities in terms of accuracy and reproducibility . Robotic technologies have had a profound impact on radiosurgery beyond operational precision. One of the most significant contributions has been in improving patient safety and comfort by enhancing the accuracy of tumor targeting and reducing the duration of radiation exposure. Modern radiosurgical systems now employ dual-loop control mechanisms, integrating primary and secondary optical encoders, to ensure real-time feedback and adjustments during treatment . This integration is crucial for maintaining alignment and positioning accuracy, especially in dynamic clinical environments where unexpected interruptions, such as patient movement or equipment malfunction, may occur . The use of dual-encoder configurations in patient positioning systems (PPS) is an exemplary advancement, offering high-resolution feedback on motor positioning and movement. This real-time monitoring is essential for the accurate delivery of gamma beams, where even minor misalignments can lead to suboptimal treatment outcomes . Such systems leverage real-time encoder data and forward kinematics to continuously adjust the patient’s position during radiation delivery, ensuring precise alignment even during long and complex procedures . Moreover, these advanced systems are designed to respond effectively to unplanned interruptions such as power outages or emergency stops. The ability to know the exact position of each motor allows the system to resume treatment accurately from the point of interruption or safely return to a home position without compromising the patient’s treatment outcome . This reliability is critical in radiosurgery, where the margin for error is minimal. Nevertheless, several challenges remain in the optimization of robotic radiosurgery, including refining real-time feedback algorithms, improving the integration of imaging technologies, and exploring ways to minimize treatment time while maximizing precision. Our study involved a thorough review of existing research on PPS, revealing the pivotal role of advanced control systems in achieving high accuracy and reliability. For instance, a study by Johnson and Patel (2022) highlights that a PPS with dual control loops is more accurate and dependable than those with a single loop , a critical consideration for ensuring patient safety and precise treatment. Another research piece by Nguyen and Garcia (2021) underscores the importance of robust safety features in PPS, which help to reduce risks and improve patient outcomes . In addition to ongoing research, major companies such as Siemens and Samsung are actively competing to deliver superior products . Two of the most competitive PPS products currently available include the 6-DoF Robotic Couch System by gKteso GmbH—boasting an absolute accuracy of 0.5 mm—and the Hexapods 6-Axis Patient Positioning Couches for Radiotherapy by PI (Physik Instrumente), known for their precision but limited travel range . Building on these insights, we have developed our own robotic patient positioning system (RPPS) that surpasses existing offerings. Our prototype achieves an absolute accuracy of up to 0.1 mm and offers a wider travel range, positioning it as a unique and advanced solution in the field of PPS. These improvements are clinically significant: a more extensive travel range eases the setup for complex treatment angles, while heightened accuracy reduces the risk of suboptimal targeting. A key differentiator of the PPS is its forward kinematics-based approach for emergency position recovery. In the event of unexpected power loss or an emergency stop, the robotic control algorithm can identify the exact configuration of each joint based on encoder readings, thus enabling rapid and precise restoration of the patient’s treatment position. This high level of precision, flexibility, and resiliency is intended to address existing limitations in patient positioning, thereby improving treatment safety, comfort, and efficiency. Furthermore, while our current focus is on intracranial procedures, the modular design of the RPPS could be adapted for broader, body-wide applications by refining its software and hardware configurations to suit different anatomical targets. This paper is structured as follows to provide a comprehensive understanding of the topic: System Structure and Collaborative Mechanism: This section provides a detailed description of the PPS, highlighting its three primary subsystems—Linear Rail System, Linkage System, and Tabletop Assembly—and their integration with control components to achieve precise patient positioning. MathematicalModeling of Forward Kinematics: Detailed mathematical equations used to calculate the forward kinematics based on encoder feedback, including how these models aid in emergency position recovery. Materials and Methods: Technical specifications and configuration of the patient positioning system, highlighting hardware components responsible for sub-millimetric accuracy and extended travel range. System Integration and Testing: The implementation of the kinematic models into the control system and subsequent testing processes, demonstrating how the system handles emergencies and power interruptions. Results and Discussion: An analysis of the system’s performance, focusing on accuracy, reliability, and recovery time across various operational conditions. Conclusion and Future Work: The summarization of findings and potential directions for enhancing precision, emergency recovery, and body-wide treatment capabilities in robotic patient positioning systems. By comparing our approach with existing technologies and emphasizing the advantages of our forward kinematics-driven emergency position recovery, we underscore the innovation and clinical significance of a system capable of sub-millimetric precision. The ability to maintain this accuracy through unexpected events is paramount in ensuring that the therapeutic outcome is not compromised, highlighting how this work contributes to the broader advancement of robotic radiosurgery for both intracranial and potentially full-body treatments.
As shown in , the robotic patient positioning system (PPS) is a six-degree-of-freedom (DOF) robotic platform for radiosurgery. It is designed to achieve precise patient alignment by integrating three primary mechanical subsystems with real-time control capabilities: Linear Rail System: Manages horizontal movement and coarse rotation, enabling the patient to be moved into or out of the treatment area. Linkage System: Provides vertical and angular adjustments, raising or tilting the table to align the patient’s target region with the treatment beam. Tabletop Assembly: Facilitates fine pitch corrections, often using a servo-driven cam mechanism, ensuring sub-millimetric orientation for accurate targeting. These mechanical subsystems work in conjunction with a high-performance motion controller and servo drivers, which issue commands to servo motors ( M1–M6 ) on each axis. Feedback loops from primary encoders ( PE ) and secondary encoders ( SE ) ensure real-time tracking and dynamic corrections. This dual-loop architecture compensates for any mechanical compliance or backlash, maintaining precision across all axes. The collaborative mechanism of the PPS can be better understood by examining the roles of its components and their responsibilities for specific movements, as shown in . The highlighted blue sections in represent the key components responsible for specific movements: Linear Movement (Lin): The Linear Rail System is responsible for horizontal translational movement, enabling coarse adjustments to position the patient into or out of the treatment area. Rotational Movement (Rot): The rotary table allows for coarse rotational adjustments, helping align the patient’s target zone with the radiation beam. Pitching Movements (Cam): The pitching cam system provides fine angular adjustments for tilting the Tabletop Assembly, ensuring precise alignment of the treatment target. Vertical and Angular Movements (Linkages): The three linkage arms (Linkage1, Linkage2, Linkage3) are responsible for vertical and angular adjustments. They collaboratively raise or lower the table and perform tilting motions to align the patient accurately. The integration of the control system enables these movements with high precision. illustrates the detailed control mechanism. The control system operates as follows: Each subsystem is controlled by a dedicated servo motor (M1–M6), with feedback from both primary (PE1–PE6) and secondary encoders (SE1–SE6). A central motion controller coordinates the movement of all components, issuing real-time commands and processing encoder feedback to ensure precise alignment. Power supplies (PS1, PS2) provide stable 24V DC power, ensuring uninterrupted operation of the system. This integration of mechanical and control subsystems ensures that the RPPS achieves the necessary precision and reliability required for safe and effective radiosurgery.
Forward kinematics, a foundational concept in robotics and mechanical systems, revolves around the determination of the end effector’s position and orientation based on given joint parameters and link lengths. Unlike its counterpart, inverse kinematics, which seeks to find joint parameters given a desired end effector position, forward kinematics is a direct mapping from the joint space to the workspace . This computational model becomes essential in scenarios where it is necessary to predict the movement outcome based on joint values . Denavit–Hartenberg (DH) One of the most important methods to manipulate a robot based on the kinematics modeling is Denavit–Hartenberg (DH); it is a prominent approach in kinematic modeling for robot manipulation and plays a vital role in achieving effective control over robotic systems . This method enables the traversal from the base frame to the end effector frame (in our study target point) by sequentially transitioning through intermediate frames, irrespective of the robot’s dynamics or specific characteristics. It delineates the transformations required, encompassing both the rotational and translational motions of the manipulator, as illustrated in , in order to obtain the transformation matrix. Starting from that, we can generate the matrix of our robot after determining the assignment of the manipulator frames—which is defined in as : (1) T i = Rot z , θ i Trans d i Trans a i Rot x , α i T i , the final matrix, will be the result of multiple of those matrices (2) T i = c θ i − s θ i 0 0 s θ i c θ i 0 0 0 0 1 d i 0 0 0 1 1 0 0 a i 0 c α i − s α i 0 0 s α i c α i 0 0 0 0 1 This equation can be translated into matrices, and by multiplying them, we obtain the final T i as follows: (3) T i = cos θ i − cos α i sin θ i sin α i sin θ i a i cos θ i sin θ i cos α i cos θ i − sin α i cos θ i a i sin θ i 0 sin α i cos α i d i 0 0 0 1 where a i : The distance between the z i and z i + 1 axes along the x i axis. α i : The angle between the z i and z i + 1 axes along the x i axis. d i : The distance between the x i and x i + 1 axes along the z i axis. θ i : The angle between the x i and x i + 1 axes along the z i axis.
One of the most important methods to manipulate a robot based on the kinematics modeling is Denavit–Hartenberg (DH); it is a prominent approach in kinematic modeling for robot manipulation and plays a vital role in achieving effective control over robotic systems . This method enables the traversal from the base frame to the end effector frame (in our study target point) by sequentially transitioning through intermediate frames, irrespective of the robot’s dynamics or specific characteristics. It delineates the transformations required, encompassing both the rotational and translational motions of the manipulator, as illustrated in , in order to obtain the transformation matrix. Starting from that, we can generate the matrix of our robot after determining the assignment of the manipulator frames—which is defined in as : (1) T i = Rot z , θ i Trans d i Trans a i Rot x , α i T i , the final matrix, will be the result of multiple of those matrices (2) T i = c θ i − s θ i 0 0 s θ i c θ i 0 0 0 0 1 d i 0 0 0 1 1 0 0 a i 0 c α i − s α i 0 0 s α i c α i 0 0 0 0 1 This equation can be translated into matrices, and by multiplying them, we obtain the final T i as follows: (3) T i = cos θ i − cos α i sin θ i sin α i sin θ i a i cos θ i sin θ i cos α i cos θ i − sin α i cos θ i a i sin θ i 0 sin α i cos α i d i 0 0 0 1 where a i : The distance between the z i and z i + 1 axes along the x i axis. α i : The angle between the z i and z i + 1 axes along the x i axis. d i : The distance between the x i and x i + 1 axes along the z i axis. θ i : The angle between the x i and x i + 1 axes along the z i axis.
This vital system plays a crucial role in facilitating precise patient alignment in a variety of medical scenarios. Its construction and functionality embody several distinctive features that contribute to its effective performance. These characteristics span across the system’s three main components: the Linear Rail System, the Linkage System, and the Table Assembly. Each component introduces unique attributes that, when combined, result in a highly efficient, versatile, and patient-friendly positioning system. 4.1. Frame Assignments for the PPS The patient positioning system (PPS) utilizes specific frame assignments to capture its distinct components and their movements, as shown in : Base Frame ( O i ): Located at the junction of the main rails. Linear Rail Frame ( T RR ): Positioned at the end of the linear rail. Lower Linkage System Frame ( T RLB ): At the midpoint of the lower linkage arm. Upper Linkage System Frame ( T LM ): At the midpoint of the upper linkage arm. Table Rod Frame ( T P ): Centered on the table rod. Tabletop Frame ( T E ): Centered on the tabletop. The linkage frames are shown in . These frame assignments are essential for precise kinematic evaluations in the PPS. 4.1.1. Linear Rail System from T0 to TRR This subsystem governs the bidirectional movement of the PPS. It facilitates both the translational motion of the main plate relative to the base plate along a rail system, and the rotational movement of the Linkage System around the main plate, courtesy of a rotary table as shown in . The linear motion enables the patient to be safely introduced and withdrawn from the operational area. In rooms equipped with a CT scanner, the rotary table allows a complete 180° rotation of the patient for imaging purposes. Encoded motors drive both movements, ensuring the precise positioning required for such medical devices. 4.1.2. Linkage System from TRLB to TLM Representing the heart of the PPS, the Linkage System comprises four pairs of two-arm linkages that connect the PPS to the plate on the rotary table. As shown in , to maintain patient stability and alignment, precision shafts interconnect two of these four pairs. The system harnesses the power of three encoded motors to maneuver three joints, positioning the PPS within a 2D plane that spans a substantial envelope size. This robust design enables the table to be adjusted vertically to accommodate patient needs, from a lowered position facilitating access for elderly patients, to an elevated state suitable for treatment procedures . 4.1.3. Table Tob TLM to TE Located at the top of the system, the Table Assembly is linked to the top plate of the Linkage Assembly. As shown in , the table, a product of Siemens, features a carbon fiber exterior filled with resin to minimize radio interference during treatment. Given the cantilevered design of the tabletop and the associated gravitational deflection, a pitch adjustment mechanism has been integrated into the Table Assembly. This mechanism actively counteracts gravity, ensuring a consistently level operating area. 4.2. Formulating a Mathematical Model 4.2.1. Linkage Subsystem Mathematical Modeling This section primarily focuses on using geometric principles to develop a kinematic model for the linkage subsystem. illustrates the points, link lengths, and joint angles associated with this system . We assumed that all the angles are measured counterclockwise: q 1 = 0 when L 1 is on L 0 . q 2 = 0 when L 2 is along the same line of L 1 . q 3 = 0 when L 5 is along the same line with L 0 . θ 3 = 0 when L 3 is along the same line of L 2 . θ 4 = 0 when L 3 is along the same line of L 4 . θ 5 = 0 when L 4 is along the same line of L 5 . We specify a coordinate frame for each joint as follows: (4) x 0 y 0 = cos q 1 − sin q 1 sin q 1 cos q 1 x 1 y 1 + 0 0 (5) x 1 y 1 = cos q 2 − sin q 2 sin q 2 cos q 2 x 2 y 2 + L 1 0 (6) x 2 y 2 = cos θ 3 − sin θ 3 sin θ 3 cos θ 3 x 3 y 3 + L 2 0 where (7) x 0 y 0 = cos q 3 − sin q 3 sin q 3 cos q 3 x 6 y 6 + L 0 0 (8) x 6 y 6 = cos θ 5 − sin θ 5 sin θ 5 cos θ 5 x 5 y 5 + L 5 0 (9) x 5 y 5 = cos θ 4 − sin θ 4 sin θ 4 cos θ 4 x 4 y 4 + L 4 0 (10) x 4 y 4 = cos 180 − sin 180 sin 180 cos 180 x 3 y 3 + L 3 0 From Equations to and , we determined each point position as follows: Point 1 (11) x 0 y 0 1 = cos q 1 − sin q 1 sin q 1 cos q 1 0 0 + 0 0 = 0 0 Point 2 (12) x 0 y 0 2 = cos q 1 − sin q 1 sin q 1 cos q 1 L 1 0 = L 1 cos q 1 L 1 sin q 1 Point 3 (13) x 1 y 1 3 = cos q 2 − sin q 2 sin q 2 cos q 2 L 2 y 2 + L 1 0 = L 2 cos q 2 + L 1 L 2 sin q 2 (14) x 0 y 0 3 = cos q 1 − sin q 1 sin q 1 cos q 1 L 2 cos q 2 + L 1 L 2 sin q 2 = L 2 cos q 1 cos q 2 + L 1 cos q 1 − L 2 sin q 1 sin q 2 L 2 sin q 1 cos q 2 + L 1 sin q 1 + L 2 cos q 1 sin q 2 Point 6 (15) x 0 y 0 6 = cos q 3 − sin q 3 sin q 3 cos q 3 0 0 + L 0 0 = L 0 0 Point 5 (16) x 6 y 6 5 = cos θ 5 − sin θ 5 sin θ 5 cos θ 5 0 0 + L 5 0 = L 5 0 (17) x 0 y 0 5 = cos q 3 − sin q 3 sin q 3 cos q 3 L 5 0 + L 0 0 = L 5 cos q 3 + L 0 L 5 sin q 3 Therefore, Point 4 (18) x 5 y 5 4 = cos θ 4 − sin θ 4 sin θ 4 cos θ 4 0 0 + L 4 0 = L 4 0 (19) x 6 y 6 4 = cos θ 5 − sin θ 5 sin θ 5 cos θ 5 L 4 0 + L 5 0 = L 4 cos θ 5 + L 5 L 4 sin θ 5 (20) x 0 y 0 4 = cos q 3 − sin q 3 sin q 3 cos q 3 L 4 cos θ 5 + L 5 L 4 sin θ 5 + L 0 0 = L 4 cos q 3 cos θ 5 + L 5 cos q 3 − L 4 sin q 3 sin θ 5 + L 0 L 4 sin q 3 cos θ 5 + L 5 sin q 3 + L 4 cos q 3 sin θ 5 4.2.2. Intersection of Two Circles In the domain of kinematics and mechanical geometry, a well-accepted technique for determining the location of a specific point within a linkage mechanism involves the intersection of two circles . As shown in , we can pinpoint the position of Point 4 using this intersecting circles methodology. The scenario involves two circles: the first centered at Point 3 with a radius of L3, and the second centered at Point 5 with a radius of L5. The intersection of these two circles yields two potential points, namely Point 4 and Point 6. Consider two circles: Circle 1, centered at Point 3 ( C 3 = ( x 3 , y 3 ) ) with radius L 3 . Circle 2, centered at Point 5 ( C 5 = ( x 5 , y 5 ) ) with radius L 5 . The equations of these circles are as follows: (21) ( x − x 3 ) 2 + ( y − y 3 ) 2 = L 3 2 (22) ( x − x 5 ) 2 + ( y − y 5 ) 2 = L 5 2 To find the intersection points, we can subtract two circle equations and simplify these. This will give a linear equation in x and y . Out of the two intersection points, the one with the greater y -coordinate is considered the upper intersection point. We choose this as Point 4, and the other intersection point is disregarded. These represent the possible locations of the mechanical joint or linkage. However, due to the physical constraints and design of our mechanical system, the existence of Point 6 would suggest a mechanically unfeasible configuration. Therefore, Point 6 can be categorically eliminated from our options, leaving us with the valid position, Point 4. After that, to ascertain the orientation of the upper arm of the linkage mechanism , we must calculate the value of ϕ 3 . This angle ϕ 3 is a critical parameter that provides insights into the inclination of the upper arm, thereby enabling a comprehensive understanding of the system’s kinematics. The determination of ϕ 3 is a crucial step in accurately modeling and analyzing the mechanical system. (23) c = ( P 4 ( x ) − P 2 ( x ) ) 2 + ( P 4 ( y ) − P 2 ( y ) ) 2 (24) Γ = arc a 2 + b 2 − c 2 2 × a × b (25) ϕ 3 = 180 ∘ − Γ 4.3. Table Pitching Mathematical Modeling Table angle or pitching angle controlled by servo motor and CAM as shown in the following . In order to transition from point TP to point TE, it is necessary to calculate the length L10, based on the angle input as per the given and . Then, after determining all needed parameters from , we can create the DH table as follows: 4.4. Utilize the Denavit–Hartenberg (DH) Parameters With the essential parameters and points now determined, we can utilize the Denavit–Hartenberg (DH) parameters. This allows us to transition from one frame to the subsequent frame, ultimately leading us to the final transformation matrix representing our PPS. 4.4.1. DH Matrix Formulation The standard DH transformation matrix, denoted as A i , is constructed using the given parameters a , α , d , and θ : A i = cos ( θ ) − sin ( θ ) 0 0 sin ( θ ) cos ( θ ) 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 0 0 1 d 0 0 0 1 1 0 0 a 0 1 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 cos ( α ) − sin ( α ) 0 0 sin ( α ) cos ( α ) 0 0 0 0 1 4.4.2. Linear Rail Subsystem The transformation matrix for the linear rail subsystem, denoted as T o , is as follows: T o = A i | a = 0 , α = 0 , d = 0 , θ = 0 T 1 = A i | a = 0 , α = 0 , d = − L + L i n , θ = 0 (26) T 0 1 = 1 0 0 0 0 1 0 0 0 0 1 Lin − L 0 0 0 1 As shown in , the transformation matrices describe the motion and positioning of the linear rail system. 4.4.3. Rotary Base to the Height of the Linkage As shown in , two transformation matrices are derived: T R representing a rotation about the θ -axis by − π 2 and no translation along the x -axis: T R = A i | a = 0 , α = 0 , d = 0 , θ = − π 2 T R R representing a rotation about the ‘Rott’ angle and a translation along the z -axis by L R : T R R = A i | a = 0 , α = 0 , d = L R , θ = R o t (27) T R R R = 1 0 0 0 0 0 1 0 0 − 1 0 0 0 0 0 1 (28) T R R R L B = cos ( R o t ) − sin ( R o t ) 0 0 sin ( R o t ) cos ( R o t ) 0 0 0 0 1 L R 0 0 0 1 4.4.4. Linkage Subsystem The linkage subsystem of the PPS consists of multiple stages of transformation: First, no translation along the x -axis and a rotation of π 2 about the x -axis: T R L B = A i | a = 0 , α = π 2 , d = 0 , θ = 0 (29) T R L B L 1 = 1 0 0 0 0 0 − 1 0 0 1 0 0 0 0 0 1 Second, a translation of L 0 2 along the x -axis, a translation of L L along the z -axis, and a rotation of π about the z -axis: T L 1 = A i | a = L 0 2 , α = 0 , d = L L , θ = { π Next, we move the first link by controlling the first motor q 1 . This involves a translation of L 1 along the x -axis and a rotation of − π + q 1 about the z -axis: T L 2 = A i | a = L 1 , α = 0 , d = 0 , θ = p i + q 1 (30) T L 1 L 2 = − 1 0 0 − L 0 2 0 − 1 0 0 0 0 1 L L 0 0 0 1 After this, we control the second link with the second motor q 2 . This step comprises a translation of L 2 along the x -axis and a rotation of q 2 about the z -axis: T L 12 = A i | a = L 2 , α = 0 , d = 0 , θ = q 2 (31) T L 3 L M = cos ( q 2 ) − sin ( q 2 ) 0 L 2 cos ( q 2 ) sin ( q 2 ) cos ( q 2 ) 0 L 2 sin ( q 2 ) 0 0 1 0 0 0 0 1 Lastly, we move to the middle of the upper arm based on the ϕ 3 value. This involves a translation of L 3 2 along the x -axis and a rotation of ϕ 3 about the z -axis: T L M = A i | a = L 3 2 , α = 0 , d = 0 , θ = ϕ 3 (32) T L M P = cos ( θ 3 ) − sin ( θ 3 ) 0 L 3 cos ( θ 3 ) 2 sin ( θ 3 ) cos ( θ 3 ) 0 L 3 sin ( θ 3 ) 2 0 0 1 0 0 0 0 1 4.4.5. Moving to the Table Pitching Frame See for an illustration of the following transition. We transition from T L M to T P . This involves a translation of L 6 along the x -axis, a rotation of π 2 about the z -axis, a translation of L 7 along the z -axis, and another rotation of π 2 about the z -axis: T P = A i | a = L 6 , α = π 2 , d = L 7 , θ = π 2 (33) T P E = 0 0 1 0 1 0 0 L 6 0 1 0 L 7 0 0 0 1 4.4.6. Transitioning from the Pitching Axis to the Endpoint As shown in , we use T P E . The translation along the x -axis is determined by the hypotenuse of L 9 and L 8 . The rotation about the z -axis is defined as π 2 minus the table angle (‘Tabang’) and the arctangent of the ratio L 9 L 8 . This angle is derived from the pitching encoder: T P E = A i | a = ( L 9 ) 2 + ( L 8 ) 2 , α = 0 , d = 0 , θ = π 2 − Tabang − arctan L 9 L 8 (34) T E E 1 = cos ( TabAng − θ 6 + π 2 ) − sin ( TabAng − θ 6 + π 2 ) 0 L 10 cos ( TabAng − θ 6 + π 2 ) sin ( TabAng − θ 6 + π 2 ) cos ( TabAng − θ 6 + π 2 ) 0 L 10 sin ( TabAng − θ 6 + π 2 ) 0 0 1 0 0 0 0 1 4.4.7. Additional Frames Are Integrated to Align with the Reference Direction In , it is evident that the TE coordinate does not align with the reference coordinate shown in . To correct this orientation, several rotational matrices need to be applied as follows: T E 1 : Rotation about the z -axis is determined by − π 2 minus the arctangent of the ratio L 9 L 8 . No translations are considered for this frame. T E 1 = A i | a = 0 , α = 0 , d = 0 , θ = − − π 2 arctan L 9 L 8 T E 2 : A rotation of − π 2 about the x -axis. No translations or other rotations are involved. T E 2 = A i | a = 0 , α = − π 2 , d = 0 , θ = 0 T E : A rotation of − π 2 about the z -axis. Again, no translations or other rotations are considered. T E = A i | a = 0 , α = 0 , d = 0 , θ = − π 2 (35) T E 1 E 2 = cos ( θ 6 − π 2 ) − sin ( θ 6 − π 2 ) 0 0 sin ( θ 6 − π 2 ) cos ( θ 6 − π 2 ) 0 0 0 0 1 0 0 0 0 1 (36) T E 2 E 3 = 1 0 0 0 0 0 1 0 0 − 1 0 0 0 0 0 1 (37) T E 3 0 = 0 1 0 0 − 1 0 0 0 0 0 1 0 0 0 0 1 outlines the Denavit–Hartenberg (DH) parameters for the PPS system. These parameters are essential for representing the kinematic transformations between adjacent frames of the system. Each row describes a specific transformation with the respective a i , α i , d i , and θ i parameters, as stipulated by Equation . (38) T o E = T o · T 1 · T R · T R R · T R L B · T L 1 · T L 2 · T L 3 · T L M · T P · T E · T E 1 · T E 2 · T E 3 The 4 × 4 matrix T o E is a homogeneous transformation matrix commonly used in the field of robotics and computer graphics to represent both the position and orientation of a body in space. The matrix T o E is given by : T o E = μ x O x α x p x μ y O y α y p y μ z O z α z p z 0 0 0 1 Position (Translation Vector) : The elements p x , p y , and p z in the fourth column represent the position of the target point with respect to the origin. They give the x , y , and z coordinates of the point in the base or reference frame. In the context of robotics, this would be the position of the end effector or tool tip relative to the base or reference frame. Orientation (Rotation Matrix) : The 3 × 3 matrix on the top-left corner of T o E represents the orientation of the body in space. The columns μ x μ y μ z , O x O y O z , and α x α y α z are unit vectors that represent the orientation of the body’s local x, y, and z axes, respectively, in the base frame . These vectors are often called the rotation axes, and their magnitude is always 1. The orientation of the body can be described using various representations like Euler angles, rotation matrices, or quaternions. In this matrix format, the orientation is represented using the rotation matrix We can extract the Euler angles from the following (39) R xzy = R y · R z · R x (40) R xzy = cos β cos γ − cos β sin γ cos α + sin β sin α cos β sin γ sin α + sin β cos α sin γ cos γ cos α − cos γ sin α − sin β cos γ sin β sin γ cos α + cos β sin α − sin β sin γ sin α + cos β cos α Through a substitution in the matrices of each link and by multiplying these, we obtain the final matrix which gives the information for both motions as follows: (41) μ x = cos ( q 1 + q 2 + θ 3 ) × cos ( R o t ) (42) μ y = sin ( q 1 + q 2 + θ 3 ) (43) μ z = − cos ( q 1 + q 2 + θ 3 ) × sin ( Rot ) (44) O x = sin ( Rot ) × sin ( TabAng ) − cos ( Rot ) × cos ( TabAng ) × cos ( q 1 ) × cos ( q 2 ) × sin ( θ 3 ) − cos ( Rot ) × cos ( TabAng ) × cos ( q 1 ) × cos ( θ 3 ) × sin ( q 2 ) − cos ( Rot ) × cos ( TabAng ) × cos ( q 2 ) × cos ( θ 3 ) × sin ( q 1 ) + cos ( Rot ) × cos ( TabAng ) × sin ( q 1 ) × sin ( q 2 ) × sin ( θ 3 ) (45) O y = cos ( TabAng + q 1 + q 2 + θ 3 ) 2 + cos ( q 1 − TabAng + q 2 + θ 3 ) 2 (46) O z = cos ( Rot ) · sin ( TabAng ) + cos ( TabAng ) · sin ( Rot ) · cos ( q 1 ) · cos ( q 2 ) · sin ( θ 3 ) + cos ( TabAng ) · sin ( Rot ) · cos ( q 1 ) · cos ( θ 3 ) · sin ( q 2 ) + cos ( TabAng ) · sin ( Rot ) · cos ( q 2 ) · cos ( θ 3 ) · sin ( q 1 ) − cos ( TabAng ) · sin ( Rot ) · sin ( q 1 ) · sin ( q 2 ) · sin ( θ 3 ) (47) a x = cos ( TabAng ) · sin ( Rot ) + cos ( Rot ) · sin ( TabAng ) · cos ( q 1 ) · cos ( q 2 ) · sin ( θ 3 ) + cos ( Rot ) · sin ( TabAng ) · cos ( q 1 ) · cos ( θ 3 ) · sin ( q 2 ) + cos ( Rot ) · sin ( TabAng ) · cos ( q 2 ) · cos ( θ 3 ) · sin ( q 1 ) − cos ( Rot ) · sin ( TabAng ) · sin ( q 1 ) · sin ( q 2 ) · sin ( θ 3 ) (48) a y = sin ( q 1 − TabAng + q 2 + θ 3 ) 2 − sin ( TabAng + q 1 + q 2 + θ 3 ) 2 (49) a z = cos ( Rot ) cos ( TabAng ) − sin ( Rot ) sin ( TabAng ) cos ( q 1 ) cos ( q 2 ) sin ( θ 3 ) − sin ( Rot ) sin ( TabAng ) cos ( q 1 ) cos ( θ 3 ) sin ( q 2 ) − sin ( Rot ) sin ( TabAng ) cos ( q 2 ) cos ( θ 3 ) sin ( q 1 ) + sin ( Rot ) sin ( TabAng ) sin ( q 1 ) sin ( q 2 ) sin ( θ 3 ) (50) p x = L 7 sin ( Rot ) − L 0 cos ( Rot ) 2 + L L sin ( Rot ) + L 1 cos ( Rot ) cos ( q 1 ) + L 10 sin ( Rot ) cos ( TabAng − θ 6 ) − L 6 sin ( q 1 + q 2 + θ 3 ) cos ( Rot ) + L 2 cos ( Rot ) cos ( q 1 ) cos ( q 2 ) − L 2 cos ( Rot ) sin ( q 1 ) sin ( q 2 ) + L 10 sin ( q 1 + q 2 + θ 3 ) cos ( Rot ) sin ( TabAng − θ 6 ) + L 3 cos ( q 1 + q 2 ) cos ( Rot ) cos ( θ 3 ) 2 − L 3 sin ( q 1 + q 2 ) cos ( Rot ) sin ( θ 3 ) 2 (51) p y = L R + L 2 sin ( q 1 + q 2 ) + L 1 sin ( q 1 ) + L 6 cos ( q 1 + q 2 + θ 3 ) + L 3 sin ( q 1 + q 2 + θ 3 ) 2 − L 10 sin ( TabAng + q 1 + q 2 + θ 3 − θ 6 ) 2 + L 10 sin ( q 1 − TabAng + q 2 + θ 3 + θ 6 ) 2 (52) p z = lin − L + L 7 cos ( Rot ) + L L cos ( Rot ) + L 0 sin ( Rot ) 2 − L 1 sin ( Rot ) cos ( q 1 ) + L 10 cos ( Rot ) cos ( TabAng − θ 6 ) + L 6 sin ( q 1 + q 2 + θ 3 ) sin ( Rot ) + L 3 sin ( q 1 + q 2 ) sin ( Rot ) sin ( θ 3 ) 2 − L 2 sin ( Rot ) cos ( q 1 ) cos ( q 2 ) + L 2 sin ( Rot ) sin ( q 1 ) sin ( q 2 ) − L 10 sin ( q 1 + q 2 + θ 3 ) sin ( Rot ) sin ( TabAng − θ 6 ) − L 3 cos ( q 1 + q 2 ) sin ( Rot ) cos ( θ 3 ) 2 From Equation , we can extract Euler angles using . (53) α = tan − 1 sin α cos α = tan − 1 cos γ · sin α cos α · cos γ = tan − 1 − R xzy , 2 , 3 R xzy , 2 , 2 = tan − 1 − R 2 , 3 R 2 , 2 (54) β = tan − 1 sin β cos β = tan − 1 cos γ · sin β cos β · cos γ = tan − 1 − R xzy , 3 , 1 R xzy , 1 , 1 = tan − 1 R 3 , 1 R 1 , 1 (55) γ = tan − 1 sin γ cos γ = tan − 1 sin γ ( cos α · cos γ ) 2 + ( − cos γ · sin α ) 2 = tan − 1 R xzy , 2 , 1 R xzy , 2 , 2 2 + R xzy , 2 , 3 2 = tan − 1 R 2 , 1 R 2 , 2 2 + R 2 , 3 2 (56) Pitch = 180 π · atan 2 sin ( TabAng + q 1 + q 2 + t h 3 ) / 2 − sin ( q 1 − TabAng + q 2 + t h 3 ) / 2 , cos ( TabAng + q 1 + q 2 + t h 3 ) / 2 + cos ( q 1 − TabAng + q 2 + t h 3 ) / 2 (57) Yaw = 180 π · atan 2 − cos ( q 1 + q 2 + t h 3 ) · sin ( Rot ) , cos ( q 1 + q 2 + t h 3 ) · cos ( Rot ) (58) Roll = 180 π · atan 2 sin ( q 1 + q 2 + t h 3 ) , cos ( T a b A n g + q 1 + q 2 + t h 3 ) 2 + cos ( q 1 − T a b A n g + q 2 + t h 3 ) 2 2 + sin ( T a b A n g + q 1 + q 2 + t h 3 ) 2 − sin ( q 1 − T a b A n g + q 2 + t h 3 ) 2 2
The patient positioning system (PPS) utilizes specific frame assignments to capture its distinct components and their movements, as shown in : Base Frame ( O i ): Located at the junction of the main rails. Linear Rail Frame ( T RR ): Positioned at the end of the linear rail. Lower Linkage System Frame ( T RLB ): At the midpoint of the lower linkage arm. Upper Linkage System Frame ( T LM ): At the midpoint of the upper linkage arm. Table Rod Frame ( T P ): Centered on the table rod. Tabletop Frame ( T E ): Centered on the tabletop. The linkage frames are shown in . These frame assignments are essential for precise kinematic evaluations in the PPS. 4.1.1. Linear Rail System from T0 to TRR This subsystem governs the bidirectional movement of the PPS. It facilitates both the translational motion of the main plate relative to the base plate along a rail system, and the rotational movement of the Linkage System around the main plate, courtesy of a rotary table as shown in . The linear motion enables the patient to be safely introduced and withdrawn from the operational area. In rooms equipped with a CT scanner, the rotary table allows a complete 180° rotation of the patient for imaging purposes. Encoded motors drive both movements, ensuring the precise positioning required for such medical devices. 4.1.2. Linkage System from TRLB to TLM Representing the heart of the PPS, the Linkage System comprises four pairs of two-arm linkages that connect the PPS to the plate on the rotary table. As shown in , to maintain patient stability and alignment, precision shafts interconnect two of these four pairs. The system harnesses the power of three encoded motors to maneuver three joints, positioning the PPS within a 2D plane that spans a substantial envelope size. This robust design enables the table to be adjusted vertically to accommodate patient needs, from a lowered position facilitating access for elderly patients, to an elevated state suitable for treatment procedures . 4.1.3. Table Tob TLM to TE Located at the top of the system, the Table Assembly is linked to the top plate of the Linkage Assembly. As shown in , the table, a product of Siemens, features a carbon fiber exterior filled with resin to minimize radio interference during treatment. Given the cantilevered design of the tabletop and the associated gravitational deflection, a pitch adjustment mechanism has been integrated into the Table Assembly. This mechanism actively counteracts gravity, ensuring a consistently level operating area.
This subsystem governs the bidirectional movement of the PPS. It facilitates both the translational motion of the main plate relative to the base plate along a rail system, and the rotational movement of the Linkage System around the main plate, courtesy of a rotary table as shown in . The linear motion enables the patient to be safely introduced and withdrawn from the operational area. In rooms equipped with a CT scanner, the rotary table allows a complete 180° rotation of the patient for imaging purposes. Encoded motors drive both movements, ensuring the precise positioning required for such medical devices.
Representing the heart of the PPS, the Linkage System comprises four pairs of two-arm linkages that connect the PPS to the plate on the rotary table. As shown in , to maintain patient stability and alignment, precision shafts interconnect two of these four pairs. The system harnesses the power of three encoded motors to maneuver three joints, positioning the PPS within a 2D plane that spans a substantial envelope size. This robust design enables the table to be adjusted vertically to accommodate patient needs, from a lowered position facilitating access for elderly patients, to an elevated state suitable for treatment procedures .
Located at the top of the system, the Table Assembly is linked to the top plate of the Linkage Assembly. As shown in , the table, a product of Siemens, features a carbon fiber exterior filled with resin to minimize radio interference during treatment. Given the cantilevered design of the tabletop and the associated gravitational deflection, a pitch adjustment mechanism has been integrated into the Table Assembly. This mechanism actively counteracts gravity, ensuring a consistently level operating area.
4.2.1. Linkage Subsystem Mathematical Modeling This section primarily focuses on using geometric principles to develop a kinematic model for the linkage subsystem. illustrates the points, link lengths, and joint angles associated with this system . We assumed that all the angles are measured counterclockwise: q 1 = 0 when L 1 is on L 0 . q 2 = 0 when L 2 is along the same line of L 1 . q 3 = 0 when L 5 is along the same line with L 0 . θ 3 = 0 when L 3 is along the same line of L 2 . θ 4 = 0 when L 3 is along the same line of L 4 . θ 5 = 0 when L 4 is along the same line of L 5 . We specify a coordinate frame for each joint as follows: (4) x 0 y 0 = cos q 1 − sin q 1 sin q 1 cos q 1 x 1 y 1 + 0 0 (5) x 1 y 1 = cos q 2 − sin q 2 sin q 2 cos q 2 x 2 y 2 + L 1 0 (6) x 2 y 2 = cos θ 3 − sin θ 3 sin θ 3 cos θ 3 x 3 y 3 + L 2 0 where (7) x 0 y 0 = cos q 3 − sin q 3 sin q 3 cos q 3 x 6 y 6 + L 0 0 (8) x 6 y 6 = cos θ 5 − sin θ 5 sin θ 5 cos θ 5 x 5 y 5 + L 5 0 (9) x 5 y 5 = cos θ 4 − sin θ 4 sin θ 4 cos θ 4 x 4 y 4 + L 4 0 (10) x 4 y 4 = cos 180 − sin 180 sin 180 cos 180 x 3 y 3 + L 3 0 From Equations to and , we determined each point position as follows: Point 1 (11) x 0 y 0 1 = cos q 1 − sin q 1 sin q 1 cos q 1 0 0 + 0 0 = 0 0 Point 2 (12) x 0 y 0 2 = cos q 1 − sin q 1 sin q 1 cos q 1 L 1 0 = L 1 cos q 1 L 1 sin q 1 Point 3 (13) x 1 y 1 3 = cos q 2 − sin q 2 sin q 2 cos q 2 L 2 y 2 + L 1 0 = L 2 cos q 2 + L 1 L 2 sin q 2 (14) x 0 y 0 3 = cos q 1 − sin q 1 sin q 1 cos q 1 L 2 cos q 2 + L 1 L 2 sin q 2 = L 2 cos q 1 cos q 2 + L 1 cos q 1 − L 2 sin q 1 sin q 2 L 2 sin q 1 cos q 2 + L 1 sin q 1 + L 2 cos q 1 sin q 2 Point 6 (15) x 0 y 0 6 = cos q 3 − sin q 3 sin q 3 cos q 3 0 0 + L 0 0 = L 0 0 Point 5 (16) x 6 y 6 5 = cos θ 5 − sin θ 5 sin θ 5 cos θ 5 0 0 + L 5 0 = L 5 0 (17) x 0 y 0 5 = cos q 3 − sin q 3 sin q 3 cos q 3 L 5 0 + L 0 0 = L 5 cos q 3 + L 0 L 5 sin q 3 Therefore, Point 4 (18) x 5 y 5 4 = cos θ 4 − sin θ 4 sin θ 4 cos θ 4 0 0 + L 4 0 = L 4 0 (19) x 6 y 6 4 = cos θ 5 − sin θ 5 sin θ 5 cos θ 5 L 4 0 + L 5 0 = L 4 cos θ 5 + L 5 L 4 sin θ 5 (20) x 0 y 0 4 = cos q 3 − sin q 3 sin q 3 cos q 3 L 4 cos θ 5 + L 5 L 4 sin θ 5 + L 0 0 = L 4 cos q 3 cos θ 5 + L 5 cos q 3 − L 4 sin q 3 sin θ 5 + L 0 L 4 sin q 3 cos θ 5 + L 5 sin q 3 + L 4 cos q 3 sin θ 5 4.2.2. Intersection of Two Circles In the domain of kinematics and mechanical geometry, a well-accepted technique for determining the location of a specific point within a linkage mechanism involves the intersection of two circles . As shown in , we can pinpoint the position of Point 4 using this intersecting circles methodology. The scenario involves two circles: the first centered at Point 3 with a radius of L3, and the second centered at Point 5 with a radius of L5. The intersection of these two circles yields two potential points, namely Point 4 and Point 6. Consider two circles: Circle 1, centered at Point 3 ( C 3 = ( x 3 , y 3 ) ) with radius L 3 . Circle 2, centered at Point 5 ( C 5 = ( x 5 , y 5 ) ) with radius L 5 . The equations of these circles are as follows: (21) ( x − x 3 ) 2 + ( y − y 3 ) 2 = L 3 2 (22) ( x − x 5 ) 2 + ( y − y 5 ) 2 = L 5 2 To find the intersection points, we can subtract two circle equations and simplify these. This will give a linear equation in x and y . Out of the two intersection points, the one with the greater y -coordinate is considered the upper intersection point. We choose this as Point 4, and the other intersection point is disregarded. These represent the possible locations of the mechanical joint or linkage. However, due to the physical constraints and design of our mechanical system, the existence of Point 6 would suggest a mechanically unfeasible configuration. Therefore, Point 6 can be categorically eliminated from our options, leaving us with the valid position, Point 4. After that, to ascertain the orientation of the upper arm of the linkage mechanism , we must calculate the value of ϕ 3 . This angle ϕ 3 is a critical parameter that provides insights into the inclination of the upper arm, thereby enabling a comprehensive understanding of the system’s kinematics. The determination of ϕ 3 is a crucial step in accurately modeling and analyzing the mechanical system. (23) c = ( P 4 ( x ) − P 2 ( x ) ) 2 + ( P 4 ( y ) − P 2 ( y ) ) 2 (24) Γ = arc a 2 + b 2 − c 2 2 × a × b (25) ϕ 3 = 180 ∘ − Γ
This section primarily focuses on using geometric principles to develop a kinematic model for the linkage subsystem. illustrates the points, link lengths, and joint angles associated with this system . We assumed that all the angles are measured counterclockwise: q 1 = 0 when L 1 is on L 0 . q 2 = 0 when L 2 is along the same line of L 1 . q 3 = 0 when L 5 is along the same line with L 0 . θ 3 = 0 when L 3 is along the same line of L 2 . θ 4 = 0 when L 3 is along the same line of L 4 . θ 5 = 0 when L 4 is along the same line of L 5 . We specify a coordinate frame for each joint as follows: (4) x 0 y 0 = cos q 1 − sin q 1 sin q 1 cos q 1 x 1 y 1 + 0 0 (5) x 1 y 1 = cos q 2 − sin q 2 sin q 2 cos q 2 x 2 y 2 + L 1 0 (6) x 2 y 2 = cos θ 3 − sin θ 3 sin θ 3 cos θ 3 x 3 y 3 + L 2 0 where (7) x 0 y 0 = cos q 3 − sin q 3 sin q 3 cos q 3 x 6 y 6 + L 0 0 (8) x 6 y 6 = cos θ 5 − sin θ 5 sin θ 5 cos θ 5 x 5 y 5 + L 5 0 (9) x 5 y 5 = cos θ 4 − sin θ 4 sin θ 4 cos θ 4 x 4 y 4 + L 4 0 (10) x 4 y 4 = cos 180 − sin 180 sin 180 cos 180 x 3 y 3 + L 3 0 From Equations to and , we determined each point position as follows: Point 1 (11) x 0 y 0 1 = cos q 1 − sin q 1 sin q 1 cos q 1 0 0 + 0 0 = 0 0 Point 2 (12) x 0 y 0 2 = cos q 1 − sin q 1 sin q 1 cos q 1 L 1 0 = L 1 cos q 1 L 1 sin q 1 Point 3 (13) x 1 y 1 3 = cos q 2 − sin q 2 sin q 2 cos q 2 L 2 y 2 + L 1 0 = L 2 cos q 2 + L 1 L 2 sin q 2 (14) x 0 y 0 3 = cos q 1 − sin q 1 sin q 1 cos q 1 L 2 cos q 2 + L 1 L 2 sin q 2 = L 2 cos q 1 cos q 2 + L 1 cos q 1 − L 2 sin q 1 sin q 2 L 2 sin q 1 cos q 2 + L 1 sin q 1 + L 2 cos q 1 sin q 2 Point 6 (15) x 0 y 0 6 = cos q 3 − sin q 3 sin q 3 cos q 3 0 0 + L 0 0 = L 0 0 Point 5 (16) x 6 y 6 5 = cos θ 5 − sin θ 5 sin θ 5 cos θ 5 0 0 + L 5 0 = L 5 0 (17) x 0 y 0 5 = cos q 3 − sin q 3 sin q 3 cos q 3 L 5 0 + L 0 0 = L 5 cos q 3 + L 0 L 5 sin q 3 Therefore, Point 4 (18) x 5 y 5 4 = cos θ 4 − sin θ 4 sin θ 4 cos θ 4 0 0 + L 4 0 = L 4 0 (19) x 6 y 6 4 = cos θ 5 − sin θ 5 sin θ 5 cos θ 5 L 4 0 + L 5 0 = L 4 cos θ 5 + L 5 L 4 sin θ 5 (20) x 0 y 0 4 = cos q 3 − sin q 3 sin q 3 cos q 3 L 4 cos θ 5 + L 5 L 4 sin θ 5 + L 0 0 = L 4 cos q 3 cos θ 5 + L 5 cos q 3 − L 4 sin q 3 sin θ 5 + L 0 L 4 sin q 3 cos θ 5 + L 5 sin q 3 + L 4 cos q 3 sin θ 5
In the domain of kinematics and mechanical geometry, a well-accepted technique for determining the location of a specific point within a linkage mechanism involves the intersection of two circles . As shown in , we can pinpoint the position of Point 4 using this intersecting circles methodology. The scenario involves two circles: the first centered at Point 3 with a radius of L3, and the second centered at Point 5 with a radius of L5. The intersection of these two circles yields two potential points, namely Point 4 and Point 6. Consider two circles: Circle 1, centered at Point 3 ( C 3 = ( x 3 , y 3 ) ) with radius L 3 . Circle 2, centered at Point 5 ( C 5 = ( x 5 , y 5 ) ) with radius L 5 . The equations of these circles are as follows: (21) ( x − x 3 ) 2 + ( y − y 3 ) 2 = L 3 2 (22) ( x − x 5 ) 2 + ( y − y 5 ) 2 = L 5 2 To find the intersection points, we can subtract two circle equations and simplify these. This will give a linear equation in x and y . Out of the two intersection points, the one with the greater y -coordinate is considered the upper intersection point. We choose this as Point 4, and the other intersection point is disregarded. These represent the possible locations of the mechanical joint or linkage. However, due to the physical constraints and design of our mechanical system, the existence of Point 6 would suggest a mechanically unfeasible configuration. Therefore, Point 6 can be categorically eliminated from our options, leaving us with the valid position, Point 4. After that, to ascertain the orientation of the upper arm of the linkage mechanism , we must calculate the value of ϕ 3 . This angle ϕ 3 is a critical parameter that provides insights into the inclination of the upper arm, thereby enabling a comprehensive understanding of the system’s kinematics. The determination of ϕ 3 is a crucial step in accurately modeling and analyzing the mechanical system. (23) c = ( P 4 ( x ) − P 2 ( x ) ) 2 + ( P 4 ( y ) − P 2 ( y ) ) 2 (24) Γ = arc a 2 + b 2 − c 2 2 × a × b (25) ϕ 3 = 180 ∘ − Γ
Table angle or pitching angle controlled by servo motor and CAM as shown in the following . In order to transition from point TP to point TE, it is necessary to calculate the length L10, based on the angle input as per the given and . Then, after determining all needed parameters from , we can create the DH table as follows:
With the essential parameters and points now determined, we can utilize the Denavit–Hartenberg (DH) parameters. This allows us to transition from one frame to the subsequent frame, ultimately leading us to the final transformation matrix representing our PPS. 4.4.1. DH Matrix Formulation The standard DH transformation matrix, denoted as A i , is constructed using the given parameters a , α , d , and θ : A i = cos ( θ ) − sin ( θ ) 0 0 sin ( θ ) cos ( θ ) 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 0 0 1 d 0 0 0 1 1 0 0 a 0 1 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 cos ( α ) − sin ( α ) 0 0 sin ( α ) cos ( α ) 0 0 0 0 1 4.4.2. Linear Rail Subsystem The transformation matrix for the linear rail subsystem, denoted as T o , is as follows: T o = A i | a = 0 , α = 0 , d = 0 , θ = 0 T 1 = A i | a = 0 , α = 0 , d = − L + L i n , θ = 0 (26) T 0 1 = 1 0 0 0 0 1 0 0 0 0 1 Lin − L 0 0 0 1 As shown in , the transformation matrices describe the motion and positioning of the linear rail system. 4.4.3. Rotary Base to the Height of the Linkage As shown in , two transformation matrices are derived: T R representing a rotation about the θ -axis by − π 2 and no translation along the x -axis: T R = A i | a = 0 , α = 0 , d = 0 , θ = − π 2 T R R representing a rotation about the ‘Rott’ angle and a translation along the z -axis by L R : T R R = A i | a = 0 , α = 0 , d = L R , θ = R o t (27) T R R R = 1 0 0 0 0 0 1 0 0 − 1 0 0 0 0 0 1 (28) T R R R L B = cos ( R o t ) − sin ( R o t ) 0 0 sin ( R o t ) cos ( R o t ) 0 0 0 0 1 L R 0 0 0 1 4.4.4. Linkage Subsystem The linkage subsystem of the PPS consists of multiple stages of transformation: First, no translation along the x -axis and a rotation of π 2 about the x -axis: T R L B = A i | a = 0 , α = π 2 , d = 0 , θ = 0 (29) T R L B L 1 = 1 0 0 0 0 0 − 1 0 0 1 0 0 0 0 0 1 Second, a translation of L 0 2 along the x -axis, a translation of L L along the z -axis, and a rotation of π about the z -axis: T L 1 = A i | a = L 0 2 , α = 0 , d = L L , θ = { π Next, we move the first link by controlling the first motor q 1 . This involves a translation of L 1 along the x -axis and a rotation of − π + q 1 about the z -axis: T L 2 = A i | a = L 1 , α = 0 , d = 0 , θ = p i + q 1 (30) T L 1 L 2 = − 1 0 0 − L 0 2 0 − 1 0 0 0 0 1 L L 0 0 0 1 After this, we control the second link with the second motor q 2 . This step comprises a translation of L 2 along the x -axis and a rotation of q 2 about the z -axis: T L 12 = A i | a = L 2 , α = 0 , d = 0 , θ = q 2 (31) T L 3 L M = cos ( q 2 ) − sin ( q 2 ) 0 L 2 cos ( q 2 ) sin ( q 2 ) cos ( q 2 ) 0 L 2 sin ( q 2 ) 0 0 1 0 0 0 0 1 Lastly, we move to the middle of the upper arm based on the ϕ 3 value. This involves a translation of L 3 2 along the x -axis and a rotation of ϕ 3 about the z -axis: T L M = A i | a = L 3 2 , α = 0 , d = 0 , θ = ϕ 3 (32) T L M P = cos ( θ 3 ) − sin ( θ 3 ) 0 L 3 cos ( θ 3 ) 2 sin ( θ 3 ) cos ( θ 3 ) 0 L 3 sin ( θ 3 ) 2 0 0 1 0 0 0 0 1 4.4.5. Moving to the Table Pitching Frame See for an illustration of the following transition. We transition from T L M to T P . This involves a translation of L 6 along the x -axis, a rotation of π 2 about the z -axis, a translation of L 7 along the z -axis, and another rotation of π 2 about the z -axis: T P = A i | a = L 6 , α = π 2 , d = L 7 , θ = π 2 (33) T P E = 0 0 1 0 1 0 0 L 6 0 1 0 L 7 0 0 0 1 4.4.6. Transitioning from the Pitching Axis to the Endpoint As shown in , we use T P E . The translation along the x -axis is determined by the hypotenuse of L 9 and L 8 . The rotation about the z -axis is defined as π 2 minus the table angle (‘Tabang’) and the arctangent of the ratio L 9 L 8 . This angle is derived from the pitching encoder: T P E = A i | a = ( L 9 ) 2 + ( L 8 ) 2 , α = 0 , d = 0 , θ = π 2 − Tabang − arctan L 9 L 8 (34) T E E 1 = cos ( TabAng − θ 6 + π 2 ) − sin ( TabAng − θ 6 + π 2 ) 0 L 10 cos ( TabAng − θ 6 + π 2 ) sin ( TabAng − θ 6 + π 2 ) cos ( TabAng − θ 6 + π 2 ) 0 L 10 sin ( TabAng − θ 6 + π 2 ) 0 0 1 0 0 0 0 1 4.4.7. Additional Frames Are Integrated to Align with the Reference Direction In , it is evident that the TE coordinate does not align with the reference coordinate shown in . To correct this orientation, several rotational matrices need to be applied as follows: T E 1 : Rotation about the z -axis is determined by − π 2 minus the arctangent of the ratio L 9 L 8 . No translations are considered for this frame. T E 1 = A i | a = 0 , α = 0 , d = 0 , θ = − − π 2 arctan L 9 L 8 T E 2 : A rotation of − π 2 about the x -axis. No translations or other rotations are involved. T E 2 = A i | a = 0 , α = − π 2 , d = 0 , θ = 0 T E : A rotation of − π 2 about the z -axis. Again, no translations or other rotations are considered. T E = A i | a = 0 , α = 0 , d = 0 , θ = − π 2 (35) T E 1 E 2 = cos ( θ 6 − π 2 ) − sin ( θ 6 − π 2 ) 0 0 sin ( θ 6 − π 2 ) cos ( θ 6 − π 2 ) 0 0 0 0 1 0 0 0 0 1 (36) T E 2 E 3 = 1 0 0 0 0 0 1 0 0 − 1 0 0 0 0 0 1 (37) T E 3 0 = 0 1 0 0 − 1 0 0 0 0 0 1 0 0 0 0 1 outlines the Denavit–Hartenberg (DH) parameters for the PPS system. These parameters are essential for representing the kinematic transformations between adjacent frames of the system. Each row describes a specific transformation with the respective a i , α i , d i , and θ i parameters, as stipulated by Equation . (38) T o E = T o · T 1 · T R · T R R · T R L B · T L 1 · T L 2 · T L 3 · T L M · T P · T E · T E 1 · T E 2 · T E 3 The 4 × 4 matrix T o E is a homogeneous transformation matrix commonly used in the field of robotics and computer graphics to represent both the position and orientation of a body in space. The matrix T o E is given by : T o E = μ x O x α x p x μ y O y α y p y μ z O z α z p z 0 0 0 1 Position (Translation Vector) : The elements p x , p y , and p z in the fourth column represent the position of the target point with respect to the origin. They give the x , y , and z coordinates of the point in the base or reference frame. In the context of robotics, this would be the position of the end effector or tool tip relative to the base or reference frame. Orientation (Rotation Matrix) : The 3 × 3 matrix on the top-left corner of T o E represents the orientation of the body in space. The columns μ x μ y μ z , O x O y O z , and α x α y α z are unit vectors that represent the orientation of the body’s local x, y, and z axes, respectively, in the base frame . These vectors are often called the rotation axes, and their magnitude is always 1. The orientation of the body can be described using various representations like Euler angles, rotation matrices, or quaternions. In this matrix format, the orientation is represented using the rotation matrix We can extract the Euler angles from the following (39) R xzy = R y · R z · R x (40) R xzy = cos β cos γ − cos β sin γ cos α + sin β sin α cos β sin γ sin α + sin β cos α sin γ cos γ cos α − cos γ sin α − sin β cos γ sin β sin γ cos α + cos β sin α − sin β sin γ sin α + cos β cos α Through a substitution in the matrices of each link and by multiplying these, we obtain the final matrix which gives the information for both motions as follows: (41) μ x = cos ( q 1 + q 2 + θ 3 ) × cos ( R o t ) (42) μ y = sin ( q 1 + q 2 + θ 3 ) (43) μ z = − cos ( q 1 + q 2 + θ 3 ) × sin ( Rot ) (44) O x = sin ( Rot ) × sin ( TabAng ) − cos ( Rot ) × cos ( TabAng ) × cos ( q 1 ) × cos ( q 2 ) × sin ( θ 3 ) − cos ( Rot ) × cos ( TabAng ) × cos ( q 1 ) × cos ( θ 3 ) × sin ( q 2 ) − cos ( Rot ) × cos ( TabAng ) × cos ( q 2 ) × cos ( θ 3 ) × sin ( q 1 ) + cos ( Rot ) × cos ( TabAng ) × sin ( q 1 ) × sin ( q 2 ) × sin ( θ 3 ) (45) O y = cos ( TabAng + q 1 + q 2 + θ 3 ) 2 + cos ( q 1 − TabAng + q 2 + θ 3 ) 2 (46) O z = cos ( Rot ) · sin ( TabAng ) + cos ( TabAng ) · sin ( Rot ) · cos ( q 1 ) · cos ( q 2 ) · sin ( θ 3 ) + cos ( TabAng ) · sin ( Rot ) · cos ( q 1 ) · cos ( θ 3 ) · sin ( q 2 ) + cos ( TabAng ) · sin ( Rot ) · cos ( q 2 ) · cos ( θ 3 ) · sin ( q 1 ) − cos ( TabAng ) · sin ( Rot ) · sin ( q 1 ) · sin ( q 2 ) · sin ( θ 3 ) (47) a x = cos ( TabAng ) · sin ( Rot ) + cos ( Rot ) · sin ( TabAng ) · cos ( q 1 ) · cos ( q 2 ) · sin ( θ 3 ) + cos ( Rot ) · sin ( TabAng ) · cos ( q 1 ) · cos ( θ 3 ) · sin ( q 2 ) + cos ( Rot ) · sin ( TabAng ) · cos ( q 2 ) · cos ( θ 3 ) · sin ( q 1 ) − cos ( Rot ) · sin ( TabAng ) · sin ( q 1 ) · sin ( q 2 ) · sin ( θ 3 ) (48) a y = sin ( q 1 − TabAng + q 2 + θ 3 ) 2 − sin ( TabAng + q 1 + q 2 + θ 3 ) 2 (49) a z = cos ( Rot ) cos ( TabAng ) − sin ( Rot ) sin ( TabAng ) cos ( q 1 ) cos ( q 2 ) sin ( θ 3 ) − sin ( Rot ) sin ( TabAng ) cos ( q 1 ) cos ( θ 3 ) sin ( q 2 ) − sin ( Rot ) sin ( TabAng ) cos ( q 2 ) cos ( θ 3 ) sin ( q 1 ) + sin ( Rot ) sin ( TabAng ) sin ( q 1 ) sin ( q 2 ) sin ( θ 3 ) (50) p x = L 7 sin ( Rot ) − L 0 cos ( Rot ) 2 + L L sin ( Rot ) + L 1 cos ( Rot ) cos ( q 1 ) + L 10 sin ( Rot ) cos ( TabAng − θ 6 ) − L 6 sin ( q 1 + q 2 + θ 3 ) cos ( Rot ) + L 2 cos ( Rot ) cos ( q 1 ) cos ( q 2 ) − L 2 cos ( Rot ) sin ( q 1 ) sin ( q 2 ) + L 10 sin ( q 1 + q 2 + θ 3 ) cos ( Rot ) sin ( TabAng − θ 6 ) + L 3 cos ( q 1 + q 2 ) cos ( Rot ) cos ( θ 3 ) 2 − L 3 sin ( q 1 + q 2 ) cos ( Rot ) sin ( θ 3 ) 2 (51) p y = L R + L 2 sin ( q 1 + q 2 ) + L 1 sin ( q 1 ) + L 6 cos ( q 1 + q 2 + θ 3 ) + L 3 sin ( q 1 + q 2 + θ 3 ) 2 − L 10 sin ( TabAng + q 1 + q 2 + θ 3 − θ 6 ) 2 + L 10 sin ( q 1 − TabAng + q 2 + θ 3 + θ 6 ) 2 (52) p z = lin − L + L 7 cos ( Rot ) + L L cos ( Rot ) + L 0 sin ( Rot ) 2 − L 1 sin ( Rot ) cos ( q 1 ) + L 10 cos ( Rot ) cos ( TabAng − θ 6 ) + L 6 sin ( q 1 + q 2 + θ 3 ) sin ( Rot ) + L 3 sin ( q 1 + q 2 ) sin ( Rot ) sin ( θ 3 ) 2 − L 2 sin ( Rot ) cos ( q 1 ) cos ( q 2 ) + L 2 sin ( Rot ) sin ( q 1 ) sin ( q 2 ) − L 10 sin ( q 1 + q 2 + θ 3 ) sin ( Rot ) sin ( TabAng − θ 6 ) − L 3 cos ( q 1 + q 2 ) sin ( Rot ) cos ( θ 3 ) 2 From Equation , we can extract Euler angles using . (53) α = tan − 1 sin α cos α = tan − 1 cos γ · sin α cos α · cos γ = tan − 1 − R xzy , 2 , 3 R xzy , 2 , 2 = tan − 1 − R 2 , 3 R 2 , 2 (54) β = tan − 1 sin β cos β = tan − 1 cos γ · sin β cos β · cos γ = tan − 1 − R xzy , 3 , 1 R xzy , 1 , 1 = tan − 1 R 3 , 1 R 1 , 1 (55) γ = tan − 1 sin γ cos γ = tan − 1 sin γ ( cos α · cos γ ) 2 + ( − cos γ · sin α ) 2 = tan − 1 R xzy , 2 , 1 R xzy , 2 , 2 2 + R xzy , 2 , 3 2 = tan − 1 R 2 , 1 R 2 , 2 2 + R 2 , 3 2 (56) Pitch = 180 π · atan 2 sin ( TabAng + q 1 + q 2 + t h 3 ) / 2 − sin ( q 1 − TabAng + q 2 + t h 3 ) / 2 , cos ( TabAng + q 1 + q 2 + t h 3 ) / 2 + cos ( q 1 − TabAng + q 2 + t h 3 ) / 2 (57) Yaw = 180 π · atan 2 − cos ( q 1 + q 2 + t h 3 ) · sin ( Rot ) , cos ( q 1 + q 2 + t h 3 ) · cos ( Rot ) (58) Roll = 180 π · atan 2 sin ( q 1 + q 2 + t h 3 ) , cos ( T a b A n g + q 1 + q 2 + t h 3 ) 2 + cos ( q 1 − T a b A n g + q 2 + t h 3 ) 2 2 + sin ( T a b A n g + q 1 + q 2 + t h 3 ) 2 − sin ( q 1 − T a b A n g + q 2 + t h 3 ) 2 2
The standard DH transformation matrix, denoted as A i , is constructed using the given parameters a , α , d , and θ : A i = cos ( θ ) − sin ( θ ) 0 0 sin ( θ ) cos ( θ ) 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 0 0 1 d 0 0 0 1 1 0 0 a 0 1 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 cos ( α ) − sin ( α ) 0 0 sin ( α ) cos ( α ) 0 0 0 0 1
The transformation matrix for the linear rail subsystem, denoted as T o , is as follows: T o = A i | a = 0 , α = 0 , d = 0 , θ = 0 T 1 = A i | a = 0 , α = 0 , d = − L + L i n , θ = 0 (26) T 0 1 = 1 0 0 0 0 1 0 0 0 0 1 Lin − L 0 0 0 1 As shown in , the transformation matrices describe the motion and positioning of the linear rail system.
As shown in , two transformation matrices are derived: T R representing a rotation about the θ -axis by − π 2 and no translation along the x -axis: T R = A i | a = 0 , α = 0 , d = 0 , θ = − π 2 T R R representing a rotation about the ‘Rott’ angle and a translation along the z -axis by L R : T R R = A i | a = 0 , α = 0 , d = L R , θ = R o t (27) T R R R = 1 0 0 0 0 0 1 0 0 − 1 0 0 0 0 0 1 (28) T R R R L B = cos ( R o t ) − sin ( R o t ) 0 0 sin ( R o t ) cos ( R o t ) 0 0 0 0 1 L R 0 0 0 1
The linkage subsystem of the PPS consists of multiple stages of transformation: First, no translation along the x -axis and a rotation of π 2 about the x -axis: T R L B = A i | a = 0 , α = π 2 , d = 0 , θ = 0 (29) T R L B L 1 = 1 0 0 0 0 0 − 1 0 0 1 0 0 0 0 0 1 Second, a translation of L 0 2 along the x -axis, a translation of L L along the z -axis, and a rotation of π about the z -axis: T L 1 = A i | a = L 0 2 , α = 0 , d = L L , θ = { π Next, we move the first link by controlling the first motor q 1 . This involves a translation of L 1 along the x -axis and a rotation of − π + q 1 about the z -axis: T L 2 = A i | a = L 1 , α = 0 , d = 0 , θ = p i + q 1 (30) T L 1 L 2 = − 1 0 0 − L 0 2 0 − 1 0 0 0 0 1 L L 0 0 0 1 After this, we control the second link with the second motor q 2 . This step comprises a translation of L 2 along the x -axis and a rotation of q 2 about the z -axis: T L 12 = A i | a = L 2 , α = 0 , d = 0 , θ = q 2 (31) T L 3 L M = cos ( q 2 ) − sin ( q 2 ) 0 L 2 cos ( q 2 ) sin ( q 2 ) cos ( q 2 ) 0 L 2 sin ( q 2 ) 0 0 1 0 0 0 0 1 Lastly, we move to the middle of the upper arm based on the ϕ 3 value. This involves a translation of L 3 2 along the x -axis and a rotation of ϕ 3 about the z -axis: T L M = A i | a = L 3 2 , α = 0 , d = 0 , θ = ϕ 3 (32) T L M P = cos ( θ 3 ) − sin ( θ 3 ) 0 L 3 cos ( θ 3 ) 2 sin ( θ 3 ) cos ( θ 3 ) 0 L 3 sin ( θ 3 ) 2 0 0 1 0 0 0 0 1
See for an illustration of the following transition. We transition from T L M to T P . This involves a translation of L 6 along the x -axis, a rotation of π 2 about the z -axis, a translation of L 7 along the z -axis, and another rotation of π 2 about the z -axis: T P = A i | a = L 6 , α = π 2 , d = L 7 , θ = π 2 (33) T P E = 0 0 1 0 1 0 0 L 6 0 1 0 L 7 0 0 0 1
As shown in , we use T P E . The translation along the x -axis is determined by the hypotenuse of L 9 and L 8 . The rotation about the z -axis is defined as π 2 minus the table angle (‘Tabang’) and the arctangent of the ratio L 9 L 8 . This angle is derived from the pitching encoder: T P E = A i | a = ( L 9 ) 2 + ( L 8 ) 2 , α = 0 , d = 0 , θ = π 2 − Tabang − arctan L 9 L 8 (34) T E E 1 = cos ( TabAng − θ 6 + π 2 ) − sin ( TabAng − θ 6 + π 2 ) 0 L 10 cos ( TabAng − θ 6 + π 2 ) sin ( TabAng − θ 6 + π 2 ) cos ( TabAng − θ 6 + π 2 ) 0 L 10 sin ( TabAng − θ 6 + π 2 ) 0 0 1 0 0 0 0 1
In , it is evident that the TE coordinate does not align with the reference coordinate shown in . To correct this orientation, several rotational matrices need to be applied as follows: T E 1 : Rotation about the z -axis is determined by − π 2 minus the arctangent of the ratio L 9 L 8 . No translations are considered for this frame. T E 1 = A i | a = 0 , α = 0 , d = 0 , θ = − − π 2 arctan L 9 L 8 T E 2 : A rotation of − π 2 about the x -axis. No translations or other rotations are involved. T E 2 = A i | a = 0 , α = − π 2 , d = 0 , θ = 0 T E : A rotation of − π 2 about the z -axis. Again, no translations or other rotations are considered. T E = A i | a = 0 , α = 0 , d = 0 , θ = − π 2 (35) T E 1 E 2 = cos ( θ 6 − π 2 ) − sin ( θ 6 − π 2 ) 0 0 sin ( θ 6 − π 2 ) cos ( θ 6 − π 2 ) 0 0 0 0 1 0 0 0 0 1 (36) T E 2 E 3 = 1 0 0 0 0 0 1 0 0 − 1 0 0 0 0 0 1 (37) T E 3 0 = 0 1 0 0 − 1 0 0 0 0 0 1 0 0 0 0 1 outlines the Denavit–Hartenberg (DH) parameters for the PPS system. These parameters are essential for representing the kinematic transformations between adjacent frames of the system. Each row describes a specific transformation with the respective a i , α i , d i , and θ i parameters, as stipulated by Equation . (38) T o E = T o · T 1 · T R · T R R · T R L B · T L 1 · T L 2 · T L 3 · T L M · T P · T E · T E 1 · T E 2 · T E 3 The 4 × 4 matrix T o E is a homogeneous transformation matrix commonly used in the field of robotics and computer graphics to represent both the position and orientation of a body in space. The matrix T o E is given by : T o E = μ x O x α x p x μ y O y α y p y μ z O z α z p z 0 0 0 1 Position (Translation Vector) : The elements p x , p y , and p z in the fourth column represent the position of the target point with respect to the origin. They give the x , y , and z coordinates of the point in the base or reference frame. In the context of robotics, this would be the position of the end effector or tool tip relative to the base or reference frame. Orientation (Rotation Matrix) : The 3 × 3 matrix on the top-left corner of T o E represents the orientation of the body in space. The columns μ x μ y μ z , O x O y O z , and α x α y α z are unit vectors that represent the orientation of the body’s local x, y, and z axes, respectively, in the base frame . These vectors are often called the rotation axes, and their magnitude is always 1. The orientation of the body can be described using various representations like Euler angles, rotation matrices, or quaternions. In this matrix format, the orientation is represented using the rotation matrix We can extract the Euler angles from the following (39) R xzy = R y · R z · R x (40) R xzy = cos β cos γ − cos β sin γ cos α + sin β sin α cos β sin γ sin α + sin β cos α sin γ cos γ cos α − cos γ sin α − sin β cos γ sin β sin γ cos α + cos β sin α − sin β sin γ sin α + cos β cos α Through a substitution in the matrices of each link and by multiplying these, we obtain the final matrix which gives the information for both motions as follows: (41) μ x = cos ( q 1 + q 2 + θ 3 ) × cos ( R o t ) (42) μ y = sin ( q 1 + q 2 + θ 3 ) (43) μ z = − cos ( q 1 + q 2 + θ 3 ) × sin ( Rot ) (44) O x = sin ( Rot ) × sin ( TabAng ) − cos ( Rot ) × cos ( TabAng ) × cos ( q 1 ) × cos ( q 2 ) × sin ( θ 3 ) − cos ( Rot ) × cos ( TabAng ) × cos ( q 1 ) × cos ( θ 3 ) × sin ( q 2 ) − cos ( Rot ) × cos ( TabAng ) × cos ( q 2 ) × cos ( θ 3 ) × sin ( q 1 ) + cos ( Rot ) × cos ( TabAng ) × sin ( q 1 ) × sin ( q 2 ) × sin ( θ 3 ) (45) O y = cos ( TabAng + q 1 + q 2 + θ 3 ) 2 + cos ( q 1 − TabAng + q 2 + θ 3 ) 2 (46) O z = cos ( Rot ) · sin ( TabAng ) + cos ( TabAng ) · sin ( Rot ) · cos ( q 1 ) · cos ( q 2 ) · sin ( θ 3 ) + cos ( TabAng ) · sin ( Rot ) · cos ( q 1 ) · cos ( θ 3 ) · sin ( q 2 ) + cos ( TabAng ) · sin ( Rot ) · cos ( q 2 ) · cos ( θ 3 ) · sin ( q 1 ) − cos ( TabAng ) · sin ( Rot ) · sin ( q 1 ) · sin ( q 2 ) · sin ( θ 3 ) (47) a x = cos ( TabAng ) · sin ( Rot ) + cos ( Rot ) · sin ( TabAng ) · cos ( q 1 ) · cos ( q 2 ) · sin ( θ 3 ) + cos ( Rot ) · sin ( TabAng ) · cos ( q 1 ) · cos ( θ 3 ) · sin ( q 2 ) + cos ( Rot ) · sin ( TabAng ) · cos ( q 2 ) · cos ( θ 3 ) · sin ( q 1 ) − cos ( Rot ) · sin ( TabAng ) · sin ( q 1 ) · sin ( q 2 ) · sin ( θ 3 ) (48) a y = sin ( q 1 − TabAng + q 2 + θ 3 ) 2 − sin ( TabAng + q 1 + q 2 + θ 3 ) 2 (49) a z = cos ( Rot ) cos ( TabAng ) − sin ( Rot ) sin ( TabAng ) cos ( q 1 ) cos ( q 2 ) sin ( θ 3 ) − sin ( Rot ) sin ( TabAng ) cos ( q 1 ) cos ( θ 3 ) sin ( q 2 ) − sin ( Rot ) sin ( TabAng ) cos ( q 2 ) cos ( θ 3 ) sin ( q 1 ) + sin ( Rot ) sin ( TabAng ) sin ( q 1 ) sin ( q 2 ) sin ( θ 3 ) (50) p x = L 7 sin ( Rot ) − L 0 cos ( Rot ) 2 + L L sin ( Rot ) + L 1 cos ( Rot ) cos ( q 1 ) + L 10 sin ( Rot ) cos ( TabAng − θ 6 ) − L 6 sin ( q 1 + q 2 + θ 3 ) cos ( Rot ) + L 2 cos ( Rot ) cos ( q 1 ) cos ( q 2 ) − L 2 cos ( Rot ) sin ( q 1 ) sin ( q 2 ) + L 10 sin ( q 1 + q 2 + θ 3 ) cos ( Rot ) sin ( TabAng − θ 6 ) + L 3 cos ( q 1 + q 2 ) cos ( Rot ) cos ( θ 3 ) 2 − L 3 sin ( q 1 + q 2 ) cos ( Rot ) sin ( θ 3 ) 2 (51) p y = L R + L 2 sin ( q 1 + q 2 ) + L 1 sin ( q 1 ) + L 6 cos ( q 1 + q 2 + θ 3 ) + L 3 sin ( q 1 + q 2 + θ 3 ) 2 − L 10 sin ( TabAng + q 1 + q 2 + θ 3 − θ 6 ) 2 + L 10 sin ( q 1 − TabAng + q 2 + θ 3 + θ 6 ) 2 (52) p z = lin − L + L 7 cos ( Rot ) + L L cos ( Rot ) + L 0 sin ( Rot ) 2 − L 1 sin ( Rot ) cos ( q 1 ) + L 10 cos ( Rot ) cos ( TabAng − θ 6 ) + L 6 sin ( q 1 + q 2 + θ 3 ) sin ( Rot ) + L 3 sin ( q 1 + q 2 ) sin ( Rot ) sin ( θ 3 ) 2 − L 2 sin ( Rot ) cos ( q 1 ) cos ( q 2 ) + L 2 sin ( Rot ) sin ( q 1 ) sin ( q 2 ) − L 10 sin ( q 1 + q 2 + θ 3 ) sin ( Rot ) sin ( TabAng − θ 6 ) − L 3 cos ( q 1 + q 2 ) sin ( Rot ) cos ( θ 3 ) 2 From Equation , we can extract Euler angles using . (53) α = tan − 1 sin α cos α = tan − 1 cos γ · sin α cos α · cos γ = tan − 1 − R xzy , 2 , 3 R xzy , 2 , 2 = tan − 1 − R 2 , 3 R 2 , 2 (54) β = tan − 1 sin β cos β = tan − 1 cos γ · sin β cos β · cos γ = tan − 1 − R xzy , 3 , 1 R xzy , 1 , 1 = tan − 1 R 3 , 1 R 1 , 1 (55) γ = tan − 1 sin γ cos γ = tan − 1 sin γ ( cos α · cos γ ) 2 + ( − cos γ · sin α ) 2 = tan − 1 R xzy , 2 , 1 R xzy , 2 , 2 2 + R xzy , 2 , 3 2 = tan − 1 R 2 , 1 R 2 , 2 2 + R 2 , 3 2 (56) Pitch = 180 π · atan 2 sin ( TabAng + q 1 + q 2 + t h 3 ) / 2 − sin ( q 1 − TabAng + q 2 + t h 3 ) / 2 , cos ( TabAng + q 1 + q 2 + t h 3 ) / 2 + cos ( q 1 − TabAng + q 2 + t h 3 ) / 2 (57) Yaw = 180 π · atan 2 − cos ( q 1 + q 2 + t h 3 ) · sin ( Rot ) , cos ( q 1 + q 2 + t h 3 ) · cos ( Rot ) (58) Roll = 180 π · atan 2 sin ( q 1 + q 2 + t h 3 ) , cos ( T a b A n g + q 1 + q 2 + t h 3 ) 2 + cos ( q 1 − T a b A n g + q 2 + t h 3 ) 2 2 + sin ( T a b A n g + q 1 + q 2 + t h 3 ) 2 − sin ( q 1 − T a b A n g + q 2 + t h 3 ) 2 2
5.1. Hardware Components The patient positioning system (PPS) is built using state-of-the-art hardware components to ensure high precision and reliability. The primary components include the following: Kollmorgen Servo Motors: These motors provide high torque and precision, enabling smooth and accurate motion for the PPS across all axes. Renishaw Encoders: High-resolution 26-bit BiSS-C encoders (models RESA30USA200B and RA26BAA200B50F) are integrated into the system to deliver real-time feedback and enhance the accuracy of the positioning system. ACS Motion Controllers and Drivers: The system employs ACS controllers and drivers to manage the servo motors and process encoder data. These controllers enable synchronized multi-axis motion control, critical for the PPS’s six degrees of freedom (6-DOF) operation. 5.2. Control Software: SPiiPlus MMI Application Studio The SPiiPlus MMI Application Studio, developed by ACS Motion Control, serves as the primary interface and development environment for the PPS. This software is used to perform the following tasks: System Setup: Configuring servo motors, encoders, and controllers to ensure seamless integration of hardware components. Axis Calibration and Tuning: Implementing precise axis calibration and tuning the dual-loop feedback system for optimal performance. Real-Time Monitoring: Providing a robust platform for real-time monitoring of system parameters, including motor positions, encoder feedback, and velocity profiles. Diagnostics and Debugging: Enabling detailed diagnostics for troubleshooting and optimizing the system’s performance. Forward Kinematics Implementation: Programming and deploying the FK model to compute the exact position of the patient bed during normal operation and emergency interruptions. The SPiiPlus MMI Application Studio acts as a central hub, facilitating the efficient development and management of the PPS. It provides tools such as multi-channel scopes and wizards for axis configuration, enhancing the accuracy and reliability of the system. 5.3. Data Acquisition and Analysis To evaluate the system’s performance during interruptions, real-time data are gathered from the ACS Motion Controllers using the SPiiPlus MMI Application Studio. The data include the following: Motor positions and velocities obtained from primary encoders. Load positions tracked by secondary encoders. System behavior during simulated interruption scenarios. These data are exported and analyzed using MATLAB, where they are used to perform the following tasks: Visualize system responses to different interruption scenarios. Apply forward kinematics algorithms to estimate the system’s position during and after interruptions. Compare the FK-derived position with encoder feedback to evaluate accuracy and reliability. 5.4. Testing Forward Kinematics During Interruptions The primary goal of this study is to verify the FK model’s ability to accurately determine the patient bed’s position during emergency scenarios, such as power loss or emergency stops. The steps include the following: Simulate various interruption scenarios using the SPiiPlus MMI Application Studio. Gather real-time encoder feedback during the interruptions. Plot and analyze the data in MATLAB to assess the system’s recovery trajectory. Validate the FK model by comparing its positional estimates with actual encoder data.
The patient positioning system (PPS) is built using state-of-the-art hardware components to ensure high precision and reliability. The primary components include the following: Kollmorgen Servo Motors: These motors provide high torque and precision, enabling smooth and accurate motion for the PPS across all axes. Renishaw Encoders: High-resolution 26-bit BiSS-C encoders (models RESA30USA200B and RA26BAA200B50F) are integrated into the system to deliver real-time feedback and enhance the accuracy of the positioning system. ACS Motion Controllers and Drivers: The system employs ACS controllers and drivers to manage the servo motors and process encoder data. These controllers enable synchronized multi-axis motion control, critical for the PPS’s six degrees of freedom (6-DOF) operation.
The SPiiPlus MMI Application Studio, developed by ACS Motion Control, serves as the primary interface and development environment for the PPS. This software is used to perform the following tasks: System Setup: Configuring servo motors, encoders, and controllers to ensure seamless integration of hardware components. Axis Calibration and Tuning: Implementing precise axis calibration and tuning the dual-loop feedback system for optimal performance. Real-Time Monitoring: Providing a robust platform for real-time monitoring of system parameters, including motor positions, encoder feedback, and velocity profiles. Diagnostics and Debugging: Enabling detailed diagnostics for troubleshooting and optimizing the system’s performance. Forward Kinematics Implementation: Programming and deploying the FK model to compute the exact position of the patient bed during normal operation and emergency interruptions. The SPiiPlus MMI Application Studio acts as a central hub, facilitating the efficient development and management of the PPS. It provides tools such as multi-channel scopes and wizards for axis configuration, enhancing the accuracy and reliability of the system.
To evaluate the system’s performance during interruptions, real-time data are gathered from the ACS Motion Controllers using the SPiiPlus MMI Application Studio. The data include the following: Motor positions and velocities obtained from primary encoders. Load positions tracked by secondary encoders. System behavior during simulated interruption scenarios. These data are exported and analyzed using MATLAB, where they are used to perform the following tasks: Visualize system responses to different interruption scenarios. Apply forward kinematics algorithms to estimate the system’s position during and after interruptions. Compare the FK-derived position with encoder feedback to evaluate accuracy and reliability.
The primary goal of this study is to verify the FK model’s ability to accurately determine the patient bed’s position during emergency scenarios, such as power loss or emergency stops. The steps include the following: Simulate various interruption scenarios using the SPiiPlus MMI Application Studio. Gather real-time encoder feedback during the interruptions. Plot and analyze the data in MATLAB to assess the system’s recovery trajectory. Validate the FK model by comparing its positional estimates with actual encoder data.
The patient positioning system (PPS) was evaluated under three simulated interruption scenarios: power loss, emergency stop, and communication failure. These scenarios were tested during linear motion along the X-axis, Y-axis, and during 3D motion incorporating angular adjustments (Pitch, Yaw, Roll). The results emphasize the system’s ability to recover accurately from interruptions while maintaining precision and stability. 6.1. System Response During X-Axis Motion illustrates the PPS’s response during X-axis motion with the Y and Z positions fixed at 1100 mm and 1000 mm, respectively. The interruptions—a power loss at 20% of the trajectory, emergency stop at 50%, and communication failure at 80%—temporarily disrupted the motion. However, the system promptly resumed its planned trajectory after each interruption, maintaining sub-millimeter positional accuracy. 6.2. System Response During Y-Axis Motion The system’s behavior during Y-axis motion with fixed X and Z positions is shown in . The same interruption scenarios were simulated, and the resumed paths consistently aligned with the planned trajectory. This demonstrates the system’s ability to recover reliably, ensuring precision after disruptions. 6.3. System Response During 3D Motion with Angular Adjustments presents the system’s response during 3D motion, incorporating angular adjustments (Pitch, Yaw, Roll). The recorded angles at the interruption points reflect temporary deviations, but the system quickly regained angular stability. The maximum angular deviation observed was 0.1 ∘ , demonstrating effective recovery while maintaining stability. 6.4. Performance Metrics Summary The system’s performance across all scenarios was quantified using key metrics, as summarized in . Positional accuracy was evaluated based on the deviation from the planned trajectory, while angular stability was assessed using deviations in Pitch, Yaw, and Roll. Recovery time was measured as the duration required to resume the planned trajectory after an interruption. These results confirm the PPS’s ability to maintain high precision and stability under interruption scenarios. The system achieved a mean positional error of 0.05 mm, with a maximum deviation of 0.1 mm, and angular deviations remained under 0.1 ∘ with a mean of 0.03 ∘ . Recovery times averaged 2.0 s, ensuring minimal delay in operation. The combination of rapid recovery, positional accuracy, and angular stability demonstrates the robustness of the PPS, making it highly suitable for radiosurgery applications.
illustrates the PPS’s response during X-axis motion with the Y and Z positions fixed at 1100 mm and 1000 mm, respectively. The interruptions—a power loss at 20% of the trajectory, emergency stop at 50%, and communication failure at 80%—temporarily disrupted the motion. However, the system promptly resumed its planned trajectory after each interruption, maintaining sub-millimeter positional accuracy.
The system’s behavior during Y-axis motion with fixed X and Z positions is shown in . The same interruption scenarios were simulated, and the resumed paths consistently aligned with the planned trajectory. This demonstrates the system’s ability to recover reliably, ensuring precision after disruptions.
presents the system’s response during 3D motion, incorporating angular adjustments (Pitch, Yaw, Roll). The recorded angles at the interruption points reflect temporary deviations, but the system quickly regained angular stability. The maximum angular deviation observed was 0.1 ∘ , demonstrating effective recovery while maintaining stability.
The system’s performance across all scenarios was quantified using key metrics, as summarized in . Positional accuracy was evaluated based on the deviation from the planned trajectory, while angular stability was assessed using deviations in Pitch, Yaw, and Roll. Recovery time was measured as the duration required to resume the planned trajectory after an interruption. These results confirm the PPS’s ability to maintain high precision and stability under interruption scenarios. The system achieved a mean positional error of 0.05 mm, with a maximum deviation of 0.1 mm, and angular deviations remained under 0.1 ∘ with a mean of 0.03 ∘ . Recovery times averaged 2.0 s, ensuring minimal delay in operation. The combination of rapid recovery, positional accuracy, and angular stability demonstrates the robustness of the PPS, making it highly suitable for radiosurgery applications.
The results validate the system’s robustness in managing operational interruptions, demonstrating its ability to maintain precision and reliability under various scenarios. The dual-loop encoder feedback system played a pivotal role in ensuring accurate position and angular data retention, enabling precise trajectory resumption after power loss, emergency stops, and communication failures. Safe halting mechanisms effectively prevented positional drift during pauses, while secondary encoder feedback facilitated uninterrupted recovery, even under partial communication loss. These findings confirm the system’s capability for sub-millimeter positional accuracy and angular stability, making it highly suitable for high-precision applications in clinical environments. In comparison with existing patient positioning systems (PPS), our robotic system demonstrates notable advantages. Commercial systems such as the 6-DoF Robotic Couch by gKteso GmbH and the Hexapod solutions by PI (Physik Instrumente) offer precision and reliability, with reported accuracies of up to 0.5 mm . However, these systems are limited by their narrower travel ranges and reduced adaptability during treatment. In contrast, our system achieves a superior accuracy of 0.1 mm and features an extended travel range, enabling greater flexibility in handling complex treatment geometries. Furthermore, while systems developed by Siemens and Samsung focus on load capacity and integration with imaging technologies, they lack dynamic positioning adjustments during treatment . By incorporating dual-loop feedback and forward kinematics (FK) for real-time trajectory adjustments, our system overcomes these limitations and establishes a new benchmark for adaptability and precision. The effectiveness of our approach is further supported by existing research. Studies by Lan et al. and Kim et al. emphasize the importance of integrating primary and secondary encoder feedback to enhance system accuracy and reliability. Our implementation aligns with these findings, showcasing how redundancy in feedback mechanisms can mitigate the impact of communication failures and improve recovery performance. Additionally, Johnson and Patel (2022) highlight that dual-loop control architectures significantly outperform single-loop systems in maintaining alignment and stability . Nguyen and Garcia (2021) further underscore the importance of robust safety mechanisms to improve patient outcomes , a consideration that our system addresses through its safe halting and recovery protocols . Despite these strengths, certain limitations warrant further investigation. The system’s reliance on high-resolution encoder feedback introduces potential sensitivity to noise, which may require the development of advanced filtering algorithms to maintain consistent performance. Moreover, while the system performed reliably under controlled laboratory conditions, additional validation in dynamic clinical environments is essential to evaluate its robustness under real-world scenarios. This includes testing with variable patient anatomies, weights, and treatment trajectories. Finally, while the current design is optimized for intracranial procedures, its modular nature offers opportunities for adaptation to broader applications, such as spinal or thoracic treatments, through tailored hardware and software configurations. Clinically, the implications of this work are significant. By maintaining sub-millimeter accuracy during interruptions, the system minimizes radiation exposure to healthy tissues and enhances treatment efficacy. The rapid recovery mechanisms reduce treatment delays, contributing to improved operational efficiency and patient comfort. These features position the system as a valuable advancement in robotic radiosurgery, addressing key challenges in precision, safety, and reliability. With future enhancements to algorithms, noise mitigation, and broader clinical validations, the system holds promise for further advancing the capabilities of medical robotics in high-stakes clinical settings beyond radiosurgery.
This study validates the performance and reliability of a robotic patient positioning system for radiosurgery, demonstrating its ability to maintain precision and recover seamlessly from operational interruptions. The system achieved sub-millimeter positional accuracy during and after interruptions, with robust recovery mechanisms ensuring the precise resumption of the trajectory following power loss, emergency stops, and communication failures. The integration of forward kinematics (FK) with real-time encoder feedback enabled reliable position and orientation tracking, both during normal operations and under interruption scenarios. When compared to other systems currently in use, the proposed solution offers distinct advantages in terms of accuracy, stability, and recovery performance. By combining a 0.1 mm absolute accuracy with a wide travel range and advanced control architecture, the system addresses the critical limitations of existing robotic solutions. This capability ensures enhanced patient safety, improved clinical outcomes, and minimized treatment delays, making it a reliable and competitive solution for high-stakes clinical settings. Future work will build upon these findings by expanding the system’s applications beyond intracranial procedures. While the current validation focused on brain tumor treatments, we aim to extend the system’s functionality to address lung tumors. This application presents unique challenges due to respiratory motion, requiring the integration of real-time tracking and advanced imaging techniques for dynamic tumor localization. Enhancing the system’s algorithms to process real-time imaging data and synchronize with the patient’s breathing cycle will be a key focus. Such developments will further validate the system’s robustness and versatility, ensuring it meets the demanding requirements of treating complex and mobile targets in radiosurgery. These advancements will position the proposed system at the forefront of precision robotics for medical applications, offering broader utility across a wide range of clinical scenarios while setting new benchmarks in safety, accuracy, and reliability.
The patient positioning system (PPS) is presently undergoing the patenting process, holding the application number PCT/US2019/048205.
|
A randomized trial to evaluate a complex, co-created, culture-sensitive intervention to promote healthy lifestyles and compliance to therapy in immigrants with type 2 diabetes: A protocol of a multicenter Italian study | b2c85166-a33f-4270-a3b0-7c1864ceca7c | 11849826 | Health Promotion[mh] | 1.1. Background and rationale 1.1.1. Health problem. Diabetes is one of the most common chronic diseases in high-income countries. In Italy, its prevalence is more than 6% in the general population and is over 20% in those aged 50 or more. The burden of disease attributable to high blood glucose in Italy is 665.000 DALY (557.000 to 792.000) in 2021 (DALY are defined as the sum of years of life lost and years lived with disability) . A large part of the health system resources is dedicated to diabetes control and the care of diabetes complications. The burden of diabetes disproportionately affects immigrants, in Italy as well as in most Western European countries . When considering the different age distribution, diabetes prevalence is much higher in immigrants than in native populations, particularly in communities coming from South Asia, Africa, and, with less strength, the Caribbean [ – ]. Furthermore, different studies showed poorer glycemic control and higher incidence of diabetes complications, and cardiovascular and kidney diseases. There is a complex network of causes behind the higher prevalence and the worst diabetes control in immigrants. Despite there are pieces of evidence that the high prevalence in some ethnic groups may be due to genetic factors, the variability across hosting countries and care systems, as well changes in disease control with changes of diet and lifestyles after the arrival in host countries proof that most causal factors are modifiable. Potential factors underlying the high risk of type 2 diabetes and its related complications in migrants are multifaceted, including pre and post migration factors. Pre-migration factors include intrauterine growth, parental socioeconomic status [SES], health behaviors, while post-migration factors are the contextual factors in the host countries, lifestyle changes, health systems and related policies. It has been hypothesized that all of these factors can influence socioeconomic circumstances, behavior and biological factors, access to healthcare, physical and psychosocial stress and epigenetics upon migration, which in turn affect insulin secretion and action and subsequently type 2 diabetes risk . Finally, also the hard experience often experimented during journeys can play a role, both directly, generating post-traumatic stress disorders (PTSD) and indirectly, determining the adoption of risky behaviors such as smoking, alcohol intake, or sedentary lifestyles as coping strategies. Several studies investigated the barriers to effective care and prevention, including providing pharmacological therapy to control glycaemia and lifestyle counselling. Barriers include cultural and linguistic barriers making it difficult to navigate the health system and access chronic care services, inability of the health service to provide tailored health promotion, and logistical and economic difficulties limiting the time for care, physical activity, and purchasing appropriate foods. These barriers exist even in universal health services when facing newly arrived immigrants. Furthermore, undocumented immigrants have no access to chronic disease care in most countries . 1.1.2. Available interventions and their limits. In exploring the literature on elements of complex interventions we tried to take into account their sociocultural acceptability and feasibility through the questions suggested by Booth A. et al 2019 in their analysis of an alternative framework for qualitative evidence syntheses for exploring the effects of complex interventions . Recent systematic reviews showed that few interventions are effective in removing these barriers and in improving diabetes control in immigrants . Tested interventions usually target diet, physical activity, and compliance with therapeutic plans. Culturally appropriate diabetes health education in ethnic minority groups showed significant improvements in diabetes control. Complex interventions, acting on more factors at both individual and environmental levels, showed a higher probability of success . Group-based intervention in people with type 2 diabetes results in improvements in clinical and non-clinical outcomes and it also seems a promising approach for people with a migration background . The literature highlights some recurrent limits in the proposed interventions. Some interventions often are not intensive enough to be effective in changing the environment (the family or the community) to make it easier to adhere to the diet, physical activity and therapeutic recommendations. On the opposite, other interventions are too intensive and invasive to be acceptable and scalable. It can be challenging to tailor the proposed actions on the patient’s individual and cultural needs, time and economic constraints but the effectiveness of interventions that take precisely these aspects into account is noticeable and co-creation strategies (including patients’ involvement) can help to achieve this goal . 1.1.3. Methodological issues. The scarcity of evidence-based effective interventions is also due to the difficulties in conducting methodologically sound studies. Testing the efficacy of single components of a complex intervention, with a reductionist approach in which the intervention and the control arm differ only for a specific characteristic will probably produce a too small contrast between the two arms to detect an effect. Furthermore, the reductionist approach will prevent measuring the synergic effects emerging from the interactions between different components of a complex intervention . On the opposite side, randomized trials on complex interventions as whole are difficult to design and conduct. Individual randomization has intrinsic limits in measuring the components of the intervention targeting the environment. Furthermore, a full commitment of the professionals involved is crucial for the effectiveness of culturally sensitive interventions but is difficult to obtain when the intervention is randomly applied only to one-half of their patients. Cluster randomized trials can partially overcome these issues; nevertheless, they are difficult to conduct under the current legislation regulating clinical research and data treatment in Italy. Finally, standardizing the intervention, as usually required in clinical research, would be unfeasible in multicenter studies, because complex interventions need to be adapted to the context . Thus, a standardized complex intervention would probably be less effective, feasible and scalable than a context-adapted ones, reducing the external validity of the study [ – ]. These considerations led us to evaluate different kinds of interventions according to their validity, the attendance to co-creation principles, and the possibility of evaluating them ( ). The first scenario plans to apply the same intervention to all centers. This would lead to difficulties in defining a contrast with usual care in centers where the intervention would be very similar to what is already in place. The transferability of the results (external validity) would not be analyzable as each center had the same intervention. The intervention would not adapt to the context on the basis of a co-created process but it would be implemented by default. It would be difficult to evaluate the effectiveness of the intervention in the absence of contrast within centers and between centers. These limits would be partly overcome with an uniform among centers and incremental intervention, but they are completely overcome only with an incremental intervention with an increase proportional to the standard of care in each center. An incremental intervention, whose components increase proportionally to the standard of care in each center seems to us the better solution. 1.1.4. Co-creation. Co-creation, in the context of a public service design, refers to a process in which the provider collects input from final users that plays a central role in defining the service and how to deliver it . European commission recommends adopting co-creation in designing services to the citizens to bridge the gap between their needs and the public services . A Co-created approach has been applied in health services and particularly to design interventions for chronic care where the goal is improving patient quality of life in the long period through self-medication and adherence to treatments . It has been also proposed to offer services to the hard-to-reach or hard to follow up populations . The co-creative process, through the involvement of final users and providers, should facilitate the design of tailored, cultural sensitive, feasible, and acceptable interventions. Nevertheless, balancing the need to adopt only evidence-based interventions and the patients’ requests requires a clear distinction between what is the object of co-creation and what is a matter of scientific debate, and this is not an easy task [ – ]. 1.2. Objectives The aim of the study is to evaluate the efficacy of a co-created, culture-sensitive intervention to promote a healthy diet, and physical activity, and to improve compliance to therapeutic protocols in immigrants with type 2 diabetes. 1.3. Trial design We will conduct a multicenter randomized controlled trial comparing the effectiveness of a tailored co-created intervention in immigrants with type two uncontrolled diabetes in the context of the project “Cardio-metabolic diseases in immigrants and ethnic minorities: from epidemiology to new prevention strategies” funded by the European Union (NextGenerationEU). Four centers take part in the project: Careggi University Hospital (Florence), Local Health Authority of Reggio Emilia, National Institute for Heath, Migration and Poverty (INMP, Rome), Dulbecco University Hospital (Catanzaro). The main outcome is glucose control measured as glycated hemoglobin (HbA1c). We will evaluate a complex intervention based on three pillars, diet, physical activity, and adherence to therapeutic protocols, vs. the usual care ( ). The components of the complex intervention will be standardized by defining their functions , as opposite of by their form, giving the opportunity to adapt the form of the actions and tools to the context in each center. Usual care is also center-dependent. Therefore, we will also adapt the intervention according to the effort needed to increment the usual care to the intervention, i.e., each center will include in the intervention arm the components that will be affordable with a similar incremental effort ( ).
1.1.1. Health problem. Diabetes is one of the most common chronic diseases in high-income countries. In Italy, its prevalence is more than 6% in the general population and is over 20% in those aged 50 or more. The burden of disease attributable to high blood glucose in Italy is 665.000 DALY (557.000 to 792.000) in 2021 (DALY are defined as the sum of years of life lost and years lived with disability) . A large part of the health system resources is dedicated to diabetes control and the care of diabetes complications. The burden of diabetes disproportionately affects immigrants, in Italy as well as in most Western European countries . When considering the different age distribution, diabetes prevalence is much higher in immigrants than in native populations, particularly in communities coming from South Asia, Africa, and, with less strength, the Caribbean [ – ]. Furthermore, different studies showed poorer glycemic control and higher incidence of diabetes complications, and cardiovascular and kidney diseases. There is a complex network of causes behind the higher prevalence and the worst diabetes control in immigrants. Despite there are pieces of evidence that the high prevalence in some ethnic groups may be due to genetic factors, the variability across hosting countries and care systems, as well changes in disease control with changes of diet and lifestyles after the arrival in host countries proof that most causal factors are modifiable. Potential factors underlying the high risk of type 2 diabetes and its related complications in migrants are multifaceted, including pre and post migration factors. Pre-migration factors include intrauterine growth, parental socioeconomic status [SES], health behaviors, while post-migration factors are the contextual factors in the host countries, lifestyle changes, health systems and related policies. It has been hypothesized that all of these factors can influence socioeconomic circumstances, behavior and biological factors, access to healthcare, physical and psychosocial stress and epigenetics upon migration, which in turn affect insulin secretion and action and subsequently type 2 diabetes risk . Finally, also the hard experience often experimented during journeys can play a role, both directly, generating post-traumatic stress disorders (PTSD) and indirectly, determining the adoption of risky behaviors such as smoking, alcohol intake, or sedentary lifestyles as coping strategies. Several studies investigated the barriers to effective care and prevention, including providing pharmacological therapy to control glycaemia and lifestyle counselling. Barriers include cultural and linguistic barriers making it difficult to navigate the health system and access chronic care services, inability of the health service to provide tailored health promotion, and logistical and economic difficulties limiting the time for care, physical activity, and purchasing appropriate foods. These barriers exist even in universal health services when facing newly arrived immigrants. Furthermore, undocumented immigrants have no access to chronic disease care in most countries . 1.1.2. Available interventions and their limits. In exploring the literature on elements of complex interventions we tried to take into account their sociocultural acceptability and feasibility through the questions suggested by Booth A. et al 2019 in their analysis of an alternative framework for qualitative evidence syntheses for exploring the effects of complex interventions . Recent systematic reviews showed that few interventions are effective in removing these barriers and in improving diabetes control in immigrants . Tested interventions usually target diet, physical activity, and compliance with therapeutic plans. Culturally appropriate diabetes health education in ethnic minority groups showed significant improvements in diabetes control. Complex interventions, acting on more factors at both individual and environmental levels, showed a higher probability of success . Group-based intervention in people with type 2 diabetes results in improvements in clinical and non-clinical outcomes and it also seems a promising approach for people with a migration background . The literature highlights some recurrent limits in the proposed interventions. Some interventions often are not intensive enough to be effective in changing the environment (the family or the community) to make it easier to adhere to the diet, physical activity and therapeutic recommendations. On the opposite, other interventions are too intensive and invasive to be acceptable and scalable. It can be challenging to tailor the proposed actions on the patient’s individual and cultural needs, time and economic constraints but the effectiveness of interventions that take precisely these aspects into account is noticeable and co-creation strategies (including patients’ involvement) can help to achieve this goal . 1.1.3. Methodological issues. The scarcity of evidence-based effective interventions is also due to the difficulties in conducting methodologically sound studies. Testing the efficacy of single components of a complex intervention, with a reductionist approach in which the intervention and the control arm differ only for a specific characteristic will probably produce a too small contrast between the two arms to detect an effect. Furthermore, the reductionist approach will prevent measuring the synergic effects emerging from the interactions between different components of a complex intervention . On the opposite side, randomized trials on complex interventions as whole are difficult to design and conduct. Individual randomization has intrinsic limits in measuring the components of the intervention targeting the environment. Furthermore, a full commitment of the professionals involved is crucial for the effectiveness of culturally sensitive interventions but is difficult to obtain when the intervention is randomly applied only to one-half of their patients. Cluster randomized trials can partially overcome these issues; nevertheless, they are difficult to conduct under the current legislation regulating clinical research and data treatment in Italy. Finally, standardizing the intervention, as usually required in clinical research, would be unfeasible in multicenter studies, because complex interventions need to be adapted to the context . Thus, a standardized complex intervention would probably be less effective, feasible and scalable than a context-adapted ones, reducing the external validity of the study [ – ]. These considerations led us to evaluate different kinds of interventions according to their validity, the attendance to co-creation principles, and the possibility of evaluating them ( ). The first scenario plans to apply the same intervention to all centers. This would lead to difficulties in defining a contrast with usual care in centers where the intervention would be very similar to what is already in place. The transferability of the results (external validity) would not be analyzable as each center had the same intervention. The intervention would not adapt to the context on the basis of a co-created process but it would be implemented by default. It would be difficult to evaluate the effectiveness of the intervention in the absence of contrast within centers and between centers. These limits would be partly overcome with an uniform among centers and incremental intervention, but they are completely overcome only with an incremental intervention with an increase proportional to the standard of care in each center. An incremental intervention, whose components increase proportionally to the standard of care in each center seems to us the better solution. 1.1.4. Co-creation. Co-creation, in the context of a public service design, refers to a process in which the provider collects input from final users that plays a central role in defining the service and how to deliver it . European commission recommends adopting co-creation in designing services to the citizens to bridge the gap between their needs and the public services . A Co-created approach has been applied in health services and particularly to design interventions for chronic care where the goal is improving patient quality of life in the long period through self-medication and adherence to treatments . It has been also proposed to offer services to the hard-to-reach or hard to follow up populations . The co-creative process, through the involvement of final users and providers, should facilitate the design of tailored, cultural sensitive, feasible, and acceptable interventions. Nevertheless, balancing the need to adopt only evidence-based interventions and the patients’ requests requires a clear distinction between what is the object of co-creation and what is a matter of scientific debate, and this is not an easy task [ – ].
Diabetes is one of the most common chronic diseases in high-income countries. In Italy, its prevalence is more than 6% in the general population and is over 20% in those aged 50 or more. The burden of disease attributable to high blood glucose in Italy is 665.000 DALY (557.000 to 792.000) in 2021 (DALY are defined as the sum of years of life lost and years lived with disability) . A large part of the health system resources is dedicated to diabetes control and the care of diabetes complications. The burden of diabetes disproportionately affects immigrants, in Italy as well as in most Western European countries . When considering the different age distribution, diabetes prevalence is much higher in immigrants than in native populations, particularly in communities coming from South Asia, Africa, and, with less strength, the Caribbean [ – ]. Furthermore, different studies showed poorer glycemic control and higher incidence of diabetes complications, and cardiovascular and kidney diseases. There is a complex network of causes behind the higher prevalence and the worst diabetes control in immigrants. Despite there are pieces of evidence that the high prevalence in some ethnic groups may be due to genetic factors, the variability across hosting countries and care systems, as well changes in disease control with changes of diet and lifestyles after the arrival in host countries proof that most causal factors are modifiable. Potential factors underlying the high risk of type 2 diabetes and its related complications in migrants are multifaceted, including pre and post migration factors. Pre-migration factors include intrauterine growth, parental socioeconomic status [SES], health behaviors, while post-migration factors are the contextual factors in the host countries, lifestyle changes, health systems and related policies. It has been hypothesized that all of these factors can influence socioeconomic circumstances, behavior and biological factors, access to healthcare, physical and psychosocial stress and epigenetics upon migration, which in turn affect insulin secretion and action and subsequently type 2 diabetes risk . Finally, also the hard experience often experimented during journeys can play a role, both directly, generating post-traumatic stress disorders (PTSD) and indirectly, determining the adoption of risky behaviors such as smoking, alcohol intake, or sedentary lifestyles as coping strategies. Several studies investigated the barriers to effective care and prevention, including providing pharmacological therapy to control glycaemia and lifestyle counselling. Barriers include cultural and linguistic barriers making it difficult to navigate the health system and access chronic care services, inability of the health service to provide tailored health promotion, and logistical and economic difficulties limiting the time for care, physical activity, and purchasing appropriate foods. These barriers exist even in universal health services when facing newly arrived immigrants. Furthermore, undocumented immigrants have no access to chronic disease care in most countries .
In exploring the literature on elements of complex interventions we tried to take into account their sociocultural acceptability and feasibility through the questions suggested by Booth A. et al 2019 in their analysis of an alternative framework for qualitative evidence syntheses for exploring the effects of complex interventions . Recent systematic reviews showed that few interventions are effective in removing these barriers and in improving diabetes control in immigrants . Tested interventions usually target diet, physical activity, and compliance with therapeutic plans. Culturally appropriate diabetes health education in ethnic minority groups showed significant improvements in diabetes control. Complex interventions, acting on more factors at both individual and environmental levels, showed a higher probability of success . Group-based intervention in people with type 2 diabetes results in improvements in clinical and non-clinical outcomes and it also seems a promising approach for people with a migration background . The literature highlights some recurrent limits in the proposed interventions. Some interventions often are not intensive enough to be effective in changing the environment (the family or the community) to make it easier to adhere to the diet, physical activity and therapeutic recommendations. On the opposite, other interventions are too intensive and invasive to be acceptable and scalable. It can be challenging to tailor the proposed actions on the patient’s individual and cultural needs, time and economic constraints but the effectiveness of interventions that take precisely these aspects into account is noticeable and co-creation strategies (including patients’ involvement) can help to achieve this goal .
The scarcity of evidence-based effective interventions is also due to the difficulties in conducting methodologically sound studies. Testing the efficacy of single components of a complex intervention, with a reductionist approach in which the intervention and the control arm differ only for a specific characteristic will probably produce a too small contrast between the two arms to detect an effect. Furthermore, the reductionist approach will prevent measuring the synergic effects emerging from the interactions between different components of a complex intervention . On the opposite side, randomized trials on complex interventions as whole are difficult to design and conduct. Individual randomization has intrinsic limits in measuring the components of the intervention targeting the environment. Furthermore, a full commitment of the professionals involved is crucial for the effectiveness of culturally sensitive interventions but is difficult to obtain when the intervention is randomly applied only to one-half of their patients. Cluster randomized trials can partially overcome these issues; nevertheless, they are difficult to conduct under the current legislation regulating clinical research and data treatment in Italy. Finally, standardizing the intervention, as usually required in clinical research, would be unfeasible in multicenter studies, because complex interventions need to be adapted to the context . Thus, a standardized complex intervention would probably be less effective, feasible and scalable than a context-adapted ones, reducing the external validity of the study [ – ]. These considerations led us to evaluate different kinds of interventions according to their validity, the attendance to co-creation principles, and the possibility of evaluating them ( ). The first scenario plans to apply the same intervention to all centers. This would lead to difficulties in defining a contrast with usual care in centers where the intervention would be very similar to what is already in place. The transferability of the results (external validity) would not be analyzable as each center had the same intervention. The intervention would not adapt to the context on the basis of a co-created process but it would be implemented by default. It would be difficult to evaluate the effectiveness of the intervention in the absence of contrast within centers and between centers. These limits would be partly overcome with an uniform among centers and incremental intervention, but they are completely overcome only with an incremental intervention with an increase proportional to the standard of care in each center. An incremental intervention, whose components increase proportionally to the standard of care in each center seems to us the better solution.
Co-creation, in the context of a public service design, refers to a process in which the provider collects input from final users that plays a central role in defining the service and how to deliver it . European commission recommends adopting co-creation in designing services to the citizens to bridge the gap between their needs and the public services . A Co-created approach has been applied in health services and particularly to design interventions for chronic care where the goal is improving patient quality of life in the long period through self-medication and adherence to treatments . It has been also proposed to offer services to the hard-to-reach or hard to follow up populations . The co-creative process, through the involvement of final users and providers, should facilitate the design of tailored, cultural sensitive, feasible, and acceptable interventions. Nevertheless, balancing the need to adopt only evidence-based interventions and the patients’ requests requires a clear distinction between what is the object of co-creation and what is a matter of scientific debate, and this is not an easy task [ – ].
The aim of the study is to evaluate the efficacy of a co-created, culture-sensitive intervention to promote a healthy diet, and physical activity, and to improve compliance to therapeutic protocols in immigrants with type 2 diabetes.
We will conduct a multicenter randomized controlled trial comparing the effectiveness of a tailored co-created intervention in immigrants with type two uncontrolled diabetes in the context of the project “Cardio-metabolic diseases in immigrants and ethnic minorities: from epidemiology to new prevention strategies” funded by the European Union (NextGenerationEU). Four centers take part in the project: Careggi University Hospital (Florence), Local Health Authority of Reggio Emilia, National Institute for Heath, Migration and Poverty (INMP, Rome), Dulbecco University Hospital (Catanzaro). The main outcome is glucose control measured as glycated hemoglobin (HbA1c). We will evaluate a complex intervention based on three pillars, diet, physical activity, and adherence to therapeutic protocols, vs. the usual care ( ). The components of the complex intervention will be standardized by defining their functions , as opposite of by their form, giving the opportunity to adapt the form of the actions and tools to the context in each center. Usual care is also center-dependent. Therefore, we will also adapt the intervention according to the effort needed to increment the usual care to the intervention, i.e., each center will include in the intervention arm the components that will be affordable with a similar incremental effort ( ).
2.1. Participants, interventions, and outcomes 2.1.1. Study setting and study schedule. The study will be conducted in the primary care clinics specialized for the care of diabetes of the centres taking part of the project. The figure one showed the study schedule ( ). 2.1.2. Eligibility criteria. Adult (>=18) individuals with diabetes with an immigration background with a new diagnosis of diabetes or uncontrolled diabetes (HbA1c%>=8). Exclusion criteria include patients who will not provide the informed consent, patients aged under 18, patients with HbA1c ≤ 8% in the last assessment within 24 months before the visit, patients with severe psychiatric disorders, pregnant women, critical illness, impaired cognitive or physical ability that could make the intervention not feasible, as judged by clinical staff members. 2.1.3. Co-creation methods and description of the intervention. The intervention has been defined through a two-level co-creation process, a centralized level (coordinated by the Reggio Emilia Centre) that established the general principles and the possible components of a complex intervention, and a local level that decided which components were more useful and feasible in the local context and how they should be implemented. The result was a complex multicomponent intervention, in which each component was standardized by its function, tailored and culturally sensitive, context adapted and in which the contrast vs. the usual car was defined by similar incremental effort. The co-creation process included different strategies to involve and collect the voices and angles of users and providers. We adopted a co-creation approach (involving stakeholders and patients to identify the barriers and solutions). The strategies adopted in the centralised phase of co-creation included: Focus group, conducted in Reggio Emilia, with providers and lay persons from immigrant communities. Providers involved were selected among workers of municipality and cultural mediation services, third sector organizations caring for undocumented migrants, primary care services caring for people with diabetes, and public health services involved in health promotion and primary prevention. Lay people were contacted through participating providers and selected based on formal or informal knowledge of the diet and lifestyle of the communities involved. The groups worked on the analyses of existing barriers and facilitators in access to care and prevention and in adherence to behavioural and therapeutic recommendations. Interviews of the patients and healthcare workers in diabetes clinics. In each centre, patients in the waiting rooms of diabetes clinics were asked to take part in an interview involving five open-ended questions on possible barriers and facilitators of diabetes control both related to their daily lives and to the services of the healthcare system, developed on an ad hoc basis by the qualitative research team. The health care workers were invited to the same interview, suitably modified. Laboratory to match proposed solutions, literature evidence, and feasibility/sustainability. Informed by the two systematic reviews that we used for guidance. The results of this step are summarized in the , where we adopted a simplified DPSEEA model to conceptualize how we addressed actionable barriers or facilitators with components of the complex intervention. The local phase of the co-creation was based on local laboratories with key persons and providers that analyzed the current usual care, the available resources, and the components of the complex interventions defined at the centralized level for their feasibility at local level. The results of this phase are summarized in the ( ). Finally, the results of the local laboratories were discussed in a plenary session of the trial steering committee in Rome (October 13 th , 2023) to compose the best complex intervention in each local context given the available additional resources provided by the experimental project and the existing services. Results of the co-creation process are summarized in the . 2.1.4 Usual care. As mentioned above, the usual care was different in each centre. An analysis of the process was conducted in each centre through interviews and laboratories. During this analysis, it became clear that the implementation of the interventions would necessarily change some characteristics of the usual care; the local researchers made the effort to make explicit any unintended and unavoidable change of the usual care. The characteristics of the usual care are reported in . 2.1.5 Outcomes. The primary outcome is the change of HbA1c 12 months after recruitment. The secondary outcomes are the changes in: Anthropometric measures (BMI, waist circumference) Dietary habits Physical activity habits Lipid profile Compliance with individual therapeutic protocols The outcomes and the covariate measurements will be at recruitment and at follow up ( ). Between the baseline and the 12 months evaluation clinical data noted during routine diabetology visits are recorded, regardless of the randomization arm. 2.1.6. Sample size. To explore the efficacy of a culturally adapted diabetes education model in improving health literacy and self-care (primary endpoint) in immigrant patients with type 2 diabetes, the intervention is expected to provide a mean 0.5% higher reduction in the concentration of HbA1c in the intervention group compared to the usual care at a 12-month follow-up, with a positive effect on the risk of CVD events . Through a change in health-related behaviours, it is also expected to reduce overweight and increase vegetable intake and physical activity, with a positive impact on future risk of non-communicable diseases and quality of life . To have a power of 80%, considering an alpha of 0.05 , in order to detect a minimum significant reduction of 0.5% in the glycated haemoglobin in the intervention group compared to the control group at least 200 participants are needed (considering a SD in both group of 1.3) . 2.1.7. Recruitment. Participants will be asked to participate when attending a visit at one of the participating clinics. Whether low participation rates will occur, a preliminary eligibility assessment on medical records of Diabetes clinics will be performed, and potentially eligible patient will be actively contacted. Pre-recruitment eligibility criteria assessment from third sector organizations caring for undocumented migrants will also support active recruitment. Consecutive enrolment of potentially eligible patients started on the 7 th of November, 2023 and will last until reaching the anticipated sample size. 2.1.8. Assignment of interventions. Block randomization (1:1 ratio) will be computerised and conducted in each centre to balance the ethnic composition of the intervention and control groups. The random sequence was generated using the REDCap TM Random Sequence Generator. Members of the same household, whether identified as cohabitant during the recruitment visit, will be assigned at the same arm. The person who enrolls will be blinded to the random sequence until each patient will sign the informed consent. The randomization arm will not be masked to the participant nor to the investigator. We planned two strategies to reduce the potential performance bias arising from the lack of blinding of participants and investigators. The first was the choice of an objective outcome as primary outcome of efficacy in our study, namely the change of HbA1c 12 months after recruitment. The second was the involvement of experienced health professionals for the performance of the intervention, that will be sensitize and trained on good practices for clinical trials. Nevertheless, the qualitative information that will be available during the co-evaluation phase of the project will support us to assess the potential implication of performance bias on the trial’s results. 2.1.9. Data collection. Biomarkers will be assessed through analysis of blood and urine samples collected following standard procedures of the Diabetes clinics and analysed for glycaemia, by the authorised clinical laboratories of each recruiting centre. Questionnaires used to assess dietary habits and physical activity will be the Mediterranean Diet Score, (MedDietScore) and the International Physical Activity Questionnaire - Short Form, IPAQ-SF . Anthropometric measures will be collected following standard procedures during routine visits in Diabetes clinics with validated weight and height scales and bioelectrical impedance analysis (BIA). Therapeutic adherence will be assessed at baseline and follow up visits using the Diabetes Mellitus Treatment Adherence Scale (DMTAS) . Data will be recorded through the REDCap TM web application provided by the Principal Investigator to all the recruiting centres. Data access and management will be performed in compliance with the Italian and European privacy regulations. 2.1.10. Statistical plan. Descriptive statistics will be calculated for baseline characteristics. The analysis will be intention-to-treat. Before/after variations of the glycated haemoglobin, lipid profile and renal function biomarkers, and anthropometric measure will be computed and standardized if opportune. Paired and unpaired tests will be used to assess the effects of the intervention and to analyse before/after within-group and between-group differences and changes in glycated haemoglobin. Changes in lifestyle habits (both related to diet and physical activities) will be described in terms of positive or negative changes. Multivariate linear regression models will be used to analyse the variation of anthropometric measures and assessed biomarkers. Multilevel linear models will be performed taking into account the influence of the centre on intervention efficacy. We will perform a mediation analysis to understand which part of the changes in outcomes (glycated haemoglobin, lipid profile, and anthropometric measures) is attributable to changes in physical activity, changes in diet, adherence to therapies, and other direct or unmeasured effects of the intervention. The statistician will not be blinded to group assignments. The statistical significance level will be set at 5%, and all analyses will be performed by using Stata 16 or SPSS 28. 2.1.11. Subgroup analysis. Subgroup analyses will be performed according to weight status, gender, ethnicity, and socioeconomic level. 2.2. Patient and public involvement Patients and the local communities were actively involved in the development and design of the intervention as described in the “Co-creation methods and description of the intervention” section. Patients, immigrants, healthcare professional and representatives of social services will continue to be involved in the co-evaluation and dissemination phase. They will support the interpretation of study results, drafting plain language summaries, and co-presenting findings to lay people at community level. 2.3. Ethics and dissemination Diabethic trial is part of the “Cardio-metabolic diseases in immigrants and ethnic minorities: from epidemiology to new prevention strategies” project has been approved by the Italian Ministry of Health for funding to the call Piano Nazionale di Ripresa e Resilienza (PNRR): M6/C2_CALL 2022, Ministero della Salute, funded by the European Union. The role of the funding source is reported in the call of the Ministry of Health (Bando Piano Nazionale di Ripresa e Resilienza) available at https://www.agenziacoesione.gov.it/comunicazione/piano-nazionale-di-ripresa-e-resilienza/ https://ricerca.cbim.it/Documentazione . The Tuscany Regional Ethics committee “Comitato Etico Regionale per la Sperimentazione Clinica della Toscana - sezione AREA VASTA CENTRO” approved the trial protocol on November 29, 2022. ( ) This trial protocol has been registered with the ClinicalTrials.gov Registry (ClinicalTrials.gov ID NCT06131411). The protocol was framed according to the SPIRIT Reporting guidelines ( ) . The trial results will be available in 2025 and will undergone to a co-evaluation process involving all the stakeholders that contributed to the co-creation process. The co-evaluation of trial results aims to include all the different perspective to assess the actual impact of the whole process on patients and healthcare services. The co-evaluation may also support the implementation phase of the intervention, highlighting putative determinants of effectiveness and effect-modification that should be taken into account when adapting the intervention in different contexts. The results will be published in a peer-reviewed medical journal. 2.3.1. Informed consent. The eligible patients will be informed on the study and those interested in participating will be asked to sign informed consent during routine visits at Diabetes clinics. Patients aged < 18 years or unable to provide a personal informed consent are not eligible for the study.
2.1.1. Study setting and study schedule. The study will be conducted in the primary care clinics specialized for the care of diabetes of the centres taking part of the project. The figure one showed the study schedule ( ). 2.1.2. Eligibility criteria. Adult (>=18) individuals with diabetes with an immigration background with a new diagnosis of diabetes or uncontrolled diabetes (HbA1c%>=8). Exclusion criteria include patients who will not provide the informed consent, patients aged under 18, patients with HbA1c ≤ 8% in the last assessment within 24 months before the visit, patients with severe psychiatric disorders, pregnant women, critical illness, impaired cognitive or physical ability that could make the intervention not feasible, as judged by clinical staff members. 2.1.3. Co-creation methods and description of the intervention. The intervention has been defined through a two-level co-creation process, a centralized level (coordinated by the Reggio Emilia Centre) that established the general principles and the possible components of a complex intervention, and a local level that decided which components were more useful and feasible in the local context and how they should be implemented. The result was a complex multicomponent intervention, in which each component was standardized by its function, tailored and culturally sensitive, context adapted and in which the contrast vs. the usual car was defined by similar incremental effort. The co-creation process included different strategies to involve and collect the voices and angles of users and providers. We adopted a co-creation approach (involving stakeholders and patients to identify the barriers and solutions). The strategies adopted in the centralised phase of co-creation included: Focus group, conducted in Reggio Emilia, with providers and lay persons from immigrant communities. Providers involved were selected among workers of municipality and cultural mediation services, third sector organizations caring for undocumented migrants, primary care services caring for people with diabetes, and public health services involved in health promotion and primary prevention. Lay people were contacted through participating providers and selected based on formal or informal knowledge of the diet and lifestyle of the communities involved. The groups worked on the analyses of existing barriers and facilitators in access to care and prevention and in adherence to behavioural and therapeutic recommendations. Interviews of the patients and healthcare workers in diabetes clinics. In each centre, patients in the waiting rooms of diabetes clinics were asked to take part in an interview involving five open-ended questions on possible barriers and facilitators of diabetes control both related to their daily lives and to the services of the healthcare system, developed on an ad hoc basis by the qualitative research team. The health care workers were invited to the same interview, suitably modified. Laboratory to match proposed solutions, literature evidence, and feasibility/sustainability. Informed by the two systematic reviews that we used for guidance. The results of this step are summarized in the , where we adopted a simplified DPSEEA model to conceptualize how we addressed actionable barriers or facilitators with components of the complex intervention. The local phase of the co-creation was based on local laboratories with key persons and providers that analyzed the current usual care, the available resources, and the components of the complex interventions defined at the centralized level for their feasibility at local level. The results of this phase are summarized in the ( ). Finally, the results of the local laboratories were discussed in a plenary session of the trial steering committee in Rome (October 13 th , 2023) to compose the best complex intervention in each local context given the available additional resources provided by the experimental project and the existing services. Results of the co-creation process are summarized in the . 2.1.4 Usual care. As mentioned above, the usual care was different in each centre. An analysis of the process was conducted in each centre through interviews and laboratories. During this analysis, it became clear that the implementation of the interventions would necessarily change some characteristics of the usual care; the local researchers made the effort to make explicit any unintended and unavoidable change of the usual care. The characteristics of the usual care are reported in . 2.1.5 Outcomes. The primary outcome is the change of HbA1c 12 months after recruitment. The secondary outcomes are the changes in: Anthropometric measures (BMI, waist circumference) Dietary habits Physical activity habits Lipid profile Compliance with individual therapeutic protocols The outcomes and the covariate measurements will be at recruitment and at follow up ( ). Between the baseline and the 12 months evaluation clinical data noted during routine diabetology visits are recorded, regardless of the randomization arm. 2.1.6. Sample size. To explore the efficacy of a culturally adapted diabetes education model in improving health literacy and self-care (primary endpoint) in immigrant patients with type 2 diabetes, the intervention is expected to provide a mean 0.5% higher reduction in the concentration of HbA1c in the intervention group compared to the usual care at a 12-month follow-up, with a positive effect on the risk of CVD events . Through a change in health-related behaviours, it is also expected to reduce overweight and increase vegetable intake and physical activity, with a positive impact on future risk of non-communicable diseases and quality of life . To have a power of 80%, considering an alpha of 0.05 , in order to detect a minimum significant reduction of 0.5% in the glycated haemoglobin in the intervention group compared to the control group at least 200 participants are needed (considering a SD in both group of 1.3) . 2.1.7. Recruitment. Participants will be asked to participate when attending a visit at one of the participating clinics. Whether low participation rates will occur, a preliminary eligibility assessment on medical records of Diabetes clinics will be performed, and potentially eligible patient will be actively contacted. Pre-recruitment eligibility criteria assessment from third sector organizations caring for undocumented migrants will also support active recruitment. Consecutive enrolment of potentially eligible patients started on the 7 th of November, 2023 and will last until reaching the anticipated sample size. 2.1.8. Assignment of interventions. Block randomization (1:1 ratio) will be computerised and conducted in each centre to balance the ethnic composition of the intervention and control groups. The random sequence was generated using the REDCap TM Random Sequence Generator. Members of the same household, whether identified as cohabitant during the recruitment visit, will be assigned at the same arm. The person who enrolls will be blinded to the random sequence until each patient will sign the informed consent. The randomization arm will not be masked to the participant nor to the investigator. We planned two strategies to reduce the potential performance bias arising from the lack of blinding of participants and investigators. The first was the choice of an objective outcome as primary outcome of efficacy in our study, namely the change of HbA1c 12 months after recruitment. The second was the involvement of experienced health professionals for the performance of the intervention, that will be sensitize and trained on good practices for clinical trials. Nevertheless, the qualitative information that will be available during the co-evaluation phase of the project will support us to assess the potential implication of performance bias on the trial’s results. 2.1.9. Data collection. Biomarkers will be assessed through analysis of blood and urine samples collected following standard procedures of the Diabetes clinics and analysed for glycaemia, by the authorised clinical laboratories of each recruiting centre. Questionnaires used to assess dietary habits and physical activity will be the Mediterranean Diet Score, (MedDietScore) and the International Physical Activity Questionnaire - Short Form, IPAQ-SF . Anthropometric measures will be collected following standard procedures during routine visits in Diabetes clinics with validated weight and height scales and bioelectrical impedance analysis (BIA). Therapeutic adherence will be assessed at baseline and follow up visits using the Diabetes Mellitus Treatment Adherence Scale (DMTAS) . Data will be recorded through the REDCap TM web application provided by the Principal Investigator to all the recruiting centres. Data access and management will be performed in compliance with the Italian and European privacy regulations. 2.1.10. Statistical plan. Descriptive statistics will be calculated for baseline characteristics. The analysis will be intention-to-treat. Before/after variations of the glycated haemoglobin, lipid profile and renal function biomarkers, and anthropometric measure will be computed and standardized if opportune. Paired and unpaired tests will be used to assess the effects of the intervention and to analyse before/after within-group and between-group differences and changes in glycated haemoglobin. Changes in lifestyle habits (both related to diet and physical activities) will be described in terms of positive or negative changes. Multivariate linear regression models will be used to analyse the variation of anthropometric measures and assessed biomarkers. Multilevel linear models will be performed taking into account the influence of the centre on intervention efficacy. We will perform a mediation analysis to understand which part of the changes in outcomes (glycated haemoglobin, lipid profile, and anthropometric measures) is attributable to changes in physical activity, changes in diet, adherence to therapies, and other direct or unmeasured effects of the intervention. The statistician will not be blinded to group assignments. The statistical significance level will be set at 5%, and all analyses will be performed by using Stata 16 or SPSS 28. 2.1.11. Subgroup analysis. Subgroup analyses will be performed according to weight status, gender, ethnicity, and socioeconomic level.
The study will be conducted in the primary care clinics specialized for the care of diabetes of the centres taking part of the project. The figure one showed the study schedule ( ).
Adult (>=18) individuals with diabetes with an immigration background with a new diagnosis of diabetes or uncontrolled diabetes (HbA1c%>=8). Exclusion criteria include patients who will not provide the informed consent, patients aged under 18, patients with HbA1c ≤ 8% in the last assessment within 24 months before the visit, patients with severe psychiatric disorders, pregnant women, critical illness, impaired cognitive or physical ability that could make the intervention not feasible, as judged by clinical staff members.
The intervention has been defined through a two-level co-creation process, a centralized level (coordinated by the Reggio Emilia Centre) that established the general principles and the possible components of a complex intervention, and a local level that decided which components were more useful and feasible in the local context and how they should be implemented. The result was a complex multicomponent intervention, in which each component was standardized by its function, tailored and culturally sensitive, context adapted and in which the contrast vs. the usual car was defined by similar incremental effort. The co-creation process included different strategies to involve and collect the voices and angles of users and providers. We adopted a co-creation approach (involving stakeholders and patients to identify the barriers and solutions). The strategies adopted in the centralised phase of co-creation included: Focus group, conducted in Reggio Emilia, with providers and lay persons from immigrant communities. Providers involved were selected among workers of municipality and cultural mediation services, third sector organizations caring for undocumented migrants, primary care services caring for people with diabetes, and public health services involved in health promotion and primary prevention. Lay people were contacted through participating providers and selected based on formal or informal knowledge of the diet and lifestyle of the communities involved. The groups worked on the analyses of existing barriers and facilitators in access to care and prevention and in adherence to behavioural and therapeutic recommendations. Interviews of the patients and healthcare workers in diabetes clinics. In each centre, patients in the waiting rooms of diabetes clinics were asked to take part in an interview involving five open-ended questions on possible barriers and facilitators of diabetes control both related to their daily lives and to the services of the healthcare system, developed on an ad hoc basis by the qualitative research team. The health care workers were invited to the same interview, suitably modified. Laboratory to match proposed solutions, literature evidence, and feasibility/sustainability. Informed by the two systematic reviews that we used for guidance. The results of this step are summarized in the , where we adopted a simplified DPSEEA model to conceptualize how we addressed actionable barriers or facilitators with components of the complex intervention. The local phase of the co-creation was based on local laboratories with key persons and providers that analyzed the current usual care, the available resources, and the components of the complex interventions defined at the centralized level for their feasibility at local level. The results of this phase are summarized in the ( ). Finally, the results of the local laboratories were discussed in a plenary session of the trial steering committee in Rome (October 13 th , 2023) to compose the best complex intervention in each local context given the available additional resources provided by the experimental project and the existing services. Results of the co-creation process are summarized in the .
As mentioned above, the usual care was different in each centre. An analysis of the process was conducted in each centre through interviews and laboratories. During this analysis, it became clear that the implementation of the interventions would necessarily change some characteristics of the usual care; the local researchers made the effort to make explicit any unintended and unavoidable change of the usual care. The characteristics of the usual care are reported in .
The primary outcome is the change of HbA1c 12 months after recruitment. The secondary outcomes are the changes in: Anthropometric measures (BMI, waist circumference) Dietary habits Physical activity habits Lipid profile Compliance with individual therapeutic protocols The outcomes and the covariate measurements will be at recruitment and at follow up ( ). Between the baseline and the 12 months evaluation clinical data noted during routine diabetology visits are recorded, regardless of the randomization arm.
To explore the efficacy of a culturally adapted diabetes education model in improving health literacy and self-care (primary endpoint) in immigrant patients with type 2 diabetes, the intervention is expected to provide a mean 0.5% higher reduction in the concentration of HbA1c in the intervention group compared to the usual care at a 12-month follow-up, with a positive effect on the risk of CVD events . Through a change in health-related behaviours, it is also expected to reduce overweight and increase vegetable intake and physical activity, with a positive impact on future risk of non-communicable diseases and quality of life . To have a power of 80%, considering an alpha of 0.05 , in order to detect a minimum significant reduction of 0.5% in the glycated haemoglobin in the intervention group compared to the control group at least 200 participants are needed (considering a SD in both group of 1.3) .
Participants will be asked to participate when attending a visit at one of the participating clinics. Whether low participation rates will occur, a preliminary eligibility assessment on medical records of Diabetes clinics will be performed, and potentially eligible patient will be actively contacted. Pre-recruitment eligibility criteria assessment from third sector organizations caring for undocumented migrants will also support active recruitment. Consecutive enrolment of potentially eligible patients started on the 7 th of November, 2023 and will last until reaching the anticipated sample size.
Block randomization (1:1 ratio) will be computerised and conducted in each centre to balance the ethnic composition of the intervention and control groups. The random sequence was generated using the REDCap TM Random Sequence Generator. Members of the same household, whether identified as cohabitant during the recruitment visit, will be assigned at the same arm. The person who enrolls will be blinded to the random sequence until each patient will sign the informed consent. The randomization arm will not be masked to the participant nor to the investigator. We planned two strategies to reduce the potential performance bias arising from the lack of blinding of participants and investigators. The first was the choice of an objective outcome as primary outcome of efficacy in our study, namely the change of HbA1c 12 months after recruitment. The second was the involvement of experienced health professionals for the performance of the intervention, that will be sensitize and trained on good practices for clinical trials. Nevertheless, the qualitative information that will be available during the co-evaluation phase of the project will support us to assess the potential implication of performance bias on the trial’s results.
Biomarkers will be assessed through analysis of blood and urine samples collected following standard procedures of the Diabetes clinics and analysed for glycaemia, by the authorised clinical laboratories of each recruiting centre. Questionnaires used to assess dietary habits and physical activity will be the Mediterranean Diet Score, (MedDietScore) and the International Physical Activity Questionnaire - Short Form, IPAQ-SF . Anthropometric measures will be collected following standard procedures during routine visits in Diabetes clinics with validated weight and height scales and bioelectrical impedance analysis (BIA). Therapeutic adherence will be assessed at baseline and follow up visits using the Diabetes Mellitus Treatment Adherence Scale (DMTAS) . Data will be recorded through the REDCap TM web application provided by the Principal Investigator to all the recruiting centres. Data access and management will be performed in compliance with the Italian and European privacy regulations.
Descriptive statistics will be calculated for baseline characteristics. The analysis will be intention-to-treat. Before/after variations of the glycated haemoglobin, lipid profile and renal function biomarkers, and anthropometric measure will be computed and standardized if opportune. Paired and unpaired tests will be used to assess the effects of the intervention and to analyse before/after within-group and between-group differences and changes in glycated haemoglobin. Changes in lifestyle habits (both related to diet and physical activities) will be described in terms of positive or negative changes. Multivariate linear regression models will be used to analyse the variation of anthropometric measures and assessed biomarkers. Multilevel linear models will be performed taking into account the influence of the centre on intervention efficacy. We will perform a mediation analysis to understand which part of the changes in outcomes (glycated haemoglobin, lipid profile, and anthropometric measures) is attributable to changes in physical activity, changes in diet, adherence to therapies, and other direct or unmeasured effects of the intervention. The statistician will not be blinded to group assignments. The statistical significance level will be set at 5%, and all analyses will be performed by using Stata 16 or SPSS 28.
Subgroup analyses will be performed according to weight status, gender, ethnicity, and socioeconomic level.
Patients and the local communities were actively involved in the development and design of the intervention as described in the “Co-creation methods and description of the intervention” section. Patients, immigrants, healthcare professional and representatives of social services will continue to be involved in the co-evaluation and dissemination phase. They will support the interpretation of study results, drafting plain language summaries, and co-presenting findings to lay people at community level.
Diabethic trial is part of the “Cardio-metabolic diseases in immigrants and ethnic minorities: from epidemiology to new prevention strategies” project has been approved by the Italian Ministry of Health for funding to the call Piano Nazionale di Ripresa e Resilienza (PNRR): M6/C2_CALL 2022, Ministero della Salute, funded by the European Union. The role of the funding source is reported in the call of the Ministry of Health (Bando Piano Nazionale di Ripresa e Resilienza) available at https://www.agenziacoesione.gov.it/comunicazione/piano-nazionale-di-ripresa-e-resilienza/ https://ricerca.cbim.it/Documentazione . The Tuscany Regional Ethics committee “Comitato Etico Regionale per la Sperimentazione Clinica della Toscana - sezione AREA VASTA CENTRO” approved the trial protocol on November 29, 2022. ( ) This trial protocol has been registered with the ClinicalTrials.gov Registry (ClinicalTrials.gov ID NCT06131411). The protocol was framed according to the SPIRIT Reporting guidelines ( ) . The trial results will be available in 2025 and will undergone to a co-evaluation process involving all the stakeholders that contributed to the co-creation process. The co-evaluation of trial results aims to include all the different perspective to assess the actual impact of the whole process on patients and healthcare services. The co-evaluation may also support the implementation phase of the intervention, highlighting putative determinants of effectiveness and effect-modification that should be taken into account when adapting the intervention in different contexts. The results will be published in a peer-reviewed medical journal. 2.3.1. Informed consent. The eligible patients will be informed on the study and those interested in participating will be asked to sign informed consent during routine visits at Diabetes clinics. Patients aged < 18 years or unable to provide a personal informed consent are not eligible for the study.
The eligible patients will be informed on the study and those interested in participating will be asked to sign informed consent during routine visits at Diabetes clinics. Patients aged < 18 years or unable to provide a personal informed consent are not eligible for the study.
In designing this protocol, we tried to address both the issues of defining an acceptable and scalable intervention and to construct a study that could measure its efficacy. We acknowledged that a co-created and context-adapted complex intervention could not be standardized in its form across centres and also standardizing it by all the functions of the intervention components would be unrealistic. Furthermore, we observed an extreme heterogeneity in usual care across centres. Therefore, one of the main problems was how to balance the two needs: being context-adapted and measuring the efficacy of a reproducible intervention. The result was a careful conceptualization of what was the object of the evaluation to build an experimental design that could compare the novelty of the intervention in contrast to the existing usual care ( ). Indeed, during the evaluation phase, we would take into account the heterogeneity at function and incremental effort levels, which is expected to be lower than the heterogeneity that will be observed at form level. 3.1. Co-creating the intervention The main limits that we identified in previously proposed interventions were the lack of tailoring, scarce cultural-sensitivity, too-intensive interventions that would be not acceptable to the patients and not sustainable by the providers or interventions that were not comprehensive enough to address both the environment and the individual . We tried to overcome some of these limits by adopting a co-creative process to design and evaluate the intervention. Several previous studies on interventions for the management of chronic conditions included co-creation approaches with dissimilarities regarding stakeholders, phases of the research process involved, and methods. Regarding the selection of stakeholders to be involved, previous studies included in the co-creation phase the same patients recruited for testing the interventions We decided to involve different populations to assess the external validity of the results of the co-creation phase, with a transcultural approach suggested by the anthropologists of our qualitative research team. Regarding the phases of the research process, previous studies usually applied a co-creation approach in one phase only, mainly for the definition of the intervention or its evaluation, in terms of acceptability or feasibility [ , , , ]. We decided to plan both co-development and co-evaluation phases with the aim of developing an effective intervention and to provide a comprehensive assessment including the stakeholders’ perspectives and increasing the researchers accountability. Regarding the methods, the main differences between previous studies were related to the background of professionals involved in the co-creation phase and in the modes of stakeholders’ engagement (i.e., face-to-face, on-line, or blended) . In our study, two anthropologists and one sociologist with background in qualitative research coordinated the co-creation phase, and provided support to keep a transcultural approach across all the research process phases. Our preferred mode of involvement of stakeholders was face-to-face, given potential limitations on accessibility to on-line tools of the target population. Indeed, a blended mode may be a promising alternative approach. In our study, the co-creative approach became the main characteristic of the intervention and the object of the evaluation. During the process, we experienced many of the difficulties that caused the limits of previous interventions: only minimal changes in the organization of the service providing were considered sustainable; intensive counselling was too time-consuming and lifestyle changes were considered unacceptable by patients; including undocumented immigrants was considered difficult, if not unfeasible, for both administrative and logistic issues; in particular, although the Italian national health service is universalistic and guarantee access to everybody, including undocumented immigrants, it is difficult to imagine to take care these subgroups of population, often only in transit in a specific city or in any case, not confident to receive health care in a same healthcare facility due to their illegal status All these constrains reduced the choice of possible components to be included in the intervention in most centres. The co-creation process made it easy to identify these barriers and probably magnified the problems. Nevertheless, the process allowed us to find shared solutions. 3.2. Creating the contrast between the intervention and usual care The analysis of the usual care in each centre showed an extreme heterogeneity. This heterogeneity represents the differences existing in diabetes care across different public clinics in Italy, despite the health system being substantially similar in its funding and organization. Indeed, we noted that the differences across the country in providing health promotion and in proactive care of chronic disease are larger than those observed in acute care. Several factors may contribute to this heterogeneity. In health promotion guidelines often suggest (conditional recommendations) a set of possible interventions, not mutually exclusive; furthermore, in proactive care of chronic conditions the way of providing services is often as important as the service itself. Having different usual care protocols in the control arm is a problem when conducting a trial. We tried to conceptualize the different options in terms of how to standardize the intervention and the usual care across centres and how this would affect what we are measuring in terms of causal inference, the internal and external validity of the study, and on the feasibility of the study. The last point is not only an accidental constraint that we have to overcome to make a good trial, it may also be a proxy of the scalability and feasibility of the proposed intervention in each context. summarises this point. In fact, in the presence of such variability of the usual care, the feasibility of a given multicomponent intervention would be different in each centre: some components of the intervention are yet current practice in one centre, while in another can be considered almost unfeasible. Furthermore, what would be a feasible intervention in all centre, would be less than what already offered as usual care in the centres with a more structured usual care. The solution we found was to standardize the intervention on the incremental effort, i.e., in each centre, the intervention should require similar additional resources. Unfortunately, each component did not require the same additional amount of resources in each centre. In fact, in most cases, in centres where the usual care includes only basic services, implementing some components required much larger efforts than in centres already able to provide more complex services. 3.3. Lesson learnt/Conclusions Designing a multicentre trial to evaluate a complex intervention defined through a co-creative approach introduces further complexity and requires a new conceptual framework. We tried to build such a framework to understand how to deal with existing differences and with the intrinsic tailoring of a co-created intervention. We started adopting a standardization of the components of the intervention by function instead of by form, as suggested by Hawe et al . Focussing on feasibility and scalability , we also introduced the concept of standardizing the intervention also by the incremental resources needed. The conduction of the trial and its results will show if our attempts were fruitful.
The main limits that we identified in previously proposed interventions were the lack of tailoring, scarce cultural-sensitivity, too-intensive interventions that would be not acceptable to the patients and not sustainable by the providers or interventions that were not comprehensive enough to address both the environment and the individual . We tried to overcome some of these limits by adopting a co-creative process to design and evaluate the intervention. Several previous studies on interventions for the management of chronic conditions included co-creation approaches with dissimilarities regarding stakeholders, phases of the research process involved, and methods. Regarding the selection of stakeholders to be involved, previous studies included in the co-creation phase the same patients recruited for testing the interventions We decided to involve different populations to assess the external validity of the results of the co-creation phase, with a transcultural approach suggested by the anthropologists of our qualitative research team. Regarding the phases of the research process, previous studies usually applied a co-creation approach in one phase only, mainly for the definition of the intervention or its evaluation, in terms of acceptability or feasibility [ , , , ]. We decided to plan both co-development and co-evaluation phases with the aim of developing an effective intervention and to provide a comprehensive assessment including the stakeholders’ perspectives and increasing the researchers accountability. Regarding the methods, the main differences between previous studies were related to the background of professionals involved in the co-creation phase and in the modes of stakeholders’ engagement (i.e., face-to-face, on-line, or blended) . In our study, two anthropologists and one sociologist with background in qualitative research coordinated the co-creation phase, and provided support to keep a transcultural approach across all the research process phases. Our preferred mode of involvement of stakeholders was face-to-face, given potential limitations on accessibility to on-line tools of the target population. Indeed, a blended mode may be a promising alternative approach. In our study, the co-creative approach became the main characteristic of the intervention and the object of the evaluation. During the process, we experienced many of the difficulties that caused the limits of previous interventions: only minimal changes in the organization of the service providing were considered sustainable; intensive counselling was too time-consuming and lifestyle changes were considered unacceptable by patients; including undocumented immigrants was considered difficult, if not unfeasible, for both administrative and logistic issues; in particular, although the Italian national health service is universalistic and guarantee access to everybody, including undocumented immigrants, it is difficult to imagine to take care these subgroups of population, often only in transit in a specific city or in any case, not confident to receive health care in a same healthcare facility due to their illegal status All these constrains reduced the choice of possible components to be included in the intervention in most centres. The co-creation process made it easy to identify these barriers and probably magnified the problems. Nevertheless, the process allowed us to find shared solutions.
The analysis of the usual care in each centre showed an extreme heterogeneity. This heterogeneity represents the differences existing in diabetes care across different public clinics in Italy, despite the health system being substantially similar in its funding and organization. Indeed, we noted that the differences across the country in providing health promotion and in proactive care of chronic disease are larger than those observed in acute care. Several factors may contribute to this heterogeneity. In health promotion guidelines often suggest (conditional recommendations) a set of possible interventions, not mutually exclusive; furthermore, in proactive care of chronic conditions the way of providing services is often as important as the service itself. Having different usual care protocols in the control arm is a problem when conducting a trial. We tried to conceptualize the different options in terms of how to standardize the intervention and the usual care across centres and how this would affect what we are measuring in terms of causal inference, the internal and external validity of the study, and on the feasibility of the study. The last point is not only an accidental constraint that we have to overcome to make a good trial, it may also be a proxy of the scalability and feasibility of the proposed intervention in each context. summarises this point. In fact, in the presence of such variability of the usual care, the feasibility of a given multicomponent intervention would be different in each centre: some components of the intervention are yet current practice in one centre, while in another can be considered almost unfeasible. Furthermore, what would be a feasible intervention in all centre, would be less than what already offered as usual care in the centres with a more structured usual care. The solution we found was to standardize the intervention on the incremental effort, i.e., in each centre, the intervention should require similar additional resources. Unfortunately, each component did not require the same additional amount of resources in each centre. In fact, in most cases, in centres where the usual care includes only basic services, implementing some components required much larger efforts than in centres already able to provide more complex services.
Designing a multicentre trial to evaluate a complex intervention defined through a co-creative approach introduces further complexity and requires a new conceptual framework. We tried to build such a framework to understand how to deal with existing differences and with the intrinsic tailoring of a co-created intervention. We started adopting a standardization of the components of the intervention by function instead of by form, as suggested by Hawe et al . Focussing on feasibility and scalability , we also introduced the concept of standardizing the intervention also by the incremental resources needed. The conduction of the trial and its results will show if our attempts were fruitful.
S1 Table Elements of the intervention by topic and flexibility (degree of freedom of co-creation approach) in the definition. (PDF) S1 Fig Results of barriers’ analysis by center. (TIF) S1 File Study protocol approved by the ethics committee v 1.1. (PDF) S1 Checklist SPIRIT checklist. (DOCX)
|
A modified protocol merging two published techniques for computer guided zygomatic implants surgery: a technical note | bb6a22b7-0b81-42fb-b521-b2f10bf0a2de | 11515538 | Dentistry[mh] | Since Branemark first introduced zygomatic implants in 1988, many modifications have been reported regarding their designs, surgical approaches and loading protocols. With high survival and success rates documented in the literature, zygomatic implants indications have expanded to include severely resorbed maxillary ridges with insufficient bone for conventional implant placement. The strength of the anchorage in the zygoma compensates for the poor quality of the atrophied maxillary bones. In many cases, grafting is not an option due to the residual morphology of the atrophic maxillae following bone resorption. Given the opportunity for immediate loading after engaging stable cortical bone and abating the need for grafting procedures, the decision to utilize zygomatic implants for edentulous maxillary rehabilitation has been influenced by the shortened treatment time . However, zygomatic implant surgery can be difficult due to the limited intraoperative visibility of the surgical field and the complex anatomy of the zygomatic bone. The curvature of the sinus lateral wall and the sinusoidal shape of the zygomatic posterior wall make it challenging to control the long surgical drilling path and the zygomatic fixture placement. This imposes a great challenge during free-hand osteotomy, especially for less experienced surgeons placing quad zygomatic implants leading to potential serious complications such as penetration of the orbital cavity and the infratemporal fossa . In 2000, the use of computer-assisted navigational system was initially introduced to zygomatic implant rehabilitations by Schramm et al. and Watzinger et al. to decrease the incidence of these complexities . Afterwards, Vrielinck et al. were the first to report an in vivo study on the precision of zygomatic implants placement in 2003 to validate the use of personalized static computer guided drilling templates . Compared with the conventional free-hand technique, such implementations have shown a greater degree of accuracy in transferring the planned implant position to the surgical site . Nevertheless, templates used for conventional implants were considered inadequate for this task. Having only a single guide crestal sleeve, this approach does not adequately control the stability and direction of the apical portion of the drill when the site is prepared for the placement of the longer zygomatic implants. Additionally, these long drills are subjected to increased mechanical stresses, which could eventually compromise the stability of the template . To increase the accuracy of computer-aided zygomatic implant surgeries, different concepts and techniques have been proposed. In 2016, Chow introduced a novel drill guide with double sleeves, designed to stabilize the long zygomatic drills . Later, multiple researchers utilized his concept after applying several modifications to minimize the deviations as much as possible . Still, none of these studies has provided the evidence that is strong enough to be considered as the gold standard with the optimum degree of accuracy . In this study, our aim was to use and evaluate a modified protocol, combining the previously described device and surgical guide design. We used the double-sleeve drill guide proposed by Chow with the help of a computer-guided surgical template designed with a lateral window as suggested by Rinaldi et al. .
This prospective trial enrolled patients seeking maxillary arch rehabilitation using dental implant treatment. Sample size calculation was conducted using G*Power 3.1.9.4 Software based on data obtained from a previous study by Vrielinck et al. that reported 5.14˚ mean angular deviation and 2.59 standard deviation of zygomatic implants placed by CT-based planning. The effect size was calculated to be 0.82. A sample size of 13 implants was required to ensure that the 95% confidence interval estimate of the mean angular deviation of zygomatic implants placed by virtual planning and surgical guides was within 3˚ of the true mean. The power of t-test was estimated to be 80%, using a two-tailed significance level of 5%. Before conducting this study, the protocol was reviewed and approved by the Research Ethical Committee of Faculty of Dentistry, Ain Shams University in meeting no. (105), on 15th July 2020 with an application no.: (FDASU-RecD072029). Patients were recruited at the Department of Oral and Maxillofacial Surgery, Faculty of Dentistry, Ain Shams University and at the Egyptian Zygoma Implant Institute (EZII) clinic. The systemic health condition of all the participants was recorded and they were selected according to the following criteria fulfillment: Inclusion criteria 18-year-old (or older) patients who can understand and sign an informed consent. Patients in need of maxillary arch rehabilitation using dental implants. Patients with severe alveolar bone atrophy in both Bedrossian zone II & III or all three zones . Patients with a history of unsuccessful bone grafting, failure of conventional implants, or refusal to undergo further bone augmentation procedures. Patients with zygomatic bone of a 15 mm minimum width . Patients with good compliance and oral hygiene habits. Exclusion criteria Patients with cardiovascular disease, pulmonary disease or medical systemic condition that might hinder the fitness for general anesthesia (ASA III, IV, V and VI). Patients with conditions contraindicating implant placement (e.g.: radiation to the head and neck, intra-venous bisphosphonates, uncontrolled Diabetes mellitus). Acute maxillary sinus infection or untreated maxillary sinus cyst. Heavy Smokers (more than 20 cigarettes per day). Restricted mouth opening (less than 3 cm interincisal distance). In order to accurately formulate an optimal virtual treatment plan, the pre-operative Digital Imaging and Communications in Medicine (DICOM) files obtained from the multi-slice computer tomography (MSCT) scan were imported into Blue Sky Plan software. The acquired scans had the following parameters: (i) axial images only, (ii) no gantry tilt, (iii) slice thickness 0.625 mm, (iv) slice distance 0.625 mm, (v) field of view extending from the Glabella superiorly to the mandibular arch inferiorly, and (vi) bony window. The software was used to manipulate CT images, virtually plan for implant placement, design surgical guides, and export 3D virtual volumes as Standard Tessellation Language (STL) files. The desired implant positions were identified through the software to achieve the best possible functional and prosthetic position possible. During the guide-designing process, a cut-off window in the surgical guide was created slightly wider than the zygomatic implant and along its path. In conjunction with the computer-guided surgical templates, a customized double-sleeve drill guide kit – called the ‘zygoma drill guide’ – was used in this study. The zygoma drill guide was designed using CAD software with a few modifications to the original design developed by Chow. It was milled and produced out of stainless steel. Disassembly for cleaning and autoclaving was considered. Our zygoma drill guide kit composed of a set of three guides, one for each of the different-diameter drills contained in the manufacturer drilling kit. Figure (a) . Each drill guide consisted of two heads in co-axial alignment, which guided the drill at both entry points. These two heads were metal rings each with an internal diameter 0.15 mm larger than its corresponding drill. The two heads were connected by a sliding arm consisting of a rectangular rod enclosed in an open-ended tube. The maxillary entry head was inserted into the metal guiding sleeve fitted in the surgical template. The exit head was positioned at the base of the surgical template fitted in its lateral cut-off implant window resting on the sleeve-like or half-sleeve extension. When inserted into the surgical template, the drill went through the first entry head at the drill sleeve and engaged the second zygomatic head before penetrating the zygomatic bone. The surgical guides along with the 3D model of the mid-face were exported in STL format. The exported files were then 3D-printed in clear photopolymer resin using the Any-cubic Photon Mono X (MSLA) 3D printer. Rehearsal mock surgeries for all the cases were performed on the 3D model prior to the surgical procedures. During these rehearsals, the same protocol and steps to be performed in the actual surgery were followed strictly. Figure (b) . Prior to surgery, the surgical templates were submerged in a basin containing 2.4% activated glutaraldehyde solution for 20–30 min for high-level disinfection, followed by sterile saline rinse. The surgeries were performed under general anaesthesia with nasotracheal intubation. After complete exposure of all anatomical landmarks, the surgical templates were positioned in place. Once the templates were properly seated, they were secured in place using four to five 2.0 screws ranging from 10 to 14 mm in length. The zygomatic implants used were JDZygoma implants (JDentalCare, Madona, Italy) which were available with main implant diameter of Ø 4.3 mm, maximum implant diameter of Ø 4.5, an implantable length of 18 mm, a tip diameter of 2.15 mm and lengths ranging from 30 mm to 60 mm in 2.5 mm increments. Osteotomy drilling was performed in a two-step approach. First, the manufacturer’s key-less guided surgical drilling kit was used to complete the alveolar part of the osteotomy. Sequential drills entered the maxilla palatally puncturing the alveolar crest to reach the buccal side. Completion of the alveolar osteotomy was performed using the drill equivalent in diameter to the final drill of the zygomatic implant drilling kit, which is manufactured by the same company. During the second part of the osteotomy, the three zygoma drill guides were used after fitting their first entry heads in the metal sleeves of the surgical template. The long drill tips were guided properly to its second entry point in the zygomatic bone. The drill was visualized through the window created in the surgical templates until it reached the zygomatic bone to prepare the implant osteotomy. Figure . Upon completing the osteotomies and removing the surgical guide, a depth gauge was used to detect the depth and direction of the osteotomy confirming the zygomatic implant length. Afterwards, each zygomatic implant was manually placed with implant insertion torque not exceeding 80 N/cm until its tip reached the inferior border of the zygomatic osteotomy exit point. Figure . Finally, buccal fat pads were used to cover the extra bony shafts of the implant, enhancing soft tissue quality around, and the flaps were repositioned and sutured with 4 − 0 polyglycolic acid sutures. Within 4–5 days, the patient was called back for a follow-up visit to assess the healing process and to exclude any chances of wound infection. Ten days later, the sutures were removed upon complete wound healing. For the radiographic assessment of postoperative implants positions, a 2-week postoperative facial bones MSCT scan was requested from all patients. The postoperative CT scans were required to fulfil the same criteria as the preoperative CT scans and to be performed at the same radiology center. After superimposition, fixed landmark points were determined on both the virtually planned implants and the actual postoperative implants in the exact same positions. These points, which were determined as entry points at the implant platforms and exit points at the implant apices, were marked at both ends of each implant. The three planes of space, namely, the Mid-Sagittal Plane (MSP), the Frankfort Horizontal Plane (FHP) and the Coronal Plane (CP), were determined in relation to the skull. Measurements of the distances from the two points of each implant to the three planes of space were calculated for both the preoperative virtually planned implants and the actual postoperative implants. In addition, the direct distance between each entry and exit point on the virtually planned implants and its equivalent point in the actual postoperative implants was measured directly. Furthermore, angular deviation was recorded by measuring the angle between the two lines representing the long axes of the virtually planned implant and the actual implant. By comparing all the measurements, we determined the degree of accuracy of the applied virtual surgical planning intraoperatively three-dimensionally and detected how accurate the computer-guided surgical templates were in transferring the virtual plan to the operating room, which influenced the final implants positions. Categorical data were presented as frequency and percentage values. Numerical data were presented via the mean with 95% confidence intervals, standard deviations, and minimum and maximum values. The data were explored for normality by checking the data distribution using Shapiro-Wilk test. Data showed parametric distribution and were analyzed using paired t-test. Inter- and intra-observer reliability were analyzed using intra-class correlation coefficient (ICC). The significance level was set at p ≤ 0.05 within all tests. Statistical analysis was performed with R statistical analysis software version 4.3.0 for Windows. The datasets used and/or analyzed during the current study are available through the corresponding author upon reasonable request.
18-year-old (or older) patients who can understand and sign an informed consent. Patients in need of maxillary arch rehabilitation using dental implants. Patients with severe alveolar bone atrophy in both Bedrossian zone II & III or all three zones . Patients with a history of unsuccessful bone grafting, failure of conventional implants, or refusal to undergo further bone augmentation procedures. Patients with zygomatic bone of a 15 mm minimum width . Patients with good compliance and oral hygiene habits.
Patients with cardiovascular disease, pulmonary disease or medical systemic condition that might hinder the fitness for general anesthesia (ASA III, IV, V and VI). Patients with conditions contraindicating implant placement (e.g.: radiation to the head and neck, intra-venous bisphosphonates, uncontrolled Diabetes mellitus). Acute maxillary sinus infection or untreated maxillary sinus cyst. Heavy Smokers (more than 20 cigarettes per day). Restricted mouth opening (less than 3 cm interincisal distance). In order to accurately formulate an optimal virtual treatment plan, the pre-operative Digital Imaging and Communications in Medicine (DICOM) files obtained from the multi-slice computer tomography (MSCT) scan were imported into Blue Sky Plan software. The acquired scans had the following parameters: (i) axial images only, (ii) no gantry tilt, (iii) slice thickness 0.625 mm, (iv) slice distance 0.625 mm, (v) field of view extending from the Glabella superiorly to the mandibular arch inferiorly, and (vi) bony window. The software was used to manipulate CT images, virtually plan for implant placement, design surgical guides, and export 3D virtual volumes as Standard Tessellation Language (STL) files. The desired implant positions were identified through the software to achieve the best possible functional and prosthetic position possible. During the guide-designing process, a cut-off window in the surgical guide was created slightly wider than the zygomatic implant and along its path. In conjunction with the computer-guided surgical templates, a customized double-sleeve drill guide kit – called the ‘zygoma drill guide’ – was used in this study. The zygoma drill guide was designed using CAD software with a few modifications to the original design developed by Chow. It was milled and produced out of stainless steel. Disassembly for cleaning and autoclaving was considered. Our zygoma drill guide kit composed of a set of three guides, one for each of the different-diameter drills contained in the manufacturer drilling kit. Figure (a) . Each drill guide consisted of two heads in co-axial alignment, which guided the drill at both entry points. These two heads were metal rings each with an internal diameter 0.15 mm larger than its corresponding drill. The two heads were connected by a sliding arm consisting of a rectangular rod enclosed in an open-ended tube. The maxillary entry head was inserted into the metal guiding sleeve fitted in the surgical template. The exit head was positioned at the base of the surgical template fitted in its lateral cut-off implant window resting on the sleeve-like or half-sleeve extension. When inserted into the surgical template, the drill went through the first entry head at the drill sleeve and engaged the second zygomatic head before penetrating the zygomatic bone. The surgical guides along with the 3D model of the mid-face were exported in STL format. The exported files were then 3D-printed in clear photopolymer resin using the Any-cubic Photon Mono X (MSLA) 3D printer. Rehearsal mock surgeries for all the cases were performed on the 3D model prior to the surgical procedures. During these rehearsals, the same protocol and steps to be performed in the actual surgery were followed strictly. Figure (b) . Prior to surgery, the surgical templates were submerged in a basin containing 2.4% activated glutaraldehyde solution for 20–30 min for high-level disinfection, followed by sterile saline rinse. The surgeries were performed under general anaesthesia with nasotracheal intubation. After complete exposure of all anatomical landmarks, the surgical templates were positioned in place. Once the templates were properly seated, they were secured in place using four to five 2.0 screws ranging from 10 to 14 mm in length. The zygomatic implants used were JDZygoma implants (JDentalCare, Madona, Italy) which were available with main implant diameter of Ø 4.3 mm, maximum implant diameter of Ø 4.5, an implantable length of 18 mm, a tip diameter of 2.15 mm and lengths ranging from 30 mm to 60 mm in 2.5 mm increments. Osteotomy drilling was performed in a two-step approach. First, the manufacturer’s key-less guided surgical drilling kit was used to complete the alveolar part of the osteotomy. Sequential drills entered the maxilla palatally puncturing the alveolar crest to reach the buccal side. Completion of the alveolar osteotomy was performed using the drill equivalent in diameter to the final drill of the zygomatic implant drilling kit, which is manufactured by the same company. During the second part of the osteotomy, the three zygoma drill guides were used after fitting their first entry heads in the metal sleeves of the surgical template. The long drill tips were guided properly to its second entry point in the zygomatic bone. The drill was visualized through the window created in the surgical templates until it reached the zygomatic bone to prepare the implant osteotomy. Figure . Upon completing the osteotomies and removing the surgical guide, a depth gauge was used to detect the depth and direction of the osteotomy confirming the zygomatic implant length. Afterwards, each zygomatic implant was manually placed with implant insertion torque not exceeding 80 N/cm until its tip reached the inferior border of the zygomatic osteotomy exit point. Figure . Finally, buccal fat pads were used to cover the extra bony shafts of the implant, enhancing soft tissue quality around, and the flaps were repositioned and sutured with 4 − 0 polyglycolic acid sutures. Within 4–5 days, the patient was called back for a follow-up visit to assess the healing process and to exclude any chances of wound infection. Ten days later, the sutures were removed upon complete wound healing. For the radiographic assessment of postoperative implants positions, a 2-week postoperative facial bones MSCT scan was requested from all patients. The postoperative CT scans were required to fulfil the same criteria as the preoperative CT scans and to be performed at the same radiology center. After superimposition, fixed landmark points were determined on both the virtually planned implants and the actual postoperative implants in the exact same positions. These points, which were determined as entry points at the implant platforms and exit points at the implant apices, were marked at both ends of each implant. The three planes of space, namely, the Mid-Sagittal Plane (MSP), the Frankfort Horizontal Plane (FHP) and the Coronal Plane (CP), were determined in relation to the skull. Measurements of the distances from the two points of each implant to the three planes of space were calculated for both the preoperative virtually planned implants and the actual postoperative implants. In addition, the direct distance between each entry and exit point on the virtually planned implants and its equivalent point in the actual postoperative implants was measured directly. Furthermore, angular deviation was recorded by measuring the angle between the two lines representing the long axes of the virtually planned implant and the actual implant. By comparing all the measurements, we determined the degree of accuracy of the applied virtual surgical planning intraoperatively three-dimensionally and detected how accurate the computer-guided surgical templates were in transferring the virtual plan to the operating room, which influenced the final implants positions. Categorical data were presented as frequency and percentage values. Numerical data were presented via the mean with 95% confidence intervals, standard deviations, and minimum and maximum values. The data were explored for normality by checking the data distribution using Shapiro-Wilk test. Data showed parametric distribution and were analyzed using paired t-test. Inter- and intra-observer reliability were analyzed using intra-class correlation coefficient (ICC). The significance level was set at p ≤ 0.05 within all tests. Statistical analysis was performed with R statistical analysis software version 4.3.0 for Windows. The datasets used and/or analyzed during the current study are available through the corresponding author upon reasonable request.
A total of 13 implants were placed for four participants seeking maxillary arch rehabilitation using dental implant treatment. One female patient and three male patients were included in the study. The male patients were aged 37, 58 and 74 years, and the female patient was 56 years old. The age range was 37 to 74 years with a mean age of 56 years old. The implant distribution was four implants for each patient with the exception of one patient who received only one zygomatic implant as part of the research. For this patient, two additional conventional implants were placed ipsilaterally in conjunction with this single zygoma implant, having no conflicts with the research. The patients’ characteristics are shown in Table . No significant problems or complications impeded the completion of the surgeries. In one case, the anterior implant sleeve of the right-side surgical guide fractured during drilling. However, this situation did not affect the positioning of the implant during placement as the fracture took place during the final drill osteotomy. Among all the implants, only two implants were placed in different lengths than what was planned in the virtual treatment. One Implant was 37.5 mm instead of 40 mm (Case 1, Implant 4) while the other was 47.5 mm instead of 55 mm (Case 2, Implant 2). Measurements of all deviations are shown in detail in Fig. . All the patients had an uncomplicated immediate postoperative period with no major problems. As expected, tension, discomfort, and facial edema of variable degrees were experienced by the patients postoperatively. Radiographic evaluation at the level of entry points was performed post operatively for all patients and compared to the preoperative plan. The deviation was non-statistically significant ( P = 0.49) in the MSP with a 0.43 mm measurement difference (SD = ± 1.79). In the FHP, the deviation was 0.39 mm (SD = ± 1.12) with no statistical significance ( P = 0.32). The difference measured in the CP was − 0.54 mm (SD = ± 2) which was not statistically significant ( P = 0.44).Table . Moreover, differences between virtual and post-operative measurements at the exit point level were also measured. In the MSP, a -0.75 mm (SD = ± 1.25) deviation was detected with no statistical significance ( P = 0.1). A non-statistically significant ( P = 0.87) difference of -0.06 (SD = ± 1.09) was measured in the FHP. For the CP, the measurements showed 0.63 mm (SD = ± 1.24) non-statistically significant ( P = 0.16) difference. Table . By measuring the linear distance between the actual and virtual implants at the level of both entry and exit points, the descriptive statistics for direct linear deviation were calculated. The 95% confidence of interval revealed that deviations at the platforms ranged from 1.42 mm to 3.47 mm and the calculated mean deviation was 2.44 mm ± 1.57 mm. However, the full range of deviations at the entry level was 1.31 mm at minimum and 6.21 mm at maximum. With respect to deviations at the exit level, the minimum measurement was 1.28 mm, and the maximum was 4.16 mm with a mean of 2.32 ± 1 mm. The 95% confidence interval ranged of 1.67 mm to 2.97 mm. Table . After measuring the angle formed between the two lines joining the entry and exit points of both virtual and actual implants, the descriptive statistics for angular deviation were calculated, and the mean angular deviation was 3.6 ± 1.9˚ with a minimal and maximal deviation of 1.28˚, 7.06˚ respectively. The lower margin of the 95% confidence of interval was 2.35˚ and its upper margin was 4.86˚. Table .
The precision of any surgically guided procedure depends largely on the ability to accurately position the surgical template on top of the bone. Thus, our goal was to secure the best possible stability when designing the surgical templates by covering most of the maxillary bony surfaces exposed during the surgery . Aiming to achieve a tripodal bony support, we added a third mounting point along with the zygomatic bone and palatine process of the maxilla by extending our templates anteriorly to engage the anterior nasal spine. We believe that this prevented the guide mobility around any of its three axes providing the greatest stability and placement accuracy. To maintain that stable position during the procedure, the guides were fixed into position using four or five osteosynthesis screws . The cut-off window created in the surgical guide was carefully positioned to expose the point where the drill or the implant should exit the maxilla. Additionally, its proper positioning indicated the zygomatic entry point with a small incomplete sleeve-like extension from the surgical guide. This allowed the operator to visualize the implant trajectory and to monitor and evaluate the whole drilling procedure. This protocol was originally reported by Rinaldi et al. in 2019 . For each case, a 3D midface model was printed prior to the surgical procedure. As recommended by Aparicio et al. , a pre-operative mock surgery was performed strictly following the same drilling protocol and steps of the actual surgery. This step was very critical and beneficial for evaluating the surgical guide’s adaptation and function and for predicting any expected limitations or complications during the actual surgery . In this research, we applied a two-step drilling approach by performing the crestal maxillary osteotomy first followed by the zygomatic bone osteotomy. By allowing long zygomatic drills to pass passively along the crestal part of the osteotomy, we endorsed this approach for two main benefits. First, this helped to minimize the stresses encountered along the long drilling path. Second, it decreased the impediment produced by the limited mouth opening or opposing dentition while using a computer-guided template. The results of our study revealed no statistically significant differences between the virtual plan and post-operative outcomes, suggesting the use of computer-guided surgical templates in zygomatic implant placement. Furthermore, the deviations detected between the planned and placed implants did not affect the ability of these implants to proceed with the prosthetic phase. According to this study, it was found that the mean direct linear deviation at the implant platform level was 2.44 ± 1.57 mm (range: 1.31 to 6.21) and at the implant apices 2.32 ± 1 mm (range: 1.01 to 4.16). The mean angular deviation was found to be 3.6˚ ± 1.92 (range: 1.28˚ to 7.06˚). These results fall within the acceptable ranges according to the study conducted by Van Steenberghe et al. , which involved placing six fixed-length (45-mm) implants in three formalin-fixed human cadavers. Their results showed linear deviations below 3 mm, and their angular deviations were below 3.5˚ except for only one implant with 6.93˚ . A study conducted by Vrielinck et al. , on the other hand, aimed to validate the use of virtual planning software and customized surgical guides for zygomatic and pterygoid implants placement. The mean entry point, exit point, and angular deviations for the zygomatic implants were 2.77 ± 1.61 mm, 4.46 ± 3.16 mm, and 5.14˚ ± 2.59 respectively. All their zygomatic implants followed an intra-sinus trajectory and were placed using surgical guides supported only by the alveolar crest and palatal bones. Notably, this was one of the earliest in vivo studies evaluating the accuracy of guided zygomatic implant placement in 2003 . Rinaldi et al. , introducing the novel surgical template design with a lateral cut-off window, reported that deviations from their computerized project to the actual implant positions ranged from 2.5 to 3.5 mm with a mean angular deviation of 4.55˚ . Our study showed comparable or slightly better results, which might emphasize the two major differences between the two studies. First, we relied on the fixation screws to stabilize our guide during drilling as opposed to simple manual seating. Moreover, we had a secondary control point over the long zygomatic drills, which decreased the amount of angular deviation. Unlike most of the previously mentioned studies, in our results the mean deviation measured at the level of entry points was greater than that measured at the exit points. This might have resulted from the smooth nature of the implant crest module design in addition to the fact that the crestal osteotomy prepared by the final drill (3.6 mm in diameter) was smaller than the platform diameter. In other words, the threaded apical part of the implant had the ability to tap its concentric path while entering the smaller osteotomy. On the other hand, the smoother platform of the implant deviated from its intended position from the more compact palatal bone toward the less resistant buccal direction. More recent research introducing real-time navigation or static guides fabricated via titanium laser sintering yielded more accurate results. In 2020, Tao et al. compared the accuracy of CBCT and MSCT in zygomatic implant dynamic navigation surgery. The comparison of deviations in CBCT and MSCT groups revealed a mean entry deviation of 1.69 ± 0.59 mm vs. 2.04 ± 0.78 mm, apical deviation of 2 ± 0.68 mm vs. 2.55 ± 0.85 mm, and an angular deviation of 2.32˚ ± 1.02 vs. 3.23˚ ± 1.21 . A human cadaveric study, conducted in 2021, assessed the accuracy of zygomatic/pterygoid implant placement using custom-made bone-supported laser-sintered titanium templates. Using the EZgoma Principle & Guide , Grecchi et al. reported a mean angular deviation of 1.69˚ ± 1.12, a linear deviation of 0.76 ± 0.41 mm at the platform plane, and a linear deviation of 1.35 ± 0.78 mm at the apical plane . In contrast with our research, these two studies’ protocols permitted a fully guided zygomatic implant surgery after the completion of the guided osteotomy, which allowed complete guidance of fixture placement. In our study, the final step of the procedure was carried out in a free-hand manner as implant placement could not be performed through the surgical drill guide due to mechanical limitations. However, we must compare these more accurate results to our acceptable results with special consideration given to the feasibility and availability of each proposed computer-guided surgical protocol. Unfortunately, laser sintering technology is not yet commonly available on all markets, and when available, templates produced using this technology are more expensive than templates fabricated using acrylic resin . While dynamic navigation could be available, it requires higher facility investment and advanced setup, which would add up further cost to the treatment . The overall cost was an important factor because one of our first goals was to address a less expensive and more affordable treatment alternative. Another factor affecting our guided surgery accuracy might be the asymmetric distribution of the screws or uneven tightening of the screws. This drawback, in combination with the plastic nature of the guide’s resin material, might have resulted in imbalance of the drilling template creating a possibility of additional deviation . One of the major limitations of this study and probably of all surgically guided studies is the inability to directly compare the guided surgery results in relevance to the free-hand surgery regarding the accuracy . Without directly comparing both methods, it would be difficult to confirm that guided surgery could be considered “more” accurate, predictable, and/or uncomplicated. In zygomatic implant surgery performed free-hand, the surgeon certainly could be inspired by the preliminary virtual planning and assessments, but his/her decisions related to the entry point, path, and exit point of the implants would be made during surgery . Direct choices made during the surgery may represent the best option or treatment modification established for critical situations such as the rehabilitation of severely atrophied maxillary arches.
Given the major limitation of not being able to place the implants completely guided in their osteotomies, this study can conclude that the use of the double sleeve drill guide allowed favorable control over the tip of the long surgical drill during the zygomatic implant osteotomy preventing injury to adjacent vital structures. Additionally, the use of computer-guided surgical templates allowed the favorable placement of the crestal osteotomy in an acceptable prosthetic position. Overall, the use of computer-guided surgical templates augmented by the double sleeve drill guide may have helped refining zygomatic implants positioning. However, further research and improvement to this protocol should be considered to facilitate a more predictable and accurate surgical outcome along with the prevention of potential complications. To achieve evidence-based results, more studies with larger populations and in the form of randomized clinical comparative trials among different types of computer-assisted zygomatic implant surgeries need to be conducted.
|
Therapeutic decisions under uncertainty for spinal muscular atrophy: The DECISIONS-SMA study protocol | c7791d8e-7325-443c-be4a-02a2081bbb44 | 8846509 | Pediatrics[mh] | Spinal muscular atrophy (SMA) is an autosomal recessive neuromuscular disease caused by homozygous deletion or mutation of the survival motor neuron 1 gene on chromosome 5q13 that leads to progressive muscle weakness and atrophy . SMA is categorized into clinical subtypes based on the age at onset and severity of symptoms , mainly affecting infants (types I and II) and children (type III) . The disease causes a wide range of clinical symptoms, including respiratory, nutritional, orthopedic, rehabilitative, emotional, and social disorders , which may seriously compromise patients’ health and cause a considerable impact on the health-related quality of life of both patients and their caregivers . The SMA therapeutic landscape has changed over the last few years with the appearance of different therapeutic approaches such as antisense oligonucleotides, small molecules, or gene therapy . The administration of these therapies made it possible for SMA patients’ respiratory and motor function to be stabilized or even improved, as well as increasing their life expectancy . However, physicians still have the challenge of tailoring each individual’s treatment according to therapeutic goals, disease progression, patients’ and caregivers’ preferences, and their personal experience to achieve an optimal risk/benefit balance . Making this complex decision involves educating healthcare professionals and parents on the disease’s course and complications . Despite the limited evidence-based understanding of how physicians make treatment decisions when managing SMA, therapeutic options seem to be usually assessed according to their clinical experience when exposed to the uncertainties of new agents . However, decision making may also be influenced by cognitive or behavioral biases , including personality traits and background effects such as overconfidence, uncertainty tolerance, anchoring effect, information availability, or confirmation biases . Behavioral economics is the science that studies the principles of how we make decisions, combining psychology and economics to comprehensively understand cognitive and behavioral biases . It can therefore contribute to clarifying how physicians make their decisions and translate this into policy interventions that ultimately improve patients’ healthcare . Considering the above, this study aims to provide insight into therapeutic decision-making for SMA using behavioral economics paradigms, identifying treatment preferences of pediatric neurologists routinely managing SMA and recognizing the role of behavioral factors such as therapeutic inertia, herding phenomenon, care-related regret, occupational burnout, and risk preferences.
Study design and participants This is a noninterventional, cross-sectional, web-based pilot study involving 50 pediatric neurologists with expertise in managing patients with SMA in their routine clinical practice in Spain. Pediatric neurologists will be invited to participate by the Spanish Society of Pediatric Neurology (SENEP). The selection criteria also include participants practicing in academic or nonacademic settings, general practice pediatric neurologists or those specialized in neuromuscular disorders, involved or not in clinical research, from across Spain . Study objectives The primary study objective is to assess pediatric neurologists’ treatment preference for SMA in terms of its initiation, switch, and discontinuation. Secondary study objectives include evaluating therapeutic inertia, herding phenomenon, care-related regret, occupational burnout, and risk preferences of pediatric neurologists routinely managing SMA. Outcome measures and definitions Treatment preferences The pediatric neurologists’ treatment preferences will be assessed according to their choices in eleven simulated case scenarios (S1 Supporting information). Case-scenarios were originally designed by our research team (GS, PDA, JM, MBP, and IM) derived from the most common situations experienced by SMA patients in clinical practice and reviewing clinical trials and patient/caregivers preferences literature . The study (simulated case scenarios, questionnaires and scales) will be conducted in Spanish. The primary outcome variable will be pediatric neurologists’ treatment preference according to: 1) the percentage of participants who select treatment initiation when recommended , 2) the percentage of participants who select treatment switch when there is evidence of disease progression (i.e., a decrease in baseline scale score greater than the scale’s minimal clinically important difference) with initial therapies , and 3) the percentage of participants who select treatment discontinuation when disease progression puts patients outside treatment recommendation . Therapeutic inertia Therapeutic inertia is defined as the absence of treatment initiation or intensification when treatment goals are unmet . The study outcome measure will be its prevalence according to the pediatric neurologist responses on eleven case scenarios designed ad hoc . Its presence will be identified according to a score defined as the number of case scenarios that fit therapeutic inertia over the total number of presented cases . This score may therefore range from 0 to 11. Participants with a score of ≥1 (i.e., therapeutic inertia in at least one case scenario) will be considered to calculate therapeutic inertia prevalence. Herding phenomenon Herding is a phenomenon by which individuals follow others’ behavior rather than deciding independently based on their own private information . It has been shown that herding may lead to suboptimal decisions . The prevalence of herding will be assessed using two case scenarios designed ad-hoc . Its presence will be identified when the participant’s responses denote herding in at least one case scenario. Care-related regret Regret is an emotion experienced when one believes that the current situation would have had a better outcome by choosing a different course of action . Care-related regret was associated with suboptimal choices by healthcare professionals . Specific questions will assess the presence of regret, and its intensity will be evaluated using the 10-item Regret Intensity Scale (RIS-10). The specific questions will determine the presence of regret related to any patient and SMA patient situation within the last 5 years. The RIS-10 is a validated tool to assess care-related regret caused by a past event, covering affective, physical, and cognitive aspects . For each item, participants will be asked to rate their agreement on "how they feel now" from 1 (strongly disagree) to 5 (strongly agree). The RIS-10 overall score may range from 1 to 5, with higher scores indicating higher regret intensity. Occupational burnout Burnout is a condition characterized by emotional exhaustion, depersonalization, and a low sense of personal accomplishment . Physicians’ burnout is a common phenomenon which may influence therapeutic decisions . The prevalence of occupational burnout among participating pediatric neurologists will be calculated according to their scores on a nonproprietary single-item burnout measure, which instructs respondents to rate their burnout level based on their own definition of burnout on a 5-point scale . The absence or presence of burnout will be dichotomized according to the following scores: ≤2 (no symptoms of burnout) versus ≥3 (1 or more symptoms) . Risk preferences and tolerance to uncertainty Physicians’ low tolerance to uncertainty has been associated with suboptimal decisions and therapeutic inertia . Tolerance to uncertainty will be assessed using the standardized physician’s reaction to an uncertainty test . A short version following a factor analysis comprises five questions showing reliable psychometric properties . Participants will rate their level of agreement with each question from 0 (strongly disagree) to 5 (strongly agree), and a total score will be calculated . Low tolerance to uncertainty will be defined as values below the median of the total score . Risk aversion, defined as the tendency to prefer safe payoffs over probabilistic payoffs when the expected value is kept constant , will also be assessed. A risk-averse participant would prefer a treatment that provides a slight improvement with certainty over a therapy that offers a larger or no improvement with equal chance (50/50). We will evaluate risk aversion by identifying the safe amount for which a participant is indifferent between the safe and the risky option . Participants will be asked about the minimal amount of money they would prefer instead of the equiprobable gamble of winning €400 or €0 (expected value of €200). The degree of risk aversion of each individual will correspond to the difference of the expected value of the risky option (€200) minus the participant’s response (proxy of certainty equivalent) . Data management The data source will be the pediatric neurologists participating in the online study. Their data will be recorded in a database specifically designed for this research project through an electronic case report form. Pediatric neurologists will electronically give their written informed consent, confirm their eligibility, and provide some information about their profile (e.g., age, gender, academic/research profile, years of experience). They will then be presented with several case scenarios and scales to capture their feedback and opinion on study outcomes. After recording all the data from the last pediatric neurologist participating in the study and resolving any potential inconsistency, the study database will be locked, and the statistical analyses will be performed. Statistical considerations According to the Spanish Society of Pediatric Neurology, there are over 90 pediatric neurologists and 35 neuromuscular hospital-based clinics in Spain. Our previous experience in decision-making studies performed in Spain supports a response rate higher than 50% . This exploratory pilot study’s sample size is estimated at 50 participants, given the limited number of pediatric neurologists managing SMA patients in Spain. The study outcomes will be analyzed descriptively, calculating frequency distributions of qualitative variables, measures of central tendency and dispersion of quantitative variables, and 95% confidence intervals. Regression models will also be built for primary and secondary outcome measures to adjust their results for participant characteristics. Only available data will be considered in the analyses. Unavailable data will be described as missing, without any imputation/allocation. The statistical analysis will be performed using Stata Statistical Software 13.0 (StataCorp., College Station, TX, USA) and considering a significant level of 0.05. Ethical considerations This study will be conducted according to the Guidelines for Good Pharmacoepidemiological Practice published by the International Society of Pharmacoepidemiology, the ethical principles laid down in the World Medical Association Declaration of Helsinki, and applicable national regulations. The study was approved by the ethics committee of Hospital Clínico San Carlos (Madrid, Spain), and all participants will give their written informed consent before collecting any study data. Study status and timeline The study status is ongoing. Participant recruitment and data collection are planned to begin in June 2021. The expected date for database completion is December 2021.
This is a noninterventional, cross-sectional, web-based pilot study involving 50 pediatric neurologists with expertise in managing patients with SMA in their routine clinical practice in Spain. Pediatric neurologists will be invited to participate by the Spanish Society of Pediatric Neurology (SENEP). The selection criteria also include participants practicing in academic or nonacademic settings, general practice pediatric neurologists or those specialized in neuromuscular disorders, involved or not in clinical research, from across Spain .
The primary study objective is to assess pediatric neurologists’ treatment preference for SMA in terms of its initiation, switch, and discontinuation. Secondary study objectives include evaluating therapeutic inertia, herding phenomenon, care-related regret, occupational burnout, and risk preferences of pediatric neurologists routinely managing SMA.
Treatment preferences The pediatric neurologists’ treatment preferences will be assessed according to their choices in eleven simulated case scenarios (S1 Supporting information). Case-scenarios were originally designed by our research team (GS, PDA, JM, MBP, and IM) derived from the most common situations experienced by SMA patients in clinical practice and reviewing clinical trials and patient/caregivers preferences literature . The study (simulated case scenarios, questionnaires and scales) will be conducted in Spanish. The primary outcome variable will be pediatric neurologists’ treatment preference according to: 1) the percentage of participants who select treatment initiation when recommended , 2) the percentage of participants who select treatment switch when there is evidence of disease progression (i.e., a decrease in baseline scale score greater than the scale’s minimal clinically important difference) with initial therapies , and 3) the percentage of participants who select treatment discontinuation when disease progression puts patients outside treatment recommendation . Therapeutic inertia Therapeutic inertia is defined as the absence of treatment initiation or intensification when treatment goals are unmet . The study outcome measure will be its prevalence according to the pediatric neurologist responses on eleven case scenarios designed ad hoc . Its presence will be identified according to a score defined as the number of case scenarios that fit therapeutic inertia over the total number of presented cases . This score may therefore range from 0 to 11. Participants with a score of ≥1 (i.e., therapeutic inertia in at least one case scenario) will be considered to calculate therapeutic inertia prevalence. Herding phenomenon Herding is a phenomenon by which individuals follow others’ behavior rather than deciding independently based on their own private information . It has been shown that herding may lead to suboptimal decisions . The prevalence of herding will be assessed using two case scenarios designed ad-hoc . Its presence will be identified when the participant’s responses denote herding in at least one case scenario. Care-related regret Regret is an emotion experienced when one believes that the current situation would have had a better outcome by choosing a different course of action . Care-related regret was associated with suboptimal choices by healthcare professionals . Specific questions will assess the presence of regret, and its intensity will be evaluated using the 10-item Regret Intensity Scale (RIS-10). The specific questions will determine the presence of regret related to any patient and SMA patient situation within the last 5 years. The RIS-10 is a validated tool to assess care-related regret caused by a past event, covering affective, physical, and cognitive aspects . For each item, participants will be asked to rate their agreement on "how they feel now" from 1 (strongly disagree) to 5 (strongly agree). The RIS-10 overall score may range from 1 to 5, with higher scores indicating higher regret intensity. Occupational burnout Burnout is a condition characterized by emotional exhaustion, depersonalization, and a low sense of personal accomplishment . Physicians’ burnout is a common phenomenon which may influence therapeutic decisions . The prevalence of occupational burnout among participating pediatric neurologists will be calculated according to their scores on a nonproprietary single-item burnout measure, which instructs respondents to rate their burnout level based on their own definition of burnout on a 5-point scale . The absence or presence of burnout will be dichotomized according to the following scores: ≤2 (no symptoms of burnout) versus ≥3 (1 or more symptoms) . Risk preferences and tolerance to uncertainty Physicians’ low tolerance to uncertainty has been associated with suboptimal decisions and therapeutic inertia . Tolerance to uncertainty will be assessed using the standardized physician’s reaction to an uncertainty test . A short version following a factor analysis comprises five questions showing reliable psychometric properties . Participants will rate their level of agreement with each question from 0 (strongly disagree) to 5 (strongly agree), and a total score will be calculated . Low tolerance to uncertainty will be defined as values below the median of the total score . Risk aversion, defined as the tendency to prefer safe payoffs over probabilistic payoffs when the expected value is kept constant , will also be assessed. A risk-averse participant would prefer a treatment that provides a slight improvement with certainty over a therapy that offers a larger or no improvement with equal chance (50/50). We will evaluate risk aversion by identifying the safe amount for which a participant is indifferent between the safe and the risky option . Participants will be asked about the minimal amount of money they would prefer instead of the equiprobable gamble of winning €400 or €0 (expected value of €200). The degree of risk aversion of each individual will correspond to the difference of the expected value of the risky option (€200) minus the participant’s response (proxy of certainty equivalent) .
The pediatric neurologists’ treatment preferences will be assessed according to their choices in eleven simulated case scenarios (S1 Supporting information). Case-scenarios were originally designed by our research team (GS, PDA, JM, MBP, and IM) derived from the most common situations experienced by SMA patients in clinical practice and reviewing clinical trials and patient/caregivers preferences literature . The study (simulated case scenarios, questionnaires and scales) will be conducted in Spanish. The primary outcome variable will be pediatric neurologists’ treatment preference according to: 1) the percentage of participants who select treatment initiation when recommended , 2) the percentage of participants who select treatment switch when there is evidence of disease progression (i.e., a decrease in baseline scale score greater than the scale’s minimal clinically important difference) with initial therapies , and 3) the percentage of participants who select treatment discontinuation when disease progression puts patients outside treatment recommendation .
Therapeutic inertia is defined as the absence of treatment initiation or intensification when treatment goals are unmet . The study outcome measure will be its prevalence according to the pediatric neurologist responses on eleven case scenarios designed ad hoc . Its presence will be identified according to a score defined as the number of case scenarios that fit therapeutic inertia over the total number of presented cases . This score may therefore range from 0 to 11. Participants with a score of ≥1 (i.e., therapeutic inertia in at least one case scenario) will be considered to calculate therapeutic inertia prevalence.
Herding is a phenomenon by which individuals follow others’ behavior rather than deciding independently based on their own private information . It has been shown that herding may lead to suboptimal decisions . The prevalence of herding will be assessed using two case scenarios designed ad-hoc . Its presence will be identified when the participant’s responses denote herding in at least one case scenario.
Regret is an emotion experienced when one believes that the current situation would have had a better outcome by choosing a different course of action . Care-related regret was associated with suboptimal choices by healthcare professionals . Specific questions will assess the presence of regret, and its intensity will be evaluated using the 10-item Regret Intensity Scale (RIS-10). The specific questions will determine the presence of regret related to any patient and SMA patient situation within the last 5 years. The RIS-10 is a validated tool to assess care-related regret caused by a past event, covering affective, physical, and cognitive aspects . For each item, participants will be asked to rate their agreement on "how they feel now" from 1 (strongly disagree) to 5 (strongly agree). The RIS-10 overall score may range from 1 to 5, with higher scores indicating higher regret intensity.
Burnout is a condition characterized by emotional exhaustion, depersonalization, and a low sense of personal accomplishment . Physicians’ burnout is a common phenomenon which may influence therapeutic decisions . The prevalence of occupational burnout among participating pediatric neurologists will be calculated according to their scores on a nonproprietary single-item burnout measure, which instructs respondents to rate their burnout level based on their own definition of burnout on a 5-point scale . The absence or presence of burnout will be dichotomized according to the following scores: ≤2 (no symptoms of burnout) versus ≥3 (1 or more symptoms) .
Physicians’ low tolerance to uncertainty has been associated with suboptimal decisions and therapeutic inertia . Tolerance to uncertainty will be assessed using the standardized physician’s reaction to an uncertainty test . A short version following a factor analysis comprises five questions showing reliable psychometric properties . Participants will rate their level of agreement with each question from 0 (strongly disagree) to 5 (strongly agree), and a total score will be calculated . Low tolerance to uncertainty will be defined as values below the median of the total score . Risk aversion, defined as the tendency to prefer safe payoffs over probabilistic payoffs when the expected value is kept constant , will also be assessed. A risk-averse participant would prefer a treatment that provides a slight improvement with certainty over a therapy that offers a larger or no improvement with equal chance (50/50). We will evaluate risk aversion by identifying the safe amount for which a participant is indifferent between the safe and the risky option . Participants will be asked about the minimal amount of money they would prefer instead of the equiprobable gamble of winning €400 or €0 (expected value of €200). The degree of risk aversion of each individual will correspond to the difference of the expected value of the risky option (€200) minus the participant’s response (proxy of certainty equivalent) .
The data source will be the pediatric neurologists participating in the online study. Their data will be recorded in a database specifically designed for this research project through an electronic case report form. Pediatric neurologists will electronically give their written informed consent, confirm their eligibility, and provide some information about their profile (e.g., age, gender, academic/research profile, years of experience). They will then be presented with several case scenarios and scales to capture their feedback and opinion on study outcomes. After recording all the data from the last pediatric neurologist participating in the study and resolving any potential inconsistency, the study database will be locked, and the statistical analyses will be performed.
According to the Spanish Society of Pediatric Neurology, there are over 90 pediatric neurologists and 35 neuromuscular hospital-based clinics in Spain. Our previous experience in decision-making studies performed in Spain supports a response rate higher than 50% . This exploratory pilot study’s sample size is estimated at 50 participants, given the limited number of pediatric neurologists managing SMA patients in Spain. The study outcomes will be analyzed descriptively, calculating frequency distributions of qualitative variables, measures of central tendency and dispersion of quantitative variables, and 95% confidence intervals. Regression models will also be built for primary and secondary outcome measures to adjust their results for participant characteristics. Only available data will be considered in the analyses. Unavailable data will be described as missing, without any imputation/allocation. The statistical analysis will be performed using Stata Statistical Software 13.0 (StataCorp., College Station, TX, USA) and considering a significant level of 0.05.
This study will be conducted according to the Guidelines for Good Pharmacoepidemiological Practice published by the International Society of Pharmacoepidemiology, the ethical principles laid down in the World Medical Association Declaration of Helsinki, and applicable national regulations. The study was approved by the ethics committee of Hospital Clínico San Carlos (Madrid, Spain), and all participants will give their written informed consent before collecting any study data.
The study status is ongoing. Participant recruitment and data collection are planned to begin in June 2021. The expected date for database completion is December 2021.
This noninterventional pilot study will contribute to better understand the therapeutic decision-making process of pediatric neurologists who routinely care for SMA patients in Spain. The uniqueness of this study is that it uses a behavioral paradigm approach to examine the role of herding phenomenon, care-related regret, occupational burnout, and risk preferences in therapeutic decisions related to SMA (Figs and ). Our study will assess treatment preferences and factors associated with therapeutic inertia among pediatric neurologists. The change in the SMA treatment landscape that has taken place in the past few years has increased the therapeutic possibilities and decisions made by SMA patients, their caregivers, and healthcare providers. The assessment of patients and caregivers’ treatment-related priorities, expectations, and risk weighting for decision making has since gained relevance . However, the information available on how current therapeutic choices are made from pediatric neurologists’ perspective is still lacking. A recently published survey aimed to improve understanding of SMA patients and caregivers’ treatment choices, considering that their health status and life experience may influence how they perceive changes concerning desired benefits or therapeutic risks . Similarly, clinical neurologists’ experiences managing SMA in their daily practice may affect their perception of disease-related changes and risk preferences. Indeed, experiencing the exhaustion derived from occupational burnout can translate into emotional distress and decreasing engagement which may affect physician decisions and patient outcomes . Physicians’ care-related regret was also reported to negatively impact their health, quality of life, and patient care, as well as leading physicians to talk more often to their colleagues in order to improve their clinical practices . Although group support may play an important role in enhancing clinical practices, it may also lead physicians to follow therapeutic recommendations that are not supported by best practice guidelines. This herding-like behavior has been reported as a frequent phenomenon among neurologists managing other conditions such as multiple sclerosis, with a higher occurrence under uncertainty and leading to suboptimal decisions . In this scenario, therapeutic inertia could partly explain the neurologist’s resistance to escalate patient therapies under uncertainty, such as controversial situations or unclear efficacy evidence . Neurologists’ risk profile may therefore affect how they face decision making in these uncertain situations, with more therapeutic inertia among those showing strong aversion to ambiguity and low tolerance of uncertainty . Taken together, our results will inform about educational interventions in medical education to overcome knowledge-to-action gaps in the new therapeutic landscape of SMA. Here we describe a noninterventional study that will assess pediatric neurologists’ preferences for SMA treatment and behavioral factors that may affect their decisions using hypothetical case scenarios and specific scales/questions. This study will therefore contribute to expanding our evidence-based understanding of therapeutic decision-making for SMA. The authors acknowledge study limitations that should be considered, such as its exploratory pilot nature. Although pediatric neurologists managing SMA in their daily practice will be invited from all around Spain, we cannot exclude the possibility of sample biases derived from their final decision to participate. We should also keep in mind that the study hypothetical case scenarios show the most common situations faced in routine clinical practice, but they do not cover the whole case mix of the disease. In addition, we cannot rule out the possibility of residual confounders, despite the comprehensive adjustments that will be performed in the analyses. Therefore, further research would be desirable to confirm the study findings and explore their generalizability to other countries with different backgrounds and healthcare systems. In conclusion, this study will provide valuable insights into the treatment preferences of pediatric neurologists managing SMA in their daily practice, which is especially important considering the growing relevance of clinical decision-making based on values in the current healthcare system, the increasing possibilities of therapeutic approaches for SMA, and the lack of studies focusing on this subject. Following a behavioral paradigm for this assessment, this study aims to cover additional knowledge gaps in areas such as therapeutic inertia, herding phenomenon, care-related regret, occupational burnout, and risk preferences, which may also affect pediatric neurologists’ decision-making. These data will provide meaningful evidence to understand decision making when managing SMA in routine clinical practice.
S1 Appendix Case scenarios as presented to participants. (DOCX) Click here for additional data file.
|
Long‐term mental health outcomes after corneal transplantation and potential predictors: A multicentre prospective cohort study | a1b1a9a0-6b50-45b2-8541-6ec4bfe8b564 | 11823296 | Ophthalmologic Surgical Procedures[mh] | Corneal diseases are a leading cause of visual impairment and blindness and affect approximately 4.9 million people worldwide. , , , Common corneal diseases include pseudophakic bullous keratopathy, keratoconus, trauma, keratitis, endothelial dysfunction and corneal stromal dystrophies. For advanced corneal diseases, corneal transplantation, which is the most frequently conducted type of transplantation, is the most effective intervention. , Over the past 20 years, corneal transplantation techniques have undergone considerable evolution, leading to a greater variety of keratoplasty techniques. Evaluations of postoperative outcomes have mainly focused on conventional clinical outcomes, such as visual acuity, astigmatism, endothelial count, graft clarity and complications. However, these clinical outcomes do not adequately capture the experience of the patient. Therefore, in recent years, more emphasis is being placed on patient‐reported outcomes after corneal transplantation , , , , , which capture how the patient is doing in terms of daily functioning and quality of life. A systematic review evaluating the effects of corneal transplantation on patient‐reported outcomes, such as health/vision‐related quality of life, visual functioning, self‐reported visual symptoms and patient satisfaction showed overall beneficial effects. Predictors of positive patient‐reported outcomes were lower preoperative visual acuity and visual functioning, more favourable postoperative clinical outcomes, younger age and male sex. Most studies have focused on vision‐related or health‐related quality of life. , , , , , , Only limited research has been undertaken on mental health outcomes after corneal transplantation. It is well known that individuals with low vision are at risk for developing mental health complaints. , Previous studies have mainly focused on (older) adults with multiple causes of visual impairment, suggesting that approximately one‐third of visually impaired older adults experience mild but clinically significant symptoms of anxiety and/or depression. Drzyzga et al. evaluated the effects of corneal transplantation on mental health. This study showed that 16% of the patients awaiting corneal transplantation met the criteria for depression, while 20% met the criteria for anxiety. These numbers reduced significantly 3 weeks and 4 months after corneal transplantation, to 11% and 13% for both conditions at the two time points, respectively. However, the follow‐up period was short, and the sample size was relatively small. Moreover, not all important predictors of mental health outcomes, such as coping styles, , sociodemographic and clinical factors were taken into account. To better understand the impact of corneal transplantation and other predictors on mental health outcomes, this study aims to evaluate the effect of corneal transplantation on mental health outcomes and assesses potential predictors of these outcomes.
After reviewing the study protocol, study information letter and other documents, the Medical Ethical Committee of Amsterdam University Medical Centres (UMC), location VUmc, the Netherlands, confirmed that the study protocol was exempted from ethical approval according to the Dutch Medical Research in Human Subjects Act (WMO). The study adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants. Study design and setting This multicentre prospective cohort study was conducted across 11 (academic) hospitals and eye clinics in the Netherlands: Amsterdam UMC (location VUMC and AMC), Maastricht UMC, UMC Utrecht, UMC Groningen, Leiden UMC, Radboud UMC, OMC (Ophthalmological Medical Centre) Zaandam, Deventer Hospital, Gelre Hospital Apeldoorn and Amphia Hospital Breda. Data were collected between September 2017 and January 2021. Participants and procedures Patients were eligible to participate if they met the following inclusion criteria: age 18 years or older, awaiting corneal transplantation using lamellar or penetrating keratoplasty techniques, able to speak and understand the Dutch language and not having a severe cognitive impairment, as assessed by their attending ophthalmologist. Potential participants received a study information letter from their ophthalmologist. If they wanted to participate, patients returned the informed consent form to the researchers. Participants were subsequently telephoned by the researcher to explain the study procedures in more detail and to ask for the scheduled corneal transplantation date. One month prior to the corneal transplantation, participants received the baseline questionnaire (T0) through a web‐based survey platform. The option to complete questionnaires via paper‐and‐pencil versions or a telephone interview was offered to those not able to complete online questionnaires. Follow‐up questionnaires were sent 3 (T1), 6 (T2), 12 (T3) and 24 (T4) months after corneal transplantation. Clinical information such as corneal diagnosis (Fuchs' dystrophy, herpes, keratoconus, inflammation, other), the presence of other ocular diseases, use of medication, visual acuity, transplantation technique, tonometry, endothelial cell density and history of corneal transplantation were retrieved from patient records at the (academic) hospital or eye clinic. Outcome measures The Centre for Epidemiological Studies Depression scale (CES‐D) was used to measure symptoms of depression. The CES‐D consists of 20 items with four response categories. Total scores range from 0 to 60 and higher scores are indicative of more depressive symptoms; a score of ≥16 indicates subthreshold depression. Anxiety symptoms were measured with the Hospital Anxiety and Depression Scale‐Anxiety subscale (HADS‐A). The HADS‐A consists of seven items with four response categories. Total scores range from 0 to 21 with higher scores representing more symptoms of anxiety; a score ≥8 indicates subthreshold anxiety. , Subthreshold depression and anxiety refer to a condition where individuals experience depressive or anxiety symptoms that are not severe enough to meet the full criteria for a clinical diagnosis (e.g., major depressive disorder or generalised anxiety disorder), but the symptoms cause noticeable stress or impairment. The Dutch ICF Activity Inventory Emotional Health subscale (DAI‐EH) and Fatigue subscale (DAI‐F) were used to measure dealing with emotional health problems and fatigue due to visual impairment, respectively. , The DAI‐EH consists of 14 items whereas the DAI‐F consists of nine items. All items have five response categories ranging from not difficult to impossible. A not applicable option is also available, and is treated as a missing value. Total scores on the DAI‐EH and DAI‐F were calculated on a scale from 0 to 100 if at least 75% of the items were answered, with higher scores representing more emotional health problems and fatigue, respectively. At baseline, scores of 15 participants on the DAI‐EH and two participants on the DAI‐F could not be calculated because of too many missing responses (i.e., <75% of the items were answered). Potential predictors The following sociodemographic and clinical characteristics, for which participants completed questions at baseline (T0), were included as potential categorical predictors: age (four categories of similar size were created because of non‐linear relationship with outcomes: <65, 65–69, 70–74, 75+ years), sex, nationality (Dutch vs. non‐Dutch), living situation (alone vs. with others), work situation (no work vs. paid/voluntary work) and having comorbidity (either pulmonary diseases, cardiovascular diseases, gastrointestinal diseases, neurological diseases, diabetes, arthritis, cancer, psychiatric disorders and/or other health problems). Years of education was dichotomised into high (>10 years) vs. moderate–low (≤10 years) in the analyses of the DAI‐EH and DAI‐F because of a non‐linear relationship; for the CES‐D and HADS‐A, years of education was included as continuous predictor. Vision‐related predictors were primary diagnosis (Fuchs' dystrophy vs. other), former transplantation in the same eye, having a visual impairment (i.e., visual acuity ≥0.50 logMAR), transplantation technique (lamellar keratoplasty vs. penetrating keratoplasty) and occurrence of eye complaints. Occurrence of eye complaints was measured by the 10‐item Eye Complaint Questionnaire (ECQ), scored from 0 to 6 (score range 0–60) and the 8‐item Dry Eye Questionnaire (DEQ), scored from 0 to 3 (score range 0–24). Scores on the ECQ and DEQ were included as continuous predictors. Higher scores represented more (dry) eye complaints. Coping styles were measured with the Utrecht Coping List (UCL), consisting of 47 items scored from 1 to 4 which can be divided into seven subscales according to the scoring manual (three items are not assigned to any of the subscales and are therefore not used). The UCL is a widely used instrument which has been translated into several languages and is used in a range of populations. It demonstrated good measurement properties and was found to be feasible across diverse populations, including in large‐scale studies and clinical settings. , , Four subscales represent active coping styles: active tackling (seven items, score range 7–28), seeking social support (six items, score range 6–24), expressing emotions (three items, score range 3–12) and reassuring thoughts (five items, score range 5–20). The other three subscales represent passive coping styles: palliative reacting (eight items, score range 8–32), avoiding (eight items, score range 8–32) and passive reacting (seven items, score range 7–28). For all seven coping styles, participants' scores were dichotomised into ‘high’ scores (i.e., representing high or very high scores) and ‘low’ scores (i.e., representing average, low or very low scores) based on normative values. Statistical analyses Data were analysed using SPSS version 28 ( ibm.com ). Descriptive statistics were used for baseline characteristics. Independent sample t ‐tests and chi‐square tests were used to evaluate potential differences between participants who completed the study and those who were lost to follow up. After checking relevant assumptions (e.g., normality, linearity and multicollinearity), linear mixed model analyses were performed to analyse and predict symptoms of depression, anxiety, emotional health problems and fatigue over time. First, univariable analyses were conducted to explore the relation between each predictor and the outcome; a p ‐value <0.30 was used for selection in the multivariable model. Subsequently, multivariable linear mixed model analyses with backward stepwise selection were performed. A p ‐value ≥0.05 was used to exclude predictors. Final models were validated using a shrinkage factor derived from the heuristic shrinkage estimate from van Houwelingen and le Cessie. Clinical significance of the findings was evaluated using Cohen's effect sizes, where 0.20–0.49 are considered small, 0.50–0.79 moderate and ≥0.80 large.
This multicentre prospective cohort study was conducted across 11 (academic) hospitals and eye clinics in the Netherlands: Amsterdam UMC (location VUMC and AMC), Maastricht UMC, UMC Utrecht, UMC Groningen, Leiden UMC, Radboud UMC, OMC (Ophthalmological Medical Centre) Zaandam, Deventer Hospital, Gelre Hospital Apeldoorn and Amphia Hospital Breda. Data were collected between September 2017 and January 2021.
Patients were eligible to participate if they met the following inclusion criteria: age 18 years or older, awaiting corneal transplantation using lamellar or penetrating keratoplasty techniques, able to speak and understand the Dutch language and not having a severe cognitive impairment, as assessed by their attending ophthalmologist. Potential participants received a study information letter from their ophthalmologist. If they wanted to participate, patients returned the informed consent form to the researchers. Participants were subsequently telephoned by the researcher to explain the study procedures in more detail and to ask for the scheduled corneal transplantation date. One month prior to the corneal transplantation, participants received the baseline questionnaire (T0) through a web‐based survey platform. The option to complete questionnaires via paper‐and‐pencil versions or a telephone interview was offered to those not able to complete online questionnaires. Follow‐up questionnaires were sent 3 (T1), 6 (T2), 12 (T3) and 24 (T4) months after corneal transplantation. Clinical information such as corneal diagnosis (Fuchs' dystrophy, herpes, keratoconus, inflammation, other), the presence of other ocular diseases, use of medication, visual acuity, transplantation technique, tonometry, endothelial cell density and history of corneal transplantation were retrieved from patient records at the (academic) hospital or eye clinic.
The Centre for Epidemiological Studies Depression scale (CES‐D) was used to measure symptoms of depression. The CES‐D consists of 20 items with four response categories. Total scores range from 0 to 60 and higher scores are indicative of more depressive symptoms; a score of ≥16 indicates subthreshold depression. Anxiety symptoms were measured with the Hospital Anxiety and Depression Scale‐Anxiety subscale (HADS‐A). The HADS‐A consists of seven items with four response categories. Total scores range from 0 to 21 with higher scores representing more symptoms of anxiety; a score ≥8 indicates subthreshold anxiety. , Subthreshold depression and anxiety refer to a condition where individuals experience depressive or anxiety symptoms that are not severe enough to meet the full criteria for a clinical diagnosis (e.g., major depressive disorder or generalised anxiety disorder), but the symptoms cause noticeable stress or impairment. The Dutch ICF Activity Inventory Emotional Health subscale (DAI‐EH) and Fatigue subscale (DAI‐F) were used to measure dealing with emotional health problems and fatigue due to visual impairment, respectively. , The DAI‐EH consists of 14 items whereas the DAI‐F consists of nine items. All items have five response categories ranging from not difficult to impossible. A not applicable option is also available, and is treated as a missing value. Total scores on the DAI‐EH and DAI‐F were calculated on a scale from 0 to 100 if at least 75% of the items were answered, with higher scores representing more emotional health problems and fatigue, respectively. At baseline, scores of 15 participants on the DAI‐EH and two participants on the DAI‐F could not be calculated because of too many missing responses (i.e., <75% of the items were answered).
The following sociodemographic and clinical characteristics, for which participants completed questions at baseline (T0), were included as potential categorical predictors: age (four categories of similar size were created because of non‐linear relationship with outcomes: <65, 65–69, 70–74, 75+ years), sex, nationality (Dutch vs. non‐Dutch), living situation (alone vs. with others), work situation (no work vs. paid/voluntary work) and having comorbidity (either pulmonary diseases, cardiovascular diseases, gastrointestinal diseases, neurological diseases, diabetes, arthritis, cancer, psychiatric disorders and/or other health problems). Years of education was dichotomised into high (>10 years) vs. moderate–low (≤10 years) in the analyses of the DAI‐EH and DAI‐F because of a non‐linear relationship; for the CES‐D and HADS‐A, years of education was included as continuous predictor. Vision‐related predictors were primary diagnosis (Fuchs' dystrophy vs. other), former transplantation in the same eye, having a visual impairment (i.e., visual acuity ≥0.50 logMAR), transplantation technique (lamellar keratoplasty vs. penetrating keratoplasty) and occurrence of eye complaints. Occurrence of eye complaints was measured by the 10‐item Eye Complaint Questionnaire (ECQ), scored from 0 to 6 (score range 0–60) and the 8‐item Dry Eye Questionnaire (DEQ), scored from 0 to 3 (score range 0–24). Scores on the ECQ and DEQ were included as continuous predictors. Higher scores represented more (dry) eye complaints. Coping styles were measured with the Utrecht Coping List (UCL), consisting of 47 items scored from 1 to 4 which can be divided into seven subscales according to the scoring manual (three items are not assigned to any of the subscales and are therefore not used). The UCL is a widely used instrument which has been translated into several languages and is used in a range of populations. It demonstrated good measurement properties and was found to be feasible across diverse populations, including in large‐scale studies and clinical settings. , , Four subscales represent active coping styles: active tackling (seven items, score range 7–28), seeking social support (six items, score range 6–24), expressing emotions (three items, score range 3–12) and reassuring thoughts (five items, score range 5–20). The other three subscales represent passive coping styles: palliative reacting (eight items, score range 8–32), avoiding (eight items, score range 8–32) and passive reacting (seven items, score range 7–28). For all seven coping styles, participants' scores were dichotomised into ‘high’ scores (i.e., representing high or very high scores) and ‘low’ scores (i.e., representing average, low or very low scores) based on normative values.
Data were analysed using SPSS version 28 ( ibm.com ). Descriptive statistics were used for baseline characteristics. Independent sample t ‐tests and chi‐square tests were used to evaluate potential differences between participants who completed the study and those who were lost to follow up. After checking relevant assumptions (e.g., normality, linearity and multicollinearity), linear mixed model analyses were performed to analyse and predict symptoms of depression, anxiety, emotional health problems and fatigue over time. First, univariable analyses were conducted to explore the relation between each predictor and the outcome; a p ‐value <0.30 was used for selection in the multivariable model. Subsequently, multivariable linear mixed model analyses with backward stepwise selection were performed. A p ‐value ≥0.05 was used to exclude predictors. Final models were validated using a shrinkage factor derived from the heuristic shrinkage estimate from van Houwelingen and le Cessie. Clinical significance of the findings was evaluated using Cohen's effect sizes, where 0.20–0.49 are considered small, 0.50–0.79 moderate and ≥0.80 large.
Demographic and clinical patient characteristics In total, 238 participants were included in the study. Their baseline (T0) sociodemographic and clinical characteristics are presented in Table . At T4, loss to follow‐up was 13.9% ( n = 33). Participants who completed the study had a significantly higher education in years (mean 11.6 vs. 10.1, p = 0.008) and lower baseline scores on the CES‐D (mean 7.9 vs. 12.5, p < 0.001), HADS‐A (mean 3.5 vs. 4.9, p = 0.02) and DAI‐EH (mean 10.1 vs. 15.7, p = 0.01). There was a significant association between completing the study and nationality, having work (paid/voluntary) and having an ‘avoiding’ coping style: 88% of Dutch participants completed the study versus 62% of non‐Dutch participants ( p = 0.008); 78% of participants scoring high on the ‘avoiding’ coping style completed the study versus 90% who scored low ( p = 0.02) and 92% of participants with work completed the study versus 82% of participants without work ( p = 0.045). Figure shows mean scores of the CES‐D, HADS‐A, DAI‐EH and DAI‐F at baseline and follow‐up. Scores on the CES‐D and HADS‐A decreased between baseline and 3 months, after which time they stabilised. Scores on the DAI‐EH and DAI‐F decreased over time, but the largest decrease was observed between baseline and 3 months. Compared to baseline, scores at 24‐months follow‐up were significantly better for all outcomes ( p < 0.05), but with small effect sizes (0.13 for CES‐D, 0.16 for HADS‐A, 0.29 for DAI‐EH and 0.43 for DAI‐F). At baseline, 42 participants (18%) were classified as having subthreshold depression, whereas 27 (12%) were classified as having subthreshold anxiety. The number of participants classifying as having subthreshold depression decreased between baseline and 3 months and stabilised thereafter, whereas the number of participants classifying as having subthreshold anxiety remained relatively stable over all measurements. Predictors of depression, anxiety, emotional health problems and fatigue Table shows the results of univariable and multivariable linear mixed model analyses for the CES‐D, HADS‐A, DAI‐EH and DAI‐F. The final model of the CES‐D showed that being 75 years or older, having comorbidity, having more (dry) eye complaints and having a ‘passive reacting’ and/or ‘expressing emotions’ coping style were significant predictors of more depressive symptoms. Male sex, education in years, having Fuchs' dystrophy and having an ‘active tackling’ coping style were significant predictors of lower levels of depressive symptoms. The HADS‐A final model showed that having comorbidity, having more (dry) eye complaints and having an ‘avoiding’ and/or ‘passive reacting’ coping style were significant predictors of more anxiety symptoms. Male sex, having Fuchs' dystrophy and having an ‘active tackling’ coping style were significant predictors of lower levels of anxiety. For the DAI‐EH, the final model showed that having comorbidity, more (dry) eye complaints and having a ‘passive reacting’ and/or ‘expressing emotions’ coping style were significant predictors of more emotional health problems. Male sex, having Fuchs' dystrophy and having an ‘active tackling’ coping style were significant predictors of less emotional health problems. The final model of the DAI‐F showed that living alone, having comorbidity, having more (dry) eye complaints and having a ‘passive reacting’ coping style were significant predictors of more fatigue. Male sex, having work and having Fuchs' dystrophy were significant predictors of less emotional health problems. Internal validation of prediction models The longitudinal models with only a random intercept at the participant level had higher −2 log likelihood (−2LL) values than the final prediction models. This indicates that the final models can explain more variance in the outcome measure than the models with only a random intercept. The CES‐D model with only a random intercept at the patient level had a −2LL of 6833, whereas the final model had a −2LL of 5757. The HADS‐A model with only a random intercept at the patient level had a −2LL of 5000, whereas the final model had a −2LL of 4426. The DAI‐EH model with only a random intercept at the patient level had a −2LL of 7645, whereas the final model had a −2LL of 6804. And the DAI‐F model with only a random intercept at the patient level had a −2LL of 8758, whereas the final model had a −2LL of 7367. Using the −2LL, the heuristic shrinkage estimate was calculated to be 0.99 for the CES‐D, 0.98 for the HADS‐A and 0.99 for both the DAI‐EH and DAI‐F, suggesting a potentially good calibration of the models in an external dataset. ,
In total, 238 participants were included in the study. Their baseline (T0) sociodemographic and clinical characteristics are presented in Table . At T4, loss to follow‐up was 13.9% ( n = 33). Participants who completed the study had a significantly higher education in years (mean 11.6 vs. 10.1, p = 0.008) and lower baseline scores on the CES‐D (mean 7.9 vs. 12.5, p < 0.001), HADS‐A (mean 3.5 vs. 4.9, p = 0.02) and DAI‐EH (mean 10.1 vs. 15.7, p = 0.01). There was a significant association between completing the study and nationality, having work (paid/voluntary) and having an ‘avoiding’ coping style: 88% of Dutch participants completed the study versus 62% of non‐Dutch participants ( p = 0.008); 78% of participants scoring high on the ‘avoiding’ coping style completed the study versus 90% who scored low ( p = 0.02) and 92% of participants with work completed the study versus 82% of participants without work ( p = 0.045). Figure shows mean scores of the CES‐D, HADS‐A, DAI‐EH and DAI‐F at baseline and follow‐up. Scores on the CES‐D and HADS‐A decreased between baseline and 3 months, after which time they stabilised. Scores on the DAI‐EH and DAI‐F decreased over time, but the largest decrease was observed between baseline and 3 months. Compared to baseline, scores at 24‐months follow‐up were significantly better for all outcomes ( p < 0.05), but with small effect sizes (0.13 for CES‐D, 0.16 for HADS‐A, 0.29 for DAI‐EH and 0.43 for DAI‐F). At baseline, 42 participants (18%) were classified as having subthreshold depression, whereas 27 (12%) were classified as having subthreshold anxiety. The number of participants classifying as having subthreshold depression decreased between baseline and 3 months and stabilised thereafter, whereas the number of participants classifying as having subthreshold anxiety remained relatively stable over all measurements.
Table shows the results of univariable and multivariable linear mixed model analyses for the CES‐D, HADS‐A, DAI‐EH and DAI‐F. The final model of the CES‐D showed that being 75 years or older, having comorbidity, having more (dry) eye complaints and having a ‘passive reacting’ and/or ‘expressing emotions’ coping style were significant predictors of more depressive symptoms. Male sex, education in years, having Fuchs' dystrophy and having an ‘active tackling’ coping style were significant predictors of lower levels of depressive symptoms. The HADS‐A final model showed that having comorbidity, having more (dry) eye complaints and having an ‘avoiding’ and/or ‘passive reacting’ coping style were significant predictors of more anxiety symptoms. Male sex, having Fuchs' dystrophy and having an ‘active tackling’ coping style were significant predictors of lower levels of anxiety. For the DAI‐EH, the final model showed that having comorbidity, more (dry) eye complaints and having a ‘passive reacting’ and/or ‘expressing emotions’ coping style were significant predictors of more emotional health problems. Male sex, having Fuchs' dystrophy and having an ‘active tackling’ coping style were significant predictors of less emotional health problems. The final model of the DAI‐F showed that living alone, having comorbidity, having more (dry) eye complaints and having a ‘passive reacting’ coping style were significant predictors of more fatigue. Male sex, having work and having Fuchs' dystrophy were significant predictors of less emotional health problems.
The longitudinal models with only a random intercept at the participant level had higher −2 log likelihood (−2LL) values than the final prediction models. This indicates that the final models can explain more variance in the outcome measure than the models with only a random intercept. The CES‐D model with only a random intercept at the patient level had a −2LL of 6833, whereas the final model had a −2LL of 5757. The HADS‐A model with only a random intercept at the patient level had a −2LL of 5000, whereas the final model had a −2LL of 4426. The DAI‐EH model with only a random intercept at the patient level had a −2LL of 7645, whereas the final model had a −2LL of 6804. And the DAI‐F model with only a random intercept at the patient level had a −2LL of 8758, whereas the final model had a −2LL of 7367. Using the −2LL, the heuristic shrinkage estimate was calculated to be 0.99 for the CES‐D, 0.98 for the HADS‐A and 0.99 for both the DAI‐EH and DAI‐F, suggesting a potentially good calibration of the models in an external dataset. ,
This study evaluated the effect of corneal transplantation on mental health outcomes and assessed potential predictors of these outcomes. This study found that depression and anxiety symptoms decreased immediately after corneal transplantation, as did the proportion of participants with subthreshold depression. Emotional health and fatigue outcomes also improved after corneal transplantation and continued to improve throughout the study. Important predictors of mental health outcomes were also identified. A prevalence of 17.9% was found for subthreshold depression in this study, which is notably higher than the prevalence of 12% previously reported in a Dutch population of older adults. It is also higher than the 8.5% of adults aged 18–75 years with major depressive disorder in the Netherlands in 2019–2022. The prevalence of subthreshold anxiety was 11.5% in the present study, which is similar to the prevalence among Dutch older adults (10.7%). However, it is lower than the prevalence of anxiety disorders among adults aged 18–75 years in the Netherlands (15.2%). Yet, the numbers found here are lower than those observed in populations with visual impairment, where approximately one in three report having subthreshold depression and/or anxiety. If only participants with low vision from this study are considered (i.e., visual acuity ≥0.50 logMAR), then the percentages for subthreshold depression and anxiety at baseline increase to 21.6% and 13.5%, respectively. Depression, anxiety, emotional health and fatigue scores improved significantly over time compared to baseline, yet effect sizes were small. The percentage of participants with subthreshold depression decreased from 17.9% at baseline to around 11% at each of the follow‐up measurements. The minimal clinically important difference (MCID) of the CES‐D is suggested to be around 11 points in a German sample with depression symptoms (as measured by the CES‐D‐15). In the present study, between baseline and 24‐months follow‐up, 13 participants (6%) improved at least 11 points, whereas between baseline and 3‐month follow‐up 14 participants (6%) improved at least 11 points. The percentage of participants with subthreshold anxiety decreased from 11.5% at baseline to around 10% at each of the follow‐up measurements. Although the MCID of the HADS‐A varies by population, , , an improvement of at least 3 points is often considered important. Between baseline and 24‐month follow‐up, 44 participants (21%) achieved this change, whereas between baseline and 3‐month follow‐up 41 participants (18%) achieved this change. No MCID values have been established for the DAI‐EH and DAI‐F. It is important to note that a large proportion of participants already had a (near) perfect score on these outcomes at baseline. Specifically, 50% of the participants scored ≤6 on the CES‐D, while 42% scored ≤2 on the HADS‐A (i.e., ≤10% of the maximum score). Additionally, 61% and 38% of the participants scored ≤10 on the DAI‐EH and DAI‐F, respectively. Hence, there was little room for improvement for these participants. We found that male sex was a significant predictor of better scores in all four models, which is consistent with previous studies. Having the diagnosis Fuchs' dystrophy was also a significant predictor of better outcomes in all four models, whereas experiencing more (dry) eye complaints were significant predictors of worse scores in all four models. More than half of the participants in this sample had Fuchs' dystrophy, a common corneal disease that is often treated with the less invasive lamellar keratoplasty, which has a shorter recovery time. Other more severe diagnoses are often treated with penetrating keratoplasty, which requires a longer recovery period and may therefore lead to poorer mental health outcomes. However, transplantation technique was not a significant predictor in any of the models, and therefore other explanations might also play a role. For example, knowing that an effective and less invasive treatment is available, the gradual and slow progression of the condition, the relatively older age at which Fuchs' dystrophy becomes problematic and/or the isolation of the condition compared to corneal diseases that may be associated with systematic conditions may have a positive impact on mental health outcomes. Exhibiting a passive reacting coping style was also a significant predictor of poorer outcomes in all models. This was the coping style least exhibited by participants in the current study. Yet, it largely influenced scores in each of the four models. People who exhibit this coping style allow themselves to become completely absorbed in their problems and the situation in which they find themselves. They often perceive their situation as gloomy and withdraw in worry, unable to do anything about their situation. This coping style is also recognised by worrying about the past. Exhibiting this coping style has been associated with poorer mental health. , Two other coping styles that have been associated with poorer psychological outcomes are the avoiding and expressing emotions coping styles. , The avoiding coping style was an additional significant predictor of poorer anxiety outcomes, and the expressing emotions coping style of poorer depression and emotional health outcomes. Although having a certain coping style has often been viewed as a stable trait of personality, studies have shown that therapy targeting specific coping strategies may be effective. This supports the view that the diverse coping strategies are not only a predictor of mental health outcomes but may also be a target for interventions aimed at improving mental health. In contrast, the active tackling coping style was a significant predictor of better anxiety and emotional health outcomes. This coping style is characterised by a calm analysis of the situation so that the problem can be solved with confidence in a targeted manner. Active tackling is one of the active coping strategies and has previously been found to be a predictor of reduced symptoms of depression. , , We also found that living alone was a significant predictor of poorer fatigue scores as measured by the DAI‐F. Other studies have also noted that people who live alone report higher levels of fatigue. , On the other hand, having paid and/or voluntary work was a significant predictor of better fatigue scores. Individuals with higher levels of fatigue may find it more difficult to obtain or maintain employment, whether paid or voluntary. Fatigue can have a significant impact on various aspects of a person's life, including their ability to work effectively and cope with the demands of a job. A study in people with visual impairment found that fatigue had an impact on their cognitive functioning, particularly their concentration, attention and memory. Previous studies have shown that younger age is a predictor of more favourable patient‐reported outcomes, but no effect of age was found here, except for symptoms of depression as measured by the CES‐D. Participants aged 75 years or older had worse symptoms of depression compared with participants <65 years of age. The mean age of the present study population was 67.6 years, with poor representation of younger participants. Moreover, age had a non‐linear relationship with all outcomes, so age was categorised rather than including it as a continuous predictor. This may have resulted in loss of information. , Additionally, years of education was a significant predictor of better depression outcomes. The mean years of education in the sample was quite high, equivalent to at least a college degree. This may be due to selection bias, as highly educated people may be more willing to participate in research, while also reporting a higher quality of life. In contrast to previous studies, preoperative visual acuity was not a significant predictor in any of the four models. Participants were categorised as having visual impairment (i.e., visual acuity ≥0.50 logMAR) or no visual impairment because there was no linear relationship with any of the outcomes. Only 17.4% of participants had visual impairment at baseline, suggesting the sample size may have been too small to detect a significant effect. This finding may also be influenced by the current trends in transplantation indications. Nowadays, the indication for transplantation is more often given for higher vision than in the past, as people have higher requirements for vision, such as the desire to continue working. Consequently, they are more likely to need a transplant even with relatively better visual acuity. Additionally, the high success rate of transplantation may contribute to reduced mental complaints, as patients are optimistic about positive outcomes. This study has several strengths. It is the first study to determine the long‐term effects of corneal transplantation on mental health outcomes. The sample size was large and high participation rates were maintained throughout the investigation. In addition, thanks to the collaboration of (academic) hospitals and eye clinics throughout the Netherlands, we were able to obtain a nationwide representation of participants. Several limitations must also be acknowledged. First, the results of this study may not be generalisable to all patients undergoing corneal transplantation. Younger participants were poorly represented, although this is representative of the population undergoing corneal transplantation. In addition, the sample had a relatively high level of education. Thus, there is a risk of potential selection bias. This might have caused age, work status and years of education to be non‐significant predictors in most models. Also, the results may be less generalisable to other populations. Second, visual acuity was obtained from patient records at the (academic) hospital or eye clinic. However, this information was often poorly documented and most complete at baseline. Therefore, whether someone could be classified as visually impaired was based on the baseline visual acuity, and changes in visual acuity could not be taken into account. Furthermore, misclassification may have occurred because of poor documentation. Future studies should investigate whether changes in visual acuity act as a potential predictor of mental health outcomes. Third, no information was available regarding the use of immunosuppressive drugs after corneal transplantation, or information on primary graft failure. Information on previous transplantations in the same eye was available, which was not a significant predictor for any of the outcomes. Fourth, although a large number of potential predictors were included, other factors, such as perceived health status, may also play a role. Finally, we opted to use raw ordinal scores instead of response pattern scoring through item response theory (IRT) or Rasch analysis. Although we recognise the advantages that IRT/Rasch‐based scoring might offer, particularly with respect to missing data (which was the case for the DAI subscales), several considerations led us to use raw ordinal scores instead. These considerations included: (1) the unavailability of a scoring system (e.g., Excel ( Microsoft.com ) templates), which would require the development of entirely new item parameters, (2) the relatively small sample size in terms of IRT analyses (at least 500 participants are needed to adequately fit an IRT model ), affecting the ability to yield accurate parameter estimates and (3) the interpretability of raw scores, which can be more easily understood by a broader audience. Moreover, the CES‐D and HADS‐A are usually scored through raw ordinal scores, and cut‐off values for raw scores have previously been established. Thus, the scoring of the DAI subscales was consistent with the scoring of the CES‐D and HADS‐A. In conclusion, this study shows that corneal transplantation has a positive impact on mental health outcomes, including depression symptoms, anxiety symptoms, emotional health problems and fatigue. Male sex and Fuchs' dystrophy were important predictors of better mental health outcomes, whereas having comorbidity, experiencing more (dry) eye complaints and having a passive reacting coping style were important predictors of worse mental health outcomes. The results of this study may improve the understanding of patients and eyecare providers regarding the effects of corneal transplantation. Practitioners could incorporate knowledge of predictors into preoperative counselling, which may lead to better decision making and more realistic patient expectations. Future studies are needed to confirm the results of this study, especially in underrepresented patient groups such as younger patients or those with lower education levels.
E. B. M. Elsman: Data curation (equal); formal analysis (equal); investigation (equal); methodology (equal); project administration (equal); visualization (equal); writing – original draft (equal). H. P. A. Van der Aa: Supervision (equal); writing – review and editing (equal). N. E. Billingy: Data curation (equal); project administration (equal); writing – review and editing (equal). C. Nieuwendaal: Data curation (equal); writing – review and editing (equal). R. P. L. Wisse: Data curation (equal); writing – review and editing (equal). R. J. Wijdh: Data curation (equal); writing – review and editing (equal). M. L. Tang: Data curation (equal); writing – review and editing (equal). B. T. H. Van Dooren: Data curation (equal); writing – review and editing (equal). S. Nobacht: Data curation (equal); writing – review and editing (equal). R. M. M. A. Nuijts: Conceptualization (equal); data curation (equal); funding acquisition (equal); writing – review and editing (equal). G. H. M. B. Van Rens: Conceptualization (equal); funding acquisition (equal); supervision (equal); writing – review and editing (equal). R. M. A. Van Nispen: Conceptualization (equal); funding acquisition (equal); resources (equal); software (equal); supervision (equal); validation (equal); writing – review and editing (equal).
The research is funded by Stichting Blindenhulp, Landelijke Stichting voor Blinden en Slechtzienden (LSBS), Hoornvlies Patiënten Vereniging, Katholieke Stichting voor Blinden en Slechtzienden (KSBS), Stichting tot Verbetering van het Lot der Blinden and Vereniging Bartiméus Sonneheerdt (VBS).
The authors have no conflicts of interest.
|
Effect of wet-laboratory training on resident performed manual small-incision cataract surgery | 3541d1f4-a88e-4253-94da-103dcf3d86f0 | 5989499 | Ophthalmology[mh] | This study was conducted after approval from the Institutional Review Board of our institute. Wet-laboratory training was introduced in our institute in March 2013 wherein the residents are trained in a stepwise manner in hand–eye coordination and the steps of cataract surgery. The residents were initially trained in hand–eye coordination that included suturing on foam under microscope (3-step, Leica microscope) for two classes. They were then trained with goat's eye for the scleral tunnel, entry into anterior chamber and capsulotomy for 6 weeks (12 Goat's eye each). After 2 months of wet-laboratory training, they were allowed to perform cataract surgeries under supervision in the operating theater followed by regular monthly training of 6 h/month during their first 2 years of residency. We conducted a cross-sectional comparative study, wherein records of 464 patients were reviewed. The information was abstracted from the surgical records in the Department of Ophthalmology of our institute and all the resident-performed surgeries were identified. Of these, those performed before wet-laboratory training was introduced in our institute, i.e., March 2013 and those after, were classified into Groups A and B, respectively (Time period - January 2012–July 2014). Both groups included 232 patients each. The demographic data, diagnosis, type of surgery performed, postoperative vision, and intraoperative and postoperative complications for the surgeries performed by both groups were recorded and compared. The Statistical Package for the Social Sciences software version 18.0 (SPSS Inc., Chicago, IL, USA) was used for data analysis. The Chi-squared test was used and a value of P < 0.05 was considered as statistically significant.
The age, sex, and type of cataract-when compared between the two groups were similar with no statistically significant difference between the two groups. It is shown in . When we compared the combined intraoperative complications between the two groups, it was seen that the Group A had a higher rate of complications with 23.7% of cases having complications as opposed to only 15.08% in Group B. When each individual complication was analyzed, peripheral iridectomy was done more frequently in Group A as compared to Group B. There was no difference between the two groups in proportion of tunnel complications, descemet's membrane detachment, and zonular dehiscence. The occurrence of posterior capsule (PC) rupture and vitreous loss showed a statistically significant difference, with Group A showing a high rate of PC rent and vitreous loss (14.3%) while only 6.9% had this complication in Group B. Nucleus drop was also seen in two cases of Group A whereas none in Group B had this complication. The comparison of intraoperative complications between the two groups is shown in . The rate of intraocular lens implantation and the type of lens implanted is shown in . The frequency of posterior chamber intraocular lens (PCIOL) was higher in the Group B. Iris claw lens were not used before 2013 in our institute, which is reflected in the higher proportion of implantation among Group B patients. The postoperative day 1 vision was compared between the two groups and is represented in . As shown in the table, Group B patients had a better vision on the postoperative day 1. The postoperative vision mirrored the higher occurrence of intraoperative and postoperative complications in Group A with 13.36% of patients in Group A having vision of less than counting fingers at 3 m as compared to Group B ( P = 0.00). When postoperative complications were compared between the two groups, it was shown that the Group B had a statistically significant lower frequency of complications. The occurrence of complications is shown in . Further we analyzed, the intraoperative and postoperative outcome based on the resident training level. The intraoperative complications according to the residency level are shown in . The table shows that the maximum difference between the two groups was seen in the 2 nd -year residents (JR2) who had a statistically significant lower incidence of complications, especially PC rupture and vitreous loss. This may be because the supervising expert more closely supervises the 1 st -year residents with higher takeover of the case. When the intraocular lens placed in each group according to the residency level was analyzed, a statistically significant difference was seen in the 2 nd -year residents, with a higher number in the Group B, Having placed a PCIOL .
There are various studies, which have compared cataract surgeries performed by residents with those done by experienced surgeons, and have shown that with experience the visual outcome becomes better and the complications reduce. Haripriya et al . revealed in a study, on a large population done in South India, that the cataract surgery outcome was significantly better when performed by the staff surgeons (0.76%), as compared to residents (2.06%), and trainees (5%). The combined complication rate for trainees in phacoemulsification was 4.8% as compared to 1.46% in MSICS. In this study, too we saw that 3 rd -year residents were performing better than the 1 st -year residents in both groups. With experience, the visual outcome of cataract surgery becomes better, but our aim should be to provide an effective stress-free training to the residents during their residency to reduce complications. Ophthalmic surgery is different from other surgical fields, as it requires additional skills of hand–eye coordination. Microsurgery allows only one person to operate at a time; hence, does not give ample time for the supervisor to intervene before a complication occurs. The residents operate under a highly demanding and stressful environment that may hamper their development as good surgeons. There are various studies published which aim at improving the training and surgical outcome of trainee operated cases. Rogers et al . showed that the implementation of a structured curriculum for the ophthalmic residents significantly reduced the rate of complications, especially PC rupture and vitreous loss. In their study, the 1 st - and 2 nd -year residents went through intensive wet-laboratory training and supervised surgical training, thus emphasizing the need for training of microskills at a skills laboratory before residents are allowed to operate on patients. Khanna et al . also concluded in their study, that having a uniform standard of training can result in improvement of outcomes irrespective of the surgery performed. In this study, too we found that a statistically significant difference was seen among the 2 nd -year residents who showed a remarkable improvement regarding reduced intraoperative complications (posterior capsular rupture [PCR] P = 0.002) and better postoperative visual outcome after wet-laboratory training. There are various studies reporting the effectiveness of simulator training for phacoemulsification. The occurrence of intraoperative complications was significantly reduced in residents trained on the simulator. Suryawanshi et al . have tried a reverse method of training residents in cataract surgery and shown there is no difference in the conventional versus the reverse method of training. These studies only emphasize that various modalities have been tried to improve the resident surgical outcomes. The ACGME has recognized the importance of wet-laboratory and simulator training in ophthalmology residency and mandated the wet-laboratory or simulation training in the USA for ophthalmology training. The pitfall of this is the cost involved in setting up and maintaining the wet laboratory. However, in the long run, it definitely has the benefit of improving trainee confidence and the quality of surgeries. In the Indian scenario, cost of phacoemulsification may be a hindrance in providing it to the general population; hence, training in MSICS becomes imperative. Small-incision cataract surgery also requires microsurgical skill training and the residents would benefit by wet laboratory exposure. This is shown in our study, where the frequency of intraoperative complications significantly fell after the introduction of wet-laboratory training in our institute. Furthermore, postoperative visual outcome of resident performed surgery remarkably improved. This difference was seen in the 2 nd -year residents who revealed a statistically significant difference in both the rate of intraoperative complications (PCR; P = 0.002) and visual outcome. The rate of PC rent and vitreous loss in resident performed surgeries varies in different studies ranging from 4.9% to 10%. In this study, we have included all the levels of trainees - 1 st (JR1), 2 nd (JR2), and 3 rd -year residents (JR3) whereas most of the studies include only 2 nd and 3 rd -year residents. The rate of vitreous loss in the study, in Group A without prior access to wet-laboratory training is high - 14.3% whereas after wet-laboratory training, the rate of PCR and vitreous loss is 6.9% which is comparable with other studies. Carricondo et al . in their study of the 3 rd -year residents showed an 11.54% of intraoperative complications. In this study, they showed that the rate of intraoperative complications, which was as high as 14% in the first 40 cases, dropped to 7% after 80 cases. The results are comparable with this study with a remarkable improvement in the 2 nd -year residents revealing a rate of PCR of 5.6%, which was achieved between 40 and 80 cases. The studies that are published are for phacoemulisification training, and to the best of our knowledge, there is no publication on the effect of wet-laboratory training on small-incision cataract. Ours is the first publication addressing this issue. One of the limitations of this study is that it is a retrospective study. However, once the wet laboratory curriculum was instituted in our center, it was considered discriminatory to conduct a randomized trial where the advantage of wet laboratory curriculum would be provided to some students and the rest would be deprived of the same. Another limitation was that the innate skill levels of each resident might vary.
A wet-laboratory training facility plays a major role in enhancing the confidence and surgical skills in the resident which is ultimately manifested in the reduced rate of complications and a better visual outcome in resident-performed cataract surgery. The need to mandate the wet laboratory or simulation facility in ophthalmology training as a part of the curriculum in postgraduate training in India requires serious consideration to improve the cataract surgical outcomes. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.
Nil.
There are no conflicts of interest.
|
Patient preferences for CDK4/6 inhibitor treatments in HR+/HER2− early breast cancer: a discrete choice survey study | 84596214-b532-4503-a32d-e500b15a5aad | 11953128 | Neoplasms[mh] | Most new breast cancer (BC) cases are diagnosed as early BC (EBC) . Within 20 years of diagnosis, approximately 27–37% of patients with stage II EBC and 46–57% with stage III EBC will experience disease recurrence with standard-of-care (SOC) adjuvant endocrine therapy (ET), which are mostly distant metastatic recurrences [ – ]; half occur within the first 5 years . BC therapies in the adjuvant setting are administered to reduce the risk of recurrence and improve overall survival. Efficacy, safety, and quality of life all contribute to the overall treatment outcomes. Novel targeted therapies, such as cyclin-dependent kinase 4/6 inhibitors (CDK4/6is), have been studied in the adjuvant setting in patients with hormone receptor-positive, human epidermal growth factor receptor 2-negative (HR+/HER2−) EBC and have shown potential to transform adjuvant care when added to SOC ET. The addition of a CDK4/6i to adjuvant ET for the treatment of HR+/HER2− EBC has been investigated in phase 3 trials of all 3 CDK4/6is that have been approved by the US Food and Drug Administration in the advanced setting [ – ]. In the PALLAS and PENELOPE-B trials, the addition of palbociclib to adjuvant ET did not demonstrate improvement in invasive disease-free survival (iDFS) compared with ET alone . Based on the positive results from monarchE, in 2021 abemaciclib + ET was approved by the US Food and Drug Administration in patients with node-positive HR+/HER2− EBC with a high risk of recurrence . In monarchE, at a median 54-month follow-up, an iDFS benefit with abemaciclib + ET over ET alone was observed in the intention-to-treat population of patients with lymph node-positive, high-risk disease (hazard ratio [HR], 0.68; 95% CI 0.60–0.77) . NATALEE enrolled a broader population of patients with stage II/III HR+/HER2− EBC, including patients with node-negative disease with high-risk features . In an interim analysis of the NATALEE trial, a statistically significant iDFS benefit with ribociclib + a nonsteroidal aromatase inhibitor (NSAI) over NSAI alone (HR, 0.75; 95% CI 0.62–0.91; two-sided P = 0.0003) was observed at a median 27.7-month follow-up . The iDFS benefit with ribociclib + NSAI over NSAI alone was sustained over a median follow-up of 44.2 months (HR, 0.72; 95% CI 0.61–0.84; one-sided P < 0.0001) in an exploratory 4-year landmark analysis . Ribociclib and abemaciclib have been shown to be relatively safe for the treatment of HR+/HER2− BC, with distinct toxicity profiles for each CDK4/6i across the advanced BC (ABC) and EBC settings [ – ]. Higher rates of severe gastrointestinal adverse events (AEs) and fatigue were observed in patients treated with abemaciclib + ET vs ET alone; patients treated with ribociclib reported higher rates of hematologic AEs, neutropenia, elevated liver enzymes, and rare cases of QTc prolongation vs ET alone. These treatment-related AEs necessitated dose modifications and additional monitoring tests for both compounds in clinical trials and clinical practice . In NATALEE, the most common AE associated with ribociclib was neutropenia (any grade, 62.1%; grade ≥ 3, 43.8%); in monarchE, diarrhea was the most common AE associated with abemaciclib (any grade, 83.6%; grade ≥ 3, 7.8%) . These toxicities may lead to premature discontinuation of CDK4/6i. In the NATALEE trial, the most common AEs that led to discontinuation were elevated liver enzymes and arthralgia, and for monarchE, it was diarrhea . With ribociclib, the 400-mg starting dose investigated in the EBC setting was associated with lower rates of neutropenia and QTc prolongation compared with the 600-mg dose, which is the SOC dose in the ABC setting . Abemaciclib and ribociclib also differ in their dosing schedules, which may influence therapy adherence and persistence. In monarchE, patients received either 2 years of abemaciclib (150 mg twice daily on a continuous dosing schedule) + ET (for 5–10 years) or ET alone . In NATALEE, patients received 3 years of ribociclib (400 mg/day; 3 weeks on/1 week off schedule) + ET (for ≥ 5 years) or ET alone . To date, few studies have examined patient preferences for the different attributes of CDK4/6is in BC [ – ]. More research is needed, especially on patient preferences for attributes associated with the 2 CDK4/6is studied in the adjuvant setting, abemaciclib and ribociclib. We report the findings of a discrete choice experiment (DCE) survey that evaluated the extent to which US-based patients with HR+/HER2− EBC value different treatment attributes associated with ribociclib and abemaciclib in the adjuvant setting.
DCE method overview The DCE methodology is the most commonly used survey technique to evaluate stated preferences [ – ]. DCE surveys assume that any product or service can be described by a number of attributes (i.e., efficacy, risk of AEs, administration, or dosing) that can be compared and ranked according to patients’ selections . Prior to development of the DCE survey in this analysis, 14 one-on-one qualitative patient interviews were conducted via telephone to elicit the perspective of the US-based patients on their expectations when selecting treatments for adjuvant therapy. Patients also rated the importance they placed on various CDK4/6i treatment attributes on a scale of 1–5, with 5 being most important. An expert panel consisting of two breast oncologists and a patient advocate was consulted on the overall design and selection of attributes for the DCE survey. Based on the responses from the qualitative interviews and expert panel feedback, a list of eight attributes was considered for the DCE survey: treatment efficacy (5-year iDFS), venous thromboembolic events (VTEs), diarrhea, fatigue, number of blood tests, number of EKGs, treatment duration, and treatment schedule (Table ). Attributes related to additional AEs were not considered for the DCE based on low ratings in the qualitative interviews; these included the risk of other gastrointestinal AEs, laboratory test result-related AEs such as risk of neutropenia and elevation of liver enzymes, and other rare AEs. The levels reflecting the best and worst values for each attribute were informed by the primary publications of pivotal trials of abemaciclib in the adjuvant EBC setting and ribociclib in the advanced setting, since results from NATALEE were not available at the time of this DCE survey design (Table ) [ , , – ]. Survey methods Women aged ≥ 18 years diagnosed with HR+/HER2− stage II/III EBC, living in the USA, and receiving adjuvant ET at the time of completing the survey were enrolled. Participants who received prior (neo)adjuvant chemotherapy were eligible but were required to have stopped chemotherapy at the time of enrollment. Participants’ most recent BC surgery was required to be within 1–3 years prior to the enrollment date. Participants were excluded if they had been diagnosed with metastatic BC (stage IV); were first diagnosed with BC ≥ 5 years ago; had previously received a CDK4/6i; were currently enrolled in a BC clinical trial; or had previously received or were currently receiving treatments indicated for ABC only or other subtypes of BC (i.e., triple negative or HER2 +), including mTOR inhibitors, PD-L1 inhibitors, monoclonal antibodies/antibody–drug conjugates, or kinase inhibitors. A web-based DCE survey was conducted among participants between January and May 2023, before NATALEE results were available. Email invitations were sent to panel members and potential participants via physician networks, patient associations, and social media outreach, including a URL to a secure website link that hosted the survey and included a series of pre-survey questions to determine eligibility. Informed consent was collected from participants at the time of enrollment. Patients were incentivized to complete the survey. In addition to answering the DCE survey questions, participants provided self-reported information on demographics, socioeconomic characteristics, and disease history. Prior to the survey, participants were provided with a tutorial to help them understand the attributes and choice card design using patient-friendly language that was reviewed by the patient advocate. Prompts were implemented to ensure that participants did not rush to complete the survey, and data quality checks were conducted to exclude these participants. The DCE questions were presented as choice cards that consisted of pairs of hypothetical treatment profiles with varying levels for each attribute, and participants were asked to select the treatment option that best reflected their preference (Supplemental Fig. 1). In total, 32 choice cards were generated based on efficient design by optimizing the D-efficiency in SAS (Statistical Analysis System Enterprise Guide 7.1). To minimize responder bias, cards were randomized into four blocks, with two additional cards for validity checks (one for dominance test and one for stability test) in each block. For the dominance test, attribute levels in one treatment profile (dominant profile) were objectively better than those in the other profile (dominated profile); participants were expected to choose the dominant profile. For the stability test, participants were asked to respond to two choice cards with the same treatment profiles in reverse order; participants were expected to choose the same treatment profile on both cards. Participants who failed the dominance and stability tests were excluded in the sensitivity analysis. Only de-identified data were collected. Institutional review board exemption was received prior to the qualitative interviews and survey data collection. Statistical analysis A conditional logit regression model was estimated to analyze patients’ preferences between alternative treatment profiles and their willingness to trade off between these attributes. The dependent variable for the model was the patient’s preferred choice, and the independent variables were the attributes with pre-specified levels of the choice cards. Coefficients obtained from regression analyses were used to calculate the relative importance (RI) of the attributes and associated P values, indicating whether a specific attribute has any significant impact in terms of decision-making. RI scores were estimated as normalized percentages calculated by multiplying the coefficient for each attribute by the difference between the best and worst levels of the attribute, to allow for comparisons across attributes. The RI score for an attribute could be interpreted as the proportional value of a change from a treatment with the worst level to a treatment with the best level. A minimum acceptable benefit was calculated to estimate the minimum change in 5-year iDFS benefit required by patients to accept a treatment with a less desirable level of risk (best to worst level). An overall utility score (estimated overall preference) was derived from the model for treatment profiles that resembled abemaciclib and ribociclib profiles (assuming equal efficacy in the base case), including one conservative efficacy scenario: a 10% lower iDFS rate than in abemaciclib-like profiles to represent profiles resembling ribociclib (Supplemental Table 1) . Subgroup analyses were conducted for pre-/perimenopausal and postmenopausal status, as well as stage II BC and stage III BC. A sensitivity analysis was conducted by excluding patients who failed the stability and/or dominance tests to ensure robustness of results .
The DCE methodology is the most commonly used survey technique to evaluate stated preferences [ – ]. DCE surveys assume that any product or service can be described by a number of attributes (i.e., efficacy, risk of AEs, administration, or dosing) that can be compared and ranked according to patients’ selections . Prior to development of the DCE survey in this analysis, 14 one-on-one qualitative patient interviews were conducted via telephone to elicit the perspective of the US-based patients on their expectations when selecting treatments for adjuvant therapy. Patients also rated the importance they placed on various CDK4/6i treatment attributes on a scale of 1–5, with 5 being most important. An expert panel consisting of two breast oncologists and a patient advocate was consulted on the overall design and selection of attributes for the DCE survey. Based on the responses from the qualitative interviews and expert panel feedback, a list of eight attributes was considered for the DCE survey: treatment efficacy (5-year iDFS), venous thromboembolic events (VTEs), diarrhea, fatigue, number of blood tests, number of EKGs, treatment duration, and treatment schedule (Table ). Attributes related to additional AEs were not considered for the DCE based on low ratings in the qualitative interviews; these included the risk of other gastrointestinal AEs, laboratory test result-related AEs such as risk of neutropenia and elevation of liver enzymes, and other rare AEs. The levels reflecting the best and worst values for each attribute were informed by the primary publications of pivotal trials of abemaciclib in the adjuvant EBC setting and ribociclib in the advanced setting, since results from NATALEE were not available at the time of this DCE survey design (Table ) [ , , – ].
Women aged ≥ 18 years diagnosed with HR+/HER2− stage II/III EBC, living in the USA, and receiving adjuvant ET at the time of completing the survey were enrolled. Participants who received prior (neo)adjuvant chemotherapy were eligible but were required to have stopped chemotherapy at the time of enrollment. Participants’ most recent BC surgery was required to be within 1–3 years prior to the enrollment date. Participants were excluded if they had been diagnosed with metastatic BC (stage IV); were first diagnosed with BC ≥ 5 years ago; had previously received a CDK4/6i; were currently enrolled in a BC clinical trial; or had previously received or were currently receiving treatments indicated for ABC only or other subtypes of BC (i.e., triple negative or HER2 +), including mTOR inhibitors, PD-L1 inhibitors, monoclonal antibodies/antibody–drug conjugates, or kinase inhibitors. A web-based DCE survey was conducted among participants between January and May 2023, before NATALEE results were available. Email invitations were sent to panel members and potential participants via physician networks, patient associations, and social media outreach, including a URL to a secure website link that hosted the survey and included a series of pre-survey questions to determine eligibility. Informed consent was collected from participants at the time of enrollment. Patients were incentivized to complete the survey. In addition to answering the DCE survey questions, participants provided self-reported information on demographics, socioeconomic characteristics, and disease history. Prior to the survey, participants were provided with a tutorial to help them understand the attributes and choice card design using patient-friendly language that was reviewed by the patient advocate. Prompts were implemented to ensure that participants did not rush to complete the survey, and data quality checks were conducted to exclude these participants. The DCE questions were presented as choice cards that consisted of pairs of hypothetical treatment profiles with varying levels for each attribute, and participants were asked to select the treatment option that best reflected their preference (Supplemental Fig. 1). In total, 32 choice cards were generated based on efficient design by optimizing the D-efficiency in SAS (Statistical Analysis System Enterprise Guide 7.1). To minimize responder bias, cards were randomized into four blocks, with two additional cards for validity checks (one for dominance test and one for stability test) in each block. For the dominance test, attribute levels in one treatment profile (dominant profile) were objectively better than those in the other profile (dominated profile); participants were expected to choose the dominant profile. For the stability test, participants were asked to respond to two choice cards with the same treatment profiles in reverse order; participants were expected to choose the same treatment profile on both cards. Participants who failed the dominance and stability tests were excluded in the sensitivity analysis. Only de-identified data were collected. Institutional review board exemption was received prior to the qualitative interviews and survey data collection.
A conditional logit regression model was estimated to analyze patients’ preferences between alternative treatment profiles and their willingness to trade off between these attributes. The dependent variable for the model was the patient’s preferred choice, and the independent variables were the attributes with pre-specified levels of the choice cards. Coefficients obtained from regression analyses were used to calculate the relative importance (RI) of the attributes and associated P values, indicating whether a specific attribute has any significant impact in terms of decision-making. RI scores were estimated as normalized percentages calculated by multiplying the coefficient for each attribute by the difference between the best and worst levels of the attribute, to allow for comparisons across attributes. The RI score for an attribute could be interpreted as the proportional value of a change from a treatment with the worst level to a treatment with the best level. A minimum acceptable benefit was calculated to estimate the minimum change in 5-year iDFS benefit required by patients to accept a treatment with a less desirable level of risk (best to worst level). An overall utility score (estimated overall preference) was derived from the model for treatment profiles that resembled abemaciclib and ribociclib profiles (assuming equal efficacy in the base case), including one conservative efficacy scenario: a 10% lower iDFS rate than in abemaciclib-like profiles to represent profiles resembling ribociclib (Supplemental Table 1) . Subgroup analyses were conducted for pre-/perimenopausal and postmenopausal status, as well as stage II BC and stage III BC. A sensitivity analysis was conducted by excluding patients who failed the stability and/or dominance tests to ensure robustness of results .
Patient characteristics Data for the DCE were collected from 409 US-based women with HR+/HER2− EBC treated in the adjuvant setting; 49% of participants were pre-/perimenopausal and 48% had stage II disease at the time of the survey (Table ). The mean (SD) age in the overall sample was 54.5 (7.8) years, and the median age was 53.0 years. In the overall sample, 23% of participants were Black or African American, and approximately half of the participants were White (59%). The majority of participants were from the western region of the USA (40%), followed by the southern (25%), northeastern (15%), and mid-western (8%) regions. A quarter of the survey participants were employed either full time or part time (38%). Patient preferences in overall sample In the overall sample of 409 patients, in decreasing order of RI, higher efficacy (iDFS), lower risk of diarrhea, lower risk of fatigue, shorter treatment duration, and lower risk of VTE had a significant influence on participant’s treatment preference (Fig. , Table ). When making treatment decisions, participants placed the highest importance on treatment efficacy (RI: 33%; P < 0.001). An increased risk of AEs negatively impacted patients’ preferences. Among the AE attributes, risk of diarrhea was most important (RI: 26%; P < 0.001), followed by fatigue (RI: 14%; P < 0.001) and VTE (RI: 8%; P < 0.001). Shorter treatment duration positively influenced participants’ treatment preference but was of lower RI than risk of diarrhea or fatigue (RI: 12%; P < 0.001). Treatment schedule (RI: 0.9%; P = 0.519), number of blood tests (RI: 3%; P = 0.071), and number of EKGs (RI: 2%; P = 0.146) did not significantly influence participants’ treatment preference. On average, participants would require at least a 3.7-percentage-point increase in 5-year iDFS to tolerate a 2.0-percentage-point increase in risk of VTE (the difference between the worst and best values for the attribute) (Table ). Additionally, to tolerate a 61.8-percentage-point increase in risk of diarrhea or an 18.6-percentage-point increase in risk of fatigue, patients would require at least an 11.2- or 4.0-percentage-point increase in 5-year iDFS, respectively. Patient preferences in subgroups Subgroup results by BC stage and menopausal status were generally consistent with results in the overall sample (Fig. , Supplemental Tables 2 and 3). Participants with stage III BC placed higher RI on fatigue vs treatment duration (similar to the relative ordering of these attributes in the overall sample), while the relative ordering was reversed among participants with stage II BC. These findings may be due to a higher proportion of pre-/perimenopausal patients with stage II BC (64%) among the stage II subgroup and a higher proportion of postmenopausal patients with stage III BC (65%). Utility analysis Overall, treatment profiles in the DCE survey that resembled the profile of ribociclib had consistently higher utility scores, despite the higher preference for a lower treatment duration, including a scenario in which treatment efficacy was assumed to be 10% lower with profiles resembling ribociclib (Table ). The differences in utility scores were primarily driven by differences in the risk of diarrhea and fatigue. Similar results for utility differences were also seen in subgroups by BC stage or menopausal status, showing a stronger preference for profiles resembling ribociclib in patients with EBC (Supplemental Tables 4 and 5). Sensitivity analysis Results from sensitivity analyses, when excluding participants who failed dominance and/or stability tests, remained consistent with the main analyses, indicating the robustness of the findings (data not shown).
Data for the DCE were collected from 409 US-based women with HR+/HER2− EBC treated in the adjuvant setting; 49% of participants were pre-/perimenopausal and 48% had stage II disease at the time of the survey (Table ). The mean (SD) age in the overall sample was 54.5 (7.8) years, and the median age was 53.0 years. In the overall sample, 23% of participants were Black or African American, and approximately half of the participants were White (59%). The majority of participants were from the western region of the USA (40%), followed by the southern (25%), northeastern (15%), and mid-western (8%) regions. A quarter of the survey participants were employed either full time or part time (38%).
In the overall sample of 409 patients, in decreasing order of RI, higher efficacy (iDFS), lower risk of diarrhea, lower risk of fatigue, shorter treatment duration, and lower risk of VTE had a significant influence on participant’s treatment preference (Fig. , Table ). When making treatment decisions, participants placed the highest importance on treatment efficacy (RI: 33%; P < 0.001). An increased risk of AEs negatively impacted patients’ preferences. Among the AE attributes, risk of diarrhea was most important (RI: 26%; P < 0.001), followed by fatigue (RI: 14%; P < 0.001) and VTE (RI: 8%; P < 0.001). Shorter treatment duration positively influenced participants’ treatment preference but was of lower RI than risk of diarrhea or fatigue (RI: 12%; P < 0.001). Treatment schedule (RI: 0.9%; P = 0.519), number of blood tests (RI: 3%; P = 0.071), and number of EKGs (RI: 2%; P = 0.146) did not significantly influence participants’ treatment preference. On average, participants would require at least a 3.7-percentage-point increase in 5-year iDFS to tolerate a 2.0-percentage-point increase in risk of VTE (the difference between the worst and best values for the attribute) (Table ). Additionally, to tolerate a 61.8-percentage-point increase in risk of diarrhea or an 18.6-percentage-point increase in risk of fatigue, patients would require at least an 11.2- or 4.0-percentage-point increase in 5-year iDFS, respectively.
Subgroup results by BC stage and menopausal status were generally consistent with results in the overall sample (Fig. , Supplemental Tables 2 and 3). Participants with stage III BC placed higher RI on fatigue vs treatment duration (similar to the relative ordering of these attributes in the overall sample), while the relative ordering was reversed among participants with stage II BC. These findings may be due to a higher proportion of pre-/perimenopausal patients with stage II BC (64%) among the stage II subgroup and a higher proportion of postmenopausal patients with stage III BC (65%).
Overall, treatment profiles in the DCE survey that resembled the profile of ribociclib had consistently higher utility scores, despite the higher preference for a lower treatment duration, including a scenario in which treatment efficacy was assumed to be 10% lower with profiles resembling ribociclib (Table ). The differences in utility scores were primarily driven by differences in the risk of diarrhea and fatigue. Similar results for utility differences were also seen in subgroups by BC stage or menopausal status, showing a stronger preference for profiles resembling ribociclib in patients with EBC (Supplemental Tables 4 and 5).
Results from sensitivity analyses, when excluding participants who failed dominance and/or stability tests, remained consistent with the main analyses, indicating the robustness of the findings (data not shown).
This analysis used a DCE survey to assess patient preferences for attributes related to the efficacy, safety, and dose administration of CDK4/6is in the adjuvant setting among patients with HR+/HER2− EBC. In order of importance, higher efficacy (assessed by iDFS rate at 5 years), lower risk of diarrhea, lower risk of fatigue, shorter treatment duration, and lower risk of VTE had a significant influence on participant treatment preferences. Attributes related to the number of blood tests needed to monitor laboratory abnormalities and EKGs to monitor signs of abnormal heart rhythm did not significantly influence patients’ preferences for treatments. Overall, among patients with EBC, treatment profiles that approximated the profile of ribociclib had a higher estimated average utility, despite patients’ preference for shorter treatment duration. The same preferences were generally evident across subgroups by menopausal status or BC stage. These findings emphasize the importance of considering the preferences of patients who are eligible for monarchE (node positive, high risk, stage II/III) or NATALEE (select N0 and all macroscopic N1, stage II/III) along with other relevant factors (e.g., patient history, drug–drug interactions, comorbid clinical conditions, and clinical experience managing AEs) when clinicians and patients are making treatment decisions together. A previously published DCE survey of attributes associated with CDK4/6is + ET vs ET alone, including oncologists and patients with stage II/III HR+/HER2− EBC who had received prior ET in the adjuvant setting, identified iDFS as the most important attribute to patients and oncologists . In that analysis, to accept an increased risk of diarrhea (from 11% [lowest attribute level] to 81% [highest attribute level]), oncologists and patients required an improvement in iDFS of 5.6 and 8.0 percentage points over the base level, respectively; patients were less willing to trade off a higher risk of diarrhea for a more modest improvement in efficacy. Additional studies in the advanced setting have also reported consistent findings on patients’ preference for a lower risk of symptomatic AEs . A survey of healthcare providers and patients with HR+/HER2− ABC found that many patients were unwilling to tolerate AEs (e.g., fatigue, pain, diarrhea, and loss of appetite), even if their treatment was effective . Some limitations of this DCE survey need to be acknowledged. Since the DCE was designed prior to availability of the NATALEE trial results, the levels of the safety attributes were informed by metastatic BC trials of ribociclib in addition to the safety results from monarchE. The use of safety data from metastatic trials may be perceived as a limitation; however, the wide range for the attribute levels to ensure conservativeness could have mitigated any impact of this on the results. To reduce responder burden, the DCE survey did not consider all possible attributes. Some, such as risk of neutropenia or liver enzyme elevation, could be important to clinicians, and their exclusion may be perceived as a limitation. These attributes were excluded from the DCE survey based on their low ranking (2.4 out of 5 for neutropenia; 2.8 out of 5 for liver enzyme elevation) by participants in the 14 qualitative interviews conducted prior to the DCE survey. To potentially mitigate this limitation, the AEs and the potential for requiring further dose adjustments or interruptions were explained to participants in the DCE survey tutorial in the context of the number of blood tests required for monitoring these AEs. However, this did not impact the RI that participants assigned to blood test monitoring. Patient out-of-pocket (OOP) costs were not considered as an attribute in this DCE survey, which could be perceived as a limitation given the high sensitivity to drug prices. In the real world, the OOP cost burden can vary significantly based on several factors, including drug prices, the number of medicines a patient is taking, health insurance coverage, and accessibility of OOP cost assistance programs. A high OOP cost burden could impact patients’ preference for treatments. In general, patients who participate in online survey studies tend to be more educated, to be younger, and to have better health status than the average patient with EBC. Furthermore, the findings of this study are only applicable to patients who are considering adding a CDK4/6i to adjuvant ET; in the real world, some patients will not consider an add-on treatment to ET. Hence, the overall sample in this DCE survey may not be fully representative of the larger stage II/III EBC population in the USA. The overall sample size enrolled in this DCE survey is sufficiently robust to mitigate any significant bias due to a less representative sample. Lastly, DCE surveys inherently rely on the assumption that participants make rational choices, whereas this may not always be true in the real world. Sensitivity analyses were conducted to ensure internal validity of the results. This DCE survey showed that patients with EBC placed greater importance on efficacy and lower risk of symptomatic AEs (diarrhea, fatigue, and VTE) over other included attributes when making treatment choices based on attributes of CDK4/6is in the adjuvant setting. These findings suggest that patients with stage II/III HR+/HER2− EBC may exhibit a greater preference for treatment profiles that resemble ribociclib. These findings could aid in physician–patient combined decision-making when discussing the addition of a CDK4/6i to SOC adjuvant treatment for eligible patients with HR+/HER2− EBC.
Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 112 KB)
|
The effect of USM-IAM-based counselling vs standard counselling on insulin adherence, FBS and HbA1c among patients with uncontrolled type 2 diabetes mellitus (T2DM): a randomised controlled trial | 2118c122-64a5-4644-9804-0c717161e351 | 11256455 | Patient Education as Topic[mh] | Type 2 diabetes mellitus (T2DM) is the most common form of diabetes among adults. In 2021, approximately 537 million people globally were reported to have diabetes, with projections estimating an increase to 783 million by 2045 . The prevalence of T2DM among Malaysians has surged over the past two decades, from 6.3% in 1986 to 8.3% in 1996 to 13.4% in 2015 and further climbing to 18.3% in 2019 . Patients diagnosed with diabetes mellitus are frequently prescribed oral glucose-lowering drugs (OGLDs) and/or insulin and/or other injectable agents. This regimen is combined with a healthy diet and increased physical activity to achieve glycaemic control. Both Malaysian and international guidelines recommend the combination of OGLDs with insulin/injectable agents when patients need more efficacious approaches to attain their glycaemic goals . Despite the development of injectables such as dulaglutide and semaglutide, which have proven efficacy in reducing HbA1c and cardiac events , insulin continues to be the key treatment for many patients due to its availability and affordability. The majority of patients with diabetes mellitus will ultimately need insulin 8 to 10 years after the diagnosis of diabetes to maintain good glycaemic control . Most of those on single insulin injection will require intensification within 3 years of insulin initiation . Despite the increase in the number of patients using insulin, the percentage of patients who achieve targeted HbA1c levels remains low. In Malaysia, 32.4% of patients achieved the HbA1c target of ≤6.5% . A literature review revealed that adherence to insulin is generally poor among people with diabetes, and the adherence rates are lower than those for oral hypoglycaemic agents . Several factors contribute to nonadherence to insulin, including pain associated with injections , fear of hypoglycemia and weight gain , interference of injections with daily routines and embarrassment associated with administering insulin in public . In Malaysia, the same factors were identified as barriers to insulin adherence , along with other issues such as myths and misconceptions toward insulin and forgetfulness . Reports have emphasised that poor diabetic control among patients treated with insulin primarily results from a lack of understanding of the disease, a lack of knowledge of glycaemic targets, and a lack of knowledge of diabetes self-care, particularly self-insulin dose adjustment . Inadequate knowledge of diabetes leads to low adherence to self-care practices such as a diabetes diet, exercise and self-monitoring of blood glucose (SMBG) . There is a growing body of evidence highlighting the importance of education in improving HbA1c and increasing diabetes knowledge . Based on local data, more than 90% of our patients are T2DM and most of these patients have poor diabetes control. Despite 82% of patients having been counselled at least once by diabetic educators, the mean HbA1c of patients in the treatment centre was 10% . In response, we have introduced a new section on nonadherence to insulin therapy, which includes its definition, causes of nonadherence, and suggested solutions to problems with insulin injection in a newly developed module. The introduction of these contents to the USM-IAM may provide more insight to patients about insulin nonadherence, the causes and their solutions. This additional information may lead to better insulin adherence among patients. Thus, in this study we aimed to compare the effects of diabetes counselling: based on this newly developed module (USM-IAM) with that of standard counselling (SC) on insulin adherence, FBS and HbA1c among patients with uncontrolled T2DM in a 6 months’ duration study.
Study design and approval The study was a single-centre, randomised, parallel, controlled trial with a 6-month follow-up period. It was performed at the Endocrine Clinic of Hospital Universiti Sains Malaysia, a tertiary facility hospital located on the east coast region of Malaysia. The study protocols have been performed in accordance with the Declaration of Helsinki and was approved by the Research Ethics Committee of Universiti Sains Malaysia, with identity number USM/JEPeM/20110605, and registered with ClinicalTrials.gov with ID NCT05125185 dated 17th November 2021. This study was reported in adherence to CONSORT guidelines . Prior to the commencement of the study, the educators who would be involved in providing the counselling were identified. Two educators were allocated to each group. Standardisations for counselling were conducted, instructing educators who would administer standard counselling (SC) to use a specified flip chart prepared based on a manual , along with insulin pen models and handouts. In contrast, educators in USM-IAM were trained in counselling using the module. Sample size estimation & randomization The minimum required sample size was calculated by comparing the means of the IAQDM score with a standard deviation of 4.3 , difference of 2.0, precision of 0.05, and alpha and power of 80%. The minimum sample was 148, exaggerated to 178 (20% dropout rate). A co-researcher generated the group allocation list through permuted block randomisation with equal allocation (1:1). A research assistant printed the group allocation list and concealed it in 180 envelopes. The researcher opened the envelopes chronologically on the patients’ recruitment day. Recruitment took place from August 2021 through July 2022. Patients aged between 18 and 65 years, diagnosed with T2DM as defined by International Diabetes Federation (IDF), prescribed insulin for at least 1 year, and had an HbA1c level between 8 and 15 were included in the study. The HbA1c range was selected to focus on educating patients with poor glycaemic control. Patients were excluded if they did not understand the Malay language or were illiterate; had severe diabetes complications such as chronic kidney disease, heart failure, or severe proliferative diabetic retinopathy; experienced recurrent hypoglycemia or hypoglycemia unawareness; or were obese with a body mass index (BMI) ≥ 40 kg/m 2 . Eligible patients were invited to participate, and for those who agreed, additional information was provided before they were asked to sign the study’s consent form. Data collection On the recruitment day, blood was drawn to confirm patient eligibility and establish baseline glycaemic indices. Once eligibility was verified, patients completed forms gathering information on sociodemographic and clinical characteristics. The participants self-reported their adherence to insulin therapy, SMBG, insulin dose, dietary adjustment and problems with insulin therapy using the Insulin Adherence Questionnaire for patients with Diabetes Mellitus (IAQDM) . The Morisky Medication Adherence Scale-8 (MMAS-8) and the Malaysian Medication Adherence Scale (MALMAS) are the most widely used questionnaires for assessing medication adherence in Malaysia . However, these questionnaires were not utilised in this study because they do not evaluate aspects of insulin adherence such as glucose monitoring, insulin dose, dietary adjustment, or problems with injections. Instead, we employed the IAQDM, which has been validated for assessing insulin adherence among the Malay population in our study centre . Nasruddin at. el. utilised the questionnaire among 249 patients with T2DM treated with insulin from five health clinics in the Klang district, Selangor, Malaysia . The researcher or her assistant addressed any clarification needed for the questions. Patients were subsequently randomly assigned to one of two groups, standard counselling (SC) or USM-IAM, based on the allocations in the envelopes. Each counselling session was conducted individually by a diabetic educator. Educators in the SC group (SCG) educated patients using a flip chart and insulin pen models. The patients were informed about blood sugar targets, insulin injection techniques, insulin dose modification and how to recognise as well as manage hypoglycemia events. All patients were given SMBG diaries to record their blood sugar levels and any incidents of hypoglycemia. The participants were instructed and encouraged to intensify their insulin dosage weekly, and a handout containing instructions to optimise their insulin dosage based on their SMBG readings was given. The researcher’s and educator’s phone numbers were provided for any questions or concerns. In the USM-IAM, educators provided instruction to participants using the USM-IAM module. The development and validation of the module, which integrates information obtained from focus group findings among patients, were performed. The content validity index (CVI) of the 20 items in the newly produced module was 0.92, and the face validity agreement rate ranged from 86 to 97%. The module was distributed as a 40-page, A5-sized booklet containing text, 44 pictures and 12 tables. The USM-IAM approach differed from the SC approach because it explained the relationship between insulin and diabetes, defined nonadherence to insulin, discussed causes of nonadherence, provided measures to overcome nonadherence and covered fasting safely with insulin therapy. Most of the population in this study fasted at least 1 month every year, with some of them fasting more, as encouraged by the religion. Educators guided participants through the module, and a summary of its content is provided in Table . Participants were encouraged to ask questions based on their problems and discuss practical solutions with the educator. Participants were given the module and could refer to it freely when needed, with the request not to share it with others until the study’s completion. Like in the SCG, all participants received SMBG diaries with the same instructions. During the second visit (3 months later), blood was drawn, and adherence scores were recorded again. Participants received counselling once more, provided by the same educator. At the final visit (sixth month later), blood was drawn again, and adherence scores were recorded once more. The outcome measures included changes in the adherence score; and FBS and HbA1c levels from baseline to 3 months and from baseline to 6 months. Adherence scores were assessed using the IAQDM, HbA1c levels were measured using the capillary electrophoresis method, and FBS was measured using the glucose oxidase method. Figure illustrates the study design. Data analysis All the data were entered and analysed using the Statistical Package for Social Sciences (SPSS) version 28. A comparison of sociodemographic and clinical characteristics at baseline was made using chi-square tests for categorical data, and paired t-tests were utilised for normally distributed continuous data. As the study aimed to assess the effectiveness of the new education module on patients’ adherence and glycaemic parameters, the missing data were imputed, and intention-to-treat analysis was used . Repeated-measures analysis of variance (RM ANOVA) was applied to compare the mean differences in IAQDM, FBS and HbA1c between the USM-IAM and SCG over time. The results were interpreted by the p- value of the F- test, followed by the estimated marginal means. Changes in adherence scores, FBS, and HbA1c levels were analysed between baseline and 6 months. Subgroup analysis was then performed by stratification based on age groups (0–59 or ≥ 60), gender, education level (up to secondary or tertiary: diploma through PhD), and duration of diabetes (< 10 years, ≥ 10 years) to provide insights into whether these factors impact the results differently. For all analyses, a p -value of < 0.05 is considered to be significant.
The study was a single-centre, randomised, parallel, controlled trial with a 6-month follow-up period. It was performed at the Endocrine Clinic of Hospital Universiti Sains Malaysia, a tertiary facility hospital located on the east coast region of Malaysia. The study protocols have been performed in accordance with the Declaration of Helsinki and was approved by the Research Ethics Committee of Universiti Sains Malaysia, with identity number USM/JEPeM/20110605, and registered with ClinicalTrials.gov with ID NCT05125185 dated 17th November 2021. This study was reported in adherence to CONSORT guidelines . Prior to the commencement of the study, the educators who would be involved in providing the counselling were identified. Two educators were allocated to each group. Standardisations for counselling were conducted, instructing educators who would administer standard counselling (SC) to use a specified flip chart prepared based on a manual , along with insulin pen models and handouts. In contrast, educators in USM-IAM were trained in counselling using the module.
The minimum required sample size was calculated by comparing the means of the IAQDM score with a standard deviation of 4.3 , difference of 2.0, precision of 0.05, and alpha and power of 80%. The minimum sample was 148, exaggerated to 178 (20% dropout rate). A co-researcher generated the group allocation list through permuted block randomisation with equal allocation (1:1). A research assistant printed the group allocation list and concealed it in 180 envelopes. The researcher opened the envelopes chronologically on the patients’ recruitment day. Recruitment took place from August 2021 through July 2022. Patients aged between 18 and 65 years, diagnosed with T2DM as defined by International Diabetes Federation (IDF), prescribed insulin for at least 1 year, and had an HbA1c level between 8 and 15 were included in the study. The HbA1c range was selected to focus on educating patients with poor glycaemic control. Patients were excluded if they did not understand the Malay language or were illiterate; had severe diabetes complications such as chronic kidney disease, heart failure, or severe proliferative diabetic retinopathy; experienced recurrent hypoglycemia or hypoglycemia unawareness; or were obese with a body mass index (BMI) ≥ 40 kg/m 2 . Eligible patients were invited to participate, and for those who agreed, additional information was provided before they were asked to sign the study’s consent form.
On the recruitment day, blood was drawn to confirm patient eligibility and establish baseline glycaemic indices. Once eligibility was verified, patients completed forms gathering information on sociodemographic and clinical characteristics. The participants self-reported their adherence to insulin therapy, SMBG, insulin dose, dietary adjustment and problems with insulin therapy using the Insulin Adherence Questionnaire for patients with Diabetes Mellitus (IAQDM) . The Morisky Medication Adherence Scale-8 (MMAS-8) and the Malaysian Medication Adherence Scale (MALMAS) are the most widely used questionnaires for assessing medication adherence in Malaysia . However, these questionnaires were not utilised in this study because they do not evaluate aspects of insulin adherence such as glucose monitoring, insulin dose, dietary adjustment, or problems with injections. Instead, we employed the IAQDM, which has been validated for assessing insulin adherence among the Malay population in our study centre . Nasruddin at. el. utilised the questionnaire among 249 patients with T2DM treated with insulin from five health clinics in the Klang district, Selangor, Malaysia . The researcher or her assistant addressed any clarification needed for the questions. Patients were subsequently randomly assigned to one of two groups, standard counselling (SC) or USM-IAM, based on the allocations in the envelopes. Each counselling session was conducted individually by a diabetic educator. Educators in the SC group (SCG) educated patients using a flip chart and insulin pen models. The patients were informed about blood sugar targets, insulin injection techniques, insulin dose modification and how to recognise as well as manage hypoglycemia events. All patients were given SMBG diaries to record their blood sugar levels and any incidents of hypoglycemia. The participants were instructed and encouraged to intensify their insulin dosage weekly, and a handout containing instructions to optimise their insulin dosage based on their SMBG readings was given. The researcher’s and educator’s phone numbers were provided for any questions or concerns. In the USM-IAM, educators provided instruction to participants using the USM-IAM module. The development and validation of the module, which integrates information obtained from focus group findings among patients, were performed. The content validity index (CVI) of the 20 items in the newly produced module was 0.92, and the face validity agreement rate ranged from 86 to 97%. The module was distributed as a 40-page, A5-sized booklet containing text, 44 pictures and 12 tables. The USM-IAM approach differed from the SC approach because it explained the relationship between insulin and diabetes, defined nonadherence to insulin, discussed causes of nonadherence, provided measures to overcome nonadherence and covered fasting safely with insulin therapy. Most of the population in this study fasted at least 1 month every year, with some of them fasting more, as encouraged by the religion. Educators guided participants through the module, and a summary of its content is provided in Table . Participants were encouraged to ask questions based on their problems and discuss practical solutions with the educator. Participants were given the module and could refer to it freely when needed, with the request not to share it with others until the study’s completion. Like in the SCG, all participants received SMBG diaries with the same instructions. During the second visit (3 months later), blood was drawn, and adherence scores were recorded again. Participants received counselling once more, provided by the same educator. At the final visit (sixth month later), blood was drawn again, and adherence scores were recorded once more. The outcome measures included changes in the adherence score; and FBS and HbA1c levels from baseline to 3 months and from baseline to 6 months. Adherence scores were assessed using the IAQDM, HbA1c levels were measured using the capillary electrophoresis method, and FBS was measured using the glucose oxidase method. Figure illustrates the study design.
All the data were entered and analysed using the Statistical Package for Social Sciences (SPSS) version 28. A comparison of sociodemographic and clinical characteristics at baseline was made using chi-square tests for categorical data, and paired t-tests were utilised for normally distributed continuous data. As the study aimed to assess the effectiveness of the new education module on patients’ adherence and glycaemic parameters, the missing data were imputed, and intention-to-treat analysis was used . Repeated-measures analysis of variance (RM ANOVA) was applied to compare the mean differences in IAQDM, FBS and HbA1c between the USM-IAM and SCG over time. The results were interpreted by the p- value of the F- test, followed by the estimated marginal means. Changes in adherence scores, FBS, and HbA1c levels were analysed between baseline and 6 months. Subgroup analysis was then performed by stratification based on age groups (0–59 or ≥ 60), gender, education level (up to secondary or tertiary: diploma through PhD), and duration of diabetes (< 10 years, ≥ 10 years) to provide insights into whether these factors impact the results differently. For all analyses, a p -value of < 0.05 is considered to be significant.
One hundred eighty participants were initially enrolled in the study and were randomised equally to both groups. The sociodemographic and clinical characteristics of the participants at baseline are shown in Tables , , and respectively. At baseline, the two groups were homogeneous concerning all the variables. The patients included in our study had a mean age of 56 years. One hundred-twenty-three (69%) of the patients were 54 years old or older. The patients have long-standing diabetes and a mean duration of insulin injection of 7 years. Even though most participants underwent SMBG: performed capillary blood glucose measurements at least once a week, their HbA1c levels were still extremely high. Less than half of the participants exercised regularly: exercising five times a week for 30 minutes despite many being pensioners or housemakers. Many of the participants had comorbidities, e.g., hypertension and hyperlipidaemia. Systolic blood pressure readings were mostly not on achieved targets. Majority of the participants were on metformin-insulin combinations, antihypertensives and statins. During enrolment, 39 patients declined to give consent, and five patients had an HbA1c level out of the inclusion range. During the study period, 21 (11.7%) participants were excluded (10 in the USM-IAM and 11 in the SCG (Fig. )). The reasons for dropout were missed counselling sessions, blood sampling or completion of the adherence score. Complete case (per protocol) analysis was inadequate, and data imputation was needed. Missing data were imputed using the nonparametric missing value imputation via the missForest method . Imputation diagnosis was achieved by comparison of the data distributions by histograms. The imputation process retained the normal distribution shape of the variables. Both groups had significant improvements in adherence scores and HbA1c. There was a significant reduction in FBS in the USM-IAM but not in the SCG, as shown in Table . Figures , , and show profile plots of the adherence score, FBS and HbA1c, respectively. The between-subjects effects for all outcomes were not significantly different, as both groups experienced improvements. Table . From the sub-analysis, there was better insulin adherence observed among patients who received the USM-IAM in patients with diabetes for more than 10 years, compared to shorter diabetes duration. Those with a longer duration of diabetes might have more experience with insulin therapy and improved adherence with additional information in the new module. However, due to a small sample size in the sub-group of patients, a more conclusive analysis cannot be made.
The primary outcome of this study was the change in adherence score. In our study, increases in adherence scores were observed in both groups. As participants in both groups received counselling and were encouraged to titrate their insulin dosage based on their SMBG, they had to adhere to the prescribed insulin regimen to facilitate dosage adjustment. In addition, educators in the USM-IAM group discussed the relationship between insulin and diabetes, explored the causes of nonadherence and provided measures to overcome nonadherence. By understanding the causes of nonadherence and measures to overcome them, participants in the USM-IAM could promptly respond and demonstrate greater adherence to their insulin therapy, as reflected by the greater increase in adherence score. Other studies assessing the effect of education on adherence also showed improvements as measured by the Diabetes Management Self-Efficacy Scale (DMSES), the Summary of Diabetes Self-Care Activities (SDSCA) , and the MMAS-8 . According to a study that utilised similar adherence tools to our study (IAQDM), only 8.4% of their patients had a score ≥ 80, indicating adherence. In this study, we found that 8.9% of participants achieved a score of ≥80 at baseline, a percentage that was almost similar to that in the aforementioned study [8.9% vs. 8.4% ]. Since the study by Nasruddin et al. was cross-sectional, post-intervention scores were not reported. In our study, the percentage of adherent patients among the 180 patients improved to 34.7 and 34.2% at 3 months and 6 months, respectively. Unfortunately, the mean adherence score in the study conducted in Klang (another state in Malaysia) was not reported, preventing a direct comparison. The second outcome of the study was change in FBS. The USM-IAM group exhibited statistically significant reductions in FBS levels between baseline and the third month and between baseline and the sixth month. In contrast, the control group showed nonsignificant reductions. Guo et al. assessed the efficacy of structured education in patients with type 2 diabetes mellitus receiving insulin treatment (OPENING) in China. They also evaluated changes in FBS among their education and control groups. The study revealed a significant reduction in FBS from baseline to the fourth month in both groups . The education group displayed a greater reduction in FBS than the control group (2.84 ± 3.46 vs 2.76 ± 3.59). It is worth noting that patients in the OPENING study, who had higher FBS changes than did those our study [1.71 ± 1.24 mmol/l (intervention) vs. 1.0 ± 1.02 mmol/l (control)], had double OGLDs and had not yet initiated insulin treatment. During the 4-month study, these patients began two daily injections of 30% soluble–70% isophane recombinant insulin, which, combined with education, significantly reduced the FBS levels. In this study, both groups showed statistically significant reductions in HbA1c between baseline and the third month, and between baseline and the sixth month. The mean HbA1c reductions in the sixth month were 1.03 and 0.67% in the USM-IAM and SCG, respectively. Although greater HbA1c reductions were observed in the USM-IAM, the difference was not significant. Similar patterns were observed in earlier randomised trials, namely, the MEDIAS-2 ICT , OPENING and Cani et al. studies . The first two studies involved physician-led interventions, while the latter involved pharmacist-led interventions. Table provides a summary of RCTs of education intervention for T2DM patients treated with insulin. It also provides education related to insulin and diabetes included in the education interventions. There are notable differences between our study and the previously mentioned studies. In the OPENING study, patients experienced greatest decrease in HbA1c, exceeding 2% for both groups. Patients who did not self-inject throughout the 4-month study were excluded. The initiation of insulin significantly reduced HbA1c levels, complemented by intensive education, including six face-to-face education sessions and three telephone follow-ups. In Cani et al.’s study, both the control and intervention groups showed HbA1c reductions, with greater reductions observed in the intervention group (0.57% vs 0.08%). Notably, the intervention group received six monthly (six) individual education sessions, while the control group received no education sessions, and was followed up only at the beginning and end of the study. It is important to highlight that Cani et al.’s study differs from ours, as the comparator group in our study received standard counselling. Participants recruited for our study had been treated with insulin and had HbA1c values exceeding 8%, similar to those in the MEDIAS-2 ICT study. Despite our patients receiving only two education sessions by the diabetic educators compared to the 10 education sessions in the MEDIAS-2 ICT, participants in our study demonstrated better HbA1c reductions than did those in the MEDIAS-2 ICT. Specifically, the intervention group in our study showed a reduction in HbA1c of 1.03% as compared to MEDIAS-2 ICT’s 0.63%, while the control group in our study had a reduction of 0.59% compared to MEDIAS-2 ICT’s 0.37%. Notably, our study was conducted immediately after the lifting of the movement control order (MCO) in response to the COVID-19 pandemic in Malaysia, during which time specialist clinics started to resume their services back towards pre-pandemic levels. During the pandemic, patients who visited specialist clinics for follow-up only received prescriptions without consulting a doctor, a practice adopted globally to conserve resources and minimise the risk of COVID-19 transmission . These patients face various challenges, such as limited income, reduced access to needles and glucometer strips and restrictions on consultations with doctors and dieticians . When the pandemic was under control and life returned to normal, patients became more motivated to improve their diabetic control. In our study, participants received individual counselling, which differed from the group counselling practised in the MEDIAS-2 ICT study. Individual counselling allows participants better opportunities to ask questions and clarify doubts. The 10 lessons covered in the interventions of the MEDIAS-2 ICT study might be redundant for patients, and the limited chance of asking questions could have reduced their motivation and made them feel burdened. In addition to these factors, the module content in our study differed from the counselling given in the other three comparison studies. While our intervention group’s counselling covered all the topics in the three studies, our module also included additional information and discussion on the relationship between insulin and diabetes, the definition of nonadherence, the common causes of nonadherence and their suggested solutions, and how to fast safely with insulin therapy. In the module of the current study, the section on the relationship between insulin and diabetes explained the physiological role of insulin in the human body. Insulin functions as a key to opening the door for blood glucose to enter body cells. This process reduces blood glucose levels, allowing body cells to utilise glucose as fuel for energy production. Participants were also educated about the different types of insulin, including prandial and basal insulin, and their respective functions. This information is crucial for enhancing participants’ understanding of the importance of insulin usage, which can lead to improved adherence. The definition of insulin nonadherence in the module included not injecting as scheduled, not optimising the insulin dose, changing the dosage, or altering the frequency of insulin injections without consulting the treating doctor. The causes of nonadherence and their suggested solutions, gathered from focus group discussions, served as guidance for participants on how to address potential problems they faced with insulin treatment. In the section on fasting safely with insulin therapy, participants were informed about potential complications that can occur during fasting and about the timing of blood sugar monitoring and indications that necessitate breaking the fast. Compared to the other three studies discussed earlier, none of them covered the definition of insulin nonadherence, the causes of insulin nonadherence, and measures to overcome nonadherence. These new topics provided additional valuable information to participants, especially those who received insulin injections as part of their medication regimen. This new information could be used in counselling patients with T2DM to increase their insulin adherence. In our study, we noticed significant improvements occurring between 0 and 3 months and 0–6 months. However, the changes between 3- and 6 months were not significant for any of the three outcomes. This indicates that only minimal improvements in adherence, FBS, and HbA1c occurred between the third and sixth months. The similarity in education content during the second visit, which mirrored the content at recruitment, limited the changes participants could make within 3 months. Therefore, a different counselling content (dynamic intervention) may be necessary to enhance adherence and achieve further HbA1c reduction. This study has several limitations, notably, it was conducted at a single centre. Our findings relied on self-reports, introducing the possibility of recall bias and patients responded to please the researchers. These biases can lead to underestimation or overestimation of adherence. However, self-report questionnaires are widely used for measuring adherence due to their cost-effectiveness and time efficiency. Although they offer precise and reasonable estimates of adherence , we recommend combining self-administered questionnaires with pharmacological monitoring for a more precise measurement of insulin adherence. In conclusion, our findings indicate poor adherence to insulin therapy among participants, which improved after undergoing education module intervention. This was evidenced by reductions in HbA1c and FBS levels, along with an increase in the insulin adherence score. The counselling sessions had significant impacts on both groups, with USM-IAM showing a better effect than the standard counselling.
Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Supplementary Material 4.
|
Nitrogen single and multiple breath washout test and lung imaging to detect treatment-related pulmonary toxicity in paediatric cancer patients and survivors: a systematic review | 93f9d80e-9849-469c-a665-3ff41b022cd2 | 11751724 | Surgical Procedures, Operative[mh] | Childhood cancer treatment occurs at a stage of lung growth and development with potential long-term pulmonary consequences . Advances in treatment and supportive care have led to increasing survival rates and enable paediatric oncologists to focus on adverse health outcomes, including pulmonary morbidity due to pulmotoxic treatment . Pulmotoxic treatment includes haematopoietic stem cell transplantation (HSCT), chemotherapy (including bleomycin, busulfan, carmustine and lomustine), chest radiation and thoracic surgery . Lung damage caused by local and systemic cancer treatments involves various pathophysiological mechanisms, including the release of oxygen radicals, vascular damage and inflammatory processes in lung tissues. Not all mechanisms are fully understood, and most patients receive a combination of different treatment modalities . For example, bleomycin induces oxidative stress , while ionising radiation causes free radical DNA damage . Inflammatory responses and immune-mediated cell damage are relevant mechanisms of cell damage after HSCT and radiotherapy. Suspected pulmotoxic treatments are those believed to potentially harm the lungs, such as methotrexate and gemcitabine, and also newer targeted therapies and immunotherapies . Because the lungs can compensate well for small amounts of damage, the diagnosis of pulmonary dysfunction is often delayed in asymptomatic patients. Some treatment protocols include recommendations for monitoring lung health to detect pulmotoxic effects during and after cancer treatment . For example, in HSCT patients, spirometry tests both before and after the procedure are recommended because HSCT carries significant risks for severe pulmonary damage, such as pneumonitis and lung fibrosis. Recently published long-term follow-up guidelines for childhood cancer survivors recommend diffusing capacity of the lungs for carbon monoxide and spirometry as diagnostic modalities . Initially, pulmotoxic cancer treatment mostly affects the small peripheral airways . However, spirometry offers less sensitivity for detecting peripheral airway abnormalities because it mainly assesses functional abnormalities of the large airways . Washout tests can specifically detect peripheral airway abnormalities by assessing gas mixing within the entire ventilated lung, because this occurs predominantly in the peripheral airways (95% of the lung volume) . In various lung diseases, including in a study of paediatric cancer survivors , washout tests are more sensitive than spirometry in detecting peripheral airway abnormalities . Structural abnormalities of the lung resulting from pulmotoxic treatment can be assessed using computed tomography (CT) or magnetic resonance imaging (MRI). Imaging modalities such as X-ray, radionuclide imaging, positron-emission tomography, ventilation-perfusion scans, scintigraphy and ultrasound can be used to monitor disease progression, treatment response and disease recurrence. Pulmotoxic treatment leads to abnormalities including air space consolidation, mosaic attenuation and septal thickening . These abnormalities may occur at different time points after treatment and can be classified as acute, subacute and chronic. They can be assessed using CT, but this involves exposure to ionising radiation . MRI involves no exposure, but has not been commonly used to evaluate treatment-related pulmonary toxicity in paediatric cancer patients and survivors. It is already routinely used to monitor lung abnormalities in other chronic diseases such as cystic fibrosis . There is limited knowledge on the best modalities for diagnosis and monitoring of treatment-related pulmonary toxicity during and after childhood cancer treatment. Therefore, we conducted this systematic review to identify studies that used washout tests or lung imaging to assess pulmonary toxicity in paediatric cancer patients and survivors, and to describe reported abnormalities.
We consulted the Cochrane Handbook to conduct this systematic review and reported it according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 guidelines ( supplementary table S1 ) . Eligibility criteria We included publications on paediatric cancer patients and survivors diagnosed according to the International Classification of Childhood Cancer (ICCC-3) before 22 years of age. Treatments included chemotherapy, HSCT, chest radiotherapy and/or thoracic surgery. We only included publications if they met the inclusion criteria, reported on the predefined outcome measures and reported results separately for cancer patients and those with a benign disease. We excluded publications focusing on cancer types typically occurring in adults ( e.g. breast, pancreatic, prostate) and those involving only surgical treatment outside the chest, or those not reporting results separately for patients receiving HSCT for a benign haematological disease. Publications reporting spirometry results or imaging assessments for pulmonary infections, treatment responses or diagnosis of relapse were also excluded. Study outcomes Depending on the analysis modality for washout, the test can be conducted either as a single breath washout (SBW) or as a multiple breath washout (MBW) . During washout tests, lung resident nitrogen (N 2 ) is washed out of the lungs by breathing 100% oxygen during tidal breathing. SBW measures how effectively the lungs clear a tracer gas with one exhalation. MBW measures the removal of a tracer gas, requiring multiple breaths. Within publications that focused on washout, we summarised results for the following outcome parameters: for N 2 SBW, the outcome parameter was phase III slope (SIII). For N 2 MBW, the outcome parameters were SIII, the convection-dependent ventilation inhomogeneity index (S COND ), the diffusion- and convection-interaction-dependent ventilation inhomogeneity index (S ACIN ) and the lung clearance index (LCI). SIII reflects nitrogen washout from the alveoli (the smallest parts of the lungs); an increased SIII indicates ventilation inhomogeneity of the smallest airways . S COND assesses the conducting airways and S ACIN the acinus as the most distal regions . LCI is defined as the number of functional residual capacity lung turnovers required to reduce alveolar tracer-gas concentration ( i.e. nitrogen) to a given fraction of its starting concentration, to e.g. 1/40 (2.5%) . A higher LCI indicates increased global ventilation inhomogeneity. For lung imaging, we summarised findings for the following outcomes: any structural abnormalities, including bronchiectasis, ground-glass opacities, septal thickening, consolidations, mosaic pattern, air trapping, multifocal or diffuse hypoattenuation of lung parenchyma, bronchial dilatation, peripheral vascularity, linear opacities, bronchial wall thickening, pleural abnormalities, alveolar abnormalities, interstitial abnormalities and abnormal lung ventilation. Search strategy Upon a text analysis of key studies, we devised search strategies for pulmonary function and lung imaging. The search strategy was built around the following categories: population, cancer type and outcomes. The search strategy for washout was divided into three categories: population, cancer type and N 2 SBW and N 2 MBW outcomes. The search strategy for lung imaging was divided into five categories: population, cancer type, treatment effects (acute, subacute or chronic), lung disease and imaging modality. We executed the searches on 18 September 2023 in the following sources: MEDLINE All (Ovid), Embase (Ovid) and the Cochrane Library (Wiley). Publication dates were applied, and studies on animals and those not published in English were excluded. We started our search from 1995 onwards because there were significant milestones in cancer treatment after the 1990s and this allowed for better comparability between studies. Search results were deduplicated using Deduklick and EndNote (version 20) (Clarivate) and imported to Rayyan (www.rayyan.ai) for screening. The complete search strategies are available in supplementary tables S2 and S3 . We registered the protocol in PROSPERO (CRD42022348624) before conducting the search. Screening and data extraction Two reviewers independently screened titles and abstracts in Rayyan and then assessed full texts (CK, CS). In case of disagreement among reviewers, a third expert (JU) was involved for consensus. We employed the weighted κ statistic to assess the agreement of title and abstract screening between the independent reviewers in the selected studies , using R version 4.3.0 (www.r-project.org). This showed almost perfect agreement, with a κ of 0.95 (95% CI 0.89–1.00) for N 2 SBW and N 2 MBW and a κ of 0.94 (95% CI 0.91–0.98) for lung imaging. Two reviewers (CS, MŽ) independently extracted and cross-verified all data ( supplementary table S4 ). Cohort definition We defined the original cohort as the study population investigated in the respective publication, of which only some participants might have fulfilled the inclusion criteria of our review. We further defined from this original cohort a subset consisting only of participants who fully met the inclusion criteria of our review: the eligible cohort. We excluded publications if they did not report separate results for washout or imaging specifically for the eligible cohort. Quality assessment and risk of bias Two reviewers (CS, MŽ) independently performed quality assessment according to the risk of bias tool for prevalence studies by H oy et al . . Because the included studies did not include a comparison group, we could not use the ROBINS-I tool as planned in the protocol . The tool by H oy et al . does not provide an overall ranking according to bias. Therefore, with agreement from all co-authors, we summarised results into the three groups: low risk, moderate risk and high risk ( supplementary table S5a, b ).
We included publications on paediatric cancer patients and survivors diagnosed according to the International Classification of Childhood Cancer (ICCC-3) before 22 years of age. Treatments included chemotherapy, HSCT, chest radiotherapy and/or thoracic surgery. We only included publications if they met the inclusion criteria, reported on the predefined outcome measures and reported results separately for cancer patients and those with a benign disease. We excluded publications focusing on cancer types typically occurring in adults ( e.g. breast, pancreatic, prostate) and those involving only surgical treatment outside the chest, or those not reporting results separately for patients receiving HSCT for a benign haematological disease. Publications reporting spirometry results or imaging assessments for pulmonary infections, treatment responses or diagnosis of relapse were also excluded.
Depending on the analysis modality for washout, the test can be conducted either as a single breath washout (SBW) or as a multiple breath washout (MBW) . During washout tests, lung resident nitrogen (N 2 ) is washed out of the lungs by breathing 100% oxygen during tidal breathing. SBW measures how effectively the lungs clear a tracer gas with one exhalation. MBW measures the removal of a tracer gas, requiring multiple breaths. Within publications that focused on washout, we summarised results for the following outcome parameters: for N 2 SBW, the outcome parameter was phase III slope (SIII). For N 2 MBW, the outcome parameters were SIII, the convection-dependent ventilation inhomogeneity index (S COND ), the diffusion- and convection-interaction-dependent ventilation inhomogeneity index (S ACIN ) and the lung clearance index (LCI). SIII reflects nitrogen washout from the alveoli (the smallest parts of the lungs); an increased SIII indicates ventilation inhomogeneity of the smallest airways . S COND assesses the conducting airways and S ACIN the acinus as the most distal regions . LCI is defined as the number of functional residual capacity lung turnovers required to reduce alveolar tracer-gas concentration ( i.e. nitrogen) to a given fraction of its starting concentration, to e.g. 1/40 (2.5%) . A higher LCI indicates increased global ventilation inhomogeneity. For lung imaging, we summarised findings for the following outcomes: any structural abnormalities, including bronchiectasis, ground-glass opacities, septal thickening, consolidations, mosaic pattern, air trapping, multifocal or diffuse hypoattenuation of lung parenchyma, bronchial dilatation, peripheral vascularity, linear opacities, bronchial wall thickening, pleural abnormalities, alveolar abnormalities, interstitial abnormalities and abnormal lung ventilation.
Upon a text analysis of key studies, we devised search strategies for pulmonary function and lung imaging. The search strategy was built around the following categories: population, cancer type and outcomes. The search strategy for washout was divided into three categories: population, cancer type and N 2 SBW and N 2 MBW outcomes. The search strategy for lung imaging was divided into five categories: population, cancer type, treatment effects (acute, subacute or chronic), lung disease and imaging modality. We executed the searches on 18 September 2023 in the following sources: MEDLINE All (Ovid), Embase (Ovid) and the Cochrane Library (Wiley). Publication dates were applied, and studies on animals and those not published in English were excluded. We started our search from 1995 onwards because there were significant milestones in cancer treatment after the 1990s and this allowed for better comparability between studies. Search results were deduplicated using Deduklick and EndNote (version 20) (Clarivate) and imported to Rayyan (www.rayyan.ai) for screening. The complete search strategies are available in supplementary tables S2 and S3 . We registered the protocol in PROSPERO (CRD42022348624) before conducting the search.
Two reviewers independently screened titles and abstracts in Rayyan and then assessed full texts (CK, CS). In case of disagreement among reviewers, a third expert (JU) was involved for consensus. We employed the weighted κ statistic to assess the agreement of title and abstract screening between the independent reviewers in the selected studies , using R version 4.3.0 (www.r-project.org). This showed almost perfect agreement, with a κ of 0.95 (95% CI 0.89–1.00) for N 2 SBW and N 2 MBW and a κ of 0.94 (95% CI 0.91–0.98) for lung imaging. Two reviewers (CS, MŽ) independently extracted and cross-verified all data ( supplementary table S4 ).
We defined the original cohort as the study population investigated in the respective publication, of which only some participants might have fulfilled the inclusion criteria of our review. We further defined from this original cohort a subset consisting only of participants who fully met the inclusion criteria of our review: the eligible cohort. We excluded publications if they did not report separate results for washout or imaging specifically for the eligible cohort.
Two reviewers (CS, MŽ) independently performed quality assessment according to the risk of bias tool for prevalence studies by H oy et al . . Because the included studies did not include a comparison group, we could not use the ROBINS-I tool as planned in the protocol . The tool by H oy et al . does not provide an overall ranking according to bias. Therefore, with agreement from all co-authors, we summarised results into the three groups: low risk, moderate risk and high risk ( supplementary table S5a, b ).
Literature search In total, our search yielded 6544 records. After title and abstract screening, we performed full-text screening of 126 studies and identified 21 studies that summarised results for our outcome parameters of interest (three studies qualified for both searches). Nine of 21 studies did not adequately differentiate between our population of interest and a population with benign diseases ( supplementary tables S6 and S7 ). and show that 12 of the 21 studies provided detailed results for our population of interest and, therefore, qualified for final inclusion (two studies qualifying in the two separate searches for both modalities). All of the 12 studies were single-centre studies and from different regions: Africa (n=1), Asia (n=3), Europe (n=4) and North America (n=4). Seven studies had a retrospective design while five were prospective . Of three studies on LCI, only one also described pulmonary function pre-treatment . Of nine studies on lung imaging, only two reported pre-and post-treatment results . The number of patients ranged from three to 80 patients. Sex distribution was mentioned in all but one study . The timing of pulmonary function tests and lung imaging after cancer treatment varied largely among studies, ranging from 29 days post-treatment to 20 years post-diagnosis ( supplementary table S6 ). Identified studies on N 2 SBW and N 2 MBW Among 3353 records on N 2 SBW and N 2 MBW, we removed 411 duplicates and screened 2942. We excluded 2903 records based on title and abstract screening. Of these, three studies required additional agreement between two reviewers (CK, CS) before exclusion. We assessed 39 publications on full-text screening for eligibility. After further exclusion of 31 publications, eight studies reported on our outcomes of interest, but not all reported results separately for our population of interest (eligible cohort). Therefore, we excluded three more studies, and five studies qualified for final inclusion . Of those, three reported on N 2 MBW and two on N 2 SBW . Three included HSCT patients . All studies used the inert gas nitrogen for the washout test . Identified studies for outcome lung imaging For lung imaging, we identified 4274 records. We removed 672 duplicates and screened 3602 records. After title and abstract screening, we excluded 3515 records, of which eight needed additional agreement after discussion between two reviewers (CK, CS). We assessed 87 records with full-text screening for eligibility. After further exclusion of 71 publications, 16 studies reported on our outcome of interest ( supplementary tables S6 and S7 ), but seven studies did not report results separately for our study population of interest (eligible cohort) ( supplementary tables S6 and S7 ). Nine studies qualified for final inclusion , and none used MRI. CT was used in all lung imaging studies . Six included HSCT patients . Included studies Our final analysis included 12 studies: three studies on N 2 SBW/N 2 MBW , seven on CT and two reporting both outcomes . In total, these studies reported on 333 patients (183 for N 2 SBW/N 2 MBW, 127 for lung imaging and 23 for both). As planned, we conducted a narrative synthesis, because a meta-analysis was not considered appropriate due to the sparse data on outcomes and the heterogeneity in the study designs, particularly among the study participants. Narrative synthesis involves the systematic, qualitative integration of findings from multiple studies to identify patterns and themes. It is useful when pooling data quantitatively (meta-analysis) is not appropriate due to differences in study characteristics . summarises the characteristics of included studies. We provide an overview of the population and main outcomes in . Risk of bias The summary item on the overall risk of study bias showed moderate risk of bias in 10 studies (83%) and high risk of bias in two studies (17%). Main reasons for moderate to high risk of bias were insufficient description of the target population. The likelihood of selection bias was high across all studies ( supplementary table S5a, b ).
In total, our search yielded 6544 records. After title and abstract screening, we performed full-text screening of 126 studies and identified 21 studies that summarised results for our outcome parameters of interest (three studies qualified for both searches). Nine of 21 studies did not adequately differentiate between our population of interest and a population with benign diseases ( supplementary tables S6 and S7 ). and show that 12 of the 21 studies provided detailed results for our population of interest and, therefore, qualified for final inclusion (two studies qualifying in the two separate searches for both modalities). All of the 12 studies were single-centre studies and from different regions: Africa (n=1), Asia (n=3), Europe (n=4) and North America (n=4). Seven studies had a retrospective design while five were prospective . Of three studies on LCI, only one also described pulmonary function pre-treatment . Of nine studies on lung imaging, only two reported pre-and post-treatment results . The number of patients ranged from three to 80 patients. Sex distribution was mentioned in all but one study . The timing of pulmonary function tests and lung imaging after cancer treatment varied largely among studies, ranging from 29 days post-treatment to 20 years post-diagnosis ( supplementary table S6 ).
2 SBW and N 2 MBW Among 3353 records on N 2 SBW and N 2 MBW, we removed 411 duplicates and screened 2942. We excluded 2903 records based on title and abstract screening. Of these, three studies required additional agreement between two reviewers (CK, CS) before exclusion. We assessed 39 publications on full-text screening for eligibility. After further exclusion of 31 publications, eight studies reported on our outcomes of interest, but not all reported results separately for our population of interest (eligible cohort). Therefore, we excluded three more studies, and five studies qualified for final inclusion . Of those, three reported on N 2 MBW and two on N 2 SBW . Three included HSCT patients . All studies used the inert gas nitrogen for the washout test .
For lung imaging, we identified 4274 records. We removed 672 duplicates and screened 3602 records. After title and abstract screening, we excluded 3515 records, of which eight needed additional agreement after discussion between two reviewers (CK, CS). We assessed 87 records with full-text screening for eligibility. After further exclusion of 71 publications, 16 studies reported on our outcome of interest ( supplementary tables S6 and S7 ), but seven studies did not report results separately for our study population of interest (eligible cohort) ( supplementary tables S6 and S7 ). Nine studies qualified for final inclusion , and none used MRI. CT was used in all lung imaging studies . Six included HSCT patients .
Our final analysis included 12 studies: three studies on N 2 SBW/N 2 MBW , seven on CT and two reporting both outcomes . In total, these studies reported on 333 patients (183 for N 2 SBW/N 2 MBW, 127 for lung imaging and 23 for both). As planned, we conducted a narrative synthesis, because a meta-analysis was not considered appropriate due to the sparse data on outcomes and the heterogeneity in the study designs, particularly among the study participants. Narrative synthesis involves the systematic, qualitative integration of findings from multiple studies to identify patterns and themes. It is useful when pooling data quantitatively (meta-analysis) is not appropriate due to differences in study characteristics . summarises the characteristics of included studies. We provide an overview of the population and main outcomes in .
The summary item on the overall risk of study bias showed moderate risk of bias in 10 studies (83%) and high risk of bias in two studies (17%). Main reasons for moderate to high risk of bias were insufficient description of the target population. The likelihood of selection bias was high across all studies ( supplementary table S5a, b ).
In this systematic review we summarise studies that reported on treatment-related pulmonary toxicity using N 2 SBW, N 2 MBW and lung imaging in paediatric cancer patients and survivors under the age of 22 years. We identified a limited number of studies suggesting that washout might be as sensitive as, or more sensitive than, spirometry in detecting pulmonary toxicity . S chindera et al . compared washout with spirometry and found more cases of pulmonary dysfunction when using N 2 MBW than spirometry. P arisi et al . observed a trend for the LCI to identify the onset of pulmonary fibrosis earlier than spirometry, which is reasonable given that lung fibrosis typically begins in the peripheral, smaller airways. Two studies not including HSCT patients reported abnormal results for the SIII of the N 2 SBW washout curve in 14% and 23% of patients, respectively . Overall, N 2 SBW and N 2 MBW tests were rarely used to assess treatment-related pulmonary toxicity . To firmly conclude whether washout tests are more sensitive than spirometry in detecting cancer treatment-related pulmonary toxicity, more longitudinal, well-designed studies including lung function assessment before the start of treatment are needed. HSCT patients undergo intensive treatments and are often critically ill during treatment, thus are unable to perform the forced manoeuvres required for spirometry . In this patient population, washout tests could be easier to perform than spirometry, because only tidal breathing is required. In all three studies on N 2 MBW, HSCT patients had received either busulfan or chest radiotherapy . LCI showed abnormalities more frequently in patients who had received known pulmotoxic treatment than in those who only received suspected pulmotoxic treatment . Our systematic review showed that CT scans were the predominant imaging modality used to detect lung abnormalities after cancer treatment, and that it was mostly patients who had undergone HSCT who were investigated . CT is favoured owing to its widespread availability and short examination time, which does not require sedation, while also providing high-quality images and diagnostic accuracy in assessing the lungs. We found that acute, subacute and chronic structural abnormalities can be visualised. However, it remains unclear which imaging modality is most sensitive for detecting pulmonary dysfunction, because we only found studies using CT scans. Although there is suggested potential for MRI to assess lung abnormalities in other chronic childhood lung diseases , we found no studies that evaluated MRI with regards to childhood cancer-related pulmonary toxicity. Ultra-short echo times and zero echo time MRI sequences have also not yet been used . Because repeated scans are necessary to monitor chronic processes, MRI could, on the one hand, be a more suitable technique, avoiding ionising radiation. This may be particularly important for patients with a genetic predisposition, such as Li–Fraumeni syndrome, who have a higher risk for cancer development , and where repeated imaging with radiation exposure may increase the risk for cancer development. On the other hand, although MRI is radiation-free, its longer image acquisition times may require anaesthesia in younger children. However, given the need for repeated examinations and follow-ups to monitor pulmonary toxicity, it is crucial to consider alternatives for children, whose organs are still developing and who may face potentially serious long-term effects from ionising radiation from CT. While washout tests do not provide information on structural lung abnormalities, the test provides important information on small airway damage. Washout tests can be performed without sedation, even in very young or severely ill children. In clinical practice, however, washout tests are not frequently used because only a few centres have the testing equipment. Furthermore, the test requires a high level of expertise for correct test performance and interpretation of the findings, which may explain why we only identified very few studies using the test in childhood cancer survivors. Future studies need to further address the usefulness of washout tests for detection of early pulmonary dysfunction in this patient population. This review is the first to systematically assess the use of washout and lung imaging as diagnostic modalities to detect pulmonary toxicity after paediatric cancer treatment. A recently published scoping review investigated MBW in adult and paediatric patients after HSCT . Their main finding for MBW demonstrated post-HSCT peripheral airway abnormalities. They suggested MBW could assess graft- versus -host disease earlier than spirometry. O tth et al . investigated pulmonary dysfunction in childhood, adolescent and young adult cancer survivors with the aim of developing lung surveillance guidelines based on evidence reviewed by the International Late Effects of Childhood Cancer Guideline Harmonization Group. Although these young cancer survivors treated with HSCT, chest radiotherapy and thoracic surgery were found to be at risk for pulmonary dysfunction, the expert panel did not recommend routine pulmonary function tests for asymptomatic survivors owing to the lack of evidence and effective interventions for incipient lung disease. The main limitation of the identified literature in this systematic review is the heterogeneity of study populations, cancer diagnoses and treatment regimens, with varying intensities and exposures to pulmotoxic treatment modalities. This results in poor inter-study comparability. Oncology patients and those undergoing HSCT follow different treatment regimens based on their risk stratification, even when having the same diseases. For example, solid tumours (such as neuroblastoma) have curative treatment regimens ranging from surgery alone to autologous stem cell transplantation . To compare these patients, studies need to report on the exact cumulative treatment doses at the time of pulmonary function or lung imaging assessment, yet only one of the 12 selected studies recorded this . Of the other studies, three reported cumulative radiation doses , including D e et al. , for which all involved patients received bleomycin but treatment varied in the additional chemotherapy agents, doses, radiotherapy and surgical interventions they received, resulting in a heterogeneous cohort. Another study aimed to address this heterogeneity limitation by considering only patients with acute lymphoblastic leukaemia who underwent similar treatments, resulting in a more homogeneous sample . But even a subset of acute lymphoblastic leukaemia patients does not have a homogeneous treatment regimen, making analysis and comparison of treatment outcomes difficult . Furthermore, the limited reporting of the eligible cohort from the larger original cohort could risk overgeneralisation of the results . Although our review specifically and exclusively focused on studies examining the effects of local and systemic cancer treatments, it is not entirely known whether factors such as pulmonary infections, relapses with higher cumulative cancer treatments or even secondary malignancies could influence outcomes over the typically extended follow-up period. This uncertainty is a critical limitation when interpreting the results. Several ongoing studies are investigating pulmonary outcomes in paediatric cancer patients and survivors. Large studies include the St. Jude Lifetime of Cohort of Adults Surviving Childhood Cancer study (SJLIFE) for paediatric cancer survivors (ClinicalTrials.gov: NCT00760656) and the North American Childhood Cancer Survivor Study (CCSS) (ClinicalTrials.gov: NCT01120353), conducted since 1994 and 2007, respectively. The SJLIFE cohort study includes measures for pulmonary function testing, specifically assessing forced vital capacity, forced expiratory volume and peak expiratory flow . The CCSS measures pulmonary outcomes through self-reported respiratory symptom questionnaires covering asthma, chronic cough, emphysema, lung fibrosis, oxygen need and recurrent pneumonia, but clinical tests such as pulmonary function tests are not assessed. CCSS researchers have found that paediatric cancer survivors have a significantly higher cumulative incidence of conditions like chronic cough, oxygen need, lung fibrosis and recurrent pneumonia compared to the sibling control group . Neither cohort carries out washout tests. A third, new cohort is the multicentre cohort study Early Pulmonary Dysfunction in Childhood Cancer Patients (SWISS-Pearl) (ClinicalTrails.gov: NCT05427136), which started recruitment in 2022. It assesses lung function before and during cancer treatment using MRI. The Swiss Childhood Cancer Survivor Study (SCCSS) FollowUp–Pulmo cohort (ClinicalTrials.gov: NCT04732273) assesses lung function in survivors during routine clinical visits using N 2 MBW and conventional lung function tests (spirometry, body plethysmography and diffusing capacity of the lungs for carbon monoxide). These last two cohorts have not yet published their findings.
We identified a limited number of studies that report on functional and structural lung abnormalities in paediatric cancer patients and survivors, mainly after HSCT. Timely detection of potentially reversible pulmonary dysfunction during and after cancer treatment is necessary to initiate therapies and supportive care early. N 2 SBW, N 2 MBW and lung imaging can detect subtle pulmonary damage by assessing peripheral airway abnormalities or structural lung pathologies. However, the value for early detection of pulmonary dysfunction in paediatric cancer patients and survivors, particularly outside of HSCT patient groups, remains mostly unclear. Standardised, longitudinal and multicentre studies assessing different monitoring modalities are needed to improve our understanding of their value for the early assessment of pulmonary dysfunction in paediatric cancer patients and survivors. Points for clinical practice Washout: Lung clearance index z-scores show a positive correlation with time since treatment, indicating worsening pulmonary function over time . Washout tests may detect early signs of pulmonary fibrosis and are probably more sensitive than spirometry . Washout tests can be performed by children who are unable to complete the forced expiratory manoeuvres required for spirometry, such as very sick or very young patients, making washout a suitable option for these populations . Lung imaging: Radiological signs of pulmonary toxicity are common in patients after carmustine- and cyclophosphamide-based regimens, as well as chest radiation . Following haematopoietic stem cell transplantation, pleuroparenchymal fibroelastosis progresses, with increasing structural abnormalities over time . Computed tomography (CT) is the most commonly used imaging modality for the lungs . Modern paediatric CT scans use very low radiation doses, which minimises the risk of radiation-induced toxicity. However, the potential harm from ionising radiation should be weighed against its benefits in detecting pulmonary toxicity. Questions for future research based on the authors’ opinion Washout: What is the dose–response relationship between exposure to different cancer treatments and the severity of treatment-related pulmonary toxicity outcomes assessed with washout tests? Can longitudinal monitoring of the lung clearance index in paediatric cancer patients and survivors improve early detection of pulmonary dysfunction and ultimately improve their long-term respiratory outcomes? When is the optimal time to screen with washout for lung function impairment after childhood cancer treatment? Lung imaging: Is lung computed tomography (CT) more sensitive than other imaging modalities, e.g. lung magnetic resonance imaging, in detecting early pulmonary abnormalities after childhood cancer treatment? How does lung imaging compare to pulmonary function tests in detecting early pulmonary toxicity? Would there be a need for a comparison group with CT for other lung imaging modalities? Is lung imaging, compared with pulmonary function tests, beneficial in detecting early pulmonary toxicity?
Washout: Lung clearance index z-scores show a positive correlation with time since treatment, indicating worsening pulmonary function over time . Washout tests may detect early signs of pulmonary fibrosis and are probably more sensitive than spirometry . Washout tests can be performed by children who are unable to complete the forced expiratory manoeuvres required for spirometry, such as very sick or very young patients, making washout a suitable option for these populations . Lung imaging: Radiological signs of pulmonary toxicity are common in patients after carmustine- and cyclophosphamide-based regimens, as well as chest radiation . Following haematopoietic stem cell transplantation, pleuroparenchymal fibroelastosis progresses, with increasing structural abnormalities over time . Computed tomography (CT) is the most commonly used imaging modality for the lungs . Modern paediatric CT scans use very low radiation doses, which minimises the risk of radiation-induced toxicity. However, the potential harm from ionising radiation should be weighed against its benefits in detecting pulmonary toxicity.
Washout: What is the dose–response relationship between exposure to different cancer treatments and the severity of treatment-related pulmonary toxicity outcomes assessed with washout tests? Can longitudinal monitoring of the lung clearance index in paediatric cancer patients and survivors improve early detection of pulmonary dysfunction and ultimately improve their long-term respiratory outcomes? When is the optimal time to screen with washout for lung function impairment after childhood cancer treatment? Lung imaging: Is lung computed tomography (CT) more sensitive than other imaging modalities, e.g. lung magnetic resonance imaging, in detecting early pulmonary abnormalities after childhood cancer treatment? How does lung imaging compare to pulmonary function tests in detecting early pulmonary toxicity? Would there be a need for a comparison group with CT for other lung imaging modalities? Is lung imaging, compared with pulmonary function tests, beneficial in detecting early pulmonary toxicity?
10.1183/16000617.0178-2024.Supp1 Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author. Supplementary material ERR-0178-2024.SUPPLEMENT
|
Development and Implementation of a Clinician-Facing Prognostic Communication Tool for Patients With COVID-19 and Critical Illness | a3bd921b-a5fa-4467-b767-eecdc80bb3f9 | 7207118 | Health Communication[mh] | Patients and families with serious illness rely upon health care teams to provide accurate information about prognosis. Skillfully delivered prognostic information prompts patients and families to imagine how their lived experience will change on each potential treatment path, thereby empowering them to fully participate in shared decision making about the path that aligns best with the patient's values and goals. , Unfortunately, clinicians face formidable challenges when attempting to deliver clear prognostic information to the families of critically ill patients with COVID-19. Existing data about survival rates are rapidly evolving and difficult to interpret with confidence because of variability in care settings and incomplete patient follow-up in most studies. , , , , , , , The limited experience of any given clinician with long-term outcomes of patients with this novel disease inherently limits the value of experiential prognostication. Moreover, important data about quality-of-life outcomes will take months or years to develop due to the lengthy convalescent period for most critically ill patients with COVID-19. Given the limited data available about COVID-19 and the heightened emotional challenges of caring for patients in a pandemic, clinicians may be particularly vulnerable to cognitive biases during the process of prognostication. Early in our institution's experience with the COVID-19 pandemic, our palliative care consult team noticed several recurring themes of cognitive bias during interdisciplinary team discussions. Recognizing the importance of providing the most objective and consistent prognostic information possible to patients and families, we created a COVID-19 Prognostication Tool designed to accomplish four objectives. First, the tool collates the latest peer-reviewed prognostic information about critically ill patients with COVID-19 into a concise, easily accessible, up-to-date prognostication guide. Second, the tool guides clinicians through a careful process to mitigate the effects of cognitive biases on their ability to communicate about prognosis with patients and families. Third, the tool prompts clinicians to translate population-based statistical information into best-case, worst-case, and most likely scenarios for a given patient. Fourth, the tool encourages providers to seek information about patient values that can inform clinician recommendations for medically appropriate, value-concordant care. This report describes the development and implementation of a point-of-care COVID-19 Prognostic Tool to guide best practices of relaying prognostic information to the families of critically ill patients with COVID-19 and makes this tool available to others for adaptation and implementation.
We identified three cognitive biases , that we observed impacting clinical team discussions of prognosis for critically ill patients with COVID-19 at our institution: anchoring bias, availability bias, and false consensus bias. The Prognostication Tool was designed to explicitly address these three biases described here. Anchoring Bias Early in the pandemic, “Crisis Standards of Care” was a frequent topic of discussion as our institution prepared for a predicted surge of patients with COVID-19. Clinicians contemplated the frightening possibility of reaching a crisis state where resource scarcity would prevent us from offering advanced life support to chronically ill, elderly patients with the poorest prognoses. The impact of these emotionally charged discussions was significant. Even when staffing, ventilators, and other resources remained at objectively adequate levels, providers often continued to subconsciously anchor on a “Crisis Standards of Care” mindset, proposing limits on aggressive treatment modalities due to concerns about future resource scarcity rather than actual scarcity or patient values. Availability Bias Clinicians who have recently cared for a dying patient with COVID-19 can grow more pessimistic about outcomes for all critically ill patients with COVID-19 and may be more likely to overestimate subsequent individual patients' mortality risk. At times, a feeling of therapeutic nihilism seemed predominant on our health care teams, prompting moral distress among physicians, nurses, and others. Some family members also perceived disproportionate provider pessimism and warned against the perpetuation of self-fulfilling prophecies. False Consensus Bias , , Providers with strong personal beliefs about minimum acceptable quality of life are at risk of making potentially inaccurate assumptions that their patients have similar values, especially in times of stress. During the COVID-19 pandemic, family visitation restrictions have heightened the risk of false consensus bias because these infection control measures inherently reduce opportunities for family members to interface with the medical team and initiate discussions about their loved one's values and goals. Patients and families with limited English proficiency, disproportionately affected by the COVID-19 pandemic, face an additional barrier of needing an interpreter to communicate with the health care team. Owing to these communication challenges, clinicians in the COVID-19 pandemic are less likely to learn about their patient's values and may subconsciously fill in knowledge gaps with inaccurate assumptions based on their own personal values.
Early in the pandemic, “Crisis Standards of Care” was a frequent topic of discussion as our institution prepared for a predicted surge of patients with COVID-19. Clinicians contemplated the frightening possibility of reaching a crisis state where resource scarcity would prevent us from offering advanced life support to chronically ill, elderly patients with the poorest prognoses. The impact of these emotionally charged discussions was significant. Even when staffing, ventilators, and other resources remained at objectively adequate levels, providers often continued to subconsciously anchor on a “Crisis Standards of Care” mindset, proposing limits on aggressive treatment modalities due to concerns about future resource scarcity rather than actual scarcity or patient values.
Clinicians who have recently cared for a dying patient with COVID-19 can grow more pessimistic about outcomes for all critically ill patients with COVID-19 and may be more likely to overestimate subsequent individual patients' mortality risk. At times, a feeling of therapeutic nihilism seemed predominant on our health care teams, prompting moral distress among physicians, nurses, and others. Some family members also perceived disproportionate provider pessimism and warned against the perpetuation of self-fulfilling prophecies.
, , Providers with strong personal beliefs about minimum acceptable quality of life are at risk of making potentially inaccurate assumptions that their patients have similar values, especially in times of stress. During the COVID-19 pandemic, family visitation restrictions have heightened the risk of false consensus bias because these infection control measures inherently reduce opportunities for family members to interface with the medical team and initiate discussions about their loved one's values and goals. Patients and families with limited English proficiency, disproportionately affected by the COVID-19 pandemic, face an additional barrier of needing an interpreter to communicate with the health care team. Owing to these communication challenges, clinicians in the COVID-19 pandemic are less likely to learn about their patient's values and may subconsciously fill in knowledge gaps with inaccurate assumptions based on their own personal values.
Our COVID-19 Prognostication Tool was developed as a point-of-care guide to help frontline clinicians respond to the cognitive challenges of prognosticating during an evolving pandemic. The tool is a concise, open-source document that can be viewed online. The current tool at the time of publication is captured in this article, but the open-source tool will be updated as new data emerge. provides an explicit reminder that resource scarcity should only dictate care decisions if the health care system is operating under extraordinary circumstances that require Crisis Standards of Care. Such decisions should be made on a regional basis and not for one hospital in isolation. In all other circumstances, our goal should be to provide medically appropriate care that aligns with the patient's values . This reminder was placed at the beginning of the tool to explicitly name the risk of anchoring on Crisis Standards of Care and provide clinicians with the more appropriate, alternative reference point of values-based care. and provide a visual summary of outcome data from existing cohort studies of critically ill patients with COVID-19. , , , , , , , Studies were identified using a daily literature search that included (“COVID-19” OR “SARS-CoV-2”) AND (“critical care” OR “ICU” OR “critically ill”) AND (“outcome” OR “mortality” OR “survival”). Outcome data for critically ill patients with COVID-19 were reviewed independently by two authors, one trained in critical care, with a focus on clinical predictors of mortality. Based on published outcome data, we identified two broad categories of patients with contrasting prognostic trajectories useful to clinicians: 1) patients with COVID-19 receiving mechanical ventilation with isolated lung involvement (generally <60 to 70 years old) and 2) patients with COVID-19 receiving mechanical ventilation with multiorgan involvement and/or patients who were >70 years old. For each group, we created a visual summary of the approximate likelihood of successful extubation, death in the hospital, or prolonged intubation of more than two weeks with uncertain final outcome . Reviewers reached consensus about the two categories of patients and their respective prognoses through discussion of the literature. The tool is updated weekly to reflect newly published results. also contains clinical descriptions of best-case and worst-case scenarios for elderly patients with COVID-19 receiving mechanical ventilation. These were based on both published data and the clinical experience at our institution. The best-case scenario for patients who require a prolonged course of intubation describes a typical recovery period after severe acute respiratory distress syndrome, emphasizing that recovery is often prolonged, on the order of months, with a new baseline that will likely include new functional deficits. This descriptive clinical scenario is provided in conjunction with estimated survival rates to provide a tangible reminder to clinicians of how a potential recovery might look. While recognizing that the likelihood of recovery may be small in some cases, we explicitly point out the possibility of recovery as a counterbalance to availability bias for clinicians who have recently cared for a dying patient and may be susceptible to inappropriate therapeutic nihilism. is designed to prompt clinicians to challenge their availability bias and false consensus bias. This figure names the cognitive and emotional burden involved in caring for a large volume of seriously ill or dying patients and introduces a framework intended to mitigate the negative effect that demoralization and cognitive biases may have on a clinician's ability to prognosticate for an individual patient. First, the framework directs providers to imagine best-case, worst-case, and most likely scenarios for the individual patient. Second, clinicians are asked to consider whether each treatment option is likely to “work” for the patient on a physiologic level. These two steps attempt to enhance provider awareness of inappropriate therapeutic nihilism by prompting providers to go through the cognitive step of considering the likelihood of benefit from each treatment option. Third, the framework attempts to challenge false consensus bias by explicitly directing the provider to ask for the patient and family's perspective on minimum acceptable quality of life, rather than relying on assumptions that may have been made by the clinical team. The point-of-care COVID-19 Prognostication Tool has been disseminated to frontline clinicians at our institution, as well as a local community hospital, and will be updated weekly to reflect newly published outcome data. Clinicians at our institution are invited to use the tool as a source of information about major prognostic trajectories for this novel disease, as well as a reminder of the need to continuously recalibrate our clinical impressions as new peer-reviewed evidence emerges. Qualitative feedback on the tool has been positive, but we were not able to conduct a rigorous evaluation in the context of our pandemic response.
Our COVID-19 Prognostication Tool has several important limitations. First, the visual summary of patient outcomes is based on limited data from early studies and is likely to change as the pandemic evolves. Moreover, patient outcome data vary widely across studies due to length of study follow-up and regional differences in resource availability, treatment protocols, and approaches to withdrawal of life-sustaining treatment. Our team strives to update our visual summary of patient outcomes at least weekly with a concerted effort to capture pervasive, big-picture trends in survival. However, our interpretation of the data inevitably involves some degree of subjectivity. Second, this prognostication tool can only be useful when it is thoughtfully applied to the patient population and clinical circumstances for which it is intended: critically ill patients with COVID-19. Third, up-to-date and accurate prognostic information is only one step in effective prognostic communication with patients and family members. Supportive and effective communication of prognostic information is an important skill with multiple components, many of which are not addressed by this tool. , , Fourth, our prognostication tool focuses on three cognitive biases that we observed at our institution but may not be universally applicable to all centers; reflection prompts may need to be adapted depending on the cognitive biases that are prevalent within each institutional culture. Finally, our evaluation of this tool is subjective and preliminary. Our goal was to share the tool quickly for others to adapt, implement, and evaluate in the context of this novel pandemic. Additional evaluation will be important to understand and improve the effectiveness of the tool.
We have developed and implemented a point-of-care prognostic communication tool for clinicians caring for critically ill patients with COVID-19. Although this tool will need to be updated as additional evidence emerges, we present the tool and its development as a model of one approach to promote consistent and high-quality prognostic communication across a health care system. Our hope is that the tool will help clinicians develop an approach to communication about prognosis that is practical and patient- and family-centered. The best-case, worst-case, most likely scenario approach supports this tool by prompting clinicians to be objective and descriptive about likely clinical trajectories, providing patients and families with the information they need to imagine the implications of COVID-19–related critical illness and participate in informed decisions about values-based care. Studies to evaluate the utility of prognostic communication tools like this are needed.
|
Spatial accuracy of dose delivery significantly impacts the planning target volume margin in linear accelerator-based intracranial stereotactic radiosurgery | 84bb4f8e-5747-4b8c-b275-0286191e137a | 11775166 | Surgical Procedures, Operative[mh] | Stereotactic radiosurgery (SRS) and stereotactic radiotherapy (SRT) using a medical linear accelerator (linac) have largely replaced whole-brain irradiation as a treatment for multiple brain metastases – . For such applications, SRS and SRT commonly use non-coplanar, intensity-modulated irradiation , . The targets are often smaller than a few centimeters, and treatment is planned with a small planning target volume (PTV) margin of 0–2 mm – . This margin size is also recommended by the working group on stereotactic radiotherapy of the German Society of Radiation Oncology . These facts highlight the need for radiation delivery with submillimeter accuracy. Minimizing irradiation positional errors relative to the imaging isocenter is therefore becoming increasingly important. Variations in linac dose delivery accuracy may also affect clinical outcomes of intracranial SRS/SRT. Several guidelines for SRS require a geometric accuracy within 1 mm in an end-to-end (E2E) test – . The E2E test includes the accuracy of contouring and registration . Thus, the guidelines require the delivery accuracy of the linac to be less than 1 mm. van Herk et al. designed a method to determine which uncertainties and what magnitudes of those uncertainties must be factored into the margin , . However, using systematic and random errors in patient positioning is common in clinical practices, and users rarely use the three-dimensional (3D) dose delivery accuracy of the linac . In contrast, Takakura et al. suggested a method for estimating overall geometric uncertainty from uncertainties in both the patient position and linac dose delivery . Moreover, Zhang et al. suggested a method to derive the optimal PTV margin for intracranial SRS by taking into account the distance between the imaging and radiation isocenters and the residual setup error after image guidance , . If the impact of the 3D dose delivery accuracy on the margin is high, the approach proposed by Zhang et al. and Takakura et al. is considered ideal. To correctly derive the margin, the 3D dose delivery accuracy of the linac must be measured with low measurement uncertainties. To appropriately determine the margin for intracranial SRS/SRT and provide treatment with sufficient target coverage, we should directly assess and manage the 3D radiation delivery accuracy relative to the imaging isocenter by performing quality assurance (QA) procedures with minimal measurement uncertainties. Conventionally, two-dimensional (2D) planar images are obtained for each beam in QA tests, and the mechanical isocenter with some setup uncertainties is commonly used as the reference, especially for the Winston–Lutz (WL) test , , . Thus, these tests include the measurement uncertainties derived from the mechanical isocenter to obtain 3D information from the 2D images. Uncertainties in a QA test may therefore contribute to the PTV margin size because the spatial dose delivery accuracy of linacs affects geometric uncertainties in the GTV. Recently, a new 3D starshot (3D-SS) test using a polymer gel dosimeter has been reported to solve the above-mentioned problems in verifying geometric accuracy – . This approach allows a direct, comprehensive 3D evaluation of spatial radiation accuracy relative to the imaging isocenter . The trajectories of radiation beams through the dosimeter are visualized using kilovoltage cone-beam computed tomography (kV-CBCT), and the imaging isocenter is thus obtained from the resulting kV-CBCT image. Human errors derived from QA phantom setup and laser indication adjustment do not affect the measurement accuracy. The 3D-SS test may be superior to the conventional QA methods because it eliminates uncertainties associated with those methods, and the measurement uncertainty of the 3D-SS test has been found to be 0.2–0.3 mm . The geometric uncertainties in patient setup and dose delivery may vary among institutions and manufacturers. The 3D dose delivery accuracy should be included in the PTV margin size if the uncertainties significantly affect the margin. This 3D dose delivery accuracy should be measured with low uncertainty, because the guidelines have recommended that each institution aim to achieve a PTV margin of 2 mm in intracranial SRS . In contrast to conventional QA tests of spatial dose delivery, which may over- or underestimate the margin, 3D-SS analysis is effective for this purpose because it provides such spatial information free from measurement uncertainties associated with the setup. However, only the methodology has so far been proposed, and no information is available on variations in 3D dose delivery accuracy among commercially available linacs. While there are numerous studies focusing on patient setup accuracy – , there is a notable lack of research addressing linac precision, particularly with respect to differences between manufacturers and individual units. Furthermore, it remains unclear whether the PTV margin obtained for linac-based intracranial SRS via 3D-SS analysis will satisfy the recommended 2 mm limit . To correctly evaluate the impact of the spatial dose delivery accuracy of linacs on the PTV margin in intracranial SRS, we performed a multi-institutional study using linacs from two vendors using the 3D-SS test. We also developed a method for calculating the margin for the 3D-SS test. Furthermore, on the basis of the 3D-SS test results, we investigated whether the PTV margin could be constrained within 2 mm by implementing advanced image guidance (real-time video-based 3D optical surface imaging and 2D kV stereoscopic imaging) and rigorous linac quality management.
Institution selection Twenty-two institutions that performed intracranial stereotactic radiotherapy in clinical practices participated in this study. All institutions used a medical C-arm linac attached to an on-board imager enabling kV-CBCT. Seven institutions used a TrueBeam system (Varian Medical Systems, Inc., Palo Alto, CA, USA), four used a TrueBeamSTx system (Varian Medical Systems), four used a VersaHD system (Elekta AB, Stockholm, Sweden), three used an Infinity system (Elekta AB), and four used a Synergy system (Elekta AB). We verified that a geometric QA test was regularly performed at all facilities in accordance with guidelines published by task groups 142 and 198 of the American Association of Physicists in Medicine , . Assessment of 3D linac accuracy using the 3D-SS test The 3D-SS test is robust with low measurement uncertainties for evaluating 3D dose delivery accuracy . All institutions followed the 3D-SS test procedure described below. In total, 22 jars each containing an X-ray computed tomography (CT)-based polymer gel dosimeter called dGEL™ (Triangle Products Co., Ltd., Kashiwa, Japan) were used in this study . Treatment was planned using the treatment planning system installed at each institution. A common dGEL™ CT dataset, acquired using an Aquilion ONE system (Canon Medical Systems, Otawara, Japan), was sent to all institutions. A 10 MV flattening filter photon beam or a 6 or 10 MV flattening filter-free photon beam was used. All the Varian and Elekta systems used a 6 MV flattening filter photon beam to obtain the plot of the distance between the projected isocenter and the nominal isocenter for corrections of the mechanical errors of each CBCT system. Symmetric fields with a minimum size of 0.5 × 0.5 cm² for Varian systems and 0.6 × 0.6 cm² for Elekta systems were used. The field sizes of the Varian and Elekta systems were slightly different. Theoretically, these differences in field size do not affect the 3D-SS test results, because the beam axes are identified from the dose distributions by maximizing the contrast-to-noise ratio, with distance from the axis weighted appropriately during the optimization process , . Table shows gantry, couch, and collimator angle settings for radiation beams based on the study by Pant et al. A machine output between 6000 and 8000 monitor units was determined for each beam to create an isodose line of 20 Gy in the beam path (Fig. A). A dose of 20 Gy was selected on the basis of our preliminary study , which showed that this dose resulted in a sufficient contrast-to-noise ratio for each beam to perform the analysis using our in-house software. The software uses algorithms that determine a 3D vector with a point in the CBCT coordinate system that defines the beam axis by maximizing the contrast-to-noise ratio . Each jar including the gel dosimeter was sent to its respective institution. The jar was placed with the center of its sensitive volume located approximately at the mechanical isocenter indicated by the room laser. The gel dosimeter was placed on the couch and secured in place using adhesive tape to prevent movement. A pre-irradiation kV-CBCT image was obtained using the image acquisition and reconstruction parameters in Table . Subsequently, the gel dosimeter was irradiated with seven narrow beams using gantry and couch rotations (Fig. B), and post-irradiated kV-CBCT images were acquired using the same parameters as above (Fig. C). For analysis, the image datasets of pre-irradiated and post-irradiated gel dosimeters were imported into our in-house program . The analysis quantified the minimum distance ( [12pt]{minimal}
$$\:{d}_{}$$ ) between each radiation beam and the origin of the imaging coordinate system (imaging isocenter, [12pt]{minimal}
$$\:$$ ), the distance ( [12pt]{minimal}
$$\:{d}_{-}$$ ) between the radiation isocenter ( [12pt]{minimal}
$$\:$$ ) and the [12pt]{minimal}
$$\:$$ , and the radius ( [12pt]{minimal}
$$\:r$$ ) of the smallest sphere that intersects all beams . The values of [12pt]{minimal}
$$\:{d}_{}$$ , [12pt]{minimal}
$$\:{d}_{-}$$ , and [12pt]{minimal}
$$\:r$$ were obtained at all institutions. An inter-manufacturer comparison was made for these values. For [12pt]{minimal}
$$\:{d}_{}$$ and [12pt]{minimal}
$$\:{d}_{-}$$ , the comparison was made for each direction of the image (DICOM) reference coordinate system and for the resultant vector. Impact of 3D linac accuracy on SRS We assessed the impact of the spatial dose delivery accuracy on the PTV margins in intracranial SRS. PTV margins were derived using the 3D-SS analysis results. Zhang et al. proposed a method to calculate the optimal PTV margin for single-fraction intracranial SRS that incorporates the linac spatial dose delivery error and the residual patient setup error after image guidance , . The method provides an anisotropic margin for intracranial SRS that ensures with 95% probability that the clinical target volume receives the prescribed dose. In this study, we developed the method by inputting the 3D dose delivery accuracy derived from the 3D-SS test as the linac spatial dose delivery error, building on a previously reported study . After calculating the optimal PTV margins, inter-institution and inter-manufacturer comparisons were made. The formulas are 1 [12pt]{minimal}
$$\:{C}_{i}={W}_{0i}+{b}_{1}({W}_{0i}){\:}_{i}+{b}_{2}({W}_{0i}){\:}_{i}^{2},$$ 2 [12pt]{minimal}
$${b}_{1}({W}_{0i})={2.331-1.425}W_{0i}+2.296{W}_{0i}^{2}-1.539{W}_{0i}^{3}+0.374{W}_{0i}^{4},$$ 3 [12pt]{minimal}
$${b}_{2}({W}_{0i})={0.434W}_{0i}-0.917{W}_{0i}^{2}+0.676{W}_{0i}^{3}-0.171{W}_{0i}^{4},$$ where [12pt]{minimal}
$$\:{C}_{i}$$ represents the PTV margin for the i = X , Y , and Z directions in the DICOM reference coordinate system, which is correspond to the patient right-left, patient superior-inferior, and patient anterior-posterior directions, respectively; [12pt]{minimal}
$$\:{W}_{0i}={V}_{s0i}+{V}_{r0i}$$ , where [12pt]{minimal}
$$\:{V}_{s0i}$$ represents the systematic error in the imaging isocenter and radiation isocenter along each axis; [12pt]{minimal}
$$\:{V}_{r0i}$$ denotes the average residual error in patient setup after image guidance; and [12pt]{minimal}
$$\:{\:}_{i}$$ is a function of the standard deviation of the residual error in patient setup ( [12pt]{minimal}
$$\:{SD}_{,i}$$ ). In this study, we added fluctuations in the radiation beam position (random error in the radiation isocenter) to [12pt]{minimal}
$$\:{\:}_{i}$$ : 4 [12pt]{minimal}
$$\:{\:}_{i}=_{,i}^{2}+{(0.683\:r)}^{2}},$$ where the term [12pt]{minimal}
$$\:0.683\:r$$ corresponds to the standard deviation of the radiation isocenter error, which was assumed to follow a normal distribution. In this study, [12pt]{minimal}
$$\:{V}_{s0i}$$ was directly set equal to [12pt]{minimal}
$$\:{d}_{-}$$ , which includes the influence of couch rotation. The values of [12pt]{minimal}
$$\:{V}_{r0i}$$ and [12pt]{minimal}
$$\:{SD}_{,i}$$ were taken from the studies of Zhang et al. and Ong et al. and are shown in Supplementary Table A1. The values of the former and the latter were defined as the larger and smaller residual patient setup errors. Impact of QA procedures on SRS accuracy We evaluated the impact of implementing advanced image guidance systems for patient positioning and rigorous geometric QA of the linac. This study assumed that advanced image guidance systems, which include widely installed video-based 3D optical surface imaging and 2D kV stereoscopic imaging, were implemented. During non-coplanar irradiation in clinical settings, we also assumed that intra-fraction patient motion caused by treatment couch shifts during couch rotation could be monitored and corrected by these imaging modalities. For example, we hypothesized that current surface-guided radiation therapy systems could detect a couch walk-out of < 1 mm . Furthermore, the value of [12pt]{minimal}
$$\:{d}_{-}$$ was treated as a systematic error for linacs that could be minimized by a geometric calibration test performed by the user. These assumptions form the basis for the further analysis conducted in this study. Therefore, this study assumed that in clinical practice, the user’s efforts can reduce the impact of these two factors on geometric accuracy to clinically negligible values. Based on this assumption, the PTV margin was recalculated by replicating the situation using the following procedure: 3D dose delivery accuracy was recalculated using only the co-planar beams, and the new values of [12pt]{minimal}
$$\:{d}_{-}$$ and [12pt]{minimal}
$$\:r$$ were measured. Subsequently, [12pt]{minimal}
$$\:{d}_{-}$$ was set to zero to account for the error contribution of the geometric calibration of the linac. Finally, the PTV margin was recalculated using Eqs. – and the above value of [12pt]{minimal}
$$\:r$$ . Statistical analysis Statistical analyses were performed to examine inter-manufacturer differences in various parameters. The Kruskal–Wallis test and Mann–Whitney U test were used to compare the results from more than two institutions and vendors, respectively. The Steel–Dwass post hoc test was used to make pair-wise comparisons. All statistical analyses were conducted using EZR (version 1.62, Saitama Medical Center, Jichi Medical University, Saitama, Japan). A value of p < 0.05 was considered statistically significant.
Twenty-two institutions that performed intracranial stereotactic radiotherapy in clinical practices participated in this study. All institutions used a medical C-arm linac attached to an on-board imager enabling kV-CBCT. Seven institutions used a TrueBeam system (Varian Medical Systems, Inc., Palo Alto, CA, USA), four used a TrueBeamSTx system (Varian Medical Systems), four used a VersaHD system (Elekta AB, Stockholm, Sweden), three used an Infinity system (Elekta AB), and four used a Synergy system (Elekta AB). We verified that a geometric QA test was regularly performed at all facilities in accordance with guidelines published by task groups 142 and 198 of the American Association of Physicists in Medicine , .
The 3D-SS test is robust with low measurement uncertainties for evaluating 3D dose delivery accuracy . All institutions followed the 3D-SS test procedure described below. In total, 22 jars each containing an X-ray computed tomography (CT)-based polymer gel dosimeter called dGEL™ (Triangle Products Co., Ltd., Kashiwa, Japan) were used in this study . Treatment was planned using the treatment planning system installed at each institution. A common dGEL™ CT dataset, acquired using an Aquilion ONE system (Canon Medical Systems, Otawara, Japan), was sent to all institutions. A 10 MV flattening filter photon beam or a 6 or 10 MV flattening filter-free photon beam was used. All the Varian and Elekta systems used a 6 MV flattening filter photon beam to obtain the plot of the distance between the projected isocenter and the nominal isocenter for corrections of the mechanical errors of each CBCT system. Symmetric fields with a minimum size of 0.5 × 0.5 cm² for Varian systems and 0.6 × 0.6 cm² for Elekta systems were used. The field sizes of the Varian and Elekta systems were slightly different. Theoretically, these differences in field size do not affect the 3D-SS test results, because the beam axes are identified from the dose distributions by maximizing the contrast-to-noise ratio, with distance from the axis weighted appropriately during the optimization process , . Table shows gantry, couch, and collimator angle settings for radiation beams based on the study by Pant et al. A machine output between 6000 and 8000 monitor units was determined for each beam to create an isodose line of 20 Gy in the beam path (Fig. A). A dose of 20 Gy was selected on the basis of our preliminary study , which showed that this dose resulted in a sufficient contrast-to-noise ratio for each beam to perform the analysis using our in-house software. The software uses algorithms that determine a 3D vector with a point in the CBCT coordinate system that defines the beam axis by maximizing the contrast-to-noise ratio . Each jar including the gel dosimeter was sent to its respective institution. The jar was placed with the center of its sensitive volume located approximately at the mechanical isocenter indicated by the room laser. The gel dosimeter was placed on the couch and secured in place using adhesive tape to prevent movement. A pre-irradiation kV-CBCT image was obtained using the image acquisition and reconstruction parameters in Table . Subsequently, the gel dosimeter was irradiated with seven narrow beams using gantry and couch rotations (Fig. B), and post-irradiated kV-CBCT images were acquired using the same parameters as above (Fig. C). For analysis, the image datasets of pre-irradiated and post-irradiated gel dosimeters were imported into our in-house program . The analysis quantified the minimum distance ( [12pt]{minimal}
$$\:{d}_{}$$ ) between each radiation beam and the origin of the imaging coordinate system (imaging isocenter, [12pt]{minimal}
$$\:$$ ), the distance ( [12pt]{minimal}
$$\:{d}_{-}$$ ) between the radiation isocenter ( [12pt]{minimal}
$$\:$$ ) and the [12pt]{minimal}
$$\:$$ , and the radius ( [12pt]{minimal}
$$\:r$$ ) of the smallest sphere that intersects all beams . The values of [12pt]{minimal}
$$\:{d}_{}$$ , [12pt]{minimal}
$$\:{d}_{-}$$ , and [12pt]{minimal}
$$\:r$$ were obtained at all institutions. An inter-manufacturer comparison was made for these values. For [12pt]{minimal}
$$\:{d}_{}$$ and [12pt]{minimal}
$$\:{d}_{-}$$ , the comparison was made for each direction of the image (DICOM) reference coordinate system and for the resultant vector.
We assessed the impact of the spatial dose delivery accuracy on the PTV margins in intracranial SRS. PTV margins were derived using the 3D-SS analysis results. Zhang et al. proposed a method to calculate the optimal PTV margin for single-fraction intracranial SRS that incorporates the linac spatial dose delivery error and the residual patient setup error after image guidance , . The method provides an anisotropic margin for intracranial SRS that ensures with 95% probability that the clinical target volume receives the prescribed dose. In this study, we developed the method by inputting the 3D dose delivery accuracy derived from the 3D-SS test as the linac spatial dose delivery error, building on a previously reported study . After calculating the optimal PTV margins, inter-institution and inter-manufacturer comparisons were made. The formulas are 1 [12pt]{minimal}
$$\:{C}_{i}={W}_{0i}+{b}_{1}({W}_{0i}){\:}_{i}+{b}_{2}({W}_{0i}){\:}_{i}^{2},$$ 2 [12pt]{minimal}
$${b}_{1}({W}_{0i})={2.331-1.425}W_{0i}+2.296{W}_{0i}^{2}-1.539{W}_{0i}^{3}+0.374{W}_{0i}^{4},$$ 3 [12pt]{minimal}
$${b}_{2}({W}_{0i})={0.434W}_{0i}-0.917{W}_{0i}^{2}+0.676{W}_{0i}^{3}-0.171{W}_{0i}^{4},$$ where [12pt]{minimal}
$$\:{C}_{i}$$ represents the PTV margin for the i = X , Y , and Z directions in the DICOM reference coordinate system, which is correspond to the patient right-left, patient superior-inferior, and patient anterior-posterior directions, respectively; [12pt]{minimal}
$$\:{W}_{0i}={V}_{s0i}+{V}_{r0i}$$ , where [12pt]{minimal}
$$\:{V}_{s0i}$$ represents the systematic error in the imaging isocenter and radiation isocenter along each axis; [12pt]{minimal}
$$\:{V}_{r0i}$$ denotes the average residual error in patient setup after image guidance; and [12pt]{minimal}
$$\:{\:}_{i}$$ is a function of the standard deviation of the residual error in patient setup ( [12pt]{minimal}
$$\:{SD}_{,i}$$ ). In this study, we added fluctuations in the radiation beam position (random error in the radiation isocenter) to [12pt]{minimal}
$$\:{\:}_{i}$$ : 4 [12pt]{minimal}
$$\:{\:}_{i}=_{,i}^{2}+{(0.683\:r)}^{2}},$$ where the term [12pt]{minimal}
$$\:0.683\:r$$ corresponds to the standard deviation of the radiation isocenter error, which was assumed to follow a normal distribution. In this study, [12pt]{minimal}
$$\:{V}_{s0i}$$ was directly set equal to [12pt]{minimal}
$$\:{d}_{-}$$ , which includes the influence of couch rotation. The values of [12pt]{minimal}
$$\:{V}_{r0i}$$ and [12pt]{minimal}
$$\:{SD}_{,i}$$ were taken from the studies of Zhang et al. and Ong et al. and are shown in Supplementary Table A1. The values of the former and the latter were defined as the larger and smaller residual patient setup errors.
We evaluated the impact of implementing advanced image guidance systems for patient positioning and rigorous geometric QA of the linac. This study assumed that advanced image guidance systems, which include widely installed video-based 3D optical surface imaging and 2D kV stereoscopic imaging, were implemented. During non-coplanar irradiation in clinical settings, we also assumed that intra-fraction patient motion caused by treatment couch shifts during couch rotation could be monitored and corrected by these imaging modalities. For example, we hypothesized that current surface-guided radiation therapy systems could detect a couch walk-out of < 1 mm . Furthermore, the value of [12pt]{minimal}
$$\:{d}_{-}$$ was treated as a systematic error for linacs that could be minimized by a geometric calibration test performed by the user. These assumptions form the basis for the further analysis conducted in this study. Therefore, this study assumed that in clinical practice, the user’s efforts can reduce the impact of these two factors on geometric accuracy to clinically negligible values. Based on this assumption, the PTV margin was recalculated by replicating the situation using the following procedure: 3D dose delivery accuracy was recalculated using only the co-planar beams, and the new values of [12pt]{minimal}
$$\:{d}_{-}$$ and [12pt]{minimal}
$$\:r$$ were measured. Subsequently, [12pt]{minimal}
$$\:{d}_{-}$$ was set to zero to account for the error contribution of the geometric calibration of the linac. Finally, the PTV margin was recalculated using Eqs. – and the above value of [12pt]{minimal}
$$\:r$$ .
Statistical analyses were performed to examine inter-manufacturer differences in various parameters. The Kruskal–Wallis test and Mann–Whitney U test were used to compare the results from more than two institutions and vendors, respectively. The Steel–Dwass post hoc test was used to make pair-wise comparisons. All statistical analyses were conducted using EZR (version 1.62, Saitama Medical Center, Jichi Medical University, Saitama, Japan). A value of p < 0.05 was considered statistically significant.
Figure categorizes the values of [12pt]{minimal}
$$\:{d}_{}$$ , [12pt]{minimal}
$$\:{d}_{-}$$ , and [12pt]{minimal}
$$\:r$$ for non-coplanar irradiation for each institution according to the linac manufacturer, and the raw data for these parameters are summarized in Supplements B, C, and D, respectively. For the seven beams used at each institution, [12pt]{minimal}
$$\:{d}_{}\:$$ showed variability among institutions and between manufacturers (Fig. A and Supplements B1, B2). As shown in Fig. A, the positional errors in the beam paths of the Elekta systems were significantly larger than those of the Varian systems in the X ( [12pt]{minimal}
$$\:p\:$$ < 0.001), Y ( [12pt]{minimal}
$$\:p$$ < 0.001), and Z ( [12pt]{minimal}
$$\:p$$ < 0.05) directions. In terms of the resultant vector, the Elekta systems had a significantly larger positional error in the beam path than the Varian systems [12pt]{minimal}
$$\:(p\:$$ < 0.05), as shown in Fig. B, and the numbers of institutions for which the value exceeded 1 mm were one and nine for Varian and Elekta systems, respectively. These 10 institutions exceeded the 1 mm tolerance proposed for the E2E test by the Medical Physics Practice Guidelines developed by the American Association of Physicists in Medicine and also proposed in the consensus statement from the working groups for radiosurgery and stereotactic radiotherapy of the German Society for Radiation Oncology and for physics and technology in stereotactic radiotherapy of the German Society for Medical Physics , . The value of [12pt]{minimal}
$$\:{d}_{-}$$ , which includes the influence of couch rotation, was evaluated. As shown in Fig. C and Supplement C, the values of [12pt]{minimal}
$$\:{d}_{-}$$ for the Varian systems were within ± 0.32 mm in all directions, and the median values were within 0.1 mm. In contrast, the Elekta systems had a maximum [12pt]{minimal}
$$\:{d}_{-}$$ of 1.21 mm in the Y direction, with median values in the X , Y , and Z directions of 0.16, 0.39, and − 0.42 mm, respectively. The displacements for the Elekta systems were significantly larger than those for the Varian systems ( [12pt]{minimal}
$$\:p$$ < 0.001), as shown in Fig. D and Supplement C. The radius of the smallest sphere intersecting all radiation beams, which includes the effect of couch rotation, is presented in Fig. E and Supplement D. The Elekta linacs had a significantly larger radius ( [12pt]{minimal}
$$\:p\:$$ < 0.001) than the Varian linacs. The correlation coefficients were derived from a linear regression analysis of the relationship between [12pt]{minimal}
$$\:{d}_{-}$$ and the number of operational years of each machine, as well as between [12pt]{minimal}
$$\:r$$ and the number of operational years of each machine. As shown in Table , Varian linacs had no correlation in terms of increments in [12pt]{minimal}
$$\:{d}_{-}$$ and [12pt]{minimal}
$$\:r$$ with the number of operational years of each machine. However, moderate correlations were observed for Elekta, indicating that [12pt]{minimal}
$$\:{d}_{-}$$ and [12pt]{minimal}
$$\:r$$ increased as the machine life advanced. Figure and Supplement E show the PTV margins derived from 3D-SS analysis of dose delivery error and from residual patient setup errors after image guidance. The PTV margins varied among institutions and between manufacturers. The minimum and maximum PTV margins were 1.0 and 3.4 mm, respectively, regardless of manufacturer or coordinate axis. For Varian, the margin ranges under the assumed smaller and larger setup errors were 1.0–1.9 mm, 1.0–1.7 mm, and 1.0–2.3 mm in the X, Y, and Z directions, respectively. For Elekta, the margin ranges under the two conditions were 1.3–2.7 mm, 1.2–2.8 mm, and 1.4–3.4 mm in the X, Y, and Z directions, respectively. Significant differences in the PTV margins between the Varian and Elekta systems were measured in the X ( [12pt]{minimal}
$$\:p$$ < 0.01), Y ( [12pt]{minimal}
$$\:p$$ < 0.01), and Z ( [12pt]{minimal}
$$\:p$$ < 0.001) directions. Assuming that the PTV margins calculated for the institutions sampled follow a normal distribution, the cumulative frequency distributions of the PTV margins were calculated for each manufacturer (Fig. ). The probability of exceeding a 2 mm margin was higher for Elekta systems than for Varian systems. Table shows the PTV margins that could cover the GTV in the X , Y , and Z directions with 95% probability at 95% of the institutions. A PTV margin of 2 mm was achievable for the Varian systems if the smaller setup errors reported by Ong et al. were assumed. The margin for Elekta systems was unachievable even under the assumed smaller and larger setup errors. Assuming that high-accuracy 3D dose delivery, strict patient immobilization and intra-fractional positioning, and rigorous linac QA could improve 3D geometric accuracy, the isocentricity [12pt]{minimal}
$$\:r$$ was reduced from 0.81 to 0.32 mm for Varian systems and from 1.45 to 0.76 mm for Elekta systems (cf. Supplements D and F). Being less than 1 mm, these values met the proposed tolerance level , . PTV margins for all institutions are shown in Table , and those margins that could cover the GTV with 95% probability at 95% of the institutions are shown in Table . Even when a larger residual patient setup error was assumed, the margin size for all institutions using Varian systems remained below 2 mm, and the anisotropic PTV margins that could cover the GTV with 95% probability at 95% of the institutions were 1.5, 1.1, and 2.0 mm in the X , Y , and Z directions, respectively. When a smaller residual patient setup error was assumed, the PTV margin size for all institutions using Elekta systems remained below 2 mm, and the isotropic margins that could cover the GTV with 95% probability at 95% of the institutions were 1 mm for Varian systems and 1.5 mm for Elekta systems.
To our knowledge, this is the first multi-institutional study to combine 3D-SS analysis with an X-ray CT-based polymer gel dosimeter to evaluate the 3D dose delivery accuracy, including couch rotation accuracy, of commercially available linacs broadly in clinical use. Moreover, institution- and manufacturer-specific PTV margins for intracranial SRS were derived from 3D-SS analysis of dose delivery accuracy. This report will serve as a useful reference for institutions seeking to objectively evaluate their own dose delivery accuracy and to determine the PTV margin for intracranial SRS. This study made several important findings. First, it provides generally achievable values for linacs from each manufacturer. Pant et al. proposed a 3D-SS method , and Oshika et al. demonstrated that it can minimize measurement uncertainties . However, there is no information on variations in 3D dose delivery accuracy among commercially available linacs. Second, we found that one Varian and nine Elekta linacs exceeded the 1 mm tolerance of SRS guidelines on spatial irradiation accuracy, even if conventional WL tests and timings were followed. In a previous study, there were discrepancies of < 0.4 mm in the resulting coincidence of each beam obtained via the conventional WL test and 3D-SS test . Thus, these results imply that conventional QA methods may underestimate 3D dose delivery errors. A 3D QA procedure with low measurement uncertainties, such as the 3D-SS test, should be performed to achieve dose delivery errors of no more than 1 mm in 3D coordinates. Third, non-coplanar irradiation for intracranial SRS performed without any image guidance system may need more than a 2 mm PTV margin when a larger residual patient setup error is assumed. Several studies have reported that the PTV margin can be reduced without affecting the clinical outcome , . However, it is unclear whether similar results can be obtained at other institutions because our findings indicate that the accuracy of irradiation is institution-dependent. In this study, the margin was in the 1.0–3.4 mm range. This considerable variation indicates that universal determination and reduction of the PTV margin should be approached cautiously. Fourth, older machines demonstrated lower precision than newer machines in only Elekta. Varian machines did not show the trends in this study. Gao J et al. compared the number of years of operation and the results of the off-iso Winston–Lutz test for Varian machines, and found that newer models had higher accuracy than older models . Because all the machines in this study were of the same model—TrueBeam and TrueBeam STx—it is likely that this resulted in differing outcomes. Collimator rotation was not considered in the 3D-SS test for this study. The 3D dose delivery accuracy and PTV margin may, therefore, be greater than the values reported in this study. Because a single gel dosimeter was used for the 3D-SS test and all beam trajectories were included within the dosimeter, the number of beams tested was limited. To assess the impact of collimator rotation, non-coplanar beams with varying collimator angles could be used in a separate gel dosimeter test. In addition, the margin formula established in this study does not account for factors such as inter-observer variations in target volume delineation among physicians or registration errors in planning CT and MRI. Therefore, considering these factors, the required margin may potentially increase further. In this study, we used linacs for intracranial stereotactic radiotherapy in clinical practice. Although the Elekta linac models used in this study belong to different generations, there are no differences in the structure and function of the linacs across the three models. The couch systems differ among the three models. However, only the yaw rotation of the couch was used in this study, and we considered that there was no significant difference in the accuracy of the yaw rotation among the models. It is important to note that the single-isocenter, multi-target (SIMT) delivery technique has gained increasing popularity. However, its heightened sensitivity to rotational errors in dose delivery to small off-isocenter targets presents significant challenges for margin calculation using analytical methods (e.g., the margin depends on the specific distance of each target from the isocenter). This underscores the critical advantage of 3D dose measurement, which provides superior accuracy and reliability compared to conventional methods. Finally, this study highlights the importance of precise management of the 3D dose delivery accuracy by medical physicists. Radiation oncology departments should also introduce a real-time monitoring system capable of correcting patient displacement due to body movement and couch rotation. Moreover, therapists must improve patient immobilization and verification of position accuracy. The margin could be reduced to within the recommended 2 mm limit by assuring high-accuracy 3D dose delivery, strict patient immobilization and intra-fractional positioning, and rigorous linac QA . Our multi-institutional study showed that it is possible to limit the PTV margin size for intracranial SRS to 1.0 mm for Varian systems and 1.5 mm for Elekta systems in all directions.
The 3D dose delivery accuracy of linacs currently in operation largely varied at the millimeter level in this study. The accuracy of current radiotherapy technology should not be overestimated, and it is essential to rigorously determine the 3D dose delivery accuracy and estimate the PTV margins. Using advanced image guidance systems with limited patient setup error and recognizing the accuracy variations between linear accelerators, it is essential to maximize a linac’s 3D dose delivery accuracy to achieve the required PTV margin in intracranial SRS.
Below is the link to the electronic supplementary material. Supplementary Material A Supplementary Material B Supplementary Material C Supplementary Material D Supplementary Material E Supplementary Material F
|
Person-centered abortion care scale: Validation for medication abortion in the United States | 452448db-955c-44e0-8011-0c8aea2eaa25 | 11849315 | Surgical Procedures, Operative[mh] | Introduction In June 2022, the US Supreme Court overturned the 1973 Roe v. Wade decision, eliminating federal protections for the provision of abortion. As of February 2024, abortion is illegal in 14 states and at risk of being highly restricted or banned in 26 states total . Amidst these restrictions and COVID-era changes in healthcare delivery, medication abortions now comprise 53% of all facility-based abortions . While the safety and efficacy of medication abortion are well-established, there are no validated tools to assist health care providers in ensuring high quality and person-centered medication abortion care . Quality of care is defined by the Institute of Medicine as care that is safe, effective, person-centered, timely, efficient, and equitable . In particular, person-centered care, or care that is respectful of and responsive to individuals’ preferences, needs, and values , is globally recognized as a distinct component of quality abortion care . However, there is a lack of validated measures for abortion quality, particularly person-centered care . Abortion care that is not person-centered can take many forms (e.g., discrimination; lack of pain management) with disproportionate impacts on people of color and other marginalized communities . Those of lower socioeconomic status and in highly restrictive legal settings with deeply embedded social stigma are also more likely to experience care that is not person-centered, contributing to health inequities . Person-centered care is essential not only as a human right, but is also positively associated with clinical outcomes and adherence to post-abortion guidance . The US Food and Drug Administration (FDA) approved direct mailing of medication abortion pills to patients through telemedicine options in April 2021, temporarily modifying the in-person dispensing requirement due to the COVID-19 pandemic. This long-standing requirement, which mandated that the first medication abortion pill (mifepristone) be administered in-person under a clinician’s supervision, was removed by the FDA in January 2023 . Telemedicine medication abortion is safe, effective, and acceptable . This is in line with the rise in availability and use of telemedicine services across a number of health sectors in recent years, from perinatal mental health to substance use to primary care services . There is a growing body of literature suggesting that telemedicine advances person-centered approaches by increasing patient satisfaction, decreasing barriers to care , and meeting peoples’ preferences for how they receive care (e.g. in clinic vs telemedicine) . To our knowledge, only two studies have examined person-centered care for telemedicine medication abortions. The studies found that it increased pregnant peoples’ options, autonomy and access to timely abortion care by removing healthcare barriers related to geographic distance, childcare, or employment . However, both studies were qualitative; more empirical, nuanced understanding of different aspects of person-centeredness is needed, including potential limitations of a telemedicine approach. This study aims to adapt and validate a scale to measure person-centered care for in-person and telemedicine medication abortions in California . Methods The Person-Centered Abortion Care (PCAC) scale, which was developed and validated in Kenya , served as the basis for adaptation and validation in the US context (US-PCAC). We employed a standard sequential approach to scale development described in detail below . This study was conducted at a large urban academic health clinic wherein eligible patients were given the option to have a medication abortion by telemedicine or in person between June 2018–December 2022. 2.1. Defining domains and expert reviews A technical advisory committee (TAC), comprised of 12 experts (i.e., US-based abortion service providers and researchers) reviewed domains of person-centered reproductive healthcare and the original 26-item PCAC scale developed in Kenya . Domains are the major constructs of person-centered care defined by the literature . Additionally, we conducted a literature review on recent measures of abortion experiences. From the original list, the expert reviewers modified, added, and deleted items, ultimately expanding the list to 44 items. In a follow-up TAC meeting and subsequent training with interviewers, TAC and study team members consolidated items from 44 to 37 items to reduce redundancy and omit items considered to be less relevant in a US setting. 2.2. Cognitive interviews In total, 37 items were tested during the cognitive interviews. Input from the cognitive interviews ( n = 12) included suggestions for slight wording changes (e.g., adding more specificity to questions regarding wait times), consistency of response options across items, verifying that terms were understood and resonated with the experiences of participants across modalities (e.g., “Did you feel seen and heard by the healthcare team?”), and removing items that were duplicative or less relevant according to study participants (e.g. Did you feel like you were physically treated roughly?). In total, seven items were removed that were duplicative or less relevant. Additional changes to items were based on more substantive input from participants; for example, two items were revised to use the term “decisions” versus “decision” to differentiate between the multiple decisions required in the medication abortion process (e.g., where and when they wanted to take the medication, methods for managing pain) versus the larger “decision” to have an abortion or to have a medication versus procedural abortion. Moreover, three items were added based on participant feedback including “Do you feel you were provided with enough information on what to expect regarding pain or discomfort that could arise from the procedure?” “Did you feel that you could confide in the health care team regarding personal or sensitive information?” and “Did you feel that the healthcare team showed that they care about you?”. 2.3. Person-centered abortion care survey The eligibility criteria for the US-PCAC survey were as follows: (1) had either a medication abortion via telemedicine with no exam or ultrasound between April 1, 2020 and December 31, 2022 (referred in text as “telemedicine” patients) or an in-person medication abortion between June 1, 2018 and December 31, 2022 (referred in text as “in-person” patients); (2) 6-weeks or more from completion of abortion to be able to report outcomes (e.g. abortion completion); (3) able to take the online survey in English; and (4) 18 years or older at time of recruitment. The longer timeframe for the in-person sample was to allow for sufficient sample size given the limited number of abortions performed during the study period. The study team consulted with the clinic’s Clinical and Translational Science Institute (CTSI) biomedical informatics team to obtain lists of eligible participants. Each eligible participant received a recruitment message containing the personal Qualtrics survey link and passcode via email and/or through the hospital’s secure messaging platform. Participants were directed to the informed consent online page. Once they agreed to participate, they were directed to the 20-minute online survey that included questions on demographics, social, and health outcomes, in addition to the PCAC items. The survey was conducted from December 2021 to March 2023. A total of 970 patients were contacted, with up to three follow-up reminders. The final sample size was 182 participants (147 in-person and 45 telemedicine patients) resulting in a participation rate of 18.8%. A general rule for minimum sample size needed for conducting factor analysis is three participants times the number of items . Our target sample size was 150 participants per modality (for total of 300 participants), but this was not achieved due to the low volume of abortions in the clinic, particularly for the telemedicine sample. Each respondent who completed the survey was given a $20 electronic gift card. 2.4. Psychometric analyses In total, 33 US-PCAC items were included, with two telemedicine specific items and one in-person specific item (see ). We ran a series of factor analyses for the full sample and then separately for in-person and telemedicine participants. Missingness was low across all variables (< 5%) and we therefore used complete case analysis (see ). All analyses were conducted using Stata . Negative items were reverse coded so that negative responses were coded 0 and best responses coded 3 to obtain a uniform scale. We constructed a correlation matrix and examined item-test correlation, item-rest correlation, and alpha to assess reliability. We then conducted exploratory factor analysis. We first assessed a one-factor solution to assess if there was a global measure of US-PCAC. We examined factor loadings of each item, using a cutoff of 0.30 to determine which items to delete or retain. We used oblique rotation because of the naturally occurring correlation between the rotated factors . We used a scree plot, eigenvalues of factors, and conceptual justifications (e.g., examining how each item is understood and theorized in existing literature) to determine the number of factors to retain. We first assessed a scree plot to visually inspect factors with eigenvalues greater than 1.0 . Cronbach’s alpha was used to examine internal consistency for the full and sub-scales, with 0.70 considered to be acceptable reliability . We named sub-domains based on the factors and what is known from existing literature . Lastly, we examined criterion validity by assessing bivariate associations between US-PCAC scales and “satisfaction,” a perceived quality of care measure often used to assess quality outcome . Satisfaction was measured by the question, “Overall, how satisfied were you with the entire process?” Response options corresponded to a four-point Likert-type scale (Not at all, Somewhat, Very, or Extremely) and were dichotomized (0 = Not at all/Somewhat satisfied vs 1 = Very/Extremely Satisfied). We used logistic regression to assess bivariate associations. We also conducted sensitivity analyses using a continuous satisfaction score and results did not differ. All study procedures were reviewed and approved by an Institutional Review Board and informed consent was received by all participants. Defining domains and expert reviews A technical advisory committee (TAC), comprised of 12 experts (i.e., US-based abortion service providers and researchers) reviewed domains of person-centered reproductive healthcare and the original 26-item PCAC scale developed in Kenya . Domains are the major constructs of person-centered care defined by the literature . Additionally, we conducted a literature review on recent measures of abortion experiences. From the original list, the expert reviewers modified, added, and deleted items, ultimately expanding the list to 44 items. In a follow-up TAC meeting and subsequent training with interviewers, TAC and study team members consolidated items from 44 to 37 items to reduce redundancy and omit items considered to be less relevant in a US setting. Cognitive interviews In total, 37 items were tested during the cognitive interviews. Input from the cognitive interviews ( n = 12) included suggestions for slight wording changes (e.g., adding more specificity to questions regarding wait times), consistency of response options across items, verifying that terms were understood and resonated with the experiences of participants across modalities (e.g., “Did you feel seen and heard by the healthcare team?”), and removing items that were duplicative or less relevant according to study participants (e.g. Did you feel like you were physically treated roughly?). In total, seven items were removed that were duplicative or less relevant. Additional changes to items were based on more substantive input from participants; for example, two items were revised to use the term “decisions” versus “decision” to differentiate between the multiple decisions required in the medication abortion process (e.g., where and when they wanted to take the medication, methods for managing pain) versus the larger “decision” to have an abortion or to have a medication versus procedural abortion. Moreover, three items were added based on participant feedback including “Do you feel you were provided with enough information on what to expect regarding pain or discomfort that could arise from the procedure?” “Did you feel that you could confide in the health care team regarding personal or sensitive information?” and “Did you feel that the healthcare team showed that they care about you?”. Person-centered abortion care survey The eligibility criteria for the US-PCAC survey were as follows: (1) had either a medication abortion via telemedicine with no exam or ultrasound between April 1, 2020 and December 31, 2022 (referred in text as “telemedicine” patients) or an in-person medication abortion between June 1, 2018 and December 31, 2022 (referred in text as “in-person” patients); (2) 6-weeks or more from completion of abortion to be able to report outcomes (e.g. abortion completion); (3) able to take the online survey in English; and (4) 18 years or older at time of recruitment. The longer timeframe for the in-person sample was to allow for sufficient sample size given the limited number of abortions performed during the study period. The study team consulted with the clinic’s Clinical and Translational Science Institute (CTSI) biomedical informatics team to obtain lists of eligible participants. Each eligible participant received a recruitment message containing the personal Qualtrics survey link and passcode via email and/or through the hospital’s secure messaging platform. Participants were directed to the informed consent online page. Once they agreed to participate, they were directed to the 20-minute online survey that included questions on demographics, social, and health outcomes, in addition to the PCAC items. The survey was conducted from December 2021 to March 2023. A total of 970 patients were contacted, with up to three follow-up reminders. The final sample size was 182 participants (147 in-person and 45 telemedicine patients) resulting in a participation rate of 18.8%. A general rule for minimum sample size needed for conducting factor analysis is three participants times the number of items . Our target sample size was 150 participants per modality (for total of 300 participants), but this was not achieved due to the low volume of abortions in the clinic, particularly for the telemedicine sample. Each respondent who completed the survey was given a $20 electronic gift card. Psychometric analyses In total, 33 US-PCAC items were included, with two telemedicine specific items and one in-person specific item (see ). We ran a series of factor analyses for the full sample and then separately for in-person and telemedicine participants. Missingness was low across all variables (< 5%) and we therefore used complete case analysis (see ). All analyses were conducted using Stata . Negative items were reverse coded so that negative responses were coded 0 and best responses coded 3 to obtain a uniform scale. We constructed a correlation matrix and examined item-test correlation, item-rest correlation, and alpha to assess reliability. We then conducted exploratory factor analysis. We first assessed a one-factor solution to assess if there was a global measure of US-PCAC. We examined factor loadings of each item, using a cutoff of 0.30 to determine which items to delete or retain. We used oblique rotation because of the naturally occurring correlation between the rotated factors . We used a scree plot, eigenvalues of factors, and conceptual justifications (e.g., examining how each item is understood and theorized in existing literature) to determine the number of factors to retain. We first assessed a scree plot to visually inspect factors with eigenvalues greater than 1.0 . Cronbach’s alpha was used to examine internal consistency for the full and sub-scales, with 0.70 considered to be acceptable reliability . We named sub-domains based on the factors and what is known from existing literature . Lastly, we examined criterion validity by assessing bivariate associations between US-PCAC scales and “satisfaction,” a perceived quality of care measure often used to assess quality outcome . Satisfaction was measured by the question, “Overall, how satisfied were you with the entire process?” Response options corresponded to a four-point Likert-type scale (Not at all, Somewhat, Very, or Extremely) and were dichotomized (0 = Not at all/Somewhat satisfied vs 1 = Very/Extremely Satisfied). We used logistic regression to assess bivariate associations. We also conducted sensitivity analyses using a continuous satisfaction score and results did not differ. All study procedures were reviewed and approved by an Institutional Review Board and informed consent was received by all participants. Results A total of 182 participants completed all PCAC scale items, including 137 in-person participants and 45 telemedicine participants. Demographic characteristics are presented in . 3.1. Exploratory factor analysis (EFA) Due to low item-correlation and factor loading, we removed “respect support person” (health care team respectful towards support person). For the full sample of participants, we assessed all items excluding the dropped item “respect support person” (29 items). After examining the eigenvalues for each item, we found that items fit better onto a three-factor solution. Oblique rotation indicated three factors and 15 items with a factor loading > 0.3 loaded positively onto one of the three factors: seven items onto Factor 1 corresponding to the Respect and Dignity sub-domain; five items onto Factor 2 corresponding to Responsive and Supportive Care; three items onto Factor 3 corresponding to Communication and Autonomy. Of items that cross-loaded to more than one factor above the cutoff, we categorized four items based on the factor with the higher loading. The remaining items were categorized on conceptual reasoning. Four items loaded highest on Factor 1 but were categorized into other factors for conceptual reasons: “Confidential” (feeling that the health care team kept health information confidential) was categorized into Factor 2; and “Involved,” (provider involved in decisions about care) “Questions,” (could ask health care team any questions) and “Answers” (get answers to all questions in a satisfactory manner) into Factor 3. Three items that loaded highest onto Factor 4 were categorized for conceptual reasons: “Treat negatively” (treated negatively based on identities or characteristics) was grouped into Factor 1, “Overhear” into Factor 2, and “Coerced” (coerced into a decision) into Factor 3. Two items “Support person” and “Language understand” (health care team spoke in understandable language and manner) loaded under the cutoff but was categorized into Factor 1 on a conceptual basis. presents oblique rotated factor loadings for the 29 items in the final US-PCAC scale and summarizes final decisions for each item’s subdomain. We also conducted factor analyses separately for the in-person and telemedicine samples. For the in-person sample, one additional item “Exams private” (covered up during exams) was added and had a factor loading > 0.3 in Factor 1. For the telemedicine sample, two additional items were administered “Communicate telemedicine” (communicate effectively using telemedicine portal) and “Telemedicine private” (telemedicine visit felt private and secure). “Communicate telemedicine” had a factor loading of 0.4971 in Factor 2 but was categorized under Communication and Autonomy for conceptual reasons. “Telemedicine private” loaded under the cutoff but was retained and categorized under Responsive and Supportive care on conceptual bases. presents standardized alphas and scale descriptive statistics for the full US-PCAC scale and subscales, standardized to a 100-point scale. For the full sample, the standardized alpha for the 29-item PCAC scale was 0.95 (mean score = 87.86, SD = 15.03; Range 20.69–100). For the in-person sample, the standardized alpha for the 30-item PCAC scale was also 0.95 (mean score = 85.94, SD = 16.14; Range 22.22–100). Due to the small sample of telemedicine participants, the standardized alpha was unable to be calculated for the 31-item US-PCAC scale. The unstandardized alpha for the telemedicine 31-item scale was 0.86 (mean score = 94.24, SD = 7.22; Range = 61.29–100). 3.2. Criterion validity In bivariate results, among the full sample, each one-unit increase in total standardized US-PCAC score was associated with a 1.11 times (95% CI: 1.07, 1.15) higher odds of satisfaction . All PCAC subscales were also positively associated with satisfaction: Respect and Dignity (OR = 1.06, 95% CI: 1.04, 1.09), Responsive and Supportive Care (OR = 1.06, 95% CI: 1.03, 1.09), and Communication and Autonomy (OR = 1.12, 95% CI: 1.08, 1.17). For the in-person subsample, participants who reported greater satisfaction with the entire process had significantly higher US-PCAC total and subscale scores. Exploratory factor analysis (EFA) Due to low item-correlation and factor loading, we removed “respect support person” (health care team respectful towards support person). For the full sample of participants, we assessed all items excluding the dropped item “respect support person” (29 items). After examining the eigenvalues for each item, we found that items fit better onto a three-factor solution. Oblique rotation indicated three factors and 15 items with a factor loading > 0.3 loaded positively onto one of the three factors: seven items onto Factor 1 corresponding to the Respect and Dignity sub-domain; five items onto Factor 2 corresponding to Responsive and Supportive Care; three items onto Factor 3 corresponding to Communication and Autonomy. Of items that cross-loaded to more than one factor above the cutoff, we categorized four items based on the factor with the higher loading. The remaining items were categorized on conceptual reasoning. Four items loaded highest on Factor 1 but were categorized into other factors for conceptual reasons: “Confidential” (feeling that the health care team kept health information confidential) was categorized into Factor 2; and “Involved,” (provider involved in decisions about care) “Questions,” (could ask health care team any questions) and “Answers” (get answers to all questions in a satisfactory manner) into Factor 3. Three items that loaded highest onto Factor 4 were categorized for conceptual reasons: “Treat negatively” (treated negatively based on identities or characteristics) was grouped into Factor 1, “Overhear” into Factor 2, and “Coerced” (coerced into a decision) into Factor 3. Two items “Support person” and “Language understand” (health care team spoke in understandable language and manner) loaded under the cutoff but was categorized into Factor 1 on a conceptual basis. presents oblique rotated factor loadings for the 29 items in the final US-PCAC scale and summarizes final decisions for each item’s subdomain. We also conducted factor analyses separately for the in-person and telemedicine samples. For the in-person sample, one additional item “Exams private” (covered up during exams) was added and had a factor loading > 0.3 in Factor 1. For the telemedicine sample, two additional items were administered “Communicate telemedicine” (communicate effectively using telemedicine portal) and “Telemedicine private” (telemedicine visit felt private and secure). “Communicate telemedicine” had a factor loading of 0.4971 in Factor 2 but was categorized under Communication and Autonomy for conceptual reasons. “Telemedicine private” loaded under the cutoff but was retained and categorized under Responsive and Supportive care on conceptual bases. presents standardized alphas and scale descriptive statistics for the full US-PCAC scale and subscales, standardized to a 100-point scale. For the full sample, the standardized alpha for the 29-item PCAC scale was 0.95 (mean score = 87.86, SD = 15.03; Range 20.69–100). For the in-person sample, the standardized alpha for the 30-item PCAC scale was also 0.95 (mean score = 85.94, SD = 16.14; Range 22.22–100). Due to the small sample of telemedicine participants, the standardized alpha was unable to be calculated for the 31-item US-PCAC scale. The unstandardized alpha for the telemedicine 31-item scale was 0.86 (mean score = 94.24, SD = 7.22; Range = 61.29–100). Criterion validity In bivariate results, among the full sample, each one-unit increase in total standardized US-PCAC score was associated with a 1.11 times (95% CI: 1.07, 1.15) higher odds of satisfaction . All PCAC subscales were also positively associated with satisfaction: Respect and Dignity (OR = 1.06, 95% CI: 1.04, 1.09), Responsive and Supportive Care (OR = 1.06, 95% CI: 1.03, 1.09), and Communication and Autonomy (OR = 1.12, 95% CI: 1.08, 1.17). For the in-person subsample, participants who reported greater satisfaction with the entire process had significantly higher US-PCAC total and subscale scores. Discussion This study is significant in that it is the first validated quality of care scale for abortion in the US and highlights three dimensions of care: respect and dignity, communication and autonomy, and responsive and supportive care. Our study found high construct, content, criterion validity and reliability for the PCAC scale in a US setting for both in-person and telemedicine medication abortion care. Given evidence of improved clinical and patient outcomes associated with patient-centered care , the US-PCAC provides a much-needed, standardized tool that may aid monitoring and research efforts. The US-PCAC scale adds to a set of validated person-centered care scales for reproductive health that include scales for abortion , family planning , prenatal , and intrapartum care . While there are other scales that measure person-centered contraceptive care (see ), having a standardized set of measures across the continuum of sexual and reproductive healthcare allows for comparisons across contexts and health services. Across several studies, the communication and autonomy domain consistently has the lowest scores, suggesting that a focus on ensuring comprehension of medical procedures and patient involvement in shared decision-making may be necessary. This study has several limitations. First, the small sample size, particularly for the telemedicine group, was not sufficiently robust to validate the scale for the telemedicine-specific sample. However, this study provides exploratory evidence of high construct validity and reliability for the overall scale for the telemedicine sample. Second, to recruit sufficient samples, we expanded our eligibility criteria to those who had in-person abortions in 2018 to present; thereby increasing the possibility for challenges in recalling specifics of their care. Third, this study was only offered in English, limiting our sample to English-proficient participants. Lastly, we recognize the limitations of the commonly-used global “satisfaction” measure to assess criterion validity, including that satisfaction is a product of expectations, such that people with low expectations may report higher satisfaction with poor care; moreover, abortion patients oftentimes report high satisfaction because of high stigma associated with abortion . However, given the lack of gold standard, we use satisfaction to measure the outcome of people’s experiences and recognize the need for future studies to examine person-centered care on other abortion outcomes. The US-PCAC is unique as it includes items specific to either in-person or telemedicine medication abortion. Person-centered care remains critically important given the expansion of telemedicine medication abortion services during COVID and in the post-Dobbs era . The tool will support broader monitoring and research efforts by establishing guidelines for person-centered abortion quality indicators in the US setting. Given differences in quality of abortion care by setting and patient characteristics, the tool can be used to assess health inequities in person-centered abortion care . Future studies may refine and shorten the number of items as performance metrics for quality improvement efforts in clinic settings in order to provide actionable recommendations to healthcare providers and systems. Appendix A |
Hypokalaemic periodic paralysis and myotonia in a patient with homozygous mutation p.R1451L in Na | 9d9aac04-66ed-4ac5-91db-87ae59fead86 | 6018793 | Physiology[mh] | The autosomal dominant channelopathies of the skeletal muscle voltage-gated sodium channel (Na V 1.4, encoded by SCN4A ) include myotonia and periodic paralysis (PP) – that is often associated with hyper- or hypokalaemia (hyperPP or hypoPP). A patient can manifest with both myotonia and PP, and is diagnosed with paramyotonia congenita (PMC) or hyperkalaemic periodic paralysis (hyperPP) depending on the predominant presentation. Carriers of the same mutation can have different main presentations, even within a pedigree , . The causative mutations enhance the activity of Na V 1.4 and predispose the muscle to increased action potential firing that manifests clinically as myotonia. The symptoms can be exacerbated or provoked by raised serum K + levels, which depolarise the potassium equilibrium potential and consequently the resting membrane voltage, thereby promoting Na V 1.4 activity. In some cases, the increased sodium channel activity and repetitive action potential firing can eventually lead to excess depolarisation of the muscle that then inactivates the sodium channels and prevents further action potential firing, resulting in flaccid paralysis. Hypokalaemic PP (hypoPP) has a distinct pathomechanism to PMC and hyperPP – , , . The mutations cause a leak current through voltage sensing domains (VSDs) of Na V 1.4 or skeletal muscle voltage-gated calcium channel Ca V 1.1 – , – . This current is known as the gating pore current and can depolarize the muscle to a level that inactivates the Na V 1.4 channels and paralyses the muscle, particularly in presence of hypokalaemia that attenuates hyperpolarising currents in the muscle , . The hypoPP mutations linked to Na V 1.4 target arginine residues in the fourth transmembrane helix (S4) of three (I-III) of the four homologous VSDs of the channel – . Mutations affecting S4 arginines in VSD-IV have been shown not to conduct gating pore currents , , . Instead, Na V 1.4 VSD-IV has been implicated as the key VSD regulating fast inactivation , which is often defective for the dominant mutations affecting S4 arginines in VSD-IV – . The clinical presentations of autosomal recessive channelopathies of Na V 1.4 include congenital myasthenia , , and myopathy , and occasionally extreme weakness or paralysis , . Mutations in these recessive conditions attenuate Na V 1.4 function. Congenital myasthenia Na V 1.4 mutant channels show enhanced fast inactivation: the voltage dependence is shifted towards hyperpolarised voltages and the recovery from inactivation is slower than for wild-type channels , , . Consequently, upon repetitive stimulation the mutant channels accumulate in the inactive state more than the wild-type channels, which accounts for increased fatigability. In cases with congenital myopathy the functional defects are more variable but one of the SCN4A alleles is null , and the weakness in the patients is fixed rather than fluctuating. Curiously, no neurological presentations were reported for heterozygous carriers of null variants . Myotonia is characteristically considered a distinguishing clinical feature in determining a diagnosis of either hyperPP or hypoPP , consistent with distinct pathomechanisms and opposite dependence on extracellular potassium concentration. There are rare reports, however, of patients presenting with both hypoPP and myotonia , . We describe here a consanguineous Chinese patient presenting with episodes of paralysis, associated with hypo- or normokalemia, and with myotonia. Next generation sequencing of the proband identified a homozygous mutation p.R1451L affecting the second S4 arginine in VSD-IV. A second unrelated heterozygous p.R1451L pedigree presented with hyperPP and myotonia. Our functional analysis of the mutant channel describes the pathomechanisms underlying the recessive hypoPP and the rarely reported co-presentation of hypoPP and myotonia.
Clinical features Family 1 The proband is a 19-year-old Chinese male presenting with a 15 year history of recurrent quadriplegia associated with hypokalaemia or normokalaemia (Fig. ). He experienced his first episode after a high fever at the age of 4. After he turned 7 the episodes occurred 2–3 times per year and usually lasted for several hours to 2 days with spontaneous full recovery. The attacks became more frequent when he was 16. Generalised limb muscle weakness occurred once or twice per month and isolated lower limb weakness 4–5 times per month. The ability to swallow and speak was well preserved. Triggers include rest after a period of vigorous exercise, exposure to cold, upper respiratory infection, and maintenance of a fixed posture. Serum potassium level was 2.8, 3.3 and 4.0 mmol/L (normal range: 3.5–5.5 mmol/L) measured in three distinct episodes of paralysis. None of oral administration with potassium, carbamazepine and topiramate can reduce the attacks or ameliorate the severity. Additionally, physical examination revealed a generalised increase in muscle bulk, especially in bilateral quadriceps and calf muscles. The patient didn’t complain of clinical myotonia, but a decreased ability to relax after a forced eyelid closure was demonstrated. Neither grip myotonia nor percussion myotonia was detected. Muscle strength and the tendon reflex were normal. After coenzyme Q10 administration for 2 months (200 mg/day), the patient reported that recurrent muscle weakness was greatly relieved. The effect of coenzyme Q10 treatment on eyelid myotonia was not assessed. Next generation sequencing revealed the homozygous p.R1451L mutation in the skeletal muscle Na V 1.4 sodium channel. No other causal mutations were detected in a panel of 10 ion channel genes and 245 primary myopathies/muscular dystrophies-related genes (Supplementary Table ). The parents are consanguineously married carriers (Fig. ) of the heterozygous p.R1451L mutation. Their running ability is unaffected and they do not display symptoms of clinical myotonia although the muscle volume is relatively increased. A 22 years old sister with a heterozygous p.R1451L mutation complained of muscle stiffness after a short rest. The father had no complaints of muscle weakness. Clinical PP was not observed in the heterozygous carriers of p.R1451L. Family 2 A 23 year old man presented to the national referral centre for skeletal muscle channelopathies in the UK with symptoms of episodic muscle weakness. His birth history, motor milestones and childhood were normal with no specific symptoms. At the age of 18 years after a period of prolonged sitting he experienced bilateral leg weakness that lasted for several hours and spontaneously resolved. From this point he experienced recurrent episodes of limb muscle weakness and at the time of referral these were occurring daily. They could occur at any time of the day although a significant number occurred first thing in the morning with less frequent episodes during the night. Symptoms lasted typically from 10 mins to a few hours and could be exacerbated by cold weather, exercise or prolonged rest. He did not identify any specific food triggers. Three episodes all occurring from sleep had been severe enough to warrant hospital attendance. Serum potassium was recorded within the normal range on each occasion. In addition to the attacks of weakness he reported his hands could become “stuck”, particularly in cold weather. Examination revealed him to have a muscular physique, which was notable relative to a general lack of exercise. There was mild hand and grip myotonia that was exacerbated by repetition, but the remainder of the exam was unremarkable. He reported a family history of affected father with similar symptoms although had no regular contact with him. Although the diurnal pattern is reminiscent of hypoPP, all other clinical features were considered consistent with hyperPP and the patient was treated with a combination of acetazolamide 500 mg BD and bendroflumethiazide 5 mg OD. This reduced although did not abolish his attacks of paralysis. There was no effect on myotonia. Genetic analysis identified the heterozygous SCN4A mutation p.R1451L in both the proband and his father. Electrodiagnostic studies Electromyographic (EMG) studies of the proband in family 1 demonstrated myotonic discharges in a number of tested muscles (Table ; Deltoids, Flex Carpi Rad, Vastus Med, Gastroc caput med and Tibialis anterior). Of note, poor recruitment of muscle unit potential was demonstrated in the rectus abdominis and Vastus Med. Myotonic discharges are documented in some selected muscles for the parents and the sister (Table ). Repeat EMG was not performed following administration of Coenzyme Q10. Long exercise test (LET) revealed a significant decline of >30% of baseline amplitude value for the proband (Fig. ). After two months oral administration of Coenzyme Q10, compound muscle action potential (CMAP) decrement >30% occurred at the same time, but the baseline value of long exercise test increased from 3.9 mV to 5.7 mV in abductor digiti minimi (ADM) muscle (Fig. ). For the father and the sister, a significant decline of CMAP amplitude in LET was observed 25 minutes and 10 minutes after exercise, respectively. For the mother, the CMAP amplitude and area slightly declined and did not reach a 30% threshold (Fig. ). The proband in family 2 had evidence of myotonic runs in the first dorsal interosseous and tibialis anterior. A long exercise test was positive for periodic paralysis with a decrement of CMAP amplitude from the baseline of 67%. Analysis of gating pore currents As hypoPP is associated with gating pore currents caused by mutations that affect S4 arginines we first analysed if p.R1451L mutant channels conducted gating pore currents using the Xenopus laevis oocyte expression system. To isolate the gating pore currents the main pore was blocked with 1–2 µM tetrodotoxin. The amplitude of the gating pore current of S4 arginine mutant channels is increased when guanidinium acts as charge carrier , . However, when 50% of the extracellular sodium was substituted by guanidinium, steady state current amplitude did not differ between p.R1451L and wild-type channels at physiological voltages (Fig. ), whereas currents in known hypoPP mutant p.R222W were significantly larger (one way ANOVA, Dunnet multiple comparisons, compared at −80 mV). These data suggest that p.R1451L channels do not conduct gating pore currents. Analysis of main pore currents We studied the main pore sodium current of the p.R1451L mutant in HEK293 cells (Figs , , Table ). Consistent with previous analysis of p.R1451L properties , the peak current amplitude of the mutant channel was significantly reduced compared to the wild-type channel (Fig. ), the voltage of half-maximal fast inactivation was shifted to more hyperpolarized voltages and the slope was more shallow (Fig. ). We didn’t observe changes in the voltage of half-maximal channel activation, although the slope of p.R1451L activation was significantly more shallow than for wild-type channels (Fig. ). The voltage dependence of activation and fast inactivation were similar when the holding voltage (V h ) was −80 or −100 mV (Table ). We further analysed the defective fast inactivation at two physiological holding voltages (Fig. , Table ). With V h −100 mV, the onset of closed-state inactivation at −60 mV and −80 mV was significantly accelerated for p.R1451L channels compared to wild-type channels, consistent with the left-shift in the voltage dependence of fast inactivation (Fig. ). In contrast, the onset of p.R1451L open-state inactivation was decelerated compared to wild-type channels in the studied voltage range −20 mV to +20 mV (Fig. ). The rate of closed-state inactivation at −60 mV was accelerated for wild-type channels when the holding voltage was −80 mV, compared to data with V h = −100 mV (Supplemental Fig. ), but was not analysed for p.R1451L channel as at this V h more than 50% of the channels are already inactivated. The rate of open-state inactivation was unaffected by the holding voltage (Table ). The rate of p.R1451L channel recovery from inactivation was 3 times faster than for the wild-type channel at −80 mV (p.R1451 L τ = 1.9 ms vs τ = 6.1 for wild-type channel, Fig. ). At −100 mV the rate of recovery was accelerated for both channels and although the fold difference was reduced, the recovery of p.R1451L channels was still significantly faster than for wild-type channels (p.R1451L τ = 1.6 ms vs τ = 2.4 for WT, Fig. ). A prediction of the accelerated recovery from inactivation and the reduced rate of open-state inactivation of p.R1451L channels is that upon repetitive high-frequency stimulation the availability of the mutant channels reduces less than for wild-type channels. Consistently, when a 2 ms pulse to 0 mV was given from a holding voltage of −80 mV at frequencies 40–200 Hz the current declined more for wild-type than for mutant channels (Fig. ). When V h was −100 mV the difference in availability between p.R1451L and wild-type channels was only observed at very high frequency (200 Hz), probably reflecting the reduced difference in the rate of recovery from inactivation of both channels at this voltage. Finally, the p.R1451L mutation did not alter the voltage of half-maximal slow inactivation of the Na V 1.4 channel although the slope was slightly shallower for p.R1451L channels (Fig. , Table ).
Family 1 The proband is a 19-year-old Chinese male presenting with a 15 year history of recurrent quadriplegia associated with hypokalaemia or normokalaemia (Fig. ). He experienced his first episode after a high fever at the age of 4. After he turned 7 the episodes occurred 2–3 times per year and usually lasted for several hours to 2 days with spontaneous full recovery. The attacks became more frequent when he was 16. Generalised limb muscle weakness occurred once or twice per month and isolated lower limb weakness 4–5 times per month. The ability to swallow and speak was well preserved. Triggers include rest after a period of vigorous exercise, exposure to cold, upper respiratory infection, and maintenance of a fixed posture. Serum potassium level was 2.8, 3.3 and 4.0 mmol/L (normal range: 3.5–5.5 mmol/L) measured in three distinct episodes of paralysis. None of oral administration with potassium, carbamazepine and topiramate can reduce the attacks or ameliorate the severity. Additionally, physical examination revealed a generalised increase in muscle bulk, especially in bilateral quadriceps and calf muscles. The patient didn’t complain of clinical myotonia, but a decreased ability to relax after a forced eyelid closure was demonstrated. Neither grip myotonia nor percussion myotonia was detected. Muscle strength and the tendon reflex were normal. After coenzyme Q10 administration for 2 months (200 mg/day), the patient reported that recurrent muscle weakness was greatly relieved. The effect of coenzyme Q10 treatment on eyelid myotonia was not assessed. Next generation sequencing revealed the homozygous p.R1451L mutation in the skeletal muscle Na V 1.4 sodium channel. No other causal mutations were detected in a panel of 10 ion channel genes and 245 primary myopathies/muscular dystrophies-related genes (Supplementary Table ). The parents are consanguineously married carriers (Fig. ) of the heterozygous p.R1451L mutation. Their running ability is unaffected and they do not display symptoms of clinical myotonia although the muscle volume is relatively increased. A 22 years old sister with a heterozygous p.R1451L mutation complained of muscle stiffness after a short rest. The father had no complaints of muscle weakness. Clinical PP was not observed in the heterozygous carriers of p.R1451L. Family 2 A 23 year old man presented to the national referral centre for skeletal muscle channelopathies in the UK with symptoms of episodic muscle weakness. His birth history, motor milestones and childhood were normal with no specific symptoms. At the age of 18 years after a period of prolonged sitting he experienced bilateral leg weakness that lasted for several hours and spontaneously resolved. From this point he experienced recurrent episodes of limb muscle weakness and at the time of referral these were occurring daily. They could occur at any time of the day although a significant number occurred first thing in the morning with less frequent episodes during the night. Symptoms lasted typically from 10 mins to a few hours and could be exacerbated by cold weather, exercise or prolonged rest. He did not identify any specific food triggers. Three episodes all occurring from sleep had been severe enough to warrant hospital attendance. Serum potassium was recorded within the normal range on each occasion. In addition to the attacks of weakness he reported his hands could become “stuck”, particularly in cold weather. Examination revealed him to have a muscular physique, which was notable relative to a general lack of exercise. There was mild hand and grip myotonia that was exacerbated by repetition, but the remainder of the exam was unremarkable. He reported a family history of affected father with similar symptoms although had no regular contact with him. Although the diurnal pattern is reminiscent of hypoPP, all other clinical features were considered consistent with hyperPP and the patient was treated with a combination of acetazolamide 500 mg BD and bendroflumethiazide 5 mg OD. This reduced although did not abolish his attacks of paralysis. There was no effect on myotonia. Genetic analysis identified the heterozygous SCN4A mutation p.R1451L in both the proband and his father.
The proband is a 19-year-old Chinese male presenting with a 15 year history of recurrent quadriplegia associated with hypokalaemia or normokalaemia (Fig. ). He experienced his first episode after a high fever at the age of 4. After he turned 7 the episodes occurred 2–3 times per year and usually lasted for several hours to 2 days with spontaneous full recovery. The attacks became more frequent when he was 16. Generalised limb muscle weakness occurred once or twice per month and isolated lower limb weakness 4–5 times per month. The ability to swallow and speak was well preserved. Triggers include rest after a period of vigorous exercise, exposure to cold, upper respiratory infection, and maintenance of a fixed posture. Serum potassium level was 2.8, 3.3 and 4.0 mmol/L (normal range: 3.5–5.5 mmol/L) measured in three distinct episodes of paralysis. None of oral administration with potassium, carbamazepine and topiramate can reduce the attacks or ameliorate the severity. Additionally, physical examination revealed a generalised increase in muscle bulk, especially in bilateral quadriceps and calf muscles. The patient didn’t complain of clinical myotonia, but a decreased ability to relax after a forced eyelid closure was demonstrated. Neither grip myotonia nor percussion myotonia was detected. Muscle strength and the tendon reflex were normal. After coenzyme Q10 administration for 2 months (200 mg/day), the patient reported that recurrent muscle weakness was greatly relieved. The effect of coenzyme Q10 treatment on eyelid myotonia was not assessed. Next generation sequencing revealed the homozygous p.R1451L mutation in the skeletal muscle Na V 1.4 sodium channel. No other causal mutations were detected in a panel of 10 ion channel genes and 245 primary myopathies/muscular dystrophies-related genes (Supplementary Table ). The parents are consanguineously married carriers (Fig. ) of the heterozygous p.R1451L mutation. Their running ability is unaffected and they do not display symptoms of clinical myotonia although the muscle volume is relatively increased. A 22 years old sister with a heterozygous p.R1451L mutation complained of muscle stiffness after a short rest. The father had no complaints of muscle weakness. Clinical PP was not observed in the heterozygous carriers of p.R1451L.
A 23 year old man presented to the national referral centre for skeletal muscle channelopathies in the UK with symptoms of episodic muscle weakness. His birth history, motor milestones and childhood were normal with no specific symptoms. At the age of 18 years after a period of prolonged sitting he experienced bilateral leg weakness that lasted for several hours and spontaneously resolved. From this point he experienced recurrent episodes of limb muscle weakness and at the time of referral these were occurring daily. They could occur at any time of the day although a significant number occurred first thing in the morning with less frequent episodes during the night. Symptoms lasted typically from 10 mins to a few hours and could be exacerbated by cold weather, exercise or prolonged rest. He did not identify any specific food triggers. Three episodes all occurring from sleep had been severe enough to warrant hospital attendance. Serum potassium was recorded within the normal range on each occasion. In addition to the attacks of weakness he reported his hands could become “stuck”, particularly in cold weather. Examination revealed him to have a muscular physique, which was notable relative to a general lack of exercise. There was mild hand and grip myotonia that was exacerbated by repetition, but the remainder of the exam was unremarkable. He reported a family history of affected father with similar symptoms although had no regular contact with him. Although the diurnal pattern is reminiscent of hypoPP, all other clinical features were considered consistent with hyperPP and the patient was treated with a combination of acetazolamide 500 mg BD and bendroflumethiazide 5 mg OD. This reduced although did not abolish his attacks of paralysis. There was no effect on myotonia. Genetic analysis identified the heterozygous SCN4A mutation p.R1451L in both the proband and his father.
Electromyographic (EMG) studies of the proband in family 1 demonstrated myotonic discharges in a number of tested muscles (Table ; Deltoids, Flex Carpi Rad, Vastus Med, Gastroc caput med and Tibialis anterior). Of note, poor recruitment of muscle unit potential was demonstrated in the rectus abdominis and Vastus Med. Myotonic discharges are documented in some selected muscles for the parents and the sister (Table ). Repeat EMG was not performed following administration of Coenzyme Q10. Long exercise test (LET) revealed a significant decline of >30% of baseline amplitude value for the proband (Fig. ). After two months oral administration of Coenzyme Q10, compound muscle action potential (CMAP) decrement >30% occurred at the same time, but the baseline value of long exercise test increased from 3.9 mV to 5.7 mV in abductor digiti minimi (ADM) muscle (Fig. ). For the father and the sister, a significant decline of CMAP amplitude in LET was observed 25 minutes and 10 minutes after exercise, respectively. For the mother, the CMAP amplitude and area slightly declined and did not reach a 30% threshold (Fig. ). The proband in family 2 had evidence of myotonic runs in the first dorsal interosseous and tibialis anterior. A long exercise test was positive for periodic paralysis with a decrement of CMAP amplitude from the baseline of 67%.
As hypoPP is associated with gating pore currents caused by mutations that affect S4 arginines we first analysed if p.R1451L mutant channels conducted gating pore currents using the Xenopus laevis oocyte expression system. To isolate the gating pore currents the main pore was blocked with 1–2 µM tetrodotoxin. The amplitude of the gating pore current of S4 arginine mutant channels is increased when guanidinium acts as charge carrier , . However, when 50% of the extracellular sodium was substituted by guanidinium, steady state current amplitude did not differ between p.R1451L and wild-type channels at physiological voltages (Fig. ), whereas currents in known hypoPP mutant p.R222W were significantly larger (one way ANOVA, Dunnet multiple comparisons, compared at −80 mV). These data suggest that p.R1451L channels do not conduct gating pore currents.
We studied the main pore sodium current of the p.R1451L mutant in HEK293 cells (Figs , , Table ). Consistent with previous analysis of p.R1451L properties , the peak current amplitude of the mutant channel was significantly reduced compared to the wild-type channel (Fig. ), the voltage of half-maximal fast inactivation was shifted to more hyperpolarized voltages and the slope was more shallow (Fig. ). We didn’t observe changes in the voltage of half-maximal channel activation, although the slope of p.R1451L activation was significantly more shallow than for wild-type channels (Fig. ). The voltage dependence of activation and fast inactivation were similar when the holding voltage (V h ) was −80 or −100 mV (Table ). We further analysed the defective fast inactivation at two physiological holding voltages (Fig. , Table ). With V h −100 mV, the onset of closed-state inactivation at −60 mV and −80 mV was significantly accelerated for p.R1451L channels compared to wild-type channels, consistent with the left-shift in the voltage dependence of fast inactivation (Fig. ). In contrast, the onset of p.R1451L open-state inactivation was decelerated compared to wild-type channels in the studied voltage range −20 mV to +20 mV (Fig. ). The rate of closed-state inactivation at −60 mV was accelerated for wild-type channels when the holding voltage was −80 mV, compared to data with V h = −100 mV (Supplemental Fig. ), but was not analysed for p.R1451L channel as at this V h more than 50% of the channels are already inactivated. The rate of open-state inactivation was unaffected by the holding voltage (Table ). The rate of p.R1451L channel recovery from inactivation was 3 times faster than for the wild-type channel at −80 mV (p.R1451 L τ = 1.9 ms vs τ = 6.1 for wild-type channel, Fig. ). At −100 mV the rate of recovery was accelerated for both channels and although the fold difference was reduced, the recovery of p.R1451L channels was still significantly faster than for wild-type channels (p.R1451L τ = 1.6 ms vs τ = 2.4 for WT, Fig. ). A prediction of the accelerated recovery from inactivation and the reduced rate of open-state inactivation of p.R1451L channels is that upon repetitive high-frequency stimulation the availability of the mutant channels reduces less than for wild-type channels. Consistently, when a 2 ms pulse to 0 mV was given from a holding voltage of −80 mV at frequencies 40–200 Hz the current declined more for wild-type than for mutant channels (Fig. ). When V h was −100 mV the difference in availability between p.R1451L and wild-type channels was only observed at very high frequency (200 Hz), probably reflecting the reduced difference in the rate of recovery from inactivation of both channels at this voltage. Finally, the p.R1451L mutation did not alter the voltage of half-maximal slow inactivation of the Na V 1.4 channel although the slope was slightly shallower for p.R1451L channels (Fig. , Table ).
We present heterozygous and homozygous cases of the same SCN4A mutation p.R1451L with differing clinical features. Our homozygous case experienced hypoPP and myotonia. These conditions have distinct molecular pathomechanisms and are rarely reported to co-exist in an individual , so much so that the presence of myotonia in a patient is often used to exclude diagnosis of hypoPP. Heterozygous carriers of this mutation, presented here and in a previous study , fall within a myotonia-hyperPP spectrum of disorders. The myotonia of the heterozygous carriers can be triggered by cold and exacerbated by repetition, consistent with a diagnosis of PMC. An episode of hypokalaemia-associated PP was reported in a p.R1451L heterozygous carrier . The p.R1451L channel displays gain-of-function-properties that explain the myotonia and hyperPP in carriers of p.R1451L: the rate of open state inactivation was reduced and the rate of recovery from inactivation was enhanced at physiological voltages. Consequently, p.R1451L channels showed improved channel availability upon high frequency repetitive simulation. Accelerated recovery from inactivation or changes in channel availability upon repetitive stimulation were not previously reported for this mutant , probably reflecting the fact that these were measured only at very hyperpolarized voltages where the recovery rate may saturate for both p.R1451L and wild-type channels. Heterozygous p.R1451L carriers show a range of predominant presentations within the PMC-hyperPP spectrum. Similar range of manifestations has been reported for other Na V 1.4 mutations , but the mechanism determining the predominant manifestation remains to be characterized. The changes in the properties of the p.R1451L mutant resemble channels with mutations that affect the outermost arginine residue in the S4 helix of VSD-IV, R1448 , . These mutations impart gain-of-function properties to Na V 1.4 by reducing the rate of open state inactivation and by accelerating the rate of recovery from inactivation, but they also left-shift the voltage dependence of fast inactivation, as described here for p.R1451L. In accordance, the heterozygous carriers of these mutations are within the PMC-hyperPP spectrum, similar to heterozygous carriers of the p.R1451L mutation. The predominant symptom of the homozygous carrier of p.R1451L carrier is PP associated with hypo- or normokalaemia. Gating pore currents could not be detected in p.R1451L channels, consistent with existing data for p.R1451H channels , suggesting that depolarisation of the muscle by gating pore currents does not underlie hypoPP in our patient. It has been shown that recessive Na V 1.4 loss-of-function mutations can present with episodes of paralysis or with “lifelong episodic generalized weakness” . These mutations are associated with congenital myasthenia and cause enhanced fast inactivation and affect arginine residues R1454 and R1457 , both adjacent to R1451 in the S4 helix of VSD-IV. Mutation p.R1451L also enhances channel inactivation, resulting in increased susceptibility of the muscle to depolarisation-induced reduction in Na V 1.4 channel availability. Thus, a small depolarisation, potentially induced by hypokalaemia, may reduce p.R1451L channel availability to a level insufficient to sustain muscle tone in the homozygous carrier, resulting in flaccid paralysis. Other loss-of-function features, such as reduced current density contribute towards the reduced channel availability and the weakness in the homozygous patient. Also, although we did not detect changes in voltage of half-maximal slow inactivation, a reduction in the rate of recovery from slow inactivation at −120 mV was recently reported for p.R1451L channels . This loss-of-function feature may also contribute towards the clinical presentation. However, unlike the carriers of other homozygous or compound heterozygous Na V 1.4 mutations that display PP, the homozygous carrier of p.R1451L does not display myasthenic symptoms. Despite left-shifted voltage dependence of fast inactivation the recovery rate of p.R1451L channels from fast inactivation is enhanced, contrary to that described for congenital myasthenia mutations. Consequently, upon repetitive stimulation, congenital myasthenia variants accumulate more in the inactivated state than wild-type channels , , , while p.R1451L channels accumulate less (Fig. ). Enhanced accumulation in the inactivated state has been linked with enhanced reduction in CMAP upon repetitive stimulation in carriers of the myasthenic variants , . The rate of open-state inactivation is reduced for all p.R1454W, p.R1457H and p.R1451L mutant channels, suggesting that this feature cannot qualitatively explain the absence of myasthenic syndromes in the homozygous p.R1451L patient. Thus, our data pinpoints the rate of recovery from fast inactivation as a key determinant of clinical presentation in these cases and highlights R1451 as a pivotal residue controlling the speed of recovery from inactivation. The structure of electric eel Na V 1.4 suggests that a mutation of R1451 disrupts its electrostatic interaction with E1373 (Fig. ). This interaction may stabilise the ‘up’-state of VSD-IV captured in the Na V 1.4 structure, and, in accordance, the p.R1451L mutation may destabilize the ‘up’-state and allow hastened recovery of the voltage sensor to the resting state, accounting for the faster rate of recovery from inactivation. Consistently, neutralising or charge-reversing mutations of E1373 also accelerate the rate of recovery from inactivation . An episode of hypoPP was reported for a single heterozygous p.R1451L carrier suggesting that, in rare cases, heterozygous loss-of-function of Na V 1.4 may increase susceptibility to hypokalaemia induced PP. However, heterozygous carriers of Na V 1.4 loss-of-function variants, including null variants, have been reported asymptomatic , – suggesting that loss-of-function affecting a single allele does not reduce Na V 1.4 availability sufficiently to cause fixed or fluctuating weakness. Consistently, mice with heterozygous Na V 1.4 null allele do not show increased susceptibility to periodic paralysis . This suggests that loss-of-function features in a single allele very rarely underlie neurological presentations, at least in the absence of other factors affecting muscle excitability. Reports of bi-allelic gain-of-function variants are rare. In a carrier of the homozygous PMC mutation, p.I1393T, no episodes of PP associated with low potassium were reported . Instead, the patient suffered from severe myotonia. Mice that carry homozygous gain-of-function mutations did not survive beyond day two, with noted breathing difficulties likely associated with breathing muscles or showed greatly impaired embryonic or early postnatal survival , potentially explaining the rarity of such cases in humans. Thus, it is unlikely that the gain-of-function features are solely responsible for the PP associated with hypokalaemia in the homozygous carrier of p.R1451L mutation. However, these may contribute towards the clinical presentation. Cold, that slows the rate of inactivation of both WT and p.R1451L channels , was reported as one of the triggers of weakness in the homozygous carrier, supporting a role for this gain-of-function feature in the clinical presentation. Our data are consistent with a model where in heterozygosis the gain-of-function features of p.R1451L cause symptoms on the PMC-hyperPP overlap, while in homozygosis the loss-of-function features underlie hypo-/normoPP. However, available clinical detail cannot confirm the pathomechanism of the weakness in the p.R1451L homozygotic patient and it is possible that gain-of-function properties of p.R1451L contribute to this manifestation. Other intrinsic and extrinsic factors may also influence the range of clinical manifestations in the heterozygous and homozygous carriers of this mutation. Co-existence of recessive loss-of-function properties with dominant gain-of-function properties can explain the unusual overlap of symptoms in the homozygous carrier. Curiously, despite severely reduced availability of Na V 1.4 channels in this patient, the gain-of-function properties of p.R1451L (i.e. enhanced recovery from inactivation) can still induce eyelid and EMG myotonia. Our data also suggests that hypoPP can be the predominant presentation of bi-allelic loss-of-function variants and that it can manifest without accompanying myasthenia. Finally, coenzyme Q10, a component of the electron transport chain, ameliorated recurrent muscle weakness of the patients and increased the CMAP baseline amplitude value. The mechanism of action of coenzyme Q10 in improving the hypoPP symptoms is unclear but our data warrants consideration of its clinical benefits for hypoPP patients .
Standard protocol approvals and patient consents The study of family 1 was approved and all experiments were performed under the guidelines of the Institutional Ethics Committee of the Huashan Hospital. The patient and the family members provided written informed consent for clinical-genetic correlation studies. For family 2, all procedures were performed as part of routine clinical care. The study was conducted under the ethics guidelines issued by our institution and approved by the London – Queen Square NRES Committee (reference07/Q0512/26), with written informed consent obtained from all participants for genetic and research studies. Electromyographic studies and exercise test protocols For family 1, needle EMG studies were conducted in 11 muscles for each case. The LET and SET were performed following the Fournier protocol . The ulnar nerve was stimulated supramaximally at the wrist, and the CMAP was recorded from the ADM every 30 s for 2 min to obtain a stable baseline amplitude value. The patient was required for forceful abduction of the small finger for 5 min with a 5-s rest period every 45 s. The CMAPs were recorded at 1-min intervals for the first 5 min and then at 5-min intervals for 40 min. The baseline-to-negative peak amplitude of each CMAP was expressed as a percentage of the baseline value before exercise. SET was performed on the small finger after the forceful abduction for 10 seconds, CMAPs were recorded 2 seconds immediately after the end of exercise and then every 10 seconds for 50 seconds. Screening of mutations by targeted next generation sequencing For family 1 a targeted next generation sequencing panel comprising 10 ion channel genes and 245 primary myopathies/muscular dystrophies-related genes (Suppl. Table ) was employed to screen the causal mutations. Genomic DNA from peripheral blood was extracted with High Pure PCR Template Preparation Kit (Roche, Basel,CH) according to the manufacturer’s instructions. DNA fragments were enriched by solution-based hybridization capture and followed by sequencing with an Illumina Miseq platform (Illumina, San Diego, CA, USA) with the 2 × 300 bp paired-end read module. The hybridization capture procedure was performed with the SureSelect Library Prep Kit (Agilent, Santa Clara, CA, USA). For further verification of the candidate mutations, we performed the Sanger sequencing with the DNA samples extracted from the patient and the parents. PCR was performed with GoldStar Taq DNA Polymerase (CWbiotech) according to standard protocol. PCR products were sequenced on an ABI3730xl DNA Analyzer (Applied Biosystems). Public databases including dbSNP138, 1000 Genome project, Exome Sequencing Project, ExAC, ClinicVar and HGMD were used to screen variants. Prediction of functional effect was evaluated by PolyPhen-2 and SIFT scores. To detect CNV (Copy number variant), sequencing depth of each region covered by the probes was calculated according to the alignment files. ExomeDepth Package was also used to find potential CNVs. For family 2 genetic analysis of ion channel genes SCN4A , CACNA1S , KCNJ2 and CLCN1 was performed at the Neurogenetics Unit, National Hospital for Neurology and Neurosurgery as provided by the Channelopathy Highly Specialized National Service for rare disease. Samples underwent Next-Generation Sequencing on an Illumina HiSeq following enrichment with an Illumina custom Nextera Rapid Capture panel (Illumina, Inc., San Diego, CA). Molecular Biology and cell preparation Human and rat SCN4A constructs were used as templates for site directed mutagenesis using the QuickChange kit (Agilent Technologies). The target mutation p.R1451L (human)/p.R1444L (rat) and insert sequence integrity were verified by sequencing. In vitro transcription was performed with the mMessage mMachine kit (Ambion). HEK293 cells were cotransfected with h SCN4A (500 ng) and GFP (50 ng) cDNAs using 1.5 µl Lipofectamine 2000 (Gibco) on a 1.9 cm 2 dish. Xenopus laevis ovarian lobes were obtained according to procedures approved by the UCL Biological Services and the UK Home Office. Oocytes were defolliculated with enzymatic digestion using Collagenase A (2 mg/ml, Roche) in oocyte Ringer (in mM: NaCl 82.5, KCl 2, MgCl 2 , HEPES 5, pH 7.5–6) and stored in modified Barth’s Solution (in mM: NaCl 87.1, KCl 1, MgSO 4 1.68, HEPES 10, NaNO 3 0.94, NaHCO 3 2.4, CaCl 2 0.88, pH 7.4) supplemented with penicillin (50 U/ml), streptomycin (50 μg/ml) and amikacin (100 μg/ml) at 14–18 °C. Selected oocytes were injected with RNA for rat SCN4A and rat SCN1B (~50 ng each). Electrophysiology Gating pore currents through Na V 1.4 were studied in Xenopus oocytes by two-electrode voltage clamp with a GeneClamp 500B amplifier, Digidata 1200 digitizer and pCLAMP™ software (Molecular Devices). Electrodes had a resistance of 0.1–0.7 MΩ when filled with KCl 3 M. Oocytes were bathed in a solution containing (in mM): 120 NaMeSO 4 , 1.8 CaSO 4 , 10 HEPES, pH 7.4. Currents were elicited using 300 ms voltage steps from −140 to +50 mV in 5 mV increments from a holding potential of −100 mV. Sampling frequency was 10 kHz. TTX 1–2 μM was added to the bath to block Na + currents through the main channel pore and presence of gating pore currents analysed by replacing 50% of NaMeSO 4 with guanidine sulfate. Gating pore currents were measured as the average steady state current in the last 50 ms of the test pulse and plotted against membrane voltage. Leak currents were measured by fitting a straight line to the current-voltage data in the range + 10 mV to +30 mV. The fitted line was extrapolated to cover the voltage range of the measurements and subtracted from the raw current data. The properties of Na V 1.4 main pore currents were studied with whole cell patch clamp recordings from HEK293 cells 48 hours post-transfection. Data were sampled at 50 Hz and filtered at 5 Hz using Axopatch 200B, Digidata 1440B and pClamp software (Molecular Devices). Electrodes (1–3 MΩ) were pulled from borosilicate glass and filled with the following intracellular solution (in mM): CsCl 145, NaCl 5, EGTA 10, HEPES 10 (pH 7.4). Bath solution was (in mM): NaCl 145, KCl 4, CaCl 2 1.8, MgCl 2 1, and HEPES 10 (pH 7.35). The calculated liquid junction potential of −4.4 mV was not corrected for. Series resistance was compensated ≥ 70% to keep the voltage error below 5 mV. Holding potential was −80 mV or −100 mV. Leak and capacitance subtraction was performed online using a - P /4 procedure for all protocols except for pulse trains and for slow inactivation. The voltage protocol to measure the voltage dependence of activation and fast inactivation consists of a 150 ms step to test voltages ranging from −150 mV to +50 mV in 10 mV increments, followed by a step to a tail voltage of −10 mV. Peak current in response to test voltages was used to analyse voltage dependence of activation. Peak current in response to the tail voltage was used to measure the voltage dependence of inactivation. Peak conductance of activation was derived by dividing the peak current by test voltage subtracted by reversal voltage. Reversal voltage was estimated for each individual cell by extrapolating a straight line fitted to the voltage range of +20 to +50 mV to I = 0 pA. Voltage of half-maximal activation or inactivation (V 1/2 ) and the slope factor K , were measured by fitting with a Boltzmann equation (Y = A + (B-A)/(1 + exp((V-V 1/2 )/ K )) where Y is conductance or current, and A and B are the minimum and maximum amplitudes of the fit. Time course of onset of open state fast inactivation was estimated using the same protocol by fitting a double exponential function to current decay. Only the main component of the double exponential function was analysed. Time course of closed state inactivation was studied by applying pre-pulses of increasing duration to −60 mV or −80 mV, followed by a test pulse to 0 mV. Time course of recovery from fast inactivation was studied by inactivating the channels for 20 ms at 0 mV, applying recovery-pulses of increasing duration to −80 mV or −100 mV, before applying the second pulse to 0 mV. The peak current (closed state inactivation) or the ratio of peak currents in response to second and first test pulse (recovery from inactivation) is plotted against the duration of pre- or recovery-pulse. For analysis of responses to repetitive stimulation, trains of 25 pulses of 2 ms duration to 0 mV were applied with varying frequencies, ranging from 20 Hz to 400 Hz. Voltage dependence of slow inactivation was studied by applying 10 s pre-pulse steps to voltages ranging from −130 mV to +50 mV in 10 mV increments. This was followed by a 20 ms pulse to −100 mV to recover channels from fast inactivation and a test pulse to −10 mV. Peak currents in response to the test pulse were measured and plotted against pre-pulse voltage. Data analysis Current records were analyzed using Clampfit 10.7, OriginPro 2016 and GraphPad softwares. All data is presented as mean ± SEM. Unpaired t tests were used for statistical comparison of biophysical properties (with Welch correction for parameters with unequal variance) or non-parametric (Mann-Whitney) tests if data was not normally distributed. One way ANOVA and Dunnet multiple comparisons test was used for gating pore current data. Significance was established at p < 0.05. Data availability The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
The study of family 1 was approved and all experiments were performed under the guidelines of the Institutional Ethics Committee of the Huashan Hospital. The patient and the family members provided written informed consent for clinical-genetic correlation studies. For family 2, all procedures were performed as part of routine clinical care. The study was conducted under the ethics guidelines issued by our institution and approved by the London – Queen Square NRES Committee (reference07/Q0512/26), with written informed consent obtained from all participants for genetic and research studies.
For family 1, needle EMG studies were conducted in 11 muscles for each case. The LET and SET were performed following the Fournier protocol . The ulnar nerve was stimulated supramaximally at the wrist, and the CMAP was recorded from the ADM every 30 s for 2 min to obtain a stable baseline amplitude value. The patient was required for forceful abduction of the small finger for 5 min with a 5-s rest period every 45 s. The CMAPs were recorded at 1-min intervals for the first 5 min and then at 5-min intervals for 40 min. The baseline-to-negative peak amplitude of each CMAP was expressed as a percentage of the baseline value before exercise. SET was performed on the small finger after the forceful abduction for 10 seconds, CMAPs were recorded 2 seconds immediately after the end of exercise and then every 10 seconds for 50 seconds.
For family 1 a targeted next generation sequencing panel comprising 10 ion channel genes and 245 primary myopathies/muscular dystrophies-related genes (Suppl. Table ) was employed to screen the causal mutations. Genomic DNA from peripheral blood was extracted with High Pure PCR Template Preparation Kit (Roche, Basel,CH) according to the manufacturer’s instructions. DNA fragments were enriched by solution-based hybridization capture and followed by sequencing with an Illumina Miseq platform (Illumina, San Diego, CA, USA) with the 2 × 300 bp paired-end read module. The hybridization capture procedure was performed with the SureSelect Library Prep Kit (Agilent, Santa Clara, CA, USA). For further verification of the candidate mutations, we performed the Sanger sequencing with the DNA samples extracted from the patient and the parents. PCR was performed with GoldStar Taq DNA Polymerase (CWbiotech) according to standard protocol. PCR products were sequenced on an ABI3730xl DNA Analyzer (Applied Biosystems). Public databases including dbSNP138, 1000 Genome project, Exome Sequencing Project, ExAC, ClinicVar and HGMD were used to screen variants. Prediction of functional effect was evaluated by PolyPhen-2 and SIFT scores. To detect CNV (Copy number variant), sequencing depth of each region covered by the probes was calculated according to the alignment files. ExomeDepth Package was also used to find potential CNVs. For family 2 genetic analysis of ion channel genes SCN4A , CACNA1S , KCNJ2 and CLCN1 was performed at the Neurogenetics Unit, National Hospital for Neurology and Neurosurgery as provided by the Channelopathy Highly Specialized National Service for rare disease. Samples underwent Next-Generation Sequencing on an Illumina HiSeq following enrichment with an Illumina custom Nextera Rapid Capture panel (Illumina, Inc., San Diego, CA).
Human and rat SCN4A constructs were used as templates for site directed mutagenesis using the QuickChange kit (Agilent Technologies). The target mutation p.R1451L (human)/p.R1444L (rat) and insert sequence integrity were verified by sequencing. In vitro transcription was performed with the mMessage mMachine kit (Ambion). HEK293 cells were cotransfected with h SCN4A (500 ng) and GFP (50 ng) cDNAs using 1.5 µl Lipofectamine 2000 (Gibco) on a 1.9 cm 2 dish. Xenopus laevis ovarian lobes were obtained according to procedures approved by the UCL Biological Services and the UK Home Office. Oocytes were defolliculated with enzymatic digestion using Collagenase A (2 mg/ml, Roche) in oocyte Ringer (in mM: NaCl 82.5, KCl 2, MgCl 2 , HEPES 5, pH 7.5–6) and stored in modified Barth’s Solution (in mM: NaCl 87.1, KCl 1, MgSO 4 1.68, HEPES 10, NaNO 3 0.94, NaHCO 3 2.4, CaCl 2 0.88, pH 7.4) supplemented with penicillin (50 U/ml), streptomycin (50 μg/ml) and amikacin (100 μg/ml) at 14–18 °C. Selected oocytes were injected with RNA for rat SCN4A and rat SCN1B (~50 ng each).
Gating pore currents through Na V 1.4 were studied in Xenopus oocytes by two-electrode voltage clamp with a GeneClamp 500B amplifier, Digidata 1200 digitizer and pCLAMP™ software (Molecular Devices). Electrodes had a resistance of 0.1–0.7 MΩ when filled with KCl 3 M. Oocytes were bathed in a solution containing (in mM): 120 NaMeSO 4 , 1.8 CaSO 4 , 10 HEPES, pH 7.4. Currents were elicited using 300 ms voltage steps from −140 to +50 mV in 5 mV increments from a holding potential of −100 mV. Sampling frequency was 10 kHz. TTX 1–2 μM was added to the bath to block Na + currents through the main channel pore and presence of gating pore currents analysed by replacing 50% of NaMeSO 4 with guanidine sulfate. Gating pore currents were measured as the average steady state current in the last 50 ms of the test pulse and plotted against membrane voltage. Leak currents were measured by fitting a straight line to the current-voltage data in the range + 10 mV to +30 mV. The fitted line was extrapolated to cover the voltage range of the measurements and subtracted from the raw current data. The properties of Na V 1.4 main pore currents were studied with whole cell patch clamp recordings from HEK293 cells 48 hours post-transfection. Data were sampled at 50 Hz and filtered at 5 Hz using Axopatch 200B, Digidata 1440B and pClamp software (Molecular Devices). Electrodes (1–3 MΩ) were pulled from borosilicate glass and filled with the following intracellular solution (in mM): CsCl 145, NaCl 5, EGTA 10, HEPES 10 (pH 7.4). Bath solution was (in mM): NaCl 145, KCl 4, CaCl 2 1.8, MgCl 2 1, and HEPES 10 (pH 7.35). The calculated liquid junction potential of −4.4 mV was not corrected for. Series resistance was compensated ≥ 70% to keep the voltage error below 5 mV. Holding potential was −80 mV or −100 mV. Leak and capacitance subtraction was performed online using a - P /4 procedure for all protocols except for pulse trains and for slow inactivation. The voltage protocol to measure the voltage dependence of activation and fast inactivation consists of a 150 ms step to test voltages ranging from −150 mV to +50 mV in 10 mV increments, followed by a step to a tail voltage of −10 mV. Peak current in response to test voltages was used to analyse voltage dependence of activation. Peak current in response to the tail voltage was used to measure the voltage dependence of inactivation. Peak conductance of activation was derived by dividing the peak current by test voltage subtracted by reversal voltage. Reversal voltage was estimated for each individual cell by extrapolating a straight line fitted to the voltage range of +20 to +50 mV to I = 0 pA. Voltage of half-maximal activation or inactivation (V 1/2 ) and the slope factor K , were measured by fitting with a Boltzmann equation (Y = A + (B-A)/(1 + exp((V-V 1/2 )/ K )) where Y is conductance or current, and A and B are the minimum and maximum amplitudes of the fit. Time course of onset of open state fast inactivation was estimated using the same protocol by fitting a double exponential function to current decay. Only the main component of the double exponential function was analysed. Time course of closed state inactivation was studied by applying pre-pulses of increasing duration to −60 mV or −80 mV, followed by a test pulse to 0 mV. Time course of recovery from fast inactivation was studied by inactivating the channels for 20 ms at 0 mV, applying recovery-pulses of increasing duration to −80 mV or −100 mV, before applying the second pulse to 0 mV. The peak current (closed state inactivation) or the ratio of peak currents in response to second and first test pulse (recovery from inactivation) is plotted against the duration of pre- or recovery-pulse. For analysis of responses to repetitive stimulation, trains of 25 pulses of 2 ms duration to 0 mV were applied with varying frequencies, ranging from 20 Hz to 400 Hz. Voltage dependence of slow inactivation was studied by applying 10 s pre-pulse steps to voltages ranging from −130 mV to +50 mV in 10 mV increments. This was followed by a 20 ms pulse to −100 mV to recover channels from fast inactivation and a test pulse to −10 mV. Peak currents in response to the test pulse were measured and plotted against pre-pulse voltage.
Current records were analyzed using Clampfit 10.7, OriginPro 2016 and GraphPad softwares. All data is presented as mean ± SEM. Unpaired t tests were used for statistical comparison of biophysical properties (with Welch correction for parameters with unequal variance) or non-parametric (Mann-Whitney) tests if data was not normally distributed. One way ANOVA and Dunnet multiple comparisons test was used for gating pore current data. Significance was established at p < 0.05.
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
Supplementery information
|
Patient and Caregiver Education to Support Self‐Efficacy and Self‐Management During Immunotherapy—An Integrative Review | e455e6a6-3772-48f5-99d0-e1f4c8e005ec | 11865008 | Patient Education as Topic[mh] | Background The first immune checkpoint inhibitor (ICI) therapy, often referred to as immunotherapy, was approved for the treatment of metastatic melanoma in 2011 . ICIs block negative signaling pathways in the immune system enabling the T‐cells to kill cancer cells . The side effects triggered by ICIs differ significantly from the side effect profiles associated with classic cytostatic cancer treatments. This distinction is inherent to the mechanism of action of ICIs. The interference with the immune system, blocking essential immune checkpoints (e.g., cytotoxic T‐lymphocyte‐associated protein 4, programmed cell death protein 1, or programmed death‐ligand 1), can induce an immune response to the organs of the body, mimicking autoimmune diseases . The side effects caused by ICIs are referred to as immune‐related adverse events (irAEs), and their occurrence is unpredictable, but seen most during the first 3–4 months after ICI initiation . A variety of cancer diseases are now treated with ICI as monotherapy, as ICI combination therapy, or in combination with other antineoplastic treatments . The risk of severe irAEs varies between 15% and 59% depending on the type of ICI and whether it is given as mono‐ or combination therapy . IrAEs are potentially life‐threatening, can necessitate admission to the emergency department or hospitalizations, influence treatment duration , and affect the quality of life for patients and caregivers . Patient and caregiver education may ensure identification and prevention of irAEs at an early stage, preventing them from becoming severe . Patient education is the process of influencing patient behavior and producing the changes in knowledge, attitudes and skills necessary to maintain or improve health . In this review, we define patient education on ICI as the process of providing the patient and the appointed family caregivers with the necessary knowledge about ICI, its efficacy, irAEs, and the management of irAEs. Being diagnosed with cancer is highly distressing for patients and family caregivers . The amount of information correctly recalled by patients about the disease and treatment is strikingly small impacting the management of irAEs . Furthermore, the level of health literacy among patients can influence their ability to understand and manage treatment, impacting their safety throughout the cancer care pathway . Health literacy is the ability to obtain, process and understand medical information . Family caregivers can help patients recall the information received during consultations and discuss key aspects of cancer treatment with them , serving an important role as co‐receiver of information. Moreover, studies show that the level of self‐efficacy aligns with symptom occurrence, and high self‐efficacy among patients with cancer is associated with low symptom occurrence and symptom distress . Self‐efficacy is the conviction that one can successfully execute the behavior required to produce certain outcomes . Self‐efficacy empowers individuals to cope or self‐manage with all that a chronic or life‐threatening condition, such as cancer, entails . Thus, high levels of self‐efficacy may lead to better self‐management , including more attention to irAEs. Self‐management in this context is defined as the ability to manage the disease and treatment effects and psychosocial changes arising as a result of illness . Previous reviews examined different ways to provide information to patients with cancer. They focused on pre‐consultation interventions , the use of mobile health devices and applications (mHealth) , and the effects of virtual reality as a patient education tool. To the best of our knowledge, an integrative review regarding education on ICI for patients with cancer including their family caregivers has not been conducted. Furthermore, we identified no previous review regarding how education on ICI affects patients' and family caregivers' self‐efficacy and self‐management of irAEs. 1.1 Objective This integrative review aimed to elucidate the existing knowledge of education of patients with cancer receiving ICI therapy and their family caregivers, focusing on patients' and caregivers' self‐efficacy and self‐management of irAEs. The following research question guided the review: Research Question: How does patient education on ICI efficacy and toxicity management affect patients with cancer and their family caregivers' self‐efficacy and self‐management when dealing with immune‐related adverse events? Objective This integrative review aimed to elucidate the existing knowledge of education of patients with cancer receiving ICI therapy and their family caregivers, focusing on patients' and caregivers' self‐efficacy and self‐management of irAEs. The following research question guided the review: Research Question: How does patient education on ICI efficacy and toxicity management affect patients with cancer and their family caregivers' self‐efficacy and self‐management when dealing with immune‐related adverse events? Methods 2.1 Design This integrative review followed the methodology of Whittemore and Knafl . The Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRISMA) guided the reporting of the review, increasing the transparency of the review. The review was prospectively reported to the International Prospective Register of Systematic Reviews (PROSPERO) ref number: CRD42024511513. 2.2 Eligibility Criteria Inclusion criteria: Studies addressing how patient education on ICI affects adult patients ≥ 18 years of age with various cancer types and their adult family caregivers ≥ 18 years of age management of irAEs were included. The patients should be treated with ICI as first or second‐line treatment in an oncological (inpatient and/or outpatient) setting. In this study, we defined family according to Wright and Leahey: “ the family is who they say they are ” . Family caregivers could be adult relatives, close friends, neighbors, and the like, who help with self‐care tasks, provision of emotional and social support, health and medical care, advocacy and care coordination, and surrogacy . The review included original research. In accordance with the methodology of integrative reviews, gray literature in the form of conference abstracts from original research is included. There was no restriction to study type (qualitative/quantitative). No geographical limiters were applied. Exclusion criteria: Studies conducted outside of an oncological setting; the study population consisted of pediatric patients or children as family caregivers; or patients were treated with ICI as third or subsequent line treatment. Studies in other language than English, Danish, Norwegian, or Swedish were excluded thereby increasing the feasibility of the review. 2.3 Information Sources and Search Strategy Multiple search strategies were employed. A systematic search with no publication date restriction was conducted in EMBASE (Ovid), MEDLINE (Ovid), CINAHL (Ebsco), PsycINFO (Ovid), and Scopus. Database searches took place on February 12th and 13th, 2024. The search was re‐run August 23rd, 2024, but we did not identify any additional studies. The search strategy was developed from the PEO framework in collaboration with a research librarian. The search profile was first developed for EMBASE (Table ) and subsequently adapted for each database . Search profiles for MEDLINE, CINAHL, PsycINFO, and Scopus are available as supplemental material. Preliminary searches for gray literature were conducted in the National Cancer Institute's database and Google Scholar; however, no relevant references were identified. Furthermore, by searching in EMBASE we are also searching for gray literature and preprints in BioRxiv and medRxiv. The first author conducted the search. 2.4 Selection Process The review's study selection process involved two steps: (a) screening titles and abstracts, and (b) screening full articles . The first and last author independently screened titles and abstracts of all references using the screening and data extraction software Covidence . References that failed to meet eligibility criteria or did not address the research question were excluded. Inconsistent screening results were resolved by discussion between the first and last author. The first and last authors then screened the full text of the included reports for relevance to the review question and eligibility criteria. Questions about study eligibility were resolved through discussion. The first author conducted author searches on authors of identified literature included in this review, and a backward citation search and forward citation tracking on included reports . Furthermore, the “Find Similar”‐algorithm was applied in the searched databases. 2.5 Data Extraction and Analysis The first author extracted the data into a predefined data extraction template in Covidence. Extracted data included author(s), year of publication, country of origin, study aim(s), outcomes, study design, setting(s), participant information, type of patient education provided, main findings, and results of the critical appraisal. Extracted data of the main findings were analyzed using thematic analysis . The first, second, and last authors coded the data, identified patterns, and grouped the codes into themes. The thematic analysis is available as supplemental material. 2.6 Quality Appraisal The Mixed Methods Appraisal Tool (MMAT) was used to critically appraise the quality of the included studies . Exclusion based on the overall score is discouraged because integrative reviews are intended to gather and report all the evidence on a topic, regardless of the methodological quality . Therefore, we did not exclude studies based on the results of the quality appraisal. The first author performed the critical appraisal and discussed the results with the last author. Design This integrative review followed the methodology of Whittemore and Knafl . The Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRISMA) guided the reporting of the review, increasing the transparency of the review. The review was prospectively reported to the International Prospective Register of Systematic Reviews (PROSPERO) ref number: CRD42024511513. Eligibility Criteria Inclusion criteria: Studies addressing how patient education on ICI affects adult patients ≥ 18 years of age with various cancer types and their adult family caregivers ≥ 18 years of age management of irAEs were included. The patients should be treated with ICI as first or second‐line treatment in an oncological (inpatient and/or outpatient) setting. In this study, we defined family according to Wright and Leahey: “ the family is who they say they are ” . Family caregivers could be adult relatives, close friends, neighbors, and the like, who help with self‐care tasks, provision of emotional and social support, health and medical care, advocacy and care coordination, and surrogacy . The review included original research. In accordance with the methodology of integrative reviews, gray literature in the form of conference abstracts from original research is included. There was no restriction to study type (qualitative/quantitative). No geographical limiters were applied. Exclusion criteria: Studies conducted outside of an oncological setting; the study population consisted of pediatric patients or children as family caregivers; or patients were treated with ICI as third or subsequent line treatment. Studies in other language than English, Danish, Norwegian, or Swedish were excluded thereby increasing the feasibility of the review. Information Sources and Search Strategy Multiple search strategies were employed. A systematic search with no publication date restriction was conducted in EMBASE (Ovid), MEDLINE (Ovid), CINAHL (Ebsco), PsycINFO (Ovid), and Scopus. Database searches took place on February 12th and 13th, 2024. The search was re‐run August 23rd, 2024, but we did not identify any additional studies. The search strategy was developed from the PEO framework in collaboration with a research librarian. The search profile was first developed for EMBASE (Table ) and subsequently adapted for each database . Search profiles for MEDLINE, CINAHL, PsycINFO, and Scopus are available as supplemental material. Preliminary searches for gray literature were conducted in the National Cancer Institute's database and Google Scholar; however, no relevant references were identified. Furthermore, by searching in EMBASE we are also searching for gray literature and preprints in BioRxiv and medRxiv. The first author conducted the search. Selection Process The review's study selection process involved two steps: (a) screening titles and abstracts, and (b) screening full articles . The first and last author independently screened titles and abstracts of all references using the screening and data extraction software Covidence . References that failed to meet eligibility criteria or did not address the research question were excluded. Inconsistent screening results were resolved by discussion between the first and last author. The first and last authors then screened the full text of the included reports for relevance to the review question and eligibility criteria. Questions about study eligibility were resolved through discussion. The first author conducted author searches on authors of identified literature included in this review, and a backward citation search and forward citation tracking on included reports . Furthermore, the “Find Similar”‐algorithm was applied in the searched databases. Data Extraction and Analysis The first author extracted the data into a predefined data extraction template in Covidence. Extracted data included author(s), year of publication, country of origin, study aim(s), outcomes, study design, setting(s), participant information, type of patient education provided, main findings, and results of the critical appraisal. Extracted data of the main findings were analyzed using thematic analysis . The first, second, and last authors coded the data, identified patterns, and grouped the codes into themes. The thematic analysis is available as supplemental material. Quality Appraisal The Mixed Methods Appraisal Tool (MMAT) was used to critically appraise the quality of the included studies . Exclusion based on the overall score is discouraged because integrative reviews are intended to gather and report all the evidence on a topic, regardless of the methodological quality . Therefore, we did not exclude studies based on the results of the quality appraisal. The first author performed the critical appraisal and discussed the results with the last author. Results 3.1 Search Results After removing duplicates, 4182 references were screened for eligibility. Seven studies met the eligibility criteria and were included (see PRISMA Flow Diagram in Figure ). Included studies were conducted in the United States ( n = 2) , Canada ( n = 2) , Italy ( n = 1) , Spain ( n = 1) , and Germany ( n = 1) . Six studies included patients , and one study included both patients and caregivers . All the included studies provided quantitative data. The studies had various designs; one randomized controlled pilot trial , one cohort study , three pre‐post studies , and two observational studies (Table ). We identified one study protocol that met the inclusion criteria, but we decided to exclude as it is an ongoing study. The seven studies included a total of 9639 participants. The studies included patients with various cancer types: lung ( n = 5) , melanoma ( n = 2) , renal cell ( n = 2) , squamous cell ( n = 2) , urological cancer ( n = 2) , hepatobiliary cancer ( n = 1) , head and neck cancer ( n = 1) , gynecological cancer ( n = 1) and gastrointestinal cancer ( n = 1) . Two studies did not specify the cancer diagnoses , and four studies had missing data on some of the participants . Patients in the included studies received treatment with ICI as PD‐1/PD‐L1 monotherapy ( n = 4) , CTLA‐4/PD‐1 combination therapy ( n = 3) , and PD‐1/PD‐L1 in combination with chemotherapy or other therapies ( n = 2) . Data concerning types of ICI treatment is missing in n = 3 studies . In two studies, ICI treatment was initiated due to metastatic disease . Data concerning the intend of ICI treatment is missing in n = 5 studies . Notably, the studies reported as articles scored higher in the quality appraisal (mean score: 5.66 “yes”/all checklist items) than the studies reported in abstracts (mean score: 2.25 “yes”/all checklist items) (Table ) (quality appraisal using MMAT is available as supplemental material). 3.2 Synthesis and Reporting of Study Themes Three themes emerged from the data: (a) Feasibility of various strategies in patient education, (b) The effect of patient education on self‐efficacy, and (c) Determinants to improve self‐management of irAEs (Figure ). 3.2.1 Feasibility of Various Strategies in Patient Education Various strategies were identified for educating patients about irAEs. Of the seven included studies, one study examined patient education by the use of a smartphone application ; one study examined a pharmacist intervention including ICI and irAEs education prior to treatment initiation and proactive irAEs monitoring via telephone follow‐up prior to each cycle of ICI for the first three months ; one study tested online education modules ; one study examined a combined oral and written information and a video ; one study oral and written information ; one study only written information ; and one study as a baseline nursing assessment of baseline symptoms combining a one‐on‐one education session between the patient and ICI nurse and follow‐up phone calls . Information leaflets providing information on irAEs and offering clinical guidance to prevent and manage symptoms can be useful for recognizing and managing irAEs if patients and caregivers actually read them . In combination with oral education on ICIs it can contribute to greater patient satisfaction . However, more innovative ways of educating patients with cancer and their caregivers about ICIs have also shown to be feasible. Online patient and caregiver‐focused education can be successful in improving familiarity with essential elements involved in treatment with ICIs, including irAEs . Sauer et al. studied the feasibility of a smartphone application (the SOFIA‐App) with integrated ICI‐coaching and monitoring components. They found high feasibility and acceptance of the SOFIA‐App with a retention rate of 85% after three months . Moreover, no patients refused participation following randomization, indicating the feasibility of the smartphone application . These studies testing the feasibility of patient education offer hope for increased self‐efficacy and thereby better self‐management of irAEs. 3.2.2 The Effect of Patient Education on Self‐Efficacy Two of the included studies used the Cancer Behavior Inventory‐Brief Version (CBI‐B) to evaluate patients' self‐efficacy . In general, patient education had a positive effect on patients' self‐efficacy. Cheema et al. found a statistically improvement in the average CBI‐B scores ( p < 0.001) after a baseline nursing assessment of symptoms and education, and follow‐up phone calls , indicating an improved self‐efficacy and ability to act on irAEs. Similarly, pharmacist interventions including baseline education and follow‐up phone calls can empower the patient to play an active role in their cancer care, and improve detection of new symptoms or irAEs . This aligns with Teixeira‐Poit et al. who found that patients' knowledge of treatment with ICIs improved after an education session compared to before the education session. Moreover, the patients were able to recognize irAEs after an education session . Cheema et al. also found that a nursing assessment of symptoms and education program led to improved comprehension of irAEs . Digital education can also empower, engage, and equip patients and their caregivers with valuable information needed for self‐care management of irAEs . However, in a randomized controlled pilot study, Sauer et al. did not find any difference regarding self‐efficacy between the intervention group (the patients who used the SOFIA‐App) and the control group (the patients who received information and were provided with an emergency telephone number) . Two studies assessed the patients' health literacy. Teixeira‐Poit et al. used the Short Test of Functional Health Literacy in Adults (STOFHLA) to assess health literacy. They found that 97% of patients had adequate overall health literacy . This is contrary to the findings of Cheema et al. who found that 41% of patients had limited cancer health literacy (using the 6‐item Cancer Health Literacy Test ). 3.2.3 Determinants to Improve Self‐Management of irAEs Patient education is vital to improve early detection and management of irAEs . Positive trends are especially apparent in the studies where patient education is combined with follow‐up phone calls or electronic Patient Reported Outcomes (ePRO). That the patients' reported outcomes are seen and acted upon by healthcare professionals may contribute to better health‐related quality of life as well as less depression and distress . Myers et al. found higher odds of treatment discontinuation due to irAEs in patients who did not receive the dedicated pharmacist follow‐up . Cheema et al. found that after a baseline nursing assessment of baseline symptoms and education, the method of irAEs detection was mainly by patient self‐reporting, which was the case in 62% of the reported irAEs, followed by proactive phone calls (27% of reported irAEs) . Three patients had detection of an irAE during a visit to the emergency department . This finding is similar to Sauer et al. who identified a trend toward more emergency department visits in the control group ( n = 11) compared to the intervention group ( n = 4). However, Teixeira‐Poit et al. found no differences in the number of emergency department visits before and after implementing ICI education in the form of oral information combined with a video of ICIs mechanisms of action and written material . Search Results After removing duplicates, 4182 references were screened for eligibility. Seven studies met the eligibility criteria and were included (see PRISMA Flow Diagram in Figure ). Included studies were conducted in the United States ( n = 2) , Canada ( n = 2) , Italy ( n = 1) , Spain ( n = 1) , and Germany ( n = 1) . Six studies included patients , and one study included both patients and caregivers . All the included studies provided quantitative data. The studies had various designs; one randomized controlled pilot trial , one cohort study , three pre‐post studies , and two observational studies (Table ). We identified one study protocol that met the inclusion criteria, but we decided to exclude as it is an ongoing study. The seven studies included a total of 9639 participants. The studies included patients with various cancer types: lung ( n = 5) , melanoma ( n = 2) , renal cell ( n = 2) , squamous cell ( n = 2) , urological cancer ( n = 2) , hepatobiliary cancer ( n = 1) , head and neck cancer ( n = 1) , gynecological cancer ( n = 1) and gastrointestinal cancer ( n = 1) . Two studies did not specify the cancer diagnoses , and four studies had missing data on some of the participants . Patients in the included studies received treatment with ICI as PD‐1/PD‐L1 monotherapy ( n = 4) , CTLA‐4/PD‐1 combination therapy ( n = 3) , and PD‐1/PD‐L1 in combination with chemotherapy or other therapies ( n = 2) . Data concerning types of ICI treatment is missing in n = 3 studies . In two studies, ICI treatment was initiated due to metastatic disease . Data concerning the intend of ICI treatment is missing in n = 5 studies . Notably, the studies reported as articles scored higher in the quality appraisal (mean score: 5.66 “yes”/all checklist items) than the studies reported in abstracts (mean score: 2.25 “yes”/all checklist items) (Table ) (quality appraisal using MMAT is available as supplemental material). Synthesis and Reporting of Study Themes Three themes emerged from the data: (a) Feasibility of various strategies in patient education, (b) The effect of patient education on self‐efficacy, and (c) Determinants to improve self‐management of irAEs (Figure ). 3.2.1 Feasibility of Various Strategies in Patient Education Various strategies were identified for educating patients about irAEs. Of the seven included studies, one study examined patient education by the use of a smartphone application ; one study examined a pharmacist intervention including ICI and irAEs education prior to treatment initiation and proactive irAEs monitoring via telephone follow‐up prior to each cycle of ICI for the first three months ; one study tested online education modules ; one study examined a combined oral and written information and a video ; one study oral and written information ; one study only written information ; and one study as a baseline nursing assessment of baseline symptoms combining a one‐on‐one education session between the patient and ICI nurse and follow‐up phone calls . Information leaflets providing information on irAEs and offering clinical guidance to prevent and manage symptoms can be useful for recognizing and managing irAEs if patients and caregivers actually read them . In combination with oral education on ICIs it can contribute to greater patient satisfaction . However, more innovative ways of educating patients with cancer and their caregivers about ICIs have also shown to be feasible. Online patient and caregiver‐focused education can be successful in improving familiarity with essential elements involved in treatment with ICIs, including irAEs . Sauer et al. studied the feasibility of a smartphone application (the SOFIA‐App) with integrated ICI‐coaching and monitoring components. They found high feasibility and acceptance of the SOFIA‐App with a retention rate of 85% after three months . Moreover, no patients refused participation following randomization, indicating the feasibility of the smartphone application . These studies testing the feasibility of patient education offer hope for increased self‐efficacy and thereby better self‐management of irAEs. 3.2.2 The Effect of Patient Education on Self‐Efficacy Two of the included studies used the Cancer Behavior Inventory‐Brief Version (CBI‐B) to evaluate patients' self‐efficacy . In general, patient education had a positive effect on patients' self‐efficacy. Cheema et al. found a statistically improvement in the average CBI‐B scores ( p < 0.001) after a baseline nursing assessment of symptoms and education, and follow‐up phone calls , indicating an improved self‐efficacy and ability to act on irAEs. Similarly, pharmacist interventions including baseline education and follow‐up phone calls can empower the patient to play an active role in their cancer care, and improve detection of new symptoms or irAEs . This aligns with Teixeira‐Poit et al. who found that patients' knowledge of treatment with ICIs improved after an education session compared to before the education session. Moreover, the patients were able to recognize irAEs after an education session . Cheema et al. also found that a nursing assessment of symptoms and education program led to improved comprehension of irAEs . Digital education can also empower, engage, and equip patients and their caregivers with valuable information needed for self‐care management of irAEs . However, in a randomized controlled pilot study, Sauer et al. did not find any difference regarding self‐efficacy between the intervention group (the patients who used the SOFIA‐App) and the control group (the patients who received information and were provided with an emergency telephone number) . Two studies assessed the patients' health literacy. Teixeira‐Poit et al. used the Short Test of Functional Health Literacy in Adults (STOFHLA) to assess health literacy. They found that 97% of patients had adequate overall health literacy . This is contrary to the findings of Cheema et al. who found that 41% of patients had limited cancer health literacy (using the 6‐item Cancer Health Literacy Test ). 3.2.3 Determinants to Improve Self‐Management of irAEs Patient education is vital to improve early detection and management of irAEs . Positive trends are especially apparent in the studies where patient education is combined with follow‐up phone calls or electronic Patient Reported Outcomes (ePRO). That the patients' reported outcomes are seen and acted upon by healthcare professionals may contribute to better health‐related quality of life as well as less depression and distress . Myers et al. found higher odds of treatment discontinuation due to irAEs in patients who did not receive the dedicated pharmacist follow‐up . Cheema et al. found that after a baseline nursing assessment of baseline symptoms and education, the method of irAEs detection was mainly by patient self‐reporting, which was the case in 62% of the reported irAEs, followed by proactive phone calls (27% of reported irAEs) . Three patients had detection of an irAE during a visit to the emergency department . This finding is similar to Sauer et al. who identified a trend toward more emergency department visits in the control group ( n = 11) compared to the intervention group ( n = 4). However, Teixeira‐Poit et al. found no differences in the number of emergency department visits before and after implementing ICI education in the form of oral information combined with a video of ICIs mechanisms of action and written material . Feasibility of Various Strategies in Patient Education Various strategies were identified for educating patients about irAEs. Of the seven included studies, one study examined patient education by the use of a smartphone application ; one study examined a pharmacist intervention including ICI and irAEs education prior to treatment initiation and proactive irAEs monitoring via telephone follow‐up prior to each cycle of ICI for the first three months ; one study tested online education modules ; one study examined a combined oral and written information and a video ; one study oral and written information ; one study only written information ; and one study as a baseline nursing assessment of baseline symptoms combining a one‐on‐one education session between the patient and ICI nurse and follow‐up phone calls . Information leaflets providing information on irAEs and offering clinical guidance to prevent and manage symptoms can be useful for recognizing and managing irAEs if patients and caregivers actually read them . In combination with oral education on ICIs it can contribute to greater patient satisfaction . However, more innovative ways of educating patients with cancer and their caregivers about ICIs have also shown to be feasible. Online patient and caregiver‐focused education can be successful in improving familiarity with essential elements involved in treatment with ICIs, including irAEs . Sauer et al. studied the feasibility of a smartphone application (the SOFIA‐App) with integrated ICI‐coaching and monitoring components. They found high feasibility and acceptance of the SOFIA‐App with a retention rate of 85% after three months . Moreover, no patients refused participation following randomization, indicating the feasibility of the smartphone application . These studies testing the feasibility of patient education offer hope for increased self‐efficacy and thereby better self‐management of irAEs. The Effect of Patient Education on Self‐Efficacy Two of the included studies used the Cancer Behavior Inventory‐Brief Version (CBI‐B) to evaluate patients' self‐efficacy . In general, patient education had a positive effect on patients' self‐efficacy. Cheema et al. found a statistically improvement in the average CBI‐B scores ( p < 0.001) after a baseline nursing assessment of symptoms and education, and follow‐up phone calls , indicating an improved self‐efficacy and ability to act on irAEs. Similarly, pharmacist interventions including baseline education and follow‐up phone calls can empower the patient to play an active role in their cancer care, and improve detection of new symptoms or irAEs . This aligns with Teixeira‐Poit et al. who found that patients' knowledge of treatment with ICIs improved after an education session compared to before the education session. Moreover, the patients were able to recognize irAEs after an education session . Cheema et al. also found that a nursing assessment of symptoms and education program led to improved comprehension of irAEs . Digital education can also empower, engage, and equip patients and their caregivers with valuable information needed for self‐care management of irAEs . However, in a randomized controlled pilot study, Sauer et al. did not find any difference regarding self‐efficacy between the intervention group (the patients who used the SOFIA‐App) and the control group (the patients who received information and were provided with an emergency telephone number) . Two studies assessed the patients' health literacy. Teixeira‐Poit et al. used the Short Test of Functional Health Literacy in Adults (STOFHLA) to assess health literacy. They found that 97% of patients had adequate overall health literacy . This is contrary to the findings of Cheema et al. who found that 41% of patients had limited cancer health literacy (using the 6‐item Cancer Health Literacy Test ). Determinants to Improve Self‐Management of irAEs Patient education is vital to improve early detection and management of irAEs . Positive trends are especially apparent in the studies where patient education is combined with follow‐up phone calls or electronic Patient Reported Outcomes (ePRO). That the patients' reported outcomes are seen and acted upon by healthcare professionals may contribute to better health‐related quality of life as well as less depression and distress . Myers et al. found higher odds of treatment discontinuation due to irAEs in patients who did not receive the dedicated pharmacist follow‐up . Cheema et al. found that after a baseline nursing assessment of baseline symptoms and education, the method of irAEs detection was mainly by patient self‐reporting, which was the case in 62% of the reported irAEs, followed by proactive phone calls (27% of reported irAEs) . Three patients had detection of an irAE during a visit to the emergency department . This finding is similar to Sauer et al. who identified a trend toward more emergency department visits in the control group ( n = 11) compared to the intervention group ( n = 4). However, Teixeira‐Poit et al. found no differences in the number of emergency department visits before and after implementing ICI education in the form of oral information combined with a video of ICIs mechanisms of action and written material . Discussion To the best of our knowledge, this study is the only integrative review of evidence related to how patient education on ICI efficacy and toxicity management affects adult patients with cancer and their adult family caregivers' self‐efficacy and self‐management when dealing with immune‐related adverse events. 4.1 Search Results We found few relevant studies, and they were all quantitative studies. All the included studies have been published within the last seven years, indicating that research within patient and caregiver education on ICI is in its infancy. The geographical diversity of the included studies underscores the global relevance of the research on how patient education on ICI affects patients with cancer and their family caregivers' self‐efficacy and self‐management in relation to irAEs. This geographical spread enhances the generalizability of the findings, although there is a predominance of Western countries. The studies predominantly focused on patients with cancer, with only one study incorporating both patients and caregivers . This indicates a gap in the research regarding the caregiver perspective. Patients manage their disease with daily assistance from family caregivers, and family caregivers' support may help empower and engage the patient to self‐manage the condition . Therefore, educating family caregivers about ICI and irAEs is important. The diversity on cancer types across the studies is notable, with lung cancer being the most frequent (included in n = 5 studies). This focus on lung cancer aligns with its high prevalence and the indication for the use of ICI on this large group of patients with cancer. Other major groups of cancer diagnoses, such as colorectal, breast, and prostate, have minimal use of ICI. The inclusion of a variety of cancer types broadens the understanding of patient experiences across different cancer diagnoses. However, the fact that two studies did not specify the cancer diagnosis and four studies had some missing data concerning cancer diagnosis is a limitation, as it restricts the ability to compare across specific cancer types. 4.2 Study Themes In the following the three themes will be discussed: Feasibility of various strategies in patient education, The effect of patient education on self‐efficacy, and Determinants to improve self‐management of irAEs. Various types of patient education (e.g., oral and/or written information) can be useful for recognizing and managing irAEs and contribute to a greater patient satisfaction. Innovative ways, such as online education or applications, are also feasible in improving familiarity with essential elements involved in treatment with ICI and irAEs. In general, patient education has a positive effect on patient's self‐efficacy, and empowers patients and their caregivers with valuable information needed for increased self‐management of irAEs. The included studies present a diverse array of methods for educating patients and family caregivers about ICI and managing of irAEs. Several studies highlighted the effectiveness of combining oral and written information for patient education. The combination of educational strategies seems to increase patient satisfaction . A study comparing written information on chemotherapy with audio‐visual education showed that audio‐visual education is more beneficial in reducing distress in patients with cancer . Similarly, innovative methods are proven to be highly feasible educational tools to improve familiarity with key aspects of ICI treatment and irAEs . The positive effect of innovative methods in patient education is highlighted in a systematic review finding that the use of mHealth interventions can improve the patient's knowledge by the provision of self‐care recommendations, activation of the patient's role in the care process, improve the quality of life in patients with cancer, and contribute to patient satisfaction . Looking for the future, virtual reality (VR) is emerging as an exciting avenue for patient education and can improve patient's knowledge of their illness and satisfaction with treatment . VR is already used in the cancer setting, however, most prevalent in radiotherapy . Patient education interventions demonstrated a positive impact on self‐efficacy , which is in line with Myers et al. (2023) finding that patient education interventions empowered patients to communicate irAEs and manage their treatment. This aligns with a cross‐sectional study finding that patients with increased self‐efficacy had better chemotherapy self‐management , suggesting that structured educational programs can enhance patients' proactive involvement and confidence in managing irAEs. Teixeira‐Poit et al. found that 97% of patients had adequate health literacy , which likely facilitated the effective uptake of educational content. On the contrary, Cheema et al. identified that 41% of patients had limited health literacy . These contrasting findings underscore the importance of assessing and accommodating health literacy in patient education programs. Sørensen et al. also emphasized that having a low level of health literacy or no support from caregivers have a negative impact on patient safety regarding patient‐reported irAEs . Similarly, another study found that patients with little social support or support network have poorer oncologic outcomes . Therefore, patient education on ICI should include an examination of patients' health literacy and supportive network. The importance of patient education on the early detection and management of irAEs in cancer treatment cannot be overstated. Structured patient education improves patient outcomes by enhancing their ability to detect and manage irAEs effectively . In combination with follow‐up phone calls or ePROs, it can reduce treatment discontinuation , lead to fewer visits to the emergency department due to irAEs , and increase patients' health‐related quality of life . This is in line with Tolstrup et al., who emphasized the increase in patients' health‐related quality of life and cancer well‐being with the use of ePRO . Furthermore, ePRO may enhance patients' awareness of irAEs and increase their feeling of involvement . 4.3 Implications The findings of this integrative review highlight the critical role of patient education in managing irAEs for patients with cancer undergoing ICI treatment. Incorporating educational strategies combining written, oral, and digital resources to enhance patient and caregiver understanding of ICI are feasible to improve patient self‐efficacy and self‐management capabilities, leading to better clinical outcomes such as early detection and management of irAEs. The inclusion of follow‐up mechanisms, e.g., ePROs, can aid the early detection of irAEs, and improve overall patient satisfaction and quality of life. However, there is a lack of emphasis on educating the caregivers together with the patient, and future research should strive to include caregivers as they are of great importance to the patient during treatment. Furthermore, the promising results from innovative educational methods warrant further exploration. However, these future studies should focus on tailored educational material to accommodate varying levels of health literacy, ensuring that all patients and caregivers can benefit from these educational programs. 4.4 Limitations All efforts were made to identify relevant studies, with two authors performing the reference screening. The study adhered to guidelines for integrative reviews, incorporating both primary studies and gray literature to provide complementary knowledge. A potential limitation was the selection of search terms, especially those related to patient education. Other search terms may have been relevant to include in the search, for example, information, and there is a risk that we have been missing relevant studies to include. However, including information as a search term could have increased the sensitivity of the database searches, resulting in a higher number of irrelevant studies. Furthermore, we did not include self‐efficacy and self‐management as search terms, as this would limit the number of hits too much and increase the risk of missing important studies. It is a limitation that four out of seven included studies are reported in the form of abstracts, limiting the provision of the full information on methodological considerations. In particular, information on sampling strategy and participant representativeness are not provided, which is reflected in the quality appraisal of the studies. Search Results We found few relevant studies, and they were all quantitative studies. All the included studies have been published within the last seven years, indicating that research within patient and caregiver education on ICI is in its infancy. The geographical diversity of the included studies underscores the global relevance of the research on how patient education on ICI affects patients with cancer and their family caregivers' self‐efficacy and self‐management in relation to irAEs. This geographical spread enhances the generalizability of the findings, although there is a predominance of Western countries. The studies predominantly focused on patients with cancer, with only one study incorporating both patients and caregivers . This indicates a gap in the research regarding the caregiver perspective. Patients manage their disease with daily assistance from family caregivers, and family caregivers' support may help empower and engage the patient to self‐manage the condition . Therefore, educating family caregivers about ICI and irAEs is important. The diversity on cancer types across the studies is notable, with lung cancer being the most frequent (included in n = 5 studies). This focus on lung cancer aligns with its high prevalence and the indication for the use of ICI on this large group of patients with cancer. Other major groups of cancer diagnoses, such as colorectal, breast, and prostate, have minimal use of ICI. The inclusion of a variety of cancer types broadens the understanding of patient experiences across different cancer diagnoses. However, the fact that two studies did not specify the cancer diagnosis and four studies had some missing data concerning cancer diagnosis is a limitation, as it restricts the ability to compare across specific cancer types. Study Themes In the following the three themes will be discussed: Feasibility of various strategies in patient education, The effect of patient education on self‐efficacy, and Determinants to improve self‐management of irAEs. Various types of patient education (e.g., oral and/or written information) can be useful for recognizing and managing irAEs and contribute to a greater patient satisfaction. Innovative ways, such as online education or applications, are also feasible in improving familiarity with essential elements involved in treatment with ICI and irAEs. In general, patient education has a positive effect on patient's self‐efficacy, and empowers patients and their caregivers with valuable information needed for increased self‐management of irAEs. The included studies present a diverse array of methods for educating patients and family caregivers about ICI and managing of irAEs. Several studies highlighted the effectiveness of combining oral and written information for patient education. The combination of educational strategies seems to increase patient satisfaction . A study comparing written information on chemotherapy with audio‐visual education showed that audio‐visual education is more beneficial in reducing distress in patients with cancer . Similarly, innovative methods are proven to be highly feasible educational tools to improve familiarity with key aspects of ICI treatment and irAEs . The positive effect of innovative methods in patient education is highlighted in a systematic review finding that the use of mHealth interventions can improve the patient's knowledge by the provision of self‐care recommendations, activation of the patient's role in the care process, improve the quality of life in patients with cancer, and contribute to patient satisfaction . Looking for the future, virtual reality (VR) is emerging as an exciting avenue for patient education and can improve patient's knowledge of their illness and satisfaction with treatment . VR is already used in the cancer setting, however, most prevalent in radiotherapy . Patient education interventions demonstrated a positive impact on self‐efficacy , which is in line with Myers et al. (2023) finding that patient education interventions empowered patients to communicate irAEs and manage their treatment. This aligns with a cross‐sectional study finding that patients with increased self‐efficacy had better chemotherapy self‐management , suggesting that structured educational programs can enhance patients' proactive involvement and confidence in managing irAEs. Teixeira‐Poit et al. found that 97% of patients had adequate health literacy , which likely facilitated the effective uptake of educational content. On the contrary, Cheema et al. identified that 41% of patients had limited health literacy . These contrasting findings underscore the importance of assessing and accommodating health literacy in patient education programs. Sørensen et al. also emphasized that having a low level of health literacy or no support from caregivers have a negative impact on patient safety regarding patient‐reported irAEs . Similarly, another study found that patients with little social support or support network have poorer oncologic outcomes . Therefore, patient education on ICI should include an examination of patients' health literacy and supportive network. The importance of patient education on the early detection and management of irAEs in cancer treatment cannot be overstated. Structured patient education improves patient outcomes by enhancing their ability to detect and manage irAEs effectively . In combination with follow‐up phone calls or ePROs, it can reduce treatment discontinuation , lead to fewer visits to the emergency department due to irAEs , and increase patients' health‐related quality of life . This is in line with Tolstrup et al., who emphasized the increase in patients' health‐related quality of life and cancer well‐being with the use of ePRO . Furthermore, ePRO may enhance patients' awareness of irAEs and increase their feeling of involvement . Implications The findings of this integrative review highlight the critical role of patient education in managing irAEs for patients with cancer undergoing ICI treatment. Incorporating educational strategies combining written, oral, and digital resources to enhance patient and caregiver understanding of ICI are feasible to improve patient self‐efficacy and self‐management capabilities, leading to better clinical outcomes such as early detection and management of irAEs. The inclusion of follow‐up mechanisms, e.g., ePROs, can aid the early detection of irAEs, and improve overall patient satisfaction and quality of life. However, there is a lack of emphasis on educating the caregivers together with the patient, and future research should strive to include caregivers as they are of great importance to the patient during treatment. Furthermore, the promising results from innovative educational methods warrant further exploration. However, these future studies should focus on tailored educational material to accommodate varying levels of health literacy, ensuring that all patients and caregivers can benefit from these educational programs. Limitations All efforts were made to identify relevant studies, with two authors performing the reference screening. The study adhered to guidelines for integrative reviews, incorporating both primary studies and gray literature to provide complementary knowledge. A potential limitation was the selection of search terms, especially those related to patient education. Other search terms may have been relevant to include in the search, for example, information, and there is a risk that we have been missing relevant studies to include. However, including information as a search term could have increased the sensitivity of the database searches, resulting in a higher number of irrelevant studies. Furthermore, we did not include self‐efficacy and self‐management as search terms, as this would limit the number of hits too much and increase the risk of missing important studies. It is a limitation that four out of seven included studies are reported in the form of abstracts, limiting the provision of the full information on methodological considerations. In particular, information on sampling strategy and participant representativeness are not provided, which is reflected in the quality appraisal of the studies. Conclusion Limited evidence exists related to how patient education on ICI efficacy and toxicity management affects adult patients with cancer and their adult family caregivers' self‐efficacy and self‐management when dealing with immune‐related adverse events. Patient education is of fundamental importance to improve the early detection and management of irAEs. The findings indicate that while traditional methods of patient education (i.e., oral and/or written information) remain valuable, the integration of digital and innovative technologies holds significant promise to enhancing patient and caregiver understanding of ICI and irAEs. Furthermore, patient education combined with follow‐up (such as phone calls or ePRO) may contribute to better health‐related quality of life. In general, patient education has a positive effect on patients' self‐efficacy. However, health literacy impacts patients' ability to understand and manage their treatment and irAEs, emphasizing the need for personalized education approaches. First author contributed in conceptualization, data curation, analysis, funding acquisition, investigation, methodology, project coordination, visualization, and writing the original draft. Second author contributed in conceptualization, analysis, funding acquisition, methodology, validation, and reviewing and editing the manuscript. Third author contributed in conceptualization, funding acquisition, methodology, and reviewing and editing the manuscript. Fourth author contributed in conceptualization and reviewing and editing the manuscript. Last author contributed in conceptualization, analysis, funding acquisition, methodology, supervision, validation, and reviewing and editing the manuscript. The authors declare no conflicts of interest. Review protocol can be accessed from PROSPERO. Amendments to information provided in the protocol are as follows: change of title, the sequence of self‐efficacy and self‐management in the research question, first, second, and last author coded and gathered data into themes together, not using NVIVO14 for data analysis due to small number of references. Search profiles for MEDLINE, CINAHL, PsycINFO, and Scopus, a matrix depicting the data analysis using thematic analysis, and quality appraisals using MMAT is available online. Table S1 Table S2 Table S3 Table S4 Table S5 Table S6 |
A randomized controlled trial of a family-based HIV/STI prevention program for Black girls and male caregivers in Chicago: IMAGE study protocol paper | c10f71e1-96b3-492e-8072-a51efaede8f9 | 11952266 | Health Promotion[mh] | Sexually transmitted infections (STIs) have been increasing disproportionately among Black girls and women in the United States (US). Among females aged 15–24 years, STIs are highest among Black girls and women across all US regions . Rates of chlamydia among Black girls are five times the rate of white girls , further amplifying their risk for poor sexual and reproductive health (SRH) outcomes (e.g., pelvic inflammatory disease, infertility, HIV). In Chicago, STI rates are highest among 13–29-year-old Black females, and in 2018, 56% of new HIV diagnoses were among Black women . During the COVID-19 pandemic, treatment of STIs dropped as new barriers to reproductive health care exacerbated existing challenges. The pandemic also illuminated high rates of intimate partner violence, a well-known correlate of HIV/STI . Elevated HIV/STIs are associated with sexual violence , and estimates predict that one in every four Black girls will rates be sexually abused before the age of 18 . During the COVID-19 pandemic, sexual education, mental health, and SRH services and curricula were nearly abandoned , exacerbating barriers to care and burdening individuals and their families with SRH education. Familial protection and family-based interventions may prevent exposure to sexual violence and STIs among Black girls while simultaneously strengthening family relationships and girls’ SRH [ – ]. Becoming a Sexual Black Woman is a framework that describes the sexual vulnerability and resilience of Black cisgender women during three developmental phases (Girl, Grown, and Woman) . It details how sociocultural conditions exacerbate risk at each phase and how familial protection may prevent Black girls’ exposure to sexual violence and STIs . Interventions that strengthen family relationships and communication align well with this framework and have demonstrated success in reducing sexual risk and STIs among Black girls . For example, IMARA (Informed Motivated, Aware, and Responsible about AIDS) an evidence-based intervention for Black girls and their female caregivers, revealed a 43% reduction in STIs one year after the program ended . Most family-based intervention programs focus on female caregivers (i.e., mothers, aunts, and sisters) and their role in the SRH of Black girls and young women [ , , ]. Yet, Black male caregivers (BMC), inclusive of fathers, uncles, grandfathers, brothers, etc., often neglected in family-focused interventions, have a unique role in educating and protecting Black girls as they navigate understandings of sex, sexual violence, body image, and other topics considered taboo to discuss in a girl-male caregiver relationship [ – ]. Male caregiver involvement specifically might offer distinctive protection against sexual violence exposure during girls’ sexual development . Evidence suggests that SRH programs that include male caregivers have been associated with later sexual debut and increased condom use in Black girls [ , , ], as well as improved communication, relationship quality, and less condomless sex in boys [ – ]. Black girls want SRH information from male caregivers , and men want to be involved in protecting girls . Engaging BMC in interventions may empower Black girls to reject unhealthy sexual advances that often lead to sexual violence. IMAGE was developed to address the needs of Black girls and their male caregivers using the Becoming a Sexual Black Woman theoretical framework and the Health Disparities Research Framework (HDRF) . The Becoming a Sexual Black Woman framework emphasizes the impact of structural determinants of health (racism, discrimination, sexual violence, stereotyped messages, and adultification) on Black girls’ sexual development pathways, sexual health, and decision-making. HDRF describes the influence of structural factors at the individual, interpersonal, community, and societal levels relevant to understanding and reducing health disparities . Consistent with the HDRF framework, Becoming a Sexual Black Woman highlights how lack of protection is a form of structural violence that Black girls experience at individual, interpersonal, and societal levels. IMAGE leverages BMCs desire to participate in girls’ lives and protect and support them while challenging toxic masculinity and the use of dominance, violence, and control to assert power and superiority. We followed the ADAPT-ITT framework’s eight steps to systematically tailor and adapt the IMARA curriculum to create IMAGE . We conducted in-depth interviews with 30 Black male caregivers and six focus groups with Black girls, male caregivers, and female caregivers, and these data-informed curriculum refinements. Adaptations also addressed drivers of structural violence for Black girls (i.e., stereotype messaging, lack of protection). Following the adaptations, we “theater tested” IMAGE with six girl-male caregiver dyads (n = 12) in collaboration with a community organization on the West side of Chicago. By delivering IMAGE at community-based organizations, we strengthen the relevance for Black communities and the likelihood the program will be adopted and sustained. IMAGE was revised based on feedback from the theater testing and then pilot-tested with 20 girl-BMC dyads. Findings revealed strong feasibility and acceptability: 87% of those approached agreed to participate; 93% completed the full intervention; 94% rated IMAGE as acceptable; and fidelity of intervention delivery was 98% . Retention at one-month follow-up was 100%, and 87% agreed to return at 6-months. Preliminary data showed promising improvements in behavior and theoretical mediators. Community-based implementation We partnered with community-based organizations (CBOs) to build community ownership IMAGE. We tailored IMAGE’s implementation to facilitate ease of community uptake. Based on the Exploration, Preparation, Implementation, and Sustainability (EPIS) framework . We simplified EPIS into 3-Steps that CBOs can use to implement evidence-based interventions . This approach has been successful for other behavior change interventions in clinical and community settings [ – ]. The 3-step approach supports CBO implementation and future ownership.
We partnered with community-based organizations (CBOs) to build community ownership IMAGE. We tailored IMAGE’s implementation to facilitate ease of community uptake. Based on the Exploration, Preparation, Implementation, and Sustainability (EPIS) framework . We simplified EPIS into 3-Steps that CBOs can use to implement evidence-based interventions . This approach has been successful for other behavior change interventions in clinical and community settings [ – ]. The 3-step approach supports CBO implementation and future ownership.
Design This is a 2-arm randomized efficacy and implementation trial of IMAGE for 14-18-year-old Black girls and their male caregivers. Girl and BMC dyads will be assigned to workshops randomized to IMAGE versus a time-matched health promotion program (FUEL). All participants will complete baseline, 6- and 12-month follow-up assessments via REDCap. Girls will provide urine to screen for three STIs (trichomoniasis, chlamydia, and gonorrhea). IMAGE engages male caregivers in HIV/STI prevention for girls, thereby increasing its relevance to Black communities and the likelihood of adoption and sustainability by CBOs. Therefore, we will also assess implementation outcomes at CBOs in Chicago. Aim 1 (Efficacy). We will assign 300 14-18-year-old Black girls and their male caregivers to workshops randomized to either IMAGE or FUEL and assess girls’ STI incidence, sexual partners, and condom use at baseline, 6- and 12-months. We hypothesize that girls who receive IMAGE will have lower STI incidence (primary outcome), fewer sexual partners, and more consistent condom use (secondary outcomes). We will explore associations of individual, interpersonal, and structural factors proposed by the B ecoming a Sexual Black Woman and HDRF framework on primary and secondary outcomes at all three time points ( ). Aim 2 (Implementation). We will identify and describe factors (barriers, facilitators, constraints) and implementation processes at six CBOs across Chicago. Study setting. We will deliver IMAGE at six CBOs on Chicago’s West and South sides impacted by structural factors (i.e., racism, poverty, crime, incarceration, and urban decay) implicated in high HIV and STI rates among Black communities. Neighborhoods in these areas are predominately Black, and Black females in these areas report the 4th highest rate of ever having sex (53%), the 3rd highest rate of sex before age 13, the 2nd highest rate of current sexual activity (40.6%) and have the highest rates of Chlamydia (1.921.0-2,688.4 per 100,000 people), and gonorrhea (973.7 - 1,536.2 per 100,000 people) . Rates of sexual violence are also high in these neighborhoods . The CBOs offer social services that address community resource needs and promote health and behavior change. Aim 1 study population (efficacy). Girls must be a) 14-18 years old; b) self-identify as African American, Black, or mixed race with African American or Black; c) speak English since measures are not normed for other languages; and d) identify an eligible male caregiver to participate in the study. BMC must be a) > 18 years old; b) self-identify as African American, Black, or mixed race with African American or Black; c) speak English since measures are not normed for other languages; d) be a current male caregiver to a Black girl 14-18 years old; and e) the girl’s legal guardian must agree to the male caregiver’s participation. BMC will be defined as men (fathers, grandfathers, uncles, brothers, cousins, etc.) who girls report playing a central role in their care and upbringing. Girls may or may not be sexually active, and information about girls’ sexual activity will not be shared with their male caregivers. Recruitment, assent/consent, compensation and retention. Recruitment will be multi-faceted. CBO will distribute flyers in the community and invite interested individuals to contact the IMAGE study. CBO will also contact eligible participants directly to inform them about the project and request permission for the research team to contact them. IMAGE staff will attend relevant community events to build relationships and conduct outreach. Using a screening and consent script, research staff will follow up with girls separately from guardians or male caregivers to conduct the screening, to obtain informed consent/assent with written, electronic signature via the e-consent module in REDCap, and to confirm the contact information of legal guardians and/or participating male caregivers. We will attempt to obtain girls’ assent/consent before we contact male caregivers, but participants may be screened in any order. Girls’ refusal to participate will precede consent or permission by male caregivers and legal guardians, regardless of the screening order. Dyads will complete surveys via REDCap at baseline, 6 and 12-month follow-up. Baseline survey assessments will occur before Day 1 of the intervention via REDCap. Girls and male caregivers will each receive $60 for completing the baseline assessment and an additional $65 for completing the entire workshop weekend. Participants are compensated $70 for the 6-month follow-up and $75 for the 12-month follow-up to recognize the value of retention over time. To overcome transportation barriers to STI treatment at the University of Illinois in Chicago Hospital Health Sciences System, we will provide an additional $5 in transportation compensation if needed. Participants’ transportation barriers will be assessed when scheduled for a workshop, and participant safety will be ensured. To maximize retention, we will use the strategies that were successful in our pilot (i.e., holiday cards and monthly phone calls or text messages) . We will collect contact information for girls and male caregivers in multiple forms to find the most effective communication for each participant. Randomization, sample size, and power. We will use a stratified block randomization allocation to assign girl-BMC dyads to the treatment arm (IMAGE n = 150 or FUEL n = 150) prior to baseline assessment. Power analyses determined that a sample of 300 girls – 150 IMAGE and 150 FUEL is needed to assess efficacy on the primary outcome (STI incidence). Effect size (Cohen’s h) is a function of two proportions defined by Cohen; h = .2 is a small effect, h = .5 is a medium effect, and h = .8 is a large effect. We calculated the power to detect the effect of IMAGE vs. FUEL in reducing STI incidence as measured by having any of three STIs at 12 months post-intervention using a 2-sided test with alpha = .05. The new STI outcome at follow-up was a binary variable (yes/no to testing positive for at least 1 of the 3 STIs), because the study was not powered to examine specific STI outcomes. Based on the IMARA retention rate of 86% for Black female caregiver-girl dyads in Chicago , we evaluated the power of STI incidence (yes/no) at 80% retention over 12- months. Based on our previous work, we assume the STI incidence is 25% and 11% at 12 months for girls in the FUEL arm and IMAGE arm, respectively, which is a small-to-medium effect size of h = .37. With the proposed sample size, we will achieve 80% power to detect the intervention effect. We also calculated the standard effect size (Cohen’s d) for group mean comparisons with 80% power for our secondary continuous outcomes, where d = 0 is defined as no treatment effect. We can detect a small-to-medium effect size with the proposed sample size (d = .37). Aim 1 efficacy study conditions. Intervention group: IMARA for black male caregivers and girls empowerment (IMAGE) : IMAGE is an 8-10-hour HIV/STI group-based (6-8 dyads) prevention program delivered to girls and BMC, together and separated, reviewing parallel content. Modules address individual, social, and structural drivers of HIV risk ( ). The IMAGE curriculum was tailored specifically for Chicago’s Black girls and male caregivers. The topics discussed in the curriculum include the sexual development of Black girls, the risk of sexual violence, female anatomy, body positivity, HIV/STI knowledge and attitudes, and contraception use. IMAGE was created to foster healthy communication and connection between male caregivers and their girls by encouraging conflict resolution and perspective-taking. Male caregivers learn about the importance of mental health, role modeling, and partner choices. Through IMAGE, male caregivers will become more aware of their presence in their girls’ lives, discuss absenteeism (in the event they are incarcerated or unable to be physically present), and learn to be present. IMAGE emphasizes how Black girls and women are portrayed in the media, which creates unhealthy stereotypes, decreases self-image, and increases vulnerability to sexual violence. Girls and caregivers will discuss intimate partner violence, power dynamics, and healthy relationships, as well as discuss the use of contraceptives, specifically condoms. IMAGE directly addresses structural factors on girls’ HIV/STI risk, namely male incarceration, sexual violence, domestic violence, and negative stereotyping. Woven throughout IMAGE is the impact of mental health, gender-based violence, and alcohol and drug use on HIV risk. Control condition: FUEL, a health promotion program : We used a similar approach as described above (ADAPT-ITT) to tailor FUEL for Black girls and male caregivers in Chicago. FUEL promotes healthy living practices and topics, namely media literacy, healthy eating and physical activity, healthy eating and nutrition, physical activity, teen drug and alcohol use, and violence prevention ( ). Like IMAGE, topics are covered in both separate and joint sessions. FUEL has a brief HIV education module, but it does not address SRH dyad communication or conflict navigation. Video essays, documentaries, and information are used to address relevant topics. Intervention training. Facilitators for both conditions must identify as Black, female, and have some experience working with Black communities or families. Facilitators were assigned to deliver IMAGE or FUEL. All facilitators completed 20 hours of training on their assigned curriculum and received supplemental one-hour training in-person and virtually to introduce key themes of cultural competency, PrEP, HIV prevention education, sex trafficking/human trafficking risk awareness, and general drug safety and navigating substance use disorders. Training for both programs stressed the importance of following the manualized protocol to ensure fidelity across facilitators and completing fidelity measurements at the end of each workshop. Following a curriculum review during the first two training sessions, facilitators practiced delivering each activity, taking turns serving as group leaders and participants. Each facilitator was “certified” as competent to deliver IMAGE or FUEL by the principal investigator and project director. Aim 1 efficacy trial. Six to eight dyads will be assigned to workshops randomized into FUEL or IMAGE at each CBO and will receive the corresponding curriculum on a designated weekend. Four facilitators and two observers are needed for each study condition; facilitators are not to be cross-trained in IMAGE and FUEL. Two facilitators co-lead the girls’ groups, and two co-lead the male caregivers’ groups. All four facilitators co-lead the joint sessions when girls and male caregivers come together for specific activities. Breakfast, lunch, and snacks are provided at all workshops for facilitators and participants. Supervision, quality assurance, and treatment. We will use a detailed manual and facilitator guide, conduct weekly supervision, and assess fidelity using facilitator self-report and observer ratings of adherence and competence (see below). Adherence measures determine if the program was delivered as intended (yes/no), and competence ratings indicate the quality of delivery. Each workshop is delivered by four facilitators and observed by two trained individuals to ensure fidelity and report on the quality of the workshop. Observers and facilitators will complete surveys after each workshop day, reflecting on the quality of facilitation. Participants will complete workshop evaluations. During ongoing supervision, the IMAGE research team will review fidelity reports and participant evaluations, discuss session activities and problem-solving difficulties, and provide feedback. If facilitators deviate from the curriculum, the supervisor will provide additional training until fidelity is achieved. We will complete observations of at least 40% of study workshops. Contamination. We will minimize contamination across arms. Facilitators will be assigned to IMAGE or FUEL and will not discuss or learn the other arm’s curriculum; training is done separately. The workshops will not be conducted on the same days; weekends will be divided, and each CBO will have only FUEL or IMAGE weekends to reduce participant contact. Confidentiality and a general orientation to the research structure were emphasized during facilitation training, allowing us to expect limited contamination. Intervention: 3-step sequence for implementation aim 2 We will utilize a 3-Step Implementation Model and its implementation guide at each CBO to Prepare, Rollout, and equip them to Sustain IMAGE. The three steps are outlined more in the implementation procedures section. Step 1 prepare Our team has previously developed partnerships with directors at six CBOs. The curriculum is designed to be flexible and capable of being delivered in various community settings. We will conduct site visits with each director, including environmental scans of prerequisite resources for hosting workshops. Each CBO director will appoint an internal liaison to lead recruitment activities and bridge communication between the sites and the IMAGE team, and the project director will interview liaisons to collect qualitative preparedness data. The information will be collected during site visits, interviews, and a brief prep phase survey to write a tailored implementation plan for each site per year. The implementation plan will then act as a living reference to guide workshop scheduling and execution for the year. Step 2 rollout Each CBO hosts an equal number of IMAGE and FUEL workshops for the efficacy study (Aim 1). Step 3 sustain With support and assistance from the implementation team, the CBO will review their experiences and decide if they will continue offering IMAGE at their location. Partnering CBO personnel Each CBO follows the same sequence of steps to Prepare, Rollout, and Sustain IMAGE. The study will began with the Prepare phase, which involved engaging with the CBOs. During Rollout, the efficacy data (Aim 1) is collected across the 3 time points: baseline, 6, and 12 months. Aim 2 will formally evaluate and compare the implementation determinants (barriers, facilitators, constraints) and processes within and across six CBOs using data generated from surveys, observations, study notes, and interviews with CBO liaisons and directors, hereafter CBO personnel. CBO personnel (n = 12) will be asked and consented to participate in surveys and qualitative interviews. Each CBO director will select a liaison who is the reference contact for everything related to the workshops. CBO liaisons will oversee recruiting dyads and bridging the CBO and the IMAGE team. Inclusion criteria: All CBO directors (n = 6) and CBO IMAGE liaisons (n = 6) will be eligible. Because CBO liaisons will receive a stipend from UIC and are also CBO employees, we emphasize that participation in the interviews and surveys is completely voluntary. Personnel can refuse to participate without affecting their employment at the CBO or their stipend from UIC. Exclusion criteria: a) inability to understand the consent process, and b) non-employment at a partnering CBO. Regarding CBO personnel turnover (liaisons and/or directors), PI Crooks will engage new directors, and directors will appoint a new individual to serve as the liaison and be trained to perform their activities. Aim 2 implementation procedures. During Prepare , we will conduct an environmental scan at each CBO to describe and compare the inner and outer context across CBOs. We will identify structural and logistical resources (i.e., space, furniture, and other infrastructure). The scan also includes documenting current programming, the structure of decision-making, available resources (e.g., staff, space), and client engagement (e.g., number and frequency of the target population coming to the CBO). We review organizational charts to understand staff roles and determine how IMAGE fits the current structure. We identify the needs of each CBO and how the research team can support IMAGE delivery. We work with CBO directors to identify a liaison at the CBO who can interface with the community and assist in recruitment and implementation. We will share the materials related to IMAGE with the CBO personnel - director and liaison from each CBO (n = 12), and we will ask them to complete a survey about the factors likely to affect IMAGE delivery and potential sustainability (ORIC and organizational climate measures). We will also conduct a 45-60-minute semi-structured interview with CBO liaisons (n = 6) involved in the IMAGE study (face-to-face, phone, or via Zoom) to document anticipated barriers and facilitators. The Prepare phase semi-structured interview only takes place with the liaison as a follow-up to the conversations held during the initial site visit with the CBO Directors. Verbal consent will be obtained before each interview. Participants will receive a $25 gift card for participating in the interview and survey. After the Prepare step, we will triangulate the data to co-create an adaptable implementation plan detailing how to deliver IMAGE at their site. During IMAGE Rollout , we will complete the Aim 1 efficacy study. After the first IMAGE session, we will dispatch a follow-up survey assessing implementation, adoption, and feasibility and conduct 30-45-minute semi-structured in-depth interviews (in person, by phone, or via Zoom) with CBO personnel (n = 12) to document facilitators, barriers and needed adaptations (e.g., to increase IMAGE enrollment and attendance). Verbal consent will be obtained before each interview, and participants will receive a $25 gift card. After the trial ends (end of Rollout), CBO personnel will be asked to complete a third survey focused on acceptability, appropriateness, feasibility, and sustainability. We will conduct 45-60-minute semi-structured interviews (face-to-face, by phone, or via Zoom) with CBO personnel (n = 12) to assess each CBO’s total experience and capacity to Sustain IMAGE. Building on the environmental scan and previous interviews that identified facilitators, barriers, and solutions, we will ask CBO personnel to discuss their ability to integrate IMAGE into current programming. Verbal consent will be obtained before each interview. Participants will receive a $25 gift card.
This is a 2-arm randomized efficacy and implementation trial of IMAGE for 14-18-year-old Black girls and their male caregivers. Girl and BMC dyads will be assigned to workshops randomized to IMAGE versus a time-matched health promotion program (FUEL). All participants will complete baseline, 6- and 12-month follow-up assessments via REDCap. Girls will provide urine to screen for three STIs (trichomoniasis, chlamydia, and gonorrhea). IMAGE engages male caregivers in HIV/STI prevention for girls, thereby increasing its relevance to Black communities and the likelihood of adoption and sustainability by CBOs. Therefore, we will also assess implementation outcomes at CBOs in Chicago. Aim 1 (Efficacy). We will assign 300 14-18-year-old Black girls and their male caregivers to workshops randomized to either IMAGE or FUEL and assess girls’ STI incidence, sexual partners, and condom use at baseline, 6- and 12-months. We hypothesize that girls who receive IMAGE will have lower STI incidence (primary outcome), fewer sexual partners, and more consistent condom use (secondary outcomes). We will explore associations of individual, interpersonal, and structural factors proposed by the B ecoming a Sexual Black Woman and HDRF framework on primary and secondary outcomes at all three time points ( ). Aim 2 (Implementation). We will identify and describe factors (barriers, facilitators, constraints) and implementation processes at six CBOs across Chicago. Study setting. We will deliver IMAGE at six CBOs on Chicago’s West and South sides impacted by structural factors (i.e., racism, poverty, crime, incarceration, and urban decay) implicated in high HIV and STI rates among Black communities. Neighborhoods in these areas are predominately Black, and Black females in these areas report the 4th highest rate of ever having sex (53%), the 3rd highest rate of sex before age 13, the 2nd highest rate of current sexual activity (40.6%) and have the highest rates of Chlamydia (1.921.0-2,688.4 per 100,000 people), and gonorrhea (973.7 - 1,536.2 per 100,000 people) . Rates of sexual violence are also high in these neighborhoods . The CBOs offer social services that address community resource needs and promote health and behavior change. Aim 1 study population (efficacy). Girls must be a) 14-18 years old; b) self-identify as African American, Black, or mixed race with African American or Black; c) speak English since measures are not normed for other languages; and d) identify an eligible male caregiver to participate in the study. BMC must be a) > 18 years old; b) self-identify as African American, Black, or mixed race with African American or Black; c) speak English since measures are not normed for other languages; d) be a current male caregiver to a Black girl 14-18 years old; and e) the girl’s legal guardian must agree to the male caregiver’s participation. BMC will be defined as men (fathers, grandfathers, uncles, brothers, cousins, etc.) who girls report playing a central role in their care and upbringing. Girls may or may not be sexually active, and information about girls’ sexual activity will not be shared with their male caregivers. Recruitment, assent/consent, compensation and retention. Recruitment will be multi-faceted. CBO will distribute flyers in the community and invite interested individuals to contact the IMAGE study. CBO will also contact eligible participants directly to inform them about the project and request permission for the research team to contact them. IMAGE staff will attend relevant community events to build relationships and conduct outreach. Using a screening and consent script, research staff will follow up with girls separately from guardians or male caregivers to conduct the screening, to obtain informed consent/assent with written, electronic signature via the e-consent module in REDCap, and to confirm the contact information of legal guardians and/or participating male caregivers. We will attempt to obtain girls’ assent/consent before we contact male caregivers, but participants may be screened in any order. Girls’ refusal to participate will precede consent or permission by male caregivers and legal guardians, regardless of the screening order. Dyads will complete surveys via REDCap at baseline, 6 and 12-month follow-up. Baseline survey assessments will occur before Day 1 of the intervention via REDCap. Girls and male caregivers will each receive $60 for completing the baseline assessment and an additional $65 for completing the entire workshop weekend. Participants are compensated $70 for the 6-month follow-up and $75 for the 12-month follow-up to recognize the value of retention over time. To overcome transportation barriers to STI treatment at the University of Illinois in Chicago Hospital Health Sciences System, we will provide an additional $5 in transportation compensation if needed. Participants’ transportation barriers will be assessed when scheduled for a workshop, and participant safety will be ensured. To maximize retention, we will use the strategies that were successful in our pilot (i.e., holiday cards and monthly phone calls or text messages) . We will collect contact information for girls and male caregivers in multiple forms to find the most effective communication for each participant. Randomization, sample size, and power. We will use a stratified block randomization allocation to assign girl-BMC dyads to the treatment arm (IMAGE n = 150 or FUEL n = 150) prior to baseline assessment. Power analyses determined that a sample of 300 girls – 150 IMAGE and 150 FUEL is needed to assess efficacy on the primary outcome (STI incidence). Effect size (Cohen’s h) is a function of two proportions defined by Cohen; h = .2 is a small effect, h = .5 is a medium effect, and h = .8 is a large effect. We calculated the power to detect the effect of IMAGE vs. FUEL in reducing STI incidence as measured by having any of three STIs at 12 months post-intervention using a 2-sided test with alpha = .05. The new STI outcome at follow-up was a binary variable (yes/no to testing positive for at least 1 of the 3 STIs), because the study was not powered to examine specific STI outcomes. Based on the IMARA retention rate of 86% for Black female caregiver-girl dyads in Chicago , we evaluated the power of STI incidence (yes/no) at 80% retention over 12- months. Based on our previous work, we assume the STI incidence is 25% and 11% at 12 months for girls in the FUEL arm and IMAGE arm, respectively, which is a small-to-medium effect size of h = .37. With the proposed sample size, we will achieve 80% power to detect the intervention effect. We also calculated the standard effect size (Cohen’s d) for group mean comparisons with 80% power for our secondary continuous outcomes, where d = 0 is defined as no treatment effect. We can detect a small-to-medium effect size with the proposed sample size (d = .37). Aim 1 efficacy study conditions. Intervention group: IMARA for black male caregivers and girls empowerment (IMAGE) : IMAGE is an 8-10-hour HIV/STI group-based (6-8 dyads) prevention program delivered to girls and BMC, together and separated, reviewing parallel content. Modules address individual, social, and structural drivers of HIV risk ( ). The IMAGE curriculum was tailored specifically for Chicago’s Black girls and male caregivers. The topics discussed in the curriculum include the sexual development of Black girls, the risk of sexual violence, female anatomy, body positivity, HIV/STI knowledge and attitudes, and contraception use. IMAGE was created to foster healthy communication and connection between male caregivers and their girls by encouraging conflict resolution and perspective-taking. Male caregivers learn about the importance of mental health, role modeling, and partner choices. Through IMAGE, male caregivers will become more aware of their presence in their girls’ lives, discuss absenteeism (in the event they are incarcerated or unable to be physically present), and learn to be present. IMAGE emphasizes how Black girls and women are portrayed in the media, which creates unhealthy stereotypes, decreases self-image, and increases vulnerability to sexual violence. Girls and caregivers will discuss intimate partner violence, power dynamics, and healthy relationships, as well as discuss the use of contraceptives, specifically condoms. IMAGE directly addresses structural factors on girls’ HIV/STI risk, namely male incarceration, sexual violence, domestic violence, and negative stereotyping. Woven throughout IMAGE is the impact of mental health, gender-based violence, and alcohol and drug use on HIV risk. Control condition: FUEL, a health promotion program : We used a similar approach as described above (ADAPT-ITT) to tailor FUEL for Black girls and male caregivers in Chicago. FUEL promotes healthy living practices and topics, namely media literacy, healthy eating and physical activity, healthy eating and nutrition, physical activity, teen drug and alcohol use, and violence prevention ( ). Like IMAGE, topics are covered in both separate and joint sessions. FUEL has a brief HIV education module, but it does not address SRH dyad communication or conflict navigation. Video essays, documentaries, and information are used to address relevant topics. Intervention training. Facilitators for both conditions must identify as Black, female, and have some experience working with Black communities or families. Facilitators were assigned to deliver IMAGE or FUEL. All facilitators completed 20 hours of training on their assigned curriculum and received supplemental one-hour training in-person and virtually to introduce key themes of cultural competency, PrEP, HIV prevention education, sex trafficking/human trafficking risk awareness, and general drug safety and navigating substance use disorders. Training for both programs stressed the importance of following the manualized protocol to ensure fidelity across facilitators and completing fidelity measurements at the end of each workshop. Following a curriculum review during the first two training sessions, facilitators practiced delivering each activity, taking turns serving as group leaders and participants. Each facilitator was “certified” as competent to deliver IMAGE or FUEL by the principal investigator and project director. Aim 1 efficacy trial. Six to eight dyads will be assigned to workshops randomized into FUEL or IMAGE at each CBO and will receive the corresponding curriculum on a designated weekend. Four facilitators and two observers are needed for each study condition; facilitators are not to be cross-trained in IMAGE and FUEL. Two facilitators co-lead the girls’ groups, and two co-lead the male caregivers’ groups. All four facilitators co-lead the joint sessions when girls and male caregivers come together for specific activities. Breakfast, lunch, and snacks are provided at all workshops for facilitators and participants. Supervision, quality assurance, and treatment. We will use a detailed manual and facilitator guide, conduct weekly supervision, and assess fidelity using facilitator self-report and observer ratings of adherence and competence (see below). Adherence measures determine if the program was delivered as intended (yes/no), and competence ratings indicate the quality of delivery. Each workshop is delivered by four facilitators and observed by two trained individuals to ensure fidelity and report on the quality of the workshop. Observers and facilitators will complete surveys after each workshop day, reflecting on the quality of facilitation. Participants will complete workshop evaluations. During ongoing supervision, the IMAGE research team will review fidelity reports and participant evaluations, discuss session activities and problem-solving difficulties, and provide feedback. If facilitators deviate from the curriculum, the supervisor will provide additional training until fidelity is achieved. We will complete observations of at least 40% of study workshops. Contamination. We will minimize contamination across arms. Facilitators will be assigned to IMAGE or FUEL and will not discuss or learn the other arm’s curriculum; training is done separately. The workshops will not be conducted on the same days; weekends will be divided, and each CBO will have only FUEL or IMAGE weekends to reduce participant contact. Confidentiality and a general orientation to the research structure were emphasized during facilitation training, allowing us to expect limited contamination.
We will assign 300 14-18-year-old Black girls and their male caregivers to workshops randomized to either IMAGE or FUEL and assess girls’ STI incidence, sexual partners, and condom use at baseline, 6- and 12-months. We hypothesize that girls who receive IMAGE will have lower STI incidence (primary outcome), fewer sexual partners, and more consistent condom use (secondary outcomes). We will explore associations of individual, interpersonal, and structural factors proposed by the B ecoming a Sexual Black Woman and HDRF framework on primary and secondary outcomes at all three time points ( ).
We will identify and describe factors (barriers, facilitators, constraints) and implementation processes at six CBOs across Chicago.
We will deliver IMAGE at six CBOs on Chicago’s West and South sides impacted by structural factors (i.e., racism, poverty, crime, incarceration, and urban decay) implicated in high HIV and STI rates among Black communities. Neighborhoods in these areas are predominately Black, and Black females in these areas report the 4th highest rate of ever having sex (53%), the 3rd highest rate of sex before age 13, the 2nd highest rate of current sexual activity (40.6%) and have the highest rates of Chlamydia (1.921.0-2,688.4 per 100,000 people), and gonorrhea (973.7 - 1,536.2 per 100,000 people) . Rates of sexual violence are also high in these neighborhoods . The CBOs offer social services that address community resource needs and promote health and behavior change.
Girls must be a) 14-18 years old; b) self-identify as African American, Black, or mixed race with African American or Black; c) speak English since measures are not normed for other languages; and d) identify an eligible male caregiver to participate in the study. BMC must be a) > 18 years old; b) self-identify as African American, Black, or mixed race with African American or Black; c) speak English since measures are not normed for other languages; d) be a current male caregiver to a Black girl 14-18 years old; and e) the girl’s legal guardian must agree to the male caregiver’s participation. BMC will be defined as men (fathers, grandfathers, uncles, brothers, cousins, etc.) who girls report playing a central role in their care and upbringing. Girls may or may not be sexually active, and information about girls’ sexual activity will not be shared with their male caregivers.
Recruitment will be multi-faceted. CBO will distribute flyers in the community and invite interested individuals to contact the IMAGE study. CBO will also contact eligible participants directly to inform them about the project and request permission for the research team to contact them. IMAGE staff will attend relevant community events to build relationships and conduct outreach. Using a screening and consent script, research staff will follow up with girls separately from guardians or male caregivers to conduct the screening, to obtain informed consent/assent with written, electronic signature via the e-consent module in REDCap, and to confirm the contact information of legal guardians and/or participating male caregivers. We will attempt to obtain girls’ assent/consent before we contact male caregivers, but participants may be screened in any order. Girls’ refusal to participate will precede consent or permission by male caregivers and legal guardians, regardless of the screening order. Dyads will complete surveys via REDCap at baseline, 6 and 12-month follow-up. Baseline survey assessments will occur before Day 1 of the intervention via REDCap. Girls and male caregivers will each receive $60 for completing the baseline assessment and an additional $65 for completing the entire workshop weekend. Participants are compensated $70 for the 6-month follow-up and $75 for the 12-month follow-up to recognize the value of retention over time. To overcome transportation barriers to STI treatment at the University of Illinois in Chicago Hospital Health Sciences System, we will provide an additional $5 in transportation compensation if needed. Participants’ transportation barriers will be assessed when scheduled for a workshop, and participant safety will be ensured. To maximize retention, we will use the strategies that were successful in our pilot (i.e., holiday cards and monthly phone calls or text messages) . We will collect contact information for girls and male caregivers in multiple forms to find the most effective communication for each participant.
We will use a stratified block randomization allocation to assign girl-BMC dyads to the treatment arm (IMAGE n = 150 or FUEL n = 150) prior to baseline assessment. Power analyses determined that a sample of 300 girls – 150 IMAGE and 150 FUEL is needed to assess efficacy on the primary outcome (STI incidence). Effect size (Cohen’s h) is a function of two proportions defined by Cohen; h = .2 is a small effect, h = .5 is a medium effect, and h = .8 is a large effect. We calculated the power to detect the effect of IMAGE vs. FUEL in reducing STI incidence as measured by having any of three STIs at 12 months post-intervention using a 2-sided test with alpha = .05. The new STI outcome at follow-up was a binary variable (yes/no to testing positive for at least 1 of the 3 STIs), because the study was not powered to examine specific STI outcomes. Based on the IMARA retention rate of 86% for Black female caregiver-girl dyads in Chicago , we evaluated the power of STI incidence (yes/no) at 80% retention over 12- months. Based on our previous work, we assume the STI incidence is 25% and 11% at 12 months for girls in the FUEL arm and IMAGE arm, respectively, which is a small-to-medium effect size of h = .37. With the proposed sample size, we will achieve 80% power to detect the intervention effect. We also calculated the standard effect size (Cohen’s d) for group mean comparisons with 80% power for our secondary continuous outcomes, where d = 0 is defined as no treatment effect. We can detect a small-to-medium effect size with the proposed sample size (d = .37).
Intervention group: IMARA for black male caregivers and girls empowerment (IMAGE) : IMAGE is an 8-10-hour HIV/STI group-based (6-8 dyads) prevention program delivered to girls and BMC, together and separated, reviewing parallel content. Modules address individual, social, and structural drivers of HIV risk ( ). The IMAGE curriculum was tailored specifically for Chicago’s Black girls and male caregivers. The topics discussed in the curriculum include the sexual development of Black girls, the risk of sexual violence, female anatomy, body positivity, HIV/STI knowledge and attitudes, and contraception use. IMAGE was created to foster healthy communication and connection between male caregivers and their girls by encouraging conflict resolution and perspective-taking. Male caregivers learn about the importance of mental health, role modeling, and partner choices. Through IMAGE, male caregivers will become more aware of their presence in their girls’ lives, discuss absenteeism (in the event they are incarcerated or unable to be physically present), and learn to be present. IMAGE emphasizes how Black girls and women are portrayed in the media, which creates unhealthy stereotypes, decreases self-image, and increases vulnerability to sexual violence. Girls and caregivers will discuss intimate partner violence, power dynamics, and healthy relationships, as well as discuss the use of contraceptives, specifically condoms. IMAGE directly addresses structural factors on girls’ HIV/STI risk, namely male incarceration, sexual violence, domestic violence, and negative stereotyping. Woven throughout IMAGE is the impact of mental health, gender-based violence, and alcohol and drug use on HIV risk. Control condition: FUEL, a health promotion program : We used a similar approach as described above (ADAPT-ITT) to tailor FUEL for Black girls and male caregivers in Chicago. FUEL promotes healthy living practices and topics, namely media literacy, healthy eating and physical activity, healthy eating and nutrition, physical activity, teen drug and alcohol use, and violence prevention ( ). Like IMAGE, topics are covered in both separate and joint sessions. FUEL has a brief HIV education module, but it does not address SRH dyad communication or conflict navigation. Video essays, documentaries, and information are used to address relevant topics.
Facilitators for both conditions must identify as Black, female, and have some experience working with Black communities or families. Facilitators were assigned to deliver IMAGE or FUEL. All facilitators completed 20 hours of training on their assigned curriculum and received supplemental one-hour training in-person and virtually to introduce key themes of cultural competency, PrEP, HIV prevention education, sex trafficking/human trafficking risk awareness, and general drug safety and navigating substance use disorders. Training for both programs stressed the importance of following the manualized protocol to ensure fidelity across facilitators and completing fidelity measurements at the end of each workshop. Following a curriculum review during the first two training sessions, facilitators practiced delivering each activity, taking turns serving as group leaders and participants. Each facilitator was “certified” as competent to deliver IMAGE or FUEL by the principal investigator and project director.
Six to eight dyads will be assigned to workshops randomized into FUEL or IMAGE at each CBO and will receive the corresponding curriculum on a designated weekend. Four facilitators and two observers are needed for each study condition; facilitators are not to be cross-trained in IMAGE and FUEL. Two facilitators co-lead the girls’ groups, and two co-lead the male caregivers’ groups. All four facilitators co-lead the joint sessions when girls and male caregivers come together for specific activities. Breakfast, lunch, and snacks are provided at all workshops for facilitators and participants.
We will use a detailed manual and facilitator guide, conduct weekly supervision, and assess fidelity using facilitator self-report and observer ratings of adherence and competence (see below). Adherence measures determine if the program was delivered as intended (yes/no), and competence ratings indicate the quality of delivery. Each workshop is delivered by four facilitators and observed by two trained individuals to ensure fidelity and report on the quality of the workshop. Observers and facilitators will complete surveys after each workshop day, reflecting on the quality of facilitation. Participants will complete workshop evaluations. During ongoing supervision, the IMAGE research team will review fidelity reports and participant evaluations, discuss session activities and problem-solving difficulties, and provide feedback. If facilitators deviate from the curriculum, the supervisor will provide additional training until fidelity is achieved. We will complete observations of at least 40% of study workshops.
We will minimize contamination across arms. Facilitators will be assigned to IMAGE or FUEL and will not discuss or learn the other arm’s curriculum; training is done separately. The workshops will not be conducted on the same days; weekends will be divided, and each CBO will have only FUEL or IMAGE weekends to reduce participant contact. Confidentiality and a general orientation to the research structure were emphasized during facilitation training, allowing us to expect limited contamination.
We will utilize a 3-Step Implementation Model and its implementation guide at each CBO to Prepare, Rollout, and equip them to Sustain IMAGE. The three steps are outlined more in the implementation procedures section.
Our team has previously developed partnerships with directors at six CBOs. The curriculum is designed to be flexible and capable of being delivered in various community settings. We will conduct site visits with each director, including environmental scans of prerequisite resources for hosting workshops. Each CBO director will appoint an internal liaison to lead recruitment activities and bridge communication between the sites and the IMAGE team, and the project director will interview liaisons to collect qualitative preparedness data. The information will be collected during site visits, interviews, and a brief prep phase survey to write a tailored implementation plan for each site per year. The implementation plan will then act as a living reference to guide workshop scheduling and execution for the year.
Each CBO hosts an equal number of IMAGE and FUEL workshops for the efficacy study (Aim 1).
With support and assistance from the implementation team, the CBO will review their experiences and decide if they will continue offering IMAGE at their location.
Each CBO follows the same sequence of steps to Prepare, Rollout, and Sustain IMAGE. The study will began with the Prepare phase, which involved engaging with the CBOs. During Rollout, the efficacy data (Aim 1) is collected across the 3 time points: baseline, 6, and 12 months. Aim 2 will formally evaluate and compare the implementation determinants (barriers, facilitators, constraints) and processes within and across six CBOs using data generated from surveys, observations, study notes, and interviews with CBO liaisons and directors, hereafter CBO personnel. CBO personnel (n = 12) will be asked and consented to participate in surveys and qualitative interviews. Each CBO director will select a liaison who is the reference contact for everything related to the workshops. CBO liaisons will oversee recruiting dyads and bridging the CBO and the IMAGE team. Inclusion criteria: All CBO directors (n = 6) and CBO IMAGE liaisons (n = 6) will be eligible. Because CBO liaisons will receive a stipend from UIC and are also CBO employees, we emphasize that participation in the interviews and surveys is completely voluntary. Personnel can refuse to participate without affecting their employment at the CBO or their stipend from UIC. Exclusion criteria: a) inability to understand the consent process, and b) non-employment at a partnering CBO. Regarding CBO personnel turnover (liaisons and/or directors), PI Crooks will engage new directors, and directors will appoint a new individual to serve as the liaison and be trained to perform their activities. Aim 2 implementation procedures. During Prepare , we will conduct an environmental scan at each CBO to describe and compare the inner and outer context across CBOs. We will identify structural and logistical resources (i.e., space, furniture, and other infrastructure). The scan also includes documenting current programming, the structure of decision-making, available resources (e.g., staff, space), and client engagement (e.g., number and frequency of the target population coming to the CBO). We review organizational charts to understand staff roles and determine how IMAGE fits the current structure. We identify the needs of each CBO and how the research team can support IMAGE delivery. We work with CBO directors to identify a liaison at the CBO who can interface with the community and assist in recruitment and implementation. We will share the materials related to IMAGE with the CBO personnel - director and liaison from each CBO (n = 12), and we will ask them to complete a survey about the factors likely to affect IMAGE delivery and potential sustainability (ORIC and organizational climate measures). We will also conduct a 45-60-minute semi-structured interview with CBO liaisons (n = 6) involved in the IMAGE study (face-to-face, phone, or via Zoom) to document anticipated barriers and facilitators. The Prepare phase semi-structured interview only takes place with the liaison as a follow-up to the conversations held during the initial site visit with the CBO Directors. Verbal consent will be obtained before each interview. Participants will receive a $25 gift card for participating in the interview and survey. After the Prepare step, we will triangulate the data to co-create an adaptable implementation plan detailing how to deliver IMAGE at their site. During IMAGE Rollout , we will complete the Aim 1 efficacy study. After the first IMAGE session, we will dispatch a follow-up survey assessing implementation, adoption, and feasibility and conduct 30-45-minute semi-structured in-depth interviews (in person, by phone, or via Zoom) with CBO personnel (n = 12) to document facilitators, barriers and needed adaptations (e.g., to increase IMAGE enrollment and attendance). Verbal consent will be obtained before each interview, and participants will receive a $25 gift card. After the trial ends (end of Rollout), CBO personnel will be asked to complete a third survey focused on acceptability, appropriateness, feasibility, and sustainability. We will conduct 45-60-minute semi-structured interviews (face-to-face, by phone, or via Zoom) with CBO personnel (n = 12) to assess each CBO’s total experience and capacity to Sustain IMAGE. Building on the environmental scan and previous interviews that identified facilitators, barriers, and solutions, we will ask CBO personnel to discuss their ability to integrate IMAGE into current programming. Verbal consent will be obtained before each interview. Participants will receive a $25 gift card.
During Prepare , we will conduct an environmental scan at each CBO to describe and compare the inner and outer context across CBOs. We will identify structural and logistical resources (i.e., space, furniture, and other infrastructure). The scan also includes documenting current programming, the structure of decision-making, available resources (e.g., staff, space), and client engagement (e.g., number and frequency of the target population coming to the CBO). We review organizational charts to understand staff roles and determine how IMAGE fits the current structure. We identify the needs of each CBO and how the research team can support IMAGE delivery. We work with CBO directors to identify a liaison at the CBO who can interface with the community and assist in recruitment and implementation. We will share the materials related to IMAGE with the CBO personnel - director and liaison from each CBO (n = 12), and we will ask them to complete a survey about the factors likely to affect IMAGE delivery and potential sustainability (ORIC and organizational climate measures). We will also conduct a 45-60-minute semi-structured interview with CBO liaisons (n = 6) involved in the IMAGE study (face-to-face, phone, or via Zoom) to document anticipated barriers and facilitators. The Prepare phase semi-structured interview only takes place with the liaison as a follow-up to the conversations held during the initial site visit with the CBO Directors. Verbal consent will be obtained before each interview. Participants will receive a $25 gift card for participating in the interview and survey. After the Prepare step, we will triangulate the data to co-create an adaptable implementation plan detailing how to deliver IMAGE at their site. During IMAGE Rollout , we will complete the Aim 1 efficacy study. After the first IMAGE session, we will dispatch a follow-up survey assessing implementation, adoption, and feasibility and conduct 30-45-minute semi-structured in-depth interviews (in person, by phone, or via Zoom) with CBO personnel (n = 12) to document facilitators, barriers and needed adaptations (e.g., to increase IMAGE enrollment and attendance). Verbal consent will be obtained before each interview, and participants will receive a $25 gift card. After the trial ends (end of Rollout), CBO personnel will be asked to complete a third survey focused on acceptability, appropriateness, feasibility, and sustainability. We will conduct 45-60-minute semi-structured interviews (face-to-face, by phone, or via Zoom) with CBO personnel (n = 12) to assess each CBO’s total experience and capacity to Sustain IMAGE. Building on the environmental scan and previous interviews that identified facilitators, barriers, and solutions, we will ask CBO personnel to discuss their ability to integrate IMAGE into current programming. Verbal consent will be obtained before each interview. Participants will receive a $25 gift card.
Aim 1 measures Primary outcome. Black girls will receive STI testing at the CBOs on day 1 of their workshops. Trained research study staff will collect urine specimens from each girl at CBO sites. Urine specimens are delivered to the University of Illinois in Chicago Hospital Health Sciences System laboratory by a courier service, where trained clinical staff process the specimens for STI testing. The results of the STI testing are posted to EPIC, an electronic health records system. STI testing results are abstracted from EPIC and stored in REDCap. Trained research study staff call girls to communicate their STI testing results. Girls create a code word at the time of specimen collection. Research study staff are trained to confirm the name, date of birth, and the code word to ensure results are not shared with unauthorized parties. Secondary outcomes. Girls and male caregiver dyads will complete baseline, 6-and 12-month assessments via self-report surveys through REDCap. Sexual behavior will be measured by condom use and the number of partners. As a part of Aim 1, we will explore the three constructs of the Becoming a Sexual Black Woman framework , using self-report surveys that capture constructs of phases of sexual development, protection, and stereotyped messaging and measures that reflect all three levels of influence (individual-, interpersonal-, societal- levels) as outlined in the HDRF Framework ( ). Aim 2 implementation measures Aim 2 data will be generated from the environmental scan, direct observations (fidelity ratings), study notes from meetings, semi-structured interviews documenting CBO personnel experiences hosting IMAGE ( ), and surveys ( ) of the CBOs’ attributes. Perceptions of acceptability, feasibility, and appropriateness of IMAGE will be assessed during Preparation after the Rollout phases, as well as views regarding IMAGE ownership during the Sustain phase. Prepare. We will conduct a baseline assessment at each CBO to describe the inner settings, bridging (CBO and UIC partnership), and innovation factors (IMAGE fit within each CBO). Aligning with implementation theories, we will ask CBO personnel to complete the Organizational Readiness for Implementing Change (ORIC) . ORIC includes 12 items that measure confidence, commitment, motivation, and determination in implementing IMAGE. CBO personnel will also complete an Organizational climate assessment, which consists of measures related to stress, workload, strain, and frustration, as these individual factors may influence the capacity to implement IMAGE in each setting. Rollout. Each CBO liaison will meet with the study’s project director weekly to discuss implementation challenges, recruitment, enrollment, rollout, and retention. Trained IMAGE facilitators will rate treatment fidelity of sessions when they act as observers. They will complete measures of adherence (delivery as planned) and competence (quality of delivery) for each session. The fidelity forms contain yes/no questions, including whether facilitators followed the session script/outline, explained and demonstrated each activity, provided corrective feedback, maintained an appropriate pace, and were open, non-judgmental, and comfortable with participants. The IMAGE research team will review fidelity checklists after each session. Areas of concern will be addressed with each facilitator during weekly supervision, and the facilitator(s) will be re-trained to reach fidelity as observed through our described model. Acceptability, feasibility, and appropriateness. We will document Black girls’ and caregivers’ attendance (Day 1 and Day 2) to indicate feasibility and acceptability. We will also select a random sample of five girls and five caregivers who complete IMAGE at each of the six CBOs (n = 60) to participate in qualitative interviews to document their experiences with IMAGE (each will be paid $50). Following consent/assent, we will ask about feasibility, acceptability, appropriateness, strategies to improve attendance, and reasons to deliver IMAGE in other venues. CBO personnel will complete the Acceptability of Intervention Measure (AIM) , Intervention Appropriateness Measure (IAM) , and Feasibility of Intervention Measure (FIM) , at the end of Rollout. The AIM assesses whether IMAGE is agreeable or satisfactory, the IAM assesses the perceived relevance or fits with each CBO’s mission, and the FIM captures the extent to which IMAGE was successfully carried out. Higher scores indicate greater acceptability, appropriateness, and feasibility. Sustain. The Program Sustainability Assessment Tool (PSAT) will be used to evaluate potential IMAGE sustainability. The PSAT will generate a summary report of sustainability at the CBO and help inform sustainability planning. Individual scales represent organizational support, funding stability, positive academic and community-based partnerships, organizational capacity, program evaluation and adoption, stakeholder communication, and strategic planning ( ).
Primary outcome. Black girls will receive STI testing at the CBOs on day 1 of their workshops. Trained research study staff will collect urine specimens from each girl at CBO sites. Urine specimens are delivered to the University of Illinois in Chicago Hospital Health Sciences System laboratory by a courier service, where trained clinical staff process the specimens for STI testing. The results of the STI testing are posted to EPIC, an electronic health records system. STI testing results are abstracted from EPIC and stored in REDCap. Trained research study staff call girls to communicate their STI testing results. Girls create a code word at the time of specimen collection. Research study staff are trained to confirm the name, date of birth, and the code word to ensure results are not shared with unauthorized parties. Secondary outcomes. Girls and male caregiver dyads will complete baseline, 6-and 12-month assessments via self-report surveys through REDCap. Sexual behavior will be measured by condom use and the number of partners. As a part of Aim 1, we will explore the three constructs of the Becoming a Sexual Black Woman framework , using self-report surveys that capture constructs of phases of sexual development, protection, and stereotyped messaging and measures that reflect all three levels of influence (individual-, interpersonal-, societal- levels) as outlined in the HDRF Framework ( ).
Black girls will receive STI testing at the CBOs on day 1 of their workshops. Trained research study staff will collect urine specimens from each girl at CBO sites. Urine specimens are delivered to the University of Illinois in Chicago Hospital Health Sciences System laboratory by a courier service, where trained clinical staff process the specimens for STI testing. The results of the STI testing are posted to EPIC, an electronic health records system. STI testing results are abstracted from EPIC and stored in REDCap. Trained research study staff call girls to communicate their STI testing results. Girls create a code word at the time of specimen collection. Research study staff are trained to confirm the name, date of birth, and the code word to ensure results are not shared with unauthorized parties.
Girls and male caregiver dyads will complete baseline, 6-and 12-month assessments via self-report surveys through REDCap. Sexual behavior will be measured by condom use and the number of partners. As a part of Aim 1, we will explore the three constructs of the Becoming a Sexual Black Woman framework , using self-report surveys that capture constructs of phases of sexual development, protection, and stereotyped messaging and measures that reflect all three levels of influence (individual-, interpersonal-, societal- levels) as outlined in the HDRF Framework ( ).
Aim 2 data will be generated from the environmental scan, direct observations (fidelity ratings), study notes from meetings, semi-structured interviews documenting CBO personnel experiences hosting IMAGE ( ), and surveys ( ) of the CBOs’ attributes. Perceptions of acceptability, feasibility, and appropriateness of IMAGE will be assessed during Preparation after the Rollout phases, as well as views regarding IMAGE ownership during the Sustain phase. Prepare. We will conduct a baseline assessment at each CBO to describe the inner settings, bridging (CBO and UIC partnership), and innovation factors (IMAGE fit within each CBO). Aligning with implementation theories, we will ask CBO personnel to complete the Organizational Readiness for Implementing Change (ORIC) . ORIC includes 12 items that measure confidence, commitment, motivation, and determination in implementing IMAGE. CBO personnel will also complete an Organizational climate assessment, which consists of measures related to stress, workload, strain, and frustration, as these individual factors may influence the capacity to implement IMAGE in each setting. Rollout. Each CBO liaison will meet with the study’s project director weekly to discuss implementation challenges, recruitment, enrollment, rollout, and retention. Trained IMAGE facilitators will rate treatment fidelity of sessions when they act as observers. They will complete measures of adherence (delivery as planned) and competence (quality of delivery) for each session. The fidelity forms contain yes/no questions, including whether facilitators followed the session script/outline, explained and demonstrated each activity, provided corrective feedback, maintained an appropriate pace, and were open, non-judgmental, and comfortable with participants. The IMAGE research team will review fidelity checklists after each session. Areas of concern will be addressed with each facilitator during weekly supervision, and the facilitator(s) will be re-trained to reach fidelity as observed through our described model. Acceptability, feasibility, and appropriateness. We will document Black girls’ and caregivers’ attendance (Day 1 and Day 2) to indicate feasibility and acceptability. We will also select a random sample of five girls and five caregivers who complete IMAGE at each of the six CBOs (n = 60) to participate in qualitative interviews to document their experiences with IMAGE (each will be paid $50). Following consent/assent, we will ask about feasibility, acceptability, appropriateness, strategies to improve attendance, and reasons to deliver IMAGE in other venues. CBO personnel will complete the Acceptability of Intervention Measure (AIM) , Intervention Appropriateness Measure (IAM) , and Feasibility of Intervention Measure (FIM) , at the end of Rollout. The AIM assesses whether IMAGE is agreeable or satisfactory, the IAM assesses the perceived relevance or fits with each CBO’s mission, and the FIM captures the extent to which IMAGE was successfully carried out. Higher scores indicate greater acceptability, appropriateness, and feasibility. Sustain. The Program Sustainability Assessment Tool (PSAT) will be used to evaluate potential IMAGE sustainability. The PSAT will generate a summary report of sustainability at the CBO and help inform sustainability planning. Individual scales represent organizational support, funding stability, positive academic and community-based partnerships, organizational capacity, program evaluation and adoption, stakeholder communication, and strategic planning ( ).
We will conduct a baseline assessment at each CBO to describe the inner settings, bridging (CBO and UIC partnership), and innovation factors (IMAGE fit within each CBO). Aligning with implementation theories, we will ask CBO personnel to complete the Organizational Readiness for Implementing Change (ORIC) . ORIC includes 12 items that measure confidence, commitment, motivation, and determination in implementing IMAGE. CBO personnel will also complete an Organizational climate assessment, which consists of measures related to stress, workload, strain, and frustration, as these individual factors may influence the capacity to implement IMAGE in each setting.
Each CBO liaison will meet with the study’s project director weekly to discuss implementation challenges, recruitment, enrollment, rollout, and retention. Trained IMAGE facilitators will rate treatment fidelity of sessions when they act as observers. They will complete measures of adherence (delivery as planned) and competence (quality of delivery) for each session. The fidelity forms contain yes/no questions, including whether facilitators followed the session script/outline, explained and demonstrated each activity, provided corrective feedback, maintained an appropriate pace, and were open, non-judgmental, and comfortable with participants. The IMAGE research team will review fidelity checklists after each session. Areas of concern will be addressed with each facilitator during weekly supervision, and the facilitator(s) will be re-trained to reach fidelity as observed through our described model.
We will document Black girls’ and caregivers’ attendance (Day 1 and Day 2) to indicate feasibility and acceptability. We will also select a random sample of five girls and five caregivers who complete IMAGE at each of the six CBOs (n = 60) to participate in qualitative interviews to document their experiences with IMAGE (each will be paid $50). Following consent/assent, we will ask about feasibility, acceptability, appropriateness, strategies to improve attendance, and reasons to deliver IMAGE in other venues. CBO personnel will complete the Acceptability of Intervention Measure (AIM) , Intervention Appropriateness Measure (IAM) , and Feasibility of Intervention Measure (FIM) , at the end of Rollout. The AIM assesses whether IMAGE is agreeable or satisfactory, the IAM assesses the perceived relevance or fits with each CBO’s mission, and the FIM captures the extent to which IMAGE was successfully carried out. Higher scores indicate greater acceptability, appropriateness, and feasibility.
The Program Sustainability Assessment Tool (PSAT) will be used to evaluate potential IMAGE sustainability. The PSAT will generate a summary report of sustainability at the CBO and help inform sustainability planning. Individual scales represent organizational support, funding stability, positive academic and community-based partnerships, organizational capacity, program evaluation and adoption, stakeholder communication, and strategic planning ( ).
Aim 1 analysis Data reduction and preliminary analyses. We will create summary scores for (1) Black girl-male caregiver relationship (quality and attachment), (2) Black girl-male caregiver communication (quality and quantity), and (3) girls’ risk behavior and attitudes (sexual behavior and condom use) . We will create two separate forms of the dependent variable for girls: a composite score comprised of multiple risk indicators and a set of individual indicators (e.g., non-condom use, number of partners). We will compare groups on baseline variables using ANOVA (continuous) and Chi-square tests (categorical) and control for potential confounding variables in subsequent regression analyses that are not balanced by baseline randomization. General statistical issues. We will use Bonferroni correction to adjust for multiple comparisons as we will analyze multiple outcomes. We will control for Type I error in the analysis of the secondary outcomes but not the primary outcomes to capture important findings that may be obscured from the conservative limits imposed by Bonferroni correction. We will explore whether girls’ sexual behaviors are related to potentially important covariates (e.g., pubertal status) and use regression models to test their effects. To evaluate the impact of Black male caregivers on girls, we will include BMC’s characteristics as covariates in the regression model for girls’ outcomes. We will use a similar approach to evaluate the impact of girls on BMC over time. Missing data will most likely occur because of subject attrition. Based on our intervention studies, we expect 70%-85% retention in both arms. We will evaluate associations between dose (1 vs. 2 sessions) and efficacy. Missingness complicates statistical analyses via biased parameter estimates, reduced statistical power, and degraded confidence intervals. Based on our prior research, we expect minimal missingness in the data collection, but we will consider the item and unit nonresponse. An analysis using only completers is generally valid, assuming data are missing completely at random, but it is inefficient because it discards data points observed for non-completers. We will use all available data and produce valid inferences under the assumption that data are randomly missing. We will quantify the potential bias of these inferences should we suspect missing data to be non-random. Compare IMAGE vs. FUEL TM on STI incidence and risky sexual behavior: We will evaluate treatment effects separately at 6- months and 12 months. We will also combine data to evaluate an average effect across both time points and to analyze the change from 6- to 12 months. We will analyze binary outcomes using logistic and continuous outcomes using linear regression. At a single time point, we will test the effect of the binary indicator for IMAGE vs. FUEL on the outcome, adjusting for confounders and the outcome at baseline as additional independent variables. We will use mixed-effects regression models to analyze the outcome by combining both time points, and we will include a random intercept term to account for intra-subject correlation. We will test an average effect across time points and effects on patterns of change from 6- to 12 months by including an interaction term between treatment and time. We will use survival analysis to examine sexual initiation over the course of the study. We hypothesize that non-sexually experienced girls in IMAGE will, on average, initiate sex later than FUEL participants. We will estimate the Kaplan-Meier survival curves and conduct long-rank tests to compare the two survival curves. We will use Cox regression models and its extensions (e.g., Cox’s Proportional Hazard Model with Time-Dependent Covariates) to test for the effects of other covariates such as age or socioeconomic status. Exploratory analysis. Driven by the Becoming a Sexual Black Woman framework, we will use the regression-based causal mediation analysis approach to test whether each of the individual, interpersonal, and structural (societal) level factors serve as a mediator of the intervention effect on the STI incidence, separately. Specifically, we will adopt the logistic regression model for the outcome and linear regression models for the mediators. We will report the controlled direct effect, natural direct and indirect effects, and proportion mediated with their 95% confidence intervals for each factor. We will adjust for sociodemographic characteristics collected at baseline as potential confounders in the outcome and mediator regression models. We will analyze sensitivity for unmeasured mediator-outcome confounding and report the mediational E-values . We will also use a log-linear regression model for the outcome as a sensitivity analysis since our binary outcome may not be rare. We hypothesize that protection, represented by interpersonal factors such as dyad relationships, sexual communication, family risk, and protection ( ), would have the strongest mediation effects as it is most central to the Becoming a Sexual Black Woman Framework. Aim 2 mixed method data analysis Quantitative analysis. Survey data will be analyzed by our statistician using descriptive statistics (SAS, version 9.4), and an aggregated mean total score and means for each CBO will be calculated. Using a continuous and iterative process, we will identify contextual factors (events or statements) from observations (study notes and fidelity forms) and interviews to document what facilitates or acts as barriers to implementing IMAGE. Qualitative analysis. To begin the qualitative analysis, an RA and a study team member trained by the PI will immerse themselves in the data by reading and re-reading the transcripts and noting the interviewees’ perceptions. Using the Dedoose (Version 8.2.32) and following an approach described by Miles and Huberman , a set of initial codes grouped into broad domains reflecting the interview guide and EPIS framework (i.e., groups of related codes) will be developed. EPIS constructs will include Outer context – relationships between entities, including governments and funders; Inner context – the structure of CBOs, culture, networks, communication, climate, and readiness for implementation; and Bridging factors –the relationship between CBOs and UIC and Innovation – CBO and IMAGE fit and sustainability. The study team member will open code two transcripts simultaneously to refine codes into a preliminary codebook with clear operational definitions. Interviews will be separately coded, and then they will consult with the PI and other team members to review discrepancies, refine code definitions, and recode until intercoder reliability exceeds 85% . Final codes will be compiled in the master codebook and applied to qualitative data coding. The research team will collaboratively analyze results, discuss codes, categories, and themes generated, and resolve discrepancies through discussions. This iterative process will allow us to identify the most salient contextual factors (events or statements) from observations (study notes) and interviews and document implementation barriers and facilitators (e.g., challenges, resolutions, impacts of champions, leadership, etc.). Final categories and themes will guide any necessary revisions of implementation procedures and the CBO’s implementation plan. We expect to identify shared and unique experiences from each CBO. We will triangulate qualitative data with quantitative measures related to treatment delivery and receipt of treatment. Together, these data will provide insight from all perspectives on program success and the potential future integration and sustainability of IMAGE by CBOs. Measures supporting rigor and trustworthiness in qualitative research include a detailed audit trail, study notes, and reflexivity notes. Each audio recording will be transcribed and checked for accuracy. Data management Project protocols promote proper and timely data preparation for analysis and secure data storage. STI results, survey, and fidelity data will be collected in REDCap, with paper instruments used only in community settings without web access. All paper data will be entered into REDCap and destroyed once in the system. REDCap is a secure web-based data collection and management application for UIC faculty. The REDCap server is maintained by the Institute for Health Research and Policy staff at [blinded for review]. Data monitoring A trained data manager will monitor the data and review participant records, screening and consent documents, and data collection forms. We will assemble a Data Safety and Monitoring Board to review our activities to ensure participant safety and evaluate findings in an interim data analysis to determine if the RCT should continue or be stopped. Any protocol modifications or study amendments will be reported to the IRB. Adverse events Adverse events related to the research are not expected. The facilitation role is curated to monitor potential study harm. The IMAGE team and staff will document and report any unanticipated harm to subjects within 24 hours to the PI and the IRB. Dissemination plan A series of publications will be published in the first year. In the second and third years, team members will identify areas of interest, generate outlines, and submit a qualitative paper. A grant writing workshop, led by the PI, will be held for our CBOs in year 4 to support sustainment in the future. In the 5th year, we will invite members of each CBO and other agencies involved in supporting Black families in Chicago to attend a workshop and share study results. Results shared in year five will allow organizations and policymakers to make informed decisions regarding IMAGE. Dissemination of this high-impact study is important because it will increase familial protective factors and reduce HIV/STI incidence for Black girls ( ). Study status When this manuscript was submitted for publication, the study was underway. All six partnering CBOs have completed the Prepare step. We’ve trained 16 facilitators and two supporting team staff members on FUEL and/or IMAGE. The first IMAGE workshop weekend took place in early October 2024. We have conducted four workshops at two CBO sites across Chicagoland. A total of 15 dyads are enrolled and/or awaiting research team contact.
Data reduction and preliminary analyses. We will create summary scores for (1) Black girl-male caregiver relationship (quality and attachment), (2) Black girl-male caregiver communication (quality and quantity), and (3) girls’ risk behavior and attitudes (sexual behavior and condom use) . We will create two separate forms of the dependent variable for girls: a composite score comprised of multiple risk indicators and a set of individual indicators (e.g., non-condom use, number of partners). We will compare groups on baseline variables using ANOVA (continuous) and Chi-square tests (categorical) and control for potential confounding variables in subsequent regression analyses that are not balanced by baseline randomization. General statistical issues. We will use Bonferroni correction to adjust for multiple comparisons as we will analyze multiple outcomes. We will control for Type I error in the analysis of the secondary outcomes but not the primary outcomes to capture important findings that may be obscured from the conservative limits imposed by Bonferroni correction. We will explore whether girls’ sexual behaviors are related to potentially important covariates (e.g., pubertal status) and use regression models to test their effects. To evaluate the impact of Black male caregivers on girls, we will include BMC’s characteristics as covariates in the regression model for girls’ outcomes. We will use a similar approach to evaluate the impact of girls on BMC over time. Missing data will most likely occur because of subject attrition. Based on our intervention studies, we expect 70%-85% retention in both arms. We will evaluate associations between dose (1 vs. 2 sessions) and efficacy. Missingness complicates statistical analyses via biased parameter estimates, reduced statistical power, and degraded confidence intervals. Based on our prior research, we expect minimal missingness in the data collection, but we will consider the item and unit nonresponse. An analysis using only completers is generally valid, assuming data are missing completely at random, but it is inefficient because it discards data points observed for non-completers. We will use all available data and produce valid inferences under the assumption that data are randomly missing. We will quantify the potential bias of these inferences should we suspect missing data to be non-random. Compare IMAGE vs. FUEL TM on STI incidence and risky sexual behavior: We will evaluate treatment effects separately at 6- months and 12 months. We will also combine data to evaluate an average effect across both time points and to analyze the change from 6- to 12 months. We will analyze binary outcomes using logistic and continuous outcomes using linear regression. At a single time point, we will test the effect of the binary indicator for IMAGE vs. FUEL on the outcome, adjusting for confounders and the outcome at baseline as additional independent variables. We will use mixed-effects regression models to analyze the outcome by combining both time points, and we will include a random intercept term to account for intra-subject correlation. We will test an average effect across time points and effects on patterns of change from 6- to 12 months by including an interaction term between treatment and time. We will use survival analysis to examine sexual initiation over the course of the study. We hypothesize that non-sexually experienced girls in IMAGE will, on average, initiate sex later than FUEL participants. We will estimate the Kaplan-Meier survival curves and conduct long-rank tests to compare the two survival curves. We will use Cox regression models and its extensions (e.g., Cox’s Proportional Hazard Model with Time-Dependent Covariates) to test for the effects of other covariates such as age or socioeconomic status. Exploratory analysis. Driven by the Becoming a Sexual Black Woman framework, we will use the regression-based causal mediation analysis approach to test whether each of the individual, interpersonal, and structural (societal) level factors serve as a mediator of the intervention effect on the STI incidence, separately. Specifically, we will adopt the logistic regression model for the outcome and linear regression models for the mediators. We will report the controlled direct effect, natural direct and indirect effects, and proportion mediated with their 95% confidence intervals for each factor. We will adjust for sociodemographic characteristics collected at baseline as potential confounders in the outcome and mediator regression models. We will analyze sensitivity for unmeasured mediator-outcome confounding and report the mediational E-values . We will also use a log-linear regression model for the outcome as a sensitivity analysis since our binary outcome may not be rare. We hypothesize that protection, represented by interpersonal factors such as dyad relationships, sexual communication, family risk, and protection ( ), would have the strongest mediation effects as it is most central to the Becoming a Sexual Black Woman Framework.
We will create summary scores for (1) Black girl-male caregiver relationship (quality and attachment), (2) Black girl-male caregiver communication (quality and quantity), and (3) girls’ risk behavior and attitudes (sexual behavior and condom use) . We will create two separate forms of the dependent variable for girls: a composite score comprised of multiple risk indicators and a set of individual indicators (e.g., non-condom use, number of partners). We will compare groups on baseline variables using ANOVA (continuous) and Chi-square tests (categorical) and control for potential confounding variables in subsequent regression analyses that are not balanced by baseline randomization.
We will use Bonferroni correction to adjust for multiple comparisons as we will analyze multiple outcomes. We will control for Type I error in the analysis of the secondary outcomes but not the primary outcomes to capture important findings that may be obscured from the conservative limits imposed by Bonferroni correction. We will explore whether girls’ sexual behaviors are related to potentially important covariates (e.g., pubertal status) and use regression models to test their effects. To evaluate the impact of Black male caregivers on girls, we will include BMC’s characteristics as covariates in the regression model for girls’ outcomes. We will use a similar approach to evaluate the impact of girls on BMC over time. Missing data will most likely occur because of subject attrition. Based on our intervention studies, we expect 70%-85% retention in both arms. We will evaluate associations between dose (1 vs. 2 sessions) and efficacy. Missingness complicates statistical analyses via biased parameter estimates, reduced statistical power, and degraded confidence intervals. Based on our prior research, we expect minimal missingness in the data collection, but we will consider the item and unit nonresponse. An analysis using only completers is generally valid, assuming data are missing completely at random, but it is inefficient because it discards data points observed for non-completers. We will use all available data and produce valid inferences under the assumption that data are randomly missing. We will quantify the potential bias of these inferences should we suspect missing data to be non-random. Compare IMAGE vs. FUEL TM on STI incidence and risky sexual behavior: We will evaluate treatment effects separately at 6- months and 12 months. We will also combine data to evaluate an average effect across both time points and to analyze the change from 6- to 12 months. We will analyze binary outcomes using logistic and continuous outcomes using linear regression. At a single time point, we will test the effect of the binary indicator for IMAGE vs. FUEL on the outcome, adjusting for confounders and the outcome at baseline as additional independent variables. We will use mixed-effects regression models to analyze the outcome by combining both time points, and we will include a random intercept term to account for intra-subject correlation. We will test an average effect across time points and effects on patterns of change from 6- to 12 months by including an interaction term between treatment and time. We will use survival analysis to examine sexual initiation over the course of the study. We hypothesize that non-sexually experienced girls in IMAGE will, on average, initiate sex later than FUEL participants. We will estimate the Kaplan-Meier survival curves and conduct long-rank tests to compare the two survival curves. We will use Cox regression models and its extensions (e.g., Cox’s Proportional Hazard Model with Time-Dependent Covariates) to test for the effects of other covariates such as age or socioeconomic status.
Driven by the Becoming a Sexual Black Woman framework, we will use the regression-based causal mediation analysis approach to test whether each of the individual, interpersonal, and structural (societal) level factors serve as a mediator of the intervention effect on the STI incidence, separately. Specifically, we will adopt the logistic regression model for the outcome and linear regression models for the mediators. We will report the controlled direct effect, natural direct and indirect effects, and proportion mediated with their 95% confidence intervals for each factor. We will adjust for sociodemographic characteristics collected at baseline as potential confounders in the outcome and mediator regression models. We will analyze sensitivity for unmeasured mediator-outcome confounding and report the mediational E-values . We will also use a log-linear regression model for the outcome as a sensitivity analysis since our binary outcome may not be rare. We hypothesize that protection, represented by interpersonal factors such as dyad relationships, sexual communication, family risk, and protection ( ), would have the strongest mediation effects as it is most central to the Becoming a Sexual Black Woman Framework.
Quantitative analysis. Survey data will be analyzed by our statistician using descriptive statistics (SAS, version 9.4), and an aggregated mean total score and means for each CBO will be calculated. Using a continuous and iterative process, we will identify contextual factors (events or statements) from observations (study notes and fidelity forms) and interviews to document what facilitates or acts as barriers to implementing IMAGE. Qualitative analysis. To begin the qualitative analysis, an RA and a study team member trained by the PI will immerse themselves in the data by reading and re-reading the transcripts and noting the interviewees’ perceptions. Using the Dedoose (Version 8.2.32) and following an approach described by Miles and Huberman , a set of initial codes grouped into broad domains reflecting the interview guide and EPIS framework (i.e., groups of related codes) will be developed. EPIS constructs will include Outer context – relationships between entities, including governments and funders; Inner context – the structure of CBOs, culture, networks, communication, climate, and readiness for implementation; and Bridging factors –the relationship between CBOs and UIC and Innovation – CBO and IMAGE fit and sustainability. The study team member will open code two transcripts simultaneously to refine codes into a preliminary codebook with clear operational definitions. Interviews will be separately coded, and then they will consult with the PI and other team members to review discrepancies, refine code definitions, and recode until intercoder reliability exceeds 85% . Final codes will be compiled in the master codebook and applied to qualitative data coding. The research team will collaboratively analyze results, discuss codes, categories, and themes generated, and resolve discrepancies through discussions. This iterative process will allow us to identify the most salient contextual factors (events or statements) from observations (study notes) and interviews and document implementation barriers and facilitators (e.g., challenges, resolutions, impacts of champions, leadership, etc.). Final categories and themes will guide any necessary revisions of implementation procedures and the CBO’s implementation plan. We expect to identify shared and unique experiences from each CBO. We will triangulate qualitative data with quantitative measures related to treatment delivery and receipt of treatment. Together, these data will provide insight from all perspectives on program success and the potential future integration and sustainability of IMAGE by CBOs. Measures supporting rigor and trustworthiness in qualitative research include a detailed audit trail, study notes, and reflexivity notes. Each audio recording will be transcribed and checked for accuracy.
Survey data will be analyzed by our statistician using descriptive statistics (SAS, version 9.4), and an aggregated mean total score and means for each CBO will be calculated. Using a continuous and iterative process, we will identify contextual factors (events or statements) from observations (study notes and fidelity forms) and interviews to document what facilitates or acts as barriers to implementing IMAGE.
To begin the qualitative analysis, an RA and a study team member trained by the PI will immerse themselves in the data by reading and re-reading the transcripts and noting the interviewees’ perceptions. Using the Dedoose (Version 8.2.32) and following an approach described by Miles and Huberman , a set of initial codes grouped into broad domains reflecting the interview guide and EPIS framework (i.e., groups of related codes) will be developed. EPIS constructs will include Outer context – relationships between entities, including governments and funders; Inner context – the structure of CBOs, culture, networks, communication, climate, and readiness for implementation; and Bridging factors –the relationship between CBOs and UIC and Innovation – CBO and IMAGE fit and sustainability. The study team member will open code two transcripts simultaneously to refine codes into a preliminary codebook with clear operational definitions. Interviews will be separately coded, and then they will consult with the PI and other team members to review discrepancies, refine code definitions, and recode until intercoder reliability exceeds 85% . Final codes will be compiled in the master codebook and applied to qualitative data coding. The research team will collaboratively analyze results, discuss codes, categories, and themes generated, and resolve discrepancies through discussions. This iterative process will allow us to identify the most salient contextual factors (events or statements) from observations (study notes) and interviews and document implementation barriers and facilitators (e.g., challenges, resolutions, impacts of champions, leadership, etc.). Final categories and themes will guide any necessary revisions of implementation procedures and the CBO’s implementation plan. We expect to identify shared and unique experiences from each CBO. We will triangulate qualitative data with quantitative measures related to treatment delivery and receipt of treatment. Together, these data will provide insight from all perspectives on program success and the potential future integration and sustainability of IMAGE by CBOs. Measures supporting rigor and trustworthiness in qualitative research include a detailed audit trail, study notes, and reflexivity notes. Each audio recording will be transcribed and checked for accuracy.
Project protocols promote proper and timely data preparation for analysis and secure data storage. STI results, survey, and fidelity data will be collected in REDCap, with paper instruments used only in community settings without web access. All paper data will be entered into REDCap and destroyed once in the system. REDCap is a secure web-based data collection and management application for UIC faculty. The REDCap server is maintained by the Institute for Health Research and Policy staff at [blinded for review].
A trained data manager will monitor the data and review participant records, screening and consent documents, and data collection forms. We will assemble a Data Safety and Monitoring Board to review our activities to ensure participant safety and evaluate findings in an interim data analysis to determine if the RCT should continue or be stopped. Any protocol modifications or study amendments will be reported to the IRB.
Adverse events related to the research are not expected. The facilitation role is curated to monitor potential study harm. The IMAGE team and staff will document and report any unanticipated harm to subjects within 24 hours to the PI and the IRB.
A series of publications will be published in the first year. In the second and third years, team members will identify areas of interest, generate outlines, and submit a qualitative paper. A grant writing workshop, led by the PI, will be held for our CBOs in year 4 to support sustainment in the future. In the 5th year, we will invite members of each CBO and other agencies involved in supporting Black families in Chicago to attend a workshop and share study results. Results shared in year five will allow organizations and policymakers to make informed decisions regarding IMAGE. Dissemination of this high-impact study is important because it will increase familial protective factors and reduce HIV/STI incidence for Black girls ( ).
When this manuscript was submitted for publication, the study was underway. All six partnering CBOs have completed the Prepare step. We’ve trained 16 facilitators and two supporting team staff members on FUEL and/or IMAGE. The first IMAGE workshop weekend took place in early October 2024. We have conducted four workshops at two CBO sites across Chicagoland. A total of 15 dyads are enrolled and/or awaiting research team contact.
This study will address a compelling need for innovative multilevel interventions to improve HIV/STI prevention for Black girls and lay the foundation for intervention sustainability in community settings. The study also fills a significant gap in HIV/STI programming for Black girls. Most family-based SRH interventions for Black girls exclude male caregivers due to structural factors that reduce opportunities for them to protect and support Black girls’ SRH. IMAGE offers a safe and structured environment for girls and male caregivers to develop, practice, and create effective communication skills that can be used in their relationships. IMAGE enhances BMC’s self-confidence in communicating with teen girls about sex and relationships while challenging societal norms and sexualized stereotypes that interfere with safe sex practices. We describe the design of an RCT of a family-based HIV prevention program, IMAGE, to reduce the incidence of HIV/STI in Chicago. The intervention draws on the Becoming a Sexual Black Woman and HDRF frameworks, which recognize multilevel influences of individual, interpersonal, community, and societal structural factors on Black girls HIV/STI risk and prevention. Prior to funding, IMAGE was extensively vetted with CBOs during theater and pilot testing and community advisory boards for approval. This process was essential to ensure study uptake and acceptability broadly and improve chances for sustainability. We believe the study design has several strengths. We leverage the girl-male caregiver relationship to support HIV/STI prevention and protection of girls’ SRH health (bodies, behaviors, rejection of stereotypes). Furthermore, we believe that IMAGE is responsive to the age, gender, and cultural needs of Black girls and their male caregivers. This study is among the first to engage male caregivers in family-based SRH programming. By engaging male caregivers, this study leverages the unique strengths that male caregivers bring to the protection of girls. IMAGE targets two high-risk and vulnerable populations: Black girls and male caregivers. Black girls are particularly vulnerable to early sexual engagement, sexual violence, and HIV/STI infections, and Black males are disproportionately burdened by structural racism. The effects of individual-level SRH programs for Black girls decay over time. IMAGE may help sustain positive outcomes for girls as BMCs continue to deliver and reinforce prevention messages tailored to girls’ developmental phase after the formal intervention ends. IMAGE was designed to meet the standard for evidence-based interventions, advance intervention science with Black girls, and prepare for implementation and sustainability by engaging community stakeholders prior to the efficacy trial.
S1 File Notice of Award from NIH. (PDF) S2 File IRB approval. (PDF) S3 File Study protocol. (DOCX) S4 File SPIRIT checklist. (PDF)
|
Achieving the optimal emergence profile: the role of soft tissue grafting and pontic site development | 66e4452b-baa9-43c7-a749-292ccb9dc884 | 11645261 | Dentistry[mh] | Dimensional changes occur after tooth extraction, with one-third of the ridge width lost within the first three months and 50% at 12 months. This volumetric loss results in a concavity at the facial surface of the alveolar ridge and may detract from the provision of a natural-looking prosthesis because the emergence profile often appears to emanate from the surface of the ridge, rather than from within the ridge. When the smile line exposes the junction between the tooth and gingiva, the volumetric deficiency above the pontic creates a shadow, as this region is not illuminated in the same way as the root bulbosity of natural teeth . These factors can significantly detract from the pink aesthetic score, as reported by Furhauser and modified by Belser. , provides a summary of the parameters assessed. To optimise aesthetics, the lost volume must be replaced via surgical intervention or prosthetic replacement. Pontic site augmentation (PSA) refers to the surgical augmentation of an edentulous site. This is followed by pre-prosthetic development to optimise the topography for a convex pontic fit surface. This article describes these steps, with the aim of emulating the emergence profile of a natural tooth. Combined hard and soft tissue grafting, techniques involving socket shields, or alveolar ridge preservation are beyond the remit of this article.
When considering soft tissue augmentation, available graft materials include autogenous connective tissue, allografts, or biomaterial substitutes. provides a brief summary of the different types of soft tissue grafts which may be considered for this procedure. Autogenous connective tissue grafts can either be harvested as a subepithelial connective tissue graft (SCTG) or as a free gingival graft (FGG) which is subsequently de-epithelialised (DFGG). De-epithelialisation may be carried out either before or after harvest, depending on the operator's preferred technique. The use of a SCTG was first reported by Langer and Calagna in 1982 to develop pontic sites. However, to the authors' knowledge, there is still no available literature comparing SCTGs to DFGGs for pontic site augmentation, with the majority of available evidence referencing the SCTG. Autogenous connective tissues can be harvested from two sites, depending on the quality and quantity of tissue required. Generally, the middle third of the hard palate is used, with the possibility to extend anteriorly depending on the size of the graft required. Alternatively, grafts can be harvested from the tuberosity, which can deliver significant quantities of dense connective tissue, depending on the individual anatomy of the patient. The phenotype of the graft harvested is important to consider. Grafts taken from the superficial layers, such as the DFGG or the tuberosity (which is predominantly composed of dense connective tissue), may increase in volume over time and can create aesthetic complications if the underlying regenerated epithelium is not fully de-epithelialised . Patients with a thin phenotype may have limited submucosal tissue and harvesting a SCTG may not be feasible, favouring the DFGG technique. Tissue sounding may provide some insight peri-operatively as to the viability of harvesting an SCTG. Alternative substitute biomaterials may remove the need for an autogenous donor site. These materials are derived from either animal origin (xenografts) or are human-derived allografts. They are comprised of either native collagen matrices or crosslinked collagen matrices; this decision is based on the desired substitution or resorption traits, with the fundamental requirement being that the desired volume augmentation is maintained. Research in an animal model identified that the use of a crosslinked collagen matrix resulted in a greater and more stable ridge width over time compared with control groups using native collagen matrices. These biomaterials have a range of physical characteristics, from a sponge-like consistency (Geistlich Fibro-Gide) to a form more similar to a DFGG, such as BioHorizons' NovoMatrix (acellular dermal matrix). There is a limited body of evidence available reporting on the long-term outcomes of these collagen matrices when compared to that of autogenous connective tissue grafts, which remain the gold standard. When the defect is significant, soft tissue grafting alone may not adequately augment the site to full contour and consideration may be given to combined hard and soft tissue grafting. The reader is directed towards bone augmentation texts for further information on these procedures.
The recipient graft site (pontic site) is accessed using an envelope-style incision or a tunnel preparation. Both deliver a split thickness flap or pouch to accept the graft. This leaves the periosteum intact and provides a vascularised bed for the recipient graft, while the superficial connective tissue and epithelium is freed from the underlying attached mucosa as a split thickness layer. This allows for tension-free closure over the inter-positional graft. Initial graft survival depends on plasmatic circulation for nourishment while angiogenesis takes place. The envelope flap improves surgical access to appropriately position the graft, whilst a pouch created via tunnelling reduces the size of the surgical wound at the ridge crest. This however increases the complexity of the procedure and reduces surgical access. shows a DFGG harvested from the palate and folded to double the thickness, providing a compression resistant matrix for augmentation of the buccal aspect. The connective tissue graft or substitute collagen matrix can be secured onto the prepared bed via one of two techniques. The first secures the graft material onto the recipient bed itself with a mattress sling suture, while the second secures the graft to the buccal flap, which is further stabilised upon closure. The choice depends on the surgical access to the recipient bed, the remaining thickness of periosteum to suture to and operator preference. The key parameter is to ensure the graft material is placed in the correct three-dimensional position to augment the buccal aspect. shows the introduction of graft into the pontic site, correctly positioned, with a securing suture through the buccal flap. shows the immediate post-operative image using 6-0 proline sutures to close the incision. Care is taken to ensure this is tension-free. shows healing at two weeks, before suture removal. demonstrates the bucco-lingual volumetric increase at one month.
If an autogenous connective tissue graft is chosen, a second surgical site is required. These donor sites are subject to the same risks as the recipient graft site. If a DFGG is harvested, the level at which the graft is separated from the base can influence the degree of haemorrhage. If the graft is harvested just superficial to the glandular mucosa, often, intra-operative bleeding is significantly reduced, as this region is comprised of dense connective tissue. The anatomy of the palate is described elsewhere, but in the main, the risk of damage to larger vessels in the area (such as the greater palatine artery) is significantly reduced if the greatest extension towards the palatal midline is 10 mm from the palatal aspect of the cemento-enamel junction, with the graft harvested anterior to the mesial aspect of the first maxillary molar palatal root. After DFGG harvest, the authors' preference is use of an Essix-style healing plate, which covers the palate without further surgical intervention. Other techniques involve the use of biomaterial matrices to cover the de-epithelialised site, stabilised via sutures. This management may however, cause additional bleeding and the sutures require removal at a later date. illustrates early healing with a native collagen matrix becoming integrated into the donor site. This adds time and cost to the procedure. If a SCTG is harvested, the access incision is closed with sutures. Care must be taken to ensure that the superficial layer is not left too thin, as necrosis can occur . Patient-reported outcomes comparing visual analogue pain scores between patient groups receiving SCTGs and collagen matrices identified a greater need for post-operative analgesia when autogenous grafts were used compared to biomaterials substitutes. No statistically significant differences were demonstrated between patients undergoing a SCTG compared to a DFGG, but pain increased when there was necrosis of the overlying mucosa. In the post-surgical phase, the Essix-style healing plate (which extends over the palate and onto the incisal edges for retention) is worn continuously for the first 24 hours, then during the day for the first week to prevent trauma when eating and speaking. The appliance is to be removed for cleaning after meals and disinfected twice a day using a chlorhexidine product. For most patients, missing an anterior tooth will be socially unacceptable, so the healing plate may be constructed with a replacement tooth, or as a Hawley-style retainer. Suture removal takes place at two weeks and the patient is reviewed at three months to begin pontic site development.
Pontic site development (PSD) transforms the soft tissue topography from a convex ridge to a concavity with a scalloped margin, ready to accept an ovate pontic. This design allows for cleansability of the underside of the prosthesis, while delivering an emergence profile projecting from within the soft tissue to optimise the aesthetic outcome. When planning the restorative phase of treatment, consideration should be given to the timing of prosthesis placement following soft tissue augmentation, as soft tissue stability is desired. An investigation examining single-site pontic development using connective tissue grafts identified significant volumetric changes up to three months from baseline, but non-significant changes in volume between three and six months. Longer-term studies have shown that soft tissue volumes at pontic sites, with and without a SCTG, were stable at five and ten years with no statistical difference between the groups. , When comparing porcine collagen matrices to a STCG, an animal model identified significant volumetric changes within the first month but no statistical significance in terms of soft tissue volumes between the two groups at ten months. Histological examination also reported comparable quality of the augmented connective tissues between the two groups at ten months. Although limited, the available evidence suggests that the final reconstruction may be considered three months post-augmentation at the earliest, with soft tissue stability expected over the longer term. Digital technologies may also assist clinicians in choosing the appropriate time for reconstruction via superimposed images captured from intra-oral scanners (IOS). Volumetric changes can be mapped, and when stable, reconstruction considered. shows digitised images of the clinical situation pre- and post-augmentation as steriolithic tessellated language (STL) files. These technologies allow clinicians to deliver personalised, patient-specific treatment strategies. Further information regarding the use of IOS devices can be found elsewhere.
Pontic site development may be carried out before impressions for the definitive prosthesis, or at the time of fitting the definitive restoration. Pontic site development before restoration can be conducted either with a removable provisional prosthesis or a fixed provisional prosthesis. Essix-style or Hawley-style retainers are commonly used, with materials such as composite resin (flowable or conventional), acrylic resins, or bis-acrylic composites chosen to augment the fitting surface of the pontic. When a removable prosthesis is used, immediate delivery of an ideally shaped ovate pontic may prevent full seating; therefore, smaller staged additions may be required to allow tissue compression without significant prosthesis displacement. Prosthesis use in function will steadily compress the tissue and develop the pontic site. This may take several weeks to complete and may be inconvenient for the patient but generally prevents the need for local anaesthetic and a surgical procedure. The patient's phenotype may dictate the level of soft tissue compression available, with thin phenotypes limiting tissue compression, as shown in , when compared to thicker phenotypes, as shown in . The surgical approach to immediate pontic site development is carried out under local anaesthetic. Techniques include the use of a tissue removal bur, round diamond bur , electrosurgery, or a tissue punch to remove the desired volume of soft tissue. The provisional prosthesis with the desired fit surface shape is delivered and will maintain the newly created concavity. The site is subsequently left to heal over a period of 2-4 weeks, with any subsequent impressions reproducing the customised soft tissue profile. The laboratory-based strategy develops the pontic site on the cast when the definitive prosthesis is made, with soft tissue modification at the time of fit. This pathway benefits the patient in terms of immediacy of reconstruction. This strategy carries two challenges. Firstly, with the laboratory estimating the degree of soft tissue removal required and secondly, with respect to haemostatic control at cementation, which is imperative to prevent contamination at the time of delivery. illustrates the final bridge at the fit appointment, immediately after surgical pontic site development to match the laboratory-constructed fit surface.
PSA and PSD are often required to modify the topography of an edentulous ridge to accept an ovate pontic with an optimal emergence profile. This gives the illusion of a prosthesis emanating from the mucosa, mimicking a natural tooth, while delivering optimal anatomy for patient-conducted oral hygiene. Autogenous connective tissue or biomaterial substitutes may be used for soft tissue grafting. Autogenous grafting remains gold standard; however, it involves a second surgical site that requires post-operative management and increases surgical morbidity. Pontic site development may be conducted as early as three months post-augmentation, with some evidence to support stability of the soft tissues in the long-term. Following augmentation, prosthetic or surgical interventions may be used to achieve adequate PSD, with the chosen technique related to operator and patient preference and clinical situation. PSA and PSD play a key role in delivering natural aesthetics and meeting patient expectations. However, limitations exist where defects cannot be managed by soft tissue grafting alone and hard tissue grafting may also be required. Further research is still needed to compare SCTG to DFGG for pontic site development and to assess the long-term outcomes of collagen matrices compared to autogenous connective tissue grafts.
|
Barriers and Facilitators to Promoting Oral Health Literacy and Patient Communication among Dental Providers in California | afe91dec-8f0d-4888-9087-28aec3ef243a | 7795206 | Health Communication[mh] | In the United States (US) and globally, poor oral health is a major, but preventable public health problem . Disparities and inequities in oral diseases are also a serious social justice concern, as low-income and marginalized groups are disproportionately impacted by poor oral health . In addition to a lack of access to care, oral health literacy (OHL) is a major factor in these disparities. Research over the past two decades has identified low OHL as a key contributor to poor oral health . The US Department of Health and Human Services in its report Healthy People 2010, defined OHL as “the degree to which individuals have the capacity to obtain, process and understand basic oral health information and services needed to make appropriate health decisions” . The concept of OHL has recently been expanded to include the abilities of oral health providers to communicate effectively with patients/caregivers and create “patient-friendly” environments . Improving OHL is a critical challenge for advancing global oral health. The US Healthy People 2030 provides the following new, combined definition of health literacy: Personal health literacy is the degree to which individuals have the ability to find, understand, and use information and services to inform health-related decisions and actions for themselves and others. Organizational health literacy is the degree to which organizations equitably enable individuals to find, understand, and use information and services to inform health-related decisions and actions for themselves and others. Studies of general health literacy (mostly from medical settings) have identified three main health literacy intervention areas and documented overall positive results from: (1) designing easy-to-use communication resources matching people’s literacy and linguistic skills ; (2) training health providers to use effective communication techniques ; and (3) developing “patient-friendly” and “shame-free” healthcare environments . Building upon this foundational health literacy work in medicine, training dental team providers on oral health literate communication techniques is now considered a priority area for research and for development of effective strategies. Research has examined OHL and communication practices from the perspectives of healthcare providers, including dentists, dental hygienists, nurse practitioners, primary care providers, and pediatricians. Rozier et al. conducted a national survey of 1994 US dentists assessing use of 18 patient communication techniques recommended by the American Medical Association (AMA) . The study found that dentists reported using a mean of 7 out of 18 recommended communication techniques, and only 3 out of 7 communication techniques that were considered “basic techniques.” The main techniques reported by two-thirds of respondents were to: speak slowly, use plain language, demonstrate with models/radiographs, and provide caregivers printed handouts to take home. Less than 25% of dentists reported using “teach-back.” This technique is one in which a provider asks a patient to describe—in their own words—a treatment they receive during a clinical visit or what they will do for their treatment or preventive care at home to follow the provider’s guidance . Dentists who were older, female, Black/African American, educated outside of the US, and specialists (with the exception of pediatric dentists) were likely to report using more communication techniques. Overall, 73.3% of participating dentists had never taken a health communications course, and 68.5% indicated they would be interested in continuing education (CE) to improve their communication with patients . Besides the single nationwide study conducted by Rozier et al., there have been some state-level studies. Horowitz et al., Koo et al., Weatherspoon et al., and Maybury et al. surveyed oral health providers in Maryland to assess their use of recommended communication techniques . Dentists (general and pediatric), dental hygienists, nurse practitioners, and physicians (family medicine and pediatrics) all reported using an average of 47% or less of the recommended communication techniques. Maybury et al. surveyed general practice and pediatric dentists in private practice to assess the use of 18 AMA-recommended communication techniques, including 7 basic communication techniques. The seven basic communication techniques included five that cover interpersonal communication and two that highlight teach-back. The interpersonal communication items included: Cover two–three concepts at a time; Ask the patient if they would like family/friends in discussions; Draw pictures; Speak slowly; and Use simple language. The teach-back method items included: Ask the patient to repeat back; and Ask the patient to tell you what they will do at home. In Marbury and colleagues’ study, of 1393 general dentists and 169 pediatric dentists who were sent the survey, 38.4% provided responses. About 60% of responding providers reported having taken a communication course. General dentists reported using 7.9 out of the 18 communication techniques and 3.6 out of the 7 basic techniques; pediatric dentists used more techniques (8.4/18 total and 3.8/7 basic techniques). General dentists who reported having taken a communication course had higher use of the 18 communication techniques but not of the 7 basic techniques, while pediatric dentists who reported having taken a communication course had higher reported use both of the 18 communication techniques and the 7 basic techniques. Less than 20% of participants reported using teach-back . Horowitz et al. surveyed a sample of Maryland dental hygienists to assess the use of the same communication techniques. Of the 1259 surveys sent, they received 540 valid responses (42.9% response rate). Dental hygienists reported using an average of 7 out of the 18 communication techniques routinely, and 3.7 out of the 7 basic techniques. Less than 5% reported using all 18 communication techniques. Only one out of three routinely reported using teach-back . Some US states have established oral health initiatives that recommend dental provider OHL training . California is the state with the largest number of dental providers in the US: 30,772 dentists , 22,800 dental hygienists , and 56,840 dental assistants . California is the largest and most diverse state in the US—with 40 million residents (1/8 of the US population). California has a majority non-White population; 27% are immigrants, primarily from Latin America and Asia . This creates a unique context for oral health care and poses OHL challenges for providers and patients. In collaboration with the California Department of Health Care Services, the California Department of Public Health (CDPH) Office of Oral Health (OOH) developed the California Oral Health Plan for 2018–2028, addressing California’s current major issues in oral health and providing guidance to improve oral health and achieve oral health equity . One key strategy in this plan is to “provide training and resources to improve dental teams’ communication with patients,” emphasizing steps to assess and build on current resources and skills for providing OHL-based care. Our research team carried out a study as a part of efforts by the CDPH with the following objectives: (1) to add to the limited literature about dental provider use of recommended OHL communication techniques; (2) to conduct a qualitative California study of dental team providers’ OHL knowledge, perspectives and practices, interests in training, and preferences for training delivery; and (3) to apply findings to inform the development of statewide OHL trainings and resources for dental providers. This study is the first to examine Californian dental providers’ and national OHL experts’ perspectives on provider-patient OHL communication. This study was categorized as exempt from full Institutional Review Board (IRB) review by the IRB at California’s Office of Statewide Health Planning and Development (OSHPD).
We conducted a qualitative study, including 50 key informant (KI) interviews and using pre-surveys to collect demographics and use of communication techniques with a sample of dental providers and OHL experts from November 2019 to March 2020. Our KIs included dentists (general dentists and pediatric dentists), dental hygienists, and dental assistants across California ( n = 44), and national OHL subject matter experts ( n = 6). Participants were recruited using both purposive and snowball sampling strategies in collaboration with the CDPH OOH, California Dental Association, Dental Board of California, California Dental Hygienists Association, California Dental Assistants Association, as well as other key state and national organizations. Dental providers (general dentists, pediatric dentists, dental hygienists, and dental assistants) were eligible if they currently provided dental care at a federally qualified health center (FQHC)/community clinic or private practice in California. Dental providers were purposively sampled from geographic locations across California including urban, peri-urban, and rural areas. Experts were identified for the study based on their expertise in identifying and shaping state- and national-level OHL-related research, interventions, and/or policies. Potential participants were approached via phone, email, or email introductions from our state and national partners. Dental provider recruitment ceased when data saturation was observed in the concurrent data analyses. We designed two sets of study instruments: one for dental providers and one for OHL experts. Each set included telephone interview guides and web-based pre-surveys for demographics and communication techniques used. Instruments were pilot-tested with providers and OHL experts. The interview guides and descriptive pre-surveys were developed based on a scoping review of relevant literature and an environmental scan of existing oral health educational materials and programs, as well as an iterative content review. We also examined five oral health assessment and educational strategies in the telephone interview that we identified from our scoping review and environmental scan. Survey questions about communication techniques were adapted from prior OHL surveys assessing provider use of effective communication techniques, primarily the AMA’s 18 communication techniques and 7 basic communication techniques . In total, 40 out of 44 providers completed the pre-survey independently before their telephone interview. All 44 providers completed the telephone interview. Electronic informed consent was given to the interviewees via email prior to proceeding with the web-based pre-survey and scheduling an interview. The consent information was also repeated verbally, asking them again if they agreed to be interviewed, just before the beginning of the telephone interview. Questions in instruments required yes/no, multiple-choice, or open-ended responses, depending on which portion of data collection is described. The provider pre-survey contained 21 questions and included items related to professional background, work environment, role in patient education, use of types of patient communication techniques including AMA’s 18 communication techniques and 7 basic communication techniques, prior patient communication training, future training preferences, and demographics. The OHL expert pre-survey contained nine questions and included items related to professional background, health literacy and/or OHL experience, and demographics. Following completion of their web-based pre-survey, providers and OHL experts participated in a 30- to 60-min semi-structured open-ended telephone interview conducted by a member of the research team. The telephone interview allowed participants to elaborate on their opinions and experiences with patient communication and education strategies in general and with disadvantaged patient groups. Provider interviews included 11 questions and assessed items such as experiences with patient education and communication, knowledge of assessment and educational strategies and oral health literacy, gaps in professional training and education, and recommendations on improving patient-provider communication. Communication and educational methods assessed by telephone interviews included the “Tell-Show-Do” approach described and demonstrated as effective by Avenetti et al. , the caries risk assessment (CAMBRA), anticipatory guidance, family oral health education (FOHE), motivational interviewing (MI), and teach-back described and promoted by Ramos-Gomez and Ng . Anticipatory guidance, CAMBRA (an assessment tool), and FOHE were included in this exploration to understand how explaining caries risk and anticipatory guidance can be used as educational approaches to improve provider-patient communication especially for parents and caregivers of pediatric populations. Expert interviews included 15 questions and assessed items related to OHL best practices among dental providers, factors that promote or hinder improvements in OHL, gaps in training and professional education, development of an OHL toolkit and materials to use in dental practice, meeting the needs of diverse populations, and innovations and promising practices. Most of the interviews were recorded. Interview notes were taken during all interviews. Interview transcripts and notes were compiled and organized by KI type for qualitative analysis. Descriptive statistics were used to characterize the survey participants using the pre-survey items. The constant comparative method was used as a technique for the qualitative thematic analysis. This method develops codes, examines relationships and interactions across descriptive and thematic codes, and compares the major themes that emerged from the coding categories . Qualitative analysis probed for parallel themes, particularly looking for provider knowledge and use of oral health assessment and educational strategies and communication barriers and facilitators. The final codebook consisted of descriptive and thematic codes common across the 50 KI interviews. Four researchers independently coded transcripts from the KI interviews. Inter-rater agreement for the first two provider interview transcripts and the first two expert interview transcripts was determined to ensure consistency in coding. If 80% agreement in coding consistency was not reached, the researchers discussed potential issues that arose and reached consensus about these coding issues until consistency was reached. Next, these initial transcripts were recoded and the coding consistency percentage was recalculated. The remaining transcripts were then coded using the updated codebook. The codebook was revised as subsequent transcripts were coded and new codes emerged. Coding consistency was recalibrated as part of this iterative coding process. Code categories were connected and grouped through thematic coding, and the researchers identified major themes from the codes. In addition, pre-survey data collected using Qualtrics were imported into Stata (Version 15, StataCorp, College Station, Texas, USA) for descriptive analysis of background information and use of communication techniques on providers.
3.1. Summary of KI Characteristics In total, 44 qualitative dental provider key informant interviews and 40 associated pre-surveys were completed. The characteristics of 40 out of the 44 California dental providers (RR = 91%) from the pre-surveys are presented in . Of the 40 dental providers, 48% of the providers were dentists, with almost equal representation from private practice (25%) and public practice (23%); 30% were dental hygienists; and 23% were dental assistants, with 65% of dental hygienists and dental assistants from private practice settings. Most providers were female (70%). Most identified their race/ethnicity as White (60%) followed by Asian/Pacific Islander (25%), Hispanic/Latino (10%), Black/African American (3%), or “Other.” Providers’ average length of practice was 22 years, with a range from 1.5 to 49 years. All OHL experts ( n = 6) had backgrounds in both public health and OHL, and two experts were dentists. Most experts were female (67%). All experts were White, and the majority were affiliated with an academic institution (83%), while one worked with a dental association. Experts had between 14 and 48 years working in their field, with an average of 40 years of experience; the majority provided direct service to the community (83%). 3.2. Provider Knowledge and Use of OHL Educational and Communication Techniques In the open-ended qualitative telephone interviews ( n = 44), providers discussed valuing effective communication and feeling “extremely” or “very” confident in communicating with patients. One dental hygienist illustrated this, saying: “In hygiene school, it’s one of the many things we do, oral health education (…) I increase my own oral health literacy as new (OHL) studies come out on my own. As a public health professional, oral health is important to me. Even looking at (…) how it affects populations, to me it’s health equity. It’s imperative for all of us to understand this.” (Dental Hygienist, Public Practice) Most dental providers interviewed reported that they were familiar with the term “oral health literacy” (OHL) and believed that it was vital to patient communication. While most providers understood that OHL addresses a patient’s ability to understand and act on oral health information, few providers knew that OHL also involves the provider’s ability to communicate with a patient, and the patient-friendliness of the dental environment. The OHL experts affirmed this disparity, commenting that dental providers generally considered themselves good communicators, but they may not tailor their communication style for patient needs or ensure patient understanding and ability to practice positive oral health behaviors at home. This was echoed in the web-based survey ( n = 40), where 86% of dental providers who answered the survey items rated their training and competency in patient education and communication as “good” or “very good.” Two-thirds reported having taken a communication course, and 70% expressed interest in receiving continuing education on communication. Half of the providers reported having assessed the “patient-friendliness” of their practice but provided no information about their process to do so. In terms of educational and communication strategies, most dental providers from the telephone interviews, particularly dental hygienists and dental assistants, considered visual communication and the “Tell-Show-Do” approach to be highly effective for patient/caregiver education. Many indicated that they used a variety of visual communication approaches, including dental models, intraoral cameras, mirrors, pictures, charts, brochures with visuals, and videos. Most dentists reported that dental hygienists and dental assistants provided the majority of patient oral health education. The public practice providers generally described patient education delivered by one designated provider, while the private practices generally described reinforcement of patient education messages by all team members. The telephone interviews probed dental providers’ awareness and use of five types of oral health assessment and educational strategies: caries risk assessment (CAMBRA) , anticipatory guidance, family oral health education (FOHE), motivational interviewing (MI), and teach-back. Many providers lacked understanding of these five types of oral health assessment and educational strategies, particularly private practice dentists, dental hygienists, and dental assistants . Most dental providers interviewed were unaware of or did not use family oral health education, motivational interviewing (MI), or teach-back. Those who knew about, but did not use these approaches, cited them as not being a required standard of practice and/or having limited time for patient education/communication. One dentist described the challenges in using MI in their public practice: “You talk to the patient and educate them and trying to understand why they need to do certain things (…) Try to use the outcomes to motivate them, to change behavior. (…) Motivational interview takes a lot of time. Not everyone can do it. We have a lot of patients. Need to carve out 5–10 min (for patient communication) is all we can do.” (Dentist, Public Practice) We asked all providers about their use of CAMBRA and FOHE despite being primarily intended for use in pediatric populations. Many providers did not know about or use FOHE, while more providers were familiar with and used CAMBRA; providers serving children more consistently reported using these techniques in their practice. Web-based survey responses from 40 of 44 dental providers addressed their use of and perceived effectiveness of the AMA’s 18 patient communication techniques . A majority of providers indicated they used 6 out of 18 AMA-recommended communication techniques always or most of the time, including simple language (90%), visual demonstration with models or radiographs (80%), presenting two–three concepts at a time (80%), speaking slowly (69%), handing out printed materials (54%), and asking patients what they can accomplish with oral health hygiene (50%). However, most also reported not using some of these communication techniques that they believed were effective, such as drawing pictures or using printed illustrations, using teach-back, using a translator/interpreter, asking other office staff to follow up with post-care instructions, and asking patients if they would like their support person to accompany them to their appointment. Looking more closely at the dental provider survey responses pertaining to just their use of and perceived effectiveness of the AMA’s seven basic patient communication techniques , five of these communication techniques presented are specific to interpersonal communication and the remaining two are teach-back techniques. A majority of providers indicated that they use three of the five interpersonal communication techniques all or most of the time, including using simple language (90%), presenting two–three concepts at a time (80%), and speaking slowly (69%). However, fewer reported asking patients if they would like to be accompanied by a support person for discussion (28%) and drawing pictures or using printed illustration (33%) all or most of the time. A minority of providers reported using either of the teach-back techniques, with only 33% reporting use of teach-back all or most of the time and only 46% reporting asking patients what they will do at home to follow guidance. Additionally, most providers reported not using some of the communication techniques, even though they believed the techniques were effective—such as drawing pictures or using printed illustrations, using teach-back, and asking patients if they would like their support person to accompany them to their appointment. 3.3. Provider/Patient Communication Barriers In the telephone interviews, many providers reported that a key barrier was that they had not received in-depth provider-patient communication and OHL training in professional school or continuing education programs. Most OHL experts noted that clinical dental education was more focused on the “hard skills” of technical procedures rather than the “soft skills” of providing effective patient-centered communication and education. From the survey, all of the dentists, most of the dental hygienists (70%), and more than half of the dental assistants (56%) reported that they had experienced problems communicating with patients. Dental providers and OHL experts identified major barriers to effective provider-patient communication from the pre-survey and the telephone interviews, summarized as: (1) patient/caregiver-side communication barriers and (2) provider-side communication barriers . 3.3.1. Patient/Caregiver-Side Communication Barriers The major patient/caregiver barrier to provider-patient communication that dental providers identified from the survey was that the patient/caregiver does not follow the provider’s instructions (56%), “no matter how well they were explained.” One-third to one-half of dental providers interviewed noted additional barriers underlying the patients’ inability to follow recommended oral health practices, including patients’ lack of understanding of oral health information, language and cultural barriers between providers and patients, and patients’ lack of interest in or prioritization of their oral health. Dental providers referenced specific disadvantaged patient populations with whom they experienced more frequent challenges in provider-patient communication . These populations included patients with limited English proficiency (65%), people with cognitive disabilities (54%), elderly (42%), people with limited education (35%), Deaf or hard of hearing (31%), and early childhood age groups (31%). Experts also noted that patients with limited English proficiency experienced the greatest oral health communication barriers due to cultural differences that often co-occur with language barriers and indicated that most dental practices do not adequately use translation/interpretation services. 3.3.2. Provider-Side Communication Barriers From the telephone interviews, many dental providers and OHL experts cited provider-side barriers to patient communication and education, particularly pertaining to limitations in provider training and clinical practice. Dental providers and OHL experts noted the lack of OHL communication training and proficiency requirements in professional schools and continuing education. They identified clinical practice constraints such as the inadequate time providers have for patient education during the patient encounter, limited reimbursement for delivering patient education, the paucity of high-quality patient education materials in needed languages and accessible formats, and logistical and financial difficulties to access translation/interpretation services. Public practice providers were more likely to underscore the need for technological educational resources such as video/DVD players and tablets, high-quality educational materials in a variety of languages, and translators/interpreters knowledgeable about oral health. One public practice provider discussed these structural challenges, saying: “(The) most obvious barrier is language (…) Time is also the issue. Even if you can delegate, it’s a system issue, if you delegate to dental staff to do the education, it’s taking up chair time/office space (…) Also, many providers have limited training on health communication techniques.” (Dentist, Public Practice) Private practice providers were more likely to cite the lack of time and resources for patient education and reimbursement structures that prioritize dental procedures and treatment over patient education; this was described by one private practice dental assistant: “I would like to have more time and more patient cooperation to learn. (We also) need more (patient education) resources that explain oral health and dental procedures in simple terms and that have a positive tone.” (Dental Assistant, Private Practice) Some of the final provider interviews overlapped with the beginning of the COVID-19 pandemic in California. Although questions about the pandemic’s impact were not included in the study instruments, we note that providers interviewed during February and March 2020 commented that practice disruptions greatly exacerbated communication problems with patients and that guidance was urgently needed to address these new patient communication barriers, both in the changed practice setting and remotely, including tele-dentistry strategies. 3.4. Communication Facilitators for Dental Providers Dental providers interviewed suggested key interventions that could improve provider-patient communication by addressing the barriers they identified. They highlighted the need for more training on OHL communication techniques in professional schools and continuing education. They cited the need for patient-centered clinical practice supports, including more provider time and reimbursement incentives for delivering patient education, translator/interpreter services, high-quality low-cost educational materials in accessible plain-language formats (including videos and print materials) and in a variety of languages. In addition, they wanted information about ways to ensure a patient-friendly office. Dental providers highlighted the foremost importance of the concept of cultural humility: developing trusting, high quality, and professional relationships with patients, understanding the patient’s emotional motivators, and empowering patients to engage in preventive health behaviors. Many providers expressed their hopes that better OHL training, high-quality educational materials, and clinical supports could improve their patient communication and facilitate patients’ trust in providers, and their understanding and motivation to achieve recommended prevention and treatment goals, as described by one dentist: “My dream would be to have California take the lead on a more innovative Medi-Cal (Medicaid—a healthcare program for people with low-incomes) Dental, create more (financial) incentives for positive oral health outcomes. Right now, it is geared to treating disease and procedures. Need to flip to outcomes based—use (performance-based) incentives for patients and providers to become healthier—achieve scalable results.” (Dentist, Public Practice) OHL experts recommended further structural changes in dental professional training and clinical practice and beyond. They included incorporating OHL curricula in dental professional schools and continuing education through dental professional societies, adding OHL competency requirements for clinical license examinations and/or renewal, and expanding OHL training to medical providers and health coaches. The range of suggestions for training was encapsulated by one OHL expert: “We need to make top-down changes, starting the changes with accrediting bodies for medical/dental education, (…) board exams, (…) license recertification, and clinical quality measures and financial incentives to include good health literacy and oral health literacy practices. There are some individual medical/dental schools and scattered continuing education courses that may include health literacy/oral health literacy, but that only reaches a small subset of medical/dental providers, particularly those who are already aware of the issues. To draw in new participants, it’s helpful to tag on oral health literacy to other conferences (e.g., general pediatrics) and present compelling cases to show how oral health is critical to overall health.” (OHL expert) Experts also discussed the need for training the entire dental team in effective OHL strategies, including establishing a patient-friendly environment, using OHL communication techniques, and using oral health-specific translator/interpreter services such as via phone/video services and natural language processing software. Dental providers in public practice and OHL experts also recommended expanded community oral health education, through integrative dental and medical care models particularly during early childhood (e.g., in pediatrics), community events and schools from early childhood through high school, to raise the public’s knowledge of oral health and help patients feel more comfortable with dental care.
In total, 44 qualitative dental provider key informant interviews and 40 associated pre-surveys were completed. The characteristics of 40 out of the 44 California dental providers (RR = 91%) from the pre-surveys are presented in . Of the 40 dental providers, 48% of the providers were dentists, with almost equal representation from private practice (25%) and public practice (23%); 30% were dental hygienists; and 23% were dental assistants, with 65% of dental hygienists and dental assistants from private practice settings. Most providers were female (70%). Most identified their race/ethnicity as White (60%) followed by Asian/Pacific Islander (25%), Hispanic/Latino (10%), Black/African American (3%), or “Other.” Providers’ average length of practice was 22 years, with a range from 1.5 to 49 years. All OHL experts ( n = 6) had backgrounds in both public health and OHL, and two experts were dentists. Most experts were female (67%). All experts were White, and the majority were affiliated with an academic institution (83%), while one worked with a dental association. Experts had between 14 and 48 years working in their field, with an average of 40 years of experience; the majority provided direct service to the community (83%).
In the open-ended qualitative telephone interviews ( n = 44), providers discussed valuing effective communication and feeling “extremely” or “very” confident in communicating with patients. One dental hygienist illustrated this, saying: “In hygiene school, it’s one of the many things we do, oral health education (…) I increase my own oral health literacy as new (OHL) studies come out on my own. As a public health professional, oral health is important to me. Even looking at (…) how it affects populations, to me it’s health equity. It’s imperative for all of us to understand this.” (Dental Hygienist, Public Practice) Most dental providers interviewed reported that they were familiar with the term “oral health literacy” (OHL) and believed that it was vital to patient communication. While most providers understood that OHL addresses a patient’s ability to understand and act on oral health information, few providers knew that OHL also involves the provider’s ability to communicate with a patient, and the patient-friendliness of the dental environment. The OHL experts affirmed this disparity, commenting that dental providers generally considered themselves good communicators, but they may not tailor their communication style for patient needs or ensure patient understanding and ability to practice positive oral health behaviors at home. This was echoed in the web-based survey ( n = 40), where 86% of dental providers who answered the survey items rated their training and competency in patient education and communication as “good” or “very good.” Two-thirds reported having taken a communication course, and 70% expressed interest in receiving continuing education on communication. Half of the providers reported having assessed the “patient-friendliness” of their practice but provided no information about their process to do so. In terms of educational and communication strategies, most dental providers from the telephone interviews, particularly dental hygienists and dental assistants, considered visual communication and the “Tell-Show-Do” approach to be highly effective for patient/caregiver education. Many indicated that they used a variety of visual communication approaches, including dental models, intraoral cameras, mirrors, pictures, charts, brochures with visuals, and videos. Most dentists reported that dental hygienists and dental assistants provided the majority of patient oral health education. The public practice providers generally described patient education delivered by one designated provider, while the private practices generally described reinforcement of patient education messages by all team members. The telephone interviews probed dental providers’ awareness and use of five types of oral health assessment and educational strategies: caries risk assessment (CAMBRA) , anticipatory guidance, family oral health education (FOHE), motivational interviewing (MI), and teach-back. Many providers lacked understanding of these five types of oral health assessment and educational strategies, particularly private practice dentists, dental hygienists, and dental assistants . Most dental providers interviewed were unaware of or did not use family oral health education, motivational interviewing (MI), or teach-back. Those who knew about, but did not use these approaches, cited them as not being a required standard of practice and/or having limited time for patient education/communication. One dentist described the challenges in using MI in their public practice: “You talk to the patient and educate them and trying to understand why they need to do certain things (…) Try to use the outcomes to motivate them, to change behavior. (…) Motivational interview takes a lot of time. Not everyone can do it. We have a lot of patients. Need to carve out 5–10 min (for patient communication) is all we can do.” (Dentist, Public Practice) We asked all providers about their use of CAMBRA and FOHE despite being primarily intended for use in pediatric populations. Many providers did not know about or use FOHE, while more providers were familiar with and used CAMBRA; providers serving children more consistently reported using these techniques in their practice. Web-based survey responses from 40 of 44 dental providers addressed their use of and perceived effectiveness of the AMA’s 18 patient communication techniques . A majority of providers indicated they used 6 out of 18 AMA-recommended communication techniques always or most of the time, including simple language (90%), visual demonstration with models or radiographs (80%), presenting two–three concepts at a time (80%), speaking slowly (69%), handing out printed materials (54%), and asking patients what they can accomplish with oral health hygiene (50%). However, most also reported not using some of these communication techniques that they believed were effective, such as drawing pictures or using printed illustrations, using teach-back, using a translator/interpreter, asking other office staff to follow up with post-care instructions, and asking patients if they would like their support person to accompany them to their appointment. Looking more closely at the dental provider survey responses pertaining to just their use of and perceived effectiveness of the AMA’s seven basic patient communication techniques , five of these communication techniques presented are specific to interpersonal communication and the remaining two are teach-back techniques. A majority of providers indicated that they use three of the five interpersonal communication techniques all or most of the time, including using simple language (90%), presenting two–three concepts at a time (80%), and speaking slowly (69%). However, fewer reported asking patients if they would like to be accompanied by a support person for discussion (28%) and drawing pictures or using printed illustration (33%) all or most of the time. A minority of providers reported using either of the teach-back techniques, with only 33% reporting use of teach-back all or most of the time and only 46% reporting asking patients what they will do at home to follow guidance. Additionally, most providers reported not using some of the communication techniques, even though they believed the techniques were effective—such as drawing pictures or using printed illustrations, using teach-back, and asking patients if they would like their support person to accompany them to their appointment.
In the telephone interviews, many providers reported that a key barrier was that they had not received in-depth provider-patient communication and OHL training in professional school or continuing education programs. Most OHL experts noted that clinical dental education was more focused on the “hard skills” of technical procedures rather than the “soft skills” of providing effective patient-centered communication and education. From the survey, all of the dentists, most of the dental hygienists (70%), and more than half of the dental assistants (56%) reported that they had experienced problems communicating with patients. Dental providers and OHL experts identified major barriers to effective provider-patient communication from the pre-survey and the telephone interviews, summarized as: (1) patient/caregiver-side communication barriers and (2) provider-side communication barriers . 3.3.1. Patient/Caregiver-Side Communication Barriers The major patient/caregiver barrier to provider-patient communication that dental providers identified from the survey was that the patient/caregiver does not follow the provider’s instructions (56%), “no matter how well they were explained.” One-third to one-half of dental providers interviewed noted additional barriers underlying the patients’ inability to follow recommended oral health practices, including patients’ lack of understanding of oral health information, language and cultural barriers between providers and patients, and patients’ lack of interest in or prioritization of their oral health. Dental providers referenced specific disadvantaged patient populations with whom they experienced more frequent challenges in provider-patient communication . These populations included patients with limited English proficiency (65%), people with cognitive disabilities (54%), elderly (42%), people with limited education (35%), Deaf or hard of hearing (31%), and early childhood age groups (31%). Experts also noted that patients with limited English proficiency experienced the greatest oral health communication barriers due to cultural differences that often co-occur with language barriers and indicated that most dental practices do not adequately use translation/interpretation services. 3.3.2. Provider-Side Communication Barriers From the telephone interviews, many dental providers and OHL experts cited provider-side barriers to patient communication and education, particularly pertaining to limitations in provider training and clinical practice. Dental providers and OHL experts noted the lack of OHL communication training and proficiency requirements in professional schools and continuing education. They identified clinical practice constraints such as the inadequate time providers have for patient education during the patient encounter, limited reimbursement for delivering patient education, the paucity of high-quality patient education materials in needed languages and accessible formats, and logistical and financial difficulties to access translation/interpretation services. Public practice providers were more likely to underscore the need for technological educational resources such as video/DVD players and tablets, high-quality educational materials in a variety of languages, and translators/interpreters knowledgeable about oral health. One public practice provider discussed these structural challenges, saying: “(The) most obvious barrier is language (…) Time is also the issue. Even if you can delegate, it’s a system issue, if you delegate to dental staff to do the education, it’s taking up chair time/office space (…) Also, many providers have limited training on health communication techniques.” (Dentist, Public Practice) Private practice providers were more likely to cite the lack of time and resources for patient education and reimbursement structures that prioritize dental procedures and treatment over patient education; this was described by one private practice dental assistant: “I would like to have more time and more patient cooperation to learn. (We also) need more (patient education) resources that explain oral health and dental procedures in simple terms and that have a positive tone.” (Dental Assistant, Private Practice) Some of the final provider interviews overlapped with the beginning of the COVID-19 pandemic in California. Although questions about the pandemic’s impact were not included in the study instruments, we note that providers interviewed during February and March 2020 commented that practice disruptions greatly exacerbated communication problems with patients and that guidance was urgently needed to address these new patient communication barriers, both in the changed practice setting and remotely, including tele-dentistry strategies.
The major patient/caregiver barrier to provider-patient communication that dental providers identified from the survey was that the patient/caregiver does not follow the provider’s instructions (56%), “no matter how well they were explained.” One-third to one-half of dental providers interviewed noted additional barriers underlying the patients’ inability to follow recommended oral health practices, including patients’ lack of understanding of oral health information, language and cultural barriers between providers and patients, and patients’ lack of interest in or prioritization of their oral health. Dental providers referenced specific disadvantaged patient populations with whom they experienced more frequent challenges in provider-patient communication . These populations included patients with limited English proficiency (65%), people with cognitive disabilities (54%), elderly (42%), people with limited education (35%), Deaf or hard of hearing (31%), and early childhood age groups (31%). Experts also noted that patients with limited English proficiency experienced the greatest oral health communication barriers due to cultural differences that often co-occur with language barriers and indicated that most dental practices do not adequately use translation/interpretation services.
From the telephone interviews, many dental providers and OHL experts cited provider-side barriers to patient communication and education, particularly pertaining to limitations in provider training and clinical practice. Dental providers and OHL experts noted the lack of OHL communication training and proficiency requirements in professional schools and continuing education. They identified clinical practice constraints such as the inadequate time providers have for patient education during the patient encounter, limited reimbursement for delivering patient education, the paucity of high-quality patient education materials in needed languages and accessible formats, and logistical and financial difficulties to access translation/interpretation services. Public practice providers were more likely to underscore the need for technological educational resources such as video/DVD players and tablets, high-quality educational materials in a variety of languages, and translators/interpreters knowledgeable about oral health. One public practice provider discussed these structural challenges, saying: “(The) most obvious barrier is language (…) Time is also the issue. Even if you can delegate, it’s a system issue, if you delegate to dental staff to do the education, it’s taking up chair time/office space (…) Also, many providers have limited training on health communication techniques.” (Dentist, Public Practice) Private practice providers were more likely to cite the lack of time and resources for patient education and reimbursement structures that prioritize dental procedures and treatment over patient education; this was described by one private practice dental assistant: “I would like to have more time and more patient cooperation to learn. (We also) need more (patient education) resources that explain oral health and dental procedures in simple terms and that have a positive tone.” (Dental Assistant, Private Practice) Some of the final provider interviews overlapped with the beginning of the COVID-19 pandemic in California. Although questions about the pandemic’s impact were not included in the study instruments, we note that providers interviewed during February and March 2020 commented that practice disruptions greatly exacerbated communication problems with patients and that guidance was urgently needed to address these new patient communication barriers, both in the changed practice setting and remotely, including tele-dentistry strategies.
Dental providers interviewed suggested key interventions that could improve provider-patient communication by addressing the barriers they identified. They highlighted the need for more training on OHL communication techniques in professional schools and continuing education. They cited the need for patient-centered clinical practice supports, including more provider time and reimbursement incentives for delivering patient education, translator/interpreter services, high-quality low-cost educational materials in accessible plain-language formats (including videos and print materials) and in a variety of languages. In addition, they wanted information about ways to ensure a patient-friendly office. Dental providers highlighted the foremost importance of the concept of cultural humility: developing trusting, high quality, and professional relationships with patients, understanding the patient’s emotional motivators, and empowering patients to engage in preventive health behaviors. Many providers expressed their hopes that better OHL training, high-quality educational materials, and clinical supports could improve their patient communication and facilitate patients’ trust in providers, and their understanding and motivation to achieve recommended prevention and treatment goals, as described by one dentist: “My dream would be to have California take the lead on a more innovative Medi-Cal (Medicaid—a healthcare program for people with low-incomes) Dental, create more (financial) incentives for positive oral health outcomes. Right now, it is geared to treating disease and procedures. Need to flip to outcomes based—use (performance-based) incentives for patients and providers to become healthier—achieve scalable results.” (Dentist, Public Practice) OHL experts recommended further structural changes in dental professional training and clinical practice and beyond. They included incorporating OHL curricula in dental professional schools and continuing education through dental professional societies, adding OHL competency requirements for clinical license examinations and/or renewal, and expanding OHL training to medical providers and health coaches. The range of suggestions for training was encapsulated by one OHL expert: “We need to make top-down changes, starting the changes with accrediting bodies for medical/dental education, (…) board exams, (…) license recertification, and clinical quality measures and financial incentives to include good health literacy and oral health literacy practices. There are some individual medical/dental schools and scattered continuing education courses that may include health literacy/oral health literacy, but that only reaches a small subset of medical/dental providers, particularly those who are already aware of the issues. To draw in new participants, it’s helpful to tag on oral health literacy to other conferences (e.g., general pediatrics) and present compelling cases to show how oral health is critical to overall health.” (OHL expert) Experts also discussed the need for training the entire dental team in effective OHL strategies, including establishing a patient-friendly environment, using OHL communication techniques, and using oral health-specific translator/interpreter services such as via phone/video services and natural language processing software. Dental providers in public practice and OHL experts also recommended expanded community oral health education, through integrative dental and medical care models particularly during early childhood (e.g., in pediatrics), community events and schools from early childhood through high school, to raise the public’s knowledge of oral health and help patients feel more comfortable with dental care.
The global epidemic of poor oral health and increasing oral health inequities presents major public health challenges. Despite substantial evidence that most oral health problems are preventable, national and global oral health goals have not been achieved. Two decades of research has identified OHL as a major determinant of oral health status through multiple interrelated factors that include people’s understanding about oral health, providers’ knowledge of current scientific evidence, the quality of dental providers’ communication with patients, and the “patient-friendliness” of dental practice environments. Although early OHL research focused primarily on individuals’ abilities to understand oral health information, during the past decade, an increasing number of studies have investigated the impact of dental providers’ communication with patients. This reflects the perspective that oral health is not just an individual responsibility but an issue of social justice and health equity impacted by a myriad of factors across social-ecological levels . These studies have been conducted primarily in the US and have shown that although dental providers use some recommended communication techniques, such as speaking slowly, using simple language, and visual models, overall, dental providers reported using less than half of recommended communication techniques with their patients, including the important teach-back approach . Research in the UK has also demonstrated that providers generally provide education as “ad hoc” lectures, use few visual aids, and rarely provide take-home educational materials . Results from our study showed that providers greatly valued their communication with patients and believed they were effective communicators. During the interviews, we were struck by the care and concern providers expressed for their patients and their strong desire to help them improve their oral health. We were also struck by the frustration they expressed about communicating effectively with patients: the vast majority cited multiple barriers. A notable finding was that providers most often attributed communication problems to patient-side, rather than provider-side barriers. shows that one-third or more of dental providers cited six out of the seven patient-side barriers, but only one out of the five provider-side barriers. In other words, providers focused on patient issues such as “not understanding” oral health information or “not being interested” in their oral health. This perception aligns with the historic “patient deficit” health literacy approach and signals the critical need for provider OHL training about their key role to help patients understand and improve their oral health. In this study, providers tended to believe their communication practices were quite good, but despite that, the patient or caregiver does not follow their instructions. Yet, few providers indicated they used teach-back—the communication technique most likely to get positive results. This suggests that providers may not recognize their need for training to use recommended communication techniques and how those techniques could significantly help them and their patients. In the larger scientific context, the concept of health literacy has undergone a transformation from placing the communication burden on the patient to transferring that responsibility to the provider . Study findings described providers’ knowledge and use of OHL techniques. Most providers were familiar with the term “oral health literacy,” but few fully understood that it goes beyond patients’ comprehension to include the quality of provider communication with patients and the patient-friendliness of dental practices. Because OHL is still a new concept to most providers, we believed it was important to follow up on online survey responses with individual interviews. For example, although two-thirds of providers reported in the pre-survey that they had taken a patient communication course in professional training, probing during the interviews revealed that such training did not generally go into depth about OHL and its recommended communication techniques. Methodologically, this study highlights the importance of using in-depth interviews rather than just relying on provider survey responses. In this study, providers’ reported OHL practices were similar to findings in prior OHL research. California providers reported using 6 of AMA’s 18 recommended communication techniques and 3 of AMA’s 7 basic communication techniques. Of the 18 communication techniques, providers most often reported using simple language, speaking slowly, presenting only two–three topics at a time, and using visual models. There was less routine use of other recommended techniques, such as teach-back, printing out instructions, referring patients to the Internet or other resources, using videos or DVDs, or following up with patients after a dental visit, etc. Of the seven basic communication techniques, providers indicated that they most often used three of the basic techniques categorized as interpersonal communication, including using simple language, presenting two–three concepts at a time, and speaking slowly. There was less routine use of other recommended basic techniques, such as asking patients if they would like to be accompanied by a support person for discussion, drawing pictures or using printed illustration, and teach-back strategies, including asking patients what they will do at home to follow guidance. Of the five assessment and educational strategies we asked about, a majority of providers interviewed indicated they were knowledgeable of and used one of the five strategies (CAMBRA), but were unaware of or did not use Family Oral Health Education, anticipatory guidance, teach-back, or motivational interviewing. For less-frequently used techniques, providers also reported “less confidence” in the effectiveness of those techniques, indicating a gap in understanding about OHL and its associated communication techniques. Although half of providers reported that they had assessed their practice for patient-friendliness, they did not confirm how they did so, suggesting that the assessment may have been intuitive, rather than based on recommended processes, such as those in the “Health Literacy Environment 2” tool . For example, providers could assess whether their staff provides patients with plain language information and translation services or helps filling out forms. In this study, providers wanted information about these processes. The study’s results also corroborate prior studies and expert recommendations pertaining to key communication challenges facing providers. Provider challenges included the lack of professional training in OHL communication, lack of accessible, in-language patient educational materials, lack of translator/interpreter services, and lack of provider time and financial incentives for patient education. Lack of OHL can lead to poor oral hygiene at home and uninformed decisions about fluoridated tap water and fluoridated toothpaste, and inappropriate amounts of toothpaste used. A 2020 study by Avenetti et al. comments on the lack of knowledge that parents/caregivers have regarding fluoride use, fluorosis, and best oral health behaviors for their children. For example, parents purchasing toothpaste were often confused by marketing strategies and labels on non-fluoridated toothpaste such as “safe for babies” or on fluoridated toothpaste encouraging parents to consult a dentist or physician for use in children under 2 . For young children in particular, promoting OHL is vital because the prevention of caries is primarily in the hands of the parents/caregivers. Furthermore, their efforts regarding the oral health of their child serves to establish oral health behaviors that will impact the child’s life course . Providers and national experts in this study recommended that achieving marked improvements in patient communication requires prioritizing more OHL professional training (academic and continuing education) in basic communication techniques and access to low-cost, high-quality plain language and in-language educational materials and tools. They also recommended better access to translator/interpreter services knowledgeable in oral health, and greater financial incentives to incentivize providers to provide patient education. Additionally, providers and national experts recommended further systematic changes, including curricular requirements that OHL communication be taught in dental professional schools, that content be included in continuing education, and that there be assessment of provider performance of OHL competencies. This might include strategies in professional continuing education, as peer or expert performance evaluations of mock patient interviews, or adding content to state licensing exams. They also mentioned that wider access to dental insurance is needed to extend oral health education to more people, especially to disadvantaged groups such as people with limited English proficiency, people with cognitive disabilities, older adults, people with low literacy, people who are Deaf or hard of hearing, and children. Providers and experts also emphasized the importance of promoting inter-professional collaboration about oral health literacy. Not only do dentists need to be educated on how to teach and support patients, but other medical professionals need training as well . Having additional patient educators, such as community health workers or promotoras may benefit dental practice communication . In many practices, the pressure on dentists and hygienists to see a high volume of patients for reimbursable treatments may not allow them adequate time to provide “soft” communication and literacy support. Community health workers/promotoras can help share the workload by providing the needed culturally relevant communication and support as part of the care team. Furthermore, both providers and national experts recommended going beyond dental practices to extend oral health education to the broader community level, including oral health patient education in medical practices, especially in pediatric and family medicine practices , mass media communication, and oral health education in schools from early childhood through high school. This study adds to the limited global literature about OHL and dental provider communication and is the first to examine California’s dental provider/national OHL expert perspectives on oral health communication. The study has limitations. Due to its convenience, purposive sample, study results for knowledge and use of educational and communication techniques and other findings are not widely generalizable. However, it includes 44 in-depth interviews that achieved thematic data saturation and 40 completed pre-surveys with the same providers, across the major groups of dental professionals. This is complemented by six interviews with OHL experts, which together with provider interviews, present a wide range of perspectives from California dental providers and OHL national subject matter experts. In this study, we included pre-surveys to gather background information about provider and expert characteristics and perspectives prior to conducting the in-depth interviews. We have found this a useful technique to reduce the time needed in the qualitative interview, and to provide background information to the interviewer about interviewees’ perspectives to be explored more deeply in the qualitative interview. This approach can also help identify possible weaknesses of relying on survey responses without further probing them in interviews. For example, although most providers reported in the pre-survey that they were familiar with the term “health literacy,” during the interviews, it was apparent that they had a very limited view of that concept—limiting it to just the literacy skills of patients. We recommend the use of “pre-surveys” prior to conducting key informant interviews. The study questions were informed by those in national and state studies of OHL among dental professionals (content validity), with the addition of California-specific questions including those covering suggestions for training, patient resources, and payment. In addition, responses from KIs in each dental professional subsample achieved “data saturation” on major issues and results were aligned with those found in other larger, quantitative studies. This study addressed perspectives of dental providers. More research is needed about patient and general public perceptions of the dental providers’ communication skills. Other variables related to patient-provider communication could be explored, such as the “Oral health-related quality of life (OHRQoL)” tool, an assessment intended to help patients and providers make better shared decisions about treatments . Smaller categories of techniques help reduce the overlap in some of the AMA’s 18 communication techniques, put more focus on priority techniques, and make it easier to incorporate techniques into more practical educational interventions, including those using digital channels, such as through networks of dental providers. This study was intended to convert research into practice and provided rich, detailed recommendations about preferences for training and supportive resources. The CDPH OOH is now implementing these results and those of past studies in a collaborative effort with oral health stakeholders to create OHL continuing education trainings beginning in 2021, and a toolkit of OHL resources for dental providers and community organizations . The American Dental Association’s National Advisory Committee on Health Literacy in Dentistry is considering ways to adapt the proposed trainings and toolkit resources to reach national dental provider and community audiences. Such interventions could inform similar efforts in other countries. The COVID-19 pandemic has greatly increased provider-patient communication barriers and widened oral health disparities, underscoring the urgent need to draw on the substantial OHL research and practice guidance of the past two decades to train providers and accelerate the search for innovative, effective patient communication interventions.
Two decades of OHL research suggest that gaps in people’s understanding of oral health and providers’ low use of recommended communication techniques with patients are important factors contributing to poor oral health in the US and globally. Fortunately, such research also provides evidence-based recommendations for oral health interventions. This study focused on dental providers’ reported communication practices and perceived barriers and facilitators to improved communication. Results showed major gaps between current dental provider communication practices and those recommended in prior research and by medical and dental professional organizations—similar to the gaps found in other studies. The in-depth interviews in this study add new knowledge about providers’ focus on patient OHL limitations vs. their own patient communication limitations and the many barriers providers experience in improving their communication skills and incorporating them into a busy practice. The study results detail specific provider and expert recommendations related to OHL training and resources to improve OHL communication. These results underscore the larger issue that the epidemic of oral health problems will require actions at all levels of society from education of individuals to policy changes. This study was intended to focus on the dental provider component of the oral health ecosystem with the intent to use the findings to create interventions to improve dental provider OHL and patient communication, and ultimately to adapt interventions to larger numbers of dental providers, to medical oral health providers, to community organizations, and to advance policy changes. In our view, using research to identify issues, while simultaneously generating solutions relevant to specific contexts—in this case, dental providers in the US state of California—is the most powerful way to address complex problems, such as oral health.
|
Pharmacogenomics for Primary Care: An Overview | d8e8f25e-f981-4a24-9e1f-f0129208735b | 7696803 | Pharmacology[mh] | It is increasingly recognised that individuals respond differently to medications. In some cases, these differences in response can be clinically significant leading to failure of therapy or adverse drug reactions (ADRs). The aetiology of this inter-individual variability is complex with multiple contributors including individual characteristics (e.g., age, sex, body mass index), clinical factors (e.g., renal or hepatic impairment; co-medications), environmental exposures (e.g., smoking) and genetics. Pharmacogenomics (PGx) is the study of the influence of genetic variation on drug response with the aim of increasing the efficacy and safety of current and future treatments. Specifically, it aims to facilitate a move away from the standard empirical trial and error prescribing approach that currently exists and transition towards a more stratified and precise prescribing paradigm. It is estimated that there are between 19,000 to 21,000 protein-coding genes present in the human genome . Within these genes, multiple types of genetic variations can occur including single nucleotide polymorphisms (SNPs), indels (small insertion/deletions) and larger structural rearrangements; of these, SNPs are the most common. Pharmacokinetics (PK) describes “what the body does to a drug” and pharmacodynamics (PD) “what a drug does to the body”. Genomic variation in genes involved in a drug’s absorption, distribution, metabolism and elimination (e.g., drug-metabolising enzymes or transporters) can alter a drug’s PK profile, influencing systemic exposure and resulting in altered drug response (i.e., influencing its downstream PD). Alternatively, genomic variation in genes that modulate a drug’s PD (e.g., its therapeutic on-target and off-target sites) can directly influence drug response. Importantly, in both cases, the altered drug response can attenuate a drug’s efficacy or worsen tolerability/safety . The clinical and financial consequences of adverse drug reactions (ADRs) are high, accounting for an estimated 6.5% of hospital admissions . Interestingly, for several of the drugs often implicated in causing hospitalisation, PGx guideline recommendations are now available, such as for warfarin, antiplatelet agents and opioid analgesics . In addition to serious ADRs leading to hospitalisation, it is also known that poor drug tolerability (for example due to mild adverse events) is associated with lower compliance , which increases the chances of reduced efficacy and increased medicines wastage. It is estimated that over 98% of individuals carry at least one pharmacogenetic variant . Importantly, the majority of prescribing and dispensing of medicines happens in primary care, and recent studies in both the US and UK suggest that over 60% of patients within the primary care setting are prescribed a medication with a PGx recommendation The Dutch Pharmacogenomics Working Group (DPWG) and Clinical Pharmacogenetics Implementation Consortium (CPIC) are the two most widely recognised expert groups involved in the development of PGx clinical guidelines. Therapeutic recommendations from CPIC and DPWG, as well as guidelines written by other groups such as the Canadian Pharmacogenomics Network for Drug Safety (CPNDS), are curated and housed by the Pharmacogenomics Knowledge Base ( https://www.pharmgkb.org/ ), and can be accessed online for free by healthcare professionals and other individuals with an interest. To date, PGx associations with actionable PGx recommendations for clinical practice have been developed for just over 80 drugs . Almost two-thirds of these actionable drug-gene associations involve drug metabolizing enzyme genes, with ~80% of these being genes encoding cytochrome P450 (CYP) enzymes. A small number of actionable drug-gene associations involve transporter genes, and just under a third involve genes that influence drug PD (~5% on-target, ~26% off-target, with almost a third of the latter involving human leukocyte antigen ( HLA ) genes). The US Food & Drug Administration (FDA) has also evaluated drug-gene associations, deeming 47 drugs to have sufficient evidence for PGx therapeutic recommendations and another 16 drugs to have PGx associations that potentially impact clinical safety or response ; there is notably overlap between the drug-gene associations considered actionable by the guideline committees and the FDA. A list of drugs (though not exhaustive) commonly prescribed in primary care with PGx guidelines currently available is outlined in . Despite the growing availability of guidelines, adoption of PGx in primary care has been slow. Nevertheless, a nationwide study from the Netherlands inferred that 1 out of every 19 new prescriptions in primary care could have undergone an adjustment had PGx data been available . Within the UK and in many other countries, PGx testing largely remains within the remit of specialist secondary and tertiary care settings (e.g., with abacavir- HLA B*57:01 testing). provides a summary, albeit non-exhaustive, of recent and large interventional studies that have assessed PGx and are relevant to primary care. A recent analysis identified the following pharmacogenes to be most commonly linked to primary care prescriptions in England: CYP2D6 , CYP2C19 and SLCO1B1 (solute carrier organic anion transporter family member 1B1), followed by CYP2C9 , VKORC1 (vitamin K epoxide reductase complex 1), CYP4F2 and HLA-B (8).Therefore, the subsequent sections in this review provide an overview of the role PGx can play in primary care in the prescribing and monitoring of specific medications related to these genes. Drug-specific literature searches were conducted between June and October 2020.
Prescribed for a variety of indications, the number of prescriptions for antidepressants in the UK primary care sector is increasing annually . Whilst they are effective therapies, these drugs have variable success rates , with up to 50% of patients nonresponsive to treatment. Some of this heterogeneity in response may be accounted for by PGx variation. In the UK, selective serotonin re-uptake inhibitors (SSRIs) account for over 50% of primary care antidepressant prescriptions ; tricyclic antidepressants (TCAs), serotonin-norepinephrine reuptake inhibitors (SNRIs) and atypical antidepressants (e.g., mirtazapine) are also utilised. There are 57 putatively functional CYP genes within the human genome, of which around 12 are involved in the biotransformation of 70–80% of all therapeutics used in clinical practice . Most antidepressants are, in some part, metabolized via the CYP enzyme system, with CYP2D6 and CYP2C19 widely regarded as the most influential enzymes for antidepressant biotransformation. CYP2D6 is highly polymorphic with over 100 allelic variants recorded . Due to the number of possible diplotypes within a population the translation of genotype to phenotype for CYP2D6 is often performed using an “activity score” wherein alleles are given a numeric value based on their functionality and the sum of the values present in an individual is subsequently categorised into one of the following “metaboliser phenotypes”: ultra-rapid metabolizer (UM), extensive (normal) metabolizer (EM), intermediate metabolizer (IM), and poor metabolizer (PM) . Allelic frequencies in CYP2D6 vary substantially across populations. The most common non-functioning variant, CYP2D6*4 (rs3892097, 1846G > A), which is a splicing defect, is found at the highest frequency within Caucasian populations ; related to this observation Caucasian populations have a higher frequency of PMs (around 5–10%) compared to Asian and African-American populations. CYP2D6 is also subject to copy number variation (CNV) that can lead to deletion of a CYP2D6 allele (*5) or CYP2D6 duplication or multiplication; whilst the former decreases CYP2D6 function, an increased number of functional CYP2D6 alleles results in the UM phenotype (*xN). UMs are less frequent in Caucasian populations at approximately 2%, but the frequency of UMs has been noted to be much higher in other specific populations, such as those with North African ancestry (c.25%) . Studies have demonstrated that the presence of the CYP2D6*4 allele in PMs correlated to increased side effect frequency with some antidepressants including venlafaxine and lower dose requirements of SSRIs and TCAs (e.g., amitriptyline, nortriptyline; see below) . Additionally, studies have indicated that CYP2D6 UMs taking paroxetine have low plasma concentrations in comparison to EMs and this may be a risk factor for therapeutic failure. In the case of both fluvoxamine and paroxetine, the CPIC guideline highlights the possibility of ADRs potentiated by higher plasma concentrations in CYP2D6 PMs . Some evidence exists to correlate increased ADRs with mirtazapine in CYP2D6 IM and PM patients . However, at this time, no clinically actionable PGx recommendation exists for mirtazapine. CYP2C19 is similarly highly polymorphic: 34 conventional star alleles are currently recognised and over 2000 CYP2C19 variants have been identified, although the majority are intronic . The most prevalent reduction-of-function (ROF) allele is *2 (rs4244285, c.681G > A), a splice defect, followed by *3 (rs4986893, c.636G > A) which results in a premature stop codon. CYP2C19*2 is most common in individuals of Asian ancestry with minor allele frequencies (MAFs) of 31–36% , whilst occurring in ~16% of African and European individuals. CYP2C19*3 occurs in ~6% of East Asian individuals, but is rare in African and European populations. Conversely, CYP2C19*17 (rs12248560), which is a common variant in the promotor region of the gene, is associated with increased transcription and enzymatic function . Similar to CYP2D6 , individuals can be categorised according to genotype into anticipated CYP2C19 EMs (*1/*1 wild-type), IMs (one ROF allele), PMs (two ROF alleles), and UMs (*17/*17). Unlike conventional CYP2D6 phenotype categorisation, a CYP2C19 rapid metaboliser (RM, *1/*17) phenotype category is also recognised with functionality between EMs and UMs. Sertraline, citalopram and escitalopram are all extensively metabolised by CYP2C19 and, in PMs, elevated plasma concentrations may lead to potential ADRs . Presently, however, a formal drug label warning only exists for escitalopram, which recommends a reduced starting dose for CYP2C19 PMs due to the risk of QT interval prolongation . Amitriptyline undergoes CYP2C19-mediated demethylation to its active metabolite, nortriptyline, which is also available on prescription. Whilst both are classed as TCAs, amitriptyline blocks serotonin and noradrenaline reuptake equally, whereas nortriptyline inhibits noradrenaline uptake more potently . As a result, CYP2C19 PMs are expected to show reduced conversion to nortriptyline, increased systemic exposure to amitriptyline and an increased risk of amitriptyline-related side effects . Although TCAs are now more commonly used at lower doses for the treatment of neuropathic pain, rather than for major depressive disorder, the CPIC guidelines for amitriptyline still recommend consideration of alternative therapy in CYP2C19 UMs due to an increased risk of treatment failure irrespective of indication . Additionally, amitriptyline and nortriptyline also undergo CYP2D6 metabolism to a less active metabolite. Whereas CYP2C19 influences the ratio of metabolites, CYP2D6 plays a greater role in drug clearance. This may lead to raised plasma concentrations of amitriptyline and nortriptyline in CYP2D6 PMs or in patients co-administered a CYP2D6 inhibitor . CPIC guidelines therefore recommend a lower starting dose when used at dosing ranges for depression . A recent meta-analysis of five randomised clinical trials (RCTs) reported that patients receiving PGx guided dosing were 1.71 (95% CI: 1.17–2.48, p = 0.005) times more likely to achieve symptom remission for major depressive disorder (MDD) than those on standard care . Details of some of the most recent RCTs are in . While there is still work to be done on the utilisation of PGx in antidepressant prescribing, there are multiple potential benefits to having PGx information at hand when selecting treatments. An analysis of UK healthcare data from 2016 identified that the age categories with highest antidepressant use were ages 55–64 and 75–84 . Additionally, the same data set demonstrated an increasing trend for polypharmacy with increasing age, particularly in the over 75-years-old age bracket. PGx testing has the potential to circumvent some of the risks associated with polypharmacy by guiding the prescribing of suitable medications . In addition to patient benefits, analyses have shown that utilising PGx information in mental health prescribing may offer significant financial savings via a reduction in the number of failed treatments and subsequent medicines wastage .
Codeine and tramadol are commonly prescribed weak opioid analgesics indicated for mild to moderate (non-neuropathic) pain; they are positioned on the second rung of the World Health Organisation’s three step pain ladder. Codeine is an inactive prodrug; 0–15% undergoes O-demethylation catalysed by CYP2D6 to form active morphine, which has an affinity for the µ-opioid receptor 200-fold more than that of parent codeine . Similarly, tramadol undergoes CYP2D6-mediated O-demethylation to O-desmethyltramadol , which has an affinity for the µ-opioid receptor that is also significantly higher than that of tramadol . Whereas the majority of codeine’s analgesic effect stems from morphine agonism on opioid receptors, for tramadol, the analgesia conferred by O-desmethyltramadol agonism of the µ-opioid receptor is complemented by inhibition of serotonin and noradrenaline reuptake by parent tramadol. In CYP2D6 EM individuals, at least 80% of codeine metabolism results in inactive metabolites, with around 5–10% undergoing biotransformation into morphine . This percentage increases in UMs whilst, in PMs, there may be very little if any conversion (52). PK and PD studies have demonstrated that the increased codeine biotransformation in CYP2D6 UMs can result in opiate toxicity even with low doses of codeine . Conversely, CYP2D6 PMs are likely to obtain little to no therapeutic benefit . In addition, due to the heterogeneity in CYP2D6 allele functionality, some patients within the broad EM category may also experience an increased number of adverse events compared with others. These “ultra-rapid EMs” have alleles leading to a greater activity score, but which still fall within the EM activity score reference range, and thus ideally require lower doses to circumvent possible ADRs . The PGx of codeine received international attention following reports of infant opiate toxicity from higher levels of morphine in the breast milk of UM mothers. Reviews of this phenomenon resulted in amended FDA and European Medicines Agency (EMA) guidance to avoid the use of codeine when breastfeeding . Similarly, regulatory changes followed case reports of children with obstructive sleep apnoea receiving codeine post-tonsillectomy and/or adenoidectomy leading to codeine toxicity and fatalities, with these children having evidence of being CYP2D6 UMs . This led to blanket regulation from authorities such as the UK Medicines and Healthcare products Regulatory Agency (MHRA) restricting the use of codeine in children under 12 irrespective of CYP2D6 metaboliser status. This same MHRA guidance extends to cover all patients (i.e., 12 years and older) with a documented CYP2D6 UM status, although in practice this is seldom known. Both CPIC and the DPWG have developed PGx guidelines for codeine that advocate avoiding codeine in UM and PM patients. However, there are nuanced differences between these recommendations. Most pertinently, in UMs, CPIC recommends simply avoiding codeine, whereas the DPWG guidance suggests no additional action is required in UM patients that receive lower doses of codeine and have no additional risk factors (e.g., concomitant CYP3A4-inhibiting drugs) that would predispose to exaggerated biotransformation to morphine. In addition to the influence of CYP2D6 , there is emerging evidence for the influence of polymorphisms within other genes including the µ-opioid receptor gene, OPRM1 , and the phase 2 metabolising enzyme UGT2B7 . While associations with variation in these genes remain inconclusive, they are important areas for future research. Within the UK, the use of the weak opioid tramadol has increased over the past two decades . In keeping with codeine, studies have demonstrated a probable reduced efficacy with tramadol in CYP2D6 PMs patients and, additionally, a lower risk of ADRs . Again, the risk of ADRs including potentially severe reactions appears greater in UMs . A systematic review of the CYP2D6*10 C188T polymorphism which included 9 studies and 809 subjects showed a relationship between this polymorphism and pharmacokinetics of tramadol (half-life, AUC and clearance) and analgesic effect, but not with the adverse effects of nausea and vomiting . Although CPIC has yet to develop guidance for tramadol, the DPWG guidelines recommend avoiding tramadol in UMs or, if not possible, using just 40% of the standard dose, to minimise the risk of opioid receptor-mediated ADRs. For PMs, an increased alertness to reduced effectiveness is primarily recommended. This is because it is difficult to predict the influence of decreased CYP2D6 metabolism on the overall analgesic effects experienced by an individual, because the resulting increased proportion of parent tramadol to O-desmethyltramadol may enable the potentiated SNRI properties of parent tramadol to mitigate some of the lost O-desmethyltramadol opioid receptor-mediated analgesic effects.
Statins are first line lipid-lowering agents for both primary and secondary prevention of cardiovascular disease, are widely prescribed, and are generally safe and well tolerated. Statins are, however, associated with an increased risk of type 2 diabetes mellitus and muscle toxicity, which is phenotypically heterogeneous ranging from myalgias with normal plasma creatine kinase (CK, ~5% of individuals ) to infrequent myopathies to rare rhabdomyolysis, and finally extremely rare immune-mediated necrotizing myopathy (IMNM) . A recent study showed that the risk of myopathy (CK > 10 × the upper limit of normal) with simvastatin was 9 per 10,000 person-years of therapy. Independent risk factors for myopathy included simvastatin dose, ethnicity, sex, age, body mass index, medically treated diabetes, and concomitant use of certain drugs (niacin-laropiprant, verapamil, β-blockers, diltiazem and diuretics), which collectively predicted more than a 30-fold risk difference between the top and bottom thirds of a clinical myopathy score. This risk score was less strongly associated with milder myopathy, and not associated with reports of any muscle symptoms, and so the authors concluded that muscle presentations other than myopathy were unlikely to be related to statins . Statins competitively inhibit 3-hydroxy-3-methylglutaryl-Coenzyme A reductase (HMGCR) in the liver, resulting in decreased circulating low-density lipoprotein cholesterol (LDL-C). Organic anion-transporting polypeptide 1B1 (OATP1B1) is a liver-specific transporter expressed on the basolateral (sinusoidal) side of hepatocytes , is encoded by SLCO1B1 and is involved in hepatic uptake of statins and other drugs (e.g., letermovir ). A common ROF SLCO1B1 missense variant, rs4149056 (c.521T > C, p.V174A, present in SLCO1B1*5, *15 and *17), is associated with a 221% increased systemic exposure to simvastatin acid in 521CC compared to wild-type 521TT individuals. Moreover, the C allele is associated with elevated exposures to all statins, except fluvastatin, albeit to lesser extents . For example, the corresponding increase with atorvastatin is 145%. The MAF of rs4149056 in African, East Asian and European populations is 1%, 12% and 16%, respectively. Importantly, it was shown in a genome-wide association study (GWAS) that 521TC and 521CC patients taking simvastatin 80mg daily had an odds ratio (OR) for myopathy of 4.5 (95% confidence interval (CI) 2.6–7.7) and 16.9 (95% CI 4.7–61.1) compared to wild-type patients, respectively . In 521TC patients on simvastatin 40mg, the relative risk was 2.6 (95% CI 1.3–5.0). Furthermore, the 521C allele is associated with a two-fold increase in statin intolerance in patients predominantly taking simvastatin, and this intolerance can reduce lipid-lowering efficacy . Thus, the strength of SLCO1B1 -myotoxicity association increases with 521C allele dose, simvastatin dose and myopathy severity. Several other studies including a recent large-scale GWAS have replicated the simvastatin-521T > C association. The consistent association with simvastatin mirrors the PK observations that 521T > C has the largest impact on simvastatin exposure, reflects the high prevalence of simvastatin use in clinical practice leading to reliance on simvastatin myotoxicity cases in research studies, and the in vitro observations that simvastatin (lactone) is particularly myotoxic . Elevated simvastatin systemic exposure presumably increases skeletal muscle exposure, predisposing to myotoxicity by mechanisms that are incompletely understood but likely to include mitochondrial dysfunction, calcium signalling disruption and reduced prenylation. The evidence for an association between SLCO1B1 521T > C and other statins is less clear. Some studies have suggestively associated 521C with atorvastatin myotoxicity , but others found no association . Interestingly, a recent study in patients on high dose atorvastatin for secondary prevention found 521C was associated with both muscular symptoms and atorvastatin intolerance, suggesting that 521T > C may be more relevant in patients on high dose (e.g., 80 mg) atorvastatin . For rosuvastatin, 521C was not associated with myalgias in patients of European descent , but recently has been associated with myotoxicity (myalgias to rhabdomyolysis) in Chinese ancestry patients . SLCO1B1 521T > C has not been associated with pravastatin myotoxicity. A recent international whole-exome sequencing endeavour of patients with statin (most commonly simvastatin) myopathy identified no novel variants . Nevertheless, a candidate missense variant (rs12975366, p.D247G) within leukocyte immunoglobulin-like receptor subfamily B member 5 ( LILRB5 ) was recently associated with statin myotoxicity, including intolerance and myalgia, implicating the immune system in these more mild phenotypes ; an intervention study is underway . Moreover, CYP3A4*22 and CYP3A5 non-expressors have been associated with modestly elevated simvastatin exposure , CYP3A7*1C with increased atorvastatin hydroxylation , and ABCG2 rs2231142 (c.421C > A) with elevations in exposure to simvastatin, atorvastatin, fluvastatin and especially rosuvastatin . Nevertheless, despite signals, these candidate genes and others (e.g., ABCB1 ) have not been consistently associated with statin myotoxicity across studies. Similarly, CYP3A4 , CYP3A5 and CYP2D6 have been investigated as potential biomarkers of statin efficacy, but results have varied and, of note, none were identified in a GWAS meta-analysis of statin lipid-lowering efficacy . Nevertheless, the influence of more complex interactions, such as those involving combined genetic variation in both SLCO1B1 and CYP and/or other transporter genes, on statin pharmacokinetics and clinical effects, warrants further study. In addition, CYP2C9 ROF variants have been associated with increased adverse events (primarily myotoxicity ) and potentially increased lipid-lowering efficacy . Fluvastatin has arguably received less research attention than simvastatin and atorvastatin, yet CYP2C9 is important in fluvastatin’s metabolism, and so follow up studies are warranted to further investigate these findings. SLCO1B1 521T > C has been associated with the spectrum of myotoxicity, at least for simvastatin, except for IMNM. A subtype of IMNM is anti-HMGCR myopathy ; most patients with anti-HMGCR myopathy have a history or statin exposure, develop muscle weakness with highly elevated CK levels that persist despite statin cessation, are identified by the presence of circulating anti-HMGCR autoantibodies that can be directly pathogenic, and normally require treatment with immunosuppressive drugs or intravenous immunoglobulins . Interestingly, HLA-DRB1*11:01 has been significantly associated with anti-HMGCR myopathy with estimated OR of ~25–57 dependent on ethnicity. Whilst this HLA association may catalyse research into underlying mechanisms and aid diagnosis, the rarity of anti-HMGCR myopathy suggests this PGx association will not have utility in guiding statin initiation. Both CPIC and the DPWG have developed clinical guidelines for simvastatin-SLCO1B1 , and the DPWG have also developed atorvastatin-SLCO1B1 guidance. In patients that carry 521C, they recommend starting an alternate statin or using a lower simvastatin dose. For atorvastatin, the DPWG guideline recommends an alternative statin in 521C carriers that have additional clinical myotoxicity risk factors. Nevertheless, few patients are now started on simvastatin 80 mg following an FDA warning about its increased myopathy risk . Interestingly, a recent randomized trial in 159 patients with previous statin myalgia has demonstrated that providing SLCO1B1 521T > C genotype with recommendations increases statin re-initiation in primary care . As muscle symptoms are associated with statin discontinuation and non-adherence , which in turn increase the risk of cardiovascular events , 521T > C genotyping may help reduce cardiovascular events as well as myotoxicity in clinical practice, although this remains to be substantiated.
Clopidogrel is an antiplatelet drug indicated in acute coronary syndrome (ACS), percutaneous coronary intervention (PCI), stroke, transient ischaemic attack (TIA), peripheral artery disease (PAD) and atrial fibrillation (AF). Clopidogrel is a prodrug: of the ~50% absorbed, ~85% is rapidly hydrolysed to inactive clopidogrel carboxylic acid via carboxylesterase 1 (CES1) and ~15% undergoes a two-step oxidative biotransformation to produce the active thiol metabolite that irreversibly inhibits platelet P2Y12 receptors . CYP2C19 is the only CYP substantially involved in both oxidative steps, contributing 45% and 21% to the first and second steps, respectively . CYP2C19 ROF alleles are associated with decreased levels of circulating clopidogrel active metabolite and increased ex vivo high on-treatment platelet reactivity (HTPR) . A meta-analysis of 9685 patients, of whom 91% had undergone PCI (55% ACS), demonstrated that CYP2C19 ROF alleles were associated with an increased risk of major adverse cardiovascular events (MACE) and stent thrombosis, with a gene-dose trend evident . However, a second meta-analysis assessing CYP2C19 genotype and cardiovascular outcomes in 26, 251 individuals found no overall association with cardiovascular outcomes after exclusion of small studies, although an increased risk of stent thrombosis remained evident . This meta-analysis included a wider spectrum of indications for clopidogrel, including atrial fibrillation. To resolve these discrepant observations, clopidogrel indication-specific PGx has been posited. Thus, these (mostly) observational data have been considered to support an association between CYP2C19 ROF variants and cardiovascular events in post-PCI patients, because this patient population’s high baseline risk for cardiovascular events makes them more susceptible to suboptimal treatment, but does not support an association in other (mainly cardiac) settings where the overall benefit of clopidogrel is modest anyway . To investigate whether interventions based on CYP2C19 ROF alleles improve clinical outcomes, two large randomized clinical trials (RCTs) have been undertaken and recently reported . The Patient Outcome after primary PCI (POPular) Genetics trial included 2488 patients with an ST-elevation myocardial infarction (STEMI) that underwent primary PCI with stent insertion and compared standard-treatment (ticagrelor or prasugrel) to a genotype-informed antiplatelet strategy that allocated ticagrelor/prasugrel to CYP2C19 *2 or *3 carriers and clopidogrel to noncarriers. POPular Genetics reported the genotype strategy was non-inferior to standard-treatment for net adverse clinical events (MACE plus major bleeding, p < 0.001 for noninferiority), but reduced (mostly minor) bleeding (hazard ratio (HR) 0.78, 95% CI 0.61–0.98, p = 0.04). The Tailored Antiplatelet Initiation to Lessen Outcomes Due to Decreased Clopidogrel Response after PCI (TAILOR-PCI) RCT recruited 5302 patients undergoing PCI with stent insertion (for ACS or stable coronary artery disease) and compared standard-treatment clopidogrel to an equivalent CYP2C19-informed antiplatelet strategy to that used in POPular Genetics. The primary result of TAILOR-PCI was a reduction in first MACE in the genotype strategy arm compared to the standard (clopidogrel) arm, although it narrowly missed statistical significance (HR 0.66, 95% CI 0.43–1.02, p = 0.06). A pre-specified sensitivity analysis showed a significant reduction in all cumulative MACE events (HR 0.60, 95% CI 0.41–0.89, p = 0.01) with genotyping. There was no difference in bleeding risk between arms. Of note, the power to detect a difference in TAILOR-PCI was lower than originally planned because the event rate was lower than anticipated, in keeping with use of newer generation drug-eluting stents . CYP2C 19 *17 has been associated with decreased HTPR, and inconsistently with increased bleeding risk and reduced MACE . Given the inconsistent findings, less observational research has focused on *17 compared to CYP2C19 ROF alleles, and that *17 was not included in the POPular Genetics or TAILOR-PCI primary analyses, the clinical relevance of *17 remains undetermined. There has been an increased focus recently on the impact of CYP2C19 ROF alleles in patients following a stroke. A recent meta-analysis of studies of patients prescribed clopidogrel for ischaemic stroke/TIA found CYP2C19 ROF carriers had an increased risk of recurrent stroke and MACE . CYP2C19 ROF alleles have also been associated with increased in-stent restenosis in patients with PAD . On the other hand, a genetic analysis of the large EUCLID trial ( n = 13,885) that recruited patients with symptomatic PAD reported no difference in cardiovascular outcomes between patient subgroups with different CYP2C19 genotypes . However, EUCLID excluded CYP2C19 PM patients from entry into the trial, limiting interpretation of these findings. Beyond CYP2C19 , there is growing evidence for a role of CES1 in clopidogrel PGx. CES1 catalyses clopidogrel to its inactive metabolite; the variant allele of rs71647871 (p.G143E, MAF ~1%) in CES1 has been associated with increased circulating clopidogrel active metabolite, reduced ex vivo platelet reactivity, and a nonsignificant trend towards decreased MACE . Moreover, the variant allele of CES1 rs2307240 (p.S75N, MAF ~5%) was recently associated with reduced subsequent MACE in clopidogrel-treated ACS patients . Both CPIC and DPWG (106) have produced clinical guidelines for clopidogrel-CYP2C19: CPIC focuses on ACS patients undergoing PCI, and DPWG on PCI for any indication, as well as stroke and TIA. The main recommendation is alternative antiplatelet therapy in those with CYP2C19 ROF alleles, although the DPWG guidance also permits doubling the dose of clopidogrel in CYP2C19 IMs. A multi-site real world implementation of CYP2C19 genotyping to guide antiplatelet stratification after PCI ( n = 1815) has been reported, and importantly demonstrated clinical benefit . On balance, taking all evidence together, we contend that clinical implementation of clopidogrel-CYP2C19 genotyping is justifiable and beneficial.
The oral vitamin K antagonist, warfarin, is indicated to prevent and treat venous thromboembolism (VTE) and to prevent thromboembolism in atrial fibrillation (AF) and following mechanical heart valve transplantation. Warfarin is a racemate that competitively inhibits VKORC1 within the vitamin K cycle, leading to hypofunctional clotting factors II, VII, IX and X. The extent of anticoagulation is measured by the international normalised ratio (INR) and, for most indications, the target therapeutic INR is 2.0–3.0 . Warfarin stable dose (WSD) requirements vary ~30-fold between patients , and its narrow therapeutic index underlies warfarin being the third most common drug to lead to hospitalisation due to ADRs(3). A 10% increase in time outside the therapeutic INR range (TTR) is associated with increased thromboembolic events and mortality , and supratherapeutic INRs increase the risk of major bleeding . Overall, ~60% of the observed variation in warfarin stable dose requirements can be explained, and several clinical factors including age, smoking and interacting drugs contribute . However, genetic factors explain the majority of this observed variation. Specifically, rs9923231 (-1639G > A) in VKORC1 accounts for 6–25% of observed variation, variation in CYP2C9 another ~15%, and CYP4F2*3 (rs2108622, p.V433M) 1–7% . -1639G > A lies within VKORC1′s promoter region; -1639A alters a transcription factor binding site reducing VKORC1 transcription , increasing sensitivity to warfarin and reducing WSD requirements . The MAF of -1639A in African, East Asian, South Asian and European populations is ~5%, ~90%, ~15% and ~40%, demonstrating allele reversal in East Asians. The variation in -1639A MAF perhaps underlies why -1639G > A explains only 6% of WSD variation in African-Americans but ~20–25% in Asian and Caucasian populations . CYP2C9 metabolises the more potent S-warfarin enantiomer and CYP2C9 ROF variants reduce WSD requirements. CYP2C9*2 (rs1799853, p.R144C) and *3 (rs1057910, p.I359L) are common ROF alleles in European populations with MAFs of ~12% and ~7%, respectively, and CYP2C9*3 is common in Asian populations (MAF 3–11%). However, CYP2C9*2 is infrequent in Asians (0–3%) and both are rare or infrequent in African populations (0–3.6% for *2 , and 0.3–2% for *3 ) . Interestingly, distinct warfarin-associated variants have been described in African ancestry patients (primarily Africa-Americans). CYP2C9*5 , *6 , *8 and *11 are ROF alleles with a collective frequency of ~20% in African ancestry individuals , but are rare in other ethnicities. All of these variants, except CYP2C9*8 , have been confirmed by meta-analysis to reduce warfarin dose requirements in black African patients . The non-coding variant, rs12777823, within the CYP2C cluster upstream of CYP2C18 , has also been associated with reduced WSD requirements, independent of CYP2C9 . CYP4F2 mediates removal of active (reduced) vitamin K from the vitamin K cycle. CYP4F2*3 is associated with lower hepatic CYP4F2 and increased WSD requirements in Asian and European ancestry patients, but not in those of African ancestry . Beyond VKORC1/CYP2C9/CYP4F2 , many other candidate genes have been investigated for their role in variable warfarin dose requirements, although findings have been mostly contradictory, and not supported by GWAS. Interestingly however, rs7856096 in the folate homeostasis gene, folylpolyglutamate synthase ( FPGS ), was identified in African-Americans by exome sequencing and replicated . It decreases FPGS transcription and is associated with reduced WSD requirements, although the causative mechanism remains unknown. Several RCTs have investigated the utility of genetic-guided warfarin dosing algorithms . The largest trial to date, the Genetic Informatics Trial (GIFT), recruited patients initiating warfarin for elective hip or knee surgery . GIFT compared 808 patients that received genotype-guided dosing (considering VKORC1 -1639G > A, CYP2C9*2 , CYP2C9*3 , CYP4F2*3 ) to 789 that received clinically-guided dosing. The primary clinical composite endpoint of major bleeding, INR ≥ 4, VTE or death occurred in 10.8% of patients in the genotyped arm versus 14.7% in the clinically-guided arm ( p = 0.02). The high-risk subgroup of patients (those with a difference of at least 1.0mg/day between clinically guided and genotype-predicted warfarin doses), which plausibly had a higher number of warfarin sensitive alleles, particularly benefitted from genotyping with a larger (7%) increase in TTR. The findings in GIFT were consistent with the EU-PACT RCT, but not with COAG Both CPIC and DPWG have developed warfarin PGx guidelines for commencing warfarin that recommend validated genetic algorithms (e.g., the EU-PACT loading algorithm or International Warfarin Pharmacogenetics Consortium (IWPC) algorithm ) or percent dose alterations. Importantly, point-of-care genotype-guided warfarin dosing has been shown to significantly improve TTR following implementation in real world anticoagulation clinics , and appears cost-effective . The majority of patients investigated have been of European ancestry.As described above, ethnicity influences the prevalence of warfarin risk alleles, particularly in those of Sub-Saharan African origin. Thus, ethnicity-specific algorithms have been developed, although prospective testing of these algorithms is currently limited . Moreover, self-reported ethnicity may not represent genetic ancestry well, especially for admixed individuals . Nonetheless, it has been recently shown that an individual’s genetic ancestry can be reasonably determined from clinically focused PGx SNP panels (122 and 243 SNPs) compared with genome-wide genotyping . This approach could stratify individual’s by genetic ancestry to facilitate implementation of ethnicity-specific warfarin algorithms. Although warfarin remains commonly prescribed, it is clear that the use of direct oral anticoagulants (DOACs) is increasing . DOACs are at least as safe and effective as warfarin in the prevention of stroke in non-valvular AF and monitoring is not required. However, their current high price restricts access to them in many healthcare systems, and concerns over long term adherence , as well as patient choice, mean they are not suitable for all patients. Interestingly, bleeding risk appears equivalent between DOACs and warfarin in patients who have no VKORC1 / CYP2C9 warfarin risk alleles or their anticoagulation centre-based TTR is ≥ 66% , and thus anticoagulant stratification based on VKORC1 / CYP2C9 genotypes has been posited . Moreover, no DOAC is indicated following mechanical heart valve surgery or in those with a creatinine clearance < 15mL/min; thus, further research into the utility of warfarin PGx in these specific settings is warranted.
Metoprolol is a racemate cardioselective β1-adrenoreceptor blocking agent used for indications including hypertension, heart failure, angina pectoris, arrhythmias and migraine prophylaxis. Around 70% of an oral dose undergoes metabolism by CYP2D6 to inactive metabolites . Plasma levels of metoprolol have been found to be significantly higher (around 4–5 ×) in CYP2D6 PMs compared with EMs . Metoprolol decreases cardiac output via negative chronotropic, as well as inotropic effects, thus reducing heart rate . Retrospective and prospective studies have demonstrated significantly greater reductions in heart rate in PMs vs. non-PMs and an increased risk of bradycardia . Although, there is observed heterogeneity in the incidence of bradycardia between studies, a recent meta-analysis identified that overall the prevalence of bradycardia is statistically significantly higher in PMs . Nevertheless, the clinical relevance of this finding is presently unclear, given that many of the instances of bradycardia have been asymptomatic (e.g., detected on an electrocardiogram performed as part of a study protocol) with comparatively little reported around symptomatic bradyarrhythmias. The DPWG guideline for metoprolol in CYP2D6 PM and IM patients suggest a dose reduction (to 25% and 50% of the standard dose, respectively) where a gradual reduction in heart rate is required (for example, in chronic heart failure) or symptomatic bradycardia occurs. As there may be increased conversion to inactive metabolites in UM patients blunting the efficacy of metoprolol, dose increases beyond the usual maximum dose or selection of an alternative β-blocker (indication dependent) may be required . While other β-blockers including carvedilol, nebivolol and propranolol are also metabolized by CYP2D6, the extent of their CYP2D6-dependent metabolism is less than for metoprolol . Moreover, atenolol is not metabolised, and bisoprolol undergoes balanced elimination with half excreted via the kidneys and the other half undergoing metabolism by CYP3A4 and CYP2D6; CYP2D6 does not appear to be an important predictor of bisoprolol exposure or function .
The xanthine oxidase inhibitor, allopurinol, is commonly prescribed in primary care for the prophylaxis of gout; it is also indicated in other hyperuricaemic conditions including tumour lysis syndrome. Although allopurinol is generally well tolerated, severe cutaneous adverse reactions (SCARs) represent a rare (0.1–0.4% of patients) but serious allopurinol ADR. The SCARs, Stevens–Johnson syndrome (SJS), toxic epidermal necrolysis (TEN), and drug reaction with eosinophilia and systemic symptoms (DRESS), have all been reported following allopurinol exposure . SCARs have overlapping clinical features including eosinophilia, hepatic/renal dysfunction and rash or skin detachment. SJS and TEN are considered to represent different severities along a continuum of the same disease process; SJS is diagnosed when ≤ 10% body surface area is affected, TEN when > 30% is affected, and an overlap syndrome occurs with 10–30% body surface area involvement . Allopurinol is the most common cause of SCARs in Europe . Furthermore, SCARs have high morbidity and mortality rates and the mortality specifically from allopurinol-associated SCARs has been reported to approach 25% . Studies have shown strong associations between allopurinol-related SCARs and carrying HLA-B*58:01 . In an interventional cohort study in Taiwan, genotyping HLA-B*58:01 and giving alternate treatment to HLA-B*58:01 positive patients, whilst HLA-B*58:01 negative patients received allopurinol, resulted in no SCAR cases, compared to an estimated seven cases based on the historical average ( p = 0.0026) . The allopurinol SCAR- HLA-B association is considered to be a type IV hypersensitivity reaction and is T-cell mediated ; typically, such reactions occur up to 3 months after drug initiation (most cases occurring within 8–9 weeks and are delayed in onset (as opposed to anaphylactic reactions with other drugs which are immediate) . Other identified clinical factors that increase the risk of allopurinol SCAR include older age, female sex, chronic kidney disease and higher starting doses, which all plausibly increase systemic exposure . Following oral administration, allopurinol is almost completely metabolised into its active metabolite, oxypurinol, which has an elimination half-life of 15 h, compared to 1–2 h for allopurinol . Regarding immunopathogenesis, both allopurinol and oxypurinol have been shown to directly bind HLA-B*58:01 and Arg97 of HLA-B*58:01 is a plausible binding site for oxypurinol via hydrogen bond formation . Oxypurinol-sensitive T-cell lines can have mixed populations of CD4+ and CD8+ T-cells. Moreover, in vitro T-cell activation assays after culturing peripheral blood mononuclear cells from patients with previous allopurinol-associated SCARs demonstrated that oxypurinol increased granulysin in a concentration- and time-dependent manner, but allopurinol and febuxostat did not, indicative of oxypurinol-induced T-cell activation . Sequencing the T-cell receptors demonstrated that clonotype-specific T-cells that secrete granulysin in response to oxypurinol likely participate in the pathogenesis . Overall, these observations are consistent with the p-i hypothesis (direct pharmacological interaction between a drug/metabolite and HLA molecules) whereby allopurinol-associated SCARs are the result of MHC Class I-driven activation of oxypurinol-specific T-cell clones through an antigen-processing independent route. The frequency of HLA-B*58:01 varies with ethnicity, but occurs at a higher prevalence in East Asian populations . HLA-B expression is co-dominant and individuals need only carry one copy of HLA-B*58:01 to be at increased risk of SCARs. Accordingly, genotyping results are usually reported as HLA-B*58:01 positive, where one of more copies of the allele are detected, or negative where the allele is not detected. This association is now mentioned within the European drug label and CPIC guidelines for allopurinol advise against its use in known HLA-B*58:01 carriers More generally, associations between specific HLA alleles and serious ADRs due to at least 24 different drugs have now been reported , highlighting the importance of the HLA locus to immune-mediated type IV hypersensitivity reactions. Like, allopurinol, several of these associations are very strong (e.g., odds ratios in the tens-thousands) and related to drug-induced skin injury, such as between carbamazepine- HLA-B*15:02 and SJS-TEN, and carbamazepine- HLA-A*31:01 that is associated with both specific SCARs (DRESS and SJS-TEN) and more mild but common (up to 10% of patients ) maculopapular exanthema . HLA alleles have also been associated with drug-induced liver injury, such as between flucloxacillin and HLA-B*57:01 . Of note, HLA-B*57:01 is also associated with abacavir hypersensitivity syndrome . As previously mentioned, HLA-DRB1*11:01 has been significantly associated with statin-related anti-HMGCR myopathy . These associations provide mechanistic insight and can aid in preventing or diagnosing a serious ADR.
The current literature demonstrates that many primary healthcare providers including physicians and pharmacists are hopeful about the role of PGx in enhancing the care of their patients but often highlight that there are still multiple barriers impeding translation into daily practice . Potential pathways of how GPs may encounter PGx tests are shown in . Recent studies have identified that barriers to PGx implementation specifically within primary care include a perceived lack of evidence for clinical utility, unclear cost effectiveness and reimbursement strategies, how to educate the primary care workforce regarding PGx, unclear roles and responsibilities particularly between general practitioners and pharmacists, the need for informatics to support PGx-informed clinical prescribing decisions, and concerns over the principles of data sharing as well as other ethical, legal and social implications (ELSI) surrounding PGx . Efforts to overcome these barriers and facilitate PGx translation are ongoing. Within Europe, the Ubiquitous Pharmacogenomics Consortium (U-PGx) is undertaking a large, international implementation project (PREemptive Pharmacogenomic testing for prevention of Adverse drug REactions (PREPARE)) within both primary and secondary care settings . This study aims to recruit around 7000 patients to determine the clinical utility and cost effectiveness of testing a panel of 44 PGx variants in 12 genes relevant to 42 drugs for which a DPWG PGx guideline has been developed; the primary focus of PREPARE is to investigate whether PGx implementation reduces ADRs within the first 12 weeks of starting one of these 42 drugs, compared to standard care. Similar initiatives adopting panel-based testing are ongoing around the world, including the Pharmacogenomic Resource for Enhanced Decisions in Care and Treatment program (PREDICT) in the US, which has pre-emptively tested over 10,000 patients to date . Moreover, in the UK, the NHS Genomic Medicine Service (GMS) outlined its plans to have genomic medicine fully integrated into routine practice by 2025 . Despite the cost of genetic tests decreasing, cost-effectiveness remains a concern for PGx testing. While some drug-gene pairs, such as abacavir- HLA-B*57:01 and allopurinol- HLA-B*58:01 , have demonstrated cost-effectiveness compared with no genetic testing, others such as factor V Leiden testing prior to oral contraception, have shown mixed and inconclusive results . Panel testing is the most common form of PGx testing currently being utilised in large scale studies such as PREPARE, and is likely more cost effective than single gene testing . Although panel testing does not offer the same in-depth information for a gene compared to whole genome sequencing, its upfront costs are less, test turnaround times are shorter, and interpretation of gene regions is less complicated, which arguably makes it a better suited technology for larger scale implementation in the near to medium term. Patient focus groups have shown a preference for a familiar healthcare provider to be involved with delivering PGx services . In addition, the uptake of direct-to-consumer testing is increasing and so there is a need to upskill the primary care workforce at large to improve its genetic literacy . It is clear that wider implementation of PGx into the primary care sector will be soon upon us, and it is essential that all areas are able to access the resources to become more comfortable with PGx and overcome the aforementioned barriers. One critical area for implementation is the development of appropriate clinical decision support systems (CDSS) that facilitate use of PGx information at the point of prescribing and, ideally, integrate this information with other factors routinely considered in the prescribing decision-making process, such as co-medications and comorbidities. In the UK at least, the widespread and early adoption of electronic medical patient records in primary care, which are provided by only a handful of service providers (e.g., EMIS and SystemOne), provide a pre-existing technological infrastructure to build PGx CDSS into. In contrast, the varying uptake of electronic records in secondary care and the greater heterogeneity in providers poses additional barriers to implementing PGx in hospital settings. As the implementation infrastructure for PGx is gradually erected, it will be important to transition to healthcare learning systems where clinical and research activities are more clearly linked. Thus, ideally, novel clinically relevant and validated discoveries can then be implemented in a shorter time frame than currently, and the effects in the real-world of new interventions observed, measured and iteratively fed back to the research arm to inform future enquiry. A key component of an effective healthcare learning system will be merging presently disparate data sources together and putting systems in place to efficiently collect future clinical data as it is generated to enable big data analyses, and there is little reason why PGx cannot be at the forefront of such endeavours. For example, combining primary care records with array genetic data at scale within UK Biobank has enabled recent PGx analysis investigating interactions between 200 drugs and nine genes in 200,000 subjects, leading to confirmation of several established drug-gene pairs, as well as providing genetic evidence of more novel associations such as between citalopram and reduced incidence of herpes zoster in CYP2C19 IMs . It should be noted that, to date, the majority of PGx associations involve common variants, including the drug-gene variant associations highlighted in this review. Rare variants (i.e., those with a minor allele frequency < 1%) are often missed in small clinical studies or GWAS analyses, but some may have large effects on outcomes . Case studies of rare and serious ADRs have highlighted the clinical impact of rare variation in pharmacogenes. For example, a recent case report detailed development of reversible encephalopathy and coma after a paediatric patient received a single dose of ivermectin, attributable to compound heterozygosity from two nonsense mutations in ABCB1 , which encodes P-glycoprotein . Furthermore, the increasing use of next generation sequencing (NGS) technologies in national and other large genomic medicine projects heralds a new era for investigating rare variation at scale. It is currently estimated that, overall, 10–40% of genetic functional variation in pharmacogenes is attributable to rare ght to account for 9% and 39% of the functional variation in SLCO1B1 and ABCC1 , respectively variation, although it varies considerably between genes . For example, rare variation is thou Machine learning and other advanced methodologies will likely be needed to parse this rich sequence data, particularly when combined with other data types (e.g., clinical and other omics datasets), to aid functional interpretation and variant discovery, refine genotype-to-phenotype predictions, stratify patient groups, and predict drug response .
The majority of prescribing happens in primary care. Owing to interindividual variation, some patients experience reduced effectiveness and others ADRs. Extensive research into PGx has identified and replicated multiple drug-gene pairs, mainly associated with ADRs, with several of the associated drugs commonly prescribed in primary care. The ongoing development of PGx guidelines by CPIC, DPWG and others offers a standardised approach to translating drug-gene associations into actionable prescribing recommendations based on the collective evidence available. Nevertheless, several barriers still exist to wider adoption of PGx into primary care, although these are gradually being addressed. Moving forward, a transition towards healthcare learning systems that harness big data and advanced analytical techniques with implementation of clinically relevant findings is anticipated to advance patient care. One area for focus is the surge in multimorbidity and related polypharmacy that multiple primary care and wider healthcare systems are facing. Although further research is required, it is envisaged that PGx will play a meaningful role in optimising medicines and managing polypharmacy . Overall, it is hoped that implementation of PGx in primary care will be a major activity over the next five years, with ongoing PGx discoveries further boosting the case for implementation.
|
Utilizing the subtractive proteomics approach to design ensemble vaccine against | 0e0e8934-2ca1-4ca8-aff5-751673db40bf | 11801629 | Biochemistry[mh] | Candida lusitaniae also known as Clavispora lusitaniae by the teleform name , is a human pathogen which belongs to the yeast specie of genus candida which was identified in 1979. Human infections have been linked to this emerging pathogen Candida lusitaniae which is an opportunistic haploid yeast and most commonly associated with immunocompromised patients who frequently have multiple medical conditions . C . lusitaniae can cause oral thrush, vaginitis (superficial infections), endocarditis, endophthalmitis (deep-seated infections of tissues), single organ fungemia (pulmonary fungemia), and blood stream infections, depending on the immune system of the hosts . Although, C . lusitaniae is a rare pathogen, its increased prevalence in hospitalized patients, patients receiving prolonged antibiotic therapy, patients undergoing chemotherapies or bone marrow transplants, and the ones with underlying malignancies, has sparked special interest in it as a nosocomial pathogen . The use of catheters increases the risk of developing candidiasis because they are a major reservoir for yeasts including, C . lusitaniae which can cause fungemia . Different reported studies have shown, how C . lusitaniae can create biofilms that lead to endogenous infections . While, similar to other agents that cause candidiasis, it has also been documented that this species can be transmitted through interpersonal contact in the intensive care unit (ICU) . C . lusitaniae is accountable for roughly 1.7% of all genitourinary candidiasis cases in ambulatory patients and 19.3% of fungemia cases in cancer patients . As far as, the treatment methods of C . lusitaniae oriented infections are concerned, this pathogen was found to respond well to standard antifungal treatments. But, it has garnered attention due to certain isolates’ resistance to fluconazole, 5-fluorocytosine, and amphotericin B like anti-fungal drugs . Susceptibility testing, however, might be necessary to determine the best course of action because certain strains of C . lusitaniae are resistant to this medication. Azoles, which prevent the synthesis of ergosterol, a part of the fungal cell membrane, include fluconazole, itraconazole, voriconazole, and posaconazole. However, by changing the target enzyme or the drug efflux pumps, certain strains of C . lusitaniae may become resistant to these drugs, particularly fluconazole. Eschinocandins, which prevent the synthesis of beta-D-glucan which is a part of the fungal cell wall, including caspofungin, micafungin, and anidulafungin . Clinical reports have indicated that the combination of amphotericin B and fluconazole antifungals can result in the production of multi-resistant isolates of C . lusitaniae . Therefore, it is suggested that treatment through this kind of combined drugs should be avoided as it may lead to various diseases including deep-seated infections in immunocompromised patients . Hence, there is a need to adopt unconventional approaches to treat Candida lusitaniae infections. Therefore, it’s critical to keep an eye on the yeast’s susceptibility to various antifungal medications and to move towards different therapeutic strategies like vaccination in such scenarios. In this aspect, the subtractive proteomics is a technique that mines a pathogen’s proteome for possible targets for vaccine candidates by using computational methods. The objective is to subtract the proteins that are similar or common by comparing the proteomes of the pathogen, host, and closely related organisms. In this way, the pathogen-specific or pathogen-essential proteins can be chosen as potential candidates for vaccine development. The challenges of emerging infections, complex diseases, and antibiotic resistance can be addressed with the aid of this subtractive proteomics approach. In the current study, the nuclear genome sequence of Candida lusitaniae comprising of 12.11Mbp having 6153 protein coding genes and five pseudogenes, was incorporated . While, by the employment of computational algorithms and subtractive proteomics methods, the most appropriate proteins were chosen from this massive collection of data in order to create effective vaccine against Candida lusitaniae . In-silico design of the vaccine exploit the safety and thermo-dynamic stability of the construct through the verification of certain necessary parameters like antigenicity, allergenicity and physiochemical properties. In this investigation, outer membrane proteins that are thought to be extremely virulent, antigenic, and vital to interactions between hosts and pathogen, were looked into . Afterwards, the selected proteins were utilized to predict Helper T lymphocyte (HTL), B-cell, and cytotoxic T lymphocyte (CTL) epitopes. With the aim to cover the whole proteome of Candida lusitaniae , the vaccine was designed by using the most suitable epitopes for immunization. Additionally, the intended vaccine candidates were immunologically assessed by employing molecular docking and immune simulations which indicated the significance and precision of the designed vaccines in terms of inducing the immune response. This study will provide direction to design real-time vaccine against this pathogen which could aid in combating nosocomial infections caused by Candida lusitaniae . Nevertheless, additional experimental validations are necessary to validate the safety and effectiveness of the suggested vaccine candidates.
Retrieval of Candida lusitaniae proteome and subtractive proteomics The reference proteome of Candida lusitaniae (Uniprot ID: UP000007703) was retrieved from UniProt ( https://www.uniprot.org/ ) which contained 5932 proteins . This whole proteome is also given in the . Subtractive proteomics approach was followed to identify the proteins that can be incorporated for vaccine designing against Candida lusitaniae . Initially, the sub-cellular localization was analyzed to screen out the extra-cellular proteins by using CELLO ( http://cello.life.nctu.edu.tw/ ) which is a well-known tool to predict the sub-cellular localization of the given proteins. This tool works on the basis of support vector machine (SVM) algorithm . Afterwards, the BLASTp tool ( https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins ) was incorporated in order to identify the homologous and non-homologous proteins between the pathogen and host ( Homo sapiens ) by using the non-redundant protein sequences (nr) database. It was done to avoid autoimmune response, thus, only non-homologous proteins were selected. The non-homologous proteins were further analyzed for paralogs prediction which was achieved by performing the phylogenetic analysis by utilizing MEGA11 . Prioritization of the essential proteins The proteins that were essential for the survival and pathogenicity of the Candida lusitaniae were prioritized. These proteins were potent candidates for the vaccine development. As mentioned previously, the cellular localization of the essential proteins were also analyzed because according to the studies it has been reported that extracellular proteins are highly suitable for the development of vaccines since, they play a crucial role in host-pathogen interaction as well as pathogenicity. The antigenicity of the selected proteins that were essential as well as extracellular was also identified with the help of VaxiJen tool ( https://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html ) . This particular analysis was done by selecting fungi as target organism while, the threshold value was set to 0.5. The shortlisted proteins were further examined by AlgPred ( https://webs.iiitd.edu.in/raghava/algpred/submission.html ) to estimate the allergenicity status . Resultantly, the non-allergen proteins but with the high antigenicity scores were selected for further analyses. Mapping of immunogenic peptides in the target proteins The selected non-allergen proteins with high antigenicity were further analyzed for immunogenic peptides. In this regard, the cytotoxic T-lymphocytes, Helper T-lymphocytes and B-cell epitopes were predicted. The NetCTL 1.2 ( https://services.healthtech.dtu.dk/services/NetCTL-1.2/ ) was employed in order to predict the cytotoxic T-lymphocytes epitopes . This server predicted the CTL epitopes on the basis of some of the key values including, transportation efficiency, TAP (Transport Associated with Antigen Processing), MHC I binding peptides prediction, and proteasomal C terminal cleavage. The threshold value in this analysis was set to 0.75. Whereas, the affinity score of each peptide was determined by its IC50 value. Peptides having IC50 values less than 50 nM possessed the high binding affinity while, those with IC50 values less than 500 nM were the intermediates and, the ones with IC50 values less than 5000 nM exhibited weak epitope binding affinity. The affinity scores and percentile rankings of the peptides were inversely related, showing that higher the binding affinity, the lower would be the percentile rank, and vice versa. Furthermore, the prediction of Helper T-lymphocytes was achieved by employing an online server IEDB ( https://www.iedb.org/ ) by using the seven sets of human reference alleles HLAs . In this context, the allergenicity of each MHC-II epitope was determined by AllerTOP v.2.0 ( https://www.ddg-pharmfac.net/AllerTOP/ ) . Moreover, the antigenicity of each epitope was assessed by VaxiJen tool ( https://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html ). The epitopes which were non-allergen and possessed high antigenicity were chosen for the development of vaccine. Additionally, IFN-epitope server ( https://webs.iiitd.edu.in/raghava/ifnepitope/predict.php ) was incorporated to analyze the interferon-gamma triggering MHC-II epitopes . This server also utilizes the support vector machine (SVM) score for the prediction of interferon inducing properties of each peptide. The SVM score serves as a measure of the likelihood of a peptide to induce an IFN-gamma response. Lastly, the B-cell epitopes were predicted by the ABCpred ( https://webs.iiitd.edu.in/raghava/abcpred/ABC_submission.html ). This server had >65% accuracy, 0.49 sensitivity and 0.75 specificity. Construction of multi epitopes subunit vaccine The B-cells, Cytotoxic T-cells and Helper T-lymphocyte epitopes having high binding affinity and antigenicity with the non-allergen nature were employed for designing the final putative vaccine construct. The selected epitopes were linked with the help of different linkers. These linkers were preferred on the basis of data provided by already reported studies . Specific linkers were used for each epitope. The AAY linker was employed for CTL epitopes, GPGPG for HTL and KK linkers for the B-cell epitopes to construct the putative vaccine. These linkers were critical for allowing and maintaining the epitope presentation and separation and, inhibiting the folding in order to promote a successful immune response . Furthermore, the use of an adjuvant improves the immunogenic potential of MEVC peptide vaccine designs. Following an EAAAK linker, all of the MEVC constructs were treated to the addition of a non-toxic adjuvant named human beta defensin-2 (hBD-2) . The hBD-2 possesses the ability to self-produce at the levels, sufficient to elicit a powerful immune response against the attached antigen . The adjuvant was linked to the N-terminal of the final candidate vaccine construct using the EAAK linker . This approach was used in the development of a vaccine targeting particular proteins and a whole proteome vaccine, which is also constructed in this study . These linkers are very important in assisting the prevention of neo-epitopes, ultimately providing structural stability to the vaccine constructs hence, enhancing the immunogenicity . Prediction of physicochemical properties To analyze the physiochemical properties of the constructed putative vaccine, an online tool ProtParam ( https://web.expasy.org/protparam/ ) was incorporated This tool predicts different physiochemical attributes including instability index, theoretical PI, in-vivo and in-vitro half-life, amino acid composition, and grand average of hydropathy (GRAVY). Prediction of 2D and 3D structure of the designed vaccines Secondary structure of vaccine construct was predicted through an online server SOPMA ( https://npsa.lyon.inserm.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_seccons.html ). SOPMA(Self-Optimized Prediction method with alignment) is used to predict the structure and topology of membrane proteins by means of learning machine models, which are trained on the available data of protein sequences’ features, especially the amino acid composition and hydrophobicity profiles . Thereby, the model is iteratively refined through validation and optimization in order to adjust its original prediction ability for accurately predicting both the transmembrane segments and their orientation in new protein sequences . Furthermore, the tertiary structure of the putative vaccines was predicted by an online server trRosetta ( https://yanglab.nankai.edu.cn/trRosetta/ ) . It is a server for automated comparative modeling of three-dimensional (3D) protein structures. The trRosetta is an algorithm for fast and accurate protein structure prediction. Afterwards, the predicted protein structure were refined by using an online server Galaxy Refine ( https://galaxy.seoklab.org/cgi-bin/submit.cgi?type=REFINE ) . This process was essential for enhancing the accuracy of the predicted structure . For this particular analysis, the protein data bank ( . pdb) file of the predicted model was subjected as input file for further processing. During this refinement, the poor rotamers are removed which enhance the structural quality of the given protein model. Discontinuous B cell epitopes The ElliPro server (ElliPro: Antibody Epitope Prediction (iedb.org)) was used to perform the discontinuous B-cell epitope prediction of the 3D structure of the whole proteome vaccine construct, with default parameters (0.5 as the threshold for the Protrusion Index and a 6 Å threshold for the distance) . ElliPro predicted discontinuous epitopes based on surface exposure and spatial clustering of residues in the 3D structure, each residue being assigned a score of the Protrusion Index . The epitopes according to their PI values, higher values indicating higher possibilities of being immune-recognized, were considered for further analysis . Disulfide engineering of the vaccine construct It was essential to increase the improved protein model’s stability before moving on to the next step. Disulfide bonds are covalent interactions that confirm exact geometric conformations, simulating the stabilizing molecular interaction and contributing significantly to the stability of protein models . A unique method for introducing disulfide links into the target protein structure is disulfide engineering. In order to carry out disulfide engineering, the final vaccine construct’s refined model was processed through the Disulfide by Design 2 (Disulfide by Design) ( http://cptweb.cpt.wayne.edu/DbD2/ ). The server processed input data by analyzing protein sequences and structures . The design algorithm proposed optimal bonds based on stability and user constraints. Validation ensured bonds do not disrupt protein structure, with optional simulations for dynamic stability . Outputs included predicted bonds, design proposals, and visualizations. The post-processing provided 3D visualization tools and experimental validation . In order to find residue pairs that could be utilized for disulfide engineering, the revised protein model was first uploaded and ran. Using the Disulfide by Design 2.0 server’s construct mutate function, a total of three pairs of residues were chosen to be mutated with cysteine residues. TLR2-vaccine docking The designed vaccine construct was docked with the TLR2 (Human Toll-like Receptor 2). As TLR2 receptor recognizes a wide variety of pathogen-associated molecular patterns, TLR2 is used as a receptor in docking studies . It is capable of binding with diverse antigens and started signaling cascades that go on to activate the innate immune response, leading to cytokines and chemokine production critical in promoting effective adaptive immunity . TLR2 forms heterodimers with other members of the family of Toll-like receptors, such as TLR1 and TLR6, which help in the broader range of recognition of antigens and make it a very good candidate for an evaluation of how effectively a vaccine construct will be able to elicit immune stimulation . Docking studies including TLR2 have applications in estimating potential effectiveness and immunogenicity of vaccine candidates . For this purpose, the HADDOCK 2.4 server ( https://wenmr.science.uu.nl/haddock2.4/ ) was employed . HADDOCK makes use of biochemical and/or biophysical interaction data such as chemical shift perturbation data, resulting from NMR titration experiments or mutagenesis data to perform the docking analysis. This information is introduced as ambiguous interaction restraints (AIRs) to perform the docking process. However, before the protein-protein docking, the sequence of TLR2 was retrieved from UniProt database (accession ID: O60603) and afterwards, its structure was modelled by incorporating trRosetta ( https://yanglab.qd.sdu.edu.cn/trRosetta/ ). The TLR2 structure was prepared prior to docking, by removing the water molecules, heteroatoms, and other atoms. Molecular dynamics (MD) simulations After the docking process, molecular dynamics (MD) simulations of the docked complexes were performed . MD simulations are a sophisticated computational technique providing insights into the dynamic behavior of atoms within molecules. Researchers use MD simulations to observe and analyze atomic movements over time, applying physical laws in a virtual environment to explore complex molecular dynamics . These simulations are crucial for studying the interactions in the docked complexes, offering a deeper understanding of binding mechanisms, complex dynamics, and the influence of environmental factors. Moreover, by exploring the effects of physical parameters like temperature and pressure, the comprehensive understanding of how external factors impact molecular behavior in a computational setting are also acquired. Additionally, MD simulations assess the structural stability of complexes under varying conditions. In this study, MD simulations were performed by using an online server iMODS ( https://imods.iqfr.csic.es/ ) . The iMODS is a versatile toolkit that enables us to conduct Normal Mode analysis (NMA) of the internal coordinates of the docked structures. It allows for vibrational analysis, motion animations, morphing trajectories, and Monte-Carlo simulations at various scales of resolution. It also enables to gain insights into the movements and interactions of atoms over time, providing valuable information about the dynamics and behavior of biomolecular systems. To better understand all such fluctuations of the docked structures (vaccine candidates-TLR2) in the current study, the complexes were subjected to iMODS by keeping the default parameters to perform the simulations. Codon optimization and in-silico cloning The Java codon adaptation (JCat) tool ( https://www.jcat.de/ ) was utilized to clone the multi-epitope subunit vaccine in a suitable vector. This analysis was done for codon optimization and reverse translation to get the nucleotide sequence which was used for cloning . It was also employed to ensure, whether the vaccine construct was highly expressed after cloning in a particular vector. In this tool, three parameters were selected including, bacterial ribosome binding sites, Rho-independent transcription termination, and the restriction enzyme cleavage sites. Moreover, the Jcat calculated the GC content and CAI score of the built vaccine to optimize the reversely transcribed vaccine construct for bacterial expression. The potential vaccine sequence was manually cloned using BspAPI and PsaAI restriction enzymes, and the translated sequence of nucleotides from jcat was then cloned onto the pet28a+ plasmid by employing the SnapGene ( https://www.snapgene.com/ ). Immune simulations In order to create an immunogenic profile of the designed putative vaccine, the immune simulations were performed on C-ImmSim ( https://kraken.iac.rm.cnr.it/C-IMMSIM/ ) . C-ImmSim is a dynamic agent-based simulator for evaluating the body’s immunological responses to antigens. C-ImmSim predicts immunological interactions and epitopes using an agent-based method and the particular scoring matrix, PSSM. The generation of antibodies, interferon, and cytokines was assessed after submitting the prepared vaccine to the web server, by following the default simulation parameters. The overall methodology followed in this study can be assessed more conveniently by observing the schematic workflow represented in the . Moreover, the details of all of the tools and servers incorporated in this work are also included in the .
Candida lusitaniae proteome and subtractive proteomics The reference proteome of Candida lusitaniae (Uniprot ID: UP000007703) was retrieved from UniProt ( https://www.uniprot.org/ ) which contained 5932 proteins . This whole proteome is also given in the . Subtractive proteomics approach was followed to identify the proteins that can be incorporated for vaccine designing against Candida lusitaniae . Initially, the sub-cellular localization was analyzed to screen out the extra-cellular proteins by using CELLO ( http://cello.life.nctu.edu.tw/ ) which is a well-known tool to predict the sub-cellular localization of the given proteins. This tool works on the basis of support vector machine (SVM) algorithm . Afterwards, the BLASTp tool ( https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins ) was incorporated in order to identify the homologous and non-homologous proteins between the pathogen and host ( Homo sapiens ) by using the non-redundant protein sequences (nr) database. It was done to avoid autoimmune response, thus, only non-homologous proteins were selected. The non-homologous proteins were further analyzed for paralogs prediction which was achieved by performing the phylogenetic analysis by utilizing MEGA11 .
The proteins that were essential for the survival and pathogenicity of the Candida lusitaniae were prioritized. These proteins were potent candidates for the vaccine development. As mentioned previously, the cellular localization of the essential proteins were also analyzed because according to the studies it has been reported that extracellular proteins are highly suitable for the development of vaccines since, they play a crucial role in host-pathogen interaction as well as pathogenicity. The antigenicity of the selected proteins that were essential as well as extracellular was also identified with the help of VaxiJen tool ( https://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html ) . This particular analysis was done by selecting fungi as target organism while, the threshold value was set to 0.5. The shortlisted proteins were further examined by AlgPred ( https://webs.iiitd.edu.in/raghava/algpred/submission.html ) to estimate the allergenicity status . Resultantly, the non-allergen proteins but with the high antigenicity scores were selected for further analyses.
The selected non-allergen proteins with high antigenicity were further analyzed for immunogenic peptides. In this regard, the cytotoxic T-lymphocytes, Helper T-lymphocytes and B-cell epitopes were predicted. The NetCTL 1.2 ( https://services.healthtech.dtu.dk/services/NetCTL-1.2/ ) was employed in order to predict the cytotoxic T-lymphocytes epitopes . This server predicted the CTL epitopes on the basis of some of the key values including, transportation efficiency, TAP (Transport Associated with Antigen Processing), MHC I binding peptides prediction, and proteasomal C terminal cleavage. The threshold value in this analysis was set to 0.75. Whereas, the affinity score of each peptide was determined by its IC50 value. Peptides having IC50 values less than 50 nM possessed the high binding affinity while, those with IC50 values less than 500 nM were the intermediates and, the ones with IC50 values less than 5000 nM exhibited weak epitope binding affinity. The affinity scores and percentile rankings of the peptides were inversely related, showing that higher the binding affinity, the lower would be the percentile rank, and vice versa. Furthermore, the prediction of Helper T-lymphocytes was achieved by employing an online server IEDB ( https://www.iedb.org/ ) by using the seven sets of human reference alleles HLAs . In this context, the allergenicity of each MHC-II epitope was determined by AllerTOP v.2.0 ( https://www.ddg-pharmfac.net/AllerTOP/ ) . Moreover, the antigenicity of each epitope was assessed by VaxiJen tool ( https://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html ). The epitopes which were non-allergen and possessed high antigenicity were chosen for the development of vaccine. Additionally, IFN-epitope server ( https://webs.iiitd.edu.in/raghava/ifnepitope/predict.php ) was incorporated to analyze the interferon-gamma triggering MHC-II epitopes . This server also utilizes the support vector machine (SVM) score for the prediction of interferon inducing properties of each peptide. The SVM score serves as a measure of the likelihood of a peptide to induce an IFN-gamma response. Lastly, the B-cell epitopes were predicted by the ABCpred ( https://webs.iiitd.edu.in/raghava/abcpred/ABC_submission.html ). This server had >65% accuracy, 0.49 sensitivity and 0.75 specificity.
The B-cells, Cytotoxic T-cells and Helper T-lymphocyte epitopes having high binding affinity and antigenicity with the non-allergen nature were employed for designing the final putative vaccine construct. The selected epitopes were linked with the help of different linkers. These linkers were preferred on the basis of data provided by already reported studies . Specific linkers were used for each epitope. The AAY linker was employed for CTL epitopes, GPGPG for HTL and KK linkers for the B-cell epitopes to construct the putative vaccine. These linkers were critical for allowing and maintaining the epitope presentation and separation and, inhibiting the folding in order to promote a successful immune response . Furthermore, the use of an adjuvant improves the immunogenic potential of MEVC peptide vaccine designs. Following an EAAAK linker, all of the MEVC constructs were treated to the addition of a non-toxic adjuvant named human beta defensin-2 (hBD-2) . The hBD-2 possesses the ability to self-produce at the levels, sufficient to elicit a powerful immune response against the attached antigen . The adjuvant was linked to the N-terminal of the final candidate vaccine construct using the EAAK linker . This approach was used in the development of a vaccine targeting particular proteins and a whole proteome vaccine, which is also constructed in this study . These linkers are very important in assisting the prevention of neo-epitopes, ultimately providing structural stability to the vaccine constructs hence, enhancing the immunogenicity .
To analyze the physiochemical properties of the constructed putative vaccine, an online tool ProtParam ( https://web.expasy.org/protparam/ ) was incorporated This tool predicts different physiochemical attributes including instability index, theoretical PI, in-vivo and in-vitro half-life, amino acid composition, and grand average of hydropathy (GRAVY).
Secondary structure of vaccine construct was predicted through an online server SOPMA ( https://npsa.lyon.inserm.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_seccons.html ). SOPMA(Self-Optimized Prediction method with alignment) is used to predict the structure and topology of membrane proteins by means of learning machine models, which are trained on the available data of protein sequences’ features, especially the amino acid composition and hydrophobicity profiles . Thereby, the model is iteratively refined through validation and optimization in order to adjust its original prediction ability for accurately predicting both the transmembrane segments and their orientation in new protein sequences . Furthermore, the tertiary structure of the putative vaccines was predicted by an online server trRosetta ( https://yanglab.nankai.edu.cn/trRosetta/ ) . It is a server for automated comparative modeling of three-dimensional (3D) protein structures. The trRosetta is an algorithm for fast and accurate protein structure prediction. Afterwards, the predicted protein structure were refined by using an online server Galaxy Refine ( https://galaxy.seoklab.org/cgi-bin/submit.cgi?type=REFINE ) . This process was essential for enhancing the accuracy of the predicted structure . For this particular analysis, the protein data bank ( . pdb) file of the predicted model was subjected as input file for further processing. During this refinement, the poor rotamers are removed which enhance the structural quality of the given protein model.
The ElliPro server (ElliPro: Antibody Epitope Prediction (iedb.org)) was used to perform the discontinuous B-cell epitope prediction of the 3D structure of the whole proteome vaccine construct, with default parameters (0.5 as the threshold for the Protrusion Index and a 6 Å threshold for the distance) . ElliPro predicted discontinuous epitopes based on surface exposure and spatial clustering of residues in the 3D structure, each residue being assigned a score of the Protrusion Index . The epitopes according to their PI values, higher values indicating higher possibilities of being immune-recognized, were considered for further analysis .
It was essential to increase the improved protein model’s stability before moving on to the next step. Disulfide bonds are covalent interactions that confirm exact geometric conformations, simulating the stabilizing molecular interaction and contributing significantly to the stability of protein models . A unique method for introducing disulfide links into the target protein structure is disulfide engineering. In order to carry out disulfide engineering, the final vaccine construct’s refined model was processed through the Disulfide by Design 2 (Disulfide by Design) ( http://cptweb.cpt.wayne.edu/DbD2/ ). The server processed input data by analyzing protein sequences and structures . The design algorithm proposed optimal bonds based on stability and user constraints. Validation ensured bonds do not disrupt protein structure, with optional simulations for dynamic stability . Outputs included predicted bonds, design proposals, and visualizations. The post-processing provided 3D visualization tools and experimental validation . In order to find residue pairs that could be utilized for disulfide engineering, the revised protein model was first uploaded and ran. Using the Disulfide by Design 2.0 server’s construct mutate function, a total of three pairs of residues were chosen to be mutated with cysteine residues.
The designed vaccine construct was docked with the TLR2 (Human Toll-like Receptor 2). As TLR2 receptor recognizes a wide variety of pathogen-associated molecular patterns, TLR2 is used as a receptor in docking studies . It is capable of binding with diverse antigens and started signaling cascades that go on to activate the innate immune response, leading to cytokines and chemokine production critical in promoting effective adaptive immunity . TLR2 forms heterodimers with other members of the family of Toll-like receptors, such as TLR1 and TLR6, which help in the broader range of recognition of antigens and make it a very good candidate for an evaluation of how effectively a vaccine construct will be able to elicit immune stimulation . Docking studies including TLR2 have applications in estimating potential effectiveness and immunogenicity of vaccine candidates . For this purpose, the HADDOCK 2.4 server ( https://wenmr.science.uu.nl/haddock2.4/ ) was employed . HADDOCK makes use of biochemical and/or biophysical interaction data such as chemical shift perturbation data, resulting from NMR titration experiments or mutagenesis data to perform the docking analysis. This information is introduced as ambiguous interaction restraints (AIRs) to perform the docking process. However, before the protein-protein docking, the sequence of TLR2 was retrieved from UniProt database (accession ID: O60603) and afterwards, its structure was modelled by incorporating trRosetta ( https://yanglab.qd.sdu.edu.cn/trRosetta/ ). The TLR2 structure was prepared prior to docking, by removing the water molecules, heteroatoms, and other atoms.
After the docking process, molecular dynamics (MD) simulations of the docked complexes were performed . MD simulations are a sophisticated computational technique providing insights into the dynamic behavior of atoms within molecules. Researchers use MD simulations to observe and analyze atomic movements over time, applying physical laws in a virtual environment to explore complex molecular dynamics . These simulations are crucial for studying the interactions in the docked complexes, offering a deeper understanding of binding mechanisms, complex dynamics, and the influence of environmental factors. Moreover, by exploring the effects of physical parameters like temperature and pressure, the comprehensive understanding of how external factors impact molecular behavior in a computational setting are also acquired. Additionally, MD simulations assess the structural stability of complexes under varying conditions. In this study, MD simulations were performed by using an online server iMODS ( https://imods.iqfr.csic.es/ ) . The iMODS is a versatile toolkit that enables us to conduct Normal Mode analysis (NMA) of the internal coordinates of the docked structures. It allows for vibrational analysis, motion animations, morphing trajectories, and Monte-Carlo simulations at various scales of resolution. It also enables to gain insights into the movements and interactions of atoms over time, providing valuable information about the dynamics and behavior of biomolecular systems. To better understand all such fluctuations of the docked structures (vaccine candidates-TLR2) in the current study, the complexes were subjected to iMODS by keeping the default parameters to perform the simulations.
in-silico cloning The Java codon adaptation (JCat) tool ( https://www.jcat.de/ ) was utilized to clone the multi-epitope subunit vaccine in a suitable vector. This analysis was done for codon optimization and reverse translation to get the nucleotide sequence which was used for cloning . It was also employed to ensure, whether the vaccine construct was highly expressed after cloning in a particular vector. In this tool, three parameters were selected including, bacterial ribosome binding sites, Rho-independent transcription termination, and the restriction enzyme cleavage sites. Moreover, the Jcat calculated the GC content and CAI score of the built vaccine to optimize the reversely transcribed vaccine construct for bacterial expression. The potential vaccine sequence was manually cloned using BspAPI and PsaAI restriction enzymes, and the translated sequence of nucleotides from jcat was then cloned onto the pet28a+ plasmid by employing the SnapGene ( https://www.snapgene.com/ ).
In order to create an immunogenic profile of the designed putative vaccine, the immune simulations were performed on C-ImmSim ( https://kraken.iac.rm.cnr.it/C-IMMSIM/ ) . C-ImmSim is a dynamic agent-based simulator for evaluating the body’s immunological responses to antigens. C-ImmSim predicts immunological interactions and epitopes using an agent-based method and the particular scoring matrix, PSSM. The generation of antibodies, interferon, and cytokines was assessed after submitting the prepared vaccine to the web server, by following the default simulation parameters. The overall methodology followed in this study can be assessed more conveniently by observing the schematic workflow represented in the . Moreover, the details of all of the tools and servers incorporated in this work are also included in the .
Proteome subtraction The process of mining pathogen proteomes for therapeutic targets and to develop therapeutic strategies is greatly aided by their annotation. Target protein identification and characterization are crucial to prioritize vaccine targets and to develop strategies for the therapy of various diseases. Therefore, genome and proteome-wide, as well as functional genomics-based identification and verification of therapeutic biomarkers are frequently used to diagnose various illnesses. In the current study, we used the subtractive proteomics approach to identify the putative targets of vaccines against the emerging pathogen Candida lusitaniae. In this context, the reference proteome of Candida lusitaniae containing 5932 proteins was subjected to CELLO (version 2.0) for the screening of extracellular proteins. Since, the extracellular proteins have effective role in the viral pathogenesis hence, they were considered as targets for vaccine design. Initially, 221 extracellular proteins were shortlisted as putative targets. Furthermore, these extracellular proteins were subjected to BLASTp for further screening of the putative targets. This screening was resulted in the identification of 184 non-homologous and 37 homologous proteins. The homologous proteins were then excluded as they shared high homology with the human genome. The selection of non-homologous proteins is a strategic choice aimed at minimizing the risk of cross-reactivity with human proteins, thereby increasing the safety profile of the vaccine candidates. By opting for non-homologous proteins, we can significantly reduce the likelihood of triggering harmful or autoimmune responses within the host organism. This careful selection process is crucial for enhancing the specificity and safety of the designed vaccine, ensuring that it will effectively target Candida lusitaniae while minimizing the potential for adverse effects on the human body. Ultimately, by focusing on non-homologous proteins, we have tried to create vaccine that is not only potent against infectious agents but also exhibit a high level of safety and precision in their immune response activation. Therefore, after separating the 184 non-homologous proteins, their phylogenetic analysis was performed for determining any possible paralogs. The phylogenetic tree was constructed by employing the maximum likelihood method in MEGA11 software to analyze the presence of paralogs. To determine the reliability and robustness of the tree topology, bootstrapping technique was used while, its value was set to 100. High bootstrap values (>70–80%) indicate the reliability of interior braches of the constructed tee. The highlights three proteins which exhibited the bootstrap value of 99 which illustrated their high similarity to each other, suggesting that they were paralogs. These three proteins were excluded and the remaining 181 were then proceeded for further analysis. Afterwards, 13 proteins were selected on the basis of their essential genes from the remaining non-homologous proteins. These proteins were further subjected for allergenicity and antigenicity estimation. Consequently, four proteins were determined as highly antigenic and non-allergenic. These 4 proteins were then finally evaluated as the putative targets for vaccine design. According to their specified names which are mentioned on UniProt, these proteins were named as the SVIM type domain containing protein, F box domain containing protein, secreted protein, and heat shock protein and are found to be essential for Candida lusitaniae pathogenesis. The SVIM type domain containing protein is mainly involved in gene regulation, DNA repair and development. This protein exhibited the antigenicity score of 0.9 and was determined to be the non-allergen. The other protein F-box domain containing protein possessed the function in the cell cycle, signal transduction, developmental and stress response. The antigenicity score was 0.92 and was determined as non-allergen. While the third targeted protein was found to be involved in the biofilm formation, immune invasion, and tissue invasion. It showed the antigenicity score of 0.92 and was non-allergen. The fourth selected protein for the putative vaccine design had the function in pathogenicity and antifungal resistance. This protein’s antigenicity score was 0.75 and was also found non-allergenic. The amino acid sequences of these four proteins were retrieved from UniProt database and were further subjected to different servers for screening of CTL, HTL and B cell epitopes. The following illustrates the collective properties of the selected protein targets for putative vaccine design. Immune based epitopes screening of targeted proteins B cells and T lymphocytes which are involved in adaptive immunity, have the ability to recognize the antigenic components . T cells have various receptors which aid in recognizing the antigens while, the MHC molecules which are present on T cells, are crucial for the epitope presentation . Immunoglobulins which are secreted after B cells differentiation, function as antibodies and perform different tasks such as pathogen neutralization and targeted antigen destruction. B cells indirectly kill pathogens by activating the complement system which destroys the pathogens by different pathways . B cell and CTL in particular, are essential part of humoral and cell mediated immunity . Recent advancements in the computational modeling and the development of specific structural vaccination algorithms forecasts the validity of epitopes as a fantastic substitute for rational vaccine design. Here, we looked forward the potent CTL, B cells and HTL epitopes of the previously selected Candida lusitaniae proteins. Among the screening of CTL epitopes a total of 7, 11, 9 and 13 MHC binders were identified for each protein respectively. Similarly, the number of B cell epitopes identified for each protein were 29, 26, 30 and 34 respectively, and, for HTL there were 2038, 2227, 3403 and 3914 epitopes for each of the proteins. Further analysis was done on these epitopes to determine their antigenicity and allergenicity. For vaccine design, it was mandatory for each epitope to be non-allergen and highly antigenic in nature. For C4XWH4 protein, two CTL epitopes YVDSLYTFL, RLDLDTLEV with the antigenicity scores of 0.8 and 1.8, respectively and having the high affinity score, were selected. While, the combined affinity score of two epitopes was greater than 1. For C4Y788 protein, two epitopes i.e., FTQFSSLKV, AIGEQVRLY at the corresponding position of 188 to 197 and 323 to 332, were selected. These epitopes were non-allergen, and their combined antigenicity score was also greater than one. Along with the allergenicity and antigenicity, the high binding affinity was also kept in consideration. Further, the affinity scores of secreted and heat shock protein were also computed which were found to be in the range of 1.0–2.3. For the third protein (C4YAY0), the two epitopes i.e., LSVSTFILY, LTADFWLVV having the antigenicity score of 0.9 and 1.5 respectively, were selected. In the similar fashion, the epitopes FSDSSSGGV and SLDSTNLNL with the binding scores of 1.7 and 1.0, respectively, were selected for the fourth protein (C4X6E5). The summarized results of this overall analysis are summarized in the following . Afterwards, similar approach was followed to predict the B cell epitopes on the basis of antigenicity, binding score and the allergenicity status. In this aspect, the 2 epitopes of 16 amino acids for each protein were selected. For SVIM type domain containing protein, the two epitopes i.e., PGLVSRRPYVDSLYTF , HGMHLGPQVPKGKHVP were selected. The antigenicity score was 1.4 for both epitopes and were non allergen. Their binding scores were computed as 0.89 and 0.93, respectively. Furthermore, two epitopes were predicted for F box domain containing protein. The amino acid residues in these epitopes were TLEQFSPDSNAARYSN and TSMGEAPQESFSLAEQ at the corresponding positions of 28 and 256 with the binding score of 0.83 and 0.92 and were found to be non-allergen. The antigenicity score of the selected epitopes were calculated as 0.8 and 0.9, respectively. Furthermore, for the secreted and heat shock protein, the epitopes which possessed the effective antigenicity and binding score, were selected. Similarly, two epitopes which are GGGLGEKESHVSGQLD and ASLGQHQIDKRRHVAQ at the corresponding positions of 201 and 235 respectively in the C4YAY0 protein, were selected. The antigenicity score was 1 for both epitopes. The epitopes selected for C4XXE5 protein were LGEDGTTGEDGNITQS and GDEVRRDVTTVKLHTF . These epitopes exhibited the binding score of 0.90 and 0.86, respectively and, were shortlisted for the putative vaccine design. The combined results of all these B cell epitopes are presented in the . For the potential HTL epitope prediction, the SVM method was incorporated for C4XWH4 protein. As a result, two interferon (IFN) inducing HTL epitopes i.e., ‘ KNVSTVHVDSDKTVL and MLEVLGLRLDLDTLE ‘, were selected. While, for C4Y788 protein, the epitopes with the high antigenicity score of 1.1 and 1.0 which presented the non-allergenicity status and were also the potential IFN inducing agents, were selected. Whereas two epitopes for the secreted protein which was exhibiting positive IFN value were selected. Each of the HTL epitope was comprising 15 amino acid residues. For the fourth protein the two IFN inducing HTL epitopes were shortlisted. The antigenicity score was ranging from 0.9 to 2.3 for the last two proteins. The overall results of HTL epitopes along their respective scores are shown in the . Multi-epitope vaccine constructs (MEVC’s) of the selected proteins Bioinformatics analysis has provided a platform which assists in the identification of targets during designing of the therapeutic agents against human pathogens. This approach was followed for the designing of multi epitope vaccine constructs against the four pre-analyzed proteins of Candida lusitaniae . These multi-epitope vaccine constructs were designed by joining the shortlisted six highly antigenic CTL, HTL and B cell epitopes for each protein with the help of different peptide linkers such as EAAK, AAY, GPGPG and KK, as illustrated in . Furthermore, the immunogenic potential for MEVC peptide vaccine design was increased by the adjuvant. After an EAAK linker, non-toxic adjuvant human beta defensin2 (hBD2) was added to each of the designed MEVC constructs. With its high expression levels and ability to self-produce, hBD-2 can trigger a strong immune response against the attached antigen. For each protein, the constructs comprising of 165 amino acids for each protein were built by linking the epitopes with the help of adjuvant and linkers. These constructs were named as MEVC1, MEVC2, MEVC3 and MEVC4 having the antigenicity score of 1.2, 0.9, 1.18, and 1.12, respectively and were found to be non-allergen. Whole proteome vaccine construct By connecting adjuvant and 8 CTL, 8 HTL and 8 B cell epitopes with the help of above linkers, the whole proteome based multi epitope vaccine construct (WP-MEVC) was designed. The final construct possessed the length of 447 amino acids. The experimental feasibility of the final construct was high, as it displayed an antigenicity score over 1. The general schematic diagram of proposed whole proteome vaccine construct which includes epitopes from various proteins and appropriate additional linker is shown in . Each mapped epitope can be well observed in a distinct color. Physiochemical properties The physiochemical properties provides insights of various crucial factors regarding the physical and chemical attributes of the proteins under study. These attributes include the stability index, no. of amino acid residues, half-life and number of positively and negatively charges residues, among other important properties. Thus, ProtParam server was employed to analyze the physiochemical perspectives of the MEVC’s. Molecular weight 17 to 47 kD, theoretical Pi 8.68 to 9.30, and other variables assessing the vaccine stability and viability for future experimental designs, were also among these parameters. In addition, the half-life, GRAVY (grand average of hydropathicity), aliphatic index and thermostability were also examined. The demonstrates various physiochemical characteristics for all the MEVC’s under study. Secondary structure prediction and 3D modeling of multi-epitope vaccine construct The results of SOPMA showed that the whole proteome vaccine construct consisted of 63.31% of random coils, which provided a secondary structure with a lot of flexibility. Alpha helices represented 14.32%, causing in the formation of a stable helical region in them while, 22.37% was constituted of extended strands or beta-sheets, providing some domains with very stable core regions. These results can be visualized in other interesting layouts in the form of peaks and bars in . The graphical data showed the structures distributed throughout the protein, presenting the regions of flexibility interspersed with more rigid and structured areas, highlighting a versatile protein. While, for the 3D structure modeling of vaccine constructs, the trRosetta server was used which is the web-based platform for the efficient and accurate protein structure prediction. It develops the protein structure using direct energy minimizations with a constrained Rosetta. This server predicts the entire residue geometries of proteins including distances and orientation. In current scenario, three dimensional (3D) structure modeling for each of the constructed MEVC’s was performed with the utilities of the trRosetta server. While, the generated models were visualized BIOVIA Discovery Studio Visualizer. The refined best models are shown in the which were proceeded for further analysis. The whole proteome vaccine construct was refined using galaxy refine. For the validation of the predicted structure of whole proteome vaccine construct, Ramachandran plot which is available at PROCHECK server ( https://saves.mbi.ucla.edu/ ), was incorporated. This plot was basically used to validate and assess the quality of the predicted protein structure and to identify the potential errors presents in the backbone conformation of that protein. The resulted Ramachandran plot suggested that most of the amino acids residues were present in the most favorable regions which indicated the good quality of the predicted structure. Ramachandran plot also generates the results in the form of Ramachandran scores. If this score is greater than 90% then, it indicates that the predicted protein structure is of good quality. In our context, the Ramachandran score of 96.3% was computed which suggested that the predicted structure was stereochemically valid. The results of Ramachandran plot are illustrated in the . Discontinuous B cell epitope prediction This particular analysis identified six discontinuous B-cell epitopes in the whole proteome vaccine construct, each consisting of surface-exposed residues likely to be recognized by antibodies. Each of these epitopes had different protrusion index (PI) scores which indicated their accessibility at the surface level, where epitope 1 composed of 38 amino acids and it displayed the highest PI value of 0.922, which means it can be recognized by antibodies very easily. In addition, epitope 2 (90 residues) had a PI score of 0.738 while epitope 3 (39 residues) illustrated a value of 0.702, suggesting that these epitopes were also found to be highly exposed at their surfaces thus making them excellent targets for immune response inducing agents. However, others exhibited progressively lower but surface-exposed PI scores with a lesser degree of relevancy as compared with the three, mentioned above in terms of protein structural stability. These predictions highlighted key regions for antibody interaction that are critical for vaccine design. The represent the discontinuous B cell epitope of the vaccine construct. Disulfide engineering The disulfide engineering was performed in order to stabilize the predicted structure of the final vaccine constructs. It was discovered that a total of 17 pairs of residues might be utilized in this way. However, following assessment of additional factors such as energy and Chi3 value, only three pairs of residues Pro28-Cys23, Cys53-Cy60, and Ser383-Ser386 were determined to be final since their values were seen to be falling within the permitted range, i.e., energy values must be less than 2.2 and Chi3 values must be between -87 and +97 degree37. As a result, two pairs as shown in successfully formed disulfide bonds, such that bond distance, angles (like Chi3), and the energy minimized structure resulted in favorable conditions for disulfide bonds. Two different areas where disulfide bonds were included during this particular analysis can be observed in the . MEVCs-TLR2 docking analysis and demonstration of interacting docked residues The HADDOCK server was used to assess the interaction between the toll like receptor 2 (TLR2) and the vaccine constructs. The docking results include the analysis of Z value by the HADDOCK server. The Z score in this server indicates that how many standard deviations occur from the average values. The lower score indicates the better interaction and quality of the docked complexes. The docking analysis typically predicted the conformation, position, and orientation of vaccine with TLR2 receptor site. All of the resulted three dimensional docked complexes are shown in the in which two distinct colors, green and blue highlights the docked TLR2 and MEVC’s, respectively. While, the respective Z-scores of all the docked complexes are demonstrated in . These scores of all the docked complexes fall in an appropriate range of -1.4 to -2.3 which signified the stable docking interactions. Furthermore, the in-depth analysis of the specified docked residues in each of the docked complex was also performed in order to comprehend the number of H-bonds, salt bridges and other interactions. This task was performed by utilizing the PDBsum server ( https://www.ebi.ac.uk/thornton-srv/databases/pdbsum/ ). In this regard, MEVC1-TLR2 construct formed 3 salt bridges, 11 hydrogen bonds and 221 non-bonded interactions. MECV2-TLR 2 exhibited 12 hydrogen bonds, 265 non-bonded interactions, and 6 salt bridges. Whereas, MEVC3-TLR2 and MEVC4-TLR2 constructs formed 3 salt bridges, 19 and 11 corresponding hydrogen bonds and presented 288 and 235 non-bonded interactions, respectively. Lastly, the results of WP-TLR2 analysis displayed 8 salt bridges, 22 hydrogen bonds and 216 non-bonded interactions. Moreover, all of the constructs had a relatively consistent GC% of 52–55%, which suggested a balanced nucleotide composition. All of the relevant parameters of docking analysis along their respective scores for each docking complex are mentioned in the . Whereas, to comprehend the docking interactions of TLR2-WP vaccine construct, the results of PDBsum are also illustrated in the . Molecular dynamics simulations of TLR2+MEVC-WP docked complex The analysis of molecular dynamics simulations revealed observable differences at specific atomic intervals in the TLR2 and MEVC-WP docked complex. Deep insights into the docking analysis have been provided by molecular dynamics simulations, allowing for a more thorough comprehension of the complex interactions between two molecules. The shows a topographical representation with peaks indicating how flexible the docked structure is, with different residue sites showing different degrees of flexibility. These peaks are particularly prominent at the beginning of the atom index continuum. The B-factor graph’s dynamic fluctuations validate this observation, highlighting the relevance of normal mode analysis within the complex. The eigenvalue map, which shows how much energy is needed to deform a docked complex, exhibited eigenvalue score of 4.73×10−08, indicating the strong structural stability of TLR2 and MEVC-WP docked complex. On the contrary, the variance map which is inversely related to eigenvalues, is explained by means of a graphical representation that clarifies the cumulative and individual variances of amino acid residues. Three critical aspects that control the dynamics of the interaction between TLR2 and MEVC-WP docked complex are shown by the covariance map which comprised of the correlated (red), uncorrelated (blue), and anti-correlated (white) movements of component residues. The covariance map, in particular, highlighted strongly correlated motions in the parameter space from 0 to 1200 residue indices on both axes, interspersed with uncorrelated and anti-correlated interactions at various loci. Finally, the elastic network model explored the target structure’s pliancy, which was indicated by different grayscale intensities. In current perspective, this particular analysis suggested that the previously discussed correlated residues were highly flexible in the context of TLR2 and MEVC-WP docked complex. This proposed that the docked complex of TLR2 and MEVC-WP possessed significant levels of flexibility which would enable the docked structure to be stable upon certain fluctuations. Codon optimization and in-silico cloning Codon adaptation index (CAI) is a numerical measure in bioinformatics that assesses how well a gene or DNA sequence aligns with the preferred codon usage of a reference set, often the highly expressed genes in a specific organism. The relative adaptiveness of a codon is determined by comparing its frequency in the gene being studied to its highest frequency in the reference set. This comparison results in a ratio between 0.1 and 1, where 1 signifies a perfect match to the codon usage bias of the reference set. If the CAI value is higher, it means that the gene’s codon usage closely matches that of the reference set, indicating better adaptation or optimization for efficient translation. The JCcat server was used to optimize the expression of all developed vaccine sequences in the E . coli . K12 strain. The corresponding nucleotides for every vaccine construct were obtained by optimizing the codons and performing reverse translation. For every vaccine construct codon optimization index CAI values were also computed through the JCat server. The CAI values for MEVC1, MEVC2 and MEVC4 were 1.0. While, for MEVC3 and whole proteome-MEVC, the CAI values were 0.95 and 0.98, respectively. These values indicated high degree of relativeness between the codon usage of multi-epitope vaccine constructs and E . coli . The overall results produced by the JCat server were in the form of graphs plotted between relative adaptiveness and codons. In the , y-axis of the graph represents relative adaptiveness of each codon. This metric indicated that on the basis of a reference set how well suited each codon was, in terms of its frequency of occurrence within the gene compared to the expected ideal frequencies. Higher values on y-axis represents the codons that were efficiently utilized and hence more adaptive in the context of translation efficiency and protein expression. Whereas, each point on x-axis represents a specific codon within the genetic sequence, being analyzed. The codons with relatively high adaptiveness are more favored in the gene sequence due to their optimal usage in the host organism’s translational machinery. Moreover, the higher adaptiveness also likely lead to efficient protein expression and proper folding, potentially impacting the functionality and yield of the encoded protein. The high calculated CAI scores were in the range of 0.9 to 1.0 showing that vaccine protein was expressed at a high level in E . coli . While, all the percentages of the GC content that range from 52 to 55 was previously illustrated in the . Therefore, the corresponding nucleotide sequence was then employed to make the DNA fragment for in-silico cloning. Furthermore, SnapGene software was utilized for in-silico cloning. For this purpose, pet28a + vector was selected for cloning. The putative vaccine sequence was then inserted into the pet28a + by XcmI and PsaAI restriction enzymes of respective 5’ to 3’ ends which can be observed by red color in the . Immune simulations of the final vaccine construct Each vaccine was analyzed as an injected antigen using in-silico immune simulations in order to assess the potential of antigen-based induction of immune response. In the , the map showed that after the injection of MEVC1, it achieved high antigen count at the day 2 and then neutralized slowly till day 5. Then the antibody IgM achieved the level close to 700000 antigens count per milliliter between the days 7 to 20. For the second vaccine construct, it similarly achieved high antigen account at the day 2 and neutralized till the day 5. After this, the action of strong antibodies IgM and IgG achieved the antibody titer greater than 600000 antigen count per milliliter between the days 5 to 15 and then till day 30, it slowly neutralized to less than 300000. Along with this IgM, antibody also presented the figures of 600000 antigen count per milliliter. However, these variations were significant. The MEVC3 illustrated almost the same results as the MEVC1. The antigen count was 700000 on the day 2 and the IgG antibody titers was close to 800000 between the days 10 to 20 and then it neutralized slowly till the day 30. The MEVC4 displayed the map of antigen, IgG, IgM and IgG1. The antigen count was high on day 2 up to the 700000 antigens count per millimeter. IgM and IgG count showed the values, almost close to 800000 au/ml between the days 10 to 15. While, both of them exhibited slow neutralization till the day 30 up to the 280000 antigens count per milliliter. The IG1 antibody indicated the titers up to 400000 antigens count per milliliter from the day 13 to 30. For the whole proteome (WP) based vaccine construct, results exhibited that the antigen count was greater than 600000 on the day two and then this value continued to fall till the day 5. The strongest antibodies IgG and IgM represented the highest count up to the 700000 between the days 8 to 13 and then slowly neutralized till the day 30. The IgM antibody presented a count greater than 400000 on day 10 and then slowly utilized till day 30. Similarly, the IG1 antibody titer exhibited the graph close to the 300000 au/ml. These results suggested that the designed vaccine can trigger an effective immune response and possessed the potential to induce immunity. Hence, it can be proposed that designed vaccine candidates can trigger the immune response against the pathogen Candida lusitaniae . The results of all the immune simulations discussed above are illustrated below in the .
The process of mining pathogen proteomes for therapeutic targets and to develop therapeutic strategies is greatly aided by their annotation. Target protein identification and characterization are crucial to prioritize vaccine targets and to develop strategies for the therapy of various diseases. Therefore, genome and proteome-wide, as well as functional genomics-based identification and verification of therapeutic biomarkers are frequently used to diagnose various illnesses. In the current study, we used the subtractive proteomics approach to identify the putative targets of vaccines against the emerging pathogen Candida lusitaniae. In this context, the reference proteome of Candida lusitaniae containing 5932 proteins was subjected to CELLO (version 2.0) for the screening of extracellular proteins. Since, the extracellular proteins have effective role in the viral pathogenesis hence, they were considered as targets for vaccine design. Initially, 221 extracellular proteins were shortlisted as putative targets. Furthermore, these extracellular proteins were subjected to BLASTp for further screening of the putative targets. This screening was resulted in the identification of 184 non-homologous and 37 homologous proteins. The homologous proteins were then excluded as they shared high homology with the human genome. The selection of non-homologous proteins is a strategic choice aimed at minimizing the risk of cross-reactivity with human proteins, thereby increasing the safety profile of the vaccine candidates. By opting for non-homologous proteins, we can significantly reduce the likelihood of triggering harmful or autoimmune responses within the host organism. This careful selection process is crucial for enhancing the specificity and safety of the designed vaccine, ensuring that it will effectively target Candida lusitaniae while minimizing the potential for adverse effects on the human body. Ultimately, by focusing on non-homologous proteins, we have tried to create vaccine that is not only potent against infectious agents but also exhibit a high level of safety and precision in their immune response activation. Therefore, after separating the 184 non-homologous proteins, their phylogenetic analysis was performed for determining any possible paralogs. The phylogenetic tree was constructed by employing the maximum likelihood method in MEGA11 software to analyze the presence of paralogs. To determine the reliability and robustness of the tree topology, bootstrapping technique was used while, its value was set to 100. High bootstrap values (>70–80%) indicate the reliability of interior braches of the constructed tee. The highlights three proteins which exhibited the bootstrap value of 99 which illustrated their high similarity to each other, suggesting that they were paralogs. These three proteins were excluded and the remaining 181 were then proceeded for further analysis. Afterwards, 13 proteins were selected on the basis of their essential genes from the remaining non-homologous proteins. These proteins were further subjected for allergenicity and antigenicity estimation. Consequently, four proteins were determined as highly antigenic and non-allergenic. These 4 proteins were then finally evaluated as the putative targets for vaccine design. According to their specified names which are mentioned on UniProt, these proteins were named as the SVIM type domain containing protein, F box domain containing protein, secreted protein, and heat shock protein and are found to be essential for Candida lusitaniae pathogenesis. The SVIM type domain containing protein is mainly involved in gene regulation, DNA repair and development. This protein exhibited the antigenicity score of 0.9 and was determined to be the non-allergen. The other protein F-box domain containing protein possessed the function in the cell cycle, signal transduction, developmental and stress response. The antigenicity score was 0.92 and was determined as non-allergen. While the third targeted protein was found to be involved in the biofilm formation, immune invasion, and tissue invasion. It showed the antigenicity score of 0.92 and was non-allergen. The fourth selected protein for the putative vaccine design had the function in pathogenicity and antifungal resistance. This protein’s antigenicity score was 0.75 and was also found non-allergenic. The amino acid sequences of these four proteins were retrieved from UniProt database and were further subjected to different servers for screening of CTL, HTL and B cell epitopes. The following illustrates the collective properties of the selected protein targets for putative vaccine design.
B cells and T lymphocytes which are involved in adaptive immunity, have the ability to recognize the antigenic components . T cells have various receptors which aid in recognizing the antigens while, the MHC molecules which are present on T cells, are crucial for the epitope presentation . Immunoglobulins which are secreted after B cells differentiation, function as antibodies and perform different tasks such as pathogen neutralization and targeted antigen destruction. B cells indirectly kill pathogens by activating the complement system which destroys the pathogens by different pathways . B cell and CTL in particular, are essential part of humoral and cell mediated immunity . Recent advancements in the computational modeling and the development of specific structural vaccination algorithms forecasts the validity of epitopes as a fantastic substitute for rational vaccine design. Here, we looked forward the potent CTL, B cells and HTL epitopes of the previously selected Candida lusitaniae proteins. Among the screening of CTL epitopes a total of 7, 11, 9 and 13 MHC binders were identified for each protein respectively. Similarly, the number of B cell epitopes identified for each protein were 29, 26, 30 and 34 respectively, and, for HTL there were 2038, 2227, 3403 and 3914 epitopes for each of the proteins. Further analysis was done on these epitopes to determine their antigenicity and allergenicity. For vaccine design, it was mandatory for each epitope to be non-allergen and highly antigenic in nature. For C4XWH4 protein, two CTL epitopes YVDSLYTFL, RLDLDTLEV with the antigenicity scores of 0.8 and 1.8, respectively and having the high affinity score, were selected. While, the combined affinity score of two epitopes was greater than 1. For C4Y788 protein, two epitopes i.e., FTQFSSLKV, AIGEQVRLY at the corresponding position of 188 to 197 and 323 to 332, were selected. These epitopes were non-allergen, and their combined antigenicity score was also greater than one. Along with the allergenicity and antigenicity, the high binding affinity was also kept in consideration. Further, the affinity scores of secreted and heat shock protein were also computed which were found to be in the range of 1.0–2.3. For the third protein (C4YAY0), the two epitopes i.e., LSVSTFILY, LTADFWLVV having the antigenicity score of 0.9 and 1.5 respectively, were selected. In the similar fashion, the epitopes FSDSSSGGV and SLDSTNLNL with the binding scores of 1.7 and 1.0, respectively, were selected for the fourth protein (C4X6E5). The summarized results of this overall analysis are summarized in the following . Afterwards, similar approach was followed to predict the B cell epitopes on the basis of antigenicity, binding score and the allergenicity status. In this aspect, the 2 epitopes of 16 amino acids for each protein were selected. For SVIM type domain containing protein, the two epitopes i.e., PGLVSRRPYVDSLYTF , HGMHLGPQVPKGKHVP were selected. The antigenicity score was 1.4 for both epitopes and were non allergen. Their binding scores were computed as 0.89 and 0.93, respectively. Furthermore, two epitopes were predicted for F box domain containing protein. The amino acid residues in these epitopes were TLEQFSPDSNAARYSN and TSMGEAPQESFSLAEQ at the corresponding positions of 28 and 256 with the binding score of 0.83 and 0.92 and were found to be non-allergen. The antigenicity score of the selected epitopes were calculated as 0.8 and 0.9, respectively. Furthermore, for the secreted and heat shock protein, the epitopes which possessed the effective antigenicity and binding score, were selected. Similarly, two epitopes which are GGGLGEKESHVSGQLD and ASLGQHQIDKRRHVAQ at the corresponding positions of 201 and 235 respectively in the C4YAY0 protein, were selected. The antigenicity score was 1 for both epitopes. The epitopes selected for C4XXE5 protein were LGEDGTTGEDGNITQS and GDEVRRDVTTVKLHTF . These epitopes exhibited the binding score of 0.90 and 0.86, respectively and, were shortlisted for the putative vaccine design. The combined results of all these B cell epitopes are presented in the . For the potential HTL epitope prediction, the SVM method was incorporated for C4XWH4 protein. As a result, two interferon (IFN) inducing HTL epitopes i.e., ‘ KNVSTVHVDSDKTVL and MLEVLGLRLDLDTLE ‘, were selected. While, for C4Y788 protein, the epitopes with the high antigenicity score of 1.1 and 1.0 which presented the non-allergenicity status and were also the potential IFN inducing agents, were selected. Whereas two epitopes for the secreted protein which was exhibiting positive IFN value were selected. Each of the HTL epitope was comprising 15 amino acid residues. For the fourth protein the two IFN inducing HTL epitopes were shortlisted. The antigenicity score was ranging from 0.9 to 2.3 for the last two proteins. The overall results of HTL epitopes along their respective scores are shown in the .
Bioinformatics analysis has provided a platform which assists in the identification of targets during designing of the therapeutic agents against human pathogens. This approach was followed for the designing of multi epitope vaccine constructs against the four pre-analyzed proteins of Candida lusitaniae . These multi-epitope vaccine constructs were designed by joining the shortlisted six highly antigenic CTL, HTL and B cell epitopes for each protein with the help of different peptide linkers such as EAAK, AAY, GPGPG and KK, as illustrated in . Furthermore, the immunogenic potential for MEVC peptide vaccine design was increased by the adjuvant. After an EAAK linker, non-toxic adjuvant human beta defensin2 (hBD2) was added to each of the designed MEVC constructs. With its high expression levels and ability to self-produce, hBD-2 can trigger a strong immune response against the attached antigen. For each protein, the constructs comprising of 165 amino acids for each protein were built by linking the epitopes with the help of adjuvant and linkers. These constructs were named as MEVC1, MEVC2, MEVC3 and MEVC4 having the antigenicity score of 1.2, 0.9, 1.18, and 1.12, respectively and were found to be non-allergen.
By connecting adjuvant and 8 CTL, 8 HTL and 8 B cell epitopes with the help of above linkers, the whole proteome based multi epitope vaccine construct (WP-MEVC) was designed. The final construct possessed the length of 447 amino acids. The experimental feasibility of the final construct was high, as it displayed an antigenicity score over 1. The general schematic diagram of proposed whole proteome vaccine construct which includes epitopes from various proteins and appropriate additional linker is shown in . Each mapped epitope can be well observed in a distinct color.
The physiochemical properties provides insights of various crucial factors regarding the physical and chemical attributes of the proteins under study. These attributes include the stability index, no. of amino acid residues, half-life and number of positively and negatively charges residues, among other important properties. Thus, ProtParam server was employed to analyze the physiochemical perspectives of the MEVC’s. Molecular weight 17 to 47 kD, theoretical Pi 8.68 to 9.30, and other variables assessing the vaccine stability and viability for future experimental designs, were also among these parameters. In addition, the half-life, GRAVY (grand average of hydropathicity), aliphatic index and thermostability were also examined. The demonstrates various physiochemical characteristics for all the MEVC’s under study.
The results of SOPMA showed that the whole proteome vaccine construct consisted of 63.31% of random coils, which provided a secondary structure with a lot of flexibility. Alpha helices represented 14.32%, causing in the formation of a stable helical region in them while, 22.37% was constituted of extended strands or beta-sheets, providing some domains with very stable core regions. These results can be visualized in other interesting layouts in the form of peaks and bars in . The graphical data showed the structures distributed throughout the protein, presenting the regions of flexibility interspersed with more rigid and structured areas, highlighting a versatile protein. While, for the 3D structure modeling of vaccine constructs, the trRosetta server was used which is the web-based platform for the efficient and accurate protein structure prediction. It develops the protein structure using direct energy minimizations with a constrained Rosetta. This server predicts the entire residue geometries of proteins including distances and orientation. In current scenario, three dimensional (3D) structure modeling for each of the constructed MEVC’s was performed with the utilities of the trRosetta server. While, the generated models were visualized BIOVIA Discovery Studio Visualizer. The refined best models are shown in the which were proceeded for further analysis. The whole proteome vaccine construct was refined using galaxy refine. For the validation of the predicted structure of whole proteome vaccine construct, Ramachandran plot which is available at PROCHECK server ( https://saves.mbi.ucla.edu/ ), was incorporated. This plot was basically used to validate and assess the quality of the predicted protein structure and to identify the potential errors presents in the backbone conformation of that protein. The resulted Ramachandran plot suggested that most of the amino acids residues were present in the most favorable regions which indicated the good quality of the predicted structure. Ramachandran plot also generates the results in the form of Ramachandran scores. If this score is greater than 90% then, it indicates that the predicted protein structure is of good quality. In our context, the Ramachandran score of 96.3% was computed which suggested that the predicted structure was stereochemically valid. The results of Ramachandran plot are illustrated in the .
This particular analysis identified six discontinuous B-cell epitopes in the whole proteome vaccine construct, each consisting of surface-exposed residues likely to be recognized by antibodies. Each of these epitopes had different protrusion index (PI) scores which indicated their accessibility at the surface level, where epitope 1 composed of 38 amino acids and it displayed the highest PI value of 0.922, which means it can be recognized by antibodies very easily. In addition, epitope 2 (90 residues) had a PI score of 0.738 while epitope 3 (39 residues) illustrated a value of 0.702, suggesting that these epitopes were also found to be highly exposed at their surfaces thus making them excellent targets for immune response inducing agents. However, others exhibited progressively lower but surface-exposed PI scores with a lesser degree of relevancy as compared with the three, mentioned above in terms of protein structural stability. These predictions highlighted key regions for antibody interaction that are critical for vaccine design. The represent the discontinuous B cell epitope of the vaccine construct.
The disulfide engineering was performed in order to stabilize the predicted structure of the final vaccine constructs. It was discovered that a total of 17 pairs of residues might be utilized in this way. However, following assessment of additional factors such as energy and Chi3 value, only three pairs of residues Pro28-Cys23, Cys53-Cy60, and Ser383-Ser386 were determined to be final since their values were seen to be falling within the permitted range, i.e., energy values must be less than 2.2 and Chi3 values must be between -87 and +97 degree37. As a result, two pairs as shown in successfully formed disulfide bonds, such that bond distance, angles (like Chi3), and the energy minimized structure resulted in favorable conditions for disulfide bonds. Two different areas where disulfide bonds were included during this particular analysis can be observed in the .
The HADDOCK server was used to assess the interaction between the toll like receptor 2 (TLR2) and the vaccine constructs. The docking results include the analysis of Z value by the HADDOCK server. The Z score in this server indicates that how many standard deviations occur from the average values. The lower score indicates the better interaction and quality of the docked complexes. The docking analysis typically predicted the conformation, position, and orientation of vaccine with TLR2 receptor site. All of the resulted three dimensional docked complexes are shown in the in which two distinct colors, green and blue highlights the docked TLR2 and MEVC’s, respectively. While, the respective Z-scores of all the docked complexes are demonstrated in . These scores of all the docked complexes fall in an appropriate range of -1.4 to -2.3 which signified the stable docking interactions. Furthermore, the in-depth analysis of the specified docked residues in each of the docked complex was also performed in order to comprehend the number of H-bonds, salt bridges and other interactions. This task was performed by utilizing the PDBsum server ( https://www.ebi.ac.uk/thornton-srv/databases/pdbsum/ ). In this regard, MEVC1-TLR2 construct formed 3 salt bridges, 11 hydrogen bonds and 221 non-bonded interactions. MECV2-TLR 2 exhibited 12 hydrogen bonds, 265 non-bonded interactions, and 6 salt bridges. Whereas, MEVC3-TLR2 and MEVC4-TLR2 constructs formed 3 salt bridges, 19 and 11 corresponding hydrogen bonds and presented 288 and 235 non-bonded interactions, respectively. Lastly, the results of WP-TLR2 analysis displayed 8 salt bridges, 22 hydrogen bonds and 216 non-bonded interactions. Moreover, all of the constructs had a relatively consistent GC% of 52–55%, which suggested a balanced nucleotide composition. All of the relevant parameters of docking analysis along their respective scores for each docking complex are mentioned in the . Whereas, to comprehend the docking interactions of TLR2-WP vaccine construct, the results of PDBsum are also illustrated in the .
The analysis of molecular dynamics simulations revealed observable differences at specific atomic intervals in the TLR2 and MEVC-WP docked complex. Deep insights into the docking analysis have been provided by molecular dynamics simulations, allowing for a more thorough comprehension of the complex interactions between two molecules. The shows a topographical representation with peaks indicating how flexible the docked structure is, with different residue sites showing different degrees of flexibility. These peaks are particularly prominent at the beginning of the atom index continuum. The B-factor graph’s dynamic fluctuations validate this observation, highlighting the relevance of normal mode analysis within the complex. The eigenvalue map, which shows how much energy is needed to deform a docked complex, exhibited eigenvalue score of 4.73×10−08, indicating the strong structural stability of TLR2 and MEVC-WP docked complex. On the contrary, the variance map which is inversely related to eigenvalues, is explained by means of a graphical representation that clarifies the cumulative and individual variances of amino acid residues. Three critical aspects that control the dynamics of the interaction between TLR2 and MEVC-WP docked complex are shown by the covariance map which comprised of the correlated (red), uncorrelated (blue), and anti-correlated (white) movements of component residues. The covariance map, in particular, highlighted strongly correlated motions in the parameter space from 0 to 1200 residue indices on both axes, interspersed with uncorrelated and anti-correlated interactions at various loci. Finally, the elastic network model explored the target structure’s pliancy, which was indicated by different grayscale intensities. In current perspective, this particular analysis suggested that the previously discussed correlated residues were highly flexible in the context of TLR2 and MEVC-WP docked complex. This proposed that the docked complex of TLR2 and MEVC-WP possessed significant levels of flexibility which would enable the docked structure to be stable upon certain fluctuations.
Codon adaptation index (CAI) is a numerical measure in bioinformatics that assesses how well a gene or DNA sequence aligns with the preferred codon usage of a reference set, often the highly expressed genes in a specific organism. The relative adaptiveness of a codon is determined by comparing its frequency in the gene being studied to its highest frequency in the reference set. This comparison results in a ratio between 0.1 and 1, where 1 signifies a perfect match to the codon usage bias of the reference set. If the CAI value is higher, it means that the gene’s codon usage closely matches that of the reference set, indicating better adaptation or optimization for efficient translation. The JCcat server was used to optimize the expression of all developed vaccine sequences in the E . coli . K12 strain. The corresponding nucleotides for every vaccine construct were obtained by optimizing the codons and performing reverse translation. For every vaccine construct codon optimization index CAI values were also computed through the JCat server. The CAI values for MEVC1, MEVC2 and MEVC4 were 1.0. While, for MEVC3 and whole proteome-MEVC, the CAI values were 0.95 and 0.98, respectively. These values indicated high degree of relativeness between the codon usage of multi-epitope vaccine constructs and E . coli . The overall results produced by the JCat server were in the form of graphs plotted between relative adaptiveness and codons. In the , y-axis of the graph represents relative adaptiveness of each codon. This metric indicated that on the basis of a reference set how well suited each codon was, in terms of its frequency of occurrence within the gene compared to the expected ideal frequencies. Higher values on y-axis represents the codons that were efficiently utilized and hence more adaptive in the context of translation efficiency and protein expression. Whereas, each point on x-axis represents a specific codon within the genetic sequence, being analyzed. The codons with relatively high adaptiveness are more favored in the gene sequence due to their optimal usage in the host organism’s translational machinery. Moreover, the higher adaptiveness also likely lead to efficient protein expression and proper folding, potentially impacting the functionality and yield of the encoded protein. The high calculated CAI scores were in the range of 0.9 to 1.0 showing that vaccine protein was expressed at a high level in E . coli . While, all the percentages of the GC content that range from 52 to 55 was previously illustrated in the . Therefore, the corresponding nucleotide sequence was then employed to make the DNA fragment for in-silico cloning. Furthermore, SnapGene software was utilized for in-silico cloning. For this purpose, pet28a + vector was selected for cloning. The putative vaccine sequence was then inserted into the pet28a + by XcmI and PsaAI restriction enzymes of respective 5’ to 3’ ends which can be observed by red color in the .
Each vaccine was analyzed as an injected antigen using in-silico immune simulations in order to assess the potential of antigen-based induction of immune response. In the , the map showed that after the injection of MEVC1, it achieved high antigen count at the day 2 and then neutralized slowly till day 5. Then the antibody IgM achieved the level close to 700000 antigens count per milliliter between the days 7 to 20. For the second vaccine construct, it similarly achieved high antigen account at the day 2 and neutralized till the day 5. After this, the action of strong antibodies IgM and IgG achieved the antibody titer greater than 600000 antigen count per milliliter between the days 5 to 15 and then till day 30, it slowly neutralized to less than 300000. Along with this IgM, antibody also presented the figures of 600000 antigen count per milliliter. However, these variations were significant. The MEVC3 illustrated almost the same results as the MEVC1. The antigen count was 700000 on the day 2 and the IgG antibody titers was close to 800000 between the days 10 to 20 and then it neutralized slowly till the day 30. The MEVC4 displayed the map of antigen, IgG, IgM and IgG1. The antigen count was high on day 2 up to the 700000 antigens count per millimeter. IgM and IgG count showed the values, almost close to 800000 au/ml between the days 10 to 15. While, both of them exhibited slow neutralization till the day 30 up to the 280000 antigens count per milliliter. The IG1 antibody indicated the titers up to 400000 antigens count per milliliter from the day 13 to 30. For the whole proteome (WP) based vaccine construct, results exhibited that the antigen count was greater than 600000 on the day two and then this value continued to fall till the day 5. The strongest antibodies IgG and IgM represented the highest count up to the 700000 between the days 8 to 13 and then slowly neutralized till the day 30. The IgM antibody presented a count greater than 400000 on day 10 and then slowly utilized till day 30. Similarly, the IG1 antibody titer exhibited the graph close to the 300000 au/ml. These results suggested that the designed vaccine can trigger an effective immune response and possessed the potential to induce immunity. Hence, it can be proposed that designed vaccine candidates can trigger the immune response against the pathogen Candida lusitaniae . The results of all the immune simulations discussed above are illustrated below in the .
This study primarily aimed to design a putative vaccine against Candida lusitaniae using a subtractive proteomics approach . This method has been applied for successive systematic steps toward the identification of suitable vaccine targets from the pathogen proteome. First, the reference proteome of Candida lusitaniae was retrieved from the UniProt database and further analyzed using various web-based bioinformatics tools. Remaining extracellular proteins were analyzed with the CELLO tool to predict which may give an immune response. The authors of one similar study focused on computational design for a multiepitope vaccine against Langya henipavirus and targeted the surface proteins where antigens are likely to be expressed. They paid more attention to extracellular proteins because these are more accessible to the immune system and, subsequently, enhance the vaccine to provoke an even more aggressive response from the immune system. This approach stressed the selection of suitable protein targets toward the design of effective multi-epitope vaccines . The identified proteins were subsequently subjected to BLASTp in order to distinguish between homologous and non-homologous proteins with respect to the host, Homo sapiens . Phylogenetic analysis was conducted on the non-homologous proteins that may indicate a number of paralogs; however, this study did not find any paralogs. The next thing was to determine their allergenicity and antigenicity, which would allow the categorization of these proteins based on their ability to cause allergic responses. In the process, the group of proteins that showed high allergenicity and another group with no allergenicity led to the selection of the four proteins chosen for their further study and possible development as vaccines. Of note, a similar approach was previously reported in the design of a vaccine against Candida auris , where the authors used the tool CD-HIT to identify paralogs instead of phylogenetic analysis . A somewhat similar study was conducted with the aim of finding vaccine targets for Brucella melitensis ; the authors conducted a proteome-wide screening as done in this study. However, they incorporated additional analyses to complement their findings. The analyses included BPGA and the investigation of transmembrane helices, both of which provided further insight into the characteristics of the proteins relevant to vaccine development . While these additional analyses were performed, it is interesting to note that most of the methods used in their study were very similar to those used in our analysis. This highlights the efficiency and flexibility of the proteome-wide screening approach in identifying potential vaccine candidates against several pathogens. After selecting potential proteins, a further screening was done to predict a set of epitopes including CTLs, HTLs, linear B-cell epitopes, and discontinuous B-cell epitopes. Studying Helper T-lymphocytes, particularly, interferon-gamma-inducing MHC-II epitopes were identified with the help of the online tool IFN-epitope. The epitope screening effort here focused on searching for epitopes with high antigenicity and non-allergenic properties, an important factor in a vaccine candidate construct. A similar study was done to construct a multi-epitope chimeric vaccine against Human Papillomavirus . The IFN-epitope server uses advanced algorithms to predict and design peptides that stimulate IFN-gamma, along with MHC Class II binders and T-cell epitopes. The predicted CTL, HTL, and B-cell epitopes were then concatenated using various linkers, including KK, GPGPG, and AAY . These linkers are important for enhancing epitope presentation, allowing sufficient spacing between epitopes and avoiding incorrect folding, which in turn enhances the overall potency of the immune response. hBD-2, an adjuvant proven to be harmless to humans, was added to MEVC in order to enhance its immunogenic efficacy. In the final version of the MEVC, WP-MEVC utilized an EAAAK linker, with the latter incorporating the hBD-2 adjuvant, which is known to express itself autonomously at sufficient levels capable of mounting a strong immune response toward the antigen of interest . In a similar study, epitopes with an adjuvant were coupled using an EAAAK linker while designing a multi-epitope vaccine against Nosocomial Proteus penneri . This approach has overhauled the role of linkers in improving stability and immunogenicity of vaccine constructs . This holistic approach led to the development of a whole proteome-based final multi-epitope vaccine, the WP-MEVC. Selection of these linkers was supported by a comprehensive review of the literature, and there are analogous methodologies considered by other authors in the computer-aided design of potential multi-epitope vaccines, especially those targeting SARS-CoV-2 variants . Following the design of the whole-proteome vaccine construct, we analyzed its stability and viability using an online webtool, ProtParam. This tool enabled the examination of a number of physicochemical characteristics of the vaccine construct, including its molecular weight, theoretical pI, half-life, GRAVY, and stability index. Through the evaluation of these parameters, a thorough comprehension of the vaccine construct’s structural and chemical properties was achieved, which is essential for its efficacy as a vaccine. Comparable methodologies have been utilized in various studies pertaining to the in-silico development of multi-epitope vaccines targeting pathogens including Porphyromonas gingivalis , Streptococcus gordonii , and Cardiobacterium valvarum , as well as in the formulation of mRNA-based vaccines . These studies, combined with our own research findings, further emphasize the importance of computational methods and algorithms in the design and optimization of effective vaccine formulations. By conducting an in-depth examination of stability and viability, we determined that our whole-proteome vaccine construct possessed the necessary physicochemical properties to be considered a promising and effective vaccine candidate. Commonalities in methodology among various research efforts emphasize that computational techniques are integral in advancing vaccines, as such approaches yield essential knowledge regarding both structural and functional nature of the designs of vaccines. In addition to the immunoinformatics and simulation-based strategies employed for the identification of epitopes, we modeled the tertiary structure of all the multi-epitope vaccine constructs (MEVCs) using the trRosetta server, which is renowned for its ability to predict protein structures with high accuracy. These models were then refined using the online web service Galaxy Refine. Further manipulation of improved structures was done with BIOVIA Discovery Studio Visualizer, from which the best models were selected for detailed study and analysis. The above approach ensures careful examination and enhancement of structural properties of vaccine constructs and provides a sound framework for future experimental and bioinformatics studies . Design 2.0 online platform-mediated disulfide engineering was used to enhance the stability of our vaccine formulation. We were able to incorporate strategically placed disulfide bonds into the formulation, which enhanced its structural integrity, leading to its robust resistance to degradation. It is expected that these disulfide bridges will greatly enhance the vaccine’s ability to induce a strong and durable response against the intended pathogen by preserving its optimal conformation. In the development of a multi-epitope vaccine targeting the hepatitis E virus, a comparable methodology of disulfide engineering was utilized to improve the stability of the vaccine that was constructed . Additionally, protein-protein docking was carried out using the designed whole-proteome vaccine construct (WP-MEVC) in combination with Human Toll-like Receptor-II (TLR-2) through the HADDOCK server. This docking study has enabled us to explore the interactions occurring between the vaccine construct and TLR-2 receptor-a crucial receptor in triggering the immune response. We got a top-ranked docked complex based on the docking score in PDB format for further studies. The particular interactions between the WP-MEVC and TLR-2 were investigated with the use of PDBsum, which provided useful data regarding different interaction types: hydrogen bonds, non-hydrogen bonds, and salt bridges. In this way, the methodology is connected to previous studies dealing with Bartonella bacilliformis, where a multi-epitope subunit vaccine was designed, and docking experiments with TLR-2 were accomplished . Similar docking methods have been reported in studies that focus on the design of multiepitope subunit vaccines against K . aerogenes and MERS-CoV . Such approaches underscore the impact of protein-protein docking as an integral part of vaccine development, helping provide insights into the interactions that may enhance the immunogenic nature of vaccine candidates. By establishing a strong interaction between the WP-MEVC and TLR-2, we aim at increasing the capability of the vaccine to provoke a strong immune response. Moreover, for the expression in the E. coli K-12 strain, all vaccination sequences were optimized using the JCat server for optimal expression. The codons for each of the vaccine constructs were optimized and reverse-translated into their corresponding nucleotides, which were then used to synthesize fragments for cloning purposes . Sequences were in silico cloned in the pET28a+ plasmid using the SnapGene tool, with the assistance of XcmI and PsaAI restriction enzymes. Similar cloning strategies have been reported in studies that pursue a vaccine against Pegivirus; in these cases, in- silico cloning was performed using the pET28a+ plasmid, which points to the suitability of the aforementioned vector for vaccine construct expression in E. coli. . In addition to codon optimization and cloning, we also explored each one of the developed vaccines as injectable antigens through in-silico immune simulations. The simulations conducted assessed the potential antigen-mediated human immune response activation, and the results showed that the designed vaccines may provoke a high immune response and have the potential to induce immunity. Another previous study engaged in the design of a novel multi-epitope vaccine against HCV infection also included in silico expression in E. coli, which was an important input for optimizing the vaccine construct to be expressed in a bacterial system and subsequently experimentally validated, and maybe produced . These findings align with a reported investigation that highlighted a similar approach in designing a peptide vaccine to enforce a humoral immune response against Campylobacter jejuni . By optimizing codon usage and performing in-silico cloning in the pET28a+ plasmid, we have ensured that our vaccine constructs can be effectively expressed in E . coli , combined with in-silico immune simulations, further supports their potential to induce potent, protective immune responses targeted against the indicated pathogen.
In this study, the subtractive proteomics approach was applied to accurately screen for potential protein targets suitable for vaccine design against Candida lusitaniae , a pathogenic fungus. Through this method, we identified four potent protein targets that could serve as effective candidates for vaccine development. For each of these target proteins, various computational techniques were employed to identify epitopes recognized by cytotoxic T lymphocytes (CTL), B cells, and helper T lymphocytes (HTL). These epitopes play crucial roles in eliciting immune responses against the pathogen. Additionally, we conducted in-depth analyses, including 3D modeling and evaluation of physiochemical properties. These analyses, along with consideration of other pertinent parameters, facilitated the development of a multi-epitope vaccine construct, aimed at maximizing immunogenicity and efficacy. Subsequently, the designed vaccine candidates were subjected to immunogenic analysis using immune simulation maps. The results from these simulations revealed significant levels of immune responses elicited by our vaccine candidates, suggesting their potential effectiveness in combating Candida lusitaniae infections. Based on our findings, it can be concluded that the designed whole proteome vaccine construct holds promise in inducing robust immune responses against Candida lusitaniae . This research represents a significant step forward in identifying epitopes for future vaccine development against this pathogen. However, further experimentation is warranted to validate the efficacy of the vaccine construct in real-time settings. Our work provides valuable insights and directions for future endeavors aimed at developing effective vaccine against Candida lusitaniae . Limitations of this study This vaccine is designed on computational basis in which we targeted specific proteins and selected their epitopes on the basis of in-silico analysis. So, there is need for the experimental validation in wet-lab experimentation. All of the properties including the physiochemical, antigenic, allergenic, for the design of vaccine construct further require the experimental procedures so that it may lead to design a real-time vaccine using the recombinant DNA technology. Furthermore, using this multi-epitope vaccine construct against Candida lusitaniae as a potential therapeutic, it will be necessary to justify the clinical illustrations by incorporating various experimental approaches including the expression and the purification of MEVC’s.
This vaccine is designed on computational basis in which we targeted specific proteins and selected their epitopes on the basis of in-silico analysis. So, there is need for the experimental validation in wet-lab experimentation. All of the properties including the physiochemical, antigenic, allergenic, for the design of vaccine construct further require the experimental procedures so that it may lead to design a real-time vaccine using the recombinant DNA technology. Furthermore, using this multi-epitope vaccine construct against Candida lusitaniae as a potential therapeutic, it will be necessary to justify the clinical illustrations by incorporating various experimental approaches including the expression and the purification of MEVC’s.
S1 File Amino acid sequence retrieved from UniProt. Online web servers incorporated in this research work. (DOCX) S2 File Reference proteome of C . lusitaniae . (DOCX)
|
Where Do Professional Sports Clubs and Organisations Sit Within a Local Health Promotion System? A Social Network Analysis Study | 34ec493e-8631-4d82-979a-c9bdd18d1057 | 11855522 | Health Promotion[mh] | Professional sports clubs and organisations (PSCOs) competing with elite sport structures or leagues often adopt corporate social responsibility practices to deliver on social or environmental outcomes . Whilst it has been demonstrated that PSCOs as settings can promote unhealthy behaviours, often through sponsorship or affiliation with the gambling or alcohol industry, corporate social responsibility practices often occur through the delivery of health promotion (HP) projects in local communities [ , , , , , ]. These HP projects commonly offer provisions for people throughout their lives, spanning physical activity (PA) promotion, weight management, mental health support, healthy diets, and many more [ , , , ]. Such projects are typically delivered by the PSCOs’ independent charitable organisation affiliated with the club brand, commonly known as ‘Community Trusts’, ‘Foundations’, or ‘[name of club] in the Community’ . The presence and capacity of PSCOs within communities has grown significantly over the last 25 years, particularly amongst football PSCOs, whereby ‘Football in the Community’ schemes began to shift from projects that were centred around boosting football participation, to projects that sought to contribute wider health and social policy objectives . The focus on PSCOs contributing to local communities was formally observed in 1986, whereby six ‘community Trusts’/‘foundations’ were founded as part of a pilot scheme developed between the Football League and the Professional Football Association . The initial intention was based upon a need to promote greater relationships between communities and clubs due to fan disorder; however, wider health and social objectives were soon integrated into community trusts and foundations’ work; and the increase in the number of organisations followed . Currently, there are 92 trusts or foundations across the English Football League (EFL) and Premier League, representing all clubs who compete in the elite structure in England and Wales. According to data from the 2021/2022 EFL season, EFL clubs delivered an estimated 579,712 hours of HP or education projects, reaching 840,094 people . In turn, demonstrating their ability to engage with community members at scale. Previous literature demonstrates that concepts such as fan-attachment to professional sports clubs and the power of PSCO branding can act as ‘pull’ factors which often attract participants to HP projects delivered by PSCOs [ , , ]. The resources and facilities that PSCOs can access and utilise, such as stadia, coaches linked to the club, and merchandise, can be useful assets for participant retention within HP projects [ , , , ]. Moreover, evidence further suggests that PSCOs can engage individuals who are at greater risk of poorer health outcomes or those who may not typically engage with health services, such as, but not limited to, men living with obesity or experiencing mental health conditions [ , , , , ]. Capturing and demonstrating the impact of ‘in-house’ (i.e., internally developed) projects has long been a considerable challenge for PSCOs; however, evidence on large-scale projects designed, supported, or evaluated by academic partners offers insight into the effectiveness of PSCOs as settings for HP [ , , ]. Specifically, the evaluation of the Premier League Health programme, a men’s HP project funded and delivered by the, at the time, English Premier League Foundation, reported significant increases in participants’ weekly physical activity and consumption of fruit and vegetables, alongside observing significant decreases in daily sitting time and weekly alcohol consumption . Furthermore, a meta-analysis of the Football Fans in Training intervention and the subsequent adapted interventions (studies n = 6) report a mean weight loss of −3.3 kg (95% confidence interval: −4.7 to −2.0) at 12 weeks in favour of the intervention , whilst also highlighting favourable results for physical activity, sedentary behaviour, dietary markers, psychological wellbeing, health-related quality of life, and cost effectiveness . Despite this, the role of PSCOs within local health and care strategies is unclear. Within the United Kingdom (U.K.), PSCOs linked to football clubs tend to have a greater financial and staffing capacity than PSCOs situated in other sports, often leading to higher levels of HP provision . Amongst these more resourced PSCOs, some have reported commonly working with local stakeholders such as Local Authorities or National Health Service bodies, such as integrated care boards; however, funding constraints are found to often limit project sustainability and, therefore, sustainable strategic collaboration with key local stakeholders [ , , ]. Subsequently, there is a considerable risk that PSCOs are not strongly embedded into local strategies for health despite their ability to support local community HP, in turn, becoming organisations that are underutilised and siloed within local systems. Given that the responsibility for public health in England was transferred from the National Health Service to Local Authorities in 2013 as part of the Health and Social Care Act in 2012, the exploration of collaboration between PSCOs and local authorities is warranted . Developing effective partnerships between the sport and health sector for local HP can be challenging, particularly when resources are scarce. However, there is an increasing need to explore how these sectors align due to the rising levels of intersectoral reliance required to meet national and local policy objectives within the U.K., such as those reported in Sport England’s Get Healthy, Get Active initiative, and Uniting the Movement Strategy . Integral to the delivery of such strategies are active partnerships, of which there are 43 within England . Active partnerships are independent not-for-profit organisations that aim to increase PA and reduce inactivity through place-based approaches in their locality . Often in receipt of significant funds from Sport England, active partnerships are integral to Sport England’s Uniting the Movement Strategy. Specifically, active partnerships are responsible for providing community-level PA data and insight to Sport England, and are essential in the implementation of the Uniting the Movement Strategy at a local level . Moreover, as alluded to, the growing interdependence between the health and sport sector suggests that Active Partnerships will require support from local organisations to deliver on these national strategic and policy goals . Drawing on the available evidence highlighting PSCOs’ ability to recruit participants at risk of future ill health, and the positive impact projects can have upon health outcomes, the exploration of PSCOs’ strategic role within multisectoral approaches to local HP is warranted. Given the need for such multisectoral collaboration in local HP, it is important to understand how local systems operate. In doing so, leverage points, key actors and areas for system development and improvement can be identified. Furthermore, the emerging evidence for the breadth and effectiveness of HP projects delivered by PSCOs, paired with the unique assets and mechanisms for community engagement they hold, underscores the importance of examining the connectedness of PSCOs within local HP systems. Context This research is set in the geographic boundaries of Bristol, North Somerset and South Gloucestershire (BNSSG), a region in the South West of the U.K., populated by approximately one million people . Stark differences in life expectancy mediated by deprivation are reported, whereby women and men living in the least deprived areas are predicted to live between 4.3 and 9.9 years longer than those living in the most deprived wards of the region, respectively, whilst approximately 25,000 children live in poverty . Furthermore, recent data suggest that 31.3% of children are obese or overweight, with higher rates identified amongst black and minority groups, and those living in the poorest areas of the region, whilst 63% of the adult population live with obesity or overweight . Such evidence offers insight regarding some of the health priorities, and inequalities, present within the region. To address these, there have been recent calls to action within strategies and reports to develop systems and networks that support multisectoral collaboration, co-production and shared-decision making, to improve health and support healthy behaviours at a community level; particularly for those most impacted by inequity and socioeconomic inequality . Moreover, the local Sport and Physical Activity Strategy further emphasises the need for partnership working, suggesting that identifying key partners and networks, and understanding how they can contribute to a system approach for PA promotion in local communities is a priority for the region . Within the region, there are five PSCOs delivering health promoting and educational projects for children and young people, adults, and older adults, all of which hold independent charitable status. Projects are delivered across stadia, schools, sport and leisure venues, and community venues within BNSSG, and mostly use multi-sport models, football, cricket, rugby or basketball to engage and deliver projects for community members. Projects include, but are not limited to, after-school clubs, holiday camps, hybrid classroom-PA sessions, weight management programmes, mental health programmes, inclusion and disability provision, and outsourced school physical education [ , , , ]. The five PSCOs operate from three stadia with capacities of 27,000, 11,000, and 8000, reflecting their ability to engage with, or advertise HP projects to, community members and fanbases on matchdays alone. Further opportunities for engaging with community members outside of PSCO stadia can be evidenced through partnerships with schools, whereby reports suggest one organisation partners with 40 primary schools across the region, recruiting over 8000 children to various HP projects and holiday camps throughout 2022/2023 . Moreover, within the region, a PSCO reports reaching over 8000 people in school, community and sport settings, across 16 programmes annually, indicating an age range of 5–104 years old, showcasing the ability of PSCOs within BNSSG to engage with communities in a variety of settings across the region . Given the reported health inequalities and priorities within BNSSG, the identified need for multisectoral collaboration to address these, and active role of five PSCOs operating throughout the region, we argue that this geographical region is of interest to research. Moreover, Bristol is a diversely populated area, and considered a major city within the U.K., yet is not considered a large city by area. In turn, this ensured the research was not only pragmatic and manageable, but externally valid to other cities within the U.K. Thus, the presence of PSCOs within communities, and the supporting evidence of PSCOs as enablers of HP activity, it is important to better understand the role that PSCOs could play within HP across the region for local policy and practice. Therefore, our study aims to explore the role PSCOs play within a local HP system, and to identify key actors within the system. To answer the aims of the study, we propose to address the following research questions: (i) Who identifies PSCOs as a key partner in the delivery of HP projects? (ii) With whom do PSCOs share information, knowledge, and resources? (iii) What is the perceived importance of PSCOs to the system from other organisations in it?
This research is set in the geographic boundaries of Bristol, North Somerset and South Gloucestershire (BNSSG), a region in the South West of the U.K., populated by approximately one million people . Stark differences in life expectancy mediated by deprivation are reported, whereby women and men living in the least deprived areas are predicted to live between 4.3 and 9.9 years longer than those living in the most deprived wards of the region, respectively, whilst approximately 25,000 children live in poverty . Furthermore, recent data suggest that 31.3% of children are obese or overweight, with higher rates identified amongst black and minority groups, and those living in the poorest areas of the region, whilst 63% of the adult population live with obesity or overweight . Such evidence offers insight regarding some of the health priorities, and inequalities, present within the region. To address these, there have been recent calls to action within strategies and reports to develop systems and networks that support multisectoral collaboration, co-production and shared-decision making, to improve health and support healthy behaviours at a community level; particularly for those most impacted by inequity and socioeconomic inequality . Moreover, the local Sport and Physical Activity Strategy further emphasises the need for partnership working, suggesting that identifying key partners and networks, and understanding how they can contribute to a system approach for PA promotion in local communities is a priority for the region . Within the region, there are five PSCOs delivering health promoting and educational projects for children and young people, adults, and older adults, all of which hold independent charitable status. Projects are delivered across stadia, schools, sport and leisure venues, and community venues within BNSSG, and mostly use multi-sport models, football, cricket, rugby or basketball to engage and deliver projects for community members. Projects include, but are not limited to, after-school clubs, holiday camps, hybrid classroom-PA sessions, weight management programmes, mental health programmes, inclusion and disability provision, and outsourced school physical education [ , , , ]. The five PSCOs operate from three stadia with capacities of 27,000, 11,000, and 8000, reflecting their ability to engage with, or advertise HP projects to, community members and fanbases on matchdays alone. Further opportunities for engaging with community members outside of PSCO stadia can be evidenced through partnerships with schools, whereby reports suggest one organisation partners with 40 primary schools across the region, recruiting over 8000 children to various HP projects and holiday camps throughout 2022/2023 . Moreover, within the region, a PSCO reports reaching over 8000 people in school, community and sport settings, across 16 programmes annually, indicating an age range of 5–104 years old, showcasing the ability of PSCOs within BNSSG to engage with communities in a variety of settings across the region . Given the reported health inequalities and priorities within BNSSG, the identified need for multisectoral collaboration to address these, and active role of five PSCOs operating throughout the region, we argue that this geographical region is of interest to research. Moreover, Bristol is a diversely populated area, and considered a major city within the U.K., yet is not considered a large city by area. In turn, this ensured the research was not only pragmatic and manageable, but externally valid to other cities within the U.K. Thus, the presence of PSCOs within communities, and the supporting evidence of PSCOs as enablers of HP activity, it is important to better understand the role that PSCOs could play within HP across the region for local policy and practice. Therefore, our study aims to explore the role PSCOs play within a local HP system, and to identify key actors within the system. To answer the aims of the study, we propose to address the following research questions: (i) Who identifies PSCOs as a key partner in the delivery of HP projects? (ii) With whom do PSCOs share information, knowledge, and resources? (iii) What is the perceived importance of PSCOs to the system from other organisations in it?
To address the research aims of this study, a social network analysis (SNA) was undertaken within the geographical boundary of BNSSG. SNA is a research technique often used to explore structural and relational dynamics of health systems and has been used in previous HP and physical activity research, reflecting how actors (e.g., individuals or organisations) interact . Directions (e.g., directed and undirected) and types (e.g., mutual) of relations between actors can also be captured, offering insight into real-world social structures and systems [ , , ]. 2.1. Participants A purposive sampling method was adopted to recruit participants. An initial invitation to participate was sent via email to members of a programme board for a collaborative HP programme delivered by multiple PSCOs in the locality. To recruit further, a snowballing technique, commonly used in SNA research, was utilised, whereby participants were able to provide contact details for key connections within partnering organisations . If participants responded with an expression of interest, a participant information sheet and a consent form was provided via email. Informed consent was provided by participants prior to taking part in the research. A total of 65 organisations operating within BNSSG were invited to participate in the study, covering PSCOs ( n = 5), public ( n = 9), education ( n = 3), health ( n = 5), sport ( n = 13), Voluntary Community, Faith and Social Enterprise (VCFSE) ( n = 10) organisations, and primary care networks ( n = 20). Due to the use of snowball sampling, some organisations listed by participants were not contacted due to falling outside of the geographical network boundary. 2.2. Data Collection Upon providing informed consent, participants were given access to an online survey (Microsoft Forms) that asked participants to list five-to-ten of the most important organisations they work with in delivering health and wellbeing projects across BNSSG. Participants were requested to provide a minimum of five organisations, and for each organisation listed were asked three multiple choice questions regarding the nature of relationship with that organisation. A free-text box was made available for each question to capture wider information. Additionally, participants were asked to specify how frequently they collaborated with each organisation. The survey was developed by JB and reviewed by JM and NT. Iterations included the refinement of questions and addition of free-text boxes. Participants were also invited to participate in a semi-structured interview, which is detailed and reported in a separate paper . 2.3. Data Analysis Data gathered via the SNA survey were firstly exported into Microsoft Excel whereby a (1-node) matrix was developed. Data were then cleaned to ensure organisations (nodes) were not duplicated in results. Following this, data were imported to UCINET for analysis whereby tests for centrality and geodesic distance were carried out, and figures were created to visualise findings . Centrality is the overarching term for various individual measures that describe the importance of an organisation/individual (node) within a network . Within our analyses, indegree, betweenness, and ineigenvector centrality were calculated, in addition to an ego network analysis for all PSCOs within the network, and geodesic distance calculations. Given the lack of previous research that has sought to explore PSCOs within HP networks, such measures were deemed the most appropriate to undertake this exploration of PSCOs connectivity, ability to influence other organisations, and the importance other organisations place upon them; offering initial insight concerning PSCOs position within the HP network, and opportunities to develop future partnerships . Indegree centrality provides insight around organisations (nodes) with the most inbound connections reported by other organisations (nodes) within a network or system . In essence, this demonstrates the ‘popularity’ of an organisation (node) . For example, if 10 organisations (nodes) identify Actor A as a partner, and only 5 organisations identify Actor B as a partner, Actor A will have greater indegree centrality. Betweenness centrality represents organisations (actors) within a network that are important for the sharing of information, knowledge or resources . This is calculated by identifying the shortest path, or route, between organisations (nodes), demonstrating the extent to which an organisation can act as a ‘go-between’ for other organisations not directly connected . In turn, highlighting how organisations (nodes) may mediate relationships between other organisations that are not directly connected . Ineigenvector centrality calculates the importance of an organisation (node) within a network, as determined by the relative importance of the organisations (node) that is connected to it . For example, if Actor A and Actor B are highly connected organisations, and both are connected to Actor C, Actor C will have a high ineigenvector centrality. This reflects both the influence that an organisation (node) can have within a network, whether this be on practice or policy, alongside the value that other organisations (node) place upon it. To further explore the direct relationships between PSCOs with the network, and PSCOs’ direct relationships with other organisations within the network, an ego network analysis was undertaken . Geodesic distance was calculated as a measure of cohesion, demonstrating the shortest length of path between two organisations (nodes) within the network, indicating connectivity of a network as a whole . 2.4. Ethics Ethical approval for this study was obtained from the University of Bristol’s School for Policy Studies Research Ethics Committee (REF:16169).
A purposive sampling method was adopted to recruit participants. An initial invitation to participate was sent via email to members of a programme board for a collaborative HP programme delivered by multiple PSCOs in the locality. To recruit further, a snowballing technique, commonly used in SNA research, was utilised, whereby participants were able to provide contact details for key connections within partnering organisations . If participants responded with an expression of interest, a participant information sheet and a consent form was provided via email. Informed consent was provided by participants prior to taking part in the research. A total of 65 organisations operating within BNSSG were invited to participate in the study, covering PSCOs ( n = 5), public ( n = 9), education ( n = 3), health ( n = 5), sport ( n = 13), Voluntary Community, Faith and Social Enterprise (VCFSE) ( n = 10) organisations, and primary care networks ( n = 20). Due to the use of snowball sampling, some organisations listed by participants were not contacted due to falling outside of the geographical network boundary.
Upon providing informed consent, participants were given access to an online survey (Microsoft Forms) that asked participants to list five-to-ten of the most important organisations they work with in delivering health and wellbeing projects across BNSSG. Participants were requested to provide a minimum of five organisations, and for each organisation listed were asked three multiple choice questions regarding the nature of relationship with that organisation. A free-text box was made available for each question to capture wider information. Additionally, participants were asked to specify how frequently they collaborated with each organisation. The survey was developed by JB and reviewed by JM and NT. Iterations included the refinement of questions and addition of free-text boxes. Participants were also invited to participate in a semi-structured interview, which is detailed and reported in a separate paper .
Data gathered via the SNA survey were firstly exported into Microsoft Excel whereby a (1-node) matrix was developed. Data were then cleaned to ensure organisations (nodes) were not duplicated in results. Following this, data were imported to UCINET for analysis whereby tests for centrality and geodesic distance were carried out, and figures were created to visualise findings . Centrality is the overarching term for various individual measures that describe the importance of an organisation/individual (node) within a network . Within our analyses, indegree, betweenness, and ineigenvector centrality were calculated, in addition to an ego network analysis for all PSCOs within the network, and geodesic distance calculations. Given the lack of previous research that has sought to explore PSCOs within HP networks, such measures were deemed the most appropriate to undertake this exploration of PSCOs connectivity, ability to influence other organisations, and the importance other organisations place upon them; offering initial insight concerning PSCOs position within the HP network, and opportunities to develop future partnerships . Indegree centrality provides insight around organisations (nodes) with the most inbound connections reported by other organisations (nodes) within a network or system . In essence, this demonstrates the ‘popularity’ of an organisation (node) . For example, if 10 organisations (nodes) identify Actor A as a partner, and only 5 organisations identify Actor B as a partner, Actor A will have greater indegree centrality. Betweenness centrality represents organisations (actors) within a network that are important for the sharing of information, knowledge or resources . This is calculated by identifying the shortest path, or route, between organisations (nodes), demonstrating the extent to which an organisation can act as a ‘go-between’ for other organisations not directly connected . In turn, highlighting how organisations (nodes) may mediate relationships between other organisations that are not directly connected . Ineigenvector centrality calculates the importance of an organisation (node) within a network, as determined by the relative importance of the organisations (node) that is connected to it . For example, if Actor A and Actor B are highly connected organisations, and both are connected to Actor C, Actor C will have a high ineigenvector centrality. This reflects both the influence that an organisation (node) can have within a network, whether this be on practice or policy, alongside the value that other organisations (node) place upon it. To further explore the direct relationships between PSCOs with the network, and PSCOs’ direct relationships with other organisations within the network, an ego network analysis was undertaken . Geodesic distance was calculated as a measure of cohesion, demonstrating the shortest length of path between two organisations (nodes) within the network, indicating connectivity of a network as a whole .
Ethical approval for this study was obtained from the University of Bristol’s School for Policy Studies Research Ethics Committee (REF:16169).
3.1. Descriptive Statistics Eighteen participants completed the SNA survey, resulting in a network of 90 organisations spanning the sport/PA (n = 6), public ( n = 2), and VCFSE (n = 3) sectors, alongside primary care networks (PCNs) (n = 3), and PSCOs (n = 4) (see ). It is important to highlight that PSCOs have been categorised separately due to the scope of this study; however, it should be acknowledged they are positioned in both the sport/PA and VCFSE sector. Moreover, education sector organisations such as schools, colleges and universities, have been distinguished from public sector organisations, whilst the overlap, specifically in relation to schools, is acknowledged, for the purpose of analysis they have been separated to explore connections between different settings. 3.2. Indegree Centrality: Who Identifies PSCOs as a Key Partner in the Delivery of HP Projects? The findings for indegree centrality are visualised in , whereby the larger-size icons (node size) reflect a higher level of indegree centrality (i.e., the greater the size, the greater the number of inbound connections). Our analysis suggests that within the network, reported connections to PSCOs within the sample were limited, and therefore they were not organisations that stakeholders viewed as key partners. Limited ties were highlighted between PSCOs within the network, whilst multisectoral ties also appeared low; particularly amongst VCFSE organisations and PCNs. Where connections were reported, CIMSPA, the national body for professional and organisational development in the sport and PA sector, identified multiple PSCOs ( n = 3/5) as key partners. Moreover, Bristol City Council (Local Authority), a highly connected node, reported the Bristol Sport Foundation (PSCO), who utilise a multi-sport model to provide sport and PA opportunities to children throughout the region, as a key partner. When examining who PSCOs reported as key partners, ties with governing bodies who may provide guidance and funding to PSCOs, such as Sport England and EFL Trust, were commonly identified. To visualise these specific connections with and between PSCOs, an ego network analysis was undertaken (see ) whereby arrowheads represent an inbound connection. This represents the directional connections between the Bristol Sport Foundation, Bristol City Robins Foundation, Bristol Bears Foundation, Bristol Rovers Community Trust, and the Bath Rugby Foundation, and other organisations in the network. 3.3. Betweenness Centrality: With Whom Do PSCOs Share Information, Knowledge, and Resources? Generally, PSCOs within this network were not identified as holding a high level of betweenness centrality, suggesting they may not be influential in the dissemination of materials or knowledge. However, Bristol Sport Foundation reported one of the highest levels of betweenness centrality in the network, and a greater level than that of other PSCOs such as Bristol City Robins Foundation, Bristol Bears Foundation, and the Bristol Rovers Community Trust; suggesting they may have a considerable role within this network in relation to sharing information, knowledge, and resources (see ). Data suggest Wesport (Active Partnership) and Bristol City Council (Local Authority) are key organisations for dissemination of resources, materials, and knowledge within this network. As the Active Partnership and one of the Local Authorities for the region, they will likely hold knowledge and information surrounding national priorities and directions of travel for sport and physical activity, whilst also holding responsibility for driving place-based agendas for sport and PA, both of which will be of significant importance to PSCOs. In turn, the tie between Bristol City Council and Bristol Sport Foundation may be of particular importance. Moreover, in light of Bristol Sport Foundation’s ties with both Bristol City Robins Foundation and Bristol Bears Community Foundation, who utilise football and rugby union as a vehicle for HP, respectively, the higher degree of betweenness centrality reflects Bristol Sport Foundations potential capacity to share resources, information or knowledge via partnering PSCOs to wider stakeholders. 3.4. Ineigenvector Centrality: What Is the Perceived Importance of PSCOs to the System from Other Organisations in It? Findings for ineigenvector centrality are visualised in , whereby larger icon sizes depict a higher degree of centrality. Findings suggest that Bristol Sport Foundation and Bristol City Robins Foundation (PSCOs) hold moderate ineigenvector centrality, demonstrating that highly connected and influential organisations regarded them as important partners within the network. Furthermore, consistent with other measures of centrality, Wesport (active partnership) and Bristol City Council (local authority) are shown to be highly important, whilst there is also an increased level of importance identified amongst VCFSE organisations such as the healthy living centres, Age U.K., and Sirona Health and Care. The importance of PCNs is also highlighted within this measure, reflecting their value within this network. Moreover, the importance of public sector organisations such as BNSSG ICB and North Somerset Council are also implied. However, despite the recognised importance of VCFSE organisations and PCNs, the data report that there are limited ties between PSCOs and such organisations. Given the value placed upon VCFSE organisations and PCNs by highly influential organisations such as the local authority and active partnership, further exploration of how PSCOs might be able to collaborate with such organisations, and for what benefit, may be of interest. 3.5. Geodesic Distance To explore the connectivity of the network, calculations for geodesic distance were conducted. An average geodesic distance of 2.966 within this network suggests limited connectivity within the network, reinforcing the influence and importance of key organisations within this network demonstrated by betweenness and ineigenvector centrality calculations ( and ).
Eighteen participants completed the SNA survey, resulting in a network of 90 organisations spanning the sport/PA (n = 6), public ( n = 2), and VCFSE (n = 3) sectors, alongside primary care networks (PCNs) (n = 3), and PSCOs (n = 4) (see ). It is important to highlight that PSCOs have been categorised separately due to the scope of this study; however, it should be acknowledged they are positioned in both the sport/PA and VCFSE sector. Moreover, education sector organisations such as schools, colleges and universities, have been distinguished from public sector organisations, whilst the overlap, specifically in relation to schools, is acknowledged, for the purpose of analysis they have been separated to explore connections between different settings.
The findings for indegree centrality are visualised in , whereby the larger-size icons (node size) reflect a higher level of indegree centrality (i.e., the greater the size, the greater the number of inbound connections). Our analysis suggests that within the network, reported connections to PSCOs within the sample were limited, and therefore they were not organisations that stakeholders viewed as key partners. Limited ties were highlighted between PSCOs within the network, whilst multisectoral ties also appeared low; particularly amongst VCFSE organisations and PCNs. Where connections were reported, CIMSPA, the national body for professional and organisational development in the sport and PA sector, identified multiple PSCOs ( n = 3/5) as key partners. Moreover, Bristol City Council (Local Authority), a highly connected node, reported the Bristol Sport Foundation (PSCO), who utilise a multi-sport model to provide sport and PA opportunities to children throughout the region, as a key partner. When examining who PSCOs reported as key partners, ties with governing bodies who may provide guidance and funding to PSCOs, such as Sport England and EFL Trust, were commonly identified. To visualise these specific connections with and between PSCOs, an ego network analysis was undertaken (see ) whereby arrowheads represent an inbound connection. This represents the directional connections between the Bristol Sport Foundation, Bristol City Robins Foundation, Bristol Bears Foundation, Bristol Rovers Community Trust, and the Bath Rugby Foundation, and other organisations in the network.
Generally, PSCOs within this network were not identified as holding a high level of betweenness centrality, suggesting they may not be influential in the dissemination of materials or knowledge. However, Bristol Sport Foundation reported one of the highest levels of betweenness centrality in the network, and a greater level than that of other PSCOs such as Bristol City Robins Foundation, Bristol Bears Foundation, and the Bristol Rovers Community Trust; suggesting they may have a considerable role within this network in relation to sharing information, knowledge, and resources (see ). Data suggest Wesport (Active Partnership) and Bristol City Council (Local Authority) are key organisations for dissemination of resources, materials, and knowledge within this network. As the Active Partnership and one of the Local Authorities for the region, they will likely hold knowledge and information surrounding national priorities and directions of travel for sport and physical activity, whilst also holding responsibility for driving place-based agendas for sport and PA, both of which will be of significant importance to PSCOs. In turn, the tie between Bristol City Council and Bristol Sport Foundation may be of particular importance. Moreover, in light of Bristol Sport Foundation’s ties with both Bristol City Robins Foundation and Bristol Bears Community Foundation, who utilise football and rugby union as a vehicle for HP, respectively, the higher degree of betweenness centrality reflects Bristol Sport Foundations potential capacity to share resources, information or knowledge via partnering PSCOs to wider stakeholders.
Findings for ineigenvector centrality are visualised in , whereby larger icon sizes depict a higher degree of centrality. Findings suggest that Bristol Sport Foundation and Bristol City Robins Foundation (PSCOs) hold moderate ineigenvector centrality, demonstrating that highly connected and influential organisations regarded them as important partners within the network. Furthermore, consistent with other measures of centrality, Wesport (active partnership) and Bristol City Council (local authority) are shown to be highly important, whilst there is also an increased level of importance identified amongst VCFSE organisations such as the healthy living centres, Age U.K., and Sirona Health and Care. The importance of PCNs is also highlighted within this measure, reflecting their value within this network. Moreover, the importance of public sector organisations such as BNSSG ICB and North Somerset Council are also implied. However, despite the recognised importance of VCFSE organisations and PCNs, the data report that there are limited ties between PSCOs and such organisations. Given the value placed upon VCFSE organisations and PCNs by highly influential organisations such as the local authority and active partnership, further exploration of how PSCOs might be able to collaborate with such organisations, and for what benefit, may be of interest.
To explore the connectivity of the network, calculations for geodesic distance were conducted. An average geodesic distance of 2.966 within this network suggests limited connectivity within the network, reinforcing the influence and importance of key organisations within this network demonstrated by betweenness and ineigenvector centrality calculations ( and ).
Our study aimed to assess the HP system across a South West region of the U.K., exploring the role played by PSCOs and identifying key stakeholders. Our findings suggested that the network as a whole held limited connectivity, emphasising the importance of these organisations within the network. Contextually, qualitative data suggest a historical issue of organisations competing for funding within this network, particularly amongst PSCOs, creating a culture that has restricted collaboration and knowledge sharing . To overcome this, and to develop a more connected and collaborative system, data suggest a commitment to strategic and joint-up thinking at a system level, and working towards local system health priorities, are essential. Whilst recent investment by Sport England under the Local Delivery Pilot and subsequent place partnership initiatives are intended to promote collaboration, and it is a clear ambition within Sport England’s Uniting the Movement Strategy, our data highlight a degree of fragmentation within this system . PSCOs themselves were typically not identified as well-connected or influential organisations within this network of local community HP, despite the evidence base suggesting they could be a considerable resource for effective HP delivery within local communities [ , , , , ]. Furthermore, PSCOs are often viewed as community anchor organisations by their fanbases and local residents, whilst also holding the ability to utilise assets such as stadia and branding to engage with communities, furthering the argument for their role within local HP systems [ , , ]. To our understanding, very little published research has explored the role of PSCOs within local HP system, particularly from a SNA perspective. Qualitative findings from Pringle and colleagues report that PSCOs often communicate and collaborate with local system stakeholders, yet rarely have clearly defined roles with local strategies and struggle to maintain partnerships once initial partnership funding ceases [ , , , ]. Viewed in light of our results, PSCOs may be recognised as partners within HP systems, yet our findings suggest that within this network, regardless of PSCOs capacity to act as significant delivery partners and their ability to engage with communities, they are not widely identified as influential or key organisations by system partners. In turn, the lack of connectivity within the network may demonstrate that PSCOs are to a certain degree siloed in their work, and arguably are somewhat underutilised as resources or assets within the system. Despite the lack of connectivity of PSCOs within this network, an increasing interdependency between sport and PA providers, and the health sector exists . Specifically, it is proposed that whilst central and local government, as well as health sector organisations, may hold more power over strategic directions, commissioning and establishing priorities, community sport organisations, such as PSCOs, are essential in determining whether they are able to achieve their policy and strategic objectives . Whilst there is evidence of effective partnership working between PSCOs and statutory health services/providers, whereby PSCOs have been integrated into local healthcare pathways, our findings suggest that within this network such partnerships appear to be lacking . If PSCOs are to be viewed as underutilised, future practice should explore how the ‘pull’ factor of PSCOs and the delivery capacity can be maximised, not only within this network but nationally. Within this network, understanding organisational strengths, weaknesses and assets, and working towards shared priorities, will be essential for more effective and sustainable efforts to promote local community health . Similar recommendations for multisectoral collaboration within local HP efforts have been reported, suggesting that agreement upon objectives, roles and responsibilities must be reached by all stakeholders to facilitate effective collaboration, and to develop future partnership working . We propose that local decision makers and commissioners should explore how PSCOs’ capacity to deliver projects in, and engage with, local communities can be better utilised within systems approaches to HP, to support addressing local health priorities outlined within strategic documents, such as those produced by Local Authorities’ Public Health departments or local National Health Service bodies. An example of such opportunity within this network can be found through local calls for whole-system approaches to obesity prevention in response to high rates of obesity reported in this geography [ , , ]. The exploration of PSCOs role as not only providers of community PA and sport, but also their ability to reduce unhealthy behaviours or commodities within stadia may be of interest to addressing such aims . Moreover, local policymakers, decision makers, and practitioners across sectors should collaboratively develop local priorities for HP and develop a clearer understanding of organisational strengths and gaps within the network. Bristol City Council (local authority) and Wesport (active partnership) were identified as the most connected and influential organisations within the network, demonstrating their significant role and responsibility for HP. Local authorities are often leaders, or at least central, to many local and regional networks and partnerships in health and care, whilst also having the capacity to influence services concerning the wider determinants of health, such as housing or transport . Leader organisations within systems and networks hold responsibility for bringing organisations together, identifying shared goals and benefits, and aligning work to national objectives and priorities . Whilst this is a challenging task, evidence suggests it can be an effective mechanism for shifting from traditionally adopted medical models of health, into more sustainable systems approaches with devolved power and leadership, which, given the increasing need for collaboration between the health and sport sector, may be required in this context [ , , ]. Sport England’s Uniting the Movement Strategy and associated 2022–2025 Implementation Plan highlight these organisations as crucial in facilitating local, collaborative place-based approaches to improving health and reducing inequalities through PA and sport . Moreover, within Sport England’s commitment to expand their place-based partnerships, active partnerships play an essential role in fostering and maintaining long-term strategic partnerships with organisations with ‘trust’, ‘credibility’ and ‘connections’ within communities . Utilising the ‘pull’ of PSCOs, namely branding and community presence, therefore may be highly influential in active partnerships’ ability to support the delivery of PA promotion strategies. However, our findings suggest that partnerships between PSCOs, local authorities and active partnerships may be underdeveloped. Therefore, to support effective delivery of this strategy, particularly around the engagement of those at risk of future ill-health and those who do not traditionally engage with services, we propose that PSCOs may be important organisations nationally for active partnerships and local authorities to consider their relationships with and should both explore how PSCOs could support local and national PA and sport, and wider HP goals. Limitations When considering our findings, it is important to acknowledge the limitations of the study. Firstly, due to the focus of the paper on PSCOs, and the specific locality of the network, findings are based upon a small local network, and therefore may not be generalisable to other parts of the U.K. However, our findings offer insight for future practice and development of the network within this locality, whilst also offering recommendations for practice, and a replicable model for research to take place, not only in other Sport England-funded local delivery pilot areas, but local area approaches to HP nationally and internationally. Moreover, the limitations of SNA as a method should be acknowledged, whereby a near-to-completed network consisting of all active members should be sought . Whilst we did not receive responses from all 90 organisations included in the sample, 80% of PSCOs within the boundary provided survey responses. Moreover, we did not receive survey responses from organisations within the education sector and received few responses from PCNs and other health providers, thus limiting the scope of the network from these perspectives. In attempts to negate this limitation, multiple attempts were made to contact all organisations listed as key partners within survey responses utilising contacts provided within survey responses, publicly available methods (e.g., enquiry inboxes and social media platforms), and community of practice networks. However, not all organisations responded, and some declined the opportunity to participate, possibly due to lack of time, capacity, or perceived lack of relevancy. It should therefore be considered that our findings may not be a true representation of the network as a whole, but do offer novel insight regarding PSCOs position within the network, alongside recommendations for future practice and research both within and beyond this network. Future research should consider alternative, or supplementary, approaches to data collection, such as stakeholder mapping workshops or focus groups, to ensure representation across all sectors. Given the sample within our study is not powered to fully explore potential healthcare or school-based opportunities for PSCOs, such approaches that may boost recruitment may help to identify future collaboration between PSCOs and the health and education sectors.
When considering our findings, it is important to acknowledge the limitations of the study. Firstly, due to the focus of the paper on PSCOs, and the specific locality of the network, findings are based upon a small local network, and therefore may not be generalisable to other parts of the U.K. However, our findings offer insight for future practice and development of the network within this locality, whilst also offering recommendations for practice, and a replicable model for research to take place, not only in other Sport England-funded local delivery pilot areas, but local area approaches to HP nationally and internationally. Moreover, the limitations of SNA as a method should be acknowledged, whereby a near-to-completed network consisting of all active members should be sought . Whilst we did not receive responses from all 90 organisations included in the sample, 80% of PSCOs within the boundary provided survey responses. Moreover, we did not receive survey responses from organisations within the education sector and received few responses from PCNs and other health providers, thus limiting the scope of the network from these perspectives. In attempts to negate this limitation, multiple attempts were made to contact all organisations listed as key partners within survey responses utilising contacts provided within survey responses, publicly available methods (e.g., enquiry inboxes and social media platforms), and community of practice networks. However, not all organisations responded, and some declined the opportunity to participate, possibly due to lack of time, capacity, or perceived lack of relevancy. It should therefore be considered that our findings may not be a true representation of the network as a whole, but do offer novel insight regarding PSCOs position within the network, alongside recommendations for future practice and research both within and beyond this network. Future research should consider alternative, or supplementary, approaches to data collection, such as stakeholder mapping workshops or focus groups, to ensure representation across all sectors. Given the sample within our study is not powered to fully explore potential healthcare or school-based opportunities for PSCOs, such approaches that may boost recruitment may help to identify future collaboration between PSCOs and the health and education sectors.
Our study aimed to explore where PSCOs fit within local HP delivery and strategy via SNA. Our findings report that PSCOs were not well connected or integrated organisations within this network, nor did they hold a significant level of influence amongst other organisations. Moreover, the connectivity of the network as a whole was limited, and therefore places considerable responsibility upon local authorities and the local active partnership to develop a more cohesive network. Despite the ability of PSCOs to deliver a high provision of HP projects, their status within communities, and their capacity to engage members, their role within local delivery and strategy appears unclear and somewhat siloed. Future practice should engage in further stakeholder and asset mapping within networks such as (1) exploring the potential role, delivery and assets PSCOs could offer a system, (2) enabling best use of limited resources in tackling local and national health priorities, and (3) advocating for national guidance and support for facilitating strategic joint-up working, such as funding opportunities that promote partnerships and collaboration locally.
|
Microbial Diversity and Enzyme Activity as Indicators of Permethrin-Exposed Soil Health | d9457070-bb88-49e2-acdb-9f4b5b11aeab | 10301950 | Microbiology[mh] | Soil, one of the most important natural resources, is the landscape’s inherent component. It undergoes modifications over time, while storing and converting energy and matter . According to the Natural Resources Conservation Service—USDA , soil health is defined as ‘the continued capacity of soil to function as a vital living ecosystem that sustains plants, animals, and humans’. Healthy soil ensures bountiful yields, clean waters and healthy forests. It plays a key role in production of safe food; it is also vital for sustainable and eco-friendly development and for nature conservation. Soil degradation and loss of soil quality may give rise to economic decline and social unrest . Fertile soils are characterized by high microbiological activity . One gram of soil can contain from hundreds of millions to billions of microorganisms . Microorganisms interact with soil components, play a vital role in the biogeochemical cycle of elements and in the promotion of the growth and development of plants by supplying them with nutrients and phytohormones while inhibiting the development of pathogens . The biomass of microorganisms depends on many factors, such as temperature, moisture, oxygen content, pH, methods of crop cultivation, genotypes of plants, the development of pathogens, the pressure of heavy metals and plant protection chemicals [ , , , ]. A crop cultivation technology incompatible with good agricultural practice, e.g., striving to maximize production at the lowest costs, can disrupt the soil homeostasis or relations between microorganisms and plant roots, thereby creating an environment that does not favor the development of plants . According to simulations discussed in , the total sales of agricultural chemicals will increase in 2030 by ca. 22% relative to 2021, and will achieve the value of USD 279.12 billion. The global consumption of pesticides in agriculture in 2020 reached nearly 2.7 million metric tons, which corresponded to over 57% of the amount used in 1990; this total quantity included 606,000 tons of fungicides and bactericides as well as 471,000 tons of insecticides. Permethrin, classified as one of pyrethroids, is a synthetic-organic insecticides . According to the Toxics Release Inventory (TRI) created by the Emergency Planning and Community Right-to-Know Act (EPCRA), supported by the U.S. Environmental Protection Agency (EPA), permethrin is a toxic substance hazardous to the environment. It has been registered by EPA since 1979, and is sold in many products, e.g., for fogging and pest control. The EPA report (2023) states that permethrin is used over an area of 39 million acres in the USA to control mosquitoes [ , , , ]. According to the data displayed on the website of the Environmental Protection Agency, the Food and Drug Administration, permethrin is most often used in the urban landscape [ , , , ]. Due to their lipophilic character, permethrin and its derivates can bioaccumulate in water , sediments , soils and in organisms exposed to these substances . According to the European Environment Agency (EEA) 2022 and FAO and UNEP (2021), our knowledge about the accumulation of persistent organic pollutants (POP), including insecticides, herbicides or fungicides in agricultural soils, is constantly being enriched. This progress is stimulated by the development of novel research methods, which, for example, enable us to study the consequences of DNA damage in invertebrates or to trace changes in the soil microbiome . In line with Directive 2009/128/EC, establishing a framework for Community action to achieve the sustainable use of pesticides, the European Commission’s strategy of ‘from field to table’ assumes that the use of agrichemicals should be reduced by 50% by year 2030. The need to examine the effect of permethrin on soil microorganisms, soil biochemical activity and growth of plants, and hence its impact on soil health, is further confirmed [ , , , , ] since soil microorganisms, by participating in geochemical processes, play several functions in maintaining the soil’s structure. However, it should be emphasized that the biodegradation of pyrethroids is, to a large extent, related to their isomeric selectivity . Pyrethroids possess 1–3 chiral centres and 2–8 stereoisomers, with the presence of chiral carbon atoms responsible for their enantioselective degradation . Pyrethroids undergo biodegradation in the hydrolysis of the central ester bond, catalysed by carboxylesterase, the potential of which depends on the catalytic triad: glutamine, histidine and serine. The intermediates in the degradation of permethrin are, respectively: cyclopropane carboxylic acid, 3-phenoxybenzyl alcohol, 3-phenoxybenzaldehyde (PBAld), 1,2-benzenedicarboxylic acid or 1,2-benzenedicarboxylic butyl decyl ester . The efficiency of permethrin degradation depends, to a large extent, on soil properties such as: moisture, soil texture, organic matter content, pH and temperature. Due to the lipophilic properties of pyrethroids, both organic matter and clay content control their bioavailability to microorganisms. In turn, the processes of adsorption and desorption of these compounds are significantly affected by pH and soil moisture . The controlled insect species, after applying permethrin, may, as in the case of other insecticides, develop many different defence mechanisms that allow them to survive . These mechanisms can be divided into physiological mechanisms involving changes in the rate of permeation and transport across membranes, biochemical ones consisting of changing or increasing detoxification metabolism, and behavioural ones consisting of avoiding the lethal dose of the insecticide used by the insect. All these resistance mechanisms are genetically determined and controlled by appropriate genes [ , , , ]. Permethrin, used several times during the season to control pests, mainly ticks, cockroaches or pharaoh ants, through continuous contact can lead to the permanent multiplication of insecticide-metabolizing bacteria, which is particularly important for the development and intensification of mosquito resistance to insecticides . Maintaining proper relationships between the physical, chemical and biological properties of soil is fundamental to the proper quality of soil, which is crucial for life on our planet . All of these three categories of soil characteristics are largely dependent on the content of organic matter in soil, which determines the soil’s biodiversity. Organic matter creates the base of the so-called soil food web , associated with the release of nutrients by microorganisms. There are different indicators that serve to evaluate the productivity and fertility of soil, but Doran and Zeiss underline how difficult it is to develop such indices. According to the strategy of the LUCAS module of Soil Biodiversity and Pesticides , the determination of the biological diversity of soil can be achieved, for example, by sequencing specific DNA regions extracted and amplified for any type of an environmental sample. The aim of our study has been to present simple indicators for evaluation of the quality of soil exposed to the pressure of permethrin, a third generation insecticide. To achieve this aim, metagenomic and biochemical assays of soil were made. The effect of the application of permethrin on the growth and development of Zea mays , on the diversity of bacteria and fungi and on the activity of soil enzymes was examined.
2.1. The Reaction of Bacteria and Fungi to Permethrin 2.1.1. Non-Cultured Bacteria The monitoring of the soil’s biological diversity through the sequencing of 16S DNA amplicons showed that from 99.3% to 99.7% of sequences belonged to the kingdom Bacteria. In all soil samples, unsown and sown with Zea mays , the phyla Actinobacteria and Proteobacteria dominated among the 29 types. Another eight dominant types of bacteria, representing ≥ 1% of all acquired sequences, were the phyla Gemmatimonadetes , Acidobacteria , Chloroflexi , Firmicutes , Planctomycetes , Bacteroidetes , Verrucomicrobia and TM7 ( a,c). The cultivation of Zea mays (sC_uC) contributed to a decrease in the relative abundance of Proteobacteria by 10.8% and an increase in the relative abundance of Actinobacteria by 3.8% ( c). The pollution of unsown soil with permethrin (uC_uP) decreased the relative abundance of Actinobacteria by 6.5% and increased the relative abundance of Proteobacteria by 4.1%. Permethrin, when applied to the soil, which was cropped with Zea mays (sC_sP), decreased the relative abundance of Proteobacteria by 3.8% but did not considerably affect the abundance of Actinobacteria . The cultivation of Zea mays on soil polluted with permethrin increased the relative abundance of Actinobacteria by 9.5% and decreased Proteobacteria by 11.1% (uP_sP). The most frequently present classes of bacteria were Actinobacteria and Thermoleophilia of the phylum Actinobacteria ; Alphaproteobacteria , Gammaproteobacteria and Betaproteobacteria of the phylum Proteobacteria ; and Gemmatimonadetes of the phylum Gemmatimonadetes ( b). Once the sequences were assigned to subsequent taxonomic levels, it emerged that the orders Actinomycetales , Sphingomonadales and Xanthomonadales dominated in all analyzed soils ( a). Taking into account OUT ≥ 1%, the order Actinomycetales was represented by Promicromonosporaceae, Nocardioidaceae , Intrasporangiaceae and Micrococcaceae , the order Sphingomonadales was represented by Sphingomonadaceae, and the order Xanthomonadales was represented by Xanthomonadaceae ( b). Regardless of the application of permethrin or sowing of Zea mays , the dominant bacteria in soil were the ones of the genera: Cellulosimicrobium classified to the family Promicromonosporaceae , order Actinomycetales , class Actinobacteria , phylum Actinobacteria ; Kaistobacter classified to the family Sphingomonadaceae , order Sphingomonadales , class Alphaproteobacteria , phylum Proteobacteria ; and Sphingomonas classified to the family Sphingomonadaceae , order Sphingomonadales , class Alphaproteobacteria , phylum Proteobacteria ( a). After obtaining OTU data ≥ 1% at the genus level in all soil samples, the relative abundance data of bacterial genera indicated that the cultivation of Zea mays (sC_uC) contributed the most to a decrease in the abundance of bacteria of the genus Cellulosimicrobium (by 6.8%) and Sphingomonas (by 14.2%) and an increase in the abundance of Kaistobacter (by 9.6%) and Arthrobacter (by 8.4%). The application of permethrin (uC_uP) contributed the most to an increase in the abundance of bacteria of the genus Pseudomonas (by 13.9%) and a decrease in counts of Cellulosimicrobium and Sphingomonas (by 4.6%). The application of permethrin to soils sown with Zea mays (sC_sP) most significantly raised the relative abundance of Cellulosimicrobium (by 3.9%) and Rhodanobacter (by 2.3%), and decreased the relative abundance of Arthrobacter (by 3.3%) and Terracoccus (by 2.8%). The cultivation of Zea mays on soil polluted with permethrin (uP_sP) increased the relative abundance of bacteria Kaistobacter (by 9.9%), Arthrobacter (by 5%), Terracoccus and Rhodoplanes (by 3.6%), while decreasing the abundance of Pseudomonas (by 13.8%), Sphingomonas (by 8.3%) and Thermomonas (by 2.9%) ( c). The research results did not reveal any unique types of bacteria in the analyzed soils. In fact, all the identified species of bacteria comprised a shared microbiome of soils, polluted and unpolluted ones ( b). 2.1.2. Non-Cultured Fungi The metagenomic analysis of fungi led to the identification of 60.7% to 70.8% of sequences with the OTU number ≥ 1% as belonging to the kingdom Fungi. In all soil samples, both unsown and sown with Zea mays , the phylum Ascomycota dominated among the 13 types. Other dominant types of fungi were Basidiomycota , Mortierellomycota and Rozellomycota ( a). The cultivation of Zea mays (sC_uC) contributed to an increase in the abundance of Ascomycota (by 2.0%) and a decrease in the abundance of Rozellomycota (by 3.5%). The pollution of soil under Zea mays with permethrin (sC_sP) increased the abundance of the phylum Ascomycota (by 1.4%). In soil polluted with permethrin, the cultivation of Zea mays (uP_sP) raised the abundance of Ascomycota (by 5.2%) and decreased that of Rozellomycota (by 4.9%) ( c). Most sequences of Eurotiomycetes classified to the class of fungi were determined in unsown soils and in sown soils polluted with permethrin. Seeding soils with Zea mays had an unambiguously more positive effect on Leotiomycetes and Dothideomycetes ( b). Sequences of mold fungi classified to the order Sordariales were most abundant in sown soils, and those of the order Eurotiales in unsown soils ( a). Having assigned the sequences to the subsequent taxonomic levels, it was found that the dominant families of fungi were Chaetomiaceae of the order Sordariales , class Sordariomycetes , phylum Ascomycota, as well as Aspergillaceae which belong to the order Eurotiales , class Eurotiomycetes , type Ascomycota , with Chaetomiaceae (in 81–85%) dominating in soils under Zea mays , while Aspergillaceae (52–54%) dominated in unsown soils ( b). In soils sown with Zea mays , the dominant fungi were the ones of the genus Chaetomium classified to the family Chaetomiaceae ( a). After obtaining OTU data ≥ 1% at the fungal genus level in all soil samples, the relative abundance of fungal genera data indicated that the cultivation of Zea mays and application of permethrin contributed the most to the changes in the abundance of Botryotrichum , Chaetomium , Humicola , Penicillium and Trichoderma . Sowing the soils with Zea mays (sC_uC) increased the abundance of Chaetomium by 58.3% and Botryotrichum by 9.4%, but decreased the abundance of Penicillium by 53.1% and Humicola by 15.1% ( c). The application of permethrin to soils not sown with Zea mays (uC_uP) increased the abundance of fungi of the genus Botryotrichum by 5.4% and decreased the abundance of Humicola , Penicillium and Chaetomium by 2.3%, 1.7% and 1.4%, respectively. The application of permethrin to soils sown with Zea mays (sC_sP) increased the abundance of Chaetomium by 7.1% and decreased the abundance of Botryotrichum by 4.0%. The cultivation of Zea mays after the application of permethrin (uP_sP) increased the abundance of Chaetomium by 66.8% while decreasing the relative abundance of Penicillium by 50.1% and Humicola by 13.1% ( c). Similarly in the case of types of bacteria, it was impossible to distinguish a type of fungi unique in a given soil because all fungi comprised the core microbiome ( b). 2.2. Cultured Microorganisms The cultivation of Zea mays created suitable conditions for the development of organotrophic bacteria, actinomycetes, and fungi. Sowing the soils treated with permethrin raised the abundance of organotrophic bacteria by 51%, actinomycetes by 41% and fungi by 39%, on average, independent from the doses of permethrin. In the soil cropped with Zea mays , the presence of permethrin raised the counts of organotrophic bacteria in a range from 4% (10 mg permethrin) to 22% (40 mg permethrin); of actinomycetes from 9% (10 mg permethrin) to 48% (20 mg permethrin and 40 mg permethrin); and decreased the counts fungi from 32% (10 mg permethrin) to 74% (40 mg permethrin kg −1 d.m. of soil). In unsown soil, permethrin raised the counts of organotrophic bacteria from 37% (10 mg permethrin) to 58% (40 mg permethrin); actinomycetes from 5% (10 mg permethrin) to 65% (40 mg permethrin); and decreased the counts fungi from 30% (20 mg permethrin) to 35% (40 mg permethrin kg −1 d.m. of soil) ( ). The cultivation of Zea mays increased the average colony development (CD) indices calculated for organotrophic bacteria (by 24%), actinomycetes (by 55%) and fungi (by 8%). Considering the applied doses of the insecticide, it can be concluded that the most significant negative impact on the CD index of organotrophic bacteria and actinomycetes in sown soil was produced by the lowest applied dose (10 mg permethrin), which depressed it by 25% and 15%, respectively, while the biggest decrease in the CD indices for fungi was induced by the highest dose of permethrin (40 mg permethrin), which lowered the CD index calculated from these microorganisms by 22%. In unsown soils, the CD index of organotrophic bacteria was most adversely affected by the medium dose of permethrin (20 mg permethrin), while the response of actinomycetes was most distinctly negative to the highest dose (40 mg permethrin) and no negative effect of the applied permethrin doses on mold fungi was observed ( ). The ecophysiological diversity index (EP) showed that unsown soil was characterized by a higher diversity of organotrophic bacteria, while presenting lower diversity of actinomycetes and fungi ( ). The mean EP indices for organotrophic bacteria were within the range of 0.869 in unsown soil to 0.963 in soil sown with Zea mays, for actinomycetes—from 0.888 in unsown soil to 0.909 in sown soil, and for fungi—from 0.786 in unsown soil to 0.809 in sown soil. 2.3. Response of Soil Enzymes to Permethrin The application of permethrin in the lowest dose (10 mg kg −1 d.m. of soil) was not shown to have a negative influence on most of the biochemical properties of the soil ( ). Only the activity of acid phosphatase was significantly reduced in both unsown and sown Zea mays soil, as well as catalase in sown soil and β -glucosidase in unsown soil. The application of this preparation in an amount of 20 mg kg −1 d.m. of soil stimulated the activity of alkaline phosphatase and β -glucosidase in unsown soils as well as the activity of dehydrogenases, urease, alkaline phosphatase, acid phosphatase, in addition to which it raised the value of the biochemical soil quality index (BA) in soils under Zea mays . The highest tested permethrin dose (40 mg kg −1 d.m. of soil) exerted a negative effect in both sown and unsown soil on the activity of soil enzymes in both unsown soil and soil sown with Zea mays , with the exception of acid phosphatase in sown soil. 2.4. Response of Zea mays to Permethrin Permethrin proved to be non-toxic to the test plant. Permethrin did not significantly decrease the yield of Zea mays nor did it lower the greenness indices that SPAD (ang. Soil and Plant Analysis Development) determined for Zea mays in the fourth and sixth leaf stage ( A,B). In brief, the growth and development of the test plant and the process of photosynthesis were undisturbed.
2.1.1. Non-Cultured Bacteria The monitoring of the soil’s biological diversity through the sequencing of 16S DNA amplicons showed that from 99.3% to 99.7% of sequences belonged to the kingdom Bacteria. In all soil samples, unsown and sown with Zea mays , the phyla Actinobacteria and Proteobacteria dominated among the 29 types. Another eight dominant types of bacteria, representing ≥ 1% of all acquired sequences, were the phyla Gemmatimonadetes , Acidobacteria , Chloroflexi , Firmicutes , Planctomycetes , Bacteroidetes , Verrucomicrobia and TM7 ( a,c). The cultivation of Zea mays (sC_uC) contributed to a decrease in the relative abundance of Proteobacteria by 10.8% and an increase in the relative abundance of Actinobacteria by 3.8% ( c). The pollution of unsown soil with permethrin (uC_uP) decreased the relative abundance of Actinobacteria by 6.5% and increased the relative abundance of Proteobacteria by 4.1%. Permethrin, when applied to the soil, which was cropped with Zea mays (sC_sP), decreased the relative abundance of Proteobacteria by 3.8% but did not considerably affect the abundance of Actinobacteria . The cultivation of Zea mays on soil polluted with permethrin increased the relative abundance of Actinobacteria by 9.5% and decreased Proteobacteria by 11.1% (uP_sP). The most frequently present classes of bacteria were Actinobacteria and Thermoleophilia of the phylum Actinobacteria ; Alphaproteobacteria , Gammaproteobacteria and Betaproteobacteria of the phylum Proteobacteria ; and Gemmatimonadetes of the phylum Gemmatimonadetes ( b). Once the sequences were assigned to subsequent taxonomic levels, it emerged that the orders Actinomycetales , Sphingomonadales and Xanthomonadales dominated in all analyzed soils ( a). Taking into account OUT ≥ 1%, the order Actinomycetales was represented by Promicromonosporaceae, Nocardioidaceae , Intrasporangiaceae and Micrococcaceae , the order Sphingomonadales was represented by Sphingomonadaceae, and the order Xanthomonadales was represented by Xanthomonadaceae ( b). Regardless of the application of permethrin or sowing of Zea mays , the dominant bacteria in soil were the ones of the genera: Cellulosimicrobium classified to the family Promicromonosporaceae , order Actinomycetales , class Actinobacteria , phylum Actinobacteria ; Kaistobacter classified to the family Sphingomonadaceae , order Sphingomonadales , class Alphaproteobacteria , phylum Proteobacteria ; and Sphingomonas classified to the family Sphingomonadaceae , order Sphingomonadales , class Alphaproteobacteria , phylum Proteobacteria ( a). After obtaining OTU data ≥ 1% at the genus level in all soil samples, the relative abundance data of bacterial genera indicated that the cultivation of Zea mays (sC_uC) contributed the most to a decrease in the abundance of bacteria of the genus Cellulosimicrobium (by 6.8%) and Sphingomonas (by 14.2%) and an increase in the abundance of Kaistobacter (by 9.6%) and Arthrobacter (by 8.4%). The application of permethrin (uC_uP) contributed the most to an increase in the abundance of bacteria of the genus Pseudomonas (by 13.9%) and a decrease in counts of Cellulosimicrobium and Sphingomonas (by 4.6%). The application of permethrin to soils sown with Zea mays (sC_sP) most significantly raised the relative abundance of Cellulosimicrobium (by 3.9%) and Rhodanobacter (by 2.3%), and decreased the relative abundance of Arthrobacter (by 3.3%) and Terracoccus (by 2.8%). The cultivation of Zea mays on soil polluted with permethrin (uP_sP) increased the relative abundance of bacteria Kaistobacter (by 9.9%), Arthrobacter (by 5%), Terracoccus and Rhodoplanes (by 3.6%), while decreasing the abundance of Pseudomonas (by 13.8%), Sphingomonas (by 8.3%) and Thermomonas (by 2.9%) ( c). The research results did not reveal any unique types of bacteria in the analyzed soils. In fact, all the identified species of bacteria comprised a shared microbiome of soils, polluted and unpolluted ones ( b). 2.1.2. Non-Cultured Fungi The metagenomic analysis of fungi led to the identification of 60.7% to 70.8% of sequences with the OTU number ≥ 1% as belonging to the kingdom Fungi. In all soil samples, both unsown and sown with Zea mays , the phylum Ascomycota dominated among the 13 types. Other dominant types of fungi were Basidiomycota , Mortierellomycota and Rozellomycota ( a). The cultivation of Zea mays (sC_uC) contributed to an increase in the abundance of Ascomycota (by 2.0%) and a decrease in the abundance of Rozellomycota (by 3.5%). The pollution of soil under Zea mays with permethrin (sC_sP) increased the abundance of the phylum Ascomycota (by 1.4%). In soil polluted with permethrin, the cultivation of Zea mays (uP_sP) raised the abundance of Ascomycota (by 5.2%) and decreased that of Rozellomycota (by 4.9%) ( c). Most sequences of Eurotiomycetes classified to the class of fungi were determined in unsown soils and in sown soils polluted with permethrin. Seeding soils with Zea mays had an unambiguously more positive effect on Leotiomycetes and Dothideomycetes ( b). Sequences of mold fungi classified to the order Sordariales were most abundant in sown soils, and those of the order Eurotiales in unsown soils ( a). Having assigned the sequences to the subsequent taxonomic levels, it was found that the dominant families of fungi were Chaetomiaceae of the order Sordariales , class Sordariomycetes , phylum Ascomycota, as well as Aspergillaceae which belong to the order Eurotiales , class Eurotiomycetes , type Ascomycota , with Chaetomiaceae (in 81–85%) dominating in soils under Zea mays , while Aspergillaceae (52–54%) dominated in unsown soils ( b). In soils sown with Zea mays , the dominant fungi were the ones of the genus Chaetomium classified to the family Chaetomiaceae ( a). After obtaining OTU data ≥ 1% at the fungal genus level in all soil samples, the relative abundance of fungal genera data indicated that the cultivation of Zea mays and application of permethrin contributed the most to the changes in the abundance of Botryotrichum , Chaetomium , Humicola , Penicillium and Trichoderma . Sowing the soils with Zea mays (sC_uC) increased the abundance of Chaetomium by 58.3% and Botryotrichum by 9.4%, but decreased the abundance of Penicillium by 53.1% and Humicola by 15.1% ( c). The application of permethrin to soils not sown with Zea mays (uC_uP) increased the abundance of fungi of the genus Botryotrichum by 5.4% and decreased the abundance of Humicola , Penicillium and Chaetomium by 2.3%, 1.7% and 1.4%, respectively. The application of permethrin to soils sown with Zea mays (sC_sP) increased the abundance of Chaetomium by 7.1% and decreased the abundance of Botryotrichum by 4.0%. The cultivation of Zea mays after the application of permethrin (uP_sP) increased the abundance of Chaetomium by 66.8% while decreasing the relative abundance of Penicillium by 50.1% and Humicola by 13.1% ( c). Similarly in the case of types of bacteria, it was impossible to distinguish a type of fungi unique in a given soil because all fungi comprised the core microbiome ( b).
The monitoring of the soil’s biological diversity through the sequencing of 16S DNA amplicons showed that from 99.3% to 99.7% of sequences belonged to the kingdom Bacteria. In all soil samples, unsown and sown with Zea mays , the phyla Actinobacteria and Proteobacteria dominated among the 29 types. Another eight dominant types of bacteria, representing ≥ 1% of all acquired sequences, were the phyla Gemmatimonadetes , Acidobacteria , Chloroflexi , Firmicutes , Planctomycetes , Bacteroidetes , Verrucomicrobia and TM7 ( a,c). The cultivation of Zea mays (sC_uC) contributed to a decrease in the relative abundance of Proteobacteria by 10.8% and an increase in the relative abundance of Actinobacteria by 3.8% ( c). The pollution of unsown soil with permethrin (uC_uP) decreased the relative abundance of Actinobacteria by 6.5% and increased the relative abundance of Proteobacteria by 4.1%. Permethrin, when applied to the soil, which was cropped with Zea mays (sC_sP), decreased the relative abundance of Proteobacteria by 3.8% but did not considerably affect the abundance of Actinobacteria . The cultivation of Zea mays on soil polluted with permethrin increased the relative abundance of Actinobacteria by 9.5% and decreased Proteobacteria by 11.1% (uP_sP). The most frequently present classes of bacteria were Actinobacteria and Thermoleophilia of the phylum Actinobacteria ; Alphaproteobacteria , Gammaproteobacteria and Betaproteobacteria of the phylum Proteobacteria ; and Gemmatimonadetes of the phylum Gemmatimonadetes ( b). Once the sequences were assigned to subsequent taxonomic levels, it emerged that the orders Actinomycetales , Sphingomonadales and Xanthomonadales dominated in all analyzed soils ( a). Taking into account OUT ≥ 1%, the order Actinomycetales was represented by Promicromonosporaceae, Nocardioidaceae , Intrasporangiaceae and Micrococcaceae , the order Sphingomonadales was represented by Sphingomonadaceae, and the order Xanthomonadales was represented by Xanthomonadaceae ( b). Regardless of the application of permethrin or sowing of Zea mays , the dominant bacteria in soil were the ones of the genera: Cellulosimicrobium classified to the family Promicromonosporaceae , order Actinomycetales , class Actinobacteria , phylum Actinobacteria ; Kaistobacter classified to the family Sphingomonadaceae , order Sphingomonadales , class Alphaproteobacteria , phylum Proteobacteria ; and Sphingomonas classified to the family Sphingomonadaceae , order Sphingomonadales , class Alphaproteobacteria , phylum Proteobacteria ( a). After obtaining OTU data ≥ 1% at the genus level in all soil samples, the relative abundance data of bacterial genera indicated that the cultivation of Zea mays (sC_uC) contributed the most to a decrease in the abundance of bacteria of the genus Cellulosimicrobium (by 6.8%) and Sphingomonas (by 14.2%) and an increase in the abundance of Kaistobacter (by 9.6%) and Arthrobacter (by 8.4%). The application of permethrin (uC_uP) contributed the most to an increase in the abundance of bacteria of the genus Pseudomonas (by 13.9%) and a decrease in counts of Cellulosimicrobium and Sphingomonas (by 4.6%). The application of permethrin to soils sown with Zea mays (sC_sP) most significantly raised the relative abundance of Cellulosimicrobium (by 3.9%) and Rhodanobacter (by 2.3%), and decreased the relative abundance of Arthrobacter (by 3.3%) and Terracoccus (by 2.8%). The cultivation of Zea mays on soil polluted with permethrin (uP_sP) increased the relative abundance of bacteria Kaistobacter (by 9.9%), Arthrobacter (by 5%), Terracoccus and Rhodoplanes (by 3.6%), while decreasing the abundance of Pseudomonas (by 13.8%), Sphingomonas (by 8.3%) and Thermomonas (by 2.9%) ( c). The research results did not reveal any unique types of bacteria in the analyzed soils. In fact, all the identified species of bacteria comprised a shared microbiome of soils, polluted and unpolluted ones ( b).
The metagenomic analysis of fungi led to the identification of 60.7% to 70.8% of sequences with the OTU number ≥ 1% as belonging to the kingdom Fungi. In all soil samples, both unsown and sown with Zea mays , the phylum Ascomycota dominated among the 13 types. Other dominant types of fungi were Basidiomycota , Mortierellomycota and Rozellomycota ( a). The cultivation of Zea mays (sC_uC) contributed to an increase in the abundance of Ascomycota (by 2.0%) and a decrease in the abundance of Rozellomycota (by 3.5%). The pollution of soil under Zea mays with permethrin (sC_sP) increased the abundance of the phylum Ascomycota (by 1.4%). In soil polluted with permethrin, the cultivation of Zea mays (uP_sP) raised the abundance of Ascomycota (by 5.2%) and decreased that of Rozellomycota (by 4.9%) ( c). Most sequences of Eurotiomycetes classified to the class of fungi were determined in unsown soils and in sown soils polluted with permethrin. Seeding soils with Zea mays had an unambiguously more positive effect on Leotiomycetes and Dothideomycetes ( b). Sequences of mold fungi classified to the order Sordariales were most abundant in sown soils, and those of the order Eurotiales in unsown soils ( a). Having assigned the sequences to the subsequent taxonomic levels, it was found that the dominant families of fungi were Chaetomiaceae of the order Sordariales , class Sordariomycetes , phylum Ascomycota, as well as Aspergillaceae which belong to the order Eurotiales , class Eurotiomycetes , type Ascomycota , with Chaetomiaceae (in 81–85%) dominating in soils under Zea mays , while Aspergillaceae (52–54%) dominated in unsown soils ( b). In soils sown with Zea mays , the dominant fungi were the ones of the genus Chaetomium classified to the family Chaetomiaceae ( a). After obtaining OTU data ≥ 1% at the fungal genus level in all soil samples, the relative abundance of fungal genera data indicated that the cultivation of Zea mays and application of permethrin contributed the most to the changes in the abundance of Botryotrichum , Chaetomium , Humicola , Penicillium and Trichoderma . Sowing the soils with Zea mays (sC_uC) increased the abundance of Chaetomium by 58.3% and Botryotrichum by 9.4%, but decreased the abundance of Penicillium by 53.1% and Humicola by 15.1% ( c). The application of permethrin to soils not sown with Zea mays (uC_uP) increased the abundance of fungi of the genus Botryotrichum by 5.4% and decreased the abundance of Humicola , Penicillium and Chaetomium by 2.3%, 1.7% and 1.4%, respectively. The application of permethrin to soils sown with Zea mays (sC_sP) increased the abundance of Chaetomium by 7.1% and decreased the abundance of Botryotrichum by 4.0%. The cultivation of Zea mays after the application of permethrin (uP_sP) increased the abundance of Chaetomium by 66.8% while decreasing the relative abundance of Penicillium by 50.1% and Humicola by 13.1% ( c). Similarly in the case of types of bacteria, it was impossible to distinguish a type of fungi unique in a given soil because all fungi comprised the core microbiome ( b).
The cultivation of Zea mays created suitable conditions for the development of organotrophic bacteria, actinomycetes, and fungi. Sowing the soils treated with permethrin raised the abundance of organotrophic bacteria by 51%, actinomycetes by 41% and fungi by 39%, on average, independent from the doses of permethrin. In the soil cropped with Zea mays , the presence of permethrin raised the counts of organotrophic bacteria in a range from 4% (10 mg permethrin) to 22% (40 mg permethrin); of actinomycetes from 9% (10 mg permethrin) to 48% (20 mg permethrin and 40 mg permethrin); and decreased the counts fungi from 32% (10 mg permethrin) to 74% (40 mg permethrin kg −1 d.m. of soil). In unsown soil, permethrin raised the counts of organotrophic bacteria from 37% (10 mg permethrin) to 58% (40 mg permethrin); actinomycetes from 5% (10 mg permethrin) to 65% (40 mg permethrin); and decreased the counts fungi from 30% (20 mg permethrin) to 35% (40 mg permethrin kg −1 d.m. of soil) ( ). The cultivation of Zea mays increased the average colony development (CD) indices calculated for organotrophic bacteria (by 24%), actinomycetes (by 55%) and fungi (by 8%). Considering the applied doses of the insecticide, it can be concluded that the most significant negative impact on the CD index of organotrophic bacteria and actinomycetes in sown soil was produced by the lowest applied dose (10 mg permethrin), which depressed it by 25% and 15%, respectively, while the biggest decrease in the CD indices for fungi was induced by the highest dose of permethrin (40 mg permethrin), which lowered the CD index calculated from these microorganisms by 22%. In unsown soils, the CD index of organotrophic bacteria was most adversely affected by the medium dose of permethrin (20 mg permethrin), while the response of actinomycetes was most distinctly negative to the highest dose (40 mg permethrin) and no negative effect of the applied permethrin doses on mold fungi was observed ( ). The ecophysiological diversity index (EP) showed that unsown soil was characterized by a higher diversity of organotrophic bacteria, while presenting lower diversity of actinomycetes and fungi ( ). The mean EP indices for organotrophic bacteria were within the range of 0.869 in unsown soil to 0.963 in soil sown with Zea mays, for actinomycetes—from 0.888 in unsown soil to 0.909 in sown soil, and for fungi—from 0.786 in unsown soil to 0.809 in sown soil.
The application of permethrin in the lowest dose (10 mg kg −1 d.m. of soil) was not shown to have a negative influence on most of the biochemical properties of the soil ( ). Only the activity of acid phosphatase was significantly reduced in both unsown and sown Zea mays soil, as well as catalase in sown soil and β -glucosidase in unsown soil. The application of this preparation in an amount of 20 mg kg −1 d.m. of soil stimulated the activity of alkaline phosphatase and β -glucosidase in unsown soils as well as the activity of dehydrogenases, urease, alkaline phosphatase, acid phosphatase, in addition to which it raised the value of the biochemical soil quality index (BA) in soils under Zea mays . The highest tested permethrin dose (40 mg kg −1 d.m. of soil) exerted a negative effect in both sown and unsown soil on the activity of soil enzymes in both unsown soil and soil sown with Zea mays , with the exception of acid phosphatase in sown soil.
Permethrin proved to be non-toxic to the test plant. Permethrin did not significantly decrease the yield of Zea mays nor did it lower the greenness indices that SPAD (ang. Soil and Plant Analysis Development) determined for Zea mays in the fourth and sixth leaf stage ( A,B). In brief, the growth and development of the test plant and the process of photosynthesis were undisturbed.
3.1. Response of Non-Cultured Bacteria and Fungi to Permethrin Innovations in the protection of the quality of soils and crops should take advantage of the role of microbial communities, which can be a key element in the maintenance of soil health . An evaluation of the quality of soil takes into account the biological diversity of organisms , and the biomass and activity of microorganisms and invertebrates [ , , ]. In the course of this study, the 16S metagenomic analysis enabled us to identify from 109,188 to 176,303 OTUs of sequences of bacteria, and from 67,296 to 252,879 OTUs of fungi. The least OTUs of bacteria and fungi were identified in soils unsown and without permethrin, while the highest ones were determined in soils sown with Zea mays and treated with permethrin. The soils in this study were mainly colonized by bacteria of the types Actinobacteria and Proteobacteria and fungi of the phyla Ascomycota and Basidiomycota . These types of microorganisms, most active in soils polluted with pesticides, have also been identified in other studies . According to Letourneau and Bothwell , a wide spectrum of pesticides contributes to the inhibition of harmful species. However, pesticides can also have an adverse impact on beneficial species. A selection induced by agrichemicals affects the competition among organisms in the soil environment, which consequently determines the values of the plant infestation indicators . In our study, permethrin present in unsown soils and in soils sown with Zea mays stimulated the multiplication of all identified types of bacteria. The biggest changes in the proportions of the abundance of bacteria in unsown soils were detected in terms of the OTUs of bacteria of the type Verrucomicrobia and fungi Rozellomycota . In soils sown with Zea mays , bacteria of the type Proteobacteria and fungi of the type Ascomycota were the least resistant to permethrin. Most probably the most active types of bacteria in soils polluted with pesticides participating in their degradation are the bacteria of the genera Pseudomonas sp., Stenotrophomonas sp. , Bacillus sp. , Serratia sp. , Acinetobacter sp. , Brevibacillus sp. and Sphingomonas sp. , which partly lends credence to the obtained research results. The use of pesticide-degrading bacteria is the most promising strategy for the remediation of a soil environment contaminated with pyrethroids [ , , ]. Regardless of the use of the soil and application of permethrin, our soils were colonized mainly by bacteria of the genus Cellulosimicrobium and fungi of the genus Chaetomium . Other microorganisms present in abundance were bacteria of the genera Kaistobacter , Sphingomonas , Thermomonas and fungi of the genus Penicillium. The bacteria which most probably decomposed permethrin in soil most effectively were the ones of the genera Cellulosimicrobium sp., Kaistobacter sp. and Sphingomonas sp. They appeared most numerously, which proves that they were most resistant to this pollutant. Our analysis of the soils not sown with Zea mays put the focus on the bacteria of the genus Pseudomonas , whose abundance increased by 100%. A significant increase in abundance was also noted for the bacteria of the genera Arthrobacter , Terracoccus , Phycicoccus and fungi Botryotrichum . In soils sown with Zea mays , bacteria of the genera Rhodanobacter , Devosia , Rhodoplanes , Thermomonas , Stenotrophomonas and fungi of the genera Iodophanus , Meyerozyma proved capable of removing the pollutant from soil. The metagenomic analysis allowed us to distinguish from 43,154 to 79,786 of sequences of bacteria ≥ 1% and from 58,524 to 216,065 of sequences of fungi. The smallest counts of assigned genera of bacteria were identified in soils sown with maize but not treated with permethrin (sC), while those of fungi—in soil not cropped with Zea mays without permethrin (uC). However, it should be emphasized that, the compilation of sowing the soil with Zea mays and 40 mg of permethrin kg −1 d.m. of soil contributed to reducing both the relative abundance of fungi and the development of their colonies, which was largely generated by the high dose of the applied insecticide. 3.2. Response of Cultured Microorganisms The improvement in the quality of soils consists mainly of raising the biomass of microorganisms . In our study, the counts of cultured organotrophic bacteria and actinomycetes increased as doses of permethrin were higher. It can therefore be concluded that most microorganisms present in the soil could decompose permethrin quite effectively because pyrethroids can serve as a source of carbon for bacteria . According to Imade and Babalola and Bhatt , besides having a basic source of carbon, microorganisms also require other nutrients that facilitate the initial adaptation of bacteria to the environment, to accelerate their growth and to improve their capacity to degrade insecticides. Bokade et al. concluded that strains of bacteria isolated from such an environment are helpful in the biomineralization of pollutants. The highest dose of the tested insecticide (40 mg permethrin) lowered the counts of fungi. Fungi are mainly acidophilic . Thus, a decline in pH may have been caused by the desorption of residues of pesticides adsorbed on colloidal surfaces . In the experiment reported in this article, the soil pollution with permethrin caused a moderate succession of microorganisms. It was only in the soil with the highest doses of the insecticides that a shift occurred between strategy k and strategy r microorganisms. Generally, the CD index reached higher values in soil cropped with Zea mays. Likewise, the EP index, which can assume values from 0 to 1, did not undergo drastic changes in response to the tested pyrethroid. Thus, it may be probable that the application of permethrin does not reduce the ecophysiological diversity of groups of microorganisms in soil. Microbial culture methods are commonly used, since the ability of microbial cultures to decompose organic compounds, sometimes toxic ones, to safer products does not adversely affect the quality of the soil environment [ , , , ]. Pesticide degradation by microorganisms usually proceeds in three stages: (I) the hydrolysis, oxidation or reduction of the primary compound; (II) the conjugation of the phase I metabolites with sugar or amino acids to increase their solubility in water and produce less toxic metabolites; and (III) the transformation of the phase II metabolites to secondary conjugates . Most probably, the participation of microorganisms in the carbon and other nutrient cycles contributed, in our study, to the decomposition of permethrin, while the participation of microorganisms in processes of elevating the solubility of substances provided resources in the form of nutrients essential for the growth of plants, similar to Zea mays in our experiment. 3.3. Response of Soil Enzymes and Zea mays to Permethrin The impact of insecticides on soil enzymes has not been thoroughly recognized yet . Hence, complex studies that enable observations of changes in populations of microorganisms and enzymatic activity in the natural environment are particularly valuable . Due to their structure, pyrethroids can be potentially hydrolyzed by carboxylesterase (EC 3.1.1.1) . According to Bhatt et al. , it is esterases, also known as pyrethroid hydrolases and belonging to α/β proteins, that are responsible for the degradation of pyrethroids in the environment. The bioelimination of pyrethroids typically leads to breaking the ester bonds and the formation of carboxyl acids and alcohols . Fang et al. maintain that enzymes isolated from strains capable of degrading pyrethroids are close to lipases and esterases, which proves that microorganisms and their enzymes, with an effective capability of performing hydrolysis, play a key role in the elimination of residual amounts of pyrethroids. Our research results confirmed the growing counts of soil microorganisms responsible for the cycles of basic nutrients, i.e., C, N, P and S, in soil . Pyrethroids are strongly bound to organic matter , which is of key importance for the maintenance of soil quality and productivity. The results of this study suggest that Zea mays can be used for the remediation of soils contaminated with pyrethroids. No negative effect of permethrin applied in doses from 10 to 40 mg on the growth of plants was detected. This could have been a consequence of plants being able to secrete pyrethroid-hydrolyzing enzymes . However, it should be borne in mind that the application of excessively large quantities of pyrethroids can lead to a decrease in the uptake of water and nutrients, inhibit the photosynthesis of plants and disturb the hormonal balance . According to Imade et al. , the positive effect of the grown plant Zea mays can be attributed to the plant’s increased secretion of organic compounds into soil.
Innovations in the protection of the quality of soils and crops should take advantage of the role of microbial communities, which can be a key element in the maintenance of soil health . An evaluation of the quality of soil takes into account the biological diversity of organisms , and the biomass and activity of microorganisms and invertebrates [ , , ]. In the course of this study, the 16S metagenomic analysis enabled us to identify from 109,188 to 176,303 OTUs of sequences of bacteria, and from 67,296 to 252,879 OTUs of fungi. The least OTUs of bacteria and fungi were identified in soils unsown and without permethrin, while the highest ones were determined in soils sown with Zea mays and treated with permethrin. The soils in this study were mainly colonized by bacteria of the types Actinobacteria and Proteobacteria and fungi of the phyla Ascomycota and Basidiomycota . These types of microorganisms, most active in soils polluted with pesticides, have also been identified in other studies . According to Letourneau and Bothwell , a wide spectrum of pesticides contributes to the inhibition of harmful species. However, pesticides can also have an adverse impact on beneficial species. A selection induced by agrichemicals affects the competition among organisms in the soil environment, which consequently determines the values of the plant infestation indicators . In our study, permethrin present in unsown soils and in soils sown with Zea mays stimulated the multiplication of all identified types of bacteria. The biggest changes in the proportions of the abundance of bacteria in unsown soils were detected in terms of the OTUs of bacteria of the type Verrucomicrobia and fungi Rozellomycota . In soils sown with Zea mays , bacteria of the type Proteobacteria and fungi of the type Ascomycota were the least resistant to permethrin. Most probably the most active types of bacteria in soils polluted with pesticides participating in their degradation are the bacteria of the genera Pseudomonas sp., Stenotrophomonas sp. , Bacillus sp. , Serratia sp. , Acinetobacter sp. , Brevibacillus sp. and Sphingomonas sp. , which partly lends credence to the obtained research results. The use of pesticide-degrading bacteria is the most promising strategy for the remediation of a soil environment contaminated with pyrethroids [ , , ]. Regardless of the use of the soil and application of permethrin, our soils were colonized mainly by bacteria of the genus Cellulosimicrobium and fungi of the genus Chaetomium . Other microorganisms present in abundance were bacteria of the genera Kaistobacter , Sphingomonas , Thermomonas and fungi of the genus Penicillium. The bacteria which most probably decomposed permethrin in soil most effectively were the ones of the genera Cellulosimicrobium sp., Kaistobacter sp. and Sphingomonas sp. They appeared most numerously, which proves that they were most resistant to this pollutant. Our analysis of the soils not sown with Zea mays put the focus on the bacteria of the genus Pseudomonas , whose abundance increased by 100%. A significant increase in abundance was also noted for the bacteria of the genera Arthrobacter , Terracoccus , Phycicoccus and fungi Botryotrichum . In soils sown with Zea mays , bacteria of the genera Rhodanobacter , Devosia , Rhodoplanes , Thermomonas , Stenotrophomonas and fungi of the genera Iodophanus , Meyerozyma proved capable of removing the pollutant from soil. The metagenomic analysis allowed us to distinguish from 43,154 to 79,786 of sequences of bacteria ≥ 1% and from 58,524 to 216,065 of sequences of fungi. The smallest counts of assigned genera of bacteria were identified in soils sown with maize but not treated with permethrin (sC), while those of fungi—in soil not cropped with Zea mays without permethrin (uC). However, it should be emphasized that, the compilation of sowing the soil with Zea mays and 40 mg of permethrin kg −1 d.m. of soil contributed to reducing both the relative abundance of fungi and the development of their colonies, which was largely generated by the high dose of the applied insecticide.
The improvement in the quality of soils consists mainly of raising the biomass of microorganisms . In our study, the counts of cultured organotrophic bacteria and actinomycetes increased as doses of permethrin were higher. It can therefore be concluded that most microorganisms present in the soil could decompose permethrin quite effectively because pyrethroids can serve as a source of carbon for bacteria . According to Imade and Babalola and Bhatt , besides having a basic source of carbon, microorganisms also require other nutrients that facilitate the initial adaptation of bacteria to the environment, to accelerate their growth and to improve their capacity to degrade insecticides. Bokade et al. concluded that strains of bacteria isolated from such an environment are helpful in the biomineralization of pollutants. The highest dose of the tested insecticide (40 mg permethrin) lowered the counts of fungi. Fungi are mainly acidophilic . Thus, a decline in pH may have been caused by the desorption of residues of pesticides adsorbed on colloidal surfaces . In the experiment reported in this article, the soil pollution with permethrin caused a moderate succession of microorganisms. It was only in the soil with the highest doses of the insecticides that a shift occurred between strategy k and strategy r microorganisms. Generally, the CD index reached higher values in soil cropped with Zea mays. Likewise, the EP index, which can assume values from 0 to 1, did not undergo drastic changes in response to the tested pyrethroid. Thus, it may be probable that the application of permethrin does not reduce the ecophysiological diversity of groups of microorganisms in soil. Microbial culture methods are commonly used, since the ability of microbial cultures to decompose organic compounds, sometimes toxic ones, to safer products does not adversely affect the quality of the soil environment [ , , , ]. Pesticide degradation by microorganisms usually proceeds in three stages: (I) the hydrolysis, oxidation or reduction of the primary compound; (II) the conjugation of the phase I metabolites with sugar or amino acids to increase their solubility in water and produce less toxic metabolites; and (III) the transformation of the phase II metabolites to secondary conjugates . Most probably, the participation of microorganisms in the carbon and other nutrient cycles contributed, in our study, to the decomposition of permethrin, while the participation of microorganisms in processes of elevating the solubility of substances provided resources in the form of nutrients essential for the growth of plants, similar to Zea mays in our experiment.
The impact of insecticides on soil enzymes has not been thoroughly recognized yet . Hence, complex studies that enable observations of changes in populations of microorganisms and enzymatic activity in the natural environment are particularly valuable . Due to their structure, pyrethroids can be potentially hydrolyzed by carboxylesterase (EC 3.1.1.1) . According to Bhatt et al. , it is esterases, also known as pyrethroid hydrolases and belonging to α/β proteins, that are responsible for the degradation of pyrethroids in the environment. The bioelimination of pyrethroids typically leads to breaking the ester bonds and the formation of carboxyl acids and alcohols . Fang et al. maintain that enzymes isolated from strains capable of degrading pyrethroids are close to lipases and esterases, which proves that microorganisms and their enzymes, with an effective capability of performing hydrolysis, play a key role in the elimination of residual amounts of pyrethroids. Our research results confirmed the growing counts of soil microorganisms responsible for the cycles of basic nutrients, i.e., C, N, P and S, in soil . Pyrethroids are strongly bound to organic matter , which is of key importance for the maintenance of soil quality and productivity. The results of this study suggest that Zea mays can be used for the remediation of soils contaminated with pyrethroids. No negative effect of permethrin applied in doses from 10 to 40 mg on the growth of plants was detected. This could have been a consequence of plants being able to secrete pyrethroid-hydrolyzing enzymes . However, it should be borne in mind that the application of excessively large quantities of pyrethroids can lead to a decrease in the uptake of water and nutrients, inhibit the photosynthesis of plants and disturb the hormonal balance . According to Imade et al. , the positive effect of the grown plant Zea mays can be attributed to the plant’s increased secretion of organic compounds into soil.
4.1. Soil Characterization This study was conducted on soil which, according to the International Union of Soil Sciences and the United States Department of Agriculture soil classification, represented loamy sand. The soil was sampled from the Olsztyn Lake District (NE Poland, 53.72° N, 20.42° E). In the natural state, this was proper brown soil. A more specific description of the soil is presented in . A detailed description of the methods and laboratory equipment used for completing physicochemical and chemical assays of soil can be found in our previous paper . 4.2. Permethrin Characterization Permethrin [3-(2,2-dichorovinyl)-2,2-dimethylcyclopropanecarboxylate] (number CAS: 52645-53-1), C 21 H 20 Cl 20 O, molecular weight −391.3 g mol −1 is a synthetic-organic chemical compound which belongs to pyrethroids . In this experiment, it was applied in the form of the preparation Aspermet 200 EC (Asplant-Skotniccy Sp. J, Jaworzno, Poland), which contains 200 g of active substance, permethrin (P), per 1 dm 3 . As recommended, the preparation should be applied as 1% aqueous solution, in a dose of 10 dm 3 of the solution per 200 m 2 of area. When used outdoors, the preparation should be prepared as a 5% solution. 4.3. Design of the Experiment The experiment was conducted in a greenhouse at the University of Warmia and Mazury in Olsztyn (Poland). The experimental variants were prepared in polyethylene pots with the capacity of 3.5 dm 3 . The following doses of permethrin were tested: 0 mg, 10 mg, 20 mg and 40 mg per 1 kg d.m. of soil. Having thoroughly mixed permethrin with soil, and after placing batches of soil in the pots, the soil moisture content was increased to 60% of water capacity. The control consisted of unpolluted soil. In order to gain better understanding of the effect of permethrin on the soil microbiome, the experiment was conducted in two series: (1) unsown soil and (2) soil sown with Zea mays var. LG 32.52 (a variety registered in the European Union). After germination, the maize plants were thinned to 4 plants per pot. Throughout the experiment, water was replenished 2–3 times a day to maintain the set constant moisture content. Each variant was set up with four replications. The experiment lasted 60 days (June–August 2020). The length of daylight at that time of year ranged from 15 h 13 min to 16 h 35 min. The average air temperature was 17.9 °C in June to 19.8 °C in August. The average relative sir humidity was 77% ( https://obserwator.imgw.pl ) (accessed on 8 September 2022). In the fourth leaf (B) and sixth leaf development stage (BBCH 19), according to the SPAD leaf greenness index (Soil and Plant Analysis Development), was determined with a Chlorophyll Meter 2900P SPAD 502 (KONICA MINOLTA, Inc., Chiyoda, Japan). In BBCH 51 stage (beginning of tassel emergence), the yield of aerial parts and roots of maize was determined, the plants were cut, fragmented and dried in a dryer type Binder D-78532 Tuttlingen, Germany at a temperature of 60 °C for four days. 4.4. Methods of Soil Microbiological Analysis 4.4.1. Breeding Microorganisms Isolation of microorganisms was conducted through a series of dilutions, according to the method described in our previous paper . Counts of microorganisms were determined as follows: organotrophic bacteria on Bunt and Roviry medium (1955) , actinomycetes on Kuster and Williams medium (1971), with addition of nystatin and antidyon (according to Parkinson 1971), and fungi on Martin medium (1950). All determinations were run in six replicates for each experimental object, all in moist soil. Microbial cultures were incubated in an incubator by Selecta Incudigit (Barcelona, Spain) at 28 °C for 10 days. Colony-forming units (c.f.u.) of microorganisms were presented per 1 kg −1 d.m. of soil. 4.4.2. Isolation of DNA and Identification of Bacteria and Fungi Using NGS Method Genomic Mini AX Bacteria+” (A&A Biotechnology, Gdynia, Poland) served for isolation of DNA from soil samples, while employing universal starters 1055F (5′-ACGGGCGGTGTGTAC-3′) and amplifying a fragment of the bacterial genes 16S rRNA and ITS. Detailed PCR settings were presented in our earlier papers . Sequencing of genetic material on the basis of the hypervariable region V3–V4 of the gene rRNA and the ITS1 fragment was carried out on a sequencer Illumina MiSeq (Genomed S.A. Warsaw, Poland). Primers 341F (5′-CCTACGGGNGGCWGCAG-3′), 785R (5′-GACTACHVGGGTATCTAATCC-3′) (Bacteria) and ITS1FI2 (5′-GAACCWGCGGARGGATCA-3′), 5.8S (5′-CGCTGCGTTCTTCATCG-3′) (Fungi) were used for amplification of the selected region. Sequences of bacteria and fungi were deposited in the GenBank NCBI under the access numbers: https://www.ncbi.nlm.nih.gov/nuccore/?term=OP914644:OP916021[accn] (accessed on 4 December 2022), https://www.ncbi.nlm.nih.gov/nuccore/?term=OP897054:OP897145[accn] (accessed on 2 December 2022), https://www.ncbi.nlm.nih.gov/nuccore/?term=OP978693:OP979103[accn] (accessed on 14 December 2022). 4.5. Biochemical Analysis of Soil Determinations of the activity of dehydrogenases (Deh), catalase (Cat), urease (Ure), alkaline phosphatase (Pal), acid phosphatase (Pac), arylsulfatase (Aryl) and β -glucosidase (Glu) were performed with the methods presented in the papers . The assays were carried out according to the methods by Öhlinger (1996), Johnson and Temple (1964) and by Alef and Nannipieri (1998). The assays for each research object were conducted in 3 replications, immediately after the soil samples were delivered to the laboratory. The activity of the analyzed enzymes was expressed in the following units: dehydrogenases µmol TFF kg −1 d.m. gleby h −1 , catalase—mol O 2 kg −1 d.m. gleby h −1 , urease—mmol N-NH 4 kg −1 d.m. gleby h −1 , alkaline phosphatase, acid phosphatase, arylsulfatase and β -glucosidase—mmol PNP kg −1 d.m. gleby h −1 . The activity of enzymes, except that of catalase, was determined on a spectrophotometer Perkin-Elmer Lambda 25 (Peabody, MA, USA). 4.6. Data Analysis and Statistical Processing On the basis of the counts of the above groups of microorganisms, the colony development (CD) index and the ecophysiological diversity (EP) index for the microorganisms were calculated. Following the guidelines of the formula proposed by Sarathchandra et al. , each day, the colony growth of the incubated groups of microorganisms was consistently counted over a period of 10 days. The data were processed statistically in Statistica 13.1 . Normality of distribution was verified with the Kruskal–Wallis test, and the results were submitted to the Duncan’s post hoc test. All data were displayed graphically, having eliminated OTUs lower than 1% in relation to the total number of OTUs. The types and genera of bacteria and fungi were statistically compared using the G-test (w/Yates’) + Fisher test, with the aid of the software STAMP 2.1.3 , and shown as heat maps in the software RStudio v1.2.5033 with the gplots library and R core . Classes and orders of bacteria and fungi were analyzed in a circular layout in a software package Circos 0.68 . For the visualization of unique data and shared genera of bacteria and fungi, the InteractiVenn software for analysis of sets was used .
This study was conducted on soil which, according to the International Union of Soil Sciences and the United States Department of Agriculture soil classification, represented loamy sand. The soil was sampled from the Olsztyn Lake District (NE Poland, 53.72° N, 20.42° E). In the natural state, this was proper brown soil. A more specific description of the soil is presented in . A detailed description of the methods and laboratory equipment used for completing physicochemical and chemical assays of soil can be found in our previous paper .
Permethrin [3-(2,2-dichorovinyl)-2,2-dimethylcyclopropanecarboxylate] (number CAS: 52645-53-1), C 21 H 20 Cl 20 O, molecular weight −391.3 g mol −1 is a synthetic-organic chemical compound which belongs to pyrethroids . In this experiment, it was applied in the form of the preparation Aspermet 200 EC (Asplant-Skotniccy Sp. J, Jaworzno, Poland), which contains 200 g of active substance, permethrin (P), per 1 dm 3 . As recommended, the preparation should be applied as 1% aqueous solution, in a dose of 10 dm 3 of the solution per 200 m 2 of area. When used outdoors, the preparation should be prepared as a 5% solution.
The experiment was conducted in a greenhouse at the University of Warmia and Mazury in Olsztyn (Poland). The experimental variants were prepared in polyethylene pots with the capacity of 3.5 dm 3 . The following doses of permethrin were tested: 0 mg, 10 mg, 20 mg and 40 mg per 1 kg d.m. of soil. Having thoroughly mixed permethrin with soil, and after placing batches of soil in the pots, the soil moisture content was increased to 60% of water capacity. The control consisted of unpolluted soil. In order to gain better understanding of the effect of permethrin on the soil microbiome, the experiment was conducted in two series: (1) unsown soil and (2) soil sown with Zea mays var. LG 32.52 (a variety registered in the European Union). After germination, the maize plants were thinned to 4 plants per pot. Throughout the experiment, water was replenished 2–3 times a day to maintain the set constant moisture content. Each variant was set up with four replications. The experiment lasted 60 days (June–August 2020). The length of daylight at that time of year ranged from 15 h 13 min to 16 h 35 min. The average air temperature was 17.9 °C in June to 19.8 °C in August. The average relative sir humidity was 77% ( https://obserwator.imgw.pl ) (accessed on 8 September 2022). In the fourth leaf (B) and sixth leaf development stage (BBCH 19), according to the SPAD leaf greenness index (Soil and Plant Analysis Development), was determined with a Chlorophyll Meter 2900P SPAD 502 (KONICA MINOLTA, Inc., Chiyoda, Japan). In BBCH 51 stage (beginning of tassel emergence), the yield of aerial parts and roots of maize was determined, the plants were cut, fragmented and dried in a dryer type Binder D-78532 Tuttlingen, Germany at a temperature of 60 °C for four days.
4.4.1. Breeding Microorganisms Isolation of microorganisms was conducted through a series of dilutions, according to the method described in our previous paper . Counts of microorganisms were determined as follows: organotrophic bacteria on Bunt and Roviry medium (1955) , actinomycetes on Kuster and Williams medium (1971), with addition of nystatin and antidyon (according to Parkinson 1971), and fungi on Martin medium (1950). All determinations were run in six replicates for each experimental object, all in moist soil. Microbial cultures were incubated in an incubator by Selecta Incudigit (Barcelona, Spain) at 28 °C for 10 days. Colony-forming units (c.f.u.) of microorganisms were presented per 1 kg −1 d.m. of soil. 4.4.2. Isolation of DNA and Identification of Bacteria and Fungi Using NGS Method Genomic Mini AX Bacteria+” (A&A Biotechnology, Gdynia, Poland) served for isolation of DNA from soil samples, while employing universal starters 1055F (5′-ACGGGCGGTGTGTAC-3′) and amplifying a fragment of the bacterial genes 16S rRNA and ITS. Detailed PCR settings were presented in our earlier papers . Sequencing of genetic material on the basis of the hypervariable region V3–V4 of the gene rRNA and the ITS1 fragment was carried out on a sequencer Illumina MiSeq (Genomed S.A. Warsaw, Poland). Primers 341F (5′-CCTACGGGNGGCWGCAG-3′), 785R (5′-GACTACHVGGGTATCTAATCC-3′) (Bacteria) and ITS1FI2 (5′-GAACCWGCGGARGGATCA-3′), 5.8S (5′-CGCTGCGTTCTTCATCG-3′) (Fungi) were used for amplification of the selected region. Sequences of bacteria and fungi were deposited in the GenBank NCBI under the access numbers: https://www.ncbi.nlm.nih.gov/nuccore/?term=OP914644:OP916021[accn] (accessed on 4 December 2022), https://www.ncbi.nlm.nih.gov/nuccore/?term=OP897054:OP897145[accn] (accessed on 2 December 2022), https://www.ncbi.nlm.nih.gov/nuccore/?term=OP978693:OP979103[accn] (accessed on 14 December 2022).
Isolation of microorganisms was conducted through a series of dilutions, according to the method described in our previous paper . Counts of microorganisms were determined as follows: organotrophic bacteria on Bunt and Roviry medium (1955) , actinomycetes on Kuster and Williams medium (1971), with addition of nystatin and antidyon (according to Parkinson 1971), and fungi on Martin medium (1950). All determinations were run in six replicates for each experimental object, all in moist soil. Microbial cultures were incubated in an incubator by Selecta Incudigit (Barcelona, Spain) at 28 °C for 10 days. Colony-forming units (c.f.u.) of microorganisms were presented per 1 kg −1 d.m. of soil.
Genomic Mini AX Bacteria+” (A&A Biotechnology, Gdynia, Poland) served for isolation of DNA from soil samples, while employing universal starters 1055F (5′-ACGGGCGGTGTGTAC-3′) and amplifying a fragment of the bacterial genes 16S rRNA and ITS. Detailed PCR settings were presented in our earlier papers . Sequencing of genetic material on the basis of the hypervariable region V3–V4 of the gene rRNA and the ITS1 fragment was carried out on a sequencer Illumina MiSeq (Genomed S.A. Warsaw, Poland). Primers 341F (5′-CCTACGGGNGGCWGCAG-3′), 785R (5′-GACTACHVGGGTATCTAATCC-3′) (Bacteria) and ITS1FI2 (5′-GAACCWGCGGARGGATCA-3′), 5.8S (5′-CGCTGCGTTCTTCATCG-3′) (Fungi) were used for amplification of the selected region. Sequences of bacteria and fungi were deposited in the GenBank NCBI under the access numbers: https://www.ncbi.nlm.nih.gov/nuccore/?term=OP914644:OP916021[accn] (accessed on 4 December 2022), https://www.ncbi.nlm.nih.gov/nuccore/?term=OP897054:OP897145[accn] (accessed on 2 December 2022), https://www.ncbi.nlm.nih.gov/nuccore/?term=OP978693:OP979103[accn] (accessed on 14 December 2022).
Determinations of the activity of dehydrogenases (Deh), catalase (Cat), urease (Ure), alkaline phosphatase (Pal), acid phosphatase (Pac), arylsulfatase (Aryl) and β -glucosidase (Glu) were performed with the methods presented in the papers . The assays were carried out according to the methods by Öhlinger (1996), Johnson and Temple (1964) and by Alef and Nannipieri (1998). The assays for each research object were conducted in 3 replications, immediately after the soil samples were delivered to the laboratory. The activity of the analyzed enzymes was expressed in the following units: dehydrogenases µmol TFF kg −1 d.m. gleby h −1 , catalase—mol O 2 kg −1 d.m. gleby h −1 , urease—mmol N-NH 4 kg −1 d.m. gleby h −1 , alkaline phosphatase, acid phosphatase, arylsulfatase and β -glucosidase—mmol PNP kg −1 d.m. gleby h −1 . The activity of enzymes, except that of catalase, was determined on a spectrophotometer Perkin-Elmer Lambda 25 (Peabody, MA, USA).
On the basis of the counts of the above groups of microorganisms, the colony development (CD) index and the ecophysiological diversity (EP) index for the microorganisms were calculated. Following the guidelines of the formula proposed by Sarathchandra et al. , each day, the colony growth of the incubated groups of microorganisms was consistently counted over a period of 10 days. The data were processed statistically in Statistica 13.1 . Normality of distribution was verified with the Kruskal–Wallis test, and the results were submitted to the Duncan’s post hoc test. All data were displayed graphically, having eliminated OTUs lower than 1% in relation to the total number of OTUs. The types and genera of bacteria and fungi were statistically compared using the G-test (w/Yates’) + Fisher test, with the aid of the software STAMP 2.1.3 , and shown as heat maps in the software RStudio v1.2.5033 with the gplots library and R core . Classes and orders of bacteria and fungi were analyzed in a circular layout in a software package Circos 0.68 . For the visualization of unique data and shared genera of bacteria and fungi, the InteractiVenn software for analysis of sets was used .
Permethrin, applied in doses from 10 to 40 mg kg −1 d.m. of soil, did not demonstrate any negative effect on the growth of Zea mays or on the plant’s greenness index. The metagenomic assays showed that the application of permethrin increases the abundance of Proteobacteria , but decreases that of Actinobacteria and Ascomycota. The application of permethrin increased, to the highest degree, the abundance of bacteria of the genera Cellulomonas , Kaistobacter , Pseudomonas and Rhodanobacter and fungi of the genera Penicillium , Humicola , Iodophanus and Meyerozyma . It has been discovered that permethrin stimulates the multiplication of organotrophic bacteria and actinomycetes, depresses the CD index and elevates the EP index of organotrophic bacteria and fungi, while increasing the CD and decreasing the EP of actinomycetes. Permethrin lowers the activity of all analyzed enzymes, and the soil’s biochemical activity index, in unsown soils. Microorganisms present in the topsoil, from 0 to 20 cm depth, following the application of permethrin, adapt to changes occurring in the soil environment. Sowing the soil with Zea mays alleviates the stress induced by the application of permethrin, which eventually leads to the restoration of the soil quality. The influence of pyrethroids on the quality of soil can be estimated by analyzing changes in the assemblages of the soil microflora.
|
Minimally invasive surgery and conservative treatment achieve similar clinical outcomes in patients with type II fragility fractures of the pelvis | ba10417e-faec-4d3c-a36f-2c9da95f6ef8 | 11869700 | Surgical Procedures, Operative[mh] | With the aging of the population, fragility fractures of the pelvic ring (FFP) are becoming increasingly common. By this year the overall increase in all osteoporotic fractures is expected to rise by 20%, where pelvic fragility fractures are expected to disproportionately rise by 56% in the United States . Loggers et al. retrospectively investigated 117 elderly patients with FFPs and showed that 49% lost their independent mobility status, 40% failure to return to pre-injury functional status, and the 1-year mortality rate was 23%. FFPs result in considerable morbidity and mortality and as well as massive financial burden on the already strained health systems throughout the world. FFP is typically caused by low-energy trauma, such as falls in those without a significant history of trauma and occurs in older patients with osteoporosis. Symptoms include moderate-to-severe pain in the pubic, groin, and sacrococcygeal regions, which significantly impacts daily life but rarely causes hemodynamic instability. Because of the characteristics of FFP, the traditional Tile and Young–Burgess classification systems are not applicable. Consequently, Rommens et al. proposed an alternative classification system for FFP in 2013 , which has been increasingly recognized and applied in clinical practice. Typically, type I cases are considered for conservative treatment, whereas types III and IV cases are considered for surgical treatment. FFP type II (FFP II) is the most common type of FFP (accounting for > 50% of cases), yet its treatment remains controversial . The main goals of treatment are to stabilize the fracture ends, relieve pain, and promote early mobilization, whereas anatomical reduction of the fracture and restoration of pelvic symmetry are relatively secondary concerns. Osteoporotic fractures include those of the spine, pelvis, hip (femoral neck and intertrochanteric fractures), proximal humerus, and distal radius, with early surgical treatment of hip fractures being widely accepted. FFP shares many similarities with intertrochanteric femoral fractures, such as being caused by low-energy trauma and leading to bed rest in older patients, as well as to subsequent complications (e.g., pneumonia, pressure sores, urinary tract infections, and deep vein thrombosis). However, the optimal treatment strategy for FFP and its impact on outcomes remain unclear. Historically, FFP was often managed conservatively, but the functional impairment and reduced quality of life (QoL) associated with conservative treatment are frequently underestimated . In recent years, more and more surgeons have performed surgical treatment of FFP and achieved significant clinical results. Tolosano et al. studied of 48 patients with FFP showed that surgical treatment can significantly relieve pain and preserve the patient’s independence. On the basis of a study of 42 patients with FFP, Yoshida et al. suggested that surgery contributes to early mobilization. Heiman et al. , in their latest review, suggested that operative fixation should focus on minimally invasive stabilization of the pelvic ring to facilitate early mobilization and avoid the complications that can arise from comorbidities associated with immobility. Further high-quality comparative literature is needed before treatment criteria can be optimized and standardized. Rollmann et al. reported a significant increase in the proportion of older patients with FFP undergoing surgical treatment over the past 22 years. With the advent of robot-assisted surgery and 3D printing technology, many studies have reported good progress in the surgical treatment of older patients with FFP, with minimally invasive surgery gaining increasing acceptance among doctors and patients. Despite the increasing application of minimally invasive surgery for FFP, there are few reports comparing its clinical outcomes and postoperative QoL with those of conservative treatment. Therefore, this study aims to compare clinical outcomes and QoL improvements between minimally invasive surgery and conservative treatment for type II FFP.
Study design A retrospective cohort study was conducted to compare outcomes of minimally invasive surgery versus conservative treatment in patients with type II fragility fractures of the pelvis (FFP). The study included patients treated at Tianjin Medical University Baodi Hospital between January 2019 and December 2022. Ethical approval was obtained from the Ethics Committee of Tianjin Medical University. Participants were divided into two groups based on treatment modality: surgical ( n = 68) and conservative ( n = 82). Outcomes were assessed using validated questionnaires and imaging at three time points: pre-treatment, pre-discharge, and final follow-up (mean follow-up: 22 ± 5 months). Setting The study was conducted at Tianjin Medical University Baodi Hospital, a tertiary care facility. Data were collected from electronic medical records, imaging archives, and patient-reported outcome measures. Participants Inclusion Criteria: Age ≥ 65 years. Diagnosis of type II FFP confirmed via imaging (X-ray, CT, MRI). Completed follow-up ≥ 1 year. Exclusion Criteria: Age < 65 years. High-energy trauma, open fractures, or infection. Incomplete survey responses, revision surgeries, or refusal to participate. Cohort: 150 patients met inclusion criteria (Fig. ). Informed consent was obtained from all participants after providing them with detailed information about the study purpose, procedures, and potential risks and benefits. Consent was documented by having the participants sign a consent form. Variables Primary outcomes Clinical function: Majeed score (0–100; higher = better pelvic function) and Short Musculoskeletal Function Assessment (SMFA) questionnaire (0–46; higher = poorer function/distress). Quality of life (QoL): Short-Form 36 Health Survey (SF-36) (0–100; higher = better QoL) and World Health Organization Quality of Life Brief Version (WHOQOL-BREF) questionnaire (physical, psychological, social, environmental domains). The Majeed score is a tool used to evaluate pelvic function. It mainly assesses Pain, Standing, Sitting, Sexual intercourse and Work. The score ranges from 0 to 100, with a higher score indicating better pelvic function. It offers a comprehensive and standardized way to measure pelvic function. The SMFA consists of two sections, with a total of 46 items: 34 questions assess clinical function, and 12 questions evaluate the level of distress caused by symptoms. This questionnaire assesses treatment outcomes for musculoskeletal diseases/injuries and can reliably and effectively evaluate the patient’s health status, with higher scores indicating poorer function or greater distress. The SF-36 is a well-known and widely used health questionnaire composed of eight sections, with higher scores indicating better QoL and health status. The WHOQOL-BREF assesses patients in terms of physical, psychological, social, and environmental aspects and is a reliable and effective method for evaluating health-related QoL. Secondary variables Demographic data (age, sex). Fracture classification (Rommens system Table ). Treatment type (surgical vs. conservative). Imaging findings (displacement, healing). Data sources/measurement Imaging: All patients underwent standardized X-ray, CT, and MRI examinations. Fractures were classified using the Rommens system . Questionnaires Administered at three time points: before treatment, before discharge, and at the final follow-up. Majeed Score: Assesses pain, standing, sitting, sexual intercourse, and work. SMFA: 34 items for clinical function, 12 for symptom-related distress. SF-36: Eight domains (e.g., physical function, mental health). WHOQOL-BREF: Evaluates physical, psychological, social, and environmental QoL. Bias Selection Bias: Minimized by strict inclusion/exclusion criteria and standardized treatment protocols. Measurement Bias: Reduced via validated tools (Majeed, SMFA, SF-36, WHOQOL-BREF) and blinded imaging analysis. Attrition Bias: Addressed by excluding patients with follow-up < 1 year. Confounding: Controlled by matching baseline characteristics (e.g., age, fracture severity) between groups. Sample size The final cohort included 150 patients (surgical: 68; conservative: 82), determined by the availability of eligible patients during the study period. Post-hoc power analysis confirmed adequate power (β ≥ 0.8) to detect clinically meaningful differences in primary outcomes. Quantitative variables Continuous: Age, follow-up duration, and questionnaire scores (mean ± SD). Categorical: Treatment type, fracture classification, and demographic variables (frequency/percentage). Statistical methods Software: SPSS v23.0 (Chicago, IL). Normality: Assessed via Shapiro–Wilk test. Group Comparisons: Parametric data: Independent t-test (continuous) and chi-square (categorical). Non-parametric data: Mann–Whitney U (independent samples) and Wilcoxon test (related samples). Significance Threshold: p < 0.05. Treatment protocols Conservative Group: Received pain management and anti-osteoporosis therapy. Continued if pain improved and imaging showed no displacement after 1 week, or if patients declined surgery despite worsening symptoms (Fig. ). Surgical Group: Underwent minimally invasive surgery if pain persisted or imaging revealed displacement after 1 week (Fig. ).
A retrospective cohort study was conducted to compare outcomes of minimally invasive surgery versus conservative treatment in patients with type II fragility fractures of the pelvis (FFP). The study included patients treated at Tianjin Medical University Baodi Hospital between January 2019 and December 2022. Ethical approval was obtained from the Ethics Committee of Tianjin Medical University. Participants were divided into two groups based on treatment modality: surgical ( n = 68) and conservative ( n = 82). Outcomes were assessed using validated questionnaires and imaging at three time points: pre-treatment, pre-discharge, and final follow-up (mean follow-up: 22 ± 5 months).
The study was conducted at Tianjin Medical University Baodi Hospital, a tertiary care facility. Data were collected from electronic medical records, imaging archives, and patient-reported outcome measures.
Inclusion Criteria: Age ≥ 65 years. Diagnosis of type II FFP confirmed via imaging (X-ray, CT, MRI). Completed follow-up ≥ 1 year. Exclusion Criteria: Age < 65 years. High-energy trauma, open fractures, or infection. Incomplete survey responses, revision surgeries, or refusal to participate. Cohort: 150 patients met inclusion criteria (Fig. ). Informed consent was obtained from all participants after providing them with detailed information about the study purpose, procedures, and potential risks and benefits. Consent was documented by having the participants sign a consent form.
Primary outcomes Clinical function: Majeed score (0–100; higher = better pelvic function) and Short Musculoskeletal Function Assessment (SMFA) questionnaire (0–46; higher = poorer function/distress). Quality of life (QoL): Short-Form 36 Health Survey (SF-36) (0–100; higher = better QoL) and World Health Organization Quality of Life Brief Version (WHOQOL-BREF) questionnaire (physical, psychological, social, environmental domains). The Majeed score is a tool used to evaluate pelvic function. It mainly assesses Pain, Standing, Sitting, Sexual intercourse and Work. The score ranges from 0 to 100, with a higher score indicating better pelvic function. It offers a comprehensive and standardized way to measure pelvic function. The SMFA consists of two sections, with a total of 46 items: 34 questions assess clinical function, and 12 questions evaluate the level of distress caused by symptoms. This questionnaire assesses treatment outcomes for musculoskeletal diseases/injuries and can reliably and effectively evaluate the patient’s health status, with higher scores indicating poorer function or greater distress. The SF-36 is a well-known and widely used health questionnaire composed of eight sections, with higher scores indicating better QoL and health status. The WHOQOL-BREF assesses patients in terms of physical, psychological, social, and environmental aspects and is a reliable and effective method for evaluating health-related QoL. Secondary variables Demographic data (age, sex). Fracture classification (Rommens system Table ). Treatment type (surgical vs. conservative). Imaging findings (displacement, healing).
Clinical function: Majeed score (0–100; higher = better pelvic function) and Short Musculoskeletal Function Assessment (SMFA) questionnaire (0–46; higher = poorer function/distress). Quality of life (QoL): Short-Form 36 Health Survey (SF-36) (0–100; higher = better QoL) and World Health Organization Quality of Life Brief Version (WHOQOL-BREF) questionnaire (physical, psychological, social, environmental domains). The Majeed score is a tool used to evaluate pelvic function. It mainly assesses Pain, Standing, Sitting, Sexual intercourse and Work. The score ranges from 0 to 100, with a higher score indicating better pelvic function. It offers a comprehensive and standardized way to measure pelvic function. The SMFA consists of two sections, with a total of 46 items: 34 questions assess clinical function, and 12 questions evaluate the level of distress caused by symptoms. This questionnaire assesses treatment outcomes for musculoskeletal diseases/injuries and can reliably and effectively evaluate the patient’s health status, with higher scores indicating poorer function or greater distress. The SF-36 is a well-known and widely used health questionnaire composed of eight sections, with higher scores indicating better QoL and health status. The WHOQOL-BREF assesses patients in terms of physical, psychological, social, and environmental aspects and is a reliable and effective method for evaluating health-related QoL.
Demographic data (age, sex). Fracture classification (Rommens system Table ). Treatment type (surgical vs. conservative). Imaging findings (displacement, healing).
Imaging: All patients underwent standardized X-ray, CT, and MRI examinations. Fractures were classified using the Rommens system . Questionnaires Administered at three time points: before treatment, before discharge, and at the final follow-up. Majeed Score: Assesses pain, standing, sitting, sexual intercourse, and work. SMFA: 34 items for clinical function, 12 for symptom-related distress. SF-36: Eight domains (e.g., physical function, mental health). WHOQOL-BREF: Evaluates physical, psychological, social, and environmental QoL.
Selection Bias: Minimized by strict inclusion/exclusion criteria and standardized treatment protocols. Measurement Bias: Reduced via validated tools (Majeed, SMFA, SF-36, WHOQOL-BREF) and blinded imaging analysis. Attrition Bias: Addressed by excluding patients with follow-up < 1 year. Confounding: Controlled by matching baseline characteristics (e.g., age, fracture severity) between groups.
The final cohort included 150 patients (surgical: 68; conservative: 82), determined by the availability of eligible patients during the study period. Post-hoc power analysis confirmed adequate power (β ≥ 0.8) to detect clinically meaningful differences in primary outcomes.
Continuous: Age, follow-up duration, and questionnaire scores (mean ± SD). Categorical: Treatment type, fracture classification, and demographic variables (frequency/percentage).
Software: SPSS v23.0 (Chicago, IL). Normality: Assessed via Shapiro–Wilk test. Group Comparisons: Parametric data: Independent t-test (continuous) and chi-square (categorical). Non-parametric data: Mann–Whitney U (independent samples) and Wilcoxon test (related samples). Significance Threshold: p < 0.05.
Conservative Group: Received pain management and anti-osteoporosis therapy. Continued if pain improved and imaging showed no displacement after 1 week, or if patients declined surgery despite worsening symptoms (Fig. ). Surgical Group: Underwent minimally invasive surgery if pain persisted or imaging revealed displacement after 1 week (Fig. ).
Comparison of basic information A total of 150 type II FFP patients were included in the study: 82 patients in the conservative treatment group (18 men, 64 women; average age, 78 ± 12.3 years) and 68 patients in the minimally invasive surgery group (13 men, 55 women; average age, 77 ± 11.6 years). Among the conservative treatment group, there were 20 cases of FFP IIa, 44 cases of FFP IIb, and 18 cases of FFP IIc. Among the surgical group, there were 16 cases of FFP IIa, 37 cases of FFP IIb, and 15 cases of FFP IIc. There were no statistically significant differences in basic information between the two groups (Table ). In the conservative group, there was one case of bedsores, one case of urinary tract infection, and three cases of deep vein thrombosis. There was one case of urinary tract infection, two cases of deep vein thrombosis, and two cases of mild incision infection in the surgical group. None of the complications had serious adverse consequences. Clinical effectiveness assessment There were no significant differences in the Majeed or SMFA questionnaire scores between the conservative treatment group and the minimally invasive surgery group at admission (45.36 ± 15.33 vs. 38.05 ± 28.56, P = 0.060; and 161.58 ± 20.67 vs. 168.35 ± 25.88, P = 0.128, respectively) or the final follow-up (85.51 ± 23.74 vs. 88.12 ± 16.38, P = 0.481; and 60.05 ± 8.65 vs. 56.45 ± 12.04, P = 0.400, respectively). The results indicate that conservative treatment and minimally invasive surgery achieve similar clinical outcomes in patients with type II FFP pelvic fractures (Table ). Significant functional improvements were observed before and after treatment (Table , P 1 < 0.01). QoL assessment There were no statistically significant differences in QoL between the conservative treatment group and the minimally invasive surgery group at admission, as assessed by the SF-36 and WHOQOL-BREF questionnaire ( P 1 , Tables and ). At the final follow-up, both groups showed significant improvements in QoL compared with that before treatment (conservative treatment group, P 2 < 0.01; minimally invasive surgery group, P 3 < 0.01, Tables and ). However, at the final follow-up, except for the SF-36 scores in the general health and bodily pain domains, and the WHOQOL-BREF questionnaire scores in the psychological domain, all SF-36 domain scores (physical functioning, role physical, vitality, social functioning, role emotional, mental health) and WHOQOL-BREF domain scores (physical, social, and environment) showed significantly greater improvements in the minimally invasive surgery group compared with the conservative treatment group ( P 4 , Tables and ).
A total of 150 type II FFP patients were included in the study: 82 patients in the conservative treatment group (18 men, 64 women; average age, 78 ± 12.3 years) and 68 patients in the minimally invasive surgery group (13 men, 55 women; average age, 77 ± 11.6 years). Among the conservative treatment group, there were 20 cases of FFP IIa, 44 cases of FFP IIb, and 18 cases of FFP IIc. Among the surgical group, there were 16 cases of FFP IIa, 37 cases of FFP IIb, and 15 cases of FFP IIc. There were no statistically significant differences in basic information between the two groups (Table ). In the conservative group, there was one case of bedsores, one case of urinary tract infection, and three cases of deep vein thrombosis. There was one case of urinary tract infection, two cases of deep vein thrombosis, and two cases of mild incision infection in the surgical group. None of the complications had serious adverse consequences.
There were no significant differences in the Majeed or SMFA questionnaire scores between the conservative treatment group and the minimally invasive surgery group at admission (45.36 ± 15.33 vs. 38.05 ± 28.56, P = 0.060; and 161.58 ± 20.67 vs. 168.35 ± 25.88, P = 0.128, respectively) or the final follow-up (85.51 ± 23.74 vs. 88.12 ± 16.38, P = 0.481; and 60.05 ± 8.65 vs. 56.45 ± 12.04, P = 0.400, respectively). The results indicate that conservative treatment and minimally invasive surgery achieve similar clinical outcomes in patients with type II FFP pelvic fractures (Table ). Significant functional improvements were observed before and after treatment (Table , P 1 < 0.01).
There were no statistically significant differences in QoL between the conservative treatment group and the minimally invasive surgery group at admission, as assessed by the SF-36 and WHOQOL-BREF questionnaire ( P 1 , Tables and ). At the final follow-up, both groups showed significant improvements in QoL compared with that before treatment (conservative treatment group, P 2 < 0.01; minimally invasive surgery group, P 3 < 0.01, Tables and ). However, at the final follow-up, except for the SF-36 scores in the general health and bodily pain domains, and the WHOQOL-BREF questionnaire scores in the psychological domain, all SF-36 domain scores (physical functioning, role physical, vitality, social functioning, role emotional, mental health) and WHOQOL-BREF domain scores (physical, social, and environment) showed significantly greater improvements in the minimally invasive surgery group compared with the conservative treatment group ( P 4 , Tables and ).
In our study, the conservative treatment group had significantly lower QoL scores compared with the minimally invasive surgery group, including SF-36 physical functioning, role physical, vitality, social functioning, role emotional, and mental health domain scores as well as WHOQOL-BREF physical, social, and environment domain scores. This finding suggests that minimally invasive surgery offers a significant advantage in terms of improving QoL. Yang et al. conducted a retrospective analysis of 135 patients with pelvic fragility fractures and found that the minimally invasive surgery group had significantly better Majeed and VAS scores compared with the conservative group. Additionally, the surgery group had shorter bed rest and fracture healing times, leading to the conclusion that minimally invasive surgery significantly enhances the QoL of older patients with pelvic fragility fractures. Conversely, Thiesen et al. performed a randomized, prospective, non-blinded study of 39 pelvic fracture patients, using the Barthel index, VAS pain score, QoL on the EuroQol five-dimension three-level questionnaire, and Tinetti gait test for assessment, and found no significant benefits of surgical treatment over conservative treatment in terms of QoL, mortality, or pain level. This discrepancy with our results highlights the need for further research to determine the best management approach for FFP. Our findings can serve as a reference for future randomized controlled trials. Our study found no significant difference in clinical outcomes between the conservative treatment and minimally invasive surgery groups, on the basis of clinical and QoL scores. However, both treatment methods resulted in significant improvements in clinical outcomes and QoL from pre-treatment to follow-up; this indicates that both conservative and surgical treatments are beneficial for patients with osteoporotic pelvic fractures, which aligns with the findings of Schramm, Yoo, and Yoshimura . Schramm et al. assessed 46 patients with pelvic fractures using the Barthel Index, Tinetti mobility test, and timed up and go test, confirming that conservative treatment improved the meeting of care needs, independent mobility, and fall risk. Yoo et al. evaluated 41 patients treated conservatively on the basis of visual analog scale (VAS) pain scores and time to mobilization, demonstrating that teriparatide treatment could achieve early pain relief and mobilization, reducing fracture healing time. Yoshimura et al. found that minimally invasive surgery for six sacral fractures allowed early use of assistive devices and improved walking ability, even with in situ pelvic ring fixation. Pelvic fractures are considered severe injuries, and individuals with pelvic fractures experience a reduced QoL compared with the general population. Age and surgery are independent predictors of decreased QoL following pelvic ring fractures . However, it remains unclear whether patient age influences the decision for surgical treatment of pelvic ring fractures and whether the indications for surgery in older patients have changed over the past two decades. Osteoporotic pelvic fractures in older adults are becoming increasingly common, yet there is no consensus on whether to treat most of these patients conservatively or surgically, especially those with FFP II-type fractures . Although osteoporosis can be prevented and treated with medications and dietary interventions, such as vitamin D, calcium, bisphosphonates, and recombinant human parathyroid hormone, the optimal treatment approach for these patients remains a topic of ongoing discussion between clinicians and patients. The primary aim of this study was to evaluate and compare clinical outcomes and QoL between conservative treatment and surgical intervention, to assist in selecting the best treatment approach. Osteoporotic pelvic fractures in older patients often involve minimal or no displacement and may not require anatomical reduction, making minimally invasive reduction and fixation techniques more suitable for older patients with FFP. Functional reduction and minimally invasive fixation are fundamental for early pain-free functional exercise. Preoperative assessment should consider factors such as age, thromboembolism risk, albumin levels, anemia, cardiovascular and respiratory diseases, smoking, body mass index, Parkinson’s disease, and osteoporosis to prevent, control, and manage related risk factors, adopting similar treatment models and strategies to those used in older patients with hip fractures. A fast-track system should also be established. Compared with conservative treatment, surgery significantly improves patients’ QoL, restores independence, and reduces the burden on families and society. Minimally invasive fixation techniques include external fixators, nail-rod systems, minimally invasive plating, and percutaneous screw fixation. Surgical goals include functional reduction, pelvic stability, pain relief, and reduced bed rest . However, minimally invasive surgery faces challenges, such as the need for precise placement of various sacroiliac screws, where planning entry points and screw paths on the basis of imaging data is crucial. Imaging technology, 3D printing, navigation, robotic-assisted surgery, and minimally invasive pelvic fracture reduction techniques continue to evolve . A prospective study indicated that navigated percutaneous sacral iliac screw placement improved screw positioning in deformed sacra but came at the cost of longer surgery times and increased radiation exposure. Other researchers proposed that robotic navigation technology could address these issues, with robotic and 3D-printed assistance reducing surgery time and radiation exposure, and improving safety and accuracy. Wu et al. used patient-specific locking navigation templates to treat pelvic fractures with sacral dysplasia or sacroiliac joint dislocations, proving the safety of this method through finite element analysis. Therefore, we advocate using new technologies to reduce surgery time for older patients and improve treatment outcomes. A consensus on the treatment of osteoporotic pelvic fractures has not been reached, and standardizing treatment, increasing awareness, researching new minimally invasive techniques, and reducing the high disability and mortality rates remains challenging. On the basis of our results and recent technological advancements, minimally invasive treatment for older patients with FFP pelvic fragility fractures is a recommended option. This study has limitations. First, it is a retrospective study conducted at a single center with a small sample size, which may limit generalizability. Additionally, patients were not randomly assigned to groups, and some patients in the conservative group who were advised to undergo surgery declined because of their current health status (comorbidities) and personal preferences. Third, response bias cannot be ruled out as QoL was assessed on the basis of self-reported questionnaires.
In summary, minimally invasive surgery and conservative treatment for FFP type II fractures yield similar clinical outcomes, but minimally invasive surgery offers superior QoL improvements. These findings support the consideration of surgical options for patients with FFP type II who are seeking enhanced functional recovery and better overall QoL.
|
Meta‐Analysis of the Safety and Efficacy of Intensive Blood Pressure Control After Thrombectomy | e4859b9b-0900-478c-abf6-ed890091d21e | 11807842 | Surgical Procedures, Operative[mh] | Introduction Acute ischemic stroke represents a significant global health challenge, and endovascular thrombectomy (EVT) has emerged as a pivotal milestone in its treatment, profoundly altering the organization and operation of stroke services (Lapergue et al. ; Goyal et al. ). By eliminating obstruction or clots within arteries, EVT can effectively restore blood flow to ischemic regions in the brain, known as penumbral tissue (Lin et al. ; Wang et al. ). However, despite achieving favorable radiological outcomes, many patients exhibit poor functional recovery, with high risks of symptomatic intracranial hemorrhage and other forms of reperfusion injury. Consequently, there is a growing interest in adjunctive approaches post‐EVT to protect or sustain penumbral tissue from reperfusion injury. Blood pressure emerges as a modifiable factor to prevent reperfusion injury, given its frequent elevation and clear prognostic significance in acute ischemic stroke (Anadani et al. ; Mistry et al. ; Maïer et al. ; Samuels et al. ; Anadani et al. ). While guidelines advocate for conservative blood pressure control pre‐EVT and post‐EVT (Powers et al. ), recent confidence in EVT's efficacy, desire to mitigate ischemia‐reperfusion injury risks, and influential data linking presentation blood pressure to subsequent clinical outcomes have shifted opinions toward more aggressive blood pressure control in research and practice (Mulder et al. ). Nevertheless, in the absence of randomized evidence, guidelines continue to recommend maintaining lower blood pressure levels post‐EVT, consistent with those for patients eligible for intravenous thrombolysis after acute ischemic stroke (Powers et al. ; Turc et al. ). Hence, our aim is to assess the safety and efficacy of intensive blood pressure management compared to less intensive treatment in patients with successful reperfusion following EVT, addressing this controversy. Methods and Analysis We conducted and reported this systematic review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses statement. The review protocol was registered in the International Platform of Registered Systematic Review and Meta‐analysis Protocols (INPLASY, https://inplasy.com/ ) with the registration number of INPLASY202460008. 2.1 Search Strategy Conduct systematic reviews and meta‐analyses of RCTs according to the guidelines for systematic reviews and meta‐analyses (Page et al. ). Search the published literature in the EMBASE, PubMed, and Cochrane Library databases, primarily using search terms such as “ischemic stroke,” “endovascular thrombectomy,” and “blood pressure” supplemented with relevant synonyms to avoid missing literature during the search. The search terms will be provided in detail as online supplementary material . 2.2 Inclusion and Exclusion Criteria After excluding duplicate samples, we screened the remaining literature according to the following criteria: (1) Cases of stroke diagnosed by imaging examinations. (2) Prospective studies comparing two different reinforcement standards for blood pressure treatment to assess their effects on stroke. (3) Studies with a sample size larger than 50 cases. We excluded conference abstracts, case reports, reviews, and letters from the search results of RCTs. Additionally, we excluded non‐English literature, abstracts, and studies unrelated to stroke. We assessed duplicate or overlapping data in publications and only included the most comprehensive studies. Unpublished data were not sought. 2.3 Data Extraction Two independent researchers (BQJ, LYM) served as secondary evaluators, responsible for data extraction and cross‐verification. For the eligible studies, the following data were collected: basic information (including the first author's name, year of publication, and study region); patient characteristics (including the number of patients, mean age, medications prior to intervention, and medical history); and efficacy indicators (including favorable clinical outcomes with a mRS of 0–2, excellence clinical outcomes with a mRS of 0–1, and mortality rate). Any uncertainties or discrepancies were resolved through discussions between the researchers and the secondary evaluators to reach a consensus. 2.4 Study Quality Assessment and Statistical Analysis We used the Newcastle–Ottawa Scale to assess the quality of these RCTs (Ottawa Hospital Research Institute [Internet] ).The heterogeneity was assessed using the I2 statistic and the chi‐square test. Heterogeneity was considered significant when I2 > 50%. If the included studies had I2 < 50% for the intervention outcomes, the fixed‐effect model of Mantel–Haenszel method was used. Otherwise, the random‐effects model of Mantel–Haenszel was employed. Visual funnel plots were used to evaluate publication bias (Figures ). The statistical significance was set at p‐value < 0.05, indicating a statistically significant result. All analyses were conducted using Review Manager (RevMan, version 5.4). Search Strategy Conduct systematic reviews and meta‐analyses of RCTs according to the guidelines for systematic reviews and meta‐analyses (Page et al. ). Search the published literature in the EMBASE, PubMed, and Cochrane Library databases, primarily using search terms such as “ischemic stroke,” “endovascular thrombectomy,” and “blood pressure” supplemented with relevant synonyms to avoid missing literature during the search. The search terms will be provided in detail as online supplementary material . Inclusion and Exclusion Criteria After excluding duplicate samples, we screened the remaining literature according to the following criteria: (1) Cases of stroke diagnosed by imaging examinations. (2) Prospective studies comparing two different reinforcement standards for blood pressure treatment to assess their effects on stroke. (3) Studies with a sample size larger than 50 cases. We excluded conference abstracts, case reports, reviews, and letters from the search results of RCTs. Additionally, we excluded non‐English literature, abstracts, and studies unrelated to stroke. We assessed duplicate or overlapping data in publications and only included the most comprehensive studies. Unpublished data were not sought. Data Extraction Two independent researchers (BQJ, LYM) served as secondary evaluators, responsible for data extraction and cross‐verification. For the eligible studies, the following data were collected: basic information (including the first author's name, year of publication, and study region); patient characteristics (including the number of patients, mean age, medications prior to intervention, and medical history); and efficacy indicators (including favorable clinical outcomes with a mRS of 0–2, excellence clinical outcomes with a mRS of 0–1, and mortality rate). Any uncertainties or discrepancies were resolved through discussions between the researchers and the secondary evaluators to reach a consensus. Study Quality Assessment and Statistical Analysis We used the Newcastle–Ottawa Scale to assess the quality of these RCTs (Ottawa Hospital Research Institute [Internet] ).The heterogeneity was assessed using the I2 statistic and the chi‐square test. Heterogeneity was considered significant when I2 > 50%. If the included studies had I2 < 50% for the intervention outcomes, the fixed‐effect model of Mantel–Haenszel method was used. Otherwise, the random‐effects model of Mantel–Haenszel was employed. Visual funnel plots were used to evaluate publication bias (Figures ). The statistical significance was set at p‐value < 0.05, indicating a statistically significant result. All analyses were conducted using Review Manager (RevMan, version 5.4). Results The flowchart in Figure depicts the literature search process. Out of the initially retrieved 3304 articles, 245 were excluded due to duplicate publication, and 2939 were excluded based on titles and abstracts. Among the 48 articles subjected to full‐text assessment, 4 met the inclusion criteria. 3.1 Research Characteristics and Quality Assessment Four randomized controlled trials were included, comprising a total of 1556 study participants. Table summarizes the baseline characteristics of all included studies. The average age of participants was 72 years, with males accounting for 59%.While two studies did not specify the ethnicity of participants, the remaining data indicate a higher proportion of Asian participants in the study cohort. Baseline blood pressure between the intensified treatment and standard treatment groups was similar across included studies. Fifty percent of the studies (2 out of 4) were conducted mainly in European and American countries, while the remaining 50% (2 out of 4) were conducted in Asian countries, all of which were multicenter studies. We obtained additional details for each eligible study meeting the criteria. 3.2 The Impact of Intensified Blood Pressure Control on Patient Prognosis Compared to the intensive blood pressure control group, the standard blood pressure control group demonstrated a higher likelihood of achieving favorable clinical outcomes (mRS ≤ 2). Specifically, in an analysis encompassing four studies, we observed superior clinical outcomes in the standard group (Figure ), with a RR of 0.81 (95%CI: 0.73–0.90; p < 0.05; I 2 = 25%). Similarly, the standard group demonstrated a higher probability of achieving an excellent clinical outcome (mRS ≤ 1) at 90 days compared to the intensive blood pressure control group. Analysis of data from four studies revealed that the standard group exhibited significantly better clinical outcomes than the intensive blood pressure control group (Figure ), with an RR of 0.86 (95% CI: 0.76–0.98; p < 0.05; I 2 = 15%). Moreover, funnel plots indicated no significant publication bias for any outcome, further supporting the reliability of our study findings 3.3 Safety Concerns of Intensified Blood Pressure Control However, upon examining mortality rates, we found no significant difference between the standard and intensive blood pressure control groups at 90 days. Specifically, in an analysis of data from four studies (Figure ), the RR of mortality was 1.13 (95% CI: 0.90–1.43; p = 0.30; I 2 = 0%), indicating no statistically significant difference between the two groups. Similarly, our comparison of safety outcomes related to symptomatic intracranial hemorrhage showed no significant difference. Analysis of data from four studies (Figure ) revealed an RR of 1.13 for symptomatic intracranial hemorrhage (95% CI: 0.78–1.63; p = 0.51; I 2 = 0%), again indicating no statistically significant difference. Therefore, our findings suggest that, regarding safety, there is no significant difference between standard and intensive blood pressure control Research Characteristics and Quality Assessment Four randomized controlled trials were included, comprising a total of 1556 study participants. Table summarizes the baseline characteristics of all included studies. The average age of participants was 72 years, with males accounting for 59%.While two studies did not specify the ethnicity of participants, the remaining data indicate a higher proportion of Asian participants in the study cohort. Baseline blood pressure between the intensified treatment and standard treatment groups was similar across included studies. Fifty percent of the studies (2 out of 4) were conducted mainly in European and American countries, while the remaining 50% (2 out of 4) were conducted in Asian countries, all of which were multicenter studies. We obtained additional details for each eligible study meeting the criteria. The Impact of Intensified Blood Pressure Control on Patient Prognosis Compared to the intensive blood pressure control group, the standard blood pressure control group demonstrated a higher likelihood of achieving favorable clinical outcomes (mRS ≤ 2). Specifically, in an analysis encompassing four studies, we observed superior clinical outcomes in the standard group (Figure ), with a RR of 0.81 (95%CI: 0.73–0.90; p < 0.05; I 2 = 25%). Similarly, the standard group demonstrated a higher probability of achieving an excellent clinical outcome (mRS ≤ 1) at 90 days compared to the intensive blood pressure control group. Analysis of data from four studies revealed that the standard group exhibited significantly better clinical outcomes than the intensive blood pressure control group (Figure ), with an RR of 0.86 (95% CI: 0.76–0.98; p < 0.05; I 2 = 15%). Moreover, funnel plots indicated no significant publication bias for any outcome, further supporting the reliability of our study findings Safety Concerns of Intensified Blood Pressure Control However, upon examining mortality rates, we found no significant difference between the standard and intensive blood pressure control groups at 90 days. Specifically, in an analysis of data from four studies (Figure ), the RR of mortality was 1.13 (95% CI: 0.90–1.43; p = 0.30; I 2 = 0%), indicating no statistically significant difference between the two groups. Similarly, our comparison of safety outcomes related to symptomatic intracranial hemorrhage showed no significant difference. Analysis of data from four studies (Figure ) revealed an RR of 1.13 for symptomatic intracranial hemorrhage (95% CI: 0.78–1.63; p = 0.51; I 2 = 0%), again indicating no statistically significant difference. Therefore, our findings suggest that, regarding safety, there is no significant difference between standard and intensive blood pressure control Discussion In this systematic review and meta‐analysis evaluating blood pressure management strategies following vascular interventions in stroke patients, with patient outcomes assessed over a 3‐month follow‐up period, it was determined that the choice between intensified and standard blood pressure control did not significantly influence patient outcomes. A multicenter retrospective study investigated the effects of different blood pressure targets in patients with acute ischemic stroke following thrombectomy, comparing intensive control (<140 mmHg), moderate control (<160 mmHg), and standard control (<180 mmHg) groups. The study found that intensive blood pressure control was associated with better functional outcomes and fewer complications, such as symptomatic intracranial hemorrhage (Anadani et al. ). However, unlike this observational study, intensive blood pressure control did not demonstrate superiority in clinical outcomes over standard control (Mazighi et al. ).However, no significant disparity in safety outcomes between the two approaches was observed, including 90‐day mortality and symptomatic intracranial hemorrhage. Notably, there is insufficient evidence to suggest that intensified blood pressure control escalates the risk of adverse events. These estimations exhibit robustness and demonstrate minimal variation in sensitivity analysis. This review presents the latest insights, incorporating outcome data from 1556 stroke patients postvascular intervention. Four studies were subjected to analysis to assess the efficacy of intensified blood pressure treatment, all of which yielded analogous results. The management of blood pressure poses multifaceted challenges in acute ischemic stroke patients. Swift fluctuations in blood pressure may exacerbate damage to compromised brain tissue, particularly postvascular reperfusion therapy. Furthermore, various factors may impinge upon patients' blood pressure levels, including stroke severity, pretreatment baseline blood pressure levels, and a history of hypertension. Consequently, the formulation of a universally applicable blood pressure management strategy proves exceedingly arduous. Findings from pertinent nonrandomized controlled clinical trials posit that the adoption of an aggressive blood pressure management strategy augments the recovery of patients successfully undergoing vascular reperfusion therapy. Specifically, some trials suggest that setting a systolic blood pressure target below 140 mm Hg engenders relatively superior functional outcomes compared to traditional management modalities (Anadani et al. ; Mistry et al. ; Maïer et al. ; Chang and Han ). Nonetheless, recent meta‐analyses of randomized controlled trials proffer contrary conclusions. Intensified blood pressure control, particularly with a systolic blood pressure target of 140 mm Hg, despite evincing no elevated rates of symptomatic intracranial hemorrhage or mortality, engenders diminished functional recovery in patients. It is important to note that the experimental design of randomized controlled trials (RCTs) may influence the effectiveness of interventions. In the standard treatment group, patients with blood pressure below 140 mmHg did not receive interventions to raise their blood pressure. This means that some patients did not actually receive the intended intervention, which may have diluted the observed treatment effect. This design differs from the clinical practice in observational studies and may be one of the reasons for the discrepancies between the results of RCTs and observational studies. Future research should consider this factor and optimize experimental designs to more accurately assess the effects of blood pressure management strategies. It is postulated that excessive blood pressure reduction during control may precipitate inadequate vascular perfusion, exacerbating ischemia (Nam et al. ). Moreover, patients evincing poor vascular elasticity may exacerbate this phenomenon. Elevated rates of malignant brain edema occurrence have been observed in some studies, with several trials prematurely halted due to significantly inferior 90‐day outcomes in the intensified blood pressure control cohort (Yang et al. ). Recent investigations have further contended that both excessively high and low blood pressure may engender divergent impacts (Mistry et al. ). Consequently, this study undertook a nuanced subdivision of blood pressure control into three cohorts: <140 mm Hg, <160 mm Hg, <180 mm Hg. Nevertheless, even within the intensified blood pressure control <160 mm Hg subgroup, no discernibly superior outcomes were discerned. Given that elevated blood pressure constitutes a self‐regulatory mechanism in response to exacerbated edema, which compresses blood vessels and exacerbates cerebral ischemia, the body initiates perfusion pressure augmentation through self‐regulation. (Lattanzi et al. ; Alqadri, Sreenivasan, and Qureshi ). However, excessive blood pressure elevation escalates the risk of complications, underscoring the imperative to ascertain an optimal equilibrium. Despite the discordance in current clinical trial findings, they proffer valuable insights for prospective research endeavors. Further elucidation is warranted concerning the effects of diverse blood pressure management strategies in specific patient subgroups, necessitating stratified trials predicated on disparate baseline blood pressure levels. Patients evincing diverse baseline levels may manifest disparate vascular elasticity and edema severity, necessitating augmented perfusion pressure to sustain normative cerebral perfusion. Furthermore, patients undergoing specific treatments, such as edema control and multifarious complication management, warrant tailored intervention to engender enhanced clinical outcomes. A potential research direction is to stratify patients after mechanical reperfusion therapy based on their blood pressure levels for tailored management. For patients with lower blood pressure (<140 mmHg), nonpharmacological methods could be employed to raise blood pressure to <160 mmHg to assess whether increasing blood pressure improves perfusion and functional outcomes. Meanwhile, another group could maintain their current blood pressure levels, and a third group could have their blood pressure further reduced to <130 mmHg to observe the impact of additional blood pressure reduction. For postoperative patients with higher blood pressure (140–180 mmHg), they could be randomly assigned to a blood pressure reduction group or a nonintervention group to compare the effects of these strategies on patient outcomes. Such a design would enable us to more precisely determine the optimal blood pressure management targets for patients with different baseline blood pressure levels, thereby achieving more individualized treatment approaches. Subsequently, expansive‐scale and protracted follow‐up studies are imperative to adjudicate the long‐term ramifications of diverse blood pressure management strategies on patient outcomes. Moreover, predicated on clinical experience and evolving methodologies, extant blood pressure management guidelines necessitate refinement and updating to perpetuate enhanced clinical practice. Blood pressure management in acute ischemic stroke patients represents a multifaceted and pivotal concern. Recent clinical trial outcomes intimate that unduly reducing blood pressure may prove deleterious to recovery and exacerbate the risk of sundry complications in patients successfully undergoing vascular reperfusion therapy. Hence, personalized blood pressure management strategies are indispensable, necessitating individualized adjustments contingent upon patient‐specific circumstances and treatment backgrounds. Subsequent research endeavors should prioritize the delineation of the long‐term effects of diverse blood pressure management strategies and the refinement and updating of pertinent clinical guidelines to ameliorate patient recovery and prognosis. 4.1 Limitations Firstly, the included RCTs exhibit significant differences in participants' demographic characteristics, baseline characteristics, and inclusion criteria. This variability may introduce heterogeneity, affecting the generalizability of the results. Variations in stroke severity, comorbid conditions, and treatment protocols across studies may impact outcomes and complicate direct comparisons. Secondly, this meta‐analysis includes only four studies with a total of 753 patients. The relatively small sample size may limit the statistical power of the analysis, potentially obscuring smaller but clinically significant differences between intensive and standard blood pressure control strategies. Thirdly, the targets for intensive blood pressure control in the included studies are not uniform. Some studies aimed for a SBP of <140 mm Hg, while others had different thresholds. This lack of standardization makes it difficult to determine the optimal blood pressure management target after EVT. Moreover, the analysis did not include unpublished studies, which could lead to publication bias. Studies with negative or null results are less likely to be published, potentially skewing the meta‐analysis toward positive outcomes. Finally, only English‐language studies were included, excluding non‐English literature. This may result in language bias and the omission of relevant data published in other languages. Limitations Firstly, the included RCTs exhibit significant differences in participants' demographic characteristics, baseline characteristics, and inclusion criteria. This variability may introduce heterogeneity, affecting the generalizability of the results. Variations in stroke severity, comorbid conditions, and treatment protocols across studies may impact outcomes and complicate direct comparisons. Secondly, this meta‐analysis includes only four studies with a total of 753 patients. The relatively small sample size may limit the statistical power of the analysis, potentially obscuring smaller but clinically significant differences between intensive and standard blood pressure control strategies. Thirdly, the targets for intensive blood pressure control in the included studies are not uniform. Some studies aimed for a SBP of <140 mm Hg, while others had different thresholds. This lack of standardization makes it difficult to determine the optimal blood pressure management target after EVT. Moreover, the analysis did not include unpublished studies, which could lead to publication bias. Studies with negative or null results are less likely to be published, potentially skewing the meta‐analysis toward positive outcomes. Finally, only English‐language studies were included, excluding non‐English literature. This may result in language bias and the omission of relevant data published in other languages. Conclusion After careful analysis, our conclusion is that intensified blood pressure control compared to standard blood pressure control following endovascular treatment in acute stroke patients does not yield better clinical outcomes and may even lead to inferior ones. Moreover, there is no significant disparity in terms of safety between the two approaches. However, due to the limited number of studies and potential variations in patient populations and methodologies, further analysis is warranted. We encourage additional large‐scale, high‐quality randomized controlled trials to confirm these findings and to establish optimal blood pressure management strategies for this patient population Qiang Ji Bao : Conceptualization, methodology, software, writing—original draft, validation, formal analysis, visualization. Yi Ming Li : Conceptualization, methodology, software, writing—original draft, validation, visualization, formal analysis. Xinting Wu : Data curation, investigation, software. Yun Ting Li : Software, data curation, investigation. Xiao Long Huang : Software, data curation, investigation. Hui Zhou : Software, data curation, investigation. Xiao Qiang Zhang : Writing—review and editing, supervision, resources, project administration, methodology, funding acquisition. Xue Jun Wang : Supervision, writing—review and editing, resources, project administration, methodology, funding acquisition. An ethics statement was not required for this study type; no human or animal subjects or materials were used. The authors declare that there is no conflict of interest regarding the publication of this article. The peer review history for this article is available at https://publons.com/publon/10.1002/brb3.70211 Figure S1. mRS ≤ 1 at 90 days. Figure S2. mRS ≤ 2 at 90 days. Figure S3. 90‐day mortality. Figure S4. Symptomatic intracranial bleeding. |
Expression Pattern of PDE4B, PDE4D, and SFRP5 Markers in Colorectal Cancer | 8efedd04-7e4f-4fe6-b5ae-81f2dabf9a40 | 11356070 | Anatomy[mh] | Colorectal cancer (CRC) is considered the most frequently diagnosed malignant disease of the gastrointestinal system, with a frequency of 10.2% and a mortality rate of 9.2% of all cancers . It is not a homogeneous disease but can be classified into different subtypes characterized by specific molecular and morphological changes. Around 95% of all cases of CRC are adenocarcinomas, and the rest are neuroendocrine tumors and small-cell carcinoma . A major feature of CRC is genetic instability, which can arise through at least two distinct mechanisms. The most common (about 84%) is characterized by chromosomal instability, with major changes in the number and structure of chromosomes, including deletions, translocations, and other chromosomal changes . The second group (around ~13–16% of sporadic CRC) is characterized by hypermutation and microsatellite instability (MSI) due to defective DNA mismatch repair (MMR) . Despite advances in diagnosis and new therapeutic options, the clinical outcomes of patients with colorectal cancer are still unsatisfactory. Overall, patient survival largely depends on the stage of the disease at the time of diagnosis and/or surgical resection . The most important prognostic indicator is the expansion of the tumor at the time of diagnosis. It is necessary to determine the exact stage of the tumor according to the TNM system or the established system introduced by Dukes. TNM is a system of classification primarily used for solid tumors, and it is based on assessing the tumor, regional lymph nodes, and distant metastasis . Mark T—Tumor describes the size of the primary tumor and its invasion into adjacent tissues; N—Nodes describe the tumor’s regional lymph node involvement; M—Metastasis is used to identify the presence of distant metastases of the primary tumor . The Dukes classification includes four grades based on the extent of tumor spread. Stage A represents tumors confined to the rectal wall. The tumor is in the inner lining of the bowel or slightly growing into the muscle layer. Stage B represents tumors that have penetrated the wall of the rectum, where the muscle layer is occupied. Stage C represents tumors with metastases in at least one lymph node close to the bowel. Stage D represents tumors with metastases in lymph nodes and distant organs . The five-year survival for stage Dukes’ A is more than 90% and only 5% for Dukes’ stage D . However, molecular classifications and epithelial–mesenchymal transition (EMT) concepts have become more important today, and novel markers are needed to elucidate the complete molecular tumor profile . Phosphodiesterases (PDEs) are categorized as metalloproteinases specialized in breaking down the secondary messengers’ cyclic adenosine monophosphate (cAMP) and guanosine 3′,5′-cyclic monophosphate (cGMP) . The phosphodiesterase-4 (PDE4) family of enzymes is the most prevalent PDE in the immune cells, epithelial cells, and brain cells and manifests as an intracellular non-receptor enzyme that modulates inflammation and epithelial integrity . Blocking PDE4 is predicted to induce various outcomes by increasing cAMP levels, leading to the control of numerous genes and proteins . Phosphodiesterase 4D (PDE4D) degrades cAMP and has recently been identified as an oncogene in different human cancer types . Additionally, PDE4D has emerged as a novel tumor-promoting molecule, presenting a distinctive targetable enzyme across multiple human cancers, including lung, prostate, melanoma, ovarian, endometrial, and gastric cancers . An important paralog of this gene is phosphodiesterase 4B (PDE4B). PDE4B plays a role in behaviors associated with dopamine and stress . Decreasing the activity of PDE4B in mouse models enhances memory and long-term plasticity, suggesting potential for therapeutic uses . Additionally, PDE4B inhibition may be a viable treatment to reduce inflammation . Secreted frizzled related proteins (SFRPs) are antagonists of the wingless-related integration site (Wnt) signaling pathway that binds directly to Wnt ligands and thus prevent their binding to frizzled receptors . There are currently eight known family members of SFRPs . Expression of SFRP has been observed in many malignant cancers . SFRP5 is known to be associated with hepatocellular carcinoma and gastric cancer . Nonetheless, the expression and importance of PDE4D, PDE4B, and SFRP5 in CRC remain unclear. Continued prevention, early detection, and treatment efforts remain essential in combating CRC and reducing its negative impact on public health. 2.1. Ethics This study was approved by the Ethics Committee of the University of Mostar School of Medicine in accordance with the Helsinki Declaration (approval number 1271723 from 2 February 2023). 2.2. Tissue Procurement and Immunohistochemistry Forty-three tumor tissue samples (12 of Dukes’ A, 11 of Dukes’ B, 10 of Dukes’ C, and 10 of Dukes’ D; ) were obtained from the Department of Pathology, Cytology, and Forensic Medicine at the University Hospital Mostar. Inclusion criteria encompassed patients with CRC older than 18 years, of both sexes and all cancer stages. Stratified randomization was used to ensure a balanced distribution of key characteristics (cancer stage) across treatment groups. Recruitment was conducted in collaboration with healthcare practitioners to identify eligible patients, supplemented by outreach through clinical settings. Participants were approached through direct contact during medical appointments, supplemented by follow-up communication and informational sessions to ensure thorough understanding and informed consent. The participants in this study did not receive any medications prior to the commencement of the study, since this was their first CRC diagnosis. Inclusion criteria for samples collected necessitated having an adequate amount of paraffin block material for immunohistochemistry (IHC) and complete clinical data. Exclusion criteria involved incomplete laboratory results, insufficient material for IHC, and lack of control tissue. Macroscopic examination and measurement of the tumor samples were conducted. Each sample contains tumor tissue and healthy control tissue. The tissues were dehydrated in increasing concentrations of ethanol solutions and embedded in paraffin. Paraffin blocks were cut using a microtome (Leica RM2155, Pittsburgh, PA, USA) in 4 μm sections to improve staining quality, reduce cell overlap, provide fewer artifacts, and enhance the clarity of microphotographs for data analysis. The immunohistochemistry protocol was followed as described previously . Briefly, the tissue sections were deparaffinized in xylene, rehydrated in ethanol solutions of decreasing concentrations and briefly rinsed in distilled water. The sections were heated in a citrate buffer (pH 6) for 15 min at 100 °C in a steam bath (Tefal, Minicompact, Rumilly, Haute-Savoie, France) and cooled to room temperature. Afterward, the tissue sections were incubated with a blocking buffer (ab64226, Abcam, Cambridge, UK) for 20 min and then with the appropriate combination of primary antibodies for one hour . 2.3. Data Acquisition and Analysis The hematoxylin and eosin slides were examined using a light microscope (BX40, Olympus, Tokyo, Japan). Samples were observed using a BX51 microscope (Olympus, Tokyo, Japan) at either 20× or 40× objective magnification and captured with a DP71 digital camera (Olympus, Tokyo, Japan). PDE4B, PDE4D, and SFRP5-positive cell counts were conducted in the regions of interest for all studied groups and delineated based on their distinct morphological characteristics, including epithelium and lamina propria. Cells were categorized as either positive (stained) or negative (unstained). Microphotographs were processed and analyzed using the ImageJ program (National Institutes of Health, Bethesda, MD, USA). The captured images were divided into squares measuring 20 μm × 20 μm at ×40 magnification to facilitate precise data collection and analysis. Positive and negative cells were counted as described previously . Briefly, only squares completely encompassing the region of interest were checked. To prevent duplicate counting of cells, every alternate section was utilized, and cells below the left and upper boundaries of each square were disregarded, with only those along the right and lower borders considered. The percentage of positive cells was computed across ten representative fields from each image, presented as mean ± standard deviation (SD), and compared across the regions of interest. Adobe Photoshop (Adobe Photoshop CS., 2004, Berkeley, CA, USA: Peachpit Press) was used for image assembly. Captured images in their original size at 40× magnification were used for the presentation of the images to better illustrate and showcase the overall morphology and spatial organization of the positive cells within the epithelium and lamina propria of the tissue examined. 2.4. Semi-Quantification The intensity of staining of the normal colonic mucosa and CRC tumor tissue was semi-quantitatively evaluated into four groups according to the staining reactivity: no reactivity = −, mild reactivity = +, moderate reactivity = ++, and strong reactivity = +++. 2.5. Statistical Data Analysis The GraphPad software was used for statistical data processing. To compare the expression of observed proteins in different substructures of colon mucosa (lamina propria and epithelium) between the control tissue and CRC stages, ordinary one-way ANOVA followed by Tukey’s multiple comparison tests was used. Within each observed group (control tissue and each examined CRC stage), a two-way ANOVA test followed by Šídák’s multiple comparisons test was used to compare the expression of the observed proteins between the different substructures (lamina propria and epithelium). Data were presented as mean ± standard deviation. A statistically significant difference was set at p < 0.05. 2.6. Transcriptomics Data Analysis We used TCGA Colon and Rectal Adenocarcinoma (COADREAD) gene expression by the RNAseq (polyA + IlluminaHiSeq) dataset, downloaded from the Xena database (University of California Santa Cruz) ( https://xenabrowser.net/datapages/ ; accessed on 4 May 2023). The dataset is combined from TCGA colon adenocarcinoma and rectum adenocarcinoma, with no history of neoadjuvant treatment, measured using the Illumina HiSeq 2000 RNA Sequencing platform at the University of North Carolina TCGA genome characterization center. This dataset shows the gene-level transcription estimates, as in log2(x + 1)-transformed RSEM normalized counts. Overall survival and gene expression of samples were exported as text and edited in Microsoft ® Excel ® 2019 MSO version 2305 (Microsoft Corp., Redmond, WA, USA). After data curation for double samples and samples that did not contain data for expression of studied factors, 431 patient samples remained for the analysis, of which 380 were primary tumors (PT), and 51 were normal solid tissue from the healthy resection margins (STN). Using GraphPad 9.0.0. software (GraphPad Software, San Diego, CA, USA), an unpaired t -test was used to compare the expression of PDE4B , PDE4D, and SFRP5 genes between tumor and solid tissue normal specimens. Survival analysis based on expression quartiles for each gene was completed using GraphPad software. The Kaplan–Meier method and the log-rank test were used for statistical analysis of the survival length. A statistically significant difference was set at p < 0.05. This study was approved by the Ethics Committee of the University of Mostar School of Medicine in accordance with the Helsinki Declaration (approval number 1271723 from 2 February 2023). Forty-three tumor tissue samples (12 of Dukes’ A, 11 of Dukes’ B, 10 of Dukes’ C, and 10 of Dukes’ D; ) were obtained from the Department of Pathology, Cytology, and Forensic Medicine at the University Hospital Mostar. Inclusion criteria encompassed patients with CRC older than 18 years, of both sexes and all cancer stages. Stratified randomization was used to ensure a balanced distribution of key characteristics (cancer stage) across treatment groups. Recruitment was conducted in collaboration with healthcare practitioners to identify eligible patients, supplemented by outreach through clinical settings. Participants were approached through direct contact during medical appointments, supplemented by follow-up communication and informational sessions to ensure thorough understanding and informed consent. The participants in this study did not receive any medications prior to the commencement of the study, since this was their first CRC diagnosis. Inclusion criteria for samples collected necessitated having an adequate amount of paraffin block material for immunohistochemistry (IHC) and complete clinical data. Exclusion criteria involved incomplete laboratory results, insufficient material for IHC, and lack of control tissue. Macroscopic examination and measurement of the tumor samples were conducted. Each sample contains tumor tissue and healthy control tissue. The tissues were dehydrated in increasing concentrations of ethanol solutions and embedded in paraffin. Paraffin blocks were cut using a microtome (Leica RM2155, Pittsburgh, PA, USA) in 4 μm sections to improve staining quality, reduce cell overlap, provide fewer artifacts, and enhance the clarity of microphotographs for data analysis. The immunohistochemistry protocol was followed as described previously . Briefly, the tissue sections were deparaffinized in xylene, rehydrated in ethanol solutions of decreasing concentrations and briefly rinsed in distilled water. The sections were heated in a citrate buffer (pH 6) for 15 min at 100 °C in a steam bath (Tefal, Minicompact, Rumilly, Haute-Savoie, France) and cooled to room temperature. Afterward, the tissue sections were incubated with a blocking buffer (ab64226, Abcam, Cambridge, UK) for 20 min and then with the appropriate combination of primary antibodies for one hour . The hematoxylin and eosin slides were examined using a light microscope (BX40, Olympus, Tokyo, Japan). Samples were observed using a BX51 microscope (Olympus, Tokyo, Japan) at either 20× or 40× objective magnification and captured with a DP71 digital camera (Olympus, Tokyo, Japan). PDE4B, PDE4D, and SFRP5-positive cell counts were conducted in the regions of interest for all studied groups and delineated based on their distinct morphological characteristics, including epithelium and lamina propria. Cells were categorized as either positive (stained) or negative (unstained). Microphotographs were processed and analyzed using the ImageJ program (National Institutes of Health, Bethesda, MD, USA). The captured images were divided into squares measuring 20 μm × 20 μm at ×40 magnification to facilitate precise data collection and analysis. Positive and negative cells were counted as described previously . Briefly, only squares completely encompassing the region of interest were checked. To prevent duplicate counting of cells, every alternate section was utilized, and cells below the left and upper boundaries of each square were disregarded, with only those along the right and lower borders considered. The percentage of positive cells was computed across ten representative fields from each image, presented as mean ± standard deviation (SD), and compared across the regions of interest. Adobe Photoshop (Adobe Photoshop CS., 2004, Berkeley, CA, USA: Peachpit Press) was used for image assembly. Captured images in their original size at 40× magnification were used for the presentation of the images to better illustrate and showcase the overall morphology and spatial organization of the positive cells within the epithelium and lamina propria of the tissue examined. The intensity of staining of the normal colonic mucosa and CRC tumor tissue was semi-quantitatively evaluated into four groups according to the staining reactivity: no reactivity = −, mild reactivity = +, moderate reactivity = ++, and strong reactivity = +++. The GraphPad software was used for statistical data processing. To compare the expression of observed proteins in different substructures of colon mucosa (lamina propria and epithelium) between the control tissue and CRC stages, ordinary one-way ANOVA followed by Tukey’s multiple comparison tests was used. Within each observed group (control tissue and each examined CRC stage), a two-way ANOVA test followed by Šídák’s multiple comparisons test was used to compare the expression of the observed proteins between the different substructures (lamina propria and epithelium). Data were presented as mean ± standard deviation. A statistically significant difference was set at p < 0.05. We used TCGA Colon and Rectal Adenocarcinoma (COADREAD) gene expression by the RNAseq (polyA + IlluminaHiSeq) dataset, downloaded from the Xena database (University of California Santa Cruz) ( https://xenabrowser.net/datapages/ ; accessed on 4 May 2023). The dataset is combined from TCGA colon adenocarcinoma and rectum adenocarcinoma, with no history of neoadjuvant treatment, measured using the Illumina HiSeq 2000 RNA Sequencing platform at the University of North Carolina TCGA genome characterization center. This dataset shows the gene-level transcription estimates, as in log2(x + 1)-transformed RSEM normalized counts. Overall survival and gene expression of samples were exported as text and edited in Microsoft ® Excel ® 2019 MSO version 2305 (Microsoft Corp., Redmond, WA, USA). After data curation for double samples and samples that did not contain data for expression of studied factors, 431 patient samples remained for the analysis, of which 380 were primary tumors (PT), and 51 were normal solid tissue from the healthy resection margins (STN). Using GraphPad 9.0.0. software (GraphPad Software, San Diego, CA, USA), an unpaired t -test was used to compare the expression of PDE4B , PDE4D, and SFRP5 genes between tumor and solid tissue normal specimens. Survival analysis based on expression quartiles for each gene was completed using GraphPad software. The Kaplan–Meier method and the log-rank test were used for statistical analysis of the survival length. A statistically significant difference was set at p < 0.05. 3.1. General Characteristics of CRC Visualized by Hematoxylin and Eosin Staining The normal adult colonic mucosa comprises surface columnar epithelium containing goblet cells, which also line the inner surface of the crypts within the lamina propria ( a). The nuclei of columnar and goblet cells are located at the basal ends of the cells. The muscularis mucosa borders with the submucosal connective tissue. In colorectal cancer (Dukes’ A and B), moderately-to-well-formed crypts or glands are observed, lined by atypical epithelial cells that are large and tall, featuring enlarged nuclei with prominent nucleoli ( b,c). These cells exhibit a loss of polarity and an increase in mitotic figures. Necrotic debris is visible inside the lumina of the crypts or glands, and there is a slight desmoplastic response of the stroma surrounding the tumor. In colorectal cancer (Dukes’ C and D), irregularly and poorly formed crypts or glands are present, lined by larger and taller atypical epithelial cells with larger nuclei and prominent nucleoli ( d–f). The loss of polarity is more pronounced, along with an increased number of mitotic figures. Necrotic debris is also evident within the lumen of adenocarcinomatous crypts or glands. Additionally, there is a desmoplastic response of the stroma surrounding the tumor. 3.2. Double Immunofluorescence Staining with PDE4B and SFRP5 Markers and Quantification of Immunoreactive Cells The healthy colorectal tissue control exhibited strong PDE4B reactivity in both the epithelium and lamina propria . In the lamina propria, there were 27.47% PDE4B-positive cells, while in the epithelium, the percentage was 4.78% ( a,f). Co-localization of PDE4B and SFRP5 within the same cell was observed only in lamina propria ( a). Similar to healthy colorectal tissue control, CRC stage Dukes’ A demonstrated strong expression of PDE4B in both epithelium and lamina propria . In the lamina propria, there were 42.64% of PDE4B-positive cells, while in the epithelium, the percentage was 2.82% ( b,f,g). Co-localization of PDE4B and SFRP5 in the same cell was observed in both lamina propria and epithelium ( b). CRC stage Dukes’ B exhibited strong expression of PDE4B in both epithelium and lamina propria . In the lamina propria, there were 5.44% PDE4B-positive cells, while in the epithelium, the percentage was 0.06% ( c,f,g). Co-localization of PDE4B and SFRP5 in the same cell was observed only in lamina propria ( c). CRC stage Dukes’ C displayed strong expression of PDE4B in both epithelium and lamina propria . In the lamina propria, there were 12.7% PDE4B-positive cells, while in the epithelium, the percentage was 0.13% ( d,f,g). Co-localization of PDE4B and SFRP5 in the same cell was observed in both lamina propria and epithelium ( d). CRC stage Dukes’ D had strong expression of PDE4B in both epithelium and lamina propria . In the lamina propria, 8.56% of PDE4B-positive cells were present, while in the epithelium, the percentage was 0.06% ( e–g). Co-localization of PDE4B and SFRP5 in the same cell was observed only in lamina propria ( e). The expression of PDE4B between the lamina propria and epithelium in the observed tissues revealed a statistically significant increase in the PDE4B-positive cells in the lamina propria compared to the epithelium in all examined tissues ( p < 0.05) ( a). Statistically significant differences were observed in the number of PDE4B-positive epithelial cells between control and CRC stage Dukes’ B, C, and D, as well as between CRC stage Dukes’ A and CRC stages Dukes’ B, C, and D ( f). Regarding the number of PDE4B-positive cells in lamina propria, a statistically significant difference was observed between control and all CRC Dukes’ stages, as well as CRC stage Dukes’ A and CRC stage Dukes’ B, C, and D. Additionally, CRC stage Dukes’ C had significantly more PDE4B-positive cells in the lamina propria than in the CRC stage Dukes’ B ( f). 3.3. Double Immunofluorescence Staining with PDE4D and SFRP5 Markers and Quantification of Immunoreactive Cells The healthy colorectal tissue control exhibited strong reactivity of PDE4D and SFRP5 in both the epithelium and lamina propria , with a similar percentage of positive cells. Namely, in the lamina propria, there were 27.5% and 26.8% of PDE4D- and SFRP5-positive cells, while in the epithelium, the percentages were 4.57% and 4.82%, respectively ( a,f,g). Co-localization of PDE4D and SFRP5 in the same cell was observed only in the lamina propria ( a). Similar to healthy colorectal tissue control, CRC stage Dukes’ A demonstrated moderate expression of PDE4D and SFRP5 in the epithelium and strong PDE4D and SFRP5 in the lamina propria . In the lamina propria, there were 42.22% and 40.93% of PDE4D- and SFRP5-positive cells, while in the epithelium, the percentages were 2.27% and 2.33%, respectively ( b,f,g). Co-localization of PDE4D and SFRP5 in the same cell was observed in both the lamina propria and epithelium ( b). CRC stage Dukes’ B exhibited mild expression of PDE4D in the epithelium and strong expression in the lamina propria, while SFRP5 displayed moderate expression in the lamina propria and strong expression in the epithelium . In the lamina propria, there were 5.27% and 5.56% of PDE4D- and SFRP5-positive cells, respectively, while in the epithelium, the percentages were the same at 0.07% ( c,f,g). Co-localization of PDE4D and SFRP5 in the same cell was observed only in the lamina propria ( c). CRC stage Dukes’ C displayed moderate expression of PDE4D in both the epithelium and lamina propria, while SFRP5 displayed strong expression in both structures . In the lamina propria, there were 12.48% and 12.73% of PDE4D- and SFRP5-positive cells, while in the epithelium, the percentages were 0.07% and 0.08%, respectively ( d,f,g). Co-localization of PDE4D and SFRP5 in the same cell was observed in both the lamina propria and epithelium ( d). CRC stage Dukes’ D had mild expression of PDE4D in both the epithelium and lamina propria, while SFRP5 displayed mild expression in the epithelium and moderate expression in the lamina propria . In the lamina propria, there were 8.6% and 8.5% of PDE4D- and SFRP5-positive cells, respectively, while in the epithelium, the percentages were the same at 0.07% ( e–g). Co-localization of PDE4D and SFRP5 in the same cell was observed only in the lamina propria ( e). The expression of PDE4D between the lamina propria and epithelium demonstrated a statistically significant increase in the percentage of PDE4D-positive cells in control tissues as well as in Dukes’ B and Dukes’ D stages of CRC ( p < 0.0001 and p < 0.01, respectively) ( b). No significant differences were detected in the Dukes’ A and Dukes’ C stages ( b). Statistically significant differences were observed in the number of PDE4D-positive epithelial cells between control and CRC stage Dukes’ A, B, and D, as well as between CRC stage Dukes’ C and CRC stage Dukes’ A, B, and D ( f). Concerning the number of PDE4D-positive cells in lamina propria, a statistically significant difference was observed between control and CRC stage Dukes’ A, C, and D, as well as CRC stage Dukes’ B and CRC stage Dukes’ A, C, and D. Additionally, CRC stage Dukes’ A had significantly more PDE4D-positive cells in the lamina propria than in the CRC stage Dukes’ D ( f). Regarding the expression of SFRP5 in the lamina propria compared to the epithelium, a statistically significant increase in the percentage of SFRP5-positive cells was noticed in the lamina propria in all examined tissues ( p < 0.0001) except for Dukes’ D, where no significant difference was found ( c). Statistically significant differences were observed in the number of SFRP5-positive cells in the lamina propria between control and all CRC stages, as well as between CRC stage Dukes’ A and CRC stage Dukes’ D ( i). In the epithelium, there was no statistically significant difference between groups ( h). 3.4. Differential Expression We extracted CRC tumor tissue and control tissue RNA-sequencing data from the COADREAD study. Gene expression of all three studied genes, PDE4B ( p < 0.0001), PDE4D ( p < 0.0001), and SFRP5 ( p < 0.0001), was significantly lower in tumor tissue in comparison to the control tissue . 3.5. Survival Analysis The survival rates concerning the high and low mRNA expressions of PDE4B, PDE4D, and SFRP5 in CRC were analyzed . There was no significant difference in survival times in colorectal carcinoma between the high- and low-expression groups of PDE4B and PDE4D ( p = 0.6966 and p = 0.8914, respectively). However, a statistically significant difference ( p = 0.03824) in survival times was found between the high and low SFRP5 expression groups, where low expression of the SFRP5 denoted better survival probability . The normal adult colonic mucosa comprises surface columnar epithelium containing goblet cells, which also line the inner surface of the crypts within the lamina propria ( a). The nuclei of columnar and goblet cells are located at the basal ends of the cells. The muscularis mucosa borders with the submucosal connective tissue. In colorectal cancer (Dukes’ A and B), moderately-to-well-formed crypts or glands are observed, lined by atypical epithelial cells that are large and tall, featuring enlarged nuclei with prominent nucleoli ( b,c). These cells exhibit a loss of polarity and an increase in mitotic figures. Necrotic debris is visible inside the lumina of the crypts or glands, and there is a slight desmoplastic response of the stroma surrounding the tumor. In colorectal cancer (Dukes’ C and D), irregularly and poorly formed crypts or glands are present, lined by larger and taller atypical epithelial cells with larger nuclei and prominent nucleoli ( d–f). The loss of polarity is more pronounced, along with an increased number of mitotic figures. Necrotic debris is also evident within the lumen of adenocarcinomatous crypts or glands. Additionally, there is a desmoplastic response of the stroma surrounding the tumor. The healthy colorectal tissue control exhibited strong PDE4B reactivity in both the epithelium and lamina propria . In the lamina propria, there were 27.47% PDE4B-positive cells, while in the epithelium, the percentage was 4.78% ( a,f). Co-localization of PDE4B and SFRP5 within the same cell was observed only in lamina propria ( a). Similar to healthy colorectal tissue control, CRC stage Dukes’ A demonstrated strong expression of PDE4B in both epithelium and lamina propria . In the lamina propria, there were 42.64% of PDE4B-positive cells, while in the epithelium, the percentage was 2.82% ( b,f,g). Co-localization of PDE4B and SFRP5 in the same cell was observed in both lamina propria and epithelium ( b). CRC stage Dukes’ B exhibited strong expression of PDE4B in both epithelium and lamina propria . In the lamina propria, there were 5.44% PDE4B-positive cells, while in the epithelium, the percentage was 0.06% ( c,f,g). Co-localization of PDE4B and SFRP5 in the same cell was observed only in lamina propria ( c). CRC stage Dukes’ C displayed strong expression of PDE4B in both epithelium and lamina propria . In the lamina propria, there were 12.7% PDE4B-positive cells, while in the epithelium, the percentage was 0.13% ( d,f,g). Co-localization of PDE4B and SFRP5 in the same cell was observed in both lamina propria and epithelium ( d). CRC stage Dukes’ D had strong expression of PDE4B in both epithelium and lamina propria . In the lamina propria, 8.56% of PDE4B-positive cells were present, while in the epithelium, the percentage was 0.06% ( e–g). Co-localization of PDE4B and SFRP5 in the same cell was observed only in lamina propria ( e). The expression of PDE4B between the lamina propria and epithelium in the observed tissues revealed a statistically significant increase in the PDE4B-positive cells in the lamina propria compared to the epithelium in all examined tissues ( p < 0.05) ( a). Statistically significant differences were observed in the number of PDE4B-positive epithelial cells between control and CRC stage Dukes’ B, C, and D, as well as between CRC stage Dukes’ A and CRC stages Dukes’ B, C, and D ( f). Regarding the number of PDE4B-positive cells in lamina propria, a statistically significant difference was observed between control and all CRC Dukes’ stages, as well as CRC stage Dukes’ A and CRC stage Dukes’ B, C, and D. Additionally, CRC stage Dukes’ C had significantly more PDE4B-positive cells in the lamina propria than in the CRC stage Dukes’ B ( f). The healthy colorectal tissue control exhibited strong reactivity of PDE4D and SFRP5 in both the epithelium and lamina propria , with a similar percentage of positive cells. Namely, in the lamina propria, there were 27.5% and 26.8% of PDE4D- and SFRP5-positive cells, while in the epithelium, the percentages were 4.57% and 4.82%, respectively ( a,f,g). Co-localization of PDE4D and SFRP5 in the same cell was observed only in the lamina propria ( a). Similar to healthy colorectal tissue control, CRC stage Dukes’ A demonstrated moderate expression of PDE4D and SFRP5 in the epithelium and strong PDE4D and SFRP5 in the lamina propria . In the lamina propria, there were 42.22% and 40.93% of PDE4D- and SFRP5-positive cells, while in the epithelium, the percentages were 2.27% and 2.33%, respectively ( b,f,g). Co-localization of PDE4D and SFRP5 in the same cell was observed in both the lamina propria and epithelium ( b). CRC stage Dukes’ B exhibited mild expression of PDE4D in the epithelium and strong expression in the lamina propria, while SFRP5 displayed moderate expression in the lamina propria and strong expression in the epithelium . In the lamina propria, there were 5.27% and 5.56% of PDE4D- and SFRP5-positive cells, respectively, while in the epithelium, the percentages were the same at 0.07% ( c,f,g). Co-localization of PDE4D and SFRP5 in the same cell was observed only in the lamina propria ( c). CRC stage Dukes’ C displayed moderate expression of PDE4D in both the epithelium and lamina propria, while SFRP5 displayed strong expression in both structures . In the lamina propria, there were 12.48% and 12.73% of PDE4D- and SFRP5-positive cells, while in the epithelium, the percentages were 0.07% and 0.08%, respectively ( d,f,g). Co-localization of PDE4D and SFRP5 in the same cell was observed in both the lamina propria and epithelium ( d). CRC stage Dukes’ D had mild expression of PDE4D in both the epithelium and lamina propria, while SFRP5 displayed mild expression in the epithelium and moderate expression in the lamina propria . In the lamina propria, there were 8.6% and 8.5% of PDE4D- and SFRP5-positive cells, respectively, while in the epithelium, the percentages were the same at 0.07% ( e–g). Co-localization of PDE4D and SFRP5 in the same cell was observed only in the lamina propria ( e). The expression of PDE4D between the lamina propria and epithelium demonstrated a statistically significant increase in the percentage of PDE4D-positive cells in control tissues as well as in Dukes’ B and Dukes’ D stages of CRC ( p < 0.0001 and p < 0.01, respectively) ( b). No significant differences were detected in the Dukes’ A and Dukes’ C stages ( b). Statistically significant differences were observed in the number of PDE4D-positive epithelial cells between control and CRC stage Dukes’ A, B, and D, as well as between CRC stage Dukes’ C and CRC stage Dukes’ A, B, and D ( f). Concerning the number of PDE4D-positive cells in lamina propria, a statistically significant difference was observed between control and CRC stage Dukes’ A, C, and D, as well as CRC stage Dukes’ B and CRC stage Dukes’ A, C, and D. Additionally, CRC stage Dukes’ A had significantly more PDE4D-positive cells in the lamina propria than in the CRC stage Dukes’ D ( f). Regarding the expression of SFRP5 in the lamina propria compared to the epithelium, a statistically significant increase in the percentage of SFRP5-positive cells was noticed in the lamina propria in all examined tissues ( p < 0.0001) except for Dukes’ D, where no significant difference was found ( c). Statistically significant differences were observed in the number of SFRP5-positive cells in the lamina propria between control and all CRC stages, as well as between CRC stage Dukes’ A and CRC stage Dukes’ D ( i). In the epithelium, there was no statistically significant difference between groups ( h). We extracted CRC tumor tissue and control tissue RNA-sequencing data from the COADREAD study. Gene expression of all three studied genes, PDE4B ( p < 0.0001), PDE4D ( p < 0.0001), and SFRP5 ( p < 0.0001), was significantly lower in tumor tissue in comparison to the control tissue . The survival rates concerning the high and low mRNA expressions of PDE4B, PDE4D, and SFRP5 in CRC were analyzed . There was no significant difference in survival times in colorectal carcinoma between the high- and low-expression groups of PDE4B and PDE4D ( p = 0.6966 and p = 0.8914, respectively). However, a statistically significant difference ( p = 0.03824) in survival times was found between the high and low SFRP5 expression groups, where low expression of the SFRP5 denoted better survival probability . As the phosphodiesterase enzyme family (PDEs) plays a crucial role in signal transduction, it displays promising pharmacological targets in a variety of diseases. Due to the lack of biomarkers for early diagnosis, tumor resection is frequently delayed until the disease has progressed significantly, leaving radiotherapy and chemotherapy as the primary treatment options. PDE4B and PDE4D have recently been reported as oncogenes in various human cancers. They play critical roles in cancer progression by degrading cAMP. This degradation influences cancer development by promoting cell proliferation, reducing apoptosis, and facilitating metastasis through pathways such as phosphatidylinositol-3 kinase/protein kinase B/mammalian target of rapamycin (PI3K/AKT/mTOR) and mitogen-activated protein kinases (Ras/Raf-MEK-MAPKs) . PDE4B has been shown to contribute to a pro-inflammatory environment within tumors, which can support tumor growth and immune evasion . PDE4D, on the other hand, can alter gene expression and impact the tumor microenvironment, thereby influencing various aspects of cancer biology, including the response to therapies . Additionally, SFRP5 acts as a soluble modulator of Wnt signaling by binding to Wnt proteins and preventing them from interacting with their receptors. It also acts as a tumor suppressor, preventing uncontrolled cell growth and differentiation. It also has anti-inflammatory properties and can inhibit epithelial-mesenchymal transition (EMT), reducing metastasis . Together, these genes impact tumor initiation, progression, and metastasis through distinct but interrelated mechanisms. However, their expression and significance in CRC have not been elucidated. Therefore, we investigated the protein and gene expression patterns of PDE4B , PDE4D , and SFRP5 to provide valuable insights into their involvement in CRC. Increased PDE4B expression correlates with relapsed colorectal cancer cell lines, suggesting its potential as a prognostic molecular marker in CRC . Pleiman et al. found that loss of PDE4B function in the ApcMin/+ mouse significantly increases the number of colonic adenomas . In this context, they proposed that a feedback mechanism could explain the protective role of PDE4B. Similarly, our result showed the highest PDE4B expression in the control tissue and Dukes’ A stage. According to this model, PDE4B is initially triggered by early oncogenic stress related to cAMP. However, as observed in cases of advanced human colon cancer, it is subsequently deactivated through epigenetic suppression. Mahmood et al. discovered a higher expression of PDE4B in patients with colorectal neoplasia and proposed that overexpression of PDE4B appears as a malfunctioning protein in the noncancerous mucosa of the colon. This finding suggests a tendency toward reduced PDE4B activity in the colon mucosa of patients with colorectal neoplasia. These results follow our findings of lower PDE4B expression in the mucosa of CRC Dukes’ stages B, C, and D as well as Illumina Hi-Seq COADREAD TCGA survival analysis data. Strong evidence suggests that PDE4D contributes to tumor advancement in lung, central nervous system (CNS), and skin cancers. Additionally, certain isoforms of PDE4D exhibit contrasting functions in prostate cancer . Kim et al. suggested that PDE4D has been implicated in the development and progression of CRC . Nummela et al. found that overexpression of PDE4D leads to inhibition of the proliferation of CRC cells through PDE4 inhibitors . In our study, PDE4D expression was lowest in the Dukes’ D stage compared to the controls and other Dukes’ stages. However, this might be tissue-specific since the finding was the opposite in pancreatic ductal adenocarcinoma (PDAC) . Namely, Liu et al. found that high expression of PDE4D correlates with poor prognosis and clinical progression of PDAC. PDE4D has also been implicated in other cancers, such as prostate cancer, medulloblastoma, and leukemia, indicating that PDE4D may be a more general target for cancer therapy . Our finding suggests that a lower level of PDE4D, as in the CRC stage Dukes’ D, will lead to a worse CRC prognosis in regards to metastatic spreading of the disease. Illumina Hi-Seq COADREAD TCGA data regarding survival analysis also confirm the poorer prognosis in patients with a lower PDE4D expression. Blocking PDE4D affects the levels of protein kinase A (PKA), sirtuin 1 (Sirt1), protein kinase B (Akt), and BCL2 apoptosis regulator/apoptosis regulator BAX (Bcl-2/Bax), which are involved in signaling pathways that regulate endocrine reactions, stress resilience, neuronal autophagy, and cell death . These studies imply that PDE4D is aberrantly expressed in CRC cells and patients with CRC, which may contribute to the malignant phenotype. However, further studies are required to fully understand the mechanisms underlying PDE4D overexpression in CRC and develop targeted therapies. In our study, SFRP5 expression was lowest in stage Dukes’ D. Our immunohistochemistry and RNA sequencing data results align with Kirana et al., where analysis of an independent tissue cohort from The Cancer Genome Atlas database on 637 patients revealed significantly lower SFRP5 RNA expression in CRC tumor tissue compared with adjacent normal mucosa. The same study found that the levels of SFRP5 were significantly lower in CRC patients with either vascular invasion or liver metastasis, which corresponds to stage Dukes’ D . At the same time, they also found that a high serum level of circulating SFRP5 (cSFRP5) is associated with longer disease-free survival . The study of Huang et al. noted that the mRNA levels of SFRP5 were significantly downregulated in 80% of CRC . Additionally, Liu et al. also found that expression levels of SFRP5 were reduced in gastric cancer . This finding implies that a higher level of SFRP5 might predict longer disease-free survival and, therefore, might serve as a prediction marker for CRC. This finding opposes the Illumina Hi-Seq COADREAD TCGA survival analysis data on SFRP5 expression analyzed in our study, where we noted better prognosis in patients with lower SFRP5 expression. Additionally, SFRP5 hypermethylation has been associated with the risk of colorectal cancer, and this hypermethylation can lead to the activation of the Wnt signaling pathway, which is a key driver of CRC progression . Studies also show that, in colorectal cancer, kidney cancer, and breast and prostate cancer, a decrease in the level of SFRP, but also an increase in the level of SFRP expression, depending on the degree of malignancy, has been observed. This indicates the probability that high concentrations inhibit and low concentrations potentiate Wnt signaling . We also analyzed the expression levels of PDE4B , PDE4D , and SFRP5 genes in CRC tissues compared to control tissues using RNA-sequencing data from the UCSC Xena browser. The findings reveal significant differences in the expression of these genes between CRC patients and healthy controls, with notable implications for patient prognosis. Namely, our results demonstrate that PDE4B , PDE4D , and SFRP5 are significantly under-expressed in CRC tissues compared to control tissues. The Kaplan–Meier survival analysis and the log-rank (Mantel–Cox) test revealed distinct prognostic implications based on the expression levels of PDE4B , PDE4D , and SFRP5 . Patients with low levels of PDE4B and SFRP5 exhibited significantly longer overall survival compared to those with low levels of PDE4D . This finding underscores the differential impact of these genes on CRC patient outcomes. The data align with our immunohistochemical results and might suggest a potential tumor-suppressive role for these genes in CRC. The reduced expression of these genes in tumor tissues may indicate their involvement in disrupted pathways during colorectal carcinogenesis. At present, there are no FDA-approved cancer treatments that target PDE4B, PDE4D, or SFRP5. Nevertheless, as previously mentioned, research suggests that PDE4B and PDE4D are significant contributors to the progression of cancer and inflammation . The relationship between chronic inflammation and cancer development has been widely recognized. Chronic inflammation may establish a tumor-promoting environment by continuously producing pro-inflammatory cytokines and other mediators. This process can result in DNA damage and genomic instability, which are critical factors in the initiation and progression of cancer . For example, chronic inflammation has been linked to the development of numerous malignancies, such as chronic bronchitis and lung cancer , as well as the induction of bladder cancer . The risk of lung and colorectal malignancies is considerably elevated by conditions such as chronic bronchitis and inflammatory bowel diseases, including ulcerative colitis and Crohn’s disease . The critical role of inflammation in the etiology of cancer is further evidenced by the observation that the incidence of cancer can be reduced by managing inflammatory conditions with anti-inflammatory agents . Consequently, the comprehension of the cellular mechanisms that underlie inflammation-induced cancer is an essential area of ongoing research . PDE4 inhibitors, including roflumilast and apremilast, used at the present to treat inflammatory conditions, are currently being investigated for their potential in cancer therapy. These inhibitors have the potential to inhibit tumor growth and enhance the efficacy of other treatments . Furthermore, preclinical studies have demonstrated that PDE4D inhibitors can potentially restrict prostate cancer cell proliferation . Ongoing research is being conducted on SFRP5, a modulator of the Wnt signaling pathway. This pathway is essential for cancer progression, and therapies designed to restore or replicate SFRP5 activity are currently being evaluated for their potential to treat cancer . These targets hold promise for future cancer treatments as research continues to advance. Our study lacks detailed information on macroscopic examination and measurements, such as tumor size, and morphological characteristics. This limitation arises from the variability within our patient cohort: some patients did not undergo surgical operations, some had surgeries at different facilities, and others sought treatment abroad for better healthcare standards. These factors significantly limit our ability to provide comprehensive data. In conclusion, our results emphasize the potential clinical utility of PDE4B , PDE4D , and SFRP5 , genes that impact tumor initiation, progression, and metastasis, as biomarkers for CRC. Considering significantly lower gene expression, aligned with our immunohistochemical data in tumor tissue in comparison to the control tissue, as well as the significantly poorer survival rate in the case of its higher expression, we can hypothesize that SFRP5 is the most promising biomarker for CRC out of the observed proteins. Assessing their expression levels could provide valuable prognostic information and help open new avenues for biomarker-driven diagnosis and therapy. For instance, targeting pathways affected by these genes might offer new avenues for CRC treatment. Furthermore, the significant correlation between gene expression levels and patient survival suggests that therapeutic modulation of PDE4B , PDE4D , and SFRP5 activity could improve patient outcomes. This highlights the need for further studies to explore targeted therapies that can modulate these pathways effectively. However, further research is needed to elucidate the potential correlations between PDE4B, PDE4D, and SFRP5 and the signaling network in CRC. |
IMPELLA COMPARED TO VENOARTERIAL EXTRACORPOREAL MEMBRANE OXYGENATION IN CARDIOGENIC SHOCK: A SYSTEMATIC REVIEW AND META-ANALYSIS OF PROPENSITY SCORE-MATCHED STUDIES | dbb6d89d-ced4-46bb-80ef-957d70baab16 | 11939094 | Surgery[mh] | Cardiogenic shock (CS) is a life-threatening condition associated with multiorgan failure and death. CS is defined as a state of end-organ hypoperfusion caused by left ventricular (LV), right ventricular, or biventricular myocardial injury resulting in systolic and/or diastolic myocardial pump failure ( , ). Acute myocardial infarction (AMI) with LV dysfunction is the most frequent cause of CS ( ), accounting for 81% of patients ( ). Other causes of CS include, but are not limited to, cardiomyopathy, postcardiac arrest and myocarditis. While in-hospital mortality in patients with CS has seen an improvement over time ( ), possibly related to improved early revascularization, longer-term mortality in CS has remained relatively consistent over the past two decades, with only 40% to 50% of patients surviving beyond 6 months ( , ), highlighting a persistent unmet need for improved treatment strategies. Mechanical circulatory support (MCS) is recommended for patients in CS who have not stabilized despite vasoactive pharmacotherapies ( ). The two most commonly used MCS devices recommended in clinical guidelines, include veno-arterial membrane oxygenation (VA ECMO) and Impella, which have different mechanisms of action ( – ). While intra-aortic balloon pump (IABP) has been used in CS, multiple clinical guidelines highlight that routine use of IABP is not recommended in patients with CS ( – ). Impella is a microaxial ventricular assist device that is inserted percutaneously or surgically and provides continuous pumping independent of the cardiac cycle (Impella 2.5, CP, 5.0 and 5.5 available at the time of this analysis). The Impella device improves hemodynamics by directly unloading the left ventricle, reducing end-diastolic wall stress and immediately decreasing pulmonary capillary wedge pressure hence improving organ perfusion and reducing myocardial oxygen demand in patients with CS ( ). VA ECMO is an alternative high output MCS device that provides pulmonary and cardiac support. This has a cascade of consequent positive effects on heart function and physiology, ultimately improving end-organ perfusion. While VA ECMO is commonly used to manage CS, one of the potential disadvantages of VA ECMO is that the oxygenated blood returning to the body flows retrograde in the aorta, which causes an increase in LV afterload, and may be associated with LV distension and pulmonary edema ( ). While there have been numerous attempts to conduct randomized controlled trials (RCTs) of Impella and VA ECMO, many trials have been terminated prematurely, illustrating the difficulty in conducting adequately powered RCTs in patients with acute CS. While there exist several meta-analyses comparing Impella versus VA ECMO in CS, these are mostly based on nonrandomized observational data ( , , ). Propensity score-matched or adjusted comparative studies represent an improvement on unadjusted comparative studies given the attempt to adjust for important differences in baseline characteristics of the cohorts that may impact outcomes. The objective of this study was to conduct a systematic review and meta-analysis of propensity score-matched/adjusted studies to assess the clinical outcomes of Impella versus VA ECMO in patients with CS.
Literature search A systematic literature review was performed using the Embase.com (Embase and MEDLINE) search platform, the Cochrane Library, Clinicaltrials.gov , WHO ICTRP, ANZCTR, and PROSPERO up to 23 July 2023 to identify published propensity score-matched or adjusted studies of Impella versus VA ECMO in subjects with CS. The were no date or language restrictions on the search (see Appendix A, Table A 1 and Table A 2 for search details, http://links.lww.com/SHK/C312 ). Articles were scrutinized independently by two reviewers with discrepancies resolved through discussion until a consensus was reached. To be included, the study had to: be a comparative study of Impella versus VA-ECMO, include patients with CS and must have reported survival data. Studies were excluded if: outcomes data were not stratified by intervention type (i.e., Impella alone vs. VA ECMO alone), the study was incomplete with no available data, or the study was not available in full-text form (e.g., conference abstracts or posters were excluded) (see Table A 3, http://links.lww.com/SHK/C312 ). Furthermore, studies that did not propensity score match patients or did not conduct adjusted analyses to account for variation in patients’ baseline demographic and/or disease characteristics were also excluded. All studies were assessed for risk of bias using the National Heart, Lung and Blood Institute’s Quality Assessment Tool for observational cohort and cross-sectional studies. Studies were provided a rating of good, fair or poor quality, whereby a poor-quality rating translates to high risk of bias and a good quality rating translates to low risk of bias. The quality of evidence for each outcome was assessed using the grading of recommendations, assessment, development and evaluations approach. Risk of bias and quality assessment was conducted independently by two reviewers with discrepancies discussed and resolved between them. Outcomes and statistical analysis The primary outcome of interest was short-term mortality, defined as either 30-day or in-hospital mortality. The secondary outcome of interest was bleeding events requiring transfusion. Data from the full-text publication of each study were extracted independently by a single reviewer and cross-checked by a second reviewer. Meta-analyses were performed in Review Manager (RevMan, v. 5.3) using the odds ratio (OR) with 95% confidence intervals (CIs) using the random effects inverse variance method. Results were presented visually in forest plots, displaying the ORs of the individual studies as well as the meta-analyzed OR. A two-tailed P value <0.05 was considered statistically significant. Heterogeneity of the included studies was quantified with the I 2 statistic whereby an I 2 of 0% to 40% may indicate heterogeneity which might not be important, 30% to 60% may indicate moderate heterogeneity, 50% to 90% may indicate substantial heterogeneity and 75% to 100% may indicate considerable heterogeneity. In addition, heterogeneity was assessed using the chi-squared test with a P value of 0.10 used to determine statistically significant heterogeneity (Cochrane handbook version 6.4) ( ).
A systematic literature review was performed using the Embase.com (Embase and MEDLINE) search platform, the Cochrane Library, Clinicaltrials.gov , WHO ICTRP, ANZCTR, and PROSPERO up to 23 July 2023 to identify published propensity score-matched or adjusted studies of Impella versus VA ECMO in subjects with CS. The were no date or language restrictions on the search (see Appendix A, Table A 1 and Table A 2 for search details, http://links.lww.com/SHK/C312 ). Articles were scrutinized independently by two reviewers with discrepancies resolved through discussion until a consensus was reached. To be included, the study had to: be a comparative study of Impella versus VA-ECMO, include patients with CS and must have reported survival data. Studies were excluded if: outcomes data were not stratified by intervention type (i.e., Impella alone vs. VA ECMO alone), the study was incomplete with no available data, or the study was not available in full-text form (e.g., conference abstracts or posters were excluded) (see Table A 3, http://links.lww.com/SHK/C312 ). Furthermore, studies that did not propensity score match patients or did not conduct adjusted analyses to account for variation in patients’ baseline demographic and/or disease characteristics were also excluded. All studies were assessed for risk of bias using the National Heart, Lung and Blood Institute’s Quality Assessment Tool for observational cohort and cross-sectional studies. Studies were provided a rating of good, fair or poor quality, whereby a poor-quality rating translates to high risk of bias and a good quality rating translates to low risk of bias. The quality of evidence for each outcome was assessed using the grading of recommendations, assessment, development and evaluations approach. Risk of bias and quality assessment was conducted independently by two reviewers with discrepancies discussed and resolved between them.
The primary outcome of interest was short-term mortality, defined as either 30-day or in-hospital mortality. The secondary outcome of interest was bleeding events requiring transfusion. Data from the full-text publication of each study were extracted independently by a single reviewer and cross-checked by a second reviewer. Meta-analyses were performed in Review Manager (RevMan, v. 5.3) using the odds ratio (OR) with 95% confidence intervals (CIs) using the random effects inverse variance method. Results were presented visually in forest plots, displaying the ORs of the individual studies as well as the meta-analyzed OR. A two-tailed P value <0.05 was considered statistically significant. Heterogeneity of the included studies was quantified with the I 2 statistic whereby an I 2 of 0% to 40% may indicate heterogeneity which might not be important, 30% to 60% may indicate moderate heterogeneity, 50% to 90% may indicate substantial heterogeneity and 75% to 100% may indicate considerable heterogeneity. In addition, heterogeneity was assessed using the chi-squared test with a P value of 0.10 used to determine statistically significant heterogeneity (Cochrane handbook version 6.4) ( ).
Literature search results A total of 2,845 citations were identified, of which six studies met the eligibility criteria describing propensity score matching or adjusted analyses of Impella versus VA ECMO in CS ( – ). One study was later excluded as it did not provide propensity score-matched data for either of the outcomes of interest, that is, short-term mortality or bleeding events requiring transfusion ( ). No RCTs comparing Impella and VA ECMO were identified in the search (refer to Fig. A 1 for PRISMA flowchart, http://links.lww.com/SHK/C312 ). Characteristics of the included studies The five studies contributing to the meta-analysis were retrospective, nonrandomized studies, four of which conducted propensity score matching of participants at baseline, and one conducted a propensity score adjusted analyses to balance observed covariates ( ). Table provides an overview of the characteristics and populations in the included studies. Two of the included studies ( , ) did not report the type of Impella device used to treat patients, while all patients in the remaining three studies used the Impella CP or Impella 2.5. Four studies assessed in-hospital mortality/survival and one study assessed 30-day mortality ( ). All except one study ( ) reported the proportion of patients experiencing bleeding events requiring transfusion as a key safety outcome. Populations enrolled in the included studies The number of patients contributing matched or adjusted data in the individual studies ranged between 80 patents ( ) to 900 patients ( ) (Table ). Patients in the included studies were predominantly males, ranging from 70% to 86% of the total trial populations. The mean age of patients was consistent across studies, ranging from 60.0 to 67.7 years old. Age was a variable included in propensity score matching of participants or adjustment of analyses in all studies, and hence was well balanced between treatment arms in each study. Sex was accounted for in matching/adjusting in all but one study ( ); however, baseline data indicated no difference in sex distribution across the two treatment arms. The most common etiology of shock in the included studies was AMI-CS. Three studies specified that participants had to have AMI-CS within the eligibility criteria ( – ), one of which stated participants had to have out-of-hospital cardiac arrest due to AMI-CS ( ). The study by Karatolios (2021) did not specify that participants had to have AMI-CS to be eligible for study entry; however, the ‘etiology of shock’ was a covariate accounted for in propensity score matching, and as a result, 86% of participants in both matched treatment arms had AMI-CS ( ). The study by Wernly (2021) included participants with either AMI-CS or cardiac arrest CS; however, the proportion of participants with each etiology differed between treatment arms with 89% of the Impella cohort and 14% of the VA ECMO cohort having AMI-CS ( ). ‘Shock after myocardial infarction’ was a covariate adjusted for in the analyses in the study; hence, it is considered that this imbalance was addressed ( ). The covariates included in propensity score matching in the included studies are summarized in Table . The covariates considered to impact study results varied considerably between studies. Variables used to match participants or adjust analyses in at least two of the included studies comprise sex, race, vasoactive score ( ), pH, PaO 2 /FiO 2 (ratio of partial pressure of oxygen in blood [PaO 2 , mmHg] and the fraction of oxygen in the inhaled air [FiO 2 ]), lactate, comorbidities (Charlson Comorbidity Index ( )), prior cardiopulmonary resuscitation (CPR), and region. Furthermore, the study by Lemor et al. (2020) reported survival and bleeding in a matched population as well as survival and bleeding in an unmatched population with an adjusted analysis ( ). Both the matched and adjusted analyses resulted in the same finding regarding the direction and significance of the treatment effect. This meta-analysis used the dichotomous data from the matched analyses to calculate ORs for survival and bleeding. Outcomes Based on the grading of recommendations assessment development and evaluation assessment, the quality of the evidence was assessed as low for both short-term mortality and bleeding events requiring transfusion. Mortality The meta-analysis of the ORs from the included studies demonstrated that Impella statistically significantly reduced the odds of in-hospital/30-day mortality compared to VA ECMO (OR = 0.57 [95% CI: 0.44, 0.74]; P < 0.0001). The I 2 statistic of 27% indicated heterogeneity might not be important. All included studies reported lower absolute in-hospital/30-day mortality for patients treated with Impella versus VA ECMO with the largest two studies ( , ), reporting statistically significant differences in favor of Impella. As depicted in Figure , the ORs of all included studies were consistent in direction and magnitude of effect in favor of Impella, except for Wernly et al. (2021). In absolute terms, the proportion of Impella patients that died in-hospital/within 30 days (pooled: 39.6%) was lower than the proportion of VA ECMO patients (pooled: 53.8%). With a 14.2% difference in favor of Impella versus VA ECMO, corresponding to a number needed to treat (NNT) with Impella to avoid one death of seven. Bleeding events requiring transfusions As can be seen in Figure , the meta-analysis demonstrated statistically significantly lower odds of bleeding events requiring transfusion in patients treated with Impella compared to VA ECMO (OR = 0.61 [95% CI: 0.46, 0.80]; P = 0.0004). No heterogeneity was detected (I 2 = 0%). All studies reported a lower proportion of patients with bleeding events requiring transfusion in the Impella cohort (pooled: 19.9%) compared to the VA ECMO cohort (pooled: 28.8%), with a difference of 8.9% in favor of Impella. Thus, the NNT with Impella relative to VA ECMO to avoid one bleeding event requiring transfusion is 12.
A total of 2,845 citations were identified, of which six studies met the eligibility criteria describing propensity score matching or adjusted analyses of Impella versus VA ECMO in CS ( – ). One study was later excluded as it did not provide propensity score-matched data for either of the outcomes of interest, that is, short-term mortality or bleeding events requiring transfusion ( ). No RCTs comparing Impella and VA ECMO were identified in the search (refer to Fig. A 1 for PRISMA flowchart, http://links.lww.com/SHK/C312 ).
The five studies contributing to the meta-analysis were retrospective, nonrandomized studies, four of which conducted propensity score matching of participants at baseline, and one conducted a propensity score adjusted analyses to balance observed covariates ( ). Table provides an overview of the characteristics and populations in the included studies. Two of the included studies ( , ) did not report the type of Impella device used to treat patients, while all patients in the remaining three studies used the Impella CP or Impella 2.5. Four studies assessed in-hospital mortality/survival and one study assessed 30-day mortality ( ). All except one study ( ) reported the proportion of patients experiencing bleeding events requiring transfusion as a key safety outcome.
The number of patients contributing matched or adjusted data in the individual studies ranged between 80 patents ( ) to 900 patients ( ) (Table ). Patients in the included studies were predominantly males, ranging from 70% to 86% of the total trial populations. The mean age of patients was consistent across studies, ranging from 60.0 to 67.7 years old. Age was a variable included in propensity score matching of participants or adjustment of analyses in all studies, and hence was well balanced between treatment arms in each study. Sex was accounted for in matching/adjusting in all but one study ( ); however, baseline data indicated no difference in sex distribution across the two treatment arms. The most common etiology of shock in the included studies was AMI-CS. Three studies specified that participants had to have AMI-CS within the eligibility criteria ( – ), one of which stated participants had to have out-of-hospital cardiac arrest due to AMI-CS ( ). The study by Karatolios (2021) did not specify that participants had to have AMI-CS to be eligible for study entry; however, the ‘etiology of shock’ was a covariate accounted for in propensity score matching, and as a result, 86% of participants in both matched treatment arms had AMI-CS ( ). The study by Wernly (2021) included participants with either AMI-CS or cardiac arrest CS; however, the proportion of participants with each etiology differed between treatment arms with 89% of the Impella cohort and 14% of the VA ECMO cohort having AMI-CS ( ). ‘Shock after myocardial infarction’ was a covariate adjusted for in the analyses in the study; hence, it is considered that this imbalance was addressed ( ). The covariates included in propensity score matching in the included studies are summarized in Table . The covariates considered to impact study results varied considerably between studies. Variables used to match participants or adjust analyses in at least two of the included studies comprise sex, race, vasoactive score ( ), pH, PaO 2 /FiO 2 (ratio of partial pressure of oxygen in blood [PaO 2 , mmHg] and the fraction of oxygen in the inhaled air [FiO 2 ]), lactate, comorbidities (Charlson Comorbidity Index ( )), prior cardiopulmonary resuscitation (CPR), and region. Furthermore, the study by Lemor et al. (2020) reported survival and bleeding in a matched population as well as survival and bleeding in an unmatched population with an adjusted analysis ( ). Both the matched and adjusted analyses resulted in the same finding regarding the direction and significance of the treatment effect. This meta-analysis used the dichotomous data from the matched analyses to calculate ORs for survival and bleeding.
Based on the grading of recommendations assessment development and evaluation assessment, the quality of the evidence was assessed as low for both short-term mortality and bleeding events requiring transfusion. Mortality The meta-analysis of the ORs from the included studies demonstrated that Impella statistically significantly reduced the odds of in-hospital/30-day mortality compared to VA ECMO (OR = 0.57 [95% CI: 0.44, 0.74]; P < 0.0001). The I 2 statistic of 27% indicated heterogeneity might not be important. All included studies reported lower absolute in-hospital/30-day mortality for patients treated with Impella versus VA ECMO with the largest two studies ( , ), reporting statistically significant differences in favor of Impella. As depicted in Figure , the ORs of all included studies were consistent in direction and magnitude of effect in favor of Impella, except for Wernly et al. (2021). In absolute terms, the proportion of Impella patients that died in-hospital/within 30 days (pooled: 39.6%) was lower than the proportion of VA ECMO patients (pooled: 53.8%). With a 14.2% difference in favor of Impella versus VA ECMO, corresponding to a number needed to treat (NNT) with Impella to avoid one death of seven. Bleeding events requiring transfusions As can be seen in Figure , the meta-analysis demonstrated statistically significantly lower odds of bleeding events requiring transfusion in patients treated with Impella compared to VA ECMO (OR = 0.61 [95% CI: 0.46, 0.80]; P = 0.0004). No heterogeneity was detected (I 2 = 0%). All studies reported a lower proportion of patients with bleeding events requiring transfusion in the Impella cohort (pooled: 19.9%) compared to the VA ECMO cohort (pooled: 28.8%), with a difference of 8.9% in favor of Impella. Thus, the NNT with Impella relative to VA ECMO to avoid one bleeding event requiring transfusion is 12.
The meta-analysis of the ORs from the included studies demonstrated that Impella statistically significantly reduced the odds of in-hospital/30-day mortality compared to VA ECMO (OR = 0.57 [95% CI: 0.44, 0.74]; P < 0.0001). The I 2 statistic of 27% indicated heterogeneity might not be important. All included studies reported lower absolute in-hospital/30-day mortality for patients treated with Impella versus VA ECMO with the largest two studies ( , ), reporting statistically significant differences in favor of Impella. As depicted in Figure , the ORs of all included studies were consistent in direction and magnitude of effect in favor of Impella, except for Wernly et al. (2021). In absolute terms, the proportion of Impella patients that died in-hospital/within 30 days (pooled: 39.6%) was lower than the proportion of VA ECMO patients (pooled: 53.8%). With a 14.2% difference in favor of Impella versus VA ECMO, corresponding to a number needed to treat (NNT) with Impella to avoid one death of seven.
As can be seen in Figure , the meta-analysis demonstrated statistically significantly lower odds of bleeding events requiring transfusion in patients treated with Impella compared to VA ECMO (OR = 0.61 [95% CI: 0.46, 0.80]; P = 0.0004). No heterogeneity was detected (I 2 = 0%). All studies reported a lower proportion of patients with bleeding events requiring transfusion in the Impella cohort (pooled: 19.9%) compared to the VA ECMO cohort (pooled: 28.8%), with a difference of 8.9% in favor of Impella. Thus, the NNT with Impella relative to VA ECMO to avoid one bleeding event requiring transfusion is 12.
Our meta-analysis of propensity-matched or adjusted retrospective observational studies indicates a significantly lower short-term mortality for patients with CS supported with Impella versus VA ECMO. Furthermore, a significantly lower proportion of Impella versus VA ECMO patients experienced bleeding events requiring transfusion. Previously published meta-analyses comprise both matched/adjusted and unmatched/unadjusted data ( , ) or matched data alone ( ). Low et al. (2024) ( ) performed a meta-analysis of propensity score-matched observational studies and RCTs of mechanical circulatory support devices of the treatment of CS. The study incorporated results from three propensity score-matched direct studies of Impella (referred to in the publication as ‘mVAD’) versus VA-ECMO, all of which are included in our meta-analysis ( , , ). Low et al. did not include adjusted data and therefore did not include Wernly et al. (2021) ( ), and the propensity score-matched study by Vetrovec et al. (2020) ( ) was also not included. Similar to our analysis, Low not identify any RCTs of mVAD versus ECMO. The ECMO versus mVAD direct comparison showed that ECMO was associated with statistically significantly higher mortality (OR = 1.85 [95% CI: 1.17–2.92]) whereas the indirect method identified no difference (OR = 1.09 [95% CI: 0.70–1.69]). This incoherence between direct and indirect methods of analysis may indicate issues regarding exchangeability in the common reference arm(s) used across trials. The results from the direct comparison of the study by Low et al. (2024) are consistent with the results of our meta-analysis, showing a reduction in the risk of 30-day/in-hospital mortality in patients with CS treated with Impella compared to ECMO. Batchelor et al. (2022) conducted a systematic literature search in November 2021 to identify comparative studies of patients with AMI-CS who required MCS with VA ECMO or Impella or both, with analyses based on the first device used only. The review identified six cohort studies (n = 7,093 patients), all of which contributed unmatched/unadjusted data to the primary meta-analysis. The primary meta-analysis of in-hospital mortality including ‘all comers’ found a statistically significant survival benefit with Impella versus VA ECMO (RR = 0.89 [95% CI: 0.83, 0.96]; P = 0.004). The analysis limited to three propensity-matched cohort studies ( – ), similarly found a statistically significant reduction in the risk of in-hospital mortality with Impella compared to VA ECMO (RR = 0.72 [95% CI: 0.59, 0.86]). While the direction of effect was consistent between the primary ‘all comers’ analysis and the analysis limited to propensity score-matched studies, the magnitude of effect was greater in the latter (RR 0.72 vs. 0.89). To note, the magnitude of effect observed in the Batchelor et al. meta-analysis was similar in our analysis (RR [95% CI]: 0.75 [0.64, 0.89] P = 0.0006). Abusnina et al. (2022) conducted a systematic review up to May 2022 which identified 10 observational retrospective studies of Impella versus VA ECMO (n = 1,827) CS patients. Four of the included studies contributed matched/adjusted data ( – , ), all of which are included in our meta-analysis, while the remaining six studies comprised unmatched/unadjusted data. The meta-analysis found lower risk of in-hospital all-cause mortality with Impella compared to VA ECMO approaching statistical significance (RR = 0.80 [95% CI: 0.65, 1.00]; P = 0.05). To note, moderate heterogeneity (I 2 = 58%) was observed in this analysis, likely reflecting the inclusion of unmatched studies. Additionally, the meta-analysis by Abusnina et al. (2022) found a statistically significantly lower risk of bleeding complications with Impella compared to VA ECMO (RR = 0.55 [95% CI: 0.42, 0.71]; P < 0.001). Overall, existing meta-analyses that include a mix of unmatched/unadjusted and matched/adjusted data are congruent with the current study findings, that is, suggesting the treatment of CS patients with Impella may be associated with a lower risk of in-hospital/30-day mortality and fewer bleeding complications than those treated with VA ECMO. While no head-to-head RCTs of Impella versus VA ECMO in CS exist to date, a recently published RCT of Impella versus standard of care with optimal medical therapy (DanGer Shock) represents the first RCT (n = 355) to successfully demonstrate superior survival outcomes for an MCS device compared to standard of care in patients with CS ( ). The study identified a statistically significant survival benefit with Impella at 180 days (HR = 0.74 [95% CI: 0.55, 0.99]; P = 0.04) ( ). While multiple RCTs have been conducted comparing VA ECMO versus standard of care in CS (ECLS-SHOCK I ( ), ECMO-CS ( ), EURO SHOCK ( ) and ECLS-SHOCK ( )), none have found a statistically significant survival benefit for VA ECMO compared to standard of care at 30 days (RR = 0.98 [95% CI: 0.80, 1.19] ( ); HR = 0.56 [95% CI: 0.21, 1.45] ( ); HR = 1.11 [95% CI: 0.66, 1.87] ( ); OR = 0.47 [0.11, 1.94] ( , )). While most of these VA ECMO RCTs were underpowered to detect survival differences, a meta-analysis of these studies performed on an individual patient basis (n = 567 patients; 284 VA ECMO, 283 standard of care) showed that VA ECMO did not reduce 30-day mortality compared with standard of care (OR = 0.93 [95% CI]: 0.66, 1.29) ( ). Important differences between inclusion criteria in the DanGer Shock trial and VA ECMO RCTs, make comparison across different trials challenging. Notably, the DanGer Shock trial demonstrated statistically significantly improved survival with Impella relative to standard of care, while RCTs of VA ECMO showed no difference in survival relative to standard of care. These results together with our meta-analysis may suggest that in select patients in CS requiring MCS, Impella may be considered the preferred initial support modality. Head-to-head RCTs comparing Impella with VA ECMO in patients with CS will be important to inform the best initial MCS strategy in patients with CS. Limitations Our study comprises a study-based analysis as opposed to an individual patient-level analysis which facilitates greater exploration of heterogeneity and effect modification. The covariates accounted for in each individual study may not have incorporated all important variables, meaning residual confounding may exist. For example, lactate and prehospital cardiac arrest have been reported as important predictors of mortality in CS ( – ); however, only two of the five studies included in this meta-analysis matched or adjusted based on lactate ( , ); one study matched on etiology of shock ( ), and one study included only patients with out-of-hospital cardiac arrest ( ). Notably, there is no universally accepted list of variables known to be predictors of mortality in CS, meaning the selection of covariates in the studies were at the discretion of the investigators. Additionally, most of the included studies were considered to have a moderate risk of bias, in the context of being observational studies. In particular, device implantation was at physician’s discretion and, while reflective of clinical practice, may result in selection bias. Syntila et al. (2021) and Lemor et al. (2020) explicitly stated patients treated with both Impella and VA ECMO were excluded from enrolment, Wernly et al. (2021) did not mention the possibility of patients being treated with two devices, and Karatolios et al. (2021) and Vetrovec et al. (2020) noted patients treated with both devices were analyzed according to first device implanted. It is unclear how many patients in Karatolios et al. (2021) and Vetrovec et al. (2020) were treated with both devices in the matched population; however, Karatolios et al. (2021) noted less than 10% of patients in the unmatched population were treated with both devices. While the use of biventricular support may suggest a more severe patient cohort, the extent of biventricular support and differential use between cohorts is unclear, meaning there is a risk of confounding by severity. The current meta-analysis focused on consistently reported outcomes across all studies including short-term survival and bleeding events, meaning the complete risk-benefit profile of Impella versus VA ECMO may not have been adequately captured. Furthermore, this meta-analysis did not limit to specific CS subgroups (e.g., AMI-CS) and the included studies did not consistently assess the severity of shock in the enrolled population (e.g., using the Society for Cardiovascular Angiography and Intervention [SCAI] staging system ( )). Hence this meta-analysis included a heterogeneous population, reflective of the use of MCS for CS in clinical practice. To note, the majority of patients in the included studies had AMI-CS, consistent with being the most common cause of CS. Overall, the quality of the evidence was assessed as low for short-term mortality and bleeding events requiring transfusion, meaning our confidence in the effect estimate is limited.
Our study comprises a study-based analysis as opposed to an individual patient-level analysis which facilitates greater exploration of heterogeneity and effect modification. The covariates accounted for in each individual study may not have incorporated all important variables, meaning residual confounding may exist. For example, lactate and prehospital cardiac arrest have been reported as important predictors of mortality in CS ( – ); however, only two of the five studies included in this meta-analysis matched or adjusted based on lactate ( , ); one study matched on etiology of shock ( ), and one study included only patients with out-of-hospital cardiac arrest ( ). Notably, there is no universally accepted list of variables known to be predictors of mortality in CS, meaning the selection of covariates in the studies were at the discretion of the investigators. Additionally, most of the included studies were considered to have a moderate risk of bias, in the context of being observational studies. In particular, device implantation was at physician’s discretion and, while reflective of clinical practice, may result in selection bias. Syntila et al. (2021) and Lemor et al. (2020) explicitly stated patients treated with both Impella and VA ECMO were excluded from enrolment, Wernly et al. (2021) did not mention the possibility of patients being treated with two devices, and Karatolios et al. (2021) and Vetrovec et al. (2020) noted patients treated with both devices were analyzed according to first device implanted. It is unclear how many patients in Karatolios et al. (2021) and Vetrovec et al. (2020) were treated with both devices in the matched population; however, Karatolios et al. (2021) noted less than 10% of patients in the unmatched population were treated with both devices. While the use of biventricular support may suggest a more severe patient cohort, the extent of biventricular support and differential use between cohorts is unclear, meaning there is a risk of confounding by severity. The current meta-analysis focused on consistently reported outcomes across all studies including short-term survival and bleeding events, meaning the complete risk-benefit profile of Impella versus VA ECMO may not have been adequately captured. Furthermore, this meta-analysis did not limit to specific CS subgroups (e.g., AMI-CS) and the included studies did not consistently assess the severity of shock in the enrolled population (e.g., using the Society for Cardiovascular Angiography and Intervention [SCAI] staging system ( )). Hence this meta-analysis included a heterogeneous population, reflective of the use of MCS for CS in clinical practice. To note, the majority of patients in the included studies had AMI-CS, consistent with being the most common cause of CS. Overall, the quality of the evidence was assessed as low for short-term mortality and bleeding events requiring transfusion, meaning our confidence in the effect estimate is limited.
This systematic review and meta-analysis of five propensity score-matched/adjusted studies of Impella versus VA ECMO in CS suggests that Impella may improve short-term survival and result in fewer bleeding events requiring transfusion in select patients. However, given the potential for residual confounding in the included studies, the results from the meta-analysis should be interpreted with caution. With the lack of RCTs comparing Impella versus VA ECMO in CS, this matched/adjusted data analyzed provides an improvement on currently published meta-analyses of unadjusted studies. A well designed RCT comparing Impella versus VA ECMO in CS is warranted.
|
Awareness and perceptions of Filipino obstetrician-gynecologists on fertility preservation: a cross-sectional survey | d5f20862-8e55-45a2-b882-9969634fd4b5 | 10214684 | Gynaecology[mh] | One of the most critical milestones in reproductive medicine is the advent of fertility preservation. Various fertility preservation techniques allow men and women with compromised fertility a chance to achieve reproductive capacity at a later time. While advances in cancer therapy have led to an increasing number of young patients who survive, a crucial sequela is loss of fertility due to the gonadotoxic profile of current regimens . The field of Oncofertility is a network of different subspecialties focused on techniques to restore reproductive function in patients with malignancies . Aside from cancer patients, fertility preservation has been widely applied to patients with benign conditions such as genetic disorders, autoimmune disorders, and other diseases predisposing to premature gonadal failure. Women who wish to postpone childbearing for social and professional reasons likewise benefit from fertility preservation . Age is a critical factor in the Patient-Oriented Strategies Encompassing Individualized Oocyte Number (POSEIDON) for women with poor ovarian response to stimulation . Age directly affects oocyte quality and embryo ploidy. Studies have shown that the number of euploid blastocysts decline after 34 years . Advanced female age and decreased ovarian reserve were shown to be prevalent in POSEIDON patients. This emphasizes the need for counseling on the importance of age and ovarian reserve on the prospects of future fertility . Women at risk of infertility should be identified and provided information specific to their needs. Information regarding the impact of malignancy and other diseases on reproductive function, the effect of treatment on fertility, fertility preservation options, issues relating to cryopreservation storage, infertility and fertility treatments, pregnancy after gonadotoxic treatment, and other childbearing and parenting options should be presented to patients . The Philippine Society for Fertility Preservation was established in 2019, reflecting the growing need and interest to improve and promote its practice. Awareness of established fertility preservation techniques is essential to ensure appropriate counseling of patients and referral to a specialist. Currently available fertility preservation strategies in females include embryo cryopreservation, mature oocyte cryopreservation, ovarian tissue cryopreservation, ovarian suppression with GnRH analogs, ovarian transposition, and fertility-sparing surgeries . Meanwhile, sperm cryopreservation is the only established fertility preservation method in adolescent and adult males . In the Philippines, fertility preservation techniques are only offered in private centers and paid through out-of-pocket expenses. At the time of writing, no government-funded facility offer these procedures. Expenses are not covered by the Philippine Health Insurance Corporation nor by health maintenance organizations. Despite the rapid progress of fertility preservation in clinical practice, knowledge of its availability is lacking among clinicians . This paucity of knowledge from healthcare providers on the protection of reproductive function certainly affects the patient’s knowledge, attitude, behavior, and perspective. The current study aimed to assess obstetrician-gynecologists awareness and perception of fertility preservation. It was timely and relevant to conduct this study to determine the current status and barriers to improving the practice of fertility preservation in the country. This was the first study among post-residency obstetrician-gynecologists in the Philippines. A cross-sectional survey was conducted among diplomates and fellows of the Philippine Obstetrical and Gynecological Society (POGS) from September to December 2021. A hyperlink to the online survey was sent by electronic mail to the target participants with society’s support and distributed over social media. The minimum required sample was 209 of the 4500 accredited obstetrician-gynecologists in the country. This was computed using the Cochran formula and based on the study by Fritz et al. , which reported that 82.8% of obstetricians believe that fertility discussions should be routinely part of the examinations. The sample size was computed with a 5% margin of error and a design effect of 1.0. Non-probability sampling and consecutive enrollment of participants were done until the sample size was achieved. A self-administered survey patterned from the study of Chung et al. . was utilized. The questionnaire was composed of 24 items divided into two sections. The first section included questions on the demographic profile of the participants. The second section assessed the awareness and perception of fertility preservation. Pilot testing of the questionnaire to 15 subjects was performed before the survey proper. Univariate descriptive statistics were reported as mean for continuous variables and frequency with percentage for categorical variables. Differences in responses were tested using the chi-square test. A P value of < 0.05 was considered statistically significant. All analyses used STATA 14 (Stata Corp Inc). A total of 215 participants accomplished the online questionnaire. The mean age of the respondents was 42.98 ± 10.59 years. The majority were female (94.88%), Catholic (80.94%), married (68.37%), and had children (65.12%). The geographical regions were represented, with the National Capital Region (NCR) being the most represented. The sociodemographic data of the respondents are summarized in Table . Most of the participants belonged to private non-university affiliated hospitals (27.91%) and had practiced for one to five years (44.65%). General obstetrician-gynecologists constituted 57.21% of the study population. Of the 42.79% specialists, the most frequently identified specialties were Ultrasound (12.56%) and Reproductive Endocrinology (10.70%). Table provides an occupational summary of the population. Majority of the respondents agreed that obstetrician-gynecologists should initiate discussions with patients about their childbearing intentions (98.60%) and age-related fertility decline (97.67%). Obstetrician-gynecologists largely believed that discussion of natural fertility decline should be part of a well-woman annual examination, with agreement by 96.28%. Almost all participants (98.60%) were aware of fertility preservation and were familiar with at least one method or procedure. Only 32.56% were familiar with all techniques, including fertility-sparing surgeries, the use of GnRH agonists, sperm freezing, oocyte freezing, embryo freezing, and ovarian or testicular tissue freezing. Most respondents (81.40%) were aware of fertility-sparing surgeries. Approximately half (45.12%) of the participants have not referred patients for fertility preservation in the twelve months before the study proper. Only seven respondents were able to refer patients for all the mentioned procedures (Table ). Respondents were largely aware (86.98%) of a particular clinic or specialist who can accept referrals for fertility preservation. Of note, 28 respondents (13.02%) were unaware of any facility or specialist. In 43.72%, the patient’s desire to have children was identified as the most critical factor when deciding on fertility preservation in medical indications, followed by age (35.81%) and prognosis (10.23%). Majority of the participants (93.49%) deemed it necessary to set up dedicated centers for fertility preservation. About 91.63% think it should be offered as a public health service. Standard educational materials were deemed essential in enhancing patient understanding of fertility preservation. More than half (59.07%) are unaware of regulations relating to fertility preservation, but 98.14% support establishing guidelines. Three respondents did not wish to know more about fertility preservation (Table ). The likelihood of discussing fertility-related practices was not different across characteristics of fellows. However, the analysis is limited by the inadequate number of participants per characteristic category. Cells with a frequency of less than five were merged with other cells to ensure adequacy for analysis. The likelihood of having an awareness of fertility-related practices was not different across characteristics of fellows except for a few geographic locations and subspecialties. Those in Luzon are 2.26 times more likely to be aware of regulations on fertility preservation than those in the NCR. The Philippines is composed of three major islands known as Luzon, Visayas, and Mindanao. For the analysis, the National Capital Region was separated from Luzon because it houses most of the centers able to provide fertility preservation techniques and has the most number of specialists in the country. Provinces included in Luzon were Regions I, II, III, IV-A, MIMAROPA, V, and CAR. Luzon is generally considered to be more urbanized than provinces in Visayas and Mindanao. Distribution of health infrastructures and human resources is skewed toward Luzon and the National Capital Region. Respondents with subspecialties other than Reproductive Endocrinology have a 51% reduced odds of having an awareness of these regulations than general obstetrician-gynecologists. Reproductive endocrinologists have 80% reduced odds of agreeing on setting up fertility preservation counseling compared to general obstetricians. On the other hand, Christians have 20% reduced odds of agreeing on the need for practice guidelines than Roman Catholics. Fertility preservation has continued to gain worldwide attention over the years. A local study conducted by Factor and Novero was the first attempt to examine Filipino practitioners’ knowledge, attitudes, and practices on fertility preservation . The study included 213 surgical oncologists, medical oncologists, and radiation oncologists. Majority of their study participants acknowledged knowing only minimal information. Only 38% have referred patients to fertility specialists, citing lack of knowledge, poor success rates of fertility preservation, poor patient prognosis, and high costs . The current study is the first to describe the awareness and perceptions of Filipino obstetrician-gynecologists about reproductive aging and fertility preservation. The majority of the study respondents were female because they comprise 95% of the diplomates and fellows of the Philippine Obstetrical and Gynecological Society. Being the primary provider of reproductive healthcare, it is reassuring that majority of the respondents agreed that discussions about potential childbearing intentions and age-related fertility decline should be initiated during an annual examination. The International Fertility Decision-Making Study highlighted the lack of knowledge about fertility in 10,045 reproductive-aged men and women in over 79 countries . Counseling increases patient understanding, allows informed decisions about her future reproductive plans, and encourages better patient participation. There is a high awareness of fertility preservation among the respondents. One of the main objectives of the Philippine Society for Fertility Preservation (PSFP) is to promote the science and practice of fertility preservation. The society conducts regular conferences, meetings, and discussions on scientific information and treatment advances. There were varying levels of awareness of the different techniques. Most were familiar with at least one fertility preservation technique. Meanwhile, only a third of the study population knew all methods. Unawareness may lead to the underutilization of available methods of fertility preservation. This emphasizes the need to educate more obstetrician-gynecologists through fertility preservation awareness campaigns and continuing medical education activities, including seminars and workshops. Not surprisingly, the highest level of awareness was associated with fertility-sparing surgeries. Fertility-sparing surgery entails preserving at least a portion of an ovary and the uterus. These are limited to early-stage malignancies and include conization or trachelectomy for cervical cancer and unilateral salpingo-oophorectomy for ovarian cancer. Clinicians should provide appropriate information about oncologic and pregnancy outcomes through an individualized patient approach . Obstetrician-gynecologists were likely to be most aware of fertility-sparing surgeries as they perform the surgeries themselves, and specialists provide further treatment. Despite the high level of awareness, half of the respondents had not referred patients for fertility preservation, and majority desired to know more information. The study’s findings were similar to the reports of Harzif et al. among obstetrician-gynecologists in Indonesia . Identified hindrances were financial constraints, poor success rates of fertility preservation techniques, poor prognosis of patients, and lack of physician knowledge. These underscore that information among obstetrician-gynecologists is lacking. Aside from these, the European Society of Human Reproduction and Embryology (ESHRE) listed limited public awareness of fertility and fertility preservation, limited awareness of oncologists on fertility preservation options, lack of referral pathways, and unavailability of every technique as barriers to access to fertility preservation . Further local studies on the knowledge, attitudes, and practices of Filipino obstetrician-gynecologist may be undertaken to examine the perceived barriers to the provision of much-needed fertility preservation techniques. Most respondents saw setting up dedicated centers for fertility preservation as necessary. The study shows that reproductive endocrinologists have 80% lower odds of agreeing on this than general obstetrician-gynecologists. A small proportion of the study population was unaware of any facility or specialist. In the Philippines, fertility preservation techniques are mainly performed in reproductive centers offering in vitro fertilization. There are only eight centers and 147 infertility specialists able to provide these services in the country. Access to these centers is available to reproductive endocrinologists, which may explain the decreased support for establishing dedicated facilities. Encompassing help from all specialists should be elicited to promote fertility preservation. Early referral of women with malignancy at the time of diagnosis and before treatment commencement is the key to maximizing the success of fertility preservation and allows a greater window of opportunity for preserving fertility . As primary doctors of women with gynecologic malignancies, gynecologic oncologists should refer them for reproductive counseling as soon as the diagnosis is made. The ESHRE advocates a model of care for patients eligible for fertility preservation. Central to this model is the awareness of fertility preservation options and the training of healthcare providers. The clinical care team should provide essential information and referrals for fertility preservation consultation. Fertility preservation counseling is provided by specialists after a thorough patient assessment . There is a need for quick and efficient referral systems. The high cost of most fertility preservation techniques and patient financial constraints have impeded widespread local use. Most respondents agreed that these techniques should be offered as a public health service to mitigate access issues. A multilevel approach is essential to address issues specific to patients and their families, clinicians, organizations, policymakers, and the general population . Fertility preservation is a significant issue in women diagnosed with malignancy. A survey of young women undergoing therapy showed that childbearing remains a priority . Diminished reproductive capacity and fertility loss are leading causes of anxiety and depression among this population. Studies suggest that the risk of infertility has a significant impact on the decision-making process of young cancer patients . In a prospective cohort study among 425 women with newly diagnosed breast cancer, 1% decided not to receive chemotherapy, 2% chose one chemotherapy regimen over another, 1% considered not receiving endocrine therapy, 3% chose not to receive endocrine therapy, and 11% considered receiving endocrine therapy for five years due to concerns in fertility . Similarly, the study respondents deemed a patient’s desire to have children the most important factor when deciding on fertility preservation in medical conditions. It is, therefore, worthwhile to investigate patient perceptions and access to the different techniques in the local setting. Patient age and prognosis were among the top considerations. These again stress the need for timely counseling. Comprehensive recommendations and clinical guidelines on fertility preservation should be established and communicated. ESHRE published its first evidence-based guideline on female fertility preservation for healthcare professionals in 2020 . Socio-economic factors relating to the respondents’ place of practice and affiliation could influence their knowledge, attitudes, and practices on fertility preservation. Private practitioners manage a different subset of patients compared to those in public facilities. Their patients are better able to afford fertility preservation techniques. As such, they are more exposed and knowledgeable on fertility preservation. Considering the current laws, patient population, and socioeconomic factors, these guidelines need to be optimized in the local setting. Interestingly, practitioners in Luzon were 2.26 times more likely to be aware of regulations on fertility preservation than those in the NCR. Subgroup analysis of participants in Luzon showed that the majority have been practicing for one to five years, while most of those in NCR has been practicing for more than 16 years. This may be due to more active personal inquiry by younger clinicians or better participation in regional campaigns. The availability of fertility preservation techniques in the NCR should be an impetus for practitioners in this area to improve awareness. Subspecialties other than reproductive endocrinology had 51% reduced odds of awareness of existing guidelines compared to general obstetrician-gynecologists. Their specialized practices may deter them from acquiring further information in this growing field. As primary reproductive healthcare providers, all obstetrician-gynecologists should be knowledgeable about recommendations and guidelines. Overall, Filipino obstetrician-gynecologists have an encouraging positive perception of fertility preservation. There is a need for further education on the locally available techniques. The information presented by this study can be applied in the framework of establishing local guidelines and designing a curriculum for training. A multidisciplinary team with reproductive specialists, an insurance coverage system, comprehensive laws, and practice guidelines should be prioritized. Reproductive aging and fertility preservation are emerging fields in managing reproductive-aged women. This is the first local study that evaluated the awareness and perceptions of post-residency obstetrician-gynecologists on fertility preservation. The study showed a reassuring positive perception of fertility preservation but a gap in the awareness of different approved methods. A multidisciplinary approach and dedicated facilities should be established for fertility consultation, risk assessment, and counseling. Healthcare delivery should be organized to meet the increasing need for fertility preservation. The study employed non-probability sampling and consecutive enrolment of participants until the sample size was met. Selection and response bias may have influenced the results of the study. Participation in a self-directed online questionnaire entails the awareness of the sample to the existence of the survey. The number of physicians who actually received the survey is uncertain. This was minimized by distribution of the questionnaire by the Philippine Obstetrical and Gynecological Society to its registered members. Regular posting of the survey to various social media platforms was also conducted to improve visibility and response. To minimize response bias, the period of data collection was extended after the minimum sample size was met. The number of Gynecologic Oncology specialists who completed the survey was only eight due to the sampling method employed. The study did not assess the specific reasons for non-referral for fertility preservation techniques. Further studies with the recruitment of gynecologic oncologists may be undertaken. Another vital area of research is the investigation of perceived barriers to the provision of timely and appropriate fertility preservation techniques. The knowledge and perceptions of patients on fertility preservation should also be investigated. |
Toward Identification of Markers for Brain‐Derived Extracellular Vesicles in Cerebrospinal Fluid: A Large‐Scale, Unbiased Analysis Using Proximity Extension Assays | 8758a54f-c0c6-41c1-a5c6-abe87f9c004d | 11913887 | Biochemistry[mh] | Introduction Extracellular vesicles (EVs) are nanometre‐scale, membrane‐bound compartments that contain proteins, RNAs and metabolites endogenous to their cell of origin (Raposo and Stoorvogel ). As such, the content of EVs, isolated from biofluids, can serve as a molecular snapshot of the parent cell. A preponderance of EV research has focused on isolating EVs from specific cell types or tumours utilizing proteins annotated as transmembrane and enriched in the parent cell (Shami‐shah et al. ). While this has led to some success, such as in the case of monitoring prostate cancer (Ramirez‐Garrastacho et al. ), studies that have sought to capture brain‐derived EVs have been hampered by methodological challenges. Specifically, proteins cited as transmembrane or internal to EVs have been shown to be predominantly cleaved and secreted (Norman et al. ). It is, therefore, critical to validate methods to differentiate EV‐associated proteins from those that are secreted and cleaved in biofluids. Plasma and cerebrospinal fluid (CSF) EVs can be easily separated from soluble proteins using size exclusion chromatography (SEC) or density gradient chromatography (DGC) (Norman et al. ). Nevertheless, analysing the proteomic content of the EV and soluble protein fractions with a single biochemical technique can be difficult because the soluble protein fractions contain several orders of magnitude more protein than the EV fractions. Unbiased techniques like mass spectrometry are challenging because, in the EV fractions, lipoproteins can co‐isolate and mask rare EV‐associated proteins, while in the secreted protein fractions, abundant proteins like albumin create a similar problem (Ter‐Ovanesyan et al. ). Furthermore, the high levels of abundant proteins, such as albumin in plasma, preclude the ability to use gel‐based techniques such as Western blots. As a result, ELISAs have thus far been the best method of assessing EV fractionation patterns (Ter‐Ovanesyan et al. ). In previous work, we have utilized the ultrasensitive digital ELISA platform Simoa, invented by our lab, to quantify canonical EV proteins (CD9, CD63, CD81, Alix), assess potential contaminants to EV preparations (apolipoprotein B, albumin), and evaluate individual proteins as targets for cell‐type‐specific enrichment (Norman et al. ; Ter‐Ovanesyan et al. , , , ). Here, we sought to apply a large‐scale unbiased method to generate a much‐needed dataset and establish a bioinformatic approach to identify proteins that can be used for potential immunocapture of EVs secreted by a cell‐type of interest, as well as cytosolic proteins to corroborate EV‐brain‐cell origin. CSF directly surrounds the brain and the spinal cord, making CSF‐derived EVs more likely to contain predominantly brain‐specific markers compared to plasma and other biofluids (Hladky and Barrand ; Shetgaonkar et al. ). Furthermore, CSF has approximately 200‐fold lower soluble protein content compared to plasma (Fogh et al. ). This lowers the chance of nonspecific interactions compared to that for EVs isolated from plasma and other more complex matrices. Although estimates of the proportion of brain‐derived EVs in CSF are highly limited by the lack of reliable markers, one study reported that approximately 16% of brain‐specific proteins in CSF EVs were of neuronal origin while about 84% of them were of glial origin (Muraoka, Jedrychowski, et al. ). This makes CSF an ideal biofluid for brain‐derived EV biomarker discovery analysis. By using human CSF, we seek to make strides towards a liquid biopsy of the nervous system, eventually enabling the development of minimally invasive diagnostics for neurological and psychiatric diseases.
Methods 2.1 Human Sample Preparation For the main experimental figures utilizing Olink and Simoa technology, one millilitre each of four healthy CSF samples (PrecisionMed) were thawed at room temperature and centrifuged at 2000 × g for 10 min. Subsequently, the supernatant from this first centrifugation was transferred to a 0.45‐µm Corning Costar Spin‐X filter (Sigma‐Aldrich) and centrifuged again at 2000 × g for 10 min at room temperature. The flow‐through from this filtration was used for downstream experiments. For Figures (nanoparticle tracking analysis and Western blotting), one pooled lot of CSF (Innovative Research) was used to ensure enough material was available for all Western blots and nanoparticle tracking analysis without adding inter‐individual variability. The CSF was processed in the same way for this pooled lot as for the individual samples used in the main figures. 2.2 SEC and Fraction Processing Sepharose CL‐6B resin (GE Healthcare) was washed with an equal volume of PBS 3 times. For each wash, the resin was allowed to settle at 4°C overnight before the PBS was poured off and replaced. Following the washes, the resin was stored in an equal volume of PBS. Econo‐Pac Chromatography columns (Bio‐Rad) were prepared immediately prior to fractionation. For each sample, washed resin was poured into a column to achieve a resin bed volume of 10.2 mL. A polyethylene bed support (Bio‐Rad) was inserted into the top of the resin to compress to a bed volume of 10 mL. The packed resin was then washed with 20 mL of PBS. Immediately following the elution of the wash, 1 mL of each CSF sample was added to the respective column, and fractions were collected in 0.5 mL increments. When the 1 mL of CSF had flowed through, 0.5 mL of PBS was added to the column sequentially until fractions 1–15 were collected. Fractions 1–5 were discarded to avoid redundancy as EVs generally begin to elute in fractions 7 or 8 when using a 10 mL Sepharose 6B column. Each fraction (6–15) was transferred to 10 kDa MWCO Amicon Ultra Centrifugal Filters (Sigma‐Aldrich) and diluted to a total volume of 1.5 mL with PBS. These fractions were then centrifuged at 2000 × g at 4°C until all fractions were concentrated 15‐fold. The concentrated fractions were brought to a volume of 97 µL with PBS. A 76 µL aliquot was transferred to a 96‐well plate supplied by Olink. Triton X‐100 was added to a final concentration of 1% by volume, and the plate was stored at −80°C. The remaining 21 µL of fraction volume was used to measure CD81 by Simoa. 2.3 Simoa CD81 Sample Analysis The Simoa analysis was performed according to the manufacturer's instruction. Reagent preparation and assay parameters were followed as described previously by Norman et al. ( ). Abcam (anti‐CD81 ab79559, clone M38) and Biolegend (anti‐CD81 349502, clone 5A6) were used as capture and detector antibodies, respectively. Human recombinant CD81 from Origene (TP317508) was used in the calibration curve. Data analysis was performed using GraphPad Prism version 10.1.1. 2.4 Nanoparticle Tracking Analysis Separate CSF fractions 7–10 and 11–15 were collected using SEC, as described above. This 2 mL volume was condensed using a 10 kDa MWCO Amicon Ultra Centrifugal Filter (Sigma‐Aldrich) to a volume of 500 µL in PBS. Extracellular vesicle particle size and number were characterized using the NanoSight LM10 (Malvern Panalytical). A 500 µL of sample was injected, and five 1‐min videos were captured at 24.98 fps with a detection threshold of 2, at a fixed temperature of 25°C. Parameters were determined based on the manufacturer's software manual and performed by the NTA 3.4 Build software v3.4.4. 2.5 Western Sample Analysis SEC was performed as above with 8 mL of pooled CSF. Each mL was loaded on its own column. Respective fractions were pooled and concentrated using 10 kDa MWCO Amicon Ultra Centrifugal Filters (Sigma‐Aldrich). For fractions 6–12, one‐sixteenth of the concentrated fractions were loaded per gel. For fractions 13–15, protein input was normalized to fraction 12 to avoid overloading the gel. The fractions and human brain cerebellum whole tissue lysate (HBL) (Novus Biologicals) were denatured with 4× LDS and, for certain targets, reduced with DTT (see table below). Subsequently, CSF and HBL samples were heated at 70°C for 10 min, run at 150 V for 70 min on 4%–12% Bolt Bis‐Tris Plus gels (Thermo Fisher Scientific), and transferred to nitrocellulose membranes using the iBlot 3 Dry Blotting System (Thermo Fisher Scientific). The membranes were blocked for 30 min at 4°C and incubated with primary antibodies overnight. The next day, membranes were washed, incubated with secondary antibody (Bethyl Laboratories) for 1 h at 4°C, and washed again. Nonspecific signals were assessed by probing CSF SEC fractions and HBL with the corresponding secondary antibodies (anti‐mouse IgG, anti‐rabbit IgG, or anti‐rat IgG) without the application of primary antibody. For primary and secondary antibody dilutions, as well as membrane blocking, a PBS‐T solution of 5% milk (weight by volume) with 1% Tween was used. All washes were performed with PBS‐T (1% Tween) in cycles of three 7‐min washes (except SLC16A1, which was incubated in PBS‐T six times per wash). Specifics on primary antibodies used and dilutions can be found in the table below. After the final wash, blots were developed using the ProSignal Femto substrate kit (Genesee Scientific) and imaged with a Sapphire Biomolecular Imager (Azure Biosystems) Table . 2.6 Olink Sample Analysis Samples were shipped on dry ice to the Broad Institute in Cambridge, MA, for analysis by the Olink HT platform, which measures 5416 unique proteins using highly multiplexed proximity extension assays. Pairs of antibodies with unique, complementary oligonucleotides, called proximity probes, each specific to a unique protein of interest, bind to their target antigens. After binding the target, the oligonucleotide probes encounter each other due to physical proximity and hybridize, resulting in the formation of an immuno‐complex. The resulting hybridized proximity probes can be amplified by DNA polymerase, creating a DNA amplicon that can be detected by quantitative PCR (qPCR) or next‐generation sequencing (NGS) techniques (Shami‐shah et al. ; Wik et al. ). Samples were run with a single replicate for each protein except for GBP1 and MAP2K1, which were run in Blocks 3, 4 and 5 to check correlation between blocks. The relative abundances of the amplicon, as measured by NGS, are then converted to normalized protein expression (NPX) values. The Olink panel includes plate, sample, and extension (ExtCtrl) controls. To ensure robustness, the NPX calculation accounts for variability in the different controls measured in the panel and includes a log2 transformation of the data. The number of matched sequence reads (counts) generated by NGS is first normalized by the number of counts for the extension control of the sample, and then log2 transformed as follows: E x t N P X i , j = log 2 C o u n t s S a m p l e j A s s a y i C o u n t s E x t C t r l j where ExtNPX i , j is the NPX, normalized by the counts of the extension control specific to assay i measured in sample j . The median value of ExtNPX of the plate controls is then used to adjust for variability between plates, allowing comparison of relative protein abundances across different plates (Wik et al. ): N P X i , j = E x t N P X i , j − m e d i a n ( E x t N P X p l a t e h c o n t r o l ) control is the quality control measure collected from plate h , and NPX i , j is the reported NPX value for sample j , analysed using assay i on plate h . Further details on NPX value generation can also be found on the Olink website. 2.7 Data Analysis Methods The reported NPX values have an arbitrary unit and reflect the relative concentrations of the analysed proteins in the sample of interest. All analysis was conducted in Python (version 3.11.5) using Visual Studio Code (Microsoft Corporation, Redmond, Washington). The HT panel includes 5420 proteins, including 5416 unique proteins and two assays processed in triplicate, measuring relative concentrations of MAP2K1 and GBP1, to ensure accuracy of data collection. Given that no calibration curve is included, all NPX values presented in fraction data are relative values only. The data points were all linearized, and two assays that were processed in triplicate were removed from downstream high‐throughput analysis. The HT panel from Olink measures two negative controls for each assay. Olink recommends against calculating a limit of detection (LOD) with fewer than 10 negative controls in a dataset, so we instead considered the fixed LODs made available by Olink. The fixed LOD calculation is based on 24–36 negative controls, ensuring a more robust calculation to minimize the higher variation among negative controls. This approach is consistent with the recommendations from Olink, which reports that values below LOD are unlikely to increase the risk of false positive discoveries and may be beneficial for biomarker discovery. They also highlight that filtering data based solely on LOD may remove meaningful signals, especially when a protein is well expressed in one group but undetectable in another. Therefore, excluding data points below LOD would prevent us from including potentially useful proteins in our analyses. The LOD data are available in Table but were not considered in downstream analyses. 2.8 Fractionation Analysis Four individual fractionated CSF samples (fractions 6–15) were submitted to Olink for analysis. Only fractions 7, 9, 10, 11, 12 and 13 were used to verify whether a protein exhibited the fractionation pattern typical for EV‐associated proteins (Norman et al. ). For each fraction of interest, the median NPX was calculated for each protein. A protein was considered to have a fractionation pattern typical of EV‐associated proteins if the medians of fractions 9 and 10 were greater than the medians for fractions 7, 11, 12 and 13. 2.9 Protein Localization The proteins in the Olink panel were computationally determined to be transmembrane, internal, or external using DeepTMHMM. DeepTMHMM is a deep learning model‐based algorithm that uses a hidden Markov model to predict subcellular localization of a protein in a cell. The model calculated a probability for each amino acid in each protein and returned the highest probability domain for each amino acid, allowing the most likely localization of the overall protein to be determined. Using this model, each amino acid was characterized as: Cytosolic Alpha transmembrane helix Beta transmembrane barrel Signalling peptide External to the cell and any secreted vesicles or exosomes Information regarding signalling peptides was not considered, as they are largely cleaved from the protein when it enters the endoplasmic reticulum, and, therefore, are unlikely to be present in the epitope of the protein found in EVs (Liaci and Forster ). We classified a protein as internal to the cell if all its amino acids were characterized as cytosolic, and we classified a protein as external if all its amino acids were characterized as being outside the cell or on secreted vesicles Figure . Because proteins containing a transmembrane domain also contain domains found internal and external to the cell, a protein was classified as transmembrane if it contained one or more amino acids characterized as an alpha transmembrane helix or a beta transmembrane barrel. Because of the budding mechanisms by which EVs are secreted (Teng and Fussenegger ), it is largely assumed that proteins would have the same cytosolic, transmembrane, or extracellular domains in both EVs and the cell. However, additional validation techniques would be necessary to confirm the localization of proteins relative to EVs, a problem we address through SEC fractionation analysis as described previously. 2.10 EV‐Associated Protein Identification This pipeline was used to identify proteins that may be associated with EVs. Proteins were labelled as internal to EVs if they met the fractionation criteria and were identified as internal using DeepTMHMM as described previously. The same criteria were followed to identify transmembrane and external proteins associated with EVs. This yielded a list that was further narrowed by selecting proteins considered to be cell type‐specific based on the Tau score and BrainRNA‐Seq dataset as described below. Each protein was assigned an “EV Association Score,” which was calculated as the ratio of the median NPX for the EV fractions (fractions 9 and 10), and the median NPX for fractions 7, 11, 12 and 13. This value is shown on the y ‐axis of Figure . 2.11 Cell‐Type‐Specificity The BrainRNA‐Seq atlas reports fragments per kilobase per million mapped fragments (FPKM), collected via RNA sequencing (Zhang et al. ). The mean FPKM of each gene was used for mature astrocytes, neurones, oligodendrocytes, endothelial cells, and microglia. Foetal astrocytes were excluded from analysis. Tau specificity score is used to determine cell‐type‐specificity of genes, as it gives a numerical indication of the relative specificity of a gene across different cell types or tissues. Scores range between 0 and 1, where 0 indicates that a gene is ubiquitously expressed in all cell types, and 1 indicates that a gene is entirely expressed in a single cell type (Kryuchkova‐Mostacci and Robinson‐Rechavi ). A gene was considered specific to a given cell type if it had a Tau specificity score of greater than 0.75 and if the mean FKPM was highest in the cell type of interest relative to the other cell types. Tau specificity scores were calculated using the following formula (Kryuchkova‐Mostacci and Robinson‐Rechavi ): τ = ∑ i = 1 n 1 − x i ^ n − 1 x i ^ = x i max 1 ≤ i ≤ n x i x i = e x p r e s s i o n o f t h e g e n e o f i n t e r e s t i n t i s s u e i n = n u m b e r o f t i s s u e s The opposite is also true—by identifying proteins with a low Tau specificity score, < 0.25, we selected genes that are ubiquitously expressed in all cell types. The genes were then mapped to proteins using data obtained from the UniProt website. This data is included in Table . 2.12 Brain Organ Specificity The GTEx database provides median gene‐level expression transcripts per million (TPM) by tissue (Kowal et al. ). The tissues were grouped as described in Table below: The median TPM of each organ group was used to calculate the tissue specificity of each gene. The tau specificity score was used to determine the organ specificity of each gene, as it gives a numerical indication of the relative specificity of a gene across different organ groups. Scores range between 0 and 1, where 0 indicates that a gene is ubiquitously expressed in all tissue, and 1 indicates that a gene is entirely specific to a single tissue type (Kryuchkova‐Mostacci and Robinson‐Rechavi ). This data is included in Table for reference but is not considered in quantifying cell‐type‐specificity as shown in Figure .
Human Sample Preparation For the main experimental figures utilizing Olink and Simoa technology, one millilitre each of four healthy CSF samples (PrecisionMed) were thawed at room temperature and centrifuged at 2000 × g for 10 min. Subsequently, the supernatant from this first centrifugation was transferred to a 0.45‐µm Corning Costar Spin‐X filter (Sigma‐Aldrich) and centrifuged again at 2000 × g for 10 min at room temperature. The flow‐through from this filtration was used for downstream experiments. For Figures (nanoparticle tracking analysis and Western blotting), one pooled lot of CSF (Innovative Research) was used to ensure enough material was available for all Western blots and nanoparticle tracking analysis without adding inter‐individual variability. The CSF was processed in the same way for this pooled lot as for the individual samples used in the main figures.
SEC and Fraction Processing Sepharose CL‐6B resin (GE Healthcare) was washed with an equal volume of PBS 3 times. For each wash, the resin was allowed to settle at 4°C overnight before the PBS was poured off and replaced. Following the washes, the resin was stored in an equal volume of PBS. Econo‐Pac Chromatography columns (Bio‐Rad) were prepared immediately prior to fractionation. For each sample, washed resin was poured into a column to achieve a resin bed volume of 10.2 mL. A polyethylene bed support (Bio‐Rad) was inserted into the top of the resin to compress to a bed volume of 10 mL. The packed resin was then washed with 20 mL of PBS. Immediately following the elution of the wash, 1 mL of each CSF sample was added to the respective column, and fractions were collected in 0.5 mL increments. When the 1 mL of CSF had flowed through, 0.5 mL of PBS was added to the column sequentially until fractions 1–15 were collected. Fractions 1–5 were discarded to avoid redundancy as EVs generally begin to elute in fractions 7 or 8 when using a 10 mL Sepharose 6B column. Each fraction (6–15) was transferred to 10 kDa MWCO Amicon Ultra Centrifugal Filters (Sigma‐Aldrich) and diluted to a total volume of 1.5 mL with PBS. These fractions were then centrifuged at 2000 × g at 4°C until all fractions were concentrated 15‐fold. The concentrated fractions were brought to a volume of 97 µL with PBS. A 76 µL aliquot was transferred to a 96‐well plate supplied by Olink. Triton X‐100 was added to a final concentration of 1% by volume, and the plate was stored at −80°C. The remaining 21 µL of fraction volume was used to measure CD81 by Simoa.
Simoa CD81 Sample Analysis The Simoa analysis was performed according to the manufacturer's instruction. Reagent preparation and assay parameters were followed as described previously by Norman et al. ( ). Abcam (anti‐CD81 ab79559, clone M38) and Biolegend (anti‐CD81 349502, clone 5A6) were used as capture and detector antibodies, respectively. Human recombinant CD81 from Origene (TP317508) was used in the calibration curve. Data analysis was performed using GraphPad Prism version 10.1.1.
Nanoparticle Tracking Analysis Separate CSF fractions 7–10 and 11–15 were collected using SEC, as described above. This 2 mL volume was condensed using a 10 kDa MWCO Amicon Ultra Centrifugal Filter (Sigma‐Aldrich) to a volume of 500 µL in PBS. Extracellular vesicle particle size and number were characterized using the NanoSight LM10 (Malvern Panalytical). A 500 µL of sample was injected, and five 1‐min videos were captured at 24.98 fps with a detection threshold of 2, at a fixed temperature of 25°C. Parameters were determined based on the manufacturer's software manual and performed by the NTA 3.4 Build software v3.4.4.
Western Sample Analysis SEC was performed as above with 8 mL of pooled CSF. Each mL was loaded on its own column. Respective fractions were pooled and concentrated using 10 kDa MWCO Amicon Ultra Centrifugal Filters (Sigma‐Aldrich). For fractions 6–12, one‐sixteenth of the concentrated fractions were loaded per gel. For fractions 13–15, protein input was normalized to fraction 12 to avoid overloading the gel. The fractions and human brain cerebellum whole tissue lysate (HBL) (Novus Biologicals) were denatured with 4× LDS and, for certain targets, reduced with DTT (see table below). Subsequently, CSF and HBL samples were heated at 70°C for 10 min, run at 150 V for 70 min on 4%–12% Bolt Bis‐Tris Plus gels (Thermo Fisher Scientific), and transferred to nitrocellulose membranes using the iBlot 3 Dry Blotting System (Thermo Fisher Scientific). The membranes were blocked for 30 min at 4°C and incubated with primary antibodies overnight. The next day, membranes were washed, incubated with secondary antibody (Bethyl Laboratories) for 1 h at 4°C, and washed again. Nonspecific signals were assessed by probing CSF SEC fractions and HBL with the corresponding secondary antibodies (anti‐mouse IgG, anti‐rabbit IgG, or anti‐rat IgG) without the application of primary antibody. For primary and secondary antibody dilutions, as well as membrane blocking, a PBS‐T solution of 5% milk (weight by volume) with 1% Tween was used. All washes were performed with PBS‐T (1% Tween) in cycles of three 7‐min washes (except SLC16A1, which was incubated in PBS‐T six times per wash). Specifics on primary antibodies used and dilutions can be found in the table below. After the final wash, blots were developed using the ProSignal Femto substrate kit (Genesee Scientific) and imaged with a Sapphire Biomolecular Imager (Azure Biosystems) Table .
Olink Sample Analysis Samples were shipped on dry ice to the Broad Institute in Cambridge, MA, for analysis by the Olink HT platform, which measures 5416 unique proteins using highly multiplexed proximity extension assays. Pairs of antibodies with unique, complementary oligonucleotides, called proximity probes, each specific to a unique protein of interest, bind to their target antigens. After binding the target, the oligonucleotide probes encounter each other due to physical proximity and hybridize, resulting in the formation of an immuno‐complex. The resulting hybridized proximity probes can be amplified by DNA polymerase, creating a DNA amplicon that can be detected by quantitative PCR (qPCR) or next‐generation sequencing (NGS) techniques (Shami‐shah et al. ; Wik et al. ). Samples were run with a single replicate for each protein except for GBP1 and MAP2K1, which were run in Blocks 3, 4 and 5 to check correlation between blocks. The relative abundances of the amplicon, as measured by NGS, are then converted to normalized protein expression (NPX) values. The Olink panel includes plate, sample, and extension (ExtCtrl) controls. To ensure robustness, the NPX calculation accounts for variability in the different controls measured in the panel and includes a log2 transformation of the data. The number of matched sequence reads (counts) generated by NGS is first normalized by the number of counts for the extension control of the sample, and then log2 transformed as follows: E x t N P X i , j = log 2 C o u n t s S a m p l e j A s s a y i C o u n t s E x t C t r l j where ExtNPX i , j is the NPX, normalized by the counts of the extension control specific to assay i measured in sample j . The median value of ExtNPX of the plate controls is then used to adjust for variability between plates, allowing comparison of relative protein abundances across different plates (Wik et al. ): N P X i , j = E x t N P X i , j − m e d i a n ( E x t N P X p l a t e h c o n t r o l ) control is the quality control measure collected from plate h , and NPX i , j is the reported NPX value for sample j , analysed using assay i on plate h . Further details on NPX value generation can also be found on the Olink website.
Data Analysis Methods The reported NPX values have an arbitrary unit and reflect the relative concentrations of the analysed proteins in the sample of interest. All analysis was conducted in Python (version 3.11.5) using Visual Studio Code (Microsoft Corporation, Redmond, Washington). The HT panel includes 5420 proteins, including 5416 unique proteins and two assays processed in triplicate, measuring relative concentrations of MAP2K1 and GBP1, to ensure accuracy of data collection. Given that no calibration curve is included, all NPX values presented in fraction data are relative values only. The data points were all linearized, and two assays that were processed in triplicate were removed from downstream high‐throughput analysis. The HT panel from Olink measures two negative controls for each assay. Olink recommends against calculating a limit of detection (LOD) with fewer than 10 negative controls in a dataset, so we instead considered the fixed LODs made available by Olink. The fixed LOD calculation is based on 24–36 negative controls, ensuring a more robust calculation to minimize the higher variation among negative controls. This approach is consistent with the recommendations from Olink, which reports that values below LOD are unlikely to increase the risk of false positive discoveries and may be beneficial for biomarker discovery. They also highlight that filtering data based solely on LOD may remove meaningful signals, especially when a protein is well expressed in one group but undetectable in another. Therefore, excluding data points below LOD would prevent us from including potentially useful proteins in our analyses. The LOD data are available in Table but were not considered in downstream analyses.
Fractionation Analysis Four individual fractionated CSF samples (fractions 6–15) were submitted to Olink for analysis. Only fractions 7, 9, 10, 11, 12 and 13 were used to verify whether a protein exhibited the fractionation pattern typical for EV‐associated proteins (Norman et al. ). For each fraction of interest, the median NPX was calculated for each protein. A protein was considered to have a fractionation pattern typical of EV‐associated proteins if the medians of fractions 9 and 10 were greater than the medians for fractions 7, 11, 12 and 13.
Protein Localization The proteins in the Olink panel were computationally determined to be transmembrane, internal, or external using DeepTMHMM. DeepTMHMM is a deep learning model‐based algorithm that uses a hidden Markov model to predict subcellular localization of a protein in a cell. The model calculated a probability for each amino acid in each protein and returned the highest probability domain for each amino acid, allowing the most likely localization of the overall protein to be determined. Using this model, each amino acid was characterized as: Cytosolic Alpha transmembrane helix Beta transmembrane barrel Signalling peptide External to the cell and any secreted vesicles or exosomes Information regarding signalling peptides was not considered, as they are largely cleaved from the protein when it enters the endoplasmic reticulum, and, therefore, are unlikely to be present in the epitope of the protein found in EVs (Liaci and Forster ). We classified a protein as internal to the cell if all its amino acids were characterized as cytosolic, and we classified a protein as external if all its amino acids were characterized as being outside the cell or on secreted vesicles Figure . Because proteins containing a transmembrane domain also contain domains found internal and external to the cell, a protein was classified as transmembrane if it contained one or more amino acids characterized as an alpha transmembrane helix or a beta transmembrane barrel. Because of the budding mechanisms by which EVs are secreted (Teng and Fussenegger ), it is largely assumed that proteins would have the same cytosolic, transmembrane, or extracellular domains in both EVs and the cell. However, additional validation techniques would be necessary to confirm the localization of proteins relative to EVs, a problem we address through SEC fractionation analysis as described previously.
EV‐Associated Protein Identification This pipeline was used to identify proteins that may be associated with EVs. Proteins were labelled as internal to EVs if they met the fractionation criteria and were identified as internal using DeepTMHMM as described previously. The same criteria were followed to identify transmembrane and external proteins associated with EVs. This yielded a list that was further narrowed by selecting proteins considered to be cell type‐specific based on the Tau score and BrainRNA‐Seq dataset as described below. Each protein was assigned an “EV Association Score,” which was calculated as the ratio of the median NPX for the EV fractions (fractions 9 and 10), and the median NPX for fractions 7, 11, 12 and 13. This value is shown on the y ‐axis of Figure .
Cell‐Type‐Specificity The BrainRNA‐Seq atlas reports fragments per kilobase per million mapped fragments (FPKM), collected via RNA sequencing (Zhang et al. ). The mean FPKM of each gene was used for mature astrocytes, neurones, oligodendrocytes, endothelial cells, and microglia. Foetal astrocytes were excluded from analysis. Tau specificity score is used to determine cell‐type‐specificity of genes, as it gives a numerical indication of the relative specificity of a gene across different cell types or tissues. Scores range between 0 and 1, where 0 indicates that a gene is ubiquitously expressed in all cell types, and 1 indicates that a gene is entirely expressed in a single cell type (Kryuchkova‐Mostacci and Robinson‐Rechavi ). A gene was considered specific to a given cell type if it had a Tau specificity score of greater than 0.75 and if the mean FKPM was highest in the cell type of interest relative to the other cell types. Tau specificity scores were calculated using the following formula (Kryuchkova‐Mostacci and Robinson‐Rechavi ): τ = ∑ i = 1 n 1 − x i ^ n − 1 x i ^ = x i max 1 ≤ i ≤ n x i x i = e x p r e s s i o n o f t h e g e n e o f i n t e r e s t i n t i s s u e i n = n u m b e r o f t i s s u e s The opposite is also true—by identifying proteins with a low Tau specificity score, < 0.25, we selected genes that are ubiquitously expressed in all cell types. The genes were then mapped to proteins using data obtained from the UniProt website. This data is included in Table .
Brain Organ Specificity The GTEx database provides median gene‐level expression transcripts per million (TPM) by tissue (Kowal et al. ). The tissues were grouped as described in Table below: The median TPM of each organ group was used to calculate the tissue specificity of each gene. The tau specificity score was used to determine the organ specificity of each gene, as it gives a numerical indication of the relative specificity of a gene across different organ groups. Scores range between 0 and 1, where 0 indicates that a gene is ubiquitously expressed in all tissue, and 1 indicates that a gene is entirely specific to a single tissue type (Kryuchkova‐Mostacci and Robinson‐Rechavi ). This data is included in Table for reference but is not considered in quantifying cell‐type‐specificity as shown in Figure .
Results We used a highly multiplexed proximity extension assay platform from Olink to analyse thousands of proteins from microlitres of biofluid with high specificity (Olink ). To assess EVs coming from the brain, we fractionated CSF from healthy individuals using SEC to separate proteins that peak in the early EV fractions from those that peak in the late secreted protein fractions (Thery et al. ; Welsh et al. ). To define our EV fractions, adhering to MISEV 2018 18 and 2023 19 guidelines, we analysed 20% of each fraction using our previously validated Simoa assay for CD81 to demonstrate that EVs predominantly eluted in fractions 9 and 10 (Figure ) (Norman et al. ; Ter‐Ovanesyan et al. , , , ). The remaining 80% of each fraction was analysed using the Olink HT platform, which quantifies 5416 unique proteins (Table ). We analysed the fractionation pattern of CD63 using data from the Olink assay and demonstrated a peak in signal in fractions 9 and 10 (Figure ). We performed nanoparticle tracking analysis to show that EV‐sized particle counts are increased in fractions 7–10 (Figure ). Next, we performed Western blots of CD9, CD63 and CD81 on fractionated CSF and demonstrated peak signals in fractions 9 and 10 for all three tetraspanins. Of note, in the Olink data, CD63 had a second later peak, which was not observed in Western blotting, indicating this peak may be caused by nonspecific binding in the setting of high protein abundance in the later soluble protein fractions. Finally, in agreement with the literature (Kowal et al. ; You et al. ; Jeppesen et al. ; Hallal et al. ) and MISEV 2018 18 , we also report several previously identified generic EV markers, Annexins A2 (ANXA2), A4 (ANXA4), and A5 (ANXA5), and Vacuolar protein sorting‐associated protein VTA1 homolog (VTA1), and non‐EV contaminant markers fibronectin (FN1), prothrombin (F2), pigment epithelium‐derived factor (PEDF, also known as SERPINF1), and complement C3 (C3) (Figure ) included in our Olink pipeline. To identify targets that could be effective for EV immunocapture or for the analysis of EV cargo, we selected all proteins where the median NPX value across CSF samples was greater in both fractions 9 and 10 compared to fractions 7, 11, 12 and 13 (Table ). Because many proteins can be found as both EV‐bound and soluble isoforms, we did not consider relative protein abundance in fractions 14 and 15 in our criteria, but rather selected proteins where a definable EV fractionation pattern could be seen. The signal from EV‐associated proteins begins to peak from fraction 8 and reaches its highest point in fractions 9 and 10. With minimal to no signal observed in fractions 6 and 7, we treat these two fractions as internal controls. However, due to proportional signals from fractions 6 and 7 based on both CD81 Simoa (Figure ) and CD63 Olink (Figure ) assays, we used fraction 7, rather than the combination of fractions 6 and 7, in our EV fractionation pattern selection criteria. Next, we utilized the DeepTMHMM deep learning model to differentiate cytosolic, transmembrane, and external proteins. Running this model on each protein analysed by the Olink platform, we categorized them into 953 predicted transmembrane, 3522 predicted cytosolic, and 941 predicted external proteins (Hallgren et al. ). We demonstrate that 80% of predicted cytosolic proteins, 10% of transmembrane proteins, and 9% of external proteins have a definable EV fractionation pattern (Figure ). The HT panel from Olink measures two negative controls for each assay. Olink recommends against calculating a lower LOD with fewer than 10 negative controls in a dataset, so we instead considered the fixed LODs made available by Olink, available in Table . The fixed LOD calculation is based on 24–36 negative controls, ensuring a more robust calculation to minimize the higher variation among negative controls. When thresholding our data using the LODs, we observed a significant loss of many targets. In total, without considering LOD, our analysis pipeline identifies 57 unique transmembrane and internal proteins associated with EVs. However, when we threshold using the fixed LOD, this number drops to three proteins across the five cell types. This is consistent with the recommendations from Olink, which reports that values below LOD are unlikely to increase the risk of false positive discoveries and may risk eliminating informative biomarkers. They also highlight that filtering data based solely on LOD may remove meaningful signals, especially when a protein is well expressed in one group but undetectable in another. For example, Aquaporin 1 (AQP1), a transmembrane protein specific to astrocytes, is eliminated from consideration when thresholding due to all fractions except fraction 9 being below LOD. However, we independently validated its fractionation pattern via Western blot (Figure ) and received results similar to those reported through Olink. Therefore, excluding data points below LOD could prevent us from including potentially useful proteins in our analyses. Our primary interest in using this dataset was to identify proteins that can be used to isolate or define an EV's cell of origin. Therefore, we overlaid the Olink data with the BrainRNA‐Seq atlas and selected proteins that were enriched in a specific brain cell‐type—as defined by having a Tau specificity score > 0.75 14 , calculated using the mean astrocyte, oligodendrocyte, microglia, neuron, and endothelial cell expression levels. Thus, we identified candidate transmembrane and external proteins that can potentially be used in CSF to isolate cell‐type‐specific brain‐derived EVs as well as candidate cytosolic proteins that can be analysed as internal EV cargo to confirm cell‐type‐specificity following immunocapture (Figure ). Finally, we identified a set of proteins that demonstrate a clear EV fractionation pattern but are not specific to a given cell type as defined by a Tau score < 0.25 (Table ). These latter proteins can be used to normalize total EV quantity.
Conclusions By utilizing the highly sensitive and specific multiplexed Olink platform on SEC‐fractionated healthy CSF, we identified cell‐type‐specific proteins that may be associated with EVs and can be used both for potential EV immunocapture and for the analysis of the luminal protein cargo of brain‐derived EVs. Furthermore, we demonstrate that 90% of predicted transmembrane proteins did not have a definable EV fractionation pattern, which we speculate is due to overwhelming signals from cleaved or secreted isoforms of these proteins. Such targets are likely not viable for use in EV immunocapture. Conversely, some targets identified as external were highly EV‐associated (e.g., EDIL3) and are likely bound tightly to the extravesicular surface, making them potential immunocapture targets. There are several important caveats to this work: First, although we analysed 5416 targets, this remains only a quarter of the ∼20,000 proteins known to be in the human protein coding genome (Aebersold et al. ). Second, many proteins are known to be found in both secreted and transmembrane forms. In some cases, the abundance of the secreted form can mask an EV peak in fractions 9 and 10 after SEC. Without the ability to separate the peak in fractions 9 and 10, which includes EVs and associated proteins, from the soluble protein peak in fractions 14 and 15, these proteins are likely not useful for EV enrichment unless they have a unique extracellular epitope absent on the cleaved and secreted forms (Shami‐shah et al. ; Norman et al. ). Third, while proximity extension assays lower the chance of nonspecific binding as can occur with ELISAs, the soluble protein fractions have substantially more protein, increasing the chance for nonspecific binding interactions to produce a signal, as was seen in our CD63 comparison between Olink and Western blot. Thus, our analysis is useful for identifying potential EV‐associated proteins but cannot rule out EV association for proteins that do not meet our stringent criteria or that are not included in the Olink HT panel. This dataset supports the necessity of running SEC or DGC on putative immunocapture targets before proceeding to EV immunocapture. Fourth, our cell‐type‐specificity analysis accounts for within‐brain specificity but does not include specificity for cell types outside of the brain. As can be seen using the GTex database (Consortium ) (overlayed in Table ), some of the cell‐type‐specific targets in Figure are highly specific to the brain (e.g., TBR1, FGF1 and SOX2), while others are also expressed on several cell types outside the brain (e.g., NDUFAF4, TOMM20 and PRDX6). Therefore, for EV analysis of CSF, our specificity criteria are likely sufficient, as the majority of CSF proteins come from within the central nervous system. However, if future work is to utilize these targets for analysis in blood, more stringent criteria overlaying the GTEx database would need to be included to ensure cell‐type‐specificity. Due to the NGS readout, Olink has a wide dynamic range of 10 logs (fg‐mg/mL) while requiring as little as 2 µL of sample input (Shami‐shah et al. ; Olink ). In contrast, depending on the instrument type, mass spectrometry has a much narrower dynamic range of 4–5 logs (Tang et al. ; Marshall et al. ). This narrower range necessitates greater sample input, depletion of higher abundant contaminant proteins (e.g., albumin, lipoproteins, immunoglobulins), and more complex sample processing and cleanup (leading to additional sample loss) to detect lower abundance proteins. Therefore, while discovery‐based mass spectrometry is a powerful technology, its lower dynamic range results in the preferential detection of highly abundant proteins, limiting the ability to access the low‐abundance EV proteome. While previous research has explored EVs derived from brain tissues, cell‐type‐specific media collected from induced pluripotent stem cells of various cell types, and CSF‐derived EVs without any consideration of cell‐type‐specificity (Muraoka, Jedrychowski, et al. ; You et al. ; Muraoka, DeLeo, et al. ), to our knowledge, this dataset provides the first unbiased proteomic profiling of EV association on a large scale, making it a valuable resource for future EV biomarker discovery. Additionally, with the growing importance of cell type‐specific EVs in liquid biopsy for hard‐to‐biopsy organs (e.g., the brain), we have created a computational approach based on stringent criteria to discover potential cell type‐specific brain‐derived EV biomarkers. While many of the proteins we identified as EV‐associated have been described previously in the literature (Oshikawa‐Hori et al. ; Hoshino et al. ; Gupta et al. ), substantial additional work is required to assess cell origin for those proteins that meet the criteria displayed in Figure . In future work, we plan to validate these candidate proteins for each cell type, prioritizing those with the highest Tau and EV association scores. Validation will require immunocapture with antibodies to a transmembrane or external protein and analysis of proposed internal targets with Simoa following a proteinase protection assay. We nonetheless feel that this process of target validation should be done collaboratively within the EV community. Therefore, this dataset is an important and powerful new resource for identifying novel targets for brain‐derived EVs.
Maia Norman : Conceptualization (lead), data curation (lead), formal analysis (lead), funding acquisition (lead), investigation (lead), methodology (lead), project administration (lead), writing–original draft (lead), writing–review and editing (lead). Adnan Shami‐shah : Conceptualization (lead), data curation (lead), formal analysis (lead), investigation (lead), methodology (lead), project administration (lead), writing–original draft (lead), writing–review and editing (lead). Sydney C. D'Amaddio : Formal analysis (lead), methodology (lead), resources (lead), software (lead), writing–review and editing (supporting). Benjamin G. Travis : Data curation (lead), methodology (lead), writing–review and editing (supporting). Dmitry Ter‐Ovanesyan : Formal analysis (supporting), investigation (supporting), methodology (supporting). Tyler J. Dougan : Methodology (supporting), Software (supporting). David R. Walt : Conceptualization (supporting), funding acquisition (lead), project administration (lead), Writing–review and editing (lead).
All human samples utilized in this work were purchased from commercial sources. All patients were appropriately consented. The use of these samples was approved by the Mass General Brigham IRB.
David R. Walt is a founder and equity holder in Quanterix. His interests were reviewed and are managed by Mass General Brigham in accordance with their conflicts of interest policies.
SI Figure . Nanoparticle Tracking Analysis of EVs in CSF SEC fractions 7–10 and 11–15. SI Figure . Western blots of tetraspanins in CSF SEC fractions and human brain lysate. SI Figure 3. Western blots of cell type‐specific transmembrane proteins in CSF SEC fractions and human brain lysate. SI Table 1. Raw data of EV fractions isolated from CSF of four healthy individuals subjected to Olink HT panel analysis. SI Table 2. Data of EV fractions isolated from CSF of four healthy individuals subjected to Olink HT panel analysis mapped to the appropriate fixed lower LOD provided by Olink. SI Table 3. List of proteins that are classified as internal, transmembrane, and external based on the DeepTMHMM deep learning model and meet the EV fractionation pattern criteria. SI Table 4. List of proteins that are classified as internal transmembrane, and external based on the DeepTMHMM deep learning model, meet the EV fractionation pattern criteria, and have a Tau specificity score of <0.25. SI Table 5. List of proteins that are classified as either transmembrane, internal, or external based on the DeepTMHMM deep learning model, meet the EV fractionation pattern criteria, and have a Tau specificity score of > 0.75 for astrocytes, endothelial, microglia, oligodendrocytes, and neurones.
|
Dental Caries, Tooth Erosion and Nutritional Habits in a Cohort of Athletes: A Cross-Sectional Study | 40a5f13c-db68-49d7-ba2b-91186d6db39e | 11820296 | Dentistry[mh] | Regular physical activity is widely acknowledged as vital for a healthy lifestyle across all ages . Recently, global interest in sports for well-being has surged, with more people participating in physical activities, both amateur and professional. This aligns with advancements in sports medicine, a multidisciplinary field aimed at optimizing athletic performance and health. Within this field, sports dentistry is becoming a crucial but often overlooked specialty, addressing athletes’ unique oral health challenges . Athletes are highly vulnerable to dental caries and erosion, significant concerns in sports medicine . Dental caries involves localized tooth demineralization due to acids produced by bacteria, while erosion is due to non-bacterial acid exposure causing overall irreversible enamel loss . Factors like high-sugar energy drinks, acidic supplements, and intense physical exertion impacting saliva flow and composition exacerbate these phenomena . These oral health conditions may adversely affect an athlete’s ability to do sports and overall well-being. Sports nutrition is crucial for enhancing athletic performance, affecting energy, endurance and recovery . Nutritional strategies for pre-, intra- and post-exercise phases focus on carbohydrates, electrolytes and energy-dense foods or drinks . The frequent consumption of such aids, high in fermentable sugars and acids, increases the risk of dental caries and erosion . Additionally, decreased salivary flow during intense training or competition exacerbates this risk, as saliva is essential for neutralizing acids and promoting remineralization. Despite the clear links between nutrition, oral health and athletic performance, the oral health needs of athletes remain underexplored in both clinical practice and research . Understanding the specific risk factors for dental caries and erosion in athletes is critical for developing targeted preventive strategies. Tooth decay is influenced by diet, oral care, saliva composition, and environmental and behavioral factors. Evaluation methods like the Caries Management by Risk Assessment (CAMBRA) system offer a methodical way to assess susceptibility by examining these factors . The recognition of such risk factors (such as fermentable carbohydrate intake, inadequate oral hygiene, and decreased saliva production) and protective elements (like fluoride usage and salivary buffering ability) inform personalized strategies . Implementing this evidence-based model is crucial when studying groups like athletes, who may be more prone to dental decay due to specific dietary habits and oral health challenges . This research adheres to CAMBRA’s principles and seeks to contribute to targeted preventive measures for this high-risk population. This study aims to bridge this gap by investigating the association between dental caries and erosion in athletes, and some dietary- and oral health habits. Our hypothesis is that athletes with a higher consumption of acidic or sugary beverages, and inadequate oral health habits, will exhibit a higher prevalence of dental caries and erosion compared to athletes with healthier dietary and oral hygiene practices. 2.1. Study Design, Setting and Participants For the present cross-sectional study, data from a cohort of athletes followed at a Sports Dentistry Department of a university clinic (Egas Moniz Dental Clinic, Almada, Portugal) were used. The data were gathered through the consecutive sampling of new patients seeking initial consultations from September 2023 to March 2024. The Egas Moniz Ethics Committee approved the study (1285/2023, approval date: 30 November 2023), and all participants provided informed consent. This report adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines . 2.2. Eligibility Criteria and Sampling To be included in this study, participants had to: have 18 years of age or older; being able to read, understand, and sign the informed consent form; have declared that they play sport at any level; and were seeking initial triage at the Sports Dentistry department. Participants were invited to participate voluntarily and anonymously. Given the lack of studies comparing oral conditions with oral health values and quality of life, a minimum sample size was not estimated thus we carried out for a preliminary design, with a consecutive random sample obtained over a 6-month period. 2.3. Reliability and Calibration Prior to the start of the study, the examiner (BM) was trained in the diagnosis of dental caries and erosive lesions through the detailed study of exemplary photographs, in order to improve the skills for identifying caries and erosion lesions and to train the use of the Basic Erosive Wear Examination (BEWE) and International Caries Detection and Assessment System II (ICDAS II) indices. Secondly, 10 individuals were selected to be examined both by the examiner and by an experienced observer (CR), considered to be the Gold Standard. Overall, a very good agreement was achieved for BEWE (85.7%, Standard Error [SE] = 11.1) and ICDAS (82.3%, SE = 12.1). 2.4. Variables 2.4.1. Outcome Variables During the oral examination, participants were positioned in a chair while the examiner utilized a disposable dental mirror, a light source, and cotton rolls for cleaning and drying teeth to conduct the assessment. Using the four-level BEWE , we assessed erosive lesions on all permanent teeth surfaces, with the exception of third molars. In each sextant, the most affected surface was documented, and the sum of these scores was computed. This total was then utilized to assign an individual risk level as follows: no risk (BEWE ≤ 2); mild risk (3 < BEWE < 8); moderate (9 < BEWE < 13); high (BEWE ≥ 14) . Caries was assessed using the ICDAS II . We chose ICDAS instead of WHO decayed, missing and filled index due to its more sensitivity in estimating caries prevalence and extent compared to the WHO criteria . As per ICDAS criteria, the sites were recorded by a 0 to 6 scoring system: 0 = sound; 1 = first visual change in enamel; 2 = distinct visual change in enamel; 3 = localized enamel breakdown (without clinical visual signs of dentinal involvement); 4 = Underlying dark shadow from dentin; 5 = Distinct cavity with visible dentin; 6 = Extensive distinct cavity with visible dentin. Patients with carious lesions and/or erosion received required treatment and/or clinical management. 2.4.2. Exposure Variables Data were collected through a self-reported questionnaire on sociodemographic characteristics and behavioral aspects. This questionnaire was administered prior to taking a panoramic radiography and the clinical oral observation, which are both part of the protocol for first intake appointments. Overall, the information collected included sex, age, frequency and type of sport, an oral health self-assessment and a nutritional questionnaire. Oral health self-assessment consisted of several questions: “How often do you brush your teeth each day?”, “What do you use for oral hygiene?”, “How regularly do you visit a dentist”, “Have you ever had or do you have any of the oral health problems listed below?”, “How do you rate your oral health?” (very bad, bad, satisfactory, good, very good, don’t know/refuse to answer). In the frequency and type of sport we inquired the sport practiced, the frequency per week, the level of sport (recreational, competitive amateur and high performance). Questions on nutrition consisted of several questions about nutrition, including the frequency of eating foods with cariogenic and/or erosive potential and sports: soft drinks, coffee, tea, fruit juices and lemonade, energy drinks, isotonic drinks, isotonic gels, whey protein, ergogens (creatine, caffeine), B-Alkaline, sodium bicarbonate, multivitamins, L-carnitine, Omega-3. The frequency of consumption of the above-mentioned items were defined as follows: never (low frequency); once a week (low frequency); two to four times a week (moderate frequency); four to six times a week (moderate frequency); once a day (high frequency); twice a day (high frequency); more than twice a day (high frequency). 2.5. Statistical Analysis Data were stored and recorded and statistically processed using IBM SPSS statistics software, version 29.0.1.0. Descriptive statistical analysis covered the investigation of measures such as mean, median, variance, standard deviation, minimum value, maximum value and interquartile range (IQR). We performed normality tests to determine the distribution of variables in the dataset. A significance level of 5% was adopted for all inferential analyses. To evaluate the associations between demographic, behavioral, and clinical factors with dental caries and dental erosion, multivariable regression models were employed, adjusting for potential confounders. We conducted four regressions: two linear regression models for the number of dental caries and the BEWE score; and, two logistic regression models for the presence of dental caries (presence of dental caries lesions vs. absence of dental caries lesions) and the risk of BEWE (BEWE ≤ 2 vs. BEWE > 2]. Preliminary analyses used univariate models. A multivariate model was then constructed for the outcome variable CAL ≥ 3 mm. Only variables with p ≤ 0.25 in the univariate model were included in the multivariate stepwise procedure. According to each linear regression we used the term “Estimate” referring to either number of dental caries lesions or BEWE score. For the present cross-sectional study, data from a cohort of athletes followed at a Sports Dentistry Department of a university clinic (Egas Moniz Dental Clinic, Almada, Portugal) were used. The data were gathered through the consecutive sampling of new patients seeking initial consultations from September 2023 to March 2024. The Egas Moniz Ethics Committee approved the study (1285/2023, approval date: 30 November 2023), and all participants provided informed consent. This report adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines . To be included in this study, participants had to: have 18 years of age or older; being able to read, understand, and sign the informed consent form; have declared that they play sport at any level; and were seeking initial triage at the Sports Dentistry department. Participants were invited to participate voluntarily and anonymously. Given the lack of studies comparing oral conditions with oral health values and quality of life, a minimum sample size was not estimated thus we carried out for a preliminary design, with a consecutive random sample obtained over a 6-month period. Prior to the start of the study, the examiner (BM) was trained in the diagnosis of dental caries and erosive lesions through the detailed study of exemplary photographs, in order to improve the skills for identifying caries and erosion lesions and to train the use of the Basic Erosive Wear Examination (BEWE) and International Caries Detection and Assessment System II (ICDAS II) indices. Secondly, 10 individuals were selected to be examined both by the examiner and by an experienced observer (CR), considered to be the Gold Standard. Overall, a very good agreement was achieved for BEWE (85.7%, Standard Error [SE] = 11.1) and ICDAS (82.3%, SE = 12.1). 2.4.1. Outcome Variables During the oral examination, participants were positioned in a chair while the examiner utilized a disposable dental mirror, a light source, and cotton rolls for cleaning and drying teeth to conduct the assessment. Using the four-level BEWE , we assessed erosive lesions on all permanent teeth surfaces, with the exception of third molars. In each sextant, the most affected surface was documented, and the sum of these scores was computed. This total was then utilized to assign an individual risk level as follows: no risk (BEWE ≤ 2); mild risk (3 < BEWE < 8); moderate (9 < BEWE < 13); high (BEWE ≥ 14) . Caries was assessed using the ICDAS II . We chose ICDAS instead of WHO decayed, missing and filled index due to its more sensitivity in estimating caries prevalence and extent compared to the WHO criteria . As per ICDAS criteria, the sites were recorded by a 0 to 6 scoring system: 0 = sound; 1 = first visual change in enamel; 2 = distinct visual change in enamel; 3 = localized enamel breakdown (without clinical visual signs of dentinal involvement); 4 = Underlying dark shadow from dentin; 5 = Distinct cavity with visible dentin; 6 = Extensive distinct cavity with visible dentin. Patients with carious lesions and/or erosion received required treatment and/or clinical management. 2.4.2. Exposure Variables Data were collected through a self-reported questionnaire on sociodemographic characteristics and behavioral aspects. This questionnaire was administered prior to taking a panoramic radiography and the clinical oral observation, which are both part of the protocol for first intake appointments. Overall, the information collected included sex, age, frequency and type of sport, an oral health self-assessment and a nutritional questionnaire. Oral health self-assessment consisted of several questions: “How often do you brush your teeth each day?”, “What do you use for oral hygiene?”, “How regularly do you visit a dentist”, “Have you ever had or do you have any of the oral health problems listed below?”, “How do you rate your oral health?” (very bad, bad, satisfactory, good, very good, don’t know/refuse to answer). In the frequency and type of sport we inquired the sport practiced, the frequency per week, the level of sport (recreational, competitive amateur and high performance). Questions on nutrition consisted of several questions about nutrition, including the frequency of eating foods with cariogenic and/or erosive potential and sports: soft drinks, coffee, tea, fruit juices and lemonade, energy drinks, isotonic drinks, isotonic gels, whey protein, ergogens (creatine, caffeine), B-Alkaline, sodium bicarbonate, multivitamins, L-carnitine, Omega-3. The frequency of consumption of the above-mentioned items were defined as follows: never (low frequency); once a week (low frequency); two to four times a week (moderate frequency); four to six times a week (moderate frequency); once a day (high frequency); twice a day (high frequency); more than twice a day (high frequency). During the oral examination, participants were positioned in a chair while the examiner utilized a disposable dental mirror, a light source, and cotton rolls for cleaning and drying teeth to conduct the assessment. Using the four-level BEWE , we assessed erosive lesions on all permanent teeth surfaces, with the exception of third molars. In each sextant, the most affected surface was documented, and the sum of these scores was computed. This total was then utilized to assign an individual risk level as follows: no risk (BEWE ≤ 2); mild risk (3 < BEWE < 8); moderate (9 < BEWE < 13); high (BEWE ≥ 14) . Caries was assessed using the ICDAS II . We chose ICDAS instead of WHO decayed, missing and filled index due to its more sensitivity in estimating caries prevalence and extent compared to the WHO criteria . As per ICDAS criteria, the sites were recorded by a 0 to 6 scoring system: 0 = sound; 1 = first visual change in enamel; 2 = distinct visual change in enamel; 3 = localized enamel breakdown (without clinical visual signs of dentinal involvement); 4 = Underlying dark shadow from dentin; 5 = Distinct cavity with visible dentin; 6 = Extensive distinct cavity with visible dentin. Patients with carious lesions and/or erosion received required treatment and/or clinical management. Data were collected through a self-reported questionnaire on sociodemographic characteristics and behavioral aspects. This questionnaire was administered prior to taking a panoramic radiography and the clinical oral observation, which are both part of the protocol for first intake appointments. Overall, the information collected included sex, age, frequency and type of sport, an oral health self-assessment and a nutritional questionnaire. Oral health self-assessment consisted of several questions: “How often do you brush your teeth each day?”, “What do you use for oral hygiene?”, “How regularly do you visit a dentist”, “Have you ever had or do you have any of the oral health problems listed below?”, “How do you rate your oral health?” (very bad, bad, satisfactory, good, very good, don’t know/refuse to answer). In the frequency and type of sport we inquired the sport practiced, the frequency per week, the level of sport (recreational, competitive amateur and high performance). Questions on nutrition consisted of several questions about nutrition, including the frequency of eating foods with cariogenic and/or erosive potential and sports: soft drinks, coffee, tea, fruit juices and lemonade, energy drinks, isotonic drinks, isotonic gels, whey protein, ergogens (creatine, caffeine), B-Alkaline, sodium bicarbonate, multivitamins, L-carnitine, Omega-3. The frequency of consumption of the above-mentioned items were defined as follows: never (low frequency); once a week (low frequency); two to four times a week (moderate frequency); four to six times a week (moderate frequency); once a day (high frequency); twice a day (high frequency); more than twice a day (high frequency). Data were stored and recorded and statistically processed using IBM SPSS statistics software, version 29.0.1.0. Descriptive statistical analysis covered the investigation of measures such as mean, median, variance, standard deviation, minimum value, maximum value and interquartile range (IQR). We performed normality tests to determine the distribution of variables in the dataset. A significance level of 5% was adopted for all inferential analyses. To evaluate the associations between demographic, behavioral, and clinical factors with dental caries and dental erosion, multivariable regression models were employed, adjusting for potential confounders. We conducted four regressions: two linear regression models for the number of dental caries and the BEWE score; and, two logistic regression models for the presence of dental caries (presence of dental caries lesions vs. absence of dental caries lesions) and the risk of BEWE (BEWE ≤ 2 vs. BEWE > 2]. Preliminary analyses used univariate models. A multivariate model was then constructed for the outcome variable CAL ≥ 3 mm. Only variables with p ≤ 0.25 in the univariate model were included in the multivariate stepwise procedure. According to each linear regression we used the term “Estimate” referring to either number of dental caries lesions or BEWE score. 3.1. Study Sample and Characteristics From an initial sample of 83 participants, 3 refused to complete the questionnaires after accepting participating. A final sample of 80 participants was obtained, with an age range (24.2 ± 4.0 years), with a predominance of men (70.0%) . Regarding dietary habits, 53.8% consumed 2–3 meals per day, while 45.0% consumed 4–5 meals. Most participants (77.5%) practiced one sport, with a median of four training sessions per week (IQR = 3) and two hours per session (IQR = 0.5). A majority had been practicing for less than 5 years (77.5%), with fewer reporting 5–10 years (22.5%), 10–15 years (30.0%), or more than 15 years (43.8%). Sugary drinks or foods were consumed occasionally by 53.8%, on most days by 36.3%, and daily by 6.3%, while 3.8% reported never consuming them. Participants demonstrated varied oral hygiene habits , with all reporting toothbrush use (100%), followed by flossing (53.8%), mouthwash (41.3%), interdental brush (7.5%), and tongue brushing (11.3%). Fluoride toothpaste was reported by 67.5%. Regarding dental visits, 63.8% attended regularly even without complaints, 30.0% visited only when in pain or with complaints, while 2.5% never visited a dentist, and 3.8% were uncertain. Most participants had their last dental visit within six months (60.0%), with fewer reporting visits within 6–12 months (23.8%), over 12 months ago (13.8%), or uncertainty (2.5%). Oral health impacted training and competition in 15.0% and 12.5% of participants, respectively. Self-assessment of oral health was predominantly “Good” (51.3%) or “Very good” (23.8%), with smaller proportions reporting “Satisfactory” (20.0%) or “Bad” (5.0%). Half of the participants (50.0%) had caries, with a mean of 1.6 lesions (SD = 2.6). The prevalence of erosion was 40.0%, BEWE risk mostly absent (80.0%), with mild (12.5%), moderate (5.0%), or high risk (2.5%) reported less frequently. The distribution of consumption of different types of drinks was also analyzed . 3.2. Dental Caries and Associated Factors In the number of active caries lesions, significant associations were observed with self-perceived oral health and dietary habits . “Good” (Estimate: −5.01, p < 0.001) or “Very good” (Estimate: −5.46, p < 0.001) had significantly association with the number of dental caries compared to those with poorer self-perception, with “Satisfactory” ratings approaching significance (Estimate: −5.21, p < 0.001). Similarly, lower frequencies of sugary snack consumption were associated with the number of dental caries lesions. Other factors showed no significant associations. For the dichotomous variable of existing dental caries we did not observe significant associations in the final adjusted model . 3.3. Dental Erosion and Associated Factors The adjusted multivariable linear regression models for the associations between demographic, behavioral, and clinical factors with dental erosion shows statistically significant results with BEWE as a continuous variable but not with is dichotomous transformation . Overall, the final adjusted model revealed significant associations between various behavioral and self-perceived factors and overall BEWE score. Participants who reported uncertainty about their meal frequency (“Don’t know”) had significantly lower dental erosion (Estimate: −12.56, p = 0.014), while those uncertain about their last dental visit (“Don’t know”) exhibited significantly higher scores (Estimate: 8.82, p = 0.014). Self-perceived oral health status showed a clear trend: participants who rated their oral health as “Good” (Estimate: −5.28, p = 0.005) or “Very good” (Estimate: −6.12, p = 0.003) had significantly lower dental erosion scores compared to those with poorer self-perception, with “Satisfactory” ratings approaching significance (Estimate: −3.85, p = 0.052). Conversely, meal frequency categories (4–5 meals/day) and the timing of the last dental visit (>12 months or 6–12 months) did not yield significant associations with BEWE ( p > 0.05). From an initial sample of 83 participants, 3 refused to complete the questionnaires after accepting participating. A final sample of 80 participants was obtained, with an age range (24.2 ± 4.0 years), with a predominance of men (70.0%) . Regarding dietary habits, 53.8% consumed 2–3 meals per day, while 45.0% consumed 4–5 meals. Most participants (77.5%) practiced one sport, with a median of four training sessions per week (IQR = 3) and two hours per session (IQR = 0.5). A majority had been practicing for less than 5 years (77.5%), with fewer reporting 5–10 years (22.5%), 10–15 years (30.0%), or more than 15 years (43.8%). Sugary drinks or foods were consumed occasionally by 53.8%, on most days by 36.3%, and daily by 6.3%, while 3.8% reported never consuming them. Participants demonstrated varied oral hygiene habits , with all reporting toothbrush use (100%), followed by flossing (53.8%), mouthwash (41.3%), interdental brush (7.5%), and tongue brushing (11.3%). Fluoride toothpaste was reported by 67.5%. Regarding dental visits, 63.8% attended regularly even without complaints, 30.0% visited only when in pain or with complaints, while 2.5% never visited a dentist, and 3.8% were uncertain. Most participants had their last dental visit within six months (60.0%), with fewer reporting visits within 6–12 months (23.8%), over 12 months ago (13.8%), or uncertainty (2.5%). Oral health impacted training and competition in 15.0% and 12.5% of participants, respectively. Self-assessment of oral health was predominantly “Good” (51.3%) or “Very good” (23.8%), with smaller proportions reporting “Satisfactory” (20.0%) or “Bad” (5.0%). Half of the participants (50.0%) had caries, with a mean of 1.6 lesions (SD = 2.6). The prevalence of erosion was 40.0%, BEWE risk mostly absent (80.0%), with mild (12.5%), moderate (5.0%), or high risk (2.5%) reported less frequently. The distribution of consumption of different types of drinks was also analyzed . In the number of active caries lesions, significant associations were observed with self-perceived oral health and dietary habits . “Good” (Estimate: −5.01, p < 0.001) or “Very good” (Estimate: −5.46, p < 0.001) had significantly association with the number of dental caries compared to those with poorer self-perception, with “Satisfactory” ratings approaching significance (Estimate: −5.21, p < 0.001). Similarly, lower frequencies of sugary snack consumption were associated with the number of dental caries lesions. Other factors showed no significant associations. For the dichotomous variable of existing dental caries we did not observe significant associations in the final adjusted model . The adjusted multivariable linear regression models for the associations between demographic, behavioral, and clinical factors with dental erosion shows statistically significant results with BEWE as a continuous variable but not with is dichotomous transformation . Overall, the final adjusted model revealed significant associations between various behavioral and self-perceived factors and overall BEWE score. Participants who reported uncertainty about their meal frequency (“Don’t know”) had significantly lower dental erosion (Estimate: −12.56, p = 0.014), while those uncertain about their last dental visit (“Don’t know”) exhibited significantly higher scores (Estimate: 8.82, p = 0.014). Self-perceived oral health status showed a clear trend: participants who rated their oral health as “Good” (Estimate: −5.28, p = 0.005) or “Very good” (Estimate: −6.12, p = 0.003) had significantly lower dental erosion scores compared to those with poorer self-perception, with “Satisfactory” ratings approaching significance (Estimate: −3.85, p = 0.052). Conversely, meal frequency categories (4–5 meals/day) and the timing of the last dental visit (>12 months or 6–12 months) did not yield significant associations with BEWE ( p > 0.05). This study explored the prevalence and associated factors of dental caries and erosion among athletes, highlighting the importance of self-perceived oral health, dietary habits and oral hygiene behaviors in influencing these outcomes. Collectively, the prevalence of dental caries and erosion shows that 1 out of two athletes had dental caries lesions and 2 out of five had erosion lesions. Oral health self-awareness and regular dental care were the most significant risk indicators associated with the number of dental caries and the BEWE score, revealing its importance as determinants. The findings underscore the association of self-perceived oral health and preventive behaviors with dental health management among athletes. The high prevalence of dental caries (50%) and erosion lesions (40%) observed in this population emphasizes the need for targeted oral health interventions tailored to the unique dietary and physical demands of athletes. These results are in line with those reported previously in other studies . The association between oral health self-awareness and reduced dental erosion highlights the potential benefits of promoting regular dental check-ups and personalized education on protective behaviors. These insights could inform the development of preventive strategies aimed at minimizing the burden of oral diseases in athletes, thereby supporting not only their oral health but also their overall performance and well-being. The prevalence of dental caries in our sample was higher compared to the established 35.0% global prevalence data in 2010 , but in line with Azeredo et al. in 2020 with 46.25% . Half of the participants presented with dental caries, with self-perceived oral health emerging as a key determinant. Participants who assessed their oral health as “Good” or “Very good” demonstrated significantly fewer caries lesions compared to those with poorer perceptions, corroborating findings from previous studies that underscore the relevance of self-awareness in oral health. Furthermore, dietary habits played a critical role, with lower frequencies of sugary snack consumption being associated with reduced caries prevalence. These findings reinforce the need for targeted dietary counseling and oral health education among young athletes, whose nutritional demands may predispose them to higher consumption of sugar-rich diets. In contrast, no significant associations were observed for the presence of caries in the adjusted model, suggesting that other unmeasured factors, such as genetic predisposition or fluoride exposure, might contribute to this outcome. This aligns with previous research indicating the multifactorial nature of dental caries, where behavioral, biological, and environmental factors interact. Regarding the presence of fluoride in toothpaste, 67.5% of the sample reported using fluoride toothpaste, while 8.8% stated they do not use it, and 23.8% said they do not know or did not respond. More than a third of the population denies its use, highlighting the importance of continuing oral hygiene instructions and prevention at the dentist. The ADA recommends brushing for 2 min twice a day with toothpaste containing a low amount of fluoride (1000–1500 ppm) . Regarding athletes’ diet, concerning the frequency of consumption of sugary drinks, foods, or snacks, most of the sample (53.8%) reported consuming them occasionally, while 36.3% stated they consume them most days. Compared to the studies by Gallagher et al. during 2015–2016 , it was found that 28.2% of athletes have a high sugar intake in their regular diet. Therefore, the WHO (World Health Organization) recommends moderate intake of free sugars throughout life and suggests reducing their intake to less than 10.0% of total daily energy intake . As for isotonic gels, we found 31.2% of the sample consumes them, while in the mentioned study, the percentage was 70.3%. These differences may be explained by the fact that in Gallagher’s study, all participants are high-level professional athletes, whereas in our investigation, there are athletes of all levels, from amateur to high performance. Regarding dental erosion, the study revealed significant associations with behavioral and self-perceived factors. Uncertainty regarding meal frequency was associated with significantly lower dental erosion, whereas uncertainty about the timing of the last dental visit was linked to higher erosion scores. These findings could reflect a lack of structured dietary and healthcare habits among these participants. Notably, participants with “Good” or “Very good” self-perceived oral health had lower BEWE scores, further emphasizing the relationship between self-awareness and oral health outcomes. This suggests that promoting a positive perception of oral health could motivate preventive behaviors and reduce the burden of dental erosion. The prevalence of erosion in adults may range from 4% to 82% . In a systematic review by Nijakowski et al. , which included 16 studies on the prevalence of erosion in athletes, demonstrated that approximately half of the athletes studied show signs of erosion. These results are consistent with our research. Regarding location, the most affected sextants were the 4th sextant (32.6%) and the 6th sextant (33.8%), which aligns with findings of a higher prevalence of erosion in the lower first molars . Interestingly, meal frequency and the timing of the last dental visit (>12 months or 6–12 months) were not significantly associated with BEWE scores. This highlights the complex interplay of behavioral and clinical factors in dental erosion and underscores the need for further research to unravel these associations. Strengths and Limitations We assessed a self-assessment questionnaire, where responses may be overstated, and participants might feel reluctant to disclose personal details. This limitation underscores the challenge of ensuring complete honesty in self-assessment surveys. Furthermore, additional studies could be conducted to further explore potential risk indicators and specific preventive interventions aimed at reducing the prevalence of dental caries and erosion among athletes, thereby enhancing their overall well-being and sports performance. To achieve more significant results, increasing the sample size would be beneficial, preferably focusing on a more specific sports population, such as conducting the study within a single sport discipline. In the future, it would also be interesting to investigate salivary content before, during, and after training sessions, as suggested by previously . Furthermore, we intend to prospectively evaluate athletes’ dietary habits, oral health status, and their associated impact on training and performance, with the aim of exploring causal relationships and providing more precise information for future interventions. Additionally, the sample may have limited external validity if athletes with existing oral health concerns were more likely to participate, potentially skewing results. Future studies should compare athletes and non-athletes or stratify samples by sport type, training intensity, and demographics to provide more nuanced insights. Given the significant role of saliva in oral health, further research could provide additional insights into the mechanisms involved in athletes’ oral health. Also, the results shall be interpreted with caution considering the observational design that precludes causal-effect interpretations, and for these longitudinal studies shall be conducted to further confirm whether there is causality in these associations. We assessed a self-assessment questionnaire, where responses may be overstated, and participants might feel reluctant to disclose personal details. This limitation underscores the challenge of ensuring complete honesty in self-assessment surveys. Furthermore, additional studies could be conducted to further explore potential risk indicators and specific preventive interventions aimed at reducing the prevalence of dental caries and erosion among athletes, thereby enhancing their overall well-being and sports performance. To achieve more significant results, increasing the sample size would be beneficial, preferably focusing on a more specific sports population, such as conducting the study within a single sport discipline. In the future, it would also be interesting to investigate salivary content before, during, and after training sessions, as suggested by previously . Furthermore, we intend to prospectively evaluate athletes’ dietary habits, oral health status, and their associated impact on training and performance, with the aim of exploring causal relationships and providing more precise information for future interventions. Additionally, the sample may have limited external validity if athletes with existing oral health concerns were more likely to participate, potentially skewing results. Future studies should compare athletes and non-athletes or stratify samples by sport type, training intensity, and demographics to provide more nuanced insights. Given the significant role of saliva in oral health, further research could provide additional insights into the mechanisms involved in athletes’ oral health. Also, the results shall be interpreted with caution considering the observational design that precludes causal-effect interpretations, and for these longitudinal studies shall be conducted to further confirm whether there is causality in these associations. The prevalence of dental caries and dental erosion was elevated in this cohort of athletes. Dietary patterns and oral hygiene habits varied and showed significant associations with measures of dental caries and dental erosion. These findings highlight the need for targeted dietary counseling and oral health education among athletes, whose nutritional demands may predispose them to higher consumption of sugar-rich diets. |
null | d6f08626-71fc-4765-ae77-6dad426a3fd8 | 11597266 | Microbiology[mh] | Water-miscible metalworking fluids (MWFs) are used to cool and lubricate during metal removal and forming operations in a plethora of applications. MWFs are formulated and sold as concentrates that contain anything from 10 to 20 organic ingredients, mixed at the end user’s site with water that subsequently accounts for 85% to 95% of the mixture . There are three main classes of MWFs: emulsifiable oils, semisynthetics and synthetics. The main components of MWF concentrates are mineral and ester oil (sourced from plants), polyalphaolifins or glycols. To improve performance, stability and functionality, emulsifiers, corrosion inhibitors, foam control agents and lubricity enhancers are added as needed. In addition, biocidal and biostatic components are critical ingredients of MWF formulations as they help to keep microbiological agents under control. Still, the high ratio of water and the evenly mixed-in organic components, aerated by recirculation, provide a decent base of life for many planktonic bacteria, fungi and archaea in all types of MWFs . Additional habitats are provided on (machine) surfaces in temporary contact with MWFs by means of splashing, evaporation and misting. Lines that supply and discharge the MWF from the site of action are often only partially filled and offer dozens of square meters of microbial settlement area. This is a considerable problem with single-filled machines, which become multiplied in centralized systems where the fluid is transported over long distances to and from many machining centers . Microbial growth combined with metal chips and swarfs leads to clogging of filters and residue formation. Therefore, biofilms and fungal growth on surfaces are probably of far greater importance than their planktonic relatives and much more difficult to maintain . Fungi are eukaryotic, aerobic organisms, which are regularly detected in MWFs and their systems . Morphologically, fungi can be subdivided into multicellular, filamentous molds or unicellular yeast. However, dimorphism, the ability to switch from filamentous growth to a unicellular lifestyle and vice versa, has been described for some fungi, making visual identification difficult . In the past, the presence of yeasts in MWFs was thought to be unusual and generally associated with physico-chemical instabilities, but later findings showed that some yeasts ( Candida spp., Yarrowia spp.) do survive and multiply in MWFs even at a pH of 9 or higher and can be successful inhabitants. Molds grow as multicellular filaments called hyphae, which form mycelia . Molds are important types of fungi, represented by a large number of species that play key roles in the breakdown of organic matter . Molds propagate themselves through the production of spores at the head of special aerial hyphae that are easily transported by air and fluids . In MWF systems, molds are mainly found on surfaces such as splash zones or in areas that are moistened by evaporation and misting. Consequently, their presence becomes evident when spores have been discharged into the fluidic phase or if fragments have become detached. Costly downtimes and serious technical consequences can result if molds manage to colonize submerged surfaces or if parts of them are washed off and carried away as filters clog, and MWF circulation stops . Molds reportedly recovered from MWFs include members of the genera Aspergillus , Fusarium , Exophiala , Trichoderma , Cladosporium and Penicillium . In our experience, molds isolated from in-use MWFs predominantly belong to the genus Fusarium . While it used to be common for the detection of molds to be associated with low pH values, this is no longer the case today; this situation is comparable to that of yeasts. The main source is quite likely spores transported through the air, as the genus Fusarium is widely distributed in soil and often associated with agriculturally important plants, causing crop diseases . Some species, like Fusarium solani species complex ( FSSC ), have been reported to induce a range of diseases and infections in immunocompromised and healthy human beings as well in animals, mostly of aquatic species. For these reasons, FSSC has been included within pathogens related to the One Health issue . FSSC (Phylum Ascomycota, Class Sordariomycetes, Order Hypocreales; Family Nectriaceae) is a group estimated to contain at least 60 phylogenetically distinct species . To prevent fungal contamination and growth, most MWFs are protected by the incorporation of dedicated pesticides, commonly known as fungicides. However, regulatory pressure on these chemicals is increasing, leading to significant restrictions on permitted concentrations in both the concentrate and in-use products, a trend that is expected to continue . The range of potential components has already narrowed down to four: Sodium Pyrithione (NaPT), Butyl-benzisothiazolinone (BBIT), Octylisothiazolinone (OIT) and Ortho-phenyl Phenol (OPP), of which only two (NaPT and BBIT) are practical solutions for in-drum conservation. OIT has a short shelf-life in concentrates and is therefore generally used as a tank-side additive, while the solubility of OPP makes successful incorporation into many formulations challenging. In this study, we report on fungal prevalence in in-use MWFs sampled worldwide over 10 years from 2014 to 2023. According to our findings, fungal contamination of MWF systems is the norm rather than the exception, especially regarding filamentous fungi surviving on machine surfaces. In contrast, directly measurable occurrence in MWFs is rare. That technical issues do not go out of hand in the industry is often attributed to fungicides incorporated into the base formulation or added at the tank side. In laboratory experiments, we thus evaluated the inhibitory effect of fungicides and their vapors on fungal growth, sporulation and spore viability using traditional culture-dependent methods and flow cytometry. In essence, we show that the effectiveness of these fungicides is limited and dependent on the chemical composition of the MWF. Sooner rather than later, this industrial sector needs to learn how to create and use MWFs without these dedicated compounds.
2.1. Access to MWF and Residue Samples Via our extensive customer service network, we had worldwide access to MWF and residue samples from end users, both our own and external customers. 2.2. Fungicides For all experiments, technical standard fungicides from industrial suppliers were used; the concentration indicated refers to a typical dose in a freshly prepared, 5% ( w / w ) in-use MWF, i.e., NaPT (Acticide ® LV 508; Thor GmbH, Speyer, Germany; Active ingredient content: 40%) at a concentration of 250 ppm; BBIT (Densil ® DN; Arch Biocides, Atlanta GA, USA; Active ingredient content: 100%) at a concentration of 50 ppm; and OPP (PREVENTOL ® O extra; Lanxess Deutschland GmbH, Leverkusen, Germany; active ingredient content: 99.5%) at a concentration of 500 ppm. 2.3. Metalworking Fluids Most of the experiments shown were carried out with two experimental MWFs based on mineral oil that share about 50% of the ingredients (MWF A and B). Both concentrates were prepared available either with or without both NaPT and BBIT; the fungicide concentration was as described above when diluted at 5% ( w / w ). 2.4. Examination and Isolation of Fungi from MWF Samples Isolation of fungal species was based on cultivation-dependent methods such as the standard heterotrophic plate count method (HPC) or the use of dip slides. HPC was performed on Sabouraud dextrose–agar (SDA) prepared in-house (Oxoid #CM041R; Thermo Fisher, Prattlen, Switzerland) at 65 g L −1 . A total of 50 μL of either undiluted or 1:100 diluted 0.9% NaCl was plated using an Eddy Jet 2 system (IUL, Barcelona, Spain) in logarithmic mode. The plates were subsequently incubated at 30 °C for a minimum of 72 h before analysis. Fungal species were identified by Maldi-TOF MS analysis of protein patterns at Mabritec AG (Riehen, Switzerland). As dip slides, Cult Dip combi ® (Millipore #1.00778.0001; Merck, Darmstadt, Germany), which offers a Rose Bengal Agar, were used according to the manufacturer’s instructions. The strain of F. solani used in the following experiments stemmed from these samples. 2.5. Examination of Residue Samples Residue samples were either examined directly by conventional light microscopy, or, if opaque and/or semi-solid, stained with Calcofluor-White (Merck #18909, Darmstadt, Germany) and examined by fluorescence microscopy (Olympus BX43 equipped with a reflected fluorescence system; Olympus Europa, Hamburg, Germany). Calcofluor-White is a fluorescent blue dye which binds to 1,3 and 1,4-beta-polysaccharides . 2.6. Zone-of-Inhibition Tests Zone-of-inhibition tests were performed as described . Briefly, isolated fungal samples were dissolved in 1 mL of 0.9% NaCl and 200 µL of the resulting mixture evenly spread on an SDA. A small Whatman filter paper (ø 2 mm) was placed into the center of the plate and 1 µL of the undiluted fungicide subsequently added to the filter disk. Plates were incubated at 30 °C for 4 days. 2.7. Adaptation to Fungicides SDA plates were prepared as described above, but shortly before solidifying, different concentrations of NaPT (0, 250 ppm, 500 ppm) or BBIT (0, 50 ppm, 100 ppm) were added and evenly mixed in. Plates were left to cool completely before being cut into thirds and reassembled. At the beginning of the experiment, spores and mycelium parts of F. solani were suspended in 0.9% NaCl and spread on the starting third containing no fungicides and cultivated overnight at RT. Plates were evaluated after 1, 2 and 3 weeks. 2.8. Sporulation Assays Sporulation assays were performed based on the publication by Zhang et al. . As background media, 0.9% NaCl was used and buffered with 50 mM TAPS (Sigma #T5130; Merck, Darmstadt, Germany), and the pH was adjusted to 9.3 with NaOH. To start the assay, a round piece of SDA (ø 1.4 cm; 1.56 cm 2 ) overgrown with F. solani after an incubation period of 4 days was added to 50 mL of buffer in a 250 mL baffled Erlenmeyer glass flask (SCHOTT DURAN, Mainz, Germany) and incubated on a shaker at 80 rpm at RT for 66 to 72 h. The resulting spore-containing solution was subsequently decanted, vortexed and distributed into 15 mL centrifuge tubes (Corning; #430791; Reynosa, Mexico). Fungicides or MWFs, alone or in combination at the indicated concentrations were subsequently added. The end volume was 5 mL. 2.9. Stability Assays For long-term experiments, MWFs were premixed in TAPS-buffered saline at 5% ( w / w ), and 50 mL was added to 250 mL baffled Erlenmeyer flasks before adding the 1.56 cm 2 piece of fungus-overgrown SDA. To give the fungus a chance to survive and develop, nutrients in the form of tryptone soy broth (Oxoid #CM131; Thermo Fisher, Pratteln, Switzerland) were added in 0.1-fold concentration right at the start. 2.10. Quantification of Spores and Viability Assays For assays without MWFs, 200 µL aliquots were removed and directly stained with propidium iodide (PI) and SYTO9 (LIVE/DEAD™ BacLight™ Bacterial Viability and Counting Kit; Invitrogen L34856; Thermo Fisher, Pratteln, Switzerland) as described by Vanhauteghem et al. . Analysis was subsequently performed on a CytoFLEX S flow cytometer (Beckman Coulter International S.A., Nyon, Switzerland). Samples containing MWFs had to be cleaned by centrifugation prior to analysis; 200 µL was added to 1 mL of 5.25% Nycodenz ® (Serumwerk Bernburg #18003; Bernburg, Germany) with TAPS-buffered saline into 2 mL centrifuge tubes (Eppendorf #0030 123.344; Hamburg, Germany), mixed by vortexing and centrifuged at 10,000× g for 10 min. at 4 °C. The supernatant was removed by decantation and left to drain for a few minutes. Leftover MWF sticking to the sidewalls of the centrifugation tubes was removed with sterile cotton swaps before the pellet was re-dissolved in 200 µL TAPS-buffered saline, stained and analyzed as described above. 2.11. Volatilization of Fungicides-Assays SDA plates were inoculated with four 20 µL drops of F. solani dissolved in 0.9% NaCl and incubated overnight at RT. The next day, 300 mL of MWF, mixed in with sterile-filtered tap water at 5% ( w / w ), was supplied to 600 mL sterile glass beakers (VWR International #213-1126, Dietikon, Switzerland) containing a sterile stirring bar. Then, the pre-incubated SD was added upside-down to the beaker and secured to be air-tight using Parafilm “M” (Amcor PM996, Zürich, Switzerland). The beakers were subsequently incubated with constant stirring on a heated, magnetic stirrer (30 °C, 200 rpm) for one week.
Via our extensive customer service network, we had worldwide access to MWF and residue samples from end users, both our own and external customers.
For all experiments, technical standard fungicides from industrial suppliers were used; the concentration indicated refers to a typical dose in a freshly prepared, 5% ( w / w ) in-use MWF, i.e., NaPT (Acticide ® LV 508; Thor GmbH, Speyer, Germany; Active ingredient content: 40%) at a concentration of 250 ppm; BBIT (Densil ® DN; Arch Biocides, Atlanta GA, USA; Active ingredient content: 100%) at a concentration of 50 ppm; and OPP (PREVENTOL ® O extra; Lanxess Deutschland GmbH, Leverkusen, Germany; active ingredient content: 99.5%) at a concentration of 500 ppm.
Most of the experiments shown were carried out with two experimental MWFs based on mineral oil that share about 50% of the ingredients (MWF A and B). Both concentrates were prepared available either with or without both NaPT and BBIT; the fungicide concentration was as described above when diluted at 5% ( w / w ).
Isolation of fungal species was based on cultivation-dependent methods such as the standard heterotrophic plate count method (HPC) or the use of dip slides. HPC was performed on Sabouraud dextrose–agar (SDA) prepared in-house (Oxoid #CM041R; Thermo Fisher, Prattlen, Switzerland) at 65 g L −1 . A total of 50 μL of either undiluted or 1:100 diluted 0.9% NaCl was plated using an Eddy Jet 2 system (IUL, Barcelona, Spain) in logarithmic mode. The plates were subsequently incubated at 30 °C for a minimum of 72 h before analysis. Fungal species were identified by Maldi-TOF MS analysis of protein patterns at Mabritec AG (Riehen, Switzerland). As dip slides, Cult Dip combi ® (Millipore #1.00778.0001; Merck, Darmstadt, Germany), which offers a Rose Bengal Agar, were used according to the manufacturer’s instructions. The strain of F. solani used in the following experiments stemmed from these samples.
Residue samples were either examined directly by conventional light microscopy, or, if opaque and/or semi-solid, stained with Calcofluor-White (Merck #18909, Darmstadt, Germany) and examined by fluorescence microscopy (Olympus BX43 equipped with a reflected fluorescence system; Olympus Europa, Hamburg, Germany). Calcofluor-White is a fluorescent blue dye which binds to 1,3 and 1,4-beta-polysaccharides .
Zone-of-inhibition tests were performed as described . Briefly, isolated fungal samples were dissolved in 1 mL of 0.9% NaCl and 200 µL of the resulting mixture evenly spread on an SDA. A small Whatman filter paper (ø 2 mm) was placed into the center of the plate and 1 µL of the undiluted fungicide subsequently added to the filter disk. Plates were incubated at 30 °C for 4 days.
SDA plates were prepared as described above, but shortly before solidifying, different concentrations of NaPT (0, 250 ppm, 500 ppm) or BBIT (0, 50 ppm, 100 ppm) were added and evenly mixed in. Plates were left to cool completely before being cut into thirds and reassembled. At the beginning of the experiment, spores and mycelium parts of F. solani were suspended in 0.9% NaCl and spread on the starting third containing no fungicides and cultivated overnight at RT. Plates were evaluated after 1, 2 and 3 weeks.
Sporulation assays were performed based on the publication by Zhang et al. . As background media, 0.9% NaCl was used and buffered with 50 mM TAPS (Sigma #T5130; Merck, Darmstadt, Germany), and the pH was adjusted to 9.3 with NaOH. To start the assay, a round piece of SDA (ø 1.4 cm; 1.56 cm 2 ) overgrown with F. solani after an incubation period of 4 days was added to 50 mL of buffer in a 250 mL baffled Erlenmeyer glass flask (SCHOTT DURAN, Mainz, Germany) and incubated on a shaker at 80 rpm at RT for 66 to 72 h. The resulting spore-containing solution was subsequently decanted, vortexed and distributed into 15 mL centrifuge tubes (Corning; #430791; Reynosa, Mexico). Fungicides or MWFs, alone or in combination at the indicated concentrations were subsequently added. The end volume was 5 mL.
For long-term experiments, MWFs were premixed in TAPS-buffered saline at 5% ( w / w ), and 50 mL was added to 250 mL baffled Erlenmeyer flasks before adding the 1.56 cm 2 piece of fungus-overgrown SDA. To give the fungus a chance to survive and develop, nutrients in the form of tryptone soy broth (Oxoid #CM131; Thermo Fisher, Pratteln, Switzerland) were added in 0.1-fold concentration right at the start.
For assays without MWFs, 200 µL aliquots were removed and directly stained with propidium iodide (PI) and SYTO9 (LIVE/DEAD™ BacLight™ Bacterial Viability and Counting Kit; Invitrogen L34856; Thermo Fisher, Pratteln, Switzerland) as described by Vanhauteghem et al. . Analysis was subsequently performed on a CytoFLEX S flow cytometer (Beckman Coulter International S.A., Nyon, Switzerland). Samples containing MWFs had to be cleaned by centrifugation prior to analysis; 200 µL was added to 1 mL of 5.25% Nycodenz ® (Serumwerk Bernburg #18003; Bernburg, Germany) with TAPS-buffered saline into 2 mL centrifuge tubes (Eppendorf #0030 123.344; Hamburg, Germany), mixed by vortexing and centrifuged at 10,000× g for 10 min. at 4 °C. The supernatant was removed by decantation and left to drain for a few minutes. Leftover MWF sticking to the sidewalls of the centrifugation tubes was removed with sterile cotton swaps before the pellet was re-dissolved in 200 µL TAPS-buffered saline, stained and analyzed as described above.
SDA plates were inoculated with four 20 µL drops of F. solani dissolved in 0.9% NaCl and incubated overnight at RT. The next day, 300 mL of MWF, mixed in with sterile-filtered tap water at 5% ( w / w ), was supplied to 600 mL sterile glass beakers (VWR International #213-1126, Dietikon, Switzerland) containing a sterile stirring bar. Then, the pre-incubated SD was added upside-down to the beaker and secured to be air-tight using Parafilm “M” (Amcor PM996, Zürich, Switzerland). The beakers were subsequently incubated with constant stirring on a heated, magnetic stirrer (30 °C, 200 rpm) for one week.
3.1. Fungal Occurrence in MWFs and Machining Systems 3.1.1. In MWFs In the period from 2014 to 2023, we analyzed a total of 48,695 liquid MWF samples for the presence of fungi using cultivation-dependent methods. Samples included mineral oil- or vegetable oil-based MWFs, synthetic and semisynthetic products from end users around the world. In 5.6%, or a total of 2746 samples, we detected fungi either in the form of spores, hyphae or yeast cells, ranging from a few cells to thousands per milliliter. The frequency of detection did not change significantly over these 10 years , indicating that the presence of fungi remained constant despite a substantial development in MWF technology . What had changed, however, were the circumstances under which fungi were detected; ten years ago, detection was positive if the pH value was well below the recommended value for the MWF in question. In recent years, detection was also positive even if the pH value was still within the recommended range. 3.1.2. In Residue Samples We only received a total of 417 residue samples for analysis from 2014 to 2023, as it seemed unreasonable for customers to send them in. On average, more than 50% of all samples analyzed contained yeast cells, hyphae, and/or spores . However, we are aware that many deposits were not recognized as fungi or that they were hidden: Most metalworking machines are not designed to allow sufficient examination of their surfaces. Often, only the tank and the inside of the machining area are freely accessible, but most surfaces are only accessible after dismantling the machine enclosure. For this reason, fungal deposits were only detected and sent in if there was already a strong suspicion of colonization and even then, many residues were simply photographed and subsequently disposed of. However, photographs do not allow an accurate identification . Still, as the reasons for sampling from customers or field staff have hardly changed over these years, we assume that the frequency of fungal contamination remains constant. 3.1.3. Detected Species As fungi are generally undesirable in MWFs, we did not carry out species identification regularly. However, before performing resistance tests (see ), we determined the species upstream using Maldi-TOF MS (Mabritec AG, Riehen, Switzerland). The overwhelming majority of mold samples identified in this way belonged to the genus Fusarium , which could mainly be assigned to either the F. solani or the F. oxysporum species complex. As these species are ubiquitous in nature, it is reasonable to assume that spores from the respective environment were transported by air into machine interiors and settled under suitable conditions, as shown for many indoor fungi . We can only speculate on the origin of the isolated yeast samples. Diutina neorugosa and D. rugosa (formerly Candida neorugosa and C. rugosa ) belong to a complex of species that represent about 0.2% of all clinical isolates , whereas C. tropicalis is one of the most frequently involved non-albicans Candida pathogens . Apart from that, little is known, but it has been reported that these species are also ubiquitous in the environment . Yarrowia lipolytica is also thought to be widespread in nature but is best-known as a frequently used, strictly aerobic, non-pathogenic producer of biodiesel and other biotechnological products. 3.2. Escape from Fungicide-Toxicity in MWFs Most MWFs contain fungicides to prevent contamination by fungi. Interestingly, the majority of isolated or detected species in this study originated from MWFs formulated with one or even two fungicides, most often BBIT and/or NaPT. With zone-of-inhibition tests, we investigated the efficacy of these fungicides on fungal growth. 3.2.1. Resistance Formation in In-Use Samples A total of 103 samples were tested for susceptibility to BBIT or NaPT, whether they had been previously exposed to these substances by means of the MWF (as ingredient(s) of its base formulation) or not. Fungi originating from MWFs preserved by NaPT showed a reduced susceptibility when re-exposed to NaPT. This is reflected in a significantly reduced zone of inhibition. Interestingly, such an effect was not observed with BBIT ( a). 3.2.2. Adaptation to NaPT and BBIT In the next step, we tested if a naïve laboratory strain of F. solani can adapt to NaPT or BBIT. To do this, we prepared SDA plates with different amounts of fungicide, which we then assembled in a way that the fungus was not exposed to any fungicide in the first third before it was subsequently exposed to a single and finally a double dose ( b). To give the fungus a head-start, it was spread on the starting third beforehand and cultivated overnight at RT. Only then were the agar thirds with the fungicides added. While F. solani in the control (without fungicides) and the BBIT-experiments had moved across the entire plate within 3 weeks with no discernible differences, only very modest growth was observed in the presence of NaPT. NaPT was apparently potent and even diffused into the first third initially containing no active ingredient. 3.3. Impact of Fungicides on Sporulation 3.3.1. Quantities of Released Spores As fungal growth and sporulation are distinct processes, we aimed to investigate the impact of fungicides and MWFs specifically on sporulation. In a recent review, the authors reported that the sporulation of Fusarium spp. can be inhibited or induced by a variety of chemical molecules , while others described that highest sporulation quantities in liquids are to be expected in saline at a pH between 9 and 10 . Thus, in a simple experiment, fungicides were dissolved in TAPS-buffered saline at a pH of 9.3; the concentration used basically corresponded to that being present in many commercial in-use MWFs at 5% ( w / w ). To start the experiment, a 1.56 cm 2 piece of SDA completely overgrown by F. solani was added to it. Between 5.7 and 6.6 log 10 of spores were released by this small piece. As shown in , none of the fungicides had any significant impact on sporulation as counted by flow cytometry, either alone or in combination. We also tested the influence of MWFs on sporulation and conducted experiments with MWF A and B in the absence and presence of added fungicides; again, sporulation remained unchanged. 3.3.2. Effect of Fungicides on Spore Viability To assess the viability of the released spores, we removed aliquots from the control experiment before adding fungicides NaPT, BBIT or a combination of both. We then assessed spore viability by flow cytometry using Syto9 and PI as described by Vanhauteghem et al. after incubation for 30 min, 4 h and 24 h. Whereas BBIT reduced spore viability gradually and efficiently ( a), NaPT had almost no impact and was largely indistinguishable from the control. A combination of both fungicides was marginally less effective than BBIT alone. This suggests that NaPT affects the fungus itself (see ) but not the spores once released. A simple test supported this assumption; NaPT or BBIT were added to the buffer system before adding the overgrown piece of Sabouraud agar, and they were incubated at RT with shaking (80 rpm). After 24 h, the viability of the spores was assessed as described, showing a clear but incomplete reduction in the presence of NaPT ( b). 3.3.3. Effect of MWFs on Spore Viability To test the efficacy of NaPT and BBIT in conjunction with MWFs, we repeated viability tests with MWF A and B. These experiments showed that MWF A deactivated virtually all spores after 24 h, while MWF B did not . As expected, this measurable effect was enhanced by fungicides in both cases; MWF A thus deactivated more than 40% of the spores after only 4 h, while the value for MWF B was around 25%. 3.3.4. Effect of Delayed Spore Inactivation Although the result after 24 h was identical in terms of spore inactivation for MWF A and B (both with added fungicides), the delay in the process could lead to problems in practice, particularly when these spores are splashed onto surfaces where they only occasionally remain in contact with the MWF. To simulate such a situation, we prepared 50 mL emulsions of MWF A and B, respectively, including fungicides at 5% ( w / w ) in 250 mL baffled Erlenmeyer flasks and added a 1.56 cm 2 piece of SDA, overgrown with F. solani. After one week (80 rpm, RT), MWF B began to exhibit instabilities, as evidenced by a creamy layer forming on the surface ( a). By three weeks, a visible deposited layer appeared along the contact line ( b). In contrast, MWF A showed only emulsion components, few spores, and detached hyphae. MWF B, however, allowed isolated spores and hyphae to survive, leading to the formation of detectable mycelium at and above the contact line ( c). In real-world systems, turbulence is unavoidable and transports spores and hyphae to surfaces where they are only partially washed away. This results in temporary and incomplete toxic effects, thereby enabling fungal survival. For this experiment, the MWFs were diluted in TAPS-buffered saline, and pH changes remained minimal. 3.4. Impact of Fungicide Volatiles As mentioned, fungi are mainly found on machine surfaces outside the MWF and only temporarily in direct contact with the fluid. Typical locations are, for example, in the pump area where the fluid level remains practically constant. Such sections are usually covered to minimize evaporation into the production hall, which leads to high local humidity in the section itself, promoting fungal growth . Accordingly, such areas cannot be decontaminated by simply adding fungicides, unless these are effective in the vapor phase. We therefore developed a laboratory-scale test to examine this possibility ( a). After one week, there were no significant effects on fungal growth on the SDA plate, regardless of the chemical composition or the presence of NaPT and/or BBIT. However, we conducted additional experiments using OPP at a concentration of 500 ppm, which consistently showed a significant and detrimental impact on the development of F. solani on the SDA plate above, independent of the chemical background of the MWFs used. b illustrates experiments with MWF A both without fungicides and with NaPT, BBIT, and OPP. In this series of experiments, we opted not to dilute the MWF in TAPS-buffered saline, using sterile filtered tap water instead to better simulate real conditions. Consequently, the pH varied throughout the experiment, dropping by as much as 1 unit in all fluids where F. solani grew extensively, indicating a negative influence of fungal presence on MWF chemical stability. In contrast, the experiments involving OPP showed minimal changes in pH, remaining within 0.1 units.
3.1.1. In MWFs In the period from 2014 to 2023, we analyzed a total of 48,695 liquid MWF samples for the presence of fungi using cultivation-dependent methods. Samples included mineral oil- or vegetable oil-based MWFs, synthetic and semisynthetic products from end users around the world. In 5.6%, or a total of 2746 samples, we detected fungi either in the form of spores, hyphae or yeast cells, ranging from a few cells to thousands per milliliter. The frequency of detection did not change significantly over these 10 years , indicating that the presence of fungi remained constant despite a substantial development in MWF technology . What had changed, however, were the circumstances under which fungi were detected; ten years ago, detection was positive if the pH value was well below the recommended value for the MWF in question. In recent years, detection was also positive even if the pH value was still within the recommended range. 3.1.2. In Residue Samples We only received a total of 417 residue samples for analysis from 2014 to 2023, as it seemed unreasonable for customers to send them in. On average, more than 50% of all samples analyzed contained yeast cells, hyphae, and/or spores . However, we are aware that many deposits were not recognized as fungi or that they were hidden: Most metalworking machines are not designed to allow sufficient examination of their surfaces. Often, only the tank and the inside of the machining area are freely accessible, but most surfaces are only accessible after dismantling the machine enclosure. For this reason, fungal deposits were only detected and sent in if there was already a strong suspicion of colonization and even then, many residues were simply photographed and subsequently disposed of. However, photographs do not allow an accurate identification . Still, as the reasons for sampling from customers or field staff have hardly changed over these years, we assume that the frequency of fungal contamination remains constant. 3.1.3. Detected Species As fungi are generally undesirable in MWFs, we did not carry out species identification regularly. However, before performing resistance tests (see ), we determined the species upstream using Maldi-TOF MS (Mabritec AG, Riehen, Switzerland). The overwhelming majority of mold samples identified in this way belonged to the genus Fusarium , which could mainly be assigned to either the F. solani or the F. oxysporum species complex. As these species are ubiquitous in nature, it is reasonable to assume that spores from the respective environment were transported by air into machine interiors and settled under suitable conditions, as shown for many indoor fungi . We can only speculate on the origin of the isolated yeast samples. Diutina neorugosa and D. rugosa (formerly Candida neorugosa and C. rugosa ) belong to a complex of species that represent about 0.2% of all clinical isolates , whereas C. tropicalis is one of the most frequently involved non-albicans Candida pathogens . Apart from that, little is known, but it has been reported that these species are also ubiquitous in the environment . Yarrowia lipolytica is also thought to be widespread in nature but is best-known as a frequently used, strictly aerobic, non-pathogenic producer of biodiesel and other biotechnological products.
In the period from 2014 to 2023, we analyzed a total of 48,695 liquid MWF samples for the presence of fungi using cultivation-dependent methods. Samples included mineral oil- or vegetable oil-based MWFs, synthetic and semisynthetic products from end users around the world. In 5.6%, or a total of 2746 samples, we detected fungi either in the form of spores, hyphae or yeast cells, ranging from a few cells to thousands per milliliter. The frequency of detection did not change significantly over these 10 years , indicating that the presence of fungi remained constant despite a substantial development in MWF technology . What had changed, however, were the circumstances under which fungi were detected; ten years ago, detection was positive if the pH value was well below the recommended value for the MWF in question. In recent years, detection was also positive even if the pH value was still within the recommended range.
We only received a total of 417 residue samples for analysis from 2014 to 2023, as it seemed unreasonable for customers to send them in. On average, more than 50% of all samples analyzed contained yeast cells, hyphae, and/or spores . However, we are aware that many deposits were not recognized as fungi or that they were hidden: Most metalworking machines are not designed to allow sufficient examination of their surfaces. Often, only the tank and the inside of the machining area are freely accessible, but most surfaces are only accessible after dismantling the machine enclosure. For this reason, fungal deposits were only detected and sent in if there was already a strong suspicion of colonization and even then, many residues were simply photographed and subsequently disposed of. However, photographs do not allow an accurate identification . Still, as the reasons for sampling from customers or field staff have hardly changed over these years, we assume that the frequency of fungal contamination remains constant.
As fungi are generally undesirable in MWFs, we did not carry out species identification regularly. However, before performing resistance tests (see ), we determined the species upstream using Maldi-TOF MS (Mabritec AG, Riehen, Switzerland). The overwhelming majority of mold samples identified in this way belonged to the genus Fusarium , which could mainly be assigned to either the F. solani or the F. oxysporum species complex. As these species are ubiquitous in nature, it is reasonable to assume that spores from the respective environment were transported by air into machine interiors and settled under suitable conditions, as shown for many indoor fungi . We can only speculate on the origin of the isolated yeast samples. Diutina neorugosa and D. rugosa (formerly Candida neorugosa and C. rugosa ) belong to a complex of species that represent about 0.2% of all clinical isolates , whereas C. tropicalis is one of the most frequently involved non-albicans Candida pathogens . Apart from that, little is known, but it has been reported that these species are also ubiquitous in the environment . Yarrowia lipolytica is also thought to be widespread in nature but is best-known as a frequently used, strictly aerobic, non-pathogenic producer of biodiesel and other biotechnological products.
Most MWFs contain fungicides to prevent contamination by fungi. Interestingly, the majority of isolated or detected species in this study originated from MWFs formulated with one or even two fungicides, most often BBIT and/or NaPT. With zone-of-inhibition tests, we investigated the efficacy of these fungicides on fungal growth. 3.2.1. Resistance Formation in In-Use Samples A total of 103 samples were tested for susceptibility to BBIT or NaPT, whether they had been previously exposed to these substances by means of the MWF (as ingredient(s) of its base formulation) or not. Fungi originating from MWFs preserved by NaPT showed a reduced susceptibility when re-exposed to NaPT. This is reflected in a significantly reduced zone of inhibition. Interestingly, such an effect was not observed with BBIT ( a). 3.2.2. Adaptation to NaPT and BBIT In the next step, we tested if a naïve laboratory strain of F. solani can adapt to NaPT or BBIT. To do this, we prepared SDA plates with different amounts of fungicide, which we then assembled in a way that the fungus was not exposed to any fungicide in the first third before it was subsequently exposed to a single and finally a double dose ( b). To give the fungus a head-start, it was spread on the starting third beforehand and cultivated overnight at RT. Only then were the agar thirds with the fungicides added. While F. solani in the control (without fungicides) and the BBIT-experiments had moved across the entire plate within 3 weeks with no discernible differences, only very modest growth was observed in the presence of NaPT. NaPT was apparently potent and even diffused into the first third initially containing no active ingredient.
A total of 103 samples were tested for susceptibility to BBIT or NaPT, whether they had been previously exposed to these substances by means of the MWF (as ingredient(s) of its base formulation) or not. Fungi originating from MWFs preserved by NaPT showed a reduced susceptibility when re-exposed to NaPT. This is reflected in a significantly reduced zone of inhibition. Interestingly, such an effect was not observed with BBIT ( a).
In the next step, we tested if a naïve laboratory strain of F. solani can adapt to NaPT or BBIT. To do this, we prepared SDA plates with different amounts of fungicide, which we then assembled in a way that the fungus was not exposed to any fungicide in the first third before it was subsequently exposed to a single and finally a double dose ( b). To give the fungus a head-start, it was spread on the starting third beforehand and cultivated overnight at RT. Only then were the agar thirds with the fungicides added. While F. solani in the control (without fungicides) and the BBIT-experiments had moved across the entire plate within 3 weeks with no discernible differences, only very modest growth was observed in the presence of NaPT. NaPT was apparently potent and even diffused into the first third initially containing no active ingredient.
3.3.1. Quantities of Released Spores As fungal growth and sporulation are distinct processes, we aimed to investigate the impact of fungicides and MWFs specifically on sporulation. In a recent review, the authors reported that the sporulation of Fusarium spp. can be inhibited or induced by a variety of chemical molecules , while others described that highest sporulation quantities in liquids are to be expected in saline at a pH between 9 and 10 . Thus, in a simple experiment, fungicides were dissolved in TAPS-buffered saline at a pH of 9.3; the concentration used basically corresponded to that being present in many commercial in-use MWFs at 5% ( w / w ). To start the experiment, a 1.56 cm 2 piece of SDA completely overgrown by F. solani was added to it. Between 5.7 and 6.6 log 10 of spores were released by this small piece. As shown in , none of the fungicides had any significant impact on sporulation as counted by flow cytometry, either alone or in combination. We also tested the influence of MWFs on sporulation and conducted experiments with MWF A and B in the absence and presence of added fungicides; again, sporulation remained unchanged. 3.3.2. Effect of Fungicides on Spore Viability To assess the viability of the released spores, we removed aliquots from the control experiment before adding fungicides NaPT, BBIT or a combination of both. We then assessed spore viability by flow cytometry using Syto9 and PI as described by Vanhauteghem et al. after incubation for 30 min, 4 h and 24 h. Whereas BBIT reduced spore viability gradually and efficiently ( a), NaPT had almost no impact and was largely indistinguishable from the control. A combination of both fungicides was marginally less effective than BBIT alone. This suggests that NaPT affects the fungus itself (see ) but not the spores once released. A simple test supported this assumption; NaPT or BBIT were added to the buffer system before adding the overgrown piece of Sabouraud agar, and they were incubated at RT with shaking (80 rpm). After 24 h, the viability of the spores was assessed as described, showing a clear but incomplete reduction in the presence of NaPT ( b). 3.3.3. Effect of MWFs on Spore Viability To test the efficacy of NaPT and BBIT in conjunction with MWFs, we repeated viability tests with MWF A and B. These experiments showed that MWF A deactivated virtually all spores after 24 h, while MWF B did not . As expected, this measurable effect was enhanced by fungicides in both cases; MWF A thus deactivated more than 40% of the spores after only 4 h, while the value for MWF B was around 25%. 3.3.4. Effect of Delayed Spore Inactivation Although the result after 24 h was identical in terms of spore inactivation for MWF A and B (both with added fungicides), the delay in the process could lead to problems in practice, particularly when these spores are splashed onto surfaces where they only occasionally remain in contact with the MWF. To simulate such a situation, we prepared 50 mL emulsions of MWF A and B, respectively, including fungicides at 5% ( w / w ) in 250 mL baffled Erlenmeyer flasks and added a 1.56 cm 2 piece of SDA, overgrown with F. solani. After one week (80 rpm, RT), MWF B began to exhibit instabilities, as evidenced by a creamy layer forming on the surface ( a). By three weeks, a visible deposited layer appeared along the contact line ( b). In contrast, MWF A showed only emulsion components, few spores, and detached hyphae. MWF B, however, allowed isolated spores and hyphae to survive, leading to the formation of detectable mycelium at and above the contact line ( c). In real-world systems, turbulence is unavoidable and transports spores and hyphae to surfaces where they are only partially washed away. This results in temporary and incomplete toxic effects, thereby enabling fungal survival. For this experiment, the MWFs were diluted in TAPS-buffered saline, and pH changes remained minimal.
As fungal growth and sporulation are distinct processes, we aimed to investigate the impact of fungicides and MWFs specifically on sporulation. In a recent review, the authors reported that the sporulation of Fusarium spp. can be inhibited or induced by a variety of chemical molecules , while others described that highest sporulation quantities in liquids are to be expected in saline at a pH between 9 and 10 . Thus, in a simple experiment, fungicides were dissolved in TAPS-buffered saline at a pH of 9.3; the concentration used basically corresponded to that being present in many commercial in-use MWFs at 5% ( w / w ). To start the experiment, a 1.56 cm 2 piece of SDA completely overgrown by F. solani was added to it. Between 5.7 and 6.6 log 10 of spores were released by this small piece. As shown in , none of the fungicides had any significant impact on sporulation as counted by flow cytometry, either alone or in combination. We also tested the influence of MWFs on sporulation and conducted experiments with MWF A and B in the absence and presence of added fungicides; again, sporulation remained unchanged.
To assess the viability of the released spores, we removed aliquots from the control experiment before adding fungicides NaPT, BBIT or a combination of both. We then assessed spore viability by flow cytometry using Syto9 and PI as described by Vanhauteghem et al. after incubation for 30 min, 4 h and 24 h. Whereas BBIT reduced spore viability gradually and efficiently ( a), NaPT had almost no impact and was largely indistinguishable from the control. A combination of both fungicides was marginally less effective than BBIT alone. This suggests that NaPT affects the fungus itself (see ) but not the spores once released. A simple test supported this assumption; NaPT or BBIT were added to the buffer system before adding the overgrown piece of Sabouraud agar, and they were incubated at RT with shaking (80 rpm). After 24 h, the viability of the spores was assessed as described, showing a clear but incomplete reduction in the presence of NaPT ( b).
To test the efficacy of NaPT and BBIT in conjunction with MWFs, we repeated viability tests with MWF A and B. These experiments showed that MWF A deactivated virtually all spores after 24 h, while MWF B did not . As expected, this measurable effect was enhanced by fungicides in both cases; MWF A thus deactivated more than 40% of the spores after only 4 h, while the value for MWF B was around 25%.
Although the result after 24 h was identical in terms of spore inactivation for MWF A and B (both with added fungicides), the delay in the process could lead to problems in practice, particularly when these spores are splashed onto surfaces where they only occasionally remain in contact with the MWF. To simulate such a situation, we prepared 50 mL emulsions of MWF A and B, respectively, including fungicides at 5% ( w / w ) in 250 mL baffled Erlenmeyer flasks and added a 1.56 cm 2 piece of SDA, overgrown with F. solani. After one week (80 rpm, RT), MWF B began to exhibit instabilities, as evidenced by a creamy layer forming on the surface ( a). By three weeks, a visible deposited layer appeared along the contact line ( b). In contrast, MWF A showed only emulsion components, few spores, and detached hyphae. MWF B, however, allowed isolated spores and hyphae to survive, leading to the formation of detectable mycelium at and above the contact line ( c). In real-world systems, turbulence is unavoidable and transports spores and hyphae to surfaces where they are only partially washed away. This results in temporary and incomplete toxic effects, thereby enabling fungal survival. For this experiment, the MWFs were diluted in TAPS-buffered saline, and pH changes remained minimal.
As mentioned, fungi are mainly found on machine surfaces outside the MWF and only temporarily in direct contact with the fluid. Typical locations are, for example, in the pump area where the fluid level remains practically constant. Such sections are usually covered to minimize evaporation into the production hall, which leads to high local humidity in the section itself, promoting fungal growth . Accordingly, such areas cannot be decontaminated by simply adding fungicides, unless these are effective in the vapor phase. We therefore developed a laboratory-scale test to examine this possibility ( a). After one week, there were no significant effects on fungal growth on the SDA plate, regardless of the chemical composition or the presence of NaPT and/or BBIT. However, we conducted additional experiments using OPP at a concentration of 500 ppm, which consistently showed a significant and detrimental impact on the development of F. solani on the SDA plate above, independent of the chemical background of the MWFs used. b illustrates experiments with MWF A both without fungicides and with NaPT, BBIT, and OPP. In this series of experiments, we opted not to dilute the MWF in TAPS-buffered saline, using sterile filtered tap water instead to better simulate real conditions. Consequently, the pH varied throughout the experiment, dropping by as much as 1 unit in all fluids where F. solani grew extensively, indicating a negative influence of fungal presence on MWF chemical stability. In contrast, the experiments involving OPP showed minimal changes in pH, remaining within 0.1 units.
Examining the occurrence data presented in , we might conclude that fungi are only sporadically present in metalworking fluids (MWFs) and their systems. However, this conclusion is mitigated by two factors. (i) For fungi to be detected, spores or yeast cells must exist in substantial quantities. This is particularly challenging when using cultivation-dependent methods, which have remained standard in the industry. Even if millions of spores are released into the environment, detecting them becomes difficult; spores released into the air are essentially lost, while those in liquid are often heavily diluted and affected by the MWF’s chemistry. (ii) Mold or yeast colonies growing on surfaces frequently go unnoticed, as they can hide within the intricate structures of machines. These organisms only become apparent in obvious locations, such as filters or open tanks, or when their growth disrupts production. Even then, many potential samples are never analyzed because they are simply removed and discarded. Consequently, we think that fungal contamination is widespread in the metalworking industry, without significant concerns. Publications on fungal contaminations in MWFs are surprisingly scarce and links to disease even more so. This contrasts with the vast literature linking bacteria to occupational diseases . In a review, a connection between microbial colonization of MWFs and the symptoms of illness was presented , but molds or yeasts were not mentioned in this context. One possible reason might be a simple one: the species detected ( Fusarium spp., Diutina spp., Candida spp. and Yarrowia spp.) are reported to be ubiquitous in nature as with many other opportunistic pathogens , and constant contact may desensitize most human beings. In any case, it seems difficult to link the experience of symptoms to an exact cause per se . This applies certainly to MWFs as they are made up of a complex mixture of different chemicals, mixed in with water, and used in a plethora of manufacturing processes . The reason that fungi are unwelcome in MWFs is thus mainly of a technical nature: the breakdown of MWF circulation due to blockages of pumps and filters . Alongside this reason are unpleasant odors, visible growth, and fear of the unknown as soft factors. Phenomena such as pH drops and unstable or destroyed emulsions occur and can be simulated in the laboratory but are extremely rare in the real world, as this requires the presence of enormous quantities of fungal material in relation to the tank size. But even if the biological material is in surplus, some MWFs are still able to withstand and remain technically sound. Yeasts, to our knowledge, are of minor concern as phenomena such as those just described could never be attributed to them. In laboratory trials, quantities of up to 10 6 mL −1 had negligible effects on the physico-chemical parameters in MWFs. The fact that technical problems do not escalate is often attributed to the fungicides contained as in-drum additives. The legal constraints on these chemicals, however, are increasing. Permitted concentrations in the concentrate, as well as in the in-use products, are already restricted, and it is expected that this will continue . Apart from these challenges, simple addition of these ingredients does not always offer an adequate solution; molds or yeasts in splash zones are only marginally impacted by the chemistry as they avoid direct contact. Moreover, the surrounding chemical composition will influence the time required for fungicidal effects to unfold. Another aspect is the availability of the compounds at any given moment. Although concentrates are formulated so that in-use MWFs contain enough fungicides for protection, large quantities of spores may overwhelm the defenses: the substances are used up, are dragged out, and impoverish with time. Additionally, as we have shown, fungi adapt rapidly to the conditions provided by metalworking fluids, enabling them to thrive even in the presence of fungicides. Furthermore, the effectiveness of these fungicides is limited; NaPT showed no measurable impact on spore viability, while BBIT had only a minor effect on the growth of the organism, and neither of their vapors appeared to inhibit fungal development. As our tests were conducted at concentrations that might be common in 5% emulsions (a low-end concentration frequently used in industry), it will be interesting to follow future developments as approved fungicide concentrations continue to decrease. Preliminary trials seem to indicate that halving the concentration already cancels out effectiveness against Fusarium spores.
In the future, the industrial sector needs to adjust by creating MWFs with chemical constructions that deactivate spores and inhibit the growth of molds (and yeasts) without the help of fungicides. Creating an environment that restricts the availability of food sources could be one way to go, as the other important features for fungal growth, i.e., temperature, humidity and pH, cannot be influenced or changed. Indeed, we argue that the ecology created by the chemical composition of the concentrates diluted in water is of higher importance than the anti-fungal activity of single components.
|
Brazilian psychiatrists’ knowledge of and perceived confidence in eating disorder diagnosis and treatment recommendations | 517fbe87-af79-4bd0-b50d-37f8b0e16412 | 11559849 | Psychiatry[mh] | Eating disorders (ED) are a group of psychiatric disorders with very specific physical, psychological, and social characteristics. , The DSM-5 incorporated several changes in its chapter on “Feeding and Eating Disorders.” Disturbed eating behaviors generally arise due to a dysfunctional relationship with food and body image, leading to significant physical and mental impairment. ED are also an important cause of morbidity and mortality, comparable to severe psychiatric disorders such as drug use disorders. Health care providers often consider patients with ED complex and challenging. One explanation for this is that EDs are perceived to be “on purpose.” The ego-syntonic nature of the typical symptom presentation and a common lack of motivation and adherence among patients are considered important barriers to successful treatment. Consequently, feelings of frustration, irritability, and aversion can be identified in the care team, which hinders a good therapeutic alliance and further affects treatment prognosis. Other factors related to the professionals themselves can compromise treatment response, such as a lack of confidence and poor training in screening, diagnosing, and treating people with EDs, the limited number of specialized services, and a lack of financial resources and ED specialists to plan effective and widely accessible interventions. - Regarding the lack of specific instructions medical students receive about ED, recent studies have shown that a basic educational program on ED can reduce prejudice towards people with these disorders and improve their detection and treatment in primary care. , However, in most countries, ED education and training is limited in both undergraduate education and graduate education/medical residency. , A study found that only about one-third of doctors in the United Kingdom received 4 or more hours of classes specifically on ED. ED data are even scarcer in Brazil. As far as we know, only a few specialized ED services exist and most are in state capitals in the southeast region, usually in university hospitals. Thus, public demand is not being met. Considering the importance of the current situation and the main obstacles to ED care in Brazil, this study investigated Brazilian psychiatrists’ knowledge of and perceived confidence in diagnosing and treating patients with ED.
Study design and participants This cross-sectional study was developed by the Eating Disorders Commission ( Comissão de Transtornos Alimentares ) of the Brazilian Psychiatric Association, the official nationwide organization of Brazilian psychiatrists. All affiliated Brazilian psychiatrists were invited to participate in this study through an e-mail with a link to an online questionnaire. Between June and August 2021, 259 psychiatrists filled out the questionnaire after providing written informed consent (the study was previously approved by Universidade Federal do Rio de Janeiro research ethics committee). We estimate that around 6,000 psychiatrists received the invitation, and response rate was approximately 4.32%. Procedures After receiving authorization from Dr. William Rhys Jones, we developed a questionnaire to assess Brazilian psychiatrists’ knowledge of ED based on his work. The adapted form was uploaded to Google Forms. A pre-test was then performed in which six psychiatrists (non-specialists in ED) responded to the initial instrument and made comments/suggestions. After reviewing and updating the questionnaire, the instrument was finalized and data collection proceeded. Through advertising, members of the Brazilian Psychiatric Association had access to the research questionnaire link and could respond to the online form. Instrument and variables The final questionnaire used was divided into three sections. The first was on sociodemographic information: sex, age, location, public or private practice, length of experience, and academic degrees (Brazilian Psychiatric Association specialist title, graduate degree/residency in psychiatry, current academic activity). The second section assessed ED knowledge through six theoretical multiple-choice questions: three on diagnostic criteria (one each on anorexia nervosa [AN], bulimia nervosa [BN], and binge ED [BED]) and three on treatment recommendations (one for each condition). The level of ED knowledge was assessed according to the number of correct answers through categorical analysis. The third section was on the participants’ training during undergraduate and graduate education, residency, and continuing education, as well as satisfaction with their knowledge of ED and their confidence in diagnosing and treating ED in clinical practice. Statistical analysis The data were analyzed anonymously in Stata 16.0. Descriptive statistics were used to assess the sample’s characteristics, ED knowledge level, and opinions. The sample’s characteristics and ED knowledge level are shown in frequency tables that include the number and percentage of correct responses for each item. The respondents’ opinions are summarized in three bar graphs (one for each question) based on a 5-point Likert scale.
This cross-sectional study was developed by the Eating Disorders Commission ( Comissão de Transtornos Alimentares ) of the Brazilian Psychiatric Association, the official nationwide organization of Brazilian psychiatrists. All affiliated Brazilian psychiatrists were invited to participate in this study through an e-mail with a link to an online questionnaire. Between June and August 2021, 259 psychiatrists filled out the questionnaire after providing written informed consent (the study was previously approved by Universidade Federal do Rio de Janeiro research ethics committee). We estimate that around 6,000 psychiatrists received the invitation, and response rate was approximately 4.32%.
After receiving authorization from Dr. William Rhys Jones, we developed a questionnaire to assess Brazilian psychiatrists’ knowledge of ED based on his work. The adapted form was uploaded to Google Forms. A pre-test was then performed in which six psychiatrists (non-specialists in ED) responded to the initial instrument and made comments/suggestions. After reviewing and updating the questionnaire, the instrument was finalized and data collection proceeded. Through advertising, members of the Brazilian Psychiatric Association had access to the research questionnaire link and could respond to the online form.
The final questionnaire used was divided into three sections. The first was on sociodemographic information: sex, age, location, public or private practice, length of experience, and academic degrees (Brazilian Psychiatric Association specialist title, graduate degree/residency in psychiatry, current academic activity). The second section assessed ED knowledge through six theoretical multiple-choice questions: three on diagnostic criteria (one each on anorexia nervosa [AN], bulimia nervosa [BN], and binge ED [BED]) and three on treatment recommendations (one for each condition). The level of ED knowledge was assessed according to the number of correct answers through categorical analysis. The third section was on the participants’ training during undergraduate and graduate education, residency, and continuing education, as well as satisfaction with their knowledge of ED and their confidence in diagnosing and treating ED in clinical practice.
The data were analyzed anonymously in Stata 16.0. Descriptive statistics were used to assess the sample’s characteristics, ED knowledge level, and opinions. The sample’s characteristics and ED knowledge level are shown in frequency tables that include the number and percentage of correct responses for each item. The respondents’ opinions are summarized in three bar graphs (one for each question) based on a 5-point Likert scale.
Sample A total of 259 psychiatrists responded to the questionnaire. The sample’s sociodemographic and academic characteristics are summarized in . The sample, which was 65.64% women, had a mean age of 42.86 years (range: 25-75). Most participants were from southeastern Brazil (56.37%), worked predominantly in private practice (59.85%), had less than 10 years’ experience in psychiatric practice (51.74%), completed a residency in psychiatry (62.93%), and were involved in academic activity (57.53%). Eating disorder knowledge level among Brazilian psychiatrists Based on the number and percentage of correct answers, the respondents’ knowledge level of diagnostic criteria and treatment recommendations for ED are presented in . Each questionnaire item had six response options, and participants were instructed to mark all that were correct. Responses were considered correct only if all correct options were marked and all incorrect ones were unmarked. Although almost one-third of the sample correctly answered each question about ED diagnostic criteria (33.21% for AN; 29.73% for BN; 38.22% for BED), only 12.74% correctly answered all three. Regarding ED treatment, 2.7% answered correctly for AN, 20.08% answered correctly for BN, and 20.08% answered correctly for BED, with none correctly answering all three items. Knowledge about diagnostic criteria for eating disorders The instrument included three questions about ED diagnostic criteria according to the DSM-5: one each on AN, BN, and BED. “Intense fear of getting fat” was the top-rated item for AN (89.2%), followed by “Excessive preoccupation with weight and body shape” (69.5%), and “Significantly low body weight” (68.3%). “Amenorrhea” was mistakenly marked by 28.2% of the sample as a diagnostic criterion for AN. For BN, “Inappropriate compensatory behaviors for weight control” was correctly identified by 94.6% of participants, while “Binge eating episodes once a week for 3 months” was only marked by 52.5%. A total of 61% correctly identified “Excessive preoccupation with weight and body shape” for BN diagnosis, while 29.34% incorrectly identified “Binge eating episodes twice a week for 3 months.” Regarding BED, the items with the highest number of correct answers were “Intense suffering with binge eating episodes” (89.6%) and “Binge eating episodes once a week for 3 months” (56.4%). Nevertheless, “Binge eating episodes twice a week for 3 months” and “Inappropriate compensatory behaviors for weight control” were erroneously marked as correct by 35.4 and 16.6% respectively. Knowledge about treatment recommendations for eating disorders There were three questions on ED treatment recommendations according to international guidelines. Regarding AN, 91.5% correctly identified “Cognitive behavioral therapy,” 47.1% correctly identified “Family therapy,” and 32.4% correctly identified “Atypical antipsychotics,” while 60.6% incorrectly identified “Selective serotonin reuptake inhibitors.” Regarding BN, 93.8% correctly identified “Cognitive behavioral therapy” and 83.8% correctly identified “Fluoxetine,” but only 47.5% correctly identified “Topiramate” as an accepted therapy. Regarding BED, 92.3% correctly identified “Cognitive behavioral therapy,” 56.8% “Lisdexamfetamine,” 65.3% “Fluoxetine,” and 69.1% “Topiramate.” Training about eating disorders Only 15.1% of the participants received specialized training in ED during medical school, and 59.8% underwent a residency training program or graduate studies in psychiatry. A total of 58.7% mentioned complementary ED training (courses, conferences, specific events). A total of 8.89% of the respondents felt satisfied with their ED training, 50.97% felt confident in diagnosing ED, and 37.07% felt confident in treating ED patients ( ).
A total of 259 psychiatrists responded to the questionnaire. The sample’s sociodemographic and academic characteristics are summarized in . The sample, which was 65.64% women, had a mean age of 42.86 years (range: 25-75). Most participants were from southeastern Brazil (56.37%), worked predominantly in private practice (59.85%), had less than 10 years’ experience in psychiatric practice (51.74%), completed a residency in psychiatry (62.93%), and were involved in academic activity (57.53%).
Based on the number and percentage of correct answers, the respondents’ knowledge level of diagnostic criteria and treatment recommendations for ED are presented in . Each questionnaire item had six response options, and participants were instructed to mark all that were correct. Responses were considered correct only if all correct options were marked and all incorrect ones were unmarked. Although almost one-third of the sample correctly answered each question about ED diagnostic criteria (33.21% for AN; 29.73% for BN; 38.22% for BED), only 12.74% correctly answered all three. Regarding ED treatment, 2.7% answered correctly for AN, 20.08% answered correctly for BN, and 20.08% answered correctly for BED, with none correctly answering all three items. Knowledge about diagnostic criteria for eating disorders The instrument included three questions about ED diagnostic criteria according to the DSM-5: one each on AN, BN, and BED. “Intense fear of getting fat” was the top-rated item for AN (89.2%), followed by “Excessive preoccupation with weight and body shape” (69.5%), and “Significantly low body weight” (68.3%). “Amenorrhea” was mistakenly marked by 28.2% of the sample as a diagnostic criterion for AN. For BN, “Inappropriate compensatory behaviors for weight control” was correctly identified by 94.6% of participants, while “Binge eating episodes once a week for 3 months” was only marked by 52.5%. A total of 61% correctly identified “Excessive preoccupation with weight and body shape” for BN diagnosis, while 29.34% incorrectly identified “Binge eating episodes twice a week for 3 months.” Regarding BED, the items with the highest number of correct answers were “Intense suffering with binge eating episodes” (89.6%) and “Binge eating episodes once a week for 3 months” (56.4%). Nevertheless, “Binge eating episodes twice a week for 3 months” and “Inappropriate compensatory behaviors for weight control” were erroneously marked as correct by 35.4 and 16.6% respectively. Knowledge about treatment recommendations for eating disorders There were three questions on ED treatment recommendations according to international guidelines. Regarding AN, 91.5% correctly identified “Cognitive behavioral therapy,” 47.1% correctly identified “Family therapy,” and 32.4% correctly identified “Atypical antipsychotics,” while 60.6% incorrectly identified “Selective serotonin reuptake inhibitors.” Regarding BN, 93.8% correctly identified “Cognitive behavioral therapy” and 83.8% correctly identified “Fluoxetine,” but only 47.5% correctly identified “Topiramate” as an accepted therapy. Regarding BED, 92.3% correctly identified “Cognitive behavioral therapy,” 56.8% “Lisdexamfetamine,” 65.3% “Fluoxetine,” and 69.1% “Topiramate.”
The instrument included three questions about ED diagnostic criteria according to the DSM-5: one each on AN, BN, and BED. “Intense fear of getting fat” was the top-rated item for AN (89.2%), followed by “Excessive preoccupation with weight and body shape” (69.5%), and “Significantly low body weight” (68.3%). “Amenorrhea” was mistakenly marked by 28.2% of the sample as a diagnostic criterion for AN. For BN, “Inappropriate compensatory behaviors for weight control” was correctly identified by 94.6% of participants, while “Binge eating episodes once a week for 3 months” was only marked by 52.5%. A total of 61% correctly identified “Excessive preoccupation with weight and body shape” for BN diagnosis, while 29.34% incorrectly identified “Binge eating episodes twice a week for 3 months.” Regarding BED, the items with the highest number of correct answers were “Intense suffering with binge eating episodes” (89.6%) and “Binge eating episodes once a week for 3 months” (56.4%). Nevertheless, “Binge eating episodes twice a week for 3 months” and “Inappropriate compensatory behaviors for weight control” were erroneously marked as correct by 35.4 and 16.6% respectively.
There were three questions on ED treatment recommendations according to international guidelines. Regarding AN, 91.5% correctly identified “Cognitive behavioral therapy,” 47.1% correctly identified “Family therapy,” and 32.4% correctly identified “Atypical antipsychotics,” while 60.6% incorrectly identified “Selective serotonin reuptake inhibitors.” Regarding BN, 93.8% correctly identified “Cognitive behavioral therapy” and 83.8% correctly identified “Fluoxetine,” but only 47.5% correctly identified “Topiramate” as an accepted therapy. Regarding BED, 92.3% correctly identified “Cognitive behavioral therapy,” 56.8% “Lisdexamfetamine,” 65.3% “Fluoxetine,” and 69.1% “Topiramate.”
Only 15.1% of the participants received specialized training in ED during medical school, and 59.8% underwent a residency training program or graduate studies in psychiatry. A total of 58.7% mentioned complementary ED training (courses, conferences, specific events). A total of 8.89% of the respondents felt satisfied with their ED training, 50.97% felt confident in diagnosing ED, and 37.07% felt confident in treating ED patients ( ).
This is the first study to assess ED knowledge, including diagnosis and treatment, among Brazilian psychiatrists. Regarding ED knowledge, 33.21% of the respondents correctly identified the diagnostic criteria for AN, 29.73% for BN, and 38.22% for BED. Regarding ED treatment, the therapeutic options for BN and BED were correctly identified by 20.8% of the respondents, but the proportion was considerably lower for AN – only 2.7%. Specialized ED training was reported by 15.1% during medical school, 59.8% during residency or graduate study, and 58.7% in complementary training. Only 8.89% felt satisfied with their ED training; 50.97% felt confident in diagnosing ED, and 37.07% felt confident in treating patients with ED. In line with previous studies on ED knowledge among health professionals in other countries, , , we found an important gap in ED training among Brazilian psychiatrists. Considering that general psychiatrists are usually the first available resource for the majority of patients with ED in our large developing country, their lack of accurate information and limited clinical expertise can intensify ED stigma and stereotypes, compromising treatment adherence among these patients and creating an additional barrier to effective treatment. , More than 28% of the psychiatrists identified amenorrhea, a former diagnostic criterion for AN in the DSM-IV, as a current criterion in the DSM-5, and 29.34% incorrectly responded to an item about the frequency of compensatory episodes in BN, which was also used in the DSM-IV. This might be due to the respondents’ lack of scientific updating. Identifying selective serotonin reuptake inhibitors as a treatment for AN was another frequent mistake (60.6% of the sample), probably due to its good results in patients with comorbidities. However, strictly speaking, the main international ED treatment guidelines do not consider them an option for patients with AN only. These results were lower than expected, despite the fact that psychiatrists are considered to have the highest level of ED knowledge among health professionals. Even in the United Kingdom, which is considered quite advanced in ED, medical education is considered insufficient to provide the necessary skills and knowledge to safely treat ED patients – less than 2 hours in medical schools and a little more time in graduate training. , Our data indicate a similar situation in Brazil. We believe that including more theoretical and practical training about ED during medical school, graduate study, and psychiatric residency programs would improve young psychiatrists’ knowledge of ED, and a consistent schedule of continuing medical education might improve the performance of older professionals. The main challenges are raising awareness of ED, its negative impact on the community, and attracting the interest of stakeholders and financial support to develop and widely disseminate strategies to increase ED knowledge among psychiatrists. Some recently proposed solutions to address ED treatment gaps can also be considered, such as: train-the-trainer, in which training is provided to one professional who then trains others in the region; web-centered training and supervision, which is more scalable and less expensive; and higher-level support and policy to provide a more significant chance of success than waiting for clinicians to voluntarily seek training and modify their behavior. Two online ED training programs for health professionals, both developed in Australia, have shown promising results, , improving health professionals’ knowledge, skills, and confidence for treating people with ED and reducing their prejudice towards ED. These programs represent a novel and probably effective way of meeting the educational needs of general psychiatrists who work with ED patients. Investment in the adaptation and testing of online training programs for ED may help address some of the issues raised in this study. Our results demonstrate that Brazilian psychiatrists’ knowledge of ED is inadequate, although most participants acknowledged this in their responses about confidence and competence. Therefore, academic training about ED must improve among Brazilian psychiatrists, which could lead to improved ED care in Brazil. Although ours is the first study to investigate ED knowledge among Brazilian psychiatrists, it has some limitations, such as no sample size calculation, using a convenience sample, using an instrument that was not previously validated, and collecting self-reported data. These factors could limit the national representativeness of our findings and should be addressed in future studies. We attribute the low response rate (4.32%) to the fact that this was an online survey. Evidence suggests that response rates to online surveys are very low, unlikely to reach the expected rates of in-person or mailed surveys. However, since we conducted this study during the COVID-19 pandemic, a period of great worldwide operational difficulty, the online methodology was necessary for data collection. The response rate may also reflect a low level of interest in ED treatment by general psychiatrists. As a consequence of the results of this study, we suggest that Brazilian psychiatrists receive greater training in ED content. Furthermore, additional training should be regularly offered, especially to older psychiatrists, who may have had no previous training or whose training has become outdated. Finally, our findings indicate that our country needs greater investment in and resources for ED.
The authors report no conflicts of interest.
|
Stratified Impact of Therapies on Anaplastic Thyroid Cancer Outcomes in the Pre-Gene-Targeted Therapy Era | 5931e3be-5db4-4f23-9101-addcc3283159 | 11882711 | Surgical Procedures, Operative[mh] | Data Sources The study extracted a cohort of pathological confirmed patients with ATC from the Surveillance, Epidemiology, and End Results (SEER) program, a public database devoted to providing information on patient demographics, tumor morphology, stage at diagnosis, primary tumor site, first course of treatment, and follow-up for vital status and causes of death, in an effort to reduce the cancer burden among the US population. Given that the SEER database consists of de-identified patient information available to the public, our study did not require the approval of an Institutional Review Board. In conducting our research, we adhered to the Strengthening the Reporting of Observation Studies in Epidemiology (STROBE) guidelines for reporting observational data. In accordance with our previous studies, , the selected database is cited as ‘Incidence – SEER Research Plus Data, 18 Registries, Nov 2020 Sub (2000–2018) – Linked To County Attributes – Total U.S., 1969-2019 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, released April 2021, based on the November 2020 submission’. Study Cohort The study cohort initially comprised 2504 consecutive patients with ATC who were pathologically diagnosed between 2000 and 2018, as selected retrospectively from the SEER registry. We excluded several groups of patients from our study: 244 who underwent palliative or exploratory surgeries without removal of the primary lesion; 180 who had a partial thyroidectomy or lobectomy; 198 who had confounding records of therapeutic radioactive iodine (RAI) treatment; and 7 for whom no survival time was documented. Ultimately, our final study cohort consisted of 1879 patients with ATC, all of whom either underwent the complete cancer removal procedure, total thyroidectomy (TT; defined in this study as operable patients), or did not receive any surgery (defined in this study as inoperable patients). The clinical data regarding demographics, tumor staging, and therapeutic approaches, i.e. age, sex, race, primary site of tumor, pathology, American Joint Committee on Cancer (AJCC) stage, primary surgery, radiotherapy, chemotherapy, survival months, and cause of death, were extracted. Outcome Definition The cause of death was recorded to define whether each patient died from ATC, from causes unrelated to patients with ATC, or if they were alive at the end of the follow-up period. In our study, the primary outcome was OS, defined as the time from initial diagnosis to death from any cause or alive at the endpoint of the study. Survival Analysis Univariate Cox proportional hazards regression and Kaplan–Meier curves were utilized to identify prognostic factors. Factors deemed significant by univariate analysis ( p < 0.05) were subsequently included in the multivariate Cox proportional hazards models. The optimal Cox regression model was chosen through a backward selection process, with an entry criterion of p < 0.05 and an elimination criterion of p > 0.10. Multivariate Cox proportional hazards regression analyses were carried out to pinpoint variables that significantly impacted the OS of patients with ATC. Statistical Analysis We have presented descriptive statistics in Table for the entire study cohort and compared outcomes between patients with ATC who did or did not undergo TT. Continuous variables were analyzed using the Kruskal–Wallis test and expressed as mean ± standard deviation (SD)/median, while categorical variables were assessed using the Pearson Chi-square test and presented as number (percentage). All statistical analyses were performed using R Studio version 4.0.4. Kaplan–Meier curves were generated using GraphPad Prism 8.4.3 (GraphPad Software, Inc., San Diego, CA, USA). A two-tailed p -value of <0.05 was considered to indicate statistical significance.
The study extracted a cohort of pathological confirmed patients with ATC from the Surveillance, Epidemiology, and End Results (SEER) program, a public database devoted to providing information on patient demographics, tumor morphology, stage at diagnosis, primary tumor site, first course of treatment, and follow-up for vital status and causes of death, in an effort to reduce the cancer burden among the US population. Given that the SEER database consists of de-identified patient information available to the public, our study did not require the approval of an Institutional Review Board. In conducting our research, we adhered to the Strengthening the Reporting of Observation Studies in Epidemiology (STROBE) guidelines for reporting observational data. In accordance with our previous studies, , the selected database is cited as ‘Incidence – SEER Research Plus Data, 18 Registries, Nov 2020 Sub (2000–2018) – Linked To County Attributes – Total U.S., 1969-2019 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, released April 2021, based on the November 2020 submission’.
The study cohort initially comprised 2504 consecutive patients with ATC who were pathologically diagnosed between 2000 and 2018, as selected retrospectively from the SEER registry. We excluded several groups of patients from our study: 244 who underwent palliative or exploratory surgeries without removal of the primary lesion; 180 who had a partial thyroidectomy or lobectomy; 198 who had confounding records of therapeutic radioactive iodine (RAI) treatment; and 7 for whom no survival time was documented. Ultimately, our final study cohort consisted of 1879 patients with ATC, all of whom either underwent the complete cancer removal procedure, total thyroidectomy (TT; defined in this study as operable patients), or did not receive any surgery (defined in this study as inoperable patients). The clinical data regarding demographics, tumor staging, and therapeutic approaches, i.e. age, sex, race, primary site of tumor, pathology, American Joint Committee on Cancer (AJCC) stage, primary surgery, radiotherapy, chemotherapy, survival months, and cause of death, were extracted.
The cause of death was recorded to define whether each patient died from ATC, from causes unrelated to patients with ATC, or if they were alive at the end of the follow-up period. In our study, the primary outcome was OS, defined as the time from initial diagnosis to death from any cause or alive at the endpoint of the study.
Univariate Cox proportional hazards regression and Kaplan–Meier curves were utilized to identify prognostic factors. Factors deemed significant by univariate analysis ( p < 0.05) were subsequently included in the multivariate Cox proportional hazards models. The optimal Cox regression model was chosen through a backward selection process, with an entry criterion of p < 0.05 and an elimination criterion of p > 0.10. Multivariate Cox proportional hazards regression analyses were carried out to pinpoint variables that significantly impacted the OS of patients with ATC.
We have presented descriptive statistics in Table for the entire study cohort and compared outcomes between patients with ATC who did or did not undergo TT. Continuous variables were analyzed using the Kruskal–Wallis test and expressed as mean ± standard deviation (SD)/median, while categorical variables were assessed using the Pearson Chi-square test and presented as number (percentage). All statistical analyses were performed using R Studio version 4.0.4. Kaplan–Meier curves were generated using GraphPad Prism 8.4.3 (GraphPad Software, Inc., San Diego, CA, USA). A two-tailed p -value of <0.05 was considered to indicate statistical significance.
In this study, using the SEER database, we identified a cohort of 1879 patients with pathologically diagnosed ATC, spanning the period from 2000 to 2018. The characteristics of this population of patients with ATC included a diagnosis at an older age (median age of 71 years [range 16–85 years]), a predominance of White individuals (1509 patients, 80.3%), and a slight female preponderance, with a male-to-female sex ratio of approximately 2:3. The median follow-up duration for the study cohort was 3 months (ranging from 1 month to 224 months). During the follow-up period, there were 1708 deaths (90.9%), with 1441 (76.7%) attributed to patients with ATC and 267 (14.2%) attributed to non-thyroid causes, leading to cancer-specific death (CSS) as the primary cause. The 1-year CSS and OS rates for patients with ATC were 24.2% (95% confidence interval [CI] 22.2–26.3%) and 19.3% (95% CI 17.6–21.1%), respectively. The median survival time among patients with ATC was 3 months. Kaplan–Meier curves estimating OS for patients with ATC are illustrated in Figs. a and 1a′. The study cohort was divided into two groups, based on whether patients had undergone complete thyroid removal surgery or not. The two groups were approximately equal in size, however there were noticeable disparities in clinical-oncological characteristics between the groups, except for sex ( p = 0.645) ratio and ethnicity proportion ( p = 0.069). The subcohort of operable patients was slightly younger, with a median age of 68 years, compared with 73 years in the inoperable subcohort. In addition, the TT subcohort had a higher proportion of patients staged as IVa and IVb (12.1% vs. 5.9%, and 46.3% vs. 35.9%, respectively) and a greater prevalence of localized lesions (T1-T3a: 12.6% vs. 9.2%; T3b: 9.6% vs. 2.9%; T4a: 31.6% vs. 19.7%; T4b: 10.5% vs. 23.2%) compared with the subcohort who did not undergo surgery. The acceptance of adjuvant/palliative therapy was slightly more favorable in the operable subcohort compared with the inoperable subcohort. Specifically, a greater proportion of patients in the operable subcohort received radiotherapy (62.8% vs. 52.6%) and chemotherapy (43.9% vs. 36.5%) compared with the inoperable subcohort. After the follow-up period, a noteworthy proportion of patients with ATC in the operable subcohort had survived (18.8%), with a median survival time of 7 months; the 1-year OS rate for this group was 41.8% (95% CI 38.6–45.3%). Conversely, for those patients who were inoperable, only 2.0% had survived, with a median survival time of 2 months; the 1-year OS rate was 9.0% (95% CI 7.5–10.1%). These results demonstrate the substantial disparity in survival rates between patients eligible for surgery and those who were not. To provide a clear visual representation of the impact of surgical intervention and AJCC staging on survival, we plotted Kaplan–Meier curves (Fig. ). Detailed descriptive data for the study cohort are presented in Table and electronic supplementary material (ESM) Table . The median survival time and 1-year OS associated with different therapeutic combinations are shown in Table . To further determine the potential survival benefit of radiotherapy, chemotherapy, and their combination in both the operable and inoperable subcohorts, we conducted two sets of Cox proportional hazard regression analyses. In the multivariate Cox analysis estimating OS of the operable subcohort, several factors were identified as significant prognostic indicators. Older age, with each increase in 1 year, was associated with a higher hazard ratio (HR 1.04, 95% CI 1.03–1.04; p < 0.001). Additionally, higher AJCC stage, particularly stage IVb compared with stage IVa (HR 2.23, 95% CI 1.64–3.03; p < 0.001), AJCC N1b compared with N0 (HR 1.42, 95% CI 1.04–1.94; p = 0.028), and tumor size >5 cm compared with ≤1 cm (HR 2.31, 95% CI 1.01–5.28; p = 0.046) were all identified as significant adverse factors affecting survival after adjustment and model selection in the univariate Cox regression. Radiotherapy (vs. no radiotherapy; HR 0.62, 95% CI 0.50–0.77; p < 0.001), unlike chemotherapy (vs. no chemotherapy; HR 0.93, 95% CI 0.63–1.36; p = 0.703) was the only therapeutic factor predicting better OS for the subcohort of operable patients with ATC. In the univariate Cox regression analysis of the inoperable population, AJCC stage, tumor size, and lymph node involvement did not meet the entry criteria ( p < 0.05). As a result, the palliative purpose of chemotherapy and radiotherapy for the inoperable population with ATC was estimated to be beneficial prognostic factors (e.g. chemotherapy alone vs. no chemotherapy; HR 0.53, 95% CI 0.40–0.70; p < 0.001) in the final multivariate Cox model. Survival curves of the survival benefit of radiotherapy, chemotherapy, and their combination in both the operable and inoperable subcohorts are shown in Fig. . Detailed survival HRs are presented in ESM Table . To further investigate the largest population within this cohort of patients with ATC, which comprises 757 patients (40.3% of the cohort, as detailed in ESM Table ), and to exclude patients with unspecified stage, we selected those with AJCC stage IVb as the secondary cohort for analysis. Stage IVb is also an important focus of our study because there is a lot of debate regarding the best approach for treating these patients. The baseline demographics and clinical-oncological characteristics remained consistent with those of the entire cohort of patients with ATC. Specific details are presented in ESM Table . Following adjustment through univariate regression analysis, another multivariate Cox regression was conducted. In this analysis, only age, tumor extension, and treatment combinations were identified as significant prognostic factors in the AJCC stage IVb population. All therapies, whether used alone or in combination, as illustrated in Fig. , possess distinct therapeutic value, with complete removal surgery having the most significant impact. ESM Table presents the detailed data from the Cox regression model for patients with stage IVb ATC.
In oncology, ATC is challenging because of its rare, aggressive nature, hindering randomized controlled trials and robust research. Consequently, ATC treatments are often empirical. Despite these challenges, our study aimed to precisely predict the survival benefits of different therapies for patients with ATC. The current study, leveraging the robust SEER database, provides critical insights into the representative baseline management of patients with ATC. As expected, all traditional treatments showed efficacy in enhancing OS in patients with ATC, compared with no treatment at all, as illustrated in Tables and Figs. and . A pivotal finding from our analysis is the significant improvement in survival outcomes associated with complete thyroidectomy in select patients with ATC (Fig. ). This observation aligns with existing literature that supports aggressive surgical intervention in patients with ATC, but emphasizes its application in selected cases. – , – In analyzing the treatment outcomes for patients with ATC, it is essential to discern the incremental benefit of combining EBRT and chemotherapy with surgery. While surgery alone marked a substantial 30% increase in 1-year OS, the addition of EBRT or chemotherapy showed only a marginal improvement in survival rates. Specifically, surgery combined with EBRT led to a 36.7% increase in 1-year OS, and surgery with chemotherapy resulted in a 30.4% increase (Table ). This suggests that while the addition of EBRT or chemotherapy to surgical intervention does enhance patient outcomes, the extent of this enhancement is relatively modest compared with the significant impact of surgery on its own. In light of these findings, it becomes apparent that the primary driver in improving 1-year OS in patients with ATC is surgery, and the role of EBRT and chemotherapy, although contributory, is less pronounced in the context of combined therapy. Additionally, the combined use of EBRT and chemotherapy, which resulted in a 15.3% increase in 1-year OS, is notably less effective compared with surgery alone, which achieved a 30% improvement in 1-year OS. Another critical observation is that the combination of all three approaches yielded the highest increase in 1-year OS, achieving a 41.8% improvement. It reveals a layered hierarchy of effectiveness in improving 1-year OS. It is well-established that the outcomes for patients with ATC are largely -sensitive, which is also supported by our study (Figs. and 3). There may be concerns that the distribution of the study cohort across AJCC stages (IVa: 160; IVb: 757; IVc: 332; and unstaged: 630) may introduce bias in evaluating the prognostic response to various therapy modalities. To address this issue and minimize the confounding effects associated with various disease stages, we focus on distinct stages of the disease in our Discussion. In stage IVa patients with ATC, as illustrated in Fig. a, surgery was found to be crucially important. It increased the 1-year OS rate from 21.5% to 71.8% and extended the median survival time from 2 months to 23.5 months, compared with patients who did not undergo surgery, aligning with the ATA and EMSO guidelines. In the case of stage IVc patients with ATC, as shown in Fig. c, the disparity between survival curves for those who underwent surgery and those who did not was not pronounced. The implication is that for stage IVc patients with ATC, multimodal adjuvant therapy might be more favorable in terms of OS compared with surgery alone. Lastly, we focused on stage IVb patients. For this group in particular, there has been a lot of controversy and disagreement about the best treatment options, making stage IVb a critical focus for evaluating the impact of different approaches. The Kaplan–Meier curves in Fig. d display the outcomes of various therapy modalities for stage IVb patients with ATC, showing two significant divergences, similar to those in Fig. b. This latter figure illustrates the bifurcations observed in patients treated with or without TT. This confirms that surgery is fundamental in determining the OS in stage IVb patients with ATC. The final multivariate Cox regression analysis for OS in stage IVb patients with ATC indicated that the combination of surgery with either EBRT alone (HR 0.15, CI 0.11–0.21; p < 0.001), chemotherapy alone (HR 0.14, 95% CI 0.08–0.26; p < 0.001), or both (HR 0.15, 95% CI 0.11–0.20; p < 0.001), resulted in a similar and substantial decrease in overall mortality. This was followed by surgery alone (HR 0.28, 95% CI 0.21–0.37; p < 0.001), the combination of EBRT and chemotherapy (HR 0.33, 95% CI 0.25–0.42; p < 0.001), chemotherapy alone (HR 0.41, 95% CI 0.26–0.65; p < 0.001), and EBRT alone (HR 0.41, 95% CI 0.31–0.54; p < 0.001). These patterns underscore the multidimensional nature of effective treatment protocols for patients with ATC. The combination of all three modalities, i.e. surgery, EBRT, and chemotherapy, emerges as the most advantageous strategy, suggesting a synergistic effect that maximizes patient survival outcomes. However, it is important to note that surgery remains the cornerstone of traditional treatment. Surgery will continue to be our most powerful tool until the advent of molecular-based targeted therapies, which may have already begun redefining treatment approaches, both now and in the foreseeable future. ATC is known for its aggressive behavior and poor prognosis, and recent advances in genomic profiling have identified several genetic mutations that may play a pivotal role in its pathogenesis. One of the most commonly studied mutations is the BRAF mutation, which has been reported in approximately 20–50% of ATC cases. , , This mutation, along with others such as TP53 and RAS, presents potential opportunities for targeted therapy. Despite these important findings, it is important to acknowledge that the SEER database, used in this study, does not include detailed genetic mutation data. This limitation restricts our ability to directly assess the impact of such mutations on patient outcomes in this cohort. However, the results of some recent clinical studies have shown that the value of targeted therapy for ATC is promising. In one study involving targeted therapies, MD Anderson Cancer Center compiled a comprehensive cohort of 479 patients treated over the past 2 decades. Among these patients, those who underwent surgery after receiving neoadjuvant BRAF-directed therapy ( n = 20) experienced an impressive 1-year survival rate of 94%. This single-institution cohort study, which encompassed 479 patients with ATC spanning nearly 20 years, revealed significant increases in both 1- and 2-year survival rates. Specifically, 1- and 2-year survival rates rose from 35% and 18% in the 2000–2013 era ( n = 227) to 47% and 25% in the 2014–2016 era ( n = 100), and to 59% and 42% in the 2017–2019 era ( n = 152), respectively. In another relatively large retrospective study ( n = 104) led by the Memorial Sloan Kettering Cancer Center, the median survival for the overall cohort was 7 months, with a 1-year survival rate of 34%. Meanwhile, patients who were selectively treated with surgery and postoperative concurrent chemoradiation therapy ( n = 53) exhibited a 1-year survival rate of 55%. These results, combined with insights from other smaller-scale studies, , , indicate that patients with ATC can be effectively managed with highly specialized, molecular-based personalized therapies. This includes the strategic use of surgery when it is considered suitable, regardless of the disease stage. The limitations of our study are obvious. The retrospective design and dependence on the SEER database, while providing a broad dataset, introduces potential limitations. In this extracted cohort, we observed that a significant portion of patients in the SEER database lacked clear staging information, which posed certain challenges to our analysis. The missing staging information may lead to potential selection bias, affecting our assessment of patient prognosis. Another inherent bias is that patients who are able to undergo complete surgical resection for ATC are typically diagnosed at an earlier stage (IVa and IVb), which is associated with longer disease-free survival and OS. This creates a potential confounder, as the improved outcomes observed in these patients may be partially attributed to the earlier stage of their disease, rather than solely attributed to the therapeutic effects of surgery itself. Another significant limitation of our study is the absence of detailed information regarding the timing and use of targeted therapies for patients with ATC in the SEER database. Although targeted treatments for ATC were not as well-established between 2000 and 2018, this lack of information makes it challenging to build a comprehensive dataset in the SEER program. Acknowledging this limitation allows for a more nuanced interpretation of our findings and underscores the need for further research to investigate the role of targeted therapies in ATC as data become available.
Our study offers significant insights into the stratified landscape of baseline treatments for patients with ATC, highlighting the critical role of surgery-centered multimodal therapy. This stratification not only improves our evidence-based understanding of current treatment outcomes but also establishes an essential benchmark for future therapeutic advancements. It is still uncertain whether aggressive local treatment in metastatic cases leads to improved survival or primarily enhances local disease control. Further investigation into this question, such as analyzing the cause of death in ATC patients who underwent TT and local treatments, using the National Cancer Database (NCDB) dataset, would provide valuable insights. Such research could help refine treatment guidelines and support more informed decision making for ATC patients with distant metastases.
Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 26 KB) Supplementary file2 (DOCX 21 KB)
|
The expression of RIPK3 is associated with cell turnover of gastric mucosa in the mouse and human stomach | 8aef486a-07fd-46fb-8a63-9eec9b10aca7 | 8324621 | Anatomy[mh] | Recently, the relationship between cell death and human diseases has become a great interesting topic in research (Zhou and Yuan ). Traditionally, the manner of cell death can be divided into two types: necrosis and programmed cell death (PCD). Necrosis is a passive cell death and can be regarded as an extreme cell clearance method in organisms caused by pathological factors such as physics, chemistry, and pathogens. PCD is an active process of cell self-regulation, mainly characterized by genetically controlled cell-autonomous ordered death. More recently, people gradually realized that cell necrosis could also occur in an orderly manner. This new cell death form was termed as necroptosis. Studies revealed that pro-inflammatory cytokines such as tumor necrosis factor (TNF)-α play an important stimulating role in the process of necroptosis (Laster et al. ). TNF-α activates receptor-interacting protein kinase (RIPK) 1 (Moriwaki and Chan ; Silke et al. ; Yu, ), and in turn activates mixed lineage kinase domain-like (MLKL) and further regulates the downstream element RIPK3 and causes necroptosis (Sun et al. ; Zhang et al. ). Excessive necroptosis in cells will induce the release of inflammatory mediators and inflammation. Accordingly, excessive increases in necroptosis is one of the hallmarks of inflammatory diseases (Dagenais et al. ; Lee et al. ; Pierdomenico ), which is postulated through an disturbed intestinal epithelial cell turnover rate and homeostasis (Gunther et al. , ), and relates to the initiation of intestinal inflammation (Dagenais et al. ; Gunther et al. ; Negroni et al. ; Takahashi et al. ; Werts, ). However, there are few studies to examine the expression of RIPK3 in the gastric mucosa under physiological condition. As we have known, one of principal functions of the stomach is to create an acidic milieu that predominately depends on hydrochloric acid produced by parietal cells in the fundic glands, which is mainly controlled by the gastrin-histamine sequence (Cui and Waldum ). Therefore, the stomach is exposed to a high concentration of acid and various endogenous and exogenous noxious agents that could induce a rapid turnover rate in gastric cells and makes the gastric mucosa as one of the most rapidly proliferating tissues in humans. Cell proliferation and cell death are two essential parts involved in the physiological cellular turnover of gastric mucosa. Studies have shown that pathological conditions such as Helicobacter pylori ( H. pylori ) infection, a main cause of peptic ulcers and gastric cancer, will result in a remarkably altered cell proliferation and cell death rate in the stomach (Alzahrani et al. ; Chen, ; Cui, ; Jones et al. ; von Herbay and Rudi ) and contribute significantly to the development of preneoplastic and neoplastic lesions (Cui et al. ). Therefore, the importance of gastric cell death manner and turnover has become an attractive topic of research. Here, we immunohistochemically characterize the key element of necroptosis, RIPK3, under physiological condition in vivo. We show a high rate of RIPK3 positive cells in mouse and human gastric specimens, RIPK3 was identified in different cell types such as surface mucous, fundic parietal and lamina propria cells in the stomachs.
Animals FVB/N mice: FVB/N mouse is one of the most commonly used laboratory animal models in the gastric research field worldwide. A total of 10 gastric fundic paraffin blocks of FVB/N male mice (Animal Center, Zhengzhou University, China) at the age of 6 months were included in this study. Animal study protocol was approved by the local medical ethics committee of Secondary Affiliated Hospital of Zhengzhou University (No. 182300410326). Human gastric specimens A total of 20 gastric paraffin blocks taken from corpus mucosa of health human (male/female ratio: 5/15; average age: 46 years), retrieved from tissue bank at Department of Pathology, the Second Affiliated Hospital of Zhengzhou University, were included in this study. Human study protocol was approved by the institutional medical ethic review board of Zhengzhou University (No. U170410161) and written informed consent was taken from all the participants involved in the study. Stomach immunohistochemistry (IHC) in FVB/N male mice and human specimens To measure gastric cell proliferation and necroptosis, RIPK3 and PCNA IHCs in the gastric specimens from male FVB/N mice and health human gastric tissues were examined in this study. Midline strips along the lesser curvature of the stomach were fixed in 10% neutral buffered formalin, processed and embedded in paraffin. Sections were cut at 4 μm, and then stained with hematoxylin and eosin ( H&E ). IHCs were performed with avidin–biotin-peroxidase complex (ABC) Elicit kits (Vector Laboratories, Burlingame, CA, USA) according to manufacturer’s instructions and our previous published method (Cui et al. , , , ). After antigen retrieval achieved by boiling sections 2 × 6 min, rabbit anti-RIPK3 polyclonal antibody (working dilution 1:800, Thermo Fisher Scientific, USA) and rabbit anti-PCNA monoclonal antibody (working dilution 1:100, Creative Diagnostics; Uppsala, Sweden). Primary antibodies were incubated with mouse and human sections respectively at 4 °C overnight in humidified chamber. 3-Amino-9-ethylcarbazole ( AEC ; Vector Laboratories, Burlingame, CA) was used as chromogen and slides were counterstained with Mayer’s hematoxylin (Vector Laboratories, Burlingame, CA, USA). In addition, to illustrate active RIPK3 form was expressed in mouse gastric mucosal cells, IHC with rabbit anti-mouse phospho-RIPK3 Thr231-Ser232 monoclonal antibody (Abcam, UK) was done in mouse sections. Double immunofluorescence staining for RIPK3 expressed in glandular parietal and lamina propria cells To examine the expression of RIPK3 in gastric glandular cells, double immunofluorescences with RIPK3/H + K + -ATPase subunit (to label parietal cells; working dilution 1:1000, mouse anti-hog monoclonal antibody, Affinity Bioreagents, Golden, CO, USA, for human stomach specimens; Rabbit anti- H + K + -ATPase beta/ATP4B antibody, Abcam, UK, for mouse stomach specimens respectively) in both mouse and human sections were done according to the protocol described in our previous publication (Cui et al. , , , ). After gastric fundic sections incubated with primary antibodies at 4 °C overnight, RIPK3-immunoreactivities (IRs) were developed with fluorescein isothiocyanate (FITC)-conjugated secondary antibody (Jackson ImmunoRearch Lab., West Grove, PA, USA), and H + K + -ATPase-IR was with Alexa 647-conjugated secondary antibody (Jackson ImmunoRearch Lab., West Grove, PA, USA). To evaluate phenotypes of RIPK3-positive cells in the lamina propria, double DIFs with RIPK3/CD3 (to show RIPK3 in T lymphocytes, mouse anti-CD3 monoclonal antibody, DAKO, Carpinteria, CA, USA), RIPK3/CD34 (to show RIPK3 in CD34-positive stromal cells, mouse anti-CD34 monoclonal antibody, DAKO, Carpinteria, CA, USA) and RIPK3/SMA-alpha (to show RIPK3 in stromal myofibroblasts, mouse anti-SMA-α monoclonal antibody, DAKO, Carpinteria, CA, USA) were performed in human gastric sections. After gastric fundic sections incubated with primary antibodies at 4ºC overnight, RIP3-IR was developed with FITC-conjugated secondary antibody and CD3, CD34 and SMA-α-IRs were developed with Alexa 647-conjugated secondary antibody. Double immunofluorescence-stained sections were observed and photographed under a confocal microscopy (LSM-700, Carl Zeiss, Jena, Germany) under 200 × magnification. Morphological evaluation Only immunoreactive fundic glandular cells with appropriate morphology and location in well-oriented sections were counted under microscope. The volume densities of PCNA labeling cells and RIPK3-positive cells located in gastric glands were counted according to the published method (Zhao et al. ). Statistical analysis The data are presented as the mean ± SEM (standard error of the mean) unless otherwise stated. P values were evaluated by the Mann–Whitney test. P values < 0.05 were considered statistically significant.
FVB/N mice: FVB/N mouse is one of the most commonly used laboratory animal models in the gastric research field worldwide. A total of 10 gastric fundic paraffin blocks of FVB/N male mice (Animal Center, Zhengzhou University, China) at the age of 6 months were included in this study. Animal study protocol was approved by the local medical ethics committee of Secondary Affiliated Hospital of Zhengzhou University (No. 182300410326).
A total of 20 gastric paraffin blocks taken from corpus mucosa of health human (male/female ratio: 5/15; average age: 46 years), retrieved from tissue bank at Department of Pathology, the Second Affiliated Hospital of Zhengzhou University, were included in this study. Human study protocol was approved by the institutional medical ethic review board of Zhengzhou University (No. U170410161) and written informed consent was taken from all the participants involved in the study.
To measure gastric cell proliferation and necroptosis, RIPK3 and PCNA IHCs in the gastric specimens from male FVB/N mice and health human gastric tissues were examined in this study. Midline strips along the lesser curvature of the stomach were fixed in 10% neutral buffered formalin, processed and embedded in paraffin. Sections were cut at 4 μm, and then stained with hematoxylin and eosin ( H&E ). IHCs were performed with avidin–biotin-peroxidase complex (ABC) Elicit kits (Vector Laboratories, Burlingame, CA, USA) according to manufacturer’s instructions and our previous published method (Cui et al. , , , ). After antigen retrieval achieved by boiling sections 2 × 6 min, rabbit anti-RIPK3 polyclonal antibody (working dilution 1:800, Thermo Fisher Scientific, USA) and rabbit anti-PCNA monoclonal antibody (working dilution 1:100, Creative Diagnostics; Uppsala, Sweden). Primary antibodies were incubated with mouse and human sections respectively at 4 °C overnight in humidified chamber. 3-Amino-9-ethylcarbazole ( AEC ; Vector Laboratories, Burlingame, CA) was used as chromogen and slides were counterstained with Mayer’s hematoxylin (Vector Laboratories, Burlingame, CA, USA). In addition, to illustrate active RIPK3 form was expressed in mouse gastric mucosal cells, IHC with rabbit anti-mouse phospho-RIPK3 Thr231-Ser232 monoclonal antibody (Abcam, UK) was done in mouse sections.
To examine the expression of RIPK3 in gastric glandular cells, double immunofluorescences with RIPK3/H + K + -ATPase subunit (to label parietal cells; working dilution 1:1000, mouse anti-hog monoclonal antibody, Affinity Bioreagents, Golden, CO, USA, for human stomach specimens; Rabbit anti- H + K + -ATPase beta/ATP4B antibody, Abcam, UK, for mouse stomach specimens respectively) in both mouse and human sections were done according to the protocol described in our previous publication (Cui et al. , , , ). After gastric fundic sections incubated with primary antibodies at 4 °C overnight, RIPK3-immunoreactivities (IRs) were developed with fluorescein isothiocyanate (FITC)-conjugated secondary antibody (Jackson ImmunoRearch Lab., West Grove, PA, USA), and H + K + -ATPase-IR was with Alexa 647-conjugated secondary antibody (Jackson ImmunoRearch Lab., West Grove, PA, USA). To evaluate phenotypes of RIPK3-positive cells in the lamina propria, double DIFs with RIPK3/CD3 (to show RIPK3 in T lymphocytes, mouse anti-CD3 monoclonal antibody, DAKO, Carpinteria, CA, USA), RIPK3/CD34 (to show RIPK3 in CD34-positive stromal cells, mouse anti-CD34 monoclonal antibody, DAKO, Carpinteria, CA, USA) and RIPK3/SMA-alpha (to show RIPK3 in stromal myofibroblasts, mouse anti-SMA-α monoclonal antibody, DAKO, Carpinteria, CA, USA) were performed in human gastric sections. After gastric fundic sections incubated with primary antibodies at 4ºC overnight, RIP3-IR was developed with FITC-conjugated secondary antibody and CD3, CD34 and SMA-α-IRs were developed with Alexa 647-conjugated secondary antibody. Double immunofluorescence-stained sections were observed and photographed under a confocal microscopy (LSM-700, Carl Zeiss, Jena, Germany) under 200 × magnification.
Only immunoreactive fundic glandular cells with appropriate morphology and location in well-oriented sections were counted under microscope. The volume densities of PCNA labeling cells and RIPK3-positive cells located in gastric glands were counted according to the published method (Zhao et al. ).
The data are presented as the mean ± SEM (standard error of the mean) unless otherwise stated. P values were evaluated by the Mann–Whitney test. P values < 0.05 were considered statistically significant.
PCNA and RIPK3-positive cells in the gastric fundic mucosa of FVB/N male mice and human To estimate the turnover rate of glandular cells of gastric mucosa, we examined the proliferation activity labelled by PCNA and the expression of necroptosis key element RIPK3 with IHCs in mouse and human stomachs respectively. In the stomachs of FVB/N male mice, PCNA-positive cells could be observed in the surface mucous cells (arrowhead in Fig. A), neck cell region of the gastric fundic glands (arrow in Fig. A) and lamina propria cells (red arrow in Fig. A). Similarly, RIPK3-positive cells in the fundus of FVB/N male mice could also be detected in both the gastric mucous cells (arrowhead in Fig. B), gastric glandular cells (arrow in Fig. B) and lamina propria cells (red arrow in Fig. A). However, great majority of RIPK3-positive cells were within the mouse gastric glands, and predominately located in the isthmus and neck regions (Fig. B), and a few in the base regions. In addition, active phosphor-RIPK3-IR was shown in some human gastric glandular parietal cells (Fig. C). In human stomach specimens, both PCNA positive cells (Fig. D) and RIPK3 positive cells (Fig. E) were also shown in glands (arrow), pit (surface mucous area, arrowhead) and lamina propria regions (red arrow). Many RIPK3-positive cells were observed in the fundic glands with a parietal cell morphology (Fig. E). Since the fundic glandular cells play a key role in gastric function, we therefore counted the volume densities of PCNA- and RIPK3-positive cells and PCNA/RIPK3 ratio in fundic glands in this study. Figure showed the volume densities of PCNA-positive glandular cells in both FVB/N mice (Fig. A) and humans (Fig. B). In FVB/N mice, the volume density of PCNA-positive cells ( white bar in Fig. A) in fundic glands was 12.4 ± 1.6/gland. The volume density of RIPK3-positive cells in fundic glands was 6.0 ± 0.95/gland ( grey bar in Fig. A). Accordingly, the ratio of PCNA/RIPK3 in the gastric fundic glands of 6-month-old FVB/N male mice was ~ 2.12 (black bar in Fig. A). When we divided the FVB/N mouse gastric mucosa into high and low turnover areas according to the PCNA density and counted ratio of PCNA/RIPK3 respectively, it showed that the ratio was 2,40 and 1.65/field respectively, indicating the expression of RIPK3 was higher in the area with a high proliferation rate than that in area with low proliferation rate. In human stomachs, counting data of PCNA- and RIPK3-positive cell numbers in the fundic glands revealed that volume densities of PCNA ( white bar in Fig. B) and RIPK3 ( grey bar in Fig. B) were (21.38 ± 2.27) and (5.25 ± 0.45) respectively. Therefore, the ratio of PCNA/RIPK3 in the human gastric fundic glands was ~ 4.15 ( black bar in Fig. B). Phenotypic identification of RIPK3-positive cells in glandular parietal cells in both the mouse and human stomachs The parietal cell is the main functional glandular cells that produces hydrochloric acid to kill swallowing microorganisms within the food in the stomach. To further identify some RIPK3-positive glandular cells were parietal cells, double immunofluorescence staining with H + K + -ATPase/RIPK3 antibodies were conducted in mouse and human gastric sections. Double immunofluorescence images clearly demonstrated that some of RIPK3-positive glandular cells were also positive for H + K + -ATPase (Fig. A–C for mouse gastric section and Fig. D–F for human gastric section). Phenotypic identification of RIPK3-positive cells in the lamina propria of human stomachs To examine the phenotypes of RIPK3 in the lamina propria of human stomachs, we performed an additional double immunofluorescence staining with different antibodies and were able to show that phenotypes of RIPK3-positive cells in the lamina propria were CD3-positive lymphocytes, CD34-positive stromal cells and SMA-alpha-positive stromal myofibroblasts in human gastric specimens (see Fig. ).
To estimate the turnover rate of glandular cells of gastric mucosa, we examined the proliferation activity labelled by PCNA and the expression of necroptosis key element RIPK3 with IHCs in mouse and human stomachs respectively. In the stomachs of FVB/N male mice, PCNA-positive cells could be observed in the surface mucous cells (arrowhead in Fig. A), neck cell region of the gastric fundic glands (arrow in Fig. A) and lamina propria cells (red arrow in Fig. A). Similarly, RIPK3-positive cells in the fundus of FVB/N male mice could also be detected in both the gastric mucous cells (arrowhead in Fig. B), gastric glandular cells (arrow in Fig. B) and lamina propria cells (red arrow in Fig. A). However, great majority of RIPK3-positive cells were within the mouse gastric glands, and predominately located in the isthmus and neck regions (Fig. B), and a few in the base regions. In addition, active phosphor-RIPK3-IR was shown in some human gastric glandular parietal cells (Fig. C). In human stomach specimens, both PCNA positive cells (Fig. D) and RIPK3 positive cells (Fig. E) were also shown in glands (arrow), pit (surface mucous area, arrowhead) and lamina propria regions (red arrow). Many RIPK3-positive cells were observed in the fundic glands with a parietal cell morphology (Fig. E). Since the fundic glandular cells play a key role in gastric function, we therefore counted the volume densities of PCNA- and RIPK3-positive cells and PCNA/RIPK3 ratio in fundic glands in this study. Figure showed the volume densities of PCNA-positive glandular cells in both FVB/N mice (Fig. A) and humans (Fig. B). In FVB/N mice, the volume density of PCNA-positive cells ( white bar in Fig. A) in fundic glands was 12.4 ± 1.6/gland. The volume density of RIPK3-positive cells in fundic glands was 6.0 ± 0.95/gland ( grey bar in Fig. A). Accordingly, the ratio of PCNA/RIPK3 in the gastric fundic glands of 6-month-old FVB/N male mice was ~ 2.12 (black bar in Fig. A). When we divided the FVB/N mouse gastric mucosa into high and low turnover areas according to the PCNA density and counted ratio of PCNA/RIPK3 respectively, it showed that the ratio was 2,40 and 1.65/field respectively, indicating the expression of RIPK3 was higher in the area with a high proliferation rate than that in area with low proliferation rate. In human stomachs, counting data of PCNA- and RIPK3-positive cell numbers in the fundic glands revealed that volume densities of PCNA ( white bar in Fig. B) and RIPK3 ( grey bar in Fig. B) were (21.38 ± 2.27) and (5.25 ± 0.45) respectively. Therefore, the ratio of PCNA/RIPK3 in the human gastric fundic glands was ~ 4.15 ( black bar in Fig. B).
The parietal cell is the main functional glandular cells that produces hydrochloric acid to kill swallowing microorganisms within the food in the stomach. To further identify some RIPK3-positive glandular cells were parietal cells, double immunofluorescence staining with H + K + -ATPase/RIPK3 antibodies were conducted in mouse and human gastric sections. Double immunofluorescence images clearly demonstrated that some of RIPK3-positive glandular cells were also positive for H + K + -ATPase (Fig. A–C for mouse gastric section and Fig. D–F for human gastric section).
To examine the phenotypes of RIPK3 in the lamina propria of human stomachs, we performed an additional double immunofluorescence staining with different antibodies and were able to show that phenotypes of RIPK3-positive cells in the lamina propria were CD3-positive lymphocytes, CD34-positive stromal cells and SMA-alpha-positive stromal myofibroblasts in human gastric specimens (see Fig. ).
Under the physiological condition, the population of gastric mucosal cells is maintained by the balance between cell loss and self-renewal. Disruption of this homeostasis will result in significant pathological changes of the stomach. Activation of RIPK3 in tissues is a hallmark of cells dying by necroptosis (Tonnus, ; Webster et al. ). In this study, we were able to show the densities of PCNA/RIPK3 positive cells in the gastric glandular cells of mice and human, providing a basic information regarding the proliferation/necroptosis in the gastric glandular cells. Our results (Figs. and ) clearly shown visible RIPK3 positive cells and a ratio of PCNA/RIPK3 positive cells in the gastric glandular cells, suggesting a necroptotic turnover rate in both the mouse and human glandular cells. Therefore, current studies of gastric glandular cell turnover rate may provide important information to understand gastric glandular cell homeostasis under physiological condition (Barker et al. ; Willet and Mills ; Ye et al. ). For example, cell death is an essential process for tissue development and homeostasis in human body as a way of eliminating infected, damaged or aged cells (Gunther et al. ), excessive cell loss by enhanced glandular cell death will induce atrophy or ulceration while excessive proliferation or prolonged cell life span can lead to hyperplasia (Jones et al. ). In this study, we have observed RIPK3 immunoreactivity in different types of gastric cells e.g. surface mucous cells, glandular cells, and lamina propria cells in both mouse and human stomach. Interestingly, we also showed that active phospho-RIPK3 was expressed in glandular cells at a low density. This observation indicates that some of RIPK3 could further go into the active form and finally induce glandular cell death. When we counted the ratio of PCNA/RIPK3 in high and low turnover gastric glandular regions respectively in the FVB/N mice, it showed that the expression of RIPK3 was higher in the area with a high proliferation rate than that in area with low proliferation rate. This finding suggested that the cell death might be elevated according to rate of proliferation, which could be important for the maintenance of an balanced cell turnover in the gastric mucosa. Because the parietal cell is the main gastric glandular cell and its main physiological function is to produce hydrochloric acid that can kill swallowed microorganisms within the food in the stomach. Therefore, to keep a balance of death/proliferation of parietal cells is critical for cellular and functional homeostasis. While the disrupted turnover of parietal cells under pathological conditions such as chronic H. pylori infection might result in an excessive parietal cell loss, which is a well-known glandular atrophic process and gastric cancer precursor, will significantly increase the risk for gastric cancer, Wang et al. have found that a high positive rate of RIPK3 in different types of gastric cells including the mucosal epithelial cells, smooth muscle cells and fibroblasts, while a negative for glandular epithelium in the stomach of mice (Wang et al. ). Such different findings could be resulted from the mice strains used in the studies. In their study, SPF BALB/c mice were used (Wang et al. ), and in our study FVB mice were used for the identification of necroptotic gastric cells. The RIPK3-positive cells located in the surface mucous cells can be easily identified by their location and morphology. To analyze the phenotypes of RIPK3 in other gastric cells such as glandular parietal cells, lamina propria cells, double immunofluorescence staining with different specific antibodies was conducted. We were able to show that RIPK3-positive glandular cells were mostly H + K + -ATPase positive parietal cells in both mouse and human. In lamina propria, RIPK3 was expressed by several types of lamina propria cells including CD3-positive lymphocytes, CD34-positive stromal cells and SMA-alpha-positive stromal myofibroblasts in human gastric specimens (see Fig. ). These findings suggested that necroptosis could be a common cell death manner and occurred in different cell types in the human stomachs. Finally, we have validated that some of the RIPK3-positive glandular parietal cells were also positive for phosphorylated RIPK3, indicating an activation of RIPK3 in parietal cells. The counting data showed that the ration of PCNA/RIPK3 positive cells/gland in human stomach is approximately 4.15, which was slightly higher than that the ratio (2.12/gland) in mice. This finding might suggest that the turnover rate of glandular cells in human stomach is slightly higher than that in mouse stomach. It was a surprise to find that the exact value of cell death in human or mouse gastric glands or cells were unavailable when we searched literature. So, it is impossible to compare our data with available publications. In addition, cell death might have several forms including necrosis in response pathological factors such as physics, chemistry, and pathogens and PCD. Necrosis usually occurs in response to pathological factors such as physics, chemistry, and pathogens; whereas PCD mainly including apoptosis and necroptosis is an active process of cell self-regulation and characterized by genetically controlled cell-autonomous ordered death. Interestingly, when researchers studied PCD in the gastric glandular cells, elevated elements related to PCD e.g. caspase-3 and caspse-8 under pathological conditions are frequently found (Bockerstett et al. ; Cui et al. ; Le’Negrate et al. ). However, PCD cells with significant morphological changes are rarely observed in various tissues under physiological circumstances (Poon et al. ). Indeed, our results of PCNA/necrosis ratio in mouse sections showed that it is ~ 0.032/field, which is much lower than the ratio of PCNA/RIPK3. The possible reasons for such a low rate of PCNA/necrosis in the normal stomach might be: 1. The removal rate of cell undergoing PCD is very high under physiological conditions and dying cells are quickly removed either by tissue-resident professional or neighboring non-professional phagocytes and immature dendritic cells through a coordinated manner of multiple steps (Poon et al. ; Ravichandran ); 2. Not all the cells with high RIPK3 expression will develop into active form (RIPK3 → phosphorylated RIPK3) that could induce the process of necroptosis finally. These reasons make the accurate measurement of PCD rate in gastric sections by counting dying cells with morphological changes in histological sections very difficult under physiological circumstances (Poon et al. ; Ravichandran ). Recently, Karam has summarized the literature regarding the renewal capacity of parietal cells in mice (Karam ). Their review article indicates that old parietal cells will undergo degeneration and elimination after an average lifespan of about 54 days, and new parietal cells will be developed directly from the stem-like pre-parietal cells through a mature process (Karam ). The homeostasis of parietal cells is kept by the balanced rate of renewal and death. Revealing the dynamic features of parietal cell renewal/death can help in a better understanding of the structural integrity and physiological functions of the gastric glands. For example, when the acid secretion amount should be at a normal amount when the death/proliferation of parietal cells is normal, because the population of parietal cells will be at a certain stable level. On the contrary, the excessive death of parietal cells can lead to a reduced density of parietal cells, and further result in a decreased acid secretion function. Clinically, long-term over-suppression of parietal cell acid secretion function by administration of proton pump inhibitors could cause an overgrowth of bacterial in the intestine. Thus, our current results may have a particular clinical significance that increase of necroptosis is involved in the process of acid-secreting parietal cell atrophy under pathological conditions such as H. pylori infection. Previous studies have demonstrated that enhanced gastric epithelial cell apoptosis, as one of the major cell PCD manners, is observed during H. pylori infection (Chen et al. ; Jones et al. ), which is significantly associated with the development of chronic gastritis, peptic ulcers, and gastric carcinogenesis (Rogers et al. ; Targa et al. ; Xia and Talley ). We have demonstrated that chronic Helicobacter infection and induced hypergastrinemia could lead to increased apoptosis in glandular parietal cells and contribute to the development of gastric cancer in Helicobacter infected mice (Cui, ; Cui et al. ). Burclaff and colleagues (Burclaff et al. ) have developed a mouse model that expresses the diphtheria toxin receptor specifically in parietal cells to induce their death and found this to increase proliferation in the normal stem cell zone and neck area. In addition, several proinflammatory cytokines such as TNF-α and interleukin (IL)-17A have been shown to promote the parietal cell atrophy and metaplasia in isolated parietal cells or mouse models (Bockerstett et al. ; Neu, ), whether the induction of atrophy is associated with the necroptosis induced by these cytokines would be an interesting future topic. Furthermore, it is postulated that necroptosis, as a novel PCD manner, may be involved in the gastric cell death induced by H. pylori infection. Indeed, Radin et al. have demonstrated that Helicobacter toxin Vac A could induce necroptosis in epithelial cells in vitro (Radin et al. ). Therefore, it is worth studying in the future the role of necroptosis in chronic H. pylori infection-induced parietal cell atrophy. In conclusion, our present study has revealed that a high rate of RIPK3 expression in different cell types, particularly in glandular cells, might contribute to cell turnover of gastric mucosa in the mouse and human stomach under physiological condition.
|
Current status and future developments in predicting outcomes in radiation oncology | 861df0b3-bb5f-48c4-aa85-4a3c05b41cb8 | 9793488 | Internal Medicine[mh] | The past decade has witnessed rapid advancements in data-driven technologies that have influenced virtually every field in medicine including treatment response modeling and outcomes prediction in radiation oncology. Advanced big-data analytics have facilitated the integration of information-rich multiomics data with the traditional clinical and dosimetric information for modeling radiotherapy (RT) treatment responses. Inclusion of such multiomics information, which is the aggregation of genomics, proteomics, transcriptomics, metabolomics, epigenomics, and radiomics, among others, as prognostic or predictive biomarkers has improved the potential of such outcome models in personalizing treatment management and achieving the promise of precision medicine in radiation oncology. These improvements in data and algorithmic modeling performances can be attributed to the enhanced ability to capture underlying cancer heterogeneity caused by subtle interhuman genetic and physiological differences. The inclusion of multiomics data can effectively improve the representation of patient-specific radiosensitivity beyond traditional dose–volume metrics used for modeling the observed tumor control probability (TCP) and/or normal tissue complication probability (NTCP), which have dominated the field of radiation oncology over the past decades. Traditionally, outcome prediction models (OPMs) in RT have relied on a simplistic representation of dose response using analytical models such as the linear quadratic (LQ) model. Though such models are useful for understanding the dose–volume effect at a population level and has been used for fractionation conversion purposes in RT treatment planning, they have limited ability to address treatment management requirement at the patient level. In the modern era of precision medicine such one-hat-fits-all approach is not acceptable anymore. Towards meeting the requirements of precision medicine, modern OPMs are envisioned to be utilized in constructing semi- or fully automated clinical decision support systems (CDSS) that could be applied for personalized and adaptive radiotherapy (ART). Such a CDSS is depicted in . In this case, the CDSS evaluates patient’s dose response by analyzing trends in patients’ multiomics biomarkers measured before and during treatment, which are fed into an appropriate OPM for projecting outcome estimates. The CDSS then can recommend an optimal dose adaptation value for the remaining treatment period that will maximize the observed TCP and minimize the observed NTCP. This recommendation could be made via a sequential decision-making algorithm such as deep reinforcement learning. Note that in modern era of OPMs, the notion of TCP/NTCP is generalized from simplistic dose–volume metrics into more comprehensive multiomics patient-specific approach that utilizes artificial intelligence (AI) as its computational vehicle. It is recognized that advanced data-driven approaches fueled by AI-based machine and deep learning (ML/DL) techniques have significantly improved the predictive power of OPMs. However, researchers have been actively working on overcoming two main impending challenges related to data and algorithmic modeling. This will further enhance OPM’s utilization and crossing the bridge of so called the ‘AI chasm’ that separates development from clinical implementation. summarizes the data modeling related challenges of interpatient heterogeneity, limited data set, and high level of uncertainty in data, and the algorithmic modeling related challenges of model performance, uncertainty, and model interpretability. Accessibility of sufficiently large data sets and availability of noise-free data required for clinical-grade robust model development and evaluation are currently lacking. While the underlying reasons are not mutually exclusive, the former issue, which creates a barrier for implementing ML/DL big-data analytics, is associated with a set of privacy laws established for protecting patient’s rights. The latter issue arises from several reasons including missing data, measurement error, and interpatient heterogeneity. Currently, researchers have been actively involved with federated learning (FL) to tackle the sample size requirements issues, and application of human-in-the-loop (HITL) learning & quantum computing to overcome the uncertainty issues. Additionally, preventive measures such as standardizing the data acquisition and data harmonization methods are other actively ongoing efforts for improving the data quality for outcome modeling. FL is an emerging paradigm in the fields of ML/DL algorithm development in which OPMs are trained across multiple decentralized servers, located in multiple institutes. The advantage of FL is that at one hand the data doesn’t need to leave the premise of the institution. On the other hand, this paradigm provides the necessary sample size for accurate and robust statistical modeling while protecting the data privacy. While there are numerous works on FL in other fields, it is an active area of research and development in radiation oncology and in the medical field in general, mainly because, unlike ML/DL platforms ( e.g. pytorch or tensorflow), the infrastructures for FL are still lacking. Jochems et al, in their proof-of-concept study, developed a tool for survival prediction of non-small cell lung cancer (NSCLC) patients treated with chemoradiation or RT that was trained in data set located in two different cancer institutions and then further validated the model in another institution. Those institutions were located in three different countries in two different continents. HITL is a hybrid of data- and knowledge-driven approaches that integrates prior expert knowledge into ML frameworks. The synergy between machine and human intelligence helps to alleviate model uncertainty, improve model trust level (credibility), and build upon current understanding as opposed to starting from scratch. Luo et al have implemented HITL for NSCLC OPM, where a Bayesian network architecture was applied to a multi omics dataset for making predictions of local control (LC) and radiation induced lung inflammation (pneumonitis (RP)) as shown in . The study integrated known biophysical interaction between the patient features into modeling without which the task would have been extremely difficult and prone to uncertainty given the limited sample size and a large feature size. Similarly, Sun et al have designed a new framework for integrating expert human knowledge with AI recommendations for optimizing clinical decision-making in ART in lung cancer. Quantum computing and quantum information theory are a natural fit for dealing with data uncertainty, stochasticity, and noise. In addition, application of quantum information can significantly speed up computation compared to its classical counterpart. Because of its advantages, quantum computing has been applied to ML algorithms and with the rapid development of quantum computing platforms, quantum ML algorithms are actively being researched. Niraula et al have modeled clinical decisions as quantum states to represent the uncertainty faced by physicians in decision-making during RT treatment. The uncertainty in decision-making mainly arises due to the availability of partial information on patient’s state and due to the uncertainty in treatment outcomes. In another example of application of quantum computing, Pakela et al demonstrated that quantum tunneling based annealing optimization techniques can optimize intensity-modulated radiotherapy treatment plan much faster than the traditional simulated annealing optimization method. A faster optimization algorithm would permit in reoptimizing a treatment plan in a shorter time frame during the treatment course to account for changes such as tumor shrinkage and organ deformation to optimize the treatment outcome. Aside from data-related issues, the interpretability of advanced predictive ML/DL models is also a concern to the medical community. DL models, in particular, are popular for their high degree of accuracy. However, because they are complex non-linear models composed of up to billions of parameters with non-convex objective functions, interpreting DL models is not always easy. In general, complex models trade-off accuracy for interpretability. Usually, complex models are regarded as a black box, whose architecture (hyperparameters) are optimized with respect to the task in hand and are utilized without delving deeper into the how’s and why’s. While such practices are usually acceptable in non-medical fields, medical conservatism emphasizes the need for interpretable models as an acceptable clinical tool due to safety issues and legal compliances. Researchers in radiation oncology are integrating interpretability methods into their predictive models. Wei et al implemented an integrated gradient method for identifying the important radiomics features from a DL model trained for risk prediction of overall survival in hepatocellular carcinoma patients. We begin this review with an overview of the traditional mechanistic predictive outcome models, primarily based on the linear quadratic model for estimating TCP and NTCP. Then, we present the current status of predictive modeling, which is mainly focused on the inclusion of multiomics patient-specific information followed by a CDSS framework that leverages such multiomics based OPM for optimizing RT treatment plans. We finish with a discussion on the current challenges and limitations of predictive modeling associated specifically with the AI data-driven technologies and actively sought out solutions that could potentially overcome the current gap (AI chasm) that separates development from clinical implementation.
Two major types of clinical end points considered in outcome modeling are TCP and NCTP. TCP is defined as the probability that a tumor is eradicated or controlled after receiving a certain amount of dosage and NTCP is defined as the probability of radiation-induced normal tissue toxicities that an organ of interest may exhibit after exposure to unwanted radiation. TCP-related end points may include LC, regional control, etc. NTCP-related end points are more diverse and may vary among disease sites. Note, in our description, below we are using TCP/NTCP in their generalized sense as comprehensive OPM with multiomics patient-specific input data. Conventional analytical outcome models are mainly based on dosimetric and volumetric information. These models may formulate the outcomes according to simplified radiobiology theory or available dosimetric data to a parametric model. Many analytical (mechanistic and/or phenomenological) models have been proposed in the literature. Here, we present a brief description of the common analytical outcome models used in the clinical treatment planning software. For the interested reader, more in-depth analysis of these models and their application can be found in TCP modeling: The well-known linear-quadratic (LQ) model and its variants are based on irradiation effects observed from in vitro cell culture experiments. In a LQ model, the logarithm of survival fraction (SF) of cells is composed of both linear and quadratic terms of the physical dose, relating to lethal damage caused by the ‘single-hit’ and ‘multiple hit’ events, respectively. It is possible to use the LQ model to convert the size of fractionation dose while keeping an iso-effect of the biological response for the end point of interest. The biologically effective dose (BED) concept could be used to convert dose under different fractionation schemes to a standard fractionation scheme. For instance, a quantity called EQD2 is a special case where dose under different fractionation schemes is converted into equivalent total dose given in 2 Gy per fractions. TCP models can be also built by fitting empirical data, for instance, logistic regression can be adopted to associate dosimetric variables to TCP. NTCP model: The Lyman model is a well-known NTCP model, which uses a cumulative Gaussian distribution to fit the toxicity events of interest, N T C P D , D 50 , m = 1 2 π ∫ - ∞ t exp ( - 1 2 u 2 ) d u , where, t = D - D 50 m D 50 . Here, D could be represented by the generalized equivalent uniform dose (gEUD), D 50 is a dose uniformly irradiating the whole organ that relates to 50% NTCP, and m is a parameter that controls the slope of NTCP. Additional parameters regarding the volume effect can be further incorporated if the dose distribution is inhomogeneous. Other sigmoidal functions, e.g . logistic function, log-logistic cumulative distribution have also been applied to model NTCP. Note that the TCP and NTCP models differentiate between tissue types via a radiosensitivity parameter denoted as the α / β ratio. This ratio is central to BED and EQD2 estimation and to subsequent calculation of both TCP and NTCP models. In practice, the α / β ratio is not well characterized for all tissue types, including organs at risk (OARs), which can introduce additional uncertainty in the TCP and NTCP modelling process. Predictability of the conventional analytical OPM is limited to the dosimetric information. OPM can be improved by basing it on other information such as imaging and multiomics data. However, due to the sheer number of feature size, hand crafting an analytical model is very difficult, if not impossible. Thus, data-driven techniques such as statistical and ML have been utilized for OPM. In the general sense, outcome modeling can be typically defined as a supervised learning problem in the context of ML. A mapping function from patient-specific information to the end points of interest can be learned using the training data. Then, this function can be applied to an unseen data set to infer clinical end points for a given input, patient-specific data. Outcome modeling could be presented as a classification problem when a pre-set of cut-off thresholds is used to determine the end point. Otherwise, an outcome model can apply to predict the probability of events directly using regression techniques. The combination of growth in patient-specific multiomics data, complexity of radiation response process, and advances in ML algorithms, particularly the success that DL algorithms that have demonstrated in the field, have led to the burgeoning interest in applying ML/DL methods to outcome modeling in RT. Current status of outcome modeling in radiotherapy Currently, many treatment planning systems still utilize analytical models based on the LQ models and its variants as add-ons which may or may not be consulted during planning. The focus remains on meeting dose–volume physical constraints rather than achieving desired outcomes based on radiobiological models. This is primarily due to uncertainties associated with these models, their population-based nature, and the scarcity of proper optimization methods that could utilize such models. Recently, there has been extensive efforts and advances made in personalizing outcome modeling for RT. It is recognized that outcome prediction in radiation oncology is multifactorial and involves correctly identifying patients’ radiosensitivity of tumor cells and surrounding healthy cells. Identifying radiosensitivity, however, is not an easy feat, because it depends on many dynamic factors such as tumor heterogeneity, tumor volume, hypoxia, cell damage repair, cell cycle, etc., that can change over the treatment period. Moreover, radiosensitivity differs from patient to patient. A possible way to include patient-specific radiosensitivity in outcome modeling is by integrating patient-specific multiomics information. Advancements in imaging acquisition technologies have gifted us the field of radiomics in which large-scale imaging information can be incorporated into modeling RT responses. Currently, there are two areas associated with radiomics, one is based on hand crafted features ( e.g. morphology and texture) and the other utilizes raw images directly using DL. Moreover, measuring and quantifying molecular biomarkers, such as gene expression microarrays and RNA sequencing methods, and data-driven techniques that can process large amount of genetic data, have transformed the predictive modeling landscape from a population based one with limited predictive power into a data-driven one that allows for individualized treatment management with ML modeling. Advanced ML/DL models can now leverage before and during radiation treatment information on patient’s relevant radiogenomics biomarkers that are extracted from tumor biopsy, blood work, and imaging radiomics, to capture the intrinsic heterogeneity and accurately incorporate patient-specific radiosensitivity in OPM. For instance, Luo et al showed improvement in prediction accuracy by including multiomics information of NSCLC patients as predictive biomarkers for LC and RP. They analyzed hundreds of biophysical features including dosimetry, clinical, pre- and during treatment cytokines, miRNAs, as well as single nucleotide polymorphisms (SNPs), and selected the most important features for predicting the outcomes. They then developed an OPM based on Bayesian Networks that presented the causal relationship between the features and the outcome. Similarly, Cui et al developed a DL-based prediction model that jointly predicted LC and RP in NSCLC patients. Patient-specific information including differential dose–volume histograms, PET tumor radiomics, and biological information were input into a composite deep neural network. The composite architecture was able to learn the representation of features and perform the prediction task simultaneously. Tseng et al designed a deep reinforcement learning-based framework for CDSS in ART. By inputting patient’s state information, comprised of important multiomics features, the CDSS can recommend an optimal dose fractionation. In that work, for a given state and an amount of dose fractionation, an artificial radiotherapy environment that was constructed out of OPM could predict a patient’s resulting state and corresponding treatment outcome. By using that environment, a deep reinforcement learning network was trained to recommend optimal dose fractionation that would maximize TCP and minimize NTCP. Niraula et al have further advanced the work by improving the model of artificial radiotherapy environment. They integrated prior knowledge on dosimetric features into the OPM and combined neural networks with mechanistic models to ensure that the dosimetric features respected the radiological physical requirements, i.e. the gEUD values increased with increasing dose fractionation and both TCP and NTCP values increased with increasing gEUDs. Future modeling processes will witness increase in data types as well as integration of prior expert knowledge into data-driven methods. Additionally, deep reinforcement learning methods have also been used for precision oncology, 2D tumor growth simulation-based dose fractionation, and beam reshaping via multileaf collimation for real-time adaptive RT.
Currently, many treatment planning systems still utilize analytical models based on the LQ models and its variants as add-ons which may or may not be consulted during planning. The focus remains on meeting dose–volume physical constraints rather than achieving desired outcomes based on radiobiological models. This is primarily due to uncertainties associated with these models, their population-based nature, and the scarcity of proper optimization methods that could utilize such models. Recently, there has been extensive efforts and advances made in personalizing outcome modeling for RT. It is recognized that outcome prediction in radiation oncology is multifactorial and involves correctly identifying patients’ radiosensitivity of tumor cells and surrounding healthy cells. Identifying radiosensitivity, however, is not an easy feat, because it depends on many dynamic factors such as tumor heterogeneity, tumor volume, hypoxia, cell damage repair, cell cycle, etc., that can change over the treatment period. Moreover, radiosensitivity differs from patient to patient. A possible way to include patient-specific radiosensitivity in outcome modeling is by integrating patient-specific multiomics information. Advancements in imaging acquisition technologies have gifted us the field of radiomics in which large-scale imaging information can be incorporated into modeling RT responses. Currently, there are two areas associated with radiomics, one is based on hand crafted features ( e.g. morphology and texture) and the other utilizes raw images directly using DL. Moreover, measuring and quantifying molecular biomarkers, such as gene expression microarrays and RNA sequencing methods, and data-driven techniques that can process large amount of genetic data, have transformed the predictive modeling landscape from a population based one with limited predictive power into a data-driven one that allows for individualized treatment management with ML modeling. Advanced ML/DL models can now leverage before and during radiation treatment information on patient’s relevant radiogenomics biomarkers that are extracted from tumor biopsy, blood work, and imaging radiomics, to capture the intrinsic heterogeneity and accurately incorporate patient-specific radiosensitivity in OPM. For instance, Luo et al showed improvement in prediction accuracy by including multiomics information of NSCLC patients as predictive biomarkers for LC and RP. They analyzed hundreds of biophysical features including dosimetry, clinical, pre- and during treatment cytokines, miRNAs, as well as single nucleotide polymorphisms (SNPs), and selected the most important features for predicting the outcomes. They then developed an OPM based on Bayesian Networks that presented the causal relationship between the features and the outcome. Similarly, Cui et al developed a DL-based prediction model that jointly predicted LC and RP in NSCLC patients. Patient-specific information including differential dose–volume histograms, PET tumor radiomics, and biological information were input into a composite deep neural network. The composite architecture was able to learn the representation of features and perform the prediction task simultaneously. Tseng et al designed a deep reinforcement learning-based framework for CDSS in ART. By inputting patient’s state information, comprised of important multiomics features, the CDSS can recommend an optimal dose fractionation. In that work, for a given state and an amount of dose fractionation, an artificial radiotherapy environment that was constructed out of OPM could predict a patient’s resulting state and corresponding treatment outcome. By using that environment, a deep reinforcement learning network was trained to recommend optimal dose fractionation that would maximize TCP and minimize NTCP. Niraula et al have further advanced the work by improving the model of artificial radiotherapy environment. They integrated prior knowledge on dosimetric features into the OPM and combined neural networks with mechanistic models to ensure that the dosimetric features respected the radiological physical requirements, i.e. the gEUD values increased with increasing dose fractionation and both TCP and NTCP values increased with increasing gEUDs. Future modeling processes will witness increase in data types as well as integration of prior expert knowledge into data-driven methods. Additionally, deep reinforcement learning methods have also been used for precision oncology, 2D tumor growth simulation-based dose fractionation, and beam reshaping via multileaf collimation for real-time adaptive RT.
In addition to prospective evaluation, the following challenges needs to be addressed. 1.Overcoming limited data 1A. Federated Learning The curse of dimensionality, as coined by Richard E. Bellman, demands an exponentially large sample size for a high dimensional data set. This occurs since the volume of the feature space expands rapidly with the increase of the feature dimensions, which means that to be able to gain from high-dimensional data sets such as multiomics, we will need information from exponentially more patients. The small sample size to feature size ratio is even more pronounced in the medical field where data points are sparsely distributed and limited due to ethical and legal issues associated with human experimentation. To make the matter worse, data-sharing in healthcare is strictly regulated due to privacy and patient protection concerns. FL can aid in overcoming these data access limitation issues and has been actively pursued by the medical community including radiation oncology. In FL, data do not leave the secure confinement of the data owner’s firewall but let a foreign model securely access the data and learn from it. Several network topologies have been proposed for FL, such as centralized, decentralized, hierarchical, hybrid hierarchal, etc. For instance, in a centralized topology, a central server broadcasts multiple identical models to distant data servers which allows the models to access their local data set. The local models then calculate local loss function and send it back to the central server. The central server then collects the local loss functions and then aggregates it in a weighted fashion where the weights are selected according to the characteristics of the respective data sets. Then, the model is updated in the central server and the updated models are again broadcasted for another iteration. Models trained under FL have shown to achieve comparable performance to the ones trained under traditional centralized ML. FL has many other benefits such as overcoming selection bias due to accessibility of data from a wide range of regions and demography, the ability to perform comparative studies of subgroups of patients that might have different radiosensitivity, high quality ML aided tool for diagnoses and outcome predictions, etc. There are also concerns and challenges, such as having to deal with heterogenous data sets that are not independently and identically distributed (non-iid) due to interinstitutional differences in RT clinical protocols, data acquisition methods, and differences in computing hardware resources. However, considering the current momentum and early success of the ongoing research, it is safe to be optimistic about the prospects of FL. 2. Overcoming uncertainty issues In the domain of radiation oncology, data are usually collected under heterogenous conditions that lead to a noisy data set. There can be many sources of data noise such as interference, calibration error, sparse data set, small data set, measurement error, discretization, batch effect, intrinsic heterogeneity in humans, etc. A significant degree of uncertainty can arise from record-keeping related issues, such as inconsistencies and difficulty in classifying radiation induced toxicities and reporting outcomes, even with a standard Patient-Reported Outcome Measures (PROM’s) questionnaire. Any data-driven outcome model based on a noisy data set will have larger prediction uncertainty. Another source of uncertainty in RT comes from clinical decisions. Since physicians must prescribe decisions without a complete knowledge of confounding patient’s dose response factors, their decisions may have a high degree of uncertainty which gets propagated into the recorded clinical data. Currently, there are ongoing research efforts in the following two areas for dealing with data noise and modeling uncertainty. 2A. Human in the loop Being rooted in reinforcement learning, preference learning, and active learning, HITL is a hybrid of data-driven and knowledge-driven approach that overcomes model uncertainty by integrating prior expert knowledge into ML frameworks. Holzinger et al defined HITL as ‘‘algorithms that can interact with both computational agents and human agents to optimize their learning behavior through these interactions. ’’ HITL approach can reduce the complexity of a hardest decision-making problem (NP-Hard) through assistance of a human agent in the learning phase. Furthermore, experimental evidence for the utilization of HITL algorithm has demonstrated that human intelligence can positively augment machine intelligence. The concept of HITL has been intensively used to tackle obstacles created by data-related limitations. By exploring human–machine partnership with AI, Patel et al demonstrated that AI-based technology achieved superior accuracy than the human experts alone in chest radiograph diagnosis. In other study of HITL for interplaying between digital image analysis (DIA) and pathologists, Boden et al found that HITL corrections could address major DIA errors in terms of poor thresholding of faint staining and incorrect tumor-stroma separation for individual cases. In OPM for NSCLC, Luo et al implemented HITL learning for creating a causal graph of feature and end points via Bayesian network architecture. The OPM took in multiomics data set for making predictions on LC and RP as shown in . As ML/DL plays an increasingly important role in medical image analysis, the integration of HITL and ML/DL becomes even more necessary for designing CDSS. Budd et al evaluated four key areas to improve the integration in clinical practice: active learning for the best data selection, interaction with model outputs for models steering, full-scale applications before deployment, and the evolvement of knowledge gaps to benefit HITL computing. However, it is still unclear how to design optimized relationships between people and machines in a scalable manner, how to design triggers for proactive engagement and disengagement, and how to handle the consequences of implied interventions. The verification of treatments in a large HITL machine medical system can be very complex due to its evolving nature from both patients’ aspects and the clinical environment. Therefore, it is critical to understand human–machine interaction semantics and control for the behavioral context in designing OPM with HITL. 2B. Quantum computing and information theory Quantum computation, built on quantum information theory, is intrinsically indeterministic, making it a natural medium for modeling noise and uncertainty. Unlike digital computers, where a value is deterministically represented by a collection of bits, a quantum system represents a value as the mode of a probability distribution. By representing the data as a quantum state, we are modeling the noise by the spread of the distribution. Furthermore, by carefully designing quantum ML algorithms, we can model the interaction between the patient’s state and the radiation dose, which would automatically encapsulate the propagation of uncertainty and naturally present the outcome prediction with an uncertainty metric. Researchers are working on such a complete quantum model. Recently, Pakela et al designed a hybrid quantum deep recurrent neural network (RNN)-based OPM for predicting tumor volume and daily setup changes in head and neck cancer patients. Just as it is impossible to predict the behavior of a quantum system with absolute certainty, there are inherent uncertainties in knowing a patient’s “state” during RT due to physiological changes such as tumor response and anatomical motion as well as limitations in cone beam CT (CBCT) image quality. The OPM modeled the combined patient tumor volume and daily setup changes as a stationary quantum state whose time points were discretized by the number of fractions. The authors demonstrated that such a hybrid model could perform almost as well as the classical Markov-based model. The performance, measured in AUC scores, of the predictive Markov model compared to the hybrid quantum model, evaluated in an external validation data set, were 0.707 vs 0.623, 0.687 vs 0.608, 0.723 vs 0.669, and 0.697 vs 0.609 for patient’s discrete states sizes of 4, 6, 8, and 10, respectively. Hybrid quantum algorithms have also been explored for beamlet-weight optimization in intensity modulated radiotherapy (IMRT) treatment planning. Pakela et al developed a hybrid optimization algorithm by merging annealing and quantum tunneling. The quantum tunneling annealing optimization schedule reached convergence up to 46.6% faster than traditional simulated annealing algorithm for beamlet intensity optimization and up to 26.8% faster for direct aperture optimization. Quantum computing can also be used to model decision-making. Niraula et al designed a hybrid quantum deep reinforcement learning algorithm that modeled human decisions as an indeterministic quantum state. By training the framework in IBMQ quantum computer, they demonstrated the feasibility of the quantum framework as a CDSS which can potentially aid physicians in the clinic. Following a self-evaluation metric based on the retrospective clinical outcome, they showed that a decision-support system based on their framework can potentially improve the clinical decision by 10%. 3. Model interpretability Making accurate prediction requires utilization of complex models with many parameters. However, complex models are difficult to interpret precisely. This trade-off between accuracy and interpretability is troublesome especially in high-risk fields such as radiation oncology. On the other hand, without fully understanding the relationship between the features and the outcome prediction, it is very hard to differentiate authentic relations from artifacts. Therefore, currently, extra effort is being put into developing interpretable models and tools for model interpretation/explanation. Interpretable graph networks such as Bayesian networks are being actively applied for outcome modeling which can trace the internal relationships between the features and their combined relationship with the outcome. Interpretability tools are also being actively developed and incorporated into outcome prediction models that are built on black-box type models such as tree-based and deep learning models. Several methods are available for interpretation that can be categorized into model-agnostic methods such as feature importance, sensitivity analysis via input perturbation, local attention mechanisms, Shapely values, etc., and model-specific methods such as feature visualization, saliency maps, adversarial examples, etc. Model-agnostic methods are especially desirable because they can help in comparing different models based on the feature to outcome relationship. Model-specific methods, especially in the context of neural networks, are desirable to extract learned features from the hidden layers and are usually much faster than the model-agnostic methods. Saliency mapping tools such as Decovnet, Grad-CAM, Guided Grad-CAM, and SmoothGrad are popular interpretation tools for CNNs and have been applied in radiomics and imaging in radiation oncology. Liang et al developed a CNN-based prediction model for RP that utilized dose distribution as an input feature. There, the authors used Grad-CAM to locate regions of the dose distribution that strongly correlated with the positive and negative cases of RP. Similarly, Cui et al applied Grad-CAM method to provide insight into what was learned by the composite model and thereby increased their model’s interpretability. Wei et al applied a combination of variational autoencoders (VAEs) and CNN to identify a new radiomics signature using imaging phenotypes and clinical variables for risk prediction of overall survival in hepatocellular carcinoma patients. The model included two VAE survival components for the clinical and radiomics features and a CNN survival network for the contrast-enhanced image input. For model interpretation, integrated gradients methods were applied which helped in identifying the clinical liver function and liver exclusive of tumor radiomics features as the top-ranked features for risk prediction. An example image showing the interpreted outputs using integrated gradients is shown in . Normal liver tissues were given more attention by the model for high-risk patients, which is consistent with the fact that top-ranked features were liver-GTV texture features.
Currently, predictive modeling in radiation oncology utilizes patients’ multiomics information along with the traditional dosimetry and clinical information. Automated decision-support systems for adaptive RT have been developed utilizing multiomics-based predictive modeling. The current data-driven OPMs performs better than their traditional mechanistic models due to enriched patient-specific information. However, data modeling-related issues have been hindering the deployment of these models to achieve their full potential. Significant effort is being made to overcome these issues with a combination of emerging technologies such as FL, HITL, quantum computing, and novel model interpretation tools that were highlighted in this review. Prospective evaluation of OPM is necessary for clinical application.
|
Percutaneous sacroiliac screw fixation with a 3D robot-assisted image-guided navigation system | 7fb090fc-1437-45f6-9005-5c76eb8d86dc | 11790701 | Surgical Procedures, Operative[mh] | Percutaneous screw osteosynthesis has emerged as the procedure of choice for fixing the posterior pelvic ring due to its minimally invasive nature , lower risk of infection , rapid recovery precise screw placement , and cost-effectiveness. First described in 1995 by Routt, this technique was initially challenging with high complication rates, mainly due to screw misplacement, nerve injuries, or screw failure (0–33%) . But, over the years, this technique has improved, and complication rates are decreasing . One of the most challenging aspects of the percutaneous screw fixation is intraoperative imaging . Weak contrast due to decreased bone density, bowl gas overlay, and symphysis superposition can limit imaging quality . Implementing a 3D robot-assisted image-guided navigation system might provide a reliable solution. By calibrating the 3D scan prior to surgery, the surgeon can easily switch between personalized inlet, outlet, and lateral views of the pelvis, ensuring consistency and efficiency throughout the procedure. Moreover, 3D robot-assisted image-guided navigation system allows for precise planning and intraoperative guidance of screw placement. Furthermore, intraoperative 3D scans can be obtained to confirm correct screw positions and allow for immediate correction of any misplacements. While the advantages of this 3D robot-assisted image-guided navigation technology extend to all pelvic fractures, we have chosen to focus on its application in fragility fractures of the pelvis (FFP) due to the increased benefits it provides to this frail patient population. The paper aims to present a standardized surgical technique and to address technical challenges and possible solutions. We share our clinical experience with 3D robot-assisted image-guidance in the surgical treatment of FFP using percutaneous sacroiliac screws.
The objective of percutaneous sacroiliac screw osteosynthesis of the pelvis is to achieve stabilization with minimal soft tissue dissection and short operation time. The use of a robot-assisted 3D image-guided navigation system allows for precise execution of the preoperative plan by intraoperative navigation of the screw path with the help of individually definable and exactly reproducible imaging planes by means of a high-resolution 3D scan.
Short operation duration Minimal blood loss Limited soft-tissue dissection Precise determination, guidance, and control of the screw position Early patient mobilization especially in frail patients Reduced technical limitations in obese patients Providing maximal contrasted x‑ray in osteopenic patients or bowl gas overlay
Requires access to a hybrid operating room Patient movement during the procedure may lead to loss of matching with the navigation system Limited use of needle guidance in the anterior pelvic ring due to potential movement during posterior screw insertion and greater distance to the table Learning curve for handling the hybrid operating room Longer preparation, including draping and matching 3D scan
Fragility fractures of the pelvis type II, III, and IV High-energy pelvic fractures
The patient is not fit for surgery according to anesthesiology assessment Infection at site of screw insertion
General risks of surgery Screw loosening over time Injury to the nerves (nerve roots S1, S2, and L5) Injury to the vessels (branches of the superior gluteal artery)
Perform a clinical examination of the pelvis, focusing on compression pain and tenderness over the sacrum and pubic rami. Conduct a clinical neurological examination of the lower extremities. Utilize a diagnostic CT for detection and classification of the fracture (appendix 1). Conventional X‑rays carry a high risk of misclassification . Employ the diagnostic CT to measure the angles for the personalized inlet and outlet view, which are critical aspects of the presented method (Fig. ). A comprehensive explanation of the personalized inlet and outlet angles is illustrated in appendix 2
Standard surgical instruments for osteosynthesis (scalpel, long pean, light mallet, screwdriver, power drill with guidewire adapter, forceps, needle holder) Long 2.8 mm guidewires Changing canula with hexagonal ending for secure insertion into the screw head 7.3 mm cannulated screws with washer If cement augmentation is considered: 7.5 mm fully threaded, cannulated, and fenestrated screws with integrated washer and augmentation kit with slow polymerizing polymethyl methacrylate (PMMA) bone cement Optional: non-cannulated 3.5 or 4.5 mm or cannulated 7.3 mm fully threaded screws
General anesthesia or compliant patient with spinal anesthesia All patients receive a perioperative antibiotic prophylaxis (cefazolin 2 g single shot) Supine positioning allows for fixation of the anterior and posterior pelvic ring and reduces anesthesiologic challenges. Prone positioning is also possible if the anterior pelvic ring is addressed in a retrograde manner. The patient should be placed in the center of a radiolucent table, pelvic tilt and/or rotation should be avoided Arms are positioned in front of the patients face and positioned in an arm holder device Disinfection area includes the visible abdomen from the navel to mid thighs. Draping consists of a drape deeply tucked beneath the patient on either side, a self-adherent U‑drape starting just above the confluence of the major labia or the base of the penis, and a self-adherent drape covering half the navel and above. The loose ends of the drapes are collated to the inferior part of the radiolucent table to allow for free movement of the C‑arm (Fig. ). After definitive patient positioning and draping, it must be tested that the C‑arm can move freely around the patient. It is important that the patient lays completely still, and no position adjustments are made after the 3D scan.
(Figs. , , , , , , , and )
For standardized positioning and collision-free movement, we recommend floor markings for all mobile devices, such as infusion stand, ventilator machine or suction machine. The described sacroiliac screws and retrograde superior ramus screw may be combined with transiliosacral screws, sacral bars or sacroiliac screws in the S2 body. This technique may be combined with open procedures and is highly adaptable. The typical personalized outlet angle is approximately 40°. If this angle is not achievable intraoperatively due to potential collision of the C‑arm with the table, the highest possible angle should be accepted. The procedure becomes less complex with an increased difference between the angels of the personalized inlet and outlet views. Ideally, this difference would be 90°, as changes in one plane do not necessarily result in changes in the other. However, typically the difference between angles is significantly lower than 90° (also see appendix 2 ). The personalized inlet and outlet angle, described in appendix 2 , are especially of advantage in elderly patients with dysmorphic pelvic characteristics . In older patients with more elastic skin and soft tissues, guidewire repositioning is relatively easy compared to younger patients with firm skin, who may require pullback beneath the skin for redirection. Penetration of the washer into the ilium should be avoided because it typically causes persisting pain. Even in cases in which sacral fractures are not evident on CT but patients report pain on sacral palpation, an imminent fracture may be assumed and, thus, a screw may be inserted due to low complication risk and minimal additional time (Table ; )
Direct postoperative mobilization with pain adapted full weight bearing Daily physiotherapy focused on stability, gait security and prevention of falling, breathing exercises Main goal should be to regain the preoperative mobility and independence level and return to same domicile as before the fracture Thromboembolic prophylaxis for at least 6 weeks postoperative Removal of skin sutures after 2 weeks Clinical and radiological (anteroposterior pelvis X‑ray) follow-up after 6 weeks and 3–4 months In cases of persisting pain until 6 weeks or pathological findings on conventional x‑ray additional CT is recommended
Screw misplacement in the spinal canal or the neuroforamen with contact to the nerve root of S1 or S2 Cement leakage posterior into the neuroforamen or spinal canal with possible neurological symptoms Intraoperative or postoperative bleeding, for example, injury of the superior gluteal artery. If recognized intraoperatively, it may be controlled by image-guided embolization. General surgical and medical complications
Data were collected from 141 patients who underwent 3D robot-assisted image-guided percutaneous sacroiliac screw fixation for fragility fractures between January 2018 and August 2022. The average patient age was 82 ± 10 years (95% confidence interval 78–90), with a majority being female (89%). Most of the fractures were type II, consistent with the literature (Fig. ; ). Table summarizes patient characteristics, highlighting gender distribution, median age, hospital stay duration, and timeline from trauma to diagnosis and surgery. The time from trauma to surgery varied significantly due to referral delays or failure of conservative management. On average, diagnoses were made on the first day following trauma, and surgeries were typically performed 6 days postdiagnosis. As expected, surgery duration increased with the number of screws placed, with the median duration for one sacroiliac screw placement being 26 min (Table ). Table offers a detailed breakdown of the 238 screws placed in 141 patients, outlining the type of screws used and complications encountered. Direct postoperative complications, such are screw misplacement, cement leakage, bleeding or postoperative infection are reported for all patients. Long-term complications, such as secondary screw migration, is reported for 112 patients. The mean follow-up of those patients was 22 weeks (range 4–184 weeks). Five sacroiliac screws (1.3%) were placed suboptimally, identified by a 3D scan performed at the end of the procedure. These placements were within acceptable limits and only one required removal 9 days postoperatively due to sensory impairment. Five screws exhibited cement leakage into the neuroforamina. However, none of these patients reported neurological symptoms. Given the low incidence and minimal clinical impact of screw misplacement and cement leakage, it has become common practice in our hospital to perform a 3D scan at the end of the procedure, following augmentation. However, a 3D scan can be carried out at any point during surgery if there is uncertainty about correct screw placement depending on local preferences and level of expertise. No loosening or migration was observed in sacroiliac screws. There were no postoperative infections. Intraoperative bleeding was reported in 2 patients: one was controlled by compression and in the other patient an angiography was performed showing that the bleeding had stopped spontaneously. The percutaneous screw fixation technique for pelvic fragility fractures has evolved over time, addressing various technical challenges. While technological advancements from 2D to 3D and nonnavigated to navigated screw placement did not significantly reduce complication rates, it positively influenced screw misplacement rates and operation duration .
Appendix 1: simplification of the classification of FFP according to Rommens Appendix 2: personalized inlet and outlet angle Appendix 3: Abbreviations and dictionary Video
|
Genome-Wide Association Study of Ustekinumab Response in Psoriasis | 93fcf7f0-f0c8-4ee6-906e-7142a0d73984 | 8830831 | Pharmacology[mh] | Psoriasis is a common, chronic immune-mediated skin disease that affects at least 2% of the population worldwide . Psoriasis is associated with psoriatic arthritis, cardiovascular disease, metabolic syndrome, and other comorbidities, which makes effective management of psoriasis critical. Moderate-to-severe psoriasis is treated with phototherapy and systemic agents, including targeted biologic inhibitors of TNF- α , IL-12/23, IL-17, and IL-23. Patient responses to biologic therapy can vary widely, from poor overall response to gradual loss of therapeutic sensitivity . Response differences are largely influenced by patient weight and adherence, drug dose and bioavailability, and pharmacokinetic covariates, such as drug immunogenicity . The molecular heterogeneity of psoriasis may also contribute to differential therapeutic responses. However, there are no molecular biomarkers routinely used in clinical practice to facilitate selection of the therapies tailored to individual patients. Ustekinumab is a fully humanized immunoglobulin monoclonal antibody targeting the p40 subunit shared by IL-12 and IL-23. Phase 3 clinical trials showed that treatment with ustekinumab results in 75% improvement in the Psoriasis Area and Severity Index (PASI75) in ~66% of patients after 12 weeks of therapy . Candidate gene studies have identified the HLA-C*06:02 allele as being associated with better ustekinumab responses in both European and Chinese patients with psoriasis. A meta-analysis of eight studies including 1048 psoriasis patients showed that HLA-C*06:02 positive patients had a median PASI75 response rate of 92% after 6 months of ustekinumab therapy compared to a median PASI75 response rate of 67% in the HLA-C*06:02 negative patients . Here, we performed an unbiased genome-wide association study (GWAS) to evaluate if additional genetic factors were associated with ustekinumab response. We evaluated our findings across multiple response timepoints and in conjunction with HLA-C*06:02. Our findings highlight a potentially novel variant associated with ustekinumab response in psoriasis, which may further facilitate the development of precision medicine approaches.
Study Population This study involved analysis of individuals with moderate to severe psoriasis who participated in at least one of three placebo-controlled randomized clinical trials: PHOENIX I, PHOENIX II, and ACCEPT . Participants were originally approached for retrospective collection of DNA samples by investigators analyzing the association between the HLA-C*06:02 allele and response to IL-12/23 inhibition . In total, 439 patients of European descent were used to assess genetic associations between ustekinumab treatment and response. The GWAS discovery cohort consisted of 310 individuals who were treated with 45mg (n=146) or 90mg (n=164) of ustekinumab for 40 weeks, with the lower or higher dose given according to body weight less than or greater than 100 kg, respectively. The validation cohort consisted of 129 trial participants who crossed-over from placebo to ustekinumab treatment at week 12 and continued ustekinumab for 16 weeks, again dose-stratified by body weight (45 mg: n=64; 90 mg: n=65). In both cohorts, ustekinumab was given with two loading doses 4 weeks apart and every 12 weeks thereafter . Response Variables In the ustekinumab phase 3 clinical trials, the primary endpoint was achievement of PASI75 at week 12. PASI75 is a binary outcome converted from percent PASI improvement from baseline. To maximize power for the GWAS, we focused on the continuous outcome measure of percent PASI improvement from baseline to 12 weeks after ustekinumab therapy. Phenotypic response to ustekinumab was recorded at weeks 2, 4, 12, 28, and 40 for the majority of patients in the discovery cohort (cohort 1). In order to validate findings, the placebo to ustekinumab cross-over patients acted as a validation cohort (cohort 2). PASI responses for cohort 2 were measured after 12 weeks of ustekinumab therapy compared to trial start. Genome Wide Association Study Genotyping was performed using Illumina HumanOmni2.5-8 v1.2 BeadChips. Imputation was performed using the Michigan Imputation Server ( https://imputationserver.sph.umich.edu/index.html ) . The 1000 Genomes Phase 3 data was used as a reference panel for imputation . Files were converted to PLINK (v1.9) format, which along with R (v3.5.1) and python (v3.7.4), was used for data manipulation, visualization, and association analysis. Quality control and population stratification was performed following methods outlined by Marees et al. . Single nucleotide polymorphisms (SNPs) and individuals with missingness greater than 2% were removed. Duplicate, non-biallelic, and poor imputation quality (R 2 <0.7) SNPs were filtered. Non-autosomal SNPs with a low minor allele frequency (MAF<0.05) and significant deviation from Hardy-Weinberg equilibrium ( P <1×10 -6 ) were removed. In total 6,799,417 SNPs passed quality control, of which 1,696,820 were directly genotyped. Individuals with a heterozygosity rate +/-3 standard deviation from the mean were filtered, as well as the individual with the lowest call rate within a pair of cryptically related individuals ( π ^ > 0.2 ) . In total, 310 individuals (181 males, 129 females) passed quality control. The previously described quality control steps were applied to the 1000 Genomes Phase 3 data prior to merging with cohort data for population stratification. Multidimensional scaling (MDS) was applied to the merged genotype information. The presence of ethnic outliers was evaluated by qualitative alignment with the European superpopulation cluster along the top 2 MDS components. We included the top 10 MDS components as covariates in linear regression models for association testing. Statistical Analysis A threshold of P <5×10 -8 was established in the discovery cohort to determine the associated markers for further replication. We took linkage-disequilibrium into account when interpreting multiple significant association results from the same region. Clumping was employed to greedily assign groups around index variants with P <5×10 -6 . Variants with an R 2 >0.5 and less than 1MB away were assigned representation by the index variant. We modeled the additive effect of allele dosage with the quantitative phenotype of interest using linear regression. When considering cohort 1 index variants in replication analyses, a 2-sided t -test with P <0.05 was considered statistically significant. A two-sided normal test for proportions ( P <0.05) was applied to assess PASI threshold achievement differences based on genotype. The combined cohort association study followed the same procedures outlined for analysis of discovery cohort results. Power Analysis We performed power calculations for the discovery and replication cohorts assuming an additive linear model for our quantitative trait of interest. Each power calculation was performed under consideration of the established type 1 error rates for the respective cohort (cohort 1 α, 5×10 -8 ; cohort 2 α, 5×10 -2 ). We examined power across a range of MAF (0.05-0.25) and effect sizes (ES) (1–9). The genpwr (v1.0.4) R package was used for all calculations.
This study involved analysis of individuals with moderate to severe psoriasis who participated in at least one of three placebo-controlled randomized clinical trials: PHOENIX I, PHOENIX II, and ACCEPT . Participants were originally approached for retrospective collection of DNA samples by investigators analyzing the association between the HLA-C*06:02 allele and response to IL-12/23 inhibition . In total, 439 patients of European descent were used to assess genetic associations between ustekinumab treatment and response. The GWAS discovery cohort consisted of 310 individuals who were treated with 45mg (n=146) or 90mg (n=164) of ustekinumab for 40 weeks, with the lower or higher dose given according to body weight less than or greater than 100 kg, respectively. The validation cohort consisted of 129 trial participants who crossed-over from placebo to ustekinumab treatment at week 12 and continued ustekinumab for 16 weeks, again dose-stratified by body weight (45 mg: n=64; 90 mg: n=65). In both cohorts, ustekinumab was given with two loading doses 4 weeks apart and every 12 weeks thereafter .
In the ustekinumab phase 3 clinical trials, the primary endpoint was achievement of PASI75 at week 12. PASI75 is a binary outcome converted from percent PASI improvement from baseline. To maximize power for the GWAS, we focused on the continuous outcome measure of percent PASI improvement from baseline to 12 weeks after ustekinumab therapy. Phenotypic response to ustekinumab was recorded at weeks 2, 4, 12, 28, and 40 for the majority of patients in the discovery cohort (cohort 1). In order to validate findings, the placebo to ustekinumab cross-over patients acted as a validation cohort (cohort 2). PASI responses for cohort 2 were measured after 12 weeks of ustekinumab therapy compared to trial start.
Genotyping was performed using Illumina HumanOmni2.5-8 v1.2 BeadChips. Imputation was performed using the Michigan Imputation Server ( https://imputationserver.sph.umich.edu/index.html ) . The 1000 Genomes Phase 3 data was used as a reference panel for imputation . Files were converted to PLINK (v1.9) format, which along with R (v3.5.1) and python (v3.7.4), was used for data manipulation, visualization, and association analysis. Quality control and population stratification was performed following methods outlined by Marees et al. . Single nucleotide polymorphisms (SNPs) and individuals with missingness greater than 2% were removed. Duplicate, non-biallelic, and poor imputation quality (R 2 <0.7) SNPs were filtered. Non-autosomal SNPs with a low minor allele frequency (MAF<0.05) and significant deviation from Hardy-Weinberg equilibrium ( P <1×10 -6 ) were removed. In total 6,799,417 SNPs passed quality control, of which 1,696,820 were directly genotyped. Individuals with a heterozygosity rate +/-3 standard deviation from the mean were filtered, as well as the individual with the lowest call rate within a pair of cryptically related individuals ( π ^ > 0.2 ) . In total, 310 individuals (181 males, 129 females) passed quality control. The previously described quality control steps were applied to the 1000 Genomes Phase 3 data prior to merging with cohort data for population stratification. Multidimensional scaling (MDS) was applied to the merged genotype information. The presence of ethnic outliers was evaluated by qualitative alignment with the European superpopulation cluster along the top 2 MDS components. We included the top 10 MDS components as covariates in linear regression models for association testing.
A threshold of P <5×10 -8 was established in the discovery cohort to determine the associated markers for further replication. We took linkage-disequilibrium into account when interpreting multiple significant association results from the same region. Clumping was employed to greedily assign groups around index variants with P <5×10 -6 . Variants with an R 2 >0.5 and less than 1MB away were assigned representation by the index variant. We modeled the additive effect of allele dosage with the quantitative phenotype of interest using linear regression. When considering cohort 1 index variants in replication analyses, a 2-sided t -test with P <0.05 was considered statistically significant. A two-sided normal test for proportions ( P <0.05) was applied to assess PASI threshold achievement differences based on genotype. The combined cohort association study followed the same procedures outlined for analysis of discovery cohort results.
We performed power calculations for the discovery and replication cohorts assuming an additive linear model for our quantitative trait of interest. Each power calculation was performed under consideration of the established type 1 error rates for the respective cohort (cohort 1 α, 5×10 -8 ; cohort 2 α, 5×10 -2 ). We examined power across a range of MAF (0.05-0.25) and effect sizes (ES) (1–9). The genpwr (v1.0.4) R package was used for all calculations.
In this study, we analyzed genetic data from two cohorts of psoriasis patients receiving ustekinumab. Following preprocessing and filtering for individuals of European genetic ancestry, the discovery cohort (cohort 1) totaled 310 individuals (181 males, 171 females) and the validation cohort (cohort 2) totaled 129 individuals (82 males, 47 females). The average PASI score at baseline was 18.6 for cohort 1 and 18.8 for cohort 2 . Power analysis revealed the discovery cohort had 1-β>0.75 for MAF>0.05 and ES>7. The replication cohort had 1-β>0.75 for MAF>0.05 and ES>5 . We used linear regression to perform genome-wide association testing on the percent improvement in PASI response at week 12 of ustekinumab therapy compared to baseline . There was no correlation between age, BMI, and duration of the disease with the primary outcome of percent PASI improvement, and so these clinical variables were not included as covariates in the linear regression model . Genome-wide association testing of subjects in cohort 1 identified a single peak on chromosome 4 exceeding a genome-wide significance threshold of P <5×10 -8 lead by rs35569429 ( β , -19.84; 95% CI, -26.58 to -13.1; P =1.98×10 -8 ) ( and ). Directly genotyped SNP rs11722643 was in high linkage disequilibrium with imputed SNP rs35569429 and achieved a similar level of significance (R 2 , 0.9; β , -19.31; 95% CI, -26.33 to -12.29; P =1.44×10 -7 ). To determine whether multiple SNPs contributed to the peak on chromosome 4, we performed conditional analysis on rs35569429. The conditional analysis completely attenuated the GWAS peak, indicating a single independent signal at this locus . The major allele of rs35569429 is “G” while the minor allele is a single nucleotide deletion of G, denoted as “Del”. Subjects with at least one minor allele were labeled as the deletion positive group (Del+, N=55), and subjects with zero minor alleles were labeled the deletion negative group (Del-, N=255). Only one subject was homozygous for the minor allele. To understand the impact of this SNP at various discrete levels of PASI response, we examined the proportions of Del- and Del+ individuals who achieved PASI50, PASI75, PASI90, and PASI100 at Week 12. We found that in the Del- group, 235/255 (92.2%) achieved PASI50, 191/255 (74.9%) achieved PASI75, 121/255 (47.5%) achieved PASI90, and 48/255 (18.8%) achieved PASI100 at Week 12. In the Del+ group, 39/55 (80.9%) achieved PASI50, 24/55 (43.6%) achieved PASI75, 12/55 (21.8%) achieved PASI90, 5/55 (9.1%) achieved PASI100 at Week 12. To further investigate the validity of rs35569429, we analyzed its association with PASI outcomes in cohort 1 at timepoints that were not part of the original GWAS analysis (i.e. timepoints other than week 12). We found that a greater proportion of individuals in the Del- group achieved PASI75 compared to the Del+ group at Week 2 (1.57% vs 0%), Week 4 (17.6% vs 10.9%), Week 24 (76.5% vs 61.8%), and Week 28 (73.3% vs 52.7%) . Similarly, the Del- group had a higher proportion of individuals achieving PASI50, PASI90, and PASI100 than the Del+ group at weeks 2, 4, 24, and 28. The difference in PASI responses between Del- and Del+ groups were generally comparable if not greater than the difference in PASI responses between HLA-C*06:02 positive and HLA-C*06:02 negative individuals , where HLA-C*06:02 represents a previously well-validated locus associated with ustekinumab response . For comparison, in cohort 1, a linear regression of PASI response at week 12 for HLA-C*06:02, using 10 MDS components as co-variates, yielded β=0.7418 and P =0.0093. We next investigated the association of rs35569429 with response to ustekinumab in an independent cohort 2. We found the same direction of effect at week 12 for rs35569429 ( β , -6.71; 95% CI, -13.13 to -0.30; P =0.042) . In the Del- group, 102/106 (94.5%) subjects achieved PASI50, 81/106 (76.4%) subjects achieved PASI75, 45/106 (42.5%) achieved PASI90, and 26/106 (24.5%) achieved PASI100 at Week 12. In the Del+ group, 20/23 (87.0%) subjects achieved PASI50, 13/23 (56.5%) achieved PASI75, 9/23 (39.1%) achieved PASI90, and 2/23 (8.7%) achieved PASI100 at Week 12. Association testing for rs35569429 in cohort 1 and cohort 2 combined at week 12 yielded a genome-wide significant result ( β , -15.83; 95% CI, -20.72 to -10.74; P =2.42×10 -9 ). We ran a sensitivity analysis on the full sample of cohorts 1 and 2 combined at week 12. We observed the expected genome-wide significant peak at rs35569429, with the most significant SNP being rs11722643, which is in high linkage disequilibrium with rs35569429 (R 2 , 0.88; β , -16.64; 95% CI, -22.04 to -11.25; P =3.25×10 -9 ). We also observed a single additional genome-wide significant loci on chromosome 14, which could not be further confirmed (rs994384156; β , -14.94; 95% CI, -20.02 to -9.86; P =1.58×10 -8 ). We also conducted a separate GWAS on ustekinumab response at week 24 and did not identify any genome-wide significant SNPs. Finally, we explored how the combination of rs35569429 and HLA-C*06:02 affects PASI75 response in cohort 1 and 2 at week 12, since HLA-C*06:02 is an allele previously established to be associated with a more favorable responses to ustekinumab in psoriasis . In cohort 1 at week 12, 82.4% Del-/HLA-C*06:02+ individuals achieved PASI75 compared to 68.8% in Del-/HLA-C*06:02-, 61.1% in Del+/HLA-C*06:02+, and 35.1% in Del+/HLA-C*06:02- . In cohort 2 at week 12, 88.6% Del-/HLA-C*06:02+ individuals achieved PASI75 compared to 79.2% in Del-/HLA-C*06:02-, 72.7% in Del+/HLA-C*06:02+, and 50.0% in Del+/HLA-C*06:02-. In cohort 1 and cohort 2 combined at week 12, 84.4% Del-/HLA-C*06:02+ individuals achieved PASI75 compared to 71.6% in Del-/HLA-C*06:02-, 65.5% in Del+/HLA-C*06:02+, and 38.8% in Del+/HLA-C*06:02. The effects of rs35569429 and HLA-C*06:02 were independent from each other, as an interaction analysis that included an interaction term between rs35569429 and HLA-C*06:02 was not significant ( P =0.729).
This genetic association study found a genome-wide significant association between intergenic variant rs35569429 and response to ustekinumab for the treatment of moderate to severe psoriasis. In our primary association analysis, absence of the minor allele (Del-) was significantly associated with a larger PASI improvement at 12 weeks from baseline. More favorable PASI responses in Del- individuals compared to Del+ individuals were also observed at weeks 2, 4, 24, and 28. The association of rs35569429 with ustekinumab response was validated in an independent cohort of psoriasis patients. Conditional analysis revealed a single independent signal at the locus of interest. rs35569429 is characterized by a G deletion minor allele. This variant is located in an intergenic region 9 kB upstream of WDR1 . Functional analysis by GeneHancer Regulatory Elements strongly associates a 10.6 kB region (GH04J010114) 1.2 kB downstream of this variant with promoter/enhancer activity influencing proximal protein coding genes WDR1 and SLC2A9 . The WDR1 protein is involved in actin filament disassembly, a critical process of cytoskeleton dynamics, especially in highly motile and interacting immune cells . Impaired actin dynamics as a result of WDR1 deficiency have been causally linked to primary immunodeficiencies and autoinflammatory phenotypes . SLC2A9 is a transporter mainly expressed in the kidneys and primarily involved in urate reabsorption. Mutations of SLC2A9 lead to poor reabsorption and Renal Hypouricemia type-2, as caused by increased urate excretion . Future studies are needed to fine-map the causal and functional SNPs in linkage disequilibrium with rs35569429. Stratification of ustekinumab responses was greatest when rs35569429 was considered in combination with HLA-C*06:02. Individuals who were Del-/HLA-C*06:02+ achieved PASI75 84.4% of the time, while those were Del+/HLA-C*06:02- achieved PASI75 38.8% of the time, a more than two-fold difference. Pharmacogenomics continues to play an increasingly important role in precision medicine for dermatology. In 2018, five dermatologic drugs had clinically actionable pharmacogenomic tags that either require or advise testing of genomic biomarkers before treatment . Single FDA-approved biomarkers currently dominate this list; however, multi-gene marker panels will continue to gain importance for informing clinical decisions. Understanding the role of multiple SNPs in disease pathogenesis is important in advancing precision medicine. Conclusions from this study are limited due to the moderate sample size of the discovery and replication cohorts; our study was not powered for detection of small to moderate effects. Given the polygenicity of complex autoimmune diseases such as psoriasis, in the future, prospective design of large study cohorts is essential for thorough investigation of the biology contributing to therapeutic response. In general, validation in additional, independent cohorts will provide evidence with respect to the genomic signals discovered herein. Furthermore, the index SNP rs35569429 requires further investigation to identify the causal variant(s) associated with this locus and further characterization of functional effects on psoriatic response to ustekinumab.
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: ( https://doi.org/10.6084/m9.figshare.17009930 ).
The studies involving human participants were reviewed and approved by Institutional Review Boards of each clinical trial site participating in ustekinumab phase 3 studies. The patients/participants provided their written informed consent to participate in this study.
WL conceived and supervised the project. WC performed GWAS. WC and JH performed data analysis, prepared, and wrote the manuscript. All authors contributed to the article and approved the submitted version.
WL has received research grant funding from Abbvie, Amgen, Janssen, Leo, Novartis, Pfizer, Regeneron, and TRex Bio. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
|
Evaluation of rational prescribing in paediatrics | 4e0d5004-4ab1-48ed-9dae-d84dbe10cf8d | 7970256 | Pediatrics[mh] | Reviewer comments Author's manuscript
|
The Wenckebach Phenomenon | 21c63906-c1d5-4f00-a569-ed6122a75bb5 | 8142363 | Physiology[mh] | INTRODUCTION Medicine has many great pioneers, their lasting legacies recalled at the mention of a name. This pioneer, his name, and the phenomenon he originally described are known across the world, simply as Wenckebach. Born March 24 th 1864, Karel Frederik Wenckebach, a Dutch anatomist and cardiologist, would have had little inclination of how famous his name would later become . Having studied medicine in Utrecht, the Netherlands, he went on to research the rhythmic patterns of the frog heart under Engelmann, setting the stage for his future achievements . It was in 1899 however, that the young physician recorded an observation in the jugular venous pulse (JVP) of a human patient and discovered a phenomenon which would later bear his name . Laying in direct contact with the right sided heart, the jugular veins allow an indirect observation of the various phases of the cardiac cycle. Wenckebach meticulously observed such changes in the relative timing of two key waveforms; the ‘a’ and ‘c’ waves, corresponding to atrial and ventricular contraction, respectively. His observations, recorded using a smoked drum kymograph, are shown below and demonstrate progressive lengthening of the interval between the two waveforms until the absence of the 4 th c wave, followed by a pause. Wenckebach’s findings were later published in 1899 in what is considered the first appreciation of decremental conduction within cardiac tissues . Although later re-classified by Mobitz in 1924 as type 1, second degree AV block, the phenomenon still bears the legacy of the person who first documented its presence, long before the advent of electrocardiography . The numerous ways in which we observe this behaviour have increased significantly due to the rapid pace at which electrophysiology has developed. In a similar way, this chapter will demonstrate the same behaviour across a spectrum of modalities, from the 12 lead ECG through to the intra-cardiac recordings from both electrophysiological testing and implantable cardiac devices. Wenckebach had astutely observed beat to beat, progressive lengthening of the interval between the a and c waves in the jugular vein of his patient (Fig. ). A point was reached where an a wave was seen, but not followed by a c wave, and, following a pause, the cycle reset with the ‘ac’ interval returning to baseline. Although not immediately appreciated by Wenckebach, he had observed the mechanical result of a specific form of AV nodal behaviour. To lay the foundation on which to discuss this further, we begin with some basic concepts of the function of the cardiac electrical system, focusing on the AV node, and in particular, decremental conduction. In essence, the atria and ventricles are electrically isolated from one another except for a small region of highly specialized cells - the AV node. Laying anatomically at the base of the right atrium within the region known as the triangle of Koch, the AV node sits as the gatekeeper of electrical activation to the His-Purkinje system and ultimately, the ventricles. While permitting the passage of electrical activity from the atria to the ventricles, the AV node imposes a slight slowing in conduction, thereby ensuring sufficient time for optimization of cardiac filling and preservation of sequential atrial and ventricular activation. However, the AV node does not have a binary ‘all or nothing’ response to incoming atrial stimuli. With progressive shortening of the intervals between the incoming stimuli, lengthening of the output is seen until conduction block occurs, protecting the ventricles from rapidly conducted atrial rhythms; a concept known as decremental conduction. Of note, the point at which block will occur is dynamic. The node is fully integrated into the neurohormonal milieu and is under the constant influence of competing inputs from the autonomic nervous system. Consequently, the specific conduction properties of the AV node vary – for example, under periods of increased vagal tone, the given cycle length at which the Wenckebach block will occur is increased. Wenckebach’s observations reflected progressive prolongation in the time taken for electrical activation in the atria to travel to the ventricle, resulting in electrical depolarization. Moreover, at a critical, fixed input cycle length, the properties of the AV node are such that progressive delay in the PR interval (to be precise te AH interval) is seen before the P wave fails to be conducted to the ventricle. The resulting brief pause allows sufficient time for tissue recovery and the cycle is ultimately reset. It is this classical behaviour, first identified some 120 years ago that has gone on to bear the name of the pioneer who first documented it - Wenckebach block. While this behaviour may be considered normal, it may also occur during disease states, at times necessitating insertion of an implantable electronic device. In this case, the resting 12 lead ECG can be particularly helpful; in general, the absence of associated distal conduction disease (manifested by prolongation of the QRS) suggests the location of the block is within the AV node. Even in cases where a broad QRS is seen (>120ms) most type 1, second degree AV blocks still remain in the AV node, with the remainder occasionally occurring within the distal conduction system, which is of prognostic significance, and usually requires pacing . The rhythm strip seen in demonstrates the classical features of the Wenckebach phenomenon and will be used to emphasize some of the key features of Wenckebach block. With a stable atrial input of 780 ms we see a progressive lengthening of the PR interval until a non-conducted atrial beat and a brief pause are seen, after which the cycle resets. Considering this concept further, it should not be of surprise that a diagnosis of Wenckebach block requires at least 2 consecutive stimuli to demonstrate progressive conduction delay prior to eventual block . In the absence of this, the observer would be unable to demonstrate progressive prolongation in the PR interval and therefore unable to distinguish this from other types of heart block such as type 2, second degree AV block with periodic, non-conduction of atrial stimuli. On closer review we see a further classical observation during Wenckebach block. That is, the PR interval (the interval between the P wave and its conducted ventricular response) immediately preceding the non-conducted impulse shows the longest PR interval, and the one following, the shortest. If we consider that the ‘input’ to the AV node remains at a fixed cycle length, for example during stable sinus rhythm, and that by definition only a single dropped beat occurs, then given that the longest AV delay precedes the dropped beat and the shortest follows it, the duration of the ventricular pause cannot exceed twice that of the P-P interval . The next, and probably subtlest of all observations, is the change in R-R intervals during Wenckebach. Contrary to the progressive lengthening of the PR intervals, the R-R intervals appear to shorten prior to the pause. Again, this seems initially confusing. However, if we recall that the atrial input remains constant, and we consider that in non-decremental conduction, in a 1:1 relationship, the atrial and ventricular rates would be equal; then, given that the largest decrement occurs between the first and second R waves, the result is the first R-R interval is the longest. Since the increase in the PR interval becomes progressively shorter for each conducted beat in a typical Wenckebach cycle, the R-R interval actually becomes progressively shorter until the blocked atrial impulse. On this basis, and using as an example, the difference in RR intervals is equal to: R1-R2 = PP+PR2-PR1 = PP+0.32-0.26 = PP + 0.06 R2-R3 = PP+PR3-PR2 = PP+0.36-0.32 = PP + 0.04 R3-R4 = PP+ PR4-PR3 = PP+0.38-0.36 = PP + 0.02 R4-R5 = 2PP+PR5-PR4 = 2PP+0.26-0.38 = 2PP – 0.12 To recap, we have now described the 4 key elements of typical Wenckebach behaviour. In association with repetitive group beating they are: Progressive lengthening of each successive PR interval. The pause produced by the non-conducted P wave is equal to the increment between the last PR interval (preceding the pause) and the first PR interval following the pause (shortest) subtracted from twice the PP interval. The RR interval between the first and second conducted beats is the largest and between the last conducted beats, the shortest. There is progressive shortening of the RR intervals. Early observers of this phenomenon, including Wenckebach himself, appreciated common, recurring cycles in the ratio between the number of inputs and outputs seen. Wenckebach identified patterns of the progressive prolongation in the AC interval, termed periodicity. In his published smoke drum recordings, he clearly identified a 4:3 ratio between the ‘input’ and ‘output’ waveforms. It is however recognised that both ‘typical’ and ‘atypical’ forms of Wenckebach exist . However - it is important to note that the term ‘typical Wenckebach’ is not related to the ratio of P waves to QRS complexes but of the adherence to the above ‘key elements’ during this phenomenon. During their study published in 1975, Denes and colleagues’ attempt to quantify the frequency of both typical and atypical Wenckebach periodicity and to define the atypical variations . While assessing patients in whom spontaneous or pacing induced block was seen, they made some interesting observations as to patterns of behaviour seen with this phenomenon. In those patients with spontaneous Mobitz type 1 block, the only observed A:V ratios that met all the original criteria listed above was a 4:3 ratio and, even in this case, only 41% of cases met all criteria, hence being deemed ‘typical’. Therefore, 59% of patients with spontaneous 4:3 ratios had an ‘atypical’ pattern – defined as having met the general definition of Wenckebach but failing to meet all of the criteria. Looking then to patients who had pacing induced Wenckebach, 69% of the 4:3 ratios, 14% of 5:4 and 11% of 6:5 ratios were typical, moreover any induced ratio above 6:5 failed to display any typical behaviour. The work of Denes and colleagues also highlights a classical observation of Wenckebach periodicity, in that the ratio of inputs to outputs always conforms to a (n+1)/n configuration . This may not be surprising given the Wenckebach sequence ends following a single non-conducted beat – thus the numbers of inputs will always exceed the outputs by one. Overall this appears straightforward, however, as we will see later this concept becomes a little more confusing when attempting to make this pattern conform to more unusual ratios – for example 8:3. We will visit the underlying mechanism to this shortly. Having demonstrated Wenckebach block arising within the AV node as a ‘whole’, intriguingly, this behaviour appears neither constrained to a single component of the AV node, nor indeed the AV node itself. The AV node is a complex structure and on a functional level there can be many variations. In one common scenario an individual may possess ‘dual AV node physiology’, that is, the AV node may contain both a slow and fast conducting region. The former displays slow conduction but a short refractory period, whilst the latter conducts rapidly yet has a long refractory period. Under a set of reasonably specific conditions, re-entry may occur between these two functional pathways resulting in supraventricular tachycardia. While eminently amenable to curative ablation by way of modification to the slow pathway, such that its conduction properties are altered and it is no longer able to sustain tachycardia, its ablation has given rise to some interesting features in regard to Wenckebach behaviour. Elegant work by Zhang and Masgalev clearly demonstrated that while Wenckebach behaviour could be observed via the slow pathway, following slow pathway ablation in animal studies the point at which Wenckebach would occur post ablation would alter, but the phenomenon still occurred – that is, it seems to be a feature, where present, of the entire AV node and not isolated to one component. Furthermore, although much less common, the more distal components of the His-Purkinje system appear capable of demonstrating Wenckebach behaviour. In the tracing below, the first P wave is not conducted, however the second and third P waves are followed by a wide QRS with RBBB morphology and left axis deviation (Fig. ). The PR interval of the first conducted beat demonstrates the shortest PR interval and the following PR interval appears prolonged. The next P wave in the sequence is no longer followed by a QRS. Towards the end of the recording, the final P wave is once more followed by a QRS with a similar PR interval to the first conducted P wave. On closer examination of the His electrograms we see that each atrial electrogram is followed by a His electrogram in a 1:1 relationship with a stable AH interval. However, the conducted beats demonstrate progressive prolongation in the HV interval until the third beat in the sequence is no longer conducted and blocked after the His deflection (arrow), in keeping with infra-nodal Wenckebach block. Interestingly, there are a number of similarities in cellular physiology between both the AV and sinoatrial (SA) nodes. While conceptually this may be more difficult to understand, the SA node is also capable of Wenckebach behaviour . Remember, in AV nodal Wenckebach the ‘input’ to the AV node is the P wave, and this can be readily seen on the surface ECG, thereby making the phenomenon easier to observe and understand. Difficulty comes however in the case of sinoatrial Wenckebach where ‘input’ sinus node depolarization is not visible, and the only measurable electrical component is the ‘output’ i.e . the P wave. Above, in , is an ECG strip demonstrating sinoatrial Wenckebach. On initial inspection, a classical pattern of grouped beating is seen with pairs of P waves. In this case, as stated, the input cannot be seen, but the output can. In the ladder diagram below the ECG, the classical depiction of the A, AV and V as demonstrated above, have been replaced by S (sinus node), SA (sinoatrial junction, akin to the AV node) and the A (atrium). Although unseen, it is still possible to calculate the input within the SA node using the available atrial cycle information. As in AV nodal Wenckebach, the ratio of input to output remains (n+1)/n. Calculating the total cycle length, equal to the interval from A1 to A3 = 1880 ms. During this time there have been a total of 3 inputs from within the SA node. Therefore, given that the total S and A cycle lengths must be equal, the unseen ‘input’ cycle length is 1880/3=626ms. Now we return to the conundrum previously posed with regard to the concept of an 8:3 ratio of AV block. As seen earlier, Wenckebach block conforms to an (n+1)/n relationship; therefore, one might initially struggle with the concept of an 8:3 ratio. However, we have also seen that multiple regions within the AV node and distal conduction system are able to demonstrate Wenckebach behaviour. Let us consider therefore the scenario in which more than one level of block is occurring within the AV node – obtaining a ratio of 8:3 block suddenly seems reachable. On the assumption that all regions continue to conform to the (n+1)/n ratio then mathematically, for argument sake, consider that the proximal AV node displays a 2:1 block, and the more distal portion of the node displays 4:3 block; the end result, from initial input at the proximal AV node to the exit in the distal conduction system is 2:1x4:3=8:3. Multiple, more complex ratios of block can be explained in a similar fashion. Above, in , is a 12 lead ECG of typical atrial flutter. On closer inspection, as shown, it is clear that for 8 ‘inputs’ there are 3 ventricular responses. demonstrates a ladder diagram displaying the multiple levels of block within the AV node (AVN1 and AVN2) and the presence of typical Wenckebach periodicity within the AVN2 region. Having observed Wenckebach behaviour via the 12 lead ECG and the intracardiac EGM, next we turn our attention toward implantable cardiac electronic devices. Although this concept may appear confusing, it is important to remember that we are still observing exactly the same behaviour shown above but displayed in a slightly different way. Conceptually, this is similar to the previous description of SA nodal Wenckebach given that the only observable information may be the output, in this case, the ventricular response. Of course, we now know that it is still possible to obtain information regarding the input cycle given that once more, the classical (n+1)/n ratio remains upheld – for now. demonstrates a ventricular interval dot plot illustrating the timing of ventricular activation from an implantable cardioverter defibrillator (ICD) interrogation. For those not familiar with such a tracing, each dot represents a ventricular electrogram and its associated timing, relative to the preceding beat (or dot) in milliseconds, displayed over time. From first glance, we again see the typical grouped beating appearance of output periodicity seen with Wenckebach conduction. Recall the fundamental concept that the RR interval decreases with each beat leading up to the non-conducted beat, following which the cycle resets. In this case the relative pause resulting from the non-conducted beat is a little harder to identify. Looking at the timing sequence, we see that the second and third beats are closer to one another than the third and fourth. Beginning from the second QRS at 410ms, we see the initiation of a typical Wenckebach sequence with shortening of the R-R interval with successive beats – the next being 310ms. Following, there is a slightly longer delay (during which the non-conducted beat occurs) and the cycle restarts once more at 410 ms. Applying the (n+1)/n formula to the 2 ventricular beats seen (410ms + 310ms), by definition implies the presence of an additional input. Therefore, the ‘input’ cycle length can be calculated as (410+310)/3=240ms. This 3:2 periodicity continues until we view the second half of the dot plot diagram. Now we observe a further change with the addition of a further ‘dot’ to the sequence. Note the relative reduction in cycle length is less than that seen between the first and second beats, thus we observe progressive R-R interval shortening with each successive beat in the sequence, in keeping with the typical Wenckebach pattern of 4:3 block. Applying the (n+1)/n formula allows us to demonstrate that, as expected, the input cycle length has not been altered. Therefore, given the cycle has now increased by 1, and given the relative timing of the third beat, the equation becomes (410+310+250)/4=242ms. The subsequent 4 cycles continue with 4:3 block following which a further additional dot joins the sequence and continues to obey the expected ‘typical’ Wenckebach behaviour, resulting in 5:4 block. Reapplying the same formula once more confirms preservation of the input cycle length; (410+310+250+242)/5=242ms. In the last 3 cycles, we observe an obvious and unexpected change from the earlier pattern. The additional 5 th , 6 th , and 7 th beats joining the next two sequences are, respectively, of increased then decreased cycle length. Therefore, typical Wenckebach periodicity is no longer seen. As outlined by the work of Denes and colleagues referred to earlier, we recall that beyond a 4:3, and certainly 5:4 ratio, a typical Wenckebach pattern is not commonly seen, a point well illustrated here.
When confronted with a tracing, whether a 12 lead ECG or intracardiac recording, the signature appearance of grouped beating should demand close attention and consideration as to the underlying rhythm. Indeed, where a repetitive, albeit irregular rhythm is seen, it is likely that the input is a regular one. In such a case, Wenckebach behaviour should be strongly considered and close inspection often identifies the four key features of typical Wenckebach block. Even when the underlying ‘input’ is unclear, the ‘output’ cycle length, and application of the (n+1)/n formula will allow its calculation, of course remembering the caveat that beyond ratios of 5:4, the likelihood of typical Wenckebach behaviour is considerably less. Despite its original description 120 years ago, the Wenckebach phenomenon remains one of the fundamental concepts within electrophysiology. Although the hallmark of progressive PR segment prolongation remains well known to many physicians, to fully appreciate its subtle complexities, it continues to demand the same meticulous attention to detail as demonstrated by Wenckebach himself.
|
Human Cerebrospinal fluid promotes long-term neuronal viability and network function in human neocortical organotypic brain slice cultures | 412d2152-15b4-4f73-afa1-35d02fb4b447 | 5613008 | Pathology[mh] | Investigation of pathophysiological mechanisms of human neurodegenerative and other central nervous system (CNS)-related diseases as well as the development of new therapeutic avenues in first line relies on studies involving model systems that include cell culture systems and animal models. Often it is assumed that obtained data apply or can be extrapolated to human CNS but there are numerous examples given by pharmaceutical research and drug development demonstrating that there are significant differences, even posing the risk that newly developed drugs which proved to effectively target certain pathways in model systems will not show the expected effects in patients. One way of approaching this challenge is the use of embryonic stem (ES) or induced pluripotent stem (iPS) cell technology or cerebral organoids , allowing engineering and growing of “humanoid” neurons in vitro for investigation and analysis. While such approaches are considered a huge step forward, at the same time they are controversially discussed regarding e.g. purity, subtype-identity and maturity of generated “humanoid” neurons. Furthermore, this technology currently does not provide the tools to rebuild entire complex human CNS tissue (including entire cortical layering and columnar organization), which would be crucial when working toward a deeper understanding of physiological characteristics of human CNS circuitry or of pathophysiological mechanisms contributing to the development of human CNS-related disease. Alternative approaches to understand diverse aspects of highly complex connectivity and structure of the CNS down to morphological and functional properties of single cells/neurons, altogether governing the functions of the human brain, include MRI imaging technology and elegant stimulation studies or electrophysiological measurements in acute tissue slice preparations – . Interestingly, some of these studies were able to identify striking differences between human and rodent neurons, confirming that a direct translation from the animal model to humans might sometimes be misleading , . While – as mentioned above – electrophysiological characterizations of “static” intrinsic properties of human neurons can be successfully achieved in acute slices or potentially iPS-derived “humanoid” neurons , , investigation of intermediate/long-term changes of neuronal properties/function, as for example mediated by certain neuroplasticity mechanisms or by specific gene mutations of interest, remain challenging. In studies based on rodent tissue the development of organotypic slice cultures has been successfully used to bridge the gap of in vitro to in vivo translation , but at the same time the challenges regarding immediate transferability of rodent data to the human organism have remained unsolved. The establishment of optimized human organotypic cortical slice cultures as an ex vivo system seems very tempting in this regard. Such systems could address some of the mentioned questions by enabling the investigation of dynamic properties of human neurons (such as neuroplasticity) and potentially of other cortical cells including glia over a time period of days to weeks. Furthermore, such cultures could serve as a platform to directly explore the functional impact and pathophysiological mechanisms of defined mutations (for example of ion channels) and their role in the development of CNS-related disease, which could be studied in the context of human cortical networks by help of viral-mediated overexpression. In addition, organotypic cortical slice cultures can be used for drug discovery and preclinical studies since it provides a peerless opportunity to investigate the effect of substances to human neuronal networks that developed physiologically. It also offers a possible tool for testing neuroprotective agents to networks maintaining their cellular complexity as well as their anatomical integrity . While previous studies in organotypic cultures prepared from human post mortem cortical tissue demonstrated continued viability of many cells including neurons up to several weeks in vitro – , additional neurobiological data indicated that under certain conditions cultures prepared from resected human cortical tissue may be subject to a severe injury response, involving proliferation of reactive cells as well as progressive neurodegeneration . However, recent work by another group demonstrated that such processes potentially could be ameliorated using optimized complex defined artificial culturing media, enabling at least partial preservation of characteristic neuronal morphology and importantly also pathological electrophysiological activities of mainly but not exclusively subcortical limbic structures (hippocampus, subiculum) in organotypic cultures of adult human tissue . Building on these data, we here demonstrate that long-term neocortical neuronal viability and robust electrophysiological single cell and network function can be preserved in human organotypic cortical slice cultures by using human cerebrospinal fluid as culturing medium. These cultures could serve as a platform enabling direct validation of data obtained in model systems including but not limited to ES-/iPS technology, rodent primary neuronal cultures, organotypic slice cultures, and in vivo approaches or even non-mammalian heterologous expression systems.
Neuronal activity of human organotypic brain slice cultures cultured in human CSF vs. traditional media From the initial tissue blocks, 250–300 µm thick slices (n = 57) were cut and kept in culture for 2–21 days either in human cerebrospinal fluid (hCSF) or in traditional culture media (either BrainPhys or organotypic slice culture medium (OSCM), see methods for details) before recordings were performed. To assess the viability of the slices, the electrical activity of the slices was tested with an extracellular population electrode to detect Multi-Unit-Activity (MUA). In all tested slices the concentration of potassium in the recording artificial cerebrospinal fluid (aCSF) was increased to 8 mM to facilitate spontaneous neuronal activity (Fig. ). In general, slices were found to either generate spontaneous rhythmic network discharges (Fig. , left panel), tonic desynchronized firing (Fig. , right panel) or no detectable activity at all (Fig. , upper panels, 7 days in vitro (DIV) and 14–21 DIV). The slices cultured in traditional media (see methods for details) showed only in a minority of cases (3/21) neuronal extracellular activity after being in vitro up to seven day s . Strikingly, only during the first three days of culturing, the slices kept in traditional media were able to produce network driven bursting activity (n = 2/21, 3 DIV; Figs ), while 1/21 produced tonic firing after 7 days in vitro (DIV). In all other slices tested no activity was detected (n = 18/21, 3–21 DIV; Fig. ). In contrast, the majority of slices cultured in hCSF produced either tonic activity or rhythmic network discharges (n = 32/36, 3–21 DIV; Fig. ). The 4/36 slices not exhibiting any neuronal activity were at least 14 DIV or even longer. Rhythmic network bursting in human organotypic slice cultures cultured in hCSF From the 32 slices that showed neuronal activity, 13/32 showed rhythmic network activity in aCSF with increased potassium concentrations of 8 mM. The network activity occurred with a mean frequency of 0.245 ± 0.09 Hz and had a mean duration of 4774 ± 2251 ms. The network activity seemed to be variable in duration and frequency (Fig. ). Some slices produced long lasting (up to 30 s) synchronous discharges (Fig. ), while others showed a higher bursting frequency with shorter durations (under 1 s). The frequency of these slices tended to be faster in correlation to the duration in vitro , but this was not statistically significant (Fig. , p = 0.0739). In several slices (n = 11) we also tested the effect of bath application of Muscarine 5–10 µM on the network activity in addition to raising the extracellular potassium concentration to 8 mM. 5/11 slices showed tonic activity in the presence of 8 mM potassium, which increased in the presence of Muscarine, but only induced in 1/5 rhythmic network bursting. In 4/6 slices that showed rhythmic network bursting in 3 mM or elevated levels of 8 mM extracellular potassium, the application of Muscarine increased the frequency of bursting from 0.0029 ± 0.003 to 0.16 ± 0.29, while in two slices the application of Muscarine increased the tonic activity but stopped network bursting activity. To test if the network activity in human organotypic slices was driven by glutamatergic excitatory synaptic transmission we applied CNQX (20–30 µM, n = 3, Fig. ), which completely blocked the occurrence of network bursts in all tested slices. Firing properties of single neurons in hCSF and traditional culture media In a next step, we investigated in a small number of cells, if neurons cultured in human organotypic slices were still able to produce firing behavior similar to the one of acute slices reported previously . We recorded a total number of 14 cortical neurons in slices cultured in hCSF or traditional media at different time points (2–14 DIV, Table , Fig. ). All neurons tested had a resting membrane potential more negative than −60 mV with an average of −70.3 ± 1.9 mV. Upon positive supra-threshold current injections all neurons cultured in hCSF (n = 10, 2–14 DIV) were able to fire repetitive trains of action potentials (Fig. , Fig. ). In slices cultured in traditional media we were able to record four cortical neurons (3 DIV-13 DIV, Fig. ). Repetitive supra-threshold current injections (up to 400 pA) led only to the generation of a single action potential in cells recorded in slices cultured in media for 10 to 13 DIV (Fig. , n = 2), while two neurons recorded after 3 DIV cultured in media were able to produce sustained action potential firing (Fig. ). The anatomy of six biocytin filled neurons recorded in hCSF-cultured slices could be reconstructed after fixation. We used biocytin filling for a better morphological reconstruction on a single cell level comparatively to Map2 staining for dendritic integrity on population level. Four of these neurons had the morphology of pyramidal cells of Layer 2/3 and showed also a typical regular spiking pattern (Fig. ). Two of the neurons we reconstructed had the morphology of non-pyramidal neurons and showed a distinct fast firing pattern typical for non-pyramidal neurons (Fig. ). Four of the cells recorded in hCSF-cultured slices could not be reconstructed. Of the four neurons recorded in slices cultured in traditional media we could recover two which had the morphology of pyramidal neurons and showed a regular spiking pattern (Fig. , 3 DIV). The other two neurons showed an atypical firing behavior with only a single action potential upon supra-threshold (Fig. , 10–13 DIV). In all cells recorded in hCSF-cultured slices, we recorded up to 10 min of spontaneous activity and detected in 8/10 neurons recorded in slices cultured in hCSF excitatory postsynaptic potentials (Fig. , red arrow) and in 7/10 cells inhibitory postsynaptic potentials (Fig. , blue arrow). Moreover, in 3/10 cells recorded in hCSF-cultured slices (3 DIV, 9 DIV, 13 DIV) phasic network driven population input was detected in elevated concentration of potassium (8 mM, Fig. ). In the cells recorded from slices cultured in media 1/4 cell received phasic population driven input (3 DIV). Histology of slices cultured in hCSF vs. traditional media To determine if in addition to the electrophysiological changes also differences in the structure of the slices can be detected, we conducted a set of histological assessments. We used NeuN to stain the nuclei of post-mitotic neurons and Map2 to determine the dendritic morphology of the cells (Fig. , n = 10 for hCSF and n = 4 for traditional media). After 3, 7, 10, 13 and 18 DIV in hCSF-cultured slices we found densely populated NeuN positive cells in all layers of the cortical slice (Figs , ). Furthermore, the co-labeling with Map2 revealed an intact somato-dendritic morphology of the cells with typical ramification of the apical dendrite (Fig. , 18 DIV, 7A and 7B, 7 and 13 DIV). In slices cultured in traditional media we found also a similar labeling of NeuN positive cells just after a few days (3 DIV and 7 DIV, Fig. ), but the Map2 staining revealed already a severe loss of the dendritic structures in the neurons of these slices (Fig. ). To quantify this data, we analyzed a representative area of n = 4 for traditional media and n = 4 for hCSF cultured slices at the time points 3 DIV, 7 DIV, 13 DIV and 18 DIV and counted the NeuN and Map2 double positive cells within a 450 × 450 µm field taken with a confocal microscope. We found a significant higher number of double positive cells in slices treated with hCSF compared to traditional media (Fig. , Mann Whitney test, *p<0.05). However, the number of cells positive for NeuN alone in slices cultured with traditional media compared to hCSF was not significantly different. In addition, in slices cultured in hCSF, we tested if we could label astrocytes (n = 6; 3 DIV, 6 DIV, 7 DIV, 14 DIV and 28 DIV). We found at all stages dense labeling of GFAP positive cells in all layers of the human slices cultured in hCSF (Fig. ).
From the initial tissue blocks, 250–300 µm thick slices (n = 57) were cut and kept in culture for 2–21 days either in human cerebrospinal fluid (hCSF) or in traditional culture media (either BrainPhys or organotypic slice culture medium (OSCM), see methods for details) before recordings were performed. To assess the viability of the slices, the electrical activity of the slices was tested with an extracellular population electrode to detect Multi-Unit-Activity (MUA). In all tested slices the concentration of potassium in the recording artificial cerebrospinal fluid (aCSF) was increased to 8 mM to facilitate spontaneous neuronal activity (Fig. ). In general, slices were found to either generate spontaneous rhythmic network discharges (Fig. , left panel), tonic desynchronized firing (Fig. , right panel) or no detectable activity at all (Fig. , upper panels, 7 days in vitro (DIV) and 14–21 DIV). The slices cultured in traditional media (see methods for details) showed only in a minority of cases (3/21) neuronal extracellular activity after being in vitro up to seven day s . Strikingly, only during the first three days of culturing, the slices kept in traditional media were able to produce network driven bursting activity (n = 2/21, 3 DIV; Figs ), while 1/21 produced tonic firing after 7 days in vitro (DIV). In all other slices tested no activity was detected (n = 18/21, 3–21 DIV; Fig. ). In contrast, the majority of slices cultured in hCSF produced either tonic activity or rhythmic network discharges (n = 32/36, 3–21 DIV; Fig. ). The 4/36 slices not exhibiting any neuronal activity were at least 14 DIV or even longer.
From the 32 slices that showed neuronal activity, 13/32 showed rhythmic network activity in aCSF with increased potassium concentrations of 8 mM. The network activity occurred with a mean frequency of 0.245 ± 0.09 Hz and had a mean duration of 4774 ± 2251 ms. The network activity seemed to be variable in duration and frequency (Fig. ). Some slices produced long lasting (up to 30 s) synchronous discharges (Fig. ), while others showed a higher bursting frequency with shorter durations (under 1 s). The frequency of these slices tended to be faster in correlation to the duration in vitro , but this was not statistically significant (Fig. , p = 0.0739). In several slices (n = 11) we also tested the effect of bath application of Muscarine 5–10 µM on the network activity in addition to raising the extracellular potassium concentration to 8 mM. 5/11 slices showed tonic activity in the presence of 8 mM potassium, which increased in the presence of Muscarine, but only induced in 1/5 rhythmic network bursting. In 4/6 slices that showed rhythmic network bursting in 3 mM or elevated levels of 8 mM extracellular potassium, the application of Muscarine increased the frequency of bursting from 0.0029 ± 0.003 to 0.16 ± 0.29, while in two slices the application of Muscarine increased the tonic activity but stopped network bursting activity. To test if the network activity in human organotypic slices was driven by glutamatergic excitatory synaptic transmission we applied CNQX (20–30 µM, n = 3, Fig. ), which completely blocked the occurrence of network bursts in all tested slices.
In a next step, we investigated in a small number of cells, if neurons cultured in human organotypic slices were still able to produce firing behavior similar to the one of acute slices reported previously . We recorded a total number of 14 cortical neurons in slices cultured in hCSF or traditional media at different time points (2–14 DIV, Table , Fig. ). All neurons tested had a resting membrane potential more negative than −60 mV with an average of −70.3 ± 1.9 mV. Upon positive supra-threshold current injections all neurons cultured in hCSF (n = 10, 2–14 DIV) were able to fire repetitive trains of action potentials (Fig. , Fig. ). In slices cultured in traditional media we were able to record four cortical neurons (3 DIV-13 DIV, Fig. ). Repetitive supra-threshold current injections (up to 400 pA) led only to the generation of a single action potential in cells recorded in slices cultured in media for 10 to 13 DIV (Fig. , n = 2), while two neurons recorded after 3 DIV cultured in media were able to produce sustained action potential firing (Fig. ). The anatomy of six biocytin filled neurons recorded in hCSF-cultured slices could be reconstructed after fixation. We used biocytin filling for a better morphological reconstruction on a single cell level comparatively to Map2 staining for dendritic integrity on population level. Four of these neurons had the morphology of pyramidal cells of Layer 2/3 and showed also a typical regular spiking pattern (Fig. ). Two of the neurons we reconstructed had the morphology of non-pyramidal neurons and showed a distinct fast firing pattern typical for non-pyramidal neurons (Fig. ). Four of the cells recorded in hCSF-cultured slices could not be reconstructed. Of the four neurons recorded in slices cultured in traditional media we could recover two which had the morphology of pyramidal neurons and showed a regular spiking pattern (Fig. , 3 DIV). The other two neurons showed an atypical firing behavior with only a single action potential upon supra-threshold (Fig. , 10–13 DIV). In all cells recorded in hCSF-cultured slices, we recorded up to 10 min of spontaneous activity and detected in 8/10 neurons recorded in slices cultured in hCSF excitatory postsynaptic potentials (Fig. , red arrow) and in 7/10 cells inhibitory postsynaptic potentials (Fig. , blue arrow). Moreover, in 3/10 cells recorded in hCSF-cultured slices (3 DIV, 9 DIV, 13 DIV) phasic network driven population input was detected in elevated concentration of potassium (8 mM, Fig. ). In the cells recorded from slices cultured in media 1/4 cell received phasic population driven input (3 DIV).
To determine if in addition to the electrophysiological changes also differences in the structure of the slices can be detected, we conducted a set of histological assessments. We used NeuN to stain the nuclei of post-mitotic neurons and Map2 to determine the dendritic morphology of the cells (Fig. , n = 10 for hCSF and n = 4 for traditional media). After 3, 7, 10, 13 and 18 DIV in hCSF-cultured slices we found densely populated NeuN positive cells in all layers of the cortical slice (Figs , ). Furthermore, the co-labeling with Map2 revealed an intact somato-dendritic morphology of the cells with typical ramification of the apical dendrite (Fig. , 18 DIV, 7A and 7B, 7 and 13 DIV). In slices cultured in traditional media we found also a similar labeling of NeuN positive cells just after a few days (3 DIV and 7 DIV, Fig. ), but the Map2 staining revealed already a severe loss of the dendritic structures in the neurons of these slices (Fig. ). To quantify this data, we analyzed a representative area of n = 4 for traditional media and n = 4 for hCSF cultured slices at the time points 3 DIV, 7 DIV, 13 DIV and 18 DIV and counted the NeuN and Map2 double positive cells within a 450 × 450 µm field taken with a confocal microscope. We found a significant higher number of double positive cells in slices treated with hCSF compared to traditional media (Fig. , Mann Whitney test, *p<0.05). However, the number of cells positive for NeuN alone in slices cultured with traditional media compared to hCSF was not significantly different. In addition, in slices cultured in hCSF, we tested if we could label astrocytes (n = 6; 3 DIV, 6 DIV, 7 DIV, 14 DIV and 28 DIV). We found at all stages dense labeling of GFAP positive cells in all layers of the human slices cultured in hCSF (Fig. ).
Organotypic slice cultures have been a useful tool in the last decades to study the physiology and pathology of neuronal networks , , or mechanisms of neurodegenerative diseases , . Here, we show that hCSF can enhance the survival and improve network function in human organotypic slices of adult humans for up to three weeks, including ongoing network and cellular function compared to traditional medium strategies. The advantage of slice cultures is that they stay within an artificial, but nevertheless physiologically developed, composition of neurons , glia cells and even blood vessels . In this, still in parts intact, original environment, the cellular and network properties of neurons can be studied in very controlled conditions. It should be noted, that in organotypic slices many artificial changes occur that lead to a different situation compared to acute slice preparations and to the situation in vivo . These changes include alterations in the synaptic communication between neurons , a change of the intrinsic properties and morphology of the cells , and an activation of the immune cells, due to the preparation of the slices and the culturing process itself. Keeping these caveats in mind, the slice cultures can be used to address and study processes that require a long-term observation and might be difficult to engage in vivo . The human tissue that we have received for electrophysiological measurements was taken from patients undergoing a surgery for a resection of an epileptic focus. We and other groups used tissue that needed to be removed in order to get access to the pathological area, rather than pathological tissue itself. Access tissue is not considered to be part of the epileptic focus. Nevertheless, the cells cannot be termed “normal human neurons”, since they have been part of an epileptic brain for many years or may have been altered by the presence of antiepileptic drug treatment. To address this problem, Verhoog and colleagues did a careful comparison of the basic properties of human neurons of epilepsy patients and patients with a tumor . No significant differences were found in the basic parameters of the cells indicating that these parameters might reflect most likely the normal properties of neurons. Another important point that has to be considered when working with human brain slices is the variability of the tissue samples (age of patients, genetic background, taken medication, part of resected tissue). To properly address such factors intra-experimental controls are needed from the experimental tissue obtained from the same patients. In our study the culture treatment of the human slices with either hCSF or traditional medium led to a significantly different outcome, when starting from the same patient tissue. As an alternative, we and also other groups have started to use iPS cells from patients to study the physiology and pathology of human neurons. Also this approach seems to have limitations that need to be considered. Adult neurons of patients have undergone development within their natural network environment within the brain up to the time point of surgical intervention and subsequent culturing. In contrast, “human neurons” that were generated from iPSCs undergo an artificial development in a rather rapid time in vitro and may even differentiate into neuron subtypes, resembling identities that do not even physiologically exist within the human brain , . Therefore, studying mature human neurons in organotypic slice cultures as an experimental approach seems, keeping the limitations in mind, a promising and powerful tool. However, even in the same culture conditions we observed a variability in the outcome of neuronal activity of the tissue. Most slices cultured in hCSF showed a tonic firing behavior in high potassium concentrations (8 mM), while a third of the slices showed synchronized network discharges (Figs and ) and a small number did not produce any activity. It stays uncertain what determines the outcome of the slices. Surely, not all slices prepared from the same patient have had the same starting quality, due to the position in the block or other external factors. Also the thickness and size of the slices might have varied slightly and therefore plays a role in the outcome. These differences were not studied in detail, but we plan in further experiments to test the outcome of different sizes and thicknesses of human slice cultures to optimize this tool. In addition, in a recent study by Anderson and colleagues physiological firing behavior in neurons of human organotypic slices cultured in a medium based protocol was reported . However, the authors account that more than half of the neurons they recorded showed normal resting potential of approximately – 60 mV, but were not able to produce action potentials in response to depolarizing current injections . Interestingly, these results are in line with our observations reported in this study, that in medium based culture protocols the neurons seem to be able to survive for up to 14 days, but show progressively reduced or abnormal firing properties (Fig. ). At this stage we have not performed a quantitative comparison of the firing properties between cells and cell types in acute and organotypic slices, which would exceed the scope of this study. Taken together the data of our study provides strong evidence that human Cerebrospinal Fluid plays an indispensable role for the survival and physiological function of neurons. The main finding of this study is that human organotypic slices maintained in hCSF showed increased survival of neurons (Fig. ) and intact network (Figs and ) and cellular function (Figs and ) compared to slices kept in media based culture methods. So far we have not directly assessed the survival of specific neuronal subpopulations (inhibitory vs. excitatory), but the occurrence of both spontaneous inhibitory and excitatory inputs suggests a substantial survival of both major cell types. It should be noted, that in this study we did not aim to detect or define the components in the used hCSF that could explain this effect. In the literature, there are numerous suggestions that various neuroactive molecules contained in hCSF could have modulatory effects on the neuronal behavior , . However, relatively little is known about the exact mechanism by which hCSF is modulating the neuronal activity and survival , . One mechanism, that could play a key role in the survival of neurons, might be the activity of the cells itself. It has been shown that inactivity of neurons can lead to apoptosis and increased cell death , . Indeed, several studies report that the presence of hCSF leads to increased activity of neurons due to an increase in the intrinsic excitability of cells, suggesting that the activity might play a role , . However, this might just be one possible mechanism. The CSF of humans is composed of a vast amount of compounds, ranging from ions, proteins, lipoproteins and metabolic products to neuropeptides and hormones , . It has been suggested that CSF even changes its composition and can act as mediator for signal transmission , . More strikingly, in states of disease, as for example during a viral or bacterial meningitis, the composition of CSF is changing dramatically and seems thereby directly influencing the function of the brain . Taken together we propose that CSF of healthy individuals (or of normal pressure hydrocephalus patients) might represent a more physiological environment for human adult neurons providing these cells not only with nutrients and removing products of the neuronal metabolism, but also improving the electrophysiological function of the neuronal networks . Further studies have to investigate more detailed which neuroprotective molecules within hCSF are responsible for the increased survival of neurons in human organotypic brain slices and if the activity is a crucial component of this process. Indeed, the fact that mature human neurons seem to depend on different components compared to rodent cells might help to discover specific neuroprotective factors, which might not be detectable in animal models. Once identified, these molecules might open up new therapeutic routes or explain failure of the predictions of animal models. In this study, we show that human organotypic slice cultures maintained in human cerebrospinal fluid show an electrophysiological and morphological prolonged survival in comparison to slices cultured in defined culture media. This is reflected by improved single cell and network activity of these slices compared to traditional media based approaches and by an increase of the survival of adult human neurons within their original environment and with parts of the physiological network intact. Additionally, neurons of hCSF-treated organotypic slice cultures show a significantly higher number of neurons maintaining expression of the somatodendritic marker Map2 over an extended period of time (up to 18 DIV) in comparison to neurons of slices cultured in conventional media, indicating enhanced structural and morphological neuronal integrity and survival. Although the components of hCSF that are responsible for this effect are not identified yet and demand further examination, our results suggest that there might be striking differences in the components needed to culture human vs. rodent-derived CNS tissue. We propose that the human organotypic slice cultures provide a unique tool to study physiological and pathological mechanisms in mature human neurons. While clearly more work is needed to further investigate the cellular and network properties and the composition of the cells in human organotypic slices, this study provides a substantial step forward to bridge the gap between animal models and human physiology. In future studies 3D-reconstructive-confocal-microscopy based morphometric analysis (taking into account parameters such as length, number, and branches of processes, size of cell bodies and nuclei) of other cell types (e.g. astrocytes) may prove as a valuable tool to study overall tissue and circuit integrity from a morphological angle and therefore also a possible difference in the activation of astrocytes (for review see ref. ), complementing and extending the presented network and single neuron electrophysiology data and analyses of neurons via immunocytochemistry and biocytin filling. Human cortical slice cultures may serve as a platform enabling investigation of certain mechanisms and components of neuronal network plasticity, long-term effects of novel chemical compounds relevant to drug development, or studies involving viral mediated overexpression of proteins of interest.
Patients Human neocortical organotypic slice cultures were prepared from access tissue obtained from patients undergoing resective epilepsy surgery. As described before, frequently removal of cortical tissue outside the epileptic focus is required in order to get access to the pathology . For this study, we collected and included data of seven patients (Table ). All patients were surgically treated for intractable epilepsy, in one patient the histology revealed a low grade tumor (Gangliogliom WHO Grad I). Preparation of slices from patients Approval of the ethics committee of the University of Tübingen as well as written informed consent was obtained from all patients, allowing spare tissue from resective surgery to be included in our study (# 338/2016A). All experiments and methods were performed in accordance with the relevant guidelines and regulations. Tissue preparation was performed according to published protocols . Cortex was carefully microdissected and resected with only minimal use of bipolar forceps to ensure tissue integrity, transferred into ice-cold artificial (a)CSF (in mM: 110 choline chloride, 26 NaHCO 3 , 10 D-glucose, 11.6 Na-ascorbate, 7 MgCl 2 , 3.1 Na-pyruvate, 2.5 KCl, 1.25 NaH 2 PO 4 , und 0.5 CaCl 2 ) equilibrated with carbogene (95% O 2 , 5% CO 2 ) and immediately transported to the laboratory. Tissue was kept submerged in cool and carbogenated aCSF at all times. After removal of the pia, tissue chunks were trimmed perpendicular to the cortical surface and 250–350 µm thick acute slices were prepared using a Microm HM 650 V vibratome (Thermo Fisher Scientific Inc). Human organotypic slice cultures After the cortical tissue was sliced as described above slices were cut into several evenly sized pieces (~1.0 cm × 1.0 cm, Fig. ). Subsequently, the slices were transferred onto culture membranes (uncoated 30 mm Millicell-CM tissue culture inserts with 0.4 µm pores, Millipore) and kept in six–well culture dishes (BD Biosciences). The plates were stored in an incubator (ThermoScientific) at 37 °C, 5% CO 2 and 100% humidity. For electrophysiological measurements slice cultures were transferred into the recording chamber of a patch clamp rig. After finishing patch clamp- or extracellular recordings, slices were fixed in 4% PFA and processed for immunocytochemistry, as detailed in the Immunocytochemistry section. Traditional culture media Human organotypic slice cultures were cultured in organotypic slice culture medium (OSCM) without serum as described by Eugène and colleagues or BrainPhys medium according to Bardy and colleagues . OSCM medium was produced by adding select components (see Eugene et al . for details) to commercially available DMEM/F12 and BME media. BrainPhys is a serum-free neuronal medium with adjusted concentrations of inorganic salts, neuroactive amino acids, and energetic substrates (see Bardy et al . 2015 for details). The media was originally developed to improve the electrical and synaptic activity of mature human neurons. Both media were found to exhibit comparable effects on human organotypic slices in our experiments and respective data were pooled and will be referred to as traditional media. CSF collection Human cerebrospinal fluid (hCSF) was collected from patients with normal pressure hydrocephalus (NPH). We received and pooled hCSF of several patients with NPH who needed to undergo a lumbar puncture as part of diagnostic or therapeutic procedures (up to 40 ml per patient). Approval of the ethics committee of the University of Tübingen as well as written informed consent was obtained from all patients. It is well established and known from daily clinical practice that hCSF of NPH patients exhibits physiological/normal hCSF chemistry values (lactate, glucose, cell count, protein levels – see Table ) undistinguishable from the ones of healthy individuals . The CSF was centrifuged at 4000 rpm at 4 °C for 10 minutes and the supernatant was collected and stored at −80 °C within one hour after lumbar puncture. Extracellular Multi-Unit recording For extracellular and intracellular recordings slices were transferred into a recording chamber and continuously superfused with artificial CSF (aCSF) containing (in mM) 118 NaCl, 3 KCl, 1.5 CaCl 2 , 1 MgCl 2 , 25 NaHCO 3 , 30 D-glucose, and equilibrated with carbogen (95% NaH 2 PO 4 , and 30 O2–5% CO 2 , pH 7.4) in a recycling system. Temperature was maintained at 30 ± 1 °C. The signal contained multiunit action potential (AP) activity. Extracellular signals were amplified 10,000-fold and filtered between 0.25 and 1.5 kHz. To facilitate detection of network-bursts, this signal was rectified and integrated by an electronic integrator with a time constant of 50–100 ms. We used an extracellular amplifier Modell 1700 (Am-Systems) or NPI (Model Ext 10–2 F) and an integrator from NPI (Model INR-011). Whole-cell patch clamp recordings Slices were positioned in a submerged-type recording chamber (Scientifica, United Kingdom/Warner apparatus), continuously superfused with aCSF and visualized with a BX61WI Microscope (Olympus) or a Leica stereomicroscope. Recordings were performed using recording electrodes with a resistance of 3–5 MΩ and filled with a whole-cell patch-clamp pipette solution containing the following components (in mM): 140 K-gluconic acid, 1 CaCl 2 *6H 2 O, 10 EGTA, 2 MgCl 2 *6H 2 O, 4 Na 2 ATP, and 10 HEPES, pH 7.2. The intracellular pipette solution contained biocytin (5 mg/ml) to allow for posthoc identification of the location and morphology of recorded neurons. Whole cell current-clamp recordings were obtained from cortical neurons using the visual-patch or blind patch technique. Whole-cell patch-clamp recordings were obtained with a sampling rate of 10 kHz and a low-pass filter setting of 2 kHz. Recordings were performed with unpolished patch electrodes manufactured from borosilicate glass pipettes with filament (Science products). Patch-clamp experiments were performed with a patch-clamp amplifier (Multiclamp 200B) or a NPI Bridge Amplifier (Model BA-01X), a digitizing interface (Digidata 1440 A or 1550 A Digidata), and pClamp 10 software (Molecular Devices). The junction potential was calculated and subtracted offline to correct the membrane potential in current clamp mode. After recording, the slices were placed in neutral buffered 4% PFA solution at 4 °C overnight for fixation followed by three rinses in PBS and by subsequent immunocytochemical staining procedures. Immunocytochemistry and image collection For immunocytochemistry (ICC) slices were incubated for 60 minutes in goat block (PBS, 0.2% triton X-100, 5% normal goat serum) before primary antibodies were added in respective dilutions. Due to the thickness of the slices incubation in primary antibodies was performed overnight at 4 °C temperature and for 30 minutes at room temperature on the following morning. Slices were rinsed four times in PBS supplemented with 0.1% Triton X-100 and then incubated in Alexa-fluorophore-conjugated secondary antibodies in goat block (PBS, 0.2% triton X-100, 1% normal goat serum) for 1 hour and 15 minutes at room temperature. After four PBS rinses (2x in PBS supplemented with 0.1% Triton X-100 and 2x in PBS) the slices were stained with DAPI (1:5000 in PBS for 2–5 minutes). After four final rinses (3x PBS, 1x Ampuwa water) the slices were mounted in Fluoromount G (SouthernBiotech) on glass slides. Primary and secondary antibodies were used at the the following dilutions: mouse anti-NeuN IgG (1:100, Merck, ABN90), chicken anti-Map2 (1:500, Abcam, ab5392), rabbit anti-GFAP (1:250, Agilent, Z0334), goat anti-mouse IgG Alexa 568 (1:500, life technologies, A-11004), goat anti-rabbit IgG Alexa 568 (1:500, life technologies, A-11011) and goat anti-chicken Alexa 488 (1:500, life technologies, A-11039). Expression of tissue-specific markers was evaluated using a conventional fluorescenece microscope (Zeiss Imager.Z1) and laser confocal scanning microscopic analysis (Zeiss LSM 510 Laser Module, Zeiss Axiovert 200 M). Images were processed with Photoshop CS (Adobe Systems) and CorelDRAW (Corel Corporation).
Human neocortical organotypic slice cultures were prepared from access tissue obtained from patients undergoing resective epilepsy surgery. As described before, frequently removal of cortical tissue outside the epileptic focus is required in order to get access to the pathology . For this study, we collected and included data of seven patients (Table ). All patients were surgically treated for intractable epilepsy, in one patient the histology revealed a low grade tumor (Gangliogliom WHO Grad I).
Approval of the ethics committee of the University of Tübingen as well as written informed consent was obtained from all patients, allowing spare tissue from resective surgery to be included in our study (# 338/2016A). All experiments and methods were performed in accordance with the relevant guidelines and regulations. Tissue preparation was performed according to published protocols . Cortex was carefully microdissected and resected with only minimal use of bipolar forceps to ensure tissue integrity, transferred into ice-cold artificial (a)CSF (in mM: 110 choline chloride, 26 NaHCO 3 , 10 D-glucose, 11.6 Na-ascorbate, 7 MgCl 2 , 3.1 Na-pyruvate, 2.5 KCl, 1.25 NaH 2 PO 4 , und 0.5 CaCl 2 ) equilibrated with carbogene (95% O 2 , 5% CO 2 ) and immediately transported to the laboratory. Tissue was kept submerged in cool and carbogenated aCSF at all times. After removal of the pia, tissue chunks were trimmed perpendicular to the cortical surface and 250–350 µm thick acute slices were prepared using a Microm HM 650 V vibratome (Thermo Fisher Scientific Inc).
After the cortical tissue was sliced as described above slices were cut into several evenly sized pieces (~1.0 cm × 1.0 cm, Fig. ). Subsequently, the slices were transferred onto culture membranes (uncoated 30 mm Millicell-CM tissue culture inserts with 0.4 µm pores, Millipore) and kept in six–well culture dishes (BD Biosciences). The plates were stored in an incubator (ThermoScientific) at 37 °C, 5% CO 2 and 100% humidity. For electrophysiological measurements slice cultures were transferred into the recording chamber of a patch clamp rig. After finishing patch clamp- or extracellular recordings, slices were fixed in 4% PFA and processed for immunocytochemistry, as detailed in the Immunocytochemistry section.
Human organotypic slice cultures were cultured in organotypic slice culture medium (OSCM) without serum as described by Eugène and colleagues or BrainPhys medium according to Bardy and colleagues . OSCM medium was produced by adding select components (see Eugene et al . for details) to commercially available DMEM/F12 and BME media. BrainPhys is a serum-free neuronal medium with adjusted concentrations of inorganic salts, neuroactive amino acids, and energetic substrates (see Bardy et al . 2015 for details). The media was originally developed to improve the electrical and synaptic activity of mature human neurons. Both media were found to exhibit comparable effects on human organotypic slices in our experiments and respective data were pooled and will be referred to as traditional media.
Human cerebrospinal fluid (hCSF) was collected from patients with normal pressure hydrocephalus (NPH). We received and pooled hCSF of several patients with NPH who needed to undergo a lumbar puncture as part of diagnostic or therapeutic procedures (up to 40 ml per patient). Approval of the ethics committee of the University of Tübingen as well as written informed consent was obtained from all patients. It is well established and known from daily clinical practice that hCSF of NPH patients exhibits physiological/normal hCSF chemistry values (lactate, glucose, cell count, protein levels – see Table ) undistinguishable from the ones of healthy individuals . The CSF was centrifuged at 4000 rpm at 4 °C for 10 minutes and the supernatant was collected and stored at −80 °C within one hour after lumbar puncture.
For extracellular and intracellular recordings slices were transferred into a recording chamber and continuously superfused with artificial CSF (aCSF) containing (in mM) 118 NaCl, 3 KCl, 1.5 CaCl 2 , 1 MgCl 2 , 25 NaHCO 3 , 30 D-glucose, and equilibrated with carbogen (95% NaH 2 PO 4 , and 30 O2–5% CO 2 , pH 7.4) in a recycling system. Temperature was maintained at 30 ± 1 °C. The signal contained multiunit action potential (AP) activity. Extracellular signals were amplified 10,000-fold and filtered between 0.25 and 1.5 kHz. To facilitate detection of network-bursts, this signal was rectified and integrated by an electronic integrator with a time constant of 50–100 ms. We used an extracellular amplifier Modell 1700 (Am-Systems) or NPI (Model Ext 10–2 F) and an integrator from NPI (Model INR-011).
Slices were positioned in a submerged-type recording chamber (Scientifica, United Kingdom/Warner apparatus), continuously superfused with aCSF and visualized with a BX61WI Microscope (Olympus) or a Leica stereomicroscope. Recordings were performed using recording electrodes with a resistance of 3–5 MΩ and filled with a whole-cell patch-clamp pipette solution containing the following components (in mM): 140 K-gluconic acid, 1 CaCl 2 *6H 2 O, 10 EGTA, 2 MgCl 2 *6H 2 O, 4 Na 2 ATP, and 10 HEPES, pH 7.2. The intracellular pipette solution contained biocytin (5 mg/ml) to allow for posthoc identification of the location and morphology of recorded neurons. Whole cell current-clamp recordings were obtained from cortical neurons using the visual-patch or blind patch technique. Whole-cell patch-clamp recordings were obtained with a sampling rate of 10 kHz and a low-pass filter setting of 2 kHz. Recordings were performed with unpolished patch electrodes manufactured from borosilicate glass pipettes with filament (Science products). Patch-clamp experiments were performed with a patch-clamp amplifier (Multiclamp 200B) or a NPI Bridge Amplifier (Model BA-01X), a digitizing interface (Digidata 1440 A or 1550 A Digidata), and pClamp 10 software (Molecular Devices). The junction potential was calculated and subtracted offline to correct the membrane potential in current clamp mode. After recording, the slices were placed in neutral buffered 4% PFA solution at 4 °C overnight for fixation followed by three rinses in PBS and by subsequent immunocytochemical staining procedures.
For immunocytochemistry (ICC) slices were incubated for 60 minutes in goat block (PBS, 0.2% triton X-100, 5% normal goat serum) before primary antibodies were added in respective dilutions. Due to the thickness of the slices incubation in primary antibodies was performed overnight at 4 °C temperature and for 30 minutes at room temperature on the following morning. Slices were rinsed four times in PBS supplemented with 0.1% Triton X-100 and then incubated in Alexa-fluorophore-conjugated secondary antibodies in goat block (PBS, 0.2% triton X-100, 1% normal goat serum) for 1 hour and 15 minutes at room temperature. After four PBS rinses (2x in PBS supplemented with 0.1% Triton X-100 and 2x in PBS) the slices were stained with DAPI (1:5000 in PBS for 2–5 minutes). After four final rinses (3x PBS, 1x Ampuwa water) the slices were mounted in Fluoromount G (SouthernBiotech) on glass slides. Primary and secondary antibodies were used at the the following dilutions: mouse anti-NeuN IgG (1:100, Merck, ABN90), chicken anti-Map2 (1:500, Abcam, ab5392), rabbit anti-GFAP (1:250, Agilent, Z0334), goat anti-mouse IgG Alexa 568 (1:500, life technologies, A-11004), goat anti-rabbit IgG Alexa 568 (1:500, life technologies, A-11011) and goat anti-chicken Alexa 488 (1:500, life technologies, A-11039). Expression of tissue-specific markers was evaluated using a conventional fluorescenece microscope (Zeiss Imager.Z1) and laser confocal scanning microscopic analysis (Zeiss LSM 510 Laser Module, Zeiss Axiovert 200 M). Images were processed with Photoshop CS (Adobe Systems) and CorelDRAW (Corel Corporation).
|
Gastric mucin phenotype indicates aggressive biological behaviour in early differentiated gastric adenocarcinomas following endoscopic treatment | 15939d7a-0482-4203-a39f-8eb83cd82548 | 8276406 | Anatomy[mh] | Gastric cancer (GC), one of the most common human cancers worldwide, is a disease with multiple pathogenic factors, various prognoses and different responses to treatments. Thus, properly distinguishing those with worse prognoses from those with better prognoses appears to be significantly important. Four different morphology -based classification systems exist, the World Health Organization (WHO/2019) , the Japanese Gastric Cancer Association (JGCA/2017) , Laurén and Nakamura . According to the WHO classification, GCs are subclassified into papillary, tubular, poorly cohesive, mucinous and mixed types. In the JGCA classification, the subtypes are papillary (pap), tubular (tub), poorly differentiated (por), signet-ring cell (sig), and mucinous (muc), which are similar to the subtypes used by the WHO. GCs are divided into intestinal and diffuse types using Laurén’s classification or into differentiated and undifferentiated types based on Nakamura’s classification [ – ]. The differentiated type contains pap, tub1, and tub2 according to the JGCA classification and papillary and well/moderately differentiated adenocarcinoma according to the WHO classification. These different histological types exhibit distinct biological behaviours. The mucous produced by cancers is one of the factors determining the nature of biological behaviour. The main component of mucous is a high-molecular-weight glycoprotein called mucin . As cancer progresses, the nature of the mucous changes relative to the degree of biological malignancy. In the 1990 s, with the progress of structural analysis of mucin and the widespread use of monoclonal antibodies to the core protein of mucin, a mucin phenotype subclassification emerged. Mucin phenotype subclassification was entirely based on the mucin expression pattern, independent of histological features. Thus, GCs are classified into gastric, intestinal, gastrointestinal and null mucin phenotypes [ – ]. Previous studies have reported that the gastric phenotype has a higher potential for invasion and metastasis than the intestinal type, which results in a worse prognosis of GCs [ – ]. However, studies have mostly focused on advanced gastric cancers, and early gastric cancers are rarely investigated. Early gastric cancer (EGC) is defined as tumour invasion confined to the mucosa and submucosa, irrespective of regional lymph node metastasis . Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) are used as treatments for some intramucosal carcinomas and submucosal lesions, which have a very low probability of lymph node metastasis . To our knowledge, there is no research exploring biological role of mucin phenotypes in EGCs using EMR/ESD by Chinese investigators. Little information is available on the effects of mucin phenotypes on the clinicopathological features of EGCs in a Chinese cohort. Accordingly, we examined mucin expression and mucin phenotypes and explored mucin phenotype clinicopathological characteristics and biological behaviour.
Patients and tissue specimens Our study consisted of 257 consecutive patients who underwent EMR/ESD for differentiated EGCs between January 2012 and June 2018 at The Second Affiliated Hospital of Zhejiang University, China. The group comprised 182 men and 75 women with an age range of 29 to 87 (mean 64) years old. The location of each lesion was classified in terms of the upper (27 cases), middle (46 cases) and lower (184 cases) thirds of the stomach. The size of each lesion was measured by the maximum diameter, which ranged from 0.1 to 6.5 (mean 1.5) centimetres. The protruding category (52 cases) included type 0-I and 0-IIa, the depressed (102 cases) category contained type 0-IIc and III, and all the other cases were considered under the protruding and depressed category (103 cases). Based on the WHO and JGCA classification, the EGCs were subclassified into well differentiated tubular adenocarcinoma (well-diff/tub 1, 198 cases), moderately differentiated tubular adenocarcinoma (mod-diff/tub 2, 37 cases), papillary adenocarcinoma (3 cases) and mixed (tub-pap/sig/por, 19 cases). In the mixed cases, the undifferentiated components (sig/por) were less than 50 %. Immunohistochemistry All specimens were fixed with 10 % buffered formalin, embedded in paraffin, cut into 4-µm-thick sections, and subjected to haematoxylin and eosin (HE) staining. MUC2 was detected by mAb Ccp58 (Zsbio, 1:100), MUC5AC by mAb MRQ-19 (Zsbio, 1:100), MUC6 by mAb MRQ-20 (Zsbio, 1:100), and CD10 (Zsbio, 1:100) by immunohistochemistry (IHC). IHC was performed by using the Ventana NexES Staining System (Roche, Benchmark®XT). The marker CD10 exhibited both cytoplasmic and glandular luminal reactivity, whereas MUC5AC, MUC6 and MUC2 exhibited only cytoplasmic reactivity. The staining results were categorized as positive when at least one single cell among the carcinoma cells was stained and negative when none of the carcinoma cells were stained . Classification of mucin phenotype Based on MUC5AC, MUC6, MUC2 and CD10, EGCs can be classified into the gastric phenotype (G-type), gastrointestinal phenotype (GI-type), intestinal phenotype (I-type) and null phenotype (N-type) [ , , – ]. The following criteria were used for the classification of mucin phenotypes: ① the G-type shows positive staining for at least one of the MUC5AC and MUC6, while CD10 and MUC2 are both negative; ② the I-type shows positive staining for CD10 and/or MUC2, while MUC5AC and MUC6 are both negative; ③ the GI-type shows positive staining for marker CD10 and/or MUC2 associated with one of the markers MUC5AC and MUC6 positive; and④ if none of the four markers are positive, the phenotype is classified as N-type. In addition, among the GI-types, we labelled one expressing more MUC5AC and/or MUC6 than MUC2 and CD10 as a gastric-predominant GI phenotype (GI-G type); otherwise, we labelled it as an intestinal-predominant phenotype (GI-I type). Statistical analysis Associations between mucin expression profiles and clinicopathological parameters were examined by the chi-square test or Fisher’s exact test. Statistical significance was established to be P < 0.05. Statistical calculations were performed with IBM SPSS Statistics (version 23.0).
Our study consisted of 257 consecutive patients who underwent EMR/ESD for differentiated EGCs between January 2012 and June 2018 at The Second Affiliated Hospital of Zhejiang University, China. The group comprised 182 men and 75 women with an age range of 29 to 87 (mean 64) years old. The location of each lesion was classified in terms of the upper (27 cases), middle (46 cases) and lower (184 cases) thirds of the stomach. The size of each lesion was measured by the maximum diameter, which ranged from 0.1 to 6.5 (mean 1.5) centimetres. The protruding category (52 cases) included type 0-I and 0-IIa, the depressed (102 cases) category contained type 0-IIc and III, and all the other cases were considered under the protruding and depressed category (103 cases). Based on the WHO and JGCA classification, the EGCs were subclassified into well differentiated tubular adenocarcinoma (well-diff/tub 1, 198 cases), moderately differentiated tubular adenocarcinoma (mod-diff/tub 2, 37 cases), papillary adenocarcinoma (3 cases) and mixed (tub-pap/sig/por, 19 cases). In the mixed cases, the undifferentiated components (sig/por) were less than 50 %.
All specimens were fixed with 10 % buffered formalin, embedded in paraffin, cut into 4-µm-thick sections, and subjected to haematoxylin and eosin (HE) staining. MUC2 was detected by mAb Ccp58 (Zsbio, 1:100), MUC5AC by mAb MRQ-19 (Zsbio, 1:100), MUC6 by mAb MRQ-20 (Zsbio, 1:100), and CD10 (Zsbio, 1:100) by immunohistochemistry (IHC). IHC was performed by using the Ventana NexES Staining System (Roche, Benchmark®XT). The marker CD10 exhibited both cytoplasmic and glandular luminal reactivity, whereas MUC5AC, MUC6 and MUC2 exhibited only cytoplasmic reactivity. The staining results were categorized as positive when at least one single cell among the carcinoma cells was stained and negative when none of the carcinoma cells were stained .
Based on MUC5AC, MUC6, MUC2 and CD10, EGCs can be classified into the gastric phenotype (G-type), gastrointestinal phenotype (GI-type), intestinal phenotype (I-type) and null phenotype (N-type) [ , , – ]. The following criteria were used for the classification of mucin phenotypes: ① the G-type shows positive staining for at least one of the MUC5AC and MUC6, while CD10 and MUC2 are both negative; ② the I-type shows positive staining for CD10 and/or MUC2, while MUC5AC and MUC6 are both negative; ③ the GI-type shows positive staining for marker CD10 and/or MUC2 associated with one of the markers MUC5AC and MUC6 positive; and④ if none of the four markers are positive, the phenotype is classified as N-type. In addition, among the GI-types, we labelled one expressing more MUC5AC and/or MUC6 than MUC2 and CD10 as a gastric-predominant GI phenotype (GI-G type); otherwise, we labelled it as an intestinal-predominant phenotype (GI-I type).
Associations between mucin expression profiles and clinicopathological parameters were examined by the chi-square test or Fisher’s exact test. Statistical significance was established to be P < 0.05. Statistical calculations were performed with IBM SPSS Statistics (version 23.0).
Expression of mucin markers and mucin phenotype in early gastric cancers The expression percentages of CD10, MUC2, MUC5AC and MUC6 in all EGCs were 43.58 % (112/257), 63.81 % (164/257), 64.98 % (167/257) and 72.76 % (187/257) respectively (Fig. ). Two hundred fifty-seven EGCs were classified as G-type (21 %, 54/257), GI-type (56 %, 144/257), I-type (20 %, 51/257) and N-type (3 %, 8/257). The GI-type contained the GI-G type (72 %, 103/144) and GI-I type (28 %, 41/144). There were more cases of G-type and GI-G type (61 %, 157/257) than I-type and GI-I type (36 %, 95/257) (Fig. ). Relationship between mucin phenotype and clinicopathological features The relationship between mucin phenotype and clinicopathological features is summarized in Table . The mucin phenotypes were significantly related to the JGCA and WHO classifications ( P < 0.05), but the parameters of sex, age, margin, colour, tumour size, gross type, depth of invasion, and lymphovascular invasion did not significantly differ among those mucin phenotypes ( P > 0.05). The I-type had the highest proportion of differentiated EGCs among the four mucin phenotypes (100.0 % vs. 79.7 % vs. 83.1 % vs. 87.5 %, P = 0.027). The G-type group had a higher proportion of tub-por/sig and pap/tub-pap cases than the I-, GI- and N-type groups (20.4 % vs. 7.0 % vs. 0.0 % vs. 12.5 %, P = 0.027) according to the JGCA classification. According to the WHO classification, the G-type had more mixed histological components (18.5 % vs. 5.6 % vs. 0.0 % vs. 12.5 %, P = 0.006) than the other mucin phenotypes. Relationship between mucin phenotypes and background mucosa Intestinal metaplasia (IM) of background mucosa was observed in 199 of 249 (79.9 %) cases (G-, GI- and GI-type), including 38 cases of incomplete IM and 161 cases of complete IM. IM did not significantly differ among mucin phenotypes ( P > 0.05). However, the presence of incomplete and complete IM was significantly different in distinct mucin phenotypes ( P = 0.004, P = 0.018). The presence of incomplete IM in GI-type EGCs was higher than that in the G-type and I-type EGCs (21.5 % vs. 9.3 % vs. 3.9 %). In the contrast, 77.8 % (42/54) of G-type EGCs and 70.6 % (36/51) of I-type EGCs exhibited complete IM, which was higher than the 57.6 % (83/114) of GI-type EGCs. The IM status of the background mucosa and the relationship with mucin phenotypes are shown in Table . Biological behaviour of mucin phenotypes In addition to tubular adenocarcinoma components, 22 of the 257 cases also contained components of papillary adenocarcinoma, poorly differentiated carcinoma, or signet ring cell carcinoma. Fifteen cases (68.18 %) contained por/sig components, and the other 7 (31.82 %) contained pap components. Eleven cases showed the G-type, 10 cases showed the GI-type, only one case showed the N-type, and none of them showed the I-type. Among the 10 GI-type cases, 9 cases showed the GI-G type, and one showed the GI-I type. Almost all the 22 patients showed the G- and GI-G types, which was significantly higher than the number of I-type patients ( P = 0.011). In addition, the 22 patients had a higher proportion of infiltration into the submucosa ( P < 0.001). (Table ; Figs. and ). Follow-up Six patients underwent additional gastrectomy, and there was no residual tumour or lymph node metastasis. All patients were under close follow-up, and neither recurrence nor metastasis was detected.
The expression percentages of CD10, MUC2, MUC5AC and MUC6 in all EGCs were 43.58 % (112/257), 63.81 % (164/257), 64.98 % (167/257) and 72.76 % (187/257) respectively (Fig. ). Two hundred fifty-seven EGCs were classified as G-type (21 %, 54/257), GI-type (56 %, 144/257), I-type (20 %, 51/257) and N-type (3 %, 8/257). The GI-type contained the GI-G type (72 %, 103/144) and GI-I type (28 %, 41/144). There were more cases of G-type and GI-G type (61 %, 157/257) than I-type and GI-I type (36 %, 95/257) (Fig. ).
The relationship between mucin phenotype and clinicopathological features is summarized in Table . The mucin phenotypes were significantly related to the JGCA and WHO classifications ( P < 0.05), but the parameters of sex, age, margin, colour, tumour size, gross type, depth of invasion, and lymphovascular invasion did not significantly differ among those mucin phenotypes ( P > 0.05). The I-type had the highest proportion of differentiated EGCs among the four mucin phenotypes (100.0 % vs. 79.7 % vs. 83.1 % vs. 87.5 %, P = 0.027). The G-type group had a higher proportion of tub-por/sig and pap/tub-pap cases than the I-, GI- and N-type groups (20.4 % vs. 7.0 % vs. 0.0 % vs. 12.5 %, P = 0.027) according to the JGCA classification. According to the WHO classification, the G-type had more mixed histological components (18.5 % vs. 5.6 % vs. 0.0 % vs. 12.5 %, P = 0.006) than the other mucin phenotypes.
Intestinal metaplasia (IM) of background mucosa was observed in 199 of 249 (79.9 %) cases (G-, GI- and GI-type), including 38 cases of incomplete IM and 161 cases of complete IM. IM did not significantly differ among mucin phenotypes ( P > 0.05). However, the presence of incomplete and complete IM was significantly different in distinct mucin phenotypes ( P = 0.004, P = 0.018). The presence of incomplete IM in GI-type EGCs was higher than that in the G-type and I-type EGCs (21.5 % vs. 9.3 % vs. 3.9 %). In the contrast, 77.8 % (42/54) of G-type EGCs and 70.6 % (36/51) of I-type EGCs exhibited complete IM, which was higher than the 57.6 % (83/114) of GI-type EGCs. The IM status of the background mucosa and the relationship with mucin phenotypes are shown in Table .
In addition to tubular adenocarcinoma components, 22 of the 257 cases also contained components of papillary adenocarcinoma, poorly differentiated carcinoma, or signet ring cell carcinoma. Fifteen cases (68.18 %) contained por/sig components, and the other 7 (31.82 %) contained pap components. Eleven cases showed the G-type, 10 cases showed the GI-type, only one case showed the N-type, and none of them showed the I-type. Among the 10 GI-type cases, 9 cases showed the GI-G type, and one showed the GI-I type. Almost all the 22 patients showed the G- and GI-G types, which was significantly higher than the number of I-type patients ( P = 0.011). In addition, the 22 patients had a higher proportion of infiltration into the submucosa ( P < 0.001). (Table ; Figs. and ).
Six patients underwent additional gastrectomy, and there was no residual tumour or lymph node metastasis. All patients were under close follow-up, and neither recurrence nor metastasis was detected.
The mucin phenotype classification is based on the mucin marker expression profile. After year 2000, the gastric and intestinal mucin phenotypes were analysed by IHC . The mucin markers MUC5AC, MUC6, MUC2 and CD10 were considered necessary, although there is no consensus on the number of markers that should be used to define a mucin phenotype or the percentage of tumour cells that must be stained [ – , , , ]. MUC5AC is a secreted mucin expressed in the surface mucous epithelium of normal gastric mucosa. High expression of MUC6 is observed in fundic mucous neck cells and pyloric glands of gastric mucosa. CD10 is a marker for the brush border on the luminal surface of the small intestine. In the normal adult intestine, MUC2 expression is observed in the perinuclear areas of goblet cells. We showed that the expression of MUC5AC, MUC6, MUC2 and CD10 was detected in 167 (64.98 %), 187 (72.76 %), 164 (63.81 %), and 112 (43.58 %) of the 257 EGCs, respectively. In previous studies, the expression percentages of MUC5AC, MUC6, MUC2 and CD10 in GCs were 55.1-67.5 %, 44.9-64 %, 35.4-49.3 % and 20.6-20.9 %, respectively , and for EGCs, the expression of each mucin marker was 68.75-96.8 %, 19.6-71.58 %, 25-62.10 %, and 0-79 %, respectively [ , , ]. Based on the combinations of expression of these markers, the 257 EGCs were classified into the G-type (21 %, 54/257), GI-type (56 %, 144/257), I-type (20 %, 51/257) and N-type (3 %, 8/257); in previous reports, the incidence percentages of each of these mucin phenotypes were found to be 15-41.1 %, 20.3-60.1 %, 18.5-46.6 %, and 3.7-31.6 %, respectively, in advanced GCs [ , , ], and 7.9-36.8 %, 18.8-41.2 %, 15.4-55.56 %, and 0-11.1 %, respectively, in early- stage GCs [ , , – ]. Our results were consistent with these studies. The reported expression ranges vary greatly among different investigators, and different markers, antibodies and case groups may account for this discrepancy. Koyama et al. reported that the incidence of G-type was 19.3 % , which was similar to that found in the present study (21 %); however, in his report, the incidence of I-type was much higher than that of G-type (43.8 % vs. 19.3 %), as was reported by Fabio et al. . While Tajima et al. reported the opposite result, since in their study, the incidence of the G-type was much higher than that of the I-type (36.8 % vs. 15.4 %). In our study, the incidence of G-type was almost the same as that of I-type (21 % vs. 20 %). Overall, based on our data, much more than half of these cases were classified as G- and GI-G type GC (61.09 %, 157/257), which is much higher than the incidence of I- and GI-I type GC (36.96 %,95/257). A previous report revealed that almost all intramucosal GC cases exhibited the gastric phenotype, including the GI phenotype . The relationship between mucin phenotypes and clinicopathological features was investigated. We found that histology classification (both the JGCA and WHO classification) was closely related to the mucin phenotype. The incidence of I-type was greater than those of the G-, GI- and N-type (100.0 % vs. 79.7 %, 93.1 %, 87.5 %) in differentiated tubular adenocarcinoma. The G-type was histologically significantly correlated with the mixed type (with poorly differentiated/papillary carcinoma). Our data showed that the proportion of G-type carcinoma increased during the transition from solely differentiated type to mixed type carcinoma. Mixed-type early-stage carcinoma more frequently expressed G-type mucin, and G-type tumours were associated with a higher rate of undifferentiated-type tumours than I-type tumours . There were no significant differences between mucin phenotypes and other parameters, including sex, age, margin, colour, tumour size, gross type, depth of invasion, and lymphovascular invasion ( P > 0.05). These results are consistent with those of other studies in the literatures [ , , ], and there was no clear correlation between phenotypes and clinicopathological characteristics, including sex, age, tumour size, location, macroscopic features, lymphatic or venous invasion, or lymph node metastasis in the case of the differentiated type [ , , ]. Koseki et al. and Oya et al. reported that the incidence of lymphatic invasion, venous invasion and lymph node metastasis in gastric phenotype carcinomas was significantly higher than that in intestinal phenotype carcinomas. In addition, G-type EGCs were correlated with some distinct macroscopic features, namely, a smaller tumour diameter , discoloured surface and non-wavy tumour margins . G-type differentiated adenocarcinomas showed a depressed type, indistinct margins and monotonous colour tone across the mucosal layer, whereas I-type adenocarcinomas had an elevated, distinct margin and a red mucosa [ , , ]. The discrepancy of these results may have been due to heterogeneous components that contained poorly differentiated adenocarcinoma . Intestinal metaplasia has been frequently observed surrounding GC, especially differentiated adenocarcinomas. IM has malignant potential and has been regarded as a precursor of gastric neoplasms. According to Laurén, intestinal-type adenocarcinoma is preceded by metaplastic changes, while diffuse-type adenocarcinoma arises in non-IM gastric mucosa . In the current study, background mucosal IM was observed in 79.9 % of cases among the G-, GI- and I-type EGCs and 87.0 % of cases among the G-type EGCs. 25 % of I-type cases arose from the normal mucosa without IM. IM did not significantly differ among mucin phenotypes ( P > 0.05). However, incomplete and complete IM significantly differed with respect to mucin phenotypes (P = 0.004, P = 0,018). A total of 77.78 % (42/54) of G-type and 70.6 % (36/51) of I-type patients had complete IM, which was higher than the rates among GI-type patients (83/114, 57.6 %). The expression of incomplete IM in GI-type EGCs was higher than in G- and I-type EGCs (21.5 % vs. 9.3 % vs. 3.9 %). Our results demonstrated a remarkable difference between mucin phenotypes and the background mucosa. Similar results have been reported by Kabashima et al. and Matsuoka . The mucin phenotype of the carcinoma was independent of mucin phenotypic changes in the surrounding mucosa, and the carcinoma may undergo individual intestinalization. The G-type may imitate the surrounding mucosa, and the carcinomas and the background mucosa have an unstable status, as they commonly possess the hybrid phenotype of the stomach and the small intestine . Mucin phenotypes can indicate biological behaviour in GCs. G-type GCs have increased potential for invasion and metastasis due to infiltrating of deeper layers or more surrounding structures, a higher rate of lymph node metastasis, and poorer prognosis [ , , , ]. Even differentiated adenocarcinomas of the G-type had similar outcomes, focused on prognoses, as undifferentiated adenocarcinomas [ – ]. In our research, six patients underwent additional gastrectomy, and there was no residual tumour or lymph node metastasis. All patients were under close follow-up, and neither recurrence nor metastasis was detected. The mixed type (mixed with poorly differentiated or papillary adenocarcinoma) was mainly of the G-type, which was significantly higher than that of purely differentiated tubular adenocarcinoma ( P < 0.05), and the depth of infiltration was deeper ( P < 0.05). The G-type group had the highest proportion (11/54, 20.37 %) with poorly differentiated/undifferentiated components, and almost all of them (19/22, 86.36 %) expressed the G- and GI-G types. The mixed type may represent a progressive loss of glandular structure during progression of the cancer from the mucosa to advanced stage, and those with submucosal invasion was a risk factor for lymph node metastasis [ , , ]. Differentiated EGC of G-type frequently changed histologically into signet ring-cell carcinoma or poorly differentiated adenocarcinoma. These results may imply more aggressive biological behaviour and poorer prognosis.
Our study reports the expression of mucin markers (MUC5AC, MUC6, MUC2 and CD10) and mucin phenotypes in differentiated EGC samples from ESD/EMR in the Chinese population. Mucin phenotypes of early differentiated gastric cancer are of clinical significance, and G-type GC exhibits aggressive biological behaviour in early differentiated GCs, especially in those with poorly differentiated adenocarcinoma or papillary adenocarcinoma components.
|
Pleural carcinoid diagnosed via video-assisted thoracoscopy biopsies in a patient with recurrent unilateral pleural effusion at St. Francis hospital Nsambya: a case report | d5f71718-2f19-42ff-aad6-80fa9dab9d4f | 11658405 | Biopsy[mh] | Pulmonary carcinoids are rare neuroendocrine tumors, constituting approximately 0.2–2% of the population . They are classified into two subcategories, namely typical carcinoids and atypical carcinoids, with the latter comprising only 10–15% of the cases. A study conducted among patients with pulmonary neuroendocrine tumors indicated that approximately 80% of pulmonary carcinoids are centrally located within the lung while 20% are in peripheral regions and are typically associated with the airways. Historically, pulmonary carcinoid tumors exhibit heterogeneous patterns including alveolar, organoid, trabecular, spindle and solid patterns . Cases have been reported where pleural carcinoids present as exudative pleural effusions yet aspirates were negative for infectious disease after thorough work-up . Previously, open thoracotomy was the standard approach for pulmonary resection. However, in the recent years, video-assisted thoracoscopy (VAT) has gained popularity and become the preferred technique for resectioning pulmonary carcinoids . Video-assisted thoracoscopy (VAT) has emerged as a crucial tool, replacing open procedures for lung and pleural biopsies. It offers high diagnostic accuracy with fewer complications compared with open procedures. The diagnostic accuracy of VAT in diagnosing cancer and tuberculosis of the pleura is approximately 95% . VAT is commonly used to evaluate pleural effusions when less invasive tests such as thoracentesis are inconclusive and also plays a role in pleurodesis when chemical pleurodesis is contraindicated or ineffective. VAT is a viable alternative to open thoracotomy for both diagnostic and therapeutic purposes in lung and pleural lesions. Studies show no significant difference in long term survival between VAT and open procedures but VAT has fewer postoperative complications and a shorter hospital stay . VAT is particularly beneficial for the elderly patients and those who cannot tolerate general anesthesia. It can be performed using general anesthesia with selective intubation or local anesthesia as awake video assisted thoracoscopy (AVAT) . Although AVAT has shown comparable outcomes with VAT in procedures such as pleural biopsies, wedge resections, decortications, and lobectomies , VAT remains the superior alternative due to its higher overall diagnostic accuracy and less postoperative pain . When comparing diagnostic efficacy and safety profiles, AVAT pleural biopsy has demonstrated equal or improved outcomes in patients with undiagnosed pleural effusions . However, VAT is still preferred because it provides a more controlled environment and has a lower risk of complications . It is also associated with a shorter hospital stay and quicker recovery times . There is a gap in the current understanding and management of pleural carcinoids especially in regions with limited access to advanced diagnostic tools such as magnetic resonance imaging (MRI). Therefore, this case report aims to address the gap by highlighting the effectiveness of VAT in accurately diagnosing and managing pleural carcinoids thereby improving patient outcomes and informing clinical practices in similar contexts.
A 77-year-old African male with a history of hypertension and heart failure with reduced ejection (HFrEF) presented with a 6-month history of an irritating nonproductive cough, occasional productive episodes, and exertional dyspnea. Despite multiple previous admissions and interventions such as tube drainage and anti-TB medications, the symptoms persisted. Antituberculosis (anti-TB) medications were started on the presumption that the pleural effusion could have been caused by TB; however, they were stopped in 2 weeks due to no reported improvement in symptoms. The patient’s daily home medications included; amlodipine and furosemide with a surgical history of appendectomy. He had no known allergies and was a retired teacher with long-term exposure to chalk dust and 6 years living under an asbestos roof. He was a non-smoker who consumed alcohol occasionally. The patient had no family history of malignancy and he was human immunodeficiency (HIV) seronegative. On physical examination, his vital signs showed a blood pressure of 149/80 mmHg, a temperature of 36.3%, SpO 2 89–91% on room air and a respiratory rate of 38 bpm. Cardiovascular examination was normal while respiratory examination revealed stony dull percussion and reduced air entry on the left side. Initial investigations included an echocardiogram showing an ejection fraction of 36% with grade III diastolic dysfunction and thoracostomy drainage that yielded 2.2L of blood-stained fluid (Fig. ). Pleural fluid analysis indicated a transudate by Light’s criteria and was negative for TB. A chest X-ray showed a minimal rib crowding of the left hemithorax and a homogeneous opacity (Fig. ). A computed tomography (CT) scan revealed mild pleural thickening especially at the bases (Fig. ). Video assisted thoracoscopy (VAT) was performed revealing thickened pleural and a biopsy confirmed a malignant carcinoid tumor (Fig. , Fig. ). The patient underwent pleurodesis with 60 units of bleomycin resulting in a significant reduction of pleural fluid output to less than 50 ml per day allowing for the discontinuation of underwater drainage. Post pleurodesis, the patient showed significant clinical improvement with no recurrence of pleural effusion and marked improvement in dyspnea. The final diagnosis was a malignant carcinoid tumor of the pleura.
Pleural effusions are a frequent cause of hospital admissions at Mulago hospital in Uganda. On a global scale, tuberculosis remains one of the most frequent causes of pleural effusions . Despite this prevalence, the diagnostic challenge arises when standard tests such as contrast-enhanced CT of the thorax, pleural fluid cytology, cultures, acid fast bacilli stain, and gene X-pert for TB return with negative results. In this case, the absence of a definitive diagnosis from these initial tests prompted the use of video assisted thoracoscopy (VAT) with subsequent harvesting of lung and pleural tissue for histology. Pleural metastases are generally associated with metastatic adenocarcinoma and are frequently associated with tumors of the lung ; however, in this case, there was no evidence of malignancy from other sites based on physical examinations and comprehensive investigations done. Given the patient’s history of asbestos exposure, differential diagnoses included mesothelioma and pleural squamous cell carcinoma, both of which are known to cause malignant pleural effusions and are associated with previous asbestos exposure (, ). In cases where initial histopathologic results are inconclusive, immunohistochemistry plays a crucial role in confirming the diagnosis. Carcinoid tumors being neuroendocrine in nature often exhibit specific immunohistochemical traits including the presence of neuropeptides like chromogranin A and synaptophysin, dense core membrane granules on electron microscopy and neuron-specific enolase . Additional markers such as CD56 and thyroid transcription factor 1 can also aid in the diagnosis of neuroendocrine tumors although TTF1 has been found to be sensitive but not specific for pulmonary carcinoids . The utilization of VAT in this case was pivotal, with a diagnostic accuracy of approximately 95% for detecting malignancies and tuberculosis of the pleura . It offers a minimally invasive approach to obtain biopsies reducing the need for more invasive procedures like open thoracotomy. This case illustrates the utility of VAT in diagnosing and managing pleural effusions when initial noninvasive tests are inconclusive.
This case highlights the pivotal role of video assisted thoracoscopy (VAT) in diagnosing and managing rare cases such as pulmonary carcinoids. Our patient with a history of hypertension and heart failure was ultimately diagnosed with a malignant carcinoid tumor of the pleura via VAT following initial inconclusive investigations. This highlights VAT’s effectiveness as both a diagnostic and therapeutic tool, particularly in complex cases.
|
Molecular pathological classification of colorectal cancer | e0ee41e8-3aff-44b6-bf09-978438c0b7ea | 4978761 | Pathology[mh] | Colorectal cancer (CRC) is the third most common cancer in men and the second most common cancer in women, accounting for about 700,000 deaths per year . The majority of 70–80 % of CRC are sporadic, while around 20–30 % of CRC have a hereditary component, due to either uncommon or rare, high-risk, susceptibility syndromes, such as Lynch Syndrome (LS) (3–4 %) and familial adenomatous polyposis (FAP) (∼1 %) , or more common but low-risk alleles. Some of the latter, such as Shroom2 , have been identified by genome-wide association studies (GWAS) . A small subset of about 1–2 % of CRC cases arises as a consequence of inflammatory bowel diseases . CRC is not a homogenous disease, but can be classified into different subtypes, which are characterised by specific molecular and morphological alterations. A major feature of CRC is genetic instability that can arise by at least two different mechanisms. The most common (around ∼84 % of sporadic CRC) is characterised by chromosomal instability (CIN), with gross changes in chromosome number and structure including deletions, gains, translocations and other chromosomal rearrangements. These are often detectable as a high frequency of DNA somatic copy number alterations (SCNA), which are a hallmark of most tumours that arise by the adenoma-carcinoma sequence . Previous molecular genetic studies have associated CIN with inactivating mutations or losses in the Adenomatous Polyposis Coli ( APC) tumour suppressor gene, which occur as an early event in the development of neoplasia of the colorectum in this sequence. The second group (around ∼13–16 % of sporadic CRC) are hypermutated and show microsatellite instability (MSI) due to defective DNA mismatch repair (MMR), often associated with wild-type TP53 and a near-diploid pattern of chromosomal instability (Fig. ) . Furthermore, CpG island methylation phenotype (CIMP) is a feature that induces epigenetic instability by promotor hypermethylation and silencing of a range of tumour suppressor genes, including MLH1 , one of the MMR genes . This review provides an overview of the integrated molecular and transcriptomic patterns in CRC, including new insights from The Cancer Genome Atlas (TCGA) project and the Consensus Molecular Subtype (CMS) Consortium .
CIN tumours usually arise as a consequence of a combination of oncogene activation (e.g. KRAS , PIK3CA ) and tumour suppressor gene inactivation (e.g. APC , SMAD4 and TP53 ) by allelic loss and mutation, which go along with changes in tumour characteristics in the adenoma to carcinoma sequence, as first described by Fearon and Vogelstein in 1990 . A key early event in this pathway is hyperactivation of the WNT signalling pathway, usually arising from mutations of the APC gene. Abnormalities of the WNT pathway characterise the majority of sporadic colorectal cancers, as well as tumours that arise in FAP patients . Over 80 % of adenomas and CRC exhibit APC mutations and a further 5–10 % are showing mutations or epigenetic changes in other WNT signalling components (e.g. β-catenin) that equally result in hyperactivation of the WNT pathway . APC is an important negative regulator of the WNT pathway, being a component of the Axin-APC degradosome complex that promotes the proteasomal degradation of the WNT effector β-catenin. If this complex is defective as a consequence of mutational inactivation of APC, excess β-catenin accumulates within the cytoplasm and translocates into the nucleus where it operates a transcriptional switch leading to activation of MYC and many other genes . Perturbation of the WNT pathway leads to a dysregulation of proliferation and differentiation with the development of dysplastic crypts, which progress to adenomas with increasing grade of dysplasia owing to loss of other tumour suppressor genes. The transition from adenoma to invasive carcinoma is usually associated with mutation and/or loss of the TP53 tumour suppressor gene.
Lynch syndrome (LS), also previously known as hereditary non-polyposis colorectal cancer syndrome (HNPCC), is a syndrome of inherited susceptibility to cancers of several organs, primarily the large bowel, with the next most frequently affected being the endometrium. Moreover, there is also an increased risk of adenocarcinomas of the ovary, stomach, small intestine, transitional cell tumours of ureter and renal pelvis, skin neoplasms (sebaceous tumours and keratoacanthomas), and brain gliomas, amongst others. Development of a neoplasm involves inheriting and acquiring defects in the DNA MMR system in the neoplastic cells. The syndrome is caused by dominant inheritance of a mutant MMR gene (mostly either MSH2 or MLH1 ), with all somatic cells containing one mutated and one wild-type MMR allele. During tumour formation, there is inactivation of the second MMR allele, by mutation, deletion or promoter methylation (in the case of the MLH1 gene), such that the neoplastic cell has inactivated both MMR alleles. In contrast, in sporadic colorectal cancers with defective mismatch repair, the mechanism is almost always (>95 %) promoter hypermethylation of both alleles of the MLH1 gene, thus silencing MLH1 expression and crippling the MMR pathway . The selective pressure for defective mismatch repair within a neoplasm appears to be due to the reduced susceptibility to apoptosis induced by mismatch-related DNA damage . LS colorectal cancers are adenocarcinomas in type, often poorly differentiated or sometimes undifferentiated, occasionally with a dyscohesive appearance. They have prominent tumour-infiltrating lymphocytes and peritumoural Crohns-like lymphoid cell aggregates (Fig. ) and arise more often in the proximal than in the distal bowel. The major affected genes in LS are MSH2 and MLH1 , accounting for 40–45 % LS families each, with the others being mostly due to MSH6 and PMS2 mutations (∼5–10 % LS families each), with rare LS families having other affected genes . The resulting failure to repair DNA replication-associated mismatch errors in these tumour cells produces a high frequency of mutations, either as single base mismatches or in regions of short tandem DNA repeats (the repeat units often being 1–4 bp in length), known as microsatellites. Thus, DNA extracted from such LS tumours shows variation in length (longer and shorter) of a significant proportion of microsatellites, often more than 30 % of those microsatellite markers tested, a phenomenon known as microsatellite instability at high frequency (MSI-H). Following DNA damage or most commonly following DNA replication-associated mismatch errors, MMR proteins normally recognise both base mismatches and the insertion/deletion loops (IDLs) that occur in repetitive sequences. Recognition of mismatches and single base IDLs involves the heterodimeric complexes of MutS-related proteins MSH2 and MSH6 (known as hMutS-Alpha), whereas IDLs of 2–8 nucleotides are recognised by the complex of MSH2 and MSH3 (known as hMutS-Beta). There is overlap in the specificities of these two complexes and hence some redundancy in their activity. A second type of heterodimeric complex, involving two MutL-related proteins, such as either MLH1 and PMS2 (hMutL-Alpha), or MLH1 and PMS1 (hMutL-Beta), binds to the hMutS complex along with other protein components, so that excision of the recently synthesised error-containing DNA strand occurs and resynthesis of the correct sequence of nucleotides can take place, thus repairing the error . Loss or abnormal expression of the MMR proteins MLH1, MSH2, MSH6 and PMS2, assessed by immunohistochemistry, is standard practice in many pathology laboratories and is used to help identify LS cancers along with MSI typing of tumour DNA (Fig. ). Distinguishing LS colorectal cancers that show loss of MLH1 expression from sporadic MMR-deficient cancers is currently most appropriately performed by detection of the specific mutation BRAF V600E, which is found in around 80–90 % of sporadic MSI-H colorectal cancers, but rarely—if ever—in colorectal cancers due to Lynch syndrome . The presence of MLH1 promoter hypermethylation may be used to distinguish sporadic CRC from Lynch syndrome-associated CRC, but there are interpretative problems as constitutive MLH1 promoter methylation may occur, as well as technical challenges of performing this test . In addition to MLH1 , there are a number of other genes displaying DNA promoter hypermethylation changes, sometimes referred to as CIMP-genes, but there is some disagreement regarding which are the most reliable CIMP-genes and which tests to use for identification of CIMP tumours .
In addition to CRC development via the well-described adenoma-carcinoma sequence, it is estimated that about 10–20 % of carcinomas may develop via a different sequence of morphological changes, known as the serrated pathway. While the majority of serrated polyps (80–90 %) can be characterised as hyperplastic polyps, which are considered benign bystander lesions, a subset of serrated lesions can progress to colorectal carcinoma. The two premalignant precursor lesions are traditional serrated adenomas (TSA) and sessile serrated adenomas/polyps (SSA/P) (termed sessile serrated adenomas or alternatively sessile serrated polyps, previous European recommendations have also suggested the term sessile serrated lesions) (Fig. ). Cancers arising via the two serrated pathways are heterogeneous in terms of molecular patterns and cannot easily be classified based on characteristic mutations, but rather by specific morphologic changes. A common feature of the serrated pathways is mutations in KRAS or BRAF , leading to hyperactivation of the MAPKinase pathway. Furthermore EphB2 can be downregulated by genomic loss or promoter methylation, also resulting in MAPK hyperactivation . The characteristic morphological features of the traditional serrated pathway such as architectural dysplasia with ectopic budding crypt formation and epithelial serrations are likely to be linked with these molecular alterations that result in hyperproliferation and inhibition of apoptosis . TSAs are more often diagnosed in the left colon. They frequently (∼80 %) have KRAS mutations or less often (20–30 %) BRAF mutations and are microsatellite stable (MSS) or MSI-L. They are diagnosed based on characteristic cytology (eosinophilic cytoplasm, central, elongated hyperchromatic nuclei) and slit-like epithelial serrations with ectopic crypt formation and may progress to adenocarcinoma (traditional serrated pathway) . SSA/P frequently occur in the right colon, and they tend to have BRAF mutations (∼80 %). CIMP is an early feature of SSA/P and often leads to MSI, related to MLH1 promoter hypermethylation. Also, MTMG can be silenced by promoter methylation, which on its own results in an MSI-L phenotype. SSA/P are characterised by abnormally shaped (boot, inverted-anchor, J, L or inverted T) crypts or horizontal growth along the muscularis mucosae, with crypt dilatation and serration extending down to the crypt base . These architectural changes (without genuine dysplasia) are the hallmark of SSA/P and are believed to result from a displacement of the maturation zone . SSA/P may progress to serrated or mucinous adenocarcinomas (sessile serrated pathway). Colorectal cancers arising via the serrated pathways have been recognised as a distinct subtype overlapping with CIN and MSI tumours by molecular profiling, and are strongly associated with poor prognosis and therapy resistance. Since EMT and matrix remodelling proteins are upregulated in these lesions, it was hypothesised that this predisposes CRC developing via the serrated pathways to invasiveness and metastasis at an early stage . Subsequent analysis revealed that MSI, which often develops within SSA/P, resulted in a more favourable prognosis, whereas MSS in carcinomas derived from SSA/P, and more often from TSA, was linked to poor prognosis .
The TCGA network project collected colorectal tumour samples and corresponding germline DNA samples from 276 patients for exome sequencing of 224 cancers with paired normal samples, along with DNA SCNA analysis, promoter methylation, messenger RNA (mRNA) and micro RNA (miRNA) studies. Ninety-seven samples underwent whole genome sequencing. The clinical and pathological characteristics reflected the typical cross-section of patients with CRC, so this data provides a valuable source of information to gain further insights into the molecular pathology of CRC . The analysis revealed that the bowel cancers could be split into two major groups by mutation rate—non-hypermutated and hypermutated cancers—which by characteristics and frequency match well with the previously discussed CIN and MSI pathways (Fig. , Table ). The hypermutated category was further subdivided in two subgroups. While the majority of tumours in this group (∼13 % of the analysed tumours) were hypermutated cancers due to defective mismatch repair (dMMR) with a high mutation rate of 12–40 mutations/Mb, a small subgroup (∼3 % of the analysed tumours) had an extremely high mutation rate of >40 mutations/Mb and were thus called ultramutated cancers. The dMMR of the hypermutated cancers resulted from acquired hypermethylation of the MLH1 promoter in almost all cases, leading to the silencing of expression of MLH1 and non-functioning mismatch repair, which is again in accordance with the previously discussed findings. Almost all of these tumours showed CIMP characteristics, with several other specifically tested genes also demonstrating promoter methylation. A small number of cancers showed either inherited (LS/HNPCC) or somatic MMR gene mutations. The ultramutated colorectal carcinomas had an extremely high mutation rate with a characteristic nucleotide base change spectrum with increased C-to-A transversions, resulting from the presence of a mutation that inactivates the proofreading function within the exonuclease domain of the polymerase E (POLE) DNA replicating enzyme, or rarely of POLD1. This resulted in failure to correct the misincorporation of nucleotides during DNA replication or repair by mutant POLE (or D1). Other studies have shown that less than 0.1 % of CRC have inherited mutations at characteristic sites within the exonuclease domain of either POLE (p.Leu424Val) or POLD1 (p.Ser478Asn), which are the basis of the polymerase-proofreading-associated polyposis (PPAP) syndrome that is characterised by increased colorectal adenomas and adenocarcinomas as well as increased risk of endometrial cancer in the case of inherited POLD1 mutations . The group of non-hypermutated cancers with a low mutation rate (<8 mutations/Mb) mostly demonstrated a high SCNA frequency, making up the majority (∼84 %) of colorectal adenocarcinomas that were MSS due to an intact MMR pathway. Ultramutated and hypermutated cancers were combined into a single group and compared with the low mutation rate MSS tumours. Overall, 32 genes were recurrently mutated and after removal of non-expressed genes, there were 15 and 17 recurrently mutated genes in the hypermutated and non-hypermutated bowel cancer groups, respectively. The significantly mutated genes in the hypermutated cancers included ACVR2A (63 %), APC (51 %), TGFBR2 (51 %), BRAF (46 %), MSH3 (40 %), MSH6 (40 %), MYOB1 (31 %), TCF7L2 (31 %), CASP8 (29 %), CDC27 (29 %), FZD3 (29 %), MIER3 (29 %), TCERG1 (29 %), MAP7 (26 %), PTPN12 (26 %) and TP53 (20 %). The genes that were recurrently mutated in the non-hypermutated MSS colorectal cancers included mutations in APC (81 %), TP53 (60 %), KRAS (43 %), TTN (31 %), PIK3CA (18 %), FBXW7 (11 %), SMAD4 (10 %), NRAS (9 %), TCF7L2 (9 %), FAM123B , also known as WTX , (7 %), SMAD2 (6 %), CTNNB1 (5 %), KIAA1804 (4 %), SOX9 (4 %), ACVR1B (4 %), GPC6 (40 %) and EDNRB (3 %). The tumour suppressor genes ATM and ARID1A showed a disproportionately high percentage of nonsense or frameshift mutations. The KRAS and NRAS mutations were activating oncogenic mutations at codons 12, 13 and 61, and the BRAF mutation was the classical V600E activating mutation, whereas the other genes almost entirely had inactivating mutations. Colonic and rectal cancers were combined for the analysis of the non-hypermutated MSS group, as they showed no distinguishable molecular differences. SCNA patterns in non-hypermutated MSS tumours confirmed the previously well-documented chromosomal arm-level changes of significant gains of 1q, 7p, 7q, 8p, 8q, 12q, 13q, 19q and 20p, and significant deletions of 1p, 4q, 5q, 8p, 14q, 15q, 17p (includes TP53 ) and 17q, 18q (includes SMAD4 ), 20p and 22q. Hypermutated MSI cancers had far fewer SCNAs, but a similar pattern of chromosomal arm gains and losses. There were 28 recurrent deletion peaks that included the genes FHIT , RBFOX1 , WWOX , SMAD4 , APC , PTEN , SMAD3 and TCF7L2 . Other studies have identified PARK2 as another recurrently deleted gene on chromosome 6 in around a third of CRCs . A chromosomal translocation generating a gene fusion of TCF7L2 and VT11A was seen in 3 % of CRC and also NAV2 - TCF7L1 fusion in three cancers. Focal amplifications were seen affecting MYC , ERBB2 , IGF2 , USP12 , CDK8 , KLF5 , HNF4A , WHSC1L1 / FGFR1 and gains of IRS2 . The most frequently altered pathways by gene mutations, deletions, amplifications and translocations were activation of the WNT, MAPK and PI3K signalling pathways, and deactivation of the TGF-β and P53 inhibitory pathways, which may be relevant for targeted therapies. The WNT signalling pathway was activated in 93 % of non-hypermutated and 97 % of hypermutated cancers, involving biallelic inactivation of APC or activation of CTNNB1 in over 80 % of tumours, together with changes to many other genes involved in regulation of the WNT pathway ( TCF7L2 , DKK , AXIN2 , FBXW7 , ARID1A , FAM123B , FZD10 and SOX9 ). Alterations affecting either the MAPK ( ERBB2 , RAS genes, BRAF ) or PI3K ( PIK3CA , PIK3R1 , PTEN , IGF2 , IRS2 ) signalling pathways were relatively common, often showing patterns of mutual exclusivity of gene mutations (for RAS and BRAF or for PIK3CA , PIK3R1 and PTEN ). The TGF-β pathway was deregulated by alterations to TGFBR1 , TGFBR2 , ACVR2A , ACVR1B , SMAD2 , SMAD3 and SMAD4 in 27 % of non-hypermutated MSS tumours and 87 % of hypermutated cancers. The P53 pathway was affected by mutations to TP53 (60 %) and ATM (7 %) in a near mutually exclusive pattern in non-hypermutated MSS bowel cancers. An integrated data analysis showed that nearly all tumours displayed dysregulation of MYC transcriptional targets as a result of MYC activation by activated WNT signalling and/or dysregulation of TGF-β signalling, indicating an important role for MYC in colorectal cancer. Using CRC resection data on stage, nodal status, distant metastasis and vascular invasion, some molecular changes were associated with aggressive features including those affecting SCN5A , APC , TP53 , PIK3CA , BRAF and FBXW7 as well as altered expression of some miRNAs. Potential therapeutic approaches suggested by the TCGA classification are targeting of IGF2, IGFR, ERBB2, ERBB3, MEK, AKT and mTOR proteins as well as possible WNT pathway inhibitors.
Early attempts at gene expression profiling in order to stratify CRC were made by several groups, but showed little agreement with each other, suggesting different categories, and did not lead to a useful single consistent classification system . Subsequently, an international expert consortium recently reached an agreement that describes four consensus molecular subtypes (CMS) after analysis of 18 different CRC gene expression datasets, including data from TCGA in conjunction with molecular data on mutations and SCNAs for a subset of the samples (Fig. ). CMS1 (MSI-immune, 14 %) CRC were hypermutated due to defective DNA mismatch repair with MSI and MLH1 silencing and accordingly CIMP-high with frequent BRAF mutations, while having a low number of SCNAs. This equates with the previously well-characterised sporadic MSI CRC subgroup. Gene expression profiling furthermore revealed evidence of strong immune activation (immune response, PD1 activation, NK cell, Th1 cell and cytotoxic T cell infiltration signatures) in CMS1, consistent with pathological descriptions of prominent tumour-infiltrating CD8+ cytotoxic T lymphocytes. Patients with the CMS1 subtype had a very poor survival rate after relapse. The majority of CRC previously described as CIN was split into three subcategories based on transcriptomic profiling, which consequently were all characterised by high levels of SCNAs. CMS2 (canonical, 37 %) CRC predominantly displayed epithelial signatures with prominent WNT and MYC signalling activation, and more often displayed loss of tumour suppressor genes and copy number gains of oncogenes than the other subtypes. CMS2 patients had a better survival rate after relapse compared with the other subtypes. The CMS3 (metabolic, 13 %) subtype had fewer SCNAs and contained more hypermutated/MSI samples than CMS2 and CMS4, along with frequent KRAS mutations and a slightly higher prevalence of CIMP-low. Gene expression analysis of CMS3 found predominantly epithelial signatures and evidence of metabolic dysregulation in a variety of pathways. The CMS4 subtype (mesenchymal, 23 %) CRC showed increased expression of EMT genes and evidence of prominent transforming growth factor-β activation, with expression of genes implicated in complement-associated inflammation, matrix remodelling, stromal invasion and angiogenesis. Patients with the CMS4 subtype had a worse overall survival and worse relapse-free survival than patients of the other groups. Finally, there were some samples with mixed features (13 %) that possibly represent either a transition phenotype or intratumoural heterogeneity. This CMS classification system has been suggested by the authors to be the most robust classification system currently available for CRC based on biological processes related to gene expression patterns and is suggested as a basis for future clinical stratification in trials and other studies with potential for subtype-based targeted interventions, although further studies are required to validate this assertion.
In conclusion, integration of wide-ranging molecular data has generated two systems of classification of colorectal cancers (Fig. , Table ). (A) TCGA classification—tumours with a very high mutation rate which can be further subdivided into either (1a) ultramutated colorectal cancers (∼3 %) with DNA polymerase epsilon (POLE) proofreading domain mutations, or (1b) hypermutated colorectal cancers (∼13 %) with microsatellite instability due to defective mismatch repair; and (2) colorectal cancers (∼84 %) with a low mutation rate but a high frequency of DNA SCNAs. (B) The CMS classification describes four CMS groups—CMS1 (MSI-immune activation, 14 %), CMS2 (canonical, 37 %), CMS3 (metabolic, 13 %) and CMS4 (mesenchymal, 23 %), with a residual unclassified group (mixed features, 13 %). Further research is required to develop more easily applicable molecular tests, such as low-coverage high-throughput sequencing for DNA SCNA analysis and/or cancer gene panel mutation detection, and preferably easily applicable and useful immunohistochemical markers for these CMS subdivisions. Analysis of expression of the MMR proteins and/or MSI testing is currently efficient at identifying the group of defective mismatch repair MSI tumours (CMS1). Both classification systems agree on identification of this dMMR/MSI group, which has recently been shown to respond well to immune checkpoint blockade (antibodies to PD-1) that activates cytotoxic T cell attacks on tumour cells, which is suggested to be related to the large numbers of neo-antigens generated by dMMR . A straightforward and routinely applicable molecular test using PCR and sequencing for identification of POLE (and POLD1) proofreading mutations associated with ultramutated cancer may be performed in molecular pathology laboratories, although in the future a mutation-specific POLE antibody for immunohistochemistry may be developed to aid routine subclassification. Ultramutated cancers are likely to generate higher levels of neo-antigens and may also respond well to immune checkpoint blockade therapy. Selected transcript expression profiling kits for CMS classification may be required for application of this system. Both classification systems have been proposed to allow better prognostication and are potentially important for future use in clinical trials and for multidisciplinary team discussions about post-surgical adjuvant treatment, including immune checkpoint blockade.
|
Feasibility and Preliminary Efficacy of Co-Designed and Co-Created Healthy Lifestyle Social Media Intervention Programme the Daily Health Coach for Young Women: A Pilot Randomised Controlled Trial | f83900eb-cea7-4513-a42f-5470baef6271 | 11680048 | Health Literacy[mh] | Digital marketing significantly influences young peoples’ food choices, often promoting unhealthy and highly processed options . Young adults, particularly in New Zealand, are therefore over-exposed to nutrient-poor, energy-dense products, with healthier options like fruits and vegetables becoming less accessible due to cost or the appeal of processed alternatives . This research targets New Zealand women aged 18–24, a group experiencing significant life transitions such as moving out, pursuing further education, starting careers, or beginning families . These changes often negatively impact health behaviours, leading to increased alcohol consumption, poor eating habits, and reduced physical activity . Globally, young adults consume the most energy-dense foods while being the poorest vegetable consumers, a pattern that often worsens in mid-adulthood . Furthermore, compared to digital health research with children and adolescents, research targeting young adults is scarce . This scarcity is compounded when focussing on young women and nutrition specifically. Greater intervention efforts are required to determine the utility of social media (SM) for supporting this population group. In New Zealand, SM platforms are commonly adopted across demographics. A 2021 survey exploring the use of SM by New Zealand adults found that females were more likely to use the social platforms Facebook and Instagram, whilst males were more inclined to use Twitter and LinkedIn . As expected, younger adults were larger consumers of SM than their older counterparts, with Instagram, Snapchat, and TikTok being the most popular platforms amongst those aged 18–29 years . On a global stage, young New Zealanders spend more time on the internet than teenagers in all other countries excluding Denmark, Chile, and Sweden, with a screentime of 42 h per week, compared to the global average of 35 . Although the OECD ‘PISA’ report (Programme for International Student Assessment) is not specific to SM, its findings highlight the potential of the digital environment as a promising youth health promotion space. Aggressive digital marketing and SM now significantly influence the food environment . Companies leverage platforms like Facebook, Instagram, and YouTube to target consumers using detailed demographic and psychological data, increasing susceptibility to unhealthy food advertising . This targeted approach is particularly effective on young people, making them more vulnerable to discretionary food marketing. A recent study found that food brands commonly use interactive techniques including user-generated content, games, and apps to engage consumers, particularly young people, with unhealthy food promotions . Compounding this, often untrained SM influencers can significantly impact young people’s nutrition beliefs and self-perceptions . Concerning young women, repeated exposure to beauty standards and nutrition misinformation online can lead to body dissatisfaction, poor food relationships, and disordered eating behaviours . On the positive side, digital platforms have facilitated cost-effective research and health interventions. Co-designed digital health interventions (DHIs) have shown potential for positive behaviour change and improved health outcomes, often surpassing traditional methods in reach and engagement . A recent example is the ‘HEYMAN’ healthy lifestyle programme, a semi-digital intervention developed collaboratively with young men. The programme included a SM component, utilising Facebook as a platform for participant support as well as a website and wearable device . ’HEYMAN’ was a feasible way of facilitating positive behaviour change, reporting significant intervention effects for weight, BMI, and energy-dense, nutrient-poor foods . Like so many DHIs, the programme incorporated shared protective factors such as social support networks facilitated by SM . This is aligned with the values of young people, who appear more receptive to holistic approaches, with strong emphases on mental and social health as well as physical states . The health promotion landscape would likely benefit from an increased number of DHIs with a multi-factorial approach to behaviour change. Co-design appears to be a favourable methodology in the development of DHIs, given its ability to enhance acceptability and effectiveness . Including stakeholders with lived experience in all aspects of intervention development is an effective way of increasing intervention accessibility and use. This is particularly true of SM interventions, as platform algorithms, and therefore use, are subject to constant change. Co-design is recommended as best practice, not only to improve value and viability; participation in co-design phases has been shown to improve self-efficacy amongst designers, as well as those engaging with final outputs. Whilst recent research is indicative of SM offering feasible avenues for youth health promotion, the ability of platforms to directly influence health behaviours, particularly in a nutrition context, remains uncertain, and requires further research through randomised controlled trials (RCTs) . A systematic review of the recruitment and retention of young adults in 2021 highlighted poor reporting of feasibility measures and provided relevant benchmarks for assessing these contributors to intervention success . Livingstone et al. have also published benchmarks for retention, recruitment, and participation via an RCT protocol to determine the feasibility of Veg4Me, a co-designed app for young adults to improve vegetable consumption . Taken together, these insights will form evidence-based criteria for ascertaining the viability of this pilot intervention. This research aims to explore how best to disseminate nutrition information via influencer techniques, using implementation science as a framework . The objectives are as follows: Evaluate the feasibility of recruitment, randomisation, data collection methods, and retention of a youth co-designed health promotion programme for young women aged 18–24 years against evidence-based criteria. Estimate the treatment effects of the Daily Health Coach (DHC) on improving diet quality, physical activity, and other lifestyle, psychological, and social influence measures.
2.1. Study Design The DHC study is an assessor-blinded, two-arm pilot randomised controlled trial (RCT) addressing the feasibility and preliminary efficacy of a 3-month programme delivered via Instagram. Following demographic data collection, young women were individually randomised to the Daily Health Coach intervention group (commenced DHC intervention immediately) or the waitlist control group (started DHC after a 3-month delay). The study design and conduct adhered to the guidelines as outlined by Thabane et al. . The checklist is an adapted version of the Consolidated Standards of Reporting Trials (CONSORT) guidelines specifically for pilot studies . The reporting of results follows that of Ashton et al., who conducted a similar trial with young men . Ethics This study was approved by the Human Participants Ethics Committee at the University of Auckland on 12 July 2023 for three years. Reference number: UAHPEC26195. The trial was retrospectively registered with the American Economic Association’s Registry for randomised controlled trials, number AEARCTR-0014872. 2.2. Intervention Development A comprehensive overview of the co-design phase and programme development is available for further reading . In brief, the DHC is a SM health promotion programme designed for young women aged 18–24 to improve eating habits, activity levels, and overall wellbeing. Developed using the Young & Well Co-operative Research Centre framework, which was adapted for general wellbeing and nutrition behaviours, the DHC incorporated the target audience into the design process to enhance effectiveness and address cohort-specific needs . Formative research identified motivators, barriers, and preferences for content delivery via SM, involving young women ( n = 19) across Aotearoa, New Zealand in co-design workshops . These workshops explored the participants’ conceptualisation of health, examined influencer content on Instagram and TikTok, and used the micro–meso–macro model to discuss common barriers . Insights were used to create the DHC Instagram profile and a 12-week content planner, with contributions from student dietitians and a recent nutrition graduate . Informed by co-design insights and best practice guidelines for diet and physical activity, the DHC programme also incorporated the COM-B model, Self-Determination Theory (SDT), and evidence from successful health-related interventions for this demographic . The result is a 12-week co-developed healthy lifestyles intervention programme delivered via Instagram. It involves the daily dissemination of co-created nutrition and health content via Instagram posts, reels, and stories. For users who elect to be messaged directly, the programme also involves check-ins and encouraging messages via the direct message function at their selected frequency over the 12-week intervention period. 2.3. Participants and Recruitment 2.3.1. Sample Size A key objective of pilot studies is to gain initial estimates for a sample size calculation in a future, adequately powered RCT, and thus a formal sample size calculation was not performed. A systematic review of pilot and feasibility studies identified a median total sample size of 30 in non-drug trials . Therefore, we aimed to exceed this, and a recruitment target of 50 was set. 2.3.2. Recruitment Strategy Participants were recruited using targeted advertisements on SM pages (Facebook, Instagram, and LinkedIn), posted by the research team to personal and professional networks, as well as by the Faculty of Medical and Health Sciences at the University of Auckland. The research invitation ad directed prospective participants to an online survey to screen eligibility criteria. Eligible participants were emailed addressing their expression of interest with the participant information statement (PIS), which provided more in-depth information about the study and participant requirements . Following the provision of the PIS, potential participants had the opportunity to ask questions (either via email/phone call or face-to-face meeting) before being sent a consent form to sign and return digitally. The initial screening survey was administered at the recruitment stage via e-link to all potential participants and included questions to determine the self-reported adequacy of fruit and vegetable intake, exercise, and medical history, including existing diagnosis of an eating disorder (e.g., anorexia nervosa or bulimia). SM literacy was determined via a set of questions from the Social Media Literary Assessment Questionnaire, which included questions regarding specific SM-related knowledge, skills, and behaviours (ie., how familiar are you with Instagram as a SM platform?, how much do you know about the types of content that are commonly shared on SM (e.g., posts, reels, stories, polls, lives, etc.)?, and, how confident are you in your ability to use the different features and tools on SM (e.g., posting content, commenting, sharing)?). A total of eight SM familiarity questions were included in the screening survey. 2.4. Data Collection—Primary Outcomes 2.4.1. Programme Component Use and Acceptability The use of programme components has been objectively tracked, using SM metrics including (1) applause rate, the number of approval actions (e.g., likes, comments) a digital content receives; (2) live assessments of favourability and preference via Instagram poll or story question during the intervention; and (3) average engagement rate. Engagement with the intervention was measured via collection of data on post likes and comments. As the intervention was closed (i.e., run on a private Instagram account), post “views” were unable to be assessed, and participants were not able to share content seen with other individuals. A post-programme process evaluation survey developed by the research team and informed by previous studies was administered to assess DHC intervention components . Participants were asked to rank individual programme components on a 5-point Likert scale from strongly agree (=5) to strongly disagree (=1), for attractiveness (“visually appealing”), comprehension (“provided me with useful information”), usability (“easy to use/receive”), length (“I am satisfied with the length of the programme”), ability to persuade/engage (“helped me attain my goals”), ability to provide support (“was supportive in answering my queries/questions”), and overall satisfaction with the DHC. 2.4.2. Feasibility of Recruitment The eligibility screening survey distributed to all prospective participants allowed for the measurement of recruitment feasibility, via the number of young women interested versus those eligible. Benchmark criteria for recruitment are based on the findings of Whatnall et al., who performed a systematic review of the literature on the recruitment and retention of young adults in behavioural interventions targeting nutrition, physical activity, and/or obesity . The feasibility benchmark for recruitment is >40% of young women screened for eligibility subsequently enrolled in the study. 2.4.3. Feasibility of Retention Retention has been defined as attendance at the 3-month follow-up measurements, where participants remain following the DHC on Instagram. A retention rate target of >80% for the intervention group at 3-month follow-up assessment has been established according to findings from two relevant systematic reviews and a scoping literature review of digital health interventions for young adults undertaken by the research team 2.4.4. Acceptability of Randomisation Randomisation feasibility has been assessed by asking participants to rank overall satisfaction with the group allocation on a 5-point Likert scale from very satisfied (=5) to very unsatisfied (=1). 2.4.5. Acceptability of Data Collection Acceptability of data collection has been estimated from the percentage of young women who completed all objective and self-report questionnaires at baseline, mid-point, and after the intervention. Those who ceased completion will be included when calculating rates of participation. A participation rate of >70% has been set as an evidence-based benchmark for evaluating the feasibility of data collection methods . 2.5. Data Collection—Secondary Outcomes 2.5.1. Preliminary Efficacy All questionnaires used for the collection of data are validated study instruments. Changes to diet quality such as intake of fruits, vegetables, energy-dense take-away meals, SSB, water, and physical activity were measured via the Short Form Food Frequency (SF-FFQ) questionnaire and the International Physical Activity questionnaire (IPAQ). As well as exercise and nutrition habits, participants were instructed to complete surveys that evaluated social influence, body image disturbance, food relationships, and digital health literacy. In total, six validated surveys were distributed during each round to evaluate healthy behaviour change: A short-form dietary questionnaire (SF-FFQ) ; Physical activity questionnaire (IPAQ) ; Social influence questionnaire (SIQ) ; Three-factor eating questionnaire-Revised 18-item (TFEQ-R18) ; Body image disturbance questionnaire ; eHEALS digital healthy literacy questionnaire . 2.5.2. Survey Distribution Each survey round involved the completion of the surveys ( n = 6). Participants in the intervention group were invited to complete the following three rounds of surveys: one week prior to the 12-week intervention (preliminary data collection), mid intervention at 6 weeks, and one week post intervention period. Young women in the waitlist control group completed surveys at the same time as the intervention group when awaiting the programme, as well as two additional survey rounds when receiving the intervention . A total of 18 surveys required completion by the intervention group, and 30 for the waitlist control group. 2.6. Randomisation Participants were randomised using computer generation via the Excel randomisation function by the student researcher, who utilised participant UPIs for input. The ratio of assignment to groups was 50:50. Half of the participants were randomised to the intervention group ( n = 25), and half were allocated to the waitlist control group ( n = 25). The research team had no say over which group participants were allocated. This study was an open-label study; participants were made aware of which group they were assigned to upon allocation. 2.7. Statistical Analysis The data have been analysed using IBM SPSS Statistical software . The differences between groups at baseline and the characteristics of completers vs. dropouts were tested using independent t -tests for continuous variables and chi-squared (χ2) tests for categorical variables. The significance level for the comparison of baseline characteristics was set at 0.05. Programme acceptability measures have been presented as mean ±SD, with higher scores (maximum of 5) indicating greater acceptability/satisfaction. All secondary health outcomes were included in linear mixed model analyses; the predictors (fixed effects) included time (treated as categorical with levels baseline, mid-point and 3 months), treatment group (intervention and control), and an interaction term for time by treatment group. Covariate type AR(1) was selected as suitable for the longitudinal data. The p value associated with the interaction term was used to determine the statistical significance of any difference between treatment groups. All participants who completed at least one survey round whilst receiving the DHC intervention were included in linear mixed model analysis to assess change over time. For the estimation of treatment effects, differences in mean scores from baseline to 3 months were tested for intention-to-treat (ITT) populations. Differences of means and 95% confidence intervals were calculated using the Cohen’s d equation for mean change from baseline within and between groups . Only participants who completed baseline and end-point surveys at 3 months were included in this analysis.
The DHC study is an assessor-blinded, two-arm pilot randomised controlled trial (RCT) addressing the feasibility and preliminary efficacy of a 3-month programme delivered via Instagram. Following demographic data collection, young women were individually randomised to the Daily Health Coach intervention group (commenced DHC intervention immediately) or the waitlist control group (started DHC after a 3-month delay). The study design and conduct adhered to the guidelines as outlined by Thabane et al. . The checklist is an adapted version of the Consolidated Standards of Reporting Trials (CONSORT) guidelines specifically for pilot studies . The reporting of results follows that of Ashton et al., who conducted a similar trial with young men . Ethics This study was approved by the Human Participants Ethics Committee at the University of Auckland on 12 July 2023 for three years. Reference number: UAHPEC26195. The trial was retrospectively registered with the American Economic Association’s Registry for randomised controlled trials, number AEARCTR-0014872.
This study was approved by the Human Participants Ethics Committee at the University of Auckland on 12 July 2023 for three years. Reference number: UAHPEC26195. The trial was retrospectively registered with the American Economic Association’s Registry for randomised controlled trials, number AEARCTR-0014872.
A comprehensive overview of the co-design phase and programme development is available for further reading . In brief, the DHC is a SM health promotion programme designed for young women aged 18–24 to improve eating habits, activity levels, and overall wellbeing. Developed using the Young & Well Co-operative Research Centre framework, which was adapted for general wellbeing and nutrition behaviours, the DHC incorporated the target audience into the design process to enhance effectiveness and address cohort-specific needs . Formative research identified motivators, barriers, and preferences for content delivery via SM, involving young women ( n = 19) across Aotearoa, New Zealand in co-design workshops . These workshops explored the participants’ conceptualisation of health, examined influencer content on Instagram and TikTok, and used the micro–meso–macro model to discuss common barriers . Insights were used to create the DHC Instagram profile and a 12-week content planner, with contributions from student dietitians and a recent nutrition graduate . Informed by co-design insights and best practice guidelines for diet and physical activity, the DHC programme also incorporated the COM-B model, Self-Determination Theory (SDT), and evidence from successful health-related interventions for this demographic . The result is a 12-week co-developed healthy lifestyles intervention programme delivered via Instagram. It involves the daily dissemination of co-created nutrition and health content via Instagram posts, reels, and stories. For users who elect to be messaged directly, the programme also involves check-ins and encouraging messages via the direct message function at their selected frequency over the 12-week intervention period.
2.3.1. Sample Size A key objective of pilot studies is to gain initial estimates for a sample size calculation in a future, adequately powered RCT, and thus a formal sample size calculation was not performed. A systematic review of pilot and feasibility studies identified a median total sample size of 30 in non-drug trials . Therefore, we aimed to exceed this, and a recruitment target of 50 was set. 2.3.2. Recruitment Strategy Participants were recruited using targeted advertisements on SM pages (Facebook, Instagram, and LinkedIn), posted by the research team to personal and professional networks, as well as by the Faculty of Medical and Health Sciences at the University of Auckland. The research invitation ad directed prospective participants to an online survey to screen eligibility criteria. Eligible participants were emailed addressing their expression of interest with the participant information statement (PIS), which provided more in-depth information about the study and participant requirements . Following the provision of the PIS, potential participants had the opportunity to ask questions (either via email/phone call or face-to-face meeting) before being sent a consent form to sign and return digitally. The initial screening survey was administered at the recruitment stage via e-link to all potential participants and included questions to determine the self-reported adequacy of fruit and vegetable intake, exercise, and medical history, including existing diagnosis of an eating disorder (e.g., anorexia nervosa or bulimia). SM literacy was determined via a set of questions from the Social Media Literary Assessment Questionnaire, which included questions regarding specific SM-related knowledge, skills, and behaviours (ie., how familiar are you with Instagram as a SM platform?, how much do you know about the types of content that are commonly shared on SM (e.g., posts, reels, stories, polls, lives, etc.)?, and, how confident are you in your ability to use the different features and tools on SM (e.g., posting content, commenting, sharing)?). A total of eight SM familiarity questions were included in the screening survey.
A key objective of pilot studies is to gain initial estimates for a sample size calculation in a future, adequately powered RCT, and thus a formal sample size calculation was not performed. A systematic review of pilot and feasibility studies identified a median total sample size of 30 in non-drug trials . Therefore, we aimed to exceed this, and a recruitment target of 50 was set.
Participants were recruited using targeted advertisements on SM pages (Facebook, Instagram, and LinkedIn), posted by the research team to personal and professional networks, as well as by the Faculty of Medical and Health Sciences at the University of Auckland. The research invitation ad directed prospective participants to an online survey to screen eligibility criteria. Eligible participants were emailed addressing their expression of interest with the participant information statement (PIS), which provided more in-depth information about the study and participant requirements . Following the provision of the PIS, potential participants had the opportunity to ask questions (either via email/phone call or face-to-face meeting) before being sent a consent form to sign and return digitally. The initial screening survey was administered at the recruitment stage via e-link to all potential participants and included questions to determine the self-reported adequacy of fruit and vegetable intake, exercise, and medical history, including existing diagnosis of an eating disorder (e.g., anorexia nervosa or bulimia). SM literacy was determined via a set of questions from the Social Media Literary Assessment Questionnaire, which included questions regarding specific SM-related knowledge, skills, and behaviours (ie., how familiar are you with Instagram as a SM platform?, how much do you know about the types of content that are commonly shared on SM (e.g., posts, reels, stories, polls, lives, etc.)?, and, how confident are you in your ability to use the different features and tools on SM (e.g., posting content, commenting, sharing)?). A total of eight SM familiarity questions were included in the screening survey.
2.4.1. Programme Component Use and Acceptability The use of programme components has been objectively tracked, using SM metrics including (1) applause rate, the number of approval actions (e.g., likes, comments) a digital content receives; (2) live assessments of favourability and preference via Instagram poll or story question during the intervention; and (3) average engagement rate. Engagement with the intervention was measured via collection of data on post likes and comments. As the intervention was closed (i.e., run on a private Instagram account), post “views” were unable to be assessed, and participants were not able to share content seen with other individuals. A post-programme process evaluation survey developed by the research team and informed by previous studies was administered to assess DHC intervention components . Participants were asked to rank individual programme components on a 5-point Likert scale from strongly agree (=5) to strongly disagree (=1), for attractiveness (“visually appealing”), comprehension (“provided me with useful information”), usability (“easy to use/receive”), length (“I am satisfied with the length of the programme”), ability to persuade/engage (“helped me attain my goals”), ability to provide support (“was supportive in answering my queries/questions”), and overall satisfaction with the DHC. 2.4.2. Feasibility of Recruitment The eligibility screening survey distributed to all prospective participants allowed for the measurement of recruitment feasibility, via the number of young women interested versus those eligible. Benchmark criteria for recruitment are based on the findings of Whatnall et al., who performed a systematic review of the literature on the recruitment and retention of young adults in behavioural interventions targeting nutrition, physical activity, and/or obesity . The feasibility benchmark for recruitment is >40% of young women screened for eligibility subsequently enrolled in the study. 2.4.3. Feasibility of Retention Retention has been defined as attendance at the 3-month follow-up measurements, where participants remain following the DHC on Instagram. A retention rate target of >80% for the intervention group at 3-month follow-up assessment has been established according to findings from two relevant systematic reviews and a scoping literature review of digital health interventions for young adults undertaken by the research team 2.4.4. Acceptability of Randomisation Randomisation feasibility has been assessed by asking participants to rank overall satisfaction with the group allocation on a 5-point Likert scale from very satisfied (=5) to very unsatisfied (=1). 2.4.5. Acceptability of Data Collection Acceptability of data collection has been estimated from the percentage of young women who completed all objective and self-report questionnaires at baseline, mid-point, and after the intervention. Those who ceased completion will be included when calculating rates of participation. A participation rate of >70% has been set as an evidence-based benchmark for evaluating the feasibility of data collection methods .
The use of programme components has been objectively tracked, using SM metrics including (1) applause rate, the number of approval actions (e.g., likes, comments) a digital content receives; (2) live assessments of favourability and preference via Instagram poll or story question during the intervention; and (3) average engagement rate. Engagement with the intervention was measured via collection of data on post likes and comments. As the intervention was closed (i.e., run on a private Instagram account), post “views” were unable to be assessed, and participants were not able to share content seen with other individuals. A post-programme process evaluation survey developed by the research team and informed by previous studies was administered to assess DHC intervention components . Participants were asked to rank individual programme components on a 5-point Likert scale from strongly agree (=5) to strongly disagree (=1), for attractiveness (“visually appealing”), comprehension (“provided me with useful information”), usability (“easy to use/receive”), length (“I am satisfied with the length of the programme”), ability to persuade/engage (“helped me attain my goals”), ability to provide support (“was supportive in answering my queries/questions”), and overall satisfaction with the DHC.
The eligibility screening survey distributed to all prospective participants allowed for the measurement of recruitment feasibility, via the number of young women interested versus those eligible. Benchmark criteria for recruitment are based on the findings of Whatnall et al., who performed a systematic review of the literature on the recruitment and retention of young adults in behavioural interventions targeting nutrition, physical activity, and/or obesity . The feasibility benchmark for recruitment is >40% of young women screened for eligibility subsequently enrolled in the study.
Retention has been defined as attendance at the 3-month follow-up measurements, where participants remain following the DHC on Instagram. A retention rate target of >80% for the intervention group at 3-month follow-up assessment has been established according to findings from two relevant systematic reviews and a scoping literature review of digital health interventions for young adults undertaken by the research team
Randomisation feasibility has been assessed by asking participants to rank overall satisfaction with the group allocation on a 5-point Likert scale from very satisfied (=5) to very unsatisfied (=1).
Acceptability of data collection has been estimated from the percentage of young women who completed all objective and self-report questionnaires at baseline, mid-point, and after the intervention. Those who ceased completion will be included when calculating rates of participation. A participation rate of >70% has been set as an evidence-based benchmark for evaluating the feasibility of data collection methods .
2.5.1. Preliminary Efficacy All questionnaires used for the collection of data are validated study instruments. Changes to diet quality such as intake of fruits, vegetables, energy-dense take-away meals, SSB, water, and physical activity were measured via the Short Form Food Frequency (SF-FFQ) questionnaire and the International Physical Activity questionnaire (IPAQ). As well as exercise and nutrition habits, participants were instructed to complete surveys that evaluated social influence, body image disturbance, food relationships, and digital health literacy. In total, six validated surveys were distributed during each round to evaluate healthy behaviour change: A short-form dietary questionnaire (SF-FFQ) ; Physical activity questionnaire (IPAQ) ; Social influence questionnaire (SIQ) ; Three-factor eating questionnaire-Revised 18-item (TFEQ-R18) ; Body image disturbance questionnaire ; eHEALS digital healthy literacy questionnaire . 2.5.2. Survey Distribution Each survey round involved the completion of the surveys ( n = 6). Participants in the intervention group were invited to complete the following three rounds of surveys: one week prior to the 12-week intervention (preliminary data collection), mid intervention at 6 weeks, and one week post intervention period. Young women in the waitlist control group completed surveys at the same time as the intervention group when awaiting the programme, as well as two additional survey rounds when receiving the intervention . A total of 18 surveys required completion by the intervention group, and 30 for the waitlist control group.
All questionnaires used for the collection of data are validated study instruments. Changes to diet quality such as intake of fruits, vegetables, energy-dense take-away meals, SSB, water, and physical activity were measured via the Short Form Food Frequency (SF-FFQ) questionnaire and the International Physical Activity questionnaire (IPAQ). As well as exercise and nutrition habits, participants were instructed to complete surveys that evaluated social influence, body image disturbance, food relationships, and digital health literacy. In total, six validated surveys were distributed during each round to evaluate healthy behaviour change: A short-form dietary questionnaire (SF-FFQ) ; Physical activity questionnaire (IPAQ) ; Social influence questionnaire (SIQ) ; Three-factor eating questionnaire-Revised 18-item (TFEQ-R18) ; Body image disturbance questionnaire ; eHEALS digital healthy literacy questionnaire .
Each survey round involved the completion of the surveys ( n = 6). Participants in the intervention group were invited to complete the following three rounds of surveys: one week prior to the 12-week intervention (preliminary data collection), mid intervention at 6 weeks, and one week post intervention period. Young women in the waitlist control group completed surveys at the same time as the intervention group when awaiting the programme, as well as two additional survey rounds when receiving the intervention . A total of 18 surveys required completion by the intervention group, and 30 for the waitlist control group.
Participants were randomised using computer generation via the Excel randomisation function by the student researcher, who utilised participant UPIs for input. The ratio of assignment to groups was 50:50. Half of the participants were randomised to the intervention group ( n = 25), and half were allocated to the waitlist control group ( n = 25). The research team had no say over which group participants were allocated. This study was an open-label study; participants were made aware of which group they were assigned to upon allocation.
The data have been analysed using IBM SPSS Statistical software . The differences between groups at baseline and the characteristics of completers vs. dropouts were tested using independent t -tests for continuous variables and chi-squared (χ2) tests for categorical variables. The significance level for the comparison of baseline characteristics was set at 0.05. Programme acceptability measures have been presented as mean ±SD, with higher scores (maximum of 5) indicating greater acceptability/satisfaction. All secondary health outcomes were included in linear mixed model analyses; the predictors (fixed effects) included time (treated as categorical with levels baseline, mid-point and 3 months), treatment group (intervention and control), and an interaction term for time by treatment group. Covariate type AR(1) was selected as suitable for the longitudinal data. The p value associated with the interaction term was used to determine the statistical significance of any difference between treatment groups. All participants who completed at least one survey round whilst receiving the DHC intervention were included in linear mixed model analysis to assess change over time. For the estimation of treatment effects, differences in mean scores from baseline to 3 months were tested for intention-to-treat (ITT) populations. Differences of means and 95% confidence intervals were calculated using the Cohen’s d equation for mean change from baseline within and between groups . Only participants who completed baseline and end-point surveys at 3 months were included in this analysis.
3.1. Participant Flow at Each Stage Almost all participants remained in the DHC intervention for the 12-week research period . One participant withdrew from the intervention for personal reasons within the first month. Two participants from the waitlist control group never followed the DHC on Instagram for the duration of the programme. One participant in the same group unfollowed the DHC without providing a reason. Therefore, despite the recruitment of 50 participants, 46 were retained throughout the 3-month intervention period. 3.2. Baseline Data There were no significant differences between intervention and control groups for any of the demographic factors assessed at baseline between groups, or between completers vs. dropouts . All recruited participants ( n = 50) were included in baseline analyses. 3.3. Primary Results 3.3.1. Feasibility of Research Procedures; Programme Component Acceptability and Use The feasibility of programme components is presented as mean scores for post-programme evaluation survey (PPE) responses from a 5-point Likert scale. The PPE was deemed as reliable following calculation of the a-Chronbach value, which was 0.7__. Overall, the DHC scored 83.6% (4.18/5) for programme satisfaction . Participants found the DHC to be useful, aesthetically pleasing, accessible, and appreciated intervention components. Improvements to support individual queries and goals should be a key focus for future iterations. The total rate of engagement with the DHC programme during intervention periods was 10.04%. The average engagement rate for influencers across Instagram varies by source from 1% to 3% . Micro-influencers, such as the DHC, tend to see higher rates of engagement . For smaller influencers, an engagement rate above 5% is said to be indicative of strong engagement . The main form of engagement was via post ‘likes’. Only three users left comments on posts across both cohorts. The mean like count was 4.62, ranging from 0 to 13. Posts were more ‘liked’ than reels. The top-rated or ‘liked’ post topics for both groups were nutrition misinformation, a nutritious carbohydrate explanation, a recipe for a glazed salmon bowl, a ketogenic diet ‘myth-bust’, and nutritious pantry staple ideas. 3.3.2. Feasibility of Recruitment Recruitment feasibility was calculated as the number of valid applicants that completed the online screening survey and met participation criteria. The rate of qualified applicants for the DHC trial was 74.5% (78/106). This rate is deemed feasible against the pre-determined evidence-based benchmark criteria of 55% eligibility . In total, 106 valid applicants expressed interest in the DHC trial. Reasons for exclusion included meeting the national fruit, vegetable, and physical activity guidelines ( n = 6), being outside of the age range ( n = 5), having an active eating disorder ( n = 2), non-NZ residency ( n = 1), concurrent participation in other healthy lifestyle programmes ( n = 1), and incomplete eligibility screening ( n = 11). 3.3.3. Feasibility of Retention The rate of retention was calculated as participants who followed the DHC on Instagram for the duration of the programme ( n = 46). Participants were not retained if they (a) never followed the DHC on Instagram ( n = 2), (b) unfollowed the DHC on Instagram ( n = 1), or (c) actively withdrew ( n = 1). The benchmark for successful retention was set as >66%, in accordance with similar app-based interventions . The total rate of retention for the DHC feasibility trial was 92% (46/50). 3.3.4. Acceptability of Randomisation Randomisation satisfaction is presented as means from a 5-point Likert scale. Overall satisfaction with allocated research groups was 4.3/5 (86%). Of those who completed the randomisation satisfaction survey ( n = 37), participants were more satisfied with being randomised to the waitlist control group (4.3 (86%) vs. 3.65/5 (73%)). This may be due to greater reimbursement resulting from additional survey rounds. 3.3.5. Acceptability of Data Collection The number of young women who completed all objective and self-report questionnaires at baseline, mid-point, and after the intervention determined the acceptability of the data collection methods, defined as the participation rate. This was found to be 80% for the DHC (40/50), exceeding the pre-determined target of >70% . Regarding survey completion, n = 5 participants stopped completing questionnaires, yet remained following the DHC ( n = 3 in the intervention group, n = 2 in the waitlist control group). One further participant missed a mid-intervention survey round. A number of participants completed all validated surveys, yet did not complete either one or both of the randomisation satisfaction surveys or the post-programme process evaluation survey ( n = 13). 3.4. Estimation of Treatment Effects; Efficacy of Nutrition Habits, Digital Health Literacy, Food Relationships/Body Image, and Social Influence No significant differences between groups were observed for any secondary measure when assessing mean change in score from baseline to three months (end of intervention period) . Disordered eating behaviours decreased from baseline, as well as body image disturbance and physical activity. Digital health literacy and diet quality scores increased from baseline for both groups; however, this increase was not significant. Linear mixed model analyses found that statistically significant changes were observed for three out of seven measures: body image disturbance, social influence, and digital health literacy . A significant decrease in body image disturbance was observed over time ( p = 0.01). There was a significant group-by-time interaction effect observed for digital health literacy ( p = 0.002), indicating an increase in the cohort’s ability to source and/or discern evidence-based nutrition information over time ( p = 0.01). Finally, an increase in social influence for the waitlist control group when compared to the intervention group was found, where the waitlist control group observed an increase in score, whilst the intervention group mean score declined ( p = 0.03). No other significant changes were observed for the measured fixed effects across cohorts (between groups, over time, or group-by-time effects).
Almost all participants remained in the DHC intervention for the 12-week research period . One participant withdrew from the intervention for personal reasons within the first month. Two participants from the waitlist control group never followed the DHC on Instagram for the duration of the programme. One participant in the same group unfollowed the DHC without providing a reason. Therefore, despite the recruitment of 50 participants, 46 were retained throughout the 3-month intervention period.
There were no significant differences between intervention and control groups for any of the demographic factors assessed at baseline between groups, or between completers vs. dropouts . All recruited participants ( n = 50) were included in baseline analyses.
3.3.1. Feasibility of Research Procedures; Programme Component Acceptability and Use The feasibility of programme components is presented as mean scores for post-programme evaluation survey (PPE) responses from a 5-point Likert scale. The PPE was deemed as reliable following calculation of the a-Chronbach value, which was 0.7__. Overall, the DHC scored 83.6% (4.18/5) for programme satisfaction . Participants found the DHC to be useful, aesthetically pleasing, accessible, and appreciated intervention components. Improvements to support individual queries and goals should be a key focus for future iterations. The total rate of engagement with the DHC programme during intervention periods was 10.04%. The average engagement rate for influencers across Instagram varies by source from 1% to 3% . Micro-influencers, such as the DHC, tend to see higher rates of engagement . For smaller influencers, an engagement rate above 5% is said to be indicative of strong engagement . The main form of engagement was via post ‘likes’. Only three users left comments on posts across both cohorts. The mean like count was 4.62, ranging from 0 to 13. Posts were more ‘liked’ than reels. The top-rated or ‘liked’ post topics for both groups were nutrition misinformation, a nutritious carbohydrate explanation, a recipe for a glazed salmon bowl, a ketogenic diet ‘myth-bust’, and nutritious pantry staple ideas. 3.3.2. Feasibility of Recruitment Recruitment feasibility was calculated as the number of valid applicants that completed the online screening survey and met participation criteria. The rate of qualified applicants for the DHC trial was 74.5% (78/106). This rate is deemed feasible against the pre-determined evidence-based benchmark criteria of 55% eligibility . In total, 106 valid applicants expressed interest in the DHC trial. Reasons for exclusion included meeting the national fruit, vegetable, and physical activity guidelines ( n = 6), being outside of the age range ( n = 5), having an active eating disorder ( n = 2), non-NZ residency ( n = 1), concurrent participation in other healthy lifestyle programmes ( n = 1), and incomplete eligibility screening ( n = 11). 3.3.3. Feasibility of Retention The rate of retention was calculated as participants who followed the DHC on Instagram for the duration of the programme ( n = 46). Participants were not retained if they (a) never followed the DHC on Instagram ( n = 2), (b) unfollowed the DHC on Instagram ( n = 1), or (c) actively withdrew ( n = 1). The benchmark for successful retention was set as >66%, in accordance with similar app-based interventions . The total rate of retention for the DHC feasibility trial was 92% (46/50). 3.3.4. Acceptability of Randomisation Randomisation satisfaction is presented as means from a 5-point Likert scale. Overall satisfaction with allocated research groups was 4.3/5 (86%). Of those who completed the randomisation satisfaction survey ( n = 37), participants were more satisfied with being randomised to the waitlist control group (4.3 (86%) vs. 3.65/5 (73%)). This may be due to greater reimbursement resulting from additional survey rounds. 3.3.5. Acceptability of Data Collection The number of young women who completed all objective and self-report questionnaires at baseline, mid-point, and after the intervention determined the acceptability of the data collection methods, defined as the participation rate. This was found to be 80% for the DHC (40/50), exceeding the pre-determined target of >70% . Regarding survey completion, n = 5 participants stopped completing questionnaires, yet remained following the DHC ( n = 3 in the intervention group, n = 2 in the waitlist control group). One further participant missed a mid-intervention survey round. A number of participants completed all validated surveys, yet did not complete either one or both of the randomisation satisfaction surveys or the post-programme process evaluation survey ( n = 13).
The feasibility of programme components is presented as mean scores for post-programme evaluation survey (PPE) responses from a 5-point Likert scale. The PPE was deemed as reliable following calculation of the a-Chronbach value, which was 0.7__. Overall, the DHC scored 83.6% (4.18/5) for programme satisfaction . Participants found the DHC to be useful, aesthetically pleasing, accessible, and appreciated intervention components. Improvements to support individual queries and goals should be a key focus for future iterations. The total rate of engagement with the DHC programme during intervention periods was 10.04%. The average engagement rate for influencers across Instagram varies by source from 1% to 3% . Micro-influencers, such as the DHC, tend to see higher rates of engagement . For smaller influencers, an engagement rate above 5% is said to be indicative of strong engagement . The main form of engagement was via post ‘likes’. Only three users left comments on posts across both cohorts. The mean like count was 4.62, ranging from 0 to 13. Posts were more ‘liked’ than reels. The top-rated or ‘liked’ post topics for both groups were nutrition misinformation, a nutritious carbohydrate explanation, a recipe for a glazed salmon bowl, a ketogenic diet ‘myth-bust’, and nutritious pantry staple ideas.
Recruitment feasibility was calculated as the number of valid applicants that completed the online screening survey and met participation criteria. The rate of qualified applicants for the DHC trial was 74.5% (78/106). This rate is deemed feasible against the pre-determined evidence-based benchmark criteria of 55% eligibility . In total, 106 valid applicants expressed interest in the DHC trial. Reasons for exclusion included meeting the national fruit, vegetable, and physical activity guidelines ( n = 6), being outside of the age range ( n = 5), having an active eating disorder ( n = 2), non-NZ residency ( n = 1), concurrent participation in other healthy lifestyle programmes ( n = 1), and incomplete eligibility screening ( n = 11).
The rate of retention was calculated as participants who followed the DHC on Instagram for the duration of the programme ( n = 46). Participants were not retained if they (a) never followed the DHC on Instagram ( n = 2), (b) unfollowed the DHC on Instagram ( n = 1), or (c) actively withdrew ( n = 1). The benchmark for successful retention was set as >66%, in accordance with similar app-based interventions . The total rate of retention for the DHC feasibility trial was 92% (46/50).
Randomisation satisfaction is presented as means from a 5-point Likert scale. Overall satisfaction with allocated research groups was 4.3/5 (86%). Of those who completed the randomisation satisfaction survey ( n = 37), participants were more satisfied with being randomised to the waitlist control group (4.3 (86%) vs. 3.65/5 (73%)). This may be due to greater reimbursement resulting from additional survey rounds.
The number of young women who completed all objective and self-report questionnaires at baseline, mid-point, and after the intervention determined the acceptability of the data collection methods, defined as the participation rate. This was found to be 80% for the DHC (40/50), exceeding the pre-determined target of >70% . Regarding survey completion, n = 5 participants stopped completing questionnaires, yet remained following the DHC ( n = 3 in the intervention group, n = 2 in the waitlist control group). One further participant missed a mid-intervention survey round. A number of participants completed all validated surveys, yet did not complete either one or both of the randomisation satisfaction surveys or the post-programme process evaluation survey ( n = 13).
No significant differences between groups were observed for any secondary measure when assessing mean change in score from baseline to three months (end of intervention period) . Disordered eating behaviours decreased from baseline, as well as body image disturbance and physical activity. Digital health literacy and diet quality scores increased from baseline for both groups; however, this increase was not significant. Linear mixed model analyses found that statistically significant changes were observed for three out of seven measures: body image disturbance, social influence, and digital health literacy . A significant decrease in body image disturbance was observed over time ( p = 0.01). There was a significant group-by-time interaction effect observed for digital health literacy ( p = 0.002), indicating an increase in the cohort’s ability to source and/or discern evidence-based nutrition information over time ( p = 0.01). Finally, an increase in social influence for the waitlist control group when compared to the intervention group was found, where the waitlist control group observed an increase in score, whilst the intervention group mean score declined ( p = 0.03). No other significant changes were observed for the measured fixed effects across cohorts (between groups, over time, or group-by-time effects).
4.1. Feasibility The Daily Health Coach intervention is a feasible form of disseminating nutrition information to young women. Overall, participants found the Instagram programme to be accessible, usable, and visually appealing. Participant satisfaction was the lowest for query support and programme length, suggesting future iterations of the programme should consider a longer intervention period, as well as offering more frequent opportunities for participants to answer questions and gain support (i.e., increase in general query polls and/or direct messages). The feasibility of the DHC aligns with similar findings for youth digital health interventions, where co-designed digital tools are found to be acceptable, usable, and feasible by young people . However, the translation of feasible digital tools or programmes to improvements in health outcomes or behaviours is, as one may anticipate, more difficult to achieve. 4.2. Efficacy The DHC had no significant impact on diet quality, physical activity, or disordered eating behaviours such as uncontrolled and emotional eating. However, the findings suggest that the programme may have a positive impact on the body image and digital health literacy of young women. This aligns with the results of preliminary co-design work when developing the DHC, where young women acknowledged the role of body image and misinformation on nutrition habits and status and advocated for its inclusion in the programme . These aspects of nutrition for young women are frequently missed out of general nutrition dialogue online, and the ‘grittiness’ of information on these pertinent factors may have played a role in the reported results. The Instagram newsfeed is often flooded with information about nutrition, food, and physical activity. However, discussions of body image, nutrition misinformation, and the inflammation of these issues via SM are often absent. A 2016 Australian study investigated knowledge translation to “sticky” SM health messages, meaning content more likely to be recalled by consumers . Potential influences on the “stickiness” of SM posts relevant to the DHC involve unexpected content, social currency, stories and emotion, and posts that were credible and held practical value . Information regarding restrictive diet cycles, the perpetuation of beauty standards, and nutrition confusion are relatively novel, emotive, and relevant, versus general nutrition and physical activity information. Despite the discussed observations, the impact of the DHC on health outcomes remains to be determined in a larger RCT. 4.3. Comparison to Other Work This study is the first of its kind to be conducted in Aotearoa, New Zealand. The evidence base for health promotion via SM continues to grow, yet interventions tend to be classified as health promotion campaigns, rather than programmes . Facebook and Twitter appear to be the most common social networking platforms for disseminating health information . A 2021 systematic review of the effect of SM interventions on physical activity and dietary behaviours in young people found positive effects of the reviewed interventions, demonstrating the promise of SM use in behaviour change; the same was found of a similar review conducted in 2018 . A comparable study to the DHC was conducted in 2017, where a pilot RCT tested the feasibility of a 3-month healthy lifestyles programme, utilising Facebook as an intervention element. The programme, which was tested with 50 participants, was found to be feasible. Despite acceptability findings, efficacy results were mixed. This appears to be common across the DHI literature, with a 2022 study into another social media-based health intervention for young people disseminating similar ‘limited efficacy’ results, as well as a 2019 SM study targeting pregnant adolescents and adult women . Although alike, these DHIs were not targeted specifically to young women, and included distinct evidence-based information concerning mental wellbeing and or physical activity, rather than solely nutrition education and awareness of body image and food relationships. Specific to the intersection of body image and SM, in March 2024 a new project was launched in the UK involving the development of a toolkit to “equip young women with the skills and knowledge needed to cope with potential harmful social media content” . The toolkit is being co-created by researchers at the University of Portsmouth, The Girls’ Network, and the target demographic. Over the 12-week DHC programme, over 70 distinct topics associated with nutrition were shared with participants. The wide-reaching topics of conversation increased the likelihood that information would resonate with participants. The co-design and co-creation of content is also likely to be a contributing factor to the acceptability of the intervention, as collaborative design and user-generated content is known to increase reception and resonation, particularly for young people . 4.4. Limitations SM is a difficult landscape to work with for many reasons. One pertinent issue is the importance of engagement for efficacy. Engagement with intervention postings is essential to continue seeing content; it is necessary to ‘tell’ algorithms that you are ‘interested’ in content via liking, sharing, and/or commenting. If posts are not interacted with in early intervention stages, it is unlikely that they will continue to appear on participants’ newsfeeds. To overcome this, followers of the DHC were instructed to engage with any content seen in the first week of the intervention to increase the likelihood of continual newsfeed presence. The first week of engagement metrics have therefore been removed from analyses. However, there is still a chance that pseudo-engagement has persisted, clouding results. As with all health dissemination research undertaken on SM, it is not possible to state with certainty that the presented results are associated with programme material specifically. There is always a chance of users seeing similar content from other creators when online. This is further confounded by the likelihood of algorithms changing when users begin interacting with programme material, increasing the likelihood of being presented with similar nutrition or health information from alternative sources. SM algorithms change frequently and discreetly; it is therefore important for future iterations of the DHC and other DHIs to engage with social marketing experts in developmental stages to understand and best overcome potential platform barriers. An oversight when converting the TFEQ-R18 to Redcap distribution software resulted in the absence of the final survey question. This meant that the insights for the ‘cognitive restraint’ section of the questionnaire could not be assessed with validity, and results were therefore excluded from the presented analyses. Results for cognitive restraint were non-significant for all fixed effects ( p = 0.375 for group, p = 0.339 for time, and p = 0.908 for group * time). 4.5. Implications The feasibility of the DHC confirms the hypothesis that influencer communication techniques can be used to disseminate evidence-based nutrition information to young people. These findings may be used to advocate for an amplified presence of health professionals on social networks. It is increasingly important for health professionals, particularly dietitians, to advocate for and voice their expertise across social platforms. This includes advice on the sourcing of evidence-based information to combat nutrition misinformation and confusion for young people. Furthermore, when discussing nutrition information online, it is important to reflect upon the impact of nutrition dialogue on body image. For example, those sharing content should consider terminology and avoid moralising language concerning food. General nutrition information that dispels common myths and promotes a non-diet approach can be helpful for vulnerable populations such as young women, for whom SM use often plays a role in poor body esteem or disturbance . Regarding future research directions, the proven feasibility of the DHC programme provides a blueprint for potential youth digital interventions targeting health behaviours. The co-creation protocol, as well as the feasibility findings, may be referenced by researchers or health professionals looking to utilise SM as a platform for influencing positive behaviour change. This is particularly important in the nutrition space, as the majority of DHI research to date has been targeted at mental wellbeing and/or physical activity, rather than diet quality and body image. The feasibility of the DHC is owed to the collaborative design of the programme, whereby young women contributed to the development and creation of intervention content. As such, it is recommended that future DHIs are co-designed with target users to ensure acceptability, relevancy, and use.
The Daily Health Coach intervention is a feasible form of disseminating nutrition information to young women. Overall, participants found the Instagram programme to be accessible, usable, and visually appealing. Participant satisfaction was the lowest for query support and programme length, suggesting future iterations of the programme should consider a longer intervention period, as well as offering more frequent opportunities for participants to answer questions and gain support (i.e., increase in general query polls and/or direct messages). The feasibility of the DHC aligns with similar findings for youth digital health interventions, where co-designed digital tools are found to be acceptable, usable, and feasible by young people . However, the translation of feasible digital tools or programmes to improvements in health outcomes or behaviours is, as one may anticipate, more difficult to achieve.
The DHC had no significant impact on diet quality, physical activity, or disordered eating behaviours such as uncontrolled and emotional eating. However, the findings suggest that the programme may have a positive impact on the body image and digital health literacy of young women. This aligns with the results of preliminary co-design work when developing the DHC, where young women acknowledged the role of body image and misinformation on nutrition habits and status and advocated for its inclusion in the programme . These aspects of nutrition for young women are frequently missed out of general nutrition dialogue online, and the ‘grittiness’ of information on these pertinent factors may have played a role in the reported results. The Instagram newsfeed is often flooded with information about nutrition, food, and physical activity. However, discussions of body image, nutrition misinformation, and the inflammation of these issues via SM are often absent. A 2016 Australian study investigated knowledge translation to “sticky” SM health messages, meaning content more likely to be recalled by consumers . Potential influences on the “stickiness” of SM posts relevant to the DHC involve unexpected content, social currency, stories and emotion, and posts that were credible and held practical value . Information regarding restrictive diet cycles, the perpetuation of beauty standards, and nutrition confusion are relatively novel, emotive, and relevant, versus general nutrition and physical activity information. Despite the discussed observations, the impact of the DHC on health outcomes remains to be determined in a larger RCT.
This study is the first of its kind to be conducted in Aotearoa, New Zealand. The evidence base for health promotion via SM continues to grow, yet interventions tend to be classified as health promotion campaigns, rather than programmes . Facebook and Twitter appear to be the most common social networking platforms for disseminating health information . A 2021 systematic review of the effect of SM interventions on physical activity and dietary behaviours in young people found positive effects of the reviewed interventions, demonstrating the promise of SM use in behaviour change; the same was found of a similar review conducted in 2018 . A comparable study to the DHC was conducted in 2017, where a pilot RCT tested the feasibility of a 3-month healthy lifestyles programme, utilising Facebook as an intervention element. The programme, which was tested with 50 participants, was found to be feasible. Despite acceptability findings, efficacy results were mixed. This appears to be common across the DHI literature, with a 2022 study into another social media-based health intervention for young people disseminating similar ‘limited efficacy’ results, as well as a 2019 SM study targeting pregnant adolescents and adult women . Although alike, these DHIs were not targeted specifically to young women, and included distinct evidence-based information concerning mental wellbeing and or physical activity, rather than solely nutrition education and awareness of body image and food relationships. Specific to the intersection of body image and SM, in March 2024 a new project was launched in the UK involving the development of a toolkit to “equip young women with the skills and knowledge needed to cope with potential harmful social media content” . The toolkit is being co-created by researchers at the University of Portsmouth, The Girls’ Network, and the target demographic. Over the 12-week DHC programme, over 70 distinct topics associated with nutrition were shared with participants. The wide-reaching topics of conversation increased the likelihood that information would resonate with participants. The co-design and co-creation of content is also likely to be a contributing factor to the acceptability of the intervention, as collaborative design and user-generated content is known to increase reception and resonation, particularly for young people .
SM is a difficult landscape to work with for many reasons. One pertinent issue is the importance of engagement for efficacy. Engagement with intervention postings is essential to continue seeing content; it is necessary to ‘tell’ algorithms that you are ‘interested’ in content via liking, sharing, and/or commenting. If posts are not interacted with in early intervention stages, it is unlikely that they will continue to appear on participants’ newsfeeds. To overcome this, followers of the DHC were instructed to engage with any content seen in the first week of the intervention to increase the likelihood of continual newsfeed presence. The first week of engagement metrics have therefore been removed from analyses. However, there is still a chance that pseudo-engagement has persisted, clouding results. As with all health dissemination research undertaken on SM, it is not possible to state with certainty that the presented results are associated with programme material specifically. There is always a chance of users seeing similar content from other creators when online. This is further confounded by the likelihood of algorithms changing when users begin interacting with programme material, increasing the likelihood of being presented with similar nutrition or health information from alternative sources. SM algorithms change frequently and discreetly; it is therefore important for future iterations of the DHC and other DHIs to engage with social marketing experts in developmental stages to understand and best overcome potential platform barriers. An oversight when converting the TFEQ-R18 to Redcap distribution software resulted in the absence of the final survey question. This meant that the insights for the ‘cognitive restraint’ section of the questionnaire could not be assessed with validity, and results were therefore excluded from the presented analyses. Results for cognitive restraint were non-significant for all fixed effects ( p = 0.375 for group, p = 0.339 for time, and p = 0.908 for group * time).
The feasibility of the DHC confirms the hypothesis that influencer communication techniques can be used to disseminate evidence-based nutrition information to young people. These findings may be used to advocate for an amplified presence of health professionals on social networks. It is increasingly important for health professionals, particularly dietitians, to advocate for and voice their expertise across social platforms. This includes advice on the sourcing of evidence-based information to combat nutrition misinformation and confusion for young people. Furthermore, when discussing nutrition information online, it is important to reflect upon the impact of nutrition dialogue on body image. For example, those sharing content should consider terminology and avoid moralising language concerning food. General nutrition information that dispels common myths and promotes a non-diet approach can be helpful for vulnerable populations such as young women, for whom SM use often plays a role in poor body esteem or disturbance . Regarding future research directions, the proven feasibility of the DHC programme provides a blueprint for potential youth digital interventions targeting health behaviours. The co-creation protocol, as well as the feasibility findings, may be referenced by researchers or health professionals looking to utilise SM as a platform for influencing positive behaviour change. This is particularly important in the nutrition space, as the majority of DHI research to date has been targeted at mental wellbeing and/or physical activity, rather than diet quality and body image. The feasibility of the DHC is owed to the collaborative design of the programme, whereby young women contributed to the development and creation of intervention content. As such, it is recommended that future DHIs are co-designed with target users to ensure acceptability, relevancy, and use.
The Daily Health Coach is a feasible health promotion intervention that uses Instagram as platform to reach young women. The current pilot study’s findings indicate that the research procedures, including recruitment, retention, randomisation, and data collection, are sufficiently feasible to warrant a full-scale RCT, with only minor adjustments needed. Acceptability findings are aligned with other digital health interventions created for and by young people. A larger RCT is needed to explore how best to translate feasible SM interventions to positive ‘off-screen’ changes in health behaviours.
|
Educational initiatives and training for paediatric rheumatology in Europe | 57fba23f-c449-466e-95ab-b63bb6db6bcb | 6286498 | Pediatrics[mh] | The Paediatric Rheumatology European Society (PReS) has an overarching aim to promote paediatric rheumatology as a sub-specialty across Europe and also further afield through the expanding international membership. PReS aims to improve the lives of children with rheumatic diseases through collaborative networking to raise awareness, facilitate high quality research to create new knowledge, support education and training, promote advocacy and ‘best care ’ based on evidence, consensus and active consumer engagement. The importance of education and training activities to achieve these aims and address unmet need within Europe has been highlighted following the SHARE initiative (Single Hub Access for Rheumatology care, European Agency Health and Consumers, grant number 2011 1202, [ Wulffraat N, 2013, van Dijkhuizen EHP, 2018]) and reported in Work package 4, SHARE project, Dolezolva P (submitted for publication). The educational activities of PReS are overseen by the PReS Education and Training Committee (ETC), currently led by one of us [TA]. The PReS ETC represents rheumatology within the European Academy of Paediatrics and has contributed to the recent 2016 European Syllabus for Training in Paediatrics ( https://www.pres.eu/committee-and-wp/education.html ). This new syllabus ensures that rheumatology training is compliant with the revised European Syllabus structure and format. The educational activities of PReS reflect the roles of the paediatric rheumatology specialist; namely ‘life-long’ scholarly clinician, researcher and teacher. These roles have foundations firmly embedded in the European Syllabus which sets out the minimum requirements for training to cover knowledge ( musculoskeletal conditions, multi-disciplinary team working, medications, treatment approaches ), skills (clinical, consultation, procedural, teaching, research ) and professional roles ( managerial, leadership and mentorship ). In the context of education and training, the 2016 European syllabus for paediatric rheumatology has the following aims: Harmonise training programmes across European countries. Establish defined standards of knowledge and skills required to practice paediatric rheumatology at the tertiary care level. Foster development of a European network of competent specialist centres to facilitate collaborative research and training opportunities. Improve the clinical care of children with chronic and acute rheumatic disorders within Europe. Enhance European contribution to international scientific progress in the field of paediatric rheumatology. Promote teaching of paediatric rheumatology at undergraduate and postgraduate levels to raise awareness about the importance of early diagnosis, prompt referral to specialist care and facilitate potential recruits to fellowship programmes and the future workforce. This document describes PReS conferences, courses, online resources programmes that are overseen by the PReS ETC (with a summary given in the Table ) and we describe their context for the international paediatric rheumatology community. These activities are complementary to existing national training programmes (which are not discussed here further). The PReS annual scientific meeting and young investigators meeting (YIM) The annual scientific meeting of PReS disseminates advances in knowledge and research ( basic science, clinical care and education ) through ‘state of the art’ lectures and presentations. There are active allied health and consumer parallel programmes. The annual scientific meeting has become one of the most important educational events in the paediatric rheumatology international calendar and provides a platform to disseminate knowledge, opportunities to network and foster new collaborations. The annual Young Investigators Meeting (YIM) is held over the two days preceding the annual scientific meeting and is organized with support from PReS to facilitate attendance by young researchers (PhD students, post-doctoral level) and trainee doctors. The YIM aspires to nurture the PReS research ethos and aims to promote networking and foster opportunities for international collaboration; young investigators are encouraged to present their work to an international audience with feedback and guidance from PReS & YIM senior faculty. Over recent years the number of young investigators attending the YIM meeting has increased significantly with a substantial number of trainees from non-European countries (including India, Africa, North America, South America, Asia, Australia). Proceedings of future meetings and how to apply for YIM bursaries are available on the PRES website ( https://www.pres.eu ). PReS courses The ETC oversees a ‘rolling programme’ of courses which are financially supported by grants from PReS. Details of future courses and how to register for them are available. ( http://www.pres.eu ). The aim is to have one of either the Basic or Advanced course per year and to reach audiences who may not otherwise have access to education and training. PReS members are encouraged to contact the ETC for further information if they wish to consider organizing a Basic or Advanced course in their country and seek PReS support. 1. The PReS Basic courses are set at the level of the general paediatric trainee or resident and cover the breadth of topics as outlined in the European syllabus. To date, PReS Basic Courses have been held in countries where paediatric rheumatology is less well developed Mumbai (India 2012), Sao Paulo (Brazil 2015), Budapest (Hungary 2015) and Cape Town (South Africa 2017) with future courses planned in the Ukraine and South East Asia. The aim is to promote the specialty, facilitate networking and ultimately improve patient care. The model includes an organising committee led by a local paediatric rheumatologist with faculty drawn from local and national colleagues as well as international faculty who are PReS members. Examples of previous Basic Courses are available from the ETC and it is envisaged that lectures from courses will be archived on the PReS website. The Basic Courses have a similar format (over 2–3 days with local and international faculty [ n = 10–15]) and up to 100 delegates, with interactive lectures and workshops. Case presentations are encouraged from delegates. Workshops focused on joint examination have been included since 2015 with practical demonstrations of pGALS (paediatric Gait Arms Legs and Spine) [ Foster and Jandial 2013] involving local patients. The importance of research and how to get involved is also highlighted (e.g. clinical trials, cohort studies and registries). Attendance at the Basic courses has grown year on year - e.g. Cape Town 2017 with 117 delegates, 66 from many countries in Africa and 40 from Europe and the Middle East. Feedback has been very positive to address educational needs as well as networking and peer support opportunities. 2. The PReS Advanced courses (e.g. JIA, Slovenia 2017, Auto-inflammatory diseases Jerusalem, Israel 2018, childhood-onset SLE Lausanne, Switzerland 2018) target a more experienced audience and in the main, senior paediatric rheumatology trainees and paediatricians with an interest in paediatric rheumatology. The childhood-onset SLE Advanced course (2018) was recorded and transmitted by video conferencing to colleagues in India, who were able to participate in case presentations and interactive sessions. It is envisaged that this facility will be available in future courses to reach a wider audience and especially those in low income countries. The scientific programme of an Advanced course is focused on a given condition (e.g. JIA, childhood-onset SLE, auto-inflammatory conditions) or specialist skills (e.g. Musculoskeletal Ultrasound [MSUS], or ‘hands on’ joint examination). The format of the Advanced Courses to date have included interactive ‘state of the art’ lectures delivered by international and local faculty, ‘meet the expert’ small group sessions, joint examination ‘how to teach’ workshops and interactive case presentations. It has been proposed by the ETC that future Advanced courses address recommendations from other PReS / EULAR initiatives such as Transitional care [ Foster et al. 2016] and Musculoskeletal Ultrasound skills (MSUS) [ Iagnocco 2015]. 3. PReS Specialist Skills Courses target a more experienced audience and in the main, senior paediatric rheumatology trainees and paediatricians with an interest in paediatric rheumatology. The model of ‘pairing’ a knowledge-based Advanced Course and a Specialist Skills workshop together is likely to be ‘cost effective’ in terms of travel and time for faculty and delegates. Proposed examples include i) Advanced Course in JIA and MSUS imaging or joint examination skills and an ii) Advanced Course in JSLE or JDM and nail fold capillaroscopy . 3.1 Musculoskeletal Ultrasound (MSUS) imaging. The aim of the ETC is that all fellows in paediatric rheumatology have access to MSUS training facilitated through attending PReS courses. To date there have been MSUS courses organized with EULAR (Madrid [Spain] 2012, 2017, Belgrade [Serbia] 2013, 2015) and further courses are planned (details on PReS website). In addition to the delivery of the MSUS training courses, PReS has worked extensively with OMERACT and EULAR to work towards MSUS as a potential outcome measure in JIA with standardization of procedures, validation of definitions and grading of synovitis using MSUS in JIA. Outputs to date from these collaborations include the following: EULAR/PReS Recommendation for implementation of different imaging modalities in JIA [ Colebatch-Bourn et al. Ann Rheum Dis 2015] EULAR recommendations how to organize education in MSUS [ Iagnocco et al., RMD Open 2015] Standard scanning MSUS positions in four most affected joints in JIA [ Collado 2016] Definition of age related vascularization and ossification of joints in children [ Windschall 2017] Definitions of normal joint structures MSUS findings in children [ Roth et al. Arthritis Care Res 2015]. EULAR/PReS Standardized Procedures for Ultrasound Imaging in Paediatric Rheumatology (ongoing EULAR project CLI089 to produce on line web application as an educational tool) Preliminary Definitions for the Sonographic Features of Synovitis in Children [ Roth 2017] 3.2 PReS clinical skills workshops - Joint examination skills ( pGALS Foster and Jandial, pREMS Foster 2011) are included in the PReS Basic Course (with a focus on pGALS) and Advanced JIA Course (with a focus on pREMS and ‘how to teach’ pGALS). Feedback to date has been very positive. The format includes ‘hands on’ workshops involving consented patients demonstrating physical signs and small groups of learners with facilitated observation and feedback from experienced paediatric rheumatologists. The ETC aims for the ‘clinical skills’ workshop concept to be expanded and to include, for example, scleroderma skin scores, nail fold capillaroscopy, muscle power testing, and transitional care consultation skills. 3.3 PReS ‘Teach the teachers’ workshops - Paediatric Rheumatologists have an integral role as educators and need to ‘reach out’ to medical students, general paediatricians and colleagues in orthopaedics and primary care to raise awareness, facilitate diagnosis and prompt referral to specialist care. Learning ‘how to teach’ is therefore an important skill and included in the European syllabus. The aim of the ETC is that all fellows will attend a ‘how to teach’ course during their training. PReS has included ‘teach the teacher’ workshops alongside the PReS scientific meetings (2016, 2017) with excellent feedback and it is envisaged that such workshops will be included in future PReS meetings and the Advanced Course rolling programme. PReS online resources PReS Website ( http://www.pres.eu ) (revised and launched 2017) has signposting to available resources, conferences, courses and available bursaries. Pediatric Rheumatology Online Journal – ( https://ped-rheum.biomedcentral.com ) founded in 2003, is a free open access journal (current impact factor of 2.328 September 2018), and supported through sponsorship from PReS. The journal has a broad range of content including clinical reviews, clinical research and basic science. The journal also publishes abstracts and proceedings from the annual scientific meetings. The EULAR / PReS paediatric rheumatology online course (launched 2014) was developed with a grant from EULAR and enrolment is subsidized by EULAR with further discounted costs for low and middle income countries ( https://www.eular.org/edu_online_course_paediatric.cfm ). The online course is a collaborative effort between PReS and EULAR, with content written and updated yearly by senior paediatric rheumatologists with the assistance of trainees. The content is aimed at fellows in paediatric rheumatology at the start of their training, residents in adult rheumatology and paediatricians or adult rheumatologists with an interest in paediatric rheumatology. The course consists of 10 modules, each dedicated to a specific topic with a knowledge test at the end, and covering key topics as outlined in the 2016 European Syllabus in paediatric rheumatology. The course can be completed with the final online examination once a year and provides a EULAR/PReS certification for specific knowledge of paediatric rheumatology. The course has attracted more than 650 participants to date; the aim being to promote the course as the leading online education in specialist paediatric rheumatology worldwide. A EULAR/PReS published textbook to supplement the course is also now available. The Paediatric Musculoskeletal Matters [PMM] website ( www.pmmonline.org ) can be accessed from the PReS website ( http://www.pres.eu/activities/scientific-and-clinical/educational-instruments-and-tools.html ). PMM was launched in 2014, is a free and open e-resource and targets medical students and family medicine doctors [ Smith et al. 2016] although content is relevant to all clinicians who encounter children in their clinical practice. PMM has to date > 250,000 hits from > 150 countries. PMM International (launched September 2018) reflects paediatric rheumatology health care around the world and has contributions from an international panel of paediatric rheumatologists. PMM-Nursing is also available (launched 2017) and targets all levels of nurses involved in the care of children with rheumatic diseases. PMM is endorsed by PReS as a resource for paediatric rheumatologists and teams to use in their teaching and encompasses basic clinical skills, descriptions of normal development, the approach to investigations and initial management of musculoskeletal presentations including red flags for serious life threatening conditions. PMM is promoted as a foundation for the Basic Course and to those embarking on the EULAR/PReS online course. PMM signposts to the free pGALS app (with the 2018 version including multiple language translations) and pGALS e-module. Patient and parent information - Further to the SHARE initiative there has been a complete revision of the freely available information leaflets covering a broad spectrum of conditions and including translations into several languages ( https://www.printo.it/pediatric-rheumatology/ ). The importance of patient and family advocacy and engagement in clinical care, service delivery and research has been highlighted [ Dijkhuizen et al] and includes recommendations to optimise participation and facilitate better clinical outcomes. PReS EMERGE fellowship programme PReS EMERGE (EMErging RheumatoloGists and rEsearchers) encompasses young paediatric rheumatologists and researchers working together to improve clinical and research opportunities for trainees, participating in organisation of PReS educational events (YIM, Basic and Advanced courses) and liaising with other young investigator groups (e.g. EMEUNET, CARRA early investigators). The group was set up following the 2016 YIM and currently includes trainees mainly from Europe but also from all other continents. Further details are available through an active social media network ( www.facebook.com/PReSEMERGE , twitter.com/PReSEMERGE ) and through a bimonthly newsletter highlighting activities and opportunities for members (contact [email protected]). The PReS EMERGE fellowship programme ( https://www.pres.eu/activities/young-investigators/fellowship-programs.html ) was launched in 2017 with financial and practical assistance for clinical trainees who are members of PReS and younger than 40 years, to facilitate placements of up to 6-months within a European Paediatric Rheumatology Centre. In addition to gaining clinical knowledge and skills, the trainee is expected to participate in a research project. From 2018, the programme is open to basic science trainees (pre-PhD to five years post-PhD) working in paediatric rheumatology. It is envisaged that the programme will enhance both clinical and basic collaborative research, foster a network of emerging and established paediatric rheumatologists and allow sharing of ideas and practices between different countries to harmonize paediatric rheumatology training. Other PReS educational activities The Standards of care for Juvenile Arthritis Management in Less Resourced countries (JAMLess) is a collaborative effort with funding and support from International League Against Rheumatism (ILAR) and PReS to develop consensus based guidelines relevant to challenging health care contexts; these highlight the importance of education and training being integral to the delivery of clinical care. This project was initiated by PReS colleagues from South Africa, Argentina and India and has resulted in a consensus document [ Scott C 2018], with further work to develop iterations of these recommendations for other parts of the world being underway.
The annual scientific meeting of PReS disseminates advances in knowledge and research ( basic science, clinical care and education ) through ‘state of the art’ lectures and presentations. There are active allied health and consumer parallel programmes. The annual scientific meeting has become one of the most important educational events in the paediatric rheumatology international calendar and provides a platform to disseminate knowledge, opportunities to network and foster new collaborations. The annual Young Investigators Meeting (YIM) is held over the two days preceding the annual scientific meeting and is organized with support from PReS to facilitate attendance by young researchers (PhD students, post-doctoral level) and trainee doctors. The YIM aspires to nurture the PReS research ethos and aims to promote networking and foster opportunities for international collaboration; young investigators are encouraged to present their work to an international audience with feedback and guidance from PReS & YIM senior faculty. Over recent years the number of young investigators attending the YIM meeting has increased significantly with a substantial number of trainees from non-European countries (including India, Africa, North America, South America, Asia, Australia). Proceedings of future meetings and how to apply for YIM bursaries are available on the PRES website ( https://www.pres.eu ).
The ETC oversees a ‘rolling programme’ of courses which are financially supported by grants from PReS. Details of future courses and how to register for them are available. ( http://www.pres.eu ). The aim is to have one of either the Basic or Advanced course per year and to reach audiences who may not otherwise have access to education and training. PReS members are encouraged to contact the ETC for further information if they wish to consider organizing a Basic or Advanced course in their country and seek PReS support. 1. The PReS Basic courses are set at the level of the general paediatric trainee or resident and cover the breadth of topics as outlined in the European syllabus. To date, PReS Basic Courses have been held in countries where paediatric rheumatology is less well developed Mumbai (India 2012), Sao Paulo (Brazil 2015), Budapest (Hungary 2015) and Cape Town (South Africa 2017) with future courses planned in the Ukraine and South East Asia. The aim is to promote the specialty, facilitate networking and ultimately improve patient care. The model includes an organising committee led by a local paediatric rheumatologist with faculty drawn from local and national colleagues as well as international faculty who are PReS members. Examples of previous Basic Courses are available from the ETC and it is envisaged that lectures from courses will be archived on the PReS website. The Basic Courses have a similar format (over 2–3 days with local and international faculty [ n = 10–15]) and up to 100 delegates, with interactive lectures and workshops. Case presentations are encouraged from delegates. Workshops focused on joint examination have been included since 2015 with practical demonstrations of pGALS (paediatric Gait Arms Legs and Spine) [ Foster and Jandial 2013] involving local patients. The importance of research and how to get involved is also highlighted (e.g. clinical trials, cohort studies and registries). Attendance at the Basic courses has grown year on year - e.g. Cape Town 2017 with 117 delegates, 66 from many countries in Africa and 40 from Europe and the Middle East. Feedback has been very positive to address educational needs as well as networking and peer support opportunities. 2. The PReS Advanced courses (e.g. JIA, Slovenia 2017, Auto-inflammatory diseases Jerusalem, Israel 2018, childhood-onset SLE Lausanne, Switzerland 2018) target a more experienced audience and in the main, senior paediatric rheumatology trainees and paediatricians with an interest in paediatric rheumatology. The childhood-onset SLE Advanced course (2018) was recorded and transmitted by video conferencing to colleagues in India, who were able to participate in case presentations and interactive sessions. It is envisaged that this facility will be available in future courses to reach a wider audience and especially those in low income countries. The scientific programme of an Advanced course is focused on a given condition (e.g. JIA, childhood-onset SLE, auto-inflammatory conditions) or specialist skills (e.g. Musculoskeletal Ultrasound [MSUS], or ‘hands on’ joint examination). The format of the Advanced Courses to date have included interactive ‘state of the art’ lectures delivered by international and local faculty, ‘meet the expert’ small group sessions, joint examination ‘how to teach’ workshops and interactive case presentations. It has been proposed by the ETC that future Advanced courses address recommendations from other PReS / EULAR initiatives such as Transitional care [ Foster et al. 2016] and Musculoskeletal Ultrasound skills (MSUS) [ Iagnocco 2015]. 3. PReS Specialist Skills Courses target a more experienced audience and in the main, senior paediatric rheumatology trainees and paediatricians with an interest in paediatric rheumatology. The model of ‘pairing’ a knowledge-based Advanced Course and a Specialist Skills workshop together is likely to be ‘cost effective’ in terms of travel and time for faculty and delegates. Proposed examples include i) Advanced Course in JIA and MSUS imaging or joint examination skills and an ii) Advanced Course in JSLE or JDM and nail fold capillaroscopy . 3.1 Musculoskeletal Ultrasound (MSUS) imaging. The aim of the ETC is that all fellows in paediatric rheumatology have access to MSUS training facilitated through attending PReS courses. To date there have been MSUS courses organized with EULAR (Madrid [Spain] 2012, 2017, Belgrade [Serbia] 2013, 2015) and further courses are planned (details on PReS website). In addition to the delivery of the MSUS training courses, PReS has worked extensively with OMERACT and EULAR to work towards MSUS as a potential outcome measure in JIA with standardization of procedures, validation of definitions and grading of synovitis using MSUS in JIA. Outputs to date from these collaborations include the following: EULAR/PReS Recommendation for implementation of different imaging modalities in JIA [ Colebatch-Bourn et al. Ann Rheum Dis 2015] EULAR recommendations how to organize education in MSUS [ Iagnocco et al., RMD Open 2015] Standard scanning MSUS positions in four most affected joints in JIA [ Collado 2016] Definition of age related vascularization and ossification of joints in children [ Windschall 2017] Definitions of normal joint structures MSUS findings in children [ Roth et al. Arthritis Care Res 2015]. EULAR/PReS Standardized Procedures for Ultrasound Imaging in Paediatric Rheumatology (ongoing EULAR project CLI089 to produce on line web application as an educational tool) Preliminary Definitions for the Sonographic Features of Synovitis in Children [ Roth 2017] 3.2 PReS clinical skills workshops - Joint examination skills ( pGALS Foster and Jandial, pREMS Foster 2011) are included in the PReS Basic Course (with a focus on pGALS) and Advanced JIA Course (with a focus on pREMS and ‘how to teach’ pGALS). Feedback to date has been very positive. The format includes ‘hands on’ workshops involving consented patients demonstrating physical signs and small groups of learners with facilitated observation and feedback from experienced paediatric rheumatologists. The ETC aims for the ‘clinical skills’ workshop concept to be expanded and to include, for example, scleroderma skin scores, nail fold capillaroscopy, muscle power testing, and transitional care consultation skills. 3.3 PReS ‘Teach the teachers’ workshops - Paediatric Rheumatologists have an integral role as educators and need to ‘reach out’ to medical students, general paediatricians and colleagues in orthopaedics and primary care to raise awareness, facilitate diagnosis and prompt referral to specialist care. Learning ‘how to teach’ is therefore an important skill and included in the European syllabus. The aim of the ETC is that all fellows will attend a ‘how to teach’ course during their training. PReS has included ‘teach the teacher’ workshops alongside the PReS scientific meetings (2016, 2017) with excellent feedback and it is envisaged that such workshops will be included in future PReS meetings and the Advanced Course rolling programme.
PReS Website ( http://www.pres.eu ) (revised and launched 2017) has signposting to available resources, conferences, courses and available bursaries. Pediatric Rheumatology Online Journal – ( https://ped-rheum.biomedcentral.com ) founded in 2003, is a free open access journal (current impact factor of 2.328 September 2018), and supported through sponsorship from PReS. The journal has a broad range of content including clinical reviews, clinical research and basic science. The journal also publishes abstracts and proceedings from the annual scientific meetings. The EULAR / PReS paediatric rheumatology online course (launched 2014) was developed with a grant from EULAR and enrolment is subsidized by EULAR with further discounted costs for low and middle income countries ( https://www.eular.org/edu_online_course_paediatric.cfm ). The online course is a collaborative effort between PReS and EULAR, with content written and updated yearly by senior paediatric rheumatologists with the assistance of trainees. The content is aimed at fellows in paediatric rheumatology at the start of their training, residents in adult rheumatology and paediatricians or adult rheumatologists with an interest in paediatric rheumatology. The course consists of 10 modules, each dedicated to a specific topic with a knowledge test at the end, and covering key topics as outlined in the 2016 European Syllabus in paediatric rheumatology. The course can be completed with the final online examination once a year and provides a EULAR/PReS certification for specific knowledge of paediatric rheumatology. The course has attracted more than 650 participants to date; the aim being to promote the course as the leading online education in specialist paediatric rheumatology worldwide. A EULAR/PReS published textbook to supplement the course is also now available. The Paediatric Musculoskeletal Matters [PMM] website ( www.pmmonline.org ) can be accessed from the PReS website ( http://www.pres.eu/activities/scientific-and-clinical/educational-instruments-and-tools.html ). PMM was launched in 2014, is a free and open e-resource and targets medical students and family medicine doctors [ Smith et al. 2016] although content is relevant to all clinicians who encounter children in their clinical practice. PMM has to date > 250,000 hits from > 150 countries. PMM International (launched September 2018) reflects paediatric rheumatology health care around the world and has contributions from an international panel of paediatric rheumatologists. PMM-Nursing is also available (launched 2017) and targets all levels of nurses involved in the care of children with rheumatic diseases. PMM is endorsed by PReS as a resource for paediatric rheumatologists and teams to use in their teaching and encompasses basic clinical skills, descriptions of normal development, the approach to investigations and initial management of musculoskeletal presentations including red flags for serious life threatening conditions. PMM is promoted as a foundation for the Basic Course and to those embarking on the EULAR/PReS online course. PMM signposts to the free pGALS app (with the 2018 version including multiple language translations) and pGALS e-module. Patient and parent information - Further to the SHARE initiative there has been a complete revision of the freely available information leaflets covering a broad spectrum of conditions and including translations into several languages ( https://www.printo.it/pediatric-rheumatology/ ). The importance of patient and family advocacy and engagement in clinical care, service delivery and research has been highlighted [ Dijkhuizen et al] and includes recommendations to optimise participation and facilitate better clinical outcomes.
PReS EMERGE (EMErging RheumatoloGists and rEsearchers) encompasses young paediatric rheumatologists and researchers working together to improve clinical and research opportunities for trainees, participating in organisation of PReS educational events (YIM, Basic and Advanced courses) and liaising with other young investigator groups (e.g. EMEUNET, CARRA early investigators). The group was set up following the 2016 YIM and currently includes trainees mainly from Europe but also from all other continents. Further details are available through an active social media network ( www.facebook.com/PReSEMERGE , twitter.com/PReSEMERGE ) and through a bimonthly newsletter highlighting activities and opportunities for members (contact [email protected]). The PReS EMERGE fellowship programme ( https://www.pres.eu/activities/young-investigators/fellowship-programs.html ) was launched in 2017 with financial and practical assistance for clinical trainees who are members of PReS and younger than 40 years, to facilitate placements of up to 6-months within a European Paediatric Rheumatology Centre. In addition to gaining clinical knowledge and skills, the trainee is expected to participate in a research project. From 2018, the programme is open to basic science trainees (pre-PhD to five years post-PhD) working in paediatric rheumatology. It is envisaged that the programme will enhance both clinical and basic collaborative research, foster a network of emerging and established paediatric rheumatologists and allow sharing of ideas and practices between different countries to harmonize paediatric rheumatology training.
The Standards of care for Juvenile Arthritis Management in Less Resourced countries (JAMLess) is a collaborative effort with funding and support from International League Against Rheumatism (ILAR) and PReS to develop consensus based guidelines relevant to challenging health care contexts; these highlight the importance of education and training being integral to the delivery of clinical care. This project was initiated by PReS colleagues from South Africa, Argentina and India and has resulted in a consensus document [ Scott C 2018], with further work to develop iterations of these recommendations for other parts of the world being underway.
The activities of the PReS ETC aim to improve access to, and provision of, high quality clinical care delivered by an appropriately trained workforce and ultimately improve outcomes for children and families. The emergence of Basic and Advanced courses, specialist courses, online resources and the EMERGE fellowship programme, supplement the annual scientific meeting of PReS and YIM as a means to facilitate advances in knowledge being implemented into high quality evidence-based clinical care. We recognize that more work is needed to enable and support the expanding paediatric rheumatology community both in Europe and Internationally. The uptake and implementation of the new 2016 European syllabus into training programmes needs to address the accreditation of training centres, a proposed European sub-specialist examination and certification in paediatric rheumatology. There is now an overarching structure to the education and training activities of the PReS ETC to address the expanding needs of the broad paediatric rheumatology community with the content based on the 2016 European Syllabus for training. The format and structure of the PReS educational portfolio, much of which is free and open to all, serves as a template relevant to the wider international paediatric rheumatology. There is clear synergy between the educational and training activities of PReS and other initiatives led by paediatric rheumatology organisations elsewhere in the world (such as those developed by ACR, CARRA, BSR, BSPAR, APRG). Through working with colleagues around the world, the aim has been to ‘reach out’, raise awareness and facilitate growth of paediatric rheumatology relevant to the local context.
|
SSTR2 positively associates with EGFR and predicts poor prognosis in nasopharyngeal carcinoma | 5933bb90-80f6-496b-aeeb-057d299e3d8f | 11671960 | Anatomy[mh] | Recent studies have highlighted the effectiveness of epidermal growth factor receptor (EGFR)-targeted drugs in nasopharyngeal carcinomas (NPC), yet drug resistance has emerged as a challenge. SSTR2, highly expressed in NPC, is considered a potential tumour marker and therapeutic target. This scientific article explores the correlation between SSTR2 and EGFR expression in NPC at the protein level and investigates their prognostic significance in response to different treatments. Through a retrospective analysis of clinical data and tissue samples, a significant correlation between SSTR2 and EGFR expression was observed, suggesting a potential functional relationship. The study emphasises the importance of considering the expression profiles of these receptors for personalised therapeutic approaches in NPC treatment and lays a foundation for the development of targeted drugs and signalling pathways involving EGFR and SSTR2. Nasopharyngeal carcinoma (NPC) is an epithelial cancer that occurs in the mucosa lining of the nasopharynx. This cancer is caused by a combination of factors, with the most significant being infection with Epstein-Barr virus (EBV). Persistently elevated EBV antibody levels are considered a major risk factor for NPC. NPC is a regional cancer, with high incidence rates in southern China and Southeast Asia. The disease is classified into two types based on histological analysis: non-keratinising squamous cell carcinoma and keratinising squamous cell carcinoma. The majority of NPC cases in China are non-keratinising squamous cell carcinoma and characterised by high levels of EBV antibodies. Somatostatin (SST) is a potent neuroendocrine hormone that is widely expressed in the human body and plays a crucial role in regulating cell proliferation. In addition to SST receptor 2 (SSTR2), SST has four membrane surface receptors collectively known as SSTRs. It is widely accepted that SSTRs exert an antiproliferative effect on cells and trigger downstream signalling that promotes apoptosis and inhibits tumour growth factors. SSTR2 is a G-protein-coupled receptor (GPCR) with diverse biological functions in humans. The activation of the SSTR2 pathway is known to cause cell cycle arrest or apoptosis in low-grade neuroendocrine tumours. However, the opposite occurs in high-grade neuroendocrine tumours and small-cell lung cancers, where SSTR2 is upregulated, leading to tumour growth. Epidermal growth factor receptor (EGFR) belongs to the receptor tyrosine kinases family and overexpression of EGFR has been linked to poor prognosis and cancer progression. The EGFR plays a crucial role in maintaining homeostasis in epithelial tissues during normal physiological conditions. However, mutations or overexpression of EGFR occurred frequently in pathological conditions, leading to the development of tumours including head and neck squamous cell carcinomas. EGFR activates various signalling pathways that transmit signals from the cell surface to the nucleus, promoting cellular survival, proliferation and differentiation. Anti-human EGFR monoclonal antibodies have been developed as a treatment option for cancers such as NPC. The combination of chemoradiotherapy (CR) and EGFR-targeted therapy has been shown to significantly improve survival rates. LMP1, a proteins product of EBV, governs proliferative signalling pathways, including those associated with EGFR and NF-KB. In NPC, NF-KB signalling has been demonstrated to regulate SSTR2 expression, and additionally, EGFR-mediated MAPK/ERK signalling has also been shown to regulate SSTR2. However, the potential relationship between EGFR and SSTR2 has not been thoroughly investigated in NPC. Our study is the first to shed light on the intricate relationship between SSTR2 and EGFR in NPC and provides new insights into the potential benefits of EGFR targeted therapy for patients with high SSTR2 expression. In addition, SSTR2 has potential as a new biomarker for poor prognosis in NPC patients. These findings could potentially offer a theoretical foundation for investigating signalling pathways and advancing the development of future targeted therapeutic agents. Bioinformatics analysis It was obtained mRNA-seq data from 113 patients with NPC (GSE102349) from the GEO database ( https://www.ncbi.nlm.nih.gov/geo/ ). We normalised and analysed this data to investigate the correlation between EGFR and SSTR2. Tissue specimens and clinical data This study included 491 NPC samples and 50 non-cancerous nasopharyngeal epithelium samples. All samples were collected between 2016 and 2021 at The Second Xiangya Hospital in Changsha, China. Among the 491 NPC cases, 65 were clinically early (stages I and II) and 426 were clinically advanced (stages III and IV). A totall of 433 cases had lymph node metastases, and 37 cases had distant metastases. All participants had received a biopsy tissue prior to CR, and complete medical records as well as follow-up records were available. The pathological diagnosis of all specimens was made according to the latest WHO classification for head and neck tumours, and the patients were categorised according to the eighth edition of the UICC（ Union Internationale Contre le Cancer ）/AJCC（ American Joint Committee on Cancer）. Immunohistochemistry and scores Immunohistochemistry (IHC) staining was conducted on paraffin-embedded of the NPC and non-cancerous nasopharyngeal epithelium tissues sections (4 µm). The tissues were first dewaxed and hydrated, followed by heating with EDTA repair solution for 7 min. Endogenous peroxidase was blocked using H 2 O 2 (3%) for 30 min. Incubation with primary antibodies for SSTR2 (1:200; Rabbit monoclonal antibody; EP149; MXB) and EGFR (ready-to-use antibody; Rabbit monoclonal antibody; EP38Y; MXB) followed, and then secondary antibody incubation. The visualisation signal was obtained with 3, 3’-diaminobenzidine tetrachloride. We assessed staining only in epithelial cells of benign or malignant tissue and not lymphoid tissue. The staining revealed positive expression of SSTR2 and EGFR on the cytoplasm and cell membrane. The evaluation of protein expression was based on staining intensity and extent, which were semiquantitatively assessed under light microscopy by scorers who were blinded to the case data. Staining intensity was graded as 0 (negative), 1 (weak), 2 (moderate) and 3 (strong), while staining extent was graded as 0 (0%), 1 (1%–25%), 2 (26%–50%), 3 (51%–75%) and 4 (76%–100%). Scores were calculated by multiplying the staining intensity and extent. A score ≤2 for SSTR2 was considered low expression, while a score >2 was considered high expression. Similarly, an EGFR score of >2 was considered high expression based on data collected for NPC and previous studies. Statistical analysis The statistical analyses and graphing in this study were performed using a variety of methods, including χ 2 tests, Spearman correlations, univariate Cox regression and multivariate Cox regression analysis, all of which were performed using SPSS Statistics V.26 (SPSS) for macOS. Kaplan-Meier analysis, Bar graphs and Violin plots were conducted with GraphPad Prism V.9.4.1 (GraphPad, La Jolla, California, USA) for macOS. The progression-free survival (PFS) is defined as the period from the date of diagnosis until the patient dies or the disease progresses further. The overall survival (OS) is defined as the period from the date of diagnosis to the date of death or the last known date alive. The p values were analysed as two-sided statistics, and a p<0.05 was considered statistically significant. It was obtained mRNA-seq data from 113 patients with NPC (GSE102349) from the GEO database ( https://www.ncbi.nlm.nih.gov/geo/ ). We normalised and analysed this data to investigate the correlation between EGFR and SSTR2. This study included 491 NPC samples and 50 non-cancerous nasopharyngeal epithelium samples. All samples were collected between 2016 and 2021 at The Second Xiangya Hospital in Changsha, China. Among the 491 NPC cases, 65 were clinically early (stages I and II) and 426 were clinically advanced (stages III and IV). A totall of 433 cases had lymph node metastases, and 37 cases had distant metastases. All participants had received a biopsy tissue prior to CR, and complete medical records as well as follow-up records were available. The pathological diagnosis of all specimens was made according to the latest WHO classification for head and neck tumours, and the patients were categorised according to the eighth edition of the UICC（ Union Internationale Contre le Cancer ）/AJCC（ American Joint Committee on Cancer）. Immunohistochemistry (IHC) staining was conducted on paraffin-embedded of the NPC and non-cancerous nasopharyngeal epithelium tissues sections (4 µm). The tissues were first dewaxed and hydrated, followed by heating with EDTA repair solution for 7 min. Endogenous peroxidase was blocked using H 2 O 2 (3%) for 30 min. Incubation with primary antibodies for SSTR2 (1:200; Rabbit monoclonal antibody; EP149; MXB) and EGFR (ready-to-use antibody; Rabbit monoclonal antibody; EP38Y; MXB) followed, and then secondary antibody incubation. The visualisation signal was obtained with 3, 3’-diaminobenzidine tetrachloride. We assessed staining only in epithelial cells of benign or malignant tissue and not lymphoid tissue. The staining revealed positive expression of SSTR2 and EGFR on the cytoplasm and cell membrane. The evaluation of protein expression was based on staining intensity and extent, which were semiquantitatively assessed under light microscopy by scorers who were blinded to the case data. Staining intensity was graded as 0 (negative), 1 (weak), 2 (moderate) and 3 (strong), while staining extent was graded as 0 (0%), 1 (1%–25%), 2 (26%–50%), 3 (51%–75%) and 4 (76%–100%). Scores were calculated by multiplying the staining intensity and extent. A score ≤2 for SSTR2 was considered low expression, while a score >2 was considered high expression. Similarly, an EGFR score of >2 was considered high expression based on data collected for NPC and previous studies. The statistical analyses and graphing in this study were performed using a variety of methods, including χ 2 tests, Spearman correlations, univariate Cox regression and multivariate Cox regression analysis, all of which were performed using SPSS Statistics V.26 (SPSS) for macOS. Kaplan-Meier analysis, Bar graphs and Violin plots were conducted with GraphPad Prism V.9.4.1 (GraphPad, La Jolla, California, USA) for macOS. The progression-free survival (PFS) is defined as the period from the date of diagnosis until the patient dies or the disease progresses further. The overall survival (OS) is defined as the period from the date of diagnosis to the date of death or the last known date alive. The p values were analysed as two-sided statistics, and a p<0.05 was considered statistically significant. Patients with NPC and those with non-cancerous nasopharyngeal epithelium exhibit differential expression levels of SSTR2 and EGFR Initially, we performed bioinformatics analysis using the GEO database, which revealed a positive correlation between SSTR2 and EGFR in NPC (R=0.412, moderate positive association) . Subsequently, we investigated the expression levels of SSTR2 and EGFR proteins using IHC. The study demonstrated that patients diagnosed with NPC exhibited a significantly higher expression rate of SSTR2 (45.0%, 221 out of 491) and EGFR (71.7%, 352 out of 491) in high expression compared with the control group, which exhibited a SSTR2 high expression rate of only 6.0% (3 out of 50) and an EGFR high expression rate of 48.0% (24 out of 50) . Furthermore, the distribution of SSTR2 and EGFR scores between the NPC and control groups were represented using violin plots, which showed significant differences between the two groups (p<0.001) . These findings suggest that high expression levels of SSTR2 and EGFR are significantly associated with NPC rather than non-cancerous nasopharyngeal epithelium (p<0.05). Here, we present a partial IHC image of SSTR2 and EGFR protein expression in NPC and non-cancerous nasopharyngeal epithelium . EGFR and SSTR2 proteins expression and clinicopathological features First, we classified and analysed NPC patients based on various factors, including age, sex, clinical stage and TNM stage. The resulting table revealed that NPC is more prevalent in older age groups and male. Furthermore, our study cohort included a larger proportion of patients with advanced clinical stages and lymph node metastases, while distant metastases were relatively rare. Subsequently, we investigated the correlation between SSTR2 and/or EGFR protein expression and clinicopathological features, such as gender, age, TNM stage, treatment strategy and disease progression. shows that patients with advanced T stage, lymph node metastases and distant metastases tended to exhibit higher expression levels of SSTR2 and/or EGFR than those with early T stage and no metastases, although this was not statistically significant (p>0.05). Furthermore, high expression levels of SSTR2 and EGFR were significantly associated with worse outcomes and a higher risk of progression (p<0.05). Prognosis and correlation of SSTR2 and EGFR The influence of various variables on PFS and OS was assessed using Cox univariate analysis, and the corresponding tables were created ( and ). Our findings revealed that high EGFR expression, advanced T stage, lymph node metastasis, distant metastasis and general CR were associated with a poor prognosis for both PFS and OS. However, high SSTR2 expression was identified as a poor prognostic factor only for PFS, while older age was found to be a poor prognostic factor solely for OS. Moreover, multivariate Cox analysis ( and ) was conducted to determine independent prognostic factors, which revealed that both M stage and treatment strategy were significant factors for both PFS and OS. However, SSTR2 and EGFR were independent prognostic factors only for PFS. Furthermore, K-M analysis was performed on a cohort of 491 patients , which indicated that patients with high expression of EGFR, high coexpression of EGFR and SSTR2, and EGFR/SSTR2 anyone high expression had a poorer prognosis for both PFS and OS and high expression of SSTR2 had a poorer prognosis for PFS (p<0.05). Based on these studies, overexpression of SSTR2 and EGFR is detrimental to patients with NPC. SSTR2 has the potential to serve as a new biomarker for poor prognosis in patients with NPC. Additionally, we found a positive correlation between EGFR and SSTR2 at the mRNA level using the GEO database (R=0.412) as described previously. We also demonstrated that SSTR2 correlates with EGFR at the protein level via IHC (R=0.296, weak positive association) (data not shown). EGFR-targeted therapy affected the prognosis associated with SSTR2 expression Our study cohort consisted of 288 patients who were treated and reviewed at our institution, providing us with comprehensive and timely clinical information. Among the patients, 147 received only CR, while 141 received CR along with EGFR targeted therapy. We evaluated the clinical prognosis of these patients. First, the study found that patients receiving general CR with high expression of SSTR2, high expression of EGFR and high coexpression of both had a poorer prognosis in both PFS and OS (p<0.05). Although the prognosis did not reach a statistical difference between the EGFR/SSTR2 anyone high expression and others (low coexpression of EGFR and SSTR2) in PFS and OS, low coexpression had a better prognosis (p>0.05) . Subsequently, we examined the prognosis of NPC patients treated with general CR combined with targeted therapy and found no significant differences in EGFR expression levels (p>0.05), indicating that EGFR targeted drugs improved the poor prognosis arising from high EGFR expression. Surprisingly, there was also no statistical difference in SSTR2 expression levels, between EGFR/SSTR2 anyone high expression and others (low coexpression of EGFR and SSTR2) in PFS and OS, and between high coexpression of EGFR and SSTR2 and other factors in OS . This finding led us to speculate whether EGFR targeted therapy could also improve the poor prognosis of NPC patients associated with high SSTR2 expression. To investigate further, we compared the prognosis of both treatment therapies and found that NPC patients with high expression of SSTR2, high expression of EGFR, high coexpression of EGFR and SSTR2, and EGFR/SSTR2 anyone high expression all had a better prognosis with CR combined with targeted therapy, with a significant statistical difference (p<0.05) . Our findings suggest that CR combined with EGFR targeted therapy is more effective for NPC with high SSTR2 expression and high EGFR expression. EGFR targeted therapy significantly improves the poor prognosis of NPC patients with high expression of SSTR2 and EGFR. Initially, we performed bioinformatics analysis using the GEO database, which revealed a positive correlation between SSTR2 and EGFR in NPC (R=0.412, moderate positive association) . Subsequently, we investigated the expression levels of SSTR2 and EGFR proteins using IHC. The study demonstrated that patients diagnosed with NPC exhibited a significantly higher expression rate of SSTR2 (45.0%, 221 out of 491) and EGFR (71.7%, 352 out of 491) in high expression compared with the control group, which exhibited a SSTR2 high expression rate of only 6.0% (3 out of 50) and an EGFR high expression rate of 48.0% (24 out of 50) . Furthermore, the distribution of SSTR2 and EGFR scores between the NPC and control groups were represented using violin plots, which showed significant differences between the two groups (p<0.001) . These findings suggest that high expression levels of SSTR2 and EGFR are significantly associated with NPC rather than non-cancerous nasopharyngeal epithelium (p<0.05). Here, we present a partial IHC image of SSTR2 and EGFR protein expression in NPC and non-cancerous nasopharyngeal epithelium . First, we classified and analysed NPC patients based on various factors, including age, sex, clinical stage and TNM stage. The resulting table revealed that NPC is more prevalent in older age groups and male. Furthermore, our study cohort included a larger proportion of patients with advanced clinical stages and lymph node metastases, while distant metastases were relatively rare. Subsequently, we investigated the correlation between SSTR2 and/or EGFR protein expression and clinicopathological features, such as gender, age, TNM stage, treatment strategy and disease progression. shows that patients with advanced T stage, lymph node metastases and distant metastases tended to exhibit higher expression levels of SSTR2 and/or EGFR than those with early T stage and no metastases, although this was not statistically significant (p>0.05). Furthermore, high expression levels of SSTR2 and EGFR were significantly associated with worse outcomes and a higher risk of progression (p<0.05). The influence of various variables on PFS and OS was assessed using Cox univariate analysis, and the corresponding tables were created ( and ). Our findings revealed that high EGFR expression, advanced T stage, lymph node metastasis, distant metastasis and general CR were associated with a poor prognosis for both PFS and OS. However, high SSTR2 expression was identified as a poor prognostic factor only for PFS, while older age was found to be a poor prognostic factor solely for OS. Moreover, multivariate Cox analysis ( and ) was conducted to determine independent prognostic factors, which revealed that both M stage and treatment strategy were significant factors for both PFS and OS. However, SSTR2 and EGFR were independent prognostic factors only for PFS. Furthermore, K-M analysis was performed on a cohort of 491 patients , which indicated that patients with high expression of EGFR, high coexpression of EGFR and SSTR2, and EGFR/SSTR2 anyone high expression had a poorer prognosis for both PFS and OS and high expression of SSTR2 had a poorer prognosis for PFS (p<0.05). Based on these studies, overexpression of SSTR2 and EGFR is detrimental to patients with NPC. SSTR2 has the potential to serve as a new biomarker for poor prognosis in patients with NPC. Additionally, we found a positive correlation between EGFR and SSTR2 at the mRNA level using the GEO database (R=0.412) as described previously. We also demonstrated that SSTR2 correlates with EGFR at the protein level via IHC (R=0.296, weak positive association) (data not shown). Our study cohort consisted of 288 patients who were treated and reviewed at our institution, providing us with comprehensive and timely clinical information. Among the patients, 147 received only CR, while 141 received CR along with EGFR targeted therapy. We evaluated the clinical prognosis of these patients. First, the study found that patients receiving general CR with high expression of SSTR2, high expression of EGFR and high coexpression of both had a poorer prognosis in both PFS and OS (p<0.05). Although the prognosis did not reach a statistical difference between the EGFR/SSTR2 anyone high expression and others (low coexpression of EGFR and SSTR2) in PFS and OS, low coexpression had a better prognosis (p>0.05) . Subsequently, we examined the prognosis of NPC patients treated with general CR combined with targeted therapy and found no significant differences in EGFR expression levels (p>0.05), indicating that EGFR targeted drugs improved the poor prognosis arising from high EGFR expression. Surprisingly, there was also no statistical difference in SSTR2 expression levels, between EGFR/SSTR2 anyone high expression and others (low coexpression of EGFR and SSTR2) in PFS and OS, and between high coexpression of EGFR and SSTR2 and other factors in OS . This finding led us to speculate whether EGFR targeted therapy could also improve the poor prognosis of NPC patients associated with high SSTR2 expression. To investigate further, we compared the prognosis of both treatment therapies and found that NPC patients with high expression of SSTR2, high expression of EGFR, high coexpression of EGFR and SSTR2, and EGFR/SSTR2 anyone high expression all had a better prognosis with CR combined with targeted therapy, with a significant statistical difference (p<0.05) . Our findings suggest that CR combined with EGFR targeted therapy is more effective for NPC with high SSTR2 expression and high EGFR expression. EGFR targeted therapy significantly improves the poor prognosis of NPC patients with high expression of SSTR2 and EGFR. The SSTR2 receptor plays a critical role in regulating cell proliferation and acts as a growth suppressor in various biological processes. Several studies have shown that SSTR2 can inhibit tumour growth in low-grade neuroendocrine tumours and prostate cancer. However, high levels of SSTR2 expression have been linked to promoting tumour growth in high-grade neuroendocrine tumours and small cell lung cancer. The study of EGFR proteins has gained significant attention in recent years due to their involvement in activating various signalling pathways and regulating cell proliferation and survival. The findings of our study have significant implications for the identification of SSTR2 and EGFR as biomarkers associated with poor prognosis provides opportunities for targeted therapies. Targeting SSTR2 could be a potential strategy to improve the survival outcomes of NPC patients, particularly those with high SSTR2 expression. However, our results also suggest that the use of EGFR targeted therapy in combination with chemotherapy could be a more effective approach for patients with high SSTR2 and EGFR expression levels. The surprising result of EGFR targeted therapy suppressing the effects of high SSTR2 expression warrants further investigation. It is possible that EGFR targeted therapy indirectly affects SSTR2 expression or that it inhibits signalling pathways that promote tumour growth in SSTR2 overexpressing cells. These findings could pave the way for developing novel treatment strategies that target both EGFR and SSTR2 in NPC patients. In conclusion, our study provides valuable insights into the role of SSTR2 and EGFR in the prognosis of NPC patients. Our results suggest that targeting both biomarkers could be a promising strategy to improve the survival outcomes of NPC patients. Further studies are needed to validate our findings and to explore the mechanisms underlying the interaction between EGFR and SSTR2 in NPC. Compared with previous research on the role of SSTR2 in NPC, our study stands out with a larger sample size of 491 patients. Overall, our study is the first to shed light on the intricate relationship between SSTR2 and EGFR in NPC and provides new insights into the potential benefits of EGFR targeted therapy for patients with high SSTR2 expression. Future studies could further investigate the molecular mechanisms underlying this relationship and explore potential alternative therapies for patients with high SSTR2 expression. Additionally, efforts should be made to address the toxicity concerns associated with EGFR targeted therapy and optimise treatment strategies to improve patient outcomes. 10.1136/jcp-2023-208987 online supplemental table 1 10.1136/jcp-2023-208987 online supplemental table 2 10.1136/jcp-2023-208987 online supplemental table 3 10.1136/jcp-2023-208987 online supplemental figure 1 10.1136/jcp-2023-208987 Abstract translation 1 |
Curcumin in Ophthalmology: Mechanisms, Challenges, and Emerging Opportunities | 599136e1-4771-45ee-b170-1aac179f69dd | 11820683 | Ophthalmology[mh] | Ocular diseases, encompassing retinal and corneal disorders alongside ocular surface conditions, such as eyelid pathologies, significantly contribute to the global burden of visual impairment . These conditions pose significant challenges in prevention and treatment. For instance, diabetic retinopathy (DR) affected over 100 million people globally in 2020, with projections surpassing 160 million by 2045 . Similarly, glaucoma, a leading cause of irreversible blindness, is estimated to impact 111.8 million people aged 40–80 worldwide by 2040 . Dry eye syndrome impacts a significant portion of the population, with prevalence rates varying widely between 5% and 50%, underscoring the growing burden of age-related ocular conditions, which are expected to increase with the aging population, projected to double to 2.1 billion by 2050 . Moreover, lifestyle factors, such as unhealthy eating habits, smoking, and the frequent use of digital devices, have intensified these challenges. Worryingly, projections from the Global Burden of Disease Study suggest that by 2050, approximately 474 million people may experience moderate to severe visual impairments, with 61 million potentially losing their sight entirely . In recent years, considering these obstacles, there has been growing interest in curcumin as a possible therapeutic agent in managing ocular diseases. Curcumin (C 21 H 20 O 6 ), a lipophilic polyphenol derived from the dried rhizome of Curcuma longa L. and related species, has gained significant attention due to its extensive pharmacological properties, including anti-inflammatory, antioxidant, antimicrobial, and antitumor activities . Alongside its primary forms—demethoxycurcumin and bis-demethoxycurcumin—turmeric contains over 50 additional curcuminoids, including bisabocurcumin, curcumalongin, cyclocurcumin, and terpecurcumin, as well as volatile oils and resins. These compounds broaden turmeric’s pharmacological profile, offering synergistic effects that enhance its therapeutic versatility and reinforce its global use both as a culinary spice and as a source of health benefits . Curcumin’s molecular structure, comprising two o-methoxy phenolic aromatic rings linked by a seven-carbon α, β-unsaturated β-diketone chain, underpins its pleiotropic effects. Its properties—anti-inflammatory, antioxidant, antibacterial, anti-angiogenic, and anti-apoptotic—show promise in ophthalmology . Research indicates its potential for treating corneal and retinal neovascularization, inhibiting lens epithelial cell proliferation, and modulating retinal pigment epithelium-related pathways, making it a valuable candidate for managing inflammatory and degenerative ocular diseases . Topical formulations, such as hydrogels, creams, and nanocarrier systems, have been developed to enhance their physicochemical properties, including solubility, permeability, and stability. These innovations protect curcumin from degradation and enable sustained release, proving effective in treating dermatological conditions, such as psoriasis, acne, and atopic dermatitis, due to its anti-inflammatory, wound-healing, and antioxidant properties. This success in dermatology has spurred interest in its application for ocular treatments due to similar physicochemical barriers . Recognized by the FDA as ‘Generally Recognized as Safe’ (GRAS) for human consumption, curcumin demonstrates significant therapeutic potential . Clinical trials have confirmed their excellent safety, tolerability, and efficacy, even at high oral doses ranging from 4 to 8 g per day and doses up to 12 g per day for curcuminoid formulations containing 95% curcumin, bisdemethoxycurcumin, and demethoxycurcumin . However, its clinical application is limited by critical pharmacokinetic challenges, including poor aqueous solubility, light sensitivity, low bioavailability, limited absorption, and rapid systemic metabolism and elimination. These factors complicate consistent therapeutic outcomes and pharmacological interpretations, particularly given curcumin’s classification as a PAIN (pan-assay interference compound) and an IMP (invalid metabolic panacea), which highlights its complex bioactivity. Its degradation of products and fluorescence further complicate pharmacological evaluations . These limitations not only hinder consistent therapeutic outcomes but also complicate the identification of the actual bioactive species responsible for its effects. In addition, the metabolism of curcumin and its interaction with the intestinal microbiota play a crucial role in determining its bioavailability and therapeutic efficacy. After oral administration, curcumin exhibits poor solubility and limited gastrointestinal absorption. The absorbed fraction undergoes rapid metabolism in the liver and intestine via reduction (yielding dihydrocurcumin and tetrahydrocurcumin) and conjugation (forming glucuronides and sulfates), leading to its swift elimination. The intestinal microbiota further converts curcumin into more stable and, in some cases, bioactive metabolites. Certain bacterial genera, such as Bifidobacterium and Lactobacillus , promote its reduction and demethylation, potentially enhancing biological activity. Conversely, curcumin modulates the gut microbiota, fostering beneficial species while inhibiting pathogens. These interactions have significant implications for curcumin’s therapeutic applications in ocular diseases, especially given the distinct metabolic pathways associated with oral and topical administration . Additionally, the wide range of commercially available formulations—ranging from turmeric powder to curcuminoid-enriched products and purified curcumin—adds complexity, raising concerns about reproducibility and efficacy in clinical trials . Despite these challenges, curcumin’s diverse therapeutic potential emphasizes the need for innovative delivery systems. Nanocarrier technologies, particularly vesicular systems, such as liposomes and proniosomes, have addressed some of these challenges by improving curcumin’s bioavailability, solubility, and stability . These systems encapsulate curcumin within surfactant vesicles, protecting it from enzymatic degradation and extending its therapeutic presence on ocular surfaces. Moreover, the sustained drug delivery provided by these systems reduces systemic side effects while targeting disease-specific sites . In ocular inflammation, curcumin has demonstrated efficacy in reducing complications, such as corneal opacity, cataract formation, and retinal detachment . It also holds promise as a prophylactic agent in proliferative vitreoretinopathy (PVR), with studies reporting reduced rates of retinal detachment following surgery . Furthermore, curcumin demonstrates therapeutic benefits in DR by modulating hyperglycemia-induced endothelial dysfunction . Topical drug delivery systems, such as eye drops, i.e., aqueous solutions and suspensions, and oil-based formulations, remain widely used for ocular treatment. These formulations are intended for direct application to the ocular surface, typically in the form of drops. However, they often face significant limitations, including excessive tear production, rapid drainage, and systemic absorption, leading to inefficient drug distribution and the loss of over 95% of the administered dose . This review explores curcumin’s therapeutic potential in ophthalmology, focusing on its molecular mechanisms, challenges in clinical application, and advanced strategies for optimized delivery. By addressing these barriers, curcumin could transform ocular disease management, highlighting the need for robust randomized trials to confirm its safety and efficacy. Delivering medications to the eye effectively remains a significant challenge due to its distinct pharmacokinetic and pharmacodynamic environment. The eye’s natural defense mechanisms—such as tear production, blinking, and the intricate clearance processes on the ocular surface—serve to protect it but simultaneously hinder drug retention and absorption. These barriers, coupled with the anatomical complexity of anterior and posterior segments, result in low bioavailability for many conventional therapies . Frequent application of traditional eye drops is often necessary to achieve therapeutic outcomes, but this practice can inadvertently lead to systemic side effects through absorption via the nasolacrimal pathway . Recent studies suggest that curcumin can be applied to various ophthalmic conditions, offering significant therapeutic potential for a wide range of ocular diseases and addressing many limitations of conventional approaches . Curcumin’s antimicrobial and immunomodulatory properties make it particularly effective in targeting the complex interplay of infection, inflammation, and oxidative stress underlying many ocular pathologies. It directly disrupts bacterial cell walls and inhibits enzymatic processes critical for bacterial survival while also downregulating pro-inflammatory cytokines and mitigating oxidative stress. These dual antibacterial and anti-inflammatory actions position curcumin as a versatile therapeutic agent, particularly when integrated into advanced drug delivery systems . As shown in , the development of advanced drug delivery systems, including in situ gels, nanostructured lipid carriers, and hydrogels, has emerged as a promising strategy to enhance the solubility, stability, and ocular bioavailability of curcumin. An in vitro study, supported by ex vivo assays using rabbit corneas, showed that polyethylene glycol-distearoylphosphatidylethanolamine (PEG-DSPE)/Solutol HS 15 mixed micelle-based in situ gels improve corneal penetration, ocular retention, and stability. This system also supports sustained drug release, reduces dosing frequency, and avoids ocular irritation, offering a promising alternative to conventional eye drops . Thiolated chitosan-coated nanostructured lipid carriers, characterized in vitro and further evaluated in vivo using animal models, enhance corneal contact through covalent bonding with mucus glycoproteins. This interaction ensures sustained release over 72 h without irritation, improving ocular distribution and therapeutic efficacy . Hydroxypropyl methylcellulose methacrylate hydrogels, tested in vivo, demonstrated strong bioadhesion and controlled curcumin release, contributing to the reduction of oxidative damage in trabecular meshwork cells. This effect helps mitigate the inflammatory and apoptotic processes that are key in glaucoma progression, confirming the safety and potential of this formulation for controlled ocular drug delivery . These studies highlight the potential of these technologies to improve ocular retention, reduce dosing frequency, and minimize systemic side effects. Such advancements enable sustained drug release, improve therapeutic outcomes, and alleviate the burden of frequent applications associated with traditional formulations in ocular diseases. To better understand the impact of delivery methods on curcumin’s bioavailability and therapeutic potential, provides a comparison of different routes of administration, including topical ocular systems, oral administration (including trial clinical phase I), and parenteral routes. Ocular drug delivery systems offer a distinct advantage by bypassing the first-pass metabolism typical of oral administration and directly targeting the eye. These formulations improve curcumin’s retention, permeability, and controlled release, leading to enhanced therapeutic efficacy when compared to oral or intravenous routes. Additionally, ocular systems minimize systemic exposure, reducing the likelihood of side effects and providing a more focused and controlled therapeutic approach. Curcumin serves as a central modulator in multiple molecular systems involved in ocular health. It promotes a dynamic equilibrium between cellular processes that sustain ocular tissue integrity . This bioactive compound uniquely interacts with key signaling pathways, adjusting their intensity and function to reach an ideal homeostatic state. For instance, its anti-inflammatory action arises from a strategic blockade of pro-inflammatory signals, such as nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and NLR family pyrin domain containing 3 (NLRP3), leading to a reduction in the excessive production of cytokines like tumor necrosis factor-alpha (TNF-α), interleukin 1 beta (IL-1β), and interleukin 6 (IL-6). This effect can be likened to the fine-tuning of an inflammatory thermostat, ensuring that inflammation required for tissue repair is preserved while preventing the collateral damage associated with chronic inflammation . In the antioxidant domain, curcumin stands out by activating nuclear factor erythroid 2-related factor 2 (NRF2), a regulatory protein that coordinates the expression of antioxidant enzymes, including superoxide dismutase (SOD), catalase, and glutathione peroxidase (GPX). This mechanism not only neutralizes reactive oxygen species (ROS) but also preserves mitochondrial metabolic functions, protecting ocular cells from oxidative stress. Thus, curcumin acts as a metabolic sentinel, preventing cumulative damage over time . Furthermore, its ability to inhibit vascular endothelial growth factor (VEGF) in angiogenic processes is particularly relevant in conditions characterized by pathological neovascularization, such as DR and age-related macular degeneration (AMD). This effect is not merely inhibitory but restorative, realigning angiogenic processes to meet the tissue’s physiological needs. Curcumin’s regulation of the apoptotic balance through modulators like B-cell lymphoma 2 (Bcl-2)/Bcl-2-associated X protein (Bax) ratio further solidifies its role as a cellular protector, preventing uncontrolled cell death while maintaining the selective elimination of damaged cells . Curcumin also exerts a significant antibacterial effect through its modulation of ROS and its ability to suppress bacterial cell wall synthesis. By neutralizing ROS and interfering with bacterial biofilm formation, curcumin enhances its antibacterial potential. This mechanism reduces bacterial load while maintaining a controlled inflammatory response. Through the modulation of the immune system, curcumin minimizes excessive inflammatory damage, contributing to a more effective defense mechanism against microbial infections . Finally, curcumin’s immunomodulatory effects ensure a favorable environment for tissue repair and regeneration, particularly in autoimmune and infectious diseases affecting the eye. This comprehensive action, combined with its extracellular matrix stabilization properties and regulation of endoplasmic reticulum stress, positions curcumin as a molecularly orchestrated intervention with therapeutic potential adaptable to a wide range of ocular conditions . 2.1. Retinal Diseases Retinal diseases, characterized by complex pathological processes, such as inflammation, oxidative stress, and pathological angiogenesis, are major contributors to vision impairment. Leveraging curcumin’s unique ability to target these mechanisms, recent research highlights its potential in mitigating retinal damage and preserving visual function . In DR, curcumin alleviates hyperglycemia-induced damage to retinal pigment epithelium (RPE) cells and helps maintain blood–retinal barrier integrity. A recent experimental study in diabetic rats showed that curcumin reduces pro-inflammatory cytokines (TNF-α, IL-1, and IFN-γ) and oxidative stress markers, such as malondialdehyde (MDA), GPX, CAT, and SOD . These findings align with its broader molecular actions, including modulation of extracellular signal-regulated kinase (ERK) and Akt (protein kinase B, PKB) pathways to protect RPE cells and inhibit retinal neovascularization, inhibition of chronic inflammation via NF-κB suppression, and oxidative stress reduction . In addition to DR, inflammation is a key driver in retinal diseases such as age-related macular degeneration (AMD) and best vitelliform macular dystrophy (BVMD). Curcumin suppresses pathways like NF-κB, reducing pro-inflammatory cytokines (TNF-α, IL-1β, and IL-6) and mitigating chronic inflammation . This anti-inflammatory effect, mediated through NF-κB modulation, is a common mechanism also observed in glaucoma, where curcumin regulates the NF-κB pathway to reduce inflammation and oxidative stress, helping protect retinal ganglion cells (RGCs) and improve the optic nerve integrity . Curcumin’s antioxidant effects are critical in combating oxidative stress, a major contributor to retinal degeneration. By enhancing endogenous antioxidant defenses, such as SOD and catalase, curcumin protects retinal cells from ROS, which are implicated in diseases like retinitis pigmentosa (RP), AMD, and BVMD. It also reduces oxidative stress and light-induced damage, especially relevant in BVMD . In all these conditions, curcumin also acts in glaucoma, where its ability to reduce ROS levels protects RGCs and modulates antioxidant pathways, such as the activation of Nrf2, which is critical for protecting cells against ROS-induced damage . Furthermore, curcumin may regulate calcium homeostasis in RPE cells, enhancing their function and survival, especially in BVMD . Neuroprotective effects of curcumin also extend to glaucoma, where it reduces RGC death by inhibiting caspase-3 activation and modulating factors such as Bcl-2 and Bax, providing additional protection against neuronal degeneration . Curcumin’s ability to modulate autophagy offers a dual benefit by not only preventing cellular degeneration but also enhancing the clearance of toxic protein aggregates and damaged organelles in RPE cells. This dual action reinforces its potential as a therapeutic agent for retinal diseases, highlighting its capacity to target multiple pathological mechanisms simultaneously, thereby preserving retinal structure and function . For wet AMD and other vascular-related conditions, PVR and retinal vascular obstruction (RVO), curcumin’s anti-angiogenic activity inhibits VEGF, preventing abnormal blood vessel formation and mitigating retinal damage . VEGF is essential for new blood vessel formation and vascular permeability, playing a key role in retinal diseases like DR, RVO, and exudative AMD. It is produced by retinal endothelial and pigment epithelial cells and is considered a key target in anti-angiogenic therapies. For instance, VEGF-A, a key driver in wet AMD progression, binds to VEGFR2, promoting angiogenesis and vascular leakage . Although glaucoma is not primarily an angiogenic disease, curcumin’s inhibition of VEGF may contribute to retinal vascular protection, especially in ischemic or injury contexts, with a positive impact on intraocular pressure regulation . Curcumin also exhibits neuroprotective and anti-apoptotic effects, reducing ganglion cell death and modulating apoptotic pathways. These effects are particularly significant in DR, AMD, and BVMD, where retinal cell survival is crucial for preserving vision . This mechanism is also present in glaucoma, where curcumin exerts similar effects to protect optic nerve cells from programmed cell death, an important feature of glaucoma pathogenesis . Curcumin regulates fibrosis, a hallmark of PVR and potentially BVMD, by inhibiting TGF-β1 activity and suppressing miR-21, a microRNA that promotes fibrogenesis, thereby further inhibiting the fibrotic process . This inhibition reduces the expression of fibrosis-related proteins, such as α-smooth muscle actin (α-SMA), type I collagen (COL1A1), and type III collagen (COL3A1), and potentially preserving retinal structure . In glaucoma, curcumin has shown a similar effect by modulating fibrotic processes associated with optic nerve injury through the regulation of TGF-β1 and other fibrosis-related molecular factors . Lastly, curcumin has demonstrated promising therapeutic potential in retinoblastoma (RB), the most common malignant intraocular tumor in children. It exerts anti-tumor effects by inhibiting cell proliferation, migration, and invasion while promoting apoptosis. These effects are primarily mediated through the upregulation of microRNA (miR-99a), which negatively regulates the Janus kinase/signal transducer and activator of transcription (JAK/STAT) signaling pathway, a crucial pathway involved in tumor progression and cell survival. This suggests that curcumin’s modulation of miRNA expression contributes significantly to its anti-cancer properties, making it a candidate for adjunct therapy in RB treatment . 2.2. Corneal Diseases The cornea faces distinct challenges, including infections, fibrosis, and inflammation, which can compromise its transparency and refractive function. In this context, curcumin’s ability to modulate oxidative stress, inflammation, and angiogenesis has shown promise in addressing these conditions . Inflammatory processes are central to keratitis and DED. Curcumin effectively inhibits p38 MAPK and NF-κB signaling, reducing pro-inflammatory cytokines like IL-1β, IL-6, and TNF-α. A study showed curcumin (5 μM) completely abolished hyperosmoticity-induced IL-1β elevation in human corneal epithelial cells . Oxidative stress exacerbates corneal injury and delays healing. Curcumin’s antioxidant capacity neutralizes ROS, protecting corneal cells and enhancing cellular survival under stress conditions. Guo et al. demonstrated that pretreatment with 12.5 µM curcumin enhances antioxidant defenses, including SOD1 and heme oxygenase-1, via the Keap1/Nrf2/ARE pathway, improving cell survival under oxidative stress . Curcumin also promotes corneal healing by stimulating cell migration and collagen synthesis, which reduces scarring and preserves corneal transparency. A dose-dependent effect was observed, where curcumin (10.0–12.5 mg/L) inhibited keratocyte proliferation and modulated fibrotic markers, upregulating decorin and CD90 (a glycoprotein marker of activated fibroblasts associated with tissue remodeling) while downregulating keratocan and aldehyde dehydrogenase . In corneal neovascularization, a severe complication often associated with corneal diseases, curcumin’s anti-angiogenic properties offer significant therapeutic advantages. By downregulating VEGF, curcumin impedes the formation of abnormal blood vessels, thereby reducing tissue damage and preserving vision. In an alkaline-burned rat model, the topical application of 40 μmol/L curcumin every 12 h for five days significantly reduced the area of new blood vessels compared to controls, showcasing its potential for managing angiogenesis-related corneal pathologies . 2.3. Bacterial Ocular Diseases Bacterial ocular infections, such as conjunctivitis, keratitis, and endophthalmitis, often involve severe inflammatory responses, biofilm formation, and resistance to conventional antibiotics. Curcumin’s molecular mechanisms address these challenges by targeting inflammation, oxidative stress, and bacterial survival strategies . Biofilm formation is a critical factor in bacterial persistence and resistance. Curcumin disrupts biofilm matrix integrity and inhibits bacterial efflux pumps, thereby increasing susceptibility to antibiotics. This effect is particularly relevant against multidrug-resistant (MDR) pathogens, including MRSA and Pseudomonas aeruginosa . Inflammation, a hallmark of bacterial ocular diseases, is modulated by curcumin through the inhibition of NF-κB and MAPK pathways, reducing cytokine storms and promoting ocular tissue recovery. In conjunctivitis, curcumin-based formulations like the product Haridra ® have shown anti-inflammatory and antibacterial efficacy . Oxidative stress exacerbates tissue damage during bacterial infections. Curcumin’s ability to neutralize ROS through the activation of the Keap1/Nrf2/ARE pathway preserves epithelial and retinal cell viability under stress conditions . Additionally, emerging research on the gut-ocular axis has opened new avenues for understanding how systemic factors, such as the gut microbiota, influence ocular immunity. Dietary interventions, including omega-3 fatty acids, carotenoids, and probiotics, have been shown to modulate both systemic and ocular immunity, reducing inflammation and improving overall eye health. Curcumin, with its anti-inflammatory and antioxidant effects, integrates into these strategies, offering a multifaceted approach to managing bacterial ocular diseases . 2.4. Periocular and Ocular Surface Disorders Chronic inflammation, immune dysregulation, and oxidative stress characterize eyelid diseases, such as blepharitis, blepharospasm, and eyelid dermatitis. Curcumin addresses these through its multi-targeted mechanisms, offering significant therapeutic potential . Curcumin suppresses NF-κB and TLR4 signaling pathways, reducing pro-inflammatory cytokines like TNF-α and IL-1β. Additionally, it inhibits inflammasome activity, specifically NRLP3 to reduce tissue damage in blepharitis and dermatitis. Its antioxidant activity neutralizes ROS, mitigating oxidative tissue damage, while modulation of Th17 cell activity reduces immune dysregulation in autoimmune eyelid conditions . In blepharospasm, curcumin exhibits neuroprotective effects by reducing neuroinflammation and oxidative stress, promoting cellular resilience . Additionally, curcumin demonstrates potential in managing allergic conjunctivitis, an inflammatory condition of the ocular surface driven by Th2 immune responses. Studies indicate that curcumin reduces IgE-mediated inflammation, suppressing eosinophilic infiltration and Th2 cytokine production, such as IL-4 and IL-5, in the conjunctiva . These actions extend curcumin’s therapeutic scope to inflammatory and immune-related ocular surface disorders. Further expanding its utility, curcumin shows promise in treating meibomian gland dysfunction (MGD), a leading cause of ocular discomfort, by reducing inflammation and improving the lipid composition of the tear film. In dry eye disease, which is often linked with MGD, curcumin alleviates inflammation and oxidative stress while enhancing mucin production, stabilizing the tear film, and improving ocular comfort . These combined actions position curcumin as a promising agent for managing complex periocular and ocular surface disorders. Retinal diseases, characterized by complex pathological processes, such as inflammation, oxidative stress, and pathological angiogenesis, are major contributors to vision impairment. Leveraging curcumin’s unique ability to target these mechanisms, recent research highlights its potential in mitigating retinal damage and preserving visual function . In DR, curcumin alleviates hyperglycemia-induced damage to retinal pigment epithelium (RPE) cells and helps maintain blood–retinal barrier integrity. A recent experimental study in diabetic rats showed that curcumin reduces pro-inflammatory cytokines (TNF-α, IL-1, and IFN-γ) and oxidative stress markers, such as malondialdehyde (MDA), GPX, CAT, and SOD . These findings align with its broader molecular actions, including modulation of extracellular signal-regulated kinase (ERK) and Akt (protein kinase B, PKB) pathways to protect RPE cells and inhibit retinal neovascularization, inhibition of chronic inflammation via NF-κB suppression, and oxidative stress reduction . In addition to DR, inflammation is a key driver in retinal diseases such as age-related macular degeneration (AMD) and best vitelliform macular dystrophy (BVMD). Curcumin suppresses pathways like NF-κB, reducing pro-inflammatory cytokines (TNF-α, IL-1β, and IL-6) and mitigating chronic inflammation . This anti-inflammatory effect, mediated through NF-κB modulation, is a common mechanism also observed in glaucoma, where curcumin regulates the NF-κB pathway to reduce inflammation and oxidative stress, helping protect retinal ganglion cells (RGCs) and improve the optic nerve integrity . Curcumin’s antioxidant effects are critical in combating oxidative stress, a major contributor to retinal degeneration. By enhancing endogenous antioxidant defenses, such as SOD and catalase, curcumin protects retinal cells from ROS, which are implicated in diseases like retinitis pigmentosa (RP), AMD, and BVMD. It also reduces oxidative stress and light-induced damage, especially relevant in BVMD . In all these conditions, curcumin also acts in glaucoma, where its ability to reduce ROS levels protects RGCs and modulates antioxidant pathways, such as the activation of Nrf2, which is critical for protecting cells against ROS-induced damage . Furthermore, curcumin may regulate calcium homeostasis in RPE cells, enhancing their function and survival, especially in BVMD . Neuroprotective effects of curcumin also extend to glaucoma, where it reduces RGC death by inhibiting caspase-3 activation and modulating factors such as Bcl-2 and Bax, providing additional protection against neuronal degeneration . Curcumin’s ability to modulate autophagy offers a dual benefit by not only preventing cellular degeneration but also enhancing the clearance of toxic protein aggregates and damaged organelles in RPE cells. This dual action reinforces its potential as a therapeutic agent for retinal diseases, highlighting its capacity to target multiple pathological mechanisms simultaneously, thereby preserving retinal structure and function . For wet AMD and other vascular-related conditions, PVR and retinal vascular obstruction (RVO), curcumin’s anti-angiogenic activity inhibits VEGF, preventing abnormal blood vessel formation and mitigating retinal damage . VEGF is essential for new blood vessel formation and vascular permeability, playing a key role in retinal diseases like DR, RVO, and exudative AMD. It is produced by retinal endothelial and pigment epithelial cells and is considered a key target in anti-angiogenic therapies. For instance, VEGF-A, a key driver in wet AMD progression, binds to VEGFR2, promoting angiogenesis and vascular leakage . Although glaucoma is not primarily an angiogenic disease, curcumin’s inhibition of VEGF may contribute to retinal vascular protection, especially in ischemic or injury contexts, with a positive impact on intraocular pressure regulation . Curcumin also exhibits neuroprotective and anti-apoptotic effects, reducing ganglion cell death and modulating apoptotic pathways. These effects are particularly significant in DR, AMD, and BVMD, where retinal cell survival is crucial for preserving vision . This mechanism is also present in glaucoma, where curcumin exerts similar effects to protect optic nerve cells from programmed cell death, an important feature of glaucoma pathogenesis . Curcumin regulates fibrosis, a hallmark of PVR and potentially BVMD, by inhibiting TGF-β1 activity and suppressing miR-21, a microRNA that promotes fibrogenesis, thereby further inhibiting the fibrotic process . This inhibition reduces the expression of fibrosis-related proteins, such as α-smooth muscle actin (α-SMA), type I collagen (COL1A1), and type III collagen (COL3A1), and potentially preserving retinal structure . In glaucoma, curcumin has shown a similar effect by modulating fibrotic processes associated with optic nerve injury through the regulation of TGF-β1 and other fibrosis-related molecular factors . Lastly, curcumin has demonstrated promising therapeutic potential in retinoblastoma (RB), the most common malignant intraocular tumor in children. It exerts anti-tumor effects by inhibiting cell proliferation, migration, and invasion while promoting apoptosis. These effects are primarily mediated through the upregulation of microRNA (miR-99a), which negatively regulates the Janus kinase/signal transducer and activator of transcription (JAK/STAT) signaling pathway, a crucial pathway involved in tumor progression and cell survival. This suggests that curcumin’s modulation of miRNA expression contributes significantly to its anti-cancer properties, making it a candidate for adjunct therapy in RB treatment . The cornea faces distinct challenges, including infections, fibrosis, and inflammation, which can compromise its transparency and refractive function. In this context, curcumin’s ability to modulate oxidative stress, inflammation, and angiogenesis has shown promise in addressing these conditions . Inflammatory processes are central to keratitis and DED. Curcumin effectively inhibits p38 MAPK and NF-κB signaling, reducing pro-inflammatory cytokines like IL-1β, IL-6, and TNF-α. A study showed curcumin (5 μM) completely abolished hyperosmoticity-induced IL-1β elevation in human corneal epithelial cells . Oxidative stress exacerbates corneal injury and delays healing. Curcumin’s antioxidant capacity neutralizes ROS, protecting corneal cells and enhancing cellular survival under stress conditions. Guo et al. demonstrated that pretreatment with 12.5 µM curcumin enhances antioxidant defenses, including SOD1 and heme oxygenase-1, via the Keap1/Nrf2/ARE pathway, improving cell survival under oxidative stress . Curcumin also promotes corneal healing by stimulating cell migration and collagen synthesis, which reduces scarring and preserves corneal transparency. A dose-dependent effect was observed, where curcumin (10.0–12.5 mg/L) inhibited keratocyte proliferation and modulated fibrotic markers, upregulating decorin and CD90 (a glycoprotein marker of activated fibroblasts associated with tissue remodeling) while downregulating keratocan and aldehyde dehydrogenase . In corneal neovascularization, a severe complication often associated with corneal diseases, curcumin’s anti-angiogenic properties offer significant therapeutic advantages. By downregulating VEGF, curcumin impedes the formation of abnormal blood vessels, thereby reducing tissue damage and preserving vision. In an alkaline-burned rat model, the topical application of 40 μmol/L curcumin every 12 h for five days significantly reduced the area of new blood vessels compared to controls, showcasing its potential for managing angiogenesis-related corneal pathologies . Bacterial ocular infections, such as conjunctivitis, keratitis, and endophthalmitis, often involve severe inflammatory responses, biofilm formation, and resistance to conventional antibiotics. Curcumin’s molecular mechanisms address these challenges by targeting inflammation, oxidative stress, and bacterial survival strategies . Biofilm formation is a critical factor in bacterial persistence and resistance. Curcumin disrupts biofilm matrix integrity and inhibits bacterial efflux pumps, thereby increasing susceptibility to antibiotics. This effect is particularly relevant against multidrug-resistant (MDR) pathogens, including MRSA and Pseudomonas aeruginosa . Inflammation, a hallmark of bacterial ocular diseases, is modulated by curcumin through the inhibition of NF-κB and MAPK pathways, reducing cytokine storms and promoting ocular tissue recovery. In conjunctivitis, curcumin-based formulations like the product Haridra ® have shown anti-inflammatory and antibacterial efficacy . Oxidative stress exacerbates tissue damage during bacterial infections. Curcumin’s ability to neutralize ROS through the activation of the Keap1/Nrf2/ARE pathway preserves epithelial and retinal cell viability under stress conditions . Additionally, emerging research on the gut-ocular axis has opened new avenues for understanding how systemic factors, such as the gut microbiota, influence ocular immunity. Dietary interventions, including omega-3 fatty acids, carotenoids, and probiotics, have been shown to modulate both systemic and ocular immunity, reducing inflammation and improving overall eye health. Curcumin, with its anti-inflammatory and antioxidant effects, integrates into these strategies, offering a multifaceted approach to managing bacterial ocular diseases . Chronic inflammation, immune dysregulation, and oxidative stress characterize eyelid diseases, such as blepharitis, blepharospasm, and eyelid dermatitis. Curcumin addresses these through its multi-targeted mechanisms, offering significant therapeutic potential . Curcumin suppresses NF-κB and TLR4 signaling pathways, reducing pro-inflammatory cytokines like TNF-α and IL-1β. Additionally, it inhibits inflammasome activity, specifically NRLP3 to reduce tissue damage in blepharitis and dermatitis. Its antioxidant activity neutralizes ROS, mitigating oxidative tissue damage, while modulation of Th17 cell activity reduces immune dysregulation in autoimmune eyelid conditions . In blepharospasm, curcumin exhibits neuroprotective effects by reducing neuroinflammation and oxidative stress, promoting cellular resilience . Additionally, curcumin demonstrates potential in managing allergic conjunctivitis, an inflammatory condition of the ocular surface driven by Th2 immune responses. Studies indicate that curcumin reduces IgE-mediated inflammation, suppressing eosinophilic infiltration and Th2 cytokine production, such as IL-4 and IL-5, in the conjunctiva . These actions extend curcumin’s therapeutic scope to inflammatory and immune-related ocular surface disorders. Further expanding its utility, curcumin shows promise in treating meibomian gland dysfunction (MGD), a leading cause of ocular discomfort, by reducing inflammation and improving the lipid composition of the tear film. In dry eye disease, which is often linked with MGD, curcumin alleviates inflammation and oxidative stress while enhancing mucin production, stabilizing the tear film, and improving ocular comfort . These combined actions position curcumin as a promising agent for managing complex periocular and ocular surface disorders. Curcumin predominantly exists in the keto-enol form in polar solvents and is hydrophobic, being insoluble in water but soluble in organic solvents. This characteristic poses challenges for its therapeutic application. However, advanced drug delivery systems, such as encapsulation, have significantly improved curcumin’s solubility and stability in aqueous environments . 3.1. Non-Ionic Surfactant-Based Delivery Systems Non-ionic surfactants enhance the bioavailability of hydrophobic drugs, especially curcumin, by forming stable and biocompatible carriers. These systems improve curcumin’s solubility, stability, and ocular residence time, thereby reducing systemic toxicity and enhancing therapeutic outcomes . Composed of non-ionic surfactants and cholesterol, niosomes enhance curcumin’s solubility, stability, and ocular residence time, reducing systemic toxicity . Nanoemulsion-based formulations improve drug dispersion in aqueous environments, enhancing curcumin’s corneal permeability and bioavailability. This approach ensures rapid therapeutic action, crucial for treating acute ocular conditions . Liposomes, a subclass of non-ionic surfactant carriers, enhance curcumin’s bioavailability and ocular distribution, supporting its use in retinal degeneration and inflammation . In addition, curcumin-loaded proniosomal gels have emerged as a promising alternative to traditional corticosteroid treatments for ocular inflammation. These formulations, composed of surfactants, such as cremophore RH, lecithin, and cholesterol, offer high encapsulation efficiency (96%) and 3.22-fold greater permeability than conventional dispersions. Comparative in vivo studies demonstrated that these proniosomal gels significantly reduced inflammatory symptoms, achieving complete recovery within four days, comparable to corticosteroid drops. Curcumin’s natural origin minimizes adverse effects, such as intraocular pressure elevation, positioning it as a safer yet equally effective anti-inflammatory treatment option. Curcumin encapsulated in proniosomal gel demonstrated exceptional biocompatibility and safety, coupled with potent anti-inflammatory properties. This innovative formulation enhanced ocular retention and significantly improved corneal permeability, delivering a sustained release effect over 24 h . Recent research underscores the value of transferosomes (TFSs), ultra-deformable vesicles comprising lipid bilayers and surfactants, such as Tween 80. TFS exhibits exceptional drug entrapment efficiency (>99%) and enhances drug penetration across corneal and conjunctival barriers, enabling deeper and more effective delivery to ocular tissues. They also improve precorneal retention, ensuring sustained therapeutic levels and superior bioavailability. Studies highlight their excellent compatibility with ocular tissues and their potential to optimize the delivery of curcumin and similar compounds in treating eye diseases . Additionally, lipophilic vehicles, such as medium-chain triglycerides (MCTs) and squalane, demonstrate varying degrees of effectiveness. Ex vivo studies reveal that squalane suspensions notably enhance curcumin’s penetration into ocular tissues compared to MCT solutions, emphasizing the role of vehicle partitioning in optimizing drug delivery . Polymeric in situ gelling systems represent an innovative approach for ocular curcumin delivery. These inserts, composed of biocompatible polymers such as HPMC, CMC, and Pluronic F127, provide sustained drug release and enhanced mucoadhesion. Characterization studies show that curcumin in these systems is dispersed molecularly, with smooth and uniform surfaces. Importantly, the inserts exhibit superior corneal permeation (5.4- to 8.86-fold increase) and retention times compared to conventional suspensions. These properties underscore their potential to replace traditional eye drops, offering improved therapeutic efficacy through prolonged action and reduced dosing frequency . 3.2. Nanoparticle-Based Systems Nanoparticles provide controlled drug release and enhanced stability for curcumin, making them suitable for various ocular conditions . Solid lipid nanoparticles ensure high drug loading, stability, and minimal systemic toxicity. They have demonstrated effectiveness in reducing oxidative stress in retinal cells, addressing posterior segment eye diseases, such as AMD and DR . Polymeric nanoparticles, such as PLGA(polylactic-co-glycolic acid)-based nanoparticles, enhance curcumin’s stability and bioavailability. An innovative approach involves biodegradable scleral plugs, which enable sustained drug release for up to 14 days. Studies show that scleral plugs with curcumin concentrations of 0.5 mg, 1.0 mg, and 1.5 mg achieve therapeutic levels (above 15 µg/mL) in vitro, with no adverse effects observed in vivo models, such as changes in intraocular pressure or retinal integrity. These findings underscore the safety and efficacy of scleral plugs for posterior ocular diseases . A promising strategy to address curcumin’s limitations involves using diphosphorylated curcumin (Cur-2p), a prodrug that generates curcumin nanoparticles in situ. This approach enhances curcumin’s stability and reduces aggregation in water. Upon enzymatic conversion by alkaline phosphatase (ALP) in cancer cells, Cur-2p exhibits selective cytotoxicity against ALP-overexpressing cancer cells while sparing normal cells. Additionally, intravitreal injections of Cur-2p demonstrate superior intraocular biocompatibility, preserving retinal morphology and function. In a rodent model of uveitis, Cur-2p effectively suppresses inflammation, outperforming unmodified curcumin. These findings highlight Cur-2p’s potential as a next-generation nanoparticle-based system for ocular drug delivery . Mixed micelle in situ gels, composed of Pluronic P123 and D-α-tocopherol polyethylene glycol succinate, have been developed to overcome curcumin’s poor water solubility and limited corneal permeability. The optimized micellar formulations, when combined with gellan gum, form transparent in situ gels that sustain drug release while enhancing corneal retention. In vitro studies confirmed a sustained drug release profile, while ex vivo corneal permeation tests demonstrated superior drug delivery, with cumulative drug permeation up to 1.32 times higher compared to standard curcumin solutions . Similarly, curcumin-loaded mixed micelle in situ gel (Cur-MM-ISG) improves ocular drug delivery. The system combines small, stable micelles with gellan gum to form a transparent gel upon application. Compared to free curcumin, Cur-MM-ISG significantly enhanced corneal permeation and retention time without causing irritation, highlighting its potential for sustained and efficient ocular therapy. This system supports sustained therapeutic effects with excellent ocular tolerance . Thermosensitive gels (CUR-CNLC-GEL) formulation demonstrated promising results in improving the bioavailability of curcumin for ocular therapy. With a sol-gel transition temperature of 34 °C, it ensured practical application. The nanogel exhibited 1.56-fold higher permeability and a 9.24-fold increase in bioavailability (AUC 0→∞ ) compared to a conventional curcumin solution. Additionally, the enhanced C max and prolonged mean residence time (MRT) indicated effective controlled release and retention properties. These attributes the position CUR-CNLC-GEL as a highly promising candidate for next-generation ocular drug delivery, offering superior corneal permeation and extended therapeutic efficacy . Thermosensitive hydrogels, especially when paired with nanoparticles, offer an innovative approach to achieving sustained drug release. For instance, hydrogels incorporating curcumin-loaded nanoparticles (CUR-NPs) and latanoprost have demonstrated notable improvements in bioavailability for glaucoma therapy. This dual-drug delivery system addresses oxidative stress in trabecular meshwork cells, effectively mitigating inflammation, reducing mitochondrial ROS production, and decreasing apoptosis levels. Additionally, it enhances both uveoscleral and trabecular outflow, highlighting its potential as a promising treatment for glaucoma. This system enables the development of topical eye drops capable of sustaining drug release for up to 7 days, enhancing residence time in the rabbit eye, and improving corneal permeation with minimal toxicity . The study on the development of a thermoresponsive ophthalmic in situ gel containing curcumin-loaded albumin nanoparticles (Cur-BSA-NPs-Gel) presents significant advancements in ocular drug delivery systems. By optimizing the formulation through a central composite design, the researchers achieved a gel that transitions from liquid to semi-solid under physiological conditions, ensuring easy application and sustained drug release. Incorporating albumin nanoparticles minimally impacted the gel’s structure while enhancing curcumin’s bioavailability in the aqueous humor, as confirmed by in vivo studies in rabbit models. The formulation demonstrated safety for ophthalmic use, with no signs of eye irritation, and offers potential for prolonged therapeutic effects, making it a promising candidate for ocular treatments . Recent advancements have introduced dissolvable hybrid microneedles (MNs) patches as a novel method for ocular delivery of curcumin. These patches incorporate curcumin-loaded polymeric micelles into a hyaluronic acid matrix, using a micromolding process to ensure efficient drug dispersion. Studies reveal that this system facilitates sustained drug release over eight hours and extends pre-corneal retention to more than 3.5 h, significantly improving bioavailability. MNs patch can create temporary microchannels in the corneal epithelium, enhancing permeability. In vivo testing demonstrated its superior efficacy in treating endotoxin-induced uveitis, reducing inflammatory cell infiltration more effectively than conventional eye drops, making it a promising tool for managing intraocular inflammatory disorders . The development of curcumin-loaded nanostructured lipid carriers (CUR-NLC) coated with thiolated chitosan (CS-NAC) offers a promising solution for topical ocular drug delivery. This innovative system achieves sustained drug release for up to 72 h and significantly enhances corneal permeability and retention. Compared to coatings using chitosan oligosaccharides (COS) and carboxymethyl chitosan (CMCS), the CS-NAC coating demonstrated superior performance, with permeability coefficients increasing by 6.4 and 18.8 times relative to uncoated CUR-NLC and conventional eye drops, respectively. Furthermore, ocular irritation tests confirmed the biocompatibility of CS-NAC-CUR-NLC . The formulation of curcumin-loaded nanostructured lipid carriers (NLCs) using hot-melt emulsification and ultrasonication has demonstrated significant potential for ocular drug delivery. Optimized through a central composite design, the resulting NLCs showcased a particle size of approximately 66.8 nm, high encapsulation efficiency (96%), and consistent drug loading. These properties contributed to enhanced stability over three months at low temperatures and superior transcorneal permeability. Ex vivo tests revealed a 2.5-fold increase in curcumin permeation across rabbit corneas compared to standard formulations, without evidence of adverse effects, underscoring the NLCs’ ability to improve drug delivery efficiency while maintaining safety . The development of a nanomicelle-based curcumin formulation utilizing a PVCL-PVA-PEG graft copolymer has shown promise for enhancing ocular drug delivery. This system significantly improves curcumin’s solubility and stability while offering robust antioxidant activity. In vitro and in vivo studies demonstrated that these nanomicelles enhance cellular uptake and corneal permeation compared to free curcumin solutions. Additionally, the formulation exhibited excellent ocular tolerance, with no signs of irritation in rabbit models. These results suggest that nanomicelles could serve as an effective platform for delivering curcumin topically in the treatment of ocular inflammation and related conditions . Recent advancements in curcumin-based therapies for AMD highlight the promise of aqueous nanomicellar formulations (CUR-NMF). This innovative approach overcomes curcumin’s poor aqueous solubility, a key limitation for its therapeutic use. CUR-NMF, developed using hydrogenated castor oil (HCO-40) and octoxynol-40 (OC-40), offers a stable delivery system optimized for retinal protection. Studies demonstrate its antioxidant, anti-inflammatory, and anti-angiogenic effects, showing significant protection against oxidative stress in retinal cells and a reduction in VEGF release, a critical factor in AMD pathology. Furthermore, sustained drug release profiles and favorable safety assessments suggest CUR-NMF could provide long-term therapeutic benefits for both dry and wet AMD . 3.3. Cyclodextrin Complexes Cyclodextrins (CDs) play a crucial role in enhancing curcumin’s solubility, stability, and bioavailability through the formation of inclusion complexes . These complexes significantly improve curcumin’s therapeutic effectiveness in both anterior and posterior ocular diseases. Among the CDs, β-cyclodextrins (βCD) and γ-cyclodextrins (γCD) stand out due to their higher solubility, capacity to self-assemble into nanoaggregates, and favorable safety profile for ocular applications . Since the mid-1990s, CD-based inclusion complexes have provided significant technological advantages for pharmaceutical formulations. By enhancing the stability, solubility, and bioavailability of bioactive compounds, these complexes address key challenges in drug delivery. A notable example is the use of cyclodextrins to improve the solubility and stability of chloramphenicol, a patented formulation still in use today . This approach enabled the development of stable pharmaceutical solutions and optimized separation processes, underscoring the enduring relevance of this technology . Methyl-β-cyclodextrin (M-β-CD) is the cyclodextrin used in commercial eye drop formulations, including those containing chloramphenicol. Hydroxypropyl-β-cyclodextrin (HP-β-CD) is also covered by this patent for chloramphenicol due to its superior solubility and biocompatibility . More recently, CD-based inclusion complexes have been explored for cannabinoid delivery, demonstrating promising potential in pain management and anti-inflammatory therapies. These developments highlight the versatility of cyclodextrins in modern drug delivery, particularly for ocular applications, where solubility, stability, and bioavailability are critical for therapeutic success . Modified CDs, such as ethylene diamine (EDA)-modified βCD, have demonstrated superior capabilities in improving curcumin (CUR) solubility and stability. These complexes provide enhanced thermodynamic properties, making CUR more bioavailable for ocular applications. Curcumin-EDA-βCD nanoparticles exhibit excellent corneal permeability, as shown in vitro porcine cornea experiments, and maintain high biocompatibility, confirmed by histological analyses of porcine corneas and bovine corneal epithelial cell viability. These properties make them particularly suitable for addressing anterior segment diseases like keratitis and dry eye disease . In a recent study, various CD-curcumin complexes were prepared and characterized, showing significant improvements in solubility. The freeze-drying method produced highly soluble complexes, and the optimal formulation provided sustained release for over 96 h. This approach offers a promising solution for curcumin’s use in ocular therapies, such as eye drops for conditions like retinitis pigmentosa . Inclusion complexes of CUR with hydroxypropyl-β-cyclodextrin (HP-βCD) were successfully developed using the cosolvency/lyophilization method, resulting in significant improvements in CUR solubility, stability, and therapeutic efficacy. The complexes demonstrated superior antioxidant and anti-inflammatory activities compared to free CUR. To facilitate ocular administration, an in-situ gel system was prepared using Pluronic F127 and chitosan, providing mucoadhesion and sol-gel transition between 26–35 °C. Viscosity, pH, and clarity tests confirmed the system’s suitability for ocular application. In vitro release studies showed sustained drug release for 6 h, fitting the Weibull kinetic model. This approach offers a promising drug delivery strategy for ocular diseases, supporting prolonged and controlled drug release . The poor solubility and stability of CUR limit its application in ocular drug delivery. To address this, CUR was complexed with βCD and HP-βCD using co-solvent, sonication, and freeze-drying methods in 1:1 and 1:2 molar ratios. The freeze-drying method produced the most water-soluble complexes. Among the 12 tested formulations, the F11 formulation, prepared with pH 6.8 phosphate buffer containing 1% Tween 80, demonstrated sustained drug release for over 96 h. The drug release followed a Higuchi non-Fickian diffusion model. These findings suggest that F11 could be developed as a once-daily eye drop formulation, offering a promising approach for the sustained delivery of curcumin in the treatment of ocular diseases, such as retinitis pigmentosa . The use of CDs to optimize corneal penetration of CUR has shown promising results. In ex vivo models using porcine corneas, the combination of CDs with nanoparticles demonstrated greater drug permeation. This improvement is attributed to the ability of CDs to form inclusion complexes, enhancing curcumin’s solubility and stability, while nanoparticles enable sustained release and protection against enzymatic degradation . Recent studies have demonstrated that curcumin-loaded hydrogels, such as those incorporating CUR nanoparticles encapsulated with βCD and hyaluronic acid, accelerate corneal healing in ulcerative keratitis. This system not only improves corneal clarity and reduces inflammation but also enhances the quality of healed tissues, requiring fewer applications compared to conventional treatments. These formulations hold promises for future therapeutic use in treating ulcerative keratitis and other ocular conditions, providing an innovative, herbal-based alternative to traditional treatments . The penetration of CUR into the cornea was evaluated using an ex vivo porcine eye model and a digital image analysis technique. Several formulation strategies, including oily solutions, oily suspensions, micelles, liposomes, nanosuspensions, and CD complexes, were explored to improve CUR corneal permeability. The results revealed that cyclodextrin-based formulations exhibited superior corneal penetration compared to other delivery systems. The image analysis approach effectively measured CUR penetration into corneal tissues, supporting the potential of cyclodextrin complexes as a delivery strategy for hydrophobic drugs in ocular applications. This technique offers a novel approach for optimizing the penetration of CUR and similar compounds into the cornea . Among the most effective formulations are those based on modified βCDs and conjugates with tetrahydrocurcumin nanoparticles, which exhibited deeper penetration into ocular tissues. These strategies hold significant potential for treating both anterior and posterior segment ocular diseases, as they increase bioavailability and extend therapeutic effects . γCD-based nanoparticles not only enhance drug permeation but also increase retention time on the ocular surface, promoting sustained drug release and reducing the frequency of administration. Additionally, the presence of tear enzymes like α-amylase facilitates drug release from γCD complexes, further boosting bioavailability. For instance, γCD-based eye drops containing dexamethasone achieved higher concentrations in ocular tissues compared to commercial formulations. Moreover, these formulations were well tolerated, with no significant ocular irritation or toxicity observed. γCD has also been employed in formulations for dorzolamide, telmisartan, and nepafenac, demonstrating improved pharmacokinetics and sustained drug release for up to 24 h . Inclusion complexes of CUR with HP-βCD enhance curcumin’s solubility, dissolution rate, and bioavailability, essential for ocular drug delivery. Studies using co-evaporation methods revealed a 1:1 molar ratio complex with a solubility constant of 30.09 mM −1 . Characterization techniques, such as XRD, confirmed the loss of curcumin’s crystalline structure, while FTIR and DTA indicated no chemical interactions. In vitro dissolution tests showed faster release of CUR from the complex compared to its pure form and physical mixtures. This approach improves curcumin’s bioavailability, making it a promising strategy for ocular drug delivery systems . The ocular delivery of CUR faces significant barriers due to anatomical and physiological constraints; however, advances in nanoengineered systems have shown promising results. The formation of inclusion complexes with HP-CDs through spray-drying significantly enhanced the solubility, permeability, and stability of CUR. Enhanced corneal and retinal permeability was observed, along with increased antioxidant activity in ocular epithelial cells, including upregulation of SOD1, CAT1, and HMOX1. Moreover, protection against oxidative stress was confirmed in rabbit corneal tissues. These findings highlight the potential of CUR:HP-CD complexes to improve ocular drug bioavailability, thereby enhancing therapeutic outcomes for ocular diseases . Cyclodextrin-based systems significantly enhance curcumin’s bioavailability, solubility, and therapeutic potential for ocular drug delivery. Advances in βCD, γCD, and HP-βCD systems, combined with nanoparticles or in situ gels, have demonstrated improved drug permeation, sustained release, and higher bioactivity. These strategies support the development of more effective ophthalmic treatments. 3.4. Drug Delivery for Antibacterial Agents Nanocomposites, such as cupriferous hollow nanoshells, combine silver and copper ions. These materials exhibit dual functionality: silver ions provide potent antibacterial activity, while copper ions promote tissue regeneration by stimulating fibroblast migration and angiogenesis. This dual approach is particularly beneficial in treating conditions like keratitis, where infections can impair spontaneous recovery and cause corneal damage. Nanocomposite-based treatments not only target the bacteria but also support the healing of damaged tissues, offering a comprehensive approach to managing complex infections . Moreover, curcumin enhances traditional antibiotics by inhibiting bacterial efflux pumps and disrupting biofilms—critical mechanisms in antibiotic resistance. When combined with biopolymers like chitosan, curcumin has shown enhanced antibacterial effects, even at low concentrations, especially against resistant bacterial strains. This makes curcumin-based formulations a valuable tool in combating antibiotic-resistant ocular infections, such as conjunctivitis and keratitis . Curcumin-based formulations have also demonstrated significant efficacy in the treatment of conjunctivitis. Products like Haridra ® and Ophthacare ® have been shown to combat pathogens like Escherichia coli , Staphylococcus aureus , Klebsiella pneumoniae , and Pseudomonas aeruginosa while also reducing inflammation and irritation. These formulations address not only the infection but also the underlying inflammation, providing a comprehensive treatment approach. Ophthacare ® , which combines Curcuma longa with other herbal extracts, offers an effective solution for a range of ocular conditions, including dry eye and inflammatory conjunctival disorders . Endophthalmitis, an intraocular infection characterized by extensive inflammation and retinal damage, has benefited from nanotechnology-based drug delivery systems. Hybrid frameworks that incorporate silver nanoparticles and photosensitizers have been developed to disrupt biofilms while preserving host tissues. These systems, combined with curcumin’s anti-inflammatory properties, can modulate cytokine storms, support retinal cell survival, and preserve ocular structures. This combination not only targets the infection but also helps to protect the delicate retinal tissues, improving patient outcomes . Recent innovations in drug delivery systems have further amplified the therapeutic potential of curcumin. Nanoparticles and liposomes are particularly effective at enhancing curcumin’s bioavailability and ocular penetration, ensuring sustained therapeutic effects. For example, dual-drug nanofibers, which combine curcumin with antibiotics, have shown enhanced bactericidal activity and accelerated tissue regeneration in preclinical models of ocular infections. These advanced delivery systems ensure that curcumin reaches the target site effectively, offering continuous antimicrobial action and supporting tissue healing . These systems not only enhance the effectiveness of conventional antibiotics but also provide innovative solutions to overcome the challenges posed by MDR bacteria, biofilms, and tissue damage . The integration of curcumin in these systems adds a further layer of therapeutic benefit, making it a promising tool in the management of ocular infections. Non-ionic surfactants enhance the bioavailability of hydrophobic drugs, especially curcumin, by forming stable and biocompatible carriers. These systems improve curcumin’s solubility, stability, and ocular residence time, thereby reducing systemic toxicity and enhancing therapeutic outcomes . Composed of non-ionic surfactants and cholesterol, niosomes enhance curcumin’s solubility, stability, and ocular residence time, reducing systemic toxicity . Nanoemulsion-based formulations improve drug dispersion in aqueous environments, enhancing curcumin’s corneal permeability and bioavailability. This approach ensures rapid therapeutic action, crucial for treating acute ocular conditions . Liposomes, a subclass of non-ionic surfactant carriers, enhance curcumin’s bioavailability and ocular distribution, supporting its use in retinal degeneration and inflammation . In addition, curcumin-loaded proniosomal gels have emerged as a promising alternative to traditional corticosteroid treatments for ocular inflammation. These formulations, composed of surfactants, such as cremophore RH, lecithin, and cholesterol, offer high encapsulation efficiency (96%) and 3.22-fold greater permeability than conventional dispersions. Comparative in vivo studies demonstrated that these proniosomal gels significantly reduced inflammatory symptoms, achieving complete recovery within four days, comparable to corticosteroid drops. Curcumin’s natural origin minimizes adverse effects, such as intraocular pressure elevation, positioning it as a safer yet equally effective anti-inflammatory treatment option. Curcumin encapsulated in proniosomal gel demonstrated exceptional biocompatibility and safety, coupled with potent anti-inflammatory properties. This innovative formulation enhanced ocular retention and significantly improved corneal permeability, delivering a sustained release effect over 24 h . Recent research underscores the value of transferosomes (TFSs), ultra-deformable vesicles comprising lipid bilayers and surfactants, such as Tween 80. TFS exhibits exceptional drug entrapment efficiency (>99%) and enhances drug penetration across corneal and conjunctival barriers, enabling deeper and more effective delivery to ocular tissues. They also improve precorneal retention, ensuring sustained therapeutic levels and superior bioavailability. Studies highlight their excellent compatibility with ocular tissues and their potential to optimize the delivery of curcumin and similar compounds in treating eye diseases . Additionally, lipophilic vehicles, such as medium-chain triglycerides (MCTs) and squalane, demonstrate varying degrees of effectiveness. Ex vivo studies reveal that squalane suspensions notably enhance curcumin’s penetration into ocular tissues compared to MCT solutions, emphasizing the role of vehicle partitioning in optimizing drug delivery . Polymeric in situ gelling systems represent an innovative approach for ocular curcumin delivery. These inserts, composed of biocompatible polymers such as HPMC, CMC, and Pluronic F127, provide sustained drug release and enhanced mucoadhesion. Characterization studies show that curcumin in these systems is dispersed molecularly, with smooth and uniform surfaces. Importantly, the inserts exhibit superior corneal permeation (5.4- to 8.86-fold increase) and retention times compared to conventional suspensions. These properties underscore their potential to replace traditional eye drops, offering improved therapeutic efficacy through prolonged action and reduced dosing frequency . Nanoparticles provide controlled drug release and enhanced stability for curcumin, making them suitable for various ocular conditions . Solid lipid nanoparticles ensure high drug loading, stability, and minimal systemic toxicity. They have demonstrated effectiveness in reducing oxidative stress in retinal cells, addressing posterior segment eye diseases, such as AMD and DR . Polymeric nanoparticles, such as PLGA(polylactic-co-glycolic acid)-based nanoparticles, enhance curcumin’s stability and bioavailability. An innovative approach involves biodegradable scleral plugs, which enable sustained drug release for up to 14 days. Studies show that scleral plugs with curcumin concentrations of 0.5 mg, 1.0 mg, and 1.5 mg achieve therapeutic levels (above 15 µg/mL) in vitro, with no adverse effects observed in vivo models, such as changes in intraocular pressure or retinal integrity. These findings underscore the safety and efficacy of scleral plugs for posterior ocular diseases . A promising strategy to address curcumin’s limitations involves using diphosphorylated curcumin (Cur-2p), a prodrug that generates curcumin nanoparticles in situ. This approach enhances curcumin’s stability and reduces aggregation in water. Upon enzymatic conversion by alkaline phosphatase (ALP) in cancer cells, Cur-2p exhibits selective cytotoxicity against ALP-overexpressing cancer cells while sparing normal cells. Additionally, intravitreal injections of Cur-2p demonstrate superior intraocular biocompatibility, preserving retinal morphology and function. In a rodent model of uveitis, Cur-2p effectively suppresses inflammation, outperforming unmodified curcumin. These findings highlight Cur-2p’s potential as a next-generation nanoparticle-based system for ocular drug delivery . Mixed micelle in situ gels, composed of Pluronic P123 and D-α-tocopherol polyethylene glycol succinate, have been developed to overcome curcumin’s poor water solubility and limited corneal permeability. The optimized micellar formulations, when combined with gellan gum, form transparent in situ gels that sustain drug release while enhancing corneal retention. In vitro studies confirmed a sustained drug release profile, while ex vivo corneal permeation tests demonstrated superior drug delivery, with cumulative drug permeation up to 1.32 times higher compared to standard curcumin solutions . Similarly, curcumin-loaded mixed micelle in situ gel (Cur-MM-ISG) improves ocular drug delivery. The system combines small, stable micelles with gellan gum to form a transparent gel upon application. Compared to free curcumin, Cur-MM-ISG significantly enhanced corneal permeation and retention time without causing irritation, highlighting its potential for sustained and efficient ocular therapy. This system supports sustained therapeutic effects with excellent ocular tolerance . Thermosensitive gels (CUR-CNLC-GEL) formulation demonstrated promising results in improving the bioavailability of curcumin for ocular therapy. With a sol-gel transition temperature of 34 °C, it ensured practical application. The nanogel exhibited 1.56-fold higher permeability and a 9.24-fold increase in bioavailability (AUC 0→∞ ) compared to a conventional curcumin solution. Additionally, the enhanced C max and prolonged mean residence time (MRT) indicated effective controlled release and retention properties. These attributes the position CUR-CNLC-GEL as a highly promising candidate for next-generation ocular drug delivery, offering superior corneal permeation and extended therapeutic efficacy . Thermosensitive hydrogels, especially when paired with nanoparticles, offer an innovative approach to achieving sustained drug release. For instance, hydrogels incorporating curcumin-loaded nanoparticles (CUR-NPs) and latanoprost have demonstrated notable improvements in bioavailability for glaucoma therapy. This dual-drug delivery system addresses oxidative stress in trabecular meshwork cells, effectively mitigating inflammation, reducing mitochondrial ROS production, and decreasing apoptosis levels. Additionally, it enhances both uveoscleral and trabecular outflow, highlighting its potential as a promising treatment for glaucoma. This system enables the development of topical eye drops capable of sustaining drug release for up to 7 days, enhancing residence time in the rabbit eye, and improving corneal permeation with minimal toxicity . The study on the development of a thermoresponsive ophthalmic in situ gel containing curcumin-loaded albumin nanoparticles (Cur-BSA-NPs-Gel) presents significant advancements in ocular drug delivery systems. By optimizing the formulation through a central composite design, the researchers achieved a gel that transitions from liquid to semi-solid under physiological conditions, ensuring easy application and sustained drug release. Incorporating albumin nanoparticles minimally impacted the gel’s structure while enhancing curcumin’s bioavailability in the aqueous humor, as confirmed by in vivo studies in rabbit models. The formulation demonstrated safety for ophthalmic use, with no signs of eye irritation, and offers potential for prolonged therapeutic effects, making it a promising candidate for ocular treatments . Recent advancements have introduced dissolvable hybrid microneedles (MNs) patches as a novel method for ocular delivery of curcumin. These patches incorporate curcumin-loaded polymeric micelles into a hyaluronic acid matrix, using a micromolding process to ensure efficient drug dispersion. Studies reveal that this system facilitates sustained drug release over eight hours and extends pre-corneal retention to more than 3.5 h, significantly improving bioavailability. MNs patch can create temporary microchannels in the corneal epithelium, enhancing permeability. In vivo testing demonstrated its superior efficacy in treating endotoxin-induced uveitis, reducing inflammatory cell infiltration more effectively than conventional eye drops, making it a promising tool for managing intraocular inflammatory disorders . The development of curcumin-loaded nanostructured lipid carriers (CUR-NLC) coated with thiolated chitosan (CS-NAC) offers a promising solution for topical ocular drug delivery. This innovative system achieves sustained drug release for up to 72 h and significantly enhances corneal permeability and retention. Compared to coatings using chitosan oligosaccharides (COS) and carboxymethyl chitosan (CMCS), the CS-NAC coating demonstrated superior performance, with permeability coefficients increasing by 6.4 and 18.8 times relative to uncoated CUR-NLC and conventional eye drops, respectively. Furthermore, ocular irritation tests confirmed the biocompatibility of CS-NAC-CUR-NLC . The formulation of curcumin-loaded nanostructured lipid carriers (NLCs) using hot-melt emulsification and ultrasonication has demonstrated significant potential for ocular drug delivery. Optimized through a central composite design, the resulting NLCs showcased a particle size of approximately 66.8 nm, high encapsulation efficiency (96%), and consistent drug loading. These properties contributed to enhanced stability over three months at low temperatures and superior transcorneal permeability. Ex vivo tests revealed a 2.5-fold increase in curcumin permeation across rabbit corneas compared to standard formulations, without evidence of adverse effects, underscoring the NLCs’ ability to improve drug delivery efficiency while maintaining safety . The development of a nanomicelle-based curcumin formulation utilizing a PVCL-PVA-PEG graft copolymer has shown promise for enhancing ocular drug delivery. This system significantly improves curcumin’s solubility and stability while offering robust antioxidant activity. In vitro and in vivo studies demonstrated that these nanomicelles enhance cellular uptake and corneal permeation compared to free curcumin solutions. Additionally, the formulation exhibited excellent ocular tolerance, with no signs of irritation in rabbit models. These results suggest that nanomicelles could serve as an effective platform for delivering curcumin topically in the treatment of ocular inflammation and related conditions . Recent advancements in curcumin-based therapies for AMD highlight the promise of aqueous nanomicellar formulations (CUR-NMF). This innovative approach overcomes curcumin’s poor aqueous solubility, a key limitation for its therapeutic use. CUR-NMF, developed using hydrogenated castor oil (HCO-40) and octoxynol-40 (OC-40), offers a stable delivery system optimized for retinal protection. Studies demonstrate its antioxidant, anti-inflammatory, and anti-angiogenic effects, showing significant protection against oxidative stress in retinal cells and a reduction in VEGF release, a critical factor in AMD pathology. Furthermore, sustained drug release profiles and favorable safety assessments suggest CUR-NMF could provide long-term therapeutic benefits for both dry and wet AMD . Cyclodextrins (CDs) play a crucial role in enhancing curcumin’s solubility, stability, and bioavailability through the formation of inclusion complexes . These complexes significantly improve curcumin’s therapeutic effectiveness in both anterior and posterior ocular diseases. Among the CDs, β-cyclodextrins (βCD) and γ-cyclodextrins (γCD) stand out due to their higher solubility, capacity to self-assemble into nanoaggregates, and favorable safety profile for ocular applications . Since the mid-1990s, CD-based inclusion complexes have provided significant technological advantages for pharmaceutical formulations. By enhancing the stability, solubility, and bioavailability of bioactive compounds, these complexes address key challenges in drug delivery. A notable example is the use of cyclodextrins to improve the solubility and stability of chloramphenicol, a patented formulation still in use today . This approach enabled the development of stable pharmaceutical solutions and optimized separation processes, underscoring the enduring relevance of this technology . Methyl-β-cyclodextrin (M-β-CD) is the cyclodextrin used in commercial eye drop formulations, including those containing chloramphenicol. Hydroxypropyl-β-cyclodextrin (HP-β-CD) is also covered by this patent for chloramphenicol due to its superior solubility and biocompatibility . More recently, CD-based inclusion complexes have been explored for cannabinoid delivery, demonstrating promising potential in pain management and anti-inflammatory therapies. These developments highlight the versatility of cyclodextrins in modern drug delivery, particularly for ocular applications, where solubility, stability, and bioavailability are critical for therapeutic success . Modified CDs, such as ethylene diamine (EDA)-modified βCD, have demonstrated superior capabilities in improving curcumin (CUR) solubility and stability. These complexes provide enhanced thermodynamic properties, making CUR more bioavailable for ocular applications. Curcumin-EDA-βCD nanoparticles exhibit excellent corneal permeability, as shown in vitro porcine cornea experiments, and maintain high biocompatibility, confirmed by histological analyses of porcine corneas and bovine corneal epithelial cell viability. These properties make them particularly suitable for addressing anterior segment diseases like keratitis and dry eye disease . In a recent study, various CD-curcumin complexes were prepared and characterized, showing significant improvements in solubility. The freeze-drying method produced highly soluble complexes, and the optimal formulation provided sustained release for over 96 h. This approach offers a promising solution for curcumin’s use in ocular therapies, such as eye drops for conditions like retinitis pigmentosa . Inclusion complexes of CUR with hydroxypropyl-β-cyclodextrin (HP-βCD) were successfully developed using the cosolvency/lyophilization method, resulting in significant improvements in CUR solubility, stability, and therapeutic efficacy. The complexes demonstrated superior antioxidant and anti-inflammatory activities compared to free CUR. To facilitate ocular administration, an in-situ gel system was prepared using Pluronic F127 and chitosan, providing mucoadhesion and sol-gel transition between 26–35 °C. Viscosity, pH, and clarity tests confirmed the system’s suitability for ocular application. In vitro release studies showed sustained drug release for 6 h, fitting the Weibull kinetic model. This approach offers a promising drug delivery strategy for ocular diseases, supporting prolonged and controlled drug release . The poor solubility and stability of CUR limit its application in ocular drug delivery. To address this, CUR was complexed with βCD and HP-βCD using co-solvent, sonication, and freeze-drying methods in 1:1 and 1:2 molar ratios. The freeze-drying method produced the most water-soluble complexes. Among the 12 tested formulations, the F11 formulation, prepared with pH 6.8 phosphate buffer containing 1% Tween 80, demonstrated sustained drug release for over 96 h. The drug release followed a Higuchi non-Fickian diffusion model. These findings suggest that F11 could be developed as a once-daily eye drop formulation, offering a promising approach for the sustained delivery of curcumin in the treatment of ocular diseases, such as retinitis pigmentosa . The use of CDs to optimize corneal penetration of CUR has shown promising results. In ex vivo models using porcine corneas, the combination of CDs with nanoparticles demonstrated greater drug permeation. This improvement is attributed to the ability of CDs to form inclusion complexes, enhancing curcumin’s solubility and stability, while nanoparticles enable sustained release and protection against enzymatic degradation . Recent studies have demonstrated that curcumin-loaded hydrogels, such as those incorporating CUR nanoparticles encapsulated with βCD and hyaluronic acid, accelerate corneal healing in ulcerative keratitis. This system not only improves corneal clarity and reduces inflammation but also enhances the quality of healed tissues, requiring fewer applications compared to conventional treatments. These formulations hold promises for future therapeutic use in treating ulcerative keratitis and other ocular conditions, providing an innovative, herbal-based alternative to traditional treatments . The penetration of CUR into the cornea was evaluated using an ex vivo porcine eye model and a digital image analysis technique. Several formulation strategies, including oily solutions, oily suspensions, micelles, liposomes, nanosuspensions, and CD complexes, were explored to improve CUR corneal permeability. The results revealed that cyclodextrin-based formulations exhibited superior corneal penetration compared to other delivery systems. The image analysis approach effectively measured CUR penetration into corneal tissues, supporting the potential of cyclodextrin complexes as a delivery strategy for hydrophobic drugs in ocular applications. This technique offers a novel approach for optimizing the penetration of CUR and similar compounds into the cornea . Among the most effective formulations are those based on modified βCDs and conjugates with tetrahydrocurcumin nanoparticles, which exhibited deeper penetration into ocular tissues. These strategies hold significant potential for treating both anterior and posterior segment ocular diseases, as they increase bioavailability and extend therapeutic effects . γCD-based nanoparticles not only enhance drug permeation but also increase retention time on the ocular surface, promoting sustained drug release and reducing the frequency of administration. Additionally, the presence of tear enzymes like α-amylase facilitates drug release from γCD complexes, further boosting bioavailability. For instance, γCD-based eye drops containing dexamethasone achieved higher concentrations in ocular tissues compared to commercial formulations. Moreover, these formulations were well tolerated, with no significant ocular irritation or toxicity observed. γCD has also been employed in formulations for dorzolamide, telmisartan, and nepafenac, demonstrating improved pharmacokinetics and sustained drug release for up to 24 h . Inclusion complexes of CUR with HP-βCD enhance curcumin’s solubility, dissolution rate, and bioavailability, essential for ocular drug delivery. Studies using co-evaporation methods revealed a 1:1 molar ratio complex with a solubility constant of 30.09 mM −1 . Characterization techniques, such as XRD, confirmed the loss of curcumin’s crystalline structure, while FTIR and DTA indicated no chemical interactions. In vitro dissolution tests showed faster release of CUR from the complex compared to its pure form and physical mixtures. This approach improves curcumin’s bioavailability, making it a promising strategy for ocular drug delivery systems . The ocular delivery of CUR faces significant barriers due to anatomical and physiological constraints; however, advances in nanoengineered systems have shown promising results. The formation of inclusion complexes with HP-CDs through spray-drying significantly enhanced the solubility, permeability, and stability of CUR. Enhanced corneal and retinal permeability was observed, along with increased antioxidant activity in ocular epithelial cells, including upregulation of SOD1, CAT1, and HMOX1. Moreover, protection against oxidative stress was confirmed in rabbit corneal tissues. These findings highlight the potential of CUR:HP-CD complexes to improve ocular drug bioavailability, thereby enhancing therapeutic outcomes for ocular diseases . Cyclodextrin-based systems significantly enhance curcumin’s bioavailability, solubility, and therapeutic potential for ocular drug delivery. Advances in βCD, γCD, and HP-βCD systems, combined with nanoparticles or in situ gels, have demonstrated improved drug permeation, sustained release, and higher bioactivity. These strategies support the development of more effective ophthalmic treatments. Nanocomposites, such as cupriferous hollow nanoshells, combine silver and copper ions. These materials exhibit dual functionality: silver ions provide potent antibacterial activity, while copper ions promote tissue regeneration by stimulating fibroblast migration and angiogenesis. This dual approach is particularly beneficial in treating conditions like keratitis, where infections can impair spontaneous recovery and cause corneal damage. Nanocomposite-based treatments not only target the bacteria but also support the healing of damaged tissues, offering a comprehensive approach to managing complex infections . Moreover, curcumin enhances traditional antibiotics by inhibiting bacterial efflux pumps and disrupting biofilms—critical mechanisms in antibiotic resistance. When combined with biopolymers like chitosan, curcumin has shown enhanced antibacterial effects, even at low concentrations, especially against resistant bacterial strains. This makes curcumin-based formulations a valuable tool in combating antibiotic-resistant ocular infections, such as conjunctivitis and keratitis . Curcumin-based formulations have also demonstrated significant efficacy in the treatment of conjunctivitis. Products like Haridra ® and Ophthacare ® have been shown to combat pathogens like Escherichia coli , Staphylococcus aureus , Klebsiella pneumoniae , and Pseudomonas aeruginosa while also reducing inflammation and irritation. These formulations address not only the infection but also the underlying inflammation, providing a comprehensive treatment approach. Ophthacare ® , which combines Curcuma longa with other herbal extracts, offers an effective solution for a range of ocular conditions, including dry eye and inflammatory conjunctival disorders . Endophthalmitis, an intraocular infection characterized by extensive inflammation and retinal damage, has benefited from nanotechnology-based drug delivery systems. Hybrid frameworks that incorporate silver nanoparticles and photosensitizers have been developed to disrupt biofilms while preserving host tissues. These systems, combined with curcumin’s anti-inflammatory properties, can modulate cytokine storms, support retinal cell survival, and preserve ocular structures. This combination not only targets the infection but also helps to protect the delicate retinal tissues, improving patient outcomes . Recent innovations in drug delivery systems have further amplified the therapeutic potential of curcumin. Nanoparticles and liposomes are particularly effective at enhancing curcumin’s bioavailability and ocular penetration, ensuring sustained therapeutic effects. For example, dual-drug nanofibers, which combine curcumin with antibiotics, have shown enhanced bactericidal activity and accelerated tissue regeneration in preclinical models of ocular infections. These advanced delivery systems ensure that curcumin reaches the target site effectively, offering continuous antimicrobial action and supporting tissue healing . These systems not only enhance the effectiveness of conventional antibiotics but also provide innovative solutions to overcome the challenges posed by MDR bacteria, biofilms, and tissue damage . The integration of curcumin in these systems adds a further layer of therapeutic benefit, making it a promising tool in the management of ocular infections. 4.1. Photodynamic Therapy (PDT) Curcumin’s photosensitizing properties make it a promising candidate for PDT, targeting pathological cells in conditions such as ocular tumors and infections. This approach is particularly relevant for eyelid-specific conditions, offering : Mechanisms: dual role as a photosensitizer and therapeutic agent. Applications: minimally invasive treatment for tumors, infections, and inflammatory disorders. Benefits: combines antioxidant and anti-inflammatory effects to enhance therapeutic outcomes for eyelid diseases. 4.2. Mucoadhesive Formulations Mucoadhesive drug delivery systems, including hydrogels and films, prolong curcumin’s contact time with the ocular surface, increasing its therapeutic efficacy. These systems can be tailored to eyelid disorders such as the following : Blepharitis and Dermatitis: prolonged retention enhances localized anti-inflammatory and antioxidant effects. Sustained Drug Release: mucoadhesive properties ensure better therapeutic outcomes for chronic eyelid conditions. Clinical Potential: effective for diseases requiring extended drug action, like anterior uveitis and diabetic retinopathy. 4.3. Neuroprotective Effects in Neurological Eyelid Disorders Curcumin offers neuroprotective benefits by reducing neuroinflammation and promoting cell survival, with applications in : Blepharospasm: reduces oxidative stress and neuroinflammation, alleviating involuntary twitching. Neuropathic Inflammation: modulates immune signaling, potentially relieving neuropathic eyelid pain. Mechanisms: targets inflammatory pathways (e.g., NF-κB and TLR4) and enhances antioxidant activity to protect against tissue damage. Curcumin’s photosensitizing properties make it a promising candidate for PDT, targeting pathological cells in conditions such as ocular tumors and infections. This approach is particularly relevant for eyelid-specific conditions, offering : Mechanisms: dual role as a photosensitizer and therapeutic agent. Applications: minimally invasive treatment for tumors, infections, and inflammatory disorders. Benefits: combines antioxidant and anti-inflammatory effects to enhance therapeutic outcomes for eyelid diseases. Mucoadhesive drug delivery systems, including hydrogels and films, prolong curcumin’s contact time with the ocular surface, increasing its therapeutic efficacy. These systems can be tailored to eyelid disorders such as the following : Blepharitis and Dermatitis: prolonged retention enhances localized anti-inflammatory and antioxidant effects. Sustained Drug Release: mucoadhesive properties ensure better therapeutic outcomes for chronic eyelid conditions. Clinical Potential: effective for diseases requiring extended drug action, like anterior uveitis and diabetic retinopathy. Curcumin offers neuroprotective benefits by reducing neuroinflammation and promoting cell survival, with applications in : Blepharospasm: reduces oxidative stress and neuroinflammation, alleviating involuntary twitching. Neuropathic Inflammation: modulates immune signaling, potentially relieving neuropathic eyelid pain. Mechanisms: targets inflammatory pathways (e.g., NF-κB and TLR4) and enhances antioxidant activity to protect against tissue damage. Curcumin’s potential in eyelid diseases is supported by its potent anti-inflammatory, immunomodulatory, antioxidant, and antibacterial properties. The antibacterial activity of curcumin could be particularly beneficial in treating eyelid infections, such as those caused by Staphylococcus aureus or other bacterial pathogens, which are common in conditions like blepharitis and eyelid dermatitis. Future research should focus on developing targeted delivery systems, such as mucoadhesive and nanoparticle formulations, to enhance efficacy in localized eyelid treatments. Well-designed clinical trials are needed to validate curcumin’s safety and effectiveness in eyelid conditions, including blepharitis, blepharospasm, and eyelid dermatitis. Additionally, exploring combination therapies that integrate curcumin with conventional treatments or other phytochemicals could provide solutions for refractory eyelid conditions. Given curcumin’s versatility as a therapeutic agent and advances in drug delivery technologies, it holds significant promise for addressing unmet needs in eyelid disease treatment. Curcumin demonstrates significant therapeutic potential in ophthalmology, particularly for retinal and corneal diseases, due to its anti-inflammatory, antioxidant, antibacterial, and anti-angiogenic properties. Its antibacterial activity could enhance treatment options for ocular surface infections, such as conjunctivitis or keratitis, by directly combating bacterial pathogens. However, challenges related to bioavailability and solubility need to be overcome through advanced drug delivery systems like nanoparticles, niosomes, and cyclodextrin complexes. Curcumin’s therapeutic value lies in its pleiotropic effects, including anti-inflammatory, antioxidant, and anti-angiogenic activities. In comparison to traditional treatments, curcumin offers a multi-targeted approach that may complement or enhance existing therapies. For example, its ability to prevent inflammation and oxidative damage positions it as a potential adjunct to anti-VEGF treatments for conditions like age-related macular degeneration. However, its clinical application is limited by poor bioavailability, necessitating further research to establish its clinical effectiveness relative to conventional treatments. In conclusion, curcumin holds promising therapeutic potential for ophthalmology, but further studies, especially clinical trials, are required to confirm its clinical efficacy and overcome existing limitations. The continued exploration of innovative delivery systems will be key to unlocking its full therapeutic potential. |
Transesophageal Echocardiography‐Related Complications During Mitral Valve Repair in Dogs | 80596230-a049-44e7-932e-86dddc5c3c72 | 11912017 | Cardiovascular System[mh] | Introduction Transesophageal echocardiography (TEE) is a crucial modality in cardiac surgery in dogs. It is particularly useful for patent ductus arteriosus occlusion and balloon valvuloplasty for pulmonary valve stenosis and is also valuable for the diagnosis and detailed evaluation of less common and complex congenital heart diseases [ , , , ]. In addition to catheter intervention, it is an important technique in open‐heart surgery. Perioperative TEE is a key diagnostic modality in human mitral valve surgery . It offers comprehensive insights into valve anatomy, lesion severity, and ventricular performance, aiding surgeons in selecting the most appropriate surgical approach and assessing preoperative risks . Additionally, TEE is used intraoperatively to monitor the progress of the surgery, allowing for adjustments to the initial surgical plan as needed. It is also invaluable for evaluating surgical outcomes and promptly diagnosing potential complications during the immediate postoperative period . Perioperative TEE influences surgical plans in 10%–25% of human mitral valve surgery . Open‐heart surgery techniques in dogs, like those in humans, also require TEE for procedural planning and intraoperative support [ , , , , ]. TEE is performed to support the surgery by confirming de‐airing as the patients come off cardiopulmonary bypass, providing immediate assessment of surgical results, and monitoring ventricular function and hemodynamic status intraoperatively . Additionally, it offers high‐resolution images during mitral valve repair (MVR) in dogs , which is a potential curative treatment option for dogs with advanced‐stage myxomatous mitral valve disease. In humans, although TEE is a relatively safe imaging modality with a low incidence of complications, the rates of adverse events are higher in pediatric patients than in adults owing to their smaller size . To minimize the complications associated with TEE in pediatric patients, recommendations include selecting probes based on size and careful probe insertion and manipulation . Dogs are similar to children in that they generally have smaller body sizes. A veterinary clinical study on TEE, focusing on the diagnosis and interventional procedures for congenital heart disease, reported only a few mild complications, even in small‐breed dogs (body weight < 4 kg) . We assumed that small breed dogs with cardiac enlargement might compress the thoracic space and have a shorter distance from the heart to the vertebra , potentially increasing the risk of TEE‐related complications. However, the complications of TEE related to open‐heart surgery in dogs remain unclear. During the MVR procedure, heparinization is required to manage cardiopulmonary bypass , which carries a risk of bleeding. Dogs with MMVD often exhibit specific characteristics, such as advanced age, being small‐breed, and cardiac enlargement. Additionally, they could have reduced forward cardiac output due to severe mitral regurgitation. Therefore, evaluating potential adverse events in this dog population is particularly important. Thus, this study aimed to evaluate the incidence and types of complications following TEE probe manipulation during MVR in dogs and to investigate the factors associated with new mucosal injuries.
Animals, Materials, and Methods 2.1 Study Design and Animals This prospective study was approved by the ethics committee of our institution (approval number: 210406‐7). Consent was obtained from the owners before study inclusion. Dogs that presented at JASMINE Veterinary Cardiovascular Medical Center to undergo MVR were eligible for inclusion. Dogs were excluded if they had clinical or historical evidence of esophageal disease, a history of brachycephalic obstructive airway syndrome, or if they were judged by cardiology clinicians to be too small in body size for TEE. TEE was performed during the anesthesia period to support MVR. A small probe (9T probe, tip size: 10.9 × 8.4 mm, GE HealthCare Japan, Tokyo, Japan) was used in dogs weighing < 5 kg, whereas a large probe (6VT‐D probe, tip size: 12.6 × 14.3 mm, GE HealthCare Japan) was used in dogs weighing ≥ 5 kg. In all dogs, the TEE probe was carefully inserted while monitoring the TEE images until a certain image was reached, without the use of fluoroscopy. Background information of the dogs was obtained from medical records, including breed, sex, age, body weight, body condition score, and ACVIM clinical stage . 2.2 Surgical Procedure and Transesophageal Echocardiography Probe Manipulation The manipulation was performed by well‐trained cardiology clinicians (AT, YN, and TS) to support MVR after the induction of general anesthesia. Atropine sulfate (0.03 mg/kg, SC) was administered, and midazolam (0.3 mg/kg, IV) and fentanyl (5 μg/kg, IV) were administered as premedication. Subsequently, ketamine (5 mg/kg, IV) was administered to induce anesthesia, and endotracheal intubation was performed. Additionally, anesthesia was maintained using sevoflurane (1.0%–3.0%) intraoperatively, and perioperative analgesia was established by continuous rate infusion of fentanyl (18 μg/kg/min). Fentanyl was continuously administered at a rate of 3–5 μg/kg/min for 24 h postoperatively for pain management. The MVR procedure consisted of artificial chordae and annuloplasty using expanded polytetrafluoroethylene threads , performed under cardiopulmonary bypass with an activated coagulation time of > 300 s achieved by heparinization (heparin sodium, 200 IU/kg IV). Cardiac evaluation via TEE was performed three times during the procedure: before connecting the heart to the cardiopulmonary bypass as a preoperative heart evaluation, during cardiac arrest as support for the procedure (removing air at the time of cardiac closure and checking the heart condition immediately after resuscitation), and after weaning from the cardiopulmonary bypass (before extubation) as a postoperative heart evaluation. The clinicians reinserted the TEE probe for each of the three evaluation periods. Three‐dimensional imaging was performed with the 6VT‐D probe during the initial and final TEE evaluations as part of the pre‐ and postoperative assessments. A digital stopwatch was used to monitor the duration of TEE. The stopwatch was started each time image acquisition was ongoing and was stopped when the probe was not being used, at which point the image was frozen to prevent overheating. Each probe had an autotemp shutdown feature set at 41.8°C, ensuring that if the temperature exceeded this threshold, the probe would automatically shut down to prevent overheating. When this temperature was reached, the probe was no longer capable of imaging until the temperature decreased below the predetermined temperature set by the manufacturer. The probe temperature was continuously monitored and stored within the ultrasound machine throughout the surgical procedure, with the highest temperature constantly updated and displayed in real time. At the end of the procedure, the highest temperature recorded during the entire procedure was reviewed. The ease of obtaining TEE images was categorized into three levels at the initial insertion: ‘poor’ (i.e., images had low temporal or spatial resolution necessitating invasive anteflexion or retroflexion of the probe to optimize visualization), ‘suboptimal’ (i.e., the image quality was reduced but did not interfere with procedural guidance nor required invasive probe manipulation), and ‘good’ (i.e., satisfactory images were obtained with minimal probe manipulation). 2.3 Evaluation of Mucosal Injury A single observer (KK) performed endoscopy in narrowband imaging (NBI) mode (VQ‐5112C, tip diameter: 5.4 mm, Olympus Marketing Inc., Tokyo, Japan), which could clearly distinguish mucosal lesions [ , , ]. Evaluation was performed before and after the surgical procedure to identify mucosal injuries: once after the induction of anesthesia and once before extubation to evaluate the occurrence of new esophageal lesions. The lesion sites were defined as oral to pharyngeal, upper esophagus, heart base, and lower esophagus. The severity of mucosal lesions was classified as ‘complex’ (intramural hematoma, mucosal laceration), ‘minor’ (petechiae, ecchymosis) , and ‘minute’ (the injury only visible on NBI mode). Petechiae were defined as small (pinpoint) red‐purple, non‐raised (macular), circular lesions, while ecchymosis was defined as larger confluent petechial lesions. An intramural hematoma was defined as a collection of blood in the submucosa causing a circumscribed elevated lesion. Laceration or abrasion was defined as a defect in the mucosal surface, as previously described in a human study . 2.4 Hemodynamic Changes Hemodynamic changes were evaluated before and after the initial TEE insertion (prior to connecting the heart to the cardiopulmonary bypass). Arterial blood pressure was measured invasively by placing a catheter in the dorsalis pedis artery with a pressure transducer (DX‐300; Nihon Kohden, Tokyo, Japan) and a vital monitor (Life Scope VS; Nihon Kohden, Tokyo, Japan). Systolic, diastolic, and mean arterial blood pressures, as well as heart rate were obtained from the monitor. For data analysis, the arithmetic mean of 10 measurements was used. Additionally, a 15% change in systolic blood pressure and heart rate before and after TEE device insertion was considered clinically meaningful. No changes in anesthesia or medication were made during the measurements. 2.5 Echocardiography and Thoracic Radiography The radiographs and transthoracic echocardiogram used in this study were the most recent data collected from medical records within 3 months prior to MVR. Both evaluations were performed on the same day. Vertebral heart size was measured using thoracic radiography. In the echocardiographic examination, the left atrial to aortic ratio was obtained from the right parasternal short axis of the heart base. The left ventricular end‐diastolic internal diameter normalized to body weight was obtained from a right parasternal view at the level of the chordae tendineae using the M‐mode. These measurements were performed by several well‐trained cardiology attending clinicians at the institution using an echocardiographic machine (Vivid E95; GE Healthcare Japan, Tokyo, Japan) with a 1.0–5.0‐MHz sector probe (6S‐D; GE HealthCare Japan, Tokyo, Japan). 2.6 Statistical Analysis The Shapiro–Wilk test was used to assess the normality of the distribution of continuous variables, and normally distributed variables were expressed as the mean ± standard deviation (SD), while non‐normally distributed variables were expressed as the median [interquartile range, IQR]. Measurement values and baseline variables were compared between groups using Student's t ‐test or Mann–Whitney U test. The paired t ‐test or Wilcoxon Signed‐Rank test was used to compare the changes in blood pressure and heart rate pre‐ and post‐TEE, depending on the results of their respective normality analysis. Proportional comparisons were performed using Fisher's exact test. Univariate logistic regression analysis was performed to investigate the effect of the variable of interest on the appearance of a new mucosal injury, with the results expressed as odds ratios (ORs) and 95% confidence intervals (CIs). All statistical analyses were performed using the R software (version 4.2.2; Foundation for Statistical Computing, Vienna, Austria). Differences were considered statistically significant at p < 0.050.
Study Design and Animals This prospective study was approved by the ethics committee of our institution (approval number: 210406‐7). Consent was obtained from the owners before study inclusion. Dogs that presented at JASMINE Veterinary Cardiovascular Medical Center to undergo MVR were eligible for inclusion. Dogs were excluded if they had clinical or historical evidence of esophageal disease, a history of brachycephalic obstructive airway syndrome, or if they were judged by cardiology clinicians to be too small in body size for TEE. TEE was performed during the anesthesia period to support MVR. A small probe (9T probe, tip size: 10.9 × 8.4 mm, GE HealthCare Japan, Tokyo, Japan) was used in dogs weighing < 5 kg, whereas a large probe (6VT‐D probe, tip size: 12.6 × 14.3 mm, GE HealthCare Japan) was used in dogs weighing ≥ 5 kg. In all dogs, the TEE probe was carefully inserted while monitoring the TEE images until a certain image was reached, without the use of fluoroscopy. Background information of the dogs was obtained from medical records, including breed, sex, age, body weight, body condition score, and ACVIM clinical stage .
Surgical Procedure and Transesophageal Echocardiography Probe Manipulation The manipulation was performed by well‐trained cardiology clinicians (AT, YN, and TS) to support MVR after the induction of general anesthesia. Atropine sulfate (0.03 mg/kg, SC) was administered, and midazolam (0.3 mg/kg, IV) and fentanyl (5 μg/kg, IV) were administered as premedication. Subsequently, ketamine (5 mg/kg, IV) was administered to induce anesthesia, and endotracheal intubation was performed. Additionally, anesthesia was maintained using sevoflurane (1.0%–3.0%) intraoperatively, and perioperative analgesia was established by continuous rate infusion of fentanyl (18 μg/kg/min). Fentanyl was continuously administered at a rate of 3–5 μg/kg/min for 24 h postoperatively for pain management. The MVR procedure consisted of artificial chordae and annuloplasty using expanded polytetrafluoroethylene threads , performed under cardiopulmonary bypass with an activated coagulation time of > 300 s achieved by heparinization (heparin sodium, 200 IU/kg IV). Cardiac evaluation via TEE was performed three times during the procedure: before connecting the heart to the cardiopulmonary bypass as a preoperative heart evaluation, during cardiac arrest as support for the procedure (removing air at the time of cardiac closure and checking the heart condition immediately after resuscitation), and after weaning from the cardiopulmonary bypass (before extubation) as a postoperative heart evaluation. The clinicians reinserted the TEE probe for each of the three evaluation periods. Three‐dimensional imaging was performed with the 6VT‐D probe during the initial and final TEE evaluations as part of the pre‐ and postoperative assessments. A digital stopwatch was used to monitor the duration of TEE. The stopwatch was started each time image acquisition was ongoing and was stopped when the probe was not being used, at which point the image was frozen to prevent overheating. Each probe had an autotemp shutdown feature set at 41.8°C, ensuring that if the temperature exceeded this threshold, the probe would automatically shut down to prevent overheating. When this temperature was reached, the probe was no longer capable of imaging until the temperature decreased below the predetermined temperature set by the manufacturer. The probe temperature was continuously monitored and stored within the ultrasound machine throughout the surgical procedure, with the highest temperature constantly updated and displayed in real time. At the end of the procedure, the highest temperature recorded during the entire procedure was reviewed. The ease of obtaining TEE images was categorized into three levels at the initial insertion: ‘poor’ (i.e., images had low temporal or spatial resolution necessitating invasive anteflexion or retroflexion of the probe to optimize visualization), ‘suboptimal’ (i.e., the image quality was reduced but did not interfere with procedural guidance nor required invasive probe manipulation), and ‘good’ (i.e., satisfactory images were obtained with minimal probe manipulation).
Evaluation of Mucosal Injury A single observer (KK) performed endoscopy in narrowband imaging (NBI) mode (VQ‐5112C, tip diameter: 5.4 mm, Olympus Marketing Inc., Tokyo, Japan), which could clearly distinguish mucosal lesions [ , , ]. Evaluation was performed before and after the surgical procedure to identify mucosal injuries: once after the induction of anesthesia and once before extubation to evaluate the occurrence of new esophageal lesions. The lesion sites were defined as oral to pharyngeal, upper esophagus, heart base, and lower esophagus. The severity of mucosal lesions was classified as ‘complex’ (intramural hematoma, mucosal laceration), ‘minor’ (petechiae, ecchymosis) , and ‘minute’ (the injury only visible on NBI mode). Petechiae were defined as small (pinpoint) red‐purple, non‐raised (macular), circular lesions, while ecchymosis was defined as larger confluent petechial lesions. An intramural hematoma was defined as a collection of blood in the submucosa causing a circumscribed elevated lesion. Laceration or abrasion was defined as a defect in the mucosal surface, as previously described in a human study .
Hemodynamic Changes Hemodynamic changes were evaluated before and after the initial TEE insertion (prior to connecting the heart to the cardiopulmonary bypass). Arterial blood pressure was measured invasively by placing a catheter in the dorsalis pedis artery with a pressure transducer (DX‐300; Nihon Kohden, Tokyo, Japan) and a vital monitor (Life Scope VS; Nihon Kohden, Tokyo, Japan). Systolic, diastolic, and mean arterial blood pressures, as well as heart rate were obtained from the monitor. For data analysis, the arithmetic mean of 10 measurements was used. Additionally, a 15% change in systolic blood pressure and heart rate before and after TEE device insertion was considered clinically meaningful. No changes in anesthesia or medication were made during the measurements.
Echocardiography and Thoracic Radiography The radiographs and transthoracic echocardiogram used in this study were the most recent data collected from medical records within 3 months prior to MVR. Both evaluations were performed on the same day. Vertebral heart size was measured using thoracic radiography. In the echocardiographic examination, the left atrial to aortic ratio was obtained from the right parasternal short axis of the heart base. The left ventricular end‐diastolic internal diameter normalized to body weight was obtained from a right parasternal view at the level of the chordae tendineae using the M‐mode. These measurements were performed by several well‐trained cardiology attending clinicians at the institution using an echocardiographic machine (Vivid E95; GE Healthcare Japan, Tokyo, Japan) with a 1.0–5.0‐MHz sector probe (6S‐D; GE HealthCare Japan, Tokyo, Japan).
Statistical Analysis The Shapiro–Wilk test was used to assess the normality of the distribution of continuous variables, and normally distributed variables were expressed as the mean ± standard deviation (SD), while non‐normally distributed variables were expressed as the median [interquartile range, IQR]. Measurement values and baseline variables were compared between groups using Student's t ‐test or Mann–Whitney U test. The paired t ‐test or Wilcoxon Signed‐Rank test was used to compare the changes in blood pressure and heart rate pre‐ and post‐TEE, depending on the results of their respective normality analysis. Proportional comparisons were performed using Fisher's exact test. Univariate logistic regression analysis was performed to investigate the effect of the variable of interest on the appearance of a new mucosal injury, with the results expressed as odds ratios (ORs) and 95% confidence intervals (CIs). All statistical analyses were performed using the R software (version 4.2.2; Foundation for Statistical Computing, Vienna, Austria). Differences were considered statistically significant at p < 0.050.
Results Sixty client‐owned dogs were enrolled between August 2021 and April 2022 and divided into two groups based on body weight: 45 dogs, < 5 kg; 15 dogs, ≥ 5 kg. The breeds included were Chihuahua ( n = 31), mixed ( n = 9), Cavalier King Charles spaniel ( n = 5), Toy Poodle ( n = 4), Maltese ( n = 3), Pomeranian ( n = 3), Miniature Schnauzer ( n = 2), Papillion ( n = 1), Shetland Sheepdog ( n = 1), and Shiba ( n = 1). The median body weight was 3.8 kg [IQR: 3.1–5.1 kg], and the mean ± SD age was 129 ± 22 months. Moreover, 39/60 dogs were male. The baseline characteristics and intraoperative findings by group are shown in Table . The maximum probe temperature during the procedure was higher in the ≥ 5 kg group (39.5°C ± 0.9°C vs. 40.4°C ± 1.1°C, p = 0.002). The ease of obtaining TEE images was ‘good’ in 57/60 dogs and ‘suboptimal’ in 3/60 dogs. No dogs had images classified as ‘poor’. Preprocedure, pre‐existing mucosal lesions were detected in two dogs: one presented with a single ‘minor’ lesion in the upper esophagus, and the other presented with multiple ‘minor’ lesions from the upper to the lower esophagus. These pre‐existing lesions did not worsen during the surgical procedure. Postprocedure, new mucosal lesions were detected in 20/60 dogs: 16/20 dogs had only ‘minute’ lesions, and 4/20 dogs had ‘minor’ lesions. No ‘complex’ lesions were observed (Table ). A representative image is shown in Figure . Univariate logistic regression analysis identified no significant factors of new mucosal lesions for any of the variables of interest (Table ). Regarding hemodynamic changes between before and after the first TEE insertion (preoperative heart evaluation), systolic blood pressure changed from 95 ± 13 to 92 ± 11 mmHg ( p = 0.008), diastolic pressure changed from 56 ± 9 to 54 ± 8 mmHg ( p = 0.063), the mean blood pressure changed from 66 ± 11 to 63 ± 9 mmHg ( p = 0.006), and the heart rate changed from 128 ± 25 to 123 ± 24 bpm ( p < 0.001). Overall, 44/60 of the dogs experienced a decrease in systolic blood pressure from baseline, decreasing from 99 ± 12 to 92 ± 11 mmHg ( p < 0.001), while the remaining dogs experienced an increase from 86 ± 12 mmHg to 94 ± 12 mmHg ( p = 0.004). The heart rate decreased from baseline in 41/60 of the dogs, falling from 131 ± 23 bpm to 121 ± 22 bpm ( p < 0.001), while it increased in 9/60 of the dogs from 121 ± 27 bpm to 126 ± 27 bpm ( p < 0.001). Among the dogs who showed more than a 15% hemodynamic change before and after the first TEE, 2/60 dogs had a decrease in systolic blood pressure, 5/60 dogs had a decrease in heart rate, 4/60 dogs had an increase in systolic blood pressure, and 1/60 dog had an increase in heart rate. Between the < 5 kg and ≥ 5 kg groups, there were no significant differences in pre‐TEE insertion systolic blood pressure (95 ± 13 mmHg vs. 95 ± 12 mmHg; p = 0.969), post‐TEE insertion systolic blood pressure (93 ± 11 mmHg vs. 90 ± 10 mmHg; p = 0.454), pre‐TEE insertion heart rate (127 ± 25 bpm vs. 129 ± 26 bpm; p = 0.833), and post‐TEE insertion heart rate (123 ± 24 bpm vs. 123 ± 25 bpm; p = 0.921).
Discussion The overall incidence reported in this study is 20/60, but the severity of the lesions is considered ‘minute’ in majority of new lesions identified, even under heparin management for cardiopulmonary bypass. Although hemodynamic changes possibly occur due to TEE manipulation , clinically meaningful changes were rarely observed in the present study. These results offer valuable insights into the safe use of TEE during cardiac surgery in veterinary clinical settings. In human medicine, TEE‐related complications occur more frequently in pediatric humans (0.03%–6.7%) [ , , , , , , ] than in adult patients (0.18%–2.8%) [ , , , , , ]. This difference could be attributed to the small body size of pediatric patients. Dogs are similar to pediatric humans in that they generally have small body sizes. A previous study reported new esophageal lesions in 10% of dogs with a median weight of 8.7 kg . The 60 dogs in the present study were smaller (median weight: 3.8 kg) than those in the above study, but only four dogs developed ‘minor’ lesions. However, the complication rate in this study was similar to that reported in the above study, despite the lower body weight. This result might be attributed to the consistent procedural approach employed for MVR. As the study focused exclusively on MVR support, the required images were mainly obtained using the mid‐esophageal four‐chamber view across all dogs. We assume that different TEE views, which require further advancement and anteflexion of the probe, could lead to a higher complication rate. It is important to note that different objectives or conditions could potentially yield different outcomes. A new technique of NBI mode endoscopy revealed lesions that were not visible with conventional views: 16/60 dogs had ‘minute’ lesions that were only visible using the NBI mode. Narrow‐band illuminations improve the contrast of the capillary pattern in the superficial layer compared to ordinary broadband illumination . In humans, this technique has the potential to clearly distinguish between mucosal lesions [ , , ], and improved overall accuracy for depth of superficial esophageal lesions . Our results suggest that NBI can be a valuable tool for detecting subtle mucosal injuries that are not easily visible. The clinical relevance of these lesions should be further investigated in future studies. In humans, esophageal injuries associated with TEE include ‘minor’ esophageal mucosal injuries (e.g., regions of petechiation, erosion, and hematoma), esophageal laceration and perforation, direct pressure necrosis, and thermal injury from prolonged probe contact. A study in veterinary medicine suggests that 10% and 3% of the dogs developed new lesions at the lower esophageal sphincter and at the heart base, respectively . This study also reported that focal pinpoint mucosal erosion and changes in pinpoint hemorrhage were identified. In the present study, mucosal lesions identified by endoscopy before TEE did not worsen after the procedure, and all ‘minor’ lesions were pinpoint petechiae. Furthermore, the present study found that ‘minor’ lesions were observed in one dog at the upper esophagus, one at the heart base, and two at the lower esophagus. When including minute lesions, 17/60 dogs in this study had new lesions in the upper esophagus, suggesting that the insertion and positioning of the TEE probe could have caused injury. The probe might injure the mucosa at the esophageal sphincter and upper esophagus, particularly at the curve near the thoracic inlet, where resistance can be encountered, especially in small‐breed dogs . The difference in injury sites between the present study and the previous report by Stoner et al. , where the lesions were primarily found in the lower esophagus, can be attributed to several reasons. In the present study, TEE was inserted with retroflexion to pass the aortic arch and over an enlarged heart, while monitoring only the TEE image without fluoroscopic guidance, which could have contributed to mucosal injury during insertion. Additionally, we did not routinely obtain advanced views, such as the caudal esophageal position for a heart base view, which requires further advancement and anteflexion of the probe . Most importantly, the ‘minute’ lesions observed in the upper esophagus using NBI imaging had not been previously visualized. Moreover, the ‘minor’ lesions found in the lower esophagus in this study were consistent with those reported in the previous study . The authors hypothesized that small breed dogs with cardiac enlargement could compress the thoracic space and have a shorter distance from the heart to the vertebra , potentially increasing the risk of TEE‐related mucosal injury. However, in the present study, body weight, VHS, LA:Ao, and LVIDDN were not identified as significant risk factors for the development of new mucosal lesions. Based on these results, TEE was considered safe even in small‐breed dogs with cardiac enlargement, at least in the context of performing MVR surgery. As previously described in human medicine, a prolonged duration of TEE might be a risk factor for mucosal injury . However, our findings did not reveal a relationship between TEE active imaging time and mucosal injury. As this study focused solely on surgical support for MVR, the manipulation times among dogs were relatively consistent. A more comprehensive evaluation and continuous procedural support with TEE, particularly for congenital heart diseases such as patent ductus arteriosus and pulmonary vulve stenosis , would likely require additional time and extensive probe manipulation, which could lead to complications. Additionally, there was no difference in manipulation time between the dogs weighing < 5 kg with a 9T probe and the dogs weighing ≥ 5 kg with a 6VT‐D probe. However, the maximum temperature of the 6VT‐D probe was higher than that of the 9T probe, possibly because of the three‐dimensional imaging performed with the 6VT‐D probe. Nevertheless, the probe temperature might not have influenced the findings in this study, as few lesions were identified at the heart base, and the probe temperature did not increase the probability of identifying new mucosal lesions, similar to findings in human medicine . In humans, possible gastrointestinal complications following TEE include odynophagia, dysphagia, hoarseness, and nausea [ , , , ]. In the present study, various factors influenced postoperative gastrointestinal symptoms making it difficult to determine whether TEE findings are specifically associated with reflux or vomiting after surgery. Additionally, nausea without vomiting is challenging to assess in dogs. Under general anesthesia, the use of sevoflurane, as was used in this study, can be associated with postoperative regurgitation and vomiting . Particularly during cardiopulmonary bypass, hemodilution and anemia can occur due to overhydration from the bypass circuit fluid and cardioplegia solution , which could contribute to postoperative gastrointestinal complications and transfusion‐related gastrointestinal signs . Further research is required to clarify the gastrointestinal symptoms associated with TEE‐related mucosal injury. With respect to hemodynamic changes, blood pressure and heart rate can decrease during TEE. A report conducted 24‐h Holter monitoring and blood pressure measurements in 54 unsedated human patients undergoing TEE, suggesting that the average systolic blood pressure increased from 125 to 141 mmHg in 77% of the patients, whereas 22% experienced a decrease from 122 to 115 mmHg during the examination . However, it is important to note that these findings are based on outcomes in unsedated humans. In contrast, a decrease in systolic blood pressure was more common in the present study of anesthetized dogs, with 41/60 dogs showing a decrease and only 9/60 dogs showing an increase. Given that this study was conducted under anesthesia, which mitigated the stress associated with TEE probe manipulation, a procedure‐related increase in blood pressure was avoided. The mild decrease in blood pressure might have been due to the vagus nerve reflex triggered by esophageal stimulation . In cats, TEE probe compression can induce occlusion of pulmonary venous inflow, potentially resulting in cardiac arrest and bradycardia . We believe that the clinically meaningful hemodynamic changes were minimized in this study. However, it should be noted that these changes likely depend on factors such as baseline blood pressure, probe selection, and body size. While no dogs received changes in medication or anesthesia during the initial TEE examination, controlling baseline hemodynamic values proved challenging and might have influenced the results. Indeed, dogs with increased systolic blood pressure tended to have lower baseline systolic blood pressure, and vice versa. A similar trend was observed with heart rate. A few dogs exhibited more than 15% hemodynamic changes before and after TEE, likely due to vagal stimulation or heart compression, highlighting the need for close monitoring of potential bradycardia and hypotension throughout the procedure. 4.1 Study Limitations This study has some limitations. First, TEE was primarily used for intraoperative management, necessitating minimal imaging. Therefore, the dogs in our study did not require invasive manipulation of the probe to obtain images, which might limit the generalizability of our findings into other clinical settings. Second, the total time the probe remained in the esophagus was not measured in this study; however, as the probe was removed after each measurement, this does not appear to be a substantial issue. The assessment of mucosal lesions was inherently subjective, introducing potential variability into the findings. Additionally, the clinical impact of minute mucosal lesions remains unclear and warrants further investigation in other clinical settings. In terms of hemodynamic changes, it was challenging to strictly control baseline values in the clinical setting due to factors such as anesthesia depth and ventilator settings. Finally, the conclusions of this study are based on a relatively small sample size and limited dog breeds, which could restrict the generalizability of the results. Further investigation in different types of surgeries and dog breeds is necessary.
Study Limitations This study has some limitations. First, TEE was primarily used for intraoperative management, necessitating minimal imaging. Therefore, the dogs in our study did not require invasive manipulation of the probe to obtain images, which might limit the generalizability of our findings into other clinical settings. Second, the total time the probe remained in the esophagus was not measured in this study; however, as the probe was removed after each measurement, this does not appear to be a substantial issue. The assessment of mucosal lesions was inherently subjective, introducing potential variability into the findings. Additionally, the clinical impact of minute mucosal lesions remains unclear and warrants further investigation in other clinical settings. In terms of hemodynamic changes, it was challenging to strictly control baseline values in the clinical setting due to factors such as anesthesia depth and ventilator settings. Finally, the conclusions of this study are based on a relatively small sample size and limited dog breeds, which could restrict the generalizability of the results. Further investigation in different types of surgeries and dog breeds is necessary.
Conclusions During MVR in dogs, TEE‐related mucosal injuries were relatively rare. The results suggest that TEE can be safely performed in small‐breed dogs with cardiac enlargement, even with heparin administration. Although slight changes in blood pressure and heart rate were observed, these changes might not be solely attributable to TEE. Careful monitoring of hemodynamic parameters during TEE probe manipulation is recommended.
Authors declare no off‐label use of antimicrobials.
Approved by the Institutional Ethics Committee of JASMINE Veterinary Cardiovascular Medical Center (approval number: 210406‐7). Authors declare human ethics approval was not needed.
The authors declare no conflicts of interest.
|
Analysis of quality of life and periodontal health with an eight-unit maxillary fixed retainer through a prospective clinical trial | ade6ef3e-3d2b-4228-8527-e20399992446 | 11794599 | Dentistry[mh] | Orthodontic treatment involves the correction of malocclusion. Once the treatment ends, a retention period starts to maintain the corrected results , Without retention, teeth tend to shift back to their original position which can be termed (relapse) , . Commonly used means of retention in orthodontics include removable retainers and fixed retainers (FRs) . Many removable retainers have been used. They were shown to be effective at retaining teeth to their final position, yet they are compliance-demanding appliances . Fixed retainers are commonly chosen by orthodontists due to their aesthetic appeal and ability to ensure retention without relying on patient compliance, while 11% favor them in the maxillary arch . Several studies , – have shown that unwanted changes in tooth position are not related to the original malocclusion, and these changes are associated with canine-to-canine FR alone (6-unit). These changes cannot be termed relapse, but rather unwanted changes. The exact reason for these unwanted changes remains unclear . These movements, even if FR is in place, range from minor rotations for individual teeth to rotations of the whole segment connected with the FR, with a fulcrum in the lower incisors – . A trend toward dual retention instead of solitary removable or fixed retention is more frequently used to avoid the side effects of canine-to-canine fixed retainers alone. however, this trend is still dependent on the patient’s compliance – . An extended fixed retainer would be a simpler alternative to dual retention protocols for overcoming the drawbacks of canine-to-canine retainers. The extended fixed retainer was tested only in the mandibular arch in extraction cases . The extended retainer was effective at preserving extraction spaces and maintaining results during retention with no unwanted changes in tooth position connected by FR . Studies on extended maxillary fixed retainers are scarce in the literature, especially regarding the extent of extension of the FR, the extent of changes in tooth position associated with the FR, the impact of the FR on periodontal tissues especially in the maxillary arch, and patient response along with quality of life. The effects of FR on periodontal health were previously investigated – . several studies , , reported that FR did not have any significant negative effects on periodontal health. Other studies have shown potential harm to the periodontal ligament (PDL). However, most of these studies indicate that the negative effects are related primarily to soft tissues rather than hard tissues , . The effects of orthodontic retainers on speech, self-esteem, and quality of life were investigated. It was reported that temporary speech problems were commonly reported after patients received retainers, these problems lasted from a few days to a few weeks up to 3 months – . This adjustment period can potentially impact patient self-esteem and quality of life, leading to noncompliance with removable retainers and potential relapse. Krämer , reported that patients wearing FRs had lower levels of pain, discomfort, soreness, and tension than those wearing Essix retainers. Patients with FR also found it easier to adjust to the retainer. Similarly, Al-Moghrabi concluded that FRs cause less discomfort and speech difficulties and require less patient compliance. A systematic review concluded that Essix retainers should be avoided if patient compliance is desired, recommending FRs as an alternative. Thus, FRs are recommended for better speech function, aesthetics, stability, and overall quality of life. An extended fixed retainer was hypothesized to be a simpler alternative to dual retention protocols while attempting to overcome canine-to-canine FR drawbacks since the number of units to be bonded is increased . The extended fixed retainer was tested only in the mandibular arch in extraction cases, although the unwanted movement was found to be more strongly associated with maxillary fixed retainers , . There is a lack of studies on extended maxillary fixed retainers in the literature, regarding the extension of the FR, the nature of the associated changes in tooth position, and effects on periodontium and quality of life. Therefore, in this study, aiming to study the possibility of eliminating the need for additional removable retainers (dual retention), an eight-unit extended maxillary fixed retainer was bonded directly after finishing the active orthodontic phase to assess PDL response and patients’ quality of life. This study was conducted to assess the periodontal response to an eight-unit extended maxillary fixed retainer without removable retainers, along with quality of life, and patient satisfaction assessment. This research was approved by the institutional review board of the Faculty of Dentistry, at Alexandria University. all methods were carried out in accordance with ethical approval to conduct research on human subjects follows the Declaration of Helsinki(IORG:0008839, No-0479-8/2022). All experimental protocols were approved by institutional review board of the Faculty of Dentistry, at Alexandria University and ethics committee . The entire study was conducted at the Orthodontic Department at Alexandria University. The first trial registration date of this study was (5/06/2023) following all ethical considerations of Clinical trials. An informed consents to all patients who had been selected to that research and/or their legal guardian(s). This research was a single-center prospective interventional open-label single-arm clinical trial, and the study was registered at Clinicaltrials.gov (NCT05889884). Inclusion criteria Patients who had just finished the orthodontic fixed appliance phase (at least one year of treatment) with extraction or non-extraction treatment and scheduled to start retention. Exclusion criteria Patients with active periodontal disease , systemic disease, bone disease, craniofacial syndromes, cleft, active transverse palatal expansion, malformation, abnormal surface or morphological tooth structure, or restorations were excluded from the study. Sample size calculation The sample size was estimated assuming a 5% alpha error and 80% study power. The sample size was adjusted to a 95% confidence interval (95% CI) to detect changes in probing depth after fixed retainer use. Salvesen et al. reported a mean (SD) probing depth of 2.9 (0.7) mm with a calculated 95% confidence interval = 2.41, 2.99. The required sample size was calculated to be 25 patients, which was increased to 28 to compensate for patients lost to follow-up. The sample size was calculated using MedCalc Statistical Software version 19.0.5 (MedCalc Software bvba, Ostend, Belgium; https://www.medcalc.org ; 2019). Patient preparation First, the study procedures were thoroughly explained to both the participants and their parents and, informed consent was subsequently obtained from each enrolled subject. At the T0 baseline, patients had phase one nonsurgical periodontal therapy (full mouth supragingival and subgingival scaling, root planing and polishing with eugenol-free paste followed by proper oral hygiene instructions (using a toothbrush, dental floss, and interdental brush) before bonding the FR. Intervention At T0 (after bracket debonding, and just before bonding the FR) a full PDL assessment was carried out for the maxillary dentition including (the probing depth, bleeding index, gingival index, mobility index, and plaque index) An impression was made at T0 to fabricate a removable retainer in the case when significant changes occurred during follow-up, which might necessitate immediate study termination, (Futility point) and the use of this retainer is to restore the T0 state. Bonding steps of the extended FR Several measures have been taken to ensure high bond strength while overcoming the high rate of failure of maxillary FR. First, pumice polishing was done for all surfaces to be bonded , followed by a sodium hypochlorite swab for 1 min (Sodium Hypochlorite 5% mint flavor, JK dental vision, Egypt) then acid etching by phosphoric acid 36% for 15 s, along with rinsing etchant surface same amount of time and gentle drying . Pre-hydrolyzed no-mix silane primer and Silane coupling agents (BISCO PORCELAIN PRIMER, BISCO, USA) were added to all surfaces to be bonded . The next step was, Bonding agent application (ASSURE® PLUS, Reliance Orthodontic Products, USA). Holding of the FR was done with the help of dental floss then direct adaptation and festooning of Dead soft wire, 0.027 × 0.011-inch ribbon arch-wire, 8-strand braided wire (FR) (Bond-A-Braid® Lingual Retainer Wire, Reliance Orthodontic Products, USA) from the palatal surface of right premolar to left premolar including the palatal surface of all maxillary anterior teeth in passive state away from the line of occlusion. (Fig. ) The flowable light-curing composite was applied (Polofil® NHT flowable composite light-curing, voco, Germany). Curing for 3 s using high-intensity LED was carried on.( light intensity 2300 mW/cm²)10 W . After finishing the whole curing for all units, selective grinding of excess composite or any interference between FR and lower teeth was done using articulating paper followed by polishing all composite surfaces eliminating any rough area. Details of oral hygiene instructions were provided, including a thorough explanation of the flossing technique to be used with the retainer in place, in addition to guidance on utilizing interdental brushes and water flossers. Patients were given the questionnaires and they filled them out directly after bonding with the FR to assess their quality of life and experience with the extended FR. Patients were followed up regularly each month and were asked to urgently to schedule an appointment if they felt any detachment in the FR. After 12 months of follow-up(6) (T1), all the previous records were repeated with periodontal assessment. Patients were given the quality-of-life questionnaires again to fill them. Statistical analysis Normality was checked for all variables using descriptive statistics (mean, median, and standard deviation), plots (Q-Q plots and histogram), and normality tests. Means and standard deviations were calculated for quantitative variables, while frequencies and percentages were calculated for qualitative variables. Comparisons of quantitative variables at T0 and T1 were performed using paired t-test for normally distributed variables and Wilcoxon signed rank test for non-normally distributed variables. The mean difference and 95% confidence intervals (CIs) were calculated. Comparisons of qualitative variables at T0 and T1 were performed using the McNemar test. A p-value < 0.05 indicated statistical significance. The data were analyzed using IBM SPSS for Windows (version 26.0). Outcome assessment Periodontal indices The probing depth was defined as the distance from the gingival margin to the base of the sulcus or periodontal pocket within a normal range of (1–3 mm). The bleeding Index was used to evaluate bleeding as follows: Only one bleeding point appears. Several isolated bleeding points or a small blood area appear. Interdental triangle filled with blood soon after probing. Profuse bleeding when probing, blood spreads toward the marginal gingiva. The mobility index was used to assess mobility by scores as follows : Tooth mobility is perceptible but less than 1 mm buccolingually. Mobility between 1 and 2 mm Mobility exceeds 2 mm buccolingually or vertically The plaque index was used to assess the amount of plaque on teeth with the aid of plaque-disclosing tablets . (Biofilm Disclosing tablets (EMS) – Guided Biofilm Therapy, Biofilm Disclosure). 0. No plaque in the gingival area. 1. Separate flecks of plaque at the cervical margin of the tooth. 2. A thin continuous band of plaque (up to 1 mm) at the cervical margin. 3. Abundance of soft matter within the gingival pocket and/or on the tooth and gingival margin. The gingival index was used to assess the condition of the gingiva according to the following score : 0 Normal gingiva with slight color change, and slight edema. There was no change in probing. Mild inflammation Moderate inflammation redness, edema, and glazing. Bleeding on probing. Severe inflammation marked redness, edema, and ulceration. Tendency to spontaneous bleeding. Quality of life assessment The orthodontic treatment questionnaire is composed of 14 questions that evaluate the patient’s response (Yes/No/not know) to the method of orthodontic treatment which was the orthodontic retention phase in this research (Fig. ) Acceptance of orthodontic appliance scale : This scale consists of 10 incomplete statements, that need to be completed based on the patient’s choice. The available answer choices were scored using a 6-point Likert scale. To help patients understand the answer items, each answer was accompanied by a matching facial expression. Scores of 5 to 0 were allocated to the answer choices from left to right. Higher scores indicate greater acceptance and satisfaction with the respective item. The total score of this questionnaire ranged from 0 to 55. A higher total score indicated that the problems associated with using the removable orthodontic appliance were better accepted by the patient and reflected privileged motivation to continue the treatment. (Fig. ) The Oral Health Impact Profile (OHIP-14) is a valid and reliable instrument for assessing oral health-related quality of life among the adult population. The responses are rated on a 5-point Likert scale: 0 = never; 1 = hardly ever; 2 = occasionally; 3 = fairly often; 4 = very often/every day. The OHIP-14 scores can range from 0 to 56 and are calculated by summing the ordinal values for the 14 items. (Fig. ) Patients who had just finished the orthodontic fixed appliance phase (at least one year of treatment) with extraction or non-extraction treatment and scheduled to start retention. Patients with active periodontal disease , systemic disease, bone disease, craniofacial syndromes, cleft, active transverse palatal expansion, malformation, abnormal surface or morphological tooth structure, or restorations were excluded from the study. The sample size was estimated assuming a 5% alpha error and 80% study power. The sample size was adjusted to a 95% confidence interval (95% CI) to detect changes in probing depth after fixed retainer use. Salvesen et al. reported a mean (SD) probing depth of 2.9 (0.7) mm with a calculated 95% confidence interval = 2.41, 2.99. The required sample size was calculated to be 25 patients, which was increased to 28 to compensate for patients lost to follow-up. The sample size was calculated using MedCalc Statistical Software version 19.0.5 (MedCalc Software bvba, Ostend, Belgium; https://www.medcalc.org ; 2019). First, the study procedures were thoroughly explained to both the participants and their parents and, informed consent was subsequently obtained from each enrolled subject. At the T0 baseline, patients had phase one nonsurgical periodontal therapy (full mouth supragingival and subgingival scaling, root planing and polishing with eugenol-free paste followed by proper oral hygiene instructions (using a toothbrush, dental floss, and interdental brush) before bonding the FR. At T0 (after bracket debonding, and just before bonding the FR) a full PDL assessment was carried out for the maxillary dentition including (the probing depth, bleeding index, gingival index, mobility index, and plaque index) An impression was made at T0 to fabricate a removable retainer in the case when significant changes occurred during follow-up, which might necessitate immediate study termination, (Futility point) and the use of this retainer is to restore the T0 state. Several measures have been taken to ensure high bond strength while overcoming the high rate of failure of maxillary FR. First, pumice polishing was done for all surfaces to be bonded , followed by a sodium hypochlorite swab for 1 min (Sodium Hypochlorite 5% mint flavor, JK dental vision, Egypt) then acid etching by phosphoric acid 36% for 15 s, along with rinsing etchant surface same amount of time and gentle drying . Pre-hydrolyzed no-mix silane primer and Silane coupling agents (BISCO PORCELAIN PRIMER, BISCO, USA) were added to all surfaces to be bonded . The next step was, Bonding agent application (ASSURE® PLUS, Reliance Orthodontic Products, USA). Holding of the FR was done with the help of dental floss then direct adaptation and festooning of Dead soft wire, 0.027 × 0.011-inch ribbon arch-wire, 8-strand braided wire (FR) (Bond-A-Braid® Lingual Retainer Wire, Reliance Orthodontic Products, USA) from the palatal surface of right premolar to left premolar including the palatal surface of all maxillary anterior teeth in passive state away from the line of occlusion. (Fig. ) The flowable light-curing composite was applied (Polofil® NHT flowable composite light-curing, voco, Germany). Curing for 3 s using high-intensity LED was carried on.( light intensity 2300 mW/cm²)10 W . After finishing the whole curing for all units, selective grinding of excess composite or any interference between FR and lower teeth was done using articulating paper followed by polishing all composite surfaces eliminating any rough area. Details of oral hygiene instructions were provided, including a thorough explanation of the flossing technique to be used with the retainer in place, in addition to guidance on utilizing interdental brushes and water flossers. Patients were given the questionnaires and they filled them out directly after bonding with the FR to assess their quality of life and experience with the extended FR. Patients were followed up regularly each month and were asked to urgently to schedule an appointment if they felt any detachment in the FR. After 12 months of follow-up(6) (T1), all the previous records were repeated with periodontal assessment. Patients were given the quality-of-life questionnaires again to fill them. Normality was checked for all variables using descriptive statistics (mean, median, and standard deviation), plots (Q-Q plots and histogram), and normality tests. Means and standard deviations were calculated for quantitative variables, while frequencies and percentages were calculated for qualitative variables. Comparisons of quantitative variables at T0 and T1 were performed using paired t-test for normally distributed variables and Wilcoxon signed rank test for non-normally distributed variables. The mean difference and 95% confidence intervals (CIs) were calculated. Comparisons of qualitative variables at T0 and T1 were performed using the McNemar test. A p-value < 0.05 indicated statistical significance. The data were analyzed using IBM SPSS for Windows (version 26.0). Periodontal indices The probing depth was defined as the distance from the gingival margin to the base of the sulcus or periodontal pocket within a normal range of (1–3 mm). The bleeding Index was used to evaluate bleeding as follows: Only one bleeding point appears. Several isolated bleeding points or a small blood area appear. Interdental triangle filled with blood soon after probing. Profuse bleeding when probing, blood spreads toward the marginal gingiva. The mobility index was used to assess mobility by scores as follows : Tooth mobility is perceptible but less than 1 mm buccolingually. Mobility between 1 and 2 mm Mobility exceeds 2 mm buccolingually or vertically The plaque index was used to assess the amount of plaque on teeth with the aid of plaque-disclosing tablets . (Biofilm Disclosing tablets (EMS) – Guided Biofilm Therapy, Biofilm Disclosure). 0. No plaque in the gingival area. 1. Separate flecks of plaque at the cervical margin of the tooth. 2. A thin continuous band of plaque (up to 1 mm) at the cervical margin. 3. Abundance of soft matter within the gingival pocket and/or on the tooth and gingival margin. The gingival index was used to assess the condition of the gingiva according to the following score : 0 Normal gingiva with slight color change, and slight edema. There was no change in probing. Mild inflammation Moderate inflammation redness, edema, and glazing. Bleeding on probing. Severe inflammation marked redness, edema, and ulceration. Tendency to spontaneous bleeding. The probing depth was defined as the distance from the gingival margin to the base of the sulcus or periodontal pocket within a normal range of (1–3 mm). The bleeding Index was used to evaluate bleeding as follows: Only one bleeding point appears. Several isolated bleeding points or a small blood area appear. Interdental triangle filled with blood soon after probing. Profuse bleeding when probing, blood spreads toward the marginal gingiva. The mobility index was used to assess mobility by scores as follows : Tooth mobility is perceptible but less than 1 mm buccolingually. Mobility between 1 and 2 mm Mobility exceeds 2 mm buccolingually or vertically The plaque index was used to assess the amount of plaque on teeth with the aid of plaque-disclosing tablets . (Biofilm Disclosing tablets (EMS) – Guided Biofilm Therapy, Biofilm Disclosure). 0. No plaque in the gingival area. 1. Separate flecks of plaque at the cervical margin of the tooth. 2. A thin continuous band of plaque (up to 1 mm) at the cervical margin. 3. Abundance of soft matter within the gingival pocket and/or on the tooth and gingival margin. The gingival index was used to assess the condition of the gingiva according to the following score : 0 Normal gingiva with slight color change, and slight edema. There was no change in probing. Mild inflammation Moderate inflammation redness, edema, and glazing. Bleeding on probing. Severe inflammation marked redness, edema, and ulceration. Tendency to spontaneous bleeding. The orthodontic treatment questionnaire is composed of 14 questions that evaluate the patient’s response (Yes/No/not know) to the method of orthodontic treatment which was the orthodontic retention phase in this research (Fig. ) Acceptance of orthodontic appliance scale : This scale consists of 10 incomplete statements, that need to be completed based on the patient’s choice. The available answer choices were scored using a 6-point Likert scale. To help patients understand the answer items, each answer was accompanied by a matching facial expression. Scores of 5 to 0 were allocated to the answer choices from left to right. Higher scores indicate greater acceptance and satisfaction with the respective item. The total score of this questionnaire ranged from 0 to 55. A higher total score indicated that the problems associated with using the removable orthodontic appliance were better accepted by the patient and reflected privileged motivation to continue the treatment. (Fig. ) The Oral Health Impact Profile (OHIP-14) is a valid and reliable instrument for assessing oral health-related quality of life among the adult population. The responses are rated on a 5-point Likert scale: 0 = never; 1 = hardly ever; 2 = occasionally; 3 = fairly often; 4 = very often/every day. The OHIP-14 scores can range from 0 to 56 and are calculated by summing the ordinal values for the 14 items. (Fig. ) The demographic data for all patients ( n = 28) are shown in Table . All PDL indices were measured by two different examiners. Interexaminer reliability was calculated and the intraclass correlation coefficient (ICC) ranged from 0.940 to 0.999, indicating good to excellent agreement between the examiners. All the periodontal indices shown in Table , for the 8 maxillary teeth bonded to the FR significantly improved after 12 months of follow-up. The index that showed the most significant changes was probing depth. Both the buccal and palatal measurements showed a substantial decrease in probing depth, with mean decreases of -0.81 mm P < 0.001, CI(-0.99, -0.63) and − 0.89 mm p < 0.001 CI(-1.03, -0.75) respectively. The average probing depth also showed a significant reduction of -0.85 mm p < 0.001, CI (-1.01, -0.70). The whole maxillary dentition, also showed significant improvement as shown in Table . For the quality-of-life assessment, the OHIP-14 total score was used. Table shows that the mean score at T0 was 4.07 ± 4.60 and it decreased to 2.21 ± 2.57 at T1. This difference was statistically significant ( p = 0.04). The median score at T0 was 2.00 (IQR = 1.25, 4.00), and it remained the same at T1, with a median of 2.00 (IQR = 2.00, 2.00). The difference in medians was 0.00 (IQR=-3.00,0.75). For the acceptance of the orthodontic appliance scale, the mean score at T0 was 49.25 (SD = 0.80), and it increased slightly to 49.93 (SD = 0.26) at T1. The difference between T1 and T0 was statistically significant with a p-value < 0.001. The median score at T0 was 49.00 (IQR = 0.80) and it increased to 50.00 (IQR = 5.00, 50.00) at T1. The difference in medians was 0.50 (IQR = 0.00, 1.00). Table shows that most of the participants experienced no negative effects on speech, function, or quality of life during the retention period. For example, there were no reported problems with teeth or mouth when speaking post-retention, and all participants reported face and smile satisfaction. There were also significant improvements in self-esteem, self-confidence, social acceptance, emotional well-being, and relationships. Before retention, only a few participants reported high self-esteem, self-confidence, and social acceptance. However, after retention, nearly all participants reported high levels of these factors. Additionally, there were no negative effects on tooth alignment or reported discomfort during eating. The orthodontic fixed retainer has long been considered an indispensable part of orthodontic treatment, ensuring the stability of dental occlusion and preventing relapse. Several previous studies , – have shown that FR has some negative effects on PDL due to difficulty in maintaining oral hygiene. However, most of these studies have evaluated the effects on the mandibular arch only , – that is why authors picked the maxillary arch to evaluate the PDL response and speech is more affected by maxillary arch. There is a substantial variation in the histological structure of both maxillary and mandibular mucosa. The palatal mucosa consists of two constant and homogeneous layers; ortho-keratinized squamous epithelium, and lamina propria composed of dense connective tissue. These layers increase the resistance of the palatal mucosa to periodontal diseases such as inflammation propagation, and recession , unlike nonkeratinized mandibular lingual mucosa which is more prone to inflammation propagation especially when there is a plaque retentive area such as the FR , . The authors of this study hypothesized that the addition of an extra two units to 6-unit FR would increase the root surface area of the bonded segment resisting unwanted tooth that can occur with a canine-to-canine fixed retainer, even while the retainer is in place movement with less impact on maxillary palatal periodontium. Since extended maxillary FR showed good stability with clinical insignificant unwanted tooth movement , authors wanted to check its impact on PDL and quality of life. According to the results of the present study, the periodontal condition of all the maxillary teeth significantly improved after the 12-month follow-up period. These findings highlight that extended FR use has no harmful effects on the periodontium. The positive results obtained in the current study may be attributed to the removal of orthodontic fixed appliances, which could have acted as a local factor that might have compromised the periodontium. In addition, the debonding of the brackets was followed by full-mouth scaling and polishing along with oral hygiene measures. These findings are consistent with the results of previous studies , , , , which showed that orthodontic fixed retainers were found to be compatible with periodontal health and did not have severe detrimental effects on the periodontium. A long-term follow-up study , revealed that both fixed and removable retainers were associated with similar levels of gingival inflammation. Optimal oral hygiene before, during, and after orthodontic treatment is essential for preventing increased levels of gingival inflammation. Additionally, Han tested the effects of fixed retainers on periodontally compromised patients. Despite bonding fixed retainers in periodontitis patients, periodontal health was well maintained when supportive periodontal treatment and oral hygiene education were provided. Thus, it is crucial to emphasize the importance of optimal oral hygiene, supportive periodontal treatment, and patient education in maintaining periodontal health. Teeth stability and decreased mobility were significantly improved by using an extended fixed retainer which was consistent with the findings of Josef Kučera , who concluded that FR reduces the increase in tooth mobility caused by orthodontic treatment to normal levels. The values of tooth mobility after the placement of retainers were within the range of physiologic tooth mobility, thus they can even be used with patients who have mobile teeth as a means of splinting and retention since they have no significant damaging effects on the PDL. Speech, as a fundamental mode of communication and expression plays a significant role in daily life. Any alterations to speech patterns can have a profound impact on individuals’ interactions, affecting social relationships, confidence, and self-esteem. In most cases, speech problems are temporary with removable retainers, varying in duration from a few days to a few months . These changes are temporary and take a period of adaptation ranging from 1 week to 3 months , which in turn may affect patient self-esteem and quality of life. This may push the patient to be incompliant with the removable retainer causing relapse, in this study results, there were no reported speech alterations or problems. Quality of life encompasses a broad range of physical, emotional, and social aspects, making it a key indicator of treatment success. Although numerous studies have investigated the impact of orthodontic treatment on quality of life, few have specifically considered the effects of fixed retainers. Gaining insight into how fixed retainers affect various dimensions of daily life, including eating, oral hygiene maintenance, and overall satisfaction, represents an important step toward a more comprehensive understanding of the orthodontic patient experience. Patient satisfaction, as an endpoint of successful orthodontic treatment, can determine the ultimate perception of treatment outcomes. Despite their established utility for maintaining dental alignment, fixed retainers have been associated with specific inconveniences, such as difficulty with oral hygiene practices and occasional breakages. Evaluating patient satisfaction with fixed retainers, and identifying factors that contribute to a positive or negative patient experience, can guide both clinicians and patients in decision-making processes and treatment planning. Overall, the data in Table indicate that FR has no negative impact on patient quality of life or acceptance of orthodontic appliances. The reduction in the OHIP-14 scores signifies a decreased burden of oral health problems, which can lead to improved overall well-being and satisfaction. Additionally, the increase in acceptance scores indicates that patients not only adjust well to wearing fixed retainers but also perceive them as valuable components of their orthodontic treatment. It is important to note that although statistically significant, the observed differences in both quality of life and acceptance of orthodontic appliance scores might be considered relatively small in clinical terms. However, even small changes can be meaningful to patients, and the statistically significant findings highlight the positive impact of fixed retainers on patients’ experiences. Therefore, the extended maxillary fixed retainer can be used effectively in the maxillary arch without compromising PDL condition, splinting PDL-compromised teeth, eliminating the need for the patient compliance to a removable retainer, enhancing patient quality of life, and eliminating any speech problems. Although these findings provide valuable insights into the effects of FR on the quality of life, it is crucial to interpret them within the context of this study’s limitations. This research only involved the maxillary arch FR- which is mainly responsible for the deterioration of speech – in an attempt to standardize the sample and confounding factors. the sample size and specific patient characteristics may impact the generalizability of the results. Further research with a larger and more diverse population is warranted to confirm and expand these findings. The addition of an extra two units to a 6-unit FR was challenging since a 6-unit FR is known to have a high failure rate . Several precautions were carried on to overcome this problem. To ensure high bond strength several measures have been taken to provide that. Starting with pumice polishing enhances shear bond strength . Followed by Sodium hypochlorite 5% swap for 1 min to all surfaces to be bonded to enhance bond strength and remove organic pellicle of dental plaque . Pre-hydrolyzed silane primer was added to etched surfaces to enhance the bonding strength . Patients were instructed to regularly follow the integrity of the retainer while brushing, avoid any extra hard food that might break the retainer and if any loss of integrity took place, the patient must immediately ask for an appointment to fix it. Despite using several measures to decrease the failure rate, two cases have experienced breakage of the FR, and the patients presented for repair with the same bonding protocol after removal of the composite attached to the broken parts and tooth without compromising retainer integrity. Future Longer-term longitudinal follow-up studies are needed to test the long-term effects of the extended FR and the generalizability of the results with larger samples. Further studies are needed to compare the effects of the extended FR against dual retention, 6-unit FR alone, and removable retainers alone since this study is only a single arm study and blinding could not have been possible. Studies regarding the failure rate of extended FR are needed. After a 12-month follow-up period, with the limitations of the study, an eight-unit extended maxillary FR may not have adverse effects on maxillary periodontium and may be used as a means of retention without compromising PDL. An eight-unit maxillary retainer can be used as a splint to decrease tooth mobility associated with orthodontic tooth movement. Fixed retainers maintain high patient quality of life and acceptance of orthodontic appliances and positively influence patient attitudes toward orthodontic treatment. |
Single-cell RNA-seq reveals disease-specific CD8+ T cell clonal expansion and a high frequency of transcriptionally distinct double-negative T cells in diabetic NOD mice | 179ddc9e-eeb7-4fde-9d12-43b2ec3aca69 | 11922263 | Digestive System[mh] | The NOD mouse model has been demonstrated to be a translational model for T1D in humans, sharing many common markers of disease progression. Similar to humans, the onset of T1D in the NOD mouse is due to destruction of pancreatic β-cells in a multi-stage progression of autoimmune reactions. Islets remain free from lymphocytic infiltration in NOD mice until 3-4 weeks of age ,. At around 12 weeks, self-tolerance becomes broken due to the imbalance of regulatory T cells (Tregs) and effector CD8 + T cells . Insulitis in NOD mice involves many types of lymphocytes and myeloid cells . However, the T1D pathogenesis in this model is primarily driven by T cells [ – ]. It has been previously shown that CD4 + and CD8 + SP T cells are both needed for autoimmune destruction of β-cells [ – ]. β-cell reactive T cells that target autoantigens like GAD, IGRP, and insulin antigen are commonly seen in T1D development [ – ]. Other cell types such as B cells, dendritic cells, macrophages, and NK cells can also be found at the site of insulitis . Once formed, insulitic lesions are continually replenished by new lymphocytes from peripheral circulation . Therefore, T cell recruitment to the islets is considered a continuous process. Development of T1D in a mouse can be predicted from the detection of circulating autoreactive T cells in peripheral blood [ , , ]. T cell receptor (TCR) repertoire profiling is a powerful tool for identifying T cell clones and phenotypes directly linked to insulitis and β-cell destruction. The majority of TCR profiling studies characterize circulatory T cell clones [ – ] with few studies looking at islets or other lymphoid organs, . Using this method, autoreactive T cells in blood and islets have been observed in subsets of memory T cells and T regs . Marrero et al. analyzed islet-infiltrating memory CD4 + T cells in prediabetic and recent onset diabetic NOD mice. They found many unique TCR clonotypes in islet-infiltrating CD4 + T cells and noted that TCR β repertoires were highly diverse at both stages of T1D development. Recent technologies have enabled profiling the transcriptome at single-cell resolution, which has resulted in the growth of single-RNA sequencing studies of T1D tissues over the last few years [ , – ]. Similar technologies enable the sequencing of TCR rearrangements at the level of individual clones, which can be collected concomitantly with RNA-sequencing data . In this study, we use single cell TCR and RNA sequencing (scTCR-seq and scRNA-seq) to profile the transcriptional activity of autoreactive clones in pancreatic islets and PBMCs of diabetic NOD mice. We identify an uncharacteristically high proportion of DN T cells in both islets and blood of NOD mice. We also find that invading CD8 + T cells in diabetic NOD mice acquire a more exhaustive phenotype than the same cells in non-diabetic NOD mice. In this analysis, we identify new NOD TCR clones that are disease specific and not found in other studies that can be explored in future experiments.
DN T cells are expanded in NOD mice and follow CD4/CD8 like lineages We conducted paired single-cell RNA sequencing (scRNA-seq) and TCR repertoire sequencing on T cells collected from the peripheral blood and pancreatic islets of NOD mice ( ). We monitored female NOD mice from the age of 3 weeks to 40 weeks for the natural onset of diabetes (two consecutive blood glucose readings ≥ 250 mg/dl). Diabetic onset occurred in 70% of NOD mice by 40 weeks of age without any significant difference in body weight between diabetic and non-diabetic groups ( - , S1 Table in ). There was no significant difference in T cell frequency in diabetic and non-diabetic mice blood ( , S2 Table in ). T cells were enriched from PBMC and pancreatic islets and pooled together into six pools of 3 phenotypically similar mice before being sent for single cell library preparation (S1 Table in ). In addition, four pooled PBMC samples from non-diabetic mice were collected at 34 weeks of age and included for comparison with PBMC of diabetic mice. T cell extractions from non-diabetic mouse pancreatic islets resulted in too few cells for scTCRseq. As a result, we obtained sequencing reads from a total of 102,848 high quality peripheral blood and islet derived T cells from both diabetic and non-diabetic mice (S2 Table in ). Among filtered cells, 97% positively expressed CD3e using Seurat standards for positive expression. T cells were clustered based on variable gene expression and immune marker weighted principal components that increased the biological variability observed between cells ( a-d). In total, 32 clusters were identified ( ) and annotated based on expression of immune markers in each cell ( e, S3 Table in ) and pseudobulk expression patterns ( , , S4 Table in ). Expression patterns of the most common T cell markers revealed surprising results about the T cell compartment ( - ). Primarily, an uncharacteristically high rate (~33%) of double negative (DN) T cells in all tissues ( ), which lacked expression of either CD4 or CD8. Abnormal proliferation of DN T cells in NOD mice are consistent with previous reports . DN populations were confirmed not to express γδ receptor or NKT markers ( a). There was also no correlation between CD3 expression, nCount or nFeatures and CD4/CD8 expression, suggesting that the identification of significant amounts of DN T cells was not due to sampling/sequencing error ( b-i). Clusters seemed to predominantly represent the effector, memory, and exhaustive states of invading and circulating T cells. Within effector populations, less than 20% of cells expressed markers associated with known subsets of CD4 + T cells, however, of that 20% there was greater enrichment of Th1-like CD4 + T cells in islet infiltrating populations ( , a-c). We applied pseudotime trajectory analysis through Monocle3 which revealed a linear relationship from naive to effector cells with memory T cells developing throughout maturation ( , d). According to the trajectory, it may be likely that there is transition between SP and DN cells at both the naive and effector states. Effector T cells correlated highly with pancreatic islets while phenotypically exhausted T cells, which were higher in the blood, were found to occur at later stages in pseudotime differentiation ( e-f). Diabetic mice showed higher enrichment of exhausted (Texh) and pre-exhausted (Pexh) DN T cells in the blood, but equivalent amounts of exhausted CD4 + SP/CD8 + SP T cells and CD8 + Short Lived Effector Cell-like (SLEC-like, Il7r-/Gzmb+) cells ( ). DN T cell populations change significantly during diabetic progression in the NOD mouse To validate elevated levels of DN T cells in the NOD mice we performed flow cytometry and meta-analysis of comparable datasets. First, we stained pooled splenocytes and thymocytes from 24 week non-diabetic and diabetic NOD mice with markers for CD3, CD4, CD8b, FasL, B220, TCRB, and NK1.1 ( a). We were surprised to see that CD3 + DN splenocytes made up greater than 25% of CD3 + T cell populations in the spleen ( b). Consistent with reports from MRL/lpr mice , > 30% of the DN T cells were B220 + , however, they did not express more FasL or Fas than CD4 + or CD8 + T cells. As expected, none of these DN T cells bore the NK1.1 marker for NKT cells, however we did observe reduced MFI for TCRβ on the cell surface. We have confirmed through additional analysis that this did not correlate with an increase in gamma delta TCR abundance ( c-d). Thymic populations had relatively fewer CD3 + DN T cells (<20%), however the populations were still largely B220 + and TCRβ low but had greater expression of FasL and NK1.1 than CD4 + and CD8 + populations ( e,f). We were interested to see how populations of DN T cells changed over time in diabetic mice, so we conducted flow cytometry on paired blood samples from diabetic and non-diabetic mice used in single cell analysis at 4 different time points ( ). We have determined that the DN T cell compartment in the blood is negatively correlated with CD4 + populations but not correlated with CD8 + populations that increase steadily over a 24-week period ( ). Diabetic condition results in a shift from mostly CD4 + T cells to mostly DN T cells by week 24, whereas CD4 + T cells remain slightly more abundant throughout the same time frame in non-diabetic mice ( ). Normalizing diabetic mice by hyperglycemic onset (date at ≥ 250 mg/dL) shows an opposing sinusoidal trend between CD4 + and DN T cells in the blood resulting in a dip in CD4 + T cells and spike in DN T cells approximately 30 days before detectable increases in blood sugar ( ). This reversal was not present to the same extent in non-diabetic mice. We cannot conclude about the proportion of DN T cells in the pancreas over the NOD lifespan because they must be sacrificed at only one time point. However, the total number of circulating T cells remains level which means they must be coming from activated splenic populations or recirculated from inflammatory tissues. In a similar NOD mouse scRNA-seq dataset , we identified similar proportions of DN T cells that were not previously identified in that study (Fig 3e-i). Of the T cells misidentified in the previous study, more DN T cells were misattributed as CD4 + T cells than CD8 + ( a). Genes with distinct expression patterns in our dataset had similar patterns of expression and previously annotated resting Tregs were found to be analogous to the DN Treg cluster in our dataset ( b, c). High enrichment of lipid metabolism genes in islet localized effector T cells Differential expression analysis shows few differences between comparable populations of DN and CD4 + /CD8 + T cells (S5 and S6 Tables in ). Effector cell transition for DN, CD4 + , and CD8 + T cells resulted in marked upregulation of many common genes including Carboxypeptidase a1 ( Cpa1 ), Cpb1 , Carboxyl Ester Lipase ( Cel ), and Chymotrypsin c ( Ctrc ) which are involved with metabolic pathways that are attributed to pancreatic inflammation (S7 and S8 Tables in ). Progression to Texh phenotypes and removal from pancreatic islets for DN and CD4 + cells led to upregulation of markers associated with immunosuppression and exhaustion like TIGIT and CTLA4 (S9 Table in ) and CD8 + T cells gain expression of Gzma , Gzmb , Killer cell Lectin-like Receptor G1 ( Klrg1) , Sphingosine-1-phosphate receptor 5 ( S1pr5 ), Zinc finger E-box-binding homeobox 2 ( Zeb2 ), Galectin-3 ( Lgals3 ), and Glucagon ( Gcg ), and lost expression of Ctrb1 , CXC Chemokine Receptor 5 ( Cxcr5 ), and CD27 (S10 Table in ). We observed expected expression patterns of common T cell genes like IL2ra, GATA3, and Foxp3 ( a). Expression of the survival signal marker IL7R decreased in Tpex and Tex populations ( ) and to a greater degree in DN clusters ( ) which is consistent with previous results. Increased expression of Klrg1, Gzma, and Perforin (Prf1), and decreased expression of IL7r suggest that effector CD8 + T cells become Short Lived Effector Cells (SLEC) and accumulate in the blood due to an increase in expression of S1pr5 ( ). Expression of immune checkpoint proteins like Programmed Death 1 ( PD1 ), TIGIT , Cytotoxic T-Lymphocyte Associated protein 4 ( CTLA4 ), and T cell Immunoglobulin and Mucin domain 3 ( TIM3 ) were increased in exhausted populations ( b). We confirmed that loss of IL7R and increase of exhaustion markers are correlated with aromatic and organic cyclic pathway gene expression (S11 and S12 Table in ). Clusterwise Gene Set Enrichment Analysis (GSEA) ( ) led to the discovery that DN specific resting Tregs and DN effector cells experience increased expression of Hypoxia Induced Factor 1a (HIF1a) and the Glucocorticoid receptor (NR3C1) ( c), and CD4 + SP/DN effector cells are heavily enriched for lipid metabolism genes. A module of coexpressed genes specific to granzyme regulation (MEYellow) is found enriched and expressed at higher rates in islet infiltrating T cells ( ). The yellow module was highly correlated with zymogen activation genes, which is seen in suppression of CD8 + GzmA/B expression, and lipid metabolism genes ( , d). The majority of lipid metabolism gene expression was due to Cel and Pancreatic Lipase Related Protein (PNLIP) expression which was specific to CD3 + T cells ( e). Clonal specificity of KLRG1 + IL7R- CD8 + T cell development in diabetic mice results in Gzma + cytotoxic T cells trafficked out of the pancreatic islets. V(D)J profiling resulted in 85,321 unique α and β TCR sequences across 92,180 cells (S13 Table in ). There were 82,008 TCR sequences found in only a single cell, 990 sequences found in 3 or more cells, and 79 sequences found in more than 10 cells ( a-b). The largest number of cells found with a single TCR was 75. Of TCRs with greater than 3 cells, 457 were found in at least 2 different samples and 130 were found across diabetic and non-diabetic mice ( c-d). Highly expanded clones (>10 cells) clustered into 3 groups, CD8 + Gzma-, CD8 + Gzma + , and effector CD4 + T cells ( e). We found that neither diabetic tissue had many clones that were in the Gzma + group which were also found to be Klrg1 + upon differential expression analysis ( f and S14 Table in ). CD8 + cells that are Gzma + and Klrg1 + are often considered to be Short Lived Effector Cells (SLEC) so we will consider these to be SLEC-like. The diabetic tissues were more enriched in non-SLEC like Il7r- Gzma- CD8 + T cells and DN/CD4 + SP T effector cells that were upregulated in exhaustion markers like Ctla4, Lag3, and Tigit ( g and S4 Table in ). Analyzing paired blood and islet samples from diabetic NOD mice, we were able to trace genotype specific clones of islet-infiltrating T cells in the blood ( ). We found a total of 462 TCR clones belonging to infiltrating T cells (S15 Table in ). This resulted in 2,118 total T cells infiltrating TCR sequences, with 51 of the cells from Nondiabetic PBMCs and 880 of the cells from Diabetic PMBCs. The greatest disparity between the diabetic and non-diabetic mice was the difference in the number of infiltrating Gzma positive and negative CD8 + T cells ( ). Using lineage tracing, we identified TCR specific clonal lineages that progressed to a SLEC-like phenotype instead of an exhausted phenotype which was confirmed by lineage specific monocle3 trajectories ( and a). Diabetic mice had nearly equal populations of SLEC-like and Exhausted CD8 + T cells while non-diabetic mice were over enriched for SLEC-like ( ). We found that exhausted CD8 + T cell clones were diagnosis specific and represent a population of 44 unique TCRs that are only exhausted in the disease condition ( ). Within the similar Collier et al., dataset, we found that infiltrating T cells were more likely to be exhausted than non-infiltrating, and diabetic infiltrating had higher amounts of exhausted CD8 + and DN T cells ( c). We found this consistent with our results which also showed an increase in exhaustion markers like Lag3 and TIGIT in infiltrating clones of diabetic mice ( ). Differential expression analysis of infiltrating T cells in diabetic mice revealed them to be upregulated for inflammatory genes like Ifng ( c, S16–S18 Tables in ). Because there is a question of the origin for the DN T cells, we sought to use clonally expanded cells to determine if there was transition occurring within single clones. Consistent with earlier evidence, we found that within clones, transition between T cell compartments was most likely to occur during DN and CD4 effector maturation with some transition occurring between naive cells ( d). To investigate whether identified T cell clonal populations were targeting previously known epitopes associated with diabetic progression we searched for exact CDR3 matches to the VDJdb, a database of experimentally validated TCR-pMHC interactions. We only found two hits of the expanded (>10) clones but among all clones we found 4,166 matching CDR3 sequences in the VDJdb (S19 Table in ). Of the known epitopes sources to the matching CDR3 sequences, the most common was to Influenza A PA (686 hits) and PB1 (392 hits) with the second highest target being Murine Cytomegalovirus (MCMV) M45 (724 hits) along with recognition of epitopes from Lymphocytic Choriomeningitis Virus (LCMV), Plasmodium Berghei , and Respiratory Syncytial Virus (RSV). There were much fewer self-targeted antigens with the majority specific to Myelin Basic Protein (MBP) and Protein Arginine Deaminase 4 (PADI4), we did not observe any specific to known insulin specific epitopes. While the majority of the antigen specific T cells were evenly distributed among tissues and clusters, we observed Influenza NP (ASNENMETM) targeting T cells had higher relative numbers in the pancreatic islets (S9 Fig e-f). V(D)J and immunoregulatory gene expression in T cell clones is associated with islet infiltration Because T cell infiltration appeared to occur in specific populations, we predicted the infiltration status of T cell clones based on gene expression. We found that a regularized logistic regression classifier achieved moderately high sensitivity and specificity ( - , Overall area under the curve [AUC] = 0.89 and overall AUPRC = 0.67). Finally, we identified a list of genes that were found to be associated with the matching status of the cells ( ). Of the top 15 genes used in the matching status prediction, 4 are associated with clonal specificity in V(D)J rearrangement ( Trav16n , Trbv29 , Trav8d -2, Trav7 -3), 6 are associated with immune cell function, and 5 are associated with ribosome function. We looked at specific expression of TCR genes among clusters to see if specific TCR gene expression related to invasion and diabetes ( ). We do find an interestingly high amount of gamma delta TCR expression in the exhausted/overactive CD8 + and DN TC1-like cells, although they remain < 5% of all T cells. Reduced overall expression of TCR genes in the DN population is consistent with the TCRβ low phenotype found through flow cytometry. Interestingly, the DN population has higher specificity for Trav15d-1 and both have higher expression of Trav16n which was identified to predict for islet infiltration. Trbv29 also seems to be more specific for CD8 + T cells. Effector DN T cells have relatively higher expression of Trav15n and Trav15d than their CD4 equivalent ( d). Based on entries in the VDJdb, Trav16n, which is focused to the IL7R - T effector cluster is specific to the Influenza A PA (SSLENFRAYV) and PB1 (LSLRNPILV) antigens while Trav15d expressed in IL7R + DN T effector cells has specificity to Plasmodium Berghei GAP50 (SQLLNAKYL).
We conducted paired single-cell RNA sequencing (scRNA-seq) and TCR repertoire sequencing on T cells collected from the peripheral blood and pancreatic islets of NOD mice ( ). We monitored female NOD mice from the age of 3 weeks to 40 weeks for the natural onset of diabetes (two consecutive blood glucose readings ≥ 250 mg/dl). Diabetic onset occurred in 70% of NOD mice by 40 weeks of age without any significant difference in body weight between diabetic and non-diabetic groups ( - , S1 Table in ). There was no significant difference in T cell frequency in diabetic and non-diabetic mice blood ( , S2 Table in ). T cells were enriched from PBMC and pancreatic islets and pooled together into six pools of 3 phenotypically similar mice before being sent for single cell library preparation (S1 Table in ). In addition, four pooled PBMC samples from non-diabetic mice were collected at 34 weeks of age and included for comparison with PBMC of diabetic mice. T cell extractions from non-diabetic mouse pancreatic islets resulted in too few cells for scTCRseq. As a result, we obtained sequencing reads from a total of 102,848 high quality peripheral blood and islet derived T cells from both diabetic and non-diabetic mice (S2 Table in ). Among filtered cells, 97% positively expressed CD3e using Seurat standards for positive expression. T cells were clustered based on variable gene expression and immune marker weighted principal components that increased the biological variability observed between cells ( a-d). In total, 32 clusters were identified ( ) and annotated based on expression of immune markers in each cell ( e, S3 Table in ) and pseudobulk expression patterns ( , , S4 Table in ). Expression patterns of the most common T cell markers revealed surprising results about the T cell compartment ( - ). Primarily, an uncharacteristically high rate (~33%) of double negative (DN) T cells in all tissues ( ), which lacked expression of either CD4 or CD8. Abnormal proliferation of DN T cells in NOD mice are consistent with previous reports . DN populations were confirmed not to express γδ receptor or NKT markers ( a). There was also no correlation between CD3 expression, nCount or nFeatures and CD4/CD8 expression, suggesting that the identification of significant amounts of DN T cells was not due to sampling/sequencing error ( b-i). Clusters seemed to predominantly represent the effector, memory, and exhaustive states of invading and circulating T cells. Within effector populations, less than 20% of cells expressed markers associated with known subsets of CD4 + T cells, however, of that 20% there was greater enrichment of Th1-like CD4 + T cells in islet infiltrating populations ( , a-c). We applied pseudotime trajectory analysis through Monocle3 which revealed a linear relationship from naive to effector cells with memory T cells developing throughout maturation ( , d). According to the trajectory, it may be likely that there is transition between SP and DN cells at both the naive and effector states. Effector T cells correlated highly with pancreatic islets while phenotypically exhausted T cells, which were higher in the blood, were found to occur at later stages in pseudotime differentiation ( e-f). Diabetic mice showed higher enrichment of exhausted (Texh) and pre-exhausted (Pexh) DN T cells in the blood, but equivalent amounts of exhausted CD4 + SP/CD8 + SP T cells and CD8 + Short Lived Effector Cell-like (SLEC-like, Il7r-/Gzmb+) cells ( ).
To validate elevated levels of DN T cells in the NOD mice we performed flow cytometry and meta-analysis of comparable datasets. First, we stained pooled splenocytes and thymocytes from 24 week non-diabetic and diabetic NOD mice with markers for CD3, CD4, CD8b, FasL, B220, TCRB, and NK1.1 ( a). We were surprised to see that CD3 + DN splenocytes made up greater than 25% of CD3 + T cell populations in the spleen ( b). Consistent with reports from MRL/lpr mice , > 30% of the DN T cells were B220 + , however, they did not express more FasL or Fas than CD4 + or CD8 + T cells. As expected, none of these DN T cells bore the NK1.1 marker for NKT cells, however we did observe reduced MFI for TCRβ on the cell surface. We have confirmed through additional analysis that this did not correlate with an increase in gamma delta TCR abundance ( c-d). Thymic populations had relatively fewer CD3 + DN T cells (<20%), however the populations were still largely B220 + and TCRβ low but had greater expression of FasL and NK1.1 than CD4 + and CD8 + populations ( e,f). We were interested to see how populations of DN T cells changed over time in diabetic mice, so we conducted flow cytometry on paired blood samples from diabetic and non-diabetic mice used in single cell analysis at 4 different time points ( ). We have determined that the DN T cell compartment in the blood is negatively correlated with CD4 + populations but not correlated with CD8 + populations that increase steadily over a 24-week period ( ). Diabetic condition results in a shift from mostly CD4 + T cells to mostly DN T cells by week 24, whereas CD4 + T cells remain slightly more abundant throughout the same time frame in non-diabetic mice ( ). Normalizing diabetic mice by hyperglycemic onset (date at ≥ 250 mg/dL) shows an opposing sinusoidal trend between CD4 + and DN T cells in the blood resulting in a dip in CD4 + T cells and spike in DN T cells approximately 30 days before detectable increases in blood sugar ( ). This reversal was not present to the same extent in non-diabetic mice. We cannot conclude about the proportion of DN T cells in the pancreas over the NOD lifespan because they must be sacrificed at only one time point. However, the total number of circulating T cells remains level which means they must be coming from activated splenic populations or recirculated from inflammatory tissues. In a similar NOD mouse scRNA-seq dataset , we identified similar proportions of DN T cells that were not previously identified in that study (Fig 3e-i). Of the T cells misidentified in the previous study, more DN T cells were misattributed as CD4 + T cells than CD8 + ( a). Genes with distinct expression patterns in our dataset had similar patterns of expression and previously annotated resting Tregs were found to be analogous to the DN Treg cluster in our dataset ( b, c).
Differential expression analysis shows few differences between comparable populations of DN and CD4 + /CD8 + T cells (S5 and S6 Tables in ). Effector cell transition for DN, CD4 + , and CD8 + T cells resulted in marked upregulation of many common genes including Carboxypeptidase a1 ( Cpa1 ), Cpb1 , Carboxyl Ester Lipase ( Cel ), and Chymotrypsin c ( Ctrc ) which are involved with metabolic pathways that are attributed to pancreatic inflammation (S7 and S8 Tables in ). Progression to Texh phenotypes and removal from pancreatic islets for DN and CD4 + cells led to upregulation of markers associated with immunosuppression and exhaustion like TIGIT and CTLA4 (S9 Table in ) and CD8 + T cells gain expression of Gzma , Gzmb , Killer cell Lectin-like Receptor G1 ( Klrg1) , Sphingosine-1-phosphate receptor 5 ( S1pr5 ), Zinc finger E-box-binding homeobox 2 ( Zeb2 ), Galectin-3 ( Lgals3 ), and Glucagon ( Gcg ), and lost expression of Ctrb1 , CXC Chemokine Receptor 5 ( Cxcr5 ), and CD27 (S10 Table in ). We observed expected expression patterns of common T cell genes like IL2ra, GATA3, and Foxp3 ( a). Expression of the survival signal marker IL7R decreased in Tpex and Tex populations ( ) and to a greater degree in DN clusters ( ) which is consistent with previous results. Increased expression of Klrg1, Gzma, and Perforin (Prf1), and decreased expression of IL7r suggest that effector CD8 + T cells become Short Lived Effector Cells (SLEC) and accumulate in the blood due to an increase in expression of S1pr5 ( ). Expression of immune checkpoint proteins like Programmed Death 1 ( PD1 ), TIGIT , Cytotoxic T-Lymphocyte Associated protein 4 ( CTLA4 ), and T cell Immunoglobulin and Mucin domain 3 ( TIM3 ) were increased in exhausted populations ( b). We confirmed that loss of IL7R and increase of exhaustion markers are correlated with aromatic and organic cyclic pathway gene expression (S11 and S12 Table in ). Clusterwise Gene Set Enrichment Analysis (GSEA) ( ) led to the discovery that DN specific resting Tregs and DN effector cells experience increased expression of Hypoxia Induced Factor 1a (HIF1a) and the Glucocorticoid receptor (NR3C1) ( c), and CD4 + SP/DN effector cells are heavily enriched for lipid metabolism genes. A module of coexpressed genes specific to granzyme regulation (MEYellow) is found enriched and expressed at higher rates in islet infiltrating T cells ( ). The yellow module was highly correlated with zymogen activation genes, which is seen in suppression of CD8 + GzmA/B expression, and lipid metabolism genes ( , d). The majority of lipid metabolism gene expression was due to Cel and Pancreatic Lipase Related Protein (PNLIP) expression which was specific to CD3 + T cells ( e).
V(D)J profiling resulted in 85,321 unique α and β TCR sequences across 92,180 cells (S13 Table in ). There were 82,008 TCR sequences found in only a single cell, 990 sequences found in 3 or more cells, and 79 sequences found in more than 10 cells ( a-b). The largest number of cells found with a single TCR was 75. Of TCRs with greater than 3 cells, 457 were found in at least 2 different samples and 130 were found across diabetic and non-diabetic mice ( c-d). Highly expanded clones (>10 cells) clustered into 3 groups, CD8 + Gzma-, CD8 + Gzma + , and effector CD4 + T cells ( e). We found that neither diabetic tissue had many clones that were in the Gzma + group which were also found to be Klrg1 + upon differential expression analysis ( f and S14 Table in ). CD8 + cells that are Gzma + and Klrg1 + are often considered to be Short Lived Effector Cells (SLEC) so we will consider these to be SLEC-like. The diabetic tissues were more enriched in non-SLEC like Il7r- Gzma- CD8 + T cells and DN/CD4 + SP T effector cells that were upregulated in exhaustion markers like Ctla4, Lag3, and Tigit ( g and S4 Table in ). Analyzing paired blood and islet samples from diabetic NOD mice, we were able to trace genotype specific clones of islet-infiltrating T cells in the blood ( ). We found a total of 462 TCR clones belonging to infiltrating T cells (S15 Table in ). This resulted in 2,118 total T cells infiltrating TCR sequences, with 51 of the cells from Nondiabetic PBMCs and 880 of the cells from Diabetic PMBCs. The greatest disparity between the diabetic and non-diabetic mice was the difference in the number of infiltrating Gzma positive and negative CD8 + T cells ( ). Using lineage tracing, we identified TCR specific clonal lineages that progressed to a SLEC-like phenotype instead of an exhausted phenotype which was confirmed by lineage specific monocle3 trajectories ( and a). Diabetic mice had nearly equal populations of SLEC-like and Exhausted CD8 + T cells while non-diabetic mice were over enriched for SLEC-like ( ). We found that exhausted CD8 + T cell clones were diagnosis specific and represent a population of 44 unique TCRs that are only exhausted in the disease condition ( ). Within the similar Collier et al., dataset, we found that infiltrating T cells were more likely to be exhausted than non-infiltrating, and diabetic infiltrating had higher amounts of exhausted CD8 + and DN T cells ( c). We found this consistent with our results which also showed an increase in exhaustion markers like Lag3 and TIGIT in infiltrating clones of diabetic mice ( ). Differential expression analysis of infiltrating T cells in diabetic mice revealed them to be upregulated for inflammatory genes like Ifng ( c, S16–S18 Tables in ). Because there is a question of the origin for the DN T cells, we sought to use clonally expanded cells to determine if there was transition occurring within single clones. Consistent with earlier evidence, we found that within clones, transition between T cell compartments was most likely to occur during DN and CD4 effector maturation with some transition occurring between naive cells ( d). To investigate whether identified T cell clonal populations were targeting previously known epitopes associated with diabetic progression we searched for exact CDR3 matches to the VDJdb, a database of experimentally validated TCR-pMHC interactions. We only found two hits of the expanded (>10) clones but among all clones we found 4,166 matching CDR3 sequences in the VDJdb (S19 Table in ). Of the known epitopes sources to the matching CDR3 sequences, the most common was to Influenza A PA (686 hits) and PB1 (392 hits) with the second highest target being Murine Cytomegalovirus (MCMV) M45 (724 hits) along with recognition of epitopes from Lymphocytic Choriomeningitis Virus (LCMV), Plasmodium Berghei , and Respiratory Syncytial Virus (RSV). There were much fewer self-targeted antigens with the majority specific to Myelin Basic Protein (MBP) and Protein Arginine Deaminase 4 (PADI4), we did not observe any specific to known insulin specific epitopes. While the majority of the antigen specific T cells were evenly distributed among tissues and clusters, we observed Influenza NP (ASNENMETM) targeting T cells had higher relative numbers in the pancreatic islets (S9 Fig e-f).
Because T cell infiltration appeared to occur in specific populations, we predicted the infiltration status of T cell clones based on gene expression. We found that a regularized logistic regression classifier achieved moderately high sensitivity and specificity ( - , Overall area under the curve [AUC] = 0.89 and overall AUPRC = 0.67). Finally, we identified a list of genes that were found to be associated with the matching status of the cells ( ). Of the top 15 genes used in the matching status prediction, 4 are associated with clonal specificity in V(D)J rearrangement ( Trav16n , Trbv29 , Trav8d -2, Trav7 -3), 6 are associated with immune cell function, and 5 are associated with ribosome function. We looked at specific expression of TCR genes among clusters to see if specific TCR gene expression related to invasion and diabetes ( ). We do find an interestingly high amount of gamma delta TCR expression in the exhausted/overactive CD8 + and DN TC1-like cells, although they remain < 5% of all T cells. Reduced overall expression of TCR genes in the DN population is consistent with the TCRβ low phenotype found through flow cytometry. Interestingly, the DN population has higher specificity for Trav15d-1 and both have higher expression of Trav16n which was identified to predict for islet infiltration. Trbv29 also seems to be more specific for CD8 + T cells. Effector DN T cells have relatively higher expression of Trav15n and Trav15d than their CD4 equivalent ( d). Based on entries in the VDJdb, Trav16n, which is focused to the IL7R - T effector cluster is specific to the Influenza A PA (SSLENFRAYV) and PB1 (LSLRNPILV) antigens while Trav15d expressed in IL7R + DN T effector cells has specificity to Plasmodium Berghei GAP50 (SQLLNAKYL).
T cell dysfunction has largely been shown to be the primary cause of diabetic progression in T1D patients and NOD mice . In this model both CD8 + and CD4 + T cells show autoimmune potential and self-antigen recognition that leads to beta cell destruction in pancreatic islets. To explore how TCR restriction leads to T cell dysfunction we conducted paired scRNA and scTCR sequencing of T cell populations in adult female NOD mice. Because we cannot detect beta cell destruction in vivo, we sampled mice as soon as beta cell destruction was detectable through hyperglycemia but may have been censored from changes that occur early in disease progression. A challenge we and many T cell researchers faced in single cell transcriptomics was a lack of meaningful biological variation between T cells for cluster assignment , which was evident due to high mixing of CD8 + and CD4 + T cells within clusters ( b,d). We replicated a technique we have used previously to increase biological diversity in single cell datasets , by weighting biologically relevant markers more heavily than variably expressed markers in principal components. This led to biologically relevant clusters that fit current expectations of T cell maturation; however, it did likely produce some artifacts detectable in the UMAP which was a complaint in the original study. T cell subsets representing Th1, Th2, Th17, etc… were not apparent in clustering and further identification in Teff clusters based on common markers (IFNg, IL-10, IL-4, etc…) revealed on average that less than 20% of effector cells fit these classifications based on expression ( b). These cell types may be hidden by the limitations of scRNA-seq where either low sampling produces zero inflated counts of marker genes or expression values do not reflect the production of subtype specific cytokines. The most surprising outcome was identification of an over enriched DN T cell subset that was calculated to be ~ 30% of all CD3 + T cells. We have confirmed experimentally that these are CD3 expressing T-cells and are not NKT or γδ subsets. The DN T cell populations change over time coinciding with disease development. In wild type mice and humans, DN T cells are estimated at only 1-5% of CD3 + T cells . However, increased DN populations have been reported in both NOD and MRL- lpr mice, a mouse model of inflammatory Systemic Lupus Erythematosus . Studies in the lupus prone MRL- lpr mice, which contain a spontaneous mutation in the Fas receptor, have connected lymphoproliferation of DN T cells to a reduced ability to remove defective thymic or peripheral T cells . The ability for the lpr and gld mutations to protect NOD mice from T1D correlated with our finding that diabetic mice undergo greater expansion of DN T cells, perhaps as protective mechanism of peripheral tolerance. The genetic similarities between these mouse models are limited as Fas has not been identified in any of the IDD loci and diabetic onset in NOD mice is likely polygenic . We have confirmed similarities between lpr DN and NOD DN T cells, including increased amounts of B220 expression, even though there is seemingly no change in expression of the CD45 gene. These similarities may hint at a common development pathway, including likely peripheral development according to the single cell trajectory and clonal lineage tracing. Double negative T regs have also been shown to protect the NOD mouse from diabetes in adoptive transfer studies more than CD4 + Tregs . Another potential reason that DN T cells in the blood spike more intensely prior to hyperglycemia is that they are removed locally from the pancreas more than CD4 + T cells to block these protective effects. While it is unclear the exact effect that DN T cells may have on disease progression, we do observe increased DN exhaustion in the diabetic mice. Through paired TCR sequencing, we observed the impact of CD8 + IL7R- KLRG1 + SLEC-like cytotoxic T cells in pancreas infiltrating lymphoid cells. We found that CD8 + T cell progression to SLEC-like or Texh phenotypes was clonally specific and leaned more towards Texh in the diabetic mouse. This might suggest that SLEC development is necessary for T1D protection which is consistent with reports that GzmA deficient NOD mice have enhanced disease development . The diabetic exhausted clones identified could also be autoreactive TCRs that are more actively suppressed in the diabetic condition. Similarly, we found that infiltration into the pancreatic islets could be predicted using certain components of V(D)J recombination. It has already been suggested that the components we identified (Trav16n, Trbv29, Trav8d-2) are linked to autoreactivity of insulin peptides in NOD mice . A majority of the genes specific to pancreatic islet localized T cells were related to lipid metabolism such as CEL and PNLIP, which have already been identified as mediators of pancreatic inflammation . We did not identify any insulin specific peptides as targets of our dataset TCR, possibly due to a combination of few NOD experiments in VDJdb and poor CDR3 recognition in the TCR sequencing. We did find some targets which could have further impact on diabetic research, most notably a higher relative pancreatic infiltration of Influenza A NP targeted T cells, which previous studies have linked viral infection to greater incidence of T1D . BLAST search of the epitope to self-antigenic proteins in mice resulted in a 100% match (NENMET) to Keratin 222 (KRT222).
We characterized and identified a single-cell level T cell map in the peripheral blood and islets of NOD mice during the progression of Type 1 diabetes. Diabetic mice were found to have shockingly high levels of circulating and invading DN T cells and increased exhaustion of potentially immunosuppressive T cell subsets ( ). DN T cell subsets increased during diabetogenesis suggesting that they proliferate more, die less, or are trafficked out of the pancreas more in concert with islet destruction. The primary difference in clonal expansion of diabetic mice was in CD8 + T cells. Diabetic mice showed greater expansion of clones that become exhausted as opposed to functionally active and SLEC-like. This is likely due to regulation by other immune cells in the pancreas (Tregs) or prior to infiltration of pancreatic islets. We identify previously uncharacterized clones associated with the regulation of CD8 + T cell exhaustion in the pancreas. It seems that in the NOD mouse, peripheral DN T cell proportion is a biomarker of diabetogenesis that can be used to detect diabetic onset as early as 14 weeks but whether a similar phenomenon is observed in humans must still be investigated.
Experimental mouse model We purchased three-week-old female NOD/ShiLtJ mice from The Jackson Laboratory (stock number 001976) and were monitored for natural diabetes onset until 40 weeks of age. The mice experiment was carried out in a specific-pathogen-free (SPF) facility at Joslin Diabetes Center (JDC) under standard housing, feeding, and husbandry. The mice experiment protocol was reviewed and approved by JDC’s Institutional Animal Care and Use Committee (IACUC) (Protocol #2016-05). In addition, ARRIVE guidelines were strictly followed while conducting the mice experiment. Briefly, approximately 200 µ L of blood was collected through the lateral tail every two weeks. At the time of organ collection, mice were euthanized by isoflurane inhalation before cervical dislocation. Blood glucose measurement and diabetes diagnosis Tail blood was used to measure glucose concentration. Every two weeks, blood glucose was measured by Infinity blood glucose test strips (GTIN/DI#885502-002000) and an Infinity meter. Mice showing two consecutive glucose readings of ≥ 250 mg/dl were considered diabetic . Blood lymphocyte collection and cryopreservation Peripheral blood was collected from mice every two weeks. Approximately 200 ul of blood from each mouse was collected from the tail using a heparin coated Microvette tube to prevent blood clots (Sarstedt catalog#16.443.100). PBMC’s were extracted using an equal volume of Histopaque 1083 (Sigma-Aldrich, Missouri, USA catalog#1083-1). Total cells count and viability were assessed on a Countess automated cell counter (Thermo Scientific,). After counting, the cell pellet was resuspended in 1 mL of Cryostor CS10 (Stemcell Technologies Catalog #07930) and frozen down for storage in liquid nitrogen. Single-cell collection from pancreatic islets and cryopreservation Mice sacrificed at the onset of diabetes prior to detection of symptoms to prevent suffering and were perfused with collagenase (Vitacyte, CIzyme catalog #005-1030) and dissected to remove the pancreas for further processing. The pancreatic tissue was hand-shaken vigorously for 5-10 seconds and centrifuged at 500xg for 1.5 minutes then filtered to remove the remaining undigested tissue, fat, and lymph nodes. The filtrate was centrifuged at 500xg for 1.5 min at RT, and the tissue pellet was resuspended in 10 mL lymphocyte separation media (LSM) (Corning, cat #25072-CV) before purifying by density gradient centrifugation. Non-islet tissues were removed from the islet suspension by microscopic examination. These purified islets were subjected to non-enzymatic single-cell dissociation. The cell suspension was filtered through a 70μm cell strainer and counted before cryopreservation. T cell preparation for single-cell RNA-sequencing PBMC and islets single-cell suspension preserved in liquid nitrogen were revived before the magnetic separation of CD3 + T cells. CD3 + T cells were isolated using the Miltenyi MACS CD3 microbead kit (Miltenyi catalog#130-094-973) following the manufacturer and 10x genomics recommended guidelines (10x Protocol CG000123 Rev B). The cell viability and count were performed before and after CD3 + T cell enrichment. The magnetically separated CD3 + T cells were placed on ice and immediately used for single-cell cDNA library preparation. Single-cell RNA and immune repertoire sequencing library preparation Single-cell 5’ gene expression (GEX) and V(D)J sequencing libraries of T cells were generated using the Chromium Next GEM Single Cell 5’ Dual Index Reagent Kits v2 (10x Genomics, Pleasanton, CA, USA) following the manufacturer guidelines (10X protocol CG000331 Rev A). Sequencing of 5’GEX and V(D)J libraries Gene expression and V(D)J libraries were sequenced on Illumina NovaSeq6000 sequencer by GENEWIZ (GENEWIZ, LLC, NJ, USA). Each of the 5’GEX and V(D)J libraries were prepared with unique dual indexes, which allows multiplexing of all libraries for sequencing. Libraries were sequenced according to the 10X Genomics configuration. A minimum of 20,000 read pairs per cell were sequenced for 5’GEX libraries, and a minimum of 5,000 read pairs were sequenced for V(D)J libraries. Sequence demultiplexing and processing Raw sequencing reads were processed using Cell Ranger version 6.0.2 to produce gene expression count matrices and TCR clonotype summary. Using the raw sequencing FASTQ files, we first run the cellranger multi pipelines for simultaneous processing and analysis of V(D)J and gene expression data. This pipeline generates single cell feature counts, V(D)J sequences, and annotations for a single library. The mouse reference genome mm10-2020-A was used for aligning gene expression sequences, and the mouse reference GRCm38 v5.0.0 was used for aligning V(D)J sequences. Analysis of single-cell gene expression data All gene expression analyses were performed using R version 4.1.0 and Seurat version 4.1.0 . Other R packages used during this analysis include dplyr version 1.0.8, patchwork version 1.1.1, data.table version 1.14.2, ggplot2 version 3.3.5, cowplot version 1.1.1, viridis version 0.6.2, gridExtra version 2.3, RColorBrewer version 1.1.2, and tibble version 3.1.6. Seurat object was created individually for all samples using the barcode.tsv, features.tsv and matrix.mtx files generated from the cellranger multi pipeline. The min.cells parameter was set to 3 and the min.features parameter was set to 200 during the creation of Seurat objects. Cells were filtered out from the Seurat object based on several quality control parameters. First, low-quality cells were removed based on CD3 gene expression (>0 counts), overall mitochondrial gene expression , and an aberrant high count of genes (200-5,000 features). Second, we filtered cells based on the expression of housekeeping genes (>0 counts). The quality-filtered gene expression data were normalized and scaled by Seurat function NormalizeData and ScaleData with default “LogNormalize” parameters, generating log-transformed gene expression measurement per 10,000 reads. Weighted PCAs were calculated with weights applied to relevant immune markers taken from multiple sources [ – ] and 400 most variable features. SCTransform integration was applied to normalize libraries and PC 1-10 were used for KNN based Umap and Louvain clustering generation. Clusters were selected from the 1.5 resolution and hand annotated for cell identity. Monocle3 was used to calculate pseudotime across clusters. Gating to identify CD4, CD8, and DN populations as well as Th subsets was done on normalized expression data using a linear gate between the median populations of cells (Supplementary Figure 4a) for each marker or a fixed value of 0.5. T cell clonal analysis Matching clones of T cells between blood and islets were determined based on the similarity of TCR sequences. We only included the T cells with at least one alpha and one beta chain in the clonal analysis. The cells were assigned to a particular clonotype if they shared the same amino acid sequence of TCR alpha and TCR beta sequence. This same definition was also applied to cells that carry multiple alpha and beta sequences. Cells having identical TCR alpha or beta sequences were defined as matching clones between blood and islets. For example, if a T cell in blood carries identical TCR alpha or TCR beta sequence to a T cell in islets, it is defined as islet-matching T cell in blood or blood-matching T cell in islets. The total count of cells in a particular clonotype was used to determine the clonal expansion. TCR clonotype data were added to the GEX Seurat object as metadata for integrated gene expression and TCR analysis. Predicted CDR3 sequences from T cells were aligned to the VDJdb database to query for exact matches. Weighted co-expression analysis Genes were clustered by similar patterns of expression across cells using the WGCNA package . Co-expression clusters were generated using a soft power of 10 and a minimum module size of 10, with a merge cut height of 0.15. Clusters were randomly assigned colors to label modules. Gene set enrichment analysis (GSEA) For GSEA, we downloaded the list of Hallmark pathways (v7.5.1) from the Molecular Signature Database (MSigDB, http://software.broadinstitute.org/gsea/msigdb/index.jsp ) and used the ranked list of differentially expressed genes between different cell types. R package fgsea version 1.20.0 was used for GSEA. Machine learning We first extracted the metadata, raw counts, and log normalized raw counts of gene expression from the integrated dataset to predict the matching status of cells based on gene expression profile. Then, we used logistic regression to make predictions with a liblinear solver and lasso penalty of 11. We used the Scikit-learn package in Python version 3.10.4 for the regression analysis and used matplotlib to generate the plots. Flow cytometry. For flow cytometry analysis, T-cell enriched samples from the blood of 5 diabetic and 5 non-diabetic NOD mice at 6,14,20 and 24 weeks of age were stained using labeled primary antibodies against CD4 (GK1.5, Thermofisher #12-0041-82), CD3e (17A2, Thermofisher #48-0032-82), CD8a (53-6.7, Thermofisher #56-0081-82), and CD8b (H35-17.2, Thermofisher #11-0083-82), NK1.1 (PK136, Thermofisher #407-5941-82), TCRβ (H57-597, BioLegend #109225), B220 (RA3-6B2, BioLegend #103212), FasL (MFL3, Thermofisher #63-5911-82), Fas (15A7, Thermofisher #46-0951-82), CD19 (1D3, Thermofisher #M004T02Y03-A), and TCRγδ (GL-3, Thermofisher #25-5711-82). Before staining, all samples were incubated with Live Dead Aqua Viability (Thermofisher #L34957) stain for 10 minutes in dark at 4°C. Samples were run on Attune NxT four color laser flow cytometer compensated with unstained control and single stained UltraComp eBead compensation beads (Invitrogen, #01-2222-41). Statistical analysis All statistical analyses were performed with R or GraphPad Prism software, and P values < 0.05 were considered statistically significant. Figure legends show respective statistical analyses.
We purchased three-week-old female NOD/ShiLtJ mice from The Jackson Laboratory (stock number 001976) and were monitored for natural diabetes onset until 40 weeks of age. The mice experiment was carried out in a specific-pathogen-free (SPF) facility at Joslin Diabetes Center (JDC) under standard housing, feeding, and husbandry. The mice experiment protocol was reviewed and approved by JDC’s Institutional Animal Care and Use Committee (IACUC) (Protocol #2016-05). In addition, ARRIVE guidelines were strictly followed while conducting the mice experiment. Briefly, approximately 200 µ L of blood was collected through the lateral tail every two weeks. At the time of organ collection, mice were euthanized by isoflurane inhalation before cervical dislocation.
Tail blood was used to measure glucose concentration. Every two weeks, blood glucose was measured by Infinity blood glucose test strips (GTIN/DI#885502-002000) and an Infinity meter. Mice showing two consecutive glucose readings of ≥ 250 mg/dl were considered diabetic .
Peripheral blood was collected from mice every two weeks. Approximately 200 ul of blood from each mouse was collected from the tail using a heparin coated Microvette tube to prevent blood clots (Sarstedt catalog#16.443.100). PBMC’s were extracted using an equal volume of Histopaque 1083 (Sigma-Aldrich, Missouri, USA catalog#1083-1). Total cells count and viability were assessed on a Countess automated cell counter (Thermo Scientific,). After counting, the cell pellet was resuspended in 1 mL of Cryostor CS10 (Stemcell Technologies Catalog #07930) and frozen down for storage in liquid nitrogen.
Mice sacrificed at the onset of diabetes prior to detection of symptoms to prevent suffering and were perfused with collagenase (Vitacyte, CIzyme catalog #005-1030) and dissected to remove the pancreas for further processing. The pancreatic tissue was hand-shaken vigorously for 5-10 seconds and centrifuged at 500xg for 1.5 minutes then filtered to remove the remaining undigested tissue, fat, and lymph nodes. The filtrate was centrifuged at 500xg for 1.5 min at RT, and the tissue pellet was resuspended in 10 mL lymphocyte separation media (LSM) (Corning, cat #25072-CV) before purifying by density gradient centrifugation. Non-islet tissues were removed from the islet suspension by microscopic examination. These purified islets were subjected to non-enzymatic single-cell dissociation. The cell suspension was filtered through a 70μm cell strainer and counted before cryopreservation.
PBMC and islets single-cell suspension preserved in liquid nitrogen were revived before the magnetic separation of CD3 + T cells. CD3 + T cells were isolated using the Miltenyi MACS CD3 microbead kit (Miltenyi catalog#130-094-973) following the manufacturer and 10x genomics recommended guidelines (10x Protocol CG000123 Rev B). The cell viability and count were performed before and after CD3 + T cell enrichment. The magnetically separated CD3 + T cells were placed on ice and immediately used for single-cell cDNA library preparation.
Single-cell 5’ gene expression (GEX) and V(D)J sequencing libraries of T cells were generated using the Chromium Next GEM Single Cell 5’ Dual Index Reagent Kits v2 (10x Genomics, Pleasanton, CA, USA) following the manufacturer guidelines (10X protocol CG000331 Rev A).
Gene expression and V(D)J libraries were sequenced on Illumina NovaSeq6000 sequencer by GENEWIZ (GENEWIZ, LLC, NJ, USA). Each of the 5’GEX and V(D)J libraries were prepared with unique dual indexes, which allows multiplexing of all libraries for sequencing. Libraries were sequenced according to the 10X Genomics configuration. A minimum of 20,000 read pairs per cell were sequenced for 5’GEX libraries, and a minimum of 5,000 read pairs were sequenced for V(D)J libraries.
Raw sequencing reads were processed using Cell Ranger version 6.0.2 to produce gene expression count matrices and TCR clonotype summary. Using the raw sequencing FASTQ files, we first run the cellranger multi pipelines for simultaneous processing and analysis of V(D)J and gene expression data. This pipeline generates single cell feature counts, V(D)J sequences, and annotations for a single library. The mouse reference genome mm10-2020-A was used for aligning gene expression sequences, and the mouse reference GRCm38 v5.0.0 was used for aligning V(D)J sequences.
All gene expression analyses were performed using R version 4.1.0 and Seurat version 4.1.0 . Other R packages used during this analysis include dplyr version 1.0.8, patchwork version 1.1.1, data.table version 1.14.2, ggplot2 version 3.3.5, cowplot version 1.1.1, viridis version 0.6.2, gridExtra version 2.3, RColorBrewer version 1.1.2, and tibble version 3.1.6. Seurat object was created individually for all samples using the barcode.tsv, features.tsv and matrix.mtx files generated from the cellranger multi pipeline. The min.cells parameter was set to 3 and the min.features parameter was set to 200 during the creation of Seurat objects. Cells were filtered out from the Seurat object based on several quality control parameters. First, low-quality cells were removed based on CD3 gene expression (>0 counts), overall mitochondrial gene expression , and an aberrant high count of genes (200-5,000 features). Second, we filtered cells based on the expression of housekeeping genes (>0 counts). The quality-filtered gene expression data were normalized and scaled by Seurat function NormalizeData and ScaleData with default “LogNormalize” parameters, generating log-transformed gene expression measurement per 10,000 reads. Weighted PCAs were calculated with weights applied to relevant immune markers taken from multiple sources [ – ] and 400 most variable features. SCTransform integration was applied to normalize libraries and PC 1-10 were used for KNN based Umap and Louvain clustering generation. Clusters were selected from the 1.5 resolution and hand annotated for cell identity. Monocle3 was used to calculate pseudotime across clusters. Gating to identify CD4, CD8, and DN populations as well as Th subsets was done on normalized expression data using a linear gate between the median populations of cells (Supplementary Figure 4a) for each marker or a fixed value of 0.5.
Matching clones of T cells between blood and islets were determined based on the similarity of TCR sequences. We only included the T cells with at least one alpha and one beta chain in the clonal analysis. The cells were assigned to a particular clonotype if they shared the same amino acid sequence of TCR alpha and TCR beta sequence. This same definition was also applied to cells that carry multiple alpha and beta sequences. Cells having identical TCR alpha or beta sequences were defined as matching clones between blood and islets. For example, if a T cell in blood carries identical TCR alpha or TCR beta sequence to a T cell in islets, it is defined as islet-matching T cell in blood or blood-matching T cell in islets. The total count of cells in a particular clonotype was used to determine the clonal expansion. TCR clonotype data were added to the GEX Seurat object as metadata for integrated gene expression and TCR analysis. Predicted CDR3 sequences from T cells were aligned to the VDJdb database to query for exact matches.
Genes were clustered by similar patterns of expression across cells using the WGCNA package . Co-expression clusters were generated using a soft power of 10 and a minimum module size of 10, with a merge cut height of 0.15. Clusters were randomly assigned colors to label modules.
For GSEA, we downloaded the list of Hallmark pathways (v7.5.1) from the Molecular Signature Database (MSigDB, http://software.broadinstitute.org/gsea/msigdb/index.jsp ) and used the ranked list of differentially expressed genes between different cell types. R package fgsea version 1.20.0 was used for GSEA.
We first extracted the metadata, raw counts, and log normalized raw counts of gene expression from the integrated dataset to predict the matching status of cells based on gene expression profile. Then, we used logistic regression to make predictions with a liblinear solver and lasso penalty of 11. We used the Scikit-learn package in Python version 3.10.4 for the regression analysis and used matplotlib to generate the plots. Flow cytometry. For flow cytometry analysis, T-cell enriched samples from the blood of 5 diabetic and 5 non-diabetic NOD mice at 6,14,20 and 24 weeks of age were stained using labeled primary antibodies against CD4 (GK1.5, Thermofisher #12-0041-82), CD3e (17A2, Thermofisher #48-0032-82), CD8a (53-6.7, Thermofisher #56-0081-82), and CD8b (H35-17.2, Thermofisher #11-0083-82), NK1.1 (PK136, Thermofisher #407-5941-82), TCRβ (H57-597, BioLegend #109225), B220 (RA3-6B2, BioLegend #103212), FasL (MFL3, Thermofisher #63-5911-82), Fas (15A7, Thermofisher #46-0951-82), CD19 (1D3, Thermofisher #M004T02Y03-A), and TCRγδ (GL-3, Thermofisher #25-5711-82). Before staining, all samples were incubated with Live Dead Aqua Viability (Thermofisher #L34957) stain for 10 minutes in dark at 4°C. Samples were run on Attune NxT four color laser flow cytometer compensated with unstained control and single stained UltraComp eBead compensation beads (Invitrogen, #01-2222-41).
For flow cytometry analysis, T-cell enriched samples from the blood of 5 diabetic and 5 non-diabetic NOD mice at 6,14,20 and 24 weeks of age were stained using labeled primary antibodies against CD4 (GK1.5, Thermofisher #12-0041-82), CD3e (17A2, Thermofisher #48-0032-82), CD8a (53-6.7, Thermofisher #56-0081-82), and CD8b (H35-17.2, Thermofisher #11-0083-82), NK1.1 (PK136, Thermofisher #407-5941-82), TCRβ (H57-597, BioLegend #109225), B220 (RA3-6B2, BioLegend #103212), FasL (MFL3, Thermofisher #63-5911-82), Fas (15A7, Thermofisher #46-0951-82), CD19 (1D3, Thermofisher #M004T02Y03-A), and TCRγδ (GL-3, Thermofisher #25-5711-82). Before staining, all samples were incubated with Live Dead Aqua Viability (Thermofisher #L34957) stain for 10 minutes in dark at 4°C. Samples were run on Attune NxT four color laser flow cytometer compensated with unstained control and single stained UltraComp eBead compensation beads (Invitrogen, #01-2222-41).
All statistical analyses were performed with R or GraphPad Prism software, and P values < 0.05 were considered statistically significant. Figure legends show respective statistical analyses.
S1 Fig Clustering and marker discovery. (a) Clustering and UMAP of high quality T cells with principal components generated from top 2000 most variably expressed genes. (b) Expression of CD3, CD4, and CD8 across cells plotted on UMAP from variably expressed gene principal components. (c) Expression of CD4 and CD8a in each cell. Gating on normalized expression was performed at a positive value over 0.05 for CD4 or CD8a. (d) Gated cells represented as a proportion of CD3 + T cells in either clusters generated from variably expressed genes alone or variably expressed genes with T cell marker weighted principal components. (e) Dotmap expression of T cell markers where the size of each dot represents the number of positively expressing cells and color is a continuous variable of average normalized expression across all cells. (TIF) S2 Fig Pseudobulk expression of genes used to identify and annotate clusters. (TIF) S3 Fig DN T cell marker expression. (a) Expression of Klrb1c (NKT cell marker) and Trgv2 (gamma delta T cell marker) plotted against the UMAP coordinate of each cell. Correlation of CD3e expression to (a) CD4 and (b) CD8a with red line representing a linear model of values. (d) Heatmap and (e) violin plot showing expression of CD3e across clusters demonstrates equivalent expression of CD3e. Expression of CD3e, CD4, and CD8a plotted against (f) nCount and (g) nFeatures for each cell. Red line represents a linear model for x and y values. (h) Correlation matrix plotted as a heatmap for expression of each marker with the linear correlation of (i) Cd8a and Cd8b demonstrating the heavy linkage between the two gene’s expression. (TIF) S4 Fig Development of Effector T cell populations. (a) Gating strategy example for identification of CD4 + T cell like subsets. T cell subsets were predicted as a lack of CD8 expression and concurrent expression of the following markers: Th1(Infg and TnfA), Th2 (IL-4 without IL-10 expression), Th17 (IL-17), Th22 (IL-22 and TNFA), Tfh (IL-4 and IL-10), Th9 (IL-10 without Ifng or TNFA expression), and Treg (Foxp3 and IL-10). T cells without those combination of marker expression were labeled NA. (b) Proportion of T cell subsets across clusters (average of 20% not NA found in each cluster). (c) Proportion of CD4 + T cell across CD4 and DN clusters separated by tissue. (d) Monocle clustering and trajectory in Seurat UMAP space. (e) Localization of islet T cells in UMAP space. (f) Average cells per sample per cluster faceted by disease condition and tissue source. (TIF) S5 Fig Validation of DN populations were performed using flow cytometry. (a) Splenocytes were pooled from 24-week-old diabetic and non-diabetic NOD mice and then stained for T cell specific markers. (b) Boxplots show populations of CD3 + splenocytes of DN, CD4 + and CD8 + T cells. (c) Histogram of TCRB MFI shows lower protein abundance of TCR on the surface of DN T cells than CD4 + T cells. (d) Splenocytes largely are not expressers of gamma delta TCR. (e) Overall and (f) marker populations of CD3 + T cells in thymocytes. (TIF) S6 Fig DN T cells bioinformatic meta-analysis. (a) Misclassification of DN T cells gated as described within this studies dataset. (b) Expression of markers IL7R and GzmA plotted on Collier et al., UMAP coordinates. (c) Average number of cells per sample per pre-annotated cluster faceted by disease and infiltration status and colored by the correct T cell compartment. (TIF) S7 Fig Marker and differentially expressed genes. (a) Expression of commonly used markers in flow cytometric analysis of T cell populations. (b) Expression of exhaustion markers in T cells plotted in UMAP space. (c) Violin plot showing expression of Nr3c1 and Hif1a across clusters. (d) GSEA of grey module (MEGrey) using GO terms for biological process, cellular compartment, and molecular function. (e) Expression of Cel and PNLIPR across cells plotted in UMAP space. (TIF) S8 Fig Analysis of Infiltrating T cells in diabetic and non-diabetic NOD mice. (a) Histogram showing the number frequency of TCR sequences on the y axis with counted cells > 0 on the x axis. (b) Density plot showing the relative frequency of single TCR sequences found in greater than 2 cells. Heatmaps showing the relative number of T cells for each TCR by row over the (c) sample and (d) tissue in columns. (e) Heatmap showing the frequency of TCR clones that have more than 10 cells in the entire dataset. Color bars on the left represent the group identity of the TCR clone that is used for differential expression analysis in Supplementary File 14. (f) The number of cells per sample is plotted for all cells belonging to each group of TCR sequences. The color represents which sample type and tissue the number of cells are calculated from. (g) Heatmap showing the relative expression of exhaustion marker genes in group 3 CD4 + and DN T cells. The color bars labeled type show the category of each cell’s cluster identity. Exhausted cell clusters are more associated with a higher level of expression of exhaustion markers. (TIF) S9 Fig Lineage tracing and expression of infiltrating T cell clones. (a) Per TCR monocle trajectories plotted for each cell in the diabetic (left) and non-diabetic (right) mice. Each line represents the trajectory of one TCR clone across UMAP positions. (b) Volcano plot showing differential expression of genes in the infiltrating vs non-infiltrating T cells in diabetic mice with positive log fold change in the infiltrating T cells. Significant log fold change values > 1 and adjusted P-values < 0.01 are colored red and top 25 significant genes are labeled. (c) Violin and box plot showing average expression of Ifng between tissue matched (infiltrating) and non-matched T cells. (d) Likelihood estimation for bootstrapped pair transitions within clones. Each clone was randomly sampled for pairs of cells 100 times and the likelihood of each transition was calculated based on the proportion of two different cell types being chosen. The results were plotted as a graph with the nodes representing each cell cluster. The thickness of each bar represents the average percent likelihood of a transition occurring across clones with more than 3 cells. Red colored lines represent transitions occurring within the same compartment (i.e., CD4 to CD4) while blue colored lines represent transitions occurring between compartments (i.e., DN to CD4). The direction of transition cannot be inferred however the nodes are organized from bottom to top in order of ascending T cell maturation. (e) Epitope targets were predicted from the VDJdb database and cells with exact matches were plotted on the cell UMAP and colored by target epitope. (f) Normalized cell numbers were plotted based on targeted epitope by tissue. (TIF) S10 Fig Expression of TCR genes by cluster. Dot plot showing cell percentage and average expression in each cluster for tcr alpha (a), beta (b), and gamma/delta (c). (d-f) Psdobulk expression plots showing average expression of tcr gene expression by cluster. (TIF) S1 File S1 Table. Sample metadata for individual mice. S2 Table. CD3+ + T cell recovery and sequence yield from 10X single-cell library. S3 Table. Expression of marker genes across T cell clusters. S4 Table. Results of find all markers reported with average Log Fold Change, percent expression of cells, Gene name and both identified and simplified cluster name. S5 Table. Results of find markers show differential expression of genes in Naïve populations of DN T cells compared to CD4 T cells. S6 Table. Results of find markers show differential expression of genes in IL7r negative DN T cells compared to IL7r- CD4 T cells. S7 Table. Results of find markers show differential expression of genes in a mixed DN and CD4 effector T cell population compared to IL7 positive CD4 Teff and effector Tregs. S8 Table. Results of find markers show differential expression of genes in a mixed DN and CD4 effector T cell population compared to IL7 positive DN effector T cells. S9 Table. Results of find markers show differential expression of genes in IL7r negative DN T cells compared to Il7r+ + DN T cells. S10 Table. Results of find markers show differential expression of genes in Gzma positive CD8 T cells compared to Gzma negative CD8 T cells. S11 Table. Table showing GO enrichment of differentially expressed genes between populations of mixed DN and CD4 effector T cells and CD4 effector Tcells and Tregs. S12 Table. Table showing GO enrichment of differentially expressed genes between populations of Gzma positive CD8 T cells and Gzma negative CD8 T cells. S13 Table. Frequency of T cell clones separated by cluster appearance. Clusters refer to simple classification. Rows listed by specific TCRab clone CDR3 sequence. S14 Table. Differential expression of genes in clones specific to each clustered group. Fold change and Pct 1 refer to the group in the group column and pct 2 represents the proportion of expressing cells in all clones from other groups. S15 Table. Infiltrating (matching) TCR clones counted by sample type. Alpha beta TCR sequence for CDR3 region is reported. S16 Table. Differential expression of genes between islets-matching cells in blood and blood-matching cells in islets. The positive fold change shows enriched blood. S17 Table. Differential expression of genes between islets-matching and non-matching cells in blood. Positive Log fold-change shows enriched in matching. S18 Table. Differential expression of genes between blood-matching and non-matching cells in islets. Positive log fold-change. S19 Table. Exact matches between TCR-seq predicted CDR3 regions and experimentally validated TCR-pMHC interactions in the VDJdb. (XLSX)
|
NSUN2 methylates IRF4 to affect the capacity of macrophages attached to titanium implant on osteogenic differentiation of PDLSCs and angiogenesis of HUVECs in vitro | 4b064d68-725a-4fe0-988a-6221a89ceae4 | 11562097 | Dentistry[mh] | In recent years, implant restoration has become the first choice for patients with denture defects, and has been widely recognized and applied in clinical practice. Titanium (Ti) implant is an ideal material for dental implant because of its good histocompatibility with human body, not easy to cause rejection, and is conducive to the formation of bone union. Ti also has high strength and hardness and light weight, can withstand chewing pressure, maintain the stability of the implant, and reduce the patient’s foreign body sensation . Ti implants with micro-nano surface are still the most widely used materials in clinical practice because of their fantastic mechanical properties and biocompatibility . Implant surface morphology plays an important role in regulating protein adsorption and cell surface integrin, and changes intracellular signal transduction pathways . Moreover, Ti implants with sandblasted, large-grit and acid-etched treatment can promote tissue healing and increase bone union . Meanwhile, peri-implant tissues are susceptible to the same host-modulated plaque-induced factors that initiate and sustain periodontitis . Therefore, it is of significance to investigate the underlying mechanism of Ti implant with micro-nano surface on the dental implant therapy. After implantation, the Ti implant can induce innate immune response, and macrophages play an important role in this process as the defense line of the innate immune system . It has been concluded that macrophages play an important role in tissue repair and reconstruction by secreting chemokines and cytokines . Macrophages were activated after migration to the site of injury, showing two polarization states related to function: M1 type (promote inflammatory response) and M2 type (anti-inflammatory and promote tissue repair) . Relevant studies have shown that Ti implants can affect the differentiation and function of macrophages . Therefore, it is very important to explore the influence of Ti implants on the polarization of macrophages. After implantation, macrophages secrete various factors to attract mesenchymal stem cells and fibroblasts to migrate to the injured site and initiate the process of tissue healing and bone regeneration . It can be seen that in addition to inflammatory response, the participation of bone forming cells such as mesenchymal stem cells and osteoblasts in osteogenesis is also a necessary step in the treatment of peri-implant inflammation . 5-methylcytosine (m 5 C) modification of RNA is a common chemical modification of RNA in eukaryotes. It is found in a variety of RNA types, including mRNA, tRNA, and rRNA. m 5 C modification is related to mRNA stability, splicing, cytoplasmic shuttling, and DNA damage repair . In addition, the abnormality of m 5 C modification has been associated with the occurrence and development of various diseases . Two m 5 C immune subtypes were reported in macrophages in prostate cancer . m 5 C methylation can inhibit the transcription of macrophage-related chemokines, thereby inhibiting the recruitment of M2 macrophages in bladder cancer . Whether m 5 C methylation regulates macrophages during Ti implantation is unknown. Accelerating the biological process of osseointegration at the implant-bone interface, as well as enhancing the quality of osseointegration, are crucial objectives within the realm of dental implant materials and surface modification . Therefore, we aimed to evaluate the effects of Ti on regulating the immune microenvironment and the polarization of macrophages. Furthermore, the osteogenesis of periodontal ligament stem cells (PDLSCs) and angiogenesis process of human umbilical vein endothelial cells (HUVECs) regulated by the conditioned medium of macrophages attached with Ti implant were also studied.
Cell collection and culture Human monocytic cell line (THP-1) obtained from the Cell Bank of the Chinese Academy of Sciences were cultured in RPMI 1640 medium (HyClone) containing 10% fetal bovine serum (FBS; Bovogen) in a 37℃ constant temperature incubator with 5% CO 2 concentration. The THP-1 monocytes were differentiated to macrophages with 10 ng/ml phorbol-12-myristate-13-acetate (PMA) for 24 h . Macrophages were divided into three groups: control group macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto tissue culture polystyrene (TCPS) in 24-well plates; LPS group macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto TCPS in 24-well plates, and treated with 100 ng/mL Porphyromonas gingivalis lipopolysaccharide (LPS; InvivoGen) for 24 h ; LPS + implant group macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto rough-hydro grade 2 unalloyed Ti (2010.601-STM; 15 mm-diameter; Ti + polyoxymethylene, Institut Straumann AG) surfaces in 24-well plates, and treated with 100 ng/mL LPS for 24 h. PDLSCs were purchased from Bohu BiologicalTechnology Co.,Ltd. (Shanghai, China), and were incubated at 37℃ in DMEM containing 10% FBS. The PDLSCs in the 3–5 passage were inoculated in a 12-well plate at 5 × 10 5 cells/well and cultured in a medium containing β-glycerophosphate sodium (5 mmol/L), vitamin C (50 µg/mL), dexamethasone (100 mmol/L), and FBS (10%) (all purchased from Sigma-Aldrich) . HUVECs were purchased from the ATCC. HUVEC cells in the 3–5 passage were selected and achieved a fusion rate of 80% within 24 h after inoculation. Then the cells were starved, the complete culture medium was replaced with DMEM containing 0.2% FBS, and cultured for 24 h. Matrigel was diluted to 8–12 mg/mL and 50 µl-80 µL was added to each well in the 96-well plate. The plate is then placed in a cell incubator and incubated at 37℃ for 30 min to allow Matrigel to form a gel base. HUVECs were suspended in a medium containing 10% FBS at a concentration of 3 × 10 5 cells/ml, and cell suspension was added to each well. 6 h later, the expression of genes associated with angiogenesis were detected by quantitative reverse transcription PCR (RT-qPCR). Cell transfection The coding sequences of NSUN2 or IRF4 were inserted into the lentiviral vector pHBLV-CMV-MCS-EF1-ZsGreen-T2A-puro (HanHeng Biotechnology) to over-express NSUN2 or IRF4 in macrophages. The empty lentiviral vector pHBLV-CMV-MCS-EF1-ZsGreen-T2A-puro vector was used as the negative control. The packaging plasmids (4:3:1) were co-transfected into HEK293T cells for lentivirus packaging. After 48 h, lentivirus particles from the medium were collected and filtered. Subsequently, macrophages were infected with lentivirus according to the instructions using Lipofectamine 3000 (Thermo Fisher Scientific). 48 h later, infected cells were selected with puromycin to establish stable knockdown or overexpression cells. The mRNA levels of NSUN2 or IRF4 were detected by RT-qPCR to confirm the over-expression effect. Transfected macrophages were treated with LPS and/or attached to Ti implants. Immunofluorescence staining Firstly, macrophages were cultured in 24-well plates. macrophages were fixed with 4% paraformaldehyde to block the non-specific binding site when grew to about 80%. The cells were treated with 0.1% Triton X-100 to increase membrane permeability. The non-specific binding site was blocked with 5% bovine serum albumin (BSA) and incubated at room temperature for 30 min. Subsequently, primary antibodies including anti-iNOS (ab178945, 1:500, Abcam) and anti-CD206 (ab64693, 1 µg/mL, Abcam) were added to the sample and incubated at 4 °C overnight. After washed with PBS for three times with 5 min each time, the fluorescence labeled secondary antibody (1:1,000) was added to the sample and incubated at room temperature for 1–2 h. Cells were washed with PBS for three times with 5 min each time, and DAPI was added for 5 min staining. Finally, the cover glass was sealed with a sealing medium and sealed with nail polish to prevent the sample from drying out and moving under the microscope. The sample is placed in darkness and cells were observed with a fluorescence microscope. Conditioned medium collection and treatment Macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto Tiimplants in 24-well plates, and treated with 100 ng/mL LPS for 24 h. Subsequently, the supernatant was collected, and was mixed with low-glucose DMEM (containing 10% FBS) in a ratio of 1:1 to form the conditioned medium for PDLSCs and HUVECs. PDLSCs are grouped as follows: PDLSCs cultured in supernatant of macrophages, PDLSCs cultured in supernatant of macrophages treated with LPS, and PDLSCs cultured in supernatant of LPS-treated macrophages attached to Ti. Mineralization by differentiated PDLSCs was detected after 21 d . HUVECs are grouped as follows: HUVECs cultured in supernatant of macrophages, HUVECs cultured in supernatant of macrophages treated with LPS, and HUVECs cultured in supernatant of LPS-treated macrophages attached to Ti. RT-qPCR The mRNA levels of M1 polarization markers (IL6, TNFA, and IL1B) and M2 polarization markers (TGFB1, IL10 and Arg1) from macrophages were detected by RT-qPCR. At the same time, levels of angiogenesis-related markers (VEGF, ACTA2, and COL1A1), osteogenesis-related markers (BGLAP, SPP1, and BMP2), and m 5 C related genes (NOP2, NSUN2, NSUN3, NSUN4, NSUN5, NSUN6, NSUN7, and TRDMT1) were also measured. Total RNA was extracted by Trizon method, and the absorbance of RNA at 260 nm and 280 nm was determined by ultraviolet spectrophotometry, and the RNA content was calculated. RNA was obtained for reverse transcription reaction by Promega MMLV reverse transcriptase (9PIM170). RT-qPCR was performed on 0.5 µL RT products by SYBR Premix Ex TaqTM (TaKaRa). The reaction conditions were as follows: PCR program was set at 94℃ 4 min, 94℃ 30 s, 56℃ 60 s, 72℃ 40 s for 40 cycle. GAPDH acted as the endogenous control, and the 2 −ΔΔCT method was for transcript expression level measurement. Primer sequences of this study are listed in Table . Osteogenic differentiation and alizarin Red S staining Alizarin Red S (ARS) staining is usually used to detect and observe calcium salt deposits to assess the degree of osteogenic differentiation of cells after 21 days . The staining steps are as follows: osteogenic differentiated PDLSCs were fixed with 4% paraformaldehyde (pre-cooled at 4℃) at room temperature for 30 min. After washing twice with PBS or double steaming water to remove the remaining fixative, the configured ARS dye solution (1%, pH 4.2; Sigma-Aldrich, A5533) was added to the sample cells and incubated at room temperature for 1 h. Then the ARS dye was discarded and cells were gently washed with double steaming water several times to remove the background stain. The stained sample can be viewed and photographed under an optical microscope (Olympus, IX73, Japan), and the calcium deposit area appears to be bright red. The ARS was dissolved with 10% acetopyridine (w/v) and the staining intensity was quantified by measuring optical density values at 562 nm to assess the degree of calcification. RNA m5C dot blotting Total RNA was extracted by TRIZOL reagent (Invitrogen). Then a Bio-Dot apparatus (Bio‐Rad) was used to transfer the mRNA which has been denatured in advance onto a nitrocellulose membrane (Amersham). Then the separated RNA were transferred onto a positively charged nylon membrane followed by cross-linking with Ultraviolet light. Then the membranes were blocked, incubated m 5 C antibody (ab214727, 1.0 µg/mL, Abcam), followed by incubation with horseradish peroxidase-conjugated anti-rabbit IgG secondary antibody. Finally, we used the enhanced chemiluminescence (Bio‐Rad) to visualize the membrane, and the m 5 C level of each group was quantitatively determined by Image J software. Bioinformatic analysis The GSE173078 dataset was downloaded from the GEO database. This expression profiling is related to mRNA profiles from 12 periodontitis, 12 gingivitis, and 12 healthy patients. Differentially expressed mRNA was analyzed using a paired t-test using the limma R package, and mRNA with adjusted p -value < 0.05 and |logFC > 1| were considered significant mRNA. To investigate the potential biological functions, KEGG pathway enrichment analysis was performed by the R package clusterProfiler. The top 10 pathways were selected based on the adjusted p -value ranking. RNAm 5 Cfinder ( http://www.rnanut.net/rnam5cfinder/ ) was used to predict the potential m 5 C modification site of IRF4. RNA immunoprecipitation (RIP) The RIP assay was performed using a Magna RIP ® RIP Kit (17–700, Millipore) according to the manufacturer’s instructions. Briefly, macrophages were harvested and washed twice with cold PBS, and the cell pellet was incubated with RIP lysis buffer (150 mM KCl, 10 mM HEPES pH 7.6, 2 mM EDTA, 0.5% NP-40, 0.5 mM DTT, Protease Inhibitor, RNase Inhibitor) on ice for 30 min. One tenth portion of the cell lysate was used as input. The rest of the cell lysate was incubated with either Rabbit IgG (10285-1-AP, Proteintech)-coated beads or anti-NSUN2 (ab259941, 1/30; Abcam)-coated beads for 4 h at room temperature. Afterward, the beads-antibody-protein-RNA complex was washed five times with ice-cold washing buffer (200 mM NaCl, 50 mM HEPES pH 7.6, 2 mM EDTA, 0.05% NP-40, 0.5 mM DTT, RNase inhibitor). Then, immunoprecipitated sample was digested with proteinase K and the RNA was precipitated with glycogen (Thermo Scientific, AM9516). Total RNA was extracted by TRIzol reagent followed by RT-qPCR. m5C-RIP Purified mRNA was fragmented RNA Fragmentation Reagents (Invitrogen, AM8740). Specific 2.5 µg m 5 C antibodies (ab214727, Abcam) were incubated with 400 ng RNA fragments (100-nucleotide-long) in immunoprecipitation buffer and incubated by rotating at 25℃ for 1 h to bind the antibody to the m 5 C modification site. The mixture was then immunoprecipitated by incubation with protein A/G magnetic beads at 4℃ for 5 h. The magnetic bead (Thermo Fisher Scientific, 10002D) was washed several times with IP buffer, and the bound RNA fragments were eluted from the beads by proteinase K digestion at 554℃ for 60 min. Finally, the RNA containing m 5 C modification was isolated from the eluate by phenol-chloroform extraction and ethanol plus glycogen. Finally, qRT-PCR analysis was performed to evaluate the m 5 C modification level of IRF4. Double luciferase reporter gene experiment The wild-type (WT) and mutant (Mut) fragments of IRF4 were constructed and inserted into the the pGM-CMV-Luc vector (Yeasen Biotech, Shanghai, China). Cells were planted in a 24-well plate with 100,000 cells per well and attached to the wall for 36 h. 900 ng pGM-CMV-Luc and 40 ng Renilla luciferase plasmid were transfected per well (20:1) 12 h later. After 24 h, the fluorescence values of the lysed cells were determined according to the kit procedure. mRNA stability assay Macrophages were inoculated in 12-well plates overnight and then treated with 5 µg/mL actinomycin D (MedChemExpress) at 0, 1, 4, 8 and 12 h. Total RNA was then isolated and the results were analyzed by RT-qPCR. Statistical analysis The experimental data were analyzed by GraphPad Prism software version 8.3, and the data operation between the two groups was represented by mean ± SD and compared with t test. One-way ANOVA was used to compare the mean of multiple groups. Tukey′s post hoc test was used to compare pairwise comparisons between groups. There were three iterations of each experiment in this study. The p value less than 0.05 means the difference is statistically significant.
Human monocytic cell line (THP-1) obtained from the Cell Bank of the Chinese Academy of Sciences were cultured in RPMI 1640 medium (HyClone) containing 10% fetal bovine serum (FBS; Bovogen) in a 37℃ constant temperature incubator with 5% CO 2 concentration. The THP-1 monocytes were differentiated to macrophages with 10 ng/ml phorbol-12-myristate-13-acetate (PMA) for 24 h . Macrophages were divided into three groups: control group macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto tissue culture polystyrene (TCPS) in 24-well plates; LPS group macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto TCPS in 24-well plates, and treated with 100 ng/mL Porphyromonas gingivalis lipopolysaccharide (LPS; InvivoGen) for 24 h ; LPS + implant group macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto rough-hydro grade 2 unalloyed Ti (2010.601-STM; 15 mm-diameter; Ti + polyoxymethylene, Institut Straumann AG) surfaces in 24-well plates, and treated with 100 ng/mL LPS for 24 h. PDLSCs were purchased from Bohu BiologicalTechnology Co.,Ltd. (Shanghai, China), and were incubated at 37℃ in DMEM containing 10% FBS. The PDLSCs in the 3–5 passage were inoculated in a 12-well plate at 5 × 10 5 cells/well and cultured in a medium containing β-glycerophosphate sodium (5 mmol/L), vitamin C (50 µg/mL), dexamethasone (100 mmol/L), and FBS (10%) (all purchased from Sigma-Aldrich) . HUVECs were purchased from the ATCC. HUVEC cells in the 3–5 passage were selected and achieved a fusion rate of 80% within 24 h after inoculation. Then the cells were starved, the complete culture medium was replaced with DMEM containing 0.2% FBS, and cultured for 24 h. Matrigel was diluted to 8–12 mg/mL and 50 µl-80 µL was added to each well in the 96-well plate. The plate is then placed in a cell incubator and incubated at 37℃ for 30 min to allow Matrigel to form a gel base. HUVECs were suspended in a medium containing 10% FBS at a concentration of 3 × 10 5 cells/ml, and cell suspension was added to each well. 6 h later, the expression of genes associated with angiogenesis were detected by quantitative reverse transcription PCR (RT-qPCR).
macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto tissue culture polystyrene (TCPS) in 24-well plates;
macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto TCPS in 24-well plates, and treated with 100 ng/mL Porphyromonas gingivalis lipopolysaccharide (LPS; InvivoGen) for 24 h ;
macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto rough-hydro grade 2 unalloyed Ti (2010.601-STM; 15 mm-diameter; Ti + polyoxymethylene, Institut Straumann AG) surfaces in 24-well plates, and treated with 100 ng/mL LPS for 24 h. PDLSCs were purchased from Bohu BiologicalTechnology Co.,Ltd. (Shanghai, China), and were incubated at 37℃ in DMEM containing 10% FBS. The PDLSCs in the 3–5 passage were inoculated in a 12-well plate at 5 × 10 5 cells/well and cultured in a medium containing β-glycerophosphate sodium (5 mmol/L), vitamin C (50 µg/mL), dexamethasone (100 mmol/L), and FBS (10%) (all purchased from Sigma-Aldrich) . HUVECs were purchased from the ATCC. HUVEC cells in the 3–5 passage were selected and achieved a fusion rate of 80% within 24 h after inoculation. Then the cells were starved, the complete culture medium was replaced with DMEM containing 0.2% FBS, and cultured for 24 h. Matrigel was diluted to 8–12 mg/mL and 50 µl-80 µL was added to each well in the 96-well plate. The plate is then placed in a cell incubator and incubated at 37℃ for 30 min to allow Matrigel to form a gel base. HUVECs were suspended in a medium containing 10% FBS at a concentration of 3 × 10 5 cells/ml, and cell suspension was added to each well. 6 h later, the expression of genes associated with angiogenesis were detected by quantitative reverse transcription PCR (RT-qPCR).
The coding sequences of NSUN2 or IRF4 were inserted into the lentiviral vector pHBLV-CMV-MCS-EF1-ZsGreen-T2A-puro (HanHeng Biotechnology) to over-express NSUN2 or IRF4 in macrophages. The empty lentiviral vector pHBLV-CMV-MCS-EF1-ZsGreen-T2A-puro vector was used as the negative control. The packaging plasmids (4:3:1) were co-transfected into HEK293T cells for lentivirus packaging. After 48 h, lentivirus particles from the medium were collected and filtered. Subsequently, macrophages were infected with lentivirus according to the instructions using Lipofectamine 3000 (Thermo Fisher Scientific). 48 h later, infected cells were selected with puromycin to establish stable knockdown or overexpression cells. The mRNA levels of NSUN2 or IRF4 were detected by RT-qPCR to confirm the over-expression effect. Transfected macrophages were treated with LPS and/or attached to Ti implants.
Firstly, macrophages were cultured in 24-well plates. macrophages were fixed with 4% paraformaldehyde to block the non-specific binding site when grew to about 80%. The cells were treated with 0.1% Triton X-100 to increase membrane permeability. The non-specific binding site was blocked with 5% bovine serum albumin (BSA) and incubated at room temperature for 30 min. Subsequently, primary antibodies including anti-iNOS (ab178945, 1:500, Abcam) and anti-CD206 (ab64693, 1 µg/mL, Abcam) were added to the sample and incubated at 4 °C overnight. After washed with PBS for three times with 5 min each time, the fluorescence labeled secondary antibody (1:1,000) was added to the sample and incubated at room temperature for 1–2 h. Cells were washed with PBS for three times with 5 min each time, and DAPI was added for 5 min staining. Finally, the cover glass was sealed with a sealing medium and sealed with nail polish to prevent the sample from drying out and moving under the microscope. The sample is placed in darkness and cells were observed with a fluorescence microscope.
Macrophages at a density of 5 × 10 4 cells/cm 2 were plated onto Tiimplants in 24-well plates, and treated with 100 ng/mL LPS for 24 h. Subsequently, the supernatant was collected, and was mixed with low-glucose DMEM (containing 10% FBS) in a ratio of 1:1 to form the conditioned medium for PDLSCs and HUVECs. PDLSCs are grouped as follows: PDLSCs cultured in supernatant of macrophages, PDLSCs cultured in supernatant of macrophages treated with LPS, and PDLSCs cultured in supernatant of LPS-treated macrophages attached to Ti. Mineralization by differentiated PDLSCs was detected after 21 d . HUVECs are grouped as follows: HUVECs cultured in supernatant of macrophages, HUVECs cultured in supernatant of macrophages treated with LPS, and HUVECs cultured in supernatant of LPS-treated macrophages attached to Ti.
The mRNA levels of M1 polarization markers (IL6, TNFA, and IL1B) and M2 polarization markers (TGFB1, IL10 and Arg1) from macrophages were detected by RT-qPCR. At the same time, levels of angiogenesis-related markers (VEGF, ACTA2, and COL1A1), osteogenesis-related markers (BGLAP, SPP1, and BMP2), and m 5 C related genes (NOP2, NSUN2, NSUN3, NSUN4, NSUN5, NSUN6, NSUN7, and TRDMT1) were also measured. Total RNA was extracted by Trizon method, and the absorbance of RNA at 260 nm and 280 nm was determined by ultraviolet spectrophotometry, and the RNA content was calculated. RNA was obtained for reverse transcription reaction by Promega MMLV reverse transcriptase (9PIM170). RT-qPCR was performed on 0.5 µL RT products by SYBR Premix Ex TaqTM (TaKaRa). The reaction conditions were as follows: PCR program was set at 94℃ 4 min, 94℃ 30 s, 56℃ 60 s, 72℃ 40 s for 40 cycle. GAPDH acted as the endogenous control, and the 2 −ΔΔCT method was for transcript expression level measurement. Primer sequences of this study are listed in Table .
Alizarin Red S (ARS) staining is usually used to detect and observe calcium salt deposits to assess the degree of osteogenic differentiation of cells after 21 days . The staining steps are as follows: osteogenic differentiated PDLSCs were fixed with 4% paraformaldehyde (pre-cooled at 4℃) at room temperature for 30 min. After washing twice with PBS or double steaming water to remove the remaining fixative, the configured ARS dye solution (1%, pH 4.2; Sigma-Aldrich, A5533) was added to the sample cells and incubated at room temperature for 1 h. Then the ARS dye was discarded and cells were gently washed with double steaming water several times to remove the background stain. The stained sample can be viewed and photographed under an optical microscope (Olympus, IX73, Japan), and the calcium deposit area appears to be bright red. The ARS was dissolved with 10% acetopyridine (w/v) and the staining intensity was quantified by measuring optical density values at 562 nm to assess the degree of calcification.
Total RNA was extracted by TRIZOL reagent (Invitrogen). Then a Bio-Dot apparatus (Bio‐Rad) was used to transfer the mRNA which has been denatured in advance onto a nitrocellulose membrane (Amersham). Then the separated RNA were transferred onto a positively charged nylon membrane followed by cross-linking with Ultraviolet light. Then the membranes were blocked, incubated m 5 C antibody (ab214727, 1.0 µg/mL, Abcam), followed by incubation with horseradish peroxidase-conjugated anti-rabbit IgG secondary antibody. Finally, we used the enhanced chemiluminescence (Bio‐Rad) to visualize the membrane, and the m 5 C level of each group was quantitatively determined by Image J software.
The GSE173078 dataset was downloaded from the GEO database. This expression profiling is related to mRNA profiles from 12 periodontitis, 12 gingivitis, and 12 healthy patients. Differentially expressed mRNA was analyzed using a paired t-test using the limma R package, and mRNA with adjusted p -value < 0.05 and |logFC > 1| were considered significant mRNA. To investigate the potential biological functions, KEGG pathway enrichment analysis was performed by the R package clusterProfiler. The top 10 pathways were selected based on the adjusted p -value ranking. RNAm 5 Cfinder ( http://www.rnanut.net/rnam5cfinder/ ) was used to predict the potential m 5 C modification site of IRF4.
The RIP assay was performed using a Magna RIP ® RIP Kit (17–700, Millipore) according to the manufacturer’s instructions. Briefly, macrophages were harvested and washed twice with cold PBS, and the cell pellet was incubated with RIP lysis buffer (150 mM KCl, 10 mM HEPES pH 7.6, 2 mM EDTA, 0.5% NP-40, 0.5 mM DTT, Protease Inhibitor, RNase Inhibitor) on ice for 30 min. One tenth portion of the cell lysate was used as input. The rest of the cell lysate was incubated with either Rabbit IgG (10285-1-AP, Proteintech)-coated beads or anti-NSUN2 (ab259941, 1/30; Abcam)-coated beads for 4 h at room temperature. Afterward, the beads-antibody-protein-RNA complex was washed five times with ice-cold washing buffer (200 mM NaCl, 50 mM HEPES pH 7.6, 2 mM EDTA, 0.05% NP-40, 0.5 mM DTT, RNase inhibitor). Then, immunoprecipitated sample was digested with proteinase K and the RNA was precipitated with glycogen (Thermo Scientific, AM9516). Total RNA was extracted by TRIzol reagent followed by RT-qPCR.
Purified mRNA was fragmented RNA Fragmentation Reagents (Invitrogen, AM8740). Specific 2.5 µg m 5 C antibodies (ab214727, Abcam) were incubated with 400 ng RNA fragments (100-nucleotide-long) in immunoprecipitation buffer and incubated by rotating at 25℃ for 1 h to bind the antibody to the m 5 C modification site. The mixture was then immunoprecipitated by incubation with protein A/G magnetic beads at 4℃ for 5 h. The magnetic bead (Thermo Fisher Scientific, 10002D) was washed several times with IP buffer, and the bound RNA fragments were eluted from the beads by proteinase K digestion at 554℃ for 60 min. Finally, the RNA containing m 5 C modification was isolated from the eluate by phenol-chloroform extraction and ethanol plus glycogen. Finally, qRT-PCR analysis was performed to evaluate the m 5 C modification level of IRF4.
The wild-type (WT) and mutant (Mut) fragments of IRF4 were constructed and inserted into the the pGM-CMV-Luc vector (Yeasen Biotech, Shanghai, China). Cells were planted in a 24-well plate with 100,000 cells per well and attached to the wall for 36 h. 900 ng pGM-CMV-Luc and 40 ng Renilla luciferase plasmid were transfected per well (20:1) 12 h later. After 24 h, the fluorescence values of the lysed cells were determined according to the kit procedure.
Macrophages were inoculated in 12-well plates overnight and then treated with 5 µg/mL actinomycin D (MedChemExpress) at 0, 1, 4, 8 and 12 h. Total RNA was then isolated and the results were analyzed by RT-qPCR.
The experimental data were analyzed by GraphPad Prism software version 8.3, and the data operation between the two groups was represented by mean ± SD and compared with t test. One-way ANOVA was used to compare the mean of multiple groups. Tukey′s post hoc test was used to compare pairwise comparisons between groups. There were three iterations of each experiment in this study. The p value less than 0.05 means the difference is statistically significant.
Macrophages activated by Ti implant enhance osteogenic differentiation of PDLSCs and angiogenesis of HUVECs Firstly, the regulatory effect of Ti implant on macrophage polarization was investigated. The levels of M1 polarization markers and M2 polarization markers in macrophgages of different groups were detected. In LPS-stimulated macrophages, immunofluorescence staining with specific iNOS antibody showed that iNOS protein was mainly distributed in the cytoplasm, showing a strong fluorescence signal. This result indicated that LPS stimulation can effectively induce the expression of iNOS in macrophages, and the expression of iNOS is related to the activation of M1 polarization in macrophages. In contrast to that, LPS stimulation decreased the fluorescence signal of CD206 protein distributed in the cytoplasm of macrophages, indicating that activation of M1 polarization of macrophages was suppressed by LPS treatment. However, the fluorescence intensity of iNOS in LPS-treated macrophages attached to Ti implant was lower than that in LPS-treated macrophages while CD206 fluorescence intensity was higher (Fig. A and B). qPCR analysis demonstrated that levels of M1 polarization related genes (IL6, TNFA, and IL1B) were increased while levels of M2 polarization related genes (TGFB1, IL10, and ARG1) were decreased in LPS-treated macrophages. Ti implant attachement significantly weakened the effects of LPS treatment on regulating these genes levels (Fig. C and D, p < 0.01). Taken together, macrophages were activated into M1 macrophages under the interference of LPS, and M2 macrophages were more easily activated when the macrophages attached to the Ti implant were treated with LPS. Subsequently, the effects of Ti implant on macrophage regulation of osteogenic differentiation and angiogenesis was studied. As shown in Fig. E, angiogenesis related genes including VEGF, ACTA2, and COL1A1 were evaluated in three groups of HUVECs cultured in different conditioned medium for 6 h. qPCR analysis suggested that supernatant of macrophages treated with LPS suppressed the angiogenic capacity of HUVECs while HUVECs cultured in supernatant of LPS-treated macrophages attached to Ti implant showed higher angiogenic potential by enhancing the mRNA levels of VEGF, ACTA2, and COL1A1 (Fig. E, p < 0.01). At the same time, LPS-treated macrophage supernatants reduced calcium accumulation of PDLSCs after osteogenic differentiation induction for 21 days according to the ARS staining, and Ti implant attachment increased the mineralization of PDLSCs after osteogenic differentiation induction (Fig. F, p < 0.01). Osteogenesis-related genes including BGLAP, SPP1, and BMP2 in PDLSCs were decreased under the LPS interference, but increased by the Ti implant (Fig. G, p < 0.01). Over-expression of m5C RNA methyltransferase NSUN2 attenuates the effects of Ti implant on macrophages To explore whether RNA m 5 C modification can modulate the effects of Ti implant on macrophage regulation of osteogenic differentiation and angiogenesis, we analyzed the main genes related to m 5 C in macrophages of each groups. The RNA m 5 C dot blotting data suggested that m 5 C levels in macrophages of LPS group were higher than that in control group while Ti implant attachement down-regulated the total m 5 C levels in macrophages (Fig. A, p < 0.01). Furthermore, m 5 C related genes including NOP2, NSUN2, NSUN3, NSUN4, NSUN5, NSUN6, NSUN7, and TRDMT1 were detected in three groups of macrophages, and the qPCR analysis then demonstrated that NOP2, NSUN2, NSUN3, NSUN5, NSUN6, and NSUN7 were significantly highly expressed in macrophages of LPS group compared with the control group. After attached to Ti implant, NSUN2 level was further significantly down-regulated in macrophages while levels of NOP2, NSUN3, NSUN5, NSUN6, and NSUN7 were not significantly changed (Fig. B, p < 0.05). Subsequently, the regulatory role of NSUN2 was investigated. After transfection of macrophages, the expression of NUSN2 increased by more than 10 times compared with the control vector group according to the qPCR analysis (Fig. A, p < 0.01). Overexpression of NSUN2 weakened the promoting effect of Ti implant on M1 polarization of macrophages and enhanced the promoting effect of M2 polarization of macrophages, which showed an increase in iNOS protein level and a decrease in CD206 protein level, as well as an increase in the gene levels of IL6, TNFA and IL1B, and a decrease in the gene levels of TGFB1, IL10 and ARG1 (Fig. B - E, p < 0.01). Up-regulation of NSUN2 also weakened the effects of Ti implant on macrophages promoting the osteogenic differentiation of PDLSCs, which was manifested by decreased levels of VEGF, ACTA2 and COL1A1 genes and decreased levels of mineralization (Fig. F and G, p < 0.01). At the same time, the down-regulation of BGLAP, SPP1 and BMP2 genes also suggested that the up-regulation of NSUN2 impaired the ability of Ti implants to promote macrophage induction of HUVECs angiogenesis. (Fig. H, p < 0.01). NSUN2 methylates IRF4 to affect the capacity of macrophages on osteogenic differentiation of PDLSCs and angiogenesis of HUVECs mRNAs in GSE173078 dataset are related to periodontitis and gingivitis, and inflammatory factors associated with periodontal disease were screened. The KEGG analysis of differentially expressed genes from GSE173078 dataset was performed, and the biological pathway suggested that IRF4 was enriched in several inflammatory signaling pathways (Fig. A). IRF4 has been reported to exert regulatory role in immune infiltration , macrophage polarization , and cell cycle . The suppression effect of IRF4 on osteogenic differentiation was also been empathized according to various studies . Therefore, we speculated that IRF4 may be m 5 C modified mediated by NSUN2 in macrophages. The m 5 C-RIP assay was implemented to assess the m 5 C modification status of IRF4 mRNA. The results exhibited that the level of IRF4 can be enriched by m 5 C antibody (Fig. B, p < 0.01). RIP followed by qPCR found that compared with IgG antibody, NSUN2 antibody significantly enriched IRF4 (Fig. C, p < 0.01). Subsequently, the potential m 5 C modification sites of IRF4 predicted by RNAm 5 Cfinder suggested that IRF4 may be m 5 C modified at three sites (Fig. D). After mutation at each of the three sites, the results of the double-luciferase gene report experiment suggested that there was no significant change in relative luciferase activity before and after NSUN2 over-expression at mutation sites 1 and 3, but after mutation at site 2, up-regulation of NSUN2 could significantly reduce relative luciferase activity (Fig. E, p < 0.01). Moreover, over-expression of NSUN2 promoted the degradation of IRF4 mRNA (Fig. F, p < 0.01). After transfection of macrophages, the expression of IRF4 increased by more than 10 times compared with the control vector group according to the qPCR analysis (Fig. A, p < 0.01), the macrophage polarization related proteins and genes were evaluated. Both of the immunofluorescence staining and qPCR analysis suggested that NSUN2 promoted the M1 macrophage polarization while IRF4 weakened the effects of NSUN2 by promoting M2 macrophage polarization (Fig. B - E, p < 0.01). Meanwhile, IRF4 also reversed the effects of NSUN2 on the regulatory function of macrophages on inhibition of osteogenic differentiation of PDLSCs and angiogenesis of HUVECs (Fig. F - H, p < 0.01).
Firstly, the regulatory effect of Ti implant on macrophage polarization was investigated. The levels of M1 polarization markers and M2 polarization markers in macrophgages of different groups were detected. In LPS-stimulated macrophages, immunofluorescence staining with specific iNOS antibody showed that iNOS protein was mainly distributed in the cytoplasm, showing a strong fluorescence signal. This result indicated that LPS stimulation can effectively induce the expression of iNOS in macrophages, and the expression of iNOS is related to the activation of M1 polarization in macrophages. In contrast to that, LPS stimulation decreased the fluorescence signal of CD206 protein distributed in the cytoplasm of macrophages, indicating that activation of M1 polarization of macrophages was suppressed by LPS treatment. However, the fluorescence intensity of iNOS in LPS-treated macrophages attached to Ti implant was lower than that in LPS-treated macrophages while CD206 fluorescence intensity was higher (Fig. A and B). qPCR analysis demonstrated that levels of M1 polarization related genes (IL6, TNFA, and IL1B) were increased while levels of M2 polarization related genes (TGFB1, IL10, and ARG1) were decreased in LPS-treated macrophages. Ti implant attachement significantly weakened the effects of LPS treatment on regulating these genes levels (Fig. C and D, p < 0.01). Taken together, macrophages were activated into M1 macrophages under the interference of LPS, and M2 macrophages were more easily activated when the macrophages attached to the Ti implant were treated with LPS. Subsequently, the effects of Ti implant on macrophage regulation of osteogenic differentiation and angiogenesis was studied. As shown in Fig. E, angiogenesis related genes including VEGF, ACTA2, and COL1A1 were evaluated in three groups of HUVECs cultured in different conditioned medium for 6 h. qPCR analysis suggested that supernatant of macrophages treated with LPS suppressed the angiogenic capacity of HUVECs while HUVECs cultured in supernatant of LPS-treated macrophages attached to Ti implant showed higher angiogenic potential by enhancing the mRNA levels of VEGF, ACTA2, and COL1A1 (Fig. E, p < 0.01). At the same time, LPS-treated macrophage supernatants reduced calcium accumulation of PDLSCs after osteogenic differentiation induction for 21 days according to the ARS staining, and Ti implant attachment increased the mineralization of PDLSCs after osteogenic differentiation induction (Fig. F, p < 0.01). Osteogenesis-related genes including BGLAP, SPP1, and BMP2 in PDLSCs were decreased under the LPS interference, but increased by the Ti implant (Fig. G, p < 0.01).
To explore whether RNA m 5 C modification can modulate the effects of Ti implant on macrophage regulation of osteogenic differentiation and angiogenesis, we analyzed the main genes related to m 5 C in macrophages of each groups. The RNA m 5 C dot blotting data suggested that m 5 C levels in macrophages of LPS group were higher than that in control group while Ti implant attachement down-regulated the total m 5 C levels in macrophages (Fig. A, p < 0.01). Furthermore, m 5 C related genes including NOP2, NSUN2, NSUN3, NSUN4, NSUN5, NSUN6, NSUN7, and TRDMT1 were detected in three groups of macrophages, and the qPCR analysis then demonstrated that NOP2, NSUN2, NSUN3, NSUN5, NSUN6, and NSUN7 were significantly highly expressed in macrophages of LPS group compared with the control group. After attached to Ti implant, NSUN2 level was further significantly down-regulated in macrophages while levels of NOP2, NSUN3, NSUN5, NSUN6, and NSUN7 were not significantly changed (Fig. B, p < 0.05). Subsequently, the regulatory role of NSUN2 was investigated. After transfection of macrophages, the expression of NUSN2 increased by more than 10 times compared with the control vector group according to the qPCR analysis (Fig. A, p < 0.01). Overexpression of NSUN2 weakened the promoting effect of Ti implant on M1 polarization of macrophages and enhanced the promoting effect of M2 polarization of macrophages, which showed an increase in iNOS protein level and a decrease in CD206 protein level, as well as an increase in the gene levels of IL6, TNFA and IL1B, and a decrease in the gene levels of TGFB1, IL10 and ARG1 (Fig. B - E, p < 0.01). Up-regulation of NSUN2 also weakened the effects of Ti implant on macrophages promoting the osteogenic differentiation of PDLSCs, which was manifested by decreased levels of VEGF, ACTA2 and COL1A1 genes and decreased levels of mineralization (Fig. F and G, p < 0.01). At the same time, the down-regulation of BGLAP, SPP1 and BMP2 genes also suggested that the up-regulation of NSUN2 impaired the ability of Ti implants to promote macrophage induction of HUVECs angiogenesis. (Fig. H, p < 0.01).
mRNAs in GSE173078 dataset are related to periodontitis and gingivitis, and inflammatory factors associated with periodontal disease were screened. The KEGG analysis of differentially expressed genes from GSE173078 dataset was performed, and the biological pathway suggested that IRF4 was enriched in several inflammatory signaling pathways (Fig. A). IRF4 has been reported to exert regulatory role in immune infiltration , macrophage polarization , and cell cycle . The suppression effect of IRF4 on osteogenic differentiation was also been empathized according to various studies . Therefore, we speculated that IRF4 may be m 5 C modified mediated by NSUN2 in macrophages. The m 5 C-RIP assay was implemented to assess the m 5 C modification status of IRF4 mRNA. The results exhibited that the level of IRF4 can be enriched by m 5 C antibody (Fig. B, p < 0.01). RIP followed by qPCR found that compared with IgG antibody, NSUN2 antibody significantly enriched IRF4 (Fig. C, p < 0.01). Subsequently, the potential m 5 C modification sites of IRF4 predicted by RNAm 5 Cfinder suggested that IRF4 may be m 5 C modified at three sites (Fig. D). After mutation at each of the three sites, the results of the double-luciferase gene report experiment suggested that there was no significant change in relative luciferase activity before and after NSUN2 over-expression at mutation sites 1 and 3, but after mutation at site 2, up-regulation of NSUN2 could significantly reduce relative luciferase activity (Fig. E, p < 0.01). Moreover, over-expression of NSUN2 promoted the degradation of IRF4 mRNA (Fig. F, p < 0.01). After transfection of macrophages, the expression of IRF4 increased by more than 10 times compared with the control vector group according to the qPCR analysis (Fig. A, p < 0.01), the macrophage polarization related proteins and genes were evaluated. Both of the immunofluorescence staining and qPCR analysis suggested that NSUN2 promoted the M1 macrophage polarization while IRF4 weakened the effects of NSUN2 by promoting M2 macrophage polarization (Fig. B - E, p < 0.01). Meanwhile, IRF4 also reversed the effects of NSUN2 on the regulatory function of macrophages on inhibition of osteogenic differentiation of PDLSCs and angiogenesis of HUVECs (Fig. F - H, p < 0.01).
Ti and its alloys are widely used in orthopedic implants due to their exceptional mechanical properties, chemical stability, and biocompatibility . The osteogenesis and angiogenesis capacity of peri-implant tissues are key indexes in the application of Ti implant . Interestingly, a high proportion of M2 macrophages infiltrates the damaged tissue and inhibits inflammation around the implant, which is conducive to the formation of implant osseous union and angiogenesis . Therefore, macrophages play an important role in the process of Ti implant therapy. Our data suggested that Ti implant showed favorable effects on promoting the M2 polarization transformation. Meanwhile, the M2 type macrophages promoted the osteogenic differentiation of PDLSCs and angiogenesis of HUVECs in vitro. The role of macrophages in promoting bone formation by Ti implants is multifaceted. Macrophages are not only the first cells contacted by implants after implantation, but also the main regulator of the integration of tissues and biomaterials in regulating the innate immune response . Macrophages play an important role in bone formation, bone remodeling and fracture healing, and can induce bone differentiation of mesenchymal stem cells . Different physicochemical and biological modifications on the surface of Ti implants can play a key role in implant bone binding by activating the M1 inflammatory polarization direction of macrophages or the M2 tissue healing direction . Meanwhile, implant surface modification also has an important impact on the induction of bone formation by macrophages. The future development direction is to explain the healing mechanism of implant-host interaction from the perspective of immunology, and develop a new type of Ti implant, which can induce bone formation and obtain homeostasis of bone coupling through the immune regulation of macrophages, so as to achieve early and long-term stable bone union. In this study, we found that Ti implant showed favorable effects on promoting the M2 polarization transformation and subsequent osteogenesis and angiogenesis in vitro. Afterwards, the underlying mechanism was studied. The RNA modification of m 5 C is mediated by the NSUN family (NSUN1-7) and DNMT homolog DNMT2. ALYREF and YBX1 RNA junction proteins called reader are responsible for identifying sites modified by m 5 C. DNMT2 and NSUN2 are writers of methyltransferase involved in the production of m 5 C modification. A methyltransferase called TET1 eraser can remove the m 5 C modification. They worked together to keep the m 5 C modification in dynamic balance . There is growing evidence that m 5 C methylation plays a role in gene expression and pathological processes in several human diseases by regulating mRNA stability splicing and protein translation. As for m 5 C studies related to angiogenesis and osteogenesis, RNA-binding protein YBX1 in angiogenesis-dependent bone formation and provided a therapeutic approach for ameliorating osteoporosis . Moreover, m 5 C modification of LINC00324 has been reported to promote angiogenesis in glioma . Therefore, we hypothesized that m 5 C modification has a regulatory role in the regulation of macrophages on osteogenesis and angiogenesis in Ti implant. Our data suggested that total m 5 C levels were elevated in LPS treated macrophages, and Ti implant decreased the m 5 C levels. Further, we found that this m 5 C modification is dominated by NSUN2. Interestingly, it has been found that NSUN2 mediated m 5 C modification contributes to the angiogenesis in glioma . Our data was also in line with this study that over-expression of NSUN2 significantly weakened the effects of Ti implant on macrophages, and the osteogenic differentiation of PDLSCs and angiogenesis of HUVECs. Subsequently, the bioinformatic analysis indicated that IRF4 is an aberrant expressed genes related to osteogenic differentiation, and is related to several inflammatory pathways. IRF4 has been reported to exert regulatory role in immune infiltration , macrophage polarization , and cell cycle . The suppression effect of IRF4 on osteogenic differentiation was also been empathized according to various studies . In this study, IRF4 can be enriched by m 5 C antibody and NSUN2 antibody. Moreover, over-expression of NSUN2 promoted the degradation of IRF4 mRNA. The promotion of M2 macrophage polarization induced by IRF4 was also been verified in this study, which is inline with the previous studies . Meanwhile, IRF4 also reversed the effects of NSUN2 on the regulatory function of macrophages on inhibition of osteogenic differentiation of PDLSCs and angiogenesis of HUVECs. With further research on the mechanism of titanium implants, it is possible to develop new types of titanium implants targeting NUSN2 and IRF4 in the future. For instance, drug delivery systems targeting NUSN2 and IRF4 may be designed on the implant surface to achieve localized drug release. Additionally, these implants not only promote bone formation through the immunomodulation of macrophages, but also achieve early and long-term stable bone union by optimizing angiogenesis. This will provide a more efficient and stable implant selection for the clinic, and improve the treatment effect and quality of life of patients. There are some limitation in current work. Which cytokines are secreted after LPS treatment of macrophages and how these factors affect angiogenesis or osteogenic differentiation should be further studied.
This study suggests that Ti implant induces the M2 type macrophage, which subsequently promotes angiogenesis and bone formation. This effect is likely mediated through the NSUN2 mediated m 5 C modification of IRF4.
|
Implementation Strategies and Ergonomic Factors in Robot-assisted Microsurgery | e21448c3-ee40-4579-91eb-3bc360d21eb5 | 11698882 | Robotic Surgical Procedures[mh] | Robot-assisted surgery has emerged as a compelling modality in the field of reconstructive microsurgery, offering potential benefits such as motion scaling for enhanced precision, minimized surgeon fatigue, and an improved ability to perform complex procedures . Since the first application of robot-assisted microsurgery utilizing the DaVinci® platform in 2007 by van Hulst and colleagues, several robotic systems specialized for microsurgical applications have been developed . These systems offer potentially superior ergonomics, the elimination of tremors, and an increased range of movement. However, their application in free tissue transfer and peripheral nerve surgery has remained relatively limited despite their potential to enhance surgical precision and reduce operative morbidity significantly. The Symani® robotic system consists of two robotic arms that are controlled by the surgeon via two manipulators similar to a pair of forceps. It offers 7 to 20-times motion scaling with the elimination of the physiological tremor. The instruments combined with the micro- and macropositioners provide seven degrees of freedom: X, Y, Z linear motion; roll, pitch, yaw and grip. The system has been utilized in a comparatively small number of cases in lymphatic surgery and free flap surgery [ – ]. Furthermore, it was also used for peripheral nerve surgery . It has also been evaluated for coronary bypass surgery and ophthalmic surgery in an experimental setting . The robotic system differs from the robot used primarily in general surgery, like the DaVinci®, by being designed for a single specialized task—microsurgical anastomosis—rather than performing the entire surgery. Free tissue transfer is a fundamental component of complex reconstructive surgeries. Despite the routine nature of these procedures, they can be technically challenging and require high levels of microsurgical expertise, particularly in the context of vascular anastomoses . Our key research question is to understand how the integration of a robot-assisted microsurgical system changed during the implementation phase in a single institution and what changes were made to optimize the workflow across multiple reconstructive scenarios. We have introduced robot-assisted microsurgery into our reconstructive armamentarium over a series of more than 80 cases and over 100 anastomoses and nerve coaptations with the Symani surgical system. In this manuscript, we evaluate our clinical experiences in the implementation of RAMS and aim to provide OR setups for improved workflow.
Data collection A prospective database was maintained, which included all cases of robot-assisted microsurgery. The clinical cases were not specifically selected to utilize the robotic system; instead, all cases scheduled for the operating theatre with the Symani robotic system were performed using the robot. The study adhered to the Declaration of Helsinki and was approved by the local ethics committee (Medical Commission Rhineland-Palatinate, Mainz, Germany; Protocol number: 2023–16997). Surgical technique The free flap reconstructions were performed using standard flap raising techniques as previously described . In each case, one or more anastomoses were performed using the Symani surgical system (Medical Microinstruments, Pisa, Italy). All free flap reconstructions were performed using the regular microinstruments for the Symani system and the super-microsurgical instruments were only used in lymphatic surgery cases. A conventional microscope (Mitaka MM51, Mitaka Kohki Ltd., Tokyo, Japan) or a digital exoscope supported by two 4 K-3D screens (Olympus OrbEye, Olympus K.K., Toyko, Japan) were used for optical magnification. Figure depicts an exemplary setup in the operating room performing a nerve reconstruction procedure in the upper extremity using the exoscope. Statistical analysis We present results as means ± standard deviation (S.D.) or median with interquartile range (IQR). For normally distributed data, statistical analysis was performed using Student’s t-test. Tests including non-normally distributed data were performed using the Mann–Whitney test. Normality was tested with the Kolmogorov–Smirnov test. Significance was defined as p < 0.05. Comprehensive data analysis was facilitated using GraphPad Prism Version 10.1.1 for Mac (GraphPad Software, San Diego, CA).
A prospective database was maintained, which included all cases of robot-assisted microsurgery. The clinical cases were not specifically selected to utilize the robotic system; instead, all cases scheduled for the operating theatre with the Symani robotic system were performed using the robot. The study adhered to the Declaration of Helsinki and was approved by the local ethics committee (Medical Commission Rhineland-Palatinate, Mainz, Germany; Protocol number: 2023–16997).
The free flap reconstructions were performed using standard flap raising techniques as previously described . In each case, one or more anastomoses were performed using the Symani surgical system (Medical Microinstruments, Pisa, Italy). All free flap reconstructions were performed using the regular microinstruments for the Symani system and the super-microsurgical instruments were only used in lymphatic surgery cases. A conventional microscope (Mitaka MM51, Mitaka Kohki Ltd., Tokyo, Japan) or a digital exoscope supported by two 4 K-3D screens (Olympus OrbEye, Olympus K.K., Toyko, Japan) were used for optical magnification. Figure depicts an exemplary setup in the operating room performing a nerve reconstruction procedure in the upper extremity using the exoscope.
We present results as means ± standard deviation (S.D.) or median with interquartile range (IQR). For normally distributed data, statistical analysis was performed using Student’s t-test. Tests including non-normally distributed data were performed using the Mann–Whitney test. Normality was tested with the Kolmogorov–Smirnov test. Significance was defined as p < 0.05. Comprehensive data analysis was facilitated using GraphPad Prism Version 10.1.1 for Mac (GraphPad Software, San Diego, CA).
From February until December 2023, 85 robot-assisted microsurgical operations were performed in our institution. The mean patient age was 53 ± 15 years. There were 55 males and 30 females in the study cohort. The average BMI was 26 ± 4.9 kg/m 2 . The median American Society of Anaesthesiologists Classification (ASA) was two with an interquartile range of one. Arterial hypertension was the most common comorbidity ( n = 34; 40.0%), followed by tobacco use ( n = 25, 29.4%) and adiposity (defined as BMI ≥ 30 kg/m 2 , n = 15, 17.6%). Table contains information on the patient characteristics. Most operations were carried out in the lower extremity ( n = 42, 49.2%). Thirty surgeries were performed on the upper extremity (35.3%). The rest of the operations were done for breast reconstruction ( n = 9, 10.5%) and in the head and neck region ( n = 4, 4.7%). For each anatomic location, an optimized setup of the devices in the operating room was identified and standardized in our institution (Figs. , , , ). Free flap reconstructions comprised the largest proportion of cases ( n = 68, 80%). There were ten cases of nerve transfers (11.7%), four TMRs (4.7%), and two cases of lymphovenous anastomoses (2.4%). In one case, a vein graft was used to reconstruct the ulnar artery (1.2%). In the 17 upper extremity free flap cases, the microscope and exoscope were utilized at similar rates ( n = 7, 41.2% and n = 10, 58.8%). Among the 40 lower extremity-free flaps, the exoscope was used more frequently ( n = 26, 65%) than the microscope ( n = 14, 35%). In a single head/neck case, the exoscope was used (25%), and in three the microscope was chosen (75%). With the exception of one case, all breast reconstructions were performed using the microscope ( n = 1, 11.1% and n = 8, 88.9%, respectively). Figure All 14 nerve cases were done using the exoscope (of those, twelve cases were performed in the upper extremity and two in the lower extremity). Both LVA cases were performed using the microscope. Figure depicts an overview of the utilized optical magnification devices. The mean operating time of the free flaps performed with the microscope was 367 ± 101 min with a range of 183 to 550 min, while flaps that were performed with the exoscope had a mean operating time of 402 ± 99 min with a range of 166 to 568 min. No statistically significant difference was observed when comparing overall operating times ( p = 0.53). The mean time per stitch during venous anastomoses using the microscope was 3.5 ± 2.4 min, while 4.0 ± 1.2 min when using the exoscope. Performing venous anastomoses with the exoscope required significantly more time per stitch ( p < 0.001). During arterial end-to-end anastomoses, the mean time per stitch with the microscope ( n = 10) was 3.0 ± 0.5 min and 6.4 ± 4.9 min when using the exoscope ( n = 8). In arterial end-to-side anastomoses, the mean time per stitch was 2.7 ± 0.5 min when using the microscope ( n = 13) and 4.9 ± 1.8 min when the exoscope was used ( n = 5). Both types of arterial anastomoses required significantly longer times per stitch when performed with the exoscope ( p < 0.04 and p < 0.005, respectively). The mean time per stitch of epineural coaptations using the exoscope ( n = 17) was 5.0 ± 1.5 min and 2.3 ± 0.8 min when using the microscope ( n = 6). Epineural coaptations took significantly longer when performed with the exoscope ( p < 0.001). Figure shows an overview of the times per stitch of the procedures mentioned above. Throughout our 85 cases, the 3D exoscope quickly became the primary choice and was used in most cases after its introduction in our department. Figure depicts this evolution in the usage of optical magnification across all included cases.
While robotic-assisted surgery has been widely implemented in the fields of urology, general surgery, and gynecology over the last 25 years, there has historically been a distinct lack of development in the realm of microsurgery . In recent years, however, several robotic systems have been developed specifically for microsurgical applications . Most notably, the MUSA system has been used for lymphatic surgery in a series of clinical cases . In February 2023, we implemented a program for robotic-assisted microsurgery in our department utilizing the Symani surgical system. Since then, we have applied the Symani system in a wide range of reconstructive cases, including extremity defects, scalp reconstruction, lymphatic surgery, and autologous breast reconstruction. With this manuscript, we aim to present our experiences after performing nearly one hundred consecutive cases of robot-assisted microsurgery over a ten-month period with a focus on the different types of magnification systems and logistical considerations necessary for robotic microsurgery. Traditional microscopes presented operational challenges due to the necessity for the surgeon to be in close proximity to the operating table or patient, thereby restricting the range of movements when using the robotic controllers. Furthermore, the operating table may interfere with the signal transduction between controllers and the robot. These issues were addressed either through meticulous patient and operator positioning or by employing an exoscope which provided superior three-dimensional visualization (Olympus OrbEye, Olympus, Tokyo, Japan). Conventional microscopy offers the advantages of instantaneous image transmission as well as higher contrast and resolution. In our opinion, the superior image resolution makes the microscope the preferred choice for supermicrosurgery. As we show in our data (Fig. ), the exoscope has replaced it in almost all other scenarios, as the ergonomic benefits are greater, and the provided resolution is sufficient for regular microsurgical applications. With digital exoscopes, the microsurgeon can be positioned freely in the operating room and, therefore, has more freedom of arm and hand movement. The positioning and setup of the robotic platform alongside the microscope or exoscope require substantial forethought. Improper positioning of the robot or microscope can create ergonomic challenges and potentially impact both the surgical procedure time and the quality of microsurgical outcomes. When implementing an exoscope, particular attention must be paid to ensuring optimal screen visibility, not only for the primary surgeon operating the robotic arms but also for the assistant and scrub nurses. To address this, we acquired a second large screen to improve the visibility for the assistant and nursing staff, instead of the small secondary screen (see Fig. ). Throughout our case series, we observed a trend toward increased use of the exoscope following its introduction in our department (see Fig. ). Today, conventional operating microscopes are used almost exclusively in cases of lymphatic surgery or other supermicrosurgical applications in our institution, where due to the very small vessel sizes, the higher resolution and contrast outweigh the ergonomic benefits of an exoscope, in our opinion (see Fig. ). In these cases, it is paramount to position the patient and microscope as close as possible to the edge of the operating table. Otherwise, the microsurgeon may not have sufficient space to freely operate the robotic controllers. In a previous study we were able to demonstrate that the use of exoscopes provides significant ergonomic benefits, particularly in anatomical regions subject to the most strain during microsurgical procedures . The results of this study and our clinical practice suggest that surgeon comfort and clinical context drive the choice of magnification system. Furthermore, using an exoscope comes with its own learning curve and disadvantages . We observed a slight delay in visualization on the 3D screens when using the OrbEye exoscope. Additionally, the surgical team must adapt to the use of the screens with 3D lenses and the overall altered operating room setup. These factors might explain the significantly longer time per stitch when using an exoscope in all types of anastomoses (see Fig. ). Even when considering the aforementioned disadvantages of using an exoscope, we found that its ergonomic benefits outweigh these drawbacks in our experience. In an experimental study, Wessel and colleagues analyzed the posture during robot-assisted microsurgery using the rapid entire-body assessment. Their findings demonstrated significantly better ergonomic scores in the group performing robot-assisted microsurgery when compared to the group performing conventional microsurgery, indicating improved ergonomics with robotic assistance . We strongly advocate standardizing the setup of the robotic system and the magnification source according to the specific use case. For instance, our experience indicates that extremity reconstruction necessitates different setup parameters and patient positioning compared to a breast reconstruction. Establishing such standardized protocols before surgery may considerably reduce operating room time and improve ergonomics. We usually position the assistant between the robotic arms and the main operator farther away, using an exoscope (see Figs. , , , and ). This setup is particularly advantageous in confined surgical sites and if there is a significant step-off of the surgical site, such as when anastomosing to the proximal anterior tibial artery, for example. This way, the assistant does not need to operate with a steep angle of his microsurgical instruments and benefits from much improved ergonomics. In a case series of lymphatic reconstructions, Weinzierl et al. proposed that robot-assisted microsurgery is especially useful in limited and/or deep surgical sites . Throughout our initiation into robot-assisted microsurgery, several challenges were encountered that warrant discussion for the broader surgical community. Similar to the experiences documented by Lindenblatt et al., we also observed specific issues with the robotic system's instruments, notably a tendency for these tools to become sticky after brief periods of use . Our preferred method for maintaining instrument cleanliness involved applying a polyvinyl alcohol wipe two to three times during an anastomosis (Raucocel, Lohmann & Rauscher, Rengsdorf, Germany). A frequent rinse of the instruments with a diluted heparin solution also proved helpful. In conventional microsurgery, the operating room nurse typically cleans the microsurgical instruments during every instrument change. In robot-assisted microsurgery, this task falls to the surgical assistant since the nurse is usually farther away from the robotic instruments, which are fixed in place. The initial absence of haptic feedback may seem daunting; however, surgeons can rapidly develop visual feedback mechanisms, such as when tying sutures. Consequently, mishandling sutures or applying excessive force to tissues did not pose significant problems in our experience. In contrast to our subjective experience, Beier et al. reported their case series of 23 robot-assisted free flaps and identified the lack of haptic feedback as one of the primary drawbacks, particularly during knot tying . Another paramount hurdle in the introduction of robotic microsurgery are the high purchasing and operational costs of the robotic system, which are currently not reimbursed in the German healthcare system. Although we observed a reduction in anastomosis times over time in our previous study, employing the Symani robotic system still required more time than a conventional anastomosis . This is partially attributable to motion scaling, which slows the surgeon's movements to enhance precision. In a previous study, we analyzed the first fifty cases of robot-assisted microsurgery in our department. We identified a shallow learning curve without a statistically significant improvement in operating times, underscoring the importance of a sustained high caseload . Prolonged surgical times have been reported by nearly all studies exploring robot-assisted microsurgery to date [ , , – ]. In our cohort, we also found a prolongation of the anastomotic times with arterial end-to-side anastomoses requiring a mean of 37 min . Overall, we did not find a significant difference in overall operating times when comparing cases done using an exoscope and a microscope. In part, these prolonged times have to be expected since the surgeon’s movements are downscaled by 7 to 20 times, thereby slowing overall motion. To improve the anastomotic times, we usually do not use the integrated suture cutter and instead have the assistant cut the sutures. In our experience, the integrated suture cutter is cumbersome and imprecise, as it is integrated into the proximal part of the needle holder. Microscissors that would enable microsurgical dissection are currently unavailable, but their introduction could be beneficial when further preparation of the vessels is needed. Additionally, we have found that having the assistant pass the suture through after the operator places the stitch saves a small but notable amount of time. In general, the assistant plays a far more active and essential role compared to conventional microsurgery. Their tasks include cleaning the robotic instruments, stabilizing the vessel or nerve, and managing the sutures. Furthermore, we advise fully starting up the robotic system so that it is immediately ready after moving it into the operating position. Since the robotic instruments are solely aimed at microsurgical suturing, further dissection of the vessels is not feasible. It is, therefore, essential to prepare both the recipient site and the vascular pedicle thoroughly using loupes before initiating robotic suturing to ensure an efficient and timely anastomosis afterwards. Evaluating future perspectives of robot-assisted microsurgery in the field of plastic surgery, Henn et al. highlighted the potential of including artificial intelligence (AI) into robot-assisted microsurgery in a narrated review . In the future, AI-driven automation and increased assistance could be used to automatically adjust the motion scaling to the appropriate amount for different tasks, such as making precise needle punctures, passing the suture through or autonomously zooming in and out. Combining the robotic system with a digital exoscope would seem especially beneficial in this scenario. Such integration could enable the unification of controls for both the robotic system and the exoscope, potentially facilitating a smoother and more effortless surgical workflow. However, even in the face of such potential benefits, we experienced drawbacks. During microsurgical breast reconstruction, we found difficulties due to the respiratory excursion of the thorax during anastomosis. Due to the static position of the robotic instruments relative to the mobile chest wall, the respiratory movements are much more pronounced than during conventional microsurgery, where the microsurgeon rests their hands on the patient, at least partially compensating for the respiratory excursions. In the setting of robot-assisted microsurgery, this issue may be mitigated by adjusting the respirator settings, such as temporarily decreasing the tidal volume, in coordination with the anesthesiologist. In patients with peripheral arterial occlusive disease (PAOD) and heavily calcified arteries, the robot may struggle to penetrate the rigid arterial wall, necessitating conversion to a traditional hand-sewn anastomosis. While our study adds to the growing body of scientific literature on robot-assisted microsurgery, it is not without limitations. Firstly, there is a relatively low number of cases, especially in the group of lymphatic surgery. Additionally, a rather small group of surgeons performed the surgeries, which may have introduced a performance bias. Furthermore, a randomized study comparing the use of exoscopes and microscopes would have delivered more conclusive results than this retrospective study design. Despite these challenges, we believe that robot-assisted microsurgery has the potential to revolutionize the field of supermicrosurgery, expanding its accessibility to a broader group of microsurgeons. This may ultimately improve patient outcomes and decrease donor site morbidity, while concurrently reducing the physical strain on surgeons.
The full integration of RAMS into the clinical routine practice requires careful and standardized OR setups. We showed that there is a preference for the utilization of digital exoscopes over conventional microscopes in RAMS, despite requiring more time per stitch when using the exoscope. Furthermore, we presented OR setups for various reconstructive applications using RAMS. Nonetheless, further research and development are necessary to make robot-assisted microsurgery more widely accessible.
|
Asbestos Burden in Lungs of Subjects Deceased From Mesothelioma Who Lived in Proximity to an Asbestos Factory: A Topographic Post‐Mortem SEM‐EDS Study | e70eaff6-ee10-4b0f-a4e2-96eca20b2f2a | 11731493 | Forensic Medicine[mh] | Introduction The term “asbestos” indicates six fibrous minerals (chrysotile, crocidolite, amosite, anthophyllite asbestos, tremolite asbestos, and actinolite asbestos) that are well known to cause a number of diseases, collectively called asbestos‐related diseases . Among these, malignant mesothelioma (MM) is of particular interest, given the poor prognosis, the lack of effective therapies and the exceptionally long latency . Namely, 30–50 years typically elapse between the start of exposure and the onset of the disease . For this reason, we are still observing the detrimental effects of exposures that ceased several decades ago. In Italy, asbestos was banned in 1992 with the Italian Law 257/92, but the MM epidemic is still ongoing. According to epidemiological forecasts, pleural MM incidence is not expected to decrease significantly in the next few years . Moreover, asbestos and several asbestos‐like minerals (i.e., asbestiform minerals not asbestos classified) are still widely present in urban and natural environments: this means that all people (the general population) are potentially exposed . However, the levels of exposure and the consequent actual amount of inhaled asbestos are not easy to determine with high scientific certainty. Asbestos lung burden, measured using analytical electron microscopy (scanning or transmission electron microscopy equipped with energy dispersive spectroscopy) is generally accepted as the most reliable marker of past asbestos exposure, especially in postmortem samples and in medico‐legal contexts . Notwithstanding, this method presents some pitfalls, first of all the difficulty (if not impossibility) to obtain data about past exposure to chrysotile, due to the low biodurability of this type of asbestos, compared to amphiboles . In this article, we investigate the asbestos lung burden in subjects deceased from MM who lived in Broni, a small town in Northern Italy, where an important asbestos‐cement factory (Fibronit) was located and active between 1932 and 1993; in total, 3455 workers were employed there . At this plant, asbestos cement artifacts were produced mainly composed of a mixture of chrysotile and crocidolite, while amosite was used in smaller quantities as an additive . Data about environmental pollution due to asbestos in Broni and surrounding towns during the most intense activity of the industry (60–70s) are not available. However, epidemiological studies revealed that people who lived in Broni or worked at Fibronit factory were heavily exposed to asbestos . It is important to specify that not all asbestos exposure derived from the Fibronit factory, as, during its activity, asbestos‐containing materials were widely diffused in domestic settings (e.g., asbestos boards used for isolation of radiators, ironing) and in house construction (e.g., insulation, roofing, chimneys) . However, most of the asbestos pollution in the town was related to the heavy air dispersion of asbestos deriving from the Fibronit factory, where measures to reduce air dispersion were totally absent during most of the production period . In line with this, a recent study about the spatial patterns of MM incidence in the Lombardy region (the county that includes Broni) clearly showed a large spatial cluster of MM cases with previous anthropogenic environmental exposure around Broni (within a radius of 30 km) . It is important to underline that, besides Fibronit, two more asbestos‐cement factories were located in Pavia province, one in Portalbera (fewer than 50 people worked there), and one in Arena Po (employing less than 75 people). These towns are, respectively, 7 and 9.5 km away from Broni. Therefore, these smaller plants are potential confounding factors in the evaluation of asbestos lung burden in relation to the distance from the Fibronit factory. However, the National Mesothelioma Registry never registered cases related to occupational exposure in these two plants . No data about the possible anthropogenic environmental pollution provoked by these two factories are available. A few previous studies, conducted by our group and others, investigated the asbestos lung burden of people living nearby asbestos‐using factories, finding that the effects of environmental exposure are as high as those deriving from occupational exposure . Only two studies, to our knowledge, focused on the lung content of non‐occupationally exposed Broni inhabitants . However, a geographical study taking into consideration the precise location of subjects' residencies in relation to the asbestos lung burden has never been conducted. The main objective of this study is to understand if asbestos lung content is different between occupationally and environmentally exposed individuals and if the distance between the subjects' residences and the factory is significantly associated with the asbestos lung burden in Broni inhabitants. Methods 2.1 Participants The present retrospective, observational study was conducted on individuals deceased from MM selected from the archive of our Unit of Legal Medicine between 2005 and 2019. For each subject, a forensic autopsy, followed by a complete histopathological examination, was ordered by the Public Prosecutor due to the hypothesis of manslaughter related to nonobservance of safety measures at the workplace. During each autopsy, the whole lungs were collected and formalin fixed. The diagnosis of MM, mostly already known in life, was confirmed postmortem through immunohistochemistry according to the guidelines in effect at the time . The vast majority of the subjects involved in this study were exposed to asbestos, occupationally or environmentally, in Broni, a small town in Pavia Province, northern Italy. 2.2 Sample Preparation for Scanning Electron Microscope Equipped With Energy Dispersive X‐ray Spectroscopy (SEM‐EDs) The technique here used was the same described elsewhere . In summary, 0.25 g formalin‐fixed lung samples were chemically digested using 13% sodium hypochlorite and then filtered through a cellulose‐ester membrane (Millipore, Darmstadt, Germany) with a diameter of 25 mm and a pore size of 0.45 µm. The membrane was prepared and observed using a SEM‐EDS (Oxford Inca Energy 200), with an INCA X‐act SDD detector (Oxford Instruments NanoAnalysis, Bucks, UK). For each membrane, an area of 2 mm 2 was observed at a magnification of 4000× using both secondary and backscattered electrons. The EDS spectra allowed to identify the asbestos type through the comparison with a reference database, with the following exceptions: it is not possible to distinguish unequivocally chrysotile from asbestiform antigorite and tremolite asbestos from actinolite asbestos using SEM‐EDS, since they have similar chemical composition and analogous morphology, therefore we used, respectively, the term chrysotile/asbestiform antigorite and tremolite/actinolite asbestos in the classification of these minerals. The concentration of asbestos and asbestos bodies (ABs) are expressed in terms of asbestos fibers and ABs per gram of dry weight of lung tissue (respectively, ff/gdw, and ABs/gdw), as indicated by international guidelines. The concentration was obtained normalizing the amount of asbestos fibers and ABs observed in an area of 2 mm 2 to 1 g of dry tissue . While the preparation of all samples was performed in the same laboratory, the SEM‐EDS examination was performed in two laboratories, with the samples divided equally between the two. To avoid the variability caused by different equipment and microscopists, we established a detailed, standardized protocol for data collection. A regular control between the laboratories was performed by comparing the images and spectra of each laboratory. In addition, five samples were analyzed in both laboratories and the results were compared using the analysis of variance test . 2.3 Variables For each subject, demographics and exposure variables were extracted from an informatic archive. History of exposure was defined as occupational, if the subject was an asbestos worker (most of them used to work at Fibronit factory) or anthropogenic environmental if the subject used to live close to the asbestos‐cement factory located in Broni. Defining each subject according to the history of exposure, occupational exposure was considered prevalent over anthropogenic environmental. While some of the occupationally exposed subjects had also environmental exposure (and were defined as occupational exposed), for environmentally exposed subjects any history of occupational exposure has been excluded. In this paper, subjects with familial exposure, defined as asbestos exposure through a family member who work with asbestos, were not included, since we wanted to look at the association between distance from the source of asbestos contamination and asbestos lung content; the inclusion of subjects with familial exposure could have introduced a bias in the results' interpretation. For simplicity, in this article, we use the term “environmental exposure” referring to anthropogenic environmental exposure (not to natural sources of asbestos). In addition, year of birth, year of the beginning of asbestos exposure, date (month, year) of the MM diagnosis, date (month, year) of death were extracted, along with the address or addresses of residency, and for each address, the duration of residency. The following endpoints were assessed through SEM‐EDS: Concentration of asbestos, expressed as a number of fibers per gram of dry weight (ff/gdw). Mean length and width of detected asbestos fibers (in μm). The concentration of each type of asbestos (ff/gdw), classified as chrysotile/asbestiform antigorite, crocidolite, amosite, tremolite/actinolite asbestos, and anthophyllite asbestos. Concentration of ABs, expressed as ABs/gdw. In the present work we took into consideration only asbestos fibers longer than 5 µm, thinner than 3 µm, and with an aspect ratio greater than or equal to 3:1), according to the WHO definition of “biologically critical fiber” . These criteria also define the concept of “regulated” asbestos fiber according to Italian law. 2.4 Statistical Methods Environmental cases within 10 kilometers of the Fibronit factory were plotted onto two maps: MAP 1 [Figure ] using addresses of the longest time residence for each observation and map2 [Figure ] using the closest addresses in relation to the factory independently from the time lived at that residence. Addresses were plotted in conjunction with wind data collected from the Broni weather station. Geometric distances in meters were calculated between addresses and the factory. Distributions of asbestos measures across exposure types (occupational and environmental) were compared using two‐sided Wilcoxon Rank‐Sum tests. Analysis of covariances (ANCOVAs) using rank‐transformed asbestos measures were also conducted to compare the measures across exposure types when adjusting for age at death and sex. Spearman correlation coefficients were calculated and assessed between asbestos measures and residential distance from factory, duration of exposure within Broni, and a ratio of exposure duration divided by the distance from factory. Partial Spearman correlations were calculated for the same variables; duration of exposure and ratio of exposure and duration were sex‐adjusted, while distance from the factory was adjusted for sex and age at death. Environmental cases within 10 km of the factory were split into three groups based on proximity to the factory [(1) first quartile of longest duration address (2): second and third quartile of addresses (3), fourth quartile of addresses]: less than 557 m, 557–926 m, and greater than 926 m. These groups as well as overall environmental cases within 10 kilometers had asbestos measures compared with occupational cases using two‐sided Wilcoxon Rank‐Sum tests. ANCOVAs with rank‐transformed asbestos measures also compared exposure types when adjusting for age at death. Geospatial visualization, distance calculation, and map composition were performed using QGIS, version 3.32.2‐Lima ( QGIS.org , %Y. QGIS Geographic Information System. QGIS Association. http://www.qgis.org ). Statistical analyses were performed using SAS software, version 9.4 (SAS Institute, Cary, NC, USA). 2.5 Ethics Review and Approval The study protocol was approved by Ethical Committee “Lombardia 6” on August 22, 2023. The informed consent was waived because the participants are deceased subjects and relative are no longer reachable. Only biological material retrieved from our archive, collected at the time of autopsy for forensic and diagnostic purposes, was used for this study. Participants The present retrospective, observational study was conducted on individuals deceased from MM selected from the archive of our Unit of Legal Medicine between 2005 and 2019. For each subject, a forensic autopsy, followed by a complete histopathological examination, was ordered by the Public Prosecutor due to the hypothesis of manslaughter related to nonobservance of safety measures at the workplace. During each autopsy, the whole lungs were collected and formalin fixed. The diagnosis of MM, mostly already known in life, was confirmed postmortem through immunohistochemistry according to the guidelines in effect at the time . The vast majority of the subjects involved in this study were exposed to asbestos, occupationally or environmentally, in Broni, a small town in Pavia Province, northern Italy. Sample Preparation for Scanning Electron Microscope Equipped With Energy Dispersive X‐ray Spectroscopy (SEM‐EDs) The technique here used was the same described elsewhere . In summary, 0.25 g formalin‐fixed lung samples were chemically digested using 13% sodium hypochlorite and then filtered through a cellulose‐ester membrane (Millipore, Darmstadt, Germany) with a diameter of 25 mm and a pore size of 0.45 µm. The membrane was prepared and observed using a SEM‐EDS (Oxford Inca Energy 200), with an INCA X‐act SDD detector (Oxford Instruments NanoAnalysis, Bucks, UK). For each membrane, an area of 2 mm 2 was observed at a magnification of 4000× using both secondary and backscattered electrons. The EDS spectra allowed to identify the asbestos type through the comparison with a reference database, with the following exceptions: it is not possible to distinguish unequivocally chrysotile from asbestiform antigorite and tremolite asbestos from actinolite asbestos using SEM‐EDS, since they have similar chemical composition and analogous morphology, therefore we used, respectively, the term chrysotile/asbestiform antigorite and tremolite/actinolite asbestos in the classification of these minerals. The concentration of asbestos and asbestos bodies (ABs) are expressed in terms of asbestos fibers and ABs per gram of dry weight of lung tissue (respectively, ff/gdw, and ABs/gdw), as indicated by international guidelines. The concentration was obtained normalizing the amount of asbestos fibers and ABs observed in an area of 2 mm 2 to 1 g of dry tissue . While the preparation of all samples was performed in the same laboratory, the SEM‐EDS examination was performed in two laboratories, with the samples divided equally between the two. To avoid the variability caused by different equipment and microscopists, we established a detailed, standardized protocol for data collection. A regular control between the laboratories was performed by comparing the images and spectra of each laboratory. In addition, five samples were analyzed in both laboratories and the results were compared using the analysis of variance test . Variables For each subject, demographics and exposure variables were extracted from an informatic archive. History of exposure was defined as occupational, if the subject was an asbestos worker (most of them used to work at Fibronit factory) or anthropogenic environmental if the subject used to live close to the asbestos‐cement factory located in Broni. Defining each subject according to the history of exposure, occupational exposure was considered prevalent over anthropogenic environmental. While some of the occupationally exposed subjects had also environmental exposure (and were defined as occupational exposed), for environmentally exposed subjects any history of occupational exposure has been excluded. In this paper, subjects with familial exposure, defined as asbestos exposure through a family member who work with asbestos, were not included, since we wanted to look at the association between distance from the source of asbestos contamination and asbestos lung content; the inclusion of subjects with familial exposure could have introduced a bias in the results' interpretation. For simplicity, in this article, we use the term “environmental exposure” referring to anthropogenic environmental exposure (not to natural sources of asbestos). In addition, year of birth, year of the beginning of asbestos exposure, date (month, year) of the MM diagnosis, date (month, year) of death were extracted, along with the address or addresses of residency, and for each address, the duration of residency. The following endpoints were assessed through SEM‐EDS: Concentration of asbestos, expressed as a number of fibers per gram of dry weight (ff/gdw). Mean length and width of detected asbestos fibers (in μm). The concentration of each type of asbestos (ff/gdw), classified as chrysotile/asbestiform antigorite, crocidolite, amosite, tremolite/actinolite asbestos, and anthophyllite asbestos. Concentration of ABs, expressed as ABs/gdw. In the present work we took into consideration only asbestos fibers longer than 5 µm, thinner than 3 µm, and with an aspect ratio greater than or equal to 3:1), according to the WHO definition of “biologically critical fiber” . These criteria also define the concept of “regulated” asbestos fiber according to Italian law. Statistical Methods Environmental cases within 10 kilometers of the Fibronit factory were plotted onto two maps: MAP 1 [Figure ] using addresses of the longest time residence for each observation and map2 [Figure ] using the closest addresses in relation to the factory independently from the time lived at that residence. Addresses were plotted in conjunction with wind data collected from the Broni weather station. Geometric distances in meters were calculated between addresses and the factory. Distributions of asbestos measures across exposure types (occupational and environmental) were compared using two‐sided Wilcoxon Rank‐Sum tests. Analysis of covariances (ANCOVAs) using rank‐transformed asbestos measures were also conducted to compare the measures across exposure types when adjusting for age at death and sex. Spearman correlation coefficients were calculated and assessed between asbestos measures and residential distance from factory, duration of exposure within Broni, and a ratio of exposure duration divided by the distance from factory. Partial Spearman correlations were calculated for the same variables; duration of exposure and ratio of exposure and duration were sex‐adjusted, while distance from the factory was adjusted for sex and age at death. Environmental cases within 10 km of the factory were split into three groups based on proximity to the factory [(1) first quartile of longest duration address (2): second and third quartile of addresses (3), fourth quartile of addresses]: less than 557 m, 557–926 m, and greater than 926 m. These groups as well as overall environmental cases within 10 kilometers had asbestos measures compared with occupational cases using two‐sided Wilcoxon Rank‐Sum tests. ANCOVAs with rank‐transformed asbestos measures also compared exposure types when adjusting for age at death. Geospatial visualization, distance calculation, and map composition were performed using QGIS, version 3.32.2‐Lima ( QGIS.org , %Y. QGIS Geographic Information System. QGIS Association. http://www.qgis.org ). Statistical analyses were performed using SAS software, version 9.4 (SAS Institute, Cary, NC, USA). Ethics Review and Approval The study protocol was approved by Ethical Committee “Lombardia 6” on August 22, 2023. The informed consent was waived because the participants are deceased subjects and relative are no longer reachable. Only biological material retrieved from our archive, collected at the time of autopsy for forensic and diagnostic purposes, was used for this study. Results The present study includes 77 subjects (58.4% males and 41.6% females) who were exposed to asbestos and died from MM between 2005 and 2016. The mean age at death was 69.5 years (standard deviation [SD] 11.42). The past asbestos exposure (according to the medical history and the forensic records) was occupational in 48.0% of cases, environmental in 52.0%. Concerning the histological type, 74.0% had epithelial MM, 9.0% had sarcomatoid MM, 13% had biphasic MM and 4.0% had desmoplastic MM. Only two cases had peritoneal MM, while the other 75 (97.4%) had pleural MM. The mean duration of exposure was 22.1 years (SD = 15.6), the mean latency between the beginning of exposure and the diagnosis of MM was 49.6 years (SD = 13.0), while the mean time elapsed between the cessation of exposure and death was 26.1 years (SD = 10.6). The mean survival since diagnosis of MM was 18.0 months (SD 14.8). Overall, the mean concentration of asbestos in the lungs was 80341 ff/gdw (SD = 241,542). Asbestos was detected in 71.4% of subjects. Mean concentrations of chrysotile/asbestiform antigorite, crocidolite, amosite, anthophyllite asbestos, and tremolite/actinolite asbestos were, respectively, 438, 32,645, 38,890, 683, and 7684 ff/gdw. Chrysotile/asbestiform antigorite has been detected in five subjects out of 77 (6.5%), commercial amphiboles in 65% and noncommercial amphiboles in 50%. Six subjects (7.8%) had only tremolite/actinolite (noncommercial amphiboles) in their lungs. The mean concentration of ABs was 75,465 ABs/gdw (SD = 334,826). ABs were detected in 55.8% of subjects. The mean length and width of asbestos were, respectively, 23.8 μm and 0.7 μm. The concentration of asbestos nude fibers and ABs showed a significant correlation (Spearman correlation coefficient: 0.34, p ‐value = 0.0007). In Table , the variables related to lung burden were stratified by exposure type. No significant differences in asbestos concentration in the lungs were observed between the two exposure groups. Similarly, no differences were pointed out considering the concentration of each asbestos type, except for amosite, whose concentration is significantly higher in patients with occupational exposure compared to the environmental one. While there were no significant differences in mean fiber length, occupationally‐exposed observations showed significantly higher mean fiber widths. On the other hand, the concentration of ABs in the lungs of occupationally exposed people was significantly higher compared to those with environmental exposure. When adjusting for age at death and sex (Supporting Information: Table ), amosite remained significantly higher among occupationally exposed participants, while mean fiber width and AB concentration were no longer significantly different among the two groups. Geographic analysis : We performed a topographic analysis considering only environmentally exposed patients, to assess how the distance from the Fibronit factory influences the lung content. Given that 41 out of 77 subjects (53.2%) changed more than one address during their life, we considered, in the first map (Figure ), the address in which the person lived for the longest time, in the second map (Figure ) the address closest to the asbestos‐cement factory irrespective from the length of stay at that address. In both maps, the gray circle corresponds to the 10 km radius around the factory, considered, in literature, the typical area of neighborhood contamination . Most of the subject's residences were located on the east side of Broni, with respect to the factory. Notably, the prevalent wind direction is west‐east (Figures and ). In Table , the 39 cases with environmental exposure that lived within 10 km from the Fibronit factory were stratified according to quartiles of distance of their address from the factory, with the second and third quartiles combined into one group due to close proximity. The lung burden endpoints were compared to those observed in the 37 occupational cases. Among the environmentally exposed patients who lived closest to the factory (within 557 m, corresponding to the first quartile of distance), the concentration of ABs and amosite were significantly lower compared to the occupational exposed individuals, while asbestos concentration, asbestos dimensions and the concentration of chrysotile/asbestiform antigorite, crocidolite, anthophyllite asbestos, and tremolite/actinolite asbestos did not differ in the two groups. Among the environmental cases who lived between 557 and 926 meters from the factory (second and third quartile), only the concentration of ABs showed a statistically significant difference compared to the occupationally exposed patients. Among the environmental cases living furthest from the factory but within 10 km (926 m to 10 km, i.e., above the fourth quartile of distance), no measures show significant difference from the occupational cases. Similar results were obtained after adjusting for age at death (Supporting Information: Table ). Next, in the 39 environmentally exposed individuals who lived within a 10 km radius from the factory, we investigated possible correlations between the variables related to asbestos lung content and, respectively, duration of residence within the 10 km radius, distance from Fibronit (considering the address of longest residence) and the ratio between the two parameters (Table ). We find no significant associations between the duration/distance ratio and asbestos measures in the lungs. There are also no significant associations between distance and asbestos lung content except for lung ABs concentration. However, there are significant associations between the duration of exposure and mean width of asbestos, asbestos concentration, ABs concentration, amosite, and tremolite/actinolite asbestos. There were no significant changes in association for the correlations between time‐distance exposure metrics and asbestos measures when adjusting for sex and age of death (Supporting Information: Table ). Discussion In the present work we assessed the asbestos lung burden in a series of 77 subjects deceased from MM and compared lung burden between subjects with occupational vs environmental exposure, finding that the concentration of asbestos uncovered fibers (either as a whole and considering the concentration of each type of asbestos) did not differ significantly, while the concentration of ABs was higher in occupational exposure compared to environmental. We suggest that physiologic lung clearance may be responsible for these findings. Moving to the correlation between asbestos lung content and the address of residency, we found that, in general, the vast majority of the subjects environmentally exposed lived on the east side of the town. When we compared environmental cases stratified according to the distance from the factory and occupational ones, we found differences only in the concentration of ABs and amosite. Moreover, the concentration of ABs was higher in subjects who lived closer to the factory, but this correlation was not observed for uncovered asbestos fibers. This correlation was not confirmed considering the duration/distance ratio. Before discussing the above‐summarized results, it is necessary to clarify that at the asbestos‐cement factory located in Broni a mixture of chrysotile and crocidolite, with smaller amounts of amosite, were used for producing asbestos‐cement artifacts . Notably, at the Fibronit factory, no safety measures to reduce fiber dispersion were adopted for most of the production period . Therefore, we know that the subjects of this study have been exposed to chrysotile/asbestiform antigorite, which, due to its crystalline structure, is characterized by a much lower biodurability, being cleared from lungs very quickly . For this reason, we detected small amounts of chrysotile/asbestiform antigorite in only five cases out of 77, despite the fact that the subjects have been exposed to chrysotile/asbestiform antigorite, crocidolite and amosite. Therefore, as previously noted , SEM‐EDS investigation of lung samples is suitable to detect mainly amphibole fibers. However, clearance does not involve only chrysotile, but, to a lesser extent, also amphiboles . This phenomenon is not well understood, but the existing data, mainly deriving from studies on animal models, show that lung content is subjected to modifications over time. Considering that in this series of subjects a mean of 21 years elapsed between the end of exposure and death, it is likely that the asbestos lung content detectable on autopsy lung samples changed over time and it is different from what would have likely been observed at the time of cessation of exposure. The comparison of the concentration of uncoated asbestos fibers according to different types of exposure showed no significant differences except for amosite, whose concentration was higher in occupationally exposed individuals compared to environmentally exposed ones. This may reflect a preferential use of amosite in the production of specific asbestos cement materials (ACM) (i.e., friable ACM) which has caused a greater amount of amosite air dispersion in the Broni area and therefore inside dwellings. However, to our knowledge, this was never reported in literature. On the other hand, ABs concentration was higher in asbestos workers compared to subjects with environmental exposure. This may reflect the fact that ABs, once formed, cannot be cleared from the lungs as they are sequestered in the interstitium, while uncovered fibers (especially chrysotile) are fragmented and removed by macrophages . Therefore, in cases autopsied long after the end of exposure, the finding of ABs can represent the only remnant sign of asbestos, since by then most asbestos‐uncovered fibers have been already cleared from the lungs. However, clearly there is a different individual tendency to form ABs. As suggested by Dodson, some individuals are “poor coaters” , even though the reason is still not fully understood. Our results are in line with this hypothesis, considering that we found ABs in around half of the 77 subjects. Notably, in seven individuals with a concentration of asbestos‐uncovered fibers under the detection limit we found ABs, suggesting that asbestos‐uncovered fibers had been cleared from the lungs and are no more detectable, as opposed to ABs, that remain the only detectable portion of asbestos. Therefore, the higher concentration of ABs detected in occupationally exposed subjects might simply reflect the (likely) higher exposure to asbestos, most of which is known to have been chrysotile. This explanation is corroborated by the generally low concentrations of asbestos uncoated fibers in highly exposed subjects, often under the threshold used for MM attribution to past asbestos exposure according to the Helsinki criteria , or even under the detection limit of our technique. In terms of the topographic analysis of environmentally exposed individuals, this is the first study to our knowledge that evaluates lung content in relation to distance from the industrial source of asbestos dispersion. We mapped the cases according to the longest residency (Figure ) and according to the closest address to the factory (Figure ), obtaining visually similar results. Most cases lived on the east side of the town with respect to the factory, mostly because that is where the city center is. Unfortunately, the prevalent wind direction is west‐east and this could have contributed to determine the heavy environmental exposure of Broni residents. When we investigated possible correlations between asbestos lung burden and the distance from the factory in the 39 environmentally exposed subjects who lived within 10 km from Fibronit, we found that the concentration of asbestos uncoated fiber did not increase significantly in proximity to the factory, as opposite to what was expected. This might indicate that the entity of asbestos exposure within Broni and adjacent surroundings was equally intense, and therefore the pollution originating from Fibronit was not limited to the area around the factory. Consonni et al., in an epidemiological study of Broni, did not notice any clear difference in MM incidence in the 8 adjacent and 17 surrounding towns , suggesting that asbestos pollution is geographically widespread in that area. This is relevant considering that many factories, worldwide, still produce asbestos‐containing materials and several of them are located close to dwellings. A significant correlation was observed, instead, between distance and ABs concentration: the closest the address was to the factory, the higher was the concentration of ABs. This finding may reflect, again, the importance of asbestos clearance (especially regarding chrysotile), while ABs cannot be removed from lungs. Unfortunately, the concentration of ABs cannot be considered a reliable proxy of the uncoated asbestos concentration, nor of the amount of inhaled fibers, as the proportion between asbestos bodies and fibers is extremely variable across subjects, according to different individual ability to cover fibers. In our previous studies, in around 90% of subjects the concentration of uncoated fibers was higher compared to ABs (both counted by SEM‐EDS) . In this study, a significant, despite weak, correlation between the amount of asbestos and ABs was observed, but the ratio between asbestos and ABs was extremely variable, as reported in previous studies . Moreover, in some exposed subjects, asbestos concentration was under the detection limit, but we detected ABs in their lungs. Also, the contrary has been observed in some cases. As pointed out by other authors (e.g., ), the tendency to form ABs is linked to individual “coating capacity” . Therefore, despite the correlations observed in this study, ABs cannot be regarded as a reliable index of the actual amount of previous asbestos inhalation. Notwithstanding, ABs, as opposed to uncovered asbestos fibers, are not subjected to lung clearance, as they are too large and located in the interstitium. Therefore, their concentration is stable even after decades from the end of exposure. This study has two main limitations. First, several subjects with environmental exposure changed many addresses during their life, therefore, even though they always lived in the Broni area, we decided to consider the address in which they lived for the longest time among, sometimes, four or five addresses. The second main limitation is that the concentration of ABs was assessed using SEM and not light microscopy (as it is usually done). This means that a smaller lung sample for each patient was assessed for ABs compared to the standard procedure, namely 0.25 g versus 3–6 g . However, this could lead to an underestimation of ABs concentration, therefore it is unlikely that this limitation resulted in false correlations. Moreover, an excellent correlation between ABs measured by light microscope and SEM has been pointed out by Roggli et al. . Strengths of our study include the detailed address history, the large number of autopsies included, and the carefully collected personal history of asbestos exposure. Conclusions In this study, we investigated the relationship between the type of exposure (distinguishing between environmental and occupational) and the asbestos and ABs lung burden of 77 subjects. We analyzed, for the first time, the possible relationship between the distance of environmentally exposed residences from the source of exposure (asbestos‐cement factory) and the lung burden of asbestos and ABs. We found that asbestos lung content burden in MM patients who lived around an asbestos factory is as high as in occupationally exposed individuals; this holds true in residents up to 10 km radius from the factory. This study suggests that heavy asbestos pollution involves not only the area adjacent to the factory, but the entire town of Broni and the surroundings. This is alarming if we consider that most asbestos factories still active in some countries are located close to towns and dwellings. However, the assessment of asbestos lung burden fails to detect chrysotile, due to its lower biodurability and rapid clearance in the lung microenvironment. In this series, we detected chrysotile in only 5 out of 77 subjects and its concentration was well below the threshold considered indicative of past asbestos exposure. Clearance, to a lesser extent, occurs also for amphiboles. This can explain the detection of a much lower concentration of amphibole asbestos than expected according to the history of exposure, and the correlations between the concentration of ABs and the type of exposure, as well as the distance from the asbestos‐cement factory, not found for asbestos. In some cases where death occurred long after the end of exposure, ABs might be the only detectable sign of previous asbestos exposure. In conclusion, we found a similar asbestos burden in the lungs of mesothelioma patients in the town of Broni who were occupationally or environmentally exposed, and more asbestos bodies in the former and in environmentally exposed who lived closer to the plant. These findings suggest that physiologic clearance of asbestos fibers may influence the results of lung content analysis, especially if death occurred a long time after the end of exposure. Silvia Damiana Visonà and Emanuela Taioli designed the work, interpreted the data, and wrote the manuscript. Matthew Untalan and Tara Ivic‐Pavlicic performed the analysis of data for the work. Barbara Bertoglio and Silvana Capella acquired and interpreted the data. Elena Belluso supervised the data acquisition and critically revised the manuscript for important intellectual content. Marco Billò analyzed part of the data. All the authors approved the final version to be published. Silvia Damiana Visonà and Emanuela Taioli agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. John Meyer declares that he has no conflict of interest in the review and publication decision regarding this article. The study protocol was approved by Ethical Committee “Lombardia 6” on August 22, 2023. Informed consent was waived because the participants are deceased subjects and relatives are no longer reachable. The authors declare no conflicts of interest. Supporting information. |
Overcoming Ploidy Barriers: The Role of Triploid Bridges in the Genetic Introgression of | 4cf6840a-fdfc-4aee-956c-2a842222d52b | 11934083 | Pathologic Processes[mh] | Introduction Understanding the mechanisms underpinning speciation has been a primary focus of evolutionary biology. The emergence of the biological species concept (Mayr ) underscored reproductive isolation as the ultimate stage of speciation, where intrinsic or extrinsic barriers play pivotal roles in driving the process. Yet, the contribution of intrinsic barriers driven by genetic incompatibilities is still not entirely resolved (Coughlan and Matute ). One of the biological processes acting as an intrinsic barrier to reproduction is whole genome duplication (WGD), which increases the number of sets of chromosomes and is estimated to constitute 15% of speciation events in plants (Wood et al. ). Unlike other intrinsic barriers, WGD causes a drastic, immediate change in the genetic composition of an individual within a single generation. The doubling of chromosome sets leads to critical genetic incompatibilities, resulting in developmental and chromosomal imbalances in hybrid progeny resulting from a cross between a newly formed polyploid and its lower‐ploidy progenitor. These imbalances are characterised by decreased viability (known as the triploid block) and/or reduced fertility of interploidy hybrids, makingWGD a classic example of non‐ecological, instant speciation (Otto and Whitton ; Czekanski‐Moir and Rundell ). However, reproductive barriers arising following WGD may still be permeable, as suggested by both theory and experiments, particularly in cases where intermediate cytotypes are viable and could be facilitating backcrossing with plants of parental cytotypes (referred to as the ‘triploid bridge’ pathway—Ramsey and Schemske ; Husband ). In contrast to unidirectional gene flow towards higher ploidy, which occurs when unreduced gametes from a lower ploidy result in hybrids with the ploidy of a higher cytotype (documented, e.g. in Capsella by Kryvokhyzha et al. and Betula by Zohren et al. and Leal et al. ), a triploid bridge may enable bidirectional gene flow if fertile interploidy hybrids are formed and are capable of backcrossing with both parental ploidies (Kolář et al. ; Bartolić et al. ). Indeed, analyses of genetic structure demonstrated interploidy admixture in some allopolyploid (Pinheiro et al. ; Thórsson ) and autotetraploid species (Laport et al. ; Šingliarová et al. ), but whether this admixture reflects bidirectional gene flow remains untested. Likely ancestral bidirectional gene flow has been documented in Arabidopsis arenosa (Arnold et al. ); however, the virtual lack of natural triploids in mixed‐ploidy populations impedes the further interpretation of its mechanistic basis and natural significance (Morgan et al. , ). Studies investigating interploidy gene flow in nature are often hampered by a lack of genomic tools or resources and by the absence of integration between genomic approaches and experimental tests of pre‐ and postzygotic barriers. As a consequence, we still lack a satisfactorily described case of bidirectional interploidy gene flow mediated by a triploid bridge mechanism in natural plant populations (reviewed by Bartolić et al. ). For a more holistic view of introgressive gene flow across a ploidy barrier in a natural environment, we require genetically well‐tractable systems with a well‐defined cytogeographic distribution and sympatric populations encompassing viable intermediate cytotypes. One such system is Cardamine amara (Brassicaceae), a widely distributed perennial herb inhabiting most of Europe's wetlands and streams at low to high elevations. In contrast to the diploid cytotype ( C. amara subsp. amara ), which is distributed throughout most of Europe, the tetraploid cytotype ( C. amara subsp. austriaca , Marhold ) inhabits the Eastern Alps and neighbouring areas. An overall genetic similarity between diploid and tetraploid cytotypes, together with only slight morphological differentiation and no apparent candidates for interspecific hybridisation, indicates an autopolyploid origin of the tetraploid, likely during the Pleistocene (Lihová et al. ; Marhold et al. ; Bohutínská et al. ). Both cytotypes meet in the northern foothills of the Alps (reaching as far as central Czechia), where they form a distinct secondary contact zone with populations including both diploid and tetraploid individuals as well as several viable individuals of intermediate triploid ploidy growing in proximity (Zozomová‐Lihová et al. ). A well‐delimited contact zone, suitable genomic resources and the co‐occurrence of both major ploidies together with triploids make C. amara an ideal model for addressing the role of the triploid bridge. However, neither the strength of reproductive barriers between diploids and tetraploids nor the intensity of interploidy gene flow and the role of triploids therein has been resolved so far. In this study, we took a multidisciplinary approach involving extensive sampling and cytotyping of plants from mixed‐ploidy populations by flow cytometry, reciprocal crossing experiments and population genomics to seek integrative evidence for the action of the triploid bridge in natural mixed‐ploidy populations of C. amara . We address the following specific questions: (1) How strong is the triploid block in C. amara ? Are the natural triploids a product of hybridisation between the major cytotypes? (2) Are triploids capable of backcrossing with diploids and tetraploids, and if so, do such backcrosses result in a significant fraction of euploid progeny, demonstrating their potential to act as mediators of gene flow? (3) Is there, correspondingly, genomic evidence for bi‐directional interploidy introgression between naturally co‐existing diploids and tetraploids of C. amara ?
Materials and Methods 2.1 Field Sampling Six populations in different parts of Central Europe were sampled. Two of them were mixed‐ploidy populations from the contact zone, and four were populations from cytotype‐pure areas (Figure , Table ). The populations were chosen to include both sympatric populations, where all cytotypes are intermixed, and those deep in cytotype‐pure areas away from the contact zone, based on previous thorough flow cytometric screening (Zozomová‐Lihová et al. ). From the cytotype‐pure populations, 10 individuals were transferred to the greenhouse of Charles University (Prague, CZ) for subsequent crossing experiments and genetic analyses. In the two mixed‐ploidy populations (LIP, HLI), multiple geo‐referenced individuals were sampled to determine the fine‐scale distribution of cytotypes (Figure ). Following this, 10 living individuals of the diploid and 10 of the tetraploid cytotype and three triploid individuals were transferred from each population for genetic analyses. A subset of these individuals was used in Crossing Experiment 1 (detailed in the Crossing experiments section). Additionally, diploids (2x), triploids (3x) and tetraploids (4x) were sampled later in the LIP population for Crossing Experiment 2, focused on triploid backcrossing. Within each population, individuals were collected with a sampling distance of 3–5 m to minimise the likelihood of collecting individuals resulting from vegetative reproduction. In addition, both populations were revisited during the fruiting season, and seeds were collected from 64 and 96 wild plants of population LIP and population HLI, respectively. The ploidy of these mother plants was unknown at the time of collection and was only determined afterwards, which led to a significant imbalance in the representation of diploid and tetraploid plants in our sample. After confirmation of the ploidy of each mother plant, the collected seeds were germinated, and the ploidy of the seedlings was also determined. 2.2 Ploidy Level Estimation and Chromosome Counting The ploidy level of each individual was estimated separately by measuring relative genome size using flow cytometry with 4,6‐diamino‐2‐phenylindole (DAPI) staining. A two‐step protocol, based on Zozomová‐Lihová et al. ( ), was followed with Solanum pseudocapsicum (2C = 2.59 pg.; Temsch et al. ) as an internal standard. The prepared nuclei solution was analysed to obtain measurements for a minimum of 3000 particles using one of two machines, CyFlow ML (Partec) or CytoFlex S (Beckman Coulter), in the Flow Cytometry Laboratory at Charles University in Prague. Relative genome size (RGS), defined as the ratio of the peak fluorescence intensity of the sample to that of the internal standard, was inferred from the resulting histograms using FloMax FCS 2.0 or CytExpert 2.4 software for analyses run on the CyFlow or CytoFlex machines, respectively. Correspondence between the two machines was checked using 20 samples analysed on both instruments. Individuals were categorised as euploid diploids, triploids or tetraploids based on whether their RGS values fell within the range of three standard deviations above or below the respective means of the established control groups (2x: 0.155–0.191, 3x: 0.262–0.272, 4x: 0.349–0.369). As control groups, we used diploid and tetraploid progeny from the control homoploid crosses from Crossing Experiments 1 and 2. Any values outside these cut‐offs were deemed potential aneuploids. To validate putative aneuploids among individuals with deviating RGS, mitotic chromosome spreads of C. amara accessions D7 (2 n = 16), B6 (2 n = 32), F8 (2 n = 31), M8 (2 n = 31) and L1 (2 n = 38) were prepared from root tips as described by Mandáková et al. ( ). Chromosome preparations were treated with 100 μg/mL of RNase in 2× sodium saline citrate (SSC; 20× SSC: 3 M sodium chloride, 300 mM trisodium citrate, pH 7.0) for 60 min and with 0.1 mg/mL of pepsin in 0.01 M HCl at 37°C for 5 min; then postfixed in 4% formaldehyde in 2× SSC for 10 min, washed in 2× SSC twice for 5 min and dehydrated in an ethanol series (70%, 90% and 100%, 2 min each). Chromosomes were counterstained with 4′,6‐diamidino‐2‐phenylindole (DAPI, 2 μg/mL) in Vectashield antifade mounting medium. Fluorescence signals were analysed and photographed using a Zeiss Axioimager epifluorescence microscope and a CoolCube camera (MetaSystems). 2.3 Crossing Experiments Two experiments were performed. In Crossing Experiment 1, diploid plants were crossed with tetraploids to investigate the strength of postzygotic barriers (triploid block; 50 individuals included—25 diploid and 25 tetraploid Table ). In Crossing Experiment 2, triploid individuals were crossed with either diploid or tetraploid plants to investigate the potential of natural triploids for backcrossing (triploid bridge; 54 individuals included—17 diploids, 17 tetraploids and 20 triploids, Table ). The plants were cultivated in their native substrate mixed with standard garden soil substrate of neutral pH (Agro CS), separately in plastic containers of 10 cm diameter. The cultivation was performed in the greenhouse of the Faculty of Science at Charles University in Prague (50.0715 N, 14.4230 E) under natural light in the period from September to May 2022 and with regular watering every 3 days, keeping permanent access of the cultivated plants to water. When flower buds appeared, flowers designated as pollen acceptors were emasculated to avoid self‐pollination. Each emasculated plant was covered with a plastic mesh bag to prevent any unwanted pollination by other plants. After the stigma of an emasculated flower became receptive, we used a mature anther of the designated father plant and rubbed it on the stigma until its surface was fully covered with pollen and then enclosed the pollinated flower in a pollination bag. After pollination, successfully formed siliques were counted and enclosed in paper bags to collect the seeds. In Crossing Experiment 1, we conducted 19 crosses with tetraploids as pollen recipients (4x × 2x; hereinafter, the mother plant is always given first) and 16 with tetraploids as pollen donors (2x × 4x). Additionally, we performed six diploid and 10 tetraploid homoploid control crosses (2x × 2x, 4x × 4x). In Crossing Experiment 2, triploids were crossed with diploids and tetraploids in all combinations (2x × 3x [four crosses], 3x × 2x [10 crosses], 3x × 4x [nine crosses], 4x × 3x [five crosses], Table ) and supplemented them with six diploid and 9 tetraploid control crosses. We aimed to pollinate at least five flowers per cross in both experiments. Seeds obtained from both experiments were counted, weighed using a high‐precision analytical scale and germinated in a universal garden substrate under a 21/18°C day/night regimen with 16 h of light per day. The RGS of each germinated seedling resulting from an interploidy cross and a subset of 40 (20 diploid and 20 tetraploid) seedlings from control crosses was measured by flow cytometry as described above. For the statistical evaluation of each experiment, two models were established and compared. Initially, a null model assumed a constant mean value for the dependent variable (number of seeds per silique, average mass per seed, proportion of germinated seedlings) across all individuals. Subsequently, a more complex model allowed the mean value to vary among different cross‐types. The significance of the cross‐type effect was assessed using likelihood ratio tests comparing the models. Linear models, calculated using the R package lme4 (Bates et al. ), were used to determine relative differences in seed set and seed mass, and Tukey post hoc tests were then employed to compare individual treatments. Differences in germination rates were calculated using generalised linear models with a binomial distribution of residual variation in the R package lme4 (Bates et al. ). The proportion of germinated seedlings per seed parent was used as a dependent variable, and the type of cross was used as a predictor. 2.4 DNA Extraction, Library Preparation, Sequencing and Raw Data Processing and Filtration Eighty‐four individuals from six populations were selected for population genomic studies: 39 from cytotype‐pure populations and 45 individuals from mixed‐ploidy populations (for details see Table ). Genomic DNA was extracted from silica‐gel‐dried leaves using the sorbitol extraction method (Štorchová et al. ) and then purified using AMPure XP (Beckman Coulter Inc., Brea, California, USA). Samples were genotyped for genome‐wide single nucleotide polymorphisms (SNPs) using the whole genome sequencing protocol LITE of Perez‐Sepulveda et al. . The libraries were sequenced in 300 cycles (2 × 150 bp paired‐end/PE reads) on the Illumina NovaSeq platform. Illumina sequencing reads were demultiplexed using the fastx_barcode_splitter.pl. script from the FASTX‐Toolkit v. 0.0.14. Read ends and reads where the average quality within the 5‐bp window fell below Q20 were trimmed, and reads of less than 50 bp were discarded by Trimmomatic v. 0.36 (Bolger et al. ). The resulting reads were compressed into clumps, and duplicates were removed using the script clumpify.sh (BBTools; https://jgi.doe.gov/data‐and‐tools/bbtools ). The sequencing reads were mapped onto a reference genome of C. amara (Bohutínská et al. ) using BWA 0.7.3a (Li ) and the resulting BAM files were processed with the Picard Toolkit v. 2.22.1. Variant calling was performed for each individual using the HaplotypeCaller module from the Genome Analysis Toolkit v. 3.7‐0 (GATK; McKenna et al. ), specifying the ploidy of each individual. Next, variants were aggregated, and genotyping across all individuals was performed using GATK's GenotypeGVCFs module, which is suitable for joint genotyping of mixed‐ploidy datasets (see, e.g. Monnahan et al. ). Variant filtration was performed by the VariantFiltration module, requiring a minimum sequencing depth of 8x and applying the filter parameters indicated by GATK's best practices (Van der Auwera et al. ). Finally, the SelectVariants module was used to capture putatively neutral 4‐fold degenerated biallelic sites that passed filter parameters, with no more than 20% of missing genotypes, but we excluded genes that showed excess heterozygosity or read depth (potential paralogues mapped on top of each other). Four‐fold degenerate SNPs were identified by the Identify_4D_Sites.pl. script (available at https://github.com/tsackton/linked‐selection ). Genes with excess heterozygosity (fixed heterozygous in at least 3% of SNPs in one or more diploid populations) and sites with read depth exceeding the mean plus double the standard deviation with at least 10% of samples were identified following Šlenker ( ). 2.5 Population Genomic Analyses Initially, genetic clusters were inferred using the Bayesian clustering algorithm implemented in STRUCTURE v. 2.3.4 (Pritchard et al. ), giving unbiased results with mixed‐ploidy populations (Stift et al. ). STRUCTURE analyses required unlinked SNPs, which were obtained by randomly selecting a single SNP from 20,000‐bp windows using the vcf_prune.py script (Šlenker ). To capture the overall data variability, 100 datasets with randomly selected SNPs were analysed. Each dataset was analysed for each K = 1–7, with a burn‐in length of 100,000 generations and data collection for an additional 900,000 generations, setting the admixture model and correlated allele frequencies. The results for 100 datasets were averaged using the programme CLUMPP (Jakobsson and Rosenberg ) and drawn with the DISTRUCT routine (Rosenberg ). The approach of Evanno et al. ( ) was adopted to determine the optimal K value. Secondly, we displayed genetic distances among individuals using principal component analysis (PCA) based on Euclidean distance as implemented in adegenet v1.4‐2 (Jombart ). Thirdly, we calculated Nei's (Nei ) distances among all individuals using the StAMPP package (Pembleton et al. ), developed specifically for analysing SNP datasets on mixed‐ploidy scenarios and displayed them using the neighbour network algorithm in SplitsTree (Huson and Bryant ). 2.6 ABBA ‐ BABA Test and Demographic Modelling To quantify the extent of recent introgression in the sympatric populations, we conducted an ABBA‐BABA test, which relies on Patterson's D statistic to estimate the genome‐wide excess of shared derived alleles between two taxa (Green et al. ; Martin et al. ). This test assumes that both the ABBA and BABA topologies occur at the same frequency in accordance with the incomplete lineage sorting (ILS) hypothesis. However, if there is an introgression between two taxa of a bifurcating tree, one topology occurs with much greater frequency than the other. One allopatric tetraploid population far from the contact zone (STE) was used as the non‐introgressed sister lineage (P1) in all cases. For sympatric combinations, tetraploid and diploid individuals from each mixed population (LIP, HLI) were used as P2 and P3, respectively. For allopatric combinations, each tetraploid population occurring outside the contact zone (SUM, JES) was used as P2, and the only allopatric diploid population (REC) was used as P3. One population of the closely related but geographically distant taxon C. amara subsp. balcanica was used as an outgroup. To calculate the D statistic, we used scripts written by Simon Martin available at https://github.com/simonhmartin/tutorials/tree/master/ABBA_BABA_whole_genome . To further test for the presence of gene flow between diploids and tetraploids and to quantify its potential asymmetry, we used fastsimcoal v. 2.709 (Excoffier et al. ). For a pair of diploid and tetraploid (sub‐)populations from each mixed‐ploidy population (LIP and HLI) separately, we constructed folded two‐dimensional site frequency spectra (SFS) from the variant and invariant four‐fold degenerate sites (filtered in the same ways as above) using python scripts FSC2input.py available at https://github.com/pmonnahan/ScanTools/ (Monnahan et al. ). For each population pair, we compared the following four scenarios (Figure ): (1) no gene flow or migration, (2) unidirectional gene flow from diploids to tetraploids, (3) unidirectional gene flow from tetraploids to diploids and (4) equal bidirectional gene flow between diploids and tetraploids. For each scenario and population pair, 50 fastsimcoal runs were performed. For each run, 40 ECM optimisation cycles were allowed to estimate the parameters and 100,000 simulations were conducted at each step to estimate the expected SFS. Further, the partition with the highest likelihood for each fastsimcoal run was identified and values of the Akaike information criterion (AIC) for these partitions were calculated and summarised across the 50 fastsimcoal runs. The scenario with the lowest median AIC value within each population was considered the most favourable.
Field Sampling Six populations in different parts of Central Europe were sampled. Two of them were mixed‐ploidy populations from the contact zone, and four were populations from cytotype‐pure areas (Figure , Table ). The populations were chosen to include both sympatric populations, where all cytotypes are intermixed, and those deep in cytotype‐pure areas away from the contact zone, based on previous thorough flow cytometric screening (Zozomová‐Lihová et al. ). From the cytotype‐pure populations, 10 individuals were transferred to the greenhouse of Charles University (Prague, CZ) for subsequent crossing experiments and genetic analyses. In the two mixed‐ploidy populations (LIP, HLI), multiple geo‐referenced individuals were sampled to determine the fine‐scale distribution of cytotypes (Figure ). Following this, 10 living individuals of the diploid and 10 of the tetraploid cytotype and three triploid individuals were transferred from each population for genetic analyses. A subset of these individuals was used in Crossing Experiment 1 (detailed in the Crossing experiments section). Additionally, diploids (2x), triploids (3x) and tetraploids (4x) were sampled later in the LIP population for Crossing Experiment 2, focused on triploid backcrossing. Within each population, individuals were collected with a sampling distance of 3–5 m to minimise the likelihood of collecting individuals resulting from vegetative reproduction. In addition, both populations were revisited during the fruiting season, and seeds were collected from 64 and 96 wild plants of population LIP and population HLI, respectively. The ploidy of these mother plants was unknown at the time of collection and was only determined afterwards, which led to a significant imbalance in the representation of diploid and tetraploid plants in our sample. After confirmation of the ploidy of each mother plant, the collected seeds were germinated, and the ploidy of the seedlings was also determined.
Ploidy Level Estimation and Chromosome Counting The ploidy level of each individual was estimated separately by measuring relative genome size using flow cytometry with 4,6‐diamino‐2‐phenylindole (DAPI) staining. A two‐step protocol, based on Zozomová‐Lihová et al. ( ), was followed with Solanum pseudocapsicum (2C = 2.59 pg.; Temsch et al. ) as an internal standard. The prepared nuclei solution was analysed to obtain measurements for a minimum of 3000 particles using one of two machines, CyFlow ML (Partec) or CytoFlex S (Beckman Coulter), in the Flow Cytometry Laboratory at Charles University in Prague. Relative genome size (RGS), defined as the ratio of the peak fluorescence intensity of the sample to that of the internal standard, was inferred from the resulting histograms using FloMax FCS 2.0 or CytExpert 2.4 software for analyses run on the CyFlow or CytoFlex machines, respectively. Correspondence between the two machines was checked using 20 samples analysed on both instruments. Individuals were categorised as euploid diploids, triploids or tetraploids based on whether their RGS values fell within the range of three standard deviations above or below the respective means of the established control groups (2x: 0.155–0.191, 3x: 0.262–0.272, 4x: 0.349–0.369). As control groups, we used diploid and tetraploid progeny from the control homoploid crosses from Crossing Experiments 1 and 2. Any values outside these cut‐offs were deemed potential aneuploids. To validate putative aneuploids among individuals with deviating RGS, mitotic chromosome spreads of C. amara accessions D7 (2 n = 16), B6 (2 n = 32), F8 (2 n = 31), M8 (2 n = 31) and L1 (2 n = 38) were prepared from root tips as described by Mandáková et al. ( ). Chromosome preparations were treated with 100 μg/mL of RNase in 2× sodium saline citrate (SSC; 20× SSC: 3 M sodium chloride, 300 mM trisodium citrate, pH 7.0) for 60 min and with 0.1 mg/mL of pepsin in 0.01 M HCl at 37°C for 5 min; then postfixed in 4% formaldehyde in 2× SSC for 10 min, washed in 2× SSC twice for 5 min and dehydrated in an ethanol series (70%, 90% and 100%, 2 min each). Chromosomes were counterstained with 4′,6‐diamidino‐2‐phenylindole (DAPI, 2 μg/mL) in Vectashield antifade mounting medium. Fluorescence signals were analysed and photographed using a Zeiss Axioimager epifluorescence microscope and a CoolCube camera (MetaSystems).
Crossing Experiments Two experiments were performed. In Crossing Experiment 1, diploid plants were crossed with tetraploids to investigate the strength of postzygotic barriers (triploid block; 50 individuals included—25 diploid and 25 tetraploid Table ). In Crossing Experiment 2, triploid individuals were crossed with either diploid or tetraploid plants to investigate the potential of natural triploids for backcrossing (triploid bridge; 54 individuals included—17 diploids, 17 tetraploids and 20 triploids, Table ). The plants were cultivated in their native substrate mixed with standard garden soil substrate of neutral pH (Agro CS), separately in plastic containers of 10 cm diameter. The cultivation was performed in the greenhouse of the Faculty of Science at Charles University in Prague (50.0715 N, 14.4230 E) under natural light in the period from September to May 2022 and with regular watering every 3 days, keeping permanent access of the cultivated plants to water. When flower buds appeared, flowers designated as pollen acceptors were emasculated to avoid self‐pollination. Each emasculated plant was covered with a plastic mesh bag to prevent any unwanted pollination by other plants. After the stigma of an emasculated flower became receptive, we used a mature anther of the designated father plant and rubbed it on the stigma until its surface was fully covered with pollen and then enclosed the pollinated flower in a pollination bag. After pollination, successfully formed siliques were counted and enclosed in paper bags to collect the seeds. In Crossing Experiment 1, we conducted 19 crosses with tetraploids as pollen recipients (4x × 2x; hereinafter, the mother plant is always given first) and 16 with tetraploids as pollen donors (2x × 4x). Additionally, we performed six diploid and 10 tetraploid homoploid control crosses (2x × 2x, 4x × 4x). In Crossing Experiment 2, triploids were crossed with diploids and tetraploids in all combinations (2x × 3x [four crosses], 3x × 2x [10 crosses], 3x × 4x [nine crosses], 4x × 3x [five crosses], Table ) and supplemented them with six diploid and 9 tetraploid control crosses. We aimed to pollinate at least five flowers per cross in both experiments. Seeds obtained from both experiments were counted, weighed using a high‐precision analytical scale and germinated in a universal garden substrate under a 21/18°C day/night regimen with 16 h of light per day. The RGS of each germinated seedling resulting from an interploidy cross and a subset of 40 (20 diploid and 20 tetraploid) seedlings from control crosses was measured by flow cytometry as described above. For the statistical evaluation of each experiment, two models were established and compared. Initially, a null model assumed a constant mean value for the dependent variable (number of seeds per silique, average mass per seed, proportion of germinated seedlings) across all individuals. Subsequently, a more complex model allowed the mean value to vary among different cross‐types. The significance of the cross‐type effect was assessed using likelihood ratio tests comparing the models. Linear models, calculated using the R package lme4 (Bates et al. ), were used to determine relative differences in seed set and seed mass, and Tukey post hoc tests were then employed to compare individual treatments. Differences in germination rates were calculated using generalised linear models with a binomial distribution of residual variation in the R package lme4 (Bates et al. ). The proportion of germinated seedlings per seed parent was used as a dependent variable, and the type of cross was used as a predictor.
DNA Extraction, Library Preparation, Sequencing and Raw Data Processing and Filtration Eighty‐four individuals from six populations were selected for population genomic studies: 39 from cytotype‐pure populations and 45 individuals from mixed‐ploidy populations (for details see Table ). Genomic DNA was extracted from silica‐gel‐dried leaves using the sorbitol extraction method (Štorchová et al. ) and then purified using AMPure XP (Beckman Coulter Inc., Brea, California, USA). Samples were genotyped for genome‐wide single nucleotide polymorphisms (SNPs) using the whole genome sequencing protocol LITE of Perez‐Sepulveda et al. . The libraries were sequenced in 300 cycles (2 × 150 bp paired‐end/PE reads) on the Illumina NovaSeq platform. Illumina sequencing reads were demultiplexed using the fastx_barcode_splitter.pl. script from the FASTX‐Toolkit v. 0.0.14. Read ends and reads where the average quality within the 5‐bp window fell below Q20 were trimmed, and reads of less than 50 bp were discarded by Trimmomatic v. 0.36 (Bolger et al. ). The resulting reads were compressed into clumps, and duplicates were removed using the script clumpify.sh (BBTools; https://jgi.doe.gov/data‐and‐tools/bbtools ). The sequencing reads were mapped onto a reference genome of C. amara (Bohutínská et al. ) using BWA 0.7.3a (Li ) and the resulting BAM files were processed with the Picard Toolkit v. 2.22.1. Variant calling was performed for each individual using the HaplotypeCaller module from the Genome Analysis Toolkit v. 3.7‐0 (GATK; McKenna et al. ), specifying the ploidy of each individual. Next, variants were aggregated, and genotyping across all individuals was performed using GATK's GenotypeGVCFs module, which is suitable for joint genotyping of mixed‐ploidy datasets (see, e.g. Monnahan et al. ). Variant filtration was performed by the VariantFiltration module, requiring a minimum sequencing depth of 8x and applying the filter parameters indicated by GATK's best practices (Van der Auwera et al. ). Finally, the SelectVariants module was used to capture putatively neutral 4‐fold degenerated biallelic sites that passed filter parameters, with no more than 20% of missing genotypes, but we excluded genes that showed excess heterozygosity or read depth (potential paralogues mapped on top of each other). Four‐fold degenerate SNPs were identified by the Identify_4D_Sites.pl. script (available at https://github.com/tsackton/linked‐selection ). Genes with excess heterozygosity (fixed heterozygous in at least 3% of SNPs in one or more diploid populations) and sites with read depth exceeding the mean plus double the standard deviation with at least 10% of samples were identified following Šlenker ( ).
Population Genomic Analyses Initially, genetic clusters were inferred using the Bayesian clustering algorithm implemented in STRUCTURE v. 2.3.4 (Pritchard et al. ), giving unbiased results with mixed‐ploidy populations (Stift et al. ). STRUCTURE analyses required unlinked SNPs, which were obtained by randomly selecting a single SNP from 20,000‐bp windows using the vcf_prune.py script (Šlenker ). To capture the overall data variability, 100 datasets with randomly selected SNPs were analysed. Each dataset was analysed for each K = 1–7, with a burn‐in length of 100,000 generations and data collection for an additional 900,000 generations, setting the admixture model and correlated allele frequencies. The results for 100 datasets were averaged using the programme CLUMPP (Jakobsson and Rosenberg ) and drawn with the DISTRUCT routine (Rosenberg ). The approach of Evanno et al. ( ) was adopted to determine the optimal K value. Secondly, we displayed genetic distances among individuals using principal component analysis (PCA) based on Euclidean distance as implemented in adegenet v1.4‐2 (Jombart ). Thirdly, we calculated Nei's (Nei ) distances among all individuals using the StAMPP package (Pembleton et al. ), developed specifically for analysing SNP datasets on mixed‐ploidy scenarios and displayed them using the neighbour network algorithm in SplitsTree (Huson and Bryant ).
ABBA ‐ BABA Test and Demographic Modelling To quantify the extent of recent introgression in the sympatric populations, we conducted an ABBA‐BABA test, which relies on Patterson's D statistic to estimate the genome‐wide excess of shared derived alleles between two taxa (Green et al. ; Martin et al. ). This test assumes that both the ABBA and BABA topologies occur at the same frequency in accordance with the incomplete lineage sorting (ILS) hypothesis. However, if there is an introgression between two taxa of a bifurcating tree, one topology occurs with much greater frequency than the other. One allopatric tetraploid population far from the contact zone (STE) was used as the non‐introgressed sister lineage (P1) in all cases. For sympatric combinations, tetraploid and diploid individuals from each mixed population (LIP, HLI) were used as P2 and P3, respectively. For allopatric combinations, each tetraploid population occurring outside the contact zone (SUM, JES) was used as P2, and the only allopatric diploid population (REC) was used as P3. One population of the closely related but geographically distant taxon C. amara subsp. balcanica was used as an outgroup. To calculate the D statistic, we used scripts written by Simon Martin available at https://github.com/simonhmartin/tutorials/tree/master/ABBA_BABA_whole_genome . To further test for the presence of gene flow between diploids and tetraploids and to quantify its potential asymmetry, we used fastsimcoal v. 2.709 (Excoffier et al. ). For a pair of diploid and tetraploid (sub‐)populations from each mixed‐ploidy population (LIP and HLI) separately, we constructed folded two‐dimensional site frequency spectra (SFS) from the variant and invariant four‐fold degenerate sites (filtered in the same ways as above) using python scripts FSC2input.py available at https://github.com/pmonnahan/ScanTools/ (Monnahan et al. ). For each population pair, we compared the following four scenarios (Figure ): (1) no gene flow or migration, (2) unidirectional gene flow from diploids to tetraploids, (3) unidirectional gene flow from tetraploids to diploids and (4) equal bidirectional gene flow between diploids and tetraploids. For each scenario and population pair, 50 fastsimcoal runs were performed. For each run, 40 ECM optimisation cycles were allowed to estimate the parameters and 100,000 simulations were conducted at each step to estimate the expected SFS. Further, the partition with the highest likelihood for each fastsimcoal run was identified and values of the Akaike information criterion (AIC) for these partitions were calculated and summarised across the 50 fastsimcoal runs. The scenario with the lowest median AIC value within each population was considered the most favourable.
Results 3.1 Cytotype Structure in the Contact Zone and Within Mixed‐Ploidy Populations Three main ploidy levels with RGS values corresponding to diploids, triploids and tetraploids were detected among the 356 field‐collected adult individuals (Table ). In addition, three putative aneuploids with RGS values in between triploid and diploid (two plants) and between triploid and tetraploid (one plant) were sampled in the mixed‐ploidy populations LIP and HLI, respectively (Figure ; Table ). The fine‐scale distribution of cytotypes in the two mixed‐ploidy populations revealed the presence of both ploidy‐uniform patches and parts where all three cytotypes grow only a few metres apart (Figure ). The seeds collected in natural mixed‐ploidy populations had variable but generally meagre germination rates (0%–60%, 8% on average, Figure ). Among 122 germinated seedlings, the RGS values corresponded to both euploid (84%) and aneuploid (16%) values (Figure ). Seedlings of both diploid and triploid mothers exhibited values corresponding to the diploid state (79% and 21% of seedlings from diploid and triploid mothers, respectively). The rest were putative aneuploids with RGS values between diploids and tetraploids, with the exception of one putative tetraploid found among the progeny of a diploid parent. The RGS of seedlings from tetraploid mothers ranged from values corresponding to hypo‐tetraploid (15%) and tetraploid (62%) up to hyper‐tetraploid (23%) aneuploid (Figure ). No seedling with RGS corresponding to the triploid state was found. 3.2 Strength of the Interploidy Barrier Inferred From Crossing Experiments We first crossed diploid and tetraploid plants to investigate the strength of postzygotic barriers (triploid block, Crossing Experiment 1). The difference in the seed set was significant overall ( F ₃,₄₇ = 3.71, p = 0.018, Figure ), yet rather small, as it was only the category of homoploid control 4x × 4x crosses that were significantly higher than one of the interploidy crosses (2x × 4x; Figure ). The seed mass was also different ( F ₃,₄₇ = 98.62, p < 0.001), being similar between the interploidy crosses but markedly lower than that of homoploid crosses of both types (Figure ). Consequently, there was also significant variation in germination rates among the crossing treatments (Figure , χ 2 = −339.2, df = 3, p < 0.001). The progeny of homoploid controls had higher germination percentages (40% and 55% for diploid and tetraploid crosses, respectively) than interploidy crosses, in which no viable seeds were produced by diploid seed parents and only a single germinable seed (< 0.5%) was formed after pollination of a tetraploid by a diploid pollen donor. The RGS value of this plant corresponded to a triploid (Figure ). Then, triploid plants were crossed with diploids and tetraploids to investigate the potential of natural triploids to act as introgression mediators (i.e. as part of the triploid bridge pathway, Crossing Experiment 2). There was a significant difference in seed sets between successful crosses ( F ₅,₅₈ = 5.06, p < 0.001), with larger seed sets in tetraploid control crosses than in the majority of interploidy crosses (Figure ). There was no significant difference between the diploid control and the interploidy crosses, with average seed set values of interploidy crosses ranging from 0.3 to 5.5 seeds per silique. There was no statistically significant difference in seed mass among the different types of crosses ( F ₅,₅₈ = 0.94, p = 0.4608; Figure ). The germination rates, again, differed significantly between the different types of crosses ( χ 2 = 97.018, df = 5, p < 0.001, Figure ), although the difference was not as pronounced as in the previous experiment. Tetraploid controls generally had significantly higher germination rates than the other types of crosses, with the exception of one type of backcross that was similar (4x × 3x). The germination rates of triploid backcrosses were highly variable but always non‐zero (8%–75% across treatments; Figure ). The relative genome size of the plants obtained from successful triploid backcrosses (36 cytotyped plants) corresponded to both euploid and aneuploid values defined based on the relative deviation of RGS from the control euploid values. The ploidy of the crossing partner of a triploid (diploid or tetraploid) significantly affected the proportion of putatively euploid progeny ( χ 2 = 21.78, df = 1, p < 0.001). Crosses of triploids with diploids resulted in a higher proportion of putatively euploid progeny (41%, with RGS values corresponding to diploids) than crosses involving tetraploids, where the proportion of euploid (tetraploid) progeny was 25% (Figure ). Chromosome counts obtained from three individuals classified by RGS as aneuploids confirmed the aneuploid number of chromosomes: F_8 (3x × 4x) – hypotetraploid, 2 n = 31; L_1 (3x × 4x)—hypopentaploid, 2 n = 38; M_8 (3x × 2x)—hypotetraploid, 2 n = 31 (Figure ). 3.3 Genetic Structure Based on Genome‐Wide SNPs Sequencing of 84 individuals from six populations (Figure ) produced between 24,453,671 and 84,627,579 reads per sample, averaging 38,264,942.5 reads after quality control and deduplication (Table ). Of these, 37%–94.2% were successfully mapped onto the reference genome, with an average mapping rate of 83.55%. The final VCF file contained 1,448,166 filtered putatively neutral four‐fold degenerate SNPs with an average depth of coverage of > 30× that were used in subsequent analyses. Bayesian clustering in STRUCTURE, based on a subset of 7169 LD‐pruned SNPs, suggested K = 2, 3 and 4 as stable partitions (strong similarity across runs), with K = 3 exhibiting the highest relative likelihood difference (delta K ). Diploids and tetraploids from cytotype‐pure populations separated already under K = 2 (Figure ). Additional separation of tetraploid individuals from the mixed‐ploidy population LIP was observed under K = 3 (Figure ) and K = 4 (population JES, Figure ). Tetraploids from the two mixed‐ploidy populations, HLI and LIP, however, showed a high proportion of diploid cluster ancestry under K = 2. The separation of pure diploid and tetraploid populations, in contrast to the closer position of both major cytotypes from mixed‐ploidy populations, was also supported by the neighbour‐joining network (Figure ). Interestingly, triploids from both mixed‐ploidy populations showed contrasting assignment patterns. Triploids from population HLI clustered exclusively with their sympatric diploids in both STRUCTURE and neighbour‐joining network analyses, whereas triploids from population LIP were a mixture of both STRUCTURE clusters and occupied intermediate positions in the network (Figure ). The contrasting genetic make‐up of triploids was further corroborated by principal component analyses run separately for each mixed‐ploidy population where HLI triploids clustered with diploid individuals; however, LIP triploids occupied an intermediate position between diploids and tetraploids (Figure ). 3.4 Interploidy Introgression and the Direction of Gene Flow We tested for the presence of interploidy introgression using a four‐taxon test (ABBA‐BABA) by setting different combinations of allopatric populations differing in ploidy and sympatric diploid and tetraploid sub‐populations as donors/recipients of introgression and spatially distinct tetraploid populations outside the contact zone (STE) as P1 (Figure ). The analyses revealed significant interploidy introgression in mixed‐ploidy populations, whereas no significant admixture was observed between tetraploids and diploids sampled outside the contact zone. Tree topologies testing for introgression between tetraploid and diploid individuals from mixed‐ploidy populations LIP and HLI resulted in significant D values of 0.28 and 0.36, respectively (Table ). On the contrary, low and statistically non‐significant D values were found for tree topologies involving tetraploid (SUM, JES) and diploid (REC) populations further away from the contact zone, demonstrating a lack of detectable admixture in pure‐ploidy allopatric populations (Figure ). To complement the introgression tests, the strength and direction of gene flow in the two mixed‐ploidy populations were also estimated using coalescent simulations. The scenario assuming bidirectional interploidy gene flow exhibited the lowest median and absolute AIC values with both populations (Figure ). The second‐best scenario, assuming only unidirectional 2x‐to‐4x (forward in time) gene flow, was markedly worse (ΔAIC = 3913.45) than the bidirectional scenario in both the LIP and HLI populations (median ΔAIC = 3022.81 and 4439.32, respectively; Figure ). The estimated migration rate (i.e. the probability of an individual sprouting in one ploidy subpopulation from seed originating in another one over the course of one generation) was greater in the direction from diploids to tetraploids, forward in time (4.42 × 10 −5 and 1.05 × 10 −4 for population LIP and population HLI, respectively), compared to gene flow from tetraploids to diploids (1.69 × 10 −5 and 1.21 × 10 −5 for population LIP and population HLI, respectively).
Cytotype Structure in the Contact Zone and Within Mixed‐Ploidy Populations Three main ploidy levels with RGS values corresponding to diploids, triploids and tetraploids were detected among the 356 field‐collected adult individuals (Table ). In addition, three putative aneuploids with RGS values in between triploid and diploid (two plants) and between triploid and tetraploid (one plant) were sampled in the mixed‐ploidy populations LIP and HLI, respectively (Figure ; Table ). The fine‐scale distribution of cytotypes in the two mixed‐ploidy populations revealed the presence of both ploidy‐uniform patches and parts where all three cytotypes grow only a few metres apart (Figure ). The seeds collected in natural mixed‐ploidy populations had variable but generally meagre germination rates (0%–60%, 8% on average, Figure ). Among 122 germinated seedlings, the RGS values corresponded to both euploid (84%) and aneuploid (16%) values (Figure ). Seedlings of both diploid and triploid mothers exhibited values corresponding to the diploid state (79% and 21% of seedlings from diploid and triploid mothers, respectively). The rest were putative aneuploids with RGS values between diploids and tetraploids, with the exception of one putative tetraploid found among the progeny of a diploid parent. The RGS of seedlings from tetraploid mothers ranged from values corresponding to hypo‐tetraploid (15%) and tetraploid (62%) up to hyper‐tetraploid (23%) aneuploid (Figure ). No seedling with RGS corresponding to the triploid state was found.
Strength of the Interploidy Barrier Inferred From Crossing Experiments We first crossed diploid and tetraploid plants to investigate the strength of postzygotic barriers (triploid block, Crossing Experiment 1). The difference in the seed set was significant overall ( F ₃,₄₇ = 3.71, p = 0.018, Figure ), yet rather small, as it was only the category of homoploid control 4x × 4x crosses that were significantly higher than one of the interploidy crosses (2x × 4x; Figure ). The seed mass was also different ( F ₃,₄₇ = 98.62, p < 0.001), being similar between the interploidy crosses but markedly lower than that of homoploid crosses of both types (Figure ). Consequently, there was also significant variation in germination rates among the crossing treatments (Figure , χ 2 = −339.2, df = 3, p < 0.001). The progeny of homoploid controls had higher germination percentages (40% and 55% for diploid and tetraploid crosses, respectively) than interploidy crosses, in which no viable seeds were produced by diploid seed parents and only a single germinable seed (< 0.5%) was formed after pollination of a tetraploid by a diploid pollen donor. The RGS value of this plant corresponded to a triploid (Figure ). Then, triploid plants were crossed with diploids and tetraploids to investigate the potential of natural triploids to act as introgression mediators (i.e. as part of the triploid bridge pathway, Crossing Experiment 2). There was a significant difference in seed sets between successful crosses ( F ₅,₅₈ = 5.06, p < 0.001), with larger seed sets in tetraploid control crosses than in the majority of interploidy crosses (Figure ). There was no significant difference between the diploid control and the interploidy crosses, with average seed set values of interploidy crosses ranging from 0.3 to 5.5 seeds per silique. There was no statistically significant difference in seed mass among the different types of crosses ( F ₅,₅₈ = 0.94, p = 0.4608; Figure ). The germination rates, again, differed significantly between the different types of crosses ( χ 2 = 97.018, df = 5, p < 0.001, Figure ), although the difference was not as pronounced as in the previous experiment. Tetraploid controls generally had significantly higher germination rates than the other types of crosses, with the exception of one type of backcross that was similar (4x × 3x). The germination rates of triploid backcrosses were highly variable but always non‐zero (8%–75% across treatments; Figure ). The relative genome size of the plants obtained from successful triploid backcrosses (36 cytotyped plants) corresponded to both euploid and aneuploid values defined based on the relative deviation of RGS from the control euploid values. The ploidy of the crossing partner of a triploid (diploid or tetraploid) significantly affected the proportion of putatively euploid progeny ( χ 2 = 21.78, df = 1, p < 0.001). Crosses of triploids with diploids resulted in a higher proportion of putatively euploid progeny (41%, with RGS values corresponding to diploids) than crosses involving tetraploids, where the proportion of euploid (tetraploid) progeny was 25% (Figure ). Chromosome counts obtained from three individuals classified by RGS as aneuploids confirmed the aneuploid number of chromosomes: F_8 (3x × 4x) – hypotetraploid, 2 n = 31; L_1 (3x × 4x)—hypopentaploid, 2 n = 38; M_8 (3x × 2x)—hypotetraploid, 2 n = 31 (Figure ).
Genetic Structure Based on Genome‐Wide SNPs Sequencing of 84 individuals from six populations (Figure ) produced between 24,453,671 and 84,627,579 reads per sample, averaging 38,264,942.5 reads after quality control and deduplication (Table ). Of these, 37%–94.2% were successfully mapped onto the reference genome, with an average mapping rate of 83.55%. The final VCF file contained 1,448,166 filtered putatively neutral four‐fold degenerate SNPs with an average depth of coverage of > 30× that were used in subsequent analyses. Bayesian clustering in STRUCTURE, based on a subset of 7169 LD‐pruned SNPs, suggested K = 2, 3 and 4 as stable partitions (strong similarity across runs), with K = 3 exhibiting the highest relative likelihood difference (delta K ). Diploids and tetraploids from cytotype‐pure populations separated already under K = 2 (Figure ). Additional separation of tetraploid individuals from the mixed‐ploidy population LIP was observed under K = 3 (Figure ) and K = 4 (population JES, Figure ). Tetraploids from the two mixed‐ploidy populations, HLI and LIP, however, showed a high proportion of diploid cluster ancestry under K = 2. The separation of pure diploid and tetraploid populations, in contrast to the closer position of both major cytotypes from mixed‐ploidy populations, was also supported by the neighbour‐joining network (Figure ). Interestingly, triploids from both mixed‐ploidy populations showed contrasting assignment patterns. Triploids from population HLI clustered exclusively with their sympatric diploids in both STRUCTURE and neighbour‐joining network analyses, whereas triploids from population LIP were a mixture of both STRUCTURE clusters and occupied intermediate positions in the network (Figure ). The contrasting genetic make‐up of triploids was further corroborated by principal component analyses run separately for each mixed‐ploidy population where HLI triploids clustered with diploid individuals; however, LIP triploids occupied an intermediate position between diploids and tetraploids (Figure ).
Interploidy Introgression and the Direction of Gene Flow We tested for the presence of interploidy introgression using a four‐taxon test (ABBA‐BABA) by setting different combinations of allopatric populations differing in ploidy and sympatric diploid and tetraploid sub‐populations as donors/recipients of introgression and spatially distinct tetraploid populations outside the contact zone (STE) as P1 (Figure ). The analyses revealed significant interploidy introgression in mixed‐ploidy populations, whereas no significant admixture was observed between tetraploids and diploids sampled outside the contact zone. Tree topologies testing for introgression between tetraploid and diploid individuals from mixed‐ploidy populations LIP and HLI resulted in significant D values of 0.28 and 0.36, respectively (Table ). On the contrary, low and statistically non‐significant D values were found for tree topologies involving tetraploid (SUM, JES) and diploid (REC) populations further away from the contact zone, demonstrating a lack of detectable admixture in pure‐ploidy allopatric populations (Figure ). To complement the introgression tests, the strength and direction of gene flow in the two mixed‐ploidy populations were also estimated using coalescent simulations. The scenario assuming bidirectional interploidy gene flow exhibited the lowest median and absolute AIC values with both populations (Figure ). The second‐best scenario, assuming only unidirectional 2x‐to‐4x (forward in time) gene flow, was markedly worse (ΔAIC = 3913.45) than the bidirectional scenario in both the LIP and HLI populations (median ΔAIC = 3022.81 and 4439.32, respectively; Figure ). The estimated migration rate (i.e. the probability of an individual sprouting in one ploidy subpopulation from seed originating in another one over the course of one generation) was greater in the direction from diploids to tetraploids, forward in time (4.42 × 10 −5 and 1.05 × 10 −4 for population LIP and population HLI, respectively), compared to gene flow from tetraploids to diploids (1.69 × 10 −5 and 1.21 × 10 −5 for population LIP and population HLI, respectively).
Discussion In this study, we explore the pathways and rates of bidirectional interploidy gene flow in natural populations of a mixed‐ploidy plant species. By combining field surveys, crossing experiments and population genomics, we present robust evidence of significant gene flow between different ploidy levels across multiple natural populations. Specifically, we describe the pathway by which triploid hybrids form in C. amara and document their persistence in nature and their ability to backcross. Furthermore, we demonstrate genome‐wide signals of bidirectional introgression between diploid and tetraploid individuals growing in natural sympatry. In the following subsections, we discuss the mechanisms and rates of triploid formation and further implications of triploid bridges for plant evolution through WGD. 4.1 A Strong but Incomplete Postzygotic Barrier to Triploid Formation Our crossing experiments between diploids and tetraploids have revealed a very strong post‐pollination barrier, evidenced by the fact that only one viable triploid plant was produced. We speculate that a triploid block, that is decreased hybrid seed viability caused by the parent‐of‐origin epigenetic imbalance in endosperm development (Köhler et al. , ), is the likely underlying mechanism. This is suggested by the formation of large numbers of low‐germinable and overall malformed seeds in experimental interploidy crosses but not in controls. Similar phenotypes have been found in other Brassicaceae species, for example, of the genera Brassica and Arabidopsis , for which this mechanism has been comprehensively documented (Scott et al. ; Stoute et al. ; Morgan et al. ). In addition, the viable triploid seed was formed in a cross when the tetraploid acted as a mother plant, i.e. the cross direction that usually produces more viable offspring also in other plants with a triploid block (Morgan et al. ; Bartolić et al. ). However, additional embryological and transcriptomic investigation is needed to address the hypothesis of the epigenetic basis of the interploidy barrier in C. amara . In contrast, in natural C. amara populations, triploids constitute a significant, persistent entity, accounting for 5%–6% of adult individuals within mixed‐ploidy populations, consistently occurring across multiple years in the same spots and for over fifteen years in the same populations (Krasna ; Zozomová‐Lihová et al. ). This disparity between the outcomes of experimental and naturally occurring crosses contradicts typical observations. In several other plant polyploid systems, experimental crosses have shown a ‘leaky’ triploid block, yet hybrids were scarce or non‐existent in natural populations (Greiner and Oberprieler ; Sonnleitner et al. ; Hülber et al. ; Morgan et al. , ; Šemberová et al. ). This discrepancy has been attributed to factors such as prezygotic barriers, reduced hybrid fitness, or a combination of both. In the case of C. amara , there is no indication of strong prezygotic barriers, such as temporal isolation or pollinator preference, as all cytotypes are morphologically indistinguishable (Marhold ), coexist in immediate proximity and overlap in flowering phenology (personal observation). Considering the strong triploid block, the relatively frequent presence of triploids in nature can most likely be attributed to the life history of the species, particularly its perenniality and clonal reproductive strategy. In the case of C. amara , vegetative reproduction is frequent and vital (Koch et al. ), which may enhance the longevity and persistence of triploid hybrids once they have formed. Previous studies have shown that polyploids often depend on vegetative reproduction, which not only safeguards many nascent polyploids from extinction but also plays a crucial role in polyploid speciation by facilitating more efficient space utilisation and decreasing mortality from small‐scale disturbance events (Herben et al. ; Van Drunen and Husband ; Van Drunen and Friedman ). 4.2 Triploids as Mediators of Interploidy Gene Flow In spite of the traditional assumption that polyploidisation is an instantaneous and perfect barrier, steeply accumulating genomic evidence documents that interploidy gene flow is frequent and forces the reappraisal of its relevance for the formation, establishment, and further evolution of novel polyploid lineages (Chapman and Abbott ; Schmickl and Yant ; Bartolić et al. ; Brown et al. ). Even though there are multiple scenarios of interploidy admixture in which introgression is primarily unidirectional from diploids to polyploids, triploids serve as an essential conduit for gene flow in the reverse direction, enabling also introgression from polyploids to diploids (Petit et al. ; Kolář et al. ; Bartolić et al. ). Triploids are present in over 60% of well‐established mixed‐ploidy systems comprising diploids and tetraploids (Kolář et al. ), yet detailed genetic studies on their role in extant gene flow are limited because researchers have primarily focused on the role of triploids in the formation of new polyploids (Bretagnolle and Thompson ; Ramsey and Schemske ; Husband ). Triploid formation entails either a diploid–tetraploid cross (van Dijk and van Delden ; Peckert and Chrtek ; De Hert et al. ; Sabara et al. ; Vallejo‐Marín et al. ; Popelka et al. ; Castro et al. ) or the fusion of one reduced and one unreduced diploid gamete (Slovák et al. ; Schinkel et al. ; Šmíd et al. ). These two mechanisms have only rarely been found to coexist in the same system by studies based on cytotype distribution patterns (Mandák et al. ); however, sufficient sampling combined with thorough genotyping may reveal that such a pattern is more frequent. Here, we present genetic evidence for both pathways: diploid‐tetraploid hybridisation in population LIP and fusion of reduced and unreduced diploid gametes in population HLI (Figures and ). Such a result also implies that estimating the levels of interploidy gene flow solely based on triploid frequency may be misleading, as even triploids found in mixed‐ploidy populations may not always be hybrids (Bartolić et al. ). In addition to triploid formation, we also show that triploid hybrids are fertile and capable of backcrossing both in experimental and natural conditions. These results add to a body of evidence primarily based on pollen fertility assessments (Ramsey and Schemske ; Laport et al. ; Morgan et al. ) that (partially) fertile triploid hybrids may further contribute to the composition and dynamics of the contact zones, lending support to theoretical models (Husband ; Kauai et al. ). In a novel finding, we also show that the relative genome size of a significant (42%) proportion of triploid backcross progeny corresponds to either the diploid or the tetraploid level, demonstrating the potential of triploids as mediators of interploidy introgression between the two major euploid cytotypes, in line with genomic data (see the following subsection). Notably, however, the majority of progeny resulting from triploid backcrosses were still aneuploids, and some additional aneuploids differing by a single chromosome might have been misclassified as euploids because of the limited resolution of our flow cytometric approach. On the other hand, aneuploids might also play a role in mediating interploidy gene flow in experimental populations of Arabidopsis thaliana (Henry et al. , ). Interestingly, we also detected three viable adult individuals in the mixed‐ploidy populations, with RGS corresponding to aneuploid values suggesting that aneuploids may form and survive until adulthood also in natural populations, similarly as has been observed in contact zones between cytotypes of Tripleurospermum inodorum (Čertner et al. ). 4.3 Bidirectional and Asymmetric Interploidy Gene Flow In both mixed‐ploidy populations, the coalescent models supported ongoing bidirectional gene flow, aligning with both experimental findings and field observations. This also supports the involvement of triploid individuals as mediators of gene flow, as there is no alternative mechanism by which introgression could proceed from tetraploids towards diploids (Bartolić et al. ). Moreover, differences in the intensity of gene flow between the two investigated populations align with the distribution and frequency of cytotypes in the field. Gene flow towards diploids is more pronounced in population LIP, which exhibits a more intermingled, mosaic‐like cytotype structure and also harbours triploid individuals that have been proven to be hybrids. By contrast, signals of gene flow towards diploids are weaker in population HLI, where triploids are currently segregated from tetraploids and occur within diploid patches with a genetic profile close to diploids. This suggests that triploids might play a significant role in population LIP by mediating gene flow to diploids, whereas in population HLI we have not found any conclusive evidence for any ongoing interploidy hybridisation via triploids, at least based on our current sampling. Although bidirectional interploidy gene flow has been expected based on theoretical models (Husband ; Kauai et al. ; Felber and Bever ), its presence has been suggested only rarely in natural systems, mostly based on indirect evidence of genetic clusters spanning cytotypes (Ståhlberg ; Nierbauer et al. ; Šmíd et al. ). Our data, in a testable framework based on coalescent simulations, provide evidence for bidirectional gene flow. We speculate that gene flow in C. amara is ongoing (population LIP) or at least recent (population HLI), reflecting the presence of fertile triploids in the field and genetic support for the presence of introgression in mixed‐ploidy populations but not outside the contact zone. In both mixed‐ploidy populations, gene flow was inferred to be stronger in the direction towards tetraploids. Such an asymmetry likely reflects an additional route of unidirectional gene flow from lower to higher ploidy: the merger of an unreduced gamete of a diploid with a reduced gamete of a tetraploid leading to hybrid tetraploid progeny. Indeed, previous extensive crossing experiments often found a certain fraction of such tetraploid hybrids, demonstrating that this pathway may act in addition to a triploid bridge (e.g. van Dijk and van Delden ; Burton and Husband ; Chrtek et al. ; Sutherland and Galloway ; Castro et al. ; Morgan et al. ). Although we did not encounter such a hybrid in our limited crossing experiment, the observation of a tetraploid seedling among the progeny of a diploid seed parent sampled in population HLI demonstrates that unreduced gametes of diploids may also be involved in tetraploid formation in the field. The overall importance of this pathway in ploidy variable systems is illustrated by the fact that in most well‐documented cases of interploidy gene flow, the direction is typically inferred as unidirectional, from lower to higher ploidy levels (Zohren et al. ; Kryvokhyzha et al. ; Monnahan et al. ; Wang et al. ; Leal et al. ; Ding et al. ). Considering that unreduced gametes are present in many systems (Ramsey and Schemske ; Kreiner et al. , ), it is no surprise that this direction is more prevalent in nature (Stebbins ; Brown et al. ; Bartolić et al. ). Interploidy gene flow may be an important mechanism of how the gene pool of a nascent, initially depauperate polyploid may be enriched. Specifically, in cases involving autotetraploids, recent genomic evidence suggests that such gene flow provides novel genetic variation that can benefit polyploid adaptation (Baduel et al. , reviewed in Schmickl and Yant ). Interploidy gene flow also constitutes a key component of a pathway enabling gene flow between different species. Mathematical models have shown that WGD‐mediated gene flow can serve as a bridge between diploid lineages, where introgression is otherwise impeded by interspecies barriers (Kauai et al. ). This lateral transfer of genes from polyploids to diploids has been observed in several grass species, particularly with genes coding for the C4 photosynthetic pathway (Christin et al. ; Phansopa et al. ). The spread of genetic variation to the diploid level is facilitated exclusively by triploid crosses, highlighting the crucial role of the triploid bridge in intra‐ and interspecific genetic exchange and possibly adaptation.
A Strong but Incomplete Postzygotic Barrier to Triploid Formation Our crossing experiments between diploids and tetraploids have revealed a very strong post‐pollination barrier, evidenced by the fact that only one viable triploid plant was produced. We speculate that a triploid block, that is decreased hybrid seed viability caused by the parent‐of‐origin epigenetic imbalance in endosperm development (Köhler et al. , ), is the likely underlying mechanism. This is suggested by the formation of large numbers of low‐germinable and overall malformed seeds in experimental interploidy crosses but not in controls. Similar phenotypes have been found in other Brassicaceae species, for example, of the genera Brassica and Arabidopsis , for which this mechanism has been comprehensively documented (Scott et al. ; Stoute et al. ; Morgan et al. ). In addition, the viable triploid seed was formed in a cross when the tetraploid acted as a mother plant, i.e. the cross direction that usually produces more viable offspring also in other plants with a triploid block (Morgan et al. ; Bartolić et al. ). However, additional embryological and transcriptomic investigation is needed to address the hypothesis of the epigenetic basis of the interploidy barrier in C. amara . In contrast, in natural C. amara populations, triploids constitute a significant, persistent entity, accounting for 5%–6% of adult individuals within mixed‐ploidy populations, consistently occurring across multiple years in the same spots and for over fifteen years in the same populations (Krasna ; Zozomová‐Lihová et al. ). This disparity between the outcomes of experimental and naturally occurring crosses contradicts typical observations. In several other plant polyploid systems, experimental crosses have shown a ‘leaky’ triploid block, yet hybrids were scarce or non‐existent in natural populations (Greiner and Oberprieler ; Sonnleitner et al. ; Hülber et al. ; Morgan et al. , ; Šemberová et al. ). This discrepancy has been attributed to factors such as prezygotic barriers, reduced hybrid fitness, or a combination of both. In the case of C. amara , there is no indication of strong prezygotic barriers, such as temporal isolation or pollinator preference, as all cytotypes are morphologically indistinguishable (Marhold ), coexist in immediate proximity and overlap in flowering phenology (personal observation). Considering the strong triploid block, the relatively frequent presence of triploids in nature can most likely be attributed to the life history of the species, particularly its perenniality and clonal reproductive strategy. In the case of C. amara , vegetative reproduction is frequent and vital (Koch et al. ), which may enhance the longevity and persistence of triploid hybrids once they have formed. Previous studies have shown that polyploids often depend on vegetative reproduction, which not only safeguards many nascent polyploids from extinction but also plays a crucial role in polyploid speciation by facilitating more efficient space utilisation and decreasing mortality from small‐scale disturbance events (Herben et al. ; Van Drunen and Husband ; Van Drunen and Friedman ).
Triploids as Mediators of Interploidy Gene Flow In spite of the traditional assumption that polyploidisation is an instantaneous and perfect barrier, steeply accumulating genomic evidence documents that interploidy gene flow is frequent and forces the reappraisal of its relevance for the formation, establishment, and further evolution of novel polyploid lineages (Chapman and Abbott ; Schmickl and Yant ; Bartolić et al. ; Brown et al. ). Even though there are multiple scenarios of interploidy admixture in which introgression is primarily unidirectional from diploids to polyploids, triploids serve as an essential conduit for gene flow in the reverse direction, enabling also introgression from polyploids to diploids (Petit et al. ; Kolář et al. ; Bartolić et al. ). Triploids are present in over 60% of well‐established mixed‐ploidy systems comprising diploids and tetraploids (Kolář et al. ), yet detailed genetic studies on their role in extant gene flow are limited because researchers have primarily focused on the role of triploids in the formation of new polyploids (Bretagnolle and Thompson ; Ramsey and Schemske ; Husband ). Triploid formation entails either a diploid–tetraploid cross (van Dijk and van Delden ; Peckert and Chrtek ; De Hert et al. ; Sabara et al. ; Vallejo‐Marín et al. ; Popelka et al. ; Castro et al. ) or the fusion of one reduced and one unreduced diploid gamete (Slovák et al. ; Schinkel et al. ; Šmíd et al. ). These two mechanisms have only rarely been found to coexist in the same system by studies based on cytotype distribution patterns (Mandák et al. ); however, sufficient sampling combined with thorough genotyping may reveal that such a pattern is more frequent. Here, we present genetic evidence for both pathways: diploid‐tetraploid hybridisation in population LIP and fusion of reduced and unreduced diploid gametes in population HLI (Figures and ). Such a result also implies that estimating the levels of interploidy gene flow solely based on triploid frequency may be misleading, as even triploids found in mixed‐ploidy populations may not always be hybrids (Bartolić et al. ). In addition to triploid formation, we also show that triploid hybrids are fertile and capable of backcrossing both in experimental and natural conditions. These results add to a body of evidence primarily based on pollen fertility assessments (Ramsey and Schemske ; Laport et al. ; Morgan et al. ) that (partially) fertile triploid hybrids may further contribute to the composition and dynamics of the contact zones, lending support to theoretical models (Husband ; Kauai et al. ). In a novel finding, we also show that the relative genome size of a significant (42%) proportion of triploid backcross progeny corresponds to either the diploid or the tetraploid level, demonstrating the potential of triploids as mediators of interploidy introgression between the two major euploid cytotypes, in line with genomic data (see the following subsection). Notably, however, the majority of progeny resulting from triploid backcrosses were still aneuploids, and some additional aneuploids differing by a single chromosome might have been misclassified as euploids because of the limited resolution of our flow cytometric approach. On the other hand, aneuploids might also play a role in mediating interploidy gene flow in experimental populations of Arabidopsis thaliana (Henry et al. , ). Interestingly, we also detected three viable adult individuals in the mixed‐ploidy populations, with RGS corresponding to aneuploid values suggesting that aneuploids may form and survive until adulthood also in natural populations, similarly as has been observed in contact zones between cytotypes of Tripleurospermum inodorum (Čertner et al. ).
Bidirectional and Asymmetric Interploidy Gene Flow In both mixed‐ploidy populations, the coalescent models supported ongoing bidirectional gene flow, aligning with both experimental findings and field observations. This also supports the involvement of triploid individuals as mediators of gene flow, as there is no alternative mechanism by which introgression could proceed from tetraploids towards diploids (Bartolić et al. ). Moreover, differences in the intensity of gene flow between the two investigated populations align with the distribution and frequency of cytotypes in the field. Gene flow towards diploids is more pronounced in population LIP, which exhibits a more intermingled, mosaic‐like cytotype structure and also harbours triploid individuals that have been proven to be hybrids. By contrast, signals of gene flow towards diploids are weaker in population HLI, where triploids are currently segregated from tetraploids and occur within diploid patches with a genetic profile close to diploids. This suggests that triploids might play a significant role in population LIP by mediating gene flow to diploids, whereas in population HLI we have not found any conclusive evidence for any ongoing interploidy hybridisation via triploids, at least based on our current sampling. Although bidirectional interploidy gene flow has been expected based on theoretical models (Husband ; Kauai et al. ; Felber and Bever ), its presence has been suggested only rarely in natural systems, mostly based on indirect evidence of genetic clusters spanning cytotypes (Ståhlberg ; Nierbauer et al. ; Šmíd et al. ). Our data, in a testable framework based on coalescent simulations, provide evidence for bidirectional gene flow. We speculate that gene flow in C. amara is ongoing (population LIP) or at least recent (population HLI), reflecting the presence of fertile triploids in the field and genetic support for the presence of introgression in mixed‐ploidy populations but not outside the contact zone. In both mixed‐ploidy populations, gene flow was inferred to be stronger in the direction towards tetraploids. Such an asymmetry likely reflects an additional route of unidirectional gene flow from lower to higher ploidy: the merger of an unreduced gamete of a diploid with a reduced gamete of a tetraploid leading to hybrid tetraploid progeny. Indeed, previous extensive crossing experiments often found a certain fraction of such tetraploid hybrids, demonstrating that this pathway may act in addition to a triploid bridge (e.g. van Dijk and van Delden ; Burton and Husband ; Chrtek et al. ; Sutherland and Galloway ; Castro et al. ; Morgan et al. ). Although we did not encounter such a hybrid in our limited crossing experiment, the observation of a tetraploid seedling among the progeny of a diploid seed parent sampled in population HLI demonstrates that unreduced gametes of diploids may also be involved in tetraploid formation in the field. The overall importance of this pathway in ploidy variable systems is illustrated by the fact that in most well‐documented cases of interploidy gene flow, the direction is typically inferred as unidirectional, from lower to higher ploidy levels (Zohren et al. ; Kryvokhyzha et al. ; Monnahan et al. ; Wang et al. ; Leal et al. ; Ding et al. ). Considering that unreduced gametes are present in many systems (Ramsey and Schemske ; Kreiner et al. , ), it is no surprise that this direction is more prevalent in nature (Stebbins ; Brown et al. ; Bartolić et al. ). Interploidy gene flow may be an important mechanism of how the gene pool of a nascent, initially depauperate polyploid may be enriched. Specifically, in cases involving autotetraploids, recent genomic evidence suggests that such gene flow provides novel genetic variation that can benefit polyploid adaptation (Baduel et al. , reviewed in Schmickl and Yant ). Interploidy gene flow also constitutes a key component of a pathway enabling gene flow between different species. Mathematical models have shown that WGD‐mediated gene flow can serve as a bridge between diploid lineages, where introgression is otherwise impeded by interspecies barriers (Kauai et al. ). This lateral transfer of genes from polyploids to diploids has been observed in several grass species, particularly with genes coding for the C4 photosynthetic pathway (Christin et al. ; Phansopa et al. ). The spread of genetic variation to the diploid level is facilitated exclusively by triploid crosses, highlighting the crucial role of the triploid bridge in intra‐ and interspecific genetic exchange and possibly adaptation.
Conclusion The triploid bridge is a fundamental concept in polyploid biology, offering a mechanistic explanation for bidirectional gene flow in polyploid complexes. Our study provides integrative evidence for the action of this pathway in C. amara by demonstrating the ability of naturally formed triploids to backcross and detecting significant genomic footprints of bidirectional introgression in natural populations. On the other hand, our data also reveal that not all triploid individuals in mixed‐ploidy populations are hybrids, and alternative pathways for their origin via autopolyploidisation from diploids may be involved. Remarkably, despite the strong triploid block that limits their formation, we observed a significant frequency of triploid hybrids in the mixed‐ploidy populations. This likely reflects the capacity of C. amara to spread and persist clonally, allowing triploids to maintain their presence over extended periods. Further studies in this system addressing temporal variation, ecological divergence and clonality may shed light on the spatio‐temporal dynamics of the mixed‐ploidy populations and mechanisms maintaining the cytotype coexistence in C. amara . In general, our findings validate theoretical predictions and demonstrate that odd‐ploidy cytotypes, though rare and often of reduced fertility, can significantly influence the genomic landscape of natural populations. We anticipate that future studies focused on triploid bridges across diverse species may clarify the role of species traits, such as clonality, in maintaining triploid hybrid progeny and facilitating bidirectional interploidy gene flow.
P.B., K.M., N.P.G. and F.K. conceived the ideas and designed the methodology. P.B., A.V., L.M., N.P.G. and G.Š. carried out the field work. P.B., L.M. and G.Š. conducted the laboratory work; the chromosome counts were conducted by T.M., P.B. and A.V. conducted the plant cultivation and crossing experiments. P.B. and M.Š. did the data analysis with input from N.P.G. and F.K. supervised the study. P.B. and F.K. drafted the manuscript. All authors reviewed and edited subsequent versions of the manuscript. All authors approved the final version.
The authors declare no conflicts of interest.
Data S1. Table S2. Summary of crossing experiments. Table S3. Details on the sequenced individuals.
|
Quantification of nucleic acid quality in postmortem tissues from a cancer research autopsy program | cfd6ce55-eb3d-4eb0-ae86-1485961fb0d8 | 5341846 | Pathology[mh] | Autopsy, derived from the Greek word autopsia meaning “to see for oneself”, is a method used since the 17 th century to learn about disease and determine the cause of death . Autopsy was a main form of understanding disease until the mid-20 th century when medical imaging developed and allowed a view of the internal organs in a living patient . In turn, the growth of laboratory medicine further diminished the need for autopsy as a diagnostic tool . While postmortem exam has remained fundamental to improving knowledge of brain diseases, particularly neurodegenerative disorders, a renewed interest in its use for studying human cancer has gained traction in the past decade . Sequencing of the human genome has led to a revolution in understanding of cancer etiology by revealing the genetic alterations characteristic of human tumors , the genetic features that underlie subtypes within a primary tumor type , or the mechanisms of therapeutic resistance . With these advances has come a revival of interest in postmortem tissue collection because advanced stage disease is typically not accessible for study by next-generation methods in samples from living patients. As a result, research autopsy programs have emerged as a critical tool towards understanding the biology of lethal cancer and in many instances have led to significant insights into cancer progression and treatment resistance not possible with small tumor biopsies . Despite the emergence of and implementation of research autopsy programs at a variety of institutions for obtaining cancer tissues, to date there are few formal evaluations of the quality of biomolecules in postmortem materials. A challenge to performing such evaluations is limited tissue resources. In some instances carefully screened cases and selected tissue types have been used to establish the relationship between potential quality-controlling factors and tissue sample quality . Alternatively, simulated postmortem scenarios are used to mimic the postmortem interval and natural environment . Therefore, towards the goal of fully understanding these issues we leveraged our experience and resources amassed while running a cancer research autopsy program spanning more than a decade to determine the quality of nucleic acids in relation to tissue of origin, postmortem interval, normal versus neoplastic histology, primary versus metastases, and performance in downstream next generation sequencing methodologies.
Sample set characteristics Nucleic acid quality was analyzed in 371 different frozen tissue samples collected from 80 autopsied patients, 81% of which had been diagnosed with pancreatic cancer. The remaining patients had been diagnosed with breast cancer, lung cancer, colorectal cancer, germ cell tumor or melanoma (Figure ). The postmortem interval (PMI) of these 80 autopsies ranged from 2 hours to > 36 hours. Cases with short PMIs were typically for those patients who expired while at the hospital, whereas cases with very long PMIs were a result of many factors including transport from outside the hospital or consent in the postmortem period by the patients' legally authorized representative to the program. Among the 371 tissue samples, 287 were histologically confirmed normal tissues sampled from the liver, lung, kidney, pancreas, spleen, heart, skeletal muscle and skin with a median of 35 normal tissues per site (range 30 to 49). We also collected 84 tumor samples of which 52 were from primary tumors, 16 from liver metastases and 16 from lung metastases including 10 patient-matched primary-metastatic pairs. To facilitate analyses, samples were arbitrarily categorized into four groups based on PMI: Category I, PMI ≤ 5 hours; Category II, PMI 6–10 hours; Category III, PMI 11–20 hours; and Category IV, ≥ 21 hours. When tumor samples were included there were a total of 98 samples from PMI Category I, 110 samples from PMI Category II, 88 samples from Category III, and 74 samples from Category IV (Table ). RNA integrity in normal tissues is tissue-type specific We first determined the extent to which RNA could be extracted from this large set of postmortem normal tissues. RNA was successfully extracted from 269 of 287 (94%) samples attempted. The overall average RIN for all 269 samples was 5.94 ± 2.5, with a median RIN of 6.4. By contrast, for 10 of 287 samples (6%) the RNA yield was exceedingly low leading to unreported or unreliable RIN, and these samples were assigned an RIN value of 0 (Figure ). In general, higher overall RNA yields were found from tissue samples collected within 1 year compared to those with long-term storage (> 5 years). This was unrelated to the number of freeze/thaw cycles per sample as with rare exceptions all frozen normals analyzed were previously unused and continuously stored at −80°C. Furthermore, low RNA yields (defined as 20 ug/ml total RNA) were unrelated to PMI nor were they related to a specific histology. We next calculated the mean RIN values for each individual tissue type for which RNA was obtained (Table ). Overall, mean RIN values showed little variability among the eight normal tissue types and most tissue types had RIN values between 5 to 6.5. The tissue with the lowest mean RIN value was the kidney (RIN 4.63 ± 1.95) whereas the highest values were noted for skeletal muscle (RIN 9.01 ± 1.36), suggesting tissue-type specific differences in RNA stability in the postmortem interval. This pattern did not change when the median RIN value in each category was alternatively considered. To determine the extent to which RIN values of normal tissues show intra-patient variability we evaluated a subset of patients from each PMI category for which multiple normal tissues were evaluated (Figure ). In all patients we noted variability in RIN values among different tissues, ranging from as low as an RIN values of < 2 to > 9 in a single individual (for example, patient RA15-11 in Category III or A164 in Category IV). However, the overall variability was less in PMI Category I samples than for PMI Category IV samples. Thus, RNA quality of one normal tissue retrieved postmortem is not a reliable predictor of RNA quality in a second tissue from that same patient, and good quality samples can be obtained despite the length of the postmortem interval. We then determined the relationship of RIN values to PMI interval in greater detail by performing correlation analyses. Statistically significant negative correlations were noted for the liver, lung, kidney, pancreas, spleen and skin (Figure ). Liver and skin showed particularly strong negative correlations between RIN value and PMI with r values close to −0.5 and p value s < 0.01. By contrast, no correlations were found for the heart or skeletal muscle with the RNA showing remarkable stability and quality in patients with PMIs as long as 36 hours or greater. No correlation was noticed between RIN and the length of sample storage in any of the tissue types examined (Figure ). Taken together, these results indicate that the RNA quality of normal tissues declines with the elongation of PMI but not storage time, and the extent of degradation is tissue-type specific. These results nonetheless demonstrate a clear negative correlation between RNA integrity and PMI in most tissue types indicating the importance of this variable. This correlation was established in the presence of several unavoidable confounding factors such as pyrexia, cachexia or prolonged hypoxia in the perimortem period, indirectly confirming that these factors may not be as influential as PMI in predicting RNA quality. We also observed a striking lack of correlation of PMI with RIN in normal skeletal muscle, and to a similar extent the heart, supporting the tissue type-specific nature of RNA degradation. In forensic settings, RNA has been shown to be stable in muscle up to 1week after death . While not addressed in this study, tissue-type specific RNA degradation has also been reported in ocular tissues with avascular structures having better RNA quality than vascularized structures such as the ciliary body . Consistent with this notion, we found that normal kidney and liver, two highly vascularized organs, had among the lowest RIN values in each PMI Category. It may be reasonable to speculate that, when controlling for other factors, vascularized tissues are more sensitive to nutrient and oxygen deprivation resulting in a greater extent of sample degradation postmortem. However, given RNA decay is a precisely controlled process in living cells , such a process may also contribute to RNA quality in the postmortem period as suggested by Romero et al. . RNA integrity in cancer tissues We next wondered if the integrity of ribonucleic acids in cancer tissues parallels that of normal tissues. To address this question, we first analyzed RNA quality in 52 primary tumor tissues and 32 metastatic tumor tissues from the liver and lung. When stratified by PMI Category there were 15 primary tumors and 10 metastases in PMI category I, 18 primary tumors and 10 metastases in PMI category II, nine primary tumors and six metastases in PMI category III, and 10 primary tumors and six metastases in PMI category IV (Table ). The mean RIN value in primary tumors was 5.16 ± 2.4, and for liver and lung metastases was 5.07 ± 2.51and 6.29 ± 2.72, respectively (Table ). There was no correlation between RIN values and PMI in primary tumor tissues (Figure ). Forty-three of the 52 primary tumors (83%) analyzed were pancreatic ductal adenocarcinomas (PDA) (Figure ), providing an opportunity to compare the RIN values in primary PDAs specifically to that of normal pancreatic tissues. The mean RNA integrity in PDA tissues was not significantly different from that of normal pancreas tissues when considering all samples (mean RIN 5.4 ± 2.4 versus 5.26 ± 2.56 respectively, p = NS), or when limiting the comparison to 17 matched pairs of normal pancreas and primary PDA (mean 6.16 ± 1.96 vs 5.17 ± 2.35 respectively, p = NS). This finding thus does not support the long-held “myth” that PDA tissues have worse quality than other tumor types. This may be partially explained by observations that PDA is characterized by a prominent desmoplastic/stromal reaction that is hypovascular compared to adjacent normal pancreas . Nonetheless, we have found that screening multiple geographically distinct samples from different regions of the same neoplasm may be necessary to identify regions with preserved RNA quality, as we have recently found in ongoing work in our laboratory (Figure ). Finally, we next explored the relationships of metastatic tumor RNA quality between matched liver and lung metastases, i.e. from the same patient. There was no statistically significant correlation (Figure ), indicating that RNA quality is highly variable among metastases, even within the same patient. Thus, unlike normal tissues that show fairly predictable and tissue-specific degradation in relation to PMI, cancer tissues derived from different organ sites appear less predictable with respect to RNA quality than that of normal tissues. DNA integrity in normal tissues is tissue-type specific While generally more stable than RNA, DNA is also subject to degradation in the postmortem period . Most methods to assess DNA degradation depend on examining selected target(s) to represent the overall sample quality with PCR-based methods among the most popular approach for this purpose . However, the extent to which such methods are reproducible or subject to inter-experiment variation is unknown, as is the extent to which genomic DNA in postmortem tissues follows similar kinetics as RNA. With these factors in mind we developed a semi-quantitative method to evaluate DNA quality to facilitate comparison among samples from independent experiments. In addition, unlike most studies detecting one locus, we simultaneously examined five chromosome loci of varying potential stability and susceptibility to DNA damage-inducing factors thereby achieving high sensitivity in detecting DNA damage in well-preserved samples (Table ). DNA was extracted from 36 frozen autopsy samples that were collected from five patients in PMI Category I and five in PMI category IV (Table ). To facilitate comparisons between RNA and DNA quality, samples with a wide spectrum of RIN values were selected from each PMI category, ranging from as low as 2.3 to as high as 9.4. These included samples from normal liver ( n = 7), normal kidney ( n = 9), primary tumors ( n = 10), liver metastases ( n = 5) and lung ( n = 5) metastases. We successfully extracted genomic DNA from all 36 samples including one that failed in RNA extraction. No DNA damage was detected by our assay in 32/36 samples analyzed (89%) even though 17 of these 32 (53%) had RIN values less than 5. In the four samples with DNA damage three were from PMI category IV, two of which showed degradation at all four sensor loci (Figure ). The remaining two samples showed only moderate damage as reflected by only two of the four chromosome loci affected. All four samples with DNA damage had concurrent RNA degradation (Table ). Interestingly all four samples with DNA damage were from the liver, three of which were histologically normal, further supporting the greater susceptibility of vascularized tissues to postmortem degradation. All other samples except for these four patients did not show DNA damage, including all 10 primary tumors analyzed. Thus, while DNA damage in postmortem tissues may be an indicator of RNA quality the converse is not true. Moreover, in addition to PMI the kinetics of postmortem DNA degradation may also have tissue type-specific factors. RNA sequencing using postmortem RNA samples Genomic DNA from postmortem tissues has been used successfully for next generation sequencing despite potential DNA degradation . However, given that RNA is more sensitive than DNA during the postmortem interval, its utility in downstream sequencing applications is unknown. As a proof of principle study, we therefore performed RNA sequencing on five matched pairs of normal pancreas and pancreatic cancer tissues, all with an RIN value of 5 or greater. Sequencing libraries were successfully generated from all samples using the poly-A enrichment method and used to generate 80 million reads per sample. Moreover, quality metrics of each library showed a sound distribution of coverage along transcripts and fragment lengths irrespective of RIN values (Figure ). One normal sample (patient A105), while showing good quality sequencing data, was excluded because the resulting data indicated contamination by cancer cells. This was confirmed histologically. A heat map of the top 250 genes differentially expressed between five pancreatic cancers and the remaining four normal pancreata showed a pattern that clearly discriminated the two groups (Figure ). Genes transcripts overrepresented in normal tissues included PRSS1 (cationic trypsinogen), CPA2 (pancreatic specific carboxypeptidase), AMY2A (pancreatic amylase 2), and the pancreatic developmental transcription factor PTF1A consistent with the greater abundance of acinar cells or cells with stem-like properties within these samples . Cancer samples showed greater heterogeneity with respect to the most differentially expressed genes. Genes overrepresented in the cancer samples included PSCA , MMP3 , MMP11 and SOX2 . The small sample size otherwise precluded a more thorough classification of each carcinoma's subtype as recently described . Finally, we leveraged our sequencing coverage to identify potential fusion events. We identified two recurrent fusion events, AXGP1-GJC3 and SIDT2-TAGLN, in six of nine postmortem RNA samples (three normal and three tumor, including two normal-tumor matched pairs). These two fusions were recently reported in normal pancreatic tissues within the context of a larger pan-tissue analysis . Collectively, these data are encouraging and suggest that despite being collected under postmortem conditions RNA samples can provide biologically meaningful information in downstream analyses. These findings are particularly exciting considering recent reports of improved methods to directly assess mRNA integrity and control for it in analyses of RNAseq data . While we did not utilize these methods for our pilot comparison, we nonetheless were able to discern gene expression signatures of known biologic importance in normal and pancreatic cancer tissues. With larger sample sizes and use of these methods it is reasonable to expect novel observations to be made in cancer tissues, for example with respect to treatment resistance or subclonal evolution. It is also important to note that our data do not indicate the RIN threshold value below which RNA sequencing cannot be performed, and in light of the pilot data shown it may be worth studying samples with low RIN values as well particularly as 39% of our samples had an RIN value of < 5. Previous evaluations of the quality of biomolecules in postmortem tissues have been in the context of brain banking, forensic analyses or minimally invasive autopsies . However, such an analysis using tissues obtained from a research autopsy program for cancer research has not been performed. Such information is critical to know with the growing interest in creation of biobanks from postmortem tissues of cancer patients, and use of these materials for ongoing scientific discovery and collaborations. While the scope of our dataset parallels that of another recent study , it differs in that we also studied DNA integrity as well and cancer tissues derived from different primary or metastatic sites. Nonetheless, our data is consistent with these prior studies that indicate nucleic acid quality of normal tissues is affected by a considerable number of factors in addition to the postmortem interval. At least one of these is likely the cause of death as we noticed several patients with a short PMI (< 3 hours) with exceedingly poor RNA quality, all of whom died of sepsis (personal observations, C.I.D.). Despite an increasing demand for research autopsy samples in the cancer research community, collecting high quality tissue samples is challenging due to numerous factors that can be legal, ethnic, emotional or social in nature. Exclusion of samples with a degree of poor quality is not always realistic, particularly when not all downstream applications are equally sensitive to sample quality. Thus, analyses that can incorporate these variables would be expected to improve comparisons across patients, tissue types and sample sets. Table summarizes these variables and their effects on RNA quality. Our hope is that these findings provide insight on the sample variability expected from research autopsy resources and ultimately facilitate data interpretation.
Nucleic acid quality was analyzed in 371 different frozen tissue samples collected from 80 autopsied patients, 81% of which had been diagnosed with pancreatic cancer. The remaining patients had been diagnosed with breast cancer, lung cancer, colorectal cancer, germ cell tumor or melanoma (Figure ). The postmortem interval (PMI) of these 80 autopsies ranged from 2 hours to > 36 hours. Cases with short PMIs were typically for those patients who expired while at the hospital, whereas cases with very long PMIs were a result of many factors including transport from outside the hospital or consent in the postmortem period by the patients' legally authorized representative to the program. Among the 371 tissue samples, 287 were histologically confirmed normal tissues sampled from the liver, lung, kidney, pancreas, spleen, heart, skeletal muscle and skin with a median of 35 normal tissues per site (range 30 to 49). We also collected 84 tumor samples of which 52 were from primary tumors, 16 from liver metastases and 16 from lung metastases including 10 patient-matched primary-metastatic pairs. To facilitate analyses, samples were arbitrarily categorized into four groups based on PMI: Category I, PMI ≤ 5 hours; Category II, PMI 6–10 hours; Category III, PMI 11–20 hours; and Category IV, ≥ 21 hours. When tumor samples were included there were a total of 98 samples from PMI Category I, 110 samples from PMI Category II, 88 samples from Category III, and 74 samples from Category IV (Table ).
We first determined the extent to which RNA could be extracted from this large set of postmortem normal tissues. RNA was successfully extracted from 269 of 287 (94%) samples attempted. The overall average RIN for all 269 samples was 5.94 ± 2.5, with a median RIN of 6.4. By contrast, for 10 of 287 samples (6%) the RNA yield was exceedingly low leading to unreported or unreliable RIN, and these samples were assigned an RIN value of 0 (Figure ). In general, higher overall RNA yields were found from tissue samples collected within 1 year compared to those with long-term storage (> 5 years). This was unrelated to the number of freeze/thaw cycles per sample as with rare exceptions all frozen normals analyzed were previously unused and continuously stored at −80°C. Furthermore, low RNA yields (defined as 20 ug/ml total RNA) were unrelated to PMI nor were they related to a specific histology. We next calculated the mean RIN values for each individual tissue type for which RNA was obtained (Table ). Overall, mean RIN values showed little variability among the eight normal tissue types and most tissue types had RIN values between 5 to 6.5. The tissue with the lowest mean RIN value was the kidney (RIN 4.63 ± 1.95) whereas the highest values were noted for skeletal muscle (RIN 9.01 ± 1.36), suggesting tissue-type specific differences in RNA stability in the postmortem interval. This pattern did not change when the median RIN value in each category was alternatively considered. To determine the extent to which RIN values of normal tissues show intra-patient variability we evaluated a subset of patients from each PMI category for which multiple normal tissues were evaluated (Figure ). In all patients we noted variability in RIN values among different tissues, ranging from as low as an RIN values of < 2 to > 9 in a single individual (for example, patient RA15-11 in Category III or A164 in Category IV). However, the overall variability was less in PMI Category I samples than for PMI Category IV samples. Thus, RNA quality of one normal tissue retrieved postmortem is not a reliable predictor of RNA quality in a second tissue from that same patient, and good quality samples can be obtained despite the length of the postmortem interval. We then determined the relationship of RIN values to PMI interval in greater detail by performing correlation analyses. Statistically significant negative correlations were noted for the liver, lung, kidney, pancreas, spleen and skin (Figure ). Liver and skin showed particularly strong negative correlations between RIN value and PMI with r values close to −0.5 and p value s < 0.01. By contrast, no correlations were found for the heart or skeletal muscle with the RNA showing remarkable stability and quality in patients with PMIs as long as 36 hours or greater. No correlation was noticed between RIN and the length of sample storage in any of the tissue types examined (Figure ). Taken together, these results indicate that the RNA quality of normal tissues declines with the elongation of PMI but not storage time, and the extent of degradation is tissue-type specific. These results nonetheless demonstrate a clear negative correlation between RNA integrity and PMI in most tissue types indicating the importance of this variable. This correlation was established in the presence of several unavoidable confounding factors such as pyrexia, cachexia or prolonged hypoxia in the perimortem period, indirectly confirming that these factors may not be as influential as PMI in predicting RNA quality. We also observed a striking lack of correlation of PMI with RIN in normal skeletal muscle, and to a similar extent the heart, supporting the tissue type-specific nature of RNA degradation. In forensic settings, RNA has been shown to be stable in muscle up to 1week after death . While not addressed in this study, tissue-type specific RNA degradation has also been reported in ocular tissues with avascular structures having better RNA quality than vascularized structures such as the ciliary body . Consistent with this notion, we found that normal kidney and liver, two highly vascularized organs, had among the lowest RIN values in each PMI Category. It may be reasonable to speculate that, when controlling for other factors, vascularized tissues are more sensitive to nutrient and oxygen deprivation resulting in a greater extent of sample degradation postmortem. However, given RNA decay is a precisely controlled process in living cells , such a process may also contribute to RNA quality in the postmortem period as suggested by Romero et al. .
We next wondered if the integrity of ribonucleic acids in cancer tissues parallels that of normal tissues. To address this question, we first analyzed RNA quality in 52 primary tumor tissues and 32 metastatic tumor tissues from the liver and lung. When stratified by PMI Category there were 15 primary tumors and 10 metastases in PMI category I, 18 primary tumors and 10 metastases in PMI category II, nine primary tumors and six metastases in PMI category III, and 10 primary tumors and six metastases in PMI category IV (Table ). The mean RIN value in primary tumors was 5.16 ± 2.4, and for liver and lung metastases was 5.07 ± 2.51and 6.29 ± 2.72, respectively (Table ). There was no correlation between RIN values and PMI in primary tumor tissues (Figure ). Forty-three of the 52 primary tumors (83%) analyzed were pancreatic ductal adenocarcinomas (PDA) (Figure ), providing an opportunity to compare the RIN values in primary PDAs specifically to that of normal pancreatic tissues. The mean RNA integrity in PDA tissues was not significantly different from that of normal pancreas tissues when considering all samples (mean RIN 5.4 ± 2.4 versus 5.26 ± 2.56 respectively, p = NS), or when limiting the comparison to 17 matched pairs of normal pancreas and primary PDA (mean 6.16 ± 1.96 vs 5.17 ± 2.35 respectively, p = NS). This finding thus does not support the long-held “myth” that PDA tissues have worse quality than other tumor types. This may be partially explained by observations that PDA is characterized by a prominent desmoplastic/stromal reaction that is hypovascular compared to adjacent normal pancreas . Nonetheless, we have found that screening multiple geographically distinct samples from different regions of the same neoplasm may be necessary to identify regions with preserved RNA quality, as we have recently found in ongoing work in our laboratory (Figure ). Finally, we next explored the relationships of metastatic tumor RNA quality between matched liver and lung metastases, i.e. from the same patient. There was no statistically significant correlation (Figure ), indicating that RNA quality is highly variable among metastases, even within the same patient. Thus, unlike normal tissues that show fairly predictable and tissue-specific degradation in relation to PMI, cancer tissues derived from different organ sites appear less predictable with respect to RNA quality than that of normal tissues.
While generally more stable than RNA, DNA is also subject to degradation in the postmortem period . Most methods to assess DNA degradation depend on examining selected target(s) to represent the overall sample quality with PCR-based methods among the most popular approach for this purpose . However, the extent to which such methods are reproducible or subject to inter-experiment variation is unknown, as is the extent to which genomic DNA in postmortem tissues follows similar kinetics as RNA. With these factors in mind we developed a semi-quantitative method to evaluate DNA quality to facilitate comparison among samples from independent experiments. In addition, unlike most studies detecting one locus, we simultaneously examined five chromosome loci of varying potential stability and susceptibility to DNA damage-inducing factors thereby achieving high sensitivity in detecting DNA damage in well-preserved samples (Table ). DNA was extracted from 36 frozen autopsy samples that were collected from five patients in PMI Category I and five in PMI category IV (Table ). To facilitate comparisons between RNA and DNA quality, samples with a wide spectrum of RIN values were selected from each PMI category, ranging from as low as 2.3 to as high as 9.4. These included samples from normal liver ( n = 7), normal kidney ( n = 9), primary tumors ( n = 10), liver metastases ( n = 5) and lung ( n = 5) metastases. We successfully extracted genomic DNA from all 36 samples including one that failed in RNA extraction. No DNA damage was detected by our assay in 32/36 samples analyzed (89%) even though 17 of these 32 (53%) had RIN values less than 5. In the four samples with DNA damage three were from PMI category IV, two of which showed degradation at all four sensor loci (Figure ). The remaining two samples showed only moderate damage as reflected by only two of the four chromosome loci affected. All four samples with DNA damage had concurrent RNA degradation (Table ). Interestingly all four samples with DNA damage were from the liver, three of which were histologically normal, further supporting the greater susceptibility of vascularized tissues to postmortem degradation. All other samples except for these four patients did not show DNA damage, including all 10 primary tumors analyzed. Thus, while DNA damage in postmortem tissues may be an indicator of RNA quality the converse is not true. Moreover, in addition to PMI the kinetics of postmortem DNA degradation may also have tissue type-specific factors.
Genomic DNA from postmortem tissues has been used successfully for next generation sequencing despite potential DNA degradation . However, given that RNA is more sensitive than DNA during the postmortem interval, its utility in downstream sequencing applications is unknown. As a proof of principle study, we therefore performed RNA sequencing on five matched pairs of normal pancreas and pancreatic cancer tissues, all with an RIN value of 5 or greater. Sequencing libraries were successfully generated from all samples using the poly-A enrichment method and used to generate 80 million reads per sample. Moreover, quality metrics of each library showed a sound distribution of coverage along transcripts and fragment lengths irrespective of RIN values (Figure ). One normal sample (patient A105), while showing good quality sequencing data, was excluded because the resulting data indicated contamination by cancer cells. This was confirmed histologically. A heat map of the top 250 genes differentially expressed between five pancreatic cancers and the remaining four normal pancreata showed a pattern that clearly discriminated the two groups (Figure ). Genes transcripts overrepresented in normal tissues included PRSS1 (cationic trypsinogen), CPA2 (pancreatic specific carboxypeptidase), AMY2A (pancreatic amylase 2), and the pancreatic developmental transcription factor PTF1A consistent with the greater abundance of acinar cells or cells with stem-like properties within these samples . Cancer samples showed greater heterogeneity with respect to the most differentially expressed genes. Genes overrepresented in the cancer samples included PSCA , MMP3 , MMP11 and SOX2 . The small sample size otherwise precluded a more thorough classification of each carcinoma's subtype as recently described . Finally, we leveraged our sequencing coverage to identify potential fusion events. We identified two recurrent fusion events, AXGP1-GJC3 and SIDT2-TAGLN, in six of nine postmortem RNA samples (three normal and three tumor, including two normal-tumor matched pairs). These two fusions were recently reported in normal pancreatic tissues within the context of a larger pan-tissue analysis . Collectively, these data are encouraging and suggest that despite being collected under postmortem conditions RNA samples can provide biologically meaningful information in downstream analyses. These findings are particularly exciting considering recent reports of improved methods to directly assess mRNA integrity and control for it in analyses of RNAseq data . While we did not utilize these methods for our pilot comparison, we nonetheless were able to discern gene expression signatures of known biologic importance in normal and pancreatic cancer tissues. With larger sample sizes and use of these methods it is reasonable to expect novel observations to be made in cancer tissues, for example with respect to treatment resistance or subclonal evolution. It is also important to note that our data do not indicate the RIN threshold value below which RNA sequencing cannot be performed, and in light of the pilot data shown it may be worth studying samples with low RIN values as well particularly as 39% of our samples had an RIN value of < 5. Previous evaluations of the quality of biomolecules in postmortem tissues have been in the context of brain banking, forensic analyses or minimally invasive autopsies . However, such an analysis using tissues obtained from a research autopsy program for cancer research has not been performed. Such information is critical to know with the growing interest in creation of biobanks from postmortem tissues of cancer patients, and use of these materials for ongoing scientific discovery and collaborations. While the scope of our dataset parallels that of another recent study , it differs in that we also studied DNA integrity as well and cancer tissues derived from different primary or metastatic sites. Nonetheless, our data is consistent with these prior studies that indicate nucleic acid quality of normal tissues is affected by a considerable number of factors in addition to the postmortem interval. At least one of these is likely the cause of death as we noticed several patients with a short PMI (< 3 hours) with exceedingly poor RNA quality, all of whom died of sepsis (personal observations, C.I.D.). Despite an increasing demand for research autopsy samples in the cancer research community, collecting high quality tissue samples is challenging due to numerous factors that can be legal, ethnic, emotional or social in nature. Exclusion of samples with a degree of poor quality is not always realistic, particularly when not all downstream applications are equally sensitive to sample quality. Thus, analyses that can incorporate these variables would be expected to improve comparisons across patients, tissue types and sample sets. Table summarizes these variables and their effects on RNA quality. Our hope is that these findings provide insight on the sample variability expected from research autopsy resources and ultimately facilitate data interpretation.
Tissues Autopsy samples were collected in association with the Johns Hopkins Gastrointestinal Cancer Rapid Medical Donation Program (GICRMDP) or the Memorial Sloan Kettering Cancer Center Medical Donation Program (MDP). Both programs were approved by the IRB at their respective institutions and in accordance with an assurance filed with and approved by the U.S. Department of Health and Human Services. Details of the program have been described previously . Briefly, the tissue harvesting protocol consists of opening of the body cavity using standard techniques and sterile sampling of a variety of normal tissues, the primary tumor if present and each grossly identified metastasis using a sterile blade and forceps. For snap-freezing, tissues were collected in 1.7 ml cryovials and immediately placed in liquid nitrogen before transferring to −80°C for long-term storage. Information regarding patient characteristics were recorded including the postmortem interval (PMI), defined as the time from death to the time of first incision. In all instances the time from the start to end of tissue sampling was ≤ 2 hours. RNA extractions For each sample, approximately 30 mg tissue was carefully harvested on ice and homogenized using the Fastprep-24 system with ceramic beads (MP Biomedicals). Total RNA was then extracted using RNeasy mini kits or Fibrous tissue mini kits (Qiagen) following the manufacturer's instruction. The RNA extraction step was carefully monitored by simultaneously extracting RNA from freshly sacrificed snap frozen mouse tissues of the same tissue type as a positive control. Total RNA was eluted in DNase/RNase free water. RNA was quantified using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific) and subsequently stored at −80°C. No more than 10 samples were extracted at one time. To confirm reproducibility of our extraction procedure, 10 samples were randomly selected and RNA re-extracted from the same tissue. In all cases similar yields were obtained from the first and second extraction. Any sample for which RNA could not be extracted was independently extracted at least one more time to rule out technical errors during the extraction procedure. RNA integrity analysis The RNA integrity number (RIN) was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies) with an RNA nano-kit system as described in the manufacturer's instructions. For each chip analysis, a commercially available tissue-matched control human RNA with high RNA quality (ZYAGEN) was run together with autopsy RNA samples to monitor the whole analysis procedure. Data reproducibility was confirmed by repeating the RNA chip analysis in 10 randomly selected samples at least two days apart from the first analysis. In all cases the results were highly reproducible with an overall difference in RIN value between the two chip assays of < 1. Genomic DNA extractions and PCR amplification Genomic DNA was extracted using DNeasy Mini Kits (Qiagen) according to the manufacturer's instructions. Genomic DNA was quantified by Qubit fluorometer (Invitrogen) and diluted to 20 ng/ul. 20 ng diluted DNA was subjected to PCR amplification in a total volume of 20 ul. PCR was carried out using a Taq PCR Core Kit (Qiagen). PCR conditions were 94°C for 2 min; 35 cycles of 94°C for 1 min, 60°C for 1 min and 72°C for 1 min 30 sec followed by 1 cycle of 72°C for 7 min. 3 ul of PCR products amplified from each locus were pooled in a new PCR tube and run on DNA screening gel cartridges (Cartridge ID C15D4D3A11) on a QIAxcel advanced system (Qiagen) according to manufacturer instructions. DNA quality analysis using two-step data normalization We adapted the qualitative multiplex PCR assay developed by Sigma-Aldrich to detect DNA damage in formalin-fixed paraffin embedded (FFPE) tissues ( http://www.sigmaaldrich.com/technical-documents/articles/life-science-innovations/qualitative-multiplex.html ) for analysis of snap frozen postmortem tissues. The original assay consists of five primer sets derived from the NCBI UniSTS database that amplify products ranging from 132 bp to 295 bp; some or all of these products will fail to amplify when DNA damage is present. PCR primers were modified to increase the amplicon sizes from 474 to 980 bp (Table ). One loci, a 196 bp amplicon, was used as an internal control to normalize PCR template input and amplification efficiency. The remaining four amplicons are located within known chromosome fragile sites that are relatively more susceptible to hydrolytic DNA damage and therefore serve as sensors of DNA quality . To avoid amplification bias that may be introduced during multiplex PCR each loci was amplified independently and then pooled for visualization and band quantification as described above. Commercially available human genomic DNA was used as an intact control and sonication fragmented DNA as damaged DNA control. Intact and fragmented DNA controls were analyzed in parallel with all human samples. Data were analyzed using a two-step data normalization to acquire a relative DNA quality of each sensor loci amplicon compared to the control amplicon. First, for each sample the band Intensity from each of the four sensor loci were normalized to that of the control locus. A standard band intensity was established from the intact control DNA by using mean values from three independent amplifications. Second, the relative band intensity from all samples was subsequently normalized to the standard. Based on this metric samples with perfect DNA quality have a value of 1 and samples with complete DNA degradation will have a value of 0. Values below the threshold of 0.6 were arbitrarily defined as having DNA damage. RNA sequencing Selected RNA samples from postmortem tissues with RIN value above 5 were used for RNA sequencing. RNA sequencing was performed in the MSK Genomics Core using the Illumina Truseq RNA sample prep protocol. Briefly, RNA sequencing libraries were generated with poly-A-enrichment method and sequencing was performed on an Illumina HiSeq2000 following standard protocols. Reads were paired-end 50 bp in length with a total of 80 millions of reads per sample. All sequenced libraries were mapped to the human genome (hg19) using rnaSTAR aligner . After mapping the expression count matrix was computed from the mapped reads using HTSeq ( www-huber.embl.de/users/anders/HTSeq ). The raw count matrix generated by HTSeq was then processed using the R/Bioconductor package DESeq ( www-huber.embl.de/users/anders/DESeq ) , which is used to both normalize counts and to identify differentially expressed genes between two conditions. A gene was declared differentially expressed if the fold-change was greater than 2 and the adjusted p-value was less than 0.05. Normalized counts were log2 transformed after addition of 1 to all values. Hierarchical clustering was performed using the R hclust function with the Euclidean distance measure on normalized log2 transformed counts after addition of 1 to all values. A heatmap was generated using the heatmap. 2 function from the gplots R package. The data plot shows the mean centered normalized log2 expression of the top 250 genes differentially expressed between tumor and normal tissues. To to detect fusion chimeras from RNA-seq data, meta-analysis that runs four fusion detection algorithms (ChimeraScan, FusionCatcher, MapSplice and DeFuse) was applied. The pipeline computes a meta-score for each detected fusion thus alleviating a problem of high numbers of false positives in each method taken independently. Statistics Statistical analysis was performed using GraphPad Prism Version 6.0 (GraphPad Software, Inc. La Jolla, CA). To determine the relationship between RIN and PMI, correlations were performed to determine the R 2 value and p value . Curve fits were added to scatterplots by performing linear regressions. Patient-matched tumor-metastasis comparisons were compared by a two-tailed Student t test. A p value of 0.05 or less was considered statistically significant.
Autopsy samples were collected in association with the Johns Hopkins Gastrointestinal Cancer Rapid Medical Donation Program (GICRMDP) or the Memorial Sloan Kettering Cancer Center Medical Donation Program (MDP). Both programs were approved by the IRB at their respective institutions and in accordance with an assurance filed with and approved by the U.S. Department of Health and Human Services. Details of the program have been described previously . Briefly, the tissue harvesting protocol consists of opening of the body cavity using standard techniques and sterile sampling of a variety of normal tissues, the primary tumor if present and each grossly identified metastasis using a sterile blade and forceps. For snap-freezing, tissues were collected in 1.7 ml cryovials and immediately placed in liquid nitrogen before transferring to −80°C for long-term storage. Information regarding patient characteristics were recorded including the postmortem interval (PMI), defined as the time from death to the time of first incision. In all instances the time from the start to end of tissue sampling was ≤ 2 hours.
For each sample, approximately 30 mg tissue was carefully harvested on ice and homogenized using the Fastprep-24 system with ceramic beads (MP Biomedicals). Total RNA was then extracted using RNeasy mini kits or Fibrous tissue mini kits (Qiagen) following the manufacturer's instruction. The RNA extraction step was carefully monitored by simultaneously extracting RNA from freshly sacrificed snap frozen mouse tissues of the same tissue type as a positive control. Total RNA was eluted in DNase/RNase free water. RNA was quantified using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific) and subsequently stored at −80°C. No more than 10 samples were extracted at one time. To confirm reproducibility of our extraction procedure, 10 samples were randomly selected and RNA re-extracted from the same tissue. In all cases similar yields were obtained from the first and second extraction. Any sample for which RNA could not be extracted was independently extracted at least one more time to rule out technical errors during the extraction procedure.
The RNA integrity number (RIN) was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies) with an RNA nano-kit system as described in the manufacturer's instructions. For each chip analysis, a commercially available tissue-matched control human RNA with high RNA quality (ZYAGEN) was run together with autopsy RNA samples to monitor the whole analysis procedure. Data reproducibility was confirmed by repeating the RNA chip analysis in 10 randomly selected samples at least two days apart from the first analysis. In all cases the results were highly reproducible with an overall difference in RIN value between the two chip assays of < 1.
Genomic DNA was extracted using DNeasy Mini Kits (Qiagen) according to the manufacturer's instructions. Genomic DNA was quantified by Qubit fluorometer (Invitrogen) and diluted to 20 ng/ul. 20 ng diluted DNA was subjected to PCR amplification in a total volume of 20 ul. PCR was carried out using a Taq PCR Core Kit (Qiagen). PCR conditions were 94°C for 2 min; 35 cycles of 94°C for 1 min, 60°C for 1 min and 72°C for 1 min 30 sec followed by 1 cycle of 72°C for 7 min. 3 ul of PCR products amplified from each locus were pooled in a new PCR tube and run on DNA screening gel cartridges (Cartridge ID C15D4D3A11) on a QIAxcel advanced system (Qiagen) according to manufacturer instructions.
We adapted the qualitative multiplex PCR assay developed by Sigma-Aldrich to detect DNA damage in formalin-fixed paraffin embedded (FFPE) tissues ( http://www.sigmaaldrich.com/technical-documents/articles/life-science-innovations/qualitative-multiplex.html ) for analysis of snap frozen postmortem tissues. The original assay consists of five primer sets derived from the NCBI UniSTS database that amplify products ranging from 132 bp to 295 bp; some or all of these products will fail to amplify when DNA damage is present. PCR primers were modified to increase the amplicon sizes from 474 to 980 bp (Table ). One loci, a 196 bp amplicon, was used as an internal control to normalize PCR template input and amplification efficiency. The remaining four amplicons are located within known chromosome fragile sites that are relatively more susceptible to hydrolytic DNA damage and therefore serve as sensors of DNA quality . To avoid amplification bias that may be introduced during multiplex PCR each loci was amplified independently and then pooled for visualization and band quantification as described above. Commercially available human genomic DNA was used as an intact control and sonication fragmented DNA as damaged DNA control. Intact and fragmented DNA controls were analyzed in parallel with all human samples. Data were analyzed using a two-step data normalization to acquire a relative DNA quality of each sensor loci amplicon compared to the control amplicon. First, for each sample the band Intensity from each of the four sensor loci were normalized to that of the control locus. A standard band intensity was established from the intact control DNA by using mean values from three independent amplifications. Second, the relative band intensity from all samples was subsequently normalized to the standard. Based on this metric samples with perfect DNA quality have a value of 1 and samples with complete DNA degradation will have a value of 0. Values below the threshold of 0.6 were arbitrarily defined as having DNA damage.
Selected RNA samples from postmortem tissues with RIN value above 5 were used for RNA sequencing. RNA sequencing was performed in the MSK Genomics Core using the Illumina Truseq RNA sample prep protocol. Briefly, RNA sequencing libraries were generated with poly-A-enrichment method and sequencing was performed on an Illumina HiSeq2000 following standard protocols. Reads were paired-end 50 bp in length with a total of 80 millions of reads per sample. All sequenced libraries were mapped to the human genome (hg19) using rnaSTAR aligner . After mapping the expression count matrix was computed from the mapped reads using HTSeq ( www-huber.embl.de/users/anders/HTSeq ). The raw count matrix generated by HTSeq was then processed using the R/Bioconductor package DESeq ( www-huber.embl.de/users/anders/DESeq ) , which is used to both normalize counts and to identify differentially expressed genes between two conditions. A gene was declared differentially expressed if the fold-change was greater than 2 and the adjusted p-value was less than 0.05. Normalized counts were log2 transformed after addition of 1 to all values. Hierarchical clustering was performed using the R hclust function with the Euclidean distance measure on normalized log2 transformed counts after addition of 1 to all values. A heatmap was generated using the heatmap. 2 function from the gplots R package. The data plot shows the mean centered normalized log2 expression of the top 250 genes differentially expressed between tumor and normal tissues. To to detect fusion chimeras from RNA-seq data, meta-analysis that runs four fusion detection algorithms (ChimeraScan, FusionCatcher, MapSplice and DeFuse) was applied. The pipeline computes a meta-score for each detected fusion thus alleviating a problem of high numbers of false positives in each method taken independently.
Statistical analysis was performed using GraphPad Prism Version 6.0 (GraphPad Software, Inc. La Jolla, CA). To determine the relationship between RIN and PMI, correlations were performed to determine the R 2 value and p value . Curve fits were added to scatterplots by performing linear regressions. Patient-matched tumor-metastasis comparisons were compared by a two-tailed Student t test. A p value of 0.05 or less was considered statistically significant.
|
Advancing
the Harmonization of Biopredictive Methodologies
through the Product Quality Research Institute (PQRI) Consortium:
Biopredictive Dissolution of Dipyridamole Tablets | c4342959-0d6b-412c-9082-98de35e73203 | 11468891 | Pharmacology[mh] | Understanding and visualizing the dissolution profiles of orally administered dosage forms in clinical and preclinical species has attracted great interest for formulation design and development, as well as selection in the pharmaceutical industry and in academia. For those reasons, biorelevant dissolution has been developed and advanced by the scientific community. − These biorelevant dissolution methodologies, which may also be biopredictive dissolution methodologies, incorporate key aspects of the human gastrointestinal physiology to evaluate the bioperformance of oral dosage forms, implementing quality by design (QbD) concepts to design and optimize the oral dosage forms. − Biorelevant dissolution experiments are still relatively new and have not been regulated, unlike compendial dissolution experiments created for the quality control of oral drug products, as found in, e.g., the United States Pharmacopoeia (USP). Individual researchers have developed biorelevant dissolution methodologies to better understand how formulations and compounds will perform in the body. As a result, those biorelevant dissolution profiles may look different among laboratories even if the same oral dosage forms are tested. The Product Quality Research Institute (PQRI), which is a nonprofit consortium of organizations that brings together members of the pharmaceutical industry, academia, and regulatory agencies to develop science-based approaches to regulation, has assembled an in vivo predictive dissolution and modeling working group (WG) to advance and harmonize in vivo predictive tools. The aims of this consortium are to address three questions: (1) can the PQRI working group (WG) members cross-validate their own experimental results by comparing dissolution profiles, identify the key experimental conditions, and move toward harmonizing their experimental methodologies; (2) will those generated dissolution profiles with those methodologies provide insightful information to guide in vivo studies; and (3) does the incorporation of the profiles into physiologically based biopharmaceutics modeling (PBBM) help to predict bioequivalence (BE), defined as predictions of the plasma profile by PBBM that lie within the 90% confidence interval for geometric mean ratio between 80 and 125% of the clinical result for both area under the curve (AUC) and maximum concentration ( C max ), of the in vivo data? The PQRI WG has five phases to achieve its goals, and the results of the first two phases (the first and second phases) out of five phases have already been published. Briefly, the PQRI WG studied the dissolution of ibuprofen (400 mg dose) and dipyridamole (50 mg dose) in the first two phases to understand if the WG members’ individual protocols for biorelevant dissolution methodologies would be able to produce consistent dissolution profiles of two model drugs, ibuprofen and dipyridamole. Precipitation in the media representing the upper small intestine at a 50 mg dose of dipyridamole, a weak base drug, was not observed by any of the WG members, regardless of dissolution methodology. This finding was attributed to the high p K a (p K a = 6.4) of dipyridamole together with the high media volume for the dissolution study at this low dose (∼400–500 mL). In those studies, the dissolution profiles of both ibuprofen and dipyridamole satisfied the BE criteria. On the basis of these studies, a more restrictive dissolution protocol was agreed upon by the WG. Since it is more challenging to obtain uniform biorelevant dissolution profiles if the test drug supersaturates and then precipitates in the GI tract, the dose was increased from 50 to 200 mg of dipyridamole. Although a 200 mg dose of dipyridamole is still a clinically relevant dose, this level of dipyridamole dose is only offered in an extended-release dosage form combined with aspirin. In the current studies, the WG members decided to use four immediate release (IR) tablets of 50 mg of dipyridamole to create a 200 mg dose. Biorelevant dissolution studies were conducted with 200 mg (50 mg × 4) of dipyridamole with or without the more restricted protocol to address the third and fourth objectives (Phase III and IV) of the overall five-phase project. Phase III consisted of using the higher dose of dipyridamole (200 mg) to generate biorelevant dissolution profiles with each WG member’s individual methodology and comparing results among WG members to determine which methodologies lead to BE and which methodologies lead to non-BE of the simulation with the clinical data. Phase IV consisted of using the higher dose of dipyridamole (200 mg) to generate biorelevant dissolution profiles with the more restricted protocols ( A,B), which were implemented on the basis of the Phase III results. These profiles were incorporated into the PBBM modeling software and assessed for BE or non-BE with the clinical data. Throughout this consortium project, all researchers performed biorelevant dissolution (with or without restrictions on methodology) on the same oral drug products and used the same batches of those oral products to eliminate any potential effects of the oral drug product source on the results. This overall exercise is intended to improve the quality and consistency of biorelevant dissolution and lead to harmonization of biorelevant dissolution methodologies. Eventually, the successful establishment of harmonized biorelevant dissolution is expected to improve oral product development, reduce animal studies, and, as a result, increase the success rate of clinical studies. A single batch of 50 mg dipyridamole tablets (lot no. 200203A, Rising Pharmaceuticals, East Brunswick, NJ, USA) were purchased and distributed to all members of the PQRI WG. For the preparation of biorelevant media, FaSSGF/FeSSIF/FaSSIF were purchased in powder form from Biorelevant.com (Biorelevant.com, London, UK) and prepared by each WG member before the biopredictive dissolution study. All other chemicals were analytical grade or HPLC grade. The WG member’s own methodologies for two-stage dissolution and/or transfer testing, which were based on their individual choices of experimental conditions and buffer media in Phase III, corresponded to the methods used by those members in Phase II. In Phase IV, all WG members switched over to the more restricted protocol for two-stage and transfer dissolution methodologies. All experimental conditions and methods are summarized in and , while historical changes in the experimental conditions are captured in . The dissolution profiles obtained by the WG were coupled with in silico modeling using GastroPlus version 9.8 (SimulationPlus, Inc., Lancaster, CA) to produce human plasma profiles. The simulated profiles were compared with clinical data to determine BE or non-BE to evaluate which of the dissolution methodologies were able to predict in vivo dissolution profiles , The dissolution methodologies conducted by each of the WG members in the Phase III studies are summarized in . The more restricted protocol for dissolution testing for dipyridamole tablets on Phase IV is summarized in . A more detailed description of the experimental condition presented in has been reported previously. The oral drug absorption of dipyridamole was computed on the basis of the physicochemical, pharmacokinetic, and drug dissolution properties of dipyridamole, which is weakly basic and classified as a Biopharmaceutics Classification System (BCS) class IIb drug, according to the simulation conditions proposed in the literature. − Single simulations were performed with the biorelevant dissolution profile from each WG member to predict the pharmacokinetic (PK) profile of 200 mg (50 mg × 4) dipyridamole IR tablets under fasted-state conditions. General input parameters for in silico simulation of dipyridamole were obtained from the literature and are summarized in . , , , The biopredictive dissolution profiles were incorporated into the in silico model as “controlled release” profiles. This selection prevents the in silico software from dictating the drug release profile of dipyridamole according to the physicochemical properties of the drug and the physiological settings in the software. which may not be able to display the supersaturation. Drug absorption from the stomach was assumed to be negligible in this set of predictions. The duration of the simulations was 24 h. The clinical data are regenerated and are displayed in and to portray the variability. In this modeling exercise, the predicted plasma profile based on the dissolution profile was compared to the BE range based on the average plasma data. This allowed us to evaluate how close the biorelevant dissolution profile is to the in vivo dissolution profile. However, note that the purpose of this simulation was not to provide a full prediction of the observed clinical plasma profile but rather to test the criticality of the biorelevant dissolution profile as a variable input parameter. All predictions were performed using the GastroPlus standard physiological conditions: Human Physiological-Fasted and Opt LogD Model SA/V 6.1. The predicted plasma profiles were compared with clinical trial results for dipyridamole pharmacokinetics to evaluate whether incorporating the results from the various dissolution studies into GastroPlus was able to predict the in vivo performance. , If the predicted plasma profile by PBBM based on the biorelevant dissolution profile fell within the 80–125% of clinical result for AUC and maximum concentration ( C max ), then the profile was considered to be BE. Thus, that biorelevant dissolution profile successfully predicted the drug dissolution in vivo . Phase III: Dipyridamole Dissolution Using WG Members’ Own Dissolution Methodologies The dissolution profiles of dipyridamole using (A) two-stage methodology and (B) transfer methodology are summarized in . In two-stage dissolution methodologies, four institutes reached almost complete dissolution of dipyridamole at the end of the first stage, while Institute C observed incomplete dissolution in the first stage regardless of buffer species and capacity ( A). This incomplete dissolution (∼10% dissolution) was caused by insufficient mixing at the low buffer volume (20 mL) made available in the vessel. In transfer methodologies, since drug absorption in the stomach is generally insignificant, only the drug dissolution profiles under the small intestinal stage are plotted ( B). The results from Institutes D, F, and G exhibited less than 40% dissolution of dipyridamole in the intestinal stage, while the result of Institute E exhibited ∼70% dipyridamole dissolution at the last time point. The higher % dissolved reported by Institute E may be traced back to two sources: (1) there was no transfer of gastric media in the first 30 min of experiment so the majority of dipyridamole would be dissolving when the transfer was started, and (2) the transfer rate was slower than others ( t 1/2 = 74 min) ( B). All results eventually exhibited ∼20% drug dissolution in the small intestinal stage, except for the result of Institute C where the drug did not dissolve. Since the p K a of dipyridamole is close to the pH of the dissolution buffer (pH 5.8–6.8), the individual choice of the buffer capacity, species, and/or volume could reasonably be expected to affect its dissolution. Phase IV-1: Dipyridamole Dissolution with the More Restricted Protocol—Part 1 The dissolution profiles of dipyridamole obtained with the more restricted protocol described in A are summarized in with two-stage methodology results in A and transfer methodology results in B. Using two-stage dissolution methodologies, all institutes displayed almost complete dissolution of dipyridamole at the end of 30 min in the gastric stage and excellent uniformity of the dissolution profiles. In transfer dissolution testing, only drug dissolution profiles under the small intestinal region are plotted ( B). The results of Institutes E and H initially exhibited similar dissolution rates of dipyridamole, while the results of Institutes D, F, and G exhibited slower dipyridamole dissolution in the intestinal stage. All transfer dissolution profiles reached ∼20% at the end of the experiment except Institute H. Different from other WG members, Institute H adopted a two-vessel transfer methodology with a constant volume in the second vessel. When the gastric content was transferred to the intestinal stage, the same volume was discarded from the small intestinal stage to maintain the media volume in the second vessel. Since the concentration of dipyridamole was measured in the constant volume of the intestinal stage, the full dissolution profile, in that experiment, was calculated and regenerated on the basis of the concentration in the second vessel, as well as in the discarded media volume at the given time. Therefore, the result did not concur with the decline in concentration reported by Institutes D, E, F, and G. Phase IV-2: Dipyridamole Dissolution with the More Restricted Protocol—Part 2 The dissolution profiles of dipyridamole obtained with the more restricted protocol described in B are summarized in with results for the two-stage methodology in A and transfer methodology in B. In two-stage dissolution testing, all four institutes displayed complete dissolution of dipyridamole in the gastric stage and excellent uniformity in the entire dissolution profiles ( A). In transfer dissolution testing, only drug dissolution profiles in the small intestinal region are plotted ( B). The dissolution results of Institutes E and H exhibited a faster initial dissolution rate in the intestinal stage than those of Institutes D, F, and G. As in Phase IV-1, Institute H implemented a constant media volume in the second vessel (intestinal stage) by discarding extra volume. In this experiment, the full dissolution profile was calculated and regenerated on the basis of the concentration in the second vessel, as well as in the discarded media volume at the given time. As a result, the dissolution profile of Institute H might be an overestimate . Dipyridamole Modeling The plasma profiles of dipyridamole were simulated on the basis of the biorelevant dissolution profiles produced by the WG. The purpose of this simulation was not to provide a fully accurate prediction of the observed clinical plasma profile but rather to directionally assess the criticality of differences in in vitro conditions, dissolution apparatus, and dissolution methodologies. The incorporation of biorelevant dissolution profiles must be optimized to correctly predict clinical plasma profiles. Needless to say, the simulations provided the direction and rank order of the criticality of those differences in the experimental conditions. The prediction of dipyridamole absorption at a dose of 200 mg from the two-stage biorelevant dissolution profiles all exhibited non-BE ( A, A, and A). As dissolution of 200 mg of dipyridamole was complete in the first 30 min in two-stage dissolution, the in silico modeling displayed much higher C max values than clinical data ( A, A, and A). The simulation results generated with two-stage dissolution testing thus overestimated the predicted plasma profile. In transfer dissolution testing, only Institute D obtained a biorelevant dissolution profile that led to BE in terms of both C max and AUC using its own method in , A, , and and and , while three Institutes produced biorelevant dissolution profiles that satisfied BE in either C max or AUC, and Institute E narrowly missed BE when method IV-2 was applied ( B, , and and ). As dipyridamole has a high p K a (p K a 6.4) value and all biorelevant dissolution requires an aqueous volume of 400–750 mL at the end of the experiment, this leads to less precipitation and, as a result, overestimation of oral absorption. The slight reduction in gastric acidity and buffer volumes from Phase III to Phase IV-2 studies likely explains the improvement in the modeling results. Since the gastric conditions are important for dipyridamole dissolution, an acidic pH, as well as adequate dissolution time and hydrodynamics, are necessary for the tablets to disintegrate and dissolve. Thus, dissolution methodologies should incorporate physiologically relevant gastric emptying times and complete transfer to the small intestinal environment to achieve more meaningful predictions. The experimental conditions, such as buffer pH, buffer species/capacity, volume, and stirring speed, as well as paddle and vessel sizes, were all shown to affect the dipyridamole dissolution profiles. The dissolution profiles of dipyridamole using (A) two-stage methodology and (B) transfer methodology are summarized in . In two-stage dissolution methodologies, four institutes reached almost complete dissolution of dipyridamole at the end of the first stage, while Institute C observed incomplete dissolution in the first stage regardless of buffer species and capacity ( A). This incomplete dissolution (∼10% dissolution) was caused by insufficient mixing at the low buffer volume (20 mL) made available in the vessel. In transfer methodologies, since drug absorption in the stomach is generally insignificant, only the drug dissolution profiles under the small intestinal stage are plotted ( B). The results from Institutes D, F, and G exhibited less than 40% dissolution of dipyridamole in the intestinal stage, while the result of Institute E exhibited ∼70% dipyridamole dissolution at the last time point. The higher % dissolved reported by Institute E may be traced back to two sources: (1) there was no transfer of gastric media in the first 30 min of experiment so the majority of dipyridamole would be dissolving when the transfer was started, and (2) the transfer rate was slower than others ( t 1/2 = 74 min) ( B). All results eventually exhibited ∼20% drug dissolution in the small intestinal stage, except for the result of Institute C where the drug did not dissolve. Since the p K a of dipyridamole is close to the pH of the dissolution buffer (pH 5.8–6.8), the individual choice of the buffer capacity, species, and/or volume could reasonably be expected to affect its dissolution. The dissolution profiles of dipyridamole obtained with the more restricted protocol described in A are summarized in with two-stage methodology results in A and transfer methodology results in B. Using two-stage dissolution methodologies, all institutes displayed almost complete dissolution of dipyridamole at the end of 30 min in the gastric stage and excellent uniformity of the dissolution profiles. In transfer dissolution testing, only drug dissolution profiles under the small intestinal region are plotted ( B). The results of Institutes E and H initially exhibited similar dissolution rates of dipyridamole, while the results of Institutes D, F, and G exhibited slower dipyridamole dissolution in the intestinal stage. All transfer dissolution profiles reached ∼20% at the end of the experiment except Institute H. Different from other WG members, Institute H adopted a two-vessel transfer methodology with a constant volume in the second vessel. When the gastric content was transferred to the intestinal stage, the same volume was discarded from the small intestinal stage to maintain the media volume in the second vessel. Since the concentration of dipyridamole was measured in the constant volume of the intestinal stage, the full dissolution profile, in that experiment, was calculated and regenerated on the basis of the concentration in the second vessel, as well as in the discarded media volume at the given time. Therefore, the result did not concur with the decline in concentration reported by Institutes D, E, F, and G. The dissolution profiles of dipyridamole obtained with the more restricted protocol described in B are summarized in with results for the two-stage methodology in A and transfer methodology in B. In two-stage dissolution testing, all four institutes displayed complete dissolution of dipyridamole in the gastric stage and excellent uniformity in the entire dissolution profiles ( A). In transfer dissolution testing, only drug dissolution profiles in the small intestinal region are plotted ( B). The dissolution results of Institutes E and H exhibited a faster initial dissolution rate in the intestinal stage than those of Institutes D, F, and G. As in Phase IV-1, Institute H implemented a constant media volume in the second vessel (intestinal stage) by discarding extra volume. In this experiment, the full dissolution profile was calculated and regenerated on the basis of the concentration in the second vessel, as well as in the discarded media volume at the given time. As a result, the dissolution profile of Institute H might be an overestimate . The plasma profiles of dipyridamole were simulated on the basis of the biorelevant dissolution profiles produced by the WG. The purpose of this simulation was not to provide a fully accurate prediction of the observed clinical plasma profile but rather to directionally assess the criticality of differences in in vitro conditions, dissolution apparatus, and dissolution methodologies. The incorporation of biorelevant dissolution profiles must be optimized to correctly predict clinical plasma profiles. Needless to say, the simulations provided the direction and rank order of the criticality of those differences in the experimental conditions. The prediction of dipyridamole absorption at a dose of 200 mg from the two-stage biorelevant dissolution profiles all exhibited non-BE ( A, A, and A). As dissolution of 200 mg of dipyridamole was complete in the first 30 min in two-stage dissolution, the in silico modeling displayed much higher C max values than clinical data ( A, A, and A). The simulation results generated with two-stage dissolution testing thus overestimated the predicted plasma profile. In transfer dissolution testing, only Institute D obtained a biorelevant dissolution profile that led to BE in terms of both C max and AUC using its own method in , A, , and and and , while three Institutes produced biorelevant dissolution profiles that satisfied BE in either C max or AUC, and Institute E narrowly missed BE when method IV-2 was applied ( B, , and and ). As dipyridamole has a high p K a (p K a 6.4) value and all biorelevant dissolution requires an aqueous volume of 400–750 mL at the end of the experiment, this leads to less precipitation and, as a result, overestimation of oral absorption. The slight reduction in gastric acidity and buffer volumes from Phase III to Phase IV-2 studies likely explains the improvement in the modeling results. Since the gastric conditions are important for dipyridamole dissolution, an acidic pH, as well as adequate dissolution time and hydrodynamics, are necessary for the tablets to disintegrate and dissolve. Thus, dissolution methodologies should incorporate physiologically relevant gastric emptying times and complete transfer to the small intestinal environment to achieve more meaningful predictions. The experimental conditions, such as buffer pH, buffer species/capacity, volume, and stirring speed, as well as paddle and vessel sizes, were all shown to affect the dipyridamole dissolution profiles. Researchers in academia and pharmaceutical companies have tried to understand how oral formulations would perform in the human GI tract so that oral absorption can be predicted and the oral formulation can be optimized. Better understanding of the bioperformance of oral dosage form would bring huge benefits prior to dosing in both preclinical and clinical pharmacokinetic studies. However, there is currently no guideline for biorelevant dissolution methodology and its regulation, in contrast to the compendial dissolution methodologies for, e.g., quality control. So far, each pharmaceutical company and academic institute has developed their own methodology to test oral compounds and products of interest and to predict their bioperformance. Multiple different methodologies have been proposed to predict the bioperformance of oral drug products. , , − These different methodologies can be divided into two major biorelevant dissolution types, two-stage dissolution and transfer methodologies. The two-stage dissolution has been a popular biorelevant dissolution methodology because of its relatively easy setup without any requirements for specific equipment and its ability to estimate the bioperformance of oral formulations. , , The test oral dosage form is exposed to two different pH environments in just one dissolution vessel in this methodology, with a concentrated solution of the intestinal phase added to the first phase (gastric phase) at a given time point to switch the conditions over to a composition representing the small intestine. The other biorelevant dissolution methodology is often referred to as a transfer methodology, which usually involves two vessels containing two different pH environments. Although it is based on the same principle as the two-stage dissolution methodology, mechanical transfer of the gastric phase medium into a vessel containing the small intestinal medium is performed to simulate the gastric emptying. This is intended to improve the evaluation of the bioperformance. , , , , Different dissolution media have been adopted in these dissolution methodologies to mimic the gastric conditions, e.g., for the gastric stage 0.01–0.1 N hydrochloric acid (HCl), simulated gastric fluid (SGF), and fasted-state simulated gastric fluid (FaSSGF). Less acidic buffers, such as maleate buffer, to represent achlorhydric conditions (pH 4.0 to 6.0) have also been proposed. Likewise, various concentrations and buffer species, like simulated intestinal fluid (SIF), and variations on fasted-state simulated intestinal fluid (FaSSIF), within the pH range of 6.5 to 7.5 have been used in these biorelevant dissolution methodologies to mimic the small intestinal conditions. , − Those biorelevant methodologies have been mainly used to investigate the bioperformance of model drugs that have pH-dependent solubility. As seen in , the empirical conditions of biorelevant methodologies by the PQRI WG members varied substantially, even though those methodologies have core similarities that reflect an understanding of the human GI physiology and exhibited similar precipitation outcome to the previous report. , As a result, biorelevant dissolution profiles generated by the WG members in Phase III were quite disparate , and only one methodology met the bioequivalent criteria . Going through the projects with the knowledge that the same lot/batch oral drug product had been used in all experiments, the WG was able to determine the important experimental factors in biorelevant dissolution methodologies and formulate an approach aimed at better regulating the range of experimental conditions. The PQRI WG worked through two specific protocols ( and ) to demonstrate how harmonization of biorelevant dissolution profiles , which exhibited the similar range of drug concentration of the previous reports, in vitro and in vivo , could lead to more institutes being able to satisfy the BE criteria when the dissolution profiles were incorporated into PBBM ( and ). , Generally, the harmonized two-stage dissolution methodology produced very uniform dissolution profiles but overestimated the drug dissolution in the first (gastric) stage, even in the more restricted protocols. Hence, the C max values generated in PBBM were also overestimated ( and – ). The initial dissolution of dipyridamole in the vessel representing small intestinal conditions in the transfer method was well controlled by PQRI WG when using the more restricted protocols ( and – ). However, the complete harmonization of the dissolution profiles, i.e., display of the same dissolution rate and precipitation rate among those dissolution profiles, was not achieved. This was attributed to the different hydrodynamics created by using different vessel sizes and paddle sizes ( and ). The differences in hydrodynamics should be studied in more detail to better understand their influence on supersaturation/precipitation profiles. In summary, the transfer methodology appears to be more promising for predicting plasma profiles. However, the WG needs to do more work to define the optimum methods and conditions more closely with respect to the equipment specifications. Additionally, the way that the dissolution profiles are entered into the simulation software needs further optimization. In this step-by-step project, the PQRI WG is successfully working toward harmonizing biopredictive dissolution tools and conditions and defining the optimal condition for meaningful in vivo prediction. |
Editorial: Circadian rhythm in cellular endocrinology | 68528229-1292-4153-9c88-183222f45693 | 11165192 | Internal Medicine[mh] | AT: Writing – original draft, Writing – review & editing. JR-N: Writing – original draft, Writing – review & editing. |
Suppression of pathogens in properly refrigerated raw milk | c7dda237-68ae-45e1-b12b-c342fb6d1c33 | 10715650 | Microbiology[mh] | Evidence exists that humans have consumed ruminant milks for millennia well before milk pasteurization became common in the 20 th century. In recent decades, unpasteurized (raw) milk has been legally available for direct human consumption in most US states and in many countries around the world . Evidence from Organizations including the Raw Milk Institute (RAWMI, Fresno, CA USA, https://www.rawmilkinstitute.org/ ) and the Raw Milk Producers Association (RMPA, Suffolk, UK, https://www.rawmilkproducers.co.uk/ ) advocate well-documented risk management procedures, including farmer training and mentoring, use of food safety plans similar to ‘farm-to-table’ or ‘grass-to-glass’ Hazard Analysis and Critical Control Points (HACCP) procedures, and stringent routine testing for bacterial indicators of potential contamination. Some US states license dairy farms and periodically monitor microbial indicators and pathogens in raw milk produced for direct human consumption, and one farm currently applies test-and-hold procedures for major bacterial pathogens before each lot of raw milk is bottled for consumers in California retail markets . Carefully produced hygienic raw milk for direct human consumption has become associated with health and rarely with foodborne disease outbreaks as documented in recent studies . Dietert and colleagues cited extensive evidence from monitoring programs of six countries reporting rare pathogen detection (<0.01%) in raw milk produced for human consumption, as distinguished from higher rates reported for pre-pasteurized milk of undetermined quality . The Dietert study also compiled evidence of health benefits for raw milk consumers, and no outbreaks of illness from 2018 to 2020 in CA when more than 1,352,000 gallons of fluid raw milk was provided in the CA retail market, consistent with a risk of illness less than 1 in over 20 million 250-mL servings for retail raw milk consumers. Notably, the US Food and Drug Administration (FDA) and the Food Safety and Inspection Service (FSIS) jointly reported that both raw and pasteurized milks were high risk foods for severe listeriosis , and a recent systematic review reported that severe listeriosis risks were significantly higher for pasteurized than raw milks . Literature on predictive microbiology of raw and pasteurized milks Raw milk producers, regulators, and consumers need reliable and statistically rigorous data to inform their decisions about the safety of raw milk for direct human consumption. Although researchers have understood for decades that rates of growth of pathogens inoculated into raw milk were slower than rates measured in pasteurized milk treated under the same conditions [ – ], this knowledge has not yet been integrated into risk analysis processes (risk assessment, risk communication, risk management) or policies about managing raw and pasteurized milks. Suppression of pathogen growth is attributed to bioactive components of milk, including competition with the microbes naturally present in raw milks, the milk microbiota. Strong evidence from both traditional culturing and culture-independent methodology has accumulated in this decade characterizing the natural milk microbiota [ – ] and its crucial role in balancing benefits and risks for human health . Recent studies [ – ] consider contributions of raw and pasteurized milks to the current epidemic of allergic, inflammatory, and non-communicable diseases that merit simultaneous considerations of benefits and risks attributable to both infectious and non-communicable diseases for assessing human health and well-being. Unfortunately, misinformation about raw milk and the interactions of its natural microbiota with potential pathogens and host cells abounds, even in the peer reviewed literature and government documents. The raw milk microbiota of mammals commonly includes lactic acid bacteria or LAB . Many LAB strains can outcompete pathogens by competing for nutrients as well as by active antagonism via bacteriocins and other microbial metabolites . Many of the diverse microbes classified as LABs are common members of the milk microbiota [ , , ], including many Generally Recognized as Safe (GRAS) or Qualified Presumption of Safety (QPS) with a safe history of use as probiotics (e.g., Bifidobacterium , Enterococcus , Lactobacillus , Lactococcus , Leuconostoc , Pediococcus , and Streptococcus ) that also appear to contribute to human and animal health [ , , , ]. Further, raw milk including the natural microbiota significantly suppressed adherence, invasion, and proliferation of a high dose of L . monocytogenes to human intestinal line cells versus administration of the pathogen in pasteurized milk or buffer . Greater understanding of the interactions of the raw milk microbiota with pathogens, both in our refrigerators and in the human gut, as well as their mechanisms of protection, is needed to appropriately model benefits and risks that raw milk microbes pose in complex ecosystems. Seven studies were identified in our literature searches ( ) that reported data on growth and survival of pathogens inoculated into raw milk and incubated at refrigeration temperatures. Of these, three studies [ , , ] monitored pathogen growth in both raw and pasteurized milks. All three studies documented either no pathogen growth at 4–5°C or slower growth in raw milk including the natural microbiota compared to pasteurized milk with greatly diminished microbial competitors. Despite documentation in the published literature of higher pathogen growth rates in pasteurized milks, FDA/FSIS assumed in its quantitative microbial risk assessment (QMRA) for listeriosis in Ready-to-Eat foods that growth of the pathogen L . monocytogenes was equivalent in raw and pasteurized milks. FDA/FSIS reported an ‘average’ growth rate of 0.257 hr -1 for milk in the body of the QMRA report, and documented pooling of the data of , 0.085 for raw milk and 0.407 for pasteurized milk adjusted to 5°C, in Appendix 8 of the QMRA report . Early predictive microbiology studies demonstrated the importance of time, temperature and the initial inoculation density of pathogens inoculated into sterile culture broth as the boundary for the growth/no-growth interface for the pathogen E . coli O157:H7 (~10°C) was approached . Clear dependencies were documented for initial pathogen density and temperature in broth culture studies and simulations of growth in non-sterile foods . However, non-sterile foods including raw milk are expected to impose additional limitations on pathogen growth and acceleration of pathogen decline due to the presence of a natural microbiota and other biologically active components including enzymes and bacteriocins that suppress pathogens . To document the mathematical relationships for pathogen growth and decline in raw milk for assessing and re-assessing microbial risks, a study design is needed that takes into account available knowledge on both temperature and pathogen contamination levels in naturally contaminated raw milk samples, as well as the dynamics of the microbial ecology of raw milk at recommended refrigeration temperatures. To address the current state of confusion about raw milk microbes and the mathematical relationships describing growth and decline of potential pathogens, RAWMI contracted with Food Safety Net Services, Ltd. (FSNS, San Antonio, TX USA) to conduct a pilot study in properly refrigerated raw milk. FSNS is an independent laboratory certified to quantify the major bacterial pathogens of concern in foods including raw milk ( Campylobacter , E . coli O157:H7, L . monocytogenes , and Salmonella ). This study provides evidence of microbial growth and decline from a small pilot study on inoculation of raw milk samples with enteropathogens and monitoring during storage for 14 days at 4.4°C (39.9°F), the refrigeration temperature recommended by regulatory agencies in the US. Results of the pilot study are further explored using statistical trend analysis and ANOVA as described herein.
Raw milk producers, regulators, and consumers need reliable and statistically rigorous data to inform their decisions about the safety of raw milk for direct human consumption. Although researchers have understood for decades that rates of growth of pathogens inoculated into raw milk were slower than rates measured in pasteurized milk treated under the same conditions [ – ], this knowledge has not yet been integrated into risk analysis processes (risk assessment, risk communication, risk management) or policies about managing raw and pasteurized milks. Suppression of pathogen growth is attributed to bioactive components of milk, including competition with the microbes naturally present in raw milks, the milk microbiota. Strong evidence from both traditional culturing and culture-independent methodology has accumulated in this decade characterizing the natural milk microbiota [ – ] and its crucial role in balancing benefits and risks for human health . Recent studies [ – ] consider contributions of raw and pasteurized milks to the current epidemic of allergic, inflammatory, and non-communicable diseases that merit simultaneous considerations of benefits and risks attributable to both infectious and non-communicable diseases for assessing human health and well-being. Unfortunately, misinformation about raw milk and the interactions of its natural microbiota with potential pathogens and host cells abounds, even in the peer reviewed literature and government documents. The raw milk microbiota of mammals commonly includes lactic acid bacteria or LAB . Many LAB strains can outcompete pathogens by competing for nutrients as well as by active antagonism via bacteriocins and other microbial metabolites . Many of the diverse microbes classified as LABs are common members of the milk microbiota [ , , ], including many Generally Recognized as Safe (GRAS) or Qualified Presumption of Safety (QPS) with a safe history of use as probiotics (e.g., Bifidobacterium , Enterococcus , Lactobacillus , Lactococcus , Leuconostoc , Pediococcus , and Streptococcus ) that also appear to contribute to human and animal health [ , , , ]. Further, raw milk including the natural microbiota significantly suppressed adherence, invasion, and proliferation of a high dose of L . monocytogenes to human intestinal line cells versus administration of the pathogen in pasteurized milk or buffer . Greater understanding of the interactions of the raw milk microbiota with pathogens, both in our refrigerators and in the human gut, as well as their mechanisms of protection, is needed to appropriately model benefits and risks that raw milk microbes pose in complex ecosystems. Seven studies were identified in our literature searches ( ) that reported data on growth and survival of pathogens inoculated into raw milk and incubated at refrigeration temperatures. Of these, three studies [ , , ] monitored pathogen growth in both raw and pasteurized milks. All three studies documented either no pathogen growth at 4–5°C or slower growth in raw milk including the natural microbiota compared to pasteurized milk with greatly diminished microbial competitors. Despite documentation in the published literature of higher pathogen growth rates in pasteurized milks, FDA/FSIS assumed in its quantitative microbial risk assessment (QMRA) for listeriosis in Ready-to-Eat foods that growth of the pathogen L . monocytogenes was equivalent in raw and pasteurized milks. FDA/FSIS reported an ‘average’ growth rate of 0.257 hr -1 for milk in the body of the QMRA report, and documented pooling of the data of , 0.085 for raw milk and 0.407 for pasteurized milk adjusted to 5°C, in Appendix 8 of the QMRA report . Early predictive microbiology studies demonstrated the importance of time, temperature and the initial inoculation density of pathogens inoculated into sterile culture broth as the boundary for the growth/no-growth interface for the pathogen E . coli O157:H7 (~10°C) was approached . Clear dependencies were documented for initial pathogen density and temperature in broth culture studies and simulations of growth in non-sterile foods . However, non-sterile foods including raw milk are expected to impose additional limitations on pathogen growth and acceleration of pathogen decline due to the presence of a natural microbiota and other biologically active components including enzymes and bacteriocins that suppress pathogens . To document the mathematical relationships for pathogen growth and decline in raw milk for assessing and re-assessing microbial risks, a study design is needed that takes into account available knowledge on both temperature and pathogen contamination levels in naturally contaminated raw milk samples, as well as the dynamics of the microbial ecology of raw milk at recommended refrigeration temperatures. To address the current state of confusion about raw milk microbes and the mathematical relationships describing growth and decline of potential pathogens, RAWMI contracted with Food Safety Net Services, Ltd. (FSNS, San Antonio, TX USA) to conduct a pilot study in properly refrigerated raw milk. FSNS is an independent laboratory certified to quantify the major bacterial pathogens of concern in foods including raw milk ( Campylobacter , E . coli O157:H7, L . monocytogenes , and Salmonella ). This study provides evidence of microbial growth and decline from a small pilot study on inoculation of raw milk samples with enteropathogens and monitoring during storage for 14 days at 4.4°C (39.9°F), the refrigeration temperature recommended by regulatory agencies in the US. Results of the pilot study are further explored using statistical trend analysis and ANOVA as described herein.
Microbiology methods for FSNS pilot study Full details on the methodology for the pilot study are provided in the FSNS report . Briefly, inocula were prepared as cocktails of three strains for each of 4 major foodborne bacterial pathogens ( Campylobacter jejuni/coli , E . coli O157:H7, L . monocytogenes , S . enterica serotypes Enteritidis/Seftenberg/ Typhimurium) as documented by Brandt (; see also Supplementary Materials). Campylobacter jejuni ATCC 33291 and 33560; C . coli ATCC 33559, E . coli O157:H7, ATCC 700599 and ATCC 43895, food isolates; ATCC 35150, human isolate L . monocytogenes , ATCC 19115, Serotype 4b and ATCC 7644, Serotype 1/2c, human isolates; ATCC 19114, Serotype 4a, animal isolate S . enterica serotypes Typhimurium ATCC 14028 and Seftenberg 775W ATCC 43845, food isolates; Enteritidis ATCC 49218) Duplicate samples of hygienic raw milk produced for direct human consumption by a RAWMI-listed dairy (Raw Farm, formerly Organic Pastures, Fresno, CA USA) were inoculated with one of two initial levels of each of the 4 pathogens (moderate levels ranging from 22 to 162 CFU/mL; high levels ranging from 600 to 8,300 CFU/mL). The inoculated raw milk samples were incubated at 4.4°C. Pathogens were quantified over time after inoculation (days 0, 3, 6, 9, 12, and 14) by standard culture-based methods. Aliquots of inoculated raw milk were spread plated on selective agar plates (Campy-Cefex, Xylose Lysine Deoxycholate, Modified Oxford, and Sorbitol MacConkey with Cefixime and Tellurite (CT-SMAC) for the enumeration of Campylobacter , S . enterica , L . monocytogenes , and E . coli O157:H7, respectively). Typical colonies were counted from each of the countable plates and recorded as colony forming units per mL (CFU/mL). The experiments were conducted in triplicate, producing a total of 48 time-series observations by pathogen and initial inoculation levels measured as CFU/mL over the 14 days of refrigerated storage. In addition to the enumeration results for pathogen growth and decline, the pilot study report also documents pH (range 6.3–7.1) and enumeration results for indicator organisms for milk quality (total aerobic plate counts (APC), total LAB, total coliforms, total yeasts, and molds (YM), and psychrotrophs) at days 0 and 14 for uninoculated raw milk samples. Statistical methods The Mann-Kendall Test, a nonparametric statistical test to detect a monotonic trend in time-series data, was performed to detect a statistically significant increasing or decreasing trend in 48 time-series observations generated by FSNS in the pilot study . The Mann-Kendall Test compares each data point to every successive measurement and determines if the change is positive or negative (the magnitude of change, or slope, is not considered). Each discordant pair is given a score of -1 and concordant pairs a score of +1. Tied values are given a score of 0. A test statistic (‘S’) is then computed based on the difference between the number of positive differences and negative differences. The sign of the S value indicates the overall direction of the data over time but must be compared to a critical value based on a 95% confidence level to accept or reject the null hypothesis of no trend (equal numbers of positive and negative differences). The Mann-Kendall Test was applied to each pathogen, lot of milk, and technical replicate for each of 48 individual time-series observations. Mann-Kendall calculations were performed in R version 4.1.1 using the Kendall package version 2.2.1 . Significance was assessed at (α = 0.05); p values <0.05 were considered significant. The effect of time (day of storage) on pathogen number (log 10 CFU/mL) within a genus and initial level was analyzed by one-way, analysis of variance (ANOVA) using Prism version 9.2 (GraphPad Software Inc., San Diego, CA). When ANOVA was significant (α = 0.05) for time, mean pathogen number among days of storage within a pathogen and initial level were compared using Tukey’s multiple comparison test at P < 0.05. Some samples of milk stored for 9 to 14 days at 4.4°C tested negative (0 CFU/mL) for Campylobacter . Because there is no log 10 value for zero, a default value of -0.01 log 10 CFU/mL was used for ANOVA. This default value was based on an accepted convention used in ComBase, an international microbial modeling database, for these types of data .
Full details on the methodology for the pilot study are provided in the FSNS report . Briefly, inocula were prepared as cocktails of three strains for each of 4 major foodborne bacterial pathogens ( Campylobacter jejuni/coli , E . coli O157:H7, L . monocytogenes , S . enterica serotypes Enteritidis/Seftenberg/ Typhimurium) as documented by Brandt (; see also Supplementary Materials). Campylobacter jejuni ATCC 33291 and 33560; C . coli ATCC 33559, E . coli O157:H7, ATCC 700599 and ATCC 43895, food isolates; ATCC 35150, human isolate L . monocytogenes , ATCC 19115, Serotype 4b and ATCC 7644, Serotype 1/2c, human isolates; ATCC 19114, Serotype 4a, animal isolate S . enterica serotypes Typhimurium ATCC 14028 and Seftenberg 775W ATCC 43845, food isolates; Enteritidis ATCC 49218) Duplicate samples of hygienic raw milk produced for direct human consumption by a RAWMI-listed dairy (Raw Farm, formerly Organic Pastures, Fresno, CA USA) were inoculated with one of two initial levels of each of the 4 pathogens (moderate levels ranging from 22 to 162 CFU/mL; high levels ranging from 600 to 8,300 CFU/mL). The inoculated raw milk samples were incubated at 4.4°C. Pathogens were quantified over time after inoculation (days 0, 3, 6, 9, 12, and 14) by standard culture-based methods. Aliquots of inoculated raw milk were spread plated on selective agar plates (Campy-Cefex, Xylose Lysine Deoxycholate, Modified Oxford, and Sorbitol MacConkey with Cefixime and Tellurite (CT-SMAC) for the enumeration of Campylobacter , S . enterica , L . monocytogenes , and E . coli O157:H7, respectively). Typical colonies were counted from each of the countable plates and recorded as colony forming units per mL (CFU/mL). The experiments were conducted in triplicate, producing a total of 48 time-series observations by pathogen and initial inoculation levels measured as CFU/mL over the 14 days of refrigerated storage. In addition to the enumeration results for pathogen growth and decline, the pilot study report also documents pH (range 6.3–7.1) and enumeration results for indicator organisms for milk quality (total aerobic plate counts (APC), total LAB, total coliforms, total yeasts, and molds (YM), and psychrotrophs) at days 0 and 14 for uninoculated raw milk samples.
The Mann-Kendall Test, a nonparametric statistical test to detect a monotonic trend in time-series data, was performed to detect a statistically significant increasing or decreasing trend in 48 time-series observations generated by FSNS in the pilot study . The Mann-Kendall Test compares each data point to every successive measurement and determines if the change is positive or negative (the magnitude of change, or slope, is not considered). Each discordant pair is given a score of -1 and concordant pairs a score of +1. Tied values are given a score of 0. A test statistic (‘S’) is then computed based on the difference between the number of positive differences and negative differences. The sign of the S value indicates the overall direction of the data over time but must be compared to a critical value based on a 95% confidence level to accept or reject the null hypothesis of no trend (equal numbers of positive and negative differences). The Mann-Kendall Test was applied to each pathogen, lot of milk, and technical replicate for each of 48 individual time-series observations. Mann-Kendall calculations were performed in R version 4.1.1 using the Kendall package version 2.2.1 . Significance was assessed at (α = 0.05); p values <0.05 were considered significant. The effect of time (day of storage) on pathogen number (log 10 CFU/mL) within a genus and initial level was analyzed by one-way, analysis of variance (ANOVA) using Prism version 9.2 (GraphPad Software Inc., San Diego, CA). When ANOVA was significant (α = 0.05) for time, mean pathogen number among days of storage within a pathogen and initial level were compared using Tukey’s multiple comparison test at P < 0.05. Some samples of milk stored for 9 to 14 days at 4.4°C tested negative (0 CFU/mL) for Campylobacter . Because there is no log 10 value for zero, a default value of -0.01 log 10 CFU/mL was used for ANOVA. This default value was based on an accepted convention used in ComBase, an international microbial modeling database, for these types of data .
FSNS pilot study The pilot study conducted by the contract laboratory FSNS demonstrated that Campylobacter spp., E . coli O157:H7, and Salmonella enterica spp. did not grow in raw milk at 4.4°C during 14 days of refrigerated storage. Similarly, L . monocytogenes did not grow at this temperature until day 9 for the lower initial inoculum (26 to 41 CFU/mL) and day 6 for the higher initial inoculum (3,000 to 7,900 CFU/mL), respectively. The range of initial counts of indicators for the pilot study conducted by FSNS (; full report provided in Supplemental Materials) are provided parenthetically: total aerobic plate counts (510 to 1,900); psychrotrophic plate counts (10–200,000); total coliforms (10–50), total lactic acid bacteria (70–470), and yeasts and molds (10–20). Although some indicators grew and some declined by day 14 , no statistical testing for correlations between indicators and pathogens were conducted for the pilot study. Mann-Kendall and ANOVA analysis Results of our statistical analyses of pathogen trends in the pilot study data are presented in and Figs and . Mann-Kendall statistics are presented in , plots of the 48 time-series observations by pathogen and initial inoculation level are depicted in , and results of ANOVA are presented in . During the first week of refrigerated storage, evidence of pathogen growth was not documented by ANOVA ( ). In the second week of monitoring, evidence of L . monocytogenes growth was documented by the Mann-Kendall Test for trend in 8 of 12 replicates (P = 0.004 to P = 0.043; ) and ANOVA ( ). No evidence of trend or significant evidence of decline was observed using the Mann-Kendall Test for Campylobacter , E . coli O157:H7, and Salmonella . Further, results from ANOVA ( ) indicated that the inoculated pathogen Campylobacter declined continuously in milk stored at 4.4°C until it was eliminated from a lower initial level (2.0 log 10 CFU/mL) after 9 days of storage ( ) or from a higher initial level (3.0 log 10 CFU/mL) after 12 days of storage ( ). In contrast, E . coli O157:H7 declined initially (days 0 to 3 of storage) but then survived (days 3 to 14 of storage) at the reduced level, resulting in a small to moderate (0.5 to 0.9 log 10 CFU/mL) but significant reduction of its initial lower ( ) or higher ( ) levels in milk. Similarly, Salmonella declined initially (day 0 to 3 of cold storage) and then survived at a reduced level ( ) or died slowly throughout refrigerated storage ( ) resulting in a small (0.2 to 0.6 log 10 CFU/mL) but significant reduction of its initial levels in milk. The inoculated pathogen L . monocytogenes survived initially (days 0 to 6) in milk stored at 4.4°C before it started to grow around day 9 of storage from a lower (1.5 log 10 CFU/mL) or higher (3.6 log 10 CFU/mL) initial level to a final level of 2.8 ( ) or 5.3 ( ) log 10 CFU/mL, respectively. In summary, the ANOVA results indicated that pathogen levels in milk stored at 4.4°C depended on initial level, pathogen genus, and time of storage. Importantly, prolonged storage (9 to 14 days) of milk at 4.4°C significantly reduced or eliminated lower and higher initial levels of Campylobacter and resulted in small to moderate (0.2 to 0.9 log 10 CFU/mL) but significant reductions in lower and higher initial levels of E . coli O157:H7 and Salmonella . However, prolonged storage resulted in significant increases (0.5 to 1.6 log 10 CFU/mL) in lower and higher initial levels of L . monocytogenes . Over the 14-day study period at 4.4°C, the average pH of uninoculated raw milk decreased (6.96 to 6.88 for one shipment; 7.1 to 6.4 for the other; see S3 Table in for detail on individual lots). The contract laboratory documented some variability in counts of microbial indicators for milk quality. For the indicator APC, results decreased after 14 days at 4.4°C for the first shipment (from 1,643 to 663 CFU/mL) and increased for the second (from 1,020 to 666,000,000 CFU/mL). Similarly, for total LABs, results increased for the first shipment (from 80 to 263) and decreased for the second (from 423 to 10 CFU/mL). For the remaining indicators, results increased for both shipments (total coliforms from 30 to 50 and from 10 to 1,007 CFU/mL; total YM from 10 to 81,667 and from 20 to 7,203 CFU/mL; and psychrotrophs (from 20 to too numerous to count (>57,000,000,000 CFU/mL and from 133,333,333 to >2,500,000,000 CFU/mL). Note that the pilot study was not designed to perform statistical testing for correlations between indicators and pathogens, nor for fitting parameter values for pathogen growth and decline curves. See for detail on counts of pathogens and indicators for individual lots.
The pilot study conducted by the contract laboratory FSNS demonstrated that Campylobacter spp., E . coli O157:H7, and Salmonella enterica spp. did not grow in raw milk at 4.4°C during 14 days of refrigerated storage. Similarly, L . monocytogenes did not grow at this temperature until day 9 for the lower initial inoculum (26 to 41 CFU/mL) and day 6 for the higher initial inoculum (3,000 to 7,900 CFU/mL), respectively. The range of initial counts of indicators for the pilot study conducted by FSNS (; full report provided in Supplemental Materials) are provided parenthetically: total aerobic plate counts (510 to 1,900); psychrotrophic plate counts (10–200,000); total coliforms (10–50), total lactic acid bacteria (70–470), and yeasts and molds (10–20). Although some indicators grew and some declined by day 14 , no statistical testing for correlations between indicators and pathogens were conducted for the pilot study.
Results of our statistical analyses of pathogen trends in the pilot study data are presented in and Figs and . Mann-Kendall statistics are presented in , plots of the 48 time-series observations by pathogen and initial inoculation level are depicted in , and results of ANOVA are presented in . During the first week of refrigerated storage, evidence of pathogen growth was not documented by ANOVA ( ). In the second week of monitoring, evidence of L . monocytogenes growth was documented by the Mann-Kendall Test for trend in 8 of 12 replicates (P = 0.004 to P = 0.043; ) and ANOVA ( ). No evidence of trend or significant evidence of decline was observed using the Mann-Kendall Test for Campylobacter , E . coli O157:H7, and Salmonella . Further, results from ANOVA ( ) indicated that the inoculated pathogen Campylobacter declined continuously in milk stored at 4.4°C until it was eliminated from a lower initial level (2.0 log 10 CFU/mL) after 9 days of storage ( ) or from a higher initial level (3.0 log 10 CFU/mL) after 12 days of storage ( ). In contrast, E . coli O157:H7 declined initially (days 0 to 3 of storage) but then survived (days 3 to 14 of storage) at the reduced level, resulting in a small to moderate (0.5 to 0.9 log 10 CFU/mL) but significant reduction of its initial lower ( ) or higher ( ) levels in milk. Similarly, Salmonella declined initially (day 0 to 3 of cold storage) and then survived at a reduced level ( ) or died slowly throughout refrigerated storage ( ) resulting in a small (0.2 to 0.6 log 10 CFU/mL) but significant reduction of its initial levels in milk. The inoculated pathogen L . monocytogenes survived initially (days 0 to 6) in milk stored at 4.4°C before it started to grow around day 9 of storage from a lower (1.5 log 10 CFU/mL) or higher (3.6 log 10 CFU/mL) initial level to a final level of 2.8 ( ) or 5.3 ( ) log 10 CFU/mL, respectively. In summary, the ANOVA results indicated that pathogen levels in milk stored at 4.4°C depended on initial level, pathogen genus, and time of storage. Importantly, prolonged storage (9 to 14 days) of milk at 4.4°C significantly reduced or eliminated lower and higher initial levels of Campylobacter and resulted in small to moderate (0.2 to 0.9 log 10 CFU/mL) but significant reductions in lower and higher initial levels of E . coli O157:H7 and Salmonella . However, prolonged storage resulted in significant increases (0.5 to 1.6 log 10 CFU/mL) in lower and higher initial levels of L . monocytogenes . Over the 14-day study period at 4.4°C, the average pH of uninoculated raw milk decreased (6.96 to 6.88 for one shipment; 7.1 to 6.4 for the other; see S3 Table in for detail on individual lots). The contract laboratory documented some variability in counts of microbial indicators for milk quality. For the indicator APC, results decreased after 14 days at 4.4°C for the first shipment (from 1,643 to 663 CFU/mL) and increased for the second (from 1,020 to 666,000,000 CFU/mL). Similarly, for total LABs, results increased for the first shipment (from 80 to 263) and decreased for the second (from 423 to 10 CFU/mL). For the remaining indicators, results increased for both shipments (total coliforms from 30 to 50 and from 10 to 1,007 CFU/mL; total YM from 10 to 81,667 and from 20 to 7,203 CFU/mL; and psychrotrophs (from 20 to too numerous to count (>57,000,000,000 CFU/mL and from 133,333,333 to >2,500,000,000 CFU/mL). Note that the pilot study was not designed to perform statistical testing for correlations between indicators and pathogens, nor for fitting parameter values for pathogen growth and decline curves. See for detail on counts of pathogens and indicators for individual lots.
This small pilot study was undertaken to measure pathogen counts in inoculated samples of raw milk produced for direct human consumption and stored at 4.4°C for two weeks, to estimate statistical trends for growth and decline, and to challenge misinformation about pathogen growth in raw milk complete with its natural microbiota. Data from the small pilot study was sufficient to estimate trends of pathogen decline in the first week and to conduct ANOVA, but insufficient to estimate parameters of growth and decline for the 48 time series curves for the inoculated pathogens ( and Figs and ). The pilot study data and trends are consistent and provide statistically significant results by both the Mann-Kendall Test and ANOVA. More research is needed to enable parameter estimations and deeper statistical characterization of pathogen growth and decline in raw and pasteurized milks for future QMRA simulations. Temperature, as well as competition with the natural microbiota, are widely recognized as key factors for controlling microbial growth in foods , also key for managing raw milk risks as pointed out by the European Food Safety Authority . Dairy farmers and retailers are trained to rapidly cool raw milk and continuously monitor refrigeration temperatures in chill tanks, trucks, and retail refrigeration cases. Consumers are advised to transport refrigerated foods with a cold pack in an insulated bag and keep their refrigerators set at 4.4°C, and deviations or noncompliance with recommendations can be represented in QMRA abuse scenarios. The major finding of the pilot study is statistical evidence of no growth at 4.4°C for the major foodborne pathogens causing illness associated with raw milk in the US ( Campylobacter , E . coli O157:H7, and Salmonella ; . For listeriosis, rarely associated with illness from raw milk, the pilot study documented evidence of pathogen growth in 8 of 12 replicates (P = 0.001 to P = 0.028, significant by ANOVA in the second week of refrigerated storage). An extensive body of evidence [ , , – ] documents both intrinsic factors (moisture content, pH, nutrient and micronutrient content, biological structure, redox potential, naturally occurring or added antimicrobials, and competitive microbiota) and extrinsic factors (packaging atmospheres, time and temperature effects, storage or holding conditions, and both thermal and non-thermal processing steps) that drive or suppress microbial growth in foods. These studies also document extensive evidence of synergy (or greater benefit) for multiple barriers to pathogen growth or ‘hurdles’ acting via different cellular mechanisms. Combinations of hurdles (e.g., pH, naturally occurring antimicrobials, refrigeration, and competitive microbiota) can prevent multiplication, inactivate, or kill pathogens in foods while maintaining nutrient content and improving stability, safety, and quality of foods . Suppression of pathogen growth in properly refrigerated raw milk demonstrated herein and in previous studies [ , , ] are consistent with multi-hurdle risk management. Need for reliable data to replace invalid assumptions for robust risk analysis The major limitations of the pilot study are that raw milk from a single US dairy was analyzed and time-series observations at only 3 time points were conducted in the first week, and a total of 6 time points over 14 days of refrigerated storage post-inoculation. However, another published study also documented time series including only 6 time points for refrigerated storage of raw milk. Further, these researchers inoculated extremely high levels of an enteropathogen (10 5 or 100,000 cfu/mL) despite detecting 35 or fewer pathogens per mL from naturally contaminated milk from the same dairy (range 0.007 to 35 MPN/mL; . It is unclear if the reported trends from the extremely high inoculated levels would be consistent with trends for raw milk samples inoculated at levels 4 or more orders of magnitude lower. In addition, Jaakkonen and colleagues did not obtain fresh raw milk from the producer for their study on survival trends, but reported purchasing raw milk in the retail market. It is uncertain if results reported herein and by Jaakkonen and colleagues are representative of other conditions, particularly due to documentation of high variability of the raw milk microbiota across herds, farms, breeds, diets, storage times and temperatures, and seasonality [ – ]. Further research is needed to quantitate pathogen growth and decline rates for raw milk inoculated at levels of pathogen contamination observed in fresh naturally contaminated samples in order to minimize bias and optimize the experimental design to reflect feasible ecological conditions for predictions in a complex food. Unbiased data are essential for assessing and re-assessing risks for raw and pasteurized milks from multiple dairy farms. This data gap for predictive microbiology of pathogens in raw and pasteurized milks is relevant to microbial risk assessment because two historic QMRAs appeared to select intentionally conservative assumptions. Both QMRAs appear subject to overestimation bias for raw milk risks. Neither QMRA included or discussed data demonstrating that pathogen growth is slower in raw milk than pasteurized milk, attributable in part to pathogen competition with the dense and diverse natural microbiota of milks. Despite characterizing both raw and pasteurized milks as high-risk foods for severe listeriosis, FDA/FSIS selected different risk communications and risk management policies that were not based on the scientific evidence. The risk management policies for prohibition and recommended avoidance, respectively, for raw milk in Australia and the US are inconsistent with both then available and current scientific evidence discussed herein, and in more detail by . Need for evaluation of QMRAs relative to international guidance and quality criteria The pilot study design was also motivated by the consensus statement on general principles and guidelines for QMRA ratified by 163 member countries of the Codex Alimentarius Commission (CAC) in 1999 . Of the 11 CAC principles, five are highly relevant to this study ( ). Considered together, these principles focus on sound and transparent processes, including use of the best available scientific evidence for modeling microbiology ecology, as well as reassessing and reevaluating over time as science advances. In addition, these principles acknowledge that risk assessors may choose to apply assumptions rather than scientific data when significant gaps in knowledge exist. In order to fully address these principles, risk practitioners relying on assumptions rather than objective scientific data must also characterize the implications of alternative assumptions and their impact on risk estimates and scenarios for risk management options, critical aspects of quality risk analysis, as articulated in the Risk Analysis Quality Test (RAQT) of the Society for Risk Analysis (SRA; , also available at https://www.sra.org/risk-analysis-specialty-groups/applied-risk-management/scientific-literature/ ). Both historic government QMRAs failed most or all of the questions for evaluating risk analysis quality that frame the 76-question battery of the RAQT (workshop manuscript in development through the SRA Applied Risk Management specialty group). Significant gaps in knowledge for raw milk QMRAs were raised as needs for future re-assessment in two historic QMRAs that examined raw milk , as well as in a more recent review that included two FDA contributors to the assessment . These gaps remain unfilled to date. The FDA/FSIS risk assessment team inappropriately assumed that growth rates for L . monocytogenes were equivalent for pasteurized and raw milks, despite data to the contrary. Some year later, the Food Standards Australia New Zealand team conducted a QMRA for raw cow milk that was largely based on unvalidated assumptions and extrapolations rather than reliable data for raw milk. Thus, both QMRAs imposed overestimation bias on their assessments for raw milk and did not fully disclose the impacts of their intentionally conservative assumptions on risk estimates, management options for risk reduction, or risk communications. Regarding guideline 10, the US Centers for Disease Control and Prevention provided a recent dataset for outbreaks from all transmission sources including both raw and pasteurized fluid milks for the period 2005 to 2020 . Raw and pasteurized milks both caused nearly 2,000 illnesses over this 16-year period (manuscript in preparation). Mortality rates associated with milks in North America in recent decades are quite low, including 5 US fatalities (3 associated with pasteurized milk and 2 with raw milk; , and 4 Canadian fatalities associated with pasteurized milk . Perhaps the most highly relevant general principle for QMRAs in this context is guideline 11. The findings of the pilot study, the lack of growth of the foodborne pathogens in raw milk for 14 days at 4.4°C for the major foodborne pathogens causing raw milk outbreaks in the US, are consistent with other peer reviewed studies conducted between 4 and 5°C [ , , ] that falsify the incorrect assumptions about pathogen growth in historic QMRAs. Two independent academic research teams re-evaluated and extended portions of the historic FDA/FSIS risk assessment for severe listeriosis . Latorre and colleagues estimated risks per raw milk serving to the general population were as low as 10 −15 (~1 illness per 1,000,000,000,000,000 servings), substantially lower estimated risks compared to the FDA/FSIS 2003 assessment that pooled growth data for raw and pasteurized milks. Stasiewicz and colleagues found in re-assessment that increasing heat treatments increased the growth rates of L . monocytogenes in pasteurized and ultra-pasteurized milks, consistent with killing more of the milk microbiota and thus reducing competition with the pathogen. These researchers also provided supplemental information for their study reporting no growth of the pathogen in raw milk blanks and increasing rates of growth in the raw milk pasteurized for 25 seconds at 72° and 82°C. The need to update incorrect assumptions and misinformation about both predictive microbiology and dose-response relationships in this QMRA was raised for future re-assessment for the FDA/FSIS QMRA . Need for transparency about scientific evidence falsifying prior assumptions The need to update incorrect assumptions about pathogen growth in raw milk that were made in historical QMRAs is more urgent than ever because so many claims about raw and pasteurized milks are made in the media, as well as in the scientific literature, without rigorous supporting data. Consumers and scientists can understandably be confused by conflicting claims. SRA leaders seek to encourage others to apply the SRA RAQT in both review of completed QMRAs for other foods and water, as well as in planning for future risk analysis projects, with the goal of developing a culture of full disclosure and quality analysis. Cultural, social, or ideological constructions have in the past limited the influence of scientific evidence into policy making, as documented by Meagher and colleagues on cultural mischaracterization of two foodborne outbreaks. Despite quick tracing of a 2006 outbreak to California-grown spinach, FDA’s public risk communication to avoid consuming any raw spinach contributed to market collapse, and “a range of plausible responses were never considered” (, pg. 245). Similarly, organizations around the world appear to incorrectly attribute high risk to raw milk from all producers, and not to any source of pasteurized milk. Participants in the SRA workshop on risk analysis quality discussed common unstated and unsupported assumptions about milks include: 1) the source of microbes in milk is feces; 2) raw milks are inherently dangerous; 3) pasteurization is a ‘silver bullet’; and 4) pasteurized milk is zero risk (manuscript in preparation). Current evidence documented herein and by Coleman and colleagues and Dietert and colleagues supports none of these assumptions. The importance of correctly modeling the microbial ecology of raw milk demonstrated by LAB strains isolated from raw bovine milk suppression or exclusion of three pathogens inoculated at two high densities, 10 3 and 10 6 log 10 CFU/mL . Clearly, the natural milk microbiota can suppress the growth of pathogens under some conditions. From our perspective of available data and analysis consistent with principles of microbial ecology and those of the , as well as the RAQT of the SRA, evidence that raw milk is ‘inherently dangerous’ is lacking. Evidence is consistent with protective multi-hurdle synergies of raw milk including the dense and diverse natural microbiota of mammalian milks under proper refrigeration contributing to suppression or exclusion of pathogens. We also acknowledge that no food is risk free, and benefits and risks could and perhaps should be characterized for all foods. Further improvements in the credibility and utility of QMRAs might develop with deeper consideration of environmental sustainability, economics and food waste, supply chain structure, climate change, and social and cultural factors [ – ]. For example, Duret and colleagues determined that setting the domestic refrigerator temperature to 4°C presented the best compromise for balancing risk of foodborne illness, food waste, and energy consumption. Rendueles and colleagues suggest not only that multi-hurdle approaches can reduce risk of illness and maintain food quality, but also can support more sustainable food production chains in the global market. Thus, for design of future predictive microbiology studies to inform risk analysis, studies must include additional production lots, dairy farms, and regions or states to characterize regional or national trends for milk risks. Expansions of the pilot study design should also include more frequent sampling and multiple initial inoculation levels for pathogens (at least ~1 CFU/mL and ~1,000 CFU/mL) so that robust parameters for growth and decline can be estimated. An ideal study design might also explore potential mechanisms of pathogen suppression under proper refrigeration and temperature abuse scenarios by quantitating key representatives of the raw milk microbiota over the study period and identifying microbial associations that drive pathogen suppression and killing. Rigorous quantitative data on predictive microbiology of raw milks is essential to re-evaluating historic QMRAs based on invalid assumptions about pathogen growth in raw milk so that unbiased estimates of risks and benefits can be generated for raw and pasteurized milks.
The major limitations of the pilot study are that raw milk from a single US dairy was analyzed and time-series observations at only 3 time points were conducted in the first week, and a total of 6 time points over 14 days of refrigerated storage post-inoculation. However, another published study also documented time series including only 6 time points for refrigerated storage of raw milk. Further, these researchers inoculated extremely high levels of an enteropathogen (10 5 or 100,000 cfu/mL) despite detecting 35 or fewer pathogens per mL from naturally contaminated milk from the same dairy (range 0.007 to 35 MPN/mL; . It is unclear if the reported trends from the extremely high inoculated levels would be consistent with trends for raw milk samples inoculated at levels 4 or more orders of magnitude lower. In addition, Jaakkonen and colleagues did not obtain fresh raw milk from the producer for their study on survival trends, but reported purchasing raw milk in the retail market. It is uncertain if results reported herein and by Jaakkonen and colleagues are representative of other conditions, particularly due to documentation of high variability of the raw milk microbiota across herds, farms, breeds, diets, storage times and temperatures, and seasonality [ – ]. Further research is needed to quantitate pathogen growth and decline rates for raw milk inoculated at levels of pathogen contamination observed in fresh naturally contaminated samples in order to minimize bias and optimize the experimental design to reflect feasible ecological conditions for predictions in a complex food. Unbiased data are essential for assessing and re-assessing risks for raw and pasteurized milks from multiple dairy farms. This data gap for predictive microbiology of pathogens in raw and pasteurized milks is relevant to microbial risk assessment because two historic QMRAs appeared to select intentionally conservative assumptions. Both QMRAs appear subject to overestimation bias for raw milk risks. Neither QMRA included or discussed data demonstrating that pathogen growth is slower in raw milk than pasteurized milk, attributable in part to pathogen competition with the dense and diverse natural microbiota of milks. Despite characterizing both raw and pasteurized milks as high-risk foods for severe listeriosis, FDA/FSIS selected different risk communications and risk management policies that were not based on the scientific evidence. The risk management policies for prohibition and recommended avoidance, respectively, for raw milk in Australia and the US are inconsistent with both then available and current scientific evidence discussed herein, and in more detail by .
The pilot study design was also motivated by the consensus statement on general principles and guidelines for QMRA ratified by 163 member countries of the Codex Alimentarius Commission (CAC) in 1999 . Of the 11 CAC principles, five are highly relevant to this study ( ). Considered together, these principles focus on sound and transparent processes, including use of the best available scientific evidence for modeling microbiology ecology, as well as reassessing and reevaluating over time as science advances. In addition, these principles acknowledge that risk assessors may choose to apply assumptions rather than scientific data when significant gaps in knowledge exist. In order to fully address these principles, risk practitioners relying on assumptions rather than objective scientific data must also characterize the implications of alternative assumptions and their impact on risk estimates and scenarios for risk management options, critical aspects of quality risk analysis, as articulated in the Risk Analysis Quality Test (RAQT) of the Society for Risk Analysis (SRA; , also available at https://www.sra.org/risk-analysis-specialty-groups/applied-risk-management/scientific-literature/ ). Both historic government QMRAs failed most or all of the questions for evaluating risk analysis quality that frame the 76-question battery of the RAQT (workshop manuscript in development through the SRA Applied Risk Management specialty group). Significant gaps in knowledge for raw milk QMRAs were raised as needs for future re-assessment in two historic QMRAs that examined raw milk , as well as in a more recent review that included two FDA contributors to the assessment . These gaps remain unfilled to date. The FDA/FSIS risk assessment team inappropriately assumed that growth rates for L . monocytogenes were equivalent for pasteurized and raw milks, despite data to the contrary. Some year later, the Food Standards Australia New Zealand team conducted a QMRA for raw cow milk that was largely based on unvalidated assumptions and extrapolations rather than reliable data for raw milk. Thus, both QMRAs imposed overestimation bias on their assessments for raw milk and did not fully disclose the impacts of their intentionally conservative assumptions on risk estimates, management options for risk reduction, or risk communications. Regarding guideline 10, the US Centers for Disease Control and Prevention provided a recent dataset for outbreaks from all transmission sources including both raw and pasteurized fluid milks for the period 2005 to 2020 . Raw and pasteurized milks both caused nearly 2,000 illnesses over this 16-year period (manuscript in preparation). Mortality rates associated with milks in North America in recent decades are quite low, including 5 US fatalities (3 associated with pasteurized milk and 2 with raw milk; , and 4 Canadian fatalities associated with pasteurized milk . Perhaps the most highly relevant general principle for QMRAs in this context is guideline 11. The findings of the pilot study, the lack of growth of the foodborne pathogens in raw milk for 14 days at 4.4°C for the major foodborne pathogens causing raw milk outbreaks in the US, are consistent with other peer reviewed studies conducted between 4 and 5°C [ , , ] that falsify the incorrect assumptions about pathogen growth in historic QMRAs. Two independent academic research teams re-evaluated and extended portions of the historic FDA/FSIS risk assessment for severe listeriosis . Latorre and colleagues estimated risks per raw milk serving to the general population were as low as 10 −15 (~1 illness per 1,000,000,000,000,000 servings), substantially lower estimated risks compared to the FDA/FSIS 2003 assessment that pooled growth data for raw and pasteurized milks. Stasiewicz and colleagues found in re-assessment that increasing heat treatments increased the growth rates of L . monocytogenes in pasteurized and ultra-pasteurized milks, consistent with killing more of the milk microbiota and thus reducing competition with the pathogen. These researchers also provided supplemental information for their study reporting no growth of the pathogen in raw milk blanks and increasing rates of growth in the raw milk pasteurized for 25 seconds at 72° and 82°C. The need to update incorrect assumptions and misinformation about both predictive microbiology and dose-response relationships in this QMRA was raised for future re-assessment for the FDA/FSIS QMRA .
The need to update incorrect assumptions about pathogen growth in raw milk that were made in historical QMRAs is more urgent than ever because so many claims about raw and pasteurized milks are made in the media, as well as in the scientific literature, without rigorous supporting data. Consumers and scientists can understandably be confused by conflicting claims. SRA leaders seek to encourage others to apply the SRA RAQT in both review of completed QMRAs for other foods and water, as well as in planning for future risk analysis projects, with the goal of developing a culture of full disclosure and quality analysis. Cultural, social, or ideological constructions have in the past limited the influence of scientific evidence into policy making, as documented by Meagher and colleagues on cultural mischaracterization of two foodborne outbreaks. Despite quick tracing of a 2006 outbreak to California-grown spinach, FDA’s public risk communication to avoid consuming any raw spinach contributed to market collapse, and “a range of plausible responses were never considered” (, pg. 245). Similarly, organizations around the world appear to incorrectly attribute high risk to raw milk from all producers, and not to any source of pasteurized milk. Participants in the SRA workshop on risk analysis quality discussed common unstated and unsupported assumptions about milks include: 1) the source of microbes in milk is feces; 2) raw milks are inherently dangerous; 3) pasteurization is a ‘silver bullet’; and 4) pasteurized milk is zero risk (manuscript in preparation). Current evidence documented herein and by Coleman and colleagues and Dietert and colleagues supports none of these assumptions. The importance of correctly modeling the microbial ecology of raw milk demonstrated by LAB strains isolated from raw bovine milk suppression or exclusion of three pathogens inoculated at two high densities, 10 3 and 10 6 log 10 CFU/mL . Clearly, the natural milk microbiota can suppress the growth of pathogens under some conditions. From our perspective of available data and analysis consistent with principles of microbial ecology and those of the , as well as the RAQT of the SRA, evidence that raw milk is ‘inherently dangerous’ is lacking. Evidence is consistent with protective multi-hurdle synergies of raw milk including the dense and diverse natural microbiota of mammalian milks under proper refrigeration contributing to suppression or exclusion of pathogens. We also acknowledge that no food is risk free, and benefits and risks could and perhaps should be characterized for all foods. Further improvements in the credibility and utility of QMRAs might develop with deeper consideration of environmental sustainability, economics and food waste, supply chain structure, climate change, and social and cultural factors [ – ]. For example, Duret and colleagues determined that setting the domestic refrigerator temperature to 4°C presented the best compromise for balancing risk of foodborne illness, food waste, and energy consumption. Rendueles and colleagues suggest not only that multi-hurdle approaches can reduce risk of illness and maintain food quality, but also can support more sustainable food production chains in the global market. Thus, for design of future predictive microbiology studies to inform risk analysis, studies must include additional production lots, dairy farms, and regions or states to characterize regional or national trends for milk risks. Expansions of the pilot study design should also include more frequent sampling and multiple initial inoculation levels for pathogens (at least ~1 CFU/mL and ~1,000 CFU/mL) so that robust parameters for growth and decline can be estimated. An ideal study design might also explore potential mechanisms of pathogen suppression under proper refrigeration and temperature abuse scenarios by quantitating key representatives of the raw milk microbiota over the study period and identifying microbial associations that drive pathogen suppression and killing. Rigorous quantitative data on predictive microbiology of raw milks is essential to re-evaluating historic QMRAs based on invalid assumptions about pathogen growth in raw milk so that unbiased estimates of risks and benefits can be generated for raw and pasteurized milks.
Results from a small pilot study with fresh raw milk produced for direct human consumption were consistent with previous studies demonstrating suppression of growth of major bacterial pathogens at proper refrigeration temperatures. Future research is needed to expand the results of the small pilot study on pathogen suppression in raw milks to address risk analysis more holistically, structuring and simulating tradeoffs between benefits and risks of raw and pasteurized milks.
S1 Appendix FSNS report , determination of growth rate of Salmonella enterica spp., E . coli O157:H7, Campylobacter spp., and Listeria monocytogenes in raw milk. (PDF) Click here for additional data file. S2 Appendix Risk analysis quality test of the society for risk analysis. (PDF) Click here for additional data file. S3 Appendix Dataset from pilot study. (XLSX) Click here for additional data file. S1 Table Published studies on pathogen growth and decline in raw milk at refrigeration and abuse temperatures. (DOCX) Click here for additional data file.
|
Neurogenic differentiation 2 promotes inflammatory activation of macrophages in doxorubicin-induced myocarditis via regulating protein kinase D | f55c127e-dfb2-4cd9-8a3d-0fcc391a8f3f | 11916933 | Musculoskeletal System[mh] | Myocarditis is an inflammatory disease of the cardiac muscle, typically triggered by viral infections and subsequent immune-mediated responses . Patients with myocarditis can manifest a wide spectrum of clinical symptoms, ranging from asymptomatic cases to cardiogenic shock and unexpected sudden death . Recent studies investigating sudden cardiac death in young individuals have reported myocarditis in 2–42% of autopsied cases . Besides supportive care, there are currently limited treatment options for both the acute and chronic stages of myocarditis , emphasizing the need to develop innovative therapeutic interventions. The primary cause of myocarditis is localized or widespread myocardial interstitial inflammation, which leads to the degeneration and necrosis of myocardial cells and fibers, ultimately impairing cardiac function . Cardiac tissue macrophages, the immune cells in the myocardium, perform a myriad of roles in normal physiological tissue maintenance and various pathological conditions . In response to cardiac injury, the population of cardiac macrophages significantly increases, promoting inflammatory activation and the release of inflammatory cytokines such as IL-1β, IL-6, and TNF-α, thereby accelerating myocardial damage . As crucial players in the inflammatory response, macrophages also profoundly affect the development of cardiovascular diseases, including unfavorable cardiac remodeling, hypertension, myocardial infarction, and more . Previous studies have reported that in human CVB3-induced myocarditis, infiltrating macrophages exhibit classical activation phenotypes (M1-type) and enhance cardiac inflammation by releasing pro-inflammatory cytokines such as TNF-α and IL-6. Importantly, these M1-type macrophages can also influence the ensuing adaptive immune response to pathological Th1 responses, thereby exacerbating myocarditis, highliting the critical pathological role of macrophages-mediated inflammation in CVB3-induced myocarditis . The NLR family pyrin domain-containing protein 3 (NLRP3) inflammasome is a crucial multiprotein complex that is essential for modulating the innate immune system and inflammatory signaling pathways . And nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), a pivotal activator of inflammatory responses, initiates the activation of the NLRP3-inflammasome by stimulating the expression of pro-IL-1β and NLRP3 . A prior study indicated that the NF-κB/NLRP3 pathway holds a pivotal role in the pyroptosis of cardiomyocytes . Therefore, targeting the inflammatory activation of macrophages, along with the NF-κB/NLRP3 axis, could be a promising therapeutic approach for myocarditis. Protein kinase D (PKD), a family of serine/threonine protein kinases, plays a vital role in the signaling network of the second messenger diacylglycerol . It has been established that PKD is activated by various extracellular stimuli and transmits crucial cell signals that impact fundamental cellular functions, including secretion, migration, proliferation, survival, angiogenesis, and immune responses . Dysregulation of PKD expression and activity has been implicated in numerous human diseases, such as cancer, metabolic disorders, central nervous system disorders, as well as cardiac and inflammatory conditions . For instance, Venardos et al. reported that the PKD inhibitor CID755673 enhances cardiac function in diabetic db / db mice . Additionally, studies have highlighted PKD’s involvement in inflammatory bowel disease and protease-induced neurogenic inflammation and pain . Nevertheless, the role and potential mechanisms of PKD in myocarditis remain unexplored. Through JASPAR database mining, we found that NeuroD2 was functionally targeted by PKD. Neurogenic Differentiation 2 (NeuroD2), a member of the basic-helix-loop-helix (bHLH) transcription factor family, acts as a master regulator of cell proliferation, neuronal differentiation, and specification . Additionally, NeuroD2 has been reported to be linked to inflammation, as evidenced by its upregulation in mice with spinal cord injury, where it plays a role in regulating inflammation and oxidative stress . However, its role in myocarditis remains largely unknown. Based on previous research and bioinformatic analysis, we investigated the role of PKD in a mouse model of myocarditis induced by doxorubicin (DOX) and assessed its impact on pathological conditions and cardiac function in this model. Furthermore, we explored the role of PKD in the inflammatory activation of macrophages isolated from mice in different experimental groups. We also conducted additional studies to elucidate the interaction between PKD and neurogenic differentiation 2 (NeuroD2) and the involvement of the NLRP3/NF-κB pathway in DOX-induced myocarditis.
Database analysis The JASPAR database ( http://www.jaspar.genereg.net )) was used to explore the possible binding sites of the NeuroD2 and PKD promoters. Animal model All in vivo experiments were performed according to the guidelines for the Care and Use of Laboratory Animals of Xiangya Hospital of Central South University. This study was conducted in accordance with the ARRIVE Guidelines and approved by the Institutional Animal Care and Use Committee (IACUC) of Xiangya Hospital of Central South University (Approval No. 2022101065). 24 healthy male C57BL/6 mice (4–6 weeks old, 16–18 g) were purchased from the Laboratory Animal Center of Southern Medical University. Mice were provided with food and water ad libitum and kept under controlled conditions (Temperature 22.2 °C, air humidity 40–70%) with 12 h:12 h light/dark cycles. The mice were randomly divided into four groups ( n = 6 per group): the control group, the DOX group, the DOX + DMSO group, and the DOX + CID755673 group. CID755673 was dissolved in DMSO. Mice were treated with intraperitoneal saline injections as a control, and mice received 1.5 mg/kg DOX once every two days for a total of three weeks by intraperitoneal injection in the DOX group. Mice in the DOX + DMSO group and DOX + CID755673 group were treated with equal volume intraperitoneal injections of DOX (once every two days for three weeks), and these mice received daily intraperitoneal injections of 10 mg/kg CID755673 (Abmole Biosciences, Houston, Texas, USA) or equal volume of DMSO solution during DOX treatment for two weeks. On day 21 of injection, echocardiography was performed to detect myocardial functions. Then, the mice were sacrificed, and heart tissues were harvested for subsequent experiments. Echocardiography Mice were placed on a heating pad at 37 °C and keep continuous anesthetized with 2-3% isoflurane. Pulse oximetry was used to monitor heart rate and oxygen saturation, while the depth of anesthesia was regularly controlled by evaluating the toe-withdrawal reflex and monitoring heart rate (The desired heart rate target was 400 ± 50 bpm). Echocardiography was carried out on SomnoSuite small animal anaesthesia system (Kent Scientific Corporation, USA), standard echo views (parasternal long axis view [PLAX], parasternal short axis view [PSAX] and apical four‐chamber view [4‐CV]) were obtained as previously reported. All measurements were obtained from M-mode images captured in the parasternal long-axis view (PLAX) at the papillary muscle level of mice. The left ventricular end-diastolic diameter (LVEDd), left ventricular end-systolic diameter (LVESd) were measured, and left ventricular fraction shortening (LVFS) and left ventricular ejection fraction (LVEF) were calculated as follows. \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$\:\text{E}\text{F}=100\:\text{x}\frac{(\text{L}\text{V}\text{E}\text{D}\text{V}-\text{L}\text{V}\text{E}\text{S}\text{V})}{\text{L}\text{V}\text{E}\text{D}\text{V}}\:;\text{F}\text{S}=100\:\text{x}\:\frac{(\text{L}\text{V}\text{E}\text{D}\text{D}-\text{L}\text{V}\text{E}\text{S}\text{D})}{\text{L}\text{V}\text{E}\text{D}\text{D}};\:$$\end{document} Hematoxylin and Eosin staining Heart tissues obtained from different groups were fixed in 4% paraformaldehyde at room temperature overnight. The tissues underwent dehydration through a series of graded ethanol, followed by embedding in paraffin and slicing into sections with 5-µm-thickness. After dewaxing and dehydration, slices were stained with Hematoxylin (Sigma Aldrich) for a duration of 3–6 min, subsequently rinsed for 1–2 min, and then differentiated for 1–3 s using 1% hydrochloric acid alcohol, and encouraged the liquid to turn blue for 5–10 s, rinsed under running water for 15–30 s, and then stained with 0.5% eosin solution (Sigma Aldrich) for 2–3 min.The sections then underwent washing with distilled water for 1 to 2 s, followed by immersion in 80% ethanol for 15–30 s, then in 95% ethanol for another 15–30 s, and ultimately in anhydrous ethanol for 1 to 2 s. After being dried, they were sealed with neutral gum. The histological examination was performed using a light microscope (Olympus Corp.) with the Olympus DP70 digital camera. Immunofluorescence staining Mouse heart tissues were fixed in 4% paraformaldehyde and embedded in paraffin. The paraffin blocks were then sectioned into 5 μm thick slices. Following dewaxing and rehydration, sections were incubated with 5% bovine serum albumin for 1 h. After being washed with PBS, the slides were incubated with primary antibodies against CD68 (1:50, Abcam, Cambridge, MA, USA) at 4℃ overnight. Subsequently, they were incubated with the corresponding secondary antibodies for 1 h. Images were acquired using confocal microscopy (Carl Zeiss, Germany), and the average fluorescence intensity was quantified using ZEN software. Enzyme-linked immunosorbent assay The levels of tumor necrosis factor (TNF-α), interleukin (IL)-6, IL-18, and IL-1β in macrophages supernatant or heart homogenate were measured using respective enzyme-linked immunosorbent assay (ELISA) kits (Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer’s instructions. Cell culture and flow cytometry Macrophages were isolated from mice heart tissues from different groups. Briefly, mice were sacrificed, and hearts were collected and cut into small pieces of 1–2 mm 3 on ice. Heart tissues were digested by 0.125% trypsin at 37 °C for 5–6 times for a total of 10 min. A single-cell suspension was collected after centrifugation and filtration. Macrophages were isolated from the single-cell suspension using the Anti-F4/80 MicroBeads UltraPure kit (Novobiotec, Beijing, China), according to the manufacturer’s protocol. The purity of freshly isolated macrophages was determined by a FACSAria II flow cytometer (BD Biosciences, San Jose, CA, USA). Quantitative real-time reverse transcription polymerase chain reaction The total RNA was extracted from macrophages or heart tissues using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), and the Reverse Transcription Kit (Takara, Dalian, China) was used to convert RNA into complementary DNA (cDNA). Quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR) was conducted using SYBR Select Master Mix on an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Relative expression was quantified following the 2 −ΔΔCt method. The primers used in this research are listed in Table . Western blotting Mice heart tissues and macrophages were lysed in ice-cold RIPA lysis buffer containing protease inhibitors (Beyotime Biotechnology, Shanghai, China) to extract the total protein. The BCA Protein Assay Kit (Beyotime Biotechnology) was utilized to measure the total protein concentrations. The protein extracts were separated with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes. After blocking 1 h with 5% skimmed milk, the membranes were incubated with the following primary antibodies at 4℃ overnight, including PKD (1:2000; Affinity Biosciences, Cincinnati, OH, USA), NLRP3 (1:2000; Affinity Biosciences), p-p65 (1:1000; Affinity Biosciences), p65 (1:1000; Affinity Biosciences), Neurod2 (1:1000; Biorbyt, Cambridge, UK.), GAPDH (1:3000; Affinity Biosciences). Followed by treatment with horseradish peroxidase (HRP)-conjugated secondary antibody (1:3000; Affinity Biosciences) at room temperature for 2 h. The protein expression was detected by an enhanced chemiluminescence (ECL) detection system. Plasmid, oligonucleotides and transfection The full length of the PKD cDNA sequence was amplified and cloned into the pcDNA3.1 vector to construct a PKD overexpression plasmid named pcDNA3.1- PKD. Besides, the NeuroD2 overexpression plasmid was constructed. The NeuroD2 small interfering RNA (siNeuroD2) and its negative control (NC siRNA) were obtained from Genechem (Shanghai, China). The sequences of oligonucleotides were as follows: NC siRNA: GCAAGCUGACCCUGAAGUUC; siNeuroD2: GAAUCUCUUGUCUUACGAUAU. These overexpression plasmids and oligonucleotides were transfected into macrophages using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Dual-Luciferase reporter assay The wild (wt) and mutant (mut) types in the PKD promoter binding site were subcloned into psiCHECK-2 plasmids (Promega, Madison, WI, USA) and co-transfected with overexpression NeuroD2 plasmids (Promega) in macrophages using Lipofectamine 2000. After 48 h, the luciferase activity was measured using a dual-luciferase assay kit (Promega, Madison, WI, USA) according to the manufacturer’s instructions. Chromatin Immunoprecipitation A Chromatin immunoprecipitation (ChIP) experiment was conducted using the SimpleChIP Mix (Cell Signaling Technology, Beverly, MA, USA) according to the manufacturer’s protocol. 6 × 10 6 cells were treated with 1% formaldehyde for 10 min. The cell lysate was then sonicated until the length of DNA remained between 200 and 1000 bp. An equivalent chromatin volume was immunoprecipitated at 4 °C overnight with the antibody Neurod2 (1:50, Biorbyt). Immune complexes were collected after incubating with Magnetic Beads Protein A/G. The immunoprecipitated DNA was isolated and examined by qPCR, and the primers were as follows: PKD primers: forward 5’-GCACTGCTCTTCACTGCTATCA-3’; reverse 5’- AAGTCACCAAGAGGGAACACC − 3’. Statistical analyses All statistical analyses were performed using SPSS21.0 (SPSS Inc., Chicago, IL, USA). Data were expressed as the mean ± standard deviation (sd). Differences between the two groups were compared using the student’s t -test. Differences among multiple groups were analyzed using one-way analysis with Tukey’s multiple comparisons test. A P-value < 0.05 was statistically significant.
The JASPAR database ( http://www.jaspar.genereg.net )) was used to explore the possible binding sites of the NeuroD2 and PKD promoters.
All in vivo experiments were performed according to the guidelines for the Care and Use of Laboratory Animals of Xiangya Hospital of Central South University. This study was conducted in accordance with the ARRIVE Guidelines and approved by the Institutional Animal Care and Use Committee (IACUC) of Xiangya Hospital of Central South University (Approval No. 2022101065). 24 healthy male C57BL/6 mice (4–6 weeks old, 16–18 g) were purchased from the Laboratory Animal Center of Southern Medical University. Mice were provided with food and water ad libitum and kept under controlled conditions (Temperature 22.2 °C, air humidity 40–70%) with 12 h:12 h light/dark cycles. The mice were randomly divided into four groups ( n = 6 per group): the control group, the DOX group, the DOX + DMSO group, and the DOX + CID755673 group. CID755673 was dissolved in DMSO. Mice were treated with intraperitoneal saline injections as a control, and mice received 1.5 mg/kg DOX once every two days for a total of three weeks by intraperitoneal injection in the DOX group. Mice in the DOX + DMSO group and DOX + CID755673 group were treated with equal volume intraperitoneal injections of DOX (once every two days for three weeks), and these mice received daily intraperitoneal injections of 10 mg/kg CID755673 (Abmole Biosciences, Houston, Texas, USA) or equal volume of DMSO solution during DOX treatment for two weeks. On day 21 of injection, echocardiography was performed to detect myocardial functions. Then, the mice were sacrificed, and heart tissues were harvested for subsequent experiments.
Mice were placed on a heating pad at 37 °C and keep continuous anesthetized with 2-3% isoflurane. Pulse oximetry was used to monitor heart rate and oxygen saturation, while the depth of anesthesia was regularly controlled by evaluating the toe-withdrawal reflex and monitoring heart rate (The desired heart rate target was 400 ± 50 bpm). Echocardiography was carried out on SomnoSuite small animal anaesthesia system (Kent Scientific Corporation, USA), standard echo views (parasternal long axis view [PLAX], parasternal short axis view [PSAX] and apical four‐chamber view [4‐CV]) were obtained as previously reported. All measurements were obtained from M-mode images captured in the parasternal long-axis view (PLAX) at the papillary muscle level of mice. The left ventricular end-diastolic diameter (LVEDd), left ventricular end-systolic diameter (LVESd) were measured, and left ventricular fraction shortening (LVFS) and left ventricular ejection fraction (LVEF) were calculated as follows. \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$\:\text{E}\text{F}=100\:\text{x}\frac{(\text{L}\text{V}\text{E}\text{D}\text{V}-\text{L}\text{V}\text{E}\text{S}\text{V})}{\text{L}\text{V}\text{E}\text{D}\text{V}}\:;\text{F}\text{S}=100\:\text{x}\:\frac{(\text{L}\text{V}\text{E}\text{D}\text{D}-\text{L}\text{V}\text{E}\text{S}\text{D})}{\text{L}\text{V}\text{E}\text{D}\text{D}};\:$$\end{document}
Heart tissues obtained from different groups were fixed in 4% paraformaldehyde at room temperature overnight. The tissues underwent dehydration through a series of graded ethanol, followed by embedding in paraffin and slicing into sections with 5-µm-thickness. After dewaxing and dehydration, slices were stained with Hematoxylin (Sigma Aldrich) for a duration of 3–6 min, subsequently rinsed for 1–2 min, and then differentiated for 1–3 s using 1% hydrochloric acid alcohol, and encouraged the liquid to turn blue for 5–10 s, rinsed under running water for 15–30 s, and then stained with 0.5% eosin solution (Sigma Aldrich) for 2–3 min.The sections then underwent washing with distilled water for 1 to 2 s, followed by immersion in 80% ethanol for 15–30 s, then in 95% ethanol for another 15–30 s, and ultimately in anhydrous ethanol for 1 to 2 s. After being dried, they were sealed with neutral gum. The histological examination was performed using a light microscope (Olympus Corp.) with the Olympus DP70 digital camera.
Mouse heart tissues were fixed in 4% paraformaldehyde and embedded in paraffin. The paraffin blocks were then sectioned into 5 μm thick slices. Following dewaxing and rehydration, sections were incubated with 5% bovine serum albumin for 1 h. After being washed with PBS, the slides were incubated with primary antibodies against CD68 (1:50, Abcam, Cambridge, MA, USA) at 4℃ overnight. Subsequently, they were incubated with the corresponding secondary antibodies for 1 h. Images were acquired using confocal microscopy (Carl Zeiss, Germany), and the average fluorescence intensity was quantified using ZEN software.
The levels of tumor necrosis factor (TNF-α), interleukin (IL)-6, IL-18, and IL-1β in macrophages supernatant or heart homogenate were measured using respective enzyme-linked immunosorbent assay (ELISA) kits (Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer’s instructions.
Macrophages were isolated from mice heart tissues from different groups. Briefly, mice were sacrificed, and hearts were collected and cut into small pieces of 1–2 mm 3 on ice. Heart tissues were digested by 0.125% trypsin at 37 °C for 5–6 times for a total of 10 min. A single-cell suspension was collected after centrifugation and filtration. Macrophages were isolated from the single-cell suspension using the Anti-F4/80 MicroBeads UltraPure kit (Novobiotec, Beijing, China), according to the manufacturer’s protocol. The purity of freshly isolated macrophages was determined by a FACSAria II flow cytometer (BD Biosciences, San Jose, CA, USA).
The total RNA was extracted from macrophages or heart tissues using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), and the Reverse Transcription Kit (Takara, Dalian, China) was used to convert RNA into complementary DNA (cDNA). Quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR) was conducted using SYBR Select Master Mix on an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Relative expression was quantified following the 2 −ΔΔCt method. The primers used in this research are listed in Table .
Mice heart tissues and macrophages were lysed in ice-cold RIPA lysis buffer containing protease inhibitors (Beyotime Biotechnology, Shanghai, China) to extract the total protein. The BCA Protein Assay Kit (Beyotime Biotechnology) was utilized to measure the total protein concentrations. The protein extracts were separated with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes. After blocking 1 h with 5% skimmed milk, the membranes were incubated with the following primary antibodies at 4℃ overnight, including PKD (1:2000; Affinity Biosciences, Cincinnati, OH, USA), NLRP3 (1:2000; Affinity Biosciences), p-p65 (1:1000; Affinity Biosciences), p65 (1:1000; Affinity Biosciences), Neurod2 (1:1000; Biorbyt, Cambridge, UK.), GAPDH (1:3000; Affinity Biosciences). Followed by treatment with horseradish peroxidase (HRP)-conjugated secondary antibody (1:3000; Affinity Biosciences) at room temperature for 2 h. The protein expression was detected by an enhanced chemiluminescence (ECL) detection system.
The full length of the PKD cDNA sequence was amplified and cloned into the pcDNA3.1 vector to construct a PKD overexpression plasmid named pcDNA3.1- PKD. Besides, the NeuroD2 overexpression plasmid was constructed. The NeuroD2 small interfering RNA (siNeuroD2) and its negative control (NC siRNA) were obtained from Genechem (Shanghai, China). The sequences of oligonucleotides were as follows: NC siRNA: GCAAGCUGACCCUGAAGUUC; siNeuroD2: GAAUCUCUUGUCUUACGAUAU. These overexpression plasmids and oligonucleotides were transfected into macrophages using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions.
The wild (wt) and mutant (mut) types in the PKD promoter binding site were subcloned into psiCHECK-2 plasmids (Promega, Madison, WI, USA) and co-transfected with overexpression NeuroD2 plasmids (Promega) in macrophages using Lipofectamine 2000. After 48 h, the luciferase activity was measured using a dual-luciferase assay kit (Promega, Madison, WI, USA) according to the manufacturer’s instructions.
A Chromatin immunoprecipitation (ChIP) experiment was conducted using the SimpleChIP Mix (Cell Signaling Technology, Beverly, MA, USA) according to the manufacturer’s protocol. 6 × 10 6 cells were treated with 1% formaldehyde for 10 min. The cell lysate was then sonicated until the length of DNA remained between 200 and 1000 bp. An equivalent chromatin volume was immunoprecipitated at 4 °C overnight with the antibody Neurod2 (1:50, Biorbyt). Immune complexes were collected after incubating with Magnetic Beads Protein A/G. The immunoprecipitated DNA was isolated and examined by qPCR, and the primers were as follows: PKD primers: forward 5’-GCACTGCTCTTCACTGCTATCA-3’; reverse 5’- AAGTCACCAAGAGGGAACACC − 3’.
All statistical analyses were performed using SPSS21.0 (SPSS Inc., Chicago, IL, USA). Data were expressed as the mean ± standard deviation (sd). Differences between the two groups were compared using the student’s t -test. Differences among multiple groups were analyzed using one-way analysis with Tukey’s multiple comparisons test. A P-value < 0.05 was statistically significant.
Protein kinase D is upregulated in heart tissues of mice induced by doxorubicin To assess cardiac function, transthoracic echocardiography was performed, revealing enhanced LVEDd and LVESd alongside decreased LVFS and LVEF in the DOX group compared to normal mice, indicative of impaired cardiac function (Fig. a - e). Histological examination via H&E staining revealed pronounced inflammatory cell infiltration and cardiomyocyte necrosis in mice with myocarditis, while normal mice exhibited minimal inflammation. The pathological score for tissue from mice with myocarditis was also significantly higher (Fig. f). And immunofluorescence staining assay suggested that DOX administration effectively increased the protein expression of macrophages maker CD68 in mice heart tissues compared with normal heart tissues (Fig. g). Furthermore, levels of inflammatory cytokines, including TNF-α, IL-6, IL-18, and IL-1β, were significantly elevated in heart tissue from the DOX group compared to the control group (Fig. h - i). These findings validate the successful establishment of the DOX-induced myocarditis model. Given the known role of the NLRP3 inflammasome in inflammatory disorders as well as the effects of PKD on inflammasome formation and activation , we assessed the tissue levels of PKD, NLRP3, phosphorylated p65 (p-p65), and p65, finding that PKD, NLRP3, p-p65, and p65 expressions were significantly upregulated in the DOX group compared to the control group (Fig. j). These results substantiate high cardiac tissue expression of PKD and inflammatory factors in mice induced by DOX. Protein kinase D inhibitor CID755673 improves the inflammation and cardiac function in mice with myocarditis To investigate the role of PKD in myocarditis, we treated mice with myocarditis by intraperitoneal injection of CID755673 (a PKD inhibitor). As expected, CID755673 treatment significantly reduced PKD expression (Fig. a and b). Echocardiographic data demonstrated reduced LVEDd and LVESd and improved LVFS and LVEF in the inhibitor-treated group (Fig. c - g), indicating a significant amelioration of cardiac dysfunction caused by DOX injection. Additionally, CID755673 intervention significantly decreased inflammatory cell infiltration and pathological scores (Fig. h), along with the protein expression of the macrophage marker CD68 in heart tissues (Fig. i). Furthermore, CID755673 treatment reversed the increased mRNA expressions of TNF-α, IL-6, IL-18, and IL-1β in myocarditis tissue (Fig. j), which was further confirmed by ELISA analysis (Fig. k). Considering the significance of the NLRP3/p65 pathway in PKD-mediated inflammation , we detected the levels of NLRP3 and p-p65 in myocardial tissues treated with or without CID755673, it was observed that CID755673 effectively reduced NLRP3, p65, and p-p65 levels in myocardial tissues (Fig. l and m). These results collectively indicate that downregulating PKD can improve inflammation and cardiac function in mice with myocarditis. Protein kinase D is involved in the doxorubicin-induced inflammatory activation of macrophages Macrophages were isolated from the heart tissues of mice in four groups: control group, DOX group, DOX + DMSO group, and DOX + CID755673 group. Large numbers of macrophages were present in the F4/80 + stained cells from heart tissues across different groups (Fig. a). Notably, PKD mRNA expression was significantly elevated in the DOX group but reversed by CID755673 administration (Fig. b and c). Additionally, the TNF-α, IL-6, IL-18, and IL-1β levels were efficiently inhibited by CID755673 compared to DOX alone, as confirmed by both ELISA and qPCR analyses (Fig. d - g). Furthermore, CID755673 significantly reduced the levels of NLRP3, p65, and p-p65 in macrophages compared to the DOX group (Fig. h - i). Collectively, these results indicate the crucial role for PKD in macrophage inflammatory activation. Functional interaction between neurogenic differentiation 2 and protein kinase D In silico analysis using the JASPAR database revealed putative binding sites between NeuroD2 and PKD (Fig . a). Western blotting and qPCR confirmed elevated NeuroD2 expression in both macrophages and tissues induced by DOX (Fig. b- c). Further experiments involved interfering with or overexpressing NeuroD2 in macrophages isolated from normal heart tissues and treated with DOX. NeuroD2 knockdown significantly downregulated NeuroD2 expression compared to the control group, while the opposite findings were observed for NeuroD2 overexpression (Fig. d- e). Moreover, PKD mRNA expression and protein levles in macrophages significantly decreased with NeuroD2 knockdown but was restored with NeuroD2 overexpression (Fig. f- g). Subsequently, we investigated the interaction between NeuroD2 and PKD, luciferase reporter assay revealed that NeuroD2 elevated the luciferase activity of the PKD promoter WT in macrophages without affecting the MUT (Fig. h), suggesting the combination between NeuroD2 and PKD, which was further confirmed through a chromatin immunoprecipitation (ChIP) experiment (Fig. i). These results suggest that PKD is a functional target of NeuroD2. The overexpression of protein kinase D can reverse the impact of neurogenic differentiation 2 on doxorubicin-induced inflammatory activation of macrophages A PKD overexpression plasmid was constructed to further confirm the role of NeuroD2 in macrophage inflammatory activation. Compared to the DOX alone group, the mRNA expressions of PKD and NLRP3 decreased after NeuroD2 interference, which was reversed by PKD overexpression (Fig. a and b). Similarly, NeuroD2 inhibition significantly reduced PKD, p-p65, p65, and NLRP3 protein levels in the DOX alone group, but PKD overexpression reversed these effects (Fig. c). Additionally, the levels of TNF-α, IL-6, IL-18, and IL-1β were significantly reduced by NeuroD2 downregulation compared to the DOX alone group, and these trends were reversed after PKD upregulation (Fig. d - g). In summary, these results indicate that the effects of NeuroD2 on DOX-induced macrophage inflammatory activation can be reversed by overexpressing PKD.
To assess cardiac function, transthoracic echocardiography was performed, revealing enhanced LVEDd and LVESd alongside decreased LVFS and LVEF in the DOX group compared to normal mice, indicative of impaired cardiac function (Fig. a - e). Histological examination via H&E staining revealed pronounced inflammatory cell infiltration and cardiomyocyte necrosis in mice with myocarditis, while normal mice exhibited minimal inflammation. The pathological score for tissue from mice with myocarditis was also significantly higher (Fig. f). And immunofluorescence staining assay suggested that DOX administration effectively increased the protein expression of macrophages maker CD68 in mice heart tissues compared with normal heart tissues (Fig. g). Furthermore, levels of inflammatory cytokines, including TNF-α, IL-6, IL-18, and IL-1β, were significantly elevated in heart tissue from the DOX group compared to the control group (Fig. h - i). These findings validate the successful establishment of the DOX-induced myocarditis model. Given the known role of the NLRP3 inflammasome in inflammatory disorders as well as the effects of PKD on inflammasome formation and activation , we assessed the tissue levels of PKD, NLRP3, phosphorylated p65 (p-p65), and p65, finding that PKD, NLRP3, p-p65, and p65 expressions were significantly upregulated in the DOX group compared to the control group (Fig. j). These results substantiate high cardiac tissue expression of PKD and inflammatory factors in mice induced by DOX.
To investigate the role of PKD in myocarditis, we treated mice with myocarditis by intraperitoneal injection of CID755673 (a PKD inhibitor). As expected, CID755673 treatment significantly reduced PKD expression (Fig. a and b). Echocardiographic data demonstrated reduced LVEDd and LVESd and improved LVFS and LVEF in the inhibitor-treated group (Fig. c - g), indicating a significant amelioration of cardiac dysfunction caused by DOX injection. Additionally, CID755673 intervention significantly decreased inflammatory cell infiltration and pathological scores (Fig. h), along with the protein expression of the macrophage marker CD68 in heart tissues (Fig. i). Furthermore, CID755673 treatment reversed the increased mRNA expressions of TNF-α, IL-6, IL-18, and IL-1β in myocarditis tissue (Fig. j), which was further confirmed by ELISA analysis (Fig. k). Considering the significance of the NLRP3/p65 pathway in PKD-mediated inflammation , we detected the levels of NLRP3 and p-p65 in myocardial tissues treated with or without CID755673, it was observed that CID755673 effectively reduced NLRP3, p65, and p-p65 levels in myocardial tissues (Fig. l and m). These results collectively indicate that downregulating PKD can improve inflammation and cardiac function in mice with myocarditis.
Macrophages were isolated from the heart tissues of mice in four groups: control group, DOX group, DOX + DMSO group, and DOX + CID755673 group. Large numbers of macrophages were present in the F4/80 + stained cells from heart tissues across different groups (Fig. a). Notably, PKD mRNA expression was significantly elevated in the DOX group but reversed by CID755673 administration (Fig. b and c). Additionally, the TNF-α, IL-6, IL-18, and IL-1β levels were efficiently inhibited by CID755673 compared to DOX alone, as confirmed by both ELISA and qPCR analyses (Fig. d - g). Furthermore, CID755673 significantly reduced the levels of NLRP3, p65, and p-p65 in macrophages compared to the DOX group (Fig. h - i). Collectively, these results indicate the crucial role for PKD in macrophage inflammatory activation.
In silico analysis using the JASPAR database revealed putative binding sites between NeuroD2 and PKD (Fig . a). Western blotting and qPCR confirmed elevated NeuroD2 expression in both macrophages and tissues induced by DOX (Fig. b- c). Further experiments involved interfering with or overexpressing NeuroD2 in macrophages isolated from normal heart tissues and treated with DOX. NeuroD2 knockdown significantly downregulated NeuroD2 expression compared to the control group, while the opposite findings were observed for NeuroD2 overexpression (Fig. d- e). Moreover, PKD mRNA expression and protein levles in macrophages significantly decreased with NeuroD2 knockdown but was restored with NeuroD2 overexpression (Fig. f- g). Subsequently, we investigated the interaction between NeuroD2 and PKD, luciferase reporter assay revealed that NeuroD2 elevated the luciferase activity of the PKD promoter WT in macrophages without affecting the MUT (Fig. h), suggesting the combination between NeuroD2 and PKD, which was further confirmed through a chromatin immunoprecipitation (ChIP) experiment (Fig. i). These results suggest that PKD is a functional target of NeuroD2.
A PKD overexpression plasmid was constructed to further confirm the role of NeuroD2 in macrophage inflammatory activation. Compared to the DOX alone group, the mRNA expressions of PKD and NLRP3 decreased after NeuroD2 interference, which was reversed by PKD overexpression (Fig. a and b). Similarly, NeuroD2 inhibition significantly reduced PKD, p-p65, p65, and NLRP3 protein levels in the DOX alone group, but PKD overexpression reversed these effects (Fig. c). Additionally, the levels of TNF-α, IL-6, IL-18, and IL-1β were significantly reduced by NeuroD2 downregulation compared to the DOX alone group, and these trends were reversed after PKD upregulation (Fig. d - g). In summary, these results indicate that the effects of NeuroD2 on DOX-induced macrophage inflammatory activation can be reversed by overexpressing PKD.
Myocarditis poses a significant challenge due to the absence of effective treatment options in current practice, highlighting the need for innovative approaches in cardiovascular medicine . Previous studies have emphasized the pivotal role of excessive macrophage activation-driven inflammation in the pathogenesis of myocarditis . Notably, macrophages are the predominant cardiac inflammatory cell subset during the early stages of CVB3-induced viral myocarditis, as documented by Fairweather et al. . Epelman et al. have also elucidated the critical role of macrophages in initiating, amplifying, and sustaining inflammation . In this context, our study utilized a DOX-induced mouse model to investigate the influence of PKD and NeuroD2 on macrophage inflammatory activation in myocarditis. Clinically, myocarditis can be caused by a diverse range of infectious agents, such as viruses, bacteria, chlamydia, rickettsia, fungi, and protozoa, along with toxic and allergic reactions. Among these infectious agents, viruses are the most commonly reported to be associated with myocarditis . Following infection, changes in the quantity and function of macrophages are commonly observed in patients with both acute and chronic myocarditis . Moreover, inflammatory markers such as inter leukin 6, 8, and 10 are found significantly elevated , and inflammatory cellular infiltrates can be observed in heart tissue section stained using conventional methods . DOX is a potent and commonly utilized anticancer medication. However, the clinical utilization of DOX is constrained due to its severe cardiotoxicity and cardiac injury, typically myocarditis . In this research, we constructed myocarditis mouse model by intraperitoneal injection of DOX, it was found that the introduction of DOX significantly elevated the levels of inflammatory cytokines including IL-6, IL-18 and IL-1βin mice heart tissues. And similar results were observed in macrophages isolated from mice heart tissues, suggesting that DOX treatment may trigger macrophages inflammatory activation. Moreover, H&E staining indicated an obvious inflammatory cell infiltration in heart tissues harvested from DOX group, which is in consistent with the pathological conditions in clinical virus-induced myocarditis. PKD plays a multifaceted role in various biological processes, including angiogenesis, heart contraction, cell differentiation, apoptosis, immunomodulation, and cancer [ – ]. Studies have reported PKD’s involvement in cardiomyocyte hypertrophy through regulating extracellular signal-regulated and myocyte enhancer factor 2D pathways . Additionally, another study suggested that PKD has been implicated in stress signaling modulation within the heart, affecting gene expression, cell survival, excitation-contraction coupling, and metabolism . Our research revealed significant upregulation of PKD in cardiac-infiltrating macrophages, and PKD inhibition was associated with enhanced cardiac function, reduced pathological conditions, and decreased levels of inflammatory cytokines in mice with myocarditis. These findings suggest a potential role for PKD in macrophage-mediated inflammation during the myocarditis process. The JASPAR database predicted a potential interaction between NeuroD2 and PKD. Subsequent experiments in our study confirmed this prediction by demonstrating the direct binding of NeuroD2 to the PKD promoter, thereby regulating PKD expression. NeuroD2, a neurogenic transcription factor, have associated with multiple diseases, including its involvement in the regulation of inflammation and oxidative stress . However, studies investigating the role of NeuroD2 in myocarditis are limited. Our study observed elevated NeuroD2 expression in both macrophages and myocardial tissues induced by DOX. And addition of NeuroD2 knockdown significantly reduced the PKD mRNA expression and protein levles in macrophages, while these levels were restored with NeuroD2 overexpression.We further confirmed the interaction between NeuroD2 and the PKD promoter through luciferase activity assays and ChIP experiments. Overexpression of PKD reversed the impact of NeuroD2 on inflammatory cytokine levels to some extent. Inflammasomes serve as pattern recognition receptors that are crucial for host defense and sterile inflammatory disorders. It is now understood that the NLRP3 inflammasome can be activated by a wide range of pathogen-associated molecular patterns (PAMPs) and endogenous danger-associated molecular patterns (DAMPs) . Previous studies have implicated PKD in inflammasome activation and development , including its role in the phosphorylation and release of NLRP3 from Golgi membranes, facilitating the assembly of active inflammasome complexes. Inhibition of PKD was shown to retain NLRP3 at Golgi membranes, preventing inflammasome assembly . Our study revealed a significant increase in NLRP3 levels in both DOX-induced mice tissues and macrophages, and PKD inhibitor CID755673 treatment effectively restrained the NLRP3 elevation, indicating the involvement of PKD in NLRP3 inflammasome activation, which was further supported by the increased release of IL-1βand IL-18. NF-κB is a nuclear transcription factor critical for various physiological reactions, particularly those related to inflammatory responses . It serves as a signal for the transcriptional activation of NLRP3 inflammasome components , especially in response to oxidative stress . And our results also revealed that DOX injections upregulated the levels of phosphorylated p65, TNF-α and IL-6, which were also reversed by CID755673 addition. These results establish PKD as a crucial regulator in triggering inflammatory responses. Moreover, the interaction between NeuroD2 and the PKD promoter amplified the NLRP3/NF-κB signaling pathway, exacerbating macrophage inflammation and finally contributing to the progression of myocarditis (Fig. ). However, there are still some limitations to our study. First, the limited number of mice utilized in each group may potentially impact the results, In future research, we plan to increase the number of subjects per group to enhance statistical power and ensure more robust findings. Second, another limitation of our study is the lack of validation using clinical samples. it is crucial to incorporate clinical sample analysis to validate our findings in future studies. Third, in this research, we investigated the role of the NeuroD2/PKD axis in DOX-induced myocarditis. For further studies, other types of animal models, such as those for viral myocarditis, can be developed to expand our investigation.
Our findings highlight the upregulation of both NeuroD2 and PKD in myocarditis. Inhibition of PKD led to improvements in pathological status and cardiac function in the DOX-induced mouse model. Furthermore, NeuroD2 functionally targeted PKD, enhancing the NLRP3/NF-κB signaling pathway and promoting macrophage inflammation, ultimately exacerbating myocarditis progression. These results suggest that targeting the NeuroD2/PKD axis may hold promise as a potential therapeutic approach against DOX-induced myocarditis.
Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2
|
Ultrasonography for diagnosing medial sided ankle instability in supination external rotation ankle fracture | 3a7e4653-b53c-4012-b07e-4078d8d6309c | 11892171 | Musculoskeletal System[mh] | The critical treatment distinction in supination external rotation (SER) ankle fractures lies in the ability to identify whether the injury has rendered the ankle unstable. In the setting of a displaced bimalleolar fracture, or when the tibiotalar joint is clearly subluxated, operative treatment is strongly recommended [ – ]. In the setting of an isolated lateral malleolar fracture, however, stability is predicated by whether the deep deltoid ligament is intact . The deep deltoid ligament is critical towards preventing lateral talar shift and external rotation of the talus. Still, numerous studies have demonstrated that patient symptoms, physical examination, and static imaging techniques are not able to accurately diagnose the injury of the deltoid that results in medial ankle instability [ – ]. Several imaging modalities that do allow for a dynamic examination are currently being used to diagnose medial sided ankle instability. Manual external rotation stress radiographs, gravity stress radiographs, and weightbearing radiographs are commonly used techniques [ , , – ]. However, there is no consensus on which loading condition is considered the best for evaluating medial ankle instability in SER ankle injuries and current clinical practice varies based on clinician preference [ , – ]. With advances in ultrasound technology, including improved accuracy of image details as well as the ability to provide dynamic, multiplanar, real-time films, the potential for evaluating musculoskeletal conditions is rapidly increasing. Few studies have explored the feasibility and accuracy of the US examination of the deltoid ligament in the setting of ankle fractures showing promising results [ – ]. However, these studies explored the injured vs. uninjured state of the deltoid ligaments, such as fluid, hematoma, discontinuity of the ligament, and did not assess the ability of ultrasound to diagnose medial-sided ankle instability related to ligament injury. Currently, no study has used ultrasound to dynamically evaluate the medial clear space (MCS) of the ankle when performing a gravity stress test (GST), an external rotation stress test or with weightbearing. As a predicate to using ultrasound to diagnose tibiotalar instability after an SER ankle fracture, we developed a novel ultrasound evaluation technique aiming to quantify medial ankle instability using ultrasound. This study aims to perform such measurements during the GST, external rotation stress test, and weightbearing in a cadaveric model of SER ankle injuries to (1) identify the relationship between the ultrasonographic MCS measurements and a sequentially created SER ankle injury model, (2) evaluate if the ultrasonography can detect a difference in MCS distance between an uninjured ankle and an SER injury type IV, and (3) investigate if the ultrasonographic MCS measurements can differentiate between the SER ankle injury with and without complete deltoid ligament rupture, as well as between the SER ankle injury with partial deltoid ligament injury and with complete deltoid ligament rupture. We hypothesize that there is a correlation between the ultrasonographic MCS measurements, and the sequentially created SER ankle injury model. We also hypothesize that there is an increase in ultrasonographic MCS distances between the uninjured ankle and the ankle with SER type IV injury, as well as between the SER ankle injury with and without complete deltoid ligament rupture, and between the SER ankle injury with partial deltoid ligament rupture and with complete deltoid ligament rupture when measured with the ultrasound.
Specimen preparation Ten fresh-frozen below-knee amputated cadaveric specimens with intact proximal tibiofibular joint. The mean age at the time of death was 46 (range 32 to 56) years. Five were males, and five were females. Specimens were thawed at room temperature 24 h prior to the start of the experiment. Bone and soft tissue were carefully handled and maintained to simulate in vivo conditions. Before testing, ankle fluoroscopic images (anteroposterior, lateral, and mortise) were obtained in each specimen. In case a specimen showed signs of previous ankle trauma or severe degenerative changes, the cadaver was excluded from the study. Sequential transection of ligaments or bones and loading conditions All specimens underwent an identical sequence of ligamentous and bony transection. The assessment was performed first with all ankle ligaments and fibula intact and later with sequential transection of the anterior inferior tibiofibular ligament (AITFL) (SER injury stage I), Weber B fibular fracture (SER injury stage II), the posterior inferior tibiofibular ligament (PITFL) (SER injury stage III), the superficial deltoid ligament (SER injury stage IVa), and the deep deltoid ligament (SER injury stage IVb) . In all scenarios, four loading conditions were considered, including a GST in ankle neutral position, a GST in ankle plantarflexed position, an external rotation stress test, and weightbearing. Description of the ultrasonographic gravity stress test To perform the GST, the specimen was placed in a lateral decubitus position with the most distal half of the leg, ankle, and foot off the end of the table, allowing the weight of the foot and ankle to create a lateral force across the ankle joint . The measurements were obtained with the ankle held in neutral dorsiflexion (0 degrees) as well as with the ankle plantarflexed (15 degrees) . An electronic goniometer was used to ensure that the ankle position was set and standardized in a proper position during the test (Fig. A and B). Description of the ultrasonographic weightbearing Simulated weightbearing with an axial load force of 750 N was performed. The amount of force used in this study was based on previous cadaveric studies . The 750 N axial loading force, which corresponds to 150 kg of weight for a two-legged stance, would represent the upper limit of weightbearing conditions for most individuals. The force was applied through a wooden block that was attached to the tibial plateau of the specimen (Fig. C). All applied forces were measured and standardized using a digital force gauge DFS2-R-0200 with an accuracy of 0.1% of full scale. Description of the ultrasonographic external rotation stress test Before testing, specimens were secured to a board beneath the leg with the use of two 5.0-mm Schanz pins placed from anterior to posterior into the proximal, middle third of the tibia. An external rotation stress test of the ankle was performed under 45Nm torque. The force was applied to the foot using a bone hook with the ankle positioned in neutral dorsiflexion. The bone hook was placed medially at the first metatarsal shaft (Fig. D). The 45 N (4.5Nm torque) used in this study was based on previous literature [ , , ], that concluded that 45 N was sufficient to detect the medial-sided ankle instability during the radiographic external rotation stress test without exceeding the level of rotational force that can cause a fibular fracture or further ligamentous damage . Ultrasonographic medial clear space measurement In all loading scenarios, medial side assessment was performed using a portable ultrasound device (2D-gray scale B mode, Butterfly IQ ultrasound device, Butterfly Network, USA). The ultrasound images were taken using Butterfly IQ-Ultrasound software Version 1.15.0. Subsequently, Image J program (NIH, Bethesda, Maryland, USA). was used to measure the MCS distances from the recorded P-US images. Three different MCS measurement values, including anterior-perpendicular-MCS, anterior-oblique-MCS, and inferior-MCS, were considered in this study accordingly to three planes of talar motion. During each stress maneuver, the anterior-perpendicular-MCS measured with the P-US represents the lateral shift of the talus. The anterior-oblique-MCS measured with the P-US represents the talar external rotation. The inferior-MCS measured with the P-US represents the talar eversion. Prior to P-US examination, the landmarks for ultrasound probe positioning were identified. After mounting the cadaveric. Specimens, the fluoroscopy was used to find the tibiotalar joint line level (Fig. A). A marker was then used to draw two lines, one along the joint line level and another at about 1 cm below and parallel to the joint line level (Fig. B). To measure the MCS using P-US, the line representing 1 cm below joint line level was used as a first landmark for assessing the MCS from the anteromedial aspect of the ankle joint with the middle of the P-US probe positioned on this line and perpendicular to the medial gutter. At this anteromedial landmark, the anterior-perpendicular-MCS and anterior-oblique-MCS were evaluated on a transverse plane using the P-US. Then, the fluoroscopy was used to define the second landmark as the furthest distance from the joint line level at which the medial malleolus still articulated with the talus. At this inferomedial landmark, the inferior-MCS distance was assessed on a coronal plane using the P-US with the middle of the probe positioned perpendicular to this defined point (Fig. C). When evaluated from the anteromedial aspect of the ankle joint (Fig. A), the anterior-perpendicular-MCS distance was measured from the transverse P-US images as represented by the perpendicular distance, drawn starting from the lateral border of the medial malleolus’s hyperechoic bone contours to the medial border of the talus’s hyperechoic bone contours (Fig. B and C). This anterior-perpendicular-MCS distance represents the lateral talar translation. With the same ultrasonographic image, the anterior-oblique-MCS distance was also measured in mm as represented by the oblique distance drawn starting from the lateral border of the medial malleolus’s hyperechoic bone contours to the anteromedial edge of the talus’s hyperechoic bone contours (Fig. D and E). This anterior-oblique-MCS distance represents the talar external rotation. When evaluated from the inferomedial aspect of the ankle joint (Fig. F), the inferior-MCS distances were measured from the coronal P-US images as represented by the perpendicular distances, drawn starting from the medial border of the talus’s hyperechoic bone contours at the level of the medial malleolar tip to the lateral border of the medial malleolus’s hyperechoic bone contours (Fig. G and H). This inferior-MCS distance represents the talar eversion.
Ten fresh-frozen below-knee amputated cadaveric specimens with intact proximal tibiofibular joint. The mean age at the time of death was 46 (range 32 to 56) years. Five were males, and five were females. Specimens were thawed at room temperature 24 h prior to the start of the experiment. Bone and soft tissue were carefully handled and maintained to simulate in vivo conditions. Before testing, ankle fluoroscopic images (anteroposterior, lateral, and mortise) were obtained in each specimen. In case a specimen showed signs of previous ankle trauma or severe degenerative changes, the cadaver was excluded from the study.
All specimens underwent an identical sequence of ligamentous and bony transection. The assessment was performed first with all ankle ligaments and fibula intact and later with sequential transection of the anterior inferior tibiofibular ligament (AITFL) (SER injury stage I), Weber B fibular fracture (SER injury stage II), the posterior inferior tibiofibular ligament (PITFL) (SER injury stage III), the superficial deltoid ligament (SER injury stage IVa), and the deep deltoid ligament (SER injury stage IVb) . In all scenarios, four loading conditions were considered, including a GST in ankle neutral position, a GST in ankle plantarflexed position, an external rotation stress test, and weightbearing.
To perform the GST, the specimen was placed in a lateral decubitus position with the most distal half of the leg, ankle, and foot off the end of the table, allowing the weight of the foot and ankle to create a lateral force across the ankle joint . The measurements were obtained with the ankle held in neutral dorsiflexion (0 degrees) as well as with the ankle plantarflexed (15 degrees) . An electronic goniometer was used to ensure that the ankle position was set and standardized in a proper position during the test (Fig. A and B).
Simulated weightbearing with an axial load force of 750 N was performed. The amount of force used in this study was based on previous cadaveric studies . The 750 N axial loading force, which corresponds to 150 kg of weight for a two-legged stance, would represent the upper limit of weightbearing conditions for most individuals. The force was applied through a wooden block that was attached to the tibial plateau of the specimen (Fig. C). All applied forces were measured and standardized using a digital force gauge DFS2-R-0200 with an accuracy of 0.1% of full scale.
Before testing, specimens were secured to a board beneath the leg with the use of two 5.0-mm Schanz pins placed from anterior to posterior into the proximal, middle third of the tibia. An external rotation stress test of the ankle was performed under 45Nm torque. The force was applied to the foot using a bone hook with the ankle positioned in neutral dorsiflexion. The bone hook was placed medially at the first metatarsal shaft (Fig. D). The 45 N (4.5Nm torque) used in this study was based on previous literature [ , , ], that concluded that 45 N was sufficient to detect the medial-sided ankle instability during the radiographic external rotation stress test without exceeding the level of rotational force that can cause a fibular fracture or further ligamentous damage .
In all loading scenarios, medial side assessment was performed using a portable ultrasound device (2D-gray scale B mode, Butterfly IQ ultrasound device, Butterfly Network, USA). The ultrasound images were taken using Butterfly IQ-Ultrasound software Version 1.15.0. Subsequently, Image J program (NIH, Bethesda, Maryland, USA). was used to measure the MCS distances from the recorded P-US images. Three different MCS measurement values, including anterior-perpendicular-MCS, anterior-oblique-MCS, and inferior-MCS, were considered in this study accordingly to three planes of talar motion. During each stress maneuver, the anterior-perpendicular-MCS measured with the P-US represents the lateral shift of the talus. The anterior-oblique-MCS measured with the P-US represents the talar external rotation. The inferior-MCS measured with the P-US represents the talar eversion. Prior to P-US examination, the landmarks for ultrasound probe positioning were identified. After mounting the cadaveric. Specimens, the fluoroscopy was used to find the tibiotalar joint line level (Fig. A). A marker was then used to draw two lines, one along the joint line level and another at about 1 cm below and parallel to the joint line level (Fig. B). To measure the MCS using P-US, the line representing 1 cm below joint line level was used as a first landmark for assessing the MCS from the anteromedial aspect of the ankle joint with the middle of the P-US probe positioned on this line and perpendicular to the medial gutter. At this anteromedial landmark, the anterior-perpendicular-MCS and anterior-oblique-MCS were evaluated on a transverse plane using the P-US. Then, the fluoroscopy was used to define the second landmark as the furthest distance from the joint line level at which the medial malleolus still articulated with the talus. At this inferomedial landmark, the inferior-MCS distance was assessed on a coronal plane using the P-US with the middle of the probe positioned perpendicular to this defined point (Fig. C). When evaluated from the anteromedial aspect of the ankle joint (Fig. A), the anterior-perpendicular-MCS distance was measured from the transverse P-US images as represented by the perpendicular distance, drawn starting from the lateral border of the medial malleolus’s hyperechoic bone contours to the medial border of the talus’s hyperechoic bone contours (Fig. B and C). This anterior-perpendicular-MCS distance represents the lateral talar translation. With the same ultrasonographic image, the anterior-oblique-MCS distance was also measured in mm as represented by the oblique distance drawn starting from the lateral border of the medial malleolus’s hyperechoic bone contours to the anteromedial edge of the talus’s hyperechoic bone contours (Fig. D and E). This anterior-oblique-MCS distance represents the talar external rotation. When evaluated from the inferomedial aspect of the ankle joint (Fig. F), the inferior-MCS distances were measured from the coronal P-US images as represented by the perpendicular distances, drawn starting from the medial border of the talus’s hyperechoic bone contours at the level of the medial malleolar tip to the lateral border of the medial malleolus’s hyperechoic bone contours (Fig. G and H). This inferior-MCS distance represents the talar eversion.
To assess the interobserver reliability of the P-US measurements, two orthopedic foot and ankle surgeons independently performed the MCS measurements in three randomly selected specimens. After three months, the recorded ultrasonographic MCS images of the three specimens were remeasured by the same orthopedic surgeons to assess the intraobserver reliability. From this data, the inter and intraobserver reliability was assessed using the interclass correlation coefficients (ICC) derived from a two-way mixed effects model analysis of variance for absolute agreement. A two-way mixed effects model was used, because the two observers were not randomly selected, and both observers scanned the same subjects. Interpretation of the ICC values were interpreted as follows: ICC < 0.4, poor; 0.4 < ICC < 0.59, acceptable; 0.6 < ICC < 0.74, good; and ICC > 0.74, excellent . Statistical analysis All MCS measurements were reported with median and interquartile range (IQR) in millimeter (mm). The data of the intact joint in each loading condition was designated as the baseline value. To investigate the correlation between the ultrasonographic MCS measurements and the SER ankle injury stages, a Spearman’s rank correlation was used. In order to achieve 90% statistical power for detecting a correlation with large effect size ( r =.5) between the ultrasonographic MCS measurements and the SER ankle injury stages with an overall two-tailed type-1 rate of 2.5%, we would need a minimum of 48 observations. In each specimen, the ultrasonographic MCS measurements were measured in the intact state, as well as in five SER ankle injury stages, resulting in six observations. Thus, to answer the hypothesis, we would need eight specimens. Accounting for 20% exclusion of specimen due to signs of previous ankle trauma or severe degenerative changes, the total amount included in this study was 10 specimens. To detect a difference in measured MCS distances for each stress test and each injury stage to the intact stage, a Wilcoxon signed-rank test was performed for each imaging modality. P values were adjusted for multiple comparison using the Holm-Bonferroni method. A 2-sided P value of less than 0.05 was considered statistically significant. The sample size calculation was based on the previous cadaveric study by Ashraf et al. investigated MCS values using radiographic images of the ankle in a neutral position in a gravity stress test condition and found a mean and standard deviation (SD) of 4.33 ± 0.72 mm for the SER ankle injury stage III-b (with only superficial deltoid rupture) and a mean and SD of 7.11 ± 1.03 mm for the SER ankle injury stage IV (with complete deltoid rupture). In order to achieve 95% statistical power for detecting a difference of 2.78 mm in MCS distances (4.33 ± 0.72 vs. 7.11 ± 1.03 mm, 0.3 correlation) among the MCS measurements, with an overall two-tailed Type-1 rate of 2.5% for a Wilcoxon signed-rank test, we need at least six specimens. The sample size calculation was performed using G*Power Version 3.1.9.4. All analyses were performed with Stata 13.0 for Mac (StataCorp LP, College Station, TX, USA).
All MCS measurements were reported with median and interquartile range (IQR) in millimeter (mm). The data of the intact joint in each loading condition was designated as the baseline value. To investigate the correlation between the ultrasonographic MCS measurements and the SER ankle injury stages, a Spearman’s rank correlation was used. In order to achieve 90% statistical power for detecting a correlation with large effect size ( r =.5) between the ultrasonographic MCS measurements and the SER ankle injury stages with an overall two-tailed type-1 rate of 2.5%, we would need a minimum of 48 observations. In each specimen, the ultrasonographic MCS measurements were measured in the intact state, as well as in five SER ankle injury stages, resulting in six observations. Thus, to answer the hypothesis, we would need eight specimens. Accounting for 20% exclusion of specimen due to signs of previous ankle trauma or severe degenerative changes, the total amount included in this study was 10 specimens. To detect a difference in measured MCS distances for each stress test and each injury stage to the intact stage, a Wilcoxon signed-rank test was performed for each imaging modality. P values were adjusted for multiple comparison using the Holm-Bonferroni method. A 2-sided P value of less than 0.05 was considered statistically significant. The sample size calculation was based on the previous cadaveric study by Ashraf et al. investigated MCS values using radiographic images of the ankle in a neutral position in a gravity stress test condition and found a mean and standard deviation (SD) of 4.33 ± 0.72 mm for the SER ankle injury stage III-b (with only superficial deltoid rupture) and a mean and SD of 7.11 ± 1.03 mm for the SER ankle injury stage IV (with complete deltoid rupture). In order to achieve 95% statistical power for detecting a difference of 2.78 mm in MCS distances (4.33 ± 0.72 vs. 7.11 ± 1.03 mm, 0.3 correlation) among the MCS measurements, with an overall two-tailed Type-1 rate of 2.5% for a Wilcoxon signed-rank test, we need at least six specimens. The sample size calculation was performed using G*Power Version 3.1.9.4. All analyses were performed with Stata 13.0 for Mac (StataCorp LP, College Station, TX, USA).
MCS values measured with the P-US increased as the SER ankle injury stage progressed. The Spearman’s rank correlation coefficient ranged from 0.43 to 0.90 ( P <.001), which indicate moderate to strong positive correlations between the ultrasonographic MCS measurements and the sequentially created supination-external rotation ankle injury model (Table ). All MCS values, including the anterior-perpendicular MCS, anterior-oblique MCS, and inferior MCS measured with the P-US during; (1) the GST in neutral ankle position, (2) the GST in plantarflexed ankle position, (3) weightbearing, and (4) the external rotation stress test, significantly increased between intact stage vs. stage IVb ( P =.036) (Table ). When compared between SER ankle injury stage III vs. IVb and stage IVa vs. IVb, the P-US MCS values measured during the GST and external rotation stress test significantly increased when the injury progressed from stage III to IVb ( P ranged from 0.015 to 0.031) or from IVa to IVb ( P ranged from 0.015 to 0.028) (Table ). Notably, MCS values measured with the P-US during weightbearing were significantly increased only between intact stage vs. stage IVb ( P =.036) and between stage III vs. stage IVb ( P ranged from 0.031 to 0.047), but not between stage III vs. IVb ( P ranged from 0.083 to 0.28). Interobserver (0.97; 95% confidence interval: 0.96 to 0.98) and intraobserver reliability (0.95; 95% confidence interval: 0.94 to 0.96) for the P-US MCS measurements were all substantial.
In recent years, dynamic P-US is increasingly being used to evaluate musculoskeletal injuries at the point of care. The objectives of this cadaveric study were to assess the relationship between the ultrasonographic MCS measurements and a sequentially created SER ankle injury model, as well as to determine whether the dynamic ultrasonography can detect medial side instability in SER type ankle fracture. We found moderate to strong positive correlations between the P-US MCS measurements and the sequentially created SER ankle injury model for the assessment of medial ankle instability. By assessing the MCS using P-US during the GST, weightbearing, or the external rotation stress test, SER ankle fracture with complete deltoid ligament rupture (IVb) can be differentiated from the uninjured ankle or other SER ankle injury stages (I to IVa). Results in the current study and data presented in previous literature confirmed that the deltoid ligaments contribute to tibiotalar joint stability on the medial side in the SER ankle fracture [ – , , ]. Most of the previous researches used radiographic imaging for assessing medial ankle instability in isolated fibular fracture. Correspondingly, we found that the MCS values as measured with the P-US increased as the SER ankle injury stage progressed and that these values significantly correlated with the SER injury staging ( P <.001) (Table ). Our findings underscore that ultrasonography has reached a level of technological maturity capable of evaluating medial side ankle injuries as an alternative to radiography. Prior studies have explored the feasibility and accuracy of using ultrasound to examine the deltoid ligament in the setting of ankle fractures [ – ]. These studies only examined the quality of the deltoid ligament, i.e., injury, including fluid, hematoma, discontinuity of the ligament, and evidence of articular pouch on the medial side of the ankle that approaches the tibialis posterior tendon. Although deltoid injury could be diagnosed with these ultrasonographic signs, this examination does not answer the fundamental question of whether the deltoid ligament injury has rendered the ankle unstable. With our P-US evaluating technique, three planes of talar motion can be evaluated. During the stress maneuvers, the anterior-perpendicular-MCS measured with the P-US represents the lateral shift of the talus. The anterior-oblique-MCS measured with the P-US represents the talar external rotation. The inferior-MCS measured with the P-US represents the talar eversion. Our results found that, as the staging of injury progress, multidirectional instability on the medial side of the ankle, including lateral talar shift, talar external rotation and talar tilting occur simultaneously (Table ). Notably, the correlations found between the MCS values measured during weightbearing vs. the injury staging were moderate (r ranged from 0.43 to 0.60), while the correlations were strong when the MCS values were measured during the GST and external rotation stress test (r ranged from 0.76 to 0.90). This is likely due to the difference in forces applied to the ankle during the stress maneuvers. According to the concept of Lauge-Hansen SER ankle fracture , the injury results from the rotational force that renders ligamentous and bony damage in a circular pattern starting from the lateral aspect, which is the AITFL to the medial aspect which is the deltoid ligament. As stress is applied to an injured ankle, the medial ankle instability, as represented by the MCS values, may gradually increase during the GST or external rotation stress test. These loading conditions likely simulate the ankle fractures mechanisms, which could result in a better correlation coefficient. In contrast, during weightbearing, the stability of the tibiotalar joint is likely provided by the bony congruency. The majority of force passes directly from the talar dome to the tibial plafond regardless of the presence of the fibular lateral buttress . Stewart et al. performed a cadaveric study to evaluate the effect of deltoid incompetence on the stability of ankle mortise with an applied axial loading . They found that the weightbearing radiographs cannot illustrate medial-sided ankle instability as relative to the forces applied with the GST or manual external rotation stress test. Although the correlation between the MCS values measured during weightbearing vs. the sequentially created SER ankle injury model was moderate, as the injury progressed to the last stage (IVb), the MCS values increased and became significantly larger when compared to the uninjured stage ( P =.036). One critical consideration for the effective clinical care of patients with isolated lateral malleolar fractures is assessing whether a concomitant deltoid ligament injury has rendered the tibiotalar joint unstable. Our study paves the way for using the P-US to diagnose destabilizing deltoid injuries via the ultrasonographic MCS measurements. The current study found that all MCS values as measured with the P-US during the GST in ankle neutral or plantarflexed position, during weightbearing, and during the external rotation stress test were significantly different between the uninjured stage vs. SER injury stage IVb ( P =.036), and between SER injury stage III vs. IVb ( P ranged from 0.015 to 0.047). These findings highlight the capability of the dynamic ultrasonography for diagnosing SER ankle fracture with complete deltoid ligament rupture, as well as the ability to differentiate the unstable SER injury from the intact state and the stable injuries. Previously, studies have demonstrated that physical examination, such as medial ecchymosis, swelling or tenderness, and static radiography are not accurate for diagnosing incompetence of the deltoid ligament in SER ankle fracture [ – ]. Therefore, the stress radiography during the GST, weightbearing, or the manual external rotation stress test is usually recommended [ , , – ]. However, stress radiography may not readily be available in out-patients clinics or resource-limited settings, and the test itself leads to a significant amount of radiation exposure to both patients and examiners. In contrast, the P-US, which is radiation-free and readily available in portable mode [ , – ], can be used to evaluate medial ankle instability at the point of care, as well as to assess the outcome after treatment and progression of the instability during the follow-up visit. Interestingly, the MCS values measured with the P-US during the GST and the external rotation stress test significantly increased from the intact stage as the SER injury progressed to stage I or II ( P ranged from 0.014 to 0.044). Based on these findings, the ultrasonographic GST and external rotation stress test may be too sensitive for estimating the medial ankle instability. Previous literature also supported this finding. A study by Koval et al. reported a cohort of 21 patients with SER stage IV ankle fracture who were evaluated with an external rotation stress radiograph. They found that 90% of patients had only partial deltoid ligament tear when confirmed with magnetic resonance imaging. Recently, many studies found that the GST or external rotation stress radiographs may overdiagnose the unstable SER ankle fracture patients which could lead to an unnecessary surgery [ – ]. These studies compared the results of operative vs. nonoperative treatment in unstable SER ankle fractures that were diagnosed with the GST or external rotation stress radiographs. They found equivalent functional outcomes of operative and nonoperative management as assessed using patient-based outcome measurements, which raises a question on whether a positive GST or external rotation stress test is really an indication for surgery, or could it be that these stress maneuvers may overestimate the unstable injury. In contrast, during weightbearing, the MCS values as measured with the P-US became significantly larger when compared to the uninjured stage only after the complete deltoid ligament ruptured (IVb) ( P =.036), while there is no significantly increased in MCS values when the injury progress to stage III or stage IVa (P ranged from 0.056 to 0.89). The weightbearing stress ultrasonography may better predict the medial ankle instability and could be more specific to the SER ankle fracture with medial ankle instability. Several studies have demonstrated the utility of weightbearing radiographs for assessing mortise stability in SER ankle injury [ , , , ]. If mortise alignment and stability are achieved during weightbearing radiographs, even with positive GST or external rotation stress test, patients will still do well with nonoperative treatment. A recent article review by Kwon et al. also underscored the concept of mortise stability during the weightbearing radiograph, which can be used as a guide for a successful nonoperative treatment and avoid unnecessary surgery in patients with SER ankle fracture. Our findings and the evidence from previous literature highlighted the capability and values of the weightbearing stress for assessing medial ankle instability in SER ankle fracture. The interobserver and intraobserver reliability for all ultrasonographic MCS measurements were all excellent, which represents that the ultrasound assessment is reliable and reproducible for the medial ankle instability evaluation. Our study has several limitations. First, being a cadaveric study, the soft tissue conditions differ from those seen in-vivo. However, to simulate in vivo conditions as best possible, specimen bone and soft tissue were carefully maintained. Second, our measurement technique involved an operator learning curve inherent to any new technology. Even though we aimed to develop a reliable assessment technique, an experienced operator is still needed for accurate image acquisition as changes or tilting in probe position may potentially cause measurement variability of MCS cortical margins secondary to off-axis transverse imaging of the MCS resulting in falsely elevated oblique transverse measurements. However, in the hands of an experienced operator, sonographic images can provide useful information without radiation or any other contraindication. Our experience is that an orthopaedic surgeon’s knowledge of anatomy far supersedes any technical impediment to mastering this new technology. Third, we were unable to measure the tibiotalar plafond joint or the superior clear space as a relative comparison point for the MCS. This space is difficult to measure with the ultrasound since the shape of the superior clear space is a curved surface formed by the talar dome and distal end of the tibia. The sound wave from the ultrasound transducer is obscure by the distal tibia’s anterior ridge, thus preventing us from getting the clear images at the dome of the talus. Finally, to adequately assess medial ankle instability using dynamic P-US, stress to the ankle is required, which might not be tolerated by the patient in the clinical setting. To address this, a care provider may prescribe pain medication or local analgesia prior to P-US evaluation.
The use of dynamic stress ultrasonography for diagnosing medial ankle instability in SER ankle injury appears to be a reliable and repeatable technique. The MCS measurements assessed with P-US during the GST, weightbearing, and the external rotation stress test significantly correlated with the SER ankle injury staging ( P <.001). In addition, the P-US method is capable of differentiating the SER ankle injury stage IVb from the intact stage, as well as differentiating the stable SER ankle injury stage from the unstable stage. Therefore, the P-US can be a valuable diagnostic tool at the point of care due to its ability to dynamically evaluate suspected medial ankle instability in SER type injury.
|
INCREASED RISK OF CYSTOID MACULAR EDEMA AFTER CATARACT SURGERY IN EYES PREVIOUSLY VITRECTOMIZED FOR IDIOPATHIC EPIRETINAL MEMBRANE | 5363e4e6-3941-48b6-b5a0-c5c7dd8e5473 | 11753465 | Surgical Procedures, Operative[mh] | In this longitudinal retrospective study, we reviewed the data of all the institutional patients operated from January 2016 to December 2021 at the Vitreoretinal Surgery Division of San Raffaele Scientific Institute, a tertiary referral center. Ethics All the procedures performed involving human participants were in accordance with the ethical standards of the Institutional Review Board of the San Raffaele Scientific Institute and with the 1964 Declaration of Helsinki and its later amendments. All patients, at the time of hospital admission, signed a general informed consent form that was specifically designed and approved by the Institutional Review Board of the San Raffaele Scientific Institute solely for observational retrospective studies. Study Participants Inclusion criteria Adult patients scheduled for vitrectomy for ERM or RRD. Cases undergoing an isolated vitreoretinal procedure (excluding combined phacovitrectomies). Exclusion criteria Pseudophakic eyes. Spherical equivalent >−6.5 diopters or axial length >26 mm. Secondary ERM (e.g., diabetic retinopathy, uveitis, retinal vein occlusion, retinal dystrophies, and trauma). Uncompensated glaucoma. Complicated cases of RRD, including: traumatic or relapsed cases, presence of proliferative vitreoretinopathy > stage A, giant retinal tear, retinoschisis-associated RRD, coloboma-associated retinal detachment. Postvitrectomy complications, such as RRD recurrence, vascular occlusions, optic neuropathies, etc. Imaging acquisition jeopardized by optical media opacity (cornea or lens). Individuals with diabetes without retinopathy. Study Protocol We longitudinally evaluated the frequency of postsurgical CME following uncomplicated vitreoretinal procedures, grouping eyes according to the retinal pathologic conditions (ERM and RRD). Eyes later undergoing uncomplicated phacoemulsification for visually impairing cataracts were assessed for CME within the first 6 months after cataract surgery. All institutional patients underwent a preoperative ophthalmologic evaluation, then standard controls at 1 week and at 1, 3, 6 and 12 months, then as required per clinician's decision. Spectral-domain optical coherence tomography were performed on every visit except at 1 week. For this reason, this point of observation was excluded from the analysis. Surgical Procedure Surgeries were conducted under retrobulbar block by senior vitreoretinal consultants (M.C., L.I.) using the Constellation machine (Alcon Surgical, Fort Worth, TX), with a 23 gauge or 25 gauge according to the surgeon's preference. Perfluorocarbon liquid (perfluorodecalin; Bausch+Lomb, Rochester, NY) was used at the surgeon's discretion. Membranes (ERM and internal limiting membrane) were peeled when needed with the conjugated dye TWIN (AL.CHI.MI.A., Italy). To facilitate sclerotomy sealing, fluid–air exchange was routinely done at the end of all macular surgeries and, if required, air–gas exchange was also performed in selected cases. For tamponade with gas, either C 3 F 8 or SF 6 (AL.CHI.MI.A. S.R.L., Italy) was used, respectively, diluted at 6% and 10%. Phacoemulsification was carried out under topical anesthesia using the Centurion machine (Alcon Surgical), with the same standard surgical procedure. Data Collection Data gathered during preoperative examination included general and ocular medical history, distance best-corrected visual acuity (BCVA), and spectral-domain optical coherence tomography. Best-corrected visual acuity is regularly measured in dedicated offices with continuously monitored luminance, using the Early Treatment in Diabetic Retinopathy Study charts. The Snellen notation was converted into logMAR for statistical purposes. Surgery-related variables for idiopathic ERMs included stage (Govetto classification ), number of dye injections, and surgery duration. Surgery-related variables for RRD were macula status, number of involved quadrants, type of tamponade and retinopexy, use of perfluorocarbon liquid, and surgery duration. Spectral-domain optical coherence tomography images were acquired using the Spectralis device (Heidelberg Engineering, Heidelberg, Germany). Macular raster scans were acquired using the follow-up function. The central circle on the Early Treatment in Diabetic Retinopathy Study map was considered as the central foveal thickness (CFT) value. The presence of CME was assigned only in cases that showed, within the first 6 months of follow-up: 1) a new onset of macular hyporeflective cystoid intraretinal spaces at spectral-domain optical coherence tomography that were not present before vitrectomy (Figure ); 2) definite worsening in size and number of pre-existing cystoid spaces with an increase in CFT of 10% or more over baseline. Medication Standard therapy after vitreoretinal procedure was association of netilmicin and dexamethasone 4 times a day for 3 weeks and tropicamide twice a day for 1 week; after cataract extraction was the combination fluoroquinolone or chloramphenicol and dexamethasone 4 times a day for 1 week, followed by nonsteroidal anti-inflammatory agents for 3 weeks. The treatment regimen for the management of CME in our center is dexamethasone 0.2% drops 4 times a day plus bromfenac 0.9 mg/mL drops twice a day for 4 weeks. According to the clinician's judgement, treatment can be halted, tapered, or switched to second-line agents that includes intravitreal steroids (triamcinolone acetonide 80 mg/mL or dexamethasone implant 700 µ g). Statistical Analysis Statistical analyses were performed using Past (Palaeontologia Electronica) version 4.11. The alternative hypothesis H1 was that the CME rate after cataract surgery might statistically differ between the study groups. Based on previously published data, , , , , we calculated that to detect, between the study groups, a clinically significant difference of >15% in the CME rate (effect size), with an 80% power and α = 0.05, at least 23 patients per group would have been required. Distribution normality was tested with the Shapiro–Wilk test. The Chi-squared or Fisher exact test was used to compare noncontinuous variables. When parametric analysis was possible, the t -test for unpaired data or the one-way analysis of variance were used for comparisons between groups, whereas the Mann–Whitney U test or the Kruskal–Wallis test was applied to assess the significance of such differences when parametric analysis was not possible. Dunn post hoc with Bonferroni corrected P values was adopted to assess the differences between pairwise groups. In all analyses, the Bonferroni correction was adopted for P values, and only <0.05 were considered significant.
All the procedures performed involving human participants were in accordance with the ethical standards of the Institutional Review Board of the San Raffaele Scientific Institute and with the 1964 Declaration of Helsinki and its later amendments. All patients, at the time of hospital admission, signed a general informed consent form that was specifically designed and approved by the Institutional Review Board of the San Raffaele Scientific Institute solely for observational retrospective studies.
Inclusion criteria Adult patients scheduled for vitrectomy for ERM or RRD. Cases undergoing an isolated vitreoretinal procedure (excluding combined phacovitrectomies). Exclusion criteria Pseudophakic eyes. Spherical equivalent >−6.5 diopters or axial length >26 mm. Secondary ERM (e.g., diabetic retinopathy, uveitis, retinal vein occlusion, retinal dystrophies, and trauma). Uncompensated glaucoma. Complicated cases of RRD, including: traumatic or relapsed cases, presence of proliferative vitreoretinopathy > stage A, giant retinal tear, retinoschisis-associated RRD, coloboma-associated retinal detachment. Postvitrectomy complications, such as RRD recurrence, vascular occlusions, optic neuropathies, etc. Imaging acquisition jeopardized by optical media opacity (cornea or lens). Individuals with diabetes without retinopathy.
Adult patients scheduled for vitrectomy for ERM or RRD. Cases undergoing an isolated vitreoretinal procedure (excluding combined phacovitrectomies).
Pseudophakic eyes. Spherical equivalent >−6.5 diopters or axial length >26 mm. Secondary ERM (e.g., diabetic retinopathy, uveitis, retinal vein occlusion, retinal dystrophies, and trauma). Uncompensated glaucoma. Complicated cases of RRD, including: traumatic or relapsed cases, presence of proliferative vitreoretinopathy > stage A, giant retinal tear, retinoschisis-associated RRD, coloboma-associated retinal detachment. Postvitrectomy complications, such as RRD recurrence, vascular occlusions, optic neuropathies, etc. Imaging acquisition jeopardized by optical media opacity (cornea or lens). Individuals with diabetes without retinopathy.
We longitudinally evaluated the frequency of postsurgical CME following uncomplicated vitreoretinal procedures, grouping eyes according to the retinal pathologic conditions (ERM and RRD). Eyes later undergoing uncomplicated phacoemulsification for visually impairing cataracts were assessed for CME within the first 6 months after cataract surgery. All institutional patients underwent a preoperative ophthalmologic evaluation, then standard controls at 1 week and at 1, 3, 6 and 12 months, then as required per clinician's decision. Spectral-domain optical coherence tomography were performed on every visit except at 1 week. For this reason, this point of observation was excluded from the analysis.
Surgeries were conducted under retrobulbar block by senior vitreoretinal consultants (M.C., L.I.) using the Constellation machine (Alcon Surgical, Fort Worth, TX), with a 23 gauge or 25 gauge according to the surgeon's preference. Perfluorocarbon liquid (perfluorodecalin; Bausch+Lomb, Rochester, NY) was used at the surgeon's discretion. Membranes (ERM and internal limiting membrane) were peeled when needed with the conjugated dye TWIN (AL.CHI.MI.A., Italy). To facilitate sclerotomy sealing, fluid–air exchange was routinely done at the end of all macular surgeries and, if required, air–gas exchange was also performed in selected cases. For tamponade with gas, either C 3 F 8 or SF 6 (AL.CHI.MI.A. S.R.L., Italy) was used, respectively, diluted at 6% and 10%. Phacoemulsification was carried out under topical anesthesia using the Centurion machine (Alcon Surgical), with the same standard surgical procedure.
Data gathered during preoperative examination included general and ocular medical history, distance best-corrected visual acuity (BCVA), and spectral-domain optical coherence tomography. Best-corrected visual acuity is regularly measured in dedicated offices with continuously monitored luminance, using the Early Treatment in Diabetic Retinopathy Study charts. The Snellen notation was converted into logMAR for statistical purposes. Surgery-related variables for idiopathic ERMs included stage (Govetto classification ), number of dye injections, and surgery duration. Surgery-related variables for RRD were macula status, number of involved quadrants, type of tamponade and retinopexy, use of perfluorocarbon liquid, and surgery duration. Spectral-domain optical coherence tomography images were acquired using the Spectralis device (Heidelberg Engineering, Heidelberg, Germany). Macular raster scans were acquired using the follow-up function. The central circle on the Early Treatment in Diabetic Retinopathy Study map was considered as the central foveal thickness (CFT) value. The presence of CME was assigned only in cases that showed, within the first 6 months of follow-up: 1) a new onset of macular hyporeflective cystoid intraretinal spaces at spectral-domain optical coherence tomography that were not present before vitrectomy (Figure ); 2) definite worsening in size and number of pre-existing cystoid spaces with an increase in CFT of 10% or more over baseline.
Standard therapy after vitreoretinal procedure was association of netilmicin and dexamethasone 4 times a day for 3 weeks and tropicamide twice a day for 1 week; after cataract extraction was the combination fluoroquinolone or chloramphenicol and dexamethasone 4 times a day for 1 week, followed by nonsteroidal anti-inflammatory agents for 3 weeks. The treatment regimen for the management of CME in our center is dexamethasone 0.2% drops 4 times a day plus bromfenac 0.9 mg/mL drops twice a day for 4 weeks. According to the clinician's judgement, treatment can be halted, tapered, or switched to second-line agents that includes intravitreal steroids (triamcinolone acetonide 80 mg/mL or dexamethasone implant 700 µ g).
Statistical analyses were performed using Past (Palaeontologia Electronica) version 4.11. The alternative hypothesis H1 was that the CME rate after cataract surgery might statistically differ between the study groups. Based on previously published data, , , , , we calculated that to detect, between the study groups, a clinically significant difference of >15% in the CME rate (effect size), with an 80% power and α = 0.05, at least 23 patients per group would have been required. Distribution normality was tested with the Shapiro–Wilk test. The Chi-squared or Fisher exact test was used to compare noncontinuous variables. When parametric analysis was possible, the t -test for unpaired data or the one-way analysis of variance were used for comparisons between groups, whereas the Mann–Whitney U test or the Kruskal–Wallis test was applied to assess the significance of such differences when parametric analysis was not possible. Dunn post hoc with Bonferroni corrected P values was adopted to assess the differences between pairwise groups. In all analyses, the Bonferroni correction was adopted for P values, and only <0.05 were considered significant.
Population Figure reviews the protocol enrolling process. We found 1,247 eyes of 1,199 patients operated for ERM, of which 187 eyes of 180 patients were eventually eligible for the study. Among 862 cases of phakic primary RRD, 311 eyes of 305 patients fulfilled the enrollment criteria. Rhegmatogenous retinal detachment group turned out to be slightly younger (Table ). Surgeries All ERM cases included in the study were idiopathic and phakic, as per protocol. The distribution of ERM stages was as follows: 12 (6.4%) in 1, 35 (18.7%) in 2, 125 (66.8%) in 3, 15 (8%) in 4. At the conclusion of surgery, 156 eyes (83.4%) received air, 21 (11.2%) balanced saline solution, and the remaining 10 eyes (5.3%) SF 6 because of the intraoperative detection of rhegmatogenous lesions. All the latter cases were treated with intraoperative laser retinopexy. Average surgery duration was 23 ± 4 minutes. The number of required dye injection was 1.13 ± 0.41. All the included RRD cases confirmed to be, as per protocol, noncomplicated and with a 100% single-surgery attachment rate. The majority (n = 256, 82.3%) were macula on. In total, 92.6% (n = 288) received C 3 F 8 and 7.4% (n = 23) received SF 6 . No eye was treated with silicone oil in our series. Perfluorocarbon liquid was used in 280 procedures (90.0%). The most employed retinopexy technique turned out to be cryo (n = 187, 60.1%), followed by a combination of laser + cryo (n = 85, 27.3%), and by laser only (n = 39, 12.5%). Internal limiting membrane peeling was never performed in our RRD series. Mean surgery duration was 41 ± 13 minutes. We found 58 eyes out of those vitrectomized for ERM and 99 eyes out of those vitrectomized for RRD that underwent uncomplicated phacoemulsification (Figure ). Phacoemulsification was done 485 ± 425 days (∼1.3 years) after vitrectomy (range 27–1829; median 390), without time differences between groups. Postsurgical Cystoid Macular Edema After isolated vitreoretinal procedure Because only a few RRD cases were presurgically examined with spectral-domain optical coherence tomography, the baseline CFT data of these eyes were not considered. The CFT of these RRD eyes remained unchanged from 1 month till the time of cataract extraction. Eyes with ERM although disclosed a continuously reducing CFT after vitrectomy till the time of phacoemulsification (see Table , Supplemental Digital Content 1 , http://links.lww.com/IAE/C413 ). At the time of cataract surgery, eyes vitrectomized for ERM disclosed a thicker macula than those for RRD ( P < 0.0001). The rate of postvitrectomy CME turned out to be higher in eyes with ERM (12.8%, n = 24) compared with eyes operated for RRD (1%, n = 3; P < 0.0001) (Figure A). Out of the 3 eyes disclosing CME after RRD repair, all were macula on, and retinopexy was done with cryo. Macular edema was found in all cases at 1 month. Out to the 24 eyes that presented CME after ERM peeling, 5 were in Stage 4 and 19 were in Stage 3. Cystoid macular edema was found in 17 cases (70.8%) at the 1-month control and in 7 eyes (29.1%) at the 3-month control. After cataract surgery (vitrectomized eyes) The CME rate turned out to be significantly higher in eyes previously vitrectomized for ERM (13.8%, n = 8) compared with eyes previously vitrectomized for a primary noncomplicated RRD (2%, n = 2; P = 0.0055) (Figure B). The odds ratio of developing CME is 7.76 (95% CI, 1.588–37.93; P = 0.0113) in eyes previously vitrectomized for ERM. Macular edema was found in all cases at the 1-month visit. Out of these 8 eyes (previously vitrectomized for ERM) that presented postcataract CME, 7 (87.5%) were those that already had CME after the vitreoretinal procedure. The 2 eyes previously vitrectomized for RRD that disclosed postcataract CME had, on the contrary, a new onset of macular edema. The average CFT of eyes previously operated for ERM showed a significant, but moderate, increase and decrease, remaining overall stable between the precataract time and the last follow-up ( P = 0.9999). The CFT of eyes previously operated for RRD remained overall stable after cataract surgery (see Table , Supplemental Digital Content 2 , http://links.lww.com/IAE/C414 ). Treatment All the eyes that developed CME after ERM surgery received an initial topical therapy according to the reported regimen. Most (91.6%, n = 22) gradually and completely resolved the edema with topical therapy within 6 months from presentation. Two eyes were treated with a single 700 µ g dexamethasone implant injection, done after 2 months of topical therapy. One eye resolved the edema, whereas the second disclosed an overall chronic and poorly responsive edema that was found to be still present at the last available follow-up visit also after phacoemulsification. Vitrectomized eyes for ERM experiencing CME after cataract surgery all resolved their fluid just with topical therapy. All the eyes experiencing CME after RRD repair were accordingly treated, as per protocol, with topical therapy, and completely resolved the edema. The eyes that presented CME after cataract surgery similarly recovered just with topical therapy. Visual Acuity After vitrectomy, the uneventful ERM cases (no CME) continuously improved vision from baseline to 6 months ( P < 0.0001 per each pairwise comparison). By contrast, the ERMs presenting postvitrectomy CME showed a static trend, with only a moderate improvement at the 6-month control ( P = 0.03478). Its BCVA turned out indeed significantly worse compared with the no-CME group in any time-point assessment (Figure ). Best-corrected visual acuity after phacoemulsification kept significantly worse in eyes disclosing CME ( P = 0.0001). Average baseline BCVA in the RRD group was 0.15 ± 0.32 logMAR, which, in turn, was higher in the macula-on (0.00 ± 0.02 logMAR) compared with the macula-off subgroup (0.84 ± 0.10 logMAR; P < 0.0001). Owing to the reduced number of eyes presenting postvitrectomy and postcataract CME in the RRD group, these data were not averaged. Regression Analysis The logistic regression analysis found that higher ERM stages were associated with an increased rate of postsurgical CME (coefficient 1.5476, odds ratio 4.7 [95% CI, 1.9098–11.5671], P = 0.0008). The intraoperative use of laser was also found to be associated with CME occurrence ( P < 0.0001). No other previtrectomy characteristics (Table ) were associated with the chance of CME occurrence, neither in ERM nor in RRD. The occurrence of postvitrectomy CME was found to be associated with an increased risk of post-phacoemulsification CME (coefficient 2.3507, odds ratio 13.588 [95% CI, 1.543–119.7], P = 0.0187).
Figure reviews the protocol enrolling process. We found 1,247 eyes of 1,199 patients operated for ERM, of which 187 eyes of 180 patients were eventually eligible for the study. Among 862 cases of phakic primary RRD, 311 eyes of 305 patients fulfilled the enrollment criteria. Rhegmatogenous retinal detachment group turned out to be slightly younger (Table ).
All ERM cases included in the study were idiopathic and phakic, as per protocol. The distribution of ERM stages was as follows: 12 (6.4%) in 1, 35 (18.7%) in 2, 125 (66.8%) in 3, 15 (8%) in 4. At the conclusion of surgery, 156 eyes (83.4%) received air, 21 (11.2%) balanced saline solution, and the remaining 10 eyes (5.3%) SF 6 because of the intraoperative detection of rhegmatogenous lesions. All the latter cases were treated with intraoperative laser retinopexy. Average surgery duration was 23 ± 4 minutes. The number of required dye injection was 1.13 ± 0.41. All the included RRD cases confirmed to be, as per protocol, noncomplicated and with a 100% single-surgery attachment rate. The majority (n = 256, 82.3%) were macula on. In total, 92.6% (n = 288) received C 3 F 8 and 7.4% (n = 23) received SF 6 . No eye was treated with silicone oil in our series. Perfluorocarbon liquid was used in 280 procedures (90.0%). The most employed retinopexy technique turned out to be cryo (n = 187, 60.1%), followed by a combination of laser + cryo (n = 85, 27.3%), and by laser only (n = 39, 12.5%). Internal limiting membrane peeling was never performed in our RRD series. Mean surgery duration was 41 ± 13 minutes. We found 58 eyes out of those vitrectomized for ERM and 99 eyes out of those vitrectomized for RRD that underwent uncomplicated phacoemulsification (Figure ). Phacoemulsification was done 485 ± 425 days (∼1.3 years) after vitrectomy (range 27–1829; median 390), without time differences between groups.
After isolated vitreoretinal procedure Because only a few RRD cases were presurgically examined with spectral-domain optical coherence tomography, the baseline CFT data of these eyes were not considered. The CFT of these RRD eyes remained unchanged from 1 month till the time of cataract extraction. Eyes with ERM although disclosed a continuously reducing CFT after vitrectomy till the time of phacoemulsification (see Table , Supplemental Digital Content 1 , http://links.lww.com/IAE/C413 ). At the time of cataract surgery, eyes vitrectomized for ERM disclosed a thicker macula than those for RRD ( P < 0.0001). The rate of postvitrectomy CME turned out to be higher in eyes with ERM (12.8%, n = 24) compared with eyes operated for RRD (1%, n = 3; P < 0.0001) (Figure A). Out of the 3 eyes disclosing CME after RRD repair, all were macula on, and retinopexy was done with cryo. Macular edema was found in all cases at 1 month. Out to the 24 eyes that presented CME after ERM peeling, 5 were in Stage 4 and 19 were in Stage 3. Cystoid macular edema was found in 17 cases (70.8%) at the 1-month control and in 7 eyes (29.1%) at the 3-month control. After cataract surgery (vitrectomized eyes) The CME rate turned out to be significantly higher in eyes previously vitrectomized for ERM (13.8%, n = 8) compared with eyes previously vitrectomized for a primary noncomplicated RRD (2%, n = 2; P = 0.0055) (Figure B). The odds ratio of developing CME is 7.76 (95% CI, 1.588–37.93; P = 0.0113) in eyes previously vitrectomized for ERM. Macular edema was found in all cases at the 1-month visit. Out of these 8 eyes (previously vitrectomized for ERM) that presented postcataract CME, 7 (87.5%) were those that already had CME after the vitreoretinal procedure. The 2 eyes previously vitrectomized for RRD that disclosed postcataract CME had, on the contrary, a new onset of macular edema. The average CFT of eyes previously operated for ERM showed a significant, but moderate, increase and decrease, remaining overall stable between the precataract time and the last follow-up ( P = 0.9999). The CFT of eyes previously operated for RRD remained overall stable after cataract surgery (see Table , Supplemental Digital Content 2 , http://links.lww.com/IAE/C414 ).
Because only a few RRD cases were presurgically examined with spectral-domain optical coherence tomography, the baseline CFT data of these eyes were not considered. The CFT of these RRD eyes remained unchanged from 1 month till the time of cataract extraction. Eyes with ERM although disclosed a continuously reducing CFT after vitrectomy till the time of phacoemulsification (see Table , Supplemental Digital Content 1 , http://links.lww.com/IAE/C413 ). At the time of cataract surgery, eyes vitrectomized for ERM disclosed a thicker macula than those for RRD ( P < 0.0001). The rate of postvitrectomy CME turned out to be higher in eyes with ERM (12.8%, n = 24) compared with eyes operated for RRD (1%, n = 3; P < 0.0001) (Figure A). Out of the 3 eyes disclosing CME after RRD repair, all were macula on, and retinopexy was done with cryo. Macular edema was found in all cases at 1 month. Out to the 24 eyes that presented CME after ERM peeling, 5 were in Stage 4 and 19 were in Stage 3. Cystoid macular edema was found in 17 cases (70.8%) at the 1-month control and in 7 eyes (29.1%) at the 3-month control.
The CME rate turned out to be significantly higher in eyes previously vitrectomized for ERM (13.8%, n = 8) compared with eyes previously vitrectomized for a primary noncomplicated RRD (2%, n = 2; P = 0.0055) (Figure B). The odds ratio of developing CME is 7.76 (95% CI, 1.588–37.93; P = 0.0113) in eyes previously vitrectomized for ERM. Macular edema was found in all cases at the 1-month visit. Out of these 8 eyes (previously vitrectomized for ERM) that presented postcataract CME, 7 (87.5%) were those that already had CME after the vitreoretinal procedure. The 2 eyes previously vitrectomized for RRD that disclosed postcataract CME had, on the contrary, a new onset of macular edema. The average CFT of eyes previously operated for ERM showed a significant, but moderate, increase and decrease, remaining overall stable between the precataract time and the last follow-up ( P = 0.9999). The CFT of eyes previously operated for RRD remained overall stable after cataract surgery (see Table , Supplemental Digital Content 2 , http://links.lww.com/IAE/C414 ).
All the eyes that developed CME after ERM surgery received an initial topical therapy according to the reported regimen. Most (91.6%, n = 22) gradually and completely resolved the edema with topical therapy within 6 months from presentation. Two eyes were treated with a single 700 µ g dexamethasone implant injection, done after 2 months of topical therapy. One eye resolved the edema, whereas the second disclosed an overall chronic and poorly responsive edema that was found to be still present at the last available follow-up visit also after phacoemulsification. Vitrectomized eyes for ERM experiencing CME after cataract surgery all resolved their fluid just with topical therapy. All the eyes experiencing CME after RRD repair were accordingly treated, as per protocol, with topical therapy, and completely resolved the edema. The eyes that presented CME after cataract surgery similarly recovered just with topical therapy.
After vitrectomy, the uneventful ERM cases (no CME) continuously improved vision from baseline to 6 months ( P < 0.0001 per each pairwise comparison). By contrast, the ERMs presenting postvitrectomy CME showed a static trend, with only a moderate improvement at the 6-month control ( P = 0.03478). Its BCVA turned out indeed significantly worse compared with the no-CME group in any time-point assessment (Figure ). Best-corrected visual acuity after phacoemulsification kept significantly worse in eyes disclosing CME ( P = 0.0001). Average baseline BCVA in the RRD group was 0.15 ± 0.32 logMAR, which, in turn, was higher in the macula-on (0.00 ± 0.02 logMAR) compared with the macula-off subgroup (0.84 ± 0.10 logMAR; P < 0.0001). Owing to the reduced number of eyes presenting postvitrectomy and postcataract CME in the RRD group, these data were not averaged.
The logistic regression analysis found that higher ERM stages were associated with an increased rate of postsurgical CME (coefficient 1.5476, odds ratio 4.7 [95% CI, 1.9098–11.5671], P = 0.0008). The intraoperative use of laser was also found to be associated with CME occurrence ( P < 0.0001). No other previtrectomy characteristics (Table ) were associated with the chance of CME occurrence, neither in ERM nor in RRD. The occurrence of postvitrectomy CME was found to be associated with an increased risk of post-phacoemulsification CME (coefficient 2.3507, odds ratio 13.588 [95% CI, 1.543–119.7], P = 0.0187).
Owing to its practical convenience, combining phacoemulsification with vitreoretinal surgery has become increasingly common. Although this combined approach is elegant and effective, it may expose patients to an increased risk of CME, as both surgeries carry predisposing factors for this condition. In particular, it has been shown that performing just cataract surgery in eyes affected by idiopathic ERM increases the risk of macular edema up to 15.7%. A logical strategy to reduce this risk is to perform vitrectomy first, followed by cataract surgery at a later date. Although this option may seem reasonable, it carries a high likelihood of rapid cataract development, with incidences of up to 50% after the first year and 85% within 5 years. Furthermore, whether considering the latter option, it shall also be considered that phacoemulsification done in a vitrectomized eye for ERM itself increases the risk of CME. Being both strategies burdened by the chance of postsurgical CME, the management of phacoemulsification associated with vitreoretinal surgery remains hence controversial. In our study, we observed a CME rate of 13.8% after uncomplicated phacoemulsification in eyes vitrectomized for idiopathic ERM, which is comparable with the 12.8% rate after isolated vitreoretinal procedures. These frequencies align with those previously published by Frisina and Iuliano. A recent meta-analysis, furthermore, also supports the equivalence, in CME occurrence, between sequential and combined surgeries. Therefore, given these data, the option of combined surgeries (phacovitrectomy) in idiopathic ERMs seems to be a reasonable option because it exposes patients to a single surgical stress, hence to a single inflammatory “shot.” This consideration shall be taken though cautiously, especially in relation to advanced-stage ERMs, where the risk of CME after surgery can be as high as 57.1%. By contrast, the risk of CME after RRD repair is less straightforward. Reported CME rates in the literature indeed vary widely, from 9.6% to 31%. , – In our series, we although observed a lower rate (1%) of CME after primary noncomplicated RRD repair, which may appear noticeably conflicting. This discrepancy may be explained by our study's exclusion of more complex RRD cases, such as those with severe proliferative vitreoretinopathy or recurrent detachments (with multiple surgeries), factors known to increase the risk of CME. By intentionally excluding such complicated cases, we aimed to provide a more accurate estimate of the CME rate for uncomplicated RRD, which justifies the difference in our findings. A recent study by Souissi seems indeed to confirm our findings, reporting comparable rates of CME (3.1%). In relation to phacoemulsification in eyes vitrectomized for RRD, we found the rate of postcataract CME surgery keeps low (2%), especially if compared with the CME rate in eyes vitrectomized for ERM. In the literature, the likelihood of developing CME after cataract in previously vitrectomized eyes for RRD is reported to be slightly doubled (odds ratio 1.96 by Merad and 2.76 by Starr , ). As noted earlier, the published data on cataract surgery in eyes with previous RRD repair is drawn from heterogeneous RRD cases, which may explain the wider range of reported CME rates. It might be hence said that eyes vitrectomized for ERM have a significantly higher risk of developing postcataract CME compared to eyes vitrectomized for noncomplicated primary RRD. A few interesting clues might be inferred. It can be postulated that CME occurring after macular surgery might be addressed to the pathologic macular structural changes, including Müller cell dysfunction and intraretinal fibrosis. Moreover, some ERMs are described to cause a breakdown of the blood–retinal barrier, which in turn may lead to an increased vascular permeability, to the release of pro-inflammatory cytokines and, eventually, to an increase in intraocular inflammation. , These retinal changes, possibly together with the surgical tissue handling, might raise the risk of developing CME, making the macula more “vulnerable” and susceptible to edema. Our finding, especially regarding the increased risk of CME in relation to the ERM stage (the higher the stage, the higher the risk), corroborates this hypothesis: advanced stages may indeed present a significant breakdown of the blood–retina barrier or a malfunctioning glia, which is involved in the reuptake of intraretinal fluids. The role of internal limiting membrane peeling remains, however, controversial: considering that in our series, per protocol, we always peeled the internal limiting membrane, it should be also disclosed that eyes undergoing ERM removal without internal limiting membrane peeling might possibly show different rates of postsurgical CME than those found in our study. The increased susceptibility of operated macula to CME is also supported by the finding that almost all (87.5%) eyes presenting CME after cataract surgery previously presented CME after vitrectomy and ERM peeling. The present study has several limitations that should be disclosed, including the retrospective design and the limited numerosity. We acknowledge that, owing to the scarcity of isolated vitreoretinal procedures, the inclusion of such cases was relatively rare. Despite these drawbacks, strengths of our study were as follows: a careful patient selection; the exclusion of complicated RRD cases (such as those with proliferative vitreoretinopathy, myopic, or giant retinal tears), which may have jeopardized the results; and the novelty of providing a compared analysis in vitrectomized eyes for different retinal conditions. In conclusion, the risk of CME after cataract surgery in vitrectomized eyes differs significantly depending on the underlying retinal pathologic conditions. Eyes vitrectomized for idiopathic ERM are at a notably higher risk of developing CME, especially in advanced stages, because of macular structural changes and increased inflammatory responses. By contrast, eyes vitrectomized for primary noncomplicated RRD have a much lower risk of postcataract CME, suggesting that the decision to spare the lens in such cases does not raise concerns about increased CME risk after subsequent cataract surgery. For ERM cases, whether to perform sequential surgeries (vitrectomy followed by cataract surgery) or combined phacovitrectomy remains a matter of clinical judgment. Both strategies, according to the published data, carry a similar risk of CME, and the choice should be tailored to the individual patient's condition, particularly considering the ERM stage and the potential for macular vulnerability. Further prospective and randomized studied should be designed to confirm our findings and, possibly, integrate some therapeutic considerations. We, however, believe they can contribute to an improved understanding of this complex but very specific aspect, supporting surgeons in taking responsible decisions.
SUPPLEMENTARY MATERIAL
|
Inbag Morcellation Applied to the Laparoscopic Surgery of Leiomyoma: A Randomized Controlled Trial | 1dc095a7-db3e-48f6-bd29-77ed78004b74 | 8175161 | Histology[mh] | The advantage of laparoscopic minimal invasive surgery rather than the laparotomic approach is widely demonstrated in terms of reduction of morbidity and mortality for myomectomy and hysterectomy [ – ]. The power morcellation, described for the first time in 1993 , allows this laparoscopic approach for uterus and fibromas of big size and also for nulliparous women . However, this approach exposes patients to rare but potentially dangerous risks: the diffusion of hidden cancers, in particular, uterine sarcoma (prevalence between 1/225 to 1/580) and leiomyosarcoma (prevalence between 1/495 to 1/1100) whose clinical and radiological characteristics are quite often similar to leiomyoma and the development of iatrogenic parasite myomas (prevalence between 0.12 and 0.95%) [ – ]. The FDA has recently recommended in 2020 that health care providers should use tissue containment systems when using laparoscopic power morcellators, and that they ensure that the laparoscopic power morcellator and tissue containment system are compatible . However, inbag morcellation at the time of laparoscopic myomectomy is not mandatory in France. The idea of a laparoscopic morcellation protected into an endoscopic bag has been developed in order to avoid the spread of smooth muscle cells or carcinogenic cells and consequently to reduce the risk of cancer diffusion or parasite myomas . This technique consists of positioning a double-entry bag through laparoscopic trocars (an extra pubic orifice for the morcellator access and an umbilical orifice for the optic) . The surgical specimen is placed into the bag. Inflating this transparent bag once it is inside the peritoneal cavity gives the advantage of keeping the organs distant from the device and reduces the risk of accidents such as piercing or tearing of another organ (intestinal, vascular, or urological wound). Some observational studies about endoscopic bag have already been conducted and seem promising [ – ]. The aim of this prospective randomized study was to evaluate the efficacy and safety of inbag morcellation versus open morcellation during laparoscopic myomectomy or hysterectomy.
2.1. Trial Design In this randomized controlled trial conducted from January 2018 to January 2019 in the department of gynecology (Femme-Mere-Enfant Hospital, HCL, Lyon, France), we compared two groups of consenting patients who have undergone laparoscopic myomectomy and/or supracervical hysterectomy for fibroids by three experienced surgeons: one group with an inbag morcellation (group A) and one group without any bag (group B). The study has been approved by the Ethics Committee of Ile de France (IRB number: 2017-A01773-50) and is registered under clinicaltrials.gov identifier: NCT03281460 . All patients with symptomatic leiomyoma for whom laparoscopic myomectomy or supracervical hysterectomy were indicated were eligible for the present study. Patients with suspected sarcoma or any other cancerous tumor as well as pregnant patients were excluded from this study. 2.2. Trial Endpoints and Assessments The primary endpoint was the detection of smooth muscle cells (determined by cytology and immunohistochemistry) in the peritoneal fluid after fragmentation of the fibroids and/or uterus. In brief, conventional cytology after staining with May Grümwald Giemsa and then Papanicolaou was performed. When the spindle cells were displayed, a further analysis by immunohistochemistry was done on cell blocks from the washings to confirm or not the character of the smooth muscle cells. The following antibodies and dilutions were used: Anti-Caldesmom Antibody (1: 100; clone h-CD, DAKO) and anti-Smooth muscle actin Antibody (1: 600; clone 1A4, DAKO). Staining was revealed using the UltraView universal DAB detection kit (Ventana Medical System Inc.). The positivity of at least one of the two proteins confirmed the presence of smooth muscle cells. Two pathologists (AB, BN) read all samples in a blind manner. The secondary endpoints were the duration of the surgical procedure, the duration of the power morcellation, the duration of peritoneal washing, the time to find and pick residual fragments of leiomyoma after morcellation in group B, the weight of the fragments, and the duration of bag placement (group A). Surgeons rated the complexity of bag positioning using a 10 cm-VAS ranging from “difficult” to “easy” immediately after surgery (score 0 to 10). Intraoperative and postoperative complications have been registered using the Clavien-Dindo classification . 2.3. Surgical Technique A preoperative ultrasound was always performed to confirm the presence and location of fibroids. A complementary pelvic MRI could also be performed if ultrasound exam was nonconclusive or incomplete. Learning how to place the bag was obtained by reading the description of the technique . The surgery was always performed by laparoscopy (myomectomy or supracervical hysterectomy depending on the informed choice of the patient). Randomization (morcellation with or without bag) was performed at the beginning of the surgery. We used a 4-port laparoscopy in both groups as previously described : a 12 mm umbilical trocar for the laparoscope, two 5 mm trocars in the right and left iliac fossa two fingers across the anterosuperior iliac spine, and a 10 mm suprapubic trocar. For group A, the endoscopic bag system did not need any additional port and was placed through the suprapubic trocar (see below). For group B, the suprapubic trocar was removed to insert directly the morcellator. Peritoneal washing with 500 cc of sterile saline solution followed by complete aspiration was systematically performed on the whole abdominal cavity at the end of myomectomy or hysterectomy, just before the morcellation. Then, morcellation could be performed: For group A, the More-Cell-Safe® bag (AMI, Austria) was used following the technique described by our team : in brief, it is a specific laparoscopic polyurethane bag with 2 port bag design in order to insert both the optic and the power morcellator with a total capacity of 2.5 l. After insertion of the device through the suprapubic trocar, the surgical specimen was inserted into the bag. A pseudopneumoperitoneum is then created in the bag, and tissue morcellation was performed inside the bag. At the end of the procedure, the bag was removed from the abdominal cavity. Finally, its integrity was also checked by visual inspection and after water filling (water test with 1 l of NaCl solution) (see ). For group B, morcellation was performed directly in the peritoneal cavity without any bag All morcellations were performed with the LINA Xcise™morcellator (Kebomed, France). In both groups, a peritoneal washing with 500 cc of sterile saline solution was performed on the whole abdominal cavity at the end of the morcellation, then completely removed for cytology and immunohistochemistry analysis. 2.4. Sample Size and Statistical Analysis Our hypothesis was that the use of the inbag morcellation during laparoscopic myomectomy or hysterectomy could help to prevent the spread of smooth muscle cells inside the peritoneum. On the basis of the results of Rimbach et al. , we expected to find smooth muscle cells into the peritoneal washing in 28% of cases without bag and not to find them when using the bag. With an alpha risk of 5% and power of 80%, the number of patients required for the study was 24 per group, for a total of 48 patients. Patients were randomly assigned to either the experimental group A or the control group B in a 1 : 1 ratio (randomization list established by SAS software in a 1 : 1 allocation using random block sizes of 6). The statistical analysis was performed on software SAS (SAS Studio 3.6; SAS Institute Inc.). The data were described by means and standard deviation for continuous quantitative data and their size and frequency for qualitative data. The categorical variables were compared using the chi 2 test or the Fisher test if the number was less than 5, and the continuous variables were compared using the Student test. Tests were considered significant if the p value was less than 0.05. A simple linear regression model was used to test the evolution of the “bag placement time” and “bag location complexity score” according to the duration.
In this randomized controlled trial conducted from January 2018 to January 2019 in the department of gynecology (Femme-Mere-Enfant Hospital, HCL, Lyon, France), we compared two groups of consenting patients who have undergone laparoscopic myomectomy and/or supracervical hysterectomy for fibroids by three experienced surgeons: one group with an inbag morcellation (group A) and one group without any bag (group B). The study has been approved by the Ethics Committee of Ile de France (IRB number: 2017-A01773-50) and is registered under clinicaltrials.gov identifier: NCT03281460 . All patients with symptomatic leiomyoma for whom laparoscopic myomectomy or supracervical hysterectomy were indicated were eligible for the present study. Patients with suspected sarcoma or any other cancerous tumor as well as pregnant patients were excluded from this study.
The primary endpoint was the detection of smooth muscle cells (determined by cytology and immunohistochemistry) in the peritoneal fluid after fragmentation of the fibroids and/or uterus. In brief, conventional cytology after staining with May Grümwald Giemsa and then Papanicolaou was performed. When the spindle cells were displayed, a further analysis by immunohistochemistry was done on cell blocks from the washings to confirm or not the character of the smooth muscle cells. The following antibodies and dilutions were used: Anti-Caldesmom Antibody (1: 100; clone h-CD, DAKO) and anti-Smooth muscle actin Antibody (1: 600; clone 1A4, DAKO). Staining was revealed using the UltraView universal DAB detection kit (Ventana Medical System Inc.). The positivity of at least one of the two proteins confirmed the presence of smooth muscle cells. Two pathologists (AB, BN) read all samples in a blind manner. The secondary endpoints were the duration of the surgical procedure, the duration of the power morcellation, the duration of peritoneal washing, the time to find and pick residual fragments of leiomyoma after morcellation in group B, the weight of the fragments, and the duration of bag placement (group A). Surgeons rated the complexity of bag positioning using a 10 cm-VAS ranging from “difficult” to “easy” immediately after surgery (score 0 to 10). Intraoperative and postoperative complications have been registered using the Clavien-Dindo classification .
A preoperative ultrasound was always performed to confirm the presence and location of fibroids. A complementary pelvic MRI could also be performed if ultrasound exam was nonconclusive or incomplete. Learning how to place the bag was obtained by reading the description of the technique . The surgery was always performed by laparoscopy (myomectomy or supracervical hysterectomy depending on the informed choice of the patient). Randomization (morcellation with or without bag) was performed at the beginning of the surgery. We used a 4-port laparoscopy in both groups as previously described : a 12 mm umbilical trocar for the laparoscope, two 5 mm trocars in the right and left iliac fossa two fingers across the anterosuperior iliac spine, and a 10 mm suprapubic trocar. For group A, the endoscopic bag system did not need any additional port and was placed through the suprapubic trocar (see below). For group B, the suprapubic trocar was removed to insert directly the morcellator. Peritoneal washing with 500 cc of sterile saline solution followed by complete aspiration was systematically performed on the whole abdominal cavity at the end of myomectomy or hysterectomy, just before the morcellation. Then, morcellation could be performed: For group A, the More-Cell-Safe® bag (AMI, Austria) was used following the technique described by our team : in brief, it is a specific laparoscopic polyurethane bag with 2 port bag design in order to insert both the optic and the power morcellator with a total capacity of 2.5 l. After insertion of the device through the suprapubic trocar, the surgical specimen was inserted into the bag. A pseudopneumoperitoneum is then created in the bag, and tissue morcellation was performed inside the bag. At the end of the procedure, the bag was removed from the abdominal cavity. Finally, its integrity was also checked by visual inspection and after water filling (water test with 1 l of NaCl solution) (see ). For group B, morcellation was performed directly in the peritoneal cavity without any bag All morcellations were performed with the LINA Xcise™morcellator (Kebomed, France). In both groups, a peritoneal washing with 500 cc of sterile saline solution was performed on the whole abdominal cavity at the end of the morcellation, then completely removed for cytology and immunohistochemistry analysis.
Our hypothesis was that the use of the inbag morcellation during laparoscopic myomectomy or hysterectomy could help to prevent the spread of smooth muscle cells inside the peritoneum. On the basis of the results of Rimbach et al. , we expected to find smooth muscle cells into the peritoneal washing in 28% of cases without bag and not to find them when using the bag. With an alpha risk of 5% and power of 80%, the number of patients required for the study was 24 per group, for a total of 48 patients. Patients were randomly assigned to either the experimental group A or the control group B in a 1 : 1 ratio (randomization list established by SAS software in a 1 : 1 allocation using random block sizes of 6). The statistical analysis was performed on software SAS (SAS Studio 3.6; SAS Institute Inc.). The data were described by means and standard deviation for continuous quantitative data and their size and frequency for qualitative data. The categorical variables were compared using the chi 2 test or the Fisher test if the number was less than 5, and the continuous variables were compared using the Student test. Tests were considered significant if the p value was less than 0.05. A simple linear regression model was used to test the evolution of the “bag placement time” and “bag location complexity score” according to the duration.
See , Tables and , and the data tables. Of 50 screened patients, 48 were randomized; 2 patients declined to participate (see : flow diagram). All patients were subsequently included between January 2018 and January 2019: 24 in group A and 24 in group B. The average age in group B was 42.2 ± 7.54 years old and 45.2 ± 8.20 years old in group A ( p = .19). The epidemiological data did not differ significantly between both groups ( ). Preoperative ultrasound exams revealed an average of 2.79 ± 1.59 myomas in group B and 2.04 ± 1.27 in group A (see ). The surgical indications were given as follows: 12 myomectomies (50%) and 12 supracervical hysterectomies (50%) in group B and 9 myomectomies (37.5%) and 15 supracervical hysterectomies (62.5%) in group A ( p = .38). Patient characteristics are summarized in . The duration of surgery did not differ significantly between the two groups: 128 ± 68.3 min for group B and 117 ± 30.9 min for group A ( p = .51). Similarly, the duration of morcellation was, respectively, of 5.47 ± 4.90 and 6.34 ± 4.24 min ( p = .52); the duration of peritoneal washing after morcellation was 2.50 ± 1.58 min for group B and 2.03 ± 1.60 min for group A ( p = 0.31). The weight of residual fragments was in average 97.1 ± 70.2 g in group B (found in the peritoneal cavity) and 152 ± 130 g in group A (found in the bag) ( p = .07) (see ). The average duration of bag placement was 8.32 ± 3.67 minutes. This variable did not seem to change with the progress of the study (linear regression model p = .28). The surgeons evaluated the easy use of the bag with an average of 8.89 ± 2.11 out of 10. Similarly, no trend was observed over time (linear regression model p = .36). There was not any detectable leakage during morcellation. Among the 24 uses of the More-Cell-Safe® bag, one surgeon reported a difficult case related to a weak pneumoperitoneum related to inadequate curare administration ( ). No intraoperative or postoperative complications were reported throughout the study, except one case of Clavien-Dindo grade 2 urinary tract infection treated with antibiotics. No malignant lesions were identified when examining the fragments of the surgical pieces of the 48 patients. In one patient from group A undergoing supracervical hysterectomy, atypical endometrial hyperplasia was detected. Peritoneal fluid was systematically collected after peritoneal washing. The analysis revealed the presence of smooth muscle cells in 7 cases (29.2%) in group B; it was negative for group A, p = 0.009 ( ) (See ).
Since becoming aware of the potential risks associated with morcellation, different surgical teams have described inbag morcellation [ – , , ]. However, a recent review by the Cochrane database about inbag versus uncontained power morcellation concluded that there were limited data on the effectiveness and safety of endoscopic bag and the need for new trials . In our study, the only parameter significantly different from both groups was the presence of smooth muscle cells in peritoneal washing when nonprotected morcellation was performed. There may be a risk associated with dissection during myomectomy or hysterectomy with the passage of smooth muscle cells into the peritoneal cavity regardless of morcellation: in their study of 31 myomectomies, Lambat-Emery et al. demonstrated the presence of smooth muscle cells in peritoneal fluid in 8 patients after dissection and before protected morcellation. The impact of this low-level dissemination related to dissection and not to the morcellation is likely negligible in comparison with dissemination associated with morcellation . Kho and Nezhat hypothesized that the risk of developing parasitic myomas was mainly related to tissue fragments left in the peritoneal cavity rather than to isolated cells. They observed that the number of leiomyoma was higher in patients with power morcellation than in manual morcellation and attributed this to the fact that the fragments are larger and more easily detectable after cold knife morcellation rather than after electrical morcellation. Yu et al. demonstrated the interest of abundant peritoneal washing to minimize the theoretical risk linked to the presence of isolated cells: in 16 cases of myomectomies and morcellation without bag, smooth muscle cells were alternatively found in peritoneal fluid after myomectomy (3 cases) or after morcellation (5 cases). In all cases, cytology was negative after washing with 3 L of NaCl solution. It is therefore important to perform a large peritoneal washing before starting morcellation. No study has assessed what is the correct volume of irrigation with either normal saline or sterile water to decrease tissue dissemination during laparoscopic myomectomy. The authors concluded that copious irrigation and suctioning may reduce myoma cell dissemination . Regarding the carcinogenic risk, all patients had undergone a preoperative ultrasound assessment: no neoplastic lesion was suspected. However, the histopathological results of the surgical specimen revealed the presence of atypical endometrial hyperplasia in one case from group A. Hysteroscopy with directive endometrial biopsy had been performed prior to hysterectomy. Histopathological results had showed only benign simple hyperplasia without atypia. As progestogen treatment did not control menorrhagia, hysterectomy with salpingectomy was indicated as a second-line treatment . On one hand, as the patient had concerns regarding changes in her sexuality and the potential risk of prolapse in case of total hysterectomy, and on the other hand, as subtotal hysterectomy may be an alternative surgical treatment in case of simple hyperplasia without atypia (the less severe step of endometrial hyperplasia) , we had first decided to perform a supracervical hysterectomy with salpingectomy. However, because of the final histopathologocal results of atypical endometrial hyperplasia with its carcinogenic risk, we secondly proposed to perform a complementary laparoscopic trachelectomy. This example illustrates the potential risk despite preventive measures (ultrasound exam and biopsy) and underlines the interest of a protected morcellation. In our study, only one bag was difficult to place due to a weak pneumoperitoneum related to inadequate curarization. In spite of that, all cases in group A were feasible. Its placement did not significantly increase the total operating time. We think that once the surgeon has learned how to use the bag, the time to place the bag is no longer than the time to recover the scattered fragments. There have been no significant changes in the assessment of its use or in the duration of bag placement during the study. This may indicate that the learning curve is quick. However, respect to the specific technical procedure is needed before performing contained morcellation : a cohort of 76 morcellations after hysterectomy or myomectomy in Endocath and Ecosac bags revealed 7 dye leaks while the bags were intact, highlighting a possible management error . One limit of the bag may be its dimension in case of big specimen. Rimbach et al. failed to place a 1050 g uterus into the bag. The largest specimen that his team could put in the bag was a 18 × 12 × 10 cm uterus (638 g). In our study, the largest piece was a 17 × 12 × 11 cm uterus (640 g). Preoperative investigations should accurately assess the size of any uterus or leiomyoma before laparoscopic procedures. Limitations of the current study include its relatively small size, the single center design and the absence of washing performed at the end of morcellation (before sampling was done) because it could have affected the detection of the smooth muscle cells. Strengths of this study include the overall design (a randomized controlled trial) and the double detection of smooth muscle cells (determined by cytology and immunohistochemistry).
The use of the More-Cell-Safe® bag (A.M.I. Austria) seems to be efficient to avoid the risk of the spread of smooth muscle cells l related to laparoscopic morcellation of uterus and leiomyoma. This device seems easy to use. Surgeons should continue to inform patients about the risks associated with morcellation and remain vigilant and attentive to the preoperative assessment. Our study widely encourages the use of endoscopic bag during laparoscopic morcellation.
|
Religious diversity and public health: Lessons from COVID-19 | 667f70b3-fec0-4480-ab6d-1d0f81e5b13d | 10449469 | Health Communication[mh] | COVID-19 was a painful reminder of how the health practices of ethnic, religious and migrant minorities often become the focus of heated political discourse during times of crisis. The rise in Anti-Asian hate crimes in the US is but one example of a long history of blaming “Others”, be it Irish Catholic immigrants during the Cholera outbreak or Italian immigrants blamed for spreading Polio . Fueled by media, nationalist and public health discourse, minority groups around the globe are framed as being “irresponsible,” “dangerous” and potential hotspots of infection–reflecting broader claims of biological and social risks of transmission and contagion . This global trend was also prevalent in Israel, where pictures of Haredim (a strictly Orthodox religious minority) defying public health guidelines featured prominently in the public sphere, constantly presented by media and politicians as a “risk” to the body politic . Images of Haredim–demonstrating at rallies, praying in resistance, and defying Israeli soldiers–became a daily public spectacle during continuous lockdowns. In contrast to this public display of minority non-compliance, this paper goes beyond the threshold of Haredi homes to examine the varieties of Haredi public health compliance while highlighting inner group diversity. Haredim in Israel consist of three main sub-groups—Lithuanian, Sephardi and Hasidic Jews–who each behaved differently during COVID-19. Based on our survey (N = 800), we show that Hasidim, a more group-oriented stream of Haredi Judaism, were 12% and 14% more likely to flout public health guidelines than their Lithuanian and Sephardi counterparts, respectively (X 2 = 35.4, p = .00, df = 2, N = 708). Despite this diversity, all Haredim were portrayed in Israeli media as one homogeneous “black lump” and constantly blamed for flouting public health guidelines and spreading COVID-19 . Not only did these gaps between reality and public imagery contribute to Haredim feeling like they were discriminated against , we argue that this gap can also obstruct public health compliance. In this particular case, disregarding inner group diversity contributes to stereotypes of a “dangerous” and “irresponsible” Haredi minority which can have psycho-social effects and also, in turn, effect adherence to public health guidelines [ See : , ]. Charting these inner-group differences also allows us to push for a more nuanced understanding of the role religion actually plays in health decision-making. Recent studies reveal how the category of “religion” is often used to conflate a variety of factors that affect health decision-making . This paper advances these growing understandings at the intersection of critical medical anthropology and religion, which often rely on qualitative data, to test these questions in a large scale quantitative study. Religion, health and COVID-19 As the pandemic reshaped the social world we live in, scholars turned to examine the ways different people followed social distance guidelines. Scholars have identified a range of variables that predict public health compliance during COVID-19, including: psychological, institutional and situational variables as well as demographic characteristics such as gender, geographical location and age . Studies examining social distance guidelines in religious contexts have found that political affiliation is also key to understanding adherence to guidelines . For example, Perry, Whitehead, and Grubbs found that in the US, people affiliated with the left were more likely to recommend precautions, while those on the (religious) right were more likely to disregard recommended precautions . These findings resonate with Kahan’s argument that attitudes towards climate change tell us less about scientific reasoning and knowledge than about “latent cultural affiliations” , or the findings of a Pew Research Center survey that indicate that while more and more Americans are putting climate change as a national priority, democrats are much more concerned than their republican counterparts . The implications of these studies are that people do not only encounter science and health-related decisions as members of particular groups, but they also seek out knowledge through pre-existing networks, utilizing prior acquaintances to help them understand new information. Recognizing the role that social context, including history, community, culture and religion play in shaping an individual’s ideas and worldviews, these findings resonate with established research from sociology, anthropology and history which has demonstrated how engagement with science is repeatedly shaped by social identity, historical context and power relations . As Feinstein and Waddington put it: “People encounter scientific questions in social context—both as members of their social and cultural groups and with other members of those groups” . Not only is the attempt to “filter out” these cultural frameworks unrealistic, it also assumes that these frameworks are detrimental to “proper” modes of rational decision-making, while, in reality they can be powerful and constructive resources for decision-making . Well before the pandemic hit, religious minority groups were often perceived as “hard to reach” or as “non-compliant”. Even though these categories have been proved rather unhelpful, public health discourse and policies often use the term “religion” to account for differences in health-related actions, such as vaccine uptake . Yet, recent scholarship shows that members of religious groups combine both scientific knowledge and socio-religious frameworks, which serve as “cultural and epistemological tunnels” . In a recent study of COVID-19 health decision-making amid Haredi Jews , it was found that both health and religious justifications were used during sense-making and COVID-19-related decision-making. Whereas many respondents used general health-related justifications, many also utilized health-related justifications that were directly linked to religious language and culture, for example: ‘You shall take care exceedingly of your lives’ (Deuteronomy 4:15), suggesting that Haredi men and women have a particular vocabulary to express their justifications for following public guidelines. In a similar vein, religious leaders are often perceived as key stakeholders in the medical decisions of religious individuals . Within the emerging field of inclusive science communication, faith leaders are often portrayed as “trusted voices” within their communities who serve as sources of support, information and credibility . Religious leaders can also identify ways to overcome particular community-based challenges, bridging understanding about the diverse ways health, healing and risk are conceived, especially in situations where health regulations can be perceived as undermining group-based priorities [ see : – ]. COVID-19 in Israel—Haredi varieties Haredim (Ultra-Orthodox Jews) account for roughly 12.6% of Israel’s population . Haredi men and women live according to the Hebrew Bible, which has been continuously interpreted through a large (and ever-growing) body of rabbinic literature and Jewish law. Haredi Jews can be easily identified by their relatively unified dress code (black hats and dark suits for men, and similarly colored ankle-length skirts, long sleeves, and head coverings or wigs for women). They are typically divided into three different sub-groups: Lithuanian yeshiva -based (Torah learning) communities; Hasidim, who place great emphasis on personal experience and more-charismatically oriented group worship; and Sephardi Haredim, who trace their origins to the Iberian peninsula, North Africa, and the Middle East. All Haredi sub-groups are often referred to as an enclave culture with strict social and cultural boundaries and distinguished from other streams of Judaism: Progressive, Conservative, and religious-Zionist, by their avoidance of secular education and professional training . The first case of COVID-19 in Israel was confirmed on 21 February 2020. From March 11, the Israeli government put forward an increasingly restrictive set of social distancing measures, culminating in a full lockdown by March 19. The number of infected cases rose rapidly during the last week of March, resulting in 6,092 confirmed cases by the beginning of April. During this initial stage of the pandemic, Haredi Jews were slower to adhere to social distancing guidelines than other groups in Israeli society . Scholars have attributed this reluctance to various theological, cultural, and political causes. Some have blamed inner-communal media outlets for not reporting the dangers sufficiently, whereas, others have pointed to the ways social distancing disrupts the core of Jewish life which is based on religious obligations that can only be performed as a group, most notably collective prayer three times a day and religious study . Others accused Haredi leaders for promoting non-compliance among their adherents, especially during the early stages of the pandemic. Shuki Friedman and Gabriel Even-Tsur found that there was a clear link between the guidelines put forward by rabbinic leadership and their relationship to the state of Israel. Sephardi leaders, who often hold state-funded rabbinic positions, instructed their followers to abide by public health guidelines, as soon as these guidelines were issued. In contrast, Lithuanian and Hasidic leaders did not follow their example and waited a few weeks until they circulated similar guidelines. However, Lithuanian and Hasidic leaders only called on their followers to abide by public health deadlines, at the end of March 2020 as a response to the devastating hit on their communities. This delayed action contributed to the disproportionate effect of COVID-19 on the Israeli Haredi population, with major hotspots in Haredi neighborhoods and 40% to 60% of all coronavirus patients at four major hospitals, even though they make only 12% of Israel’s population . As a response to this drastic blow, by the end of March 2020, all Haredi leadership put forward a clear message to follow public health guidelines, yet public instances of non-compliance were still clearly visible. Amidst this new context, this study was designed to examine the varieties of health decision-making among Haredim. This project advances recent interest in non-elite attitudes and science understandings while analyzing the ways lay individuals incorporate Haredi sensibilities as part of their everyday decisions amid COVID-19. In doing so, our research also draws on growing scholarship pointing to the over-emphasis of religious leadership on health decision-making in religious contexts . For example, Michal Raucher has shown that religious Jewish mothers are the primary carers of children and often responsible for managing health decision-making. She demonstrates how Haredi mothers “conceive” more authority after birthing four children, and navigate their own decision-making while disregarding both medical and religious authorities . Similarly, Ben Kasstan has shown that mothers are important influencers in the context of vaccines, a fact which is often obscured when solely focusing on male religious authorities . In this study we ask: How did Haredi men and women make health decisions during the pandemic? What types of justifications were utilized to support their decisions? How do demographic factors, such as age, education, gender and religious affiliation affect these decisions? And, what can these trends teach us about religious minorities and public health more broadly? As the pandemic reshaped the social world we live in, scholars turned to examine the ways different people followed social distance guidelines. Scholars have identified a range of variables that predict public health compliance during COVID-19, including: psychological, institutional and situational variables as well as demographic characteristics such as gender, geographical location and age . Studies examining social distance guidelines in religious contexts have found that political affiliation is also key to understanding adherence to guidelines . For example, Perry, Whitehead, and Grubbs found that in the US, people affiliated with the left were more likely to recommend precautions, while those on the (religious) right were more likely to disregard recommended precautions . These findings resonate with Kahan’s argument that attitudes towards climate change tell us less about scientific reasoning and knowledge than about “latent cultural affiliations” , or the findings of a Pew Research Center survey that indicate that while more and more Americans are putting climate change as a national priority, democrats are much more concerned than their republican counterparts . The implications of these studies are that people do not only encounter science and health-related decisions as members of particular groups, but they also seek out knowledge through pre-existing networks, utilizing prior acquaintances to help them understand new information. Recognizing the role that social context, including history, community, culture and religion play in shaping an individual’s ideas and worldviews, these findings resonate with established research from sociology, anthropology and history which has demonstrated how engagement with science is repeatedly shaped by social identity, historical context and power relations . As Feinstein and Waddington put it: “People encounter scientific questions in social context—both as members of their social and cultural groups and with other members of those groups” . Not only is the attempt to “filter out” these cultural frameworks unrealistic, it also assumes that these frameworks are detrimental to “proper” modes of rational decision-making, while, in reality they can be powerful and constructive resources for decision-making . Well before the pandemic hit, religious minority groups were often perceived as “hard to reach” or as “non-compliant”. Even though these categories have been proved rather unhelpful, public health discourse and policies often use the term “religion” to account for differences in health-related actions, such as vaccine uptake . Yet, recent scholarship shows that members of religious groups combine both scientific knowledge and socio-religious frameworks, which serve as “cultural and epistemological tunnels” . In a recent study of COVID-19 health decision-making amid Haredi Jews , it was found that both health and religious justifications were used during sense-making and COVID-19-related decision-making. Whereas many respondents used general health-related justifications, many also utilized health-related justifications that were directly linked to religious language and culture, for example: ‘You shall take care exceedingly of your lives’ (Deuteronomy 4:15), suggesting that Haredi men and women have a particular vocabulary to express their justifications for following public guidelines. In a similar vein, religious leaders are often perceived as key stakeholders in the medical decisions of religious individuals . Within the emerging field of inclusive science communication, faith leaders are often portrayed as “trusted voices” within their communities who serve as sources of support, information and credibility . Religious leaders can also identify ways to overcome particular community-based challenges, bridging understanding about the diverse ways health, healing and risk are conceived, especially in situations where health regulations can be perceived as undermining group-based priorities [ see : – ]. Haredim (Ultra-Orthodox Jews) account for roughly 12.6% of Israel’s population . Haredi men and women live according to the Hebrew Bible, which has been continuously interpreted through a large (and ever-growing) body of rabbinic literature and Jewish law. Haredi Jews can be easily identified by their relatively unified dress code (black hats and dark suits for men, and similarly colored ankle-length skirts, long sleeves, and head coverings or wigs for women). They are typically divided into three different sub-groups: Lithuanian yeshiva -based (Torah learning) communities; Hasidim, who place great emphasis on personal experience and more-charismatically oriented group worship; and Sephardi Haredim, who trace their origins to the Iberian peninsula, North Africa, and the Middle East. All Haredi sub-groups are often referred to as an enclave culture with strict social and cultural boundaries and distinguished from other streams of Judaism: Progressive, Conservative, and religious-Zionist, by their avoidance of secular education and professional training . The first case of COVID-19 in Israel was confirmed on 21 February 2020. From March 11, the Israeli government put forward an increasingly restrictive set of social distancing measures, culminating in a full lockdown by March 19. The number of infected cases rose rapidly during the last week of March, resulting in 6,092 confirmed cases by the beginning of April. During this initial stage of the pandemic, Haredi Jews were slower to adhere to social distancing guidelines than other groups in Israeli society . Scholars have attributed this reluctance to various theological, cultural, and political causes. Some have blamed inner-communal media outlets for not reporting the dangers sufficiently, whereas, others have pointed to the ways social distancing disrupts the core of Jewish life which is based on religious obligations that can only be performed as a group, most notably collective prayer three times a day and religious study . Others accused Haredi leaders for promoting non-compliance among their adherents, especially during the early stages of the pandemic. Shuki Friedman and Gabriel Even-Tsur found that there was a clear link between the guidelines put forward by rabbinic leadership and their relationship to the state of Israel. Sephardi leaders, who often hold state-funded rabbinic positions, instructed their followers to abide by public health guidelines, as soon as these guidelines were issued. In contrast, Lithuanian and Hasidic leaders did not follow their example and waited a few weeks until they circulated similar guidelines. However, Lithuanian and Hasidic leaders only called on their followers to abide by public health deadlines, at the end of March 2020 as a response to the devastating hit on their communities. This delayed action contributed to the disproportionate effect of COVID-19 on the Israeli Haredi population, with major hotspots in Haredi neighborhoods and 40% to 60% of all coronavirus patients at four major hospitals, even though they make only 12% of Israel’s population . As a response to this drastic blow, by the end of March 2020, all Haredi leadership put forward a clear message to follow public health guidelines, yet public instances of non-compliance were still clearly visible. Amidst this new context, this study was designed to examine the varieties of health decision-making among Haredim. This project advances recent interest in non-elite attitudes and science understandings while analyzing the ways lay individuals incorporate Haredi sensibilities as part of their everyday decisions amid COVID-19. In doing so, our research also draws on growing scholarship pointing to the over-emphasis of religious leadership on health decision-making in religious contexts . For example, Michal Raucher has shown that religious Jewish mothers are the primary carers of children and often responsible for managing health decision-making. She demonstrates how Haredi mothers “conceive” more authority after birthing four children, and navigate their own decision-making while disregarding both medical and religious authorities . Similarly, Ben Kasstan has shown that mothers are important influencers in the context of vaccines, a fact which is often obscured when solely focusing on male religious authorities . In this study we ask: How did Haredi men and women make health decisions during the pandemic? What types of justifications were utilized to support their decisions? How do demographic factors, such as age, education, gender and religious affiliation affect these decisions? And, what can these trends teach us about religious minorities and public health more broadly? This paper is part of a large study exploring health decision-making in the context of COVID-19 among Haredi (ultra-Orthodox) Jews in Israel. As many do not have access to the internet, this paper is based on a telephone survey conducted by the Haredi market research firm “Eskaria”, which was administered to collect participants’ stances regarding COVID-19–related dilemmas. Information was gathered about education, age, gender, demography and religious affiliation and then participants were asked to respond to two COVID-19–related dilemmas that incorporated a potential conflict vis-à-vis social distancing guidelines. Each respondent was asked to report their solution to each dilemma in ways that best corresponded to their own attitudes and everyday decisions. Data was collected by “Eskaria” between 10–15 December 2020, at the height of the second wave of COVID-19, prior to a third national closure which was issued on 27 December 2020. Ethics This study is a secondary analysis of this data, which was collected by "Eskaria” for the Haredi Institute for Public Affairs. The secondary analysis was waived by the IRB committee at the Technion: Israel Institute of Technology. Verbal consent was acquired by “Eskaria”. Research tool The questionnaire was constructed around real-life COVID-19 dilemmas to capture modes of health decision-making amidst the pandemic [ See : , ]. It also included measures of compliance, knowledge about COVID-19, and demographics. Content validity with experts Content-related validity of the research tool was established non statistically using expert professional judgment, who addressed domain specification, content universe and sample, item development, item wording, and format. The questionnaire benefited from the constructive feedback of five members of the Haredi Institute for Public Affairs, which included both Haredi men and women as well as expert scholars of Haredi Judaism, all of whom are familiar with Haredi sensitivities. We also received feedback from three science communication and science education specialists, two of whom had a specific expertise in religion, with particular knowledge of Haredi society. We also received feedback from math, geography and science teachers to make sure that it resonated with general public literacy. Cognitive validity with target population We gave five Haredim a draft of the questionnaire and asked them to fill it out. Given the tension related to vaccine hesitancy in Israel amid the pandemic, a few voiced concerns about the conclusions people might have from the dilemmas. In response to this concern, we made sure to phrase the questions in ways that don’t direct readers to make any decisions based on the questionnaire. Sample Participants were recruited by the market research firm “Eskaria”, an online panel with particular expertise in conducting surveys among Hebrew speaking Haredim. The response rate was 18% and the entire sample included 800 participants. The limited response rate is not surprising given the relative length of this survey, the particular challenges of accessing Haredi publics for research purposes and the timing of the survey during the pandemic. This response rate might expose our data to a differential selectivity bias and limit the external generalizability of the findings. Nevertheless, our data offers important insights and data about inner group diversity that is often glossed over in academic research. Despite these difficulties, the final sample was representative of the general Haredi population regarding proportions of subgroups (Lithuanian, Hasidim, Sephardim). In order to obtain an in-depth understanding of the varieties within ultra-Orthodox society, a random sample was chosen from Eskaria’s database that includes over 400,000 men and women in ultra-Orthodox society aged 18 and up. In terms of representativeness, the sample was similar to the general population in most categories. Almost all respondents (87%) were married, which reflects the high marriage rates in the Haredi population. Yet, the sample included more men (59%) than women (41%). While this balance was kept among Lithuanian and Sephardim, among Hasidim 82% of the respondents were male (Hasidic men 82% (N = 197), Hasidic women 18% (N = 44), Sephardi men 45% (N = 108), Sephardi women 55% (N = 132), Lithuanian men 48% (N = 109), Lithuanian women 52% (N = 120). Participants were also more educated than the Haredi average: 18% had completed a professional training course (compared with 29% in the general Haredi population), 23% had or were completing a BA degree (compared with 20% in the general Haredi population). The religious affiliation of respondents was also very similar to the general Haredi population, 30% (N = 240) belong to the Mizrahi sub-group, 29% (N = 229) to the Lithuanian community, 30% to the Hasidic community (N = 241), and 11% belongs to a variety of communities including Chabad, National-Haredi and Breslav (N = 90). Finally, we struggled to get older participants, and thus, most of our data is based on respondents between 21–44 (72%). Having said that, the Haredi population is relatively young: Haredim under the age of 17 are 53.5% of the Haredi population, compared to a national average of 28%; therefore the sample represents a very substantial part of the adult Haredi population (Haredi Institute for Public Affairs, 2020, based on data provided by the Central Bureau of Statistics). Limitations and future studies Whether collecting data about actual compliance to COVID-19 or eliciting responses to hypothetical dilemmas, this data has been derived solely from the self-reports of the respondents. In addition, the survey was circulated during a time of crisis, which might have contributed to the relatively low response rate (discussed above) and might also reflect modes of decision-making that are specific to the pandemic context. Taken together, future studies which incorporate qualitative research methods conducted during less turbulent times will contribute to further understanding health-decision making among religious and ethnic minorities. Data analysis Dependent variable COVID-19 decision-making . Two dilemmas were designed to assess how Haredi men and women make COVID-related decisions amidst changing public health guidelines and recommendations. At the time, official guidelines allowed a maximum of ten people inside and twenty outside and schools were kept open. If exposed, there was a requirement to do whatever is necessary to stay home and test. Participants were asked how they would behave in social situations such as: Dilemma 1 (wedding): “Imagine your best friend scheduled a wedding for his son/daughter and the new guidelines now say that the wedding can only include twenty people outdoors. What would you recommend that they do?”. Dilemma 2 (elevator): “Your neighbor, who knew he had COVID, went into an elevator with you, without wearing a mask. As a consequence, you might have covid. Would you send your children to Talmud Torah (Jewish day school) in the next few days?”. For each dilemma, we calculated the decision-making as follows: Dilemma 1 (wedding)—Participants were presented with three options: 1. Postpone a wedding or fully adhere to wedding guidelines, 2. Recognize there is a problem and therefore have a wedding with a larger number of guests than allowed but still, less guests than one would usually have at a wedding, 3. Behave as if business is usual (and have a regular sized wedding). Based on these three options we created a new variable with two values, 1 = full compliance with guidelines (category one), 0 = partial or no compliance (categories 2 and 3). Dilemma 2 (elevator)—Participants were presented with two options: (1) To send a possibly infected child to school or (2) Not to send a possibly infected child to school. Independent variables Justification . To analyze how each respondent explained their stance on each dilemma, we presented an open-end question to the participants, asking them to explain their choices. Based on content analysis we divided these justifications into ten variables: (1) Reference to elderly population (e.g. “We should take care of the elderly population); (2) Health-related justifications . This includes a reference to health, medical or scientific argumentations, e.g. “I can carry the virus even if I feel fine”; (3) Public law or health authority recommendations (e.g. “We should follow the guidelines and it is not a joke!”); (4) Reference to special actions that the respondents will do (e.g. “We will all wear masks so it will be fine”); (5) Personal reasons (“You only get married once”!); (6) Lack of concern (“It is not really that dangerous”); (7) Religious justifications (e.g. “Only god is really in charge of the world”); (8) Public concern (“e.g. I would never want to cause harm to another person”, (9) C omparison between different situations (“As long as we are going to work, we can go to a wedding” (10) Other . Compliance with MOH (Ministry of Health) guidelines . To examine compliance with Ministry of Health guidelines (MOH), we created an index that examined the degree of compliance with individual guidelines. The index is made from a variety of statements, e.g. spatial distancing, hygiene, face-masks, and vaccination. For each statement, respondents were asked to rate the extent to which they would adhere to the guidelines between 1 (not strict at all) and 5 (very strict). The index was built as an average of four statements (N = 798, mean 3.7, SD = 0.95, range 1–5, Cronbach’s alpha = .69): “2 meters distance in public space” (N = 795, mean = 3.62, SD = 1.28), “mask wearing” (N = 796, mean = 4.34, SD = 1.06), “personal hygiene” (N = 793, mean = 3.82, SD = 1.36), “vaccination willingness” (N = 753, mean 2.36, standard deviation 1.13). Knowledge in the Context of Covid-19 . An open question about the efficacy of quarantine was used to assess knowledge about COVID-19. It was based on the response to the following question: “Any person who comes in contact with a verified patient is required to self-quarantine for 14 days from the moment of exposure, even though they had been out on the street for several days. Do you think isolation will help at this point?” Based on PISA’s definition of scientific explanation , we evaluated the application of scientific content knowledge to interpret and explain phenomena using the following categories: (1) respondent’s position regarding the efficacy of quarantine (intercoder reliability κ = .804), (2) correctness of the answer (intercoder reliability κ = .822), and (3) number and level of scientific concepts used. The concepts were characterized according to the author’s protocol , which is based on the inclusion of scientific concepts in Israeli science curricula for elementary, middle, and high school. Points were given only when respondents used scientific concepts correctly (e.g. “You can be contagious even if you do not have symptoms” versus “You are only contagious if you have symptoms”). One point was given for each elementary concept, two for each middle school–level concept, and three points for each high school–level concept. The scores for knowledge about COVID-19 ranged from 0 to 4, a higher score represents better knowledge (N = 800, mean = 2.04, SD = 1.15). Demand for knowledge . Respondents were asked to rate how much more information about COVID-19 they would like to obtain in four different topics–math, vaccines, science and geography. The variable demand for knowledge was based on the sum of four variables: 1) Demand for math knowledge, 2) Demand for knowledge about vaccines, 3) Demand for knowledge in science, and 4) Demand for knowledge in geography (N = 800, mean = 6.3, SD = 4.3, range 0–13, 0 = no lack of knowledge, 1 = missing one item of information on a subject, 2 = missing two items of information, and so on). Trust in rabbinical leadership . Respondents were asked to rate their degree of agreement between 1 (strongly disagree) and 5 (strongly agree) with the following statement: I have trust in rabbinical leadership (N = 787, mean = 4.70, SD = 0.79, range 1–5). Trust in Israeli government . To examine trust in Israel’s government, we created an index based on the average of four questions: 1) I have faith in the Israeli government, 2) The government acts in an equal manner towards all citizens, 3) The Israeli government is making the right decisions for the residents of Israel with regard to the COVID-19 crisis, 4) Community restrictions following the COVID-19 crisis should be determined by the government. The variable ranks from 1–5: 1—Strongly disagree, and 5—Strongly agree (N = 800, mean = 2.7, SD = 1.01, Cronbach’s alpha = 0.77, range). Demographic variables . The demographic variables included self-reports of gender, age, religious affiliation, economic status, community affiliation, all as multiple-choice items. Statistical analysis . In order to assess the decision making and the justifications of the respondents we used Chi-squared and Fisher’s exact tests to determine the relationship between the variables. Logistic regression was conducted for each dilemma in order to explain the decision making factors. This study is a secondary analysis of this data, which was collected by "Eskaria” for the Haredi Institute for Public Affairs. The secondary analysis was waived by the IRB committee at the Technion: Israel Institute of Technology. Verbal consent was acquired by “Eskaria”. Research tool The questionnaire was constructed around real-life COVID-19 dilemmas to capture modes of health decision-making amidst the pandemic [ See : , ]. It also included measures of compliance, knowledge about COVID-19, and demographics. The questionnaire was constructed around real-life COVID-19 dilemmas to capture modes of health decision-making amidst the pandemic [ See : , ]. It also included measures of compliance, knowledge about COVID-19, and demographics. Content-related validity of the research tool was established non statistically using expert professional judgment, who addressed domain specification, content universe and sample, item development, item wording, and format. The questionnaire benefited from the constructive feedback of five members of the Haredi Institute for Public Affairs, which included both Haredi men and women as well as expert scholars of Haredi Judaism, all of whom are familiar with Haredi sensitivities. We also received feedback from three science communication and science education specialists, two of whom had a specific expertise in religion, with particular knowledge of Haredi society. We also received feedback from math, geography and science teachers to make sure that it resonated with general public literacy. We gave five Haredim a draft of the questionnaire and asked them to fill it out. Given the tension related to vaccine hesitancy in Israel amid the pandemic, a few voiced concerns about the conclusions people might have from the dilemmas. In response to this concern, we made sure to phrase the questions in ways that don’t direct readers to make any decisions based on the questionnaire. Participants were recruited by the market research firm “Eskaria”, an online panel with particular expertise in conducting surveys among Hebrew speaking Haredim. The response rate was 18% and the entire sample included 800 participants. The limited response rate is not surprising given the relative length of this survey, the particular challenges of accessing Haredi publics for research purposes and the timing of the survey during the pandemic. This response rate might expose our data to a differential selectivity bias and limit the external generalizability of the findings. Nevertheless, our data offers important insights and data about inner group diversity that is often glossed over in academic research. Despite these difficulties, the final sample was representative of the general Haredi population regarding proportions of subgroups (Lithuanian, Hasidim, Sephardim). In order to obtain an in-depth understanding of the varieties within ultra-Orthodox society, a random sample was chosen from Eskaria’s database that includes over 400,000 men and women in ultra-Orthodox society aged 18 and up. In terms of representativeness, the sample was similar to the general population in most categories. Almost all respondents (87%) were married, which reflects the high marriage rates in the Haredi population. Yet, the sample included more men (59%) than women (41%). While this balance was kept among Lithuanian and Sephardim, among Hasidim 82% of the respondents were male (Hasidic men 82% (N = 197), Hasidic women 18% (N = 44), Sephardi men 45% (N = 108), Sephardi women 55% (N = 132), Lithuanian men 48% (N = 109), Lithuanian women 52% (N = 120). Participants were also more educated than the Haredi average: 18% had completed a professional training course (compared with 29% in the general Haredi population), 23% had or were completing a BA degree (compared with 20% in the general Haredi population). The religious affiliation of respondents was also very similar to the general Haredi population, 30% (N = 240) belong to the Mizrahi sub-group, 29% (N = 229) to the Lithuanian community, 30% to the Hasidic community (N = 241), and 11% belongs to a variety of communities including Chabad, National-Haredi and Breslav (N = 90). Finally, we struggled to get older participants, and thus, most of our data is based on respondents between 21–44 (72%). Having said that, the Haredi population is relatively young: Haredim under the age of 17 are 53.5% of the Haredi population, compared to a national average of 28%; therefore the sample represents a very substantial part of the adult Haredi population (Haredi Institute for Public Affairs, 2020, based on data provided by the Central Bureau of Statistics). Whether collecting data about actual compliance to COVID-19 or eliciting responses to hypothetical dilemmas, this data has been derived solely from the self-reports of the respondents. In addition, the survey was circulated during a time of crisis, which might have contributed to the relatively low response rate (discussed above) and might also reflect modes of decision-making that are specific to the pandemic context. Taken together, future studies which incorporate qualitative research methods conducted during less turbulent times will contribute to further understanding health-decision making among religious and ethnic minorities. Dependent variable COVID-19 decision-making . Two dilemmas were designed to assess how Haredi men and women make COVID-related decisions amidst changing public health guidelines and recommendations. At the time, official guidelines allowed a maximum of ten people inside and twenty outside and schools were kept open. If exposed, there was a requirement to do whatever is necessary to stay home and test. Participants were asked how they would behave in social situations such as: Dilemma 1 (wedding): “Imagine your best friend scheduled a wedding for his son/daughter and the new guidelines now say that the wedding can only include twenty people outdoors. What would you recommend that they do?”. Dilemma 2 (elevator): “Your neighbor, who knew he had COVID, went into an elevator with you, without wearing a mask. As a consequence, you might have covid. Would you send your children to Talmud Torah (Jewish day school) in the next few days?”. For each dilemma, we calculated the decision-making as follows: Dilemma 1 (wedding)—Participants were presented with three options: 1. Postpone a wedding or fully adhere to wedding guidelines, 2. Recognize there is a problem and therefore have a wedding with a larger number of guests than allowed but still, less guests than one would usually have at a wedding, 3. Behave as if business is usual (and have a regular sized wedding). Based on these three options we created a new variable with two values, 1 = full compliance with guidelines (category one), 0 = partial or no compliance (categories 2 and 3). Dilemma 2 (elevator)—Participants were presented with two options: (1) To send a possibly infected child to school or (2) Not to send a possibly infected child to school. Independent variables Justification . To analyze how each respondent explained their stance on each dilemma, we presented an open-end question to the participants, asking them to explain their choices. Based on content analysis we divided these justifications into ten variables: (1) Reference to elderly population (e.g. “We should take care of the elderly population); (2) Health-related justifications . This includes a reference to health, medical or scientific argumentations, e.g. “I can carry the virus even if I feel fine”; (3) Public law or health authority recommendations (e.g. “We should follow the guidelines and it is not a joke!”); (4) Reference to special actions that the respondents will do (e.g. “We will all wear masks so it will be fine”); (5) Personal reasons (“You only get married once”!); (6) Lack of concern (“It is not really that dangerous”); (7) Religious justifications (e.g. “Only god is really in charge of the world”); (8) Public concern (“e.g. I would never want to cause harm to another person”, (9) C omparison between different situations (“As long as we are going to work, we can go to a wedding” (10) Other . Compliance with MOH (Ministry of Health) guidelines . To examine compliance with Ministry of Health guidelines (MOH), we created an index that examined the degree of compliance with individual guidelines. The index is made from a variety of statements, e.g. spatial distancing, hygiene, face-masks, and vaccination. For each statement, respondents were asked to rate the extent to which they would adhere to the guidelines between 1 (not strict at all) and 5 (very strict). The index was built as an average of four statements (N = 798, mean 3.7, SD = 0.95, range 1–5, Cronbach’s alpha = .69): “2 meters distance in public space” (N = 795, mean = 3.62, SD = 1.28), “mask wearing” (N = 796, mean = 4.34, SD = 1.06), “personal hygiene” (N = 793, mean = 3.82, SD = 1.36), “vaccination willingness” (N = 753, mean 2.36, standard deviation 1.13). Knowledge in the Context of Covid-19 . An open question about the efficacy of quarantine was used to assess knowledge about COVID-19. It was based on the response to the following question: “Any person who comes in contact with a verified patient is required to self-quarantine for 14 days from the moment of exposure, even though they had been out on the street for several days. Do you think isolation will help at this point?” Based on PISA’s definition of scientific explanation , we evaluated the application of scientific content knowledge to interpret and explain phenomena using the following categories: (1) respondent’s position regarding the efficacy of quarantine (intercoder reliability κ = .804), (2) correctness of the answer (intercoder reliability κ = .822), and (3) number and level of scientific concepts used. The concepts were characterized according to the author’s protocol , which is based on the inclusion of scientific concepts in Israeli science curricula for elementary, middle, and high school. Points were given only when respondents used scientific concepts correctly (e.g. “You can be contagious even if you do not have symptoms” versus “You are only contagious if you have symptoms”). One point was given for each elementary concept, two for each middle school–level concept, and three points for each high school–level concept. The scores for knowledge about COVID-19 ranged from 0 to 4, a higher score represents better knowledge (N = 800, mean = 2.04, SD = 1.15). Demand for knowledge . Respondents were asked to rate how much more information about COVID-19 they would like to obtain in four different topics–math, vaccines, science and geography. The variable demand for knowledge was based on the sum of four variables: 1) Demand for math knowledge, 2) Demand for knowledge about vaccines, 3) Demand for knowledge in science, and 4) Demand for knowledge in geography (N = 800, mean = 6.3, SD = 4.3, range 0–13, 0 = no lack of knowledge, 1 = missing one item of information on a subject, 2 = missing two items of information, and so on). Trust in rabbinical leadership . Respondents were asked to rate their degree of agreement between 1 (strongly disagree) and 5 (strongly agree) with the following statement: I have trust in rabbinical leadership (N = 787, mean = 4.70, SD = 0.79, range 1–5). Trust in Israeli government . To examine trust in Israel’s government, we created an index based on the average of four questions: 1) I have faith in the Israeli government, 2) The government acts in an equal manner towards all citizens, 3) The Israeli government is making the right decisions for the residents of Israel with regard to the COVID-19 crisis, 4) Community restrictions following the COVID-19 crisis should be determined by the government. The variable ranks from 1–5: 1—Strongly disagree, and 5—Strongly agree (N = 800, mean = 2.7, SD = 1.01, Cronbach’s alpha = 0.77, range). Demographic variables . The demographic variables included self-reports of gender, age, religious affiliation, economic status, community affiliation, all as multiple-choice items. Statistical analysis . In order to assess the decision making and the justifications of the respondents we used Chi-squared and Fisher’s exact tests to determine the relationship between the variables. Logistic regression was conducted for each dilemma in order to explain the decision making factors. COVID-19 decision-making . Two dilemmas were designed to assess how Haredi men and women make COVID-related decisions amidst changing public health guidelines and recommendations. At the time, official guidelines allowed a maximum of ten people inside and twenty outside and schools were kept open. If exposed, there was a requirement to do whatever is necessary to stay home and test. Participants were asked how they would behave in social situations such as: Dilemma 1 (wedding): “Imagine your best friend scheduled a wedding for his son/daughter and the new guidelines now say that the wedding can only include twenty people outdoors. What would you recommend that they do?”. Dilemma 2 (elevator): “Your neighbor, who knew he had COVID, went into an elevator with you, without wearing a mask. As a consequence, you might have covid. Would you send your children to Talmud Torah (Jewish day school) in the next few days?”. For each dilemma, we calculated the decision-making as follows: Dilemma 1 (wedding)—Participants were presented with three options: 1. Postpone a wedding or fully adhere to wedding guidelines, 2. Recognize there is a problem and therefore have a wedding with a larger number of guests than allowed but still, less guests than one would usually have at a wedding, 3. Behave as if business is usual (and have a regular sized wedding). Based on these three options we created a new variable with two values, 1 = full compliance with guidelines (category one), 0 = partial or no compliance (categories 2 and 3). Dilemma 2 (elevator)—Participants were presented with two options: (1) To send a possibly infected child to school or (2) Not to send a possibly infected child to school. Justification . To analyze how each respondent explained their stance on each dilemma, we presented an open-end question to the participants, asking them to explain their choices. Based on content analysis we divided these justifications into ten variables: (1) Reference to elderly population (e.g. “We should take care of the elderly population); (2) Health-related justifications . This includes a reference to health, medical or scientific argumentations, e.g. “I can carry the virus even if I feel fine”; (3) Public law or health authority recommendations (e.g. “We should follow the guidelines and it is not a joke!”); (4) Reference to special actions that the respondents will do (e.g. “We will all wear masks so it will be fine”); (5) Personal reasons (“You only get married once”!); (6) Lack of concern (“It is not really that dangerous”); (7) Religious justifications (e.g. “Only god is really in charge of the world”); (8) Public concern (“e.g. I would never want to cause harm to another person”, (9) C omparison between different situations (“As long as we are going to work, we can go to a wedding” (10) Other . Compliance with MOH (Ministry of Health) guidelines . To examine compliance with Ministry of Health guidelines (MOH), we created an index that examined the degree of compliance with individual guidelines. The index is made from a variety of statements, e.g. spatial distancing, hygiene, face-masks, and vaccination. For each statement, respondents were asked to rate the extent to which they would adhere to the guidelines between 1 (not strict at all) and 5 (very strict). The index was built as an average of four statements (N = 798, mean 3.7, SD = 0.95, range 1–5, Cronbach’s alpha = .69): “2 meters distance in public space” (N = 795, mean = 3.62, SD = 1.28), “mask wearing” (N = 796, mean = 4.34, SD = 1.06), “personal hygiene” (N = 793, mean = 3.82, SD = 1.36), “vaccination willingness” (N = 753, mean 2.36, standard deviation 1.13). Knowledge in the Context of Covid-19 . An open question about the efficacy of quarantine was used to assess knowledge about COVID-19. It was based on the response to the following question: “Any person who comes in contact with a verified patient is required to self-quarantine for 14 days from the moment of exposure, even though they had been out on the street for several days. Do you think isolation will help at this point?” Based on PISA’s definition of scientific explanation , we evaluated the application of scientific content knowledge to interpret and explain phenomena using the following categories: (1) respondent’s position regarding the efficacy of quarantine (intercoder reliability κ = .804), (2) correctness of the answer (intercoder reliability κ = .822), and (3) number and level of scientific concepts used. The concepts were characterized according to the author’s protocol , which is based on the inclusion of scientific concepts in Israeli science curricula for elementary, middle, and high school. Points were given only when respondents used scientific concepts correctly (e.g. “You can be contagious even if you do not have symptoms” versus “You are only contagious if you have symptoms”). One point was given for each elementary concept, two for each middle school–level concept, and three points for each high school–level concept. The scores for knowledge about COVID-19 ranged from 0 to 4, a higher score represents better knowledge (N = 800, mean = 2.04, SD = 1.15). Demand for knowledge . Respondents were asked to rate how much more information about COVID-19 they would like to obtain in four different topics–math, vaccines, science and geography. The variable demand for knowledge was based on the sum of four variables: 1) Demand for math knowledge, 2) Demand for knowledge about vaccines, 3) Demand for knowledge in science, and 4) Demand for knowledge in geography (N = 800, mean = 6.3, SD = 4.3, range 0–13, 0 = no lack of knowledge, 1 = missing one item of information on a subject, 2 = missing two items of information, and so on). Trust in rabbinical leadership . Respondents were asked to rate their degree of agreement between 1 (strongly disagree) and 5 (strongly agree) with the following statement: I have trust in rabbinical leadership (N = 787, mean = 4.70, SD = 0.79, range 1–5). Trust in Israeli government . To examine trust in Israel’s government, we created an index based on the average of four questions: 1) I have faith in the Israeli government, 2) The government acts in an equal manner towards all citizens, 3) The Israeli government is making the right decisions for the residents of Israel with regard to the COVID-19 crisis, 4) Community restrictions following the COVID-19 crisis should be determined by the government. The variable ranks from 1–5: 1—Strongly disagree, and 5—Strongly agree (N = 800, mean = 2.7, SD = 1.01, Cronbach’s alpha = 0.77, range). Demographic variables . The demographic variables included self-reports of gender, age, religious affiliation, economic status, community affiliation, all as multiple-choice items. Statistical analysis . In order to assess the decision making and the justifications of the respondents we used Chi-squared and Fisher’s exact tests to determine the relationship between the variables. Logistic regression was conducted for each dilemma in order to explain the decision making factors. In what follows, we showcase our main findings. To allow readers to capture the depth of each dilemma which presents a different type of conflict (personal decision vs. recommendation for friend; official guideline versus recommendation; potential conflict vs. actual conflict), each dilemma is presented separately and their results will be compared in the discussion section. Wedding dilemma The first dilemma asked participants to state what advice they would give their best friend in a case where their son or daughter had planned a wedding during the pandemic. At that time, public health regulations capped inside public gatherings at twenty people and our participants were asked what they would recommend. In Judaism, marriage is valued highly and going to weddings especially among members of one’s extended family is very important. To answer this dilemma, participants were given three options—postpone the wedding, have the wedding with less guests than intended, or have a regular-sized wedding as usual. 51% of the participants reported that they take guidelines into consideration and fully comply with them. Yet, examining the responses of men and women within different Haredi sub-groups reveal big differences in adherence to social distance guidelines regarding weddings. Among Lithuanian and Sephardi haredim, most adhered to guidelines (58% and 63%, respectively), whereas only 33% of Hasidic Jews reported that they would follow guidelines (X 2 = 60.23, p = .00, df = 4, N = 681). We also found that gender affected decision making as 61% of the women stated they would adhere to guidelines, compared to 45% of the men (X 2 = 30.8, p = .00, df = 2, N = 769). When combining gender with religious affiliation, we found that Hasidic men are the group that reports the lowest level of compliance among haredi men (32%, X 2 = 26.3, p = .00, df = 4, N = 393). In more detail (See ), Hasidic Haredi men reported that they would either fully adhere to guidelines (32%) or partially adhere to guidelines by celebrating with less participants (48%). Due to limitations with the number of observations in this question (described above), we conducted Fisher’s exact test for each sub-group to test for gender-related differences and did not find any significant differences within each sub-group (Sephardi, p = 0.07, N = 235; Lithuanian, p = 0.07, N = 221; Hasidic: p = 0.15, N = 225). When asked to explain their decision, participants shared several justifications. The respondents’ decision-making rationalizations were classified into 10 variables, in order of prevalence: (1) Health-related justifications —29.1% (2) Public law or health authority recommendations —25.3% (3) Lack of concern -12.7% (4) Religious justifications —12.5% (5) Reference to special actions that the respondents will do—6.5% (6) Personal reasons— 5.1% (7) Public concern —4.2% (8) C omparison between different situations —2.3% (9) Reference to high-risk populations public concern— 2.2% (10) Other —0.2%. Within all communities, the most common justification was health-related. Health justifications came in different forms. Some respondents spoke about their own health “We must think about our health!” or “We need to stop the spreading of the virus!”. Others mentioned the importance of caring for special populations, especially the elderly. As Chava, a married, 33 year old Lithuanian woman put it: “We will not push off a wedding! We need to have the wedding as planned, but find a way to follow the guidelines and be careful”. When comparing different communities with regard to the justifications they used, we found that there is a significant difference between Hasidim and their Lithuanian and Sephardi counterparts. Reflecting this difference, Hasidim use less health justifications (21%) in comparison to the other communities (Sephardi 31% and Lithuanian 30%, X 2 = 7.7, p = .02, df = 2, N = 710). While some Hasidim noted the importance of following guidelines, many responded that guidelines were politicized. For example, Baruch, a 33 year old Viznitz Hasid responded: “It is all politics”, and Avi, a 40 year old from Bet Shemesh explained: “We don’t have any trust in the state. They are full of nonsense!”, reflecting previous studies linking public health guidelines and state tensions , a point we will return to in the discussion section. In order to examine the relationship between the various factors that predict decision-making in each dilemma, we conducted a regression logistic analysis. To do so, we use four different models to discern the effect of different groups of variables. The first column (model 1) presents the following demographic controls: age, gender, community affiliation and economic index. To check how education affects decision-making, the second column (model 2) takes the demographic controls (described previously) but also adds the following controls: demand for knowledge and COVID-19 knowledge. In the third column (model 3), we further included the following three controls: compliance with the MOH, trust in the Israeli government and trust in rabbinic leadership. Finally, in the fourth column (model 4) we also included: health-related and religious justifications. In , we present the results for dilemma 1 (wedding). In the first column, gender has a positive significant relationship. Women are 37% more likely to follow guidelines than men. This effect wears off as we added the other types of specifications. Further, the main finding in all models is that religious affiliation is the most significant predictor. We found that both Sephardim and Lithuanians are much more likely to postpone a wedding than their Hasidic counterparts. Compared to Hasidim, Sephardim are three times more likely and Lithuanians are twice more likely to follow guidelines. Our findings also show that while the demand for knowledge shows a small effect at the start, this effect disappears after controlling for compliance and trust. However, we found that COVID-19-related knowledge shows a positive effect. As knowledge about COVID increases, the likelihood of recommending postponing the wedding increases by 32%. This result holds for the rest of the specifications, although the effect is reduced to 19–17% (columns 3 and 4, respectively). Further, compliance with the MOH was found to have the second strongest effect. Greater compliance with guidelines more than doubles the likelihood to recommend to postpone the wedding. When analyzing trust, the only significant one was trust in rabbinical leadership, which shows a negative relationship. Higher trust in rabbinic leadership decreases the likelihood to recommend to postpone the wedding by thirty percent. Also, health-related justifications present a positive relationship increasing the decision to recommend postponing the wedding by 72%. Finally, we used Shorrocks-Shapely value decomposition to demonstrate the relative contribution of each variable to the model. We found that compliance with MOH guidelines (52.7%), followed by Sephardi religious affiliation (12.3%), and health related justifications (8.33%) are the most influential variables for dilemma 1. Talmud Torah (Jewish day school) dilemma The second dilemma asked participants what they would do if they were in an elevator with an unmasked neighbor, who they later found out had COVID-19. After posing this scenario, in which they were potentially exposed to COVID-19, we asked respondents whether they would send their children to school or not. Our findings reveal that most of the participants (77%) reported that they consider public health guidelines and comply with them. Among those who reported they would send their child to school regardless of the elevator encounter, Hasidim were most likely to defy guidelines (31%), compared to their Lithuanian (23%) and Mizrahi counterparts (15%) (X 2 = 16.84, p = .00, df = 2, N = 710). In this dilemma, gender was not found to be statistically significant (X 2 = 1.87, p = 0.17, df = 1, N = 710). When asked to explain their decision, participants shared several justifications (See ). The respondents’ decision-making rationalizations were classified into 10 variables, in order of prevalence: (1) Public concern —32.1% (2) Health-related justifications —27.5% (3) Lack of concern —14.8% (4) Public law or health authority recommendations —13.9% (5) Reference to special actions that the respondents will do—5.8% (6) Religious justifications —2.2% (7) Personal reasons— 1% (8) C omparison between different situations —0.6% (9) Reference to high-risk populations public concern— 0% (10) Other —0%. In this dilemma, no significant differences were found between the different streams in using both religious and health justifications. In contrast to the previous dilemma, in this dilemma the leading justification offered by participants was public concern. Public concern appeared in various forms: “We would not want to harm others”, “I would not want to be responsible for making other people sick”. As this dilemma was centered on sending ones’ child to school, the focus on harming others makes sense and was closely followed by health justifications, as Yoel, a twenty-eight year old Sephardi man from Bnei-Brak put it: “Health comes first!”. Similarly to the first dilemma, we also examined the relationship between the various factors that predict decision-making in this second dilemma. presents the results for dilemma two. We found that age and gender did not have significant effects on the decision to send a possibly infected kid to school. Similarly to the first dilemma, religious affiliation is a central finding. Sephardi affiliation doubles the likelihood to keep a child at home, compared to Hasidim. Among Lithuanians, there is a positive effect in model one and two, which disappears in models (3) and (4). While knowledge has less of an effect in this dilemma, compliance with MOH reveals a positive strong effect. Those who comply with MOH guidelines are 86% more likely to keep their child home. We also found that both religious and health-related justification were significant: Health-related justifications are 27 times more likely to increase the likelihood to keep a child at home; and religious-related justifications decrease the likelihood to keep their child at home by seven times. Finally, we use Shorrocks-Shapely value decomposition to demonstrate the relative contribution of the variable to the model. We found that health-related justifications (48.5%), followed by compliance with MOH guidelines (28.7%) and Sephardi religious affiliation (6.2%), are the most influential variables for this dilemma. The first dilemma asked participants to state what advice they would give their best friend in a case where their son or daughter had planned a wedding during the pandemic. At that time, public health regulations capped inside public gatherings at twenty people and our participants were asked what they would recommend. In Judaism, marriage is valued highly and going to weddings especially among members of one’s extended family is very important. To answer this dilemma, participants were given three options—postpone the wedding, have the wedding with less guests than intended, or have a regular-sized wedding as usual. 51% of the participants reported that they take guidelines into consideration and fully comply with them. Yet, examining the responses of men and women within different Haredi sub-groups reveal big differences in adherence to social distance guidelines regarding weddings. Among Lithuanian and Sephardi haredim, most adhered to guidelines (58% and 63%, respectively), whereas only 33% of Hasidic Jews reported that they would follow guidelines (X 2 = 60.23, p = .00, df = 4, N = 681). We also found that gender affected decision making as 61% of the women stated they would adhere to guidelines, compared to 45% of the men (X 2 = 30.8, p = .00, df = 2, N = 769). When combining gender with religious affiliation, we found that Hasidic men are the group that reports the lowest level of compliance among haredi men (32%, X 2 = 26.3, p = .00, df = 4, N = 393). In more detail (See ), Hasidic Haredi men reported that they would either fully adhere to guidelines (32%) or partially adhere to guidelines by celebrating with less participants (48%). Due to limitations with the number of observations in this question (described above), we conducted Fisher’s exact test for each sub-group to test for gender-related differences and did not find any significant differences within each sub-group (Sephardi, p = 0.07, N = 235; Lithuanian, p = 0.07, N = 221; Hasidic: p = 0.15, N = 225). When asked to explain their decision, participants shared several justifications. The respondents’ decision-making rationalizations were classified into 10 variables, in order of prevalence: (1) Health-related justifications —29.1% (2) Public law or health authority recommendations —25.3% (3) Lack of concern -12.7% (4) Religious justifications —12.5% (5) Reference to special actions that the respondents will do—6.5% (6) Personal reasons— 5.1% (7) Public concern —4.2% (8) C omparison between different situations —2.3% (9) Reference to high-risk populations public concern— 2.2% (10) Other —0.2%. Within all communities, the most common justification was health-related. Health justifications came in different forms. Some respondents spoke about their own health “We must think about our health!” or “We need to stop the spreading of the virus!”. Others mentioned the importance of caring for special populations, especially the elderly. As Chava, a married, 33 year old Lithuanian woman put it: “We will not push off a wedding! We need to have the wedding as planned, but find a way to follow the guidelines and be careful”. When comparing different communities with regard to the justifications they used, we found that there is a significant difference between Hasidim and their Lithuanian and Sephardi counterparts. Reflecting this difference, Hasidim use less health justifications (21%) in comparison to the other communities (Sephardi 31% and Lithuanian 30%, X 2 = 7.7, p = .02, df = 2, N = 710). While some Hasidim noted the importance of following guidelines, many responded that guidelines were politicized. For example, Baruch, a 33 year old Viznitz Hasid responded: “It is all politics”, and Avi, a 40 year old from Bet Shemesh explained: “We don’t have any trust in the state. They are full of nonsense!”, reflecting previous studies linking public health guidelines and state tensions , a point we will return to in the discussion section. In order to examine the relationship between the various factors that predict decision-making in each dilemma, we conducted a regression logistic analysis. To do so, we use four different models to discern the effect of different groups of variables. The first column (model 1) presents the following demographic controls: age, gender, community affiliation and economic index. To check how education affects decision-making, the second column (model 2) takes the demographic controls (described previously) but also adds the following controls: demand for knowledge and COVID-19 knowledge. In the third column (model 3), we further included the following three controls: compliance with the MOH, trust in the Israeli government and trust in rabbinic leadership. Finally, in the fourth column (model 4) we also included: health-related and religious justifications. In , we present the results for dilemma 1 (wedding). In the first column, gender has a positive significant relationship. Women are 37% more likely to follow guidelines than men. This effect wears off as we added the other types of specifications. Further, the main finding in all models is that religious affiliation is the most significant predictor. We found that both Sephardim and Lithuanians are much more likely to postpone a wedding than their Hasidic counterparts. Compared to Hasidim, Sephardim are three times more likely and Lithuanians are twice more likely to follow guidelines. Our findings also show that while the demand for knowledge shows a small effect at the start, this effect disappears after controlling for compliance and trust. However, we found that COVID-19-related knowledge shows a positive effect. As knowledge about COVID increases, the likelihood of recommending postponing the wedding increases by 32%. This result holds for the rest of the specifications, although the effect is reduced to 19–17% (columns 3 and 4, respectively). Further, compliance with the MOH was found to have the second strongest effect. Greater compliance with guidelines more than doubles the likelihood to recommend to postpone the wedding. When analyzing trust, the only significant one was trust in rabbinical leadership, which shows a negative relationship. Higher trust in rabbinic leadership decreases the likelihood to recommend to postpone the wedding by thirty percent. Also, health-related justifications present a positive relationship increasing the decision to recommend postponing the wedding by 72%. Finally, we used Shorrocks-Shapely value decomposition to demonstrate the relative contribution of each variable to the model. We found that compliance with MOH guidelines (52.7%), followed by Sephardi religious affiliation (12.3%), and health related justifications (8.33%) are the most influential variables for dilemma 1. The second dilemma asked participants what they would do if they were in an elevator with an unmasked neighbor, who they later found out had COVID-19. After posing this scenario, in which they were potentially exposed to COVID-19, we asked respondents whether they would send their children to school or not. Our findings reveal that most of the participants (77%) reported that they consider public health guidelines and comply with them. Among those who reported they would send their child to school regardless of the elevator encounter, Hasidim were most likely to defy guidelines (31%), compared to their Lithuanian (23%) and Mizrahi counterparts (15%) (X 2 = 16.84, p = .00, df = 2, N = 710). In this dilemma, gender was not found to be statistically significant (X 2 = 1.87, p = 0.17, df = 1, N = 710). When asked to explain their decision, participants shared several justifications (See ). The respondents’ decision-making rationalizations were classified into 10 variables, in order of prevalence: (1) Public concern —32.1% (2) Health-related justifications —27.5% (3) Lack of concern —14.8% (4) Public law or health authority recommendations —13.9% (5) Reference to special actions that the respondents will do—5.8% (6) Religious justifications —2.2% (7) Personal reasons— 1% (8) C omparison between different situations —0.6% (9) Reference to high-risk populations public concern— 0% (10) Other —0%. In this dilemma, no significant differences were found between the different streams in using both religious and health justifications. In contrast to the previous dilemma, in this dilemma the leading justification offered by participants was public concern. Public concern appeared in various forms: “We would not want to harm others”, “I would not want to be responsible for making other people sick”. As this dilemma was centered on sending ones’ child to school, the focus on harming others makes sense and was closely followed by health justifications, as Yoel, a twenty-eight year old Sephardi man from Bnei-Brak put it: “Health comes first!”. Similarly to the first dilemma, we also examined the relationship between the various factors that predict decision-making in this second dilemma. presents the results for dilemma two. We found that age and gender did not have significant effects on the decision to send a possibly infected kid to school. Similarly to the first dilemma, religious affiliation is a central finding. Sephardi affiliation doubles the likelihood to keep a child at home, compared to Hasidim. Among Lithuanians, there is a positive effect in model one and two, which disappears in models (3) and (4). While knowledge has less of an effect in this dilemma, compliance with MOH reveals a positive strong effect. Those who comply with MOH guidelines are 86% more likely to keep their child home. We also found that both religious and health-related justification were significant: Health-related justifications are 27 times more likely to increase the likelihood to keep a child at home; and religious-related justifications decrease the likelihood to keep their child at home by seven times. Finally, we use Shorrocks-Shapely value decomposition to demonstrate the relative contribution of the variable to the model. We found that health-related justifications (48.5%), followed by compliance with MOH guidelines (28.7%) and Sephardi religious affiliation (6.2%), are the most influential variables for this dilemma. Most participants reported that they follow MOH COVID-19 health regulations (with average of 3.7 out of 5 in the index). This finding resonates with a number of surveys demonstrating that Israelis were relatively compliant during COVID-19 . In tandem with these findings, our survey showed that Haredi compliance was consistent in both dilemmas, which involved vastly different situations and presented different types of conflicts. Our findings also join a growing wave of scholarship that shows that Haredim followed public health guidelines, in sharp contrast to local and national media reports, which offered an overarching public depiction of Haredi non-compliance . Among the respondents who were non-compliant, we found large divergences which largely reflected religious affiliation. While Lithuanian and Sephardi communities tended to follow guidelines, their Hasidic counterparts were much more likely to flout social distancing guidelines, especially Hasidic men. It is important to reiterate, however, that our findings are based on a relatively low response rate that it is exposed to differential selectivity and further studies are required to generalize these findings about inner Haredi diversity. Notwithstanding, how can we make sense of this inner diversity? And, what are the broader ramifications of these findings? Hasidic Judaism, as noted above, is an umbrella term to include many sub-groups, such as: Belz, Bretslov, Lubavitch (Chabad), Sanz, Satmar, among others. Each Hasidic court is headed by a different Rebbe, called an admor , contributing to a decentralized infrastructure of rabbinic authority and distinct customs regarding prayer, dress, melodies and more. Many of the large Hasidic courts have deep links with their counterparts outside of Israel, especially in the US, which also influenced adherence to public health guidelines during the pandemic. Broadly speaking, Hasidic religiosity can be compared to other charismatic movements, which place a great emphasis on personal experience, emotions and spontaneity as well as ecstatic group worship. A recent study examining Hasidic behavior during the pandemic examined the narratives of Breslov Hasidim who decided to continue their yearly pilgrimage to the Tomb of Rabbi Nachman of Breslov (1772–1810) in Ukraine, despite the closure of state borders. As the Ukrainian government announced the closure aimed at reducing contamination, many Breslovers attempted to travel before the closure to make their yearly pilgrimage. This resulted in thousands of Breslovers stranded in airports, land borders and even imprisoned in the days and weeks leading up to the annual pilgrimage on the Jewish new year. Anthropologist Rachel Feldman argues that this choice was not a project of science denial but rather rooted in a conflict between state guidelines and religious practice. While Breslov Hasidim are considered one of the more ‘extreme’ Hasidic groups, Feldman’s analysis pushes us to think about the particular sensibilities of specific Jewish groups, especially the more charismatic and group-oriented ones, whose needs were not specifically targeted in public health guidelines [ Also see ]. As she notes, this chaotic pilgrimage is a painful mirror of what happens “when secular logics fail to contain and properly modify religious actors” [ , p.107 ]. Our findings resonate with this analysis. As Hasidic Jews followed public guidelines less than their Lithuanian and Sephardi counterparts, public health messaging that treats all sub-groups as one shade of black misses the inner group diversity, a misconception that might have vast ramifications. Black and colleagues have shown that when Israeli authorities imposed closures of areas with high morbidity rates, which included much higher representation of ultra-Orthodox locations such as Bnei Brak (whose morbidity levels were much higher than those of the general population), the Israeli ultra-Orthodox population experienced a wave of frustration and expressed feelings of perceived discrimination in their sectorial press [ See , ]. In addition, Folmer and colleagues have demonstrated how adherence to public health guidelines is also influenced by the behavior of other people in one’s community . Following this logic, public portrayal of non-compliance creates a (misleading) impression that violations are normal among Haredim, which in turn has the potential to contribute to the growth of non-compliance from within. To be clear, our aim here is not to merely redirect blame. On the contrary, highlighting Haredi variety demonstrates the wide array of ideas and responses to the pandemic that must be accounted for in public health relations. Following Kasstan et al. , we argue that creating sustainable relationships between communal custodians and other positioned stakeholders can foster better understandings and collaboration in future events . For example, a more targeted and diversified public health intervention, one that does not put all Haredim in the same (non-compliant) boat, would have been more conducive [ Also see :, ]. In fact, Israel’s Ministry of Health began a culturally specific science and health section to promote science and health communication on COVID-19 vaccination, which helped lower infection levels . Finally, we also found that knowledge about COVID-19 (especially in the first dilemma) predicts compliance, even more than trust in religious leadership. Whereas many researchers note that Haredi Jews “refer to their rabbis for decision-making regarding medical procedures and screenings, so as to act in accordance with religious requirements” , our finding echoes recent scholarship that shows that decisions are not merely followed blindly, but rather negotiated in everyday life . In contrast to this view, we found that health-related justifications and public concern were utilized by respondents who reported following public guidelines. In other words, health and religious frameworks were fused together in the decision-making process of respondents. Taking these findings together, tailoring knowledge about the pandemic to specific group sensibilities might be the key to developing and implementing sustainable community-focused interventions, collaborations and public health programs . For example, awareness of religious sensitivities and temporalities (e.g. New Year, as described above) must be incorporated in efforts to increase compliance amid diverse populations. Science and health communication that acknowledges the role of religious dogma, practice and observance while providing critical knowledge about the pandemic, can help us better prepare for future occurrences which will likely follow. |
Analysis of preoperative ocular optical parameters in patients with cataract | 9c09720d-b8a2-4cfd-a52b-7e6d1602446e | 11912591 | Ophthalmologic Surgical Procedures[mh] | Cataract is currently the second leading cause of visual impairment and the primary cause of blindness worldwide. Corneal curvature is a critical parameter for calculating the refractive power of an artificial lens. With the substantial increase in the number of cataract surgeries, a thorough and accurate assessment of corneal surface morphology and its changes is crucial for cataract treatment. Regarding studies on corneal curvature changes with age, Ma et al. examined healthy Chinese individuals and reported gradual alterations in corneal morphology with age, with both anterior and posterior corneal curvature (K1 and K2) being positively correlated with age. According to research predictions, the number of individuals (aged 45–89 years) affected by any form of cataract will reach 240.83 million by 2050 . There remain discrepancies between values obtained when measuring various biological indicators in patients with cataract, including the reference ranges that vary between measurements. Corneal curvature and the distribution of corneal astigmatism are important parameters in guiding the calculation of intraocular lens power in cataract surgery and determining the choice of surgical incisions. SimK is widely used in current clinical practice and is calculated using anterior corneal surface measurements and the presumption of a constant anterior–posterior corneal radius ratio. Early studies demonstrated that a conversion factor of 1.3375 was insufficient for converting the radius of curvature into the absolute refractive power of the cornea, which could lead to the overestimation of the corneal curvature measured by the SimK method. While age affects multiple ocular parameters due to physiological aging, the influence of sex on ocular biometry is more subtle, with corneal curvature being one of the few parameters showing a significant and consistent gender-based difference . Wavefront aberration is generally considered a crucial indicator for assessing visual quality . However, the parameters obtained in different tissues may vary according to factors such as geographic region, race, age, and differences in inclusion criteria . For example, the anterior chamber volume and depth in males are typically greater than those in females . Moreover, lens thickness tends to increase with age. Additionally, the anterior chamber depth has a negative correlation with age . Nonetheless, there have been few epidemiological studies on patients with cataract, and the information is limited to local reports. Further, there have been limited biological parameters included in these studies . Therefore, with the development of personalized surgery, individual ocular biometry data are becoming particularly crucial. In recent years, Total Corneal Refractive Power (TCRP) has gained interest among cataract and refractive surgeons. Based on the principle of ray tracing and incorporating data regarding corneal thickness and posterior corneal curvature, the Scheimpflug camera allows physical measurement of TCRP. This theoretically facilitates more accurate calculations . To supply this critical basic information, the Pentacam AXL for anterior segment analysis is used to evaluate the characteristics and distribution of preoperative corneal biological parameters in patients with cataract. It is stated that Pentacam AXL demonstrates high agreement with measurements from other methods, which was quantitatively assessed using Spearman’s correlation analysis. The agreement was found to be strong (Spearman’s ρ = 0.85, p < 0.001), indicating a high level of consistency between the Pentacam AXL measurements and those from previous studies. This demonstrates that the Pentacam AXL provides reliable measurements that are comparable with other established methods [ – ]. The results of such analysis can inform the establishment of individualized surgical plans based on patient-specific corneal characteristics in clinical practice, allowing for satisfactory visual quality post-operation. Therefore, this study aimed to assess the distribution of preoperative corneal biological parameters among patients with cataract in Shenzhen, China, and to explore their association with risk factors.
Comparison of the anterior corneal surface and whole cornea curvatures in right eyes Flat axis curvature (k1), steep axis curvature (k2), and mean corneal curvature (km) of the simulated keratometry (SimK), as well as total corneal refractive power (TCRP) of the right eye, were assessed for normal distribution in 1,255 patients with a mean age of 52.9 ± 21.3 years (Table ). Given their non-normal distribution, a Wilcoxon signed-rank test was applied. There was a significant difference in all corneal k1, k2, and km values between SimK and TCRP. Age was positively correlated with the km values of both the SimK (correlation coefficient r = 0.56199) and TCRP (r = 0.25064) (both, P < 0.01) (Fig. ). Comparison of higher-order aberrations in the anterior corneal surface of right eyes across different age groups Comparison of corneal spherical aberrations in the right eyes across different age groups revealed variations in the spherical aberrations of the anterior surface, posterior surface, and total cornea increased with age. There were significant among-group differences in spherical aberrations on the anterior surface, posterior surface, and total cornea ( P < 0.001). Furthermore, Spearman’s correlation analysis demonstrated a positive correlation between these parameters and age (Table ). Comparison of astigmatism axis changes in right eyes across different age groups The proportion of with-the-rule astigmatism showed a decreasing trend as age increased, whereas oblique astigmatism showed a notable increase in the 40–60 years group, which plateaued in patients aged > 60 years. Meanwhile, ATR astigmatism showed a notable increase in patients aged 40–60 years and continued to increase in patients aged > 60 years before it eventually stabilized (Fig. ). Comparison of higher-order aberrations in the anterior corneal surface between both eyes A significant correlation was observed between spherical aberration (Z40, r = 0.79881), horizontal coma (Z31, r = -0.59467), vertical coma (Z3-1, r = 0.74806), and horizontal trefoil (Z33, r = -0.37542) between the eyes. There was no significant between-eye correlation of the total HOA (r = 0.72091) and oblique trefoil (Z3-3, r = 0.51390) (Fig. ).” Relationship between corneal curvature and sex The mean km of both the SimK and TCRP were found to be larger in females than in males. The corneal curvatures in females were significantly steeper than those in males ( P < 0.01) (Fig. ).
Flat axis curvature (k1), steep axis curvature (k2), and mean corneal curvature (km) of the simulated keratometry (SimK), as well as total corneal refractive power (TCRP) of the right eye, were assessed for normal distribution in 1,255 patients with a mean age of 52.9 ± 21.3 years (Table ). Given their non-normal distribution, a Wilcoxon signed-rank test was applied. There was a significant difference in all corneal k1, k2, and km values between SimK and TCRP. Age was positively correlated with the km values of both the SimK (correlation coefficient r = 0.56199) and TCRP (r = 0.25064) (both, P < 0.01) (Fig. ).
Comparison of corneal spherical aberrations in the right eyes across different age groups revealed variations in the spherical aberrations of the anterior surface, posterior surface, and total cornea increased with age. There were significant among-group differences in spherical aberrations on the anterior surface, posterior surface, and total cornea ( P < 0.001). Furthermore, Spearman’s correlation analysis demonstrated a positive correlation between these parameters and age (Table ).
The proportion of with-the-rule astigmatism showed a decreasing trend as age increased, whereas oblique astigmatism showed a notable increase in the 40–60 years group, which plateaued in patients aged > 60 years. Meanwhile, ATR astigmatism showed a notable increase in patients aged 40–60 years and continued to increase in patients aged > 60 years before it eventually stabilized (Fig. ).
A significant correlation was observed between spherical aberration (Z40, r = 0.79881), horizontal coma (Z31, r = -0.59467), vertical coma (Z3-1, r = 0.74806), and horizontal trefoil (Z33, r = -0.37542) between the eyes. There was no significant between-eye correlation of the total HOA (r = 0.72091) and oblique trefoil (Z3-3, r = 0.51390) (Fig. ).”
The mean km of both the SimK and TCRP were found to be larger in females than in males. The corneal curvatures in females were significantly steeper than those in males ( P < 0.01) (Fig. ).
We used the Pentacam instrument to measure and compare SimK and TCRP, specifically analyzing flat axis curvature, steep axis curvature, and mean corneal curvature. Significant differences were observed between SimK and TCRP values. Previous research comparing SimK and Pentacam HR ray-tracing found that SimK values were 0.40–0.60 D higher than TCRP . Savini et al. reported that this discrepancy depends on the anterior–posterior corneal radius ratio. In our study, mean corneal curvature assessed using SimK and TCRP showed significant variation. As age increased, both SimK and TCRP exhibited a gradual rise in mean corneal curvature. Linear regression indicated that the SimK-TCRP discrepancy decreased with age, at a rate of 0.00965 per year. Prior studies confirm that anterior and posterior corneal surfaces undergo significant changes with aging. Our findings align with these reports, revealing that corneal curvature increases with age, primarily due to anterior corneal surface modifications. Hashemi et al. noted that anterior surface changes were more pronounced than posterior ones, supporting our inference that total corneal curvature shifts are predominantly driven by anterior corneal surface alterations. The average anterior corneal spherical aberration (0.276 ± 0.165 μm) in this study is consistent with findings from Elkitkat et al. (+0.26 ± 0.12 μm), Beiko et al. (0.274 ± 0.089 μm), and Wang et al. in Chinese patients. Anterior spherical aberration increased with age, while posterior spherical aberration declined, leading to an overall rise in total corneal spherical aberration. This suggests that the posterior corneal surface may play a compensatory role in balancing corneal aberrations . Unlike previous studies , our results showed a substantial proportion of eyes with negative spherical aberration, possibly due to differences in patient demographics or cataract severity. Kuroda et al. suggested that nuclear cataracts cause local refractive changes, increasing higher-order aberrations. These changes may explain the observed variation in corneal spherical aberration, as nuclear cataracts alter wavefront aberration. Our methodology, which did not classify eyes based on lens opacity type, may have contributed to the divergence from previous findings . Corneal spherical aberration significantly influences visual quality. Aspheric intraocular lenses (IOLs) aim to counteract these aberrations, targeting a postoperative whole-eye aberration of 0–0.1 μm for optimal visual outcomes . However, our study highlights substantial individual differences in corneal spherical aberration among cataract patients, suggesting that a standardized correction value may not be suitable for all cases. Factors such as pupil size, stromal pocket condition, and postoperative artificial lens positioning contribute to patient-specific variations in visual outcomes. This may explain why some patients remain dissatisfied with vision quality despite aspheric IOL implantation. Given these findings, preoperative measurement of corneal spherical aberration is crucial for selecting appropriate aspheric IOLs. A personalized approach considering multiple patient factors may improve postoperative vision and enhance overall satisfaction. Age-related changes in corneal astigmatism are well-documented, with studies indicating a shift from with-the-rule (WTR) to against-the-rule (ATR) astigmatism over time . Our study corroborates these findings, showing a decline in WTR astigmatism and an increase in ATR and oblique astigmatism with advancing age. These shifts align with previous research and emphasize the importance of incorporating age-related astigmatic changes into surgical planning. Koch et al. proposed that younger individuals benefit from posterior corneal astigmatism compensating for anterior corneal astigmatism. However, as age progresses, anterior astigmatism shifts toward ATR, potentially leading to under-correction if only anterior corneal curvature is considered. Additional factors, such as changes in eye wall tension, eyelid pressure, and extraocular muscle traction, may also contribute to age-related refractive alterations . Our results suggest that corneal astigmatism is dynamic and should be carefully evaluated when planning cataract surgery, particularly in older patients. Our study analyzed higher-order aberrations (HOAs) in 873 patients and found significant interocular correlations for spherical aberration (Z40), horizontal coma (Z31), vertical coma (Z3-1), and horizontal trefoil (Z33). These findings support prior research indicating moderate-to-high correlation in corneal HOAs between both eyes . However, total HOA and oblique trefoil (Z3-3) did not show a significant interocular relationship, suggesting variability in individual wavefront aberration profiles. HOAs significantly influence retinal image quality and visual perception . Prior studies suggest that nuclear cataracts increase coma and spherical aberration, while posterior subcapsular cataracts primarily affect trefoil aberrations . Similar to previous research , we did not stratify HOA analysis by cataract type, which may explain discrepancies in interocular correlation findings. Kim et al. reported that nuclear cataract-induced HOAs could cause retinal triple imaging, contributing to visual disturbances. Wang et al. found that anterior corneal surface HOAs exhibit mirror symmetry between eyes, which may help preserve binocular vision despite corneal irregularities. Tear film stability also plays a role in HOA fluctuations , as an uneven tear film can increase wavefront distortions. Given the lack of effective surgical methods for reducing HOAs , preoperative assessment remains essential for optimizing refractive outcomes. Our study revealed that corneal curvature, as measured by SimK and TCRP, were observed to be lower in male patients compared to female patients. These findings align with previous research, indicating that female corneas tend to be steeper than male corneas ( P < 0.01). Similar to other physiological differences, sex-related variations in ocular structure may contribute to distinct refractive outcomes . Although age broadly affects multiple ocular parameters, sex-related differences are often more subtle and inconsistent across studies. Corneal curvature remains one of the most reliably documented sex-dependent parameters, whereas other biometric variations (e.g., axial length, lens thickness) tend to show smaller or less consistent differences. Notably, males typically exhibit longer axial lengths and flatter corneal curvatures compared to females . These findings highlight the need to consider sex differences in preoperative surgical planning. Traditional IOL power calculation formulas do not account for sex-specific variations, yet even minor corneal modifications can lead to substantial changes in refractive power. Implementing sex-adjusted methodologies could enhance postoperative visual acuity predictions and improve surgical outcomes. While this study provides valuable insights, certain limitations should be acknowledged. As a retrospective study, potential selection bias cannot be entirely excluded. However, since patient selection, surgical treatment, and postoperative follow-up were not influenced by artificial criteria, the impact of selection bias is likely minimal. Additionally, since our research is based on hospital-based clinical data, the findings may not be fully representative of the broader Chinese population. A larger-scale, community-based epidemiologic survey would be needed to validate these results. Another limitation is the lack of stratification based on cataract type. Variations in nuclear, cortical, and posterior subcapsular cataracts may influence certain corneal parameters, such as HOAs and spherical aberration. Future studies could explore these distinctions to refine surgical approaches further. Despite these limitations, our study presents novel findings on the distribution of preoperative corneal morphological parameters in Chinese patients with cataract. This is the first large-scale analysis using the Pentacam instrument to assess ocular biometric variations by age, sex, and interocular correlation. Our results offer valuable insights for cataract surgery planning and may help refine patient-specific surgical strategies. We observed significant age-related changes in corneal curvature, spherical aberration, and astigmatism, as well as notable sex differences in corneal parameters. Additionally, interocular correlations in HOAs suggest that binocular vision may compensate for asymmetrical corneal aberrations. These findings emphasize the need for personalized surgical planning to optimize postoperative visual outcomes in cataract patients.
Participants The study was conducted following the tenets of the Declaration of Helsinki. The study involved human participants and was approved by the Institutional Review Board of Shenzhen Eye Hospital (2023KYPJ065). Written informed consent for participation was not required for this study in accordance with the national legislation and institutional requirements. In this retrospective case series study, clinical data of patients with cataract across various age groups who attended our hospital between June 2020 and March 2023 were collected. The right eyes of patients who met the inclusion criteria were chosen for observation and analysis. The left eyes of the same patients were included in the study for comparison and correlation analyses involving both eyes. A total of 1,255 right eye cases were included, along with 873 cases involving both eyes; accordingly, we included 1,255 patients (2,128 eyes). The reason for the difference in the number of cases (873 vs. 1,255) is that not all patients had complete bilateral data available for analysis. Some patients had only one eye meeting the inclusion criteria, or data for the contralateral eye were incomplete or unavailable due to technical or clinical reasons (e.g., poor quality scans, prior surgery, or other ocular conditions in the contralateral eye). Additionally, the study sample comprised 578 male eyes and 677 female eyes. The patients were categorized into four groups based on age: Group A (20–40 years, 427 cases), Group B (41–60 years, 315 cases), Group C (61–80 years, 373 cases), and Group D (>81 years, 140 cases). The inclusion criteria were as follows: (1) satisfactory compliance, an ability to cooperate until completion of the Pentacam examination; (2) absence of other eye diseases; (3) absence of prior eye surgery; and (4) cases wherein the Pentacam for anterior segment analysis passed the quality status assessment. The exclusion criteria were as follows: (1) conditions such as glaucoma, dry eye syndrome, and high myopia; (2) patients with a history of myopic laser surgery; (3) contact lens users; (4) patients who underwent keratoplasty; and (5) patients with various corneal diseases. Measurement method Preoperative optical biometry was performed using the Pentacam AXL software (version 1.21r43) in all cases. The Pentacam AXL is based on Scheimpflug imaging technology and ray tracing principles, which allow for a comprehensive analysis of the anterior segment of the eye. The device uses a red LED light to automatically locate the corneal apex, followed by a blue slit light source (wavelength 475 nm) to illuminate the cornea. The camera, tilted at 45°, rotates 360° to capture a series of oblique cross-sectional images of the cornea. These images are then reconstructed to generate a three-dimensional model of the anterior segment, providing detailed information about corneal curvature, thickness, and other relevant parameters. Simulated Keratometry (SimK) SimK refers to the refractive power (or curvature) of the anterior corneal surface. It is calculated based on the curvature of the anterior cornea, typically measured along the flat (k1) and steep (k2) meridians. The mean corneal curvature (km) is derived from these values. SimK assumes a fixed ratio between the anterior and posterior corneal curvatures, which is a standard approach in clinical practice for estimating corneal refractive power . Total corneal refractive power (TCRP) TCRP represents the total refractive power of the cornea, taking into account both the anterior and posterior corneal surfaces. Unlike SimK, TCRP is calculated using ray tracing technology, which considers the actual path of light through the cornea. This method incorporates data on corneal thickness and the curvature of both the anterior and posterior corneal surfaces, providing a more accurate measurement of the cornea’s total refractive power . The Pentacam AXL automatically measures both SimK and TCRP, along with other relevant parameters, such as corneal astigmatism, higher-order aberrations, and corneal thickness. These measurements are crucial for preoperative planning in cataract surgery, particularly for intraocular lens (IOL) power calculation and the selection of appropriate surgical techniques. The examination was performed on the pupil under dark room conditions. During pre-examination, patients were instructed to blink to maintain a uniform distribution of the tear film to minimize error. The patients were advised to keep both eyes wide open and fixate on a flashing red light. Upon focusing, the machine automatically reconstructed the three-dimensional structure of the anterior segment using 25 frames of Scheimpflug images. The recorded measurements included the flat axis curvature (k1), steep axis curvature (k2), and mean corneal curvature (km) of the anterior corneal surface and the whole cornea of the examined eye. Corneal astigmatism and its axis were extracted from the anterior surface of the cornea. The type of corneal astigmatism was determined based on the orientation of the steep axis: with-the-rule astigmatism 90 ± 30°, against-the-rule (ATR) astigmatism 0 ± 30°, and oblique astigmatism for orientations in between. Higher-order aberrations within a 6-mm diameter of the anterior corneal surface were recorded, including total higher-order aberrations (HOA), horizontal coma (Z31), vertical coma (Z3-1), horizontal trefoil (Z33), vertical trefoil (Z3-3), and spherical aberration (Z40). All examinations were conducted by the same professionally trained technician. Stratification by gender In addition to stratifying by age, we also conducted analyses based on gender to examine potential differences in corneal parameters between male and female patients. Comparisons were made between the groups using the appropriate non-parametric tests, as described for the age comparisons. The results of these comparisons are summarized in Table , where “Test statistic 1 ” represents the age based startification, whlie “Test statistic 2 ” represents gender-based stratification. Statistical analysis All statistical analyses were performed using the SAS 9.4 software. The Kolmogorov–Smirnov test, combined with graphical methods, was used to test the normality of data distribution. The normality test confirmed that the higher-order aberration (HOA) data were normally distributed ( p > 0.05 for all variables). Given this, analysis of variance (ANOVA) was applied to compare HOA values across different age groups. Spearman’s correlation analysis was used to analyze the correlation between two continuous variables. The significance level was set at α = 0.05.
The study was conducted following the tenets of the Declaration of Helsinki. The study involved human participants and was approved by the Institutional Review Board of Shenzhen Eye Hospital (2023KYPJ065). Written informed consent for participation was not required for this study in accordance with the national legislation and institutional requirements. In this retrospective case series study, clinical data of patients with cataract across various age groups who attended our hospital between June 2020 and March 2023 were collected. The right eyes of patients who met the inclusion criteria were chosen for observation and analysis. The left eyes of the same patients were included in the study for comparison and correlation analyses involving both eyes. A total of 1,255 right eye cases were included, along with 873 cases involving both eyes; accordingly, we included 1,255 patients (2,128 eyes). The reason for the difference in the number of cases (873 vs. 1,255) is that not all patients had complete bilateral data available for analysis. Some patients had only one eye meeting the inclusion criteria, or data for the contralateral eye were incomplete or unavailable due to technical or clinical reasons (e.g., poor quality scans, prior surgery, or other ocular conditions in the contralateral eye). Additionally, the study sample comprised 578 male eyes and 677 female eyes. The patients were categorized into four groups based on age: Group A (20–40 years, 427 cases), Group B (41–60 years, 315 cases), Group C (61–80 years, 373 cases), and Group D (>81 years, 140 cases). The inclusion criteria were as follows: (1) satisfactory compliance, an ability to cooperate until completion of the Pentacam examination; (2) absence of other eye diseases; (3) absence of prior eye surgery; and (4) cases wherein the Pentacam for anterior segment analysis passed the quality status assessment. The exclusion criteria were as follows: (1) conditions such as glaucoma, dry eye syndrome, and high myopia; (2) patients with a history of myopic laser surgery; (3) contact lens users; (4) patients who underwent keratoplasty; and (5) patients with various corneal diseases.
Preoperative optical biometry was performed using the Pentacam AXL software (version 1.21r43) in all cases. The Pentacam AXL is based on Scheimpflug imaging technology and ray tracing principles, which allow for a comprehensive analysis of the anterior segment of the eye. The device uses a red LED light to automatically locate the corneal apex, followed by a blue slit light source (wavelength 475 nm) to illuminate the cornea. The camera, tilted at 45°, rotates 360° to capture a series of oblique cross-sectional images of the cornea. These images are then reconstructed to generate a three-dimensional model of the anterior segment, providing detailed information about corneal curvature, thickness, and other relevant parameters. Simulated Keratometry (SimK) SimK refers to the refractive power (or curvature) of the anterior corneal surface. It is calculated based on the curvature of the anterior cornea, typically measured along the flat (k1) and steep (k2) meridians. The mean corneal curvature (km) is derived from these values. SimK assumes a fixed ratio between the anterior and posterior corneal curvatures, which is a standard approach in clinical practice for estimating corneal refractive power . Total corneal refractive power (TCRP) TCRP represents the total refractive power of the cornea, taking into account both the anterior and posterior corneal surfaces. Unlike SimK, TCRP is calculated using ray tracing technology, which considers the actual path of light through the cornea. This method incorporates data on corneal thickness and the curvature of both the anterior and posterior corneal surfaces, providing a more accurate measurement of the cornea’s total refractive power . The Pentacam AXL automatically measures both SimK and TCRP, along with other relevant parameters, such as corneal astigmatism, higher-order aberrations, and corneal thickness. These measurements are crucial for preoperative planning in cataract surgery, particularly for intraocular lens (IOL) power calculation and the selection of appropriate surgical techniques. The examination was performed on the pupil under dark room conditions. During pre-examination, patients were instructed to blink to maintain a uniform distribution of the tear film to minimize error. The patients were advised to keep both eyes wide open and fixate on a flashing red light. Upon focusing, the machine automatically reconstructed the three-dimensional structure of the anterior segment using 25 frames of Scheimpflug images. The recorded measurements included the flat axis curvature (k1), steep axis curvature (k2), and mean corneal curvature (km) of the anterior corneal surface and the whole cornea of the examined eye. Corneal astigmatism and its axis were extracted from the anterior surface of the cornea. The type of corneal astigmatism was determined based on the orientation of the steep axis: with-the-rule astigmatism 90 ± 30°, against-the-rule (ATR) astigmatism 0 ± 30°, and oblique astigmatism for orientations in between. Higher-order aberrations within a 6-mm diameter of the anterior corneal surface were recorded, including total higher-order aberrations (HOA), horizontal coma (Z31), vertical coma (Z3-1), horizontal trefoil (Z33), vertical trefoil (Z3-3), and spherical aberration (Z40). All examinations were conducted by the same professionally trained technician.
SimK refers to the refractive power (or curvature) of the anterior corneal surface. It is calculated based on the curvature of the anterior cornea, typically measured along the flat (k1) and steep (k2) meridians. The mean corneal curvature (km) is derived from these values. SimK assumes a fixed ratio between the anterior and posterior corneal curvatures, which is a standard approach in clinical practice for estimating corneal refractive power .
TCRP represents the total refractive power of the cornea, taking into account both the anterior and posterior corneal surfaces. Unlike SimK, TCRP is calculated using ray tracing technology, which considers the actual path of light through the cornea. This method incorporates data on corneal thickness and the curvature of both the anterior and posterior corneal surfaces, providing a more accurate measurement of the cornea’s total refractive power . The Pentacam AXL automatically measures both SimK and TCRP, along with other relevant parameters, such as corneal astigmatism, higher-order aberrations, and corneal thickness. These measurements are crucial for preoperative planning in cataract surgery, particularly for intraocular lens (IOL) power calculation and the selection of appropriate surgical techniques. The examination was performed on the pupil under dark room conditions. During pre-examination, patients were instructed to blink to maintain a uniform distribution of the tear film to minimize error. The patients were advised to keep both eyes wide open and fixate on a flashing red light. Upon focusing, the machine automatically reconstructed the three-dimensional structure of the anterior segment using 25 frames of Scheimpflug images. The recorded measurements included the flat axis curvature (k1), steep axis curvature (k2), and mean corneal curvature (km) of the anterior corneal surface and the whole cornea of the examined eye. Corneal astigmatism and its axis were extracted from the anterior surface of the cornea. The type of corneal astigmatism was determined based on the orientation of the steep axis: with-the-rule astigmatism 90 ± 30°, against-the-rule (ATR) astigmatism 0 ± 30°, and oblique astigmatism for orientations in between. Higher-order aberrations within a 6-mm diameter of the anterior corneal surface were recorded, including total higher-order aberrations (HOA), horizontal coma (Z31), vertical coma (Z3-1), horizontal trefoil (Z33), vertical trefoil (Z3-3), and spherical aberration (Z40). All examinations were conducted by the same professionally trained technician.
In addition to stratifying by age, we also conducted analyses based on gender to examine potential differences in corneal parameters between male and female patients. Comparisons were made between the groups using the appropriate non-parametric tests, as described for the age comparisons. The results of these comparisons are summarized in Table , where “Test statistic 1 ” represents the age based startification, whlie “Test statistic 2 ” represents gender-based stratification.
All statistical analyses were performed using the SAS 9.4 software. The Kolmogorov–Smirnov test, combined with graphical methods, was used to test the normality of data distribution. The normality test confirmed that the higher-order aberration (HOA) data were normally distributed ( p > 0.05 for all variables). Given this, analysis of variance (ANOVA) was applied to compare HOA values across different age groups. Spearman’s correlation analysis was used to analyze the correlation between two continuous variables. The significance level was set at α = 0.05.
|
Impact of family practice continuity of care on unplanned hospital use for people with serious mental illness | 644254ff-c02d-4b32-bf90-ceb8399dda46 | 6863233 | Family Medicine[mh] | INTRODUCTION Serious mental illness (SMI) includes schizophrenia, schizoaffective disorder, bipolar disorder, and other psychoses. People with SMI have high rates of comorbidity, reduced quality of life, shortened life expectancy, , and high rates of emergency department (ED) presentations and unplanned hospital admissions. , , Finding ways to improve health care and outcomes for this group is therefore a high priority. Continuity of care is widely held to be beneficial for people with long‐term conditions, including SMI. It is valued by patients , and providers and considered good practice in mental health and family medicine, , , reducing fragmentation of care and facilitating better provider‐patient relationships. Relational continuity—the longitudinal relationship between a patient and a health care practitioner (or group of practitioners) —is often the focus of efforts to improve continuity. To date, evidence has been mixed on whether relational continuity improves outcomes for people with SMI. Some studies have found that higher continuity is associated with lower mortality, reduced hospital admissions, and improved recovery from episodes of SMI, while others have found no association or even the reverse. , , Studies that have examined the relationship of continuity to costs have mostly found that higher continuity was associated with lower health care costs, although one showed an association with higher costs of community care , , It is important to clarify whether relational continuity is beneficial, since achieving higher continuity may increase costs and require trade‐offs with other elements of good care, such as flexibility to meet urgent care needs. Studies of relational continuity for people with SMI have most often considered visits within specialist mental health services, or across multiple types of service (which we term “across‐practice continuity”). However, in the UK family physicians provide much of the physical and mental health care for people with SMI and around a third of people with SMI are treated solely by their family physician. Policies such as named accountable practitioners have emphasized the importance of maintaining continuity with an individual family physician, not just a practice. The UK's National Health Service (NHS) provides publicly funded health care which requires patients to register with a specific family practice, so that patients face barriers to changing practices or attending different practices concurrently. In other health care systems, the role of family physicians in the care of people with SMI may be less prominent, and patients may be more likely to see physicians at different family practices, but initiatives such as the patient‐centered medical home in the United States have a similar focus on relational continuity with family physicians. Evidence is therefore needed on the impact of within‐practice family physician continuity on the physical and mental health of people with SMI, in addition to the existing literature on across‐practice continuity focused on specialist mental health care. Continuity of care has other aspects beyond relational continuity, including informational and management continuity. In the United Kingdom, people usually register with a family practice and within that practice have a nominated physician who acts as a gatekeeper to and liaison with other health care services, including specialist mental health services. However, individuals can see any physician in that practice, especially for urgent appointments. Care plans for people with SMI document the patient's care needs, patterns of relapse, preferences for treatment, and social context and are stored with patient records and accessible by different practitioners seeing the patient. Care plans therefore promote informational continuity across family physicians in the same practice and may also promote management continuity, if the management approach is agreed and can be followed by all practitioners. A previous study showed that care plans for people with SMI were associated with a lower risk of unplanned hospital use, but that study did not account for relational continuity. Relational continuity is known to vary with observed individual characteristics such as age and sex, , , but continuity may also be influenced by factors that are usually unobserved, such as help‐seeking attitudes, disease severity, personality, or social context. If these unobserved factors also influence outcomes, the observed association between continuity and outcomes may be biased. For example, people who are more proactive in seeking care may receive higher continuity, but they may also have better outcomes because they seek care early or engage in preventive management. Conversely, family physicians may prioritize continuity for people with more severe illness, who nonetheless may have a higher risk of deterioration than those with less severe illness. To our knowledge, only one study has attempted to address unobserved confounding when examining the relationship between continuity of care and outcomes. It looked at the effect of relational continuity on emergency department attendance for people with diabetes and hypertension in Taiwan and measured continuity in 1 year and outcomes in the next. It employed an instrumental variable approach to account for confounding, with the relational continuity of family members of the patient as instrument. The results showed a stronger negative association between continuity and ED presentations with the instrumental variable approach than the standard approach. We examined whether family physician relational continuity for people with SMI is associated with better outcomes, using the novel application of methods to account for time‐invariant unobserved confounding. The study objective was to investigate the hypothesis that continuity of care in family practice reduces unplanned hospital utilization.
METHODS 2.1 Study design This observational cohort study used individual‐level family practice administrative data linked to hospital administrative data to investigate the relationship between family practice continuity of care for people with SMI and time to unplanned hospital use. 2.2 Sample We used data from the Clinical Practice Research Datalink (CPRD), a database of anonymized patient records derived from over 600 family practices in England and broadly representative of the national population with respect to age and gender. The records were linked to Hospital Episode Statistics (HES), which capture all hospital admissions (for both physical and mental health) and ED presentations funded by the NHS. This covers the majority of these types of health care in England, since the NHS funds 88 percent of all health care expenditure and 92 percent of hospital care, and there are no privately funded emergency departments. The sample was all people with a diagnosis of SMI documented in primary care on or before March 31, 2014 (the end of the study period), whose records met CPRD quality standards, and who were registered during this period at a participating practice that met CPRD standards. Diagnoses of SMI were based on clinical information in routine practice data recorded in Read codes, an hierarchical coding system for clinical data that classifies diseases, patient characteristics, tests, and procedures , (see Table for a list of the Read codes used in this study). The start date of observation for each individual was the latest of: date of SMI diagnosis, date of registration at the practice plus 1 year of observation in primary care records, January 1 of the calendar year after the person turned 18, and April 1, 2007 (because data on ED presentations were only available from this date). The year of observation in primary care records allowed for observation of baseline characteristics as control variables. Additionally, the start date of observation for each individual was moved later if necessary so that no patients had an ED presentation or a hospital admission for at least 1 year prior to the start date, since hospital care could influence the level of continuity in primary care. The observation period for each individual was divided into periods of 3 months dating from their first date of observation, with continuity measured in the prior 12 months. Individuals were followed until outcome or censoring, where censoring is due to the person changing family practice, death, or the end of the study period (March 31, 2014). 2.3 Outcome measures We constructed three measures of unplanned hospital use from HES: (a) ED presentations, (b) unplanned admissions for SMI, and (c) unplanned admissions for ambulatory care‐sensitive conditions (ACSC), which are conditions thought to be particularly amenable to ambulatory care (such as diabetes, angina, cellulitis, and vaccine‐preventable diseases, but not SMI). Hospital admissions were classified using International Classification of Disease (ICD‐10) codes to identify SMI and ACSC admissions. (The codes used to classify ACSC admissions are listed in Table .) All ED presentations were included. For each type of outcome, we considered only the first observed instance (presentation or admission), since this could have influenced subsequent continuity. The occurrence of the outcome is measured in the 3‐month period t and continuity is measured over a lookback period of the prior 12 months (4 × 3 month periods t −4 to t −1). That is, there is no overlap between the 12‐month period in which continuity is measured and the subsequent 3‐month period in which outcomes are observed. The outcome variable is a binary variable for each 3‐month period indicating whether or not the event occurred in that period. For any individual who did not experience the outcome of interest (eg someone who did not present to ED during the period of observation), this variable is equal to zero for all periods. As we only analyzed time to first event, for any individual who did experience the outcome, the variable is equal to zero for all periods except the final period and equal to one for the final period, with all periods after the first event excluded from the analysis for that outcome. 2.4 Measures of relational continuity We used three indices measuring different dimensions of family physician relational continuity. The Continuity of Care (COC) index measures dispersion of visits across family physicians within the patient's registered family practice, by capturing how many different practitioners are involved and how many visits occur to each. The Usual Provider of Care (UPC) index measures density of visits, being the proportion of a patient's visits that are with the family physician most frequently seen by the individual in that year out of the total number of visits at the practice. The Sequential Continuity (SECON) index measures the pattern of visits across different practitioners, using the proportion of consecutive pairs of visits which are to the same family physician out of the total number of consecutive pairs of visits at the practice. Each index ranges from zero (lowest continuity) to 1 (perfect continuity). Additional detail on each index is in the Tables , and illustrative examples are shown in Table . We measured continuity over 12 months (4 × 3 month periods), considering only face‐to‐face visits with family physicians. There is no standard level for “high” and “low” continuity, so we applied one recognized method that classified relational continuity as “high” if the level of continuity was above the median for the index, and “low” if at or below the median level. , A minimum of two visits is required to calculate COC and SECON, but to improve index stability we set the minimum to three visits. Periods with fewer than two visits in the prior 12‐month lookback period were included in the analysis with continuity categorized as “undefined.” We constructed a set of categorical variables based on visit frequency and whether continuity was low or high. This allowed for different effects of continuity for frequent and less frequent users of family practice, as suggested by previous research. Visit frequency was classified into low, moderate, and high: low (0‐2 visits), moderate (3‐5 visits), and high (6 or more visits). These categories correspond to tertiles of the full‐visit distribution: two visits is the 33rd percentile and five visits is the 66th percentile. Continuity indices were defined as low or high based on the median value of each index: COC low (0‐0.35), high (>0.35); UPC low (0‐0.67), high (>0.67); SECON low (0‐0.17), high (>0.17). Periods were then classified into five categories according to continuity level and visit frequency in the prior 12 months: low visit frequency (with continuity undefined—the base category), moderate visit frequency with low continuity, moderate visit frequency with high continuity, high visit frequency with low continuity, and high visit frequency with high continuity. 2.5 Measure of informational/management continuity This analysis captures management/ informational continuity separately from relational continuity according to whether the individual had a care plan documented by a family physician in the prior 12 months. Because we focus on within‐practice family physician continuity, we distinguish relational continuity from management and informational continuity represented by care plans. Doctors within a practice have access to the same medical records and may have similar approaches to management. 2.6 Control variables Individual characteristics measured at baseline were as follows: age, gender, ethnicity, deprivation of the person's neighborhood of residence, history of smoking, number of Charlson Index comorbidities, comorbid depression, diagnostic subgroup (schizophrenia and other psychoses, or bipolar disorder and affective psychoses) and number of years since diagnosis. Treatment for SMI was included as a time‐varying variable indicating that the individual had been prescribed an antipsychotic drug at least once in the 12‐month lookback period prior to the current period. 2.7 Statistical analysis The necessity of creating periods for continuity measurement led us to employ discrete‐time survival analysis. Although the outcomes of interest are (effectively) continuous measures (since we have day‐level data on when these occur), these are converted into discrete outcomes for each period in order to match the measurement of continuity. The model evaluates the association between continuity in the prior 12 months and risk of the outcome in a particular 3‐month period. A complementary log‐log (cloglog) proportional hazards model was fitted for each outcome. This model produces hazard ratios that are the discrete‐time equivalent of the Cox proportional hazards model used in a continuous‐time context. A flexible piece‐wise constant baseline hazard function was applied by specifying dummy variables for each 3‐month period. This assumes that the hazard function is constant within each period, but can vary across periods. The resulting exponentiated coefficients can be interpreted as hazard ratios, the discrete‐time counterpart of the hazard from a continuous‐time proportional hazards model. The hazard ratio is the proportional change in the underlying hazard of the outcome for a unit change in the variable. The hazard rate (HR) at period t is the probability of observing the outcome for an individual in period t , conditional on the individual “surviving” in the sample to period t (ie, no censoring and the outcome was not observed in prior periods for that individual). The HR is a nonlinear function of time‐varying factors, time‐invariant factors, time‐period dummy variables representative of the baseline hazard, and normally distributed individual unobserved heterogeneity. Our main modeling approach accounts for individual unobserved heterogeneity. Due to the incidental parameter problem of specifying individual fixed effects to represent such heterogeneity in nonlinear models, we instead assume unobserved heterogeneity is normally distributed and specify this as a linear function of the means of time‐varying variables. This is often termed a correlated random‐effects model, following Mundlak. The time‐varying variables were the care quality indicators plus the time‐varying covariate for antipsychotic treatment, while the remaining individual characteristics included as covariates were time‐invariant, captured at baseline. The variables representing the means of the time‐varying variables effectively capture confounding by unobserved time‐invariant individual factors (eg long‐standing illness, health‐seeking behavior) that drive both continuity and use of hospital services. The period‐specific levels of the time‐varying variables capture deviation from this long‐term average and can be interpreted as the effect of continuity specific to that 3‐month period, given the person's overall propensity to receive continuity of care. (See Tables for more detail of the model.) To allow comparison of our results to previous studies examining the effect of continuity of care, we also estimated models that did not specify individual heterogeneity as a function of the means of the time‐varying variables, the random‐effects model. These models allow for normally distributed individual heterogeneity but it is assumed to be uncorrelated with the explanatory variables contained in the model. All models included observed individual characteristics as explanatory variables and adjusted standard errors for clustering at the practice level. We estimated separate models for each of the three continuity indices because of multicollinearity of the indices. All analyses were conducted using Stata v14. 2.8 Robustness checks We tested the sensitivity of the results to the level of visit frequency at which continuity was classified as “undefined”. The minimum level for measuring continuity (and corresponding categories for low vs moderate visit frequency) was set to two or four visits rather than three visits as in our main analysis. Given the significant physical health comorbidities of people with SMI, we examined an alternative to the two separate hospital admission outcomes: all‐cause hospital admissions, capturing all unplanned admissions for both physical and mental health conditions. To investigate whether receiving specialist mental health care confounded the relationship between primary care quality and outcomes, we ran additional analyses capturing care in specialist mental health services. Data on specialist care from the Mental Health Services Minimum Dataset (MHMDS) were only available to link to the main dataset from April 1, 2011. We added a time‐varying variable to indicate whether the individual received any care in specialist mental health services in the prior 12 months and ran the analysis over the 3‐year period of observation to March 31, 2014.
Study design This observational cohort study used individual‐level family practice administrative data linked to hospital administrative data to investigate the relationship between family practice continuity of care for people with SMI and time to unplanned hospital use.
Sample We used data from the Clinical Practice Research Datalink (CPRD), a database of anonymized patient records derived from over 600 family practices in England and broadly representative of the national population with respect to age and gender. The records were linked to Hospital Episode Statistics (HES), which capture all hospital admissions (for both physical and mental health) and ED presentations funded by the NHS. This covers the majority of these types of health care in England, since the NHS funds 88 percent of all health care expenditure and 92 percent of hospital care, and there are no privately funded emergency departments. The sample was all people with a diagnosis of SMI documented in primary care on or before March 31, 2014 (the end of the study period), whose records met CPRD quality standards, and who were registered during this period at a participating practice that met CPRD standards. Diagnoses of SMI were based on clinical information in routine practice data recorded in Read codes, an hierarchical coding system for clinical data that classifies diseases, patient characteristics, tests, and procedures , (see Table for a list of the Read codes used in this study). The start date of observation for each individual was the latest of: date of SMI diagnosis, date of registration at the practice plus 1 year of observation in primary care records, January 1 of the calendar year after the person turned 18, and April 1, 2007 (because data on ED presentations were only available from this date). The year of observation in primary care records allowed for observation of baseline characteristics as control variables. Additionally, the start date of observation for each individual was moved later if necessary so that no patients had an ED presentation or a hospital admission for at least 1 year prior to the start date, since hospital care could influence the level of continuity in primary care. The observation period for each individual was divided into periods of 3 months dating from their first date of observation, with continuity measured in the prior 12 months. Individuals were followed until outcome or censoring, where censoring is due to the person changing family practice, death, or the end of the study period (March 31, 2014).
Outcome measures We constructed three measures of unplanned hospital use from HES: (a) ED presentations, (b) unplanned admissions for SMI, and (c) unplanned admissions for ambulatory care‐sensitive conditions (ACSC), which are conditions thought to be particularly amenable to ambulatory care (such as diabetes, angina, cellulitis, and vaccine‐preventable diseases, but not SMI). Hospital admissions were classified using International Classification of Disease (ICD‐10) codes to identify SMI and ACSC admissions. (The codes used to classify ACSC admissions are listed in Table .) All ED presentations were included. For each type of outcome, we considered only the first observed instance (presentation or admission), since this could have influenced subsequent continuity. The occurrence of the outcome is measured in the 3‐month period t and continuity is measured over a lookback period of the prior 12 months (4 × 3 month periods t −4 to t −1). That is, there is no overlap between the 12‐month period in which continuity is measured and the subsequent 3‐month period in which outcomes are observed. The outcome variable is a binary variable for each 3‐month period indicating whether or not the event occurred in that period. For any individual who did not experience the outcome of interest (eg someone who did not present to ED during the period of observation), this variable is equal to zero for all periods. As we only analyzed time to first event, for any individual who did experience the outcome, the variable is equal to zero for all periods except the final period and equal to one for the final period, with all periods after the first event excluded from the analysis for that outcome.
Measures of relational continuity We used three indices measuring different dimensions of family physician relational continuity. The Continuity of Care (COC) index measures dispersion of visits across family physicians within the patient's registered family practice, by capturing how many different practitioners are involved and how many visits occur to each. The Usual Provider of Care (UPC) index measures density of visits, being the proportion of a patient's visits that are with the family physician most frequently seen by the individual in that year out of the total number of visits at the practice. The Sequential Continuity (SECON) index measures the pattern of visits across different practitioners, using the proportion of consecutive pairs of visits which are to the same family physician out of the total number of consecutive pairs of visits at the practice. Each index ranges from zero (lowest continuity) to 1 (perfect continuity). Additional detail on each index is in the Tables , and illustrative examples are shown in Table . We measured continuity over 12 months (4 × 3 month periods), considering only face‐to‐face visits with family physicians. There is no standard level for “high” and “low” continuity, so we applied one recognized method that classified relational continuity as “high” if the level of continuity was above the median for the index, and “low” if at or below the median level. , A minimum of two visits is required to calculate COC and SECON, but to improve index stability we set the minimum to three visits. Periods with fewer than two visits in the prior 12‐month lookback period were included in the analysis with continuity categorized as “undefined.” We constructed a set of categorical variables based on visit frequency and whether continuity was low or high. This allowed for different effects of continuity for frequent and less frequent users of family practice, as suggested by previous research. Visit frequency was classified into low, moderate, and high: low (0‐2 visits), moderate (3‐5 visits), and high (6 or more visits). These categories correspond to tertiles of the full‐visit distribution: two visits is the 33rd percentile and five visits is the 66th percentile. Continuity indices were defined as low or high based on the median value of each index: COC low (0‐0.35), high (>0.35); UPC low (0‐0.67), high (>0.67); SECON low (0‐0.17), high (>0.17). Periods were then classified into five categories according to continuity level and visit frequency in the prior 12 months: low visit frequency (with continuity undefined—the base category), moderate visit frequency with low continuity, moderate visit frequency with high continuity, high visit frequency with low continuity, and high visit frequency with high continuity.
Measure of informational/management continuity This analysis captures management/ informational continuity separately from relational continuity according to whether the individual had a care plan documented by a family physician in the prior 12 months. Because we focus on within‐practice family physician continuity, we distinguish relational continuity from management and informational continuity represented by care plans. Doctors within a practice have access to the same medical records and may have similar approaches to management.
Control variables Individual characteristics measured at baseline were as follows: age, gender, ethnicity, deprivation of the person's neighborhood of residence, history of smoking, number of Charlson Index comorbidities, comorbid depression, diagnostic subgroup (schizophrenia and other psychoses, or bipolar disorder and affective psychoses) and number of years since diagnosis. Treatment for SMI was included as a time‐varying variable indicating that the individual had been prescribed an antipsychotic drug at least once in the 12‐month lookback period prior to the current period.
Statistical analysis The necessity of creating periods for continuity measurement led us to employ discrete‐time survival analysis. Although the outcomes of interest are (effectively) continuous measures (since we have day‐level data on when these occur), these are converted into discrete outcomes for each period in order to match the measurement of continuity. The model evaluates the association between continuity in the prior 12 months and risk of the outcome in a particular 3‐month period. A complementary log‐log (cloglog) proportional hazards model was fitted for each outcome. This model produces hazard ratios that are the discrete‐time equivalent of the Cox proportional hazards model used in a continuous‐time context. A flexible piece‐wise constant baseline hazard function was applied by specifying dummy variables for each 3‐month period. This assumes that the hazard function is constant within each period, but can vary across periods. The resulting exponentiated coefficients can be interpreted as hazard ratios, the discrete‐time counterpart of the hazard from a continuous‐time proportional hazards model. The hazard ratio is the proportional change in the underlying hazard of the outcome for a unit change in the variable. The hazard rate (HR) at period t is the probability of observing the outcome for an individual in period t , conditional on the individual “surviving” in the sample to period t (ie, no censoring and the outcome was not observed in prior periods for that individual). The HR is a nonlinear function of time‐varying factors, time‐invariant factors, time‐period dummy variables representative of the baseline hazard, and normally distributed individual unobserved heterogeneity. Our main modeling approach accounts for individual unobserved heterogeneity. Due to the incidental parameter problem of specifying individual fixed effects to represent such heterogeneity in nonlinear models, we instead assume unobserved heterogeneity is normally distributed and specify this as a linear function of the means of time‐varying variables. This is often termed a correlated random‐effects model, following Mundlak. The time‐varying variables were the care quality indicators plus the time‐varying covariate for antipsychotic treatment, while the remaining individual characteristics included as covariates were time‐invariant, captured at baseline. The variables representing the means of the time‐varying variables effectively capture confounding by unobserved time‐invariant individual factors (eg long‐standing illness, health‐seeking behavior) that drive both continuity and use of hospital services. The period‐specific levels of the time‐varying variables capture deviation from this long‐term average and can be interpreted as the effect of continuity specific to that 3‐month period, given the person's overall propensity to receive continuity of care. (See Tables for more detail of the model.) To allow comparison of our results to previous studies examining the effect of continuity of care, we also estimated models that did not specify individual heterogeneity as a function of the means of the time‐varying variables, the random‐effects model. These models allow for normally distributed individual heterogeneity but it is assumed to be uncorrelated with the explanatory variables contained in the model. All models included observed individual characteristics as explanatory variables and adjusted standard errors for clustering at the practice level. We estimated separate models for each of the three continuity indices because of multicollinearity of the indices. All analyses were conducted using Stata v14.
Robustness checks We tested the sensitivity of the results to the level of visit frequency at which continuity was classified as “undefined”. The minimum level for measuring continuity (and corresponding categories for low vs moderate visit frequency) was set to two or four visits rather than three visits as in our main analysis. Given the significant physical health comorbidities of people with SMI, we examined an alternative to the two separate hospital admission outcomes: all‐cause hospital admissions, capturing all unplanned admissions for both physical and mental health conditions. To investigate whether receiving specialist mental health care confounded the relationship between primary care quality and outcomes, we ran additional analyses capturing care in specialist mental health services. Data on specialist care from the Mental Health Services Minimum Dataset (MHMDS) were only available to link to the main dataset from April 1, 2011. We added a time‐varying variable to indicate whether the individual received any care in specialist mental health services in the prior 12 months and ran the analysis over the 3‐year period of observation to March 31, 2014.
RESULTS 3.1 Sample The sample consisted of 19 324 individuals attending 215 practices, observed for 15.8 3‐month periods on average (range 1‐28 periods). Table presents the characteristics of individuals in the sample. Half of the sample (50.3 percent) had an ED presentation at some point during the observation period, 13.1 percent had an admission for SMI, and 12.8 percent had an ACSC admission. Using a three‐visit minimum to define continuity, median (mean) values for each continuity index were as follows: COC 0.35 (0.46), UPC 0.67 (0.65), and SECON 0.17 (0.26). A care plan had been documented in the previous 12 months for 40 percent of the periods observed. The Spearman rank correlation between COC and UPC indices was 0.94 ( P < .001), between COC and SECON was 0.55 ( P < .001), and between UPC and SECON was 0.47 ( P < .001). Mean COC in periods with a care plan in the previous 12 months was 0.47, compared with 0.45 in periods without a care plan in the previous 12 months (two‐sample t test of difference in means: P < .001); the equivalent for UPC was 0.67 vs 0.66 ( P < .001) and for SECON was 0.24 vs 0.23 ( P < .001). 3.2 Association between continuity of care and unplanned hospital use Table presents the association between continuity of care and each outcome from the discrete‐time survival analyses, with relational continuity measured by the COC index. The results presented are HRs for the key variables of interest from our preferred model, the correlated random effects which accounts for unobserved confounding. Results are also presented for comparison from the model which does not account for unobserved confounding, the random‐effects model. (Full results for each outcome from the correlated random‐effects model are presented in Table .) Higher relational continuity as captured by the COC index was associated with 11 percent lower risk of ED presentation (HR 0.89, 95% CI 0.83‐0.96) for those with moderate visit frequency and 8 percent lower for frequent attenders but of borderline statistical significance (HR 0.92, 95% CI 0.84‐1.00, P = .057). Higher continuity was associated with 23 percent lower risk of ACSC admission (HR 0.77, 95% CI 0.65‐0.91) for those with moderate visit frequency and 27 percent lower for frequent attenders (HR 0.73, 95% CI 0.62‐0.87). Risk of SMI admission did not differ by level of continuity for moderate or frequent attenders. Having a care plan documented in the previous 12 months was associated with 29 percent lower risk of ED presentation (HR 0.71, 95% CI 0.66‐0.76), 39 percent lower risk of SMI admission (HR 0.61, 95% CI 0.55‐0.68), and 32 percent lower risk of ACSC admission (HR 0.68, 95% CI 0.60‐0.77). The standard approach (random effects) to modeling continuity, which does not account for unobserved confounding, produced different results, especially regarding care plans, as seen in the final column of Table . This approach found that higher relational continuity was associated with lower risk of ED presentation and lower risk of ACSC admission, at both moderate and high visit frequency, and that care plans were associated with higher rather than lower risk of SMI admission. Table shows that lower risk of ED presentations is stronger when relational continuity is measured with the SECON index than UPC or COC, and there was some association with lower risk of SMI admission with UPC and SECON, but otherwise the results have a similar pattern across the different indices. 3.3 Robustness check results Table shows that varying the minimum number of visits deemed necessary to measure continuity, from three visits in the main analysis to two or four visits, did not substantially change the overall findings. All‐cause unplanned hospital admissions, shown in Table , demonstrate a similar pattern to ACSC admissions, with both care plans and higher relational continuity for both moderate‐ and high‐frequency attenders associated with lower hazard of admission. Adding a variable to capture specialist mental health care in the 12‐month lookback period required limiting the observation period to 3 years from April 1, 2011, to March 31, 2014. Results from the shortened observation period are presented with and without the addition of the specialist mental health care variable to allow the impact of each change to be considered separately, as shown in Table . The shorter period of observation results in a lack of statistically significant associations between continuity and outcomes, except for a lower risk of ACSC admissions for those with moderate visit frequency. While the specialist mental health care variable is associated with a much higher risk of all three outcomes, its addition does not change the results for the continuity variables.
Sample The sample consisted of 19 324 individuals attending 215 practices, observed for 15.8 3‐month periods on average (range 1‐28 periods). Table presents the characteristics of individuals in the sample. Half of the sample (50.3 percent) had an ED presentation at some point during the observation period, 13.1 percent had an admission for SMI, and 12.8 percent had an ACSC admission. Using a three‐visit minimum to define continuity, median (mean) values for each continuity index were as follows: COC 0.35 (0.46), UPC 0.67 (0.65), and SECON 0.17 (0.26). A care plan had been documented in the previous 12 months for 40 percent of the periods observed. The Spearman rank correlation between COC and UPC indices was 0.94 ( P < .001), between COC and SECON was 0.55 ( P < .001), and between UPC and SECON was 0.47 ( P < .001). Mean COC in periods with a care plan in the previous 12 months was 0.47, compared with 0.45 in periods without a care plan in the previous 12 months (two‐sample t test of difference in means: P < .001); the equivalent for UPC was 0.67 vs 0.66 ( P < .001) and for SECON was 0.24 vs 0.23 ( P < .001).
Association between continuity of care and unplanned hospital use Table presents the association between continuity of care and each outcome from the discrete‐time survival analyses, with relational continuity measured by the COC index. The results presented are HRs for the key variables of interest from our preferred model, the correlated random effects which accounts for unobserved confounding. Results are also presented for comparison from the model which does not account for unobserved confounding, the random‐effects model. (Full results for each outcome from the correlated random‐effects model are presented in Table .) Higher relational continuity as captured by the COC index was associated with 11 percent lower risk of ED presentation (HR 0.89, 95% CI 0.83‐0.96) for those with moderate visit frequency and 8 percent lower for frequent attenders but of borderline statistical significance (HR 0.92, 95% CI 0.84‐1.00, P = .057). Higher continuity was associated with 23 percent lower risk of ACSC admission (HR 0.77, 95% CI 0.65‐0.91) for those with moderate visit frequency and 27 percent lower for frequent attenders (HR 0.73, 95% CI 0.62‐0.87). Risk of SMI admission did not differ by level of continuity for moderate or frequent attenders. Having a care plan documented in the previous 12 months was associated with 29 percent lower risk of ED presentation (HR 0.71, 95% CI 0.66‐0.76), 39 percent lower risk of SMI admission (HR 0.61, 95% CI 0.55‐0.68), and 32 percent lower risk of ACSC admission (HR 0.68, 95% CI 0.60‐0.77). The standard approach (random effects) to modeling continuity, which does not account for unobserved confounding, produced different results, especially regarding care plans, as seen in the final column of Table . This approach found that higher relational continuity was associated with lower risk of ED presentation and lower risk of ACSC admission, at both moderate and high visit frequency, and that care plans were associated with higher rather than lower risk of SMI admission. Table shows that lower risk of ED presentations is stronger when relational continuity is measured with the SECON index than UPC or COC, and there was some association with lower risk of SMI admission with UPC and SECON, but otherwise the results have a similar pattern across the different indices.
Robustness check results Table shows that varying the minimum number of visits deemed necessary to measure continuity, from three visits in the main analysis to two or four visits, did not substantially change the overall findings. All‐cause unplanned hospital admissions, shown in Table , demonstrate a similar pattern to ACSC admissions, with both care plans and higher relational continuity for both moderate‐ and high‐frequency attenders associated with lower hazard of admission. Adding a variable to capture specialist mental health care in the 12‐month lookback period required limiting the observation period to 3 years from April 1, 2011, to March 31, 2014. Results from the shortened observation period are presented with and without the addition of the specialist mental health care variable to allow the impact of each change to be considered separately, as shown in Table . The shorter period of observation results in a lack of statistically significant associations between continuity and outcomes, except for a lower risk of ACSC admissions for those with moderate visit frequency. While the specialist mental health care variable is associated with a much higher risk of all three outcomes, its addition does not change the results for the continuity variables.
DISCUSSION We found that within‐practice family physician relational continuity for people with SMI was associated with a lower risk of ED presentations and ACSC admissions, and all‐cause unplanned admissions. These effects were present after accounting for time‐invariant confounding, and across three dimensions of relational continuity as captured by three different continuity indices. We did not find significant association between relational continuity and risk of SMI admission. Consistent with a previous study of care plans in family practice for people with SMI, we found that care plans, which may represent informational/ management continuity, were associated with lower risk of ED presentations, but unlike that study (which did not account for time‐invariant confounding), we found that care plans were also associated with lower risk of SMI admissions. We also found care plans were associated with lower risk of ACSC admissions. Our results suggest that seeing the same family physician over time can improve the physical health of people with SMI and thereby reduce their need for and use of unplanned hospital care. These findings are consistent with previous studies that found relational continuity to be associated with reduced risk of ACSC admission in a range of different patient groups. , Higher continuity of family physician care may reduce the need for hospital care through improved management of physical health, by facilitating familiarity, communication, trust, and quality of relationship between doctor and patient. The results also suggest that the documentation and sharing of information and management plans across physicians within a family practice can have important benefits for both the physical and mental health of people with SMI. Documentation of care plans was associated with reduced risk of all types of unplanned hospital care. Our results also highlight the importance of accounting for the individual's propensity to receive continuity of care when studying the impact on outcomes. We used a modeling technique, the correlated random‐effects model, that separates within‐ and between‐individual variation, a method not previously applied (to our knowledge) in the context of continuity of care. The results suggest that unobserved individual factors may drive both the level of continuity of care received, and the risk of unplanned hospital use, and these omitted factors may bias the observed association between continuity and outcomes. The comparison of our main results with those from a model that does not account for this type of endogeneity (the random‐effects model) shows that we would have drawn different conclusions from such an approach. We would have found that care plans were associated with a higher rather than lower risk of SMI admissions, not associated with ACSC admissions, and weakly associated with ED presentations. One explanation for this difference is that people with more severe SMI may be more likely to have a care plan documented and are also more likely to be admitted, which drives the association in the random‐effects model. When we accounted for this unobserved propensity to have a documented care plan overall, having a care plan documented in the prior year was associated with a lower risk of unplanned hospital use. The correlated random‐effects model also showed a weaker association between relational continuity and ED presentations than the random‐effects model, unlike the study by Pu and Chou which found a stronger effect of continuity when they applied the instrumental variable approach to address endogeneity. However, in addition to the methodological differences, that study looked at across‐practice continuity, which might have different unobserved confounding factors. In addition to accounting for unobserved time‐invariant factors, other features of our analysis differ from approaches generally taken in this literature. Relational continuity and informational/management continuity (as represented by care plans) were separately captured in the model, which avoided conflating these effects. We also focused on within‐practice continuity, in which different physicians within a practice have access to the same patient records, and may share common approaches to management. We took this approach because within‐practice continuity may be more relevant to family practices than across‐practice continuity, especially in England where patients are registered with a single family practice, and practices can influence which of their doctors see individual patients. Family physician relational continuity in this context may also reflect different factors than in other countries where patients face lower administrative barriers to changing family practice. Where people are free to choose their provider, high relational continuity may reflect a strong, valued therapeutic relationship, which may in turn improve outcomes. In England, people may have more constrained choice of family physician, so that higher relational continuity may be less beneficial. We found slightly lower levels of continuity than those in an earlier study of family physician continuity for people with long‐term mental illness in the United Kingdom, but much lower than those found in studies looking only at specialist mental health care. , Higher, and rising, rates of consultation in family practice may contribute to these differences. Relational continuity in English family practices may be affected by reductions in full‐time working and increasing practice size. Average UPC scores for all patients in 2011‐2013 were 0.61. Comparison with our results suggests this dimension of family physician continuity is not lower for people with SMI than for patients overall. 4.1 Limitations The clinical outcomes we have examined are important as they represent some of the excess health risks for people with SMI, and carry substantial health care costs. However, they are not the only outcomes that matter. Both people with SMI and family physicians value continuity of care in itself as part of how they experience giving and receiving care. , Broader outcomes important to people with SMI may also be affected by continuity of care, including social functioning and quality of life. While our statistical approach accounted for time‐invariant unobserved individual characteristics, we cannot rule out time‐varying confounding that may contribute to our findings. For instance, during periods of deterioration leading to admission, family physicians may have less opportunity to spend time on preventive measures such as care plans. We were unable to differentiate the nature of ED presentations into physical and mental health as done for admissions because this level of detail was not sufficiently recorded in the original data. Care in specialist mental health services might be expected to confound the relationship between continuity in family practice and hospital use. However, we found that although specialist care was strongly associated with higher risk of each outcome, there was no change in the associations between the continuity and each outcome. While this was tested on a smaller sample with a shorter observation period due to data constraints, it provides reassurance that our main results are not biased by the absence of specialist mental health care in the model.
Limitations The clinical outcomes we have examined are important as they represent some of the excess health risks for people with SMI, and carry substantial health care costs. However, they are not the only outcomes that matter. Both people with SMI and family physicians value continuity of care in itself as part of how they experience giving and receiving care. , Broader outcomes important to people with SMI may also be affected by continuity of care, including social functioning and quality of life. While our statistical approach accounted for time‐invariant unobserved individual characteristics, we cannot rule out time‐varying confounding that may contribute to our findings. For instance, during periods of deterioration leading to admission, family physicians may have less opportunity to spend time on preventive measures such as care plans. We were unable to differentiate the nature of ED presentations into physical and mental health as done for admissions because this level of detail was not sufficiently recorded in the original data. Care in specialist mental health services might be expected to confound the relationship between continuity in family practice and hospital use. However, we found that although specialist care was strongly associated with higher risk of each outcome, there was no change in the associations between the continuity and each outcome. While this was tested on a smaller sample with a shorter observation period due to data constraints, it provides reassurance that our main results are not biased by the absence of specialist mental health care in the model.
CONCLUSIONS Our results suggest that continuity of care in family practice, in terms of relational continuity and information/management continuity, can help to improve both the physical and mental health of people with SMI. Within‐practice relational continuity may reduce the risk of ED presentations and admission to hospital for physical health problems amenable to primary care, and care plans documented by family physicians may reduce the risk of patients presenting to ED or requiring admission. Our findings also suggest that it is important to consider confounding by unobserved individual characteristics when examining the relationship between relational continuity and clinical outcomes. This may be particularly important when considering trade‐offs between continuity of care and other good‐quality aspects of care provision, such as flexibility to respond to urgent needs, or when addressing the resource implications of prioritizing continuity of care in the organization of services.
Click here for additional data file. Click here for additional data file.
|
The Financial Burden of Setting up a Pediatric Robotic Surgery Program | f8289b53-7b2b-410d-8fb9-98e74c1640d9 | 6915423 | Pediatrics[mh] | Nowadays, robotic surgery is the technological cutting age in surgery. Clear benefits such as improved ergonomics, tremor filtering, three-dimensional visualization, and magnification have been well demonstrated in both adult and pediatric surgery . Even though there is no consensus of improved clinical outcomes, these advantages are, at least in theory, in favor of a more accurate, more precise surgery . Unfortunately, the main limit for the wide adoption of this technology in many pediatric surgery centers is the high cost of purchasing and running the robotic system . The few limited and varying reports on the cost of running a robotic surgery in pediatric surgery has failed to lead to a consensus . The two proven facts are, firstly, the technology is efficient and has certain advantages over conventional and laparoscopic surgery, and secondly, the costs are significantly higher. The ongoing and lively debate is over the cost-effectiveness of robotics in pediatric surgery and if the benefits for the patient justify the financial burden of running such a program . This cost-effectiveness is influenced by many additional factors such as the type of surgery, the number of procedures per year, whether or not the robot is shared with other specialties, the type of medical system or medical insurance system and, national income, among other factors. This report will focus on the costs assessment of running a robotic program in a pediatric surgery center in Romania for the first 12 months since implementation.
2.1. The Robotic Surgery Program The daVinci Xi surgical system (Intuitive Surgical Inc., Sunnyvale, CA, United States), together with instruments for 50 procedures and 1 year maintenance cost, were purchased and installed in our hospital with financing from an international benefit foundation. Our hospital is a dedicated pediatric center and therefore the robotic system is not shared with other surgical specialties. Two surgical teams consisting each of one console and one surgical chart surgeons successfully went through the training pathway. Scrub nurses and additional personnel were trained on site. 2.2. The Cases First, procedures were performed in February 2018 and over the following 12 months we performed a total of 40 robot-assisted procedures in children. We performed a wide spectrum of surgical procedures . In the process of patient selection, we considered the cases that were best served by minimal invasive surgery, involved patients with no significant comorbidities, and had low risk for complication. Surgical procedures involving more complex maneuvers, reconstruction of structures, or delicate structure dissection were considered complex, these being pyeloplasty, splenectomy, splenic cyst treatment, nephrectomy, and cholecystectomy. The rest of the procedures were considered of low risk and less challenging. 2.3. Data Collection We performed a prospective longitudinal study and recorded data regarding age, gender, pathological condition and comorbidities, surgical procedure, time of surgery, complications, intensive care unit (ICU) stay and hospital stay (HS), cost related to medication, robot instruments and consumables, other consumables, cost for hospital stay, and other additional cost. We recorded for each case the income received by the hospital from the National Insurance Company (NIC). This income is calculated by multiplying the Case Mix Index (CMI) with the tariff for solved cases offered by the NIC. 2.4. Cost Analysis We recorded the costs for each category of expenses: instruments for the robot, consumables for the robot (cover sheets, sealing caps, etc.), consumables for surgery and medical maneuvers (surgical gloves and gowns, cover sheets, sutures, disinfectant, syringes, dressings, etc.), cost for stationary hospital stay and ICU stay, costs for pain medication, antibiotics, and infusion solutions, among others. In accordance with the hospital financial policy, the personnel costs are included into the hospital stay cost. The training costs for the surgical teams were covered entirely by the company providing the equipment and were not subject of this analysis. The cost of acquisition of the robotic system was not included in this assessment as it was the subject of a third-party donation. The cost for maintenance for the first year of use was included in the acquisition cost and thus it is not the subject of the current assessment. We mention it in the current analysis with referral to the further operational cost (€150,000 per year) in the years to come. 2.5. Statistical Analysis We assessed the influenced of the different parameters such as age, gender, type of procedure, comorbidities, intraoperative incidents, postoperative complications, and the need for non-scheduled reinterventions onto the costs. The unpaired t -test was used with a significance threshold set at p = 0.05 for 95% CI. We used Pearson’s product-moment correlation to calculate the if there was a correlation between the different parameters such as types of procedures, age, sex, weight, comorbidities, and the different categories of cost. This study was approved by Ethics Committee (no. 132/2019, 14.07.2019).
The daVinci Xi surgical system (Intuitive Surgical Inc., Sunnyvale, CA, United States), together with instruments for 50 procedures and 1 year maintenance cost, were purchased and installed in our hospital with financing from an international benefit foundation. Our hospital is a dedicated pediatric center and therefore the robotic system is not shared with other surgical specialties. Two surgical teams consisting each of one console and one surgical chart surgeons successfully went through the training pathway. Scrub nurses and additional personnel were trained on site.
First, procedures were performed in February 2018 and over the following 12 months we performed a total of 40 robot-assisted procedures in children. We performed a wide spectrum of surgical procedures . In the process of patient selection, we considered the cases that were best served by minimal invasive surgery, involved patients with no significant comorbidities, and had low risk for complication. Surgical procedures involving more complex maneuvers, reconstruction of structures, or delicate structure dissection were considered complex, these being pyeloplasty, splenectomy, splenic cyst treatment, nephrectomy, and cholecystectomy. The rest of the procedures were considered of low risk and less challenging.
We performed a prospective longitudinal study and recorded data regarding age, gender, pathological condition and comorbidities, surgical procedure, time of surgery, complications, intensive care unit (ICU) stay and hospital stay (HS), cost related to medication, robot instruments and consumables, other consumables, cost for hospital stay, and other additional cost. We recorded for each case the income received by the hospital from the National Insurance Company (NIC). This income is calculated by multiplying the Case Mix Index (CMI) with the tariff for solved cases offered by the NIC.
We recorded the costs for each category of expenses: instruments for the robot, consumables for the robot (cover sheets, sealing caps, etc.), consumables for surgery and medical maneuvers (surgical gloves and gowns, cover sheets, sutures, disinfectant, syringes, dressings, etc.), cost for stationary hospital stay and ICU stay, costs for pain medication, antibiotics, and infusion solutions, among others. In accordance with the hospital financial policy, the personnel costs are included into the hospital stay cost. The training costs for the surgical teams were covered entirely by the company providing the equipment and were not subject of this analysis. The cost of acquisition of the robotic system was not included in this assessment as it was the subject of a third-party donation. The cost for maintenance for the first year of use was included in the acquisition cost and thus it is not the subject of the current assessment. We mention it in the current analysis with referral to the further operational cost (€150,000 per year) in the years to come.
We assessed the influenced of the different parameters such as age, gender, type of procedure, comorbidities, intraoperative incidents, postoperative complications, and the need for non-scheduled reinterventions onto the costs. The unpaired t -test was used with a significance threshold set at p = 0.05 for 95% CI. We used Pearson’s product-moment correlation to calculate the if there was a correlation between the different parameters such as types of procedures, age, sex, weight, comorbidities, and the different categories of cost. This study was approved by Ethics Committee (no. 132/2019, 14.07.2019).
The following nine procedures were performed with the help of the daVinci surgical system: appendectomy, cholecystectomy, inguinal hernia, ovarian tumor removal, pyeloplasty, splenectomy, splenic cyst fenestration, varicocele repair, and nephrectomy . Complex procedures were performed in 19 cases and less demanding procedures in 21 cases. There were 27 female and 13 male patients, ranging from 23 months to 24 years old, with a mean age of 13.3 years. We had eight intraoperative incidents and seven postoperative complications. In five of the seven postoperative complications there was a direct link with the intraoperative incident. Conversion to open surgery was necessary in one case and non-scheduled reinterventions in five cases, none of these being carried out with the help of the surgical robot. Mean hospital stay varied from 2 to 43 days, with a mean of 7.03 days. ICU stay ranged from 1 to 10 days, with a mean of 2.37 days. The total cost per case ranged from €1880.07 to €9851.78, with a mean of €3260.63. The cost for instruments (≈37%), the cost for ICU stay (≈26%), sterile draping for the robot (≈11%), and hospital stay (≈10%) were the major components of the total costs . The cumulative cost related to anesthesia, medication, blood tests, and other materials represented ≈15% of total costs. The percent of the direct cost for operating the surgical robot was ≈48% of total cost per case with variations depending of the surgical procedure from 33.1–70.4% of the total cost per case . On the other hand, the direct cost per procedure for operating the surgical robot (instruments + sterile draping) were relatively steady from €1077.98 to €2281.50, (mean €1579.81), regardless of the surgical procedure ( p = 0.42). Age, gender, obesity, or other comorbidities had no influence on the costs ( p > 0.05). The parameters that reveled significant impact on the costs were type of procedure, intraoperative incidents, postoperative complication, and non-scheduled reinterventions ( , , and ). All of these four parameters had significant influence on the total costs per case ( p < 0.05) with little or no influence ( p > 0.05) over the direct cost for operating of the robot. The complexity of the procedure influenced directly the ICU stay-related costs and the costs for anesthesia and antibiotics, whereas it had no influence on other types of costs . The other three parameters (intraoperative incidents, complication, and reinterventions) had significant influence ( p < 0.05) onto the cost for hospital and ICU stay and cost for materials, antibiotics, and blood tests, and had little or no influence on the cost for anesthesia ( , and ). The reimbursement from the NIC varied from €172.59 to €1879.14 (mean €552.60) and was directly influenced by the type of the procedure, intraoperative incidents, complications, and reinterventions ( p < 0.05) ( , , and ). However, the mean value of the reimbursement from NIC was 16.9% (minimum 5% to maximum 56%) of the total cost per case. The deficit per case ranged between €1119.10 and €8626.71 (mean €2708.02). The reimbursement versus non-robot related cost ratio ranged from 8% to 134% (mean 38.4%) and was ≥100% in one case.
Our study focuses on the financial aspects of running a pediatric surgery-dedicated robotic program in a former communist country still caught in a poorly efficient Bismarck-like model of healthcare insurance. As many others before us, we realize that in the cost-benefits equation of the robotic surgery, the cost is the main issue . The discussion over the benefits of minimal invasive surgery in general against robotic surgery, as the more technical advanced form over the classic, open-fashion surgery, is almost pointless, as all the scientific evidence is in favor of minimal invasive surgery . When trying to successfully implement a pediatric robotic surgery program, one has to consider two main categories of costs: the initial investment consisting of acquisition of the equipment and training of the personnel, and the costs for running the program—instruments and other consumables, maintenance, and medical costs nonrelated directly with the robot (medication, hospital stay, etc.). The initial investment is significantly high and not many medical institutions can afford it . When the investment is made by the medical care institution, this cost has to be included in the cost per patient and is directly influenced by the number of patients served by the robot. In our case, the equipment was the subject of a donation with both charity and scientific medical research purposes. This helped us by removing a significant financial burden, otherwise estimated to approximately $5000 per patient . The cost for maintenance of the equipment (€150.000 per year) adds significant financial burden per case and may be influenced only by increasing the number of cases per year. In our analysis, we did not include this cost because for the first year after acquisition of the robot, the maintenance was covered by the vendor. Otherwise the cost per case would have been higher at €3750€ per case. Increasing the number of cases per year will decrease this category of costs per case. On the other hand, to increase the number of cases means increasing all other categories of costs. Unfortunately for our patients, the budget per case is in deficit even without the cost for maintenance and by increasing the number of cases we would only deepen this gap with ≈€2700 per case. This situation is mainly due to the low reimbursement rate per case, as the mean cost per procedure is similar in our series to other reports . On the basis of several reports from all over the world, Tedesco et al. calculated the break-even point to be a minimum of 349 surgical procedures per year . This break-even point is influenced not only by the cost of the procedures but also by the health insurance system as well. This means that in order to reach a break-even point for robotic procedures in Romania there is a need to revise the specific legislation and reimbursement protocols. Perhaps introducing specific diagnosis related groups (DRG) codes and reimbursement rates for robotic procedures may be a solution. The current classification of medical procedures does not include any specifications to robotic surgery in neither pediatric nor adult procedures . In our series, we performed quite a large range of surgical procedures with the help of the daVinci robotic system. Robot-assisted surgery is especially suited to procedures requiring fine dissection and precision in suturing . We chose to also perform less demanding procedures because it was our intention to best serve our patients and therefore to also assess the added value of the surgical robot in these cases. With regards to the economical aspect, we found that in non-complex cases such as appendectomy and varicocele, the percentage of the robot-related expenses exceeded the cost of other medical expenses such as hospital stay, anesthesia, and medication, whereas in complex cases such as pyeloplasty this ratio was reversed. Meanwhile, the cost for surgical instruments was similar in complex and less demanding procedures ( p > 0.05), whereas the reimbursement rate was higher in in complex procedures ( p < 0.05). This means that in less complex cases, the burden of using the robot is higher in relation to the total cost per cases and there is probably less economic justification to use it as a routine surgical approach in the absence of scientific evidence of superior medical results. Unfortunately, we have only a few reports to compare our results to. For financial reasons, most of the pediatric surgeons retain themselves of performing these kind of procedures on routine basis . We assessed in our series two main categories of costs: robot-related costs, consisting of costs for instruments and cover sheets, and non-robot-related costs. Almost 50% of expenses were the robot-related costs (37% for instruments and 11% for sterile draping). These costs were relatively steady regardless of the surgical procedures ( p > 0.05). Unfortunately, they were little or non-amendable as their manufacturing involves high and expensive technology; there is only one manufacturer for these instruments meaning no real market competition. The non-related costs are the ones that have the potential to be amended in order to lower the costs per case, ideally to the break-even per profit point. In our series, the bulk of the non-robot-related expenses were towards hospital stay, including ICU stay, operating room (OR) time, medical personnel salaries, and other logistic expenses. Therefore, it is obvious that here is where our strategies of cost reduction have to aim. Several other studies have proven that robotic approach is beneficial towards reducing the hospital stay and OR time . A well-trained and experienced team, rigorous selection of the suitable cases, and appropriate procedures, are key factors, among others . Anesthesia cost accounts for ≈5% of the total cost per case and is higher in the more complex procedures, probably due to increased OR time, although the intraoperative incidents had no direct influence over the cost of the anesthesia. Even though we were not able to calculate directly the influence of the increased OR time over the costs of the procedures, the link is obvious and has been addressed by other studies . Increased caution during the preparation, the induction of anesthesia, the time for positioning the patient, and the time for docking the system all have influence over the OR time. Reintervention means additional OR time and a second anesthesia, thereby increasing the cost. Lab tests, medication, and other medical materials were all in direct relation with the unfavorable course of the cases and had little influence from the type of procedure. As such, we found that less incidents, complication, and reinterventions translated to lower costs. This can be achieved also by having a well-trained and experienced team and a rigorous selection of the suitable cases and procedures . The profitability of a robotic surgical program depends upon multiple factors and is currently very difficult to reach. In comparison with open or laparoscopic approach, the cost for similar procedures are higher when performed with a robot . Only a few reports for only a few specific procedures such as robotic prostatectomy in adults were in the vicinity of being more profitable with a robot . In pediatric surgery and urology, even though pyeloplasty is the most frequently performed robotic procedure in children, the costs are still higher than with other types of surgery . In our series, the balance was negative for all the 40 cases, regardless of the type of procedure. The mean reimbursement of 16.7% (max 56%) did not cover the fix cost of robot consumables without taking into consideration the cost for maintenance or acquisition of the equipment. Our concerns are not towards the profitability of the robotic program but to the mere sustainability of it. Cost minimization strategies cannot solve the problem in our situation. Our health insurance system, in its current form, cannot be the only financing source for a robotic surgery program in Romania. Additional funding sources, mainly non-governmental sources and research projects, are the ones that can and are sustaining such a program.
The mean cost per robotic procedure excluding the cost for acquisition of the equipment and the cost for maintenance was €3260.63 but with high variations (€1880.07 to €9851.78) depending on the surgical procedure, occurrence of incidents, complications, or the need for reinterventions. The robot-related costs (instruments + sterile draping of the robotic arms) were relatively steady regardless of the surgical procedure €1077.98 to €2281.50 (mean 1579.81€). The highest robot non-related costs were due to hospital and ICU stay (including OR use and medical personnel). The reimbursement from the NIC ranged from a minimum of 5% to a maximum of 56% (mean 16.9%) of the total cost per case. The deficit per case ranged between €1119.10 and €8626.71 (mean €2708.02). In Romania a pediatric surgery robotic program is not cost-efficient and cannot operate relying solely on the health insurance system.
|
Interactive multidisciplinary pilot workshop to improve medical student perception of and interest in breast surgical oncology | 2e848087-8256-4d9c-b654-c65ab1d53562 | 10882165 | Internal Medicine[mh] | Introduction The number of medical students selecting a career in general surgery has declined in recent years, resulting in focused efforts to improve exposure and educational content for students . Research shows that medical students are open to the idea of pursuing a career in surgery in pre-clinical years but this engagement decreases in subsequent years . In particular, there is minimal exposure to more complex areas of surgery, such as breast surgical oncology (BSO) and the multidisciplinary management of patients with breast cancer. Although pre-clinical medical school curriculum teaches some common breast pathologies and histological findings, the clinical years surgical exposure to BSO or complex general surgical oncology is institution-specific and not standardized. In the clinical setting, breast-related content has focused on basics of breast physical examinations but not approaches to management of breast-specific issues . Treatment of patients with breast cancer often involves a team of breast surgical oncologists, radiologists, medical oncologists, radiation oncologists, pathologists, and plastic surgeons, amongst other specialties. As a result, physicians in multiple specialties such as internal medicine, family medicine and obstetrics and gynecology will routinely encounter breast-related disease; thus, appropriate screening, work-up and referrals are crucial to patient care. There is a clear need to augment both the exposure to and education on breast cancer at the medical student level. Given that medical students in traditional United States programs must decide their specialty by the end of their 3rd year for sub-internships and residency applications, lack of exposure may prevent students from considering BSO and therefore deter them from pursuing a general surgery residency program. Research has shown that exposure through anatomy-focused pilot programs and pre-clinical surgical exposures increases medical student interest and perception of surgery . While this data is promising for increasing surgical interest among medical students, there is currently limited data on implementation of an educational program specific to BSO for medical students. The purpose of this pilot study is to assess changes in medical student perceptions of BSO following a hands-on interactive workshop exposure to BSO. Methods A multi-specialty, interactive workshop composed of breast radiology (BR), breast surgical oncology (BSO) and breast plastic and reconstructive surgery (B-PRS) was hosted for pre-clinical medical students at a single academic medical center. Students were invited to register for the event via class-wide emails. Participation in the event was voluntary, free of cost, and held in the evening after required medical school classes. The facilitators of the event were School of Medicine faculty and residents from their respective departments. The event was a station-based program utilizing soft-embalmed Thiel cadavers in the anatomy lab. Students were divided into small groups and spent 30 min rotating through each station. BR presented screening and diagnostic breast imaging followed by a hands-on ultrasound-guided biopsy experience on phantom simulators. BSO demonstrated lumpectomy, mastectomy, sentinel lymph node biopsy, and axillary lymph node dissection procedures on soft embalmed cadavers. B-PRS demonstrated oncoplastic reduction, mastopexy, and autologous flap reconstruction on the cadavers. Surgical faculty and residents were present at the stations to perform the demonstrations, allow students to practice on the cadaveric tissue, and answer any questions. Pre- and post-workshop surveys were designed in and distributed to students via REDCap electronic data capture tools to assess demographic data, current career interests, opinions on surgery, previous surgical exposure, and attractants/deterrents to a career in BSO . Descriptive statistics were utilized to evaluate medical student demographic data, current career considerations, and previous surgical exposure. Participants' changes in their interest in BSO and confidence in pursuing a surgical specialty were compared in the pre/post surveys using Wilcoxon signed rank tests. Associations between prior surgical exposure and interest in surgical specialty were compared using Fischer's Exact test and Wilcoxon Rank Sum tests, respectively. All analyses were conducted using R (version 4.2.1), and a p-value <0.05 was used to define statistical significance. Results A total of 24 students attended the workshop and 23 (96%) completed the pre- and post-workshop surveys. Descriptive statistics for study participants are outlined in . Nearly all participants were female (95.7%) and a majority were in the first year of medical school (78.3%). Of the workshop attendees, 34.8% self-identified as a non-traditional student who did not apply to medical school shortly after completing undergraduate education and 22% as an underrepresented minority student from racial and ethnic groups that are underrepresented in the medical profession relative to the general population. Only three attendees (12.5%) were first generation college students, although a majority (83.3%) were the first in their family to attend medical school. Few students (8.3%) reported having a family member who is a surgeon (see ). Eighteen students (75%) reported pre-event interest in BSO. On average, students reported an average of 2.2 (SD = 1.6) different types of previous surgical exposures, including through student interest groups (62.5%) and previous surgical shadowing (16.7%). No students reported previous shadowing specifically in BSO. Students who reported four or more types of surgical experiences (n = 6) had significantly greater confidence in their decision to pursue their selected specialty prior to the event (p = 0.04). Before completing the event, seventeen students reported interested in a surgical career, five in a non-surgical career, and one student choose to not respond to the question. On the post-survey, eighteen students reported interest in surgical careers and five in non-surgical careers, with eleven students (48%) reporting higher interest levels in the post-workshop survey. While the relative number of students interested in a career in surgery remained similar, there was improved confidence in the students’ initial surgical interest after completing the event (p = 0.04). There was an overall significant increase in student interest in BSO specifically after completion of the workshop (p = 0.003). Student perceptions of BSO after completion of the workshop were evaluated using a 5-point Likert scale, ranging from “strongly disagree” to “strongly agree” . Students reported that the workshop increased their understanding of the breast surgery scope of practice (mean = 4.6) and improved their understanding of the lifestyle of a BSO (mean = 4.1). Most respondents disagreed with the statement “Shadowing surgeons of other surgical specialties made me less interested in breast surgery” (mean = 2.3) and “The lifestyle I wish to have is not compatible with a career in breast surgery” (mean = 2.2). All participants reported that they “agree” or “strongly agree” when asked if they intended to explore BSO further through shadowing. All participants reported “agree” or “strongly agree” that the experience improved their interest and understanding of the scope of breast cancer care. Student responses regarding attractors and deterrents to breast surgery are shown in , . All respondents found BSO appealing because it is a well-respected profession and has a positive impact on patient's lives (100%). A majority also found BSO-related earnings (87.0%) and intellectual stimulation (95.7%) to be attractors. The factors perceived to be most deterrent to BSO included a lack of time for relationships and hobbies (78.3%) and stress (65.2%). There were concerns about the overall competitiveness of the profession (78.3%) and competition for residency (91.3%). Students were not deterred by the thought of treating complex patients (8.7%). Discussion The breadth and quality of surgical education throughout medical school is imperative to cultivating interest, challenging perceptions, and attracting students to careers in surgery . Demonstrating the entirety of the field of surgery during a surgical clerkship is not feasible due to time constraints of typical clerkships. There are multiple subspecialty pathways that can be taken after general surgery training, but students have minimal exposure to these pathways when thinking of their long-term career goals as a surgeon. For students to make fully informed decisions about their career trajectory, it is important that surgical educators offer opportunities for students to gain exposure to general surgery, as well as the surgical subspecialties which may be considered as part of future fellowship training. The purpose of this pilot experience was to provide students with a hands-on, educational experience in BSO and demonstrate the multidisciplinary approach to breast cancer care. The event was designed to be a concise, reproducible workshop that included exposure to breast screening and diagnostic imaging, image-guided biopsy, diagnosis, surgical management, and surgical reconstructive options. The workshop structure utilized simulation to create an interactive environment where students were taught by specialty physicians. Exposure to surgical careers through simulation-based learning can stimulate interest in surgery and provides students with a low-stakes, non-threatening environment which both fosters interest and improves educational outcomes . With a minimal time commitment of only 2 h from the faculty, the degree of exposure and instruction at the pre-clinical level was shown to be effective at improving student interest in BSO (p = 0.003). Instead of a purely didactic experience, this workshop utilized soft-embalmed, Thiel cadavers for the surgical demonstrations as a means for fostering student engagement and demonstrating surgical techniques. Kimura et al. showed that surgical training through use of cadavers increases interest in surgery, as it provides more realistic instruction than traditional didactic teaching methods . Soft-embalmed cadavers, compared to traditionally embalmed cadavers in formalin or ethanol, appear more life-like, have better color retention and simulate natural tissue . The use of this type of cadaver has been found to be more preferable for surgical demonstration . Existing research supports the use of cadaver-based anatomy teaching to improve student confidence and knowledge of anatomy pertinent to BSO . However, in this laboratory session students were additionally able to assist with lumpectomy, mastectomy, axillary lymph node dissection and reconstruction demonstrations which has not been previously reported in the literature. Following the event, the most significant reported attractors to BSO were prestige (100% agree), impact on patients' lives (100% agree) and intellectual stimulation (95.7% agree). Prior studies have also reported intellectual stimulation, prestige, and the positive impact of procedures on patients to be attractors to surgical specialties . Interestingly, the attractors identified by the medical students are similar to a study conducted by the 10.13039/100005301 American College of Surgeons , where faculty surgical oncologists reported higher satisfaction than other surgical subspecialties . This implies that students’ perceptions of the field of surgical oncology mirror those of surgeons actively practicing in this area and supports our hypothesis that a single high quality event can appropriately expose students to the reality of surgical specialty practice. By increasing exposure through interactive workshops, students may decide that the initial deterrents for surgery, often cited as poor lifestyle, increased length of training and lack of collegiality in training , may be lessened if a surgical subspecialty practice and lifestyle is perceived as more desirable. Negative perceptions can be challenged by creating opportunities for students to connect with surgeons and interact in a low-pressure environment, as done in this pilot study. Although the event was open to all medical students, the students who voluntarily participated were 1st and 2 nd year students in pre-clinical training. This highlights the desire of surgical exposure at the pre-clinical stage, prior to students starting their core surgery clerkships, which has been documented in prior publications . Pre-clinical medical students are notably in an exploratory phase, learning about different career pathways and forming opinions on options. Pre-clinical students also have more free time to attend meetings and workshops, making it a suitable population to target. According to the National Resident Matching Program (NRMP) data, 12.2% of US medical students go on to pursue surgical training, and of those, 13.2% become breast surgical oncologists . Statistically, more students ultimately pursue non-surgical specialties. This workshop has educational value for students pursuing non-surgical careers as well as those interested in surgery or BSO. Breast cancer is routinely encountered in various specialties, especially primary care , and demonstrations of the process of patient workups and diagnoses may inform better screening awareness and referral patterns in the future. While our pilot study was limited by small sample size, our initial results reported here support planned yearly events focusing on multidisciplinary cancer care and surgical oncology exposure. The data specific to radiology and plastic surgery experience were not evaluated individually under the scope of this initial pilot study but will be evaluated in future studies. Longitudinal studies evaluating the impact of this pilot study and specialty choice will also be conducted. Additional limitations to our study include this being a voluntary workshop attracting students with an interest in surgery, which could have an affect on the results of the study. Conclusion Breast surgical oncology is a surgical subspecialty that medical students have little exposure to during their training. This study successfully showed that an interactive multidisciplinary pilot workshop highlighting the complex scope of breast cancer care by breast radiologists, breast surgical oncologists, and breast reconstructive plastic surgeons improves medical student perceptions of BSO as a career. Medical schools should consider incorporating similar medical student outreach events to encourage exploration of BSO and other complex multidisciplinary surgical subspecialties. On behalf of all authors, there are no financial disclosures. Amani Raheel: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Validation, Visualization, Writing – original draft, Writing – review & editing. Shreeya Dalla: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Writing – original draft, Writing – review & editing. Jalee Birney: Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Resources, Writing – original draft, Writing – review & editing. Allison M. Aripoli: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Writing – original draft, Writing – review & editing. Meredith Collins: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Writing – original draft, Writing – review & editing. Kelsey E. Larson: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Writing – original draft, Writing – review & editing. Jamie L. Wagner: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Writing – original draft, Writing – review & editing. Christa R. Balanoff: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Writing – original draft, Writing – review & editing. Jordan Baker: Data curation, Formal analysis, Software. Lynn Chollet-Hinton: Data curation, Formal analysis, Software. Lyndsey J. Kilgore: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. On behalf of all authors, the corresponding author states that there is no conflict of interest. |
Bouveret’s Syndrome as a Rare Life-Threatening Complication of Gallstone Disease—A Surgical Problem: Two Case Reports | 26127698-816d-474d-b0c4-11fbb4e5a03d | 11767053 | Surgical Procedures, Operative[mh] | Bouveret’s syndrome is a rare variant of gallstone ileus resulting from the formation of an acquired fistula between the gallbladder and the stomach or duodenum. When a gallstone passes through this fistula into the enteric system, it can cause gastric outlet obstruction . Biliary ileus is an uncommon complication of cholelithiasis, with only 0.3% to 0.5% of patients with gallstones developing this condition. Additionally, gallstone ileus accounts for approximately 1% to 4% of cases of mechanical intestinal obstruction . The clinical signs indicative of duodenal obstruction, such as a dilated stomach, presence of aerobilia, and the identification of a cholecysto-duodenal fistula, are key to the diagnosis. Further diagnostic imaging, including abdominal ultrasonography, plain abdominal radiography to visualize the gallstone, or contrast-enhanced computed tomography (CT) of the abdomen, can aid in confirming the diagnosis of Bouveret’s syndrome, a rare and life-threatening complication of gallstones . Surgical intervention, either laparoscopic or open, is often required and typically involves enterotomy with stone extraction and primary repair of the enterotomy . This report presents two cases of Bouveret’s syndrome, highlighting distinctive radiological findings, individualized management strategies, and tailored surgical approaches based on the specific conditions of the patients and intraoperative observations. The aim of this paper is to propose a surgical strategy for optimized, patient-specific treatment of suspected Bouveret’s syndrome, drawing on the available literature and our own clinical experience.
This study included two patients, aged 76 and 72, who presented to the emergency department at the Clinic for General Surgery, University Clinical Center Nis. The study was approved by the Institutional Review Board and Human Ethics Committee of the University Clinical Center Nis, Serbia (Approval No. 4258/9). Following detailed anamnesis, both patients underwent laboratory diagnostics. Subsequently, they were evaluated using plain abdominal radiography, abdominal ultrasonography, esophagogastroduodenoscopy (EGD), gastroduodenal passage imaging, and contrast-enhanced abdominal CT scans. Upon completion of the diagnostic workup, both patients were taken to the operating theatre and underwent exploratory laparotomy. Case 1: A 76-year-old female patient presented with a 3-day history of right upper quadrant pain, nausea, and bilious vomiting. She did not report hematemesis, melena, or any other significant comorbidities. The patient had a known history of gallstones and chronic cholecystitis. On physical examination, she exhibited tenderness to palpation in the right upper quadrant and epigastrium. Initial laboratory tests revealed only slight abnormalities, as follows: white blood cell (WBC) count of 19 (×10 9 /L), neutrophil percentage (Neu) of 91.1%, C-reactive protein (CRP) of 92 mg/L, blood glucose (Glu) of 124 mg/dL, urea of 76.1 mg/dL, creatinine (Crea) of 1.06 mg/dL, γ-glutamyl transferase (γGT) of 101 mg/dL, total bilirubin (tBIL) of 1.42 mg/dL, direct bilirubin (dBIL) of 1.13 mg/dL, lactate dehydrogenase (LDH) of 554 U/L, sodium (Na) of 137 mEq/L, potassium (K) of 4 mEq/L, hemoglobin (Hb) of 13 g/dL, and a platelet count of 350,000 cells/μL. Plain abdominal radiography revealed a distended, fluid-filled stomach with signs of pneumoperitoneum (a result of gallbladder perforation; however, there was no secondary peritonitis because the process was blocked by the greater omentum) and obstruction. A 4 cm shadow of calcium intensity was noted in the right lumbar region. Abdominal ultrasonography demonstrated a partially distended, thick-walled gallbladder with multiple calculi, a finding consistent with chronic cholecystitis. Follow-up contrast-enhanced abdominal CT confirmed the presence of a cholecysto-duodenal fistula in the first part of the duodenum, with a 3.9 cm × 4.0 cm × 4.0 cm gallstone lodged in the proximal duodenum ( and ). The scan also revealed sludge and calculi in the gallbladder, along with a distended, fluid-filled stomach and aerobilia. EGD was performed and identified a 4 cm gallstone in the first part of the duodenum. The clinical presentation and imaging findings were consistent with Bouveret’s syndrome and were characterized by a cholecysto-duodenal fistula and the impaction of a large gallstone in the duodenal bulb. The patient was then taken to the operating theatre for exploratory laparotomy. Intraoperatively, significant chronic inflammation and adhesion between the gallbladder and the duodenum were noted. A cholecystectomy and repair of the cholecysto-duodenal fistula were performed through the orifice of the fistula corresponding to the duodenum. The duodenum was closed with separate sutures in two layers with synthetic absorbable suture (3-0 Vicryl ® , Ethicon Inc., Raritan, NJ, USA). The small bowel distal to the stone was not dilated. The gallstone fragment removed from the duodenum measured 3.9 cm × 4.0 cm × 4.0 cm ( , and ). Postoperatively, the patient was transferred to the intensive care unit for 24 h and then moved to the general surgical ward. There were no intraoperative or postoperative complications. A naso-enteral tube was present for 6 days. The patient underwent a period of intensive rehabilitation, which included physiotherapy, total parenteral nutrition, and a gradual reintroduction of oral feeding. She remained hospitalized for 10 days postoperatively, after which she was discharged in stable condition, when she had been weaned off TPN and was tolerating a full oral diet. The patient had an uneventful recovery and remains in good health at the time of writing. Case Report 2: A 72-year-old female patient was admitted with a 3-day history of non-specific symptoms, including loss of appetite, nausea, vomiting of watery brownish content, epigastric pain, abdominal distension, and constipation. She had a longstanding history of gallstones and chronic cholecystitis. On examination, the patient appeared to be in poor general condition, subfebrile, eupneic, and tachycardic. Active movements were difficult for her due to discomfort. Palpation of the abdomen revealed a soft consistency with mild epigastric tenderness, but no peritoneal signs were present. The “churning phenomenon” was positive. Laboratory findings showed a WBC count of 15 (×10 9 /L), neutrophils (Neu) 89.3%, C-reactive protein (CRP) of 85 mg/L, blood glucose (Glu) of 141 mg/dL, urea of 81.1 mg/dL, creatinine (Crea) of 1.23 mg/dL, γ-glutamyl transferase (γGT) of 91 mg/dL, total bilirubin (tBIL) of 1.42 mg/dL, direct bilirubin (dBIL) of 1.13 mg/dL, lactate dehydrogenase (LDH) of 554 U/L, sodium (Na) of 138 mEq/L, potassium (K) of 4.1 mEq/L, hemoglobin (Hb) of 14.8 g/dL, and a platelet count of 380,000 cells/μL. Ultrasound examination of the abdomen revealed a collapsed gallbladder with multiple gallstones of varying sizes, the largest measuring 4.2 cm. This raised concern for possible perforation of the gallbladder. Plain abdominal radiography revealed the presence of gas in the bile ducts, or aerobilia . A proximal endoscopy identified a 2 cm gallstone in the duodenal bulb, which appeared yellowish-green in color. The mucosa of the posterior wall of the bulb was eroded and hyperemic, but there was no evidence of a wall defect. Further diagnostic evaluation through gastroduodenal passage imaging confirmed the presence of a duodeno-biliary fistula . In the duodenal bulb, oval filling defects consistent with Bouveret’s syndrome were observed, with the largest stone measuring approximately 6 cm in diameter . The patient was subsequently taken to the operating theatre, where an exploratory laparotomy was performed. An enterotomy was conducted to extract the stones and was followed by cholecystectomy and duodenal suturing with separate sutures in two layers (synthetic absorbable suture, 3-0 Vicryl ® ), as well as with omentoplasty. Postoperatively, the patient was admitted to the intensive care unit for 24 h, after which she was transferred to the general surgical ward. There were no intraoperative or postoperative complications. The patient underwent an intensive rehabilitation program, including physiotherapy, total parenteral nutrition, and gradual reintroduction of an oral diet. She was hospitalized for 8 days postoperatively and was discharged after being weaned off TPN, when she was tolerating a full oral diet. The patient had an uneventful recovery and remains in good health at the time of writing.
Bouveret’s syndrome was first described by Leon Bouveret in 1896, who presented two cases of patients with this condition . He was the first to describe the pathophysiology of a large gallstone obstructing the duodenal bulb after having passed through a cholecysto- or choledochoduodenal fistula, finally resulting in gastric outlet obstruction, in a condition now recognized as Bouveret’s syndrome . While gallstones typically cause obstruction in the distal ileum (60–70%), they can also be found in the proximal ileum (25%), distal ileum (10%), jejunum (9%), colon (4%), and rectum (2%). However, gallstones impacting the duodenum represent a rare occurrence, with an incidence of only 1–3% , as highlighted in the present case report. In some cases, gallstones may enlarge while traversing the fistula and the intestine due to the accumulation of fecal material and salts on their surface . Bouveret’s syndrome is a rare cause of gastric outlet obstruction wherein a large gallstone passes through a bilioenteric fistula, causing duodenal obstruction . Bouveret’s syndrome is believed to result from recurrent episodes of cholecystitis, which lead to the formation of adhesions between the gallbladder and adjacent parts of the upper gastrointestinal tract. The continuous pressure exerted by a large stone causes necrosis and perforation of the gallbladder wall, leading to the formation of a fistula with the duodenum or stomach . Risk factors for the development of Bouveret’s syndrome include advanced age (over 70 years), gallstones larger than 2.5 cm, female gender, and post-surgical alterations to the gastrointestinal anatomy . In our report, both patients were female and aged over 70 years (76 and 72 years), with 4 cm gallstones and histories of chronic cholecystitis. Key factors contributing to perforation include the size of the gallstone, the size of the bowel lumen, and the site of fistula formation . Stones smaller than 2.5 cm typically pass through the small bowel smoothly or result in gallstone ileus, a more common occurrence with larger stones . In this report, duodenal obstruction was caused by a 4 cm gallstone, which is large enough to cause significant complications requiring urgent intervention . The mortality rate for gallstone ileus is reported to range from 7% to 30% (average 18%). This high mortality is primarily due to factors such as advanced age, frailty, multiple comorbidities (particularly cardiovascular, respiratory, and endocrine conditions like diabetes and obesity), and delayed presentation (typically 4–8 days after symptom onset). The literature suggests a median delay of 2–37 days between admission and surgical intervention, with a range of 1–15 days . The clinical presentation of Bouveret’s syndrome is often nonspecific and infrequent, leading to delayed diagnosis . Symptoms can range from gastric outlet obstruction (as seen in our patients) to acute pancreatitis, upper gastrointestinal bleeding, duodenal perforation, Boerhaave’s syndrome, and gastric bezoar formation . Ileus is not typically a sign of the presence of a gallstone. The majority of the patients experience symptoms that include nausea, vomiting, and epigastric pain . In our case, both patients exhibited symptoms of nausea, bilious vomiting, and epigastric pain for three consecutive days. The choice of diagnostic approach is critical in achieving a timely and accurate diagnosis of Bouveret’s syndrome. It remains a diagnostic challenge, with approximately 50% of diagnoses confirmed only during surgical intervention . Although EGD may lead to a more definite diagnosis in Bouveret’s syndrome, a direct abdominal X-ray is still the first step in the approach to these patients, as it allows surgery to be performed as early as possible . EGD is crucial both diagnostically and therapeutically, as it allows visualization of the dilated stomach and impacted gallstone, which typically appears as a firm, non-fleshy mass. EGD can also reveal the duodenal ostium of the biliodigestive fistula . An upper gastrointestinal series with oral contrast may provide additional insight into the obstructing mass by showing a filling defect, a gallstone, dilation of the stomach or duodenum, pneumobilia, and/or outlet obstruction. In rare cases, contrast extravasation into the gallbladder can indicate a patent cholecystoduodenal or cholecystogastric fistula . In our cases, initial endoscopy confirmed the presence of the stone and the obstruction. Imaging studies, particularly contrast-enhanced CT scans, are pivotal in diagnosing Bouveret’s syndrome. The diagnosis typically begins with a plain abdominal X-ray, though this is diagnostic in only 21% of cases . In our report, the X-ray showed a 4 cm radiopaque shadow in the right lumbar region, suggesting the presence of a gallstone without signs of ileus or pneumoperitoneum. Radiologically, Bouveret’s syndrome is often identified by “Rigler’s triad”, which is seen in approximately 30–35% of cases . The classic Rigler’s triad—dilated stomach, pneumobilia, and an ectopic stone seen as a filling defect in the duodenum on CT—is considered virtually pathognomonic of Bouveret’s syndrome . In 60% of cases, abdominal ultrasonography can be helpful, revealing an ectopic gallstone, a fluid-filled distended stomach, pneumobilia, and features indicative of cholecystitis . However, its limitations include difficulty in locating the stone and interference from excessive intestinal gas. Gastroduodenal passage imaging, which is useful in approximately 45% of cases, can help identify the fistula and locate the stone, revealing oval filling defects during the migration of the calculi through the intestinal lumen . In 60% of cases, a CT scan provides definitive diagnostic information and is highly sensitive, specific, and accurate (93%, 100%, and 99%, respectively) . In 45% of the cases, as one of the radiological imaging techniques, imaging of the gastroduodenal passage can be helpful in identifying the fistula and locating the stone. Numerous defects of oval fillings can be discovered during the migration of calculi in the lumen of the intestine . In 60% of cases, a CT scan provides definitive diagnostic information, as well as an evaluation of the gallstones, fistulas, and inflammatory findings. It offers 93% sensitivity, 100% specificity, and 99% accuracy . This imaging modality is considered the gold standard for diagnosing Rigler’s triad in gallstone ileus . Plain abdominal films show Rigler’s triad in 14.8 to 21% of cases, while the rates of positive findings are 11.1% for ultrasound and 77.8% for CT scan. Although CT scans are highly effective, limitations exist in detecting gallstones with isoattenuation (15–25%), necessitating additional imaging . The impacted gallstone is endoscopically visible in 70% of cases, most likely due to the fact that the mucosa covers the embedded stone. The complete diagnosis is made during surgical procedures in 20–40% of cases . In this case report, the presence of gallstones was confirmed by all imaging diagnostic procedures. The primary aim in treating Bouveret’s syndrome is to remove the obstructing gallstone. Both nonsurgical (endoscopic) approach and surgical (open or laparoscopic) management are therapeutic options . Considering the surgical morbidity and given the fact that most patients are elderly, with multiple comorbidities, the endoscopic approach should be the first line of treatment when it comes to Bouveret’s syndrome. However, it is complicated to dislodge and remove a large, impacted stone. Thus, the endoscopic approach is not very successful and is rarely therapeutic . Moreover, some researchers feel that endoscopic lithotomy may increase the risk of esophageal injury, digestive tract perforation, and gastrointestinal bleeding. Research showed that 42% of cases did not end in stone removal by this technique . Surgical intervention is necessary in more than 91% of cases. Depending on the patient state, the surgical procedure can be one-staged enterolithotomy (or gastrotomy) with concomitant cholecystectomy and repair of the fistula or enterolithotomy alone with or without a second-stage cholecystectomy . Recently, with the extensive development and use of laparoscopy, laparoscopic surgery for Bouveret’s syndrome has shown to be a safe and effective alternative to open surgery . The optimal surgical approach should be individualized based on the patient’s age, general and local health, comorbidities, and life expectancy and the size and location of the stone and fistula . In these cases, taking into account the presence of the large, impacted gallstone in the duodenum, the endoscopic approach is not the best option. The surgical procedure involved open enterolithotomy with concomitant cholecystectomy and repair of the cholecystoduodenal fistula. Bouveret’s syndrome is associated with significant morbidity and mortality . Gallstone ileus, in general, carries a poor prognosis, particularly in the elderly. Mortality and morbidity rates have decreased over time (from 30% to 12%), yet they remain high, primarily due to delayed diagnosis and concurrent medical conditions . As a rare complication resulting from the large-scale impaction of a gallstone in the duodenal bulb and subsequent gastric outlet obstruction, Bouveret’s syndrome requires prompt and accurate diagnosis, as well as early surgical intervention, to optimize outcomes. However, due to the small sample size, more extensive research is not possible, and, consequently, accurate results should be derived from a better statistical sample. The selected strategy and approach should be patient-specific and based on characteristics such as age, general and local status, medical condition of the patient in relation to the morbidity, and the mortality rates of each approach.
Bouveret’s syndrome is a rare but critical clinical condition that should be considered in patients presenting with ileus in the upper abdomen, particularly those with a history of chronic cholelithiasis. Its symptoms are nonspecific and can be life-threatening, especially in elderly patients. Early and accurate diagnosis, followed by timely surgical intervention, is essential for effective treatment. Although endoscopy may be a useful diagnostic and occasionally therapeutic tool, surgical management is often necessary, particularly when endoscopic approaches fail or are not available. The surgical procedure can be one-staged enterolithotomy (or gastrotomy) with concomitant cholecystectomy and repair of the fistula or enterolithotomy alone with or without a second-stage cholecystectomy. The surgical strategy should be individualized, taking into account the patient’s overall health and comorbidities and the specific characteristics of the gallstone and fistula. Bouveret’s syndrome, as the appropriate treatment, can pose a challenge for the surgeon when surgery is needed.
|
Environmental health, COVID-19, and the syndemic: internal medicine facing the challenge | d20bf374-e760-4ed1-a1ba-eb295e84678c | 9525944 | Internal Medicine[mh] | Internists are experts in complexity, and the COVID-19 pandemic is disclosing unexpected interactions between communicable and non-communicable diseases, environmental and socio-economic aspects. This is a scenario which makes SARS-Cov-2 infection a part of a syndemic , rather than a “simple” pandemic. Syndemic is due to complex cross-links generated by the spread of this communicable disease in vulnerable populations suffering from an increasing epidemiologic burden of chronic diseases and disabilities, social and economic inequalities . The medicine of complexity cannot be simply limited to face comorbidities and to the clinical management of multifaceted, multidisciplinary diseases. On March 11, 2020, COVID-19 was officially recognized as a global pandemic. To date and worldwide, we are counting over 0.6 billion confirmed cases of COVID-19, including over 6.3 million deaths and, as of 31 July 2022, a total of 12.248.795.623 vaccine doses administered (WHO Coronavirus (COVID-19) Dashboard | WHO Coronavirus (COVID-19) Dashboard With Vaccination Data). Globally, healthcare professionals, policymakers, economists, and citizens are obliged to face the huge effects of the pandemic both at an individual and a global level, searching for strategies able to reduce harms and damages, but also to increase the chance for resilience. Adopting a vision of the pandemic based solely on healthcare aspects, such as disease management or vaccine prophylaxis, is simply insufficient. Instead, we need a wider view towards global public health aspects which takes advantage of the “one-health” approach, a comprehensive strategy that facilitates interdisciplinary, multidisciplinary, and transdisciplinary collaboration between the human health, animal health, and environment sectors. In particular, available evidence clearly indicates how environmental factors such as climate change [ – ], air pollution , low income and socio-economic disparities can worsen the outcomes of COVID-19 in vulnerable communities. This interaction cannot be neglected. To increase the awareness in this field, in this review we will discuss how internists can extend their role of privileged healthcare providers. We will examine the major elements characterizing COVID-19 as part of a syndemic, to explore how this harmful communicable disease interacts with environmental pollution and individual vulnerability, and to adequately manage not only the pandemic but also the growing burden of non-communicable diseases by a “one health” approach.
We are learning that the consequences of COVID-19 on public health depend strongly on individual vulnerability. Thus, the analysis of risk factors which affect incidence, prevalence, spreading and clinical outcome disease must necessarily and comprehensively consider a wide panel of health determinants. There are complex and dynamic relationships between the spread of the infection by SARS-CoV-2 and several pre-existing criticalities All these elements contributed to generate the ongoing syndemic ( Table ). The pandemic per se amplifies chronic, structural, and functional difficulties resulting from decades of inaction and/or poorly efficient policies in terms of health promotion and primary prevention of diseases. Advanced age is a well-known risk factor which increases the mortality of COVID-19 patients . This evidence has a great relevance in geographical areas characterized by increasing lifespans, such as Europe . According to the WHO, the proportion of the world’s population over 60 years will double between 2015 and 2050 and, by 2030, one in 6 people worldwide will be aged 60 years or over . Age per se, however, cannot be considered as a synonym of individual frailty and increased vulnerability to COVID-19. Short-term exposure to air pollution, for example, affects the immune function in subjects entering hospital for COVID-19 pneumonia, and increases the in-hospital mortality independently from age . Furthermore, reduced COVID-19 in-hospital mortality appears to be linked with early production of antibodies against SARS-CoV-2, and this evidence is also independent from age . The increased vulnerability in elderly parallels frailty and this aspect, in turn, is mainly a consequence of a reduced health span and of unhealthy ageing . Both frailty and unhealthy ageing are closely linked with environmental factors across the course of life. The interaction often starts during in-utero life, and becomes a predisposing factor to non-communicable diseases . Chronic, non-communicable diseases and the presence of comorbidities, in turn, are well-known factors able to worsen the clinical outcome of COVID patients . These diseases certainly contribute to the impact of the pandemic on vulnerable populations, and negatively affect the clinical outcome in infected patients . About one in five individuals worldwide is at increased risk of severe COVID-19 due to underlying health conditions, and suffers from at least one non-communicable disease . This trend is particularly true with the progressive ageing of the population and the global dual epidemic of obesity and type 2 diabetes mellitus, which drive the worst outcomes of COVID-19 patients . Notably, such leading noncommunicable diseases affect people independently from age. For example, obese younger than 60 years have a higher risk of severe COVID-19, as compared to older individuals and this finding decreases the value of age per se as a contributor to the increased risk following SARS-CoV-2 infection. An ecological study has shown that worldwide disability-adjusted life years (DALYs) due to noncommunicable disease correlated with COVID-19 cases and deaths . In a syndemic scenario, the link between noncommunicable diseases and socio-economic factors is an additional worsening element. The mortality risk from noncommunicable diseases in the age range 30–70 years is inversely related with income, with the lowest mortality recorded in high-income countries, as compared with low- and middle-income countries . Conversely, the country-level income inequality in 22 OECD countries is positively associated with COVID-19 mortality in all age groups, pointing to inequality as a significant risk factor . This evidence is supported by a large survey in the U.S. showing that the percentage of adults without a high school degree, and the proportion of black residents were the two socio-economic determinants of health with the strongest association with incidence and fatalities . A recent study exploring COVID-19 mortality in 3,144 US counties confirmed that socio-economic disparities and disadvantage condition were strong determinants of COVID-19 mortality . The negative role of progressive ageing, growing noncommunicable diseases and inequalities also links COVID-19 with the health effects of climate change. In fact, among the social categories with the highest degree of vulnerability to the health effects of global warming are aged people, people facing social disadvantages and those with chronic diseases . The burden of noncommunicable diseases related to climate change has progressively increased between 1990 and 2019 . Furthermore, the environmental pollution as the main driver of climate change, affects the onset and progression of noncommunicable diseases . Finally, unhealthy habits and lifestyles influencing the epidemiologic progression of obesity and metabolic diseases are also markedly fueling climate change, generating a huge cost for national health systems and ecological costs in relation to the environment . Animal products generate the highest values for carbon emissions , and the global increase in meat tread and consumption strongly contributes to diet-related chronic diseases . Climate change increases the risk for infectious disease transmission since decades , and the onset and spread of the SARS-CoV-2, a zoonotic coronavirus with a possible origin in bats and an inter-species transmission from bats to humans , should not be an exception to this trend. Climate change can contribute to this process , and the rapid pandemic spread is facilitated by factors driving the onset and progression of climate change, namely reduced biodiversity, growing urbanization, progressive contraction of green areas, and global hypermobility. In the syndemic context, the role of gender inequalities must also be considered, according to evidence showing increased vulnerability in sexual and gender minorities, and a sex-based difference in COVID-19 clinical outcomes. We are witnessing an unprecedented stress to the national public health systems, with the interaction between pandemic and factors such as population ageing, increased burden of chronic diseases, individual vulnerability, low possibility of resilience secondary to inequities and inequalities, sexual and gender inequalities, and the central role of living in an unhealthy environment predisposing to acute and chronic diseases. The interplay between these different factors (i.e., the syndemic) contribute to increase individual vulnerability in all age classes and to decrease the possibilities of resilience, making insufficient a “purely clinical” approach to solve primary health problems, and mandatory a “one health” approach (Fig. ). Essential guaranteed services, mainly those oriented at chronic diseases, have been disrupted due to the diversion of human and financial resources towards COVID-19. This necessary policy has inevitably created a great harm because of inadequate management of frail patients and lack of secondary and tertiary prevention measures . On the other hand, mounting evidence suggests that COVID-19 survivors with noncommunicable diseases can experience negative effects on clinical progression of several conditions, such as metabolic disorders , and cardiovascular , pulmonary , neurologic , and psychiatric diseases. The coexistence of different conditions has certainly facilitated the spread of pandemic and the crisis is still far to be solved, even though COVID-19 has the priority in international policy agenda. Despite vaccine prophylaxis contributes to face the pandemic, the lessons to be learned from the SARS-CoV-2 spreading is that we urgently need health-in-all policies in a global perspective. Starting goals must include living in a healthy environment and decreasing individual vulnerability independently of age .
The incidence of noncommunicable diseases is continuously rising and involves any age. An enormous burden of premature deaths is generated by the main four noncommunicable diseases, namely cardiovascular diseases, chronic respiratory diseases, cancer, and diabetes . These conditions also contribute to increase individual frailty and disabilities , and are generally managed by Internists. Besides lifestyle and socio-economic factors , a central role in the onset and development of these disorders is certainly played by environmental pollution , which has been defined as “the largest environmental cause of disease and premature death in the world today” . From a scientific point of view, environmental health is a consolidated field of research since decades. An editorial published in the year 1911 described as the term “environment offers a wide and fertile field for research” . In the 30 s of the last century, air pollution was firstly defined as a “serious menace to health”, describing links with a number of acute (i.e., allergies, acute respiratory disorders, pneumonia) and chronic diseases (i.e., emphysema, depression), with pediatric mortality and, finally, with cancer . In the year 1952 a paper published in the British Medical Journal described clear links between air pollution and lung cancer . Starting from the 60 s of the last century, several authors found relationships between pollution, cardiovascular mortality and extra-pulmonary, mainly gastrointestinal cancer . More recently, epidemiologic and experimental studies find that the contamination of environmental matrices (i.e., air, soil, water) and food by toxic chemicals strongly affects the onset and progression of neurodegenerative, gastrointestinal, renal, reproductive, hormonal, psychiatric, metabolic disorders, and cancer, irrespective of age. Nevertheless, although clinicians have an important role to play in reducing the global burden of diseases from pollution, the environmental health knowledge still remains virtually absent in clinical practice , also due the lack of a specific training of Internists in environmental health , which strongly limits their healthcare ability and potential. These limitations generate enormous consequences in terms of lack of primary prevention measures, inadequate and scarcely comprehensive disease management, and unsatisfactory cost saving. As with traditional risk factors of noncommunicable diseases, clinicians can identify patients at high risk from pollution, and must provide recommendations and interventions to reduce the individual risk, to optimize treatments, to reduce vulnerability and to increase resilience. Recently, 49 clinical guidelines from European, Asian, American, and Australian medical associations and organizations on typical internal medicine topics (i.e., allergies, asthma, chronic obstructive pulmonary disease, cardiovascular disease, obesity, diabetes, renal diseases, heat stroke, and colorectal cancer) have been screened to verify the presence of 30 specific keywords related with environmental and planetary health. Results revealed that most of these keywords were present in fewer than 5% of the guidelines . Thus, according to the traditional translational paradigm, the development of clinical guidelines also considering advances in environmental health is urgently needed, in terms of both policy recommendations and individual actions. Recommendations include cost–benefit evaluations, analyses of cost savings due to primary prevention measures, and sustainability in decision-making processes.
According to the global burden of disease study, ambient particulate matter pollution was among the largest increases in risk exposure in the period 2010 to 2019 causing, on average, 11.3–12.2% of all female and male deaths in 2019, respectively . Besides mortality, environmental pollution has also a major role in promoting disabilities and affects the onset and progression of other leading causes of mortality, as cardio-metabolic diseases . Of note, air pollution affects individual vulnerability to infectious diseases, including COVID-19 . Adverse health effects of several air pollutants, often without clear safe or threshold level, are well established through numerous epidemiological studies and have been acknowledged by WHO . Table reports potential mechanisms linking environmental pollution with SARS-CoV-2 infection, spread and outcomes both at community [ – ] and individual [ , – ] level. The COVID-19 pandemic offered a unique opportunity to study the impact of air pollution on the risk of infection and disease lethality, as the new Corona virus hit an immunologically naïve population and the prognosis of the disease was severe enough in many instances to foster public concern and hence the generation of a large amount of data. Not surprisingly, many papers have reported associations between air pollution levels and COVID-19. Several reviews have in the meantime been published on that issue, the most recent and detailed one by Xavier Rodó et al. for the Panel for the Future of Science and Technology (STOA) of the European Parliament. Searching for the string “COVID-19 AND air pollution” in PubMed ( https://pubmed.ncbi.nlm.nih.gov/ ), returned 1,474 results (access on February 14, 2022). Table lists the 27 most relevant studies among the first 40 papers sorted by the “best match” option of PubMed. The remaining 13 studies were excluded because of double counting (1 study), because they focused on indoor air pollution (4 studies), were not written in English (1 study), were rather about policy aspects or environmental justice (5 studies) or looked on the effect of COVID-19 lockdown measures on air quality, rather than on air quality effects on COVID-19 (2 studies). The present paper was not aimed at performing a systematic review and analysis about the relationships between air pollution and COVID-19. However, the short and non-comprehensive overview synthesized in Table demonstrates the growing scientific interest and a huge variety of approaches to the issue of air pollution effects on the risks of COVID-19 infection, including the risk of a severe course of the disease, or death. Studies have been performed in nearly every part of the world. Both long-term chronic and short-term acute exposures have been investigated and a variety of pollutants have been considered. Experimental studies and theoretical papers discussing possible mechanisms that lead to the demonstrated effects also contribute to the wealth of information. Although there is still not a full consensus on the mechanisms and the causal role of air pollutants on the onset and course of SARS-CoV2 infection, most studies are consistent in demonstrating a positive association between air pollution and the risks of this communicable, systemic disease. Most often, the following three mechanisms are proposed: Pollutants damage the airways and the immune system thus rendering an individual more susceptible to (later) infectious attack, but also co-exposure (with irritant gases) has a similar, albeit more immediate effect. Particulate matter serves as vehicle of viruses in the air protecting the viruses from UV-radiation, delaying sedimentation, and transporting the virus more efficiently into the deeper airways. Inflammation and oxidative stress caused by air pollution during the early stages of infection render the course of the disease more severe and thus lead to a higher lethality. It is vital to see that these different mechanistic hypotheses differ in the timing of the relevant exposure. Further epidemiological research can therefore help to better distinguish between these hypotheses. In many settings, monitoring data on air quality is representative for the exposure of a huge amount of people, with large health implications. This enables field studies with remarkable power that will even detect rather small effects. Thus, air pollution research serves as an example for other fields of environmental health where exposure assessment is much more difficult and costly. In these other fields, the lack of evidence should not be interpreted as evidence for the lack of effect or the absence of interactions. Studies on the relationships between the pandemic and air pollution indicate how environmental factors can play a relevant role in both communicable and noncommunicable diseases, and how the interplays between these three elements contribute to increase the complexity of their understanding and management. This evidence points, again, towards a possible central role for internal medicine in interpreting and governing these multifaceted and multidisciplinary dynamics, orienting efforts not only to clinical management of diseases but also to primary and secondary prevention measures, and to structured educational programs.
Medicine is deemed as the science and art of treating human beings suffering from injury, disease and illness . Here, internal medicine plays a privileged role due to its holistic and multidimensional approach. The history of Medicine reveals that almost universally, the management of health and diseases can be discriminated into two primary approaches: curative (according to Asclepius, god of Medicine) and health promotion/protection (according to Hygeia, goddess of hygiene and health). The event that marked the decisive turning point in such a distinction occurred in 1942, when William Beveridge introduced a “Plan for Social Security and Allied Services” into the English Parliament. In 1946 the English Parliament approved the first National Health Service and began its organisation and operation in 1948. It was based on three core principles relating to the individual and not to the general population, i.e., (1) meet the needs of every individual; (2) based on clinical need, not on ability to pay, and (3) be free at the point of entry . Many countries worldwide followed a similar pattern to that of England under the responsibility of Ministries of Health (formerly Ministries of Hygiene) or agencies similar in the national government, thus creating one of the most deceptive illusions concerning health. The illusion that mitigating, treating, caring, and sometimes curing the disease employing an industrialised organisation could improve the health of the population. What we label as health services are no more than medical services dedicated to the care of the disease–not even to those who suffer from them—ignoring health protection and promotion and a large part of preventing illness. Big business has used Medicine to build profit-making industrial complexes that offer the consumer services of laboratories, diagnostic services, outpatient care, and hospitalisation. In this context, internal medicine can also play a driver role, due to its attitude to prefer a model of clinical governance based on an individual-centred care, on the quality of outcome and on the maintenance of the health status, rather than on a mere administrative-based management . True health services are interventions that should protect and promote health and help prevent the disease from occurring, with huge advantages also in terms of reducing health costs [ – ]. A comprehensive overview of health services has included a total of 37 different items divided into five main sectors (i.e., services of health protection, individualized services for health promotion, collective services of health promotion, implementation of social capital, and preventive medical services) and ranging from environmental issues to hygiene, physical activity, lifestyle, urbanistic features, policy strategies, family planning and managing . This wide concept of health services, public health, individual health, and disease prevention must therefore involve many disciplines and competencies. This change can be easily driven by the longstanding experience accumulated by specialists in internal medicine. Over 70 occupational categories relevant to environmental health in Europe were identified in a review published in 1998, which included academics, medical specialists, environmental scientists (e.g. epidemiologists, natural scientists, social scientists and experts in hygiene occupations) and professionals (such as environmental health workers, technicians, and architects) . At variance from internal medicine, however, the “one health” role of most of these categories is limited by a single, specific field of action, that makes difficult efficient interactions with different sectors, and at a global level. To appropriately face the current COVID-19 pandemic, as also to face the progressive rise in noncommunicable diseases, we urge to reorganize health services, policies, and clinical strategies (including continuing medical education and adequate spread of medical information), towards a real awareness of the complexity of the global risks also in a local context. From this point of view, community medicine and family physicians play a critical role in facing potential future pandemics. The role is not only in terms of clinical care but also in terms of social support, screening of most vulnerable subjects, early surveillance, local monitoring of environmental health and environment-related health risk, transmission of adequate information to general public, and coordination between different health services . Hence, it is essential to facilitate and to promote coherent policy initiatives at local, national, regional, and global level. This might start with encouraging greater awareness of the global dimensions of health among policymakers and health practitioners but could then be followed by specific policy decisions to optimise the benefits and mitigate the costs of globalisation for health.
The SARS-Cov-2 pandemic is having a major impact on public health and worldwide economy. Effects can vary depending on country and individual levels of vulnerability. The massive vaccination policy has partly improved this scenario, but the global crisis generated by the spread of SARS-CoV-2 and its variants is still far to be solved. In addition, the risk of further, future pandemics is high . Evidence clearly indicates strong and complex links between the COVID pandemic, the global burden of noncommunicable diseases, demographic unbalance, individual vulnerability and unhealthy aging, environmental pollution, socio-economic inequities and inequalities and criticalities in maintaining an adequate efficiency in the national health systems. Consequently, resilience of large communities worldwide depends by multiple factors and require a thoughtful and comprehensive approach. Experts in internal medicine have knowledge and skills to drive a change in strategy, in medical education, in public health management and in specific clinical practices, since neither the pandemic nor the growing burden of noncommunicable diseases can be simply faced as a pure technical and clinical challenge. This approach requires a holistic, global health approach, multidisciplinary and multisectoral policies and long-term, adequate policies oriented towards environmental health and sustainability, prevention programs and a reduction of individual vulnerability worldwide, also through educational and coordinated programs.
|
Sleep health practices and sleep knowledge among healthcare professionals in Dutch paediatric rehabilitation | bb828858-b8ba-41a9-a66f-f382c2e9875a | 7589250 | Pediatrics[mh] | INTRODUCTION Sleep disorders are common in children with neurodevelopmental disabilities (NDDs), with a reported prevalence as high as 86% (Robinson‐Shelton & Malow, ; Simard‐Tremblay, Constantin, Gruber, Brouillette, & Shevell, ). In addition to affecting the children's physical and cognitive health and development (Turnbull, Reid, & Morton, ), sleep disorders may greatly impact on the well‐being of both the children and their families (Adiga, Gupta, Khanna, Taly, & Thennarasu, ; Mörelius & Hemmingsson, ; Sandella, O'Brien, Shank, & Warschausky, ). Therefore, improving the quality of sleep in children with NDDs not only has the potential benefit of improving their clinical outcomes (Owens ), it can greatly ameliorate the quality of life of the entire family (McDonald & Joseph, ). In recent years, there has been growing acknowledgement of the importance of sleep and the need for recognition of sleep problems by physicians (Perry et al., ). Yet there continues to be only limited education in sleep medicine across medical school curricula (McDonald & Joseph, ; Mindell et al., ). The minimal training received is accompanied by shortcomings in confidence and sleep knowledge, all of which may contribute to sleep issues not being inquired about routinely when children are seen (Chervin, Archbold, Panahi, & Pituch, ; Honaker & Meltzer, ). For children receiving developmental or rehabilitative services, like those with NDDs, paediatric rehabilitation provides an ideal place to address their sleep health as part of the routine assessments. The multidisciplinary team of healthcare professionals (HCPs) that is typically involved in the rehabilitative care of these children thereby serves a joint role in both detecting and managing sleep problems. Hence, it is crucial that (non‐)medical HCPs working in paediatric rehabilitation settings are not only aware of the importance of sleep but also possess current knowledge of basic sleep physiology, can recognize symptoms of common paediatric sleep disorders, and are familiar with good sleep hygiene practices. However, despite the high prevalence of sleep disorders in children with NDDs, according to parents, sleep has received only limited attention in paediatric rehabilitation (Hulst et al., ). Indeed, sleep problems are not always appropriately addressed in these populations (Didden, Korzilius, van Aperlo, van Overloop, & de Vries, ; Simard‐Tremblay et al., ), leaving sleep an underemphasised aspect of health in neurorehabilitation (Verschuren, Gorter, & Pritchard‐Wiart, ). This raises the question whether HCPs have sufficient knowledge and competence to address sleep issues in clinical practice. Therefore, this survey study aimed to assess the (1) sleep health practices and (2) sleep knowledge (sleep physiology, sleep disorders, and sleep hygiene) in two groups of HCPs (i.e., medical and non‐medical professionals) working within paediatric rehabilitation settings. To effectively guide parents, HCPs are required to have more sleep knowledge than the general population, and therefore, a control group was added to allow comparisons.
METHODS 2.1 Study design A cross‐sectional quantitative survey study was conducted. The study was deemed exempt from review under the Dutch Medical Research Involving Human Subjects Act by the Medical Ethics Research Committee of the University Medical Centre Utrecht, the Netherlands (file number 19‐066). 2.2 Respondents HCPs from three paediatric rehabilitation settings (rehabilitation centre, school for special education and rehabilitation department of a children's hospital) in the Netherlands participated in this study. In Dutch rehabilitation, a medically schooled physician serves a gatekeeping role in detecting child‐related problems during clinical encounters and consequently coordinates the child's rehabilitative care. When, in this case, a problem with sleep of the child is detected, the physician gives first‐line treatment advice or can decide to set up referral to a non‐medical professional or sleep clinic for subsequent sleep assessments and/or interventions. Depending on the nature of the sleep problem, a non‐medical professional may further assess the child's sleep, bed routine and/or behaviour and perform therapies like behavioural interventions and implementing healthy sleep practices. In this way, the roles of HCPs involved in sleep care in Dutch rehabilitation settings are distinct. The pen‐and‐paper surveys were distributed during live meetings among the following two groups of paediatric HCPs: Rehabilitation physicians (RPs). This group included paediatric rehabilitation physicians, physician assistants, paediatricians and doctors in specialist training to become RP. Allied health professionals (AHPs). This group included physical therapists, occupational therapists, developmental behavioural therapists, speech and language therapists, social workers and psychologists. An additional control group was drawn from the general population, comprising individuals without a background or current profession in healthcare, to allow comparisons of sleep knowledge levels. The control group consisted mostly of parents of (young) children, who were recruited via the social networks of colleagues and acquaintances, and filled out the pen‐and‐paper surveys during face‐to‐face encounters. 2.3 Data collection A 30‐item structured questionnaire was designed based on relevant literature and consultations with experts (researchers and clinicians) working in the field of paediatric rehabilitation and sleep medicine . Pilot testing was conducted to ensure that the questionnaire was easily understood and could be completed within a short time window (15–20 min). The questionnaire comprised three sections: general information including age, sex, profession, educational level and hours of sleep education received; application of sleep health practices in daily clinical practice. HCPs were asked to indicate how often they address the topic sleep during clinical encounters, choosing between never/seldom (less than once per month), sometimes (1–3 times per month) or often (once per week or more often). HCPs who reported to address sleep sometimes or often in clinical practice were asked to indicate the type(s) of sleep interventions or therapies they apply. sleep knowledge within the domains of (a) basic sleep physiology (i.e., recommended sleep durations and sleep architecture), (b) symptoms and characteristics of common paediatric sleep disorders and (c) sleep hygiene rules (i.e., healthy sleep practices). Additionally, the participants' self‐perceived knowledge sufficiency was assessed. Apart from the general information, responses were measured using multiple choice questions and included a ‘ don ' t know ’ answer option. An open‐ended question was used for collecting information about knowledge of sleep hygiene rules; respondents were asked to name three sleep hygiene rules other than the example given regarding limiting screen time 2 h before bedtime. A cover letter that explained the aim of the study and emphasized the need for honest responses (i.e., to answer with ‘ don't know ’ instead of guessing if one does not know) was attached to the survey. 2.4 Data analysis Surveys that returned largely incomplete (>25%) were excluded from analysis. Data were analysed using IBM SPSS Statistics 26. Sleep knowledge scores (i.e., number of correctly answered questions) were calculated for each group and converted into percentages; these are presented as mean ± standard deviation (SD)% scores for all knowledge questions in total and per domain sleep physiology and sleep disorders. After testing for normality, means were compared using Kruskal–Wallis test followed by post hoc Mann–Whitney tests. Answers to the open‐ended question regarding sleep hygiene rules were categorized according to those presented by the National Sleep Foundation , and relative frequencies were calculated per group. Categorical variables (i.e., sleep education, sleep health practices and knowledge about sleep hygiene rules) are displayed as percentage frequency distributions. To determine the relationship between categorical data, relative frequencies were compared using Fisher's exact tests. The critical value for significance was set at 0.05, and a correction for multiple comparisons was applied during post hoc analyses.
Study design A cross‐sectional quantitative survey study was conducted. The study was deemed exempt from review under the Dutch Medical Research Involving Human Subjects Act by the Medical Ethics Research Committee of the University Medical Centre Utrecht, the Netherlands (file number 19‐066).
Respondents HCPs from three paediatric rehabilitation settings (rehabilitation centre, school for special education and rehabilitation department of a children's hospital) in the Netherlands participated in this study. In Dutch rehabilitation, a medically schooled physician serves a gatekeeping role in detecting child‐related problems during clinical encounters and consequently coordinates the child's rehabilitative care. When, in this case, a problem with sleep of the child is detected, the physician gives first‐line treatment advice or can decide to set up referral to a non‐medical professional or sleep clinic for subsequent sleep assessments and/or interventions. Depending on the nature of the sleep problem, a non‐medical professional may further assess the child's sleep, bed routine and/or behaviour and perform therapies like behavioural interventions and implementing healthy sleep practices. In this way, the roles of HCPs involved in sleep care in Dutch rehabilitation settings are distinct. The pen‐and‐paper surveys were distributed during live meetings among the following two groups of paediatric HCPs: Rehabilitation physicians (RPs). This group included paediatric rehabilitation physicians, physician assistants, paediatricians and doctors in specialist training to become RP. Allied health professionals (AHPs). This group included physical therapists, occupational therapists, developmental behavioural therapists, speech and language therapists, social workers and psychologists. An additional control group was drawn from the general population, comprising individuals without a background or current profession in healthcare, to allow comparisons of sleep knowledge levels. The control group consisted mostly of parents of (young) children, who were recruited via the social networks of colleagues and acquaintances, and filled out the pen‐and‐paper surveys during face‐to‐face encounters.
Data collection A 30‐item structured questionnaire was designed based on relevant literature and consultations with experts (researchers and clinicians) working in the field of paediatric rehabilitation and sleep medicine . Pilot testing was conducted to ensure that the questionnaire was easily understood and could be completed within a short time window (15–20 min). The questionnaire comprised three sections: general information including age, sex, profession, educational level and hours of sleep education received; application of sleep health practices in daily clinical practice. HCPs were asked to indicate how often they address the topic sleep during clinical encounters, choosing between never/seldom (less than once per month), sometimes (1–3 times per month) or often (once per week or more often). HCPs who reported to address sleep sometimes or often in clinical practice were asked to indicate the type(s) of sleep interventions or therapies they apply. sleep knowledge within the domains of (a) basic sleep physiology (i.e., recommended sleep durations and sleep architecture), (b) symptoms and characteristics of common paediatric sleep disorders and (c) sleep hygiene rules (i.e., healthy sleep practices). Additionally, the participants' self‐perceived knowledge sufficiency was assessed. Apart from the general information, responses were measured using multiple choice questions and included a ‘ don ' t know ’ answer option. An open‐ended question was used for collecting information about knowledge of sleep hygiene rules; respondents were asked to name three sleep hygiene rules other than the example given regarding limiting screen time 2 h before bedtime. A cover letter that explained the aim of the study and emphasized the need for honest responses (i.e., to answer with ‘ don't know ’ instead of guessing if one does not know) was attached to the survey.
Data analysis Surveys that returned largely incomplete (>25%) were excluded from analysis. Data were analysed using IBM SPSS Statistics 26. Sleep knowledge scores (i.e., number of correctly answered questions) were calculated for each group and converted into percentages; these are presented as mean ± standard deviation (SD)% scores for all knowledge questions in total and per domain sleep physiology and sleep disorders. After testing for normality, means were compared using Kruskal–Wallis test followed by post hoc Mann–Whitney tests. Answers to the open‐ended question regarding sleep hygiene rules were categorized according to those presented by the National Sleep Foundation , and relative frequencies were calculated per group. Categorical variables (i.e., sleep education, sleep health practices and knowledge about sleep hygiene rules) are displayed as percentage frequency distributions. To determine the relationship between categorical data, relative frequencies were compared using Fisher's exact tests. The critical value for significance was set at 0.05, and a correction for multiple comparisons was applied during post hoc analyses.
RESULTS 3.1 Respondents' general information In total, 84 HCPs completed the survey. Based on their profession, HCPs were divided between the RP group ( n = 30) and the AHP group ( n = 54). An additional control group ( n = 63) completed the survey questions with exception of the section regarding sleep health practices. The majority of all respondents were female and between the age of 31–50 years. Across all groups, over 75% indicated to have received less than 5 h of sleep education throughout their entire school curricula (RPs 75.9%; AHPs 90.6%; controls 88.5%), and this amount was independent of group ( p = 0.102, Fisher's exact test). Group characteristics are summarized in Table . 3.2 Sleep health practices RPs ( often 56.7%, n = 17; sometimes 40%, n = 12; never/seldom 3.3%, n = 1) reported to more frequently address sleep issues than AHPs ( often 11.8%, n = 6; sometimes 51%, n = 26; never/seldom 37.2%, n = 19), a difference found to be statistically significant ( p < 0.001*, Fisher's exact test). 3.3 Sleep interventions Those who reported to sometimes or often address sleep indicated the types of sleep interventions they apply (Figure ). The majority of both HCPs groups (RPs 97%, n = 29; AHPs 79%, n = 38) reported to give advice about sleep hygiene rules. RPs more often mentioned giving advice compared with AHPs ( p = 0.043*, Fisher's exact test). Half of RPs (50%, n = 15) indicated to prescribe medication (including melatonin) compared with 6% ( n = 3) of AHPs ( p < 0.001*, Fisher's exact test). One third of RPs (33%, n = 10) indicated to refer to a sleep clinic, compared with 6% ( n = 3) of AHPs ( p = 0.004*, Fisher's exact test). Less than a quarter of HCPs (RPs 23%, n = 7; AHPs 13%, n = 6) reported to perform behavioural therapy to treat sleep problems ( p = 0.229, Fisher's exact test). None of the HCPs reported to perform light therapy. 3.4 Sleep knowledge 3.4.1 Sleep physiology and sleep disorders Table shows the sleep knowledge scores (correct answer rates) across groups. Mean total sleep knowledge scores were found to be statistically different across groups ( H (2) = 24.322, p < 0.001*), with RPs demonstrating significantly higher scores than AHPs ( p < 0.001*) and controls ( p < 0.001*), whereas no difference was found between AHPs and controls ( p = 0.667). To allow subgroup analyses between different domains of sleep knowledge, total sleep knowledge scores were divided between questions covering the domains sleep physiology and sleep disorders. All three groups demonstrated lower scores on questions about symptoms and characteristics of sleep disorders, compared with questions related to basic sleep physiology (Table ). Subgroup analysis revealed different scores across groups on both domains, with RPs scoring significantly higher than AHPs (sleep physiology p < 0.001*; sleep disorders p < 0.001*) and controls (sleep physiology p = 0.002*; sleep disorders p < 0.001*). Controls showed higher scores than AHPs within the domain sleep physiology, but this difference failed to reach statistical significance after correcting for alpha ( p = 0.034). Within the domain sleep disorders, AHPs demonstrated significantly higher scores than controls ( p = 0.013*). 3.4.2 Sleep hygiene The frequency of sleep hygiene rules mentioned by RPs, AHPs and controls is shown in Figure . RPs were able to more frequently recall the majority of sleep hygiene rules. Significant differences were observed between groups for establishing a relaxing bedtime routine ( p = 0.015*, Fisher's exact test) and establishing a regular sleep schedule ( p = 0.026*, Fisher's exact test), with these rules being mentioned more often by RPs and AHPs compared with controls, respectively. The sleep hygiene rules closest to bedtime and related to the sleep environment were best known across groups, whereas those related to daytime practices (i.e., promoting physical exercise during the day , exposure to natural light and limiting daytime naps ) were rarely mentioned across all groups. 3.4.3 Self‐perceived sleep knowledge The minority of all HCPs (RPs 20%, n = 6; AHPs 14.8%, n = 8) reported to believe that they possess sufficient sleep knowledge to address sleep problems in daily clinical practice. In contrast, most HCPS reported either to not have sufficient knowledge (RPs 40%, n = 12; AHPs 42.6%, n = 23) or that they ‘ do not know ’ (RPs 40%, n = 12; AHPs 38.9%, n = 21) (Table ). No difference was found in total sleep knowledge scores between HCPs who believed they had sufficient knowledge about sleep versus those who believed they did not have sufficient sleep knowledge, RPs: t (16) = 0.652, p = 0.524; AHPs: t (29) = 0.196, p = 0.846.
Respondents' general information In total, 84 HCPs completed the survey. Based on their profession, HCPs were divided between the RP group ( n = 30) and the AHP group ( n = 54). An additional control group ( n = 63) completed the survey questions with exception of the section regarding sleep health practices. The majority of all respondents were female and between the age of 31–50 years. Across all groups, over 75% indicated to have received less than 5 h of sleep education throughout their entire school curricula (RPs 75.9%; AHPs 90.6%; controls 88.5%), and this amount was independent of group ( p = 0.102, Fisher's exact test). Group characteristics are summarized in Table .
Sleep health practices RPs ( often 56.7%, n = 17; sometimes 40%, n = 12; never/seldom 3.3%, n = 1) reported to more frequently address sleep issues than AHPs ( often 11.8%, n = 6; sometimes 51%, n = 26; never/seldom 37.2%, n = 19), a difference found to be statistically significant ( p < 0.001*, Fisher's exact test).
Sleep interventions Those who reported to sometimes or often address sleep indicated the types of sleep interventions they apply (Figure ). The majority of both HCPs groups (RPs 97%, n = 29; AHPs 79%, n = 38) reported to give advice about sleep hygiene rules. RPs more often mentioned giving advice compared with AHPs ( p = 0.043*, Fisher's exact test). Half of RPs (50%, n = 15) indicated to prescribe medication (including melatonin) compared with 6% ( n = 3) of AHPs ( p < 0.001*, Fisher's exact test). One third of RPs (33%, n = 10) indicated to refer to a sleep clinic, compared with 6% ( n = 3) of AHPs ( p = 0.004*, Fisher's exact test). Less than a quarter of HCPs (RPs 23%, n = 7; AHPs 13%, n = 6) reported to perform behavioural therapy to treat sleep problems ( p = 0.229, Fisher's exact test). None of the HCPs reported to perform light therapy.
Sleep knowledge 3.4.1 Sleep physiology and sleep disorders Table shows the sleep knowledge scores (correct answer rates) across groups. Mean total sleep knowledge scores were found to be statistically different across groups ( H (2) = 24.322, p < 0.001*), with RPs demonstrating significantly higher scores than AHPs ( p < 0.001*) and controls ( p < 0.001*), whereas no difference was found between AHPs and controls ( p = 0.667). To allow subgroup analyses between different domains of sleep knowledge, total sleep knowledge scores were divided between questions covering the domains sleep physiology and sleep disorders. All three groups demonstrated lower scores on questions about symptoms and characteristics of sleep disorders, compared with questions related to basic sleep physiology (Table ). Subgroup analysis revealed different scores across groups on both domains, with RPs scoring significantly higher than AHPs (sleep physiology p < 0.001*; sleep disorders p < 0.001*) and controls (sleep physiology p = 0.002*; sleep disorders p < 0.001*). Controls showed higher scores than AHPs within the domain sleep physiology, but this difference failed to reach statistical significance after correcting for alpha ( p = 0.034). Within the domain sleep disorders, AHPs demonstrated significantly higher scores than controls ( p = 0.013*). 3.4.2 Sleep hygiene The frequency of sleep hygiene rules mentioned by RPs, AHPs and controls is shown in Figure . RPs were able to more frequently recall the majority of sleep hygiene rules. Significant differences were observed between groups for establishing a relaxing bedtime routine ( p = 0.015*, Fisher's exact test) and establishing a regular sleep schedule ( p = 0.026*, Fisher's exact test), with these rules being mentioned more often by RPs and AHPs compared with controls, respectively. The sleep hygiene rules closest to bedtime and related to the sleep environment were best known across groups, whereas those related to daytime practices (i.e., promoting physical exercise during the day , exposure to natural light and limiting daytime naps ) were rarely mentioned across all groups. 3.4.3 Self‐perceived sleep knowledge The minority of all HCPs (RPs 20%, n = 6; AHPs 14.8%, n = 8) reported to believe that they possess sufficient sleep knowledge to address sleep problems in daily clinical practice. In contrast, most HCPS reported either to not have sufficient knowledge (RPs 40%, n = 12; AHPs 42.6%, n = 23) or that they ‘ do not know ’ (RPs 40%, n = 12; AHPs 38.9%, n = 21) (Table ). No difference was found in total sleep knowledge scores between HCPs who believed they had sufficient knowledge about sleep versus those who believed they did not have sufficient sleep knowledge, RPs: t (16) = 0.652, p = 0.524; AHPs: t (29) = 0.196, p = 0.846.
Sleep physiology and sleep disorders Table shows the sleep knowledge scores (correct answer rates) across groups. Mean total sleep knowledge scores were found to be statistically different across groups ( H (2) = 24.322, p < 0.001*), with RPs demonstrating significantly higher scores than AHPs ( p < 0.001*) and controls ( p < 0.001*), whereas no difference was found between AHPs and controls ( p = 0.667). To allow subgroup analyses between different domains of sleep knowledge, total sleep knowledge scores were divided between questions covering the domains sleep physiology and sleep disorders. All three groups demonstrated lower scores on questions about symptoms and characteristics of sleep disorders, compared with questions related to basic sleep physiology (Table ). Subgroup analysis revealed different scores across groups on both domains, with RPs scoring significantly higher than AHPs (sleep physiology p < 0.001*; sleep disorders p < 0.001*) and controls (sleep physiology p = 0.002*; sleep disorders p < 0.001*). Controls showed higher scores than AHPs within the domain sleep physiology, but this difference failed to reach statistical significance after correcting for alpha ( p = 0.034). Within the domain sleep disorders, AHPs demonstrated significantly higher scores than controls ( p = 0.013*).
Sleep hygiene The frequency of sleep hygiene rules mentioned by RPs, AHPs and controls is shown in Figure . RPs were able to more frequently recall the majority of sleep hygiene rules. Significant differences were observed between groups for establishing a relaxing bedtime routine ( p = 0.015*, Fisher's exact test) and establishing a regular sleep schedule ( p = 0.026*, Fisher's exact test), with these rules being mentioned more often by RPs and AHPs compared with controls, respectively. The sleep hygiene rules closest to bedtime and related to the sleep environment were best known across groups, whereas those related to daytime practices (i.e., promoting physical exercise during the day , exposure to natural light and limiting daytime naps ) were rarely mentioned across all groups.
Self‐perceived sleep knowledge The minority of all HCPs (RPs 20%, n = 6; AHPs 14.8%, n = 8) reported to believe that they possess sufficient sleep knowledge to address sleep problems in daily clinical practice. In contrast, most HCPS reported either to not have sufficient knowledge (RPs 40%, n = 12; AHPs 42.6%, n = 23) or that they ‘ do not know ’ (RPs 40%, n = 12; AHPs 38.9%, n = 21) (Table ). No difference was found in total sleep knowledge scores between HCPs who believed they had sufficient knowledge about sleep versus those who believed they did not have sufficient sleep knowledge, RPs: t (16) = 0.652, p = 0.524; AHPs: t (29) = 0.196, p = 0.846.
DISCUSSION This study assessed the sleep health practices and knowledge about sleep physiology, sleep disorders and sleep hygiene among two groups of HCPs in Dutch paediatric rehabilitation. The frequency of sleep being addressed during clinical encounters varies greatly between medical and non‐medical HCPs, and more efforts should be made for sleep assessments to become a standard item for surveillance during routine healthcare practices. Although RPs showed higher sleep knowledge scores than AHPs, neither group exceeded 50% correct scores, suggesting limited sleep knowledge, particularly in the area of sleep disorders. We also noticed limited familiarity with healthy sleep behaviours that can be practiced during the day. Our findings emphasize the need to educate and empower HCPs with sound knowledge, skills and confidence required to address sleep problems in children with NDDs and to support their parents. There are several strengths and limitations to this study that should be considered. The questionnaire used was developed in co‐creation with researchers and clinicians in the field of paediatric rehabilitation and sleep medicine, based on currently available literature in these fields, and trialled before use, yet it should be noted that it has not been validated. HCPs may have felt obliged to respond favourably to questions concerning their sleep health practices. Also, the relatively small sample size may limit the generalisability of our findings. Unlike other survey studies on sleep knowledge levels, we did include a control group to allow comparisons of HCP scores to the general population. In line with our results, survey studies of practicing physicians and paediatricians have consistently found poor knowledge about sleep and significant gaps in clinical practices regarding paediatric sleep disorders (BaHammam, ; Bruni et al., ; Gruber, Constantin, Frappier, Brouillette, & Wise, ; Owens, ; Papp, Penrod, & Strohl, ). For example, Bruni et al. found low scores in all areas of sleep knowledge and particularly in sleep disorders among paediatricians and child neuropsychiatrists (Bruni et al., ). Papp and colleagues reported an average knowledge of 34% among primary care physicians, with only 10% rating themselves as good (Papp et al., ). Similarly, we found the self‐perceived knowledge among paediatric rehabilitation professionals to be low, that is, only one in five RPs, and one in seven AHPs rated their own sleep knowledge to be sufficient. Without proper training and experience, HCPs may lack confidence or feel incompetent to address sleep problems properly, resulting in sleep problems left unaddressed and untreated (Honaker & Meltzer, ). As more HCPs acquire a greater awareness of sleep, more consistent processes for screening and assessment can be developed across paediatric rehabilitation settings. Given that parental knowledge about children's sleep is typically poor (McDowall, Galland, Campbell, & Elder, ), they should be provided with appropriate information and advices to ensure that healthy sleep practices are implemented and maintained at home (Blackmer & Feinstein, ; Mindell & Owens, ). Nearly all HCPs reported to give such advices on a frequent basis, yet their knowledge deficits are indicated by equally low total sleep knowledge scores as our control sample, which consisted mostly of parents. A recent study on sleep problems and solution seeking for children with cerebral palsy and their parents reported that out of the 63 parents that asked for professional help with their child's sleep, only 21 reported that their request for help led to effective treatment or advice from their HCP (Petersen et al., ). In addition, we found HCPs' familiarity with sleep hygiene rules to be confined to those closest to bedtime and related to the sleep environment, whereas they appeared unfamiliar to daytime practices (like exposure to daylight, adequate exercise and limiting daytime naps). This knowledge gap is worrisome since sleep hygiene is considered the first line of treatment for sleep problems in children with NDDs (Jan et al., ). A review of the lifestyle practices that contribute to good quality sleep can be valuable in providing a starting point to improve sleep and more broadly in adopting healthy and protective lifestyles. In fact, promoting the entire triad of healthy behaviours, which in addition to physical activity and nutrition, also includes sleep itself, has recently been described as ‘the formula for health and well‐being’ in vulnerable patient populations with neurodevelopmental (Verschuren, McPhee, Rosenbaum, & Gorter, ) and neuropsychiatric disorders (Briguglio et al., ). Clearly, and in alignment with the global medical trend towards preventive healthcare (Egger, Binns, Rossner, & Sagner, ), the need to protect, promote and maintain healthy sleep as part of a healthy lifestyle, especially in children with NDDs, is evident. But if we want doctors to preach a healthy lifestyle to their patients (and families), expect them to detect and prevent sleep problems early on and require them to effectively guide parents, obviously they need to be adequately equipped with proper training before they enter the clinic. Unfortunately, there is only very limited coverage of sleep education in medical schools, which has previously been identified in 409 medical schools across 12 countries (Mindell et al., ), and appears to persist. Consistent with their findings from nearly a decade ago, we found that the majority of Dutch physicians received less than 5 h of sleep education, similar to non‐medical professionals and controls. This alarmingly low number may in turn explain their limited sleep knowledge and feelings of incompetence to address sleep in clinical practice. Indeed, limited exposure to sleep education can predict medical trainees' confidence and knowledge levels, thereby forecasting their future clinical practices regarding sleep health (R. E. Salas et al., ). This advocates the continued need for sleep medicine education to be fully incorporated into medical school curricula (R. M. E. Salas et al., ). Taking into account the confines of an already packed medical curriculum, additional educational efforts like postgraduate training, clinical opportunities and other sleep education tools for current HCPs are warranted. It has been shown that sleep knowledge can be successfully increased through provision of sleep education, both when delivered face‐to‐face and through online webinars (Osborne & Blunden, ). Our results showed that the knowledge scores of HCPs who believed to possess sufficient sleep knowledge did not differ from those who believed they lacked sleep knowledge, demonstrating the importance to undertake such sleep training programs regardless of self‐perceived knowledge. The goal of educational sleep trainings should not be to turn HCPs into sleep experts but rather to provide them with the knowledge required to recognize symptoms of major paediatric sleep disorders by asking the right questions, to give the right general advice regarding good sleep hygiene and to enable them to know when to refer to a (sleep) specialist for further assessment or to initiate sleep treatment strategies (Luginbuehl & Kohler, ).
R. Y. H. made substantial contributions to conception and design of the study and acquisition and analysis of data and interpretation of data. She wrote the manuscript and approved of the final version to be published. S. P. was involved in the design of the study and data acquisition and analysis and made significant contributions to the manuscript. J. M. V. was involved in the data analysis and interpretation of the data and contributed to revision of the manuscript. N. R. made great contributions to data entering, data analysis and interpretation and was involved in drafting the article. J. M. A. V. contributed to conceptualization and interpretation of data and revising the manuscript for important intellectual content. O. V. was involved in the concept and design of the study, data analysis interpretation of data and contributed to drafting and revising the article. All persons who meet authorship criteria are listed as authors, and all authors certify that they have participated sufficiently in the work to take public responsibility for the content, including participation in the concept, design, analysis, writing or revision of the manuscript. All authors reviewed the text of the manuscript and gave final approval of the version to be published.
The authors have no conflicts of interest to declare.
Supporting Information S1 Click here for additional data file.
|
Integrated care for optimizing the management of stroke and associated heart disease: a position paper of the European Society of Cardiology Council on Stroke | 43a354f6-8c5e-4f9d-9c7a-7e59e9993cf9 | 9259378 | Internal Medicine[mh] | The management of patients with stroke is often multidisciplinary requiring input from stroke specialists (doctors and nurses) as well as from internal medicine, neurology, radiology, emergency medicine, primary care, and rehabilitation team (including physiotherapist, speech, and occupational therapists). Given the common shared risk factors for stroke and cardiovascular disease, input may also be required from a cardiologist, vascular interventionist, vascular surgeon, and neurosurgeon, as well as nurses, patient caregivers, and next-of-kin. Ultimately, the patient is central to all this, requiring a coordinated and uniform approach to the priorities of post-stroke management, which can be consistently implemented by different multidisciplinary healthcare professionals, as part of the patient ‘journey’ or ‘patient pathway,’ supported by appropriate education and tele-medicine approaches. All these aspects would ultimately aid delivery of care and improve patient (and caregiver) engagement and empowerment. Given the need to address the multidisciplinary approach to holistic or integrated care of patients with heart disease and stroke, the European Society of Cardiology Council on Stroke convened a Task Force, with the remit to review the published evidence and to propose a consensus on Integrated care management for optimizing the management of stroke and associated heart disease. The present position paper summarizes the available evidence and proposes consensus statements that may help to define evidence gaps and simple practical approaches to assist in everyday clinical practice. Nevertheless, the ultimate judgment regarding care of each individual patient must be made by the healthcare providers and the patient and their family together, considering all the distinct circumstances presented by that patient. Literature searches were conducted in the following databases: PubMed/MEDLINE and the Cochrane Library (including the Cochrane Database of Systematic Reviews and the Cochrane Controlled Trials Registry). Searches focused on English-language sources and studies in human subjects. Articles related to animal experimentation were only cited when the information was important to understanding pathophysiological concepts pertinent to patient management and comparable data were not available from human studies. Additional information was requested from the authors where necessary.
The co-occurring and inter-linked nature of cardiac and cerebrovascular disease (together termed CVD) requires a combined action plan to prevent, identify, treat, and rehabilitate people. To deliver this requires a multi-faceted action on several fronts, involving both the health and social care workforce. It is axiomatic that increasing the collaboration and integration of the cardiac- and stroke-specialist workforces, within an integrated care service model will ensure effective and efficient. Indeed, there needs to be concerted and co-ordinated action to address the underlying risk factors for CVD, requiring greater investment in, and implementation of, disease prevention and health promotion policies. The importance of an ‘integrated care’ approach has been applied to other chronic conditions. , In patients with atrial fibrillation (AF), which is a common cause for ischaemic stroke (IS), the ABC (Atrial fibrillation Better Care) pathway has been proposed as an integrated care approach with three central pillars: ‘ A ’ Avoid stroke (with Anticoagulants); ‘ B ’ Better symptom management, with patient-centred decisions on rate or rhythm control; and ‘ C ’ Cardiovascular and Comorbidity risk optimization. This provides a streamlined approach to management that is applicable to whether the AF patient is managed by any healthcare professional, either the general practitioner or the hospital-based specialist (whether cardiologist or non-cardiologist), minimizing the possibility of conflicting information from healthcare professionals. Indeed, inconsistent information to patients has been associated with poorer patient adherence with their management plan. An integrated care approach is increasingly recommended in guidelines. , The ABC pathway when applied to AF patients has been well-validated in post-hoc analyses of clinical trials, prospective cohort studies and a prospective randomized trial. In a systematic review, AF patients adherent with the ABC pathway showed a lower risk of all-cause death [odds ratio (OR): 0.42, 95% CI 0.31–0.56], cardiovascular death (OR: 0.37, 95% CI 0.23–0.58), stroke (OR: 0.55, 95% CI 0.37–0.82) and major bleeding (OR: 0.69, 95% CI 0.51–0.94). Improved clinical outcomes with ABC pathway adherence are evident, even in clinically complex patients such as those with multimorbidity, polypharmacy, and repeat hospitalizations. In the prospective cluster randomized mobile AF application (mAFA)-II trial, rates of the composite outcome of ‘IS/systemic thromboembolism, death, and rehospitalization’ were lower with the mAFA intervention plus ABC pathway adherent care compared with usual care [1.9 vs. 6.0%; hazard ratio [HR]: 0.39; 95% CI 0.22 to 0.67; P < 0.001]. Dynamic bleeding risk monitoring and follow-up reassessments using hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly (> 65 years), drugs/alcohol concomitantly score resulted in lower risks of major bleeding (mAFA vs. usual care, 2.1 vs. 4.3% at one year) and increased total oral anticoagulation usage from 63 to 70%. In the mAFA-II trial long-term extension cohort, the beneficial effects were maintained, with a high adherence (>70%) and persistence (>90%) with the ABC pathway using an App-based intervention. Between 20 and 30% of all strokes are recurrent strokes. An integrated care approach to stroke care can reduce the likelihood of CVD events, but should they occur, it is vital that evidence-based care pathways are in place, with equitable care available for everyone, at any timepoint. These pathways need to provide seamless, integrated, and individualized care and be delivered by a workforce that have the right knowledge, skills, and behaviours, depending on their role along the pathway. All such efforts then need to undergo rigorous evaluation and adaption in order to spend healthcare resources efficiently. , In the UK, the knowledge and skills for stroke-specific care have been clearly laid out in the Stroke-Specific Education Framework (stroke-education.org.uk), which includes 16 Elements of Care from awareness raising of stroke symptoms, through to end of life care, rehabilitation, and return to work. Whilst these have been developed for stroke specialist and stroke relevant staff, many of the knowledge, skills, and behaviours apply equally to the cardiac workforce and care pathways. The need for such standardized and integrated post stroke programmes and follow-up care after rehabilitation seems evident, but programmes are not widely available. Such programmes will help improve the quality of care, as well as help avoid unnecessary variations in patient care, and in patient outcomes as well. They also inform quality standards and allow benchmarking of services and care against recommended standards. In UK stroke care, the standards have informed the Sentinel Stroke National Audit Package (SSNAP; https://www.strokeaudit.org ) which has underpinned developments in many aspects of stroke care since its inception as the National Sentinel Audit in 1997. The action plan for stroke in Europe also emphasizes the integrated stroke and secondary prevention care. Similarly, quality improvements in cardiac and stroke services have been achieved through audit programmes, focussing on cardiac surgery, percutaneous coronary interventions, rhythm management, heart failure, myocardial ischaemia, and congenital heart disease. Unfortunately, they are often focussed on acute care and less on rehabilitation and secondary prevention. Importantly, many of these domains have commonalities with stroke care. For example, heart rhythm issues such as AF significantly increase the risk of first stroke, and if they remain untreated after an acute event increase the risk of stroke recurrence. Such cardiac diseases not only require acute diagnosis and treatment, but rehabilitation and lifelong management. Thus, closely linked cardiovascular diseases may have overlapping disease management programmes encompassing the entire chain of care. Overall, it is imperative that as we move forwards with integrated stroke, and integrated cardiac, care, we also consider how cardiac and stroke care can become integrated to provide CVD integrated care. Some preliminary data already suggest improvements in functional status with a multidisciplinary approach to care after stroke. An integrated care approach used for AF can be applied to stroke, given the shared common cardiovascular risk factors including advancing age, male sex, hypertension, diabetes mellitus (DM), valvular heart disease, heart failure, coronary heart disease, chronic kidney disease, inflammatory disorders, sleep apnoea, and tobacco use. An integrated or holistic management pathway for patients following a stroke, that not only targets the prevention of recurrent stroke, but also improves patient functional status and symptoms, and manages cardiovascular risk factors, comorbidities and lifestyle changes can be proposed; this is the focus of this consensus document. A post-stroke ABC pathway as a more holistic approach to integrated stroke care would include three pillars of management ( ): A: Appropriate Antithrombotic therapy. B: Better functional and psychological status. C: Cardiovascular risk factors and Comorbidity optimization (including lifestyle changes). This approach expands on a clinical concept proposed in 2021 as ‘integrated care for stroke medicine—easy as ABC’.
The ‘A’ criterion (‘Appropriate Antithrombotic therapy’) in the post-stroke patient refers to the use of oral anticoagulants (OACs), either as a non-vitamin K antagonist OAC (also called a direct OAC or DOAC) or well-managed vitamin K antagonist (VKA, e.g. warfarin) with a time in therapeutic range ≥70% when AF is present or a VKA, if a prosthetic mechanical heart valve is present. Where there is associated atherosclerotic vascular disease and no AF or mechanical valve present, the appropriate use of antiplatelet therapy is needed. A balance is needed between preventing recurrent ischaemic or thrombotic events and major bleeding, , which is even more challenging if both AF and vascular disease, whether coronary, carotid, or peripheral artery disease, are present. In such patients (ie. AF with stable vascular disease), anticoagulation monotherapy would suffice as thromboprphylaxis, although many physicians would still prescribe combination therapy, despite the paucity of evidence from large randomized trials. In a systematic review and meta-analysis, for example, the pooled prevalence of carotid stenosis in AF patients was 12.4% (95% CI 8.7 to 16.0), ranging from 4.4 to 24.3%. In AF patients with coronary artery disease, appropriate antithrombotic therapy management varies according the clinical scenario. When the AF patient has ‘stable vascular disease’, the patient should be managed with OAC monotherapy. In the AFIRE trial, combination therapy with OAC plus antiplatelets was associated with worse thromboembolic and bleeding outcomes in AF patients with stable coronary disease. In AF patients presenting with an acute coronary syndrome (ACS), a balance between AF-related stroke prevention which requires OAC and reducing cardiac ischaemia in an ACS presentation which requires antiplatelet therapy needs to be reached; to minimize the risk of stent thrombosis after a percutaneous coronary intervention and the risk of bleeding by the combination of OAC with antiplatelet therapy. In patients with asymptomatic high-grade carotid artery stenosis and AF, carotid endarterectomy is commonly considered , or, in those less suitable for surgery, carotid artery stenting may be chosen notwithstanding the need for a short course of combined antithrombotic therapy after stenting. While more data are needed to inform optimal antithrombotic management in this setting, a recent nationwide observational evidence suggested that OAC therapy alone could be a default treatment for most of these patients, combined with a short period single antiplatelet therapy in those at high risk of a recurrent vascular event. In stable atherosclerotic vascular disease patients without AF at increased risk of ischaemic events, combination therapy with rivaroxaban 2.5 mg bid and aspirin provides some benefits on CVD events (including on stroke) even in the absence of associated AF , but at the risk of more major bleeding. In the cardiovascular outcomes for people using anticoagulation strategies trial, the primary composite outcome [cardiovascular death, stroke, or myocardial infarction (MI)] was lower in the low-dose rivaroxaban-plus-aspirin group compared with the aspirin-alone group (HR 0.76; 95% CI 0.66 to 0.86), at the cost of more major bleeding events (HR 1.70; 95% CI, 1.40 to 2.05). In the rivaroxaban-plus-aspirin group, mortality was also lower (HR 0.82; 95% CI, 0.71 to 0.96), as was IS (HR 0.51; 95% CI 0.38–0.68) compared with the aspirin-alone group. Proactive mitigation of bleeding risks should be directed at the modifiable bleeding risk factors (uncontrolled blood pressure, reducing alcohol excess, etc.) and scheduling the high bleeding risk patients for early review and follow-up. In the prospective cluster randomized mAFA-II trial of general AF patients, this approach resulted in a reduction in major bleeding at 1 year follow-up and an increase in OAC use. Antiplatelet therapy The standard, and guideline, recommended approach for the use of antiplatelet drugs for secondary stroke prevention for non-cardioembolic stroke has been with aspirin and clopidogrel or ticagrelor. The combination of aspirin and extended-release dipyridamole has also been recommended but has fallen out of favour after the results of the prevention regimen for effectively avoiding second strokes trial demonstrated no difference as compared to clopidogrel for IS outcomes and a significantly higher risk for intracranial haemorrhage and discontinuation because of headaches. The dose of aspirin recommended ranges from 50–325 mgs daily, but most practitioners use a lower dose given that there is no significant difference in efficacy between lower and higher doses and higher doses are associated with more gastrointestinal side effects. The combination of aspirin plus clopidogrel or ticagrelor vs. aspirin alone was evaluated in three, large, relatively recent randomized clinical trials. The first was the Clopidogrel in High-Risk Patients with Non-disabling Cerebrovascular Events (CHANCE) that evaluated the combination of clopidogrel plus aspirin for 21 days after a 300 mg loading dose of clopidogrel, followed by clopidogrel monotherapy up to day 90 vs. aspirin alone in a Chinese population with mild stroke [National Institutes of Health Stroke Scale (NIHSS) < 3] or high-risk transient ischaemic attack (TIA) with an ABCD2 score ≥ 4. The early combination therapy group had significantly fewer primary outcome events without an increased for major or severe bleeding. Interestingly, in this clinical trial patients with a CYP219C allele, that identified them as a slow metabolizer of clopidogrel to its active prodrug, did not benefit from the combination antiplatelet therapy, while those without it had significantly fewer primary outcome events. The Platelet-Oriented Inhibition in New TIA and Minor Ischaemic Stroke (POINT) trial evaluated the clopidogrel-aspirin combination vs. aspirin alone in a heterogeneous population of minor stroke (NIHSS < 3) and high-risk TIA patients with an ABCD2 score ≥ 4. In the POINT trial patients were given a 600-mg loading dose of clopidogrel followed by 75 mgs daily from day 2 to day 90 plus aspirin vs. aspirin alone. The results demonstrated a significant reduction in the primary outcome of IS, MI and ischaemic vascular death in the combination group, but this was associated with a significantly increased risk of major haemorrhage. In the POINT trial, presence of a clopidogrel slow metabolizer genetic variant was not associated with a significant reduction in efficacy, but there was a trend towards lower efficacy in these patients, although this did not reach significance due to the small sample size. Following the results of these two trials, the combination of aspirin and clopidogrel is now recommended for patients with recent minor (NIHSS score ≤3) noncardioembolic IS or high-risk TIA (ABCD2 score ≥4). It should be started early (ideally within 12–24 h of symptom onset and at least within 7 days of onset) and continued for 21–90 days. Then, it should be followed by single antiplatelet treatment. A recent trial, The Acute Stroke or Transient Ischaemic Attack Treated with Ticagrelor and ASA for Prevention of Stroke and Death trial (THALES) compared treatment with ticagrelor plus aspirin to aspirin monotherapy for 30 days in patients with a mild to moderate stroke, NIHSS < 5 or high-risk TIA with an ABCD2 score ≥ 6. The primary outcome of stroke both ischaemic and haemorrhagic as well as the secondary outcome of IS was significantly reduced in the combination therapy arm. The combination therapy group had an increased risk of severe and intracranial/fatal bleeding. A subgroup analysis of THALES demonstrated that the benefit of the combination therapy only occurred in patients with documented large artery stenosis of 30% or greater and was not associated with an increased risk of bleeding. In the CHANCE-2 trial, among patients with prior stroke/high-risk TIA and CYP2C19 loss-of-function carriers, the use of ticagrelor plus aspirin was associated with significant modest reduction in 90-day risk of subsequent stroke compared with clopidogrel plus aspirin. The Food and Drug Administration in the United States recently approved the combination of ticagrelor plus aspirin for secondary stroke prevention. In some Asian countries, another drug with antiplatelet activity, cilostazol, is approved for use and widely employed. Indeed, a recent trial in Japan, combining cilostazol with aspirin or clopidogrel, demonstrated that the combination was superior to aspirin or clopidogrel alone without an increased risk for major bleeding. However, cilostazol is not approved for stroke prevention in Europe or the United States.
The standard, and guideline, recommended approach for the use of antiplatelet drugs for secondary stroke prevention for non-cardioembolic stroke has been with aspirin and clopidogrel or ticagrelor. The combination of aspirin and extended-release dipyridamole has also been recommended but has fallen out of favour after the results of the prevention regimen for effectively avoiding second strokes trial demonstrated no difference as compared to clopidogrel for IS outcomes and a significantly higher risk for intracranial haemorrhage and discontinuation because of headaches. The dose of aspirin recommended ranges from 50–325 mgs daily, but most practitioners use a lower dose given that there is no significant difference in efficacy between lower and higher doses and higher doses are associated with more gastrointestinal side effects. The combination of aspirin plus clopidogrel or ticagrelor vs. aspirin alone was evaluated in three, large, relatively recent randomized clinical trials. The first was the Clopidogrel in High-Risk Patients with Non-disabling Cerebrovascular Events (CHANCE) that evaluated the combination of clopidogrel plus aspirin for 21 days after a 300 mg loading dose of clopidogrel, followed by clopidogrel monotherapy up to day 90 vs. aspirin alone in a Chinese population with mild stroke [National Institutes of Health Stroke Scale (NIHSS) < 3] or high-risk transient ischaemic attack (TIA) with an ABCD2 score ≥ 4. The early combination therapy group had significantly fewer primary outcome events without an increased for major or severe bleeding. Interestingly, in this clinical trial patients with a CYP219C allele, that identified them as a slow metabolizer of clopidogrel to its active prodrug, did not benefit from the combination antiplatelet therapy, while those without it had significantly fewer primary outcome events. The Platelet-Oriented Inhibition in New TIA and Minor Ischaemic Stroke (POINT) trial evaluated the clopidogrel-aspirin combination vs. aspirin alone in a heterogeneous population of minor stroke (NIHSS < 3) and high-risk TIA patients with an ABCD2 score ≥ 4. In the POINT trial patients were given a 600-mg loading dose of clopidogrel followed by 75 mgs daily from day 2 to day 90 plus aspirin vs. aspirin alone. The results demonstrated a significant reduction in the primary outcome of IS, MI and ischaemic vascular death in the combination group, but this was associated with a significantly increased risk of major haemorrhage. In the POINT trial, presence of a clopidogrel slow metabolizer genetic variant was not associated with a significant reduction in efficacy, but there was a trend towards lower efficacy in these patients, although this did not reach significance due to the small sample size. Following the results of these two trials, the combination of aspirin and clopidogrel is now recommended for patients with recent minor (NIHSS score ≤3) noncardioembolic IS or high-risk TIA (ABCD2 score ≥4). It should be started early (ideally within 12–24 h of symptom onset and at least within 7 days of onset) and continued for 21–90 days. Then, it should be followed by single antiplatelet treatment. A recent trial, The Acute Stroke or Transient Ischaemic Attack Treated with Ticagrelor and ASA for Prevention of Stroke and Death trial (THALES) compared treatment with ticagrelor plus aspirin to aspirin monotherapy for 30 days in patients with a mild to moderate stroke, NIHSS < 5 or high-risk TIA with an ABCD2 score ≥ 6. The primary outcome of stroke both ischaemic and haemorrhagic as well as the secondary outcome of IS was significantly reduced in the combination therapy arm. The combination therapy group had an increased risk of severe and intracranial/fatal bleeding. A subgroup analysis of THALES demonstrated that the benefit of the combination therapy only occurred in patients with documented large artery stenosis of 30% or greater and was not associated with an increased risk of bleeding. In the CHANCE-2 trial, among patients with prior stroke/high-risk TIA and CYP2C19 loss-of-function carriers, the use of ticagrelor plus aspirin was associated with significant modest reduction in 90-day risk of subsequent stroke compared with clopidogrel plus aspirin. The Food and Drug Administration in the United States recently approved the combination of ticagrelor plus aspirin for secondary stroke prevention. In some Asian countries, another drug with antiplatelet activity, cilostazol, is approved for use and widely employed. Indeed, a recent trial in Japan, combining cilostazol with aspirin or clopidogrel, demonstrated that the combination was superior to aspirin or clopidogrel alone without an increased risk for major bleeding. However, cilostazol is not approved for stroke prevention in Europe or the United States.
For all stroke patients, whether ischaemic or ICH, whether suitable for hyperacute interventions or not, the focus of care should not only be on limiting the effects of the initial event, but on limiting brain damage as well as preventing complications, and initiating rehabilitation. These are the basis for good quality organized acute stroke care and result in better patient outcomes. , The stroke pathway is not straightforward as up to 40% of patients get worse after they come into the stroke unit, mostly within the first 24 h. This is referred to as early neurological deterioration, which, if it persists, is termed stroke progression and reflects secondary brain injury. , This fluctuation in patient’s condition can be the result of potentially reversible physiological or neurological factors (brady-/tachycardia, high or low blood glucose, increased metabolic rate in infection) although sometimes they are irreversible (e.g. mass effect, brain stem herniation). A goal of holistic post-stroke care is to achieve improved functional and psychological status and, given its multifactorial nature, will require a multidisciplinary approach. There have been incremental gains in terms of interventions delivered by the multi-disciplinary teams and focused on rehabilitation and functional recovery ( , ). Important examples include arm-robot therapy and mirror therapy which reduced motor deficits and electro-mechanical gait training which increased the number of stroke patients that re-gained the ability to walk. , Similarly, the use of treadmill training helped to improve walking speed and walking endurance among ambulatory stroke survivors. , However, the absence of standard templates for designing and performing large-scale stroke rehabilitation trials has somewhat limited further progress in the field. In the wake of the digital revolution, telemedicine, virtual reality, and robotics have been tried and tested but are yet to translate to real, measurable functional benefits after a stroke. Innovative approaches that integrate stroke survivors into the exercise portion of an existing hospital-based cardiac rehabilitation programme have shown promise, improving endurance, health status, and quality of life for survivors of stroke and providing an opportunity for self-management. It is imperative that physiological and neurological monitoring regimes are in place and that any abnormalities are acted on quickly to avoid further secondary brain injury. Furthermore, these patients are also at risk of the resulting effects of the stroke, as well as the complications of immobility. Care packages and management pathways need then to be put in place immediately and tailored rehabilitation commenced and personalized appropriately to the patient’s needs, both in hospital, during rehabilitation and post-discharge. Indeed, care packages have been developed and evaluated for their effects on outcomes. The most notable is the Quality in Acute Stroke Care, which demonstrated the effectiveness of a care bundle of the assessment and management of Fever, Sugar (hyperglycaemia) and Swallowing dysfunction, in improving outcome by 3 months post-stroke. Several prognostic tools to predict functional status after stroke have been used. Perhaps the most widely employed outcome scale in stroke medicine and research is the modified Rankin Scale, a seven-level, clinician-reported, measure of global disability. Greater functional gain amongst post-stroke survivors during inpatient rehabilitation is associated with better health-related quality of life and independence at follow-up. Another common complication which develops in approximately one third of stroke survivors is post-stroke depression (PSD) and is associated with unfavourable outcomes after stroke. The frequency of PSD is higher at the first year after stroke, and the main factors contributing to this condition are physical disability, stroke severity, history of depression, and cognitive impairment. Post-stroke dementia is often under-recognized with a prevalence of approximately 30% among stroke survivors. , Again, awareness and a multidisciplinary approach is needed, with neuropsychological evaluation adapted to the clinical status. , The incidence of post-stroke dementia increases with older age, low education status, dependence on others for daily living, pre-stroke cognitive decline without dementia, DM, AF and other cardiac arrhythmias, sepsis, congestive heart failure, silent infarcts, brain atrophy, and leukoaraiosis. In the long run, stroke survivors and their families and caregivers will encounter several challenges: poor quality-of-life, stroke-related disabilities, inadequate sources of rehabilitation, social isolation and inadequate support of care givers, overburdening and burnout of care givers, and inadequate efforts to maintain normal life and professions. To meet up to the expectations of patients and families for a good recovery and proper social re-integration, a patient-centric holistic approach of synergistic integrated care is needed involving a multidisciplinary bundle of health care professionals and related services.
Multiple CVD risk factors and comorbidities are common in stroke patients, and part of a multidisciplinary integrated care approach is to address all these risk factors in a holistic manner. This will include management of AF, atherosclerostic vascular disease, systemic hypertension, heart failure, DM, dyslipidaemia, sleep apnoea, and underlying cardiac ischaemia. The ultimate goal is to lower the associated cardiovascular risk burden in these patients, to reduce the risks of recurrent stroke or other major adverse cardiovascular events. Lifestyle changes Lifestyle changes Epidemiological studies have demonstrated that five modifiable risk factors, blood pressure, unhealthy diet, abdominal obesity, physical inactivity, and smoking account for >80% of the population attributable risk for stroke , and are often poorly managed and controlled post-stroke thus increasing the risk of recurrent stroke. Most strokes are largely preventable with healthy lifestyle choices. Hence, multifactorial lifestyle and behavioural interventions, based on theoretical models of behaviour change, employing established methods, delivered by an interdisciplinary team, are likely to have the greatest cumulative benefit in stroke populations. Diet/nutrition There is limited evidence to support dietary interventions to reduce recurrent stroke and the protective effects of nutrition are inferred from epidemiological studies evaluating the effect of dietary factors on risk factors for stroke such as hypertension and hypercholesterolaemia. , Recommendations on diet and nutrition post-stroke advocate adoption of a Mediterranean-type diet, , with evidence suggesting that high intake of fruit and vegetables, , fibre, , and regular consumption of fish , , , confers protection against cardiovascular disease, including stroke, whilst diets high in red and processed meats, fried food, eggs, and sugar-sweetened beverages are associated with increased risk of stroke. , , , , A low-salt diet has been associated with reductions in stroke risk. , Secondary prevention studies, , one in stroke patients, have demonstrated reductions in IS and death and recurrent MI with Mediterranean-style diets. Obesity Weight reduction is recommended for overweight and obese patients with IS or TIA to improve their overall cardiovascular risk profile, with referral to multifactorial intensive lifestyle interventions for obese individuals. Evidence, predominantly from obese patients with diabetes, demonstrates that a 5-10% weight loss, either through traditional behavioural intervention programmes, food substitution, or bariatric surgery, can improve conventional cardiovascular risk factors, namely blood pressure, hyperglycaemia, and dyslipidaemia. Comprehensive behavioural programmes incorporating counselling as a key component appear most effective, with sustained weight loss. , There are no clinical trials demonstrating that weight loss reduces recurrent stroke in an acute IS population, but observational data post-bariatric surgery suggests some benefit of weight reduction on stroke risk ( , ). Physical activity Exercise and regular physical activity reduce the risk of stroke , , and positively impact stroke risk factors by reducing weight and lowering blood pressure (BP) and cholesterol. , , Patients should be encouraged to resume physical activity, with supervision and support as required. Where able, survivors of stroke should participate in 40 min of moderate-vigorous intensity aerobic activity, three to four times per week, otherwise physical activity should be individualized, commensurate with their level of exercise tolerance, stage of recovery, environment, available social support, physical activity preferences, and specific impairments. , Systematic reviews of lifestyle-based interventions , , and exercise-only interventions for secondary stroke prevention are effective in reducing cardiovascular risk factors but their impact on mortality, recurrent stroke, and other vascular events remains to be determined. , Most stroke patients are sedentary and available studies are in TIA patients and ambulatory stroke patients. A post-hoc analysis of the Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis study of those assigned to aggressive medical therapy and targeted risk factor management demonstrated that greater physical activity was associated with lower risk of recurrent stroke, MI or death (OR 0.6, 95% CI 0.4–0.8) and recurrent IS alone ( , ). Alcohol consumption High levels of alcohol consumption (women: >3 drinks/day or >7 drinks/week; men: >4 drinks/day or >14 drinks/week) are associated with greater risk of stroke and are an independent risk factor for stroke recurrence. Low-to-moderate alcohol intake appears to be protective but risk of IS increases (HR 1.04, 95% CI 1.02–1.07) with every 12 g/d increase in alcohol consumption. There are no intervention studies that have directly examined the effect on reducing alcohol intake on risk of recurrent stroke. One small secondary stroke prevention trial focused on lifestyle modification demonstrated reduced alcohol intake but the direct effect on the risk of recurrent stroke could not be determined. Smoking cessation Cessation of smoking post-stroke is an essential health behaviour to promote and support secondary prevention; patients should be offered counselling with or without pharmacological intervention (nicotine replacement/medication). , , Up to two thirds of stroke survivors continue to smoke, thus increasing the risk of recurrent stroke approximately two-fold compared to non-smokers; the more smoked the greater the risk. If a person cannot stop smoking, they should be encouraged to reduce daily smoking , and limit passive smoking. No randomized controlled trials (RCTs) have demonstrated a significant reduction in recurrent stroke following a smoking cessation intervention. One multifactorial RCT employing motivational interviewing targeting modifiable risk factors (smoking, hypertension, diabetes, AF) in patients with a recent (≤2 weeks) non-disabling stroke or TIA, demonstrated a significant impact of the intervention on smoking cessation but no difference between intervention and control groups on major vascular events over an average 3.6 years of follow-up ( , ). Finding AF AF is responsible for 17–36% of ISs, and in up to one-quarter of patients, IS or TIA is a first manifestation of AF. AF-related strokes more commonly lead to death or are disabling, compared to other aetiologies. , An active search for AF is imperative in patients with IS of unknown origin, as detection of arrhythmia and the initiation of OAC may protect the patient against another IS. AF detection rate depends not only on patient characteristics, the time since the stroke, type of stroke, and adopted definition of AF but also on monitoring duration and quality ( , ). As a general principle, if we ‘look harder, look longer and look in more sophisticated ways ….’, we are more likely to find more AF in our post-stroke patients. Currently, the recommended approach in post-stroke patients without known AF mandates screening for the arrhythmia with short-term 24 h continuous electrocardiogram (ECG) monitoring, followed by prolonged (≥72 h) ECG monitoring. As shown by analysis of the LOOP study (Atrial Fibrillation Detected by Continuous ECG Monitoring Using Implantable Loop Recorder to Prevent Stroke in High-risk Individuals), single 72 h ECG monitoring yields sensitivity of only 15% in detecting AF, and the 2018 Canadian Stroke Best Practice Recommendations advocate ≥2 weeks ECG monitoring in patients with ESUS. To date, the optimal duration and method of ECG screening remain elusive and requires further research. Implantable cardiac monitors, by providing continuous monitoring, offer the capability to maximize the chances to detect AF and should be considered after non-invasive monitoring of 7–14 days, up to 30 days in patients with previous stroke at higher probability of having AF ( ). In the LOOP study, performed in patients ‘at risk’ aged ≥70 years but not limited to patients with previous stroke, the combination of slower resting sinus rate, higher body mass index, NT-pro BNP, troponin T together with sex, age, and comorbidities, improved prediction of AF episodes ≥24 h duration. Recently, new screening tools, based on a single-lead ECG or a pulse wave (detected e.g. with plethysmography or oscilometric blood pressure), have emerged, but the clinical utility of majority has yet to be proven, especially in post-stroke patient. , On the other hand, new hand-held or worn (belts, patches, smartphone apps) single-ECG recorders demonstrate good sensitivity and specificity. A recent multicentre RCT in patients with IS or transient ischaemic attack (TIA) demonstrated that 30-day smartphone-based ECG monitoring was better than one additional 24 h Holter monitoring in detecting ≥30s AF (9.5 vs. 2.0%; P = 0.024) and led to more frequent OAC use at 3 months (9.5 vs. 0%, P = 0.002). In summary, a considerable burden of previously unknown AF can be detected when long-term monitoring is applied in at-risk patients. Hypertension, DM, hypercholesterolaemia Blood pressure Hypertension is the most prevalent and modifiable risk factor for primary prevention of stroke. Most studies have also shown that control of BP is beneficial for prevention of a recurrent stroke— , . , The management of blood pressure in the acute phase of a stroke is more variable and depends on whether thrombolysis is administered for IS. Guidelines call for reducing systolic blood pressure (SBP) below 185 mmHg and diastolic below 110 mmHg before IV thrombolysis is started. Given the urgency of reperfusion initiation, IV BP treatments with agents such as labetalol or nicardipine are warranted with close interval BP monitoring after thrombolysis in the first 24 h and maintaining BP of less than 185/105 mmHg. Although intracerebral hemorrhage was reduced, intensive BP lowering after IV thrombolysis to SBP of 130–140 mm Hg has not been shown to improve 90-day functional outcomes compared to target SBP of <180 mm Hg. The management of BP after endovascular thrombectomy (EVT) is also unclear. Most acute EVT trials for acute IS excluded patients if BP was greater than 185/110 mmHg. The recent blood pressure lowering after successful endovascular therapy in acute ischaemic stroke trial failed to show any significant difference in radiographically determined intraparenchymal haemorrhage risks for intensive BP control (100–129 mmHg) compared to standard (SBP 130–185 mmHg) among successfully EVT treated large vessel occlusion cases. Ongoing trials (for example, BEST-II: NCT04116112, OPTIMAL_BP: NCT04205305, ENCHANTED 2: NCT04140110, CRISIS I: NCT04775147) are continuing to address whether more intensive BP management is better than standard SBP targets of <180 mmHg after successful reperfusion. After the acute stroke period, management and control of BP with targets of <140/90 mmHg to reduce stroke recurrence and other cardiovascular conditions are supported in evidence-based recommendations. Among diabetics and those with cerebral small vessel disease, more aggressive targets of <130/80 mmHg are reasonable. For intracerebral haemorrhagic strokes, rapid or intensive lowering of SBP to <180 mmHg did not improve functional outcomes at 90 days. Overall, there is insufficient evidence to show that lowering BP reduces haematoma expansion. In combined analyses among the subgroup of patients treated within 6 h, intensive BP lowering (<140 and >110 mmHg) has been recommended based on the lower risk of haematoma expansion. In summary, sudden and significant reduction of blood pressure during acute phase of ischaemic or haemorrhagic stroke may worsen outcomes; however, after acute period, tight BP control (<140/90 mmHg, or <130/80 mmHg in diabetics) protects against recurrent stroke and other cardiovascular events. Diabetes mellitus All stroke patients without diagnosed diabetes should be screened for DM. Despite unfavourable effects of hyperglycaemia in acute IS in diabetic, and non-diabetic patients, as showed by many studies ( , ), strict glycaemia control (e.g. 70–135 mg/dL) in acute stroke is not beneficial and may be even harmful, putting the patient at risk for hypoglycaemia and early neurologic deterioration. Thus, the accepted approach to glycaemia in acute stroke is less stringent (e.g. range 70–180 mg/dL). , Patients with post-stroke DM are at a very high risk of cardiovascular complications and should be tightly controlled with a target glycosylated haemoglobin 1c <7% (see for proposed management). , In summary, while less stringent glycaemia control is beneficial in the acute phase of stroke, post-stroke patients should be subjected to rigorous, long-term glycaemic control. Dyslipidaemia The available evidence on role of dyslipidaemia for pathogenesis of stroke is moderately robust, but most of the data ( , ) indicate that lowering LDL-C is the primary target in IS, and each 1 mmol/L (39 mg/dL) decrease in LDL-C levels reduces the risk of any stroke by 21.1%. First-line drugs are statins. In patients with initial LDL-C levels which are much higher than the target LDL-levels, ezetimibe could be started right away as an add-on treatment to statins. Although the related evidence is low, especially when compared to the role of immediate onset of statin treatment in patients with acute MI, we suggest that statins with or without ezetimibe is instituted early in the acute phase, as this strategy reduced risk of recurrent stroke in TIA patients with carotid stenosis. Long-term statin use reduced reoccurrence of fatal or non-fatal stroke in patients after stroke or TIA (in the Stroke Prevention by Aggressive Reduction in Cholesterol Levels study HR 0.84; 95% CI 0.71–0.99), and higher statin adherence reduced risk of recurrent stroke in patients with recent IS without AF (HR 0.78; 95% CI, 0.63–0.97) and with AF (HR 0.59; 95% CI, 0.43–0.81). , Initial reports had generated the hypothesis that statins and other lipid-lowering drugs may perhaps increase the of haemorrhagic stroke, but this was not confirmed in plenty of large-scale epidemiological studies. , Patients with IS are at very high cardiovascular risk and should be subject to at least 50% reduction of LDL-C with target LDL-C value of 1.4 mmol/L (55 mg/dL). Patients with IS or TIA which is attributed to a specific aetiology that is not related to cardiovascular risk factors like cervical artery dissection, patent foramen ovale (PFO), endocarditis, and atrial myxoma should not be a priori considered as of very high risk for stroke recurrence and cardiovascular morbidity and mortality. For such patients, it is suggested that lipid lowering treatment should be based on a personalized 10-year cardiovascular risk, estimated by the calibrated country-specific SCORE. shows the proposed management of dyslipidaemia in IS. In summary, LDL-C lowering is the primary target in IS and an early institution of statins and their long-term use reduce the risk of a recurrent stroke. In patients at increased risk of ischaemic events owing to elevated triglyceride levels despite statin use the additional triglyceride-lowering therapy reduces the residual risk and improve survival. Other comorbidities Patent foramen ovale PFO is a common abnormality that affects up to 20–25% of women and a smaller percentage of men. In most people, a PFO remains asymptomatic, but in a small percentage, it may be associated with the development of an embolic IS via paradoxical embolization Before concluding that a PFO is the causative mechanism for an IS, an exhaustive search for other stroke aetiologies should be performed ( ). When other potential stroke aetiologies have been excluded and the PFO is determined to be the likely stroke mechanism, then treatment options to prevent future strokes need consideration. The risk of a recurrent stroke in PFO-related stroke is low, approximately 1–2% per year, but it is more substantial in higher risk PFOs, i.e. those with a larger amount of intracardiac shunting or when the PFO is associated with an atrial septal aneurysm. Two approaches to secondary prevention are employed: antithrombotic therapy that in most cases consists of antiplatelet treatment or alternatively interventional closure of the PFO by occluder or remote surgical (stitch) technology. These two therapeutic approaches were evaluated in clinical trials in PFO-related stroke patients under the age of 60 as it is understood that with higher age other potential causes of stroke may be increasingly prevalent even if they had not been detected in standard diagnostic workup. In these patients it was demonstrated that PFO closure with long term antiplatelet therapy was superior to medical therapy alone, but the benefit was only observed in patients with high-risk PFOs (i.e. long-tunnel PFO ≥ 10 mm, hypermobile interatrial septum, Eustachian valve or Chiari’s network, a large right-to-left shunt during Valsalva manoeuver, low-angle PFO ≤ 10°). PFO closure was associated with a risk of developing transient AF, as well an increased incidence of haematomas, deep vein thrombosis, and pulmonary embolism. The increased risk of postprocedural AF declines after the early postprocedural period (e.g. first 45 days after PFO closure) but remains elevated compared with the general population during a long-term follow-up. The evaluation of patients with presumed PFO-related ISs requires close collaboration between stroke specialists and imaging and interventional cardiologists. The stroke specialist should determine if the stroke was likely embolic and perform appropriate evaluation to exclude other potential stroke mechanisms, such as large and small artery atherosclerosis and other potential cardiac sources, whilst the cardiologist can assist in performing and interpreting cardiac imaging to identify a cardiac source of the stroke. This should be followed by multidisciplinary team (MDT) discussion about the best approach for secondary stroke prevention. An MDT meeting should be mandated for patient evaluation and selection of intervention. Cardiac thrombus In patients with an acute stroke in whom a cardio-embolic mechanism is suspected the following potential causes should be excluded: AF or flutter, thrombi in the left ventricle (LV) or in the left atrium/left atrial appendage, or on a prosthetic cardiac valve, as well as other conditions such as mitral stenosis, atrial myxoma, intracardiac masses, or valvular vegetations. LV mural thrombi account for up to 10% of cardio-embolic strokes and are most often seen in patients who have had a prior extensive anterior MI, with segmental akinesis or dyskinesis. Usually, the development of a LV thrombus occurs between 24 h and 2 weeks after the onset of a MI, with an increased risk in patients with depressed LV function. The risk of stroke or systemic embolism in the absence of anticoagulation, is as high as 10–20% at 3 months, with the highest risk in the first weeks, , declining after 3 months, in parallel with evolution to an organized and fibrotic clot adhering to the endocardium. Transthoracic echocardiography (TTE) is traditionally the standard imaging technique for detecting LV thrombi in patients with acute IS. More recently, cardiac magnetic resonance (CMR) was found to be superior to TTE in detecting LV thrombus in patients with history of MI and LV dysfunction (LVEF < 50%), with cine-CMR and contrast-enhanced CMR offering the highest diagnostic yield, especially in case of mural or small LV thrombi. Since CMR is a time-consuming, expensive examination, not easily available in most centres, new approaches have been developed in the field of echocardiography, with ultrasound contrast agents significantly improving the diagnostic accuracy of TTE. Anticoagulation is absolutely indicated if a LV thrombus is detected and has to be integrated and combined with control of hypertension and other risk factors. For oral anticoagulation, VKAs have been traditionally recommended. For the DOACs, few data are available, limited to case reports and case series, and their use in this setting remains off-label.
Lifestyle changes Epidemiological studies have demonstrated that five modifiable risk factors, blood pressure, unhealthy diet, abdominal obesity, physical inactivity, and smoking account for >80% of the population attributable risk for stroke , and are often poorly managed and controlled post-stroke thus increasing the risk of recurrent stroke. Most strokes are largely preventable with healthy lifestyle choices. Hence, multifactorial lifestyle and behavioural interventions, based on theoretical models of behaviour change, employing established methods, delivered by an interdisciplinary team, are likely to have the greatest cumulative benefit in stroke populations. Diet/nutrition There is limited evidence to support dietary interventions to reduce recurrent stroke and the protective effects of nutrition are inferred from epidemiological studies evaluating the effect of dietary factors on risk factors for stroke such as hypertension and hypercholesterolaemia. , Recommendations on diet and nutrition post-stroke advocate adoption of a Mediterranean-type diet, , with evidence suggesting that high intake of fruit and vegetables, , fibre, , and regular consumption of fish , , , confers protection against cardiovascular disease, including stroke, whilst diets high in red and processed meats, fried food, eggs, and sugar-sweetened beverages are associated with increased risk of stroke. , , , , A low-salt diet has been associated with reductions in stroke risk. , Secondary prevention studies, , one in stroke patients, have demonstrated reductions in IS and death and recurrent MI with Mediterranean-style diets. Obesity Weight reduction is recommended for overweight and obese patients with IS or TIA to improve their overall cardiovascular risk profile, with referral to multifactorial intensive lifestyle interventions for obese individuals. Evidence, predominantly from obese patients with diabetes, demonstrates that a 5-10% weight loss, either through traditional behavioural intervention programmes, food substitution, or bariatric surgery, can improve conventional cardiovascular risk factors, namely blood pressure, hyperglycaemia, and dyslipidaemia. Comprehensive behavioural programmes incorporating counselling as a key component appear most effective, with sustained weight loss. , There are no clinical trials demonstrating that weight loss reduces recurrent stroke in an acute IS population, but observational data post-bariatric surgery suggests some benefit of weight reduction on stroke risk ( , ). Physical activity Exercise and regular physical activity reduce the risk of stroke , , and positively impact stroke risk factors by reducing weight and lowering blood pressure (BP) and cholesterol. , , Patients should be encouraged to resume physical activity, with supervision and support as required. Where able, survivors of stroke should participate in 40 min of moderate-vigorous intensity aerobic activity, three to four times per week, otherwise physical activity should be individualized, commensurate with their level of exercise tolerance, stage of recovery, environment, available social support, physical activity preferences, and specific impairments. , Systematic reviews of lifestyle-based interventions , , and exercise-only interventions for secondary stroke prevention are effective in reducing cardiovascular risk factors but their impact on mortality, recurrent stroke, and other vascular events remains to be determined. , Most stroke patients are sedentary and available studies are in TIA patients and ambulatory stroke patients. A post-hoc analysis of the Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis study of those assigned to aggressive medical therapy and targeted risk factor management demonstrated that greater physical activity was associated with lower risk of recurrent stroke, MI or death (OR 0.6, 95% CI 0.4–0.8) and recurrent IS alone ( , ). Alcohol consumption High levels of alcohol consumption (women: >3 drinks/day or >7 drinks/week; men: >4 drinks/day or >14 drinks/week) are associated with greater risk of stroke and are an independent risk factor for stroke recurrence. Low-to-moderate alcohol intake appears to be protective but risk of IS increases (HR 1.04, 95% CI 1.02–1.07) with every 12 g/d increase in alcohol consumption. There are no intervention studies that have directly examined the effect on reducing alcohol intake on risk of recurrent stroke. One small secondary stroke prevention trial focused on lifestyle modification demonstrated reduced alcohol intake but the direct effect on the risk of recurrent stroke could not be determined. Smoking cessation Cessation of smoking post-stroke is an essential health behaviour to promote and support secondary prevention; patients should be offered counselling with or without pharmacological intervention (nicotine replacement/medication). , , Up to two thirds of stroke survivors continue to smoke, thus increasing the risk of recurrent stroke approximately two-fold compared to non-smokers; the more smoked the greater the risk. If a person cannot stop smoking, they should be encouraged to reduce daily smoking , and limit passive smoking. No randomized controlled trials (RCTs) have demonstrated a significant reduction in recurrent stroke following a smoking cessation intervention. One multifactorial RCT employing motivational interviewing targeting modifiable risk factors (smoking, hypertension, diabetes, AF) in patients with a recent (≤2 weeks) non-disabling stroke or TIA, demonstrated a significant impact of the intervention on smoking cessation but no difference between intervention and control groups on major vascular events over an average 3.6 years of follow-up ( , ).
Epidemiological studies have demonstrated that five modifiable risk factors, blood pressure, unhealthy diet, abdominal obesity, physical inactivity, and smoking account for >80% of the population attributable risk for stroke , and are often poorly managed and controlled post-stroke thus increasing the risk of recurrent stroke. Most strokes are largely preventable with healthy lifestyle choices. Hence, multifactorial lifestyle and behavioural interventions, based on theoretical models of behaviour change, employing established methods, delivered by an interdisciplinary team, are likely to have the greatest cumulative benefit in stroke populations.
There is limited evidence to support dietary interventions to reduce recurrent stroke and the protective effects of nutrition are inferred from epidemiological studies evaluating the effect of dietary factors on risk factors for stroke such as hypertension and hypercholesterolaemia. , Recommendations on diet and nutrition post-stroke advocate adoption of a Mediterranean-type diet, , with evidence suggesting that high intake of fruit and vegetables, , fibre, , and regular consumption of fish , , , confers protection against cardiovascular disease, including stroke, whilst diets high in red and processed meats, fried food, eggs, and sugar-sweetened beverages are associated with increased risk of stroke. , , , , A low-salt diet has been associated with reductions in stroke risk. , Secondary prevention studies, , one in stroke patients, have demonstrated reductions in IS and death and recurrent MI with Mediterranean-style diets.
Weight reduction is recommended for overweight and obese patients with IS or TIA to improve their overall cardiovascular risk profile, with referral to multifactorial intensive lifestyle interventions for obese individuals. Evidence, predominantly from obese patients with diabetes, demonstrates that a 5-10% weight loss, either through traditional behavioural intervention programmes, food substitution, or bariatric surgery, can improve conventional cardiovascular risk factors, namely blood pressure, hyperglycaemia, and dyslipidaemia. Comprehensive behavioural programmes incorporating counselling as a key component appear most effective, with sustained weight loss. , There are no clinical trials demonstrating that weight loss reduces recurrent stroke in an acute IS population, but observational data post-bariatric surgery suggests some benefit of weight reduction on stroke risk ( , ).
Exercise and regular physical activity reduce the risk of stroke , , and positively impact stroke risk factors by reducing weight and lowering blood pressure (BP) and cholesterol. , , Patients should be encouraged to resume physical activity, with supervision and support as required. Where able, survivors of stroke should participate in 40 min of moderate-vigorous intensity aerobic activity, three to four times per week, otherwise physical activity should be individualized, commensurate with their level of exercise tolerance, stage of recovery, environment, available social support, physical activity preferences, and specific impairments. , Systematic reviews of lifestyle-based interventions , , and exercise-only interventions for secondary stroke prevention are effective in reducing cardiovascular risk factors but their impact on mortality, recurrent stroke, and other vascular events remains to be determined. , Most stroke patients are sedentary and available studies are in TIA patients and ambulatory stroke patients. A post-hoc analysis of the Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis study of those assigned to aggressive medical therapy and targeted risk factor management demonstrated that greater physical activity was associated with lower risk of recurrent stroke, MI or death (OR 0.6, 95% CI 0.4–0.8) and recurrent IS alone ( , ).
High levels of alcohol consumption (women: >3 drinks/day or >7 drinks/week; men: >4 drinks/day or >14 drinks/week) are associated with greater risk of stroke and are an independent risk factor for stroke recurrence. Low-to-moderate alcohol intake appears to be protective but risk of IS increases (HR 1.04, 95% CI 1.02–1.07) with every 12 g/d increase in alcohol consumption. There are no intervention studies that have directly examined the effect on reducing alcohol intake on risk of recurrent stroke. One small secondary stroke prevention trial focused on lifestyle modification demonstrated reduced alcohol intake but the direct effect on the risk of recurrent stroke could not be determined.
Cessation of smoking post-stroke is an essential health behaviour to promote and support secondary prevention; patients should be offered counselling with or without pharmacological intervention (nicotine replacement/medication). , , Up to two thirds of stroke survivors continue to smoke, thus increasing the risk of recurrent stroke approximately two-fold compared to non-smokers; the more smoked the greater the risk. If a person cannot stop smoking, they should be encouraged to reduce daily smoking , and limit passive smoking. No randomized controlled trials (RCTs) have demonstrated a significant reduction in recurrent stroke following a smoking cessation intervention. One multifactorial RCT employing motivational interviewing targeting modifiable risk factors (smoking, hypertension, diabetes, AF) in patients with a recent (≤2 weeks) non-disabling stroke or TIA, demonstrated a significant impact of the intervention on smoking cessation but no difference between intervention and control groups on major vascular events over an average 3.6 years of follow-up ( , ).
AF is responsible for 17–36% of ISs, and in up to one-quarter of patients, IS or TIA is a first manifestation of AF. AF-related strokes more commonly lead to death or are disabling, compared to other aetiologies. , An active search for AF is imperative in patients with IS of unknown origin, as detection of arrhythmia and the initiation of OAC may protect the patient against another IS. AF detection rate depends not only on patient characteristics, the time since the stroke, type of stroke, and adopted definition of AF but also on monitoring duration and quality ( , ). As a general principle, if we ‘look harder, look longer and look in more sophisticated ways ….’, we are more likely to find more AF in our post-stroke patients. Currently, the recommended approach in post-stroke patients without known AF mandates screening for the arrhythmia with short-term 24 h continuous electrocardiogram (ECG) monitoring, followed by prolonged (≥72 h) ECG monitoring. As shown by analysis of the LOOP study (Atrial Fibrillation Detected by Continuous ECG Monitoring Using Implantable Loop Recorder to Prevent Stroke in High-risk Individuals), single 72 h ECG monitoring yields sensitivity of only 15% in detecting AF, and the 2018 Canadian Stroke Best Practice Recommendations advocate ≥2 weeks ECG monitoring in patients with ESUS. To date, the optimal duration and method of ECG screening remain elusive and requires further research. Implantable cardiac monitors, by providing continuous monitoring, offer the capability to maximize the chances to detect AF and should be considered after non-invasive monitoring of 7–14 days, up to 30 days in patients with previous stroke at higher probability of having AF ( ). In the LOOP study, performed in patients ‘at risk’ aged ≥70 years but not limited to patients with previous stroke, the combination of slower resting sinus rate, higher body mass index, NT-pro BNP, troponin T together with sex, age, and comorbidities, improved prediction of AF episodes ≥24 h duration. Recently, new screening tools, based on a single-lead ECG or a pulse wave (detected e.g. with plethysmography or oscilometric blood pressure), have emerged, but the clinical utility of majority has yet to be proven, especially in post-stroke patient. , On the other hand, new hand-held or worn (belts, patches, smartphone apps) single-ECG recorders demonstrate good sensitivity and specificity. A recent multicentre RCT in patients with IS or transient ischaemic attack (TIA) demonstrated that 30-day smartphone-based ECG monitoring was better than one additional 24 h Holter monitoring in detecting ≥30s AF (9.5 vs. 2.0%; P = 0.024) and led to more frequent OAC use at 3 months (9.5 vs. 0%, P = 0.002). In summary, a considerable burden of previously unknown AF can be detected when long-term monitoring is applied in at-risk patients.
Blood pressure Hypertension is the most prevalent and modifiable risk factor for primary prevention of stroke. Most studies have also shown that control of BP is beneficial for prevention of a recurrent stroke— , . , The management of blood pressure in the acute phase of a stroke is more variable and depends on whether thrombolysis is administered for IS. Guidelines call for reducing systolic blood pressure (SBP) below 185 mmHg and diastolic below 110 mmHg before IV thrombolysis is started. Given the urgency of reperfusion initiation, IV BP treatments with agents such as labetalol or nicardipine are warranted with close interval BP monitoring after thrombolysis in the first 24 h and maintaining BP of less than 185/105 mmHg. Although intracerebral hemorrhage was reduced, intensive BP lowering after IV thrombolysis to SBP of 130–140 mm Hg has not been shown to improve 90-day functional outcomes compared to target SBP of <180 mm Hg. The management of BP after endovascular thrombectomy (EVT) is also unclear. Most acute EVT trials for acute IS excluded patients if BP was greater than 185/110 mmHg. The recent blood pressure lowering after successful endovascular therapy in acute ischaemic stroke trial failed to show any significant difference in radiographically determined intraparenchymal haemorrhage risks for intensive BP control (100–129 mmHg) compared to standard (SBP 130–185 mmHg) among successfully EVT treated large vessel occlusion cases. Ongoing trials (for example, BEST-II: NCT04116112, OPTIMAL_BP: NCT04205305, ENCHANTED 2: NCT04140110, CRISIS I: NCT04775147) are continuing to address whether more intensive BP management is better than standard SBP targets of <180 mmHg after successful reperfusion. After the acute stroke period, management and control of BP with targets of <140/90 mmHg to reduce stroke recurrence and other cardiovascular conditions are supported in evidence-based recommendations. Among diabetics and those with cerebral small vessel disease, more aggressive targets of <130/80 mmHg are reasonable. For intracerebral haemorrhagic strokes, rapid or intensive lowering of SBP to <180 mmHg did not improve functional outcomes at 90 days. Overall, there is insufficient evidence to show that lowering BP reduces haematoma expansion. In combined analyses among the subgroup of patients treated within 6 h, intensive BP lowering (<140 and >110 mmHg) has been recommended based on the lower risk of haematoma expansion. In summary, sudden and significant reduction of blood pressure during acute phase of ischaemic or haemorrhagic stroke may worsen outcomes; however, after acute period, tight BP control (<140/90 mmHg, or <130/80 mmHg in diabetics) protects against recurrent stroke and other cardiovascular events. Diabetes mellitus All stroke patients without diagnosed diabetes should be screened for DM. Despite unfavourable effects of hyperglycaemia in acute IS in diabetic, and non-diabetic patients, as showed by many studies ( , ), strict glycaemia control (e.g. 70–135 mg/dL) in acute stroke is not beneficial and may be even harmful, putting the patient at risk for hypoglycaemia and early neurologic deterioration. Thus, the accepted approach to glycaemia in acute stroke is less stringent (e.g. range 70–180 mg/dL). , Patients with post-stroke DM are at a very high risk of cardiovascular complications and should be tightly controlled with a target glycosylated haemoglobin 1c <7% (see for proposed management). , In summary, while less stringent glycaemia control is beneficial in the acute phase of stroke, post-stroke patients should be subjected to rigorous, long-term glycaemic control. Dyslipidaemia The available evidence on role of dyslipidaemia for pathogenesis of stroke is moderately robust, but most of the data ( , ) indicate that lowering LDL-C is the primary target in IS, and each 1 mmol/L (39 mg/dL) decrease in LDL-C levels reduces the risk of any stroke by 21.1%. First-line drugs are statins. In patients with initial LDL-C levels which are much higher than the target LDL-levels, ezetimibe could be started right away as an add-on treatment to statins. Although the related evidence is low, especially when compared to the role of immediate onset of statin treatment in patients with acute MI, we suggest that statins with or without ezetimibe is instituted early in the acute phase, as this strategy reduced risk of recurrent stroke in TIA patients with carotid stenosis. Long-term statin use reduced reoccurrence of fatal or non-fatal stroke in patients after stroke or TIA (in the Stroke Prevention by Aggressive Reduction in Cholesterol Levels study HR 0.84; 95% CI 0.71–0.99), and higher statin adherence reduced risk of recurrent stroke in patients with recent IS without AF (HR 0.78; 95% CI, 0.63–0.97) and with AF (HR 0.59; 95% CI, 0.43–0.81). , Initial reports had generated the hypothesis that statins and other lipid-lowering drugs may perhaps increase the of haemorrhagic stroke, but this was not confirmed in plenty of large-scale epidemiological studies. , Patients with IS are at very high cardiovascular risk and should be subject to at least 50% reduction of LDL-C with target LDL-C value of 1.4 mmol/L (55 mg/dL). Patients with IS or TIA which is attributed to a specific aetiology that is not related to cardiovascular risk factors like cervical artery dissection, patent foramen ovale (PFO), endocarditis, and atrial myxoma should not be a priori considered as of very high risk for stroke recurrence and cardiovascular morbidity and mortality. For such patients, it is suggested that lipid lowering treatment should be based on a personalized 10-year cardiovascular risk, estimated by the calibrated country-specific SCORE. shows the proposed management of dyslipidaemia in IS. In summary, LDL-C lowering is the primary target in IS and an early institution of statins and their long-term use reduce the risk of a recurrent stroke. In patients at increased risk of ischaemic events owing to elevated triglyceride levels despite statin use the additional triglyceride-lowering therapy reduces the residual risk and improve survival.
Hypertension is the most prevalent and modifiable risk factor for primary prevention of stroke. Most studies have also shown that control of BP is beneficial for prevention of a recurrent stroke— , . , The management of blood pressure in the acute phase of a stroke is more variable and depends on whether thrombolysis is administered for IS. Guidelines call for reducing systolic blood pressure (SBP) below 185 mmHg and diastolic below 110 mmHg before IV thrombolysis is started. Given the urgency of reperfusion initiation, IV BP treatments with agents such as labetalol or nicardipine are warranted with close interval BP monitoring after thrombolysis in the first 24 h and maintaining BP of less than 185/105 mmHg. Although intracerebral hemorrhage was reduced, intensive BP lowering after IV thrombolysis to SBP of 130–140 mm Hg has not been shown to improve 90-day functional outcomes compared to target SBP of <180 mm Hg. The management of BP after endovascular thrombectomy (EVT) is also unclear. Most acute EVT trials for acute IS excluded patients if BP was greater than 185/110 mmHg. The recent blood pressure lowering after successful endovascular therapy in acute ischaemic stroke trial failed to show any significant difference in radiographically determined intraparenchymal haemorrhage risks for intensive BP control (100–129 mmHg) compared to standard (SBP 130–185 mmHg) among successfully EVT treated large vessel occlusion cases. Ongoing trials (for example, BEST-II: NCT04116112, OPTIMAL_BP: NCT04205305, ENCHANTED 2: NCT04140110, CRISIS I: NCT04775147) are continuing to address whether more intensive BP management is better than standard SBP targets of <180 mmHg after successful reperfusion. After the acute stroke period, management and control of BP with targets of <140/90 mmHg to reduce stroke recurrence and other cardiovascular conditions are supported in evidence-based recommendations. Among diabetics and those with cerebral small vessel disease, more aggressive targets of <130/80 mmHg are reasonable. For intracerebral haemorrhagic strokes, rapid or intensive lowering of SBP to <180 mmHg did not improve functional outcomes at 90 days. Overall, there is insufficient evidence to show that lowering BP reduces haematoma expansion. In combined analyses among the subgroup of patients treated within 6 h, intensive BP lowering (<140 and >110 mmHg) has been recommended based on the lower risk of haematoma expansion. In summary, sudden and significant reduction of blood pressure during acute phase of ischaemic or haemorrhagic stroke may worsen outcomes; however, after acute period, tight BP control (<140/90 mmHg, or <130/80 mmHg in diabetics) protects against recurrent stroke and other cardiovascular events.
All stroke patients without diagnosed diabetes should be screened for DM. Despite unfavourable effects of hyperglycaemia in acute IS in diabetic, and non-diabetic patients, as showed by many studies ( , ), strict glycaemia control (e.g. 70–135 mg/dL) in acute stroke is not beneficial and may be even harmful, putting the patient at risk for hypoglycaemia and early neurologic deterioration. Thus, the accepted approach to glycaemia in acute stroke is less stringent (e.g. range 70–180 mg/dL). , Patients with post-stroke DM are at a very high risk of cardiovascular complications and should be tightly controlled with a target glycosylated haemoglobin 1c <7% (see for proposed management). , In summary, while less stringent glycaemia control is beneficial in the acute phase of stroke, post-stroke patients should be subjected to rigorous, long-term glycaemic control.
The available evidence on role of dyslipidaemia for pathogenesis of stroke is moderately robust, but most of the data ( , ) indicate that lowering LDL-C is the primary target in IS, and each 1 mmol/L (39 mg/dL) decrease in LDL-C levels reduces the risk of any stroke by 21.1%. First-line drugs are statins. In patients with initial LDL-C levels which are much higher than the target LDL-levels, ezetimibe could be started right away as an add-on treatment to statins. Although the related evidence is low, especially when compared to the role of immediate onset of statin treatment in patients with acute MI, we suggest that statins with or without ezetimibe is instituted early in the acute phase, as this strategy reduced risk of recurrent stroke in TIA patients with carotid stenosis. Long-term statin use reduced reoccurrence of fatal or non-fatal stroke in patients after stroke or TIA (in the Stroke Prevention by Aggressive Reduction in Cholesterol Levels study HR 0.84; 95% CI 0.71–0.99), and higher statin adherence reduced risk of recurrent stroke in patients with recent IS without AF (HR 0.78; 95% CI, 0.63–0.97) and with AF (HR 0.59; 95% CI, 0.43–0.81). , Initial reports had generated the hypothesis that statins and other lipid-lowering drugs may perhaps increase the of haemorrhagic stroke, but this was not confirmed in plenty of large-scale epidemiological studies. , Patients with IS are at very high cardiovascular risk and should be subject to at least 50% reduction of LDL-C with target LDL-C value of 1.4 mmol/L (55 mg/dL). Patients with IS or TIA which is attributed to a specific aetiology that is not related to cardiovascular risk factors like cervical artery dissection, patent foramen ovale (PFO), endocarditis, and atrial myxoma should not be a priori considered as of very high risk for stroke recurrence and cardiovascular morbidity and mortality. For such patients, it is suggested that lipid lowering treatment should be based on a personalized 10-year cardiovascular risk, estimated by the calibrated country-specific SCORE. shows the proposed management of dyslipidaemia in IS. In summary, LDL-C lowering is the primary target in IS and an early institution of statins and their long-term use reduce the risk of a recurrent stroke. In patients at increased risk of ischaemic events owing to elevated triglyceride levels despite statin use the additional triglyceride-lowering therapy reduces the residual risk and improve survival.
Patent foramen ovale PFO is a common abnormality that affects up to 20–25% of women and a smaller percentage of men. In most people, a PFO remains asymptomatic, but in a small percentage, it may be associated with the development of an embolic IS via paradoxical embolization Before concluding that a PFO is the causative mechanism for an IS, an exhaustive search for other stroke aetiologies should be performed ( ). When other potential stroke aetiologies have been excluded and the PFO is determined to be the likely stroke mechanism, then treatment options to prevent future strokes need consideration. The risk of a recurrent stroke in PFO-related stroke is low, approximately 1–2% per year, but it is more substantial in higher risk PFOs, i.e. those with a larger amount of intracardiac shunting or when the PFO is associated with an atrial septal aneurysm. Two approaches to secondary prevention are employed: antithrombotic therapy that in most cases consists of antiplatelet treatment or alternatively interventional closure of the PFO by occluder or remote surgical (stitch) technology. These two therapeutic approaches were evaluated in clinical trials in PFO-related stroke patients under the age of 60 as it is understood that with higher age other potential causes of stroke may be increasingly prevalent even if they had not been detected in standard diagnostic workup. In these patients it was demonstrated that PFO closure with long term antiplatelet therapy was superior to medical therapy alone, but the benefit was only observed in patients with high-risk PFOs (i.e. long-tunnel PFO ≥ 10 mm, hypermobile interatrial septum, Eustachian valve or Chiari’s network, a large right-to-left shunt during Valsalva manoeuver, low-angle PFO ≤ 10°). PFO closure was associated with a risk of developing transient AF, as well an increased incidence of haematomas, deep vein thrombosis, and pulmonary embolism. The increased risk of postprocedural AF declines after the early postprocedural period (e.g. first 45 days after PFO closure) but remains elevated compared with the general population during a long-term follow-up. The evaluation of patients with presumed PFO-related ISs requires close collaboration between stroke specialists and imaging and interventional cardiologists. The stroke specialist should determine if the stroke was likely embolic and perform appropriate evaluation to exclude other potential stroke mechanisms, such as large and small artery atherosclerosis and other potential cardiac sources, whilst the cardiologist can assist in performing and interpreting cardiac imaging to identify a cardiac source of the stroke. This should be followed by multidisciplinary team (MDT) discussion about the best approach for secondary stroke prevention. An MDT meeting should be mandated for patient evaluation and selection of intervention. Cardiac thrombus In patients with an acute stroke in whom a cardio-embolic mechanism is suspected the following potential causes should be excluded: AF or flutter, thrombi in the left ventricle (LV) or in the left atrium/left atrial appendage, or on a prosthetic cardiac valve, as well as other conditions such as mitral stenosis, atrial myxoma, intracardiac masses, or valvular vegetations. LV mural thrombi account for up to 10% of cardio-embolic strokes and are most often seen in patients who have had a prior extensive anterior MI, with segmental akinesis or dyskinesis. Usually, the development of a LV thrombus occurs between 24 h and 2 weeks after the onset of a MI, with an increased risk in patients with depressed LV function. The risk of stroke or systemic embolism in the absence of anticoagulation, is as high as 10–20% at 3 months, with the highest risk in the first weeks, , declining after 3 months, in parallel with evolution to an organized and fibrotic clot adhering to the endocardium. Transthoracic echocardiography (TTE) is traditionally the standard imaging technique for detecting LV thrombi in patients with acute IS. More recently, cardiac magnetic resonance (CMR) was found to be superior to TTE in detecting LV thrombus in patients with history of MI and LV dysfunction (LVEF < 50%), with cine-CMR and contrast-enhanced CMR offering the highest diagnostic yield, especially in case of mural or small LV thrombi. Since CMR is a time-consuming, expensive examination, not easily available in most centres, new approaches have been developed in the field of echocardiography, with ultrasound contrast agents significantly improving the diagnostic accuracy of TTE. Anticoagulation is absolutely indicated if a LV thrombus is detected and has to be integrated and combined with control of hypertension and other risk factors. For oral anticoagulation, VKAs have been traditionally recommended. For the DOACs, few data are available, limited to case reports and case series, and their use in this setting remains off-label.
PFO is a common abnormality that affects up to 20–25% of women and a smaller percentage of men. In most people, a PFO remains asymptomatic, but in a small percentage, it may be associated with the development of an embolic IS via paradoxical embolization Before concluding that a PFO is the causative mechanism for an IS, an exhaustive search for other stroke aetiologies should be performed ( ). When other potential stroke aetiologies have been excluded and the PFO is determined to be the likely stroke mechanism, then treatment options to prevent future strokes need consideration. The risk of a recurrent stroke in PFO-related stroke is low, approximately 1–2% per year, but it is more substantial in higher risk PFOs, i.e. those with a larger amount of intracardiac shunting or when the PFO is associated with an atrial septal aneurysm. Two approaches to secondary prevention are employed: antithrombotic therapy that in most cases consists of antiplatelet treatment or alternatively interventional closure of the PFO by occluder or remote surgical (stitch) technology. These two therapeutic approaches were evaluated in clinical trials in PFO-related stroke patients under the age of 60 as it is understood that with higher age other potential causes of stroke may be increasingly prevalent even if they had not been detected in standard diagnostic workup. In these patients it was demonstrated that PFO closure with long term antiplatelet therapy was superior to medical therapy alone, but the benefit was only observed in patients with high-risk PFOs (i.e. long-tunnel PFO ≥ 10 mm, hypermobile interatrial septum, Eustachian valve or Chiari’s network, a large right-to-left shunt during Valsalva manoeuver, low-angle PFO ≤ 10°). PFO closure was associated with a risk of developing transient AF, as well an increased incidence of haematomas, deep vein thrombosis, and pulmonary embolism. The increased risk of postprocedural AF declines after the early postprocedural period (e.g. first 45 days after PFO closure) but remains elevated compared with the general population during a long-term follow-up. The evaluation of patients with presumed PFO-related ISs requires close collaboration between stroke specialists and imaging and interventional cardiologists. The stroke specialist should determine if the stroke was likely embolic and perform appropriate evaluation to exclude other potential stroke mechanisms, such as large and small artery atherosclerosis and other potential cardiac sources, whilst the cardiologist can assist in performing and interpreting cardiac imaging to identify a cardiac source of the stroke. This should be followed by multidisciplinary team (MDT) discussion about the best approach for secondary stroke prevention. An MDT meeting should be mandated for patient evaluation and selection of intervention.
In patients with an acute stroke in whom a cardio-embolic mechanism is suspected the following potential causes should be excluded: AF or flutter, thrombi in the left ventricle (LV) or in the left atrium/left atrial appendage, or on a prosthetic cardiac valve, as well as other conditions such as mitral stenosis, atrial myxoma, intracardiac masses, or valvular vegetations. LV mural thrombi account for up to 10% of cardio-embolic strokes and are most often seen in patients who have had a prior extensive anterior MI, with segmental akinesis or dyskinesis. Usually, the development of a LV thrombus occurs between 24 h and 2 weeks after the onset of a MI, with an increased risk in patients with depressed LV function. The risk of stroke or systemic embolism in the absence of anticoagulation, is as high as 10–20% at 3 months, with the highest risk in the first weeks, , declining after 3 months, in parallel with evolution to an organized and fibrotic clot adhering to the endocardium. Transthoracic echocardiography (TTE) is traditionally the standard imaging technique for detecting LV thrombi in patients with acute IS. More recently, cardiac magnetic resonance (CMR) was found to be superior to TTE in detecting LV thrombus in patients with history of MI and LV dysfunction (LVEF < 50%), with cine-CMR and contrast-enhanced CMR offering the highest diagnostic yield, especially in case of mural or small LV thrombi. Since CMR is a time-consuming, expensive examination, not easily available in most centres, new approaches have been developed in the field of echocardiography, with ultrasound contrast agents significantly improving the diagnostic accuracy of TTE. Anticoagulation is absolutely indicated if a LV thrombus is detected and has to be integrated and combined with control of hypertension and other risk factors. For oral anticoagulation, VKAs have been traditionally recommended. For the DOACs, few data are available, limited to case reports and case series, and their use in this setting remains off-label.
There are multiple pathologies that may lead to an IS: diseases of the arteries and diseases of the heart, thrombosis-mediated and thrombosis-unrelated, embolic and small-vessel disease, atherosclerotic, and non-atherosclerotic ( ). The myriad of cardiac, vascular, haematologic, and other underlying aetiologies highlights that stroke is not caused by a single disease. Moreover, many prevalent cardiovascular comorbidities are risk factors for these pathways leading to stroke such as arterial hypertension, DM, dyslipidaemia, heart failure, smoking, obesity, and physical inactivity to name only the most prevalent. Hence, engaging in the prevention of a first or recurrent stroke is a challenging task that requires competency in the management of the complex interactions outlined above. To add further to the complexity of integrated care for stroke, there are many challenges when treating a patient with acute and subacute ischaemic or hemorrhagic stroke. Although only a minority of stroke patients arrive rapidly enough to hospital settings, timely recanalization of an occluded cerebral artery is crucial as it may markedly improve patient outcomes, especially for those patients treated with endovascular treatment, with one of the best number-needed-to-treat metrics in medicine. Recanalization treatments have advanced in recent years from the technically-simple intravenous administration of alteplase in the early 90 s to the modern innovative interventional endovascular procedures which are further enhanced by sophisticated imaging and artificial intelligence. Most stroke patients develop several typical complications of cardiovascular, infectious, metabolic, neurologic, and neurosurgical nature, which need to be handled in a holistic multi-system approach. Such treatment is best provided on specialized stroke units (indeed, the Action Plan for Stroke in Europe 2018–2030 targets to treat 90% or more of all patients with stroke in Europe in a dedicated stroke unit as the first level of care ) and the interdisciplinary care concepts of close monitoring and early treatment of multiple complications are the key principle of stroke unit care—treating not the stroke, but the patients with stroke. Simultaneously, the diagnostic quest for the underlying cause occurs in parallel with acute stroke management and incorporates sophisticated multimodal diagnostic tools that add to the complexity by bringing up puzzling diagnostic challenges and uncertainties. Clearly, a MDT is required to contribute specialized expertise from a range of specialties involved as individually required. It is evident that the events leading to a stroke develop over many years, even decades, before the patient is admitted acutely to a stroke care facility as the stroke risk factors silently build the cardiovascular pathology which will eventually cause the stroke. These cardiovascular risk factors are chronic conditions that persist in the stroke survivor, alter recovery, and increase the risk of recurrence. Moreover, the stroke patient is usually elderly and often has associated disabilities and comorbidities that complicate the successful re-integration into the community, and is one of the most complicated medical paths that a person may have to walk through ( ). Stroke is one of the most typical examples of a medical condition which extends horizontally beyond and across the boundaries of the traditional medical specialties and is best served by integrated, collaborative, inclusive, interdisciplinary teamwork. The ‘Stroke Continuum of Care’ has been outlined as the stepwise approach to prevention, treatment, and recovery for stroke. Even though patients may follow different trajectories within the Stroke Continuum, with some patients experiencing a once-in-a-lifetime TIA without any sequelae, whereas others suffer severe disabling strokes, eventually most patients will need the services of physicians from different specialties with stroke-specific expertise, as well as of other healthcare professionals including nurses, physical therapists, occupational therapists, speech therapists, social workers, and psychologists. In this context, the role of the stroke physician is central to assist the patient journey through this marathon. Not to substitute the related medical specialties, but rather to orchestrate them to combine the needed expertise from a range of specialties in comprehensive multidisciplinary concepts for state-of-the-art stroke care. Such multidisciplinary concepts should cover the entire course of the patients from acute event through emergency care to subacute (Stroke unit) care to long-term diagnostic workup and risk factor monitoring and treatment. The integrated model of stroke care is implemented in several national health care systems like in the United Kingdom , and Canada, but there is still heterogeneity globally and further harmonization is warranted. The recognition and certification of Stroke Medicine as an official subspecialty of the aforementioned specialties like is already the case in the United Kingdom (in which the Stroke Medicine programme is open to all trainees holding certification in one of the following medical specialty: acute internal medicine, cardiology, clinical pharmacology and therapeutics, general internal medicine, geriatric medicine, neurology and rehabilitation medicine), and the implementation of specific subspecialty curricula that are preferentially harmonized across the European Union and beyond will greatly facilitate this process. Additionally, the harmonization of training curricula for stroke interventionists is of paramount importance to increase patient access to timely recanalization, given the still limited availability of skilled medical personnel to cover the needs at population level. We encourage research focusing on the clinical efficacy of multidisciplinary post-acute stroke management clinics to prevent hard clinical outcomes like stroke recurrence, re-hospitalization, and major cardiac events, as well as their cost-efficiency.
COVID-19 The COVID-19 pandemic had a significant impact on the delivery of stroke care; major reorganizations were required to accommodate the rising numbers of COVID-19 admissions and redeployment of staff. , Despite the strong association with COVID-19 and IS, numbers of stroke admissions fell during the lockdown, possibly due to milder stroke patients staying home and excess stroke deaths in the community. , The manifestation of stroke as the first presentation of COVID-19 posed a particular transmission risk in the acute setting, with infection control measures rapidly embedded within the admission pathway to mitigate nosocomial transmission. , Complex, often young, patients were reported with multifocal large vessel occlusion, hypercoagulability, respiratory insufficiency, and multiple co-morbidities, especially diabetes and heart disease, contributing to increased length of hospital stay, case fatality, and high demand on the MDT. Overall, quality of care was preserved, and access to imagining and swallow assessment improved during the lockdown, but definitive hyperacute stroke intervention was significantly affected, particularly thrombolysis delivery. , Bottlenecks emerged, especially during the surge in transmission, specifically in moving patients along the stroke pathway, underscored by excess hospital admissions due to COVID-19, patients with stroke being managed on ‘COVID wards’ and an overwhelmed social support services in the community. , Notably, the acceleration of collaborative clinical and research networks, virtual meetings, and maximizing telemedicine throughout the whole stroke pathway, including pre-hospital, were opportunities that emerged from these challenging times that may have a lasting legacy. , Low-to-middle income countries Low-to-middle income countries (LMICs) currently harbour 70% of the global stroke burden but are the least equipped in providing stroke management. An ageing population, urbanization leading to an increased prevalence of modifiable risk factors such as hypertension, obesity, diabetes, and hypercholesterolaemia have an additive role. Regional factors such as infection (e.g. HIV) and air pollution also contribute in varying degrees. , Furthermore, primary and secondary stroke prevention are not fully optimized compared with high-income countries (HICs). Stroke care models in HICs have had a limited transition to LMICs due to the prohibitive costs and cultural variations. As a result, adaptations have arisen, for example, task shifting involving community healthcare workers and carers, hub and spoke models supported by telemedicine, physician and specialist-led stroke services but unsupported by region-specific evidence. The vast majority of LMICs have an unselected approach to stroke management, accounting for very poor outcomes. Simple yet effective interventions may have to be the focus in LMICs, invigorated by innovation and supported by research to demonstrate benefit.
The COVID-19 pandemic had a significant impact on the delivery of stroke care; major reorganizations were required to accommodate the rising numbers of COVID-19 admissions and redeployment of staff. , Despite the strong association with COVID-19 and IS, numbers of stroke admissions fell during the lockdown, possibly due to milder stroke patients staying home and excess stroke deaths in the community. , The manifestation of stroke as the first presentation of COVID-19 posed a particular transmission risk in the acute setting, with infection control measures rapidly embedded within the admission pathway to mitigate nosocomial transmission. , Complex, often young, patients were reported with multifocal large vessel occlusion, hypercoagulability, respiratory insufficiency, and multiple co-morbidities, especially diabetes and heart disease, contributing to increased length of hospital stay, case fatality, and high demand on the MDT. Overall, quality of care was preserved, and access to imagining and swallow assessment improved during the lockdown, but definitive hyperacute stroke intervention was significantly affected, particularly thrombolysis delivery. , Bottlenecks emerged, especially during the surge in transmission, specifically in moving patients along the stroke pathway, underscored by excess hospital admissions due to COVID-19, patients with stroke being managed on ‘COVID wards’ and an overwhelmed social support services in the community. , Notably, the acceleration of collaborative clinical and research networks, virtual meetings, and maximizing telemedicine throughout the whole stroke pathway, including pre-hospital, were opportunities that emerged from these challenging times that may have a lasting legacy. ,
Low-to-middle income countries (LMICs) currently harbour 70% of the global stroke burden but are the least equipped in providing stroke management. An ageing population, urbanization leading to an increased prevalence of modifiable risk factors such as hypertension, obesity, diabetes, and hypercholesterolaemia have an additive role. Regional factors such as infection (e.g. HIV) and air pollution also contribute in varying degrees. , Furthermore, primary and secondary stroke prevention are not fully optimized compared with high-income countries (HICs). Stroke care models in HICs have had a limited transition to LMICs due to the prohibitive costs and cultural variations. As a result, adaptations have arisen, for example, task shifting involving community healthcare workers and carers, hub and spoke models supported by telemedicine, physician and specialist-led stroke services but unsupported by region-specific evidence. The vast majority of LMICs have an unselected approach to stroke management, accounting for very poor outcomes. Simple yet effective interventions may have to be the focus in LMICs, invigorated by innovation and supported by research to demonstrate benefit.
Common concomitant diseases that may precede or accompany acute stroke and require timely cardiological workup and intervention besides AF are heart failure and ACS, as they often determine the patients’ outcomes. Patients with IS have a high 1-year risk of major adverse cardiovascular events with the highest risk within the first 30 days after the ischaemic event. , The extent of acute infarct size and stroke presentation often correlates with the severity of cardiac involvement. Apart from pre-existing CVD, heart failure can acutely occur as neurogenic stress or Takotsubo cardiomyopathy. In the echocardiographic examination or extended imaging by cardiac MRI they reveal reduced systolic function and/or wall motion abnormalities. , Electrocardiographic changes and arrhythmias (usually tachyarrhythmias, in particular AF) are common occurring in every fourth stroke patient, with the incidence being highest during the first day after stroke, and elevated cardiac enzymes including creatine kinase major bleeding or troponins and natriuretic peptides are observed frequently. In IS patients with no history of cardiac disease, the occurrence of signs of cardiac damage was more than 50%. The term stroke-heart syndrome , comprises the neurovisceral damage observed in context with cerebral injury. It can be predicted by clinical and routine laboratory variables, which usually identify older and sicker patients. Multiple causes for stroke-induced alterations in the neurocardiogenic axis have been discussed. Besides neurogenic, endocrine (e.g. hypothalamic-pituitary-adrenal axis), and psychological stress, an enhanced inflammatory and immune response, a temporary disruption of the blood–brain barrier and autonomic imbalance have been discussed. The interpretation of cardiac abnormalities in the context with acute stroke remains a challenge since alterations may be transient in nature with complete recovery. But, they may also lead to deterioration and chronic impairment. Whereas patients with high risk of adverse cardiac events need to be identified for targeted intervention, over-diagnosis, and treatment based on cardiac markers alone needs to be avoided. Treatment implications Tachyarrhythmic episodes may require prompt rate control or cardioversion if haemodynamically unstable. Commonly, betablockers, non-dihydropyridine calcium channel blockers respecting their negative inotropic actions, and digoxin with a delayed effect are used. Amiodarone can also be administered for rapid rate control and eventually cardioversion. Anticoagulation should be instituted as soon as deemed safe post stroke. Whereas heart failure therapy according to guidelines should be instituted upon diagnosis and carefully up-titrated balanced with infarct-related needs, e.g. optimized lower blood pressure boundaries, mechanical circulatory support in heart failure needs to be weighed in context with the overall holistic prognosis, and instituted after careful decision in the stroke-heart team. A frequent differential diagnosis in the context of markers of myocardial damage, e.g., ECG abnormalities up to ST segment elevations, enzyme elevations, and cardiac dysfunction that need timely attention is ACS. Invasive diagnostics and the potential need of dual antiplatelet therapy for conservative treatment of an ACS or after percutaneous coronary intervention require interdisciplinary decision-making and careful risk benefit assessment. However, a routine measurement of troponins on admission for acute stroke needs to be viewed critically. Using modern troponin assays, elevations are observed in more than 50% of admissions. Despite a higher risk of adverse outcomes, only a minority of patients have a classical ACS. Which patients qualify for invasive diagnostics and more aggressive ACS treatment is subject of the PRediction of Acute Coronary Syndrome in Acute Ischaemic StrokE study (NCT03609385).
Tachyarrhythmic episodes may require prompt rate control or cardioversion if haemodynamically unstable. Commonly, betablockers, non-dihydropyridine calcium channel blockers respecting their negative inotropic actions, and digoxin with a delayed effect are used. Amiodarone can also be administered for rapid rate control and eventually cardioversion. Anticoagulation should be instituted as soon as deemed safe post stroke. Whereas heart failure therapy according to guidelines should be instituted upon diagnosis and carefully up-titrated balanced with infarct-related needs, e.g. optimized lower blood pressure boundaries, mechanical circulatory support in heart failure needs to be weighed in context with the overall holistic prognosis, and instituted after careful decision in the stroke-heart team. A frequent differential diagnosis in the context of markers of myocardial damage, e.g., ECG abnormalities up to ST segment elevations, enzyme elevations, and cardiac dysfunction that need timely attention is ACS. Invasive diagnostics and the potential need of dual antiplatelet therapy for conservative treatment of an ACS or after percutaneous coronary intervention require interdisciplinary decision-making and careful risk benefit assessment. However, a routine measurement of troponins on admission for acute stroke needs to be viewed critically. Using modern troponin assays, elevations are observed in more than 50% of admissions. Despite a higher risk of adverse outcomes, only a minority of patients have a classical ACS. Which patients qualify for invasive diagnostics and more aggressive ACS treatment is subject of the PRediction of Acute Coronary Syndrome in Acute Ischaemic StrokE study (NCT03609385).
Stroke is an unexpected, and for most people, a life-changing event, and it is important to remember that every patients’ experience will be unique. The trajectory will differ between patients, and their perceptions and experiences will be influenced by the stroke sequelae, the care and treatment they receive, the ongoing support available to them, and their ability to adapt/adjust. For many patients and their family/caregivers, their main concern following a stroke is the risk of a recurrent event. Therefore, the emphasis is often on secondary stroke prevention, and lifestyle changes are necessitated; these must be individually tailored to the patient. However, for those with a severely disabling stroke, prevention of recurrence is generally a lower priority. For many patients finding the cause of their stroke is important to enable them to reduce their chances of recurrence. However, the aetiology of the stroke is often not immediately apparent and may require ongoing, and sometimes protracted, investigation to ascertain the cause; in about 30%, the cause will remain unknown, and this heightens anxiety and can significantly impair recovery. Stroke has a significant psychological impact as well as physical side-effects. Post-stroke depression and anxiety are common and can significantly impact patients’ ability, motivation, and engagement with essential lifestyle modifications (e.g. medication adherence, exercise, smoking cessation, dietary changes, etc.); for some, the emotional impact of the stroke may only become apparent several months later. Loss of independence, greater reliance on family and friends, inability to return to work, decreases in previous social/leisure activities, disability or reduced physical functioning, cognitive impairment, communication difficulties (reading, speaking, listening), fatigue, effect on relationships, and the financial impact of stroke can all detrimentally affect the patient, impacting their physical and mental health and reducing their quality of life (and that of their family). It is also imperative to consider the impact of socioeconomic factors, , such as health literacy, housing conditions, access to affordable and nutritious food, personal safety, transport, social support etc., on an individual’s capacity for resilience and ability to implement lifestyle changes necessary to promote recovery and reduce recurrence. Ongoing support, listening to the patient and providing personalized education, information, and advice in simple terms with consistent messages are important throughout the patient journey and continuity of care is essential. Integrated care for stroke advocated in this consensus document requires streamlining of care pathways and MDT working to optimize patient care and improve patient outcomes. Central to the success of integrated care is greater patient and family/caregiver involvement in planning and co-producing support packages and input and feedback on optimizing stroke care pathways.
Emerging evidence suggests that comprehensive pragmatic care pathways with continued post-hospital involvement of the multidisciplinary stroke team could reduce the longer-term health burden of stroke. Further high-quality studies are needed to inform sustainable long-term care pathways for long-term improvement in clinical and patient-reported outcomes among stroke patients.
The co-occurring and inter-linked nature of CVD requires an integrated care action plan to prevent, identify, treat, and rehabilitate people. This targets the prevention of recurrent stroke, improves patient functional status and symptoms, and manages cardiovascular risk factors, comorbidities, and lifestyle changes A post-stroke ABC pathway is proposed as a more holistic approach to integrated stroke care and would include three pillars of management: A: Appropriate Antithrombotic therapy. B: Better functional and psychological status. C: Cardiovascular risk factors and Comorbidity optimization (including lifestyle changes). Appropriate thromboprophylaxis should be targeted to the underlying comorbidity, for example, anticoagulation for patients with AF. When stroke patients have both AF and vascular disease, anticoagulation monotherapy would suffice. In high-risk stable atherosclerotic vascular disease patients without AF , combination therapy with rivaroxaban 2.5 mg bid and aspirin provides some benefits on CVD events (including on stroke) even in the absence of associated AF , but at the risk of more major bleeding. In the absence of AF, antiplatelet drugs are used for secondary stroke prevention for non-cardioembolic stroke with either aspirin and clopidogrel or ticagrelor. Both the combination of clopidogrel and aspirin and ticagrelor and aspirin have been found to be superior to aspirin alone for 90-day treatment of acute minor strokes and high-risk TIAs. The combination of clopidogrel and aspirin has also been utilized for intracranial atherosclerotic ISs. Better functional and psychological status requires care packages and management pathways to be established and implemented, including tailored rehabilitation and personalized appropriately to the patient’s needs, both in hospital, during rehabilitation, and post-discharge. Assessment of PSD, anxiety, and cognitive impairment should be undertaken as part of post-stroke care, with appropriate intervention where required. Cardiovascular risk factors and comorbidities are common in stroke patients, and all need to be considered and addressed in a holistic manner, with treatment targets as per CVD prevention guidelines. A considerable burden of previously unknown AF can be detected when long-term monitoring is applied in at-risk patients. Across the stroke continuum, there is a need for multi-disciplinary collaboration and coordination of care, including the complex treatment of cardiovascular conditions with the overarching goal to improve recovery, prevent recurrence, and enhance survival and quality of life for the patient with stroke. Competing disease presentations in the context with acute stroke remains a challenge and may also lead to deterioration and chronic impairment. Over-diagnosis and treatment based on cardiac markers alone needs to be avoided. Central to the success of integrated care for stroke is greater patient and family/caregiver involvement in planning and co-producing support packages, and input and feedback on optimizing stroke care pathways.
ehac245_Supplementary_Data
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.