Current challenges for clinical trials of cardiovascular medical devices.

IJCA-18193; No of Pages 8 International Journal of Cardiology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

Review

Current challenges for clinical trials of cardiovascular medical devices Faiez Zannad a,b,⁎, Wendy Gattis Stough c, Ileana L. Piña d, Roxana Mehran e,f, William T. Abraham g, Stefan D. Anker h, Gaetano M. De Ferrari i, Andrew Farb j, Nancy L. Geller k, Robert S. Kieval l, Cecilia Linde m, Rita F. Redberg n, Kenneth Stein o, Alphons Vincent p, Holger Woehrle q,r, Stuart J. Pocock s a

INSERM, Centre d'Investigation Clinique 9501 Unité 1116, Centre Hospitalier Universitaire, France Department of Cardiology, Université de Lorraine, Nancy, France Campbell University College of Pharmacy and Health Sciences, Buies Creek, NC, USA d Department of Medicine, Division of Cardiology, Montefiore Medical Center, Bronx, NY, USA e Cardiovascular Research Foundation, New York, NY, USA f Mount Sinai Medical Center, New York, NY, USA g Division of Cardiovascular Medicine, The Ohio State University, Columbus, OH, USA h Applied Cachexia Research, Department of Cardiology, Charité Medical School, Campus Virchow-Klinikum, Berlin, Germany i Department of Cardiology, Fondazione IRCCS Policlinico San Matteo, Pavia, Italy j U.S. Food and Drug Administration, Silver Spring, MD, USA k National Heart Lung and Blood Institute, Bethesda, MD, USA l CVRx, Inc., Minneapolis, MN, USA m Department of Cardiology, Karolinska University Hospital, Karolinska Institute, Stockholm, Sweden n University of California, San Francisco, CA, USA o Boston Scientific Corporation, St. Paul, MN, USA p Medtronic, Tolochenaz, Switzerland q ResMed Science Center, ResMed, Martinsried, Germany r Sleep and Ventilation Center Blaubeuren/Lung Center, Ulm, Germany s Department of Medical Statistics, London School of Hygiene and Tropical Medicine, London, United Kingdom b c

a r t i c l e

i n f o

Article history: Received 3 February 2014 Received in revised form 8 May 2014 Accepted 11 May 2014 Available online xxxx Keywords: Cardiovascular devices Device approval Clinical trial Research design

a b s t r a c t Several features of cardiovascular devices raise considerations for clinical trial conduct. Prospective, randomized, controlled trials remain the highest quality evidence for safety and effectiveness assessments, but, for instance, blinding may be challenging. In order to avoid bias and not confound data interpretation, the use of objective endpoints and blinding patients, study staff, core labs, and clinical endpoint committees to treatment assignment are helpful approaches. Anticipation of potential bias should be considered and planned for prospectively in a cardiovascular device trial. Prospective, single-arm studies (often referred to as registry studies) can provide additional data in some cases. They are subject to selection bias even when carefully designed; thus, they are generally not acceptable as the sole basis for pre-market approval of high risk cardiovascular devices. However, they complement the evidence base and fill the gaps unanswered by randomized trials. Registry studies present device safety and effectiveness in dayto-day clinical practice settings and detect rare adverse events in the post-market period. No single research design will be appropriate for every cardiovascular device or target patient population. The type of trial, appropriate control group, and optimal length of follow-up will depend on the specific device, its potential clinical benefits, the target patient population and the existence (or lack) of effective therapies, and its anticipated risks. Continued efforts on the part of investigators, the device industry, and government regulators are needed to reach the optimal approach for evaluating the safety and performance of innovative devices for the treatment of cardiovascular disease. © 2014 Elsevier Ireland Ltd. All rights reserved.

Abbreviations: CE, Conformité Européenne; CEC, Clinical Events Committee; EMA, European Medicines Agency; E.U., European Union; FDA, Food and Drug Administration; HRQOL, health-related quality of life; IDE, investigational device exemption; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; OMT, optimal medical therapy; OPC, objective performance criteria; PRO, patient reported outcomes; PROBE, prospective open with blinded evaluation; SCD-HeFT, Sudden Cardiac Death in Heart Failure Trial; STICH, Surgical Treatment for Ischemic Heart Failure; U.S., United STATES. ⁎ Corresponding author at: Coordinateur Général, Centre d'Investigation Clinique Inserm, CHU, Université de LorraineCoordinateur Scientifique EU FP7 HOMAGECoordinateur Scientifique EU FP7 FIBROTARGETS UMR 1116 Chef de l'Unité HTA et Insuffisance Cardiaque Pôle de Cardiologie Président, Réseau Insuffisance Cardiaque ICALOR Institut Lorrain du Coeur et des Vaisseaux CHU de Nancy 4, rue du Morvan54500 VANDOEUVRE-LES-NANCY, France. Tel.: +33 383 15 73 22; fax: +33 383 15 73 24. E-mail address: [email protected] (F. Zannad).

http://dx.doi.org/10.1016/j.ijcard.2014.05.021 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.

Please cite this article as: Zannad F, et al, Current challenges for clinical trials of cardiovascular medical devices, Int J Cardiol (2014), http:// dx.doi.org/10.1016/j.ijcard.2014.05.021

2

F. Zannad et al. / International Journal of Cardiology xxx (2014) xxx–xxx

1. Introduction

2.1. Blinding considerations

Numerous medical devices are available that prolong survival, decrease morbidity, reduce symptoms, and improve functional status and/or health-related quality of life (HRQOL) in patients across the spectrum of cardiovascular disease [1]. However, devices have different considerations than drugs, and the design of device clinical trials may not necessarily follow the patterns established for pharmacologic studies [2–4]. For example, in device clinical trials, blinding may be more challenging, making the use of subjective endpoints (e.g., quality of life or threshold for revascularization) less reliable, as they are more likely to show a powerful placebo effect, even from a sham procedure. Unlike drug studies, operator or procedural learning curves play an important role in device trials, which may result in poorer outcomes during the early phases of study. Patient selection, implant and surgical technique, and stratified follow-up are key elements for successful device therapy, but they may not be fully understood at the time a trial is designed. Fewer patients are usually enrolled in device trials than in drug trials; therefore, device trials are more prone to be underpowered for major morbid events and mortality. Conversely, inadequate patient adherence to the assigned intervention is not a problem in the evaluation of an implanted device, whereas it can be problematic in drug trials. As technology rapidly changes, new models of a device may be released during a trial, prior to the presentation of data for regulatory approval, or, in the case of devices already approved, new models can be approved via the pre-market approval (PMA) supplement process. Changes can also include updated software algorithms that dictate the response to certain measurements, which in reality, can significantly change the device function. There are no precise rules governing what degree of change is significant enough to require new clinical studies. Most approvals processed through the PMA supplement pathway follow a supplement category that does not require clinical data to support the approval [5]. This process is in contrast to the extensive pre-clinical and early phase trials that precede pivotal trials for pharmacologic therapies. Regulatory requirements for approval of medical devices in the United States are vastly different from those in Europe, causing difficulties in harmonizing research requirements globally. Finally, achieving reimbursement and ultimate adoption of a novel device may require data beyond that needed to assure safety and effectiveness. Given these factors, alternative trial designs and novel endpoints that accurately reflect safety and effectiveness are needed for the evaluation of cardiovascular devices. However, departure from the ideal randomized, double-blind, controlled trial to create a feasible environment for conducting trials and supporting innovation is associated with real concerns in terms of scientific validity and confidence in research results. The importance of these trade-offs should not be underemphasized. During the 9th Global Cardiovascular Clinical Trialists Forum held in Paris, France, in December 2012, alternative approaches to randomized, double-blind, controlled trials of cardiovascular devices were discussed, highlighting the strengths and limitations of the regulatory environment, and considering methods to achieve quality science while ensuring feasibility, patient safety, and encouraging innovation. This paper summarizes the key challenges facing cardiovascular device research and development on the path to regulatory approval.

Blinding reduces bias in clinical trials, and regulatory agencies encourage blinding to the fullest extent possible. Blinding a device trial may require a sham procedure (e.g., a procedure that simulates the device implantation, but no device is implanted), a sham device (which in most cases would violate ethical principles of research because of the lack of potential for individual benefit, although both sham procedures and devices can have a marked placebo effect which might restore the balance between potential risks and potential benefits), or implantation of a device without actuation. While sham approaches are possible for some device trials (e.g., renal artery nerve radiofrequency ablation to treat hypertension where all patients undergo renal angiography to determine anatomic eligibility and randomization and treatment is performed at the time of renal angiography [6] or cardiac resynchronization therapy), they are not for others (e.g., mechanical circulatory support). Sham procedures may be viewed by some as unethical because the risks associated with the procedure could outweigh any potential benefit gained from participating in the research. Additionally, devices are often linked to specific management or follow-up care that cannot be separated from the pure device effect. Lack of blinding has the potential to bias outcome assessment by overestimating the treatment effect [7]. Truly large treatment effects can overwhelm such bias, but if treatment effects are more modest, the potential bias makes the results difficult to interpret. Therefore, when blinding is not possible, studies should be designed with hard objective endpoints (e.g., all-cause mortality versus cause-specific mortality; all hospitalizations versus heart failure hospitalizations; all revascularization versus urgent revascularization). Any cause-specific or subjective endpoints should be adjudicated by a clinical events committee (CEC) that is blinded as to each patient's allocated treatment (i.e., prospective open with blinded evaluation [PROBE] design) [8]. It is important to note that the CEC may be unblinded by diagnostic tests (e.g., computed tomography [CT] scans or chest X-rays) where devices are evident. Special efforts to avoid unblinding in these situations (e.g., redaction of progress notes, exclusion of imaging reports from documentation provided to CEC members unless absolutely required to classify events) need to be undertaken. Open-label device trials can bias study subjects completing patient reported outcomes (PRO) or HRQOL questionnaires. Thus, such endpoints are usually inadequate as primary endpoints for pivotal trials. However, they are often important secondary or ancillary outcomes used by patients to make decisions about treatment preferences or by payers to support reimbursement decisions. Therefore, such endpoints should be rigorously collected and potential sources of bias acknowledged to facilitate data interpretation. Instruments that assess devicerelated burden independently from HRQOL may be more informative and reduce the influence of bias, since the assessment is more objectively determined. Instruments should be as device-specific as possible. Unfortunately, not many instruments have been developed or validated that capture specific device-related issues. Both the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) have released specific documents to guide the development of PROs to support regulatory claims [9–11]. Bias can also originate from unblinded investigators who may adjust therapies and alter patient management during an ongoing study [12]. The CardioMEMS Heart Sensor Allows Monitoring of Pressure to Improve Outcomes in New York Heart Association (NYHA) class III Heart Failure Patients (CHAMPION) study is a good example of a recent trial in which regulators and FDA Advisory Panel members were initially uncertain about how to interpret the data in the context of the potential confounding influence of unblinded study staff who made treatment recommendations to study investigators [13]. Even in a blinded trial, accidental unblinding can occur, which may affect care differentially in the treatment arms. Investigators may tend to report adverse events more frequently in the investigational arm of open-label trials, or they may

2. Pre-market clinical trial designs to support device approval Approval of therapeutic innovations should be based on a reasonable assurance of safety and effectiveness ideally demonstrated by randomized, controlled study designs that provide unbiased data for clinical evaluation and regulatory decision-making. However, there are multiple challenges to utilizing this ideal approach for studies of cardiovascular devices.



under-report adverse events if the investigator bias favors the study device. When it is not possible to blind patients or physicians, other study personnel (e.g., core labs, clinical research monitors, clinical research nurses, coordinating and data center personnel) should be blinded whenever possible. For device trials, the operator should not participate in study-related patient assessments after the initial procedure. When only single blinding is possible, independent adjudication of events by blinded CECs, and hard objective endpoints should enhance the quality of the results, but these measures do not completely mitigate the potential for bias. 2.1.1. Blinded placebo control A blinded placebo control group (that is treated with optimal medical therapy [OMT]) is the most rigorous comparator in a randomized trial (Table 1). The analogous control group in a device trial is a sham procedure plus OMT, although some obvious differences exist between placebo in a drug trial and sham in a device trial. In the case of a sham device, the implantation procedure itself may exert either a greater placebo effect or a greater negative effect on the patient's condition than would administration of a placebo “drug”; the magnitude of the placebo effect may be proportional to the invasiveness of the sham procedure. These clinical effects could impact the validity of conclusions drawn from trial results. For example, in a randomized trial that uses a shamdevice control group, patients who receive a device and are randomized to device “off” still experience the device implantation procedure and the presence of the device in the body, even though they do not experience active therapy. A between-group comparison between device “on” and device “off” patients only evaluates the effect of active therapy; it ignores possible effects of the implantation procedure and the device presence. A comparison of patients who receive a device to patients who do not receive a device may be required to assess the totality of risk versus benefit. Obtaining a post-procedure baseline assessment may be suggested to quantify the incremental benefit of device function above the immediate effect of the procedure, but this is not a true baseline. Since placebo effects often wane over time as chronic diseases such as cardiovascular disease progress, extended follow-up may help assess true treatment effects. Conversely, post-implantation factors including anesthesia effects, wound healing, pain, temporary increased adherence to background medical therapy, or the Hawthorne Effect [14] may skew post-procedural assessments, favoring a pre-procedural baseline. The inconvenience of the sham device or procedure without the potential benefits gained from active treatment may lead to a higher rate of

3

discontinuation (drop outs) or drop-ins (switch to “on” mode) in the sham arm, or it may negatively impact the chronically diseased patient, which raises problems with statistical power, data quality, safety, and the interpretation of study results. Thus, the effects of sham procedures on study outcomes may be unpredictable (Table 2). 2.1.2. Non-blinded optimal medical therapy control An open-label comparison of OMT to the investigational device is appropriate in circumstances where no device or surgical comparator exists, and a sham is not feasible or ethical. Examples include the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) of conventional therapy versus amiodarone versus an implantable cardioverter defibrillator in patients with NYHA class II or III heart failure and a left ventricular ejection fraction (LVEF) of ≤35% [15] and the Surgical Treatment for Ischemic Heart Failure (STICH) trial of medical therapy versus coronary artery bypass graft surgery in patients with heart failure, LVEF ≤35%, and coronary artery disease [16]. This approach can be used to evaluate a new device versus OMT or the incremental benefit of the new device when added to conventional best therapy. Endpoints should be rigorously defined and adjudicated by a blinded CEC to minimize the influence of bias. It is important to note that threshold for endpoints such as “urgent coronary revascularization” may be affected by such designs and should be avoided in non-blinded studies. For example, some patients randomized to OMT alone may crossover to intervention if the investigator has a low threshold for recommending invasive treatment. Some device trials, such as the Resynchronization Reverses Remodeling in Systolic Left Ventricular Dysfunction (REVERSE) study, have established stringent protocol procedures to minimize unnecessary crossovers (e.g., approval by multiple principal investigators or steering committee members) [17]. Less objective endpoints such as functional assessments and instruments such as HRQOL are particularly subject to bias in non-blinded trials. Further, functional assessments and HRQOL measures are often not performed in many patients leading to uninterpretable results [18]. Efforts that improve adherence to functional assessments reduce missing data and are important in such trials. 2.1.3. Non-blinded active device control In circumstances where an existing device is already a key component of the standard of care, an investigational device should be compared to a control group receiving the existing device. As with all positive control trials, a large sample size may be needed to demonstrate either superiority or non-inferiority. Achieving the necessary

Table 1 Types of randomized control groups. Control group

Description

Strengths

Limitations

OMT plus sham controla

Allows for blinding of device or procedure, particularly when subjective endpoints used

Minimizes bias, especially for subjective endpoints

OMT alone

When device or surgical comparator is not indicated due to efficacy or safety limitations and no active comparator exists When a similar drug or device is the standard of care

Tests against an optimal pharmacologic regimen to determine incremental benefit above currently known effective agents Tests against current standard of care

Ethical issues related to risk of harm with no chance of benefit; difficulties maintaining the blind, low threshold for crossover (e.g., to activate the device in the inactive arm), possible clinical effects of device implantation and physical presence that could confound trial results May not represent real-world since many patients do not receive OMT

When surgical therapy is the standard of care and subjects are acceptable surgical candidates

Tests against current standard of care

OMT plus active comparator

OMT plus surgical therapy

a

Requires large sample size to demonstrate superiority or non-inferiority; enrollment may be slow if approved alternative is available to patients who may then be reluctant to participate in study Requires a large sample size to demonstrate superiority or non-inferiority since between group difference is likely to be small; enrollment may be slow if approved alternative is available to patients who may then be reluctant to participate in the study

OMT = optimal medical therapy.


4


Table 2 Potential approaches to mitigate problems in device trials. Problem

Suggested approach and potential advantages

Potential disadvantages

Placebo effect for sham control arm

• The potential influence of immediate post-procedural factors (e.g., pain, wound healing, anesthesia, temporary increased compliance to background medical therapy, Hawthorne effect) should be considered. Establishing a post-procedural baseline after the patient has stabilized from the procedure may be warranted. • Consider randomizing to sham in only part of the control group, to quantify the placebo effect in a given trial. This approach might provide information about the influence of sham procedures in device trials, and it would contribute to the totality of information about the need for blinding. • Use primary endpoints that are less susceptible to bias and do not require adjudication, such as all-cause mortality. “Total” or “all-cause” endpoints are more objective endpoints and are preferred over subjective endpoints (e.g., all revascularization preferred over urgent revascularization; total hospitalization preferred over causespecific hospitalization). If cause-specific events are captured, then a blinded aCEC should be used for endpoint adjudication. • Instruments that assess device-related burden independently from bHRQOL may be preferrable. • Instruments should be administered by blinded study personnel, or be completed by the patient independently of study personnel in trials in which the patient is blinded but investigator is not. • HRQOL or cPROs should not be used as pivotal endpoints in open-label trials, except possibly in patient populations for whom no effective therapies exist. • A greater threshold for defining a “meaningful difference” in HRQOL or PRO instruments may be appropriate for open-label trials. • Routine addition or titration of therapies should be prespecified in the protocol, rather than at the discretion of the investigator to prevent differential use of therapies in the investigational and control arms. “Off-protocol” therapy adjustments should only occur if warranted by a change in the patient's clinical condition that is not addressed in the protocol. • Study personnel responsible for ascertaining and reporting adverse events should be blinded to the extent possible. • Physicians/operators who are not blinded to treatment assignment should not participate in study-related assessments or evaluations after the index procedure. • Registry source should be of superior quality, with minimal selection bias. • Pre-specified analysis plan for derivation of control group • Use of dOPC or performance goals rather than registrybased controls may be considered once sufficient data have accrued to generate the treatment goal.

• Post-procedural baseline is not a true baseline, since it is influenced by the fact that a procedure has occurred. • The treatment difference between the active arm and sham control should minimize concerns regarding the placebo effect, except that any potential benefit of the active arm might be underestimated because of the placebo effect, and sham devices may exert real effects. • Randomizing only part of the control group to a sham procedure would likely require a larger sample size, since it would create a 3-arm study.

Bias introduced to endpoint assessments in openlabel trials

Bias introduced to patient reported outcome or HRQOL endpoints in open-label trials

Investigator bias in open-label trials

Matching risk profiles for registry-derived control groups in non-randomized trials

a b c d

• Use of all-cause mortality may require a larger sample size

• Validated instruments that are accepted by regulatory agencies are lacking (i.e., to support an indication for improvement in HRQOL).

• Monitoring investigator adherence to these protocol specifications is difficult.

• Data quality in a registry may be problematic. • Registry data often need extensive data auditing for accuracy.

CEC = clinical events committee. HRQOL = health-related quality of life. PRO = patient reported outcome. OPC = objective performance criteria.

sample size in a reasonable time frame may be impossible for some cardiovascular devices leading to considerations of alternative (nonrandomized) designs, although this option is not preferred since the study results will be less reliable.

when the pre- and post-intervention periods are compared within each patient [19]. However, this study design may overestimate effectiveness since patients are more likely to undergo the intervention after experiencing a highly morbid clinical event.

2.1.4. Internal control In some extreme cases, it may be ethically difficult to have a control group subjected to no intervention if the intervention appears very beneficial. In this situation, the individual study subject prior to the intervention serves as his/her own control. For example, left cardiac sympathetic denervation has been performed in patients with arrhythmogenic cardiomyopathies with drug-refractory arrhythmias [19]. This intervention has been considered effective because of a dramatic reduction in the incidence of life-threatening arrhythmias,

2.2. Randomization Randomization reduces potential bias by ensuring that confounders are evenly distributed on average among treatment groups. Selection bias is problematic in the absence of randomization, which limits data interpretation and the confidence one has in the results. Despite the strengths of randomization, it is not feasible for some device trials (e.g., conditions with small candidate populations such as mechanical circulatory support for bridge-to-transplant) [20]. In addition,



randomized trials are time consuming, which may be incompatible with many device life-cycles; i.e., a new generation of the device may be engineered by the time the trial has concluded, rendering the technology studied obsolete [21]. For next-generation device iterations, equipoise sufficient to support randomization between old and new devices may be absent if evidence from laboratory, animal, or early phase clinical testing suggests that improvements in technology will lower adverse event rates or improve device effectiveness (e.g., less bleeding, improved reliability, and longer life of the device) [22]. Randomization is preferred whenever possible, and there are many examples of rigorously conducted randomized device trials [15,17, 22–24]. However, the FDA recognizes that circumstances exist where a randomized trial is not possible, and they can consider alternative designs that are “based on sound arguments that they will provide sufficiently robust data” [25]. However, the FDA does not provide specific examples of arguments to justify alternative study designs. Rather, FDA encourages early direct interactions with sponsors and utilizes a flexible approach to consider study design proposals on a case by case basis to determine whether an alternative design will likely generate sufficient evidence of device safety and effectiveness [2]. These study design recommendations may be device-specific; that is, an approach that is acceptable for one device type may not be applicable to other device types. 2.3. Alternative approaches 2.3.1. Non-randomized control groups Using a non-randomized control group is an alternative to the randomized, controlled trial, although it may introduce bias. Several approaches may be used, and each has its strengths and weaknesses (Table 1). Control groups derived from observational registry data have been used in device studies. For registries to serve in this capacity, the data collection procedures should be comprehensive, complete, and ideally concurrent. Even with rigorous selection procedures, achieving a population with a similar risk profile to those treated with the study device remains a challenge [26]. The Evaluation of the HeartWare Left Ventricular Assist Device for the Treatment of Advanced Heart Failure (ADVANCE) study used a control group derived from patients enrolled in the Interagency Registry for Mechanical Assisted Circulatory Support (INTERMACS). A pre-specified propensity score analysis was performed to achieve comparability of the derived INTERMACS control group to the treatment group [26]. The groups were similar for most characteristics, but more control patients were severely ill compared to investigational device patients [27]. The ADVANCE trial exemplifies the difficulties in using registryderived control groups. Even with sophisticated statistical methods, imbalances in patient characteristics may persist, and imbalances in unknown factors are not assured as they are in a randomized trial. Although contemporary registries confer significant advantages over historical controls because they are more likely to reflect current practice patterns, research questions extend quickly into new patient populations that may not yet be represented in such registries (e.g., mechanical circulatory support in less severely ill populations). Furthermore, registries may not collect subjective data (e.g., functional capacity and HRQOL) as rigorously as do randomized trials, and they may use different definitions for important adverse events, such as major bleeding. Objective performance criteria (OPC) are another alternative approach. According to the FDA, OPC refers “to a numerical target value derived from historical data from clinical studies and/or registries and may be used by FDA for the comparison of safety or effectiveness endpoints” [2]. OPC are developed for an established device technology using patient-level data collected from pooled data from all studies (preferably contemporary and representative) on a particular device. Performance goals derived from meta-analyses of the published literature are an alternative, less rigorous approach to developing control

5

data. Ventricular assist devices, surgical prosthetic heart valves, and some catheter-based interventions for peripheral vascular disease are examples of device categories where performance goals or OPC have been applied [20,21,28]. Using the OPC as a comparator substantially reduces the sample size because it assumes a fixed number with no variability [20]. It should be noted that performance goals or OPC are not viewed as adequate to support claims of superiority by regulators. Performance goals and OPC represent minimal acceptable outcomes, and the results of the study device either meet or do not meet the performance goals or OPC. 2.3.2. Observational studies Non-randomized observational studies that do not use pre-specified comparator data or test pre-specified statistical hypotheses are used as an alternative when randomized trials are not feasible. These studies may also provide data on device performance in a day-to-day clinical practice environment, outside the constraints of the typical randomized, controlled trial. A large, well-designed observational registry with structured data collection, augmented with a small randomized or observational study may meet the need for an indication expansion or a new device approval [29]. Due to their lower complexity and study burden, patient enrollment is often faster in observational studies, which partially offsets their limitations. However, the evidence generated by these studies is less robust than data from randomized trials and can raise questions about the benefits and risks of a new drug or device versus the current standard of care, and it is not fully informative with regard to the mechanisms of effectiveness. For example, clinical outcomes in subjects enrolled in observational studies may appear to be better than those seen in randomized trials [30], or they may differ completely [31]. Non-randomized, observational studies can provide useful data on procedural and material device safety and can provide early insights into potential device-related problems. Selection bias is a major drawback in observational studies. Considerations leading to a treatment choice for each patient may have a greater influence on outcome than the therapy itself, and even complex adjustment methods may be inadequate to resolve the influence of these factors. Although statistical methods (e.g., propensity scores) can help reduce bias, unmeasured confounders and the presence of residual confounding may still compromise the scientific validity of the study conclusions [32]. Another major limitation is the lack of a control group for comparison. 3. Post-marketing device studies 3.1. Post-marketing surveillance Post-marketing studies assess product performance and safety in day-to-day clinical practice settings, where patient populations, background therapies, important comorbidities, or other factors may differ from those reflected in pre-market clinical trials. Post-marketing surveillance can also provide long-term information that may be unavailable in a clinical trial, and it is also an important detection method for rare events. Fewer patients are usually enrolled in pre-market approval device trials than in drug trials [33]. Therefore, device trials often utilize composite primary endpoints to increase the total number of events, but these studies are often underpowered to assess differences between treatment groups for individual clinically important events. Thus, post-marketing surveillance studies can provide increased precision around point estimates for rates of key events such as mortality, stroke, and re-intervention. Regulators may have additional questions that were not fully addressed by the pivotal trials (but that do not warrant withholding approval), and post-marketing studies provide an opportunity to address these issues. The FDA has the authority to require postapproval studies for class III (highest-risk) devices as a condition of pre-market approval. Working with the FDA, sponsors may


6


prospectively design clinical trials conducted under an investigational device exemption (IDE) nested within post-approval studies that can generate data to support new indications for use. Non-randomized, observational studies are commonly used for post-market surveillance. They contribute to the evidence base for a given therapy, although inherent limitations (e.g., unmeasured confounders, selection bias) may prevent their findings from being definitive. They can enroll large numbers of patients, which enhances statistical power. Also, registries facilitate exploratory analyses of clinically relevant subgroups because of their large sample sizes and less restrictive enrollment criteria. Observational studies are an opportunity to confirm the findings of randomized controlled trials in patients that are representative of current clinical practice [34], and they may help to inform reimbursement and health policy decisions. Observational studies may also generate new hypotheses to test in future trials. Establishing common practices and standards for post-marketing surveillance are two major foci for medical device regulatory reform [29,35]. Analyzing time trends and interpreting the totality of evidence are difficult because of inconsistent approaches to post-marketing surveillance. Timely detection of unanticipated adverse events remains problematic even in the best designed observational registries, especially when adverse event reporting is voluntary. Promising results have been obtained from studies using automated tools for early detection of such events, particularly with device therapies capable of remote monitoring [36,37]. The FDA maintains vigilance over post-approval marketing studies and encourages groups of investigators and patients to report adverse events that were not noted prior to approval to FDA's MedWatch system. The FDA is making significant progress towards its goal of achieving comprehensive and timely post-market surveillance with the development of the Sentinel Initiative [38] and the implementation of the Unique Device Identifier (UDI) system. The Sentinel System will utilize active surveillance methods to augment FDA's existing post-market safety monitoring systems for drugs and devices. Data partners in the Sentinel System will include academic medical centers, health care systems, and health insurance companies. As envisioned, the Sentinel System will allow FDA to pose safety questions that have emerged from pre-market or post-market safety data sources (e.g., clinical trial data, spontaneous adverse event reports) to the participating data partner organizations. Analyses of electronic health care data are expected to greatly enhance FDA's monitoring of the safety of drugs and devices in near real-time, including boosting signals of serious adverse events (e.g., MI and unexpected death) and assessing events in key patient subgroups (e.g., the elderly, women, and minorities) throughout the US [39]. In September 2013, FDA issued a final rule establishing a system to adequately identify medical devices through their distribution and use. This rule requires the label of medical devices to include a UDI, unless the device is subject to an exception or alternative. The UDI system is expected to offer benefits for patients, the health care system, and the device industry. It will improve the visibility of devices as they traverse the distribution chain from the manufacturer through the point of patient use and provide an enhanced ability to quickly identify marketed devices when needed for recalls and reporting and analyzing adverse events. The system will be designed to document device use in electronic health records, clinical information systems, claims data sources, and registries. It is hoped that this information will aid in the ongoing assessments of the benefits and risks of medical devices. The use of UDI's on high-risk Class III devices will be required by September 2014 [40]. 3.2. Post-market comparative effectiveness In the U.S., the Congress has indicated that a comparative effectiveness trial with an active comparator is not necessarily required for FDA's assessment of safety and effectiveness. Rather, the evidence required may vary according to the characteristics of the device, its conditions of use, the existence and adequacy of warnings and other

restrictions, and the extent of experience with its use. While this approach may be justified, comparative effectiveness data are often needed to understand the benefits and risks of a new device in comparison to other devices or non-device alternatives, particularly for payers. Even for devices where comparative effectiveness has been established in a clinical trial, it is desirable to determine if benefits are consistent across all subgroups, or whether there are subsets in which benefits differ, particularly for costly interventions. Although subgroup analyses of clinical trials have limitations, subgroup findings that are consistent across several studies or meta-analyses are more convincing, and they may be sufficient to guide clinical decision-making or to justify hypothesis testing in a new clinical trial [41]. Post-marketing comparative effectiveness data are informative to characterize treatment responses in daily clinical practice. Women, the elderly, patients with significant comorbidities, and racial minorities are often underrepresented in clinical trials, and post-marketing registries contribute important data for such understudied groups. Further, patients in clinical device trials receive more optimized medical regimens than patients in usual clinical practice. Post-approval registries provide an opportunity to evaluate the effectiveness of a new technology under the conditions of routine practice. Medical devices offer unique challenges for comparative effectiveness research. Therapeutic success and complication rates are linked with operator experience, procedural techniques, and skilled followup [42]. Effectiveness is likely to improve over time as procedures are refined and experience accumulates. Also, agreement has not been reached on the optimal endpoints to assess comparative effectiveness. Should one focus only on morbidity and mortality? Can PROs or HRQOL endpoints be used, recognizing the challenge of achieving reliable, unbiased, and reproducible assessments of such measures? Consensus is needed on best practices to address these issues as they pertain to cardiovascular devices.

4. Regulatory pathways for device approval Global differences in the standards for regulatory approval are another challenge facing medical device research and development. A comprehensive review of regulatory requirements is outside the scope of this manuscript, but the processes have been reported in detail elsewhere [2,43]. Briefly, for device approval, the FDA requires demonstration of a reasonable assurance of safety and effectiveness for class III (highest risk) devices. Clinical studies are needed to establish safety and effectiveness unless the device is a modification of a previously approved device. In the latter case, the FDA in discussion with the sponsor, may then determine that a clinical study is not required, based on information about the approved device, the extent of changes in the modified device, an appropriate non-clinical assessment of the modified device, and the application of benefit–risk principles [44,45]. It may be difficult to determine the clinical impact of even small modifications without clinical studies, especially when there are multiple modifications. The regulatory process differs in the European Union (EU), where medical device approvals are processed by Notified Bodies who are designated by the government Competent Authority in each member country. The Notified Bodies generally determine approval (and grant the Conformité Européenne [CE] mark) based on reliability testing and a demonstration that the device performs as intended, rather than on safety and effectiveness data generated from a clinical trial [45]. Both systems have been criticized for different reasons, including being too burdensome, stifling innovation and delaying access to potentially life-saving therapy versus not being adequately rigorous, resulting in the exposure of patients to unnecessary risk and unsafe or ineffective devices. These concerns have led to efforts to reform medical device regulation [43,44,46–48]. Although consensus has not yet been reached on the exact nature of the reforms [49,50], one consequence will likely be a



greater alignment of EU and FDA approaches to medical device approval and surveillance. 5. Conclusion Cardiovascular device therapy differs in important ways from cardiovascular drug therapy, and these differences often necessitate a modified approach to clinical trials. For many cardiovascular devices, adherence to standard practices of blinding and control groups is not possible. However, accurate assessment of risk and benefit is equally, if not perhaps more important, for a device that is often permanently implanted. Both randomized trials and non-randomized observational registry studies play key roles in the overall life-cycle evaluation of medical devices. Prospective, randomized, controlled trials that are carefully designed and adequately powered provide the greatest strength of evidence on which pre-market approval, clinical, and policy decision making should be based. Investigators should be keenly aware of the potential for bias when blinding or randomization is not possible, and implement measures to reduce bias such as use of hard endpoints and blinding CEC members and other study personnel (Table 2). Highquality post-marketing registry surveillance studies complement the evidence base and help fill the gaps left unanswered by randomized clinical trials. No single solution or research design will be appropriate for every cardiovascular device or target patient population; limitations will exist regardless of the path chosen. A combination of different trial designs that complement each other might be a viable option. The type of trial, appropriate control group, optimal length of follow-up, and extent to which post-market observational studies should be used will depend on the specific device, its potential benefits, the target patient population and the existence (or lack) of effective therapies, and its anticipated risks. Researchers, regulators, and industry sponsors must continue to refine new approaches to determine the safety and effectiveness of innovative devices for the treatment of cardiovascular disease.

7

AstraZeneca, Boston Scientific, Covidien, CSL Behring, Janssen Pharmaceuticals, Maya Medical, Merck, Regado Biosciences, and Sanofi-Aventis. William Abraham: Biotronik (consulting fees as a member of the EchoCRT Executive Committee); CardioMEMS (consulting fees as a coprincipal investigator of the CHAMPION trial); Medtronic (consulting fees as a member of the REVERSE Trial Steering Committee); St. Jude Medical (consulting fees as principal investigator of the LAPTOP-HF trial); Abbott Vascular (consulting fees as a member of the COAPT Patient Eligibility Committee); Respicardia (consulting fees as principal investigator for the remede pilot study and general consulting); Sunshine Heart (consulting fees as co-principal investigator of the C-Pulse Pilot and Pivotal trials); and Cardiokinetix (consulting fees as co-principal investigator of the PARACHUTE IV pivotal trial). Stefan Anker: Lonestar (Steering Committee member), consultant to CardioMems, BioVentrix, Impulse Dynamics, and Bosch GmbH; and Medical Sensible (Advisory Board member). Gaetano DeFerrari: Boston Scientific (steering committee member of a cardiovascular device trial sponsored by Boston Scientific). Andrew Farb: Nothing to disclose. Nancy Geller: Nothing to disclose. Robert Kieval: CVRx (employee). Cecilia Linde: Medtronic (grant, consultant, speaker fees, principal investigator of REVERSE and MiracleEF); St. Jude (speaker fees); Cardiomems (consultant); and Respicardia (consultant). Rita Redberg: No financial conflicts to disclose. Other disclosures: Editor, JAMA of Internal Medicine; Commissioner of the Medicare Payment Advisory Commission; Chairperson of the Medicare Evidence Development and Coverage Advisory Committee; Member of the Medical Policy and Technology Assessment Committee; Member of the California Technology Assessment Forum; and Member of the Medical Advisory Panel of Bluecross and Blueshield. Kenneth Stein: Boston Scientific (employee and shareholder). Alphons Vincent: Medtronic (employee). Holger Woehrle: ResMed (employee). Stuart Pocock: Nothing to disclose.

Sources of funding Acknowledgment This work was generated from discussions during the Ninth Global Cardiovascular Clinical Trialists (CVCT) Forum held in Paris, France, in December 2012. CVCT was organized by the Clinical Investigation Center (CIC) Inserm, CHU, and University Henri Poincaré of Nancy, France and funded by an unrestricted educational grant from Association de Recherche et d'Information en Cardiologie (ARISC), a non-profit educational organization, in Nancy, France. ARISC had no involvement in preparation, review, or approval of the manuscript for publication. Relationships with industry Faiez Zannad: Boston Scientific (fees for serving as steering committee member), Resmed (fees for serving as steering committee member), Gambro (fees for serving as steering committee member), Bayer (fees for serving as steering committee member), Servier (advisory board), Pfizer (steering committee member), Novartis (steering committee member and advisory board), Takeda (steering committee member), and Janssen (steering committee member). Wendy Gattis Stough: Medtronic (consulting); INSERM, Centre d'Investigation Clinique, Centre Hospitalier Universitaire, Nancy, France (travel expense reimbursement to attend CVCT 2012 and professional time related to preparation of this paper). Ileana Piña: Consultant and salary support from Food and Drug Administration, Center for Devices and Radiological Health. Roxana Mehran: Institutional research grants from The Medicines Company, Bristol-Myers Squibb/Sanofi-Aventis, and Lilly/Daiichi Sankyo; Consulting and/or Advisory Board fees from Abbott Vascular,

The following individuals were speakers or panelists discussing the topic of this manuscript at the December 2012 9th Global Cardiovascular Clinical Trialists Forum, Paris, France: Mark Carlson, John Jarcho, Alice Mascette, and Jay Yadav.

References [1] McMurray JJ, Adamopoulos S, Anker SD, et al. ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure 2012: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association (HFA) of the ESC. Eur Heart J 2012;33:1787–847. [2] U.S.Department of Health and Human Services, Food and Drug Administration, Center for Devices and Radiological Health, Center for Biologic Evaluation and Research. Design considerations for pivotal clinical investigations for medical devices: guidance for industry, clinical investigators, institutional review boards, and Food and Drug Administration staff. http://www.fda.gov/downloads/MedicalDevices/ DeviceRegulationandGuidance/GuidanceDocuments/UCM373766.pdf; 11-7-2013. [3] Campbell G. Statistics in the world of medical devices: the contrast with pharmaceuticals. J Biopharm Stat 2008;18:4–19. [4] Schoen F, Zuckerman B, Farb A. Device development for cardiovascular therapeutics: concepts and regulatory implications. In: Antman E, editor. Cardiovascular therapeutics: a companion to Braunwald's heart disease. 4th ed. Elsevier; 2013. p. 41–52. [5] Rome BN, Kramer DB, Kesselheim AS. FDA approval of cardiac implantable electronic devices via original and supplement premarket approval pathways, 1979–2012. JAMA 2014;311:385–91. [6] Kandzari DE, Bhatt DL, Sobotka PA, et al. Catheter-based renal denervation for resistant hypertension: rationale and design of the SYMPLICITY HTN-3 trial. Clin Cardiol 2012:35. [7] Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias. Dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA 1995;273:408–12.


8


[8] Zannad F, Garcia AA, Anker SD, et al. Clinical outcome endpoints in heart failure trials: a European Society of Cardiology Heart Failure Association consensus document. Eur J Heart Fail 2013;15:1082–94. [9] Burke LB, Kennedy DL, Miskala PH, Papadopoulos EJ, Trentacosti AM. The use of patient-reported outcome measures in the evaluation of medical products for regulatory approval. Clin Pharmacol Ther 2008;84:281–3. [10] Committee for Medicinal Products for Human Use (CHMP). Reflection paper on methodological issues in confirmatory clinical trials planned with an adaptive design. http://www.emea.europa.eu/pdfs/human/ewp/245902enadopted.pdf; 2007. [11] US Department of Health and Human Services Food and Drug Administration. Guidance for industry: patient-reported outcome measures: use in medical product development to support labeling claims. http://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/UCM193282.pdf; 2009. [12] Schulz KF, Grimes DA. Blinding in randomised trials: hiding who got what. Lancet 2002;359:696–700. [13] Loh JP, Barbash IM, Waksman R. Overview of the 2011 Food and Drug Administration Circulatory System Devices Panel of the Medical Devices Advisory Committee Meeting on the CardioMEMS Champion Heart Failure Monitoring System. J Am Coll Cardiol 2013;61:1571–6. [14] Braunholtz DA, Edwards SJ, Lilford RJ. Are randomized clinical trials good for us (in the short term)? Evidence for a “trial effect”. J Clin Epidemiol 2001;54:217–24. [15] Bardy GH, Lee KL, Mark DB, et al. Amiodarone or an implantable cardioverterdefibrillator for congestive heart failure. N Engl J Med 2005;352:225–37. [16] Velazquez EJ, Lee KL, Deja MA, et al. Coronary-artery bypass surgery in patients with left ventricular dysfunction. N Engl J Med 2011;364:1607–16. [17] Linde C, Abraham WT, Gold MR, St John SM, Ghio S, Daubert C. Randomized trial of cardiac resynchronization in mildly symptomatic heart failure patients and in asymptomatic patients with left ventricular dysfunction and previous heart failure symptoms. J Am Coll Cardiol 2008;52:1834–43. [18] Mann DL, Acker MA, Jessup M, Sabbah HN, Starling RC, Kubo SH. Clinical evaluation of the CorCap cardiac support device in patients with dilated cardiomyopathy. Ann Thorac Surg 2007;84:1226–35. [19] Wilde AA, Bhuiyan ZA, Crotti L, et al. Left cardiac sympathetic denervation for catecholaminergic polymorphic ventricular tachycardia. N Engl J Med 2008;358:2024–9. [20] Parides MK, Moskowitz AJ, Ascheim DD, Rose EA, Gelijns AC. Progress versus precision: challenges in clinical trial design for left ventricular assist devices. Ann Thorac Surg 2006;82:1140–6. [21] Grunkemeier GL, Jin R, Starr A. Prosthetic heart valves: objective performance criteria versus randomized clinical trial. Ann Thorac Surg 2006;82:776–80. [22] Slaughter MS, Rogers JG, Milano CA, et al. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med 2009;361:2241–51. [23] Rose EA, Gelijns AC, Moskowitz AJ, et al. Long-term use of a left ventricular assist device for end-stage heart failure. N Engl J Med 2001;345:1435–43. [24] Smith CR, Leon MB, Mack MJ, et al. Transcatheter versus surgical aortic-valve replacement in high-risk patients. N Engl J Med 2011;364:2187–98. [25] Chen E, Sapirstein W, Ahn C, Swain J, Zuckerman B. FDA perspective on clinical trial design for cardiovascular devices. Ann Thorac Surg 2006;82:773–5. [26] Aaronson KD, Slaughter MS, Miller LW, et al. Use of an intrapericardial, continuousflow, centrifugal pump in patients awaiting heart transplantation. Circulation 2012;125:3191–200. [27] Stevenson LW, Hunt SA. A bridge far enough? Circulation 2012;125:3069–72. [28] Geraghty PJ, Matsumura JS, Conte MS. Premarket assessment of devices for treatment of critical limb ischemia: the role of objective performance criteria and goals. J Vasc Surg 2009;50:1459–61. [29] Food and Drug Administration. Strengthening our national system for medical device postmarket surveillance. http://www.fda.gov/aboutFDA/CentersOffices/ OfficeofMedicalProductsandTobacco/CDRH/CDRHReports/ucm301912.htm; 2012.

[30] Dahabreh IJ, Sheldrick RC, Paulus JK, et al. Do observational studies using propensity score methods agree with randomized trials? A systematic comparison of studies on acute coronary syndromes. Eur Heart J 2012;33:1893–901. [31] Hulley S, Grady D, Bush T, et al. Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. Heart and Estrogen/Progestin Replacement Study (HERS) Research Group. JAMA 1998;280:605–13. [32] Yue LQ. Statistical and regulatory issues with the application of propensity score analysis to nonrandomized medical device clinical studies. J Biopharm Stat 2007;17:1–13. [33] Dhruva SS, Bero LA, Redberg RF. Strength of study evidence examined by the FDA in premarket approval of cardiovascular devices. JAMA 2009;302:2679–85. [34] Mauri L. Why we still need randomized trials to compare effectiveness. N Engl J Med 2012;366:1538–40. [35] Laskey W, Awad K, Lum J, et al. An analysis of implantable cardiac device reliability. The case for improved postmarketing risk assessment and surveillance. Am J Ther 2012;19:248–54. [36] Hauser RG, Mugglin AS, Friedman PA, et al. Early detection of an underperforming implantable cardiovascular device using an automated safety surveillance tool. Circ Cardiovasc Qual Outcomes 2012;5:189–96. [37] Resnic FS, Gross TP, Marinac-Dabic D, et al. Automated surveillance to detect postprocedure safety signals of approved cardiovascular devices. JAMA 2010;304:2019–27. [38] Platt R, Wilson M, Chan KA, Benner JS, Marchibroda J, McClellan M. The new Sentinel Network—improving the evidence of medical-product safety. N Engl J Med 2009;361:645–7. [39] U.S.Department of Health and Human Services, U.S.Food and Drug Administration, Center for Drug Evaluation and Research, Office of Medical Policy. The Sentinel Initiative: an update on FDA's progress in building a national electronic system for monitoring the postmarket safety of FDA-approved drugs and other medical products. http://www.fda.gov/downloads/safety/fdassentinelinitiative/ucm233360. pdf; 2010. [40] Unique device identification system. Final rule. Fed Regist 2013;78:58785–828. [41] Sipahi I, Carrigan TP, Rowland DY, Stambler BS, Fang JC. Impact of QRS duration on clinical event reduction with cardiac resynchronization therapy: meta-analysis of randomized controlled trials. Arch Intern Med 171, 2011, 1454-1462. [42] Altman RK, Parks KA, Schlett CL, et al. Multidisciplinary care of patients receiving cardiac resynchronization therapy is associated with improved clinical outcomes. Eur Heart J 2012;33:2181–8. [43] Fraser AG, Daubert JC, Van de Werf F, et al. Clinical evaluation of cardiovascular devices: principles, problems, and proposals for European regulatory reform. Report of a policy conference of the European Society of Cardiology. Eur Heart J 2011;32:1673–86. [44] Challoner DR, Vodra WW. Medical devices and health—creating a new regulatory framework for moderate-risk devices. N Engl J Med 2011;365:977–9. [45] Kramer DB, Xu S, Kesselheim AS. Regulation of medical devices in the United States and European Union. N Engl J Med 2012;366:848–55. [46] Cingolani E, Stevens S, Shehata M, Diamond GA, Kaul S. Medical device regulatory reform: insights from the watchman left atrial appendage closure technology for stroke prophylaxis in atrial fibrillation. Arch Intern Med 2011;171:1670–2. [47] Institute of Medicine. Medical devices and the public's health: the FDA 510(k) clearance process at 35 years. The National Academies Press; 2011. [48] Thompson M, Heneghan C, Billingsley M, Cohen D. Medical device recalls and transparency in the UK. BMJ 2011;342:d2973. [49] Curfman GD, Redberg RF. Medical devices—balancing regulation and innovation. N Engl J Med 2011;365:975–7. [50] Feder N. Medical devices—balancing regulation and innovation. N Engl J Med 2012;366:280.


Clinical research challenges in the era of cardiovascular medical devices.

Current Status and Challenges of Cancer Clinical Trials in Korea.

Medical devices for the anesthetist: current perspectives.

Current-limiting challenges for all-spin logic devices.

Pathologic Evaluation of Cardiovascular Medical Devices.

Implementation of clinical research trials using web-based and mobile devices: challenges and solutions.

Specific barriers to the conduct of randomised clinical trials on medical devices.

Event Rates in Randomized Clinical Trials Evaluating Cardiovascular Interventions and Devices.

Pivotal clinical trials of novel ophthalmic drugs and medical devices: retrospective observational study, 2002-2012.

Pediatric ventricular assist devices: current challenges and future prospects.

Muscular dystrophy: new challenges and review of the current clinical trials.

Methodological choices for the clinical development of medical devices.

Post-approval studies in France, challenges facing medical devices.

Molecular landscape of pancreatic cancer: implications for current clinical trials.

Do the results of randomized clinical trials of cardiovascular drugs influence medical practice? The SAVE Investigators.

Online Journal of Current Clinical Trials.

Editorial: Medical devices and the clinical chemist.

Current clinical trials for advanced osteosarcoma and soft tissue sarcoma.

Current clinical challenges in prostate cancer.

Uterine fibroids and current clinical challenges.

Trends in Consent for Clinical Trials in Cardiovascular Disease.

Clinical trials: past, current, and future for atypical Parkinsonian syndromes.

Current Immunotherapies for Sarcoma: Clinical Trials and Rationale.

Opportunities and challenges for the development of polymer-based biomaterials and medical devices.