Author + information
- Emily P. Zeitler, MD∗ ( and )
- Eric D. Peterson, MD, MPH
- ↵∗Reprint requests and correspondence:
Dr. Emily P. Zeitler, Duke University Clinical Research Institute, 2400 Pratt Street, Durham, North Carolina 27705.
Calls for increased transparency in the reporting and analysis of clinical trials have become commonplace (1). These have generally focused on “opening up” trial databases so that external investigators can revalidate the primary study findings and/or use the trial data for secondary hypothesis generation (2,3). At its core, however, the major goal of the open-science movement is to glean as much knowledge as is possible from every trial. This aim is certainly laudable, but this learning process may not necessarily require outside investigation. In fact, in this issue of JACC: Heart Failure, the members of the Study Steering Committee from the COMPANION (Comparison of Medical Therapy, Pacing and Defibrillation in Heart Failure) trial (4) “open up” and share reflections from their own trial experiences as well as new insights and hypotheses based on subsequent trial analysis.
It is now accepted that cardiac resynchronization therapy (CRT) with or without a defibrillator can both be lifesaving and impart improvement in functional outcomes in appropriately selected patients. However, this was not fully known or appreciated prior to the COMPANION trial. This pivotal trial’s large size and rigorous methods demonstrated without a doubt that CRT could improve heart failure outcomes and paved the way for broad acceptability of CRT. Now, Bristow et al. (4) have pulled back the curtain on COMPANION even further to provide fascinating insights into how the trial was designed, conducted, and analyzed. In so doing, they have provided important lessons regarding the challenges faced by the COMPANION investigators that remain highly relevant to the design and conduct of modern clinical trials.
When COMPANION was devised, CRT was a nascent if not an innovative technology. To fully understand and appreciate its potential benefits and applications, the trial investigators and sponsors incorporated input from the U.S. Food and Drug Administration to propose a trial that was designed and powered to assess mortality in addition to functional outcomes. Although this was highly commendable at its inception, this collaborative process did not consider the fact that competing CRT manufacturers were designing and implementing parallel CRT trials that were smaller in size because they sought to gain market approval for CRT with a pacemaker (CRT-P) and then CRT with a defibrillator (CRT-D) on the basis of functional endpoints alone (5). As the investigators demonstrate, approvals of these rival devices would devastate enrollment in COMPANION, at least in part because of loss of clinical equipoise (see Figure 2 in Bristow et al. ). After 3 competing devices were approved, enrollment in COMPANION dropped from >100 patients/month to 10, and crossover from the optimal pharmacological therapy (OPT) arm to the device arms was rampant. These challenges faced by COMPANION should serve as a call to action regarding the regulatory decisions for a class of devices. On one hand, all manufacturers could have been required to meet a similarly high standard for approval (e.g., effectiveness against a hard endpoint in addition to functional outcomes). If so, it is unlikely that COMPANION would have seen such enrollment delays or extensive crossover. Alternatively, COMPANION (and other CRT studies) could have included distinct phases, with the first phase designed to gain market approval on the basis of effect on functional outcomes and a subsequent post-market phase to continue patient accrual until enough clinical outcomes had occurred to allow analyses of hard endpoints. Finally and most powerful (but least plausible in the present competitive market environment), the competing CRT manufacturers could have worked collaboratively to evaluate the class effect of CRT-D versus CRT-P versus OPT in various populations.
The investigators describe how they overcame potential loss to follow-up challenges. Specifically, vital status was obtained from withdrawn patients by revisiting the consent processes, allowing contact with those patients who dropped out of treatment to at least obtain vital status. However, such rescue was time intensive. Thus the investigators suggest that future trials include prospective informed consent for determining vital status even when patients withdraw.
The COMPANION investigators acknowledge some internal trial challenges that are somewhat unique to new and evolving device technologies. For example, they report the impact the operator learning curve had on patient safety and clinical trial progress. Had this not been identified early, the complications could have led investigators and sponsors to halt the trial. However, swift recognition by the Data and Safety Monitoring Committee of procedure-related complications led to enhanced operator training. These experiences clearly demonstrate the importance of a vigilant data and safety monitoring committee.
Secondary and Subgroup Analyses
The COMPANION dataset also provides rich soil for hypothesis generation and testing. As Bristow et al. (4) note, the population was an “unambiguously advanced, high risk” heart failure population, so the secondary analyses presented give us the opportunity to test some closely held axioms of device therapy in heart failure. For one, CRT-P was superior to OPT for all endpoints that included hospitalization, but the effect sizes were greatest for cardiovascular endpoints. For example, on a background of OPT, CRT-P reduced the hazard of all-cause death or hospitalization by 18%, but the hazard of the combined endpoint of cardiovascular mortality or heart failure hospitalization was reduced by 36%. This finding supports the mechanism of benefit of CRT as specific to the sequelae of advanced heart failure.
Second, the investigators demonstrate that the greatest effect of CRT was seen in those patients with the most neurohumoral inhibition, suggesting that CRT may have an additive and/or synergistic interaction with neurohumoral inhibition. This highlights the complementary roles that drugs and devices play in treating heart failure and underscores the importance of multispecialty collaboration for the management of patients with heart failure.
Third, Bristow et al. (4) confirm the beneficial effect of CRT-D for the prevention of sudden cardiac death in an advanced heart failure population with reduced left ventricular ejection fractions. However, the investigators highlight an apparent increased risk for pump failure in CRT-D patients compared with CRT-P patients on a background of double neurohumoral inhibition (beta-blockers and angiotensin-converting enzyme inhibitors or angiotensin receptor blockers). As evidence for this, the investigators point to the increased number of cardiac transplantations that occurred in the CRT-D arm (n = 15) versus the CRT-P (n = 6) and OPT (n = 5) arms. It is proposed that defibrillations in the CRT-D arm contributed to increased pump failure, thus partially nullifying CRT-D benefits against sudden cardiac death. On the basis of this reasoning, the investigators suggest that CRT-P may be noninferior to CRT-D for mortality reduction in the studied population. However, contemporary CRT and implantable cardioverter-defibrillator technologies, along with modern implantation techniques and programming, reduce inappropriate therapies, thus minimizing potential deleterious effects of CRT-D on pump function over CRT-P (6). Nonetheless, ethical constraints limit the ability to conduct a CRT-P versus CRT-D randomized noninferiority trial on the basis of our current understanding and application of primary prevention defibrillators in this population, which was not established at the time of COMPANION.
A decade after reporting the main findings from COMPANION, the present report by Bristow et al. (4) demonstrates how much there still is to be learned from the trial conduct and the data it generated. In this instance, the investigators have given us context with which we may deepen our understanding of this landmark clinical trial and the breakthrough technology at its center. Moving forward, collaborations with trialists, patients, clinical experts, device manufacturers, and regulatory bodies may broaden our quest to apply the lessons learned from COMPANION to the evolution of a nimble and adaptable cardiovascular device clinical trial paradigm that values transparency from both external validation of results and honest reporting of challenges faced by trial insiders.
↵∗ Editorials published in JACC: Heart Failure reflect the views of the authors and do not necessarily represent the views of JACC: Heart Failure or the American College of Cardiology.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2016 American College of Cardiology Foundation
- Committee on Strategies for Responsible Sharing of Clinical Trial Data,
- Board on Health Sciences Policy,
- Institute of Medicine
- Ross J.S.,
- Lehman R.,
- Gross C.P.
- Bristow M.R.,
- Saxon L.A.,
- Feldman A.M.,
- Mei C.,
- Anderson S.A.,
- DeMets D.L.
- Moynahan M.,
- Faris O.P.,
- Lewis B.M.