Reflections on early stopping of a clinical trial
Article Outline
Large mortality-based trials in cardiovascular disease have made a major impact on the care of afflicted patients over the past 2 decades. Several developments have conspired to make their conduct and completion increasingly complicated. These include greater emphasis on efficient clinical care, the advent of multiple evidence-based therapies as background treatments, an increased regulatory burden, and intense competition for eligible patients to participate in trials. Moreover, the enormous expense required to bring new drugs to the market and the paucity of novel products in the investigative pipeline have stimulated consolidation of multinational pharmaceutical companies. Hence, well more than half of the new products that are now generated are derived from 10% of the potential industrial sponsors.
An important new development in the clinical trials arena has been the requirement by the consort group of medical editors to register all relevant elements of clinical trials before they commence.1 Failure to do so in the current era will either preclude or seriously impair the capacity to publish the results in a high-quality peer-reviewed journal. This development has evolved in part to reduce the likelihood that negative trials, especially those sponsored by the industry, will go unreported and that any protocol modification of trials in progress will acquire increased transparency. Notable lapses in the reporting of important data (eg, in the cyclooxygenase-2 inhibitor experience) have intensified the desire for increased surveillance and transparency of all elements of clinical trials.2
An additional issue associated with commercially sponsored clinical trials has been the desire of sponsors to terminate them before their originally planned conclusion.3, 4, 5 Although most contractual clauses associated with the construct of such trials identify the right of the sponsor to discontinue the trial for administrative reasons at any time, the appropriateness of executing this option is both contentious and complex.
Our purpose is to provide a perspective on this issue in the context of our current engagement in a clinical trial that will soon end before its planned conclusion. At the time of this writing, we remain blinded to the treatment assignments and have no knowledge of the results.
The APEX-AMI trial is a randomized, double blind, and placebo-controlled clinical trial of pexelizumab that intended to enroll 8500 patients undergoing percutaneous coronary intervention for ST-elevation myocardial infarction. Patients are randomized within 6 hours of symptom onset to a novel humanized monoclonal antibody that binds specifically to the C5 complement (pexelizumab) or to placebo; the primary efficacy analysis is based on all-cause mortality at 90 days. The design details of the trial have been previously published and are registered on www.clinicaltrials.gov.6 The APEX-AMI trial represents a major academic industrial collaboration and has randomized 5745 patients from 17 countries at approximately 300 sites. The first patient was enrolled on July 13, 2004; the last patient, on May 11, 2006. The APEX-AMI trial already stands as the largest trial ever conducted with patients with ST-elevation myocardial infarction for whom primary percutaneous coronary intervention is the reperfusion strategy.
In September 2005, when 25% of the anticipated total number of events had been accumulated (at the first planned interim review by the data safety monitoring board), there were 132 deaths derived from a total of 3092 randomized patients, which translated to an overall composite mortality of 4.3%. This overall mortality was substantially lower than that anticipated, given that we expected a 6.5% mortality in the placebo group and a 24% relative reduction achieved by treatment resulting in a 4.9% mortality in patients receiving pexelizumab. Given the mortality data in the APEX-AMI trial, which was designed as an event-driven trial, the number of patients required to achieve its planned 487 deaths was projected to exceed 11
000 patients. As we made efforts to enrich the patient cohort by encouraging enrollment of higher-risk patients at entry (especially restricting those with inferior myocardial infarction to patients >60 years), the sponsors concluded a parallel investigation of pexelizumab in 4250 patients undergoing coronary bypass surgery in the PRIMO-CABG II trial. To fulfill their fiduciary responsibility, in a November 2005 press release, Alexion Pharmaceuticals (one of the sponsors) reported that the PRIMO-CABG II trial did not achieve its primary end point.7 The failure of the PRIMO-CABG II trial, which was intended to be a definitive confirmatory study likely leading to regulatory approval, to meet the primary end point was a major disappointment to both Alexion Pharmaceuticals, a biotechnology company that created pexelizumab, and its development alliance partner, Procter & Gamble Pharmaceuticals. Moreover, these results raised serious concerns about the commercial viability of pexelizumab. One option at this juncture that received serious consideration by the sponsors was to immediately discontinue further development of pexelizumab. After extensive dialogue and encouragement by the executive academic steering committee to continue enrollment for as long as possible, the sponsors (Alexion Pharmaceuticals in partnership with Procter & Gamble Pharmaceuticals) decided in January 2006 to continue the APEX-AMI trial only to the point at which 50% of the anticipated events had been acquired and before its planned total enrollment.
Although none of the prior trials evaluating pexelizumab reached a statistically significant reduction in mortality alone, a pooled analysis of all 5 placebo-controlled randomized trials of pexelizumab (N = 9233) in both patients with acute myocardial infarction and those undergoing coronary artery bypass surgery revealed a 24% relative risk reduction (P = .009) of mortality at 30 days.8 The APEX-AMI trial sponsors responded to the argument that the APEX-AMI trial should continue at least until approximately half of the anticipated number of events was accumulated, thereby providing 50% power to confirm a 24% relative reduction in mortality if it was present. The decision to conclude the APEX-AMI trial before its anticipated end was then communicated to the Food and Drug Administration, which participated in an extensive dialogue regarding the probabilities of a positive result in the APEX-AMI trial. In particular, the potential impact of prior evidence regarding the survival benefit of pexelizumab was considered especially because it might influence the interpretation of the APEX-AMI trial results. Would, for example, a strong trend in the APEX-AMI trial results combined with a highly significant effect when the APEX-AMI trial and all prior trials were pooled justify a regulatory submission to the Food and Drug Administration to explore the approvability of pexelizumab?
Other considerations regarding the potential future development of pexelizumab were that there was no apparent safety concern associated with its use in any prior trial and that it represented a novel product to address a heretofore unmet need. Because the prior trials focused on a primary end point at 30 days, the original primary end point of the APEX-AMI trial was altered from 90 to 30 days. Importantly, 90 days would remain as the principal secondary end point in the APEX-AMI trial for assessment of other morbid events that would continue to be acquired and adjudicated by an independent blinded adjudication committee as previously planned. Because the data safety monitoring board has been unblinded at the time of its first review, it was informed but not consulted about the decision to cease enrollment.
Some of the most important underpinnings of the APEX-AMI trial are key mechanistically based substudies designed to characterize the electrocardiographic, angiographic, and biologic effects of pexelizumab. Unlike the case of some other trials that concluded earlier than expected, the sponsors agreed to ensure that these were to be appropriately funded as planned irrespective of the primary result.
It is unfortunate that the APEX-AMI trial was stopped early. We recognize that many factors influenced this decision and believe that the responsibilities to all stakeholders deserve discussion.
Given that the combined data to date have shown an important reduction in mortality with pexelizumab, why did the sponsors decide to stop the trial early? Whereas its sponsors' decision to discontinue the CONVINCE study for commercial reasons was unaccompanied by further explanation, that of the sponsors in the case of the APEX-AMI trial took a more measured approach and included a rationale.5, 9 This included the lower-than-expected blinded mortality rates that would necessitate a much larger sample size and anticipated investment to complete the trial as well as the disappointing and unexpected results of the PRIMO-CABG II trial. A positive result from the latter trial was assumed to lead to approval and marketing of pexelizumab (perhaps leading to market sales even before completion of the APEX-AMI trial) for broad use in conjunction with coronary bypass surgery. From the investigators' position, stopping such a trial early (for development and commercial reasons) potentially diminishes the value of their efforts and compromises their social contract with patients, which includes the knowledge that their participation will result in a properly powered study to reliably estimate the treatment effect and has the potential to guide and/or improve future care.4
On the other hand, the APEX-AMI trial as modified will provide important generalizable information, which when framed with prior pexelizumab trials will provide substantial information to support or deny the regulatory approvability and commercial viability of a potentially attractive, safe, and lifesaving therapy.8 Because the safety of the agent is not in question and because the APEX-AMI trial patients enrolled are receiving excellent evidence-based care based on our intratrial monitoring of quality, we are confident that their interests are protected. Although the opportunity cost associated with the potential for participating in other ST-elevation myocardial infarction trials is an issue, irrespective of the primary result in the APEX-AMI trial, the multiple substudies of the APEX-AMI trial promise to reveal important insights into improving the process of ST-elevation myocardial infarction care and into how the inflammatory milieu of ST-elevation myocardial infarction relates to outcomes.6 Although a futility analysis–guided approach may have allowed continuation of the trial, thereby providing some protection from the sponsors to stop the trial in the event of little likelihood of showing significant benefit, this was not deemed acceptable.
Early stopping of trials that are designed to develop promising treatments to improve clinical outcomes for commercial reasons has been roundly criticized and should be vigorously resisted as has been argued by others.3, 4, 5 Nevertheless, the costs of clinical development are enormous and a variety of both external and internal circumstances to the trial may ultimately make it impossible to complete a trial. Although the best approach in the APEX-AMI trial, from a public health perspective, would have been to continue the trial until approximately 11000 patients were enrolled, this would likely have required an additional 2 years and substantial additional cost that was deemed unavailable to the sponsors. Moreover, even the achievement of a statistically significant benefit with a new compound that fails to produce either an approximate 1% absolute or a 15% relative mortality reduction in acute myocardial infarction seems unlikely to be a viable proposition.10 Hence, early stopping of the APEX-AMI trial in an orderly way, at a time when approximately half of the anticipated events are achieved, is a compromise that aims to achieve balance between scientific, ethical, and commercial interests. It is our hope that by articulating the elements that comprised this decision in advance of knowing the trial result, we will enhance the transparency of why and how our original plan has been modified.
References
- Clinical trial registration—a statement from the International Committee of Medical Journal Editors. JAMA. 2004;292:1363–1364
- . Arthritis medicines and cardiovascular events—“House of Coxibs”. JAMA. 2005;293:366–368
- Premature discontinuation of clinical trial for reasons not related to efficacy, safety or feasibility. BMJ. 2001;322:603–605
- . Stopping trials early for commercial reasons: the risk-benefit relationship as a moral compass. J Med Ethics. 2005;31:410–414
- . Stopping medical research to save money—a broken pact with researchers and patients. JAMA. 2003;289:2128–2130
- Assessment of Pexelizumab in Acute Myocardial Infarction (APEX-AMI). A multicenter, randomized, double-blind parallel-group placebo-controlled study of pexelizumab in patients with acute myocardial infarction undergoing primary percutaneous coronary intervention. Am Heart J. 2005;149:402–407
- Press release—Alexion reports phase III results for PRIMO-CABG2 study. http://www.prnewswire.com/cgi-bin/stories.pl?ACCT=104andSTORY=/www/story/11-23-2005/0004222006&EDATE(accessed May 2006)
- Mahaffey K, Van de Werf F, Shernan S, et al. Effect of pexelizumab on mortality in patients with acute myocardial infarction or undergoing coronary artery bypass surgery: a systematic overview. AHJ (In press).
- Principal results of the Controlled Onset Verapamil Investigation of Cardiovascular End Points (CONVINCE) Trial. JAMA. 2003;289:2073–2082
- . An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. N Engl J Med. 1993;329:673–682
Conflict of Interest: Dr Paul Armstrong and Dr Christopher Granger are in receipt of research grants from Procter & Gamble Pharmaceuticals, Cincinnati, OH, and Alexion Pharmaceuticals, Cheshire, CT.
PII: S0002-8703(06)00554-0
doi:10.1016/j.ahj.2006.06.006
© 2006 Mosby, Inc. All rights reserved.
