Author + information
- Barry M. Massie, MD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Barry M. Massie, Cardiology Division (111C), VA Medical Center, 4150 Clement Street, San Francisco, California 94121
Over the last 3 decades, clinical trials have become increasingly globalized. This has especially been the case for cardiovascular outcomes trials. There are many reasons for this phenomenon, but foremost is the growing size and length of clinical heart failure trials. For example, the modern era of heart failure therapy with neurohormonal antagonists was ushered in by the Captopril Multicenter Trial published in 1983, in which 92 patients with refractory heart failure were randomized to captopril or placebo with a primary endpoint of improvement in treadmill exercise time after 3 months of treatment (1). Surprisingly, even with this small number of patients, a subsequent post-hoc analysis showed a significantly lower mortality rate in the captopril group compared with the placebo group (4% vs. 21%) (2). The captopril trial was followed by the 253-patient CONSENSUS I (Cooperative North Scandinavian Enalapril Survival Study) (3), which was stopped in 1986 by its safety monitoring committee after a mean follow-up period of 188 days, at which point, 66 patients in the placebo group and 44 patients in the enalapril group had died, representing a 40% reduction in mortality. Importantly, these trials, which were small and short in length, led to approval of the study drugs for the treatment of heart failure.
However, as time has passed, heart failure trials have become increasingly large, to the point that many are now deservedly termed “megatrials,” with most recent Phase 3 heart failure trials enrolling several thousand patients. The major reasons for this are listed in Table 1. Serial advances in therapy with drugs and, more recently, devices have improved the prognosis of heart failure patients, prevented clinical deterioration, and improved symptoms and quality of life. Demonstrating meaningful benefit with an additional agent on current guideline-recommended therapies has become increasingly difficult. Mortality reductions rarely approach the 20% to 30% range over a 1- to 3-year period, necessitating larger and longer trials that can detect much more modest, albeit clinically meaningful, improvements. Given advances in therapy, it has become increasing difficult to include placebo groups, so often an active comparator with known efficacy is utilized, making it more difficult to show an incremental benefit. Alternatively, a noninferiority design may be utilized to ensure that the new therapy is at least as good as another therapy with proven efficacy. This type of trial requires even larger numbers of patients depending on the noninferiority margin selected. With the current focus on drug safety, even larger trials may be required to exclude unexpected safety problems when a novel agent is studied.
All of these factors have led to ever-larger trials. At the same time, driven by the growing administrative and regulatory burdens, the costs of clinical trials, both at the site level and to the sponsors, have grown even faster than the size of the trials. This is particularly the case in the United States. As a result, major trials have become increasingly international, often employing large numbers of sites in multiple countries and continents. Many now involve not only North America and Western Europe, but also countries in Eastern Europe, Latin America, and both East and South Asia. Patients enrolled in the United States typically represent a minority of the study population. However, this globalization ensures a very heterogeneous study population, with potential major differences in genetics, environmental factors, diet, behaviors, comorbid conditions, concomitant medications, utilization of invasive procedures and devices, and other practice patterns (4–6). For instance, average lengths of stay for a patient with heart failure vary from as low as 4 to 5 days in the United States, to 2 to 3 weeks in Russia, resulting in a greater impact on readmission rates, which are frequently a component of the primary endpoint in heart failure trials.
It is in this context of globalization and practice variation that the paper by O'Connor et al. (7) in this issue of the Journal assumes considerable importance. Their report specifically examines many of the major beta-blocker heart failure trials. Two of these trials enrolled patients in both the United States and outside North America. MERIT-HF (Metoprolol Controlled-Release Randomized Intervention Trial in Heart Failure) enrolled 3,991 patients with New York Heart Association functional class II to IV heart failure and ejection fraction ≤0.40 (mean ejection fraction: 0.28; 1,071 in the United States) who were randomized to metoprolol succinate, a β1-selective agent, or placebo (8). The COPERNICUS (Carvedilol Prospective Randomized Cumulative Survival) trial studied 2,249 patients with severe systolic heart failure (mean ejection fraction: 0.20; 482 in the United States) who were randomized to carvedilol, a nonselective β1/β2-blocker with some α-blocking activity, or placebo (9). BEST (Beta-Blocker Evaluation of Survival Trial) was conducted entirely in North America. It enrolled 2,708 patients with New York Heart Association functional class III or IV heart failure and ejection fraction ≤0.35 (mean: 0.23; all 2,708 from the United States or Canada) who were randomized to bucindolol, a nonselective beta-blocker with possible intrinsic sympathomimetic activity (10,11). The results of the CIBIS II (Cardiac Insufficiency Bisoprolol Study II) trial (12), which studied bisoprolol, a β1-selective agent, were also included in this analysis, but since none of the patients were enrolled in North America, its results will not be discussed in this commentary.
The provocative major finding of the O'Connor et al. (7) analysis is that there appears to be a lesser survival benefit of each of the 3 beta-blockers that were studied in North American patients than in patients enrolled in the rest of the world. However, 81% (841 of 1,035) of the North American deaths occurred in the BEST, with only 194 occurring in MERIT and COPERNICUS combined. The deaths in the latter 2 trials are evenly split, with wide confidence intervals that do not exclude a meaningful mortality benefit of ≥20% in the U.S. patients.
The analyses in the O'Connor et al. (7) paper are provocative and possibly hypothesis generating, but they should not be interpreted as demonstrating a lesser beta-blocker benefit in North American patients, given the wide confidence intervals, particularly in COPERNICUS and MERIT-HF. Subsequent experience has solidified the powerful role that beta-blockers can play in preventing and even reversing myocardial systolic dysfunction in heart failure patients. Based on the totality of data with beta-blockers and their experience, few heart failure physicians would withhold carvedilol or metoprolol from their patients.
Nor should these findings be completely dismissed, but alternative explanations must be considered. First, as noted in the previous text, the finding of lesser benefit from beta-blockers in North America is largely driven from the BEST results and cannot be extrapolated with confidence to carvedilol or metoprolol succinate, the beta-blockers commonly used in the United States for heart failure patients, because of the small numbers of events and wide confidence around the effect estimates.
Lastly, there are important differences between BEST and the other large beta-blocker trials that were conducted in the United States. The BEST trial included a higher proportion (23%) of African Americans, who more frequently carry polymorphisms that are associated with a lesser benefit from bucindolol (13). Other researchers have reported that bucindolol has intrinsic sympathomimetic activity, which may abrogate the benefit of beta-blocker therapy (11). However, most important, one cannot exclude the play of chance in these findings.
Dr. Massie has reported that he has no relationships relevant to the contents of this paper to disclose.
↵⁎ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
- American College of Cardiology Foundation
- Captopril Multicenter Study Group
- O'Connor C.M.,
- Fiuzat M.,
- Swedberg K.,
- et al.
- Aiyar N.,
- Andreka P.,
- Olson L.C.,
- et al.