Author + information
- Received November 22, 2010
- Revision received January 20, 2011
- Accepted January 26, 2011
- Published online August 9, 2011.
- ↵⁎Reprint requests and correspondence:
Dr. Venu Menon, Cleveland Clinic, Department of Cardiovascular Medicine, Desk F15, 9500 Euclid Avenue, Cleveland, Ohio 44195
Objectives The aim of this study was to describe international enrollment and participation in National Institutes of Health (NIH)–sponsored cardiovascular randomized controlled trials (RCTs).
Background RCTs provide the evidence base for major societal guidelines and profoundly influence patient care in the United States. Increased international involvement in clinical trials has been observed, but the rate of international enrollment in NIH-sponsored cardiovascular RCTs has not been described.
Methods The NIH registry of clinical trials was searched for phase III or IV cardiovascular RCTs funded by the National Heart, Lung, and Blood Institute. Studies with outcomes of myocardial infarction, stroke, or death published between 1997 and 2009 were included. Rates of international enrollment were obtained from published data or personal communication with corresponding authors.
Results Twenty-four studies met all inclusion criteria. Nineteen trials including 151,682 patients had international participation (IP), with median IP of 9.5% (range 0% to 100%). Coronary artery disease trials (11 studies) had nearly 50% international enrollment. High-risk trials and trials testing acute interventions tended to have higher rates of IP.
Conclusions Cardiovascular RCTs sponsored by the National Heart, Lung, and Blood Institute have substantial rates of international enrollment, particularly coronary artery disease trials. Given questions of applicability and ethical and financial considerations, IP in U.S. clinical trials deserves further scrutiny.
As part of its mission as the “steward of medical and behavior research for the Nation,” the National Institutes of Health (NIH) strive to provide “leadership and direction to programs designed to improve the health of the nation by conducting and supporting research in the causes, diagnosis, prevention, and cure of human diseases” (1). Subject to intense competition and peer review, NIH-supported cardiovascular (CV) randomized controlled trials (RCTs) represent the pinnacle of clinical research and seek to answer critical mechanistic and strategic questions pertinent to the practice of clinical cardiology in the United States. The results of these studies profoundly influence patient care both in the United States and worldwide and universally result in Class 1 recommendations in the American College of Cardiology and American Heart Association treatment guidelines. These trials need to be of sufficient sample size to adequately test new developments in the treatment of CV disease, and often, insufficient enrollment of U.S. patients leads to the enrollment of international patients. Our observation of high rates of international enrollment in recent landmark, NIH-sponsored CV trials led us to explore the overall rates of international enrollment in these studies.
This study is an analogue analysis of our prior research, which investigated the participation of women in NIH-sponsored trials (2). For that study, we searched the NIH registry of clinical trials for phase III or IV RCTs under the category of CV diseases funded by the National Heart, Lung, and Blood Institute (NHLBI). Studies with primary outcomes of myocardial infarction, stroke, or death published between 1997 and 2006 were included. We updated this search for the present analysis and included all eligible RCTs that published data by December 1, 2009. From the primary reports, supplementary methods reports, and personal communications with corresponding authors, we obtained the total number of patients randomized, the total number of patients recruited from outside of the United States, and the sites of international recruitment.
The median proportion of international participation (IP) per category and the total number of international patients enrolled per category divided by the total number of randomized patients in that category (Σ n international)/(Σ n randomized) are reported. Trials were then categorized according to the CV disease processes studied: coronary artery disease (CAD), electrophysiologic disease, congestive heart failure, hypertension, or primary prevention of CV disease. Trials were further categorized by level of invasiveness of the intervention being tested (invasive or noninvasive) and the level of acuity of the patients when recruitment for the trial occurred. Trials were considered acute if the intervention occurred during the acute phase of illness, such as myocardial infarction, decompensated heart failure, acute arrhythmia, or surgery.
In our original search strategy, there were 11,918 clinical trials sponsored by the NIH, 1,488 of which were funded by the NHLBI. Of the 982 studied CV disease, 141 were completed phase III or IV RCTs in adults, of which 53 had primary outcomes of myocardial infarction, stroke, or death. Of the 53, 26 were not published within the pre-specified time frame, and 5 had no published data. Since our original search, completed on November 16, 2007, 3 more trials were published meeting all inclusion criteria, resulting in 25 trials overall. Complete information was available for 24 of them (Table 1).
There were 11 CAD trials, 3 congestive heart failure trials, 6 electrophysiologic disease trials, 1 hypertension trial, and 3 primary prevention trials. Nineteen trials had IP (Table 2), and the overall median IP was 9.5% (range 0% to 100%) (Table 3). Fifteen percent of all enrollees in NHLBI-sponsored CV RCTs are from outside the U.S. CAD trials had the highest proportion of IP, and individual trials are illustrated in Figure 1. IP in trials of other categories of CV disease was much smaller (0.3% to 15.8%), but there was a trend toward increased IP in higher risk trials (physically invasive interventions) and trials performed during the acute phase of illness, although the statistical significance of these findings were not explored. Canada was the most commonly listed international site and enrolled patients for 16 of 24 trials (Table 2).
Although there is heterogeneity in international enrollment in NHLBI-sponsored CV RCTs, enrollment outside the United States in these federally funded trials appears substantial, especially for those trials studying atherosclerotic CAD. Nine of the 11 CAD trials had IP, and 50% of all patients enrolled in NIH-sponsored CAD trials are from outside the United States. In the most recent STICH (Surgical Treatment for Ischemic Heart Failure) trial, 80% of subjects were recruited internationally. In light of these findings, the reasons and consequences of increasing IP in CV trials deserve exploration.
IP in federally funded clinical trials has a number of potential benefits. Despite a decline in age-adjusted CV morbidity in developed nations, the overall global burden of CV disease continues to rise exponentially, particularly in developing countries (3,4). By participating in NIH trials, nations that do not have the robust research infrastructure of the United States have an opportunity to participate in rigorously conducted research protocols. Ideally, clinical trialists in these nations would be exposed to the scrutiny, operational rigor, and ethical considerations that govern these well-coordinated studies, creating an international pool of trialists well versed in the ethical conduct of research. Trial participation in these environments also enables research subjects to obtain access to cutting-edge treatment and interventions in areas of clinical equipoise. Smaller nations that do not have the overall population to perform well-powered clinical trials are able to contribute to the evidence gathering process in these studies, giving them a gauge of the effectiveness of the tested interventions in their individual health care environments. Increasing international enrollment may also facilitate the timely completion of sample size goals for large studies, thus leading to statistically definitive trial results.
Increasing IP in clinical trials may also be perceived negatively. The ethical and transparency concerns regarding the conduct of clinical trials in developing countries have been recently highlighted (5,6). Increased IP in NIH clinical trials may also challenge the generalizability of the overall results and cloud the anticipated treatment effect and physician acceptance when applied in the U.S. population. The treatment effect of an intervention in clinical trials may be strongly influenced by baseline patient characteristics, background medical therapy, ascertainment of clinical events, and the completeness and rigor of clinical follow-up, all of which show geographic variability. Multinational large CAD trials have shown significant differences in baseline characteristics as well as outcomes between U.S. and non-U.S. populations (7). In the HERO-2 (Hirulog and Early Reperfusion or Occlusion 2) trial, a significant variation in 30-day mortality was observed that ranged from 6.7% in the heparin arm of Western countries to 10.3% in Latin America and 13.7% in Russia and Asia, differences that persisted despite adjustment for baseline characteristics, treatment, national health, and economic statistics. In contrast, rates of reinfarction on heparin were highest in the Western nations at 5.1%, compared with 2.8% in Eastern Europe, likely reflecting inadequate enzymatic and electrocardiographic surveillance (8). The observed treatment effect in the PURSUIT (Platelet IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy) trial was greatest in North America, with no observed benefit for integrilin in Latin America and Eastern Europe on univariate analysis. The geographical differences in outcome in this study were largely explained by differences in patient characteristics but were also significantly influenced by the definition of nonfatal myocardial infarction (9). In contrast to the overall study results, an analysis of PLATO (Study of Platelet Inhibition and Patient Outcomes) showed numerically higher rates of clinical outcomes for all efficacy measures with ticagrelor over clopidogrel, including myocardial infarction, stroke, all-cause mortality, and CV death for U.S. participants in the trial (10).
We acknowledge that Canada was the most common non-U.S. site listed among the trials included in our study. Some would argue that Canada is a developed country with similar demographic characteristics to the United States, but previous investigations into the differences in patterns of care for patients in these 2 countries have shown otherwise. Canada has a universal health care system and spends less money per capita for health care expenditures (U.S. $5,293  vs. U.S. $6,700 ). In a review of 3 large trials of acute coronary syndromes that included comparisons between Canada and the United States (13), Canadian patients were more likely to be lighter, less likely to have hypertension or hypercholesterolemia, and less likely to have undergone coronary revascularization procedures. Furthermore, more Canadian patients had anterior myocardial infarctions than U.S. patients, and Canadians were more likely to present with higher Killip class scores. Time from symptom onset to thrombolytic or percutaneous intervention was also longer in Canadian patients.
Besides study interpretation and acceptance, IP in NIH trials does raise other concerns. There is a federal mandate for the equal inclusion of women and minorities in all clinical trials conducted by the NIH (14). We wonder whether the high rates of IP in CAD trials are contributing to lower rates of enrollment of women and minorities in these trials (2,15,16). The wide heterogeneity in the enrollment of international patients in large CV RCTs does not appear to be a fleeting trend. The FREEDOM (Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease) trial recently completed enrollment, and its results are eagerly awaited by the medical community. U.S. enrollment in this potentially landmark study is 360 of 1,901 (19%) (Michael Farkouh, MD, personal communication, November 2010).
The decreased U.S. participation in NHLBI CV trials is cause for concern. In the United States, we are fortunate to have outstanding medical research institutions, universities with medical schools, and tertiary hospitals across the length and breadth of our nation. Access to clinical trial participation should not be an issue. The lack of U.S. participation is certainly not due to clinical relevance of the trial being tested. These trials in CAD supported by the NIH are peer reviewed and represent vital issues in CV health care delivery that warrant study and randomization. Why, then, is an important trial such as STICH able to enroll only 200 patients in a country of 308 million, when Poland, with a population of 38 million, enrolls 288 subjects? Similarly, overall U.S. enrollment in OAT (Occluded Artery Trial) is 488 at 83 sites, compared with 450 at 27 sites in Canada. What is the cause of these low rates of United States enrollment in these pivotal clinical trials? Is it hesitancy on the part of U.S. patients to participate in a research question considered a top priority by experts in the field? Is it a lack of desire on the part of U.S. physicians to randomize their patients to answer a clinical question for which a suitable answer is yet unavailable? Is it failure of medical institutions to foster an environment that bolsters research activity? Are these the consequences of current regulatory and compliance issues that have gone overboard to make research cumbersome?
We acknowledge that patient enrollment in clinical trials in the U.S. can be a difficult task and that many of the trials in this study opened recruitment outside the U.S. because of inadequate enrollment at U.S. sites. We believe that research into the reasons for the underenrollment of U.S. patients in these critical clinical trials is urgently warranted. We also suggest that academia and tertiary medical institutions should be more supportive, with every effort made to recruit U.S. patients into these NIH trials.
Dr. Kim is a consultant for Philips Ultrasound and a recipient of an American College of Cardiology Foundation/GE Career Development Award recipient (GE provides contrast media for granted study). All other authors have reported that they have no relationships to disclose.
- Abbreviations and Acronyms
- coronary artery disease
- international participation
- National Heart, Lung, and Blood Institute
- National Institutes of Health
- Occluded Artery Trial
- randomized controlled trial
- Received November 22, 2010.
- Revision received January 20, 2011.
- Accepted January 26, 2011.
- American College of Cardiology Foundation
- National Institutes of Health
- Kim E.S.,
- Carrigan T.P.,
- Menon V.
- Yusuf S.,
- Reddy S.,
- Ounpuu S.,
- Anand S.
- Yusuf S.,
- Reddy S.,
- Ounpuu S.,
- Anand S.
- DeMaria A.N.
- Akkerhuis K.M.,
- Deckers J.W.,
- Boersma E.,
- et al.
- Marciniak T.A.
- ↵NIH Organization Functional Statements. Chapter 1125 http://oma.od.nih.gov/ms/organization/function/index.html. Accessed July 25, 2008.
- Centers for Disease Control, National Center for Health Statistics
- ↵National Institutes of Health Revitalization Act of 1993. Pub. L. No. 103-43 (June 10, 1993).
- Leaf A.,
- Albert C.M.,
- Josephson M.,
- et al.
- Schulman S.P.,
- Thiemann D.R.,
- Ouyang P.,
- et al.
- O'Hara G.E.,
- Charbonneau L.,
- Chandler M.,
- et al.
- Hamilton Health Sciences/McMaster University Population Health Research Institute