Education and debate

Randomised controlled trials in cardiovascular medicine: past achievements, future challenges

Salim Yusuf, professor of cardiology. 

Department of Medicine, McMaster University, Hamilton, Ontario, Canada

Correspondence to: S Yusuf, Hamilton General Hospital, Hamilton, Ontario L8L 2X2, Canada yusufs{at}fhs.mcmaster.ca

Clinical trials have played a crucial role in the development of treatment strategies for cardiovascular disease: the earliest trials were conducted in the 1950s, but it was not until the 1970-80s that the results of clinical trials had a major impact on the choice of treatments. Relieving symptoms and improving quality of life have always been treatment goals. Over the centuries, the choice of treatment has evolved through several periods (box 1). The approaches are not necessarily distinct from each other, and current treatment strategies for cardiovascular disease are based on a mixture of goals aimed at the improvement of symptoms (for example, breathlessness), correcting markers of disease (improved ventricular function, etc), and improving clinical outcomes (fewer admissions to hospital, prolonged survival, etc). What characterises the current era is the expectation that theory, observations in animals, and human physiological studies alone are not enough to determine the value of a treatment. Rather, these observations should be verified by providing unequivocal evidence of net clinical benefit on the basis of reliable studies using the methods of randomised controlled trials.

	The increasing impact of trials

Over the past 40 years, the results of randomised controlled trials have had an increasing impact on treatment choices (box 2). Firstly, the design and conduct of trials have improved so that results have become more reliable.^1-3 Secondly, the increasing acceptance of meta-analysis as a valid and useful methodology has meant that data from all trials, even trials too small to be reliable on their own, could contribute usefully towards the overall evidence.⁴ Thirdly, better designed trials and well conducted meta-analyses have established that many existing simple and inexpensive treatments are effective in reducing mortality and morbidity. Equally important, these trials and meta-analyses showed that many commonly used treatments were either useless or harmful, despite promising data from experimental studies, epidemiological observations, or small trials indicating a favourable impact on surrogate outcomes. Fourthly, over the past 40-50 years a vigorous effort by the pharmaceutical industry has led to the development of numerous compounds, which have been subject to rigorous randomised controlled trialspartly to satisfy the requirements for regulatory approval and partly because the medical community has demanded reliable evidence that only such trials could provide. Consequently in many countries the payers (drug benefit plans funded by private insurance or governments) have demanded not only evidence of benefit with an acceptable safety profile but also that costs matched effectiveness. The increasing reliance on randomised controlled trials for evaluating treatments was due to several factors: sound methodological principles, the need to translate discoveries in basic science reliably and rapidly to improve clinical outcomes, and social forces such as regulatory and economic factors.

	Making trials more reliable

The most fundamental advance that has made trials more reliable is randomisation. This has allowed similarities in measured, unmeasured, and unknown risk factors to be identified between active and comparator groups. Any difference in outcomes (provided it was measured in an unbiased way) was then due to differences in the treatments compared. In the 1970-80s, rediscovery that these principles were the dominant aspects of the validity of controlled trials suggested that substantial simplicity was possible in a trial's design (minimal data collection, and little or no standardisation as variations in other factors balanced themselves out between the groups) without compromising the main goals of the study. Also, acceptance that moderately sized differences in treatment (for example, a reduction in risk of one fifth or one sixth), major morbidity (for example, myocardial infarction or strokes), or mortality were likely and worthwhile made the detection of such differences important. Detecting such differences required larger studies in which several hundred or even a thousand events were observed. If the trial design was simple, large trials could be conducted at an affordable cost thereby making the evaluation of the treatments more efficient and reliable.⁵ One of the first large, simple randomised controlled trials was the US polio vaccine trial in the 1950s, which included 400 000 children, and which showed in one season the efficacy of the vaccine.⁶ Key aspects of this trial were the lack of detailed data collection on each subject, and only passive ascertainment of cases of paralytic polio when those children who had been randomised were admitted to hospital. Simplification made this large study practical at a comparatively modest cost. The value of the polio vaccine could only be established reliably by such a large trial, and the global impact of the vaccine in preventing morbidity and mortality from polio has been profound.

In cardiology, the principles of large trial sizes and simplicity were first applied to the ISIS (international study of infarct survival) and GISSI (Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico) series of trials.^7-9 These trials, which showed the value of  blocking drugs, aspirin, thrombolytic therapy, and angiotensin converting enzyme inhibitors in acute myocardial infarction, have altered the management of patients with this condition. Large trial sizes and simplicity have been adapted and applied to long term trials in heart failure (for example, the studies of left ventricular dysfunction (SOLVD)¹⁰ and Digitalis Investigation Group (DIG) trials¹¹), in secondary prevention (for example, the long term intervention with pravastatin in ischaemic disease (LIPID)¹² and the heart outcomes prevention evaluation (HOPE) trials¹³), and in primary prevention.^14-16

Another major advance is the use of a factorial design whereby more than one intervention is evaluated within a trial. This strategy has been successfully used several times in cardiology, and it allows the simultaneous evaluation of generic, inexpensive treatments (for example, aspirin¹⁷ or a vitamin¹³ or generic intervention¹⁸ for which there may not be significant funding), and newer compounds (for which there may be funding from the pharmaceutical industry). Despite concerns about "interactions" when more than one treatment is simultaneously evaluated, these have not been commonly observedpartly because drugs that may interact are not tested in the same trial and most patients are already on multiple drugs, and partly because interactions, even if they exist, may be only moderate in size.

(Credit: MARK HUDSON)

The development of principles and methods of meta-analysis have also had a major impact.¹⁹ Meta-analysis shares two key principles with large simple trials: large numbers of events are needed to reliably detect moderately sized differences, and assessment of only key data, random allocation, and unbiased outcomes are essential. Meta-analysis also emphasises the importance of making decisions on the data or outcomes from randomised controlled trials. Both large trials and meta-analyses have had a major impact on cardiovascular disease because they have provided persuasive answers that were not available by other means. Meta-analyses of existing trials showed that aspirin was effective in preventing vascular deaths, myocardial infarctions, and strokes in high risk patients.^20-22 In other situations meta-analysis emphasised what was apparent in some, but not all, trials viewed in isolation (for example,  blockers after myocardial infarction)¹⁹ or led to renewed interest in old treatments (thrombolytic therapy, for example),²³ which were confirmed by further well designed randomised trials.²⁴ In other cases the results of meta-analysis led to large trials that disproved hypotheses, such as the value of magnesium in myocardial infarction. ^{25 26}

Although randomised controlled trials have contributed substantially to improving the management of patients, large simple trials have some limitations. By themselves they cannot shed light on the mechanisms by which an intervention works. Therefore large trials should be complemented by other forms of mechanistic research and small physiological studies. Also, many trials include only a small proportion of patients with the disease of interest, so that the applicability of results to a broad group of patients is sometimes uncertain. Trials should be designed to be more relevant by including a broad range of patients with the disease of interest and excluding only those with clear indications or contraindications for the treatment being evaluated.²⁷ Indeed the relevance of randomised controlled trials to clinical practice may be enhanced by minimising details of inclusion and exclusion criteria, and leaving substantial judgment to participating physicians. As long as a reasonable description of the characteristics of those in the trial is available, such a simple and flexible approach to patient entry will enhance the value of the trial by mimicking the "real world" and including a broader range of patients. Individual trials have generally had only a modest impact on clinical practice, and it usually takes several studies with the same result to convince practitioners to adopt a new treatment or abandon a commonly used treatment. Even after several trials with favourable results, the adoption into clinical practice is often slow.²⁸ Progress in reducing the care gapthat is, the gap between those eligible for a treatment and those actually receiving itis critical to ensuring that patients benefit fully.

	Future challenges

Randomised controlled trials in cardiovascular disease face many new challenges. Currently, 60% of cardiovascular disease occurs in developing countries, and by the year 2020 this is expected to increase to 80%.²⁹ There is a major need to conduct trials of simple widely applicable treatments in these populations.

Many current trials have tended to underrepresent or even exclude elderly people (the mean age of patients with heart failure in the community is 70 to 75 years, yet in most trials the mean age of participants has been 60 to 65 years). Elderly people usually have the worst prognosis and are also likely to be at the greatest risk of adverse events. Given that elderly people are the fastest growing segment of the population and that they have high rates of cardiovascular disease and related drug use, future trials should facilitate enrolment of large numbers of such patients.

As advances in treatment continue, the potential for incremental benefit from new treatments tends to decline, the potential for side effects tends to increase, and there is a greater likelihood of drug interactions because patients are receiving several drugs. These factors may stimulate more emphasis on the concept of "primordial prevention" whereby societal and lifestyle factors could be modified to prevent the development of risk factors for common disease. Hypotheses regarding approaches to primordial prevention strategies at the community level will need evaluation in randomised trials conducted over decades and therefore require innovative study designs such as cluster randomisation, low intensity interventions, and passive ascertainment of outcomes.

A major impetus for randomised controlled trials in evaluating cardiovascular treatments has been the roles of regulatory bodies (by insisting on data from well designed trials for regulatory approval) and pharmaceutical companies (by developing new compounds and funding trials). These same influences, however, now also pose the greatest dangers to progress in cardiovascular treatments. The bureaucratisation of the conduct of clinical trials has made trials so expensive that the really large trials that are often required may never be done (box 3). These regulations have spawned huge bureaucracies within regulatory bodies, companies, and organisations of clinical trials, with very little scientific or medical value and little improvement in the reliability and validity of controlled trials. Multiple checks were designed in response to the rare instance of fraud, but have never been shown to reduce fraud or improve the reliability of trials. This has led to a culture of blind "rule followers," rather than a community that understands the critical principles of good trial design.

A second major unfavourable impact has been the overcommercialisation of trials. Although the reasonable costs of conducting any research should be met, the increasingly large per patient reimbursements that pharmaceutical companies provide and some investigators demand have made participation by some investigators more of a business than a scientific or medical endeavour. Conduct of less expensive trials of cheap generic treatments, such as a vitamin or a drug at the end of its patent life, are in jeopardy because the amount of compensation that investigators expect corresponds to commercial rates, which no peer review bodyfor example, the National Institutes of Health or the Medical Research Councilcan justify. These two influences could potentially lead to the decline of randomised trials as they become unrealistically expensive. These problems can, however, be avoided by simplification of government regulations and by their implementation by pharmaceutical companies. With this reduction in complexity, investigators should be prepared to participate both in trials that are commercially drivenwhere they are well rewardedand in trials of important scientific questions funded at more modest levels by peer review bodies. Forging partnerships between government, academia, and industry to facilitate the conduct of more trials with factorial designs will allow efficient and simultaneous evaluation of generic questions that are of societal importance.

	Conclusion

As the next millennium approaches and we are poised to make substantial further advances in treating and preventing diseases using the gains from emerging technologies and molecular biology, we will need to rely increasingly on well designed and efficient randomised trials to distinguish between worthwhile, useless, and harmful interventions. We must support the conduct of important trials of good questions of relevance to public health by ensuring adequate support from both government and industry; by making randomised controlled trials more efficient by reducing waste, unnecessary bureaucracy, and regulatory demands; and, as investigators, by being willing to participate for fair or little financial compensation. Such an approach will ensure continuing progress in the battle against cardiovascular disease and other diseases, and will ensure that the fruits of basic science can be rapidly applied to human populations.

	Acknowledgments

I thank M Gent, J Pogue, L Friedman, and J Wittes for their comments, and Karin Dearness, who typed the various drafts of the manuscript. This article is based on a presentation delivered at the British Medical Journal and Medical Research Council's meeting on the 50th anniversary of the first randomised clinical trial, 30 October 1998. SY is the holder of a research chair funded by the Heart and Stroke Foundation of Ontario and recipient of a senior scientist award from the Medical Research Council of Canada.

	Footnotes

Competing interests: None declared.

	References

1.	Friedman LM, Furberg C, DeMets DL. Fundamentals of clinical trials 3rd ed. St Louis, MS: Mosby , 1996.
2.	Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 1977; 35: 1-39[Medline].
3.	Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. I. Introduction and design. Br J Cancer 1976; 34: 585-612[Medline].
4.	Pogue J, Yusuf S. Overcoming the limitations of current meta-analysis of randomised controlled trials. Lancet 1998; 351: 47-52[Medline].
5.	Yusuf S, Collins R, Peto R. Why do we need some large, simple randomized trials? Stats Med 1984; 3: 409-420[Medline].
6.	Meldrum M. "A calculated risk:" the Salk polio vaccine field trials of 1954. BMJ 1998; 317: 1233-1236[Free Full Text].
7.	ISIS-1 (First International Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous atenolol among 16 027 cases of suspected acute myocardial infarction: ISIS-1. Lancet 1986; i: 57-66.
8.	GISSI (Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico). Long-term effects of intravenous thrombolysis in acute myocardial infarction: final report of the GISSI study. Lancet 1987; ii: 871-874.
9.	Flather MD, Farkouh ME, Yusuf S. Large simple trials in cardiovascular disease: their impact on medical practice. In: Califf RM, Mark B, Wagner GS, eds. Acute coronary care. St Louis, MS: Mosby, 1995:131-144.
10.	The SOLVD (studies of left ventricular dysfunction) Investigators. Effect of enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure. N Engl J Med 1991; 325: 293-302[Abstract].
11.	The Digitalis Investigation Group. The effect of digoxin on mortality and morbidity in patients with heart failure. N Engl J Med 1997; 336: 525-533[Abstract/Free Full Text].
12.	The Long-term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. N Engl J Med 1998; 339: 1349-1357[Abstract/Free Full Text].
13.	The Heart Outcomes Prevention Evaluation Study Investigators. The HOPE (heart outcomes prevention evaluation) study: the design of a large, simple randomized trial of an angiotensin converting enzyme inhibitor (ramipril) and vitamin E in patients at high risk of cardiovascular events. Can J Cardiol 1996; 12: 127-137[Medline].
14.	Steering Committee of the Physicians' Health Study Research Group. Final report on the aspirin component of the ongoing physicians' health study. N Engl J Med 1989; 321: 129-135[Abstract].
15.	Rodgers A, MacMahon S, Gamble G, Slattery J, Sandercock P, Warlow C, for the United Kingdom Transient Ischaemic Attack Collaborative Group. Blood pressure and risk of stroke in patients with cerebrovascular disease. BMJ 1996; 313: 147[Free Full Text].
16.	Tardif JC, Cote G, Lesperance J, Bourassa M, Lambert J, Doucet S, et al, on behalf of the Multivitamins and Probucol Study Group. Probucol and multivitamins in the prevention of restenosis after coronary angioplasty. N Engl J Med 1997; 337: 365-372[Abstract/Free Full Text].
17.	ISIS-2 (Second International Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17 187 cases of suspected acute myocardial infarction: ISIS 2. Lancet 1988; i: 349-360.
18.	A clinical trial comparing primary coronary angioplasty with tissue plasminogen activator for acute myocardial infarction. The Global Use of Strategies to Open Occluded Coronary Arteries in Acute Coronary Syndromes (GUSTO IIb) Angioplasty Substudy Investigators. N Eng J Med 1997; 336: 1621-1628[Abstract/Free Full Text].
19.	Yusuf S, Peto R, Lewis J, Collins R, Sleight P. Beta blockade during and after myocardial infarction: an overview of the randomized trials. Prog Cardiovasc Dis 1985; 27: 335-371[Medline].
20.	Antiplatelet Trialists' Collaboration. Collaborative overview of randomised trials of antiplatelet therapy. I: prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ 1994; 308: 81-106[Abstract/Free Full Text].
21.	Antiplatelet Trialists' Collaboration. Collaborative overview of randomised trials of antiplatelet therapy. II: maintenance of vascular graft of arterial patency by antiplatelet therapy. BMJ 1994; 308: 159-168[Abstract/Free Full Text].
22.	Antiplatelet Trialists' Collaboration. Collaborative overview of randomised trials of antiplatelet therapy. III: reduction in venous thrombosis and pulmonary embolism by antiplatelet prophylaxis among surgical and medical patients. BMJ 1994; 308: 235-246[Abstract/Free Full Text].
23.	Yusuf S, Collins R, Peto R, Furberg C, Stampfer MJ, Goldhaber SZ, et al. Intravenous and intracoronary fibrinolytic therapy in acute myocardial infarction: overview of results on mortality, reinfarction and side-effects from 33 randomized controlled trials. Eur Heart J 1985; 6: 556-585[Abstract/Free Full Text].
24.	Fibrinolytic Therapy Trialists' (FTT) Collaborative Group. Indications for fibrinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity results from all randomised trials of more than 1000 patients. Lancet 1994; 343: 311-322[Medline].
25.	Teo KK, Yusuf S, Collins R, Held PH, Peto R. Effects of intravenous magnesium in suspected acute myocardial infarction: overview of randomised trials. BMJ 1991; 303: 1499-1503.
26.	ISIS-4 (Fourth International Study of Infarct Survival) Collaborative Group. ISIS-4: a randomised factorial trial assessing early oral captopril, oral mononitrate, and intravenous magnesium sulphate in 58,050 patients with suspected acute myocardial infarction. Lancet 1995; 345: 669-685[Medline].
27.	Yusuf S, Held P, Teo KK, Toretsky ER. Selection of patients for randomized controlled trials: implications of wide or narrow eligibility criteria. Stats Med 1990; 9: 73-86[Medline].
28.	Montague T, Montague P, Dzavik V, Teo K, for the Clinical Quality Improvement Network (CQIN) Investigators. The promise and practice of cardiovascular risk reduction: a disease management perspective. Can J Cardiol 1996; 12: 995-999[Medline].
29.	Reddy KS, Yusuf S. Emerging epidemic of cardiovascular disease in developing countries. Circulation 1998; 97: 596-601[Free Full Text].

(Accepted 30 April 1999)