Simple Principles of Clinical Trials Remain Powerful

In this issue of JAMA, 2 articles from the Clinical Trial of Reviparin and Metabolic Modulation in Acute Myocardial Infarction Treatment Evaluation (CREATE) greatly enhance the knowledge base for the treatment of acute ST-segment elevation myocardial infarction (STEMI).^{1 - 2} Yet, as with all good clinical trials, this study raises new questions and perspectives that deserve further consideration.

The factorial design of the CREATE trial proves 2 points: (1) low-molecular-weight heparin reduces mortality and ischemic events in STEMI and (2) the age-old argument about glucose, insulin, and potassium (GIK) therapy can be put to rest—this treatment does not work in STEMI. The CREATE trial is powerful and definitive on both counts. However, the conduct and results of the trial stimulate thoughts about several questions.

Do the Principles of Pragmatic Clinical Trials^{3 - 4} Still Hold in This Era of “Personalized Medicine”? Despite ongoing efforts to develop more targeted approaches to therapeutics using biomarkers and genomics, the majority of therapies for complex diseases act through mechanisms that are not specifically targeted. Accordingly, a trial of adequate size to show modest but important effects on true clinical outcomes must remain the standard for the evidence on which therapies are based. In fact, it is easy to postulate many mechanistic reasons why either of the therapies tested in CREATE would or would not “work,” but the definitive proof requires an accounting of the risks and benefits using the “big 4” measures: length of life, quality of life, discrete negative events, and costs.

The CREATE investigators used a technique that is all too often avoided: a factorial design.⁵ In essence, this approach allows 2 questions to be simultaneously answered for little more than the effort of 1. The most common reason for eschewing factorial designs is the fear that in a regulatory setting the adverse effects of one drug (or device) will be attributed to the other being tested. The timidity of sponsors and investigators to embrace factorial designs must be overcome if researchers are to answer the multitude of critical questions about old and new therapies, especially the interactions among them. Given the increasing evidence that the drug and device research and development system is not producing the evidence that is needed to guide practice,⁶ regulators would be well served to encourage factorial-design studies.

Even more important, the CREATE trial eliminated useless, redundant approaches to clinical trial management, permitting a clear answer to 2 important questions for the global community of patients with myocardial infarction. Data forms were brief, redundant monitoring visits were not done, and adverse event reporting was limited to critical issues. By making the trial less burdensome on the investigators, the CREATE investigators encouraged and accomplished broad participation without need for external funding. Today’s overly complex “regulatory trials” tend to lead to data-intensive studies with inadequate numbers of patients and end points that do not measure length or quality of life. More trials like CREATE are needed, with fewer of these “regulatory” trials.

What Does This Trial Show About GIK Infusion? This “generic” therapy has survived decades of clinical practice without a definitive test of its safety and effectiveness. The conceptual basis for its use is sound, and its rationale has been delineated in a variety of publications cited by the CREATE investigators. Unfortunately, regardless of its scientific rationale and the positive results of small studies, this definitive trial, combined with a previous overview⁷ that showed only a modest potential benefit, answers the question beyond reasonable doubt: there is no benefit of GIK therapy. Societal resources would be better spent on evaluating other approaches in clinical trials and using other therapies in practice.

What Does This Trial Show About Reviparin? Antithrombotic therapy has become a mainstay of treatment of acute coronary syndromes (ACS), but the accrual of definitive evidence has been hampered by the fact that unfractionated heparin (UFH) became standard before the era of definitive clinical outcome trials. Thus, the CREATE investigators are technically correct in their attack on UFH in the discussion section of their article. However, directly comparative trials of either low-molecular-weight heparin or direct thrombin inhibitors have shown only modest improvements over UFH in a variety of settings, and overviews of small trials in STEMI show benefit; therefore, the inference that UFH is beneficial is strong.^{8 - 10} CREATE advances the field by demonstrating that the recommendation for antithrombotic therapy can now be made with confidence that the evidence is not built like a house of cards on a series of neutral “equivalence trials” or small outcome trials. Given the modest advantage of low-molecular-weight heparin in general in the broader range of ACS, including non–ST-elevation ACS, CREATE enhances the case for its preference over UFH. A nuance that deserves a follow-up article involves the more than 1000 patients in CREATE who were treated with placebo antithrombotic therapy and direct percutaneous coronary intervention; clear understanding is needed regarding how these patients were treated and how their outcomes affected the overall trial results.

How Do Such Trials Fit With Evolving Concepts of Medical Quality? The United States appears to be in the early phases of what will be a dramatic shift in payment for health care services.¹¹ This shift is toward “pay for performance,”¹² in which reimbursement will be based on demonstrated reliable delivery of health care that is of high quality and devoid of frequent medical errors. Quality has been defined by the Institute of Medicine as the degree to which the delivery of health care is consistent with practices that have been shown to improve clinical outcomes.^{13 - 14} A medical error is defined as either having the wrong plan or failing to properly execute the correct plan. Combining these 2 constructs leads to the conclusion that use of an antithrombotic drug is a standard of care in STEMI and that failure to use a proven drug at a dose that has been shown in a rigorous clinical trial to improve clinical outcomes is a medical error (except in the context of a clinical trial evaluating an alternative). Clinicians and hospitals who reliably administer an antithrombotic agent at the right dose should and will be paid more than those who do not. Conversely, administering GIK will be viewed as a waste of resources and energy that diverts from the focus on effective treatments.

Can Results of Trials Conducted in Very Different Cultures Answer Pragmatic Questions About Medical Care in the United States? There has been a dramatic shift in clinical research into the burgeoning populations of eastern Europe, South America, India, and China. To the extent that these trials answer questions of primary interest to the populations of these regions and foster positive relationships among clinician investigators across the world, this change is desirable. The CREATE trial clearly meets these criteria. Although not the motivation for the CREATE investigators, in many other cases this shift in research represents an avoidance of the increasingly irrational regulations in the United States and western Europe that are driving up the cost and reducing the efficiency of clinical research with little to no documented improvement in quality or protection of human research participants.

Even more worrisome is the potential for exploiting human research participants in economically disadvantaged countries by having them serve as the basis for inexpensive studies of expensive treatments that will not be available to most of the populations of these countries or in long-term care of the trial participants.¹⁵ There is little doubt that this is the motivation for too much of the shift in the location of both drug and device development research. The CREATE investigators deserve enormous credit for “doing it right” in contrast with other investigators.

Given the differences in standards of care across regions and countries, questions arise about the generalizability of trial results across national boundaries. When background therapy in a trial is vastly different from common therapy in the United States, it is difficult to know how to apply the results to country-specific standards of quality. This issue has become especially important with the advent of the era of “pay for performance”; the measures used to judge quality will carry increasing importance, and errors in devising these measures could have major consequences.¹⁶

Could the Principles of the Pragmatic Trials Approach Be Used in More Long-term Medical Care Situations to Extricate Clinicians and Patients From the Drug Safety Crisis? The massive uncertainty about the safety of drugs in Western society^{17 - 18} stems from a series of revelations that demonstrate the flaw in reasoning when powerful treatments are administered over a long term to large populations based on short-term studies or surrogate measures. A drug simply cannot be declared “safe” without measuring the balance of benefits and risks in a randomized controlled trial over an appropriate period of time in a large population representing those who will use the treatment in practice.¹⁹ Although the specific new knowledge imparted by the CREATE trial is important, its more important lasting legacy may be the demonstration that clear answers to the balance of benefit and risk of therapies can be achieved at a fraction of the current accepted cost of clinical trials.²⁰

At a time when so many new technologies with powerful effects, both beneficial and detrimental, are being pushed into practice, application of these concepts to trials of long-term medical care is an urgent matter for the public health.