Using composite endpoints in clinical trials of cardiovascular therapies can be both problematic and misleading, a new review suggests.
The practice is often complicated by wide variations in the importance of composite endpoints to patients and the magnitude of treatment effects across the component endpoints, concluded an international team led by Ignacio Ferreira-González, research fellow in the Department of Medicine at the Universitat Autònoma de Barcelona in Spain. Higher event rates and more pronounced treatment effects associated with less important components may give the wrong impression of the impact of treatment, they warned.
The researchers conducted a systematic review of cardiovascular controlled trials published in six prominent medical journals between 1 January 2002 and 30 June 2003. Of the 114 randomised controlled trials that included a composite endpoint of importance to patients, 68% reported complete component data for the primary composite endpoint, while in 98% of cases the primary composite endpoint included a fatal endpoint, they found.
Of the 84 composite endpoints for which component data were available, 54% (45) showed large or moderate gradients in both importance to patients and the magnitude of effects across components. When these were analysed by categories of importance to patients, the most significant components were associated with lower event rates in the control group (medians of 3.3 -3.7% for fatal, critical and major outcomes, 12.3% for moderate outcomes and 8.0% for minor outcomes) and with more modest treatment effects than less important components (relative risk reduction of 8% for death but of 33% for components of minor importance to patients).
Reporting these results in the British Journal of Medicine (BMJ 2007; 334:786), the research team pointed out that clinical trials, particularly in cardiology, often use composite endpoints to reduce sample-size requirements and capture the overall impact of therapeutic interventions. While composite endpoints may increase the event rate and therefore statistical power of the study, though, they may also mislead if component endpoints are of widely differing importance to patients, the number of events in the components of greater importance is small, and the magnitude of effect differs markedly across components.
“For example,” the authors commented, “a statement that an intervention reduces a composite of cardiovascular mortality, myocardial infarction and revascularisation is problematic if most of the events were revascularisation procedures and investigators found a large apparent effect of treatment on revascularisation but not on death or infarction.”
A confident interpretation of composite endpoint outcomes calls for relatively small gradients of importance to patients and similar relative risk reductions across components, they wrote. Conversely, the review findings indicated that most composite endpoints used in randomised controlled trials of cardiovascular therapies have substantial gradients in both importance to patients and treatment effects across component endpoints. Moreover, less important outcomes contribute more substantially to the composite endpoint rate and show larger treatment effects.
A problem with mortality outcomes
In particular, the authors noted, “mortality outcomes, present in almost all cardiovascular composite endpoints, provide the lowest event rate and show the smallest treatment effects. Thus, an important and plausible risk of misleading conclusions associated with the use of composite endpoints is to attribute reductions in mortality to interventions that do not, in fact, reduce death rates”.
To guard against these misconceptions, the researchers suggested, trialists should report complete data on individual component endpoints and clinicians should view the results of cardiovascular studies using composite endpoints “with caution”.
In an editorial on the review, Nick Freemantle and Mel Calvert from the University of Birmingham’s Health Care Evaluation Group said composite outcomes “potentially provide an opportunity for sponsors to ‘game’ their trials”. They have statistical advantages, as in time-to-event analysis the statistical power of the study is driven by the number of events that accrue, rather than the number of patients randomised.
Composite endpoints can help in avoiding arbitrary decisions between different candidate outcomes when pre-specifying the primary outcome of a trial, Freemantle and Calvert acknowledged. However, a positive result for a composite outcome “applies only to the cluster of events included in the composite, and not to the individual components”.
Regulatory practice may have led to ‘death’ being added to many composite primary endpoints used in clinical trials, “and it is our experience that the [US] Food and Drug Administration has actively promoted the use of such composite outcome measures in heart failure trials”, they commented, adding that the European drug licensing process “seems to follow the FDA lead”.
In this light, Freemantle and Calvert highlighted two areas of concern. First, regulators should ensure that primary outcome variables in trials really do “provide a valid and reliable measure of some clinically relevant and important treatment benefit in the patient population”, as stipulated in guidance from the International Conference on Harmonisation. Second, the “difficult problem of the appropriate interpretation of composite outcome measures should be dealt with”.
One solution, they proposed, might be to use a corollary, such as a health warning, to make it “clear that on their own the individual components of a composite have not been shown to be affected by the experimental treatment”.