CORRESPONDENCE: LETTER TO THE EDITOR
Keeping Apples and Oranges Separate: Reassessing Clinical Trials That Use Composite End Points as Their Primary Outcome
Paul S. Chan, MD, MSc*,
Brahmajee K. Nallamothu, MD, MPH and
Rodney A. Hayward, MD
* Ann Arbor Veterans Affairs Health Services, Research and Development, P.O. Box 130170, Ann Arbor, Michigan 48113-0170 (Email: paulchan{at}umich.edu).
It has become increasingly common for clinical trials in cardiovascular medicine to use composite primary end points that often have varying clinical importance. Time-to-event analyses, in which mortality is but one of several outcomes, are particularly problematic as the impact of death is clearly not equivalent to other nonfatal outcomes such as rehospitalization (1). Another concern is that indiscriminately combining mortality with other outcomes in survival analyses may lead to biases related to "competing risk" (2). Specifically, patients dying early in a clinical trial are unable to experience future nonfatal outcomes.
To overcome these concerns, Cleland et al. (3) have used a novel approach that relies on a composite outcome, namely "days lost due to death or hospitalization," over their eight-month follow-up period. (They also separately reported one-year all-cause mortality.) By developing this composite outcome, the investigators have avoided directly equating rehospitalization with death as in a traditional event-free survival analysis, while at the same time allowing for patients to have repeated rehospitalizations. This approach allows clinicians to better appreciate the effect of home telemonitoring on their composite end point and its overall clinical and economic consequences.
Because of recent advances in cardiovascular medicine and lower short-term mortality rates, it has become necessary for many clinical trials to incorporate composite end points into their protocols as a primary outcome in order to demonstrate biological efficacy and statistical significance (1,2). We applaud Cleland et al. (3) for using an innovative and rational composite end point to evaluate a clinically relevant outcome in high-risk patients with heart failure. Their meticulous reporting of each component of their composite end point is also important and should be commended.
Of course, we wonder whether additional methodologies such as "weighting" individual components of a composite end point—days dead are counted more heavily than days hospitalized, for example—may further improve these approaches. Such a strategy of weighting outcomes was used effectively in the recent A-HeFT (African-American Heart Failure Trial) study (4). Perhaps at a minimum, separate reporting of individual components of a composite primary end point should become routine and even integrated into the CONSORT (Consolidated Standards of Reporting Trials) guidelines (1).
 |
References
|
|---|
- Freemantle N, Calvert M, Wood J, Eastaugh J, Griffin C. Composite outcomes in randomized trialsgreater precision but with greater uncertainty?. JAMA 2003;289:2554-2559.[Abstract/Free Full Text]
- Lauer MS, Topol EJ. Clinical trials—multiple treatments, multiple end points, and multiple lessons JAMA 2003;289:2575-2577.[Free Full Text]
- Cleland JG, Louis AA, Rigby AS, Janssens U, Balk AH. Noninvasive home telemonitoring for patients with heart failure at high risk of recurrent admission and deaththe Trans-European Network–Home-Care Management System (TEN–HMS) study. J Am Coll Cardiol 2005;45:1654-1664.[Abstract/Free Full Text]
- Franciosa JA, Taylor AL, Cohn JN, et al. African-American Heart Failure Trial (A-HeFT)rationale, design, and methodology. J Card Fail 2002;8:128-135.[CrossRef][ISI][Medline]
|
Top
References