Background: Composite endpoints have become a popular strategy for assessing the impact of interventions in clinical trials, and it is important to clarify their merits and problems.

Objectives: To review systematically the literature pertaining to merits and demerits of composite endpoint use.

Methods: We searched, using a systematic approach, MEDLINE, EMBASE, and the Science Citation Index from inception until February 2006, and reviewed potentially informative textbooks. Eligible articles provided a commentary, analysis or discussion of composite endpoints for any of the following areas: a) rationale; b) interpretation or meaning; c) advantages; d) limitations or conceptual problems; d) recommendations for use. Fifteen articles and a textbook specifically addressed the issue. Efficiency of clinical trials to detect small effects and the potential capacity to assess the net clinical benefit of a therapy were the most commonly cited advantages of using composite endpoints. Risk of misinterpretation when heterogeneity exists among components in either patient importance or magnitude of effects of therapy was the most salient disadvantage. There were discrepancies between authors concerning the interpretation of the results in the presence of competitive risks among components and when the direction of the effect of therapy diverges among components.

Conclusions: Theoretical writings about composite endpoints provide insight into their strengths and their considerable weaknesses.