Use and misuse of composite endpoints in randomised clinical trials
Joan Siquier-Padilla, Rafael González Manzanares, Xavier Cabrer Rosselló
Abstract
Composite endpoints are widely used in large randomised cardiovascular outcome trials. They are frequently referred to as major adverse cardiovascular event (MACE), although there is no consensus around this definition. In essence, composite endpoints are single measures of effect encompassing multiple individual events, so that if any of them occurs, the patient is considered to present the composite endpoint. Their popularity has grown because of their methodological advantages, such as statistical efficiency and better ability to capture multiple clinically relevant outcomes. However, its use comes at a cost. Many times, composite endpoints are driven by the less meaningful event, or simply dilute a potential treatment effect by including outcomes that are not affected by the intervention. This and other limitations are often overlooked, therefore having a direct impact on the interpretation of clinical practice-changing trials. This review discusses key aspects related to the definition, interpretation, use and misuse of composite endpoints. Alternatives to composite endpoints are also discussed. Essential concepts are illustrated through examples based on key landmark studies, as well as topical trials. This work aims to help future trialists in the design and reporting of cardiovascular trials, and to assist readers in developing a critical understanding of them.