Composite Endpoint

A composite endpoint treats the first occurrence of any one of several specified events as a single outcome, so a participant contributes an event whether they had a heart attack, a stroke, or died of cardiovascular causes. The best-known version is three-point MACE: cardiovascular death, non-fatal myocardial infarction and non-fatal stroke. The motivation is arithmetic, since combining components raises the event rate and the required sample size falls roughly in proportion to how common the endpoint is.

Metabolic drug development leans on it heavily. The SELECT trial randomised more than 17,000 people with obesity and established cardiovascular disease to semaglutide 2.4 mg or placebo and reported roughly a 20 percent relative reduction in a composite MACE endpoint, published in 2023 and supporting a label expansion for cardiovascular risk reduction the following year. Earlier programmes such as LEADER and EMPA-REG OUTCOME used the same structure.

A composite is defensible when its components are of similar severity, similar frequency, and expected to move in the same direction for the same biological reason. Where that holds, the summary estimate fairly describes the treatment. Where it does not, the composite can be driven almost entirely by its mildest and most frequent component while the headline implies something about the most severe.

The classic misreading is inferring a mortality benefit from a composite that merely included mortality. If death is one of four components and the effect sits in hospitalisation or revascularisation, nothing has been shown about survival, yet promotional summaries routinely say lives were saved. A related trap mixes a hard event with an investigator-triggered one such as unplanned revascularisation, which is vulnerable when blinding is imperfect.