Clinical research & EBM · Reference
What is a composite endpoint?
A composite endpoint is a single outcome that combines several individual events — counting a participant as having met it if any of the component events occurs. Composites increase the number of events and statistical power, but they must be interpreted with care when the components differ in importance.
Combining several events into one outcome
Instead of analysing several outcomes separately, a trial can define a composite that treats the first occurrence of any of them as the event of interest. The methodological motivations are practical: combining events yields more events overall, which increases the statistical power to detect a difference without needing an even larger trial, and it sidesteps the multiplicity problem of testing many outcomes one by one. It can also express a broader notion of benefit — capturing the overall burden of a condition rather than any single manifestation of it.
Interpretation cautions
The main pitfall is that the components of a composite are often unequal in both frequency and importance. A composite may be driven largely by its most common but least serious component, while a more serious but rarer component contributes little to the total — so an apparently positive composite can mask the fact that the events most important to patients were unaffected. Careful reading therefore inspects the individual components alongside the composite, checking that the effect is reasonably consistent across them and not concentrated in the least meaningful one.
Good practice in reporting
Sound use of a composite endpoint depends on defining it and its components in advance in the trial protocol, and on reporting each component separately as well as the combined result. Components should ideally be of broadly similar importance and plausibly affected in the same direction by the intervention. The appraisal of a composite outcome — like that of any endpoint — is part of judging a study’s relevance, and good reporting under standards such as CONSORT makes the breakdown transparent so readers can weigh it themselves.
Key facts
At a glance
- Definition: One outcome combining several component events
- Met when: Any one component event occurs
- Benefit: More events and greater statistical power
- Main pitfall: Driven by the commonest, least serious component
- Good practice: Report each component separately
- Ideal: Components of similar importance and direction
Common questions
FAQ
What is a composite endpoint in a clinical trial?+
A composite endpoint is a single outcome made by combining several events, where a participant counts as having reached it if any one component occurs. It is used to accumulate more events and increase statistical power, and to capture an overall measure of benefit rather than a single manifestation.
Why can composite endpoints be misleading?+
Because the components often differ in how common and how serious they are, a composite can be driven mainly by its most frequent but least important component. An apparently positive result may then mask the fact that the events most important to patients were not affected, which is why components should be reported separately.
How should a composite endpoint be reported?+
It should be defined in advance in the protocol, with components of broadly similar importance that are plausibly affected in the same direction. Good reporting presents each component separately alongside the combined result so readers can see whether the effect is consistent across them.
Going deeper
Related on CASRAI
- Surrogate endpoint →
- Randomized controlled trial →
- Hazard ratio →
- Levels of evidence →
- Clinical research hub →
Sources
The step most authors miss
Doing CRediT right? Don’t stop at the statement.
A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.
Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.
