A composite endpoint contains more than one outcome

A trial headline may say a treatment reduced cardiovascular events. The underlying endpoint may combine several kinds of event into one analysis. To understand the claim, read the definition of that combined outcome before interpreting its size.

SELECT's original abstract describes a primary endpoint combining cardiovascular death, nonfatal myocardial infarction and nonfatal stroke, analyzed as time to a first event. That definition is more specific than a broad promise of “heart protection.” It tells readers what the primary result actually counted.

Keep the components visible

A result for the combined endpoint is not automatically the result for each component separately. If a summary claims that every component fell by the same amount, look for the relevant analyses. The primary composite alone cannot establish that statement.

The distinction also matters when two studies use different endpoint definitions. Similar labels can hide different combinations, follow-up periods or populations. Read the full definition rather than assuming that every cardiovascular outcome measure is interchangeable.

Distinguish the event proportion from the time-based comparison

The SELECT abstract reports primary events in 6.5% of the semaglutide group and 8.0% of the placebo group, alongside a hazard ratio of 0.80. The percentages describe observed event proportions, while the hazard ratio belongs to the time-to-event analysis. They are related summaries, not interchangeable labels for the same calculation.

Our relative and absolute risk guide discusses this example with the population and follow-up retained. It should not be used to calculate a personal probability or predict benefit for someone outside the studied group.

Ask what the abstract leaves unresolved

This desk's linked summary uses the original abstract. It does not establish the full pattern of component outcomes, subgroup findings or sensitivity analyses. Those require the relevant paper sections and, where needed, supplementary materials.

Writing “not assessed in this reading” is more accurate than assuming that an unreported component had no effect or the same effect as the composite. A missing detail in a short abstract is a boundary on the summary, not proof about the trial itself.

Watch the claim as it moves into advertising

A provider may cite a cardiovascular trial while selling access to a different product or service. The endpoint does not verify that provider's own outcomes. The study population, preparation and use remain part of the evidence boundary.

Use the marketing-claim guide to follow that transfer. A precise endpoint description makes a finding more useful because it preserves the question the researchers actually answered, without expanding it into every possible cardiovascular benefit.

Return to the evidence notes →