Surrogate Endpoint

A surrogate endpoint is a laboratory or imaging marker that a study measures instead of directly measuring an outcome that matters to the patient (heart attack, death, fracture, quality of life), such as cholesterol, HbA1c, a blood vitamin level or bone density. The Cochrane Handbook counts reliance on surrogate outcomes in the absence of patient-important outcome data as a source of indirectness that lowers the certainty of the evidence. ^[1]

Example: homocysteine

In a Cochrane review covering 15 RCTs (71,422 participants) that gave B6, B9 or B12 supplements to lower homocysteine levels, RR was 1.02 (95% CI 0.95–1.10; 12 trials) for myocardial infarction and 1.01 (95% CI 0.96–1.06; 11 trials) for death from any cause; for stroke, RR 0.90 (95% CI 0.82–0.99; 10 trials) was reported. ^[2] In other words, an effect on a marker cannot be assumed to carry over in the same way to the intended outcomes; reading the conclusion sentence together with the values in the summary table also reveals the small reduction in stroke.

Use on the site

If a claim rests on a surrogate marker, the grade belongs only to that marker and the row states that it is a biomarker. For example, in the NMN guide, the rise in blood NAD+ level is a separate “biomarker” row; this does not mean that a health outcome has improved. For how evidence grades are assigned: Levels of evidence.

For the general reading order: How to read a study?.