Objective:To use meta‐analysis to estimate the benefits of drug treatment to lower cholesterol levels in the primary and secondary prevention of coronary heart disease (CHD) events.Data sources:MEDLINE search from 1967 to 1990; bibliographies of review articles.Study selection:Nine trials met the entry criteria: they were monofactorial, randomised and controlled.Data extraction:Two independent, un‐blinded observers.Data synthesis:The odds ratio (and 95% CI) for death from CHD was 0.85 (0.64, 1.14) in primary prevention and 0.84 (0.75, 0.95) in secondary prevention studies when calculated by the method of Peto. The event rate in the secondary prevention studies was higher than that in the primary prevention studies, and the absolute risk reduction achieved by therapy in the former (3.2%) was much higher than that in the latter (0.1%). The number of subjects needing to be treated to prevent one death from CHD was 38 in secondary prevention and 675 in primary prevention. Results with the method of DerSimonian and Laird were similar.Conclusions:The benefits of cholesterol lowering to prevent death from CHD are substantially greater in the secondary prevention setting than in primary prevention.