Stroke is a leading cause of death in the industrialized world, and hypercholesterolemia may be a risk factor for stroke. To determine whether reducing cholesterol levels with HMGcoA (3-hydroxy-3-methylglutaryl coenzyme A) reductase inhibitors or other antilipidemic interventions reduces risk for nonfatal and fatal stroke. A systematic search in the MEDLINE and EMBASE databases of the English-language and non-English-language literature published from 1966 through October 1996. All randomized, controlled trials of any cholesterol-lowering intervention that reported data on nonfatal and fatal strokes, on death from coronary heart disease, and on overall mortality were included. Whether treatment effects differed according to the type of cholesterol-lowering intervention used was investigated. Trials were reviewed for methods, inclusion and exclusion criteria, and outcomes. 28 trials (for a total of 49,477 study participants in the intervention group and 56,636 participants in the control group) were included. The risk ratio for nonfatal and fatal stroke with HMGcoA reductase inhibitors was 0.76 (95% CI, 0.62 to 0.92; test of heterogeneity, P > 0.2). The risk ratios for nonfatal and fatal stroke with fibrates, resins, and dietary interventions were all close to 1.0, and the difference between the HMGcoA reductase inhibitor effect and the pooled estimate for all other interventions would, under the null hypothesis, be unlikely to occur by chance (P = 0.01). Trials with HMGcoA reductase inhibitors also showed reductions in rates of death from coronary heart disease and overall mortality. This meta-analysis of randomized, controlled trials suggests that in hyperlipidemic patients who have not previously had stroke, HMGcoA reductase inhibitors reduce the incidence of stroke.