Since the last comprehensive review of anticoagulation in acute myocardial infarction four additional randomized control trials have been reported. The overwhelming majority of all trials favored anticoagulation. Rates of thromboembolism were higher in the control, and hemorrhagic complications in the anticoagulated group. Pooling of all randomized control trials gives mean case fatality rates of 19.6 per cent for the control and 15.4 per cent for the anticoagulated group, a relative reduction of 21 per cent (P<0.05 or <0.001, depending on the analytic method). Five of six randomized control trials reported "no effect" because the difference favoring anticoagulation was not statistically significant. However, sample sizes in these "negative" papers were too small to protect against missing a 21 per cent reduction in true case fatality rate due to anticoagulation (β<0.10). All patients who present no specific contraindication should receive anticoagulants during hospitalization for infarction. (N Engl J Med 297:1091–1096, 1977)