We gave sodium nitroprusside by intravenous infusion to 163 randomly selected patients during the first 24 hours after hospitalization for typical acute myocardial infarction, and we studied its effects on mortality at one week, on the incidence of cardiogenic shock, on clinical signs of left ventricular failure, and on peak levels of creatine kinase isoenzyme MB. A control group of 165 patients received standard medical treatment and infusion of 5 per cent glucose. The end point of the study was a significant reduction in mortality in the nitroprusside group; this was reached when five deaths had occurred in this group, as compared with 18 among the controls (P less than 0.05). The incidence of cardiogenic shock, clinical signs of left-heart failure, and mean peak levels of creatine kinase isoenzyme MB were all reduced (P less than 0.05). The results indicate that infusion of nitroprusside in the early phase of acute infarction limits complications, possibly by reducing infarct size. The drug was particularly effective in anterior-wall infarction.