Intravenous rtPA (total dose, 100 mg over 3 h) was compared with placebo in a prospective, randomized, double-blind trial in 5,011 patients with suspected AMI of less than 5 h duration. No ECG or enzymatic confirmation of the diagnosis was required for study entry. At 1 month 9.8% of patients given placebo had died compared with 7.2% of those who received rtPA (2.6% actual reduction, 26% relative reduction, with 95% confidence intervals of 11-39%). The majority of deaths occurred in patients who had an in-hospital diagnosis of MI (72% in both groups), with a 1-month infarct mortality of 13.1% in the placebo limb and 9.4% in the rtPA limb (relative reduction 28%, 95% CI, 14-41%). Approximately 18% of patients in both groups had a normal ECG on entry to the trial, and at 1 month the fatality was 1.6% in the rtPA group and 3.0% in the placebo group. Treatment with rtPA did not reduce the number of patients with normal ECGs from developing MI (28% rtPA vs 24% placebo). Treatment with rtPA was associated with significantly more bleeding episodes, the vast majority of which were clinically minor. The risk of all strokes in the rtPA group was similar to that in the placebo group (1.1% vs 1.0%). Treatment with rtPA was unaccompanied by either allergic or hypotensive episodes, and, among rtPA treated patients, there was no increase in clinically important ventricular dysrhythmias. Neither age nor time from onset of symptoms reduced the benefit from rtPA.