Three hundred fifty-two patients with suspected acute myocardial infarction (AMI) were randomized to placebo (175) or tissue-type plasminogen activator (rt-PA) (177). Patients were eligible if evaluated within 165 minutes from onset of chest pain and if age was less than 75 years. Electrocardiographic criteria were not required. A mobile coronary care unit with a cardiologist present was used to initiate treatment at home in 29% of the patients. Primary endpoints were infarct size (serum lactate dehydrogenase isoenzyme activity), left ventricular function (radioangiography) and exercise capacity at 30 days. AMI was diagnosed in 59% of all randomized patients. The incidence was similar in the 2 groups (placebo, 108, rt-PA, 101). Among all randomized patients, rt-PA was associated with significantly decreased infarct size and an increased ejection fraction. Among rt-PA-treated patients there were significantly fewer Q-wave infarctions. No difference in exercise capacity could be detected. No benefit was found in subgroups of patients without ST-segment elevation on the initial electrocardiogram. There were 18 (10.3%) and 11 (6.2%) deaths (p = 0.23) within 30 days in the placebo and rt-PA groups, respectively. Adverse reactions were similar in both groups with no excess of complications in the home-treated group. Very early treatment with rt-PA in patients with a strong suspicion of AMI and ST-segment elevation limits infarct size and improves left ventricular function. The infarct pattern is shifted from Q-wave to non-Q-wave infarcts by rt-PA. The study suggests that thrombolysis can be given before hospital admission without additional risk. Furthermore, electrocardiographic records are useful for selection of patients.