We treated 3262 patients with intravenous recombinant tissue plasminogen activator (rt-PA) within four hours of the onset of chest pain thought to be caused by myocardial infarction. Of these patients, 1636 were then randomly assigned to treatment according to an invasive strategy consisting of coronary arteriography 18 to 48 hours after the administration of rt-PA, followed by prophylactic percutaneous transluminal coronary angioplasty (PTCA) if arteriography demonstrated suitable anatomy; 1626 patients were randomly assigned to treatment according to a conservative strategy, as part of which arteriography and PTCA were to be performed only in patients with spontaneous or exercise-induced ischemia. In the group assigned to the invasive strategy, PTCA was attempted in 928 of the 1636 patients (56.7 percent); the procedure was anatomically successful in 93.3 percent. In the group assigned to the conservative strategy, 216 patients (13.3 percent) underwent clinically indicated PTCA within 14 days of the onset of symptoms. Reinfarction or death within 42 days, the primary end point, occurred in 10.9 percent of the group assigned to the invasive strategy and in 9.7 percent of those assigned to the conservative strategy (P not significant). There were no significant differences between the two groups in the ejection fraction at rest or during exercise, either at hospital discharge or six weeks after randomization. Eleven of 582 patients (1.9 percent) who received 150 mg of rt-PA and 15 of 2952 patients (0.5 percent) who received 100 mg of rt-PA had intracranial hemorrhage. A subgroup of 1390 patients who were eligible for short-term intravenous beta-blockade were randomly assigned to receive 15 mg of intravenous metoprolol immediately, followed by oral metoprolol, or oral metoprolol begun on day 6. The ejection fraction and the incidence of death in the two groups were similar during the hospital period. Total mortality within the first 6 days and at 42 days was also similar. However, in the group that received intravenous metoprolol, 16 patients had nonfatal reinfarctions and 107 patients had recurrent ischemic episodes by six days after entry into the study, as compared with 31 and 147 patients, respectively, among those randomly assigned to deferred oral beta-blockade (P = 0.02 and P = 0.005, respectively); the latter comparison was considered statistically significant according to the study criteria.(ABSTRACT TRUNCATED AT 400 WORDS)