Patients presenting within four hours of the onset of acute myocardial infarction were randomly assigned to receive 80 to 100 mg of recombinant human-tissue plasminogen activator (t-PA) intravenously over a period of three hours (n = 72) or placebo (n = 66). Administration of the study drug was followed by coronary arteriography, and candidates for percutaneous transluminal coronary angioplasty were randomly assigned either to undergo angioplasty on the third hospital day (n = 42) or not to undergo angioplasty during the 10-day study period (n = 43). The patency rates of the infarct-related arteries were 66 percent in the t-PA group and 24 percent in the placebo group. No fatal or intracerebral hemorrhages occurred, and episodes of bleeding requiring transfusion were observed in 7.6 percent of the placebo group and 9.8 percent of the t-PA group. As compared with the use of placebo, administration of t-PA was associated with a higher mean (+/- SEM) ejection fraction on the 10th hospital day (53.2 +/- 2.0 vs. 46.4 +/- 2.0 percent, P less than 0.02), an improved ejection fraction during the study period (+3.6 +/- 1.3 vs. -4.7 +/- 1.3 percentage points, P less than 0.0001), and a reduction in the prevalence of congestive heart failure from 33 to 14 percent (P less than 0.01). Angioplasty improved the response of the ejection fraction to exercise (+8.1 +/- 1.4 vs. +1.2 +/- 2.2 percentage points, P less than 0.02) and reduced the incidence of postinfarction angina from 19 to 5 percent (P less than 0.05), but did not influence the ejection fraction at rest. These data support an approach to the treatment of acute myocardial infarction that includes early intravenous administration of t-PA and deferred cardiac catheterization and coronary angioplasty.