To determine 1) whether the effect of intravenous nitroglycerin (NG) therapy during acute myocardial infarction on creatine kinase infarct size is influenced by infarct location (anterior vs. inferior), timing (therapy less than 4 hours vs. greater than or equal to 4 hours after onset of pain), and dose response (mean blood pressure greater than or equal to 80 mm Hg vs. less than 80 mm Hg during the first 12 hours) and 2) whether NG therapy modifies infarct expansion, 310 patients were randomly allocated to NG (n = 154) and control (n = 156) groups. NG infusion was titrated to lower mean blood pressure by 10% in normotensive and 30% in hypertensive patients, but not below 80 mm Hg, and was maintained for 39 hours. Measurements included clinical variables, creatine kinase infarct size (geq) as well as left ventricular (LV) asynergy, LV ejection fraction, expansion index, and thinning ratio on serial two-dimensional echocardiography. Compared with controls, creatine kinase infarct size was less in the NG group (41 vs. 55 geq, p less than 0.001), in anterior (44 vs. 58 geq, p less than 0.05), and inferior (39 vs. 53 geq, p less than 0.025) NG subgroups, and in early than late NG subgroups (43% vs. 22% decrease). Other indexes of infarct size also improved (p less than or equal to 0.05) with NG compared with controls. Thus, by 10 days, LV asynergy was 40% less, LV ejection fraction was 22% more, and Killip class score was 41% less. A negative effect of mean blood pressure less than 80 mm Hg with NG was reflected in these indexes. In addition, expansion index increased (p less than 0.001) by 31% and thinning ratio decreased (p less than 0.001) by 17% in controls by 10 days but remained unchanged with NG. Infarct-related major complications were less frequent in the NG than the control groups: infarct expansion syndrome (2% vs. 15%, p less than 0.0005), LV thrombus (5% vs. 22%, p less than 0.0005), cardiogenic shock (5% vs. 15%, p less than 0.005), and infarct extension (11% vs. 22%, p less than 0.025). Mortality was less in NG than in control groups in-hospital (14% vs. 26%, p less than 0.01), at 3 months (16% vs. 28%, p less than 0.025) and 12 months (21% vs. 31%, p less than 0.05), but this advantage was only found in the anterior subgroups.(ABSTRACT TRUNCATED AT 250 WORDS)