OBJECTIVE--To assess the possible benefits of intravenous isosorbide dinitrate in acute myocardial infarction and oral isosorbide mononitrate in subacute myocardial infarction. METHODS--316 patients presenting with acute myocardial infarction were entered into double blind placebo controlled clinical trials assessing infarct size by enzyme release, ventricular size and function by echocardiography, reperfusion by continuous 12 lead ST segment monitoring and late potentials by high resolution electrocardiography. RESULTS--301 patients, of whom 292 (97%) received thrombolytic treatment, were randomised on admission to intravenous isosorbide dinitrate or placebo. Overall, there was no significant effect of treatment on infarct size, ST segment resolution, ventricular remodelling, or late potentials at day 3. A trend was observed towards a reduction in infarct size in patients with non-Q wave infarction treated with isosorbide dinitrate. Heterogeneity of nitrate effect was observed in relation to the degree of ST segment elevation on presentation with a clear benefit of isosorbide dinitrate in patients with moderate ST segment elevation (472 U/l v 704 U/l, P = 0.003) and a trend towards a deleterious effect in patients with marked ST segment elevation (1152 U/l v 1058 U/l, P = 0.2). ST segment re-elevation was more common among patients receiving nitrate treatment than in those assigned to placebo (29 v 16, P < 0.05). Some 160 patients underwent a further randomisation to sustained release isosorbide mononitrate or placebo on day 3. Echocardiographic volumes after 6 weeks of treatment were similar in the two groups. CONCLUSIONS--No benefit was observed with administration of nitrates in the treatment groups as a whole for either acute or subacute infarction. There was, however, evidence of heterogeneity of effect in the different subgroups of acute infarction, and the possibility that nitrates may have differing actions in different groups of patients should be considered.