The occurrence of ventricular late potentials in survivors of acute myocardial infarction treated with intravenous streptokinase was compared with that in a conservatively treated group and the relation between ventricular late potentials and patency of the infarct related artery was examined. Of 115 patients admitted with a first infarct, 55 were treated with intravenous streptokinase (streptokinase group) and 60 were treated conservatively (non-streptokinase group). A signal averaged electrocardiogram was recorded in all patients and coronary angiography was performed in 45 (81.8%) of the streptokinase group and in 21 (35%) of the non-streptokinase group. At a 40 Hz filter setting ventricular late potentials were significantly less common in patients treated with streptokinase (9 (16.4%) of 55) than in those who were not (26 (43.3%) of 60). A total of 66 patients underwent angiography. Of the 26 who had closed infarct-related arteries, 17 had ventricular late potentials at a 40 Hz filter setting (sensitivity 65.4%, specificity 95%) and 38 of the 40 patients with a patent infarct-related artery did not have ventricular late potentials (sensitivity 80.9%, specificity 89.5%). Patients with acute myocardial infarction treated with intravenous streptokinase were significantly less likely to have ventricular late potentials than conservatively treated patients and the absence of ventricular late potentials at 40 Hz filter setting was a good non-invasive predictor that the infarct-related artery was patent.