The relative value of exercise testing, late potentials and simple clinical assessment in predicting ischemic and arrhythmic events during follow-up after acute myocardial infarction (AMI) was investigated prospectively in a population of 176 consecutive patients surviving to 7 days after AMI. During 15 +/- 9 (range 3 to 24) months of follow-up, there were 23 ischemic events (2 fatal reinfarctions, 6 nonfatal reinfarctions and 16 patients who underwent coronary artery bypass grafting, 1 after reinfarction) and 11 arrhythmic events (7 symptomatic ventricular tachycardias and 4 sudden cardiac deaths). Stepwise multiple regression analysis showed that out of 11 variables, including exercise testing, late potentials and clinical data, exercise testing was the only independent variable predicting the occurrence of ischemic events (p less than 0.05 not including coronary artery bypass grafting and p less than 0.002 including it). Arrhythmic events were predicted, in order of importance, by Killip class (p less than 0.05), late potentials (p less than 0.005), previous AMI (p less than 0.009), occurrence of in-hospital complications (p less than 0.005) and non-Q-wave AMI (p less than 0.02). The presence of late potentials provided independent prognostic information from the Killip class and the result of exercise testing in predicting both arrhythmic and ischemic events. Exercise testing, late potentials and clinical assessment provide complementary prognostic information in postinfarction patients.