In patients with a strongly positive exercise electro-cardiogram, the workload achieved during the test allows the identification of subsets with good or poor survival rates. To determine whether the same criteria also predict acute ischemic heart events such as unstable angina and myocardial infarction, fatal and nonfatal acute manifestations were documented in 241 patients medically treated during an 8-year follow-up. All patients had a Bruce protocol treadmill exercise test with ST-segment depression greater than or equal to 2 mm and coronary angiographic studies. There were 52 deaths; of these 44 were due to coronary artery disease. There were 41 episodes of unstable angina and 21 myocardial infarcts documented as first morbid events. As expected, survival improved with increased workload achieved; patients terminating their exercise at stage I (5.1 METs) had an 8-year survival rate of 45 +/- 9% while those reaching stage IV or more (10 METs) had a survival rate of 93 +/- 6%. In a multivariate analysis, the duration of exercise and the number of narrowed coronary arteries and of left ventricular segment abnormalities correlated significantly with survival. In contrast, nonfatal acute events occurred in about 20 to 35% of patients whatever the stage of the exercise test. Furthermore, neither variables during the exercise test nor angiographic findings predicted nonfatal events. Thus, although the workload achieved did identify patients with different mortality rates, it failed to predict subsets of patients with different morbid event rates.