Purpose—To determine if attenuated early treadmill stress testing is more or less accurate than cinecoronary arteriography in the identification of high-risk acute myocardial infarction (AMI) survivors. Patients and Methods—In a prospective study 96 non-selected and consecutive AMI survivors underwent cinecoronary arteriography in the 4th week of hospitalization and were followed up for a period of 3 to 39 (21.7 ± 11.0) months. One-half of the patients were randomly submitted to an attenuated heart rate-limited treadmill stress testing in the 3rd week. Seven patients were lost in the follow-up. Of the remaining 89 patients 5 died of cardiac causes. Cardiac events (death, reinfarction, angina and heart failure) ocurred in 26 patients. Multivessel coronary artery disease was found in 67 patients (75%) and mean left ventricular ejection fraction was 49.1 ± 15.6%. Positive treadmill stress testing for myocardial ischemia ocurred in 20 of 43 patients (46%). Results—All patients who died of cardiac causes or who had a cardiac event had multivessel coronary artery disease, the mean left ventricular ejection fraction was 38.6 ± 16.6% and 45.2 ± 16.4%, respectively, and two-thirds of patients had positive stress testing. For cardiac death, multivessel coronary artery disease had better sensitivity than positive stress testing and left ventricular ejection fraction less than 0.3 (100% vs 67% vs 20%, respectively) while ejection fraction had better specificity than stress testing and multivessel coronary artery disease (87% vs 55% vs 21%, respectively). Stress testing had a better positive predictive value...