Different exercise ECG lead systems were evaluated for the detection of multivessel coronary disease in 118 male survivors of a transmural myocardial infarction. Patients were classified according to the location of myocardial infarction and angina functional class and test results were correlated with angiographic findings. The ECG criteria for a positive test in any lead were horizontal or downsloping ST-segment depression ≥ 1 mm for 0.08 second, and a slow, upsloping ST segment depressed ≥ 1.5 mm at 0.08 second after the J point compared with the rest tracing.The sensitivity of the test for multivessel disease was greater using 14 ECG leads (72%) or leads CC5, CM, and V, (64%) than lead V5alone (50%) (p< 0.05). Sensitivity was less when the site of infarction was anterior (64%) vs inferior (77%). The predictive value of a positive test ranged from 50–100% and that of a negative test from 24–80%, depending on angina functional class and lead positivity during exercise. Leads CC5, CM5and V5gave data similar to data from 14 ECG leads. The maximum diagnostic value of exercise testing after infarction was for patients in angina class 0 or 1. A positive test in this subset increased the likelihood of multivessel disease from 50–55% to 80–100%, and a negative test reduced the risk of three-vessel disease to less than 10%. In patients with more severe angina, the post-test risk was only slightly more than the prevalence of multivessel disease, and false-negative tests were common. A positive test predicted multivessel disease more frequently when the site of infarction was inferior (89%) than anterior (73%). Exercise-induced chest pain, the number of positive ECG leads, ST-segment elevation, treadmill work time and maximum depth of ST-segment depression provided additional diagnostic information.We conclude that probability statements for multivessel disease from exercise test results after infarction are most useful in patients with few or no symptoms and that satisfactory results can be obtained by recording leads CC5, CM5, and V5,.