The purpose of this study was to determine the usefulness of exercise electrocardiography in predicting the site of myocardial ischemia. Fifty-two patients were studied who had angiographically documented 1-vessel coronary artery disease (CAD) and exercise-induced reversible thallium-201 perfusion defects. The patients were divided into 2 groups: group I (28 patients) had left anterior descending CAD and group II (24 patients) had left circumflex or right CAD. There were no significant differences between the 2 groups in severity of coronary stenosis, heart rate and systolic blood pressure during exercise. The size of the perfusion defect was larger in group I than II (28 +/- 12% vs 19 +/- 10%, p less than 0.02). There was no significant difference between the 2 groups in the frequency of ST depression in the anterior, inferior or lateral electrocardiographic leads. ST depression occurred in 16 patients (57%) in group I and 11 patients (46%) in group II (difference not significant). The sensitivity of the exercise electrocardiogram was 52% using 12 leads, 50% using 3 leads (V3, V5 and aVF) and 50% using V5 alone (difference not significant). Thus, the site of ST depression during exercise is not a good predictor of the site of exercise-induced perfusion defect or anatomic site of CAD. The use of 12 leads does not improve the sensitivity of exercise electrocardiography in patients with CAD.