The poor sensitivity and the poor predictive value of ST-segment depression have limited the usefulness of the exercise electrocardiogram (ECG) in the diagnosis and evaluation of coronary artery disease (CAD). The QT dispersion (QTD), recorded as the difference between maximal and minimal QT intervals on a 12-lead exercise ECG, is sensitive to myocardial ischemia and may improve the accuracy of exercise testing in patients with CAD who do not show an ST-segment depression. Exercise ECGs were analyzed in 50 subjects who had undergone coronary angiography for clinical indications. None of them showed an ST-segment depression during or after exercise: There were 25 patients with significant coronary artery stenosis and 25 without significant stenosis. The QTD measured before, immediately after, and 1 min after exercise was similar in the 2 groups. The QTD at 3 and 5 min after exercise was significantly greater in patients with CAD than in the controls, and the most marked difference in QTD was observed at 3 min after exercise. A QTD at 3 min after exercise of >60 ms had a sensitivity of 80% and specificity of 88% regarding the diagnosis of CAD. When a ΔQTD (post-exercise QTD minus QTD at rest) at 3 min after exercise of >0 ms was added to a QTD of >60 ms as a condition for positivity, the specificity increased to 96%. QTD measured at 3 min after exercise increases the accuracy of exercise testing in patients with CAD who do not show an ST-segment depression. (Jpn Circ J 1999; 63: 517 - 521)