A rate-related change in ST-segment depression with exercise (ST/HR slope) of 6.0 microV/beat/min or more has been proposed as an accurate predictor of 3-vessel coronary artery disease (CAD). To further assess the accuracy and functional correlates of this method, exercise electrocardiograms were compared with radionuclide rest and exercise left ventricular (LV) ejection fraction (EF) and angiography in 35 patients with stable angina. The ST/HR slope was significantly increased in patients with 3-vessel CAD. An ST/HR slope of 6.0 or more identified 3-vessel CAD with a sensitivity of 89% and specificity of 88%. The predictive value for 3-vessel CAD was 73% owing to the presence of 3 false-positive slopes. The patients from whom these slopes were derived had functionally severe 2-vessel CAD, with an average decrease in exercise LVEF of 13%. Two of these 3 had additional left main CAD and the third has unsuspected additional aortic regurgitation. For the entire group, the exercise ST/HR slope was linearly related to the exercise change in LVEF (r = -0.55, p less than 0.001). Mean exercise change in LVEF for stable angina patients with ST/HR slopes of 4.5 or more was significantly different from that for patients with lower ST/HR slopes (-12 +/- 1% vs + 2 +/- 2%, p less than 0.0001). Thus, the ST/HR slope is both sensitive and specific for the identification of 3-vessel CAD, and high ST/HR slopes in patients with less extensive anatomic disease may predict functionally severe ischemia.