Electrocardiograms were recorded during and at 2-min intervals following maximal treadmill exercise in 100 patients and were correlated with coronary arteriograms, left ventricular cineangiograms, and resting and exercise cardiac hemodynamics. The incidence and extent of exercise-induced “ischemic’ S-T segment depression increased significantly (P< 0.01) with increasing extent of coronary artery disease (CAD). A criterion of abnormality of 1.0 mm or greater S-T depression most accurately predicted the presence of CAD with a specificity (true negative) of 89% and a sensitivity (true positive) of 62%. Criteria of ≧0.75 and ≧0.5 mm S-T depression offered improved sensitivities (68 and 84%, respectively) but markedly reduced specificities (78 and 57%, respectively).A significant correlation (P< 0.005) was found between increasing extent of exercise-induced S-T segment depression and the pulmonary artery wedge pressure during exercise. There were no correlations between the presence or extent of exercise-induced S-T depression and the resting or exercise pulmonary artery pressure, cardiac index, stroke-volume index, or exercise factor, or the resting left ventricular dp/dt, left ventricular end-diastolic pressure, left ventricular ejection fraction, or presence of asynergy on the left ventricular cineangiogram.The maximal treadmill exercise test provided a noninvasive method for predicting the presence and extent of significant CAD in the individual patient with chest pain. Although a normal treadmill exercise test provided little insight into the underlying resting and exercise hemodynamics, an abnormal treadmill test (≧1.0 mm S-T depression) was associated with an abnormal increase in left ventricular filling pressure with exercise in 90% of the patients.