First-pass radionuclide angiocardiography was used to access the left ventricular response to submaximal exercise in 150 patients with coronary artery disease (CAD). To test the hypothesis that resting systolic left ventricular function as determined by left ventricular ejection fraction (LVEF) was a predictor of the hemodynamic response to exercise, the study group included patients with a wide range of resting LVEF (0.12-0.82). The influences of resting LVEF, resting end-diastolic volume, the number of diseased vessels, exercise end point and exercise double product on the change in LVEF (delta LVEF) during exercise were tested using multiple linear regression analyses. Resting LVEF was a highly significant predictor of the delta LVEF (p = 0.0001). Exercise duration was not related to either the resting LVEF or the delta LVEF. For the 112 patients in whom coronary anatomy was known, resting LVEF retained its significance as a predictor of delta LVEF (p = 0.002) even after adjustment for the significance of the extent of CAD (p = 0.0007) and the exercise end point (p = 0.06). Patients with normal resting LVEF showed the most profound decreases in LVEF, the highest frequency of new regional dysfunction and the largest relative increase in end-diastolic volume during exercise. As rest LVEF decreased, the magnitude of the delta LVEF and the frequency of new regional dysfunction decreased. Therefore, left ventricular function at rest is an important determinant of the direction and magnitude of change in left ventricular function during exercise.