To investigate the role of left ventricular (LV) diastolic function in the maintenance of exercise capacity in patients with systolic dysfunction, symptom-limited cardiopulmonary exercise testing combined with radionuclide ventriculography was performed in 24 patients with an LV ejection fraction <35% after anterior myocardial infarction. The ratio of pulmonary artery wedge pressure (PAWP) to LV end-diastolic volume (EDV), an index of global diastolic function, correlated significantly with peak oxygen consumption at peak exercise (r = −0.55; p = 0.006), whereas ejection fraction at peak exercise did not. The change in PAWP/EDV ratio from rest to peak exercise was related to the increases in stroke volume (r = −0.54; p = 0.006) and cardiac output (r = −0.51; p = 0.01) during exercise, but the change in ejection fraction was not. Resting hemodynamics did not differ between patients with preserved exercise capacity (group 1,n = 8) and those with exercise impairment (group 2,n = 16). At peak exercise, stroke volume, cardiac output, and EDV were significantly higher, and PAWP and PAWP/EDV ratio were significantly lower in group 1 than in group 2, but ejection fraction and end-systolic volume were similar in both groups. Although the incidences of hypertension, LV hypertrophy, and infarct-related coronary artery lesions did not differ between the two groups, group 2 had a significantly higher incidence of non-infarct-related coronary artery lesions than group 1 (p < 0.05). Thus in patients with LV systolic dysfunction after anterior myocardial infarction, the major cause of exercise impairment and failure to increase LV performance during exercise was diastolic dysfunction associated with the presence of non-infarct-related coronary artery lesions with the potential for exercise-induced ischemia of the noninfarcted areas. (Am Heart J 1997;133:87-93.)