Echocardiographic and Doppler-derived measurements of left ventricular (LV) function at rest were examined as predictors of maximal bicycle exercise capacity in a homogeneous group of 115 patients with mild to moderate LV dysfunction (ejection fraction 22 to 56%, median 43%) participating in the DEFIANT study of nisoldipine after acute myocardial infarction. Although the relations were not exact, peak exercise work load 7 weeks after infarction correlated with measurements of diastolic LV function at rest. Exercise work load was inversely related to peak late diastolic transmitral blood flow velocity (A wave) (slope -86.6; 95% confidence interval -120.9 to -52.2) and directly to the E/A ratio (slope 20.5; 95% confidence interval 6.0 to 35.1). The relations between exercise work load and peak late diastolic flow velocity remained significant after correction for age, sex, heart rate at rest, and use of beta-blocking drugs or nisoldipine. There was no relation between peak exercise work load and peak early diastolic transmitral flow velocity (E wave), isovolumic relaxation period or deceleration time. Measurements of systolic LV function (LV end-diastolic and end-systolic volumes, and ejection fraction, stroke volume and cardiac index) were also not significant as predictors of exercise capacity.