Estimating left ventricular wall stress has recognized applications, but formulae for global stress cannot be applied to ischemic ventricles. A mathematic method for estimating regional stress in infarcted ventricles has been described. The hypothesis tested was that exercise-induced ischemia increases end-systolic wall stress. Subcostal four-chamber echocardiograms were recorded at rest and during peak symptom-limited exercise in 19 controls and 41 patients with chest pain undergoing coronary arteriography. Centerline regional wall motion and regional end-systolic wall stress were measured at rest and at peak exercise. The normal controls had increased wall motion with exercise, but wall stress remained low. All 32 of the patients with coronary artery disease (> or = 50% diameter narrowing) had wall motion abnormalities with exercise, but the sensitivity of identifying right coronary artery obstructions was poor. Patients with coronary disease had higher regional stress at peak exercise than did the controls. The sensitivity of identifying lesions in all three coronary arteries (0.95 to 1.0) was better than that for wall motion (p < 0.04). The specificity of wall stress needs to be tested in a larger population. Exercise-induced ischemia causes increased regional end-systolic wall stress that reflects its distribution in patients with coronary artery disease. These changes can be measured non-invasively during exercise echocardiography.