The exercise ECGs of 30 patients with pure aortic regurgitation (AR) were compared with functional and geometric variables measured by echocardiography and radionuclide cineangiography. The 10 patients with positive ECG responses to exercise (greater than or equal to 0.1 mV additional downsloping or horizontal ST segment depression) were similar to the 20 patients with negative tests with respect to mean age, exercise duration, peak work load achieved, symptoms, and resting ECG findings. Patients with positive exercise tests had significantly reduced left ventricular (LV) ejection fractions at rest (44 +/- 4% vs 52 +/- 2%) and during peak exercise (38 +/- 3% vs 48 +/- 2%), lower fractional shortening at rest (27 +/- 1% vs 34 +/- 2%), higher end-systolic wall stress (150 +/- 18 vs 99 +/- 8 dynes/cm2 X 10(3], lower left ventricular relative wall thickness (0.26 +/- 0.01 vs 0.30 +/- 0.01), and greater end-systolic diameter (5.7 +/- 0.3 vs 4.5 +/- 0.2 cm) than patients with negative tests (p less than 0.05 for all comparisons). Among the 18 asymptomatic patients, positive tests were associated with lower resting fractional shortening, lower exercise ejection fraction, higher wall stress, and greater end-systolic diameter. These data demonstrate that a positive exercise ECG in aortic regurgitation identifies patients, even when asymptomatic, who have developed markedly abnormal left ventricular functional and geometric responses to volume load, while a negative exercise ECG identifies a subset of asymptomatic patients who are unlikely to have severe ventricular dysfunction.