The effect of aortic valve replacement on exercise capacity was evaluated in 28 patients with aortic regurgitation using repeat determinations of peak oxygen uptake (VO2) and the anaerobic threshold (AT), based on a nonlinear increase in the plot of pulmonary ventilation versus VO2. Although the AT was on average 68% of peak VO2 before and 1 year after surgery, the test-to-test correlation coefficient was 0.92 for the AT but only 0.68 for peak VO2 postoperatively. 11 patients (39%) improved in AT, usually within the first 6 months. The preoperative echocardiographic left ventricular end-diastolic dimension and its postoperative change correlated with that in the AT (r = 0.64 and -0.60, p less than 0.001) but not with that in peak VO2. The patients likely to improve had a lower AT (p less than 0.05) preoperatively, however. Thus repeat determinations of AT using respiratory measurements give additional data on changes in functional status. Sequential exercise testing does not require the patient to be exercised much past his AT.