WHEN TO RECOMMEND aortic valve replacement for mildly symptomatic or asymptomatic patients with moderate-to-severe aortic regurgitation is a difficult clinical problem. Severe chronic aortic regurgitation is associated with a generally favorable prognosis.1' 2 Approximately 75% of the patients are alive 5 years and 50% 10 years after the diagnosis.2 However, when symptoms develop, the patient's condition often deteriorates rapidly. Without surgery, the patient usually dies within 5 years after the onset of angina and 2 years after developing heart failure. In this issue of Circulation, Henry and associates report the results of serial M-mode echocardiographic studies in 50 consecutive symptomatic patients who underwent aortic valve replacement for isolated, moderate-to-severe aortic regurgitation.10' 11 In these studies, several echocardiographic indices were predictive of late death due to congestive heart failure in symptomatic patients who underwent otherwise successful aortic valve replacement. Specifically, a preoperative LV end-systolic dimension (ESD) of > 55 mm and percent fractional dimension shortening (%,AD) < 25% indicated an increased incidence of perioperative morbidity and late postoperative congestive heart failure and death compared with patients with similar symptoms but less depressed LV systolic function. A combination of the two findings (ESD > 55 mm and %zAD < 25%) was the most powerful predictor; 11 of 13 patients (85%) with these findings either died (nine patients) or had evidence of severe congestive heart failure postoperatively. Five of the nine deaths occurred in the immediate postoperative period. This high perioperative mortality risk may not be applicable to patients currently undergoing aortic valve replacement, owing to newer operative techniques for preserving myocardial function.The data of Henry and associates are consistent with the LV cineangiographic studies by Borow and co-workers,8 who found that the immediate and longterm results of aortic valve replacement were poor in the patients in whom the LV end-systolic volume index was markedly elevated (> 90 ml/m2). In the latter study the preoperative LV ejection fraction was of less predictive value. In contrast, Clark and associates reported no perioperative mortality in 17 patients with symptomatic aortic regurgitation and impaired LV performance.9 Also, they found no difference in late mortality in patients with a preoperative cineangiographic end-systolic volume of > 110 ml/m2 compared with those who had a smaller end-systolic volume. However, the 5-year survival in patients with an ejection fraction of < 45% was 33 ± 19%, compared with 94 ± 6% (p < 0.001) in those with an ejection fraction > 45%. However, not all patients with preoperative impairment of LV performance do poorly after aortic valve replacement; late postoperative LV systolic function improves in as many as 50% of the patients.5-9Henry and co-workers also reported serial M-mode echocardiographic data in 37 initially asymptomatic patients with aortic regurgi...