Out of 160 prospectively followed patients with aortic regurgitation, the clinical courses of 53 patients with pure, severe, and chronic aortic regurgitation and without coronary artery disease who were selected for surgery on the basis of predefined criteria is discussed. Surgical criteria were either unequivocal symptoms or documentation of impaired left ventricular dysfunction (defined as angiographic ejection fraction of less than 50% plus and end-systolic volume index greater than 60 ml/m2). According to preoperative status, patients were divided as follows: 11 asymptomatic patients (group A), 30 patients with moderate (classes II to III) symptoms (group B), and 12 patients with dyspnea at rest and pulmonary edema when first seen (group C). Surgical mortality was one patient (from group C). Late death occurred in four patients (one from group B, three from group C). At the end of follow-up (minimum 1 year, mean 3.6 years) 41 patients were in functional class I, four patients in class II, and one patient in class III. All patients except one in functional classes II and III belonged to group C. Before surgery, patients from groups A and B had similar ventricular dimensions and ejection fractions, whereas patients from group C had larger end-systolic diameters and volumes and lower ejection fractions. End-diastolic and end-systolic diameters decreased significantly at 1 and 2 years after surgery. Patients from group C continued to have dilated hearts as did those patients from groups A and B who had preoperative end-systolic diameters greater than 55 mm.(ABSTRACT TRUNCATED AT 250 WORDS)