Forty-five patients underwent aortic valve replacement (AVR) for severe isolated aortic regurgitation from 1973 to 1979. There were two (4.4%) hospital deaths, both functional class IV. Six patients with mechanical prosthesis not receiving anticoagulants were excluded from further analysis. These data relate to 39 patients; the two operative deaths, 35 patients with mechanical prosthesis receiving anticoagulants, and two with bioprosthesis. There were three late cardiac deaths with 5-year survival 85%; average annual mortality rate of 3%. The 5-year survival with pre-AVR left ventricular (LV) ejection fraction greater than or equal to 0.45 was 87% vs 54% less than 0.45, (p less than 0.04); cardiac index greater than or equal to 2.5 L/min/m2 92% vs 66% less than 2.5 (p less than 0.04); mean VCF greater than or equal to 0.75 vs less than 0.75 circ/sec (p less than 0.09); end-diastolic pressure less than or equal to 20 vs greater than 20 mm Hg (p less than 0.08). Late survival was not significantly different between pre-AVR functional class I and II vs class III and IV; LV end-diastolic volume index greater than or equal to 210 vs less than 210 ml/m2; LV end-systolic volume index greater than or equal to 110 vs less than 110 ml/m2; and LV mass greater than or equal to 240 vs less than 240 gm/m2. With ejection fraction greater than or equal to 0.50 there was only one operative death (functional class IV) and no late cardiac deaths. Thus late survival following aortic valve replacement for severe isolated aortic regurgitation is better predicted preoperatively by the LV systolic pump function variables of ejection fraction and cardiac index than by LV diastolic parameters and clinical status.