Three series of patients undergoing isolated aortic valve replacement with the standard glutaraldehyde-preserved Hancock porcine bioprosthesis, the standard Björk-Shiley spherical disc prosthesis, and the Lillehei-Kaster tilting disc prosthesis were compared during a 15-year follow-up. From March 1970 to December 1985, aortic valve replacement was performed in 506 patients, 379 men and 109 women, whose ages ranged from 17 to 71 years (mean, 47 +/- 12 years); 196 had a Hancock bioprosthesis, 147 a Björk-Shiley prosthesis, and 163 a Lillehei-Kaster prosthesis. There were no significant differences among the three groups in terms of mean age, preoperative New York Heart Association class, valve lesion, associated intracardiac procedures, and operative mortality (11.7% for the Hancock bioprosthesis group, 8.8% for the Björk-Shiley prosthesis group, and 9.2% for the Lillehei-Kaster prosthesis group). Cumulative duration of follow-up is 1140 patient-years (mean, 6.6 +/- 1.5 years, 100% complete) for Hancock bioprosthesis patients, 878 patient-years (mean, 5.9 +/- 0.4 years, 98% complete) for Björk-Shiley prosthesis patients, and 1184 patient-years (mean, 7.3 +/- 0.3 years, 97% complete) for Lillehei-Kaster prosthesis patients. Actuarial survival rates at 15 years are 30% +/- 15% for Hancock bioprosthesis, 49% +/- 6% for Björk-Shiley prosthesis, and 44% +/- 7% for Lillehei-Kaster prosthesis patients. There were eight embolic episodes in Hancock valve recipients (0.7% +/- 0.1%/pt-yr), nine in Björk-Shiley prosthesis recipients (1.0% +/- 0.3%/pt-yr), and nine (0.7% +/- 0.1%/pt-yr) in Lillehei-Kaster prosthesis recipients (p = no significant difference). Actuarial freedom from emboli at 15 years is 89% +/- 4% for Hancock, 92% +/- 3% for Björk-Shiley, and 93% +/- 2% for Lillehei-Kaster prosthesis recipients (p = no significant difference). A significant difference was found in the incidence of the following complications. (1) Anticoagulant-related hemorrhage: 0.1% +/- 0.1%/pt-yr for Hancock bioprosthesis, 1.3% +/- 0.3%/pt-yr for Björk-Shiley prosthesis, and 1.9% +/- 0.4%/pt-yr for Lillehei-Kaster prosthesis patients (p less than 0.001), with an actuarial freedom at 15 years of 97.6% +/- 1% for Hancock, 85% +/- 5% for Björk-Shiley, and 68.8% +/- 10% for Lillehei-Kaster recipients (p less than 0.001). (2) Structural deterioration: 3.7% +/- 0.6%/pt-yr for Hancock and none for Björk-Shiley and Lillehei-Kaster patients (p less than 0.001), with an actuarial freedom of 100% for Björk-Shiley prosthesis and Lillehei-Kaster prosthesis patients and 38.4% +/- 12% for Hancock bioprosthesis patients at 15 years.