The investigational Carpentier-Edwards supra-annular valve was implanted in 592 patients from November, 1981, to February, 1984 (aortic valve replacement in 286, mitral valve replacement in 259, and multiple valve replacement in 47, for a total of 638 prostheses). A previous cardiac operation had been performed in 77 patients (13%). Concomitant procedures were performed in 202 patients (34.1%), including coronary artery bypass in 163 patients. The patient evaluation was 98.6% complete. The early mortality was 7.4% (44 patients) and the late mortality was 6.2% per patient-year (41 patients). The valve-related causes of late mortality were thromboembolism (five), anticoagulant-related hemorrhage (one), and prosthetic valve endocarditis (one). The overall patient survival, including operative death, was 85% at 2 years. The linearized occurrence rate for valve-related complications was 5.6% per patient-year (37 events)--thromboembolism 2.7% per patient-year (18) anticoagulant-related hemorrhage 1.2% (eight), prosthetic valve endocarditis 0.8% (five), and periprosthetic leak 0.9% per patient-year (six). There were no cases of primary tissue failure or structural failure. At 2 years, the freedom from valve-related complications was 86.9%, from valve-related mortality, 98.7%, and from valve-related mortality and reoperation, 97.7%. This valve is fixed in glutaraldehyde at low pressure and is designed to improve durability. It has provided a low incidence of valve-related complications without structural failure. The structural design of the prosthesis does not always conform to the anatomy of bicuspid aortic valves.