Between January 1985 and March 1990, isolated Omnicarbon valve replacement operations were performed on 90 patients aged 25-72 years. There were 53 aortic valve replacements (AVR) and 37 mitral valve replacements (MVR). The cumulative follow-up was 320 patient-years, with a mean(s.d.) follow-up of 3.7(1.4) years. There were three operative and hospital deaths (3.3%) resulting from retrograde aortic dissection during cardiopulmonary bypass, postoperative renal failure and postoperative mediastinitis. Seven patients died during the late postoperative period, four from valve-related causes. Two of these patients died from prosthetic valve endocarditis, and the others died from thromboembolism and valve thrombosis. The mean(s.d.) actuarial survival rate at 6 years was 86.2(4.3)% (98.8(0.8)% for AVR, 82.1(4.8)% for MVR). The mean(s.d.) actuarial survival rate of freedom from all valve-related mortality at 6 years was 93.5(2.6)% (100% for AVR, 88.1(2.9)% for MVR). There were two thromboembolic events (one mesenteric artery thrombosis and one valve thrombosis). The standardized incidence of thromboembolism was 0.63% per patient-year. Prosthetic valve endocarditis occurred in three patients (0.94% per patient-year). One patient (0.31% per patient-year) was found to have a paravalvular leak resulting from aortitis syndrome. The mean(s.d.) actuarial rate of freedom from all valve-related complications at 6 years was 89.2(2.0)% (98.6(1.0)% for AVR, 86.4(2.2)% for MVR). There were no instances of anticoagulant-related haemorrhage, valve-related haemolysis, or structural failure. Results of a follow-up period of 6 years indicated that good clinical results and a low incidence of valve-related complications can be demonstrated with the Omnicarbon valve.