We reviewed clinical data on 29 patients (17 acquired disease and 12 congenital disease) who underwent tricuspid valve replacement using a Carpentier-Edwards porcine valve or a St. Jude Medical valve at Niigata University Hospital between January 1978 and October 1995, and analyzed risks related to the outcome. Age ranged from 8 to 62 years, and 17 females were included. New York Heart Association functional class was IV in 9 patients (31%). Previous valve surgery had been performed in 10 patients (34%). An urgent surgery was required in 4 patients and simultaneous mitral (with aortic) valve replacement in 13 (45%). Mean follow-up period was 6.7 years. Hospital mortality occurred in 5 patients (17.2%) and all of them had acquired valvular diseases with functional class IV. There were 6 late deaths and none of them related to the tricuspid prosthesis. Nonfatal morbidity occurred in 2 patients (a thrombosed valve and a chronic hemolytic anemia) with a St. Jude valve. Late functional class was I or II in all patients except 3 with severe left-ventricular dysfunction or bioprosthetic failure. Late postoperative echocardiographic examination showed moderate tricuspid stenosis or regurgitation in 3 patients with a Carpentier-Edwards valve. By the Kaplan-Meier method, functional class (P < 0.01) and baseline diseases (p < 0.05) were found to affect probability of freedom from cardiac- and/or valve-related mortality including hospital death. In conclusion, the present study indicates that similarly good outcome would be expected whether using a St. Jude Medical valve or a Carpentier-Edwards porcine valve, and that both functional class IV and combined acquired valvular disease are risk factors for mortality after tricuspid valve replacement. Patient-specific selection of the prosthesis would be preferable in tricuspid valve replacement.