A series of 146 consecutive patients who underwent tricuspid valve replacement at the University Brugmann Hospital between 1967 and 1987 was reviewed. Mean age at operation was 51.4 years (+/- 12.1 years). Different types of prostheses were implanted including porcine and bovine pericardial bioprostheses and older and bileaflet mechanical valves. Most patients were severely disabled by their cardiac disease before operation, with 30.1% in New York Heart Association functional class III and 69.9% in class IV. Operative mortality and hospital mortality rates (30 days) were high (16.4%). Incremental risk factors for hospital death included icterus (p < 0.005), preoperative hepatomegaly (p = 0.012), and New York Heart Association functional class IV (p = 0.013). Multivariate analysis only selected preoperative icterus (p < 0.01) as being independently significantly related to hospital mortality. The hospital survivors were followed up for a median of 94 months. A complete follow-up was available for all patients except two for 30 months or more. At 30 months the only two significant parameters were the type of myocardial protection (p = 0.024) and the year of operation (before 1977 or after [precardioplegia era or after], p = 0.011). There were 70 late deaths during the entire follow-up period. The univariate (log-rank statistics) incremental risk factor for late death was the type of tricuspid prosthesis (Smeloff-Cutter and Kay-Shiley versus St. Jude Medical versus bioprosthesis) (p = 0.04). A trend was observed for the type of operative myocardial protection (normothermia and coronary perfusion) (p = 0.06) and preoperative New York Heart Association functional class IV (p = 0.055). Actuarial survival was 74% at 60 months and 23.4% at 180 months. Cumulative follow-up added up to 1015 patient-years. In a more detailed analysis of the effect on survival of the type of tricuspid prosthesis, a significant difference was observed between the bioprostheses and some older mechanical prostheses (Smeloff-Cutter and Kay-Shiley) (p = 0.04) but not between the bioprostheses and the bileaflet valves (p = 0.15). When the follow-up period was stratified according to less than 7 years and more than 7 years of follow-up, no difference was observed for the first period, but for the late follow-up the new mechanical prostheses did better than the bioprostheses (p = 0.05), suggesting a degradation of the bioprostheses after 7 years and favoring mechanical prostheses for those patients with a good long-term prognosis.