In order to evaluate selective differences of biological porcine valves versus pericardial valves and to analyze various valve models, 8 different bioprostheses (4 porcine valves, 4 pericardial valves) were studied in a 12-year follow-up. From 1978 to 1990, 476 porcine bioprostheses (Carpentier-Edwards: n = 104, Carpentier Edwards Supraanular: n = 59, Hancock I: n = 41, Hancock II: n = 272) and 647 pericardial valves (Hancock-Extracorporeal: n = 479, Ionescu-Shiley: n = 76, Carpentier-Edwards: n = 57, Mitroflow: n = 35) were implanted. At time of implantation, the patient age ranged from 21-85 years, mean 57.1 +/- 12.4 years. 831 patients were analyzed in the long-term follow-up (62.3 +/- 18.6 months, cumulative follow-up of 6632 patient-years). The incidences of thromboembolic complications (TE), endocarditis (E), primary tissue failure (PTF), rate of reoperation (ReOp), and late mortality due to prosthesis dysfunction were analyzed, calculated, and compared within the different valve models (actuarial data, chi 2 test, log rank analysis). The incidences of TE and E were lower for pericardial valves when compared with porcine bioprostheses (TE: 0.88 vs. 1.8%/patient year; E: 0.24 vs. 0.5%/patient year); within the 2 groups, the different valve models did not show any major differences. However, the incidence of PTF was significantly higher in the pericardial valve group, being 36 +/- 6.5%, 68 +/- 10% and 86 +/- 19.5% after 6, 8, and 10 years; the respective figures for the porcine valves were 6 +/- 3.5%, 18 +/- 7.1%, and 60 +/- 13.1% (p < 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)