The prognosis of valvular heart disease is favorably influenced by surgical treatment in spite of the possible incurrence of early or late postoperative complications. The hemodynamic properties of currently implanted bioprostheses and mechanical prostheses are essentially similar. It can be assumed that the fate of the patient, after having survived the early-postoperative phase, is predominantly dependent on the underlying disease and on complications of the prosthetic valve, itself, or the inherent necessity for chronic anticoagulation. To characterize the longterm course more closely, in this study in patients with prosthetic aortic or mitral valves, analysis of the survival rate as well as incidence of valvular dysfunction, late endocarditis, thromboembolism and bleeding associated with anticoagulant therapy was performed separately for patients with Björk-Shiley prostheses and for those with bioprostheses. The study encompasses 407 patients after isolated aortic valve replacement, 226 with Björk-Shiley prosthesis and 181 with bioprosthesis, as well as 269 patients after isolated mitral valve replacement, 142 with Björk-Shiley prosthesis and 127 with bioprosthesis (Table 2). All patients had valve replacement between 1975 and 1983 which was performed in the Department of Cardiovascular Surgery at the German Heart Center in Munich under comparable conditions, since 1977 with employment of cardioplegia. The preoperative characteristics of the groups compared did not differ significantly from each other (Table 1). Continuous follow-up was obtained in 95% of the patients. The mean observation period of all four compared patient groups ranged between 52.1 and 61.3 months. The term "major complication" was used to designate late-postoperative death, hemodynamic derangement due to prosthetic dysfunction, late endocarditis as well as major thromboembolism or bleeding. Prosthetic dysfunction in the case of Björk-Shiley valves was designated as mechanical or due to thrombosis, in the case of bioprostheses as valve degeneration or thrombosis. Late endocarditis was diagnosed only on documentation of - with a minimum interval of 30 days between valve implantation and onset of symptoms - changes in auscultatory findings and a morphological substrate on the implanted valve. Thromboembolism was assumed, after ruling out bleeding, only if a definitive or highly-probable transient or permanent neurologic deficit or a well-established suspicion of acute ischemia in an extremity or visceral organ was present. A bleeding complication was considered as such only if treatment by a physician was required.(ABSTRACT TRUNCATED AT 400 WORDS)