A consecutive series of 188 Mitroflow pericardial bioprostheses were inserted in 166 patients between 1st January 1983 and 31st December 1985. Twenty-two valves had to be removed from 16 patients after a mean follow up period of 78 months (range 58-92 months) for aortic, 73 months (65-79 months) for mitral and 78 months (48-103 months) for double valve replacements. All but one reoperations for primary tissue failure were carried out as elective surgical procedures. The most important cause of failure was collagen degeneration, seen in all explanted valves. The areas of degeneration were the major sites of origin of calcification, which was seen in 11 valves (50%). Ten valves (45%) showed features suggestive of lipid infiltration, extensive fatty acid deposition being identified in one and a typical atheromatous reaction in another. In contrast to the Ionescu-Shiley valve, the mode of failure was tear originating at the top of the commissure, associated with major structural changes in the tissue. Excessive pannus ingrowth was observed in 11 valves (50%). The universal presence of tissue degeneration in the glutaraldehyde treated pericardial leaflets of the Mitroflow bioprostheses explanted and examined in this study questions the adequacy of the methods employed in the harvesting and/or processing and/or preservation of this valve. We have, therefore, discontinued using the Mitroflow bioprosthesis. However, the slow deterioration of the Mitroflow bioprosthesis permits elective reoperation; preventive removal of functioning valves is therefore not indicated.