We examined the long-term clinical function and fate of allograft aortic valves, preserved by two differing methods, for aortic valve replcement: (1) fresh allografts stored at 4°C refrigeration (group I) and (2) viable allografts cryopreserved in liquid nitrogen at -196°C (group II). A total of 316 aortic valve replacement operations were performed, 124 in group I (December 1969 to May 1975) and 192 in group II (June 1975 to December 1986). Concomitant surgical procedures (eg, coronary artery bypass grafting and mitral valve replacement) were necessary in 120 operations. The overall immediate 30 day mortality was 6.0% (confident limits 4.6% to 7.7%), and the survival rates were 83% at 4 years and 60% at 15 years. Endocarditis and technical factors concerning reoperation showed no difference between the two groups according to parametric estimates and hazard function analyses. However, there was a marked difference in reoperation for valve degeneration: 23 patients in group I and O patients in group II. The freedom from reoperation for valve degeneration at 10 years was 89% for group I and 100% for group II, and at 15 years it was 59% for group I. The hazard function for group I reflected the late rising risk of degeneration. The freedom from thromboembolism of both groups (aortic valve replacement with or without bypass grafting) was 97% at 10 years and 96% at 15 years. Group II explanted valves (operation for technical malalignment reasons) has consistently shown evidence of persisting viability on tissue culture, on metabolic studies, and on histologic appearances. Chromosomal studies have shown the donor origin of these cells. The superior results with the cryopreserved valve is considered to be due to persisting viability, wich appears to be the key to durability.