The Veterans Administration Cooperative Study on Valvular Heart Disease was organized to compare survival and incidence of valve-related complications between patients receiving a bioprosthesis (the Hancock porcine heterograft) and a mechanical prosthesis (the Björk-Shiley spherical disc valve). Five hundred seventy-five patients undergoing single aortic or mitral valve replacement were randomized at surgery to one of the two valve types. At an average follow-up of 5 years (range 3 to 8) there are no statistically significant differences in survival between patients with the two valve types in the aortic valve replacement group. There is a statistically nonsignificant trend toward improved survival in patients undergoing mitral valve replacement with a bioprosthesis compared with a mechanical prosthesis (5 year survival probability was 0.70 +/- 0.05 and 0.58 +/- 0.06, respectively). Fatal and nonfatal valve-related complications occurred significantly less frequently in patients with a bioprosthesis compared with a mechanical prosthesis for both mitral and aortic valve replacement. Five year complication-free probability was 0.67 +/- 0.05 and 0.45 +/- 0.06, respectively, for patients with mitral valve replacement and 0.63 +/- 0.04 and 0.53 +/- 0.04, respectively, for those with aortic valve replacement. The difference in overall complication rates was largely due to the increased number of clinically significant but nonfatal bleeding episodes in patients receiving a mechanical prosthesis. Adjustment for differences in baseline characteristics between patients receiving a mitral mechanical prosthesis and a mitral bioprosthesis reduced the statistical significance of the difference in both mortality and complications.