To determine the association of various antigen-independent factors with long-term cadaveric kidney transplant failure.Cohort analytic study.Kidney transplant centers (N=131) in the United States.A total of 31 515 patients who received cadaveric kidney transplants between October 1987 and December 1991. Patients with unknown or uninterpretable vital status or graft survival time (n=264 [0.8%]) were excluded.Graft failure, estimated at 2 extremes, depending on whether the death of a patient with a functioning graft was censored ("censored graft failure") or not ("uncensored graft failure").During the 62-month study period, 5883 patients required the reinstitution of dialysis because of graft failure, 2404 patients died with graft failure, and 2041 patients died with a functioning graft. The relative risks of censored and uncensored graft failure were significantly associated with donor age, sex, and race and recipient body surface area, after adjusting for recipient age, sex, race, diabetes, cold ischemia time, panel cross-reactivity, pretransplant blood transfusions, previous renal transplantation, functional status, and HLA antigen mismatch.In cadaveric kidney transplantation, selected demographic and anthropometric factors are significantly related to long-term graft outcomes, even after adjusting for well-known antigen-dependent risk factors. These results support the hypothesis that the supply of viable donor nephrons and the physiologic demands of the transplant recipient are important determinants of long-term graft failure. Antigen-independent factors such as donor age should be incorporated into organ allocation algorithms to optimize equity and efficiency.