Centers may restrict the use of some donor kidneys on the belief that overall graft survival is improved by giving older kidneys to older recipients and vice versa. The prevalence and the effect on graft survival (determined by death, return to dialysis, or retransplantation) of this practice among 74,297 first cadaver kidney transplantations in 1988 to 1998 was examined by using data from the United States Renal Data System. Giving older kidneys to older recipients is common; recipients > or =55 yr old received donor kidneys that were > or =55 yr old 46.2% more often than expected, but they received kidneys that were 18 to 29 yr old 33.6% less often than expected (chi(2) P < 0.0001). Both recipient and donor age have important effects on graft survival, although the effects of donor age are much stronger than those of recipient age. Compared with recipients 18 to 29 yr old, recipients > or =55 yr old were 25% (95% confidence interval, 15 to 35%, P < 0.0001) more likely to have graft failure (adjusted for donor age and other risk factors). On the other hand, donor kidneys > or =55 yr old were 78% (95% confidence interval, 58 to 99%, P < 0.0001) more likely to fail compared with kidneys 18 to 29 yr old. However, giving older kidneys to older recipients had little independent effect on graft survival, once the intrinsic effects of recipient and donor age were taken into account. For example, transplanting donor kidneys > or =55 yr old into recipients > or =55 yr old reduced the risk of graft failure only -6% (95% confidence interval, -18 to 8%, P = 0.3923) after the independent effects of donor and recipient age per se were taken into account. Thus, giving older kidneys to older recipients is a common practice that does not improve overall graft survival.