Sirolimus, lacking known nephrotoxicity, appeared to be an ideal immunosuppressive agent in the setting of delayed graft function (DGF) after renal transplantation. Coincident with our use of sirolimus however, we noticed prolongation of DGF. To investigate possible causes of prolonged DGF, extensive donor, recipient, transplant, and post-transplant data were collected on 132 consecutive cases of DGF at the University of California, San Francisco between 1/1/97 and 6/30/01. Cox proportional hazards analysis of time to graft function was used in univariate and multivariate models to identify factors that prolong DGF. Sirolimus had a large and highly significant effect on time to graft function (hazard ratio 0.48, p = 0.0007). The hazard ratio indicates that a recipient on sirolimus is half as likely to resolve DGF or twice as likely to remain on dialysis as a recipient without sirolimus. Two other factors had less potent but still significant association with DGF duration: recipient sensitization (hazard ratio 0.66, p = 0.037), and Novartis score (hazard ratio 0.93 per 1.0 increase; p = 0.034). Sirolimus retained its profound negative association with time to graft function in all multivariate models. Because sirolimus appears to prolong DGF, it may not be the optimal immunosuppressive choice in the DGF setting.