To define the role of mammalian target of rapamycin inhibitors in kidney transplantation, we compared efficacy and safety of two immunosuppressive regimens-a calcineurin inhibitor-free regimen with depletive induction versus a calcineurin inhibitor-based regimen. De novo renal allograft recipients were randomized before transplantation to receive sirolimus (SRL; n=71, group A) or tacrolimus (n=70, group B). All patients received mycophenolate mofetil and corticosteroids. In group A, patients received rabbit antithymocyte globulin induction. In group B, antithymocyte globulin therapy could be given in case of delayed graft function. The estimated glomerular filtration rate (GFR) (Nankivell's formula) at month 12 was the primary endpoint. GFR showed no significant difference at month 12, with 56.1 in group A versus 58.4 mL/min/1.73 m in group B. In functioning grafts, renal function was significantly better in the SRL group, with higher GFR values at months 1, 2, 3, 6, and 9 (P<0.05). At month 12, patient survival and incidence of biopsy-proven rejection were not different between groups (95.8% vs. 97.1%, and 16.9% vs. 12.9%, respectively). However, proportion of graft loss was higher with SRL at months 6 and 12 (11.3% vs. 0.0%, P=0.004; 14.1% vs. 4.3%, P=0.044, respectively). Adverse events and premature withdrawals were more frequent with SRL (P<0.001 and P<0.05, respectively), whereas cytomegalovirus infections were more frequent with tacrolimus (P<0.001). Patients treated with induction plus SRL, mycophenolate mofetil, and corticosteroids may obtain good renal function but have a higher risk of adverse events, drug withdrawal, and graft loss.