Target of rapamycin inhibitors (TOR-I) (sirolimus, everolimus) are immunosuppressive agents with a novel mode of action but an uncertain clinical role. To investigate the benefits and harms of immunosuppressive regimens containing TOR-I when compared to other regimens as initial therapy for kidney transplant recipients. We searched the Cochrane Central Register of Controlled Trials (CENTRAL) (in The Cochrane Library, issue 2, 2005), MEDLINE (1966-June 2005), EMBASE (1980-June 2005), the specialised register of the Cochrane Renal Group (June 2005)., and contacted authors and pharmaceutical companies to identify relevant studies. All randomised controlled trials (RCTs) and quasi-RCTs where drug regimens containing TOR-I were compared to alternative drug regimens in the immediate post-transplant period were included, without age restriction, dosage or language of report. Two reviewers independently assessed trials for eligibility and quality, and extracted data. Results are expressed as relative risk (RR) or weight mean difference (MD) with 95% confidence intervals (CI). Thirty three trials (142 reports) were included (sirolimus (27), everolimus (5), head-to-head (1)). When TOR-I replaced CNI there was no difference in acute rejection, but serum creatinine was lower (MD -18.31 micromol/L, -30.96 to -5.67), and bone marrow more suppressed (leucopenia: RR 2.02 1.12 to 3.66; thrombocytopenia: RR 6.97 2.97 to 16.36; anaemia: RR 1.67, 1.27 to 2.20). When TOR-I replaced antimetabolites, acute rejection (RR 0.84, 0.71 to 0.99) and cytomegalovirus infection (CMV) (RR 0.49; 0.37 to 0.65) were reduced, but hypercholesterolaemia was increased (RR 1.65, 1.32 to 2.06). For low versus high-dose TOR-I, with equal CNI dose, rejection was increased (RR 1.23, 1.06 to 1.43) but calculated GFR higher (MD 4.27 mL/min, 1.12 to 7.41), and for low-dose TOR-I/standard-dose CNI versus higher-dose TOR-I/reduced CNI, acute rejection (RR 0.67, 0.52 to 0.88) and calculated GFR (MD -9.46 mL/min, -12.16 to -6.76) were reduced. There was no significant difference in mortality, graft loss or malignancy risk for TOR-I in any comparison. TOR-I have been evaluated in four different primary immunosuppressive algorithms; as replacement for CNI and for antimetabolites, in combination with CNI at low and high dose and with variable dose of CNI. Generally, surrogate endpoints for graft survival favour TOR-I (lower risk of acute rejection and higher GFR) and surrogate endpoints for patient outcomes are worsened by TOR-I (bone marrow suppression, lipid disturbance). Long-term hard-endpoint data from methodologically robust RCTs are still needed.