BACKGROUND Three large-scale clinical trials conducted in North America, Europe, and Australia showed that mycophenolate mofetil (MMF) decreases the incidence of acute renal allograft rejection in the first 6 months after transplant compared with placebo or azathioprine. This study extends the randomized, prospective, double-blind trial of MMF conducted by the Tricontinental Mycophenolate Mofetil Renal Transplantation Study Group. METHODS Patients (n=503) were randomized to receive 100-150 mg of azathioprine (AZA) (n=166), 2 g of MMF (n=173), or 3 g of MMF (n=164) per day, in conjunction with cyclosporine and prednisone from the time of transplantation. RESULTS During the first 6 months, the incidence of biopsy-proven acute graft rejection (BPR) was reduced by approximately 50% in the MMF 2 g (19.7%) and MMF 3 g (15.9%) groups compared with the AZA group (35.5%). The incidence of treatment failure during the first 6 months, including BPR, death, graft loss, and early withdrawal without prior BPR, was significantly decreased: AZA, 50%, compared with MMF 2 g, 38.2% (P=0.0287), and MMF 3 g, 34.8% (P=0.0045). At 3 years after transplant, both intent-to-treat and on-study (censoring at 90 days after treatment) analyses of graft and patient survival showed a trend toward advantage for MMF 2 g and 3 g vs. AZA (intent-to-treat: 81.9% and 84.8% vs. 80.2%; on-study: 84.0% and 86.4% vs. 82.7%), although this trend did not reach statistical significance. Rejection was the principal cause of graft loss in all groups: AZA, 9.9%; MMF 2 g, 5.8%; and MMF 3 g, 3.0%. Graft function (intent-to-treat and on-study) was comparable in all three groups at 3 years. Gastrointestinal toxicity, leukopenia, and tissue-invasive cytomegalovirus disease were more common in the MMF 3 g group both during and after the first posttransplant year. Lymphoproliferative disorders were diagnosed in one AZA (0.6%), two MMF 2 g (1.2%), and three MMF 3 g (1.8%) patients. Other (non-lymphoproliferative disorders, noncutaneous) malignancies occurred in six AZA (3.7%), four MMF 2 g (2.3%), and nine MMF 3 g (5.5%) patients. Mortality was comparable in all three groups (AZA, 8.6%; MMF 2 g, 4.7%; MMF 3 g, 9.1%) by 3 years of follow-up. CONCLUSIONS MMF significantly reduced the incidence of rejection in the first 6 months, but there was not a significant improvement in graft survival throughout the 3 years after cadaver kidney transplantation.