The aim of this randomized, open-labeled trial was to compare the incidence of acute rejection after an early (3 mo posttransplantation) withdrawal of cyclosporine (CsA) or mycophenolate mofetil (MMF) in renal transplantation. Among 218 eligible recipients, 108 nonsensitized, rejection-free patients who were under a triple drug regimen (CsA-MMF-prednisone) and had received a first kidney from a deceased donor were enrolled. At 3 mo after graft, they were gradually withdrawn from CsA (MMF group, n = 54) or MMF (CsA group, n = 54). A graft biopsy and a pharmacokinetic study of CsA and mycophenolic acid were systematically performed before the randomization. At 1 yr, graft and patient survival rates were 100% in each group. Renal function was improved in the MMF group compared with the CsA group (Cockcroft calculated clearance 64.7 +/- 18.7 versus 56.5 +/- 18.0 ml/min; P = 0.023). However, the probability of acute rejection was higher in the MMF group (18.5 versus 5.6%; P = 0.045). The 10 patients who developed acute rejection after CsA withdrawal had a significantly higher incidence of borderline changes on the randomization biopsy than the 44 rejection-free patients (five of 10 versus eight of 44; P = 0.034), and they displayed a lower area under the curve of mycophenolic acid (43 +/- 9 versus 58 +/- 22 mg/h per L; P = 0.045). Multivariate analysis confirmed that borderline changes and area under the curve of mycophenolic acid were significant risk factors of acute rejection after CsA discontinuation. It is concluded that a systematic graft biopsy and a pharmacokinetic study of mycophenolic acid are needed to reduce the risk for acute rejection after CsA withdrawal.