Chronic allograft nephropathy (CAN) represents the main cause of renal allograft loss after 1 yr of transplantation. Calcineurin inhibitor (CNI) use is associated with increased graft expression of profibrotic cytokines, whereas rapamycin inhibits fibroblast proliferation. The aim of this randomized, prospective, open-label, single-center study was to evaluate the histologic and clinical effect of rapamycin on biopsy-proven CAN. Eighty-four consecutive patients who had biopsy-proven CAN and received a transplant were randomized to receive either a 40% CNI reduction plus mycophenolate mofetil (group 1; 50 patients) or immediate CNI withdrawal and rapamycin introduction with a loading dose of 0.1 mg/kg per d and a maintaining dose aiming at through levels of 6 to 10 ng/ml (group 2; 34 patients). The follow-up period was 24 mo. At the end of follow-up, 25 patients (group 1, 10 patients; group 2, 15 patients) underwent a second biopsy. CAN lesions were graded according to Banff criteria. alpha-Smooth muscle actin (alpha-SMA) protein expression was evaluated in all biopsies as a marker of fibroblast activation. Graft function and Banff grading were superimposable at randomization. Graft survival was significantly better in group 2 (P = 0.0376, chi2 = 4.323). CAN grading worsened significantly in group 1, whereas it remained stable in group 2. After 24 mo, all group 1 biopsies showed an increase of alpha-SMA expression at the interstitial and vascular levels (P < 0.001); on the contrary, alpha-SMA expression was dramatically reduced in group 2 biopsies (P = 0.005). This study demonstrates that rapamycin introduction/CNI withdrawal improves graft survival and reduces interstitial and vascular alpha-SMA expression, slowing down the progression of allograft injury in patients with CAN.