This observational study was undertaken in maintenance renal transplant recipients to assess the relationship between cyclosporine C(2) and the long-term risk of chronic renal allograft dysfunction (CRAD). Pharmacokinetic profiling was undertaken twice yearly in 79 patients with stable graft function receiving cyclosporine microemulsion (Neoral) and steroids. Mean time since transplantation at study entry was 56 +/- 49 months posttransplant. At the end of the observational period (mean 43 +/- 14 months) 24 patients (30%) had developed CRAD, defined as proteinuria > 500 mg/24 hours associated with a rising serum creatinine and confirmed by graft biopsy. There were no significant differences at baseline between patients who did vs did not develop CRAD, except for reduced incidence of acute rejection in CRAD-free patients. The cyclosporine AUC was significantly higher among patients without CRAD (5707 ng. h/mL vs 3994 ng. h/mL, P =.0007). Mean C(2) was also significantly higher: 1001 ng/mL in the CRAD-free group versus 640 ng/mL in the CRAD group (P =.002). There was no significant difference in C(0). Regression analysis showed that the best predictors for the occurrence of CRAD were a low AUC (relative risk [RR] 1.54, P <.0001), a low C(2) (RR 1.30, P <.0001) and proteinuria (RR 4.95, P <.0001). Probability of freedom from CRAD was 90% for C(2) > 900 ng/mL. C(2) appears to be a superior strategy to C(0) monitoring of cyclosporine in stable renal transplant patients with regard to the risk of CRAD. C(2) values above 900 ng/mL are appropriate to minimize the risk of CRAD among patients receiving cyclosporine microemulsion and steroids.