To examine the evolution of renal allograft function in kidney transplant recipients receiving long-term cyclosporine therapy, we evaluated 50 cadaveric and 30 living-related renal transplant recipients having graft survival greater than or equal to 12 months and an opportunity for 5 years of follow-up. Linear analysis of long-term allograft function in each patient was undertaken by plotting reciprocal serum creatinine (1/Crs) values vs. time. Mean follow-up was 49 +/- 18 months. Actual 3-year and 5-year allograft survivals were 83.8% (n = 80) [corrected] and 73.3% (n = 75) [corrected], respectively. Collective analyses of values of 1/Crs measured at yearly intervals and of the slopes of the curves obtained by plotting 1/Crs vs. time for each patient suggested that long-term use of CsA is associated with impaired but generally stable allograft function 1-5 years posttransplant. The aggregate rate of decline of renal allograft function in the study population did not differ from that of a historical control group consisting of 59 renal transplant recipients treated with a conventional prednisone-azathioprine immunosuppressive regimen. Donor source, diabetes, and diastolic hypertension (diastolic BP greater than 95 mmHg in more than half the follow-up readings) were not correlated with a more rapid rate of decline of allograft function as reflected in the slopes of the 1/Crs vs. time curves between 12 months posttransplant and the end of follow-up. In contrast, a significantly greater rate of decline of cadaveric allograft function was observed in patients with 12-month Crs values greater than 2.5 mg% and recipients of greater than 2 HLA-A,B-mismatched cadaveric kidneys. The data do not support an indication for routine conversion from CsA to azathioprine following successful renal transplantation.