Between February and November 1983, 108 recipients of cadaveric renal transplants entered a randomized multicenter trial and were treated either with cyclosporine (CsA) and prednisone (n = 55) or with conventional treatment based on azathioprine (Aza) and glucocorticoids (n = 53). The graft survival probability at 3 years was 76% for CsA patients and 48% for Aza patients (P less than 0.001). The cumulative number of acute rejections was significantly lower in the CsA group (32 vs. 104, P less than 0.001). Incidence of early posttransplant anuria was similar in both groups and did not affect renal function after three years. Nephrotoxicity in CsA patients, when present, was handled by reducing the dose of CsA, but in 12/55 patients a change to conventional therapy was thought to be necessary. However, in this group of 12, one patient lost the allograft because of irreversible rejection and one patient died 14 months later because of an infection. Mean creatinine clearance after three years was significantly lower in the CsA patients (54.7 +/- 2.6 ml/min) than in Aza patients, (67.2 +/- 4.9 ml/min, P less than 0.05). Considering only patients with grafts functioning after three years and still on the original randomized therapy, the mean creatinine clearance was similarly and significantly decreased from 1 to 3 years in both groups. There were no significant differences in occurrence of severe infections. Side effects such as hypertension, hypertrichosis, tremor and gum hyperplasia were more frequent in CsA patients.