Immunosuppression with cyclosporin after renal transplantation is associated with better graft survival than is azathioprine treatment. However, nephrotoxicity and other side-effects have led to regimens that change treatment to azathioprine shortly after transplantation. Conversion has beneficial effects in the short term on renal function and hypertension. We report long-term follow-up (minimum 5 years) of 128 patients who had received a first or second cadaveric kidney graft and were treated with cyclosporin and prednisone; they were randomly assigned 3 months after transplantation to groups continuing to receive cyclosporin (n = 68) or changing to azathioprine (n = 60). 8 years after transplantation, patient survival was 75.3% in the cyclosporin group and 85.9% in the azathioprine group (p = 0.14) and graft survival was 64.0% and 76.6%, respectively (p = 0.38). The frequency of cardiovascular death with a functioning graft was 8% higher in the cyclosporin group (95% CI -1 to 18). The relative risk of graft loss after conversion to azathioprine compared with graft loss after conversion to azathioprine compared with cyclosporin maintenance was 0.71 (0.37-1.38) and the relative risk of patient death was 0.57 (0.23-1.41). The cyclosporin group had poorer mean creatinine clearance (17.8 mL/min [8.1-27.5] lower than azathioprine group) and a higher proportion needed hypertensive drugs (20% [4-36] more). Gout was found in 9 cyclosporin-treated patients and 1 azathioprine-treated patient (difference 12% [3 to 20]). Elective conversion from cyclosporin to azathioprine 3 months after transplantation is safe and cost-effective.