The incidence of posttransplant diabetes mellitus (PTDM) was compared in two groups of renal allograft recipients. These were all nondiabetic patients who had been transplanted between 1979 and 1987 and received either azathioprine-methylprednisolone (group 1) or cyclosporine-methylprednisolone (group 2) therapy as maintenance immunosuppression. The incidence of PTDM in group 1 was 9.1% vs. 18.6% in group 2 (P less than .05). The mean daily dose of methylprednisolone during the initial 2 months posttransplant was not greater among the PTDM patients of groups 1 or 2. Cyclosporine levels and mean daily CsA doses during the initial 2 posttransplant months were also not different among the CsA-PTDM and euglycemic CsA patients. Posttransplant diabetes mellitus occurred rapidly (less than 2 months) and required insulin therapy in the majority of cases. Increased age (greater than 40 years) was associated with a higher risk for PTDM, however, the greater incidence accompanying increased body weight only approached significance. Patient gender and donor source were not associated with significant risk for PTDM. The development of PTDM was accompanied by a significant decrease in graft survival at 3 years in the entire PTDM population and at 4 years in the CsA-PTDM subgroup. Actuarial patient survival was not adversely affected. The current study suggests that CsA may be diabetogenic when administered with methylprednisolone to renal allograft recipients. The adverse effect on allograft survival requires further investigation. These results may also have important implications for pancreatic and islet cell transplantation.