The incidence of posttransplant diabetes mellitus (PTDM) was compared in three groups of renal transplant recipients: Nondiabetic patients who had been randomized between 1980 and 1983 to receive anti-lymphoblast globulin (ALG), azathioprine (AZA), and prednisone (P) (group 1) or cyclosporine (CsA) plus prednisone (group 2). Group 3 consisted of a more recent (1984-85) cohort who were given a combination of azathioprine, cyclosporine, and prednisone (±ALG). PTDM developed in 20 of 173 previously nondiabetic 18-55-year-old patients. Three of 47 patients (6.4%) in group 1, 4 of 58 patients (6.9%) in group 2, and 13 of 68 patients in group 3 (19.1%) developed PTDM. Thus in the two groups composing the concurrent prospective randomized trial (groups 1 and 2) the incidence of PTDM did not differ. The subsequent patients who were given a combination of ALG, azathioprine, cyclosporine, and prednisone developed a significantly greater incidence of PTDM even though the total dose of cyclosporine and prednisone were lower than in groups 1 and 2. PTDM usually occurred within two months of transplantation, and 11 of 17 patients who initially required insulin are still dependent upon exogenous insulin. The incidence of PTDM was not significantly affected by sex of the recipient, HLA-type, primary renal disease, rejection episodes, primary vs. secondary transplant, or prior splenectomy. The incidence of PTDM is greater in patients older than 45 (34.2% vs. 5.2%), and heavier than 70 kg (21.1% vs. 5.1%); in recipients of cadaveric allografts (15.7% vs. 4.6%); and in patients who were hospitalized for infections (22.4% vs. 4.7%). CsA levels tended to be higher in the group 2 and 3 patients who developed PTDM than in those who remained nondiabetic. One-year actuarial patient survival in those with PTDM was 83.3% vs. 98% (P<.01) in the nondiabetic and graft survival was 77.1% vs. 87.1% (NS). The combination of Minnesota ALG, azathioprine, cyclosporine, and prednisone appears to predispose older, heavier recipients of cadaver allografts to the development of PTDM. The risk of PTDM must be weighed against the more usual results of improved patient and graft survival using this combination of immunosuppression.