Risk factors associated with death were identified in a cohort of patients who received 2 or more kidney transplants. Data on 19 variables were collected by chart review on 774 patients who received allografts between 1973 and 1980 at any one of 3 hospitals in Philadelphia. 124 of the patients received two or more transplants and were followed for a minimum of 1.5 years. Modified life table analyses of single variables indicated that 7 factors--splenectomy, donor source, age, transplant hospital, number of HLA mismatches, donor sex, and survival time of the prior graft--were significantly related to patient survival. Using all 19 variables, the proportional hazards model was fit to the data. The characteristics most related to survival were splenectomy (P less than .001), donor source (P = .0022), and age (P = .0015). The other 4 factors that were significant on univariate analysis were not significant in this multivariate analysis. The relative risk of death was 5.5 for patients who had had a splenectomy compared with those who had not had a splenectomy. Patients who had received more than one transplant were compared with patients who had received only one transplant, and a subset of recipients of primary transplants who returned to dialysis after primary graft failure. Survival of patients who had received one transplant was approximately the same as that of the retransplanted population. When the proportional hazards model was fit to the populations that received one transplant and compared with the model for the retransplanted group, only age and donor source were common to all three models. The effect of splenectomy on survival was significant for the total population of primary transplant recipients but had no effect on the survival of the subset of recipients whose kidney grafts had failed and were returned to hemodialysis. Infection accounted for 45% of the deaths among splenectomized, retransplanted patients. A higher percentage of septic deaths occurred in patients whose grafts were functioning at the time of death when compared with patients who had returned to dialysis after secondary graft failure. Although retransplantation alone is not associated with an increased mortality, retransplantation in splenectomized patients carries a high risk of death.