The results of cadaveric donor pancreas transplantation at a single institution using the bladder drainage technique have been analyzed according to several factors that may impinge on outcome. Both multivariate and univariate statistical methods were used, with emphasis on solitary (pancreas after kidney and pancreas transplant alone) as opposed to simultaneous pancreas/kidney transplants. Of the 444 pancreas transplants performed at our institution from December 1966 through December 1991, we analyzed 249 bladder-drained cadaver donor pancreas transplants from November 1984 through August 1991. The factors that had a significant impact on outcome in the Cox multivariate analysis included retransplantation, age, preservation time, and degree of HLA mismatching. The results of solitary pancreas transplants improved with time as the factors that have an impact on graft survival rates were deliberately manipulated. During the 1988 to 1991 era, pancreas graft functional survival (insulin-independent) rates were not significantly different among the 3 recipient categories. Solitary pancreas transplant recipients less than 45 years old receiving primary grafts had a 1-year function rate of 61% in the pancreas transplant alone group (n = 32) and 74% in the pancreas after kidney group (n = 24). By placing emphasis on minimizing HLA mismatches, by giving adequate immunosuppression, and by detecting and treating rejection episodes early based on a decline in urine amylase, the results with solitary pancreas transplantation can be as good as those with simultaneous pancreas kidney transplantation. There are limitations to the interpretations that can be given to retrospective studies using inhomogeneous factors, as is the case in the analyses presented here. We cannot identify risk factors with certainty because the protocols changed over time, eg, immunosuppressive regimens, policies on HLA matching, and choice of duct-management techniques. Thus, in the analysis of all cases, not only was there a higher proportion by the bladder-drainage techniques in the later period, but the proportion of cases with good HLA matches was also higher: yet the earlier cases (more poorly matched, performed by the other techniques, and with worse results), were in the model. Retransplantation is also a problem for the analysis. The number performed was proportionately greater in the later period, and the possibility of retransplantation differed according to the patient's age. Over such a long time, we cannot evaluate our gain in experience statistically.(ABSTRACT TRUNCATED AT 400 WORDS)