Although 4000 pancreas transplants have now been done, alone or in combination with a kidney transplant, the risk/ benefit profile of the procedure has not been established by controlled studies. A solo pancreas transplant abolishes the need for daily insulin but requires chronic immunosuppression, has high failure rates, and is not proved to lessen the chronic complications of diabetes. Thus, it is probably justified only in those diabetic patients with incapacitating disease. For uraemic diabetic patients, combined pancreas and kidney transplantation often removes dependence on both insulin and dialysis, and has lower rejection rates than pancreas transplant alone. However, it needs more immunosuppression than kidney transplant alone, has no proven benefit on chronic complications of diabetes, and carries an increased risk of rejection, infection, and cancer. Living-related-donor kidney transplantation followed by cadaver pancreas transplantation is a possible alternative. Transplantation of pancreatic islets could offer the advantages of strict metabolic control without the drawbacks of immunosuppressive therapy. Thus, research efforts should concentrate on immune-protected islet transplantation. An alternative approach to avoiding long-term immunosuppression is the promotion of allograft tolerance.