Isolated Pancreas Transplantation for Type 1 Diabetes: Title and subTitle BreakA Doctor's Dilemma

The primary goal for organ replacement therapy, achieved by allografting, should be to preserve life. Some transplantation procedures, such as those for the heart, liver, and lung, are necessary to preserve life, while others (eg, those for the cornea and pancreas) are not. Type 1 diabetes, a uniformly fatal disease prior to the introduction of insulin therapy approximately 80 years ago, represents an interesting case for transplantation. Insulin therapy reduced the occurrence of the metabolic catastrophe ketoacidosis and of death by inanition that characterized diabetes in the preinsulin era; however, until recently, insulin therapy also achieved grossly nonphysiologic metabolic control that was associated with long-term complications, including vision loss, amputations, and kidney failure.¹

The development of kidney failure shortened life-span substantially and kidney transplantation was the first transplant procedure to have a profound effect in patients with diabetes. Although randomized clinical trials comparing dialysis and transplantation have not been performed, observational studies have suggested that transplantation is the preferred mode of renal replacement therapy for most patients with renal failure, including those with type 1 diabetes.^{2 - 3} Kidney transplantation provides a better quality of life than does dialysis.² More importantly, studies comparing the outcomes of patients with renal failure who have received a kidney transplant with the outcomes of those who remain on the waiting list, receiving dialysis, have shown a survival benefit that begins approximately 8 months after the transplantation.³ The delay in benefit presumably reflects a relatively greater short-term mortality in the postoperative period, which ultimately gives way to decreased long-term mortality, compared with hemodialysis.

The current care environment in which pancreas transplantation is being performed is quite different than in 1966, when Kelly et al⁴ demonstrated the technical ability to perform simultaneous kidney and pancreas transplantation in patients with type 1 diabetes. One of the 2 recipients in the study by Kelly et al died from postoperative complications, but after an acute kidney-rejection crisis requiring high doses of steroids, the second recipient maintained near-normal glycemia for approximately 2 months without receiving exogenous insulin (F.C. Goetz, oral communication, October 27, 2003).

Today, whole-organ pancreas transplantation is highly successful. Enteric or bladder drainage of the exocrine secretions of the transplanted organ and greatly improved immunosuppression regimens have led to a high level of graft survival, defined as freedom from the need for exogenous insulin, with normal fasting blood glucose concentrations and normal or only minimally elevated levels of glycosylated hemoglobin. The vast majority of pancreas transplants have been performed in conjunction with kidney transplantation, with the idea that the risks associated with major surgery and immunosuppression are subsumed by the kidney transplant. There is a modest increase in surgical risk, as well as increased postoperative infections and complications, and a need for more intensive immunosuppression, attending simultaneous pancreas-kidney transplantation^{5 - 7} compared with kidney transplantation alone. These risks have been balanced by improved quality of life^{8 - 9} and by the potential for reduced complications, including a decrease and even reversal of neuropathy,¹⁰ as well as by decreased pathologic changes in the transplanted kidney.¹¹ The data to balance the risks and benefits of pancreatic transplantation are limited, owing to the absence of controlled trials. Most of the empirical data are retrospective, poorly controlled, and based on small case series at single centers.

In addition to the improvements in transplantation that have occurred, the routine treatment of type 1 diabetes also has advanced. Intensive treatment of diabetes is now widely available, including more physiologic delivery of insulin with external pumps or combinations of long- and short-acting insulin formulations and analogues, guided by self-monitoring of blood glucose levels. Although such therapies place a large burden on the patient and are associated with an increased frequency of life-disrupting and potentially dangerous hypoglycemia, they result in a profound decrease in long-term complications.¹² Recently published results from the follow-up of the Diabetes Control and Complications Trial (DCCT) cohort demonstrate 60% to 84% reductions in retinopathy and nephropathy that persist for as long as 15 years after the initiation of intensive therapy.^{13 - 14} Of note, the salutary effects of intensive therapy in the DCCT did not become manifest until approximately 3 years after its initiation; during the first 3 years, there was a transient worsening of retinopathy¹⁵ and an increased frequency of hypoglycemia, with no demonstrable benefits to counterbalance these adverse effects. There was, however, no added risk of mortality with intensive therapy; in fact, intensive therapy was projected to add an average of 5.1 years to the life span.¹⁶

The potent effect of intensive diabetes therapy aimed at achieving normoglycemia on long-term complications, albeit at the expense of increased demands on the patient and with attendant hypoglycemia, made whole-organ pancreas transplantation an increasingly attractive alternative during the past 10 years. The risk/benefit arguments that supported simultaneous pancreas-kidney transplantation have been applied to pancreas transplant alone and pancreas-after-kidney transplantation. These procedures have become increasingly popular, despite the recognition that graft survival is less than that achieved with simultaneous pancreas-kidney transplantation.⁶ In addition, earlier transplantation, prior to the development of advanced complications, was proposed as a more effective means of improving long-term outcomes. Finally, for the small proportion of the population with type 1 diabetes in which frequent and severe hypoglycemia and hyperglycemia severely affected quality of life, pancreas transplant alone was proposed as a means to ameliorate the unbearable consequences of brittle diabetes. The risks of hypoglycemia with insulin therapy would be replaced by the risk of surgery and immunosuppression. Whatever the rationale, the use of pancreas transplant alone and pancreas-after-kidney transplantation have increased 5-fold during the last 10 years, while simultaneous pancreas-kidney transplantation has been static.¹⁷ As with simultaneous pancreas-kidney transplantation, the absence of controlled trials limits a clear understanding of the relative risks and benefits.

In this issue of THE JOURNAL, Venstrom and colleagues¹⁷ examine the survival of patients who received simultaneous pancreas-kidney transplants and of patients with "preserved" kidney function who received pancreas transplants alone or pancreas-after-kidney transplants, compared with patients who remained on the waiting list for those procedures. This retrospective cohort study used a large database that represents the overall current US experience, included clinically important, quantitative end points, and was as carefully controlled as possible. The United Network for Organ Sharing policy that provides access to cadaver donor pancreata based only on waiting time and blood type enabled the investigators to compare outcomes between apparently similar groups of patients, even in the absence of randomization.

This study demonstrates 2 major points. First, the enthusiasm for simultaneous pancreas-kidney transplantation is probably justified, because patients receiving such transplants had an overall survival benefit (relative mortality risk, 0.43) after an initial 3-month period of postoperative risk, compared with patients remaining on the transplant list and receiving dialysis. How much of the benefit can be ascribed to the pancreas transplant vs the benefit of kidney transplantation (compared with ongoing dialysis) is impossible to tell from this analysis. Previous studies of diabetic patients (including those with type 1 and type 2 diabetes) receiving kidney transplantation alone, compared with otherwise similar patients on the waiting list and receiving continued dialysis, have shown a 73% reduction in the relative risk of death 18 months after transplantation.³ This salutary effect of kidney transplantation translates into a projected 11 additional years of life compared with dialysis. Of note, the benefit in the population with diabetes did not become apparent until 144 days after transplantation. Given the potent effect of kidney transplantation on survival, an independent benefit of the simultaneously added pancreas transplant is difficult to establish.

The second and more sobering finding is that for pancreas-after-kidney transplantation and pancreas transplant alone, the relative increase in mortality in the 3-month period after the transplant was not balanced by an improvement in survival over the next 4 years. For pancreas-after-kidney transplantation, the relative risk of death was approximately 42% higher than for the comparable patients with a previous kidney transplant who were waiting for a pancreas (P = .03). For pancreas transplant alone, the relative risk was 57% higher than for the patients awaiting a pancreas (P = .06).

Because the results of the study by Venstrom et al address safety, they must be taken seriously; however, the limitations of the study also must be considered. The lack of controlled clinical trial data, and reliance on retrospective analyses, heighten the potential for confounding. Were the groups that remained on the waiting list and those that received transplants truly comparable at baseline, as suggested by the authors, or were there unmeasured differences that might have explained the results? Furthermore, the 57% increased mortality in the group receiving pancreas transplant alone did not achieve statistical significance, perhaps owing to the relatively small numbers. Finally, any potential benefits of isolated pancreas transplantation may require more than the 4 years of the current study to become manifest. A delay in benefit of this duration must be compared with the benefits of intensive management of diabetes, simultaneous pancreas-kidney transplantation, and kidney transplantation alone, which occur during a shorter period of time.

This cautionary study has several important implications. First, the absence of controlled clinical trials represents a major barrier to understanding and balancing the risks and benefits of these procedures.^{6 ,18} Second, the increased early mortality in the postoperative period, without an apparent longer-term survival benefit, should temper enthusiasm for pancreas transplant alone and pancreas-after-kidney transplantation. There may be some patients in whom quality of life is so severely compromised that the potential benefits of pancreas transplant alone and pancreas-after-kidney transplantation outweigh the risks. However, such patients must be selected very carefully and be fully informed of the results of the current study. The criteria recommended by the American Diabetes Association¹⁹ need to be followed assiduously, as does the guideline proposed more than a decade ago by the Pancreas Transplant Evaluation Committee at the University of Michigan,²⁰ which states that "Patients should be selected to maximize potential benefit from pancreatic transplantation without risks that exceed those of standard treatment." Similar considerations will need to be applied to isolated islet transplantation when it becomes a clinical treatment option. Any loosening of the criteria will subject patients to risks that are not justifiable. Moreover, the severe limitation of donor pancreata means that pancreas transplant alone and pancreas-after-kidney transplantation decreases the available organs for simultaneous pancreas-kidney procedures, which are more justifiable. Finally, at least part of the explanation for the differences in outcome between the recipients of pancreas transplant alone and pancreas-after-kidney transplant and the corresponding patients on the waiting list is that improvements in routine care of diabetes have improved the long-term outlook for all diabetic patients.

George Bernard Shaw wrote in his "Preface on Doctors" in The Doctor's Dilemma that "There is a fashion in operations as there is in sleeves and skirts: the triumph of some surgeon who has at last found out how to make a once desperate operation fairly safe is usually followed by a rage for that operation not only among the doctors, but actually among their patients."²¹ The dilemma with isolated pancreas transplantation will be to resist new and attractive procedures that carry increased risk when safer, conventional therapy is available.