Contraindications:

TECHNOLOGY FACT SHEET

Indications for use:

Preuremia or nonuremia with propensity toward diabetic complications which are predicted to be worse than the undesirable side effects of immunosuppression

or

Extremely labile diabetes, refractory to any type of exogenous insulin regimen or other therapeutic approach, with severe metabolic complications and incapacitating clinical or emotional problems

Applicable populations served:

Type I diabetics with the following characteristics:

• Age of 18 to 50 years

• Lack of systemic complications

• ABO group compatibility

• Presence of overt secondary diabetic complications (nephropathy, retinopathy, or neuropathy)

• Ability to tolerate surgical procedure and post-transplant immunosuppressive routine

• Psychiatric stability and strong family support network

• Ability to understand investigative nature and risks of procedure

Procedure:

The recipient's pancreas is not removed unless diseased. The donor pancreas, either whole or segmental, is placed within the peritoneum or the abdomen. Blood vessels from the donor pancreas are connected with the iliac vessels. For managing exocrine drainage, either neoprene or prolamine is injected into the exocrine duct (duct occlusion or polymer injection method) or the duct is connected with the urinary bladder (bladder drainage), the lower part of the small intestine (enteric drainage), or the ureter (ureter drainage). The bladder-drainage technique is the most commonly used method in the United States.

Number of procedures performed:

In Minnesota from 1987 to 1995: 108

In U.S.A. from 1987 to 1993: 181

Outside of U.S.A. from 1987 to 1993: 49

Complications/adverse effects:

• Thrombosis 10.90%
• Bleeding 0.70%
• Infection 2.00%
• Pancreatitis 3.40%
• Total 17.00%

One-year survival:

Contraindications:

TECHNICAL REPORT

Cadaver Donor Pancreas Transplantation for Poorly Conrolled Type I Diabetes

Diabetes mellitus is a chronic disorder characterized by impaired metabolism of carbohydrates, protein, and fats. There are two major forms of the syndrome, insulin-dependent (Type I) and non-insulin-dependent (Type II) diabetes mellitus. Although onset can occur at any time for Type I diabetes, the emphasis of this report, the peak age of onset is typically 12 years. Type I diabetes is caused by the lack of insulin production by the beta cells of the islets of Langerhans in the pancreas (Sutherland et al., 1993a) and is associated with viral infections, autoimmune responses, and genetic factors. This lack of insulin results in protein wasting, production of ketone bodies, and hyperglycemia (excessive sugar in the blood), which leads to excessive urine excretion, excessive sugar excretion in the urine, dehydration, and acidosis accompanied by ketone bodies (ketoacidosis). Ketoacidosis, if untreated, can lead to nausea and vomiting, stupor, coma, and death. Symptoms of diabetes include excessive eating, excessive thirst, itching or prickling sensations, lassitude, weight loss, blurred vision, and irritability. The disorder is also associated with a high incidence of vascular disease involving the small and large blood vessels (micro- and macroangiopathy) from which stem a number of late diabetic complications (Abecassis and Corry, 1993).

Over 50% of the diabetic patients who survive for more than 20 years after diagnosis will experience late complications that affect the eyes (retinopathy), nerves (neuropathy), and/or kidneys (nephropathy) (Sutherland et al., 1993a). Diabetic retinopathy, which usually occurs between 10 to 20 years following the onset of Type I diabetes, can lead to total blindness within two years if left untreated (Königsrainer et al., 1991a). In fact, diabetes is the leading cause of blindness in the United States (Zehrer and Gross, 1991). Moreover, compared with the general population, individuals with diabetes are two times as likely to experience coronary heart disease or stroke, five times as likely to develop gangrene, and 17 times as likely to experience kidney failure (Zerher and Gross, 1991), and the average life-span for diabetics is significantly shorter than that for the general population (Sutherland et al., 1993a). In addition, diabetes is the third leading cause of death by disease in the United States (Zehrer and Gross, 1991).

Control of carbohydrate metabolism and blood pressure is necessary to prevent the secondary complications of diabetes (Königsrainer et al., 1991a). It has been proven that the development of neuropathy, retinopathy, and nephropathy depends upon the degree to which glycemia is controlled (Sutherland et al., 1993a), and the administration of exogenous insulin has been the cornerstone of diabetic treatment since its discovery about 70 years ago (Bartucci et al., 1992). Treatment options include a blend of intensive insulin therapy, diet, exercise, and management of complications. Recent developments, such as different types of insulin and insulin-delivery systems, as well as home blood glucose monitoring systems, have contributed to an improved clinical state for diabetics, who now can monitor their own blood glucose, meal plan, and level of activity and can make decisions regarding insulin doses during each day to achieve normal, or at least near normal, blood sugar levels (Bartucci et al., 1992). However, it has been found that exogenous insulin does not perfectly control diabetes, as dose-calculations are not always precise and it is not uncommon for diabetics to experience hypoglycemia (insufficient blood-sugar) from overdosing or ketoacidosis from underdosing. Only perfect control of blood sugar levels will prevent complications, but a perfect insulin delivery system has yet to be developed (Sutherland et al., 1993a). The only means of achieving a near physiologic metabolic control without the occurrence of hypoglycemic incidents is pancreas transplantation, which provides a self-regulated source of insulin (Bandello et al., 1991).

Organ transplantation became a clinical reality after the first human kidney transplantation in the 1950s followed by the development of the drug azathioprine and its combined use with the steroid prednisone in the 1960s. Although kidney transplantation evolved quickly as an effective treatment for end-stage renal disease, patients with diabetes were generally considered as high-risk, therefore ineligible for kidney transplantation. Because uremic diabetics undergoing dialysis had a high mortality rate, surgeons at the University of Minnesota, considering transplantation the lower risk option when compared with dialysis, initiated not only kidney transplantation but also pancreas and islet transplantation for the treatment of uremia and hyperglycemia in these patients. Thus, the first human pancreas transplant, using a segment of the pancreas along with simultaneous kidney transplantation, was performed in 1966 at the University of Minnesota center on a diabetic patient. In this case, as well as in 13 additional pancreas or combined pancreas/kidney transplants performed over the next few years, grafts functioned immediately and patients became insulin-independent; however, all grafts except one were rejected within one year (Sutherland et al., 1993a). Due to the high incidence of complications related to pancreas transplantation, it was thought by some to be unnecessary to correct diabetes while attempting to correct uremia; thus, in 1968, the first kidney transplant in a diabetic without a simultaneous grafting of the pancreas was performed at the University of Minnesota center, followed by 45 additional kidney transplants in diabetics over the next several years. The procedure proved to be effective, with graft survival rates similar to those of nondiabetic patients. This practice has since been adopted worldwide and is performed routinely for treatment of diabetic nephropathy (Sutherland et al., 1993a). Meanwhile, with improvements in surgical techniques and immunosuppressive regimens, work continued on pancreas transplantation. By mid-1993, 4,799 pancreas transplants were reported to the International Pancreas Transplant Registry, with over 3000 of these performed in the United States (Sutherland et al., 1994). Results have been encouraging, with graft survival rates for combined pancreas and kidney transplantation comparable to those of solitary kidney transplantation (American Diabetes Association, 1992b). Thus, pancreas transplantation is now considered a viable option for the treatment of Type 1 diabetes (Milde et al., 1992).

DESCRIPTION OF PANCREAS TRANSPLANTATION

Objectives

Pancreas transplantation may now be performed alone (PTA) or in combination with a kidney graft, either after successful kidney transplantation (PAK) or simultaneously (SPK). The main objective of kidney transplantation in diabetics is to ameliorate end-stage kidney disease (Sutherland et al., 1993a). The primary purpose of pancreas transplantation is to maintain glucose levels in the near-normal range by providing continuous regulation of glucose metabolism with an adequate insulin-producing pancreas. Additional objectives are to delay, prevent, or reverse the progression of diabetic complications, such as nephropathy, retinopathy, neuropathy, and cardiovascular disease (Boudreaux et al., 1991; Zehrer and Gross, 1991). However, while it is possible that complete correction of the diabetic state may prevent the development of these secondary complications, complications involving the eyes and nerves are usually far advanced in uremic diabetics and the effects of pancreas transplantation on the course of established diabetic complications have been variable; therefore, it is the overall effect on quality of life that is most important. Although all transplantation procedures require life-long immunosuppression, which is associated with adverse effects and may influence quality of life, it is a generally accepted concept that compared with patients who are not immunosuppressed but require dialysis or exogenous insulin, those who are immunosuppressed but dialysis-free or insulin-independent experience an improved quality of life (Sutherland, 1994). Thus, quality of life is considered an important objective for both kidney and pancreas transplantation.

Organ Procurement

Organs may be obtained from a cadaver or from a living related donor in the case of kidney grafts or segmental pancreas grafts. There are guidelines for the selection of donor organs; however, since there is a scarcity of donor organs, these guidelines are not considered strict requirements and often a balance must be reached between the desire to use only ideal donors and the desire to minimize mortality as prospective recipients wait for organs. Considerations for organ donation eligibility include (Schwartz et al., 1989):

• ABO blood group compatibility

• Lymphocyte cross matching

• Travel time between the donor and recipient hospitals

• Approximate size matching

• Etiology of brain death

• Absence of hepatitis B, HTLV-III, and cytomegalovirus

In the case of pancreas transplantation, a history of diabetes mellitus is an absolute contraindication for organ donation. Organs may be procured from donors up to the age of 60 years, but a thorough evaluation of possible atherosclerosis should be performed in any potential donor over the age of 40 years (Abecassis and Corry, 1993). Potential cadaver donors must be brain-dead with a beating heart, and organs are harvested in conjunction with multiple organ procurements (Taylor et al., 1994). Although matching of human leukocyte antigen (HLA) is not considered an absolute requirement for organ donation, it is associated with improved graft survival rates. In the past, HLA-matching was often not performed due to the time required for testing; however, improved methods of preserving harvested organs now allow storage of the harvested pancreas for up to 30 hours with no detrimental effects, which allows time for HLA-testing. HLA-matching can also be achieved with a living related donor, but these potential donors run a significant risk of developing diabetes themselves since they are relinquishing up to half of their own islet cells for segmental grafts; therefore, living related donors are not used routinely for pancreas transplantation (Robertson and Sutherland, 1992).

The surgical procedure for pancreas transplantation is not standardized. In addition to some differences in surgical technique for the various recipient categories of transplantation (PTA, PAK, and SPK), the pancreas graft itself may involve use of the whole organ or only parts of the organ (segmental graft), and the graft may be placed within the peritoneum or abdomen. The managment of exocrine secretions also varies and may be handled by injecting neoprene or prolamine into the exocrine duct (duct occlusion or polymer injection method) or by draining exocrine secretions into the bladder (bladder drainage method), the intestines (enteric drainage method), or the ureter (ureter drainage method). The bladder-drainage technique is the most widely used method in the United States (Sutherland et al., 1994). The immunosuppressive approach also varies but tends to follow the protocol for solitary kidney transplantation (Abecassis and Corry, 1993). Typically, this involves a triple drug regimen using azathioprine, cyclosporin, and steroids or a quadruple drug regimen that uses these same agents but adds either antilymphocyte globulin (ALG), antithymocyte globulin (ATG), or monoclonal antibodies (OKT3). The most common approach is to administer an anti-T-cell agent, such as OKT3, during the first two weeks after grafting for induction immunosuppression followed by the triple drug regimen for maintenance immunosuppression (Robertson and Sutherland, 1992). Doses differ and are adjusted individually according to parameters such as white blood count, platelet count, and serum levels of the immunosuppressive agent (Taylor et al., 1994). Further, anticoagulation therapy, which is not always used and varies when it is used, may involve the administration of aspirin, heparin, dipyridamole, antithrombin III, and/or dextran. Each of the variables in the overall transplantation procedure can influence patient and graft survival rates and health outcomes.

Rejection-Monitoring and Treatment

Monitoring of rejection episodes may involve a combination of different tests, including biopsies, radiological and nuclear studies, and measurements of enzyme levels and urine output. Kidney graft function is monitored by measurement of the serum creatinine level, which, when elevated, indicates a rejection episode. If rejection is confirmed through biopsy, antirejection therapy is warranted (Abecassis and Corry, 1993). Monitoring of pancreas graft function has been more problematic. For recipients of SPK transplants, kidney rejection can be used as a marker for pancreas rejection since there is good evidence that rejection usually occurs in both organs simultaneously (Robertson and Sutherland, 1992) though there have been isolated cases of pancreas rejection without kidney rejection (Abecassis and Corry, 1993). Because an increase in blood sugar following pancreas transplantation would indicate a non-functioning pancreas, blood glucose levels are monitored; however, increased blood sugar is considered a late marker and earlier indications of rejection are necessary for graft salvage (Robertson and Sutherland, 1992). Measurement of pancreatic exocrine secretions in the urine are valuable markers of graft function, but these are only possible in pancreas recipients who have undergone the bladder drainage method, which is the primary reason this method is the most commonly used. A decrease in the level of one of these enzymes, urinary amylase, is indicative of pancreas rejection, which may be confirmed by percutaneous or transcystoscopic core biopsy. Several experts are reluctant to perform these biopsies due to potential hazards, but others feel they are necessary to improve success rates for PTA or PAK until improved modalities for detecting pancreas rejection are confirmed. Preliminary studies on levels of serum anodal trypsinogen and plasma pancreatic secretory trypsin inhibitor suggest that these biochemical markers may be valuable in detecting pancreas graft rejection (Abecassis and Corry, 1993). When a rejection episode is suspected or confirmed, treatment usually consists of a temporary increase in steroids and/or administration of an anti-T-cell agent, such as OKT3, ALG, or ATG (Robertson and Sutherland, 1992; Taylor et al., 1994).

Complications

The benefits of pancreas transplantation must be weighed against the potential risks of the surgical procedure and chronic immunosuppression. These risks fall into a few general categories, including pancreatitis, dehydration, and infectious, urologic, and vascular complications. Pancreatitis, an inflammation of the pancreas due to autodigestion of pancreatic tissue by its own enzymes, can be related to the development of pseudocysts, rejection, or infection (Taylor et al., 1994). It is felt by some experts that pancreatitis is directly associated with handling, trauma, and injury related to organ procurement. Most patients experience some degree of pancreatitis, which can usually be resolved with non-surgical procedures; however, some proceed to a more severe form of pancreatitis that can lead to multiorgan dysfunction and sepsis or to pancreas graft loss in 20% to 100% of the patients (Grewal et al., 1993). In addition to pancreatitis, fluid management often poses a problem for diabetic pancreas recipients due to pre-existing hypotension related to diabetes. Severe dehydration and metabolic acidosis can result from the loss of sodium and bicarbonate from pancreatic exocrine secretions. Multiple hospital readmissions may be required for certain patients who experience problems with losing fluid through the urinary tract and maintaining an adequate oral intake (Taylor et al., 1994).

As with all solid organ transplants, there is a tendency for opportunistic fungal, viral, and bacterial infections to develop, though these are usually managed easily and do not often contribute to patient or graft loss. However, with pancreas transplantation, there is an increased incidence of urinary tract infections associated with the bladder drainage technique and an increased risk of developing abscesses in the area of the pancreas or within the abdomen. Many of the urinary tract infections observed after pancreas transplantation have been caused by organisms not usually associated with urinary tract infections. It is believed that the combination of bladder drainage and immunosuppression has predisposed the bladder to these infections, making them difficult to eradicate. Moreover, the combination of tissue breakdown from pancreatitis and exposure to bacterial contamination can lead to the formation of abscesses within the abdomen or around the pancreas. In addition to urinary tract infections, the bladder-drainage technique is associated with an increased incidence of other urologic complications, including hematuria, perforation of the duodenal segment, urethritis, and urethral stricture (Taylor et al., 1994). Another serious infectious complication is the development of cytomegalovirus (CMV) in 50% to 70% of pancreas/kidney recipients. While symptoms in most patients may include leukopenia, low-grade fever, and lethargy, in 10% to 30% of these patients, significant CMV disease may develop, leading to intravascular coagulation, major organ damage, and death (Bartucci et al., 1992). The overall incidence of other infectious complications among over 2000 pancreas recipients in the USA was 2% for 2,112 SPK recipients, 3.90% for 205 PAK recipients, and 2% for 147 PTA recipients (Sutherland et al., 1993b).

Vascular complications, including thrombosis, hemorrhage, and embolism, may also develop involving arteries or veins (Taylor et al., 1994). The overall rates of thrombosis and bleeding episodes among the over 2,000 patients included in the International Pancreas Transplant Registry report were, respectively, 6.30% and 0.76% for SPK recipients, 12.7% and 0.98% for PAK recipients, and 10.9% and 0.75% for PTA recipients (Sutherland et al., 1993b). In addition, pancreas graft thrombosis reportedly occurs in 10% to 30% of the cases and is attributed to surgical technique in procurement and placement, acute rejection, and the vasoconstricting effects of the immunosuppressive agent cyclosporin. When graft thrombosis occurs, graft removal is the only option (Grewal et al., 1993). Though rare, recurrence of diabetes has been reported and has been linked to two different circumstances. In a few cases involving segmental donor grafts from nondiabetic siblings to nonimmunosuppressed identical twins, autoimmune isletitus, the original cause of Type I diabetes, developed resulting in beta-cell destruction and disease recurrence (Sutherland et al., 1993b). Type II diabetes has also developed in some patients in response to hyperglycemia resulting from increased glucose production by the liver in response to steroids. In these cases, the functioning pancreas graft secretes additional insulin while peripheral cells remain resistant and require more insulin (Bartucci et al., 1992).

Immunosuppressive or adjuvant therapy may cause additional complications and increase the risk of infections and malignant disease (Remuzzi et al., 1994). The most frequent side effect with azathioprine is severe leukopenia, observed in over 50% of the patients given this drug and usually occurring late in the course of administration but reversing upon dose reduction or withdrawal. Cyclosporin is associated with a 25% incidence of nephrotoxicity or renal dysfunction (PDR, 1995), and OKT3 is associated with a nine-fold increase in lymphoproliferative disorders. Cyclosporin and steroids also increase blood pressure and alter the lipid profile, which may offset the cardiovascular benefits of achieving normal blood glucose levels. High-dose steroids have been linked to a high incidence of peripheral vascular disease among pancreas graft recipients, which has resulted in amputation in 20% of these patients. Steroids have also been responsible for gastrointestinal bleeding and perforation (Remuzzi et al., 1994) and a number of other adverse effects, including fluid and electrolyte disturbances, loss or muscle or bone mass or muscle weakness, dermatologic disturbances, menstrual irregularities, growth suppression, insulin disorders, convulsions, vertigo, headache, cataracts, and glaucoma (PDR, 1995).

Hospital Stay

When compared with kidney transplantation alone (KTA), which is performed routinely for the treatment of diabetic nephropathy, pancreas transplantation is associated with a prolonged postoperative hospital stay and an increased incidence of readmissions for complications. According to a study conducted by the United States Renal Data System (1992) and involving over 3,168 solitary kidney and simultaneous pancreas/kidney procedures in Type I diabetics with end-stage renal disease, the average post-operative hospitalization was 15 days for KTA and 23 days for SPK while the average number of readmission days per year was 20.13 days for KTA recipients and 31.47 days for pancreas recipients.

In 1989, it was estimated that the number of diagnosed diabetics in Minnesota ranged from 90,000 to 107,000 and that about 10% of these (9,000 to 10,700) were insulin-dependent Type 1 diabetics. At that time, the incidence and prevalence of diabetes related end-stage renal disease in Minnesota were approximately 230 and 800, respectively. These figures have increased over the previous ten years and are expected to continue to increase (Roesler, Minnesota Diabetes Surveillance Project, 1989).

Information provided by Minnesota data available in January, 1996 indicates that, from 1987 to 1995, a total of 562 pancreas transplantation procedures had been performed in Minnesota, including 331 SPK procedures, 108 PTA procedures, 122 PAK procedures, and 1 simultaneous pancreas/liver (SPL) procedure. It is unclear if the SPL procedure was performed in a diabetic. Not including the SPL procedure, TABLE 1 shows the frequency and distribution of each type of pancreas transplantation procedure for each year for both the University of Minnesota transplantation center, which performed 87.55% of the procedures and the Mayo Clinic, which performed 12.45%, as well as the overall totals. Based upon these figures, the yearly average from 1987 to 1995 is 62.4 procedures, although the number of procedures varied each year. It appears that the frequency of pancreas transplantation has been steadily increasing, with the exception of 1995, for which the figures may as yet be incomplete. While these procedures primarily used cadaver organs, living related donor (LRD) organs were used in 3 (.9%) of the SPK procedures, 10 (8.1%) of the PAK procedures, and 9 (8.3%) of the PTA procedures. All procedures involving LRD organs were performed at the University of Minnesota center. Among the total number of pancreas transplantations performed in Minnesota, only 128 procedures involved residents of the state. Among Minnesota residents, pancreas transplantations included 95 SPK procedures, 23 PAK procedures, and 10 PTA procedures. The University of Minnesota and the Mayo Clinic performed 75 and 20 SPK procedures, respectively; 22 and 1 PAK procedures, respectively; and 9 and 1 PTA procedures, respectively.

TABLE 1

* This figure includes one pancreas/liver transplantation procedure not shown in the table.

** One additional pancreas transplantation procedure was performed at the University of Minnesota, changing the total of procedures for 1995 in that institution to 55, the total of procedures in that institution to 493, and the total of procedures in Minnesota to 563. However, the type of procedure was not designated, therefore is not included in the table.

FDA-APPROVAL

Immunosuppressive Agents

The immunosuppressive agents approved by the FDA for use with organ transplantation are cyclosporin, azathioprine, polyclonal antithymocyte globulin (equine), muromonab-CD3 (monoclonal antibodies), and tacrolimus (FK506). Cyclosporin (Sandimmune®, manufactured by Sandoz) is indicated for prophylaxis of organ rejection in kidney, liver, and heart allogeneic transplantations. Azathioprine (Imuran®, manufactured by Glaxo-Wellcome) and polyclonal antithymocyte globulin (ATGAM®, manufactured by Upjohn) are FDA-approved to prevent rejection in renal allografts only. Muromonab-CD3 (ORTHOCLONE OKT®3, manufactured by Ortho Biotech) is indicated for treatment of acute allograft rejection in renal transplant patients and for treatment of steroid-resistant acute allograft rejection in heart and liver transplant patients. FK506 (PROGRAF^TM, manufactured by Fujisawa) has been approved by the FDA for the prophylaxis of organ rejection in patients receiving allogeneic liver transplants (PDR, 1995). None of these agents has been specifically approved for pancreas transplantation.

Corticosteroids

Prednisone and methylprednisolone are corticosteroids approved for several indications, including: endocrine disorders, rheumatic disorders, collagen diseases, dermatologic diseases, allergic states, ophthalmic diseases, respiratory diseases, hematologic disorders, neoplastic diseases, edematous states, and gastrointestinal disorders. Both corticosteroids are manufactured under a variety of names by different manufacturers; however, neither is specifically approved for use in organ transplantation (PDR, 1995).

Heparin

Heparin sodium, manufactured by Wyeth-Ayerst Laboratories, Upjohn, and Lilly, is FDA-approved for the following indications: prevention of clotting in arterial and heart surgery, prevention of postoperative deep venous thrombosis and pulmonary embolism in patients undergoing abdomino-thoracic surgery or patients who for other reasons are at risk for developing thromboembolic disease, anticoagulant therapy, and maintenance of patency of indwelling venipuncture devices used for intermittent injection or infusion therapy (PDR, 1995).

PUBLIC AGENCIES AND PROFESSIONAL ASSOCIATIONS

Health Care Financing Administration (HCFA)

In a telephone call on May 4, 1995, a spokesperson from HCFA stated that coverage for pancreas transplantation is not approved because it is considered an experimental procedure.

Agency for Health Care Policy and Research (AHCPR), Division of Office of Health Technology Assessment (OHTA)

The AHCPR has recently completed an assessment report on pancreas transplantation, entitled "Simultaneous Pancreas-Kidney and Sequential Pancreas-after-Kidney Transplantation" (AHCPR, 1995). After reviewing letter responses from transplant centers and over 100 published articles on pancreas transplantation, the AHCPR reported that, although surgical techniques have improved and both simultaneous and sequential combined pancreas/kidney transplantation are associated with relatively low mortality rates, these procedures are also associated with a significantly higher rate of morbidity, resulting in a greater frequency of readmissions and more prolonged hospitalization, than kidney transplantation alone (KTA). In addition, while most complications occur within the first year after transplantation, "it has not yet been determined whether the more intensive immunosuppression required in combined transplants vs. [sic] that required in KTA will be associated with additional deleterious long-term effects" (page 46). Moreover, the published data did not provide clear support of claims that SPK or PAK prevented or ameliorated the secondary diabetic complications of nephropathy, retinopathy, and neuropathy; thus, "It cannot be concluded with reasonable certainty that prevention, improvement, amelioration, or mitigation of secondary complications are likely to result from SPK or PAK or provide a valid rationale for the procedures" (page 46). Further, evaluations of post-transplant quality of life, often proclaimed as a valid objective for pancreas transplantation, suffer from methodologic flaws. Although the available evidence suggests that quality of life is improved after combined pancreas/kidney transplantation, the use of subjective impressions, the lack of prospective studies, and the lack of consistent evidence from objective quality-of-life measures do not allow a clear position on quality of life.

Noting that the majority of SPK transplantations use single donor cadaver organs, the AHCPR stated that this procedure cannot provide the improved renal graft protection associated with a living related donor (LRD) kidney transplant; therefore, a potential recipient who chooses SPK over LRD/KTA must accept a 40% to 70% reduction in expected kidney half life in return for a 65% probability of being insulin-independent at three years post-transplant. Although PAK allows the possibility of a LRD kidney graft and the consequent improvement in kidney graft survival, the patient must undergo two surgical procedures and accept the probability of lower long-term pancreas graft survival associated with PAK (about 35% at three years) as compared with SPK. Moreover, the agency stated that, due to the lack of valid and reliable evidence regarding the benefits of SPK and PAK and the lack of accessible data regarding the costs of these procedures, it was not possible to confidently compare differences in the clinical efficacy of SPK, PAK, and KTA or to determine the cost-effectiveness of combined pancreas/kidney transplantations. Although a cost-effectiveness model, based upon limited cost data and assumptions regarding quality of life, was constructed and demonstrated that SPK was likely to be as cost effective as KTA when considering the annual cost of treating insulin-dependence in KTA recipients, this model did not allow for definitive conclusions regarding cost-effectiveness; it emphasized instead concerns regarding the appropriateness of combined transplant in patients with relatively mild or uncomplicated diabetes.

In fact, the lack of detailed descriptions of patient selection criteria was noted by the AHCPR as the most disturbing characteristic in the published literature. Although frequent references to contraindications were made in the literature, it was not clear if the severity of diabetes, the intensity of the required insulin regimen, and the complications associated with blood glucose control played a significant role in patient selection; several transplant centers informed OHTA that these factors were not considered important criteria in choosing recipients. In addition, the AHCPR expressed doubt regarding the clinical utility of some of the pretransplant tests, such as ultrasound gallbladder studies, prophylactic cholecystectomy in asymptomatic patients, mammography, PSA screening, etc., stating that "these studies might be viewed as variations in the clinical judgement in the face of uncertainty" (page 47).

The AHCPR concluded that in order to form any logical and scientific conclusions regarding combined pancreas/kidney transplantation, additional objective data from well-designed large-scale, long-term prospective studies are needed, especially regarding the benefits and efficacy of the procedures, patient selection criteria, quality of life, and costs. The agency also emphasized that retransplantation is a significant issue that requires careful evaluation. While the national frequency of retransplantation is low, the retransplant rate in those centers performing the highest numbers of SPK and PAK procedures is not insignificant. Because graft survival is lower in retransplantion as compared with primary transplantation procedures, this practice raises a number of issues regarding the medical benefit, cost-effectiveness, and the ethical appropriateness of the use of scarce donor organs.

National Institutes of Health (NIH)

A telephone call on April 18, 1995 revealed that this agency has not conducted an assessment on pancreas transplantation for diabetes.

United Network for Organ Sharing (UNOS)

The United Network for Organ Sharing (UNOS) is a registry for all organ transplantations performed in the United States. This organization administers the National Organ Procurement and Transplant Network. Through a subcontract with the International Pancreas Transplant Registry (IPTR), data from worldwide cases of pancreas transplantation are reported to UNOS. While both organizations record and provide statistical data on the number of procedures performed, patient demographics, procedural details (recipient category, duct management techniques), donor-organ details, and outcomes with respect to patient and graft survival rates, neither forms conclusions regarding the procedures. However, a number of analyses have been performed on the statistical data provided by these organizations. Because these analyses consist of the largest patient populations among published articles on pancreas transplantation, details of the compiled data have been incorporated throughout this report, particularly in the REVIEW OF EVIDENCE section, and emphasized as either UNOS or IPTR statistics.

American Diabetes Association (ADA)

In a position statement entitled "Pancreas Transplantation for Patients with Diabetes Mellitus" (1992a), the American Diabetes Association presents its guidelines based upon its technical review (1992b), which has the same title and provides further details. The agency states that, while pancreas transplantation has been performed in thousands of diabetic patients and has become accepted therapy in certain cases, it is surrounded by issues regarding its indications and respective risks and benefits. Successful transplantation can significantly improve the quality of life of diabetics by eliminating the need for exogenous insulin, blood glucose measurements, dietary restrictions, and complications associated with exogenous insulin therapy, but there is insufficient evidence at this time regarding its effects on the long-term complications of diabetes. Because critical questions such as whether or not pancreas grafting has any positive effect on the development or progression of vascular disease have not yet been clarified, it is too early to determine the ultimate impact of transplantation on life expectancy (1992b). Therefore, prevention or attenuation of chronic diabetic complications and increased life expectancy are not indications for pancreas transplantation (1992a).

The ADA recognizes that pancreas graft survival rates for combined pancreas/kidney transplantation have been higher than those of solitary pancreas transplantation and appear to be due to the more efficient methods for detecting episodes of kidney rejection and the resultant changes in immunosuppression that protect the pancreas graft as well. Based upon survival data, the ADA recommends that combined pancreas/kidney transplantation, either SPK or PAK, be considered an acceptable alternative to insulin therapy for diabetic patients with end-stage renal disease who meet the criteria for kidney transplantation, have significant problems with insulin therapy, and do not have excessive risk for the surgical procedure, and that Medicare and other insurers should include coverage for pancreas procedures meeting these criteria. Moreover, the agency recommends that solitary pancreas transplantation should be considered only for a select few patients who have a history of severe, metabolic complications necessitating medical attention, who have severe incapacitating clinical and emotional problems with exogenous insulin therapy, and for whom other therapeutic approaches to alleviate the situation have failed. If these criteria are met, as well as institutional guidelines for assuring an objective multidisciplinary patient evaluation and determination of eligibility for transplantation, coverage by insurers is appropriate (1992a).

Further, it is recommended that centers that perform pancreas transplantations be tertiary care centers with an active kidney transplant program and the ability to adequately handle the medical and psychosocial needs of the patients. The ADA also adds that, although islet cell transplantation holds potential advantages over whole-gland transplantation, it is considered experimental at this time (1992a).

American Medical Association (AMA)

A telephone call on April 18, 1995 revealed that the Diagnostic and Therapeutic Technology Assessment (DATTA) of pancreatic transplantation completed by the AMA in 1990 has not been revised or updated. According to this report, no consensus was reached by the 26 DATTA panelists on the safety and effectiveness of pancreas transplantation alone or of pancreas transplantation in combination with a kidney transplant. In categorizing both the safety and efficacy of combined pancreas/kidney transplantation, the panelists' conclusions were divided primarily between "promising" and "investigational" ratings. However, two panelists rated the safety of the combined procedure as doubtful, one rated the efficacy as doubtful, and two rated the efficacy as established. In categorizing the safety and efficacy of solitary pancreas transplantation, about half of the panelists (50% and 46%, respectively), regarded this procedure as investigational; six rated the procedure as doubtful, and two felt the procedure was unacceptable. The safety and efficacy of solitary pancreas transplantation was rated as promising by five and six panelists, respectively. The panelists felt that, while improvements in surgical technique, graft preservation, and immunosuppression have resulted in improved patient and graft survival rates, and pancreas grafts have been beneficial in terms of neuropathy, nephropathy, and quality of life, the toxicity of immunosuppression and the monitoring of pancreas rejection in solitary pancreas transplants remain problematic (Brown, 1990).

REVIEW OF EVIDENCE: PATIENT SELECTION CRITERIA

Quality of Available Evidence

Much of the evidence included throughout the report is provided by analyses of data from the United Network of Organ Sharing (UNOS) and the International Pancreas Transplant Registry (IPTR). While providing statistical data regarding outcomes of pancreas transplantation (alone or combined) for a large patient population and determining the recipient categories, duct management techniques, immunosuppressive regimens, patient characteristics, etc., which result in the best outcomes, these analyses do not provide conclusions regarding the safety, efficacy, or appropriateness of pancreas transplantation per se. A number of additional studies have provided such conclusions, but these are limited by small patient samples, incomplete reporting of results, or insufficient data on patient selection criteria.

General Guidelines

Potential organ recipients must undergo a pretransplant evaluation and must qualify for transplantation, satisfying not only strict medical criteria but also psychological criteria that indicate the ability to comply with an elaborate regimen of postoperative care. Organ transplantation is generally only considered in cases of end-stage disease when no other medical or surgical therapy would be beneficial and prognosis for one-year survival is poor (Schwartz et al., 1989). However, pancreas transplantation differs from other organ transplantations because it is not performed to save or prolong life; thus, it is more difficult to determine indications for its use, and patient selection is often determined by balancing the potential benefits of transplantation with the risks of surgery and prolonged immunosuppression. Criteria for pancreas transplantation differ from center to center.

In its assessment report on pancreas transplantation (Brown, 1990), the American Medical Association suggests the following general guidelines for any pancreas transplantation:

• Age of 15 years to 50 years
• ABO blood group compatibility
• absence of systemic complications (gangrene, gastroenteropathy, coronary artery disease)

Additional general guidelines used at most centers performing pancreas transplantation also include (Taylor et al., 1994):

• Type I diabetes
• Predicted ability to tolerate the surgical procedure as well as post-transplant immunosuppression and potential complications
• Emotional and psychiatric stability including a strong family support network
• Ability to understand investigative nature and risks of the procedure
• Presence of overt secondary diabetic complications, especially nephropathy but including retinopathy, and neuropathy

Simultaneous Pancreas/Kidney (SPK) and Pancreas After Kidney (PAK) Transplantation

In an assessment report prepared by the Institute for Clinical Systems Integration (1994), it was stated that major transplant centers, including the University of Wisconsin, University of California/San Francisco, Massachusetts General Hospital, Pacific Presbyterian Medical Center/San Francisco, Ohio State University, and Kaiser-Permanente, all accept the following general patient selection criteria for SPK:

• Age between 18 years to 50 years
• Type I diabetes
• End-stage renal disease not necessarily requiring dialysis
• Mild diabetic complications with no prior multiple abdominal surgeries

Not all transplant centers, however, use the same criteria. With respect to diabetic complications, The University of Michigan center specifies only established diabetic nephropathy or autonomic neuropathy as indications for pancreas transplantation (Brown, 1990), and the University of Iowa does not require renal disease to have progressed to end-stage as long as uremia is present (Abecassis and Corry, 1993).

The criteria that apply to SPK also apply to PAK, except that potential recipients have already ameliorated renal disease with a previous kidney transplantation (Institute for Clinical Systems Integrated, 1994). Patients who have undergone kidney transplantation are already committed to life-long immunosuppression; therefore, the addition of pancreas transplantation adds only the surgical risk of the procedure since immunosuppression does not usually need to be altered (Abecassis and Corry, 1993). The American Diabetes Association (1992a) states that SPK or PAK is appropriate for diabetic patients with end-stage renal disease who meet the criteria for kidney transplantation, have significant problems with insulin therapy, and do not have excessive risk for the surgical procedure.

Though indications vary from center to center, the guidelines followed by the leading transplant centers have been modified by clinical experience. While not all diabetics with kidney failure are appropriate candidates for SPK, those between the ages of 20 to 40 years who have minimal secondary complications of diabetes are considered optimal candidates at the University of Nebraska center. The degree of renal dysfunction and the degree of cardiovascular risk are also primary determinants of patient selection (Stratta et al., 1993).

The timing of SPK relative to the degree of nephropathy is one of the controversial issues of patient selection. At the University of Nebraska, patients with creatinine clearance of 45 mL/min are considered eligible for SPK because it is felt that the procedure can be performed safely and effectively in the absence of uremia, which not only facilitates rehabilitation but also provides the possibility of arresting the progression of diabetic complications prior to the development of end-stage renal disease. Patients with a creatinine clearance of over 70 mL/min would be considered for PTA, rather than SPK (Stratta et al., 1993). Not all experts agree with these specifics. Taylor et al. (1994) assert that SPK is indicated by a creatinine clearance of < 40 mL/min and that a creatinine clearance of > 45 mL/min may better be treated with PTA.

Pancreas Transplantation Alone (PTA)

The selection criteria for PTA are less clear but are generally based upon exogenous insulin failure or the presence of early diabetic complications. The American Medical Association (Brown,1990) states that PTA is indicated in cases of preuremia with lesions that predict progression to end-stage renal disease. In addition, the diabetic complications that are predicted if no transplant is performed must outweigh the potential side effects from immunosuppression. At the University of Minnesota transplant center, potential PTA recipients must have nonrenal complications of diabetes not only refractory to exogenous insulin but also judged to become more severe than the side effects of immunosuppression and correctable by pancreas transplantation (Robertson and Sutherland, 1992).

Ideally, PTA should be performed prior to the development of secondary diabetic complications; however, before the earliest signs of these complications are present, there are no reliable markers that predict which patients will development them, although they can be detected early (Stratta et al., 1993; Abecassis and Corry, 1993). Because there are no accurate methods to predict the development of secondary complications, the Institute of Clinical Systems Integration (1994) does not recommend PTA at all.

At the University of Nebraska, insulin independence in itself is not considered justification for the requirement of chronic immunosuppression associated with PTA unless a positive effect on diabetic complications is possible; therefore, potential PTA recipients must demonstrate a propensity toward these complications, which are predicted to be worse than the undesirable side effects of immunosuppression (Stratta et al., 1993). It is unclear whether this propensity must be proven, but the University of Iowa center does require documented evidence of early secondary complications, such as the presence of albuminuria or reduced creatinine clearance or biopsy proof of nephropathy (Abecassis and Corry, 1993).

The University of Minnesota center, which performs more PTA procedures than other centers, uses PTA primarily to treat patients with extremely labile diabetes, defined as oscillation "between the extremes of hyper- and hypoglycemia, no matter what insulin regimen is used, even when under the supervision of a diabetologist or other physician experienced in diabetic management." (page 1602). Most potential recipients at this institution also experience hypoglycemic unawareness but, in general, have less advanced secondary complications than SPK recipients (Gruessner et al., 1994). In addition, these patients were unable to live independently, requiring constant presence of a friend or relative, due to significant problems with insulin control (Sutherland et al., 1994).

The University of Michigan center defines labile diabetes more specifically (Brown, 1990):

• Diabetes refractory to treatment with exogenous insulin with hospitalization for 90 or more days in the past year for glucose control problems
• Occurrence of six or more hypoglycemic episodes in one month requiring emergency medical attention
• Occurrence of four or more episodes of unexplained ketoacidosis during the past year

It has been argued that most patients with this degree of lability are either noncompliant or have significant family or individual psychopathology, and that transplantation may be inappropriate for patients such as this (Goldon and Orr, 1989). But proponents of pancreas transplantation for labile diabetes, while agreeing with the necessity for compliance, contend that successful transplantation is associated with patient satisfaction with glucose control, insulin-independence, and improved sense of well-being, which contribute to an improved attitude and compliance (Vinek and Dafoe, 1989).

The American Diabetes Association (1992a) recommends that solitary pancreas transplantation be considered only for a select few patients who have a history of severe, metabolic complications necessitating medical attention, who have severe incapacitating clinical and emotional problems with exogenous insulin therapy, and for whom other therapeutic approaches to alleviate the situation have failed.

Most transplant centers consider the following circumstances contraindications to pancreas transplantation (Taylor et al., 1994):

• Significant coronary artery disease
• Obesity
• Presence of active infection or malignancy
• Active substance abuse or dependence
• History of noncompliance

Further exclusion criteria are applied at some centers. The University of Michigan considers psychiatric illness, peptic ulcer, and advanced nephropathy with creatinine clearance of < 1.00 mL/sec as contraindications (Brown, 1990), and the University of Nebraska considers active smoking a relative contraindication (Stratta et al., 1993).

While some centers feel that blindness or peripheral vascular disease are relevant exclusion criteria, others do not agree. Following an analysis of 319 pancreas transplantation procedures (all types) performed at the University of Minnesota center to determine recipient risk factors, Gruessner et al. (1994) concluded that blindness, peripheral vascular disease, and duration of diabetes did not negatively affect patient and graft outcomes in any recipient category; however, age 45 years and cardiac disease were risk factors for SPK and PAK recipients.

Although PAK is considered in diabetic patients with a well-functioning kidney since these patients are already committed to immunosuppression, poor results with PAK at the University of Nebraska have been attributed to previous history of significant soft tissue infection, which has often resulted in reinfection following transplantation and, subsequently, pancreatitis caused by the same organism responsible for soft tissue infection. Thus, at this institution, a history of soft tissue infection is considered a contraindication to PAK, as well as a relative contraindication to PTA (Stratta et al., 1993).

CONCLUSIONS: PATIENT SELECTION CRITERIA

While it must be noted that patient selection criteria for pancreas transplantation are not standardized or static and specific definitions and limitations apply at different centers, the guidelines provided in TABLE 2 appear to be generally accepted at most transplant centers. Specific indications for SPK or PAK include the presence of renal disease and mild secondary diabetic complications, and the lack of prior multiple abdominal surgeries. Indications for PTA include preuremia or nonuremia with a propensity toward diabetic complications, which are predicted to be worse than the undesirable side effects of immunosuppression, or extremely labile diabetes, refractory to any type of exogenous insulin regimen or other therapeutic approach, with severe metabolic complications and incapacitating clinical or emotional problems.

TABLE 2

INDICATIONS FOR PANCREAS TRANSPLANTATION

1. General guidelines for any pancreas transplantation:
   • Age of 18 years to 50 years
   • ABO blood group compatibility
   • Lack of systemic complications
   • Type I diabetes
   • Predicted ability to tolerate the surgical procedure as well as post-transplant immunosuppression and potential complications
   • Emotional and psychiatric stability including a strong family support network
   • Ability to understand investigative nature and risks of the procedure
   • The presence of overt secondary diabetic complications, especially nephropathy but including retinopathy, and neuropathy

2. Indications for SPK and PAK:
   • Renal disease (no consensus on degree of renal dysfunction)
   • Mild secondary diabetic complications
   • No prior multiple abdominal surgeries

3. Indications for PTA:
   • Preuremia or nonuremia with propensity toward diabetic complications which are predicted to be worse than the undesirable side effects of immunosuppression

     or

   • Extremely labile diabetes, refractory to any type of exogenous insulin regimen or other therapeutic approach, with severe metabolic complications and incapacitating clinical or emotional problems

4. Contraindications:
   • Absolute contraindications
     
     • Significant coronary artery disease
     • Presence of active infection or malignancy
     • Active substance abuse or dependence
     • History of noncompliance
     • Psychiatric illness
     • Peptic ulcer
     • Advanced nephropathy with creatinine clearance of < 1.00 mL/s
     • Obesity

   • Relative contraindications
     • Active smoking
     • Blindness
     • Peripheral vascular disease
     • History of soft tissue infection

REVIEW OF EVIDENCE: TREATMENT OUTCOMES

Patient and Graft Survival

Sutherland et al. (1993b) presented an analysis of patient and graft survival rates for all pancreas transplantations reported to the United Network of Organ Sharing (UNOS) from October 1987 to November 1993, including those reported through a subcontract with the International Pancreas Transplant Registry (IPTR). The analysis excluded all cases with insufficient data or no follow-up, all cases involving simultaneous transplantation with the heart or liver in nondiabetic patients, and, due to the small number involved, all transplants involving a living related donor or involving duct drainage methods other than the bladder drainage method for the USA cases. According to the definitions used in this analysis, pancreas graft survival refers only to pancreas grafts that are functioning to the point of insulin-independence for patients, and partial graft function or patient death with a functioning graft is considered a graft failure.

Results of the analysis revealed that, among the 2,464 remaining cadaveric pancreas transplants involving the bladder drainage technique (96%) in Type I diabetics in the USA, the overall one-, two-, and three-year survival rates were 91%, 87%, and 84%, respectively, for patients and 72%, 67%, and 62%, respectively, for grafts. For the 93 retransplant cases, one-year graft survival was significantly lower than that for the 2,371 primary transplant cases (42% versus 73%, respectively). Data for the 3,569 worldwide cases are presented in TABLE 3. For solitary pancreas transplantation, the graft survival rates were significantly higher in the USA than in Europe (48% versus 26%, respectively). The lower pancreas-graft survival rate in Europe is attributed to the methods used for exocrine duct management. While 96% of the pancreas transplants performed in the USA used the bladder drainage technique, only 62% of the European cases used bladder drainage with the remaining cases managed with duct injection (28%), enteric drainage (9%), or other methods (7 patients).

TABLE 3

OVERALL ONE-YEAR PATIENT AND GRAFT SURVIVAL RATES (Reported to IPTR* October 1987 to November 1993)
Survival Category	USA	Europe	Other	Overall (p Values)
A. Patient (3569 patients)	91% (2,573 patients)	92% (921 patients)	86% (75 patients)	91% (<0.08)
B. Pancreas Graft	71%	68%	78%	72% (0.02 WC, 0.38 LR)
C. Kidney Graft (SPK only)	84%	84%	83%	84% (NS)

* IPTR: International Pancreas Transplant Registry (reported by Sutherland et al., 1993b)

Further analyses were performed on the USA data. Analysis of the bladder drainage cases according to recipient category (SPK, PAK, PTA) demonstrated that, while one-year patient survival rates did not differ (91%, 92%, and 91%, respectively), one-year pancreas graft survival was significantly higher in the SPK category at 76% than in the PAK or PTA categories at 47% and 48%, respectively. One-year kidney graft survival was 85% in the SPK group. Although sex was not found to influence outcome, age made a significant difference. One-year rates for patient survival and insulin-independence were 92% and 72%, respectively, for recipients under 45 years of age and 85% and 69%, respectively, for those above 45 years. These differences, however, were observed only in the SPK and PAK categories. With respect to race, outcomes were slightly better for African Americans than for whites with one-year patient survival rates of 95% versus 91%, respectively, and one-year graft survival rates of 84% versus 75%, respectively, although only 6% of the pancreas transplantation procedures were performed on African Americans. Graft-preservation time was also analyzed. While no differences in graft survival were found in the SPK category for storage times of less than 12 hours to 30 hours or more, outcomes in the PAK and PTA category were variable and no consistent pattern could be determined (Sutherland et al., 1993b).

The effects of different immunosuppressive regimens and mismatching of different HLA antigens were also compared. Although maintenance immunosuppression was found to be similar for most transplants, induction immunosuppression consisted of either ALG/ATG, OKT3, or neither of these. In the SPK category, one-year graft survival rates did not differ according to the immunosuppressive protocol; however, these rates were 58%, 26%, and 53%, respectively, for PAK recipients and 53%, 49%, and none [sic (page 68)], respectively, for PTA recipients. With respect to HLA-mismatching, one-year graft survival was actually better with 2 to 6 mismatched HLA antigens (76%) than with 0 to 1 mismatched HLA antigen in the SPK category. However, graft survival was significantly higher in the PAK category with 0 to 1 mismatched HLA antigen (66%) than for 2 to 6 mismatches (44%) and higher, though not significantly, in the PTA category for 0 to 1 mismatch (56%) than for 2 to 6 mismatches (48%) (Sutherland et al., 1993b).

Grafts were considered unsuccessful if they were nonfunctioning or partially functioning or if patient death occurred even with a functioning graft. The total number of cases classified as graft failures was 770. Among 542 grafts listed as nonfunctioning, graft failure was due to technical reasons, including thrombosis, bleeding, infection, anastomotic leak, and pancreatitis, or acute or chronic rejection. Analysis of the UNOS data of bladder-drained cadaver pancreas transplants revealed an overall technical failure rate of 12.5%

The authors concluded that, while, overall, recipient category was the most significant factor influencing the outcomes of pancreas transplantation with higher graft survival rates for SPK than for PAK or PTA, other variables also affected results with respect to recipient category. Minimizing HLA mismatches was associated with higher graft survival in the PAK and PTA categories, and lower graft survival was observed in the PAK category for retransplantation procedures as compared with primary transplants. Moreover, the risk of graft loss was significantly lowered in the PTA category with the use of anti-T-cell induction immunosuppression (Sutherland et al., 1993b).

After analyzing data on 435 diabetic patients receiving either a cadaver or living related donor kidney transplant (KTA), a living related donor kidney transplant followed by a cadaver pancreas transplant (PAK), or cadaver SPK at the University of Minnesota transplant center, Cheung and colleagues (1993) determined that kidney transplant from a living related donor followed by pancreas transplant from a cadaver donor was associated with significantly higher long-term kidney graft survival and equivalent pancreas graft survival as compared with cadaver KTA, living related donor KTA, or cadaver SPK (94%, 69%, 89%, and 63%, respectively, at three years). Therefore, the best option for those eligible for combined pancreas/kidney transplantation is to first receive a living related donor kidney followed by pancreas transplantation of either a whole or segmental cadaver organ or a segmental living related donor organ (PAK). When organ donation from a living relative is not possible, SPK is recommended at the University of Minnesota (Goetz et al., 1991).

Metabolic or Hormonal Control

Graft survival rates reported by Sutherland et al. (1993b) in the analysis of the International Pancreas Transplant Registry (IPTR) refer only to grafts functioning to the point of insulin-independence for patients. Among a total of 3,569 worldwide cases, insulin independence was achieved by 71% of 573 USA patients, 68% of 921 European patients, and 78% of 75 patients transplanted elsewhere. (See TABLE 3). For retransplant cases, one-year graft survival was significantly lower than that for primary transplant cases (42% versus 73%, respectively). Although improved surgical techniques and immunosuppressive protocols have resulted in improved graft survival since the inception of the United Network of Organ Sharing (UNOS), graft survival was found to be influenced by a number of factors, in particular, recipient category (SPK, PAK, PTA). Among over 2,000 transplants performed in the USA, 96% of which were SPK, one-year pancreas graft survival was significantly higher in the SPK category at 76% than in the PAK or PTA categories at 47% and 48%, respectively. For PTA, the graft survival rates were significantly higher in the USA than in Europe (48% versus 26%, respectively), which has been attributed to the higher usage in the USA of the bladder drainage method of exocrine duct management. One-year graft survival rates in the USA were 76% for the bladder drainage method and 57% for the enteric drainage method. While the number of procedures involving the duct injection method was too few to include in the analysis for non-European locations, the one-year graft survival rate for this method in Europe was 61%. (For further details on recipient category and duct management techniques, see REVIEW OF EVIDENCE: ALTERNATIVE TECHNOLOGIES.) While gender was not found to influence graft survival and no conclusions were drawn regarding the effects of different organ storage times, graft survival was found to be influenced by age, race, immunosuppression, and HLA-mismatching with respect to recipient category.

Additional studies, while on a much smaller scale, provide more detailed data on metabolic control, complications of pancreas transplantation, and reasons for graft loss. Results from these studies are provided in TABLE 4.

TABLE 4

PANCREAS ALONE OR PANCREAS/KIDNEY TRANSPLANTATION OUTCOMES
Study	Patient Characteristics	Procedure Protocol	Results	Conclusions	Comments
Morel et al. (1992) 86 subjects Retrospective study to analyze effect of pancreas transplant on metabolic control	31 Type I diabetics Mean age: 32 yrs. Mean duration of diabetes: 20 yrs. 55 healthy controls	3 SPK, 12 PAK, 16 PTA IV glucose tolerance tests (IVGTT), oral glucose tolerance tests (OGTT), plasma glucose determinations before and after meals and in the evening, determination of glycosated fraction of hemoglobin (total A1)	IVGTT: similar in study and control groups early in study but significantly higher in study group at 72 mos post-transplant OGTT: similar in both groups Plasma glucose: only values immediately after lunch and dinner were significantly higher in study group, though mean glucose level well below what is considered diabetic Total A1: within normal range throughout study and no difference between groups	Though pancreas transplantation does not restore an entirely normal metabolism, it allows a close to normal metabolism that is more nearly normal than that afforded by any other form of diabetic therapy. The initial quality of metabolic control achieved with pancreas transplantation is sustained over a prolonged period of time.	Small study sample
Nyberg et al. (1991a) 98 subjects Prospective study to assess effects of pancreas transplantation on glycemic control and serum lipids	27 Type 1 diabetics with uremia awaiting SPK 23 Type 1 diabetics with kidney transplant 23 Nondiabetics with kidney transplant 30 Type 1 diabetics with pancreas/kidney transplants Similar demographics in all groups; transplanted patients on immunosuppression	Determination of serum HDL cholesterol and triglycerides, IV Glucose tolerance test (IVGTT), Determination of glomerular filtration rate	Better metabolic control in recipients of grafts than in diabetics w/o transplants though higher than upper normal limits (hyperinsulinemia) Triglycerides increased in uremic patients and in nondiabetic kidney-grafted patients but normal in diabetic kidney-grafted patients w or w/o pancreas graft Total cholesterol increased in uremic diabetic group HDL normal for all	Hyperinsulinemia is due mostly to insulin delivery to the systemic circulation or denervation of the pancreas and partly to the use of corticosteroids. Renal insufficiency is known to cause increased lipid and triglyceride levels, and findings of this study concur with this. It is not known whether hyperinsulinemia following pancreas transplantation is harmful.
Di Carlo et al. (1991) 32 subjects Retrospective study to report experience	Type I diabetes and uremia with vasculopathy, neuropathy, and retinopathy Mean age: 38.7 yrs. Mean duration of diabetes: 24.6 yrs. Mean duration of dialysis: 17.4 mos.	SPK in all patients plus one retransplant (total 33), including 26 segmental grafts with neoprine duct occlusion and 7 duodeno-pancreatic graft with bladder drainage Immuno: azathioprine, prednisone, and ALG or cyclosporin Anticoagulation: heparin acenocoumarol until Mar. 1988, Dextran Dipyridamole or antithrombin III Acenocoumarol until Sep. 1989, Calcium Heparin Dipyridamole through end of study	Patient survival: 1-yr: 93%, 2-yr: 89% Kidney graft survival: 1-yr: 93%, 5-yr: 72% Pancreas graft survival: 1-yr: 68%, 5-yr: 58% Mortality: 6% (due to myocardial infarction, one due to septic bleeding) Complications: Surgical: 51% Rejection/effects from immunosuppression: 71% Nonsurgical infection: 43% Pancreas graft failure: 13 (39%) due to thrombosis (20%), rejection (12%), primary nonfunction (3%), or death with functioning graft (3%) Kidney graft failure: 6 (18%) due to rejection (12%), thrombosis (3%), or death with functioning graft (3%)	Technical failures, which hinder pancreas graft survival, require further improvement. The authors suggest more aggressive post-operative monitoring and treatment and more intense cardiac work-up for patient selection.	Small patient sample
Martin et al. (1991) 182 subjects Retrospective study to compare different surgical, drainage, and immunosuppressive techniques	Type I diabetes and neuropathy in all patients, with retinopathy (178), macroangiopathy (71), and/or myocardiopathy (62) Mean age: 37.2 yrs. (1976-85) 36.9 yrs. (1985-90) Mean duration of diabetes: 22.0 yrs. (1976-85) 36.9 yrs. (1985-90) Patients on dialysis: 46 (1976-85) 111 (1985-90) Duration on dialysis: 10. mos. (1976-85) 16. mos. (1985-90)	PA for all patients; segmental grafts with neoprene duct occlusion, duodeno-pancreatic grafts with enteric diversion, or duodeno-pancreatic grafts with bladder drainage Immuno: variety of regimens using different combinations of azathioprine, prednisone, cyclosporin, ALG, and OKT3	Graft survival: 40.6% (74 out of 182) Overall 5-yr survival of pancreas transplants (not clear if this is graft or patient survival): 45% Graft losses due to: Rejection (39%) Venous or arterial thrombosis (42 cases) Pancreatitis or hemorrhage (3 cases) Mortality: 13.1% (due mostly to cardiovascular complications) Surgical complications requiring re-operation: 6 in patients with enteric drainage, 1 in patients with duct occlusion	The incidence of rejection was lowered with improved immunosuppressive techniques. The use of urinary amylases as a marker for pancreas rejection in patients with bladder drainage was disappointing. The incidence of venous thrombosis, a major cause of graft loss, decreased since the use of University of Wisconsin solution. Although a higher rate of surgical complications occurred after enteric drainage, graft function is similar no matter which surgical technique is used. Longer follow-up is necessary to determine the best surgical technique.	Some results not reported clearly
Königsrainer et al. (1991b) 54 subjects Retrospective study to report experience	Type I diabetes Mean age: 36.3 yrs (Group I: delayed duct occlusion) 36.9 yrs (Group II: bladder drainage) Mean duration of diabetes: 29.5 yrs (Group I) 23.0 yrs (Group II) Mean time on dialysis: 26.0 mos (Group I) 26.5 mos (Group II) HLA mismatches: AB: 2.7 (Group I) 3.1 (Group II) DR: 1.5 (Group I) 1.5 (Group II)	49 SPK (same donor), 5 PTA, 2 retransplant; segmental graft with tail and parts of head of the gland, including delayed duct (DD) occlusion in 16 patients and bladder drainage (BD) in 38 patients Immuno: cyclosporin, azathioprine, steroids	1-yr patient survival: 94% (DD), 90% (BD) 1-yr pancreas graft survival: 72 % (DD), 74% (BD) 1-yr renal graft survival: 93% (DD), 89% (BD) Insulin independence: 75% of surviving patients Mortality: 12.5% (2) in DD group due to myocardial infarction or amputation of both legs), 15.7% (6) in BD group due to myocardial infarction, cerebral hemorrhage, cytomegalovirus or bacterial pneumonia) Graft loss due to chronic rejection (4) or pancreatitis (1) in DD group and rejection (7), venous thrombosis (3), or pancreatitis (1) in BD group Other complications: severe systemic herpes (1), wound infection or fistulae (7), reoperation for hemorrhage (2) in DD group, and reop. for hemorrhage (2) or leakage at pancreatocystomstomy (2) in BD group severe	The best graft site is the intraperitoneal position. Segmental grafting is preferred as it appears to include a sufficient number of islets, is associated with a lower incidence of urinary tract infection, urethritis, balanitis, bleeding and perforation, and transcytoscopical biopsies are easier to perform as compared with whole organ grafting. The duct occlusion method is preferred to the bladder drainage method as it prevents further damage during the early postoperative period and allows precise monitoring of the exocrine functions. exocrine	Small patient sample
Brekke (1991) 75 subjects Retrospective study to compare techniques	Diabetics (type not specified) Prior to 1988, patients with history of cerebrovascular insult, coronary heart disease, and gangrene included; since then excluded	SPK-same donor (63), PTA (9), and PAK (5) Segmental graft with neoprene duct occlusion before 1988, whole organ graft with duodeno-pancreatic cystostomy since 1988 Immuno: triple drug regimen (agents not specified) Anticoagulation: Dextron 70 then aspirin	1-yr patient survival: PAK: 100% SPK: 93% - 4 deaths due to myocardial infarction (2), cytomegalovirus (1), fungal sepsis (1) PTA: 100 % (1 death 5 yrs post-transplant) 1-yr pancreas graft survival: PTA: 33% (6 lost function) PAK: 0% (all due to chronic rejection) SPK: Not specified	PAK and PTA are less successful than SPK. While the bladder drainage technique allows for monitoring of urinary amylase excretion to detect pancreas rejection, it is associated with a high rate of graft loss; however, 2-year graft survival rates were improved with whole, as opposed to segmental, grafts using the bladder drainage method. Due to unsatisfactory results in nonuremic patients, the author suggests pancreas transplantation should be reserved for uremic patients and SKP should be the procedure performed.	Small patient sample Insufficient data on patient selection criteria Incomplete reporting of results
Esmatjes et al. (1991) 46 subjects Retrospective study to assess effects of pancreas transplantation on metabolic control	31 Type 1 diabetics with uremia Mean age: 35 yrs. Mean duration of diabetes: 19 yrs. 7 healthy controls 7 nondiabetic patients with kidney graft and immunosuppression	SPK for all diabetics Oral glucose tolerance test (OGTT), IV glucose tolerance test (IVGTT), Determination of plasma insulin and C-peptide concentrations Follow-up up to 7 yrs.	Insulin-independence for all SPK patients Insulin levels in SPK group higher than normal but not as high as kidney-grafted group Patients in both transplanted groups demonstrated mild insulin resistance, probably due to immunosuppression since prednisone is known to induce peripheral insulin resistance and cyclosporin is known to reduce plasma clearance of prednisone, resulting in increased prednisone concentration in plasma.	Pancreas transplantation results in long-term normalization of glucose metabolism but with a degree of insulin resistance attributed to immunosuppression.	Small patient sample
Hopt et al. (1991) 40 subjects Retrospective study to report experience	Type I diabetes with end-stage renal disease (ESRD) and advanced retinopathy in all patients, neuropathy and peripheral microangiopathy in 30 patients Mean age: 39.5 yrs Mean duration of diabetes: 24.5 yrs Mean duration on dialysis: 1.9 yrs	SPK (same donor) using pancreaticoduodenal graft and bladder drainage method Immuno: triple drug (azathioprine, cyclosporin, prednisolone) in all patients; addition of ATG in last 16 patients (quadruple drug regimen) Anticoagulation: Dextran (5 days) and heparin (10 days) followed by aspirin beginning at day 10	Normoglycemia reached in 37 patients (93%) immediately; 3 patients required exogenous insulin in early post-operative period. 3-yr patient survival: 87.5% (2 early deaths due to pyelonephritis and virus-myocarditis; three late post-transplant deaths due to pulmonary embolus, systemic infection, or micro- and macroangiopathy) 1-yr graft survival (pancreas and kidney): 85% 3-yr graft survival: Pancreas: 85% Kidney: 75% Complications: Pancreatitis: 10 (25%) Reoperation due to bleeding or hematoma: 6 (15%) Rejection episodes: 30 (75%) Improvement in quality of life reported by > 90% of the patients bleeding	The frequency of rejection episodes was lower with quadruple drug therapy. Though patient survival could be improved, late deaths were due to preexisting secondary diabetic lesions. While postoperative bleeding is low, the authors recommend the use of anticoagulation therapy after pancreas transplantation. SPK using pancreaticoduodenal grafts and the bladder drainage method is successful and is recommended for treatment of patients with Type I diabetes and ESRD.	Small patient sample
Boudreaux et al. (1991) 15 subjects Retrospective study to report experience	Type I diabetes with renal failure, retinopathy, and some degree of neuropathy All patients dialysis-dependent	SPK for 14 patients including 1 who required retransplant of pancreas, PAK for 1 patient Whole organ grafts with bladder drainage technique Immuno: cyclosporin, azathioprine, prednisone, and ALG Anticoagulation: Dextran (4 days) and aspirin (3 mos) in early transplants; aspirin and dipyridamole in later transplants	All patients normoglycemic and insulin-independent in recovery room and up to 20 months follow-up 1-20 month survival rates: Patient: 87% Kidney graft: 87% Pancreas graft: 81% Complications: rejection episodes (17); re-operation for bleeding, bladder leak, or conversion to enteric drainage (3), thrombosis (1), hematuria (5), bladder leak (2), wound infection (2), pneumonia (1), pancreatitis (3), pseudocyst (1), late deaths (2 due to myocardial infarction or cytomegalovirus)	Combined pancreas/ kidney transplantation (SPK) is the procedure of choice for insulin-dependent patients with end-stage renal disease and no significant cardiovascular disease.	Small patient sample

CONCLUSIONS: TREATMENT OUTCOMES

Among all the studies reported in this assessment, the one-year patient survival rate ranges from 86% to 100%, the one-year kidney graft survival rate ranges from 84% to 93%, and the one-year pancreas graft survival rate ranges from 68% to 78%, though one small study reported as low as 0% pancreas graft survival. Retransplantations are associated with a significantly lower one-year pancreas graft survival (42%) as compared with primary transplantations (73%). Besides retransplantation, pancreas graft survival is also influenced by recipient category, exocrine management techniques, immunosuppressive regimens, number of HLA mismatches, and age over 45 years. Complications reported include: patient death (up to 14%), surgical complications (up to 51%), rejection (12% to 75%), and nonsurgical infections (up to 43%). Some of the specific complications among these categories are wound infection, fistulae, re-operation for hemorrhage (15%), anastomotic or bladder leak, pancreatitis (25%), and pseudocyst. Reasons for pancreas graft failure (22% to 32%) are thrombosis, chronic rejection, pancreatitis, hemorrhage, and patient death. Mortality (0% to 14%) has been attributed to cardiovascular complications (especially myocardial infarction but also including pulmonary embolus, cerebral hemorrhage, and micro- and macro-angiopathy), amputation, cytomegalovirus, bacterial pneumonia, fungal myocarditis, and systemic infection. It has been noted that the incidence of venous thrombosis has decreased since the use of University of Wisconsin solution, and the incidence of rejection has decreased with quadruple drug immunosuppressive therapy. It has also been emphasized that technical factors require further improvement and more intense cardiac work-up for patient selection is recommended.

Insulin-independence was achieved in 68% to 78% of the cases transplanted. Although pancreas transplantation does not restore a perfectly normal glucose metabolism, it restores it to a degree that is closer to normal than other forms of diabetic therapy, and this metabolic control is sustained over a prolonged period of time. Either hyperinsulinemia or mild insulin resistance has been noted in many pancreas recipients. These conditions have been attributed to insulin delivery to the systemic circulation, denervation of the pancreas, and immunosuppression. It is not known at this time if hyperinsulinemia following pancreas transplantation is harmful.

Although complications are numerous, patient survival rates suggest that pancreas transplantation is a relatively safe procedure, and pancreas graft survival rates indicate that the procedure is effective in achieving near normal glucose control.

REVIEW OF EVIDENCE: ALTERNATIVE TECHNOLOGIES

Insulin Therapy

The American Diabetes Association (ADA) (1993) evaluated the results of the Diabetes Control and Complications Trial (DCCT), which compared blood-glucose control and diabetic complications between diabetic subjects treated either with conventional therapy or tightly monitored and controlled insulin therapy administered by multiple injections or insulin pump (intensive treatment). Over a study period of about seven years, results of the trial revealed that, compared with conventional treatment, intensive insulin therapy was associated with an approximate 60% reduction in risk for retinopathy, neuropathy, and nephropathy and with a delay in the onset and a slowing of progression of these complications. However, normalization of blood-glucose values was not achieved with intensive insulin therapy, which resulted in mean glucose values of about 40% above normal limits, a three-fold increase in the incidence of hypoglycemia, and a significant weight gain, which could have adverse effects. The ADA asserted that, although tight control may not be appropriate for all diabetics, patients should aim for the best level of glucose control they can achieve without undue risks or hazards.

No studies could be found in the available literature comparing insulin therapy, either conventional or tightly controlled, with pancreas transplantation. While numerous studies report on the differences between successful pancreas grafts with the achievement of insulin-independence, and failed grafts, necessitating exogenous insulin, it is possible that these differences are influenced by the administration of immunosuppressants, which are associated with insulin problems. While not specifically comparing pancreas transplantation with insulin therapy, the findings of the DCCT demonstrate that improved insulin control results in a delay or slowing of diabetic complications. Moreover, these findings emphasize the imperfect control of insulin with intensive insulin therapy.

Simultaneous Pancreas/Kidney Transplantation (SPK) Versus Kidney Transplantation Alone (KTA)

In the 1980s, renal transplantation became one of the preferred methods for treating diabetic nephropathy. Since then, combined pancreas/kidney transplantation has also emerged as a treatment option for this indication. However, because the combined procedure remains somewhat controversial due to the increased morbidity rate (Stratta et al., 1993a), a few investigators have compared results of the different procedures. Results from these studies are provided in TABLE 5

Overall, the combined results of these studies indicate that, although one-year patient survival has been similar for SPK and KTA, one-year kidney graft survival rates have been higher with SPK (83.2% to 96%) than those of KTA (71.2% to 80%). However, SPK is associated with significantly increased hospitalization and rates of readmission and morbidity.

TABLE 5

KIDNEY ALONE OR PANCREAS/KIDNEY TRANSPLANTATION OUTCOMES
Study	Patient Characteristics	Procedure Protocol	Results	Conclusions	Comments
USRDS* (1992) 3,168 subjects Retrospective study to compare SPK and KTA * USRDS: United States Renal Data System	Type I diabetics with ESRD, aged 18-45 yrs. (possibility of some Type II diabetics in KTA group due to imprecise or unavailable data) SPK patients had shorter delay between ESRD and transplant and were younger than KTA patients.	SPK for 380 patients KTA for 2,788 patients Shorter cold ischemia times and improved preservation for donor kidney in SPK group More HLA mismatches in SPK group	1-yr. patient survival: 90.2% for KTA vs 90.3% for SPK 1-yr. kidney graft survival: 74.8% for KTA vs 83.2% for SKP Hospitalization: mean 15 days for KTA vs mean 23 days for SPK Readmissions: 1.82 times per year and 20.13 total days for KTA vs 2.88 times per year and 31.47 for SPK	While patient survival is similar in both groups, kidney graft survival is significantly higher in SPK recipients than in KTA recipients; however, SPK is associated with significantly higher post-transplant hospitalization rates, suggesting higher post-transplant morbidity for SPK recipients than for KTA recipients.	Effects of pancreas transplant itself not assessed Results not adjusted for factors related to transplant success (HLA-matching, time from onset of ESRD to transplant, etc.)
Cheung et al. (1992) 128 subjects Retrospective study to compare SPK and KTA	Type I diabetics with chronic renal failure Higher proportion of patients > 45 years of age in KTA group	59 primary cadaver KTA 69 primary cadaver SPK More aggressive induction immunosuppression for SPK due to tendency for high incidence of acute rejections	1-yr patient survival: 87% for SPK vs 95% for KTA 1-yr graft survival: 80% for SPK vs 86% for KTA Acute rejection episodes: 72% for SPK vs 39% for KTA Hospitalization: 23.3 days for SPK vs 13.5 days for KTA, Readmissions: 2.78 for SPK vs 1.71 for KTA Significantly higher incidence of complications related to wound problems, bacterial and fungal infections, and urinary tract infections in SPK group	For diabetic patients eligible for renal transplant, decision for additional pancreas transplant should rest with patient; however, increased early morbidity and hospital time should be weighed against improved quality of life.	Effects of pancreas transplant itself not assessed Increase hospitalization for SPK patients attributed to prolonged induction immunosuppression regimen and higher rates of rejection and complications
Stratta et al. (1993a) 123 subjects Prospective study to compare SPK and KTA	92 Type I diabetics either on dialysis (30) or w/o dialysis (62) and 31 nondiabetics Age-matched	KTA for 31 nondiabetics (KTA:non-dm) and 31 diabetics (KTA:dm) SPK for 31 diabetics w/o dialysis (SPK:nd) and 30 diabetics on dialysis (SPK:d) Induction therapy with OKT3 differed individually	One-year patient, kidney, and pancreas survival, respectively: SPK:ND: 96.8%, 96.8%, 93.5% SPK:D: 100%, 96.7%, 93.3% KTA:DM: 87.1%,71% KTA:non-DM: 100%,74.2% Morbidity: Hospitalization, readmissions, incidence of rejection, infections, reoperations all higher in SPK groups though not significant Quality of life improved in 90% of those in SPK groups	SPK eliminates the need for dialysis and exogenous insulin, achieves superior metabolic control, and improves quality of life. Although associated with increased morbidity as compared with KTA, SPK was associated with higher kidney graft survival than KTA. Proper patient selection can overcome much morbidity associated with SPK with excellent patient and graft survival and potential for complete rehabilitation.	Study limited by possible selection bias

Pancreas Recipient Categories (SPK, PAK, PTA)

Although not technically alternative technologies, the different recipient categories of pancreas transplantation are associated with differences in technique, outcome, and complications, thus, have been included in this section.

An analysis of 3,569 cases reported to the International Pancreas Registry (IPTR) is reported in TABLE 6. Although patient survival did not differ significantly among the recipient categories, pancreas graft survival was significantly higher in the SPK recipients in both the USA and Europe (75% and 71%, respectively) than in the PAK recipients (48% and 37%, respectively) and the PTA recipients (48% and 26%, respectively). Analysis of the 2,112 bladder drainage cases in the USA according to recipient category (SPK, PAK, PTA) demonstrated that, while one-year patient survival rates did not differ (91%, 92%, and 91%, respectively), one-year pancreas graft survival was significantly higher in the SPK category at 76% than in the PAK or PTA categories at 47% and 48%, respectively. One-year kidney graft survival was 85% in the SPK group (Sutherland et al., 1993b).

The incidence of complications, excluding rejection episodes, was found to be higher in the PAK recipient category. TABLE 7 provides details on technical complications.

TABLE 6

OVERALL ONE-YEAR PATIENT AND GRAFT SURVIVAL RATES (Reported to IPTR October 1987 to November 1993)
Survival Category	USA	Europe	Other	p Values
A. Patient (3569 patients)
SPK	90% (2,215 patients)	92% (844 patients)	84% (68 patients)	0.01 WC, 0.07 LR
PAK	92% (208 patients)	96% (37 patients)	---	0.03 WC, 0.08 LR
PTA	91% (150 patients)	89% (40 patients)	100% (7 patients)	NS
All	91% (2,573 patients)	92% (921 patients)	86% (75 patients)	<0.08
B. Pancreas Graft
SPK	75%	71%	79%	0.02 WC, 0.35 LR
PAK	48%	37%	---	NS
PTA	48%	26%	71%	<0.01
All	71%	68%	78%	0.02 WC, 0.38 LR
C. Kidney Graft (SPK only)	84%	84%	83%	NS

* UNOS results reported by Sutherland et al. (1993b)

Additional studies have been included for more details. Results from these studies are included in TABLE 8.

TABLE 8