History and Early Results

On December 20, 1893, P. Watson Williams grafted three pieces of a sheep pancreas into the subcutaneous tissues of a diabetic child, who died 3 days later of unrelenting diabetic ketoacidosis (Williams, 1894). This first attempt to treat diabetes with transplantation, although unsuccessful, preceded decades of animal experimentation, in which investigators developed the methods necessary to perform a vascularized pancreas transplant; pancreas transplantation was subsequently used as a model to study diabetes and glucose homeostasis.

The first clinical vascularized pancreas transplantation was performed on December 17, 1966, by William Kelly and Richard Lillehei at the University of Minnesota. The patient had temporary insulin independence but eventually required graft removal and ultimately died of postoperative complications (Kelly et al, 1967). The subsequent early experience with pancreas transplantation at Minnesota and at a few other centers was characterized by some technical success, but no graft functioned beyond 1 year, and the enthusiasm for this procedure dwindled.

During 1975, only six pancreas transplantations were performed worldwide. However, the introduction of cyclosporine as an immunosuppressive medication and further technical refinements allowed for better outcomes following pancreas transplantation such that throughout the 1980s and early 1990s, the number of pancreas transplants performed increased dramatically: almost 1500 pancreas transplantations were performed in the United States during 2004. This number has declined slightly for the past several years, with fewer than 1200 pancreas transplantations performed in the United States during 2010.

Indications and Patient Selection

The majority of patients who undergo pancreatic transplantation have type 1 diabetes mellitus and renal failure. In these individuals, pancreas transplantation is performed with a simultaneous kidney transplantation (SPK; 72% of cases) or after a successful kidney transplantation (PAK; 20% of cases). The normal glucose control achieved by the pancreas transplant may protect the transplanted kidney from recurrent diabetic nephropathy and is beneficial from an overall quality of life perspective. In a small proportion of patients with diabetes that is very brittle and difficult to manage, but with preserved renal function, pancreas transplantation alone (PTA; 8% of cases) is performed. It should be noted that SPK and PAK recipients require immunosuppressive therapy to protect both the kidney and pancreas from rejection, whereas in the case of PTA recipients, the need for immunosuppression is solely for the pancreas itself. This difference becomes important when weighing the risk/benefit ratio for each type of recipient.

Potential pancreas recipients are carefully screened for contraindications to transplantation, such as an ongoing infectious process or malignancy. These candidates almost always have medical comorbidities as a result of secondary complications from diabetes; therefore a thorough assessment of a candidate’s cardiovascular status is essential. Cardiac contraindications to pancreas transplantation include the presence of noncorrectable coronary artery disease, significantly decreased ejection fraction, or myocardial infarction within the preceding 6 months. Recipient age is also important, and in some programs, recipients older than 50 years are not considered candidates because of an increased risk of perioperative complications.

Donor Operation

Selection of an appropriate deceased pancreas donor includes standard donor selection criteria. In addition, a bias exists toward using organs from younger, leaner, and more hemodynamically stable deceased donors. Donors with hemodynamic instability or those that require high doses of vasopressors are considered higher risk. In addition, pancreata with significant steatosis are usually avoided, because they are associated with a greater likelihood of postoperative complications, such as pancreatitis, peripancreatic fat necrosis, and infection. Based on these selection criteria, which are relatively stringent compared with those applied to the liver or kidney, only a fraction of deceased donors are deemed suitable for whole-organ pancreas donation. In the United States, there were 8022 deceased donors during 2009. Of these, only 1233 pancreata were transplanted (15%), compared with 6101 livers (76%) and 10,442 kidneys.

The procurement of the pancreas must be performed in conjunction with the liver procurement, carefully delineating the blood supply to the liver to ensure that both organs can be removed and transplanted. In the vast majority of cases, blood supply should not preclude the transplantation of both organs. Initial dissection involves division of the gastrocolic ligament to expose the anterior surface of the pancreas. Visual inspection is considered to be an important element to the pancreas procurement process, because this is an opportunity to assess the organ for the presence of infiltrating fat or hematoma that might preclude transplantation. The portal triad is carefully dissected, with division of the common bile duct and the gastroduodenal artery. In addition, the common hepatic, left gastric, and splenic arteries are all dissected free of surrounding lymphatic tissues. Further dissection of the pancreas includes a Kocher maneuver, to mobilize the head, and division of the lienophrenic and lienocolic ligaments, to mobilize the body and tail. Of note, the spleen is left in continuity with the pancreas to serve as a handle to minimize manipulation of the gland itself. The duodenum is decontaminated by flushing povidone-iodine antifungal/antibiotic solution through a nasoduodenal tube.

After the patient has been systemically heparinized, the abdominal aorta is ligated at its bifurcation and is cannulated in a retrograde direction for perfusion. The liver team may also elect to place a cannula for portal perfusion through the inferior mesenteric vein. The supraceliac aorta is cross-clamped, the vena cava is vented, and the abdominal organs are flushed in situ with preservation solution, typically University of Wisconsin (UW) solution, at 4° C. In addition, topical cooling is also used.

Once the organs have been adequately flushed, the liver and pancreas are either separated in situ and removed individually from the donor, or both organs are removed en bloc and divided at the back bench. In the in situ separation situation, the liver is removed first by dividing the portal vein 1 cm cephalad to the superior margin of the pancreatic head, approximately the level of the coronary vein, and dividing the splenic artery about 3 mm beyond its origin, thus preserving the entire celiac axis with the liver.

Next, removal of the pancreas proceeds. The proximal duodenum just beyond the pylorus and the distal duodenum are divided with gastrointestinal anastomosis stapler. The small bowel mesentery that lies inferior to the pancreas is divided, and the superior mesenteric artery is divided at its origin from the aorta. Long segments of donor iliac vessels are removed to use for vascular reconstruction during back-bench preparation of the pancreas.

Back-Table Preparation of the Pancreas

Relative to other solid organs, the pancreas requires extensive preparation prior to implantation into the recipient. This back-bench preparation is performed in ice-cold preservation solution to minimize any further ischemic injury to the organ. The duodenum is often shortened with a gastrointestinal anastomosis stapler, being careful to exclude any gastric tissue and also being careful not to compromise the opening of the ampulla of Vater. The division of the small bowel mesentery in the donor is shortened by firing a stapler across this mesentery and then reinforcing the staple line with a running vascular suture. The spleen is removed by dividing the vessels in the splenic hilum, being careful not to injure the tail of the gland. Finally, the arterial inflow to the graft must be reconstructed, because the organ has two major sources of blood supply that are not in continuity: the splenic artery supplies the body and tail, and branches of the superior mesenteric artery supply the head.

In most instances, arterial reconstruction can be performed using the donor iliac artery as a bifurcated Y graft. The internal iliac artery is joined to the splenic artery, and the external iliac artery is joined to the superior mesenteric artery. The common iliac artery of the donor Y graft can then be anastomosed to the recipient iliac artery, serving as the arterial inflow to the pancreas. In some instances, it is necessary to create a portal vein extension graft on the back bench using donor iliac vein; however, this technique is avoided if possible, because it may increase the risk of venous thrombosis of the graft.

Recipient Operation

The techniques used for transplanting the pancreas have changed dramatically over the past few decades. Partial segmental grafts that contained only the body and tail were once common; however, this technique is rarely used today. Exocrine secretions were previously managed by pancreatic duct ligation or by injection of a polymer that would cause duct obliteration, however, now exocrine secretions are handled by internal drainage.

In two areas of pancreas transplantation, current-day techniques differ: in the drainage of exocrine secretions and the venous drainage of the graft. Exocrine drainage is performed either via the intestinal tract or via the urinary tract. Throughout most of the 1980s and 1990s, drainage of the pancreatic secretions into the recipient bladder was the most common form of exocrine drainage. This technique is convenient for monitoring organ function by measurement of amylase levels in the urine; however, problems with hematuria, cystitis, bicarbonate loss, and dehydration are all associated with bladder drainage. These complications necessitate surgical revision to enteric drainage in up to 20% of bladder-drained pancreas recipients (Stratta, 2005). Based on these issues, and on the lower rejection rates observed with newer immunosuppressive medications, the majority of transplant centers now perform enteric drainage of the exocrine secretions. This enteric drainage is either directly into a loop of jejunum in a side-to-side fashion or into a Roux-en-Y limb of jejunum.

The venous drainage of the graft is either to the systemic circulation, via an iliac vein or the inferior vena cava, or to the portal circulation. Portal venous drainage has the theoretic advantage of delivering insulin in a more physiologic manner, because insulin undergoes a “first pass” through the liver, and the hyperinsulinemia that results from systemic drainage is avoided (Gaber et al, 1995). There is also an immunologic advantage of portal drainage that has been observed in several experimental studies, in which the delivery of foreign antigen via the portal system results in diminished immune responses. Despite these potential advantages, no demonstrable difference has been found in outcomes between human transplants drained by the portal vein and those drained by the systemic vein, and systemic venous drainage is how the majority of pancreas transplantations are performed.

There are two common locations in the abdomen where the transplant is placed based on the type of venous drainage planned: either in the pelvis, most commonly the right side, for systemic venous drainage or in the midabdomen for portal venous drainage. When the graft is placed in the pelvis, the donor portal vein is anastomosed to the external iliac vein, the common iliac vein, or the inferior vena cava. In this pelvic position, the graft is oriented with the duodenum in an inferior position, if bladder drainage is planned (Fig. 101.1A), or with the duodenum in either the superior (Fig. 101.1B) or inferior position if enteric drainage is planned. Alternatively, for portal venous drainage, the pancreas is placed in the mid-abdomen below the transverse colon with the duodenum oriented superiorly. The portal vein of the pancreas is anastomosed to a major branch of the superior mesenteric vein, found in the small intestine mesentery, in an end-to-side fashion (Fig. 101.2). Enteric drainage for exocrine secretions must be used with the portal venous drainage technique. With either venous drainage technique, the donor arterial conduit to the pancreas graft is anastomosed in an end-to-side manner to the recipient common or external iliac artery.

FIGURE 101.1 Whole-organ pancreas transplantation with systemic venous drainage. The pancreas is placed in the right pelvis with anastomosis of the donor iliac artery Y graft to the recipient iliac artery to provide arterial inflow to the graft and anastomosis of the donor portal vein to the recipient iliac vein for venous drainage. Simultaneous kidney transplantation can be performed using the left-sided iliac vessels. A, The duodenum is oriented in the inferior direction to allow anastomosis to the bladder (shown) or to the small intestine (not shown). B, Alternatively, the duodenum is oriented superiorly for anastomosis to the intestine.

FIGURE 101.2 Whole-organ pancreas transplantation with portal venous drainage. The pancreas is placed in the mid-abdomen with anastomosis of the donor portal vein to a major branch of the superior mesenteric vein and anastomosis of the donor iliac artery Y graft to the recipient iliac artery. The duodenum is oriented in the superior direction for anastomosis to the intestine. A simultaneous kidney transplant can be performed using either the left (shown) or right iliac vessels.

Complications

The major complications following pancreas transplantation are often technical in nature. Pancreas graft thrombosis, arterial or venous, is more frequent after pancreas transplantation than after other solid-organ transplants, and the reported incidence is approximately 10%. Thrombosis usually occurs within the first week following transplantation and likely reflects the relatively low blood flow through the organ. In most instances of thrombosis, graft removal is necessary (Humar et al, 2000). Early pancreatitis occurs in 10% to 20% of cases and is largely a reflection of ischemic damage to the gland during preservation and reperfusion injury. Hyperamylasemia and graft edema are characteristic, and graft pancreatitis is usually treated nonsurgically with octreotide. Leakage at the site of pancreatic exocrine drainage is another early complication, with management often dictated by the method of drainage. Bladder-drained transplants with a small leak at the duodenocystostomy can often be managed by Foley catheter drainage of the bladder, allowing the site of leakage to heal over time. Enteric-drained transplants with a leak at the duodenojejunostomy will often result in peritonitis and usually require operative intervention to control the leak.

Rejection following pancreas transplantation was once a very common occurrence. Pancreas rejection is often difficult to diagnose, and a variety of indicators are used to help make the diagnosis. These include tenderness over the area of the pancreas allograft, increased serum amylase and lipase, decreased urinary amylase excretion (if bladder drainage is used), biopsy of the pancreas, and hyperglycemia. It is noteworthy that hyperglycemia is a late indicator of rejection, and the pancreas is often difficult to salvage once hyperglycemia has occurred. If SPK transplantation is performed, renal allograft dysfunction can often assist with the diagnosis of pancreatic rejection, because the rejection process is often occurring in both organs. As a result of the high incidence of rejection and the difficulty in making the diagnosis, pancreas transplant recipients usually receive potent induction immunosuppression with a T–cell depleting agent and maintenance therapy with tacrolimus, mycophenolate mofetil, and corticosteroids.

The total amount of immunosuppression pancreas recipients receive is among the highest of any solid-organ transplantation. As a result, they are more susceptible to the complications of immunosuppressive therapy. These complications include infection with opportunistic bacteria, viruses, and fungi; malignancy; gastrointestinal complications; and others. The high incidence of these complications makes effective prophylaxis strategies important.

Results

Patient and graft survival following pancreas transplantation have improved significantly in recent years. Depending upon the type of transplantation—SPK, PAK, or PTA—patient survival is approximately 96% to 98% at 1 year and 85% to 89% at 5 years after transplantation. Graft survival varies according to the type of transplantation performed (Fig. 101.3). For SPK recipients, 1-, 5-, and 10-year pancreas graft survival rates are 85%, 73%, and 55%, respectively. For PAK recipients, the respective rates are 80%, 53%, and 37%. Finally, for PTA recipients, rates of pancreas graft survival are 76%, 52%, and 35%, respectively (Axelrod et al, 2010).

FIGURE 101.3 Pancreas graft survival rates (unadjusted) at 1 year, 3 years, 5 years, and 10 years following whole-organ transplantation by transplant category.

(Modified from Axelrod DA, et al, 2010: Kidney and pancreas transplantation in the United States, 1999-2008: the changing face of living donation. Am J Transplant 10:987-1002, Copyright 2010 Blackwell Publishing, with permission.)

The impact of a successful pancreas transplantation on the complications associated with diabetes mellitus is debated. Because the successful transplantation restores euglycemia and normal hemoglobin A1c levels, most proponents argue that diabetic complications should cease and perhaps reverse. Neuropathy appears to stabilize and even slowly improve following pancreas transplantation, whereas retinopathy progression is slowed after several years of graft function. The development of diabetic nephropathy in the transplanted kidney of SPK and early PAK recipients appears to be prevented by successful pancreas transplantation. In PTA recipients, diabetic nephropathy appears to stabilize following transplantation (Fioretto et al, 1998); however, the renal benefit is likely outweighed by the detriment to renal function caused by the immunosuppressive agents, specifically tacrolimus and cyclosporine.

Risk/Benefit Considerations

As outlined above, considerable evidence suggests that successful pancreas transplantation can retard, and in some cases reverse, the secondary complications of diabetes. Despite this evidence, the procedure is considered appropriate for only a subset of patients with type 1 diabetes because of its highly invasive nature, compared with insulin therapy, and the potential for significant surgical complications; thus considerations of risk to benefit are of utmost importance in pancreas transplantation and play an essential role in both the donor and recipient selection processes.

One analysis evaluated the survival benefit of pancreas transplantation by examining a large transplant database to compare the survival of pancreas recipients versus those patients listed for transplantation but who did not receive a transplant. Using this study design, the two patient groups were ensured to be as similar as possible. Because the indications, risks, and benefits differ, depending on whether the patient had underlying renal failure, the analysis was stratified based on whether the patient was listed for or transplanted with an SPK, PAK, or PTA (Venstrom et al, 2003). This analysis revealed a marked survival benefit for those patients who received a combined pancreas and kidney transplant; a 57% reduction in mortality rate was observed over 4 years compared with similar patients who remained on the waiting list.

In contrast to the clear survival benefit of SPK transplantation, the PAK and PTA recipients experienced an increased mortality following transplantation. During the 4-year follow-up period, PAK recipients experienced a 42% increase in mortality rate, and PTA recipients experienced a 57% increase in mortality rate compared with the comparable group of patients who remained on the waiting list. A similar analysis was reported that challenged the survival disadvantage of PAK and PTA (Gruessner et al, 2004a). Their analysis found again, as expected, that the SPK procedure conferred a marked survival advantage; however, unlike the Venstrom study, no survival disadvantage was evident in PAK or PTA transplant recipients.

There are two important caveats to these risk/benefit studies by Venstrom and Gruessner: The first is that the impact of pancreas transplantation on quality of life was not evaluated in these nonrandomized retrospective analyses and would likely heavily favor the group with successful PAK and PTA recipients, but whether this factor would outweigh any survival disadvantage in the short-term, and the significant cost of these complex procedures, has not been evaluated. A second critical outcome measure that is lacking from these studies is the long-term impact of successful pancreas transplantation on the secondary complications of diabetes. As noted above, evidence suggests that diabetic nephropathy, neuropathy, and retinopathy may be positively impacted by the long-term restoration of normoglycemia conferred by transplantation; however, these studies have not included well-controlled trials with large numbers of patients.

These results substantiate the marked clinical benefit observed when caring for patients who undergo the SPK procedure (Rayhill et al, 2000). Importantly, the relative contribution of the transplanted kidney versus the transplanted pancreas to overall patient survival has not been differentiated in a controlled study. This issue is significant given the marked patient survival benefit that has been attributed to kidney transplantation (Wolfe et al, 1999).

A recent retrospective database analysis lends important insight into the role that pancreas graft function contributes towards overall patient survival (Weiss et al, 2009). This analysis examined outcomes in individuals on the SPK waiting list who underwent either an SPK or transplantation of the kidney alone. Recipients of an SPK transplant who had pancreas allograft function 1 year after transplantation had significantly greater patient survival at 7 years (89%) compared with SPK recipients who lost their pancreas transplant within the first year after transplantation (74%) or to those recipients that underwent a kidney transplant alone from either a living donor (80%) or a deceased donor (65%).

Although the survival benefit for SPK recipients is clear, the survival benefit for PAK and PTA recipients has not been conclusively demonstrated (Sutherland et al, 2009). The same analysis by Weiss demonstrates improved patient survival at 5 years for diabetic recipients who underwent a PAK compared with kidney transplantation alone, although this did not reach statistical significance and was subject to selection bias. A recent single-center report demonstrates excellent outcomes 3 years after PAK, in which both patient and graft survival are similar to SPK recipients at that center (Fridell et al, 2009). In this report, PAK recipients had a 90% pancreas allograft, 92% kidney allograft, and a 92% patient survival 3 years after PAK transplantation, compared with 83% pancreas allograft, 86% kidney allograft, and 88% patient survival in SPK recipients. If these excellent PAK outcomes persist with further follow-up and can be replicated at other centers, it is reasonable to assume that a survival benefit to PAK could be demonstrated in future data analyses.

Currently, the most meaningful conclusion that can be drawn is that PAK transplantations should be performed in carefully selected recipients with well-preserved kidney allograft function, and these procedures should be performed at transplant centers with excellent outcomes. The field remains dynamic due to improvements in surgical technique and perioperative care and the frequent introduction of new immunosuppressive agents with more favorable safety profiles; however, improving outcomes with pancreatic islet transplantation may ultimately challenge the need to perform whole-organ pancreas transplantation for the treatment of type 1 diabetes.

History and Early Results

The original descriptions in rodents of successful islet isolation (Lacy & Kostianovsky, 1967) and subsequent transplantation (Ballinger & Lacy, 1972) caused great excitement in the medical community, in the hope that this type of therapy might be applied to patients with type 1 diabetes. Although the ability to obtain normal glucose control in diabetic rodents with an islet transplant was described decades ago (Reckard et al, 1973), translating this success to humans has been more difficult. Various centers worldwide have attempted human islet transplantation, starting in 1974, with 445 recipients receiving islets between 1974 and 2000. The vast majority of these recipients also received a kidney transplant, either prior to or at the same time as the islet transplant. Analysis of the reported cases between 1990 and 2000 demonstrated that only 19% of patients were off insulin for more than 1 week, and at 1-year follow-up, only 11% of recipients were insulin independent (Brendel et al, 2001).

These discouraging results were attributed to a number of factors, including the possibility that recurrent autoimmunity was causing progressive islet damage following transplantation, because the diabetes in these recipients was the autoimmune form. Support for this hypothesis came from a series of patients at the University of Pittsburgh who underwent upper abdominal exenteration, including total pancreatectomy, followed by liver and islet allotransplantation. In this group of 11 patients, who did not have autoimmune diabetes, 6 (55%) exhibited sustained insulin independence, a success rate far greater than what had been achieved in type 1 diabetic recipients (Carroll et al, 1995).

Islet Autotransplantation

Autotransplantation of pancreatic islets was developed for patients with chronic pancreatitis, in whom total pancreatectomy once provided the best option for treatment. Surgically induced diabetes is frequently brittle, characterized by an absence of both insulin and counterregulatory hormones, with a high rate of long-term morbidity. Treatment involves processing of the surgically removed pancreas in the islet isolation laboratory, followed by infusion of either purified islets or crude pancreatic digest into the liver via the portal vein.

Although the overall experience is limited, the success rate in recipients of autotransplantation has been far greater than in those who received allogeneic islets for the treatment of type 1 diabetes. The largest reported series is from the University of Minnesota, where between the years 1977 and 2007, 173 patients with chronic pancreatitis underwent total, near total (95%), or completion pancreatectomy with islet autotransplantation (Sutherland et al, 2008). These investigators achieved a 65% rate of graft function and a 32% rate of insulin independence, rates of success that correlate closely with the average yield of islets from the removed pancreatic tissue following digestion and processing. This group previously reported higher rates of successful islet autotransplantation in patients who had no prior surgery, or only a prior Whipple procedure, whereas those that had a prior Puestow procedure or distal pancreatectomy had lower chances of success (Gruessner et al, 2004b). Similar results have been reported by another group in a series of 22 patients, where 68% of patients showed acceptable graft function, and 41% were insulin independent (Rodriguez Rilo et al, 2003). Interestingly, both of these groups infused unpurified pancreatic digests, where the islets had not undergone the additional purification step performed during deceased-donor pancreas processing for allotransplantation.

The Edmonton Protocol and Beyond

Interest in islet transplantation was piqued by the Edmonton investigators reporting that, for the first time, diabetes in patients could be consistently reversed by transplantation of isolated pancreatic islets (Shapiro et al, 2000). The Edmonton approach relied on a novel immunosuppression regimen that completely avoided steroids and that utilized a unique combination of induction therapy with an anti–interleukin (IL)-2 receptor antibody and maintenance therapy with sirolimus and tacrolimus. The rationale for this combination was based on the desire to avoid agents with known β-cell toxicity. Although this particular drug combination was likely an important factor in the trial’s success, perhaps of even greater impact was the large number of islets that patients received; these investigators infused a larger number of islets than were administered in prior islet transplantation trials, using two or three infusions per patient, acquired from multiple deceased donors.

Whereas previous trials in the late 1990s had often achieved partial success with single infusions of islets, as evidenced by a reduced insulin requirement and increased C-peptide levels, in no trial had the further step been taken to retransplant such patients to increase the net islet mass (Hering & Ricordi, 1999). In the Edmonton series, patients who showed evidence of partial success from the first infusion received an additional infusion, sometimes two, until they had accumulated a sufficient islet mass to gain insulin independence; in general, this occurred when a threshold of 8000 to 10,000 islets per kilogram of recipient body weight had been infused.

After the landmark report by the Edmonton investigators, almost 500 recipients received islet transplants at one of approximately 40 centers worldwide (Harlan et al, 2009). In the first concerted effort to replicate the provocative results of the Edmonton Protocol, a multicenter trial sponsored by the Immune Tolerance Network enlisted 10 centers in North America and Europe to each perform four islet transplants using the Edmonton regimen (Shapiro et al, 2006). Despite a diligent attempt to ensure uniformity in technique between centers, the rate of success varied dramatically, depending on the extent of experience at the transplanting site: at the three most experienced centers, reversal of hyperglycemia was routinely achieved; however, less experienced teams gained insulin independence in only a fraction of cases (approximately 20%). Thus, although these results confirmed the efficacy of Edmonton’s approach, they also revealed the exacting nature of the technique and demonstrated the inherent difficulty in replicating an identical protocol at all centers.

A number of single-center reports have now validated the Edmonton Protocol (Goss et al, 2002; Markmann et al, 2003), and although this protocol represents a tremendous leap forward for the field, the Edmonton results also delineate the major problem that remains: the need for islets procured from multiple deceased donors to gain insulin independence in a single recipient. This requirement markedly increases costs and represents a distinct disadvantage when compared with whole-organ transplantation; however, it should be remembered that, thus far, the pancreata utilized for isolated islet transplantation have only been organs that were unsuitable for whole-organ transplantation. It seems reasonable to assume that these organs are inferior to those used in whole-organ transplantation, and that improved results may be achieved if islet transplantation candidates had access to the best organs.

The need for such a large number of islets to achieve insulin independence in clinical transplantation was unexpected in light of the finding that as little as 10% of the total islet mass can maintain normoglycemia following partial pancreatectomy. Because the total islet mass in normal individuals is estimated at 1 to 2 million, it would follow that transplanting only 200,000 islets should regain normoglycemia in the majority of diabetic recipients; instead, accumulating evidence from the islet transplant studies by Edmonton and others suggest that for an average individual, 600 to 700,000 islets are required. The nearly two-thirds loss of potency after transplantation has not been fully explained in patients, but experimental data from small animal studies indicates that the majority of transplanted islets fail to engraft (Davalli et al, 1995). Understanding and overcoming the loss of islet mass during the engraftment phase has been an important focus of recent islet transplantation clinical trials.

Some centers have reported improved success at gaining insulin independence with islets from a single donor by refining the Edmonton approach (Hering et al, 2004, 2005; Markmann et al, 2003). The most successful reports are from the University of Minnesota, where investigators have used potent induction immunosuppression and high-quality islets. In their first study, four of six consecutive patients were rendered insulin independent with single-donor islet infusions using induction therapy with the anti-CD3 monoclonal antibody hOKT3γ1 (alanine-alanine) (Hering et al, 2004). This was combined with the selection of relatively larger donors and somewhat smaller recipients with low daily insulin requirements, thus maximizing the islet mass delivered per recipient body weight.

An even more impressive result was reported by the same group using polyclonal antithymocyte globulin (Thymoglobulin) for induction therapy; in a series of 10 consecutive patients treated with Thymoglobulin induction therapy, 9 were rendered insulin independent with a single infusion of islets (Hering et al, 2005). The 1 patient who did not become insulin independent still had a marked reduction in daily insulin requirement. Collectively, these studies suggest that insulin independence can be achieved consistently with single-donor islet infusions under the correct circumstances. Overcoming the single-donor hurdle will help in the establishment of isolated islet transplantation as a more widely applicable therapy for patients with type 1 diabetes.

A second, more daunting obstacle is the realization that the durability of function after islet transplantation is less than whole-organ pancreas grafts. Whereas it is clear that very experienced islet centers can gain insulin independence in most islet recipients, and that short-term (1-year) insulin-free survival may be close to that seen with whole-organ transplants, the intermediate (2- to 4-year) insulin-free survival is less encouraging. In a report by Edmonton investigators, with follow-up on patients in their ongoing series transplants of islets alone (Ryan et al, 2005), fewer than 25% of patients remained free from exogenous insulin administration at 3 years (Fig. 101.4B). Although a somewhat disappointing result, more encouraging was the observation that more than 85% of these patients had evidence of some islet graft function by C-peptide measurement (Fig. 101.4A). Similar findings of declining islet function over time have been reported by the multicenter Collaborative Islet Transplant Registry: of the recipients who achieved insulin independence, approximately 50% were back on insulin 2 years later (Alejandro et al, 2008).

FIGURE 101.4 Islet graft survival rates following isolated islet transplantation at the University of Alberta, Edmonton (Ryan et al, 2005). A, Survival analysis for islet function as documented by C-peptide secretion over time, demonstrating 85% graft survival at 3 years. B, Survival analysis for insulin independence over time, demonstrating that fewer than 25% of patients are free from exogenous insulin therapy at 3 years.

(From Ryan EA, et al, 2005: Five-year follow-up after clinical islet transplantation. Diabetes 54:2060-2069. Copyright 2005 American Diabetes Association.)

Despite the eventual need for most recipients to resume doses of insulin, the partial graft function that persists is relevant, because many patients who were transplanted with the indication of hypoglycemia unawareness remain cured of this devastating complication. It is also evident that a partially functioning graft can significantly improve glycemic control as measured by serial hemoglobin A1c determinations. A small group of islet recipients who have resumed insulin therapy have been given supplemental islet infusions several years after the initial islet transplant and have regained durable insulin independence in most cases (Koh et al, 2010).

The mechanisms responsible for the gradual loss of islet function remain unclear. One hypothesis is that inadequate control of the deleterious alloimmune and autoimmune responses compromise islet function over time. The absence of a reliable measure of an antiislet immune reaction and the difficulty in obtaining biopsy tissue for pathologic diagnosis contribute to the complexity in clarifying this issue. One study reported durable insulin independence (>3 years) in islet recipients when more potent induction immunosuppression was used, compared with what was used in the Edmonton Protocol (Bellin et al, 2008). An alternative explanation is that the currently used immunosuppressive agents are toxic to the islets over time. This explanation may fit better with the pace of graft dysfunction and the fact that graft function often stabilizes without antirejection therapy. In addition, studies have documented very high immunosuppressive drug levels in the portal circulation, an expected consequence of the drugs being administered orally (Desai et al, 2003). The role of immune responses and medication toxicity to the loss of islet graft function should be clarified by the application of newer immunosuppressive agents in the next generation of islet transplantation studies.

Complications and Risk/Benefit Considerations

The demonstration that the procedure is relatively safe for the recipient has been important in the early trials of isolated islet transplantation. The most feared potential complications include bleeding and portal vein thrombosis. The Edmonton experience in 65 patients demonstrates that bleeding occurred in more than 20% of recipients, most of whom required either transfusion or surgical intervention. Main portal vein thrombosis, the most potentially troublesome complication of portal islet infusion, has fortunately been rare (<1%); however, in the Edmonton experience, 7% of recipients had evidence of segmental portal vein thrombus that resolved without sequelae (Ryan et al, 2005).

The risk of bleeding and thrombosis may be related in part to the technique used to perform the infusion. The procedure can be performed by an interventional radiologist, gaining access to the portal vein via a percutaneously placed transhepatic catheter and releasing the islets into the portal system (Froud et al, 2004). Alternatively, the procedure can be performed via a small laparotomy to gain access to a small mesenteric vein, through which a catheter can be advanced into the main portal vein for islet infusion (Gaber et al, 2004). The small vein is then ligated, eliminating the potential for hemorrhage. Thus the operative approach allows anticoagulation to be administered without the risk of bleeding from a puncture site in the hepatic parenchyma, although it usually requires a general anesthetic.

The relative safety advantage of islet transplantation over whole-organ pancreas transplantation was clearly evident in a comparison of the two approaches (Frank et al, 2004). In this study, major complications that included bleeding that required transfusion and reoperation were more common in the whole-organ group, although a number of minor complications were frequent in the islet transplant group. Some of these minor complications included immunosuppression-related toxicity (mouth ulcerations, edema), periportal hepatic steatosis, and mild liver function test abnormalities. A comparison of the major advantages and disadvantages of whole-organ pancreas and isolated islet transplantation can be seen in Table 101.1.

Table 101.1 Comparison of Whole Pancreas and Islet Transplantation

SPK, simultaneous pancreas and kidney transplantation

Islet transplantation also incurs the risk of infusion-related infection. This complication is minimized by screening potential donors for infection, administration of prophylactic antibiotics to the recipient, and by assessment of the islet preparation for bacterial contamination by microscopic examination and endotoxin testing. As with all forms of transplantation, the need for chronic immunosuppression carries with it a risk of opportunistic infection and a small but discernible risk for the development of malignancy, especially lymphoma and skin cancer. Similarly, exposure of the recipient to allogeneic tissue may induce an immune response in the recipient and generate cytotoxic antibodies against the HLA antigens of the donor. This exposure may be especially detrimental in islet recipients who receive transplants from multiple donors, thereby broadening their antigenic exposure. The presence of anti-HLA antibodies could hamper future attempts at islet transplantation or transplantation of lifesaving organs such as the kidney.

One of the most common and concerning potential complications of islet transplantation using the Edmonton protocol is calcineurin inhibitor–associated nephrotoxicity. Chronic renal failure in the setting of nonrenal solid-organ transplantation is well described, with an approximate 20% incidence at 10 years (Ojo et al, 2003). Similar results in islet transplant recipients would represent a particularly morbid complication given the well-established detrimental survival impact of renal failure in the diabetic population (Allen & Walker, 2003). A recent report of renal function in islet transplant recipients indicates that renal function remained stable in an appropriately selected and managed cohort of patients despite their being placed on potentially nephrotoxic immunosuppressive medications (Leitao et al, 2009). This study demonstrates that renal function can be adequately preserved in islet recipients under the correct circumstances.

Because of the side effects of chronic immunosuppression, the benefit of isolated islet transplantation, like that of isolated pancreas transplantation, may be for only select individuals. For this reason, most islet trials to date have enrolled type 1 diabetic participants with the most labile and dangerous form of the disease, specifically, diabetics attended by frequent episodes of hypoglycemia unawareness; patients are thought to be at greatest risk of severe morbidity or death from these events. Islet transplantation, even if only partially successful, has been universally found to be highly effective in reducing the frequency of these hypoglycemic episodes (Alejandro et al, 2008).

An alternative strategy has been to select recipients who are already on immunosuppressive therapy to support another organ transplant, such as a kidney (Kaufman et al, 2002). In this situation, the additional risk to the recipient relates mainly to the islet infusion procedure itself, which should be quite safe. In this setting, recipients usually lack severe hypoglycemia unawareness, although glucose control is usually suboptimal, it can be expected to improve following islet transplantation. An additional major concern in patients receiving an islet after kidney (IAK) transplantation is the function of the lifesaving renal allograft that is already in place. The experience to date suggests that IAK transplantation can be performed without jeopardizing the kidney graft. An important study in this area, although not conducted in a randomized manner, found evidence that IAK transplantation is not only safe, but it likely has a beneficial impact on renal allograft function (Fiorina et al, 2003). They studied 36 IAK recipients and stratified them based on the success of the islet transplant at 1 year, defined by continued C-peptide production. There were 24 successful cases and 12 unsuccessful cases; the successful IAK recipients demonstrated superior renal graft survival up to 7 years after islet transplantation, and the unsuccessful islet recipients developed increased microalbuminuria over time, likely reflecting diabetic damage in the transplanted kidney.

In addition, successful IAK recipients had improved cardiovascular function in terms of ejection fraction and degree of diastolic dysfunction when compared with kidney-only recipients (Fiorina et al, 2005). Finally, an improvement in patient survival was noted in successful IAK recipients compared with recipients with failed IAK transplantation and with diabetic recipients who had undergone kidney transplantation alone. A major caveat to these findings is that they are derived from a small number of IAK patients, therefore they must be substantiated in larger cohorts. In addition, these results are not based on recipients randomized to either treatment arm, thus any benefit in secondary complications attributable to IAK transplantation is subject to selection bias and requires further investigation in properly controlled studies.

Future Directions in Islet Transplantation

There are many areas of active investigation in clinical and experimental islet transplantation, including an important set of collaborative studies being supported by the National Institutes of Health. These trials have the opportunity to secure a place for islet transplantation as a standard therapeutic modality. For isolated islet transplantation to gain equal footing with or to surpass whole-organ pancreas transplantation as the preferred therapy for patients with diabetes, a number of conditions will need to be met. First, improvements must occur so that reversal of diabetes is readily accomplished with the islets from a single donor in the majority of cases. This will require not only advances in isolation techniques that allow greater recovery of healthy islet tissue, it will also require a greater understanding of the events early after transplantation that are responsible for engraftment of only a fraction of the delivered islet mass (Harlan et al, 2009). Second, clarification is needed of the immunologic and physiologic mechanisms that contribute to islet dysfunction over time that lead to the need for reinstitution of low doses of exogenous insulin in a majority of islet transplant subjects.

Finally, should the hurdles mentioned above be overcome, the reliance on deceased donors as the sole source of islets will be insufficient to treat the large number of patients with type 1 diabetes who could benefit from islet transplantation. Living donors are one alternate source of islet tissue as shown in a report of living donor islet transplantation from Japan (Matsumoto et al, 2005); however, the potential for donor morbidity makes this approach highly controversial, until the long-term durability and efficacy of islet transplantation has been firmly established. Ultimately, derivation of β-cells from xenogeneic or stem cell sources promises to provide a limitless supply of transplantable β-cells. The ongoing trials of islet transplantation will provide a critical foundation for these future therapies by defining the optimal site for implantation, the best means to monitor survival, and the most conducive immunosuppression that avoids autoimmune and alloimmune damage without pharmacologic β-cell toxicity.