Pancreas Allograft Biopsies in the Management of Pancreas Transplant Recipients: Histopathologic Review and Clinical Correlations

There is evidence that pancreas transplantation halts the progression and, to an extent, partially reverses the microvascular and macrovascular complications of diabetes.1,2 Clinical studies have shown improvement of cardiac function, neuropathy, and autonomic dysfunction in patients who received kidney-pancreas allografts compared with those who only underwent a kidney transplant.3–6 

The most common form of pancreas transplantation is a simultaneous kidney-pancreas transplantation (SPK), followed by a pancreas after a previous kidney transplantation (PAK). Pancreas alone transplantation (PTA) is done with the least frequency, being offered to a selected subgroup of nonuremic diabetics who experience recurring or severe acute metabolic complications. According to the United Network for Organ Sharing registry, approximately 900 SPKs, 400 PAKs, and more than 100 PTAs were performed in the United States in the year 2004.7 The deceased donor pancreas allograft is procured with an attached segment of the duodenum and the spleen. In the early era of pancreas transplantation, the duodenal segment was anastomosed to the urinary bladder to drain the exocrine secretions, while the blood supply was reestablished from the iliac artery and drained into the iliac vein (ie, systemic bladder drainage). In more recent years, several modifications of the procedure have been established. Enteric drainage of the pancreas (draining the pancreas exocrine secretion to the small intestine) in lieu of bladder drainage became popular, comprising up to 85% of all transplanted pancreas allografts in the year 2004. Venous drainage of the pancreas is either systemic or into the portal vein (or one of its tributaries), simulating the natural circulation. Technical details have been described elsewhere and are beyond the scope of this review.8 It is important, however, for pathologists involved in the care of the transplant patients to be aware of these techniques, since their understanding would allow for comprehensive interpretation of the laboratory tests in conditions of graft dysfunction.

Patient and graft survival statistics following pancreas transplantation have been gradually improving during the past 2 decades, with 1-year survival rates for pancreas allografts performed in the United States between 2000 and 2005 reaching 87.9% for SPK, 80.8% for PAK, and 77.2% for PTA.9 The increase in the pancreas allograft survival rates in recent years is the result of improved surgical techniques, adoption of more potent modern immunosuppression, careful management of patients, and aggressive monitoring of graft function after transplantation. Pancreas allografts are monitored by laboratory tests that assess the adequacy of the endocrine and exocrine functions, and by radiologic techniques that determine blood flow into the graft. The specificity of these markers in establishing the diagnosis of acute rejection is suboptimal at best. Clinical features for the diagnosis of acute rejection, such as elevated serum amylase or lipase levels, 50% decrease in urinary amylase for bladder-drained grafts, unexplained fever, and hyperglycemia were associated with a positive predictive value of only 75%.10 Urinary hypoamylasemia in instances of bladder-drained grafts had a false-positive predictive value for pancreas rejection in up to 53% of cases.11 Hyperglycemia could result from multiple etiologies that could be extremely difficult to discern by clinical criteria alone. While hyperglycemia could be encountered in acute rejection, it is usually of later onset and indicates a severe acute rejection. Posttransplantation diabetes is more frequently caused by calcineurin inhibitor islet toxicity, use of steroids, and increased peripheral resistance to insulin (type 2 diabetes mellitus). Progressive insulin dependency to control hyperglycemia is usually a sign of chronic rejection.

Pancreas allograft biopsy is the gold standard for the diagnosis of acute rejection. Kidney biopsy, for easier accessibility, has been used as a surrogate for a pancreas biopsy in patients with a SPK transplant and dysfunction. It should be noted, however, that there is some level of dysconcordance in the morphologic findings between kidney and pancreas biopsies that will justify biopsies of both organs when indicated. Timed biopsies per protocol have been used in some transplant centers for early surveillance of the pancreas allografts, particularly in the management of isolated pancreas allografts (ie, pancreas after kidney or pancreas alone).12,13 

The standard procedure to obtain a pancreas allograft biopsy is the percutaneous needle biopsy technique. Biopsies are performed under local anesthesia, using an 18-guage needle, and are guided by either computerized tomography scan or ultrasound.14 Cystoscopically guided needle biopsies of bladder-drained pancreas allografts were used in the early days of pancreas transplantation but now have been almost completely replaced by percutaneous biopsies.15,16 Biopsies of the pancreas allografts also can be performed laparoscopically.17,18 Radiologic visualization of the pancreas is essential to enhance the yield of adequate tissue and to ensure the safety of the procedure. The pathway of the needle biopsy is carefully guided by ultrasound visualization in order to avoid puncturing blood vessels, pancreatic duct and other vital structures and to ensure passage into the substance of the pancreas. Biopsies of the pancreas are generally safe and have a reported success rate for adequate tissue yield of 90%.14,19 Complications of biopsy are usually minor and self-limiting. Transient rise of serum amylase levels, indicating postbiopsy pancreatitis, is the most common complication of the procedure, occurring with a frequency of 6%.20 More serious complications of the percutaneous biopsy technique are rare, as indicated by an analysis of a series of 426 pancreas allograft biopsies. In this large series, the incidence of bleeding or inadvertent biopsy of other organs (eg, bowel, liver, or kidney) was 2.8%. Surgical intervention was only required in 1.2% of the patients that underwent a biopsy.21 

There are no definitive guidelines as to the number of cores of tissue needed for an adequate biopsy or whether or not multiple-site biopsies should be obtained. In one center experience, there appeared to be some difference in the histology between biopsies of the head versus the tail of the pancreas.19 The pancreatic tail was better for identification of cellular rejection; however, the number of patients in that study was small and there was no later confirmation of these findings by others. The tissue sample(s) is collected in 10% buffered formalin, then processed for paraffin embedding by standard techniques. Routine light microscopy analysis then is performed on hematoxylin-eosin–stained multiple sections. Other stains that could be used for evaluation are Masson trichrome to determine the extent of fibrosis and chronicity in older grafts, and periodic acid–Schiff stain to highlight acinar architecture and to better define acinar inflammation in a manner similar to assessing tubular inflammation in kidney allograft biopsies.22 A sample of the pancreas composed of at least 2 lobules of acinar tissue associated with 2 or 3 septal areas is considered adequate for evaluation.23 

Careful evaluation of the anatomic compartments in the pancreas, including the acini, islets, septal regions with veins, arteries, and ducts, is important to establish the diagnosis of acute rejection and to determine the severity of the rejection acceding to the standardized semiquantitive guidelines described below. Similar to other allografts, acute rejection in the pancreas is characterized by inflammatory cellular infiltrates associated with features of target tissue injury. The inflammatory cells are typically a mix of mononuclear cells, T lymphocytes, and varying number of plasma cells and eosinophils (Figure 1). Immunoblasts and activated lymphocytes could be detected in appropriate histologic preparations and could provide evidence of immunologically mediated graft damage. Inflammation commences in the interlobular septae in mild rejection, and it then progressively involves the lobules and the islets in more the severe grades of rejection. Venous endothelitis is important to diagnose acute rejection, and a requirement for the diagnosis of very mild rejection, when inflammation is restricted to the interlobular septae. Mononuclear cells in the wall of the small veins accompanied by hypertrophy of the endothelial lining are key to the recognition of endothelitis (Figure 2). Mononuclear cells adhering to the endothelium are often found in endothelitis and should initiate careful search for inflammatory cells in the vessel wall. If pancreatic ducts are present in the biopsy, they frequently exhibit inflammatory cells in the wall, a pattern identified as ductulitis (Figure 3). Inflammatory cells permeate into the connective tissue surrounding the acini and produce acinar inflammation. Inflammation in the acinar tissue in its mild form is associated with inflammatory cells situated in between acinar cells or between acinar cells and basement membrane, whereas the severe form presents with destruction of acinar cells and confluent acinar and lobular necrosis. Apoptotic cells have been identified with high frequency in pancreas biopsies exhibiting acute rejection.24 The diagnostic and prognostic significance of this finding is unclear, but it provides useful information regarding the immune response during rejection of the pancreas. Intimal arteritis, defined by inflammation in the intima of arteries, is a specific feature for acute rejection, yet it is not frequently seen in the biopsies (Figure 4).25 Inflammation of the islets of Langerhan, islitis, is easily overlooked in the routine hematoxylin-eosin–stained sections, but it is better illustrated with immunohistochemical staining for T lymphocytes and mononuclear cells (Figure 5).

Drachenberg and colleagues23 developed a grading system for pancreas acute rejection (Table). In this system, four grades of acute rejection are defined (grades 2–5). Grade 1 in this schema, inflammation of undetermined significance, describes isolated inflammation in the septal areas without any of the other histologic features of rejection, and thus is inconclusive for the diagnosis of rejection. Higher grades of rejection are determined by the involvement of the arteries and by the diffuse and destructive nature of the inflammatory reaction. The histologic grades correlate with rejection severity as determined by clinical degree of graft dysfunction and treatment response.26 Diffuse moderate to severe acinar inflammation with necrosis, intimal arteritis, and venulitis correlates with graft failure.27 

Antibody-mediated acute rejection of pancreas allografts is poorly understood, mostly due to our inability to establish the diagnosis. Deposition of complement fragment C4d in allografts, in conjunction with evidence of tissue and microvascular injury, and/or detection of anti-HLA or anti–donor-specific antibodies, is diagnostic for antibody-mediated rejection. C4d is a complement split product generated during antibody-mediated immune response. It covalently binds to capillary endothelium at time of injury and probably remains bound for approximately 2 weeks after the insult. This property makes it an ideal marker for antibody-mediated rejection. Detection of C4d in tissue biopsies has been routinely performed on heart transplant biopsies and, more recently, kidney allograft biopsies.28,29 Indirect immunofluorescence staining technique using a monoclonal antibody against C4d has been the standard methodology for testing for C4d testing. This technique is less than ideal for pancreas allograft biopsies since it requires frozen tissue, and most pancreas biopsies are collected in formalin. Recently developed polycolonal antibodies against C4d allow for testing paraffin-embedded biopsies using immunohistochemical techniques (Figure 6).30,31 It is reasonable to expect that cases with antibody-mediated acute rejection of pancreas allografts will be diagnosed and better studied in the near future. Melcher et al32 reported a case of antibody-mediated rejection of the pancreas that developed 1 month after SPK. The patient had donor-specific HLA-DR alloantibodies along with strong C4d immunofluorescence staining in the periacinar capillaries without inflammation. Figure 6 shows C4d deposition in a solitary pancreas transplant detected on a 2-week posttransplantation surveillance biopsy. The patient was sensitized by two previous islet transplants and had detectable anti-HLA antibodies.

As the number of pancreas transplants surviving beyond the first year increases, chronic rejection is becoming increasingly common. The rate of pancreas loss to chronic rejection was 8.8% in 914 pancreas transplants followed for 3 years. Chronic rejection was highest in the PAK (11.6%) and PTA (11.3%) and lowest for SPK (3.7%).33 The histologic features of chronic rejection are the result of vascular sclerosis and progressive fibrosis, with loss of functioning structures. Variable mononuclear inflammatory infiltrates are detected in these late biopsies. Fibrosis causes expansion of the interlobular septae and erodes the periphery of the lobules and then the central zones, causing distortion of the architecture and loss of acini (Figure 7). The lobules become encircled by dense fibrous tissue and display a pattern similar to hepatic cirrhosis. Eventually, the graft becomes progressively replaced by fibrous tissue. Vascular lesions of chronic rejection are similar to those in other organs and include fibrocellular intimal hyperplasia with or without foam cells and lumen compromise (Figure 8).25 These vessels are not necessarily present in the needle biopsy and are more likely to be detected in pancreatectomy specimens.

A 3-tier grading system devised for chronic rejection has reasonable reproducibility.34 These chronic rejection grades correlate appropriately with the time elapsed since transplantation (ie, the higher the grade, the longer the posttransplantation period) but did not correlate with graft loss.

Thrombosis continues to be the major source of pancreas graft failure in the first 2 weeks after transplantation.35 The low flow conditions in the pancreas, aggravated by ischemia-reperfusion injury, preconditions the pancreas to thrombotic events. Reperfusion of the pancreas, even under optimal preservation, impairs the microcirculation and enhances leukocyte-endothelial interaction, which creates a thrombogenic and proinflammatory microenvironment.36–38 The diagnosis of graft thrombosis is suspected in the event of primary nonfunction or when recipients develop acute insulin dependency in the early days following transplantation. Loss of blood flow into the pancreas is confirmed by angiographic studies. Once thrombosis is diagnosed, pancreatectomy is indicated. Resected grafts usually show thrombi in arteries and veins, along with proportionate degree of coagulation and hemorrhagic necrosis. In rare cases, thrombi are either venous or arterial and may be focal in distribution.35 Thrombosis dmay also occur in the context of chronic rejection (Figure 9).

Pancreas allografts procured from older donors or obese donors have a higher incidence of delayed graft function and pancreatitis.38–40 It has been shown in experimental pancreas transplantation that ischemia-reperfusion injury causes edema, graft inflammation, and necrosis, with upregulation of adhesion molecules and increased apoptosis.41 Similar changes were reported in clinical transplantation.42 Biopsies of pancreas allografts taken at 30 to 90 minutes after transplantation show margination of leukocytes in blood vessels and in perivascular connective tissue, reduction in the density of zymogen granules, and large autophagolysosomes in the acinar cells. A few hours later, biopsies show defects in the cell membranes of acinar cells, leading to release of zymogen granules in the intercellular space, and different degrees of hemorrhage and fibrinoid necrosis. At time of laparotomy, coagulation and hemorrhagic necrosis become dominant.

Reflux pancreatitis is an infrequent complication, more apt to occur in segmental pancreas transplants and in bladder-drained pancreatic grafts. Reflux of urine into the pancreatic duct during voiding leads to a chemical pancreatitis characterized by edema and temporary dysfunction.43 Motility of sphincter of Oddi is dysregulated and could be the grounds for late graft pancreatitis.44 This form of pancreatitis is characterized by abrupt and transient hyperamylasemia and graft tenderness. It usually occurs a few months after transplantation and could be recurrent. Mental stress, alcohol consumption, infection, certain food, and mechanical pressure were described as possible precipitating factors for late pancreatitis.45 Differentiation between pancreatitis and rejection could be problematic and difficult; however, venulitis and a predominately monocytic graft cellular infiltrates favor rejection (Figure 10). In rare cases, biopsies may show localized fat necrosis and autodigestion of the pancreas without significant inflammation (Figure 11). This type of pancreatitis could result from mechanical injury to the pancreas or reflux of biliary secretions (Figure 12).

Cytomegalovirus allograft pancreatitis is a rare complication, characterized histologically by multifocal, predominately mononuclear acinar inflammation in association with the diagnostic cytomegalovirus-cytopathic changes. Confirmation of cytomegalovirus-pancreatitis could be achieved by immunohistochemistry. In rare instances, cytomegalovirus-pancreatitis could be associated with acute rejection.46 

Islet inflammation by mononuclear cells is seen typically in rejection, along with inflammation of the septal areas and acinar tissue. It is usually observed in higher grades of rejection (grades 3–5).47 The integrity of the islets is usually well preserved, and there is not apparent necrosis of islet cells, at least by light microscopic evaluation. The use of immunohistochemistry to identify monocytes and T cells may yield more islet inflammation during rejection. Inflammation of the islets also has been a feature of recurrent diabetes in an allograft. Mononuclear cellular inflammation of the islets in these cases was associated with selective loss of β cells. Immunohistochemical tissue examination using antibodies directed against β cells and α cells could be helpful in establishing the diagnosis of recurrent autoimmune diabetes mellitus in a patient and in differentiating it from islet inflammation secondary to rejection.25 Similarly to chronic pancreatitis, islets are more resilient and survive longer than the exocrine acinar tissue. A unique lesion has been identified in the islet of Langerhan during episodes of calcineuin inhibitor (tacromlimus or cyclosporine A) toxicity. The islet cells display vacuolar degeneration and significant swelling of the cytoplasm, with formation of pseudoglandular spaces within the islet (Figure 12). Clearing of the cytoplasm is more prominent in the center of the islets rather than in the peripheral zone. Occasionally, evidence of cellular necrosis, such as condensation of the cytoplasm, apoptosis, and cellular dropout, could be identified. During these toxicity episodes, staining for β cells reveals decreased density of the intracytoplasmic granules (Figure 13, A and B).48