Chapter 27. Autoimmune Pancreatitis

• Presentation with obstructive jaundice.
• Diffuse swelling and enlargement of the pancreas, especially the head, the latter mimicking carcinoma of the pancreas.
• Diffuse irregular narrowing of the pancreatic duct on ERCP or MRCP.
• Elevated serum IgG4 level.
• Extrapancreatic and intrahepatic bile duct strictures.
• Resolution or marked improvement in pancreatic and extrapancreatic manifestations after corticosteroid treatment.
• Pancreatic biopsies reveal extensive fibrosis and lymphoplasmacytic infiltration.
• HISORT criteria used to establish the diagnosis.

Autoimmune pancreatitis is a rare disorder of presumed autoimmune causation with characteristic chemical, histologic, and morphologic findings. It is referred to by various names, including nonalcoholic destructive pancreatitis, tumefactive pancreatitis, and sclerosing pancreatitis, depending in part on specific pathologic findings and on the presence of extrapancreatic manifestations. However, it is believed that the pathologic heterogeneity may reflect different stages or manifestations of the same disease.

Autoimmune pancreatitis has been described as a primary pancreatic disorder and is also associated with other disorders of presumed autoimmune etiology, including primary sclerosing cholangitis, primary biliary sclerosis, retroperitoneal fibrosis, rheumatoid arthritis, and Sjögren syndrome. As a result, it has been proposed that autoimmune pancreatitis represents a systemic autoimmune disease.

Symptoms and Signs

Mild symptoms, usually abdominal pain, are present but attacks of acute pancreatitis are unusual. Furthermore, autoimmune pancreatitis is not a common cause of idiopathic recurrent pancreatitis. In the United States, 50–65% of patients with autoimmune pancreatitis present with obstructive jaundice. Weight loss and new onset of diabetes may also occur. Patients who have associated autoimmune conditions have corresponding clinical features. In this regard, Sjögren syndrome, rheumatoid arthritis, retroperitoneal fibrosis, ulcerative colitis, autoimmune thyroiditis, tubulointerstitial nephritis, and mediastinal adenopathy have all been reported in patients with autoimmune pancreatitis (Tables 27–1 and 27–2).

Laboratory Findings

An obstructive pattern on liver tests is common (ie, disproportionately elevated serum alkaline phosphatase and minimally elevated serum aminotransferases). Elevated serum levels of immunoglobulin G4 (IgG4) provide a marker for the disease, particularly in Western populations. Serum IgG4 normally accounts for only 5–6% of the total IgG4 in healthy patients but is elevated in those with autoimmune pancreatitis.

A large study compared serum IgG4 levels in 20 patients diagnosed with autoimmune pancreatitis based on clinical features and response to corticosteroids with 20 age- and sex-matched controls and 104 patients with pancreatic cancer, chronic pancreatitis, primary biliary sclerosis, primary cholangitis, and Sjögren syndrome. The median serum IgG4 concentration in the patients with autoimmune pancreatitis was 663 mg/dL compared with 51 mg/dL in healthy controls (in whom normal values ranged from 8–140 mg/dL). Using a cutoff of 280 mg/dL, the sensitivity and specificity of the serum IgG4 value for distinguishing autoimmune pancreatitis from pancreatic cancer was 95–97%.

It has also been demonstrated that IgG4 levels may decline during treatment with corticosteroids, and this can be used as a parameter to follow the response of patients to such therapy. These and other observations suggest that serum levels of IgG4 may be useful to distinguish autoimmune pancreatitis from pancreatic cancer. Additionally, other antibodies such as antinuclear antibodies and rheumatoid factor may be present.

A serologic marker has been identified that is present in most patients with autoimmune pancreatitis. Antibodies against the plasminogen binding protein (PBP) peptide were detected in 19 of 20 patients with autoimmune pancreatitis (95%) but also in 4 of 40 patients with pancreatic cancer (10%). Reactivity was not detected in patients with alcohol-induced chronic pancreatitis or intraductal papillary mucinous neoplasms. While the antibody was detected in most patients with autoimmune pancreatitis, it was also found in 10% of patients with pancreatic cancer making it an imperfect test to distinguish between these two conditions.

Imaging Studies

Table 27–3 summarizes the imaging findings observed in two large series, one from Massachusetts General Hospital (MGH) and the other from the Mayo Clinic. Pancreatic abnormalities were present in the majority of the patients and included diffuse enlargement, focal enlargement, and a distinct enlargement at the head of the pancreas. Noteworthy was the lack of vascular encasement calcification or pancreatic fluid; however, enlarged pancreatic lymph nodes were noted (Figures 27–1 and 27–2).

Endoscopic retrograde cholangiopancreatography (ERCP) or magnetic resonance pancreatography (MRCP) reveals strictures in the bile duct in over one third of patients with autoimmune pancreatitis; these may be common bile duct strictures, intrahepatic bile duct strictures, or proximal bile duct strictures, with accompanying narrowing of the pancreatic bile duct (Figure 27–3). Rim like enhancement of the pancreatic head can be found on magnetic resonance imaging (MRI) examination. Endoscopic ultrasound is increasingly being used to support a diagnosis of autoimmune pancreatitis, and characteristic findings appear to be diffuse pancreatic enlargement and focal irregular hypoechoic masses.

Histopathologic Findings

Characteristic findings include extensive lymphoplasmacytic infiltrates with dense fibrosis, and infiltration of polymorphic leukocytes extending transmurally, ultimately resulting in an obliterative phlebitis (see Table 27–1). Inflammatory cells cluster in the walls of the small veins and nerves. Varying degrees of pancreatic parenchymal atrophy may be present, and islets may be encased with intralobular fibrosis. Importantly, pancreatic ducts exhibit no calcification or plugging but are slit like or star shaped.

To date, there have been few published reports describing cytologic diagnosis of autoimmune pancreatitis. It has been suggested that smears rich in inflammatory cells (lymphocytes, plasma cells, and polymorphonuclear leukocytes with sparse epithelial cells lacking atypia) may be useful in supporting the diagnosis. It is important to emphasize that the diagnosis of autoimmune pancreatitis can be missed in core biopsies because of the sampling area or patchy involvement of the parenchyma.

Autoimmune pancreatitis can be classified as type 1 and type 2 based on histologic patterns and clinical profiles. Type 1 autoimmune pancreatitis is characterized by greater than 10 IgG4 plasma cells and a lymphoplasmacytic sclerosing pancreatitis with periductal lymphoplasmacytic infiltrate, obliterative phlebitis, and dense fibrosis. Type 2 autoimmune pancreatitis or idiopathic duct centric pancreatitis is characterized by a granulocytic epithelial lesion with minimal IgG4 plasma cells in the pancreatic parenchyma. In a study of 97 patients with autoimmune pancreatitis, patients with type 1 were found to have a high relapse rate (47%) compared to patients with type 2 (0%) and was more frequently associated with inflammatory bowel disease.

Autoimmune pancreatitis should be considered in the differential diagnosis of patients presenting with the varied symptoms described earlier that are referable to the pancreas and biliary tract, particularly in conjunction with other autoimmune conditions. It was formerly believed that the sine qua non for establishing a diagnosis of autoimmune pancreatitis was pancreatic histopathologic abnormalities, including extensive lymphoplasmacytic infiltrates with dense fibrosis. It is now recognized that in addition to histologic findings (H), a diagnosis of autoimmune pancreatitis can be made based on pancreatic imaging (I) and serology (S) (IgG4 level), other organ involvement (O) and response to corticosteroid treatment (R) have been determined; these have been termed the HISORT criteria. Utilizing this concept, the proposed Mayo Clinic criteria suggest that autoimmune pancreatitis can be diagnosed with at least one of three abnormalities: (1) diagnostic histology, (2) characteristic findings on computed tomography (CT) and pancreatography, with elevated IgG4 levels; and (3) response to corticosteroid therapy, with improvement in pancreatic and extrapancreatic manifestations (Table 27–4).

Patients with autoimmune pancreatitis sometimes present with features that mimic pancreatic cancer, including a mass on the head of the pancreas. Although autoimmune pancreatitis is diagnosed much less frequently than pancreatic cancer, making the correct diagnosis is necessary to avert the consequences of unnecessary surgery. As detailed in the treatment section, response to corticosteroids within a 2–4 week period is an important parameter by which the diagnosis of autoimmune pancreatitis can be supported. Autoimmune pancreatitis is usually first suggested by imaging procedures such as contrast-enhanced CT or MRI. Of note, characteristic features more suggestive of pancreatic carcinoma (eg, vascular encasement, peripancreatic changes, etc) are usually lacking. Imaging differences between autoimmune pancreatitis and pancreatic cancer are detailed in Table 27–5. A characteristic feature of autoimmune pancreatitis is rim like enhancement of the pancreatic head on MRI. Endoscopic ultrasound is also being used to diagnose autoimmune pancreatitis, and some of the findings that have been noted are listed in Table 27–3.

IgG4 can also be useful in diagnosing pancreatic and extrapancreatic involvement in autoimmune pancreatitis. In one study that examined both pancreatic and extrapancreatic tissues (gallbladder, stomach, duodenum, colon, salivary glands, and kidney), high levels of IgG4-immunostaining plasma cells (>10 IgG4 plasma cells per high-power field [HPF]) were found in 17 of 20 (85%) specimens from patients with autoimmune pancreatitis compared with 1 of 175 (0.6%) specimens from controls. Immunostaining of involved tissue for IgG4 may be particularly useful when autoimmune pancreatitis is suspected clinically but the serum IgG4 level is normal.

Caveats

At the outset, several issues should be raised regarding treatment. First, there have been no controlled trials of treatment approaches in autoimmune pancreatitis. Second, patients groups described in existing reports have been heterogeneous, with over 50% of the patients having undergone a Whipple procedure. Third, until recently, uniform diagnostic criteria were lacking, with the exception of typical pathologic changes in resected pancreatic specimens. Fourth, criteria for initiating treatment and assessing response often have not been uniformly delineated. Nonetheless, after a diagnosis of autoimmune pancreatitis has been made based on the previously outlined criteria, treatment with corticosteroids usually is initiated.

Corticosteroid Therapy

Corticosteroids have shown efficacy in alleviating symptoms, decreasing the size of the pancreas, and reversing histopathologic features in patients with autoimmune pancreatitis. Patients may respond dramatically to corticosteroid therapy within a 2- to 4-week period. Prednisone is usually administered at an initial dose of 40 mg/day for 4 weeks followed by a taper of the daily dosage by 5 mg/week based on monitoring of clinical parameters. Relief of symptoms, serial changes in abdominal imaging of the pancreas and bile ducts, decreased serum γ-globulin and IgG4 levels, and improvements in liver tests are suggested parameters to follow.

Improvement in histologic findings after steroid therapy has also been reported. In the Mayo Clinic series mentioned earlier, the median duration of treatment was 12 weeks. A poor response to corticosteroids over a 2- to 4-week period should raise suspicion of pancreatic cancer or other forms of chronic pancreatitis.

In this regard, Moon et al evaluated the utility of a 2-week trial of glucocorticoids in 22 patients who had atypical imaging findings for autoimmune pancreatitis and did not meet characteristic imaging criteria for pancreatic cancer. All patients received prednisolone (0.5 mg/kg per day). At two weeks, steroid responsiveness was defined as a reduction in the pancreatic mass and marked improvement in pancreatic duct narrowing. All 15 patients who responded were confirmed as having autoimmune pancreatitis while the presence of pancreatic cancer was confirmed at operation in 6 of 7 patients who did not respond.

In most reports, 50–70% of patients responded to corticosteroids, but about 25% required a second course of treatment while a smaller number required maintenance treatment with prednisone at a dosage of 5–10 mg/day. In the Mayo Clinic series, 15 of 29 patients had resolution of biliary strictures or pancreatic masses, or both, with steroid treatment.

Immunomodulatory Drugs

Several recent studies have reported on the use of immunomodulatory drugs (azathioprine, mycophenolate, mofetil, cyclophosphamide) in autoimmune pancreatitis patients who either flared or relapsed while on or after withdrawal of corticosteroid therapy. Raina et al found that 9 of 15 patients who responded initially to corticosteroids relapsed after steroid withdrawal but all 9 responded and are being maintained on azathioprine. In another report by Sandanayake et al, in 10 of 13 patients treated with azathioprine, remission was achieved and maintained by 7 patients with azathioprine monotherapy at a median follow-up of 14 months. Successful treatment with rituximab, a monoclonal antibody, was reported in two patients with autoimmune pancreatitis, one of whom was refractory to glucocorticoid and 6-mercaptopurine. For additional information on immunomodulatory treatment autoimmune pancreatitis see section below.

Immunoglobulin G4-Associated Cholangitis

IgG4-associated cholangitis (IAC), which can mimic primary sclerosing cholangitis, is the biliary manifestation of a steroid-responsive multisystem fibroinflammatory disorder (ie, autoimmune pancreatitis). The affected organs have a lymphoplasmacytic infiltrate rich in IgG4-positive cells. A recent study described clinical, serologic, and imaging characteristics and treatment response in 53 IAC patients. Among the study participants, 85% were male, 77% presented with obstructive jaundice, 92% had associated autoimmune pancreatitis, 74% showed increased serum IgG4 levels, and 88% had IgG4-positive cells in bile duct biopsy specimens (>10/HPF). Bile duct strictures were common, with isolated intrapancreatic bile duct strictures in 51% of patients, proximal extrahepatic or intrahepatic bile duct strictures in 49%, and multifocal structures in 32%.

Thirty patients received prednisone at a dose of 40 mg/day for 4 weeks, which was then tapered to 5 mg/week for a total of 11 weeks; 18 patients underwent surgical resection; and 5 patients were treated conservatively and had spontaneous resolution of their disease. The following treatment responses were observed. Of the 30 patients treated with prednisone, 29 (97%) had an initial response, and 18 (60% of patients) had resolution of strictures and normalization of liver enzymes. Twelve of 17 patients (71%) with proximal bile duct strictures relapsed less than 6 months after prednisone was discontinued; in these patients, the presence of proximal extrahepatic or intrahepatic stricture was predictive of relapse. Fifteen patients were retreated with steroids for relapse, and 7 patients required additional immunomodulating drugs (azathioprine, cyclophosphamide, or mycophenolate mofetil) to remain in steroid-free remission during more than 6 months of follow-up. Eight of the 15 patients retreated with steroids (44%) relapsed.

The authors of the study drew the following conclusions from these findings: first, IAC should be suspected in patients with unexplained biliary strictures associated with increased IgG4 and unexplained pancreatic disease; second, relapses are common after steroid withdrawal, especially in patients with proximal strictures; and third, the role of immunomodulating drugs for relapses needs to be explored through further study.

Hirano and colleagues reported outcomes and clinical features in 42 patients with autoimmune pancreatitis. Nineteen patients were treated with corticosteroids (prednisolone) and 23 without. The average observation period was 25 months but ranged up to 7 years. Unfavorable events (sclerosing cholangitis, distal bile duct stenosis, and retroperitoneal fibrosis) developed in 16 of the 23 patients (70%) not treated with corticosteroids but in only 3 of 19 patients (32%) treated with prednisolone. These data support the recommendation that patients with autoimmune pancreatitis, especially those with obstructive jaundice and bile duct strictures, should be treated and maintained on corticosteroids.