Recognizing autoimmune pancreatitis

The term autoimmune pancreatitis (AIP) is used for two rare, steroid-responsive pancreatitides that present most commonly with a pancreatic mass and occasionally with acute pancreatitis. They are the only pancreatic diseases with a potential cure through effective medical therapies. Although their increased recognition has led to the avoidance of unnecessary surgery in numerous patients, it has also led to overdiagnosis, unnecessary exposure to steroid therapy or, worse, inappropriate treatment with potentially toxic immunosuppressive agents and surgical delay for pancreatic cancer. Diagnosing AIP is analogous to the ability to recognize the “hoof beats of a zebra in the midst of horses.” It requires a thorough understanding of the clinical, radiologic, serologic and pathologic features of AIP and the diseases they mimic.

It is well known, but less well appreciated, that the term AIP encompasses two very different but uncommon diseases, currently designated type 1 and type 2 AIP. However, this terminology has not really helped distinguish the two diseases. Therefore, we have suggested that AIP without a qualifier or the term lymphoplasmacytic sclerosing pancreatitis (LPSP) be used for type 1 AIP and type 2 AIP be replaced with the term idiopathic duct centric chronic pancreatitis (IDCP).

LPSP is a disease of older men (75 percent of individuals are aged > 50 years and men) and is very rare in children. Although LPSP is a form of pancreatitis, it is usually painless and most commonly presents with obstructive jaundice and diffuse or focal pancreatic enlargement mimicking pancreatic cancer. However, only about 2 to 3 percent of patients who undergo resection for suspected pancreatic cancer have AIP. LPSP is the pancreatic manifestation of a multiorgan disease called immunoglobulin G4 (IgG4)-related disease and, in more than 50 percent of cases, is associated with other organ involvement and, in approximately 75 percent of cases, serum IgG4 elevation. On core biopsy or resection from any affected organ, it shows typical histologic features: lymphoplasmacytic infiltrate, storiform fibrosis, obliterative phlebitis and increased number of IgG4-positive plasma cells.

By contrast, IDCP is much less common than LPSP, localized to the pancreas and not associated with IgG4 elevation. It is best suspected in children or young adults presenting with painless jaundice, and it affects males and females equally. In patients with pancreatitis or pancreatic mass with associated inflammatory bowel disease, IDCP remains a consideration, although the possibility of drug-induced or other causes of acute pancreatitis are often more likely. It is histologically characterized by the presence of granulocyte epithelial lesion, i.e., focal duct disruption and destruction due to neutrophil invasion.

The diagnostic strategy to distinguish AIP/LPSP from pancreatic cancer utilizes (in temporal order) imaging, other organ involvement, serology, histology and response to steroids (features form the mnemonic HISORt). In a patient with a pancreatic mass, the appearance of the pancreas on computed tomography can be classified into three categories: highly suggestive/diagnostic, supportive of AIP, or highly suggestive of pancreatic cancer.1 A diffusely enlarged gland with featureless borders and delayed enhancement, especially with a capsule-like rim, is highly suggestive of AIP (nearly 50 percent of cases of AIP; rare in pancreatic cancer). A focally enlarged pancreas without low-density mass, dilated upstream duct, or atrophy is suggestive of AIP, although it is statistically more common in pancreatic cancer. Features that are typical of pancreatic cancer (low-density mass, dilated upstream duct, atrophy) should be managed as cancer. Collateral evidence for AIP includes history of other current organ involvement (e.g., kidneys, proximal biliary system, retroperitoneal fibrosis), serology (elevated serum IgG4, especially twofold), and findings on pancreatic core biopsy with features suggestive of AIP.

In patients with typical imaging features, any collateral evidence is diagnostic of LPSP. Patients with imaging features not typical for AIP but with strong collateral evidence (≥ 2 highly suggestive features) can also be confidently diagnosed with AIP. These two scenarios cover 70 percent of cases of AIP. All other patients need exclusion of cancer by fine needle aspiration. Histologic diagnosis of AIP will require core biopsy. If there are two or more collateral pieces of evidence, then careful patient selection and a monitored steroid trial may be appropriate. We do not recommend a steroid trial in the complete absence of collateral evidence, and biopsy is recommended. A very small subset of patients will undergo surgery despite the clinician’s best efforts to preoperatively diagnose the condition.

A word of caution regarding steroid trials: All that responds to steroids is not AIP. Steroids uplift mood, resolve arthritic pain, increase appetite, and cause weight gain even in patients with pancreatic cancer, thereby confounding the ability to judge “response.” Follow-up imaging results may show improvement in pancreatitis caused by the cancer, mimicking steroid response. This often prolongs the trial. If you are using a steroid trial, beware of its pitfalls.

Both LPSP and IDCP are characterized by their brisk response to steroid treatment, although frequent relapses tend to be the norm for LPSP and the exception for IDCP. However, in both LPSP and IDCP, radiologic relapse while on at least 20 mg of prednisone usually suggests that the diagnosis needs reconsideration. Patients with LPSP require close follow-up with repeat liver biochemistries checked every three months for at least the first two to three years. Patients who have frequent relapses or those whose condition becomes steroid dependent often require the addition of immunomodulators such as azathioprine to maintain remission, which we suggest taking for up to three years. Others who have contraindications to prednisone or whose disease is intolerant to immunomodulators can be given rituximab for induction or remission maintenance. The frequency, duration and end points of rituximab require further investigation.

The confusion between LPSP and IDCP needs mention. Not uncommonly we see young patients with idiopathic pancreatitis diagnosed as having AIP based on a mildly elevated serum IgG4 level and treated with steroids and, when pancreatitis recurs, azathioprine and sometimes rituximab. Just as pancreatic cancer in an older patient carries a poor prognosis, recurrent pancreatitis in a young patient is similarly devastating. Although the tendency to overdiagnose AIP is understandable, use of potentially toxic immunosuppressants in the absence of disease is problematic.

Key take away points

  • AIP is an important cause of a pancreatic mass that presents with obstructive jaundice. A systematic approach using an experienced multidisciplinary team can lead to a noninvasive diagnosis and prevent unnecessary surgery.
  • Steroids are the mainstay of initial treatment in patients with both LPSP and IDCP. However, relapses are common in LPSP and may require initiation of immunomodulators such as azathioprine or rituximab. Those intolerant to immunomodulation or steroids may require rituximab.
  • We do not recommend using a steroid trial as a surrogate for further testing, because a delayed diagnosis of cancer can significantly affect patient outcomes. A steroid trial should only be used in a very carefully selected group of patients after a thorough work-up has ruled out malignancy.
  • Diagnosing AIP as a cause of acute idiopathic pancreatitis or idiopathic recurrent acute pancreatitis has to be done very carefully as it is a rare manifestation of a very rare disease.

Dr. Nagpal has no conflicts to disclose. Dr. Chari has no conflicts to disclose. Dr. Chari is the president of the International Association of Pancreatology.

1. Chari, S.T., Takahashi, N., Levy, M.J. et al, A diagnostic strategy to distinguish autoimmune pancreatitis from pancreatic cancer. Clin Gastroenterol Hepatol. 2009;7:1097-1103.

Join the discussion

Your email address will not be published. Required fields are marked *