Autoimmune pancreatitis—also a Western disease

During the five decades after Hashimoto described struma lymphomatosa of the thyroid gland featuring diffuse lymphocytic infiltration, fibrosis, and parenchymal atrophy,¹ autoimmunity became increasingly recognised, such that when similar histological changes were described affecting the pancreas, an autoimmune mechanism was postulated.² Autoimmune pancreatitis then remained among a number of plausible explanations for the appreciable minority of patients with chronic pancreatitis who give no history of alcohol abuse.³ More recently, Japanese workers again took up the baton, describing series of patients who had undergone pancreatic resection for suspected neoplasia but whose histology demonstrated lymphocytic infiltration of the exocrine parenchyma, with periductular and interlobular fibrosis.^4,5 Further study demonstrated a T cell predominance to the infiltrate,⁵ autoantibodies directed against carbonic anhydrase II,⁶ an enzyme present in pancreatic as well as salivary ductal epithelium, and frequently high serum levels of IgG4.⁷ These findings have been confirmed in the European population, as reported by Aparisi et al in this issue of Gut,⁸ who describe a small series within a larger group of patients with non-alcoholic chronic pancreatitis (see page 703). Although not stated, such a group is highly likely to include patients with mutations of the cystic fibrosis transmembrane conductance regulator (CFTR)⁹ and serine protease inhibitor Kazal type 1 (SPINK1)¹⁰ which both increase the risk of chronic pancreatitis, as well as mutations of cationic trypsinogen (protease, serine, 1 (trypsin 1): PRSS1) which cause hereditary pancreatitis.¹¹

Confidence in the entity autoimmune pancreatitis is predicated not only on circumstantial clinical evidence that includes HLA restriction,¹² but also substantial indirect proof from animals that is perhaps as strong as for other autoimmune gastrointestinal diseases that include Sjögren’s syndrome, achalasia, atrophic gastritis, pernicious anaemia, autoimmune hepatitis, primary biliary cirrhosis, sclerosing cholangitis, inflammatory bowel disease, and autoimmune gastroenteropathy.¹³ Spontaneous development of autoimmune pancreatitis is a characteristic of the autoimmune disease prone MRL/+ mouse¹⁴ and Wistar Bonn/Kobori (WBN/Kob) rat,¹⁵ and also occurs in German shepherd and rough coated collie dogs.¹⁶ Neonatal thymus transplantation corrects the autoimmune pancreatitis and sialoadenitis of adult MRL/+ mice,¹⁷ indicating a T cell dependent process. Adoptive transfer of CD4 (helper) T cells from MRL/+ mice with autoimmune pancreatitis into unaffected syngeneic animals induces autoimmune pancreatitis.¹⁴ Autoimmune pancreatitis can also be induced by adoptive transfer of CD4 T cells from mice immunised by repeated administration of carbonic anhydrase II and/or lactoferrin in Freund’s adjuvant.¹⁸ Both CD4 and CD8 (suppressor/cytotoxic) T cells infiltrate pancreatic exocrine tissue in the MRL/+ mouse,¹⁴ WBN/Kob rat,¹⁵ German shepherd, rough coated collie,¹⁶ and humans,^5,19 suggesting a mechanism akin to that in autoimmune beta cell destruction of the non-obese diabetic (NOD) mouse, BioBreeding/Worcester rat, and humans.²⁰ It has been suggested that diabetes mellitus complicating autoimmune pancreatitis may be due to bystander injury,⁵ but the relatively large arterial flow to pancreatic islets and the widespread distribution of beta cells within islets would tend to minimise this; also, immune destruction of islets is highly specific.²¹ Direct T cell mediated damage of beta cells in autoimmune pancreatitis is suggested by the presence of CD8 cells infiltrating islets in patients with the disease, in association with a diminished proportion of beta cells.²²

The autoimmune spectrum extends from organ specific autoimmune disease, as in autoimmune thyroiditis and beta cell destruction causing type 1 diabetes mellitus, through intermediate forms such as primary biliary cirrhosis, to systemic autoimmune disease including systemic lupus erythematosus and rheumatoid arthritis, both of which can feature organ specific autoimmune disease.²⁰ Patients with autoimmune pancreatitis may present with or develop Sjögren’s syndrome, primary biliary cirrhosis, a form of sclerosing cholangitis, inflammatory bowel disease, as well as type 1 diabetes mellitus.^5,19,22 Local inflammation may extend to the gall bladder and duodenum²³; furthermore, periportal and/or retroperitoneal fibrosis¹⁹ may complicate the picture, perhaps because of determinant spreading, broadening the antigenic targets of autoimmunity.²⁰ At the other end of the spectrum, systemic lupus erythematosus has been described featuring both acute²⁴ and chronic²⁵ pancreatitis not attributable to drugs. Systemic autoimmunity may have a somewhat different pathogenesis, with damage initiated by antibody and complement responses to multiple products of apoptosis²⁰; the pancreas, therefore, may be subject to more than one form of autoimmune attack.

Interestingly, an autoimmune component in acute pancreatitis has increasing been recognised.²⁶ Cases of autoimmune pancreatitis now on record include patients presenting with focal as well as more recognisably typical diffuse pancreatic masses, and occasionally pseudocysts, parenchymal atrophy, and/or calcification.⁵ These latter apparently atypical findings have been used by some to exclude the diagnosis,¹⁹ as in the current study by Aparisi and colleagues,⁸ who also separated patients with Sjögren’s syndrome from those with suspected autoimmune pancreatitis. While this approach has some specificity, it does not have sensitivity; consensus on diagnostic criteria is yet to be achieved. As with other forms of pancreatic injury extending over prolonged periods, there is likely to be a range of clinical findings that includes acute as well as the more prevalent chronic forms, the full extent of which awaits complete characterisation. Further unrecognised autoimmune pancreatitis may be present in patients with other autoimmune diseases where the condition is not considered, or the predominant symptoms are not pancreatic, or where treatment with steroids or other immunosuppressive agents masks or alters the progression of autoimmune pancreatitis.

Increasing acceptance of autoimmune pathology affecting the exocrine pancreas has prompted a call for the term autoimmune pancreatitis²⁷ to replace the obscure terms sclerosing pancreatitis, primary inflammatory pancreatitis, lymphoplasmacytic pancreatitis, sclerosing pancreaticocholangitis, idiopathic tumefactive chronic pancreatitis, lymphoplasmacytic sclerosing pancreatitis, and idiopathic duct destructive pancreatitis.²⁸ For clarity we would diagnose all cases as autoimmune pancreatitis without differentiation of “primary” from “secondary” forms depending on the absence or presence of other autoimmune disease,²⁷ because the association with other autoimmune disease has yet to be fully defined, and when present with other autoimmune disease, the pathogenetic mechanism is not necessarily “secondary”. It is of interest to reflect on the parallels with autoimmune thyroiditis, a condition that had many names, including chronic thyroiditis, lymphocytic thyroiditis, lymphadenoid goitre, and the original struma lymphomatosa,¹ which for some time was a diagnosis made on the pathological assessment of resected specimens; notably, the condition used to be considered an uncommon disease.

Autoimmune pancreatitis can present with a non-specific history of pain and/or anorexia with weight loss that is often accompanied by jaundice from biliary obstruction; the condition may be heralded by an initial attack of acute pancreatitis.²³ Other autoimmune diseases of either the patient and/or family should draw attention to the possible diagnosis. As well as, or as an alternative to, elevated levels of IgG4 and positive anticarbonic anhydrase II titres, antilactoferrin, antinuclear, antismooth muscle, and rheumatoid factor antibodies may be identifiable.⁵ Germline gene analysis must also be undertaken to identify patients with CFTR, SPINK1, or PRSS1 mutations,^9–11 as there is overlap in the clinical features associated with all forms of chronic pancreatitis.^3,11 Radiological imaging by computerised tomography (CT), magnetic resonance imaging (MRI), or endoscopic ultrasound (EUS) may show a sausage shaped swollen pancreas or focal masses that can be single or even multiple.²⁹ Endoscopic retrograde cholangiopancreatography (ERCP) may identify biliary strictures from pancreatic compression or a pattern typical of sclerosing cholangitis, as well as pancreatic ductal stenoses²⁹; stenting may be required. The presentation can be indistinguishable from that of pancreatic cancer or much rarer pancreatic lymphoma, which should be considered for resection and/or chemotherapy. Transduodenal EUS guided biopsy offers the best prospect for unravelling this conundrum,^5,27 but it must be emphasised that a trial of steroid therapy cannot be recommended if substantial doubt remains as to the diagnosis in an operable patient. Such a trial of treatment could reduce inflammation surrounding a cancer and be interpreted as appropriate, only to be followed by progression to unresectability. In any case, optimal treatment regimens are undefined and a morphological response may take a number of weeks if not months^5,19; currently, resection is the commonest route to diagnosis. While recognising the importance of surgery when indications remain, and the acceptable results of surgery in tertiary centres when autoimmune pancreatitis is the final diagnosis,^3,30 recent experience at the Mayo Clinic has been that of a diagnosis of autoimmune pancreatitis in an increasing number of patients, with a decreasing percentage of patients undergoing surgery (Dr S Chari, personal communication, 2004). Increasing recognition of this disease has led to the identification of more cases where there is little or no indication for surgery but some cases may have to be teased out from an otherwise appropriate surgical pathway.

It is recommended that all patients with a possible or suspected diagnosis of autoimmune pancreatitis be referred to a pancreatologist, usually at a regional pancreas centre. This recommendation is all the more important as there are no published guidelines for the management of this condition. If surgery is deemed inappropriate, pancreatic exocrine and endocrine function should be formally assessed at an early stage. Suggestions as to therapy are preliminary, but if there is justifiable confidence in the diagnosis and symptoms significant, a reducing dose of prednisolone is reasonable. With this regimen, resolution of pancreatic pain as well as exocrine and endocrine insufficiency has been seen; partial or complete resolution of biliary strictures of an apparently atypical form of sclerosing cholangitis has been witnessed by us and an increasing number of workers.^5,29 Supplementary exogenous pancreatic enzyme supplementation may be required for some weeks, months or indefinitely, while treatment for diabetes mellitus will require joint care with a diabetologist; rheumatological or other expertise may also be required to help tailor treatment in patients with other autoimmune disease. Measures to monitor treatment response include assessment of symptoms, pancreatic exocrine and endocrine function, as well as full blood count, liver function tests, EUS with or without biopsy, CT, or MRI; in patients with endoprostheses, long term antibiotics may be justified and a further ERCP will be needed to remove the stent at an early opportunity. How long to continue steroids will depend on the response, side effects, and the effect of treatment withdrawal, as well as the presence and need for treatment of other autoimmune disease. Whether there is a role for alternative therapies remains to be defined, as does the place of autoimmune pancreatitis in the gamut of autoimmune diseases.