Introduction
Type 1 autoimmune pancreatitis (AIP) represents the most common manifestation of IgG4-related disease. Type 1 AIP responds well to glucocorticoids, but relapses are common requiring multiple treatment courses and ultimately leading to end-organ damage and treatment-related morbidity. To date, scant reports exist on the long-term outcomes of these patients. Moreover, the role of maintenance treatment in preventing disease relapse is still debated.
Aims & Methods
We aimed to evaluate the long-term outcomes of a large European cohort of patients with type 1 AIP and to assess the role of maintenance treatment in preventing disease relapse. Patients with type 1 AIP fulfilling ICDC criteria were included in this multicentric retrospective study within the PrescrAIP framework involving 40 European university Hospital. Relapses were defined as new onset of either AIP related symptoms or radiological abnormalities after a successful induction of remission. Patients without either complete or partial remission were excluded. The primary endpoint was time to relapse. Predictors of relapse were identified by multivariable logistic regression. A bootstrapping technique was performed to internally validate the model. Regression coefficients from the multivariable model were employed to develop a score for computing the risk of relapse for each patient. Patients were divided into four categories according to the score, from the lowest to highest risk of relapse.
Results
477 patients were included in this analysis (325 male, 68%). Mean age at baseline was 59 (IQR 40-68). 60% of the patients presented with jaundice while 7% of the patient with acute pancreatitis. Other organs involvement was present in 238 patients (48%), being biliary involvement the most frequent one (40%). At baseline, diabetes mellitus and pancreatic exocrine insufficiency were present in 30% and 27% of patients, respectively. After a mean follow up of 32 months (IQR 13-68), 150 (31%) AIP relapses occurred. After accounting for confounders, biliary tree involvement (OR 1.62 95%CI 1.1-2.2), acute pancreatitis at baseline (OR 2.1 95%CI 1.1-4.2), narrowing of the main pancreatic duct (OR 1.6, 95%CI 1.1-2.1) were independently associated with relapse, while AIP presenting as a focal mass resulted in a protective factor (OR 0.65 95%CI 0.46-0.92). Increased serum IgG4 level (OR 1.2 95%CI 0.9-1.7) pointed to higher relapse rate without reaching statistical significance. Notably, any maintenance treatment reduced the relapse rate after correcting for the above-mentioned risk factors (OR 0.51, 95%CI 0.37-0.72). There was no difference in term of relapse free survival between maintenance treatment with any anti-rheumatic drugs (azathioprine, methotrexate, mycophenolate) and low-dose GCs. Maintenance treatment reduced relapse rate only in patients with higher relapse risk. At the longest follow up, DM and PEI were recorded in 33% and 29% of the patients, respectively. Only one pancreatic cancer was detected.
Conclusion
Several disease features increase the risk of relapse in type 1 AIP. Any of the maintenance treatment reduce relapse rate in type 1 AIP. Patients showing multiple risk factors for relapse benefit the most from maintenance treatment. The development of a score predicting disease relapse enhances a personalized approach toward AIP treatment. In European AIP patients the development of pancreatic cancer during follow up appears rare.