Incidence and outcome of lung involvement in IgG4-related autoimmune pancreatitis

Immunoglobulin G4-related disease (IgG4-RD) is a recently described systemic fibroinflammatory condition associated with an elevated serum IgG4 level and abundant IgG4-positive plasma cell infiltration involving multiple organs.1, 2 Type 1 autoimmune pancreatitis (AIP) is the prototypical form of systemic IgG4-RD, and various clinical findings have been documented for this disease. Early intervention using corticosteroids has been shown to improve IgG4-related organ dysfunction, although relapse of the disease is common. In addition, various types of lung involvement in IgG4-RD, including interstitial pneumonia, inflammatory pseudotumours and lymphadenopathy, have been described.3-7 In contrast, the clinical characteristics of lung involvement in IgG4-RD, in particular the response to treatment and prognosis, remain unclear. In this study, we retrospectively evaluated the incidence and outcomes of lung involvement in 35 patients with AIP.

Between December 1996 and March 2011, a total of 35 patients with type 1 AIP diagnosed at our university hospital according to the consensus diagnostic criteria8, 9 were consecutively enrolled in this study (24 males and 11 females, average age: 67 ± 9), with a median follow-up time of 52 ± 33 months (range: 5–160 months). All patients were examined using chest and abdominal computed tomography (CT) at the time of diagnosis and relapse, and underwent chest X-ray examinations at least every 6 months, regardless of the presence or absence of respiratory symptoms. In addition, all patients regularly received medical examinations by a physician using abdominal ultrasound and assessments of biochemical parameters, including serum IgG4 levels. Lung involvement in AIP was diagnosed based on findings of chest CT abnormalities accompanied by compatible X-ray findings associated with IgG4-related lung disease, with or without a favourable response to corticosteroid therapy. Definitive diagnosis for a relapse of AIP and lung involvement was according to the CT findings since re-elevation of the serological levels alone without any abnormal CT findings is not considered to indicate a relapse.10-13 We retrospectively reviewed the following data: (i) respiratory symptoms and pulmonary radiological findings on chest CT; (ii) clinical and laboratory findings of the patients with and without lung involvement; (iii) pulmonary and pancreatic responsiveness to corticosteroid therapy; and (iv) radiological findings of the lungs and pancreas at the time of recurrence of AIP. The data were obtained from the patients' medical records, and the clinical and laboratory findings were collected at the time of the diagnosis, before starting treatment. The following serum data were also evaluated: IgG (normal range: 863–1589 mg/dL), IgG4 (4–108 mg/mL), IgG4/IgG (3–6%) and amylase (0–132 mg/dL).

The study was approved by the Human and Animal Ethics Review Committee in our university (H26-038). All results are reported as the mean ± standard deviation and were compared using the Mann–Whitney U-test or Fisher's exact test, as appropriate. P values of <0.05 were considered to be statistically significant.

Fourteen of the 35 AIP patients (40%) were diagnosed with lung involvement in AIP (L-AIP) (male/female: 9/5, average age: 68 ± 9 years), while the other 21 patients were found to be without lung involvement (NL-AIP) (male/female: 15/6, average age: 67 ± 8 years). The details of the clinical and laboratory findings for the L-AIP and NL-AIP groups are shown in Table 1 and Supplementary Table S1. The 14 L-AIP patients were followed up for a median of 64 months (range: 10–160 months). At the time of diagnosis of AIP, eight of the 14 L-AIP patients had respiratory symptoms (57%), including exertional dyspnoea in four patients and/or coughing in six patients, and four patients (29%) had complicating allergic disorders, such as allergic conjunctivitis in one case and bronchial asthma in three cases. The major chest CT findings in L-AIP included lymphadenopathy in 13, ground-glass attenuation in 7, bronchial wall thickening in 5, nodule in 3 and consolidation in 2. Elevations of the serum level of IgG4, IgG, IgG4/IgG and amylase were detected in 92%, 71%, 100% and 43% of L-AIP. However, clinical and laboratory data such as observation period, age, gender, complication allergic history, serum IgG4, IgG, IgG4/IgG and amylase level did not differ significantly between the L-AIP and NL-AIP group. A histological examination of the lungs was performed in three of the 14 patients using a transbronchial lung biopsy, and numerous lymphoplasmacytic infiltrates with IgG4-positive plasma cells (IgG4/IgG > 40%) were observed in all of these patients.

During the observation period, one of the L-AIP patients and two of the NL-AIP were monitored under a wait-and-see policy without treatment, as the symptoms were minor and no radiological progression was observed in either the lungs or pancreas. The remaining 13 L-AIP patients and 19 NL-AIP patients received initial therapy with prednisolone (0.51 ± 0.15, 0.53 ± 0.11 mg/kg/day, P = 0.78, respectively), and improvements were subsequently observed in both the AIP and lung involvement several months after starting treatment with prednisolone. Relapse of the pancreatic lesions occurred during tapering of the dose of corticosteroids or after the completion of corticosteroid therapy in five of the 14 L-AIP patients (36%) (dose range of prednisolone at the time of relapse: 3–10 mg/day). In contrast, no cases of radiological relapse of lung involvement were noted among all the 14 L-AIP patients. Furthermore, although six of the 21 patients with NL-AIP (29%) demonstrated relapse of the pancreatic lesions (dose range of prednisolone at the time of relapse: 3–10 mg/day), none of these patients exhibited new lung lesions after starting corticosteroid treatment for the observation period.

The results of the present study showed a high frequency of lung lesions as a complication of AIP (40%) and all the patients who received corticosteroids exhibited radiological improvements. Interestingly, although relapse of pancreatic lesions was observed in five of fourteen L-AIP patients with lung involvement, no relapse of lung lesions was noted (36% vs 0%). This is the first report to assess the incidence and outcome of lung involvement in AIP over a long observation period.

Lung involvement with AIP has been reported to be a systemic complication of AIP (10–51.2%),11-14 which is consistent with the present findings. Although the efficacy of corticosteroids for the treatment of lung involvement associated with AIP remains unclear, early intervention using corticosteroids has been shown to improve AIP (82–98%).10, 15, 16 Furthermore, it is generally thought that most cases of IgG4-related lung disease exhibit a good response to corticosteroids.6^,7, 17 Our results also show that lung lesions associated with AIP improve following treatment with corticosteroids. While the relapse of IgG4-related lung disease or lung involvement in AIP has not yet been comprehensively evaluated, AIP has been reported to frequently relapse when the dose of corticosteroids is tapered or discontinued10, 15, 16 (22–85%), with most cases of relapse occurring within three years.10 Our results are the first to show that lung lesions have a lower probability of relapse compared with pancreatic lesions (0% vs 36%).

There are two limitations associated with the present study. First, the design was retrospective, observational and included a small sample size. Second, some of the lung lesions were not confirmed histopathologically, and the cases may have included conditions other than IgG4-related lung disease as the cause of lung involvement in AIP. However, numerous infiltrates with IgG4-positive plasma cells were histopathologically observed in the lung tissue specimens obtained from three patients.

In conclusion, our results indicate that lung involvement in AIP is not rare and exhibits improvements on steroids in addition to a low probability of relapse.