Methods

A retrospective survey of AIP treatment, focusing on steroid treatment, was conducted in the 17 centres that participated in this study. The majority of these centres are major referral centres across Japan with established expertise in the diagnosis and management of AIP. Tokyo Metropolitan Komagome Hospital and Tohoku University served as the coordinating centres.

The diagnosis of AIP was made according to the Asian diagnostic criteria for AIP.11 To make the diagnosis of AIP, the imaging criterion, consisting of enlargement of the pancreas and irregular narrowing of the main pancreatic duct, must be present, together with the serological criterion (elevated serum IgG or IgG4 levels, or detection of autoantibodies) and/or the histopathological criterion (lymphoplasmacytic sclerosing pancreatitis). AIP can be also diagnosed with fulfilment of both the imaging criterion and a good response to steroid treatment.

Data regarding induction of remission by steroid treatment, maintenance steroid treatment and relapse of AIP were analysed. Remission was defined as the disappearance of clinical symptoms and resolution of the pancreatic and/or extrapancreatic manifestations on imaging studies.8 12 13 14 For follow-up after remission, laboratory tests and imaging studies were performed periodically, usually every 3–6 months in the first year. Relapse of AIP was defined as reappearance of symptoms with the development or reappearance of pancreatic and/or extrapancreatic (including bile duct, salivary gland and retroperitoneum) abnormalities on imaging studies and/or marked elevation of serum IgG or IgG4 levels.8 12 13 14 Re-elevation of serological levels alone without clinical symptoms or abnormal imaging was not considered to be relapse.

Statistical analysis was performed using Fisher’s exact test and Mann–Whitney’s U test. Differences with p values of <0.05 were considered significant. The period from the start of steroid treatment to relapse was evaluated using the Kaplan–Meier curve.

After analysis of the data, a consensus meeting was held involving members of the 17 centres to propose a consensus regarding steroid treatment for AIP.

Results

Study subjects

A total of 563 cases of AIP (439 men and 124 women, average age 63.0 years) were confirmed to fulfil the Asian diagnostic criteria for AIP, and they were enrolled in the analyses for steroid treatment of AIP. Of these, 459 (82%) patients with AIP (374 men and 85 women, average age 62.3 years) received steroid treatment. Of the others, 56 patients underwent surgical procedures, and 48 patients were followed-up conservatively.

No patients received any other immunosuppressive treatments such as azathioprine or ulsodeoxycholic acid.

Induction of remission by steroid treatment

The remission rate of patients with AIP was significantly higher in patients who received steroid treatment (98%, 451/459) than in those not given steroid treatment (74%, 77/104; p<0.001) (table 1).

Steroid treatment was administered mainly for obstructive jaundice (247/459 patients, 60%), abdominal pain (51 patients, 11%), associated extrapancreatic lesions such as retroperitoneal fibrosis (50 patients, 11%), diffuse enlargement of the pancreas (45 patients, 10%) and confirmation or differentiation of the diagnosis after a negative investigation of pancreatic cancer (24 patients, 5%).

In patients with DM, before steroid administration, blood glucose levels were controlled using insulin in 104 patients and using oral antidiabetic medicines in 39 patients. Endoscopic or transhepatic biliary drainage was performed in 242 (77%) of 314 patients with obstructive jaundice due to associated sclerosing cholangitis. Endoscopic or transhepatic biliary drainage was performed for patients showing hyperbilirubinaemia of >3 mg/dl in two-thirds of centres.

The initial oral prednisolone dose was 20 mg/day (n = 8, 2%), 30 mg/day (n = 283, 62%), 40 mg/day (n = 160, 35%), 60 mg/day (n = 4, 1%) and others (n = 4, 1%). The initial dose was administered for 2 weeks in three-quarters of cases, and for 3–4 weeks in the remainder. The initial dose was gradually tapered by 5 mg every 1–2 weeks to the maintenance dosage, based on changes in the clinical manifestations, biochemical blood tests (such as serum liver enzymes and IgG or IgG4 levels) and repeated imaging findings. The dose was tapered more gradually, such as 2.5 or 5 mg every 2–8 weeks, after the dose reached 15 mg/day.

The period necessary to achieve remission from the start of initial administration was 6.82 (6.11) months (mean (SD)) in the patients treated with an initial prednisolone dose of 30 mg/day, which was not significantly different from the period (6.34 (8.13) months) in those treated with an initial prednisolone dose of 40 mg/day (p = 0.401) (table 2).

At remission, the enlarged pancreas returned to near-normal size in 239 (80%) of 300 patients. It became atrophic in 58 patients (20%), and showed persistent focal enlargement in 3 patients. Elevated serum IgG4 levels decreased in all patients after the start of steroid treatment, but they failed to normalise (<135 mg/dl) in 115 (63%) of 182 patients. At remission, irregularity of the pancreatic ducts and/or some degree of bile duct stenosis remained in 67 (58%) of 115 patients with persistent elevation of serum IgG4 levels, while it remained in only 18 (27%) of 67 patients with normalised serum IgG4 levels (p<0.001).

Maintenance steroid treatment

Maintenance steroid treatment was performed after remission in 377 (82%) of 459 patients treated with steroid. The maintenance oral prednisolone dose was 10 mg/day (n = 27, 7%), 7.5 mg/day (n = 13, 3%), 5 mg/day (n = 238, 63%), 2.5 mg/day (n = 78, 21%) and others. Of the 377 patients who underwent maintenance treatment, the maintenance treatment was stopped in 104 patients (28%) in whom complete radiological and serological improvement was obtained.

Relapse of AIP

The relapse rate of patients with AIP was significantly lower in those who received steroid treatment (24%, 110/451) than in those not given steroid treatment (42%, 32/77; p = 0.003) (table 1). In the patients who received steroid treatment, relapse occurred in the pancreas (n = 57, 52%), bile duct (n = 37, 34%) and extrapancreatic lesions (n = 19; salivary gland swelling (n = 10), interstitial pneumonia (n = 4), interstitial nephritis (n = 2) and others).

There was no correlation between the relapse rate and the initial prednisolone dose (40 mg/day: 19% (31/160) vs 30 mg/day: 23% (65/283), p = 0.402) (table 2). As regards the period from the start of steroid treatment to relapse, 32% (32/99) relapsed within 6 months, 56% (55/99) relapsed within 1 year, 76% (75/99) relapsed within 2 years and 92% (91/99) relapsed within 3 years after starting medication (fig 1). The relapse rate of patients with AIP on maintenance treatment was 23% (63/273), which was significantly lower than that of patients who stopped maintenance treatment (34%, 35/104; p = 0.048). The doses of prednisolone at the time of relapse were 10 mg/day (n = 10, 16%), 7.5 mg/day (n = 7, 11%), 5 mg/day (n = 29, 46%), 2.5 mg/day (n = 8, 13%) and others.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Relapse rate of autoimmune pancreatitis and the period from the start of steroid treatment to relapse.

The relapse rate of AIP was significantly higher in patients with persistent elevation of serum IgG4 levels (30%, 34/115) than in those with normalised serum IgG4 levels (10%, 7/69; p = 0.003). Although serum IgG4 levels fluctuated by >30 mg/dl in 94 (55%) of 172 patients during maintenance treatment, re-elevation of serum IgG4 levels was detected in 37 (69%) of 54 patients who relapsed during maintenance treatment.

Of the 89 relapsed patients, 83 (93%) of 89 received steroid re-treatment (prednisolone: 60 mg/day (n = 4, 5%), 40 mg/day (n = 19, 23%), 30 mg/day (n = 39, 47%), 20 mg/day (n = 9, 11%) and others). Steroid re-treatment was effective in 91 (97%) of 94 relapsed patients. Of the 77 patients initially managed without steroid treatment relapses occurred in 32 (42%), and the relapses were treated with steroid with a 100% response rate.

Steroid-related complications

After steroid treatment, mildly or moderately worse glucose tolerance occurred in several patients, but they could be controlled by oral antidiabetic medication or insulin injection. Osteoporosis developed in 10 patients, in whom compression fractures of lumbar vertebrae (n = 5) and avascular necrosis of the femoral head (n = 3) occurred. They were treated with reduction of dosage or cessation of medication. Pneumonia occurred in 3 patients, and they were treated with antibiotics. There were no deaths attributable to complications of steroid treatment.

Discussion

In the analysis of 563 patients with AIP, 82% received oral steroid treatment. The remission rate of patients with AIP was significantly higher in those who received steroid treatment (98%) than in those who did not receive steroid treatment (74%). The relapse rate was significantly lower in patients who received steroid treatment than in those not given steroid treatment. Therefore, the administration of oral steroid appears to be standard treatment for inducing remission in AIP. However, it is most important to distinguish AIP from pancreatic cancer before starting steroid treatment. Facile use of steroids for cases in which the diagnosis of AIP is questionable should be prohibited.15 16

The indications for steroid treatment in patients with AIP are thought to be symptoms such as obstructive jaundice due to sclerosing cholangitis, abdominal pain and hydronephrosis due to associated retroperitoneal fibrosis. DM is often (67%17 to 76%18 of cases) observed in patients with AIP. Pancreatic exocrine function is also impaired in 88%19 to 91%20 of AIP patients. However, as pancreatic exocrine or endocrine dysfunction improves in some patients with AIP after steroid treatment,17 19 20 steroid treatment may be indicated in patients showing diffuse enlargement of the pancreas, even if they are asymptomatic.

In patients with DM, blood glucose levels should be controlled using insulin before starting steroid treatment. The major presenting complaint of patients with AIP is obstructive jaundice due to associated sclerosing cholangitis (65%18 to 86%21 of cases). As steroid treatment may trigger or worsen cholangitis, jaundice is usually managed by endoscopic or transhepatic biliary drainage in patients with obstructive jaundice (usually total bilirubin ⩾3 mg/dl) due to associated sclerosing cholangitis before steroid administration. In the literature, most patients with AIP were treated with 30 or 40 mg/day as the initial prednisolone dose.5 6 7 8 9 10 There was no relationship between the period necessary to achieve remission and the initial prednisolone dose. In Japan, 30 mg/day is usually used as an initial dose in patients with standard body weight (∼50 kg; 0.6 mg/kg) and 40 mg/day is used in larger patients (body weight >60 kg; 0.67 mg/kg). Therefore, we would recommend that a general initial dose of prednisolone should be 0.6 mg/kg/day.

Since pancreatic enlargement begins to improve from 1 to 2 weeks after medication,10 12 morphological and serological evaluation for effectiveness of steroid treatment should be performed 2 weeks after starting steroid treatment. A poor response to steroid treatment should raise the possibility of pancreatic cancer and the need for re-evaluation of the diagnosis. When steroid treatment is effective, the dose is tapered by 5 mg every 1–2 weeks until the dose reaches 15 mg/day, while carefully monitoring the patient’s symptoms, as well as the biochemical, serological and imaging findings. After that, the dose is tapered more gradually, and the amount of steroid is reduced to a maintenance dose over a period of 3–6 months.

Relapse occurred in 24% of patients with AIP treated with steroid, and it occurred within 6 months after medication in 32%, within 1 year in 56% and within 3 years in 92%. Although it still remains unclear what are useful predictive findings for relapse, it has been reported that markedly elevated serum IgG4 levels and the presence of bile duct stenosis were predictive factors for relapse of AIP.8 To our knowledge, patients complicated with extrapancreatic lesions such as stenosis of the proximal extrahepatic or intrahepatic bile duct or retroperitoneal fibrosis seem to take longer to achieve remission. In the present study, the patients with elevation of serum IgG4 levels during remission showed persistent abnormalities of the pancreatic and/or bile ducts and more frequent relapses than in those with normalised serum IgG4 levels. Furthermore, in 69% of relapsed patients during maintenance treatment, re-elevation of serum IgG4 levels was detected. Serum IgG4 levels at remission and during follow-up may accordingly be useful to predict or detect relapse earlier.

To prevent relapse, maintenance treatment (5 mg/day) is recommended in almost all patients treated with steroid for at least about 6 months. However, as patients with AIP are typically elderly and are at high risk of developing steroid-related complications, such as osteoporosis, DM and pneumonia, cessation of the medication should be tried. In patients showing complete improvement of cholangiopancreatogram, 1 year after initial administration of steroid, maintenance treatment can be withdrawn. Stopping maintenance treatment should be planned within at least 3 years, because of a lower relapse rate after 3 years. After stopping medication, patients should be followed-up for relapse of AIP. In most recurrent cases, re-administration or dose-up of steroid is effective. In these cases, longer maintenance treatment is necessary to prevent repeated relapse. Therefore, we think that it is now necessary to reduce the relapse rate. Careful, long-term follow-up is also necessary, since pancreatic cancer developed in three patients 3, 3.5 and 5 years after onset of AIP, respectively. On the other hand, 23% of patients with AIP relapsed despite maintenance treatment. We need to identify more useful alternative approaches to maintain remission.

In the Japanese diagnostic criteria for AIP,15 seronegative AIP cases without histological examination cannot be diagnosed as AIP even if they fulfil the imaging criterion. Kim et al22 reported that a steroid trial was useful for differentiating seronegative AIP from pancreatic cancer, based on marked improvement of pancreatic ductal narrowing, which is evident as early as 2 weeks after the beginning of steroid treatment. When response to steroid treatment is added to the diagnostic criteria, the diagnostic sensitivity is increased. However, we are concerned that the facile use of steroid trials will result in delaying pancreatic cancer surgery, which could lead to cancer progression in some cases. Therefore, as described in the Asian diagnostic criteria,11 a steroid diagnostic trial is not generally recommended and it should only be performed by pancreatologists with extreme caution in limited cases after a negative investigation for pancreatic cancer, including endoscopic ultrasound-guided fine needle aspiration (EUS-FNA; box 1).

In conclusion, oral steroid is a standard treatment for AIP. Indications for steroid treatment in patients with AIP are symptoms such as obstructive jaundice due to sclerosing cholangitis, abdominal pain, and hydronephrosis due to associated retroperitoneal fibrosis. We recommend that the initial dose of prednisolone be 0.6 mg/kg/day, and that it be reduced to a maintenance dose over a period of 3–6 months. To prevent relapses, continued maintenance treatment with low-dose prednisolone for 6 months to 3 years is recommended, but it dose not eliminate relapse entirely.