AIM: To identify clinical and pathological differences between serum immunoglobulin G4 (IgG4)-positive (SIP) and IgG4-negative (SIN) type 1 autoimmune pancreatitis (AIP) in South Korea.

METHODS: AIP was diagnosed by the international consensus diagnostic criteria. The medical records and pathology were retrospectively reviewed and IgG4-positive cells were counted in a high power field (HPF). Type I AIP was defined as a high serum level of IgG4 or histological finding. SIN type 1 AIP was defined as a histological evidence of type 1 AIP and a normal serum IgG4 level. The clinical and pathological findings were compared between the two groups. The analysis was performed using Student’s t test, Fischer’s exact test and Mann-Whitney’s U test. A P value of < 0.05 was considered statistically significant. As repeated comparison was made, P values of less than 5% (P < 0.05) were considered significant.

RESULTS: Twenty five patients with definite type 1 AIP (19 histologically and six serologically diagnosed cases) were enrolled. The mean tissue IgG4 concentrations were significantly higher in SIP than SIN group (40 cells per HPF vs 18 cells per HPF, P = 0.02). Among eight SIN patients, the tissue IgG4 concentrations were less than 15 cells per HPF in most of cases, except one. The sensitivity of serum IgG4 was 68% (17 SIP and eight SIN AIP). Other organ involvement was more frequently associated with SIP than SIN AIP (59% vs 26%, P = 0.016). However, the relapse rate and diffuse swelling of the pancreas were not associated with serum IgG4 level. The concentrations of IgG4-positive cells per HPF were higher in SIP than SIN AIP (40 vs 18, P = 0.02).

CONCLUSION: The sensitivity of serum IgG4 was 68% in type 1 AIP. High serum IgG4 level was associated with other organ involvement and tissue IgG4 concentration but did not affect the relapse rate in type 1 AIP.

Autoimmune pancreatitis (AIP) is a type of chronic pancreatitis with irregular narrowing of the pancreatic duct and systemic fibroinflammatory disease. AIP is characterized by a remarkable response to steroid therapy. According to a multicenter nationwide study in Korea, the prevalence of AIP was 2.0% among 814 patients with chronic pancreatitis[1]. An early report from Japan that proposed the term lymphoplasmacytic sclerosing pancreatitis (LPSP) described some specific morphological features of AIP, such as diffuse lymphoplasmacytic infiltration with marked interstitial fibrosis and obliterative phlebitis[2].

The Japan Pancreas Society proposed diagnostic criteria for the first time in 2002, and the characteristic features of AIP were defined as the elevation of serum immunoglobulin G4 (IgG4) and LPSP on pathology[3]. However, emerging evidence suggests the presence of two AIP types that have different clinical profiles and outcomes. In 2003, a Mayo clinic group found two distinct histological patterns, which were designated LPSP and idiopathic duct-centric chronic pancreatitis (IDCP)[4]. IDCP was characterized by inflammatory infiltrates that were denser in the lobules than in interlobular fibrotic areas.

Recently, the expert panel in the international consensus study has agreed that there are two distinct histopathological types of AIP[5]. Type 1 AIP has dense periductal lymphoplasmacytic infiltrate with storiform fibrosis and obliterative phlebitis, whereas type 2 is distinguished from type 1 by granulocyte epithelial lesions, less prominent lymphoplasmacytic infiltrate, and less prominent storiform fibrosis. Recently, international consensus diagnostic criteria (ICDC) for AIP were developed based on the agreement of an international panel of experts and ICDC include both types 1 and 2 AIP[6].

According to the ICDC, the radiological imaging and the response to steroids are common features of both types 1 and 2 AIP. However, typical serological abnormalities, such as serum IgG4 elevation and other organ involvement, can be seen only in type 1. Thus, for a definitive diagnosis of type 2 AIP, histological confirmation is always necessary. Type 2 AIP is associated with inflammatory bowel disease and affects younger patients without a gender predilection[7]. Both types of AIP respond to steroid very well, but type 2 AIP has a lower relapse rate than type 1 AIP[7].

Although elevation of serum IgG4 is the one of the characteristic features in type 1 AIP, the sensitivity of serum IgG4 is variable. The initial Japanese study reported that the sensitivities of IgG4 were 90.9%[8]. However, other studies reported the sensitivity of IgG4 as approximately 70.0%[9-12]. The problem of the previous studies was that there was no clear classification of AIP type because the study was performed before the concept of type 2 AIP was established. If the study population had included more type 2 AIP, the sensitivity of IgG4 would have been low. However, a recent multicenter study also showed that the sensitivity was only 63.0% among histologically proven type 1 AIP[13]. Type 1 AIP is considered as the pancreatic manifestation of IgG4-related systemic disease in which tissue infiltration of IgG4-positive plasma cells is a characteristic feature[14-16]. However, the reason for the variable level of serum IgG4, the relation between serum level and tissue concentration of IgG4, and clinical significance of serum IgG4 level in type 1 AIP is unknown and remains an interesting issue.

The aim of this study was to find clinical and pathological differences between serum IgG4-positive (SIP) and serum IGg4-negative (SIN) type 1 AIP in Korea.

IgG4-related disease was recognized as a systemic disease since 2003[17] and AIP was proposed as one of the IgG4-related sclerosing diseases in 2006[18]. Since then, two histopathological subtypes, LPSP and IDCP, have been recognized[19]. Type 1 AIP is now considered as the pancreatic manifestation of an IgG4-related systemic fibroinflammatory diseases involving the salivary gland, bile duct, and retroperitoneum. Thus, serum IgG4 is a useful marker for the diagnosis of type 1 AIP and most diagnostic criteria of AIP include serum IgG4 elevation as one of the criteria[6,9,20]. However, the sensitivity of serum IgG4 is variable and different among countries. If the study population includes more type 2 AIP, the sensitivity of IgG4 may be low because serum IgG4 is not usually elevated in type 2 AIP. A recent international multicenter study, which enrolled 713 patients with AIP from eight countries, reported that sensitivity of serum IgG4 was only 63% among 204 patients with histologically proven type 1 AIP[13]. The relatively low sensitivity of serum IgG4 can make the diagnosis of AIP in the clinical setting confusing. In our study, four patients with segmental type AIP underwent unnecessary surgical resection.

We questioned why the serum IgG4 test was not sufficiently sensitive if type 1 AIP is a type of IgG4-related systemic disease. Therefore, we conducted our study to analyze the clinical and pathological differences between SIP and SIN type 1 AIP. Unfortunately, there have been few studies concerning the normal serum IgG4 AIP[21,22]. One study included 58 AIP patients including 13 normal serum IgG4 AIP[21] but histology was available in only 14 cases (six cases among 13 SIN AIP). Another study included 27 patients with AIP, including seven SIN AIP[22]. Histology was not available in any cases because endoscopic ultrasonography guided fine needle aspiration was performed in 26 cases using 22 gauge needle, not to diagnose AIP, but to exclude pancreatic malignancy. Thus, it could not be determined that all of the enrolled patients were really type 1 AIP in both studies. To exclude possible type 2 AIP, our study enrolled 19 patients with histologically proven type 1 AIP and six patients who were clinically diagnosed as type 1 AIP with elevated serum IgG4 levels. Of course, there is the possibility of type 2 AIP despite the elevated serum IgG4 level among six patients, because serum IgG4 elevation was detected in 23% among 47 patients with histologically proven type 2 AIP according to a recent study[13]. However, the possibility might be very low, because the serum IgG4 level was relatively high (213, 250, 279, 300, 4000 mg/dL), except in one case (146 mg/dL) and type 2 AIP was reported to be relatively rare in Asian countries, including South Korea[13]. In addition, five patients had other organ involvement, which is rarely be seen in type 2 AIP[7].

The surgical resection rate was higher in the SIN than the SIP group. One reason could be a difficult diagnosis of AIP. If the lesion is in the body/tail and serum IgG4 is normal, the clinicians would not suspect the possibility of AIP and would not hesitate to perform a surgical resection. Another reason might be selection bias of this study, because we excluded eight patients with normal serum IgG4 and no histology. The eight patients received steroid treatment and their steroid responsiveness was 100%. One patient experienced a relapse.

In this study, the clinical profiles of type 1 AIP were similar to another recent multicenter study including 327 Asian patients[12]. The higher mean age (over 60 year), male predominance, common other organ involvement, especially sclerosing cholangitis and frequent relapse, are common features of Asian patients of AIP that are similar to our study. Interestingly, the important clinical difference between SIP and SIN type 1 AIP was the frequency of other organ involvement. Other organ involvement was significantly more common in SIP than SIN type 1 AIP (59% vs 26%). Only one patient among the SIN group had retroperitoneal fibrosis. This result implies that other organ involvement can affect the serum IgG4 level. Mikulicz’s disease refers to idiopathic symmetrical swelling of the lacrimal, submandibular gland and is an IgG4-related systemic disease. A recent study reported that the serum IgG4 level is very high (894 mg/dL) in Mikulicz’s disease and significantly higher in patients with extrasalivary gland involvement[23]. More frequent other organ involvement in our SIP type 1 AIP is similar to the results of previous studies[21,22,24].

Another reason for variable serum IgG4 levels may be the number of IgG4-positive plasma cells in the tissue. As expected, the mean tissue IgG4 concentration was significantly low in SIN compared with SIP type 1 AIP. All patients in the SIP group had high IgG4 concentrations (over 25 cells per HPF), except one case (15 cells per HPF). However, the patients in the SIN group had very low IgG4 concentrations (below 15 cells per HPF), except one case (80 cells per HPF). The data might lead us to conclude that the IgG4 concentration of pancreatic tissue can influence the sensitivity of serum IgG4 in type 1 AIP. However, the serum level of IgG4 had no correlation with tissue IgG4 concentration in SIP type 1 AIP. Thus the serum level may be influenced by not only tissue concentration, but also other factors, such as the size of the involved pancreas and other organ involvement. We think that this is the first study to investigate the correlation between serum and tissue IgG4 concentration in type 1 AIP.

Other possible clinical roles of serum IgG4, other than as a diagnostic marker, are uncertain and are a interesting issue in type 1 AIP. The clinical use of serum IgG4 may be relevant in three settings: monitoring of therapy, monitoring for disease relapse and prediction of relapse. The large multicenter study in Japan reported that IgG4 levels failed to normalize in 115/182 (63%) of the patients treated with steroids[25]. The study suggested that serial IgG4 levels are helpful in identifying early relapse. However, only 30% of patients with persistent IgG4 elevation relapsed, whereas relapse was also seen in 10% of patients with normal IgG4 levels. The results regarding the value of initial serum IgG4 levels in predicting relapse vary among studies, some reporting higher relapse rate in patients with elevated serum IgG4 levels[22,26], whereas others failed to observe any association[7,27-29]. To clarify the role of serum IgG4 in predicting relapse, type 2 AIP should be excluded in the normal serum IgG4 group, because type 2 AIP is known for rare relapse[7]. The positive study might include some patients with type 2 AIP. In our study, the relapse rate was similar between the two groups of type 1 AIP. Thus, our data supports the view that initial serum IgG4 levels cannot predict relapse in type 1 AIP.

In conclusion, the sensitivity of serum IgG4 was not sufficient (68%) for definite diagnosis of type 1 AIP. The demographic findings were similar between SIP and SIN type 1 AIP, but other organ involvement was significantly more common in SIP than SIN type 1 AIP. High serum IgG4 level was associated with other organ involvement and tissue IgG4 concentration, but did not affect the relapse rate in type 1 AIP.