Comparison of clinical and histopathological features between serum IgG4-negative and serum IgG4-positive patients with type 1 autoimmune pancreatitis: A single-center, retrospective study

INTRODUCTION

Autoimmune pancreatitis (AIP), first proposed by Yoshida et al.^[1], is classified into types 1 and 2 and, recently, “type 3 (unspecified).”^[2,3] Type 1 AIP, also referred to as lymphoplasmacytic sclerosing pancreatitis, is a key manifestation of immunoglobulin G4-related disease (IgG4-RD), defined by periductal lymphoplasmacytic infiltration, storiform fibrosis, occlusive phlebitis, and often elevated serum IgG4.^[2,4,5] It is more common in Asians than Europeans/ Americans,^[6] and while AIP accounts for only 6% of chronic pancreatitis,^[7,8] its prevalence has risen (e.g., the incidence of type 1 AIP in Japan has increased from 0.8 to 3.1/100,000, 2007-2016),^[9] warranting clinical attention.

The 2011 International Consensus Diagnostic Criteria (ICDC) integrates serum IgG4 levels, imaging, extrapancreatic involvement, pathology, and glucocorticoid response for AIP diagnosis.^[10] Long recognized that as the core serological biomarker for type 1 AIP, serum IgG4 is a subtype of IgG: It maintains immune homeostasis by participating in immune regulation and suppressing excessive inflammation, and its abnormal elevation is closely linked to the immune-mediated inflammatory mechanism of type 1 AIP.^[11] A study has shown that the sensitivity and specificity of serum IgG4 for type 1 AIP diagnosis are 72 and 93%, respectively.^[12] However, 10-30% of type 1 AIP patients have normal serum IgG4 levels (<135 mg/dL), defined as serum IgG4-negative (IgG4-N).^[13,14] In contrast, patients with serum IgG4 concentrations ≥135 mg/dL are classified as serum IgG4-positive (IgG4-P). These serum IgG4-N patients face significant diagnostic dilemmas: their misdiagnosis rate is notably higher. In addition, for those with pancreas-limited disease, the rate of unnecessary surgical interventions is significantly higher than that in serum IgG4-P patients, due mainly to the difficulty in distinguishing AIP from pancreatic cancer.^[15] Studies on differences between serum IgG4-N and serum IgG4-P patients are conflicting (e.g., recurrence rate discrepancies).^[13,14] Pathology is critical for diagnosis (ICDC requires >10 IgG4-P cells/high-power field [HPF]),^[10] with most tissue obtained through endoscopic ultrasound-guided fine-needle aspiration/biopsy (EUS-FNA/FNB).^[16,17] However, existing research focuses on EUS-FNA/FNB method/needle type effects on diagnosis,^[16-19] with few studies comparing histopathological features (e.g., tissue IgG4 count, confirmation rate) between serum IgG4-N and serum IgG4-P patients or exploring serum IgG4-tissue IgG4 correlation,^[15] a gap limiting type 1 AIP typing and management.

This study retrospectively analyzed 78 type 1 AIP patients (Changhai Hospital Affiliated with Naval Medical University, Dec 2019-2023) stratified by serum IgG4 (serum IgG4-N: <135 mg/dL; serum IgG4-P: ≥135 mg/dL). Clinical baseline, laboratory/imaging data, diagnosis-treatment strategies, and histopathology were compared between groups, and serum IgG4-tissue IgG4 correlations were explored. The goal was to clarify the association of serum IgG4 typing with type 1 AIP clinicopathological features, providing a basis for accurate diagnosis and stratified management.

MATERIAL AND METHODS

RESULTS

Distribution of different clinical features in the serum IgG4-N and serum IgG4-P groups

There were no significant differences in gender distribution between the serum IgG4-N group (serum IgG4 <135 mg/dL) and the serum IgG4-P group (serum IgG4 ≥135 mg/dL) (P = 0.724) [Figure 1a(a)].

Close

Figure 1: Distribution of clinical features between serum IgG4-N (<135 mg/dL) and serum IgG4-P (≥135 mg/dL) groups (a) Distribution of categorical clinical features (including baseline characteristics, imaging findings, diagnosis-treatment strategies and histopathological features) in the serum IgG4-N and serum IgG4-P groups (percentage stacked bar charts; P values of intergroup comparisons are labeled in each subplot) (b) Distribution of quantitative clinical baseline characteristics and laboratory indicators in the serum IgG4-N and serum IgG4-P groups (box plots; P values of intergroup comparisons are labeled in each subplot). IgG4: Immunoglobulin G4, IgG4-N: IgG4-negative, IgG4-P: IgG4-positive, PP: Pancreatic parenchyma, PD: Pancreatic duct, MT: Malignant tumor, EUS-FNA: EUS-guided fine-needle aspiration, EUS-FNB: EUS-guided fine-needle biopsy, HPF: High-power field.

Export to PPT

In terms of clinical symptoms, abdominal discomfort (P = 0.758) [Figure 1a(b)] and weight loss (P = 0.210) [Figure S1a(a)] were more common in the serum IgG4-N group; in contrast, obstructive jaundice (P = 0.230) [Figure 1a(c)], abdominal discomfort combined with obstructive jaundice (P = 1.000) [Figure S1a(b)], and diarrhea (P = 0.567) [Figure S1a(c)] were more prevalent in the serum IgG4-P group. The proportion of patients with diabetes was higher in the serum IgG4-P group (P = 0.211) [Figure S1a(d)]; however, the prevalence of smoking history was slightly higher in the serum IgG4-N group (P = 0.518), while there were no significant differences in the incidence of hypertension (P = 1.000) or alcohol consumption history (P = 1.000) between the two groups [Figure S1a(e-g)]. A family history of pancreatic diseases was rare in both groups (P = 0.244), which might be attributed to the small sample size [Figure S1a(h)].

SUPPLEMENTARY FILE

Download

In terms of imaging findings, the distribution of imaging-related malignant tumor signs (imaging malignant tumor [MT]) was comparable between the two groups (P = 0.286) [Figure 1a(d)]. With respect to CT/MRI findings, focal masses were more prominent in the serum IgG4-P group (P = 0.880) [Figure S1a(i)], whereas the proportion of PD stenosis was greater in the serum IgG4-N group (P = 0.289) [Figure S1a(j)].

EUS was performed more often in the serum IgG4-P group (P = 0.010) [Figure S1a(k)], and the distribution of EUS punctures was comparable between the two groups (P = 0.010) [Figure S1a(l)]. Diffuse pancreatic enlargement and focal masses were more prominent in the serum IgG4-P group (P = 0.043) [Figure 1a(e)]. There were significant differences in the distribution of PD (EUS-PD) signs between the two groups (P = 0.003) [Figure 1a(f)]. The proportions of bile duct stenosis (P = 0.047; Figure 1a(g)) and kidney involvement (P = 0.567; Figure 1a(h)) were greater in the serum IgG4-P group, whereas the distribution of retroperitoneal fibrosis was similar between the two groups (P = 1.000) [Figure 1a(i)].

The surgical rate was higher in the serum IgG4-N group (P = 0.010) [Figure 1a(j)]. In contrast, EUS-FNA/FNB was more commonly used in the serum IgG4-P group (P = 0.024) [Figure 1a(k)]. Both groups preferred 22G needles for puncture, whereas 19G and 25G needles were less frequently used (P = 0.013) [Figure 1a(l)]; this pattern was also observed in the supplementary analysis of needle type distribution (P = 0.042) [Figure S1a(m)]. In addition, the histopathological confirmation rate was significantly higher in the serum IgG4-P group (P = 0.007) [Figure 1a(m)]; and the proportion of patients with high-level pathological evidence (Level 1 and Level 2) was also significantly higher in this group (P = 0.024) [Figure 1a(n)]. Furthermore, a higher proportion of patients in the serum IgG4-P group had >10 IgG4-P cells/HPF (P = 0.083) [Figure 1a(o)]. After further stratifying the IgG4-P cell counts into more detailed subgroups (0–10, 10–20, 20–50, >50 cells/HPF), the trend of intergroup differences in cell count distribution reflected by Figure 1a(p) was consistent with that of Figure 1a(o) (P = 0.221) [Figure 1a(p)].

No significant differences in age distribution were observed between the two groups, with both groups predominantly composed of middle-aged and elderly individuals (P = 0.495) [Figure 1b(a)]. Liver function indicators (TB, ALT, AST) were similar between the serum IgG4-N and serum IgG4-P groups (P = 0.334, P = 0.784, P = 0.903, respectively), while elevated CA19-9 levels were more common in the serum IgG4-N group (P = 0.035) [Figure 1b(b-e)].

Correlation between serum IgG4 levels and tissue IgG4-P plasma cell counts

We further analyzed the correlation between serum IgG4 levels and tissue IgG4-P plasma cell counts. First, serum IgG4 levels were categorized into serum IgG4-N (serum IgG4 <135 mg/dL) and serum IgG4-P (serum IgG4 ≥135 mg/dL) groups; no obvious correlation was observed between the groups and the tissue IgG4-P plasma cell count [Figure 2a].

Close

Figure 2: Correlation between serum IgG4 and tissue IgG4-positive cell counts (a) Scatter plot with a fitted trend line (serum IgG4-negative group [<135 mg/dL], serum IgG4-positive group [≥135 mg/ dL]): correlation between the two indices (Spearman’s rho = 0.113, P = 0.322). (b) Scatter plot with a fitted trend line: visualization of the overall correlation between the two indices (Spearman’s rho = 0.113, P = 0.322). IgG4: Immunoglobulin G4; HPF: High-power field.

Export to PPT

Statistical analyses were subsequently performed to explore the association between serum IgG4 levels and tissue IgG4-P plasma cell counts. Neither the serum IgG4 level nor the tissue IgG4-P plasma cell count conformed to a normal distribution [Figures S2a and S2b]. Spearman correlation analysis and linear regression analysis were therefore conducted to investigate their relationships. The Spearman correlation coefficient (rho) was 0.113 (P = 0.322), indicating an extremely weak positive correlation between the serum IgG4 level and the tissue IgG4-P plasma cell count, which was not statistically significant. Linear regression analysis revealed that the regression coefficient for serum IgG4 was 0.004 (P = 0.538), further confirming that the serum IgG4 level could not significantly predict changes in the tissue IgG4-P plasma cell count. These observations indicated that there was no clear linear association between the two indicators [Table 7].

Table 7. Results of Spearman correlation analysis and linear regression for serum IgG4 and tissue IgG4-positive plasma cell count.

Analysis method	Indicator	Correlation/ Regression Coefficient	P	Significance
Spearman Correlation analysis	Correlation Coefficient (rho)	0.113	0.322	Not Significant
Linear regression	regression Coefficient (IgG4)	0.004	0.538	Not Significant

IgG4: Immunoglobulin G4

The distribution of serum IgG4 levels across the two groups (<135 mg/dL and ≥135 mg/dL) and their relationship with tissue IgG4-P plasma cell counts are shown in a scatter plot in Figure 2b. However, the tissue IgG4-P plasma cell count did not show an obvious trend with increasing serum IgG4 levels, which was consistent with the statistical results shown in Table 7. These findings further verified the lack of a clear association between serum IgG4 levels and tissue IgG4-P plasma cell counts from a visual perspective.

DISCUSSION

In this study, clinical data from 78 patients with type 1 AIP admitted to Changhai Hospital Affiliated with Naval Medical University were analyzed, and the serum IgG4 level was used as the core stratification index. We systematically compared clinical characteristics, imaging findings, diagnosis-treatment strategies, and histopathological features between the serum IgG4-N (serum IgG4 <135 mg/dL) and serum IgG4-P (serum IgG4 ≥135 mg/dL) groups and explored the association between serum IgG4 levels and the number of tissue IgG4-P plasma cells.

Notably, this study fills a critical gap in the diagnosis and treatment of type 1 AIP in Chinese populations. Existing international studies include Japanese studies (including nationwide surveys^[14] and prospective multicenter trials^[23]), Korean single-center analyses,^[15] Western cohort studies on pancreatic damage,^[24] and discussions on relapse risk factors,^[25] all have confirmed the value of serum IgG4 stratification, but their findings have limitations for Chinese clinical practice: Katz et al.^[24] focused on pancreatic damage (exocrine/endocrine insufficiency) in Western populations and lacked data on serum IgG4-stratified puncture strategies; Jaber et al.^[25] highlighted the association of high serum IgG4/IgA with relapse but lacked an analysis of IgG4-negative AIP diagnosis and Chinese-specific misdiagnosis factors. Domestic studies on Chinese populations have focused mainly on differential diagnosis from pancreatic cancer,^[26] stratification by pancreatic enlargement type^[27] or single-dimensional correlation between serum IgG4 and clinical features^[28] and lack the systematic integration of “serum IgG4 stratification+EUS puncture+serum-tissue IgG4 correlation” as well as analyses of CA19-9-related misdiagnosis clues. The present study, which is based on 78 Chinese type 1 AIP patients, is the first to address these core clinical dilemmas, providing targeted evidence for Chinese clinical practice.

Compared with the serum IgG4-N group, the serum IgG4-P group had a higher incidence of obstructive jaundice (50.85% vs. 31.58%, P = 0.230) and bile duct stenosis (55.93% vs. 26.32%, P = 0.047), which aligns with findings from Japanese, American, and Chinese scholars and relevant international studies.^{[14,24,25,28-30]} A nationwide study by Japanese scholar Sano et al.^[14] revealed that type 1 AIP patients with serum IgG4 ≥135 mg/dL had a 40.7% biliary involvement rate, which was higher than the 30.6% rate in serum IgG4-N patients; Zhou et al.^[28] and Katz et al.^[24] also confirmed that high serum IgG4 levels are associated with more active immune status, higher biliary involvement, and increased pancreatic damage (exocrine/endocrine insufficiency), collectively supporting our results. The serum IgG4-N group also had a significantly higher median level of the tumor marker CA19-9 (32.14 U/mL vs. 5.97 U/mL in the serum IgG4-P group, P = 0.035), which is consistent with the findings of Chinese scholars.^[26] Chong et al.^[31] further verified that elevated CA19-9 in IgG4-RD is linked to the disease itself rather than to serum IgG4 interference. These findings supplement the high misdiagnosis rate of focal AIP (47.06%) reported by Han et al.^[27] and address the diagnostic challenge of serum IgG4-N AIP highlighted by Jaber et al.,^[25] clarifying that “serum IgG4 negativity + CA19-9 elevation” is the core driver of misdiagnosis in Chinese patients, especially those with focal lesions. Owing to the lack of serological diagnostic support, the serum IgG4-N group had a higher proportion of patients with suspected malignancy (MT) by imaging (47.37% vs. 30.51% in the serum IgG4-P group). Since AIP and pancreatic cancer have overlapping imaging features, clinicians often use surgery to rule out malignancy, which explains why the serum IgG4-N group had a significantly higher surgical rate (47.37% vs. 15.25% in the serum IgG4-P group, P = 0.010). This finding is consistent with the conclusions of the Korean scholar Paik et al.^[15] (75% vs. 26%, P = 0.018) and the Japanese scholars Sano et al.^[14] (9.2% vs. 4.6%) and Katz et al.^[24] (8% pancreatic resection rate in AIP patients), confirming that serum IgG4-N type 1 AIP has a globally higher surgical rate, which is particularly prominent in Chinese populations due to the high misdiagnosis risk.

For the serum IgG4-P group, serological indicators suggested the presence of AIP, so clinicians tended to use EUS to confirm typical imaging features and perform EUS-FNA/ FNB for histopathological evidence, avoiding surgical trauma. In contrast, the serum IgG4-N group often presented seronegativity and PD strictures; their imaging results overlapped strongly with those of pancreatic cancer and chronic pancreatitis, making noninvasive diagnosis difficult and leading to higher surgical rates. Notably, both groups predominantly used 22G needles for EUS puncture (60% in the serum IgG4-N group; 78% in the serum IgG4-P group). The puncture sites were concentrated in the pancreatic body or head, with no significant intergroup differences (P > 0.050). This finding was likely because 22G needles balance tissue sampling quality and safety, reducing the risk of complications such as bleeding and pancreatic fistula, which is consistent with international studies confirming the superiority of 22G needles in terms of diagnostic yield and safety.^[1,19,23] More importantly, our results highlighted that the applicability of 22G needles was not affected by serum IgG4 levels (no significant intergroup differences in puncture effectiveness, P > 0.050). These findings are consistent with those of international studies,^[19,23,32] which confirm the superiority of 22G needles in terms of diagnostic yield and safety. This observation fills the gap in the “serum IgG4-stratified puncture strategy” in Chinese clinical practice^[27] and complements the diagnostic limitations mentioned by Jaber et al.^[25] (lack of high-sensitivity autoantibodies for serum IgG4-N AIP), as domestic studies^[28,33] and Western studies^[24] have not addressed needle type selection for different serum IgG4 subtypes.

As a key diagnostic criterion for type 1 AIP, the association between histopathology and serum IgG4 typing was a core focus of the present study. The serum IgG4-P group had a significantly higher histopathological confirmation rate (71.19% vs. 36.84% in the serum IgG4-N group, P = 0.015), a higher median number of tissue IgG4-P cells (36 cells/HPF vs. 8 cells/HPF, P = 0.035), a lower proportion of 0 cells/HPF (16.95% vs. 47.37%, P = 0.013), and a lower proportion of suspected type 1 AIP (13.56% vs. 47.37%, P = 0.004). This finding suggested a potential link between serum IgG4 levels and tissue inflammatory infiltration: the serum IgG4-P group had more significant systemic immune activation, leading to more prominent pancreatic pathological features (e.g., dense IgG4-P plasma cell infiltration, storiform fibrosis, and occlusive phlebitis) that make it easier to meet the ICDC pathological confirmation standards (Level 1 or Level 2).^[14,15] Consistent with the findings of Katz et al.,^[24] the serum IgG4-P patients in our study also showed more extensive organ involvement, which might be associated with greater disease severity. We also observed “serum-tissue IgG4 dissociation:” 42.11% (8/19) of serum IgG4-N patients still had tissue IgG4 >10 cells/HPF, and 38.98% (23/59) of serum IgG4-P patients had tissue IgG4 ≤10 cells/HPF, which is consistent with results from Sano et al.,^[14] Paik et al.,^[15] and Katz et.al.,^[24] but was quantified for the 1^st time in Chinese populations. This finding emphasizes that even when serum IgG4 is negative, EUS-FNA/FNB should be performed if type 1 AIP is clinically suspected (e.g., EUS showing diffuse pancreatic enlargement) to avoid missed diagnosis, addressing the unmet diagnostic need for serum IgG4-N AIP mentioned by Jaber et.al.;^[25] conversely, patients with elevated serum IgG4 require pathological confirmation to avoid misdiagnosis of other IgG4-RD as pancreatic involvement.^[34]

Through Spearman correlation analysis and linear regression, we further confirmed that there was no significant association between serum IgG4 levels and tissue IgG4-P plasma cell counts (rho = 0.113, P = 0.322; regression coefficient = 0.004, P = 0.538), which is supported by studies from different countries and regions.^[15,24,28] Unlike international studies that have only qualitatively described this dissociation,^[14,15] our study provides quantitative data in Chinese populations, negating the simplified logic of “using serum IgG4 levels to replace histopathological examination.” This finding reinforces that type 1 AIP diagnosis must follow the “multidimensional integration” principle of the ICDC, combining serology, imaging, pathology, and treatment response, which is especially critical for Chinese clinical practice with high misdiagnosis rates^[26,27] and addressing the need for comprehensive diagnostic strategies highlighted by Jaber et.al.^[25]

This study has three key innovations. First, it is the first systematic analysis of serum IgG4 stratification in Chinese type 1 AIP patients, complementing domestic studies stratified by enlargement type^[27] and Western studies focusing on pancreatic damage^[24] to form a “subtype + morphology + puncture strategy” triple-layer diagnostic framework. Second, it quantifies the “serum-tissue IgG4 dissociation” phenomenon in Chinese patients, providing more actionable diagnostic guidance than international qualitative descriptions^[14,15,24] and addressing the diagnostic challenge of serum IgG4-N AIP mentioned by Jaber et.al.^[25] Third, it confirms that 22G needles are the preferred choice for EUS puncture in Chinese patients regardless of serum IgG4 status, adapting international needle selection recommendations^[19,32] to the needs of Chinese primary hospitals (prioritizing safety and operability), which is not covered in Western cohort studies.^[24]

The strengths of this study are as follows. This study focuses on serum IgG4 stratification in type 1 AIP, a topic with limited data in Chinese populations, and it systematically clarifies differences in clinical, imaging, and pathological features between groups, as well as the lack of correlation between serum IgG4 and tissue IgG4-P cells, providing practical guidance for clinical diagnosis. However, as a single-center retrospective study, this study has several limitations. First, owing to the rarity of type 1 AIP, the small sample size (78 cases) may have prevented some statistical indicators from reaching significance, and single-center data may introduce selection bias; future multicenter prospective studies are needed to verify the results. Second, the lack of long-term follow-up data precluded the ability to evaluate the association between serum IgG4 typing and patient prognosis (e.g., recurrence rate, glucocorticoid dependence rate, incidence of pancreatic insufficiency), which is consistent with the research gap mentioned by Jaber et.al.^[25] and Katz et.al.^[24] and should be addressed by extending the follow-up duration. Third, we did not include the ICDC diagnostic indicator of glucocorticoid treatment response and did not explore the association between IgA levels and relapse;^[25] some patients with suspected AIP may not have received trial glucocorticoid treatment, limiting further diagnosis confirmation and potentially affecting typing accuracy. Fourth, compared with Han et.al.’s^[27] analysis of pancreatic enlargement types and Katz et.al.^[24] in the assessment of pancreatic exocrine/endocrine function, we did not explore the interaction between serum IgG4 type and enlargement type, nor did we systematically evaluate EPI or diabetes, which could be directions for future research.

This study adds to the available evidence by highlighting that serum IgG4 typing reflects differences in clinical features and diagnosis-treatment needs in patients with type 1 AIP. This finding emphasizes the importance of pathological examination regardless of serum IgG4 levels, helping avoid missed diagnoses or misdiagnoses and reducing unnecessary surgery, which may improve patient care by guiding more targeted diagnostic strategies. Future studies are needed that include expanded sample sizes and multicenter prospective approaches to reduce bias, explore the association between serum IgG4 typing and prognosis, and establish a combined diagnostic model integrating “serum-imaging-pathology-treatment response” to improve the accurate diagnosis and stratified management of type 1 AIP. Given the rising mortality and substantial public health burden of pancreatic cancer in China,^[35] the targeted diagnostic insights from this study are particularly meaningful for optimizing the differential diagnosis of pancreatic diseases and reducing overtreatment.

SUMMARY

In summary, this study compared serum IgG4 typing in 78 single-center patients with type 1 AIP, confirming that the serum IgG4-P (serum IgG4 ≥135 mg/dL) group was more likely to have bile duct stricture and diffuse pancreatic enlargement and tended to be diagnosed through EUS puncture, whereas the serum IgG4-N (serum IgG4 <135 mg/ dL) group had higher CA19-9 levels, a higher surgical rate, and a lower histopathological confirmation rate, requiring strengthened clinical preoperative puncture evaluation. This study also revealed no significant correlation between serum IgG4 levels and tissue IgG4-P plasma cell count, with pathological examination remaining among the most important diagnostic criteria. These findings provide practical evidence to optimize the clinical diagnostic pathway of type 1 AIP, supporting differentiated diagnosis and treatment strategies based on serum IgG4 typing to reduce misdiagnosis, missed diagnosis, and overtreatment and improve patient prognosis.