Autoimmune pancreatitis (AIP) is a distinct form of pancreatic inflammatory disease that responds well to glucocorticoid therapy (1). AIP can be classified into types 1 and 2 based on clinical and pathological findings. Type 1 AIP is the pancreatic manifestation of IgG4-related disease (IgG4-RD), which is characterized by an elevation of serum IgG4 levels and infiltration of IgG4-positive plasmacytes (2, 3), whereas type 2 is more localized in the pancreas, with normal serum IgG4 levels and the presence of neutrophil infiltration (2, 3). As a relatively newly identified disease, the knowledge on the diagnosis, treatment, and clinical outcomes of AIP has rapidly evolved over the past two decades.

Bibliometric analysis enables the qualitative and quantitative profiling of publications (4) and allows researchers to identify not only the productive countries/regions, institutions, and authors but also the research focus and emerging topics within a specific field (5, 6). Additionally, bibliometric analysis has been applied in research on autoimmune digestive diseases, including inflammatory bowel disease and primary biliary cholangitis (7, 8). However, a bibliometric analysis of AIP has not been reported in the literature thus far.

In the current study, bibliometric analysis was utilized to assess the research status of AIP over the past two decades, as well as identify the research focus and emerging topics.

Herein, we conducted a bibliometric analysis of AIP-related publications over the last 20 years. The annual number of publications showed an upward trend from 2002 to 2009 and has remained relatively stable since 2010. Leading countries/regions, active institutions and authors, core journals and references, and keywords were evaluated. Some landmark articles were identified ( Figure 7 ). To our knowledge, this is the first bibliometric analysis of AIP reported.

Japan was the leading country with respect to AIP research, contributing over 40% of studies on AIP. Seven out of the ten most productive institutions and eight out of the ten most productive authors were from Japan. Furthermore, six out of the top 10 most cited articles were first authored by Japanese researchers. Collaboration among Tokyo Metropolitan Komagome Hospital, Kansai Medical University, Shinshu University, Tohoku University, and Nagoya City University was relatively close. The Japan Pancreas Society had provided timely updates on the diagnostic criteria for AIP (11–16).

Diagnosis of AIP is currently the research focus. The Japan Pancreas Society proposed the first diagnostic criteria in 2002 (11). Since then, much progress has been made. Kamisawa et al. (17) proposed that AIP might be a pancreatic manifestation of a chronic fibroinflammatory condition currently known as IgG4-RD. Notohara et al. (18) and Zamboni et al. (19) summarized the histopathological findings of AIP and reported two subtypes—namely, lymphoplasmacytic sclerosing pancreatitis and idiopathic duct-centric pancreatitis. In 2006, Chari et al. (20) introduced the HISORt diagnostic criteria, which are based on histology, imaging of the pancreas using computed tomography or magnetic resonance imaging, serum IgG4 levels, other organ involvement, and response to steroid therapy. The integration of histology, radiology, serology, and follow-up formed a pathway for the diagnosis of AIP, which remained in further studies. In 2011, Shimosegawa et al. (2) proposed the ICDC, which are the most widely used diagnostic criteria in clinical practice. According to the ICDC, AIP can be subclassified into types 1 and 2 (2). AIP types 1 and 2 share similar radiological findings such as sausage-like pancreatic enlargement, rim-like enhancement around the lesion, delayed homogenous enhancement in the pancreatic parenchyma, and long or multiple strictures without marked upstream dilatation in the main pancreatic duct (21–23). However, AIP types 1 and 2 differ in terms of serology and histopathology. As the pancreatic manifestation of IgG4-RD, type 1 AIP exhibits elevated serum IgG4 levels (3, 24, 25). Histopathologically, type 1 AIP corresponds to lymphoplasmacytic sclerosing pancreatitis, with abundant IgG4-positive plasma cell infiltration (2). Other organ involvement, including sclerosing cholangitis, retroperitoneal fibrosis, and sclerosing sialadenitis, is common in type 1 AIP (26–29), with IgG4-related sclerosing cholangitis being the most common organ involvement, occurring in up to 80% of type 1 AIP cases (30, 31). In contrast, other organ involvement is less common in type 2 AIP (26). Type 2 AIP accounts for <5% of AIP cases in Eastern countries and is more common in Western countries, accounting for up to 10%–20% of AIP cases in Western countries (32, 33). Approximately 30% of type 2 AIP cases are estimated to be associated with inflammatory bowel disease, particularly ulcerative colitis (2). Type 2 AIP has an earlier onset (in individuals in their 30s) and a lower incidence of disease relapse after initial induction of remission than type 1 AIP (32). Histopathologically, type 2 AIP corresponds to idiopathic duct-centric pancreatitis and usually exhibits infiltration of no or very few IgG4-positive plasma cells (2). Serum IgG4 levels are often within the normal range (3). Owing to the paucity of reliable serum biomarkers, the diagnosis of type 2 AIP heavily relies on pancreatic histopathology (26). Moreover, both types of AIP share a dramatic response to glucocorticoid therapy (32).

Much progress has been made in the treatment of AIP. Glucocorticoids are the first-line therapy. Indications for glucocorticoid therapy are symptoms such as obstructive jaundice, abdominal pain, and other organ involvement (34). As for the induction of remission, Kamisawa et al. (35) recommended an initial dose of 0.6 mg/kg/day of oral prednisolone for 2–4 weeks. Symptoms are anticipated to be relieved within days after commencing the treatment (36). Assessment of the response to initial treatment with biochemical, serological, and radiological work-up at weeks 2–4 is recommended (36). Subsequently, glucocorticoids should be gradually tapered off, usually 5 mg every 1–2 weeks. When glucocorticoid therapy fails to relieve AIP-related symptoms, a reevaluation of diagnosis should be in order (15, 16). Clinicians should be particularly cautious about pancreatic cancer misdiagnoses. There are disputes over glucocorticoid maintenance therapy. In Western countries, glucocorticoid therapy is generally limited to the induction of remission without maintenance (37) because prolonged administration may increase the risk of infections, diabetes, osteoporosis, and cataracts (34). However, Masamune et al. (38) conducted the first AIP-related randomized controlled trial, the results of which favored maintenance therapy. Maintenance therapy with prednisolone at a dose of 5–7.5 mg/day was continued for 3 years. Compared with the cessation group, which had withdrawn at 26 weeks since the initial glucocorticoid therapy, the maintenance group achieved better 3-year relapse-free survival (42.1% vs. 76.7%, p=0.007) (38). Moreover, no major glucocorticoid-related complications requiring treatment cessation were found in the maintenance group (38). Thus, the Japanese consensus guidelines advocate a 3-year maintenance therapy to prevent disease relapse (16). Immunosuppressants such as azathioprine, methotrexate, and mycophenolate mofetil may be beneficial for patients with AIP. A recent meta-analysis had suggested that azathioprine was effective in preventing AIP relapse (39). B-cell depletion therapy has been proposed as a treatment for recurrent type 1 AIP. CD20 is a B-cell surface marker involved in calcium channel activation, cell proliferation, and B-cell differentiation (40). Rituximab is a monoclonal antibody targeting human CD20. Rituximab can induce complement activation and cell-mediated cytotoxicity, leading to B-cell depletion (41). Hart et al. (42) reported rituximab as a treatment for recurrent AIP. A decrease in serum IgG4 concentration and the extinction of pancreatic hypermetabolic signal on positron emission tomography were achieved in type 1 AIP after rituximab treatment (43). Other therapies for type 1 AIP, including rilzabrutinib (Bruton tyrosine kinase inhibitor), belimumab (B-cell activating factor inhibitor), and inebilizumab (anti-CD19 monoclonal antibody), are under investigation (44).

Keyword burst detection is capable of tracing research frontiers (9, 10). Some emerging topics had been recognized, including “risk,” “malignancy,” and “outcome.” Patients with AIP are at a high risk for malignancy (45). A recent meta-analysis of 17 studies involving 2,746 patients revealed that the overall prevalence of malignancy in patients with AIP was 9.6% (46). The top 5 most prevalent malignancies in patients with AIP were gastric, colorectal, bladder, prostate, and pancreatic cancers (46). The majority of pancreatic cancer cases in patients with AIP occurred at no less than 2 years after an AIP diagnosis (47). Other than malignancy, the long-term outcomes of AIP include diabetes mellitus (DM) and pancreatic exocrine insufficiency (PEI) (48, 49). Chronic inflammation in patients with AIP may cause damage to the pancreatic β-cells and acinar cells, leading to DM and PEI (50). By alleviating the inflammation and swelling of pancreatic tissues with glucocorticoid therapy, both endocrine and exocrine functions are supposed to be restored (50, 51). However, DM and PEI are often described at the time of diagnosis and during follow-up. The pooled prevalence of DM and PEI in patients with AIP at the time of diagnosis is 36.5% and 45.2%, respectively (52). Moreover, the pooled prevalence of DM during follow-up is 40.9% (52). The prevalence of PEI during follow-up ranged from 23.8% to 72.7% (49, 53, 54). Studies concerning malignancy, DM, and PEI in patients with AIP are mostly retrospective. More high-quality prospective cohort studies are required to better understand the long-term outcomes of patients with AIP.

In addition to clinical outcomes, pancreatic tissue acquisition is now gaining attention. A “22-gauge needle” and “fine-needle aspiration” have been recognized as research hotspots since the late 2010s by keyword burst analysis. Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) and endoscopic ultrasound-guided fine-needle biopsy (EUS-FNB) have been utilized for pancreatic tissue acquisition in patients with AIP. When pathology specimen collection is necessary for diagnosis or when malignancy is suspected, EUS-FNA or EUS-FNB should be taken into consideration (16). EUS-FNA is designed for the aspiration of cells from the target lesion using a conventional straight needle. Tissue core samples acquired by EUS-FNA are usually limited in size, making the histopathological diagnosis of AIP less than satisfactory. With the help of recently developed core needles, EUS-FNB is capable of obtaining a large amount of tissue core samples with preserved tissue architecture (16, 55). According to the ICDC, the histological findings of AIP can be categorized into levels 1 and 2 (2). A recent meta-analysis demonstrated that EUS-FNB had better diagnostic yield than EUS-FNA (56). The pooled diagnostic yield for level 1 and 2 histological findings was 55.8% for EUS-FNA and 87.2% for EUS-FNB (p=0.03) (56). As for the needle size, a 19-gauge needle exhibited a better pooled diagnostic yield for level 1 and 2 histological findings than a 22-gauge needle (88.9% vs. 60.6%, p=0.023) (56). However, studies investigating the diagnostic yield of EUS-FNA or EUS-FNB had mainly focused on type 1 AIP, rather than type 2 AIP, possibly because these studies were largely conducted in Eastern countries where type 2 AIP is quite rare (<5% of total AIP cases) (32, 33). Moreover, histological diagnosis is more essential for type 2 AIP than for type 1 AIP, as there are no reliable serological markers. Further studies should recruit more patients with type 2 AIP by inviting more European and American centers.

Research cooperation among countries/regions, institutions and authors has been identified by the cooperation network in our study. Active collaborations were noted among Japan, the United States, China, and South Korea. Due to the limited cases of type 2 AIP reported, its genetic predisposition, relationship with inflammatory bowel disease, and long-term outcome have not been characterized in detail (57). This might be because the current knowledge base of AIP is mainly generated from Eastern-population-driven information, especially in Japan, where type 2 AIP is less common. Therefore, further collaborative global research is essential to understand AIP comprehensively.

This study has some limitations. First, the data were retrieved exclusively from the WoSCC, rather than other databases such as Embase and MEDLINE. The WoSCC is the most commonly applied database in bibliometric analysis because it provides timely and comprehensive updates on citation network. Moreover, software that were applied in our bibliometric analysis had difficulties in integrating data from different resources. Second, the increase in the number of publications on AIP from 2002 to 2009 might have resulted from the overall increase in scientific outputs in the medical field over the past two decades. Finally, some new emerging topics related to AIP might not have been identified because of the sensitivity of algorithms applied in the analysis. Although multiple software/packages have been used in our study, the information provided in our study is still constricted by algorithms applied in bibliometric analysis. Further development in the methodology of bibliometric analysis might be helpful in resolving these limitations.

Over the past two decades, Japan was the leading country in AIP research, with more than half of the top 10 most productive institutions and top 10 most productive authors being from Japan. Research papers were mainly published in specialized journals. Diagnosis of AIP was the research focus. Long-term outcomes and pancreatic tissue acquisition are recognized as research frontiers for AIP.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

X-DZ: conceptualization, methodology, software, formal analysis, resources, data curation, visualization, and writing – original draft. YZ: methodology, software, validation, formal analysis, resources, data curation, visualization, and writing – original draft. Y-ZZ: methodology, software, validation, formal analysis, resources, data curation, visualization, and writing – original draft. C-HZ: conceptualization, supervision, funding acquisition, and writing – review and editing. D-WZ: conceptualization, supervision, project administration, and writing – review and editing. All authors contributed to the article and approved the submitted version.