Abstract
Autoimmune pancreatitis is a rare type of chronic pancreatitis associated with immune disorders. Co-occurrence of regional portal hypertension and gastric variceal bleeding associated with autoimmune pancreatitis is exceedingly rare. This case report aimed to highlight the dual role of endoscopic ultrasonography in the diagnosis and treatment of this condition and discuss the choice of medication for inducing remission. A 44-year-old man presented with melena and anemia. Imaging revealed pancreatic swelling, splenic vein occlusion, and gastric varices. Endoscopic ultrasonography–guided fine-needle biopsy confirmed autoimmune pancreatitis. Endoscopic ultrasonography was used to successfully manage gastric variceal bleeding via coil placement and tissue glue injection. The patient was treated with rituximab, resulting in improved hemoglobin levels, reduced B lymphocyte count, and resolution of pancreatic swelling. Follow-up indicated successful management and patient well-being. Additionally, we reviewed published cases of autoimmune pancreatitis patients with portal hypertension. In conclusion, endoscopic ultrasonography plays a crucial role in the diagnosis and treatment of autoimmune pancreatitis complicated by gastric variceal bleeding. Rituximab may be more effective than glucocorticoids in treating autoimmune pancreatitis associated with gastric variceal bleeding.
Keywords
Introduction
Autoimmune pancreatitis (AIP) is a rare and distinct type of chronic pancreatitis with increasing incidence and prevalence. In 2016, the prevalence of AIP in Japan was 10.1 per 100,000 persons, with an incidence of 3.1 per 100,000 persons. 1 Compared with data from the 2011 nationwide epidemiological survey in Japan, 2 both incidence and prevalence have increased. AIP is typically classified into types 1 and 2. Type 1 AIP is characterized by immunoglobulin G4 (IgG4)–related pancreatic involvement and elevated serum IgG4 levels. Type 2 AIP, known as idiopathic duct-centric pancreatitis, is generally not characterized by elevated serum IgG4 levels. 3 Patients with type 1 AIP far outnumber those with type 2 AIP. 4 Extrapancreatic involvement is possible in AIP, affecting organs such as the bile ducts, lymph nodes, salivary glands, and lungs. 5 However, the co-occurrence of regional portal hypertension (RPH) and gastric variceal bleeding (GVB) associated with AIP is exceedingly rare. This report highlights the dual role of endoscopic ultrasonography (EUS) in the diagnosis and treatment of such complex cases and discusses the selection of an optimal induction therapy.
Case presentation
The reporting of this case follows the Case Report (CARE) guidelines. 6
A man in his 40s presented to our hospital with a 2-month history of melena, occurring every 1–3 days, accompanied with mild abdominal pain. He had no medical history of peptic ulcer, viral hepatitis, or liver cirrhosis. He had no history of heavy alcohol consumption and no relevant family history. Physical examination revealed an anemic appearance and nontender abdomen. Laboratory tests indicated anemia (hemoglobin level, 49 g/L) as well as decreased serum iron (10.0 µg/dL) and ferritin (7.0 ng/dL) levels. Serum IgG4 concentrations were within the normal range on two separate tests (922 mg/L and 939 mg/L). The peripheral blood CD19+ B lymphocyte count was 275 cells/μL. Liver function, renal function, coagulation profile as well as the levels of carcinoembryonic antigen, carbohydrate antigen 19-9, amylase, and lipase were within normal limits. Contrast-enhanced abdominal computed tomography (CT) revealed significant swelling of the pancreatic body and tail with a sausage-like appearance and reduced enhancement, along with pancreatic duct narrowing in these regions (Figure 1(a)). Additionally, there was occlusion of the mid-to-distal splenic vein; multiple tortuous collateral veins in the splenic region, around the esophagogastric junction (Figure 1(b)), peripancreatic area, and superior mesenteric region; and splenomegaly. Positron emission tomography–CT showed thickening of the pancreatic body and tail, with increased fluorodeoxyglucose (FDG) uptake (standardized uptake value = 7.5) (Figure 1(c)). FDG uptake was not elevated in the salivary or lacrimal glands. EUS revealed a hypoechoic and markedly enlarged pancreatic tail (Figure 2(a)), approximately 3.3 cm in thickness, with punctate hyperechoic foci but no pancreatic ductal dilatation. Under EUS guidance, fine-needle biopsy (FNB) was initially performed using 22-G AcquireTM (Boston Scientific, Marlborough, MA, USA) through the gastric fundus to obtain a pancreatic tissue sample for diagnostic evaluation. No ulcers were found. Anechoic tubular structures were observed in the gastric fundus region, and Doppler ultrasound confirmed blood flow (Figure 2(b)), with the largest vessel measuring approximately 10 mm in diameter. Then, a 19-G fine-needle aspiration (FNA) needle (Expect™, Boston Scientific, Marlborough, MA, USA) was used to implant a coil (0.9 mm, 10 mm; Nester, Cook Medical, Bloomington, IN, USA) into the nourishing vessel, and 1 mL of Histoacryl tissue glue was injected (Figure 2(c)). Doppler ultrasound revealed significantly decreased flow around the fundus region (Figure 2(d)). The patient has not experienced melena ever since. Pathological examination revealed marked pancreatic fibrosis with lymphoplasmacytic infiltration; immunohistochemical analysis showed 10–30 IgG4+ plasma cells per high-power field (HPF) (Figure 3(a)). According to the 2008 Asian Diagnostic Criteria, 7 a definite diagnosis of AIP was established. Type 1 AIP was further confirmed based on the infiltration of IgG4+ plasma cells into the pancreatic tissue. Gastroscopy showed significant improvement in gastric varices 1 month after EUS (Figure 3(b)). Subsequently, the patient received 600 mg intravenous rituximab. Two weeks later, the patient was readmitted for a second infusion of rituximab. By this time, his hemoglobin level had increased to 103 g/L, and his CD19+ B lymphocyte count had decreased to 5 cells/μL. Follow-up abdominal CT indicated reduced swelling in the pancreatic tail (Figure 3(c)). The patient remained in good health with regular follow-up visits at our hospital.

CT and PET-CT findings before treatment. (a) The enhanced CT arterial phase showed diffuse enlargement of the pancreatic body and tail, with a “sausage-like” appearance and reduced enhancement; splenomegaly was also noted. (b) Varices in the gastric fundus and perigastric veins and (c) the PET-CT indicated thickening of the pancreatic body and tail with increased FDG uptake as well as splenomegaly. CT: computed tomography; PET-CT: positron emission tomography-computed tomography; FDG: fluorodeoxyglucose.

EUS findings. (a) EUS showed a hypoechoic enlargement in the pancreatic tail. (b) EUS showed an anechoic tubular structure in the gastric fundus, with visible blood flow signals on Doppler imaging. (c) EUS-guided coil placement and tissue glue injection and (d) reduced blood flow signals in the gastric fundus after EUS-guided coil placement and tissue glue injection. EUS: endoscopic ultrasonography.

Pathological manifestations of the pancreas and post-treatment findings on gastroscopy and CT.
Discussion
To the best of our knowledge, this is the first reported case worldwide wherein EUS-guided coil placement was used to treat gastric varices secondary to AIP, and EUS was used for diagnosing AIP in the same patient.
AIP diagnosis can be challenging, particularly for focal-type AIP that requires differentiation from pancreatic cancer. Imaging examinations and serum IgG4 testing serve as important noninvasive tools; additionally, EUS has become increasingly valuable as it enables detailed visualization of the pancreas and bile ducts and facilitates tissue biopsy. Typical EUS features of AIP include hypoechoic lesions, punctate hyperechoic foci within the lesions, peripancreatic hypoechoic rim, and thickening of the bile duct wall.8,9 EUS-FNA/FNB is a minimally invasive method for obtaining pancreatic tissue. The patient in this case underwent EUS-FNB; the pathology revealed fibrosis and infiltration of IgG4+ cells (>10 cells/HPF). According to a 2016 nationwide survey in Japan, tissue sampling using EUS-FNA was performed in 85.5% of AIP patients. 1 However, a meta-analysis including 309 patients with AIP who underwent EUS-FNA and 131 who underwent EUS-FNB showed that the diagnostic yield for level 1 or level 2 histology criteria of AIP was lower with FNA than with FNB (55.8% vs. 87.2%).10,11 A prospective multicenter study in Japan demonstrated that the sensitivity of EUS-FNA (22-G aspiration needle) for AIP diagnosis was only 7.9%, indicating the limited efficiency of EUS-FNA in establishing a pathological diagnosis of AIP. 12 Takahashi et al., 13 used EUS-FNA with a 22-G aspiration needle to determine whether the pancreatic mass lesions were benign or malignant; they reported a specificity of 100% at both cytopathologic and histopathologic levels, indicating that EUS-FNA is highly effective in ruling out pancreatic cancer rather than diagnosing AIP. Notably, an elevated IgG4 level is not a prerequisite for diagnosing type 1 AIP. In the present case, for example, the patient had type 1 AIP and normal serum IgG4 levels. In a study conducted by German researchers, 44.7% (17/38) of patients with type 1 AIP had nonelevated serum IgG4 levels (<140 mg/L). 14 Another study reported that 7.4% of patients with chronic pancreatitis caused by other reasons had increased IgG4 levels. 15 Elevated IgG4 can also be observed in patients with pancreatic cancer.16,17
The patient presented with GVB secondary to RPH. The mechanism of RPH may involve compression of the posterior splenic vein by the enlarged pancreatic body and tail, leading to impaired blood flow, subsequent formation of perigastric and gastric fundus varices, and splenomegaly. 18 As per a previous study, among 101 patients with type 1 AIP, 8.9% (9/101) had gastric fundus varices with red signs, and 33.7% (34/101) had perigastric collateral vessels. 19 Splenic vein involvement may be the trigger for the occurrence of RPH. Splenic vein involvement was reported in 42.6% (43/101) of patients with type 1 AIP . 19 Reports have indicated that glucocorticoid (GC) therapy can improve gastric varices in some AIP patients with this complication.20,21 Another report documented the case of a patient with AIP who had RPH, systemic lymphadenopathy, and splenomegaly and underwent splenectomy due to upper gastrointestinal bleeding. Postoperatively, in the absence of GC therapy, the pancreas normalized, gastric varices improved, and there were no enlarged lymph nodes in the retroperitoneum. 22 Relevant patient information obtained from our literature review is summarized in Table 1. In our case, the patient had GVB and severe anemia, making GC therapy unsuitable. Considering the trauma associated with splenectomy, we chose to perform coil placement combined with glue injection following FNB during a single EUS procedure. Follow-up gastroscopy 1 month later demonstrated a marked improvement in the gastric varices.
A review of autoimmune pancreatitis patients with portal hypertension.
DM: diabetes mellitus; GCs: glucocorticoids; IgG: immunoglobulin G; IgG4: immunoglobulin G4; PMH: past medical history; PPI: proton pump inhibitors; SMV: superior mesenteric vein; SV: splenic vein; UGIB: upper gastrointestinal bleeding.
GCs are recognized as first-line agents for both induction and maintenance therapy in AIP. A 2016 nationwide survey in Japan showed that 98.6% of type 1 AIP patients responded to GC therapy. 1 The Japanese consensus guidelines recommend oral prednisolone at a dose of 0.6 mg/kg/day for 2–4 weeks to induce remission and steroids for 3 years to prevent relapse. 23 Additionally, B cell–depleting agents such as rituximab can be used to treat AIP. Nikolic et al. 24 reported that of the 12 patients with type 1 AIP who were treated with rituximab, 8 achieved complete remission and 4 achieved partial remission; 1 patient discontinued treatment due to fever and tuberculosis, while the remaining 11 showed no recurrence during follow-up, with a median follow-up duration of 17 months. In the study by Carruthers et al., 25 30 patients with IgG4-related disease (RD) received 2 doses of 1000 mg rituximab, administered 15 days apart. Among them, 26 initially received only rituximab, and 97% responded to treatment. At 6 months, 14 patients (47%) continued to show complete remission. A meta-analysis, including 18 studies with a total of 374 patients and an average follow-up duration of 23.4 months, found that rituximab appeared to effectively induce remission in IgG4-RD, and prophylactic maintenance therapy with rituximab may help prevent IgG4-RD recurrence. 26 The European Guideline on IgG4-Related Digestive Disease has indicated that rituximab should be considered for patients who are either resistant or intolerant to high-dose steroid therapy for maintaining remission or for those who have not responded to immunosuppressant medications. 27 In this case, the patient presented with GVB. Given that GCs can increase the risk of gastrointestinal bleeding, we chose rituximab for induction therapy because it poses a relatively lower risk of bleeding.
Conclusion
We reported a case of AIP complicated by GVB, wherein a biopsy sample was collected and hemostasis was achieved under EUS guidance, followed by induction of remission with rituximab. This case demonstrates the efficient dual role of EUS in the diagnosis and treatment of AIP. Furthermore, rituximab may serve as an ideal agent for remission induction in such cases.
Footnotes
Acknowledgements
None.
Author contributions
W. Xi: Data curation, formal analysis, and writing–original draft. W. Shi: Conceptualization, funding acquisition, and writing–original draft. Y. Feng: Conceptualization, funding acquisition, and writing–review & editing. X. Cao: Formal analysis and writing–review & editing. X. Chang: Methodology and writing–review & editing. A. Yang: Conceptualization, supervision, funding acquisition, and writing–review & editing.
Consent to participate
Written informed consent was obtained from the patient.
Data availability statement
All the obtained data have been presented in this case report. For any further requests regarding the data, please contact the corresponding authors via email.
Declaration of conflicting interests
The authors have no conflicts of interest to disclose.
Funding
This study was supported by the Beijing Natural Science Foundation (L232120), National Natural Science Foundation of China (32370946, 82470700), and CAMS Innovation Fund for Medical Sciences (CIFMS) (2023-I2M-2-002, 2024-I2M-C&T-B-016).
