Gastric Varices Rupture due to Splenic Vein Obstruction Associated with Autoimmune Pancreatitis

Abstract

A 60-year-old man with a high IgG4 level was found to have pancreatic tail enlargement on computed tomography (CT), and autoimmune pancreatitis (AIP) was confirmed by a histological diagnosis. He was treated with prednisolone for one year and seven months, at which point his treatment finished. Four months later, however, he had hematemesis from gastric varices. CT showed recurrence of pancreatic tail enlargement with obstruction of the splenic artery and vein and formation of collateral blood vessels to the gastric fornix. Endoscopic injection sclerotherapy was performed, and he underwent splenectomy. This case highlights the importance of paying attention to peripancreatic vascular abnormalities during follow-up of AIP patients.

Background

The concept of autoimmune pancreatitis (AIP) was first proposed by Yoshida et al. (1) in 1995. AIP is characterized by localized or diffuse enlargement of the pancreas and is often diagnosed with obstructive jaundice. Its pathological features are severe fibrosis and infiltration of lymphocytes and plasma cells, and its clinical features are significantly responsive to steroids (2). Because asymptomatic cases can be observed without treatment (3), steroid therapy for asymptomatic cases is controversial.

We herein report initially asymptomatic AIP in a case of splenic vein occlusion with collateral angiogenesis leading to gastric varices rupture after discontinuation of steroid treatment. Although formation of varices associated with AIP has been reported in a few cases, it is not yet sufficiently recognized, and there is no definite mention of it in the guidelines for AIP. Therefore, we present a case of venous aneurysm formation associated with AIP.

Case Report

A 60-year-old man presented to the Department of General Medicine in our hospital with a primary complaint of numbness and pain in his fingers. He had no history of diabetes, hepatitis or liver cirrhosis, and his alcohol intake was approximately 20 g daily. A blood test showed normal liver enzyme levels and a mildly elevated IgG4 level (145 mg/dL).

Contrast-enhanced computed tomography (CT) performed for systemic screening showed an enlarged pancreas, mainly at the pancreatic tail, with progressively enhanced membrane-like structures at the margins (Fig. 1A). Endoscopic ultrasound (EUS) showed diffuse enlargement of the pancreatic tail and a surrounding hypoechoic area (capsule-like rim) (Fig. 2). Based on these imaging findings and the high IgG4 levels, AIP was suspected.

Figure 1.

Contrast-enhanced CT images. Before treatment (A), the pancreatic tail and surrounding membrane-like structures were enlarged. The enlargement of the pancreatic tail was improved after four months of treatment (B) compared with before, and the membrane-like structures disappeared.

Figure 2.

EUS image of the pancreatic tail. The pancreatic parenchyma was hypoechoic overall, and a capsule-like rim was observed around the parenchyma (arrowheads). The splenic artery (red arrow) and splenic vein (yellow arrow) were visible, and no stenosis of the splenic artery was observed at this point.

Fine-needle aspiration to the pancreas was performed for a pathological diagnosis. The pathological examination revealed 3 findings: prominent infiltration of lymphocytes and plasma cells and the presence of fibrosis, IgG4-positive cells >10 cells/high-power field and storiform fibrosis (Fig. 3). The diagnosis of AIP was confirmed based on the Japanese Society of Pancreatology Diagnostic Criteria for Autoimmune Pancreatitis 2020 (4).

Figure 3.

The pathology images were obtained by endoscopic ultrasound-fine needle aspiration. (A) Hematoxylin and Eosin staining. Arrowheads: prominent infiltration of lymphocytes and plasma cells and the presence of storiform fibrosis. (B) IgG4 immunostaining. Arrowheads: IgG4-positive plasma cells. More than 10 IgG4-positive plasma cells are present in each high-power field.

The patient was asymptomatic but selected treatment with prednisolone after a discussion about the need for treatment. Two months after his initial visit, treatment with 40 mg (0.6 mg/kg) of prednisolone was initiated, tapering off by 5 mg every 2 weeks. After starting prednisolone treatment, the pancreatic tail swelling improved over time (Fig. 1B), and his IgG4 levels normalized. After 10 months of prednisolone treatment, the dose was reduced to 5 mg. The patient then requested to discontinue treatment with prednisolone, so prednisolone was further tapered off, and after 19 months of treatment, it was discontinued altogether.

Four months after the discontinuation of prednisolone, the patient had hematemesis and visited the emergency department of our hospital. Contrast-enhanced CT showed recurring enlargement of the pancreatic tail, and the involved splenic artery and vein were disrupted (Fig. 4A). In addition, collateral blood vessels had developed toward the gastroesophageal area, suggesting the formation of gastric varices (Fig. 4B). A large amount of hyperdense material suspected of being hemorrhaging had accumulated in the stomach, and upper gastrointestinal bleeding was suspected. The patient was therefore transferred to the Department of Gastroenterology, and emergency upper gastrointestinal endoscopy was performed on the same day.

Figure 4.

Contrast-enhanced CT images. (A) Arrowheads: The pancreatic tail was enlarged, and a membrane-like structure was observed around the pancreas, similar to that noted before the treatment. Arrowheads: The splenic vein was involved in the peripancreatic membrane-like structure and was obstructed. (B) Arrowhead: Increased growth of collateral blood vessels was observed from the gastric fornix to the cardiac region, suggesting gastric varices formation.

Gastric varices were observed at the gastric fornix (Lg-f, F1, Cb, RC+), and eruptive bleeding occurred (Fig. 5A). Endoscopic injection sclerotherapy (EIS) using n-butyl-2-cyanoacrylate (0.5 mL) and ethyl ester of iodinated poppy-seed oil fatty acid (0.3 mL) was performed to stop the bleeding (Fig. 5B). Two weeks later, the varices ruptured again and bled, and EIS was performed once more. Hemostasis was temporarily achieved; however, we determined that treatment of the collateral vessels was necessary to achieve complete resolution of the varices. We discussed treatment with transcatheter vascular embolization procedures, such as balloon-occluded retrograde transvenous obliteration (BRTO) and partial splenic embolization (PSE). However, we did not select BRTO because a treated collateral vessel could not be confirmed by contrast-enhanced CT. Furthermore, since the targeted artery in the spleen was considered inaccessible by a catheter due to complete occlusion of the splenic artery, we also excluded the PSE procedure.

Figure 5.

Gastrointestinal endoscopy images. (A) Eruptive bleeding from gastric varices was observed. (B) After hemostasis with endoscopic injection sclerotherapy (n-butyl-2-cyanoacrylate 0.5 mL+ethyl ester of iodinated poppy-seed oil fatty acid 0.3 mL). (C) One month after splenectomy. Arrowheads: residual gastric varices were observed; however, they were smaller than those observed before splenectomy. (D) Six months after splenectomy. The gastric varices had completely disappeared.

After a conference with our department and the radiology and surgery departments, it was decided that endovascular treatment would be difficult, so splenectomy would be performed instead. The patient requested the surgery be performed at another hospital, and three weeks after the second EIS procedure, splenic artery embolization and splenectomy were performed at the referring hospital.

After surgery, the gastric varices disappeared over time, and complete endoscopic resolution was observed six months postoperatively (Fig. 5C, D). The patient experienced epigastric pain, and CT showed residual enlargement of the pancreatic tail (Fig. 6A); therefore, prednisolone (40 mg) was initiated 1 year postoperatively. Since then, the patient's AIP has improved over time (Fig. 6B), and 5 mg/day prednisolone is currently being continued without recurrence.

Figure 6.

Contrast-enhanced CT images. Before (A) and after (B) the resumption of steroids after splenectomy. Swelling of the pancreatic body tail observed before the resumption of steroids improved after the resumption of steroids. Arrow: after splenic artery coil embolization.

Discussion

AIP was first reported as chronic inflammatory sclerosis by Sarles et al. (5) in 1961. The concept of AIP was proposed by Yoshida et al. (1) in 1995, and the involvement of autoimmunity was suggested by the presence of hypergammaglobulinemia, autoantibodies, histological lymphocytic infiltration, and the effectiveness of steroids.

The standard treatment for AIP is steroid therapy (4). A 2016 national epidemiologic survey showed that 98.6% of patients responded to steroid treatment (6), indicating that steroids are a highly effective treatment for AIP. However, while AIP does respond to steroids, a certain percentage of cases relapse. The aforementioned national survey (6) found that the cumulative incidence of relapse among treated patients with AIP was 14% at 3 years and 25% at 5 years. Kubota et al. (7) also reported a cumulative incidence of relapse of 10% at 1 year, 25.8% at 3 years, and 35.1% at 5 years in a retrospective study of 479 steroid-treated patients with AIP at 22 centers. AIP should be followed up with full recognition of the possibility of relapse in the long term after remission. In addition, a previous study reported a significantly higher incidence of relapse in the steroid discontinuation group than in the steroid maintenance group (45.2% vs. 26.1%, p=0.023), indicating the need to be particularly alert for relapse after steroid discontinuation (8). Indeed, the patient in the present report relapsed four months after discontinuation of steroids, triggered by hematemesis, suggesting that the patient relapsed relatively early after steroid discontinuation.

In the present case, the gastric varices were thought to have formed due to the development of collateral blood vessels to the stomach caused by AIP-induced splenic vein obstruction. Bernades et al. (9) prospectively reviewed 266 patients with chronic pancreatitis, followed up for an average of 8.2 years, and found that 22 (8%) had splenic vein obstruction, which is common in cases of chronic pancreatitis. The pathogenesis of variceal formation due to splenic vein obstruction differs from that due to portal hypertension, particularly in patients with liver cirrhosis. In patients with variceal formation due to splenic vein obstruction, obstruction of the splenic vein increases venous pressure on the splenic side, forming a collateral blood passage via the short gastric vein (10). In contrast, in patients with variceal formation due to portal hypertension with liver cirrhosis, collateral vessels are formed by increasing pressure in the entire portal system. Therefore, the pathways of collateral vessels are widely distributed in the portal system, including the left gastric vein, splenic-retroperitoneal vein, and inferior mesenteric vein (11).

Ishikawa et al. (12) retrospectively reviewed 54 patients with AIP and found that 22 (41%) had stenosis of the splenic vein (41%), predicting a higher incidence of vascular abnormalities associated with AIP than with usual chronic pancreatitis. Hao et al. (13) reported a case of AIP with gastric varices due to splenic vein obstruction, in which the AIP improved with steroid therapy, resulting in an improvement of the gastric varices. Ishikawa et al. (12) treated 16 patients with gastric varices caused by AIP-induced splenic vein obstruction with steroids, and the gastric varices were improved in 14. Steroid therapy for AIP may also be effective in patients with peripancreatic vascular obstruction or stenosis. Furthermore, Goto et al. (14) reported three cases of AIP with gastric varices, two of which did not improve with steroid therapy; therefore, a certain number of cases of AIP-induced splenic vein stenosis and obstruction may not improve with steroids.

In the present case, despite endoscopic treatment, the patient bled twice in a short period, and transcatheter vascular embolization was difficult to perform; therefore, we determined that early treatment of the collateral vessels was necessary, and surgical resection was performed. As a result, the varices resolved rapidly without rebleeding. If transcatheter vascular embolization had been feasible, the first choice would have been BRTO or PSE followed by steroid treatment to allow the varices to regress. However, surgical resection must be considered for cases in which transcatheter vascular embolization is deemed difficult or that require early treatment. Kamio et al. (15) reported a case of splenic rupture due to AIP-induced splenic vein obstruction that was treated with splenectomy, as in the case in the present report. To avoid similar results, we must pay close attention to the effects of AIP on the peripancreatic vasculature in the follow-up of patients with AIP.

The AIP in the present patient improved after one year and seven months of treatment with steroids; however, he developed gastric varices due to splenic vein obstruction approximately four months after steroid discontinuation. Steroids should be administered in symptomatic cases of AIP because they are expected to improve symptoms; however, the treatment of asymptomatic cases is controversial. A meta-analysis of AIP in 2022 showed an increased prevalence of diabetes after steroid treatment compared with before (16); therefore, steroid treatment may be a risk factor for diabetes exacerbation. In addition, because a certain number of asymptomatic cases resolve without treatment (7), treatment of asymptomatic cases should be carefully considered.

However, as mentioned above, obstruction and stenosis of the peripancreatic vasculature are common in AIP; therefore, steroid therapy should be administered to patients with vascular abnormalities, such as splenic vein stenosis or obstruction, even if they are asymptomatic, considering the possibility of varicose vein formation in the future. Juarez et al. (17) similarly indicated that early intervention was necessary if vascular abnormalities were evident. In addition, they revealed that noninvasive screening modalities, such as contrast-enhanced CT and abdominal ultrasonography, should be performed for follow-up.

However, there are no reports indicating which patients are prone to develop vascular abnormalities or how long such abnormalities take to appear. Given that the gastric varices developed four months after steroid discontinuation in our case, some cases of AIP with severe inflammation may require imaging examinations be performed every few months. There is still no consensus concerning the specific indications for treatment of AIP with vascular lesions, and data on the timing of treatment initiation and methods and intervals of follow-up examinations are needed. Further reports involving more cases are thus necessary.

In the case in the present report, contrast-enhanced CT at the initial visit showed that the AIP extended around the splenic artery and vein; however, no vascular abnormalities were observed, and the same was true six months later. Of note, because contrast-enhanced CT was not performed for 19 months until rupture of the varices, the peripancreatic vascular involvement could not be evaluated. Since AIP had spread around the splenic vein prior to treatment, this patient was at high risk for splenic vein obstruction and associated gastric varices formation, so contrast-enhanced CT should have been performed to evaluate the peripancreatic vasculature prior to steroid discontinuation. If AIP had still been present around the splenic vein, similar to the situation before treatment, steroid treatment should have been continued or increased in order to reduce the inflammation of AIP. In addition, if contrast-enhanced CT or abdominal ultrasound had been performed early after steroid discontinuation, recurrence of AIP or vascular abnormalities might have been able to be observed in advance. Therefore, regular evaluations of the portal vein system, including the splenic vein, using contrast-enhanced CT are important during the follow-up of patients with AIP.

Conclusion

This report emphasizes the need to consider that AIP can cause inflammatory spread to the portal vein system, including the splenic vein, resulting in stenosis and obstruction. In such instances, the use of steroid therapy should be considered, even if the patient is asymptomatic.

Informed consent was obtained from the patient.

The authors state that they have no Conflict of Interest (COI).