Successful Endoscopic Sealing of Gastric Perforation With Tissue Adhesive Following Endoscopic Treatment for Gastric Fundal Variceal Bleeding: A Case Report

ABSTRACT

Gastric perforation after endoscopic treatment of gastric fundal variceal bleeding is extremely rare. In carefully selected patients with early, small perforations and limited peritoneal contamination, endoscopic tissue adhesive injection combined with conservative management may provide a feasible nonsurgical rescue strategy.

Abbreviations

CT

computed tomography

EGD

esophagogastroduodenoscopy

EVB

esophageal and gastric variceal bleeding

EVL

endoscopic variceal ligation

TIPS

transjugular intrahepatic portosystemic shunt

1. Introduction

Esophageal and gastric variceal bleeding is one of the most severe complications of portal hypertension in patients with cirrhosis and remains a potentially fatal gastrointestinal emergency [1]. It is characterized by abrupt onset, rapid blood loss, and high mortality, with reported in‐hospital mortality rates ranging from 10% to 20% despite advances in treatment [2]. Rapid diagnosis and timely intervention are therefore crucial to improve outcomes.

Current treatment strategies for acute variceal bleeding include hemodynamic resuscitation, vasoactive therapy, antibiotic prophylaxis, endoscopic intervention, interventional radiologic procedures, and surger [3]. Among these, endoscopic treatment has become a cornerstone of management because of its high efficacy, minimally invasive nature, and role in both emergency hemostasis and secondary prevention.

Endoscopic variceal ligation (EVL) is the first‐line treatment for esophageal varices. By mechanically ligating variceal columns, EVL interrupts blood flow, induces thrombosis and ischemic necrosis, and ultimately promotes fibrosis and obliteration of the varices. The initial hemostasis rate of EVL is reported to be approximately 90%, while complications—including transient dysphagia, chest pain, ulceration, stricture, infection, and rarely, perforation—occur in 2%–20% of patient [4]. Post‐banding ulcer bleeding has been reported in 2.8%–15% of cases and is associated with substantial mortality [5, 6].

For gastric fundal varices, endoscopic cyanoacrylate injection is generally regarded as more effective than EVL or endoscopic injection sclerotherapy, owing to its high immediate hemostatic rate and lower rebleeding rate [7, 8, 9]. However, this approach is not free of complications. Reported adverse events include ulcer formation, local tissue necrosis, bacteremia, inflammatory reaction, early glue extrusion, and ectopic embolization [10, 11, 12]. Gastric perforation following endoscopic treatment for gastric variceal bleeding is exceedingly rare, and there is no consensus regarding its optimal management.

Herein, we report a rare case of gastric perforation after endoscopic treatment for gastric fundal variceal bleeding that was successfully managed by repeat endoscopic tissue adhesive injection combined with conservative therapy. This case may provide useful insight into the management of similar complications.

2. Case History

A 53‐year‐old Chinese man was admitted on January 11, 2020, with a chief complaint of three episodes of melena during the preceding 3 days. He had a 30‐year history of chronic hepatitis B and had undergone laparoscopic splenectomy and portal azygos vein disconnection 7 years earlier for severe hypersplenism and esophageal varices.

On admission, the patient was conscious and hemodynamically stable but appeared fatigued. Physical examination revealed palmar erythema and telangiectasia over the chest and neck. The abdomen was mildly distended, without visible abdominal wall varices or ascites. Bowel sounds were normal. There was no abdominal tenderness, no rebound pain, no lower‐extremity edema, and no asterixis. Laboratory findings are summarized in Table 1. Based on the clinical presentation and endoscopic findings, a diagnosis of liver cirrhosis with portal hypertension and esophagogastric variceal bleeding was established. Abdominal ultrasound performed on admission showed mild ascites, which resolved with diuretic therapy.

TABLE 1.

Main laboratory observations.

Characteristics	Time point					Reference range
	January 11	January 18	January 19	January 20	February 05
Hb (g/dL)	7.3	7.9		6.8	7.1	11.5–15.0
ALB (g/L)	32.6	30.9				40.0–55.0
TBIL (μmol/L)	30.0	33.9				0.0–23.0
PT (s)	15.7	16.1				10.0–13.0
FOBT			Negative		Negative

Abbreviations: ALB, albumin; FOBT, fecal occult blood test; Hb, hemoglobin; PT, prothrombin time.

3. Methods

3.1. Differential Diagnosis

The differential diagnosis included peptic ulcer bleeding, portal hypertensive gastropathy, Mallory‐Weiss tear, and gastric malignancy. Based on the clinical history of chronic hepatitis B, prior splenectomy for esophageal varices, physical findings of chronic liver disease (palmar erythema, telangiectasia), and subsequent endoscopic confirmation of esophagogastric varices, a diagnosis of liver cirrhosis with portal hypertension and esophagogastric variceal bleeding was established.

3.2. Investigations

3.2.1. Initial Endoscopy

After initial treatment with fluid replacement, proton pump inhibition, and gastric mucosal protection, esophagogastroduodenoscopy (EGD) was performed on January 13, 2020. Endoscopy showed four blue, tortuous esophageal varices in a snake‐like pattern approximately 20 cm from the incisors, with the largest measuring about 10 mm in diameter and showing a red wale sign (Figure 1A). In the gastric fundus, a blood‐mucus pool was observed, and a varix extending from the esophagus to the lesser curvature of the cardia was identified (Figure 1B).

FIGURE 1.

Ligation therapy (2020‐01‐13). (A) Esophageal varicose veins; (B) Varicose veins on the lesser gastric curvature side of the cardia; (C) After esophageal varicose ligation; (D) After ligation treatment for varicose veins on the lesser gastric curvature side of the cardia.

3.2.2. Repeat Endoscopy After Rebleeding

On the evening of January 20, the patient again passed black stool and experienced two episodes of dizziness, prompting readmission. On readmission, his hemoglobin level was 6.8 g/dL. Immediate fluid resuscitation and acid suppression therapy were initiated. After stabilization of vital signs, repeat EGD was performed on January 21. This examination revealed four blue esophageal varices approximately 20 cm from the incisors, with the thickest measuring about 10 mm in diameter and positive red wale signs. Several sites of ligation‐band detachment were noted, but there was no active esophageal bleeding (Figure 2A). In the gastric fundus, dark red blood was present, and a varix extending from the esophagus to the lesser curvature of the cardia showed a detached ligation site with an intervening bleeding point (Figure 2B).

FIGURE 2.

Injection of lauromacrogol combined with tissue glue after bleeding (2020‐01‐21). (A) Ulcer formation after esophageal ligation ring detaching; (B) The ulcer formation after the ligation ring falled off on the lesser gastric curvature side of the cardia; (C) After treatment of varicose veins on the lesser gastric curvature side of the cardia by injection of lauromacrogol and tissue glue.

3.2.3. Emergency Assessment of Perforation

On the night of January 22, the patient developed sudden abdominal pain. Physical examination demonstrated abdominal rigidity, mild tenderness, and rebound tenderness. Emergency abdominal computed tomography (CT) revealed a small amount of free intraperitoneal gas and focal discontinuity of the gastric wall along the lesser curvature of the stomach, raising strong suspicion for gastric perforation (Figure 3A).

FIGURE 3.

Glue sealing treatment of perforation of lesser gastric curvature (2020‐01‐22). (A) CT images indicated a small amount of pneumoperitoneum; (B) Wound of lesser gastric curvature side of the cardia (arrow indicates the site of suspected gastric perforation); (C) After injection of tissue glue.

Urgent bedside endoscopy showed no significant interval change in the esophageal varices. In the gastric fundus, two previous tissue adhesive injection sites were visible. An active bleeding point was seen near the posterior wall, and white gel‐like adhesive material was attached to the anterior wall. After irrigation and clearance of the field, a defect was observed near the anterior wall of the gastric fundus/body region, though definitive identification of a full‐thickness perforation was limited by surrounding edema and blood clots. The estimated size of the defect was approximately 3–5 mm in diameter (Figure 3B).

3.3. Treatment

3.3.1. Initial Endoscopic Variceal Ligation

EVL was performed using a COOK ligation device, and a total of six bands were deployed successfully (Figure 1C,D). Given the stable hemodynamic status and absence of active bleeding at initial endoscopy, EVL was chosen for primary prophylaxis against rebleeding from esophageal varices. At that time, cyanoacrylate injection was not deemed necessary as there was no active gastric fundal variceal hemorrhage. After reevaluation on hospital Days 7 and 8, the patient was discharged on the morning of January 20.

3.3.2. Endoscopic Tissue Adhesive Injection for Gastric Fundal Variceal Bleeding

Because active bleeding from the gastric fundal variceal region was identified, endoscopic tissue adhesive injection was performed. Two injection points were selected in the gastric fundal varix. At each point, a sclerosant/tissue adhesive combination was administered according to the institutional protocol, followed by visible obliteration of the variceal lumen (Figure 2C). The patient was then placed on fasting status and received standard postprocedural supportive treatment.

3.3.3. Endoscopic Sealing of Gastric Perforation

Endoscopic clipping was attempted but failed because the local tissue was markedly swollen and friable, precluding effective approximation of the defect.

Given the small size of the suspected perforation and the patient's relatively stable condition, an endoscopic salvage approach was adopted. A sequential injection protocol consisting of 3 mL of 50% glucose solution, followed by 1.5 mL of N‐butyl‐2‐cyanoacrylate tissue adhesive, and then another 3 mL of 50% glucose solution was administered. Two injection sets were delivered to the anterior wall and one set to the posterior wall. After treatment, no active bleeding was observed at either site, and the defect appeared sealed (Figure 3C). The patient's abdominal pain lessened during the procedure, and abdominal rigidity improved.

3.3.4. Conservative Management

Following endoscopic management, the patient was treated conservatively with 72 h of fasting, intravenous levofloxacin, intravenous esomeprazole, and continuous somatostatin infusion to suppress gastric acid secretion and reduce portal pressure. His abdominal symptoms gradually resolved.

3.4. Outcome and Follow‐Up

3.4.1. Outcome

Following endoscopic sealing and conservative management, the patient's abdominal symptoms gradually resolved. His abdominal pain improved promptly after the procedure, and abdominal rigidity resolved.

3.4.2. Follow‐Up

Follow‐up CT on Day 25 showed complete resolution of the pneumoperitoneum (Figure 4A). Repeat endoscopy demonstrated posttreatment changes at the gastric fundus without significant residual gastric varices (Figure 4B). Additional EVL sessions were subsequently performed for the esophageal varices (Figure 4C,D). Before discharge, his hemoglobin level was 7.1 g/dL, and fecal occult blood testing was negative.

FIGURE 4.

Endoscopic sequential treatment for varicose veins (2020‐02‐05). (A) CT images indicated peritoneal gas absorption; (B) The change of gum emesis and the disappearance of varicose veins at the lesser gastric curvature side of the cardia; (C) Esophageal varicose veins; (D) After esophageal varicose ligation.

At 1‐month follow‐up, the patient remained asymptomatic with no evidence of recurrent variceal bleeding or abdominal complications. Repeat EGD confirmed complete resolution of posttreatment changes without residual varices. He continued secondary prophylaxis with nonselective beta‐blockers and scheduled EVL sessions every 2 weeks. A detailed timeline outlining the clinical course progression is summarized in Table 2.

TABLE 2.

Timeline of clinical course.

Day	Date	Clinical event
Day 0	January 11	Patient admitted with 3‐day history of melena
Day 0–1	January 11–12	Initial resuscitation, proton pump inhibition, gastric mucosal protection, laboratory evaluation
Day 2	January 13	EGD revealed severe esophageal varices and gastric fundal varices; EVL performed with six bands
Day 7–8	January 18–19	Reevaluation performed; patient clinically stable
Day 9 (morning)	January 20	Discharged from hospital
Day 9 (evening)	January 20	Recurrent melena and dizziness; readmitted
Day 10	January 21	Hb 6.8 g/dL; repeat EGD showed gastric fundal variceal bleeding; tissue adhesive injection performed
Day 11 (night)	January 22	Sudden abdominal pain with rigidity and rebound tenderness
Day 11	January 22	Emergency CT showed pneumoperitoneum and suspected gastric wall discontinuity
Day 11	January 22	Emergency bedside endoscopy identified suspected perforation; clipping failed; tissue adhesive sealing performed
Day 12–14	January 23–25	Conservative management with fasting, antibiotics, esomeprazole, and somatostatin
Day 25	—	Follow‐up CT showed resolution of pneumoperitoneum
Day 25	—	Repeat endoscopy showed posttreatment changes without significant residual gastric varices; additional EVL performed
Before discharge	—	Hb 7.1 g/dL; fecal occult blood test negative
Follow‐up plan	1 month later	Return visit advised for further endoscopic treatment as secondary prophylaxis

4. Conclusions

Gastric perforation after endoscopic treatment of gastric fundal variceal bleeding is a rare but potentially life‐threatening complication. This case demonstrates that in carefully selected patients with a small, early perforation, limited pneumoperitoneum, and no evidence of generalized peritonitis, endoscopic tissue adhesive injection combined with strict conservative management may provide a feasible nonsurgical rescue option. Further studies are needed to clarify the indications, safety profile, and long‐term outcomes of this strategy.

5. Discussion

This case describes a rare but serious complication of endoscopic treatment for gastric fundal variceal bleeding, namely gastric perforation after endoscopic injection therapy, which was successfully treated without surgery. The case is notable not only because of the rarity of the complication, but also because it demonstrates the feasibility of endoscopic tissue adhesive sealing as a rescue approach in a carefully selected patient.

In this patient, rebleeding occurred 8 days after the index endoscopic therapy, which is within the reported time frame for post‐banding ulcer bleeding [13, 14]. At repeat endoscopy, no active esophageal bleeding was identified; however, active bleeding was present in the gastric fundal variceal region, associated with a detached ligation site. This finding suggests that EVL may be inadequate for gastric fundal varices in some patients, particularly when varices extend from the esophagus into the gastric cardia or fundus. Consistent with prior reports, cyanoacrylate injection is generally more suitable than EVL for gastric fundal variceal bleeding because of its higher immediate hemostatic efficacy and lower rebleeding rate [15, 16, 17].

The reported immediate hemostatic rate of cyanoacrylate injection for gastric fundal varices is high, commonly exceeding 90% [11]. Rebleeding rates vary according to gastric variceal subtype [10]. Despite these favorable outcomes, tissue adhesive injection may cause important complications, including local inflammation, mucosal necrosis, ulceration, bacteremia, early glue extrusion, and systemic embolization.

The mechanism of gastric perforation in the present case was likely multifactorial. First, previous band ligation may have caused local ischemia, ulceration, and weakening of the gastric wall. Second, the injected sclerosant may have aggravated endothelial injury and sterile inflammatory necrosis, particularly in tissue already compromised by recent ligation and rebleeding. Third, repeated intervention at the same vulnerable site may have further increased local tissue fragility. The temporal relationship between injection therapy and sudden abdominal pain, together with CT findings of gastric wall discontinuity and pneumoperitoneum, strongly supports the diagnosis of procedure‐related perforation.

Perforation after endoscopic treatment of esophagogastric variceal bleeding is extremely uncommon, and there is currently no standardized treatment algorithm. Conventionally, acute gastric perforation is managed surgically. However, with recent advances in therapeutic endoscopy, selected iatrogenic gastrointestinal perforations can be managed nonoperatively when diagnosed early and when contamination is limited [18, 19].

In the present case, several factors supported an endoscopic and conservative rather than immediate surgical approach. First, CT demonstrated only a small pneumoperitoneum, suggesting a limited perforation. Second, the patient had already been fasting, which likely minimized leakage of gastric contents into the peritoneal cavity. Third, there was no fever, hemodynamic collapse, or evidence of diffuse peritonitis or sepsis, indicating that the perforation had been recognized early. Fourth, the defect was thought to be small and anatomically localized near the lesser curvature, where support from the lesser omentum may facilitate containment. Finally, immediate endoscopic access allowed direct assessment and attempted closure.

The decision for nonoperative management was made after multidisciplinary discussion involving gastroenterology and surgical teams. Given the patient's underlying cirrhosis (Child‐Pugh class B), prior splenectomy, and limited peritoneal contamination, he was considered a high‐risk candidate for surgery. Endoscopic clipping was attempted initially but failed because the tissue surrounding the defect was too edematous and friable. Under these circumstances, tissue adhesive injection was used to achieve physical sealing of the defect. The subsequent clinical course was favorable: the patient's abdominal pain improved promptly, follow‐up CT confirmed resolution of free air, and later endoscopy showed healing of the treated area without recurrent gastric variceal bleeding. These findings suggest that endoscopic tissue adhesive sealing may be a feasible rescue option in highly selected patients with early, small perforations and minimal peritoneal contamination.

While endoscopic tissue adhesive injection achieved successful sealing in this case, its use in ischemic or edematous tissues carries potential risks. Prior studies have reported localized necrosis, foreign body granuloma formation, and delayed intra‐abdominal abscess following cyanoacrylate injection. In our patient, the risk of such complications was weighed against the benefits of avoiding emergency surgery. Close monitoring for signs of infection and delayed imaging follow‐up were implemented to mitigate these concerns. This case also carries important clinical implications. First, tissue adhesive injection may be preferable to EVL for active gastric fundal variceal bleeding. Second, in patients with varices extending from the esophagus to the stomach, individualized combined strategies may be required rather than reliance on ligation alone. Third, close postprocedural monitoring is essential, particularly in cases of rebleeding after EVL. Fourth, when local tissue damage is already present, caution should be exercised in selecting sclerosants that may exacerbate inflammation or tissue necrosis. The mechanism by which tissue adhesive facilitated closure likely involves both mechanical sealing and biological healing processes. N‐butyl‐2‐cyanoacrylate polymerizes rapidly upon contact with bodily fluids, forming a solid plug that physically occludes the defect. This is followed by a localized inflammatory response that stimulates fibroblast proliferation and collagen deposition, promoting granulation tissue formation. Concurrent use of proton pump inhibitors and somatostatin reduces gastric acidity and splanchnic blood flow, creating a favorable microenvironment for mucosal healing.

To our knowledge, reports describing the use of endoscopic tissue adhesive injection to manage gastric perforation after endoscopic treatment for gastric variceal bleeding are extremely limited. Nonetheless, this experience should be interpreted cautiously. This approach should not replace surgery in patients with large perforations, diffuse peritonitis, uncontrolled sepsis, hemodynamic instability, or extensive contamination. Multidisciplinary evaluation and availability of emergency surgery remain essential.

Author Contributions

Tianwen Liu: methodology, conceptualization. Shumin Qin: conceptualization, investigation, data curation, writing – original draft, writing – review and editing. Jingyi Xu: data curation. Xiaofeng Lin: data curation. Yanfeng Lin: methodology, data curation. Shuting Wen: data curation. Zhaoli Fu: methodology, data curation.

Funding

The authors have nothing to report.

Ethics Statement

The authors have nothing to report.

Consent

Written informed consent was obtained from the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor‐in‐Chief of this journal.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are delivered by the corresponding author, upon reasonable request.