Comparative safety and efficacy of endoscopic band ligation and variceal obturation for gastric varices: Systematic review and meta-analysis

Abstract

BACKGROUND

Gastric variceal hemorrhage (GVH) is among the most life-threatening complications of portal hypertension. Endoscopic variceal ligation (EVL) and cyanoacrylate injection (endoscopic variceal obturation) are the main endoscopic options for hemostasis, but their comparative efficacy and safety remain debated due to variable evidence.

AIM

To investigate the safety and efficacy of endoscopic band ligation vs endoscopic variceal obturation in the management of gastric varices.

METHODS

PubMed, ClinicalTrials.gov, and EMBASE were systematically searched up to September 20, 2025. Eligible studies were randomized controlled trials (RCTs) comparing EVL with cyanoacrylate therapy in patients with GVH. Primary outcomes included initial hemostasis, rebleeding, treatment failure, mortality, variceal obliteration, and adverse events. Pooled risk ratios (RRs) with 95% confidence intervals (CIs) were calculated using a random-effects model (RevMan 5.4), with subgroup analyses for cyanoacrylate injection vs obturation.

RESULTS

Seven RCTs (n = 631) were included. There was no significant difference between EVL and cyanoacrylate in achieving initial hemostasis (RR = 0.96, 95%CI: 0.88-1.04) or overall mortality (RR = 1.29, 95%CI: 0.84-1.98). However, rebleeding risk was significantly higher with EVL (RR = 1.53, 95%CI: 1.03-2.26), particularly when compared with cyanoacrylate injection (RR = 2.27, 95%CI: 1.18-4.40). No significant differences were found in treatment failure, variceal obliteration, or fatal bleeding. Post-procedural fever was more common following cyanoacrylate use (RR = 0.55, 95%CI: 0.31-0.94).

CONCLUSION

Cyanoacrylate injection provides superior rebleeding control in active GVH without increasing mortality. EVL remains a reasonable alternative in settings where cyanoacrylate or operator expertise is limited. These findings support cyanoacrylate as the preferred endoscopic therapy where resources and training permit.

Key Words: Gastric varices; Portal hypertension; Endoscopic band ligation; Endoscopic variceal obturation; Upper gastrointestinal bleeding; Hemostasis; Rebleeding

Core Tip: This meta-analysis of seven randomized controlled trials establishes that while endoscopic variceal ligation (EVL) and endoscopic variceal obturation (EVO) achieve similar initial hemostasis rates for gastric varices, EVO demonstrates superior long-term durability. Specifically, cyanoacrylate injection significantly reduces the risk of rebleeding compared to EVL without increasing overall mortality. These findings support prioritizing cyanoacrylate as the first-line intervention for gastric variceal hemorrhage, reserving EVL as an alternative only when resources for obturation are unavailable.

INTRODUCTION

The formation of gastroesophageal varices (GOV), which are abnormal blood vessel connections, is one of the several complications of advanced liver cirrhosis[1]. These varices are the most dangerous from a clinical standpoint because they tend to bleed. Healthcare professionals have consistently and exhaustively documented the treatment of esophageal varices. However, treating gastric varices remains a challenge due to the unpredictable nature of bleeding of the varices, the diversity of the varices[1] location, and the generally less effective treatment alternatives. A rupture in the gastric varices results in a medical crisis as it causes significant blood loss. Bleeding gastric varices pose a risk of fatality higher than that associated with bleeding esophageal varices.

Gastric varices occur less frequently than esophageal varices; research suggests that the prevalence of gastric varices is between 20% and 25% in patients with cirrhosis and portal hypertension[2]. However, the impact of gastric varices after a bleeding incident is very serious[3-5]. A considerable portion of patients, between 4% and 65%, are likely to suffer a second bleeding incident within two years. These bleeding episodes cause severe hypotension, with death rates ranging from 15% to 30% and reaching 20% within the first six weeks after the initial bleeding event[6]. Repeat bleeding happens frequently during the period immediately after the first bleed, with the highest risk occurring in the following six weeks, making it essential to achieve complete initial bleeding control and implement strong prevention strategies.

Sarin’s classification system[7] has significantly altered doctors perception of gastric varices by categorizing them into GOVs (GOV1 that travel along the lesser curvature of the stomach; GOV2 that reach into the fundus) and isolated gastric varices (IGV) (IGV1 only present in the fundus; IGV2 found in the body, antrum, or pylorus)[8]. In addition to Sarin’s classification, the Hashizume classification categorizes gastric varices based on venous drainage patterns: Type I, without a major portosystemic shunt; Type II, with a gastrorenal shunt; Type III, with a gastrocaval shunt[8]. This classification is not merely descriptive but holds clinical significance. Accumulating evidence indicates that various subtypes possess different bleeding risks and show diverse treatment responses, with IGV1 lesions exhibiting the most aggressive behavior and elevated bleeding and rebleeding rates.

Present international guidelines endorse endoscopic intervention as the main strategy following initial stabilization through fluids and blood transfusions[9,10]. Nonetheless, this agreement falters when it comes to gastric varices, as there is a lack of clear evidence-based guidelines and ongoing discussions regarding the optimal endoscopic technique.

Endoscopic sclerotherapy, once the first treatment choice for variceal bleeding, is no longer used for gastric varices because of extremely high treatment failure and rebleeding rates[2]. These poor results have driven the search for better endoscopic methods, mainly endoscopic band ligation and tissue glue injection using cyanoacrylate. Endoscopic band ligation, which has proven highly effective for esophageal varices and is recommended by the Baveno V consensus meeting[11], works by cutting off the blood supply through mechanical compression. While it works reasonably well for GOV1 and small GOV2 lesions, its use for large fundal varices is limited by low success rates of only 26%-51% and high rebleeding rates of 60%-90%.

In contrast, N-butyl-2-cyanoacrylate injection has become an attractive alternative, achieving immediate sealing of varices through rapid hardening and showing success rates above 90% for different types of gastric varices. Current expert guidelines recommend cyanoacrylate as the preferred treatment for GOV1, GOV2, and IGV1 lesions when the necessary skills and materials are available[12]. Despite its effectiveness, cyanoacrylate remains unavailable in many resource-limited hospitals, where endoscopic band ligation becomes the default treatment because of its simpler technique and lower equipment needs.

A previous meta-analysis[13] revealed cyanoacrylate to be superior in many aspects in the management of gastric varices. Nevertheless, the small number of clinical trials and patients included in the meta-analysis confounds the certainty of the results. The available research includes several randomized controlled trials (RCTs) that show mixed and sometimes conflicting results when comparing these two endoscopic techniques[14-20]. Additionally, detailed analyses based on Sarin’s classification system[7] are notably missing, which limits how well existing evidence can be applied to specific variceal types. This systematic review and meta-analysis aims to collect and combine all available comparative evidence examining cyanoacrylate injection vs endoscopic band ligation for acute gastric variceal bleeding, to establish clear, evidence-based treatment guidelines for this life-threatening complication of portal hypertension.

MATERIALS AND METHODS

Study design and protocol registration

This systematic review and meta-analysis were conducted according to the guidelines of the Cochrane Handbook of Systematic Reviews for Interventions[21] and reported in accordance with the preferred reporting items for systematic review and meta-analyses (PRISMA) guidelines[22]. The protocol for the study has been registered and is available on PROSPERO with the ID: No. CRD420251164433

Search strategy and databases

An electronic search of PubMed and Cochrane databases, EMBASE, and the Clinical trials registry was carried out systematically, including all entries having the relevant keywords and MeSH terms, from inception till 20^th August, 2025, by 2 independent authors (Ihtasham A and Khattak F). All articles relevant to the safety and efficacy of cyanoacrylate injection and band ligation in the treatment of gastric variceal bleeding using appropriate terms were screened using the following keywords: “Cyanoacrylate injection”, “band ligation”, “gastric variceal bleeding”, etc. The detailed search strategy for each database is presented in the Supplementary Table 1. We also manually screened the references of the included articles to identify any studies that we could include.

Eligibility criteria

We included studies that were according to the following inclusion criteria (PICOT): P (population): Patients of any age with active gastric variceal bleeding. I (intervention): Procedures involving band ligation. C (control): Procedures involving the administration of cyanoacrylate injection. O (outcomes): Initial hemostasis/control of active bleeding, re-bleeding (post-treatment event), etc. T (type of study): RCTs. Studies published in the English language were included only.

Furthermore, the exclusion criteria encompassed the following: Studies with insufficient data or unclear reporting making it challenging to extract the necessary information or assess the study’s validity and reliability; Studies involving animal models or experiments, non-randomized studies, open-label designs, case reports, editorials, commentaries, letters, conference abstracts, reviews, opinion pieces, and other non-peer-reviewed publications; Studies observed to have a high risk of bias or poor methodological quality after the quality assessment process.

Study selection

The articles retrieved after the electronic search were imported to Rayyan software for screening. The software automatically detected duplicated entries and were removed by the author (Iman A). The remaining articles were screened based on the title and abstracts by two independent reviewers (Iman A and Tufail U) with a third author (Amin F) independently resolving any arising conflicts. Studies not fulfilling the inclusion criteria were removed, and a comprehensive full-text screening of the remaining articles was performed in accordance with our inclusion criteria. The selection process is detailed in the PRISMA flowchart for our study in Figure 1.

Figure 1 Preferred reporting items for systematic review and meta-analyses flow diagram of study selection process. Flowchart summarizing the identification, screening, eligibility assessment, and inclusion of randomized controlled trials comparing endoscopic variceal ligation vs cyanoacrylate therapy for gastric variceal hemorrhage.

Data extraction and outcomes

After the study selection process, the data of the included studies was extracted by two authors (Tufail U and Khalid E) into a pre-piloted standardized Google spreadsheet, with the third author (Khattak F) resolving any discrepancies independently. The sheet included three sections: Baseline characteristics: Study ID, year of publication, number of patients, age, sex, hepatocellular carcinoma presence, Child-Pugh score, albumin, bilirubin, hemoglobin, platelets, creatinine, prothrombin, ascites, encephalopathy, sepsis, gastric varices (GV) type, GV size, concomitant esophageal varices, active bleeding, hemodynamic instability and blood transfusion units. Outcomes: Initial hemostasis/control of active bleeding, re-bleeding (post-treatment event), death (overall mortality), number of sessions required for variceal obliteration, blood transfusion units (mean units/patient during specific events), treatment failure, variceal obliteration rate, fever (complication), recurrence rate (post-eradication), fatal/ulcer bleeding (complication related to procedure/ulcer), chest pain (complication), dysphagia (complication), hepatic encephalopathy (long-term complication), hepatorenal syndrome (long-term complication), sepsis (complication).

Quality assessment and risk of bias

The risk of bias in our study was evaluated using the revised Cochrane collaboration tool for RCTs (ROB2.0)[23] for all the RCTs. The evaluation indicators included randomization sequence generation, allocation concealment, blinding, incomplete outcome data, and selective reporting. This tool classified the risk of bias in the included studies as low, some concerns, or high. Traffic light plots and summary plots were generated using the Robvis tool[24].

Statistical analysis

The data were statistically analyzed using RevMan (Computer Program), Version 5.4 (Nordic Cochrane Centre, Copenhagen, Denmark). The findings were displayed in terms of risk ratios (RRs) for the dichotomous outcomes and MD for all continuous outcomes, each with 95% confidence intervals (CIs). Effect sizes were pooled using a random-effects model, while the results of the pooled analyses were visualized using forest plots. The heterogeneity among the studies was assessed by I² statistics (I² values < 50 % = low heterogeneity; I² values 50%-75% = moderate heterogeneity; I² values > 75% = significant heterogeneity). For heterogeneity lower than 50%, the fixed effects model was employed[25]. Publication bias could not be assessed using funnel plots or statistical tests, as the pooled studies were < 10, as recommended by the Cochrane collaboration. A P value of < 0.05 represented statistical significance[26].

RESULTS

Study selection

The initial database search yielded 165 records. After removing 27 duplicates, 138 unique records underwent title and abstract screening. Of these, 131 records were excluded for irrelevance, leaving 7 studies for full-text review. All 7 RCTs met the inclusion criteria and were incorporated into both the qualitative synthesis and quantitative meta-analysis (Figure 1, PRISMA 2020 flow diagram).

Study characteristics

The included RCTs were published between 2001 and 2025 and collectively enrolled 631 participants [315 in the endoscopic variceal ligation (EVL) group and 316 in the cyanoacrylate group]. Study settings spanned Egypt, China, Taiwan, and Pakistan, with sample sizes ranging from 58 to 154 patients. Interventions compared EVL (EVL/gastric variceal ligation) against endoscopic variceal obturation (EVO) or cyanoacrylate obturation. All studies evaluated adult patients presenting with acute gastric variceal hemorrhage (GVH) secondary to portal hypertension. All included studies used endoscopic confirmation of GVH at enrollment and comparable definitions for hemostasis and rebleeding. The baseline demographic characteristics and methodological details are summarized in Supplementary Tables 2 and 3. Prevalence of cirrhosis, non-cirrhotic portal hypertension, and ascites in included studies is provided in Supplementary Table 4.

Risk of bias assessment

Two studies (Tan et al[15]; Shi et al[18]) were rated as low risk of bias, two (El Amin et al[16] and Hassan et al[27]) had some concerns, and the remaining three were judged at high risk (Figure 2). Most trials demonstrated adequate randomization (5/7), but blinding was poorly reported. Missing outcome data were minimal in five studies, though outcome measurement and reporting bias were frequent sources of concern. In total, over one-third of included trials were considered at moderate-to-high overall risk of bias.

Figure 2 Risk of bias summary of included studies. Evaluation of methodological quality using the RoB2.0 tool across five domains: Randomization, deviations from intended interventions, missing outcome data, measurement of outcomes, and selective reporting. Generated using the robvis visualization tool. Low risk = green; some concerns = yellow; high risk = orange.

Primary outcomes

Initial hemostasis: All seven RCTs (n = 538 treatment arms) reported data on initial hemostasis. A total of 236 events occurred in the EVL group and 259 in the cyanoacrylate group. The pooled analysis demonstrated no significant difference between EVL and cyanoacrylate for achieving initial hemostasis (RR = 0.96; 95%CI: 0.88-1.04). Substantial heterogeneity was detected (I² substantial; τ² = 0.01; Q = 24.96, P < 0.05), prompting the use of a random-effects model (Figure 3). Subgroup analysis: Cyanoacrylate technique. To evaluate the influence of cyanoacrylate delivery methods, a predefined subgroup analysis was conducted comparing cyanoacrylate injection and obturation techniques. Cyanoacrylate injection subgroup (Figure 4): Three studies showed a non-significant trend favoring cyanoacrylate injection over EVL (RR = 0.92; 95%CI: 0.79-1.06), with moderate heterogeneity. Obturation subgroup (Figure 4): Four studies demonstrated no significant difference between the two treatments (RR = 1.00; 95%CI: 0.91-1.09) with substantial heterogeneity. The test for subgroup differences was non-significant, indicating that the method of cyanoacrylate application did not influence efficacy relative to EVL.

Figure 3 Forest plot comparing initial hemostasis between endoscopic variceal ligation and cyanoacrylate therapy. Pooled risk ratio (RR) with 95% confidence interval (CI) calculated using a random-effects model. No significant difference was observed between groups (RR = 0.96; 95%CI: 0.88-1.04; I² = 67%). Generated using a random-effects model. CI: Confidence interval.

Figure 4 Subgroup analysis of initial hemostasis by cyanoacrylate delivery technique (injection vs obturation). Forest plot comparing endoscopic variceal ligation vs cyanoacrylate subgroups. Both methods achieved similar hemostasis rates without significant subgroup differences (P > 0.05). Injection and obturation defined as per included randomized controlled trial protocols. CI: Confidence interval.

Overall rebleeding: Seven studies (n = 631 participants) provided data on overall rebleeding. The pooled analysis showed a significantly higher risk of rebleeding in the EVL group compared with cyanoacrylate (RR = 1.53; 95%CI: 1.03-2.26; P < 0.05). Heterogeneity was low (I² approximately 0%) (Figure 5), indicating consistent findings across studies. This suggests that cyanoacrylate offers superior efficacy in maintaining hemostasis and preventing recurrent bleeding.

Figure 5 Forest plot comparing overall rebleeding between endoscopic variceal ligation and cyanoacrylate therapy. Cyanoacrylate therapy was associated with a significantly lower rebleeding rate (risk ratio = 1.53; 95% confidence interval: 1.03-2.26; I² = 0%). Positive risk ratio favors cyanoacrylate. CI: Confidence interval.

Subgroup analysis: According to Sarin’s classification (Figure 6). Subgroup analysis was carried out for the rebleeding rate according to Sarin’s classification of GV. A subgroup analysis was conducted to evaluate the efficacy of ligation vs cyanoacrylate injection or obturation based on the type of gastric varices: IGV1, GOV1, and GOV2.

Figure 6 Impact of gastric varices type on the efficacy of ligation vs cyanoacrylate injection/obturation. The forest plot displays risk ratios (RR) for isolated gastric varices type 1 (IGV1) and gastroesophageal varices types 1 and 2 (GOV1, GOV2). Cyanoacrylate was associated with a statistically significant reduction in events overall (RR = 1.72, P = 0.02) and specifically in the GOV1 subgroup (RR = 1.83, P = 0.04). CI: Confidence interval; IGV: Isolated gastric varices; GOV: Gastroesophageal varices.

The overall analysis across all subgroups favored cyanoacrylate injection or obturation, demonstrating a statistically significant RR of 1.72 (95%CI: 1.10-2.70; P = 0.02). Low overall heterogeneity was observed (I² = 3%, P = 0.40).

The specific findings for each subgroup are detailed below: GOV1 subgroup: This was the only subgroup to show a statistically significant difference between treatments. The analysis of 90 total participants resulted in a RR of 1.83 (95%CI: 1.03-3.27; P = 0.04), indicating a higher risk of events in the ligation group compared to the cyanoacrylate group. No heterogeneity was detected within this subgroup (I² = 0%, P = 0.66). IGV1 subgroup: While the RR was high at 3.61, the results were not statistically significant due to a very wide CI (95%CI: 0.39-33.64; P = 0.26). This subgroup also exhibited substantial heterogeneity (I² = 75%, P = 0.05). GOV2 subgroup: No significant difference was found between the two interventions in this group, with a RR of 1.07 (95%CI: 0.43-2.64; P = 0.89). Heterogeneity was non-existent (I² = 0%, P = 0.74). Despite the varying levels of significance within individual subgroups, the test for subgroup differences (χ² = 1.49, df = 2, P = 0.47, I² = 0%) suggests that the type of gastric varices did not significantly modify the treatment effect. Subgroup analysis: Cyanoacrylate technique (Figure 7). Injection subgroup: EVL was associated with a significantly higher rebleeding risk (RR = 2.27; 95%CI: 1.18-4.40; I² = 0%). Obturation subgroup: No significant difference was observed between treatments (RR = 1.24; 95%CI: 0.710-2.17; I² moderate). The test for subgroup differences was non-significant (P > 0.05), confirming comparable protection against rebleeding between the two cyanoacrylate techniques.

Figure 7 Subgroup analysis of rebleeding by cyanoacrylate delivery method. Cyanoacrylate injection demonstrated superior prevention of rebleeding [risk ratios (RR) = 2.27; 95% confidence interval (CI): 1.18-4.40; I² = 0%], while obturation showed no significant difference (RR = 1.24; 95%CI: 0.71-2.17; I² = 43%). Test for subgroup differences was non-significant (P > 0.05). CI: Confidence interval.

Overall mortality: Six studies (n = 637 participants) reported data on all-cause mortality. The pooled estimate showed no statistically significant difference between EVL and cyanoacrylate (RR = 1.29; 95%CI: 0.84-1.98; P > 0.05), with low heterogeneity (I² < 25%) (Supplementary Figure 1). This indicates that neither treatment confers a survival advantage. Subgroup analysis (Supplementary Figure 2). Injection subgroup: RR = 0.82 (95%CI: 0.13-5.03), with high heterogeneity. Obturation subgroup: RR = 1.28 (95%CI: 0.97-1.70), with no heterogeneity. No statistically significant difference was observed between subgroups (P > 0.05).

Secondary outcomes

Number of sessions required: Five studies (n = 518) compared the number of sessions needed for variceal obliteration. No significant difference was found between EVL and cyanoacrylate (MD = 0.31 sessions; 95%CI: -0.19 to 0.80). Heterogeneity was high (I² substantial; P < 0.001) (Supplementary Figure 3).

Fatal bleeding: Three studies (n = 270) reported fatal bleeding events. No significant difference was observed between groups (RR = 0.80; 95%CI: 0.08-8.35; moderate heterogeneity), implying comparable safety profiles (Supplementary Figure 4).

Treatment failure: Four studies (n = 367) reported treatment failure. The pooled estimate showed no significant difference (RR = 1.42; 95%CI: 0.66-3.09; I² < 25%), suggesting similar overall success rates (Supplementary Figure 5).

Variceal obliteration rate: Three studies (n = 308) reported variceal obliteration. The pooled RR = 1.03 (95%CI: 0.96-1.10), indicating no difference between EVL and cyanoacrylate (I² = 0%) (Supplementary Figure 6).

Post-procedural fever: Four studies (n = 458) evaluated post-procedural fever. EVL was associated with a significantly lower risk of fever compared with cyanoacrylate (RR = 0.55; 95%CI: 0.31-0.94; I² = 0%), reflecting fewer inflammatory or embolic events (Supplementary Figure 7).

Blood transfusion requirements: Three studies (n = 171) reported transfusion requirements. No significant difference was found (MD = 1.78 units; 95%CI nonsignificant), though extreme heterogeneity (I² > 75%) warrants cautious interpretation (Supplementary Figure 8).

DISCUSSION

This systematic review with meta-analysis reports the latest comparative effectiveness between EVL and EVO to control acute GVH. By pooling data from seven RCTs that enrolled a combined total of 631 participants, our study establishes an evidence-based preference between the long-term success of one modality over the other. Any generalized conclusions, however, must be mindful of the methodologically based risk of bias as well as the technical heterogeneity found across included trials.

Analysis of primary outcomes revealed a critical distinction between immediate hemorrhage control and long-term recurrence prevention. In the setting of acute bleeding, initial hemostasis rates were statistically comparable between EVL and cyanoacrylate (RR = 0.96; 95%CI: 0.88-1.04). This parity confirms that either technique can successfully achieve immediate stabilization. Nevertheless, this hemostasis endpoint displayed considerable statistical variability (I² approximately 76%), likely reflecting the wide range of technical approaches and the inclusion of diverse gastric varices subtypes across the contributing centers[14,15,18].

The finding concerning recurrent hemorrhage carries greater clinical weight. The pooled data showed that EVL had a statistically significant risk of rebleeding significantly higher compared to cyanoacrylate (RR = 1.53; 95%CI: 1.03-2.26; P < 0.05). Although heterogeneity for this outcome was statistically low across all seven trials (I² approximately 0%), the modest magnitude of the effect and the documented risk of bias necessitate cautious interpretation. Furthermore, subgroup analysis showed that the observed superiority was not uniform: The benefit was clearly concentrated in the injection subgroup (RR = 2.27; 95%CI: 1.18-4.40), whereas the obturation subgroup showed a non-significant trend (RR = 1.24; 95%CI: 0.71-2.17). This confirms that cyanoacrylate, particularly delivered via injection, provides a robust, statistically verifiable advantage in maintaining durable hemostasis. Evaluation of secondary outcomes found that all-cause mortality (RR = 1.29; 95%CI: 0.84-1.98) and the total number of sessions required for obliteration (MD = 0.31 sessions; 95%CI: -0.19 to 0.80) were statistically equivalent. Conversely, EVL showed a lower risk of post-procedural fever (RR = 0.55; 95%CI: 0.31-0.94), a non-fatal safety benefit suggestive of fewer inflammatory complications compared to EVO.

Our updated results strengthen the provisional recommendations made by the earlier meta-analysis by Qiao et al[13]. The inclusion of larger, modern RCTs, such as Shi et al[18] and Hassan et al[27], provided the necessary power to substantiate the long-term superiority of obturation. The consistent divergence in long-term efficacy is a consequence of the interventions’ differing mechanical approaches to GVH pathophysiology. Gastric varices, particularly fundal lesions (GOV2 and IGV1), are anatomically complex, with deep submucosal vessels, high pressures, and large shunting vessels[7]. Cyanoacrylate achieves obliteration by instant intravariceal polymerization, creating a permanent, rigid cast that isolates the varix from the high portal pressure system[28]. EVL, however, relies on superficial strangulation. While this stops acute flow, it does nothing to tackle the vessels in their deepest root. This shallower form of intervention is prone to the acidic, highly contractile environment of the stomach, with early band detachment that presents as big, deep ulcers[14,15]. This failure mechanism often leads to catastrophic rebleeding. Shi et al[18] highlighted this danger, noting that among recurrence cases, EVL patients required significantly higher average blood transfusion volumes (6.80 ± 1.30 units vs 2.33 ± 0.58 units; P = 0.024). This evidence establishes that EVL’s failure to prevent recurrence is less a technical flaw and more a predictable biological limitation when applied to the unique anatomy of high-flow gastric varices.

Endoscopic ultrasound-guided cyanoacrylate injection offers a more precise approach by allowing clinicians to see and target both the gastric varix and the specific “feeding” veins behind it. By using doppler flow to confirm that the entire vascular structure is fully occluded, endoscopists can achieve better obliteration rates and lower rebleeding risks than they would with standard endoscopic injection[29]. When endoscopic methods aren’t enough or only provide a temporary fix, salvage options like transjugular intrahepatic portosystemic shunt, balloon-occluded retrograde transvenous obliteration, or transhepatic embolization are indicated to achieve definitive hemostasis[30,31]. Additionally, when looking at different gastric varix subtypes via Sarin’s classification, cyanoacrylate consistently reduced rebleeding risk (RR = 1.72; 95%CI: 1.10-2.70), while results for IGV1 weren’t statistically significant, likely reflecting limited sample size rather than loss of treatment effect.

The analysis included studies spanning Egypt, China, Taiwan, and Pakistan, which helps interpret the process outcomes that displayed heterogeneity. Crucially, clinical decision-making must incorporate variceal subtype analysis. The trial by El Amin et al[16] evaluated solely the junctional GOV1 lesions, homologically similar to esophageal varices. The one study implicated that EVL could be a realistic substitute for this particular subpopulation of smaller size. This, however, remains one individual observation that needs to be verified during future RCTs in pre-specified subgroup analysis. On the contrary, the uniform evidence indicates superior recurrence control obtained with the use of cyanoacrylate cements, the EVO as the standard method of management. The meta-analysis is consistent with the current practice statements (Baveno[32], American Association for the Study of Liver Diseases[33]), recommending the use of cyanoacrylate injection as one of the first-line endoscopic treatments of bleeding GOV2 and IGV1. Efficacy is balanced by safety: Systemic embolization, including pulmonary embolism[34], remains a rare but potentially fatal risk. The presence of large, high-flow portosystemic shunts is the primary risk factor[34], a danger underscored by the fatality (1/154 patients) noted in the Shi et al[18] trial. Therefore, clinical implementation mandates aggressive pre-procedural imaging for shunt assessment to mitigate embolic risk[33].

This meta-analysis is subject to limitations inherent to the original evidence, notably the methodological quality, where over one-third of included RCTs were rated at moderate-to-high risk of bias. Further, technical heterogeneity persists, specifically regarding the varied EVO injection protocols and the types of EVL devices used across trials (early vs large-volume ligators). Future multicenter RCTs should prioritize: (1) Direct comparison of optimized EVL techniques against standard EVO for GOV1 lesions; (2) Effectiveness studies pitting advanced modalities, such as endoscopic ultrasound-guided coiling combined with cyanoacrylate, against conventional EVO for managing high-risk shunts; and (3) Collection of robust, long-term durability and late recurrence data (beyond one year), a gap consistently observed in the literature.

CONCLUSION

In summary, this meta-analysis provides statistically consistent evidence of the mechanical inferiority of EVL compared to cyanoacrylate injection in preventing recurrent bleeding from gastric varices. Although rates of acute hemostasis are similar, EVO offers a strong, highly reproducible benefit to prevent rebleeding to justify acceptance as a first-line endoscopic choice for GOV2/IGV1 in high-volume centers. End clinical management is judicious, combining aggressive risk assessment with particular screening to exclude portosystemic shunts, to avoid the significant, though uncommon, embolic dangers inherent in the obturation therapies.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the contributions of all investigators, clinical staff, and supporting institutions involved in this study. Their assistance in data collection, analysis, and manuscript preparation is sincerely appreciated.