Aggressive or conservative cyanoacrylate therapy for gastric varices: is more glue better?

Gastric variceal bleeding remains one of the most severe complications of portal hypertension. While cyanoacrylate injection has become established as the cornerstone of endoscopic treatment for gastric varices, an important clinical question remains: should endoscopists inject glue into every visible varix, or restrict injection to those considered high risk (e.g. with stigmata of recent hemorrhage [SRH])? The randomized controlled trial (RCT) by Prajapati et al. published in this issue of Endoscopy provides the first head-to-head comparison of these two strategies in patients experiencing their first variceal bleed [1].

“A targeted injection strategy that focuses on high-risk and recently bleeding varices seems adequate for secondary prophylaxis following a first bleeding episode.”

In this single-center study, 145 patients with cirrhosis and gastroesophageal varices (GOV1 or GOV2) or isolated gastric varices (IGV1) were randomly assessed to receive either an aggressive endoscopic treatment consisting of cyanoacrylate injection into all gastric varices, in addition to endoscopic variceal ligation (EVL) of esophageal varices, or a conservative endoscopic approach, in which glue was reserved for varices showing SRH or considered high risk based on size and location (IVG1 or GOV2 with varix size of >10 mm and red wale mark), with EVL for esophageal varices and GOV1 with SRH. During follow-up in the aggressive group, endoscopic treatment was repeated monthly until complete obliteration of all varices was confirmed. During follow-up in the conservative group, patients underwent EVL for esophageal varices and GOV1 showing SRH, but glue injection was performed only for GOV2 and IVG1 with SRH. All patients also received nonselective beta-blockers as part of standard medical therapy (if no contraindication).

At 12 months’ follow-up, the rebleeding rate did not differ significantly between the groups. Competing-risk analysis showed a rebleeding incidence of 18.2% in the aggressive group compared with 15.0% in the conservative group, with no statistically significant difference (subdistribution hazard ratio [HR] 1.20, P = 0.66). All-cause mortality or liver transplantation occurred in 22.2% of patients in the aggressive group compared with 32.9% in the conservative group (HR 0.63, P = 0.15), a difference that did not reach statistical significance. Notably, time to obliteration of GOV1 varices was shorter in the aggressive group, with fewer sessions required, but this benefit did not translate into improved clinical outcomes. No cases of clinically significant embolism were observed in either group. To summarize, aggressive therapy accelerated variceal obliteration but failed to reduce rebleeding or death.

International guidelines, such as Baveno VII, acknowledge the role of cyanoacrylate in managing gastric varices but do not offer detailed guidance on the extent of injection required [2]. Similarly, the European Society of Gastrointestinal Endoscopy guideline cautiously recommends endoscopic ultrasound-guided glue or coil injection as an option but does not take a definitive stance on whether all varices should be treated regardless of appearance [3]. Observational studies have suggested that complete obliteration may reduce early rebleeding [4], but randomized data have so far been limited to comparisons between different modalities rather than different intensities of glue therapy [5] [6]. In this regard, the RCT by Prajapati at al. addresses a meaningful and previously unresolved clinical question.

The study’s strengths include its randomized design, which reduces the risk of selection bias, and the use of competing-risk analysis, a statistically robust approach given the high mortality rate of patients with cirrhosis. The trial also reflects routine practice, with most patients presenting with coexisting esophageal varices. The conservative strategy in this study was far from minimalist. It mandated treatment of any gastric varix showing SRH or large size, particularly in the cardiofundal region. This reflects existing data indicating that high-risk stigmata, rather than the mere presence of varices, are the most relevant predictors of early rebleeding [7]. The study suggests that once these high-risk lesions are treated, further “cosmetic” injections may provide further protection, while increasing procedure time and exposure to cyanoacrylate.

However, some limitations deserve consideration. This RCT was underpowered to detect small but potentially meaningful differences in rebleeding rate. Its single-center design may limit applicability, and the primary end point was modified after peer review. There is also a difference between the two groups for GOV1 treatment, which is not explicitly explained in the article. In addition, a major limitation is that none of the patients eligible for transjugular intrahepatic portosystemic shunt (TIPS) received it. These limitations, however, do not undermine the key clinical message: more glue did not translate into better outcomes.

Glue injection treats only one manifestation of portal hypertension. Despite adequate endoscopic management, overall mortality or liver transplantation at 1 year remained 22%–33% in both groups, reinforcing the need for a holistic management plan. This includes optimization of medical therapy, such as nonselective beta blockers, consideration of early TIPS in appropriate patients, and timely evaluation for liver transplantation when indicated.

For routine practice, these results offer a pragmatic message. A targeted injection strategy that focuses on high-risk and recently bleeding varices seems adequate for secondary prophylaxis following a first bleeding episode. This strategy may help reduce procedural burden and cyanoacrylate use without compromising efficacy. An aggressive strategy could still be considered when there is limited confidence in reliably identifying high-risk lesions, but it should probably be seen as the default. Further studies are needed to confirm and expand upon these findings. Multicenter trials with larger sample sizes could help detect modest but clinically significant differences. Future research should also incorporate endoscopic ultrasound-guided coil/glue or thrombin [3] and/or early TIPS [8] to establish their place relative to conventional injection. Future trials should also assess cost-effectiveness and explore patient-reported outcomes such as procedure-related pain and anxiety. Moreover, the potential application of conservative strategies in the setting of primary prophylaxis remains unexplored and could represent an important next step.

In conclusion, the study by Prajapati et al. provides timely and clinically relevant evidence. It challenges the assumption that “more glue is better” and supports a more nuanced, patient-centered approach to managing gastric varices. In endoscopic therapy, as in much of medicine, doing precisely what is necessary – and no more – may be the most efficient strategy.

Publication History

Article published online:
19 August 2025

• References

• 1 Prajapati R, Nandwani S, Mehta R. et al. Aggressive versus conservative endotherapy for gastric varices in cirrhosis: a randomized controlled trial. Endoscopy 2025; 57
• 2 de Franchis R, Bosch J, Garcia-Tsao G. et al. Baveno VII – renewing consensus in portal hypertension. J Hepatol 2022; 76: 959-974
• 3 Gralnek IM, Camus Duboc M, Garcia-Pagan JC. et al. Endoscopic diagnosis and management of esophagogastric variceal hemorrhage: ESGE Guideline. Endoscopy 2022; 54: 1094-1120
• 4 Seewald S, Ang TL, Imazu H. et al. A standardized injection technique ensures success and safety of N-butyl-2-cyanoacrylate for fundal varices. Gastrointest Endosc 2008; 68: 447-454
• 5 Lo GH, Lai KH, Cheng JS. et al. Butyl cyanoacrylate injection versus band ligation in bleeding gastric varices: a randomized trial. Hepatology 2001; 33: 1060-1064
• 6 Mishra SR, Sharma BC, Kumar A. et al. Cyanoacrylate injection versus beta-blocker for secondary prophylaxis of gastric variceal bleed: an RCT. Gut 2010; 59: 729-735
• 7 Hu Y, Zhou M, Liu D. et al. Risk factors for rebleeding after endoscopic injection of cyanoacrylate glue for gastric varices: a systematic review and meta-analysis. Dig Dis Sci 2024; 69: 2890-2899
• 8 Biswas S, Vaishnav M, Gamanagatti S. et al. Endoscopic Glue Injection vs Glue Plus BRTO or TIPSS for Preventing Gastric Variceal Bleeding: A Randomized Controlled Trial. Clin Gastroenterol Hepatol 2025; 23: 954-964.e10