Subscribe to RSS
Download PDF
DOI: 10.1055/a-2639-1875
Aggressive versus conservative endotherapy for gastric varices in cirrhosis: a randomized controlled trial
Authors
Clinical Trial:
Registration number (trial ID): CTRI/2022/02/040188, Trial registry: Clinical Trials Registry India (http://www.ctri.nic.in/Clinicaltrials), Type of Study: Open labelled, Prospective, Randomized, Single-center study
Abstract
Background
Gastric variceal bleeding in cirrhosis remains a challenging clinical problem with variations in management strategies. This randomized controlled trial compared aggressive versus conservative cyanoacrylate therapy for gastric varices in patients with cirrhosis presenting with their first variceal bleed.
Methods
Patients with cirrhosis and large gastric varices (GOV1, GOV2, IGV1) and a first episode of variceal bleeding were randomized to receive aggressive endotherapy (Group A) or conservative endotherapy (Group B). In Group A, all visible gastric varices were obliterated using cyanoacrylate glue; in Group B, only varices with stigmata of recent hemorrhage or high-risk features were treated. The primary outcome was variceal rebleeding at 1 year. Secondary outcomes included all-cause mortality, variceal obliteration time, and adverse events.
Results
145 patients were analyzed (Group A 72; Group B 73). At 1 year, the cumulative incidence of rebleeding was comparable between groups (18.2% vs. 15.0%). All-cause mortality at 1 year was also similar (22.2% vs. 32.9%), with a hazard ratio of 0.63 (95%CI 0.33–1.18; P = 0.15), suggesting a nonsignificant 37% reduction in mortality risk with aggressive endotherapy. Time to obliteration of GOV1 varices was shorter in Group A, with a median (range) of 4 (4–20) vs. 8 (4–116) weeks in Group B. Similarly, the number of endoscopic sessions required for GOV1 obliteration was fewer in Group A (1 [1–4] vs. 2 [1–5] sessions). Adverse event rates were comparable across both groups.
Conclusion
Aggressive endotherapy resulted in rebleeding and mortality rates similar to those of conservative therapy.
Introduction
Acute variceal bleeding is a serious complication of cirrhosis, associated with significant morbidity and mortality [1]. While management protocols for esophageal varices are well established, the treatment of gastric varices, particularly gastroesophageal varices type 1 (GOV1), remains less clear and is often subject to variation in clinical practice. Gastric varices tend to have a higher propensity for severe bleeding and are treated with different approaches, including endoscopic variceal ligation (EVL) or cyanoacrylate glue injection, depending on the variceal type and risk factors [2] [3]. Previous studies have demonstrated that cyanoacrylate glue is highly effective in managing gastric varices and controlling gastric variceal bleeding, with rebleeding rates ranging from 0% to 58%, depending on the completeness of variceal obliteration [4] [5]. However, there is an ongoing debate about whether glue should be used for all visible gastric varices or only for those with stigmata of recent hemorrhage (SRH). Baveno VII guidelines suggest either EVL or glue for GOV1, but the optimal approach for patients presenting with a first episode of variceal bleeding remains unsettled [6].
The primary aim of this study was to compare aggressive vs. conservative cyanoacrylate therapy for patients with gastric varices in terms of rebleeding at 1 year. By addressing this critical question, our goal is to provide evidence that can inform future guidelines and improve outcomes for cirrhosis patients with variceal bleeding. Recently, the use of endoscopic ultrasound (EUS)-guided coil plus cyanoacrylate injection has been proposed as an alternative treatment for gastric varices. However, the European Society of Gastrointestinal Endoscopy (ESGE) guidelines (2022) [7] provide only a weak recommendation for this approach due to limited supporting evidence. As our study was designed prior to these recommendations, EUS-guided therapy was not incorporated into the protocol.
Methods
Study settings
This study was conducted from February 2022 to September 2023 at the Surat Institute of Digestive Sciences, a tertiary care hospital in Surat, India, specializing in the treatment of liver and gastrointestinal disorders. The hospital is a referral center for patients with cirrhosis and variceal bleeding, providing advanced endoscopic and hepatology services.
The study protocol was approved by the Institutional Review Board of Surat Institute of Digestive Sciences, and all procedures were conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants prior to enrollment. The RCT did not undergo external monitoring.
This study was conducted in accordance with the CONSORT 2010 guidelines [8], and the manuscript has been revised to ensure full compliance. A CONSORT-compliant flow diagram (see Fig. 1s in the online-only Supplementary material) has been included to illustrate patient enrollment, randomization, follow-up, and analysis.
Patients
The study included patients with cirrhosis who presented with their first episode of acute gastroesophageal variceal bleeding. All selected cases had one or more gastroesophageal varices (GOV1 or GOV2) or isolated gastric varices type 1 (IGV1).
Inclusion criteria were age ≥18 years, cirrhosis, and a first episode of acute variceal bleeding. Exclusion criteria were 1) noncirrhotic portal hypertension, 2) IGV2, 3) only esophageal varices, no gastric varices, 4) past history of variceal bleeding, 5) past endotherapy as primary prophylaxis, 6) contraindications to beta blockers and cyanoacrylate injection, 7) pregnancy, 8) hepatocellular carcinoma, 9) portal vein thrombosis, and 10) refusal to participate in the study.
All patients underwent a thorough medical history review, routine laboratory investigations, and abdominopelvic ultrasonography. If hepatocellular carcinoma or portal vein thrombosis was detected during the ultrasound after initial endoscopic management of bleeding, the patient was excluded from the study.
Randomization and treatment
This was an open-label, randomized controlled trial. All patients with cirrhosis and suspected gastrointestinal bleeding were required to sign an informed consent form prior to study inclusion. Patients then underwent endoscopic evaluation, and those found to have nonvariceal bleeding (e.g. peptic ulcer bleeding) or isolated esophageal varices were excluded and managed according to standard clinical guidelines. Only patients meeting the inclusion criteria were randomized. Patients were randomized in a 1:1 ratio to either Group A or Group B at the time of endoscopy if GOV1, GOV2, or IGV1 varices were identified. Randomization was performed after confirming eligibility in order to prevent bias. A computer-generated randomization sequence was used, and allocation was concealed using sequentially numbered, opaque, sealed envelopes, which were opened only after eligibility confirmation; this ensured allocation concealment. No stratification was applied during randomization. The randomization sequence was maintained by the study statistician and was not accessible to investigators prior to intervention assignment.
During endoscopy, the size and nature of the varices and the presence of high-risk features – such as red wale marks, cherry-red spots, white nipples, adherent blood clots, or erosions – were documented.
The endoscopic management procedures for groups A and B during the index endoscopy are outlined in [Table 1]. At index endoscopy, patients in Group A underwent endoscopic cyanoacrylate injection for all visible GOV1, GOV2, and IGV1 varices, along with EVL for esophageal varices. In contrast, patients in Group B received EVL for esophageal varices and GOV1 varices with SRH. For GOV2 and IGV1 varices in Group B, cyanoacrylate injections were performed if SRH was present, or if the varices were deemed high risk, based on size (>10 mm) and red wale marks, even in the absence of SRH. For EVL, bands were applied to each varix in a step-ladder pattern, progressing up to 5 cm above the gastroesophageal junction. In Group B, GOV1 varices were included in the band ligation. All endoscopic procedures were performed by four experienced endoscopists, each with more than 7 years of experience in variceal endotherapy. The endoscopists followed a standardized protocol for cyanoacrylate injection and variceal ligation to ensure uniformity in treatment delivery. All procedures were performed under sedation, ensuring patient comfort and procedural feasibility. Endoscopic variceal treatment was conducted using standard techniques, with sedation protocols tailored according to patient tolerance and institutional guidelines.
For patients in both groups receiving cyanoacrylate glue injections, 0.5 mL aliquots of glue were injected into the varix, followed by 1.5 to 3 mL of distilled water, depending on the size of the varix. The procedure was repeated as needed until the varix was completely obliterated. For larger varices, glue injection started from the far end of the gastroesophageal junction, working toward the area with the maximum bulge. Additional glue injections were used to control any active bleeding.
The end point for hemostasis was the complete obliteration of the varices and the absence of bleeding within the first 48 hours. Any upper gastrointestinal bleeding episode occurring after the initial achievement of hemostasis was classified as a rebleed. Initial hemostasis was defined by stable vital signs and the absence of rebleeding within 48 hours after treatment. Rebleeding was identified either as active bleeding from the treated varices on endoscopy or as the occurrence of hematemesis and/or melena, accompanied by a hemoglobin drop of more than 2 g/dL, necessitating hospitalization. Early rebleeding was defined as rebleeding occurring within 30 days of achieving initial hemostasis, while late rebleeding referred to rebleeding events occurring after 30 days [9].
Transjugular intrahepatic portosystemic shunt (TIPS) was considered as a rescue therapy for patients with persistent or refractory gastric variceal bleeding despite endoscopic therapy.
Follow-up
Follow-up endoscopy was performed monthly until all varices were obliterated in Group A, and until esophageal varices were obliterated in Group B. During follow-up endoscopy, Group A continued to receive cyanoacrylate injections for all visible GOV1, GOV2, and IGV1 varices, with EVL for esophageal varices. Group B patients again underwent EVL for esophageal varices and GOV1 varices with SRH, while cyanoacrylate injections were given for GOV2 and IGV1 varices only if SRH was present. Subsequently, follow-up endoscopies were conducted at 3 months, 6 months, and 1 year. Additional follow-up endoscopies were performed in cases of rebleeding, and management followed the protocol outlined in Table 1.
All patients in both groups were prescribed nonselective beta-blockers (carvedilol or propranolol), unless contraindicated. Carvedilol was initiated at 3.125–6.25 mg twice daily, and propranolol at 20 mg twice daily, with subsequent titration based on heart rate and blood pressure. Compliance with beta-blocker therapy was monitored at each follow-up visit through structured patient interviews and medication logs.
All patients were monitored regularly for decompensating events and rebleeding, through outpatient department visits until the study conclusion, with a minimum follow-up duration of 1 year. Routine blood investigations, including complete blood count, renal function tests, and liver function tests, were performed at baseline, 1 month, 6 months, and 1 year. Abdominal ultrasonography was conducted at baseline, 6 months, and 1 year. Changes in Child–Turcotte–Pugh score, Model for End-Stage Liver Disease (MELD) score, and the development of new complications – such as variceal hemorrhage, hepatic encephalopathy, hepatorenal syndrome, sepsis, ascites, spontaneous bacterial peritonitis, hepatopulmonary syndrome, and hepatocellular carcinoma – were also monitored.
All study outcomes were assessed by two independent investigators (A.K. and M.K.), both of whom were blinded to treatment allocation and not involved in patient care during the study. Outcome data, including rebleeding events, mortality, and secondary end points, were anonymized and independently verified by A.K. and M.K. to ensure unbiased and reliable data analysis.
Study end points
The primary end point was variceal rebleeding at 1 year. Secondary end points included all-cause mortality or liver transplantation at 1 year, time to variceal obliteration, adverse events, and need for salvage therapy.
Patients who experienced rebleeding underwent repeat endotherapy following the study protocol. In cases where rebleeding persisted despite standard treatment, additional interventions such as TIPS, balloon-occluded retrograde transvenous obliteration, or EUS-guided glue plus coil embolization were available at the treating physician’s discretion. All-cause mortality was initially considered a co-primary end point but was later designated as a secondary end point because the sample size calculation was based on variceal rebleeding at 1 year. This change was made after receiving feedback from the peer review process to ensure alignment with the study’s primary objective and sample size assumptions.
Sample size calculation
In Group B, it was anticipated that 23% of patients would experience variceal rebleeding at the 1-year follow-up. Based on our own preliminary data and findings from other studies, we expected the rebleeding rate in Group A to be 6% [10] [11]. Using an alpha error of 0.05 (one-sided) and a study power of 80%, the required sample size was calculated to be 66 patients per group, amounting to a total of 132 patients. To account for an estimated 10% loss to follow-up, the final target sample size was set at 145 patients.
Statistical analysis
Statistical analyses were conducted using SPSS for Windows, version 26.0 (IBM Corp., Armonk, New York, USA) and R software (version 4.4.2; R Foundation for Statistical Computing, Vienna, Austria), with RStudio as the integrated development environment. All hypothesis tests were two sided, with statistical significance set at α = 0.05.
Continuous variables were summarized as medians with ranges and compared between groups using the Mann–Whitney U test. Categorical variables were expressed as frequencies and percentages, with comparisons performed using the chi-squared test or Fisher’s exact test, as appropriate.
The study was analyzed using both intention-to-treat and per-protocol principles. The intention-to-treat analysis included all randomized patients. Patients lost to follow-up were included in the intention-to-treat analysis until their last known follow-up date and were censored for time-to-event analyses. The per-protocol analysis included all patients who completed follow-up and adhered to the assigned treatment protocol. Patients who underwent liver transplantation, developed hepatocellular carcinoma, or were lost to follow-up were excluded from the per-protocol analysis.
The primary end point, variceal rebleeding at 1 year, was analyzed using the Fine–Gray competing risk regression model (cmprsk package in R). This method accounted for all-cause mortality and liver transplantation as competing events. Cumulative incidence functions were plotted for each treatment group to visualize the probability of rebleeding over time in the presence of competing risks. The subdistribution hazard ratio with 95%CIs was calculated to estimate the relative risk of rebleeding between groups.
The secondary end point of all-cause mortality was analyzed using Kaplan–Meier survival analysis (survival package in R), with group differences assessed using the log-rank test. Additionally, a Cox proportional hazards model was applied to estimate hazard ratios with 95%CIs. The proportional hazards assumption was tested using Schoenfeld residuals and was found to be satisfied (global P = 0.54). The model concordance index was 0.569, indicating moderate discriminatory ability.
Graphical plots were generated in R using the survminer, ggplot2, and cmprsk packages. Cumulative incidence function plots were used to illustrate competing risk events, while Kaplan–Meier survival curves with 95%CIs were used for mortality analysis.
All baseline characteristics and categorical outcomes were analyzed using complete case analysis. No missing baseline data were observed. Missing continuous outcome data, if any, were not imputed. Adjustments for multiplicity were not performed for secondary outcomes, which were considered exploratory; instead, 95%CIs were reported to provide estimates of effect size.
Results
Patients
Between February 2022 and September 2023, 242 consecutive cirrhosis patients presenting with their first episode of variceal bleeding were assessed for eligibility. A total of 97 patients were excluded for the following reasons: absence of gastric varices (only esophageal varices present; n = 86), presence of IGV2 varices (n = 0), contraindications to beta blockers (n = 1), hepatocellular carcinoma (n = 4), portal vein thrombosis (n = 2), restricted mouth opening preventing EVL (n = 1), and refusal to participate (n = 3). Consequently, 145 patients were enrolled in the trial, with 72 patients randomized to Group A and 73 patients to Group B.
During follow-up, five patients (three from Group A and two from Group B) underwent liver transplantation after acquiring the necessary resources. Additionally, six patients (three from each group) developed hepatocellular carcinoma. Eight patients (three from Group A and five from Group B) were lost to follow-up. The patient flow through the trial is depicted in the CONSORT diagram (Fig. 1s).
Baseline characteristics
The median age of the included patients was 51 years (range 24–81), and 123 (84.8%) were male. The most common cause of cirrhosis was alcohol related (49.0%), followed by nonalcoholic fatty liver disease (38.6%), hepatitis B virus infection (5.5%), autoimmune causes (3.4%), and hepatitis C virus infection (3.4%). The median MELD score was 9 (range 6–23), and the median Child–Turcotte–Pugh score was 7 (range 5–14). The majority of patients were in Child–Pugh class A (44.8%), followed by class B (37.2%) and class C (17.9%). Baseline characteristics were similar between the two groups ([Table 2]).
Regarding variceal characteristics, GOV1 varices were the most common, present in 88.9% of patients in Group A and 93.2% in Group B. GOV2 varices were found in 15.3% of Group A and 12.3% of Group B, while IGV1 varices were seen in 11.1% and 9.6%, respectively. Most patients also had esophageal varices, with esophageal varices present in 97.2% of Group A and 100% of Group B. Portal hypertensive gastropathy was noted in all patients in Group A and Group B (mild in most cases). There were no significant differences in baseline variceal characteristics between the groups ([Table 3]).
Treatment
Patients in Group A and Group B were treated as mentioned in [Table 1]. Four patients had active bleeding on endoscopy, with two in each group (P > 0.99). Hemostasis was successfully achieved during the index endoscopy for all patients in both Group A and Group B. The median amount of glue used during the index endoscopy was 1 mL (range 0.5–4 mL) in Group A and 1 mL (range 1–2 mL) in Group B.
Early TIPS (within 72 hours) was not required in any patient, as all cases were initially managed successfully with endoscopic intervention.
Follow-up and study end points
The median follow-up duration was 19 months (range 0–32 months). In the subgroup analysis, the median follow-up was 20.5 months (range 0–32 months) for Group A and 18 months (range 0–32 months) for Group B.
Rebleeding
By 1 year of follow-up, the cumulative incidence of rebleeding was 18.2% in Group A and 15.0% in Group B, as determined by the Fine–Gray competing risk analysis ([Table 4]). Two patients with refractory rebleeding were advised to undergo elective TIPS placement, but neither consented to the procedure. All patients received beta blockers (65 carvedilol and 80 propranolol), none of which were discontinued during follow-up. Only one patient experienced nonvariceal bleeding during the study period.
To account for mortality and liver transplantation as competing risks, a Fine–Gray competing risk regression analysis was performed ([Fig. 1]). The subdistribution hazard ratio for rebleeding in the aggressive treatment group was 1.20 (95%CI 0.52 to 2.77; P = 0.66), indicating a nonsignificant 20% increased risk of rebleeding in the aggressive group after accounting for competing events. The cumulative incidence function curves showed similar rebleeding probabilities over time between both groups, consistent with the competing risk analysis.
Baseline bilirubin, international normalized ratio, creatinine, albumin, ascites, Child–Turcotte–Pugh score, and MELD score did not differ significantly between patients who experienced rebleeding and those who did not at 1-year follow-up.
All-cause mortality or liver transplantation
A total of 46 patients (31.7%) died or underwent liver transplantation during the study. Rates of mortality or liver transplantation were 11.7% (n = 17) at 1 month, 22.1% (n = 32) at 6 months, 27.6% (n = 40) at 1 year, and 31.7% (n = 46) by study completion ([Table 4]). At 1 year, in Group A, 10 patients had died (including 4 due to variceal bleeding, 1 of whom had hepatocellular carcinoma), 3 had undergone liver transplantation, 3 had developed hepatocellular carcinoma (1 of whom died from bleeding), and 3 were lost to follow-up. In Group B, 17 patients had died, 2 had undergone liver transplantation, 3 had developed hepatocellular carcinoma, and 5 were lost to follow-up.
Mortality or liver transplantation at 1 year was 22.2% (16/72) in Group A and 32.9% (24/73) in Group B, resulting in 1-year survival rates of 77.8% (56/72) in Group A and 67.1% (49/73) in Group B. While there was a numerical trend favoring Group A, the difference was not statistically significant (Fisher’s exact test, P = 0.12). Transplant-free 1-year survival rates were 81.2% (56/69) in Group A and 69.0% (49/71) in Group B, again without a statistically significant difference (Fisher’s exact test, P = 0.12).
All-cause mortality was assessed using Kaplan–Meier survival analysis and Cox proportional hazards regression. The Kaplan–Meier curves showed a numerical trend toward better survival in Group A compared with Group B over 1 year, although this difference was not statistically significant (log-rank P = 0.15) ([Fig. 2]). The hazard ratio for mortality in the aggressive treatment group was 0.63 (95%CI 0.33 to 1.18; P = 0.15), suggesting a 37% relative reduction in mortality risk, although not statistically significant. The Cox model’s proportional hazards assumption was validated (global P = 0.54), and the model’s concordance index was 0.561, indicating modest discriminatory power.
Of the 46 patients, 9 (19.5%) died due to variceal rebleeding (4 in Group A and 5 in Group B). The remaining deaths were attributable to nonbleeding causes, including hepatic decompensation, infections, and myocardial infarction.
Across all time points (1 month, 6 months, 1 year, and study end), there was no statistically significant difference in all-cause mortality between the two groups. However, a numerical trend toward improved survival in the aggressive therapy group (Group A) was observed, though this did not reach statistical significance.
Decompensation events
At the 1-year follow-up, eight patients (11.1%) in Group A and seven patients (9.6%) in Group B experienced decompensation events unrelated to rebleeding, with a risk difference of 1.8% (95%CI –6.8% to 10.5%). New-onset or worsening ascites occurred in three patients in Group A and six patients in Group B (risk difference: –4.4%, 95%CI –13.0% to 4.1%). Spontaneous bacterial peritonitis was observed in one patient from Group A and three patients from Group B (risk difference: –2.9%, 95%CI –8.7% to 2.9%). Hepatic encephalopathy developed in five patients from Group A and two patient from Group B (risk difference: 5.8%, 95%CI –0.7% to 12.3%).
Decompensation events, including new-onset or worsening ascites, spontaneous bacterial peritonitis , and hepatic encephalopathy at 1 month, 6 months, and study end, are summarized in [Table 4].
Variceal obliteration
A total of 66 patients (91.6%) in Group A and all 73 patients (100%) in Group B underwent EVL for high-risk varices or active bleeding. By the 1-year follow-up, esophageal variceal obliteration was achieved in 60 patients in Group A and 54 patients in Group B (risk difference: 11.8%, 95%CI 0.8% to 22.7%).
The median time to achieve esophageal variceal obliteration was 8 weeks (range 4–24) in Group A vs. 12 weeks (range 8–48) in Group B. The total number of sessions required was 2 (range 1–5) in Group A and 3 (range 1–5) in Group B.
For GOV1 varices, obliteration was achieved in 62 patients in Group A and 50 patients in Group B (risk difference 22.1%, 95%CI 8.6% to 35.6%). The median time to achieve GOV1 obliteration was 4 weeks (range 4–20) in Group A and 8 weeks (range 4–116) in Group B. The number of sessions required for GOV1 obliteration was also lower in Group A (1 [range 1–4] sessions) compared with Group B (2 [range 1–5] sessions).
Adverse events
No major adverse events, such as pulmonary embolism, perforation, or stricture, were observed in either study group. Routine computed tomography scans to detect nonclinical embolization were not performed due to cost constraints. However, 22 patients voluntarily underwent computed tomography scans, and none showed evidence of glue embolism.
Discussion
In this randomized controlled trial, we compared the efficacy and safety of aggressive vs. conservative cyanoacrylate therapy for gastric varices in patients with cirrhosis presenting with their first episode of variceal bleeding. A total of 145 patients were enrolled and included in the final analysis. Our study found no significant differences between the groups in terms of rebleeding rates at the 1-year follow-up. However, aggressive therapy resulted in faster obliteration of GOV1 varices, requiring fewer treatment sessions. Importantly, no major adverse events were observed in either group, indicating that both approaches were safe. These findings suggest that while aggressive cyanoacrylate therapy may expedite variceal obliteration, a more conservative approach may be sufficient for preventing rebleeding and reducing mortality in this patient population. The study adhered to the intention-to-treat principle, as recommended in randomized trials, and incorporated competing risk methods to enhance the robustness of the outcome analysis.
By including a sizeable cohort of patients with cirrhosis with first-time variceal bleeding, this study provides valuable insights into the real-world applicability of both treatment strategies. The comprehensive follow-up period of 1 year allows for the assessment of both short-term and long-term outcomes, including mortality, rebleeding rates, and variceal obliteration success. Furthermore, the focus on clinically relevant end points – such as rebleeding, mortality, and decompensation events – ensures that the findings are directly translatable to improving patient care. Finally, the absence of major adverse events strengthens the conclusion that both treatment approaches are safe in this high-risk population.
Our study builds on existing evidence by comparing aggressive vs. conservative cyanoacrylate therapy for gastric varices, with a particular focus on patients with cirrhosis presenting with their first variceal bleed. The results demonstrated no significant differences in 1-year rebleeding rates or all-cause mortality between the two treatment approaches, although aggressive therapy facilitated faster obliteration of GOV1 varices. Previous studies have reported similar outcomes, with rebleeding rates after gastric variceal obliteration ranging from 0% to 58%, highlighting the variability in rebleeding across studies, which may be influenced by the completeness of variceal obliteration [4] [5] [12] [13]. For instance, Seewald et al. demonstrated that complete obliteration of fundal varices significantly reduced early rebleeding, with a cumulative rebleeding-free rate of 94.5% at 1 year [9]. Our findings are consistent with this, as patients in both groups experienced low early rebleeding rates, likely due to the thorough obliteration of high-risk varices in both the aggressive and conservative arms.
Several studies have excluded patients with large coexisting esophageal varices from their analysis, despite the fact that in clinical practice, gastric and esophageal varices frequently coexist. For example, Garg et al. reported that 88% of patients with gastric varices also had esophageal varices [14]. In our study, nearly all patients (95.5%) had large esophageal varices that required EVL. This coexistence makes it difficult to ascertain whether the source of bleeding is esophageal or gastric, thus supporting our decision to calculate overall rebleeding rates rather than differentiating between esophageal and gastric variceal bleeding. This approach more closely reflects real-world clinical practice.
Successful obliteration of varices is key to reducing rebleeding. Early rebleeding is seen in 0–20.5% patients and is more common in varices that are not completely obturated [9] [15]. In our study, while the overall obliteration rates were lower in Group B, the results were still acceptable. Previous research has emphasized the importance of achieving complete obliteration, as incomplete treatment can lead to higher rebleeding rates [9] [11] [16] [17] [18] [19]. Notably, the conservative approach in Group B, although focused only on high-risk varices, achieved satisfactory results due to complete obliteration of targeted varices, including those in the cardiofundal region. This may explain why the rebleeding rates in our study were comparable between the two groups.
Thromboembolic complications are a known risk of cyanoacrylate glue injection, occurring at rates of 0.5% to 4.3% depending on the technique and the volume of glue used. Our study did not encounter any major adverse events, such as pulmonary embolism or distal embolization, likely due to careful attention to technique and the use of smaller glue volumes (mean 1.32 [SD 0.96] mL in Group A). This aligns with previous findings that smaller glue volumes and undiluted glue significantly reduce the risk of embolization [20] [21]. Our study protocol included standardized retreatment with repeat endotherapy for rebleeding cases. In refractory cases, escalation to TIPS, balloon-occluded retrograde transvenous obliteration, or EUS-guided glue plus coil embolization was performed based on physician discretion. The availability of multiple therapeutic options may have contributed to the relatively low mortality observed in rebleeding cases.
Finally, it is important to note that adherence to beta-blocker therapy and regular follow-up may have contributed to the relatively low rebleeding rates observed in both treatment groups. As the management of varices continues to evolve, our findings suggest that a tailored approach to obliteration, focusing on high-risk varices, may offer an effective alternative to aggressive treatment without compromising patient outcomes.
This study has several limitations that must be considered when interpreting the results. First, the primary outcome was modified during the revision process in response to reviewer feedback to ensure clarity and alignment with the outcome for which the study was originally powered. Another important limitation is the sample size. The initial sample size calculation underestimated the number of patients required to detect a clinically meaningful difference in rebleeding rates. The study was originally powered using a one-sided hypothesis test, but a two-sided hypothesis test (α = 0.05) should have been applied. A total of 91 patients per group (182 in total) would have been required, but only 145 patients were recruited, making the study underpowered to detect smaller but clinically relevant differences. Another limitation is the lack of stratification during randomization, which may have led to imbalances in baseline characteristics between the groups. Although competing-risk regression was used to account for mortality in the rebleeding analysis, other potential confounders, such as beta-blocker adherence and portal hemodynamics, were not adjusted for and may have influenced the results. Additionally, the study was conducted at a single center, which may limit the generalizability of the findings to broader patient populations with different treatment protocols. The follow-up period was limited to 1 year, meaning longer-term outcomes, such as recurrence of gastric varices, quality of life, and potential liver function changes, were not assessed. Regarding treatment approaches, EUS-guided coil therapy was not included in the protocol, as the ESGE guidelines recommending this approach were published after the study commenced [7]. However, the evidence supporting this technique remains limited, and it is not yet widely adopted as standard practice. Future studies incorporating EUS-guided therapies may help determine their comparative efficacy against cyanoacrylate-based approaches. Finally, multiple secondary end points were analyzed without formal multiplicity corrections. While confidence intervals were reported to provide effect size estimates, the potential for type I error must be considered when interpreting these findings. Future larger, multicenter studies with an adequately powered design and appropriate statistical adjustments are needed to confirm these results and refine treatment strategies for gastric varices.
Conclusion
This randomized controlled trial comparing aggressive and conservative endotherapy for gastric varices in patients with cirrhosis found no significant differences in overall mortality, variceal rebleeding rates, or decompensation events between the two treatment approaches. Our findings suggest that aggressive endotherapy is as effective as conservative therapy, with comparable rebleeding and mortality rates. While faster variceal obliteration was observed in the aggressive endotherapy group, this was a secondary outcome that was not adjusted for multiple comparisons. Therefore, these results should be interpreted with caution. Larger, multicenter studies with appropriate statistical adjustments are needed to confirm these findings and further refine treatment strategies for gastric varices.
Conflict of Interest
The authors declare that they have no conflict of interest.
-
References
- 1 Sarin SK, Kumar A, Angus PW. et al. Diagnosis and management of acute variceal bleeding: Asian Pacific Association for Study of the Liver recommendations. Hepatol Int 2011; 5: 607-624
- 2 Sarin SK, Lahoti D, Saxena SP. et al. Prevalence, classification and natural history of gastric varices: a long-term follow-up study in 568 portal hypertension patients. Hepatology 1992; 16: 1343-1349
- 3 Sarin SK, Kumar A. Gastric varices: profile, classification, and management. Am J Gastroenterol 1989; 84: 1244-1249
- 4 Wang TJ, Ryou M. Treatment of gastric varices. Curr Hepatol Rep 2024; 23: 287-293
- 5 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: 2890-2903
- 6 de Franchis R, Bosch J, Garcia-Tsao G. et al. Baveno VII Faculty. Baveno VII – renewing consensus in portal hypertension. J Hepatol 2022; 76: 959-974
- 7 Gralnek IM, Camus Duboc M, Garcia-Pagan JC. et al. Endoscopic diagnosis and management of esophagogastric variceal hemorrhage: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy 2022; 54: 1094-1120
- 8 Moher D, Hopewell S, Schulz KF. et al. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. BMJ 2010; 340: c869
- 9 Seewald S, Ang TL, Imazu H. et al. A standardized injection technique and regimen ensures success and safety of N-butyl-2-cyanoacrylate injection for the treatment of gastric fundal varices (with videos). Gastrointest Endosc 2008; 68: 447-454
- 10 Bick BL, Al-Haddad M, Liangpunsakul S. et al. EUS-guided fine needle injection is superior to direct endoscopic injection of 2-octyl cyanoacrylate for the treatment of gastric variceal bleeding. Surg Endosc 2019; 33: 1837-1845
- 11 Lo GH, Lai KH, Cheng JS. et al. A prospective, randomized trial of butyl cyanoacrylate injection versus band ligation in the management of bleeding gastric varices. Hepatology 2001; 33: 1060-1064
- 12 Mishra SR, Chander Sharma B, Kumar A. et al. Endoscopic cyanoacrylate injection versus beta-blocker for secondary prophylaxis of gastric variceal bleed: a randomised controlled trial. Gut 2010; 59: 729-735
- 13 Mishra SR, Sharma BC, Kumar A. et al. Primary prophylaxis of gastric variceal bleeding comparing cyanoacrylate injection and beta-blockers: a randomized controlled trial. J Hepatol 2011; 54: 1161-1167
- 14 Garg M, Gupta T, Goyal S. Cyanoacrylate glue for gastroesophageal varices: a single centre experience from North India. Arq Gastroenterol 2022; 59: 434-438
- 15 Marques P, Maluf-Filho F, Kumar A. et al. Long-term outcomes of acute gastric variceal bleeding in 48 patients following treatment with cyanoacrylate. Dig Dis Sci 2008; 53: 544-550
- 16 Lee YT, Chan FK, Ng EK. et al. EUS-guided injection of cyanoacrylate for bleeding gastric varices. Gastrointest Endosc 2000; 52: 168-174
- 17 D’Imperio N, Piemontese A, Baroncini D. et al. Evaluation of undiluted N-butyl-2-cyanoacrylate in the endoscopic treatment of upper gastrointestinal tract varices. Endoscopy 1996; 28: 239-243
- 18 Kind R, Guglielmi A, Rodella L. et al. Bucrylate treatment of bleeding gastric varices: 12 years’ experience. Endoscopy 2000; 32: 512-519
- 19 Huang YH, Yeh HZ, Chen GH. et al. Endoscopic treatment of bleeding gastric varices by N-butyl-2-cyanoacrylate (Histoacryl) injection: long-term efficacy and safety. Gastrointest Endosc 2000; 52: 160-167
- 20 Kumar A, Singh S, Madan K. et al. Undiluted N-butyl cyanoacrylate is safe and effective for gastric variceal bleeding. Gastrointest Endosc 2010; 72: 721-727
- 21 Hwang SS, Kim HH, Park SH. et al. N-butyl-2-cyanoacrylate pulmonary embolism after endoscopic injection sclerotherapy for gastric variceal bleeding. J Comput Assist Tomogr 2001; 25: 16-22
Correspondence
Publication History
Received: 20 December 2024
Accepted after revision: 17 June 2025
Accepted Manuscript online:
17 June 2025
Article published online:
19 August 2025
© 2025. Thieme. All rights reserved.
Georg Thieme Verlag KG
Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany
-
References
- 1 Sarin SK, Kumar A, Angus PW. et al. Diagnosis and management of acute variceal bleeding: Asian Pacific Association for Study of the Liver recommendations. Hepatol Int 2011; 5: 607-624
- 2 Sarin SK, Lahoti D, Saxena SP. et al. Prevalence, classification and natural history of gastric varices: a long-term follow-up study in 568 portal hypertension patients. Hepatology 1992; 16: 1343-1349
- 3 Sarin SK, Kumar A. Gastric varices: profile, classification, and management. Am J Gastroenterol 1989; 84: 1244-1249
- 4 Wang TJ, Ryou M. Treatment of gastric varices. Curr Hepatol Rep 2024; 23: 287-293
- 5 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: 2890-2903
- 6 de Franchis R, Bosch J, Garcia-Tsao G. et al. Baveno VII Faculty. Baveno VII – renewing consensus in portal hypertension. J Hepatol 2022; 76: 959-974
- 7 Gralnek IM, Camus Duboc M, Garcia-Pagan JC. et al. Endoscopic diagnosis and management of esophagogastric variceal hemorrhage: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy 2022; 54: 1094-1120
- 8 Moher D, Hopewell S, Schulz KF. et al. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. BMJ 2010; 340: c869
- 9 Seewald S, Ang TL, Imazu H. et al. A standardized injection technique and regimen ensures success and safety of N-butyl-2-cyanoacrylate injection for the treatment of gastric fundal varices (with videos). Gastrointest Endosc 2008; 68: 447-454
- 10 Bick BL, Al-Haddad M, Liangpunsakul S. et al. EUS-guided fine needle injection is superior to direct endoscopic injection of 2-octyl cyanoacrylate for the treatment of gastric variceal bleeding. Surg Endosc 2019; 33: 1837-1845
- 11 Lo GH, Lai KH, Cheng JS. et al. A prospective, randomized trial of butyl cyanoacrylate injection versus band ligation in the management of bleeding gastric varices. Hepatology 2001; 33: 1060-1064
- 12 Mishra SR, Chander Sharma B, Kumar A. et al. Endoscopic cyanoacrylate injection versus beta-blocker for secondary prophylaxis of gastric variceal bleed: a randomised controlled trial. Gut 2010; 59: 729-735
- 13 Mishra SR, Sharma BC, Kumar A. et al. Primary prophylaxis of gastric variceal bleeding comparing cyanoacrylate injection and beta-blockers: a randomized controlled trial. J Hepatol 2011; 54: 1161-1167
- 14 Garg M, Gupta T, Goyal S. Cyanoacrylate glue for gastroesophageal varices: a single centre experience from North India. Arq Gastroenterol 2022; 59: 434-438
- 15 Marques P, Maluf-Filho F, Kumar A. et al. Long-term outcomes of acute gastric variceal bleeding in 48 patients following treatment with cyanoacrylate. Dig Dis Sci 2008; 53: 544-550
- 16 Lee YT, Chan FK, Ng EK. et al. EUS-guided injection of cyanoacrylate for bleeding gastric varices. Gastrointest Endosc 2000; 52: 168-174
- 17 D’Imperio N, Piemontese A, Baroncini D. et al. Evaluation of undiluted N-butyl-2-cyanoacrylate in the endoscopic treatment of upper gastrointestinal tract varices. Endoscopy 1996; 28: 239-243
- 18 Kind R, Guglielmi A, Rodella L. et al. Bucrylate treatment of bleeding gastric varices: 12 years’ experience. Endoscopy 2000; 32: 512-519
- 19 Huang YH, Yeh HZ, Chen GH. et al. Endoscopic treatment of bleeding gastric varices by N-butyl-2-cyanoacrylate (Histoacryl) injection: long-term efficacy and safety. Gastrointest Endosc 2000; 52: 160-167
- 20 Kumar A, Singh S, Madan K. et al. Undiluted N-butyl cyanoacrylate is safe and effective for gastric variceal bleeding. Gastrointest Endosc 2010; 72: 721-727
- 21 Hwang SS, Kim HH, Park SH. et al. N-butyl-2-cyanoacrylate pulmonary embolism after endoscopic injection sclerotherapy for gastric variceal bleeding. J Comput Assist Tomogr 2001; 25: 16-22