Bicentric comparison of 3-year clinical outcomes between endoscopic ultrasound-guided embolization with coil and hemostatic adjunct versus balloon-occluded retrograde transvenous obliteration for treatment of bleeding gastric varices: A propensity-score matched study
In brief
EUS-guided coils had 3% adverse events versus 37% with BRTO
In a propensity-matched comparison of 60 patients with cirrhosis and bleeding gastric varices, adverse events occurred in 3% treated with EUS-guided coils and 37% treated with BRTO. Treatment success was similar, while further liver decompensation was less frequent with coils; transfusions, repeat procedures, hospital stay, and mortality did not differ. The small, retrospective study needs confirmation in larger prospective comparisons.
- Journal
- Gastrointestinal endoscopy (Q1)
- Published
- 17 October 2026
- Study design
- Prospective / inception cohort
- Evidence level
- Level 2, Moderate (CEBM 2b)
- Authors
- Noppachai Siranart, Thomas J Wang, Peter Konyn, Ahmad N Bazarbashi, Pichamol Jirapinyo, Christopher C Thompson, et al.
- PMID
- 42892545
- DOI
- 10.1016/j.gie.2026.10.010
Why clinicians should know about it
- Picked for Gastroenterology (paper of the day, 21 October 2026): EUS‑COIL vs BRTO for bleeding gastric varices
Abstract
BACKGROUND AND AIMS: Endoscopic ultrasound-guided coil injection (EUS-COIL) and balloon-occluded retrograde transvenous obliteration (BRTO) performed by interventional radiology (IR) are both effective obliterative procedures to treat gastric varices. Comparative studies are lacking. Our aim was to assess clinical outcomes for EUS-COIL versus BRTO for treatment of acute gastric variceal bleeding. METHODS: This retrospective cohort study included all patients admitted to two tertiary care centers between 2009-2024 who underwent either EUS-COIL or BRTO for bleeding gastric varices. The primary outcomes were technical/clinical success and adverse events. Secondary outcomes included transfusions, reintervention, length of hospital stay, further hepatic decompensation (esophageal variceal hemorrhage, new ascites, or new hepatic encephalopathy (HE)), and all-cause mortality. Subgroup analyses were performed in patients with cirrhosis, including with propensity score matching (PSM) which was conducted 1:1 for age, sex, Child-Pugh classification, MELD-Na score, and hemoglobin level on admission. Outcomes over the 36-months post-intervention were assessed using time-to-event analysis. RESULTS: A total of 94 patients were included (48 EUS-COIL, 46 BRTO), of whom 60 patients with cirrhosis (30 per group) were analyzed after propensity score matching. In the matched cohort, technical success was 100% in both groups and clinical success was 100% versus 93.3% (p=0.49). The EUS-COIL group experienced significantly fewer adverse events (3% vs 37%; p=0.003) and had a longer time-to-reintervention (65 ± 52 days vs 10 ± 9.6 days; p=0.022) over a median follow-up 900 ± 790 days. The incidence of further hepatic decompensation was lower in the EUS-COIL group (10% vs 40%; p=0.007), as was the likelihood of further hepatic decompensation (HR 0.28 (95%CI: 0.08-0.984), p = 0.047) compared to BRTO. No significant differences were observed with regard to post-procedural transfusion requirements, rates of reintervention, length of hospital stay, and mortality in cirrhosis patients. In the unmatched cohort, technical (100% vs 80%; p=0.001) and clinical (100% vs 89%; p=0.03) success were higher with EUS-COIL than BRTO. CONCLUSIONS: In patients with gastric variceal hemorrhage, EUS-COIL demonstrated higher rates of technical/clinical success and lower rates of adverse events (both peri-procedural and on long-term follow-up) compared to BRTO. EUS-COIL had significantly lower rates of further hepatic decompensation. While larger prospective studies are required, these data suggest that EUS-COIL, when available, may be the preferred initial therapy for bleeding gastric varices as part of a step-up strategy.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.