To assess the feasibility, safety, and effectiveness of endoscopic ultrasound–guided shunt occlusion as a novel, minimally invasive therapeutic alternative for managing refractory hepatic encephalopathy in patients with large portosystemic shunts who are ineligible for TIPS, BRTO, or liver transplantation.

We retrospectively analyzed three male patients (ages 41–66) with ethanolrelated cirrhosis and recurrent overt hepatic encephalopathy (OHE) despite optimal medical therapy. All had large splenorenal or splenoportal shunts documented on CT, with significant portofugal flow on doppler. Due to financial constraints or high MELD, liver transplantation, BRTO, and TIPS were not feasible. Under deep sedation, an echoendoscope was positioned at the gastric cardia to visualize the splenoportal axis. The afferent feeder to the shunt was identified using Doppler, and a 19G EUS-FNA needle was used for direct puncture. Technique involved deployment of 16–20 mm coils folowed by n-butyl cyanoacrylate glue injection, in single or staged sessions, depending on residual flow. Coil number ranged from 2 to 6 per patient. Post-procedure, patients were monitored for bleeding, ascites, hypoxia, and worsening decompensation. CT portography was repeated after 4–6 weeks. Clinical folow-up included MELD score, HE episodes, and functional status assessment.

Case 1 (66-year-old male): Large splenoportal shunt with recurrent HE. Single session EUS guided shunt occlusion with two coils (18mm x 14cm and 20mm x 14cm) and 2 ml glue achieved complete obliteration. No complications. No further HE over 3 months; improved cognition reported.

Case 2 (57-year-old male): Large posterior gastric vein afferent shunt. Initial coil (20mm and 16mm) + and 2ml glue session partially occluded flow. A second session next day achieved complete thrombosis. In the second session, 2 coils of 20mm with 2ml glue was injected. Developed mild hemoperitoneum and ascites, managed with correction of coagulopathy. After discharge, only one minor HE episode occurred at 3 weeks, and none thereafter. CT at 35 days showed complete non-opacification of shunt without thrombus extension.

Case 3 (41-year-old male): a 2cm portosystemic shunt was seen. EUS guided shunt coiling was done using 4 coils of 20mm, 20mm, 18mm and 18mm, alongwith 2ml glue. Mild post-procedure hypoxia occurred, managed conservatively. CT showed partial obliteration of the shunt. No episodes of OHE in the followed up 1 month. Improved sensorium and liver function was noted.

Across all cases, technical success was 100%, clinical control of OHE was achieved in all three patients, there was no extension of portal vein or systemic thrombus, and no mortality was observed.

This case series demonstrates that EUS-guided shunt occlusion is a feasible and effective therapeutic alternative for patients with recurrent HE and contraindications to TIPS, BRTO, or transplantation. The ability to directly visualize and selectively embolize the shunt afferent offers distinct procedural precision over radiologic methods. While mild complications occurred, a were manageable, and neurocognitive improvement was consistent across patients. EUS guided shunt occlusion may represent an emerging interventional paradigm in hepatology, particularly within resource-constrained settings. A multicentric prospective study is needed to establish standardized protocols, optimal coil–glue combinations, and long-term outcomes beyond encephalopathy control, including survival and portal hemodynamics.