Endoscopic Glue Injection vs Glue Plus BRTO or TIPSS for Preventing Gastric Variceal Bleeding: A Randomized Controlled Trial

Abstract

Background and aimsTo compare the efficacy of endoscopic injection sclerotherapy with N-butyl cyanoacrylate glue (EIS-CYA) vs EIS-CYA plus a radiologic intervention (either transjugular intrahepatic portosystemic shunt (TIPSS) or balloon-occluded retrograde transvenous obliteration (BRTO)) for secondary prophylaxis in patients with liver cirrhosis who presented with acute variceal bleeding (AVB) from cardiofundal varices. Primary outcome measure was gastric varix (GV) rebleed rates at 1 year. MethodsConsecutive cirrhosis patients with AVB from cardiofundal varices were randomized into two arms (45 in each) after primary hemostasis by EIS-CYA. In the ‘endoscopic intervention’ (EI) arm, EIS-CYA was repeated at regular intervals (1, 3, 6 and 12 months), while in the ‘radiological intervention’ (RI) arm, patients underwent TIPSS or BRTO followed by endoscopic surveillance. ResultsGV rebleed rates at 1 year were higher in the EI arm compared to the RI arm: 11 (24·4%; 95% CI: 12·9%-39·5%) versus 1 (2·2%; 95% CI: 0·1%-11·8%); (p=0·004) [ARD: 22.2% (95% CI: 8.4%-36.6%)]. GV rebleed related mortality in the EI arm [8 (17·8%; 95% CI: 8·0%-32·1%)] was significantly higher than in the RI arm [1 (2·2%; 0·1%-11·8%)] (p=0.030) [ARD: 15.6 (95% CI: 2.9%-29.2%)], however, there was no difference in all-cause mortality between the two groups (12 [26·7%; 95% CI: 14·6 to 41·9] versus 7 [15·6%; 95% CI: 6·5 to 29·5]). Numbers needed to treat (NNT) to prevent one GV-related rebleed at 1 year was 4.5. ConclusionRadiological intervention for secondary prophylaxis reduces rebleeding from gastric varices and GV rebleeding related mortality in patients with gastric variceal hemorrhage. (CTRI/2021/02/031396)