904 EUS-Guided Treatment of Gastric Varices: A Systematic Review and Meta-Analysis With Indirect Comparison to Endoscopic Therapy

Abstract

INTRODUCTION: Gastric variceal (GV) bleeding carries a significant mortality in the setting of portal hypertension. Emergency band ligation and/ or transjugular intrahepatic portosystemic shunt is the current recommended treatment. Other treatment options include direct endoscopic (END) glue injection and endoscopic ultrasound (EUS) guided glue and/ or coil injection. The role of EUS in the treatment of GV is not established and remains an active area of investigation. The aim of this study was to describe the performance of EUS in the treatment of GV. METHODS: We conducted a comprehensive search of several databases (inception to Dec 2018) to identify studies that evaluated END-glue and EUS in the treatment of GV. Our goals were to estimate the pooled rates of treatment efficacy, GV obliteration, GV recurrence, early re-bleeding, late re-bleeding and adverse events with EUS guided therapy in GV. The pooled outcomes of END-glue was used as a comparator group. RESULTS: Total of 35 studies (2777 patients) were included. Mean age range was between 40 and 65 years, with 68% males. N-butyl-2-cyanoacrylate was the most commonly used glue. Use of lipiodol varied among the studies. 28% had GOV1, 48% had GOV2, and 24% had IGV1 (Sarin classification of GV). 63% of the patients had cirrhosis. 30% were due to alcohol and 31% due to viral hepatitis. 11% of the included patients had concomitant hepatocellular carcinoma. GV obliteration with EUS-guided therapy was 84.7% (95% CI 66.8-93.8, I2 = 84.2), and with END-glue was 57.1% (38.2-74.1, I2 = 94.3), P = 0.03. GV recurrence with EUS was 6.1% (95% CI 2.6-13.7, I2 = 51.4, and with END-glue was 22.6% (13.8-34.9, I2 = 77.1), P = 0.006. Late re-bleeding with EUS was 10.6% (95% CI 6.3-17.3, I2 = 45.7), and with END-glue was 20% (14.7-26.6, I2 = 89.30), P = 0.03. The pooled rates of immediate treatment efficacy and early re-bleeding were comparable. On subgroup analysis, EUS-coil/glue combination showed superior outcomes when compared to END-glue (Table 1). The pooled rate of overall adverse events with EUS-glue was 16.5% (9.7-26.9). The rates of adverse event subtypes, and how it compared to END-glue is given in Table 2. CONCLUSION: Based on our meta-analysis, EUS-guided therapies seems to demonstrate benefit in terms of GV obliteration, GV recurrence and late GV re-bleeding. They were significantly superior when indirectly compared to END therapy. The immediate treatment efficacy and early re-bleeding rates were comparable. EUS-coil/glue, in particular, seems to be the best modality.