Abstract
BACKGROUND AND OBJECTIVES: Gastric varices (GVs) are associated with a higher risk of uncontrolled bleeding and death when compared with esophageal varices. While endoscopic glue injection therapy has been traditionally used for secondary prophylaxis in GV, data regarding primary prophylaxis continue to emerge. Recently, EUS-guided therapies have been used in GV bleeding. METHODS: We conducted a comprehensive search of several major databases from inception to June 2022. Our primary goals were to estimate the pooled rates of treatment efficacy, GV obliteration, GV recurrence, and rebleeding with EUS-guided therapy in primary and secondary prophylaxis. Overall adverse events and technical failures were assessed. Random-effects model was used for our meta-analysis, and heterogeneity was assessed using the I(2) % statistics. RESULTS: Eighteen studies with 604 patients were included. In primary prophylaxis, pooled rate of GV obliteration was 90.2% (confidence interval [CI], 81.1-95.2; I(2) = 0). With combination EUS-glue and coil therapy, the rate was 95.4% (CI, 86.7%-98.5%; I(2) = 0). Pooled rate of posttherapy GV bleeding was 4.9% (CI, 1.8%-12.4%; I(2) = 0). In secondary prophylaxis, pooled rate of treatment efficacy was 91.9% (CI, 86.8%-95.2%; I(2) = 12). With EUS-glue, EUS-coil, and combination EUS-glue and coil, the rates were 94.3% (CI, 88.9%-97.1%; I(2) = 0), 95.5% (CI, 80.3%-99.1%; I(2) = 0), and 88.7% (CI, 76%-95.1%; I(2) = 14), respectively. Pooled rate of GV obliteration was 83.6% (CI, 71.5%-91.2%; I(2) = 74). With EUS-glue, EUS-coil, and combination EUS-glue and coil, the rates were 84.6% (CI, 75.9%-90.6%; I(2) = 31), 92.3% (CI, 81.1%-97.1%; I(2) = 0), and 84.5% (CI, 50.8%-96.7%; I(2) = 75), respectively. Pooled rates of GV rebleeding and recurrence were 18.1% (CI, 13.1%-24.3%; I(2) = 16) and 20.6% (CI, 9.3%-39.5%; I(2) = 66), respectively. CONCLUSION: Our analysis shows that EUS-guided therapy for GVs is technically feasible and clinically successful in both primary and secondary prophylaxis of GV.