870 The Efficacy and Safety of Endoscopic Submucosal Dissection for Early Gastric Cancer in Patients With Cirrhosis: A Systematic Review and Meta-Analysis

Abstract

INTRODUCTION: Cirrhosis is known to increase the risk of bleeding and mortality from procedures. Endoscopic submucosal dissection (ESD) is an approved minimally invasive treatment of early gastric neoplasm. The safety and complications of ESD in cirrhotics in this setting has not been mentioned in the guidelines. We aimed to address it in this meta-analysis. METHODS: We searched the databases of MEDLINE, EMBASE, and CENTRAL from inception to May 2019. A manual review of included studies and references were performed. The inclusion criteria were published studies comparing the efficacy and safety of ESD for gastric cancer in cirrhosis and non-cirrhotics. The outcomes were En bloc resection, R0 resection, post-ESD bleeding, perforation, and mortality rate. Data from each study were combined using the random-effects, generic inverse variance method of DerSimonian and Laird to calculate pooled risk ratio (RR) and 95% confidence intervals (CI). RESULTS: Of 14 included observational studies, four included compensated cirrhosis patients and ten did not mention cirrhosis status. Major etiologies of cirrhosis were hepatitis B infection (49%). Mean platelet counts in compensated cirrhotics was 113000/uL. Cirrhotics have higher risk of post-ESD bleeding compared to others with pooled RR (95% CI) of 1.43 (1.02-2.00, I 2 = 0%) (Figure 1). In subgroup analysis of compensated cirrhosis, the bleeding risk is not different between the groups, pooled RR (95% CI) of 1.22 (0.70-2.12, I 2 = 0%) (Figure 2). The data of decompensated cirrhosis is not available. On the contrary, perforation risk is comparable between the groups with pooled RR (95% CI) of 2.08 (0.51-8.43, I 2 = 0%). ESD procedure time is not different between two groups. We found that the efficacies, En bloc and R0 resection, are not different between cirrhosis and non-cirrhosis with pooled RR (95% CI) of 1.01 (0.97-1.04, I 2 = 12.1%) and 0.98 (0.92-1.04, I 2 = 0%), respectively. The procedure related mortality rate in both cirrhosis and non-cirrhosis patients are 0%. CONCLUSION: Post-ESD bleeding in cirrhosis is a concerning complication as expected in early gastric cancer; but the risk of bleeding in compensated cirrhosis is similar to non-cirrhotics. Perforation and mortality are not different compared to controls. In terms of efficacy, no differences are found between patients with and without cirrhosis. Further studies on subgroup analysis of decompensated cirrhosis are warranted.