SAT-087-Screening endoscopy in patients with advanced chronic liver disease beyond portal hypertension: Higher prevalence of upper gastrointestinal neoplasia in comparison to a healthy screening population

BACKGROUND AND AIMS: The Baveno VI criteria were proposed to help identify patients with compensated advanced chronic liver disease (cACLD) who could safely avoid screening esophagogastroduodenoscopy (EGD) for clinically significant varices. Although several studies have validated those criteria, all failed to acknowledge the possible role of the avoided EGD for the same time screening of upper GI neoplasia. We aimed to evaluate the presence, type and pathological characteristics of upper GI neoplasia in patients with cACLD who underwent screening EGD in a country with intermediate gastric cancer risk. METHOD: This retrospective case-control study enrolled all asymptomatic cACLD patients who underwent EGD for varices screening from January 2008 to June 2018. Cases were matched with asymptomatic healthy individuals who underwent EGD for gastric cancer screening at the same time as colonoscopy performed for colorectal cancer screening. RESULTS: We included 1974 subjects (610 patients, 1364 controls), 46.3% male, with a median age of 58 (50-66) years. Besides a male predominance in cases (69.0% vs. 45.6%, p < 0.001), no other demographic characteristic differed between groups. The leading aetiology of ACLD was alcoholic liver disease (53.3%), followed by chronic hepatitis C (16.2%), and non-alcoholic fatty liver disease (7.9%). Of the 610 patients with cACLD, 1 (0.2%) had squamous-cell esophageal carcinoma, 13 (2.1%) gastric neoplasia [gastric cancer, n = 10; high-grade dysplasia (HGD), n = 2; low-grade dysplasia (LGD), n = 1], 1 (0.2%) duodenal neoplasia (LGD) and 3 (0.5%) ampullary lesions (LGD). Compared to controls, cACLD patients had a higher prevalence of gastric neoplasia [2.1% vs.1%, p = 0.044; gastric cancer 1.6% vs. 0.8%, p = 0.08) and ampullary lesions (0.5% vs. 0%, p = 0.01). The prevalence of Helicobacter pylori infection was lower in patients compared to controls (36.2% vs. 47.2%, p = 0.004). In the 18 cACLD patients in whom were detected neoplastic lesions, 9 (50%) were diagnosed on the initial screening endoscopy. Five of these patients had transient elastography < 20 kPA and platelet count > 150000 before EGD. CONCLUSION: The prevalence of esophagogastroduodenal neoplastic lesions, particularly gastric neoplasia, in cACLD patients who underwent screening EGD is significant, being higher compared to controls. EGD should therefore still be considered in patients with ACLD from countries with intermediate gastric cancer risk, independent of the Baveno criteria.