Validation of the Baveno Endoscopy Avoidance Criteria in Japanese Patients With Chronic Liver Disease: A Real‐World Cohort Study

BACKGROUND: The Baveno endoscopy avoidance criteria are widely used to reduce unnecessary screening endoscopy in patients with chronic liver disease. However, evidence regarding their performance in Japanese populations remains limited, and detection-oriented strategies for high-risk varices (HRV) may also be clinically relevant in settings with broad endoscopic availability. METHODS: We retrospectively analyzed 1675 Japanese patients with chronic liver disease who underwent liver stiffness measurement (LSM) and upper gastrointestinal endoscopy at a single center. The performance and safety of the Baveno criteria were evaluated overall and by etiology. Patients were temporally divided into derivation and validation cohorts. Based on independent predictors of HRV, a simple high-risk varices detection score (HRV-DS) incorporating LSM, platelet count, and alcohol-related etiology was developed. RESULTS: The Baveno criteria classified 79.4% of patients as suitable for endoscopy avoidance, with an overall HRV miss rate of 0.39%. However, safety was modestly reduced in patients with alcohol-related liver disease. The HRV-DS achieved 100% sensitivity for HRV detection at the lowest threshold (score ≥ 1), while recommending endoscopy to 27.0% of patients. At a higher threshold (score ≥ 2), sensitivity remained approximately 87% with improved positive predictive value. These findings were consistently reproduced in both derivation and validation cohorts. CONCLUSIONS: In this large real-world Japanese cohort, the Baveno endoscopy avoidance criteria demonstrated excellent safety for excluding HRV. The exploratory HRV-DS also showed high sensitivity for identifying HRV and may provide a complementary detection-oriented approach in clinical practice.