Extending and validating the Baveno VI criteria for the exclusion of high-risk varices.

Abstract

The updated Baveno VI guidelines recommend that screening for high-risk varices (HRV) by esophago-gastro-duodenoscopy (EGD) can be avoided in patients with compensated advanced chronic liver disease (cACLD) who have liver stiffness LSM<20 kPa and platelet count PLT>150,000/L. The aims of this study were to validate extended Baveno VI criteria in patients with chronic liver disease and to establish cut-off values for our cohort. This retrospective study included 839 patients with liver cirrhosis evaluated by Transient Elastography (TE), biological tests, and upper endos-copy, all within one year. The Baveno VI criteria were validated on a sub-group of 728 patients (Cohort 1, randomly selected from the study sample) and tailored cut-off points were determined. The remaining 111 patients comprised the validating set (Cohort 2) for these specific cut-off values. In Cohort 1, Baveno VI criteria had 86.2% accuracy. The calculated cut-offs to rule-in HRV were PLT<150,000/mm3 and LSM >35.3 kPa; while to rule-out HRV they were PLT >150.000/mm3 and LSM <19.6 kPa. In patients in the "grey-zone", by multivariate analysis, albumin was independently associated with HRV at a cut-off of ˂3.4 g/dl. In the validation cohort, the calculated rule out cut-offs had 100% accuracy. The Baveno VI criteria had a good accuracy for exclusion of HRV in this large cohort of cirrhotic patients. Adding an albumin-related threshold increased performance and broadened applicability. Using the calculated rule-out criteria for HRV, all unnecessary EGD could be excluded.