Is Transient Elastography Needed for Noninvasive Assessment of High-Risk Varices? The REAL Experience.

Abstract

The Baveno VI consensus guidelines and an expanded algorithm suggest that transient elastography (TE) and platelet (PLT) count can be used to identify patients with cirrhosis who can avoid esophagogastroduodenoscopy (EGD). The primary aims of this study were to assess the ability of a simple algorithm, which uses only laboratory parameters, to predict medium/large esophageal varices (EV) in patients with hepatitis C virus (HCV) and cirrhosis from the Rete Sicilia Selezione Terapia-HCV (RESIST-HCV) cohort and to compare the performance of the algorithm with Baveno VI and Expanded Baveno VI criteria. The secondary aim was to assess the role of TE in ruling out large EV. In total, 1,381 patients with HCV-associated cirrhosis who had EGD and TE within 1 year of starting treatment with direct-acting antivirals were evaluated. Using multivariate logistic analysis, laboratory variables were selected to determine which were independently associated with medium/large EV to create the RESIST-HCV criteria. These criteria were tested in a training cohort with patients from a single center (Palermo) and validated with patients from the 21 other centers of the RESIST-HCV program (validation cohort). In the entire cohort, medium/large EV were identified in 5 of 216 patients (2.3%) using the Baveno VI criteria and 13 of 497 patients (2.6%) using the Expanded Baveno VI criteria. PLT count and albumin level were independently associated with medium/large EV. The best cut-off values were a PLT count greater than 120 × 10 cells/μL and serum albumin level greater than 3.6 g/dL; negative predictive values (NPVs) were 97.2% and 94.7%, respectively. In the training cohort of 326 patients, 119 (36.5%) met the RESIST-HCV criteria and the NPV was 99.2%. Among 1,055 patients in the validation cohort, 315 (30%) met the RESIST-HCV criteria and the NPV was 98.1%. Adding TE to the RESIST-HCV criteria reduced the avoided EGDs for approximately 25% of patients and the NPV was 98.2%. The "easy-to-use" RESIST-HCV algorithm avoids EGD for high-risk EV screening for more than 30% of patients and has the same performance criteria as TE. Using these criteria simplifies the diagnosis of portal hypertension.