Abstract
Background
In the setting of hepatitis C virus (HCV) active infection, liver stiffness (LS)–based strategies identify patients with low risk of developing esophageal variceal bleeding (VB) episodes, in whom unnecessary upper esophagogastroduodenoscopy (UGE) screening can be safely avoided. However, after sustained virological response (SVR), data on the accuracy of the criteria predicting this outcome in HCV-infected patients with cirrhosis, with or without human immunodeficiency virus (HIV) coinfection, are very limited.
Methods
This was a multicenter prospective cohort study, where HCV-monoinfected patients and HIV/HCV-coinfected individuals were included if they had (1) SVR with direct-acting antiviral–based therapy; (2) LS ≥9.5 kPa previous to treatment; and (3) LS measurement at the SVR time-point ≥14 kPa. Diagnostic accuracy of HEPAVIR, expanded Baveno VI, and HIV cirrhosis criteria, at the time of SVR, was evaluated. Missed VB episodes, negative predictive values (NPVs), and number of spared UGEs were specifically assessed.
Results
Four hundred thirty-five patients were included, 284 (65%) coinfected with HIV. Seven (1.6%) patients developed a first episode of VB after SVR. In patients without a previous VB episode, HEPAVIR, expanded Baveno VI and HIV cirrhosis criteria achieved NPV for first VB episode after SVR of 99.5% (95% confidence interval [CI], 97.1%–100%), 100% (95% CI 97.8%–100%), and 100% (95% CI 98%–100%) while sparing 45%, 39%, and 44% of UGEs, respectively. When considering HIV coinfection, the performance of the 3 criteria was similar, both in HCV-monoinfected and HIV/HCV-coinfected individuals.
Conclusions
After SVR, predictive LS-based strategies accurately identify HCV-infected patients, HIV coinfected or not, with low risk of developing VB during follow-up. In these specific patients, using HIV cirrhosis criteria maximize the number of spared UGEs while missing no VB episode.
Prediction of liver stiffness by serum indexes in HCV-infected patients with or without HIV coinfection
