Dynamics of liver stiffness-based risk prediction model during antiviral therapy in patients with chronic hepatitis B

Objective The liver stiffness-based risk prediction models predict hepatocellular carcinoma (HCC) development. We investigated the influence of antiviral therapy (AVT) on liver stiffness-based risk prediction model in patients with chronic hepatitis B (CHB). Methods Patients with CHB who initiated AVT were retrospectively recruited from 13 referral Korean institutes. The modified risk estimation for hepatocellular carcinoma in chronic hepatitis B (mREACH-B) model was selected for the analysis. Results Between 2007 and 2015, 1034 patients with CHB were recruited. The mean age of the study population (639 men and 395 women) was 46.8 years. During AVT, the mREACH-B score significantly decreased from the baseline to 3 years of AVT (mean 9.21 → 7.46, P < 0.05) and was maintained until 5 years of AVT (mean 7.23, P > 0.05). The proportion of high-risk patients (mREACH-B score ≥11) was significantly reduced from the baseline to 2 years of AVT (36.4% → 16.4%, P < 0.001) and was maintained until 5 years of AVT (12.2%, P > 0.05). The mREACH-B scores at baseline and 1 year of AVT independently predicted HCC development (hazard ratio = 1.209-1.224) (all P < 0.05). The cumulative incidence rate of HCC was significantly different at 5 years of AVT among risk groups (high vs. high-intermediate vs. low-intermediate vs. low) from baseline (4.5% vs. 3.2% vs. 1.5% vs. 0.8%) and 1 year (11.8% vs. 4.6% vs. 1.8% vs. 0.6%) (all P < 0.05, log-rank tests). Conclusions The mREACH-B score was dynamically changed during AVT. Thus, repeated assessment of the mREACH-B score is required to predict the changing risk of HCC development in patients with CHB undergoing AVT.