Long-term outcomes and significance of HBeAg Seroclearance in Chronic Hepatitis B: Novel Predictive Scores for HCC and HBsAg seroclearance

Abstract

Background and Aims: We aimed to determine the clinical outcomes, the factors and predictive scores for hepatocellular carcinoma (HCC) and hepatitis B surface antigen (HBsAg) seroclearance of a large cohort of patients undergoing HBeAg seroclearance (ESC). Method: Patients with documented ESC were followed up 3–6 monthly. Baseline characteristics and longitudinal laboratory results were recorded. Predictive scores for HCC (HCC-ESC) and HBsAg seroclearance (HBsAg-ESC) were derived from multivariate Cox regression models. Results: A total of 723 patients underwent ESC with a median ESC age and follow-up of 36.0 and 18.3 years respectively. Only 3.5% and 3.0% had persistently normal ALT and HBV DNA <2logsIU/ml respectively after ESC. For patients with 100%, 100–90%, 90–50%, 50–10%, 10–0%, and 0% normal ALT after HBeAg seroclearance, the rate of HCC was 4.3%, 2.2%, 3.6%, 3.9%, 17.3%, and 37.2% at 20 years after ESC respectively (p < 0.001). At 20 years after ESC, the cumulative incidence of HCC and HBsAg seroclearance was 7.9% and 13.5% respectively, with an overall survival of 91.5%. ESC age, male sex, cirrhosis, hypoalbuminemia, viral load, and ALT were significant factors for HCC, whereas ESC age, male sex, viral load, and antiviral therapy were significant factors for HBsAg seroclearance. The HCCESC score to predict the HCC risk for up to 20 years after ESC was calculated using age (years) + 20*sex (male = 1; female = 0) + 29*cirrhosis (presence = 1; absence = 0) + 5*DNA (logIU/ml) + 31*ALT group (flares or persistently abnormal ALT = 1; otherwise = 0) + 23*hypoalbuminemia (<39 g/l, presence = 1; absence = 0). With an optimal cut-off of 129, 121, and 114, the AUROC for predicting development of HCC at 5, 10 and 20 years after HBeAg seroclearance was 0.95, 0.91, and 0.92 respectively. The HBsAg-ESC score to predict HBsAg seroclearance up to 20 years was calculated using age (years) + 15*sex (male = 1; female = 0)−6*DNA (logIU/ml)−40*history of treatment (presence = 1; absence = 0). Using optimal cut-offs of 19, 17, and 12 to predict HBsAg seroclearance at 5, 10, and 20 years after HBeAg seroclearance were associated with an AUROC of 0.88, 0.84, and 0.74 respectively. Conclusion: Male gender, older age at ESC, ALT, and higher level of HBV DNA were associated with higher rates of HCC after ESC. HCCESC and HBsAg-ESC predictive scores can determine the likelihood of developing HCC and achieving HBsAg seroclearance.