On-treatment viral factors affect subsequent hepatitis B surface antigen seroclearance in patients treated with nucleos(t)ide analogs for > 10 years

BACKGROUND: Long-term nucleos(t)ide analog (NUC) treatment improves the outcomes of patients with chronic hepatitis B virus (HBV) infection. However, only a limited number of patients treated with NUC can achieve hepatitis B surface antigen (HBsAg) seroclearance, the so-called "functional cure." However, it remains unclear how on-treatment viral factors affect HBsAg seroclearance during long-term NUC treatment. We aimed to investigate whether the baseline and on-treatment HBV markers can predict HBsAg seroclearance and reduction in patients treated with long-term NUC treatment. METHODS: This study included two independent cohorts consisting of 843 patients in the derivation cohort and 1781 patients in the validation cohort. RESULTS: HBsAg seroclearance was infrequent (3.7-6.2%), with annual rates of 4-10/1,000 person-years in the derivation and validation cohorts. In baseline hepatitis B e-antigen (HBeAg)-positive patients, early on-treatment HBeAg loss strongly predicted subsequent HBsAg seroclearance and a reduction of < 10 IU/mL, whereas delayed or absent HBeAg loss rarely followed HBsAg seroclearance. A simple predictive model for HBsAg seroclearance based on earlier HBeAg loss, sex, and age was developed (SLOPES50). In baseline HBeAg-negative patients, low baseline or on-treatment HBsAg levels (< 100 IU/mL) were key predictors of HBsAg seroclearance or a reduction of < 10 IU/mL. Landmark analysis and time-dependent Cox regression analyses confirmed these associations in both cohorts. A substantial number of patients remained HBsAg ≥ 100 even after 15 years of NUC treatment in both cohorts. CONCLUSIONS: Functional cure during prolonged NUC treatment of > 10 years depends on early virological responses in both HBeAg-positive and HBeAg-negative patients.