Dynamics of Circulatory <scp>HBcrAg</scp> and <scp>HBsAg</scp> After Nucleos(t)ide Analogue Cessation Using <scp>iTACT</scp> Assays to Predict Clinical Flares and <scp>HBsAg</scp> Loss

BACKGROUND: Circulatory HBsAg and HBcrAg levels may fluctuate after nucleos(t)ide analogue (NA) withdrawal. AIMS: We aimed to characterize their dynamics and evaluate their performance in predicting clinical flares (CF) and HBsAg loss. METHODS: Non-cirrhotic patients with chronic hepatitis B discontinuing NA according to Asia-Pacific criteria were enrolled. CF was defined as ALT elevation > 5 times upper limit of normal following viral relapse. Circulatory HBsAg and HBcrAg levels were quantified using iTACT assays at treatment cessation and every 6 months thereafter until retreatment (up to 3 years). Predictive performance was evaluated using time-dependent models. RESULTS: Of the 206 patients (median age, 50.0 years; 78.6% male) with a median treatment duration of 37.6 months (47.6% on entecavir), 68 experienced a CF (median follow-up, 36.1 months) and 30 achieved HBsAg loss (median follow-up, 70.3 months). Both HBcrAg and HBsAg decreased over time in patients with favourable outcomes (absence of CF or achievement of HBsAg loss). HBcrAg significantly outperformed HBsAg in predicting CF, with higher time-dependent AUROCs (69.7%-71.3% vs. 57.8%-65.2%). Multivariable analyses confirmed that HBcrAg dynamics predicted CF more accurately (adjusted sub-distribution hazard ratio [aSHR], 1.71 per log U/mL; 95% CI, 1.48-1.98), whereas HBsAg dynamics were superior in predicting HBsAg loss, with lower levels indicating a higher likelihood of clearance (aSHR, 0.33 per log IU/mL; 95% CI, 0.23-0.47). CONCLUSIONS: HBcrAg dynamics more accurately predict CF, while HBsAg dynamics better predict HBsAg loss after NA cessation, supporting their complementary roles in clinical practice and suggesting distinct biological mechanisms underlying their production.