Quantification of HBsAg to predict low levels and seroclearance in HBeAg-negative patients receiving nucleos(t)ide analogues

BACKGROUND HBeAg-negative chronic hepatitis B patients require long-term nucleos(t)ide analogues(NAs) because loss of surface antigen (HBsAg) is unusual. Low quantitative HBsAg (qHBsAg) levels can identify patients with higher probability of seroclearance. The aim of our study was to evaluate qHBsAg in HBeAg-negative patients receiving NAs to predict a reduction of HBsAg levels and seroclearance. METHODS Retrospective analysis of qHBsAg in HBeAg-negative patients before and at years 1, 3, 5, 8 and over of NAs treatment. RESULTS From 1999 to 2015, HBsAg was quantified in 358 serum samples from 95 HBeAg-negative patients. Low qHBsAg (<120 IU/mL) was identified at baseline or during follow-up in 14% of patients and HBsAg loss in 4%. No baseline variables predicted seroclearance and only treatment duration predicted low qHBsAg. The annual decline of qHBsAg was -0.102 log IU/mL and the median time to HBsAg loss was 6.04 years. The decline was greater in patients achieving low HBsAg levels (-0.257) than in those who did not (-0.057)(p<0.001). The diagnostic accuracy (ROC curve, 95%CI) of qHBsAg delta at year 3 was 0.89 (0.81-0.97), with cut-off >0.3 log IU/mL showing a positive and negative predictive value of 42% and 100% to identify patients achieving low levels of HBsAg. CONCLUSIONS Reduction of qHBsAg is slow in HBeAg-negative patients receiving NAs, although low levels or faster qHBsAg decline may occur in 14%. A qHBsAg reduction >0.3 log IU/mL at year 3 can identify patients with a higher probability of achieving low levels and HBsAg seroclearance.