A novel baseline hepatitis B virus sequencing-based strategy for predicting adefovir antiviral response

Adefovir dipivoxil (ADV) is used as first-line monotherapy or rescue therapy in chronic hepatitis B (CHB) patients. In this study, we sought to identify nucleotide changes in the reverse transcriptase (RT) of hepatitis B virus (HBV) at baseline and explore their predictive value for ADV antiviral response. Ultra-deep pyrosequencing (UDPS) was utilized to determine HBV genetic variability within the RT region at baseline and during a 48-week ADV therapy. According to the viral load at the end of ADV treatment, all patients were classified into responders (HBV DNA level reduction of ⩾3 log 10 IU/mL) and suboptimal responders (HBV DNA level reduction of <3 log 10 IU/mL). Based on UDPS data at baseline, we identified 11 nucleotide substitutions whose combination frequency was significantly associated with the antiviral response among 36 CHB patients in the study group. However, the baseline distribution and frequency of rt181 and rt236 substitutions known to confer ADV resistance was a poor predictor for the antiviral response. Compared with baseline serum HBeAg, HBV-DNA and ALT levels, the baseline HBV sequence-based model showed higher predictive accuracy for ADV response. In an independent cohort of 31 validation patients with CHB, the sequence-based model provided greater predictive potency than the HBeAg/HBV-DNA/ALT and the HBeAg/HBV-DNA/ALT/sequence combinations. Taken together, we confirm the presence of ADV resistance variants in treatment-naïve patients and firstly unravel the predictive value of the baseline mutations in the HBV RT region for ADV antiviral response.