Hepatitis B Virus RNA as a Biomarker for Safe Antiviral Discontinuation: A Prospective Study of Nucleos(t)ide Analogue Withdrawal

Abstract

Background: Withdrawal of nucleos(t)ide analogue therapy is associated with hepatitis B surface antigen (HBsAg) loss and sustained off-therapy partial cure (normal alanine aminotransferase [ALT] ≤30 U/L for males and ≤20 U/L for females with hepatitis B virus [HBV] DNA <2000 IU/mL) but should be offered only to those most likely to benefit. HBV RNA may be useful for risk stratification.

Methods: The Hepatitis B Research Network Immune-Active Trial prospectively evaluated treatment with tenofovir disoproxil fumarate (TDF) for 192 weeks ± peginterferon alfa-2a for the initial 24 weeks, followed by protocolized withdrawal of TDF among eligible participants (ClinicalTrials.gov NCT01369212). HBV RNA was evaluated as a predictor of ALT flares and sustained partial cure (HBV DNA <2000 IU/mL) 48 weeks after TDF withdrawal.

Results: Of 93 participants discontinuing TDF (n = 52, TDF + peginterferon alfa-2a; n = 41, TDF alone), 52 (55.9%) had unquantifiable HBV RNA at end of treatment. ALT flares >5 times the upper limit of normal at 48 weeks off therapy occurred in 33.3%, with pretreatment age (≥35 years) and quantifiable HBV RNA at end of treatment the best predictors (area under the receiver operating characteristic curves, 0.74 and 0.85 for training and test sets, respectively). A total of 26 (28.3%) had sustained partial cure, 3 (11.5%) with ALT flare. Nonquantifiable HBV RNA and quantitative HBsAg <100 IU/mL at end of treatment were the best predictors of sustained partial cure (area under the receiver operating characteristic curves, 0.84 and 0.93 for training and test sets). If HBV RNA was quantifiable at end of treatment, the likelihood of sustained partial cure was only 3%, whereas if HBV RNA was unquantifiable and quantitative HBsAg was <100 IU/mL, this likelihood was 73%.

Conclusions: HBV RNA is a useful biomarker in predicting likelihood of achieving sustained partial cure and safe withdrawal of nucleos(t)ide analogues.

Keywords: ALT flare; HBsAg loss; functional cure; partial cure; peginterferon.

Graphical Abstract

Figure 1.

Relationship between HBV RNA levels and study outcomes. The relationship between HBV RNA levels at EOT and (A) frequency of ALT flares, (B) frequency of partial cure, and (C) safe NA withdrawal. The figures were generated using LOESS. A higher proportion of ALT flares is evident with HBV RNA levels ≥3 log₁₀ IU/mL at EOT. The proportion of sustained partial cure increased with lower HBV RNA levels, particularly with levels ≤2 log₁₀ IU/mL at EOT. There was an inverse linear relationship between safe NA withdrawal and HBV RNA levels. ALT, alanine aminotransferase; EOT, end of treatment; HBV, hepatitis B virus; LOESS, locally estimated scatterplot smoothing; NA, nucleos(t)ide analogue; ULN, upper limit of normal.

Figure 2.

Prediction of key outcomes after NA discontinuation. A, ALT flares >5 times the ULN. The best classification tree identified age at baseline and HBV RNA as the best predictors of ALT flares within 48 weeks following TDF withdrawal. For participants aged <35 years, there was only a 11% probability of an ALT flare, as opposed to a 62% probability for participants aged ≥35 years when accompanied with quantifiable HBV RNA levels at the time of withdrawal. B, Sustained partial cure after NA discontinuation. The best classification tree identified quantifiable vs nonquantifiable HBV RNA and qHBsAg ≥100 IU/mL as the best predictors of sustained partial cure at 48 weeks following TDF withdrawal. If HBV RNA was quantifiable at EOT, the likelihood of a sustained partial cure was only 3%, whereas if HBV RNA was unquantifiable and qHBsAg was <100 IU/mL, this likelihood was 73%. Participants with unquantifiable HBV RNA and qHBsAg ≥100 IU/mL had an intermediate likelihood (29%) of a sustained partial cure. C, Prediction of safe withdrawal of NA therapy. The best classification tree identified HBV RNA at EOT and age at baseline as predictors of safe withdrawal at 48 weeks after NA discontinuation. For those aged <35 years, there was a 93% chance of no adverse consequences with stopping NA. For those aged ≥35 years, unquantifiable HBV RNA at EOT predicted a 59% likelihood of safe withdrawal. This latter group accounted for 33% of the study population. ALT, alanine aminotransferase; EOT, end of treatment; HBV, hepatitis B virus; NA, nucleos(t)ide analogue; qHBsAg, quantitative hepatitis B surface antigen; TDF, tenofovir disoproxil fumarate; ULN, upper limit of normal.