Antiviral drug-resistant HBV: Standardization of nomenclature and assays and recommendations for management


  • Potential conflict of interest: Dr. Lok receives research support and serves on advisory board of GlaxoSmithKline, Gilead, Bristol Myers Squibb, Idenix, Roche Molecular Diagnostics and Innogenetics. Dr. Zoulim serves on advisory board of Gilead, Idenix, Bristol Myers Squibb, Innogenetics and receives research support from bioMerieux. Dr. Locarnini receives research support and serves on advisory board of Gilead Sciences, Bristol-Myers Squibb, Pharmasset, LG Sciences, Innogenetics and Evivar Pty Ltd. Dr. Bartholomeusz is a Consultant to Evivar Medical Pty Ltd. Dr. Pawlotsky receives research support and/or serves on advisory board of GlaxoSmithKline, Gilead, Bristol-Myers Squibb, Idenix and Novartis. Dr. Liaw receives research support and serves on advisory board and speaker bureau of Roche, Bristol-Myers Squibb, GlaxoSmithKline, and Idenix.


Substantial advances have been made in the treatment of chronic hepatitis B in the past decade. Approved treatments for chronic hepatitis B include 2 formulations of interferon and 4 nucleos(t)ide analogues (NAs). Sustained viral suppression is rarely achieved after withdrawal of a 48-week course of NA therapy, necessitating long, and in many cases, indefinite treatment with increasing risk of development of drug resistance. Antiviral resistance and poor adherence are the most important factors in treatment failure of hepatitis B. Thus, there is a need to standardize nomenclature relating to hepatitis B antiviral resistance, and to define genotypic, phenotypic, and clinical resistance to NA therapy. (HEPATOLOGY 2007;46:254–265.)

1. Introduction

Substantial advances have been made in the treatment of chronic hepatitis B in the past decade. Approved treatments for chronic hepatitis B have expanded from just one agent to a total of six agents: standard interferon (IFN), pegylated IFN, lamivudine (LAM), adefovir dipivoxil (ADV), entecavir (ETV), and telbivudine (LdT). In addition to the four approved nucleos(t)ide analogues (NAs) for chronic hepatitis B, tenofovir disoproxil fumarate (TDF), the prodrug of tenofovir and the coformulation of TDF and emtricitabine, approved treatments for human immunodeficiency virus (HIV) infection, have activity against hepatitis B virus (HBV).

Although NAs are more convenient than IFN-based therapies and have fewer side effects, sustained viral suppression is usually not achieved after withdrawal of a 48-week course of NA therapy, necessitating long, and in many cases, indefinite treatment. Unfortunately, a long duration of NA treatment is associated with an increasing risk of development of drug resistance. Antiviral resistance and poor adherence are the most important factors in treatment failure. Thus, there is a need to standardize nomenclature relating to HBV antiviral resistance, and to define genotypic, phenotypic, and clinical resistance to NA therapy. In this review, we propose definitions of terminologies, briefly describe available methods for detecting and quantifying drug resistance, and discuss the interpretation of drug resistance data and its current and future application in clinical practice.

(a) Clinical Classification of Antiviral Resistance

At the 2006 National Institutes of Health Workshop on HBV, standardized definitions of response to antiviral therapy were proposed (Table 1).1 More detailed clarification of terminologies that are commonly used in describing antiviral resistance is provided here. A sensitive and specific HBV DNA assay with a wide dynamic range of quantification, calibrated to express results in WHO international units per ml (IU/ml), should be used to quantify serum HBV DNA levels prior to treatment, to assess response, and to detect virologic breakthroughs.

(i) Primary Antiviral Treatment Failure (or Nonresponse).

Primary nonresponse is defined as the inability of NAs to reduce serum HBV DNA by ≥1 log10 IU/ml after the first 6 months of treatment. This definition was chosen as it exceeds variability in HBV DNA assays and reflects a true virologic response but a decrease in HBV DNA level by 1 log10 IU/ml is not a meaningful clinical response. Primary nonresponse may be due to factors related to the host, virus, or drug. Polymorphisms in enzymes involved in converting prodrugs to the active compounds or in phosphorylating NAs to their triphosphates (active moiety) may contribute to primary nonresponse.2, 3 Certain viral strains may be less susceptible to one or more antiviral therapies as suggested by a recent case report on primary nonresponse to adefovir.4 Potency and dose of the antiviral therapy may also be important. For example, the approved dose of adefovir (10 mg daily) is not as potent as higher doses and may be the most important factor in the high rate of primary nonresponse to adefovir.5

Monitoring for primary nonresponse is important because a high residual viral level after the first 6-12 months of therapy has been demonstrated to be associated with increased risk of antiviral resistance.6–8

(ii) Secondary Antiviral Treatment Failure (or Virologic Breakthrough).

Virologic breakthrough, which is usually associated with drug resistance, is defined as a ≥1 log10 IU/ml increase in serum HBV DNA level from nadir in two consecutive samples 1 month apart in patients who have responded and have been compliant with antiviral medication(s).9, 10 Confirmation of the increase in serum HBV DNA in a second sample is not necessary in patients with accompanying flare in aminotransferase level. Serum HBV DNA levels tend to be low initially because most antiviral-resistant HBV mutants have decreased replication fitness compared to wild-type HBV.11, 12 However, compensatory mutations that can restore replication fitness frequently accumulate during continued treatment leading to viral rebound — progressive increase in serum HBV DNA level that may exceed pretreatment value.10

(iii) Biochemical Breakthrough.

Biochemical breakthrough is defined as elevation in serum aminotransferase level during treatment in a patient who had achieved initial normalization. Serum aminotransferases may remain normal for a few weeks or a few years after virologic breakthrough. Biochemical breakthrough often coincides with a viral rebound, in some cases a marked increase in aminotransferases occur resulting in a hepatitis flare (aminotransferase >5 times the upper limit of normal) and rarely hepatic decompensation.13, 14

Table 1. Nomenclature for Antiviral Resistance
Primary treatment failure (nonresponse)Inability of nucleoside/tide analogue treatment to reduce serum HBV DNA by ≥1 log10 IU/ml after the first 6 months of treatment
Secondary treatment failure (virologic breakthrough)Increase in serum HBV DNA by ≥1 log10 above nadir on ≥2 occasions 1 month apart, while on treatment, after achieving initial response in a medication compliant patient
Biochemical breakthroughElevation in serum alanine aminotransferase (ALT) while on treatment, after achieving normalization in a medication compliant patient
Genotypic resistanceDetection of viral populations bearing amino acid substitutions in the reverse transcriptase region of the HBV genome that have been shown to confer resistance to antiviral drugs in phenotypic assay, during antiviral therapy. These mutations are usually detected in patients with virologic breakthrough but they can also be present in patients with persistent viremia and no virologic breakthrough.
Phenotypic resistanceDecreased susceptibility of an HBV polymerase to an antiviral treatment in vitro
Cross resistanceDecreased susceptibility to more than one antiviral drug conferred by the same amino acid substitution or combination of amino acid substitutions

ADV, Adefovir; LAM, lamivudine; EC50, effective concentration50; LdT, telbivudine; ETV, entecavir; NA, nucleoside/tide analogue; HBV, hepatitis B virus; PCR, polymerase chain reaction; HIV, human immunodeficiency virus; RFLP, restriction fragment length polymorphism; IFN, interferon; TDF, tenofovir.

2. Definition of Genotypic Antiviral Resistance

A fundamental issue in antiviral resistance is the criteria for defining drug-resistant mutations. When a mutation occurs during replication, it results in a nucleotide substitution. The substitution can be synonymous (not associated with an amino acid change) or non-synonymous (associated with an amino acid change). A resistance mutation induces an amino acid change that decreases the sensitivity to an antiviral drug.

Genotypic antiviral resistance designates the presence of unique nucleotide and corresponding deduced amino acid mutations in the drug target gene (the HBV polymerase gene in the case of HBV treatment with NAs) that have been previously demonstrated to be associated with antiviral resistance. Ideally, to identify potential genotypic resistance, the nucleotide and deduced amino acid sequence of HBV isolated from the patient during virologic breakthrough should be compared to the sequence of HBV isolated from a pretreatment sample from the same patient. When pretreatment samples are not available for analysis, sequence data at the time of virologic breakthrough should be compared to consensus published sequence(s) of the same HBV genotype.

Primary drug resistant mutations cause an amino acid substitution that result in reduced susceptibility to an antiviral agent while secondary compensatory mutations cause amino acid substitutions that restore functional defects in viral polymerase activity (i.e., replication fitness) associated with primary drug resistance.

For example, primary lamivudine resistance associated changes occur at codon 204 and result in amino acid changes within the tyrosine-methionine-aspartate- aspartate (YMDD) motif - rtM204V/I (methionine to valine or isoleucine substitution). These changes cause a greater than 100-fold decrease in susceptibility to lamivudine in phenotypic assays. The most common compensatory mutation associated with lamivudine resistance, rtL180M (leucine to methionine substitution) restores replication fitness of HBV polymerase that harbors the rtM204V/I mutation.12

While no true competitive replication fitness assay has been developed, the current in vitro assays can determine relative replication yield phenotype to compare HBV DNA encoding specific mutations to the reference, consensus, or wild-type genetic framework.15, 16 For example, HBV replication competent clones encoding the rtV173L (valine to leucine substitution) + rtL180M + rtM204V were demonstrated to have an increased replication yield phenotype but no change in sensitivity to lamivudine relative to HBV clones encoding rtL180M + rtM204V.17 Therefore, the rtV173L change should be considered a compensatory mutation in the HBV DNA genome.

New HBV nucleotide and corresponding deduced amino acid mutations detected in 2 or more patients undergoing the same antiviral treatment with virologic breakthrough despite medication compliance can be defined as a “putative” resistance mutation, and given a provisional status until confirmatory in vitro phenotype testing can be performed. Thus, there need to be sufficient data to determine if the selection of a particular nucleotide mutation and deduced amino acid change is unique to the antiviral selection pressure. In addition, the mutation should not be detected in patients who have continued to respond to the same antiviral treatment or be present prior to treatment using readily available assays that detect mutants comprising >5% of the viral population. Although antiviral-resistant mutations can be detected using ultra-sensitive assays (see below) from pretreatment samples and/ or before virologic breakthrough, the likelihood of HBV encoding the mutation(s) as the dominant quasispecies in pretreatment samples without selection pressure is very low.

Confirmation of genotypic resistance is based on 2 methods: (a) in vitro phenotypic analysis and (b) virtual phenotypic analysis which is the correlation of patient treatment and response data with HBV sequence data.

(i) In Vitro Phenotypic Analysis.

In vitro phenotypic assay is the “gold standard” to confirm genotypic antiviral resistance. Unfortunately, the methodology is time consuming and labor intensive due to the lack of a convenient cell culture system and the need to use specific HBV replication competent clones. In addition, multiple substitutions or sequences elsewhere in the HBV genome may influence the result.

(ii) Virtual Phenotypic Analysis: Correlation of Patient Treatment and Response Data with HBV Sequence Data.

This method relies on relational HBV databases with both clinical, virological, and HBV sequence information that are integrated and analyzed statistically via linkage.18 Large numbers of patients with virologic breakthrough during treatment are required. The input clinical and sequence data are compared to the database to determine the best match and most likely treatment response.

(a) Nomenclature

After the initial detection and reporting of antiviral resistance, there was confusion in the naming of drug-resistant mutations as the HBV genotypes vary in genomic length. In 2001, Stuyver and colleagues overcame this problem by dividing the HBV polymerase into four different functional units and re-numbering each functional unit.19 The reverse transcriptase (rt) region of the polymerase gene is common for all genotypes. Mutations within this region are prefixed with the letters rt followed by the original deduced amino acid, the codon number relative to the start of the rt region, followed by the deduced amino acid derived by the mutation. For example, the primary LAM resistance associated changes are defined as rtM204I and rtM204V using this nomenclature (substitution of methionine at codon 204 in the reverse transcriptase region of the HBV polymerase gene for isoleucine or valine). These LAM-resistant changes should not be referred to as “a YMDD mutant” as only the methionine of the conserved YMDD locus is changed.

(b) Technical Issues Associated with the Genotypic Identification of New Antiviral Resistance Mutations

The accepted standard for characterizing the sequence of HBV polymerase in order to identify known or new resistance mutations is double-stranded (or bi-directional) sequencing of the PCR amplified product. Reports on newly identified antiviral resistance mutations should include changes in nucleotide and amino acid sequences in the reverse transcriptase region of the HBV polymerase as well as any deduced amino acid change in the overlapping envelope reading frame and the sequences should be deposited in Genbank. Where more than one nucleotide change (i.e., mixed HBV population) is detected at the same position, both deduced amino acids using the International Union of Pure and Applied Chemistry (IUPAC) codes should be listed. For example, a mixed population of wild type sequence and LAM-resistant mutation would be reported as rtM204M/V and the corresponding change from isoleucine at codon 195 of the overlapping HBV surface protein to a mixture of methionine and isoleucine indicated as sI195M/I.

(c) Methods to Detect Genotypic Resistance Mutations

Assays available to identify resistance mutations include direct sequencing of PCR products, PCR amplification followed by sequence analysis of multiple clones derived from the amplicon, real time PCR formats with specific probes including allele specific PCRs, hybridization methods such as the line probe assay, restriction fragment length polymorphism (RFLP), and more recently matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) based restriction fragment mass polymorphism (RFMP) which can detect mutants that comprise <1% of the viral quasispecies.20, 21 The more sensitive assays can detect HBV DNA encoding the resistance mutations that represent 5%-10% of the entire HBV quasispecies enabling earlier identification of patients with genotypic resistance mutations prior to or at the time of virologic breakthrough. However, the utility of ultra-sensitive assays that can detect mutants comprising <1% of the viral population in predicting drug resistance remains to be determined.

(i) Direct PCR Sequencing.

Direct PCR sequencing is the least sensitive in detecting minor populations of resistant mutants as HBV DNA encoding the resistance mutations needs to reach approximately 20% of the total HBV quasispecies pool before these mutations can be observed. Direct sequencing is also not amenable to high throughput screening. However, direct PCR sequencing does allow for all mutations to be identified, including additional potential compensatory mutations and new undefined mutations. For new therapies, or the identification of new mutations associated with resistance to existing therapies, in vitro phenotypic confirmatory assays are required.

(ii) RFLP Analyses.

RFLP analyses can detect viral mutants that constitute as little as 5% of the total viral population.22 However, separate sets of endonuclease reactions must be designed specifically for each mutant of interest. Some mutations result in a new restriction site and RFLP is therefore an easy method; some other mutations destroy a restriction site and in this case RFLP analysis should be used with caution as lack of enzyme digestion may be due to loss of a restriction site or technical problems with the assay. RFLP analysis may not be possible for all resistant mutations as endonucleases specific for such sequences may not exist.

(iii) Reverse Hybridization Assay LiPA DR.

The commercially available reverse hybridization assay LiPA DR (Innogenetics, Belgium) can detect single nucleotide mismatches and the assay contains a series of short membrane-bound oligonucleotide probes. LiPA assays can detect emerging viral resistance when HBV encoding the resistance mutations constitute 5% or more of the total viral population.21, 23–25 Their major limitation is that new sets of specific probes are required for every mutant and due to genotype variability a number of probes may be required to detect a single nucleotide change.

(iv) Sequencing with Microchip-Based Technology.

Sequencing with microchip-based technology using oligonucleotide microarrays may be used to detect “new” mutants. This technology is expensive and not widely available.26


MALDI-TOF MS is based on mass spectrometric analysis of small DNA fragments containing sites of variation.20 This assay has been shown to be very sensitive and can detect mutants that constitute only 1% of the viral population.21 However, a new set of primers must be designed to detect each new mutation and access to mass spectrometer is required.

(vi) Single Genome Sequencing.

Single genome sequencing has been used in studying antiviral-resistant HIV mutations.27, 28 This method is tedious and does not provide information on the absolute frequency of the mutation. In addition, given the high rate of spontaneous mutations during HBV replication, the clinical significance of mutants that may be present in <0.1% of the viral population is uncertain.

With four approved NAs for chronic hepatitis B and several more in development, designing assays that would permit detection of all the mutations known to confer resistance to these medications is increasingly difficult. Any of the above assays can be used for research as long as the method is specified in the report. The most commonly used methods in clinical practice include direct sequencing and line probe assay.

3. Definition of In Vitro Phenotypic Antiviral Resistance

Demonstration that a given amino acid substitution confers resistance is based on the use of phenotypic assays that demonstrate in vitro a reduced susceptibility of a replication competent clone with that substitution to an antiviral agent against HBV as compared to a replication competent clone with the “wild-type” sequence (without the substitution). In vitro phenotype testing is based on the determination of changes to the effective concentration of the drug required to inhibit 50% of the target (EC50or IC50) relative to the “wild-type” reference HBV.

Clinically, antiviral drug resistance is commonly described as high- (>100-fold increase), intermediate- (10-99–fold increase) or low-level (2-9–fold increase) (Table 2). Unfortunately, such ranking of drug resistance as determined from in vitro phenotypic tests does not readily translate to what is observed clinically. For example, a small decrease in in vitro ADV susceptibility (2-9–fold increase in EC50) may confer resistance in vivo.

Table 2. In vitro and in vivo Significance of Antiviral-Resistant Mutations
  • *

    (+ L180M + M204I/V); NA = Not Available

  • 2– 9 fold → no or low level of resistance; 10– 99 fold → medium level of resistance; >100 fold → high level of resistance

  • References:

  • a

    Chin et al.53;

  • b

    Delaney et al. (2001)34;

  • c

    Ono et al.73;

  • d

    Sozzi et al.74;

  • e

    Tenney et al.47;

  • f

    Brunelle et al.48;

  • g

    Sheldon et al.58;

  • h

    Delaney et al. (2006).59

HBV-Fold Resistance-Fold Resistance-Fold Resistance-Fold Resistance-Fold Resistance-Fold Resistance
M204l>100 ah>100 a4 d1 d<1-8abc<1d
L180M + M204V>100 ah>100 aNA5-6 ef<1-4abc3-6 fg
A181 T/V1-2 dhNA5-6 d1-4 d1-3 d1 d
N236T1 fhNA3 d<1f3 f5 f
I169T + V173L + M250V>1000 dhNA>1000 d>700 e1 d<1d
T184G + S202l *>1000 dhNA35 d>700 e2 d6 d

(a) Methods Used for Phenotyping

Antiviral susceptibility testing involves assaying the activity of the polymerase enzyme or of HBV replication. These methods are time consuming and require a high level of technical expertise.29

(i) Phenotyping Based on Enzymatic Assays.

Currently, the only assays for the study of HBV polymerase activity are based on the polymerase expressed in insect cells using a baculovirus vector, and on HBV polymerase expression in purified HBV nucleocapsids.30, 31 A cell-free assay has been developed for duck hepatitis B virus (DHBV) but HBV polymerase gene contains a large insert not present in DHBV or other hepadnavirus polymerase genes, and this insertion may affect the interpretation of phenotype testing results.32, 33

(ii) Phenotyping Based on Transient Transfection of Hepatocyte-Derived Cell Lines.

Two approaches based on transient transfection have been used for phenotyping. The first relies on site-directed mutagenesis to generate point mutations that may be associated with drug resistance in recombinant, well-characterized “laboratory” HBV replication competent clones. This approach may be an advantage for research purposes as specific mutation(s) can be determined as being or not being associated with reduced antiviral susceptibility. However, in some instances multiple mutations and/or the broader genetic framework of the HBV DNA from a patient may be required to confer antiviral resistance. The second method relies on amplified full-length HBV genomes from clinical isolates (rather than laboratory generated mutants).16 However, as a number of mutations and/ or clonal variants may be present; the effect may be a culmination of all the changes in the reverse transcriptase region as well as other aspects of the HBV genome sequence unique to that particular patient and not simply the effect of one particular mutation.

(iii) Phenotyping Based on Transduction of Recombinant Baculovirus/HBV into Hepatocyte-Derived Cell Lines.

Recombinant baculovirus/HBV has also been used for phenotyping.15, 34 In this system, HBV replication is driven by endogenous promoters and therefore studies on relative replication yield phenotype can be performed; however, this technique is tedious.15

(iv) Phenotyping Based on Continuous Cell Lines Containing Stably-Integrated HBV Genomes.

Stably-transfected HBV-expressing cell lines have been created specifically for the investigation of drug resistance.35 The advantage of using stable cell lines for phenotype testing is the ability to perform cross-resistance testing in a consistent environment but a new cell line needs to be created for each new mutant.36 The integration site of HBV DNA within the cellular chromosome may affect HBV replication and cellular function; therefore, these cell lines cannot be used to determine the relative replication efficiency of HBV encoding the antiviral resistance mutations.

4. HBV Resistance Databases and Virtual Phenotyping

Virtual Phenotyping involves assigning a phenotype for a clinical isolate based on the correlations from large databases containing genotypic, phenotypic, and clinical information.

The analytical program searches linked databases for the best matches among sequences known to confer particular phenotypes. Virtual phenotyping should be considered as an adjunct and not a substitute to in vitro phenotype testing.

The first program developed to correlate HBV patient clinical, virological, and HBV sequence information is SeqHepB Registered users can access the program online, and input either genomic (nucleic acid) or amino acid sequences of clinical HBV isolates for analysis. SeqHepB defines all amino acid variations within the input sequences compared to reference sequence(s) of the same genotype as mutations37 and correlate the results with a database of HBV genotype and phenotype data and clinical histories. Various data mining algorithms and functions are being developed that facilitate rapid and efficient identification of new markers of drug resistance. One such function includes the localization of deduced amino acid changes on a molecular model of the HBV reverse transcriptase.38 This function can provide insights into the significance of a mutation in relation to antiviral drug resistance and potential mechanism for antiviral resistance.38, 39

The European Network of Excellence on antiviral drug resistance management, ViRgil, is also developing a database using a common clinical record form and a centralized virology laboratory.

The Hepatitis Virus Database (HVDB) in Japan is open to the public It is updated 4 times a year using the newest release of DDBJ (DNA Data Bank of Japan) and currently contains 10,892 HBV entries. The HVDB is mainly geared towards phylogenetic analyses.

If properly designed, a database can greatly facilitate the tracking of known resistance mutations and the development of new ones. It would also allow more fundamental research on mutation interactions, epidemiological studies on the spread of resistant mutations, and systematic study of the risk factors associated with the emergence of resistant mutations and the clinical outcomes of patients with antiviral resistance. Additional advantages are an easily accessible overview of known resistance mutations for treating physicians and timely dissemination of information on resistance profile changes.

5. Mutations Associated with Antiviral Resistance

The incidence of genotypic resistance is related to viral (pretreatment serum HBV DNA level, pre-existing antiviral-resistant mutations), host (immune status, pharmacodynamics), and treatment characteristics (potency, genetic barrier to resistance [number of mutations required to produce a marked decrease in susceptibility to the antiviral drug], and duration of treatment). The incidence of genotypic resistance also varies with the sensitivity of the methods used for detection of resistant mutations (see section 2: Definition of Genotypic Antiviral Resistance) and the patient population being studied. Thus, clinical studies have varied from testing samples from all patients with detectable serum HBV DNA by PCR assay using sensitive methods such as RFLP or reverse hybridization to testing only samples from patients with viral rebound (such as >5 log10 copies/ml or >4.3 log10 IU/ml) using less sensitive methods such as direct sequencing (Table 3). These approaches result in reports of LAM-associated resistance mutations after 1 year of therapy varying from 7% to 23%.14, 40 Figure 1 illustrates mutations associated with approved HBV therapies.

Table 3. Rates of Antiviral-Resistant HBV Mutations Reported in Clinical Trials
LAM-R ptsRates of genotypic resistancePatients testedMethod used
  • Abbreviations: LAM-R, lamivudine resistance; pts, patients.

  • References:

  • a

    Dienstag et al.75;

  • b

    Lai et al. (1998)76;

  • c

    Schalm et al.77;

  • d

    Lok et al.14;

  • e

    Westland et al.78;

  • f

    Hadziyannis et al.79;

  • g

    Fung et al.57;

  • h

    Lee et al.80;

  • i

    Colonno et al.62;

  • j

    Colonno et al.81;

  • k

    Lai et al. (2005).82

LAM15-30% after 1 yr abc 70% after 5 yr dAll pts PCR+RFLP
ADV0% after 1 yr e 29% after 5 yr fAll pts PCR+Direct sequencing
 LAM-R pts∼20% after 2 yr ghAll pts PCR+Line-probe, MALDI-TOF
ETV0% after 1 yr <1% after 2 & 3 yr iAll pts PCR+Direct sequencing
 LAM-R pts1%, 9%, ∼17% after 1, 2 & 3 yr jPts with viral reboundDirect sequencing
LdT2-3% after 1 yr kPts with viral reboundDirect sequencing
(LAM)7-8% after 1 yr k Direct sequencing
Figure 1.

The hepatitis B virus (HBV) polymerase open reading frame showing the conserved domains and location of the primary antiviral drug resistance mutations within the conserved domains. Rt, reverse transcription, LAM, lamivudine, ADV, adefovir dipivoxil, ETV, entecavir, LdT, telbivudine.

(a) Resistance to Monotherapies

(i) Lamivudine and Other L-Nucleosides.

The primary LAM resistance mutation maps to codon rtM204V/I in the YMDD motif.12, 41–43In vitro studies showed that these mutations decrease sensitivity to lamivudine by >100-fold (Table 3). The molecular mechanism of LAM resistance is steric hindrance caused by the β-branched side group of the valine or isoleucine amino acids colliding with the oxathiolane ring of LAM within the dNTP binding site.38 The rtL180M is the main compensatory change.12 Other compensatory mutations include the rtV173L and rtL80I changes.9, 17 The rtA181T change has been reported to occur in the absence of rtM204I/V and is considered a primary resistance mutation.44, 45 The rtA181T change is also selected during ADV treatment.46 LAM-resistant mutations rtM204V/I +/− L180M decrease susceptibility to ETV.36, 47

The primary LAM resistance mutations −rtM204V/I are cross-resistant with other L-nucleosides such as emtricitabine (FTC), telbivudine (beta-L-thymidine, LdT), beta-L-2′-deoxycytidine (LdC), elvucitabine and clevudine (L-FMAU; 2′-fluoro-5-methyl-beta-L-arabinofuranosyluracil), and the rtM204V/I changes have been selected during treatment with these compounds except for telbivudine where only rtM204I but not rtM204V has been observed.7, 48–51

Telbivudine is not active against HBV encoding rtM204I or rtM204V + rtL180M in cell culture (www. Lamivudine, telbivudine, and clevudine exhibit modest activity against HBV encoding rtM204V alone in cell culture52, 53 but rtM204V mutation alone is rarely detected in patients. Thus, neither clevudine nor telbivudine is expected to be efficacious in patients with lamivudine-resistant HBV.

(ii) Acyclic Phosphonates.

Adefovir (ADV) — The primary ADV-resistance mutations are rtN236T and /or rtA181T/V.39, 46, 54 Isolates of HBV with the rtN236T change are susceptible to LAM while isolates with rtA181T/V changes have decreased susceptibility to LAM (Table 3). These ADV-associated mutations result in only a modest (2-9–fold) increase in EC50 but viral rebound, hepatitis flares and hepatic decompensation have been observed in patients.55 ADV-associated mutations are partially cross-resistant with tenofovir. Decrease in serum HBV DNA levels observed when patients with virologic breakthrough due to ADV resistance are switched to tenofovir treatment is likely related to the higher dose of tenofovir than adefovir (300 mg versus 10 mg) used in clinical practice. The mechanism by which rtN236T confer resistance to ADV is thought to be due to indirect perturbation of the tri-phosphate binding site of the HBV pol.38, 54 Isolates of HBV with rtN236T and rtA181V changes are susceptible to entecavir in vitro.39, 48, 56 Case reports have confirmed the in vivo efficacy of lamivudine and entecavir in the suppression of adefovir resistant HBV.39, 55, 57

Tenofovir (TDF) — There was a recent report of resistance to TDF associated with HBV encoding changes in HBV polymerase at rtL180M + rtA194T (alanine to threonine substitution) + rtM204V in 2 HIV / HBV co-infected patients.58 HBV DNA was persistently detected in these 2 patients during TDF and LAM therapy but only 1 patient had an increase in serum HBV DNA. Aminotransferase levels remained normal in both patients and results of phenotypic studies are conflicting.59 Data on TDF resistance in patients who received tenofovir monotherapy are not available because tenofovir has been used predominantly in patients with HIV/HBV coinfection, in combination with lamivudine or emtricitabine.

(iii) Entecavir.

Two different ETV resistance genotypic profiles have been reported and confirmed in vitro.47 The first pattern of ETV resistance includes: rtI169T + rtL180M + rtM204V + rtM250V and the second pattern includes: rtL180M + rtT184G + rtS202I +rtM204V (Table 3). Other patterns including triple ETV-associated mutations have also been observed.60 In the absence of the LAM-associated resistance mutations, the rtM250V change causes a 9-fold increase in EC50 while the changes at rt 169 (I169T), 184 (T184A/F/G/I/L/S) or 202 (S202G/I) changes have little effect ( However, in the presence of LAM-associated resistance mutations, the ETV-associated mutations decrease susceptibility to ETV by >100-fold, particularly when 2 or more ETV-associated mutations are present.47 HBV isolates with ETV-associated mutations are sensitive to adefovir and tenofovir in vitro and in vivo data confirming the efficacy of adefovir has been reported in one patient.61

Recent data suggest that pre-existence of mutations such as rtL180M + rtM204V even when present in <0.1% of the viral population in patients who have not received lamivudine treatment may increase the risk of selection for ETV-associated mutations in patients receiving entecavir treatment.62 In addition, virologic breakthrough may occur in nucleoside-naïve patients receiving entecavir treatment due to selection of LAM-associated mutations alone. These data suggest that resistance to entecavir occurs through a 2-hit mechanism. Initially, LAM-associated mutants (rtM204V/I) are selected because they are less sensitive to ETV than wild type HBV. Virologic breakthrough usually occurs only after emergence of additional ETV-associated mutations, but rarely virologic breakthrough may occur after emergence of LAM-associated mutations alone.62

(b) Multidrug Resistance

Sequential treatment with NA monotherapy has resulted in the sequential selection of mutations conferring resistance to the initial therapy and subsequently the rescue therapy. For example, sequential resistance to lamivudine and later adefovir has been reported in patients who were switched to adefovir monotherapy for LAM-resistant HBV.60 Recent studies reported that multidrug resistance changes can be detected in patients who received sequential NA monotherapy and clonal analysis showed that in most instances, the mutations associated with both treatments reside in the same clone.56, 60 The collocation of mutations associated with resistance to different treatments on the same genome is worrisome because in vitro analysis of antiviral sensitivities revealed that replicating clones with LAM and ADV- associated mutations had >50-fold reduced susceptibility to combination of LAM and ADV indicating that combination therapy of the 2 drugs may not be effective in suppressing multidrug resistant HBV.54, 56

6. Monitoring and Treatment of Antiviral-Resistant HBV

(a) Monitoring for Virologic Response and Breakthrough

All patients receiving NA therapy for hepatitis B should be closely monitored for virologic response and breakthrough during treatment and for durability of response and viral relapse after treatment has stopped. Serum HBV DNA should be tested prior to treatment and then every 3 months during treatment. Patients with primary nonresponse should be considered for alternative treatment to facilitate clinical response and to minimize subsequent antiviral resistance. Patients with virologic breakthrough should be questioned about medication compliance. Tests for antiviral-resistant mutations should be performed whenever possible to confirm genotypic resistance, and to determine the pattern of mutations. The latter is particularly important as an increasing number of patients have been exposed to more than one anti-HBV NA.

(b) Treatment of Antiviral-Resistant HBV (Table 4)

Recommendations on treatment of patients with antiviral-resistant HBV depend on knowledge of the history of HBV treatments, virologic response to these treatments, the pattern of mutations detected at the time of virologic breakthrough, and in vitro data on antiviral activity of various HBV NAs against HBV isolates that harbor the mutations detected. Recent data suggest that initiating rescue therapy when virologic breakthrough is detected is more effective than delaying rescue therapy until viral rebound or biochemical breakthrough.63

Table 4. Strategies for treating Antiviral-Resistant HBV
Type of ResistanceRescue Therapy
  • Abbreviation: MDR, multidrug resistance.

  • *

    Limited in vivo data, available data indicate that addition of rescue therapy is less likely to result in sequential drug resistance than switching to rescue therapy.

  • #

    In vivo data lacking

Lamivudine/telbivudine resistance•Add adefovir or tenofovir
 •Switch to emtricitabine + tenofovir
 •Switch to entecavir (risk of subsequent entecavir resistance and multidrug resistance)
Adefovir resistance*•Add lamivudine or switch to emtricitabine + tenofovir
 •Add entecavir (if no prior lamivudine resistance)
Entecavir resistance*•Add adefovir or tenofovir
Multidrug resistance#•MDR to LAM + ADV: consider tenofovir + emtricitabine, tenofovir + entecavir
 •MDR to LAM + ETV: consider tenofovir or tenofovir + emtricitabine

(i) Lamivudine, Telbivudine and Other L-Nucleoside Resistance.

In vitro studies have demonstrated that adefovir, tenofovir, and entecavir have antiviral activity against LAM- and other L-nucleoside-resistant HBV mutants, but the activity of entecavir against these mutants is substantially lower than for wild-type HBV (Table 3).36 A pilot study in patients with lamivudine-resistant HBV reported that adefovir monotherapy resulted in similar rates of decrease in serum HBV DNA levels as combination therapy of lamivudine and adefovir.64 However, combination of lamivudine and adefovir is more effective in preventing subsequent adefovir resistance.57, 65 Tenofovir has also been reported in clinical studies to be effective in suppressing lamivudine-resistant HBV.66, 67 Although tenofovir is more potent than adefovir, it is best used in combination with lamivudine or emtricitabine to prevent drug resistance. Entecavir has been shown in clinical trials to be effective in suppressing lamivudine-resistant HBV but a higher dose (1.0 mg daily) should be used.68 Pre-existing LAM-resistant mutations increase the risk of entecavir resistance47, 62; therefore, entecavir is not an optimal treatment for patients with lamivudine-resistant HBV. If entecavir is used, lamivudine should be discontinued.

Based on in vitro data and the detection of rtM204I in patients with telbivudine resistance, the approach described above can be applied to patients with telbivudine-resistant HBV.

(ii) Adefovir.

In vitro studies demonstrated that lamivudine and entecavir have antiviral activity against ADV-resistant HBV mutants (Table 3).39, 54 Case studies confirmed that lamivudine is effective in suppressing serum HBV DNA levels in patients with ADV-resistant HBV.39, 54 However, the durability of response, particularly in patients with prior lamivudine resistance is unknown. Furthermore, in the latter patients, a rapid re-emergence of LAM-resistant mutations has been observed on reintroduction of lamivudine.59, 60 Case studies have reported that patients with primary nonresponse to adefovir experience further viral suppression when treatment is switched to tenofovir69 presumably due to the higher dose of tenofovir used in clinical practice. For the same reason, tenofovir may result in some degree of viral suppression in patients with ADV-resistant HBV but the efficacy is likely limited due to in vitro evidence of cross resistance. Entecavir has been reported to be efficacious in 2 patients with ADV-resistant HBV.57

(iii) Entecavir.

In vitro studies demonstrated that adefovir and tenofovir have antiviral activity against ETV-resistant HBV mutants47, 48 (Table 2) but clinical data on the efficacy of these treatments in patients with ETV-resistant HBV are not yet available.

(iv) Multidrug-Resistant HBV.

The most effective treatment of multidrug-resistant HBV is prevention through judicious use of NA therapy and avoidance of sequential NA monotherapies. Thus, patients with minimal disease and those who are unlikely to achieve sustained response (such as inactive carriers and HBeAg-positive patients in the immune tolerance phase) should not be treated with NA therapy, particularly if they are young. When possible, the most potent NA with the lowest rate of genotypic resistance should be administered and compliance reinforced. Response should be closely monitored and modification of treatment considered in patients with primary nonresponse. De novo combination therapy of lamivudine plus pegylated interferon or adefovir has been shown to be associated with lower rates of virologic breakthrough compared to lamivudine monotherapy,70–72 but resistance was not completely prevented. Studies on other de novo combination therapies are needed to determine the optimal combination of drugs and its cost-effectiveness.

7. Conclusions

Antiviral resistance and noncompliance to therapy are the most important cause of treatment failure in patients with hepatitis B. As more treatments become available, the complexity of antiviral-resistant mutations and the options for primary as well as rescue therapy increase. Therefore, there is an urgent need for standardization of (i) nomenclature on antiviral resistance, (ii) assays used in detection or confirmation of resistance, and (iii) format for reporting both in vitro and in vivo resistance data. This document represents the collaborative efforts of investigators from North America, Europe, Asia, and Australia. The group recognizes the importance of information sharing not only among investigators but also between investigators and practicing physicians such that new information generated from research can be rapidly disseminated and observations in clinical practice can be validated. To this end, the group hopes to establish a database on hepatitis B virus drug resistance that will be freely accessible and to develop means for sharing technology, clinical samples, and HBV isolates.


The authors express their appreciation for the valuable comments from other members of the Hepatitis B Virus Drug Resistance Working Group, in particular Drs. Rajen Koshy, T. Jake Liang and William Mason.