Long-term monitoring shows hepatitis B virus resistance to entecavir in nucleoside-naïve patients is rare through 5 years of therapy


  • Potential conflict of interest: The studies were performed at Bristol-Myers Squibb by employees of Bristol-Myers Squibb. Study participants have been informed of potential conflicts of interest.Drs. Wilber, Eggers, Colonno, Baldick, Tenney, Pokornowski, and Xu own stocks in Bristol-Myers Squibb.


Patients with chronic hepatitis B virus (HBV) infection who develop antiviral resistance lose benefits of therapy and may be predisposed to further resistance. Entecavir (ETV) resistance (ETVr) results from HBV reverse transcriptase substitutions at positions T184, S202, or M250, which emerge in the presence of lamivudine (LVD) resistance substitutions M204I/V ± L180M. Here, we summarize results from comprehensive resistance monitoring of patients with HBV who were continuously treated with ETV for up to 5 years. Monitoring included genotypic analysis of isolates from all patients at baseline and when HBV DNA was detectable by polymerase chain reaction (≥300 copies/mL) from Years 1 through 5. In addition, genotyping was performed on isolates from patients experiencing virologic breakthrough (≥1 log10 rise in HBV DNA). In vitro phenotypic ETV susceptibility was determined for virologic breakthrough isolates, and for HBV containing novel substitutions emerging during treatment. The results over 5 years of therapy showed that in nucleoside-naïve patients, the cumulative probability of genotypic ETVr and genotypic ETVr associated with virologic breakthrough was 1.2% and 0.8%, respectively. In contrast, a reduced barrier to resistance was observed in LVD-refractory patients, as the LVD resistance substitutions, a partial requirement for ETVr, preexist, resulting in a 5-year cumulative probability of genotypic ETVr and genotypic ETVr associated with breakthrough of 51% and 43%, respectively. Importantly, only four patients who achieved <300 copies/mL HBV DNA subsequently developed ETVr. Conclusion: Long-term monitoring showed low rates of resistance in nucleoside-naïve patients during 5 years of ETV therapy, corresponding with potent viral suppression and a high genetic barrier to resistance. These findings support ETV as a primary therapy that enables prolonged treatment with potent viral suppression and minimal resistance. (HEPATOLOGY 2009.)

Approximately 400 million people worldwide have chronic hepatitis B virus (HBV) infections, with a risk for chronic, life-threatening liver disease.1 Antiviral therapy for HBV can provide suppression of viral replication and halt disease progression.2, 3 However, therapeutic benefits are diminished with the emergence of drug-resistant virus, which occurs most often with prolonged therapy and incomplete viral suppression.4

Resistance to nucleoside/nucleotide antivirals arises through substitutions in the HBV polymerase reverse transcriptase domain (RT), that arise spontaneously through low-fidelity replication and are enriched through drug-selective pressure.2, 3 Antiviral therapies are characterized by their barrier to resistance, which includes three components: (1) the potency of the antiviral in suppressing viral replication, (2) a “genetic barrier”, i.e., the number of genetic changes required to effectively reduce drug susceptibility that results in virologic breakthrough, and (3) the replication fitness of the resistant virus. These factors act together to determine the levels of resistance which emerge during therapy. Other factors related to the particular binding site or mechanism of activity also contribute to determining the overall barrier to resistance.

Lamivudine (LVD) resistance (LVDr) arises with changes in the HBV RT tyrosine-methionine-aspartate-aspartate (YMDD) nucleotide-binding motif, in ∼20% of treated patients annually.5 Because virologic breakthrough can occur subsequent to emerging genotypic resistance, rates of resistance with virologic breakthrough are typically lower. Telbivudine resistance also arises at the YMDD motif and has been reported in the context of virologic breakthrough, at 22% and 9% over 2 years in patients who are positive and negative, respectively, for the hepatitis B e antigen (HBeAg).6, 7 YMDD motif changes reduce susceptibility to LVD or telbivudine by >100-fold.8 Adefovir (ADV) resistance (ADVr) arises through HBV RT A181 and N236 changes9 and occurs in 29% of patients who are HBeAg-negative after 5 years. ADVr in patients who are HBeAg-positive was reported in the context of virologic breakthrough and occurred in 25% of patients followed for 110-279 weeks of therapy.9 For LVD, ADV, and telbivudine, the genetic barrier to resistance in antiviral-naïve patients can be as low as a single substitution.

Several factors contribute to the high barrier to resistance with entecavir (ETV). ETV is potent, resulting in a higher proportion of patients achieving undetectable HBV DNA than those treated with LVD10, 11 or ADV.12 Marked (>70-fold) reductions in ETV susceptibility requires substitutions at residues T184, S202, or M250 in LVDr HBV with changes at M204I/V ± L180M.13, 14 Thus, the genetic barrier to ETVr involves multiple substitutions. In vitro studies demonstrated that the highest levels of phenotypic resistance, leading to virologic breakthrough, require both the M204V and L180M LVDr substitutions with at least one ETVr substitution.13, 15 Additionally, ETVr HBV exhibits impaired replication fitness.14 Thus, the finding that ETVr has been rarely observed in nucleoside-naïve patients16, 17 is likely due to a combination of a high genetic barrier, potent viral suppression, and reduced fitness of resistant viruses.

The development of viral resistance often predisposes patients to resistance for subsequent antivirals. LVDr substitutions M204V/I ± L180M result in complete functional cross resistance (>100-fold) to L-nucleoside analogs LVD, telbivudine, emtricitabine, and clevudine.8, 15, 18–21 However, nucleotide phosphonates ADV and tenofovir are not cross-resistant to M204 YMDD mutant HBV. Nevertheless, ADVr is more likely to emerge in patients with LVDr HBV, at rates of 18% and 25% after 1 and 2 years, respectively.22–24 Recently, evidence for cross-resistance of ADV and tenofovir with LVDr HBV with A181V or T substitutions was reported.25

LVDr also reduces the barrier to resistance with ETV. ETV exhibits ∼8-fold reduced potency against LVDr HBV.13, 21 Although 48 weeks of ETV therapy in LVD-refractory patients suppressed HBV DNA by a mean of 5.1 log10 copies/mL,26, 27 this potency was reduced relative to that in naïve patients.10, 11 Additionally, because LVDr M204 substitutions are a subset of the changes required for ETVr,8 the genetic barrier to ETVr is reduced with LVDr HBV relative to wild-type virus.

Previous analyses showed ETVr is rare in nucleoside-naïve patients, but increases over time in LVD-refractory patients.15, 16, 28, 29 Here, we report a comprehensive assessment of resistance in both patient populations treated for up to 5 years with ETV.


ADV, adefovir; ADVr, adefovir-resistant; AS-PCR, allele-specific, single-nucleotide polymorphism polymerase chain reaction; EC50, median effective concentration; ETV, entecavir; ETVr, entecavir-resistant, entecavir-resistance; HBV, hepatitis B virus; LVD, lamivudine; LVDr, lamivudine-resistant, lamivudine-resistance; nucleoside-naïve, nucleoside-treatment-naïve; RT, reverse transcriptase domain of the HBV polymerase; WT, wild type; YMDD, tyrosine-methionine-aspartate-aspartate.

Patients and Methods


Patients from six phase 2 and 3 clinical studies of the safety and efficacy of ETV were monitored for resistance through Year 5 (week 240). Serum samples were obtained from patients in studies ETV-02211 (www.clinicaltrials.gov identifier: NCT00035633) and ETV-02710 (NCT00035789) of nucleoside-naïve HBeAg-positive and HBeAg-negative patients with compensated liver disease, respectively, as well as in LVD-refractory patients (those with continued viremia or identified resistance while on LVD) in phase 3 study ETV-02627 (NCT00036608) and phase 2 study ETV-01426 in patients with compensated liver disease, and phase 2 study ETV-01515 in orthotopic liver transplant recipients. All patients provided written informed consent, and study protocols conformed to the 1975 Declaration of Helsinki and were approved by appropriate Institutional Review Boards.

Patients initially randomized to treatment with the approved ETV daily dosages (0.5 mg for nucleoside-naïve and 1.0 mg for LVD-refractory) were eligible for inclusion in the resistance surveillance. Patients from naïve trials ETV-022 and ETV-027, conducted in patients who were HBeAg-positive and HBeAg-negative, respectively, were treated from 1 to 2 years, depending on the clinical response at week 48.10, 11 Patients whose HBV DNA was suppressed below the limit of detection by the branched DNA assay (0.7 MEq/mL) and were either HBeAg-negative (for HBeAg-positive patients) or had normalized alanine aminotransferase (for HBeAg-negative patients) at week 48 were to discontinue treatment at week 52. Patients whose HBV DNA was less than 0.7 MEq/mL but did not lose HBeAg or normalize alanine aminotransferase were allowed to continue treatment through week 96. Patients who did not achieve either endpoint were offered alternative therapy in the rollover study ETV-901. All patients who remained on treatment were transferred to the rollover study at week 96. Most of the 555 patients who left the studies, therefore, did so as protocol-defined Responders, and 89% had HBV DNA < 300 copies/mL by polymerase chain reaction (PCR), and therefore were unlikely to develop genotypic resistance and/or virologic breakthrough.

Because the ETV-901 study was blinded and included patients rolling over from both ETV and LVD arms of other studies, they initially received the combination of 100 mg LVD and 1.0 mg ETV. Subsequently, a protocol amendment changed ETV-901 to monotherapy with 1.0 mg ETV. Results from the nucleoside-naïve resistance cohort reflect the use of 1.0 mg ETV for 147 of 149 patients in Year 3, and all patients in Years 4 and 5, rather than the 0.5 mg dosage.

Patients who received extended therapy in study ETV-901 were included in the resistance cohort if they received “continuous treatment,” with interruptions of ≤35 days between the end of dosing in the original study and the start of dosing in ETV-901. When treatment gaps exceeded 35 days, the end of dosing in the primary study was considered the last on-treatment observation, and their subsequent virologic profile was excluded from the long-term resistance survey.

Annual visits for Years 1 through 5 occurred at 48-week intervals. These timepoints were “windowed” to include annual patient visits collected between ≥42 to ≤58 weeks for Year 1, between ≥90 to ≤102 weeks for Year 2, between ≥132 to ≤156 weeks for Year 3, between ≥180 to ≤204 weeks for Year 4, and between ≥228 to ≤252 weeks for Year 5 of therapy.

All patients receiving ETV for ≥24 weeks were monitored for resistance using HBV DNA quantification and nucleotide sequence analysis. HBV DNA quantification used the Roche COBAS Amplicor PCR (version 2, lower limit of quantification 300 copies/mL; 57 IU/mL). Baseline and on-treatment isolates from patients with PCR-detectable HBV DNA (≥300 copies/mL) at the end of each yearly interval or at end of dosing within each year were sequenced. These included patients experiencing a virologic breakthrough (defined as a ≥1 log10 increase in HBV DNA from the on-treatment nadir, confirmed by two sequential measurements, or unconfirmed when it was the last on-treatment assessment in each year). When multiple samples were available at the end of the yearly interval, the sample closest to the end of each year was used (i.e., weeks 48, 96, 144, 192, and 240).


The HBV RT was PCR amplified from patient serum HBV DNA, sequenced directly, and analyzed.15, 16


Phenotypic ETV susceptibility was determined for HBV with novel emerging substitutions as well as for paired baseline and breakthrough isolates from patients experiencing a virologic breakthrough. Virus susceptibility assays used either patient RT population quasispecies or individual isolates cloned into a laboratory HBV genotype D expression plasmid. Susceptibility was determined by quantification of nucleocapsid-associated HBV DNA released from transfected HepG2 cells, in the presence of 0.2 nM to 15 μM ETV.15

Cumulative Probabilities.

The cumulative probability of resistance was calculated2 using the formula PTotal = 1 − (1 − Pyr1) × (1 − Pyr2) × (1 − Pyr3) × (1 − Pyr4) × (1 − Pyr5), where P = probabilityyr(i) = (number of patients with events at Year i)/(number of patients at risk at Year i) for i = 1, 2, 3, 4, 5. A resistance “event” was defined as genotypic ETVr or genotypic ETVr with virologic breakthrough, in a yearly interval. Patients “at risk” were those during Year i who did not develop resistance during Year (i-1). Patients who discontinued ETV treatment in Year i were assumed to be in follow-up for the entire year.


Resistance Surveillance in Nucleoside-Naïve Patients Treated with ETV.

In Year 1, 663 nucleoside-naïve patients treated with ETV were monitored for resistance (Table 1). At the end of Year 1, HBV DNA was successfully amplified and sequenced from 243 patients, including those 128 of the 663 (19%) patients with PCR-detectable HBV DNA, and a subset of the 535 patients with undetectable virus, who were analyzed because the patients' HBV DNA status was blinded at the time of analysis. ETVr substitution S202G with LVDr M204V+L180M changes were detected in a single HBeAg-negative patient at week 48 (patient #16). Despite being nucleoside-naïve, baseline virus from patient #16 harbored 34% LVDr HBV (L80L/I, M204M/I), and subsequently developed L180M, M204V, and ETVr substitution S202G at the time of virologic breakthrough (Table 3).16 Eight other nucleoside-naïve patients had baseline LVDr, detected in three patients using standard nucleotide sequencing and in the other five using a more sensitive single-nucleotide polymorphism allele-specific (AS) PCR assay (Fang et al., manuscript submitted). Two patients developed LVDr on ETV that was not detected at baseline; one breakthrough patient in Year 1 with 11 substitutions emerging, who left the study before confirmation of the changes; and one patient with LVDr emerging at the last treatment visit in Year 5. Neither exhibited phenotypic resistance seen for patients with ETVr (see below). Aside from patient #16, none of the nucleoside-naïve patients with LVDr at baseline subsequently developed ETVr.

Table 1. Nucleoside-Naïve Patients in ETV Resistance Cohort
Study (Patient Population)*Treated and Monitored Patients
Year 1Year 2Year 3Year 4Year 5
  • *

    Included time points in rollover study ETV-901 as outlined in Patients and Methods.

  • Time point nearest to each year end window or the end-of-dosing.

ETV-022 (HBeAg Positive)345234141115103
ETV-027 (HBeAg Negative)31844865
All Treated and Monitored663278149121108
No. < 300 copies/mL HBV DNA (%)535 (81)232 (83)132 (89)109 (90)100 (93)
Table 3. Resistance Genotypes in Patients with ETVr
inline image

Fourteen patients treated with ETV experienced virologic breakthrough during Year 1, six with HBeAg-positive HBV and eight with HBeAg-negative HBV. The HBV DNA rise was confirmed in sequential visits in only two patients, one of whom was ETVr patient #16.

Among the 384 patients discontinuing treatment at the end of Year 1 (and who were not included in the long-term resistance monitoring), 92% (352) had HBV DNA <300 copies/mL.

A total of 278 nucleoside-naïve patients were monitored in Year 2, with the majority (232, 83%) achieving HBV DNA <300 copies/mL. Genotyping was successful in 44 of the 46 patients with detectable HBV DNA. ETVr (M204V+L180M and S202G) was detected in one patient who was HBeAg-positive at week 84 (patient #23); this patient discontinued treatment at week 97 with 3.7 log10 copies/mL HBV DNA, without having experienced a virologic breakthrough. All three resistance substitutions appeared simultaneously and sequencing of cloned isolates showed that all three resistance substitutions were genetically linked and that viruses with subsets of resistant substitutions were not detected. After >7 weeks without therapy, this patient's HBV DNA reached 6.7 log10 copies/mL and ETV treatment was reinitiated; however, HBV DNA levels were not suppressed.

Eight patients in Year 2 experienced unconfirmed virologic breakthroughs. One had LVDr substitutions only at the time of breakthrough; however, the patient's baseline isolate also harbored low levels (0.13%) of LVDr HBV by the AS-PCR method. Resistance at LVDr or ETVr residues was not found in any other patient.

Of the 129 nucleoside-naïve patients discontinuing in Year 2, 107 (83%) had <300 copies/mL HBV DNA. A total of 149 patients continued therapy into Year 3 and were monitored for resistance. At the end of Year 3, 17 patients (11%) had HBV DNA ≥ 300 copies/mL and genotyping was successful on 15 patient isolates. Two patients exhibited confirmed virologic breakthrough in Year 3. One patient (#44) had genetically linked M204V+L180M and S202G resistance emerge at week 139 and experienced virologic breakthrough at week 148 and the addition of a T184T/I change. This patient had rolled over into study ETV-901 at week 100 and completed 16 weeks of combination treatment before resistance emerged with breakthrough. No ETVr or LVDr was found in any other patient.

Of the 28 patients who discontinued without resistance prior to therapy in Year 4, 25 (89%) had undetectable HBV DNA (<300 copies/mL). A total of 121 nucleoside-naïve ETV patients received therapy and were monitored for resistance in Year 4. Twelve (10%) had HBV DNA ≥ 300 copies/mL and genotyping was successful in 10 of these patients. One patient experienced a virologic breakthrough. Analysis of these patients failed to detect resistance.

Of the 14 patients who discontinued without resistance prior to therapy in Year 5, 12 (86%) had undetectable HBV DNA (<300 copies/mL). Among 108 nucleoside-naïve ETV patients monitored in Year 5, eight (7%) had HBV DNA ≥ 300 copies/mL and genotyping was successful for seven patients. No ETV-resistance was found in these patients, including two who experienced virologic breakthrough.

The overall results in nucleoside-naïve patients are summarized in Fig. 1. Through 5 years of ETV therapy, the cumulative probability of developing genotypic ETVr or genotypic ETVr accompanied by a virologic breakthrough was 1.2% and 0.8%, respectively.

Figure 1.

Cumulative probabilities of ETVr in nucleoside-naïve patients. Bars indicate the cumulative probabilities of genotypic ETVr (open) and genotypic ETVr with virologic breakthrough (filled) that occurred over the course of 5 years.

Cell culture susceptibility testing also showed that 22 of the 27 breakthroughs experienced by nucleoside-naïve ETV patients through Year 5 were unrelated to resistance, because resistance substitutions were not detected and ETV susceptibility was relatively unchanged from the wild type (Fig. 2). Isolates from the two patients with virologic breakthroughs associated with ETVr substitutions (patients #16 and #44) showed decreased ETV susceptibility of 2790-fold and 32-fold, respectively. LVDr isolates were only 2.1-fold to 26.2-fold less susceptible than the wild type (Fig. 2). This correlates with the finding that the two patients with LVDr who were further treated with ETV achieved HBV DNA <300 copies/mL on continued ETV.

Figure 2.

Phenotypic analysis of isolates from nucleoside-naïve patients experiencing virologic breakthrough. Each diamond (◊) represents the ETV susceptibility of the population isolate at baseline and is paired with a circle (○) representing the susceptibility of the virologic breakthrough population isolate for the same patient. Shapes are filled according to resistance genotype, as indicated. The wild-type (WT)/LVDr and LVDr/ETVr indicate mixtures. Susceptibility levels are expressed relative to the mean of 10 WT HBV populations with no substitutions (EC50 = 1.5nM), tested in parallel. Breakthrough isolates with ETVr from patients #16 and #44 are the filled circles in Year 1 and Year 3, respectively.

LVD-Refractory Patients.

The LVD-refractory resistance cohort included 187 patients with documented resistance or recurrent/persistently detectable HBV DNA (>300 copies/mL) while receiving LVD (Table 2). Baseline sequence evidence of genotypic LVDr was detected in 84.5% of patients. As reported, 5% of these patients also harbored ETVr substitutions at T184, S202, or M250 at study entry.15

Table 2. LVD-Refractory Patients in ETV Resistance Cohorts
Study*Treated and Monitored Patients
Year 1Year 2Year 3Year 4Year 5
  • *

    Included time points in rollover study ETV-901 as outlined in Patients and Methods.

  • Time point nearest to each year-end window or the end-of-dosing.

All Treated and Monitored187146805233
No. < 300 copies/mL HBV DNA (%)38 (20)38 (26)27 (34)24 (46)20 (61)

Emerging ETVr at T184, S202, and/or M250 was observed by nucleotide sequencing in 11, 12, 16, 6, and 2 patients during Years 1 through 5 of ETV, respectively. Virologic breakthrough with ETVr, including those resistant at baseline, occurred in 2, 14, 13, 9, and 1 patients in Years 1 through 5, respectively. Importantly, through Year 5, only three LVD-refractory patients achieving <300 copies/mL HBV DNA on ETV subsequently developed ETVr, and only two of them experienced a subsequent virologic breakthrough.

The presence of ETVr did not always result in virologic breakthrough because only 68% (39 of 57) of ETVr patients experienced a virologic breakthrough through Year 5. Various substitutions were found (Table 3). Substitutions at residue I169 emerged in 13 of the 57 (22%) patients with ETVr. All the I169 changes occurred subsequent to or simultaneously with primary ETVr at T184, S202 or M250. Only 9 of 39 (23%) breakthrough isolates had I169 changes. Three patients with I169 changes did not experience breakthrough, and one breakthrough patient developed an I169 substitution after breakthrough. This, along with the finding that I169 substitutions do not consistently alter the phenotypic susceptibility to ETV,8 suggests that the I169 change is an adaptive or accessory substitution and not a primary ETVr change.

In general, only a subset of ETVr changes was found in patients experiencing virologic breakthrough, including T184A/C/F/G/L/M, S202G, or M250V. Other substitutions at T184, S202, or M250 were found in breakthrough isolates, but only as part of a combination with these restricted changes. Occasionally, further therapy resulted in an evolution of the ETVr changes to other residues (unpublished observations). Patients with T184I/S, S202C, or M250I/L tended not to experience virologic breakthrough unless these substitutions shifted to a more highly resistant substitution listed above. Often there was a delay between the appearance of genotypic ETVr and virologic breakthrough (unpublished observations), consistent with observations that ETVr variants are replication impaired.8, 14 Through 5 years of therapy with 1.0 mg ETV in LVD-refractory patients, the cumulative probabilities of genotypic ETVr substitutions and virologic breakthrough with ETVr increased to 51% and 43%, respectively (Fig. 3).

Figure 3.

Cumulative probabilities of ETVr in LVD-refractory patients. Bars indicate the cumulative probabilities of genotypic ETVr (open) or genotypic ETVr accompanied by a virologic breakthrough (filled) that occurred over the course of 5 years.

The phenotypic ETV susceptibility of virologic breakthrough isolates from the LVD-refractory patients is summarized in Fig. 4. In contrast to the results with most nucleoside-naïve breakthrough patients (Fig. 2), breakthrough isolates from LVD-refractory patients consisted predominantly of ETVr genotypes and exhibited substantial reductions in ETV susceptibility relative to the wild type (285-fold median [range = 29-4464]). A range of susceptibilities were observed due to differences in the proportions of resistant viruses in the quasispecies and from the different ETVr changes present; however, 82% (32 of 39) had median effective concentration (EC50) for ETV greater than 100 nM, which is >66.7 times the EC50 of the wild type.

Figure 4.

Phenotypic analysis of isolates from LVD-refractory patients experiencing virologic breakthrough. Each diamond (◊) represents the ETV susceptibility of the population isolate at baseline and is paired with a circle representing the susceptibility of the breakthrough population isolate from the same patient. Shapes are filled according to resistance genotype, as indicated. The wild-type (WT)/LVDr and LVDr/ETVr indicate mixtures. Susceptibility levels are expressed relative to the mean of 10 WT HBV populations with no substitutions (EC50 = 1.5 nM), tested in parallel. The WT baseline isolate results from one Year 4 patient are shown without their on-treatment breakthrough isolate, which was unable to be amplified. Another Year 3 patient whose on-treatment breakthrough isolate was also unable to be amplified, and their baseline (WT) isolate, are both not shown.

No changes other than those at T184, S202, or M250, in association with LVDr substitutions at M204I/V ± L180M, emerged as candidates for primary ETVr substitutions during the 5-year resistance monitoring. We identified 92 other novel substitutions at 76 positions in nucleoside-naïve patients, and 22 novel changes at 20 residues in LVD-refractory patients without primary ETVr; however, none of these occurred in >3 patients (<2%), and none correlated with virologic responses or reproducibly reduced phenotypic ETV susceptibility, similar to results after 2 years of ETV.15, 16 One variant contained an F88Y substitution in a LVDr M204I/V+L180M background (Fig. 4, Year 4, second patient) and exhibited elevated ETV EC50 levels; however, this change did not reproducibly confer resistance in other cloned isolates from the same patient or when the change was engineered into laboratory background viruses. Altogether, these analyses suggested that various novel changes emerging on ETV resulted from random genetic drift rather than resistance selection.

A181 Substitutions.

A181 substitutions are associated with both LVD-resistance and ADV-resistance, as well as resistance to LVD and ADV combination therapy.25, 30 Fifteen patients in the long-term resistance cohort had A181 substitutions at some time during ETV therapy, two from study ETV-022, 12 from ETV-026, and one from ETV-015. These include changes of A181 to C, G, S, T, or V. Seven of the 15 patients had baseline A181 substitutions. Baseline A181 substitutions disappeared during ETV therapy in four patients, two others achieved and maintained undetectable HBV DNA, and the A181 substitution was maintained in one patient (data not shown). Eight other patients had A181 changes emerge on ETV. These substitutions disappeared on continued therapy in three patients, three others achieved and maintained undetectable HBV DNA, and the substitutions remained in two patients. As reported,15, 16 phenotyping of the A181 substitutions emerging in patients without ETVr showed no reduced susceptibility in the wild-type or LVDr HBV. All these results argue against a role for A181 substitutions in ETV susceptibility or resistance.


This report details the results of long-term HBV resistance monitoring over 5 years of ETV treatment. Our study methods closely adhered to recent guidelines2, 31 and involved genotypic analysis of patients with detectable HBV DNA by a sensitive PCR method (limit of quantification 300 copies/mL) and phenotypic susceptibility testing of isolates from all patients experiencing virologic breakthrough or novel emerging amino acid substitutions.

The most compelling finding was that a high barrier to resistance was maintained through 5 years of ETV therapy in nucleoside-naïve patients, where the cumulative probability of genotypic ETVr and virologic breakthrough due to ETVr was 1.2% and 0.8%, respectively. This barrier to resistance likely results from (1) potent suppression of viral replication, because 93% of patients achieved undetectable HBV DNA (<300 copies/mL) during this 5-year period, (2) a high genetic barrier to resistance, and (3) impaired replication of the ETVr variants. In the end, through 5 years, ETVr substitutions were detected in only three nucleoside-naïve patients (3 of 663), with one patient having preexisting LVDr at baseline.

There are particular strengths of our resistance study, such as the long-term observation period, the global enrollment of a large number patients in whom all relevant HBV genotypes were represented, and the practice of genotyping all patients with PCR-detectable virus, as well as the phenotyping of virologic breakthrough isolates irrespective of recognizable genotypic resistance substitutions. Study limitations include the loss of patients over time due to protocol design, which mandated treatment discontinuation when patients met certain response definition.10, 11 Nevertheless, a high percentage of patients (89%) had undetectable HBV DNA (<300 copies/mL) at the time of treatment discontinuation. Observation of those patients who continued on treatment despite not achieving these patient management endpoints suggests that the patients who discontinued would have had a low potential for developing resistance had they continued therapy. Through 5 years of surveillance, one nucleoside-naïve patient developed genotypic ETVr after achieving HBV DNA <300 copies/mL, and only three LVD-refractory patients did.

Another limitation of our resistance surveillance was introduced by the increase in ETV dosage from 0.5 to 1.0 mg daily when the patients entered the ETV-901 rollover study. Importantly, the results of a parallel survey of surveillance conducted in Japan in which nucleoside-naïve patients received the 0.5 mg dosage of ETV for 3 years yielded only one of 66 patients who developed genotypic resistance (1.7%).17

The three nucleoside-naïve study patients, with emerging ETVr, all developed M204V+L180M and S202G substitutions. Two of these patients with wild-type virus at baseline developed M204V+L180M and S202G simultaneously, and these resistance substitutions were linked genetically, in that individual clones did not have different subsets of changes. This indicates that ETVr did not emerge in a stepwise manner; which would have resulted in a subset of the viral population with LVDr changes only, accompanied by a proportion with additional ETVr substitutions. Simultaneous emergence of all three resistance substitutions has been noted in other reports of ETVr.32 The requirement to simultaneously develop multiple resistance substitutions in the nucleoside-naïve population most likely contributes to the high genetic barrier to ETVr.

Although LVDr substitutions reduce ETV susceptibility, our results do not suggest that ETV selects for LVDr substitutions de novo. Instead ETV probably maintains or enriches LVDr variants that are already present. In all but two patients, LVDr that was found during ETV treatment could be detected at baseline. The observation that LVDr was detected less frequently over time in nucleoside-naïve patients receiving ETV, with none emerging in Years 3, 4, and 5, also suggests that ETV did not cause de novo emergence of LVDr but enriched preexisting variants. Variants with ETVr may also have been selected from rare preexisting viruses as the incidence of ETVr also diminished over time, with none found in Years 4 and 5. The simultaneous emergence of three genetically-linked resistance substitutions in the two nucleoside-naïve patients infected with the wild type who developed ETVr is also consistent with the hypothesis that preexisting variants were archived at minute levels prior to therapy. Importantly, these studies included a subset of patients with preexisting LVDr changes, as has been reported for nucleoside-naïve patients. Nevertheless, the incidence of ETVr was low, suggesting that the potency of ETV was sufficient to overcome these rare occurrences.

In contrast to nucleoside-naïve patients, genotypic ETVr emerged in LVD-refractory patients with a cumulative probability of 6%, 15%, 36%, 47%, and 51% in Years 1 through 5, respectively, and the cumulative probability of virologic breakthroughs due to ETVr in 1%, 11%, 27%, 41%, and 43% over the same time frame. These results are consistent with a reduction in potency of ETV in LVD-refractory patients compared with nucleoside-naïve patients, as well as a diminished genetic barrier because patients with LVDr already have a subset of the substitutions required for ETVr.

Another important finding from this surveillance was that LVD therapy can result in selection of ETVr substitutions prior to ETV therapy. The fact that the ETVr variants were most frequently found as a subpopulation of the LVDr quasispecies suggests that they do not impart a significant replication advantage during LVD treatment or that impaired replication prevents their dominating the LVDr majority.

There were various substitutions at positions T184, S202, and M250 in the 60 patients with ETVr identified in these studies (57 in LVD-refractory and three in nucleoside-naïve patients). At breakthrough or the end of dosing, 21 had S202, 21 had T184, and six had M250 substitutions. Twelve others had substitutions at multiple ETVr positions, with the most common (nine of 12) being combinations of T184 and S202 substitutions. These population sequencing results do not indicate genetic linkage of these changes, only that virus(es) with the changes are present in the patient quasispecies.

Virus from the 41 ETVr virologic breakthrough patients all contained the LVDr substitutions M204V+L180M, either alone or mixed with M204I±L180M. Substitutions at L80 and V173 were present in some patients. In the patients with virologic breakthrough, a single ETVr substitution at either residue T184 (n = 13), S202 (n = 17), or M250 (n = 1) was found in 31 of 41 isolates, whereas 10 had mixtures of multiple ETVr substitutions at residues T184, S202, or M250 (Table 3). Other reports of patients with ETVr have all included the M204V+L180M LVDr substitutions,8, 17, 33, 34 suggesting that in the overwhelming majority of patients with ETVr, at least three substitutions were required for ETVr, two associated with LVDr (M204I/V+L180M) and another at T184, S202, or M250. In this survey, all patients with virologic breakthrough with ETVr had more than two resistance substitutions.

Only three patients had evidence of ETVr changes in the presence of the M204I change without M204V, and all but one of these had additional resistance substitutions at L80, V84, I169, L180, and/or A181 substitutions. The patient with the M204I substitution only with ETVr achieved undetectable HBV DNA (<300 copies/mL) on ETV. Consistent with these findings is that ETVr changes combined with only the M204I LVDr substitutions displayed lower levels of phenotypic ETV resistance.13

Not all patients with genotypic ETVr substitutions experienced breakthrough. ETVr breakthrough isolates showed varied levels of phenotypic susceptibility, reflecting the proportion of resistant virus in the quasispecies and the particular resistance substitution.13, 14 For example, ETVr substitutions M250I/L or T184I in a M204I LVDr background, or T184S and S202C changes in a M204V+L180M backbone, do not lead to breakthrough without additional primary ETVr changes.13-15

The emergence of ADVr and ETVr both occur more frequently in patients with LVDr than in nucleoside- naïve patients.22-24 Combination therapy is expected to be useful in managing patients failing prior therapy with resistance.2, 3 Potential combinations should include agents with nonoverlapping resistance profiles to create a barrier to resistance that is the sum of both therapies. ETV displays a unique cross-resistance pattern, with antiviral activity against both LVDr and ADVr HBV.21, 35-38 Conversely ETVr HBV is susceptible to ADV and tenofovir.8, 35 This profile suggests a role for ETV as a cornerstone of HBV therapy.

In conclusion, long-term monitoring showed low rates of resistance in nucleoside-naïve patients during 5 years of ETV therapy, corresponding with potent viral suppression, a high genetic barrier to resistance, and impaired fitness of ETVr viruses. These findings support the selection of ETV as a primary therapy that enables prolonged treatment with potent viral suppression and minimal resistance. Despite diminished susceptibility of LVDr virus to ETV and a reduced genetic barrier compared to wild-type virus, ETV exhibits potent antiviral activity and a unique resistance profile that merits investigation as a component of combination antiviral therapy.


We thank Patricia Poundstone and Bernadette Kienzle for skilled and dedicated nucleotide sequencing. We also acknowledge Drs. Bruce Kreter, Ricardo Tamez, Hong Tang, David Butcher, Carol Davis, and Esther Race for helpful comments on the manuscript and Dr. Ulysses Diva for patient information.