Assessment of current criteria for primary nonresponse in chronic hepatitis B patients receiving entecavir therapy

Authors

  • Young-Joo Yang,

    1. Department of Internal Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea
    Search for more papers by this author
  • Ju Hyun Shim,

    Corresponding author
    1. Department of Gastroenterology, Asan Liver Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea
    • Address reprint requests to: Ju Hyun Shim, M.D., Asan Medical Center, University of Ulsan College of Medicine, 86 Asanbyeongwon-gil, Songpa-gu, Seoul, 138-736, Korea. E-mail: s5854@medimail.co.kr; fax: +82-2-485-5782.

    Search for more papers by this author
  • Kang Mo Kim,

    1. Department of Gastroenterology, Asan Liver Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea
    Search for more papers by this author
  • Young-Suk Lim,

    1. Department of Gastroenterology, Asan Liver Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea
    Search for more papers by this author
  • Han Chu Lee

    1. Department of Gastroenterology, Asan Liver Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

Abstract

A primary nonresponse to oral drugs against hepatitis B virus (HBV) is a generally accepted criterion for interrupting treatment. We investigated whether the concept of primary nonresponse suggested by current American (AASLD) and European (EASL) guidelines is appropriate for treatment with entecavir (ETV). The study included 1,254 treatment-naïve patients who had pretreatment HBV DNA levels of >2,000 IU/mL and received ETV 0.5 mg/day for over 6 months. “Primary nonresponse” was defined as a <2 log drop in HBV DNA after 6 months of therapy by AASLD and as a <1 log drop after 3 months by EASL. The cumulative probability of virological response (VR; HBV DNA of <15 IU/mL) was compared in patients with and without primary nonresponse. Median time to achieve VR was significantly shorter in primary responders by AASLD than nonresponders (12 versus 24 months; P = 0.004), but the cumulative probability of achieving a VR at 54 months was similar in the two groups (95.8% versus 100%). Time to achieve a VR and cumulative probability of VR over time did not differ between primary responders and nonresponders by EASL. On-treatment virological breakthrough occurred in 18 patients with a cumulative rate of 5.6% at 72 months. ETV resistance was detected in 13 of these 18 patients (72.2%), who were all classified as primary responder according to both guidelines. Conclusion: Long-term ETV therapy generally leads to a VR in treatment-naïve patients, although the time to achieve it is delayed in primary nonresponders. The current recommendation to change therapy in primary nonresponders needs to be modified to reflect drug differences in antiviral potency and resistance risk. (Hepatology 2014;59:1303-1310)

Abbreviations
AASLD

American Association for the Study of Liver Diseases

ALT

alanine aminotransferase

APASL

Asian-Pacific Association for the Study of the Liver

CHB

chronic hepatitis B

EASL

European Association for the Study of the Liver

ETV

entecavir

HBeAg

hepatitis B e antigen

HBV

hepatitis B virus

HCC

hepatocellular carcinoma

NA

nucleoside/nucleotide analog

RFMP

restriction fragment mass polymorphism

TDF

tenofovir

VR

virological response

The presence of hepatitis B e antigen (HBeAg) is a well-established factor linked to progression of hepatitis B virus (HBV)-related liver disease. In addition, a large community-based cohort study in Taiwan has revealed that chronic elevation of hepatitis B viral replication carries increased risk of liver cirrhosis, hepatocellular carcinoma (HCC), and associated liver-related mortality.[1-3] The ultimate goal of treatment of chronic hepatitis B (CHB) is to suppress active viral replication and so minimize liver damage. Therefore, it is particularly important to achieve durable viral suppression, defined as reduction of serum HBV DNA to undetectable levels, especially after nucleoside/nucleotide analog (NA) therapy.[4-6]

Current guidelines on the management of CHB depend in part on the timepoints of the on-therapy measurements. They also depend on the extent of the decrease in HBV DNA level stipulated in the standardized definitions of categories of interim virological response (VR) to antiviral for deciding on an adaptation of treatment.[7-9] Given the substantial evidence that early on-treatment HBV DNA responses can strongly predict long-term virological outcomes, modification of the therapeutic regimen at early times is widely recommended not only for primary nonresponders but also for partial virological responders. However, the publications supporting this strategy are mostly based on data for less potent NAs such as lamivudine and telbivudine.[7, 9-14] Indeed, long-term therapy with newer and more potent antivirals (i.e., entecavir [ETV] and tenofovir [TDF]) can lead to virological remission in the vast majority of CHB patients, regardless of basal HBV DNA levels and HBeAg status.[15-17] A recent ETV study has suggested that, in contrast to the traditional guidelines, treatment adjustment is not essential even when patients have a partial VR at week 48, especially if their viral load at week 48 is low.[18]

Given this situation, the aim of this study of a large single-center cohort of NA-naïve CHB patients treated with ETV monotherapy was to assess whether the definitions of primary nonresponses used to guide the CHB treatment algorithm suggested by the current universal guidelines are optimal or need to be refined on the basis of newer data.

Patients and Methods

Study Population

We retrospectively reviewed data from 1,362 consecutive patients with well-preserved hepatic function (i.e., Child-Pugh class A) in whom hepatitis B surface antigen persisted for at least 6 months before the initiation of treatment; all were initially treated with ETV monotherapy for CHB or associated cirrhosis between January 2005 and December 2011 at Asan Liver Center (Fig. 1). We excluded any patient who underwent liver transplantation or suffered from HCC prior to the ETV treatment. Of the 1,362 patients, 1,298 had an HBV DNA level at the start of ETV treatment of at least 2,000 IU/mL, which is the viral load indicating antiviral treatment according to the more stringent European guidelines for both hepatitis B e antigen (HBeAg)-positive and HBeAg-negative CHB.[7] The ETV treatment was continued for more than 6 months. Adherence to treatment was assessed from the number of pills left from the previous refill, as well as by taking individual medication histories at each clinic visit. If a patient was not fully adherent to the ETV medication, patient education and counseling on the importance of medication adherence was always provided. Forty-four noncompliant patients were also excluded from the study due to their potential biasing effect on virological outcomes,[19, 20] and a total of 1,254 patients were thus eligible. The mean rate of medication adherence measured as the percent of days in which a patient took the prescribed ETV during the study period[21, 22] was 61.3% (standard deviation [SD], 17.8%) in these noncompliant patients. In most cases drug compliance was poor in the first year of ETV therapy and improved thereafter as a result of intensive patient education on medication at clinic visits. The pretreatment characteristics of noncompliant patients are summarized in Supporting Table 1; mean age was 44.3 years (SD, 10.4 years), 84.1% were male, and 29.5% had cirrhotic livers. Thirty-two patients (72.7%) were positive for HBeAg, and mean levels of serum HBV DNA and alanine aminotransferase (ALT) were 7.3 log10 IU/mL (SD, 1.5 log10 IU/mL) and 179.1 IU/L (SD, 150.3 IU/L), respectively. Primary VR by American Association for the Study of Liver Diseases (AASLD; 81.8% versus 98.7%; P < 0.001), although not by European Association for the Study of the Liver (EASL; 97.5% versus 90.0%; P = 0.41), was less common among these patients than among the compliant patients, but the final rate of HBV DNA negativity (<15 IU/mL) did not differ between the two groups (81.8% versus 86.4%; P = 0.38). No ETV resistance was detected in the 44 noncompliant patients. This study was approved by the Institutional Review Board of our institution.

Figure 1.

Patient selection algorithm according to the AASLD and EASL guidelines.

Table 1. Baseline Characteristics and On-Treatment Characteristics of the Study Population
  AASLD-Based Criteria (n = 1,254)EASL-Based Criteria (n = 1,143)
 All Patients (n = 1,254)Primary Responders (n = 1,238)Primary Nonresponders (n = 16)P ValuePrimary Responders (n = 1,129)Primary non-Responders (n = 14)P Value
  1. Data are presented as mean ± SD or n (%) unless stated otherwise.

  2. a

    Defined as achieving serum HBV DNA to undetectable level (<15 IU/mL by real-time PCR), irrespective of HBeAg clearance in HBeAg-positive patients, during the on-treatment follow-up period

  3. b

    This difference was no longer significant after further adjustment for covariates using multivariate logistic regression analysis by the backward elimination method (odds ratio, 1.45; 95% confidence interval, 0.11-0.97; P = 0.73).

  4. c

    This disparity was no longer significant after further adjustment for covariates using multivariate logistic regression analysis by the backward elimination method (odds ratio, 1.01; 95% confidence interval, 0.99-1.01; P = 0.18).

  5. d

    This disparity was no longer significant after further adjustment for covariates using multivariate logistic regression analysis by the backward elimination method (odds ratio, 3.19; 95% confidence interval, 0.99-2.26; P = 0.35).

Age47.5 ± 10.247.4 ± 10.248.8 ± 11.40.6147.1 ± 10.251.0 ± 7.50.16
Male786 (62.7)776 (62.7)10 (62.5)0.99707 (62.6)8 (57.1)0.67
ALT (IU/L)191.8 ± 274.3191.8 ± 274.3191.8 ± 283.50.13197.7 ± 279.586.4 ± 78.80.006c
Baseline HBV DNA (log10 IU/mL)6.9 ± 1.46.9 ± 1.45.7 ± 1.7<0.001b6.9 ± 1.46.0 ± 1.40.01d
Presence of cirrhosis524 (41.8)517 (41.8)7 (43.8)0.87462 (40.9)9 (64.3)0.08
HBeAg-positive693 (55.3)683 (55.2)10 (62.5)0.56633 (56.1)8 (57.1)0.94
HBeAg seroconversion or loss200 (28.9)198 (28.9)2 (20.0)0.73185 (29.2)1 (12.5)0.50
Median duration of treatment, months (interquartile range)24.0 (18.0-42.0)24.0 (18.0-42.0)27 (13.5-40.5)0.7830.0 (18.0-42.0)21.0 (18.0-25.5)0.08
Virological responsea1,084 (86.4)1,073 (86.7)11 (68.8)0.05983 (87.1)12 (85.7)0.70

Method for Response Monitoring

All patients were regularly followed at intervals of 3 or 6 months by routine biochemical assessment such as the standard liver panel including serum ALT using standard laboratory procedures, serologic marker analyses including HBeAg, and anti-HBe employing commercially available enzyme immunoassays (Abbott Laboratories, Chicago, IL), and virological measurements such as quantification of serum HBV DNA using a commercially available real-time polymerase chain reaction (PCR) assay (Abbott Laboratories) with a linear dynamic detection range of 1.5 × 101 to 1 × 109 IU/mL with no change of assay methods over the study period. Genotype analysis for ETV resistance by restriction fragment mass polymorphism (RFMP) analysis with sufficient analytical sensitivity to detect genetic variants at viral loads as low as 100 IU/mL as described previously,[23] was performed on the subset of patients showing virological breakthrough, defined as a confirmed increment of HBV DNA level of >1 log compared to the lowest value during treatment.[9] The sequences of forward and reverse primers used in the PCR to amplify viral DNA for RFMP genotyping are included in Supporting Table 2, and described in our previous article.[24]

Table 2. Host and Viral Factors Associated With VRa in ETV-Treated Patients
 Univariate AnalysisMultivariate Analysis
VariableHR (95% CI)P ValueHR (95% CI)P Value
  1. a

    Defined as reducing serum HBV DNA to an undetectable level (<15 IU/mL), irrespective of HBeAg clearance in HBeAg-positive patients.

Age1.01 (1.00-1.02)<0.001 
Female1.18 (1.04-1.34)0.008 
Liver cirrhosis1.64 (1.45-1.85)<0.0011.14 (1.01-1.30)0.042
HBeAg positivity0.40 (0.35-0.46)<0.0010.57 (0.49-0.65)<0.001
Baseline HBV DNA (log10 IU/mL)0.71 (0.68-0.74)<0.0010.76 (0.72-0.80)<0.001
Baseline ALT (IU/L)1.00 (1.00-1.00)0.67 
Primary VR by AASLD1.96 (1.08-3.55)0.0273.03 (1.66-5.53)<0.001
Primary VR by EASL1.24 (0.70-2.20)0.46 

Predefined Virological Responses

The two global consensus guidelines define the VRs used to assess the need for early therapeutic adaptation during NA treatment as follows: “primary nonresponse” is defined as <2 log decrease in serum HBV DNA level from baseline after at least 6 months of NA therapy according to the AASLD guidelines; however, according to the guidelines of the EASL it is defined as <1 log decrease at 3 months of therapy. In the latter case, the “primary nonresponse” is considered to represent “primary treatment failure” according to the proposal of the Asian-Pacific Association for the Study of the Liver (APASL).[7-9]

Statistical Analysis

The primary endpoint of this study was a VR, defined by reducing serum HBV DNA to an undetectable level (<15 IU/mL), irrespective of HBeAg clearance in HBeAg-positive patients. Secondary outcomes were virological breakthrough and on-treatment emergence of associated ETV resistance mutations. Tertiary endpoints were development of HCC and liver-related death. Cumulative probabilities of VR were compared by Kaplan-Meier analysis and subsequent log-rank test. Independent factors predicting a long-term VR and emergence of ETV resistance were identified using a multivariate Cox proportional hazard model. P < 0.05 was considered statistically significant.

Results

Baseline and On-Treatment Characteristics

The baseline characteristics of the study populations classified according to the two current guidelines are shown in Table 1. Of the 1,254 subjects in this study, 786 (62.7%) were males and the mean age was 47.5 years (SD, 10.2 years). Liver cirrhosis, which was diagnosed by imaging methods such as computed tomography or ultrasonography in all but seven patients, in whom it was diagnosed by histology, was present in 524 patients (41.8%), and 55.3% of the patients were positive for HBeAg. Mean baseline levels of serum HBV DNA and ALT were 6.9 log IU/mL (SD, 1.4 log10 IU/mL) and 191.8 IU/L (SD, 274.3 IU/L), respectively.

In terms of on-treatment factors (Table 1), virological evaluation at 6 months was possible in all patients, but measurements of viral load at 3 months were available in only 1,143 patients (91.1%). In univariate analysis, pretreatment serum HBV DNA levels were higher in patients achieving primary VR than in those not achieving VR as defined by either AASLD (P < 0.001) or EASL (P = 0.01) guidelines. Pretreatment serum ALT levels were higher in patients achieving primary VR than in those not achieving VR as defined by the EASL guidelines (P = 0.006). However, after adjustment for covariates by multivariate logistic regression analysis, these differences were no longer significant in either the AASLD or EASL models. Other baseline clinical and laboratory factors also did not differ according to primary VR as defined by AASLD or EASL.

Virological Outcomes According to Primary Responses Defined by the Two Guidelines

A total of 1,084 patients (86.4%), consisting of 550 (50.7%) in the HBeAg-positive group and 534 (49.3%) in the HBeAg-negative group, achieved VR during a median 30 months of follow-up (range, 6 to 72 months). For HBeAg-positive patients the cumulative probability of HBeAg seroconversion or loss was 47.6% at 72 months (Fig. 2). In the Kaplan-Meier analysis, the cumulative probabilities of achieving an ultimate VR at 54 months were 95.8% and 100% for primary responders and nonresponders, respectively, by AASLD, although the median time to achieve a VR was 12 months in the primary responders and 24 months in the primary nonresponders (P = 0.004; Fig. 3A). On the other hand, by EASL there was no significant difference in the median time to achieve a VR between primary responders and nonresponders, and also the cumulative probabilities (93.3% and 100%, respectively) at 30 months did not differ significantly (P = 0.35; Fig. 3B).

Figure 2.

Kaplan-Meier curve showing the cumulative rate of HBeAg loss during ETV therapy in HBeAg-positive patients.

Figure 3.

Kaplan-Meier analysis of the cumulative rates of achieving a VR in the primary responders and nonresponders defined by (A) AASLD and (B) EASL.

Impact of Primary Responses on Virological Outcomes

After adjustment for covariates significant in the univariate analyses (age, female gender, cirrhosis, HBeAg positivity, pretreatment HBV DNA levels, and primary response by AASLD), primary response at 6 months by AASLD (hazard ratio [HR], 3.03; 95% confidence interval [CI], 1.66-5.53; P < 0.001; Table 2), was an independent predictor of VR, as were cirrhosis (HR, 1.14; 95% CI, 1.01-1.30; P = 0.042), HBeAg positivity (HR, 0.57; 95% CI, 0.49-0.65; P < 0.001), and baseline HBV DNA levels (HR, 0.76; 95% CI, 0.72-0.80; P < 0.001). However, primary response at 3 months by EASL was not an independent predictor of VR.

Impact of Primary Responses on Resistance Mutations

During the study period, virological breakthrough occurred in 18 patients despite their good adherence to medication: one was designated as a primary nonresponder by AASLD, none by EASL. The cumulative rate of virological breakthrough over time in all patients was 5.6% at 72 months (Fig. 4). After genotypic analysis for resistance of these 18 patients, ETV resistance mutations were detected in 13 of them (72.2%; Table 3). The clinical characteristics of these patients are presented in Table 3; mean age was 42.2 years (SD, 8.9 years), and all but two were male. All of them were positive for HBeAg and the mean levels of serum HBV DNA and ALT were 8.1 log10 IU/mL (SD, 0.8 log10 IU/mL) and 145.1 IU/L (SD, 81.2 IU/L), respectively. Viral breakthrough occurred after a median of 30 months of ETV monotherapy (range, 6-54 months): at that time, ETV-specific resistance mutations were detected at amino-acid residues 184, 202, and 250, and emerging lamivudine-resistance at positions 180 and 204. All the ETV-resistant patients were classified as primary responders as defined by both the AASLD and EASL guidelines. Pretreatment viral load was the only independent predictor of emergence of ETV resistance in a multivariate Cox model adjusting for covariates (HR, 2.39; 95% CI, 1.34-4.26; P = 0.003).

Figure 4.

Kaplan-Meier curve for the probability of virological breakthrough in the entire patient population.

Table 3. Characteristics of 13 Patients Developing Genotypic Mutations to ETV
 Patient 1Patient 2Patient 3Patient 4Patient 5Patient 6Patient 7Patient 8Patient 9Patient 10Patient 11Patient 12Patient 13
Age (year)48452544463347293555504349
SexMaleMaleMaleMaleMaleMaleMaleFemaleMaleMaleFemaleMaleMale
HBeAg statusPositivePositivePositivePositivePositivePositivePositivePositivePositivePositivePositivePositivePositive
Baseline HBV DNA (log10 IU/ml)8.26.78.08.18.86.58.09.08.37.59.09.08.7
ALT (IU/L)25112166215942974326819512011392146
Month of resistance54485448302442303024421836
ETV mutational patternT184A, S202GM250VT184IT184IT184I, S202GS202GM250IS202GT184IS202GS202GS202GT184A, S202G
Primary VR by AASLDPresentPresentPresentPresentPresentPresentPresentPresentPresentPresentPresentPresentPresent
Primary VR by EASLPresentPresentPresentPresentPresentPresentPresentPresentPresentPresentPresentPresentPresent

Impact of Primary Responses on HCC Development and Liver-Related Mortality

During the observation period, 17 (1.4%) of the entire 1,254 patients developed HCC, and the cumulative incidence rates of HCC were 1.2% at year 2, 1.6% at year 3, 2.5% at year 4, and 2.5% at year 5, as shown in the Kaplan-Meier graph of Fig. 5. All these patients achieved primary responses by the two criteria, and did not experience virological breakthrough or develop ETV resistance. During the same period, there were only two liver-related deaths, at 30 months and 40 months, respectively, after starting ETV treatment, and these patients were negative for HBeAg, had underlying liver cirrhosis, and did not develop HCC before death. Both of the patients were designated as primary responders based on the AASLD and EASL criteria, and maintained suppression of serum HBV DNA to an undetectable level. They died from massive variceal bleeding, resulting in hepatic failure accompanied by overwhelming sepsis.

Figure 5.

Kaplan-Meier curve showing the cumulative rate of HCC development during ETV therapy in the entire patient population.

Discussion

The principle goals of treatment of CHB with oral NA are primarily to ensure sustained suppression of HBV virological activity as measured by quantification of HBV DNA, to alleviate hepatic inflammation, and to prevent progression to cirrhosis, HCC, and decompensated liver disease requiring liver transplantation. Successful treatment would thus decrease mortality.[1-6] We found that prolonged ETV monotherapy led to a VR in the vast majority of CHB patients, even in those that were HBeAg-positive, although the time to achieve a VR was clearly longer in patients not showing a primary response at 6 months as defined by AASLD criteria. Fortunately, the latter made up only 1.28% (16 patients) of the overall cohort. However, primary response at 3 months based on EASL did not predict long-term on-treatment virological outcomes. In our series, only 13 patients (1.0%) experienced virological breakthrough linked to the development of drug resistance during ETV treatment, and, moreover, none of these patients had a primary nonresponse to ETV. Interestingly, none of the 17 patients (1.4%) who developed HCC during ongoing treatment were primary nonresponders, or developed ETV resistance.

Given that early viral kinetic responses may be predictive of better outcomes and a reduced risk of viral resistance, primary on-treatment nonresponse or treatment failure is an indication for a change in therapy according to current evidence-based practice guidelines including the roadmap concept, providing compliance has been good. However, these guidelines are based on data from studies of less potent drugs with a higher risk of antiviral resistance, and which vary in terms of monitoring strategy.[7, 9-14] Although monotherapy with ETV or TDF is generally considered the initial choice of NA therapy for CHB, very little data are available about on-treatment management and, in particular, the applicability of the roadmap concept to CHB patients receiving this efficient drug. Hence, it seemed desirable to reappraise the current criteria for primary VR.

A multicenter European study from Zoutendijk et al.[18] suggested that in the great majority of patients with CHB, HBV DNA is reduced to undetectable levels after prolonged (3-year) treatment with ETV even in those with partial VR, defined as a reduction in HBV DNA of >1 log but presence of HBV DNA detectable by real-time PCR assay at 48 weeks of therapy. Thus, they recommended that ETV treatment should be continued in these patients, especially those who have a low viral load (<1,000 IU/mL) at week 48, rather than adapting the treatment as recommended by EASL.[7] These observations were also confirmed in our single-center cohort of Asian ETV-treated patients. However, the Zoutendijk et al. report compared final inhibitory effects on viral replication with interim virological outcomes measured at a single timepoint (48 weeks), which is not long and corresponds to on-treatment “partial VR” by the definition of the EASL guidelines, rather than to initial “primary VR.”

In our cohort of NA-naive patients, median time to achieve a VR was significantly longer for patients with primary nonresponse to ETV at week 24 by AASLD, whereas there was no difference in the time to VR when results were evaluated at week 12 by EASL. Importantly, the cumulative probability of achieving a VR after 54 months was around 95% in both primary responders and nonresponders as defined at 24 weeks, despite the difference in the rate at which VR was achieved. In view of the positive correlation between inflammatory activity of hepatocytes and viral replication activity,[25] the longer persistence of detectable HBV DNA in the primary nonresponders may lead to greater host tissue injury. However, it is important to emphasize that viral load was maintained at very low levels during most of the on-treatment time even in our primary nonresponders. It is also possible that the between-group difference in baseline viremic burden, which was independently predictive of a VR in our analysis, could have a biasing effect on the on-treatment virological kinetics, although primary response by AASLD was strongly associated with ultimate achievement of VR. It is interesting that, in contrast to the current view that CHB patients with primary treatment failure are at risk of developing genotypic resistance,[26-28] there were no primary nonresponders among the 13 patients who were subsequently found to harbor ETV-resistant mutants. This observation supports the conclusion that there is no need for early treatment adjustment in slow responders. Studies are needed to clarify the long-term effects of the rate at which VR is achieved, and to investigate the benefit of early switching to TDF in patients who do not respond primarily to ETV. In addition, the effect of heterogeneity of genetic backgrounds on responses to treatment, especially slow responses, should be studied.

It has been shown that HBV suppression induced by antiviral therapy can reduce the risk of HCC.[29, 30] In our ETV-treated series, the 5-year cumulative HCC incidence rate was 2.5%, which was about average compared with data from recent long-term ETV studies, although initial host and virus factors differed between the studies.[15, 30, 31] For example, the most recently published Japanese study by Hosaka et al.[30] reported that the cumulative HCC incidences were 3.7% and 13.7% at year 5 in the matched ETV-treated and control groups, respectively, and this difference was statistically significant. Because there were no primary nonresponders among our subset of patients diagnosed with HCC, we believe it would not be appropriate to individualize surveillance strategies for HCC based on the presence or absence of early on-treatment VR following ETV.

In terms of the on-treatment outcomes of the noncompliant patients excluded from the study, we found that although primary responses by AASLD, though not by EASL, were less frequent than in the final eligible compliant patients—probably due to poor adherence mainly in the first year of ETV treatment—optimal viral suppression was ultimately achieved in most of the cases. Note that this observation strongly supports our main conclusion from the patients finally recruited to the study. Interestingly, ETV-resistance mutations were never detected in the patients considered noncompliant, as was also observed in a previous smaller study.[22] These surprising outcomes are presumably the result of the intensive educational interventions on medication adherence by clinicians in response to poor patient compliance[21] as well, perhaps, as of the high anti-HBV potency of ETV per se.

An important limitation of this study is that our data were confined to ETV, and thus it remains to be established whether the current guidelines are clinically applicable to CHB patients initially receiving TDF, the other current first-line NA.

In conclusion, our validation analysis demonstrates that CHB patients with primary nonresponses at 6 months after the initiation of ETV treatment display a slower rate of reduction of viremia, but the vast majority of them finally become negative for HBV DNA in line with the primary responders. Also, a primary nonresponse to ETV is not associated with the selection of resistance mutations. Careful consideration is needed before modifying antiviral therapy in CHB patients with early suboptimal virological responses to ETV.

Ancillary