Dr M. H. Nguyen, Division of Gastroenterology and Hepatology, 750 Welch Road, Suite 210, Palo Alto, CA 94304, USA. E-mail: firstname.lastname@example.org
Background One of the most important factors in treatment failure using nucleos(t)ide analogues in chronic hepatitis B is anti-viral resistance. Primary drug resistance refers to amino acid changes in the hepatitis B virus polymerase/reverse transcriptase (rt) that result in reduced susceptibility to anti-viral agents. Pre-existing drug resistance mutations may occur in untreated patients and may affect their treatment outcomes.
Aim To determine the prevalence of hepatitis B DNA polymerase mutations in treatment-naïve patients.
Methods We used a direct PCR sequencing test to detect DNA polymerase mutations in 472 consecutive treatment-naïve patients at two community gastroenterology clinics in Northern California.
Results A majority of patients were Asians (>95%), had either genotype B or C (95%) and had no evidence of cirrhosis or liver cancer (94%). Mean age was 45 ± 13 and mean hepatitis B virus DNA was 5.3 ± 1.8 log10 IU/mL. Most patients did not have any detectable mutations (82.4%). Some (16.7%) had mutations of unknown clinical significance (rtV207M/L/I) and only 4 patients had rtA181A/S, rtA194S or M250I.
Conclusions No rtM204V/I or rtN236T mutations were observed in our study. Less than 1% of our patients had mutations that can be associated with primary resistance to existing anti-viral therapies for hepatitis B virus.
Hepatitis B virus (HBV) is the leading cause of chronic liver disease worldwide.1, 2 Baseline HBV DNA is highly predictive of disease progression with development of cirrhosis and liver cancer.3 At least in patients with advanced fibrosis, treatment with anti-viral therapy can decrease the risk of disease progression to cirrhosis and incidence of hepatocellular carcinoma.4
Hepatitis B virus life cycle depends on reserve transcription, catalysed by a reverse transcriptase (rt), a type of DNA polymerase. Currently available oral anti-HBV therapies are either nucleoside or nucleotide analogues targeting the HBV DNA polymerase.
Primary resistance mutations associated with drug failure for HBV involve eight codons in four different domains in the HBV polymerase gene: codons 169, 180, 181, and 184 in B-domain; codons 202 and 204 of C-domain; codon 236 of D-domain; and codon 250 of E-domain.5
Mutations of HBV DNA polymerase are common when patients are treated with certain nucleoside analogues such lamivudine (LAM) or telbivudine (LDT).5–9 Viral mutations also occur in patients treated with adefovir (ADV) although at a lower frequency.8, 9 Underlying HBV DNA polymerase mutations from previous exposure to a related nucleoside analogue such as LAM also predispose patients treated with newer agents such as entecavir (ETV) to develop genotypic and eventually phenotypic viral resistance with clinical virological breakthrough.10–12 With long-term therapy, viral resistance can eventually develop with most anti-viral agents as monotherapy.8, 9 HBV DNA polymerase mutations may also occur well before any observation of clinical viral rebound which can bring deleterious consequences especially in patients with advanced disease.6 Identifying cases with genotypic mutations prior to development of clinical virological breakthrough may allow care providers to modify the patients’ anti-viral regimens and/or avoid development of fulminant liver failure because of virological breakthrough.
Hepatitis B virus DNA polymerase mutations have also been reported in variable frequencies and with unclear clinical significance among treatment-naïve patients in published studies of relatively small sample size or studies reported only as abstracts to date.11, 13–30 Patients from these studies generally were hospital-based or recruited from tertiary care centres from Europe or Asia and may not represent the general patient population with CHB in the United States. The definition of treatment-naïve CHB in these studies is also variable and can include patients whose exposure to anti-HBV nucleoside analogues was more than 1 year from the time of mutation testing.
Our goal was to determine the prevalence of HBV DNA polymerase mutations in a large community-based patient population with CHB and no history of previous treatment for HBV in the United States.
Materials and methods
Study design and patients
We performed a cross-sectional study of 472 consecutive patients with treatment-naïve CHB and genotypic mutation analysis done between 11/2005 and 6/2008 at 2 community gastroenterology clinics in Northern California, USA, using electronic query with ICD-9 code diagnosis for chronic hepatitis B and test order for HBV genotype and mutation analysis. The medical records of all study patients were reviewed using a case report form.
Laboratory tests including quantitative serum HBV DNA polymerase chain reaction (COBAS AmpliPrep/COBAS TaqMan System, Roche Diagnostics, Indianapolis, IN, USA) were performed at local community clinical laboratories operated by either Quest Diagnostics (San Juan Capistrano, CA, USA) or LabCorp (Burlington, NC, USA).
Hepatitis B virus genotype and mutation analysis was performed with the Nichols Institute ‘Hepatitis B Virus Drug Resistance, Genotype Test’ (test code 10529×).21, 22 Automatic sequencing of a portion of the polymerase gene was used to detect resistance mutations for anti-HBV agents approved at the end of the study in June 2008: LAM, ADV, ETV, LDT. When there is a mixed base population of wild-type and mutant virus, then the lower limit of detection for minor species is approximately 20–25%.21, 22 When there is an unmixed population or mutant virus, the mutated codon corresponding to drug resistance will always be detected.
All liver biopsies were obtained via percutaneous techniques using a 16-gauge biopsy needle and histological examination was performed according to the Batts-Ludwig scoring system at a single pathology laboratory. Cirrhosis was defined by histology (stage 4) or by clinical signs of portal hypertension such as splenomegaly, thrombocytopenia, ascites, encephalopathy or varices.
Analysis of various demographic and clinicopathological characteristics of patients was carried out for the total study sample and by hepatitis B e antigen (HBeAg) status. The results were reported as mean and standard deviation and/or median and range for continuous variables and frequency counts and percentages for categorical variables. Categorical variables were evaluated using chi-square (χ2) tests. The student t-test was applied to normally distributed non-categorical variables, while non-parametric statistics were applied to all others.
A two-tailed P value of ≤0.05 was considered statistically significant. All analyses were performed using the software package stata, version 9.0 (Stata Corporation, College Station, TX, USA).
The study was approved by the Western Institutional Review Board, Olympia, Washington and the Institutional Review Board at Stanford University.
Approximately half of the patients were men and the majority (86%) of patients were older than 30 (Table 1). Most patients (94%) were either Vietnamese or Chinese. The remaining patients were Filipino (n = 11), Korean (n = 2), South Indian (n = 5), Cambodian (n = 2), Caucasian (n = 2), Hispanic (n = 3), or other ethnicity (n = 1). None of the patients received any treatment for CHB prior to HBV mutation analysis.
Table 1. Patient clinical characteristics, overall and by hepatitis B e-antigen (HBeAg) status
All treatment-naïve patients (n = 472)*
HBeAg+ patients (n = 125)*
HBeAg− patients (n = 332)*
* Total number of patients unless otherwise indicated by subheadings. HBeAg status unknown in 15 patients.
Mean albumin, total bilirubin, prothrombin time and platelet count were generally within the normal range (Table 1). A total of 17 patients of 472 had cirrhosis and none had hepatocellular carcinoma at the time HBV mutation analysis.
Hepatitis B e antigen-negative patients were generally older and had lower mean HBV DNA levels (Tables 1 and 2). HBeAg-negative patients also had higher proportion of genotype B and lower proportion of genotype C, compared with HBeAg-positive patients (79% and 17% vs. 58% and 38%) (Table 2). The majority of the study patients had either genotype B or C with only few having genotype A, D and F. Otherwise, there were no other significant differences between HBeAg-negative and HBeAg-positive patients with regard to their clinical and virological characteristics (Tables 1 and 2).
Table 2. Virological characteristics, overall and by hepatitis B e-antigen (HBeAg) status
All treatment-naïve patients (n = 457)*
HBeAg+ patients (n = 125)*
HBeAg− patients (n = 332)*
* Total number of patients unless otherwise indicated by subheadings.
HBV DNA quantity (log IU/mL) (n = 455, 125/330)
5.3 ± 1.8
7.2 ± 1.5
4.5 ± 1.3
Patients in HBVDNA range (log IU/mL) (n = 455, 125/330)
HBV Genotype (n = 457, 125/332)
Prevalence of HBV DNA polymerase mutations
Out of a total of 472 patients, a vast majority (99.1%) either did not have any detectable HBV DNA polymerase mutations (82.4%) or only had a single substitution mutation rtV207 M/L/ or I (16.7%) that has no known clinical significance (Figure 1).
No YMDD motif with rtM204V/I or rtN236T mutation was detected in any of our 472 study patients.
Only 4 patients had mutations at positions that may be associated with viral resistance to anti-HBV therapy: rtA181A/S (n = 2), rtA194S (n = 1), and rtM250I (n = 1). A detailed profile of these four patients was described in Table 3. All of these patients had HBeAg-negative CHB with age ranging from 42 to 74, either had genotype B or C and none seemed to have decompensated liver disease.
Table 3. Profiles of patients with HBV DNA polymerase mutations other than rtV207M/L/I substitutions
Baseline HBV DNA
Histology Grade/ stage
* In association with rtV207M.
rt, reverse transcriptase.
Patient #2–4 have not been treated with any anti-HBV therapy to date. Patient #1 was started on a nucleoside analogue (LDT) in July 2007 with complete viral suppression with HBV DNA < 60 IU/mL and normalization of ALT within 2 months of treatment initiation and has remained so at his 20-month follow-up.
No significant correlation between baseline mutations and family history of chronic hepatitis B was found in our sample: A family history of infection with HBV was present in 19.8% of patients with no baseline mutations and 13.3% of those with any baseline DNA polymerase mutations, P =0.165.
Hepatitis B virus DNA polymerase primary resistance mutations in treatment-naïve patients have been reported in variable frequencies and/or with unclear clinical significance in mostly small or preliminary studies.11, 13–30Table 4 summarizes results of some of these studies where sufficient data on patient characteristics can be found.13, 14, 16–20, 22, 23, 28–30 Most were conducted in hospitals or referral centres in Europe and Asia and many included fewer numbers of patients. To our knowledge, our study is the largest study to date (n = 472) that examines the prevalence of HBV DNA polymerase primary resistance mutations in treatment-naïve patients and the first to include U.S. community-based patients with CHB, most of whom were asymptomatic (only 17 or 3.6% with cirrhosis and none with HCC).
Table 4. Prevalence of DNA polymerase mutations in treatment-naïve patients
A vast majority of our 472 consecutive patients with CHB and mutation analysis testing did not have any detectable HBV DNA polymerase mutations. Approximately 17% had a single substitution mutation that has no known clinical significance (rtV207M/L/I). Only four patients (0.9%) had mutations that have been known to be associated with anti-HBV drug resistance: two patients with rtA181A/S, one with rtA194S and one with rtM250I. None had rtM204V/I or rtN236T mutations that can be associated with primary resistance to the anti-HBV nucleoside analogues (LAM, LDT, clevudine, and ETV) or nucleotide analogues (ADV and TDF) respectively.5
Mutations with rtA181 codon changes are associated with the shared pathway whereby treatment with either nucleosides (LAM, LDT, clevudine, ETV) or acyclic phosphonates (ADV, TDF) can lead to selection of HBV quasispecies. This is seen in about 40% of treatment failure with ADV and 5% with LAM.5 Mutations with rtA194 may be associated with TDF resistance and M250 codon changes can be associated with ETV resistance but usually only in association with rtM204V and rtL180M mutations.
The next largest study examining prevalence of HBV DNA polymerase mutation was recently reported as an abstract presentation by Ludwig et al.23 This retrospective study also used a direct sequencing technique to detect HBV DNA polymerase gene mutation that had similar sensitivity (20–25%) as the one used in our study. It was conducted at a university centre in Duesseldorf, Germany and included 288 treatment-naïve CHB patients. The authors found that about 10% of all of their patients had primary resistance mutations against LAM, LDT and possibly ETV and TDF including 1.4% with M204I/V and 1.3% with rtN236D or A181S/V/D. This prevalence of potential primary resistance is much higher than that seen in our population. It remains unclear if this difference may be because of the different patient populations between their study and ours: Europeans vs. Asian Americans, university-based vs. community-based, and largely HBV genotype A and D (86%) vs. HBV genotype B and C (95%). Huang et al. reported a rate of 26.9% of YMDD mutations in 104 CHB patients from Asia. This study also used a PCR-based technique to detect HBV DNA polymerase mutations, but they define ‘treatment-naïve’ as not having LAM for at least 1 year and thus may have included many patients with prior exposure to LAM given its common use in Asia.18
Some of the studies examining baseline mutations in treatment-naïve CHB patients employed indirect techniques such as the line probe assay INNO-LiPA HBV DRv2 and DRv3 by InnoGenetics (Ghent, Belgium).22, 24–27 Although the line probe assay cannot detect novel mutations and polymorphisms within the regions of the probes can prevent annealing and yield indeterminate results, the line probe assay can detect mixed mutant viral population occurring at lower proportion than sequencing techniques. Baseline prevalences of various HBV DNA polymerase gene mutations as reported in abstracts from these studies also vary greatly.16, 28, 29 Feeney et al. reported a baseline prevalence of 5.5% for rtM204I/V in 108 patients with acute or chronic hepatitis B seen at a university hospital in Ireland.16 A much higher baseline prevalence of rtM204V/I mutations (12%) was recently reported by a study of 209 untreated mostly Asian patients (85%) with CHB seen at a university centre in Canada.28 In this study, no rtA181V/T or rtN236T mutations were found. In contrast, Lampertico et al. reported a pre-treatment prevalence of 3% for M204V/I and 8% for A181T for 74 patients with CHB seen at 9 Liver Units in Italy.29 Mutations associated with ETV resistance such as S202 and M250 codon changes were also seen and the majority of the observed mutations were single mutations.
Although our study includes a large patient sample not drawn from a tertiary referral centre, it has some limitations. Given the study location in the San Francisco South Bay, our study patients included largely Asian Americans and patients with HBV genotypes B and C; therefore, our results may not be generalizable to non-Asians with CHB or patients with HBV genotypes other than B and C. Our study is also retrospective, but the retrospective design in this case would most likely bias towards higher baseline mutation rates because of the higher possibility of inclusion of patients with previous anti-HBV treatment if this information had not been rigorously obtained by the treating clinicians. In fact, our study reported some of the lowest prevalences of baseline mutations to date.
Results of HBV DNA polymerase mutations in our study were obtained via direct sequencing techniques, which is considered the gold-standard and can detect novel mutations, though its sensitivity is poorer than that of the line probe assays.22, 25–27 However, it remains unclear if and how the mutant viral population existing in low frequency, which would not have been detected by the sequencing test, as compared to the wild-type population and without anti-viral selection pressure, would affect treatment outcomes of patients with CHB. According to results from a recent study, many of these mutant viruses may be defective with stop codons or lethal G to A hypermutation.30 In this study, ultra-deep pyrosequencing (UDPS) of HBV quasispecies was performed in 17 untreated patients and can detect reverse transcriptase mutations down to a level of 1–2%. Two patients were found to have nucleoside-resistance mutations detectable only by UDPS including rtM204I at a level of 1.3% in one patient and rtA181T and rtM204I at levels of 1.0% in another patient; however, when clonal sequencing was performed to confirm the presence of these mutations, two of three clones with these mutations had stop codons and evidence for lethal G to A hypermutation. At a clinical level, Lampertico et al. did not find that baseline nucleoside-associated mutations detected by the more sensitive line probe assay influence virological response to ETV in their patient cohort.29 Therefore, while the primary limitation of the direct sequencing method is the relatively low level of sensitivity, it remains unclear what the advantage of higher sensitivity with other methods for mutation detection such as INNO LiPA testing would be as the clinical significance of viral quasispecies present in low quantity remains unknown.
In conclusions, no baseline rtM204V/I or N236T mutations were observed in our 472 consecutive untreated, almost all non-cirrhotic and community-based patients with CHB. Less than 1% of our patients had mutations that can be associated with primary resistance to ADV, LAM or ETV; however, the clinical significance of these HBV quasispecies when occur as single mutations in untreated patients remains to be determined. Large prospective studies examining the prevalence and clinical significance of baseline HBV DNA polymerase mutations in the general population of untreated CHB patients are needed.
Declaration of personal interests: Mindie H. Nguyen has served as consultant or advisory board member and has received research support from Roche Laboratories, Novartis, Bristol-Myers Squibb, and Gilead Sciences. Ruel T. Garcia has served as consultant or advisory board member for Roche Laboratories, Gilead Sciences, Novartis and has received research support from Roche Laboratories and Novartis. Huy Trinh has served as consultant or advisory board member for Roche Laboratories, Novartis, Idenix, Bristol-Myers Squibb, and Gilead Sciences and has received research support from Roche Laboratories, Bristol-Myers Squibb, and Gilead Sciences. Huy Nguyen and Khanh Nguyen have served as consultant or advisory board member for Bristol-Myers Squibb, and Gilead Sciences. Long Nguyen and Brian Levitt: none to disclose. Declaration of funding interests: The study is supported in part by an investigator-initiated grant from Novartis Pharmaceuticals.