Correspondence Fabien Zoulim, INSERM, U871, 151 Cours Albert Thomas, 69003 Lyon, France Tel: +33 4 72 68 19 70 Fax: +33 4 72 68 19 71 e-mail: email@example.com
The management of hepatitis B virus resistance to antivirals has evolved rapidly in recent years. The definition of resistance is now well established, with the importance of partial response and the improvement of assays to detect genotypic resistance and virological breakthrough. Data on phenotypic resistance have allowed to define the cross-resistance profile for the main resistant mutants, providing a rationale for treatment adaptation. Clinical studies have shown that an early treatment intervention in case of a virological breakthrough or a partial response with the addition of a second drug having a complementary cross-resistance profile allows one to maintain the majority of patients in clinical remission. The prevention of resistance should rely on the use of the most potent antivirals with a high genetic barrier to resistance as a first-line therapy. The future perspectives are to design strategies to hasten the HBsAg clearance, which should become a new treatment endpoint, to prevent drug resistance and to decrease the incidence of complications of chronic hepatitis B.
Chronic hepatitis B virus (HBV) infection is a serious clinical problem and a major cause of severe liver-related morbidity and premature mortality. The primary treatment goal for chronic hepatitis B (CHB) is the suppression of HBV replication and the prevention of active disease in the long term, thereby reducing necroinflammatory activity and preventing progression to decompensated liver disease, cirrhosis and hepatocellular carcinoma (1). The development and introduction of nucleos(t)ide analogues has played a major part in the substantial advances in CHB treatment that have occurred over the past decade. Chronic conditions such as CHB require long-term therapy, and resistance to therapy is a frequent consequence of treatment duration. The emergence of drug resistance during long-term monotherapy with nucleos(t)ide analogues is almost inevitable and represents a clinical challenge. As was the case with the human immunodeficiency virus (HIV), the management of HBV infection is dynamic and subject to change as new information becomes available.
Because the emergence of resistance limits the efficacy of antiviral therapy, managing resistance has become a key issue in clinical practice. The availability of multiple nucleos(t)ide analogues and the development of highly sensitive monitoring techniques, coupled with an improved understanding of the virological factors affecting resistance, have led to a rapid improvement of HBV resistance management (2).
Goals of antiviral therapy
The short-term goals of antiviral therapy include (i) the achievement of HBeAg seroconversion, (ii) reduction in HBV DNA levels below those associated with liver disease and (iii) persistent normalization of serum aminotransferase levels. HBsAg loss and seroconversion to anti-HBsAb is another long-term goal of treatment.
Ideally, patients should become HBV DNA undetectable by real-time polymerase chain reaction (PCR) assays during therapy, because persistent viraemia has been associated with a higher rate of viral resistance. Long-term goals such as improved survival, prevention of cirrhosis and disease complications are even more important and are achievable if treatment results in a durable virological response with the lowest level of HBV DNA possible. HBsAg seroconversion occurs infrequently with current antiviral therapy, but is a desirable virological endpoint because it is associated with a lower rate of relapse, the prevention of hepatitis delta superinfection, a reduced risk of complications and improved survival in cirrhotic patients (3).
Mechanism of selection of hepatitis B virus drug resistance mutants
Unfortunately, long-term therapy exposes patients to the risk of selection of drug-resistant mutants (2). Because of the spontaneous viral genome variability, the pharmacological pressure may select for viral species that exhibit the best replication capacity in this new treatment environment. Mutations conferring resistance to nucleoside analogues are located in the viral polymerase gene. They may directly result in conformational changes of the viral polymerase, leading to steric hindrance between the nucleoside analogue triphosphate and the substituted amino acid (4, 5). The rapidity of selection of drug-resistant mutants depends on their replication capacity and fitness, their level of resistance and free liver space available for infection by these mutants (6, 7). This may explain, at least in part, the differences in the rate of resistance for the different drugs that are clinically available.
Different mechanisms are involved in drug resistance under antiviral therapy (2, 8). First, a complex mixture of genetically distinct variants develops under selective pressure. A pre-existing or a newly acquired mutation conferring a selective advantage to a variant will give rise to virions, which are more viable and can spread more rapidly in the liver. This mutant will accumulate and become the dominant species in the patient's infected liver, under the pressure of the antiviral drug. The kinetics of replacement of the wild-type virus in liver cells by a dominant mutant are generally slow. Indeed, resistant mutants mainly infect uninfected cells but can also super-infect with a lower rate of already infected cells. The efficient spreading of the dominant mutant depends on its intrinsic fitness and the availability of free liver space for its propagation and replication (6, 9). During antiviral therapy, several months may be needed for the immune system to clear wild-type infected hepatocytes and to generate new cells that are susceptible to infection by viral drug-resistant mutants. On the other hand, the specific infectivity of drug-resistant mutants may have a major impact on the rapidity of selection of these strains during therapy. Indeed, some mutations in the viral polymerase gene may result in nucleotide changes in the overlapping surface gene, which in turn may lead to reduced viral fitness, owing to impaired assembly, secretion or infectivity (10). The level of resistance to a drug conferred by a given mutation may have a profound implication on the fitness of the mutant. This may explain the difference in drug resistance rates observed with the different antivirals.
Problem of multidrug resistance with sequential monotherapy
Based on clinical experience, the knowledge of cross-resistance and viral quasi-species evolution during therapy, it was clearly shown that sequential therapy exposes patients to the risk of selection of multidrug-resistant strains. One example is the use of sequential treatment with drugs sharing common cross-resistance profiles such as lamivudine, followed by entecavir (11, 12). Another is the development of dual resistance to both lamivudine and adefovir, especially when cross-resistance mutations are selected (13, 14). It was shown by clonal analyses that multidrug resistance may occur by the sequential addition of resistance mutations on the same viral genome, leading to resistance to both drugs. Results of phenotypic analysis of the mutants during the selection process demonstrated that the cumulative addition of these mutations conferred full resistance to both drugs.
Another situation is the use of an add-on strategy with drugs having a complementary cross-resistance profile, which may lead to the selection of multidrug-resistant strains if the add-on strategy does not induce a complete viral suppression, especially if there is a large replication space available for the mutants to spread. This was demonstrated by the longitudinal clonal analysis and phenotypic analysis of the main variants in a patient who selected a multidrug-resistant strain, after liver transplantation, harbouring mutations in the overlapping polymerase and surface that conferred resistance to both lamivudine and adefovir as well as a decreased recognition by anti-HBsAb (14).
Because of the risk of development of multidrug resistance that may not be rescued by currently available drugs, treatment decision and the choice of first-line treatment should be made with caution. This implies avoiding treatment in patients who do not meet the indication. In patients who have a treatment indication, the choice of first-line therapy should rely on the more potent drug with the highest barrier to resistance. Furthermore, in specific patient populations with a high risk of resistance development and in whom drug resistance may not be tolerated, de novo combination therapy may be considered to minimize the risk of resistance development.
Persistence of detectable viraemia during prolonged therapy (risk of viral resistance)
Undetectable HBV DNA during therapy
Risk of resistance when VL was still detectable after 24 weeks (LMV, TBV) or 48 weeks (ADV) of therapy. Requires treatment adaptation
Detection of mutations in the viral genome known to confer resistance to an antiviral drug
No detectable resistance mutation
Need a standardized definition for new resistance mutations
Increase of at least 1 log10 copies/ml compared with the lowest value during treatment
Long-term viral load suppression
Confirmation of VL increase by a second HBV DNA test is not mandatory when patients also have ALT elevation
Decreased susceptibility to an anti-HBV drug of the patient strain relative to the wild-type virus in tissue culture
Determine the role of a given amino acid substitution in antiviral drug resistance
Used for research purpose, clinical trials and definition of resistance to new drugs
Decreased susceptibility to more than one antiviral drug conferred by the same amino acid substitution or a combination of amino acid substitutions
Choice of drugs for treatment adaptation. Rationale for add-on and combination therapy
Mandatory data for second-line treatment recommendation
The failure to achieve a 1 log10 copies/ml decline in viral load after 12 weeks of therapy is considered as a primary non-response. It indicates that either there is a compliance issue or the medication does not exhibit its antiviral activity in a given patient; one large study demonstrated that a suboptimal response may be owing to a host pharmacological effect or patient compliance but not a reduced susceptibility to adefovir as measured in vitro (16). When a suboptimal response is identified, antiviral treatment should be modified. A switch to a more potent nucleoside analogue is recommended at this time (17). The week 12 time point is therefore important to determine the antiviral activity of the treatment regimen. A partial response after 6–12 months depending on the drugs and the kinetics of viral load decline should lead to treatment adaptation.
A partial response corresponds to the failure to achieve a viral load decline to a threshold that translates to an improvement in liver histology and/or to a minimum risk of resistance. When considering liver histology as an endpoint, antiviral therapy should lead to a decrease in HBV DNA levels to below 10 000 copies/ml. The risk of drug resistance, however, is associated with even much lower viral loads.
The antiviral response at week 24 of therapy was also found to be a predictor of resistance in patients treated with telbivudine or lamivudine in the Globe trial (18). Higher rates of resistance at 2 years were observed when week 24 viral load was >1000 copies/ml compared with patients with a lower viral load at the same time point, whatever their initial HBeAg status.
Adefovir dipivoxil was shown to suppress viraemia levels with a slower effect in comparison with other nucleoside analogues, i.e. lamivudine, entecavir or telbivudine. Therefore, the week 48 time point may be used for predicting resistance to adefovir dipivoxil therapy (19). Interestingly, it was demonstrated in HBe-negative patients treated with adefovir dipivoxil for 192 weeks that patients with HBV DNA levels >1000 copies/ml after 48 weeks of therapy had a higher risk of developing adefovir resistance at 4 years.
With the newer generation of drugs, i.e. tenofovir and entecavir, there are no data on the impact of persistent viraemia on the subsequent development of resistance. When using these new drugs, maximum viral suppression is advised to reduce the risk of resistance. Several situations can be depicted depending on the baseline viral load and kinetics of viral decay. The timing of treatment adaptation depends on the drug used and on the kinetics of viral load decay, especially in patients starting from a very high viral load who may need additional weeks of therapy to reach the threshold of 1000 copies/ml. If viral load is still detectable after 48 weeks of therapy and viral load evolution shows a continuous decline, therapy may be continued until the next check point. By contrast, if the kinetics show a plateau level of viral load, then antiviral treatment should be adapted to prevent the emergence of resistance.
The definition of a partial response is evolving with the availability of more sensitive HBV DNA assays and more potent drugs. It can be defined by the failure to achieve viral suppression with the most sensitive assays after 12 months of therapy and should lead to treatment adaptation to maximize viral suppression and minimize the subsequent risk of resistance (20).
Genotypic resistance is defined by the detection of mutations in the viral genome that are known to confer resistance to an antiviral drug (21). The ability to detect these mutations depends on the sensitivity of the assay. Direct sequencing of PCR products allows one to detect any known or novel mutations; it is less sensitive than a line probe assay, which, however, can detect only known mutations. Other assays are not used in a clinical routine (DNA chip assays, MALDI-TOF, and restriction fragment length polymorphism assay).
Virological breakthrough is defined by an increase of at least 1 log10 copies/ml compared with the lower value during treatment, confirmed by a second test, in a treatment-compliant patient. It is usually associated with the presence of resistance mutations and follows genotypic resistance (detection of resistance mutations) (22–24). Confirmation by a second HBV DNA test is not necessary when patients also show an alanine aminotransferase (ALT) elevation.
In the absence of treatment adaptation, the increase in viraemia levels may be accompanied in the following weeks or months by an increase in ALT levels (biochemical breakthrough) and subsequently by progression of liver disease (clinical breakthrough).
Phenotypic resistance is assessed by in vitro assays that measure the inhibitory effect of an inhibitor (IC50) on the patient strain relative to the wild-type virus, in tissue culture (25). An increase in the ratio of patient's strain IC50/wild-type IC50 defines the fold resistance conferred by the viral mutation in phenotypic assays. This allows one to determine the role of a given amino acid substitution within the viral polymerase in antiviral drug resistance.
Cross-resistance refers to the situation in which a decreased susceptibility to more than one antiviral drug is conferred by the same amino acid substitution or a combination of amino acid substitutions (26). This is assessed by in vitro phenotypic assays. It has major clinical implications for the choice of drug to adapt antiviral therapy in case of resistance.
Impact of cross-resistance on the management of drug resistance
Results of in vitro cross-resistance testing have shown that rtM204V or rtM204I mutants are the most frequently observed lamivudine-resistant mutants. They may also be selected by other l-pyrimidine analogues such as emtricitabine, clevudine and telbivudine. These mutants remain susceptible to purine analogues such as adefovir and tenofovir (27–30); these mutants also have an intermediate susceptibility to entecavir (11, 31). In vitro studies demonstrated that an rtN236T adefovir-resistant mutant is susceptible to lamivudine, emtricitabine, telbivudine and entecavir, while the rtA181V mutant has a reduced susceptibility to lamivudine (14, 32, 33). Although some level of cross-resistance has been observed in vitro, rtN236T adefovir-resistant strains remain sensitive to tenofovir in vivo at least initially, most likely because of a greater exposure to active drug, which may overcome the decreased susceptibility to tenofovir. The A181V mutant resistant to adefovir was shown to be as sensitive to tenofovir in vitro as wild-type HBV (34). Further trials of a longer duration are needed to see whether tenofovir may maintain a sustained response in patients with adefovir-resistant strains. The rtA181T mutant also confers some level of resistance to lamivudine and adefovir (35), and also results in a stop codon in the overlapping surface gene leading to truncated surface proteins. This mutant may be responsible for the slow kinetics of viral load increase and virus accumulation in infected hepatocytes (10). Two mutants are subject to a controversy: the rtI233V mutation was shown to confer primary resistance to adefovir in one study but not in another one (36, 37); the rtA194T has been associated with the emergence of HBV resistance to tenofovir in one study but this was not confirmed by another (28, 38).
Current data suggest that, in patients, with lamivudine failure the development of entecavir resistance follows a ‘two hits’ model, with the first selection of primary resistance mutations at position rt204, followed by the addition of secondary resistance mutations (at position rt184, rt202 or rt250) conferring higher resistance to entecavir (11, 31). This may explain why the development of entecavir resistance is more rapid in patients with lamivudine failure who already have selected the primary resistance mutations, in comparison with the nucleoside-naive patients in whom the whole process of selection of primary and secondary mutations needs to take place. Although strains harbouring classic lamivudine resistance mutations exhibit an intermediate susceptibility to entecavir, they remain sensitive initially to entecavir in vivo, when the latter is administered at a higher dose (1.0 mg daily). If additional secondary mutations occur, resistance to entecavir is observed and is followed by viral breakthrough. This suggests that entecavir may not be considered as an optimal treatment for patients infected with lamivudine-resistant HBV, but a good treatment option in nucleoside-naive patients. If entecavir has to be prescribed in patients with lamivudine failure, lamivudine should be discontinued, and entecavir prescribed at a double dose (1.0 mg daily).
Telbivudine is ineffective in vitro against the rtM204I mutant as well as the rtL180M+M204V mutant, but remains active against the rtM204V single mutant. This may explain why the single rtM204I mutant was the only resistant mutant detected during the phase III trials of telbivudine.
Add-on therapy with drugs lacking cross-resistance
In the past, salvage therapy was proposed for patients with lamivudine resistance and clinical breakthrough (high viral load and ALT elevation). There was a debate on whether adefovir dipivoxil switch or add-on to ongoing lamivudine was the best strategy. The knowledge of cross-resistance data and the results of long-term studies advocate for an add-on therapy at an early stage, i.e. viral breakthrough, to control rapidly viral replication and prevent clinical deterioration. Several studies have shown that switching from lamivudine to adefovir monotherapy was associated with a high incidence of adefovir resistance. A large multicentre Italian cohort study showed that rescue therapy with the addition of adefovir after development of virological breakthrough in HBeAg-negative patients treated with lamivudine led to viral suppression for 3 years in most patients (39). None of the patients receiving the add-on strategy developed genotypic resistance to adefovir in contrast to those who received adefovir monotherapy. Another Italian study compared two groups of patients with known resistance to lamivudine: in one group, adefovir was added at the time of clinical breakthrough with high viraemia levels, while in the other group, adefovir was introduced earlier at the time of viral load increase with detectable resistance mutations (40). The results clearly showed that treatment efficacy was improved when adefovir was started earlier, at the early phase of virological escape. All these clinical results support an early add-on therapy to maintain an antiviral effect on the main mutants of the viral quasi-species. This concept, which was also validated in the treatment of HIV infections, should be the landmark with the new classes of antivirals to prevent the future development of strains resistant to entecavir and tenofovir.
Prevention of drug resistance
The main strategy to prevent drug resistance development is to choose the best antivirals as a first-line therapy (1). In young patients who are HBeAg positive with high ALT and moderate HBV DNA levels, pegylated interferon should be considered; this will allow a treatment with finite duration and no risk of selection of resistant strains. In all other patients, the use of the most potent antivirals with a high barrier to resistance (tenofovir or entecavir) should be the preferred choice as this should lead to the control of viral replication and liver disease in the majority of patients for several years.
Early treatment adaptation
In case of a partial virological response, treatment should be adapted to prevent the subsequent emergence of drug-resistant mutants; in this situation, the addition of an antiviral with a complementary cross-resistance profile should be considered. Indeed, a partial response may already indicate the selection of minor virus species that are resistant to the first drug (1, 15).
Combination nucleoside analogue therapy
With the availability of new antivirals, the concept of combination therapy for CHB has been debated over the past several years. Despite the lessons drawn from the antiretroviral therapy of HIV infection, the situation is clearly different. Indeed, the antiretroviral drugs belong to several different classes of compounds that target different steps of the viral life cycle (nucleos(t)ide and non-nucleos(t)ide reverse-transcriptase inhibitors, protease inhibitors, integrase inhibitors and fusion inhibitors). It was therefore easier to demonstrate in relatively short-term clinical trials that the combination of antiretrovirals can exert an additive effect on viral load suppression. Furthermore, HIV drug resistance emerges very rapidly during monotherapy. The beneficial effect of combinations was, therefore, assessed in short-term trials showing the added value of combination in terms of viral load decline, prevention of drug resistance and a decrease in mortality rate (41). By contrast, in the setting of CHB, antivirals belong to the same class of nucleos(t)ide analogues and target only the viral polymerase (2). This might be the reason why the combination of nucleoside analogues did not show any additive effect in terms of viraemia decline compared with the most potent antiviral drug in the combination (42–44). However, the issue of prevention of drug resistance by combination therapy is critical. Several drugs with different cross-resistance profiles are now approved. Clinical experience has shown that the combination of nucleoside analogues with complementary cross-resistance profiles is an effective strategy to manage resistance (40). The new generation of inhibitors also has an improved resistance profile with very low rates of resistance in nucleoside-naive patients during the first 5 years of therapy. The benefit of combination therapy is therefore difficult to demonstrate in short-term trials. An urgent priority is to identify the patients who have the highest risk of resistance (e.g. long-standing infection and high viraemia levels associated with more complex viral populations before therapy, together with high ALT levels that are associated with a more rapid hepatocyte turnover that in turn generates a wider replication space) (6, 45, 46). It will be important to assess in the future whether the determination of the size of the viral quasi-species and the ultrasensitive detection of pre-existing mutants may help to predict resistance and may therefore represent an indication for combination therapy. Another important group for preferred use of combination therapy concerns the patients who can least afford to develop antiviral drug resistance from a clinical perspective (e.g. patients with liver cirrhosis and/or with HBV recurrence after liver transplantation). Strategy trials comparing combination therapy vs. monotherapy followed by early add-on therapy in case of a partial response should be evaluated as soon as possible in these patient groups. In this view, the characterisation of the cross-resistance profile of these antivirals by phenotypic assays (47) will be mandatory for the choice of drugs to be included in these combinations.
Perspectives: towards an improved control of viral replication and clearance of HBsAg?
Major breakthroughs in the treatment of CHB have occurred in the last 10 years. The development of nucleos(t)ide analogues with a high barrier to resistance and potent antiviral efficacy provides a new perspective. Indeed, one may expect that long-term viral suppression overcoming antiviral drug resistance may lead to a progressive depletion of intrahepatic viral cccDNA and subsequently to a decline in serum HBsAg titres. The ultimate goal will be to achieve HBsAg clearance. To this aim, new inhibitors targeting other steps of the viral cycle or the immune response may be required in combination with nucleoside analogues to hasten the viral clearance process.
This work was supported in part by grants from the European Community (ViRgil network of excellence; ViRgil LSHM-CT-2004-503359).
Conflicts of interest
The authors have not declared any conflicts of interests.