Correspondence Patrick Marcellin, Service d'Hépatologie, INSERM-CRB3, Hôpital Beaujon, APHP University of Paris 7, Clichy, France. Tel: +33 1 40 87 53 38 Fax: +33 1 47 30 94 40 e-mail: email@example.com
Since its approval for the treatment of chronic hepatitis B in 1998, lamivudine (LAM) has been used extensively throughout the world, because of its relatively low costs and favourable tolerability. However, clinical trials and cohort studies have demonstrated that a high rate of resistance to this drug develops and, as a result, it is no longer included as a first-line therapy in most current treatment guidelines. Nevertheless, because of its low cost, this drug continues to be used in many countries and the pool of patients who have developed resistance to LAM continues to increase. Thus, there is a clear need to develop coherent management strategies to treat such patients as well as limit the emergence of resistance in the first instance. The purpose of this review is to highlight the need to aim for long-term treatment success while limiting the emergence of drug resistance and its consequences for the future. In addition to add-on/switch strategies with other nucleos(t)ide analogs, currently available data suggest that interferon-based therapies, with their potential to induce a sustained response, are worthy of consideration not only for reducing de novo resistance but as an option for the management of those patients in whom drug resistance has already developed.
Lamivudine (LAM) was approved for chronic hepatitis B (CHB) in 1998 and has since been used extensively worldwide, largely because of its relatively low cost (available generically in some countries) and favourable tolerability. However, clinical trials and cohort studies have shown that resistance to LAM develops rapidly, with around 70% of patients harbouring LAM-resistant YMDD mutant viral strains after 5 years of treatment (1, 2). This can have serious clinical consequences because the virological breakthrough that follows the development of LAM resistance is associated with a severe, acute and sometimes fatal exacerbation of liver disease (3, 4). LAM resistance is especially important in CHB patients co-infected with human immunodeficiency virus (HIV) and those with fibrosis or cirrhosis (5).
Given the high incidence of LAM resistance and the possibility of the development of subsequent resistance to other nucleos(t)ide analogs (NAs), it is not surprising that LAM is no longer included as a first-line therapy in most current treatment guidelines and that more recent NAs, adefovir (ADV), entecavir (ETV), telbivudine (LdT) and tenofovir (TDF) are preferred (6–8), alongside pegylated interferon (pegIFN). There is unlikely to be any clinical benefit to continuing LAM monotherapy in patients who have developed resistance. Despite this, the use of LAM remains common around the world, particularly in some parts of Asia, where the prevalence of CHB is high and generics are available at a low cost.
There is a need to develop coherent management strategies to treat the increasing number of patients with LAM-resistant virus (including early detection of resistance) as well as to limit the emergence of resistance in the first instance. The purpose of this review is to highlight the need to aim for long-term treatment success while limiting the emergence of drug resistance and its consequences for the future. The approaches adopted so far were reviewed, including the use of NAs and pegIFN for managing resistance once it has emerged. The optimal treatment approach for these patients is currently still under investigation; however, the ability to tailor therapeutic regimens to individual patients, a major goal in hepatitis B virus (HBV) therapy, is likely to play an important part (9–11).
To detect the possibility and risk of resistance in patients undergoing treatment with NAs, HBV DNA should be assessed at baseline and every 3 months for the first year of therapy to ensure stability (12). Three-monthly testing should continue in patients who remain HBV DNA positive on therapy, but can be reduced to every 6 months in patients who have undetectable HBV DNA (below the detection level of the most sensitive assays) (12). Given the risks of exacerbations of liver disease on the development of resistance, rebound or breakthrough in patients with more advanced fibrosis or cirrhosis, long-term monitoring every 3 months is the preferred option, with additional liver function testing and careful clinical assessment (12). Although the need for constant and regular monitoring has implications in terms of cost and patient inconvenience, it is an essential part of patient management in patients with CHB undergoing treatment with NAs (8, 12, 13).
Strategies to avoid the development of resistance to nucleos(t)ide analogs
There are two main strategies for avoiding the incidence of NA resistance: postponing treatment until absolutely necessary (6–8, 12) and then careful choice of therapy in those patients who require treatment. Current first-line treatment options for both HBeAg-positive and HBeAg-negative disease include pegIFN as well as NAs.
Choice of therapy
Pegylated IFN has the potential to deliver a sustained response following a finite course of treatment without the risk of drug resistance development. The safety and efficacy of pegIFNα-2a, alone or in combination with LAM, has been established in large clinical trials (14, 15). In HBeAg-positive disease, 48 weeks of treatment with pegIFNα-2a monotherapy was associated with a significantly higher rate of HBeAg seroconversion and HBV DNA levels <100 000 copies/ml compared with LAM monotherapy 6 months post-treatment. HBsAg clearance was reported in 16 (3%) of patients receiving pegIFNα-2a at 6 months post-treatment compared with none in the LAM group (P=0.001) (14). Alanine aminotransferase (ALT) normalization rates were also higher in pegIFNα-2a-treated patients compared with LAM-treated patients. At 12 months post-treatment, the HBeAg response was sustained in 83% of Asian patients with HBeAg seroconversion at 6 months post-treatment (16). Among Asian patients without HBeAg seroconversion at 6 months post-treatment, 15% of patients went on to achieve seroconversion by 12 months post-treatment. Similar results have been reported with pegIFNα-2b in HBeAg-positive patients (17). Treatment with pegIFNα-2b alone or in combination with LAM for 52 weeks resulted in HBeAg and HBsAg negativity in 37 and 7% of patients at 24 weeks, and 37 and 11% after a mean of 3 years of follow-up (18). In HBeAg-negative disease, 6 months post-treatment, the percentages of patients with normalization of ALT levels and HBV DNA levels <20 000 copies/ml were significantly higher with pegIFNα-2a monotherapy (59 and 43% respectively) than that with LAM alone (44 and 29% respectively) (15). Twelve (3%) patients treated with pegIFNα-2a cleared HBsAg compared with none of those treated with LAM (P=0.007). Baseline ALT and HBV levels, patient age and gender and HBV genotype all significantly influenced the durability of response 6 months post-treatment (19). Long-term follow-up demonstrated that at 3 years post-treatment, the rates of biochemical and HBV DNA response remained significantly higher in patients treated with pegIFNα-2a compared with those treated with LAM monotherapy and the rate of HBsAg clearance increased to 9% in pegIFNα-2a-treated patients (20). The higher rates of virological and biochemical response in pegIFNα-2a-treated patients were still apparent 5 years after completion of therapy. Importantly, the rate of HBsAg clearance in pegIFNα-2a-treated patients increased further to 12.2%, highlighting the durability of the post-treatment response (21). Taken together, these long-term data on the durability of response, high rate of HBsAg clearance, the fixed duration of treatment and the lack of resistance support the use of pegIFNα as first-line therapy for patients with HBeAg-positive or -negative disease. PegIFNα has been shown to be effective in patients with HBeAg-positive disease with fibrosis (22, 23) and, in the absence of any contraindications, can also be considered for first-line use in patients with compensated CHB cirrhosis (24).
Patients embarking on a course of interferon (IFN)-based therapy need to be made aware of the possibility of treatment-related side effects, the most commonly encountered of which include tiredness, headaches, fever and myalgia. Some minor side effects can often be managed with the use of simple analgesics such as paracetamol (acetaminophen), while more persistent or severe effects may require dose modification. Overall, the tolerability of pegIFNα-2a in patients with hepatitis B appears to be better than in those with hepatitis C, with a lower incidence of common IFN-related adverse events (AEs) and a significantly lower incidence of depression (25). Patients with decompensated disease, as well as those who do not respond to IFN-based therapy or are unable to tolerate it, are more suitable candidates for treatment with NAs.
Nucleos(t)ide analogs currently approved for the treatment of CHB include LAM, ADV (26, 27), ETV (28–30), LdT (31) and TDF (32). Another NA, clevudine, was recently withdrawn due to toxicity problems after a limited period of availability only in Korea (33, 34).
The generation and multiplication of NA-resistant mutations is dependent on viral replication, and so, once initiated, these oral antivirals need to suppress replication as rapidly and completely as possible. Treatment-resistant mutations may therefore be limited by using the most potent NAs available. However, resistance development is a class effect and increases with the length of treatment, although the rate and degree of development varies between different NAs. While NAs are easy to take (given their oral formulation) and well tolerated, it is noteworthy that compliance is essential to ensure complete suppression of HBV DNA replication. Patients who miss doses run the risk of allowing drug levels to decline, thereby facilitating emergence of drug-resistant variants.
Lamivudine has been widely used in the treatment of CHB and is effective in suppressing HBV replication and in improving liver disease in both HBeAg-positive and HBeAg-negative patients (6, 7). After 1 year of therapy, HBeAg-positive patients showed a 4–5 log10 copies/ml reduction in the serum HBV DNA level with 18–20% HBeAg seroconversion (35). After 5 years of therapy, HBeAg seroconversion was reported in 35–65%, with the highest response in those patients with high [>5 × upper limit of normal (ULN)] ALT levels at baseline. In HBeAg-negative patients, 60–70% showed HBV DNA suppression to below detectable limits after 1 year of therapy (6). However, around half of the HBeAg-positive patients and up to 90% of HBeAg-negative patients who respond to treatment relapsed if LAM was withdrawn. Loss of HBsAg, the ultimate serological goal of therapy, is very rare (<1%) in LAM-treated patients (14, 15, 20, 31, 36, 37). The major drawback associated with its long-term use is the high risk of LAM-resistant mutations, which arise in 60–70% of patients after 5 years (6). Although long-term therapy can reduce the incidence of complications of HBV-related liver disease in patients with advanced liver disease, this is only true for as long as viral suppression is maintained and the development of resistance leads to virological rebound and biochemical breakthrough (38). The long-term safety profile of LAM is good in the absence of resistance development, but patients with LAM-resistant mutations are at risk of worsening liver disease (39). Despite the risk of resistance and its associated problems, LAM remains the mainstay of treatment in some regions due to its low cost. However, initial savings may be negated when salvage therapy is required to manage resistance. It is possible that restricting first-line LAM to patients most likely to respond may be an option in such regions. A number of factors have been found to improve the likelihood of response to LAM, including high baseline ALT, female gender and younger age (35). A high pretreatment serum HBV DNA level and a high level of the residual virus after initiation of treatment are associated with an increased risk of the development of LAM-resistant mutations (6). Given this, it may be possible to select patients with favourable characteristics who may be more suitable for first-line LAM monotherapy. There is also evidence that on-treatment response to LAM can predict the treatment outcome. Serial HBeAg levels have been shown to be useful in predicting YMDD mutant emergence in HBeAg non-seroconverted patients, with a pattern of descending and then increasing HBeAg levels on-treatment during LAM treatment predicting the emergence of YMDD mutations with a sensitivity of 66% and a specificity of 100% (40). Recently, a Roadmap concept of HBV treatment management has been proposed, which uses 24-week virological responses to individualize ongoing management in patients in an attempt to minimize long-term resistance and improve efficacy (Fig. 1) (41). This approach is thought to be especially useful for treatment-naïve patients treated with NAs with a low genetic barrier to resistance, such as LAM, ETV and LdT. The ability to predict suboptimal response to a particular NA means that extended monotherapy with the associated risk of resistance development may be avoided. However, whether or not the Roadmap concept is followed, given the potential detrimental effects of resistance emergence, regular monitoring of serum HBV DNA should be considered essential for early detection and the initiation of prompt salvage therapy (35).
Treatment of CHB patients with ADV results in less marked reductions in HBV DNA compared with LAM (42, 43). In HBeAg-positive disease, 39% of 65 patients who received ADV for 5 years had HBV DNA <1000 copies/ml and 66% had normalized ALT (27). The estimated probability of HBeAg seroconversion at year 5 of treatment was 30%. HBsAg loss/seroconversion was observed in four patients during the course of this study (2%) (27). In HBeAg-negative disease, viral suppression was sustained in only 8% of patients who stopped ADV after 1 year of treatment and, although the majority of patients who continued treatment up to 5 years maintained their response, there was no incremental response after the first year (44). Although long-term treatment with ADV is associated with a decreased fibrosis score, 2% of patients developed HCC, indicating that long-term antiviral treatment does not completely prevent HCC (26). The use of ADV is associated with a risk of renal toxicity and so creatinine monitoring should be undertaken at least every 6 months during ADV therapy. However, in a long-term study, only 3% of patients showed slight elevations in creatinine over 5 years (26).
In clinical trials, ETV showed significantly higher rates of histological, biochemical and virological response compared with LAM in both HBeAg-positive (28) and -negative disease (29). Around 70% of patients maintain HBeAg loss after 1 year of therapy, although long-term durability is currently unknown. A recent article highlighted the need for continuous therapy with NAs in patients with HBeAg-negative disease only 3% of patients treated with ETV and 5% of patients treated with LAM who had achieved a protocol-defined response at week 48 still had undetectable HBV DNA 6 months after therapy was stopped (30).
Telbivudine is more potent than LAM (45) and ADV (46) in suppressing HBV replication; however, its resistance profile is suboptimal, and LdT resistance mutations are cross-resistant with LAM. LdT monotherapy may be an effective option in patients with relatively low baseline HBV DNA levels (<7 log10 copies/ml in those with HBeAg-negative disease and <9 log10 copies/ml in those with ALT >2 × ULN) and a confirmed complete virological response at 24 weeks (47–49). In these patients, the response rate is high and the risk of resistance is low.
In head-to-head comparisons, TDF provided greater suppression of HBV DNA than ADV in HBeAg-positive and -negative patients after 48 weeks of treatment and the rates of viral suppression were maintained over 2 years (50). The rates of HBeAg seroconversion on treatment were similar for TDF and ADV; however, in HBeAg-positive patients, the rates of HBsAg clearance were significantly higher in patients treated with TDF compared with ADV (3 vs. 0%, respectively; P=0.02). It should be noted, however, that this study was performed in North America, with a high proportion of genotype A-infected patients who appear to have a higher probability of HBsAg clearance compared with other HBV genotypes. No cases of resistance to TDF have been observed so far (51–53).
Currently available NAs therefore have limitations, particularly associated with the need for long-term therapy and/or resistance development when used as monotherapy. In addition, the sequential use of NAs to address resistance development is beginning, as it has in other chronic viral infections, to result in the development of multi-drug resistant HBV (54). While the Roadmap concept goes some way towards addressing this (41), additional strategies could also have potential.
The potential advantages of combination therapy are additive or synergistic antiviral effects, and diminished or delayed resistance; however, potential disadvantages include possible increased toxicity, drug interactions and increased costs (6). Several combinations have been evaluated in patients with CHB. To date, no combination of NAs has proven to be clinically superior to monotherapy with regard to efficacy. The main benefit of combination therapy appears to be the potential to reduce the emergence of resistance. Potent NAs with a good resistance profile are now available; however, the need for combinations of NAs to prevent resistance is questionable. To date, the only combinations that have demonstrated a synergistic effect in terms of HBV DNA suppression during therapy are pegIFNα+LAM (14, 15, 55).
Nucleos(t)ide analog-based combination therapy
The potential for successful NA-based combination therapy is complicated by the possibility that combinations of NAs may target the same viral mechanisms, and so may be antagonistic, and by the cost implications associated with long-term therapy using more than one drug.
The lack of cross-resistance in vitro suggests that the combination of LAM and ADV may potentially be beneficial. A study comparing the efficacy of LAM plus ADV vs. LAM alone showed a similar initial reduction in HBV DNA up to week 16 in both groups (56). By week 104, differences were apparent, with a greater HBV DNA reduction in the combination group as well as higher rates of HBV DNA suppression and ALT normalization. However, these differences in HBV DNA reduction were not associated with differences in HBeAg seroconversion (56). Although genotypic resistance was less common in the combination group, a substantial percent still had mutation in the YMDD motif (15 vs. 43% in the LAM monotherapy group) (57). In addition, no obvious improvement in ALT normalization and virological suppression or reduction in the development of ADV-resistant mutations has been seen with ADV plus LAM compared with ADV monotherapy (58). Data on the ability of other combinations to prevent resistance are currently lacking. In one study, naïve patients were randomized to receive ADV plus placebo or ADV plus emtricitabine (FTC) for 96 weeks. Although greater HBV DNA suppression and ALT normalization were found in the combination group, the rates of HBeAg seroconversion at week 96 were similar [2/14 (14.3%) vs. 4/16 (25.0%) in the combination and monotherapy groups, respectively; P=NS], although patient numbers were small. No patients developed resistance mutations during this study (59).
Interferon/nucleos(t)ide analog combinations
The combination of IFNs with NAs is attractive, given the immunomodulatory effects of IFN and its potential to act with the antiviral action of NAs. There is also some evidence that IFN/NA combination therapy reduces the risk of LAM resistance. In a Japanese study, fewer patients in the IFN–LAM combination arm (8%) had YMDD mutations at week 52 compared with the LAM monotherapy arm (30%; P=0.047) (60). There is some discussion over whether it is better to start LAM and IFN simultaneously or to have a lead in phase with LAM monotherapy. While sequential therapy has been shown to be more effective than LAM monotherapy, simultaneous therapy has some benefits over sequential therapy – although these are not always significant (61–64).
Peg interferon/nucleos(t)ide analog combinations
In HBeAg-positive patients, end-of-treatment results demonstrate a synergy between pegIFNα-2a and LAM used in combination. Overall, 86% of patients treated with pegIFNα-2a plus LAM for 48 weeks had HBV DNA <100 000 copies at the end of treatment, compared with 52% who received pegIFNα-2a monotherapy (14). Patients receiving combination therapy also showed a more rapid and profound decline in HBV DNA levels compared with LAM monotherapy and pegIFNα-2a monotherapy (mean reduction of 7.2 logs vs. 5.8 and 4.5 logs respectively). Combining pegIFN with LAM can also reduce the development of LAM resistance. Of 256 HBeAg-positive patients treated with pegIFNα-2a plus LAM, nine (4%) developed YMDD mutations during 48 weeks of treatment vs. 69 of 254 patients (27%; P<0.001) who received LAM monotherapy (14). As expected, given the effect on resistance, the proportions of patients in the combination group who had achieved HBeAg seroconversion (27%) and HBV DNA <400 copies/ml (14%) 6 months post-treatment were significantly higher than that in the LAM monotherapy group (19 and 5% respectively). Similarly, in HBeAg-negative patients treated with pegIFNα-2a plus LAM (n=179) or pegIFNα-2a monotherapy (n=181) for 48 weeks, significantly more patients in the combination group had HBV DNA <20 000 copies/ml after 24 weeks of follow-up than those who received LAM alone (44 vs. 29%, P=0.003) (65). There was also a profound reduction in LAM resistance in the combination arm vs. the LAM monotherapy arm (<1 vs. 18%; P<0.001). The end-of-treatment results from these studies therefore clearly demonstrate a benefit of pegIFNα-2a plus LAM vs. monotherapy during treatment. In HBeAg-positive patients, a staggered combination regimen consisting of initial treatment with pegIFNα-2b for 8 weeks, followed by 24 weeks of combination therapy with LAM and LAM monotherapy for a further 28 weeks, resulted in a higher rate of virological response and more substantial reductions of HBV DNA compared with 52 weeks of LAM monotherapy (60 vs. 28% and 3.91 vs. 2.83 log10 copies/ml respectively), although interpretation of these data is complicated by the longer treatment period received by the combination therapy group (60 weeks) (55). However, the staggered combination regimen did result in a lower incidence of LAM-resistant mutants compared with the monotherapy arm (21 vs. 40% respectively). Sarin and colleagues found that when patients were treated with LAM monotherapy for 4 weeks and then switched to pegIFNα-2b for a further 24 weeks, the combined endpoint of HBV DNA suppression and HBeAg clearance and seroconversion at 24 weeks post-treatment was attained by significantly more patients than those who received placebo for 4 weeks before switching to pegIFNα-2b (4/27 vs. 13/36; P=0.086) (66). Although the authors concluded that this approach can improve the efficacy of the IFN and would also reduce the frequency of NA resistance, the use of a combined endpoint in this study is controversial and the small size of this study makes interpretation difficult.
Rudin et al. (67) recently concluded that, in comparable populations, pegIFN monotherapy is equally or more efficacious than LAM plus conventional IFN combination therapy (either simultaneous or sequential). Thus, the optimal regimen for the use IFN-based therapy in combination with NAs has yet to be determined; data are required regarding combination with the newer NAs, and on the use of sequential vs. concomitant therapy, to determine how best to build on the benefits of immune control induced by IFN and to translate on-treatment synergistic effects into a sustained post-treatment advantage.
Nucleos(t)ide analog-based strategies for dealing with lamivudine resistance
Adefovir and TDF are potential candidates as salvage therapy in LAM-resistant CHB. Like LAM, LdT selects for mutations in the YMDD motif, and although it is associated with a lower rate of drug resistance than LAM, the resistance rate is substantial and increases after the first year of treatment particularly in LAM-resistant patients (6). Resistance to ETV, although rare in NA-naïve patients, occurs in around 40% of patients with LAM-resistant YMDD mutants following 4 years of treatment (68).
The majority of reports on salvage therapy for LAM-resistant CHB involve the use of ADV, and ADV is the current standard of care for these patients. ADV-resistant mutations are distinct from those selected by LAM or ETV. However, a number of ADV-resistant mutations also confer a reduction in in vitro LAM susceptibility, although many appear to remain susceptible to ETV, at least in vitro (69–71). The clinical importance of the in vitro cross-resistance observed for ADV-resistant HBV remains to be determined.
Initial studies in patients with HBeAg-positive compensated liver disease and genotypic and clinical resistance showed that ADV alone, or in combination with LAM, resulted in a rapid reduction in serum HBV DNA of 2.5 log10 copies/ml after 4 weeks and 3.59–4.04 log10 copies/ml after 48 weeks of treatment (72). Around half of all patients treated with ADV showed normalization of ALT levels; three patients became HBeAg negative at week 48, and one became HBsAg negative. Other studies have also demonstrated the efficacy of ADV monotherapy and combination therapy in LAM-resistant patients, including those with decompensated disease (73, 74), with similar response rates (75).
Switching LAM-resistant patients to ADV is associated with an up to 21% rate of ADV resistance after 1–2 years (75–79). Adding ADV to LAM appears to minimize the risk of genotypic resistance to ADV, and prevents both virological rebound and clinical resistance over up to 3 years of therapy (79–81). Larrat et al. (82) have demonstrated that this combination can be effective when resistance mutations to both drugs are present, with patients experiencing sustained viral suppression lasting nearly 2 years, while Akyildiz (83) has reported that it is not necessary to continue LAM when adding ADV. Gaia et al. (84) recommend adding ADV to LAM at the first suggestion of resistance, although not all patients see an improvement in efficacy. Overall, treatment guidelines support the addition of ADV to existing LAM therapy, rather than ADV monotherapy in patients with LAM-resistant mutations (6, 7).
Not all LAM-resistant patients respond optimally to ADV or achieve full suppression of HBV replication (75, 78, 85). A number of baseline factors are associated with virological and clinical response in LAM-resistant patients treated with ADV, including infection with HBV genotype D, female gender, HBeAg status, serum HBV DNA levels and ALT levels (74, 86, 87). Recently, it has also been suggested that the YMDD rtL80V/I variant mutation in LAM-resistant patients may be associated with a particularly poor response to subsequent ADV therapy (88). Patients who respond poorly to ADV at the approved dose of 10 mg/day have been found to respond to a dose of 20 mg/day without significant AEs, which suggests that 10 mg/day is insufficient to suppress persistent LAM-resistant strains, possibly due to the selective advantage given to LAM-resistant mutants over wild-type HBV by continued combination therapy (89, 90). Confirmation of these initial findings is required in larger studies before an increase in the dose of ADV can be recommended in clinical practice, especially given the potential risks of renal toxicity at the higher dose.
Despite the low level of ETV resistance in NA-naïve patients, the cumulative probability of virological breakthrough in LAM-resistant patients on ETV therapy is around 10% after 1 year, 25% after 2 years (91) and 39.5% at 4 years (68), with one small study showing an incidence of 75% after a period of over 36 months (92). Switching LAM-resistant patients to ETV at a daily dose of 1 mg rather than the 0.5 mg recommended for naïve patients results in improved histological improvement, viral load reduction and ALT normalization compared with continuing LAM (93, 94). However, a recent study has shown that resistance to ETV arises in a subset of patients harbouring pre-existing LAM-resistant/ETV-resistant variants, with around half of these patients experiencing a virological rebound on ETV (95). This high resistance rate despite an increase in dose should be taken into account when considering switching LAM-resistant patients to ETV. The most recent APASL consensus statement (7) and the EASL (8) and AASLD guidelines (6) identify switching to ETV as an option for LAM-resistant patients where adding ADV is not appropriate.
As it has only recently been approved for use in CHB, data on TDF in LAM-resistant patients are very scarce. In one recent, very small study, six patients with cirrhosis and resistance to LAM and ADV were treated with TDF plus LAM. After 12 months of treatment, all patients had achieved undetectable HBV DNA and four of six had normalized ALT. These results were sustained to the end of this study (mean 16.5 months) (96). In a subanalysis of the pivotal studies comparing the efficacy of ADV and TDF, 90% of patients previously treated with LAM had HBV DNA levels <400 copies/ml after 48 weeks of treatment with TDF, compared with 88% of naïve patients (32). The efficacy of TDF in LAM-experienced patients was durable, with 92% of patients treated for 96 weeks having HBV DNA <400 copies/ml at 96 weeks (97).
Interferon-based strategies for dealing with Lamivudine resistance
In terms of administration, IFNs are not as convenient for the patient as NAs and are associated with a known side effect profile, including flu-like symptoms such as pyrexia, myalgia and fatigue, although the convenience, efficacy and AE profile of IFN have been improved by pegylation. In addition, the tolerability of pegIFN seems to be better in HBV-infected patients compared with HCV-infected patients, with a lower incidence of common IFN-related AEs and a significantly lower incidence of depression (98). Overall, the poorer tolerability profile of IFN compared with NAs needs to be balanced against the greater possibility of achieving an off-treatment response with a finite course of therapy (usually 6–12 months). Nucleos(t)ide analogs need to be given as long-term therapy to ensure constant viral suppression, with an associated increase in the risk of inducing development of drug-resistant variants. In contrast with the need for extended treatment with NAs to ensure constant viral suppression, a limited period of treatment with IFN can induce a sustained immunological response, as illustrated by HBeAg loss/seroconversion and HBsAg seroconversion, in a portion of patients. HBsAg seroconversion is of particular significance because it is associated with improved long-term outcome and is considered to be the closest to cure (99). Thus, HBsAg clearance is often considered an appropriate long-term surrogate endpoint for determining response to therapy. While increasing rates of HBsAg clearance are seen during post-treatment follow-up of patients treated with IFN-based therapy, HBsAg clearance during treatment with NAs is far less common (14, 15).
In a long-term, single-centre study comparing IFN, pegIFN, LAM and ADV in monotherapy and combination, four of 35 patients (including one with LAM resistance) treated with a finite course of IFN/pegIFNα-2a showed clearance of serum HBsAg after a mean follow-up treatment duration of 40 months (range 24–84 months) compared with no patient treated continually with NAs for a similar period (100). A study comparing inactive carriers and patients treated with IFN has demonstrated that IFN treatment accelerates the rate at which the HBsAg concentration declines (101).
Initial studies using pegIFN as follow-on monotherapy or in combination with continued LAM in LAM-resistant patients showed modest efficacy and were limited by their small size (102). More recently, switching from NAs to pegIFNα-2a has been shown to be safe and effective in HBeAg-positive and -negative CHB. In a large retrospective cohort study (PEGaLAM), treatment with pegIFNα-2a was found to be as effective in both HBeAg-positive and HBeAg-negative patients who had received prior NA therapy as in NA-naïve patients in randomized clinical trials (103). In patients who had received prior NA treatment, two thirds of HBeAg-negative and over half of HBeAg-positive patients showed HBV DNA suppression <100 000 copies/ml during treatment with pegIFNα-2a. Around half of all patients also had normal ALT levels at their latest off-treatment visit (103). HBeAg loss and HBeAg seroconversion occurred in around one-third of HBeAg-positive patients, and importantly, HBsAg loss occurred in 7% of HBeAg-positive patients, of whom two developed anti-HBs antibodies (Table 1). These results in NA-pretreated patients were similar to those seen in NA-naïve patients (14, 15). This observation also held true when considering specifically the Asian patients within this cohort (104). A similar pattern was reported for Chinese patients with HBeAg-positive CHB in a population consisting of 255 patients who were treatment naïve or who had failed on LAM or IFN (105).
Table 1. Comparison of efficacy of pegylated interferonα-2a in HBeAg-positive patients with and without pretreatment with lamivudine
In a study of LAM-resistant Chinese patients, 24 weeks of pegIFNα-2a, with an initial 12 weeks of continued therapy with LAM, resulted in sustained HBeAg seroconversion and HBV DNA <105 copies/ml 6 months post-treatment in five (31.2%) patients and HBsAg loss in two (12.5%) patients (106). In a similar study, Chinese patients with LAM-resistant CHB were treated with pegIFNα-2a for 48 weeks (n=155) or ADV for 72 weeks (n=60). All patients received an initial 12-week combination therapy with LAM (107). At week 72, HBeAg seroconversion was achieved in 12% of those in the pegIFNα-2a arm compared with 4% in the ADV arm (P=0.045). The rate of sustained response seen here with pegIFNα-2a appears to be lower than those in treatment-naïve patients; nevertheless, it is worth noting that HBsAg clearance was observed in three of these challenging patients and, importantly, cessation of LAM was not associated with any case of decompensation.
Leemans et al. (108) conducted a retrospective comparative cohort study of HBeAg-positive LAM-resistant patients who were subsequently treated with pegIFNα or pegIFNα plus LAM for at least 20 weeks with at least 16 weeks of follow-up. Among 16 patients who met the inclusion criteria, two (12.5%) achieved HBeAg seroconversion and sustained HBV DNA suppression at 6 months post-treatment. When Inno-Lipa assays for all 16 patients were compared, these two patients had much weaker signals for YMDD mutations than the other patients. The authors concluded that patients with emerging resistance to LAM may gain most benefit from treatment with pegIFN (108).
Taken together, these studies suggest that switching patients to pegIFNα-2a from LAM is safe and beneficial in a significant proportion of patients with LAM resistance, with rates of response similar to those in treatment-naïve patients. Clearly, in a patient with decompensated liver disease in whom viral breakthrough has occurred, NAs rather than IFN-based therapy would be indicated.
Summary and conclusions
The development of resistance to NAs is an issue affecting all currently available NAs to a greater or a lesser extent and is related to the need to maintain viral suppression through prolonged treatment. The reduction of HBV DNA levels leads to an improvement of inflammation and fibrosis scores and lowers the risk for cirrhosis and HCC (109). The continuous suppression of serum HBV DNA to undetectable levels is necessary to reduce the risk of resistance to NAs. In contrast, such compete suppression of HBV DNA levels to undetectable levels is not required with pegIFN therapy, because viral variants with resistance to IFN do not arise. Furthermore, clinical improvements may be achieved with lesser degrees of viral suppression (109, 110). In summary, the current approaches to minimize resistance during HBV therapy are to consider using pegIFN as the first-line approach to achieve a sustained response without the risk of resistance, and to avoid, where possible, first-line use of LAM in favour of more potent NAs with a higher barrier to resistance.
Despite the high risk of drug resistance, LAM remains a first-line strategy in some areas, largely as a result of economic constraints. The selection of patients with baseline characteristics associated with a higher likelihood of response and lower likelihood of resistance development to LAM may improve the overall outcome in such areas. In addition, assessment of the 24-week virological response to NA monotherapy, especially for NAs where the barrier to resistance is low, may also allow individualization of treatment, preventing prolonged treatment where optimal response is unlikely and thus minimizing the likelihood of resistance development. Given that around one-third of patients achieve a sustained response to a finite course of pegIFN, that there are no issues around resistance and that such therapy does not affect the efficacy of subsequent treatment with NAs, pegIFN can and should be considered for first-line therapy. Recent evidence suggests that on-treatment quantification of HBsAg during treatment with pegIFN may be helpful as an early predictor of response to identify those patients most likely to benefit from this approach and this could find clinical utility in patient management.
For those patients who are treated with NAs first line, regular monitoring is required for early detection of emerging mutations. Should such mutations arise, there are two possible options: to add on or switch to a non-cross-resistant NA, or to switch to pegIFN-based therapy, either in combination with a more potent NA or as monotherapy. As the response rates in LAM-resistant patients are similar to those obtained in naïve patients, pegIFN monotherapy for finite treatment duration can be considered as a potential rescue therapy.
The use of combination therapies is attractive, given its success in other viral diseases such as HIV. Although new data are emerging, however, the best use of combination regimens remains to be fully established in HBV infection. To date, combining currently available NAs has not shown any beneficial effects compared with monotherapy, probably reflecting the fact that they are directed at single target: the polymerase. Although theoretically attractive because of their different modes of action, no optimal combination or sequence of treatment has been identified for the combination of newer more potent NAs with pegIFN. Such strategies must await justification from further clinical studies and cost–benefit analyses.