Therapy of hepatitis B — Viral suppression or eradication?


  • Potential conflict of interest: Nothing to report.


The practicing clinician is currently faced with a number of treatment options for chronic hepatitis B. Beginning in 1998 with the licensing of lamivudine and subsequently adefovir, the treatment paradigm shifted from 4 to 6 months of conventional alfa interferon to a year of nucleoside analog therapy. However, prolonged treatment with nucleoside analogs is often needed to optimize virological response. Recently, a 48-week regimen of pegylated interferon for hepatitis B e antigen (HBeAg)-positive and HBeAg-negative chronic hepatitis B has been shown to be effective, and long-term nucleoside analog therapy has been demonstrated to maintain viral suppression. These findings have added to the complexity of decision-making and have raised questions about whether a finite course of pegylated interferon or nucleoside analog therapy, with possible long-term maintenance, is better as first-line therapy. Each of these fundamentally different approaches has advantages and limitations, and both have a place in the therapeutic armamentarium against chronic hepatitis B. Long-term therapy with nucleoside analogs, however, raises a number of practical concerns that have not been fully addressed as of yet. I will present evidence in support of the recommendation that antiviral therapy should ideally be directed toward achieving the highest rate of viral clearance with the shortest interval of treatment. (Hepatology 2006;43:S182–S193.)

Before 1998, the treatment of chronic hepatitis B (CHB) was relatively straightforward as there was only a single licensed drug, interferon alfa-2b. A 4- to 6-month course of interferon resulted in a sustained loss of hepatitis B e antigen (HBeAg) and nondetectable hepatitis B virus (HBV) DNA in approximately 30% of patients.1–3 Further studies demonstrated that response rates could be improved with a longer course (32 weeks), and at least 12 months was often required for the HBeAg-negative form of the disorder.4–6 Posttreatment follow-up in the early clinical trials was generally short (6-9 months), but a number of subsequent studies demonstrated favorable long-term outcomes.7–9 Interferon responses were found to be generally durable, and the rate of hepatitis B surface antigen (HBsAg) seroconversion (a complete response) increased during prolonged follow-up of virological responders. A handful of direct hybridization assays were commercially available to measure serum HBV DNA, and although these assays only detected serum HBV DNA of at least 105 copies/mL, nondetectable HBV DNA during posttreatment follow-up correlated with clinical, histological, and biochemical improvement.

The approach to treatment of CHB, however, has dramatically changed over the past decade. Perhaps the most obvious difference can be found in the number of therapeutic options available. As of January 2006, five approved drugs are available for CHB in the United States, including two formulations of interferon and three oral nucleos(t)ide agents. Pegylated interferon alfa-2a has replaced conventional interferon alfa-2b because of increased antiviral potency, better patient acceptance, and changing drug technology.10 Moreover, a number of drugs have been licensed for HIV-1, such as tenofovir and emtricitabine, which have antiviral activity against HBV.11–13 Virological monitoring has become more complex due to an expanded array of commercially available tests for serum HBV DNA, most often incorporating polymerase chain reaction (PCR). The more sensitive PCR assays can detect 50 HBV genomic copies/mL or less.

The added complexity of treating CHB has led to the development of evidence-based treatment guidelines for the management of CHB.14–16 These documents advocate the use of either interferon or nucleoside analogs as initial therapy and establish alanine aminotransferase (ALT) and HBV DNA thresholds for treatment. For most clinical situations, however, they do not make specific recommendations as to whether interferon or nucleoside analogs should be used as preferred first-line therapy, because both approaches have proved effective, and both have advantages and limitations (Table 1).

Table 1. Advantages and Limitations of Currently Available Antiviral Agents
  • *

    Based on data with conventional as well as pegylated interferon.

  • See Table 3.

InterferonFinite duration of treatmentGiven by injection
 Durable treatment responseFrequent side effects
 Loss of HBsAg (3%-8%*)Expensive
 ImmunomodulatoryUnpredictable immunological effects
 No drug resistanceLower response rate with high-level viremia
  Virological response depends on genotype
Nucleos(t)ideOral deliveryDrug resistance
AnalogsNegligible side effectsLong or indefinite treatment
 Potent inhibition of virus replicationLow rate of HBsAg disappearance
 Less expensive than interferonExpensive when given long-term
  Potential for multidrug resistant organisms when used sequentially

Although interferon responses are often durable, disadvantages to its use include general poorer tolerability when compared with nucleoside analogs, high cost, need for subcutaneous injection, and a potential to precipitate immunological flares and liver failure in patients with advanced liver disease.17 Although nucleoside analogs are extremely well tolerated by comparison, disadvantages to their use include a high rate of relapse on withdrawal, frequent need for prolonged or maintenance therapy, relatively high expense, and drug resistance when used as monotherapy.18, 19 Incremental virological response has been demonstrated with an increasing duration of therapy, but so have increasing rates of viral resistance.20–23 This outcome requires the use of another nucleoside analog that has antiviral activity against both the wild-type and drug-resistant HBV.

The excellent tolerability of nucleoside analogs has made it relatively easy to use them for prolonged periods to maintain viral suppression. Such an effect has important therapeutic implications because prolonged use of one of these agents, lamivudine,has been recently shown to forestall disease progression and reduce the rate of hepatocellular carcinoma.24 One may assume that even greater benefit would be observed with other nucleoside analogs that have a lower rate of drug resistance. These types of findings have led some authorities to recommend long-term treatment with lamivudine or other nucleoside analogs in patients with advanced fibrosis, with the option to switch to or add a second drug should resistance develop.24, 25

Thus, the past decade has witnessed a shift in the treatment paradigm from time-limited therapy with interferon to exploration of short-term nucleoside analog treatment, and now the relatively common practice of long-term HBV suppression with nucleoside analogs for a significant proportion of patients. Unfortunately, long-term therapy with nucleoside analogs raises a number of practical issues for patients and may be associated with a number of clinically important concerns.

Treating clinicians need to understand the level of benefit that can be expected from short-term versus prolonged antiviral therapy and in what clinical situations each of these approaches is feasible or necessary. Answers to these questions require consideration of how the available drugs act to affect viral clearance; how host and viral features can help predict a response to short-term therapy (whether this is interferon or nucleoside analog based); and what the benefits and limitations are of current finite and long-term antiviral treatment regimens. Accordingly, I review current data for each of these issues before making personal recommendations on preferred first-line therapy in various clinical situations.

Response Definition and Relationship to Viral Eradication

When using the term “finite therapy” in this paper, I refer to treatment that is short-term—1 year or less—whereas “prolonged therapy” refers to treatment beyond this interval as long as there is an intention to discontinue treatment when a virological response has been obtained. “Long-term suppression” or “maintenance therapy,” however, will be used to imply that maintaining a viral response requires indefinite therapy, and no clearly identifiable timeline or stopping point exists.

The rationale for a finite course of therapy, whether this involves interferon or nucleoside analogs, is to achieve a durable virological remission off treatment and, by doing so, prevent disease progression. With HBeAg-positive hepatitis B, virological remission is indicated by HBeAg loss with or without seroconversion to anti-HBe as well as low-level or nondetectable HBV DNA by PCR. No uniformly therapeutic endpoint in HBeAg-negative CHB has been agreed upon, and some authorities advocate that using prolonged nondetectability of HBV DNA by PCR may be the best therapeutic endpoint.26

A further treatment objective, which is often not obtainable, but more desirable, is loss of HBsAg or HBsAg seroconversion. Although this event is extremely important, even nondetectable HBsAg does not signify that true viral eradication has occurred or that clinical problems related to HBV will not occur in the future. Several studies have shown that small amounts of HBV DNA can be found in hepatocytes and other tissues years after a patient with CHB has undergone HBsAg seroconversion.27–30 Strong evidence for persistent infection after HBsAg clearance also comes from the well-described transmission of infection from organ donors who are positive for anti-HBc alone,31–33 as well as occasional descriptions of reactivated hepatitis B in HBsAg-negative patients during chemotherapy. Finally, HBV maintains its pro-oncogenic properties in cirrhosis with occult infection.34 Covalently closed circular HBV DNA (cccDNA), viral transcripts, and integrated HBV have been detected in patients with occult HBV infection.29, 30 Thus, eradication of infection is more appropriately reserved for instances in which HBV DNA can no longer be detected in hepatocytes and extrahepatic reservoirs of infection. Currently, no evidence has been found that this process occurs either naturally or through treatment, no matter how prolonged the treatment. Literally, to have hepatitis B may be to have it forever, even though in most cases further disease sequelae do not occur because of tight control of residual virus by an efficient and persistent cellular immune response.35, 36


CHB, chronic hepatitis B; HBeAg, hepatitis B e antigen; HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; PCR, polymerase chain reaction; ALT, alanine aminotransferase; cccDNA, covalently closed circular HBV DNA; ULN, upper limit of normal.

HBsAg Clearance and Reduction in cccDNA During Antiviral Therapy

Although clearance of HBsAg occurs infrequently enough as to not be a practical therapeutic endpoint, it remains an important event because it generally signals persistent loss of serum HBV DNA by PCR and has been shown to lead to improved survival as well as a diminished risk of hepatocellular carcinoma in patients with cirrhosis.37 Also, HBV infection is much less likely to reactivate when patients who have become HBsAg-negative are given chemotherapy.38

Understanding the likelihood of HBsAg conversion with the available treatments for hepatitis B is important. A meta-analysis of interferon trials before 1992 demonstrated that loss of HBsAg occurs 6% more frequently than the rate observed in untreated controls.3 Recent phase III trials of pegylated interferons alfa-2a and 2b for CHB have shown rates of HBsAg seroconversion that vary between 3% and 7% in the interferon treatment arms versus 0% in patients treated with 48 to 52 weeks of lamivudine monotherapy.39–41 Although the rate of HBsAg seroconversion may seem small, it constitutes approximately 10% to 20% of all virological responders who undergo HBeAg seroconversion. The higher rate of HBsAg clearance with interferon is at least partially explainable by the immunoregulatory effects of this drug (vide infra). By contrast, disappearance of HBsAg rarely occurs with nucleoside analogs, and even when these drugs are given for 3 or more years, the rate of HBsAg seroconversion in most studies does not differ from what would be anticipated to occur spontaneously.22, 42

Detection of HBsAg in serum and tissue is attributable to persistent infection of hepatocytes with cccDNA, a unique episomal replicative intermediate responsible for all viral transcripts. This genomic template for viral transcription has been difficult to totally eliminate from hepatocytes with either interferon or nucleoside analogs, and its persistence, generally in low copy numbers, is considered to be instrumental in later reactivation of HBV infection. Using real-time PCR, cccDNA has been detected in hepatocytes at levels ranging over 3 orders of magnitude, and the amount has been demonstrated to vary in different phases of the natural history of CHB.29 Changes in cccDNA level were associated with a reduction in serum HBsAg titer, with the lowest levels of cccDNA being observed in patients who cleared HBsAg (0.002 copies/hepatocyte vs 1.4, 0.01, and 0.02 copies/hepatocyte in HBeAg-positive CHB, HBeAg-negative CHB, and inactive carriers, respectively). Also, in a recent study of pegylated interferon alfa-2b with lamivudine versus lamivudine alone, a log cccDNA decline of −0.80 copies/genome at the end of treatment was seen to be predictive of sustained HBeAg seroconversion.43 This relationship was less evident in patients treated with lamivudine monotherapy and is consistent with in vitro studies indicating that treatment with nucleoside analogs results in a very slow depletion of intracellular pools of cccDNA.44, 45 Taken together, these studies imply that interferon-induced HBsAg seroconversion signifies a greater reduction in hepatocyte cccDNA than does HBeAg seroconversion, and when the latter event results from interferon treatment, it is likely to be associated with a quantitatively greater depletion of intracellular cccDNA as compared with nucleoside analogs. These differences may be explained by the fact that the antiviral activity of interferon is complex, involving multiple immunoregulatory actions as well as the activation of intracellular genes that down-regulate viral growth and affect apoptosis of hepatocytes.46, 47 Apoptosis of infected hepatocytes has been considered to be an important pathway in the elimination of the cccHBV DNA. By contrast, nucleoside analogs have a modest and transient immune-enhancing effect,48, 49 and their action is primarily exerted through inhibition of HBV DNA polymerase function. Further studies are needed to validate these concepts.

Optimizing the Duration of Treatment: Interferon and Nucleoside Analog Therapy

The optimal duration of interferon therapy needed to achieve a durable virological remission is not well established. Studies from North America and Europe have reported HBeAg loss in 30% to 35% of patients who are treated with 4 to 6 months of conventional alpha interferon.1, 2, 9, 50, 51 Moreover, 85% to 95% of responders have remained HBeAg negative during 5 to 10 years of follow-up, and as many as 50% to 70% of sustained responders have lost HBsAg.8, 9 A multicenter trial from Europe demonstrated added benefit of continuing therapy for 32 weeks in instances in which HBeAg remained detectable by the end of 16 weeks and only low levels of HBV DNA (<10 pg/mL) were detected.4 Most studies have shown that responses tend to remain durable in HBeAg-positive CHB. By contrast, relapse has occurred in most patients with HBeAg-negative CHB when conventional interferon is given for as long as 12 months,5 and although higher rates of response are observed with 24 months of treatment, sustained virological responses are only evident in approximately 30% of patients.6

Pegylated interferon alfa-2a recently has been used for 48 weeks in both HBeAg-positive and HBeAg-negative CHB.39, 41 At the end of treatment, HBeAg seroconversion occurred in 27% of patients, and increased to 32% at the end of a 6-month posttreatment observation period.41 The results were less favorable in patients with HBeAg-negative CHB in that serum HBV DNA was nondetectable by PCR in 63% of patients at the end of treatment and in only 19% of patients at the end of a 6-month posttreatment interval.39 In that study, a high rate of virological relapse was also observed when the study's co-primary endpoint of <20,000 copies of serum HBV DNA/mL was used to evaluate response (end of treatment: 81%; end of follow-up: 43%). Based on 24-month treatment data with conventional interferon,6 a higher rate of sustained virological response likely would be observed if pegylated interferon is given for longer than 48 weeks in HBeAg-negative CHB.

Whereas the licensed duration of treatment with pegylated interferon alfa-2a is 48 weeks, several lines of evidence suggest that a shorter course of treatment may be effective in HBeAg-positive CHB. In the phase III clinical trial cited above, less than a −1 log difference in the level of HBV DNA suppression was observed at the 24- and 48-week treatment intervals.41 Also, although the patient populations may not be strictly comparable, the rate of HBeAg seroconversion with pegylated interferon alfa-2a did not differ from the rate of HBeAg loss found in historical cohorts who were treated with shorter courses of conventional alpha interferon.1, 3, 7 Additionally, in a phase II study using the same dose (180 μg) of pegylated interferon alfa-2a for 24 weeks, the rate of sustained HBeAg loss did not differ from that observed in the larger clinical trial using 48 weeks of treatment (33% vs. 32%, respectively).10 Because pegylated interferon alfa-2a has been shown to be associated with significantly greater inhibition of serum HBV DNA than conventional interferon, only marginal benefit may result from extension of treatment beyond 24 weeks in HBeAg-positive CHB. No comparative head-to-head studies directly address this issue.

The optimal duration of nucleoside analog therapy also is not well established, but these drugs often need to be given for periods in excess of 1 year to achieve a durable virological response. In one study, 3-year durable responses were observed in only 64% of patients who achieved HBeAg seroconversion after 52 weeks of lamivudine.52 This result contrasts with those of another study in which more than 80% of patients who had been treated for 24 months or longer after HBeAg seroconversion continued to have a virological response at 2 years posttreatment.53 A Korean study suggested that relapse is significantly less likely if patients remain on treatment with lamivudine for at least 3 months after HBeAg seroconversion is achieved.54 These types of observational studies have led to recommendations to continue treatment for 6 months beyond the point of HBeAg seroconversion.14, 15, 25 Durable responses have been particularly difficult to achieve with prolonged nucleoside analog therapy in HBeAg-negative CHB. In one study, 50% of patients had a virological relapse when withdrawn from lamivudine at treatment month 24 despite non-detectable HBV DNA by PCR for 9 to 12 months while receiving treatment.55 Also, patients who switch from adefovir to placebo after 1 or 2 years of treatment frequently relapse, whereas those who continue to take adefovir do not.22 Even though patients maintained on adefovir for 3 years demonstrated a progressively greater decline in serum HBV DNA level, 21% of these patients still had more than 1,000 copies of serum HBV DNA/mL and would be likely to relapse on discontinuation of treatment. The results of long-term viral suppression with other more potent nucleoside analogs such as entecavir or tenofovir will provide interesting comparisons.

In summary, pegylated interferon has been used successfully for 48 weeks, but further studies appear warranted to address whether shorter courses of treatment might be equally beneficial for HBeAg-positive CHB. By contrast, longer courses of pegylated interferon are likely to be necessary to treat HBeAg-negative CHB. Most HBeAg-positive patients do not have a sustained virological response after 1 year of treatment with nucleoside analogs, and prolonged treatment beyond 1 year often is necessary. Long-term viral suppression with nucleoside analogs is particularly necessary in HBeAg-negative CHB, a disorder in which firm clinical and laboratory endpoints are currently lacking.

Prolonged Treatment: Benefits and Limitations

Interferon has only been given for periods longer than 1 year in HBeAg-negative CHB, and all clinical trials have used conventional interferon.26 In 1 study, 101 consecutive patients with HBeAg-negative CHB were treated with 6 million units of interferon alfa-2b thrice weekly for 24 months.6 Thirty percent had a sustained virological response, 15 cleared HBsAg, and the severity of liver disease improved in responders; moreover, the estimated 8-year complication-free survival was longer for responders. Prolonged interferon therapy is associated with a number of adverse effects, however, and 16% of patients in this study had to discontinue treatment because of adverse effects. No reports have been made of maintenance interferon therapy in either HBeAg-positive or HBeAg-negative CHB.

Prolonged treatment with lamivudine seems to be generally well tolerated, and the range of adverse events does not differ from placebo. In patients with HBeAg-negative CHB, prolonged lamivudine therapy (3.8 ± 1.4 years) has been associated with improved survival and reduction in the risk of liver-related complications that was judged to be equivalent to that observed in sustained responders to interferon therapy.56 In a multinational trial in Asia, 651 patients with advanced fibrosis and viral replication were treated with lamivudine for a median interval of 32 months. Patients treated with lamivudine had significantly less disease progression and a lower rate of primary hepatocellular carcinoma.24 These findings are not unanticipated in view of the strong relationship between persistent viremia, progression to cirrhosis, and development of hepatocellular carcinoma.57–60

A major problem complicating prolonged or long-term therapy with lamivudine, however, is the progressive increase in drug resistance. Resistance has been reported to occur in 23% of HBeAg-positive patients after 1 year of treatment; 46% at year 2, 55% at year 3, 71% at year 4, and 65% at year 5.61 Hepatitis flares have been reported with lamivudine-resistant HBV and ultimately can lead to hepatic decompensation in patients with severe fibrosis.61 Lamivudine resistance has been shown to be associated with a partial loss of clinical and histological benefit when compared with those who maintain wild-type HBV.62, 63 Lamivudine resistance can be successfully treated with adefovir,64, 65 entecavir,66 or interferon.67 However, lamivudine resistance leads to selection pressure for HBV with additional mutations in the DNA polymerase gene that reduces the susceptibility to entecavir.68

Prolonged treatment with adefovir has been reported to be associated with incremental virological response in HBeAg-positive and HBeAg-negative CHB and has been found to be generally safe. Adefovir treatment for 3 years results in an increase in the number of patients with HBeAg seroconversion (12% at year 1, 29% at year 2, and 43% at year 3).21 Three years of continuous treatment also has been associated with a progressive decline in serum HBV DNA and an increase in the percentage of patients becoming HBV DNA negative by PCR in HBeAg-negative CHB.22 Slight elevations in serum creatinine have been seen to occur in less than 2% of patients treated with 10 mg daily for 2 to 3 years and only rarely require discontinuation of drug.22 The resistance profile of adefovir is much better than that of lamivudine, and drug resistance has been observed in approximately 2% of patients at year 2, 6% at year 3, and 18% at year 4.23 Based on in vitro studies and limited clinical data, adefovir-resistant HBV appears to be successfully managed with lamivudine, tenofovir, entecavir, or interferon.69–71 Clinical trials that address the virological efficacy and safety of long-term treatment with entecavir and tenofovir are not yet available. In theory, however, prolonged treatment with these newer more potent nucleoside analogs should lead to minimal rates of drug resistance and a correspondingly higher rate of achieving virological endpoints.

Predictors of Response to Antiviral Therapy

A number of host and viral factors are associated with a higher likelihood of virological response to treatment. These differ slightly for interferon and nucleoside analogs and appear to differ to some extent between conventional and pegylated forms of alpha interferon. They are important to consider because they assist in the determination of whether a patient is likely to benefit from a year of interferon or nucleoside analog, or alternatively, whether several years or indefinite treatment with a nucleoside analog is likely to be required (Table 2).

Table 2. Predictors of Response to Pegylated Interferon and Nucleoside Analog Therapy in HBeAg-Positive CHB
High LikelihoodIntermediate LikelihoodLow Likelihood
  1. Abbreviation: HAI, histological activity index.

Pegylated Interferon  
Baseline ALT > 5 × ULNBaseline ALT > 2-5 × ULNBaseline ALT ≤ 2 × ULN
Baseline HBV DNA ≤ 109 copies/mLBaseline HBV DNA 109-10 copies/mLBaseline HBV DNA > 1010 copies/mL
Genotype A or BGenotype CGenotype D
Nucleoside Analogs  
Baseline ALT > 5 × ULNBaseline ALT > 2-5 × ULNBaseline ALT ≤ 2 × ULN
HAI ≥ 10HAI 5-9HAI 0-4

Baseline serum aminotransferase levels, baseline HBV DNA, degree of histological activity, age at acquisition, and sex all have been found to be predictive of a response to conventional interferon in HBeAg-positive CHB.1, 72, 73 Of these, the strongest correlations are with baseline ALT/aspartate aminotransferase and serum HBV DNA, with ALT values >2 × upper limit of normal (ULN) and HBV DNA < 200 pg/mL (approximately 50 × 106 copies/mL) being more frequently associated with a virological response. Infection with HIV also has been associated with a higher rate of treatment failure.74

Multivariate analysis of HBeAg-positive patients treated with 48 weeks of pegylated interferon alfa-2a have indicated that high ALT (>5 × ULN), low HBV DNA (<109 copies/mL), and low baseline HBeAg concentrations (≤ 186 IU/mL) predict a high rate of HBeAg seroconversion.75 Among patients with low HBV DNA levels, HBeAg seroconversion was observed in more than 50% of patients.75 Similarly high rates of HBeAg seroconversion (41%) were observed in patients with baseline ALT values > 5 × ULN. Pretreatment predictors of response to conventional or pegylated interferon in HBeAg-negative CHB, however, have not been as well defined. Multivariate analysis of a phase III study using pegylated interferon alfa-2a has found that high baseline ALT, low baseline HBV DNA, younger age, and female sex are predictive of a sustained virological response when evaluated at the 24 weeks posttreatment interval.76 Patients with genotype D responded poorly to interferon monotherapy in this study. Longer treatment duration and an early biochemical response to therapy may be associated with a greater likelihood of a sustained virological response.5 Patients who become HBV DNA negative by PCR after 12 weeks of treatment with pegylated interferon alfa-2a have been shown to have a better chance for a sustained virological response.77

Considerable attention has been given recently to the role of HBV genotype in predicting response to alpha interferon. Early studies on the relationship of HBV genotype to conventional interferon response were limited by small sample sizes and primarily focused on patients with HBeAg-positive CHB.78–80 These studies linked genotype B to a higher rate of response than genotype C and suggested that patients with genotype A responded more frequently than those with genotype D. These findings recently have been confirmed in larger studies. In a study in 165 patients (119 with HBeAg-positive CHB and 46 with HBeAg-negative CHB), multivariate analysis indicated that genotype was a strong independent predictor of response.81 HBV genotype A was associated with a significantly higher rate of sustained response than genotype D (49% vs. 26%, P <.005). HBeAg status did not negatively impact on response to genotype A, suggesting that low rates of sustained response with HBeAg-negative CHB in past studies may be explainable by inclusion of a disproportionate number of patients with genotype D.

The relationship between viral genotype and response to pegylated interferon is similar to that observed with conventional interferon. In a study involving 266 HBeAg-positive patients who were treated with 52 weeks of pegylated interferon alfa-2b, individuals with genotype A demonstrated significantly higher rates of response when compared with genotypes B, C, and D (47%, 44%, 28%, 25%, respectively).40 The rate of HBsAg loss also varied according to genotype being observed in 28% of virological responders with genotype A, 20% with genotype B, 0% with genotype C, and 8% with genotype D.82 Data from two phase III trials of pegylated interferon alfa-2a also have indicated a higher rate of virological response in HBeAg-positive and HBeAg-negative CHB when patients with genotype A are compared with those with genotype D, but responses rates in genotypes B and C were equivalent in these studies.39, 41, 75 Whether the discrepancies between the pegylated interferon studies can be attributed to differences in the antiviral potency of the particular interferon preparations, differences in other patient characteristics, or a combination of both factors is not known.

Baseline ALT level is the best independent predictor of HBeAg loss and HBeAg seroconversion in patients treated with nucleoside analogs. This association has been most clearly defined for lamivudine but also applies to adefovir and entecavir83, 84 (Data on File, Gilead Sciences). In contrast to the situation with interferon, baseline HBV DNA level has not been found to be predictive of response to lamivudine, although higher levels of viral replication have been shown to be associated with a higher cumulative rate of viral resistance.85 A lower rate of HBeAg seroconversion (13% vs. 31%) has been demonstrated, however, in entecavir-treated patients with high-level HBV DNA (≥1010 copies/mL) when compared to those with low-level HBV DNA (<108 copies/mL).84 Thus, additional studies with other nucleoside analogs need to be done comparing HBeAg seroconversion rates across various levels of baseline HBV DNA before concluding that virological efficacy is not impacted by pretreatment HBV DNA level. Several studies have shown that virological response to treatment with adefovir or entecavir is not affected by viral genotype.86, 87

Cost of Care: Short-Term Versus Long-Term

One of the important practical issues that needs to be addressed when evaluating how to approach the treatment of CHB is the long-term cost of care. In addition to the impact on the financial feasibility of providing care for an individual patient long-term cost of care may also influence institutional or regional health care policy and it has become an increasingly important consideration in countries in which drug costs are partially or totally subsidized by the government. In this situation, the choices of drugs are often fewer, and cost, rather than the degree of virological suppression or resistance profile, may be the major determinant of whether a drug is used at all or approved as first-line therapy.

A number of cost-effectiveness analyses of various forms of treatment for CHB have been reported. Early studies looked at the cost-effectiveness of conventional interferon alfa-2b when compared with a “do nothing” strategy and found that treatment should prolong life and lower costs for patients with HBeAg-positive and HBeAg-negative CHB.88, 89 Subsequent analyses from the perspective of a third-party payer concluded that lamivudine is more cost-effective than interferon because it allows more patients to be successfully treated within a fixed budget.90, 91 Interferon is definitely a more expensive therapeutic option when treatment cost for 1 year is compared with nucleoside analog therapy for 2 to 3 years (depending to some extent on the nucleoside analog). However, when one considers the use of nucleoside analog therapy beyond 2 years, the cost can easily exceed the cost of interferon given for 1 year (Table 3). A cost–utility study recently compared the cost of using interferon, lamivudine, or adefovir as first-line therapy versus lamivudine with a switchover to adefovir if resistance to the former occurred.92 As with all such studies, there were limitations from the assumptions made in that pegylated interferon was not assessed and the model required the presence of elevated ALT levels and the absence of cirrhosis. Nonetheless, this study found that neither lamivudine nor adefovir monotherapy is cost-effective as initial therapy because of the high rate of resistance with lamivudine and the long length of treatment that is generally necessary with these agents. Additional findings were that reserving adefovir for lamivudine-associated viral resistance may be highly cost-effective, and interferon therapy was particularly cost-effective in HBeAg-negative CHB. Although controversial, these data provide an impetus to consider that specific therapies for certain patient subsets can be justified economically as well as from a virological efficacy standpoint.

Table 3. Cost of Currently Available Antivirals for the Treatment of Hepatitis B
 Cost Per DoseMonthly CostYearly Cost
  • NOTE. Cost in USD and discounted as advertised in

  • *

    Based on 52 weeks of treatment.

  • Based on 48 weeks of treatment.

  • Not licensed for treatment of HBV.

Lamivudine (100 mg)6.49194.932,368.85*
Adefovir (10 mg)19.34580.206,498.24
Tenofovir (300 mg)16.23486.905,453.28,
Entecavir (0.5 mg)24.21726.308,134.56
Pegylated interferon alfa-2a361.211,444.8417,338.08
Recombinant interferon alfa-2b (10 million unit dose)120.961,451.5217,418.24

Treatment Advisements: Which Way to Go?

Making a single treatment recommendation that applies to all patients with CHB is not possible, nor is it likely to become possible because of the complexity of viral–host interactions in this disorder. Instead, the key to successful treatment of hepatitis B resides in careful patient selection and individualized treatment decisions. Certain assumptions can be made, however. First, whenever possible, aiming to achieve a durable virological response with the shortest course of therapy, thus minimizing the potential for adverse events and limiting costs, is desirable. Second, a chosen therapy should provide the least likelihood of promoting drug resistance or drug failure if a second drug ultimately has to be used. Third, attempting to achieve HBsAg seroconversion is worthwhile because it is associated with greater certainty of long-term benefit. Fourth, treatment decisions should not be made solely on projections about virological efficacy without consideration of a number of patient-specific features that determine feasibility of administering a particular drug. Factors such as drug cost, insurance limitations, the practicality of intensive versus a more relaxed need for monitoring, as well as other comorbid disorders and the potential to tolerate interferon side effects all require evaluation and can be key in making a decision as to which approach to use (Fig. 1).

Figure 1.

Factors involved in making a decision on choice of drug for first-line therapy in chronic hepatitis B. Each of the factors seen here may have to be weighed against projections for treatment efficacy when determining whether treatment with interferon or a nucleoside analog is feasible or well suited for a patient. Even greater importance of these factors is assumed when strong predictors of a virological response are absent. *Particularly pre-therapy HBV DNA level and genotype if interferon is being considered. **Includes time off from work and travel to and from medical center. ¥Includes prevention of disease progression, diminished risk for hepatocellular carcinoma, and transplant-free survival.

Seen in this way and based on the available data, a 48-week course of pegylated interferon should be the preferable first-line approach for several patient groups with HBeAg-positive and HBeAg-negative CHB (Table 4). Particular emphasis should be given to selection of individuals with genotype A or B, low-to-moderate levels of HBV DNA (≤109 copies/mL), moderate ALT (<2-3 × ULN), and otherwise healthy young to middle-aged persons in whom interferon-related adverse effects are unlikely to become a major issue. The data with regard to genotype C are conflicting, and individualized decisions about the appropriateness of interferon in these patients should be based on the age of the patient, extent of disease, social situational factors as listed, as well as predictors of response. Whereas only 5% to 7% or less of interferon-treated patients achieve HBsAg seroconversion with currently recommended regimens, this subgroup constitutes as much as 20% of all virological responders, and the percentage of those undergoing HBsAg seroconversion can be expected to increase with time off treatment.

Table 4. Preferred Initial Treatment Strategies According to Various Patient Features
InterferonNucleoside Analogs
  1. *Consideration can be given for nucleoside analog therapy in cases of severe acute hepatitis B with protracted viral replication (10-12 weeks after onset of illness) and treatment of pregnant mothers with high level serum HBV DNA (>1010 copies/mL) during the last trimester of pregnancy.

Age < 60, otherwise healthyAny age adult, nonserious comorbid illness
Baseline HBV DNA ≤ 109 copies/mLBaseline HBV DNA > 1010 copies/mL
Baseline ALT > 2-3 × ULNBaseline ALT > 5 × ULN
Genotype A or BAny genotype
No cirrhosisCirrhosis, with or without decompensation
 HBsAg-positive chemotherapy

A commitment to nucleoside analogs is likely to be a long-term commitment for many patients, and there is often reluctance to withdraw therapy as long as clinical benefit is maintained. Also, long-term monotherapy ultimately results in drug resistance to a variable extent. A switch in therapy can salvage patients, but this approach could potentially lead to multidrug-resistant HBV, and maintaining both drugs leads to incremental costs.70, 93, 94 These concerns should not be interpreted to mean that there is only a limited role for primary therapy with nucleoside analogs (Table 4). Based on data achieved with lamivudine and entecavir, a greater than 50% rate of HBeAg seroconversion may be expected when baseline ALT is greater than 5 × ULN, and the primary use of these agents in this circumstance will be better tolerated and more cost-effective than interferon.83, 84 Because response to a nucleoside analog appears less likely to be dependent on the baseline level of serum HBV DNA, patients with high levels of HBV DNA (>1010 copies/mL) are probably better treated with nucleoside analog therapy rather than interferon in most clinical circumstances (Table 4). In other subgroups of patients, for example, those with hepatic decompensation, nucleoside analog is the only safe decision that can be made. Nucleoside analogs often lead to clinical stabilization in these patients and have been shown to be life-saving in some cases.95 Furthermore, reasonable evidence suggests that cirrhosis can be safely managed with long-term nucleoside analog therapy with the expectation that it will forestall disease progression and reduce the frequency of hepatocellular carcinoma.

Many patients do not fall into easily decipherable subpopulations as described (for example, HBV DNA >109 but < 1010 copies/mL and ALT 1-2 × ULN). In the author's opinion, whenever strong predictors of response are absent, patient age, liver disease severity, genotype, and the other patient-specific factors listed in this paper assume greater importance in making individualized decisions about whether to proceed with interferon or nucleoside analog–based therapy.


Irrespective of the drug used to treat CHB, a sustained virological response is associated with improved biochemical, histological, and clinical outcomes. Decisions on which drug to choose as first-line therapy should not be primarily dictated by patient (or physician) convenience and acceptance; nor is it appropriate to provide undue emphasis on factors of cost. Instead, primary consideration should be given to the prospect and likelihood for improved long-term outcomes. Interferon responses tend to be durable, and there is a small but definite chance of HBsAg seroconversion, which is of major importance in younger individuals with many years of disease activity ahead. Interferon offers the additional advantage of not promoting HBV resistance. The treating clinician should recognize that treatment with a nucleoside often involves a long-term commitment, and stopping therapy may be difficult if clinical benefit is observed without achieving a firm virological endpoint. Any cost advantage to the use of these agents becomes a moot point in this circumstance.

Pegylated interferon is a particularly good choice as first-line therapy for genotype A and B patients with mild to moderate levels of serum HBV DNA and moderately elevated ALT at baseline. Tolerability remains something of an issue with interferon; however, personal experience has taught me that if a prescribing physician is enthusiastic about the use of interferon, patients are apt to be as well. The frequency of depression using pegylated interferon alfa-2a has been reported to be much lower than that historically observed with chronic hepatitis C (5% vs. 16%-20%, respectively).39, 41, 96, 97 Although the reasons for these differences have not been defined, they may reflect differences in the baseline psychiatric profiles of hepatitis B and C patients rather than in the instruments used to measure depression.

Should a virological response not be achieved with pegylated interferon, nucleoside analog therapy can be used without increasing the probability for later drug failure. The same cannot be said for sequential nucleoside analog therapy, because selection pressures for drug-resistant HBV mutants during treatment with one drug (for example, lamivudine) may facilitate resistance to another (for example, entecavir). Patients in whom nucleoside analogs are particularly well suited for first-line therapy include individuals with high baseline ALT, high levels of serum HBV DNA; patients for whom chemotherapy is planned; decompensated cirrhosis; and instances of HBeAg-negative chronic hepatitis B with low levels of serum HBV DNA, in which long-term therapy is less likely to be associated with viral resistance. Although the data are extremely limited, use of nucleoside analog therapy to treat HBsAg-positive mothers with high-level viremia (>1010 copies/mL) during the last trimester of pregnancy as an adjunct to vaccination98, 99 or severe acute hepatitis with protracted viral replication100 may be appropriate. Many of these clinical situations would only require short-term (3-6 months) nucleoside analog therapy, and, therefore, are not likely to not be associated with viral resistance. Lamivudine is a relatively good choice in these situations because of moderately high antiviral potency, proven safety record, and greatly reduced cost compared with other agents.

The problem of viral resistance to nucleoside analogs will continue to be vexing as long as monotherapy is used. Although experience with HIV infection has indicated that resistance can be overcome by using two drugs initially, this substantially increases the cost of care. Recent studies have shown that combining pegylated interferon with lamivudine results in greater on-treatment HBV DNA suppression and lower rates of lamivudine resistance, but the former outcome is not associated with a higher rate of sustained virological response when compared with pegylated interferon alone.39–41 In theory, drugs that work through different mechanisms, such as pegylated interferon and nucleoside analogs, should have additive effects against HBV. This seeming paradox may be explained by deficiencies in the design of some of these studies.

Although a considerable amount of progress has been made in the last decade with regard to antiviral therapy of CHB, there is still a great unmet need for a drug therapy that results in a durable virological remission in most patients after a finite period of treatment. Until this is achievable, individualized treatment decisions will remain key to maximizing efficacy, and chronic HBV infection will continue to be treated primarily as a liver condition rather than as an infectious disease.

In the next several years, several other nucleos(t)ide analogs will become licensed for hepatitis B, such as telbivudine and tenofovir. Two properties will characterize these newer drugs: higher antiviral potency than current agents and low rates of drug resistance. Despite these advantages, the available data suggest that these drugs will have to be given for prolonged intervals as well and resistance will ultimately emerge to some degree. Relatively good evidence indicates that combinations of nucleoside analogs will prevent problems with resistance, as already has been demonstrated for HIV infection. Thus, one of the challenges for the future will be to decide in what type of patient is combination therapy indicated as an initial approach. One consideration would be to use combination treatment in patients for whom resistance would have greater clinical implications, for example, patients with active cirrhosis or those with decompensated disease. Future studies will also clarify the role for combination therapy with pegylated interferon and a nucleoside analog. This seems warranted because all of the studies in this area have shown a greater degree of viral suppression during combination treatment when compared with either agent alone. Thus, by the end of the decade the treating physician will have a greatly expanded array of therapeutic choices and management decisions to make.