Chapter  174:  Management of Chronic Hepatitis B

Key Questions

Consensus conference question 1. What is the natural history of Hepatitis B?

EPC question 1. What is the evidence that the following population characteristics or clinical features associated with hepatitis B are predictive of hepatocellular carcinoma, liver failure, cirrhosis, liver-related death, and all-cause mortality?

Consensus conference question 2. What are the benefits and risks of the current therapeutic options for hepatitis B with defined or continuous courses of treatment?

EPC question 2a. What is the efficacy (or effectiveness) of interferon therapy, oral therapy, and various combinations in treating hepatitis B with defined or continuous courses of treatment?

EPC question 2b. What are the known harms of interferon therapy, oral therapy, and various combinations in treating hepatitis B with defined or continuous courses of treatment?

Surrogate outcomes of interest. Alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) levels, HBV viral load, change in Hepatitis B e antigen (HBeAg) status, hepatitis B surface antigen (HBsAg) conversion, liver biopsy findings (necroinflammatory activity or stage of fibrosis), and drug resistance.

Clinical outcomes of interest include hepatocellular carcinoma, liver failure, cirrhosis, liver-related death, all-cause mortality.

Consensus conference question 3. Which persons with hepatitis B should be treated?

EPC question 3a. Are there differences in efficacy/effectiveness of treatments for treatment naïve versus drug-resistant patients, chronic HBeAg-positive versus HBeAg-negative patients, or for other subpopulations (as defined previously)?

EPC question 3b. Is there evidence that specific subpopulations do not require treatment for hepatitis B (i.e., that the surrogate and/or clinical outcomes are equivalent or superior when not exposed to treatment?)

Consensus conference question 4. What measures are appropriate to monitor therapy and assess outcomes?

EPC question 4. What is the evidence that changes in surrogate endpoints in response to treatment are reliable predictors of long-term resolution or slowed progression of disease? Patient Population: Adults (≥ 18 years of age), including elderly and members of racial/ethnic minority populations.

EPC Question 1. What is the evidence that the following population characteristics or clinical features associated with hepatitis B are predictive of hepatocellular carcinoma, liver failure, cirrhosis, liver-related death, and all-cause mortality?

Forty-one articles met inclusion criteria,¹⁹–⁵⁹ including 14 publications representing eight unique populations within the United States.

Chronic carriers of HBsAg had substantially higher rates of hepatocellular carcinoma, cirrhosis, and death than people who have never been chronically HBsAg-positive.^{21 24 30 39 45} The annual incidence of hepatocellular carcinoma (HCC) was only 0.1 percent in asymptomatic HBsAg individuals, 1 percent in patients with CHB, but increased to 3–10 percent in patients with cirrhosis.⁶⁰ Patients with CHB developed cirrhosis at a rate of 2 percent per year. Reports have shown large differences in clinical event rates across diagnostic groups such as inactive HBsAg carriers, CHB without cirrhosis, and CHB with cirrhosis. A U.S. cohort study followed 400 HBsAg patients (70 percent born in Asia) for over 7 years.⁴⁷ Among 110 inactive carriers, none developed HCC or died of a liver-related disease, and only one died of any cause. Among patients with CHB but no cirrhosis, 6 percent developed HCC and died from it, while another 2 percent died from nonliver related causes. Among those with CHB and cirrhosis, 16 percent were diagnosed with HCC and 42 percent died during followup (all from liver-related causes).

Increased age was generally associated with small to moderately increased clinical outcomes; however, the evidence was inconclusive regarding whether the association between age and clinical outcomes is explained by duration of infection, age of infection, comorbidities in older individuals, and other factors that might be different between older and younger patients. Likewise, there was inconclusive evidence that geographic location or race/ethnicity contribute meaningfully for the prediction of clinical outcomes. There was high confidence that males have greater than twofold increased rates of clinical outcomes compared to women. A positive family history of HCC was associated with an increased risk of HCC, but the extent this was independent of age of infection and duration of disease is unclear. Estimates regarding coinfection and clinical outcomes could only be made with low confidence due to the paucity or inconsistency of the data; coinfection with either human immunodeficiency virus (HIV) or hepatitis delta virus (HDV) appeared associated with strongly increased liver-related mortality, and coinfection with hepatitis C virus (HCV) appeared associated with moderately increased HCC risk. Cirrhosis is a strong predictor of HCC and death. There was little to no evidence regarding the impact of nonalcoholic liver disease or alcohol consumption on future development of cirrhosis, HCC, or death.

Increased HBV deoxyribonucleic acid (DNA) viral load was strongly associated with increased HCC and liver-related mortality after accounting for baseline cirrhosis, HBeAg status, and ALT levels. There was no evidence regarding whether reduction in HBV DNA viral load was associated with better outcomes. HBV genotypes may be associated with differing risk of clinical outcomes. HBsAg loss was associated with a reduction in risk of cirrhosis, but data were sparse. There was no evidence as to whether HBsAg loss was associated with other improved outcomes. HBeAg-positive status was associated with poorer outcomes independent of other disease factors. Reversion or multiple switches in HBeAg status was associated with increased HCC; however, the mechanism of this is unclear. Basal core promoter mutations (T1762/A1764) and the precore (PC) mutation (A1896) were associated with increased HCC and basal core promoter mutations may be associated with small increases in liver-related death rates. ALT was modestly associated with associated with increased risk of HCC and cirrhosis after accounting for baseline cirrhosis, HBeAg status and HBV viral load.

Questions 2 and 3

Ninety-three articles represented 60 unique randomized trials of interferon alfa-2b,⁶¹–⁹² peginterferon alfa-2a,⁹³–⁹⁷ peginterferon alfa-2b,⁹⁸–¹⁰⁹ adefovir,^{10, 110}–¹²⁰ entecavir,¹²¹–¹²⁶ lamivudine,^{64, 67, 95, 96, 119, 127}–¹⁴² or telbivudine.^{109, 120, 127, 143} Treatment duration averaged 44±22 weeks and followup post-treatment 98±158. Most enrollees were Asian (64 percent) or white (30 percent) ethnicity/race. ^{61 63, 66, 69 81, 83, 84, 86, 87, 90,}

Sixteen articles reporting on mortality, HCC, hepatic decompensation, or cirrhosis were not of sufficient size or duration to adequately assess the effect of treatments on these outcomes.^{70, 83, 85, 86, 90, 91, 96, 106, 111, 121, 122, 124}–^{126, 132, 141} Most studies reported on serologic, virologic, or histologic outcomes with marked variation in patients enrolled, dose or duration of interventions and comparators, time to evaluate outcomes at the end of or at followup off therapies, and definitions of outcomes. When treatment effects were noted, they were rarely reassessed or reported in similar patient populations, and/or drug combinations, doses, or durations. No study assessed outcomes according to the multiple patient and disease characteristics frequently used to determine treatment strategies (e.g., according to HBeAg plus HBV DNA plus ALT plus cirrhosis status). There was a low level of evidence from individual studies or inconsistent results from several studies for most outcomes.

Question 2a. What is the efficacy (or effectiveness) of interferon therapy, oral therapy, and various combinations in treating hepatitis B with defined or continuous courses of treatment?

Question 2b. What are the known harms of interferon therapy, oral therapy, and various combinations in treating hepatitis B with defined or continuous courses of treatment?

Nucleos(t)ide analogues were well tolerated during the duration studied with safety profiles and withdrawal comparable to placebo. Adverse events were usually mild, including fatigue, headache, abdominal pain, nausea, and diarrhea. Pegylated interferon therapy, alone or combined with lamivudine, was not as well tolerated as lamivudine monotherapy. Subjects treated with combined or monotherapy were more likely to withdraw from a study or require dose modification due to an adverse event compared to lamivudine. Adverse events associated with pegylated interferon include flu-like illness, hair loss, anorexia, and less commonly depression. Pegylated interferon and conventional interferon therapy had comparable safety profiles.

Similar incidences of Grade 3 or 4 laboratory abnormalities were observed for adefovir and placebo with the exception of increases in ALT and AST levels. Subjects with or at risk of impaired renal function may develop nephrotoxicity with adefovir. Twenty-five percent of lamivudine subjects had an ALT level at least three times the baseline level compared to 8 percent of placebo subjects during the post-treatment period. One trial noted greater incidences in Grade 1–4 creatine kinase (CK) elevations with telbivudine compared to lamivudine. Higher frequencies of Grade 3–4 elevations in ALT and AST occurred with lamivudine compared to telbivudine. ALT flares occurred in 24 percent and 9 percent of the lamivudine and entecavir groups, respectively. Laboratory abnormalities were higher in the peginterferon alfa-2a monotherapy and combined therapy groups compared to lamivudine. Overall, dose modification, due mainly to laboratory abnormalities, was required for 46 and 47 percent of peginterferon mono and combined therapy recipients, respectively. Neutropenia and thrombocytopenia were cited as the most common abnormalities.

Question 3a. Are there differences in efficacy/effectiveness of treatments for treatment naïve versus drug-resistant patients, HBeAg-positive versus HBeAg-negative patients, or for other subpopulations (as defined previously)?

Potential modifiers of treatment effectiveness and harm include patient, disease, viral, biochemical and therapeutic factors. Fifteen studies examined treatment effects among patient subpopulations immediately^{61, 64, 75, 93, 105, 107, 124, 127, 130, 132, 133, 135, 140, 141, 143} and at followup off active drugs (n=23 studies).⁶¹–^{65, 72, 73, 75, 84, 85, 90, 93, 96, 97, 99, 100, 102, 104, 106, 108, 109, 114, 126} No RCTs directly compared patients with eAg+ versus eAg-, treatment naive versus prior treated, or drug resistant with baseline cirrhosis versus no-cirrhosis. Results from studies enrolling relatively pure populations indicate that there is inconsistent data that baseline treatment status, eAg status, or cirrhosis influence histological, virological, or biochemical end points.

Younger patient age was associated with enhanced HBV DNA clearance and ALT normalization in patients treated with pegylated interferon versus lamivudine.^{93, 109}

Baseline body weight was not associated with HBV DNA clearance and ALT normalization.⁹³

Disease progression or treatment induced sustained ALT normalization and HBV DNA clearance did not vary by gender (five studies, three antiviral agents used as monotherapy).^{72, 93, 109, 132, 141}

Patients with longer duration of hepatitis responded to therapy 2.5 times less frequently compared to those with shorter duration of the disease. Sustained virologic response at 48 weeks off therapy (HBeAg and HBV DNA loss) to interferon alfa-2b combined with lamivudine was greater in those with an estimated duration of hepatitis of 10 years or less after adjustment for patient gender and age.⁶³

Treatment induced followup histology, HBeAg loss or DNA clearance and ALT normalization did not clearly vary by baseline histology severity.^{63, 64, 97, 127} HBeAg loss was higher per unit increase in baseline histological activity index (HAI) score.⁶⁴ Lamivudine improved histology compared to placebo among patients with moderate or severe hepatitis but failed in those with mild hepatitis.¹⁴⁵ Interferon alfa-2b increased post-treatment HBeAg loss compared to placebo among patients with pretreatment HAI score 5–9 but not in patients with pretreatment HAI score 0–4 or >10.⁶⁴ Interferon alfa-2b combined with lamivudine compared to placebo increased post-treatment HBeAg clearance in patients with pretreatment HAI score 5–9 with no significant effects in those with pretreatment HAI score 0–4 or >10.⁶⁴ Off treatment virologic response to interferon alfa-2b plus lamivudine increased in those with a baseline inflammation score of seven or more, independent of gender and age.⁶³ Presence of steatosis did not modify the effect of peginterferon alfa-2a combined with lamivudine on post-treatment response defined as HBV DNA disappearance and ALT normalization in both HBeAg-positive and negative patients.⁹⁷ Adjusted rates of post-treatment response were greater per increase in baseline Knodell HAI.⁹⁷

It was difficult to draw conclusions on the effect of viral load on outcomes off therapy due to varying assays and cut offs of baseline DNA. There were inconsistent effects with no dose-response relationship observed. Compared to placebo, lamivudine reduced disease progression regardless of baseline viral load. Compared to lamivudine, peginterferon plus lamivudine was more effective for combined end points regardless of baseline viral load. No studies reported subgroups with very low viral load. Treatment induced HBeAg loss, ALT normalization, or histology improvement varied with baseline viral load. At followup post treatment, interferon alfa-2b increased loss of HBV DNA and HBeAg among patients with baseline HBV DNA 2-99 pg/ml but failed among those with higher baseline HBV DNA.⁸⁴ There was not a significant HBV DNA unit dose-response versus no treatment.⁶¹ Interferon alfa-2b increased off treatment rates of HBeAg loss among patients with baseline HBV DNA <10pg/ml but not in those with higher viral loads.⁶¹ Interferon alfa-2b with steroid pretreatment increased post-treatment treatment rates of HBV and HBeAg loss among patients with baseline HBV DNA 2-99 pg/ml but failed in those with HBV DNA >100 pg/ml.⁸⁴ Combined administration of interferon alfa-2b with lamivudine resulted in greater off treatment HBV DNA clearance and HBeAg seroconversion in patients with baseline HBV DNA >10⁷ copies/mL.⁷⁵ Peginterferon alfa 2-a provided greater sustained response compared to lamivudine in patients with baseline HBV DNA range in the 25–75 percentile^{93, 96} with random differences at other percentiles.

Low quality evidence indicates that treatment effects may vary by baseline HBeAg status.^{124, 132, 143} Lamivudine versus placebo decreased overall disease progression among HBeAg-positive¹³² but failed in HBeAg-negative patients.¹³² Telbivudine versus lamivudine improved outcomes among HBeAg-positive with random differences in HBeAg-negative patients.^{109, 143} Patients who were HBeAg-negative at baseline experienced improvement in biochemical, virological, and histological outcomes after adefovir therapy and pegylated interferon alfa 2-a monotherapy or combination with lamivudine.^{10, 71, 74, 76, 79, 81, 91, 93, 95, 110, 111} Adefovir^{10, 110} and pegylated interferon alfa 2-a with lamivudine⁹⁵ improved off-treatment viral clearance in HBeAg-negative patients.

Treatment induced ALT normalization and HBV DNA clearance or HBeAg seroconversion varied by HBV DNA genotype. There was better response among patients with genotype B and C at the end of treatments⁹³ and at followup off therapies.^{65, 93, 96, 99, 100, 108, 109} Patients with genotype A had lower adjusted odds of response compared to patients with genotype C.⁹³ Off treatment response to the same treatments also differed with greater adjusted odds of success among patients with genotype B versus D and with genotype C versus D.⁹³

Baseline ALT levels. Treatment induced HBeAg clearance and seroconversion, HBeAg loss or virologic clearance varied by baseline ALT levels with inconsistent evidence of better response among patients with elevated baseline ALT (ten studies; three medications used as mono or combination therapy).

HBeAg seroconversion after peginterferon alfa-2a alone or in combination with lamivudine was higher versus lamivudine alone among patients naïve to lamivudine,⁹⁶ with no significant differences among patients previously treated with lamivudine. Five RCTs enrolled lamivudine resistant patients.^{118, 119, 124, 125, 141} Adefovir plus lamivudine versus lamivudine increased ALT normalization and HBV DNA clearance but not HBeAg clearance or seroconversion in lamivudine-resistant patients¹¹⁹ without improvement in outcomes compared to adefovir monotherapy.¹¹⁹ Entecavir increased HBV DNA and HBeAg clearance and normalization of ALT in lamivudine-refractory HBeAg-positive patients compared to lamivudine^{125, 124} and improved necroinflammatory Knodell scores and Ishak fibrosis scores in lamivudine resistant patients.¹²⁵ Patients who failed previous interferon therapy did not benefit from adding lamivudine.⁹²

Question 3b. Is there evidence that specific subpopulations do not require treatment for hepatitis B (i.e., that the surrogate and/or clinical outcomes are equivalent or superior when not exposed to treatment?)

Studies did not demonstrate improvement in clinical outcomes. However, RCTs were not adequately designed to accurately assess clinical outcomes. Evidence in key question 1 demonstrates that the clinical course of CHB is asymptomatic and indolent in most adults. Therefore, the majority would be unlikely to benefit from treatment for many years. Treatment to reduce viral transmissibility is of potential immediate and long-term public health benefit. Patient, disease, and comorbidity factors are of limited value in assessing prognosis in order to make treatment decisions in an individual patient. A key exception is the presence of cirrhosis where there was high confidence that this led to a large increased risk of poor clinical outcomes. Therefore, clinicians may decide to initiate therapy in these individuals because of a poor natural history.

Specific subpopulations would not require treatment if their clinical outcomes (and possibly validly defined surrogate measures) were equivalent or superior to similar populations not receiving treatment or if harms of therapy outweighed benefits. The effects of eligible drugs on asymptomatic carriers have not been published in RCTs. Monotherapy with interferon alfa-2b or lamivudine and a combination of interferon alfa-2b with steroids failed to sustain virologic response in patients with CHB. Individuals who failed previous interferon alfa-2b therapy did not benefit after combined interferon and lamivudine treatment. Patients with HBeAg did not experience greater off treatment HBeAg seroconversion after interferon alfa-2b combined with lamivudine. Interferon alfa-2b did not improve histology or increase rates of resolved hepatitis.⁸²–^{84, 89, 91} Interferon alfa-2b combined with lamivudine compared to placebo failed to increase HBeAg clearance or sustained HBeAg seroconversion in patients treated with lamivudine⁶⁴ and in nonresponders to the previous interferon therapy.⁶⁷ Lamivudine compared to placebo failed to sustain HBeAg seroconversion in interferon nonresponders⁶⁷ and in treatment naïve patients.¹³⁶ Lamivudine did not sustain HBsAg loss, HBV DNA clearance, or ALT normalization.^{67, 136, 139}

We assessed whether certain patient or hepatitis characteristics were associated with risk of serious adverse events or noncompliance that might lead to a decision not to initiate treatment. Few data were available. Several adverse effects were specific for patients with different HBeAg baseline status. Only HBeAg-negative patients experienced dose modification due to neutropenia or thrombocytopenia.⁹⁵ Combined therapy did not prevent worsening of fibrosis scores in HBeAg-negative patients.⁹⁵ In HBeAg-positive patients depression, diarrhea, dizziness, nausea, pruritus, rash, or rigors were more common after combined therapy with lamivudine compared to lamivudine alone.⁹⁶ YMDD mutations were more common in HBeAg-positive patients after combined therapy compared to peginterferon alfa-2a alone.⁹⁶ Pyrexia was more prevalent after peginterferon alfa-2a compared to lamivudine.^{95, 96} Only HBeAg-positive at baseline patients experienced ≥ 1 serious adverse event,⁹⁶ while only HBeAg-negative patients needed dose modification due to neutropenia or thrombocytopenia.⁹⁵ The rates of YMDD mutations were lower after interferon compared to lamivudine in patients with HBeAg-positive CHB.⁹⁶

Question 4: What is the evidence that changes in surrogate endpoints in response to treatment are reliable predictors of long-term resolution or slowed progression of disease?

Studies were not adequately designed to assess the effectiveness of treatments on clinical outcomes, a necessary prerequisite for determining surrogates. Treatments did not improve all-cause mortality, liver-related death, hepatic carcinoma, or hepatic decompensation. We found even fewer studies that assessed the association of baseline ‘surrogates’ with clinical outcomes. We did not find any RCTs that evaluated whether change in a clinical outcome was explained by a treatment related change in a potential surrogate. We found associations of intermediate markers with clinical outcomes and advise caution against calling them surrogates. Four included studies were either long-term followup of prior RCTs, with randomization no longer preserved, or cohort studies of once-treated patients, where surrogate markers were assessed in relation to long-term clinical outcomes. There was lack of uniformity in surrogate and endpoint measurement, timing of measurement, definitions, and measurement of effect controlling for relevant effect. We have low confidence whether any of the listed biochemical, histologic, or virologic measures are adequate surrogate markers. Patients who are positive for HBsAg are considered to be capable of transmitting hepatitis B virus to uninfected individuals. Clearance of HBsAg, HBV DNA, or HBeAg seroconversion could be considered an appropriate clinical outcome from the perspective of transmission prevention and public health rather than or in addition to possibly being a surrogate for clinical outcomes in infected patients.

There is limited information on the association of potential surrogates of ALT normalization, detectable HBV DNA, worsening histology, and change in HBeAg on the composite endpoint of decompensation, cirrhosis and HCC, and all-cause mortality among patients treated with peginterferon alpha-2a plus lamivudine, interferon alpha-2a or 2b, or lamivudine. Among HBeAg-positive patients treated with interferon alpha-2a or 2b, a 2-point increase in HAI score at the end of treatment may be a potential surrogate for liver complications. Among HBeAg-positive patients treated with lamivudine alone or in combination with peginterferon interferon alpha-2a, HBeAg seroconversion may be an incomplete surrogate for decompensation. There are no available data that assess HBsAg seroconversion among treated patients on clinical outcomes. There are no data that assess drug resistance among treated patients or following treatment with adefovir or telbuvidine on clinical outcomes.

Key Questions

Consensus conference question 1. Which persons with hepatitis B should be treated?

EPC question 1. What is the evidence that the following population characteristics or clinical features associated with hepatitis B are predictive of hepatocellular carcinoma, liver failure, cirrhosis, liver-related death, and all-cause mortality?

Consensus conference question 2. What are the benefits and risks of the current therapeutic options for hepatitis B with defined or continuous courses of treatment?

EPC question 2a. What is the efficacy (or effectiveness) of interferon therapy, oral therapy, and various combinations in treating hepatitis B with defined or continuous courses of treatment?

EPC question 2b. What are the known harms of interferon therapy, oral therapy, and various combinations in treating hepatitis B with defined or continuous courses of treatment?

Surrogate outcomes of interest. Alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) levels, HBV viral load, change in hepatitis B e antigen (HBeAg) status, HBsAg conversion, liver biopsy findings (necroinflammatory activity or stage of fibrosis), and drug resistance.

Clinical outcomes of interest: hepatocellular carcinoma, liver failure, cirrhosis, liver-related death, all-cause mortality.

Consensus conference question 3. Which persons with hepatitis B should be treated?

EPC question 3a. Are there differences in efficacy/effectiveness of treatments for treatment naïve vs. drug-resistant patients, HBeAg-positive vs. HBeAg-negative patients, or for other subpopulations (as defined previously)?

EPC question 3b. Is there evidence that specific subpopulations do not require treatment for hepatitis B (i.e., that the surrogate and/or clinical outcomes are equivalent or superior when not exposed to treatment?)

Consensus conference question 4. What measures are appropriate to monitor therapy and assess outcomes?

EPC question 4. What is the evidence that changes in surrogate endpoints in response to treatment are reliable predictors of long-term resolution or slowed progression of disease? Patient Population: Adults (≥ 18 years of age), including elderly and members of racial/ethnic minority populations.

Eligibility

Three investigators independently decided on the eligibility of the studies according to recommendations from the Cochrane manual for systematic reviews.¹⁴⁸ The algorithm to define eligibility of the studies was developed for each research question (Appendix D). We reviewed abstracts to exclude secondary data analysis, reviews, letters, comments, case reports, and clinical trials of healthy populations to prevent hepatitis B. We confirmed eligible target populations of adults with chronic hepatitis B. The full texts of the original epidemiologic studies published in English after 1989 were examined to include studies with adult patients diagnosed with CHB. Eligible outcomes were defined as overall and liver-specific mortality, incidence of hepatocellular carcinoma (HCC) or liver failure, prevalence and incidence of cirrhosis, surrogate measures of HBeAg or HBsAg presence or seroconversion, viral load of hepatic virus B deoxyribonucleotide acid (HBV DNA), aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels, and histological necroinflammatory and fibrosis scores¹⁶ (operational definitions in Appendix D).

For question 1, we included studies if they: (1) were original research articles; (2) reported at least one of the following: hepatocellular carcinoma, liver failure, cirrhosis, liver-related death, and all-cause mortality; (3) had at least 1 year of either prospective or retrospective followup between the measurement of predictive factors and at least one of the outcomes of interest; or (4) reported results for a hepatitis B only population. Since the focus of this report is to provide evidence most relevant for a U.S. population, all studies meeting the previous criteria were included if the study reported results from a U.S. population. Only large studies (at least 1,000 participants) of populations outside of the United States were included. For questions 2–4 randomized controlled clinical trials (RCTs) of the drugs approved by the FDA for CHB, including interferon alfa-2b, pegylated interferon alfa-2a, lamivudine, adefovir, entecavir, tenofovir, and telbivudine¹⁷ were eligible for questions 2, 3, and 4. We included publications from the multinational HBV 99-01 Study Group of pegylated interferon alfa-2b that has been intensively examined in patients with CHB but not yet approved in the United States.¹⁸ Observational studies of more than 50 treated adults with more than 1 year followup that examined surrogate predictors of clinical outcomes were eligible for question 4.

Exclusion criteria included the following:

• Studies with target population as children and adolescents, healthy adults, adults with HCC, HIV, undergoing transplantation or chemotherapy, pregnant women, CHB populations mixed with other hepatitis patients (e.g., hepatitis C, CHB carriers, pregnant women with CHB, or individuals undergoing chemotherapy, if results were not separately provided for designated eligible cohort of CHB adults).

• Interventions of drugs not approved in the United States as of June 2008.

• Studies that reported not eligible outcomes including intra-hepatic concentrations of HBV DNA, acute pharmacokinetics measures, cardiovascular markers, or visual evoked potentials.

• Studies that evaluated cost effectiveness of different treatment options.

• Case series with small numbers of cases and no control comparison.

• Clinical trials of reverse transcriptase inhibitor that included less than 50 patients or examined active treatments for less than 24 weeks. Trials evaluating interferon for at least 12 weeks were eligible.

• Secondary data analysis with multiple reporting of the same outcomes.

• Data from randomized clinical trials that were reported ignoring randomization.

Quality Assessment and Rating the Body of Evidence

We analyzed study quality using the following criteria: subject selection, length and loss of followup, adjustment for confounding factors in observational studies and intention to treat principle in clinical trials, masking the treatment status, randomization scheme and adequacy, allocation concealment, and justification of sample sizes in RCTs.¹⁴⁹ The level of evidence for all studies was estimated using a subset of the U.S. Preventive Services Task Force criteria.

For all questions, evidence tables were developed identifying the purpose of the study, sample, design, independent and dependent variables, and findings (Appendix E). Baseline data were compared in different studies to test differences in the target population and unusual patterns in the data.^{150, 151} Standard deviations, regression coefficients, and 95 percent confidence intervals (CI) were calculated from reported event rates, means, standard errors, and sample size.^{152, 153} The protocol for the meta-analyses was created according to recommendations for meta-analysis of randomized controlled trials.^{154, 155} We assessed the level of evidence based on GRADE Working Group criteria.^{156, 157} We determined low level of evidence and confidence when data were from small RCTs or observational studies or from RCTs/observational studies with serious flaws in design/analysis and from post hoc subgroup analysis, moderate level of evidence, and confidence when a single large multinational study or several small RCTs/observational studies reported consistent effect of the same drugs or associations with factors and outcomes, and high level of evidence from multiple high quality studies in applicable patients reporting consistent sustained effects (post therapy at least 6 months).

Applicability of the population was estimated by evaluating the selection of the subjects in observational studies and clinical trials.¹⁵⁸ Large observational cohorts based on nationally representative samples had high applicability. Applicability of the intervention duration was high for studies with followup 1 year or more and acceptable for studies with followup of 6–12 months.¹⁵⁹ We evaluated baseline patient characteristics including age, gender, HBeAg status, previous treatment, and the presence of cirrhosis for generalizability.

We assumed the presence of publication bias and did not use statistical tests for bias defined as the tendency to publish positive results and to predict association when all conducted (published and unpublished) studies are analyzed.^{148, 160}–¹⁶² We used several strategies to reduce bias, including a comprehensive literature search of published and unpublished evidence in several databases, reference lists of systematic reviews, contacts with experts for additional references they might provide, and agreement on the eligibility status by several investigators.

Data extraction. Evaluations of the studies and data extraction were performed independently by five researchers. The data abstraction forms are shown in Appendix F. Errors in data extractions were assessed by a comparison with the established ranges for each variable and the data charts with the original articles.¹⁴⁸ Any discrepancies were detected and discussed. We abstracted the number of events among treatment groups to calculate rates, relative risk, odds ratios, and absolute risk differences (ARD).¹⁵² We abstracted the number randomized to each treatment group as the denominator to calculate estimates applying intention to treat principle.¹⁵² Means and standard deviations of continuous variables were abstracted to calculate mean differences with a 95 percent CI. We abstracted the time when the outcomes were assessed as weeks from randomization and the time of followup post treatments. We defined sustained response as 6 months or more post therapy. We extracted author reported adjustments for patient age, race, gender, and comorbidities. We prioritized clinical outcomes in the assessment of treatment benefits and harms. Sustained resolved hepatitis B was considered the next most relevant outcome.

Data synthesis. For questions 2 and 3 we summarized the results of individual studies in evidence tables to analyze differences in the outcomes among treatment groups. The definitions of the outcomes are presented below:

Clinical outcomes (clinical events) included death from all causes, liver related death, HCC or liver failure, and incidence of cirrhosis.

EPC Question 1. What is the evidence that population characteristics (age, age at infection, geographic location, race/ethnicity, gender, positive family history) or clinical features (presence of coinfections, HBV viral load, change in HBeAg status, genotype, nonalcoholic, fatty liver disease, alcohol consumption, AST/ALT level, liver biopsy finding) associated with hepatitis B are predictive of hepatocellular carcinoma, liver failure, cirrhosis, liver-related death, and all-cause mortality?

Objectives. We outlined the evidence to which the above mentioned population characteristics and clinical features predict HCC, liver failure, cirrhosis, liver-related death, and all-cause mortality in people with hepatitis B.

Description of study characteristics. Our search strategy identified 614 articles from abstracts or full articles that were obtained to determine study eligibility. Additionally we included six articles that were found through hand-searching other articles or identified by members of our TEP. Each article was read by one of three extractors and included for further review if the article either appeared to meet the inclusion criteria or if inclusion was uncertain. In cases where inclusion was not obvious, consensus by the other reviewers was used to decide.

Figure 1

Table 1

Factors associated with increased risk of selected (more...)

Table 1

Factors associated with increased risk of selected outcomes in adults with chronic hepatitis B

Risk Factor	All cause Mortality	Liver Mortality	Hepatocellular Carcinoma	Cirrhosis
Increased Age (~10 years)	3 studies32,51,56 Low confidence Moderate effect	1 study48 Low confidence Moderate effect	6 studies22,23,47–49,53 Medium confidence Small effect	2 studies29,53 Medium confidence Small effect
Male	4 studies24,47,48,56 High confidence Moderate effect	4 studies24,27,41,47 High confidence Moderate effect	8 studies20,22–24,30,42,46,49 High confidence Moderate effect	1 study29 Medium confidence Moderate effect
Geographic location and Asian race/ethnicity, early age of infection			3 studies35,48,49 Inconclusive
Family history of hepatocellular carcinoma			3 studies28,34,54 Low confidence Moderate effect
Nonalcoholic fatty liver disease
Modest alcohol consumption (drinkers average ~1 or fewer drinks per day)			5 studies23,28,30,50,53 Low confidence Small effect	1 study29 Inconclusive
Heavy alcohol consumption
Cirrhosis (present vs. absent various types of detection)		2 studies47,48 Medium confidence Strong effect	5 studies22,23,46,48,53 High confidence Strong effect	N/A
Genotype C (vs. other [mostly A, B, D])			6 studies22,33,49,55,57,59 High confidence Moderate effect
Genotype F (vs. mostly A D)			1 study33 Low confidence Strong effect
Precore mutation (A1896)			3 studies33,48,49 Low confidence Moderate effect
Basal core promoter mutation (T1762/A1764)		1 study26 Low confidence Small effect	4 studies33,48,49,57 Low confidence Moderate effect
High HBV DNA load (<10^4 copies mL, >10^5)	1 study57 Low confidence Small to moderate effect	3 studies25,48,58 High confidence Strong effect	6 studies22,23,49,52,59,168 High confidence Strong effect	1 study29 Medium confidence Strong effect
HBsAg loss				1 study53 Low confidence Small effect
HBeAg-positive status			8 studies23,37,48,52,53,55,57,59 Medium confidence Moderate effect	2 studies29,53 Medium confidence Small effect
Coinfection with HCV			2 studies20,43 Low confidence Moderate effect
Coinfection with HIV	2 studies19,31 Low confidence Small effect	3 studies19,31,44 Low confidence Strong effect
Coinfection with HDV	2 studies19,31 Inconclusive	3 studies19,31,44 Low confidence Strong effect
Elevated ALT level (>45 U/L)			3 studies23,53,57 High confidence Moderate effect	2 studies29,53 Medium confidence Small effect

Studies with references providing data for each outcome according to risk factor; level of confidence in estimate based on quality, quantity and consistency of evidence for the estimate of the relative risk magnitude is rated as “Inconclusive” (evidence insufficient to permit estimation of effect), “Low” (further research is likely to change the estimate), “Medium” (further research may change the estimate), “High” (further research is very unlikely to change the estimate); blank cells indicate no evidence available or does not apply. Magnitude of relative risk increase (RR) due to each factor for each outcome is estimated according to ranges from studies as “Small” (RR=1–2), “Moderate” (RR=2–5); and strong (RR=5 or greater)

Figure 1

Prior reports from Asia have estimated that the annual incidence of HCC is only 0.1 percent in asymptomatic HBsAg individuals, 1 percent in patients with CHB, but increases to 3–10 percent in patients with cirrhosis.⁶⁰ In this same report, patients with CHB developed cirrhosis at a rate of 2 percent per year.

Figure 3. Survival by hepatitis status, modified from (more...)

   Figure 3. Survival by hepatitis status, modified from Tong, 2006⁴⁸

Figure 3

Clinical features

Presence of coinfections: Human immunodeficiency virus (HIV), Hepatitis C virus (HCV), Hepatitis delta virus (HDV). Coinfection with HIV, HCV or HDV has been found to be associated with poorer clinical outcomes. However, the number of studies reporting this issue for any one type of coinfection is small, and associations are not consistent across different types of coinfection, so there is a low level confidence in the magnitude of these associations.^{19, 31, 43, 44}

Among HIV patients in Europe, Argentina, and Israel, HBV coinfection increased all-cause and liver-related death rates 1.5 and 3.6 times, respectively, above that of HIV infection alone.³¹ Thio and colleagues⁴⁴ found that among the large U.S. Multicenter AIDS Cohort Study (MACS) of 5,293 men who had sex with men, HIV status dramatically increased the rate of liver-related mortality in men positive for HBsAg. The liver-related mortality rate was 14.2 per 1,000 person years, which was approximately ten-fold higher than men with only HBsAg or HIV alone. However, the MACS study does not provide evidence on the extent to which the dramatically higher rate of liver-related death is due to more severe hepatitis B disease in the men with coinfection with HIV.

Among Japanese blood donors positive for HBV, those with coinfection with HCV had a 3-fold increase in HCC independent of age, sex, and ALT level.⁴³ However, a study by Amin and colleagues from Australia found similar rates of HCC in people with both HBV and HCV compared to those with HBV alone.²⁰

In a U.S. study of 231 developmentally disabled patients with chronic HBsAg-positive status living in institutional facilities, 65 patients were also antiHDV-positive.¹⁹ In multivariable models, patients positive for antiHDV were nearly 12 times (95 percent CI 1.4; 97.8) more likely to die of liver-related causes, but all-cause mortality was not significantly increased. The evidence for this association is weak because there were only eight liver-related deaths and it is uncertain how generalizable the results from an institution are to other environments.

In conclusion, estimates regarding coinfection and clinical outcomes could only be made with low confidence due to the paucity or inconsistency of the data. Coinfection with either HIV or HDV was associated with strongly increased liver-related mortality. Coinfection with HCV was associated with moderately increased HCC risk.

HBV viral load. Higher HBV viral load has been consistently shown to be associated with poorer clinical outcomes, particularly when comparing very low or undetectable levels of DNA to levels above 10⁵ or 10⁶ copies/mL. However, having low or undetectable DNA does not eliminate the risk of clinical outcomes.⁵² Furthermore, much less well known is the extent to which reductions in viral load lead to improvements in clinical outcomes.

The evidence for the association between HBV viral load and clinical outcomes was primarily from several large studies in Taiwan^{23, 29, 52, 54} and China.^{22, 25} However there were two articles from the United States that also found increased HCC cases⁴⁹ and increased non-HCC liver-related death in those with high viral loads.⁴⁸

The Taiwan REVEAL Study found that in multivariable models adjusted for age, gender, smoking, alcohol use, HBeAg status, ALT level, and cirrhosis the risk of HCC began increasing slightly for people with >10⁴ to <10⁵ copies/mL and the risk of HCC was around 6-fold higher for people with viral loads above 10⁵ or >10⁶ compared to people with undetectable viral loads.²³ This same study reported a similar association between viral load and risk of cirrhosis.²⁹ Additional reports from this study have also shown a strongly increased rate of liver-related mortality, that in turn leads to a modest (approximately 2-fold) multivariable adjusted increased all-cause death rate in those with HBV DNA ≥10⁵ compared to those with HBV DNA <10^5.58 There was no significant increase in the nonliver related death rate for those with elevated HBV DNA level. Another study from Taiwan found a similar HCC association with some increase in risk beginning above 10⁴ copies/mL and a substantial 7-fold increase in risk above 10⁶ copies/mL.^{54, 59} HCC death and chronic liver disease death have also been reported in a study from China to follow a similar trend.²⁵

In the REVEAL study the risk for HCC appeared to increase more steeply along the viral load gradient for groups with lower baseline risk of clinical outcomes.²³ For example, among the subset of people with normal ALT levels, no liver cirrhosis and negative for HBeAg there was a 4.5-fold increased risk at >10⁴ to <10⁵ copies/mL and a greater than 11 fold increased risk above >10⁵ copies/mL compared to people with normal ALT levels, no liver cirrhosis, negative for HBeAg and no detectable HBV DNA. It is likely that this steeper gradient of relative risk is driven largely by the much lower absolute rate of HCC in the low risk reference group. In a study from Hong Kong where the “low risk” HBV viral load group was defined more broadly as having levels <10^4.5 copies/mL and the “high risk” group was define as >10^6.5 copies/mL, only around a 2-fold increase in rate of HCC was found, after accounting for age, gender, cirrhosis, and albumin (additional adjustment for HBV genotype did not substantially alter the association).²² Another study from Taiwan also found that among people positive for HBsAg and negative for HBeAg, HCC cases were much more likely than controls to have elevated DNA; however, the greatest absolute proportion of both cases and controls had undetectable HBV DNA.⁵²

Results from the United States are consistent with the results from Asian countries, showing an increased rate of HCC and liver-related death across a gradient of HBV viral load.^{48, 49} In one U.S. study of 101 HCC cases of HBsAg-positive individuals, increased viral load was strongly associated with increased likelihood of HCC; however, none of the chronic inactive HBV controls had viral loads in the “high viral load group” (>10⁶ copies/mL), so the magnitude of effect due to “high viral load” could not be estimated.⁴⁹ Another U.S. cohort study from the same group of researchers found that among 400 chronic HBsAg patients high baseline HBV DNA viral load significantly increased the odds of nonHCC related liver death by nearly 5-fold (OR, 95 percent CI 4.7, 1.2; 20.4) independent of age and gender.⁴⁸

In conclusion, increased HBV DNA viral load is strongly associated with increased HCC (high confidence) and liver-related mortality (high confidence) even after accounting for baseline cirrhosis, HBeAg status, and ALT levels. However, there was only low confidence of a small to moderate association with all cause mortality. We also found a strong association between HBV viral load and cirrhosis (medium confidence). We found no evidence from these large observational studies regarding whether reduction in HBV DNA viral load is associated with better outcomes.

HBV genotype. Evidence for the impact of HBV genotypes on clinical outcomes for HBV is limited. It is clear that the prevalence of different genotypes varies substantially by geographic location, but more research is needed to determine the extent to which HBV genotype modifies the natural history of HBV related outcomes. What is available indicates that there likely are some differences in at least HCC rates according to genotype. Among a U.S. case control study of HCC in HBsAg-positive individuals, patients with HBV genotype C had 4-fold greater odds of HCC compared to other genotypes (genotypes A, B, and D). However, this association remained strong but was not statistically significant after accounting for age, gender, and basal core and precore mutations (OR 3.3, 95 percent CI 0.9; 12.1).⁴⁹ A large study from Taiwan found similar associations of 3–6-fold increased risk of HCC among people with the C genotype only compared to people with the B genotype only.^{57, 59} People with both B and C genotypes were at an intermediate risk.⁵⁹ The results from this study remained statistically significant and only modestly attenuated following multiple adjustment. In a large study from Hong Kong the HBV C genotype was associated with only a modest 1.5-fold (95 percent CI 1.2; 2.0) increased rate of HCC compared to genotype B after accounting for age, gender, cirrhosis, viral DNA load, and albumin.²² Finally, among Alaska Native people the odds of HCC were 4.7 times greater in patients with the A genotype (95 percent CI 1.4; 16.0) and 11.7 times greater in patients with the F genotype (95 percent CI 5.4; 25.4) compared to those with the D genotype.³³

In conclusion, HBV genotypes may be associated with differing risk of clinical outcomes. Genotype C moderately increases risk of HCC compared to genotypes A, B, and D (high confidence), and genotypes A (moderate effect) and F (strong effect) may increase risk compared to D (low confidence).

HBsAg loss. Only one study was identified that reported HBsAg loss and clinical outcomes, and this study had low power to detect meaningful differences in risk.⁵³ In a large Taiwanese study of asymptomatic carriers at baseline followed for an average of 7 years, those with HBsAg loss had a 40 percent reduction in risk of cirrhosis, but this was not statistically significant (95 percent CI 79 percent reduction ranging to a 64 percent increase in risk of cirrhosis), after adjusting for age, HBeAg status and AST/ALT levels.⁵³

In conclusion, HBsAg loss may be associated with a reduction in risk of cirrhosis (low confidence). There is no evidence whether or not HBsAg loss is associated with other clinical outcomes.

HBeAg status. HBeAg-negative status in a population study tends to be a marker of inactive carrier status, particularly when ALT levels are normal and HBV viral load is low (Figure 1). However, HBeAg-negative CHB can also occur (it is indicated by elevated HBV DNA and ALT). Therefore, it becomes more difficult to interpret the association between HBeAg status and outcomes without also using ALT and HBV DNA levels to help to classify people into either inactive carrier status or HBeAg-negative status. Since it is well known that inactive carriers have lower rates of clinical outcomes than those with either HBeAg-positive or negative chronic active hepatitis, the most interesting research questions may be to determine the impact of HBeAg status in people with active hepatitis and the effect of HBeAg reversion on clinical outcomes. Unfortunately, we found few studies that classified people into groups of chronic inactive hepatitis and chronic active hepatitis and then looked at the effect of HBeAg within those groups.

While several studies have reported a consistently higher rate of outcomes among people who are HBeAg-positive compared to HBeAg-negative,^{23, 29, 37, 53, 57, 59} we were unable to assess the effect of the HBeAg independent of its role as a marker of chronic active versus chronic inactive hepatitis. One study in Taiwan found the incidence rate for HCC was 3.6 times higher in HBsAg-positive people who were also HBeAg-positive compared to those who were HBeAg-negative.⁵² From the REVEAL study in Taiwan this increased risk of HCC (HR 2.6, 95 percent CI 1.6; 4.2) and cirrhosis (RR 1.7, 95 percent CI 1.3; 2.9), for HBeAg-positive people persisted following adjustment for age, gender, HBV viral load, and ALT level.^{23, 29} A third large Taiwanese study also reported 2–3-fold increased risk of HCC among people with HBeAg-negative CHB.^{57, 59}

Among Alaska Natives, reversion to HBeAg positivity or multiple switches in HBeAg status was associated with increased risk for hepatocellular carcinoma (HR 2.6, 95 percent CI 1.3; 5.4), after adjustment for potential confounders.³⁷ Another U.S. study by Tong and colleagues that classified all patients into “inactive carriers,” “chronic hepatitis,” or “cirrhotic,” found that patients who were positive for HBeAg at baseline had similar rates of HCC and all cause death as patients antiHBeAg at baseline.⁴⁷

In conclusion, HBeAg-positive status is associated with moderately increased HCC (medium confidence) and small increases in cirrhosis (medium confidence) independent of other disease factors such as HBV viral load and ALT level.

Basal core promoter (T1762/A1764) or precore mutation (A1896). Only a few recent studies have attempted to look at the extent to which basal core promoter (BCP) mutations and precore (PC) mutations impact clinical outcomes.^{26, 33, 48, 49, 57} This is one area where much of the information came from U.S. based studies.

Among Alaska Natives there was no significant association between either BCP or PC mutations and HCC. However, the basal core mutations did vary significantly by HBV genotype.³³ Among a U.S. case control study of HCC in HBsAg-positive individuals, the A1896 PC mutation was associated with a nearly 4-fold increase in HCC and the T1762/A1764 mutation was associated with an 11-fold increase in HCC compared to wild types for both of these factors, independent of age, gender, race, and HBV genotype.⁴⁹ In a U.S. cohort study of 400 chronic HBsAg patients the odds of developing HCC were 2.9 times greater (95 percent CI 1.2; 7.6) for those with the BCP mutation and 4.2 times greater (95 percent CI 1.5; 19.6) for those with the A1896 PC mutation compared to those with wild type basal and PC mutations, respectively.⁴⁸ In a large study out of China the HCC death rate was 1.40 (95 percent CI 1.06; 1.85) times greater in those with 1762T/A1764 BCP mutations compared to other HBsAg-positive subjects.²⁶ Likewise a large study from Taiwan found a 1.92-fold (95 percent CI 1.14; 3.25) increased risk of HCC, independent of HBV genotype, ALT level and HBeAg status.

In conclusion, the BCP mutations (T1762/A1764) and the PC mutation (A1896) are associated with moderately increased HCC rates and BCP is associated with increased liver-related death rates (low confidence).

Cirrhosis. Cirrhosis has been shown to be a consistently strong predictor of HCC development and death in many studies. It has been reported for decades even within the United States that survival is greatly reduced in patients with cirrhosis compared to patients without cirrhosis.^{47, 51} As early as 1984 Weissberg and colleagues were reporting that the 5-year survival rate among patients with CHB could range from 97 percent in patient without cirrhosis to 55 percent in patients with chronic active hepatitis and cirrhosis.⁵¹ In a study by Tong and colleagues, biopsy determined cirrhosis was associated with a 3.6-fold (95 percent CI 1.6; 8.9) increased odds of developing HCC independent of age, serum albumin, and baseline platelets. In the same study, the independent association was even stronger for all-cause death and nonHCC liver-related death (OR 14.2, 95 percent CI 3.4; 111.8 and 7.3, 95 percent CI 1.3; 69.56 respectively).

The findings from U.S. studies are consistent with the large studies from Taiwan and China which have consistently reported much higher rates of HCC and death in cirrhotic individuals.^{22, 23, 53} Rates are often nearly 10-fold greater in people with cirrhosis even after adjustment for other markers of disease severity such as elevated ALT or HBV viral load. Few large studies had biopsies in all of their patients and instead relied on ultrasound detected cirrhosis which still strongly predicted increased rates of clinical outcomes.

In conclusion, cirrhosis is a strong predictor of HCC (high confidence) and liver-related death (medium confidence).

Nonalcoholic fatty liver disease. No studies were identified that reported the impact of nonalcoholic fatty liver disease on clinical outcomes in people with chronic HBV.

Alcohol consumption. Alcohol consumption was not frequently reported as an important factor in models predicting clinical outcomes from HBV. Studies that did include measures of alcohol consumption tended to use variables that indicate any consumption or years of consumption and did not try to isolate people with heavy alcohol consumption. The association between alcohol consumption and clinical outcomes reported in the identified studies appeared modest at best with effect sizes around 1.5-fold increased risk of HCC. In a large Taiwanese study of over 2,000 people, alcohol consumption and duration of alcohol use were only weakly associated with HCC development. Compared to people who never drank alcohol, those who drank for over 20 years only had a 1.33-fold increased risk of HCC (95 percent CI 0.75; 2.43) adjusting for age, family history of HCC, HCV status, baseline liver function, ethnicity, and education.⁵⁰ However, there did appear to be a potential interaction with smoking status such that those with increased alcohol and smoking use had elevated HCC.⁵³ Similarly, size associations were reported in two other studies of 1.5 and 1.6-fold increased risk of HCC in those who consumed about two drinks per day in one study³⁰ or reported any alcohol consumption in the other study.²³

While modest consumption of alcohol does not appear to be a strong predictor of clinical outcomes related to HBV, cirrhosis was a consistently strong predictor of HCC and death. So while the studies identified did not break out causes of cirrhosis, it might be reasonable to assume that heavy drinking that leads to liver cirrhosis may be an important factor in clinical outcomes, even if modest drinking is not.

In conclusion, moderate alcohol consumption in people chronically infected with HBV appeared to be a weak predictor of increased HCC. There is low confidence in this association. Little evidence exists regarding the association between heavy alcohol use and clinical outcomes in people with chronic HBV.

AST and ALT levels. Few studies reported associations between elevated aminotransferase levels and clinical outcomes. Those that did tended to report increased risk of outcomes. This increased risk may be in part explained by other factors. Among a large Taiwanese study of asymptomatic carriers at baseline followed an average of 7 years, those with either elevated AST or ALT levels had a 3.1-fold (95 percent CI 1.0; 10.0) increased risk of HCC and a 3.7 fold (95 percent CI 2.3; 6.0) increased rate of cirrhosis, independent of age, HBeAg status, and baseline cirrhosis (for the HCC results).⁵³ Another study from Taiwan, also found a similar association with HCC 2.5-fold (95 percent CI 1.1; 4.3).⁵⁷ Also from Taiwan, the REVEAL study reported an unadjusted 4-fold increased risk of HCC with ALT levels >45 U/L, but after adjusting for age, gender, smoking, alcohol, HBeAg, cirrhosis, and HBV viral load the association was completely attenuated (HR 1.1, 95 percent CI 0.7; 1.7).²³ In the same study, the association between elevated ALT and cirrhosis remained significant but only modest in strength after multiple adjustment (HR 1.5, 95 percent CI 1.1; 2.1).²⁹

In conclusion, ALT is moderately associated with increased risk of HCC (high confidence) and weakly associated with cirrhosis (low confidence). These associations appear to be largely explained by accounting for baseline cirrhosis, HBeAg status, and HBV viral load (low confidence).

EPC Question 2a. What is the efficacy (or effectiveness) of antiviral therapy in treating adults with chronic hepatitis B?

Table 2

Treatments of hepatitis B: Overview of randomized controlled trials (more...)

Table 2

Treatments of hepatitis B: Overview of randomized controlled trials

Study Characteristic	Percent or Mean (Range)	Number of Subjects	Number of Trials Reporting
All studies (# subjects)	20–1,367	11,144	59
Weighted mean age	37 (24–58)	7,884	40
Gender, male (%)	78	8,408 / 10,721	58
Race, Asian (%)	64 (0–100)	5,097 / 7,954	27
Race, White (%)	30 (0–98)	2,219 / 7,954
Race, Black (%)	1 (0–24)	111 / 7,954
Race, Other (%)	5 (0–8)	184 / 5,380
e Antigen-positive (%)	81 (2–100)	7,453 / 9,160	48
e antigen-negative (%)	64 (19–100)	2,828 / 4,434	17
Mean ALT level (IU/L)	139 (77–284)	6,917	33
Median ALT level (IU/L)	Range 56–170	1,327	7
Study duration (weeks), therapy and followup combined	69 (17–208)	10,606	56
Treatment naive	All subjects	2,388	13
Treatment resistant	All subjects	1,241	11
Study withdrawals (%)	7 (0–35)	712 / 10,199	50
Withdrawals due to adverse events (%)	3 (0–16)	199 / 7,697	36
Cirrhosis (%)	21 (5–65)	1,258 / 6,047	31
Studies ≥1 biopsy	All subjects	8,466	43
HBV genotype
A	13 (0–34)	609 / 4,800	11
B	19 (0–32)	913 / 4,800
C	42 (15–100)	2,002 / 4,800
D	19 (0–46)	906 / 4,800

Figure 1

In 16 reports investigators reported outcomes for individuals who were naïve to antiviral drugs patients.^{68, 72, 75, 76, 85, 90, 94, 106}–^{108, 115, 120, 122, 126, 136, 139} Seven reports enrolled patients independent of previous treatment status or tested new drugs on patients resistant to previous treatments.^{67, 77, 92, 118, 119, 125, 141} Cirrhosis was assessed at baseline in 32 studies⁶¹–^{64, 67, 73, 74, 77, 80}–^{83, 85, 88, 91, 92, 94}–^{96, 99, 103}–^{105, 110, 121, 122, 132, 135, 136, 138, 139, 142} and was noted in 21 percent these enrollees. Authors reported HBV genotype in 13 studies.^{94, 96, 99, 104, 107, 109, 114, 121, 122, 124, 125, 139, 143} Genotype C was the most common (42 percent).

Table 3

Effects of drug therapies for chronic hepatitis B on (more...)

Table 3

Effects of drug therapies for chronic hepatitis B on clinical outcomes

Active Treatments vs. Control Treatment	Treatment Duration/Followup Off Therapy, Weeks	Studies/Subjects Enrolled	Estimates (95% CI) *Dose Response Heterogeneity: p Value/I Squared, %	Level of Evidence/Comments
Liver related death
Lamivudine vs. placebo132	130/0	1/651	0.00 (-0.01; 0.01) (RD)	Low. Sparse data (0 events in both groups) No effect of LAM on liver related death
Mortality
Lamivudine vs. placebo132	130/0	1/651	0.00 (-0.01; 0.01) (RD)	Low. Sparse data (0 events in control group) No effect of LAM on mortality
Interferon alfa-2b vs. placebo83	16/48–64	1/40	NS at the end of treatment and after followup	Low. Sparse data (small N of events) No effect of Interferon alfa 2 B on mortality
Adefovir dipivoxil111	114 vs. 240/0	1/125	0.33 (0.01; 8.10) (RR)	Low. Sparse data (0 events at second time point, no formal control) Length of adefovir therapy did not affect mortality
Entecavir124	48/0	1/89	NS among all compared doses No *	Low. Sparse data (small number of events) No dose response effect on mortality
Entecavir vs. lamivudine121,122,124–126	48–96/0–28	5/2476	NS in all studies -0.003 (-0.008;0.002) (RD) 0.7/0%	Low. Sparse data (small N of events) No differences between entecavir vs. lamivudine on mortality
Interferon alfa 2b+corticosteroid vs. interferon alfa 2b86	24/0	1/37	-0.11 (-0.27; 0.06) (RD)	Low. Sparse data (0 events in active group) No differences of pretreatment with steroid and interferon alfa-2b vs. interferon alfa-2b on mortality
Interferon alfa 2b+corticosteroid vs. symptomatic treatment70	24/48	1/20	-0.10 (-0.34; 0.14) (RD)	Low. Sparse data (0 events in active group) No differences of pretreatment with steroid and interferon alfa-2b on mortality
Interferon alfa 2b86,90	24–48/0–24	2/76	NS among all compared doses No *	Low. Sparse data (0 events in active group) No dose response effect on mortality
Peginterferon alfa-2a+placebo vs. lamivudine96	48/8	1/543	0.00 (-0.01; 0.01) (RD)	Low. Sparse data (0 events in both groups) No differences between peginterferon alfa-2a and lamivudine on mortality
Peginterferon alfa-2a+lamivudine vs. lamivudine96	48/8	1/543	0.01 (0.00; 0.03) (RD)	Low. Sparse data (0 events in control group) No differences between peginterferon alfa-2a combined with lamivudine vs. lamivudine alone on mortality
Peginterferon alfa-2a+lamivudine vs. peginterferon alfa-2a96	48/8	1/543	0.01 (0.00; 0.03) (RD)	Low. Sparse data (0 events in control group) No differences between peginterferon alfa-2a combined with lamivudine vs. peginterferon alfa-2a alone on mortality
Peginterferon alfa-2b+lamivudine vs. amivudine106	50–60/56–64	1/100	0.02 (-0.03; 0.07) (RD)	Low. Sparse data (0 events in control group) No differences between peginterferon alfa-2b combined with lamivudine vs. lamivudine on mortality
Incident cirrhosis
Interferon alfa-2b vs. no treatment83	16/48–64	1/40	-0.05 (-0.21; 0.11) (RD)	Low. Sparse data (small number of events) No effect of interferon alfa-2b on cirrhosis
Interferon alfa 2b+corticosteroid vs. interferon alfa 2b85	24/24	1/56	-0.06 (-0.24; 0.11) (RD)	Low. Sparse data (small number of events) No differences of pretreatment with steroid and interferon alfa-2b on incidence of cirrhosis
Hepatic decompensation
Lamivudine vs. placebo141	80/0	1/74	0.05 (-0.11; 0.22) (RD)	Low. Sparse data (small number of events) No effect of lamivudine on liver decompensation
Lamivudine vs. no treatment141	80/0	1/74	0.00 (-0.12; 0.12) (RD)	Low. Sparse data (small number of events) No effect of lamivudine on severe liver decompensation
Peginterferon alfa-2b+lamivudine vs. lamivudine106	52–60/57–72	1/100	0.00 (-0.04; 0.04) (RD)	Low. Sparse data (0 events in both groups) No differences between combined peginterferon alfa-2b with lamivudine vs. lamivudine alone on liver decompensation
Entecavir vs. lamivudine122,126	52–96/0–24	2/709	NS after different treatment duration	Low. Sparse data (0 events in active group) No differences between entecavir vs. lamivudine on liver decompensation
HCC
Lamivudine vs. placebo132	130/0	1/651	-0.04 (-0.07; 0.00) (RD)	Low. Significant protective effects of active drug after adjustment for country, sex, baseline Alanine aminotransferase level, Child-Pugh score, and Ishak fibrosis score (HR = 0.49, 95% CI 0.25; 0.99) No effect of lamivudine on HCC
Interferon alfa-2b vs. placebo91	96/0	1/42	0.05 (-0.07; 0.17) (RD)	Low. Sparse data (0 events in control group) No effect of Interferon alfa-2b on HCC
Adefovir dipivoxil111	114 vs. 240/0	1/250	0.03 (-0.01; 0.07) (RD)	Low. Sparse data (small number of events at first time point, no formal control) Length of adefovir therapy did not affect HCC
Interferon Alfa 2b+corticosteroid vs. interferon alfa 2b85	24/24	1/56	-0.02 (-0.28; 0.24) (RD)	Low. No difference on active hepatitis between interferon alfa-2b with pretreatment using corticosteroid vs. interferon alfa-2b

Bold - significant association at 95% confidence level; RD - absolute risk difference; RR - relative risk

Figure 2

Figure 3

The remaining studies were small and short term. They assessed use of corticosteroids or different doses or duration of therapy. None demonstrated a mortality difference between treatment approaches. One small RCT from Egypt of 40 HBeAg-positive patients (40 percent with cirrhosis) found no difference in mortality after 16 weeks of interferon alfa-2b compared to placebo and 48–64 weeks of followup.⁸³ Interferon alfa-2b with corticosteroid pretreatment compared to symptomatic therapy without antiviral drugs failed to reduce mortality in a small RCT of 20 HBeAg-positive South African patients.⁷⁰ Steroid withdrawal and low dose of interferon alfa-2b for 24 weeks in 56 HBeAg-positive patients did not reduce mortality rates (ARD -0.11 95 percent CI -0.27; 0.06).⁸⁶ Two RCTs^{86, 90} of interferon alfa-2b did not find a dose- response effect on mortality among HBeAg-positive patients.^{86, 90}

Dose or duration of the therapy did not affect mortality. Prolongation of adefovir administration did not reduce mortality in 125 HBeAg-negative Greek patients.¹¹¹ Entecavir in different doses did not decrease mortality in patients with lamivudine resistant hepatitis.¹²⁴

In conclusion, antiviral medications did not reduce mortality versus placebo, compared to other antiviral medications, or in combination with corticosteroids regardless of HBeAg or cirrhosis status. Studies reporting mortality evaluated different patient populations and drug combinations, thus generally precluding pooling. Level of evidence and confidence in effect estimate is low. Studies assessing mortality had inadequate size and duration to detect significant differences.

Figure 4

In conclusion, sparse data suggest no effects of interferon alfa-2b alone or in combination with steroids on short-term incident cirrhosis. The long-term effects of interferon alfa-2b alone or in combination with steroids on clinical outcomes are unknown. No data were available for other antiviral drugs. Overall level of evidence and confidence in effect estimate is low.

Hepatic decompensation. Hepatic decompensation was reported in three studies;^{122, 126, 141} one small RCT compared lamivudine to placebo¹⁴¹ and two assessed outcomes after 52–96 weeks of entecavir versus lamivudine administration.^{122, 126} Studies reported very few cases of hepatic decompensation. Eighty weeks of lamivudine treatment did not affect the development of hepatic decompensation in 74 Korean patients with lamivudine-resistant mutant CHB (ARD 0.05, 95 percent CI -0.11; 0.22).¹⁴¹ The Benefits of Entecavir for Hepatitis B Liver Disease (BEHoLD) study evaluated 715 HBeAg-positive patients of which 8 percent had cirrhosis at baseline.¹²² An American study¹²⁶ also assessed 709 HBeAg-positive patients, though they did not report the number of subjects with cirrhosis at baseline. Neither found a difference in hepatic decompensation between entecavir compared to lamivudine.^{122, 126} There were only two cases reported both in the lamivudine group.

In conclusion, there is insufficient evidence regarding the relative effects of entecavir versus lamivudine in preventing hepatic decompensation over 1–2 years among HBeAg-positive patients. Effects of other antiviral drugs or in different patient populations are unknown.

Hepatocellular carcinoma was reported in four studies. None demonstrated a statistically significant difference. Two studies compared placebo to lamivudine¹³² or interferon alfa-2b.⁹¹ One compared the addition of corticosteroids to interferon alfa-2b versus interferon monotherapy,⁸⁵ and one examined the effects of prolonged adefovir therapy.¹¹¹ Incidence of hepatocellular carcinoma did not differ between lamivudine (130 weeks, 17/436) and placebo (16/215) in a multicenter study of 651 Asian patients (58 percent HBeAg-positive) with confirmed cirrhosis (61 percent) or advanced fibrosis (ARD -0.04, 95 percent CI -0.07; 0.00).¹³² A further analysis that adjusted for country, sex, baseline ALT level, Child-Pugh score, and Ishak fibrosis score found a borderline significant effect (HR 0.49, 95 percent CI 0.25; 0.99, p=0.047 borderline significant).¹³²

Interferon alfa-2b for 96 weeks compared to placebo failed to prevent hepatocellular carcinoma in 42 HBeAg-negative Italian patients.⁹¹

The French Multicenter Group did not find protective effects of adding corticosteroids to interferon alfa-2b compared to interferon alone at the end of 24 weeks of therapy and at 24 weeks of followup in treatment naïve HBeAg-positive patients with CHB (ARD -0.02, 95 percent CI -0.28; 0.24).⁸⁵

In conclusion, study number, design, and duration were inadequate to accurately assess the impact of treatments on hepatocellular cancer. Limited low level evidence from one multinational RCT suggested that 130 weeks of lamivudine may reduce the incidence of hepatocellular carcinoma in Asian adults with hepatitis B and cirrhosis or advanced fibrosis. Results come from a single trial that noted no significant differences in crude rates and reported a nonstatistically significant increase in all-cause mortality with lamivudine. Protective effects on HCC with lamivudine were significant only after adjustment for baseline variables and after excluding five individuals who developed hepatocellular cancer within the first year of the study. Interferon alfa-2b monotherapy was not protective in a single small short-term study reporting very few events (low confidence). Addition of corticosteroids to alfa interferon was not superior to alfa interferon alone in a single, small short-term study with few events (low confidence). There are no data evaluating other antiviral agents.

Figure 5

Figure 6

Comparative effectiveness of interferon and reverse transcriptase inhibitors on HBsAg loss at the end of the drug administration did not differ in any of the ten RCTs that examined the association.^{67, 69, 71, 84, 99, 111, 119, 122, 126}

Entecavir and lamivudine resulted in similar rates of HBsAg loss and seroconversion.^{122, 126} Combination of interferon alfa-2b with lamivudine did not increase HBsAg loss compared to lamivudine alone in HBeAg-positive⁶⁷ and negative patients.⁷¹

Figure 7

Figure 8

Figure 9

In conclusion, interferon alfa-2b alone and with steroid pretreatment increased HBsAg loss by about 10–15 percent at the end of drug administration (moderate level of evidence). However, sustained effects of interferon alfa-2b on HBsAg loss beyond 48 weeks off treatment have not been examined. Additionally, interferon alfa-2b failed to increase rates of several criteria of resolved hepatitis B. The effects of other drugs and their combinations on composite criteria of resolved hepatitis B including HBsAg loss have not been investigated. Comparative effectiveness of evaluated active treatments on short-term intermediate outcomes (loss of HBsAg) was similar at the end of the therapy and at short-mid duration followup off treatment in the populations studied (moderate evidence).

Studies obtained assays with different sensitivity to detect HBV DNA. Viral load was measured using polymerase chain reaction assay,^{63, 74, 87, 90, 110, 113, 117, 119}–^{123, 126, 127, 129, 139, 143} reverse transcription polymerase chain reaction assay,^{107, 124} or solution hybridization assay.^{62, 66}–^{69, 71, 72, 75, 77, 84}–^{86, 136, 145} Obtained assay methods had different detection limits and units to measure viral load: <200 copies/mL,¹¹⁷ <300 copies/mL,^{113, 121, 122, 126} <400 copies/mL,^{96, 102, 116} <500 copies/mL,¹¹⁷ <1,000 log copies/ml,^{10, 111} <3 log10 copies/mL,¹⁴³ <1.6pg/mL,^{131, 140} <2.5 pg/mL,¹⁴² <3 pg/mL,^{62, 67} or <6pg/mL.⁷⁷ We explored heterogeneity in drug effects across the studies using the assay to measure HBV DNA loss and did not convert units of cut offs.

We reviewed 43 studies that examined HBV DNA clearance after interferon and reverse transcriptase inhibitors.^{10, 57, 62, 63, 66}–^{69, 71, 72, 75, 77, 80, 85}–^{88, 90, 92, 96, 106, 107, 110}–^{113, 117, 119}–^{124, 126, 127, 129, 131, 133, 136, 137, 139, 142, 145} Twenty-eight publications included HBeAg-positive.^{63, 65, 68, 69, 72, 75, 77, 80, 84}–^{88, 90, 92, 96, 102, 106, 107, 112, 119, 120, 126, 127, 129, 131, 136, 140} Five reports assessed HBeAg-negative patients;^{10, 71, 74, 110, 111} the rest of the studies included patients with chronic active hepatitis B independent of HBeAg baseline status. Twenty trials examined the effects of interferon alfa-2b,^{62, 63, 65}–^{69, 71, 72, 74, 75, 77, 80, 84}–^{88, 90, 92} one trial examined peginterferon alfa-2a,⁹⁶ eight publications reported HBV DNA loss after adefovir,^{10, 110}–^{113, 117, 119, 120} five articles¹²¹–^{124, 126} examined the effects of entecavir on HBV DNA clearance, and 39 analyzed the effects of lamivudine.^{57, 62, 63, 67, 68, 71, 72, 74, 75, 77, 96, 102, 106, 107, 119, 121}–^{124, 126, 127, 129, 131, 133, 136, 137, 139, 140, 142, 143, 145}

Figure 10

Interferon alfa-2b for 16 weeks compared to no antiviral treatment ⁶⁹ (ARD 0.45, 95 percent CI 0.22; 0.68) increased HBV DNA loss in HBeAg-positive patients. The same RCT of HBeAg-positive patients reported a significant increase in HBV DNA loss after 16 weeks of interferon alfa-2b combined with corticosteroid (ARD 0.25, 95 percent CI 0.04; 0.46).⁶⁹

Lamivudine for 12–104 weeks compared to placebo or usual care^{67, 129, 131, 133, 136, 139, 145} increased HBV DNA clearance with consistent across-the-studies relative risk (pooled RR 3.79, 95 percent CI 2.71; 5.30). The pooled absolute risk difference was significant (pooled ARD 0.48, 95 percent CI 0.31; 0.66), but inconsistent, with evidence of statistical heterogeneity that could not be explained by length of treatment or control rate of HBV DNA loss (metaregression p value >0.05). The effects of baseline HBeAg status, assay to measure viral load, or the proportion of patients with baseline cirrhosis, could not explain variability in the results. A valid metaregression was not possible because not all studies reported this information.

Interferon alfa-2b combined with lamivudine for 52 weeks compared to placebo⁶⁷ increased the rate of undetectable HBV DNA (<3pg/mL—a measure used to define resolved hepatitis B) (ARD 0.48, 95 percent CI 0.33; 0.63) with random differences (ARD 0.05, 95 percent CI -0.09; 0.18) in sustained HBV DNA response (no two consecutive detectable HBV DNA on treatment) in predominantly HBeAg-positive patients (98 percent).

Table 5

Absolute risk difference in tested nonclinical outcomes (more...)

Table 5

Absolute risk difference in tested nonclinical outcomes after antiviral drugs for chronic hepatitis B in adults

Comparison	HBsAg [RCTs/Patients]	HBeAg [RCTs/Patients]	HBV DNA Clearance [RCTs/Patients]	Histology Improved [RCTs/Patients]	ALT Normal [RCTs/Patients]	Relapse/Mutation [RCTs/Patients]
Adefovir vs. placebo	SC: NS [1/120] L	Loss: 0.11 (0.06; 0.16) [2/995] M SC: 0.05 (0.01; 0.09) [2/700] H	0.38 (0.23; 0.53) [4/1002] H 20.41 (6.79; 61.32) [4/1002] 0.59 (0.46; 0.72)* [1/120] L	Fibrosis: 0.20 (0.14; 0.26) [2/699] M Necroinflammation scores: 0.26 (0.17; 0.34) [3/819] H	0.40 (0.30; 0.49) [5/1342] 2.97 (2.38; 3.69) H 0.26 (0.19; 0.33)* [2/600] M	NS [2/1055] L NS* [1/140] L
LAM vs. placebo	Loss: NS [1/175] L NS* [3/1068] L	Loss:0.13 (0.04; 0.22) [4/1349] M 0.15 (0.05; 0.24)* [2/318] M SC: 0.05 (0.001; 0.10) [6/1638] H 1.70 (1.05; 2.74) NS * [2/318] L	0.48 (0.31; 0.66) [7/1305] 3.79 (2.71; 5.30) H 0.08 (0.00; 0.15)* [1/136] L	Necroinflammation: 2.09 (1.60; 2.74) M 0.25 (0.13; 0.38) [4/580] M	0.22 (0.13; 0.31) [7/1602] 2.42 (1.94; 3.01) M 0.21 (0.04; 0.38)* [1/136] L	YMDD mutation: 0.43 (0.38; 0.48) [2/826] H
Adefovir + LAM vs. LAM	Loss: NS [1/39] L	Loss: 0.12 (0.03; 0.21) [2/134] M SC: NS [2/134] L	0.25 (0.10; 0.39) [2/134] L		0.32 (0.13; 0.52) [2/13] M	YMDD:-0.33 (-0.50; -0.17) [1/95] L Wild type mutation: NS [1/95] L
Adefovir+ LAM vs. adefovir	Loss: NS [1/39] L	Loss: NS [1/39] L SC: NS[1/39] L	NS:[1/39] L		NS [1/39] L
Entecavir vs. LAM	Loss: NS [2/1117] M SC: NS [1/408] L	Loss: NS [3/1112] L SC: NS [3/1185] M	0.23 (0.11; 0.35) [4/1636] 1.64 (1.22; 2.22) L/M NS*1/709] L	Necroinflammation: 0.14 (0.04; 0.24) [3/1633] M Fibrosis: NS [2/995] M	0.22 (0.11; 0.32) [6/2423] 1.62 (1.28; 2.06) H	NS [0/1347] L -0.16 (-0.20; -0.12)* [1/709] L
LAM vs. adefovir		Loss: NS [1/38] L SC:NS [1/38] L	-0.26 (-0.47; -0.06) [1/38] L		-0.42 (-0.67; -0.18) [1/38] L
LAM vs. telbivudine		Loss: NS [1/63] L SC: NS [1/63] L	-0.30 (-0.55; -0.04) [1/63] L		NS [1/85] L	NS [1/63] L
Telbivudine vs. adefovir		Loss: NS [1/135] L SC: 6.03 (2.20; 16.52) [1/136] L	0.28 (0.12; 0.44) [1/136] L		NS [1/135] L
Telbivudine+LAM vs. LAM		Loss: NS[1/60] L SC: NS [1/60] L	NS [1/60] L		NS [1/101] L	NS [1/60] L
Telbivudine+LAM vs. telbivudine		Loss: NS [1/85] L SC: NS [1/85] L	NS [1/85] L		NS [1/101] L	NS [1/85] L
Interferon alfa-2b vs. placebo	Loss: NS [3/166] M NS* [4/247] L SC: NS* [2/82] L	Loss: 0.55 (0.29; 0.81) [1/40] L 2.52 (1.55; 4.10) 0.28 (0.07; 0.50)* [3/351] M SC:NS [1/40] L 0.12 (0.03; 0.21) * [2/304] M	0.45 (0.22; 0.68) [1/34] L 0.44 (0.27; 0.60)* [3/168] L	Total scores: NS* [1/40] L HAI scores:0.24 (0.00; 0.48) [1/72] L	0.31 (0.17; 0.44)* [2/131] M	Relapse: NS* [5/378] H
Interferon alfa-2b+lamivudine vs. placebo	Loss: 0.06 (0.00; 0.13) [1/119] L NS* [1/119] L	Loss: NS [1/118] L NS* [2/450] M SC: NS [1/119] L NS* [2/450] L	0.48 (0.33; 0.63) [1/119] L NS* [1/119] L	HAI scores NS [1/119] L	NS [1/119] L	YMDD mutation NS [1/118] L
Interferon alfa-2b+ corticosteroid vs. no treatment	Loss: 0.11 (0.02; 0.20) [2/103] M		0.25 (0.04; 0.46) [1/34] L NS* [2/121] M		0.25 (0.06; 0.43)* [1/87] L	Relapse NS* [1/87] L
Interferon alfa-2b vs. LAM		Loss: NS [1/151] L NS* [2/625] M SC: NS[1/151] L NS* [3/776] M	NS [1/76] L NS* [1/151] L	Knodell scores: NS* [1/151] L	NS [1/151] L NS* [2/151] L	YMDD mutation -0.23 (-0.33; -0.14)* [1/151] L
Interferon Alfa 2b+ LAM vs. interferon alfa-2b		Loss: NS [1/144] L NS* [2/347] M SC: NS [1/144] L NS* [3/482] L	NS [1/144] L NS* [2/278] L	HAI scores 0.54 (0.28; 0.79) [1/48] L Knodell scores NS* [1/144] L	NS [1/144] L NS* [2/192] L	YMDD mutation NS* [1/144] L
Interferon alfa-2b+ LAM vs. LAM	Loss: NS [2/262] L NS* [3/495] L	Loss: NS [3/414] M NS* [5/1167] M SC: NS [4/565] H NS* [3/490] M	NS [7/786] H NS* [4/365] M	HAI scores NS [3/327] M necroinflammation NS [2/389] L Knodell scores NS* [1/157] L	NS [5/626] M NS* [6/751] M	Relapse: NS [4/326] H NS* [2/158] L YMDD mutation: -0.18 (-0.35; -0.01) [6/721] M 0.42 (0.16; 1.09) M -0.23 (-0.32; -0.14)* [1/157] L
Interferon alfa-2b+ corticosteroid vs. IFN alfa-2b	Loss: NS [2/125] M NS* [3/141] L	Loss: NS [2/77] L NS* [3/122] L SC: NS* [2/85] L	NS [2/77] L NS* [6/322] H		NS* [3/170] M	Relapse: NS* [2/141] L
Peginterferon alfa-2a vs. LAM		Loss: 0.08 (0.01; 0.16) M 0.13 (0.05; 0.20)* [1/543] M SC: NS [1/543] L 0.13 (0.06; 0.20) [1/814]* M	-0.15 (-0.22; -0.07) [1/543] M 0.09 (0.04; 0.14) [1/543]* L	Necroinflammation 0.12 (0.02; 0.22) [1/552]* L Fibrosis: NS* [1/552] L HAI: NS [2/1366]* M	-0.29 (-0.42; -0.17) [2/905] 0.57 (0.46; 0.70) [2/905] M 0.13 (0.07; 0.20)* [2/905] H	YMDD mutation -0.25 (-0.31; -0.20) [1/543] L
Peginterferon alfa-2a+LAM vs. LAM		Loss: NS [1/543] L 0.07 (0.00; 0.15)* [1/543] M SC: NS [1/543] L 0.08 (0.01; 0.15) *[1/814] L	0.29 (0.21; 0.37) [1/543] M 0.09 (0.04; 0.13) [1/543]* L	Total scores: NS [2/1366]* H	-0.20 (-0.29; -0.10) [2/905] H 0.13 (0.06; 0.19) [2/905]* H	YMDD mutation -0.22 (-0.28; -0.16) [1/543] L
Peginterferon alfa-2a+LAM vs. peginterferon alfa-2a		Loss: NS [1/542] L NS [1/542]* M SC: NS [1/542] L NS [1/814]* L	0.44 (0.36; 0.51) [1/542] M NS[1/542]* L	Total scores: NS [1/96]* L	NS [1/542] L NS [1/542]* L	YMDD mutation: 0.03 (0.01; 0.06) [1/542] L
Peginterferon alfa-2b vs. interferon alfa-2b	SC: NS* [1/230]	Loss: 0.10 (0.00; 0.21) [1/230]* L			NS [1/230]* L
Peginterferon alfa-2b+LAM vs. LAM	Negative HBVDNA+ HBsAg SC 0.32 (0.14; 0.50) [1/100]	Loss: 0.34 (0.16; 0.52) [1/100] M SC: 0.32 (0.14; 0.50) [1/100] L NS: [1/100]* L	NS [1/100] L NS [1/100]* L	HAI scores NS: [1/100] L	NS [1/100] L	NS: [1/100]* L YMDD mutation: NS [1/100] L
Peginterferon alfa-2b+LAM vs. peginterferon alfa-2b	Loss: NS [1/307]	Loss: 0.12 (0.01; 0.22) [1/307] M NS [2/614]* M SC: NS [1/307]L NS: [1/307]* L		fibrosis scores: NS [1/307]* L necroinflammation scores: NS [1/307] L	0.14 (0.03; 0.24) [1/307] L NS [1/307]* L	YMDD mutation: 0.09 (0.04; 0.14) [1/307] L

SC = seroconversion; NS = not significant; italic = relative risk; *= outcomes off treatments; LAM = lamivudine Level of evidence: L = low; M = moderate; H = high

Telbivudine resulted in greater rates of HBV DNA loss compared to adefovir at 24 (ARD 0.28, 95 percent CI 0.12; 0.44) but not 52 weeks of treatment¹²⁰ in a multinational study of 135 HBeAg-positive, naïve to antiviral drugs patients. Only one drug demonstrated a significant dose response increase in rates of HBV DNA loss, 100–300 mg of lamivudine resulted in greater viral clearance compared to 25–100 mg (pooled ARD 0.21, 95 percent CI 0.10; 0.31).^{133, 137, 145} The length of treatment was not associated with greater response increase in HBV DNA loss.¹³¹ Larger doses or duration of administration of adefovir did not result in larger viral clearance.^{10, 111}–^{113, 120} Only one RCT of adefovir (Adefovir Dipivoxil 437 Study Group) reported greater HBV DNA loss after 30 versus 10 mg (ARD 0.18, 95 percent CI 0.08; 0.27).¹¹² Entecavir did not show a dose response association with HBV DNA loss in a single 24-week, double-blind, multicenter, phase II clinical trial.¹²³ Limited evidence suggested that lamivudine was less effective than adefovir (ARD -0.26, 95 percent CI -0.47; -0.06) in patients with lamivudine-resistant CHB¹¹⁹ and less effective than telbivudine in HBeAg-positive patients with compensated (upper limit of normal for serum ALT was 48 U/L for men and 37 U/L for women) CHB (ARD -0.30, 95 percent CI -0.55; -0.04).¹²⁷

Figure 11

In conclusion, lamivudine and adefovir resulted in HBV DNA clearance that was large in magnitude and maintained for periods up to 24 weeks after the treatment in patients with CHB (moderate to high evidence). Interferon alfa-2b resulted in off treatment HBV DNA loss for 8–24 weeks, the effect was attenuated at longer followup off the treatment (low evidence). Entecavir and adefovir were more effective than lamivudine at the end of the treatment (low). However, sustained differences off the treatments were not significant (entecavir) or have not been examined in RCTs. HBV DNA clearance was greater after combined therapy of adefovir and lamivudine compared to lamivudine alone at the end of the treatment (low to moderate). Long-term sustained effect off therapies has not been examined.

Figure 12

Figure 13

Figure 13

Figure 14

In conclusion, monotherapy with interferon alfa-2b and peginterferon alfa-2a increased off treatment HBeAg loss and seroconversion compared to placebo (moderate evidence and confidence). Lamivudine monotherapy increased rates of HBeAg loss at the end of the treatments and at followup (moderate to high) but did not maintain sustained HBeAg seroconversion (low). Limited low level evidence suggested that peginterferon alfa-2a increased HBeAg loss and seroconversion at followup. Combined therapy with lamivudine was more effective than lamivudine alone (low level of evidence and confidence).

Table 4

Effects of drug therapies for chronic hepatitis B on (more...)

Table 4

Effects of drug therapies for chronic hepatitis B on combined outcomes

Active Treatments vs. Control Treatment	Treatment Duration/Followup Off Therapy, Weeks	Studies/Subjects Enrolled	Estimates (95% CI) *Dose Response Heterogeneity: p Value/I Squared, %	Level of Evidence/Comments
Combined outcomes (virological, histological, biochemical) at end of treatment
Interferon alfa 2b+ prednisone vs. no treatment84	24/0	1/87	0.29 (0.13; 0.46) (RD)	Low. Interferon alfa 2b+ prednisone vs. no treatment increased rates of negative HBV DNA and HBeAg
Interferon alfa-2b vs. no treatment84	24/0	1/84	0.30 (0.13; 0.46) (RD 35MU/week) 0.10 (-0.04; 0.24) (RD 7MU/week)	Low. Interferon alfa-2b vs. no treatment increased rates of negative HBV DNA and HBeAg after 35 but not 7 MU/week
Lamivudine vs. placebo139	96/0	1/136	0.46 (0.32; 0.59) (RD)	Low. Lamivudine vs. placebo increased rates of negative HBV DNA, normal ALT
Interferon alfa-2b vs. no treatment81,91	16–96/0	2/92	0.36 (0.20; 0.51) (RD) 0.473/0%	Low. Interferon alfa-2b vs. no treatment increased rates of negative HBV DNA and normal ALT
Entecavir vs. lamivudine126	52–96/0	2/1418	0.03 (-0.01; 0.08) (RD) 0.49/0%	Moderate. No differences between entecavir vs. lamivudine on negative HBV DNA and HBeAg
Interferon alfa-2b +lamivudine vs. lamivudine75	52/0	1/75	0.13 (-0.05; 0.31) (RD)	Low. No differences between interferon alfa-2b +lamivudine vs. lamivudine alone on loss of HBV DNA+HBeAg seroconversion+HBeAg loss
Lamivudine vs. telbivudine127	52/0	1/63	-0.06 (-0.30; 0.19) (RD)	Low. No differences between lamivudine vs. telbivudine on loss of HBV DNA+HBeAg loss
Peginterferon alfa-2b+lamivudine vs. lamivudine107	60/0	1/100	0.32 (0.14; 0.50) (RD)	Low. Peginterferon alfa-2b+lamivudine vs. lamivudine increased rate of loss of HBV DNA+HBsAg seroconversion
Telbivudine+lamivudine vs. lamivudine127	52/0	1/60	-0.07 (-0.30; 0.16) (RD)	Low. No differences between telbivudine+lamivudine vs. lamivudine alone on loss of HBV DNA and HBeAg
Telbivudine+lamivudine vs. telbivudine127	52/0	1/85	-0.12 (-0.31; 0.06) (RD)	Low. No differences between telbivudine+lamivudine vs. telbivudine on loss of HBVDNA and HBeAg
Entecavir vs. lamivudine122,126	52–96	2/1418	0.03 (-0.01; 0.08) (RD) 0.49/0%	High. No differences between entecavir vs. lamivudine on negative HBV DNA and HBeAg
Interferon alfa 2b, 5MU/day vs. interferon alfa 2b, 1MU/day84	24/0	1/82	0.20 (0.01; 0.38) (RD)*	Low. Interferon alfa-2b in dose 35MU/week increased HBV DNA and HBeAg loss compared to 7MU/week
Interferon alfa 2b+lamivudine vs. lamivudine, 10075	52/0	1/75	0.13 (-0.05; 0.31) (RD)	Low. No differences between interferon alfa-2b+lamivudine vs. lamivudine on loss of HBV DNA+HBeAg seroconversion+HBeAg loss
Interferon alfa 2b, 5MU/day+ prednisolone. interferon alfa 2b, 1MU/day84	24/0	1/85	0.19 (0.01; 0.38) (RD)	Low. Interferon alfa 2b, 35MU/week+pretreatment with prednisone vs. interferon alfa-2b alone, 7MU/ week increased rate of loss of HBV DNA+HBeAg loss
Lamivudine vs. telbivudine127	52/0	1/63	-0.06 (-0.30; 0.19) (RD)	Low. No differences between lamivudine vs. telbivudine on loss of HBV DNA+HBeAg loss
Peginterferon alfa-2b + lamivudine, vs. lamivudine107	60/0	1/100	0.32 (0.14; 0.50) (RD)	Low. Peginterferon alfa-2b+lamivudine vs. lamivudine increased rate of loss of HBV DNA+HBsAg seroconversion
Telbivudine+lamivudine vs. lamivudine127	52/0	1/60	-0.07 (-0.30; 0.16) (RD)	Low. No differences between telbivudine +lamivudine vs. lamivudine on loss of HBV DNA and HBeAg
Telbivudine+lamivudine vs. telbivudine127	52/0	1/85	-0.12 (-0.31; 0.06) (RD)	Low. No differences between telbivudine+lamivudine vs. telbivudine on loss of HBV DNA and HBeAg
Combined outcomes (virological, histological, biochemical) end of followup
Interferon alfa 2b+prednisone vs. no treatment73	24/24	1/76	0.00 (-0.05; 0.05) (RD)	Low. Sparse data (0 events). Interferon alfa 2b+prednisone compared to no treatment did not increase HBV DNA, HBsAg, and HBeAg loss
Interferon alfa-2b vs. no treatment61,73,87,91	16–96/24–144	4/282	0.22 (0.08; 0.36) (RD) 0.042/63.4%	Moderate. Interferon alfa-2b vs. no treatment increased rates of negative HBV DNA+HBeAg with consistent results in multiplicative scale
2.96 (1.40; 6.25) (RR) 0.313/16%	Control rate or duration of treatment and followup could not explain statistical heterogeneity in absolute rate
Lamivudine vs. placebo139	96/24	1/136	0.07 (-0.08; 0.21) (RD)	Low. Lamivudine vs. placebo did not increase loss of HBV DNA and HBeAg at time of followup
Interferon alfa-2b vs. no treatment82,89	16–96/24–48	2/116	0.03 (-0.03; 0.10) (RD) 1/0%	Moderate. Interferon alfa-2b did not increase rate of negative HBV DNA, normal ALT, HBsAg and HBeAg loss
Interferon alfa-2b vs. no treatment82	16/40	1/58	0.19 (-0.01; 0.39) (RD)	Low. Interferon alfa-2b vs. no treatment did not increase rates of negative HBV DNA and HBeAg seroconversion
Interferon alfa-2b vs. no treatment89	16/40	1/58	0.16 (-0.04; 0.35) (RD)	Low. Interferon alfa-2b vs. no treatment did not increase rates of negative HBV DNA, HBeAg loss
Interferon alfa-2b vs. no treatment82	16/40	1/58	0.27 (0.10; 0.43) (RD)	Low. Interferon alfa-2b vs. no treatment increased rated of negative HBV DNA, HBeAg loss and normal ALT
Interferon alfa-2b vs. no treatment81,91	16–96/40–144	2/92	0.28 (0.14; 0.42) (RD) 0.968/0%	Low. Sparse data (small number of events) Interferon alfa-2b vs. no treatment increased rates of negative HBV DNA and normal ALT
Entecavir vs. lamivudine122,125	52–63/76–87	2/995	Random differences in all comparisons	Moderate. Entecavir, 0.50 or 0.1mg vs. lamivudine did not increase rate of HBV DNA and HBeAg loss
Interferon alfa 2b+prednisone vs. interferon alfa 2b73,85	24/48	1/56	-0.17 (-0.42; 0.07) (RD)	Low. Interferon alfa-2b+pretreatment with prednisone vs. interferon alfa-2b alone did not increase rates of HBV DNA loss, HBeAg loss and seroconversion
24/48	1/79	0.07(-0.10;0.24)(RD)
Interferon alfa-2b+lamivudine vs. lamivudine75	52/75	1/75	0.21 (0.06; 0.35) (RD)	Low. Interferon alfa 2b+lamivudine vs. lamivudine alone increased rates of loss of HBVDNA+ HBeAg seroconversion+HBeAg loss
Peginterferon alfa-2b+lamivudine vs. lamivudine106	60/24	1/100	0.20 (0.05; 0.35) (RD)	Low. Peginterferon alfa-2+lamivudine vs. lamivudine alone increased rates of los of HBV DNA+HBeAg loss at two time points of followup
	60/40	1/100	0.12 (-0.01; 0.25) (RD)
60/57	1/100	0.20 (0.05; 0.35) (RD)

Figure 15

Comparative effects of interferon alfa-2b, 35 MU/week with steroid pretreatment on negative HBV DNA and HBeAg was larger compared to interferon, 7 MU/week alone.⁸⁴ Interferon alfa-2b combined with lamivudine did not improve HBV DNA and HBeAg clearance and seroconversion compared to lamivudine alone.⁷⁵ Entecavir or telbivudine did not improve combined outcomes at the end of monotherapy or in combination with lamivudine.^{126, 127}

Figure 13

Figure 16

In conclusion, administration of interferon alfa-2b alone and in combination with lamivudine resulted in off treatment response in combined biochemical and virological outcomes in patients with HBeAg-positive CHB that was large in magnitude. Lamivudine alone increased rates of HBV and HBeAg loss at the end of the treatments but not at followup. The long-term effects of adefovir and telbivudine on combined outcomes have not been evaluated in RCTs.

Figure 17

Effects of drugs on histological outcomes at followup off treatment. Histological improvement in necroinflammatory scores at 24 weeks of followup off treatment (ARD 0.12, 95 percent CI 0.02; 0.22) was reported in only one RCT⁹⁵ after a 48 week administration of peginterferon alfa-2a compared to lamivudine in 552 HBeAg-negative patients.⁹⁵

In conclusion, low-moderate quality evidence suggested improvement in histological scores at the end of monotherapy with adefovir or lamivudine. Off treatment improvement was reported only in HBeAg-negative patients after treatment with peginterferon alfa-2a compared to lamivudine. A higher level of certainty is not possible because reporting is from a relatively small number of short term, small studies, there is inconsistency in findings, and there are limitations in using liver biopsy findings to accurately assess overall histological changes due to treatments. The histological improvement in necroinflammatory scores reported with peginterferon compared to lamivudine was from only one study and at 24 weeks off therapy.

Figure 18

Lamivudine for 12–96 weeks increased ALT normalization with consistent effect size in relative risk compared to placebo or no antiviral treatment (pooled RR 2.42, 95 percent CI 1.94; 3.01).^{67, 131, 136, 139}–^{141, 145} Heterogeneity in pooled absolute risk (pooled ARD 0.22, 95 percent CI 0.13; 0.31) could not be explained by the length of treatment, control rate, or the proportion of HBeAg-positive patients. Comparative effectiveness of entecavir on ALT normalization was greater compared to lamivudine with significant heterogeneity in relative (pooled RR 1.62, 95 percent CI 1.28; 2.06) and absolute risk (pooled ARD 0.22, 95 percent CI 0.11; 0.32).^{90, 121}–¹²⁶ Heterogeneity could not be explained by the dose of entecavir, the duration of treatments, or the proportion of HBeAg-positive patients. The effect of entecavir on absolute risk of ALT normalization was lower in RCTs with higher rates of outcomes after lamivudine administration (meta-regression p value=0.005).

Lamivudine was less effective compared to adefovir (ARD -0.42, 95 percent CI -0.67; -0.18) in 38 American adults with compensated liver disease (Child-Pugh-Turcotte score <7) and lamivudine-resistant hepatitis B virus;¹¹⁹ however, lamivudine administration for 48 weeks was more effective in normalizing ALT compared to peginterferon alfa-2a in HBeAg-positive and negative patients (the Peginterferon Alfa-2a Chronic Hepatitis B Study Group, pooled RR for Peginterferon alfa-2a versus lamivudine 0.57, 95 percent CI 0.46;0.70).^{95, 96} The absolute risk difference for peginterferon alfa-2a versus lamivudine was larger (ARD -0.36, 95 percent CI -0.45; -0.26) in HBeAg-negative patients⁹⁵ than in HBeAg-positive patients (ARD -0.23, 95 percent CI -0.31; -0.15)⁹⁶ (pooled ARD -0.29, 95 percent CI -0.42; -0.17). The same study reported that monotherapy with lamivudine resulted in greater ALT normalization compared to combined treatment (pooled ARD for peginterferon alfa-2a + lamivudine versus lamivudine -0.20, 95 percent CI -0.29; -0.10).^{95, 96} In contrast, a combination of lamivudine with adefovir compared to monotherapy with lamivudine increased the rate of ALT normalization in lamivudine-resistant patients with compensated CHB (pooled ARD 0.32 95 percent CI 0.13; 0.52).^{117, 119}

Figure 19

In conclusion, adefovir and lamivudine monotherapy resulted in ALT normalization that was maintained for up to 24 weeks. Longer term effects are not known. Entecavir and adefovir were more effective than lamivudine at the end of the treatment, while sustained differences have not been investigated. Peginterferon alfa-2b alone and combined with lamivudine normalized ALT at followup off the treatments when compared to lamivudine alone.

In conclusion, limited evidence from a single trial suggested that a prolongation of lamivudine administration increased virological relapse in HBeAg-positive patients. Viral relapse post treatment was lower after entecavir compared to lamivudine in patients HBeAg-positive and naïve to nucleoside analogue. Lamivudine and adefovir increased the incidence of resistant HBV YMDD mutations, while combined therapy with interferon resulted in lower rate of mutation at the end of therapy and at followup. The effect on clinical outcomes including hepatocellular carcinoma and mortality is not known.

Figure 4. Off treatment effectiveness of monotherapy (more...)

   Figure 4. Off treatment effectiveness of monotherapy with interferon compared to no treatment (results from individual studies and pooled analysis with random effects model)

Figure 5. Off treatment effectiveness of combined therapy (more...)

   Figure 5. Off treatment effectiveness of combined therapy with interferon compared to placebo (results from individual studies and pooled analysis with random effects model)

Figure 4

Figure 5

Figure 6. Off treatment effectiveness of reverse transcriptase inhibitors (more...)

   Figure 6. Off treatment effectiveness of reverse transcriptase inhibitors compared to placebo (results from individual studies and pooled analysis with random effects model)

Figure 7. Off treatment comparative effectiveness of (more...)

   Figure 7. Off treatment comparative effectiveness of monotherapy with interferon or reverse transcriptase inhibitors (results from individual studies and pooled analysis with random effects model)

Figure 6

Figure 7

Figure 8. Off treatment comparative effectiveness of (more...)

   Figure 8. Off treatment comparative effectiveness of combined with interferon or reverse transcriptase inhibitors (results from individual studies and pooled analysis with random effects model)

Figure 8

Figure 9. Off treatment comparative effectiveness of (more...)

   Figure 9. Off treatment comparative effectiveness of monotherapy with pegylated interferon alfa-2a compared to lamivudine (results from individual studies and pooled analysis)

Figure 10. Off treatment comparative effectiveness (more...)

   Figure 10. Off treatment comparative effectiveness of combined therapy with pegylated interferon alfa-2a and lamivudine (results from individual studies and pooled analysis)

Figure 9

Figure 10

In conclusion, measures of viral load, ALT levels, and histological scores at followup off the antiviral medications were evaluated in few studies without consistent differences between compared treatments.

EPC Question 2b. What are the known harms associated with treatments of hepatitis B?

Literature search and review strategy. RCTs were identified through an Ovid MEDLINE^® search, using the search terms used to assess efficacy. Additional harms data were obtained from the FDA website and industry prescribing information. We excluded interferon trials that did not report type of interferon (2a or 2b) from the adverse events and lab abnormalities analyses.^{75, 77, 181}

To address question 3 regarding particular patient populations that should not be treated, we describe whether adverse events and withdrawals varied by patient and/or disease characteristics and whether these seemed to be at high enough frequency or severity to preclude treatment.

Table 6

Subjects withdrawing from treatment and experiencing (more...)

Table 6

Subjects withdrawing from treatment and experiencing adverse events from randomized controlled trials

Adverse Event	Number of Studies	Treatment n / N (%)	Control n / N (%)	Absolute Risk Difference [95% CI]	Relative Risk Ratio [95% CI]	Trial(s) Duration
A. Adefovir monotherapy
vs. placebo95,110
Subjects not completing study / treatment	1	26 / 345 (7.5) (10 and 30 mg)	13 / 170 (7.6)	0 [-5 to 5]	0.99 [0.52 to 1.87]	48 weeks
Any adverse event	1	94 / 123 (76.4)	45 / 61 (73.7)	3 [-11 to 16]	1.04 [0.87 to 1.24]
Severe adverse event (grade III or IV)	2	24 / 294 (8.2)	19 / 228 (8.3)	0 [-6 to 6]	0.95 [0.45 to 2.01]
AE leading to discontinuation of study drug	2	4 / 294 (1.4)	1 / 228 (<1)	1 [-1 to 3]	2.34 [0.37 to 14.75]
vs. lamivudine, subjects with lamivudine resistance119
Subjects not completing study / treatment	1	1 / 20 (5)	1 / 19 (5.3)	0 [-14 to 14]	0.95 [0.06 to 14.13]	48 weeks
Any adverse event	1	18 / 19 (94.7)	19 / 19 (100)	-5 [-19 to 8]	0.95 [0.82 to 1.09]
Serious adverse event	1	3 / 19 (15.8)	1 / 19 (5.3)	11 [-9 to 30]	3.0 [0.34 to 26.3]
AE leading to discontinuation of study drug	1	0 / 19	0 / 19	0	-
vs. telbivudine120
Subjects not completing study / treatment	1	2 / 45 (4.4)	2 / 45 (4.4)	0 [-9 to 9]	1.00 [0.13 to 7.43]	52 weeks
Any adverse event	1	27 / 44 (61.4)	34 / 45 (75.6)	-14 [-33 to 5]	0.81 [0.61 to 1.08]
AE leading to discontinuation of study drug	1	0 / 44	0 / 45	0	-
B. Lamivudine monotherapy
vs. placebo136,139,145
Subjects not completing study / treatment	2	33 / 374 (8.8)	16 / 120 (13.3)	-1 [-7 to 4]	0.87 [0.51 to 1.49]	52–104 weeks
Any adverse event	1	224 / 285 (78.6)	56 / 73 (76.7)	2 [-9 to 13]	1.02 [0.89 to 1.18]	52 weeks
Serious adverse event	2	18 / 374 (4.8)	6 / 120 (5)	2 [-1 to 4]	1.24 [0.53 to 2.93]	52–104 weeks
vs. placebo, subjects refractory interferon therapy67
Subjects not completing study / treatment	1	9 / 119 (7.6)	10 / 56 (17.9)	-10 [-21 to 1]	0.42 [0.18 to 0.98]	68 weeks
AE leading to discontinuation of study drug	1	1 / 119 (<1)	4 / 56 (7.1)	-6 [-13 to 1]	0.12 [0.01 to 1.03]
vs. placebo, subjects with advanced liver disease132
Any adverse event	1	335 / 436 (76.8)	178 / 215 (82.8)	-6 [-12 to 0]	0.93 [0.86 to 1.01]	32 months (median)
Serious adverse event	1	54 / 436 (12.4)	38 / 215 (17.7)	-5 [-11 to 1]	0.70 [0.48 to 1.03]
vs. placebo, HBV antigen-negative/ HBV DNA-positive (precore mutant) patients142
Any adverse event	1	40 / 65 (61.5)	28 / 60 (46.7)	15 [-2 to 32]	1.32 [0.95 to 1.84]	26 weeks
vs. pegylated interferon-α-2a monotherapy95,96
Subjects not completing treatment / study	2	71 / 456 (15.6)	45 / 453 (9.9)	6 [1 to 10]	1.57 [1.10 to 2.22]	72 weeks
Any adverse event	2	238 / 453 (52.5)	395 / 448 (88.2)	-36 [-43 to -29]	0.59 [0.51 to 0.69]
Serious adverse event	2	10 / 453 (2.2)	21 / 448 (4.7)	-2 [-5 to 0]	0.47 [0.22 to 0.99]
AE leading to discontinuation of study drug	2	2 / 453 (<1)	21 / 448 (4.7)	-5 [-10 to 1]	0.13 [0.20 to 0.90]
Dose modification due to AE	2	0 / 453	33 / 448 (7.4)	-7 [-10 to -5]	0.03 [0.00 to 0.22]
vs. combined pegylated Interferon-α-2a and Lamivudine95,96
Subjects not completing treatment / study	2	71 / 456* (15.6)	49 / 457* (10.7)	5 [1 to 9]	1.45 [1.03 to 2.03]	72 weeks
Any adverse event	2	238 / 453 (52.5)	395 / 450 (87.8)	-35 [-41 to -29]	0.60 [0.52 to 0.68]
Serious adverse event	2	10 / 453 (2.2)	28 / 450 (6.2)	-4 [-7 to -1]	0.36 [0.18 to 0.73]
AE leading to discontinuation of study drug	2	2 / 453 (<1)	19 / 450 (4.2)	-4 [-6 to -2]	0.13 [0.03 to 0.47]
Dose modification due to AE	2	0 / 453	48 / 450 (10.7)	-10 [-15 to -6]	0.02 [0.00 to 0.15]
C. Telbivudine monotherapy
vs. adefovir (see above)
vs. lamivudine, attributed to study drug (SEBIVO INSERT - 007 GLOBE)143
Subjects not completing treatment / study	1	18 / 680 (2.6)	32 / 687 (4.7)	-2 [-4 to 0]	0.57 [0.32 to 1.00]	52 weeks
Any adverse event		NR	NR
Serious adverse event	1	18 / 680 (2.6)	33 / 687 (4.8)	-2 [-4 to 0]	0.55 [0.31 to 0.97]
AE leading to discontinuation of study drug	1	2 / 680 (<1)	5 / 687 (<1)	0 [-1 to 0]	0.40 [0.08 to 2.08]
AE leading to discontinuation, possibly related to study drug	1	1 / 680 myopathy	1 / 687 urticaria	0 [0 to 0]	1.01 [0.06 to 16.12
D. Entecavir monotherapy (acyclic guanosine derivative)
0.5 mg dose vs. lamivudine, nucleoside-naïve subjects121,122
Subjects not completing study / treatment	2	37 / 688 (5.4)	58 / 675 (8.6)	-3 [-8 to 2]	0.64 [0.33 to 1.26]	E 56–75 weeks
Any adverse event	2	552 / 679 (80.3)	545 / 668 (81.1)	0 [-6 to 6]	1.00 [0.92 to 1.08]	L 56–65 weeks
Serious adverse event	2	48 / 679 (7.1)	54 / 668 (8.1)	-1 [-4 to 2]	0.88 [0.60 to 1.27]
AE leading to discontinuation of study drug	2	7 / 679 (1.0)	18 / 668 (2.7)	-2 [-3 to 0]	0.33 [0.06 to 1.86]
0.5 mg dose vs. lamivudine, nucleoside-naïve subjects. Patient information sheet (Bristol Myers Squibb)
Any Grade 2–4 adverse event	1	102 / 679 (15)	120 / 668 (18)	-3 [-7 to 1]	0.84 [0.66 to 1.06]	Through 2 years
1 mg dose vs. lamivudine in lamivudine-refractory subjects. Patient information sheet (Bristol Myers Squibb)
Any Grade 2–4 adverse event	2	40 / 183 (21.9)	44 / 190 (23.2)	-1 [-10 to 7]	0.94 [0.87 to 1.14]	Through 2 years
vs. lamivudine in lamivudine-refractory subjects124,125
Subjects not completing study / treatment (48 weeks)	2	39 / 283 (13.9)	38 / 191 (19.9)	-12 [-30 to 6]	0.54 [0.36 to 0.81]	48 weeks
AE leading to discontinuation of study drug	2	11 / 277 (4.0)	14 / 190 (7.4)	-5 [-9 to 1]	0.43 [0.12 to 1.54]
Any adverse event	2	225 / 277 (81.2)	155 / 190 (81.6)	0 [-12 to 11]	0.99 [0.87 to 1.14]
Serious adverse event	2	22 / 277 (7.9)	14 / 190 (7.4)	1 [-4 to 6]	1.18 [0.62 to 2.27]
E. Pegylated interferon-α-2a monotherapy (See lamivudine)
F. Combination pegylated interferon-α-2a and lamivudine therapy
vs. lamivudine (see above)95,96
vs. pegylated interferon-α-2a monotherapy95,96
Subjects not completing treatment / study	2	49 / 457 (10.7)	45 / 453 (9.9)	1 [-4 to 5]	1.08 [0.71 to 1.66]	72 weeks
Any adverse event	2	395 / 450 (87.8)	395 / 448 (88.2)	0 [-58 to 4]	1.00 [0.95 to 1.05]
Serious adverse event	2	28 / 450 (6.2)	21 / 448 (4.7)	2 [-1 to 4]	1.33 [0.77 to 2.30]
AE leading to discontinuation of study drug	2	19 / 450 (4.2)	21 / 448 (4.7)	-1 [-6 to 4]	0.90 [0.33 to 2.48]
G. Combination pegylated interferon-α-2b and lamivudine therapy (interferon)
vs. pegylated interferon-α-2b monotherapy99
Subjects not completing treatment / study	1	38 / 152 (25)	37 / 155 (23.9)	1 [-8 to 11]	1.05 [0.71 to 1.55]	78 weeks
Serious adverse event	1	32 total, 17 probably related to therapy
AE leading to discontinuation of study drug	1	12 / 152 (7.9)	11 / 155 (7.1)	1 [-5 to 7]	1.11 [0.51 to 2.44]
I. Pegylated interferon-α-2b vs. interferon-α-2b109
Subjects not completing treatment / study	1	7 / 115 (6.1)	20 / 115 (17.4)	-11 [-19 to -3]	0.35 [0.15 to 0.80]	72 weeks
AE leading to discontinuation of study drug	1	0 / 115	4 / 115 (3.5)	-3 [-7 to 0]	0.11 [0.01 to 2.04]
Any adverse event	75% of patients in each treatment group experienced various clinical forms of drug-related adverse effects.
K. Interferon-α-2b monotherapy (Interferon)
Prolonged (32 weeks) vs. standard (16 weeks) duration61
Dose reduction due to AE		11.5% (7/61) in the prolonged group. Not reported in standard group.
AE leading to discontinuation of study drug		4.9% (3/61) in the prolonged group. Not reported in standard group.
Phase A - all subjects prior to randomization		Dose modification due to AE: 16/162 (10%)
91 IFN (n=21) vs. no treatment (n=21). Study duration was 104 weeks. 5/21 (23.8%) IFN subjects withdrew from study. Dose in reduction in two subjects.
vs. no treatment
90 6 months (n=19) vs. 12 months (n=19). “Treatment was well tolerated by all subjects who finished the study, and no dose modification was needed.”
83 IFN (n=20) vs. no treatment (n=20). Study duration was 68 weeks. 4 IFN and 5 NC subjects did not complete study.
82 IFN (n=30) vs. no treatment (n=28). Study duration was 10 months. One subject with a pre-existing depressive state converted to overt depression and was taken off treatment.
81 IFN (n=25) vs. no treatment (n=25). Study duration was 52 weeks. IFN therapy well tolerated, no serious AE observed.

Lamivudine monotherapy versus placebo. Several placebo-controlled RCTs evaluated the efficacy and safety of orally administered lamivudine (generally 100 mg/day) over 1 to 2 years.^{136, 139, 145} Similar to adefovir, lamivudine was generally well tolerated with no adverse events significantly greater than placebo. Fewer subjects randomized to placebo were likely to complete treatment compared to lamivudine, 13 percent versus 9 percent (ARD -1 percent, 95 percent CI -7; 4) in two trials^{139, 145} and 18 percent versus 8 percent (ARD -10 percent, 95 percent CI -21; 1) in one trial enrolling subjects refractory to interferon.⁶⁷ These differences were not statistically significant. The most common events were upper respiratory tract infections or symptoms, asthenia, abdominal pain, and headache. Overall, adverse events were similar between lamivudine and placebo in one study that enrolled subjects with advanced liver disease (histologically confirmed cirrhosis or advance fibrosis) with the exception of a greater incidence of diarrhea in the placebo group and cough in the lamivudine group over a median duration of treatment of 32 months.¹³² Rate of serious adverse events was slightly higher among the placebo group (18 percent) compared to the lamivudine group (12 percent), with an ARD of -5 percent (95 percent CI -11; 1). There were two deaths during therapy in the lamivudine group, one from lymphoma and one drowning due to a myocardial infarction (high level of evidence).

Lamivudine monotherapy versus peginterferon-alfa-2a monotherapy. Two RCTs, one with HBeAg-positive and one HBeAg-negative subjects, compared lamivudine 100 mg/day (n=456) to subcutaneously administered peginterferon-alfa-2a 180 μg/week monotherapy (n=453).^{95, 96} Subjects were treated for 48 weeks and then followed up for an additional 24 weeks. Significantly more subjects assigned lamivudine discontinued treatment compared to the subjects assigned peginterferon-alfa-2a, 16 percent to 10 percent (RR 1.57, 95 percent CI 1.10; 2.22). However, more subjects treated with peginterferon alfa-2a were more likely to withdraw from a study due to an adverse event, 5 percent compared to <1 percent.(ARD -5 percent, 95 percent CI -10; 1). Dose modification due to an adverse event was required for 7 percent of peginterferon alfa-2a subjects and none of the lamivudine group. Most adverse events were significantly more frequent with peginterferon alfa-2a therapy. An initial flu-like illness was commonly associated with peginterferon alfa-2a treatment, noted by pyrexia, fatigue, myalgia, and headache. Approximately 18 percent of peginterferon alfa-2a subjects reported hair loss compared to 2 percent of lamivudine subjects. Anorexia was reported in 16 percent of peginterferon alfa-2a subjects. Events attributed solely to peginterferon alfa-2a included rigors and an injection-site reaction (due to the subcutaneously administration). Depression was reported in 5 percent and 2 percent of peginterferon alfa-2a and lamivudine subjects, respectively, (RR 0.31, 95 percent CI 0.10; 0.93). One subject receiving lamivudine developed hepatic decompensation after cessation of therapy and died.⁹⁶ There was one death in the peginterferon-alfa-2a group.⁹⁵ This subject had developed thrombotic thrombocytopenic purpura (moderate level of evidence).

Lamivudine monotherapy versus combination peginterferon alfa-2a and lamivudine therapy. Combined peginterferon alfa-2a and lamivudine had a similar adverse event profile as peginterferon alfa-2a when compared to lamivudine monotherapy in two trials reporting.^{95, 96} More lamivudine subjects did not complete treatment but combination subjects were more likely to discontinue treatment due to an adverse event. Fifty-three percent of lamivudine subjects reported any adverse event compared to 88 percent of the combination subjects (ARD -35 percent, 95 percent CI -41; -29). Nearly 11 percent of the combination group required dose modification due to an adverse event, slightly higher than peginterferon monotherapy. The higher frequency of flu-like symptoms (pyrexia, fatigue, myalgia, and headache) was also observed with combination therapy compared to lamivudine as well as significantly greater incidences of hair loss (22 percent versus 2 percent), anorexia (13 percent versus 2 percent), and depression (6 percent versus 2 percent). There were three deaths in the combination group during the treatment period.⁹⁶ These deaths were reported as accidental and unrelated to the study drug (low to moderate level of evidence).

Lamivudine monotherapy versus combination peginterferon alfa-2b and lamivudine therapy. One trial assessed combination peginterferon alfa-2b (1.5 μg/kg of body weight per week up to 100 μg) and lamivudine therapy (n=50) compared to lamivudine monotherapy (n=50) over 52 weeks.¹⁰⁷ Combination therapy resulted in higher frequencies of transient flu-like adverse events, hair loss, and anorexia. Four subjects assigned combination therapy had serious adverse events, including one case of bipolar disorder requiring antidepressant treatment and one case of a severe local reaction. Peginterferon treatment was stopped for all cases.

Lamivudine monotherapy versus combination conventional interferon alfa-2b and lamivudine therapy. Six trials evaluated conventional interferon alfa-2b combined with lamivudine.^{62, 63, 67, 71, 72, 74} Adverse event data were provided primarily by three of the studies.^{62, 63, 67} Combined therapy had a similar safety profile to the pegylated formulation, with high frequencies of pyrexia, headache, fatigue, and myalgia compared to lamivudine monotherapy. Similar to the pegylated formulation, there were significantly higher incidences of alopecia (30–40 percent versus <1–10 percent) and anorexia (19–40 percent versus <1–5 percent) compared to lamivudine monotherapy in two RCTs reporting.^{62, 67} The trial with subjects refractory to interferon reported a significantly lower incidence of depression in the monotherapy group compared to combined therapy, 3 percent to 18 percent (ARD -15 percent, 95 percent CI -25; -5)⁶⁷ (moderate level of evidence).

Telbivudine monotherapy. The GLOBE study evaluated orally administered L-nucleoside analog telbivudine 600 mg/day (n=683) against lamivudine 100 mg/day (n=687) in both HBeAg-positive and negative chronic hepatitis subjects over 52 weeks¹⁴³ (www.fda.gov/cder/foi/label/2006/022011lbl.pdf). Compared to lamivudine, significantly fewer subjects receiving telbivudine withdrew from treatment or were noted to have a serious (Grade 3 or 4) adverse event. Less than 1 percent of subjects in both groups discontinued treatment due to an adverse event. One case of myopathy presumed to be treatment-related occurred in the telbivudine group. Comparable to adefovir and lamivudine, telbivudine is generally well tolerated; most adverse events are typically mild to moderate in severity. Incidences of adverse events were low in both groups (www.fda.gov/cder/foi/label/2006/022011lbl.pdf). In the trial versus adefovir, no serious adverse events were reported and no subject withdrew from therapy due to an adverse event¹²⁰ (low level of evidence).

Adverse events after acyclic guanosine derivative.

Entecavir monotherapy. Two RCTs enrolling nucleoside-naive subjects compared entecavir 0.5 mg/day, an acyclic guanosine derivative was to lamivudine 100 mg/day.^{121, 122} One trial enrolled HBeAg-positive subjects,¹²² and the other HBeAg-negative subjects.¹²¹ All subjects in both trials had not received treatment with a nucleoside analogue. Mean exposure to therapy was 56 to 75 weeks for entecavir and 56 to 75 weeks for lamivudine. Numbers of subjects not completing treatment, reporting any or serious adverse events were similar between treatments. More subjects in the lamivudine group were likely to discontinue treatment due to an adverse event. Rates of individual adverse events were not reported in the trials, but the most frequent events cited included headache, upper respiratory tract infection, upper abdominal pain, nasopharyngitis, dyspepsia, fatigue, back pain, arthralgia, diarrhea, insomnia, cough, and nausea. These events were noted to be mostly mild to moderate. There were four deaths considered unrelated to study therapy, two in both treatment groups. A slightly higher percentage of subjects randomized to lamivudine reported any Grade 2 to 4 adverse event through two years, 18 percent versus 15 percent for entecavir (Bristol Myer package insert). Pooled analysis of two studies enrolling subjects refractory to lamivudine found rates of any Grade 2 to 4 adverse events were similar between the entecavir 1 mg dose group and lamivudine group, 22 percent and 23 percent, respectively (Bristol Myer package insert). Three deaths were reported in one trial enrolling lamivudine-refractory subjects, none were deemed related to the study medication by the investigator¹²⁵ (low to moderate evidence).

Adverse events after interferons.

Pegylated interferon alfa-2a versus lamivudine monotherapy and combination therapy (see above).

Combination of peginterferon alfa-2a and lamivudine therapy versus peginterferon alfa-2a monotherapy. Combined peginterferon alfa-2a and lamivudine had a similar safety profile and withdrawal rates compared with peginterferon alfa-2a monotherapy.^{95, 96} The most common treatment-related adverse events in both groups included flu-like symptoms (pyrexia, fatigue, myalgia, and headache). Depression was reported by 6 percent (n=16) of the monotherapy group and 5 percent (n=13) of the combination group in one study enrolling HBeAg-positive subjects.⁹⁶ Four subjects died during the study periods, one in the peginterferon alfa-2a group⁹⁵ and three in the combined therapy group⁹⁶ (low to moderate evidence).

Combination of pegylated interferon alfa-2b and lamivudine therapy versus pegylated interferon monotherapy. One 52 week (26 week followup period) trial enrolling 307 subjects evaluated combined pegylated interferon alfa-2b 100 μg/week and lamivudine therapy in comparison to pegylated interferon alfa-2b monotherapy.⁹⁹ Approximately one-fourth of all subjects did not complete treatment. Between 7 and 8 percent discontinued treatment due to an adverse event. Overall, rates of adverse events were comparable between groups and the most common events were flu-like symptoms. There were 32 serious adverse events reported for both treatment arms. Seventeen events were likely to be attributed to therapy and included hepatitis flare (4), depression (3), severe neutropenia (3), and one case each of psychosis, seizures, pancreatitis, anxiety, dizziness, diarrhea, and syncope. All serious adverse events were reversible after treatment cessation. A study which followed these subjects an additional 26 weeks concluded that the most important predictors of dose reduction or study withdrawal were pre-existing cirrhosis and neutropenia¹⁰³ (low level of evidence).

Pegylated interferon alfa-2b versus interferon alfa-2. One Chinese trial (N=230) compared pegylated interferon alfa-2b 1.0 μg/kg/week monotherapy to interferon alfa-2b 3 MU/week monotherapy.¹⁰⁹ Significantly more subjects receiving conventional interferon did not complete treatment and followup compared to the pegylated interferon group, 17 percent versus 6 percent (ARD -11 percent, 95 percent CI -19; -3). Seventy-five percent of patients in each group reported drug-related adverse events, mainly flulike symptoms and fever. Adverse events lead to four subjects (4 percent) in the conventional interferon group to discontinue treatment.

Conventional interferon alfa-2b and lamivudine. One trial of interferon nonresponders to interferon compared 24 weeks of combined interferon alfa-2b 10 MU/week and lamivudine therapy (n=63) versus 52 weeks of placebo (n=56).⁶⁷ The percentages of subjects not completing treatment were comparable. There were significantly more flu-like adverse events observed with interferon therapy in the combined group. A multinational trial comparing combined therapy (n=76) to interferon monotherapy (n=70) found rates of adverse events were similar between groups with the exception of headache, which had significantly higher incidence in the combined group (93 percent to 67 percent, ARD 26 percent, 95 percent CI 14; 39).⁶² A Turkish study of 49 subjects reported a significantly higher incidence of mouth dryness in the combined therapy group (76 percent) compared to the interferon monotherapy group (33 percent) (ARD 57 percent, 95 percent CI 33; 81).⁶⁸

Interferon alfa-2b. Several RCTs compared different regimens of interferon alfa-2b therapy to no treatment. The trials reported that treatment was generally well-tolerated, but most subjects developed transient mild flu-like symptoms. In addition, Chung (N=65) reported anorexia/nausea in 22 percent of all subjects.⁶⁶ Janssen noted dose reduction was required in 12 percent of subjects in the “Prolonged Treatment” group due to depression, fatigue, hair loss, and headache⁶¹ (low to moderate evidence).

Adefovir. In a pooled analysis of two trials, similar incidences of Grade 3 or 4 laboratory abnormalities were observed for adefovir and placebo with the exception of significant increases in ALT and AST levels (www.fda.gov/medwatch/safety/2006/Oct_PIs/Hepsera_PPI.pdf). Over 40 percent of placebo subjects had ALT levels more than five times the upper limit of normal (ULN) in the placebo group compared to 20 percent of the adefovir 10 mg group (RR 0.49, 95 percent CI 0.37; 0.65). AST levels greater than five times the ULN were observed in 23 percent and 8 percent of the placebo and adefovir subjects, respectively. An increase in serum creatinine ≥0.3 mg/dL from the baseline level was observed in 4 percent of adefovir subjects versus 2 percent of placebo subjects with adequate renal function at week 48 of treatment. No subject developed an increase ≥0.5 mg/dL at week 48. After extended adefovir treatment of an additional 48 weeks, two subjects had increases in serum creatinine ≥0.5 mg/dL from baseline, leading to discontinued treatment in one subject.¹⁰ A black box warning from the prescribing information states subjects with or at risk of impaired renal function may develop nephrotoxicity with chronic administration of adefovir (www.fda.gov/medwatch/safety/2006/Oct_PIs/Hepsera_PPI.pdf). An analysis of renal safety utilizing the study population of the trial by Marcellin⁹⁵ found a greater occurrence of Grade 1 and 2 hematuria and proteinuria in the adefovir 30 mg group compared to placebo.¹¹⁶ In the trial versus telbivudine, one adefovir subject had an elevated serum creatinine level that returned to normal range after switching to telbivudine after study cessation.¹²⁰ Grade 3 or 4 neutropenia was reported for one subject in each treatment arm. Each case resolved without dose reduction or treatment interruption. In one 48 week trial of lamivudine-refractory subjects, there were seven Grade 3 events (37 percent) compared to two Grade 3 and 4 events (10 percent) in the combined adefovir/lamivudine group¹¹⁹ (moderate to high evidence).

Lamivudine versus placebo. A one year placebo-controlled trial of 385 Chinese subjects reported 10 subjects had abnormal liver function tests considered to be of major clinical concern, five in the lamivudine group (2 percent; four received 100 mg and one 25 mg) versus five in the placebo group (7 percent) (RR 0.26, 95 percent CI 0.08; 0.86).¹⁴⁵ A trial of 143 American subjects found the frequency of Grade 3 or 4 lab abnormalities similar between lamivudine and placebo during the course of treatment.¹³⁶ However, 25 percent of lamivudine subjects had an ALT level at least three times the baseline level (Grade 3 or 4 abnormality) compared to eight percent of placebo subjects during the 16 week post-treatment period (p=0.01). In subjects with advanced liver disease, 12 percent of subjects receiving lamivudine had elevations in serum ALT at least three times the level at baseline compared to one-fourth of the subjects receiving placebo (ARD -13 percent, 95 percent CI -20; -7)¹³² (moderate evidence).

Lamivudine versus peginterferon alfa-2a monotherapy. Comparable to the adverse event profile, rates of lab abnormalities were significantly higher in the peginterferon alfa-2a monotherapy group compared to lamivudine. In a pooled analysis of two RCTs (N=901), dose modification was required for 46 percent of peginterferon alfa-2a recipients versus none of the lamivudine recipients (ARD -46 percent, 95 percent CI -51; -42.^{95, 96} Approximately 37 percent (95 percent CI 32; 41) of peginterferon alfa-2a subjects required dose medication due to a lab abnormality (79 percent of all dose modifications), with neutropenia and thrombocytopenia cited as the most common causes (moderate evidence).

Lamivudine versus combination peginterferon alfa-2a and lamivudine therapy. Combined peginterferon alfa-2a and lamivudine had a similar lab abnormality profile as peginterferon alfa-2a monotherapy in two trials reporting (N=903).^{95, 96} No subject assigned lamivudine required dose modification, while 47 percent of combined therapy subjects need alterations in the therapy regimen (ARD -47 percent, 95 percent CI -52; -43). Lab abnormalities accounted for 78 percent of all dose modifications, primarily due to neutropenia, thrombocytopenia and elevated ALT (moderate evidence).

Lamivudine versus combination peginterferon alfa-2b and lamivudine therapy. One Chinese trial randomizing 100 subjects assessed combination peginterferon alfa-2b and lamivudine therapy (n=50) compared to lamivudine monotherapy (n=50) over at least 78 weeks.¹⁰⁷ Dose reduction was required for five (10 percent) pegylated interferon subjects due to anemia (one patient), neutropenia (three patients), and/or thrombocytopenia (four patients). One combined therapy subject had peginterferon withheld for two doses due to a severe hepatic flare-up. No lamivudine subject required a reduction of dose. There were no significant differences in the incidence of lab abnormalities between groups (low evidence).

Lamivudine versus combination conventional interferon alfa-2b and lamivudine therapy. Several trials evaluated conventional interferon alfa-2b combined with lamivudine compared to lamivudine monotherapy. One Italian trial (N=151) with a study duration of 100 weeks found no significance in rates of lab abnormalities between groups.⁶³ One trial of interferon-refractory subjects noted similar frequencies of lab abnormalities between groups.⁶⁷ Elevated ALT (≥2 times the baseline level) was significantly greater in the combined therapy group (n=63) during primary treatment, 48 percent versus 26 percent for the lamivudine group (n=119) (ARD -22 percent, 95 percent CI -36; -7). Incidence of neutropenia was also significantly greater among subjects assigned combined therapy, 16 percent versus 1 percent of lamivudine recipients. A Turkish trial (N=80) reported four cases of neutropenia occurred with combined therapy.⁷¹ Two cases each required temporary and permanent dose modification, respectively. Thrombocytopenia occurred in 11 cases (28 percent) in the combined therapy group versus three subjects in the lamivudine group (8 percent) (ARD -20 percent, 95 percent CI -36; -4) (moderate to high evidence and confidence).

Telbivudine. Significantly greater incidences in creatine kinase (CK) elevations were associated with telbivudine therapy compared to therapy with lamivudine. Data from the 52 week GLOBE trial showed 68 percent of telbivudine recipients (n=680) had a Grade 1–4 CK elevation compared to 39 percent of lamivudine subjects (n=687) (ARD 29 percent (95 percent CI 24; 34) product monograph (www.fda.gov/cder/foi/label/2006/022011lbl.pdf). Grade 3–4 CK elevations were reported for 7.5 and 3 percent of the telbivudine and lamivudine subjects, respectively. CK elevations decreased spontaneously to Grade 2 or lower in two-thirds of the telbivudine recipients and approximately three-fourths of the lamivudine subjects by the next clinical visit.¹⁴³ In the telbivudine group, two subjects required discontinuation and three subjects required interruption of treatment due to CK toxicity product monograph (www.fda.gov/cder/foi/label/2006/022011lbl.pdf). There were higher frequencies of Grade 3–4 elevations in ALT and aspartate aminotransferase (AST) in the lamivudine group compared to the telbivudine group. ALT levels greater than three times the baseline level occurred in 6 percent of the lamivudine-assigned subjects versus 3 percent of the telbivudine subjects (RR 0.59, 95 percent CI 0.36; 0.95).¹⁴³ Analysis of categories of ALT flares (≥2 times the baseline level) after 24 weeks of treatment found ALT flares were more likely to occur with lamivudine therapy (5 percent) than telbivudine (1 percent) product monograph (www.fda.gov/cder/foi/label/2006/022011lbl.pdf) (low to moderate evidence).

Laboratory abnormalities/toxicities after acyclic guanosine derivative.

Entecavir monotherapy. Elevations in ALT occurred more frequently in the lamivudine group (n=668) compared to the entecavir group (n=679), particularly during post-treatment.^{121, 122} During the 24 week followup period, ALT flares (ALT >2 times the baseline level and >5 times the ULN) occurred in 24 percent and 9 percent of the lamivudine and entecavir groups, respectively (absolute risk difference -14 percent, 95 percent CI -21; -6). Elevations in ALT were also observed more frequently in the lamivudine group compared to entecavir 1 mg in a pooled analysis of two trials assessing lamivudine-refractory subjects through 2 years of study duration (N=373) [Patient information sheet, (Bristol Myers Squibb http://www.fda.gov/medwatch/safety/2007/Baraclude_PI_jul2407.pdf)]. ALT flares >5 times the ULN occurred in 24 percent of subjects assigned lamivudine versus 12 percent assigned entecavir (ARD -12 percent, 95 percent CI -20; -4). AST levels >5 times the ULN were also significantly greater in the lamivudine group (17 percent) compared to the entecavir group (5 percent) (ARD -12 percent, 95 percent CI -18; -6) (moderate evidence).

Laboratory abnormalities/toxicities after interferons.

Pegylated interferon alfa-2a versus lamivudine monotherapy and combination therapy (see above).

Combination peginterferon alfa-2a and lamivudine therapy versus peginterferon alfa-2a monotherapy. Combined peginterferon alfa-2a and lamivudine therapy and peginterferon alfa-2a monotherapy had similar laboratory abnormality profiles.^{95, 96} Over 45 percent of both groups required dose modification. Nearly 80 percent of dose modifications were due to a lab abnormality, mainly neutropenia and thrombocytopenia. Elevated ALT levels occurred more frequently in the monotherapy group, 9 percent versus 4 percent (ARD -5 percent, 95 percent CI -10; 0)⁹⁵ (low to moderate evidence).

Combination pegylated interferon alfa-2b and lamivudine therapy versus pegylated interferon monotherapy. No significant differences in dose modifications were reported between the treatment groups, and nearly 70 percent remained on full-dose treatment at the end of therapy.⁹⁹ Frequencies of hematologic events, neutropenia (21–26 percent) and thrombocytopenia (11–13 percent), were also similar in the combined and monotherapy groups (low evidence).

Pegylated interferon alfa-2b versus interferon alfa-2. The trial by Zhao (N=230) reported four subjects (6 percent) in the conventional interferon group with elevated ALT levels and /or increased bilirubin levels discontinued treatment.¹⁰⁹ No subjects in the pegylated group discontinued therapy due to a lab abnormality (low evidence).

Conventional interferon alfa-2b and lamivudine. In a trial evaluating subjects refractory to interferon treatment, frequencies of abnormal ALT or AST, abnormal enzymes and neutropenia were not significantly different from combined conventional interferon alfa-2b and lamivudine therapy (n=63) versus placebo (n=56) through the 68 treatments and followup duration.⁶⁷ During the 52 week treatment period, 48 percent of combined therapy subjects had ALT levels at least two times the baseline level compared to 20 percent of placebo subjects (absolute risk difference 28 percent, 95 percent CI 12; 440. The Schalm trial found hepatic flares (ALT levels at least 500 IU/L and greater than two times the baseline level) were observed more frequently in the interferon monotherapy group (11 percent; 8/70 subjects) compared to combined therapy (0 percent; 0/75 subjects) during the 24 week treatment period.⁶² There was no difference in the incidence of flares during the 40-week post-treatment period (low evidence).

Interferon alfa-2b. Few of the small studies comparing different regimens of interferon alfa-2b therapy or to no treatment reported lab abnormalities. Low incidences (up to 6 percent) of thrombocytopenia and neutropenia were observed in three trials^{61, 66, 81} (low evidence).

EPC Question 3a. Are there differences in efficacy/effectiveness of treatments for treatment naïve versus drug-resistant patients, HBeAg-positive versus HBeAg-negative patients, or for other subpopulations (as defined previously)?

Clinical outcomes. Information is very limited because none of the studies reported mortality, liver related mortality, or hepatocellular carcinoma. Reported subgroup analyses appear to be exploratory in nature, varied in their definitions of outcomes and predictors and lack confirmatory findings. The French Multicenter Group⁸⁵ conducted subgroup analysis of incident cirrhosis after 24 weeks of interferon alfa-2 with steroid administration among patients with baseline ALT more than three times the upper limit of normal and did not find significant protective effects of combined therapy compared to interferon monotherapy (RR 0.40, 95 percent CI 0.04; 4.19). One small Korean trial¹⁴¹ analyzed odds ratio of hepatic decompensation defined as an increase in Child-Turcotte-Pugh score of two or more points in patients with lamivudine-resistant mutants and found that gender, baseline HBeAg-positive status, and elevated ALT or viral load were not associated with progressive hepatic functional deterioration. However, patients with baseline platelet count less than versus greater than 65,000/μl experienced hepatic decompensation less often (RR 0.98, 95 percent CI 0.97; 0.99). Investigators for the Cirrhosis Asian Lamivudine Multicentre Study Group conducted subgroup analysis of overall disease progression, defined by hepatic decompensation, hepatocellular carcinoma, spontaneous bacterial peritonitis, bleeding gastroesophageal varices, or death related to liver disease in males and females treated with lamivudine or placebo and found significant protective effects of lamivudine in both genders (RR in males 0.53, 95 percent CI 0.33; 0.84 and in females 0.12, 95 percent CI 0.03; 0.58).¹³²

In conclusion, little evidence is available regarding effectiveness of antiviral agents on clinical outcomes, especially to determine which groups should be treated or whether clinical outcomes vary according to key patient/disease characteristics. Clinical outcomes in Asian patients treated with lamivudine compared to placebo did not vary according to gender.

The vast majority of the studies described nonclinical outcomes among patient subgroups with different baseline liver enzymes, virological, or histological status. Even in these situations, studies did not provide power calculations and did not analyze baseline similarities in such subgroups. Analyses were exploratory in nature, lacked justification for subpopulation thresholds utilized, often were inconsistent in findings or thresholds employed, combined several different outcomes into a global effectiveness measure, and may have been selectively reported. Therefore, an accurate assessment and clear/concise summary is difficult (insufficient evidence).

Age (two studies, two antiviral agents, lamivudine and peginterferon, used as monotherapy). Limited evidence from two studies suggested that increased patient age was associated with lower sustained response to pegylated interferon alfa-2a or lamivudine as defined by HBV DNA clearance and ALT normalization. Younger patients had higher rates of HBV DNA clearance and ALT normalization (adjusted RR 1.26, 95 percent CI 1.00; 1.50) per 10 year decrease in age after 48 weeks of treatment and 24 weeks of treatment-free followup with peginterferon alfa-2a or lamivudine.⁹³ Sustained response to pegylated interferon alfa-2a was lower (adjusted RR 0.39, 95 percent CI 0.16; 0.92) among patients above 25 years compared to those below.¹⁰⁹

In conclusion, low levels of evidence suggested that decreased patient age was associated with enhanced treatment efficacy as measured by HBV DNA clearance and ALT normalization. No evidence was available for clinical outcomes.

Body weight (one study, Peginterferon Alfa-2a HBeAg-Negative Chronic Hepatitis B Study Group, two antiviral agents used as monotherapy). Patient body weight was not associated with sustained HBV DNA loss and ALT normalization after 48 weeks of treatment and 24 weeks of treatment-free followup with pegylated interferon alfa or lamivudine (adjusted RR 1.03, 95 percent CI 0.81; 1.3 per 10 kg increase in body weight).⁹³

In conclusion, low levels of evidence suggested that baseline body weight is not associated with treatment efficacy as measured by HBV DNA loss and ALT normalization. No evidence was available for clinical outcomes.

Duration of hepatitis (one study- Lamivudine Italian Study Group Investigators, two antiviral agents used as combination therapy). Sustained virologic response at 48 weeks off therapy (suppression of serum levels of HBeAg and HBV DNA) to interferon alfa-2b combined with lamivudine was greater in those with an estimated duration of hepatitis of 10 years or less after adjustment for patient gender and age (adjusted OR 2.55, 95 percent CI 1.26; 19).⁶³

In conclusion, low levels of evidence suggested that patients with longer duration of hepatitis responded to lamivudine therapy 2.5 times less frequent compared to those with shorter duration of the disease. No evidence was available for clinical outcomes or other therapies.

Gender (five studies, three antiviral agents, peginterferon, interferon, and lamivudine, used as monotherapy). Patient gender demonstrated inconsistent associations in five studies that evaluated this factor.^{72, 93, 109, 132, 141} One study reported that adjusted odds ratios of sustained combined response (ALT normalization and HBV DNA level <20,000 copies/ml) 24 weeks off peginterferon alfa-2a or lamivudine therapy was greater among women compared to men (OR 1.93, 95 percent CI 1.1; 3.4).⁹³ The association was not significant at year 1 off treatment (OR males versus females 0.68, 95 percent CI 0.34; 1.37).⁹³ Multivariate adjusted odds ratio of sustained 24 week off therapy combined response (HBeAg-negative, HBV DNA <5 log 10 copies/mL, and normal ALT level) to peginterferon alfa-2b or interferon alfa-2b was not significant (OR 0.59, 95 percent CI 0.22; 01.6) in males compared to females.¹⁰⁹

In conclusion, low levels of evidence suggested that disease progression or treatment induced sustained ALT normalization and HBV DNA clearance did not vary by gender.

Interferon alfa-2b increased maintained HBeAg loss off treatment compared to placebo among patients with pretreatment HAI score 5–9 (RR 5.76, 95 percent CI 1.48; 22.42) but failed in patients with pretreatment HAI score 0–4 or >10.⁶⁴ The same trial reported that interferon alfa-2b combined with lamivudine compared to placebo increased sustained HBeAg clearance in the same subpopulation with pretreatment HAI score 5–9 (RR 5.32, 95 percent CI 1.51; 18.72) with no effects in those with pretreatment HAI score 0–4 (RR 3.39, 95 percent CI 0.89; 12.87) or >10 (RR 1.79, 95 percent CI 0.89; 3.59).⁶⁴ Off treatment virologic response to interferon alfa-2b combined with lamivudine increased in those with a baseline HAI Knodell inflammation score of seven or more, independent of gender and age (adjusted RR 2.91, 95 percent CI 1.04; 8.22).⁶³ Baseline fibrosis scores were not associated with better sustained response to this treatment.⁶³ Presence of steatosis did not modify the effect of peginterferon alfa-2a combined with lamivudine on sustained response defined as HBV DNA disappearance and ALT normalization in both HBeAg-positive and negative patients;⁹⁷ however, the adjusted rates of sustained response were greater per increase in baseline Knodell HAI (adjusted OR 14.97, 95 percent CI 2.43; 92.28).⁹⁷

In conclusion, there was a low level of evidence that treatment induced followup histology, HBeAg loss or DNA disappearance and ALT normalization varies by baseline histology severity. There was no evidence for clinical outcomes.

At followup off the treatment, interferon alfa-2b, 5MU/day compared to no treatment increased HBV DNA and HBeAg loss among patients with baseline HBV DNA 2-99 pg/ml (RR 5.24, 95 percent CI 1.22; 22.50) but failed among those with baseline HBV DNA 100–200 pg/ml or >200 pg/ml⁸⁴ without a significant HBV DNA unit dose response association (RR 1.87, 95 percent CI 0.59; 5.87 per one unit increase in HBV DNA).⁶¹ Interferon alfa 2b, 10 MU three times per week versus no antiviral treatment increased sustained rates of HBeAg loss among patients with baseline HBV DNA <10pg/ml (RR 3.18, 95 percent CI 1.25; 8.05) but failed in those with higher viral load HBV DNA >10pg/ml.⁶¹ Interferon alfa-2b with steroid pretreatment increased sustained off treatment rates of HBV and HBeAg loss among patients with baseline HBV DNA 2-99 pg/ml (RR 5.38, 95 percent CI 1.26; 22.84) but failed in those with elevated baseline viral load (HBV DNA >100 pg/ml).⁸⁴ While monotherapy with interferon alfa-2b was more effective in patients with lower compared to elevated baseline HBV DNA, combined administration of interferon alfa-2b with lamivudine resulted in greater sustained HBV DNA loss and HBeAg seroconversion in patients with elevated baseline HBV DNA >10⁷ copies/mL.⁷⁵ Peginterferon alfa-2a resulted in greater sustained response compared to lamivudine in patients with baseline HBV DNA range of 25–75 percentile^{93, 96} with random differences among those below 25 percent or above the 75th percentile. Sustained combined response was increased by one log 10 unit (copies/ml) decrease in baseline HBV DNA (adjusted OR 1.28, 95 percent CI 1.10; 1.40).⁹³ Baseline mean viral load (copy/mL) was not associated with sustained response to the combined therapy with peginterferon alfa-2a plus lamivudine versus lamivudine alone.⁹⁷

In conclusion, there was a low level of evidence that treatment induced HBeAg loss, DNA normalization, or histology varies with baseline viral load. There was no evidence for clinical outcomes.

Baseline HBeAg status (Appendix E Tables 9 and 10)

Evidence from trials that combined patients with HBeAg-negative and positive baseline status. Lamivudine at the end of 130 week administration decreased disease progression defined by hepatic decompensation, hepatocellular carcinoma, spontaneous bacterial peritonitis, bleeding gastroesophageal varices, or death related to liver disease compared to placebo among patients HBeAg-positive at baseline (RR 0.30, 95 percent CI 0.16; 0.55).¹³² There were no significant effects among HBeAg-negative patients (RR 0.72, 95 percent CI 0.36; 1.43).¹³² Entecavir in a dose of 1mg/day compared to lamivudine resulted in higher rates of undetectable HBV DNA and normal ALT level among patients with HBeAg-negative baseline status (RR 18.38, 95 percent CI 1.18; 285.96) with random differences in HBeAg-positive patients.¹²⁴ Telbivudine compared to lamivudine reduced the rates of detectable HBV DNA (RR 0.67, 95 percent CI 0.54; 0.82) and improved necroinflammatory scores, with no worsening in the Knodell fibrosis score (RR 1.15, 95 percent CI 1.03; 1.27) among patients with HBeAg-positive baseline status with random differences in HBeAg-negative patients.¹⁴³ Telbivudine compared to lamivudine for 52 weeks in Chinese patients with compensated hepatitis B resulted in better outcomes in HBeAg-positive patients with no difference in small subsample of HBeAg-negative patients. ⁵⁷

Table 7

Effects of antiviral drugs on HBeAg-negative patients (relative (more...)

Table 7

Effects of antiviral drugs on HBeAg-negative patients (relative risk from individual RCTs)

Treatments	Clinical or Combined Outcomes	Biochemical	Virological (HBV DNA, HBsAg)	Histological	Mutation
Adefovir (dose, time)		NS10	HBsAg seroconversion-NS10 HBV DNA loss- NS10	Improved histology-NS10	NS10
Adefovir vs. placebo		2.45 (1.61; 3.73)1.79 (1.07; 3.00)*10,110	HBsAg seroconversion10 1.52 (0.06; 36.46) loss HBV DNA11063.50 (4.00; 1009.28) loss HBV DNA 10* 8.83 (2.94; 26.52)	Failure1100.11 (0.04; 0.27) Improved10,110 Significant improvement in necroinflammatory or fibrosis scores with NS changes in total scores	NS10,110
Interferon alfa 2b+lamivudine vs. lamivudine	NS71	Flare-NS71,74 Normalization of ALT: 1.29 (0.89; 1.86)74 1.30 (0.46; 3.71)74 1.08(0.82;1.41)761.45 (1.02; 2.05)76*	HBV DNA loss-NS71,74,76 Relapse 0.30 (0.09;0.93)74 HBsAg loss-NS71,74,76		0.03 (0.00; 0.55)0.18 (0.04; 0.73)76
Interferon alfa 2b vs. no treatment	Loss of HBV DNA and normalization of ALT81* 5.50 (1.36; 22.32) Loss of HBV DNA and normalization of ALT91* 13(0.78; 217.03) Relapse - NS81* 2.00(0.40; 9.95)	Flare 0.27 (0.11; 0.67)91	HBsAg loss - NS 91	Improved histology-NS 91 Resistance-NS91
Peginterferon alfa-2a+lamivudine vs. lamivudine	Normalization of ALT and loss of HBV DNA95* 2.64 (1.36; 5.11)	Normalization ALT950.66 (0.55; 0.79) Normalization ALT*951.34 (1.09; 1.64)	HBV DNA loss951.17 (1.06; 1.30) HBV DNA loss*952.92 (1.57; 5.44)	Failure NS95 Improved histology-NS95
Peginterferon alfa-2a+placebo vs. lamivudine	Normalization of ALT and HBV DNA loss*952.36 (1.20; 4.64)	Normalization ALT950.51 (0.41; 0.63) Normalization ALT*951.31 (1.07; 1.61)	HBV DNA Loss950.84 (0.73; 0.97)2.83 (1.52; 5.29)*	Failure NS95 Improved necroinflammatory scores95* 1.39 (1.06; 1.82) Improved HAI-NS*95 Improved fibrosis-NS95

* off treatment; NS = not significant

Evidence from trials that included exclusively patients with HBeAg-negative status (11 studies; four antiviral agents used as mono or combination therapy). We reviewed the drug effects reported in the RCTs that enrolled all HBeAg-negative patients,^{10, 71, 74, 76, 79, 81, 91, 93, 95, 110, 111} including the Peginterferon Alfa-2a HBeAg-Negative Chronic Hepatitis B Study Group^{93, 95} and the Adefovir Dipivoxil 438 Study Group^{10, 110, 111} (Table 7). Adefovir improved biochemical, virological, and histological outcomes in HBeAg-negative patients at the end of drug administration and at followup off the treatment without development of genetic mutations.^{10, 110} Interferon alfa-2b combined with lamivudine was not more effective compared to lamivudine alone to improve combined virologcial with biochemical⁷¹ or virological^{71, 74, 76} outcomes in HBeAg-negative patients but lowered the rates of relapse defined as reappearance of detectable serum HBV DNA by polymerase chain reaction (PCR) after an initial virologic response⁷⁴ and genetic viral mutations.⁷⁶ Interferon alfa-2b compared to no antiviral treatments increased sustained HBV DNA clearance and ALT normalization in a large RCT⁸¹ with random differences in a small study;⁹¹ however, it failed to increase HBsAg loss and did not improve histological scores.⁹¹ Peginterferon alfa-2a compared to lamivudine improved sustained biochemical and virological outcomes and necroinflammatory scores but failed to improve fibrosis scores.⁹⁵ Peginterferon alfa-2a combined with lamivudine compared to lamivudine improved sustained biochemical and virological outcomes with no differences on liver histology.⁹⁵

In conclusion, low level of evidence suggested that lamivudine monotherapy decreased disease progression among patients HBeAg-positive at baseline but not in HBeAg-negative patients. Telbivudine increased viral clearance and improved histology compared to lamivudine in HBeAg-positive but not in HBeAg-negative patients. Entecavir was more effective compared to lamivudine in HBeAg-negative patients to increase viral clearance and ALT normalization. Patients without HBeAg at baseline experienced improvement in biochemical, virological, and histological outcomes after adefovir therapy and pegylated interferon alfa-2a monotherapy or combination with lamivudine. There was no evidence for clinical outcomes.

Table 8

Effects of antiviral drugs for chronic hepatitis B (more...)

Table 8

Effects of antiviral drugs for chronic hepatitis B on off treatment outcomes in patient subpopulations

Outcome	Comparison	Weeks of Treatments/ Followup off the Treatment	N Studies/ Enrolled	Author	Relative Risk or Odds Ratios (95% CI)	Level of Evidence From Individual Studies is Low for All Comparisons/Comments
Baseline ALT
HBV DNA loss	Interferon alfa 2b+corticosteroid vs. interferon alfa 2b	24/24	1/115	Wai, 200265	1.22 (1.05; 1.42)	HBV DNA loss was more frequent among patients with elevated baseline ALT
HBV DNA and HBeAg loss	Interferon alfa 2b+corticosteroid vs. no treatment	24/24	1/43	Perrillo, 199084	7.82 (1.02; 59.88)	Loss of HBV DNA and HBeAg was greater among patients with baseline ALT <100U/L with random differences among those with baseline ALT 100–200 and >200U/L
Odds ratio of HBeAg and HBV DNA loss independent of gender and age	Interferon alfa 2b+lamivudine vs. lamivudine	24/48	1/150	Barbaro, 200163	3.12 (1.43; 6.82)	Adjusted odds of virologic response were higher in patients with baseline ALT >150UL
HBeAg loss	Interferon alfa 2b+lamivudine vs. placebo	24/28	1/331	Perrillo, 200264	2.90 (1.35; 6.27)	HBeAg loss was higher among patients with ALT >2 but <5ULN with random differences among those with <1ULN or 1–2ULN
HBeAg seroconversion	Interferon alfa 2b+lamivudine vs. placebo	24/28	1/331	Perrillo, 200264	2.70 (1.10; 6.58)	HBeAg seroconversion was greater among patients with >2–<5 ULN
3.27 (1.03; 10.39)	HBeAg seroconversion was greater among patients with ALT >5 ULN with random differences among those with baseline ALT<1ULN or 1–2ULN
Adjusted for treatment status odds ratio of HBeAg seroconversion, HBV DNA loss and ALT normalization in HBeAg (+) patients. In HBeAg (-) patients: loss of HBV DNA+ALT normalization	Peginterferon alfa-2a+lamivudine vs. lamivudine	9/24	1/140	Cindoruk, 200797	10.32 (9.71; 10.97)	Sustained response was greater per increase in 1unit (U/L) in baseline ALT
HBeAg seroconversion	Peginterferon alfa-2a+lamivudine vs. lamivudine	48/24	1/542	Lau, 200596	1.93 (1.01; 3.69)	Response was greater among patients with baseline ALT >5ULN, random differences among those with baseline ALT <2 or 2–5ULN
HBeAg seroconversion	Peginterferon alfa-2a+placebo vs. lamivudine	48/24	1/542	Lau, 200596	1.81 (1.07; 3.04)	HBeAg seroconversion was greater in patients with baseline ALT >2 but <5ULN. Random differences among those with baseline ALT <2 or >5ULN
Adjusted odds ratios of ALT normalization and an HBV DNA loss	Peginterferon alfa-2a vs. lamivudine	48/24	1/1036	Bonino, 200793	Random association per 1 U/L increase in baseline ALT
Adjusted for treatment allocation, HBV genotype and log HBV DNA odds ratio of HBeAg loss	Peginterferon alfa-2b+lamivudine vs. lamivudine	60/24	1/100	Chan, 2006108	Random association per 1 U/L increase in baseline ALT
Adjusted odds ratio of sustained HBeAg loss	Peginterferon alfa-2b+lamivudine vs. peginterferon alfa-2b	52/26	1/310	Janssen, 200599	Random differences among patients with elevated vs. normal baseline ALT
Adjusted odds ratio of HBeAg and HBV DNA loss, and normal ALT level	Peginterferon alfa-2b+vs. interferon alfa 2b	24/24	1/230	Zhao, 2007109	1.23 (0.51; 2.92)	RR, random differences between patients with baseline ALT level >3.4 vs. <3.4 ULN
Baseline histology
HBeAg loss	Interferon alfa 2b vs. placebo	24/28	1/264	Perrillo, 200264	5.76 (1.48; 22.42)	Interferon alfa 2b vs. placebo increased HBeAg loss among patients with pretreatment HAI score 5–9 but failed among patients with pretreatment HAI score 0–4 or >10
HBeAg loss	Interferon alfa 2b+lamivudine vs. placebo	24/28	1/331	Perrillo, 200264	5.32 (1.51; 18.72)	Interferon alfa 2b+lamivudine vs. placebo increased HBeAg loss in patients with pretreatment HAI Score 5–9 but failed among patients with pretreatment HAI score 0–4 or >10
Odds ratio of HBeAg and HBV DNA loss independent on gender and age	Interferon alfa 2b+ lamivudine vs. lamivudine	24/48	1/151	Barbaro, 200163	2.91 (1.04; 8.22)	The rate of sustained response after interferon alfa 2b+ lamivudine vs. lamivudine was increased by an increase in baseline inflammation scores
2.58 (0.88; 7.60)	The rate of sustained response after interferon alfa 2b+ lamivudine vs. lamivudine was not increased by an increase in baseline fibrosis scores
Adjusted for treatment status odds ratio of HBeAg seroconversion, HBV DNA loss and ALT normalization in HBeAg (+) patients. In HBeAg (-) patients: loss of HBV DNA, ALT normalization	Peginterferon alfa-2a+lamivudine vs. lamivudine	9/15	1/160	Cindoruk, 200797	Presence of steatosis did not modify the effect of peginterferon alfa-2a + lamivudine vs. lamivudine on sustained response
14.97 (2.43; 92.28)	The adjusted rates of sustained response were increased per increase in baseline Knodell HAI
Adjusted relative risk of HBeAg seroconversion and HBV DNA loss	Peginterferon alfa-2b+lamivudine vs. peginterferon alfa-2b	52/78	1/310	Buster, 2007100	0.98 (0.17; 5.23)	Presence of advanced fibrosis- fibrosis score of 4–6 (HAI) did not change adjusted relative risk of HBV DAN loss and HBeAg seroconversion
Baseline viral load
HBV DNA loss with persistent HBeAg	Interferon alfa 2b vs. no treatment	32/20–52	1/118	Janssen, 199961	1.87 (0.59; 5.87)	No association between baseline positive HBV DNA (per 1 unit increase) and the effects of interferon alfa 2b vs. no treatment
HBV DNA and HBeAg loss	Interferon alfa 2b vs. no treatment	24/24	1/169	Perrillo, 199084	5.24 (1.22; 22.50)	interferon alfa 2b, 5MU/day vs. no treatment increased rates of HBV DNA and HBeAg loss among patients with baseline HBV DNA 2-99pg/ml. Random differences after interferon 1MU/day and after interferon 1 or 5 MU/day among the patients with baseline HBV DNA 100–200 pg/ml or >200 pg/ml
HBeAg loss	Interferon alfa 2b vs. no treatment	32/20	1/118	Janssen, 199961	3.18 (1.25; 8.05)	RR, Interferon Alfa 2b, 10 MU three times per week vs. no treatments increased rates of HBeAg loss among the patients with baseline HBV DNA <10pg/ml. Random differences among the patients with baseline HBV DNA >10pg/ml
HBV DNA and HBeAg loss	Interferon alfa 2b+corticosteroid vs. no treatment	24/24	1/169	Perrillo, 199084	5.38 (1.26; 22.84)	interferon alfa 2b+corticosteroid vs. no treatment increased rates of HBV and HBeAg loss among patients with baseline HBV DNA 2-99pg/ml
					8.80 (0.49; 158.66)	interferon alfa 2b+corticosteroid vs. no treatment did not increase the rates of HBV DNA and HBeAg loss among patients with baseline HBV DNA 100–200 pg/ml
0.98 (0.06; 15.13)	interferon alfa 2b+corticosteroid vs. no treatment did not increase the rates of HBV DNA and HBeAg loss among patients with baseline HBV DNA >200 pg/ml
HBeAg loss	Interferon alfa 2b+corticosteroid vs. interferon alfa 2b	24/24	1/183	Wai, 200265	1.10 (1.03; 1.17)	Interferon alfa 2b+corticosteroid vs. interferon alfa 2b increased the rates of HBeAg loss in patients with low baseline HBV-DNA level
1.10 (1.01; 1.21)	Interferon alfa 2b+corticosteroid vs. interferon alfa 2b increased the rates of HBeAg loss in patients with low baseline HBV-DNA and elevated baseline ALT
Odds ratio of HBeAg and HBV DNA loss	Interferon alfa 2b+lamivudine vs. lamivudine	24/48	1/151	Barbaro, 200163	7.23 (2.71; 19.57)	Odds of sustained suppression of serum levels of HBeAg and HBV DNA was significant in those with baseline viral load of 200 pg/ml or less independent of gender and age
HBeAg loss	Interferon alfa 2b+lamivudine vs. lamivudine	52/24	1/75	Sarin, 200575	3.89 (1.20; 12.69)	Interferon alfa 2b+lamivudine vs. lamivudine resulted in increase rates of HBeAg loss in patients with baseline HBV DNA >107 copies/mL
HBeAg loss and seroconversion	Interferon alfa 2b+lamivudine vs. lamivudine	52/24	1/75	Sarin, 200575	4.87 (1.14; 20.74)	Interferon alfa 2b+lamivudine vs. lamivudine resulted in increase rates of HBeAg seroconversion and HBV DNA loss in patients with baseline HBV DNA >107 copies/mL
Viral breakthrough -reappearance of serum HBV-DNA	Interferon alfa 2b+lamivudine vs. lamivudine	176/192	1/83	Jang, 200472	Random association with baseline HBV DNA levels (1 unit increase)
Adjusted for treatment status odds ratio of HBeAg seroconversion, HBV DNA loss and ALT normalization in HBeAg (+) patients. In HBeAg (-) patients: loss of HBV DNA, ALT normalization	Peginterferon alfa-2a+lamivudine vs. lamivudine	9/6	1/140	Cindoruk, 200797	1.05 (0.13; 8.14)	Baseline mean viral load (copy/mL)was not associated with sustained response to the therapy
Sustained combined response: ALT normalization and an HBV DNA loss	Peginterferon alfa-2a+lamivudine vs. lamivudine	48/24	1/76	Bonino, 200793	2.24 (1.31; 3.83)	Peginterferon alfa-2a+lamivudine vs. lamivudine increased sustained response among patients with baseline HBV DNA <6.12 log10 copies/ml
					1.78 (1.11; 2.84)	Peginterferon alfa-2a+lamivudine vs. lamivudine increased sustained response among patients with baseline HBV DNA >6.12–8.42 log 10 copies/ml
1.37 (0.67; 2.80)	Peginterferon alfa-2a+lamivudine vs. lamivudine did not increase sustained response among patients with baseline HBV DNA >8.42 log 10 copies/ml
HBeAg seroconversion	Peginterferon alfa-2a+lamivudine vs. lamivudine	48/24	1/543	Lau, 200596	0.84 (0.47; 1.48)	Peginterferon alfa-2a+lamivudine vs. lamivudine did not increase HBeAg seroconversion among patients with baseline HBV DNA levels ≤9.07 (log copies/ml)
					1.91 (1.16; 3.15)	Peginterferon alfa-2a+lamivudine vs. lamivudine increased HBeAg seroconversion among patients baseline HBV DNA levels >9.07–10.26 (log copies/ml)
2.01 (0.82; 4.90)	Peginterferon alfa-2a+lamivudine vs. lamivudine did not increase HBeAg seroconversion among patients with baseline HBV DNA levels >10.26 (log copies/ml)
HBeAg seroconversion	Peginterferon alfa-2a+lamivudine vs. peginterferon alfa-2a	48/24	1/542	Lau, 200596	0.54 (0.32; 0.91)	The rates of HBeAg seroconversion were lower after peginterferon alfa-2a+lamivudine vs. peginterferon alfa-2a among the patients with baseline HBV DNA levels ≤9.07 (log copies/ml)
					1.03 (0.68; 1.54)	Random differences among patients with baseline HBV DNA levels >9.07–10.26 (log copies/ml)
1.27 (0.59; 2.75)	Random differences among patients with baseline HBV DNA levels>10.26 (log copies/ml)
Sustained combined response: ALT normalization and an HBV DNA level of <20,000 copies/ml	Peginterferon alfa-2a+placebo vs. lamivudine	48/24	1/96	Bonino, 200793	1.27 (0.71; 2.30)	Random differences among patients with baseline HBV DNA <6.12 log10 copies/ml
					3.87 (2.55; 5.88)	Peginterferon alfa-2a+ placebo vs. lamivudine increased the rates of sustained response among patients with baseline HBV DNA >6.12–8.42 log 10 copies/ml
1.80 (0.91; 3.57)	Random differences among patients with baseline HBV DNA >8.42 log 10 copies/ml
HBeAg seroconversion	Peginterferon alfa-2a+placebo vs. lamivudine	48/24	1/543	Lau, 200596	1.55 (0.95; 2.51)	Random differences among patients with baseline HBV DNA levels ≤9.07 (log copies/ml)
					1.86 (1.13; 3.08)	Peginterferon alfa-2a+placebo vs. lamivudine increased rates of HBeAg seroconversion among patients with baseline HBV DNA levels >9.07–10.26 (log copies/ml)
1.58 (0.62; 4.01)	Random differences among patients with baseline HBV DNA levels >10.26 (log copies/ml)
Adjusted odds ratios of sustained combined response: ALT normalization and an HBV DNA level of <20,000 copies/ml	Peginterferon alfa-2a vs. lamivudine	48/24	1/1036	Bonino, 200793	1.06 (0.93; 1.21)	Baseline HBV DNA (Log10) was not associated with sustained response to therapy
Adjusted for treatment allocation, hepatitis B virus (HBV) genotype, baseline ALT odds ratio of persistent HBeAg loss	Peginterferon alfa-2b+lamivudine vs. lamivudine	60/0	1/100	Chan, 2006108	0.70 (0.38; 1.30)	Baseline HBV DNA (log10) was not associated with sustained response to therapy
Adjusted for treatment allocation, HBV DNA genotype, IL-1b-511 polymorphism, baseline ALT odds ratio of persistent HBeAg loss and had less than 2 occasions with HBV DNA <100,000 copies/mL	Peginterferon alfa-2b+lamivudine vs. lamivudine	60/0	1/100	Chan, 2006108	0.65 (0.35; 1.20)	Baseline HBV DNA (log10) was not associated with sustained response to therapy
Adjusted odds ratio of sustained HBeAg loss	Peginterferon alfa-2b+lamivudine vs. peginterferon alfa-2b	52/0	1/307	Janssen, 200599	1.60 (1.30; 1.80)	Peginterferon alfa-2b+lamivudine vs. peginterferon alfa-2b increased the rates of HBeAg loss among patients with low baseline viral load
Multivariate adjusted odds ratio of sustained combined response: HBeAg-negative, HBV DNA <5 log10 copies/mL, and normal ALT level	Peginterferon alfa-2b vs. interferon alfa 2b	24/0	1/230	Zhao, 2007109	0.53 (0.22; 1.28)	Random difference among patients with baseline HBV DNA >8.1 vs. <8.1 log 10 copies/mL
Genotype, outcomes at followup off treatment
Adjusted for age, gender, baseline ALT, HBV DNA, and histology, precore G1896A mutation, core promoter A1762T, G1764A, and treatment with interferon with and without prednisone pretreatment odds ratios of sustained HBV DNA loss	Interferon alfa 2b+corticosteroid vs. interferon alfa 2b	24/0	1/115	Wai, 200265	1.28 (1.06; 1.42)	Patients with HBV genotype B vs. C had better sustained response to the therapy
1/68	Wai, 200265	1.47 (1.18; 1.82)	Patients with HBV genotype B vs. C and elevated baseline ALT had better sustained response to the therapy
Sustained combined response: ALT normalization and an HBV DNA level of <20,000 copies/ml	Peginterferon alfa-2a+lamivudine vs. lamivudine	48/24	1/126	Bonino, 200793	2.09 (1.29; 3.40) - C	Peginterferon alfa-2a+lamivudine vs. lamivudine increased the rates of sustained response among patients with genotype C or genotype D
3.33 (1.53; 7.27) - D
HBeAg seroconversion	Peginterferon alfa-2a+lamivudine vs. lamivudine	48/24	1/543	Lau, 200596	1.34 (0.30; 5.92)	Random difference among patients with HBV genotype A
					1.42 (0.78; 2.58)	Random difference among patients with HBV genotype B
					1.49 (0.96; 2.31)	Random difference among patients with HBV genotype C
0.67 (0.11; 3.97)	Random difference among patients with HBV genotype D
HBeAg seroconversion	Peginterferon alfa-2a+lamivudine vs. peginterferon alfa-2a	48/24	1/542	Lau, 200596	0.33 (0.11; 1.02)	Random difference among patients with HBV genotype A
					1.04 (0.60; 1.80)	Random difference among patients with HBV genotype B
					0.86 (0.59; 1.25)	Random difference among patients with HBV genotype C
1.00 (0.14; 7.05)	Random difference among patients with HBV genotype D
Sustained combined response: ALT normalization and an HBV DNA level of <20,000 copies/ml	Peginterferon alfa-2a+placebo vs. lamivudine	48/24	1/19	Bonino, 200793	2.18 (0.27; 17.32)	Random differences among patients with genotype A
					1.14 (0.70; 1.85)	Random differences among patients with genotype B
					2.22 (1.36; 3.63)	Peginterferon alfa-2a+placebo vs. lamivudine increased the rates of sustained response among patients with genotype C
1.47 (0.59; 3.69)	Random differences among patients with genotype D
HBeAg seroconversion	Peginterferon alfa-2a+placebo vs. lamivudine	48/24	1/543	Lau, 200596	4.01 (1.15; 14.07)	Peginterferon alfa-2a+placebo vs. lamivudine increased the rates of e Ag seroconversion among patients with HBV genotype A
					1.36 (0.74; 2.48)	Random differences among patients with HBV genotype B
					1.73 (1.13; 2.65)	Peginterferon alfa-2a+placebo vs. lamivudine increased the rates of e Ag seroconversion among patients with HBV genotype C
0.67 (0.11; 3.97)	Random differences among patients with HBV genotype D
Adjusted odds ratios of sustained combined response: ALT normalization and an HBV DNA level of <20,000 copies/ml	Peginterferon alfa-2a vs. lamivudine	48/24	1/1036	Bonino, 200793	2.58 (0.73; 9.20)	Random difference between genotypes (A vs. D)
					3.69 (1.54; 8.79)	Rates of sustained response were higher among patients with genotype B vs. D
5.46 (2.46; 12.10)	Rates of sustained response were higher among patients with genotype C vs. D
Adjusted for treatment allocation, HBV genotype, baseline ALT, and log HBV DNA odds ratio of persistent HBeAg loss at any time up to week 76 of post-treatment	Peginterferon alfa-2b+lamivudine vs. lamivudine	60/24	1/100	Chan, 2006108	10.37 (1.11; 96.96)	Rates of response were higher among patients with interleukin (IL)-1b-511 baseline genotype C/T vs. C/C
					Random differences in patients with genotype C vs. B
					Random differences in patients with Haplotype -511/-31 of interleukin (IL)-1b C-T vs. T-C
					Random differences in patients with interleukin (IL)-1b-511 baseline genotype T/T vs. C/C
					Random differences in patients with interleukin (IL)-1b-31 baseline genotype C/T vs. T/T or C/C vs. T/T
					Random differences in patients with IL-1 receptor antagonist genotype IL-1RN 1/2 vs. 1/1
					Random differences in patients with interleukin (IL)-1b-511 baseline genotype C/T and T/T vs. C/C
Random differences in patients with interleukin (IL)-1b-31 baseline genotype C/T and C/C vs. T/T
Adjusted relative risk of HBeAg seroconversion and HBV DNA <10,000 copies/ml.	Peginterferon alfa-2b+lamivudine vs. peginterferon alfa-2b	52/26	1/307	Buster, 2007100	11.30 (1.38; 92.57)	Adjusted rates of sustained response were higher among patients with genotype A vs. C
					4.28 (1.39; 13.21)	Adjusted rates of sustained response were higher among patients with genotype A vs. D
					12.13 (1.24; 118.30)	Adjusted rates of sustained response were higher among patients with genotype B vs. C
4.59 (1.14; 18.43)	Adjusted rates of sustained response were higher among patients with genotype B vs. D
Adjusted odds ratio of sustained HBeAg loss	Peginterferon alfa-2b+lamivudine vs. peginterferon alfa-2b	52/26	1/307	Janssen, 200599	2.40 (1.30; 4.60)	Adjusted rates of sustained response were higher among patients with HBV genotype A vs. D
					3.60 (1.40; 8.90)	Adjusted rates of sustained response were higher among patients with HBV genotype A vs. C
2.20 (0.70; 7.00)	Random difference among patients with HBV genotype B vs. C
Multivariate adjusted odds ratio of sustained HBeAg loss, HBV DNA <5 log 10 copies/mL, and normal ALT level	Peginterferon alfa-2b vs. interferon alfa 2b	24/24	0/250	Zhao, 2007109	0.19 (0.08; 0.46)	RR, response was lower among patients with genotype C vs. B
Previous treatment, outcomes at followup off treatment
HBeAg seroconversion	Peginterferon alfa-2a+lamivudine vs. lamivudine	48/24	1/543	Lau, 200596	1.52 (1.08; 2.12)	Rates of HBeAg seroconversion were higher among patients with no previous exposure to lamivudine
Random differences among those with previous LAM or Interferon therapy
HBeAg seroconversion	Peginterferon alfa-2a+lamivudine vs. peginterferon alfa-2a	48/24	1/542	Lau, 200596	Random differences among all patients with and without previous treatment
HBeAg seroconversion	Peginterferon alfa-2a+placebo vs. lamivudine	48/24	1/543	Lau, 200596	1.58 (1.11; 2.23)	Peginterferon alfa-2a+placebo vs. lamivudine increased the rates of HBeAg seroconversion among patients with no previous anti-HBV therapy
					1.43 (0.55; 3.71)	Random differences among patients with previous treatment: LAM
					1.72 (1.24; 2.38)	Peginterferon alfa-2a+placebo vs. lamivudine increased the rates of HBeAg seroconversion among patients with no previous exposure to lamivudine
					3.26 (1.08; 9.88)	Peginterferon alfa-2a+placebo vs. lamivudine increased the rates of HBeAg seroconversion among patients with previous treatment: IFN
1.55 (1.12; 2.14)	Peginterferon alfa-2a+placebo vs. lamivudine increased the rates of HBeAg seroconversion among patients with no previous exposure to conventional interferon
HBeAg loss	Peginterferon alfa-2b+lamivudine vs. peginterferon alfa-2b	52/26	2/307	Janssen, 200599	2.20 (1.10; 4.50)	Peginterferon alfa-2b+lamivudine vs. peginterferon alfa-2b increased the rates of HBeAg loss among patients without previous interferon therapy
			Flink, 2006102	0.94 (0.63; 1.40)	Random differences among naïve to any treatments patients
1/307	Flink, 2006102	Random differences among patients with previous IFN, LAM, and combined therapy
HBV DNA loss, normalization of ALT					Random differences among patients naïve to any antiviral treatment

Baseline liver function (ten studies, three medications, peginterferon, interferon, and lamivudine were used as mono or combination therapy) (Appendix E Tables 9 and 10). The effects of lamivudine on disease progression defined as hepatocellular carcinoma, spontaneous bacterial peritonitis, bleeding gastroesophageal varices, or death related to liver disease or liver decompensation at the end of the treatment did not differ among patients with different baseline ALT levels (Appendix E Figure 20).^{93, 107, 132, 141} Comparative effectiveness of peginterferon alfa-2a versus lamivudine on combined response defined as ALT normalization and an HBV DNA level of <20 000 copies/ml at the end of the treatment was greater per increase by 1 log 10 unit (IU/1) in baseline ALT levels (Table 8).⁹³ Lamivudine increased HBeAg clearance and seroconversion among patients with elevated baseline ALT with no effects in those with ALT <2 ULN (Appendix E. Figure 21).⁶⁴

At 24 weeks of followup off the administration of interferon alfa-2b with corticosteroid pretreatment, compared to interferon alfa-2b alone, HBV DNA loss was more frequent among patients with elevated baseline ALT (RR 1.22, 95 percent CI 1.05; 1.42)⁶⁵ (Appendix E Tables 9 and 10). Adjusted odds of sustained virologic response to interferon alfa-2b combined with lamivudine compared to lamivudine were higher in patients with baseline ALT >150UL (RR 3.12, 95 percent CI 1.43; 6.82).⁶³ Sustained HBeAg loss or seroconversion after interferon alfa-2b combined with lamivudine compared to lamivudine alone was higher among patients with elevated ALT with nonsignificant differences among those with baseline ALT <1 ULN or 1–2 ULN.⁶⁴ Sustained response to peginterferon alfa-2a combined with lamivudine compared to lamivudine alone was greater per increase in one unit (U/L) in baseline ALT (RR 10.32, 95 percent CI 9.71; 10.97).⁹⁷ Sustained HBeAg seroconversion after peginterferon alfa-2a combined with lamivudine compared to lamivudine alone was greater among patients with baseline ALT >5 ULN (RR 1.93, 95 percent CI 1.01; 3.69) with random differences among those with baseline ALT <2 or 2–5 ULN.⁹⁶ However, several studies reported no association between dose response increase in baseline ALT and sustained response to peginterferon alfa-2a compared to lamivudine⁹⁶ or peginterferon alfa-2b combined with lamivudine versus lamivudine.¹⁰⁸

In conclusion, the low level of evidence indicated that treatment induced effects on disease progression and liver decompensation do not vary by baseline ALT levels. The low level of evidence indicated that treatment induced HBeAg clearance and seroconversion, HBeAg loss, or virologic loss vary by baseline liver function with inconsistent across the studies evidence of better response among patients with elevated baseline ALT.

Genotype of HBV DNA (eight studies, two antiviral agents, peginterferon and lamivudine used as mono or combination therapy) (Appendix E Tables 9 and 10). The effects of antiviral drugs on intermediary outcomes among patients with different HBV DNA genotype have been evaluated at the end of treatments⁹³ and at followup off the therapies (seven studies).^{65, 93, 96, 99, 100, 108, 109} (Appendix E Tables 9 and 10). Adjusted odds ratios of ALT normalization and reduction of HBV DNA level <20,000 copies/ml after 48 weeks of peginterferon alfa-2a administration compared to lamivudine were significantly greater among patients with genotype B versus D (OR 2.31, 95 percent CI 1.30; 4.20) and genotype C versus D (OR 2.90, 95 percent CI 1.70; 5.00). Patients with genotype A had lower adjusted odds of response defined as ALT normalization and an HBV DNA level of <20,000 copies/ml compared to patients with genotype C (OR 0.33, 95 percent CI 0.10; 0.90).⁹³ No differences in treatment effects were found between patients with genotype A versus B, A versus D, or B versus C.⁹³

Off treatment response to the same treatments also differed with greater adjusted odds of success among patients with genotype B versus D (OR 3.69, 95 percent CI 1.54; 8.79) and with genotype C versus D (OR 5.46, 95 percent CI 2.46; 12.10).⁹³ Patients with genotype B HBV DNA experienced sustained clearance of serum HBV DNA after interferon alfa-2b combined with steroid pretreatment compared to interferon alfa-2b alone more often than those with genotype C (adjusted OR 1.28, 95 percent CI 1.06; 1.42).⁶⁵ One RCT reported no significant differences in sustained HBeAg seroconversion after peginterferon alfa-2a combined with lamivudine compared to peginterferon alfa-2a or lamivudine alone among all genotypes of HBV DNA.⁹⁶

In conclusion, the low level of evidence indicated that treatment induced ALT normalization and HBV DNA clearance or HBeAg seroconversion vary by HBV DNA genotype with better response among patients with genotype B and C. There was no evidence for clinical outcomes.

Previous treatment status (seven studies, four antiviral agents, peginterferon, interferon, adefovir, and lamivudine used as mono or combination therapy) (Appendix E Tables 9 and 10). Off-treatment rates of HBeAg seroconversion after administration of peginterferon alfa-2a plus lamivudine compared to lamivudine were higher among patients with no previous exposure to lamivudine (RR 1.52, 95 percent CI 1.08; 2.12)⁹⁶ with no significant differences among previously treated patients. Off-treatment HBeAg seroconversion was better after monotherapy with peginterferon alfa-2a compared to lamivudine in naïve to lamivudine patients (RR 1.72, 95 percent CI 1.24; 2.38).⁹⁶ Random differences were observed in patients previously treated with lamivudine.⁹⁶

Five RCTs enrolled lamivudine resistant patients,^{118, 119, 124, 125, 141} and one enrolled interferon resistant patients.⁹² Adefovir combined with lamivudine for 48 weeks did not result in better outcomes, including rates of HBV DNA reduction to less than <2,000 copies/ml and ALT normalization compared to adefovir alone in lamivudine resistant patients.¹¹⁸ Adefovir combined with lamivudine compared to lamivudine alone resulted in greater rates of ALT normalization (ARD 0.45, 95 percent CI 0.21; 0.69) and HBV DNA clearance (ARD 0.35, 95 percent CI 0.13; 0.57) in patients with lamivudine-resistant CHB with random differences in HBeAg clearance or seroconversion;¹¹⁹ however, the same trial reported that combined treatment did not improve outcomes compared to adefovir alone.¹¹⁹ Adefovir monotherapy improved ALT normalization compared to lamivudine alone (ARD after lamivudine versus adefovir -0.42, 9 percent CI -0.67; -0.18) with random differences in HBeAg clearance and seroconversion.¹¹⁹ The BEHoLD Study Group enrolled 182 lamivudine-refractory patients to start entecavir administration or continued on lamivudine for 48 weeks.¹²⁴ Mortality after 48 weeks of therapy and at 28 weeks of followup did not differ among treatment groups.¹²⁴ The largest dose of entecavir (1mg/day) increased HBV DNA loss in combination with normalization of ALT level at 48 weeks of the treatment (ARD 0.67, 95 percent CI 0.41; 0.92).¹²⁴ Viral clearance (HBV DNA <400 copies/mL) or normalization of ALT levels was greater after all doses of entecavir compared to lamivudine.¹²⁴ Entecavir increased the rates of HBV DNA and HBeAg clearance and normalization of ALT level in lamivudine-refractory, HBeAg-positive CHB patients compared to lamivudine, the effects were significant after 1mg/day (ARD 0.67, 95 percent CI 0.41; 0.92) or 0.5mg/day (ARD 0.40, 95 percent CI 0.14; 0.66).¹²⁵ The BEHoLD Study Group enrolled HBeAg-positive patients with persistent viremia or documented YMDD mutations after previous lamivudine therapy that were randomized to switch to entecavir 1 mg daily or continue lamivudine for a minimum of 52 weeks.¹²⁵ Entecavir increased rates of HBV DNA clearance and ALT normalization (ARD 0.50, 95 percent CI 0.42; 0.59), HBV DNA loss (<300 copies/mL by PCR) (ARD 0.18, 95 percent CI 0.11; 0.25), and HBeAg clearance (ARD 0.06, 95 percent CI 0.01; 0.12).¹²⁵ Improvement in necroinflammatory Knodell score (ARD 0.26, 95 percent CI 0.16; 0.37) and Ishak fibrosis scores (ARD 0.17, 95 percent CI 0.07; 0.26) was greater after entecavir therapy.¹²⁵ Discontinuation of lamivudine in 74 patients with lamivudine-resistant mutants resulted in the same rates of hepatic decompensation, ALT normalization, or HBeAg seroconversion compared to continuous administration of lamivudine.¹⁴¹ A combination of interferon alfa-2b with lamivudine in 20 patients who failed previous interferon therapy did not improve HBV DNA or HBeAg clearance and ALT normalization.⁹²

In conclusion, the low to moderate level of evidence indicated that treatment induced HBeAg seroconversion, ALT normalization, HBV DNA clearance, and improved histology are greater in patients naïve to antiviral medications. Entecavir improved virological and biochemical outcomes in lamivudine resistant patients without differences in mortality.

YMDD mutation (two studies, one antiviral agent used as monotherapy) (Appendix E Tables 9 and 10). One study examined the effect of 48 week lamivudine treatment versus placebo on liver histology in Chinese patients with and without YMDD mutations.¹³⁰ The authors reported improvement in necroinflammatory but not in fibrosis scores in patients with and without mutations; however, lamivudine improved the outcome of “worsened histology” only among patients without YMDD mutation (ARD 0.15, 95 percent CI 0.04; 0.54) (there was no protective effect in patients with YMDD mutations).¹³⁰ A study of discontinuing lamivudine in lamivudine resistant patients reported no association between YMDD mutations and hepatic decompensation, defined as change in the Child-Turcotte-Pugh score of two or more points after adjustments for ALT, baseline viral load, sex, YMDD variant, platelet, bilirubin, and albumin.¹⁴¹

In conclusion, low level of evidence indicated that treatment induced changes in liver histology are associated with baseline YMDD mutation. There was no evidence that this mutation was associated with differences in clinical outcomes.

Outcomes across studies with different patient populations. We analyzed the differences in off treatment outcomes across the studies that included different proportions of patients with HBeAg-positive status, baseline cirrhosis, and previous antiviral treatments. Clinical outcomes were not reported. Interferon alfa-2b combined with lamivudine compared to interferon alfa-2b alone increased HBV DNA clearance in studies that enrolled treatment naïve patients with random differences in the studies of previously treated patients (Appendix E. Figure 22).^{62, 68, 69, 72, 74, 75, 84, 87} However, the effects of interferon alfa-2b compared to no treatment or after combined interferon alfa-2b+lamivudine compared to lamivudine alone on HBV DNA loss did not show a clear pattern according to patient populations across the studies. Sustained HBeAg clearance after compared treatments was the same across the studies (Appendix E. Figure 23).⁶²–^{64, 67, 80, 83, 87, 88, 91, 98, 99, 136} Sustained HBeAg seroconversion was greater after interferon alfa-2b combined with lamivudine versus lamivudine monotherapy in one study;⁶³ however, another study⁶⁷ that assessed interferon nonresponders and a European trial of previously untreated patients reported opposite association with better effects from lamivudine monotherapy therapy. ⁶⁷ Sustained ALT normalization after compared treatments was the same across the studies (Appendix E. Figure 24).^{10, 62, 63, 72, 74}–^{76, 84, 87, 95, 96, 113}

Summary. There is no high quality evidence that clinical outcomes of all-cause or disease specific mortality, hepatocellular carcinoma or hepatic decompensation are improved with currently approved and investigated therapeutic strategies (i.e., drug, dose, duration, patient population). Changes in biochemical, virologic, and histologic measures at the end of treatment or off treatment are frequently used to assess therapeutic effectiveness. However, these measures have not been demonstrated to be accurate surrogates for determining long-term clinical outcomes due to treatments. All treatments are associated with harms and immediate pharmaceutical costs, though most drugs are well tolerated and adverse effects are relatively mild. Therefore, it is difficult to determine what patients would derive clinical benefit. There is low evidence that treatment improves HBsAg clearance and measures of hepatitis resolution. This is beneficial because loss of HBsAg likely eliminates viral transmission to noninfected individuals and defines resolved hepatitis B. Individuals who are inactive carriers also have persistence of HBsAg positivity and a very favorable long-term prognosis. Therefore, for these individuals treatment is unlikely to provide a small benefit beyond risk of transmission.