NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Wilt TJ, Shamliyan T, Shaukat A, et al. Management of Chronic Hepatitis B. Rockville (MD): Agency for Healthcare Research and Quality (US); 2008 Oct. (Evidence Reports/Technology Assessments, No. 174.)

Cover of Management of Chronic Hepatitis B

Management of Chronic Hepatitis B.

Show details

2Methods

Literature Search and Eligibility Criteria

We searched MEDLINE® via PubMed®, the Cochrane library,12 Medwatch,13 and United Kingdom Current Problems in Pharmacovigilance.14 We used the European Public Assessment Report15 to find original epidemiologic studies of adults with CHB published in English that reported mortality, incidence of hepatocellular carcinoma (HCC), or liver failure, prevalence and incidence of cirrhosis, HBeAg or HBsAg presence or seroconversion, viral load of hepatic virus B deoxyribonucleotide acid (HBV DNA), ALT levels, histological necroinflammatory and fibrosis scores,16 and adverse events after antiviral drugs approved by the Food and Drug Administration (FDA) for CHB, including interferon alfa-2b, pegylated interferon alfa-2a, lamivudine, adefovir, entecavir, tenofovir, and telbivudine.17 The search strategies for the four research questions are described in Appendix A *. Excluded references are shown in Appendix B. All work was conducted under the guidance of a Technical Expert Panel (TEP), whose members are identified in Appendix C.

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, 160162 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.

Intermediate outcomes

  • Complete response (resolved hepatitis B) included HBsAg loss or seroconversion in combination with undetectable HBV DNA and normal ALT.
  • Biochemical outcomes included changes in ALT levels, the rates of ALT normalization, and flare of hepatitis B as intermittent elevations of aminotransferase activity to more than ten times the upper limit of normal and more than twice the baseline value.
  • Virological outcomes included HBsAg clearance or seroconversion, HBeAg clearance in a person who was previously HBeAg-positive, HBeAg seroconversion defined as loss of HBeAg and detection of antiHBeAg in a person who was previously HBeAg-positive and antiHBeAg-negative, viral load of HBV DNA, and the rates of HBV DNA loss or reduction.
  • Histological outcomes included histological scores of inflammation or fibrosis and the rates of improvement in necroinflammatory scores without worsening in fibrosis scores.
  • Resistance was defined as worsening of histological scores or persistent HBV DNA load, or rates of genetic mutations.
  • Relapse was defined as reappearance of HBV DNA or active necroinflammatory disease of the liver in a person known to have the inactive HBsAg carrier state or resolved hepatitis B.
  • Harm effects included any adverse effects, serious adverse events, discontinuation of treatment, or decrease in dose independent of author's judgments of causality between drug therapies and events.

For question 3 we synthesized the results from subgroup analyses when the authors reported outcomes among patients according to age, gender, body mass index (BMI), baseline ALT, viral load, HBeAg status, pretreatment history, or histological activity. We synthesized the evidence of effect measure modification when authors compared the effects of baseline patient characteristics on the effects of the drug therapies. We compared the effects of the same drugs on different patient populations across the RCTs that included patients with only positive or negative HBeAg status.

Pooling criteria included the same operational definitions of outcomes and the same risk factors or clinical interventions.155 Meta-analysis was used to assess the consistency of the association between treatments and outcomes with random effects models.163 We conducted analyses separately for clinical, biochemical, virological, and histological outcomes and for relative risk and absolute risk differences. Assumptions underlying meta-analysis included valid measurements of the outcomes and similarity in study and target populations.

We tested consistency in the results comparing the direction and strength of the association. Chi squared tests were used to assess heterogeneity.164, 165 Significant heterogeneity means the effects of interventions on the outcomes were not consistent in the studies. We explored heterogeneity with meta-regression and sensitivity analysis and reported the results from random effects models. We analyzed whether duration of treatments or followup, doses of the drugs, proportion of the patients with HBeAg-positive baseline status, proportion of the patients with baseline cirrhosis, or control rates of the outcomes could explain heterogeneity between studies. Calculations were performed using STATA software at the 95 percent confidence level.166 We calculated the number needed to treat and the number of the events attributable to the treatments per 1,000 treated.167

Appendixes and evidence tables cited in this report are provided electronically at http://www.ahrq.gov/downloads/pub/evidence/ pdf/hepb/hepb.pdf

Footnotes

*

Appendixes and evidence tables cited in this report are provided electronically at http://www.ahrq.gov/downloads/pub/evidence/ pdf/hepb/hepb.pdf

PubReader format: click here to try

Views

  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...