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

Chou R, Carson S, Chan BKS, et al. Drug Class Review: Pegylated Interferons for Chronic Hepatitis C Infection: Final Report [Internet]. Portland (OR): Oregon Health & Science University; 2007 May.

Cover of Drug Class Review: Pegylated Interferons for Chronic Hepatitis C Infection

Drug Class Review: Pegylated Interferons for Chronic Hepatitis C Infection: Final Report [Internet].

Show details

Methods

Literature Search

To identify articles relevant to each key question, we searched Medline (1966 to July Week 4 2006), the Cochrane Central Register of Controlled Trials (3rd Quarter 2006), the Cochrane Database of Systematic Reviews (2nd Quarter 2006), and the Database of Abstracts of Reviews of Effects (3rd Quarter 2006) (See Appendix A for search strategies). We also searched reference lists of included review articles for additional citations. Pharmaceutical manufacturers were invited to submit dossiers, including citations. All citations were imported into an electronic database (Endnote 9.0).

Study Selection

All citations were reviewed for inclusion using the criteria shown in Table 3. Full-text articles of potentially relevant citations were retrieved and inclusion criteria were re-applied. Title and abstract review was conducted by two independent reviewers (Carson and Care); review of full-text articles was conducted by one reviewer (Carson) and checked by a second (Chou). Disagreements were resolved by consensus.

Table 3. Study inclusion criteria.

Table 3

Study inclusion criteria.

We defined a sustained virologic response (SVR) as the absence of detectable HCV RNA in the serum six months after the end of a course of therapy.21 SVR is the best short-term predictor of long-term virologic remission rates and is associated with improvements in fibrosis and inflammation.22 End-of-treatment response (ETR) was defined as no detectable virus at the end of a course of therapy. We did not consider ETR a primary outcome since it is not as reliable as SVR for predicting long-term remission. Some trials also measure early virologic response (EVR), which is usually defined as absence of detectable HCV RNA in serum or >2.0 log copy/ml reduction in serum HCV after 12 weeks of therapy. Although assessing EVR is helpful for determining whether to complete a full course of therapy (patients without an EVR are unlikely to achieve an SVR), it is less accurate than ETR for predicting long-term remission. We included head-to-head trials reporting EVR because no head-to-head trials reporting longer-term outcomes are currently available.

We defined a sustained biochemical response (SBR) as normalization of liver transaminases six months after the end of a course of therapy. Some trials also report end-of-treatment biochemical response. Definitions for histological response are less standardized than definitions for reporting virologic outcomes, however traditionally a histologic response has been defined as a 2-point or greater decrease in the inflammatory score or fibrosis score, or a 1-point decrease in the fibrosis score.21

Because dual therapy with pegylated interferon has only been available since 2001 and assessment of effects on rates of cirrhosis, hepatocellular cancer, need for liver transplant, and mortality would require studies with extended (a decade or more) follow-up, we believed studies evaluating these outcomes would probably not be available. However, we did search for studies reporting these important clinical outcomes.

We included non-randomized studies as well as randomized trials reporting adverse events (withdrawal due to adverse events, serious adverse events, overall adverse events, hematologic adverse events, flu-like symptoms, and depression) associated with dual therapy with pegylated interferon.

Data Abstraction

The following data were abstracted from included trials: study design, setting, population characteristics (including sex, age, ethnicity, and HCV genotype), eligibility and exclusion criteria, and interventions (dose and duration); numbers screened, eligible, enrolled, and lost to follow-up; method of outcome ascertainment; and results for each outcome (including SVR, ETR, SBR, histological response rates, quality of life, other clinical outcomes, and adverse events). We recorded intention-to-treat results when available. Results were entered into a relational database (Microsoft Access 2003).

Quality Assessment

We assessed internal validity (quality) of controlled clinical trials using predefined criteria adapted from those developed by the US Preventive Services Task Force23 and the National Health Service Centre for Reviews and Dissemination (Appendix B).24 For each included study, we assessed methods used for randomization; allocation concealment; blinding of participants, investigators, and assessors of outcomes; similarity of comparison groups at baseline; adequate reporting of attrition, crossover, adherence, and contamination; post-allocation exclusions; and use of intention-to-treat analysis.

These criteria were then used to categorize trials into "good", "fair", and "poor" quality. Studies that had a serious flaw or combination of flaws in design or implementation that seriously compromised the validity of results were categorized as "poor" quality. For example, an open-label study that used improper randomization techniques and failed to use intention-to-treat analysis would be rated poor-quality. Results of poor-quality studies are at least as likely to be due to design flaws or biases as to be due to true effects. Studies which met all quality criteria were rated "good"; the rest were rated "fair". As the "fair" quality category is broad, studies with this rating vary in their strengths and weaknesses.

We did not formally rate quality of non-randomized studies reporting adverse events. Optimal methods for rating quality of such studies is uncertain.25 In addition, all of the non-randomized studies included in this report were uncontrolled series of patients exposed to dual therapy with pegylated interferon. Such studies are generally considered much less reliable than well-designed randomized controlled trials.

Data Synthesis

We constructed evidence tables showing the study characteristics, quality ratings, and results for all included studies. Trials that evaluated dual therapy with one pegylated interferon against another provided direct evidence of comparative effectiveness and safety. Where possible, these data are the primary focus. We also performed indirect comparisons when direct head-to-head evidence was sparse or unavailable. In theory, trials that compare dual therapy with pegylated interferon to dual therapy with non-pegylated interferon or another common comparator can provide indirect evidence about effectiveness and safety if treatment effects are consistent across all of the trials.26, 27 Indirect comparisons usually agree with direct comparisons, though large discrepancies have been reported in some cases.28, 29 In addition, indirect comparisons also result in less precise estimates of treatment effects compared to the same number of similarly sized head-to-head trials because methods for indirect analyses incorporate additional uncertainty from combining different sets of trials.26, 27

We performed meta-analysis to estimate pooled relative risks and 95% confidence intervals using the DerSimonian-Laird method in a random effects model.30 We chose the random effects model because trials evaluating the same interventions and outcomes differed in patient populations, dosing of drugs, and other factors. Heterogeneity was assessed by calculating the Q-statistic and the percent of the total variance due to between study variability (I2 statistic).31 Subgroup analysis was performed to assess differences in estimates of effect in HIV co-infected versus non-HIV infected populations and for different HCV genotypes. We also performed sensitivity analysis on poor-quality studies, outlier trials, specific populations (such as patients with thalassemia), and unpublished trials to evaluate stability of estimates and conclusions. Funnel plots were produced to assess the likelihood of publication bias if there were an adequate number of studies (at least seven) to plot. Relative risks and confidence intervals were calculated and funnel plots were produced using the meta package in R.32 Forest plots were generated using RevMan 4.2.8 (Review Manager 4.2 for Windows, The Nordic Cochrane Centre, Copenhagen, Denmark). We used the method described by Bucher et al to perform indirect analyses.27

Copyright © 2007, Oregon Health & Science University, Portland, Oregon.
Bookshelf ID: NBK10666
PubReader format: click here to try

Views

Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...