Release Date: November 2012

By Roger Chou, MD; Daniel Hartung, PharmD, MPH; Basmah Rahman, MPH; Ngoc Wasson, MPH; Erika Barth Cottrell, PhD; and Rongwei Fu, PhD

The information in this article is intended to help clinicians, employers, policymakers, and others make informed decisions about the provision of health care services. This article is intended as a reference and not as a substitute for clinical judgment.

This article may be used, in whole or in part, as the basis for the development of clinical practice guidelines and other quality enhancement tools, or as a basis for reimbursement and coverage policies. AHRQ or U.S. Department of Health and Human Services endorsement of such derivative products may not be stated or implied.

This article was first published in Annals of Internal Medicine on November 27, 2012 (Ann Intern Med 2012; http://www.annals.org).

Background: Multiple treatments are available for chronic hepatitis C virus (HCV) infection.

Purpose: To compare benefits and harms of antiviral regimens for chronic HCV infection in treatment-naive adults.

Data Sources: English-language literature from MEDLINE (1947 to August 2012), the Cochrane Library Database, Embase, Scopus, PsychINFO, and clinical trial registries.

Study Selection: Randomized trials of antiviral treatments and cohort studies examining associations between sustained virologic response (SVR) after therapy and clinical outcomes.

Data Extraction: Several investigators abstracted study details and quality by using predefined criteria.

Data Synthesis: No trial evaluated effectiveness of treatment on long-term clinical outcomes. Dual therapy with pegylated interferon alfa-2b plus ribavirin was associated with a lower likelihood of SVR than was pegylated interferon alfa-2a plus ribavirin (absolute difference, 8 percentage points [95% CI, 3 to 14 percentage points]) on the basis of 7 poor- to fair-quality trials. For genotype 2 or 3 infection, dual therapy for 12 to 16 weeks was associated with a lower likelihood of SVR than was therapy for 24 weeks, and lower doses of pegylated interferon alfa-2b were less effective than standard doses (2 to 4 fair-quality trials). For genotype 1 infection, fair-quality trials found that triple therapy with pegylated interferon, ribavirin, and either boceprevir (2 trials) or telaprevir (4 trials) was associated with a higher likelihood of SVR than was dual therapy (absolute difference, 22 to 31 percentage points). Compared with dual therapy, boceprevir triple therapy increased risk for hematologic adverse events and telaprevir triple therapy increased risk for anemia and rash. A large well-designed cohort study and 18 smaller cohort studies found that an SVR after antiviral therapy was associated with lower risk for all-cause mortality than was no SVR.

Limitations: Trials involved highly selected populations. Observational studies did not always adequately control for confounders.

Conclusion: SVR rates for genotype 1 infection are higher with triple therapy that includes a protease inhibitor than with standard dual therapy. An SVR after antiviral therapy appears associated with improved clinical outcomes.

Primary Funding Source: Agency for Healthcare Research and Quality.

Chronic hepatitis C virus (HCV) infection is a leading cause of complications from chronic liver disease, including cirrhosis, liver failure, hepatocellular carcinoma, and death^1,2. The goal of antiviral treatment is to eradicate viremia and prevent long-term complications. Genotype 1 infection predominates in the United States (about 75% of cases) but is more difficult to treat than genotype 2 or 3 infection.

In the early 2000s, dual therapy with the combination of pegylated interferon plus ribavirin became the standard HCV treatment^3-6. Pegylation refers to the cross-linking of polyethylene glycol molecules to the interferon molecule, which delays renal clearance, permitting once-weekly dosing⁷. Two pegylated interferons are available: alfa-2a and alfa-2b. Interferon-based treatment is associated with a high rate of adverse effects, including influenza-like symptoms, fatigue, and neuropsychiatric and hematologic adverse effects⁸. In 2011, the U.S. Food and Drug Administration approved the first direct-acting antiviral agents, boceprevir⁹ and telaprevir¹⁰, for chronic genotype 1 infection.

Understanding the effectiveness of antiviral regimens is critical for making informed treatment decisions for HCV infection. This review focuses on comparative effectiveness in antiviral-naive patients and also examines how effectiveness varies depending on clinical and demographic characteristics.

We developed a review protocol and analytic framework (Appendix Figure 1) that included the following key questions:

1. What is the comparative effectiveness of antiviral treatment in improving health outcomes in patients with HCV infection, and does it vary according to patient subgroup characteristics (including, but not limited to, HCV genotype, age, race, sex, stage of disease, or genetic markers)?

2. What is the comparative effectiveness of antiviral treatments on the rate of sustained virologic response (SVR), and does it vary according to patient subgroup characteristics?

3. What are the comparative harms associated with antiviral treatments, and do they vary according to patient subgroup characteristics?

4. Have improvements in SVR been shown to reduce the risk for or rates of adverse health outcomes from HCV infection?

The protocol was developed by using a standardized process with input from experts and the public. Details, including full search strategies, inclusion criteria, and evidence tables and quality ratings, are provided in the full report, as are results of studies comparing induction versus fixed-dose regimens and study outcomes related to quality of life and histologic changes¹¹.

Data Sources and Searches

A research librarian searched Ovid MEDLINE from 1947 to August 2012, the Cochrane Library Database (through the first quarter of 2012), Embase (1976 to August 2012), Scopus (1960 to August 2012), PsychINFO (1806 to August 2012), clinical trials registries, and grants databases.

Study Selection

At least 2 reviewers independently evaluated studies for inclusion. For the first 3 questions, we included randomized trials of antiviral-naive patients that compared dual therapy with pegylated interferon alfa-2b plus ribavirin versus pegylated interferon alfa-2a plus ribavirin; triple therapy with pegylated interferon (alfa-2a or -2b), ribavirin, and either telaprevir or boceprevir versus dual therapy; or different doses or durations of dual or triple therapy. Dose and duration comparisons of dual therapy focused on genotype 2 or 3 infection. For the last question, we included cohort studies that reported adjusted risk estimates for the association between an SVR after antiviral treatment versus no SVR and clinical outcomes. Clinical outcomes were mortality, cirrhosis, hepatic decompensation, hepatocellular carcinoma, and need for transplantation. SVR, the primary intermediate outcome, was defined as the absence of detectable HCV RNA in the serum 6 months after the end of a course of therapy⁴. Harms included withdrawals due to adverse events, serious adverse events, neutropenia, anemia, psychological adverse events, influenza-like symptoms, and rash.

We restricted inclusion to English-language articles and included studies published as conference abstracts only in sensitivity analyses. We excluded studies of pregnant women¹², patients who received a transplant, HIV-infected patients, patients undergoing hemodialysis, and previously treated patients. We excluded regimens with antiviral drugs not approved in the United States for HCV infection.

Data Extraction and Quality Assessment

One investigator abstracted details about the study design, population, setting, interventions, analysis, follow-up, and results. A second investigator reviewed data for accuracy. Two investigators independently applied predefined criteria^13-15 to assess study quality as good, fair, or poor. Discrepancies were resolved through consensus.

Data Synthesis and Analysis

We assessed the overall strength of each body of evidence as “high,” “moderate,” “low,” or “insufficient” in accordance with the AHRQ “Methods Guide for Effectiveness and Comparative Effectiveness Reviews”¹⁶ on the basis of the quality of studies, consistency between studies, precision of estimates, and directness of evidence.

We performed meta-analyses of trials that evaluated similar populations, interventions, comparisons, and outcomes to estimate pooled relative risks (RRs) using the DerSimonian–Laird method in a random-effects model¹⁷. Heterogeneity was assessed with the I² statistic¹⁸. Statistical heterogeneity was explored through sensitivity and subgroup analyses based on study quality, differences in dosing or drugs, and outlier trials. We did not produce funnel plots because of small numbers¹⁰ of studies¹⁹, but we performed sensitivity analyses that included studies published only as abstracts. Analyses were performed with Stata software, version 11.0 (Stata Corp, College Station, Texas).

Role of the Funding Source

AHRQ's Effective Health Care Program funded this work. Investigators worked with AHRQ staff to develop and refine the scope, analytic framework, and key questions. AHRQ staff had no role in study selection, quality assessment, synthesis, or development of conclusions. AHRQ staff provided project oversight and reviewed the draft report and manuscript. The investigators are solely responsible for the manuscript's content and the decision to submit it for publication.

Appendix Figure 2 shows the search and selection results and Appendix Table 1 shows the strength of evidence ratings. No study evaluated the comparative effectiveness of current antiviral treatments on long-term clinical outcomes. Three trials found no differences between various dual- or triple-therapy regimens in short-term (6 months after regimen completion) mortality but reported few deaths (20 total)^20-22.

Virologic Outcomes

Ten trials (n = 66 to 3070) compared dual therapy with pegylated interferon alfa-2b plus ribavirin versus dual therapy with pegylated interferon alfa-2a plus ribavirin^6,21,23-30 (Appendix Table 2). Four trials were restricted to genotype 1 infection^21,27,28,30. The prevalence of baseline cirrhosis ranged from less than 5% to 20%^23,29,31,32, and the prevalence of elevated aminotransferase levels ranged from 60% to 100%^{23-25,29,30,32}. Eleven trials (n = 60 to 1465)^33-43 compared different durations of dual therapy, 6 trials (n = 53 to 602)^39,44-48 compared different doses of pegylated interferon as part of dual therapy, and 4 trials (n = 97 to 1831)^35,49-51 compared different doses of ribavirin as part of dual therapy for genotype 2 or 3 infection (Appendix Table 2). One trial was rated as good quality⁴⁰, 4 trials as poor quality^24,30,38,47, and the remainder as fair quality. Methodologic shortcomings included open-label design or inadequately described blinding^{23-25,27-29,33-39,42-52}, high or unclear attrition^{21,23,24,29,35,38,51}, and unclear or inadequate randomization or methods for allocation concealment^{24,25,27-30,34,36-39,41-48}.

Dual therapy with standard-dose (1.5 mcg/kg per week) pegylated interferon alfa-2b was associated with a slightly lower likelihood of SVR than was dual therapy with standard-dose (180 mcg per week) pegylated interferon alfa-2a (pooled relative risk [RR], 0.87 [95% CI, 0.80 to 0.95]; I² = 27%.) (Figure 1), with a pooled absolute difference of 8 percentage points (CI, 3 to 14 percentage points), on the basis of 7 trials (5 fair-quality and 2 poor-quality trials)^{21,23-25,27,29,30}. Results were similar when the meta-analysis included a trial³¹ that evaluated triple-therapy regimens, a trial⁶ published only as an abstract, and 2 trials that evaluated nonstandard doses of pegylated interferon alfa-2b^26,28 or when the analysis excluded poor-quality trials^24,30.

The largest trial (n = 3070), the Individualized Dosing Efficacy vs. Flat Dosing to Assess Optimal Pegylated Interferon Therapy (IDEAL) study, found no difference in likelihood of SVR for genotype 1 infection between 2 doses of pegylated interferon alfa-2b (1.0 mcg/kg per week or 1.5 mcg/kg per week) plus ribavirin, 800 to 1400 mg/d, or pegylated interferon alfa-2a, 180 mcg per week, plus ribavirin, 1000 to 1200 mg/d (range, 38% to 41%)²¹. Excluding IDEAL because of differential ribavirin dosing had little effect on the pooled estimate but eliminated statistical heterogeneity (6 trials; pooled RR, 0.83 [CI, 0.76 to 0.90]; I² = 0%)^{23-25,27,29,30}.

Duration Effects

Two fair-quality trials found no difference between 48 and 24 weeks of dual therapy in the likelihood of SVR in genotype 2 or 3 infection (pooled RR, 0.97 [CI, 0.84 to 1.1]; I²= 43%)^35,43. Four trials (1 good-quality trial and 3 fair-quality trials) found that 24 weeks of dual therapy was associated with a higher likelihood of SVR than was 12 to 16 weeks (pooled RR, 1.2 [CI, 1.0 to 1.3]), but the lower limit of the CI nearly crossed 1 and statistical heterogeneity was present (I²= 80%) (Figure 1)^36,38,40,42. The 1 trial that found no difference (RR, 1.0 [CI, 0.93 to 1.1]) reported high overall SVR rates (94% to 95%), was restricted to genotype 2 infection, and used a somewhat different ribavirin dosing regimen⁴². Excluding this trial reduced statistical heterogeneity, but the estimate was similar (3 trials; pooled RR, 1.2 [CI, 1.1 to 1.3]; I²= 47%)^36,38,40.

Three fair-quality trials of rapid virologic responders (undetectable HCV RNA by week 4) found no difference in the likelihood of SVR between 24 and 12 to 16 weeks of dual therapy (pooled RR, 0.99 [CI, 0.86 to 1.1]; I²= 66%) ^34,39,41. Absolute differences ranged up to 10 percentage points in either direction.

Dose Effects

Lower-dose pegylated interferon alfa-2b as part of dual therapy was associated with a lower likelihood of SVR than was a higher dose (typically 1.5 mcg/kg per week) in genotype 2 or 3 infection, although the upper limit of the CI nearly crossed 1.0 (pooled RR, 0.90 [CI, 0.81 to 0.99]; I²= 20%), on the basis of 6 trials (4 fair-quality trials and 2 poor-quality trials) (Figure 1)^38,44-48. Excluding the poor-quality trials^38,47 or 1 trial that evaluated an atypical dosing regimen⁴⁶ had little effect on the pooled estimate.

Two fair-quality trials found no clear difference between induction regimens of pegylated interferon alfa-2b (higher initial doses followed by lower doses) plus ribavirin versus standard fixed-dose dual therapy^53,54.

Two fair-quality trials of pegylated interferon alfa-2a found no difference between 1000 to 1200 and 800 mg of ribavirin daily (n = 492), or between 400 mg and 800 mg daily, in likelihood of SVR (n = 282)^35,49. One fair-quality trial (n = 1831) of pegylated interferon alfa-2b found no difference between ribavirin, 800 mg/d (flat dose), and 800 to 1400 mg/d (weight-dosed)⁵¹.

One fair-quality trial (n = 60) that primarily enrolled patients with advanced fibrosis or cirrhosis found pegylated interferon alfa-2a plus ribavirin, 600 to 800 mg/d, to be associated with a lower likelihood of SVR than was ribavirin, 1000 to 1200 mg/d (45% versus 72%; RR, 0.62 [CI, 0.40 to 0.98])⁵⁰.

Triple Therapy

Two fair-quality trials (n = 1097 and 520) compared triple therapy with boceprevir, pegylated interferon alfa-2b, and ribavirin versus dual therapy for antiviral treatment–naive patients with genotype 1 infection (Appendix Table 3)^22,55. Seven percent to 10% of patients had cirrhosis or severe fibrosis at baseline. Methodologic shortcomings included open-label design⁵⁵ or high attrition²². A 48-week boceprevir regimen (4 weeks of dual-therapy lead-in followed by 44 weeks of triple therapy) was associated with a higher likelihood of SVR than was 48 weeks of dual therapy (pooled RR, 1.8 [CI, 1.6 to 2.1]; I²= 0%), with a pooled absolute increase of 31 percentage points (CI, 23 to 39 percentage points) (Figure 2)^22,55. Other triple-therapy regimens evaluated in the trials (28 weeks with or without dual-therapy lead-in, 48 weeks without dual-therapy lead-in, or response-guided triple therapy for 28 or 48 weeks) were associated with lower or similar SVR rates compared with the 48-week regimen with lead-in.

One trial (n = 75) found that triple therapy with weight-based ribavirin, 400 to 1000 mg/d, was associated with a trend toward lower likelihood of SVR compared with triple therapy with standard-dose (800 to 1400 mg/d) ribavirin (36% versus 50%; RR, 0.71 [CI, 0.39 to 1.3])⁵⁵.

Six randomized trials compared triple therapy with telaprevir, pegylated interferon, and ribavirin versus dual therapy for genotype 1 infection (Appendix Table 3)^20,31,56-59. One trial used pegylated interferon alfa-2b⁵⁷, 1 evaluated regimens with pegylated interferon alfa-2a or alfa-2b³¹, and the remainder used pegylated interferon alfa-2a. The prevalence of baseline cirrhosis ranged from 0% to 11%. One trial⁵⁸ was rated as good quality and the remainder as fair quality. Methodologic shortcomings included open-label design or unclear blinding procedures^31,56,59, unclear randomization methods^56,58, and unclear attrition^57,58. In all triple-therapy regimens, telaprevir was administered with pegylated interferon plus ribavirin for the first 8 to 12 weeks. For regimens longer than 12 weeks, dual therapy was continued to the end of treatment.

Three trials (n = 189 to 323) found a 24-week fixed-duration telaprevir regimen was associated with a higher likelihood of SVR than was 48 weeks of dual therapy (pooled RR, 1.5 [CI, 1.3 to 1.8]; I²= 0%) (Figure 2), with an absolute increase of 22 percentage points (CI, 13 to 31 percentage points)^56-58. Excluding a trial that evaluated pegylated interferon alfa-2b instead of alfa-2a had no effect on the estimate⁵⁷. Two trials found no difference between 12 weeks of triple therapy and 48 weeks of dual therapy^56,58, and 1 trial found no difference between 48 and 24 weeks of telaprevir triple therapy⁵⁸.

One trial (n = 1088) found response-guided triple therapy with telaprevir (triple therapy for 8 or 12 weeks followed by dual therapy for a total of 24 or 48 weeks, depending on extended rapid virologic response) to be associated with a higher likelihood of SVR than was dual therapy for 48 weeks (RR, 1.6 [CI, 1.4 to 1.9]), with an absolute increase of 25 to 31 percentage points²⁰.

One trial found similar SVR rates (81% to 85%) for response-guided triple-therapy regimens that varied on telaprevir dose (750 mg 3 times daily versus 1125 mg 2 times daily) and type of pegylated interferon (alfa-2a versus alfa-2b)³¹. Another trial of extended rapid virologic responders to initial triple therapy with telaprevir reported similar, high SVR rates with 24- and 48-week regimens (92% and 88%, respectively)⁵⁹.

Effectiveness in Subgroups

In patients with genotype 1 infection, 1 trial of dual therapy with pegylated interferon alfa-2b versus alfa-2a²¹, 2 trials of 48 weeks of triple therapy with boceprevir and dual-therapy lead-in versus 48 weeks of dual therapy^22,55, and 2 trials of triple therapy with telaprevir (response-guided or fixed duration) versus 48 weeks of dual therapy^20,57 found no clear differences in RR estimates based on race, sex, age, baseline fibrosis, and weight. For boceprevir, the RR estimate was higher with a baseline HCV RNA viral load greater than 600 to 800,000 IU/mL (pooled RR, 2.0 [CI, 1.7 to 2.3]; I²= 0%) than with a lower viral load (pooled RR, 1.3 [CI, 1.0 to 1.5]; I²= 0%)^22,55, but there was no clear difference in RR estimates for telaprevir triple therapy versus dual therapy according to baseline viral load in 2 trials^20,57. Across regimens, absolute SVR rates were lower in older patients, black patients, patients with more advanced fibrosis, and patients with higher viral load. Four trials of dual therapy with pegylated interferon alfa-2b versus alfa-2a found no clear difference in RR estimates according to genotype, although absolute SVR rates were lower by 24% to 42% with genotype 1^6,23,24,29.

Harms of Antiviral Treatments

Six head-to-head trials of dual therapy with pegylated interferon alfa-2b versus alfa-2a found no difference in risk for withdrawal due to adverse events (6 trials; pooled RR, 1.1 [CI, 0.73 to 1.7]; I²= 42%)^{21,23,24,28-30}. Excluding 1 outlier trial (RR, 4.2 [CI, 1.5 to 12])²³ eliminated statistical heterogeneity, but the pooled estimate was similar (5 trials; pooled RR, 0.88 [CI, 0.7 to 1.1]; I²= 0%).

Two trials found dual therapy with pegylated interferon alfa-2b to be associated with lower risk for serious adverse events than was dual therapy with pegylated interferon alfa-2a (pooled RR, 0.76 [CI, 0.61 to 0.95]; I²= 0%) ^21,29. There were no differences between dual-therapy regimens in risk for anemia, thrombocytopenia, depression, fatigue, myalgia, or influenza-like symptoms (Appendix Table 4). Dual therapy with pegylated interferon alfa-2b was associated with higher risk for headache (3 trials; pooled RR, 1.1 [CI, 1.1 to 1.2]; I²= 0%)^21,23,28 and lower risk for rash (2 trials; pooled RR, 0.79 [CI, 0.71 to 0.88]; I²= 0%)^21,28 and neutropenia (5 trials; pooled RR, 0.61 [CI, 0.46 to 0.83]; I²= 38%). In the largest study (the IDEAL trial), dual therapy with either pegylated interferon was associated with serious adverse events in about 4% of patients, fatigue in 65%, headache in 45%, nausea in 40%, myalgia in 25%, neutrophil count less than 500 cells/mm³ in 5%, and hemoglobin level less than 85 g/L in 3%²¹.

Excluding the low-dose pegylated interferon alfa-2b group from the IDEAL trial had little effect on pooled estimates, except that pegylated interferon alfa-2b became associated with increased risk for depression (3 trials; pooled RR, 1.2 [CI, 1.0 to 1.4]; I²= 0%)^21,23,28. Excluding 2 poor-quality trials had little effect on pooled estimates^24,30.

Two trials found a 48-week boceprevir regimen with dual-therapy lead-in was associated with higher risk for neutropenia (pooled RR, 1.8 [CI, 1.5 to 2.3]; I² = 0%), dysgeusia (pooled RR, 2.5 [CI, 2.0 to 3.2]; I² = 0%), anemia (pooled RR, 2.0 [CI, 1.4 to 2.8]; I² = 0%), and thrombocytopenia (pooled RR, 3.2 [CI, 1.2 to 8.2]; I² = 0%) than dual therapy for 48 weeks (Appendix Table 4)^22,55. About 25% of patients receiving triple therapy experienced anemia (4% to 5% severe, defined as hemoglobin level less than 80 or less than 85 g/L) and about 33% neutropenia (8% to 15% severe, defined as neutrophil count < 500 cells/L). There were no differences in risk for withdrawal due to adverse events, serious adverse events, or other adverse events.

A 24-week regimen of triple therapy with telaprevir was associated with higher risk for anemia (3 trials; pooled RR, 1.3 [CI, 1.1 to 1.5]; I² = 0%) and rash (3 trials; pooled RR, 1.4 [CI, 1.1 to 1.7]; I² = 0%) than was dual therapy for 48 weeks, but there were no statistically significant differences in risk for serious adverse events, withdrawal due to adverse events, neutropenia, depression, fatigue, headache, chills/rigors, or influenza-like symptoms (Appendix Table 4)^56-58. Triple therapy was also associated with increased risk for thrombocytopenia in 1 trial (RR, 1.8 [CI, 1.2 to 2.5])⁵⁷. About half of the patients randomly assigned to telaprevir experienced rash (severe rash in 7% to 10%) and about half had anemia (severe anemia in 4% to 11%)^56-58.

One trial found response-guided therapy with telaprevir for 24 to 48 weeks was associated with higher risk for withdrawal due to adverse events (RR, 3.8 [CI, 2.6 to 5.7]), anemia (RR, 2.0 [CI, 1.6 to 2.5]), rash (RR, 1.5 [CI, 1.2 to 1.8]), and severe rash (5% versus 1%; RR, 4.6 [CI, 1.6 to 13]) than dual therapy for 48 weeks²⁰.

No trial reported harms in patient subgroups. Three trials of dual therapy with pegylated interferon alfa-2b versus alfa-2a for genotype 1 infection reported pooled estimates for harms similar to the estimates based on all trials^21,30,31.

Association Between SVR and Clinical Outcomes

Nineteen cohort studies (n = 105 to 16,864) evaluated the association between an SVR after antiviral therapy and mortality or complications of chronic HCV infection (Appendix Table 5)^60-78. Duration of follow-up ranged from 3 to 9 years. Ten studies were conducted in Asia^{60,67-72,75,77,78}. Eight^{64-66,72,75-78} were rated as poor quality and the remainder as fair quality. Although all studies reported adjusted risk estimates, only 8^{60,21,63,67-70,73} evaluated 5 key confounders (age, sex, genotype, viral load, and fibrosis stage). No study clearly described assessment of outcomes blinded to SVR status.

The largest study (n = 16,864) had the fewest methodologic shortcomings⁶¹. It adjusted for multiple potential confounders, including age, sex, viral load, presence of cirrhosis, multiple comorbid conditions, aminotransferase levels, and others. It also stratified results by genotype. In a predominantly male, Veterans Affairs population, SVR after antiviral therapy was associated with lower risk for all-cause mortality than was no SVR, after a median of 3.8 years (adjusted hazard ratio, 0.71 [CI, 0.60 to 0.86], 0.62 [CI, 0.44 to 0.87], and 0.51 [CI, 0.35 to 0.75] for genotypes 1, 2, and 3, respectively). Mortality curves began to separate as soon as 3 to 6 months after SVR assessment.

Eighteen other cohort studies also found SVR to be associated with decreased risk for all-cause mortality (adjusted hazard ratios, 0.07 to 0.39)^{60,69,72,73,75-78}, liver-related mortality (adjusted hazard ratios, 0.04 to 0.27)^{60,62,63,69,70,72,74,76,77}, hepatocellular carcinoma (adjusted hazard ratios, 0.12 to 0.46)^{60,62,63,67,68,71,73-76,78}, and other complications of end-stage liver disease versus no SVR, with effects larger than in the Veterans Affairs study. The subgroup of studies that focused on patients with advanced fibrosis or cirrhosis at baseline^{62,63,65-68,74-76} or that were conducted in Asia^{60,67-72,75,77,78} reported similar ranges of risk estimates.

Antiviral therapy for chronic HCV infection continues to evolve. No study evaluated comparative effectiveness of antiviral regimens on long-term clinical outcomes. Such trials are a challenge to carry out because of the long time course over which complications of HCV infection develop.

In lieu of direct evidence on long-term clinical outcomes, SVR rates are the primary outcome measure with which to evaluate comparative effectiveness. For treatment-naive patients, dual therapy with pegylated interferon alfa-2b is associated with a lower likelihood of SVR than is dual therapy with pegylated interferon alfa-2a (absolute difference, about 8 percentage points). Although there was no difference between dual-therapy regimens in risk for withdrawals due to adverse events, pegylated interferon alfa-2b was associated with a lower risk for serious adverse events, suggesting potential tradeoffs between benefits and harms. However, serious adverse events were reported in only 2 trials^21,29, the absolute difference was only about 1%, and antiviral-related adverse events are generally self-limited.

For genotype 2 or 3 infection, standard doses and durations (24 weeks) of pegylated interferon as part of dual therapy are more effective than shorter regimens or lower doses, lending support to current dosing guidance^4,79,89. Evidence on differential effects of ribavirin dose is limited, although differences were small in most studies.

The relative ineffectiveness of dual therapy for genotype 1 infection has led to ongoing efforts to identify more effective treatments. Recent trials found triple therapy with boceprevir or telaprevir superior to dual therapy, with SVR rates approaching the 70% to 80% observed in trials of dual therapy for genotype 2 or 3 infection^{20,22,31,55-59}. This has important implications for treatment, as well as for screening, because screening benefits depend in part on the effectiveness of available treatments⁸¹.

Triple therapy for genotype 1 infection is also associated with shorter duration of treatment, an important consideration given the high frequency of adverse effects associated with interferon-based therapy. However, triple therapy is also associated with increased risk for hematologic adverse events (neutropenia, anemia, and thrombocytopenia) with boceprevir and anemia and rash (including severe rash in less than 10% of patients) with telaprevir, although there was no clear increase in risk for serious adverse events overall. Across all antiviral regimens, absolute treatment response rates are lower in older patients; black patients; and patients with higher baseline viral load, genotype 1 infection, or more advanced fibrosis.

The strongest evidence on the association between virologic and clinical outcomes is a large Veterans Affairs cohort study that found SVR to be associated with a 30% to 50% reduction in mortality risk, after adjustment for many confounders⁶¹. The rapid separation of mortality curves in this study suggests possible residual confounding, given the typically protracted course of HCV infection. Therefore, estimates of benefit may be exaggerated, although it is not possible to determine to what degree. Eighteen other cohort studies also found that SVR was associated with decreased risk for serious complications of chronic HCV infection, but these studies had more methodologic shortcomings than did the Veterans Affairs study.

Our study has limitations. We excluded non–English-language articles. We did not perform formal analyses for publication bias because of the small numbers of trials, but analyses of abstracts and searches of clinical trials registries did not suggest publication bias. Meta-analyses were performed by using the DerSimonian–Laird random-effects model, which results in CIs that are slightly too narrow when heterogeneity is present, so that pooled estimates with 95% CIs close to 1.0 should be interpreted cautiously⁸². Estimates and conclusions based on small numbers of trials should also be interpreted cautiously. For example, pooled estimates based on 2 trials can be unreliable, particularly when statistical heterogeneity is present. The trials generally met criteria for efficacy studies, which could limit their applicability because of exclusion of patients with comorbid conditions, and greater adherence than typically observed in clinical practice. Almost all of the randomized trials were funded by pharmaceutical companies^83,84.

Additional research would help clarify the comparative effectiveness of antiviral treatments. Studies are needed to understand the long-term clinical outcomes associated with different antiviral treatments, the long-term harms of telaprevir and boceprevir, the comparative effectiveness of triple therapy with telaprevir versus boceprevir, and effective strategies to improve adherence⁸⁵. Other direct-acting antiviral agents, including second-generation protease inhibitors, polymerase inhibitors, NS5A inhibitors, and others are in active development, with all-oral, interferon-sparing regimens expected within the next few years⁸⁶.

Source: Agency for Healthcare Research and Quality.

Acknowledgment: The authors thank Robin Paynter, MLIS; Rose Campbell, MLIS; AHRQ Task Order Officer Christine Chang, MD, MPH; and USPSTF Medical Officer Iris Mabry-Hernandez, MD, MPH. They also thank Tracy Dana, MLS; Christina Bougatsos, MPH; and Ian Blazina, MPH, from Oregon Health & Science University, who assisted in data extraction and quality checking.

Grant Support: By the Agency for Healthcare Research and Quality (Contract Number 290-2007-10057-I, Task Order 8), Rockville, Maryland.

Potential Conflicts of Interest: Dr. Chou: Grant (money to institution): AHRQ; Support for travel to meetings for the study or other purposes (money to institution): AHRQ. Dr. Hartung: Grant (money to institution): AHRQ (Institutional career development award 1K12HS019456-01). Dr. Wasson: Grant (money to institution): AHRQ. Dr. Cottrell: Grant: AHRQ. Dr. Fu: Grant (money to institution): AHRQ. Disclosures can be viewed at www.acponline.org/authors/icmje/ConflictOfInterestForms.do?msNum=M11-1085.

Requests for Single Reprints: Roger Chou, MD, 3181 SW Sam Jackson Park Road, Mail Code BICC, Portland, OR 97239; e-mail, chour@ohsu.edu.

Current author addresses and author contributions are available at http://www.annals.org.

Key Questions:
1a. What is the comparative effectiveness of antiviral treatment in improving health outcomes in patients with HCV infection?
1b. How does the comparative effectiveness of antiviral treatment for health outcomes vary according to patient subgroup characteristics?
2a. What is the comparative effectiveness of antiviral treatments on intermediate outcomes on the rate of SVR?
2b. How does the comparative effectiveness of antiviral treatment for intermediate outcomes vary according to patient subgroup characteristics?
3a. What are the comparative harms associated with antiviral treatments?
3b. Do these harms differ according to patient subgroup characteristics?
4. Have improvements in SVR been shown to reduce the risk or rates of adverse health outcomes from HCV infection?

*Including but not limited to HCV genotype, age, race, sex, stage of disease, or genetic markers.

* Some studies applied to more than 1 key question. Studies of induction versus fixed-dose regimens and outcomes related to quality of life and histologic changes are not reported here but can be found in the full report.¹¹

HCV = hepatitis C virus; HR = hazard ratio; RR = relative risk; SVR = sustained virologic response.
* Outcomes related to quality of life and histologic changes are not reported in this publication but can be found in the full report¹¹.
† Details about strength of evidence for subgroup effects for specific drug comparisons are available in the full report¹¹.

Cirrhosis = METAVIR (Meta-analysis of Histologic Data in Viral Hepatitis) F4, Ishak 5–6, or equivalent; Minimal or no fibrosis = METAVIR F0–F1, Ishak 0–2, or equivalent; CLEO = Club Epatologi Ospedalieri; HCV = hepatitis C virus; IDEAL = Individualized Dosing Efficacy vs. Flat Dosing to Assess Optimal Pegylated Interferon Therapy; WIN-R = Weight-based dosing of pegINterferon α-2b and Ribavirin.
* Treatment duration varied according to genotype: typically 48 wk for genotype 1 or 4 and 24 wk for genotype 2 or 3.
† Published as abstract only; only included in sensitivity analysis.
‡ Severe fibrosis or cirrhosis.
§ More than 2 times the upper limit of normal.
|| Sample sizes and results are restricted to patients with genotype 2 or 3 infection.
¶ Patients who had undetectable virus by wk 4 were randomly assigned to 12 or 6 wk of ribavirin.
** Treatment for 12 wk if HCV RNA undetectable at 4 wk and for 24 wk if detectable.
†† Alanine aminotransferase level >40 IU/L.

Ibid = twice daily; HCV = hepatitis C virus; ILLUMINATE = Illustrating the Effects of Combination Therapy with Telaprevir; PROVE = Protease Inhibition for Viral Evaluation; SPRINT = Serine Protease Inhibitor Therapy; tid = thrice daily. Cirrhosis = METAVIR (Meta-analysis of Histologic Data in Viral Hepatitis) F4, Ishak 5–6, or equivalent. Minimal or no fibrosis = METAVIR F0–F1, Ishak 0–2, or equivalent.
* Dosing recommended by the U.S. Food and Drug Administration for boceprevir in antiviral-naive patients with cirrhosis at baseline.
† Severe fibrosis or cirrhosis.
‡ Response-guided duration: 28 wk of pegylated interferon/ribavirin if HCV RNA is negative from wk 8–24. Patients not meeting these criteria continued until wk 48.
§ Response-guided duration: 24 wk of pegylated interferon/ribavirin if HCV RNA negative from week 4 through week 12. Patients not meeting these criteria continued until week 48.
‖Dosing regimen recommended by the U.S. Food and Drug Administration for telaprevir.
¶ Response-guided duration: 24 weeks of treatment with pegylated interferon/ribavirin if HCV RNA is negative from wk 4–20. Patients not meeting these criteria continued until wk 48.
** Patients with undetectable HCV RNA at wk 4 and wk 12 randomly assigned to 24 or 48 wk of dual therapy.

RR, relative risk
* Intervention 1: interferon α-2b; intervention 2: interferon α-2a.
† Intervention 1: triple therapy with pegylated interferon and ribavirin for 48 wk with boceprevir from weeks 5 to 24; intervention 2: dual therapy for 48 wk.
‡ Intervention 1: triple therapy with telaprevir, pegylated interferon α-2, and ribavirin for 12 wk followed by dual therapy for 12 wk; intervention 2: dual therapy for 48 wk.

HCV = hepatitis C virus; HR = hazard ratio; NR = not reported; SVR = sustained virologic response; UK = United Kingdom.
* SVR defined in all studies as undetectable HCV RNA in serum 6 mo after the end of antiviral therapy, except as noted.
† Study primarily evaluated patients who received pegylated interferon plus ribavirin.
‡ SVR defined as undetectable HCV RNA on at least 2 occasions at least 2 years after completion of therapy.
§ Hepatic decompensation, complications of portal hypertension, hepatocellular carcinoma, liver transplantation, and liver-related mortality.
^|| Calculated from estimates for SVR vs. untreated and no SVR vs. untreated.
¶ Hepatic decompensation, upper gastrointestinal bleeding secondary to rupture of esophageal or gastric varices, hepatocellular carcinoma, liver transplantation, and liver-related or liver-unrelated mortality.
** Duration of undetectability to meet criteria for SVR not reported.
†† Decompensated liver disease (ascites, variceal bleeding, hepatic encephalopathy, spontaneous bacterial peritonitis), hepatocellular carcinoma, liver transplantation, and liver-related mortality.