Dual Therapy Regimens With Pegylated Interferon and Ribavirin

In lieu of direct evidence on long-term clinical outcomes, sustained virologic response (SVR) rates are the primary outcome to assess comparative benefits of different antiviral regimens. In trials of treatment-naïve patients, the likelihood of achieving an SVR was slightly lower with dual therapy with pegylated interferon alfa-2b plus ribavirin compared with dual therapy with pegylated interferon alfa-2a plus ribavirin (pooled RR 0.87, 95% CI 0.80 to 0.95; I²=27%), with a difference in absolute SVR rates of about 8 percentage points.^{20–23, 53, 55, 58} Although the largest study, the IDEAL trial, found no difference in SVR rates between dual therapy with pegylated interferon alfa-2a compared with dual therapy with pegylated interferon alfa-2b, excluding the IDEAL trial from pooled analyses resulted in similar effect estimates.²² Although there was no difference between dual therapy regimens in risk of withdrawals due to adverse events, dual therapy with pegylated interferon alfa-2b plus ribavirin was associated with a lower risk of serious adverse events than dual therapy with pegylated interferon alfa-2a plus ribavirin (pooled RR 0.76, 95% CI 0. 61 to 0.95, I²=0%), suggesting a potential tradeoff between greater benefits and harms. However, serious adverse events were only reported in two trials;^{22, 23} the rate of serious adverse events was relatively low (about 4 percent overall in IDEAL), with an absolute difference of about one percent; and adverse events with antiviral treatments generally resolve following discontinuation of therapy.

Trials found no clear differences in estimates of relative effectiveness of dual therapy with pegylated interferon alfa-2a compared with dual therapy with pegylated interferon alfa-2b in patient subgroups stratified by age, sex, race, viral load, fibrosis stage, and genotype, although absolute response rates were lower in older patients, Black patients, patients with high viral load, patients with more advanced fibrosis or cirrhosis, and genotype 1 infection.^{20–23, 56} SVR rates ranged from 24–42 percent lower in patients with genotype 1 infection compared with patients with genotype 2 or 3.

In patients with genotype 2 or 3 infection, dual therapy for 12 to 16 weeks appears to be associated with lower likelihood of SVR compared with dual therapy for 24 weeks, with no differences between 24 weeks and longer courses of therapy.^{63, 64, 66, 68, 70, 71} Standard doses of pegylated interferon alfa-2b were more effective than lower doses (no trials compared different doses of pegylated interferon alfa-2a).^{66, 73–77} Although trials comparing different ribavirin doses found no clear differences, with the exception of one trial that found lower doses associated with lower SVR rates in patients with advanced fibrosis,⁸¹ they evaluated different dose comparisons, precluding firm conclusions.^{63, 80, 82}

Triple Therapy Regimens With Pegylated Interferon, Ribavirin, and Either Boceprevir or Telaprevir

The relatively low SVR rates with pegylated interferon plus ribavirin dual therapy for genotype 1 infection (present in about three-quarters of U.S. patients with HCV infection) has led to ongoing efforts to identify more effective treatment alternatives. Recent trials found triple therapy regimens with pegylated interferon (alfa-2a or alfa-2b), ribavirin, and either boceprevir or telaprevir associated with substantially higher SVR rates than standard dual therapy with pegylated interferon (alfa-2a or alfa-2b) plus ribavirin in treatment-naïve patients with genotype 1 infection.^{30–32, 51, 59, 85–87} SVR rates with triple therapy approached the 70–80 percent rates observed with dual therapy in patients with genotype 2 or 3 infection. Trials that evaluated the telaprevir regimen recommended by the FDA (12 weeks of triple therapy with telaprevir followed by response-guided duration of 12 or 36 weeks of dual therapy) reported SVR rates of 75–80 percent.^{51, 59} Trials that evaluated the FDA-recommended boceprevir regimen for antiviral-naïve patients with cirrhosis (4 weeks of dual therapy lead-in followed by 44 weeks of triple therapy with boceprevir) reported SVR rates of 66–75 percent.^{30, 32} Trials that evaluated other regimens in antiviral naïve patients, including fixed duration telaprevir regimens, shorter fixed duration triple therapy with boceprevir, and boceprevir without dual therapy lead-in, reported similar or lower SVR rates. The SVR rates with various antiviral regimens or placebo are summarized in Table 13.

Table 13

Sustained virologic response rates with different antiviral regimens for hepatitis C virus infection.

The boceprevir regimen recommended by the FDA⁸⁴ for antiviral-naïve patients without baseline cirrhosis (4 weeks of dual therapy lead-in with pegylated interferon plus ribavirin followed by the addition of boceprevir for 24 weeks for virologic responders at weeks 8 to 24, or 4 weeks of dual therapy lead-in followed by the addition of boceprevir for 32 weeks and then 12 additional weeks of dual therapy for late virologic responders) has not been evaluated in a trial of antiviral-naïve patients. Rather, the FDA recommendation was based on a trial of previous partial responders to pegylated interferon plus ribavirin which found slightly higher SVR rates in late virologic responders who received 32 weeks of triple therapy followed by 12 weeks of dual therapy versus those who received 44 weeks of triple therapy, each following 4 weeks of dual therapy lead-in (SVR rates 79 vs. 73 percent).¹²³

As in the head-to-head trials of dual therapy with pegylated interferon alfa-2a plus ribavirin versus pegylated interferon alfa-2b plus ribavirin, relative risk estimates were similar (or there was no clear difference) in patient subgroups based on age, sex, and race, although absolute SVR rates were lower in older patients and Black patients. Triple therapy with either boceprevir or telaprevir was no more effective than dual therapy in the subgroup of patients with lower HCV-RNA viral load (<600,000 or <800,000 IU/mL).^{30, 32, 51} There was insufficient evidence to evaluate relative effectiveness of triple compared with dual therapy based on fibrosis stage.

In addition to higher likelihood of SVR, another advantage of triple therapy regimens in patients with genotype 1 infection is the potential for shorter duration (24 or 28 weeks in patients with early virologic response compared with the standard 48 weeks of dual therapy with pegylated interferon plus ribavirin). Shorter courses of treatment would probably be appealing to patients of high relevance to patients, given the high frequency of bothersome flulike symptoms associated with interferon-based therapy. Triple therapy regimens were associated with increased risk of certain harms, in particular hematological adverse events (neutropenia, anemia, and thrombocytopenia) with boceprevir and anemia and rash (including severe rash in <10 percent of patients that could result in treatment discontinuation) with telaprevir. However, there was no clear increase in risk of serious adverse events with use of protease inhibitors, and the adverse events appear to be self-limited following drug discontinuation.

Sustained Virologic Response After Antiviral Therapy and Clinical Outcomes

The strongest evidence on the association between an SVR after antiviral therapy and improved clinical outcomes is a large VA cohort study (n=16,864) that adjusted for many confounders.⁸ The VA study found decreased risk of all-cause mortality in patients who achieved an SVR compared with those who didn’t achieve an SVR across groups stratified by genotype (adjusted HR 0.71 [0.60–0.86], 0.62 [0.44–0.87] and 0.51 [0.35–0.75] for genotypes 1, 2, and 3, respectively). Despite controlling for important confounders, the possibility of residual confounding is suggested by the very rapid separation of mortality curves for patients with an SVR versus those without an SVR, which was observed at three months after assessment for SVR. This is more rapid than expected given the typically prolonged natural history of HCV infection. Therefore, estimates of effects of SVR on clinical outcomes from this study may be exaggerated, though it is not possible to determine to what degree.

Eighteen other cohort studies also found an SVR after antiviral therapy associated with decreased risk of all-cause mortality and complications of chronic HCV infection, including studies specifically of patients with baseline cirrhosis, but had more methodological shortcomings. In addition, 10 of the 19 studies were conducted in Asia, where the incidence of HCC in patients with chronic HCV infection is higher than in the United States,⁴⁶ potentially limiting their generalizability. Other studies found an SVR after antiviral therapy associated with better scores on various measures of quality of life than no SVR, but those studies focused on short-term outcomes, and typically did not adjust for confounders or blind patients to SVR status when assessing outcomes.

Findings in Relationship to What Is Already Known

Our findings regarding the comparative effectiveness of dual therapy with pegylated interferon alfa-2b plus ribavirin compared with dual therapy with pegylated interferon alfa-2a plus ribavirin are consistent with recent systematic reviews that also found the former associated with a lower likelihood of SVR.^{18, 124} Our findings of no clear difference in comparative effectiveness between 12 to 16 weeks compared with 24 weeks of response-guided dual therapy with pegylated interferon plus ribavirin in hepatitis C genotype 2 or 3 infection with rapid virologic response are discordant with a recent systematic review, which found a shorter duration of treatment associated with a lower likelihood of achieving an SVR.¹²⁵ The discrepancy may be explained by the inclusion in the other systematic review of a study that we excluded because it evaluated a nonstandard dose of pegylated interferon,⁶⁵ as well as its inclusion of subgroup analyses from trials of patients randomized to different fixed durations of therapy prior to assessment of rapid virologic response,^{64, 68, 70} which we considered separately because they did not represent randomized comparisons of response-guided treatment.

Because telaprevir and boceprevir are so new, we are unaware of other published systematic reviews on the comparative benefits and harms of regimens including these drugs, compared with standard dual therapy. Our findings on the association between achieving an SVR and reduced risk of mortality or complications associated with chronic HCV infection are consistent with a recent review that used some systematic methods.¹²⁶

Applicability

The trials included in this review generally met criteria for efficacy studies, based on the exclusion of patients with common comorbidities (such as serious psychiatric conditions or recent or ongoing substance abuse) who may receive treatments in clinical practice. In addition, the trials may have overestimated efficacy compared with what would be seen in typical practice due to improved adherence as a result of closer followup, effects of trial participation, selection of patients, or other factors. A separate review funded by AHRQ will focus on issues related to adherence in the treatment of HCV infection.¹²⁷

The severity of baseline liver disease in the patients enrolled in the trials suggests that they enrolled a broad range of patients. In trials of triple therapy with boceprevir or telaprevir, the proportion of patients with cirrhosis at enrollment ranged from <1–11 percent.^{30–32, 51, 59, 85, 87} Trials that reported the proportion of patients with minimal or no fibrosis reported rates of 27–39 percent.^{31, 51, 59, 87}

Evidence to evaluate potential differences in comparative benefits or harms in patient subgroups based on age, sex, race, and other clinical factors was relatively limited, precluding strong conclusions in these specific subgroups. The strongest evidence on the association between an SVR versus no SVR after antiviral therapy and reduced mortality comes from a study performed in a VA population, which might limit generalizability to other settings.⁸ As described above, studies conducted in Asia on the association between an SVR after antiviral therapy and risk of clinical outcomes may be of limited applicability to U.S. populations because of a higher incidence of HCC in Asian patients with chronic HCV infection.⁴⁶ However, HCC incidence is increasing in the United States in HCV-infected patients,¹²⁸ which may attenuate such concerns regarding applicability.

The results of this CER are not applicable to populations excluded from the review, including patients previously treated with antiviral therapies and excluded populations such as patients with HIV coinfection, post-transplant patients, or hemodialysis patients. Antiviral therapy is not recommended in patients following kidney transplant, and ribavirin is not recommended in those with more severe (stage 3 to 5) kidney disease since it is cleared via renal function and associated with increased risk of hemolytic anemia in this setting. Such patients were typically excluded from randomized trials of antiviral treatment.¹⁵

Implications for Clinical and Policy Decisionmaking

Our review has potential implications for clinical and policy decisionmaking. For patients with genotype 1 infection, triple therapy regimens with pegylated interferon (alfa-2a or alfa-2b), ribavirin, and telaprevir or boceprevir may be considered an alternative to dual therapy with pegylated interferon alfa-2a or alfa-2b plus ribavirin as standard treatment due to substantially superior efficacy for achieving SVR compared with dual therapy with pegylated interferon alfa-2a or alfa-2b, as well as a shorter duration of treatment. Factors that may affect decisions to utilize regimens with boceprevir or telaprevir include cost and specific harms associated with use of these drugs (such as hematologic adverse events with boceprevir and anemia and rash with telaprevir). Dual therapy with pegylated interferon alfa-2a plus ribavirin appears to be associated with higher likelihood of achieving SVR compared with dual therapy with pegylated interferon alfa-2b plus ribavirin, but absolute differences were relatively small and may be offset in part by a small increase in serious (but generally self-limited) adverse events. Therefore, decisions about which pegylated interferon to use may be affected by other considerations, such as cost, patient preferences, or other factors. For genotype 2 or 3 infection, standard doses and duration (24 weeks) of pegylated interferon as part of dual therapy are more effective than shorter regimens or lower doses, lending support to dosing guidance from the FDA and clinical practice guidelines.^{15, 33, 34} Evidence on differential effects of ribavirin dose are too limited to draw strong conclusions about optimal dosing of this component of antiviral regimens, though differences appeared relatively small.

The findings that absolute SVR rates are lower in certain subgroups (such as older patients, Black patients, patients with worse baseline fibrosis, and patients with high viral load) can be used to inform individualized decisionmaking. Patients who are less likely to achieve a SVR may make different informed decisions about therapy compared to those more likely to achieve an SVR, given the adverse effects associated with treatment.

The findings of the review are also relevant to screening recommendations, which are based in part on the effectiveness of treatments in patients found through screening to have HCV infection. Important new evidence that may affect assessments regarding potential benefits of screening include stronger evidence on the link between achieving an SVR and improvement in clinical outcomes, as well as evidence showing substantially higher SVR rates with newer triple therapy regimens with boceprevir or telaprevir in patients with genotype 1 infection, the predominant type of HCV infection in the United States.

Limitations of the Comparative Effectiveness Review Process

Our review had some potential limitations. We excluded non–English-language articles, which could result in language bias, although a recent systematic review found little empirical evidence that exclusion of non–English-language articles leads to biased estimates for noncomplementary or alternative medicine interventions.¹²⁹

We did not formally assess for publication bias with funnel plots due to small numbers (<10) of studies for all comparisons. Small numbers of studies can make interpretation of funnel plots unreliable, and experts suggest 10 studies as the minimum number of studies to perform funnel plots.⁵⁰ We included some studies which were published only as abstracts and found that their inclusion or exclusion from analyses did not change conclusions. In addition, we searched trial registries and solicited drug manufacturers for additional unpublished trials and identified none.

Another potential limitation is that we included cohort studies to evaluate the association between SVR and mortality or hepatic complications associated with chronic HCV infection. Such studies are susceptible to confounding if factors associated with SVR (such as age, race, viral load, or fibrosis stage) are also associated with these outcomes. Therefore, we only included studies that reported adjusted risk estimates, and we evaluated how well studies addressed key potential confounders as part of our quality assessment. Nonetheless, residual confounding is a possibility even in cohort studies that adjust for potential confounding.

Limitations of the Evidence Base

We identified several important limitations of the evidence base. First, studies assessing important long-term clinical outcomes associated with current antiviral treatments for chronic HCV infection are not available. In the case of antiviral regimens involving newly approved antiviral drugs, such studies are not possible yet because of the extended followup required to adequately evaluate effects on clinical outcomes. Second, no trials directly compared regimens with boceprevir compared with regimens with telaprevir. Given the increased efficacy of these regimens in patients with genotype 1 infection, trials directly comparing their effects would be helpful for informing treatment choices between these drugs. In addition, few trials have evaluated the specifically FDA-approved regimens for these drugs, limiting confidence in conclusions regarding estimates of benefits and harms for the regimens likely to be used in clinical practice. Third, almost all of the randomized trials were funded by pharmaceutical companies. Studies have shown that such studies tend to report more favorable results for drugs produced by the funder than studies funded by governmental or other sources.^{130, 131} Fourth, there was relatively limited information on effects of newer triple therapy regimens with a protease inhibitor in subgroups defined by age, body weight, baseline fibrosis stage, and other important factors. Such information would be helpful for individualizing treatment decisions with these regimens. Finally, few methodologically rigorous studies conducted in settings applicable to U.S. populations evaluated the association between achieving an SVR and improvements in clinical outcomes. Such studies would be very helpful for confirming the results of the recent, large, well-conducted VA cohort study showing an association between achieving an SVR and reduced mortality risk.⁸

Future Research

Evaluating the comparative effectiveness of current antiviral regimens on clinical outcomes in randomized trials or cohort studies is a challenge due to the long lead-time and large samples necessary to adequately assess these outcomes. This might be more feasible if the studies were to focus on populations at higher risk for complications from chronic HCV infection (e.g., patients with baseline cirrhosis, high viral load, or other risk factors for progression).

For all trials of antiviral treatments, studies that enroll broader populations with medical and psychological comorbidities, as frequently encountered in clinical practice, are needed to better understand comparative effectiveness, rather than just comparative efficacy. Studies designed using an effectiveness paradigm would also be helpful for understanding real-world effects of antiviral regimens, including effects related to the poorer treatment adherence than expected from efficacy trials. Studies that evaluate the usefulness of genomics and other methods for individualizing treatment decisions in patients with HCV infection are also needed.

Trials directly comparing triple therapy with telaprevir compared with triple therapy with boceprevir would be very helpful for understanding comparative effectiveness of these two protease inhibitors. In addition, trials evaluating the boceprevir regimen by the FDA in antiviral-naïve patients without baseline cirrhosis are needed to verify that results from studies of previously treated patients were appropriately generalized. Prolonged followup of patients exposed to telaprevir and boceprevir is needed to understand the long-term harms associated with these medications. A number of other protease inhibitors and other newer drugs for treatment of hepatitis C virus infection are currently in active development and further studies with new drugs and drug regimens are expected, including regimens without interferon.⁸⁸

It is critical that future studies that evaluate clinical outcomes in patients with an SVR versus no SVR after antiviral therapy adequately control for other factors that influence clinical outcomes in chronic HCV infection. Studies on effects of achieving an SVR on long-term quality of life would be very helpful for understanding other potential clinical benefits of antiviral therapy, but a significant challenge is whether it is possible to ethically blind patients to virologic status, which may have an important impact on assessments of quality of life.