Potential conflict of interest: Peter Ferenci is a member of the global advisory board and of the speaker's bureau of Roche, Basel, Switzerland, and of Rottapharm-Madaus, Monza, Italy. He is also advisor to Boehringer Ingelheim, Vertex/Tibotec, Pfizer, and MSD Austria. He has also received an unrestricted research grant from Roche Austria. Harald Hofer, Andreas Maieron, and Rudolf Stauber serve as speakers for Roche Austria and MSD Austria. Petra Steindl-Munda serves as speaker for Roche Austria. All other authors have no financial disclosures to report.
The IL28B genotype is the most important pretreatment predictor of treatment outcome in patients with chronic hepatitis C. The impact of the rs12979860 genotype on relapse was retrospectively evaluated in genotype 1/4 patients who received response-guided therapy with peginterferon alpha-2a 180 μg/week plus ribavirin 1,000/1,200 mg/day in a large, randomized, multicenter study. Patients with a rapid virologic response (RVR: hepatitis C virus [HCV] RNA <50 IU/mL) at week 4 were treated for 24 weeks; those with a slow virologic response (no RVR but undetectable HCV RNA or ≥2-log10 decrease at week 12) were randomized to 48 (group A) or 72 weeks of treatment (group B). Relapse rates were compared by rs12979860 genotype (C/C versus combined T/C or T/T [T/*]) in patients with confirmed end-of-treatment response and known end-of-follow-up status (sustained virologic response [SVR] or relapse). The rs12979860 genotype was determined for 340/551 study participants. In patients with RVR and C/C or T/* genotype, relapse rates were similar (10.7% versus 15.2%). In patients randomized to groups A and B, relapse rates were similar in patients with C/C genotype randomized to group A (26.9%) and group B (20.0%). In contrast, relapse rates in T/* patients differed markedly between groups A and B, overall (42.9% and 18.8%; P < 0.025, respectively) and in those with low (<400,000 IU/mL: 37.5% versus 18.8%) and high (≥400,000 IU/mL: 45.0% versus 18.8%) baseline viral loads. Conclusion: The results suggest that the benefits of extended therapy are restricted to patients with a T allele. Relapse rates are highest in patients with T/* genotype and are markedly higher in slow responders treated for 48 weeks compared with 72 weeks. (HEPATOLOGY 2011;)
The likelihood that an individual patient with chronic hepatitis C virus (HCV) infection will achieve a sustained virologic response (SVR) after treatment with pegylated interferon plus ribavirin is highly variable. Baseline host and viral characteristics, and the early viral kinetic response during antiviral treatment, exert a significant influence on the outcome of treatment.
Response-guided therapy has become standard practice for patients infected with HCV genotype 1 or 4. This patient management strategy involves measurement of the viral kinetic response at weeks 4 and 12 of treatment with dynamic adjustment of the duration of treatment. Patients with a rapid virologic response (RVR) clear HCV RNA and achieve consistently higher SVR rates (i.e., ≈80% and higher) whether treated for 24 or 48 weeks, than patients without an RVR.1-7 Patients without an RVR who clear the virus by week 12 have lower SVR rates and may profit from extending treatment to 72 weeks, although the evidence in favor of this approach is less robust than that supporting abbreviated treatment in patients with an RVR.4, 6-12
The ability to predict SVR in patients with HCV infection has increased markedly with the discovery of a single nucleotide polymorphism (SNP) that influences the response to pegylated interferon plus ribavirin. A genome-wide association study (GWAS) showed that an SNP (rs12979860) in proximity to the promoter for the IL28B gene on chromosome 19 (19q13.13) had a highly significant impact on SVR rates in HCV genotype 1 patients enrolled in a large, randomized, multicenter trial (IDEAL).13 Patients with a homozygous C/C genotype at this locus had significantly higher SVR rates than those who carried the T allele.13 Spontaneous clearance of acute HCV infection was also shown to be more likely in individuals with the C/C genotype.14 These findings have since been confirmed in several independent analyses in genotype 1 patients,15-17 and evidence is accumulating for a role of IL28B genotype in the response to therapy in patients infected with other HCV genotypes.18-21 Determination of rs12979860 genotype is thus a useful prognostic factor in patients infected with HCV genotype 1 and, with the availability of a commercial test in the U.S., it is now a tool that can be used in routine clinical practice for informing treatment decisions.22
Unfortunately, the rs12979860 genotype does not have a 100% positive predictive value for SVR; indeed, RVR is a better predictor of SVR than IL28B genotype.16, 17 It remains to be determined whether the combination of rs12979860 genotype and early viral kinetics can be used to optimize treatment duration and SVR rates. The database from a recently completed, large, randomized, multicenter trial in which the primary objective was to compare relapse rates in patients randomized to 48 and 72 weeks of treatment provides the opportunity to examine the impact of the rs12979860 genotype on outcomes in HCV genotype 1 and 4 patients undergoing response-guided therapy.12
EVR, early virologic response; GWAS, genome-wide association study; HCV, hepatitis C virus; RVR, rapid virologic response; SNP, single nucleotide polymorphism; SVR, sustained virologic response.
Patients and Methods
This Austrian multicenter study included treatment-naive chronic hepatitis patients with HCV genotype 1 or 4. All genotype 1 patients were Caucasians; all genotype 4 patients originated from Egypt. The study design and primary results of the study are published elsewhere.3, 12 Briefly, all patients initiated treatment with 180 μg/week peginterferon alpha-2a (PEGASYS, Roche, Basel, Switzerland) and ribavirin 1,000 mg/day (body weight ≤75 kg) or 1,200 mg/day (body weight >75 kg) (COPEGUS, Roche). The duration of treatment was determined on the basis of the on-treatment virologic response at treatment weeks 4 and 12. All patients with undetectable HCV RNA (<50 IU/mL by qualitative polymerase chain reaction [PCR] assay, COBAS AMPLICOR HCV Test, v. 2.0, Roche Diagnostics, Branchburg, NJ) at week 4 (RVR) were treated for 24 weeks (group D). Patients with detectable HCV RNA (≥50 IU/mL) at week 4 and either undetectable HCV RNA or ≥2 log10 drop in HCV RNA at week 12 (early virologic response [EVR]) were randomized to complete a total of 48 weeks (group A) or 72 weeks (group B) of treatment. Patients with <2 log10 drop in HCV RNA at week 12 (virologic nonresponders) continued treatment until week 24 (group C) and, if they had undetectable HCV RNA at week 24, they were allowed to continue therapy until week 72. Treatment was stopped in patients with detectable HCV RNA at week 24 (nonresponders).
Patients with undetectable HCV RNA at the end of the planned course of treatment (end-of-treatment [EoT] responders) were followed up for 24 weeks. SVR was defined as undetectable HCV RNA (50 IU/mL) at end of follow-up. Conversely, virologic relapse was defined as detection of HCV RNA (≥50 IU/mL) at the end of follow-up in a patient with an EoT virologic response.
Quantitative serum HCV RNA tests were done with the COBAS AMPLICOR HCV Monitor Test, v. 2.0 (limit of quantification 600 IU/mL). Qualitative tests were done with the COBAS AMPLICOR HCV Test, v. 2.0 (limit of detection 50 IU/mL). Samples with undetectable HCV RNA by the qualitative test were retested with the more sensitive Roche TaqMan assay (limit of detection 10 IU/mL).
Testing for rs12979860.
Whole blood samples obtained and stored in ethylene diamine tetraacetic acid (EDTA)-containing collection tubes were used for IL28B genotype testing. DNA was subsequently isolated and the rs12979860 SNP in the region of the IL28B gene was analyzed by the StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA) with a custom TaqMan SNP Genotyping Assay developed in collaboration with Applied Biosystems as described.23 Gene sequences were obtained from the NCBI Entrez SNP Database (http://www.ncbi.nlm. nih.gov/sites/entrez). GCCTGTCGTGTACTGAACCA was used as the forward and GCGCGGAGTGCAAT TCAA as the reverse primer in the genotyping assay for rs12979860.
All statistical calculations were done with SigmaPlot v. 11 (Systat Software, Erkrath, Germany). Treatment outcome (SVR or relapse) was analyzed by χ2 test in the various treatment groups by IL28B genotype (C/C versus T/C or T/T). Only patients who completed treatment as per protocol and with known EoT and end-of-follow-up (SVR or relapse) results were included in the analysis of relapse and SVR. This ensured that the analysis of outcome by treatment duration was not confounded by the inclusion of patients who withdrew prematurely and received less than the planned duration of treatment.
All patients included in this analysis provided informed written consent to rs12979860 genotype testing.
The IL28B rs12979860 polymorphism was determined for 340 of 551 (61.7%) study participants overall. Across the four treatment groups the proportion of patients represented in the rs12979860 genotype analysis ranged from 60% to 67% of the original intention-to-treat (ITT) population (Fig. 1). The overall rs12979860 genotype frequency was C/C: 115 (33.8%), T/C: 175 (51.5%), and T/T: 50 (14.7%). The baseline characteristics of these patients are shown in Table 1 and the rs12979860 genotype frequencies are presented by treatment group in Fig. 2. Twelve patients with known genotypes (C/C: n = 4, T/C: n = 7, T/T: n = 1) dropped out before week 12, were not randomized and, thus, were not included in further analyses. Of 328 patients who were assigned to a treatment group, 223 had a baseline HCV RNA level ≥400,000 IU/mL (84 C/C, 108 T/C, 31 T/T) and 105 had a baseline HCV RNA level <400,000 IU/mL (27 C/C, 60 T/C, 18 T/T).
Table 1. Baseline Characteristics of Patients in the Trial Overall and by Treatment Group Assignment
All Patients with an rs12979860 Genotype Result
Patients Without an RVR at Week 4 and with an EVR at Week 12
Patients Without an EVR
Patients with an RVR at Week 4
12 dropouts excluded.
EVR = undetectable HCV RNA (<50 IU/mL) or ≥2-log10 drop in HCV RNA at week 12; RVR = undetectable HCV RNA (<50 IU/mL) at week 4.
Patients with an rs12979860 genotype result, n (%)
The rs12979860 genotype was determined for 97 of 150 (64.7%) patients with an RVR assigned to group D. The majority of these patients (60 [61.9%]) had the homozygous C/C genotype, and 37 individuals carried the T allele (35 had the T/C genotype, 36.1%; 2 had the T/T genotype, 2.1%).
Of 97 patients with an RVR assigned to group D and with a known rs12979860 genotype, 93 (95.9%) achieved an EoT response, of whom four were lost to follow-up. Among the 89 patients with known end-of-follow-up results, 78 patients (87.6%) achieved an SVR and 11 (12.4%) relapsed. SVR rates exceeded 80%, regardless of rs12979860 genotype (Fig. 3A). Relapse rates were numerically lower in patients with the C/C genotype, but did not differ significantly from those in patients carrying the T allele (T/C and T/T combined) overall (Fig. 3B). The results were similar when the analysis was restricted to genotype 1 patients (Fig. 3C,D). Only one of the 17 HCV genotype 4 patients with an RVR relapsed. This individual had the C/C genotype.
Among individuals with the C/C genotype and baseline HCV RNA levels of <400,000 IU/mL and ≥400,000 IU/mL, respectively, 5.0% (1/20) and 13.9% (5/36) of patients relapsed. Among those patients with T allele (T/C or T/T genotype) and baseline HCV RNA level <400,000 IU/mL the relapse rate was 11.5% (3/26). Only seven patients with T allele and a baseline HCV RNA level ≥400,000 IU/mL achieved an RVR: five achieved an SVR and two relapsed.
Patients with an EVR.
The rs12979860 genotype was determined for 183 of 289 (63.3%) patients without an RVR who achieved an EVR at week 12 and were randomized to groups A or B. Fifty (27.3%) patients had the homozygous C/C genotype and 133 individuals carried the T allele (99 [54.1%] had the T/C genotype, and 34 [18.6%] had the T/T genotype).
The distribution of rs12979860 genotypes was similar in groups A and B (Fig. 1). Among patients with known rs12979860 genotypes in groups A and B, respectively, 82/93 (88.2%) and 63/90 (70.0%) achieved an EoT response, of whom 51/82 (62.2%) and 51/63 (81.0%) achieved an SVR, and 31/82 (37.8%) and 12/63 (19.0%) patients relapsed.
SVR rates were numerically higher in patients treated for 72 weeks regardless of rs12979860 genotype, although the positive impact of extended treatment was magnified in patients who carried a T allele (Fig. 4A). Relapse rates, the primary outcome in the original study, were numerically lower in patients treated for 72 weeks (20.0%, 95% confidence interval [CI] = 10.2-30.9) compared with 48 weeks (26.9%, 95% CI = 27.3-49.2; odds ratio [OR] = 2.58; 95% CI = 0.32-6.83), and were markedly lower in patients who carried a T allele (48 versus 72 weeks: 42.9%, 95% CI = 29.7-56.8 versus 18.8%, 95% CI = 8.9-32.7; OR = 3.55; 95% CI = 1.33-7.97; P < 0.025; Fig. 4B). The results were similar when the analysis was restricted to genotype 1 patients (48 versus 72 weeks: 43.8% versus 19.5% [OR = 3.21; 95% CI 1.228-8.38]; P < 0.025; Fig. 4C,D).
None of the 15 HCV genotype 4 patients in groups A or B had the C/C genotype. Among the HCV genotype 4 infected patients who carried a T allele, three of eight individuals in group A relapsed and one of seven patients in group B relapsed.
Among patients with a T allele, extended treatment was associated with lower relapse rates in patients with a baseline HCV RNA level <400,000 IU/mL (18.8% versus 37.5% in patients treated for 48 weeks) and in those with a baseline HCV RNA level ≥400,000 IU/mL (18.8% versus 45.0% in patients treated for 48 weeks) (Fig. 5). Among patients with a C/C genotype and a high baseline HCV RNA level, relapse rates were similar among those randomized to 48 weeks (6/27, 26%) and 72 weeks of treatment (3/14, 21.4%).
Relapse rates and SVR rates were similar in patients who were heterozygous (T/C) or homozygous (T/T) for the T allele (data not shown).
Patients Without an EVR.
The rs12979860 genotype results were available for 48 of 78 (61.5%) patients without an EVR. Only one patient (2.1%) (body mass index 27.5; no cirrhosis, baseline viral load: 66,600 IU/mL) had the C/C genotype, 34 (70.8%) had the T/C genotype, and 13 (27.1%) had the T/T genotype. Four patients (all T/C) were negative at week 24, of whom three achieved an EoT response. Two of these individuals achieved an SVR and one relapsed.
Overall SVR Rate.
If the results are subjected to an ITT analysis including all patients in whom the rs12979860 genotype was determined, the SVR rates in patients with an RVR assigned to group D were 83.3% (50/60; 95% CI: 71.5-91.7) in those with a homozygous CC genotype and 75.7% (28/37; 95% CI: 58.8-88.2) in those who carried a T allele (T/C or T/T). Among patients with an EVR randomized to 48 and 72 weeks, SVR rates in patients with the homozygous C/C genotype were 70.4% (19/27; 95% CI: 49.8-86.2) and 52.2% (12/23; 95% CI: 30.6-73.2, n.s.), respectively, and SVR rates in patients who carried a T allele (T/C or T/T) were 48.5% (32/66; 95% CI: 36.0-61.1) and 58.2% (39/67; 95% CI: 45.5-70.1; n.s.). In group C, the one patient with a C/C genotype did not achieve an SVR and two of 47 patients who carried a T allele achieved an SVR.
The results of this study extend what is known about IL28B polymorphisms in patients with chronic hepatitis C by providing insight into the relationship between rs12979860 genotype and relapse, and into the impact of rs12979860 genotype on response-guided therapy for HCV genotype 1 or 4 patients. Relapse rates were lower among patients with a C/C genotype compared with those who carried a T allele. This observation was not only apparent in patients without an RVR who achieved an EVR (slow responders) and who were randomized to 48 or 72 weeks of treatment, but also among those with an RVR. The overall incidence of relapse was approximately 2-fold greater in patients with a C/C genotype who were slow responders than in those who were rapid responders (≈24% versus 11%). Among slow responders treated with a standard 48-week regimen, the relapse rate was considerably higher in carriers of a T allele compared with those with the C/C genotype (42.9% versus 26.9%). To our knowledge, such clear evidence of an association between relapse and rs12979860 genotype has not been reported previously. McCarthy et al.15 reported, to the contrary, that relapse was not influenced by rs12979860; however, comparatively few relapsers (n = 29) were included in their diverse cohort of patients that included individuals with all HCV genotypes and both treatment-naive and previously treated patients.
This analysis also provides clear insight into how rs12979860 modifies the impact of treatment duration on relapse rates. In slow responders with a C/C genotype the incidence of relapse was lower in those randomized to 72 weeks as compared with 48 weeks (20.0% versus 26.9%), although the magnitude of the difference is modest and the number of patients included in these calculations is too small to be statistically significant. The impact of treatment duration on relapse, however, was much more dramatic in patients who carried the T allele. The overall relapse rate was reduced from 42.9% in slow responders who were randomized to 48 weeks of treatment to 18.8% among those randomized to 72 weeks. Remarkably, the relapse rate in slow responders with a T allele treated for 72 weeks approached that in patients with an RVR who were treated for 24 weeks (18.8% versus 15.2%, respectively). The benefits of extended treatment on relapse rates were particularly evident when baseline HCV RNA level was considered. Among patients with a T allele treated for 48 weeks, relapse rates were increased with baseline HCV RNA level. In contrast, in patients randomized to the 72-week regimen the relapse rate was identical in patients with low and high baseline HCV RNA levels. This suggests that the 72-week regimen is optimal in terms of minimizing relapse for slow responders who carry a T allele.
These findings suggest that the benefits of an extended 72-week treatment regimen are primarily limited to patients who carry a T allele, and may explain in part the inconsistent findings of the impact of extended treatment durations in slow responders.4, 6-12
A small number of HCV genotype 4 patients were included in this analysis. Relapse was uncommon in genotype 4 patients who achieved an RVR, and of 15 patients with an EVR none had a C/C genotype. Among genotype 4 patients who were slow responders and who carried a T allele, relapse rates were numerically lower in group B (1/7) than in group A (3/8). Although consistent with the results in individuals infected with HCV genotype 1, the low number of patients prevents us from drawing firm conclusions.
This study also confirms several previously reported phenomena related to rs12979860 genotype in patients with chronic hepatitis C.15-17 Individuals with a C/C genotype were more likely to achieve an RVR and an SVR than patients who carry the T allele. However, not all patients with a C/C genotype achieve an RVR and an SVR, and not all patients with a T allele are slow responders. Overall, approximately 50% of patients with a C/C genotype achieved an RVR and, regardless of the on-treatment response, approximately 83% of those with an EoT response achieved an SVR. Among patients with a T allele, approximately 16% achieved an RVR and the overall SVR rate in those with an EoT response was 72%. Although a much higher proportion of patients with a C/C genotype achieved an RVR compared with carriers of a T allele, SVR rates were similar in patients with an RVR regardless of genotype.
Interestingly, among patients with an RVR the majority of those with a C/C genotype (64%) had a baseline HCV RNA level ≥400,000 IU/mL and the majority of those with a T allele (79%) had a baseline HCV RNA level <400,000 IU/mL. This is consistent with previous reports that patients with C/C genotype have higher mean viral loads than patients who carry a T allele.13, 17
These findings confirm that achievement of an RVR at week 4 is the best predictor of SVR in patients receiving pegylated interferon plus ribavirin therapy.24 Although IL28B genotype is the best pretreatment predictor of SVR, the addition of this variable results in, at best, marginal improvement in the positive predictive value of RVR for SVR.24 The IL28B polymorphism explains only part of the response to interferon-based therapies. It has recently been suggested that increased expression of interferon-stimulated genes is a better predictor of nonresponse than IL28B polymorphism alone.25, 26 Thus, there may be scope to further improve the ability to predict response before treatment is initiated.
The results suggest that extension of treatment to 72 weeks in HCV genotype 1 and 4 patients with a slow response may decrease relapse rates in patients who carry a T allele, whereas patients with a C/C genotype derived little benefit from treatment extension. Relapse rate was the primary endpoint of the parent study. Calculation of SVR rates by ITT analysis is difficult because patients not completing the assigned treatment had to be considered treatment failures. As in the parent study, by ITT analysis, lower relapse rates did not result in significantly higher SVR rates. Indeed, when the data were subjected to an ITT analysis the SVR rates were actually lower in patients with a C/C genotype who achieved an EVR but no RVR and were randomized to extended treatment (52.2% versus 70.4% in those randomized to 48 weeks). This suggests that the higher withdrawal rate with extended therapy (with patients being imputed as SVR failures) is not offset by an increased SVR rate in C/C patients. In contrast, the ITT analysis shows that the SVR rate tended to be higher in patients who carried a T allele and who were randomized to extended treatment (58.2% versus 48.5% in those randomized to 48 weeks). Thus, if an extended 72-week regimen is being contemplated for an individual patient, the decision should be informed by rs12979860 genotype status and should only be generally considered for slow responders who carry a T allele.
Interest in extended-duration therapy is waning with the approval of the first direct-acting antiviral agents (DAAs). However, because slow virologic responders have suboptimal response rates to peginterferon/ribavirin/DAA triple therapy,28-31 response-guided therapy remains an important treatment strategy. A response-guided approach was used in the phase 3 studies of protease inhibitors by extending treatment to 48 weeks in patients with detectable HCV RNA at week 4 (telaprevir) or 8 (boceprevir).27, 28 These trials confirmed the potential for these agents to increase overall SVR rates and decrease treatment duration for many genotype 1 patients. A treatment extension beyond week 48 was not investigated so far. It is not clear, however, whether all patients will require a DAA to optimize treatment outcomes. For example, the addition of boceprevir after 4 weeks of treatment with peginterferon plus ribavirin improved SVR rates only in individuals with a <2-log10 drop in HCV RNA after the 4-week run-in period in a phase 2 study in treatment-naive genotype 1 patients.29 Furthermore, as shown in the present study, regardless of the IL28B genotype more than 80% of patients with an RVR on peginterferon/ribavirin achieve an SVR after just 24 weeks of treatment. Thus, the benefit of adding a DAA to standard peginterferon plus ribavirin therapy in these patients requires further evaluation.
Like all published studies on the role of IL28B polymorphism, this analysis was carried out retrospectively. Patients were recruited for the parent trial between 2003 and 2008, long before the importance of the IL28B genotype in the treatment of chronic hepatitis C was established. Therefore, not all patients who were enrolled in the study were represented in the genotype analysis; thus, the results must be regarded with caution. Patients were recalled and asked to participate in this analysis; thereby only 62.3% of the study participants could be tested. This compares well with other retrospective trials (i.e., 65% in the study by Mangia et al.32 and 52.2% in the study by Thompson et al.17).
The overall results of the parent study were not different from this subanalysis. Furthermore, patients were not stratified by rs12979860 genotype. It is reassuring that a similar proportion (≈60% to 67%) of patients were genotyped in each group, and that the relapse rates reported in groups A and B in the original study (33% and 17%, respectively) are generally similar to those reported in this analysis (38% and 19%, respectively). At present not a single prospective study stratified by rs12979860 genotype has been published, although studies are ongoing and preliminary data have been reported.33
In conclusion, the results of this analysis demonstrate the importance of IL28B genotype in HCV genotype 1 and 4 patients undergoing response-guided therapy. In particular, our findings suggest that the combination of virologic response and IL28B genotype are both important considerations in optimizing treatment duration. In the future, the combination of virologic response and IL28B genotype may be helpful in identifying patients who may not benefit from the addition of a DAA, and alternatively, individuals who require prolonged DAA-based triple therapy.
The authors thank Elisabeth Eder, Claudia Willheim, and Kerstin Zinober for technical help, and Drs. Bernhard Bauer, Fritz Renner, Martin Bischof, and Ludwig Kramer for referring their patients. Author contributions: Thomas-Matthias Scherzer: data collection and analysis, writing of the article. Harald Hofer, Petra Steindl-Munda, Andreas Maieron, Christian Datz. Hermann Laferl, Michael Strasser, and Rudolf Stauber: acquisition of data, critical revision of the article for important intellectual content. Albert Friedrich Stättermayer and Emina Dulic-Lakovic: acquisition of data. Peter Ferenci: study concept and design, principal investigator, analysis and interpretation of data, outlining and revising the article. Writing assistance: Support for third-party writing assistance for this article was provided by F. Hoffmann-La Roche Ltd. The study sponsor (F. Hoffmann-La Roche Ltd) was not involved in the analysis and interpretation of data.