fax: (43) 140 400 4735
Treatment of chronic hepatitis C genotype 1 patients at an academic center in europe involved in prospective, controlled trials: Is there a selection bias?†
Article first published online: 30 NOV 2011
Copyright © 2011 American Association for the Study of Liver Diseases
Volume 55, Issue 1, pages 30–38, January 2012
How to Cite
Beinhardt, S., Staettermayer, A. F., Rutter, K., Maresch, J., Scherzer, T. M., Steindl-Munda, P., Hofer, H. and Ferenci, P. (2012), Treatment of chronic hepatitis C genotype 1 patients at an academic center in europe involved in prospective, controlled trials: Is there a selection bias?. Hepatology, 55: 30–38. doi: 10.1002/hep.24671
Potential conflict of interest: Dr. Ferenci is a consultant for, advises, is on the speakers' bureau of, and received grants from Roche. He consults for, and is on the speakers' bureau of, Madaus. He also consults for Merck, Vertex, and Boehringer-Ingelheim.
- Issue published online: 21 DEC 2011
- Article first published online: 30 NOV 2011
- Accepted manuscript online: 19 SEP 2011 09:04AM EST
- Manuscript Accepted: 25 AUG 2011
- Manuscript Received: 13 JUL 2011
Pegylated interferon-alpha2/ribavirin (peg-IFN/RBV) is the standard of care (SOC) for patients with chronic hepatitis C (CHC) infection. Currently, direct-acting antiviral agents (DAAs) are evaluated in clinical trials. The aim of this study was to compare baseline characteristics and sustained virologic response (SVR) rates in patients included in clinical trials to those receiving SOC. Medical records of all 503 treatment-naïve patients with CHC, genotype (GT) 1, referred over a 4-year period (January 2006-December 2009) were reviewed. Only 310 of 503 (62%) patients received antiviral therapy, 141 were enrolled in randomized, controlled trials (“study patients”; 101 in DAA studies), and 169 received SOC. At baseline, viral load and platelet count were higher and bilirubin was lower in study patients than in SOC patients. History of psychiatric disorders was more common in SOC patients (43 [25%] versus study patients with 18 [13%]; P < 0.01). Liver biopsy was obtained in 98% of study patients, but only in 59% of SOC patients. Twenty-nine (21%) and 40 (40%) study and SOC patients, respectively, had advanced fibrosis (F3/4; P = 0.001). By intent-to-treat analysis, SVR rates were higher in DAAs (64%; 95% confidence interval [CI]: 53.4-74.4) than in SOC patients (46%; 95% CI: 37.9-53.7; P < 0.01), but not different when calculated on a treated-per-protocol (TPP) basis. Interleukin (IL)28B GT was equally distributed in both cohorts. By chance, more patients treated with IFN/RBV had rs12979860 C/C-GT (up to 44%) than DAA-treated patients. If analyzed according to the IL28B polymorphism, TPP SVR rates did not reach statistically significant differences among study and SOC patients. Conclusions: Baseline characteristics slightly favored study patients, but IL28B GT and treatment adherence were the most important factors determining outcome. Thus, the applicability of the results of controlled studies has to be tested in a “real-world” setting. (HEPATOLOGY 2012
The current standard of care (SOC) for patients infected with hepatitis C virus (HCV), genotype (GT) 1, is a response-guided combination therapy with pegylated interferon (peg-IFN) alpha2 and weight-based ribavirin (RBV).1-3 This treatment may cure about 50% of these patients.4-6 The discovery of direct-acting antiviral agents (DAAs) in 2002 led to the development of a plethora of small molecules able to block the replication of HCV, including novel antiviral targets.7, 8 Based on the impressive improvements in sustained virologic response (SVR) rates in phase III trials,9-14 the first two protease inhibitors were recently approved by the U.S. Food and Drug Administration. The results of these trials will be used for clinical decision making and counseling patients in the near future, but little is known about their applicability in “real-life” conditions.
Furthermore, polymorphisms in the region of the interleukin (IL)28B gene on chromosome 19 are associated with early and sustained virologic response (SVR)15-17 and may effect therapy strategies as well as the design and interpretation of clinical studies in the future.18
The aim of our study was to analyze baseline characteristics, including IL28B GT distribution (rs12979860) and therapy outcome of treatment-naïve patients with chronic hepatitis C (CHC), GT-1, treated with SOC or within study settings to explore whether there are differences in treatment outcome resulting from possible selection bias. Additionally, we investigated reasons for noninclusion and nontreatment of patients referred to our tertiary referral center.
Patients and Methods
Medical records of all 503 treatment-naïve patients with CHC, GT-1, referred to our center from January 1, 2006 to December 31, 2009 were reviewed retrospectively. At referral, patients had a blood test, including HCV GT and viral load, and received an appointment to see a hepatologist. Twenty-two patients had contraindications for peg-IFN/RBV-based therapy; the remaining 481 were evaluated for the feasibility for antiviral therapy. All patients were informed about the option to participate in ongoing studies (DAAs [n = 101]: telaprevir, danoprevir, TMC435, BI201355, mericitabine, balapiravir, and IDX 184; or IFN-based treatments [n = 40]: albIFN alpha-2b or response-guided treatment2). Every patient, for whatever reason, not eligible or willing to be included into a study was offered SOC therapy: 171 patients did neither opt to take part in a study nor had SOC resulting from several reasons (Fig. 1), 169 received SOC, and 141 were treated within a study regimen. For analysis of the IL28B GT, patients were either tested at one of the follow-up visits or were recalled for testing. All patients gave informed consent for genetic testing. In 79% of all treated patients, the IL28B rs12979860 GT could be determined.
Medical History and Clinical Status.
History of intravenous drug abuse (IVDA), alcohol consumption, nicotine abuse, drug-substitution therapy, history of psychiatric disorders, hypertension, diabetes mellitus, coronary artery disease, concomitant medication intake, mode of infection, country of origin, sex, age, and body mass index (BMI; calculated by dividing weight [kg] divided by height2 [m2]) were assessed.
Routine laboratory parameters at baseline included hemoglobin, white blood cell count (WBC), platelet count, aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma glutamyle transferase (γ-GT), bilirubin, blood glucose, cholesterol, triglycerides, creatinine, alpha-fetoprotein (AFP), viral load at screening time point (quantified by the real-time polymerase chain reaction COBAS TaqMan HCV Test; Roche Diagnostics, Pleasanton, CA), and HCV GT (determined by Versant genotype assay LiPA 2.0; Bayer HealthCare LLC, Subsidiary of Bayer Corporation, Tarrytown, NY). IL28B genotyping was assayed as described previously.17
A single pathologist graded necroinflammatory activity (A0-A3) and fibrosis stage (F0-F4) according to METAVIR score.19 Stage of fibrosis of nonbiopsied patients was evaluated, in some cases, by transient elastography (FibroScan; EchoSens SA, Paris, France; n = 2 in study cohort) or by noninvasive markers, such as the aspartate aminotransferase platelet ratio index score.
Quantitative variables were expressed as mean ± standard deviation (SD). Data were analyzed by using the Student t test for Gaussian variables, the Mann-Whitney U test for non-Gaussian variables, as well as the chi-square test. To determine whether variables were normally distributed, the Kolmogorow-Smirnow test was applied. Statistical analysis was performed by Microsoft Excel 2010 SP2 MSO (Microsoft Corporation, Redmond, WA) and SPSS 2006 for Windows (version 16; SPSS, Inc., Chicago, IL).
Within the 4-year observation period, only 310 (62%) of the 503 treatment-naïve GT-1 patients referred to our unit received antiviral treatment. Figure 1 shows treatment assignments of the patients: 169 patients (34%) received peg-IFN-alpha2a/RBV combination therapy (referred to as “SOC patients”). This group included 26 potentially eligible patients. Sixteen patients fulfilled all criteria and were willing to participate in a trial with DAA, but could not be included because of stringent timelines for recruitment and restriction on the number of patients per study center. Five patients refused liver biopsy, and 5 did not want to participate in a study.
Eleven patients had study exclusion criteria (e.g., hemochromatosis, age >65 years, status post–brain trauma, alcohol abuse, kidney disease, iron deficiency anemia, depression, ongoing substitution therapy, mixed GT [i.e., 1+2a+2c], too low viral load or too low platelets, or suspicion of autoimmune hepatitis).
The remaining 132 patients received SOC because of the lack of an ongoing study, and some patients opted for SOC. Finally, 141 (28%) patients participated in a prospective trial (referred to as “study patients”).
Comparison of Baseline Characteristics According to Treatment Group.
Patient characteristics and baseline laboratory findings are fully shown in Tables 1 and 2. SOC and study patients did not differ regarding age, BMI, sex, GT-1 subtype, mode of infection, hemoglobin, WBC, AST, ALT, γ-GT, blood glucose, cholesterol, triglycerides, AFP, history of IVDA, frequency of comorbidities, nicotine abuse, alcohol consumption before therapy, drug-substitution therapy, and country of origin.
|Characteristic||SOC Patients||Study Patients||DAA Patients|
|Age (years, mean [SD])||46.4 ± 12.2||46.8 ± 11.5||47.0 ± 11.3|
|BMI (kg/m2, mean [SD])||25.3 ± 4.1||24.8 ± 3.6||24.7 ± 3.5|
|Sex (N, F/M)||169 (76/93)||141 (60/81)||100 (40/60)|
|Race, N (%)|
|Caucasian||165 (98)||139 (99)||97 (97)|
|Austrian||123 (73)||97 (69)||72 (72)|
|Non-Austrian||42 (25)||42 (30)||28 (28)|
|African||1 (0.6)||1 (0.7)||1 (1)|
|Asian||3 (1.8)||2 (1.4)||2 (2)|
|History of drug abuse, N (%)||67 (40)||44 (31)||28 (28)|
|History of alcohol consumption >40 g/day, N (%)||73 (43)||58 (41)||37 (37)|
|Nicotine abuse, N (%)||75 (44)||64 (46)||42 (42)|
|Drug-substitution therapy, N (%)||25 (15)||11 (8)||5 (5)§|
|History of psychatric disorder, N (%)||43 (25)||18 (13)†||9 (9 )|
|Comorbidities, N (%)|
|Hypertension||27 (16)||21 (15)||13 (13)|
|Diabetes mellitus||4 (2)||3 (2)||2 (2)|
|Coronary artery disease||3 (2)||0||0|
|Not specified||73 (43)||49 (35)||33 (33)|
|Comedication, N (%)|
|Psychiatric||41 (24)||12 (9)*||8 (8)*|
|Other, nonpsychiatric||77 (46)||45 (32)§||31 (31)§|
|Mode of infection, N (%)|
|Blood/blood products||36 (21)||23 (16)||19 (19)|
|Plasma donation||7 (4)||10 (7)||8 (8)|
|Nosocomial||0||2 (1)||2 (2)|
|IVDA||48 (28)||35 (25)||21 (21)|
|Mixed mode of infection||7 (4)||10 (7)||8 (8)|
|Unknown||69 (41)||60 (43)||42 (42)|
|Parameter||SOC Patients||All Study Patients||DAA Study Patients|
|HCV RNA (MIU/mL)*||2.92 ± 4.43||4.94 ± 8.25†||6.08 ± 9.32†|
|HCV RNA ≥800,000 IU/mL, N (%)||49 (60)||58 (71)§||81 (81)†|
|HCV 1 subtype, N (%)|
|1a||66 (39)||46 (33)||33 (33)|
|1b||92 (54)||88 (62)||64 (64)|
|Indeterminate||7 (4)||5 (4)||2 (2)|
|Mixed (1a/1b)||3 (1.8)||2 (1.4)||1 (1)|
|Liver biopsy available, N (%)||99 (59)||138 (98)||100 (100)|
|F1||8 (8)||17 (12)||13 (13)|
|F2||51 (52)||92 (67)||60 (62)|
|F3||12 (12)||17 (12)‡||15 (15)‡|
|F4||28 (28)||12 (9)‡||9 (9)‡|
|APRI score*1||0.79 ± 0.78||0.62 ± 0.60∥||0.60 ±0.06∥|
|Female patients||13.52 ± 1.2||13.83 ± 0.95||13.81 ± 1.0|
|Male patients||15.14 ± 1.3||15.35 ± 1.30||15.52 ± 1.0|
|White blood cell count (g/L)*||6.56 ± 2.11||6.9 ± 1.94||7.0 ± 2.02|
|Platelet count (g/L)*||214 ± 73||237 ± 65‡||237 ± 67‡|
|AST (U/L)*||56.4 ± 39.1||55.5 ± 39.8||53.4 ± 39.7|
|ALT (U/L)*||81.1 ± 66.6||81.7 ± 63.6||75.9 ± 63.8|
|γ-GT (U/L)*||80.0 ± 80.0||64.1 ± 66.1||62.8±69.1∥|
|Serum bilirubin (mg/dL)*||0.78 ± 0.4||0.63 ± 0.4†||0.59±0.31†|
|Blood glucose (mg/dL)*||96.6 ± 18.3||93.2 ± 26.1||94.3 ± 27.3|
|Cholesterol (mg/dL)*||185 ± 38||182 ± 93.8||179 ± 91.4|
|Triglycerides (md/dL)*||114 ± 63||110 ± 72||100 ± 58∥|
|Serum creatinine (mg/dL)*||0.87 ± 0.14||0.84 ± 0.22||0.84 ± 0.24|
|AFP*||6.3 ± 23.4||5.8 ± 18.0||6.24 ± 20.9|
|IL-28B (rs12979860)||N = 115||N = 129||N = 92|
Study patients had slightly higher platelet count and baseline viral load as well as lower bilirubin levels than SOC patients (Table 2). In addition, patients assigned to DAA studies had lower γ-GT (62.8 ± 69.1 versus 80.0 ± 80.0 IU/L; mean [SD]; P < 0.05) and triglycerides (100 ± 58 versus 114 ± 63 mg/dL; P < 0.05).
IL28B rs12979860 was tested in 115 consenting (68%) SOC patients and 129 (91%) study patients. Distribution of GTs was not different among SOC and study patients (see Table 2).
Overall, 137 (97%) study patients, but only 99 (59%) SOC patients, had a liver biopsy. In those with liver biopsy, advanced fibrosis (F3/F4) was present only in 29 (21%) study patients, but in 40 (40%) SOC patients (P = 0.001). SOC patients presented more frequently with history of a psychiatric disorder (18 [13%] versus 43 [25%]; P < 0.01) and more often received psychiatric medication (12 [9%] versus 41 [24%]; P < 0.001) as well as other nonpsychiatric concomitant medication (45 [32%] versus 77 [46%]; p < 0.02; Table 1). More SOC than DAA patients were on drug-substitution therapy (25 [15%] versus 5 [5%]; P < 0.05) and had more comorbidities (73 [43%] versus 33 [33%]; P < 0.05; Table 1). These differences reflect stringent inclusion and exclusion criteria in studies investigating DAAs.
All SOC and study patients reached treatment endpoint (Table 3). Twelve patients, participating in the prematurely terminated balapiravir study,20 and the patient within the phase I study (IDX184), were excluded from further analysis. All of the remaining 87 patients in DAA studies were eligible for treatment-outcome analysis.
|Outcome||SOC||Study (All)||IFN Studies||DAA Studies*|
|Treatment endpoint reached, N (%)||169 (100)||127 (100)||40 (100)||87 (100)|
|Dropout, N (%)||22 (13)||1 (1)†||1 (3)||0‡|
|Lost to FU, N (%)||19 (12)||7 (6)||0§||7 (8)|
|Withdrawal/AE, N (%)||3 (2)||3 (2)||2 (5)||1 (1)|
Because of the small number of patients per study group (mean, 10) and confidentiality issues in yet unpublished trials, the outcome was only analyzed for the whole group of DAA patients.
Both on an intent-to-treat (ITT) or a treated-per-protocol (TPP) basis, SVR rates were highest in the DAA study group (ITT: 64%, 95% confidence interval [CI]: 53.4-74.4; TPP: 71%; 95% CI: 59.6-80.6) and lowest in SOC patients (ITT: 46%, 95% CI: 37.9-53.7; P < 0.01; TPP: 63%, 95% CI: 53.4-74.4; Table 3). Drop-out (13% versus 1%; P < 0.001) and lost-to-follow-up rates (12% versus 6%; not significant) were higher in SOC than study patients. The same was true in patients receiving peg-IFN/RBV within a study regimen (Table 3).
In 79% of all treated patients, the IL28B rs12979860 GT was determined and allowed us to analyze SVR rates accordingly. There were no differences in total distribution of IL28B polymorphism between the SOC group and overall study cohort (Table 2). IL28B GT had a major effect on SVR, irrespective of the treatment given (Fig. 2). For example, more patients receiving an IFN/RBV-based treatment within study settings carried the favorable C/C-GT than patients within DAA study settings (44% versus 30%; not significant), explaining their high SVR rates. On the other hand, the unfavorable T/T-GT was present only in 6% of IFN study patients, but in 19% of patients receiving DAAs. Considering only telaprevir studies the frequency of C/C-GT was just 25%, in contrast to 33% in other DAA studies.
With each new drug approved, the question comes up of whether data generated in randomized, controlled trials are representative for an average patient treated outside a major expert center. This is also true for the recently licensed protease inhibitors, which markedly improved SVR rates in patients with CHC, GT-1.9-14 As far as we know, this is the first study to assess a potential selection bias in patients with CHC treated within randomized, controlled studies. At least, the outcome of cohort studies in hepatitis C may be influenced by a selection bias.21
Not surprisingly, some baseline characteristics, adherence to treatment, and the IL18B phenotype had the biggest effect on treatment outcome. Patients receiving SOC had more advanced liver disease, a more frequent history of psychiatric disorders, and a substantial proportion was on drug-substitution therapy. Differences in baseline characteristics22-25 reflect inclusion and exclusion criteria and may create “perfect to treat” patients. Whether the slight differences documented in this study are clinically relevant remains unknown, but history of psychiatric disorders, need of drug-substitution therapy, and a higher grade of liver fibrosis may influence treatment outcome, as well.
Because of major differences in access to medical care, data from a European study center cannot be directly applied to the U.S. population of HCV patients. It may also be difficult to draw conclusions from the largely homogenous population of HCV patients seen in Vienna to the largely heterogeneous population of HCV patients seen in the United States. In particular, in this study, except for 7 patients (2.3%; 5 Asians and 2 black Africans), all patients were Caucasians, whereas in the United States, approximately 19% and 40% of treatment eligible or noneligible patients, respectively, are black.26
Analysis of data from the National Health and Nutrition Examination Survey, conducted in 2005-2008, revealed that HCV-positive patients were less likely to be insured than HCV-negative individuals (61.2% versus 81.2%). Of all HCV-positive patients, 66.7% were eligible for anti-HCV treatment, but only 54.3% of HCV-positive treatment candidates had any type of insurance coverage, and only 36.3% of treatment-eligible patients had health insurance.27 In contrast, approximately 99% of all Austrians are insured (including jobless people), and treatment for chronic hepatitis B and C is fully covered. A similar insurance coverage rate was reported from Taiwan.28 Thus, there is no direct financial benefit for patients to participate in a study, and no patient contacted our center for the sole purpose of being enrolled in a clinical trial. All patients were referred from general practitioners for evaluation of CHC and were offered treatment, if there were no contraindications. Our center treats approximately one third of all HCV patients from Vienna as well as the surrounding suburbs. These represent the “typical mix” of patients with CHC (36% with history of drug abuse, 19% infected by blood or blood products, and 42% of unknown etiology). Income, education, and ability to travel to the medical center played no role in whether patients could participate in a trial or not.
The major limitations in recruiting patients into trials were the requirement of liver biopsy and the need of drug substitution. A liver biopsy was obtained in 98% of study participants, but in only 59% of SOC patients. Among biopsied patients, advanced fibrosis was more common in SOC patients (40%) than in study patients (21%). This may be an underestimate, because patients in the SOC group may have been cirrhotic, but did not undergo biopsy, because they had obvious clinical, laboratory, or radiographic features of cirrhosis.
Most studies excluded patients on drug-substitution therapy, representing a substantial proportion of patients in “real life.” In SOC patients on drug-substitution therapy lost to follow-up, the rate was 56%, but only 9% in study patients.
The limitations of this study were its retrospective nature and the variations within each treatment protocol. A direct comparison of the SVR rates in SOC and study patients was not possible. SOC patients represent an inhomogeneous group consisting of persons not willing or not being able (because of stringent timelines for recruitment or the presence of exclusion criteria) to participate in randomized, controlled trials. Furthermore, most SOC patients were treated by a response-guided treatment concept. Treatment extension may have slightly improved outcome on peg-IFN/RBV-based treatment in our SOC cohort. Also, the design of the DAA studies varied considerably and did not allow a comparison between different regimes. Except for one study, DAA trials included a placebo arm; the PROVE-2 trial9 also evaluated patients not receiving RBV. Because all phase II and III studies with telaprevir were published,9, 11, 14 we could analyze the outcome of patients treated with telaprevir (Table 4; Fig. 2). All other studies are as yet unpublished and their results cannot be presented separately because of confidentially issues. The balapiravir study was prematurely stopped because of severe side effects,20 and therefore, patients participating in that trial were not included for further SVR analysis.
|Placebo DAA studies|
|TPP = ITT||13/21||62||4/4||100||9/17||53|
|All (including placebo)†|
|All (excluding placebo)†,‡|
In unblinded studies, the outcome of the DAA or placebo groups could be compared. The improved SVR rates on DAA became only detectable in the ITT analysis. Similarly, ITT-SVR rates in patients receiving peg-IFN/RBV within a study setting (63%) with stringent inclusion and exclusion criteria were higher than in SOC patients (46%; P < 0.02), reinforcing the role of adherence to optimize treatment outcome. Overall, participation in well-controlled, prospective trials may increase adherence by a strict visit schedule and also allows an early treatment of side effects, resulting in lower drop-out rates.
In addition, study patients are, in general, better informed, having detailed discussion about study design, treatment procedures, and potential side effects before signing informed consent. Thus, adherence to therapy is crucial to optimize SVR rates. Pill burden negatively correlates with adherence and compliance.29 Simple dosing (i.e., one pill once-daily) helps to maximize adherence, particularly when combined with frequent reinforcing visits.30 Unfortunately, the recently licensed HCV protease inhibitors will increase pill burden substantially.
IL28B GT was the most important determinant for SVR. Irrespective of treatment, C/C homozygotes had the highest SVR rates (DAA: TPP, 96%; ITT, 92%; SOC: TPP, 89%, ITT, 65%; Table 4; Fig. 2). This does not mean that triple therapy with DAAs does not confer benefit for C/C homozygotes. Though the TPP-SVR in the C/C-GT was not different among patients receiving a DAA or SOC, overall ITT-SVR was higher (DAA: 92%; SOC: 65%; P < 0.025). In T-allele carriers, SVR rates were higher in DAA patients (DAA: TPP, 62% versus 46%; P < 0.01; ITT, 57% versus 36%; P < 0.01). The overall genotype distribution between SOC and study patients was similar (Table 1), but there were differences in subgroups (data not shown). Because of small sample size, the observed differences were not significant, but they may have affected the final outcome and may explain the high SVR rates of patients on IFN/RBV. The impact of IL28B polymorphism in triple therapy is controversial; two recently presented analyses of phase III trials yielded conflicting data.31, 32
In summary, inclusion and exclusion criteria in randomized, controlled trials slightly favor patients receiving DAA over those on SOC. Patients in DAA studies were less likely to have advanced liver fibrosis or to be intravenous drug abusers.
Irrespective of the chosen treatment, the most important factors to obtain SVR were IL28B GT and better treatment adherence. These findings have to be considered when deciding which patient to treat first with DAAs in the future, because a scarcity in disposability and side-effect management is suspected.33 Motivation of patients to adhere to treatment depends largely on the experience of the treatment setting, as shown recently in a study from New Mexico.34
- 12Efficacy of boceprevir, an NS3 protease inhibitor, in combination with peginterferon alfa-2b and ribavirin in treatment-naïve patients with genotype 1 hepatitis C infection (SPRINT-1): an open-label, randomized, multicenter phase 2 trial. Lancet 2010; 376: 705-716., , , , , , et al.
- 13SPRINT-2 Investigators. Boceprevir for untreated chronic HCV genotype 1 infection. N Engl J Med 2011; 31: 364: 1195-1206., , , , , , et al.;
- 20Phase IIB study of balapiravir (RG1626; Nucleoside analogue inhibitor of HCV polymerase) plus peginterferon alfa-2a (40KD) and ribavirin for CHC genotype 1: final results. J Hepatol 2010; 52 ( Suppl. 1): S302-S303., , , , , , et al.
- 31Telaprevir substantially improved SVR rates across all IL28B genotypes in the Advance Trial. J Hepatol 2011; 54( Suppl. 1): S542: 1369., , , , , , et al.
- 32The Aspire Trial: TMC435 in treatment-experienced patients with genotype 1 HCV infection who have failed previous PEGIFN/RBV treatment. J Hepatol 2011; 54( Suppl. 1): S546: 1376., , , , , , et al.