Detection of low HCV viraemia by repeated HCV RNA testing predicts treatment failure to triple therapy with telaprevir

Alimentary Pharmacology and Therapeutics Detection of low HCV viraemia by repeated HCV RNA testing predicts treatment failure to triple therapy with telaprevir B. Maasoumy*, B. Cobb, B. Bremer*, K. Luk, P. Halfon, S. Aslam, M. P. Manns*, M. Cornberg* & H. Wedemeyer* *Department of Gastroenterology, Hepatology and Endocrinology, Hannover Medical School, Hannover, Germany. Roche Molecular Systems, Pleasanton, CA, USA. Laboratoire ALPHABIO, Hopital Ambroise Pare, Marseille, France. Correspondence to: Prof. Dr H. Wedemeyer, Carl-Neuberg Strasse 1, 30625 Hannover, Germany. E-mail: wedemeyer.heiner@mh-hannover.de Publication data Submitted 14 August 2013 First decision 12 September 2013 Resubmitted 3 October 2013 Accepted 12 October 2013 EV Pub Online 10 November 2013 SUMMARY Background Early on-treatment virological response is one of the most important predictors for sustained virological response (SVR) to treatment of chronic hepatitis C virus (HCV) genotype 1 infection with triple therapy including HCV protease inhibitors (PI). Treatment duration (24 vs. 48 weeks) is based on HCV RNA results at weeks 4 and 12 of PI therapy when HCV RNA must be undetectable to allow shorter therapy. Aim To analyse the reliability of HCV RNA measurements at key decision time points (weeks 4 and 12) and the predictive value of concordant or discordant assay results for SVR. Methods Weeks 4 and 12 samples of patients receiving telaprevir-containing triple therapy were initially tested with the AmpliPrep/COBAS-TaqMan_HCV-Test-v1.0 (limit of detection; LOD = 15IU/mL) and retested with the AmpliPrep/CO- BAS-TaqMan_HCV-Test-v2.0 (LOD = 15IU/mL) and the High_Pure/COBAS- TaqMan_HCV-Test-v2.0 (LOD = 20IU/mL). Results Concordance among the three test results in classifying samples as HCV RNA undetectable or detectable was only 55% at week 4, but 85% at week 12. Retesting of undetectable week 4 samples with the respective other assays revealed positive HCV RNA results in 32 50%. In 30%, HCV RNA was undetectable by all three tests at week 4 and all of these patients achieved SVR. In contrast, treatment failure occurred in 62% of patients with at least one detectable result, including cases with one or two other undetectable tests at week 4. Conclusions A single undetectable HCV RNA result at week 4 is not always associated with achieving SVR. Repeated testing in difficult-to-treat patients may identify those at risk for treatment failure. Aliment Pharmacol Ther 2014; 39: 85 92 85 doi:10.1111/apt.12544

B. Maasoumy et al. INTRODUCTION Hepatitis C virus (HCV) infection is an important cause for liver cirrhosis and hepatocellular carcinoma. 1, 2 About 150 million people worldwide have persistent HCV infection. 3 Significant improvements have been achieved in treating hepatitis C with the development of direct-acting anti-virals (DAAs). 4, 5 First-generation DAAs, the protease inhibitors (PIs), telaprevir (TVR) and boceprevir (BOC), were approved for treatment of chronic HCV genotype 1 infection in 2011. Combination therapy containing one of these PIs, pegylated interferon alpha (Peg-IFN) and ribavirin (RBV) (triple therapy) is now the new standard of care for treatment of chronic HCV genotype 1 infection in many countries. 6 Treatment duration, as well as stopping criteria of PI-containing triple therapy, strongly depends on the virological response during treatment, determined by measurements of on-treatment HCV RNA levels. Accurate measurement of HCV RNA during therapy has become extremely relevant, in particular, the differentiation between undetectable HCV RNA (target not detected; TND) and HCV RNA detectable/not quantifiable (below the lower limit of quantification; BLOQ). The opportunity to shorten treatment duration from 48 to 24/28 weeks in noncirrhotic treatment-na ıve or relapse patients is one of the most important advantages of triple therapy. However, patients are only eligible for a shorter therapy if HCV RNA is undetectable at weeks 4 and 12 of TVR and weeks 8 and 12 of BOC-including regimens. All patients with any detectable HCV RNA, even those with levels BLOQ, require a full treatment course of 48 weeks, potentially leading to additional adverse effects and costs. 6, 7 Recent subanalyses of the pivotal trials of BOC and TVR demonstrated that the predictive value for achieving SVR is greater for a TND result in contrast to an HCV RNA level BLOQ at week 4 of PI therapy. 8 However, despite this enormous impact of detectable borderline viraemia, there are many uncertainties concerning the interpretations of HCV RNA measurements. There is no strict cut-off until HCV RNA levels are reported as detectable. The so-called limit of detection (LOD) is defined as the lowest HCV RNA level that is still detected by the assay in >95% of the cases. However, this does not necessarily mean that an HCV RNA level <LOD is reported as undetectable. Samples may also produce detectable results at even lower HCV RNA levels than the defined LOD, just with a lower likelihood, which, if detected, will be reported as detectable/bloq. Furthermore, it is, so far, not well studied whether different assays that are used for HCV RNA measurements have an impact on the HCV RNA result. The COBAS TaqMan HCV Test for Use with the High Pure System, v2.0 () has been used in most clinical trials. However, several other assays are available for HCV RNA monitoring (Table 1), which may differ in their ability to detect borderline levels of HCV RNA. As a result, on-treatment decisions may not exclusively depend on the virological response, but may be influenced by the assay that is used or sometimes even simply on chance considering, i.e. samples with HCV RNA levels below the LOD. In this study, plasma samples of 41 patients drawn 4 and 12 weeks after triple therapy with telaprevir, which were initially assessed with the COBAS AmpliPrep/CO- BAS TaqMan HCV Test, v1.0 (), were retested with two other HCV RNA assays, the COBAS AmpliPrep/COBAS TaqMan HCV Test, v2.0 ( Table 1 CE-IVD HCV RNA viral load test analytical performance characteristics CE-IVD HCV test Viral load range (IU/mL) BLOQ reported result cut-off* Genotype 1 LOD (IU/mL) High pure TaqMan HCV Test, v2.0 25 391 000 000 <25 IU/mL, HCV RNA detected 20 CAP-TaqMan HCV Test, v1.0 43 69 000 000 <43 IU/mL, HCV RNA detected 15 CAP-TaqMan HCV Test, v2.0 15 100 000 000 <15 IU/mL, HCV RNA detected 15 Abbott RealTime HCV Test 12 100 000 000 <12 IU/mL, HCV RNA detected 12 VERSANT HCV RNA 3.0 Assay (bdna) 615 7 690 000 <615 IU/mL, HCV RNA detected 1000 * BLOQ, below the lower limit of quantification; test detects HCV RNA, but cannot quantify it. Results below the lower LOQ or LOD are reported as reflected by the BLOQ cut-off. LOD, limit of detection (below this level, HCV RNA can still be detected by assay), LOD is defined as the point at which HCV RNA is detected at >95%. Overall, LOD is most conservative estimate across genotype and matrix (plasma and serum). Assay LOD at 1000 IU/mL (95% detection rate), LLOQ and range determined per package insert, 615 IU/mL is <95% detection rate. 86 Aliment Pharmacol Ther 2014; 39: 85-92

Repeated HCV RNA testing predicts SVR ) and the TAQMAN Test, v2.0 (). Reliability of HCV RNA measurements at week 4 and 12 and predictive value of concordant or discordant test results were analysed. PATIENTS AND METHODS Patients. A total of 41 patients with chronic hepatitis C treated with pegylated interferon alpha 2a (88%) or 2b (12%), ribavirin and telaprevir were included in this study. Baseline characteristics of the study cohort are shown in table 2. There was no preselection of patients, as all consecutive individuals were included where sufficient plasma volume was available. No formal sample size calculation was performed, as this was a pilot study without reliable estimations for concordance and discordance between the different assays. For this reason, no statistical comparisons in detection rates between the assays were performed. RNA viral load measurements. The COBAS AmpliPrep/ COBAS TaqMan HCV Test, v1.0 () was used as part of routine clinical testing. Leftover EDTA lavender-top plasma from baseline, week 4 and week 12 samples were retested with the COBAS TaqMan HCV Test for Use with the High Pure System, v2.0 ( ) and the COBAS AmpliPrep/COBAS TaqMan HCV Test, v2.0 (). Results that were undetectable were reported target not detected (TND). Any detected results either below the lower limit of quantification (BLOQ) or below the limit of detection (BLOD) were reported <43 ( Table 2 Baseline characteristics of the study cohort Patients (n) 41 Age (mean s.d.) 53 10.4 Previous antiviral treatment Treatment-na ıve 5 (12%) Treatment-experienced 36 (88%) HCV genotype 1a 8 (20%) 1b 33 (80%) IL28B genotype CC 5 (12%) Non-CC 34 (83%) n/a 2 (5%) Cirrhotics 17 (41%) Platelets (/nl; mean s.d.) 164.8 70.7 Albumin (g/l; mean s.d.) 41.3 4.3 n, number; s.d., standard deviation; n/a, not available. ), <15 () or <25 () respectively. Quantifiable results (>LOQ) were reported as absolute numbers. Analysis of the virological outcome. An undetectable HCV RNA result 12 weeks after end of anti-viral treatment (SVR12) was chosen as the primary virological end point. One of the 41 selected patients died at week 16 of treatment and was therefore excluded from analysis. In one patient, there was no week 4 sample available and, in another patient, there was no week 12 sample. As a result, for each time point (weeks 4 and 12), data from 39 patients were available. Ethical Statement. This study was performed according to principals of good clinical practice as well as the declaration of Helsinki and approved by the local ethic committee of Hannover Medical School. RESULTS Concordance in week 4 samples HCV RNA was undetectable (TND) in 47.5 65% and detectable/not quantifiable (BLOQ) in 25 45% of the samples in the three assays. The was the assay that classified the most samples as HCV RNA TND (65%) and the, the assay with the most BLOQ results (45%) (Figure 1a). In almost all patients with quantifiable HCV RNA, viral load was still <100 IU/mL in all three tests except for one single patient, who even met the recommended stopping criteria according to HCV RNA result in the. Overall concordance through all three test results in classifying week 4 samples to be HCV RNA TND, BLOQ or >LOQ was 45% (Figure 1b). Overall concordance in either not detecting or detecting any HCV RNA was 55% at week 4 (Figure 1c). Concordance in week 12 samples At week 12, the majority of samples were tested HCV RNA TND ranging from 82.5% () to 90% () (Figure 1d). Only one patient had quantifiable HCV RNA at this stage. Concordance in classifying samples as HCV RNA detectable or undetectable was higher compared with week 4 (85%) (Figure 1e). Reliability of TND and BLOQ results at week 4 At week four, in 31 of the 40 selected samples, HCV RNA was reported as undetectable (TND) in at least one of the three test results. Less than half (42%) of these Aliment Pharmacol Ther 2014; 39: 85-92 87

B. Maasoumy et al. (a) 2.5% 10% 25% 15% 47.5% 45% 52.5% 65% 37.5% TND BLOQ >LOQ N = 40; 1 n/a Detectable HCV RNA Undetectable vs. Detectable/BLOQ vs. >LOQ (b) Concordance Discordance 45% 55% N = 40 20% 20% 15% 45% 17.5% 17.5% 22.5% (c) Undetectable vs. Detectable Concordance Discordance 55% 45% 20% 12.5% 12.5% N = 40 55% 20% 12.5% 12.5% (d) 5% 12.5% 5% 10% 5% 5% (e) Undetectable vs. Detectable Concordance 85% 15% 2.5% Discordance 5% 7.5% 82.5% N = 40; 1 n/a TND 85% BLOQ >LOQ Detectable HCV RNA 90% 7.5% N = 40 85% 2.5% 5% Figure 1 Overall assay concordance of results in patients on triple therapy. Distribution of HCV RNA viral load results by category ( Undetectable : Target not detected = TND; Detectable : Either Detected/Below the limit of quantification = BLOQ or Detected/Quantifiable HCV RNA = >LOQ) with the three different assays at week 4 of triple therapy (a). Concordance among the three assays in classifying week 4 samples as either HCV RNA TND, BLOQ and >LOQ (b) or simply as HCV RNA Undetectable or Detectable (c). Distribution of HCV RNA viral load results by category ( Undetectable : TND; Detectable : Either BLOQ or >LOQ) with the three different assays at week 12 of triple therapy (d). Concordance among the three assays in classifying week 12 samples as HCV RNA Undetectable or Detectable (e). samples were concordantly classified as TND through all three measurements. The highest level of discordance was documented between the and (38%) (Figure 2a). Conclusively, there was high chance that in samples, although classified as TND by one method, HCV RNA was detectable at least once in the same sample if tested with the other two assays. Likelihood ranged between 32% for samples with a TND result in the and 50% for samples previously classified as TND by the (Figure 2a). Almost one of four samples (23%) with undetectable HCV RNA in the, the assay used in the pivotal trials, had detectable HCV RNA in both other measurements. Similarly, in 69% of the samples once classified as detectable, but BLOQ, HCV RNA was not detected by one of the other two methods (Figure 2b). Reliability of TND and BLOQ results at week 12 At week 12, HCV RNA was not detected at least with one method in 37 samples. In contrast to the week 4 88 Aliment Pharmacol Ther 2014; 39: 85-92

Repeated HCV RNA testing predicts SVR (a) TND in (n = 21) (n = 26) (N = 19) Any (n = 31) 35% 21% 16% Detectable in 19% 29% 38% 10% 38% 1 other assay 50% 58% both other assays 23% 32% 5% (b) BLOQ in (n = 18) (n = 10) (N = 15) Any (n = 26) 40% 50% TND in 22% 30% 1 assay 56% 17% 50% 20% 40% 60% 73% 27% 69% 2 assays (c) TND in (n = 33) (n = 34) (N = 36) Any (n = 37) 6% 11% BLOQ/Detectable in 3% 3% 8% 1 other assay 6% 0% 3% 9% 0% 14% 16% both other assays 6% (d) BLOQ in (n = 5) (n = 4) (N = 2) Any (n = 7) 25% 40% TND in 80% 80% 40% 75% 1 assay 75% 50% 50% 50% 86% 2 assays 25% 50% Figure 2 Reliability of TND and BLOQ results at weeks 4 and 12 of triple therapy. Probability to be HCV RNA detected, although TND in one test at week 4 (a) and 12 (c) of triple therapy; Probability of a TND result by any assay in patients with an HCV RNA detected/bloq result in one test at week 4 (b) or week 12 (d) of triple therapy. samples, TND results were mostly confirmed in all three tests at week 12 (84%), Samples with HCV RNA TND showed detectable results in one of the respective other two tests in only 6 14% (Figure 2c). There were only 7 week 12 samples with detectable/ not quantifiable HCV RNA (BLOQ). The far majority (86%) of these BLOQ results were not confirmed, but reported TND in at least one of the respective other two tests (Figure 2d). Predictive value of week 4 HCV RNA results for the virological outcome Week 4 HCV RNA result had a high positive predictive value for the virological outcome in the here-analysed patient cohort (Figure 3a). In 12 (32%) of the 38 HCV genotype 1-infected patients with available week 4 and 12 samples (excluding one patient who died during treatment), HCV RNA was concordantly TND at week 4 in all three tests (ctnd). All of these patients achieved SVR12. In contrast, in the 18 patients with at least one TND result, but which was not confirmed by the two other measurements (discordant TND: dtnd), there were a significant number of virological breakthroughs and relapses. Overall, SVR rate was not higher than 50% in the dtnd group (Figure 3b). Chance to achieve SVR increased with each additional TND test result in the same week 4 sample, starting at 33% (n = 3/9) in patients with only one TND result, to 67% (n = 6/9) if HCV RNA was undetectable in two of three tests and finally up to the 100% (n = 12/12) in the ctnd group. On the other side, in the remaining eight patients, in whom HCV RNA was concordantly detectable through all three tests (notnd), only a single patient achieved SVR (13%) (Figure 3b). Of note, all of these patients still had HCV RNA results of less than 100 IU/mL in all three measurements, below the recommended stopping criteria, and 50% (n = 4) were even concordantly tested BLOQ. Any quantifiable HCV RNA (>LOQ) at week 4 in any single test had a negative predictive value of 100% (n = 4). Patients achieving SVR despite at least one detectable HCV RNA test result at week 4 (dtnd or notnd patients) were more likely to be younger (median age 45.5 yrs vs. 55 years). Furthermore, both patients with a detectable HCV RNA result at week 4, but carrying the favourable IL28B CC genotype, also achieved SVR. In contrast, all four patients infected with HCV GT1a and without ctnd experienced a virological treatment fail- Aliment Pharmacol Ther 2014; 39: 85-92 89

B. Maasoumy et al. (a) Percentage (%) 100 80 60 40 20 SVR12 according to week 4 85 76 83 28 30 47 0 0 0 0 TND BLOQ >LOQ n = 17/20 19/25 15/18 5/18 3/10 7/15 0/0 0/3 0/5 (b) SVR12 according to week 4 100 100 Percentage (%) 80 60 50 40 20 13 0 ctnd dtnd notnd n = 12/12 9/18 1/8 (c) Treatment failure according to week 12 HCV RNA 100 86 80 Percentage (%) 60 41 40 30 20 0 All ctnd Any detectable n = 15/37 9/30 6/7 (d) On treatment virological breakthrough according to week 12 HCV RNA result 80 71 Percentage (%) n = 60 40 20 0 22 10 All ctnd Any detectable 8/37 3/30 5/7 Figure 3 Treatment outcome according to HCV RNA results at key decision time points. Treatment outcome (achieving SVR12) according to HCV RNA viral load result at week 4 across assays (a) and in results combined (ctnd: concordant TND = TND results in all three assays; dtnd: discordant TND = one or two assays reported a TND result; notnd: no TND result) (b). HCV RNA viral load results at week 12 and subsequent treatment failure (c) and on-treatment virological breakthrough (d). ure, although three of them had at least one HCV RNA TND result at week 4. Predictive value of week 12 HCV RNA measurement for virological outcome Overall, there were 30 patients with week 12 samples in which HCV RNA level was concordantly undetectable in all the three assays. Not achieving ctnd at week 12 had a negative predictive value of 86%. There were nine patients with at least one detectable HCV RNA test result at week 12, including two meeting the recommended stopping criteria. Only a single patient (14%) with one or more detectable HCV RNA test results at week 12 achieved SVR (Figure 3c), while the frequency of virological breakthrough was 71% (n = 5/7) in those with at least one BLOQ result compared to only 10% (3/ 30) in those with ctnd at week 12 (Figure 3d). DISCUSSION Reliability of HCV RNA testing at key decision time points of triple therapy is absolutely crucial for optimal patient management, in particular, in differentiating undetectable from detectable results, including HCV RNA levels below the assays limit of quantification (BLOQ). Week 4 represents the most important time point for HCV RNA testing as it determines treatment duration in patients treated according to response-guided therapy (RGT). Furthermore, week 4 testing has been identified as the most valuable on-treatment SVR predictor. Here, we show that reliability of undetectable HCV RNA test results at week 4 of PI treatment may be limited and that retesting of samples can increase the positive predictive value for achieving SVR. The relatively high discordance of assay results at week 4 observed in our study may at least partly be explained by the way different assays classify samples with borderline HCV RNA. Truly HCV RNA-negative samples are unlikely to be misclassified as HCV RNA detectable, as specificities of modern HCV RNA PCR assays are extremely high. 9, 10 However, an undetectable (target not detected; TND) result does not necessarily mean that HCV RNA is really absent. While samples with HCV RNA levels above the assay s so-called limit of detection (LOD) give an HCV RNA signal in more than 95% of test results, things are entirely different if HCV RNA is still present, but yields detectable levels 90 Aliment Pharmacol Ther 2014; 39: 85-92

Repeated HCV RNA testing predicts SVR below the assay s LOD. In these cases where the actual HCV RNA titre is below the LOD, the results will fluctuate between a target not detected and a detectable result. For example, for the test with a LOD of 15 IU/mL, it has been shown that the test can detect HCV RNA levels of 2.5 IU/mL in approximately 50%. 11 Thus, discordance of replicate results should be expected (even with the same assay) in samples with low levels of HCV RNA. As a result, it may simply depend on chance whether a sample is classified as detectable/bloq or TND. Therefore, the likelihood to detect HCV RNA below the LOD increases by additional testing. It needs to be discussed whether different performances of the assays used here may have contributed to the observed discordance. In this study, we used three different versions of the COBAS Taqman assays: the COBAS AmpliPrep/COBAS TaqMan HCV Test, v1.0 (), the COBAS AmpliPrep/COBAS TaqMan HCV Test, v2.0 (), and the COBAS TaqMan HCV Test for Use with the High Pure System, v2.0 (). The is commonly used in drug trials, including the pivotal studies of TVR and BOC. In contrast, the and are widely used in routine HCV RNA testing. Analytical sensitivity of all three tests was recently compared and shown to have similar LODs (overlapping 95% confidence intervals). 12 In our study, the number of undetectable results was also comparable for all three assays. Thus, reliability of HCV RNA TND results at week 4 seems to be limited, no matter which assay is used. The second important finding of our study was that accuracy of SVR prediction can be improved by retesting of samples at key decision time points, in particular, at week 4. In a retrospective investigation of TVR and BOC phase 2 and 3 trials, an undetectable HCV RNA in contrast to a BLOQ week 4 result was an important SVR predictor. In the REALIZE study, patients with HCV RNA TND at week 4 had a 22% higher SVR rate than those with an HCV RNA BLOQ after TVR-containing triple therapy. Still, about one of five patients did not achieve a SVR despite a week 4 TND result determined by one testing. 8 Data of our study widely confirmed these findings as long as only one HCV RNA test was applied. Overall SVR rates were lower in our study, but a higher frequency of cirrhotics and/or previous nonresponders were included. However, the positive predictive value of an undetectable HCV RNA at week 4 still increased to 100% by retesting of samples, even in this difficult-to-treat cohort. In contrast, patients with virological failure despite a TND result at week 4 were all identified by additional testing. In addition to low viraemia at week 4 of therapy, other predictive factors for SVR may be considered to individualise therapy, in particular, if only one HCV RNA measurement is applied. In our study, positive SVR predictors were younger age and IL28B CC, while HCV genotype 1a was associated with treatment failure. However, value of such predictors under the circumstances of discordant TND and/or BLOQ week 4 HCV RNA results needs further prospective validation in larger cohorts. Our study has obvious limitations. We used three different assays for retesting of the samples, which may lead to a slight overestimation of discordance after repeated testing. On the other side, these also allowed demonstrating that no major differences occur among the three methods. Another limitation lies in the relatively small and heterogeneous study cohort as well as the retrospective study design. Therefore, data certainly need to be confirmed in larger, adequately powered and more homogenous cohorts in prospective studies before strict recommendations on treatment decisions and need for regular retesting can be established. Furthermore we here only compared three different versions of the Roche CTM assay. There are several other assays available and our data may not directly be transferred to these assays. Thus, future studies need to address the comparison of other assays and the significance of discordant results for the treatment outcome. However, findings presented here may still have important clinical implications to optimise management of PI-containing triple therapy. Recent studies in real-life cohorts demonstrated that safety and efficacy of Peg-IFN/RBV/PI triple therapy are limited in patients with liver cirrhosis. Serious adverse events of therapy, including severe infections and deaths, were reported in this specific patient cohort. 13, 14 In addition, SVR chances are also lower in cirrhotics, but, at the same time, these patients are in most urgent need for treatment. To avoid unnecessary side effects, it is therefore crucial to identify reliable SVR predictors. Here, confirming TND week 4 results may give important additional information to ensure a favourable risk/benefit ratio. Findings of this study results may also be relevant for patients considered as easy-to-treat. We here show that the decision for shortening of treatment is simply based on chance in a significant number of patients, as reliability of TND and BLOQ results is limited at week 4, the key decision time point of RGT. However, simple retesting in patients with otherwise beneficial response factors would most likely lead to overtreatment if all patients with discordant TND results would be treated for 48 weeks. Aliment Pharmacol Ther 2014; 39: 85-92 91

B. Maasoumy et al. Vermehren et al. presented preliminary data where, in a more easy-to-treat cohort, the majority of patients with a TND result in the, but a detectable HCV RNA result after retesting with the Abbott RealTime assay, still achieved SVR despite a shortening of treatment according to RGT criteria. 15 Thus, a more precise and reliable system is needed to define the optimal treatment duration. A possible approach would be to include several predictors in addition to RVR, such IL28B polymorphisms and HCV subgenotypes (1a vs. 1b). In conclusion, reliability of week 4 HCV RNA levels stated as TND or undetectable is limited if only a single HCV RNA measurement is performed. Patients likely to be truly HCV RNA-negative at week 4, defined by confirmed TND results, have a very high chance to achieve a SVR, even those with unfavourable baseline predictors. AUTHORSHIP Guarantor of the article: Benjamin Maasoumy. Author contributions: BM, BC, BB, SA and KL collected and analysed the data. BM, PH, MPM, MC and HW designed the research study. BM, MPM, MC and HW wrote the paper. All authors revised the manuscript critically. All authors approved the final version of the manuscript. ACKNOWLEDGEMENTS Declaration of personal interests: BM received lecture fees from Roche Diagnostics, Roche Pharma, Merck/MSD, travel support from Janssen-Cilaq and consulting fees from Abbott Molecular. BC, KL and SA are employees of Roche Diagnostics. BB has no conflict of interest. PH received consulting fees from Roche Diagnostics and lecture fees from Merck and Vertex. MPM received consulting fees from Roche, BMS, Gilead, Boehringer Ingelheim, Novartis, Tibotec, Vertex, GSK, Merck/MSD, grant/research support from Roche, BMS, Gilead, Boehringer Ingelheim, Novartis, Merck/MSD and lecture fees from Merck/MSD, Roche, BMS, Gilead. MC received lecture fees from Merck, Roche Pharma, Roche Diagnostics, Gilead, BMS, Novartis, consulting fees from Merck, Roche Pharma, Roche Diagnostics, Gilead, Novartis and grant support from Merck, Roche. HW received consulting fees from Transgene, MSD, Roche, Gilead, Abbott, BMS, Falk, grant/research support from MSD, Novartis, Gilead, Roche, Abbott and lecture fees form BMS, MSD, Novartis, ITF. Declaration of funding interests: This work was funded, in part, by the IFB-Tx a project of the BMBF (German Federal Ministry of Education and Research). Roche diagnostics provided the HCV RNA tests. REFERENCES 1. EASL Clinical Practice Guidelines. Management of hepatitis C virus infection. J Hepatol 2011; 55: 245 64. 2. Maasoumy B, Wedemeyer H. Natural history of acute and chronic hepatitis C. Best Pract Res Clin Gastroenterol 2012; 26: 401 12. 3. Hanafiah KM, Groeger J, Flaxman AD, Wiersma ST. Global epidemiology of hepatitis C virus infection: new estimates of age-specific antibody to hepatitis C virus seroprevalence. Hepatology 2012; 57: 1333 42. 4. Sarrazin C, Hezode C, Zeuzem S, Pawlotsky JM. Antiviral strategies in hepatitis C virus infection. J Hepatol 2012; 56(Suppl. 1): S88 100. 5. Dusheiko G, Wedemeyer H. New protease inhibitors and direct-acting antivirals for hepatitis C: interferon s long goodbye. Gut 2012; 61: 1647 52. 6. Ghany MG, Nelson DR, Strader DB, Thomas DL, Seeff LB. An update on treatment of genotype 1 chronic hepatitis C virus infection: 2011 practice guideline by the American Association for the Study of Liver Diseases. Hepatology 2011; 54: 1433 44. 7. Sarrazin C, Berg T, Cornberg M, et al. [Expert opinion on boceprevir- and telaprevir-based triple therapies of chronic hepatitis C]. Z Gastroenterol 2012; 50: 57 72. 8. Harrington PR, Zeng W, Naeger LK. Clinical relevance of detectable but not quantifiable hepatitis C virus RNA during boceprevir or telaprevir treatment. Hepatology 2012; 55: 1048 57. 9. Chevaliez S, Pawlotsky JM. How to use virological tools for optimal management of chronic hepatitis C. Liver Int 2009; 29(Suppl. 1): 9 14. 10. Chevaliez S, Pawlotsky JM. Virological techniques for the diagnosis and monitoring of hepatitis B and C. Ann Hepatol 2009; 8: 7 12. 11. Zitzer H, Heilek G, Truchon K, et al. Second-generation Cobas AmpliPrep/ Cobas TaqMan HCV quantitative test for viral load monitoring: a novel dualprobe assay design. J Clin Microbiol 2013; 51: 571 7. 12. Wedemeyer H, Kessler HH, Michel- Treil V, et al. The ExPECT study: an international multicenter clinical trial evaluating low HCV RNA viremia with the COBAS AmpliPrep/COBAS TaqMan HCV Test, v2.0 compared to assays used in routine clinical practise and in drug trials. Hepatology 2012; 56 (Suppl.): 676A 7A. 13. Maasoumy B, Port K, Markova AA, et al. Eligibility and safety of triple therapy for hepatitis C: lessons learned from the first experience in a real world setting. PLoS ONE 2013; 8: e55285. 14. Hezode C, Fontaine H, Dorival C, et al. Triple therapy in treatment-experienced patients with hcv-cirrhosis in a multicentre cohort of the french early access programme (anrs co20-cupic) nct01514890. J Hepatol 2013; 59: 434 41. 15. Vermehren J, Aghemo A, Pfeiffer KH, et al. Undetectable HCV-RNA in telaprevir-treated patients: low concordance between two highly sensitive real-time pcr assays. J Hepatol 2013; 58(Suppl.): 208. 92 Aliment Pharmacol Ther 2014; 39: 85-92