An OPTIMIZE Study Retrospective Analysis for Management of Telaprevir-Treated Hepatitis C Virus (HCV)-Infected Patients by Use of the Abbott RealTime HCV RNA Assay

Abstract

Protease inhibitor (PI)-based response-guided triple therapies for hepatitis C virus (HCV) infection are still widely used. Noncirrhotic treatment-naive and prior relapser patients receiving telaprevir-based treatment are eligible for shorter, 24-week total therapy if HCV RNA is undetectable at both weeks 4 and 12. In this study, the concordance in HCV RNA assessments between the Roche High Pure System/Cobas TaqMan and Abbott RealTime HCV RNA assays and the impacts of different HCV RNA cutoffs on treatment outcome were evaluated. A total of 2,629 samples from 663 HCV genotype 1 patients receiving telaprevir/pegylated interferon/ribavirin in OPTIMIZE were analyzed using the High Pure System and reanalyzed using Abbott RealTime (limits of detection, 15.1 IU/ml versus 8.3 IU/ml; limits of quantification, 25 IU/ml versus 12 IU/ml, respectively). Overall, good concordance was observed between the assays. Using undetectable HCV RNA at week 4, 34% of the patients would be eligible for shorter treatment duration with Abbott RealTime versus 72% with the High Pure System. However, using <12 IU/ml for Abbott RealTime, a similar proportion (74%) would be eligible. Of the patients receiving 24-week total therapy, 87% achieved a sustained virologic response with undetectable HCV RNA by the High Pure System or <12 IU/ml by Abbott RealTime; however, 92% of the patients with undetectable HCV RNA by Abbott RealTime achieved a sustained virologic response. Using undetectable HCV RNA as the cutoff, the more sensitive Abbott RealTime assay would identify fewer patients eligible for shorter treatment than the High Pure System. Our data confirm the <12-IU/ml cutoff, as previously established in other studies of the Abbott RealTime assay, to determine eligibility for shortened PI-based HCV treatment. (The study was registered with ClinicalTrials.gov under registration no. NCT01241760.)

INTRODUCTION

Hepatitis C virus (HCV) infection is a major public health problem and a leading cause of chronic liver disease. Analysis of HCV RNA levels in plasma or serum is critical for assessing the efficacy of antiviral therapy for chronic HCV infection (1, 2). The primary goal of therapy is achievement of a sustained virologic response (SVR), typically defined as HCV RNA levels of <25 IU/ml 12 (SVR12) or 24 weeks after completion of treatment.

The approval of direct-acting antiviral (DAA) drugs has led to a new standard of care for the treatment of HCV genotype 1 infection. DAAs in combination with pegylated interferon (Peg-IFN) and ribavirin (RBV) significantly improve SVR rates compared with Peg-IFN/RBV alone (3–7). For HCV treatment with these DAAs, a rapid HCV RNA decline does allow shorter treatment duration without significantly compromising efficacy. Therefore, the use of HCV RNA measurements that describe rapid and early viral response to guide treatment duration has become a key component of patient management.

Telaprevir is a potent and selective inhibitor of the NS3/4A HCV protease and is approved in combination with Peg-IFN and RBV for the treatment of chronic HCV genotype 1 in adult patients with compensated liver disease (8, 9). Telaprevir-based therapy consists of 12 weeks of telaprevir in combination with Peg-IFN/RBV, followed by 12 or 36 weeks of Peg-IFN/RBV treatment. In treatment-naive patients and prior relapsers without cirrhosis, treatment can be shortened from 48 to 24 weeks if the patients achieve an extended rapid virologic response (eRVR), which is defined as undetectable HCV RNA at weeks 4 and 12 of therapy. In pivotal telaprevir studies, HCV RNA was assessed using the manual Roche High Pure System/Cobas TaqMan (HPS) assay. Using this assay, a result of “HCV RNA < 25 IU/ml, target not detected”—also referred to as “undetectable” HCV RNA—is required at weeks 4 and 12 for a patient to be eligible for shorter treatment. For the Roche HPS assay, a result of “detectable but below the limit of quantification (LOQ)” for HCV RNA should not be considered equivalent to an “undetectable” HCV RNA result because of higher relapse rates after shortened treatment duration in phase II clinical studies (10).

Clinical studies to date have used the Roche HPS assay to determine response-guided treatment rules. A number of HCV RNA assays, like Roche Cobas TaqMan assays (HPS and Cobas AmpliPrep version 2 tests), Abbott RealTime (ART), Siemens kPCR, and the Qiagen Artus assay, with differing limits of detection (LOD) and LOQ thresholds, are commercially available internationally and used in routine clinical practice (11–15). Since the response-guided-decision rules were based on one specific HCV RNA assay, there is a need for evaluation of clinical HCV RNA cutoffs for optimal patient management.

Previous studies have demonstrated significant differences in clinical sensitivity between the Roche HPS and ART assays when using undetectable HCV RNA as a cutoff. A lower rate of undetectable HCV RNA at early time points was observed with ART than with HPS (16–18).

As newer DAAs are not approved in many countries worldwide and telaprevir-based triple therapy will still be used as the standard of care, the objective of this analysis was to evaluate the correlation and concordance in HCV RNA results between the HPS and ART assays using samples from chronic HCV genotype 1-infected, treatment-naive patients enrolled in the telaprevir phase III OPTIMIZE study. (The study was registered with ClinicalTrials.gov under registration no. NCT01241760.) The impact of response-guided-treatment cutoffs on treatment outcome using the different HCV RNA assays was explored.

MATERIALS AND METHODS

Study design and patient population.

The methodology for OPTIMIZE, a randomized, open-label, multicenter, phase III study in treatment-naive, HCV genotype 1-infected patients, has been described fully elsewhere (19). Patients (n = 740) received telaprevir, 1,125 mg twice daily or 750 mg every 8 h, in combination with Peg-IFN/RBV. At week 12, telaprevir dosing ended and the patients continued on standard Peg-IFN/RBV treatment. If a patient achieved an RVR, defined as undetectable HCV RNA at week 4 by HPS, the total treatment duration was 24 weeks; otherwise, the total treatment duration was 48 weeks. The study protocol was reviewed and approved by the appropriate review boards or institutional ethics committees and health authorities. The study was conducted in accordance with the Declaration of Helsinki, the good clinical practice guidelines, and applicable regulatory requirements.

Patient samples.

Stored plasma samples from patients who consented to exploratory research were used in this retrospective analysis. Samples from 663 of 740 individual patients (90%) were included in the analysis, including baseline (n = 536), week 4 (n = 586), week 12 (n = 497), end of treatment (n = 590), and follow-up week 12 (SVR12 assessment sample; n = 475) time points. The baseline characteristics and treatment outcomes of the analyzed patient population were similar to those of the overall study population (Table 1).

TABLE 1.

Baseline characteristics and treatment outcomes of the analyzed versus the overall patient populations

Parameter	Value for population:
	Analyzed (n = 663)	Overall (n = 740)
Baseline characteristics
Baseline log10 HCV RNA [mean IU/ml (SD)]	6.50 (0.65)	6.49 (0.66)
Baseline HCV RNA [n (%)]
<800,000 IU/ml	93 (14)	111 (15)
≥800,000 IU/ml	570 (86)	629 (85)
HCV genotype subtypea [n (%)]
1a	380 (58)	419 (57)
1b	279 (42)	317 (43)
IL28B genotype [n (%)]
CC	191 (29)	211 (29)
CT	368 (56)	414 (56)
TT	104 (16)	115 (16)
Stage of fibrosis [n (%)]
No cirrhosis (F0 to F3)g	564 (85)	636 (86)
Cirrhosis (F4)h	98 (15)	103 (14)
Treatment outcome
SVR12
Overall [n (%)]	486 (73)	544 (74)
Subjects with RVRb [n/Ni (%)]	366/422c (87)	434/506c (86)
Relapsed [n (%)]	38 (7)	42 (7)
On-treatment virologic failuree [n (%)]	67 (10)	74 (10)
Otherf [n (%)]	72 (11)	80 (11)

^aHCV genotype subtype determined using the NS3 assay; for four patients no subtype could be generated.

^bEligible for a shorter treatment duration of 24 weeks in the OPTIMIZE study.

^cIncluding two patients who received 48 weeks of treatment.

^dThe denominator for relapse is the number of patients with HCV RNA at <25 IU/ml at the planned end of treatment or a missing HCV RNA assessment at the planned end of treatment and HCV RNA at <25 IU/ml during follow-up from the planned end of treatment onward, with n = 531 and n = 593 for the analyzed and overall populations, respectively.

^ePatients who met a virologic stopping rule and/or viral breakthrough during treatment.

^fPatients with HCV RNA at ≥25 IU/ml at the planned end of treatment but who did not have viral breakthrough, patients with HCV RNA at <25 IU/ml at the planned end of treatment with a missing HCV RNA assessment during follow-up, and patients who were lost to follow-up.

^gMetavir score of F0 to F3 (no/mild/portal/bridging fibrosis).

^hMetavir score of F4 (cirrhosis).

ⁱN = total number of subjects with RVR.

HCV RNA assays.

In the OPTIMIZE study, plasma HCV RNA values were quantified using the HPS HCV test version 2.0 (LOQ, 25 IU/ml; LOD, 15.1 IU/ml, according to a greater than 95% detection rate for HCV genotype 1) (20). Analysis of selected samples by the ART assay (LOQ, 12 IU/ml; LOD, 8.3 IU/ml, according to a greater than 95% detection rate for HCV genotype 1) (21) was performed in a blinded way. Samples below the LOQ might be positive or negative for HCV RNA and were reported as “<LOQ, detected” or “<LOQ, target not detected,” which was also referred to as “undetectable HCV RNA” for both assays.

Data analysis.

Data were analyzed using SAS version 9.2. The concordance between the assays of HCV RNA levels above the LOD was graphically explored. To quantify the correlation, the Pearson and concordance correlation coefficients were calculated. Cross-tabulations of ART results versus HPS results were used to analyze the concordance/discordance around the LOD and LOQ. The data were categorized as follows: HCV RNA < LOQ, target not detected; HCV RNA < LOQ, detected; HCV RNA ≥ 12 IU/ml and < 25 IU/ml (only for ART); HCV RNA ≥ 25 IU/ml and < 1,000 IU/ml; and HCV RNA ≥ 1,000 IU/ml. Analyses were performed overall and by time point. The concordance in HCV RNA values between HPS and ART, below or above the LOD, was assessed graphically and characterized by Pearson and concordance correlation coefficient calculation. In addition, frequencies were calculated to study the relationships between different HPS and ART cutoffs at weeks 4 and 12, and the effect of these cutoffs on determining response-guided treatment was explored. Treatment outcome analyses were performed by treatment duration (24 or 48 weeks) based on the viral response measured with the HPS assay in the OPTIMIZE study. Patients were classified as achieving SVR12, relapse, on-treatment virologic failure, or “other.” “Other” patients included patients with HCV RNA at ≥25 IU/ml at the planned end of treatment but who did not have viral breakthrough, patients with HCV RNA at <25 IU/ml at the planned end of treatment with a missing HCV RNA assessment during follow-up, and patients who were lost to follow-up. The SVR12 rates determined in the OPTIMIZE study were used and were defined as plasma HCV RNA levels of <25 IU/ml (HPS) using the last available HCV RNA assessment 12 weeks after the last planned dose of study drugs (19). Comparison of SVR rates between populations was done using logistic regression modeling. Additionally, a non-virologic-failure-censored analysis, which excluded all patients who discontinued treatment due to reasons other than virologic failure, was performed.

RESULTS

Comparison of quantitative HCV RNA values.

HCV RNA levels were quantified in a total of 2,629 samples from 663 of the 740 patients included in OPTIMIZE, using the HPS and ART assays. Overall correlation was calculated using 691 samples from all time points with quantifiable results (results below the LOQ were not included). The corresponding Pearson correlation coefficient and concordance correlation were 0.99 and 0.94, respectively.

Comparison of viral responses.

Overall, 267 of 1,797 (15%) samples from different time points that had undetectable HCV RNA by HPS (HCV RNA < 25 IU/ml, target not detected) had detectable HCV RNA by ART. Conversely, 20 of 1,550 (1%) samples with undetectable HCV RNA by ART (HCV RNA < 12 IU/ml, target not detected) had detectable HCV RNA by HPS.

At the week 4 time point, 422 of 586 (72%) samples had undetectable HCV RNA by HPS, while 200 of 586 (34%) samples had undetectable HCV RNA by ART and 433 of 586 (74%) had HCV RNA levels of <12 IU/ml (i.e., <12 IU/ml, target not detected plus detected) by ART (Table 2 and Fig. 1A). Of the 422 samples that had undetectable HCV RNA by HPS, 191 (45%) were also undetectable by ART, while 180 (43%) had results of “HCV RNA < 12 IU/ml, detected” and 51 (12%) had HCV RNA concentrations of ≥12 IU/ml (Table 2). Of the last 51 samples, 30 had HCV RNA levels of ≥12 IU/ml and <25 IU/ml, while 21 had HCV RNA levels of >25 IU/ml by ART, ranging from 26 to 117 IU/ml. Of the 200 samples with undetectable HCV RNA by ART, 191 (96%) were also undetectable by HPS, and only 9 (5%) had detectable HCV RNA (all <25 IU/ml).

TABLE 2.

Concordance between HCV RNA results of the Roche HPS and Abbott ART assays at week 4

Viral response at wk 4 by Roche TaqMan (HPS) (IU/ml)	No. of samples with viral response at wk 4 by ART (IU/ml) of:
	<12, TND	<12, detected	≥12 to <25	≥25 to <1,000	≥1,000	Total
<25, TNDa	191	180	30	21	0	422
<25, detected	9	53	25	33	0	120
≥25, <1,000	0	0	3	21	0	24
≥1,000	0	0	0	6	14	20
Total	200	233	58	81	14	586

^aTND, target not detected.

FIG 1.

Viral response at week 4 (n = 586) and week 12 (n = 497) (A) and weeks 4 and 12 combined (n = 456) (B) for the HPS and ART assays.

The eRVR rates (i.e., undetectable HCV RNA at weeks 4 and 12) were 330 of 456 (72%) by HPS and 168 of 456 (37%) by ART. However, if an HCV RNA level of <12 IU/ml was used as the cutoff, the eRVR rate was 76% (345/456) by ART, similar to that of HPS (Fig. 1B). At week 12, the proportions of patients with undetectable HCV RNA by both assays were similar, with 94% (465/497) and 90% (447/497) for the HPS and ART assays, respectively. For 135 of 180 patients with undetectable HCV RNA by HPS and an “HCV RNA <12 IU/ml, detected” result by ART at week 4, data were also available at week 12. At week 12, 99% (134/135) of samples had undetectable HCV RNA (n = 130) or an “HCV RNA at <12 IU/ml, detected” result (n = 4), and 1% (n = 1) had HCV RNA at ≥12 IU/ml by ART.

There was no difference in SVR12 assessment (defined as an HCV RNA level below the LOQ using the last available HCV RNA assessment 12 weeks after the last planned dose of drugs) between the two assays. Similar proportions (89%; 422/475) of patients with HCV RNA levels below the LOQ were identified by both assays at 12 weeks follow-up.

In the OPTIMIZE study, 422 patients achieved RVR by HPS. Of these 422, 420 received treatment for 24 weeks, while two patients with RVR received 48 weeks of treatment and were excluded from the analysis. When HCV RNA at <12 IU/ml was used as the cutoff for ART, response rates were similar between the HPS and ART assays. Among the patients who received 24 weeks of treatment, 365 of 420 (87%; confidence interval [CI], 83.3 to 90.0%) with undetectable HCV RNA by HPS achieved SVR12 compared to 323 of 370 (87%; CI, 83.5 to 90.5%) with HCV RNA at <12 IU/ml at week 4 by ART. In the same population, 174 of 190 (92%; CI, 86.7 to 95.1%) patients with undetectable HCV RNA by ART achieved SVR12, while 149 of 180 (83%; CI, 76.5 to 88.0%) patients with HCV RNA at <12 IU/ml but still detectable by ART at week 4 achieved SVR12 (Fig. 2A).

FIG 2.

Treatment outcome in patients who received 24 or 48 weeks of treatment by viral response at week 4 overall (A) and excluding patients who discontinued treatment due to other reasons than virologic failure (B). Treatment outcomes were as determined in the OPTIMIZE study. *, two patients with undetectable HCV RNA by HPS (one patient with undetectable HCV RNA and one patient with HCV RNA at >12 IU/ml by ART) received 48 weeks of treatment and were not included in this analysis; ‡, one patient with an “HCV RNA < 25 IU/ml, detected” result by HPS and an “HCV RNA < 12 IU/ml, detected” result by ART received 24 weeks of treatment and was not included in this analysis. VF, virologic failure; TND, target not detected.

The difference in the SVR rates achieved by patients with undetectable HCV RNA by ART (92%) and patients with “HCV RNA < 12 IU/ml, detected” results (83%) was statistically significant (P = 0.0126; logistic regression) (Fig. 2A) and reflected a higher virologic failure rate (3% versus 0%) and an “other' category (7% versus 3%)” for patients with “HCV RNA < 12 IU/ml, detected” results and undetectable HCV RNA, respectively (Fig. 2A). Four (2%) patients (three HCV genotype 1a and one genotype 1b) with “HCV RNA < 12 IU/ml, detected” results by ART at week 4 who completed all treatment experienced viral breakthrough at the end of the 24-week treatment period. These patients had the IL28B genotype CT (n = 3) or CC (n = 1) and had no or minimal fibrosis (n = 2), bridging fibrosis (n = 1), or cirrhosis (n = 1). Thirteen (7%) of the patients with “HCV RNA < 12 IU/ml, detected” results were categorized as “other.” However, when a non-virologic-failure-censored analysis was performed, excluding all patients who discontinued treatment due to reasons other than virologic failure, only one (1%) patient was still categorized as “other,” and the SVR12 rates in the two groups were similar (P = 0.1462; logistic regression): 93% (CI, 88.2 to 96.6%) and 88% (CI, 81.7 to 93.2%) for patients with undetectable HCV RNA and “HCV RNA <12 IU/ml, detected” results by ART, respectively (Fig. 2B). The treatment outcome by viral response at week 4 in patients who received 48 weeks of treatment is also shown in Fig. 2.

DISCUSSION

Overall, good concordance between the HPS and ART assays was observed for HCV RNA measurements in samples from the OPTIMIZE study. In general, higher quantifiable HCV RNA results were measured using the HPS assay than using the ART assay, with a mean difference of 0.44 log₁₀ IU/ml. The difference in measured HCV RNA was smaller in the lower range of HCV RNA levels. A similar difference in measured HCV RNA levels between the two assays was previously reported (17, 22).

Discordant results between the assays were mostly observed in samples with values bordering the LOQ and LOD. As previously reported, the ART assay demonstrated higher sensitivity for detection of the HCV load in the low range than the HPS assay (16, 17, 23). Overall, for 15% of patients who had no viral RNA detected by HPS, RNA was still detectable by ART. Therefore, when undetectable HCV RNA at week 4 was used as the criterion for response-guided treatment, only 34% (200/586) of patients would be eligible for a shorter treatment using the ART assay in comparison with 72% (422/586) based on the HPS assay. When using the LOQ instead of undetectable HCV RNA as the criterion for the ART assay, a similar proportion of 74% (433/586) of patients would be eligible for a 24-week treatment duration. Furthermore, similar SVR12 rates (87%) were observed in patients with HCV RNA at <12 IU/ml by ART and those with undetectable HCV RNA by HPS. The SVR12 rate (83%) obtained in patients with “HCV RNA <12 IU/ml, detected” results by ART at week 4 was statistically different (P = 0.0126) from the SVR12 rate (92%) in patients with undetectable HCV RNA concentrations by ART at week 4. However, SVR12 rates in the two groups were similar (88% for HCV RNA at <12 IU/ml and 93% for undetectable HCV RNA; P = 0.1462) when a non-virologic-failure-censored analysis was performed, excluding all patients who discontinued treatment due to reasons other than virologic failure.

In this study, 51 patients who achieved RVR based on HPS had HCV RNA levels above the LOQ by ART, and these patients would not have been selected for a shorter treatment even if a cutoff of 12 IU/ml had been used for the ART assay for determination of treatment duration. The SVR rate in these patients, who received 24 weeks of treatment, was 84% (42/50; 1 patient received 48 weeks of treatment and was not included in the treatment outcome analyses). Of the eight patients who did not achieve SVR12, two patients (4%) relapsed, two patients (4%) experienced viral breakthrough during the Peg-IFN/RBV treatment phase, and four patients (8%) were classified as “other.” Two of the “other” patients discontinued treatment early, with HCV RNA levels of <25 IU/ml at the end of treatment and HCV RNA at >25 IU/ml thereafter; two patients had HCV RNA levels of <25 IU/ml at the end of treatment but no data available for an SVR12 assessment. This discrepancy in eligibility for shorter treatment is explained by the probabilities of detection of HCV RNA at the LOD of HCV RNA assays. In another study, Maasoumy et al. showed that rates of detectable and undetectable HCV RNA at week 4 of triple therapy showed considerable variation of approximately 30% just by measuring the same samples several times using different versions of the Roche Cobas AmpliPrep/Cobas HCV test and HPS assay (24). Braun et al. also demonstrated different rates of detectability by replicate testing of WHO standards at low viral loads using HPS or ART (25).

In contrast, 62 patients who did not achieve RVR based on HPS and were not eligible for shorter treatment had HCV RNA levels below the LOQ by ART at week 4. Data at week 12 were also available for 52 out of the 62 patients, and 98% of them maintained HCV RNA levels below the LOQ by ART (undetectable HCV RNA [n = 48] or “HCV RNA <12 IU/ml, detected” [n = 3]). Although these patients received 48 weeks of treatment (one patient received 24 weeks of treatment and was not included in the treatment outcome analysis), only 39 of 61 (64%) achieved SVR12. Two (3%) patients relapsed, 12 (20%) experienced virologic failure (4 patients met a virologic stopping rule after 32 or 40 weeks of treatment, and 8 patients had viral breakthrough during the first 24 weeks of treatment), and 8 (13%) discontinued therapy before or at week 24 and had a missing SVR12 assessment (Fig. 2A). When a non-virologic-failure-censored analysis was performed, i.e., excluding all patients who discontinued treatment for reasons other than virologic failure, the SVR12 rate increased to 76% (28/37) (Fig. 2B).

Although our study did not include repeated measurements of samples to address assay variability and the treatment duration decision was predefined based on the HPS assay used in the OPTIMIZE study, the data presented here are in line with the results of the comparison of the HPS and ART assays on samples from patients treated with other DAAs combined with Peg-IFN/RBV therapy (16, 17). These studies also suggested that 12 IU/ml is a more appropriate cutoff for response-guided treatment for a DAA-based triple therapy when the ART assay is used for HCV RNA quantification. Our results are also in line with the outcome of a similar study comparing the first- and second-generation Roche Cobas AmpliPrep/Cobas TaqMan HCV assays and the ART assay on samples from European and Japanese HCV genotype 1 patients treated with telaprevir-based therapy, respectively (23, 26). In these studies, high correlation between the two assays and greater sensitivity of the ART assay were also observed.

In conclusion, the higher rate of detectable HCV RNA with the ART assay implies that fewer patients might be eligible for a shorter duration of treatment if evaluated with that assay using the existing treatment guidelines, which were established using the less sensitive HPS assay. These data confirm a clinical cutoff of <12 IU/ml when using the ART assay to determine eligibility for a shortened 24 weeks of telaprevir-based therapy.