Article Text

Download PDFPDF

Insulin resistance and response to telaprevir plus peginterferon α and ribavirin in treatment-naïve patients infected with HCV genotype 1
  1. Lawrence Serfaty1,
  2. Xavier Forns2,
  3. Tobias Goeser3,
  4. Peter Ferenci4,
  5. Frederik Nevens5,
  6. Giampiero Carosi6,
  7. Joost P Drenth7,
  8. Isabelle Lonjon-Domanec8,
  9. Ralph DeMasi9,
  10. Gaston Picchio9,
  11. Maria Beumont10,
  12. Patrick Marcellin11
  1. 1Service d'Hépatologie and INSERM UMRS 938, Hôpital Saint Antoine, Université Pierre&Marie Curie, Paris, France
  2. 2Liver Unit, Hospital Clinic, IDIBAPS, Ciberehd, University of Barcelona, Barcelona, Spain
  3. 3Klinikum der Universität zu Köln, Köln, Germany
  4. 4Department of Internal Medicine 3, Medical University of Vienna, Vienna, Austria
  5. 5Department of Hepatology, University Hospital Gasthuisberg, Leuven, Belgium
  6. 6Clinic of Infectious and Tropical Diseases, University of Brescia, Brescia, Italy
  7. 7Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands
  8. 8Janssen Pharmaceuticals, Paris, France
  9. 9Tibotec Inc, Titusville, New Jersey, USA
  10. 10Tibotec BVBA, Beerse, Belgium
  11. 11Service d'Hépatologie and INSERM CRB3, Hôpital Beaujon, APHP, Université Paris-Diderot, Clichy, France
  1. Correspondence to Dr Lawrence Serfaty, Hôpital St Antoine, 184 Rue du Faubourg Saint-Antoine, 75571 Paris Cedex 12, France; lawrence.serfaty{at}sat.aphp.fr

Abstract

Objective Insulin resistance is a predictor of poor response to peginterferon/ribavirin in patients infected with the chronic hepatitis C virus (HCV). There are no data on direct-acting antivirals. This exploratory analysis assessed the effect of metabolic factors and insulin resistance, measured by homoeostatic model assessment (HOMA), on virological response to telaprevir in Study C208.

Design Overall, 161 HCV genotype 1-infected, treatment-naïve patients received 12 weeks of telaprevir plus peginterferon/ribavirin, then 12/36 weeks of peginterferon/ribavirin depending on on-treatment response criteria. The prognostic significance of several factors, including HOMA-insulin resistance (HOMA-IR), on virological response at weeks 4 and 12, end of treatment and 24 weeks after treatment was explored by multiple regression analysis.

Results Baseline HOMA-IR data were available for 147 patients; baseline characteristics were consistent with the overall population. Baseline HOMA-IR <2, 2–4 and >4 was seen in 54%, 30% and 16% of patients, respectively. Neither response rates (any time point) nor week 4 viral load decline were significantly influenced by baseline HOMA-IR. In multivariate analyses, fibrosis stage and low-density lipoprotein cholesterol level were predictive of sustained virological response (OR 0.47 and 1.02, respectively). After the end of treatment, HOMA-IR was significantly lower in patients with sustained virological response than in those without (0.61 vs 1.34 for relapsers and 1.15 for non-responders; p<0.05).

Conclusion In this study, baseline HOMA-IR was not predictive of virological response to telaprevir in HCV genotype 1-infected, treatment-naïve patients, while sustained virological response was associated with improved HOMA-IR. These results suggest that metabolic factors and insulin resistance do not have a significant effect on telaprevir-based treatment efficacy.

  • Vx-950
  • homeostatis model assessment
  • LDL-cholesterol
  • fibrosis
  • direct-acting antiviral agent
  • hepatitis C
  • liver transplantation
  • chronic hepatitis
  • chronic liver disease
  • hepatic encephalopathy
  • Wilson's disease
  • haemochromatosis
  • haemodynamics in cirrhosis
  • hepatocellular carcinoma
  • hepatitis B
  • hepatic haemodynamics
  • portal hypertension
  • cirrhosis
  • liver failure
  • liver transplantation
  • chronic viral hepatitis
  • autoimmune hepatitis
  • hepatitis D
  • cirrhosis
View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known about this subject?

  • Pegylated interferon α combined with ribavirin (peginterferon/ribavirin) is the standard treatment of chronic hepatitis C virus (HCV) infection.

  • Insulin resistance (IR) at baseline is associated with poor response to peginterferon/ribavirin treatment, and homoeostatic model assessment (HOMA)-IR is an independent predictor.

  • HCV itself may promote IR, and successful therapy results in improvements in IR.

  • No data have been reported on IR and triple therapy with direct-acting antiviral drugs.

What are the new findings?

  • In HCV genotype 1-infected treatment-naïve patients, baseline HOMA-IR was not predictive of response to treatment with the HCV protease inhibitor, telaprevir, in combination with peginterferon/ribavirin.

  • Sustained virological response (SVR) was associated with improved HOMA-IR, with significantly lower HOMA-IR in patients who achieved an SVR than in those who did not.

  • In multivariate analyses, only low fibrosis stage and high low-density lipoprotein cholesterol were predictive of SVR.

How might it impact on clinical practice in the foreseeable future?

  • Our results suggest that, in contrast with peginterferon/ribavirin, telaprevir may overcome the potential negative effect of metabolic factors and IR on treatment efficacy.

  • With the introduction of telaprevir in clinical practice, this finding may provide a potential improvement in antiviral efficacy in subgroups of patients with IR and other negative metabolic factors.

Introduction

Pegylated interferon α2a or α2b combined with ribavirin (peginterferon/ribavirin) has been the standard treatment of chronic infection with the hepatitis C virus (HCV) for several years. Metabolic disorders are known to play a major role in the response to peginterferon/ribavirin treatment. For example, obesity, steatosis and insulin resistance (IR) at baseline are associated with poor response.1–3 Numerous studies have shown that homoeostatic model assessment (HOMA), which is a simple way to measure IR, independently predicts response to peginterferon/ribavirin, irrespective of HCV genotype.3–8 Epidemiological data indicate a strong risk of developing IR, and ultimately overt diabetes mellitus, in patients with chronic HCV infection.9–12 Besides metabolic disorders such as obesity and/or metabolic syndrome, clinical and experimental data have suggested that HCV itself may promote IR.9–12 In keeping with this finding, several studies have shown that successful therapy results in improvements in IR, and decreases the risk of diabetes.3 ,13

To our knowledge, no data have been reported so far regarding IR and triple therapy with direct-acting antivirals (DAAs). This exploratory analysis aimed to assess whether telaprevir (a protease inhibitor acting on the NS3/NS4 region of the HCV polyprotein) added to peginterferon/ribavirin in treatment-naïve, genotype 1-infected patients (1) affects the effect of metabolic disorders on response to treatment and (2) restores insulin sensitivity. To achieve this, HOMA-IR and other data collected during the phase II C208 clinical trial were analysed. In this study,14 HCV-infected patients received telaprevir plus peginterferon/ribavirin, and sustained virological response (SVR) rates ranged from 81% to 85% across treatment arms by intent-to-treat (ITT) analysis. No statistically significant differences in SVR rates were observed between treatment groups. Adverse events were consistent with those observed in other telaprevir clinical trials.15 ,16

Patients and methods

Full methodology for this clinical trial is reported in detail elsewhere14; the main points and methodology specific to this analysis are briefly summarised.

Study population and clinical trial design

A total of 160 treatment-naïve patients aged 18–65 years with chronic HCV genotype 1 infection were included in this exploratory, prospective, multicentre, randomised, open-label, phase II clinical trial, conducted at 30 centres in Austria, Belgium, France, Germany, Italy, Spain and the Netherlands. Liver fibrosis status had to be documented with a liver biopsy or transient elastography within 18 months. Patients were excluded if they had documented cirrhosis, any evidence of another significant liver disease (including hepatitis B and hepatocellular carcinoma) or HIV co-infection. History of diabetes mellitus and concomitant administration of oral antidiabetic drugs were not exclusion criteria. Informed consent was obtained in writing from each patient, and the study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki. The trial was registered with ClinicalTrials.gov (NCT00528528).

Patients who met the inclusion criteria were randomly assigned to one of four treatment groups: (1) telaprevir 750 mg every 8 h (q8h) plus peginterferon α2a/ribavirin (q8h; α2a); (2) telaprevir 750 mg q8h plus peginterferon α2b/ribavirin (q8h; α2b); (3) telaprevir 1125 mg every 12 h (q12h) plus peginterferon α2a/ribavirin (q12h; α2a); or (4) telaprevir 1125 mg q12h plus peginterferon α2b/ribavirin (q12h; α2b). Peginterferon α2a was administered at 180 μg/week with ribavirin at 1000–1200 mg/day; peginterferon α2b was administered at 1.5 μg/kg/week with ribavirin at 800–1200 mg/day.

Patients received telaprevir plus peginterferon/ribavirin for 12 weeks followed by peginterferon/ribavirin alone for 12 or 36 weeks, based on on-treatment virological response criteria. Patients with undetectable plasma HCV RNA from week 4 until week 20 were assigned to receive a total of 24 weeks of treatment. Other patients were assigned to receive 48 weeks of treatment overall. Treatment modifications were prespecified according to the following rules: telaprevir was discontinued if HCV RNA was >1000 IU/ml at weeks 4, 6 or 8, in which case peginterferon/ribavirin was continued until week 48. Peginterferon/ribavirin was discontinued if HCV RNA levels had not decreased by >2 log10 from baseline by week 12, or if HCV RNA was confirmed detectable at weeks 24 or 36.

Study assessments

Clinical assessments

The following data were recorded at baseline: history of diabetes or high blood pressure, body mass index (BMI) and blood pressure. BMI was calculated as body weight in kg divided by height squared in m (kg/m2). A BMI below 25 kg/m2 was considered normal, 25–30 kg/m2 overweight, and >30 kg/m2 obese.

Biochemical assessments

At baseline, venous blood samples were collected after a fast of at least 8 h to assess haematology, coagulation and serum chemistry. Serum levels of the following relevant variables were determined: alanine aminotransferase, aspartate aminotransferase, γ-glutamyltranspeptidase, triglycerides, total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol and plasma glucose concentration. Blood samples for HOMA-IR analysis were collected during treatment at weeks 8, 20 and 24 weeks after the end of therapy (follow-up 24 weeks (FU24), SVR assessment time point). Serum insulin was measured on frozen fasting serum samples (stored at –20°C) by simultaneous one-step monoclonal immunoenzymatic assay (Access Ultrasensitive Insulin assay; Covance, Beerse, Belgium). The IR index was calculated on the basis of fasting values of glycaemia and insulinaemia, according to the HOMA method.17 The formula for the HOMA model is as follows: HOMA-IR = fasting insulinaemia (μU/ml) × fasting glycaemia (mmol/l)/22.5. IR determined by this method correlates closely with more complex techniques, such as the euglycaemic clamp method, in both diabetic and non-diabetic individuals.18

Steatosis assessment

Steatosis was assessed in patients with liver biopsy at baseline. Steatosis was expressed as the percentage of affected hepatocytes and graded as absent or present when affecting >5% or <5% of hepatocytes, respectively.

Liver fibrosis assessments

Fibrosis stage was assessed by liver biopsy or transient elastography (FibroScan, Echosens, Paris, France) and graded using the Metavir score. In patients investigated by transient elastography, fibrosis stage was assessed according to liver stiffness.19 Patients with liver stiffness >9 kPa were excluded.

Virological assessments

Blood samples were collected for HCV RNA analysis at the following time points: before dose on day 1, during treatment on days 1, 2, 3, 4 and 8, and at weeks 2, 3, 4, 6, 8, 10, 12, 14, 16, 20, 24, 36 and 48, as described previously.14 The COBAS TaqMan HCV assay (version 2.0; linear dynamic range 25 IU/ml to 3.91×108 IU/ml) was used to quantify plasma HCV RNA. Results were reported as either HCV RNA <25 IU/ml ‘detectable’ if the calculated IU/ml fell below the lower limit of quantification of the assay, or ‘undetectable’ if no HCV signal was detected.14 Having undetectable HCV RNA at week 4 was defined as a rapid virological response (RVR), and having undetectable HCV RNA 24 weeks after the end of treatment was defined as an SVR. Viral breakthrough was defined as either an HCV RNA value >100 IU/ml if previously <25 IU/ml, or an increase in HCV RNA >1 log10 compared with the nadir. Two consecutive samples taken ≤4 weeks apart were used to confirm viral breakthrough. Patients with undetectable HCV RNA at the end of treatment, but with subsequent positive HCV RNA were considered to be relapsers.

Statistical analysis

A draft statistical analysis plan outline was developed before initiation of these post hoc analyses. The study was not powered to show significance across these analyses. Analyses were pooled over treatment arms because of the similarity of response across dosing regimen and interferon groups. Multiple logistic regression analyses were used to assess the prognostic significance of baseline factors on baseline HOMA-IR, where the latter was dichotomised into low (≤2) and high (>2) groups. Sensitivity analyses were also performed using different cut-off points for baseline HOMA-IR (> median of 1.84, >3 and >4), and also using the log-transformed continuous value, as well as restrictions to patients with lower baseline viral load (< median value of 6.5 log10 HCV RNA IU/ml). Variables were selected for inclusion in the regression analyses based on the investigators' clinical judgement as to the relevance of the factor on response, and also on the observed frequency distribution; dichotomous variables with at least 10 observations in each category were included, which precluded statistical evaluation of race and clinical diabetes. Given the relatively larger number of covariates evaluated, the predictor variables were separated into clinically distinct domains to reduce the likelihood of false positive findings. For analyses of HOMA-IR and virological response, the primary variable of interest was baseline HOMA-IR. Ancillary (controlling) variables included: demographics (age, sex, BMI and history of hypertension); disease characteristics (fibrosis stage, presence of steatosis and baseline HCV RNA); liver function tests; lipid variables; and peginterferon treatment group. Variables used in calculation of the HOMA-IR score (ie, glucose and insulin) were not directly modelled when factors associated with HOMA-IR were being evaluated, because these were already included through the HOMA-IR score itself; however, these were included in the univariate logistic regression analyses for RVR and SVR.

In the first stage of logistic regression modelling, univariate regression analysis was used to assess the strength of association of a single variable on the response variable; those factors identified as significant were then included in the second stage, which involved fitting an initial multivariable model. After this initial fitted multivariable model, in the third and final stage, the non-significant terms from the initial model were omitted and the final multivariable regression model was presented, in which the retained terms were statistically significant at the 5% significance level. For evaluating SVR, the final model was then used to derive estimates of SVR for dichotomous splits of each predictor variable.

Fisher and Wilcoxon tests were used to test binary and continuous factors, respectively, across defined groups. If a patient was diabetic, HOMA-IR was set to >4 in categorical analyses, and imputed as 4.01 for continuous analyses. Log HOMA-IR (natural logarithm; LN) was used to normalise data distribution for some correlation analyses, where it was treated as a continuous variable. Methods for calculating virological responses have been reported previously.14

Results

Demographics and baseline characteristics

A subset of 147 patients of the ITT patient population (n=161) who had available baseline HOMA-IR measurements were included in this subanalysis. Baseline characteristics and demographics of patients with available baseline HOMA-IR data were similar to those of the total population (table 1). Of the 147 patients with available HOMA-IR data, fibrosis stage was assessed by liver biopsy in 92 patients and transient elastography in 55 patients. Four patients in the total ITT population had cirrhosis and were enrolled in the trial in error but allowed to continue, including one patient from the subset of patients with available HOMA-IR data who had cirrhosis.14 Metabolic variables in the study population are shown in table 2. HOMA-IR was <2 in 54% of patients, 2–4 in 30% of patients, and >4 in 16% of patients. Steatosis data were available in 73/92 patients with liver biopsy. Baseline characteristics of patients with or without steatosis assessment were not significantly different (data not shown). Accordingly, steatosis was present in 41% of patients. At baseline, a significant correlation between LDL cholesterol and HCV viral load was identified (r=0.25, p=0.0032).

Table 1

Baseline characteristics of patients with an available HOMA-IR assessment versus the overall ITT population

Table 2

Baseline metabolic parameters in the analysis population

Factors associated with baseline HOMA-IR

Many factors, several of which were intercorrelated, appeared to be associated with baseline HOMA-IR in the univariate analyses. Without steatosis included in the model (since data were only available in half of the patients), BMI (p<0.0001), baseline fibrosis stage (p=0.0240) and triglycerides (p=0.0430) were significantly associated with baseline HOMA-IR in multivariate analyses (table 3). Without the fibrosis stage included in the model, baseline HCV viral load was prognostic, but when both were included, the fibrosis stage had a higher prognostic value and was retained (p=0.06 for baseline HCV viral load). The greatest magnitude of effect was seen for the association with baseline fibrosis stage (OR 1.67, 95% CI 1.07 to 2.61) and BMI (OR 1.28, 95% CI 1.13 to 1.44). The magnitude of effect for triglycerides was small (OR 1.01, 95% CI 1.00 to 1.02). When steatosis was included in the model, independent factors associated with baseline HOMA-IR were steatosis (OR 6.66, 95% CI 2.06 to 21.27; p=0.003) and BMI (OR 1.22, 95% CI 1.005 to 1.48; p=0.001).

Table 3

Factors associated with baseline HOMA-IR >2 in univariate and multivariate analysis

Baseline HOMA-IR and virological response to treatment

Baseline LN HOMA-IR was not associated with achieving RVR or SVR (figure 1). When grouped together, patients who reached RVR showed a similar distribution of baseline LN HOMA-IR to non-RVR patients. A similar pattern was seen for those achieving SVR versus non-SVR patients.

Figure 1

Log baseline HOMA-IR distribution according to (A) RVR and (B) SVR. HOMA-IR, homoeostatic model assessment-insulin resistance; LN, natural logarithm; RVR, rapid virological response; SVR, sustained virological response.

The proportion of patients with undetectable HCV RNA (<10 IU/ml) at week 4, week 12, end of treatment and FU24 (SVR) was similar across baseline HOMA-IR categories (figure 2). Although the proportion of patients achieving SVR was slightly higher among those with baseline HOMA-IR <2, multiple regression analyses showed that baseline HOMA-IR did not correlate with SVR status (table 4). This was unaffected by the HOMA-IR cut-off (> median, >2, >3, or >4) used in the logistic regression model, use of the continuous value, or restriction of the analysis to patients with lower baseline viral loads (< median value of 6.5 log10 HCV RNA IU/ml). In this multivariate analysis, only fibrosis stage and baseline LDL cholesterol were predictive of SVR (p=0.0050; OR 0.47, 95% CI 0.28 to 0.80 for fibrosis stage; p=0.0122; OR 1.02, 95% CI 1.00 to 1.04 for baseline LDL cholesterol). BMI, triglycerides, baseline log10 HCV RNA, steatosis and having received peginterferon α2a were not significantly associated with SVR.

Figure 2

Virological response according to baseline HOMA-IR category. HOMA-IR, homoeostatic model assessment-insulin resistance; HCV, hepatitis C virus; SVR, sustained virological response.

Table 4

Factors associated with RVR and SVR in univariate and multivariate analysis

SVR was observed in 49 of 65 (75%) patients with baseline LDL ≤100 mg/dl, compared with 67 of 76 (88%) of patients with baseline LDL >100 mg/dl (p<0.05). SVR rates by fibrosis stage were 85% (93 of 109 patients) for fibrosis score 0–2 versus 74% (28 of 38 patients) for fibrosis score 3–4.

There was no distinct association between decline in log10 HCV RNA at week 4 and baseline HOMA-IR category (figure 3). Multiple regression analyses showed that baseline HOMA-IR did not correlate with RVR status (table 4). In this multivariate analysis, only baseline log10 HCV RNA was significantly associated with achieving RVR (p=0.00173; OR 0.44, 95% CI 0.22 to 0.86).

Figure 3

Viral load decline at week 4 according to baseline HOMA-IR category. HOMA-IR, homoeostatic model assessment-insulin resistance.

HOMA-IR course and virological response during anti-HCV treatment

The course of HOMA-IR during the study is displayed in figure 4. LN HOMA-IR decreased by a mean of 0.03 from baseline to FU24 in patients who achieved SVR. Conversely, LN HOMA-IR increased from baseline to FU24 in patients with viral relapse (0.86). In non-responders, there was no clear pattern in changes in HOMA-IR over the course of the study, although, at FU24, LN HOMA-IR was 0.39 above baseline. At FU24, mean LN HOMA-IR was significantly lower in patients who achieved SVR (0.61) than in relapsers (1.34) and non-responders (1.15; p<0.05 for comparison between SVR and non-SVR patients). An improvement (ie, numerical decrease) in HOMA-IR at FU24 relative to baseline was observed in 63 (52%) of the 121 SVR patients, compared with seven (27%) of the 26 non-SVR patients (p<0.05).

Figure 4

Association between log HOMA-IR and virological response during telaprevir-based treatment. HOMA-IR, homoeostatic model assessment-insulin resistance; LN, natural logarithm; SVR, sustained virological response; T0, baseline; W, week; FU24, follow-up week 24.

Discussion

This study in HCV genotype 1 infected patients treated with telaprevir-based regimens showed that IR assessed by HOMA-IR had no substantial effect on response to treatment, suggesting that telaprevir may overcome the potentially negative effect of IR on treatment efficacy. Treatment-naïve patients received triple therapy with telaprevir plus peginterferon/ribavirin for 12 weeks, then peginterferon/ribavirin for 12 or 36 weeks according to their on-treatment response. The rate of SVR was very high (>80%) irrespective of the type of peginterferon used or frequency of telaprevir administration. Neither RVR nor SVR was significantly influenced by the baseline HOMA-IR level. Among other metabolic factors, only LDL cholesterol was predictive of SVR. Extensive fibrosis was also significantly associated with poorer response in univariate and multivariate analysis. In accordance with previous studies, HOMA-IR level was related to steatosis and to a lesser degree to fibrosis stage and baseline HCV viral load, and was more often improved in patients with SVR.3 ,9 ,20

Baseline predictive factors of response to telaprevir-based triple therapy are not fully established. To our knowledge, this is the first study to address the effect of IR on the efficacy of a DAA in combination with peginterferon plus ribavirin. Given the negative effect of IR on virological response in patients treated with peginterferon/ribavirin,3–8 we retrospectively investigated the relationship between baseline HOMA-IR, as well as metabolic factors, and SVR in the C208 study. As expected, the prevalence of metabolic disorders and IR assessed by HOMA-IR was high, and we observed an association between HOMA-IR and viral load, suggesting HCV-related IR.7

In genotype 1-infected patients treated with peginterferon/ribavirin, Romero-Gomez et al have shown that SVR rates decreased from 60% for patients with HOMA-IR <2 to 40% and 20% for patients with HOMA-IR between 2 and 4, and HOMA-IR >4, respectively.3 Similar results have been found in patients infected with genotype 2 or 3, with a sixfold higher risk of non-response in patients with HOMA-IR >2.8 In our study, the lack of a control group showing HOMA-IR as a predictor of response to standard peginterferon/ribavirin treatment represents a limitation of our results. However, the relationship between HOMA and SVR in patients treated with standard peginterferon/ribavirin therapy has been recently confirmed by two meta-analyses.21 ,22 In one of these metaanalysis involving 2732 HCV treatment-naïve patients treated with peginterferon/ribavirin, we have shown that IR negatively influences the rate of SVR, with a mean decrease of 20% regardless of genotype.22 When considering HOMA>2 to define IR, a 22% decrease in SVR rates was still found, demonstrating that this cut-off is a robust negative predictive factor for SVR. In the present study, although 46% of patients had baseline HOMA-IR of >2, we did not find any relationship between HOMA-IR and response to telaprevir-based treatment. The tendency of higher SVR rate among patients with HOMA-IR <2 may be explained by their lower fibrosis stage. In multivariate analysis, baseline viral load was the only independent predictor of RVR, and fibrosis stage and baseline LDL cholesterol level were the only independent predictors of SVR. Fibrosis stage was assessed by liver biopsy or according to liver stiffness, which is now a validated method in patients with HCV.19 The relationship between fibrosis stage and virological response has been confirmed in telaprevir phase III studies, although to a lesser degree than in standard peginterferon/ribavirin treatment.23 Baseline higher LDL cholesterol and lower triglyceride concentrations have been identified as predictors of higher SVR in patients treated with peginterferon/ribavirin.24 One mechanism may be the relationship between LDL cholesterol level and interleukin (IL)28B polymorphism that is implicated in response to interferon in HCV-infected patients, as recently shown.25 Indeed, LDL cholesterol was significantly higher in patients with a favourable IL28B genotype (CC) compared with others (CT or TT), suggesting that LDL cholesterol may be a marker of host endogenous interferon response to HCV. Therefore we cannot exclude the possibility that the effect of LDL cholesterol on SVR may be the result of IL28B distribution in our study population. The significant correlation between LDL cholesterol and HCV viral load at baseline in our patients (r=0.25, p=0.0032) probably suggests more complex relationships with response to treatment and deserves further investigation. Other metabolic factors such as BMI, diabetes and hypertension were not significantly associated with response to telaprevir-based treatment in univariate or multivariate analysis in our study. Steatosis has been identified as a predictor of response in patients treated with peginterferon/ribavirin.2 We investigated steatosis in 73 patients with available data. As expected, steatosis was closely related to HOMA index at baseline.20 However, similarly to HOMA-IR, steatosis was not a predictor of either RVR or SVR in univariate or multivariate analyses.

The lack of association between IR and response to treatment in our study may have several explanations. Lack of power cannot be ruled out, as only 147 patients had available baseline HOMA-IR measurements, while the rate of SVR was high (>80%). However, despite this high rate of SVR, several other known predictive factors were identified in univariate and multivariate analysis. Moreover, a post hoc power calculation indicates that our sample size was sufficient to detect a 20% difference in SVR rate between patients with HOMA-IR <2 and patients with HOMA-IR >2. Characteristics of our study population such as young age (mean 44 years) and no severe liver disease in the majority of cases may also partly explain our results. Indeed, it has been suggested that the HOMA-IR prediction may be related to the severity of liver disease.21 Improvement in IR secondary to the reduction in HCV virus as a result of the direct antiviral effect of telaprevir may be another explanation. The development of IR in patients with HCV infection is thought to be due to a combination of both host- and virus-mediated pathways.12 Besides metabolic risk factors such as obesity and diabetes, experimental studies have shown that HCV proteins may directly induce IR by impairing the insulin signalling pathway.26 In addition, it has recently been shown that a 14-day course of DAA monotherapy was able to induce HOMA-IR improvement concomitantly with a decrease in viral load, confirming the existence of HCV-related IR.27 Although this relationship was less evident in the setting of telaprevir-based triple therapy, the higher proportion of SVR patients with HOMA-IR improvement may partly explain our results. Validation in a larger number of patients treated with telaprevir is needed before definitive conclusions can be drawn.

Although it is well established that IR negatively affects SVR rates in patients treated with peginterferon/ribavirin, some studies did not find such an association.28 ,29 Discrepant results in the relationship between baseline HOMA-IR and SVR may be explained by different proportions of patients with severe liver disease or with metabolic risk factors in the various studies.21 ,28 In our study population, BMI, steatosis and, to a lesser degree, viral load were independent risk factors for IR, suggesting that there are two subgroups of insulin-resistant patients: host-mediated and viral-mediated. It is possible that the telaprevir-based treatment may overcome viral-mediated but not host-mediated IR. Because our study was performed in a predominantly European cohort with small numbers of patients with obesity and diabetes, we were not able to test the latter hypothesis.

In conclusion, this subanalysis suggests that telaprevir overcomes metabolic factors and IR as negative predictors of efficacy of HCV therapy. These results warrant further confirmation in larger trials and in difficult-to-treat patients such as non-responders to peginterferon/ribavirin.

Acknowledgments

We thank all of the patients, investigators and study centre staff involved in the C208 clinical trial. This clinical trial was sponsored by Tibotec BVBA and Vertex Pharmaceuticals Inc. We also thank Bruce Coate and Karen Convery (Tibotec Inc) for statistical programming, and Emily de Looze (Gardiner-Caldwell Communications) for editing/styling and review coordination support (funded by Janssen Pharmaceuticals).

References

View Abstract

Footnotes

  • Funding This clinical trial was sponsored by Tibotec BVBA and Vertex Pharmaceuticals Inc. The authors received general editing/styling and coordination support from Gardiner-Caldwell Communications, which was funded by Janssen Pharmaceuticals.

  • Competing interests LS: consulting, advisory committees or review panels (Schering-Plough, Tibotec, Merck Sharp & Dohme, Roche, Bristol-Meyers Squibb (BMS), Gilead, GlaxoSmithKline (GSK) and Axcan Pharma); grant/research support (Roche and Schering-Plough); speaking and teaching (Roche, Schering-Plough, Gilead, BMS, Tibotec and Axcan Pharma). XF: advisory committees or review panels (Schering-Plough, Tibotec, and Merck Sharp & Dohme); grant/research support (Schering-Plough). TG: speaking and teaching (Roche, Essex, BMS, Gilead, Novartis and Falk); advisory committees or review panels (BMS); grant/research support (Roche). PF: speaking and teaching (Roche, Gilead, and Salix); advisory committees or review panels (Roche, Salix, Madaus Rottapharm, Tibotec, Novartis, Vertex and Boehringer Ingelheim); grant/research support (Roche and Madaus Rottapharm); patent held/filed (Madaus Rottapharm). FN: consulting (Ipsen, Gilead, BMS, CAF-DCF, Hepa Wash and Gambro); grant/research support (Roche, Merck Sharp & Dohme and Novartis). GC: advisory and committees or review panels (Janssen Cilag); speaking and teaching (Schering Plough, Roche and BMS). JD: declares no conflict of interest. PM: grant support (Roche, Schering-Plough and Gilead); investigator (Roche, Schering-Plough, Gilead, BMS, Vertex, Novartis, Tibotec, Merck Sharp & Dohme, Boehringer Ingelheim, Biolex and Intermune); speaker and/or expert (Roche, Schering-Plough, Gilead, BMS, Vertex, Novartis, Pharmasset, Tibotec, Merck Sharp & Dohme, Biolex, Intermune and Zymogenetics). IL-D, RDeM, GP and MB: employment (Janssen Pharmaceuticals/Tibotec/Johnson and Johnson).

  • Patient consent Obtained.

  • Ethics approval The study protocol was reviewed and approved by the appropriate institutional ethics committees/health authorities.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Digest
    Emad El-Omar William Grady Alexander Gerbes