SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Aliment Pharmacol Ther 2011; 33: 1162–1172

Summary

Background  Hepatitis C virus (HCV) is a major cause of chronic liver disease, cirrhosis and hepatocellular carcinoma and the identification of the predictors of response to antiviral therapy is an important clinical issue.

Aim  To determine the independent contribution of factors including IL28B polymorphisms, IFN-gamma inducible protein-10 (IP-10) levels and the homeostasis model assessment of insulin resistance (HOMA-IR) score in predicting response to therapy in chronic hepatitis C (CHC).

Methods  Multivariate analysis of factors predicting rapid (RVR) and sustained (SVR) virological response in 280 consecutive, treatment-naive CHC patients treated with peginterferon alpha and ribavirin in a prospective multicentre study.

Results  Independent predictors of RVR were HCV RNA <400 000 IU/mL (OR 11.37; 95% CI 3.03–42.6), rs12980275 AA (OR 7.09; 1.97–25.56) and IP-10 (OR 0.04; 0.003–0.56) in HCV genotype 1 patients and lower baseline γ-glutamyl-transferase levels (OR = 0.02; 0.0009–0.31) in HCV genotype 3 patients. Independent predictors of SVR were rs12980275 AA (OR 9.68; 3.44–27.18), age <40 years (OR = 4.79; 1.50–15.34) and HCV RNA <400 000 IU/mL (OR 2.74; 1.03–7.27) in HCV genotype 1 patients and rs12980275 AA (OR = 6.26; 1.98–19.74) and age <40 years (OR 5.37; 1.54–18.75) in the 88 HCV genotype 1 patients without a RVR. RVR was by itself predictive of SVR in HCV genotype 1 patients (OR 33.0; 4.06–268.32) and the only independent predictor of SVR in HCV genotype 2 (OR 9.0, 1.72–46.99) or genotype 3 patients (OR 7.8, 1.43–42.67).

Conclusions  In HCV genotype 1 patients, IL28B polymorphisms, HCV RNA load and IP-10 independently predict RVR. The combination of IL28B polymorphisms, HCV RNA level and age may yield more accurate pre-treatment prediction of SVR. HOMA-IR score is not associated with viral response.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Hepatitis C virus (HCV) infects up to 180 million people worldwide1 and is a major cause of chronic liver disease, cirrhosis and hepatocellular carcinoma.2 The current treatment, based on the combination of peginterferon alpha and ribavirin, leads to a sustained virological response (SVR) in ∼40–50% in patients with HCV genotype 1 and in ∼80% of those with genotype 2 or 3.3 Several baseline and on-treatment variables affect the likelihood of achieving SVR.4 Older age, advanced stage of fibrosis, African-American ethnicity and HCV- related factors, including HCV genotype 1 and high viral load at baseline, predict poor response to anti-viral therapy. Furthermore, metabolic factors, such as high body mass index (BMI), presence and severity of liver steatosis and increasing homeostasis model assessment of insulin resistance (HOMA-IR) score have been reported as negative predictors of response.5–9 On the other hand, early on-treatment kinetics of HCV RNA, e.g. undetectable HCV RNA at week 4, has a high positive predictive value of SVR.10, 11

Among the baseline predictors of response, the pre-treatment activation of IFN-stimulated genes (ISG) and the host genetic polymorphisms have been the subject of recent, major studies. Regarding ISG, it has been shown that low levels of intrahepatic and systemic CXC chemokine IFN-gamma inducible protein 10 kDa (IP-10, or CXCL10), a valid surrogate marker of ISG activation, predict a more pronounced first phase decline of HCV RNA during anti-viral therapy12 and increased SVR rates.13, 14 On the other hand, several independent studies have consistently shown that single nucleotide polymorphisms (SNPs) near IL28B, which encodes the type III interferon IFN-λ3 are strongly associated with the response to treatment of chronic hepatitis C.15–21 In particular, the homozygous genotypes TT at marker rs8099917, CC at marker rs12979860 and AA at marker rs12980275 are all associated with favourable treatment outcomes. These data have been confirmed in populations of different ancestry and HCV genotypes, and in various clinical scenarios.22, 23 However, very few studies are available where all these novel pre-treatment variables are analysed together in multivariate models.

We sought to extend our understanding of the impact of these recently reported genetic polymorphisms on treatment outcome by investigating the independent association of candidate genetic (including the rs8099917, rs12979860 and rs12980275 markers), host (including IP-10 levels and HOMA-IR score) and viral factors on virological response in a cohort of treatment-naïve, nondiabetic Caucasian patients with chronic hepatitis C.

Patients and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Patient population

Patients were recruited in an observational prospective cohort study, the Italian Hepatitis C Cohort Study (ITAHECS), designed to evaluate factors that influence response to anti-viral therapy, including host metabolic factors. Overall, 500 consecutive, treatment-naïve patients with chronic hepatitis C were enrolled and treated between June 2006 and June 2007 at seven tertiary referral Italian liver units. Eligibility criteria for therapy included Caucasian ancestry, age between 18 and 65 years, detectable HCV RNA in serum by polymerase chain reaction (PCR) (see below) and alanine aminotransferase (ALT) above the upper limit of normal within 6 months of treatment initiation. Patients were not considered for therapy if pregnant or breast-feeding, had decompensated cirrhosis or any other contraindication to therapy with the combination of pegylated IFN-α and ribavirin, were positive for hepatitis B surface antigen or anti-HIV antibodies, or reported a high daily alcohol intake (>40 g/day). All patients were asked to completely abstain from alcohol during anti-viral treatment. Therapy was based on the combination of pegylated IFN-α2a (180 μg/week) or pegylated IFN-α2b (1.5 μg/kg/week) plus ribavirin (800–1200 mg/day) for 48 weeks in HCV genotype 1 or 4 patients, and for 24 weeks in HCV genotype 2 or 3 patients. A rapid virological response (RVR) was defined as undetectable HCV RNA in serum at week 4 of therapy. Early virological response (EVR) was defined as serum HCV RNA negativity or any >2 log10 decline in HCV RNA levels at week 12 of therapy compared with baseline. Patients with sustained virological response (SVR) were those with undetectable HCV RNA in serum 24 weeks after stopping therapy. Treatment failures included patients who had a < 2 log10 drop in viral load at week 12 as compared to baseline, those whose HCV RNA was still detectable at week 24 (i.e. nonresponders, according to international guidelines3), and those who had undetectable HCV RNA at the end of therapy but detectable HCV RNA at 24 weeks after cessation of therapy (i.e. relapsers). The study was approved by all local ethical committees and all patients gave informed consent at enrolment in accordance with the Helsinki declaration.

For the present analysis, we included all patients who achieved a SVR, independent of the effective duration of therapy and of the dose of drugs received. On the other hand, patients with a treatment failure who had not received at least 80% of the recommended dose of pegylated IFN-α and ribavirin for at least 80% of the intended duration of treatment and subjects with missing information on treatment outcome were excluded from the analysis. In addition, we excluded also patients with diabetes, those infected with HCV genotype 4, due to their small number (n = 15) and those for whom genomic DNA was not available for study (Figure 1). Given the small number of non-adherent patients (n = 42), no intent-to-treat analysis was performed.

image

Figure 1.  Flow of the recruited patients.

Download figure to PowerPoint

Anthropometric and laboratory evaluations

Body mass index (BMI) was calculated as weight divided by the square of the height (kg/m2). A BMI >25 but ≤30 kg/m2 was considered as overweight, while obesity was defined as a BMI >30 kg/m2. Waist circumference was measured to the nearest 0.5 cm at the shortest point below the lower rib margin and the iliac crest. Venous blood samples were taken in the fasted state. Insulin levels were measured centrally on stored serum samples by electrochemoluminescence immunoassay (Insulina Immulite 2000, Medical System SpA, Genova, Italy; inter-assay coefficient of variation of 5%) using an autoanalyzer. The HOMA-IR score was calculated as reported.24 HOMA-IR scores were considered as continuous or dichotomous variable (i.e. < or ≥2, because this threshold has been reported in the literature as capable of discriminating SVR vs. non-SVR).7, 8 HCV genotyping was performed by INNO-LiPA HCV II assay (Innogenetics, Zwijndrecht, Belgium). Serum HCV RNA was quantified at baseline and at week 12 of therapy by reverse transcription-PCR using Cobas Amplicor HCV Monitor Test, v 2.0 (Roche, Basel, Switzerland). Qualitative HCV RNA assessment was performed at weeks 4, 12, 24, 48 during treatment and 24 weeks after stopping therapy using Cobas Amplicor HCV, v2.0 (Roche; limit of detection: 50 IU/mL). High baseline viral load was defined as HCV RNA levels >400 000 IU/mL.

Liver biopsy

Liver biopsy specimens were coded and scored by a single pathologist (MG at the University of Padua, Italy) who was blinded to clinical and biologic data, using the METAVIR score.25 Advanced fibrosis was defined as the presence of F3 or F4. Steatosis was recorded according to the criteria proposed by Brunt et al.26 and classified as absent, minimal (<5%), mild (5–33%), moderate (>33–66%) and severe (>66%).

IL28B genotyping

Genotyping was conducted in a blinded fashion relative to baseline characteristics and treatment outcome of patients. DNA samples from patients were genotyped for the IL28B rs8099917, rs12979860 and rs12980275 polymorphisms with TaqMan SNP genotyping assays (Applied Biosystems Inc., Foster City, CA, USA), using the ABI 7500 Fast real time thermocycler, according to manufacturers’ recommended protocols. TaqMan probes and primers were designed and synthesised by Applied Biosystems Inc. Automated allele calling was performed using sds software from Applied Biosystems Inc. Positive and negative controls were used in each genotyping assay. Primers and probes were reported previously.18

IP-10 quantification

Quantification of IP-10 was performed using Quantikine (R&D Systems, Minneapolis, MN, USA), a solid phase ELISA, on pre-treatment samples stored at −20 °C until assayed.

Statistical analysis

Continuous variables were summarised as mean ± standard deviation (s.d.) and categorical variables as frequency and percentage. Individual characteristics between groups were compared using the analysis of variance, Student t-test or the Mann–Whitney U-test for continuous variables, and contingency tables and the chi-squared test or the Fisher’s exact test for categorical data. Two-sided P values <0.05 were considered statistically significant. However, because of the multiple comparisons between subjects with and without RVR or SVR, we considered of interest and commented only the results of statistical tests with a P-value <0.01, to reduce the risk of false positive results due to chance only. Multivariate logistic regression analysis was used to identify factors significantly associated with RVR or SVR (dependent variable coded as 0 = absent or 1 = present). Variables significantly associated with the dependent variable on univariate analysis were included in the multivariate logistic regression model at the first step, and removed if not significantly associated with the dependent variable at the 0.1 P-value. Subsequently, a multivariate logistic regression analysis with backward selection was used to identify factors independently associated with RVR or SVR at a 0.05 P-value. The predictive value of RVR towards SVR was evaluated fitting a separate logistic regression model with SVR as the response variable and RVR as the only predictor. Sensitivity, specificity, positive and negative predictive values and likelihood ratios of a combination of variables associated with RVR or SVR in HCV genotype 1 patients were estimated.

The effect of IL28B polymorphisms was evaluated comparing the TT vs. TG/GG genotypes for marker rs8099917, CC vs. CT/TT genotypes for marker rs12979860 and AA vs. AG/GG genotypes for marker rs12980275. Patients’ data were collected in a computerised central database and analysed using the R statistical package, version 2.11.0 [R Development Core Team (2010). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org].

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Patients’ characteristics

Two-hundred and eighty patients were included in this analysis. Their baseline demographic, biochemical and virological characteristics are reported in Table 1. There were 166 (59%) male patients and the mean age was 46 ± 11 years. Mean BMI was 24.9 ± 3.8, 130 (46%) patients had BMI >25 and 90 (32%) had central obesity. The mean HOMA-IR score was 2.9 ± 3.0. HCV infection was due to genotype 1 in 121 (43%) cases, to genotype 2 in 104 (37%) and to genotype 3 in 55 (20%). HCV RNA >400 000 IU/mL was found in 189 (67%) patients. A pre-treatment liver biopsy was available in 177 patients. Liver histology showed moderate/severe (Metavir scores A2-A3) activity in 78 (44%) and advanced fibrosis (F3-F4) in 29 (16%) patients. Steatosis was moderate/severe in 25 (15%) cases.

Table 1.   Baseline characteristics of patients with chronic hepatitis C
 All patients (n = 280)Genotype 1 (n = 121)Genotype 2 (n = 104)Genotype 3 (n = 55)P*
  1. ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; HOMA-IR, homeostasis model of assessment of insulin resistance.

  2. Data are expressed as mean ± s.d. or number (%).

  3. P values for comparison among HCV genotypes 1, 2 and 3.

Age (years)46 ± 1146 ± 1149 ± 1141 ± 80.03
>40 years old212 (76)95 (79)84 (81)33 (60)0.01
Male gender166 (59)78 (64)47 (45)41 (75)0.0005
Body mass index (kg/m2)24.9 ± 3.825.2 ± 4.124.9 ± 3.624 ± 3.10.07
Waist circumference >102 cm (men) or >88 cm (women)90 (32)37 (30)38 (36)8 (14)0.009
HOMA-IR2.9 ± 3.03.4 ± 3.72.6 ± 2.43.0 ± 2.20.05
HOMA-IR > 2147 (53)73 (61)48 (46)26 (47)0.06
HCV RNA (log 10 IU/mL5.77 ± 0.875.79 ± 0.845.84 ± 0.796.0 ± 1.090.32
HCV RNA ≥400 000 IU/mL189 (67)85 (70)72(69)31 (57)0.25
METAVIR A2/A3 (n = 177)78 (44)47 (48)18 (35)13 (46)0.26
METAVIR F3/F4 (n = 177)29 (16)21 (22)2 (4)6 (21)0.006
Moderate/severe steatosis (n = 166)25 (15)6 (7)8 (16)11 (41)0.0001
ALT (U/L)93 ± 8392 ± 6175 ± 93129 ± 930.0001
AST/ALT ratio0.7 ± 0.20.7 ± 0.20.7 ± 0.21.0 ± 0.20.0005
γ-glutamyl-transferase (U/L)48 ± 4559 ± 5035 ± 3845 ± 380.004
Platelet count (× 103/μL)218 ± 59215 ± 61225 ± 59213 ± 560.89
Bilirubin (mg/dL)0.77 ± 0.30.8 ± 0.30.75 ± 0.31.0 ± 0.30.36
Ferritin (ng/mL)208 ± 208262 ± 262143 ± 111202 ± 1700.003
Total cholesterol (mg/dL)173 ± 39170 ± 30190 ± 36144 ± 420.0001
HDL cholesterol (mg/dL)53 ± 1653 ± 1754 ± 1451 ± 160.44
Triglyceride (mg/dL)89 ± 5089 ± 4498 ± 5774 ± 430.0002
Glucose (mg/dL)88 ± 1090 ± 1187 ± 987 ± 110.01
Insulin (μU/mL)13.4 ± 12.815.3 ± 15.711.9 ± 10.212 ± 9.40.11

The frequency of the favourable rs8099917 TT genotype tended to be higher in HCV genotype 3 patients compared to HCV genotype 1 or 2 patients, although the statistical significance was borderline (P = 0.049) (Table 2). There was a nonsignificant trend in the same direction for the favourable rs12979860 CC (P = 0.19) and rs12980275 AA genotypes (P = 0.21).

Table 2.   Frequency of IL28B genotypes and baseline levels of IP-10
 All patients (n = 280)Genotype 1 (n = 121)Genotype 2 (n = 104)Genotype 3 (n = 55)P*
  1. Data are expressed as mean ± s.d. or number (%).

  2. P values for comparison among HCV genotypes 1, 2 and 3.

rs8099917
TT158 (56)65 (54)54 (52)39 (71)0.049
TG107 (38)50 (41)41 (39)16 (29) 
GG15 (6)6 (5)9 (9)0 
rs12979860
CC109 (39)42 (35)39 (38)28 (51)0.19
CT136 (49)60 (50)52 (50)24 (44) 
TT35 (12)19 (15)13 (12)3 (5) 
rs12980275
AA109 (39)42 (35)39 (38)28 (51)0.21
AG136 (49)61 (50)51 (49)24 (44) 
GG35 (12)18 (15)14 (13)3 (5) 
Log IP-10 (pg/mL)2.46 ± 0.292.55 ± 0.272.37 ± 0.312.45 ± 0.270.01

Quantification of IP-10 was performed in 226 out of 280 patients: their baseline features are shown in Table S1. HCV genotype 1 patients had higher baseline IP-10 levels compared to patients infected with genotype 2 or 3 (P = 0.01) (Table 2). In HCV genotype 1 patients, significantly higher baseline IP-10 levels were observed in carriers of one or two of the risk G at rs12980275 as compared to homozygous carriers of the favourable AA at rs12980275 (P = 0.01) (Table 3); there was a nonsignificant trend in the same direction in carriers of one or two of the risk T at rs12979860 (P = 0.05) or of the risk G at rs8099917 (P = 0.07). No significant association was noted between basal HCV RNA and IL28B SNP variants in HCV genotype 1 patients (Table 3). In HCV genotype 2 or genotype 3 patients, the IL28B allele distribution showed no correlation with baseline IP-10 levels or HCV RNA (data not shown).

Table 3.   Baseline characteristics of HCV genotype 1 patients according to IL28B genetic variants
 rs8099917rs12979860rs12980275
TT (n = 65)TG (n = 50)GG (n = 6)PCC (n = 42)CT (n = 60)TT (n = 19)PAA (n = 42)AG (n = 61)GG (n = 18)P
  1. Data are expressed as mean ± s.d.

HCV RNA (log10 IU/mL)5.83 ± 0.855.74 ± 0.855.73 ± 0.720.855.9 ± 0.915.72 ± 0.885.73 ± 0.510.565.86 ± 0.915.76 ± 0.885.72 ± 0.520.80
IP-10 (log 10 pg/mL)2.49 ± 0.292.62 ± 0.232.56 ± 0.200.072.48 ± 0.262.56 ± 0.262.68 ± 0.250.052.45 ± 0.272.57 ± 0.252.69 ± 0.260.01

The baseline characteristics of patients infected with HCV genotype 4 (n = 15) and those with diabetes (n = 28), excluded from further analyses, are shown in Table S2.

Predictors of RVR

A RVR was observed in 172/280 (61%) patients, i.e. 33 (27%) with HCV genotype 1, 95 (91%) with genotype 2 and 44 (80%) with genotype 3.

By univariate analysis of HCV genotype 1 patients, a significant association was found between RVR and the homozygous carriers of the genotype TT at rs8099917 (odds ratio (OR) = 3.75; 95% confidence interval, CI: 1.42–10.6), CC at rs12979860 (OR = 6.8; 95% CI 2.6–18.11) or AA at rs12980275 (OR = 8.35; 95% CI 3.13–22.88). Since the AA at rs12980275 polymorphism showed the strongest association with RVR, only this marker of IL28B was considered for the multivariate analysis. Other factors associated with RVR in HCV genotype 1 patients were: lower basal IP-10, lower pre-treatment HCV RNA, HCV RNA <400,000, lower AST/ALT ratio and lower GGT (Table S3). No predictive factors of RVR were identified in HCV genotype 2 patients. In HCV genotype 3 patients, RVR was associated only with lower GGT.

By multivariate analysis of HCV genotype 1 patients, independent predictors of RVR were the baseline HCV RNA <400 000 IU/mL (OR = 11.37; 95% CI 3.03–42.6), rs12980275 type AA (vs. AG/GG) (OR 7.09; 95% CI 1.97–25.56) and IP-10 (OR 0.04; 95% CI 0.003–0.56) (Table 4). No factor predicting RVR was identified in genotype 2 infected patients. In HCV genotype 3 patients, RVR was predicted by lower baseline GGT levels (OR = 0.02 for one unit of increasing GGT; 95% CI 0.0009–0.31) (Table 4).

Table 4.   Multivariate logistic regression analysis of factors predictive of rapid virological response
PatientsnVariableOR95% CIP
HCV genotype 1121HCV RNA <400 000 IU/mL11.373.03–42.60.00031
  rs12980275 type (AA vs. AG/GG)7.091.97–25.560.0027
  IP-10 (log 10 pg/mL)0.0440.003–0.560.016
HCV genotype 355γ-glutamyl-transferase (U/L)0.0170.0009–0.310.006

Predictors of SVR

Sustained virological response was achieved in 209/280 (75%) patients, i.e. 65 (53%), 96 (92%) and 48 (87%) of patients infected with HCV genotype 1, 2, or 3, respectively.

By univariate analysis of HCV genotype 1 patients, SVR was associated with the favourable genotype TT at marker rs8099917 (OR = 4.05; 95% CI 1.78–9.3), CC at marker rs12979860 (OR = 8.17; 95% CI 3.02–24.18) and AA at marker rs12980275 (OR = 10.34; 95% CI 3.65–33.05). Since the AA at rs12980275 polymorphism showed the strongest association with SVR, only this marker of IL28B was considered for the multivariate analysis. Other factors associated with SVR in HCV genotype 1 patients were: lower basal IP-10, age <40 years, HCV RNA <400 000 IU/mL, lower GGT and RVR (Table S4). In HCV genotype 2 patients no factors were associated with SVR. In HCV genotype 3 patients, SVR was associated with lower ferritin level and higher total cholesterol. IL28B genetic variations did not predict SVR in HCV genotype 2 or 3 patients.

By multivariate analysis of HCV genotype 1 patients, independent predictors of SVR among baseline variables were marker AA rs12980275 (vs. AG/GG) (OR = 9.68; 95% CI 3.44–27.18), age <40 years (OR = 4.79; 95% CI 1.50–15.34) and HCV RNA <400 000 IU/mL (OR 2.74; 95% CI 1.03–7.27) (Table 5). RVR was by itself a very strong predictor of SVR in patients with HCV genotype 1, since 32 of the 33 subjects with RVR developed SVR (97%) (OR = 33.0; 95% CI 4.06–268.32).

Table 5.   Multivariate logistic regression analysis of factors predictive of sustained virological response in HCV genotype 1 patients
PatientsnVariableOR95% CIP
HCV genotype 1121rs12980275 type (AA vs. AG/GG)9.683.44–27.180.000017
  Age <404.791.50–15.340.008
  HCV RNA <400 000 IU/mL2.741.03–7.270.04
HCV genotype 1 without RVR88rs12980275 type (AA vs. AG/GG)6.261.98–19.740.0017
  Age <405.371.54–18.750.0083

In the 88 HCV genotype 1 patients who did not achieve a RVR, predictors of SVR were AA rs12980275 (vs. AG/GG) (OR = 6.26; 95% CI 1.98–19.74) and age <40 years (OR 5.37; 95% CI 1.54–18.75) (Table 5). RVR was by itself the only independent predictor of SVR in HCV genotype 2 patients (OR 9.0, 95% CI 1.72–46.99; P = 0.009) and in HCV genotype 3 patients (OR 7.8, 95% CI 1.43–42.67; P = 0.01).

Performance of IL28B type and HCV RNA for predicting RVR and SVR

To assess the performance of IL28B type compared with baseline HCV RNA and IP-10 levels for predicting RVR in our cohort of genotype 1 patients, we calculated sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and positive and negative likelihood ratios (Table 6). The rs12980275 marker and HCV RNA had similar NPV, PPV, specificity and sensitivity. However, the combination of rs12980275 marker and HCV RNA levels had rather high PPV (87%) and high specificity (98%) though low sensitivity (39%). The addition of baseline IP-10 levels to the model show similar specificity, higher sensitivity but lower PPV (76%).

Table 6.   Performance of IL28B rs12980275 genotype, baseline HCV RNA and IP-10 in predicting rapid virological response in HCV genotype 1 patients
Factors considered in the modelSensitivity %Specificity %PPV %NPV %LR+LR−
  1. PPV, positive predictive value; NPV, negative predictive value; LR+, likelihood ratio positive; LR−, likelihood ratio negative.

  2. * The IP-10 level was available in 92 patients only.

rs12980275 AA707855873.230.39
HCV RNA <400 000 IU/mL618256853.330.48
Log IP-10*209771766.70.82
rs12980275 AA and HCV RNA <400 000 IU/mL3998878117.330.62
rs12980275 AA, HCV RNA <400 000 IU/mL and Log IP-10*529476948.710.51

In HCV genotype 1 patients, the rs12980275 AA genotype, HCV RNA levels <400 000 IU/mL and the combination of rs12980275 AA genotype and HCV RNA levels <400 000 IU/mL were present in 35%, 30% and 12% of patients, respectively.

Among the 88 HCV genotype 1 patients without RVR, the PPV for SVR were 68%, 67% and 64% for rs12980275 marker, age ≤40 years and combining rs12980275 marker and age, respectively (Table 7). The combination of the two variables had a sensitivity of 61% and a specificity of 80%.

Table 7.   Performance of IL28B rs12980275 AA genotype and age in predicting sustained virological response in HCV genotype 1 patients without rapid virological response
Factors considered in the modelSensitivity %Specificity %PPV %NPV %LR+LR−
  1. PPV, positive predictive value; NPV, negative predictive value; LR+, likelihood ratio positive; LR−, likelihood ratio negative.

rs12980275 AA vs. non-AA398968713.610.68
Age ≤40 vs. >40 years309067683.330.76
rs12980275 AA and age ≤40 years618064773.030.49

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The present study reports on the relative impact of baseline features of chronic hepatitis C patients on the outcome of therapy with peginterferon alpha and ribavirin in the ‘real world’, i.e. outside clinical trials. Our results confirm previous observations that IL28B polymorphisms are independently associated with both RVR and SVR15–21 and that they have the strongest predictive values for SVR among HCV genotype 1 patients without RVR. Furthermore, baseline HCV RNA and IP-10 levels added to the predictiveness of IL28B variants, and may thus contribute to the genomic-based personalisation of therapy, even at the time direct-acting anti-virals (DAA) will be introduced. Vice versa, some metabolic variables, such as HOMA-IR score, BMI and steatosis (especially in HCV non-3 genotypes), failed to predict viral response independent of the above.

In our study, the combination of rs12980275 marker (AA genotype) and HCV RNA levels had a very high PPV (87%) and high specificity (98%) for RVR in the relatively difficult-to-treat genotype 1 patients. Adding baseline IP-10 levels to the model had comparable specificity but higher sensitivity increasing essentially the NPV (94 vs. 81%). Since 97% of patients with RVR and genotype 1 proceeded to attain SVR (Table 6) and taking into account that the combination of rs12980275 marker (AA genotype) and HCV RNA levels had a PPV of 87% for RVR, about 84% of patients with HCV genotype 1, rs12980275 AA genotype and HCV RNA <400 000 IU/mL may be cured with the currently available therapy. On the other hand, the predictive value of IL28B genotyping was very strong also in patients failing to reach RVR. Here, 33 out of 88 (38%) HCV genotype 1 patients without RVR achieved SVR: the rs12980275 AA genotype was able to identify two-thirds of them (PPV 68%). Age had similar PPV, and putting together IL28B genotype and age did not increase the PPV any further, but had a sensitivity of 61% and a NPV of 77%. Thus, if using a 4-week lead-in strategy with the standard dual combination, failure to reach RVR would certainly indicate the addition of at least one direct active anti-viral (DAA), with the possible exception of patients with <1 Log HCV RNA reduction, at high risk of selecting strains resistant to protease inhibitors (∼40%).27, 28

Favourable IL28B variants were not associated with higher levels of baseline HCV RNA, in contrast to previous studies.15, 18 Conversely, they were associated with lower levels of IP-10, consistent with the fact that good virological responders have a counterintuitive lower level of endogenous activation of the innate, interferon-associated immune response.29–31 Baseline levels of the chemokine IP-10 are an independent negative predictor of RVR, as reported previously.12–14 IP-10 levels predict especially the first-phase decline of HCV RNA during treatment32 and are considered a valid surrogate marker of innate immune response activation. Although its negative predictive value of virological response seems counterintuitive, it is in agreement with the fact that ISG are preactivated in nonresponders, who also fail to further activate genes aimed at establishing an effective anti-viral state.31 Higher IP-10 levels were found in patients with the risk allele G at rs12980275, whereas the significance was border-line for alleles T at rs12979860 or G at rs8099917, in keeping with other works.32, 33 Thus, it is clear that both IP-10 and IL28B genotype are independently associated with virological response and that IP-10 are not merely surrogate indicators of unfavourable IL28B genotypes, but that they add to their predictive value, and they indeed impair the benefit of favourable IL28B variants.34 On the other hand, we could not identify a cutoff level of IP-10 capable to discriminate responders from nonresponders with sufficient accuracy.

A surprising finding of our study is the lack of predictive value of the HOMA-IR. The clinical significance of determining the HOMA-IR score before therapy is debated. Initial studies showed that chronic hepatitis C patients with HOMA-IR score <2 had significantly better chances of achieving SVR, independent of the HCV genotype.6, 8, 35, 36 The predictive value of HOMA-IR was further reported in patients of Asian35, 37, 38 or Middle East ancestry.36, 39 However, recent works have failed to confirm these findings,40–42 and a similar controversy exists for chronic hepatitis C patients coinfected with HIV.43–45 Higher HOMA-IR scores and/or insulin levels are inversely correlated with the HCV RNA decay occurring during the first days46–48 or weeks49, 50 of therapy. In our study, we observed an association between HOMA-IR and RVR only in the genotype 3 subgroup (by univariate analysis), but no association was seen when carrying out the multivariate analysis for RVR or with SVR for any viral genotype. It is possible, however, that associations between HOMA-IR and SVR rate be confounded by several factors (e.g. adherence). In addition, it is possible that discrepancies among studies be accounted for by differences among the patients’ populations, especially in terms of prevalence of central obesity. Furthermore, the accuracy of HOMA-IR score assessment has been criticised and recent work suggests that, especially in lean and/or non-obese patients, HOMA-IR may be burdened by lack of standardisation.51 Thus, the future significance of assessing IR by HOMA-IR before treatment seems questionable, especially at a time when patients’ stratification before therapy is better achieved by genomic-based assays.

Our study has limitations. First, it was conducted exclusively on Caucasian patients. However, the vast majority of the data appeared so far in the literature have been produced among Caucasians, and limited data are available on other selected ethnical groups, such as Japanese,17 Hispanics and African-Americans.15 The overall ITAHECS patients population rarely belongs to ethnical groups other than Caucasians, since they come from the Middle East or Sub-Saharan Africa, for which no solid host gene predictive data are available. More extensive studies are warranted in this area. Second, our population contained few patients with metabolic disturbances such as central obesity: thus, data may not be generalisable to other study populations from different geographical regions and more diverse metabolic profiles.

In conclusion, genetic variation in IL28B and pretherapy levels of IP-10 and HCV RNA may be useful as first-line tools to identify the majority of HCV genotype 1 patients achieving RVR with the currently available dual therapy. This may be used to stratify patients, especially when novel DAAs will be available. It is expected that additional advances in pharmacogenomics may further improve predictive models.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Declaration of personal interests: None. Declaration of funding interests: This study was supported in part by the Ministero dell’Istruzione, dell’Università e della Ricerca, PRIN (Programmi di Ricerca Scientifica di Rilevante Interesse Nazionale) number 2005064913, Italy, and by the Swiss National Science Foundation grant number 314730-130498 (to F.N.).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • 1
    Shepard CW, Finelli L, Alter MJ. Global epidemiology of hepatitis C virus infection. Lancet Infect Dis 2005; 5: 55867.
  • 2
    Seeff LB. Natural history of chronic hepatitis C. Hepatology 2002; 36(Suppl. 1): S3546.
  • 3
    Ghany MG, Strader DB, Thomas DL, et al. Diagnosis, management, and treatment of hepatitis C: an update. Hepatology 2009; 49: 133574.
  • 4
    Zeuzem S. Heterogeneous virologic response rates to interferon-based therapy in patients with chronic hepatitis C: who respond less well? Ann Intern Med 2004; 140: 37081.
  • 5
    Bressler BL, Guindi M, Tomlinson G, Heathcote J. High body mass index is an independent risk factor for nonresponse to antiviral treatment in chronic hepatitis C. Hepatology 2003; 38: 63944.
  • 6
    Romero Gomez M, Del Mar Viloria M, Andrade RJ, et al. Insulin resistance impairs sustained response rate to peginterferon plus ribavirin in chronic hepatitis C patients. Gastroenterology 2005; 128: 63641.
  • 7
    Poynard T, Ratziu V, McHutchison J, et al. Effect of treatment with peginterferon or interferon alfa-2b and ribavirin on steatosis in patients infected with hepatitis C. Hepatology 2003; 38: 7585.
  • 8
    Poustchi H, Negro F, Hui J, et al. Insulin resistance and response to therapy in patients infected with chronic hepatitis C virus genotypes 2 and 3. J Hepatol 2008; 48: 2834.
  • 9
    Durante-Mangoni E, Zampino R, Portella G, et al. Correlates and prognostic value of the first-phase hepatitis C virus RNA kinetics during treatment. Clin Infect Dis 2009; 10: 439.
  • 10
    Yu ML, Dai CY, Huang JF, et al. Rapid virologic response and treatment duration for chronic hepatitis C genotype 1 patients: a randomized trial. Hepatology 2008; 47: 188493.
  • 11
    Poordad FF. Review article: the role of rapid virological response in determining treatment duration for chronic hepatitis C. Aliment Pharmacol Ther 2010; 31: 125167.
  • 12
    Askarieh G, Alsiö A, Pugnale P, et al. Systemic and intrahepatic interferon-gamma-inducible protein 10 kDa predicts the first-phase decline in hepatitis C virus RNA and overall viral response to therapy in chronic hepatitis C. Hepatology 2010; 51: 152330.
  • 13
    Romero AI, Lagging M, Westin J, et al. Interferon (IFN)-gamma inducible protein-10: association with histological results, viral kinetics, and outcome during treatment with pegylated IFN-alpha 2a and ribavirin for chronic hepatitis C virus infection. J Infect Dis 2006; 194: 895903.
  • 14
    Lagging M, Romero AI, Westin J, et al. IP-10 predicts viral response and therapeutic outcome in difficult-to-treat patients with HCV genotype 1 infection. Hepatology 2006; 44: 161725.
  • 15
    Ge D, Fellay J, Thompson AJ, et al. Genetic variation in IL28B predicts hepatitis C treatment-induced viral clearance. Nature 2009; 461: 399401.
  • 16
    Suppiah V, Moldovan M, Ahlenstiel G, et al. IL28B is associated with response to chronic hepatitis C interferon-a and ribavirin therapy. Nat Genet 2009; 41: 11004.
  • 17
    Tanaka Y, Nishida N, Sugiyama M, et al. Genome-wide association of IL28B with response to pegylated interferon-α and ribavirin therapy for chronic hepatitis C. Nat Genet 2009; 41: 11059.
  • 18
    Rauch A, Kutalik Z, Descombes P, et al. Genetic variation in IL28B is associated with chronic hepatitis C and treatment failure: a genome-wide association study. Gastroenterology 2010; 138: 133845.
  • 19
    Mc Carthy JJ, Li HJ, Thompson A, et al. Replicated association between an IL28B gene variant and a sustained response to pegylated interferon and ribavirin. Gastroenterology 2010; 138: 230714.
  • 20
    Thompson AJ, Muir AJ, Sulkowski MS, et al. Interleukin-28B polymorphism improves viral kinetics and is the strongest pretreatment predictor of sustained virologic response in hepatitis C virus-1 patients. Gastroenterology 2010; 139: 1209, e18.
  • 21
    Mangia A, Thompson AJ, Santoro R, et al. IL28B polymorphism determines treatment response of patients with hepatitis C genotypes 2 or 3 who do not achieve a rapid virologic response. Gastroenterology 2010; 139: 8217, e1.
  • 22
    Rallón NI, Naggie S, Benito JM, et al. Association of a single nucleotide polymorphism near the interleukin-28B gene with response to hepatitis C therapy in HIV/hepatitis C virus-coinfected patients. AIDS 2010; 24: F239.
  • 23
    Fukuhara T, Taketomi A, Motomura T, et al. Variants in IL28B in liver recipients and donors correlate with response to peg-interferon and ribavirin therapy for recurrent hepatitis C. Gastroenterology 2010; 139: 157785, e1-3.
  • 24
    Bonora E, Targher G, Alberiche M, et al. Homeostasis model assessment closely mirrors the glucose clamp technique in the assessment of insulin sensitivity: studies in subjects with various degrees of glucose tolerance and insulin sensitivity. Diabetes Care 2000; 23: 5763.
  • 25
    Bedossa P, Poynard T. An algorithm for the grading of activity in chronic hepatitis C. Hepatology 1996; 24: 28993.
  • 26
    Brunt EM, Janney CG, Di Bisceglie AM, et al. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999; 94: 246774.
    Direct Link:
  • 27
    Poordad F, McCone J, Bacon BR, et al. Boceprevir (BOC) combined with peginterferon alfa-2b/ribavirin (P/R) for treatment-naïve patients with hepatitis C virus (HCV) genotype (G) 1: SPRINT-2 Final Results. Hepatology 2010; 52(Suppl.): 107A.
  • 28
    Bacon BR, Gordon SC, Lawitz E, et al. HCV RESPOND-2 final results: high sustained virologic response among genotype 1 previous non-responders and relapsers to peginterferon/ribavirin when retreated with boceprevir plus PEGINTRON (Peginterferon alfa-2b)/Ribavirin. Hepatology 2010; 52(Suppl.): 430A.
  • 29
    Chen L, Borozan I, Feld J, et al. Hepatic gene expression discriminates responders and nonresponders in treatment of chronic hepatitis C viral infection. Gastroenterology 2005; 128: 143744.
  • 30
    Asselah T, Bieche I, Narguet S, et al. Liver gene expression signature to predict response to pegylated interferon plus ribavirin combination therapy in patients with chronic hepatitis C. Gut 2008; 57: 51624.
  • 31
    Sarasin-Filipowicz M, Oakeley EJ, Duong FH, et al. Interferon signaling and treatment outcome in chronic hepatitis C. Proc Natl Acad Sci USA 2008; 105: 70349.
  • 32
    Lagging M, Askarieh G, Bochud P, et al. IP-10 is associated with IL28B variation and predicts the first phase decline of HCV RNA and outcome of therapy in chronic hepatitis C. Hepatology 2010; 52(Suppl.): 766A.
  • 33
    Honda M, Sakai A, Yamashita T, et al. Hepatic ISG expression is associated with genetic variation in interleukin 28B and the outcome of IFN therapy for chronic hepatitis C. Gastroenterology 2010; 139: 499509.
  • 34
    Darling JM, Aerssens J, Fanning G, et al. Quantitation of pretreatment serum IP-10 improves the predictive value of an IL28B gene polymorphism for hepatitis C treatment response. Hepatology 2010; 52(Suppl.): 382A.
  • 35
    D’Souza R, Sabin CA, Foster GR. Insulin resistance plays a significant role in liver fibrosis in chronic hepatitis C and in the response to antiviral therapy. Am J Gastroenterol 2005; 100: 150915.
    Direct Link:
  • 36
    Moucari R, Ripault MP, Martinot-Peignoux M, et al. Insulin resistance and geographical origin: major predictors of liver fibrosis and response to peginterferon and ribavirin in HCV-4. Gut 2009; 58: 16629.
  • 37
    Dai CY, Huang JF, Hsieh MY, et al. Insulin resistance predicts response to peginterferon-alpha/ribavirin combination therapy in chronic hepatitis C patients. J Hepatol 2009; 50: 7128.
  • 38
    Mizuta T, Kawaguchi Y, Eguchi Y, et al. Whole-body insulin sensitivity index is a highly specific predictive marker for virological response to peginterferon plus ribavirin therapy in chronic hepatitis C patients with genotype 1b and high viral load. Dig Dis Sci 2010; 55: 1839.
  • 39
    Khattab M, Eslam M, Sharwae MA, et al. Insulin resistance predicts rapid virologic response to peginterferon/ribavirin combination therapy in hepatitis C genotype 4 patients. Am J Gastroenterol 2010; 105: 19707.
  • 40
    Petta S, Cammà C, Di Marco V, et al. Time course of insulin resistance during antiviral therapy in non-diabetic, non-cirrhotic patients with genotype 1 HCV infection. Antivir Ther 2009; 14: 6319.
  • 41
    Angelico F, Francioso S, Del Ben M, et al. Clinical trial: low plasma cholesterol and oxidative stress predict rapid virological response to standard therapy with peginterferon and ribavirin in HCV patients. Aliment Pharmacol Ther 2009; 30: 44451.
  • 42
    Akuta N, Suzuki F, Hirakawa M, et al. Amino acid substitutions in the hepatitis C virus core region of genotype 1b are the important predictor of severe insulin resistance in patients without cirrhosis and diabetes mellitus. J Med Virol 2009; 81: 10329.
  • 43
    Merchante N, de los Santos-Gil I, Merino D, et al. Insulin resistance is not a relevant predictor of sustained virological response to pegylated interferon plus ribavirin in HIV/HCV co-infected patients. J Hepatol 2009; 50: 68492.
  • 44
    Cacoub P, Carrat F, Bédossa P, et al. Insulin resistance impairs sustained virological response rate to pegylated interferon plus ribavirin in HIV-hepatitis C virus-coinfected patients: HOMAVIC-ANRS HC02 Study. Antivir Ther 2009; 14: 83945.
  • 45
    Ryan P, Resino S, Miralles P, et al. Insulin resistance impairs response to interferon plus ribavirin in patients coinfected with HIV and hepatitis C virus. J Acquir Immune Defic Syndr 2010; 55: 17681.
  • 46
    Neumann AU, Bain VG, Yoshida EM, et al. Early prediction of sustained virological response at day 3 of treatment with albinterferon-alpha-2b in patients with genotype 2/3 chronic hepatitis C. Liver Int 2009; 29: 13505.
  • 47
    Durante-Mangoni E, Zampino R, Portella G, et al. Correlates and prognostic value of the first-phase hepatitis C virus RNA kinetics during treatment. Clin Infect Dis 2009; 49: 498506.
  • 48
    Bortoletto G, Scribano L, Realdon S, et al. Hyperinsulinaemia reduces the 24-h virological response to PEG-interferon therapy in patients with chronic hepatitis C and insulin resistance. J Viral Hepat 2010; 17: 47580.
  • 49
    Nasta P, Gatti F, Puoti M, et al. Insulin resistance impairs rapid virologic response in HIV/hepatitis C virus coinfected patients on peginterferon-alfa-2a. AIDS 2008; 22: 85761.
  • 50
    Grasso A, Malfatti F, De Leo P, et al. Insulin resistance predicts rapid virological response in non-diabetic, non-cirrhotic genotype 1 HCV patients treated with peginterferon alpha-2b plus ribavirin. J Hepatol 2009; 51: 98490.
  • 51
    Lam KD, Bacchetti P, Abbasi F, et al. Comparison of surrogate and direct measurement of insulin resistance in chronic hepatitis C virus infection: impact of obesity and ethnicity. Hepatology 2010; 52: 3846.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Table S1. Baseline characteristics of 226 patients with chronic hepatitis C tested for IP-10.

Table S2. Baseline characteristics of chronic hepatitis C patients with diabetes or HCV genotype 4 who were excluded from the analysis.

Table S3. Baseline features of patients with chronic hepatitis C according to rapid virological response (RVR).

Table S4. Features of patients with chronic hepatitis C according to sustained virological response (SVR).

FilenameFormatSizeDescription
APT_4635_sm_tS1-4.doc251KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.