Vitamin A deficiency is associated with hepatitis C virus chronic infection and with unresponsiveness to interferon-based antiviral therapy

Authors

  • Davide Bitetto,

    1. Department of Medical Sciences Experimental and Clinical, Medical Liver Transplantation Unit, Internal Medicine, University of Udine, Italy
    Search for more papers by this author
  • Nadia Bortolotti,

    1. Department of Medical Sciences Experimental and Clinical, Medical Liver Transplantation Unit, Internal Medicine, University of Udine, Italy
    Search for more papers by this author
  • Edmondo Falleti,

    1. Department of Medical Sciences Experimental and Clinical, Medical Liver Transplantation Unit, Internal Medicine, University of Udine, Italy
    Search for more papers by this author
  • Sara Vescovo,

    1. Department of Medical Sciences Experimental and Clinical, Medical Liver Transplantation Unit, Internal Medicine, University of Udine, Italy
    Search for more papers by this author
  • Carlo Fabris,

    1. Department of Medical Sciences Experimental and Clinical, Medical Liver Transplantation Unit, Internal Medicine, University of Udine, Italy
    Search for more papers by this author
  • Giovanna Fattovich,

    1. Gastroenterology Clinic, Department of Medicine, Azienda Ospedaliero-Universitaria, University of Verona, Italy
    Search for more papers by this author
  • Annarosa Cussigh,

    1. Department of Medical Sciences Experimental and Clinical, Medical Liver Transplantation Unit, Internal Medicine, University of Udine, Italy
    Search for more papers by this author
  • Sara Cmet,

    1. Department of Medical Sciences Experimental and Clinical, Medical Liver Transplantation Unit, Internal Medicine, University of Udine, Italy
    Search for more papers by this author
  • Ezio Fornasiere,

    1. Department of Medical Sciences Experimental and Clinical, Medical Liver Transplantation Unit, Internal Medicine, University of Udine, Italy
    Search for more papers by this author
  • Elisa Ceriani,

    1. Department of Clinical and Experimental Medicine, Università del Piemonte Orientale “A. Avogadro,” Novara, Italy
    Search for more papers by this author
  • Mario Pirisi,

    1. Department of Clinical and Experimental Medicine, Università del Piemonte Orientale “A. Avogadro,” Novara, Italy
    2. Interdepartmental Research Centre for Autoimmune Diseases (IRCAD), Novara, Italy
    Search for more papers by this author
  • Pierluigi Toniutto

    Corresponding author
    1. Department of Medical Sciences Experimental and Clinical, Medical Liver Transplantation Unit, Internal Medicine, University of Udine, Italy
    • Department of Medical Sciences Clinical and Experimental, Internal Medicine, Medical Liver Transplantation Unit, University of Udine, Italy
    Search for more papers by this author
    • fax: +390432559490


  • Potential conflict of interest: Nothing to report.

Abstract

Recent data suggest that vitamin A modulates the expression of type I interferon receptor enhancing the antireplication effect of interferon-α on hepatitis C virus (HCV). This study aimed to investigate the prevalence of vitamin A deficiency among patients with chronic HCV infection and to assess whether vitamin A deficiency could be associated with unresponsiveness to interferon-based antiviral therapy. The analysis included 199 consecutive treatment-naïve chronic HCV patients in whom pretreatment serum vitamin A and 25-OH vitamin D were measured; 119 healthy blood donors were used as controls. Median (interquartile range) serum vitamin A in HCV-positive patients was significantly lower than in controls: 256 ng/mL (128-440) versus 742 (624-942, P < 0.0001). Overall sustained viral response was achieved in 122/199 patients, 46/109 infected by difficult to treat HCV genotypes. In these latter, 39/104 (37.5%) were nonresponders. At multivariate analysis, nonresponse to antiviral therapy was predicted by carriage of interleukin (IL)-28B T/* genotypes, baseline serum levels of γGT >60 IU/mL, of HCV RNA >600,000 IU/mL, of vitamin A ≤100 ng/mL, and a cumulative dose of ribavirin ≤80%. Seventeen patients (9.0%) had both serum levels of vitamin A ≤100 ng/mL and of vitamin D ≤20 ng/mL; the presence of a combined vitamin A and D deficiency was found to be a strong independent predictor of nonresponse to antiviral therapy. Conclusion: A high percentage of patients with chronic HCV infection have serum vitamin A deficiency. This condition is associated with nonresponse to antiviral therapy. (HEPATOLOGY 2013)

New specifically targeted direct antiviral agents (DAAs) against hepatitis C virus (HCV) have recently become available in many countries. However, they will not substitute, at least for several years, the interferon (IFN) plus ribavirin-based antiviral therapies. This is mainly due to the fact that although DAAs are very potent in inhibiting HCV replication, they are prone to favor the development of viral resistances if used in monotherapy. Triple antiviral therapy significantly improved the sustained viral response (SVR) rates in HCV genotype 1 naïve-infected patients compared to IFN plus ribavirin standard therapy. When treated with triple antiviral therapy, patients previously nonresponders to IFN plus ribavirin dual antiviral regimen achieved significantly lower SVR rates compared to relapsers.1, 2 The results suggest that the sensitivity of the host to the biological action of IFN is a prerequisite for the eradication of the infection, even using DAA triple therapy. Therefore, it would be important to understand if interferon sensitivity in the host could be modified prior to antiviral therapy in order to maximize the possibility to achieve SVR.

Several not-modifiable and modifiable factors have been identified to help clinicians in predicting, prior to antiviral treatment and in individual patients, the probability of achieving SVR.3, 4 Among the modifiable factors, insulin resistance,5 body mass index, and iron deposition in the liver have been previously investigated. More recently, serum vitamin D levels emerged as a new modifiable predictor of SVR.6 Vitamin D deficiency was associated with lower SVR rates in HCV-positive patients treated with IFN plus ribavirin in comparison to patients with normal serum vitamin D levels; this suggests that vitamin D supplementation could be helpful in enhancing the responsiveness to antiviral therapy.

Vitamin A deficiency has been found to be an important factor in conditioning a more severe course of viral infections such as measles.7 Vitamin A can up-regulate the expression of type I IFN receptor, enhancing the anti-HCV replication effect of IFN-α.8 In a cohort of previous nonresponder patients with HCV chronic infection,9 all-transretinoic-acid (ATRA) demonstrated a direct antiviral and a strong additive or synergistic effect with pegylated IFN. Nevertheless, only few studies are available regarding the prevalence of vitamin A deficiency in HCV chronically infected patients10, 11; furthermore, the potential effect of vitamin A in modifying the antiviral action of IFN and ribavirin has never been studied.

The aims of the present study were: (1) to investigate the prevalence of vitamin A deficiency among patients with chronic HCV infection; (2) to assess whether vitamin A deficiency could be associated with the absence of responsiveness to IFN plus ribavirin-based antiviral therapy; and (3) to evaluate the possible additive effect of vitamin A and vitamin D deficiency in influencing nonresponse.

Abbreviations

ATRA, all-transretinoic-acid; cEVR, complete early viral response; DAA, direct antiviral agent; EOT, end of treatment viral response; HCV, hepatitis C virus; HOMA, homeostasis model assessment; IFN, interferon; IL-28B, interleukin 28B; RVR, rapid viral response; SVR, sustained viral response.

Patients and Methods

Patients.

The study population included a total of 199 consecutive, HCV-positive treatment-naïve patients of Caucasian ethnicity who received antiviral therapy at one of three academic centers in northern Italy (Medical Liver Transplantation Unit, University of Udine [N = 67; 33.7%], Department of Gastroenterology, University of Verona [N = 85; 42.7%], Department of Clinical and Experimental Medicine, University of Novara [N = 47; 23.6%]) from September 2005 to October 2009. Chronic HCV hepatitis was defined by the presence of anti-HCV antibodies, serum HCV RNA positivity, and the persistent elevation of alanine aminotransferase (ALT) for at least 6 months. In addition, 131 patients had a liver biopsy performed within the 6 months preceding the start of antiviral therapy. Exclusion criteria were: (i) decompensated liver cirrhosis (Child-Pugh score >6); (ii) the presence of hepatocellular carcinoma (HCC); (iii) HIV coinfection; (iv) HBV coinfection; (v) autoimmune liver disease, defined according to validated diagnostic criteria12; (vi) genetic liver disease (e.g., Wilson's disease, hemochromatosis); (vii) concomitant use of drugs known to affect serum vitamin D and A concentration; and (viii) active intravenous drug use. The main clinical and demographic characteristics of the studied population are reported in Table 1. The study was conducted according to the principles of the Declaration of Helsinki and approved by the hospital Institutional Review Board and Ethical Committees. All of the patients signed a written informed consent to participate in the study. An overnight fasting blood sample was drawn to determine the baseline blood tests, including HCV RNA quantification, using real-time polymerase chain reaction (PCR) (TaqMan, Roche) and HCV genotype, detected using the InnoLipa genotyping kit (Innogenetics).

Table 1. Baseline Main Demographic and Clinical Characteristics of the Studied Population (N=199) and of Control Subjects (N=119)
 HCV+ControlsP
N=199N=119
  1. Continuous variables are presented as medians (interquartile range) and categorical variables as frequencies (%). The statistical analysis was performed by means of Pearson chi-square test for categorical variables and Mann-Whitney test for continuous variables. ALT, alanine aminotransferase; γGT, gamma-glutamyl transpeptidase; HOMA, homeostasis model assessment. *Available in 136 patients.

Age years48 (40-60)49 (42-55)0.927
Male gender102 (51.3%)73 (61.3%)0.080
Body mass index kg/m224.2 (22.3-27.6)24.0 (22.7-25.1)0.079
Alcohol consumption g/day0 (0-10)0 (0-10)0.525
Cholesterol mg/dL170 (145-194)209 (186-238)<0.001
ALT IU/mL68 (41-120)19 (16-24)<0.001
γGT IU/mL38 (21-70)21 (16-32)<0.001
HOMA1.96 (0.96-3.46)1.57 (0.97-2.65)0.208
Vitamin A ng/mL256 (128-442)742 (623-946)<0.001
25 OH vitamin D ng/mL20.6 (13.5-29.8)28.4 (22.6-33.0)<0.001
HCV genotypes195 (47.7%)
258 (29.1%)
332 (16.1%)
4-514 (7.0%)
HCV RNA (x 103 IU/mL)700 (293-1.710)
Use of PEG-interferon α-2b140 (70.4%)
Grading pretreatment*2 (2-4)
Staging pretreatment*2 (1-3)

Controls.

A group of 119 blood donors was used as healthy controls. All were tested at the moment of blood withdrawal for all viral and bacterial transmissible disease; they had normal serum levels of transaminases. The main demographic characteristics did not differ significantly from those of HCV-positive patients (Table 1).

Vitamin A Assay.

Vitamin A concentration was determined in fasting serum samples using high-performance reversed-phase liquid chromatography (HPLC) with ultraviolet (UV) detection.13 This assay focuses on the measurement of retinol only and not of retinyl esters. All samples were protected from light and stored at −70°C until they were assayed. Samples were prepared for analysis as follows: after serum protein precipitation with ethanol and subsequent three-time extraction with hexane, supernatants were pooled together and evaporated to dryness. The remaining residues were redissolved with methanol and injected into the Agilent Eclipse XDB-C18 (4.6 × 250 mm, 5 μm particle size) chromatographic column. The mobile phase was methanol with a flow rate of 1.3 mL/min. Vitamin A was detected at 325 nm and quantified by mean of a vitamin A external standard. Retinyl acetate was used as internal standard. The retention time for vitamin A was 3.5 minutes. A Beckman-Coulter (Fullerton, CA) HPLC system, equipped with a 125S pump, a 508 autosampler, and a 166 UV detector with variable wavelength were used. Ethanol, methanol, and exane were HPLC grade and purchased from Carlo Erba Reagents (Milan, Italy); vitamin A and retinyl acetate were purchased from Sigma (St. Louis, MO). The internal quality assurance was performed using serum controls from RECIPE (Munich, Germany); they are based on human serum and are available with mean values traceable to the Standard Reference Material from the National Institute of Standards and Technology 968d (NIST-SRM 968d). The precision for measurement of duplicate samples was 10% and 7.2% when samples containing vitamin A concentrations of 776 and 1267 ng/mL were analyzed. An anonymous code number was used to identify the tubes containing serum samples from patients and control subjects.

Vitamin D Assay and Homeostasis Model Assessment (HOMA) Score.

For all 199 patients, a serum sample, collected before starting antiviral therapy, was separated and stored at −80°C until used. This sample was available to measure pretreatment 25-OH vitamin D, glucose, and insulin serum levels, as described,6 and to calculate the HOMA score.

Molecular Biology.

Genotyping for the IL-28B rs12979860 C>T polymorphism was performed in the genomic DNA extracted from whole blood, by PCR-based restriction fragment length polymorphism assay. The detailed methodology of the IL-28 B genotyping has been described by our group elsewhere.14

Histology.

In 136 out of 199 patients (68.3%) a liver biopsy before starting therapy was performed. Grade and stage were scored according to the Ishak system.15

Antiviral Therapy Schedule and Outcomes.

All of the patients were treated with a combination therapy of pegylated (PEG) IFN plus ribavirin. In all, 140 patients (70.4%) received PEG IFN α-2b (PEG-Intron, Schering-Plough) at a dosage of 1.5 μg/kg per week, and 59 patients (29.6%) received PEG IFN α-2a (Pegasys, Roche) at a dosage of 180 μg per week. In patients infected by HCV genotypes 1, 4, and 5, ribavirin (Rebetol, Schering-Plough or Copegus, Roche) was administered according to the body weight (1,000 mg/d for patients weighing <75 kg, 1,200 mg/d for those weighing ≥75 kg); in patients infected by HCV genotypes 2 and 3, a flat ribavirin dose of 800 mg/d was used. The duration of antiviral therapy was 48 weeks for genotypes 1, 4, and 5 and 24 weeks for genotypes 2 and 3 infected patients, respectively. The definition of rapid viral response (RVR) was based on undetectable HCV RNA at week 4; complete early viral response (cEVR) was based on HCV RNA undetectable at week 12; end of treatment viral response (EOT) was based on HCV RNA undetectable at the end of antiviral therapy; SVR was based on HCV RNA undetectable 6 months after completing the scheduled period of therapy. Relapsers were defined as patients with HCV RNA reappearance after having reached EOT. Nonresponders were considered patients in whom HCV RNA dropped less than 2 log from baseline at week 12 (null responders) or those in whom HCV RNA dropped more than 2 log from baseline at week 12 (partial responders) but was still detectable at week 24.16 A stopping rule was applied in nonresponder patients.

Statistical Analysis.

Statistical analysis of the data was performed using the BMDP dynamic statistical software package 7.0 (Statistical Solutions, Cork, Ireland). Continuous variables are presented as median (interquartile range) and categorical variables as frequencies (%). The associations between categorical variables were evaluated using the Pearson chi-squared test and, when appropriate, the chi-squared test for linear trend. Differences for continuous variables between patients and controls were evaluated using the Mann-Whitney test. The correlation between vitamin A and vitamin D serum levels was assessed by means of Spearman's rank correlation coefficient. Logistic regression analysis was performed to identify independent predictors of nonresponse to antiviral therapy. The initial model comprised all variables to be tested; those with a nonsignificant coefficient were then removed with a backward approach.

Results

Vitamin A Levels.

A significant difference in the median (interquartile range) serum vitamin A levels was recorded between patients and controls: 256 ng/mL (128-440) versus 742 ng/mL (624-942, P < 0.0001, Fig. 1). This difference persisted after adjusting for age and body mass index (BMI) (P < 0.001). Eighty-four patients (42.2%) had vitamin A deficiency defined as serum level ≤200 ng/mL; 39 patients (19.6%) had serum levels ≤100 ng/mL, identifying severe vitamin A deficiency. None of the controls had vitamin A serum levels <200 ng/mL. BMI was found to be associated with vitamin A serum levels: patients with BMI ≤25 kg/m2 presented less frequently severe vitamin A deficiency (Table 2). A season-related significant difference in serum vitamin D levels was detected, with higher levels (>20 ng/mL) in summer and early autumn in comparison to winter and spring (42/61 versus 63/138, P = 0.002). On the contrary, no association was found between vitamin A serum levels >100 ng/mL and the season of the sampling (47/61 versus 113/138, P = 0.428). No significant association was found between vitamin A and vitamin D serum levels (P = 0.170).

Figure 1.

Vitamin A serum levels in healthy controls (N = 119) and in patients with chronic hepatitis C (N = 199).

Table 2. Association Between Clinical and Demographic Variables and Baseline Serum Vitamin A Severe Deficiency (Vitamin A Serum Levels ≤/>100 ng/mL)
 Serum Vitamin A 
≤100 ng/mL N=39 (19.6%)>100 ng/mL N=160 (80.4%)P
  1. The statistical analysis was performed by means of Pearson chi-square test. HOMA, homeostasis model assessment; ALT, alanine aminotransferase; γGT, gamma-glutamyl transpeptidase; IL-28B, interleukin 28B. *Available in 136 patients.

Age >50 years18 (46.2%)70 (43.7%)0.786
Female gender19 (48.7%)78 (48.7%)0.997
Body mass index >25 kg/m226 (66.7%)54 (33.7%)<0.001
HOMA index >4.08 (20.5%)33 (20.6%)0.988
Cholesterol <200 mg/dL31 (79.5%)127 (79.4%)0.988
ALT >60 IU/mL25 (64.1%)89 (55.6%)0.337
γGT >60 IU/mL15 (38.5%)45 (28.1%)0.207
HCV RNA >600.000 IU/mL24 (61.5%)89 (55.6%)0.504
HCV genotype 1-4-522 (56.4%)87 (54.4%)0.819
Grading pre treatment >4*6 (19.4%)22 (21.0%)0.847
Staging pre treatment >4*6 (19.4%)14 (13.3%)0.405
25 OH vitamin D >20 ng/mL19 (48.7%)86 (53.7%)0.573
Alcohol consumption >20 g/day2 (5.1%)18 (11.3%)0.254
IL-28B rs12979860 C/T or T/T genotypes28 (71.8%)103 (64.4%)0.381

Viral Response.

Ninety-five patients (47.7%) achieved RVR, 140 (70.4%) cEVR, 147 (73.9%) EOT, and 122 (61.3%) SVR. In the 90 patients infected by HCV genotypes 2-3 the following frequencies were observed: RVR 66 (73.3%), cEVR 84 (93.3%), EOT 82 (91.1%), and SVR 76 (84.4%). In HCV genotypes 1-4-5 (N = 109), 29 patients (26.6%) attained RVR, 56 (51.4%) cEVR, 65 (59.6%) EOT, and 46 (42.2%) SVR. Seventeen patients dropped out, for an overall rate of 8.5%. To assess nonresponse rate, patients who dropped out before the completion of the 12th week of therapy and, in the case of partial response, before the completion of the 24th week of therapy were excluded. Thus, nonresponse was detected in 41 of the remaining 190 patients (21.6%), 39/104 (37.5%) infected by difficult-to-treat, and 2/86 (2.3%) by easy-to-treat HCV genotypes.

Vitamin A and Viral Response.

Considering patients altogether, a highly significant association was found between the presence of severe vitamin A deficiency (≤100 ng/mL) and the condition of nonresponse to antiviral therapy (36.1% versus 18.2%, P = 0.019, Fig. 2). In patients infected by difficult-to-treat HCV 1-4-5 genotypes, nonresponse was detected in 61.9% of those with vitamin A ≤100 ng/mL, in 33.3% of those with vitamin A in the interval >100-200 ng/mL, and in 31.0% of those with vitamin A >200 ng/mL (P = 0.015, Fig. 3). The association between nonresponse to antiviral treatment and the main clinical and demographic variables is reported in Table 3. The absence of response to antiviral treatment was significantly influenced by the HCV genotype, the IL-28B rs12979860 C>T polymorphism, the baseline gamma-glutamyltranspeptidase (γGT) levels, presence of cirrhosis, having taken more than 80% of the total scheduled dose of ribavirin, and by the baseline serum levels of 25-OH vitamin D. In difficult-to-treat HCV genotypes the association between nonresponse and the demographic variables gave similar results (Table 3). In these patients, independent predictors of nonresponse to antiviral treatment were: a baseline serum level of γGT >60 IU/mL, carriage of IL-28B T/* genotypes, having taken less than 80% of the scheduled dose of ribavirin, and a baseline vitamin A serum level ≤100 ng/mL (Table 4).

Figure 2.

Association between baseline vitamin A serum levels (≤100/>100-200/>200 ng/mL) and antiviral responses considering all patients (N = 190). RVR = rapid viral response, cEVR = complete early viral response, EOT = end of treatment viral response, SVR = sustained viral response, REL = relapsers, NORES = nonresponse. The statistical analysis was carried out by means of Pearson chi-square test (#patients with vitamin A ≤100/>100 ng/mL) or chi-square test for linear trend (##patients with vitamin A ≤100/>100-200/>200 ng/mL).

Figure 3.

Association between baseline vitamin A serum levels (≤100/>100-200/>200 ng/mL) and antiviral responses considering the subgroup of difficult-to-treat 1-4-5 HCV genotypes (N = 104). RVR = rapid viral response, cEVR = complete early viral response, EOT = end of treatment viral response, SVR = sustained viral response, REL = relapsers, NORES = nonresponse. The statistical analysis was carried out by means of Pearson chi-square test (#patients with vitamin A ≤100/>100 ng/mL) or chi-square test for linear trend (##patients with vitamin A ≤100/>100-200/>200 ng/mL).

Table 3. Baseline Demographic and Clinical Characteristics of HCV-Positive Patients (N=190) and in Difficult-to-Treat HCV Genotypes (N=104) in Relationship to the Condition of Nonresponse to Antiviral Therapy With PEG-Interferon and Ribavirin
 All Genotypes (N=190)Genotypes 1-4-5 (N=104)
 Nonresponders (N=41)Responders (N=149)PNonresponders (N=39)Responders (N=65)P
  1. The statistical analysis was performed by means of Pearson chi-square test or chi-square test for linear trend when appropriate. BMI, body mass index; HOMA, homeostasis model assessment; ALT, alanine aminotransferase; γGT, gamma-glutamyl transpeptidase. *Available in 131 patients and in 82 infected by difficult to treat HCV genotypes.

Age >50 years17 (41.5%)66 (44.3%)0.74615 (38.5%)21 (32.3%)0.523
Male gender21 (51.2%)78 (52.3%)0.89820 (51.3%)38 (58.5%)0.475
Body mass index >25 Kg/m217 (41.5%)57 (38.3%)0.70915 (38.5%)21 (32.3%)0.523
HOMA index<2.020 (48.8%)79 (53.0%)0.55520 (51.3%)34 (52.4%)0.610
2.0-4.012 (29.2%)43 (28.9%)11 (28.2%)22 (33.8%)
>4.09 (22.0%)27 (18.1%)8 (20.5%)9 (13.8%)
Cholesterol >200 mg/dL4 (9.8%)33 (22.1%)0.0764 (10.3%)11 (16.9%)0.349
ALT >60 IU/mL27 (65.9%)81 (54.4%)0.18826 (66.7%)32 (49.2%)0.083
γGT >60 IU/mL23 (56.1%)33 (22.1%)<0.00122 (56.4%)14 (21.5%)<0.001
HCV RNA >600.000 IU/mL26 (63.4%)80 (53.7%)0.26725 (64.1%)32 (49.2%)0.140
HCV genotype 2-32 (4.9%)84 (56.4%)<0.001---
Use of PEG-interferon α-2b28 (68.3%)104 (69.8%)0.85327 (69.2%)41 (63.1%)0.523
*Grading pre treatment >44 (14.3%)23 (22.3%)0.3514 (14.3%)9 (16.7%)0.779
*Staging pre treatment >49 (32.1%)10 (9.7%)0.0039 (32.1%)5 (9.3%)0.009
Cumulative dose of ribavirin ≤80%12 (29.3%)15 (10.1%)0.00212 (30.8%)10 (15.4%)0.063
Cumulative dose of interferon ≤80%6 (14.6%)17 (11.4%)0.5756 (15.4%)8 (12.3%)0.656
Alcohol consumption >20 g/day4 (9.8%)15 (10.1%)0.9534 (10.3%)8 (12.3%)0.751
IL-28B C>T polymorphismC/C2 (4.9%)64 (43.0%)<0.0012 (5.2%)29 (44.6%)<0.001
C/T32 (78.0%)70 (47.0%)30 (76.9%)28 (43.1%)
T/T7 (17.1%)15 (10.0%)7 (17.9%)8 (12.3%)
25 OH vitamin D in ng/mL≤1011 (26.8%)19 (12.8%)0.03811 (28.2%)5 (7.7%)0.010
>10/≤2012 (29.3%)45 (30.2%)11 (28.2%)20 (30.8%)
>2018 (43.9%)85 (57.0%)17 (43.6%)40 (61.5%)
Table 4. Logistic Regression Analysis to Assess the Independent Predictors of Nonresponse to Antiviral Therapy in Patients Infected by HCV Genotypes 1-4-5 (N=104)
Model AO.R.95% C.I.P
  1. Model (A) comprises all the variables reported in Table 3 (except for histology which was available only in 82 patients) plus vitamin A serum levels. Model (B) comprises the variables of model A) plus the interaction between vitamin A and vitamin D serum levels: vitamin D ≤20 ng/mL alone, vitamin A ≤100 ng/mL alone, both vitamin D ≤20 ng/mL and vitamin A ≤100 ng/mL vs. both vitamin D >20 ng/mL and vitamin A >100 ng/mL. Model (C) comprises the variables of model A) plus the interaction between IL-28B genotypes and vitamin A serum levels: IL-28B CC and vitamin A ≤100 ng/mL, IL-28B CT-TT and vitamin A >100 ng/mL, IL-28B CT-TT and vitamin A ≤100 ng/mL vs. IL-28B CC and vitamin A >100 ng/mL. Variables with a coefficient close to 0, having a minimal influence to the remaining coefficients were excluded. OR, odds ratio; CI, confidence interval; *OR = 13.0, 95% CI 1.88-90.1, P = 0.007 when patients with combined Vitamin D/A ≤20/≤100 ng/mL were compared to the remaining patients. IL-28B, interleukin 28B; γGT, gamma-glutamyl transpeptidase.

IL-28B genotypes: T/* vs. CC26.34.34-159<0.001
γGT: >60 IU/mL vs. ≤60 IU/mL5.241.69-16.20.003
HCV RNA: >600.000 IU/mL vs. ≤600.000 IU/mL4.391.39-13.80.009
Cumulative dose of ribavirin: ≤80% vs. >80%3.200.91-11.20.059
Vitamin A: ≤100 vs. >100 ng/mL3.681.03-13.20.038
Vitamin D: ≤20 vs. >20 ng/mL1.670.58-4.760.318
Model BO.R.95% C.I.P
IL-28B genotypes: T/* vs. CC22.83.92-133<0.001
γGT: >60 IU/mL vs. ≤60 IU/mL6.331.92-20.80.002
HCV RNA: >600.000 IU/mL vs. ≤600.000 IU/mL4.871.48-16.00.008
Cumulative dose of ribavirin: ≤80% vs. >80%3.891.05-14.40.037
Vitamin D/A: ≤20/>100 ng/mL vs. normal Vitamin D/A0.970.29-3.260.070*
Vitamin D/A: >20/≤100 ng/mL vs. normal Vitamin D/A1.040.16-6.64
Vitamin D/A: ≤20/≤100 ng/mL vs. normal Vitamin D/A12.91.73-96.7
Model CO.R.95% C.I.P
γGT: >60 IU/mL vs. ≤60 IU/mL5.031.64-15.40.004
HCV RNA: >600.000 IU/mL vs. ≤600.000 IU/mL4.431.42-13.80.008
Cumulative dose of ribavirin: ≤80% vs. >80%3.420.94-12.40.052
IL-28B/Vitamin A: CC/≤100 ng/mL vs. CC/>100 ng/mL3.290.11-96.30.001
IL-28B/Vitamin A: CT-TT/>100 ng/mL vs. CC/>100 ng/mL26.52.91-242
IL-28B/Vitamin A: CT-TT/≤100 ng/mL vs. CC/>100 ng/mL1079.00-1280 

Vitamin A and Vitamin D Deficiencies and Viral Response.

Eighty-seven (45.8%) out of 190 patients and 47 (45.2%) out of those infected by the difficult-to-treat HCV genotypes had baseline vitamin D deficiency (≤20 ng/mL). Seventeen patients (9.0%) had both serum levels of vitamin A ≤100 ng/mL and of vitamin D ≤20 ng/mL (group A), 19 (10.0%) had isolate serum levels of vitamin A ≤100 ng/mL (group B), 70 (36.8%) had isolate serum levels of vitamin D ≤20 ng/mL (group C), and 84 (44.2%) did not present vitamin A or vitamin D deficiency (group D). A significant linear trend for decreasing frequencies of nonresponse to antiviral therapy was found in all treated patients starting from group A (9/17, 52.9%) to group B (4/19, 21.1%) to group C (14/70, 20.0%) to group D (14/84, 16.7%; P = 0.005). The same finding was detected in difficult-to-treat HCV genotypes: group A (9/11, 81.8%) to group B (4/10, 40.0%) to group C (13/36, 36.1%) to group D (13/47, 27.7%; P = 0.002). The multivariate approach highlighted the role of vitamin A deficiency as an independent predictor of nonresponse, while vitamin D deficiency alone did not reach statistical significance. Nevertheless, combined vitamin A and D deficiency was found to be an even stronger predictor of nonresponse in comparison to single vitamin A deficiency (Table 4).

Vitamin A, IL-28B rs12979860 C>T Polymorphism, and Viral Response in Difficult-to-Treat HCV Genotypes.

Thirty-one patients (29.8%) carried the C/C genotype, 57(54.8%) carried at least one T allele and had serum vitamin A >100 ng/mL, 16 (15.4%) carried at least one T allele and had vitamin A ≤100 ng/mL. Nonresponse was found to occur with increasing frequencies from C/C patients (2/31) to T/* vitamin A >100 ng/mL patients (25/57) to T/* vitamin A ≤100 ng/mL patients (12/16), with a significant linear trend (P < 0.001) (Fig. 4). With the multivariate approach, the interaction between IL-28B T/* genotypes and vitamin A ≤100 ng/mL was found to be the strongest independent predictor of nonresponse (Table 4).

Figure 4.

Association between the interaction of IL-28B rs12979860 C>T polymorphism with baseline vitamin A serum levels (≤100/>100 ng/mL) and antiviral responses considering the subgroup of difficult-to-treat 1-4-5 HCV genotypes (N = 104). RVR = rapid viral response, cEVR = complete early viral response, EOT = end of treatment viral response, SVR = sustained viral response, REL = relapsers, NORES = nonresponse. The statistical analysis was carried out by means of Pearson chi-square test for linear trend.

Discussion

Vitamin A has been demonstrated to have pleiotropic influences, ranging from eyesight to organogenesis, and regulation of metabolism and immune response. Vitamin A acting by way of ATRA plays an important role in the regulation of innate and cell-mediated immunity and in antibody-mediated responses, as recently reviewed by Hall et al.17 It is well known that children with vitamin A deficiency are at increased risk to develop respiratory diseases and that vitamin A deficiency affects morbidity and mortality associated with diarrheal diseases and measles infection in low-income countries.18

Recently it has been suggested that vitamin A can also be involved in the antiviral response to hepatitis C virus. In an in vitro HCV transfection model, up-regulation of type I IFN receptors was demonstrated after addition of 9-cis retinoic acid to HCV transfected liver tumor cell lines; this fact emphasizes the antiviral action of IFN-α in vitro.8 It might be presumed that retinoic acid binding to its receptor mediates the expression of RIG-1 gene (retinoic acid induced gene-1). RIG, similar to Toll-like receptor (TLR)-3, represents an essential step in the innate immunity response to many viruses, acting as a double-stranded RNA (dsRNA) cytosol sensing receptor.19 After its binding with dsRNA, RIG together with CARDIF forms the RIG dimer/CARDIF complex that activates IKK-ϵ, which in turn activates IRK-3 or IRF-7, determining the final transcription of Type I IFNs.20 As proof of the potential additive effect between retinoic acid and IFN-α in the antiviral response to HCV, a recent clinical study performed in HCV-positive patients demonstrated that the addition of ATRA to PEG-IFN-α was associated with a higher decrease in serum HCV RNA compared to ATRA monotherapy.9

The analysis of the present data first found a strong association between vitamin A deficiency and chronic HCV infection. This observation has been suggested by others studies. In 2001 Rocchi et al.21 demonstrated that in patients with chronic liver disease plasma but not liver tissue vitamin A concentrations were low; unfortunately, no data about retinol in normal liver tissue were available. Interestingly, the authors found a direct association between liver tissue content of retinol and aminotransferase serum levels. Concerning vitamin A and HCV chronic infection, two reports have been published: the first one pertaining to a cohort of drug users with HIV and HCV coinfection10 and the second one to a cohort of chronic HCV monoinfected patients with different stages of liver disease.11 In the former group the authors found an association between retinol deficiency and HCV but not HIV infection; it should be emphasized, however, that the concomitant drug abuse could represent a confounding factor. In the latter group vitamin A deficiency was significantly correlated with the stage of HCV liver disease more than with the presence of HCV infection itself: progressively higher rates of vitamin A deficiency were observed starting from HCV mild hepatitis to cirrhosis and hepatocellular carcinoma. The present study represents the first analysis of a cohort of chronically HCV mono-infected patients who underwent antiviral therapy at different stages of liver disease severity. Regardless of the staging and the grading of liver disease, evaluated with liver biopsy, a strong association was found between HCV infection and vitamin A deficiency. Interestingly, vitamin A deficiency was found to be associated with higher BMI values but not with the serum levels of cholesterol, triglycerides, or vitamin D, which has been reported to be severely decreased in chronic HCV infection. Altogether, these findings suggest that alternative and possibly independent mechanisms could be involved in determining vitamin D and vitamin A deficiencies in patients with chronic HCV infection.

The real novelty and probably the most important finding of this study is the association between serum vitamin A deficiency and the condition of nonresponse to antiviral therapy, suggesting that vitamin A could be an important and modifiable factor interfering with IFN sensitivity in patients with chronic hepatitis C. This finding, together with the data suggesting an antiviral activity against HCV of ATRA, suggests that vitamin A supplementation and normalization of its serum levels, before antiviral treatment, could enhance the responsiveness to INF-based antiviral therapy. These considerations seem to confirm those derived from in vitro experiments that provided evidence of a pivotal role of retinol in enhancing the expression of IFN receptor and IFN signaling, linking vitamin A deficiency to IFN unresponsiveness. The fact that vitamin A and vitamin D serum levels are not reciprocally influenced suggests that they can exert an additive and probably synergistic effect on viral response. Indeed, the analysis showed that a concomitant vitamin A and D deficiency strongly impairs the responsiveness to antiviral therapy and that its impact is not so far from that exerted by IL-28B polymorphisms. The major and obvious difference is that, instead of IL-28B polymorphisms, vitamins serum levels might be modified. Moreover, it is important to note that a strong additive effect in determining nonresponse was observed in patients with concomitant carriage of the IL-28B T/* genotype and vitamin A serum levels ≤100 ng/mL.

A possible concern in terms of the use of vitamin A supplementation in clinical practice is represented by its possible hepatotoxicity.22 However, the experiences concerning the use of polyprenoic acid, a synthetic vitamin A derivate, in the prophylaxis of HCC in patients with chronic viral hepatitis23 and the study by Bocher et al.9 did not support this assumption.

The main limitations of the present study lie in its retrospective design and in the lack of data concerning the dietary intake of both vitamin A and D. It is conceivable that vitamin D serum levels could be greatly influenced by the season, since sunlight exposure is recognized as a key factor in determining vitamin D synthesis. Nevertheless, it cannot be excluded that season-related dietary variations could influence vitamin A intake and serum levels. Nevertheless, the multicenter design of the study supported the external validation of data that have been confirmed in each center.

In conclusion, a high percentage of patients with chronic HCV infection presented serum vitamin A deficiency. This condition is strongly associated with nonresponse to antiviral therapy, suggesting that vitamin A serum levels could modulate the responsiveness to IFN-based antiviral therapy. It will be of great importance to verify if vitamin A supplementation could restore the IFN sensitivity in nonresponders because the success of new DAAs (i.e., boceprevir and telaprevir) is still conditioned by IFN responsiveness.

Ancillary