Low vitamin D serum level is related to severe fibrosis and low responsiveness to interferon-based therapy in genotype 1 chronic hepatitis C

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

25-Hydroxyvitamin D (25[OH]D) can potentially interfere with inflammatory response and fibrogenesis. Its role in disease progression in chronic hepatitis C (CHC) and its relation with histological and sustained virological response (SVR) to therapy are unknown. One hundred ninety-seven patients with biopsy-proven genotype 1 (G1) CHC and 49 healthy subjects matched by age and sex were consecutively evaluated. One hundred sixty-seven patients underwent antiviral therapy with pegylated interferon plus ribavirin. The 25(OH)D serum levels were measured by high-pressure liquid chromatography. Tissue expression of cytochrome (CY) P27A1 and CYP2R1, liver 25-hydroxylating enzymes, were assessed by immunochemistry in 34 patients with CHC, and in eight controls. The 25(OH)D serum levels were significantly lower in CHC than in controls (25.07 ± 9.92 μg/L versus 43.06 ± 10.19; P < 0.001). Lower levels of 25(OH)D were independently linked to female sex (P = 0.007) and necroinflammation (P = 0.04) by linear regression analysis. CYP27A1, but not CYP2R1, was directly related to 25(OH)D levels (P = 0.01), and inversely to necroinflammation (P = 0.01). Low 25(OH)D (odds ratio [OR], 0.942; 95% confidence interval [CI], 0.893–0.994) and cholesterol (OR, 0.981; 95%CI, 0.969–0.992) levels, older age (OR, 1.043; 95%CI, 1.002–1.085), high ferritin (OR, 1.003; 95%CI, 1.001–1.005), and necroinflammation (OR, 2.235; 95%CI, 1.014–4.929) were independently associated with severe fibrosis (F3–F4) by multivariate logistic analysis. Seventy patients (41%) achieved SVR. By multivariate analysis, hepatic steatosis (OR, 0.971; 95%CI, 0.944–0.999), lower cholesterol (OR, 1.009; 95% CI, 1.000–1.018), and 25(OH)D levels (OR, 1.039; 95%CI, 1.002–1.077) were independently associated with no SVR. Conclusion: G1 CHC patients had low 25(OH)D serum levels, possibly because of reduced CYP27A1 expression. Low vitamin D is linked to severe fibrosis and low SVR on interferon (IFN)-based therapy. (HEPATOLOGY 2010.)

T the activated hormonal form of vitamin D, 1-25-dihydroxyvitamin D, is essential for calcium and bone homeostasis.1, 2 Vitamin D 25 and 1α-hydroxylation occurs in the liver and in the kidney, respectively, involving different isoforms of cytochrome P450 (CYP), namely CYP2R1, CYP27A1, and others in the liver, and CYP27B1 in the kidney.3

Vitamin D deficiency is associated with many common and serious pathological conditions, including cancer, autoimmune disease, cardiovascular disease, insulin resistance (IR), and diabetes.1, 4, 5 There is also an association between vitamin D status and both cholestatic and noncholestatic chronic liver diseases.6–10 In patients with noncholestatic chronic hepatitis and cirrhosis, some studies8–10 have reported normal serum levels of 25-hydroxyvitamin D (25[OH]D), the liver-hydroxylated form of vitamin D and the best estimate of overall vitamin D status.1, 2 Conversely, other studies have found low serum 25(OH)D levels in patients with chronic hepatitis and cirrhosis of different origins.11–13 Low 25(OH)D levels have been related to poor liver function because of the association between vitamin D status and hepatic function indexes11 or the stage of cirrhosis.11, 14, 15 In keeping with these studies, several reports describe reduced bone mineral density in patients with chronic liver disease8, 9, 13, 15, 17, 18 and cirrhosis.9, 10, 12, 13, 19 Recently, Targher et al.18 observed lower 25(OH)D serum levels in patients with biopsy-proven nonalcoholic fatty liver disease, identifying an independent association between the histological characteristics of nonalcoholic fatty liver disease and low 25(OH)D levels. Experimental evidence also suggested the potential ability of vitamin D, through interaction with its nuclear receptor (vitamin D receptor), to interfere with inflammatory response and fibrogenesis.4

The aim of our study was to evaluate serum levels of 25(OH)D in a cohort of patients with biopsy-proven genotype 1 (G1) chronic hepatitis C (CHC), and to investigate the potential relationships between 25(OH)D, the histological features of disease, and the response to antiviral therapy.

Abbreviations

25(OH)D, 25-hydroxyvitamin D; ALT, alanine aminotransferase; AUC, area under the curve; CHC, chronic hepatitis C; CI, confidence interval; CY, cytochrome; G1, genotype 1; GGT, gamma-glutamyltransferase; HCV, hepatitis C virus; IR, insulin response; OR, odds ratio; SVR, sustained virological response.

Patients and Methods

Patients.

From January 2007 to December 2008, a total of 197 consecutive patients with G1 CHC, resident in Sicily and recruited at the Gastrointestinal and Liver Unit at the University Hospital in Palermo, fulfilling all inclusion and exclusion criteria detailed later, were assessed. Patients were included if they had a histological diagnosis of CHC (any degree of fibrosis, including cirrhosis) on a liver biopsy performed within 6 months before enrollment. G1 CHC patients were characterized by the presence of anti–hepatitis C virus (HCV) and HCV RNA, with persistently abnormal alanine aminotransferase (ALT), and by alcohol consumption of less than 20 g/day in the last year or more, evaluated by a specific questionnaire. Exclusion criteria were (1) advanced cirrhosis (Child-Pugh B and C); (2) hepatocellular carcinoma; (3) other causes of liver disease or mixed causes (excessive alcohol consumption, hepatitis B, autoimmune liver disease, Wilson's disease, hemochromatosis, α1-antitrypsin deficiency); (4) cancer or chronic intestinal diseases; (5) human immunodeficiency virus infection; (6) therapy with medications known to affect vitamin D3 metabolism, including vitamin/mineral supplements; (7) previous treatment with antiviral therapy, immunosuppressive drug, or regular use of steatosis-inducing drugs; and (8) active intravenous drug addiction.

Forty-nine randomly-selected, nondiabetic, healthy blood donors of the same ethnic group as CHC patients and living in Sicily, recruited from January 2008 to December 2008, matched for age and sex, were enrolled as controls. Alcohol consumption of more than 20 g/day during the previous year or therapy with medications known to affect vitamin D3 metabolism (calcium, vitamin D supplementation, hormonal therapy, alendronate) were additional exclusion criteria. All had normal ALT values (<30 UI/L), and no evidence of viral infection (anti-HCV, anti–human immunodeficiency virus, and hepatitis B surface antigen negative) or steatosis, verified by ultrasound scan.

The study was performed in accordance with the principles of the Declaration of Helsinki and its appendices. Approval was obtained from the hospital's Institutional Review Board and Ethics Committee, and written informed consent was obtained from all cases and controls.

Clinical and Laboratory Assessment.

Clinical and anthropometric data were collected at the time of liver biopsy. Body mass index was calculated on the basis of weight in kilograms and height (in meters). Waist circumference was measured at the midpoint between the lower border of the rib cage and the iliac crest. The diagnosis of arterial hypertension was based on the following criteria: systolic blood pressure at least 135 mm Hg or diastolic blood pressure 85 mmHg or more (measured three times within 30 minutes, in the sitting position and using a brachial sphygmomanometer), or use of blood-pressure–lowering agents. The diagnosis of type 2 diabetes was based on the revised criteria of the American Diabetes Association, using a value of fasting blood glucose at least 126 mg/dL on at least two occasions.20 In patients with a previous diagnosis of type 2 diabetes, current therapy with insulin or oral hypoglycemic agents was documented.

A 12-hour overnight fasting blood sample was drawn at the time of biopsy to determine serum levels of ALT, gamma-glutamyltransferase (GGT), total cholesterol, high-density lipoprotein and low-density lipoprotein cholesterol, triglycerides, ferritin, plasma glucose concentration, and platelet count. Serum insulin was determined by a two-site enzyme enzyme-linked immunosorbent assay (Mercodia Insulin ELISA, Arnika). IR was determined with the homeostasis model assessment method.21

The analysis of serum 25(OH) D was performed using a Chromosystem reagent kit and a chromatographic system equipped with a Waters 1525 Binary high-pressure liquid chromatography pump connected to a photo diode array detector, and detection was carried out at 265 nm. In accordance with the kit's instructions, a serum 25(OH)D concentration of 30 μg/L was considered the threshold value for identifying low levels of vitamin D.

All patients were tested at the time of biopsy for HCV-RNA by qualitative polymerase chain reaction (Cobas Amplicor HCV Test version 2.0; limit of detection: 50 IU/mL). HCV RNA positive samples were quantified by Versant HCV RNA 3.0 bDNA (Bayer Co. Tarrytown, NY) expressed in IU/mL. Genotyping was performed by INNO-LiPA, HCV II, Bayer.

Histology.

Slides were coded and read by one pathologist (D.C.) who was unaware of the patient's identity and history. A minimum length of 15 mm of biopsy specimen or the presence of at least 10 complete portal tracts was required.22 Biopsy specimens were classified according to the Scheuer numerical scoring system.23 The percentage of hepatocytes containing macrovescicular fat was determined for each 10× field. An average percentage of steatosis was then determined for the entire specimen. Steatosis was assessed as the percentage of hepatocytes containing fat droplets (minimum 5%) and evaluated as a continuous variable. Steatosis was classified as mild at 5% to 30% or moderate-severe at 30% or more.

CYP27A1 and CYP2R1 Liver Immunohistochemical Evaluation.

Immunohistochemistry was performed on liver biopsy tissue sections by means of the streptavidin-biotin-peroxidase method. All samples were fixed for 24 hours with 10% buffered formalin, paraffin-embedded, and cut in serial sections of 3 μg. Tissue morphology was evaluated by hematoxylin-eosin staining.

Immunohistochemical detection of CYP2R1 and CYP27A1 was performed using anti-human CYP2R1 (C-15) and CYP27A1 (P-17) (goat polyclonal antibody, Santa Cruz Biotechonology, Inc.). After dewaxing, the endogenous peroxidase activity was inhibited by the pretreatment of the tissue section with 3% hydrogen peroxide before incubation with the primary antibody. The sections were washed twice, after all incubation steps, in phosphate-buffered saline for 5 minutes. Then slides were microwave-oven heated three times for 5 minutes in Tris/ethylenediaminetetra-acetic acid pH 9.0 buffer (heat-induced epitope retrieval), and washed with phosphate-buffered saline. Sections were subsequentially incubated in the presence of the primary antibody overnight at −4°C. The specimens were then incubated with the LSAB HRP detection kit (Universal DakoCytomation LSAB+ System HRP) at room temperature, according to the manufacturer's instructions. As a chromogen, 3-amino-9-ethylcarbazole was used for 5 minutes at room temperature with subsequent nuclear counterstaining with hematoxylin. Normal mouse serum was used instead of primary antibodies as a negative control.

A four-grade semiquantitative scoring system (in other words, score 0-3), performed by one pathologist (C.T.) unaware of other variables, was adopted for the evaluation of CYP2R1 and CYP27A1 immunohistochemical expression. The score was graded according to the intensity of the staining: score 0 was defined as the absence of significant reactivity, scores 1 and 2 as slight and moderate reactivity, respectively, and score 3 as intense reactivity. Because a slight degree of variation could be observed in the immunohistochemical expression of CYP2R1 and CYP27A1 among different areas of the same sample, the most intense reactivity observed in the biopsy specimen was recorded as the summary score.

Immunohistochemical analysis of CYP450 was performed for control purposes using a specific primary monoclonal antibody (clone 1A2, Abcam, MA) and evaluated by the same four-grade semiquantitative scoring system adopted for CYP27A1 and CYP2R1 assessment.

Antiviral Treatment Schedule and Outcomes.

Patients were treated with pegylated interferon α-2a (Pegasys, Roche, Basel, Switzerland) 180 μg /week or pegylated interferon 2b (Peg-Intron, Schering) 1.5 μg/kg/week plus ribavirin at a dosage of 1000 or 1200 mg/day according to body weight (1000 mg/day for a body weight of <75 kg, 1200 mg/day for a body weight of >75 kg) for 48 weeks. Patients were withdrawn from treatment if they did not achieve a virological response, defined as undetectable serum HCV RNA by polymerase chain reaction, within 24 weeks after start of treatment.24 Sustained virological response (SVR) was defined as negative serum HCV RNA at polymerase chain reaction 6 months after stopping antiviral therapy.

Statistics.

Continuous variables were summarized as mean ± standard deviation, and categorical variables as frequency and percentage. The Student t test and analysis of variance were used when appropriate. Multiple linear regression analysis was performed to identify independent predictors of 25(OH)D serum levels as a continuous dependent variable. As candidate risk factors for low serum levels of 25(OH)D, we selected age, sex, body mass index, waist circumference, baseline ALT, platelet count, GGT, ferritin, total cholesterol, high-density lipoprotein cholesterol, triglycerides, blood glucose, insulin, homeostasis model assessment score, diabetes, arterial hypertension, HCV-RNA levels, steatosis, and activity score.

Multiple logistic regression models were used to assess the relationship of both fibrosis and SVR to the demographic, metabolic, and histological characteristics of patients. In the first model, the dependent variable was severe fibrosis coded as 1 = F3 to F4 in the fibrosis score versus 0 = F1 to F2. Because fibrosis grade is nonlinear, we also performed ordinal logistic regression with fibrosis F0 to F4 as the dependent variable. In the second model, the dependent variable was SVR coded 1 = present versus 0 = absent. As candidate risk factors we selected the same independent variables included in the 25(OH)D model and added 25(OH)D serum levels as an additional independent variable. In this model, we included all patients who received at least one dose of pegylated interferon (intention-to-treat analysis).

Variables associated with the dependent variable at univariate analyses (probability threshold, P ≤ 0.10) were included in the multivariate regression models.25 Regression analyses were performed by SAS.

Results

Patient Features and Histology.

The baseline features of the 197 patients are shown in Table 1. Most of our patients were in the overweight to obesity range. One patient in four had fibrosis of at least 3 by Scheuer score, with a high prevalence of moderate/severe necroinflammation (grading 2-3). Half of the cases had histological evidence of steatosis, though of moderate/severe grade in only 23 cases (11.7%).

Table 1. Baseline Demographic, Laboratory, Metabolic, and Histological Features of 197 Patients with Chronic Hepatitis C
VariableGenotype 1 Chronic Hepatitis C (n = 197)
  1. IU, international units; HOMA, homeostasis model assessment; HDL, high-density lipoprotein; HCV-RNA, hepatitis C virus ribonucleic acid. Data are given as mean ± standard deviation or as number of cases (%).

Mean age, years52 ± 12
Sex 
 Male/female104 (53)/93 (47)
Mean body mass indexkg/m227 ± 4
Body mass indexkg/m2 
 <2551 (25)
 25–29.9109 (55)
 ≥3037 (20)
Waist circumference, cm93 ± 11
Arterial hypertension 
 Absent/present155 (79)/42 (21)
Type 2 diabetes
 Absent/present187 (95)/10 (5)
Alanine aminotransferase, IU/L88 ± 71
Platelet count × 103/mmc155 ± 100
Gamma glutamyl transferase, IU/L60 ± 51
Cholesterol, mg/dL177 ± 37
HDL cholesterol, mg/dL54 ± 17
Triglycerides, mg/dL104 ± 65
Ferritin, ng/mL238 ± 215
Blood glucose, mg/dL93 ± 25
Insulin, μU/mL12 ± 6
HOMA score3.0 ± 1.9
25(OH)D μg/L25.0 ± 9.9
Serum 25-hydroxyvitamin D <30 μg/L 
 No53 (27)
 Yes144 (73)
HCV-RNAIU/mL × 103900 ± 367
Histology at biopsy 
 Steatosis: Continuous variable (percentage of total cells) Categorical variable10 ± 15
  <5%95 (48)
  ≥5% to <30%79 (40)
  ≥30%23 (12)
 Stage of fibrosis 
  145 (23)
  297 (49)
  337 (19)
  418 (9)
Grade of inflammation 
  136 (18)
  2122 (62)
  339 (20)

The control subjects (25 women and 24 men, mean age of 53.7 ± 12.8 years) were comparable for body mass index with the HCV population (26.1 ± 3.5 kg/m2). Six had arterial hypertension.

Serum 25(OH)D Levels.

Mean serum values of 25(OH)D in G1 CHC patients were significantly lower than in controls (25.1 ± 9.9 μg/L versus 43.1 ± 10.2 μg/L; P < 0.0001; Fig. 1). Accordingly, 25(OH)D serum levels of less than 30 μg/L were found in 144 (73%) G1 CHC patients and in only three control subjects (6%, P < 0.001). Advanced age (P = 0.004), female sex (P < 0.001), high waist circumference (P < 0.06), high ALT (P = 0.09), and low high-density lipoprotein levels (P = 0.01), the severity of necroinflammatory activity (P = 0.01), and the severity of fibrosis (P < 0.001) were associated with lower 25(OH)D levels in G1 CHC, though only female sex (P = 0.007), the severity of necroinflammatory activity (P = 0.04), and the severity of fibrosis (P = 0.009) were independent factors in multiple linear regression analysis (Table 2). Figure 1 also shows the distribution of serum 25(OH)D levels in relation to sex. A significant difference was observed in CHC patients (27.60 ± 9.39 μg/L in men versus 22.23 ± 9.77 μg/L in women; P = 0.0001) (Fig. 1). Figure 2 shows the distribution of 25(OH)D according to necroinflammatory activity.

Figure 1.

25(OH)D serum levels in healthy controls and in patients with genotype 1 chronic hepatitis C. The figure also shows the distribution of 25(OH)D serum levels according to sex in both chronic hepatitis C patients and healthy controls.

Table 2. Univariate and Multivariate Analysis of Factors Associated with 25(OH)D as Continuous Variable in 197 Patients with Genotype 1 Chronic Hepatitis C, by Linear Regression Model
VariableUnivariate AnalysisMultivariate Analysis
βSEP ValueβSEP Value
  1. β, β coefficient; SE, standard error of β; IU, international units; HOMA, homeostasis model assessment; HDL, high-density lipoprotein; HCV-RNA, hepatitis C virus ribonucleic acid.

Mean age, years−0.2060.0570.004−0.0520.0600.49
Male sex−0.2711.367<0.001−0.2421.7420.007
Mean body mass index, kg/m2−0.0120.1730.87  
Waist circumference, cm0.1360.0640.060.0400.0670.59
Alanine aminotransferase, IU/L0.1220.0100.090.1360.0100.06
Platelet count × 103/mmc0.0490.0010.49  
Gamma glutamyl transferase, IU/L0.0590.0140.40  
Cholesterol, mg/dL−0.0960.0190.18  
HDL cholesterol, mg/dL−0.1740.0430.01−0.0610.0480.46
Triglycerides, mg/dL0.0210.0110.77  
Ferritin, ng/mL−0.0020.0030.98  
Blood glucose, mg/dL−0.0740.0280.30  
Insulin, μU/mL−0.0830.1080.24  
HOMA score−0.0840.3630.24  
Diabetes−0.0891.0510.21  
Arterial hypertension−0.0350.9490.62  
HCV-RNA–IU/mL × 103−0.1021.0000.36  
Histology at biopsy      
 Steatosis−0.0850.0470.23  
 Grade of inflammation−0.1781.1310.01−0.1621.2620.04
 Stage of fibrosis−0.2780.780<0.001−0.2160.9450.009
Figure 2.

Distribution of 25(OH)D serum levels according to necroinflammatory activity.

Interestingly, women 55 years of age or older had lower 25(OH)D serum levels than their counterparts of younger than 55 years (20.2 ± 9.5 versus 24.5 ± 9.7; P = 0.03). Conversely, in men we observed no difference in 25(OH)D serum levels between patients 55 or older and younger than 55 years of age (26.72 ± 9.08 versus 28.52 ± 9.72; P = 0.33). To account for possible interaction between sex and age, a term for the product of the two variables was included in the linear multivariate model, and showed that the interaction between the two risk factors was significant (P = 0.002).

Considering 25(OH)D as a categorical variable, low vitamin D levels (<30 μg/L) were independently associated with high necroinflammatory activity (odds ratio [OR], 1.99; 95% confidence interval [CI], 1.16–3.42, P = 0.01) and with the interaction term between sex and age (OR, 1.015; 95%CI, 1.005–1.026, P = 0.005).

CYP27A1 and CYP2R1 Liver Evaluation.

In a random sample of 34 patients (19 men [55%]; mean age, 50 ± 12.7 years; 13 (38%) with severe fibrosis; 23 (67%) with moderate-severe necroinflammatory activity; mean 25(OH)D levels 25.96 ± 9.90 μg/L), with baseline features not significantly different from the entire group (data not shown), we evaluated the immunohistochemical expression of CYP27A1 and CYP2R1 with a four-grade semiquantitative scoring system. The same analysis was performed in eight control samples from subjects, without liver disease, who underwent cholecystectomy.

CYP27A1 was expressed, with a score of 3 in 75% (6/8) of controls versus 0% (0/34) of cases (P < 0.001), with a score of 2 in 25% (2/8) of controls versus 12% (4/34) of cases (P = 0.42), with a score of 1 in 0% (0/8) of controls versus 25% (12/34) of cases (P = 0.10), and with a score of 0 in 0% (0/8) of controls versus 53% (18/34) of cases (P = 0.04). Similarly, CYP2R1 was expressed with a score of 3 in 50% (4/8) of controls versus 0% (0/34) of cases (P < 0.001), with a score of 2 in 50% (4/8) of controls versus 15% (5/34) of cases (P = 0.10), with a score of 1 in 0% (0/8) of controls versus 50% (17/34) of cases (P = 0.05), and with a score of 0 in 0% (0/8) of controls versus 35% (12/34) of cases (P = 0.10). According to these data, the overall expression of both CYP27A1 and CYP2R1 was significantly down-modulated (P = 0.0001 for both CYP27A1 and CYP2R1) in G1 CHC samples (Supporting Document 1).

The degree of expression of CYP2R1 proved to be neither significantly related to 25(OH)D serum levels nor associated with biochemical, anthropometric, and histological features. Conversely, a significant association was found between a decreased expression of CYP27A1 and low 25(OH)D serum levels (P = 0.01) (Fig. 3A). Moreover, CYP27A1 expression was negatively associated with the degree of necroinflammatory activity (P = 0.031) (Fig. 3B). No significant associations were found between CYP27A1 expression and the various biochemical, anthropometric, and histological features, other than inflammation grade (data not shown).

Figure 3.

Distribution of 25(OH)D serum levels according to immunohistochemical score of CYP27A1 liver expression. (A): 25(OH)D serum levels significantly increase (P = 0.01), from patients with score 0 (22.04 ± 8.96; n = 18), to those with scores of 1 (28.88 ± 8.06; n = 12) and 2 (30.78 ± 5.73; n = 4). Distribution of necroinflammatory activity severity according to immunohistochemical score of CYP27A1 liver expression. (B) Among the 18 patients with score 0, six had grading 3, nine grading 2, and three grading 1; among the 12 patients with score 1, four had grading 3, four grading 2, and four grading 1; and among the four patients with score 2, all had grading 1. This inverse relation between immunohistochemical CYP27A1 score and severity of necroinflammatory activity was statistically significant (P = 0.031).

Case and control samples showed an overlapping picture because CYP450 proved to be diffusely expressed with a moderate intensity in both groups (Supporting Document 2). This suggests a maintained expression of cytochromes other than those involved in vitamin D metabolism in G1 CHC samples.

Variables Related to Severe Fibrosis.

The univariate and multivariate comparison of variables between patients with and without severe fibrosis (F3–F4) are reported in Table 3. Older age, male sex, low platelet count, high baseline values of ALT, high GGT, low cholesterol, high ferritin, low 25(OH)D, steatosis, and high necroinflammatory activity were associated with severe fibrosis (P < 0.10). Multivariate logistic regression analysis showed that the following features were independently linked to severe fibrosis (Scheuer score ≥3): older age (OR, 1.043; 95% CI, 1.002–1.085, P = 0.03), low cholesterol (OR, 0.981; 95%CI, 0.969–0.992, P = 0.001), high ferritin (OR, 1.003; 95%CI, 1.001–1.005, P = 0.007), low 25(OH)D (OR, 0.942; 95%CI, 0.893–0.994, P = 0.02), and high necroinflammatory activity (grading) (OR, 2.235; 95%CI, 1.014–4.929; P = 0.04). The overall area under the curve (AUC) of this model was good (AUC, 0.870). Figure 4, showing the distribution of 25(OH)D serum levels according to the stage of fibrosis, highlighted a significant trend in 25(OH)D reduction among the four fibrosis stages (P < 0.0001). However, even patients with fibrosis F1 had mean serum 25(OH)D levels significantly lower than control subjects (29.5 ± 10.9 μg/L versus 43.1 ± 10.2 μg/L; P < 0.0001).

Table 3. Univariate and Multivariate Analysis of Risk Factors Associated with Severe Fibrosis (F3-F4) in 197 Patients with Genotype 1 Chronic Hepatitis C, by Logistic Regression Analysis
VariableNo Severe Fibrosis (Scheuer Score 1–2) n = 142Severe Fibrosis (Scheuer Score 3–4) n = 55Univariate Analysis P ValueMultivariate Analysis OR (95% CI)P Value
  1. IU, international units; HOMA, homeostasis model assessment; HDL, high density lipoprotein; HCV-RNA, hepatitis C virus ribonucleic acid. Data are given as mean ± standard deviation or as number of cases.

Age, years50 ± 1357 ± 100.0011.043 (1.002–1.085)0.03
Sex     
 Male versus female69/7335/200.050.686 (0.258–1.826)0.45
Body mass index, kg/m227 ± 427 ± 30.45 
Waist circumference, cm93 ± 1193 ± 100.60 
Alanine aminotransferase, IU/L81 ± 70109 ± 720.021.003 (0.997–1.009)0.38
Platelet count × 103/mmc212 ± 56178 ± 54<0.0011.000 (1.001–1.003)0.19
Gamma glutamyl transferase, IU/L53 ± 4979 ± 540.0021.004 (0.996–1.012)0.29
Cholesterol, mg/dL183 ± 34162 ± 41<0.0010.981 (0.969–0.992) 0.001 
HDL cholesterol, mg/dL55 ± 1750 ± 160.07 
Triglycerides, mg/dL106 ± 7399 ± 380.51 
Ferritin, ng/mL189 ± 154368 ± 290<0.0011.003 (1.001–1.005) 0.007 
Blood glucose, mg/dL92 ± 2696 ± 210.36 
Insulin, μU/mL13 ± 713 ± 50.86 
HOMA score3.0 ± 2.13.0 ± 1.50.91 
Arterial hypertension     
 Absent versus present112/3043/120.91 
Type 2 diabetes     
 Absent versus present137/550/50.11 
25(OH)D μg/L26.3 ± 10.021.9 ± 8.80.0050.942 (0.893–0.994)0.02
HCV-RNA, IU/mL × 103987 ± 1066874 ± 4770.44 
Histology at biopsy     
 Steatosis9 ± 1414 ± 160.030.996 (0.971–1.021)0.74
 Grade of inflammation 1/2/334/90/182/32/21<0.0012.235 (1.014–4.929)0.04
Figure 4.

Distribution of 25(OH)D serum levels according to different stages of fibrosis.

Excluding steatosis and grading from the model, older age (OR, 1.043; 95%CI, 1.004–1.083; P = 0.03), cholesterol (OR, 0.980; 95%CI, 0.968–0.991; P = 0.001), and ferritin (OR, 1.003; 95%CI, 1.001–1.005; P = 0.004) were the noninvasive predictors of severe fibrosis. The overall AUC of this model was similarly good (AUC, 0.854).

Comparing patients with significant fibrosis (F2–F4) with subjects with no significant fibrosis (F1), we confirmed that low serum 25(OH)D levels were independently related to significant fibrosis (data not shown).

The model having fibrosis as an ordinal dependent variable by multiple logistic regression analysis included older age (P = 0.001), low platelets (P = 0.03), low cholesterol (P = 0.001), high ferritin (P = 0.007), low 25(OH)D (P = 0.0006), and high necroinflammatory activity (=P = 0.0002).

Factors Associated with SVR.

One hundred sixty-seven patients underwent and completed the antiviral treatment program. SVR was achieved in 70 individuals (41.9%). Among 97 patients (58.1%) who did not achieve SVR, nine were lost to follow-up, and 14 withdrew from antiviral therapy because of side effects. High GGT, low cholesterol, low 25(OH)D, greater steatosis, and severe necroinflammatory activity were associated with lack of SVR (P < 0.10) (Table 4). By logistic regression, low cholesterol (OR, 1.009; 95%CI, 1.000–1.018; P = 0.04), low 25(OH)D levels (OR, 1.039; 95%CI, 1.002–1.077; P = 0.03), and greater steatosis (OR, 0.971; 95%CI, 0.944–0.999; P = 0.04) were the only independent negative predictors of SVR.

Table 4. Univariate and Multivariate Analysis of Risk Factors Associated with Sustained Viral Response (SVR) in 167 Patients with Genotype 1 Chronic Hepatitis C, by Logistic Regression Analysis
VariableNo Sustained Viral Response N = 97Sustained Viral Response N = 70Univariate Analysis P ValueMultivariate Analysis OR (95% CI)P Value
  1. IU, international units; HOMA, homeostasis model assessment; HDL, high-density lipoprotein; HCV-RNA, hepatitis C virus ribonucleic acid. Data are given as mean ± standard deviation or as number of cases.

Age, years54 ± 1252 ± 110.27 
Sex     
 Male versus female47/5043/270.10 
Body mass index, kg/m227 ± 428 ± 50.41 
Waist circumference, cm93 ± 1194 ± 110.53 
Alanine aminotransferase, IU/L88 ± 6379 ± 480.29 
Platelet count × 103/mmc205 ± 62200 ± 490.59 
Gamma glutamyl transferase, IU/L68 ± 5650 ± 450.020.995 (0.988–1.003)0.19
Cholesterol, mg/dL173 ± 33184 ± 410.061.009 (1.000–1.018)0.04
HDL cholesterol, mg/dL53 ± 1651 ± 140.30 
Triglycerides, mg/dL110 ± 8495 ± 350.15 
Ferritin, ng/mL241±218232±2090.79 
Blood glucose, mg/dL95 ± 2791 ± 180.32 
Insulin, μU/mL13 ± 613 ± 70.88 
HOMA score3.1 ± 1.82.9 ± 1.80.62 
Arterial hypertension     
 Absent versus present75/2256/140.67 
Type 2 diabetes     
 Absent versus present92/569/10.20 
25(OH)D μg/L23.7 ± 9.226.6 ± 9.60.051.039 (1.002–1.077)0.03
HCV-RNA, IU/mL × 103974±1,108910±1,0070.80 
Histology at biopsy     
 Steatosis13 ± 187 ± 110.010.971 (0.944–0.999)0.04
 Grade of inflammation 1/2/310/70/1723/33/140.030.816 (0.466–1.429)0.47
 Stage of fibrosis 1/2/3/421/47/20/920/34/10/60.29 

A per protocol analysis of 144 patients confirmed that low cholesterol (OR, 1.012; 95% CI 1.002–1.022; p=0.02) and low 25(OH)D levels (OR, 1.048; 95%CI, 1.008–1.080; P = 0.02), as well as greater steatosis (OR, 0.970; 95%CI, 0.941–1.000; P = 0.04), were negative independent predictors of SVR.

Discussion

We have shown that the biochemical profile of G1 CHC patients is characterized by lower-than-normal serum 25(OH)D levels, and that a low 25(OH)D level is independently related to severe fibrosis and a low likelihood of SVR after standard-of-care antiviral therapy.

Lower levels of serum 25(OH)D have been previously reported in populations heterogeneous for cause and severity of chronic liver disease.11, 19 We confirmed a 25(OH)D reduction in a homogeneous cohort of patients with G1 CHC, at low prevalence of F4 fibrosis. Although a significant trend in 25(OH)D levels reduction was observed with increasing stage of fibrosis, a significant reduction was also observed in the subgroup of patients with mild fibrosis (F1), making it unlikely that low 25(OH)D levels could be entirely explained by reduced liver function.

Our study shows that low 25(OH)D levels are independently associated with female sex and with severity of necroinflammatory activity. Although the study was not designed to clarify the correlation between female sex and lower 25(OH)D levels, because of the observed reduction in women older than 55 years, but not in men of the same age range, and because of the significant interaction between sex and age, we can speculate that hormonal alterations in postmenopausal women likely modulate the vitamin D status. Our results also underline an inverse relationship between 25(OH)D and the severity of necroinflammatory activity. The cross-sectional design of our study is unable to dissect the temporal relation between changes in 25(OH)D and necroinflammation. However, CYP27A1 liver expression was directly related to serum 25(OH)D levels, and inversely associated with the severity of necroinflammatory activity. Therefore, the hepatic necroinflammatory activity caused by the HCV infection could be responsible for 25 (OH)D levels reduction by different mechanisms, such as a selectively reduced liver expression of enzymes involved in liver hydroxylation of vitamin D3.

This study also offers the first evidence that low 25(OH)D serum levels, together with known risk factors for fibrosis severity, such as older age, low cholesterol levels, and high necroinflammatory activity,26 are independently associated with the presence of severe fibrosis. We were not able to confirm IR as a risk factor for fibrosis severity, as reported by others.26, 27 The lack of this association could be attributable to differences in the mean age, alcohol use, and prevalence of obesity and diabetes. In any case we confirmed the important effect of metabolic alterations in fibrosis severity, providing evidence for an association between severe fibrosis and high ferritin levels, a surrogate marker of IR and metabolic syndrome.28

Interestingly, considering only noninvasive parameters, the AUC of the model that includes vitamin D levels to predict severe fibrosis remains good. This suggests the potential use of serum 25(OH)D levels as a noninvasive marker of liver fibrosis, a use that needs to be tested and validated in large prospective cohort studies in patients with CHC of all genotypes, and in chronic liver disease of other origins.

It is conceivable that a reduced 25(OH)D level might by itself favor progression of fibrosis. Different experimental models showed that vitamin D, via interaction with vitamin D receptor, protects against oxidative stress production,29 can influence the migration, proliferation, and gene expression of fibroblasts,30, 31 and reduces the inflammatory and fibrogenic activity of liver stellate cells.32, 33 However further prospective cohort studies will be needed to determine the causal association between vitamin D deficiency and fibrosis in patients with CHC.

Another interesting finding of this study is the evidence that lower 25(OH)D serum levels were an independent negative risk factor for SVR. Again, this observation will require further validation in different, large cohorts of patients, but is supported by experimental data that suggest a role of vitamin D in the modulation of the immune response,32, 33 and by recent clinical data36 reporting higher early virological response rate in CHC treated with standard of care plus vitamin D, compared with those treated with standard of care only. Finally, in line with data from the literature, we found that steatosis35 and lower cholesterol levels, a known surrogate marker of fibrosis severity,26 were independently associated with lower SVR rate. We did not find any association between IR and SVR, in keeping the conflicting data reported in the literature on the role of IR as a predictor of SVR.36

The main limitation of this study lies in its cross-sectional nature and its inability to dissect the temporal relation between 25(OH)D and fibrosis. A further methodological drawback is the potentially limited external validity of the results for different populations and settings. Another limitation of this study is the lack of data on the potential confounders that may influence the levels of vitamin D, such as exposure to sunshine, dietary intake, and the prevalence of osteoporosis. However, all of the subjects involved in this study lived in Sicily, where sunshine is abundant, even if we cannot rule out differences in sun exposure between healthy controls and CHC patients. However, data on the prevalence of osteoporosis were not available, nor was the dietary intake of vitamin D, which, however, remains a rather crude measure of vitamin D status because of recall bias. The lack of these data in both cases and controls makes a systematic error very unlikely. Also, data on osteoporosis were not available in both groups. Lack of data on polymorphisms of vitamin D hydroxylating enzymes, and on other variables involved in vitamin D metabolism, such as parathyroid hormone, and in vitamin D signaling regulation also could affect the interpretation of our findings.

In conclusion, this study, showing low levels of serum 25(OH)D and of their liver hydroxylating enzymes in G1 CHC patients, and suggesting a relation of vitamin D status with the severity of liver disease and response to therapy, opens a new area of research on the potential use of vitamin D in patients with CHC.

Acknowledgements

The authors thank Warren Blumberg for his help in editing this article.

Ancillary