The association between serological and dietary vitamin D levels and hepatitis C-related liver disease risk differs in African American and white males

Authors

  • D. L. White,

    Corresponding author
    1. Section of Gastroenterology and Hepatology, Department of Medicine, Baylor College of Medicine and Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    2. Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX, USA
    • Clinical Epidemiology and Outcomes Program, Houston VA Health Services Research and Development Center of Excellence, Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    Search for more papers by this author
  • S. Tavakoli-Tabasi,

    1. Hepatitis C Clinic, Section of Infectious Diseases, Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    2. Section of Infectious Diseases, Department of Medicine, Baylor College of Medicine, Houston, TX, USA
    Search for more papers by this author
  • F. Kanwal,

    1. Clinical Epidemiology and Outcomes Program, Houston VA Health Services Research and Development Center of Excellence, Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    2. Section of Gastroenterology and Hepatology, Department of Medicine, Baylor College of Medicine and Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    3. Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX, USA
    Search for more papers by this author
  • D. J. Ramsey,

    1. Clinical Epidemiology and Outcomes Program, Houston VA Health Services Research and Development Center of Excellence, Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    2. Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX, USA
    Search for more papers by this author
  • A. Hashmi,

    1. Section of Gastroenterology and Hepatology, Department of Medicine, Baylor College of Medicine and Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    Search for more papers by this author
  • J. Kuzniarek,

    1. Clinical Epidemiology and Outcomes Program, Houston VA Health Services Research and Development Center of Excellence, Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    2. Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX, USA
    Search for more papers by this author
  • P. Patel,

    1. Clinical Epidemiology and Outcomes Program, Houston VA Health Services Research and Development Center of Excellence, Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    2. Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX, USA
    Search for more papers by this author
  • J. Francis,

    1. Section of Infectious Diseases, Department of Medicine, Baylor College of Medicine, Houston, TX, USA
    Search for more papers by this author
  • H. B. El-Serag

    1. Clinical Epidemiology and Outcomes Program, Houston VA Health Services Research and Development Center of Excellence, Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    2. Section of Gastroenterology and Hepatology, Department of Medicine, Baylor College of Medicine and Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, USA
    3. Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX, USA
    Search for more papers by this author

Correspondence to:

Dr D. L. White, Michael E. DeBakey VA Medical Center, 2002 Holcombe Blvd. (MS 152), Houston, TX 77030, USA.

E-mail: dwhite1@bcm.edu

Summary

Background

Vitamin D may affect the severity of HCV-related liver disease.

Aim

To examine the association between serum vitamin D levels and advanced liver disease in a multiethnic US cohort of HCV patients, and account for dietary and supplemental intake.

Methods

We measured serum 25-hydroxyvitamin D levels and used FibroSURE-ActiTest to assess hepatic pathology in a cohort of HCV-infected male veterans. We estimated and adjusted for daily intake of vitamin D from diet using a Dietary History Questionnaire, and dispensed prescriptions prior to study enrolment. We used race-stratified logistic regression analyses to evaluate the relationship between serum vitamin D levels and risk of advanced fibrosis (F3/F4–F4) and advanced inflammation (A2/A3–A3).

Results

A total of 163 African American (AA) and 126 White non-Hispanics were studied. Overall, ~44% of AAs and 15% of Whites were vitamin D deficient (<12 ng/mL) or insufficient (12–19 ng/mL); 4% of AAs and 9% of White patients had an elevated level (>50 ng/mL). Among AAs, patients with elevated serum vitamin D levels had significantly higher odds of advanced fibrosis (OR = 12.91, P = 0.03) than those with normal levels. In contrast, AAs with insufficient or deficient levels had > two-fold excess risk of advanced inflammation (P = 0.06). Among White males there was no association between vitamin D levels and advanced fibrosis (F3/F4–F4) or inflammation (A2/A3–A3) risk.

Conclusions

We observed potential differences in the association between vitamin D levels and degree of HCV-related hepatic fibrosis between White and African American males. Additional research is necessary to confirm that high serum vitamin D levels may be associated with advanced fibrosis risk in African American males, and to evaluate whether racial differences exist in HCV-infected females.

Background

Vitamin D is an essential fat-soluble micronutrient in mammals. In addition to playing a key role in maintenance of skeletal health, there is increasing recognition that vitamin D is involved in multiple other important physiological functions including immune response and wound healing.[1] The liver plays a critical role in vitamin D metabolism by hydroxylating vitamin D synthesised endogenously secondary to sun exposure (as Vitamin D3) or else obtained directly from dietary and supplemental sources (as Vitamin D2 or D3) into its key metabolic intermediate, 25-hydroxyvitamin D or 25(OH)D. Exported from the liver bound to Vitamin D Binding Protein, 25-hydroxyvitamin D is then further hydroxylated (primarily by the kidneys) to its bioactive form, 1, 25(OH)2D which is the target ligand for the nuclear hormone vitamin D receptor.

Vitamin D may affect the severity of HCV-related liver disease. The vitamin D3 intermediate hydroxyvitamin-D3 has been shown to reduce HCV core antigen levels in a dose-dependent manner in HCV-infected HuH-7 cell lines.[2] Several[3-6] although not all[7, 8] epidemiological studies performed in HCV-infected populations undergoing treatment have shown that lower serum hydroxyvitamin D levels are associated with reduced likelihood of achieving a sustained virological response (SVR) with interferon and ribavirin combination antiviral treatment. Findings of improved SVR were also reported in some[9, 10] although not in all[11] studies of patients receiving Vitamin D supplementation.

An association between lower serum vitamin D levels and increased risk of advanced HCV-related liver fibrosis has also been shown in several[3, 5, 12] although not in all[7, 8] case–control and cross-sectional studies, many performed in European populations. The only nested case–control study was performed among prospectively recruited participants from the multicentre HALT-C treatment trial in the US.[13] Although no association was found between fibrosis progression and serum hydroxyvitamin D levels overall, a potentially increased fibrosis progression risk was observed with higher baseline levels in certain sub-groups, including diabetics and also possibly African Americans (AA) and individuals using vitamin D supplements.

Few studies have been performed in ethnically diverse general clinic populations. Another smaller US cross-sectional study in HCV patients who received antiviral therapy reported decreased risk of advanced fibrosis with higher serum D levels only in Caucasians.[14] Therefore, the association between serum vitamin D levels and the risk of hepatic fibrosis in the general HCV patient population, many of whom are contraindicated for antiviral therapy, remains unclear. Furthermore, none of these studies evaluated customary dietary intake as a possible explanatory co-factor. Some recent research suggests vitamin D2, which is obtained in many fortified foods and supplements, may not be as bioavailable or effective as vitamin D3, which is obtained endogenously or from a limited number of foods like cod liver oil.[15] Finally, the association between serum D levels and hepatic inflammatory activity risk is unclear.

To address the gaps in the HCV– vitamin D literature, and with a working hypothesis that low serum vitamin D is associated with more advanced liver disease in HCV-infected patients, we performed a cross-sectional study to determine if there were potential racial differences in the association between serologically determined total vitamin D levels and advanced liver disease in a large and multiethnic HCV patient cohort not on antiviral therapy in the US. Our study also included a novel exploratory evaluation of how customary dietary and also supplemental intake of vitamin D influence observed associations between total serum vitamin D levels and advanced liver disease risk.

Methods

Study design and population

We performed a cross-sectional study in chronically hepatitis C-infected veterans who were seen at the Michael E. DeBakey VA Medical Center (MEDVAMC) in Houston, Texas. We prospectively recruited consecutive veterans with a confirmed or possible HCV diagnosis at the dedicated Hepatitis C clinic at MEDVAMC prior to a routine clinic appointment between (January 5, 2009–January 4, 2012). All patients with a possible HCV are scheduled for a mandatory education and clinical evaluation in the HCV clinic in accordance with VA policy. To help ensure the internal validity and reliability of our findings, veterans were eligible for our current study analysis if they met the following criteria: (i) they were self-identified African American (AA) or White non-Hispanic males aged 18–70 years at recruitment as they comprised almost 90% of the underlying patient population; (ii) had neither self- nor medical record-reported history of liver transplant, decompensated liver disease, hepatocellular carcinoma, or gastric bypass surgery; (iii) had serologically confirmed HCV viremia and concomitantly tested negative for both HIV and hepatitis B virus (HBV) surface antigen; and (iv) were not currently receiving anti-viral therapy as <0.5% were receiving therapy at the time of study recruitment. This study was approved jointly by the Institutional Review Boards for Baylor College of Medicine and the Michael E. DeBakey VA Medical Center.

Data collection and study measures

Veterans completed a research assistant-administered comprehensive risk factor and dietary survey, had anthropometric measures taken, completed a fasting blood draw for serology, and consented to a comprehensive VA electronic medical record and pharmacy utilisation review at recruitment.

Serological measures

A fasting blood draw was performed at recruitment and was analysed by the MEDVAMC's CLIA-certified Central Laboratory Service for: HCV antibodies, genotype and quantitative viral load.

Baseline serum vitamin D level was assessed by measuring serum 25-hydroxyvitamin D levels (combined 25-hydroxyvitamin D2 and 25-hydroxyvitamin D3 levels) as it is recognised as the only clinically useful measure[16] as well as being a stable long-term predictor of vitamin D status.[17] We used DiaSorin's validated LIAISON automated immunochemiluminometric assay (ICMA),[18] with levels being considered clinically ‘normal’ if between 20 and 50 ng/mL, ‘deficient’ if between 12 and 19 ng/mL, ‘insufficient’ if below 12 ng/mL and ‘elevated’ if above 50 ng/mL based on Institute of Medicine (IOM) guidelines.[19] We created a dichotomous variable for season as a proxy variable for potential direct sunlight exposure classified as Winter if the blood was drawn during the 4 months with the shortest daylight hours (November–February), or otherwise as Non-Winter.

Surveys

Trained research assistants (RAs) administered a computerised sociodemographic, clinical and risk factor survey, which interrogated self-reported race/ethnicity, lifetime history of alcohol use and presence of comorbid diagnoses including diabetes.

Dietary and nutritional supplemental vitamin D intake measurement

We estimated average annual daily intake of vitamin D from diet and from nutritional supplements (combined dietary oral intake of Vitamin D2 and D3) using participant's self-reported responses to the National Cancer Institute's (NCI) validated Dietary History Questionnaire (NCI-DHQ).[20] All dietary analyses were restricted to the ~77% of participants who returned surveys with average daily caloric intake between 500 and 5000 kcal/day.

Electronic medical record (EMR) review

A single physician (HA) performed structured EMR review to determine the receipt of either out-patient or in-patient prescriptions for vitamin D by the VA pharmacy, including for multivitamins or ergocalciferol (Vitamin D2) prescribed alone or in conjunction with calcium. Information on dose, duration, and timing, including whether the patient had received a VA-dispensed vitamin D prescription within the year prior to study enrolment, was also collected.

FibroSURE-ActiTest

We estimated degree of hepatic fibrosis and inflammation using the FibrosSURE(FibroTest)-ActiTest (FS-AT). The FS-AT has been validated against hepatic biopsy in multiple study populations including in HCV+ populations.[21-24] It provides METAVIR biopsy-based equivalent hepatic fibrosis (F0, no fibrosis present – F4, cirrhosis) and inflammatory activity (A0, no inflammatory activity – A3, severe inflammatory activity). We classified study participants as advanced fibrosis cases if the FS-AT is F3/F4–F4 (cirrhosis), or else as mild fibrosis controls if F0–F3; and as advanced inflammatory activity cases if A2/A3–A3 (severe activity), or else as mild activity controls if A0–A2.

Statistical analyses

All analyses were race-stratified. We compared serum hydroxyvitamin D levels (as well as sociodemographic, dietary, and clinical characteristics) in advanced fibrosis cases (F3/F4–F4) vs. mild fibrosis controls (F0–F3) and in advanced inflammation cases (A2/A3–A3) vs. mild inflammation controls (A0–A2) using T-tests for means and the Brown-Mood test for medians respectively, and the χ2 and Fisher's exact test for large and small sample categorical variables respectively.

We performed binary logistic regression to evaluate the association between serum hydroxyvitamin D levels and risk of both advanced hepatic fibrosis (F3/F4-F4 vs. F0-F3) and advanced inflammatory activity (A2/A3-A3 vs. A0-A2). Our baseline multivariate models included adjustment for age, history of alcohol abuse, presence of overweight or obesity, viral load, and receipt of a VA-dispensed vitamin D therapy during the year prior to baseline blood draw. Season of blood draw (Winter/non-Winter) was not included in multivariate analyses because it was not significantly associated with either serum vitamin D levels or advanced inflammatory activity or fibrosis risk in Whites or AAs in any analyses.

To assess how self-reported vitamin D intake influenced observed associations between serum hydroxyvitamin D and risk of advanced liver disease, we performed two sets of exploratory multiple logistic regression analyses that included additional adjustment of our baseline multivariate model for either DHQ-determined average daily dietary intake of vitamin D or for average annual total vitamin D intake combining both dietary and nutritional supplement intake. We also performed sensitivity analyses where we recalculated risk estimates after removing all participants on VA-prescribed vitamin D therapy.

All model-based parameter estimates are reported as odds ratios with associated 95% confidence intervals, with all analyses conducted using SPSS 18.

Results

A total of 289 male veterans with chronic HCV (163 African American (AA) and 126 White non-Hispanic) fulfilled our study eligibility criteria (overall participation rate = 82%). The mean age of AAs was almost two years greater than that of Whites (57.4 vs. 55.7 years, P = 0.005). (Table 1) The proportions of advanced fibrosis (F3/F4–F4) and advanced inflammatory activity (A2/A3–A3) were both nonsignificantly higher in Whites (43% vs. 39% advanced fibrosis prevalence and 39% vs. 26% advanced activity prevalence in White and AA HCV+ males respectively). Median serum 25(OH)D levels were higher in Whites (28.9 ng/mL vs. 21.6 ng/mL in Whites vs. AAs, P < 0.001), although median values for both sub-groups were in the normal range (20–50 ng/mL). (Table 2) Overall, ~44% of AAs and ~15% of White HCV+ males were either vitamin D deficient or insufficient, with 4% of AAs and 9% of Whites with elevated levels.

Table 1. Demographic and clinical characteristics of 289 male veterans with chronic hepatitis C according to race/ethnicity
CharacteristicAfrican American = 163White = 126P-value
  1. a

    Chronic alcohol abuse defined as greater than 3 drinks a day for at least 10 years.

  2. b

    Determined by FibroSURE-ActiTest (FS-AT).

Age in years, mean (s.d.)57.4 (4.2)55.7 (5.5)0.005
Chronic alcohol abusea, n (%)
Negative101 (62.0)55 (43.7)0.004
Positive61 (37.4)70 (55.6) 
Missing1 (0.6)1 (0.7) 
HCV viral genotype, n (%)
1155 (95.1)86 (68.3)<0.001
21 (0.6)19 (15.1) 
30 (0)14 (11.0) 
42 (1.2)0 (0) 
Missing5 (3.1)7 (5.6) 
HCV viral load (log 10 in eq/mL), mean (s.d.)6.53 (0.73)6.41 (0.93)0.23
Diabetes mellitus, n (%)
Absent124 (76.1)94 (74.6)0.77
Present39 (23.9)32 (25.4) 
BMI, mean (s.d.)28.3 (5.04)28.8 (5.7)0.48
Fibrosis stageb
Mild fibrosis (0–3)100 (61.3)72 (57.1)0.55
Advanced fibrosis (3/4–4)63 (38.7)54 (42.9) 
Inflammatory activity gradeb
Mild activity (0–2)121 (74.2)77 (61.1)0.02
Advanced activity (2/3–3)42 (25.8)49 (38.9) 
Table 2. Vitamin D measures based on serology, prescriptions and dietary and supplement intake in 289 male veterans with chronic hepatitis C according to race/ethnicity
 African American (= 163)White (= 126)P-value
  1. IQR, interquartile range; NCI-DHQ, National Cancer Institute Food Frequency Questionnaire.

  2. a

    Based on Institute of Medicine (IOM) guidelines.

  3. b

    Based on physician-performed electronic medical review for dispensed vitamin D prescriptions including multivitamins for period up to 15 years.

  4. c

    Restricted to the 77% of participants (n = 100 White and n = 123 African Americans) who returned validly completed NCI-DHQ.

Serology-based
Serum vitamin D, median (IQR)21.6 (13.4)28.9 (13.0)<0.001
Serum vitamin D categorya, n (%)
Deficient (<12 ng/mL)16 (9.8)2 (1.6)<0.001
Insufficient (12–<20 ng/mL)55 (33.7)17 (13.5) 
Normal (20–50 ng/mL)86 (52.8)96 (76.2) 
Elevated (>50 ng/mL)6 (3.7)11 (8.7) 
Season blood drawn, n (%)
Non-Winter (March–October)100 (61.3)79 (62.7)0.82
Winter (November–February)63 (38.7)47 (37.3) 
Serum vitamin D, median (IQR)
Non-Winter (March–October)22.0 (13.6)30.0 (13.6)0.46 (White), 0.56 (African American)
Winter (November–February)20.5 (13.5)28.3 (12.3) 
Correlation viral load and vitamin D, Spearman rho0.030.080.18 (White), 0.72 (African American)
Serum vitamin D by diabetes diagnosis history, median (IQR)
Negative21.7 (14.1)28.6 (13.5)0.54 (White), 0.97 (African American)
Positive21.6 (12.4)29.9 (11.2) 
Serum vitamin D by fibrosis category, median (IQR)
Mild fibrosis control (F0–F3)21.8 (12.4)29.3 (12.1)0.59 (White), 0.80 (African American)
Advanced fibrosis case (F3/F4–F4)20.7 (16.1)28.1 (12.3) 
Serum vitamin D by activity category, median (IQR)
Mild activity control (A0–A2)22.7 (13.7)29.2 (12.0)1.0 (White), 0.12 (African American)
Advanced activity case (A2/A3–A3)19.4 (12.8)28.6 (14.1) 
Prescriptions from electronic medical record reviewb
Vitamin D prescription ever in past 15 years by type, n (%)
Multivitamin with Vitamin D214 (8.6)11 (8.7)0.46
Calcium/vitamin D2 combination0 (0)1 (0.8) 
Vitamin D2 alone0 (0)1 (0.8) 
None149 (91.4)113 (89.7) 
Vitamin D prescription in year pre-blood draw, n (%)
Yes7 (4.3)12 (9.5)0.08
No156 (95.7)114 (90.5) 
Duration vitamin D therapy months, median (IQR)30 (63)60 (48)0.71
Dietary and supplement intake prior year based on NCI-DHQc
Daily vitamin D intake (μg/day) dietary sources only, median (IQR)3.64 (3.78)4.01 (4.09)0.20
Daily dietary vitamin D intake (μg/day), n (%)
Tertile 1 (≤2.9)42 (34.1)33 (33.0)0.78
Tertile 2 (>2.9–5.3)42 (34.1)31 (31.0) 
Tertile 3 (>5.3)39 (31.8)36 (36.0) 
Daily vitamin D intake from nutritional supplements only (μg/day), n (%)
None78 (63.4)51 (51.0)0.02
Low supplement use (<10 μg vitamin D per day)36 (29.3)30 (30.0) 
High supplement use (≥10 μg vitamin D per day)9 (7.3)19 (19.0) 

Median serum 25(OH)D levels were minimally lower in the advanced fibrosis cases (F3/F4–F4) and advanced inflammatory activity cases (A2/A3–A3) compared with their respective mild disease controls in bivariate analysis (20.7 vs. 21.8 ng/mL for advanced vs. mild fibrosis and 19.4 vs. 22.7 ng/mL for advanced vs. mild activity in AAs, and 28.1 vs. 29.3 ng/mL for advanced vs. mild fibrosis and 29.2 vs. 28.6 ng/mL for advanced vs. mild activity in Whites). (Table 2) There were no significant associations between serum vitamin D levels and season of blood draw, diabetes status or HCV viral load in either AA or White HCV+ males.

Medical record review indicated that 27 participants (n = 14 AA and n = 13 White) had received a prescription containing vitamin D that was dispensed by the VA pharmacy for a period of up to 15 years. (Table 2) The proportion of those ever receiving a vitamin D prescription was modestly higher in Whites than in AA (10.3% vs. 8.6% in Whites and AAs respectively), while that of current users defined as on prescribed vitamin D therapy during the one year prior to blood draw was substantially higher in Whites (9.5% vs. 4.3% in Whites and AAs respectively, P = 0.08). Vitamin D prescriptions dispensed by the VA pharmacy were typically (~92% of time) for a daily multivitamin/mineral supplement, which contained 400 IU of ergocalciferol (D2) along with 66 mg of calcium. The median duration for vitamin D therapy was 30 months in Whites and 60 months in African Americans, although this difference was not significant. (Table 2) The proportion of HCV+ males receiving a VA-prescribed vitamin D prescription in the prior year was similar among those with high (>50 ng/mL) vs. nonhigh (≤50 ng/mL) serum D in both subgroups (8% vs. 9% in Whites males with high vs. nonhigh serum D levels respectively, and 0% vs. 5% in AA males with high vs. nonhigh serum D levels respectively). (data not shown).

Most participants (n = 223, ~77%) returned completed NCI-DHQ dietary surveys with average dietary intake between 500 and 5000 kcal/day. The DHQ-estimated median daily dietary intake of vitamin D (i.e. from food and beverage intake only) was modestly lower in AAs than in Whites (3.64 vs. 4.01 μg/day in AAs and Whites respectively, P = 0.20). AAs were also significantly less likely to self-report obtaining additional vitamin D via use of nutritional supplements (36.6% vs. 49.0% in AAs and Whites respectively, P = 0.02) (Table 2). However, AAs who had high serum D levels were significantly more likely than those with nonhigh levels (≤50 ng/mL) to report using supplements that contained vitamin D in the prior year (83% vs. 34%, P = 0.01). In contrast, in Whites, there were neither strong nor significant differences (55% reported use of over-the-counter supplements with vitamin D in the prior year vs. only 48% in those with nonhigh serum levels, N.S.). (data not shown)

Race-stratified logistic regression models for the association between serum hydroxyvitamin D levels and advanced hepatic fibrosis risk (F3/F4–F4) are reported in Table 3. In our baseline multivariate analysis, compared with AA HCV+ males with normal levels, those with elevated serum vitamin D levels had highly elevated odds of advanced fibrosis (OR = 12.91, P = 0.03). AA HCV+ males who were hydroxyvitamin D insufficient or deficient (<20 ng/mL) had a nonsignificant elevation in relative risk of advanced fibrosis compared to AA HCV+ males with normal serum levels (OR = 1.76, P = 0.12). The associations observed in Whites differed with no excess advanced fibrosis risk with elevated serum vitamin D levels (OR = 0.84, P = 0.82) and only a slight nonsignificant excess advanced fibrosis risk with insufficient or deficient levels (OR = 1.18, P = 0.78).

Table 3. Race-stratified logistic regression models for association between total serum vitamin D and other measures of vitamin D status and risk of FibroSURE-ActiTest-determined advanced hepatic fibrosis (F3/F4 and F4) in HCV+ males
 Univariate modelsMultivariate model-baselinea
OR95% CIP-valueOR95% CIP-value
  1. CI, confidence interval; NCI-DHQ, National Cancer Institute Dietary History Questionnaire; OR, odds ratio.

  2. a

    All multivariate models including baseline model adjust for age, diabetic status, HCV viral load, chronic alcohol abuse, adiposity (BMI >25) and whether on prescribed vitamin D therapy on year prior to study enrolment.

  3. b

    Based on Institute of Medicine guidelines.

African American (n = 163)
Total serum vitamin D categoryb
Deficient (<12 ng/mL) or insufficient (12–<20 ng/mL)1.36(0.78–2.60)0.361.76(0.86–3.56)0.12
Normal (20–50 ng/mL)1.0 (Ref)1.0 (Ref)
High(>50 ng/mL)9.83(1.10–88.08)0.0412.91(1.30–128.16)0.03
White (n = 126)
Total serum vitamin D categoryb
Deficient (<12 ng/mL) or insufficient (12–<20 ng/mL)1.21(0.45–3.24)0.711.18(0.37–3.69)0.78
Normal (20–50 ng/mL)1.0 (Ref)1.0 (Ref)
High(>50 ng/mL)0.77(0.21–2.79)0.690.84(0.20–3.59)0.82

Race-stratified logistic regression models evaluating the association between serum hydroxyvitamin D levels and advanced hepatic inflammatory activity risk (A2/A3–A3) are reported in Table 4. In our baseline multivariate analysis, AA HCV+ males who were hydroxyvitamin D insufficient or deficient (<20 ng/mL) had greater than two-fold excess risk that approached significance (P = 0.06) compared to AA HCV+ males with normal levels. AA with high serum D levels had only slightly elevated risk, although this association was not significant (OR = 1.43, P = 0.76). The results again differed in Whites; compared with Whites males with normal serum D levels, those with either insufficient or deficient levels or with elevated levels had no evidence of excess risk (OR = 0.70, P = 0.54 and OR = 0.73, P = 0.66 for White males with insufficient/deficient and elevated serum D levels respectively).

Table 4. Association between total serum Vitamin D and other measures of Vitamin D status and risk of FibroSURE-ActiTest-determined advanced hepatic inflammatory activity (A2/A3 and A3) in male veterans with chronic hepatitis C infection
 Univariate modelsMultivariate model-baselinea
OR95% CIP-valueOR95% CIP-value
  1. CI, confidence interval; NCI-DHQ, National Cancer Institute Dietary History Questionnaire; OR, odds ratio.

  2. a

    All multivariate models including baseline model adjust for age, diabetic status, HCV viral load, chronic alcohol abuse, adiposity (BMI>25) and whether on VA prescribed vitamin D therapy on year prior to study enrolment.

  3. b

    Based on Institute of Medicine guidelines.

African American (n = 163)
Total serum vitamin D categoryb
Deficient (<12 ng/mL) or Insufficient (12–<20 ng/mL)2.07(1.01–4.27)0.0482.16(0.98–4.79)0.06
Normal (20–50 ng/mL)1.0 (Ref)1.0 (Ref)
High(>50 ng/mL)0.81(0.09–7.41)0.851.43(0.15–14.11)0.76
White (n = 126)
Total serum vitamin D categoryb
Deficient (<12 ng/mL) or Insufficient (12–<20 ng/mL)0.70(25–2.01)0.510.70(0.23–2.18)0.54
Normal (20–50 ng/mL)1.0 (Ref)1.0 (Ref)
High(>50 ng/mL)1.27(0.36–4.46)0.710.73(0.18–2.96)0.66

Our sensitivity analyses where we removed participants on vitamin D therapy in the prior year (current users) indicated that our observed associations between serum vitamin D levels and risk of advanced hepatic fibrosis (F3/F4–F4) and of advanced inflammatory activity (A2/A3–A3) were robust. In further exploratory analyses, we observed that additional adjustment for NCI-DHQ-determined daily intake of vitamin D (whether dietary only or total including from supplements) only modestly attenuated observed associations between serum D levels and risk of advanced fibrosis or inflammatory activity (data not shown).

Discussion

We evaluated the association between serum vitamin D levels and risk of advanced hepatic fibrosis (F3/F4–F4) and inflammatory activity (A2/A3–A3) in a large and multiethnic population mono-infected with chronic HCV. We observed racial differences in the association between vitamin D levels and degree of HCV-related hepatic fibrosis in our cohort of HCV-infected male veterans.

We found AA males with clinically elevated serum levels of vitamin D (>50 ng/mL) have a higher risk of advanced fibrosis (F3/F4–F4) than those with normal vitamin D levels. These results are consistent with results of a study performed in the HALT-C cohort, which found that AAs whose liver fibrosis clinically progressed over a 4-year period had higher baseline vitamin D levels compared with AAs whose fibrosis had not progressed (32.7 vs. 25.2 ng/mL serum vitamin D in AA fibrosis progressors vs. nonprogressors respectively, P = 0.08).[13] Also consistent with the HALT-C study was our finding of no association between serum vitamin D levels and degree of hepatic fibrosis in White males with chronic HCV. In contrast, our results differ from those from a smaller cross-sectional study performed in Memphis, Tennessee that found no association between serum vitamin D levels and degree of fibrosis in AAs (n = 106), and also reported significantly increased risk of advanced fibrosis with low vitamin D levels in Whites (n = 65).[14] However, given known gender-based differences in vitamin D metabolism, the divergence of our findings and those of the HALT-C study from those of the smaller Memphis, Tennessee HCV study may be attributable to differences in gender ratio among studies (100% vs. 73% vs. 27% male in our cohort vs. the HALT-C and Tennessee HCV study cohorts respectively). Notably, all three studies reported inter-racial differences in association between serum D levels and HCV-related liver disease. Together, these three studies are consistent with a growing number of reports suggestive of variation in vitamin D metabolism and its association with health outcomes in populations of Caucasian and African ancestry.[25-27] However, given differences among these three studies in population and design, additional research in larger prospectively followed cohorts is necessary to confirm the presence of racial differences in the association between serum D and advanced HCV-related liver disease risk.

We also found an association between serum vitamin D levels and excess advanced inflammatory activity risk (A2/A3–A3) only in African American males who were vitamin D deficient or insufficient (<20 ng/mL); they had over 2-fold excess risk that closely approached significance in multivariate analyses (P = 0.06). In contrast, we found neither strong nor significant associations between serum vitamin D levels and advanced inflammatory activity risk in Whites. Few studies have examined the association between serum vitamin D levels and degree of hepatic inflammatory activity in the background of chronic HCV[2, 6, 28]; none was performed in US populations and none specifically examined for interethnic differences. Our findings in Whites are consistent with those from the Swiss Hepatitis C Cohort Study (P = 496), which reported no association between serum vitamin D levels and necroinflammatory activity in the unadjusted analysis (P = 0.9).[28] However, our findings differ from reports from two other groups: a multicentre Italian cohort study that found a strong and significant association between pre-treatment hepatic inflammatory grade and low serum vitamin D levels in 144 HCV+ biopsied patients (ORunadjusted = 3.42, P = 0.004)[6], and a study performed in the Australian and New Zealand CHARIOT cohort (n = 274) that found higher prevalence of vitamin D deficiency in those with higher hepatic activity grade (21% vs. 11% serum levels <50 nmol/L for grade A0/A1 vs. A2/A3 respectively, P = 0.03).[2] Although we cannot readily explain this discordance in findings among Whites, explanatory factors probably include differences in underlying study populations (e.g. the other studies included participants eligible to receive antiviral therapy vs. our study in a general HCV clinic population that included many individuals with contraindications for treatment; all other studies included both genders vs. ours, which was restricted to males only, and only our study adjusted for potential confounders of serum D levels like BMI).

Our study has multiple strengths including its focus on inter-ethnic differences; our prospective consecutive recruitment; our larger sample size particularly for AAs; and adjustment for known confounders of vitamin D status like BMI. Another was our novel evaluation of vitamin D intake from dietary and from nutritional supplement intake as potential confounders. Although our findings suggest that adjustment for oral vitamin D intake is of marginal benefit as associations between serum D levels and advanced liver disease risk were only minimally altered, future research is needed to confirm this finding, given reduced power for our dietary analyses. Finally, we measured and controlled for VA-dispensed vitamin D prescriptions over a 15-year time period and demonstrated that our observed associations were robust to removal of individuals who received vitamin D prescriptions.

Our study also has several limitations that merit discussion. Our primary limitation is our cross-sectional design, which precludes causal inferences. We also had a limited number of AA males with high serum vitamin D levels resulting in reduced power, particularly in multivariate analyses. However, the consistency in our results with those reported for the prospectively followed HALT-C study cohort that found potentially substantial excess fibrosis risk for AAs with high vitamin D levels is supportive of our study findings. We had only a single (not repeated) measure for our primary exposure, DiaSorin ICMA-determined serum hydroxyvitamin D levels, and outcomes, FibroSURE-ActiTest-determined fibrosis and inflammatory activity levels. However, these measurements were obtained for all study participants using validated tests in certified laboratories with personnel blinded to participant characteristics helping to ensure the internal validity and reliability of our findings. We do not have molecular marker measures for vitamin D binding protein or the vitamin D receptor that also have been potentially implicated in risk of HCV-related liver pathology,[4, 29, 30] either in conjunction with or instead of serum vitamin D levels. There were some clinical differences between our AA and White HCV male patients including a higher prevalence of genotype 1 in AA (95% vs. 68% in Whites) that may contribute to observed differences in the relationship with vitamin D and advanced liver disease risk. Additionally, our study was performed in a single clinic population in a subtropical humid climate where strong seasonal variations in serum vitamin D levels were not observed. Finally, as our study was limited to White and AA males, it is unclear if our findings are generalizable to HCV+ males from other race/ethnic groups or to HCV+ females.

There has been considerable interest recently in the role of vitamin D supplementation as a preventive or palliative measure for HCV-related liver disease. Our preliminary findings in a general HCV clinic population not on antiviral therapy suggested an increased advanced fibrosis risk among African American males with chronic HCV who had elevated serum vitamin D levels. Our results in conjunction with those from the HALT-C trial which suggested that there may be excess fibrosis risk in AA males and also among individuals using vitamin D supplements underscore the need for larger prospective studies to confirm these findings in other multiethnic HCV-positive male cohorts, and to assess if similar differences exist in HCV-positive females.

Authorship

Guarantor of the article: D. White.

Author contributions: D. White: Conception, design, analysis, interpretation results, manuscript writing, decision to publish. S. Tavakoli-Tabasi: Conception, data collection, editing manuscript, decision to publish. F. Kanwal: Analysis, editing manuscript, decision to publish. D. Ramsey: Data collection, analysis, decision to publish. A. Hashmi: Analysis, conception, decision to publish. J. Kuzniarek: Data collection, analysis, decision to publish. P. Patel: Data collection, analysis, decision to publish. J. Francis: Data collection, analysis, decision to publish. H. El-Serag: Editing manuscript, design, analysis, interpretation result, decision to publish. All authors approved the final version of the article, including the authorship list.

Acknowledgements

Declaration of personal interests: The authors declare no conflict of interest. The US Department of Veterans Affairs, the National Institutes of Health, and the National Institute of Diabetes and Digestive and Kidney Disease played no role in design, implementation, analysis, interpretation or decision to report these results.

Declaration of funding interests: This material is based upon work supported in part by a VA Clinical Research and Development Merit Review Award (H-22934, PI: H. El-Serag, MD, MPH), and the NIH/National Institute of Diabetes and Digestive and Kidney Diseases (K24 DK081736-01, K01 DK078154-03 and P30 Center Grant DK56338, PIs, D. White, H. El-Serag and M. Estes respectively) and the Houston VA HSR&D Center of Excellence (HFP90-020).

Ancillary