Low vitamin D serum concentration is associated with high levels of hepatitis B virus replication in chronically infected patients


  • Potential conflict of interest: Dr. Kronenberger consults, advises, is on the speakers' bureau for, and received grants from Roche, Bristol-Myers Squibb, Gilead, Merck, and Janssen. Dr. Zeuzem consults for and advises Abbott, Achillion, Bristol-Myers Squibb, Boehringer Ingelheim, Gilead, Idmir, Janssen, Novartis, Merck, Presidio, Roche, Santaris, and Vertex.

  • C.M.L. is supported by the Deutsche Forschungsgemeinschaft (LA 2806/2-1), Johann Wolfgang Goethe University (Förderung Nachwuchsforscher 2012), and the Deutsche Leberstifung (S163/10087/2012).

Address reprint requests to: Christian M. Lange, M.D., Department of Internal Medicine 1, Johann Wolfgang Goethe University Hospital, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany. E-mail: Christian.Lange@kgu.de; fax +49 69 6301 6448


Vitamin D is an important immune modulator that plays an emerging role in inflammatory and metabolic liver diseases, including infection with hepatitis C virus (HCV). In contrast, the relationship between vitamin D metabolism and chronic hepatitis B (CHB) is less well characterized. Therefore, we quantified 25(OH)D3 serum levels in a cohort of 203 treatment-naïve patients with chronic hepatitis B virus (HBV) infection and tested for their association with clinical parameters of CHB. Of 203 patients, 69 (34%), 95 (47%), and 39 (19%) had severe vitamin D deficiency (25(OH)D3 <10 ng/mL), vitamin D insufficiency (25(OH)D3 ≥10 and <20 ng/mL), or adequate vitamin D serum levels (25(OH)D3 ≥20 ng/mL), respectively. In both uni- and multivariate analyses, HBV DNA viral load (log10 IU/mL) was a strong predictor of low 25(OH)D3 serum levels (P = 0.0007 and P = 0.000048, respectively) and vice versa. Mean 25(OH)D3 serum concentrations in patients with HBV DNA <2,000 versus ≥2,000 IU/mL were 17 versus 11 ng/mL, respectively (P < 0.00001). In addition, hepatitis B early antigen (HBeAg)-positive patients had lower 25(OH)D3 serum levels than HBeAg-negative patients (P = 0.0013). Finally, 25(OH)D3 and HBV DNA serum levels showed inverse seasonal fluctuations. Conclusion: Low 25(OH)D3 serum levels are associated with high levels of HBV replication in patients with CHB. This represents a major difference from chronic hepatitis C, where numerous previous studies have shown a lack of correlation between HCV viral load and vitamin D serum levels. Inverse seasonal fluctuations of 25(OH)D3 and HBV DNA serum levels are suggestive of a functional relationship between both variables. (Hepatology 2013;58:1270–1276)


alanine aminotransferase


body mass index


chronic hepatitis B


chronic hepatitis C


hepatitis B early antigen


hepatitis B surface antigen


hepatitis B virus


hepatocellular carcinoma


hepatitis C virus


hepatitis delta virus


human immunodeficiency virus




nucleoside/nucleotide analog inhibitors


nonalcoholic steatohepatitis


pegylated interferon alpha


standard deviation


sustained virologic response


vitamin D receptor.

Despite the availability of potent vaccines, infection with hepatitis B virus (HBV) still represents one of the most significant infectious diseases worldwide.[1] Patients with chronic hepatitis B (CHB) who have HBV DNA levels >2,000 IU/mL, elevated serum alanine aminotransferase (ALT) levels, and at least moderate fibrosis or necroinflammation in liver biopsies require antiviral therapy to avoid progression of liver diseases to cirrhosis and its complications.[2-4] Though these criteria for treatment initiation are generally accepted, current knowledge on the natural history of CHB is incomplete, for example, with respect to the long-term prognosis and risk of hepatocellular carcinoma (HCC) development in immune-tolerant patients with high viral load.[2, 3] Treatment with pegylated interferon alpha (Peg-IFN-α) for 48 (or more) weeks can result in long-lasting immune-mediated control of the disease, but treatment failure occurs in the majority of treated patients.[5, 6] In contrast, potent nucleoside/nucleotide analog inhibitors (NAs) of HBV polymerase-like entecavir or tenofovir display a high antiviral efficacy and genetic barrier to viral resistance development, leading to sustained suppression of HBV replication in the vast majority of treated patients.[1, 2, 7, 8] However, it was estimated that most patients will depend on life-long therapy with NAs, a strategy of thus far unknown safety and efficacy.[3, 7] In addition, long-term treatment with NAs usually does not lead to clearance of HBV-infected hepatocytes, indicating that this treatment modality may not necessarily eliminate risk of HCC development. In view of these facts, the development of novel therapeutic strategies, especially of novel immunomodulatory approaches to control HBV infection, appears to be highly relevant.

Cholecalciferol is the precursor of the bioactive vitamin D metabolite, calcitriol.[9] Nutritional sources of cholecalciferol are rare and its largest proportion in humans is synthesized in the skin during exposure to ultraviolet light.[9] To get bioactivated, cholecalciferol is hydroxylated to 25(OH)D3 at position 25 in the liver and subsequently at position 1 in the kidneys. The resulting bioactive vitamin D metabolite, 1,25(OH)2D3, which is also called calcitriol, exerts its biological functions predominantly by signaling through a nuclear vitamin D receptor (VDR), which serves as a ligand-activated transcription factor.[10] Importantly, clinical assays to quantify calcitriol are generally characterized by poor reliability.[9] Therefore, the stable, easy-to-quantify metabolite, 25(OH)D3, is usually measured in clinical routine to assess a patient's vitamin D status.[9]

By induction or repression of expression of hundreds of genes, calcitriol serves as an important modulator of numerous signaling pathways related to both innate and adaptive immunity.[9, 11-13] In patients with tuberculosis, a therapeutic value of the immunomodulatory effect of vitamin D has already been proven in randomized, controlled clinical trials.[14, 15] Importantly, vitamin D supplementation in these studies resulted in an increased activity of intrinsic interferon-alpha (IFN-α) signaling.[16] Furthermore, vitamin D plays an emerging role in inflammatory and metabolic liver diseases, including infection with hepatitis C virus (HCV).[17-20] For example, it was shown that patients with chronic hepatitis C (CHC) frequently suffer from severe vitamin D deficiency.[17, 19-21] Although the stage of liver fibrosis was a determinant of vitamin D deficiency in CHC patients, even patients without any relevant degree of liver fibrosis had a significantly higher risk of vitamin D deficiency, compared to healthy controls.[19] In contrast to these well-documented findings, it currently remains conflicting whether 25(OH)D3 serum levels can be considered as a predictor of treatment outcome in patients with CHC.[17, 19-21]

Generally, the relationship between vitamin D metabolism and CHB is less well characterized. Therefore, we performed a comprehensive analysis of the relationship between vitamin D serum levels and clinical determinants of CHB.

Patients and Methods


Serum samples of consecutive patients with chronic hepatitis who were followed between January 2009 and December 2012 at the outpatient liver clinic of the Johann Wolfgang Goethe University Hospital (Frankfurt, Germany) were routinely stored and used for the present, retrospective study. Inclusion criteria for the present study were chronic infection with HBV, defined as detectable hepatitis B surface antigen (HBsAg) and HBV DNA ≥6 months, age ≥18 years, treatment-naïve status for both Peg-IFN-α or NAs, and sufficient serum storage. Patients were excluded from the present analyses if they were coinfected with HCV, human immunodeficiency virus (HIV), or hepatitis delta virus (HDV), if excessive alcohol consumption (>40 g/day) had been reported, if they had received a liver allograft, or if a malignant disease, including HCC, was diagnosed. Demographic and clinical characteristics, including age, sex, histological grade, and stage, ALT serum levels, body mass index (BMI), and presence or absence of diabetes, were extracted from clinical databases. Presence versus absence of diabetes was included in our analyses, because, in individuals without hepatitis B, a significant association between diabetes mellitus and low vitamin D serum levels has been described.[13] Liver biopsies were evaluated by experienced local pathologists. Liver fibrosis was classified according to Ishak's score. The study protocol was approved by local ethical committees.

Control Group

An age-matched and sex-matched control group of treatment-naïve HCV-infected patients was derived from a previously published cohort.[19] From this cohort, patients were selected for the present study as follows: For each 3 HBV-infected patients included in the present study, 2 HCV-infected patients were randomly matched according to age and sex. Before matching, both patient cohorts were stratified in groups according to age (18-29, 30-39, 40-49, 50-59, and >60). Within these groups, patients were randomly matched for sex, and supernumerous HCV patients were randomly removed. Randomization was performed based on a numerical order of a random identification number, which had been assigned in the original clinical trial in which all HCV-infected patients had been included.[22]

Quantification of 25(OH)D3 Serum Levels

Serum levels of 25(OH)D3 were measured as described previously.[19] Vitamin D concentrations of <20 and <10 ng/mL were defined as insufficiency and deficiency, respectively, whereas concentrations ≥20 ng/mL were considered as adequate.[9]

Quantification of HBV DNA and HBsAg Serum Levels

Serum HBV DNA levels were quantified by a COBAS Amplicor HBV Monitor (Roche Molecular Systems, Pleasanton, CA) with a detection limit of 12 IU/mL. HBsAg titers were determined with the Architect HBs-Antigen QT assay (Abbott Laboratories, Wiesbaden, Germany) based on an automated chemiluminescent microparticle immunoassay, following the manufacturer's recommendation. The Architect HBs-Antigen QT assay measures a range of HBsAg from 0.05 to 250 IU/mL. Samples with higher HBsAg titer required dilution to bring them in the range of the calibration curve.

Statistical Analyses

Associations between dichotomic (e.g., hepatitis B early antigen [HBeAg]-positive versus negative) or linear (e.g., HBV DNA serum concentration) variables and 25(OH)D3 serum levels were assessed by logistic or linear regression models, respectively. After univariate analyses, multivariate analyses were performed for significant associations. Multivariate models were obtained by backward selection, using a P value >0.15 for removal from the model. Only patients with complete data for the remaining covariates were included in multivariate analyses. Group differences were assessed by means of chi-square contingency tables or Wilcoxon-Mann-Whitney's U tests, as appropriate.


Patient Characteristics

A total of 203 patients were selected for the present study according to the above-described criteria. Baseline characteristics of these patients are summarized in Table 1. Because both IFN-α and some NAs may have an effect on vitamin D metabolism,[19, 23, 24] the main analyses of the present study concentrated on 25(OH)D3 serum levels in treatment-naïve CHB, generally using serum samples taken at patients' first presentation in our outpatient clinic or—in a minority of patients (n = 29)—at baseline of antiviral treatment initiation. The majority of included patients were HBeAg negative (n = 177; 87%).

Table 1. Baseline Characteristics of Included Patients
 HBeAg NegativeHBeAg PositiveCombined
 (n = 177)(n = 26)(n = 203)
Male, n (%)98 (55)9 (35)107 (53)
Female, n (%)79 (45)17 (65)96 (47)
Age, years
Mean (range)40 (19-71)34 (21-56)39 (19-71)
Missing, n000
Race, n (%)   
Caucasian157 (89)20 (77)177 (87)
Asian19 (11)6 (23)25 (12)
African1 (1)0 (0)1 (0)
HBV genotype, n (%)
A27(15)4 (15)31 (15)
B12 (6)4 (15)16 (8)
C14 (8)0 (0)14 (7)
D66 (37)7 (27)73 (36)
E10 (6)2 (8)12 (6)
Missing, n48957
Diabetes, n (%)
Yes8 (5)2 (8)10 (5)
Missing, n17320
HBV DNA, log10 IU/mL
Mean (range)3.19 (1-7)5.38 (3-9)3.47 (1-9)
Missing, n000
Fibrosis stage, n (%)
097 (55)13 (50)110 (54)
140 (23)6 (23)46 (23)
22 (1)3 (12)5 (2)
311 (6)3 (12)14 (7)
41 (1)1 (2)2 (1)
HBsAg, log10 IU/mL
Mean (range)3.2 (1-5)3.7 (2-5)3.3 (1-5)
Missing, n000
Antiviral therapy, n (%)
Yes16 (10)13 (50)29 (14)
Missing, n000
BMI, kg/m2
Mean (range)22 (18-40)27 (19-34)22 (18-40)
Missing, n28432
Mean (range)39 (10-439)50 (18-102)40 (10-439)
Missing, n000

25(OH)D3 Serum Levels

Mean 25(OH)D3 serum concentration of the entire cohort was 14.40 ng/mL (standard deviation [SD]: 7.90). Of 203 patients of the entire cohort, 69 (34%), 95 (47%), and 39 (19%) had severe vitamin D deficiency [25(OH)D3 <10ng/mL], vitamin D insufficiency (25(OH)D3 ≥10 and <20 ng/mL) or normal vitamin D serum levels (25(OH)D3 ≥20 ng/mL), respectively. 25(OH)D3 serum levels in Caucasian and non-Caucasian HBV-infected patients were similar (14.38 and 14.59 ng/mL, respectively; P = 0.7).

Hence, vitamin D deficiency and insufficiency was highly prevalent in CHB patients. Overall, 25(OH)D3 serum concentration in HBV-infected patients appeared to be somewhat lower, compared to an age-matched and sex-matched cohort of HCV-infected patients (mean 25(OH)D3 concentration =18.15 ng/mL [SD = 11.27]; details in Supporting Table 1), and significantly lower compared to healthy individuals of a previously published cohort of the same geographic region.[25]

Relationship Between 25(OH)D3 Serum Levels and Virologic Parameters of CHB

25(OH)D3 and HBV DNA serum levels showed a significant, inverse correlation (P = 0.0003, Fig. 1). Therefore, we therefore performed uni- and multivariate logistic regression analyses of determinants of 25(OH)D3 serum levels in patients with CHB. In both uni- and multivariate analyses, HBV DNA was the strongest determinant of low 25(OH)D3 serum concentration in our cohort (P = 0.0007 and P = 0.000048, respectively; Table 2). Conversely, 25(OH)D3 serum concentration was independently associated with HBV DNA serum levels in a multivariate regression model, together with degree of liver fibrosis, ALT serum levels, BMI, and presence of HBeAg (Table 3). Interestingly, HBsAg serum levels were not associated with 25(OH)D3 serum levels. To further characterize the clinical relevance of the relationship between 25(OH)D3 and HBV DNA serum levels, we stratified patients according to HBV DNA serum concentration <2,000 versus ≥2,000 IU/mL. Patients with HBV DNA viral load below this threshold, which is generally considered as relevant for clinical decision making, had substantially higher mean 25(OH)D3 serum levels, compared to patients with HBV DNA ≥2,000 IU/mL (17 versus 11 ng/mL, respectively; P < 0.00001). We then stratified patients according to HBV DNA <3 (n = 84), 3 to <5 (n = 101), 5 to <8 (n = 15), and ≥8 log10 IU/mL (n = 3); 25(OH)D3 serum levels in these patients were 16.68, 13.33, 10.43, and 9 ng/mL, respectively.

Figure 1.

Correlation between 25(OH)D3 and HBV DNA serum concentrations.

Table 2. Factors Associated With 25-Hydroxyvitamin D Serum Levels in Patients With CHB
VariableP Value, UnivariateP Value, MultivariateStandard Beta, Multivariate
  1. Abbreviation: GGT, gamma-glutamyltransferase.

Age, (years, continuous)0.3600  
Male sex0.1300  
Race (Cacausian versus non-Caucasian)0.6000  
Fibrosis (F1-F2 versus F3-F4)0.1100  
ALT (U/L, continuous)0.1900  
GGT (U/L, continuous)0.6300  
BMI (kg/m2, continuous)0.5100  
Platelets (count/nL, continuous)0.2900  
Diabetes (presence versus absence)0.7000  
HBeAg (positive versus negative)0.0040  
HBV DNA (log10 IU/mL, continuous)0.00070.000048−0.31
HBsAg (log10 IU/mL, continuous)0.5800  
Table 3. Factors Associated With HBV DNA Serum Concentration (log10 IU/mL)
VariableP Value, UnivariateP Value, MultivariateStandard Beta, Multivariate
  1. Abbreviation: GGT, gamma-glutamyltransferase.

Age (years, continuous)0.320000  
Male sex0.430000  
Race (Caucasian versus non-Caucasian)0.100000  
Fibrosis (F1-F2 versus F3-F4)0.0500000.070000.12
ALT (U/L, continuous)0.0006000.000270.24
GGT (U/L, continuous)0.450000  
BMI (kg/m2, continuous)0.0500000.01500−0.15
Platelets (count/nL, continuous)0.990000  
Diabetes (presence versus absence)0.640000  
25(OH)D3 (ng/mL, continuous)0.0016000.001300−0.21
HBeAg (present versus absent)<0.000001<0.0000010.45
HBsAg (log10 IU/mL, continuous)0.410000  

In addition to HBV DNA serum concentration, presence of HBeAg was significantly associated with low 25(OH)D3 serum levels in univariate, but not in multivariate, analysis (Table 2). Furthermore, HBeAg-positive patients had a higher risk of severe vitamin D deficiency than HBeAg-negative patients (P = 0.013; Fig. 2). However, because 25(OH)D3 serum concentration and presence of HBeAg showed a significant association only in univariate, but not in multivariate, analysis (Table 2), it appears likely that this association is driven by the substantial association between HBV viral load and low 25(OH)D3 serum concentration.

Figure 2.

HBeAg-positive hepatitis B is associated with poor 25(OH)D3 status. Mean 25(OH)D3 serum levels according to presence versus absence of HBeAg are shown on the left side. On the right, the proportions of patients with HBeAg-negative or -positive CHB suffering from significant vitamin D deficiency (25(OH)D3 serum concentration <10ng/mL) are shown.

Because 25(OH)D3 serum levels strongly fluctuate according to seasonal changes of sunlight exposure, we assessed whether HBV DNA serum levels show any fluctuation according to season as well, which might be considered as a hint for a functional effect of vitamin D metabolism on HBV replication. HBV DNA serum levels were indeed significantly lower when serum samples were taken in spring or summer, compared to autumn or winter (P = 0.01), which represented a reciprocal seasonal variation of both variables (Fig. 3).

Figure 3.

25(OH)D3 and HBV DNA serum levels are characterized by inversed seasonal variations. 25(OH)D3 and HBV DNA serum levels, measured in the same serum sample, are shown according to season when serum samples were taken for quantification.

Figure 4.

25(OH)D3 serum levels before and during antiviral therapy of CHB. Data were available for 21 patients. Mean 25(OH)D3 serum levels at baseline and follow-up were 10.3 (SD = 5.5) and 13.0 (SD = 6.6), respectively (P = 0.18). Mean follow-up time was 20.2 months (range, 5-36). Seasons of blood sampling were balanced between baseline and follow-up.

Longitudinal 25(OH)D3 Serum Levels Before and During Antiviral Therapy of CHB

In 21 patients in whom antiviral treatment was initiated after quantification of vitamin D serum levels, 25(OH)D3 serum levels could be determined at follow-up. Two of these patients received monotherapy with Peg-IFN-α, and 19 were treated with NAs (entecavir, N = 4; tenofovir, n = 13; lamivudine, n = 2). This subgroup of treated patients had a mean age of 39 years (range, 18-61), mean HBV DNA serum concentration of 5.48 log10 (SD = 1.67) at baseline of antiviral therapy, and 6 of 21 patients (29%) were HBeAg positive. Mean 25(OH)D3 serum levels at baseline and follow-up were 10.3 (SD = 5.5) and 13.0 ng/mL (SD = 6.6), respectively (P = 0.18; Fig. 4). Mean follow-up time was 20.2 months (range, 5-36). All patients had HBV DNA serum levels <2,000 IU at time of follow-up.


The present study demonstrates a profound association between higher levels of HBV replication and low 25(OH)D3 serum levels in CHB patients. At least in patients without advanced liver disease, the predominant phenotype of our cohort, HBV DNA viral load appears to be the strongest determinant of low 25(OH)D3 serum levels. In this regard, the clinically relevant threshold of HBV DNA serum concentration (≥2,000 IU/mL) identifies patients with particularly low 25(OH)D3 serum levels, compared to patients with lower levels of HBV replication. Importantly, inversed seasonal fluctuations of 25(OH)D3 and HBV DNA serum levels are suggestive for a (yet to be proven) functional relationship between both variables.

Vitamin D deficiency is common in healthy world-wide populations.[9] Nevertheless, patients with liver diseases, such as CHC or nonalcoholic steatohepatitis (NASH), are at substantially higher risk for severe vitamin D deficiency than healthy individuals.[17-20, 26, 27] In the present study, we demonstrate that patients with CHB are at high risk of vitamin D deficiency as well. Importantly, neither in the present study nor in previous studies in hepatitis C or NASH patients, was advanced liver fibrosis a prerequisite for low 25(OH)D3 serum levels, suggesting that other mechanisms, such as immune-mediated suppression of liver 25-hydroxylases (e.g., CYP2R1), contribute to vitamin D deficiency.[17-20, 26, 27] In this regard, it is important to note that only a few patients in our study had higher degrees of liver fibrosis. Therefore, we can clearly state that HBV-infected patients are at high risk of vitamin D deficiency, even in the absence of significant liver fibrosis, but we could not assess the effect of advanced liver fibrosis or cirrhosis on vitamin D serum levels in HBV-infected patients.

In view of the above-described data, vitamin D deficiency in liver diseases does not appear to be specific for distinct entities. Nevertheless, our study highlights an important specific feature of vitamin D metabolism in CHB, because the strong association between 25(OH)D3 and HBV DNA serum concentration is in sharp contrast to results of numerous previous studies in patients with CHC, which failed to show an association between vitamin D serum levels and HCV viral load.[17-21, 27-30] In this regard, the observed inversed seasonal fluctuations of 25(OH)D3 and HBV DNA serum levels might be considered as an important hint for a functional relationship between both variables.

The possible causal relationship between vitamin D metabolism and HBV replication, which needs to be proven by future studies, may offer attractive therapeutic opportunities for treatment of CHB. As highlighted above, the addition of relatively high doses of cholecalciferol to standard therapy was superior to the addition of placebo in the treatment of patients with lung tuberculosis.[14] In this study, as well as, for example, in studies in diabetes patients, vitamin D supplementation resulted in a profound change in serum and cellular cytokine profiles.[16, 31] For treatment of chronic hepatitis, the development of novel immunomodulatory agents appears highly relevant. Such agents might be suitable add-ons to both treatment of with Peg-IFN-α or NAs to increase the chance of long-lasting immune control of HBV infection. In this regard, it needs to be emphasized that we did not observe an association between vitamin D and HBsAg serum levels, suggesting that therapeutic administration of vitamin D alone may probably not promote the important clinical goal of HBsAg loss.

Our study has several limitations. Most important, clinical association studies cannot prove causal relationships. Thefore, a suggestive functional link between vitamin D metabolism and HBV replication remains elusive. Nevertheless, our study is of hypothesis-generating value for the design of future in vitro studies or prospective clinical interventional trials. Furthermore, the cohort investigated in our study is not representative for all phases of CHB. Especially, patients with advanced liver fibrosis and high ALT serum levels are underrepresented in our cohort. Furthermore, only few patients in our cohort were HBeAg positive. Although the majority of HBV patients treated in our center is HBeAg negative, the inclusion criterion “treatment-naïve state at time of vitamin D quantification” of our study appears to account for a selection of overall younger patients with limited degrees of liver fibrosis, as well as of HBeAg-negative patients. Yet, it appears unlikely that the profile of the present cohort affects the main finding of our study, that low 25(OH)D3 serum levels are associated with HBV DNA viral load. In addition, the number of patients in our study who received antiviral therapy after initial vitamin D quantification, and in whom vitamin D serum levels could be quantified during follow-up, appears to be too small to draw final conclusions. Although 25(OH)D3 serum levels were slightly higher during follow-up, compared to baseline, which may indicate that HBV replication affects vitamin D metabolism, this difference did not reach statistical significance. Larger studies are necessary to fully address this important question, as well as to assess the effect of other possible determinants of vitamin D serum levels, such as BMI.

In conclusion, we demonstrate a significant association between low 25(OH)D3 serum levels and high levels of HBV replication in chronically infected patients. Future studies to evaluate a therapeutic value of vitamin D and its analogs in HBV infection may be justified.


The authors thank Doris Kärger for her expert technical assistance.