Clinical Nutrition xxx (2015) 1e8

Contents lists available at ScienceDirect

Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu

Original article

Vitamin D supplementation reduces depressive symptoms in patients

with chronic liver disease
Caroline S. Stokes a, *, Frank Grünhage a, Crystal Baus b, Dietrich A. Volmer c,
Stefan Wagenpfeil d, Matthias Riemenschneider b, Frank Lammert a
a
Department of Medicine II, Saarland University Medical Center, Homburg, Germany
b
Department of Psychiatry and Psychotherapy, Saarland University Medical Center, Homburg, Germany
c
Institute of Bioanalytical Chemistry, Saarland University, Saarbrücken, Germany
d
Institute of Medical Biometry, Epidemiology and Medical Informatics, Saarland University, Campus Homburg, Germany

a r t i c l e i n f o s u m m a r y

Article history: Background and aims: Vitamin D deficiency and depression frequently occur in patients with chronic
Received 9 March 2015 liver diseases (CLD). Depression has recently been inversely associated with vitamin D in a meta-analysis,
Accepted 8 July 2015 and vitamin D receptor is expressed in brain. This pilot study investigates whether vitamin D replace-
ment ameliorates depressive symptoms in CLD patients and consists of a cross-sectional and an inter-
Keywords: ventional analysis.
Beck Depression Inventory
Methods: Overall, 111 patients with CLD were included in the cross-sectional analysis. The Beck
Cirrhosis
Depression Inventory II (BDI-II) was used to assess depression. Chemiluminescence immunoassay and
Mood
Supplementation
LC-MS/MS quantified serum 25-hydroxyvitamin D levels. For the interventional analysis, 77 patients
Women with inadequate vitamin D concentrations received 20,000 IU vitamin D per week for six months. The
25-Hydroxyvitamin D final follow-up was carried out six months post supplementation.
Results: In the cross-sectional analysis, 81% of patients (median age 55 years, 47% women) had inade-
quate baseline vitamin D levels (<30 ng/ml), and 31% presented with depressive symptoms (BDI-II score

14). Depression severity correlated inversely with vitamin D level in depressed patients (b ¼ 0.483,
P ¼ 0.004). Depression scores improved significantly from baseline in depressed patients after three and
six months (P ¼ 0.003 and P ¼ 0.004, respectively) of supplementation, with vitamin D levels increasing
to normal (P < 0.0001). Subgroup analyses revealed this anti-depressant effect of vitamin D to occur
predominantly in women. The final follow-up showed increases in median BDI-II scores in the setting of
decreased vitamin D levels.
Conclusions: Vitamin D levels correlated with BDI-II scores, and vitamin D replacement significantly
improved depressive symptoms in women with CLD. Adjuvant vitamin D may be considered in these
patients.
Registration No: DRKS00007782 German Clinical Trials Registry (DRKS)
© 2015 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

1. Introduction of vitamin D. The first vitamin D hydroxylation step occurs in the

liver, yielding 25-hydroxyvitamin D (the primary circulating form
Chronic liver diseases (CLD) are often associated with vitamin D and accepted marker of vitamin D status). The second oxidation
deficiency, which contributes to the manifestation of metabolic step in kidney produces the most biologically active vitamin D
bone disease and other systemic complications [1,2]. This might be, metabolite, 1,25-dihydroxyvitamin D, which subsequently enters
in part, attributed to the involvement of the liver in the metabolism all target cells, exerting its functions mainly through the vitamin D
nuclear receptor (VDR) [3].
Given the broad expression of VDR in many organs, numerous
* Corresponding author. Department of Medicine II, Saarland University Medical non-skeletal benefits for vitamin D have been observed. In brain,
Center, Kirrberger Str. 100, 66421 Homburg, Germany. Tel.: þ49 6841 16 23222;
vitamin D binds to VDR in many areas implicated in depression
fax: þ49 6841 16 23267.
E-mail address: [email protected] (C.S. Stokes). pathophysiology, such as neurons and glial cells of the limbic

http://dx.doi.org/10.1016/j.clnu.2015.07.004
0261-5614/© 2015 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

Please cite this article in press as: Stokes CS, et al., Vitamin D supplementation reduces depressive symptoms in patients with chronic liver
disease, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.07.004
2 C.S. Stokes et al. / Clinical Nutrition xxx (2015) 1e8

system [4]. Indeed, VDR gene variants are reported to influence dose of 20,000 IU cholecalciferol/vitamin D3 (Dekristol®, Jena-
susceptibility to depressive symptoms [5]. Neurosteroidal effects of pharm, Jena, Germany) contained in one capsule, daily for the first
vitamin D have recently been illustrated in an experimental study, seven days, then weekly thereon. Patients were further stratified
in which adult mice displayed behavioural and brain alterations into subgroups based on whether depressive symptoms were
upon the induction of vitamin D deficiency [6]. In humans, cross- present. All patients were followed up after three (T1) and six
sectional studies have shown serum 25-hydroxyvitamin D months (T2), after which vitamin D supplementation was stopped.
concentrations to be inversely correlated with the severity of A final follow-up took place 12 months from baseline (T3) after a
depressive symptoms [7]. A recent meta-analysis by Anglin and co- six-month pause from vitamin D. Compliance was determined from
workers underscored this relationship [8]. increases in serum 25-hydroxyvitamin D concentrations, together
Co-morbid depression often affects patients with CLD, with a with the discrete interval pill-count method. Patients abstained
reported prevalence of approximately 30% [9,10]. Standard anti- from taking other nutritional supplements during the study, unless
depressant therapy, however, is not always feasible. Certain anti- medically indicated. Patients on anti-depressant therapy were only
depressant medications are contraindicated in advanced CLD due recruited if their dose had remained stable for the preceding eight
to their side effects. Selective serotonin reuptake inhibitors, for weeks, and were instructed to inform the study coordinator
instance, can induce gastrointestinal bleeding, possibly because of immediately upon any dosage or medication changes.
increased gastric acidity or via defective platelet aggregation due to
reduced platelet serotonin [11]. Moreover, patients might be 2.3. Study procedures
dissuaded by treatments for depression because of unwanted side-
effects, namely fatigue and the aggravation of features of the Serum 25-hydroxyvitamin D analyses were conducted by
metabolic syndrome, such as weight gain and dyslipidemia [12]. chemiluminescence immunoassay LIAISON® 25-OH Vitamin D
Consequently, non-pharmacological options warrant further TOTAL Assay (DiaSorin, MN, USA). A subset of samples was vali-
investigation, particularly if they are free from major side effects dated by liquid chromatography-tandem mass spectrometry (LC-
and also deliver additional health benefits. MS/MS), 96-well micro-extraction (AC Extraction Plate, Tecan,
Vitamin D supplementation may fulfil these criteria and has Ma €nnedorf, Switzerland), 25-hydroxyvitamin D isotope calibration
previously been reported to ameliorate depressive symptoms in and quality control kits (Recipe, Munich, Germany) and 5500
overweight subjects [13]. Whether vitamin D has direct benefits on QTRAP electrospray ionisation LC-MS/MS (Sciex, Concord, ON,
depressive symptoms in patients with CLD has to date not been Canada), as previously reported [17].
investigated. The primary aim here was to assess for anti- Other standard biochemical analyses in our central clinical
depressant effects of vitamin D in CLD patients. We hypothesised laboratory included: calcium, phosphate, creatinine, urea, PTH as
that depressive symptoms can be mitigated by correcting inade- well as standard liver function tests (LFTs): alanine aminotrans-
quate vitamin D concentrations in these patients. ferase (ALT), aspartate aminotransferase (AST), alkaline phospha-
tase (AP), and gamma glutamyl transpeptidase (g-GT) activities,
2. Patients and methods bilirubin, albumin, international normalised ratio (INR), and uri-
nary calcium and phosphate concentrations. We monitored plasma
2.1. Study patients and 24-h urine calcium levels and renal function at three-month
intervals or more frequently, as indicated.
We prospectively recruited adult patients attending the Hep- Clinical examinations included lifestyle and health indicators.
atology Outpatient Clinic (Department of Medicine II, Saarland Bioelectrical impedance analysis assessed body composition
University Medical Center). Patients with CLD (diagnosed
six (Tanita BC-418 MA, Sindelfingen, Germany). Physical activity was
months previously) were invited to participate, unless they had any documented through self-report questions enabling classification
of the following exclusion criteria: severe hepatic encephalopathy into one of four physical activity index groups (inactive; moderately
as quantified by Critical Flicker Frequency (CFF; HEPAtonorm inactive; moderately active; active) as per the European Prospective
analyzer; Accelab, Kusterdingen, Germany) and defined by CFF Investigation into Cancer and Nutrition (EPIC) study questionnaires
<35 Hz [14]; current interferon therapy; and specifically for the [18]. Patients completed the Alcohol Use Disorders Identification
intervention phase: presence of hypercalciuria (urinary calcium Test (AUDIT) to evaluate alcohol intake [19].
>8.0 mmol/d); hypercalcaemia (any value above the upper normal Diagnosis of depression and assessment of depressive symp-
limit, serum calcium >2.7 mmol/l), alone or in combination with toms was conducted using the International Classification of Dis-
hyperparathyroidism (parathyroid hormone [PTH] >65.0 pg/ml) eases (ICD-10) system [20] and the validated German version of the
[15]; history of calcium-containing kidney stones; allergy or hy- revised Beck Depression Inventory II (BDI-II) [21]. This self-report
persensitivity to any of the supplement ingredients (gelatine, instrument comprises 21 questions on a four-point likert scale.
peanuts, soy); sarcoidosis; stage IV or V chronic kidney disease (as Final scores range from 0 (no depression) to 63 (severe depression);
defined by the National Kidney Foundation [16]), or pregnancy. scores
14 represent the cut-off for depression. In addition, pa-
Patients with adequate serum 25-hydroxyvitamin D levels (
30 ng/ tients were asked whether any traumatic event may have occurred
ml) did not receive vitamin D but were monitored during the to significantly influence their score.
intervention. The local research ethics committee approved the
€
study protocol (Arztekammer des Saarlandes, Ref. 57/11), which 2.4. Outcome parameters
conforms with the declaration of Helsinki. All patients provided
written informed consent. The primary outcome assessed absolute changes in depressive
symptoms after three and six months of vitamin D supplementa-
2.2. Study design tion. In contrast to depression, no formal recommendations for
interpreting the BDI-II in CLD patients exist, however a consensus
Eligible patients were included in a cross-sectional analysis statement from IMMPACT (Initiative on Methods, Measurement
assessing the association of vitamin D with depression. Patients and Pain Assessment in Clinical Trials [22]) suggests a reduction of
with serum 25-hydroxyvitamin D concentrations <30 ng/ml then five points as being clinically meaningful based on their systematic
received six-month replacement therapy. This consisted of an oral analysis of BDI in chronic pain studies. This recommendation was

Please cite this article in press as: Stokes CS, et al., Vitamin D supplementation reduces depressive symptoms in patients with chronic liver
disease, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.07.004
 C.S. Stokes et al. / Clinical Nutrition xxx (2015) 1e8 3

also endorsed by other researchers [23]. We therefore interpreted a Table 1

change (reduction) of five points on the BDI-II scale as being clin- Baseline characteristics of the cross-sectional study population.

ically meaningful, enabling the classification of a patient as Entire Depression No

‘responder’. Secondary outcomes included liver biochemistry cohort depression
changes after vitamin D supplementation. Sociodemographic
N (male/female) 111 (59/52) 34 (14/20) 77 (45/32)
2.5. Statistical analyses Age, years, median (range) 55 (20e81) 56 (30e77) 54 (20e81)
Lifestyle and health indicators (%)
PAI (inactive/moderately 13/27/26/34 18/18/37/27 10/31/22/37
Statistical analyses were performed with SPSS 20.0 (IBM, inactive/moderately
Ehningen, Germany). Non-parametric statistics were used, since active/active)
the primary outcome variable was not normally distributed. We BMI, kg/m2, median (range) 26.1 27.1 25.1
evaluated differences between categorical variables with c2 tests (17.0e54.5) (17.0e36.0) (18.5e54.5)*
Indicators of CLD
and the ManneWhitney U test was employed for continuous un- Primary liver disease, n (%)
paired variables. Differences between blood parameters and clin- Chronic hepatitis C 39 (35) 14 (41) 25 (33)
ical characteristics were examined using Kruskall-Wallis or virus infection
ManneWhitney tests, where appropriate. We examined the asso- Chronic hepatits B 21 (19) 4 (12) 17 (22)
virus infection
ciation of season sampling on both vitamin D levels and depressive
Autoimmune hepatitis 10 (9) 1 (3) 9 (12)
symptoms by comparing patients with blood taken during summer Alcoholic liver disease 8 (7) 4 (12) 4 (5)
and autumn months with winter and spring months. Serum 25- NAFLD 12 (11) 4 (12) 8 (10)
hydroxyvitamin D concentrations were further categorised based Other 21 (19) 7 (20) 14 (18)
on thresholds reflecting clinical cut-offs for vitamin D inadequacy Liver cirrhosis, n (%) 35 (32) 17 (50)** 18 (24)**
Depression based on BDI-II category, n (%)
(<30 ng/ml), deficiency (<20 ng/ml) and severe deficiency (<10 ng/ None (0e13 points) 77 (69) 0 77 (100)
ml) [24]. We investigated whether 25-hydroxyvitamin D levels Mild (14e19 points) 18 (16) 18 (53) 0
were associated with depression severity using uni- and multi- Moderate (20e28 points) 11 (10) 11 (32) 0
variate analyses after adjusting for possible confounders including Severe (29e63 points) 5 (5) 5 (15) 0
BDI-II score, median (range) 10 (0e41) 19 (14e41) 6 (0e13)
age, sex, body mass index (BMI), season of blood sampling, physical
Biochemical serum markers, median (range)
activity or presence of cirrhosis. The presence of depressive 25-hydroxyvitamin D (ng/ml) 16.9 17.4 16.3
symptoms was dichotomised and defined as clinically relevant for (4.0e49.4) (4.3e39.3) (4.0e49.4)
BDI-II
14 and was included as the dependent variable in all ana- Calcium (mmol/l) 2.4 (2.0e2.7) 2.4 (2.0e2.7) 2.4 (2.0e2.7)
lyses. Significant variables (P value set to 0.1) in univariate analysis Phosphate (mg/dl) 3.2 (1.3e4.5) 3.4 (1.7e4.5)* 3.1 (1.3e4.1)*
PTH (pg/ml) 40 (14e90) 39 (18e90) 40 (14e89)
were included in the multivariate model with backward logistic AST (U/l) 37.5 43.0 35.0
regression (cross-checked with forward logistic regression, and (15.0e263.0) (20.0e183.0) (15.0e263.0)
only reported the results if they differed). Non-significant cova- ALT (U/l) 39.5 49.0 38.0
riates with an established influence on depression (BMI, season) (8.0e340.0) (11.0e276.0) (8.0e340.0)
g-GT (U/l) 63.0 80.0 52.5
were also included in this multivariate model.
(12.0e914.0) (19.0e914.0) (12.0e502.0)
Related samples (e.g. pre and post-intervention BDI-II scores) AP (U/l) 84 (36e303) 91 (44e256) 82 (36e303)
were analysed using Wilcoxon signed rank test. Given that both Creatinine (mg/dl) 0.8 (0.5e1.9) 0.8 (0.5e1.9)* 0.9 (0.5e1.8)*
increasing age and women have a higher predisposition to Urea (mg/dl) 31.0 30.0 31.0
depression, we carried out subgroup analyses excluding patients (4.6e77.0) (16.0e77.0) (4.6e58.0)

over 65 years of age, and comparing men with women. For the Significant differences between depressed (BDI-II
14) and non-depressed patients
intervention study data, we used per-protocol analysis in addition (BDI-II <14) are denoted with * (P < 0.05) and ** (P < 0.001).
Abbreviations: ALT, alanine aminotransferase; AP, alkaline phosphatase; AST,
to intention-to-treat analysis (ITT), where the last observation
aspartate aminotransferase; BDI-II, Beck Depression Inventory-II; BMI, body mass
carried forward method was implemented to compare absolute index; CLD, chronic liver disease; g-GT, gamma glutamyl transpeptidase; NAFLD,
changes in depression scores (a zero value was imputed for change non-alcoholic fatty liver disease; PAI, physical activity index; PTH, parathyroid
score in drop-outs). Patients who returned for the first or second hormone.
follow-up were retained in the per-protocol analysis. A two-sided P
value < 0.05 was set as the threshold for statistical significance, and
significant results are indicated with an asterisk. Results are pre- concentrations (<30 ng/ml) were identified in 81% of patients
sented as medians and ranges as well as absolute and relative fre- (n ¼ 90). Of these, 52% (n ¼ 47) exhibited moderate (20e10 ng/ml)
quencies, unless stated otherwise. and 24% (n ¼ 22) presented with severe vitamin D deficiency
(<10 ng/ml); of these, 38% were sampled in summer/autumn and
3. Results 62% in winter/spring.
A third of patients (n ¼ 34) had mild to moderate depressive
3.1. Cross-sectional results symptoms (median score 19, range 14e41); of these, 16 (14%) were
on a stable dose of anti-depressants and had a significantly
3.1.1. Patient characteristics (P ¼ 0.005) higher median BDI-II score than those not on anti-
Table 1 summarises the main characteristics of the outpatient depressants, however vitamin D levels did not differ between
cohort (n ¼ 111), for which the median age was 55 years and 52 these groups. Figure 1 shows the number of patients without and
were women. CLD resulted from chronic viral hepatitis in over half with depression in the different vitamin D categories. Overall, 91%
of the cohort with 40% of patients (n ¼ 41) having a normal BMI versus 77% of depressed and non-depressed patients had inade-
(<25 kg/m2). The remainder were either overweight (40%, BMI quate vitamin D levels, however the distribution across vitamin D
25.0e29.9 kg/m2) or obese (20%, BMI
30 kg/m2). All but two pa- categories did not differ significantly between the two groups.
tients had a normal AUDIT score assessing alcohol intake. A third of Significantly more depressed than non-depressed patients a base-
patients had developed liver cirrhosis, classified as Child-Pugh line had liver cirrhosis and displayed higher serum phosphate and
stages A (n ¼ 30) or B (n ¼ 5). Inadequate vitamin D serum lower creatinine concentrations as well as a higher BMI (all

Please cite this article in press as: Stokes CS, et al., Vitamin D supplementation reduces depressive symptoms in patients with chronic liver
disease, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.07.004
4 C.S. Stokes et al. / Clinical Nutrition xxx (2015) 1e8

Fig. 2. Correlation of serum 25-hydroxyvitamin D concentrations with depressive

symptoms. Baseline 25-hydroxyvitamin D levels were significantly associated with
depressive symptoms in depressed patients (b ¼ 0.483, 95% CI 0.710 to 0.149,
Fig. 1. Clinical classification of serum 25-hydroxyvitamin D levels based on patients P ¼ 0.004, n ¼ 34), in contrast to the non-depressed patients, in whom no association
without and with depression. was observed between these two variables (b ¼ 0.008, CI 0.085 to 0.079, P ¼ 0.947,
n ¼ 77).

P < 0.05). No other differences were detected in an exploratory

analysis of the groups. Table 2
Predictors of depression with univariate and multivariate binary logistic regression
analysis.
3.1.2. Vitamin D levels were inversely associated with depressive
Factor OR 95% CI P
symptoms
Baseline 25-hydroxyvitamin D levels were significantly associ- (A) Univariate analysis with presence of depression as dependent variable
Age 1.01 0.98e1.05 0.549
ated with depression severity in depressed patients (b ¼ 0.483,
BMI 1.05 0.97e1.13 0.248
95% CI 0.710 to 0.149, P ¼ 0.004; Fig. 2), where lower vitamin D Cirrhosis 3.28 1.39e7.71 0.006
levels correlated with severity of depressive symptoms (i.e. higher Physical activity 1.00 0.98e1.01 0.759
BDI-II score). In contrast, no association was observed between Season 1.74 0.77e3.97 0.185
Sex 2.01 0.89e4.56 0.095
these two variables in non-depressed patients (Fig. 2). Uni- and
Vitamin D level 0.98 0.94e1.02 0.317
multivariate logistic regression analysis (Table 2A and B) revealed (B) Multivariate analysis with presence of depression as dependent variable
the presence of cirrhosis and female sex as independent predictors Cirrhosis 4.01 1.61e9.97 0.003
of depression, and a post-hoc analysis showed a significant inter- Sex 2.62 1.07e6.39 0.035
action of sex and BMI as independent predictors of depression (C) Multivariate analysis with presence of depression as dependent variable
and with sex assessed as interaction with BMI
(Table 2C).
Cirrhosis 3.41 1.32e8.80 0.011
Sex*BMI 1.05 1.01e1.08 0.008
3.2. Interventional results Abbreviations: BMI, body mass index; CI, confidence interval; OR, odds ratio.
Bold values represents the significant values.
3.2.1. Allocation to intervention study
All patients included in the cross-sectional analysis were brief, 70% of patients with depression were women, compared to
screened for the intervention study (Fig. 3). We excluded 34 pa- 40% without depression (P ¼ 0.02). The only baseline differences
tients for not meeting the predetermined supplementation criteria between the two groups were that significantly more depressed
or because they declined our invitation. Of these, 20 patients had patients had liver cirrhosis (P ¼ 0.02), higher serum ALT (P ¼ 0.02),
adequate baseline vitamin D levels (only three with depression), and lower serum creatinine levels (P ¼ 0.003).
but were still followed up at all time points. Consequently, 77 pa-
tients (24 with and 53 without depression) received vitamin D
supplementation (Fig. 3). Treatment data were obtained for 92% 3.2.2. Depressive symptoms improved during the treatment period
and 85% of patients, respectively, who were included in the per- The significant improvement in depressive symptoms after
protocol analysis. In each group, 79% were available for the final vitamin D supplementation is illustrated in Fig. 4 (panel A), with
observation analysis at T3. reductions in BDI-II scores in depressed patients after three (13 vs.
The baseline characteristics of the intervention groups, which 18; P ¼ 0.003) and six months (14 vs. 18; P ¼ 0.004). Of note, this
are summarised in Table 3, resemble the overall characteristics of effect was not due to anti-depressant medication as the majority of
the entire cohort (Table 1). Table 3 includes patients who suc- benefits were seen in patients not taking anti-depressants (77% vs.
cessfully completed treatment until at least three months (T1). In 23%). No change was observed in patients without depression. The

Please cite this article in press as: Stokes CS, et al., Vitamin D supplementation reduces depressive symptoms in patients with chronic liver
disease, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.07.004
 C.S. Stokes et al. / Clinical Nutrition xxx (2015) 1e8 5

Fig. 3. Flow chart of patient participation in the study.

Table 3
Characteristics of patients completing the intervention study. treatment response coincided with the restoration of serum vitamin
Depression No depression D levels, which increased to normal after three and six months in
both depressed (38.1 and 38.2 ng/ml, respectively) and non-
Sociodemographic
N (male/female) 23 (7/16)* 50 (30/20)* depressed (36.3 and 35.1 ng/ml, respectively) patients (all
Age, years, median (range) 55 (30e77) 56 (20e81) P < 0.0001). Improvements in depressive symptoms remained sig-
Lifestyle and health indicators (%) nificant when carrying out an ITT analysis (data not shown). Figure 4
PAI (inactive/moderately 17/17/35/31 10/32/20/38 further illustrates the delta (absolute change) of BDI-II scores in
inactive/moderately active/active)
BMI, kg/m2, median (range) 27.0 (17e36) 25.7 (19.4e54.5)
depressed patients, with marked improvements in the majority af-
Indicators of CLD ter both three (panel B) and six months (panel C). Specifically, half of
Primary liver disease, n (%) the patients with depression displayed improvements of five points
Chronic hepatits C virus infection 10 (43) 17 (34) or more (responders), synonymous with a clinically meaningful
Chronic hepatits B virus infection 3 (13) 12 (24)
reduction in depression [22]. With regard to secondary outcomes,
Autoimmune hepatitis 1 (4) 8 (16)
Alcoholic liver disease 1 (4) 2 (4) LFTs did not change during the study in either groups. As expected
NAFLD 3 (13) 5 (10) given the safety of vitamin D supplementation at the dose used and
Other 5 (23) 6 (12) the sample size, no major study-related adverse events occurred.
Liver cirrhosis, n (%) 10 (43)* 9 (18)* One patient complained of general malaise and withdrew from the
Depression based on BDI-II category, n (%)
None (0e13 points) 0 50 (100)
study; another patient was withdrawn due to acute hearing loss,
Mild (14e19 points) 13 (57) 0 which was assessed as unrelated to the study.
Moderate (20e28 points) 7 (30) 0
Severe (29e63 points) 3 (13) 0
BDI-II score, median (range) 18 (14e39) 6 (0e13)
Biochemical serum markers, median (range) 3.2.3. Subgroup analyses identified women as responders
25-hydroxyvitamin D (ng/ml) 17.0 (4.3e29.3) 12.2 (4.0e27.7) Depression is known to be more prevalent in females [25]. We
Calcium (mmol/l) 2.4 (2.0e2.6) 2.4 (2.0e2.6)
therefore carried out a subgroup analysis comparing response to
Phosphate (mg/dl) 3.3 (1.7e4.5) 3.1 (1.3e4.1)
PTH (pg/ml) 40 (18e90) 39.5 (16.0e89.0) vitamin D therapy between sexes. Vitamin D levels significantly
AST (U/l) 40 (26e183) 35 (16e263) (P < 0.001) increased to normal after three and six months in both
ALT (U/l) 54 (11e276)* 35 (8e340)* sexes, but non-significantly higher 25-hydroxyvitamin D concen-
g-GT (U/l) 80 (20e914) 49.5 (12.0e502.0) trations were displayed in women (41.1 and 39.5 ng/ml) as
AP (U/l) 91 (44e256) 79.5 (36.0e303.0)
Creatinine (mg/dl) 0.7 (0.6e1.0)** 0.9 (0.5e1.8)**
compared to men (35.1 and 36.1 ng/ml). During the intervention
Urea (mg/dl) 29.5 (16.0e60.0) 32.0 (4.6e58.0) period, a distinct response on the BDI-II questionnaire was
observed in women with scores returning to normal levels at three
Significant differences between depressed (BDI-II
14) and non-depressed patients
(BDI-II <14) are denoted with * (P < 0.05) and ** (P < 0.001). and six months (13 and 11 points, both P ¼ 0.002) as compared to a
Abbreviations: see Table 1.

Please cite this article in press as: Stokes CS, et al., Vitamin D supplementation reduces depressive symptoms in patients with chronic liver
disease, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.07.004
6 C.S. Stokes et al. / Clinical Nutrition xxx (2015) 1e8

Fig. 4. Changes in depressive symptoms during the study. Depressive symptoms, as assessed with the BDI-II score, demonstrated significant improvements (panel A) after 3 (T1)
and 6 (T2) months of vitamin D supplementation, as compared to baseline (T0); there was a non-significant deterioration 6 months post supplementation, but this value was still
significantly lower than baseline. Horizontal dashed line denotes cut-off for normal level. Depressed patients were also categorised based on deltas (absolute changes) of BDI-II
scores with vitamin D supplementation for (panel B) 3 months and (panel C) 6 months. A negative change denotes an improvement (i.e. reduction of BDI-II score), and a posi-
tive change is a deterioriation (i.e. a higher score).

score of 21 at baseline, whereas no change was observed in men (18 after only three months of vitamin D supplementation, which
and 19 points), in comparison to a score of 18 at baseline. coincided with establishing normal serum 25-hydroxyvitamin D
The BDI-II can be subdivided into cognitive and somatic sub- concentrations in most patients. Interestingly, a subgroup analysis
scales and both subscales improved equally in women. In depressed revealed marked improvements in depression to occur predomi-
patients, a markedly higher BMI was illustrated for the 11 female nantly in women.
responders as compared to the five non-responders (30.8 vs. The cross-sectional analysis identified 31% of patients as hav-
26.3 kg/m2); we observed no differences between the two male ing depression. This finding is in line with those of others, with
responders and the five non-responders. Moreover, no within one in every three CLD patients presenting with depressive
subject changes in BMI were detected during the study. All results symptoms [10]. Eighty percent of the cohort also had inadequate
still remained significant after removing patients over 65 years. vitamin D levels, further reinforcing previous studies reporting
vitamin D deficiency in these patients [2]. Furthermore, our re-
3.3. Observational (post-intervention) results sults concur with earlier studies and a meta-analysis showing
inverse associations of depressive symptoms with vitamin D
After six months without vitamin D supplementation (T3), both levels in healthy populations [8,26]. Though such an association
groups showed declines in vitamin D levels, which were significant does not prove causality, studies investigating the efficacy of
for both depressed (21.3 ng/ml, P ¼ 0.0004) and non-depressed vitamin D replacement therapy for depression look e albeit
patients (21.0 ng/ml, P < 0.0001). This coincided with increased inconsistently e promising [27]. A recent meta-analysis of rand-
BDI-II scores in patients with depression (median score 17; Fig. 4, omised controlled trials (RCTs) in fact reported the favourable
panel A). For individual patient data, a clinically significant dete- effect of
800 IU/day vitamin D supplementation for the man-
rioration (þ5 points on the BDI-II) was observed in 21% of patients. agement of depression to be similar to that of anti-depressant
medication when accompanied by an increase in serum 25-
4. Discussion hydroxyvitamin D concentrations [28].
One of the main criticisms regarding contradictory findings in
The findings herein support the primary study hypothesis that association and intervention studies is, however, that many do not
the correction of vitamin D deficiency ameliorates depressive correct existing vitamin D deficiencies but rather administer
symptoms in patients with CLD. In our outpatient cohort, we found blanket supplementation across cohorts who are not all vitamin D
a reduction in depressive symptoms as assessed with the BDI-II depleted [29]. It might be more relevant to assess whether

Please cite this article in press as: Stokes CS, et al., Vitamin D supplementation reduces depressive symptoms in patients with chronic liver
disease, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.07.004
 C.S. Stokes et al. / Clinical Nutrition xxx (2015) 1e8 7

correcting vitamin D deficiency provides such improvements, and Funding and competing interests
the meta-analysis [28] controlled for this and found significant
benefits. Heaney highlighted [29] that nutrients such as vitamin D The present study did not receive external funding. All authors
act dissimilarly to drugs, and that a null response to supplemen- declare that no support was received from any organisation for the
tation may be seen in those with almost adequate stores of the submitted work.
nutrient in question. We achieved very high success rates in terms
of improving vitamin D concentrations by supplementing all pa- Contributorship statement
tients with a deficiency with 20,000 IU cholecalciferol per week.
This dose appears to circumvent vitamin D metabolism-related CS, CB, MR and FL designed the study. CS collected the study
defects that might be expected in patients with CLD, given the data. All authors contributed to the analysis and interpretation of
inherent role of the liver in its activation. the work. All authors revised the manuscript for important intel-
The strength of this study is its prospective nature. Moreover, lectual content and the final version has been approved by all au-
vitamin D levels were cross-checked with a gold standard LC-MS/ thors for publication.
MS method. A limitation is that the intervention was not an RCT,
a design we avoided also based on ethical grounds as we did not
Conflict of interest
want to withhold treatment for vitamin D deficiency. A further
limitation is the relatively small sample size in the respective
None declared.
groups. Given that vitamin D therapy significantly improved
depressive symptoms to normal levels predominantly in women,
the functional mechanisms underlying this preliminary finding Acknowledgements
warrant further investigation, particularly because of the small
sample size in this subgroup. The authors would like to thank the team in the Department of
There are numerous plausible mechanisms by which vitamin D Medicine II for CFF tests. CSS is also very grateful to the European
supplementation may reduce depressive symptoms. Firstly, it is Association for the Study of the Liver (EASL) for a full bursary to
known that not only 1,25-dihydroxyvitamin D but also 25- present some of the findings of this paper at the 2014 International
hydroxyvitamin D crosses the blood brain barrier [30]. Recently, Liver Congress.
vitamin D was reported to regulate the synthesis of serotonin (5- This manuscript has been presented, in part, at the Annual
hydroxyltryptamine [5-HT]) in brain known to affect mood [31]. Meeting of the European Association for the Study of the Liver
Of major interest are the recently identified vitamin D response (EASL), London, April 2014, and published in abstract form in the
elements (VDREs) on two tryptophan hydroxylase (TPH) genes that Journal of Hepatology 2014; 60(1):S226.
have been associated with serotonin synthesis [32]. An experi-
mental study in which rats with hepatic steatosis received chole- References
calciferol reported improved cognitive effects and reduced TNF-a
levels in brain [33]. Therefore, the observed benefits may have [1] Crawford BA, Labio ED, Strasser SI, McCaughan GW. Vitamin D replacement
for cirrhosis-related bone disease. Nat Clin Pract Gastroenterol Hepatol
resulted from the anti-inflammatory properties of vitamin D [34], a 2006;3:689e99.
finding that supports the inflammatory hypothesis for some types [2] Arteh J, Narra S, Nair S. Prevalence of vitamin D deficiency in chronic liver
of depression. disease. Dig Dis Sci 2010;55:2624e8.
[3] Zún~ iga S, Firrincieli D, Housset C, Chignard N. Vitamin D and the vitamin D
We have previously demonstrated that vitamin D deficiency receptor in liver pathophysiology. Clin Res Hepatol Gastroenterol 2011;35:
increases mortality in those with advanced cirrhosis, and this as- 295e302.
sociation was also found in patients with alcoholic-related liver [4] Eyles DW, Smith S, Kinobe R, Hewison M, McGrath JJ. Distribution of the
vitamin D receptor and 1 alpha-hydroxylase in human brain. J Chem Neuro-
disease [35,36]. Thus, there are many justifications for correcting
anat 2005;29:21e30.
vitamin D concentrations in CLD and, coupled with the fact that [5] Kuningas M, Mooijaart SP, Jolles J, Slagboom PE, Westendorp RG, van
these patients often suffer from co-morbid depression, vitamin D Heemst D. VDR gene variants associate with cognitive function and depressive
symptoms in old age. Neurobiol Aging 2009;30:466e73.
substitution could play a more prominent part in future clinical
[6] Groves NJ, Kesby JP, Eyles DW, McGrath JJ, Mackay-Sim A, Burne TH. Adult
care of patients presenting with CLD and inadequate vitamin D vitamin D deficiency leads to behavioural and brain neurochemical alterations
levels. in C57BL/6J and BALB/c mice. Behav Brain Res 2013;241:120e31.
In conclusion, a large majority of patients with CLD present [7] Ozkayar N, Altun B, Ulusoy S, Yildirim T, Halil M, Yilmaz R, et al. Relationship
between vitamin D levels and depressive symptoms in renal transplant re-
with inadequate vitamin D levels, which appear to correlate with cipients. Int J Psychiatry Med 2014;47:141e51.
depressive symptoms. No study has to date evaluated the [8] Anglin RE, Samaan Z, Walter SD, McDonald SD. Vitamin D deficiency and
possibility of vitamin D correction in ameliorating depressive depression in adults: systematic review and meta-analysis. Br J Psychiatr J
Ment Sci 2013;202:100e7.
symptoms in CLD. Interestingly, the results herein indicate that [9] Stewart CA, Enders FT, Mitchell MM, Felmlee-Devine D, Smith GE. The
vitamin D substitution might significantly improve depressive cognitive profile of depressed patients with cirrhosis. Prim Care Companion
symptoms in CLD in a manner which is clinically meaningful, and in CNS Disord 2011:13.
[10] Mullish BH, Kabir MS, Thursz MR, Dhar A. Review article: depression and the
particular in women with low baseline vitamin D levels. Larger use of antidepressants in patients with chronic liver disease or liver trans-
RCTs are needed to substantiate this finding and to further explore plantation. Aliment Pharmacol Ther 2014;40:880e92.
the possible gender disparities. Vitamin D nevertheless, holds po- [11] Anglin R, Yuan Y, Moayyedi P, Tse F, Armstrong D, Leontiadis GI. Risk of upper
gastrointestinal bleeding with selective serotonin reuptake inhibitors with or
tential as a cost-effective adjunctive measure with minimal side without concurrent nonsteroidal anti-inflammatory use: a systematic review
effects. and meta-analysis. Am J Gastroenterol 2014;109:811e9.
[12] Hung CI, Liu CY, Hsiao MC, Yu NW, Chu CL. Metabolic syndrome among
psychiatric outpatients with mood and anxiety disorders. BMC Psychiatry
Ethical standards
2014;14:185.
[13] Jorde R, Sneve M, Figenschau Y, Svartberg J, Waterloo K. Effects of vitamin D
The authors declare that all procedures related to this study supplementation on symptoms of depression in overweight and obese sub-
comply with the ethical standards of the respective national and jects: randomized double blind trial. J Intern Med 2008;264:599e609.
[14] Kircheis G, Wettstein M, Timmermann L, Schnitzler A, Haussinger D. Critical
institutional committees on human experimentation and with the flicker frequency for quantification of low-grade hepatic encephalopathy.
Declaration of Helsinki. Hepatology 2002;35:357e66.

Please cite this article in press as: Stokes CS, et al., Vitamin D supplementation reduces depressive symptoms in patients with chronic liver
disease, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.07.004
8 C.S. Stokes et al. / Clinical Nutrition xxx (2015) 1e8

[15] Antoniucci DM, Sellmeyer DE, Bilezikian JP, Palermo L, Ensrud KE, [25] Karger A. Gender differences in depression. Bundesgesundheitsblatt,
Greenspan SL, et al. Elevations in serum and urinary calcium with parathyroid Gesundheitsforsch Gesundheitsschutz 2014;57:1092e8.
hormone (1-84) with and without alendronate for osteoporosis. J Clin Endo- [26] Chung HK, Cho Y, Choi S, Shin MJ. The association between serum 25-
crinol Metab 2007;92:942e7. hydroxyvitamin D concentrations and depressive symptoms in Korean
[16] K/DOQI clinical practice guidelines for chronic kidney disease. evaluation, adults: findings from the Fifth Korea National Health and Nutrition Exami-
classification, and stratification. American journal of kidney diseases. Official J nation Survey 2010. PLoS One 2014;9:e99185.
Natl Kidney Found 2002;39:S1e266. [27] Annweiler C, Rastmanesh R, Richard-Devantoy S, Beauchet O. The role of
[17] Geib T, Meier F, Schorr P, Lammert F, Stokes CS, Volmer DA. A simple micro- vitamin D in depression: from a curious idea to a therapeutic option. J Clin
extraction plate assay for automated LC-MS/MS analysis of human serum 25- Psychiatry 2013;74:1121e2.
hydroxyvitamin D levels. J Mass Spectrom 2014;50:275e9. [28] Spedding S. Vitamin D and depression: a systematic review and meta-analysis
[18] Wareham NJ, Jakes RW, Rennie KL, Schuit J, Mitchell J, Hennings S, et al. comparing studies with and without biological flaws. Nutrients 2014;6:
Validity and repeatability of a simple index derived from the short physical 1501e18.
activity questionnaire used in the European Prospective Investigation into [29] Heaney RP. Vitamin Debaseline status and effective dose. N. Engl J Med
Cancer and Nutrition (EPIC) study. Public Health Nutr 2003;6:407e13. 2012;367:77e8.
[19] Babor TF, Higgins-Biddle JC, Saunders JB. The alcohol use disorders identifi- [30] Kesby JP, Eyles DW, Burne TH, McGrath JJ. The effects of vitamin D on brain
cation test. Guidelines for use in primary health care. Second ed. Geneva: development and adult brain function. Mol Cell Endocrinol 2011;347:
World Health Organisation; 2001. 121e7.
[20] World Health Organisation. International statistical classification of diseases [31] Stockmeier CA. Involvement of serotonin in depression: evidence from
and related health problems. Geneva: World Health Organisation; 2011. postmortem and imaging studies of serotonin receptors and the serotonin
[21] Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for transporter. J Psychiatr Res 2003;37:357e73.
measuring depression. Arch Gen Psychiatry 1961;4:561e71. [32] Zhang X, Beaulieu JM, Sotnikova TD, Gainetdinov RR, Caron MG. Tryptophan
[22] Dworkin RH, Turk DC, Wyrwich KW, Beaton D, Cleeland CS, Farrar JT, et al. hydroxylase-2 controls brain serotonin synthesis. Science 2004;305:217.
Interpreting the clinical importance of treatment outcomes in chronic pain [33] Erbas O, Solmaz V, Aksoy D, Yavasoglu A, Sagcan M, Taskiran D. Cholecalciferol
clinical trials: IMMPACT recommendations. J Pain Off J Am Pain Soc 2008;9: (vitamin D 3) improves cognitive dysfunction and reduces inflammation in a
105e21. rat fatty liver model of metabolic syndrome. Life Sci 2014;103:68e72.
[23] Hiroe T, Kojima M, Yamamoto I, Nojima S, Kinoshita Y, Hashimoto N, et al. [34] Vojinovic J. Vitamin D receptor agonists' anti-inflammatory properties. Ann N.
Gradations of clinical severity and sensitivity to change assessed with the Y Acad Sci 2014;1317:47e56.
Beck Depression Inventory-II in Japanese patients with depression. Psychiatry [35] Stokes CS, Krawczyk M, Reichel C, Lammert F, Grünhage F. Vitamin D defi-
Res 2005;135:229e35. ciency is associated with mortality in patients with advanced liver cirrhosis.
[24] Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, Eur J Clin Invest 2014;44:176e83.
Heaney RP, et al. Evaluation, treatment, and prevention of vitamin D defi- [36] Trepo E, Ouziel R, Pradat P, Momozawa Y, Quertinmont E, Gervy C, et al.
ciency: an endocrine society clinical practice guideline. J Clin Endocrin Metab Marked 25-hydroxyvitamin D deficiency is associated with poor prognosis in
2011;96:1911e30. patients with alcoholic liver disease. J Hepatol 2013;59:344e50.

Please cite this article in press as: Stokes CS, et al., Vitamin D supplementation reduces depressive symptoms in patients with chronic liver
disease, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.07.004