Pharmacological Research 187 (2023) 106605

Contents lists available at ScienceDirect

Pharmacological Research
journal homepage: www.elsevier.com/locate/yphrs

Vitamin D protects against depression: Evidence from an umbrella

meta-analysis on interventional and observational meta-analyses
Vali Musazadeh a, b, Majid Keramati a, b, Faezeh Ghalichi a, c, Zeynab Kavyani a, b,
Zohre Ghoreishi d, Kamar Allayl Alras e, Naryman Albadawi f, Abdullah Salem e,
Mohamed Ismail Albadawi g, Raghad Salem h, Ahmed Abu-Zaid e, i, *, 1, Meysam Zarezadeh a, c,
Rania A. Mekary j, **
a
Student Research Committee, Nutrition Research Center, School of Nutrition and Food Sciences, Tabriz University of Medical Sciences, Tabriz, Islamic Republic of Iran
b
School of Nutrition and Food Sciences, Tabriz University of Medical Sciences, Tabriz, Islamic Republic of Iran
c
Nutrition Research Center, School of Nutrition and Food Sciences, Tabriz University of Medical Sciences, Tabriz, Islamic Republic of Iran
d
Department of Clinical Nutrition, Faculty of Nutrition and Food Science, Tabriz University of Medical Sciences, Tabriz, Islamic Republic of Iran
e
College of Medicine, Alfaisal University, Riyadh, Saudi Arabia
f
Department of Medicine, King Faisal Specialist Hospital and Research Center, Riyadh, Saudi Arabia
g
College of Medicine, Almaarefa University, Riyadh, Saudi Arabia
h
College of Medicine, Princess Nourah Bint Abdul Rahman University, Riyadh, Saudi Arabia
i
College of Graduate Health Sciences, University of Tennessee Health Science Center, Memphis, TN, USA
j
School of Pharmacy, MCPHS University, Boston, MA 02115, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Meta-analyses of interventional and observational studies investigating the efficacy and the relationship between
Vitamin D vitamin D and depression provided inconsistent results. The current umbrella meta-analysis was conducted to
Depression assess the available evidence and provide a conclusive outcome in this regard. The following international da­
25-hydroxyvitamin D
tabases were systematically searched till March 2022: PubMed, Scopus, Embase, Web of Science, and Google
Umbrella meta-analysis
Scholar. Random-effects model was carried out to calculate the pooled point estimates and their respective 95 %
confidence intervals (CI). Ten meta-analyses of randomised controlled trials (RCTs) revealed significant reduc­
tion in depression symptoms comparing participants on vitmain D supplements to those on placebo (Pooled
standardised mean difference: − 0.40; 95 % CI: − 0.60, − 0.21, p < 0.01: I2 = 89.1 %, p < 0.01). Four meta-
analyses of cohort studies (with one having two subgroups) revealed that participants with lower levels of
serum vitamin D were at increased odds of depression than those with higher levels of serum vitamin D (Pooled
odds ratio: 1.60; 95 % CI: 1.08, 2.36, p < 0.01; I2 = 91.3 %, p < 0.01). The present umbrella meta-analysis
confirms the potential benefits of vitamin D supplementation and higher serum vitamin D levels in reducing
the development and symptoms of depression.

1. Introduction number of studies have demonstrated a rise in depression psychopa­

thology and suicide tendencies across a variety of countries [4].
Depression is a mental disorder that causes disabling effects of mood Depressive symptoms were the most common mental health condition
and anxiety disorders. Depression has also become a leading global during the COVID-19 pandemic, ranging from 14.6 % to 48.3 % across
cause of disease burden [1,2]. Based on the evidence from the World all populations, according to a systematic review [5]. People who suffer
Health Organization, more than 264 million people are affected by from depression may feel sad, anxious, hopeless, helpless, irritable,
depression worldwide [3]. In the wake of the COVID-19 outbreak, a worthless, guilty, or ashamed [6]. There might also be decreased

* Correspondence to: College of Graduate Health Sciences, University of Tennessee Health Science Center, Memphis, TN 38163, USA.
** Correspondence to: School of Pharmacy, MCPHS University, 179 Longwood Avenue, Boston, MA 02115, USA.
E-mail addresses: [email protected] (A. Abu-Zaid), [email protected] (R.A. Mekary).
1
https://orcid.org/0000-0003-2286-2181.

https://doi.org/10.1016/j.phrs.2022.106605
Received 26 August 2022; Received in revised form 22 November 2022; Accepted 7 December 2022
Available online 9 December 2022
1043-6618/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
V. Musazadeh et al. Pharmacological Research 187 (2023) 106605

appetite or overeating, and inability to exercise, or even suicide among settled by consensus with the third reviewer (MZ).
them [7]. Antidepressants, which have been used for many years to treat The following data were extracted from the selected papers: publi­
depression, have raised concerns about their effectiveness and tolerance cation year, sample size, the dosage and duration of the intervention of
[8–10]. Furthermore, the failure of depression to respond to a wide vitamin D supplementation in RCTs, follow-up duration in observational
range of pharmaceutical treatments [11] indicates that other mecha­ studies, and the SMD or OR and their 95 % confidence intervals.
nisms are involved in the pathogenesis of depression, such as those
affecting neuroendocrine, immunological, neurotrophic, and metabolic 2.4. Quality assessment
systems [12]. In spite of these challenges, complementary treatments for
depression appears to be helpful. Two reviewers (ZK and MK) independently assessed the methodo­
Vitamin D is a unique neurosteroid hormone that may play a role in logical quality of the qualifying papers using the assessment of multiple
depression [13]. Vitamin D has numerous functions in the brain such as systematic reviews (AMSTAR2) questionnaire. The AMSTAR2 ques­
neuroimmunomodulation, regulation of neurotrophic factors, neuro­ tionnaire includes 16 items that asks reviewers to reply ‘Yes’ or ‘Partial
protection, neuroplasticity, and brain development [14]. Also, vitamin Yes’ or ‘No’ or ‘No Meta-analysis’. The AMSTAR2 checklist was cat­
D receptors can be found on the neurons and glia in many parts of the egorised into “critically low quality”, “low quality”, “moderate quality”,
brain, including the cingulate cortex and the hippocampus [15]. and “high quality” [25]. Also the third reviewer (RM) solved any
Vitamin D deficiency may have played a significant role in stress-related disagreements.
depression during the COVID-19 pandemic, according to a growing body
of literature [16]. Vitamin D is thought to influence the serotoninergic 2.5. Data synthesis and statistical analysis
system and contribute to the maintenance of circadian rhythms, both of
which are associated with depressive symptoms [16,17]. As a result, it is The overall effect sizes were calculated by pooling the point esti­
biologically plausible that vitamin D might play an important role in the mates and their respective 95 % CIs for observational and RCT studies,
treatment of depressive disorders [13]. However, evidence is accumu­ separately using the random effects model by DerSimonian and Laird
lating that vitamin D may have beneficial effects regarding the depres­ [26]. To detect statistical heterogeneity, the I2 index and Cochrane’s Q
sive disorders. In this context, many meta analyses of RCTs and test were utilised. An I2 value of more than 40 % or a P < 0.1 for the
observational studies have been published over the last few years. Q-test was considered as significant between-study heterogeneity. When
Several RCT studies showed a beneficial effect of vitamin D supple­ feasible, we conducted subgroup analyses based on vitamin D dosage (<
mentation on depression [18–21]. On the other hand, other studies did 4000; 4000–5000; > 5000 IU/day), intervention duration (≤ 20, > 20
not report a significant effect [22–24]. weeks), and average age (≤ 50/> 50 years) to detect possible hetero­
Therefore, as the results are conflicting and no definite conclusion geneity sources. We also conducted a sensitivity analysis in which each
could be obtained from the existing meta-analyses, the present umbrella study was excluded to examine the impact of that study on the pooled
meta-analysis was conducted to propose whether vitamin D supple­ point estimate. For outcomes with at least 10 meta-analyses, the
mentation or higher serum vitamin D levels had a protective role against small-study effect was examined performing the formal tests of Egger’s
depression and could hence be considered as a reliable therapeutic [27] and Begg’s [28] and if these last were significant, a visual evalu­
approach. ation of funnel plots was conducted. If an asymmetry was found in the
funnel plot and contingent on publication bias being the reason, the
2. Methods trim-and-fill method was used to detect the effect of the potentially
missing small studies on the overall effect. We used version 16.0 of
2.1. Search strategy and study selection STATA to conduct all statistical analyses (Stata Corporation, College
Station, TX). Unless otherwise specified, significant level was defined as
EMBASE, Scopus, Web of Science, Cochrane Central Library, and a p-value < 0.05.
PubMed scientific databases in addition to Google Scholar were checked
for relevant papers published up to March, 2022. The search strategy is 3. Results
shown in Suppl. Table 1.
To improve the sensitivity of our search strategy, we used the wild- 3.1. Study selection
card term"*". Only studies published in English were included in the
current study. Additionally, the reference list of related articles was Following a thorough search of electronic databases, a total of 300
checked for any missing eligible studies. papers were found. After removing 61 duplicates, 239 studies were
discarded due to their irrelevant titles and abstracts (n = 179), animal
2.2. Inclusion and exclusion criteria studies (n = 20), and review studies (n = 17). In the end, 23 full texts
were reviewed. Eight studies were excluded for lack of required infor­
In the current study, we included meta-analyses of randomised mation. Finally, 14 meta-analyses (One meta-analysis reported separate
controlled trials (RCTs) and of observational studies (cohort and cross- pooled point estimates for RCT studies and for observational studies,
sectional) that investigated the effect of vitamin D supplementation on finally, 10 effect size for RCTs and five effect size for observational
depression symptoms considering the following criteria: reporting studies) met all of our inclusion criteria. Note that four of these meta-
standardised mean difference (SMD), or odds ratio (OR) and their cor­ analyses reported separate pooled point estimates for cohort studies
responding confidence intervals (CI) for vitamin D supplementation on and for cross-sectional studies.The study selection process is schemati­
depression symptoms. Other studies were excluded from this review, cally depicted in the PRISMA study flow chart in Fig. 1.
including original experimental studies, case reports, in vitro, ex-vivo,
and in vivo investigations. 3.2. Study characteristics

2.3. Data extraction In the umbrella meta-analysis of RCTs, there were 10 meta-analyses
(24,510 participants from 49 RCTs) that reported weighted mean dif­
Two independent reviewers (ZK and MK) screened papers based on ferences in depression risk score comparing the vitamin D arm to the
the eligibility criteria. In the first step, the reviewers reviewed papers by placebo arm. The included meta-analyses were performed between 2014
titles and abstracts. Then, they evaluated the full texts of relevant papers and 2021. The number of subjects ranged between 66 and 42,226. The
to determine suitability for the meta-analysis. Any disagreement was average age of participants ranged between 37 and 57 years.

2
V. Musazadeh et al. Pharmacological Research 187 (2023) 106605

were performed between 2013 and 2021. The number of subjects

ranged between 383 and 12,648. The average age of participants ranged
between 28 and 73 years. Follow-up duration was between 6 months
and 3.5 years.
In the umbrella meta-analysis of cross-sectional studies, three meta-
analyses reported OR (66,409 participants from 13 cross-sectional
studies). The included meta-analyses were conducted between 2013
and 2021. The number of participants ranged from 796 to 43,137 with
an average age of 25–63 years.

3.3. Risk of bias assessment

The findings of the AMSTAR2 questionnaire-based quality assess­

ment was shown in Supplementary Table 2. Out of ten meta-analyses of
RCTs, eight [29–36] were of high-quality and two [37,38] were of
moderate-quality. Also, four meta-analyses [13,39–41] of observational
studies had high-quality and one [37] had moderate-quality.

3.4. Effects of vitamin D supplementation on depression according meta-

analyses of RCTs

Vitamin D supplementation had (Table 2) a significant effect on

decreasing depression symptoms (ESSMD: − 0.40; 95 % CI: − 0.60,
− 0.21, p < 0.01), according to the pooled analysis of 10 meta-analyses
Fig. 1. PRISMA flowchart diagram. (Fig. 2A). There was a significant between-study heterogeneity (I2
= 89.1 %, p-heterogeneity < 0.01) which the dosage of vitamin D,
Intervention duration was between 8 and 74 weeks. Vitamin D dosages sample size, and duration of intervention were determined as its sources,
used varied between 2500 and 6000 IU/day (Table 1). Jamilian et al. after performing subgroup analyses (Table 3). Vitamin D supplementa­
[29] presented data as weighted mean difference (WMD) (− 3.9;1 95 % tion in dosage of 4000–5000 IU/day, or intervention duration of ≤ 20-
CI; − 5.15, − 2.66), which was converted to SMD ( − 0.17; 95 % CI; weeks appeared to have a stronger reduction in depression symptoms
− 0.23, − 0.12) based on statistical methods. compared to other respective subgroups (Table 3). Based on the one-
In the umbrella meta-analysis of cohort studies, there were five meta- study removal analysis, no significant change was observed after
analyses (38,237 participants from 16 cohort studies) that reported removing a single study at a time (Suppl. File. 1). There was no signif­
weighted OR for depression comparing lower serum levels of vitmain D icant small-study effect according to the results of Egger’s (p = 0.21)
(vitamin D deficiency) to higher levels. The included meta-analyses and Begg’s tests (p = 0.78). However, visual inspection of the funnel
plot showed an asymmetric distribution of studies indicating a potential

Table 1
Characteristics of included meta-analyses of RCT studies.
Study, year, country No. of No. of Mean Duration of Intervention/daily Quality Measured outcomes and
studies in the participants in age intervention dose assessment results
meta-analysis the meta-analysis scale and
outcome

Li, 2014, Canada 6 1203 (healthy and with 57 74 weeks Vit D/3700 IU/day NR Depression → ↓
depression)
Spedding et al. 2014 15 42,226 (healthy and with NR 41 weeks Vit D/5200 IU/day Yes Depression → ↓
Australia depression) (Cochrane)
10/15 high
Shaffer et 7 3191 (healthy and with 47 34 weeks Vit D/2500 IU/day Yes Depression → not
al. 2014 Columbia depression) (Cochrane) significant effect
3/7 high
Gowda et al. 2015 9 4923 with depression 45 61 weeks Vit D/3070 IU/day Yes Depression → not
Australia (Cochrane) significant effect
7/9 high
Firth et 4 948 with unipolar NR NR Vit D/4300 IU/day NR Depression → ↓
al. 2019 Australia depression
Vellekkatt et al. 2019 4 948 with major depression NR 20 weeks Vit D/4800 IU/day Yes Depression → ↓
India (Cochrane)
2/4 high
Jamilian et al. 2019 9 1347 with psychiatric 40 13 weeks Vit D/6000 IU/day NR Depression → ↓
Iran disorders
Cheng et al. 2020 25 9840 (healthy and with 47 32 weeks Vit D/4200 IU/day NR Depression → ↓
Taiwan depression)
Jeremiah et al. 2020 2 66 with depression 37 8 weeks Vit D/4250 IU/day Yes Depression → ↓
Ireland (Cochrane)
0/2 high
Nicoláse et al. 2021 10 1393 with depression 46 20 weeks Vit D/5250 IU/day Yes Depression → not
Spain (Cochrane) significant effect
8/10 high

IU: International units; NR, not reported; vit D, vitamin D.

3
V. Musazadeh et al. Pharmacological Research 187 (2023) 106605

Table 2
Characteristics of the included meta-analyses for observational studies.
Study, year Study design Participants Gender Mean age, Follow-up Outcome
years duration

Meta-analyses of cohort studies

†Anglin, 2013, Canada Cohort 8815 Both genders 73 3.6 Years ↑ risk of depression
†Ju, 2013, Korea Cohort 12648 Both genders 65 NR ↑ risk of depression
Li, 2014, Canada Cohort 383 Female 73 3.5 years No statistically significant association
§Wang,2018, China Cohort 8470 Female 28 6 months ↑ risk of postpartum depression
§Wang, 2018, China Cohort 8470 Female 28 6 months No statistically significant association during pregnancy
†Tan, 2021, China Cohort 7921 Both genders 28 NR ↑ risk of depression
Meta-analyses of cross-sectional studies
†Anglin, 2013, Canada Cross-sectional 22,476 Both genders 63 NR ↑ risk of depression
†Ju, 2013, Korea Cross-sectional 43,137 Both genders 50 NR ↑ risk of depression
†Tan, 2021, China Cross-sectional 796 Both genders 25 NR ↑ risk of depression

NR, not reported; vit D, vitamin D.

† These meta-analyses reported separate pooled point estimates for cohort studies and for cross-sectional studies.
‡ Only the subgroup of original studies that included participants not on antidepressants was included.
§ Wang et al. was stratified in two groups, one for pregnant women and one for women post-delivery.

Fig. 2. Forest plot (A) funnel plot with mean difference and 95 % confidence intervals (CIs) (B) publication bias in the studies the effects of vitamin D supple­
mentation on depression symptoms.

presence of small study effect (Fig. 2B). Therefore, the trim and fill suggesting an increased hazard of depression in the lowest serum vit­
method was carried out with 12 studies (two imputed studies) and re­ main D compared to the highest (pooled HR = 2.21; 95 % CI; 1.40,
sults remained significant (ESSMD: − 0.33; 95 % CI: − 0.52, − 0.13, 3.49).
p < 0.05).
3.6. Association between serum vitamin D and depression according to
3.5. Association between vitamin D and depression according to meta- meta-analyses of cross-sectional studies
analyses of cohort studies
The association between serum vitamin D and depression was re­
The association between Vitamin D and protection against depres­ ported in three meta-analyses of cross-sectional studies with 66,411
sion risk was examined in four meta-analyses with five effect sizes that participants.The summary effect size for overall depression indicated no
included 38,237 participants. Our findings revealed a significant pro­ significant protective association between serum vitamin D and overall
tective association between serum vitamin D and overall depression risk depression (Pooled ESOR: 1.19; 95 % CI: 0.95, 1.49, p = 0.14; I2
(Pooled ESOR: 1.60; 95 % CI: 1.08, 2.36, p < 0.01) (Fig. 3). The level of = 53.9 %, p = 0.11) (Fig. 4). No subgroup analysis was performed on
heterogeneity was high (I2 = 91.3 %, p-heterogeneity < 0.01). Sub­ these meta-analyses.
group analysis showed that the protective association between higher
levels of serum vitamin D and depression was stronger among partici­ 4. Discussion
pants aged ≤ 50 years old than their older peers (Table 3). Furthermore,
after excluding the study of Tan et al. [40], and Wang et al. [41], using The current umbrella meta-analysis summarised 15 meta-analyses,
one-study removal analysis, the significance was lost (ESOR: 1.43; 95 % which included 65 RCTs, and 31 observational (cohort and cross-
CI: 0.98, 2.10), and (ESOR: 1.41; 95 % CI: 0.95, 2.09) (Suppl. File. 1). sectional) studies. According to the results, vitamin D supplementation
One meta-analysis by Anglin et al. [13] presented a pooled hazard ratio was efficient in alleviating symptoms of depression and an inverse as­
(HR) from three cohort studies and was hence not included in our um­ sociation was observed between higher serum levels of vitamin D intake
brella meta-analysis of cohort studies which reported mainly OR. The and overall depression. Based on sub-group analyses, vitamin D sup­
direction of the results of the HR were in line with our findings plementation in studies using dosage of > 5000 IU/day, and

4
V. Musazadeh et al. Pharmacological Research 187 (2023) 106605

Table 3 [32]. In general, vitamin D prevents the onset of depression by taking

Subgroup analyses for the effects of vitamin D supplementation on depression role in six main pathways: 1) Controlling the expression of calcium
symptoms. homoeostasis genes; 3) Controlling serotonin synthesis via alleviating
Variables No. Pooled point P-value I2 P- tryptophan hydroxylase 2 (TPH2) expression and repressing tryptophan
studies estimate (%) heterogeneity hydroxylase1 (TPH1); 4) Controlling inflammation by reducing the
(95 % CI) expression of inflammatory cytokines; 5) Controlling the expression of
Meta-analysis of RCTs (SMD) mitochondrial proteins that preserve normal mitochondrial respiration;
Overall 10 -0.40 < 0.001 89.1 < 0.001 and 6) Preventing the hypermethylation of gene promotors such as
(− 0.60,
Jumonji domain-containing protein 1A and 3 (JMJD1A, JMJD3) and
− 0.21)
Dosage (IU/ lysine-specific demethylase 1 and 2 (LSD1, LSD2). These genes have a
day) significant role in the activation of GABAergic neurons [29,42]. Fig. 5
0.234 exhibits the mechanism of action of vitamin D in preventing and
< 4000 3 -0.09 < 0.001 42.0 0.178 lowering symptoms of depression.
4000–5000 4 (− 0.23, < 0.001 51.1 0.105
> 5000 3 0.06) 71.5 0.030
Although the majority of RCTs have reported beneficial effects of
-0.59 vitamin D supplementation on depression and observational studies
(− 0.68, have confirmed the inverse association between serum vitamin D levels
− 0.50) and depression, a significant between-study heterogeneity was observed
-0.18
among the included studies and a few studies observed contradictory
(− 0.23,
− 0.12) results. Explanation for the inconsistent results in cohort studies can be
Intervention found below: First, various study populations were included (tubercu­
duration losis, diabetes, fatigue, dialysis, and bipolar depression). Second,
(weeks) various scales were administered for measuring depressive symptoms.
4 -0.24 < 0.001 93.4 < 0.001
Third, some interventions had a shorter duration than others. Fourth,
≤ 20 5 (− 0.29, 0.004 70.1 0.010
> 20 1 − 0.19) < 0.001 – – there were differences in baseline vitamin D level, dose, and study
NR -0.19 design [38]. In Gowda et al.’s study, subjects had a low depression level
(− 0.32, at baseline and not all of the patients were diagnosed as clinically
− 0.06)
depressed. Whereas, vitamin D is considered beneficial for depressed
-0.58
(− 0.71, individuals rather than healthy ones. In fact, individual’s baseline
− 0.44) vitamin D level was not considered in the majority of studies. Hence,
Meta-analysis of cohort studies (OR) according to findings, it is possible that individuals who have low serum
Overall 5 1.60 (1.08, < 0.001 91.3 < 0.001 25(OH)D level are expected to show greater benefit of vitamin D sup­
2.36)
plementation on depressive symptoms. Longitudinal studies have indi­
Sex
cated that low vitamin D level was associated with developing
3 1.55 (1.13, 0.007 69.7 0.037 depression in the future. Also, Vitamin D deficiency is common among
Female 2 2.13) < 0.001 97.3 < 0.001 the elderly, adolescents, obese individuals, people who are homebound
Both 1.19 (1.12,
and have limited sun exposure, and those with chronic illness. These
1.26)
Age (years)
individuals are already at high risk for developing depression. Multiple
3 2.02 (1.71, < 0.001 51.8 0.126 cross-sectional studies had unrepresentative samples, used self-reports
≤ 50 2 2.39) < 0.001 0.0 0.973 of depression, and had small sample sizes [13]. Furthermore, hetero­
> 50 1.12 (1.05, geneity attributed to analytic strategies and participant characteristics,
1.19)
various diagnostic criteria for depression and different methods for
IU: international unit; RCT; randomized control trial, NR; not reported; OR: odds measuring 25(OH)D were also mentioned as sources of bias in
ratio; SMD: standardized mean difference. cross-sectional studies [39].
In RCT studies, variability in the dose of vitamin D administered is an
intervention duration of ≤ 20-weeks exhibited stronger effects in important factor affecting the results of studies. Few studies used high
lowering symptoms of depression. Moreover, the inverse association doses of vitamin D (more than the tolerable upper-level intake) and few
between lower serum vitamin D levels and depression was stronger studies administered lower doses which was not sufficient enough to
among participants aged ≤ 50 years. achieve desirable results. It is evident that depression develops gradu­
Recent studies have shown a significant association between vitamin ally, continues for several years, and symptoms change over time. Thus,
D insufficiency or deficiency and depressive disorders [32]. Receptors of in order to observe changes in symptoms of depression, organised lon­
vitamin D and 1-alphahydroxylase enzymes, involved in the hydroxyl­ gitudinal studies are recommended. The small number of studies with
ation of 25-hydroxy vitamin D (25OHD) to the active form 1,25-dihy­ different study designs, substantial heterogeneity, and uncertain allo­
droxy vitamin D, are present on neurons and glia in multiple regions cation concealment, along with limitations in blinding were other
of the brain, including prefrontal cortex, substantia nigra, cingulate important factors for the high risk of bias observed in several RCTs [13,
cortex and hippocampus and hypothalamus which have an important 32,39] In addition, other factors such as administrating various doses of
role in the pathophysiology of depression [13,30,34,39]. Vitamin D is vitamin D for participants with and without clinical depression symp­
involved in the synthesis of neurotrophic factors and neurotransmitters toms were also mentioned for the inconsistent results observed in pre­
(serotonin, dopamine, adrenalin, and noradrenaline) through VDRs in vious studies. In a few studies, vitamin D supplementation was
the adrenal cortex and due to its steroidal structure, modulates the administered alongside fluoxetine therapy (known as a selective sero­
hypothalamic-pituitary-adrenal axis and GABA-A receptors activity [30, tonin reuptake inhibitor). The different sources of vitamin D supple­
34]. mentation, either administered via intramuscular injection, capsule, or
During depression, inflammatory markers increase [34]. Meanwhile, food were another source of heterogeneity [31,37].
vitamin D displays antioxidant effects in the central nervous system, Two of the meta-analyses of cohort studies [40,41] summarised the
enhances nerve growth factors and the gene expression of antioxidant association between vitamin D deficiency and antepartum and post­
agents, down-regulates cytokines and inflammatory mediators such as partum depression. Several physiological and methodological factors
nuclear factor-kB, which is linked to psychosocial stress and depression have been mentioned for the inconsistent findings. During pregnancy,

5
V. Musazadeh et al. Pharmacological Research 187 (2023) 106605

Fig. 3. Forest plot with mean difference and 95 % confidence intervals (CIs), the relationship between vitamin D and depression symptoms according to meta-
analyses of cohort studies.

Fig. 4. Forest plot with mean difference and 95 % confidence intervals (CIs), the relationship between vitamin D and depression symptoms according to meta-
analyses of cross-sectional studies.

vitamin D-binding proteins enhance and affect the concentration of depression. Vitamin D decreases the production of pro-inflammatory
measured vitamin D. Also, due to physiological adaptations, maternal cytokines. In addition, the sudden drop in oestrogen level after de­
25-(OH) D concentration changes in order to supply the foetus with livery reduces maternal calcium deposits and influence the
appropriate amounts of calcium for bone mineralisation. The short gonadotropin-releasing hormone (GnRH) through the HPA axis. GnRH
duration of the intervention and the small number of participants, pre­ plays a significant role in the physiological regulation of neuronal ac­
vious history of depression, variations in the origin and period of tivity and fertility cycle and decreases oestrogen levels [41]. However,
depression, different cut-offs and methods for measuring vitamin D, and due to limited outdoor activities and exposure to sunlight, less nutritious
moderate-class quality of the included studies were mentioned as food consumption, and less physical activity engagement, pregnant
methodological biases. Moreover, most observational studies used a women are prone to vitamin D deficiency. Thus, in cross-sectional
scaled cut-off instead of a clinical depression diagnosis and did not studies, the reverse causality in which patients who have less exposure
adjust for covariates such as life stress, social support and exercise. to sun end up having lower serum vitamin D levels is not ruled out [13,
Additionally, in a meta-analysis by Tan et al., 2021, the therapeutic 40].
effects of both 25(OH)D2 and 25(OH)D3 were not analysed separately Our subgroup analyses for meta-analyses of RCTs indicated that
[40]. Due to the vascular changes during pregnancy, the maternal ce­ when studies administered vitamin D in dosage of greater than 5000 IU/
rebral environment is sensitive to inflammation. Hence, the inflamma­ day for a duration of ≤ 20-weeks, stronger results were obtained.
tory nature of depression and anti-inflammatory and Findings indicated that studies which used less than the recommended
immunomodulatory features of vitamin D bring a connection between tolerable upper-level intake of vitamin D (< 4000 IU/d) didn’t observe
vitamin D deficiency and depression. The mechanism of action is mainly desirable effects. One study claimed that lower doses of vitamin D were
related to the hypothalamic–pituitary–adrenal (HPA) axis, the levels of not sufficient to cause any change in the occurrence and symptoms of
estradiol, and pro-inflammatory cytokines involved in postpartum depression [32]. When vitamin D is administered at insufficient doses,

6
V. Musazadeh et al. Pharmacological Research 187 (2023) 106605

Fig. 5. The mechanism of action of vitamin D in preventing and declining symptoms of depression.

the upregulation of the level of 25OHD from deficient to sufficient is not declined capacity for hydroxylation to produce adequate amounts of
expected. However, higher doses of vitamin D are beneficial when calcitriol, older adults are at major risk of developing vitamin D defi­
administered for patients with vitamin D deficiency (< 50 nmol/L ciency [39]. Ju et al., reported significant differences in the association
serum levels of vitamin D at baseline) [30]. Another study indicated that between 25(OH)D levels and depression. In fact, they indicated a sig­
vitamin D supplementation had a large effect in major depressive dis­ nificant 4 % reduction in depression risk in individuals less than 60
orders; whereas, vitamin D did not affect emotions in healthy subjects. years and a 10 % in elderly greater than 60 years. Despite the comor­
Several studies claimed that the positive effects of vitamin D on emo­ bidities in elderly, due to the effectiveness of vitamin D supplementation
tions were expressed when administered for more than eight weeks [29, among depressed patients with vitamin D deficiency and the higher
32,34]. Depression develops gradually, continues for several years, and doses administered to this age group, vitamin D supplementation is
symptoms change over time; supplementation must be applied for at rather beneficial for this age group. However, because of the limited
least eight weeks and according to sub-group analyses, superior effects number of studies conducted on this age group, findings should be
are observed when administered for ≤ 20-weeks, because compliance is interpreted with caution [39]. In this umbrella meta-analysis, only the
weak in long duration interventions. Moreover, vitamin D is known as a subgroup of original studies that included participants not on antide­
secosteroid hormone and steroid-like elements act by transcription in pressants were included. Comparing the efficiency of vitamin D sup­
the nucleus; Hence, it requires several weeks to take effect [32]. Cheng plementation among females or both genders did not indicate any major
et al., investigated the effect of using antidepressants along with vitamin differences because significant effects were observed in both categories.
D administration on depression. In this regard, they separated studies This may explain the benefits of vitamin D supplementation in both
into three categories: taking antidepressants, not taking antidepressants, genders. Based on sub-group analyses, cohort studies expressed an in­
not mentioning taking antidepressants. The significant effect on nega­ verse relationship between vitamin D deficiency and depression better
tive emotion was only observed in the last situation. However, it must be than cross-sectional studies. In cross-sectional studies, the biases caused
noted that the number of studies mentioning the last item was higher. by reverse causality (eg, less outdoor activity/nutrient intake, and thus
Thus, for studies not reporting using antidepressants or not, the effect of low Vit D) were not ruled out. Moreover, cross-sectional studies had
vitamin D was significant [34]. small sample sizes with misleading samples, unadjusted data (life stress,
Subgroup analyses for meta-analyses of cohort studies indicated that social support, exercise), and self-reported depression. In contrast,
the association between lower serum vitamin D levels and depression cohort studies are methodologically of better quality compared to
was stronger in participants ≤ 50 years. One study mentioned that cross-sectional studies [13].
vitamin D supplementation was not effective in individuals aged more The limitations of the present study are the various ranges of study
than 65 years; however, subjects aged 18–65 years achieved more populations with different characteristics such as maternal depression
benefits [34]. Jeremiah et al., also confirmed more beneficial effects of along with other types of depression (moderate and/or severe depres­
vitamin D supplementation for patients younger than 50 years than for sion). Moreover, environmental factors such as sunlight, altitude, or diet
those older than 50 years. This might be due to the fact that older in­ on serum 25(OH)D status were not considered in the included meta-
dividuals reveal chronic courses of depression and respond less to an­ analyses. Also, individuals baseline serum vitamin D level was not
tidepressant therapies in comparison to younger ones. Depression has a measured and reported in all of the trials, as some participants had low
bidirectional relationship with insulin resistance; therefore, impairment serum vitamin D level which could affect the symptoms of depression
in β-cell function and adaptation to insulin resistance caused by aging and the treatment with vitamin D supplementation. Despite these limi­
might limit the effectiveness of vitamin D supplementation in elderly tations, several strengths could be attributed to the present umbrella
depressed subjects [35]. Moreover, due to diminished dietary intake, meta-analysis. The inclusion of multiple high and/or moderate quality
limited sun exposure, restriction of outdoor activities, and kidneys’ observational studies and RCTs according to the AMSTAR2

7
V. Musazadeh et al. Pharmacological Research 187 (2023) 106605

questionnaire is the most notable strength of this study. Another [10] I. Kirsch, B.J. Deacon, T.B. Huedo-Medina, A. Scoboria, T.J. Moore, B.T. Johnson,
Initial severity and antidepressant benefits: a meta-analysis of data submitted to
strength was performing sub-group analyses for meta-analyses of RCTs
the Food and Drug Administration, PLoS Med. 5 (2) (2008), e45.
and meta-analyses of cohort studies. [11] M. Fava, Diagnosis and definition of treatment-resistant depression, Biol.
Psychiatry 53 (8) (2003) 649–659.
[12] R. Strawbridge, A.H. Young, A.J. Cleare, Biomarkers for depression: recent
5. Conclusion
insights, current challenges and future prospects, Neuropsychiatr. Dis. Treat. 13
(2017) 1245–1262.
The present umbrella meta-analysis confirms the potential benefits [13] R.E. Anglin, Z. Samaan, S.D. Walter, S.D. McDonald, Vitamin D deficiency and
of vitamin D supplementation in reducing symptoms of depression and depression in adults: systematic review and meta-analysis, Br. J. Psychiatry 202 (2)
(2013) 100–107.
an inverse relationship between higher serum levels of vitamin D and [14] D.F. de Abreu, D. Eyles, F. Feron, Vitamin D, a neuro-immunomodulator:
overall depression. Vitamin D supplementation in studies using dosage implications for neurodegenerative and autoimmune diseases,
of > 5000 IU/day and intervention duration of ≤ 20-weeks exhibited Psychoneuroendocrinology 34 (2009) S265–S277.
[15] D.W. Eyles, S. Smith, R. Kinobe, M. Hewison, J.J. McGrath, Distribution of the
better effects in lowering depression symptoms. Moreover, a greater risk vitamin D receptor and 1α-hydroxylase in human brain, J. Chem. Neuroanat. 29 (1)
of depression was shown among participants aged ≤ 50 with lower (2005) 21–30.
serum vitamin D levels. [16] G. Ceolin, G.P.R. Mano, N.S. Hames, L.d.C. Antunes, E. Brietzke, D.K. Rieger, J.
D. Moreira, Vitamin D, depressive symptoms, and Covid-19 pandemic, Front.
Neurosci. 15 (2021), 670879.
Funding [17] G.A. Casseb, M.P. Kaster, A.L.S. Rodrigues, Potential role of vitamin D for the
management of depression and anxiety, CNS Drugs 33 (7) (2019) 619–637.
[18] C.H. Wilkins, Y.I. Sheline, C.M. Roe, S.J. Birge, J.C. Morris, Vitamin D deficiency is
None.
associated with low mood and worse cognitive performance in older adults, Am. J.
Geriatr. Psychiatry 14 (12) (2006) 1032–1040.
Declaration of Interest [19] H.T. May, T.L. Bair, D.L. Lappé, J.L. Anderson, B.D. Horne, J.F. Carlquist, J.
B. Muhlestein, Association of vitamin D levels with incident depression among a
general cardiovascular population, Am. Heart J. 159 (6) (2010) 1037–1043.
The authors declare that there is no conflict of interests. [20] G.B. Parker, H. Brotchie, R.K. Graham, Vitamin D and depression, J. Affect. Disord.
208 (2017) 56–61.
[21] J.F. Gould, A.J. Anderson, L.N. Yelland, L.G. Smithers, C.M. Skeaff, R.A. Gibson,
Data Availability
M. Makrides, Association of cord blood vitamin D at delivery with postpartum
depression in Australian women, Aust. N. Z. J. Obstet. Gynaecol. 55 (5) (2015)
Data will be made available on request. 446–452.
[22] R. Chan, D. Chan, J. Woo, C. Ohlsson, D. Mellström, T. Kwok, P. Leung, Association
between serum 25-hydroxyvitamin D and psychological health in older Chinese
Acknowledgements men in a cohort study, J. Affect. Disord. 130 (1–2) (2011) 251–259.
[23] A. Pan, L. Lu, O.H. Franco, Z. Yu, H. Li, X. Lin, Association between depressive
The research protocol was approved and Supported by Student symptoms and 25-hydroxyvitamin D in middle-aged and elderly Chinese, J. Affect.
Disord. 118 (1–3) (2009) 240–243.
Research Committee, Tabriz University of Medical Sciences (Registra­ [24] N.O. Nielsen, M. Strøm, H.A. Boyd, E.W. Andersen, J. Wohlfahrt, M. Lundqvist,
tion code: 70169). A. Cohen, D.M. Hougaard, M. Melbye, Vitamin D status during pregnancy and the
risk of subsequent postpartum depression: a case-control study, PLoS One 8 (11)
(2013), e80686.
Appendix A. Supporting information [25] B.J. Shea, B.C. Reeves, G. Wells, M. Thuku, C. Hamel, J. Moran, D. Moher,
P. Tugwell, V. Welch, E. Kristjansson, AMSTAR 2: a critical appraisal tool for
Supplementary data associated with this article can be found in the systematic reviews that include randomised or non-randomised studies of
healthcare interventions, or both, bmj 358 (2017).
online version at doi:10.1016/j.phrs.2022.106605. [26] R. DerSimonian, N. Laird, Meta-analysis in clinical trials, Control. Clin. Trials 7 (3)
(1986) 177–188.
References [27] M. Egger, G.D. Smith, M. Schneider, C. Minder, Bias in meta-analysis detected by a
simple, graphical test, Bmj 315 (7109) (1997) 629–634.
[28] C.B. Begg, M. Mazumdar, Operating characteristics of a rank correlation test for
[1] N.J. Kassebaum, M. Arora, R.M. Barber, Z.A. Bhutta, J. Brown, A. Carter, D.
publication bias, Biometrics (1994) 1088–1101.
C. Casey, F.J. Charlson, M.M. Coates, M. Coggeshall, Global, regional, and national
[29] H. Jamilian, E. Amirani, A. Milajerdi, F. Kolahdooz, H. Mirzaei, M. Zaroudi,
disability-adjusted life-years (DALYs) for 315 diseases and injuries and healthy life
A. Ghaderi, Z. Asemi, The effects of vitamin D supplementation on mental health,
expectancy (HALE), 1990–2015: a systematic analysis for the Global Burden of
and biomarkers of inflammation and oxidative stress in patients with psychiatric
Disease Study 2015, Lancet 388 (10053) (2016) 1603–1658.
disorders: a systematic review and meta-analysis of randomized controlled trials,
[2] V. Musazadeh, M. Zarezadeh, A.H. Faghfouri, M. Keramati, P. Jamilian,
Prog. Neuro-Psychopharmacol. Biol. Psychiatry 94 (2019), 109651.
P. Jamilian, A. Mohagheghi, A. Farnam, Probiotics as an effective therapeutic
[30] S. Spedding, Vitamin D and depression: a systematic review and meta-analysis
approach in alleviating depression symptoms: an umbrella meta-analysis, Crit. Rev.
comparing studies with and without biological flaws, Nutrients 6 (4) (2014)
Food Sci. Nutr. (2022) 1–9.
1501–1518.
[3] S.L. James, D. Abate, K.H. Abate, S.M. Abay, C. Abbafati, N. Abbasi, H. Abbastabar,
[31] J.A. Shaffer, D. Edmondson, L.T. Wasson, L. Falzon, K. Homma, N. Ezeokoli, P. Li,
F. Abd-Allah, J. Abdela, A. Abdelalim, Global, regional, and national incidence,
K.W. Davidson, Vitamin D supplementation for depressive symptoms: a systematic
prevalence, and years lived with disability for 354 diseases and injuries for 195
review and meta-analysis of randomized controlled trials, Psychosom. Med. 76 (3)
countries and territories, 1990–2017: a systematic analysis for the Global Burden of
(2014) 190.
Disease Study 2017, Lancet 392 (10159) (2018) 1789–1858.
[32] U. Gowda, M.P. Mutowo, B.J. Smith, A.E. Wluka, A.M. Renzaho, Vitamin D
[4] C.K. Ettman, S.M. Abdalla, G.H. Cohen, L. Sampson, P.M. Vivier, S. Galea,
supplementation to reduce depression in adults: meta-analysis of randomized
Prevalence of depression symptoms in US adults before and during the COVID-19
controlled trials, Nutrition 31 (3) (2015) 421–429.
pandemic, JAMA Netw. Open 3 (9) (2020) (e2019686-e2019686).
[33] J. Firth, S.B. Teasdale, K. Allott, D. Siskind, W. Marx, J. Cotter, N. Veronese,
[5] J. Xiong, O. Lipsitz, F. Nasri, L.M. Lui, H. Gill, L. Phan, D. Chen-Li, M. Iacobucci,
F. Schuch, L. Smith, M. Solmi, The efficacy and safety of nutrient supplements in
R. Ho, A. Majeed, Impact of COVID-19 pandemic on mental health in the general
the treatment of mental disorders: a meta-review of meta-analyses of randomized
population: a systematic review, J. Affect. Disord. 277 (2020) 55–64.
controlled trials, World Psychiatry 18 (3) (2019) 308–324.
[6] J.L. Kim, J. Cho, S. Park, E.-C. Park, Depression symptom and professional mental
[34] Y.C. Cheng, Y.C. Huang, W.L. Huang, The effect of vitamin D supplement on
health service use, BMC Psychiatry 15 (1) (2015) 1–12.
negative emotions: a systematic review and meta-analysis, Depress. Anxiety 37 (6)
[7] S. Gawlik, L. Waldeier, M. Müller, A. Szabo, C. Sohn, C. Reck, Subclinical
(2020) 549–564.
depressive symptoms during pregnancy and birth outcome—a pilot study in a
[35] O.J. Jeremiah, G. Cousins, F. Boland, B.P. Kirby, B.K. Ryan, Evaluation of the effect
healthy German sample, Arch. Women’s Ment. Health 16 (2) (2013) 93–100.
of insulin sensitivity-enhancing lifestyle-and dietary-related adjuncts on
[8] J.C. Jakobsen, K.K. Katakam, A. Schou, S.G. Hellmuth, S.E. Stallknecht, K. Leth-
antidepressant treatment response: a systematic review and meta-analysis, Heliyon
Møller, M. Iversen, M.B. Banke, I.J. Petersen, S.L. Klingenberg, J. Krogh, S.E. Ebert,
6 (9) (2020), e04845.
A. Timm, J. Lindschou, C. Gluud, Selective serotonin reuptake inhibitors versus
[36] F. Vellekkatt, V. Menon, Efficacy of vitamin D supplementation in major
placebo in patients with major depressive disorder. A systematic review with meta-
depression: a meta-analysis of randomized controlled trials, J. Postgrad. Med. 65
analysis and Trial Sequential Analysis, BMC Psychiatry 17 (1) (2017) 58.
(2) (2019) 74.
[9] N.A. Khin, Y.F. Chen, Y. Yang, P. Yang, T.P. Laughren, Exploratory analyses of
[37] G. Li, L. Mbuagbaw, Z. Samaan, M. Falavigna, S. Zhang, J.D. Adachi, J. Cheng,
efficacy data from major depressive disorder trials submitted to the US Food and
A. Papaioannou, L. Thabane, Efficacy of vitamin D supplementation in depression
Drug Administration in support of new drug applications, J. Clin. Psychiatry 72 (4)
in adults: a systematic review, J. Clin. Endocrinol. Metab. 99 (3) (2014) 757–767.
(2011) 464–472.

8
V. Musazadeh et al. Pharmacological Research 187 (2023) 106605

[38] C.C. Nicoláse, M. Hernández-Viadele, Efficacy of vitamin D in the treatment of [41] J. Wang, N. Liu, W. Sun, D. Chen, J. Zhao, W. Zhang, Association between vitamin
depression: a systematic review and meta-analysis, Actas Esp. Psiquiatr 49 (1) D deficiency and antepartum and postpartum depression: a systematic review and
(2021) 12–23. meta-analysis of longitudinal studies, Arch. Gynecol. Obstet. 298 (6) (2018)
[39] S.-Y. Ju, Y.-J. Lee, S.-N. Jeong, Serum 25-hydroxyvitamin D levels and the risk of 1045–1059.
depression: a systematic review and meta-analysis, J. Nutr. Health Aging 17 (5) [42] M.J. Berridge, Vitamin D and depression: cellular and regulatory mechanisms,
(2013) 447–455. Pharm. Rev. 69 (2) (2017) 80–92.
[40] Q. Tan, S. Liu, D. Chen, Poor vitamin D status and the risk of maternal depression: a
dose–response meta-analysis of observational studies, Public Health Nutr. 24 (8)
(2021) 2161–2170.