Vitamin D and depression in older adults: lessons learned from observational and clinical studies

Abstract

Depression is a mental disorder triggered by the interaction of social, psychological and biological factors that have an important impact on an individual’s life. Despite being a well-studied disease with several established forms of treatment, its prevalence is increasing, especially among older adults. New forms of treatment and prevention are encouraged, and some researchers have been discussing the effects of vitamin D (VitD) on depression; however, the exact mechanism by which VitD exerts its effects is not yet conclusive. In this study, we aimed to discuss the possible mechanisms underlying the association between VitD and depression in older adults. Therefore, we conducted a systematic search of databases for indexed articles published until 30 April 2021. The primary focus was on both observational studies documenting the association between VitD and depression/depressive symptoms, and clinical trials documenting the effects of VitD supplementation on depression/depressive symptoms, especially in older adults. Based on pre-clinical, clinical and observational studies, it is suggested that the maintenance of adequate VitD concentrations is an important issue, especially in older adults, which are a risk population for both VitD deficiency and depression. Nevertheless, it is necessary to carry out more studies using longitudinal approaches in low- and middle-income countries to develop a strong source of evidence to formulate guidelines and interventions.

Introduction

Depression is a mental disorder that causes clinically significant suffering and/or impairment in social, professional, economic and other important areas of an individual’s life, and is the main cause of suicide in more severe cases^(1–4). The prevalence of depression in 2015 was estimated to be 4·4 % globally, with a higher prevalence among those between 55 and 77 years of age. Women appear to be more affected (7·5 %) than men (5·5 %)⁽⁵⁾. Among those over 60 years old, depression occurs in 7·0 % of the general older population⁽⁶⁾. According to the Global Burden of Disease, Injuries, and Risk Factors – GBD survey, depression is among the top three causes of disability⁽⁷⁾. There has been a significant increase in the global burden of disease in years lived with disabilities (YLDs) in the past 20 years due to depressive disorders. In 1990, depression occupied the fourth position, moving to the third in 2007 with an increase of 33·4 %, and remained in the third position between 2007 and 2017; however, it has increased by 14·3 %⁽⁸⁾. Moreover, as many people with depressive symptoms are undiagnosed, the prevalence of depressive disorders is probably higher than reported⁽⁴⁾.

Mental disorders are among the main problems in public health, and mood disorders are diseases with higher costs to health systems worldwide^(2,3,9,10) . According to the Mental Health Atlas of the World Health Organization (WHO), low- and middle-income countries spend less than $1 per year per capita in the treatment and prevention of mental disorders, compared with an average of >$80 in high-income countries, owing to socioeconomic issues⁽¹¹⁾. Therefore, there is an urgent need to identify the modifiable risk factors associated with the aetiology of depression, helping with the treatment and prevention of this disorder, especially in low- and middle-income countries^(12,13) .

Depression is a complex disease triggered by the interaction between social, psychological and biological factors^(4,14) . In older adults, depression can be triggered by a series of factors such as limitations in daily activities, cognition, mobility and social changes such as retirement, social isolation and relocation to long-term institutions⁽¹⁵⁾. Among the biological factors, genetic predisposition, neurotransmitter and neuroendocrine system imbalance, functional and structural brain anatomy, and cognition are the most studied mechanisms^(16,17) . Recently, nutritional factors have shown an important relationship with the evolution, prevention and treatment of mental disorders⁽¹⁸⁾. The association between VitD and depression has emerged in scientific scenarios, and this nutrient seems to be relevant in the prevention of depressive symptom development. However, the mechanism by which VitD exerts its effects remains unclear^(19,20) .

Many clinical trials have been conducted to investigate the potential therapeutic effect of VitD on patients with depression, but the results remain inconclusive due to methodological issues⁽²¹⁾. VitD is a fat-soluble vitamin that is present in two forms: VitD2 (ergosterol) and VitD3 (cholecalciferol). It is obtained from diet, supplementation and sun exposure^(22,23) . VitD has a well-established role in mineral bone metabolism, but its effects are not restricted to bone health and are also important in maintaining many biological processes, such as the regulation of gene expression, cell proliferation and differentiation, and immune system regulation^(24–26). In the central nervous system (CNS), the presence of nuclear (vitamin D receptor, VDR) and membrane (protein disulphide isomerase family A member 3, PDIA3) receptors for VitD and some enzymes (cytochrome P450 family enzymes CYP27a1, CYP27b1 and CYP24a1) responsible for converting its active form has raised the hypothesis that VitD may be involved in the pathophysiology of depression^(27–31).

Low serum VitD concentrations [25-hydroxycholecalciferol, 25(OH)D] have been considered a public health problem worldwide, especially in the elderly⁽³²⁾. For older adults, the prevalence of 25(OH)D deficiency (<50 nmol/l or <20 ng/ml) was 36 % in the United States⁽³³⁾, 19 % in Canada⁽³⁴⁾, 36 % in China(35 and 4–89 % in European countries⁽³²⁾. In low- and middle-income countries, the prevalence was approximately 41 % for older adults in Brazil⁽³⁶⁾, 91 % in India⁽³⁷⁾ and 46 % in Guatemala⁽³⁸⁾. However, different cut-off points have been suggested, and a single value to define VitD deficiency or insufficiency has been debated⁽³⁹⁾. Moreover, the establishment of desirable serum VitD concentrations is based on bone health to maintain mineral and skeletal homoeostasis^(39,40) .

It is important to mention that VitD levels via skin synthesis and intestinal absorption are influenced by various factors such as skin pigmentation, latitude, season, age, obesity and inflammatory bowel diseases, among others^(41–43). Due to reduced sun exposure, decreased skin synthesis and dietary intake, and intestinal malabsorption, the elderly are among the top risk groups for VitD deficiency^(41,44) . They also present significant complications related to low VitD concentrations (<20 ng/ml), such as the risk of fractures due to fragility and bone loss, which contribute to age-related muscle weakness and sarcopenia^{(28,43,45,46)} . In addition, VitD concentrations <20 ng/ml have been associated with an increased risk of all-cause mortality⁽⁴⁷⁾.

In this review, we aimed to update the role of VitD in depression, discussing the metabolism of VitD, its mechanism of action in the brain and the main evidence of pre-clinical, clinical and observational studies, especially those involving older adults, a population risk for both conditions, in an attempt to highlight the potential preventive and therapeutic effects of this nutrient. Also, we aimed to suggest future directions for new studies. To this end, we conducted a systematic search for articles published until 30 April 30 2021. The databases used were PubMed, Scopus, Embase, Science Direct and Web of Science (details are presented in the supplementary material).

Vitamin D: synthesis and metabolism

The synthesis of VitD (Fig. 1) by epidermal epithelial cells begins when the exposure to ultraviolet B radiation (UVB, 290–315 nm) promotes the non-enzymatic transformation of 7-dehydrocholesterol (7-DHC or pro-VitD) in pre-VitD3^(48,49) . A photolytic break forms a secosteroid molecule, which then undergoes an isomerisation reaction induced by heat to transform it into VitD3 (or cholecalciferol), a process that takes about 8 h^(48–50). Keratinocytes are the main cells of the epidermis that have the enzymatic machinery to metabolise VitD in its active form and express the vitamin D receptor (VDR)^(22,51) . In contrast, the synthesis of the active form of VitD from either food or supplementation begins with incorporation into micelles and absorption through the enterocyte membrane by apical membrane transporters or by passive diffusion⁽⁵²⁾. A fraction of VitD is incorporated into the chylomicrons, which are transported to the lymphatic system and then to the venous system by vitamin D binding protein (DBP)⁽⁵⁰⁾. The other fraction is incorporated into adipose tissue and skeletal muscles⁽⁵³⁾.

Fig. 1. Vitamin D synthesis, metabolism and target tissue actions. ⁽¹⁾ The synthesis of VitD from sunlight initiates in the skin when 7-DHC is converted in pre-VitD3 and then VitD3 [25(OH)D3 or cholecalciferol] and is carried by DBP through blood circulation. ⁽²⁾ The VitD from dietary intake (VitD2/ergocalciferol and D3/cholecalciferol) is absorbed in the small intestine and packed into chylomicrons to reach the systemic circulation. Both VitD3 and VitD2 are also transported through blood circulation by DBP to the liver, where they are converted to 25-hydroxyvitamin D [calcidiol or 25(OH)D] by the action of 25-hydroxylases. (3) 25(OH)D coupled to DBP is transported to the target organs such as kidney, bones, adipose tissue, muscle and brain, and cells such as in the immune system containing the enzyme 1-α-hydroxylase, which convert 25(OH)D to 1,25-dihydroxyvitamin D [calcitriol or 1,25(OH)₂D3], the active form of VitD. VitD act through both genomic and non-genomic pathways. In the genomic pathway, VitD active form enters the nucleus linked to the VDR where it binds to the RXR and then binds to the VDRE, resulting in modulation of target gene expression. In the non-genomic pathway, the VitD active form binds to the PDIA3 and starts signalling cascades, including the activation of phospholipase A2 activating protein (PLAA), phospholipase A2 (PLA2), phospholipase C (PLC) and opening Ca²⁺ channels that results in the activation of secondary messengers. This figure was made using BioRender (license: YN235V4QZA)

Both VitD2 and VitD3 are transported in the blood by DBP and must undergo activation through two consecutive enzymatic hydroxylation reactions in the liver and kidneys. In the liver, VitD2 and VitD3 are converted into 25-hydroxyvitamin D (calcidiol or 25(OH)D) by the action of 25-hydroxylases (cytochrome P450 enzymes group, CYP2R1 or CYP27A1)^(54–56). The 25(OH)D coupled with DBP is transported to various tissues with cells containing the enzyme 1-α-hydroxylase (CYP27B1), as in the kidney, where it converts 25(OH)D to 1,25-dihydroxyvitamin D (calcitriol or 1,25(OH)₂D3), the active form of VitD^(54–56).

The conversion of 1,25(OH)₂D3 in the kidney is regulated by several factors, including circulating concentrations of parathyroid hormone (PTH) in the parathyroid glands, serum phosphorus, calcium, fibroblast growth factor 23 (FGF-23) in the bone and its self-regulation. 1,25(OH)₂D3 decreases its own synthesis by negative feedback; it decreases the secretion of parathyroid hormone and increases the expression of 24-hydroxylase⁽⁵⁷⁾. This self-regulation by the expression of 24-hydroxylase is found in most tissues and is essential for the catabolism of 25(OH)D and 1,25(OH)₂D3⁽⁵⁸⁾.

The biological effects of 1,25(OH)₂D3 are largely mediated by VDR, which is expressed in almost all human cells^(59,60) . The VDR belongs to a subfamily of nuclear receptors, which contains two sites for ligand binding called the genomic pocket (VDR-GP), which binds in a bowl-like configuration for gene transcription, and the alternative pocket (VDR-AP), which connects in a planar-like configuration for quick responses⁽⁶⁰⁾. When VDR-GP binds to 1,25(OH)₂D3, it enters the cell nucleus and binds to the retinoid X receptor (RXR). This complex then binds to the vitamin D responsive element (VDRE) in the promoter regions of the target genes by recruiting co-activator or co-repressor complexes that regulate the transcription of genes either positively or negatively^(53,60) . The other suggested VitD receptor is PDIA3, also known as endoplasmic reticulum protein (ERp60, ERp57 and Grp58) or VitD membrane-associated rapid-response steroid-binding protein (1,25-MARRS)⁽⁶¹⁾. PDIA3 is present in caveolae (lipid rafts) and is linked to the rapid responses of 1,25(OH)₂D3 by activating signalling cascades, where it physically interacts with downstream mediators^(61,62) , including the activation of phospholipase A2 activating protein (PLAA), phospholipase A2 (PLA2), phospholipase C (PLC) and opening Ca²⁺ channels that result in the activation of secondary messengers⁽⁶³⁾. PDIA3 is involved in the function of immune and musculoskeletal systems as well as mammary gland growth and development, and participates in the intestinal uptake of calcium and phosphate⁽⁶³⁾. PDIA3 also mediates the effect of 1,25(OH)₂D3 on the regulation of osteoblasts and chondrocytes⁽⁶⁴⁾.

Vitamin D: mechanism of action in the brain

The first evidence of the role of VitD in brain function began with autoradiographic findings of the presence of VDR in the brain tissue of laboratory animals⁽⁶⁵⁾. VDR is found in neurons and glial cells in most regions of the brain, including the cortex (temporal, frontal, parietal and cingulate); deep grey matter (thalamus, basal ganglia, hypothalamus, hippocampus and amygdala); cerebellum, nuclei of the brain stem and substantia nigra (an area abundant in dopaminergic neurons); spinal cord; and ventricular system⁽⁶⁶⁾. In addition, an alternative mechanism was observed in post-mortem human brain tissue samples. It was demonstrated that 1,25(OH)₂D3 can be activated locally through the expression of the enzyme 1α-hydroxylase, which is classically expressed in the kidney and is responsible for catalysing the conversion of 25(OH)D into 1,25(OH)₂D3, showing that both forms (VitD and 25(OH)D) can pass through the blood–brain barrier^(67,68) .

It has been proposed that within the neurovascular unit, the machinery for conversion of both VitD forms involves the cytochrome P450 family enzymes CYP27a1, CYP27b1 and CYP24a1 which are expressed in neurons, and CYP27a1 which is expressed in all neural cell types and is highly expressed in endothelial cells⁽³¹⁾. The active form of VitD triggers genomic actions associated with VDR or non-genomic actions related to PDIA3, which is expressed in small amounts in extra-cerebral tissues such as the liver and kidney. On the other hand, PDIA3 is highly expressed in the brain and appears to be the main brain receptor for VitD in neural tissue (Fig. 2).

Fig. 2. The role of vitamin D in depression. ⁽¹⁾ In the brain, both active and inactive VitD is carried through blood circulation binding to DBP and can permeate the blood–brain barrier. All brain cells (endothelial cells (A), astrocytes (B), neurons (C), oligodendrocytes (D) and microglia (E)) have the machinery to transform VitD. VitD is turned into 25(OH)D by CYP27a1 in endothelial cells and neurons, and it is metabolized to 1,25(OH)₂D3 by CYP27b1 in neurons or microglia. All brain cells can express VDR, but it is highly expressed by astrocytes. When it enters the cell, 1,25(OH)₂D3 can bind to VDR, and then to the RXR in the nucleus. The complex VDR–RXR binds to the VDRE and initiates gene transcription or can be inactivated when in excess by CYP24a1. All brain cells can express PDIA3, but it is highly expressed in endothelial cells where 1,25(OH)₂D3 can bind it, and PDIA3 physically interacts with downstream mediators to initiate rapid responses and induce signalling cascades. (2) VitD regulates the expression of many processes related to depression. It maintains Ca²⁺ homoeostasis, activates the expression of many antioxidant genes, regulates the formation of serotonin and dopamine, and reduces inflammation by reducing the expression of inflammatory cytokines. TPH2, tryptophan hydroxylase 2; SERT, serotonin reuptake transporter; GDNF, glial cell-derived neurotrophic factor; COMT, catechol-O-methyltransferase; NRF2, nuclear factor-erythroid-2-related factor 2; γ-GT, γ-glutamyl transpeptidase; GCLC, glutamate-cysteine ligase; GR, glutathione reductase; GPx, glutathione peroxidase; NF-κB, nuclear factor-kappa B. This figure was made using BioRender (license: FB235V4MBD)

VitD is known as a neurosteroid because of its important role in the CNS in processes related to cell differentiation, production and release of neurotrophic factors, synthesis of neurotransmitters, intracellular calcium homoeostasis, influence on the redox state, function and metabolism of neuronal cells and cognition (Fig. 2)^(29,69) . The active form of VitD stimulates the synthesis of nerve growth factor (NGF) which acts on cholinergic neurons, and positively regulates the synthesis of neurotrophic factors derived from the glial cell line (GDNF), which acts on dopaminergic neurons, and neurotrophin 3 (NT-3), which is key to neuronal promotion, survival, differentiation and plasticity⁽⁶⁶⁾. Due to its involvement in several brain functions, observational studies in humans subjects have linked low serum VitD concentrations with some brain disorders such as schizophrenia, failure in synaptic plasticity related to learning and memory, cognitive decline and mood disorders^(27,29,70) .

Vitamin D and depressive symptoms: evidence from pre-clinical and clinical studies

Pre-clinical studies

Depression is a multifactorial disease, which makes it challenging to identify the precise biological mechanisms that link VitD to depression. However, some hypotheses have been proposed based on the experimental research data. Calcium homoeostasis, glutamatergic/GABAergic and monoaminergic system modulation, influence on circadian rhythm, anti-inflammatory properties and redox balance modulation are among the most investigated mechanisms.

The homoeostasis of intracellular and extracellular calcium (Ca²⁺) is an important factor responsible for driving the onset of depression, which links VitD with the development of depressive symptoms because of its interaction with excitatory synapses⁽²⁷⁾. The imbalance in intracellular Ca²⁺ is caused by an elevation in glutamate and by activation of the phosphoinositide signalling pathway that generates inositol triphosphate (IP3) which releases Ca²⁺ from internal stores^(27,71,72) . The elevation of Ca²⁺ can affect both ionotropic (N-methyl-d-aspartate) and metabotropic (mGluR) receptors⁽⁷³⁾. This change in neural activity drives excitatory neurons and is responsible for the decline in the activity and the number of GABAergic inhibitory neurons, as well as modulation of the activity of other neurotransmitter systems, including the inhibition of the serotonergic system and the release of norepinephrine and dopamine⁽⁷⁴⁾. However, 1,25(OH)₂D can act in this pathway by inducing the expression of proteins related to the maintenance of Ca²⁺ homoeostasis, such as calbindin, parvalbumin, Na⁺/Ca²⁺ exchanger (NCX1) and pump Ca²⁺-ATPase (PMCA). It also regulates Ca²⁺ concentrations by reducing the expression of the CaV1·2 calcium channel^(27,75) .

Concerning other neurotransmitter systems, it has been proposed that depression could result from a deficiency of serotonin (5-HT) in the synaptic cleft^(76–78). 5-HT is derived from the essential amino acid tryptophan. To produce 5-HT in the brain, tryptophan must first be transported across the blood–brain barrier and then metabolised by the enzyme tryptophan hydroxylase 2 (TPH2). VDR activation by 1,25(OH)₂D3 can induce the expression of the TPH2 gene in serotonergic neurons^(79,80) . In addition, 1,25(OH)₂D3 could act in the repression of the serotonin reuptake transporter (SERT or 5-HTT), and the mitochondrial enzyme responsible for 5-HT catabolism, monoamine oxidase-A, resulting in potentiated serotonergic transmission⁽⁸¹⁾.

In the dopaminergic system, VitD is involved in the maturation of dopaminergic neurons. VDR is present in the nucleus of positive neurons for tyrosine hydroxylase (TH), and can stimulate glial cell line-derived neurotrophic factor (GDNF) in dopaminergic neurons⁽⁸²⁾. VDR also modulates metabolism through the genomic regulation of catechol-O-methyl transferase (COMT) expression, a key enzyme involved in dopamine turnover^(82,83) . In addition, in a rat model of depression, VitD appears to produce therapeutic effects comparable to antidepressant drugs such as fluoxetine, improving anhedonia-like symptoms, probably by regulating the effect of dopamine-related actions on the nucleus accumbens⁽⁸⁴⁾.

From a chronobiological perspective, a growing body of evidence suggests that VitD participates in the mechanisms orchestrating the circadian rhythm, suggesting that hypovitaminosis D might play a role in sleep disorders⁽⁸⁵⁾. VitD has been associated with the regulation and maintenance of optimal sleep⁽⁸⁶⁾. The mediating role of VitD in the circadian rhythm is supported by studies demonstrating the association between lower concentrations of VitD and sleep^(87,88) . In addition, a circadian oscillation pattern can be equally observed in plasma 1,25(OH)₂D3 concentration and DBP, which corroborates the association between VitD and the circadian system⁽⁸⁷⁾.

Because sunlight partially regulates the synthesis of VitD and is the main zeitgeber in the regulation of the circadian rhythm, it is conceivable that VitD might contribute to the transduction of signs regulating it^(89,90) . The suprachiasmatic nucleus (SCN) is a hypothalamic structure found directly above the optic chiasm, and its strategic anatomical position allows prompt central response to sunlight stimuli through the retina. SCN is the main oscillator, which accounts for the control of circadian rhythms by regulating several body functions during a 24-h cycle, sending peripheral signals through neurohumoral mechanisms⁽⁹¹⁾. For this reason, the authors postulated that VitD is likely involved in the regulation of the sleep/wake rhythm⁽⁹⁰⁾.

Melatonin is a neurohormone involved in the regulation of mammalian circadian rhythms and sleep. It is released in response to darkness and is synthesised by the pineal gland⁽⁹²⁾. Its synthesis occurs from the metabolism of serotonin⁽⁹³⁾, which, in turn, is also regulated by VitD. Along with VDR, 1,25(OH)₂D triggers the central expression of TPH2, the gene responsible for encoding the enzyme catalysing the conversion of tryptophan into 5-hydroxytryptophan, which is then metabolised into serotonin and subsequently as melatonin^(67,79) . Therefore, it is thought that the combination of deficits in serum VitD levels and circadian rhythm impairments could induce a robust increase in depressive symptoms and/or act as an interplay variable in the pathophysiology of major depressive disorder.

Regarding anti-inflammatory pathways, it is also relevant to point out that both melatonin and VitD mediate the mitochondrial function in homoeostasis, such as down-regulating mechanistic target of rapamycin (mTOR), inducible nitric oxide synthase (iNOS) and nuclear factor kappa B (NF-κB) pathways, and up-regulating Sirtuin-1 (SIRT-1) and adenosine monophosphate-activated protein kinase (AMPK) pathways, which are critical mechanisms to avoid anomalous inflammatory responses related to oxidative stress and apoptosis⁽⁹⁴⁾.

Pro-inflammatory cytokines, interleukins and other inflammatory markers, such as prostaglandins and acute-phase C-reactive protein, have been implicated to play role in the pathophysiology of depression^(95–97). Inflammation leads to increased blood–brain barrier permeability, allowing easier entry of inflammatory molecules into the CNS⁽⁹⁸⁾. At a cellular level, it has been observed that tumour necrosis factor α (TNF-α) can induce glutamate release by activated microglia in vitro, leading to excitotoxic damage to neurons⁽⁹⁹⁾. Some cytokines can directly increase enzymatic activity for converting tryptophan to kynurenine and decreasing the production of serotonin^(100–102). Considering that macrophages, dendritic cells and activated B and T lymphocytes express 1α-hydroxylase and VDR, VitD could act by modulating the immune response and regulating cytokine expression^(97,103) . Moreover, it was demonstrated that the activity of NF-κB, a transcription factor involved in the synthesis of pro-inflammatory cytokines, was inhibited by 1,25(OH)₂D3, which helps to maintain the balance of T-helper (Th) cells, inhibiting the production of Th1 and Th17 cytokines and increasing Th2 cytokine synthesis⁽⁷⁵⁾.

Interestingly, Boontanrart et al. (2016) reported that activated microglia were associated with an increased expression of VitD receptor and Cyp27b1, which encodes the 1α-hydroxylase enzyme for converting 25(OH)D into its active form, thereby enhancing their responsiveness to 25(OH)D. Moreover, activated microglia exposed to 25(OH)D had reduced expression of pro-inflammatory cytokines, interleukin (IL)-6, IL-12 and TNF-α, and increased expression of IL-10. The decrease in pro-inflammatory cytokines was dependent on IL-10 induction of suppressor of cytokine signalling-3 (SOCS3). Therefore, 25(OH)D increases the expression of IL-10, creating a feedback loop via SOCS3 which reduces the pro-inflammatory immune response by activated microglia and probably protects the CNS from damage⁽¹⁰⁴⁾. In agreement with these findings, Lee et al. (2020) showed that VitD signalling in neurons elicits an anti-inflammatory state in microglia. Moreover, the partial deletion of VDR in neurons during early life exacerbates CNS autoimmunity in adult mice. Therefore, by changing the immune response of microglia, VitD may be an interesting mechanism for avoiding a prolonged inflammatory state in the CNS⁽¹⁰⁵⁾.

In addition, VDR activation stimulates the expression of many antioxidant genes, such as the nuclear factor erythroid-2 (NRF2), γ-glutamyl transpeptidase (γ-GT), glutamate-cysteine ligase (GCLC), glutathione reductase (GR) and glutathione peroxidase (GPx)⁽²⁷⁾. VitD negatively regulates the expression of iNOS in monocyte-derived cells, and increases the activity of γ-GT, an important enzyme in the glutathione pathway^(106,107) . Reinforcing the modulation of oxidative stress as a mechanism associated with the antidepressant-like effect of VitD, repeated administration of this compound (2·5, 7·5 and 25 µg/kg for 7 d) prevented depressive-like behavior and brain oxidative stress induced by chronic administration of corticosterone (21 d) in male and female mice^(108,109) . It has been demonstrated that reactive oxygen species (ROS) trigger a variety of molecular cascades that increase the permeability of the blood–brain barrier, allowing inflammatory cytokines to enter the CNS⁽¹¹⁰⁾. Moreover, it has been well established that inflammation and oxidative stress, which mutually amplify each other, play an important role in the pathophysiology of depression and can be a target for therapeutic strategies⁽¹¹¹⁾.

Clinical studies

Nineteen randomised clinical trials using VitD supplementation for depressive symptoms in adults were published up to 2020 (Table 1). Nine studies were double-blinded, and twelve included individuals aged >65 years. Most of the studies were conducted in high-income countries (13/19). Seven studies were conducted with community-dwelling, healthy volunteers or individuals with no specification, and three studies only with VitD-deficient individuals^(112–114). Six included only individuals with the diagnosis of depression, and two with individuals with VitD deficiency and diagnosed depression^(115,116) . Considering only the studies that included individuals with a diagnosis of depression (with or without VitD deficiency), 4/8 presented improvement in depressive symptoms after VitD supplementation.

Table 1 Vitamin D supplementation and depression/depressive symptoms: clinical trials with older adults

NA, not assessed; RCT, randomised controlled trial; VitD, vitamin D; PHQ-8, patient health questionnaire depression scale; MINI, mini-international neuropsychiatric interview; HAMD-17, Hamilton depression rating scale-17; RSAS, revised social anhedonia scale; RPAS, revised physical anhedonia scale; HAMA-14, Hamilton anxiety rating scale-14; HDRS-17, Hamilton depression rating scale-17; BDI, Beck depression inventory; PANAS, positive and negative affect schedule; DASS-21, 21-item depression; CES-D, Center for Epidemiological Studies Depression; GDS-15, 15-item geriatric depression scale; MDI, major depression inventory; BDI-II, Beck depression inventory-II; FCPS, Fawcett–Clark pleasure capacity scale; GDS-LF30, long form 30-item GDS; HDRS-24, Hamilton depression rating scale-24; BDI-21, Beck depression inventory-21; HADS-14, hospital anxiety and depression scale; MADRS, Montgomery–sberg depression rating scale.

Seven (7/19) studies reported an improvement in depressive symptoms after VitD supplementation, eleven reported no improvement and one study lacked the power to assess due to sampling size⁽¹²³⁾. Considering the studies that observed depressive symptom improvement, five of seven were conducted with individuals with depression, and one of these (1/7) reported individuals with concomitant depression and VitD deficiency. VitD doses ranged from 600 to 300 000 IU, and the majority (6/7) used VitD doses above the dietary reference intake (DRI) (> 4000 IU/d). VitD doses of 600–4000 IU were used on a daily basis; 20 000–50 000 IU were used weekly; and the effect of a single dose of 150 000–300 000 IU was evaluated.

Compared with the seven studies with positive results, the eleven studies that did not report improvements tended to use lower VitD doses (<4000 IU) and longer periods (from 6 months to 5 years of supplementation). Of the eleven negative studies, only four used higher doses: Sanders et al. (2011) used a single dose of 500 000 IU in the winter for 3–5 years; Dean et al. (2011) used 5000 IU/d for 6 weeks; Kjægaard et al. (2012) used 20 000 IU/week for 6 months; and Gugger et al. (2019) used 24 000 IU or 60 000 IU for 12 months^{(113,120,128,129)} . The age range was higher in the studies that did not observe any improvement in depressive symptoms (individuals >70 years).

Two meta-analyses have shown controversial results in clinical trials with VitD supplementation. Spedding et al. (2014) showed that VitD supplementation (daily doses of ≥800 IU) could have an effect comparable to that of antidepressants in depressive symptoms⁽²¹⁾. Due to the methodological variability of the studies, the other meta-analysis conducted by Gowda et al. (2015) showed results that did not support this hypothesis⁽¹³¹⁾. In addition, a 5-year follow-up study found no potential effect of VitD on the incidence of depression⁽¹¹⁷⁾. Comparing the findings of the published meta-analysis with the studies searched in the present review, we observed that studies that did not observe improvements in depressive symptoms were conducted with older people with no diagnosis of depression, with lower VitD doses and for longer periods of follow-up. On the contrary, studies with positive results were conducted with younger populations with a diagnosis of depression and higher VitD doses for short periods of follow-up.

Key points of pre-clinical and clinical studies

Pre-clinical studies have pointed to the potential and possible effect of vitD on depression. However, despite a considerable number of clinical studies, it has not yet been possible to prove whether VitD can prevent or be used as an adjuvant treatment in depression. The data remain controversial. In addition, it is not possible yet to define which doses/amount of vitamin D would be most appropriate for depression.

Vitamin D and depressive symptoms: evidence from observational studies

Table 2 summarises the information from forty-four observational studies that investigated the relationship between VitD and depression/depressive symptoms in both adults and older adults since 2006.

Table 2. Vitamin D supplementation and depression/depressive symptoms: observational studies with older adults

95 % CI, 95% confidence interval; OR, odds ratio; SE, standard error; ES, effect size; IRR, incidence rate ratio; B, unstandardised beta; RR, relative risk; 25(OH)D, 25-hydroxycholecalciferol; 25(OH)D2, 25-hydroxyvitamin D2; 25(OH)D2, 25-hydroxyvitamin D3; DSM-IV, diagnostic and statistical manual of mental disorders; CIDI, composite international diagnostic interview; IDS-SR, self-report version of the inventory of depressive symptoms; GDS-15, geriatric depression scale with 15 items; CES-D, Center for Epidemiologic Studies Depression scale; HAM-D-17, Hamilton depression rating scale-17 items; HAD, hospital anxiety and depression scale; PHQ-9, patient health questionnaire-9; BDI, Beck depression inventory; CIDI, composite international diagnostic interview; IDS-SR, inventory of depressive symptoms – self-report; SCL, symptom checklist; HDRS, Hamilton depression rating scale; HICDA, hospital international classification of disease adaptation; DASS21, depression anxiety stress scale.

From over 15 years of research published, we observed that most studies included a mixed population with adults and older adults (27/44), were composed of people from cohort studies (27/44) and high-income economies countries (38/44), and used screening scales of depressive symptoms (37/44). The majority of studies performed a cross-sectional (27/44), followed by both a cross-sectional and longitudinal (10/44), and, finally, a longitudinal analysis (7/44). Considering the studies that included only older adults (≥60 years, 17/44), most were composed of people from a cohort (14/17) and performed a cross-sectional (10/17), followed by both a cross-sectional and longitudinal (4/17) and, finally, a longitudinal analysis (3/17). Moreover, only three studies were performed in low- or middle-income countries. This is an important issue because, according to the Mental Health Action Plan 2013–2030, there is an imbalance between research in high- and low/middle-income countries that needs to be corrected to ensure that they have appropriate cultural and economic strategies to respond to mental health needs and priorities⁽¹³⁾. One of their main goals is to strengthen information systems, evidence and research on mental health, and it suggests the development of more studies from low/middle-income countries.

It is difficult to compare the main differences between the studies because each study was different in terms of the method used to analyse data, the cut-off point for the classification of serum VitD concentrations and the screening for depressive symptoms or diagnosis for depression. However, an increasing number of studies have found an association between VitD and both depressive symptoms (32/44) and depression (7/44), specifically in those with cross-sectional analyses (24/44 and 7/44, respectively). Considering the studies in which researchers stratified the analysis by sex (7/44), the association was divergent because some authors^(147,169) found an association in both sexes, while other studies found an association for women^(163,175) or men^{(139,154,161)} . In studies that included both adults and older adults, only five (5/27) reported no association^{(136,138,157,171,172)} .

Among the studies that exclusively analysed data of older adults, those that performed a cross-sectional analysis (10/17) found an association between VitD and either depression^(161,173) or depressive symptoms^{(135,146,154–156,165,170,174)} , but two studies that stratified the analysis by sex found an association only for men^(154,161) . In studies that performed either longitudinal or cross-sectional and longitudinal analyses combined, the results are controversial. In the longitudinal analysis, one⁽¹⁵³⁾ did not find any effect of VitD on the course of depression or remission, while another found a decrease in the score of depression with an increase in VitD⁽¹³⁴⁾, and another⁽¹⁶⁹⁾ found an increase in depression score for a low level of VitD at 3 and 6 years follow-up in women and 3 years follow-up for men. In the cross-sectional and longitudinal combined analysis, some found a cross-sectional but not longitudinal association^(159,167) , another study⁽¹⁶⁰⁾ did not find an association at baseline and 1 year follow-up, just one found an association at 4 years follow-up and another found a cross-sectional association only for women and not in the follow-up⁽¹⁶³⁾. Nevertheless, most of these studies found a higher risk for depression when considering VitD concentrations below 20 ng/ml or 50 nmol/l^{(135,146,154,159,160,163,169,174)} . Other studies found higher risk when concentrations were below 10 ng/ml or 30 nmol/l^{(156,161,165,170)} , and two studies found a lower risk for depression in concentrations >36·7 nmol/l⁽¹⁷⁶⁾ and 92 nmol/l⁽¹⁶⁷⁾. Moreover, a meta-analysis with a mixed population showed that an increase of 10 ng/ml in individuals with low serum concentrations of 25(OH)D had a protective effect against depression, with a decrease of 4 % in the risk of depression in cross-sectional studies, and a decrease of 8 % in the incidence of depression in cohort studies⁽¹⁷⁷⁾. In studies involving only the elderly population, the same 10 ng/ml increase in serum 25(OH)D level was associated with a 12 % reduction in the risk of depression⁽¹⁷⁸⁾.

Key points of observational studies

Despite the controversial results from observational studies, the majority have pointed to a higher risk of depression with low levels of VitD (20 ng/ml or 50 nmol/l). However, the variability in methodology between studies is important to note. At this moment, it is not possible to suggest a possible VitD cut-off point specific for depression. Few studies were carried out with only older adults, as well as in low- and middle-income countries. Few longitudinal studies were carried out to demonstrate causality of depression due to low levels.

Future perspective

Older adults are considered a risk group for both depression and vitamin D deficiency, which justifies further studies to focus on this population. The ageing process is associated with a reduced ability to sustain homoeostasis, which could make elderly people more susceptible to pathological alterations, including neuropsychiatric disorders^(179,180) . Also, women in menopausal transition are at risk of depression due to a lot of changes (i.e. hormone-related context, stressful events in life)⁽¹⁸¹⁾. Moreover, older adults with depression present a higher risk of mortality⁽¹⁸²⁾, especially in low- and middle-income countries, and have difficulties accessing treatment^(4,183) . Another important factor is related to the adverse effects caused by antidepressant medications and the polypharmacy common in the elderly owing to the concomitance of several pathologies, which can facilitate the discontinuation of treatment^(184–186).

Facing the urgency to identify the modifiable risk factors associated with the aetiology of depression, helping with the treatment and prevention of this disorder, it is important to carry out more studies following a proper methodology since we have an important background related to pre-clinical studies. As highlighted by the WHO, these studies need to be developed especially in low- and middle-income countries, since these places have higher prevalence of depression^(12,13) . Further, observational studies have pointed to the preventive effect of adequate serum vitamin D concentrations on the development of depressive symptoms. More longitudinal studies have been suggested^(178,187) to better elucidate the preventive effects of VitD on depression/depressive symptoms.

Besides, the variability in the diagnosis of depression, differences in VitD cut-off reference values and methods for serum VitD analysis could influence those findings that are still controversial^(21,188,189) . Recently, the use of the standardised measurement of VitD proposed by the VitD Standardization Program (VDSP) has been recommended to improve clinical and public health practice, and it is important for future studies to apply this in their methodology^(190,191) . Considering the randomised control trial (RCT) that included the elderly population (>65 years), most of them did not present any improvements in depressive symptoms after VitD supplementation. This could be due to the lower VitD doses used in those studies, and because they were not performed in older individuals diagnosed with depression. This is an important aspect to be addressed in future RCTs.

Conclusion

Overall, this updated review suggests that the monitoring and maintenance of adequate VitD concentrations is crucial, especially in older adults, a population at risk for both VitD deficiency and depression. Several pre-clinical, clinical and observational studies have suggested that VitD could have a beneficial effect on depression/depressive symptoms due to its genomic and non-genomic actions in many pathways involved in the pathophysiology of depression.

Although studies presented controversial results, clinical studies have shown that older adults with depression/depressive symptoms could benefit from higher doses of VitD supplementation for short periods. However, more RCTs are needed to confirm which doses and for how long the treatment is needed to achieve the greatest benefit. From the observational studies, the results are still controversial, but the majority have reported an association between low serum concentrations of VitD and high risk for depression/depressive symptoms in older adults, pointing to a possible preventive effect of VitD. Additional studies with prospective designs, especially in low- and middle-income countries, will possibly help to better elucidate the impact of deficient VitD status for mental health in adulthood and, consequently, for the elderly.