Vitamin D and calcium metabolism in bipolar disorder: a scoping review of contradictory evidence

Abstract

Background

This scoping review synthesized evidence on vitamin D, parathyroid hormone (PTH), and serum calcium in bipolar disorder (BD) to evaluate their potential clinical utility.

Methods

A systematic PubMed search identified original studies examining calcium metabolism biomarkers (vitamin D, PTH, serum calcium) exclusively in BD populations. Findings were synthesized narratively due to substantial methodological heterogeneity.

Results

Fourteen studies met inclusion criteria, with small sample sizes (median n = 55) and predominantly cross-sectional designs. Vitamin D comparisons between BD patients and controls yielded contradictory results: three studies reported significantly lower levels in BD patients, two found significantly higher levels, and three found no differences. Vitamin D deficiency definitions varied widely (< 25 to < 50 nmol/L), precluding meaningful comparisons. Four cognition studies showed inconsistent associations with vitamin D, with negative correlations, age-dependent effects, or no associations reported. Two small vitamin D supplementation studies in bipolar spectrum disorders yielded contradictory results in distinct populations, with one in youth and the other in adults. Data on PTH and calcium were sparse and inconsistent.

Limitations

Study limitations included a single database search, substantial study heterogeneity, and inadequate control for confounders, including seasonal variation.

Conclusions

Evidence on calcium metabolism biomarkers in BD is contradictory and methodologically limited. Fundamental inconsistencies in vitamin D status between BD patients and controls, combined with conflicting supplementation data, preclude clinical recommendations. Routine vitamin D screening specifically for BD management cannot be supported. Large-scale, standardized studies are needed before clinical application.

Introduction

Bipolar disorder (BD) is a chronic recurrent mood disorder characterized by episodes ranging from acute depression to mania, affecting approximately 40 million people worldwide (GBD 2019 Mental Disorders Collaborators 2022, Rowland and Marwaha 2018). Despite advances in treatment, many patients continue to experience residual symptoms and functional impairment, highlighting the need for novel therapeutic approaches and biomarkers.

Recent research has increasingly focused on calcium metabolism in psychiatric disorders, particularly in relation to vitamin D and its regulatory systems (Cereda et al. 2021). Calcium homeostasis involves a complex network of serum calcium, parathyroid hormone (PTH), and vitamin D metabolites. When serum calcium decreases, PTH secretion increases, enhancing calcium resorption and stimulating production of active vitamin D [1,25(OH)₂D], which binds to vitamin D receptors (VDRs) and increases intestinal calcium absorption (Peacock 2010).

Vitamin D synthesis requires two sequential hydroxylations: first in the liver, producing 25(OH)D, then in the kidneys, forming the active hormone 1,25(OH)₂D (Bikle et al. 2021). Beyond calcium homeostasis, vitamin D exhibits diverse neurobiological functions including neuroprotection, inflammation suppression, and modulation of serotonin synthesis (Lang et al. 2019, Patrick and Ames 2015). In addition to that, VDRs are widely expressed in the central nervous system and contribute to immunomodulatory and neuroprotective effects (Menéndez and Manucha 2024).

The intersection of calcium signaling and BD pathophysiology has garnered considerable attention. Meta-analytic evidence suggests elevated intracellular calcium concentrations in patients with BD (Harrison et al. 2021), while genetic studies have implicated calcium channel genes in BD susceptibility (Smedler et al. 2020). Patient-derived stem cell studies suggest altered calcium signaling in BD neurons that responds to lithium, which is frequently used as mood-stabilizer in BD treatment (Chen et al. 2014). Additionally, stress-induced cortisol elevation may increase synaptic glutamate and intracellular calcium, potentially contributing to manic episodes (Moghaddam et al. 1994).

Previous research has identified associations between vitamin D deficiency and psychiatric conditions, including BD (Cereda et al. 2021, Berk et al. 2007). Studies suggest that insufficient vitamin D levels may negatively influence cognition through VDRs in brain regions responsible for memory and executive function (Laughlin et al. 2017). However, research on calcium metabolism in BD remains methodologically inconsistent, with studies employing heterogeneous populations, varying diagnostic criteria, different vitamin D assays, and inconsistent deficiency thresholds.

The present scoping review systematically synthesizes research conducted exclusively in BD populations, focusing on vitamin D, PTH, and serum calcium. Our objective was to evaluate current evidence regarding these biomarkers in BD pathophysiology, clinical presentation, and therapeutic applications. However, it is important to note that the peripheral biomarkers such as 25(OH)D, PTH, and serum calcium do not necessarily reflect potential mechanisms in the central nervous system that are implicated in BD. Therefore, we do not presume a direct correlation between them. We hypothesized that hypovitaminosis D, elevated PTH, and hypocalcemia would be associated with greater symptom severity and cognitive impairment in BD, and explored whether vitamin D supplementation represents a viable adjunctive therapeutic strategy.

Methods

This scoping review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines (Tricco et al. 2018). The protocol was registered on the Open Science Framework at 10.17605/OSF.IO/GQ6JU.

A systematic search was conducted in PubMed from database inception to 26 August 2024 using the following search string: “bipolar disorder” AND (“vitamin D” OR “parathyroid hormone” OR “serum calcium”). No publication date limits were applied. The search was restricted to English-language articles, and no additional filters were used. Reference lists of included studies were screened for potentially relevant publications. While additional databases such as Embase and PsycINFO would strengthen the search, resource constraints limited the search to PubMed, representing a methodological limitation that may have resulted in incomplete evidence synthesis. Furthermore, the incomplete capture of the evidence base could have contributed to the exclusion of endocrinology-focused or non-psychiatric journals that in turn may have influenced heterogeneity and contradictory findings in the final review.

Studies were included if they were original research articles examining vitamin D, parathyroid hormone, or serum calcium in individuals with BD and published in English. Exclusion criteria were non-original research, case reports (n < 10), animal studies, mixed psychiatric populations without extractable BD-specific data, and multivitamin studies without specific vitamin D data.

Study selection and data extraction

Figure 1 presents the PRISMA-ScR study selection process (Page et al. 2021). The PubMed search yielded 171 records. After removing 12 duplicates, 159 records were screened, and 143 were excluded based on title and abstract. Sixteen full-text articles were assessed, with 2 excluded (multivitamin study, mixed diagnoses), resulting in 14 included studies.

Fig. 1.

PRISMA-ScR study selection process

Two reviewers independently screened titles, abstracts, and full-text articles. Disagreements were resolved through discussion. Inter-rater agreement was substantial (κ = 0.78 for screening, κ = 0.82 for full-text review). Data were extracted on study characteristics, biomarker levels, vitamin D deficiency definitions, and clinical outcomes. Values were standardized using conversion factors (vitamin D: 1 ng/mL = 2.5 nmol/L; PTH: 1 pg/mL = 0.106 pmol/L; calcium: 1 mg/dL = 0.25 mmol/L). Two studies (Leser et al., 2023; Späth et al., 2023) utilized overlapping populations but examined different outcomes and were both included.

Due to heterogeneous study populations, biomarker measurement methods, and vitamin D deficiency definitions (ranging from < 25 to < 50 nmol/L), meta-analysis was not feasible. Results are presented as narrative synthesis organized by: (1) vitamin D levels in BD vs. controls, (2) vitamin D and cognition, (3) vitamin D supplementation, and (4) PTH and calcium in BD.

Results

Fourteen studies met inclusion criteria, with significant heterogeneity in study populations, methodological approaches, and findings. Tables 1, 2, 3 and 4 present detailed study characteristics and biomarker data. The evidence base was characterized by small sample sizes (median n = 55, range n = 16–199), predominantly cross-sectional designs (n = 11), and substantial methodological diversity in vitamin D deficiency definitions, mood states assessed, and outcome measures used.

Table 1.

Vitamin D Status and disease activity in bipolar disorder

Study	BD type/state	(BD/HC)	BD Vitamin D (nmol/L)	HC Vitamin D (nmol/L)	Deficiency Definition	Deficiency %	Main Finding	Associations/Correlations with clinical variables
Higher Vitamin D in BD
Li et al., 2022	Drug-naive, mixed	100/50	46.10 ± 20.15	40.95 ± 11.30	< 30 nmol/L	NR	BD > HC (p= 0.041, NS after correction)	No correlation with HAM-D, YMRS, HCL
Li et al., 2023	Mixed states	102/51	45.90 ± 17.68	39.05 ± 9.15	< 30 nmol/L	NR	BD > HC (p= 0.043)	No correlation with HAM-D, YMRS
Lower Vitamin D in BD
Altunsoy et al., 2018	Acute mania	26/26	37.90 ± 18.70	55.78 ± 22.00	< 25 nmol/L	30.8%	BD < HC (p= 0.002)	YMRS: r=−0.641, p< 0.001; CGI-S: r=−0.559, p= 0.003
İmre et al., 2023	Acute mania	34/30	Pre-tx: 26.15 ± 12.33; Post-tx: 39.88 ± 14.25	41.08 ± 13.20	< 50 nmol/L	55.8%	BD < HC (p< 0.001); Improvement Post-tx (p < 0.001)	NR
Astaneh et al., 2024	Mania/mixed	50/50	NR	NR	NR	NR	BD < HC (p< 0.05, males only)	CGI-S: r=−0.311, p= 0.028; YMRS: r=−0.464, p= 0.001; HAM-D: r=−0.393, p= 0.005
No Difference
Leser et al., 2023	Euthymic	86/93	56.37 ± 23.64	57.29 ± 23.95	< 50 nmol/L	40.7%	No difference (p> 0.05)	No correlations with mood clinical scales
Späth et al., 2023	Mixed	170/175	57.8 ± 24.3	61.5 ± 29.3	< 50 nmol/L	9.6%	No difference (p> 0.05)	24,25(OH)₂D vs. YMRS: r=−0.154, p= 0.040; VMR vs. YMRS: r=−0.238, p= 0.015
Van Rheenen et al., 2023	Mixed states	55/22	38.25 ± 8.50	33.00 ± 8.00	< 50 nmol/L	~ 26%	No difference (p> 0.05)	YMRS: F(1,49) = 0.18, p= 0.672; MADRS: F(1,49) = 0.008, p= 0.929
Latent Profile Analysis
Zheng et al., 2024	Bipolar depression	155/0	Low: 22.5 ± 4.5; Medium: 37.8 ± 6.0; High: 62.3 ± 12.8	N/A	< 50 nmol/L	32.9% (low profile)	Three vitamin D profiles identified	High vitamin D profile associated with better treatment response (OR = 7.00, p= 0.017)

24,25(OH)₂D − 24,25-dihydroxyvitamin D; BD - Bipolar Disorder; CGI-S - Clinical Global Impression Scale - Severity; HAM-D - Hamilton Depression Rating Scale; HC - Healthy Controls; HCL - Hypomania Check List; MADRS - Montgomery Åsberg Depression Rating Scale; OR - Odds Ratio; Pre-tx - Pre-Treatment; Post-tx - Post-Treatment; NR - Not Reported; NS - Not Significant; VMR - Vitamin D Metabolite Ratio; YMRS - Young Mania Rating Scale

Table 2.

Studies examining vitamin D and cognitive function in bipolar disorder

Study	Population (n)	Design	Vitamin D level (nmol/L)	Deficiency definition	Cognitive measure	Key finding
Chen et al., 2022	BD: 100	Cross- sectional	41.15 (72% deficient)	< 50 nmol/L	BAC-A	Age-dependent effects
Leser et al., 2023	BD: 86 HC: 93	Cross- sectional	BD: 56.37 ± 23.64	< 50 nmol/L	CVLT, SCWT, TMT	No associations
Li et al., 2023	BD: 102 HC: 51	Cross- sectional	BD: 45.90 ± 17.68 (Higher than HC)	< 30 nmol/L	RBANS	Negative correlation with memory
Van Rheenen et al., 2023	BD: 55 HC: 22	Cross- sectional	BD: 38.25 ± 8.50	< 50 nmol/L	MCCB	No associations

BAC-A - Brief Assessment of Cognition in Affective Disorders; BD - Bipolar Disorder; CVLT - California Verbal Learning Test; HC- Healthy Controls; MCCB - MATRICS Consensus Cognitive Battery; RBANS - Repeatable Battery for the Assessment of Neuropsychological Status; SCWT - Stroop Color and Word Test; TMT - Trail Making Test

Table 3.

Vitamin D supplementation studies in bipolar spectrum disorders

Study	Population	Design	n	Intervention	Duration	Primary outcome	Result
Marsh et al., 2017	BSD adults, depressed	RCT, double-blind	25 completed	5000 IU/day vs. placebo	12 weeks	MADRS, YMRS, HAM-A	No significant differences
Sikoglu et al., 2015	BSD youth, manic	Open-label Trial	16	2000 IU/day	8 weeks	YMRS, CDRS	Significant improvement

BSD - Bipolar Spectrum Disorder; CDRS - Children's Depression Rating Scale; HAM-A - Hamilton Anxiety Rating Scale; IU - International Units; MADRS - Montgomery Åsberg Depression Rating Scale; RCT - Randomized Controlled Trial; YMRS - Young Mania Rating Scale

Table 4.

Studies examining parathyroid hormone and serum calcium in bipolar disorder

Study	Population	n	PTH (pmol/L)	Calcium (mmol/L)	Key finding
de Filippis et al., 2022	BD outpatients	100	Evening type: 6.59 ± 1.34	Evening type: 2.38 ± 0.13	PTH mediates chronotype-mood relationship
Li et al., 2022	Drug-naive BD	100	BD: 4.36 ± 1.79 HC: 3.45 ± 1.40	BD: 2.27 ± 0.17 HC: 2.40 ± 0.39	Higher PTH, lower calcium in BD
Li et al., 2023	BD mixed	102	NR	BD: 2.28 ± 0.17 HC: 2.42 ± 0.41	Lower calcium in BD (p= 0.002)
Steardo et al., 2020	BD mixed	199	4.83 ± 2.29	2.36 ± 0.19	Higher PTH → younger onset, more hospitalizations

BD - Bipolar Disorder; HC - Healthy Controls; NR - Not Reported; PTH - Parathyroid Hormone

Vitamin D status and disease activity

Studies examining vitamin D status in BD patients compared to mentally healthy controls, along with associations between vitamin D and mood symptoms, yielded highly inconsistent results (Table 1). Three studies reported significantly lower vitamin D levels in BD patients: Altunsoy et al. (Altunsoy et al. 2018) found lower levels in acute mania patients compared to controls (37.90 ± 18.70 vs. 55.78 ± 22.00 nmol/L, p = 0.002), Astaneh et al. (Astaneh et al. 2024) reported lower levels in manic and mixed episodes (p < 0.05), and İmre et al. (İmre et al. 2023) observed lower levels in manic patients (26.15 ± 12.33 vs. 41.08 ± 13.20 nmol/L, p < 0.001).

Among the studies finding lower vitamin D levels in BD patients, several demonstrated significant associations with mood symptom severity. Altunsoy et al. (Altunsoy et al. 2018) found strong negative correlations between vitamin D levels and both Young Mania Rating Scale (YMRS) scores (r = −0.641, p < 0.001) and Clinical Global Impression Scale - Severity (CGI-S) scores (r = −0.559, p = 0.003) in patients with acute mania. The remission group did not differ significantly from either the healthy control or the acute mania episode groups in vitamin D levels. Astaneh et al. (Astaneh et al. 2024) reported comprehensive associations between vitamin D and multiple symptom domains, with vitamin D levels negatively associated with CGI-S (r = −0.311, p = 0.028), YMRS (r = −0.464, p = 0.001), and Hamilton Depression Rating Scale (HAM-D) scores (r = −0.393, p = 0.005).

Notably, gender-specific analyses revealed important differences. Astaneh et al. (Astaneh et al. 2024) found that after gender-specific analysis, the difference between patient groups and controls was significant only in males, despite both groups showing lower vitamin D levels overall. The control group in this study consisted of first-degree relatives of the patients. In the subgroup with disease duration of ≥ 10 years, mixed BD patients had significantly lower vitamin D levels compared to acute-phase bipolar mania patients (p < 0.05).

İmre et al. (İmre et al. 2023) provided unique longitudinal data, comparing vitamin D levels in BD patients during manic episode before treatment and in post-treatment remission. Vitamin D levels in post-treatment euthymic patients were significantly increased compared to the pre-treatment patient group (p < 0.001). However, this finding is confounded by the fact that patients with low vitamin D levels received an unspecified regimen of vitamin D supplementation during inpatient treatment (n = 19, 55.8% of the sample).

In direct contradiction, two studies found higher vitamin D levels in BD patients. Li et al. (2023)(Li et al. 2023) reported significantly higher levels in BD patients compared to controls (45.90 ± 17.68 vs. 39.05 ± 9.15 nmol/L, p = 0.043), while Li et al. (2022) (Li et al. 2022) initially found higher levels (46.10 ± 20.15 vs. 40.95 ± 11.30 nmol/L, p = 0.041), though statistical significance was lost after Bonferroni correction. Li et al. (2022) reported that vitamin D levels of males in the BD group were significantly higher than those of females (p < 0.001), indicating important sex differences in vitamin D status among BD patients. Notably, neither Li et al. study found significant correlations between vitamin D levels and mood rating scales (HAM-D, YMRS, Hypomania Checklist).

Three studies found no significant differences between groups in vitamin D levels. Leser et al. (Leser et al. 2023) reported similar levels in patients with BD and controls (56.37 ± 23.64 vs. 57.29 ± 23.95 nmol/L), Van Rheenen et al. (Rheenen et al. 2023) found no differences, and Späth et al. (Späth et al. 2023) reported no overall group differences. However, Späth et al. (Späth et al. 2023) found negative correlations between YMRS scores and vitamin D metabolites, including 24,25(OH)₂D (r = −0.154, p = 0.040) and vitamin D metabolite ratio (VMR) (r = −0.238, p = 0.015), despite the absence of group differences. Van Rheenen et al. (Rheenen et al. 2023) found no associations between vitamin D levels and either YMRS (F(1,49) = 0.18, p = 0.672) or Montgomery-Åsberg Depression Rating Scale (MADRS) scores (F(1,49) = 0.008, p = 0.929).

The most sophisticated analysis was conducted by Zheng et al. (Zheng et al. 2024), who used latent profile analysis to examine the association between vitamin D and depression severity in bipolar depression. The sample was stratified into three vitamin D profiles: low-level profile (32.9%), medium-level profile (51.0%), and high-level profile (16.1%) based on baseline vitamin D concentrations. Depression severity was assessed using the Zung Self-Rating Depression Scale (SDS). The three profiles showed significant differences in both baseline vitamin D and post-treatment vitamin D levels (p < 0.001). Critically, higher SDS improvement over two weeks of treatment was associated with higher odds of belonging to the high-level vitamin D profile compared to the low-level profile (OR = 7.00; 95% CI: 1.23, 39.78; p-trend = 0.017). This suggests an association between higher baseline vitamin D levels and better treatment response in bipolar depression, though this finding requires replication.

The findings on vitamin D status in BD were inconsistent, with studies reporting lower, higher, or comparable levels relative to controls. In addition to that, reported associations between vitamin D levels and mood symptom scales were mixed, with different studies reporting both significant and non-significant findings.

As demonstrated in Table 1, these contradictory findings likely reflect substantial methodological heterogeneity, including varying vitamin D deficiency definitions (ranging from < 25 to < 50 nmol/L), different BD subtypes and mood states, seasonal timing variations, geographic differences, gender-specific effects, and the complex relationship between vitamin D status and mood symptom severity.

Vitamin D and cognitive function

Four studies examined associations between vitamin D and cognitive performance, yielding inconsistent results (Table 2). Li et al. (2023) (Li et al. 2023) reported negative correlations between vitamin D levels and memory performance on the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) in patients with BD, including total scores (r = −0.224, p = 0.023), immediate memory (r = −0.207, p = 0.037), and delayed memory (r = −0.281, p = 0.004). Paradoxically, this same study found that patients with BD had higher vitamin D levels than controls.

Chen et al. (Chen et al. 2022) found age-dependent effects in euthymic bipolar type I disorder (BD-I) patients using the Brief Assessment of Cognition in Affective Disorders (BAC-A). In older patients (46–65 years), vitamin D levels were positively associated with composite scores (β = 0.347, p = 0.020) and verbal fluency (β = 0.252, p < 0.001). However, in younger patients (20–45 years), these associations were negative (composite scores: β = −0.344, p = 0.040; verbal fluency: β = −0.230, p < 0.001).

Two studies found no significant associations. Leser et al. (Leser et al. 2023) reported no correlations between vitamin D metabolites and attention, memory, or executive function domains in BD patients or controls. Similarly, Van Rheenen et al. (Rheenen et al. 2023) found no association between vitamin D status and global cognitive performance on the MATRICS Consensus Cognitive Battery (F(1,67) = 0.72, p = 0.410).

The findings on vitamin D and cognitive function in BD were inconsistent, with studies reporting negative correlations with memory, age-dependent effects, or no associations.

As demonstrated in Table 2, the conflicting nature of these findings, combined with different cognitive assessment tools and age-related effects, precludes definitive conclusions about vitamin D’s role in BD-related cognitive dysfunction.

Vitamin D supplementation

Only two small studies examined vitamin D supplementation in bipolar spectrum disorders (BSD), yielding contradictory results (Table 3). Marsh et al. (Marsh et al. 2017) conducted a randomized controlled trial in 33 adults with BSD experiencing depressive symptoms. The study found no significant differences in depression (MADRS), mania (YMRS), or anxiety (HAM-A) scores between vitamin D supplementation (5000 IU daily) and placebo groups over 12 weeks (p = 0.89, p = 0.51, p = 0.89, respectively). Only 25 of 33 participants completed the trial, and there was no correlation between vitamin D level changes and symptom improvement.

In contrast, Sikoglu et al. (Sikoglu et al. 2015) reported positive findings in an open-label trial of 16 youth with BSD. After 8 weeks of vitamin D supplementation (2000 IU daily), significant improvements were observed in YMRS scores (t = −3.66, p = 0.002) and Children’s Depression Rating Scale scores (t = −2.93, p = 0.01), along with increased anterior cingulate cortex γ-aminobutyric acid (GABA) levels (t = 3.18, p = 0.007).

The limited reported data on vitamin D supplementation is inconclusive, with seemingly contradictory results.

As shown in Table 3, the conflicting results between these studies, differences in population (adults vs. youth), study design (randomized controlled vs. open-label), dosing regimens, and small sample sizes preclude any conclusions about vitamin D supplementation efficacy in BD.

Parathyroid hormone and serum calcium

Few studies examined PTH and serum calcium, with limited and inconsistent findings (Table 4). Two studies suggested associations between elevated PTH and illness characteristics. Steardo et al. (Steardo et al. 2020) found that higher PTH levels were associated with younger age of onset (β = −0.289, p = 0.032), more hospitalizations (β = 0.160, p = 0.017), higher childhood trauma scores (β = 1.276, p = 0.001), and lithium treatment (β = 0.179, p = 0.013).

De Filippis et al. (Filippis et al. 2022) reported the statistical mediation effect of PTH levels between evening chronotype and mood symptom severity, accounting for 25.5% of the effect on depression scores (indirect effect: −5.15, p = 0.003) and 27.5% of the effect on anxiety scores (indirect effect: −3.16, p = 0.01).

Serum calcium findings were inconsistent. Li et al. (2022, 2023) (Li et al. 2023, Li et al. 2022) reported significantly lower calcium levels in BD patients compared to controls, while Steardo et al. (Steardo et al. 2020) found positive correlations between calcium levels and anxiety scores. However, İmre et al. (İmre et al. 2023) and Sikoglu et al. (Sikoglu et al. 2015) found no significant differences in calcium levels between groups or over time.

Few studies have examined PTH and serum calcium in BD. Several studies reported higher PTH levels and associations with clinical variables. Lower serum calcium levels were reported in some studies, while others found no significant differences.

As demonstrated in Table 4, the sparse and contradictory nature of these findings, combined with the potential confounding effect of lithium treatment on calcium metabolism, limits interpretation of PTH and calcium roles in BD.

Summary of key findings

The evidence base was characterized by substantial heterogeneity and contradictory findings. Vitamin D status comparisons between BD patients and controls showed inconsistent patterns, with some studies suggesting associations between vitamin D levels and mood symptom severity while others found no relationships. Cognitive function studies yielded highly inconsistent results with potential age-dependent effects. The limited evidence of supplementation was insufficient to support clinical recommendations, and PTH/calcium data were too sparse and inconsistent to draw meaningful conclusions. Gender-specific effects and methodological heterogeneity further complicated interpretation of findings.

Discussion

This scoping review examined calcium metabolism biomarkers in BD, revealing a complex and largely inconsistent evidence base that does not support definitive conclusions about their clinical utility. Contrary to our hypothesis that hypovitaminosis D, elevated PTH, and hypocalcemia would be consistently associated with greater symptom severity, the evidence demonstrates substantial contradictions and methodological limitations.

The most striking finding is the fundamental contradiction in vitamin D status between BD patients and healthy controls. While three studies reported lower vitamin D levels in BD patients (Altunsoy et al. 2018, Astaneh et al. 2024, İmre et al. 2023), two studies by the same author found significantly higher levels (Li et al. 2023, Li et al. 2022), and three studies found no differences (Leser et al. 2023, Rheenen et al. 2023, Späth et al. 2023). Contributing factors include heterogeneous vitamin D deficiency definitions ranging from < 25 to < 50 nmol/L, varied mood states at assessment, and lack of control for seasonal and geographic variations (Wakeman 2021).

Several factors may contribute to these contradictory findings. Publication bias may favor studies showing significant associations while null results remain unpublished (Spedding 2014). Medication effects, particularly mood stabilizers, may confound vitamin D metabolism through effects on cytochrome P450 enzymes (Wakeman 2021, Kovacs et al. 2022). Additionally, timing of blood sampling relative to mood episodes and lifestyle factors including altered sunlight exposure and physical activity in BD patients may independently affect vitamin D status (Cereda et al. 2021). Sex-specific effects also complicate interpretation, as Astaneh et al. (Astaneh et al. 2024) found significant differences only in males, while Li et al. (2022) (Li et al. 2022) reported higher vitamin D levels in BD males compared to females.

Cognitive function findings are similarly inconsistent. Li et al. (2023) (Li et al. 2023) paradoxically reported both higher vitamin D levels in BD patients and negative correlations with cognitive performance, while Chen et al. (Chen et al. 2022) found opposite associations depending on age. Two other studies found no associations (Leser et al. 2023, Rheenen et al. 2023) with cognitive test levels. Despite vitamin D receptors being abundantly expressed in brain regions implicated in mood regulation and cognitive function (Eyles et al. 2005), the relationship between peripheral vitamin D levels and central nervous system activity may be more complex than assumed.

The evidence for vitamin D supplementation is insufficient for clinical recommendations, contrasting with previous systematic review suggesting potential benefits (Spedding 2014). Only two small studies with contradictory results in different populations using different designs are available (Marsh et al. 2017, Sikoglu et al. 2015). The negative findings from the higher-quality randomized trial contrast with positive open-label findings in youth, but small sample sizes and methodological differences prevent meaningful conclusions. PTH and calcium findings are too sparse and inconsistent to inform practice, with potential lithium confounding inadequately addressed (Kovacs et al. 2022).

This review has several significant limitations. Our search was limited to PubMed, which may have resulted in an incomplete synthesis of evidence. The observed inconsistencies between studies may have been caused in part by the failure to include relevant studies published in journals that are indexed in other databases. The included studies had small sample sizes (median n = 55), predominantly cross-sectional designs, and varying methodological rigor. Heterogeneity in populations, methods, and assessments prevented meta-analysis. Many studies failed to control for seasonal variation, medications, and comorbidities that could influence calcium metabolism (Bikle et al. 2021).

There was a significant methodological heterogeneity in the included studies. Vitamin D measurement methods varied across studies, including chemiluminescent immunoassays (Astaneh et al. 2024, İmre et al. 2023, Chen et al. 2022, Marsh et al. 2017, Steardo et al. 2020, Filippis et al. 2022), electrochemiluminescence assays (Altunsoy et al. 2018, Li et al. 2023, Li et al. 2022), enzyme-linked immunosorbent assays (Rheenen et al. 2023), and liquid chromatography–tandem mass spectrometry (Leser et al. 2023, Späth et al. 2023), while some studies did not report assay methods (Zheng et al. 2024, Sikoglu et al. 2015). It should be noted that, Astaneh et al. (Astaneh et al. 2024) described their method as a “radioimmunoassay technique utilizing the Architect i2000 (Abbott Laboratories)”. However, Architect i2000 operates on chemiluminescence and was therefore classified accordingly. Definitions of vitamin D deficiency differed substantially, ranging from < 25 nmol/L (Altunsoy et al. 2018, İmre et al. 2023) to < 30 nmol/L (Li et al. 2023, Marsh et al. 2017) and < 50 nmol/L (Leser et al. 2023, Rheenen et al. 2023, Späth et al. 2023, Zheng et al. 2024, Chen et al. 2022), with several studies not specifying a predefined threshold (Astaneh et al. 2024, Li et al. 2022, Sikoglu et al. 2015, Steardo et al. 2020, Filippis et al. 2022). Exclusion or control of vitamin D supplementation was not explicitly stated by several studies (Astaneh et al. 2024, Li et al. 2023, Li et al. 2022, Rheenen et al. 2023, Chen et al. 2022, Sikoglu et al. 2015, Steardo et al. 2020). While the exact definition of vitamin D supplementation as an exclusion criterion varied across the remaining studies. Lithium exposure differed across studies, with studies on drug-naïve BD patients (Li et al. 2023, Li et al. 2022) and studies including lithium treatment BD patients (Steardo et al. 2020). Study populations varied in mood states, including acute manic, depressive, mixed, and euthymic states, as well as in BD types (BD I, BD II, BD NOS, BSD). Clinical outcome measures were also varied, with mood symptom severity and cognitive function evaluated with different tools, limiting direct comparability across studies. In addition to that, the included studies differed in age distribution and geographic origin. This methodological heterogeneity limits the direct comparison of studies and may influence the inconsistency in reported findings.

The observed inconsistencies may stem from both true biological heterogeneity and methodological artefacts. Biological heterogeneity may be influenced by BD subtype, mood states, illness duration or sex-specific differences. On the other hand, methodological artefacts may arise from the use of different vitamin D assays, varying thresholds for deficiency, geographical location, season and timing of sampling, lithium treatment. Clearly distinguishing between the biological and methodological influences requires further research.

Notably, lithium-related alterations in calcium metabolism require detailed attention in future research as it is known to influence calcium and PTH regulation. In addition to that, the potential influence of sex-specific differences in BD requires future research to be designed to compare not only the BD patients and healthy controls but also the subgroups within them based on sex.

Although findings from peripheral biomarker studies are inconsistent, calcium-related pathways have been implicated in BD research and therefore further methodologically rigorous research is needed to investigate potential clinical utility of peripheral calcium metabolism biomarkers in BD. Vitamin D receptors regulate over 1000 genes involved in neurotransmitter synthesis and immune function (Lang et al. 2019), while meta-analytic evidence shows elevated intracellular calcium in BD (Harrison et al. 2021). However, the findings in our review were heterogeneous and derived predominantly from cross-sectional studies, making it difficult to draw reliable conclusions and limiting the assessment of more complex patterns of association. The sophisticated latent profile analysis by Zheng et al. (Zheng et al. 2024) reported an association between higher vitamin D levels and a better treatment response in bipolar depression patients. However, the validity of this association needs to be replicated by further research to confirm its potential predictive value. The inconsistencies may also reflect BD’s heterogeneous nature, as different subtypes and mood states may have varying relationships with calcium metabolism (Rowland and Marwaha 2018).

Based on current evidence, routine vitamin D screening as a disorder-specific biomarker for the management of BD cannot be recommended. However, this conclusion does not contradict general medical indications for vitamin D screening and supplementation. While vitamin D deficiency is common in psychiatric populations, our findings do not support its use as a BD biomarker or treatment predictor (Cereda et al. 2021). Similarly, vitamin D supplementation cannot be recommended as adjunctive BD treatment based on limited and contradictory evidence.

Future research requires large-scale, longitudinal studies with standardized assays, consistent deficiency definitions, and adequate confounder control. Studies should stratify by BD subtype and mood state while controlling for seasonal, geographic, and medication factors. Investigation of threshold effects and central nervous system vitamin D markers may provide insights beyond peripheral levels. Gender-specific analyses should be routinely conducted given emerging evidence of sex differences in vitamin D metabolism in BD.

In conclusion, current evidence does not support the clinical utility of vitamin D, PTH, or serum calcium as BD biomarkers. More rigorous, standardized research is required before clinical recommendations can be justified. Until then, clinicians should focus on evidence-based BD treatments while considering vitamin D status only for general health maintenance.

Abbreviations

ACC

Anterior Cingulate Cortex

BAC

A-Brief Assessment of Cognition in Affective Disorders

BD

Bipolar Disorder

BSD

Bipolar Spectrum Disorder

CDRS

Children’s Depression Rating Scale

CGI

S-Clinical Global Impression Scale-Severity

CVLT

California Verbal Learning Test

GABA

γ-aminobutyric acid

HAM

A-Hamilton Anxiety Rating Scale

HAM

D-Hamilton Depression Rating Scale

HC

Healthy Controls

IU

International Units

MADRS

Montgomery-Åsberg Depression Rating Scale

MCCB

MATRICS Consensus Cognitive Battery

NR

Not Reported

NS

Not Significant

OSF

Open Science Framework

PRISMA

ScR-Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews

PTH

Parathyroid Hormone

RBANS

Repeatable Battery for the Assessment of Neuropsychological Status

RCT

Randomized Controlled Trial

SDS

Zung Self-Rating Depression Scale

SCWT

Stroop Color and Word Test

TMT

Trail Making Test

VDRs

Vitamin D Receptors

VMR

Vitamin D Metabolite Ratio

YMRS

Young Mania Rating Scale

1,25(OH)₂D − 1,25

dihydroxyvitamin D

24,25(OH)₂D − 24,25

dihydroxyvitamin D

25(OH)D − 25

hydroxyvitamin D

Author contributions

Amirzhan Kulmagambetov: Conceptualization and review planning, Literature screening and selection, Data extraction and analysis, Writing - original draft preparationAdelina Tmava-Berisha: Conceptualization and review planning, Literature screening and selection, Critical analysis and interpretation, Writing - review and editing, SupervisionFrederike Fellendorf: Critically revised the final version.Melanie Lenger: Critically revised the final versionTatjana Stross: Critically revised the final versionMarko Stijic: Critically revised the final versionEva Fleischman: Critically revised the final versionJulia Ilic: Critically revised the final versionAlexander Finner: Critically revised the final versionAnna Ramirez-Obermayer: Critically revised the final versionJohanna Georgi: Critically revised the final versionAlexander Maget: Critically revised the final versionAmrei Lässer: Critically revised the final versionClaudia Mittmansgruber: Critically revised the final versionStefan Smolle: Critically revised the final versionAlfred Häussl: Critically revised the final versionJonas Schuller: Critically revised the final versionDino Hasic: Critically revised the final versionSusanne Bengesser: Critically revised the final versionRobert Queissner: Critically revised the final versionNina Dalkner: Critically revised the final version, SupervisionEva Z. Reininghaus: Conceptualization and review planning, Project administration, SupervisionAll authors have approved the final version.

Funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. Open Acces was funded by “Das Land Steiermark, Department of Economy, Tourism, Science and Research”.

Data availability

No new data were generated or analyzed in this study. All data supporting the findings of this review are derived from previously published studies, which are cited in the reference list.

Declarations

Ethics approval and consent to participate

This scoping review synthesizes data from previously published studies. No primary data were collected from human participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.