Causal relationship between circulating micronutrient levels and bipolar A Mendelian randomization study

Authors: Ting You, Xiuwen Yang, Huaili Feng, Zhaoyi Liu, Chengzhi Chen, Jingfu Qiu

Abstract

Bipolar disorder (BD) is a significant mental health condition characterized by alternating episodes of mania and depression, impacting millions worldwide. Genetic factors are known to play a major role in BD, but the influence of circulating micronutrient levels remains unclear. Two-sample Mendelian randomization (MR) and multiple MR were used to assess the impact of 15 circulating micronutrients, including copper, selenium, zinc, calcium, iron, magnesium, potassium, carotene, folate, and vitamins A, B6, B12, C, D, and E on BD risk. Genetic data were sourced from Genome-Wide Association Studies, and BD case–control data were obtained from the FinnGen consortium. Inverse variance weighted was the main analysis method. The heterogeneity of the instrumental variables was assessed using inverse variance weighted and MR-Egger, simple mode, weighted median, and weighted mode, and the horizontal pleiotropy of the instrumental variables was assessed using MR-Egger and MR-PRESSO. Higher genetically predicted levels of vitamin B12 (odds ratio = 1.81, 95% confidence 1.19–2.74, P = .005) and carotene (odds ratio = 1.49, 95% confidence 1.07–2.08, P = .017) are associated with an increased risk of BD. No significant causal associations were found between BD and the levels of other circulating micronutrients. Sensitivity analyses confirmed the robustness of these findings, with no evidence of heterogeneity or horizontal pleiotropy detected. This study suggests that elevated levels of vitamin B12 and carotene may increase the risk of BD, pointing to potential nutritional targets for intervention. Further research is needed to elucidate these findings and understand their underlying mechanisms.

1. Introduction

Bipolar affective disorder, also known as bipolar disorder (BD), is a chronic mental health condition characterized by extreme mood swings that include emotional highs (mania or hypomania) and lows (depression).^[1]^ These mood episodes can last days to weeks and significantly impact an individual’s daily functioning and quality of life. According to the World Health Organization, BD affects approximately 45 million people worldwide, with a lifetime prevalence of about 1% to 2% in the general population.^[2]^ The disorder typically manifests in late adolescence or early adulthood and follows a recurring course throughout a person’s life.^[3]^ The etiology of BD is complex and multifactorial, involving a combination of genetic, environmental, and neurobiological factors. Family studies have shown that genetic factors contribute significantly to the risk of developing BD, with heritability estimates ranging from 60% to 80%.^[4]^ Environmental factors, such as stressful life events, substance abuse, and disturbances in sleep–wake cycles, also play a critical role in triggering and exacerbating mood episodes.^[5,6]^ Neurobiologically, BD has been associated with dysregulation in neurotransmitter systems (including serotonin, dopamine, and norepinephrine), abnormalities in brain structure and function, and altered circadian rhythms.^[7,8]^

Micronutrients, including vitamins and trace elements, are essential for the proper functioning of the brain and nervous system.^[9–11]^ They play vital roles in neurotransmitter synthesis, myelin formation, and neuronal signaling, all of which are crucial for maintaining mental health. Deficiencies or imbalances in micronutrients can impair brain function and contribute to the development and progression of psychiatric disorders.^[12]^ Multiple micronutrients have been linked to mental health. For example, vitamin D is known to influence brain development and function, and its deficiency has been linked to mood disorders, including depression and BD.^[13,14]^ Similarly, B vitamins, such as B6, B12, and folate, are involved in the synthesis of neurotransmitters and homocysteine metabolism, and their deficiencies have been associated with cognitive impairments and mood disorders.^[15,16]^ Previous studies have explored the relationship between micronutrient levels and various psychiatric conditions. Observational studies have suggested that deficiencies in vitamins D, B12, and folate are more common in individuals with depression and BD. Moreover, randomized controlled trials (RCTs) have investigated the effects of micronutrient supplementation on mood and cognitive function, yielding mixed results. Some trials have shown beneficial effects of omega-3 fatty acids, vitamin D, and B vitamins on mood stabilization and cognitive performance, while others have reported no significant improvements.^[17,18]^ Despite these findings, the causal relationships between micronutrient levels and psychiatric disorders remain unclear. Observational studies are often confounded by various factors, such as dietary habits, lifestyle, and socioeconomic status, making it difficult to establish causality. Additionally, RCTs on micronutrient supplementation have limitations, including small sample sizes, short intervention durations, and variability in supplementation dosages and formulations. In principle, although larger and more rigorous RCTs could overcome these limitations, their high cost and operational complexity often render them impractical. Nevertheless, RCTs remain the gold standard for evaluating the health effects of micronutrients.

Because genetic variants are randomly assigned at conception, Mendelian randomization (MR) can mitigate confounding and reverse causality that often plague observational studies.^[19]^ Moreover, compared with RCTs, MR studies can generate data quickly and at lower cost, making them a useful complement for exploring the causal relationships between micronutrient levels and psychiatric disorders. The primary objective of this study is to determine whether genetically predicted levels of 15 micronutrients, including copper, selenium, zinc, calcium, iron, magnesium, potassium, carotene, folate, vitamin A, vitamin B6, vitamin B12, vitamin C, vitamin D, vitamin E, are causally associated with the risk of developing bipolar affective disorder. We aim to provide more reliable evidence on the potential roles of these micronutrients in the etiology of BD. The findings could have significant implications for public health and clinical practice, highlighting the importance of adequate micronutrient intake for mental health and providing a basis for targeted nutritional strategies in individuals at risk for BD.

2. Methods

2.1. Study design

This study aimed to investigate the potential genetic association between 15 circulating micronutrients (copper, selenium, zinc, calcium, iron, magnesium, potassium, carotene, folate, vitamin A, vitamin B6, vitamin B12, vitamin C, vitamin D, and vitamin E) and the risk of BD using a robust 2-sample MR analysis. The primary objective was to utilize summary statistics derived from Genome-Wide Association Studies (GWAS) as instrumental variables (IVs) to explore potential causal relationships between variations in these micronutrients and susceptibility to BD. The study’s workflow is illustrated in Figure 1.

Figure 1.: A schematic summary of the study design.

2.2. GWAS data source

2.2.1. Micronutrients

We conducted a comprehensive search for published GWAS involving individuals of European ancestry through the GWAS Catalog, PubMed, and the IEU OpenGWAS project. To minimize exposure mis-classification, the study focused on individuals of European descent and gave priority to GWAS datasets that measure serum micronutrient concentrations, which more closely reflect true body levels than the intake-based estimates used in conventional epidemiological studies. Accordingly, this criterion constituted a central component of our data-source selection strategy.

We found no relevant GWAS for vitamins B1, B2, B3, B5, B7, sulfur, iodine, chloride, and fluoride. Additionally, GWAS for vitamin K, sodium, cobalt, chromium, and molybdenum were excluded due to a lack of significant genome-wide results.^[20–22]^ Ultimately, we identified 15 micronutrients of copper, selenium, zinc, calcium, iron, magnesium, potassium, carotene, folate, vitamin A, vitamin B6, vitamin B12, vitamin C, vitamin D, and vitamin E were obtained from the IEU OpenGWAS project (https://gwas.mrcieu.ac.uk/), which includes larger sample sizes and more recent data. Sample sizes across datasets ranged from approximately 2600 to 418,000 individuals. In the UK Biobank sub-studies, fasting morning blood was collected once during 2006 to 2010. For the micronutrient-specific cohorts, samples were drawn under the same fasting protocol between 1997 and 2004. All measurements were performed on high-sensitivity platforms, and the original GWAS had already adjusted for batch and seasonal effects, allowing the data to be used directly in downstream MR analyses.

Then we screened single nucleotide polymorphisms (SNPs) significantly associated with circulating micronutrient levels using a genome-wide significance threshold (P < 5 × 10^-6)^. SNPs were then clumped based on linkage disequilibrium, defined as r^2^ < 0.001 with a clumping window > 10,000 kb.

2.2.2. Bipolar disorder

Summary data on BD from genome-wide association studies were acquired from the FinnGen consortium. In the FinnGen 10th release, bipolar-disorder cases were identified from Finland’s national health registries using an ICD-code algorithm with diagnostic codes ICD-10 F30–F31, ICD-9 296, or ICD-8 296.1–296.3. Case ascertainment was further stratified by linkage to inpatient and specialized outpatient individuals with at least 2 specialist contacts (including at least 1 hospitalization) were classified as the “strict” endpoint (BIPOLAR_STRICT), whereas those with at least 1 hospitalization or 2 outpatient visits were assigned to the “broad” endpoint (BIPOLAR_BROAD). The 10th release of the FinnGen consortium provided data comprising 7569 BIPOLAR_STRICT cases and 359,290 controls. This dataset excluded individuals with ambiguous gender, high genotype deletion rates (>5%), excessive heterozygosity (± 4 standard deviation [SD]), and non-Finnish ancestry to ensure data integrity and reliability.

2.3. Statistical analysis

In this MR study, we utilized several methods (inverse variance weighted [IVW], MR-Egger, simple mode, weighted median, and weighted mode) to investigate the causal relationship between circulating micronutrient levels and the risk of BD. Detailed descriptions of these methodologies can be found in previous publications.^[23–26]^

IVW is a key and widely used approach, estimating the average effect of genetic variants on causality through a weighted linear regression model. The results from IVW are unbiased provided there is no horizontal pleiotropy.^[27]^ To assess heterogeneity in the results, Cochrane Q test was applied; if heterogeneity was detected (P < .05), a random-effects model was used to ensure the reliability of the IVW results. MR-PRESSO analysis was conducted to detect and remove outliers, as their presence can indicate horizontal pleiotropy.^[28]^ Additionally, we performed a horizontal pleiotropy analysis using MR-Egger regression. A significant MR-Egger intercept (P < .05) suggests the presence of horizontal pleiotropy, indicating a violation of the MR analysis assumptions. A sensitivity analysis was also conducted to ensure the robustness of the findings. This included a leave-one-out analysis, a common technique where each SNP is sequentially excluded to evaluate the reliability of the remaining SNPs. Additionally, we conducted a multivariable MR (MVMR) analysis. The MVMR model used the lead SNP of each of the 2 micronutrients (clumping r^2^ < 0.001, kb = 10 000) as IVs and included additional genetically predicted values of related nutrients as covariates to control for potential multinutrient multicollinearity.

3. Results

3.1. MR estimates

To explore the causal association between SNPs related to copper, selenium, zinc, calcium, iron, magnesium, potassium, carotene, folate, vitamin A, vitamin B6, vitamin B12, vitamin C, vitamin D, and vitamin E and the risk of BD, we conducted a 2-sample MR study using genetic data from individuals of European ancestry. A total of 178 SNPs associated with the 15 micronutrients and BD were identified (Table S1, Supplemental Digital Content, https://links.lww.com/MD/Q935). We employed 5 different methods to estimate the causal relationship between the IVs and BD: IVW, MR-Egger, simple mode, weighted median, and weighted mode. By examining the interactions between these variables and outcomes, we aimed to determine the validity of inferring a causal relationship. The MR estimates derived using the IVW method indicated that predicted circulating concentrations of vitamin B12 (OR = 1.81, 95% confidence interval [CI]: 1.19–2.74, P = .005) and carotene (OR = 1.49, 95% CI: 1.07–2.08, P = .017) were suggestively associated with an increased risk of BD (Fig. 2A and B). However, no significant associations were found between the risk of BD and the predicted concentrations of copper (OR = 1.02, P = .655), selenium (OR = 1.01, P = .742), zinc (OR = 1.02, *P *= .633), calcium (OR = 0.96, P = .822), iron (OR = 1.13, *P *= .542), magnesium (OR = 1.33, P = .093), potassium (OR = 1.07, P = .722), folate (OR = 1.14, P = .577), vitamin A (OR = 0.95, P = .990), vitamin B6 (OR = 1.07, P = .667), vitamin C (OR = 1.31, P = .270), vitamin D (OR = 1.04, P = .864), and vitamin E (OR = 0.94, P = .730) (Table 1).

Figure 2.: MR analysis on the association between circulating micronutrient levels and BD. (A) Circle diagram for correlation analysis of 15 circulating micronutrient levels with BD. (B) The MR analysis results between carotene, vitamin B12 and BD. BD = bipolar disorder, MR = Mendelian randomization.

We found no evidence that micronutrient levels of copper (OR = 1.02, P = .655), selenium (OR = 1.01, P = .742), zinc (OR = 1.02, P = .633), calcium (OR = 0.96, P = .822), iron (OR = 1.13, P = .542), magnesium (OR = 1.33, P = .093), potassium (OR = 1.07, P = .722), folate (OR = 1.14 P = .577), vitamin A (OR = 0.95, P = .990), vitamin B6 (OR = 1.07, P = .667), vitamin C (OR = 1.31, P = .270), vitamin D (OR = 1.04, P = .864), and vitamin E (OR = 0.94, P = .730) are associated with the risk of BD (Table 1).

3.2. Sensitive analysis

Heterogeneity was assessed using Cochrane Q test from both the IVW and MR-Egger methods, with a significance threshold of P < .05. The MR-Egger intercept was employed to detect horizontal pleiotropy. Both Cochrane Q test and the MR-Egger intercept indicated no evidence of heterogeneity or horizontal pleiotropy in the analyses of the 15 micronutrient levels (Tables 2–5). Additionally, scatter plots, forest plots, funnel plots, and leave-one-out plots demonstrated that the results from the MR analyses of the 15 micronutrients and the risk of BD were robust across various sensitivity analyses (Fig. 3).

Figure 3.: Scatter plots, forest plots, funnel plots, and leave-one-out plots of the 15 micronutrients and the risk of BD. BD = bipolar disorder.

3.3. Multivariable MR analysis

Through the above analysis, we found that the 2 micronutrients, carotene and vitamin B12, were associated with BD. Therefore, we further used multivariate MR analysis to explore whether both micronutrients, carotene and vitamin B12, had independent causal effects on the disease (Fig. 4). The results of MR analysis showed that each 1-SD increase in carotene was associated with a 39.8 % higher risk of BD (OR = 1.398, 95 % CI: 1.09–1.79, P = .007), whereas each 1-SD increase in vitamin B12 was linked to an 82.4 % higher risk (OR = 1.824, 95 % CI: 1.36–2.45, P < .001). Cochrane Q test found no significant heterogeneity among the instruments, the MR-Egger intercept indicated no directional pleiotropy, and leave-one-out analyses revealed that no single SNP drove the estimates, confirming the robustness of the independent causal effects for both nutrients.

Figure 4.: Multivariable MR analysis on the association between micronutrients (vitamin B12 and carotene) and BD. BD = bipolar disorder, MR = Mendelian randomization.

4. Discussion

Our study explored the causal relationships between genetically predicted circulating micronutrient levels and the risk of BD using MR methods. Among 15 micronutrients examined, MR analysis showed that genetically predicted higher levels of carotene and vitamin B12 were significantly associated with increased BD risk. Specifically, higher levels of carotene and vitamin B12 were found to have independent causal effects on BD risk. These findings suggest that carotene and vitamin B12 may play a role in the etiology of BD and warrant further investigation into their mechanisms of action in the brain.

Meta-analyses of case–control and cohort studies have consistently revealed that low serum vitamin B12 or carotene levels are associated with higher mood-symptom scores, longer hospital stays, or an earlier age at onset.^[29–38]^ In other words, previous observational work has almost uniformly reported that low B12 or low carotene relates to increased depression, cognitive decline, or greater bipolar-disorder severity, whereas our MR results indicate that high levels raise BD risk. This directional discrepancy is unlikely to be due to random error and warrants mechanistic investigation. Although this study observed relatively large odds ratios, the lifetime baseline risk of BD is only about 1%, so the absolute change in risk is small. Based on MR estimates, raising vitamin B12 from the 10th to the 90th percentile of the population distribution would increase lifetime risk from 1% to approximately 1.8%, roughly 8 additional cases per 1000 individuals. This increment is insufficient to justify avoiding foods rich in B12 or carotene, especially given the well-established benefits of these nutrients for hematopoiesis, vision, and immune function. Supraphysiological levels may act as a risk modifier only in genetically susceptible subgroups, whereas for the general population maintaining the usual recommended intake remains both safe and advantageous.

Chronic elevation of vitamin B12 potentiates the methionine cycle, raising S-adenosyl-methionine levels, multiple animal studies have confirmed that this induces genome-wide DNA hypermethylation.^[39]^ Independent postmortem investigations have further demonstrated hypermethylation of BDNF promoter I and GRIN2B in the prefrontal cortex and hippocampus of patients with BD, with methylation levels correlating with illness duration.^[40,41]^ However, no animal model has yet replicated this BDNF/GRIN2B methylation signature in the bipolar-disorder brain by means of chronic high-cobalamin exposure. Future studies should therefore establish an integrated animal paradigm linking cobalamin excess, epigenetic modification, manic-like behavior to validate this mechanistic chain. In vitro studies have demonstrated that when the β-carotene concentration exceeds 1 μmol/L or when the partial pressure of oxygen is elevated, its antioxidant capacity declines and it undergoes auto-oxidation, generating lipid peroxyl radicals that propagate lipid peroxidation.^[42]^ Independent clinical investigations have reported significantly elevated lipid peroxidation products in the serum of BD patients during manic episodes.^[43]^ However, no study has directly attributed these products to β-carotene auto-oxidation. The potential mechanisms by which elevated vitamin B12 and carotene levels may increase the risk of BD remain to be further validated through experimental studies in the future.

This study makes several important contributions to the field of psychiatric epidemiology and nutritional neuroscience. First, it provides robust evidence supporting the causal role of carotene and vitamin B12 in BD risk, which could inform future dietary guidelines and interventions aimed at preventing BD. Second, the use of MR in this context exemplifies how genetic epidemiology can enhance our understanding of complex mental health disorders and their nutritional determinants. Lastly, by identifying specific micronutrients associated with BD, this research opens new avenues for exploring targeted nutritional strategies and personalized medicine approaches for individuals at risk for BD.

Despite its strengths, this study has several limitations. The MR approach, while powerful, relies on the availability and validity of genetic instruments. For some micronutrients, the genetic variants used as instruments may not capture all the variation in nutrient levels, potentially leading to biased estimates.^[44]^ Additionally, our study focused on individuals of European descent, which may limit the generalizability of the findings to other populations.^[45,46]^ Future research should aim to replicate these findings in diverse cohorts and explore the underlying biological mechanisms linking carotene and vitamin B12 to BD. Furthermore, the cross-sectional nature of the genetic data used in MR studies means that we cannot capture the dynamic changes in micronutrient levels over time and their interaction with other environmental and genetic factors. Longitudinal studies and RCTs investigating micronutrient supplementation in high-risk populations could provide more comprehensive insights into the preventive and therapeutic potential of these nutrients in BD.^[47]^

5. Conclusions

In conclusion, our MR study provides evidence for the causal role of carotene and vitamin B12 in increasing the risk of BD. These findings highlight the importance of adequate micronutrient intake for mental health and suggest potential targets for nutritional interventions. Further research is needed to confirm these associations and to understand the underlying mechanisms, paving the way for improved prevention and management strategies for BD.

Acknowledgments

We sincerely thank all the participants for their valuable contributions to this study, which utilized publicly available data from prior research studies.

Author contributions

**Data ** Ting You, Xiuwen Yang, Huaili Feng.

**Formal ** Ting You, Xiuwen Yang, Huaili Feng.

Investigation: Ting You, Huaili Feng.

**Writing – original ** Ting You, Xiuwen Yang.

**Writing – review & ** Zhaoyi Liu, Chengzhi Chen, Jingfu Qiu.

Causal relationship between circulating micronutrient levels and bipolar disorder: A Mendelian randomization study