Abstract
Zinc regulates dopaminergic signaling, and reduced serum zinc levels have been reported in individuals with ADHD. However, genetic associations between zinc and ADHD remain unclear. We examined this link using large-scale GWAS and molecular analyses across three cohorts: iPSYCH (14,584 ADHD cases and 22,492 controls), FAMHES (n = 1798), and the Hamamatsu Birth Cohort (n = 726). Two-sample Mendelian randomization revealed bidirectional associations between low serum zinc levels and ADHD diagnosis. Genetic correlation and polygenic risk score analyses supported this association. In the birth cohort, lower cord blood zinc were associated with higher ADHD symptom scores at ages 8–9. Zinc levels negatively correlated with IL-6 and maternal depressive symptoms. Directed acyclic graph analysis indicated that maternal stress increased IL-6, which reduced fetal zinc levels, linking to ADHD symptoms. These findings suggest low prenatal zinc may contribute to ADHD pathophysiology in genetically vulnerable children, potentially mediated by maternal stress and inflammation.
Subject terms: Genetic association study, Neurodevelopmental disorders, ADHD, Risk factors
Introduction
Attention Deficit Hyperactivity Disorder (ADHD) is the most prevalent neurodevelopmental disorder, beginning in childhood and often persisting across the lifespan. Both genetic and environmental factors are implicated in the disorder. Among several biomarkers1, reduced serum zinc levels have been repeatedly reported in subjects with ADHD2. Although the findings are not consistent across studies, a recent meta-analysis attributes this inconsistency to methodological issues such as small sample sizes, different measurements, and study designs3. ADHD is also associated with sleep disorders and altered melatonin production4. Zinc is known to regulate dopaminergic activity via its role in the production and metabolism of melatonin5, suggesting that zinc could be a potential target biomarker underlying the neurobiological mechanisms of ADHD. Despite its potential importance, to date, no genetic correlation or association between serum zinc levels and ADHD diagnosis has been explored.
Importantly, neuroinflammation has recently been proposed as contributing to the etiology of ADHD. In our previous study, we demonstrated that increased cytokines, such as IL-6, in cord bloods are associated with increased ADHD symptoms later in childhood6. Furthermore, elevated levels of pro-inflammatory cytokines are reported to reduce serum zinc7. Taken together, we may hypothesize that reduced levels of zinc may be involved in the development of ADHD in children in conjunction with neuroinflammation.
Perinatal stress is known to be an important mediator of child development8, including ADHD9. It is also considered to be a cause of perinatal neuroinflammation, such as increased levels of IL-6 in the brains of offspring10. Taken together, these data raise the possibility that zinc could have a role in the development of ADHD [in childhood] through its association with the inflammation that accompanies maternal perinatal stress.
Using multiple genetic and molecular biological analyses, this study aimed to investigate the relationship between serum zinc and ADHD in the context of neuroinflammation associated with maternal perinatal stress.
Methods
Subjects
Data from the Lundbeck Foundation Initiative for Integrative Psychiatric Research (iPSYCH) and the Psychiatric Genomics Consortium, the Queensland Institute of Medical Research (QIMR), the Fangchenggang Area Male Health Examination Survey (FAMHES) and the Hamamatsu Birth Cohort for Mothers and Children (HBC study) were analyzed.
Mendelian randomization analyses
Two-sample Mendelian Randomization (MR) analyses with regression analyses were conducted using iPsych and QIMR with standard variant harmonization procedures (https://mrcieu.github.io/TwoSampleMR/). We used a P value threshold of 5 × 10−8 to select single nucleotide polymorphisms (SNPs) associated with serum zinc levels and ADHD diagnosis. For the main analysis we used MR Eggar and Inverse Variance Weighted (IVW) per exposure-outcome. Potential horizontal pleiotropy was evaluated using Cochrane’s Q statistic test. Further details are provided in the Supplementary Note 1.
Genetic correlation analysis
Genetic correlations (rg) between serum levels of zinc and ADHD were calculated by LD (linkage disequilibrium) score regression using iPsych and FAMHES. Since the difference in the ethnic population between serum zinc and ADHD genome-wide association study (GWAS), we utilized the “Popcorn” program, which can calculate trans-ethnic genetic correlation.
Polygenic risk score analysis
Polygenic risk score (PRS) for subjects in HBC study was generated by PRS-CSx11 using a iPSYCH ADHD GWAS study12 for ADHD-PRS and FAMHES zinc GWAS study13 for zinc-PRS. Genome-wide genotyping of subjects in HBC study was performed using the Japonica Array 2.0 platform. Detailed information about quality controls and imputation of HBC genotypng data were described in the Supplementary Note 1. The threshold for selecting SNPs was set at a P-value of 0.05. To account for population stratification, 4 principal components (PCs) calculated with PLINK 1.9 were used. The criteria for SNP clumping were r2 > 0.1 within a 2 Mb window. PRS scores were calculated with P-value thresholds at 0.05. Standardized PRS scores (mean = 0 and standard deviation = 1) were used for the analyses. Further details are provided in the Supplementary Methods section.
Molecular analysis
Molecular analyses were performed to investigate the association between cord blood zinc levels and ADHD symptoms as assessed by the ADHD-RS in children aged 8-9 years. Serum zinc and IL-6 levels were measured by enzyme-linked immunosorbent assays (ELISA) according to the manufacturer’s protocol. Maternal stress was measured using the Edinburgh Postnatal Depression Scale (EPDS) two weeks after delivery. Analyses were limited to this subsample from the larger HBC cohort, in whom all variables of interest were available. 726 participants included in the molecular analyses had both cord blood samples (zinc and IL-6) and genome-wide genotyping data, ADHD symptom scores assessed at 8–9 years of age as well as maternal EPDS scores.
Statistical analysis
All analyses were conducted using Stata 16 software (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC). Directed Acyclic Graph (DAG) analysis was used to account for the relationship between variables. Statistical significance was set at P < 0.05, and all tests were 2-tailed.
Ethical considerations
The study protocol was approved by the Hamamatsu University School of Medicine and University Hospital Ethics Committee (Ref No. 20-233). Written informed consent was obtained from each caregiver (usually the mother) for her own and her children’s participation. This study followed the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) reporting guideline and STROBE-MR guideline.
Results
Characteristics of subjects
14,584 individuals with ADHD and 22,492 controls were gathered from iPSYCH and PGC data. GWAS data for serum zinc were obtained from the 2603 individuals in the QIMR cohort and the 1798 individuals in the FAMHES cohort. A total of 726 participants (373 males, 353 females) with both genotyping data and ADHD symptoms, measured using the parent form of ADHD-RS scale at 8-9 years old, cord blood zinc and iL-6 were included in the analysis. The study population used for each analysis is summarized in Table 1 and detailed sample characteristics of HBC study are described in the Supplementary Table 1.
Table 1.
Summary of the cohorts used in the analyses
| Analysis method | Cohort1 | Outcome | Cohort2 | Outcome |
|---|---|---|---|---|
| Mendelian Randomization | QIMR, N = 2603 | Serum zinc | iPSYCH (N = 37526) | ADHD diagnosis |
| Genetic correlation | FAMHES, N = 1798 | Serum zinc | iPSYCH (N = 37526) | ADHD diagnosis |
| Polygenic Risk Score analysis | iPSYCH (N = 37526) | ADHD diagnosis | HBC (N = 726) | ADHD symptoms |
| FAMHES, N = 1798 | Serum zinc | HBC (N = 726) | ADHD symptoms |
QIMR Queensland Institute of Medical Research, iPSCH Integrative Psychiatric Research, FAMHES the Fangchenggang Area Male Health Examination Survey, ADHD attention deficit hyperactivity disorder, HBC Hamamatsu Birth Cohort for Mothers and Children.
Bidirectional association between zinc and ADHD
A bidirectional association was observed between low serum zinc levels and ADHD diagnosis. In the analysis of serum zinc to ADHD diagnosis, the association was significant (MR-Eggar beta [SE] = −0.049[0.009], P = 2.462 × 10−8 and IVW beta [SE] = −0.018[0.035], P = 2.405 × 10−7). The association of ADHD diagnosis to serum zinc was also significant (MR-Eggar beta [SE] = −0.965[0.238], P = 8.839 × 10−5, IVW beta [SE] = −0.093[0.003], P = 7.619 × 10−4).
Genetic correlation and PRS analyses
Genetic correlation analysis showed that serum zinc levels and ADHD diagnosis are genetically correlated (rg[SE] = −0.218 [0.004], P < 0.001). PRS analysis showed that Zinc-PRS was negatively associated with serum zinc (beta [SE] = −0.841[0.011], P < 0.001) and ADHD symptoms in children (beta [SE] = −0.484[0.033], P < 0.001). ADHD-PRS was negatively associated with serum zinc (beta [SE] = −2.036[0.796], P = 0.011) and ADHD symptoms in children as previously reported (beta [SE] = −0.907 [0.005], P < 0.001)14.
Cord Blood Zinc and ADHD Symptoms
Serum zinc levels were negatively associated with ADHD symptoms in children at age 8–9 (r[SE] = −0.101[0.038], P = 0.004). Serum zinc levels were also negatively correlated with IL-6 levels (r[SE] = −0.292[0.112], P < 0.001) (Figs. 1, 2). To address the potential influence of outliers, we conducted a sensitivity analysis excluding values above the 95th percentile for IL-6. The negative association between IL-6 and cord blood zinc remained statistically significant (β = −2.99, SE = 1.22, p = 0.015), indicating the robustness of the finding. We also conducted a regression analysis adjusting for maternal age, education level, and prenatal antidepressant medication use. The association between cord blood zinc levels and ADHD-RS scores remained significant (β = −0.102, SE = 0.038, p = 0.008), further supporting the robustness of the finding.
Fig. 1. Correlation analysis between serum zinc and IL-6 in cord blood.
Red line indicates best fitted line in regression analysis.
Fig. 2. Directed acyclic graph between mother’s perinatal stress, serum zinc and A DHD symptoms.
Numbers represents standardized beta by regression analysis. Numbers in [] represents standard error. *P < 0.05, **P < 0.01, ***P < 0.001.
Maternal Stress, Inflammation, and Zinc Levels
Mothers’ perinatal stress level as measured by the total EPDS score were negatively correlated with serum zinc levels (r[SE] = −0.132[0.056], P = 0.004). DAG analysis showed that maternal stress affects serum zinc levels via modulation of IL-6 levels. Reduced zinc levels, in turn, were associated with increased ADHD symptoms in children aged 8–9 (Fig. 2).
Discussion
The present study showed that low serum zinc levels in cord blood are genetically and biologically associated with ADHD diagnosis and symptoms at age 8, possibly induced by maternal perinatal stress. Furthermore, our data proposes that IL-6 is one of the key molecules to connect maternal stress and zinc in the mechanism of ADHD pathophysiology. While zinc is known to influence dopaminergic signaling pathways, our genetic analyses did not specifically focus on dopamine-related SNPs. Instead, we employed genome-wide polygenic risk scores and Mendelian randomization based on large-scale GWAS summary statistics for ADHD and serum zinc. Multiple genetic analyses indicate a bidirectional relationship between genetic risk for ADHD and serum zinc levels. These findings suggest that lower serum zinc levels in cord blood may act as a biological vulnerability factor for the development of ADHD symptoms in genetically susceptible individuals. Although this raises the possibility that prenatal zinc supplementation could be beneficial, interventional studies are needed to test this hypothesis15.
A few limitations should be considered in this study. First, we used the EPDS score two weeks after delivery as a proxy for perinatal stress among pregnant individuals. Different measures of stress could have strengthened the validity of the construct. Second, although our model is statistically robust, the biological validity should be confirmed experimentally. The present findings should be replicated in independent cohorts and the benefit of zinc supplements should be examined in biological model systems (e.g., maternal-immune-activation models). Third, although we restricted our analyses to participants with complete data on cord blood zinc and IL-6 levels, ADHD symptoms, and genome-wide genotyping (n = 726), this subset may not be fully representative of the entire cohort. Therefore, selection bias cannot be entirely ruled out. Finally, the PRS construction was based on GWAS summary statistics from different ethnic backgrounds (European and Chinese), and while we used trans-ethnic methods (PRS-CSx), the generalizability to Japanese populations may still be limited.
Supplementary information
Acknowledgements
This work was supported by grants from the Ministry of Education, Culture, Sports, Science, and Technology in Japan (grant number 25K10877 to NT). The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The authors are grateful to all the study participants. The authors thank Dr. Tetsuo Kato of the Kato Maternity Clinic, Drs. K Sugihara, M Sugimura, K Takeuchi, K Suzuki, Y Murakami, Y Kohmura, Y Miyabe, K Hirai, Y Nakamura, R Koizumi, H Murakami, Y Kobayashi-Kohmura, K Muramatsu-Kato, Prof. N Kanayama, and all attending obstetricians of Hamamatsu University School of Medicine for their full support with participant enrollment. The authors also thank the chief midwife, Ms. Kiyomi Hinoki, and all midwives and staff at the maternity clinic of Hamamatsu University School of Medicine for their support with participant enrollment. We thank the HBC Study team, including former members, Ms. Emi Higashimoto, Noriko Kodera, Atsuko Nakamura, Yumeno Kugizaki, Yukiko Suzuki, Riyo Takabayashi, Maiko Honda, Hiroko Muraki, Makiko Narumiya, Eriko Sato, Drs. Yuka Omori, Hanae Tainaka, Asako Tokizawa, Damee Choi, Takanobu Horikoshi, Keisuke Wakusawa, Yosuke Kameno, Daisuke Kurita, Kiyokazu Takebayashi, Masamichi Yokokura, Tomoyasu Wakuda, Ryosuke Asano, Thanseem Ismail, Keiko Iwata, Yasuhide Iwata, Emiko Kawai, Masayoshi Kawai, Yujin K Kuroda, Kaori Matsumoto, Hideo Matsuzaki, Tsuruko Mori, Kyoko Nakaizumi, Kazuhiko Nakamura, Anitha A Pillai, Yui Seno, Chie Shimmura, Shiro Suda, Genichi Sugihara, Katsuaki Suzuki, Kohei Yamada, Shigeyuki Yamamoto, Yujiro Yoshihara, Yusaku Endoh, Koichi Hirano, Teruhiko Suzuki, Professor Norio Mori, Professor Nori Takei, Professor Hitoshi Kuwabara, Professor Hidenori Yamasue, Professor Masatsugu Tsujii, and Professor Toshirou Sugiyama, for data collection and management, and administrative and academic supports. The authors are also grateful to Mr. Charles Davey for their professional editing support in English.
Author contributions
N.T. and K.J.T. conceived and designed the study. N.T., A.K., T.H., T.I., T.N. acquired, analyzed and interpreted the data. N.T. and K.J.T. drafted the manuscript. All authors critically reviewed and revised the manuscript. N.T. and R.S. performed the statistical analysis.
Data availability
The data generated for this study are subject to the following licenses/restrictions: Privacy and Confidentiality of Participants. Requests to access these datasets should be directed to N.T., nagahide.takahashi@ncnp.go.jp.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
The online version contains supplementary material available at 10.1038/s44184-025-00149-3.
References
- 1.Takahashi, N., Ishizuka, K. & Inada, T. Peripheral biomarkers of attention-deficit hyperactivity disorder: current status and future perspective. J. Psychiatr. Res.137, 465–470 (2021). [DOI] [PubMed] [Google Scholar]
- 2.Villagomez, A. & Ramtekkar, U. I. ron Magnesium, vitamin D, and zinc deficiencies in children presenting with symptoms of attention-deficit/hyperactivity disorder. Children (Basel)1, 261–279 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ghoreishy, S. M., Ebrahimi Mousavi, S., Asoudeh, F. & Mohammadi, H. Zinc status in attention-deficit/hyperactivity disorder: a systematic review and meta-analysis of observational studies. Sci. Rep.11, 14612 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Takahashi, N. et al. Association between genetic risk of melatonin secretion and attention deficit hyperactivity disorder. Psychiatry Res. Commun.4, 100188 (2024). [Google Scholar]
- 5.Chen, M. D., Lin, P. Y. & Sheu, W. H. Zinc coadministration attenuates melatonin’s effect on nitric oxide production in mice. Biol. Trace Elem. Res.69, 261–268 (1999). [DOI] [PubMed] [Google Scholar]
- 6.Takahashi, N. et al. Interaction of genetic liability for attention deficit hyperactivity disorder (ADHD) and perinatal inflammation contributes to ADHD symptoms in children. Brain Behav. Immun. Health30, 100630 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Rink, L. & Gabriel, P. Zinc and the immune system. Proc. Nutr. Soc.59, 541–552 (2000). [DOI] [PubMed] [Google Scholar]
- 8.Tainaka, H. et al. Long-term effect of persistent postpartum depression on children’s psychological problems in childhood. J. Affect Disord.305, 71–76 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sujan, A. C., Oberg, A. S., Quinn, P. D. & D’Onofrio, B. M. Annual Research Review: Maternal antidepressant use during pregnancy and offspring neurodevelopmental problems—a critical review and recommendations for future research. J. Child Psychol. Psychiatry60, 356–376 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Usui, N. et al. Social communication of maternal immune activation-affected offspring is improved by Si-based hydrogen-producing agent. Front Psychiatry13, 872302 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ruan, Y. et al. Improving polygenic prediction in ancestrally diverse populations. Nat. Genet.54, 573–580 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Demontis, D. et al. Discovery of the first genome-wide significant risk loci for attention deficit/hyperactivity disorder. Nat. Genet.51, 63–75 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hemani, G. et al. The MR-Base platform supports systematic causal inference across the human phenome. eLife7, e34408 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Takahashi, N. et al. Exploration of Sleep Parameters, Daytime Hyperactivity/Inattention, and Attention-Deficit/Hyperactivity Disorder Polygenic Risk Scores of Children in a Birth Cohort in Japan. JAMA Netw. Open5, e2141768 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bilici, M. et al. Double-blind, placebo-controlled study of zinc sulfate in the treatment of attention deficit hyperactivity disorder. Prog. Neuro-Psychopharmacol. Biol. Psychiatry28, 181–190 (2004). [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data generated for this study are subject to the following licenses/restrictions: Privacy and Confidentiality of Participants. Requests to access these datasets should be directed to N.T., nagahide.takahashi@ncnp.go.jp.