Inverse associations between cord vitamin D and attention deficit hyperactivity disorder symptoms: A child cohort study

Abstract

Objective:

To examine the association between cord 25-hydroxyvitamin D₂₊₃ (25(OH)D) and attention deficit hyperactivity disorder symptoms in toddlers, using Child Behaviour Checklist for ages 1.5–5.

Method:

In a population-based birth cohort, a Child Behaviour Checklist for ages 1.5–5 questionnaire was returned from parents of 1233 infants with mean age 2.7 (standard deviation 0.6) years. Adjusted associations between cord 25(OH)D and Child Behaviour Checklist–based attention deficit hyperactivity disorder problems were analysed by multiple regression.

Results

The median cord 25(OH)D was 44.1 (range: 1.5–127.1) nmol/L. Mean attention deficit hyperactivity disorder problem score was 2.7 (standard deviation 2.1). In adjusted analyses, cord 25(OH)D levels >25 nmol/L and >30 nmol/L were associated with lower attention deficit hyperactivity disorder scores compared to levels ⩽25 nmol/L (p = 0.035) and ⩽30 nmol/L (p = 0.043), respectively. The adjusted odds of scoring above the 90th percentile on the Child Behaviour Checklist–based attention deficit hyperactivity disorder problem scale decreased by 11% per 10 nmol/L increase in cord 25(OH)D.

Conclusion:

An inverse association between cord 25(OH)D and attention deficit hyperactivity disorder symptoms in toddlers was found, suggesting a protective effect of prenatal vitamin D.

Keywords

Hyperactivity disorder vitamin D association Child Behaviour Checklist 1.5–5 cohort study

Introduction

Attention deficit hyperactivity disorder (ADHD), defined as functional impairment due to persistent attention deficits and hyperactivity, is one of the most common psychiatric disorders in childhood, with a reported prevalence of 3.4–7.1% (Polanczyk et al., 2015; Thomas et al., 2015). Despite the high prevalence and its extensive socioeconomic impact (RFF, 2014), knowledge of the specific causes of the disorder is limited. A multifactorial aetiology is suggested with a wide variety of environmental factors moderating a primary genetic susceptibility (Thapar et al., 2013). The disorder is diagnosed based on Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) or International Classification of Diseases, Tenth Edition (ICD-10) criteria after a thorough assessment in a multidisciplinary setting (Atkinson and Hollis, 2010). ADHD is most frequently diagnosed in boys of school age (Polanczyk et al., 2014). However, early signs of psychopathology, including ADHD symptoms, can be assessed by using psychometric instruments in younger children, e.g. The Child Behaviour Checklist 1.5–5 (CBCL 1.5–5; ASEBA, n.d.).

A high prevalence of vitamin D insufficiency and deficiency in pregnancy and neonates has been reported worldwide (Bodnar et al., 2007; Cadario et al., 2015; Pratumvinit et al., 2015; Rodriguez-Dehli et al., 2015; Vinkhuyzen et al., 2015). While the impact on bone growth and mineralization has been known for more than a century, more recent research has focused on extra-skeletal manifestations of hypovitaminosis D in children (Christesen et al., 2012; Wacker and Holick, 2013). Recent findings include positive associations between levels of maternal pregnancy serum 25-hydroxyvitamin D (s-25(OH)D) and psychomotor function and language development in infants and children (Morales et al., 2012; Whitehouse et al., 2012).

Hypovitaminosis D has been linked to ADHD in children (Bener and Kamal, 2014; Goksugur et al., 2014; Kamal et al., 2014; Shang-Guan and Zhao, 2015). The amount of research concerning maternal or neonatal vitamin D status and ADHD symptoms is, however, limited and contradictory. A recent Spanish cohort study found an inverse association between maternal 25(OH)D levels in early pregnancy and ADHD-like symptoms in 4-year-old children (Morales et al., 2015). In contrast, a Danish cohort and a Swedish case-control study found no association in older children (Gustafsson et al., 2015; Strom et al., 2014).

Neonatal vitamin D associations to ADHD symptoms in toddlers have not been studied, but may be anticipated if the foetal environment is critical for the development of early signs of ADHD. We, therefore, examined the association between cord 25(OH)D concentrations and parent-rated ADHD symptoms in 2- to 4-year-old children, using CBCL 1.5–5 in the Odense Child Cohort (OCC) study.

Method

Study population

The study population was drawn from the OCC, a population-based observational mother-and-child cohort, including newly discovered pregnancies in the Danish Municipality of Odense between 1 January 2010 and 31 December 2012 (Kyhl et al., 2015). Serial questionnaires and blood samples were obtained. Of 6707 pregnant women eligible for participation, 2549 (38%) enrolled in OCC (Figure 1).

Figure 1.

Flowchart.

Participants were included in this study if cord blood was donated with available s-25(OH)D measurements, and the parents had filled in a CBCL 1.5–5 questionnaire when the child was 2–4 years old (n = 1555). Exclusion criteria were multiple pregnancies, children with severe congenital malformations or syndromes and lack of information on covariates except alcohol consumption during pregnancy, maternal educational level and vitamin D supplementation. Parental psychiatric disorders were treated as a covariate.

Parameters studied

In short, following storage at −80°C, concentrations of s-25(OH)D₂₊₃ were analysed using liquid chromatography-tandem mass spectrometry on a TurboFlow column (Thermo Scientific San Jose, CA, USA), as previously described (Andersen et al., 2013). The sum of D2 and D3 was reported as total 25(OH)D.

The parent-report questionnaire CBCL 1.5–5 consists of 100 questions. All items are rated on a 3-point Likert scale: 0 (not true), 1 (somewhat/sometimes true) and 2 (very true/often true). The purpose is to measure emotional and behavioural problems, including ADHD symptoms, in children between 1.5 and 5 years. A standardized version of CBCL 1.5–5 translated into Danish is available (Kristensen et al., 2010). In this study, a DSM-oriented ADHD problem scale (CBCL 1.5–5 extracted) consisting of 6 questions with a maximum score of 12 (Table 1) was used as outcome. Good reliability and validity have been reported for the CBCL 1.5–5 in both clinical and research settings (Ivanova et al., 2010; Kristensen et al., 2010).

Table 1.

CBCL 1.5–5 ADHD problem items used.

		Not true	Somewhat/sometimes true	Very often/often true
5	Can’t concentrate, can’t pay attention for long	0	1	2
6	Can’t sit still, restless or hyperactive	0	1	2
8	Can’t stand waiting; wants everything now	0	1	2
16	Demands must be met immediately	0	1	2
36	Gets into everything	0	1	2
59	Quickly shifts from one activity to another	0	1	2
ADHD problem score		0–12

ADHD: attention deficit hyperactivity disorder; CBCL 1.5–5: Child Behaviour Checklist 1.5–5.

Covariates included a priori were maternal age, gestational age, sex of child, birth weight and season of year at birth. Parity, maternal smoking habits, alcohol consumption, educational level, vitamin D supplementation and pre-gestational height and weight for calculation of body mass index (BMI) were extracted from the OCC questionnaires as self-reported data. Parental psychiatric diagnoses were obtained from the Danish National Patient Register.

Statistical analyses

Descriptive data were presented as numbers (%), mean (standard deviation [SD]) or median (25–75% percentiles), and p-values were acquired with chi-square test or linear regression. Regression models were initially performed as linear, spline and cubic spline for categorical values.

Cord s-25(OH)D and ADHD scores were treated as both categorized and continuous variables. Cord s-25(OH)D was categorized into quartiles and routine cut-offs according to Institute of Medicine (IOM) (Ross et al., 2011) and Danish recommendations, i.e. s-25(OH)D <50 nmol/L, defined as insufficiency and s-25(OH)D <25 nmol/L defined as deficiency. ADHD score was split at the 90th percentile. Categorizing ADHD score in other percentiles, e.g. at the 97.5th percentile, led to too small numbers of children allowed for statistical analysis (possible type II error).

In the adjusted models, the initially included covariates were associated with either the ADHD score or cord s-25(OH)D. All self-reported data as well as parity, season of birth and parental psychiatric diagnoses were analysed as categorical variables. Maternal age and pre-pregnancy BMI were analysed as continuous variables. BMI was also categorized as <18.5, 18.5–24.9, 25–30 and >30 mg/m². Results were expressed as beta coefficients and odds ratios (OR) with 95% confidence intervals (CI) using linear and logistic regression, respectively. Missing data on covariates (maternal educational level, alcohol and vitamin D supplementation) were included in the statistical analysis as a missing categorical variable. A p-value < 0.05 was considered statistically significant, and p-values between 0.05 and 0.10 were considered trends.

A power analysis showed that our study would detect a true change of 0.02 ADHD scores per 10 nmol/L change in cord s-25(OH)D (α: 0.05, β = 0.8, 25(OH)D SD 22 nmol/L, ADHD score SD 0.6, n = 1200).

All data analyses were performed using STATA software, versions 13.1 and 14.0 (StataCorp, College Station, TX, USA).

Ethics

All women gave written informed consent to participate. The study complied with the Helsinki Declaration and was approved by the Regional Scientific Ethical Committee of Southern Denmark, no. S-20090130, and the Danish Data Protection Agency.

Results

The study included 1233 mothers and their children, of whom 654 (53.0%) were boys. The mean (SD) maternal age was 30.4 (4.4) years, and the mean child age when parents returned the CBCL questionnaires was 2.7 (0.6) years. The median cord s-25(OH)D concentration was 44.1 (25–75%: 30.4–59.3, range: 1.5–127.1) nmol/L. Mean (SD) CBCL-based ADHD problem score was 2.7 (2.1).

Table 2 presents the maternal, offspring and seasonal characteristics in relation to ADHD score, split by the 90th percentile. In univariate analysis, a score at or above the 90th percentile was significantly associated with lower cord s-25(OH)D concentrations, lower maternal age, primiparity, BMI <18.5 or ⩾30 kg/m², maternal smoking and alcohol consumption during pregnancy and lower maternal educational level. Trends were seen for parental psychiatric diagnoses, male sex and vitamin D supplementation.

Table 2.

Distribution of maternal, offspring and environmental factors in CBCL 1.5–5 scores.

Maternal, offspring and seasonal characteristics	Total n	Reported ADHD score (CBCL 1.5–5) in offspring		p-value
Maternal, offspring and seasonal characteristics		⩽90% percentile	>90% percentile	p-value
Cord 25(OH)D (nmol/L)	1233			0.025
0–24.9	212	181 (14.7)	31 (2.5)
25–49.9	533	473 (38.4)	60 (4.9)
⩾50	488	449 (36.4)	39 (3.2)
Maternal age (years)		30.6 ± 4.3	28.2 ± 3.8	<0.001
Parity				0.04
0	673	591 (47.9)	82 (6.7)
⩾1	560	512 (41.5)	48 (3.9)
Pre-gestational BMI (kg/m²)				0.027
<18.5	43	36 (2.9)	7 (0.6)
18.5–24.99	758	680 (55.2)	78 (6.3)
25.0–29.99	299	276 (22.4)	23 (1.9)
⩾30	133	111 (9.0)	22 (1.8)
Maternal smoking habits				0.007
Yes	50	39 (3.2)	11 (0.9)
No	1183	1064 (86.3)	119 (9.7)
Maternal alcohol consumption				0.027
Yes	116	103 (8.4)	13 (1.1)
No	839	763 (61.9)	76 (6.2)
Missing	278	237 (19.2)	41 (3.3)
Maternal education level^a				0.038
Lower	133	115 (9.3)	18 (1.5)
Intermediate	577	522 (42.3)	55 (4.5)
Higher	247	229 (18.6)	18 (1.5)
Missing	276	237 (19.2)	39 (3.2)
Parental psychiatric diagnosis				0.067
None	1035	934 (75.8)	101 (8.2)
Disposition from one parent	180	155 (12.6)	25 (2.0)
Disposition from both parents	18	14 (1.1)	4 (0.3)
Season of Birth				0.112
May until October	555	505 (41.0)	50 (4.1)
November until April	678	598 (48.5)	80 (6.5)
Sex				0.062
Male	654	575 (46.6)	79 (6.4)
Female	579	528 (42.8)	51 (4.1)
Children’s age^b		2.7 ± 0.6	2.6 ± 0.6	0.245
Vitamin D supplementation				0.057
⩽10 µg/day	105	99 (8.0)	6 (0.5)
>10 µg/day	625	565 (45.8)	60 (4.9)
Missing	503	439 (35.6)	64 (5.2)

BMI: body mass index; CBCL 1.5–5: Child Behaviour Checklist 1.5–5.

The values are numbers (%) or median ± SD. The p-values were calculated using Chi-square test or linear regression.

Lower: (high school or less). Intermediate: (high school + 1–3 years). Higher: (high school + > 4 years).

Age when CBCL 1.5–5 questionnaires were received.

We identified no continuous association between cord s-25(OH)D concentrations and ADHD score, unadjusted (p = 0.180), adjusted (p = 0.320), Table 3.

Table 3.

Associations between cord 25(OH)D and CBCL-based ADHD scores (n = 1233).

Cord 25(OH)D	Beta coefficients (95% CI)
Cord 25(OH)D	Unadjusted	p-value	Adjusteda	p-value
Continuous s-25(OH)D	−0.038 (0.094;0.017)	0.180	−0.031 (−0.092;−0.030)	0.320
⩽50 vs >50 nmol/L	−0.087 (−0.331;0.156)	0.482	−0.038 (−0.293;0.216)	0.768
⩽25 vs >25 nmol/L	−0.402 (−0.716;−0.087)	0.012	−0.352 (−0.678;−0.025)	0.035
⩽25 vs >50 nmol/L	−0.385 (−0.729;−0.042)	0.028	−0.321 (−0.686;0.043)	0.084
⩽30 vs >30 nmol/L	−0.321 (−0.596;−0.046)	0.022	−0.297 (−0.584;−0.01)	0.043
⩽30 vs >60 nmol/L	−0.218 (−0.557;0.122)	0.209	−0.200 (−0.566;0.166)	0.284

25(OH)D: 25-hydroxyvitamin D; CBCL: The Child Behaviour Checklist; CI: confidence interval; ADHD: attention deficit hyperactivity disorder.

The data were adjusted for maternal age, pre-gestational BMI, smoking status, alcohol consumption, educational level, parity, parental psychiatric disease, sex of child, child age and season of birth.

The crude association was not strengthened by spline (p = 0.277) or cubic spline (p = 0.184) regression. Split by routine cut-offs, cord s-25(OH)D >25 nmol/L had significantly lower ADHD scores compared to levels ⩽25 nmol/L, p = 0.035. Split by quartiles, cord s-25(OH)D >30 nmol/L had significantly lower ADHD scores compared to levels ⩽30 nmol/L, p = 0.043.

The adjusted odds of being in ⩾90th percentile decreased by 11% per 10 nmol/L increase in cord s-25(OH)D(p = 0.039), see Table 4. Split by routine cut-offs, the adjusted odds of an ADHD score ⩾90th percentile were decreased for cord s-25(OH)D >50 nmol/L compared to ⩽50 nmol/L (trend, p = 0.071) and >50 nmol/L compared to ⩽25 nmol/L (trend, p = 0.053). Split by quartiles, the adjusted odds of an ADHD score ⩾90th percentile were decreased for cord s-25(OH)D >30 nmol/L compared to ⩽30 nmol/L (trend, p = 0.065) and >60 nmol/L compared to ⩽30 nmol/L (trend, p = 0.062).

Table 4.

Associations between cord 25(OH)D and CBCL-based ADHD scores split by the 90th percentile (n = 1233).

Cord 25(OH)D	Beta coefficients (95% CI)
Cord 25(OH)D	Unadjusted	p-value	Adjusteda	p-value
Continuous s-25(OH)D	0.988 (0.079;0.997)	0.008	0.989 (0.979;0.999)	0.039
⩽50 vs >50 nmol/L	0.421 (0.421; 0.926)	0.035	0.677 (0.442;1.03)	0.071
⩽25 vs >25 nmol/L	0.627 (0.406; 0.967)	0.035	0.713 (0.442;1.148)	0.164
⩽25 vs >50 nmol/L	0.507 (0.307; 0.838)	0.008	0.575 (0.328;1.008)	0.053
⩽30 vs >30 nmol/L	0.612 (0.415;0.903)	0.013	0.666 (0.433;1.026)	0.065
⩽30 vs >60 nmol/L	0.547 (0.326;0.920)	0.023	0.572 (0.319;1.028)	0.062

CBCL: The Child Behaviour Checklist; CI: confidence interval; ADHD: attention deficit hyperactivity disorder.

In addition, lower maternal age (p = 0.004), psychiatric disease in parents (p = 0.014), maternal smoking during pregnancy (p = 0.047) and male sex (p = 0.006) were independently associated with increasing ADHD score.

The adjusted odds of being in the top 10 percentile of the ADHD scale were associated with lower maternal age (p = 0.002), maternal smoking during pregnancy (trend, p = 0.061), male sex (trend, p = 0.072) and missing alcohol data (trend p = 0.064).

We performed a post hoc analysis without adjustment for ‘season of birth’ in the model as it has been suggested that season of birth may not have impact on an eventual brain development during pregnancy. However, this did not change the results in the linear models, i.e. no significant association between continuous s-25(OH)D and ADHD score (β-coefficient [95% CI]: −0.047 [−0.10;0.010], p = 0.109) but strengthened the association found in the logistic models, the odds of being in the ⩾90th percentile decreased 14% per 10 nmol/L increase in cord s-25(OH)D (p = 0.004).

Discussion

This prospective cohort study of a population-based Danish birth cohort showed novel inverse associations between cord s-25(OH)D concentrations and lower parent-rated ADHD problems in toddlers. Other variables showed significant associations with higher ADHD problem scores, including lower maternal age, male sex, smoking during pregnancy and parental psychiatric disorders. These variables have previously been associated with clinically assessed ADHD (Gustafsson et al., 2015; Thapar and Cooper, 2015), supporting the hypothesis that toddlers with high CBCL 1.5–5-based ADHD scores actually represent an ADHD high-risk group.

The association between early s-25(OH)D and ADHD symptoms has only been studied sparsely, and the association between cord s-25(OH)D and CBCL-based ADHD problem scores has, to our knowledge, not been studied previously. Discrepancies in previous studies may, in addition to methodological issues and age of the children, be a matter of timing of blood sampling in relation to a possible time window of the impact of the exposure on the developing brain and differences in cut-offs for s-25(OH)D or ADHD evaluation scores.

Our data did not support a linear correlation throughout the whole range of exposure, and outcome parallels Strom et al. (2014), who found no association between maternal pregnancy s-25(OH)D and prescription of ADHD medication in late adolescents (n = 850, ADHD, n = 24), and, Gustafsson et al. (2015), who also found no association between low cord s-25(OH)D and ADHD in children between 5 and 17 years of age (n = 404). Whitehouse et al. (2012) found no association between maternal s-25(OH)D in pregnancy and behaviour in 2-year-old children when using CBCL 2–3 (n = 743 children).

In contrast, one large study has supported an association between low s-25(OH)D and ADHD-like symptoms. Morales et al. (2015) found that the adjusted risk of ADHD symptoms in children aged 4.8 years (n = 1650) decreased by 11% per 25 nmol/L increase in maternal pregnancy s-25(OH)D. However, Keim et al. (2014) found no association between maternal or cord s-25(OH)D and behaviour in children aged between 8 months and 7 years, including hyperactive behaviours in 4-year-old children (n = 3896).

The study by Morales et al. (2015) is the study most comparable to our study regarding age of assessment and sample size and it supports our findings. For other studies, methodical differences, most importantly ADHD assessment tools and age of evaluation, hamper direct comparison with our results.

Existing research suggests that vitamin D is crucial to the foetal neuroanatomical development through a variety of pathways, including neuronal differentiation and regulation of reactive oxidative species (Eyles et al., 2011). However, the reason why higher vitamin D concentrations potentially protect against ADHD symptoms remains uncertain. The preventing component could be that vitamin D, through its anti-inflammatory abilities, inhibits harmful environmental triggers related to potential neurodevelopmental anomalies, as discussed by Morales et al. (Morales et al., 2015). Moreover, vitamin D deficiency in pregnant mice has been linked to increased glucocorticoid exposure in the foetus (Tesic et al., 2015), which may in turn, through increased foetal brain exposure, lead to neurodevelopmental anomalies. Sufficient vitamin D concentrations may protect the foetal brain against excessive glucocorticoid exposure, thus protecting against neurodevelopmental disorders like ADHD.

The major strengths of this study include the relatively large population-based and prospective study design and use of the gold standard method for analysis of s-25(OH)D (Farrell et al., 2012). Furthermore, the cord vitamin D analysis and the CBCL assessment process were blinded from each other, and the statistical analyses performed were thorough, including search for optimal cut-offs and inclusion of missing data.

The limitations include the observational nature of the study, allowing no causative conclusions; minor discrepancies between cohort participants and the background population and the use of only parents as informers of ADHD symptomatology. We have no information on mother–child interactions, which could have led to confounding due to possible mediation from, e.g. daily activities and sun exposure.

However, most Danish children in this age group are in kindergartens with planned daily out- and indoor activities.

The Danish population is rather homogenous and so is this cohort, with only 3.5% being non-Caucasians. This group has lower vitamin D levels, but due to its small size, it did not alter the outcomes significantly (results not shown). Furthermore, the young age of assessment in this study may lead to higher CBCL-scores due to possible natural over-activity or restless children. Last, only 18% of the eligible participants were included which could have an impact on the generalizability.

Conclusion

In conclusion, associations were detected between higher levels of cord s-25(OH)D and lower ADHD problem score in children aged on average 2.7 years, suggesting that higher cord s-25(OH)D levels may protect against early ADHD symptoms in toddlers. The key findings in this study were summarized in Figure 2. The study further underlines the multifactorial aetiology of the disorder, identifying other independently associated factors such as low maternal age and parental psychiatric diagnoses. A follow-up at school age will be necessary to further validate the possible psychopathologic impact of vitamin D deficiency in late pregnancy and the development of ADHD.

Figure 2.

Key findings.

Clinical implications

Based on our results, we suggest that women use vitamin D supplementation during pregnancy, as it seems likely that it may reduce the risk of developing ADHD symptoms in toddlers.

Footnotes

Acknowledgements

The authors thank the participants in the cohort study, the staff at Odense Child Cohort and our collaborators at the Department of Clinical Biochemistry and Pharmacology at Odense University Hospital.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

This work was funded by The Region of Southern Denmark, the University of Southern Denmark, Aase & Ejner Danielsen’s Foundation, the King Christian IX and Queen Louise’s Jubilee Grant, the A.P. Møller Foundation for the Advancement of Medical Science, Torben and Alice Frimodt’s Foundation, Odense University Hospital, The National Board of Social Services and Mental Health Research Fund of Southern Denmark.

References

Andersen

Abrahamsen

Dalgard

. (2013) Parity and tanned white skin as novel predictors of vitamin D status in early pregnancy: A population-based cohort study. Clinical Endocrinology 79: 333–341.

ASEBA (n.d.). Preschool (Ages 1½-5) Assessments. Available at: www.aseba.org/preschool.html

Atkinson

Hollis

(2010) NICE guideline: Attention deficit hyperactivity disorder. Archives of Disease in Childhood: Education and Practice 95: 24–27.

Bener

Kamal

(2014) Predict attention deficit hyperactivity disorder? Evidence -based medicine. Global Journal of Health Science 6: 47–57.

Bodnar

Simhan

Powers

. (2007) High prevalence of vitamin D insufficiency in black and white pregnant women residing in the northern United States and their neonates. Journal of Nutrition 137: 447–452.

Cadario

Savastio

Magnani

. (2015) High prevalence of vitamin D deficiency in native versus migrant mothers and newborns in the north of Italy: A call to act with a stronger prevention program. PLoS ONE 10: e0129586.

Christesen

Elvander

Lamont

. (2012) The impact of vitamin D in pregnancy on extraskeletal health in children: A systematic review. Acta Obstetricia et Gynecologica Scandinavica 91: 1368–1380.

Eyles

Burne

McGrath

(2011) Vitamin D in fetal brain development. Seminars in Cell & Developmental Biology 22: 629–636.

Farrell

Martin

McWhinney

. (2012) State-of-the-art vitamin D assays: A comparison of automated immunoassays with liquid chromatography-tandem mass spectrometry methods. Clinical Chemistry 58: 531–542.

10.

Goksugur

Tufan

Semiz

. (2014) Vitamin D status in children with attention-deficit-hyperactivity disorder. Pediatrics International 56: 515–519.

11.

Gustafsson

Rylander

Lindh

. (2015) Vitamin D status at birth and future risk of attention deficit/hyperactivity disorder (ADHD). PLoS ONE 10: e0140164.

12.

Ivanova

Achenbach

Rescorla

. (2010) Preschool psychopathology reported by parents in 23 societies: Testing the seven-syndrome model of the child behavior checklist for ages 1.5–5. Journal of the American Academy of Child and Adolescent Psychiatry 49: 1215–1224.

13.

Kamal

Bener

Ehlayel

(2014) Is high prevalence of vitamin D deficiency a correlate for attention deficit hyperactivity disorder? Attention Deficit Hyperactivity Disorder 6: 73–78.

14.

Keim

Bodnar

Klebanoff

(2014) Maternal and cord blood 25(OH)-vitamin D concentrations in relation to child development and behaviour. Paediatric and Perinatal Epidemiology 28: 434–444.

15.

Kristensen

Henriksen

Bilenberg

(2010) The child behavior checklist for ages 1.5–5 (CBCL/1(1/2)-5): Assessment and analysis of parent- and caregiver-reported problems in a population-based sample of Danish preschool children. Nordic Journal of Psychiatry 64: 203–209.

16.

Kyhl

Jensen

Barington

. (2015) The Odense child cohort: Aims, design, and cohort profile. Paediatric and Perinatal Epidemiology 29: 250–258.

17.

Morales

Guxens

Llop

. (2012) Circulating 25-hydroxyvitamin D3 in pregnancy and infant neuropsychological development. Pediatrics 130: e913–920.

18.

Morales

Julvez

Torrent

. (2015) Vitamin D in pregnancy and attention deficit hyperactivity disorder-like symptoms in childhood. Epidemiology 26: 458–465.

19.

Polanczyk

Salum

Sugaya

. (2015) Annual research review: A meta-analysis of the worldwide prevalence of mental disorders in children and adolescents. Journal of Child Psychology and Psychiatry 56: 345–365.

20.

Polanczyk

Willcutt

Salum

. (2014) ADHD prevalence estimates across three decades: An updated systematic review and meta-regression analysis. International Journal of Epidemiology 43: 434–442.

21.

Pratumvinit

Wongkrajang

Wataganara

. (2015) Maternal vitamin D status and its related factors in pregnant women in Bangkok, Thailand. PLoS ONE 10: E0131126.

22.

RFF (2014) ADHD koster ubehandlet knap 3 mia. kroner om året. Available at: http://da.rff.rw-fonden.inforce.dk/files/RFF-site/Publikations%20upload/Newsletters/Dansk/Nyhedsbrev%20september%202014%20ny.pdf

23.

Rodriguez-Dehli

Riano Galan

Fernandez-Somoano

. (2015) Prevalence of vitamin d deficiency and insuffiency and associated factors in pregnant women of northern Spain. Nutrición Hospitalaria 31: 1633–1640.

24.

Ross

Manson

Abrams

. (2011) The 2011 report on dietary reference intakes for calcium and vitamin D from the Institute of Medicine: What clinicians need to know. The Journal of Clinical Endocrinology & Metabolism 96: 53–58.

25.

Shang-Guan

Zhao

(2015) Serum levels of 25-hydroxyvitamin D in children with attention deficit hyperactivity disorder. Zhongguo Dang Dai Er Ke Za Zhi 17: 837–840.

26.

Strom

Halldorsson

Hansen

. (2014) Vitamin D measured in maternal serum and offspring neurodevelopmental outcomes: A prospective study with long-term follow-up. Annals of Nutrition and Metabolism 64: 254–261.

27.

Tesic

Hawes

Zosky

. (2015) Vitamin D deficiency in BALB/c mouse pregnancy increases placental transfer of glucocorticoids. Endocrinology 156: 3673–3679.

28.

Thapar

Cooper

(2015) Attention deficit hyperactivity disorder. The Lancet. Epub ahead of print 16 September. DOI: 10.1016/S0140-6736(15)00238-X.

29.

Thapar

Cooper

Eyre

. (2013) What have we learnt about the causes of ADHD? Journal of Child Psychology and Psychiatry 54: 3–16.

30.

Thomas

Sanders

Doust

. (2015) Prevalence of attention-deficit/hyperactivity disorder: A systematic review and meta-analysis. Pediatrics 135: e994–1001.

31.

Vinkhuyzen

Eyles

Burne

. (2015) Prevalence and predictors of vitamin D deficiency based on maternal mid-gestation and neonatal cord bloods: The generation R study. The Journal of Steroid Biochemistry and Molecular Biology. Epub ahead of print 15 September. DOI: 10.1016/j.jsbmb.2015.09.018.

32.

Wacker

Holick

(2013) Vitamin D – Effects on skeletal and extraskeletal health and the need for supplementation. Nutrients 5: 111–148.

33.

Whitehouse

Holt

Serralha

. (2012) Maternal serum vitamin D levels during pregnancy and offspring neurocognitive development. Pediatrics 129: 485–493.