QuestionDoes higher daily vitamin D₃ supplementation up to age 2 years decrease the risk of psychiatric symptoms at ages 6 to 8 years compared with the standard recommended dose?

FindingsIn this secondary analysis of a randomized clinical trial including 346 children, those randomized to higher vitamin D₃ supplementation were less likely to have clinically significant internalizing problems.

MeaningThis study found that higher than standard vitamin D₃ supplementation up to age 2 years decreased the risk for internalizing problems in later childhood.

ImportanceVitamin D is associated with neurodevelopment, but causality, critical windows, and potentials for modification remain unknown.

ObjectiveTo determine the impact of high-dose (1200 IU) vs standard-dose (400 IU) vitamin D₃ supplementation during the first 2 years on psychiatric symptoms at ages 6 to 8 years and whether the impact is different in children with lower vs higher maternal vitamin D₃ levels; lower vs higher levels were defined as 25-hydroxyvitamin D (25[OH]D) less than 30 ng/mL vs 30 ng/mL or greater.

Design, Setting, and ParticipantsThis study was a long-term follow-up of the double-blind randomized clinical trial (RCT) Vitamin D Intervention in Infants (VIDI) conducted at a single center in Helsinki, Finland, at 60 degrees north latitude. Recruitment for VIDI took place in 2013 to 2014. Follow-up data for secondary data analysis were collected 2020 to 2021. VIDI originally included 987 term-born infants; 546 of these individuals participated in the follow-up at ages 6 to 8 years, among whom 346 individuals had data on parent-reported psychiatric symptoms. Data were analyzed from June 2022 to March 2023.

InterventionsThere were 169 infants randomized to receive 400-IU and 177 infants randomized to receive 1200-IU oral vitamin D₃ supplementation daily from ages 2 weeks to 24 months.

Main Outcomes and MeasuresPrimary outcomes were internalizing, externalizing, and total problems scores, with clinically significant problems defined as T scores of 64 or greater in the Child Behavior Checklist questionnaire.

ResultsAmong 346 participants (164 females [47.4%]; mean [SD] age, 7.1 [0.4] years), the vitamin D₃ dose was 400 IU for 169 participants and 1200 IU for 177 participants. Clinically significant internalizing problems occurred in 10 participants in the 1200-IU group (5.6% prevalence) compared with 20 participants (11.8%) in the 400-IU group (odds ratio, 0.40; 95% CI, 0.17-0.94; P = .04) after adjustment for sex, birth season, maternal depressive symptoms at birth, and parental single status at follow-up. In a post hoc subgroup analysis, 48 children in the 400-IU group with maternal 25(OH)D concentrations less than 30 ng/mL had higher internalizing problems scores compared with children in the 1200-IU group, including 44 children with maternal 25(OH)D concentrations below 30 ng/mL (adjusted mean difference, 0.49; 95% CI, 0.09-0.89; P = .02) and 91 children with maternal concentrations above 30 ng/mL (adjusted mean difference, 0.37; 95% CI, 0.03-0.72; P = .04). Groups did not differ in externalizing or total problems.

Conclusions and RelevanceThis randomized clinical trial found that higher-than-standard vitamin D₃ supplementation in the first 2 years decreased risk of internalizing problems at ages 6 to 8 years.

Trial RegistrationClinicalTrials.gov Identifiers: NCT01723852 (VIDI) and NCT04302987 (VIDI2)

Vitamin D insufficiency and deficiency are estimated to occur among almost half and more than one-tenth of the global population, respectively, across all ages.^1,2 In addition to its well-known role in skeletal health, vitamin D also plays a role in neurodevelopment. Receptors and metabolizing enzymes for vitamin D are present in various areas of the human brain,³ and experimental animal studies have linked vitamin D deficiency to abnormal brain development.⁴

Approximately one-eighth of children in high-income countries have mental disorders,⁵ and much is still unknown regarding their etiology. Results from previous studies, which were primarily observational,^6,7 suggested that lower childhood vitamin D levels, measured as serum 25-hydroxyvitamin D (25[OH]D), were associated with autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD). Lower childhood vitamin D levels have also been associated with increased levels of depressive symptoms⁸ and internalizing and externalizing problems⁹ later in childhood. Causality, however, can be verified only using randomized clinical trials (RCTs). Our 2021 publication of a part of the double-blind interventional RCT Vitamin D Intervention in Infants (VIDI) study¹⁰ did not show benefits of a higher-than-standard vitamin D₃ supplementation (1200 IU) between ages 2 weeks and 2 years compared with standard recommended supplementation (400 IU) on internalizing, externalizing, or dysregulation problems; competencies; or developmental milestones up to age 2 years. Conversely, the study found a potential small negative impact of higher-than-standard supplementation on externalizing problems. Behavioral and psychiatric problems may not be fully manifested during the early years but may become more evident when environmental demands increase.¹¹ For example, the estimated earliest peak age at onset for anxiety and fear–related disorders is 5.5 years.¹¹ Accordingly, the primary aim of this follow-up study was to build on our previous study and extend the inquiry to childhood psychiatric symptoms at ages 6 to 8 years. This period is characterized by increased demand for self-regulating skills, important in mitigating potential internalizing and externalizing problems, combined with a still-developing prefrontal cortex.¹²

Lower 25(OH)D levels in pregnancy have been associated with unfavorable neurobehavioral and mental health outcomes in offspring, including negative affectivity in infancy,¹³ attention-deficit/hyperactivity disorder,^14,15 autism spectrum disorder,¹⁶ and depression.¹⁷ Therefore, our secondary aim was to explore whether a higher-than-standard childhood vitamin D₃ supplementation modified the potential impact of maternal 25(OH)D levels during pregnancy on child mental health outcomes.

Participating children’s parents signed informed consent forms at recruitment and at the 6 to 8–year follow-up for this RCT secondary analysis. Children gave written consent to participate at the 6 to 8–year follow-up. The study was approved by the ethics committee at the Hospital District of Helsinki and Uusimaa and registered with ClinicalTrials.gov (NCT01723852 [VIDI] and NCT04302987 [VIDI2]); it follows the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.

The VIDI study (see trial protocol in Supplement 1) is a double-blind, interventional RCT described previously in detail.^18-20 The study originally comprised 987 families (492 female and 495 male infants) recruited from the Kätilöopisto Maternity Hospital in Helsinki, Finland, at 60 degrees north latitude between January 1, 2013, and June 30, 2014. All infants had Northern European ancestry. Infants were randomized to receive oral vitamin D₃ supplementation at 400 IU (10 μg; 495 individuals) or 1200 IU (30 μg; 492 individuals) from ages 2 weeks to 2 years. Of recruited families, 12 did not meet inclusion criteria (Figure). Randomization was performed in blocks of 50 infants by a pharmacist at Helsinki University Hospital without relation to the study.¹⁸ Supplements were prepared by Orion Pharmaceuticals, and both groups received 5 drops daily. Parents received information of group membership after the 2-year intervention concluded. Self-administered questionnaires were used to collect information regarding parental health, lifestyle, and demographics. Information on gestation, delivery, and child demographics was derived from hospital records. Maternal serum samples were collected during routine maternity clinic follow-up visits at 6 to 27 weeks of gestation (mean [SD] gestation time, 11.3 [1.9] weeks) and stored in the Finnish Maternity Cohort serum bank as organized by the Finnish Institute for Health and Welfare. Samples were used to analyze maternal 25(OH)D concentrations.

The follow-up study was initiated in November 2019; we invited 817 families who remained in the original VIDI study until completion of the intervention at age 2 years of the child and had available home address info to participate. Of 546 families who participated in the follow-up study (55.3% of the original study population), 456 families completed online questionnaires regarding psychological and cognitive outcomes between September 2020 and May 2021. Personnel conducting follow-up were blinded to group membership. There was 1 participant excluded owing to diagnosis of a rare genetic disorder. The population of this long-term follow-up study consisted of 346 children whose parents completed the Child Behavior Checklist (CBCL) questionnaire (63.4% of those in the follow-up study); of the original 400-IU group, 169 children (34.6%) were included, and of the original 1200-IU group, 177 children (36.4%) were included. These children had somewhat more beneficial baseline characteristics regarding breastfeeding duration, maternal 25(OH)D level, smoking status, and parental education level compared with nonparticipants (eTable 1 in Supplement 2).

Serum 25(OH)D concentrations were analyzed at the Pediatric Research Centre, University of Helsinki, using a fully automated IDS-iSYS immunoassay system with chemiluminescence detection (Immunodiagnostics System). Biochemical analyses are described in detail in the eAppendix in Supplement 2 and elsewhere.^18,19 In this study, 30 ng/mL (to convert to nanomoles per liter, multiply by 2.496) was used as the cutoff point for maternal 25(OH)D concentration. This was previously suggested as a cutoff point for suboptimal 25(OH)D levels.²¹

Childhood psychiatric symptoms were assessed at a mean (SD; range) age of 7.1 years (0.4; 6.3-8.2) years using CBCL, a standardized questionnaire comprising 113 items scored using a 3-point Likert scale (0 = absent; 1 = occurs sometimes; 2 = occurs often).²² Parent-reported questionnaires were used to calculate composite scores of internalizing, externalizing, and total problems (Cronbach α was 0.81, 0.89, and 0.93, respectively). Achenbach System of Empirically Based Assessment (ASEBA-PC) software version 3.0.136.0 (T.M. Achenbach) was used to convert raw scores to age and sex–standardized T scores; T scores of 64 or greater are considered to reflect clinically significant internalizing, externalizing, and total problems. We used Z scores of square root–transformed raw scores (0 = mean; 1 = 1 SD) and dichotomized T scores for internalizing, externalizing, and total scores as outcome measurements.

Potential covariates were assessed based on known association with childhood neurodevelopment or vitamin D₃ levels. They were child age at CBCL assessment,²² child sex,²³ maternal 25(OH)D level,^13-17 birth season,^24,25 gestational duration,²⁶ maternal age at delivery,²⁷ breastfeeding duration,²⁸ parental educational level,²⁹ parental single status at follow-up,³⁰ maternal smoking status,³¹ and maternal depressive symptoms assessed at the maternity ward after childbirth³² (eAppendix in Supplement 2).

Longitudinal secondary data analyses followed the intention-to-treat principle; 2-tailed independent-sample t tests and Pearson χ² tests were used to compare follow-up characteristics between intervention groups. Differences between groups in psychiatric problems were assessed using linear and logistic regression analysis. Results are presented without adjustment (model 1) and after adjustment for sex, maternal depressive symptoms at birth, birth season, and parental single status at follow-up (model 2). Model 2 covariates were associated with at least 1 psychiatric problems score (P < .05) or differed significantly in parental single status between supplementation groups. No other assessed potential covariates were associated with outcome variables or group membership and thus were not included in analyses (eTable 2 in Supplement 2). Comparing supplementation groups with α set at .05, we had a power of 0.80 to exclude or confirm effect sizes greater than 0.30 in mean difference (MD) for continuous variables and odds ratios (ORs) smaller than 0.42 in dichotomous variables. The potential impact of attrition bias was tested using inverse probability weighting estimation. In supplemental analyses, we tested associations of 1-year and 2-year 25(OH)D levels with psychiatric outcomes using regression analyses presented previously.

To test the potential interaction between prenatal 25(OH)D level and supplementation degree, children were grouped by maternal pregnancy 25(OH)D level. A 30-ng/mL cutoff point was selected a priori based on previous studies.²¹ Initial analyses testing for interaction between intervention group status and the dichotomized maternal 25(OH)D concentration variable did not show interactions at the P < .05 level (eAppendix in Supplement 2). However, exploratory post hoc linear and logistic regression analyses were conducted with adjustments as presented previously to examine differences in psychiatric problems among 4 subgroups created based on intervention group status and maternal 25(OH)D concentration. For reference groups, we first used the group with 400-IU supplementation and maternal 25(OH)D levels less than 30 ng/mL and then the group with 1200-IU supplementation and maternal 25(OH)D levels of 30 ng/mL or greater.

Analyses were conducted with sexes combined. Additional information on study characteristics stratified by sex and sex × intervention group interaction analysis is given in the eAppendix and eTables 3 and 4 in Supplement 2. Statistical analyses were performed using SPSS statistical software version 28 (IBM) and Stata statistical software version 17 (StataCorp). Data were analyzed from June 2022 to March 2023.

The study population comprised 346 children (164 females [47.4%]; mean [SD] age at follow-up, 7.1 [0.4] years), including 169 children (80 females [47.3%]) in the 400-IU and 177 children (84 females [47.5%]) in the 1200-IU supplementation group. Table 1 presents baseline characteristics. Concentrations of 25(OH)D were significantly higher in the 1200-IU group compared with the 400-IU group at 1 year (MD, 13.2 ng/mL; 95% CI, 11.0 to 15.4 ng/mL; P < .001) and 2 years (MD, 12.7 ng/mL; 95% CI, 10.7 to 14.6 ng/mL; P < .001) (Table 2). Differences between groups were of the same magnitude as those seen for the original study population at 1 year (805 individuals) and 2 year (814 individuals) follow-up.¹⁸ In attrition analysis, there were no differences in externalizing or internalizing behavior at 2 years between participants in this study and nonparticipants with 2-year follow-up data (eAppendix in Supplement 2).

There were 20 children (11.8%) in the 400-IU group and 10 children (5.6%) in the 1200-IU group with clinically significant internalizing problems (P = .04). The corresponding numbers for externalizing and total problems were 16 children (9.5%) vs 16 children (9.0%) (P = .83) and 9 children (5.3%) vs 8 children (4.5%) (P = .90), respectively (Table 2). The OR for clinically significant internalizing problems was 0.40 (95% CI, 0.17-0.94; P = .04) after adjustment for sex, birth season, maternal depressive symptoms at birth, and parental single status at follow-up. Results remained significant after correction for attrition bias. We conducted sensitivity analyses by repeating model 1 analyses and restricting the sample size to 318 children with available CES-D scores. Additionally, model 2 analyses were repeated using means substitution for 28 children missing CES-D scores to retain the entire sample in the analyses. Sensitivity analysis results were in line with primary results (eAppendix in Supplement 2). No effect of supplementation was found for externalizing (OR, 0.89; 95% CI, 0.42-1.91; P = .77) or total (OR, 0.81; 95% CI, 0.29-2.23; P = .68) problems (Table 3).

Table 3 presents MDs for Z scores of square root–transformed internalizing (−0.20; 95% CI, −0.41 to 0.01; P = .07), externalizing (0.01; 95% CI, −0.21 to 0.22; P = .94), and total (−0.08; 95% CI, −0.29 to 0.14; P = .48) problem scores. Score distributions are presented in the eFigure in Supplement 2.

Higher 25(OH)D levels at ages 1 and 2 years resulted in lower risk for clinically significant internalizing problems (1 year: OR, 0.93; 95% CI, 0.90 to 0.97; P = .001; 2 years: OR, 0.95; 95% CI, 0.91 to 0.98; P = .01) and lower internalizing problem scores (1 year: MD, −0.010; 95% CI, −0.019 to −0.001; P = .04; 2 years: MD, −0.012; 95% CI, −0.021 to −0.002; P = .02) in the unadjusted model (eTable 5 in Supplement 2). After adjustment, the effect of 1-year 25(OH)D levels was attenuated for internalizing problem scores but remained for clinically significant internalizing problems (OR 0.94; 95% CI, 0.89 to 0.98; P = .01); the effect remained at 2 years for both outcomes after adjustment (internalizing problem score: MD, −0.011; 95% CI, −0.021 to −0.002; P = .02; internalizing problems: OR, −0.95; 95% CI, 0.91 to 0.99; P = .01). There was no effect of child 25(OH)D levels on externalizing or total problems.

Among 291 families with data on maternal 25(OH)D levels, there were 96 children whose mothers had levels less than 30 ng/mL (33.0%) compared with 208 of 517 nonparticipating families with these data (40.2%). We found no effect of maternal 25(OH)D levels on child psychiatric problems (eTable 2 in Supplement 2). The proportion of children with clinically significant problems for each of 4 maternal 25(OH)D and infant vitamin D₃ supplementation dose subgroups is given in eTable 6 in Supplement 2. The risk of clinically significant internalizing problems was significantly lower among 91 children in the 1200-IU group with maternal 25(OH)D levels of 30 ng/mL or greater compared with 44 children in the 400-IU group with maternal 25(OH)D levels less than 30 ng/mL (adjusted OR, 0.21; 95% CI, 0.06-0.78; P = .02) (Table 4).

In a post hoc subgroup analysis, 48 children in the 400-IU group with maternal 25(OH)D concentrations less than 30 ng/mL had higher internalizing problems scores compared with children in the 1200-IU group, including 44 children whose mothers had 25(OH)D concentrations below (adjusted MD, 0.49; 95% CI, 0.09-0.89; P = .02) and 91 children whose mothers had concentrations above the 30-ng/mL cutoff (adjusted MD, 0.37; 95% CI, 0.03-0.72; P = .04) (Table 4). No significant differences were found between 400-IU and 1200-IU groups in children whose mothers had 25(OH)D concentrations of 30 ng/mL or greater (eTable 7 in Supplement 2).

In this secondary analysis of the VIDI RCT, we explored the potential impact of higher-than-standard vitamin D₃ supplementation between ages 2 weeks and 2 years on psychiatric symptoms at mean age 7.1 years. We found a 5.6% prevalence of clinically significant internalizing problems in children who received 1200-IU oral vitamin D₃ supplementation compared with 11.8% among those who received the standard recommended dose of 400 IU daily. No differences were found for total or externalizing problems. Although statistically significant differences at the P < .05 level were observed only for the dichotomous internalizing problems outcome, findings from the linear regression analysis followed the same direction, supporting the logistic regression findings.

Furthermore, internalizing problem scores were significantly higher for children from the 400-IU group with maternal 25(OH)D levels less than 30 ng/mL, compared with children from the 1200-IU group regardless of maternal 25(OH)D status. This should be interpreted cautiously, however, and be considered only as hypothesis generating given the absence of interactions between maternal 25(OH)D level and supplementation status.

Although previous studies^6-9,13,33 have suggested that higher 25(OH)D levels during fetal life and early childhood may lower the risk of childhood psychopathology, to our knowledge, this is the first RCT to assess the potential impact of high-dose vitamin D₃ supplementation in healthy infants and up to age 2 years on psychiatric symptoms during late preschool and early school age. We previously studied this population up to age 2 years and found no evidence of systematic benefits in child neurodevelopment of higher-than-standard supplementation.¹⁰ Conversely, a potential increase in externalizing problems among children in the 1200-IU supplementation group could not fully be excluded. In this study, we found no differences in externalizing problems between groups.

The potential associations of maternal and childhood 25(OH)D levels with later neurodevelopmental and mental health outcomes were previously comprehensively summarized.^6,7,34 Only a few studies, however, explored the potential association between childhood vitamin D levels and features of internalizing behaviors, such as depression. In a prospective birth cohort study⁸ among 2759 children, higher 25(OH)D concentrations at mean age 9.8 years were associated with lower levels of depressive symptoms at age 13.8 years but not age 10.6 years, suggesting a sustained beneficial outcome increasing over time. A prospective cohort study⁹ among 273 children found that those with 25(OH)D levels less than 30 ng/mL (10.3% of participants) at age 5 to 12 years had higher internalizing and externalizing scores after a median follow-up of 6 years. Conversely, a 2022 cross-sectional study³⁵ among 704 children and adolescents aged 11 to 16 years found no association between 25(OH)D concentration and depressive symptoms.

Studies focusing on maternal pregnancy 25(OH)D levels have reported inconsistent findings. A study³³ among 487 mother-child pairs found an inverse association between first-trimester 25(OH)D levels and externalizing but not internalizing symptoms. In a pregnancy cohort study among 743 mother-child pairs, no associations were found between maternal 25(OH)D concentrations at the 18th pregnancy week and total, internalizing, or externalizing symptoms measured in the child at ages 2, 5, 8, 10, 14, or 17 years.³⁶ In a study¹³ using data from VIDI and the Dutch Generation R cohort study (777 and 1505 mother-child pairs, respectively), lower early to midpregnancy 25(OH)D concentration was associated with higher infant negative affectivity. Negative affectivity is a temperament trait associated with increased risk for internalizing problems in childhood and adolescence.³⁷ Although inconclusive, our subgroup analysis findings may suggest that exposure to maternal 25OHD levels lower than 30 ng/mL during pregnancy combined with standard dose supplementation could increase the risk for later internalizing problems compared with receiving higher-than-standard supplementation during early childhood and may potentially partly explain earlier inconsistencies in the literature.

This study has several strengths, among them the double-blind RCT setting, standardized data collection, and well-characterized study population. Outcomes were assessed using CBCL, a widely used, validated questionnaire^38,39 that allows for assessment of various aspects of childhood behavioral symptoms and potential emerging or manifest signs of psychopathology. CBCL was previously used in studies from Finland.^40,41

This study has several limitations as well. Of the original study population (987 families), 546 families (55.3%) remained in the 6 to 8–year follow-up study, and of these, 346 families (63.4%) completed the CBCL questionnaire. Baseline characteristics were somewhat more beneficial among the study population compared with those lost to follow-up, potentially limiting generalizability to a more diverse population. Furthermore, the proportion of children with maternal 25(OH)D levels less than 30 ng/mL was lower among study participants than in nonparticipants (33.0% vs 40.2%), influencing the number of children available for subgroup analyses. However, attrition rates were similar between supplementation groups, baseline characteristics for nonparticipants did not differ between groups, and inverse probability weighting estimation revealed no indication of attrition bias. Externalizing and internalizing behaviors at age 2 years did not differ between participants and nonparticipants, suggesting that study participants were representative of the 2-year assessment participants in this regard.¹⁰

Questionnaires were collected from September 2020 to May 2021, concurrently with the global SARS-CoV-2 pandemic. The pandemic may have directly and indirectly negatively influenced mental health in children.^42,43 We have no reason, however, to suspect that the burden of the pandemic would have differed between supplementation groups.

Since 2003, milk products and fat spreads in Finland have been fortified with vitamin D, resulting in improved population vitamin D levels.^44,45 This may limit comparisons with studies performed in Finland before 2003 and with countries lacking systematic vitamin D fortification. Furthermore, whether our findings generalize to children living at other geographical latitudes needs to be investigated.

This secondary analysis of an RCT found that a higher-than-standard vitamin D₃ supplementation (1200 IU daily vs 400 IU) between ages 2 weeks and 2 years reduced the risk of internalizing problems later in childhood at ages 6 to 8 years. Results from the exploratory, post hoc subgroup analysis were inconclusive and need to be verified in future studies; further studies may suggest that early life higher-dose vitamin D₃ supplementation is associated with benefits for children exposed to lower pregnancy 25(OH)D levels. Furthermore, study findings need to be interpreted in context with outcomes related to children’s somatic health (eg, growth and allergies), for which lower doses were found to be more beneficial during infancy^20,46; these findings also need to be repeated and assessed for general safety.

Accepted for Publication: April 6, 2023.

Published: May 19, 2023. doi:10.1001/jamanetworkopen.2023.14319

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2023 Sandboge S et al. JAMA Network Open.

Corresponding Author: Samuel Sandboge, MD, PhD, Psychology/Welfare Sciences, Faculty of Social Sciences, University of Tampere, Kalevantie 4, 33100 Tampere, Finland (samuel.sandboge@tuni.fi).

Author Contributions: Drs Sandboge and Heinonen had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Sandboge, Räikkönen, Mäkitie, Andersson, Heinonen.

Acquisition, analysis, or interpretation of data: Sandboge, Räikkönen, Lahti-Pulkkinen, Hauta-alus, Holmlund-Suila, Girchenko, Kajantie, Andersson, Heinonen.

Drafting of the manuscript: Sandboge.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Sandboge, Lahti-Pulkkinen, Girchenko, Heinonen.

Obtained funding: Räikkönen, Hauta-alus, Mäkitie, Andersson, Heinonen.

Administrative, technical, or material support: Hauta-alus, Holmlund-Suila.

Supervision: Kajantie, Mäkitie, Andersson, Heinonen.

Conflict of Interest Disclosures: Dr Kajantie reported receiving grants from the Novo Nordisk Foundation outside the submitted work. No other disclosures were reported.

Funding/Support: This work was supported by the Academy of Finland, Sigrid Jusélius Foundation, Signe and Ane Gyllenberg Foundation, Finland Foundation for Pediatric Research, Yrjö Jahnsson Foundation, Novo Nordisk Foundation, Finska Läkaresällskapet, Finland Special Governmental Subsidy for Clinical Research, Juho Vainio Foundation, and Päivikki and Sakari Sohlberg Foundation. Orion Pharmaceuticals is acknowledged for providing vitamin D₃ supplements gratis for the study.

Role of the Funder/Sponsor: The funders 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.

Data Sharing Statement: See Supplement 3.

Additional Contributions: We would like to express our sincerest gratitude to all children and families who participated in the study. We would also like to thank Jenni Rosendahl, MD, PhD; Otto Helve, MD, PhD; Saara Valkama, MD; and Maria Enlund-Cerullo, MD, and research nurses Nea Boman, BA; Sirpa Nolvi, BA; Rhea Paajanen, BA; and Päivi Turunen, BA (Helsinki Children’s Hospital), for their assistance in data collection and laboratory technician Sari Lindén, BA (Helsinki Children’s Hospital), for work on this study. These contributors have been compensated with salary payments for this work.