Table 4.
Review and meta-analysis studies on vitamin D during pregnancy and autism 3.2. Study Design and Population.
| Ref. | Design | Studies Included | Location and Year | Participants | Sample Size (Case with ASD/Control) | ASD Diagnosis | Findings | Major Findings | LE | GR | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Common | Different | Design | ||||||||||
| E. Kocovska et al. (2012) [73] | Review | Humble et al. (2010) | No controls, cross-sectional | Sweden | European adults psychiatric disorder incl. autism | N = 117 (117/0) | Significant difference (p < 0.005) The median (25th–75th percentile) is in autistic patients, 31.5 ng/mL (23–39); and in all patients, 45 ng/mL (31–60) |
Vitamin D deficiency in pregnancy or early childhood may be an environmental trigger for ASD in people genetically susceptible to autism. | 2++ | B | ||
| Meguid et al. (2010) | Case-control, cross-sectional | Egypt | Children with or without autism | n = 112 (70/42) | DSM-IV for clinical diagnosis | Significant difference (p < 0.005) In autistic children, 28.5 ng/mL Controls: typically developing children 40.1 ng/mL In addition, children with autism had significantly lower 25(OH)D (p < 0.00001) and 1,25(OH)(2)D (p < 0.005) as well as lower calcium (p < 0.0001) serum values than the controls. |
||||||
| Molloy et al. (2010) | Case-control, cross-sectional | USA | White male children aged 4–8 years with ASD | n = 89 (49/40) | DSM-IV/ADOS | No significant difference (p = 0.4) Cases: 20 ng/mL All Participants <31 ng/mL. |
||||||
| Fernell et al. (2010) | Case-control, cross-sectional | Sweden | Mothers of Somali or Swedish origin with a child with or without autism | n = 80 (40/40) | Significant difference Somali vs. Swedish mothers Cases: Somali mothers with autism child 6.7 ng/mL; Swedish mothers with autistic child 24.8 ng/mL Controls: Somali mothers with nonautistic child 9.6 ng/mL; Swedish mothers with nonautistic child 20.7 ng/mL. |
|||||||
| H. Mazahery et al. (2016) [72] | Review | Humble et al. (2010) | No controls, cross-sectional | Sweden | European adults psychiatric disorder incl. autism | n = 117 (117/0) | Significant difference (p < 0.005) In autistic patients, 31.5 ng/mL (23–39) In all patients, 45 ng/mL (31–60). |
Low vitamin D status in utero, postnatal, and early childhood might affect different brain regions, and that situation might cause different neurodevelopmental and cognitive outcomes in infants. | 2++ | B | ||
| Stubbs et al. (2016) | Prospective cohort (uncontrolled) | USA | Pregnant mothers of autistic children aged 18 months–3 years | n = 20 (no control) | MCHAT and PDDBI | Although autism recurrence rate reported in the literature is 20%, recurrence rate of autism in newborns was 5%. | ||||||
| Saad et al. (2015) | Open-label intervention | Egypt | Autistic children with 25(OH)D <75 nmol/L |
n = 222 (122/100) |
DSM-IV/CARS, ABC 2 | 13% (n = 16): normal serum 25(OH)D concentration 57% (n = 70): vitamin D deficiency, 30% (n = 36): vitamin D insufficiency Mean 25(OH)D levels in patients with severe autism were significantly lower than patients with mild or moderate autism (p < 0.0001). Serum 25(OH)D levels showed significant negative correlation with CARS scores (r = −0.502, p < 0.0001). |
||||||
| Azzam HME et al. (2015) | RCT | Egypt; from January 2009 until June 2010 |
Children with ASD | n = 21 (21/0) | DSM-IV | No significant difference (three groups: 2.7 ± 1.9 y for vitamin D–supplemented group, 5.8 ± 2.9 y for unsupplemented autistic group, 5 ± 1.8 y for neurotypical control group) Mean presupplementation CARS scores for the two autistic groups were 33.9 ± 2.9 and 33 ± 3 (p > 0.05). Mean VABS was 51.4 ± 16 and 55.7 ± 20 (p > 0.05). |
||||||
| Ucuz İİ et al. (2015) | Open-label intervention | Turkey | Toddlers aged 2–5 years with developmental delay without and with ASD | n = 66; 25(OH)D <50 nmol/L (n = 11 cases) and ≥50 nmol/L (n =10 controls) | ABC 1 Denver II | No significant difference between groups with normal and low vitamin D levels in terms of ADSI, Denver II, CBCL, and ABC scores, and NGF and GDNF levels (p > 0.05). | ||||||
| Jia F et al. (2015) | Case | China | 32-month-old male toddler with ASD | ABC 1 CARS |
Some autism symptoms improved significantly after vitamin D3 supplementation. | |||||||
| Feng J et al. (2016) | Open-label intervention | China | Children with ASD | n = 500 (215/285) | DSM-IV/ADOS | Vitamin D3 may play an important role in etiology of ASD. Serum levels of 25(OH)D were significantly lower in ASD children than control group (p < 0.05). |
||||||
| T. Wang et al. (2016) [74] | Systematic review and meta-analysis | Fernell et al. (2015) | Sibling–control | Sweden, 2005–2008 | Gothenburg catchment area group and Stockholm Somali group; children aged 4+ years | n = 116 (58/58) | Clinical report; ASD diagnosis | Collapsed group of children with ASD had significantly lower vitamin D levels (Mean = 24.0 nM, SD = 19.6) than siblings (Mean = 31.9 nM, SD = 27.7) (p = 0.013). | Decreased vitamin D levels in patients, decreased maternal vitamin D levels during pregnancy, and decreased exposure to solar UVB might increase the risk for ASD. | 2++ | B | |
| Meguid et al. (2010) | Case-control, cross-sectional | Egypt | Children with or without autism | n = 112 (70/42) | DSM-IV for clinical diagnosis | Significant difference (p < 0.005) In autistic children, 28.5 ng/mL Controls: typically developing children 40.1 ng/mL In addition, children with autism had significantly lower 25(OH)D (p < 0.00001) and 1,25(OH)(2)D (p < 0.005) as well as lower calcium (p < 0.0001) serum values than the controls. |
||||||
| Saad et al. (2015) | Open-label intervention | Egypt | Children with ASD | n = 222 (122/100) | DSM-IV/CARS, ABC 2 | 13% (n = 16): normal serum 25(OH)D concentration 57% (n = 70): vitamin D deficiency, 30% (n = 36): vitamin D insufficiency Mean 25(OH)D levels in patients with severe autism were significantly lower than patients with mild or moderate autism (p < 0.0001). Serum 25(OH)D levels showed significant negative correlation with CARS scores (r = −0.502, p < 0.0001). |
||||||
| Adams et al. (2011) | Randomized, double-blind, placebo-controlled | U.S. | Two groups, an Arizona group aged 5–16 years and a National group aged 3–60 y; children and adults with autism | n = 141 (53/88) | DSM-V | No significant differences between ASD and control group in vitamin D3 [25(OH)D in plasma; p = 0.04]. |
||||||
| Mostafa et al. (2012) | Cross-sectional | Saudi Arabia, April–September 2012 | Children aged 5–12 years with and without ASD |
n = 80 (50/30) | DSM-IV | Autistic children had significantly lower serum levels of 25(OH)D than healthy children (p < 0.001). In addition, serum 25(OH)D levels had significant negative correlations with CARS (p < 0.001) and serum levels of anti-MAG autoantibodies (p < 0.001). |
||||||
| Neumeyer et al. (2013) | Cohort | U.S. | Boys aged 8–14 years with and without ASD | n = 37 (18/19) | DSM-IV | Serum 25(OH)D levels were lower in boys with ASD than in controls (p = 0.05). | ||||||
| Uğur et al. (2014) | Case-control | Turkey | ASD (recruited for the study); non-ASD healthy controls | n = 108 (54/54) | DSM-IV | No difference between ASD and healthy controls. | ||||||
| Gong et al. (2014) | Case-control | China, January–December 2012 | Children with and without ASD; consecutive patients with ASD admitted to Dept. of Neurology | n = 96 (48/48) | DSM-IV, CARS (all cases) | Significant negative association between serum 25(OH)D levels and CARS scores (p = 0.000). | ||||||
| Bener et al. (2014) | Case-control | Qatar, June 2011–May 2013 | Children aged 3–8 years; children with and without ASD | n = 508 (254/254) | DSM-IV | Significant difference found in mean values of vitamin D between autism (18.39 ± 8.2 ng/mL with median 18) and versus control children (21.59 ± 8.4 ng/mL) (p < 0.0001) and with median 21 (p = 0.004). | ||||||
| DU et al. (2015) | Cohort | China | Children with and without ASD | n = 226 (117/109) | Lower 25(OH)D in ASD group (p < 0.01). | |||||||
| Tostes et al. (2012) | Cohort | Brazil | Children with and without ASD | n = 48 (24/24) | DSM-IV | Serum levels of 25(OH)D were lower in children with autism (26.48 ± 3.48 ng/mL) than in typically developing subjects (40.52 ± 3.13 ng/mL) (p < 0.001). | ||||||
| A.M. Garcia-Serna and E. Morales (2019) [71] | Systematic review and meta-analysis | Stubbs et al. (2016) | Prospective cohort (uncontrolled) | U.S. | Pregnant mothers of autistic children; 18 months–3 years | n = 20 (no control) | MCHAT and PDDBI | Although the autism recurrence rate reported in the literature is 20%, the recurrence rate of autism in newborns was 5%. | Meta-analysis showed higher levels of prenatal 25(OH)D to be associated with a lower risk of autistic traits. | 2++ | B | |
| Fernell et al. (2015) | Sibling–control | Sweden, 2005–2008 | Gothenburg catchment area group and Stockholm Somali group; children aged 4+ years | n = 116 (58/58) | Clinical report; ASD | Collapsed group of children with ASD had significantly lower vitamin D levels (M = 24.0 nM, SD = 19.6) than their siblings (M = 31.9 nM, SD = 27.7) (p = 0.013). | ||||||
| Morales et al. (2012) |
Cohort | Spain, 2003–2008 | Mothers (13 weeks) 25(OH)D; children aged 14 months | n = 1820 | Psychologist report; BSID | Infants of mothers with 25(OH)D3 concentrations in pregnancy >30 ng/mL showed higher mental score (β = 2.60; 95% CI, 0.63 to 4.56) and higher psychomotor score (β = 2.32; 95% CI, 0.36 to 4.28) than those of mothers with 25(OH)D3 concentrations <20 ng/mL. The median plasma value of 25(OH)D (3) in pregnancy was 29.6 ng/mL (interquartile range, 21.8–37.3). |
||||||
| Morales et al. (2015) | Cohort | Spain; 1997–2008 | Mothers (13 weeks) 25(OH)D; children aged 4–8 years | n = 1650 | DSM-IV | Total ADHD-like symptoms in children decreased by 11% per 10-ng/mL increment of maternal 25(OH)D3 concentration (IRR = 0.89; 95% CI, 0.80 to 0.98). Similarly, number of symptoms in the ADHD subscales decreased in relation to higher maternal 25(OH)D3 concentration (IRR per 10-ng/mL increment = 0.89; 95% CI, 0.79 to 0.99 for inattention scale; and IRR = 0.88; 95% CI, 0.78 to 0.99 for hyperactivity-impulsivity scale). Using diagnostic criteria, researchers found that association of increasing maternal 25(OH)D3 with a lower risk of ADHD DSM-IV (relative risk ratio per 10-ng/mL increment = 0.87; 95% CI, 0.72, 1.06) and ICD-10 hyperkinetic disorder (relative risk ratio = 0.72; 95% CI, 0.49, 1.04) in children. |
||||||
| Zhu et al. (2015) | Cohort | China, 2008 | Children aged 16–18 months | n = 363 | Certified examiners; BSID | Toddlers in lowest quintile of cord blood 25(OH)D exhibited a deficit of 7.60 (95% CI, 212.4 to 22.82; p = 0.002) and 8.04 (95% CI, 212.9 to 23.11; p = 0.001) points in the MDI and PDI scores, respectively, compared with the reference category. Mean 25(OH)D (range): 4.9 (2.2–44.4). |
||||||
| Chen et al. (2016) | Case-control | China, January 2014–December 2015 | Mothers (1st trimester) 25(OH)D; children aged 3–7 years | n = 136 (68/68) | DSM-V | Mothers in autistic group had significantly lower maternal serum levels of 25(OH) D than in typically developing group [19.2 (IQR: 15.8–22.9) ng/mL vs. 24.3 (19.3–27.3) ng/mL, p < 0.001] with 55.9% and 29.4% being vitamin D deficient, respectively (p < 0.001). Levels of 25(OH) D increased with decreasing severity of ASD as defined by CARS score (r = −0.302, p < 0.001). Maternal first-trimester serum levels of 25(OH)D in the lower 3 quartiles (quartile 1, 2, 3) (compared to the highest quartile) were associated with increased odds of ASD diagnosis in offspring [OR (95% CI) Q1: 1.36 (0.84 to 2.58, p = 0.25); Q2: 2.68 (1.44 to 4.29, p = 0.006); Q3: 3.99 (2.58 to 7.12, p < 0.001)]. |
||||||
| Magnusson et al. (2016) | Cohort | Sweden, 2001–2011 | Stockholm Youth Cohort; Swedish-born individuals aged 4–17 years (no control) | N = 509,639 (9882 ASD cases; 2476 with intellectual disability, 7406 without intellectual disability) | National and regional registers; ASD diagnosis | Maternal vitamin D deficiency was associated with offspring risk of ASD with, but not without, intellectual disability (aORs 2.51; 95% CI, 1.22 to 5.16 and 1.28, 95% CI, 0.68 to 2.42). | ||||||
| Darling et al. (2017) | Cohort | UK, 1991–1992 | Mothers (29 weeks) 25(OH)D; children aged 6–42 months; 7–9 years | n = 7065 | Maternal report; SDQ | Children of vitamin D–deficient mothers (<50.0 nmol/L) were more likely to have scores in the lowest quartile for gross motor development at 30 months (OR 1.20; 95% CI, 1.03 to 1.40), fine-motor development at 30 months (OR 1.23; 95% CI, 1.05 to 1.44), and social development at 42 months (OR 1.20; 95% CI; 1.01 to 1.41) than vitamin D–sufficient mothers (≥50.0 nmol/L). Median (IQR) 25(OH)D: 24.5 (17.2–33.9). |
||||||
| Chawla et al. (2017) | Prospective | U.S., 2009–2011 | Newborn Epigenetic Study; mothers who measured 25(OH)D concn in plasma samples in first or second trimester; children aged 12–24 months | n = 218 mother–infant pairs | Maternal report; ITSEA | Black mothers had much lower 25(OH)D concentrations than white and Hispanic mothers. | ||||||
| Gould et al. (2017) | RCT of DHA supplementation in pregnancy | Australia; 2005–2008 |
Children aged 18 months–4 years | 299–323 | Psychologist report; BSID-III | 25(OH)D was not associated with cognitive, motor, social-emotional or adaptive behavior scores at 18 months or cognitive score at age 4 years. | ||||||
| Laird et al. (2017) | Cohort | Republic of Seychelles, 2001 | Mothers 25(OH)D; children aged 5 years | n = 189 | CBCL; total t score | Maternal 25(OH)D concentrations were not associated with child anthropometric or neurodevelopmental outcomes. Mean (SD) 25(OH)D: 22.4 (11.4) ng/ml |
||||||
| Mossin et al. 2017) | Cohort | Denmark, 2010–2012 | Newborns (cord blood) 25(OH)D; children aged 1.5–5 y | n = 1233 | Parental report; CBCL; ADHD score | Cord blood 25(OH)D levels > 25 nmol/L and >30 nmol/L were associated with lower attention deficit hyperactivity disorder scores than levels = 25 nmol/L or <25 nmol/L (p = 0.035) and =30 nmol/L or <30 nmol/L (p = 0.043), respectively. | ||||||
| Daraki et al. (2018) | Cohort | Greece, 2006–2007 | Mothers (13 weeks) 25(OH)D; children aged 4 y | n = 487 | SDQ | Children of mothers with high 25(OH)D levels had also fewer total behavioral difficulties (β-coeff: −1.25, 95% CI, −2.32, to –0.19) and externalizing symptoms (β-coeff: −0.87, 95% CI to −1.58, −0.15) at preschool age. | ||||||
| Vinkhuyzen et al. (2017) | Cohort | Netherlands, April 2002 and January 2006 | Embedded in the Generation R Study; mothers (mid-gestation) 25(OH)D; children aged 6 y | n = 3895, n = 2870 | Clinical records; ASD diagnosis | Individuals in the 25(OH)D-deficient group at mid-gestation had more than twofold increased risk of ASD (OR = 2.42, 95% CI, 1.09 to 5.07, p = 0.03) compared with the 25(OH)D-sufficient group. | ||||||
| Vinkhuyzen et al. (2018) | Prospective cohort | Netherlands, April 2002 and January 2006 | Embedded in the Generation R Study; mothers (mid gestation) 25(OH)D; children aged 6 y | n = 4229 | Parental report; SRS: autism-related traits | Compared with the 25OHD sufficient group (25OHD >50 nmol/L), those who were 25(OH)D deficient had significantly higher (more abnormal) SRS scores (mid- gestation n = 2866, β = 0.06, p < 0.001; cord blood n = 1712, β = 0.03, p = 0.01). | ||||||
| Wang et al. (2018) | Cohort | China, 2012–2013 | Newborns (cord blood) 25(OH)D; children aged 2 y | n = 552 | Parental report; ASQ: gross and fine-motor skills | Median of the 25(OH)D concentration in cord blood was 22.4 ng/mL (IQR, 27.3–8.6). Infants born in winter had lower 25(OH)D concentration. 25(OH)D deficiency was not associated with weight z-score (mean difference, 0.07; 95% CI, −0.09 to 0.23), length z-score (mean difference, 0.01; 95% CI, −0.19 to 0.21), head circumference z-score (mean difference, −0.06; 95% CI, −0.27 to 0.15) and BMI z-score (mean difference, 0.09; 95% CI, −0.07 to 0.25) or neurodevelopment during infancy, adjusting for sex, socioeconomic position, prepregnancy maternal BMI, and maternal and neonatal characteristics. | ||||||
| Veena et al. (2017) | Cohort | India, 1997–1998 | Mothers (30 weeks) 25(OH)D; children aged 9–10 and 13–14 y | n = 468; n = 472 | Psychologist report; KABC | No evidence of an association between maternal vitamin D status and offspring cognitive function. | ||||||
| Tylavsky et al. (2015) | Cohort | U.S., 2006–2011 | CANDLE Study Mothers (2nd trimester) 25(OH)D; children aged 2 years |
n = 1020 mother; 16–28 weeks of gestation | BSID | Gestational 25(OH)D was positively associated with cognitive scaled scores, receptive language, and expressive language (p < 0.001) Mean (range) 25(OH)D: 22.3 (5.9–68.4) ng/ml |
||||||
| Gale et al. (2008) | Cohort | UK (Southampton), 1991–1992 | Mothers (32 weeks) 25(OH)D; children aged 9 y | n = 178 | Mother report; SDQ | No statistically significant associations between maternal 25(OH)D concn and full-scale, verbal or performance IQ, assessed by the Wechsler Abbreviated Scale of Intelligence (p > 0.005) Median (IQR) 25(OH)D:25 (15–30.1) ng/ml |
||||||
| Hanieh et al. (2014) | Cohort | Vietnam, 2010–2012 | Mothers (32 weeks) 25(OH)D; children aged 6 months | n = 960 | Psychologist report; BSID | Infants born to women with 25(OH)D deficiency (<37.5 nmol/L) had lower developmental language scores than infants born to women who were vitamin D replete (≥75 nmol/L) (mean difference, −3.48; 95% CI, −5.67 to –1.28). | ||||||
| Whitehouse et al. (2012) | Cohort | Perth, Western Australia, 1989–1991 | Mothers (18 weeks) 25(OH)D; children aged 2–17 y | 412–652 | Parental report; CBCL: total behavior, internalizing behavior, externalizing behavior | No significant associations between maternal 25(OH)D serum quartiles and offspring behavioral/emotional problems at any age. | ||||||
| Whitehouse et al. (2013) | Cohort | Australia, May 1989 and Nov. 1991 | Raine Study Mothers gestational age 16–20 wks; children aged 5–17 y | n = 406 | Parental report; ASD diagnosis; autism spectrum quotient | Offspring of mothers with low 25(OH)D concentrations (<49 nmol/L) were at increased risk for “high” scores (≥2 SD above mean) on the Attention Switching subscale (OR, 5.46; 95% CI, 1.29 to 23.05). | ||||||
| Strøm et al. (2014) | Cohort | Denmark, 1988–1989 |
Mothers (>30 weeks) 25(OH)D; children aged 22 y | n = 850 | Population-based registry; prescription for medication or hospital admission for: ADHD | Direct association between maternal vitamin D status and offspring depression (ptrend = 0.01). |
||||||
| Gustafsson et al. (2015) | Cohort | Sweden, 1978–2000 | Newborns (cord blood) 25(OH)D3; children aged 5–17 y | n = 404 (202/202) | DSM-III-R used before 1994 and DSM-IV used after 1994 | No differences in cord blood vitamin D concentration were found between children with ADHD (median 13.0 ng/mL) and controls (median 13.5 ng/mL) (p = 0.43). | ||||||
| Keim et al. (2014) | Case-control | U.S., 1959–1965 | Collaborative Perinatal Project; n = 55,000 pregnant women; children aged 8 month and 4–7 y | 3146–3587 | Psychologist report; BSID | IQ at age 7 was associated with both maternal and cord blood 25(OH)D (β for 5-nmol/L increment of maternal 25(OH)D = 0.10; 95% CI, 0.00 to 0.19). | ||||||
ASD, autism spectrum disorder; RCT, randomized controlled trial; 95% CI, 95% confidence interval; ABC 1, Autism Behavior Checklist; ABC 2, Aberrant Behaviour Checklist; ADHD, Attention deficit–hyperactivity disorder; ADOS, Autism Diagnostic Observation Schedule; aOR, Adjusted odds ratio; ASQ, Ages and Stages Questionnaire; BMI, body mass index (kilograms per square meter of body surface area); BSID, Bayley Scales of Infant Development; CARS, Childhood Autism Rating Scale; CBCL, Child Behavior Checklist; DSM, Diagnostic and Statistical Manual; GDNF, glial cell line–derived neurotrophic factor; IQR, interquartile range; IRR, incidence rate ratio; ITSEA, Infant–Toddler Social and Emotional Assessment; KABC, Kaufmann Assessment Battery for Children; anti-MAG, anti- myelin-associated glycoprotein; DHA, docosahexaenoic acid; MCHAT, Modified Checklist for Autism in Toddlers; MDI, mental development index; NGF, nerve growth factor; OR, odds ratio; PDDBI, Pervasive Developmental Disorder Behavior Inventory; PDI, Psychomotor development index; SDQ, Strengths and Difficulties Questionnaire; SD, standard deviation; SRS, Social Responsiveness Scale.