Table 3.
Autism factor mean score regression on maternal 25(OH)D adjusted for season and 20-week gestation
| Analysis | Model | Complete record analysis | Multiple imputation analysis | ||
|---|---|---|---|---|---|
| Mean difference* | 95% CI | Mean difference | 95% CI | ||
| Observationala | Unadjusted | −0.01 | (−0.01, 0.00) | −0.01 | (−0.02, −0.00) |
| Adjusted 1 | −0.01 | (−0.01, 0.00) | −0.01 | (−0.01, 0.00) | |
| Adjusted 2 | 0.00 | (−0.01, 0.00) | −0.01 | (−0.01, 0.00) | |
| Adjusted 3 | −0.01 | (−0.01, 0.00) | −0.01 | (−0.01, 0.00) | |
| Fully adjusted | −0.01 | (−0.01, 0.00) | −0.01 | (−0.01, 0.00) | |
| Mendelian randomizationb | Adjusted for 10 principal components | −0.04 | (−0.13, 0.04) | – | – |
| Adjusted for 10 principal components and offspring risk score for 25(OH)D | −0.05 | (−0.20, 0.09) | – | – | |
aMean difference in autism factor mean score per 10 nmol/L change in adjusted 25(OH)D
bMean difference in autism factor mean score per 10 nmol/L change in adjusted 25(OH)D per unit increase in standardised maternal 25(OH)D genetic score
*Causal mean difference for Mendelian randomization
Adjusted 1 = adjusted for offspring sex;
Adjusted 2 = adjusted for financial difficulties, maternal education, and maternal occupational class;
Adjusted 3 = adjusted for parity, maternal age at birth, pre-pregnancy BMI and smoking status during pregnancy;
Fully adjusted = adjusted for all variables in adjusted 1–3
Complete record analysis N = 4913; multiple imputation analysis N = 7689
Mendelian randomization analysis, unadjusted N = 4462, adjusted N = 2937