Maternal selenium status and neuropsychological development in Spanish preschool children

Background: The relationship between maternal selenium (Se) status and child neurodevelopment has been scarcely assessed. In a previous study we observed an inverse U-shaped association between maternal Se concentrations and infant neurodevelopment at 12 months of age. In this study, this non-linear association was explored at preschool age. The effect modification by breastfeeding, child's sex and cord blood mercury was also evaluated. Methods: Study subjects were 490 mother-child pairs from the Spanish Childhood and Environment Project (INMA, 2003-2012). Child neuropsychological development was assessed at around 5 years of age by the McCarthy Scales of Children's Abilities (MSCA). Sociodemographic and dietary characteristics were collected by questionnaire at the first and third trimester of gestation and at 5 years of age. Se was measured in serum samples by ICP-MS at the end of the first trimester of pregnancy (mean +/- standard deviation (SD) = 12.4 +/- 0.6 weeks of gestation). Results: The mean +/- SD of maternal serum Se concentrations was 79.9 +/- 8.1 mu g/L. In multivariate analysis, no linear association was found between Se concentrations and the nine MSCA scales. Generalized additive models indicated inverted U-shaped relationships between Se concentrations and the verbal and global memory scales. When assessing the influence of effect modifiers, breastfeeding played a role: the association between Se and neuropsychological development was inverted U-shaped for the quantitative, general cognitive, working memory, fine motor, global motor and executive function scales only for non-breastfed children. Conclusion: Low and high maternal Se concentrations seem to be harmful for child neuropsychological development, however further studies should explore this non-linear relationship.