Prenatal and postnatal exposure to air pollution and emotional and aggressive symptoms in children from 8 European birth cohorts

Background: The association between air pollution exposure and emotional and behavioural problems in children is unclear. We aimed to assess prenatal and postnatal exposure to several air pollutants and child's depressive and anxiety symptoms, and aggressive symptoms in children of 7-11 years. Methods: We analysed data of 13182 children from 8 European population-based birth cohorts. Concentrations of nitrogen dioxide (NO2), nitrogen oxides (NOx), particulate matter (PM) with diameters of <= 10 mu m (PM10), <= 2.5 mu m (PM2.5), and between 10 and 2.5 mu m (PMcoarse), the absorbance of PM2.5 filters (PM(2.5)abs), and polycyclic aromatic hydrocarbons (PAHs) were estimated at residential addresses of each participant. Depressive and anxiety symptoms and aggressive symptoms were assessed at 7-11 years of age using parent reported tests. Children were classified in borderline/clinical range or clinical range using validated cut offs. Region specific models were adjusted for various socio-economic and lifestyle characteristics and then combined using random effect meta-analysis. Multiple imputation and inverse probability weighting methods were applied to correct for potential attrition bias. Results: A total of 1896 (14.4%) children were classified as having depressive and anxiety symptoms in the borderline/clinical range, and 1778 (13.4%) as having aggressive symptoms in the borderline/clinical range. Overall, 1108 (8.4%) and 870 (6.6%) children were classified as having depressive and anxiety symptoms, and aggressive symptoms in the clinical range, respectively. Prenatal exposure to air pollution was not associated with depressive and anxiety symptoms in the borderline/clinical range (e.g. OR 1.02 [95%CI 0.95 to 1.10] per 10 mu g/m(3) higher NO2) nor with aggressive symptoms in the borderline/clinical range (e.g. OR 1.04 [95%CI 0.96 to 1.12] per 10 mu g/m(3) higher NO2). Similar results were observed for the symptoms in the clinical range, and for postnatal exposures t