Effects of prenatal exposure to particulate matter air pollution on corpus callosum and behavioral problems in children

Objective: Air pollution (AP) may affect neurodevelopment, but studies about the effects of AP on the growing human brain are still scarce. We aimed to investigate the effects of prenatal exposure to AP on lateral ventricles (LV) and corpus callosum (CC) volumes in children and to determine whether the induced brain changes are associated with behavioral problems. Methods: Among the children recruited through a set of representative schools of the city of Barcelona, (Spain) in the Brain Development and Air Pollution Ultrafine Particles in School Children (BREATHE) study, 186 typically developing participants aged 8-12 years underwent brain MRI on the same 1.5 T MR unit over a 1.5-year period (October 2012-April 2014). Brain volumes were derived from structural MRI scans using automated tissue segmentation. Behavioral problems were assessed using the Strengths and Difficulties Questionnaire (SDQ) and the criteria of the Attention Deficit Hyperactivity Disorder DSM-IV list. Prenatal fine particle (PM2.5) levels were retrospectively estimated at the mothers' residential addresses during pregnancy with land use regression (LUR) models. To determine whether brain structures might be affected by prenatal PM2.5 exposure, linear regression models were run and adjusted for age, sex, intracranial volume (ICV), maternal education, home socioeconomic vulnerability index, birthweight and mothers' smoking status during pregnancy. To test for associations between brain changes and behavioral outcomes, negative binomial regressions were performed and adjusted for age, sex, ICV. Results: Prenatal PM2.5 levels ranged from 11.8 to 39.5 mu g/m(3) during the third trimester of pregnancy. An interquartile range increase in PM2.5 level (7 mu g/m(3)) was significantly linked to a decrease in the body CC volume (mm(3)) (beta = -53.7, 95%CI [-92.0, -15.5] corresponding to a 5% decrease of the mean body CC volume) independently of ICV, age, sex, maternal education, socioeconomic vulnerability index at home, birthweight and mothers smoking status during the third trimester of pregnancy. A 50 mm(3) decrease in the body CC was associated with a significant higher hyperactivity subscore (Rate Ratio (RR) = 1.09, 95%CI [1.01, 1.17) independently of age, sex and ICV. The statistical significance of these results did not survive to False Discovery Rate correction for multiple comparisons. Conclusions: Prenatal exposure to PM2.5 may be associated with CC volume decrease in children. The consequences might be an increase in behavioral problems.