Association between Early Life Exposure to Air Pollution and Working Memory and Attention

Abstract

Background: Although previous studies have reported negative associations between exposure to air pollution and cognition, studies of the effects of prenatal and postnatal exposures in early childhood have been limited.

Objectives: We sought to assess the role exposure to fine particulate matter ([Formula: see text]) during different prenatal and postnatal windows may play in children's cognitive development at school age.

Methods: Within the Brain Development and Air Pollution Ultrafine Particles in School Children (BREATHE) Project, we estimated residential [Formula: see text] exposures by land use regression for the prenatal period and first seven postnatal years of 2,221 children from Barcelona, Spain. The participants ([Formula: see text]) completed computerized tests assessing working memory, attentiveness, and conflict network during four visits in 2012–2013. We used linear mixed effects and distributed lag models to assess the period of exposure to [Formula: see text] in association with cognitive development.

Results: Inverse associations were identified between [Formula: see text] exposure during the fifth and sixth postnatal years and working memory, with boys showing much higher vulnerability. Regarding attention functions, exposure to higher [Formula: see text] levels during the prenatal period and from the fourth postnatal year were associated with a reduction in conflict network performance, though we found no association with attentiveness. The overall estimated cumulative effect of a [Formula: see text] increase in [Formula: see text] resulted in a reduction in the working memory [Formula: see text] score of [Formula: see text] [95% confidence interval (CI): [Formula: see text], [Formula: see text]] points and an increase in the conflict attentional network of 11.31 (95% CI: 6.05, 16.57) milliseconds, indicating a poorer performance.

Conclusions: Early life exposure to [Formula: see text] was associated with a reduction in fundamental cognitive abilities, including working memory and conflict attentional network. https://doi.org/10.1289/EHP3169.

Figure 1.

Concentrations at different time windows that the participants ($n=\mathrm{2,221}$) are exposed to during prenatal time and childhood. The concentrations were estimated at home address for each year from the temporally adjusted LUR model for Barcelona. Lower and upper bound of the boxes represent 25th and 75th percentile, respectively; central line corresponds to the median; bars outside the box represent the $1.5\times \text{interquartile range}$; and circles are outliers. The notch displays the confidence interval around the median. Note: $\mathrm{y}=\text{year}$.

Figure 2.

Association between yearly ${\mathrm{PM}}_{2.5}$ levels estimated at home address for the pregnancy period (year 0) and the first 7 years of life and different cognitive outcomes at school age (7–10 y old): working memory, attentiveness, and conflict network from the independent linear mixed effect models. Legend: Lower $d\prime$ and higher HRT-SE and conflict scores indicate impairment. Models were adjusted for age, sex, maternal education, and residential neighborhood socioeconomic status; school and individual as nested random effects. Solid lines show the difference in the outcomes for an IQR increase in ${\mathrm{PM}}_{2.5}$ [Interquartile range (IQR) exposure contrasts are reported in Table S4]. Dashed lines indicate 95% CIs. Note: $n=\text{number of children}$.

Figure 3.

Associations between yearly ${\mathrm{PM}}_{2.5}$ levels estimated at home address over the pregnancy period (year 0) and the first 7 years of life and different cognitive outcomes at school age (7–10 y old): working memory; attentiveness, and conflict network from the DLM models. Legend: Lower $d\prime$ and higher HRT-SE and conflict scores indicate impairment. The associations are presented for all the population and stratified by sex. Models were adjusted for age, sex (only in the model including all population), maternal education, and residential neighborhood socioeconomic status; school and individual included as nested random effects. Solid lines show the predicted difference in the outcomes associated with an increase of $10{\mathrm{\mu }\mathrm{g}\mathrm{m}}^{-3}$ of ${\mathrm{PM}}_{2.5}$. Grey areas indicate 95% CIs. Note: $n=\text{number of children}$.