@ARTICLE{10.3389/fendo.2018.00771, AUTHOR={Fleisch, Abby F. and Aris, Izzuddin M. and Rifas-Shiman, Sheryl L. and Coull, Brent A. and Luttmann-Gibson, Heike and Koutrakis, Petros and Schwartz, Joel D. and Kloog, Itai and Gold, Diane R. and Oken, Emily}, TITLE={Prenatal Exposure to Traffic Pollution and Childhood Body Mass Index Trajectory}, JOURNAL={Frontiers in Endocrinology}, VOLUME={9}, YEAR={2019}, URL={https://www.frontiersin.org/articles/10.3389/fendo.2018.00771}, DOI={10.3389/fendo.2018.00771}, ISSN={1664-2392}, ABSTRACT={Background: Limited evidence suggests an association between prenatal exposure to traffic pollution and greater adiposity in childhood, but the time window during which growth may be most affected is not known.Methods: We studied 1,649 children in Project Viva, a Boston-area pre-birth cohort. We used spatiotemporal models to estimate prenatal residential air pollution exposures and geographic information systems to estimate neighborhood traffic density and roadway proximity. We used weight and stature measurements at clinical and research visits to estimate a BMI trajectory for each child with mixed-effects natural cubic spline models. In primary analyses, we examined associations of residential PM2.5 and black carbon (BC) exposures during the third trimester and neighborhood traffic density and home roadway proximity at birth address with (1) estimated BMI at 6 month intervals through 10 years of age, (2) magnitude and timing of BMI peak and rebound, and (3) overall BMI trajectory. In secondary analyses, we examined associations of residential PM2.5 and BC exposures during the first and second trimesters with BMI outcomes.Results: Median (interquartile range; IQR) concentration of residential air pollution during the third trimester was 11.4 (1.7) μg/m3 for PM2.5 and 0.7 (0.3) μg/m3 for BC. Participants had a median (IQR) of 13 (7) clinical or research BMI measures from 0 to 10 years of age. None of the traffic pollution exposures were significantly associated with any of the BMI outcomes in covariate-adjusted models, although effect estimates were in the hypothesized direction for neighborhood traffic density and home roadway proximity. For example, greater neighborhood traffic density [median (IQR) 857 (1,452) vehicles/day x km of road within 100 m of residential address at delivery] was associated with a higher BMI throughout childhood, with the strongest associations in early childhood [e.g., per IQR increment natural log-transformed neighborhood traffic density, BMI at 12 months of age was 0.05 (−0.03, 0.13) kg/m2 higher and infancy peak BMI was 0.05 (−0.03, 0.14) kg/m2 higher].Conclusions: We found no evidence for a persistent effect of prenatal exposure to traffic pollution on BMI trajectory from birth through mid-childhood in a population exposed to modest levels of air pollution.} }