LAUSR

Abstract Although increasing evidence has documented the association of exposure to air pollution and childhood obesity, a limited number of studies have reported the relationship of air pollution exposure to early childhood body mass index (BMI) growth trajectories, which are strong predictors of future health. We aimed to examine the associations of prenatal and postnatal (first-year and second-year) exposure to nitrogen dioxide (NO2) and indicators of exposure to major roadways with obesity at 2 years old and BMI-for-age Z-score (BMIZ) growth trajectories between 0 and 2 years old. The length and weight of the children were repeatedly measured at birth and at 1 year (±1 month) and 2 years (±1 month) of age. Sex-specific BMIZs were calculated, and children with BMIZs greater than the 95th percentile were identified as cases of obesity (BMIZ cut-offs of childhood obesity were 1.78 for boys and 1.60 for girls). Exposure to NO2 was evaluated using a land use regression model. Two indicators for major roadways exposure (length of major roadways surrounding the residence and residential distance to the nearest major roadway) were calculated according to the residential address during pregnancy. We explored the prenatal NO2 exposure windows associated with incident obesity at 2 years old and found that the third trimester was the prenatal NO2 exposure window with the significant higher risk. Each 10 μg/m3 incremental change in prenatal exposure to NO2 was associated with a relative risk (RR) of 1.10 for childhood obesity (95% confidence interval: 1.01, 1.20). However, postnatal (first-year and second-year) exposure to NO2 was not associated with obesity risk at 2 years old. No significant associations were observed between indicators of exposure to major roadways and risk of obesity at 2 years old. In addition, prenatal exposure to the highest quartile of NO2 was significantly associated with a faster overall growth rate of the BMIZ from birth to the age of 2 years old (the slope of the BMIZ trajectories, β3 (standard error), for the 4th quartile vs. the 1st quartile: 0.88 (0.15), P value