Maternal Exposure to Greenness during Pregnancy and Birth Weight

Abstract

Background/Aim: Exposure to greenness in the residential neighbourhood, measured through the normalized difference vegetation index (NDVI), has been shown to affect human health. In this study, we study the association between maternal exposure to NDVI during pregnancy and birth weight with adjustment for air pollution as well as other individual and neighbourhood confounders. Methods: For 684,266 births between 2001 and 2012 in Massachusetts, each birth was assigned an average NDVI during pregnancy using the 250 * 250 m grid cell for which the mother lived in. We regress birth weight against NDVI, particulate air pollution (PM2.5 at 1 * 1 km), and other covariates in generalized linear models. We adjust for nonlinearity in effect using splines. The full list of covariates are as follows: marital status, maternal education, government-supported prenatal care, maternal/paternal age and race, maternal cigarette smoking before/during pregnancy, parity, Kotelchuck Index for prenatal care, gestational/other diabetes and hypertension, mode of delivery, infant sex, clinical gestational age, and census-tract percent black population and median income. Results: In the linear model with NDVI and covariates, an interquartile range (IQR) increase in NDVI (0.225) was associated with a 13.87 (95% CI: 11.98, 15.75) gram increase in birth weight. After adjusting for PM2.5, an IQR increase in NDVI was associated with a 14.90 (12.91, 16.90) gram increase in birth weight. Including interactive terms between NDVI and each of government-supported prenatal care, maternal education, and infant sex did not significantly change the NDVI effect estimates. Conclusions: Maternal exposure to greenness measured through NDVI appears beneficial to newborn health as it is positively associated with birth weight, even after adjusting for highly-resolved particulate air pollution predictions and many individual / neighbourhood health and socioeconomic variables.