In-utero exposure to PM2.5 and adverse birth outcomes in India: Geostatistical modelling using remote sensing and demographic health survey data 2019-21

Abstract

This study investigates the influence of air quality on birth weight and preterm birth. Utilizing data from the national family health survey and raster images, the study employs various statistical analyses and spatial models to elucidate the connection between in-utero exposure to air pollution and birth outcomes, both at the individual and district levels. It was observed that approximately 13% of children were born prematurely, and 17% were born with low birth weight. Increased ambient particulate matter 2.5 concentrations during pregnancy were associated with higher odds of low birth weight (AOR: 1.4; 95% CI: 1.29-1.45). Mothers exposed to particulate matter 2.5 during pregnancy had a heightened likelihood of delivering prematurely (AOR: 1.7; 95% CI: 1.57-1.77) in comparison to unexposed mothers. Climatic factors such as rainfall and temperature had a greater association with adverse birth outcomes. Children residing in the Northern districts of India appeared to be more susceptible to the adverse effects of ambient air pollution. Employing a distributed spline approach, the study identified a discernible upward trend in the risk of adverse birth outcomes as the level of exposure increased, particularly following an exposure level of 40 particulate matter 2.5 ug/m3. A 10 μg m - 3 increase in particulate matter 2.5 exposure was associated with a 5% increase in the prevalence of low birth weight and a 12% increase in preterm birth. Among the different spatial models used in this study, the multiscale geographically weighted regression spatial model showed the best fit to the actual scenario, effectively capturing the spatial relationships between particulate matter 2.5 exposure and adverse birth outcomes. In addition to addressing immediate determinants such as nutrition and maternal healthcare, it is imperative to collaboratively address distal factors encompassing both indoor and outdoor pollution to attain lasting enhancements in child health.

Fig 1. (a) Spatial distribution of in-utero exposure to PM_2.5, (b) low birth weight and (c) preterm birth across in India.

Source: The map was created by the authors using ArcMap, NFHS data, and shapefiles.

Fig 2. Weighted percentage of low birth weight and preterm birth by background characteristics.

Fig 3. Susceptibility to (a) low birth weight, and (b) preterm birth due to the in-utero exposure to PM_2.5.

Fig 4. Bivariate LISA map showing the spatial association between in-utero exposure to PM_2.5 and preterm birth (a) and low birth weight (b).

Source: The map was created by the authors using ArcMap, NFHS data, and shapefiles.

Fig 5. Predicted low birth weight (LBW) and preterm birth (PTB) results from Ordinary List Square (OLS), Geographically Weighed Regression (GWR), Multiscale Geographically Weighed Regression (MGWR), Spatial Lag Model (SLM) and Spatial Error Model (SEM) after adjusting environmental, Socioeconomic maternal and child characteristics.

Source: The map was created by the authors using ArcMap, NFHS data, and shapefiles.