Prenatal Exposure to Wildfire Smoke and Infant Health: Evidence from a Nationwide Analysis

• In Event: Wildfires, Air Pollution, and Acid Rain: Harmful Effects on Humans

Abstract

Wildfire smoke plumes are among the most significant public health risks associated with climate change. Infants appear to be particularly vulnerable to exposure during the prenatal period. However, existing estimates of the impact of wildfire smoke on birth outcomes come from a handful of states (i.e. California, Arizona, Colorado). These estimates may not generalize to other states due to heterogeneity in smoke composition. The pollution content of smoke varies due to fire intensity, fuel sources, and a plume’s chemical evolution as it drifts from its source. Plumes ejected into the upper atmosphere can travel vast distances, but may remain high above the surface and pose less risk at ground level. This paper estimates the impact of prenatal smoke exposure on health at birth across the entire continental United States.  

We count the number of medium or thick smoke-days at the county level using satellite data from NOAA’s Hazard Mapping System (HMS). We complement these data with an atmospheric model that generates county-by-day estimates of pollutant concentrations (i.e. PM2.5) overall, and the amount attributable to smoke. We merge these data to 34.8 million restricted birth records that capture gestational age and birthweight, for infants conceived between 2011 and 2020. Exposure is defined as the number of smoke-days in the prenatal period and is measured over a 32-week period starting from the week of conception. We fix smoke exposures over a 32-week period because the count of smoke-days is mechanically related to gestational length. To estimate the impact of prenatal smoke we regress outcomes on prenatal smoke-days, county-by-conception month fixed effects and state-by-birth year fixed effects. The model leverages plausibly exogenous year-to-year variation in smoke for infants conceived in the same county and calendar month.    

Our results suggest that each additional day of prenatal smoke exposure increases the pre-term birth rate by 1.0 per 1000 births (p<0.001) and low-birthweight by 0.72 per 1000 births (p<0.001). Results are robust to alternative fixed effects specifications and suggest a linear relationship between smoke-days and increased risk of pre-term birth. Our coefficients suggest that at the median number of smoke-days (0.32 days), wildfire smoke increased the relative risk of pre-term birth by 0.3%. For births at the 90th percentile (7 days), we estimate a 7.5% increased risk due to smoke. On average across the United States, the marginal impact of a prenatal smoke-day is larger than previous estimates from California. In particular, infants born in regions of the US that do not experience large smoke burdens are more sensitive to the marginal smoke-day. Ongoing work investigates if the heterogeneity in effects we observe stems from differences smoke composition or differences in precautionary behavior.

Wildfire smoke is an increasingly important health risk and the burden will continue to grow as the climate warms. This paper presents the first national assessment of the extent of harms to prenatal smoke exposure. Our estimates will be useful to communities developing mitigation efforts and will provide new evidence to state and federal policymakers that must trade-off climate investments with other priorities.

Author

• Maranna Yoder, Indiana University
  Presenting Author