The impact of wildfire smoke exposure on cardiovascular health in the western US

Cardiovascular diseases (CVD) are the primary disease burden in the US. Extensive research has established exposure to air pollutants such as PM2.5 and ozone as cause or likely cause of increased CVD mortality and morbidity. While the US has made steady progress in reducing air pollution from industrial and vehicle emissions, recent increase in wildfires has slowed or even reversed this progress particularly in the Western US. Massive quantities of air pollutants in fire smoke can travel hundreds of kilometers affecting highly populated areas in states far from the original fire. There is an important public health need to establish the effects of smoke exposure, which has a distinctly different chemical composition from ambient air pollution, on CVD risks. Understanding the key distinctions between the cardiotoxic effects of smoke constituents is important to design effective emergency response measures to fire smoke, and to assess the long-term health care needs in the Western US. We will investigate the relationship between CVD risks and acute or chronic exposure to major air pollutants in smoke, including major PM2.5 constituents (i.e., sulfate, nitrate, organic carbon, elemental carbon), and gas pollutants such as ozone and formaldehyde (HCHO). Specifically, we will first estimate daily smoke and non-smoke PM2.5 constituents, ozone and HCHO at 1 km resolution from 2001 to 2020 in the Western US using EPA’s Community Multi-scale Air Quality Model (CMAQ) calibrated by ground- level EPA measurements and satellite data in a machine learning model. Second, we will link this rich exposure dataset with ED visits and hospitalization data from six Western US states to assess the adverse health effects of acute exposure to fire smoke on various CVD outcomes including acute myocardial infarction, stroke, heart failure, atrial fibrillation, and all CVD events. Third, we will link this exposure dataset with Medicare data in Western US to characterize the effect of chronic exposure (i.e., previous weeks, months or year) to fire smoke on these CVD outcomes. As a sub-analysis, we will develop an enhanced machine learning model by including available low-cost sensor measurements for 2016 – 2020 to estimate daily PM2.5 and use this dataset to conduct similar epidemiological analysis to the main analysis. Finally, we will enhance estimates of wildfire source fluxes and their impacts on atmospheric composition from 2002 to 2025 using deep-learning based algorithms, providing higher-quality atmospheric chemistry data to support health exposure assessments. With large study populations and high-quality exposure estimates, we expect to generate robust estimates of the associations of acute and chronic exposure to wildfire smoke and CVD risks, as well as economic burdens of CVD due to wildfire smoke.