Journal of Infection and Public Health
A three-year study reveals a strong association between NO
By Rahul Raveendran Nair, Negin Nazari and Patrick R. Stephens.
Abstract
Background
Air pollution has been linked to adverse COVID-19 outcomes. However, studies of the US to-date have generally focused on the US as a whole or large cities on the east and west coasts. Few studies have considered the Midwest, and no comprehensive study of the four most common gaseous pollutants has considered Oklahoma. Oklahoma, a fossil fuel-dependent state with a relatively high respiratory disease burden, provides a unique context to evaluate pollutant impacts on COVID-19 case numbers.
Methods
We analyzed county-level COVID-19 case numbers (2020–2022) in Oklahoma in relation to satellite-derived concentrations of nitrogen dioxide (NO 2 ), sulfur dioxide (SO 2 ), ozone (O 3 ), and carbon monoxide (CO), adjusting for population density (PD), county area, and percentage of elderly. We employed boosted regression trees (BRT), random-effects regression trees (RE-EM tree), and hierarchical generalized additive models (GAM) to capture nonlinear associations between COVID-19 case numbers and the predictor variables.
Results
Among the confounding factors, PD had the strongest association with COVID-19 case numbers. Among the four gaseous pollutants we considered, all three models indicated that NO 2 had the strongest correlation with COVID-19 case numbers. CO also showed a significant negative linear association with COVID-19 case numbers, whereas positive non-linear relationships were observed for PD, NO 2 , O 3 , and area. NO 2 had the strongest association with case numbers when PD values ranged from 48 to 75.