Evidence regarding environmental factors associated with disease severity of COVID-19 remained scarce. This study aimed to investigate the association of residential greenness exposure with COVID-19 severity applying a retrospective cross-sectional study in Wuhan, China. We included 30,253 COVID-19 cases aged over 45 years from January 1 to February 27, 2020. Residential greenness was quantitatively assessed using normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI). A multilevel generalized linear model using Poisson regression was implemented to analyze the association between greenness exposure and disease severity of COVID-19, after adjusting for potential covariates. A linear exposure-response relationship was found between greenness and COVID-19 severity. In the adjusted model, one 0.1 unit increase of NDVI and EVI in the 1000-m buffer radius was significantly associated with a 7.6% (95% confidence interval (CI): 4.0%, 11.1%) and 10.0% (95% CI: 5.1%, 14.7%) reduction of the prevalence of COVID-19 severity, respectively. The effect of residential greenness seemed to be more pronounced among participants with lower population density and economic levels. Air pollutants mediated 0.82~12.08% of the greenness and COVID-19 severity association, particularly to nitrogen dioxide. Sensitivity analyses suggested the robustness of the results. Our findings suggested that residential greenness exposure was beneficial to reduce the prevalence of COVID-19 severity. Graphical Abstract ga1 Highlights • The study is the first to discuss the associations of greenness with COVID-19 severity. • Greenness was associated with reduction of the prevalence of COVID-19 severity. • The exposure-response relation between greenness and COVID-19 severity was linear. • Air pollutants mediated the associations of greenness with COVID-19 severity.
【저자키워드】 COVID-19, disease severity, air pollutant, greenness, 【초록키워드】 severity, COVID-19 severity, covariates, association, Analysis, Generalized linear model, reduction, 95% confidence interval, COVID-19 case, participant, Poisson regression, Retrospective cross-sectional study, buffer, environmental factor, Wuhan, China, significantly, remained, linear, adjusted, suggested, reduce, normalized, prevalence of COVID-19, vegetation, with COVID-19, 【제목키워드】 COVID-19 patient, Wuhan, China,