key: cord-1014182-al6255yb
authors: Zhang, Xinhan; Tang, Mengling; Guo, Fanjia; Wei, Fang; Yu, Zhebin; Gao, Kai; Jin, Mingjuan; Wang, Jianbing; Chen, Kun
title: Associations between Air Pollution and COVID-19 epidemic during quarantine period in China()
date: 2020-10-20
journal: Environ Pollut
DOI: 10.1016/j.envpol.2020.115897
sha: a41ef6452bc0f3d25d536bd3bbec2397beee30f2
doc_id: 1014182
cord_uid: al6255yb

The coronavirus disease (COVID-19) has become a global public health threaten. A series of strict prevention and control measures were implemented in China, contributing to the improvement of air quality. In this study, we described the trend of air pollutant concentrations and the incidence of COVID-19 during the epidemic and applied generalized additive models (GAMs) to assess the association between short-term exposure to air pollution and daily confirmed cases of COVID-19 in 235 Chinese cities. Disease progression based on both onset and report dates as well as control measures as potential confounding were considered in the analyses. We found that stringent prevention and control measures intending to mitigate the spread of COVID-19, contributed to a significant decline in the concentrations of air pollutants except ozone (O(3)). Significant positive associations of short-term exposure to air pollutants, including particulate matter with diameters ≤2.5μm (PM(2.5)), particulate matter with diameters ≤10μm (PM(10)), and nitrogen dioxide (NO(2)) with daily new confirmed cases were observed during the epidemic. Per interquartile range (IQR) increase in PM(2.5) (lag0-15), PM(10) (lag0-15), and NO(2) (lag0-20) were associated with a 7% [95% confidence interval (CI): (4 - 9)], 6% [95% CI: (3 - 8)], and 19% [95% CI: (13 - 24)] increase in the counts of daily onset cases, respectively. Our results suggest that there is a statistically significant association between ambient air pollution and the spread of COVID-19. Thus, the quarantine measures can not only cut off the transmission of virus, but also retard the spread by improving ambient air quality, which might provide implications for the prevention and control of COVID-19.

Preliminary data edition was performed in association analysis to ensure that the authentic 123 associations between air pollution and COVID-19 infection were emerged. Firstly, we shift the new 124 confirmed cases to ten days earlier due to the country-level gap between the onset date and report 125 date (Epidemiology Working Group for NCIP Epidemic Response, 2020). Concerning the potential 126 lag effect between air pollution and infection, we calculated several combinations of moving average 127 concentrations of air pollutants (lag0-5, lag0-10, lag0-15, lag0-20, lag0-25). 128 and dependent variables. Briefly, the model was demonstrated as follows: 136

dewpoint temperature t,i + dow t + lockdown t + city i 138 were controlled for the possible confounding effect, including a smoothing spline of mean 148 temperature t,i (℃) with 3 degrees freedom (df2) and a linear term of mean dewpoint 149 temperature t,i (℃); (4) day of week t was included as dummy variable to control for the week effect; (5) 150 and lockdown t , effectively retarding the rapid spread of COVID-19, worked as categorical variable 151 adjusting the national lock (since Jan 23 th , 2020) and unlock (until 15 March 2020) effect; (6) city i 152 variable adjusted for the fixed effects such as spatial correlation and population density. S(.) mean 153 the smoothing spline to control for the non-linear relationship. Besides, relative risks (RRs) and 95% 154 confidence intervals (95%CIs) were calculated by the coefficients and standard variances from 155 into Hubei and other six Chinese natural areas (Middle south, East, Northeast, Southwest, Northwest, 158 North). 159

The elimination of 10-day lag between COVID-19 onset day and report day was performed as 160 sensitivity analyses to test the stability of the effects which was shown in supplementary. 161 Furthermore, we adjusted a smoothing spline of mean wind speed t,i (℃) with 3 degrees freedom in 162 the Poisson model ( Figure S4 Spearman correlations among AQI, all air pollutants and meteorological variables are shown in Table  185 S1. In 235 urban areas, strong correlations were found for AQI and During the quarantine (From Jan. 23 th to Mar. 11 th ), most regions witnessed substantial diminution of 196 ambient air pollutants except for O 3 . The AQI decreased by 12.7%, and five air pollutants (ie. PM 2.5 , 197 PM 10 , NO 2 , SO 2 and CO) decreased by 13.4%, 19.4%, 43.1%, 15.9%, 10.8%, respectively. There 198 was an obvious valley point for most regions from Feb. 14 th to Feb. 17 th . The concentration of air 199 pollutants after quarantine were well below the China national air quality standards (Table S3) . COVID-19 had a significant association with highly polluted cities, and transmission dynamics of 292 COVID-19 is due to mainly to the mechanism of airborne viral infectivity rather than 293 "human-to-human transmission" (measured with density of population)(Coccia, 2020), which might 294 explain the possible mechanisms for the association. 295

Our results suggested that exposure to ambient NO 2 may contribute to the spread of COVID-19. The discordance resulted in the overall negative association came from differed measures of control 336 and diversity of population density and medical resource in disparate areas. 337

To date, several studies have examined the association between ambient air pollution and the 338 incidence or mortality of COVID-19 around world, through time-series, regression or correlation test 339 (Table 3) . Although the studies has some notable differences as compared to our work, sepecially in 340 Our study still has several limitations. First, although 235 cities were covered in our analysis, 356 considering the different control measures the findings cannot be completely interpreted as extent to 357 other countries. Secondly, our study is an ecological study, which relied on aggregate data and did 358 not contain the demographic information. Thus, the problem of ecological fallacy was inevitable. No. of cases Cumulative Cases Date 1 -J a n 1 5 -J a n 2 9 -J a n 1 2 -F e b 2 6 -F Table S1 to S2; Figure S1 to S4. 575 576 

Associations of greenness with gestational diabetes mellitus: The Guangdong

Air quality during the COVID-19: PM2.5 481 analysis in the 50 most polluted capital cities in the world

Association of COVID-19 global 484 distribution and environmental and demographic factors: An updated three-month study

The impact of COVID-19 as a necessary evil on air 487 pollution in India during the lockdown