key: cord-1011671-wjmu0p9u
authors: Jiang, Ying; Wu, Xiao-Jun; Guan, Yan-Jun
title: Effect of ambient air pollutants and meteorological variables on COVID-19 incidence
date: 2020-05-11
journal: Infection control and hospital epidemiology
DOI: 10.1017/ice.2020.222
sha: 4a15502a48c04209e37eae968aeb8f1d2116f644
doc_id: 1011671
cord_uid: wjmu0p9u

OBJECTIVE: To determine whether ambient air pollutants and meteorological variables are associated with daily COVID-19 incidence. DESIGN: A retrospective cohort from January 25 to February 29, 2020. SETTING: Cities of Wuhan, Xiaogan, and Huanggang, China. PATIENTS: The COVID-19 cases detected each day. METHODS: We collected daily data of COVID-19 incidence, 8 ambient air pollutants (particulate matter of ≤2.5 µm [PM(2.5)], particulate matter ≤10 µm [PM(10)], sulfur dioxide [SO(2)], carbon monoxide [CO], nitrogen dioxide [NO(2)], and maximum 8-h moving average concentrations for ozone [O(3)-8h]) and 3 meteorological variables (temperature, relative humidity, and wind) in China’s 3 worst COVID-19–stricken cities during the study period. The multivariate Poisson regression was performed to understand their correlation. RESULTS: Daily COVID-19 incidence was positively associated with PM(2.5) and humidity in all cities. Specifically, the relative risk (RR) of PM(2.5) for daily COVID-19 incidences were 1.036 (95% confidence interval [CI], 1.032–1.039) in Wuhan, 1.059 (95% CI, 1.046–1.072) in Xiaogan, and 1.144 (95% CI, 1.12–1.169) in Huanggang. The RR of humidity for daily COVID-19 incidence was consistently lower than that of PM(2.5), and this difference ranged from 0.027 to 0.111. Moreover, PM(10) and temperature also exhibited a notable correlation with daily COVID-19 incidence, but in a negative pattern The RR of PM(10) for daily COVID-19 incidence ranged from 0.915 (95% CI, 0.896–0.934) to 0.961 (95% CI, 0.95–0.972, while that of temperature ranged from 0.738 (95% CI, 0.717–0.759) to 0.969 (95% CI, 0.966–0.973). CONCLUSIONS: Our data show that PM(2.5) and humidity are substantially associated with an increased risk of COVID-19 and that PM(10) and temperature are substantially associated with a decreased risk of COVID-19.

virus particles can attach. 11 This hypothesis is supported by the influenza-PM 2.5 correlation. 10, 12, 13 Although SARS-CoV-2 is known to be transmitted human to human by infectious secretions, 3 these secretions can be transferred in many different ways. Thus, whether ambient air pollutants could affect the transmission of SARS-CoV-2 is an urgent question. Additionally, meteorological factors such as humidity and temperature have also been suggested to enhance air pollution 14, 15 and to promote transmission of infectious disease. 14 Their role in SARS-CoV-2 transmission, however, remains largely unknown.

Wuhan is one of the most populated cities in China, and it has a subtropical humid monsoon climate. Due to rapid industrialization and urbanization, both the concentration and composition of air pollutants have become higher and more complicated, 16, 17 and pollution has already caused severe health problems for the local population. 18 A previous study demonstrated that a tuberculosis surge in Wuhan was associated with worsening air pollution and weather. 19 Nevertheless, no study has explored the association of Wuhan local air quality and/or meteorological data with COVID-19 incidence.

Due to the severe air pollution, high COVID-19, and the citywide lockdown, Wuhan is a good research setting in which to investigate this issue. We hypothesized the potential association of air pollutants and meteorological variables with daily COVID-19 incidence, and we sought to discover factors influencing the SARS-CoV-2 pandemic. We also enrolled Xiaogan and Huanggang in the current study because they were the second-and third-worst COVID-19-stricken cities in the province of Wuhan, respectively.

The daily air quality index (AQI) and average concentration data for 6 ambient air pollutants data were collected from the AQI platform website (https://www.aqistudy.cn). These pollutants included particulate matter of ≤2.5 μm (PM 2.5 ), particulate matter ≤10 μm (PM 10 ), sulfur dioxide (SO 2 ]), carbon monoxide (CO), nitrogen dioxide (NO 2 ), and maximum 8-h moving average concentrations for ozone (O 3 -8h). We calculated their daily concentrations by averaging the hourly concentrations from all stations in the city to represent citywide pollution exposure. The website uses the standard for particulate matter from the US Environmental Protection Agency (EPA). 20 The daily temperature, relative humidity, and wind (in Beaufort number) were also obtained from the AQI platform, and we averaged the hourly data from all stations in the city to represent the citywide weather condition.

The study population included a retrospective cohort of COVID-19 patients in Wuhan, Xiaogan and Huanggang, China, between January 25 and February 29, 2020. The population of each city was obtained from the China National Bureau of Statistics. Because all 3 cities were under a citywide lockdown during the entire study period, the population size was considered stable.

The daily COVID-19 incidence was obtained from the website of Health Commission of Hubei Province from January 25, 2020, onward. 21 The COVID-19 diagnoses were made solely based on nucleic acid test from January 25 to February 11. Beginning February 12, clinical symptoms along with chest computed tomography (CT) scans were used for COVID-19 diagnoses. 22, 23 When COVID-19 was diagnosed, the patient was immediately placed in a single isolation ward. These remained in quarantine until they fully recovered or died. A previous epidemiological study demonstrated that the median incubation period of SARS-CoV-2 is 4 days. 24 Thus, we applied this correlation to analyze the relationship between COVID-19 incidence per day and the air pollution data 4 days prior.

We conducted a time-series analysis to examine the associations of the pollutants and meteorological variables with COVID-19 incidence per day. We first performed a descriptive analysis that provided the information for COVID-19 incidence, 6 air pollutants, and 3 meteorological variables in all 3 cities. We then used multivariate Poisson regression models to evaluate the association of pollutants and meteorological variables with COVID-19 incidence in 3 cities on a daily basis. All analyses were conducted using SPSS version 20.0 software (SPSS, Chicago, IL). Bonferroni correction was used for data correction, and a P value <.0056 was considered as statistically significant. Table 1 lists the descriptive statistics for daily COVID-19 incidence, air pollutant concentration, and meteorological variables in Wuhan, Xiaogan, and Huanggang. Starting January 26, Wuhan had 80 newly diagnosed COVID-19 cases, Xiaogan had 45, and Huanggang had 32. All cities reached their peak daily incidence between February 12 and 13. Thereafter, the daily incidences started to drop, and they reached their lowest by the end of the study.

Next, multiple linear relationships were used to evaluate the association of the daily COVID-19 incidence with air pollutants and meteorological variables in each city ( Table 2 ). Due to the change in case definition on February 12, the COVID-19 incidence surged on February 12 and 13. 22, 23 Additionally, Xiaogan and Huanggang showed negative COVID-19 incidences on February 17 and 19. Thus, we excluded data from these dates in our multiple linear regression analysis. First, in Wuhan (Table 2 ), all air pollutants (except SO 2 ) were strongly associated with daily COVID-19 incidence (all P < .001). Among them, PM 10 exhibited a negative association with COVID-19 incidence (relative risk [RR], 0.964; 95% confidence interval [CI], 0.961-0.967). Among the remaining pollutants, PM 2.5 was a main effector on incidence increase (RR, 1.036; 95% CI, 1.032-1.039). Among the 3 meteorological variables, temperature (RR, 0.969; 95% CI, 0.966-0.973) and humidity (RR, 1.009; 95% CI, 1.007-1.011) were significantly associated with daily COVID-19 incidence (both P < .001).

In XiaoGan, all air pollutants were notably associated with daily COVID-19 incidence (all P < .001). Similar to that of Wuhan, both PM 10 (RR, 0.961; 95% CI, 0.95-0.972) and temperature (RR, 0.89; 95% CI, 0.871-0.911) remained negatively correlated with daily COVID-19 incidence. Among the remaining air pollutants, COVID-19 incidence increase was also significantly associated with PM 2.5 (RR, 1.059; 95% CI, 1.046-1.072). For 3 meteorological variables, humidity (RR, 1.013; 95% CI, 1.007-1.019) and temperature (RR, 0.89; 95% CI, 0.871-0.911) were the 2 main factors associated with COVID-19 incidence.

In Huanggang, all ambient air pollutants (expect NO 2 ) were significantly associated with daily COVID-19 incidence (all P < .001). Similar to Wuhan and Xiaogan, PM 10 remained negatively correlated with the daily COVID-19 incidence (RR, 0.915; 95% CI, 0.896-0.934). Among the remaining air pollutants, PM 2.5 had the strongest correlation with COVID-19 incidence (RR, 1.144; 95% CI, 1.12-1.169). All meteorological variables were also significantly associated with daily COVID-19 incidence (both P < .001). Similar to the previous 2 cities, a negative association was detected between COVID-19 and temperature (RR, 0.738; 95% CI, 0.717-0.759) and between COVID-19 and wind level (RR, 0.79; 95% CI, 0.697-0.896).

Ambient air pollution and meteorological variables can affect viral transmission; however, they have not yet been examined for the presence of SARS-CoV-2. In the current study, we have provided an initial assessment of the potential effects of ambient air pollutants and meteorological variables on the incidence of COVID-19 day by day from January 25 to February 29, 2020. Overall, based on data from China's 3 worst COVID-19-stricken cities, COVID-19 incidence was negatively associated with PM 10 and temperature.

Meanwhile, a positive association was observed between COVID-19 and PM 2.5 and relative humidity.

Particulate matter is a mixture of both solid particles and liquid droplets suspended in the air. Although it is not the sole cause of respiratory illness, previous studies have demonstrated that particulate matter is a strong environmental determinant of viral transmissions. For example, a close correlation was discovered between human influenza cases and PM 2.5 concentrations in epidemiological studies based in big cities such as Beijing, 12 Hong Kong, 25 and Brisbane. 26 Its potential mechanisms are believed to damage bronchial immunity 27 and epithelial cell integrity, 28 which enhances viral attachment to and replication in the bronchus. 28 In the current study, although our data could not reveal whether particulate matter promotes SARS-CoV-2 transmission, we observed that PM 2.5 is the only air pollutant that is consistently associated with increased COVID-19 incidence in all cities. This finding agrees with the data from a previous influenza study. Contrary to PM 2.5 , increased PM 10 concentration was associated with decreased COVID-19 incidence. We believe this phenomenon to be a diameter-related effect. As shown by previous study, the receptor for both SARS-CoV-2 and SARS-CoV binding is the angiotensin-converting enzyme 2, which concentrates on type II alveolar cells. 29 However, type II alveolar cells are located in the alveoli, which can only be reached by particles with a diameter <5 μm. 30 Thus, small airborne pollutants, such as PM 2.5 , are particularly harmful because they are more likely to penetrate the respiratory tract all the way to the alveolar region unfiltered. 31, 32 For larger particulate matter, penetration ability decreases dramatically when size increases; starting at 20 μm and beyond, particulate matter is not able to penetrate below the trachea. 32 Although all particulate matter might provide condensation nuclei for viral attachment, 11 PM 2.5 deliver SARS-CoV-2 to their target cells in the alveoli, which is unreachable to PM 10 . This size factor might also explain our finding that PM 2.5 correlates with COVID-19 incidence in a positive pattern while that of PM 10 is negative. Future epidemiological studies in other countries and areas with similar population density but different PM 2.5 and PM 10 concentrations could be used to validate this result. So far, a number of studies have indicated a negative effect of ambient temperature and humidity on the viral viability and transmission. For example, low humidity increases influenza viability in the aerosol 30 and impairs the innate antiviral defense of the host, 33 which promotes influenza transmission. Meanwhile, temperature has the same effect as humidity on influenza. 34 Similar to influenza, epidemiological studies have also demonstrated that the transmission of SARS-CoV-2 is more efficient in low temperatures and at low humidity. 35, 36 This hypothesis was supported by the results from a laboratory study; SARS-CoV was more likely to be inactivated in higher temperature and humidity. 37, 38 In the current study, our data agree with previous findings that temperature is negatively associated with COVID-19 incidence. However, our results regarding humidity do not agree with those from previous studies. Recently, an outbreak of COVID-19 was reported in a public bath center in China. 39 Since the bath center was extremely humid, SARS-CoV-2 should have been inactivated quickly; however, this was not the case. Thus, our results support this finding and suggest that the study of differences between SARS-CoV-2 and other viruses is warranted in future studies.

Whether SARS-CoV-2 can be transmitted via airborne droplets remains controversial. 40 Recently, Luo et al 39 reported a COVID-19 outbreak in a public bath center in China, which suggests the role of ambient droplet in SARS-CoV-2 transmission in conditions of high relative humidity. Later, van Doremalen et al 41 also demonstrated that, under laboratory conditions, SARS-CoV-2 is viable in aerosol droplets for at least 3 hours, which provides theoretical evidence for how SARS-CoV-2 transmission in that environment. However, some researchers dispute ambient droplet transmission because SARS-CoV-2 is not present in air samples obtained from the rooms of hospitalized COVID-19 patients. 41 Most studies thus far have also reported that healthy people are infected by touching contaminated objects 40 and that SARS-CoV-2 is stable on copper, cardboard, plastic, and stainless-steel surfaces for up to 72 hours. 41 In the current study, we have shown that ambient air pollutants, especially PM 2.5 , are closely associated with COVID-19 incidence, which supports the hypothesis regarding airborne droplet transmission of SARS-CoV-2. However, the viability of SARS-CoV-2 attaching to PM 2.5 remains largely unknown. Further studies are required to investigate the extent of SARS-CoV-2 transmission via airborne droplets.

Winter features overlap multiple viral seasons at various degrees. Previous studies have suggested that epidemics of cocirculating viral infections can interfere with each other. 42, 43 For example, influenza A could overlap coronavirus outbreaks. 42 Thus, it is reasonable to hypothesize that COVID-19 incidence might also be significantly affected by other cocirculating viruses. Additionally, 1 study concluded that 2 SARS-CoV-2 strains are cocirculating in Italy. 44 Whether these 2 strains affect the infectious ability of the other remains unknown. Further study is required to fully elucidate this issue.

The current study has several limitations. First, because the ambient PM 2.5 is a mixture of solid particles and liquid droplets, the exact components of PM 2.5 that may promote coronavirus transmission remain unknown. Second, due to the relatively short study period and imperfect daily reporting practices, our results are vulnerable to changes resulting from the emergence of more detailed data. Third, we studied only 9 variables in the current study. A number of other variables could potentially have affected COVID-19 transmission; thus, our study can only provide preliminary information on the association between COVID-19 and air pollutants or other meteorological variables. In a future study, we plan to gather data from more countries and regions to remedy these data gaps. Fourth, because detailed information from confirmed COVID-19 patients was not available, we were unable to determine how underlying health problems might promote COVID-19 infection. Thus, a longer study period might have affected our results. Moreover, winter overlapped the entire study period, which limited all meteorological variables (especially temperature) to a narrow band. Thus, further study is urgently needed further elucidate these issues.

In conclusion, our findings consistently suggest that increased temperature and airborne PM 2.5 concentration are associated with increased daily COVID-19 incidence. Meanwhile, a negative association was observed between COVID-19 incidence and relative humidity and airborne PM 10 concentration. In the context of the worldwide COVID-19 pandemic, our findings might provide some information for COVID-19 personal protection and transmission reduction.

A novel coronavirus from patients with pneumonia in China

Going global-travel and the 2019 novel coronavirus

Early transmission dynamics in Wuhan, China, of novel coronavirus-infected pneumonia

Coronavirus disease 2019 (COVID-19), 2020. Centers for Disease Control and Prevention website

Ambient particulate air pollution and daily mortality in 652 cities

Health effects of particulate air pollution: a review of epidemiological evidence

Short-term exposure to ambient fine particulate matter and out-of-hospital cardiac arrest: a nationwide case-crossover study in Japan

On the association between outdoor PM(2.5) concentration and the seasonality of tuberculosis for Beijing and Hong Kong

Short-term elevation of fine particulate matter air pollution and acute lower respiratory infection

The short-term effects of air pollutants on influenza-like illness in Jinan, China

Exposure to combustion generated environmentally persistent free radicals enhances severity of influenza virus infection

PM 2.5 in Beijing-temporal pattern and its association with influenza

Impact of ambient fine particulate matter (PM 2.5 ) exposure on the risk of influenza-like-illness: a time-series analysis in Beijing

Seasonality of respiratory viral infections

Estimation of short-term effects of air pollution on stroke hospital admissions in Wuhan

Causes of ozone pollution in summer in Wuhan, central China

Characterization and source identification of PM( 2.5 )-bound polycyclic aromatic hydrocarbons in urban, suburban, and rural ambient air, central China during summer harvest

Association of daily cause-specific mortality with ambient particle air pollution in Wuhan, China

Seasonal variation of newly notified pulmonary tuberculosis cases from

What are the Air Quality Standards for PM? US Enviromental Protection Agency website

The realtime coronovirus epidemiological situationof Hubei province

COVID-19: China once again changes case definition in Hubei Province

How many coronavirus cases in China? Officials tweak the answer. The New York Times website

Clinical characteristics of coronavirus disease 2019 in China

Effects of coarse particulate matter on emergency hospital admissions for respiratory diseases: a time-series analysis in Hong Kong

Air pollution, temperature, and pediatric influenza in

Long-term exposure to PM 2.5 lowers influenza virus resistance via down-regulating pulmonary macrophage Kdm6a and mediates histone modification in IL-6 and IFN-β promoter regions

Diesel exhaust enhances influenza virus infections in respiratory epithelial cells

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Transmission of influenza A in human beings

Aerosol transmission of influenza A virus: a review of new studies

Aerosol Technology, second ed

Low ambient humidity impairs barrier function and innate resistance against influenza infection

Influenza virus transmission is dependent on relative humidity and temperature

COVID-19 transmission in mainland China is associated with temperature and humidity: a time-series analysis

Impact of meteorological factors on the COVID-19 transmission: a multicity study in China

Effects of air temperature and relative humidity on coronavirus survival on surfaces

The effects of temperature and relative humidity on the viability of the SARS coronavirus

Possible transmission of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in a public bath center in Huai'an

Aerosol and surface stability of SARS-CoV-2 as compared with SARS-CoV-1

Early occurrence of influenza A epidemics coincided with changes in occurrence of other respiratory virus infections

Possible interference between seasonal epidemics of influenza and other respiratory viruses in Hong Kong

Whole genome and phylogenetic analysis of two SARS-CoV-2 strains isolated in Italy in

Acknowledgments. We thank all medical staff and workers who risk their lives to save the general population from SARS-CoV-2.Financial support. No financial support was obtained for the current study.Conflicts of interest. All authors report no conflicts of interest relevant to this article.