Table 3.

Summary of studies on associations between exposures to air pollution and COVID-19 incidence or mortality in different regions.

Study	Study location	Time period	Methodology	Statistics	Air pollutants	Estimate
Zhu et al. (2020)	120 cities in China	Jan. 23 to Feb. 29, 2020	Time series, Generalized Additive Model (GAM)	Percentage daily confirmed cases change of 10-μg/m3 increase on lag 0-14	PM2.5	2.24% (95% CI: 1.02 to 3.45)
					PM10	1.76% (95% CI: 0.89 to 2.63)
					NO2	6.94% (95% CI: 2.38 to 11.51)
					O3	4.76% (95% CI: 1.99 to 7.52)
					SO2	7.79% (95% CI: −14.57 to −1.01)
Xu et al. (2020)	33 cities in China	Jan. 29 to Feb. 15, 2020	Time-series Poisson regression model	RR on confirmed cases for each unit on lag 3	AQI	1.0008 (95% CI:1.0003, 1.0012)
Bashir et al. (2020)	California, USA	Mar. 4 to Apr. 24, 2020	Spearman rank correlation tests	Correlation coefficient (R) on daily confirmed cases and deaths	PM2.5	−0.267a; −0.350b
					PM10	−0.339a; −0.429b
					NO2	−0.485a; −0.731b
					SO2	−0.309a; −0.397b
Hendryx and Luo (2020)	3143 counties in the US	Data up to May 31, 2020	Linear multiple regression	Estimate (SE) on prevalence and death rates	PM2.5	Prevalence: 23.5 (10.3) ∗ Death rates: 1.08 (0.54) ∗
Fattorini and Regoli (2020)	71 provinces in Italy	Data up to Apr. 27, 2020	Pearson correlation	coefficient of determination (R2)	PM2.5	0.340∗∗
					PM10	0.267∗
					NO2	0.247∗∗
					O3	0.264∗
Zoran et al. (2020)	Milan, Italy	Jan. 1 to Apr. 25, 2020	Pearson coefficient correlation	Coefficient (R) on daily new cases	PM2.5	0.25
					PM10	0.35
Zoran et al. (2020)	Milan, Italy	Jan. 1 to Apr. 30, 2020	Pearson coefficient correlation	Coefficient (R) on daily new cases	NO2	−0.35∗∗
					O3	0.50∗∗
Ogen (2020)	66 regions in European countries (Italy, Spain, France and Germany)	up to Mar. 19, 2020	Ecological, descriptive analysis	/	NO2	83% of all fatalities were associated with NO2 > 100 μmol/m2

NOTE: a: on daily confirmed cases; b: on deaths; ∗: p < 0.05; ∗∗: p < 0.01.