Table 2.

Causal direct correlation (odds ratio and 95% CI) between air pollutants and the morbidity of overweight, obesity, and severe obesity^a: RD estimates.^b

Variables	AQI	PM2.5	PM10	SO2	CO	NO2	O3
Panel 1: Air pollutants on overweight (d = 2, polynomial = 3)c
OR(OW)	1.109***	1.173***	1.053***	0.972***	0.986***	1.052	1.026
95%CI	1.027–1.315	1.094–1.252	1.030–1.076	0.965–0.979	0.980–0.992	0.982–1.092	0.989–1.043
Panel 2: Air pollutants on obesity (d = 2, polynomial = 3)c
OR(OB)	1.032***	1.022***	1.008***	0.997***	0.998***	1.000	1.001
95%CI	1.006–1.217	1.016–1.028	1.006–1.010	0.996–0.998	0.997–0.999	0.998–1.002	0.997–1.002
Panel 3: Air pollutants on severe obesity (d = 2, polynomial = 3)c
OR(SO)	1.069***	1.035***	1.013***	0.994***	0.999***	1.005	1.004
95%CI	1.014–1.208	1.015–1.055	1.007–1.019	0.991–0.997	0.998–0.999	0.998–1.012	0.996–1.007

^a Individuals were defined as OW, OB, and SO if their BMI ≥ 24, BMI ≥28, and BMI ≥30, respectively. AQI was air quality index, OW was overweight, OB was obesity, and SO was severe obesity. ^b Selected the sample where d = 2, polynomial = 3 to report. All the covariates were controlled by command ‘covs’ in the ‘rdrobust,’ using a fuzzy RD model with an instrumental variable of whether the individual was located north of the QHB. ^c*p < 0.05, **p < 0.01, ***p < 0.001.