Table 3.
Results of the sensitivity analyses for a 10μg/m3 increase in previous 3 years average of PM2.5 exposure (lag1-3)
| No. | Model* | B (95%CI)[mmol/L] |
|---|---|---|
| Model 1 | Main analysis | 0.146 (0.045,0.248) |
| Model 2 | Main analysis excluding age variable | 0.148 (0.047,0.250) |
| Model 3 | Main analysis excluding sex variable | 0.151 (0.049,0.253) |
| Model 4 | Main analysis excluding smoking status | 0.148 (0.047,0.250) |
| Model 5 | Main analysis excluding drinking status | 0.146 (0.044,0.248) |
| Model 6 | Main analysis excluding marital status | 0.148 (0.047,0.249) |
| Model 7 | Main analysis excluding income | 0.141 (0.040,0.242) |
| Model 8 | Main analysis excluding education years | −0.053 (−0.136,0.030) |
| Model 9 | Main analysis excluding exercise status | 0.149 (0.046,0.252) |
| Model 10 | Main analysis excluding staple food intake | 0.151 (0.050,0.252) |
| Model 11 | Main analysis excluding residence | 0.135 (0.037,0.233) |
| Model 12 | Main analysis excluding BMI | 0.146 (0.045,0.247) |
| Model 13 | Current smoking and drinking status | 0.144 (0.042,0.246) |
| Model 14 | Adding the intake of fruit | 0.144 (0.043,0.246) |
| Model 15 | Adding temperature and humidity | 0.260(0.144,0.375) |
| Model 16 | Linear regression model | 0.216 (0.112,0.319) |
*
In the main analysis we controlled for age, sex, marital status, education years, family income, residence, smoking and drinking status, exercise status, staple food intake, and BMI as the fixed effect. In the linear regression model, we excluded the random effects of individual from the main analysis, and changed the county as the fixed effect; other covariates were as the same as the main analysis.