. 2013 Oct 18;8(10):e76759. doi: 10.1371/journal.pone.0076759

Table 3. Percent increase of daily mortality associated with an IQR increase of CO, NO₂ and PM₁₀ with single model and principal component analysis in Beijing (mean and 95% CI), using 8 df/year.^a .

	Nonaccidental mortality	Cardiovascular mortality	Respiratory mortality
Single model
CO	2.55 (1.99, 3.11)^*	2.88 (2.10, 3.65)^*	2.39 (0.68, 4.09)^*
NO₂	2.54 (2.00, 3.08)^*	2.63 (1.87, 3.39)^*	1.79 (0.11, 3.47)^*
PM₁₀	1.80 (1.21, 2.40)^*	1.72 (0.88, 2.55)^*	2.07 (0.21, 3.92)^*
After-adjusting collinearity by principal component analysis
CO	0.97 (0.77, 1.17)^*	1.01 (0.73, 1.29)^*	0.89 (0.27, 1.51)^*
NO₂	1.04 (0.82, 1.25)^*	1.08 (0.78, 1.38)^*	0.95 (0.29, 1.61)^*
PM₁₀	1.07 (0.85, 1.30)^*	1.12 (0.81, 1.43)^*	0.99 (0.30, 1.67)^*

We applied current-day (lag 0 day) temperature and relative humidity and 2-day moving average of air pollutant concentrations (lag01), and applied 8 df per year for time, 3 df to temperature, humidity and barometric pressure.

P<0.05.

Table 3. Percent increase of daily mortality associated with an IQR increase of CO, NO2 and PM10 with single model and principal component analysis in Beijing (mean and 95% CI), using 8 df/year.a .

Table 3. Percent increase of daily mortality associated with an IQR increase of CO, NO₂ and PM₁₀ with single model and principal component analysis in Beijing (mean and 95% CI), using 8 df/year.^a .