Skip to main content
. Author manuscript; available in PMC: 2022 Oct 1.
Published in final edited form as: Environ Res. 2021 Jun 24;201:111553. doi: 10.1016/j.envres.2021.111553

Table 5.

The adjusted associations between short-term exposure to air pollutants (prediction models) and temperature with the 5 factors of ICA (multi-pollutant models).

24 hour- window, PM2.5 24 hour- window, NO2 24 hour- window, O3 24 hour- window, Temperature
ICA Factors Beta (SE) P-value FDR P-value Beta (SE) P-value FDR P-value Beta (SE) P-value FDR P-value Beta (SE) P-value FDR P-value
ICA_factor1 0.003(0.006) 0.61 0.99 −0.002(0.003) 0.48 0.99 0.001(0.003) 0.78 0.99 0.001(0.005) 0.80 0.99
ICA_factor2 −0.001(0.007) 0.92 0.99 −0.019(0.004) 9.30E-07 5.58E-05 −0.001(0.003) 0.77 0.99 0.006(0.006) 0.27 0.99
ICA_factor3 0.015(0.006) 0.02 0.99 0.004(0.003) 0.21 0.99 −0.004(0.003) 0.20 0.99 0.004(0.005) 0.39 0.99
ICA_factor4 −0.001(0.006) 0.88 0.99 0.004(0.003) 0.27 0.99 −0.002(0.003) 0.49 0.99 −0.001(0.005) 0.81 0.99
ICA_factor5 0.002(0.005) 0.73 0.99 −0.002(0.003) 0.47 0.99 0.002(0.002) 0.52 0.99 −0.004(0.004) 0.33 0.99
7-day window, PM2.5 7-day window, NO2 7-day window, O3 7-day window, Temperature
ICA Factors Beta (SE) P-value FDR P-value Beta (SE) P-value FDR P-value Beta (SE) P-value FDR P-value Beta (SE) P-value FDR P-value
ICA_factor1 0.005(0.008) 0.53 0.99 −0.001(0.004) 0.87 0.99 −0.002(0.004) 0.65 0.99 0.004(0.005) 0.41 0.99
ICA_factor2 −0.002(0.011) 0.83 0.99 −0.019(0.004) 3.09E-05 0.002 −0.005(0.005) 0.33 0.99 0.009(0.007) 0.16 0.99
ICA_factor3 0.018(0.01) 0.06 0.99 0.011(0.004) 0.008 0.25 −0.008(0.004) 0.05 0.99 0.009(0.006) 0.12 0.99
ICA_factor4 −0.001(0.01) 0.89 0.99 0.008(0.004) 0.06 0.99 −0.0002(0.004) 0.95 0.99 −0.002(0.006) 0.68 0.99
ICA_factor5 0.0001(0.008) 0.99 0.99 −0.0001(0.004) 0.99 0.99 −0.001(0.003) 0.77 0.99 −0.0002(0.005) 0.96 0.99
30-day window, PM2.5 30-day window, NO2 30-day window, O3 30-day window, Temperature
ICA Factors Beta (SE) P-value FDR P-value Beta (SE) P-value FDR P-value Beta (SE) P-value FDR P-value Beta (SE) P-value FDR P-value
ICA_factor1 −0.0001(0.011) 0.99 0.99 −0.0001(0.004) 0.99 0.99 0.001(0.005) 0.76 0.99 0.003(0.005) 0.61 0.99
ICA_factor2 −0.003(0.014) 0.85 0.99 −0.02(0.005) 7.92E-05 0.005 −0.005(0.006) 0.44 0.99 0.009(0.007) 0.22 0.99
ICA_factor3 0.025(0.013) 0.05 0.99 0.008(0.005) 0.11 0.99 −0.013(0.005) 0.01 0.70 0.006(0.006) 0.32 0.99
ICA_factor4 −0.003(0.013) 0.82 0.99 0.009(0.005) 0.05 0.99 0.002(0.005) 0.67 0.99 −0.003(0.006) 0.66 0.99
ICA_factor5 −0.007(0.01) 0.51 0.99 −0.002(0.004) 0.68 0.99 −0.0004(0.004) 0.93 0.99 −0.003(0.005) 0.57 0.99

These models were linear mixed-effect regression models (LMEM) with random participant-specific intercepts and simultaneously adjusted for PM2.5, NO2, O3, and temperature (multi-pollutant models) for the same exposure window.

All models were adjusted for age (years), body mass index (kg/m2), cigarette pack-years, alcohol intake (< or ≥2 drinks per day), socioeconomic status (income payment and years of education), season (warm/cold), and relative humidity.

ICA_factor2 rank represents the rank of the corresponding metabolite contributing to factor 2 from the independent component analysis (ICA). Higher rank (and higher weights) means higher contribution to factor 2 of the ICA. We show along slide the weights and ranks of each metabolite that contributes to factor 2 from the ICA only because it was the only significant factor.

Range of facto2- ICA rank is from 1 to 1158 i.e., number of examined metabolites. Range of facto2- ICA weight is from −1 to 1.

Abbreviations: ENT: Effective/independent number of tests; PM2.5, particulate matter ≤2.5 micrometers; NO2, nitrogen dioxide; and O3, ozone; ICA, independent component analysis; FDR, Benjamini-Hochberg false discovery rate; SE, standard error.