Supplementary Material for: Impacts of air pollution on dynamic lung function in asthma patients: the modifying effect of phenotype susceptibility and possible relationship with airway microbiota

Introduction
Air pollution and meteorological factors have consistently been reported to adversely affect asthma patients.
Methods
We used a linear mixed-effects model to explore the relationship between air pollution and the dynamic lung function of 58 adult asthma patients with different asthma phenotypes in Beijing, China. We conducted a follow-up panel study of these patients with repeated lung function tests every day in the morning and evening by supervised spirometry from November 2020 to December 2021. Induced sputum from these subjects was collected and analyzed for the microbiome composition and associations between microbiome and lung function indices.
Results
We found that a 10 μg/m3 increase in PM2.5 was associated with decreases 0.79% in FEF25-75 (95% CI: 0.31 ~ 3.25), a range of decreases of 0.76%~0.83% PEF, 1.58%~1.93% for FEF25 and 1.79%~2.31% for FEF50 in the morning or evening on different lag days.
Compared with the PM2.5 effect in the fall, PM2.5 in spring and winter had significant effects on FEV3, FEsV6, and FVC. O3 had significant effects on FEF25, FEF50, FEV3, FEsV6, FVC and FEV1/FEV1-predicted in summer. Significant reduction of lung function indices in range of 0.31% to 1.29% reduction for lung function indices associated with the summer Rh or spring Rh compared with fall Rh.
PM2.5 had a larger adverse effect on Th2 phenotype asthmatic patients than on non-Th2 phenotype asthmatic patients. For a 10 μg/m3 increase in PM2.5, there was a 0.75% (95% CI: 0.026%, 1.52%), 0.70% (95% CI: 0.010%, 1.50%),0.75% (95% CI: 0.013%, 1.49%) and 0.37 L (95% CI: 0.062%, 0.80%) on FEF50 (evening), FEF50 (morning), FEF75 (evening) and FEV1 (morning), respectively for the Th2 phenotype compared with those for the non-Th2 phenotype.
Significant differences in the sputum microbiome composition were observed between the two inflammatory phenotypes. The linear relationships between sputum microbiome and lung function indices were observed.
Conclusion
Our study demonstrated the possibility of phenotype–environment interactions.