Figure 3.

Bar graphs showing the association of BDR group with chest CT scan findings in participants with GOLD stage 0 disease (ie, normal spirometry findings; n = 1,481). We categorized tobacco-exposed participants with normal spirometry findings based on BDR variability into three groups: consistent BDR when it is present at every visit; inconsistent BDR when it is present at some, but not all, visits; and never BDR when it is not present at any visit. Multivariate linear regression models with BDR group as the independent variable and Pi10, % PRM^fSAD, % emphysema, and % gas trapping as the dependent variables. All models included the following covariates: age, sex, race, smoking status and pack-years smoked, and FEV₁ % predicted after bronchodilator administration at first visit, as well as number of visits. Based on these models, we calculated the least square mean (LSM). Pairwise comparisons using Tukey’s method correction for LSM were used. Values in the figures are presented as LSM with 95% CI. ^aP < .05. ^bP < .01. ^cP < .001. ATS-BDR = increase in FEV₁, FVC, or both of ≥ 12% and ≥ 200 mL after bronchodilator administration; BDR = bronchodilator responsiveness; FEV₁-BDR = increase in FEV₁ of ≥ 12% and ≥ 200 mL after bronchodilator administration; FVC-BDR = increase in FVC of ≥ 12% and ≥ 200 mL after bronchodilator administration; Pi10 = square root of the airway wall area for a hypothetical airway with an internal perimeter of 10 mm; PRM^fSAD = parametric response mapping functional small airways disease.