Oscillometry-defined Small Airway Dysfunction in Tobacco-exposed Adults with Impaired or Preserved Airflow

Abstract

Rationale: Small airway dysfunction (SAD) is a key feature of chronic obstructive pulmonary disease and might present in tobacco-exposed adults with normal spirometry. So far, the role of oscillometry-defined SAD in this population is largely unexplored. Objective: To investigate the prevalence of oscillometry-defined SAD and its associations with airway structural changes, quality of life (QoL), metabolic disease, and cardiovascular disease (CVD) in tobacco-exposed adults with impaired airflow or preserved airflow (PA). Methods: In a subcohort (n = 1,628) nested within a lung cancer screening trial, we assessed airway disease using pre-bronchodilator spirometry, oscillometry, and artificial intelligence-powered computed tomography. Impaired airflow included airflow obstruction (AFO) and preserved ratio impaired spirometry (PRISm). Subjects with PA, defined as FEV₁ and FEV₁:FVC greater than the lower limit of normal, were further stratified as PA with SAD (PA-SAD) or normal lung function. SAD was defined as the frequency dependence of resistance or reactance area greater than the upper limit of normal. Computed tomography biomarkers included airway wall thickness, luminal diameter, branch count, and emphysema. QoL was measured using the euroqol 5-dimension 5-level (EQ-5D-5L). Measurements and Main Results: The overall prevalence of SAD was 39%. SAD was present in 26% of subjects with PA and in 60% of those with impaired airflow. The frequency of AFO, PRISm, and PA-SAD was 21%, 15%, and 16%, respectively. Similar to those with impaired airflow, subjects with PA-SAD had lower EQ-5D-5L scores, greater airway wall thickness, narrower lumen, lower branch count, and higher rate of metabolic disease and CVD than those with normal lung function (P < 0.01 for all). However, they had minimal emphysema and significantly higher branch count than those with AFO. Subjects with AFO or PRISm and concurrent SAD had greater structural changes and more frequent CVD than those with AFO or PRISm alone. SAD was associated with CVD (odds ratio, 1.91 [95% confidence interval, 1.55-2.36]), even after adjusting for confounders and metabolic disease. Conclusions: SAD is highly prevalent among tobacco-exposed adults and is associated with airway structural changes, impaired QoL, and an increased rate of CVD, even among those with PA. PA-SAD is distinct from AFO by its preserved airway count and minimal emphysema.

Keywords: PRISm; SAD; airway wall thickness; emphysema; oscillometry.

Figure 1.

Ridgeline plots comparing key computed tomography markers across the study groups. (A) AWT-Pi10 (mm) was higher in subjects with airflow obstruction (AFO), preserved ratio impaired spirometry (PRISm), and preserved airflow with small airway dysfunction (PA-SAD) (4.08 ± 0.72, 4.05 ± 0.70, and 3.83 ± 0.68, respectively) compared with those with normal lung function (3.46 ± 0.66), with all post hoc P values <0.05. (B) Luminal diameter was lower in subjects with AFO, PRISm, and PA-SAD (3.0 ± 0.27, 3.07 ± 0.31, and 3.12 ± 0.32, respectively) compared with those with normal lung function (3.29 ± 0.34), with all P values <0.05. The mean airway luminal diameter (mm) was narrower in AFO than in PA-SAD (P < 0.05) but did not differ between PRISm and PA-SAD or between PRISm and AFO (P > 0.05). (C) Total branch count was significantly lower in AFO, PRISm, and PA-SAD (471 ± 170, 566 ± 230, and 599 ± 212, respectively) compared with normal lung function (642 ± 183), with all P values <0.05. The total branch count was significantly lower in AFO than in PRISm or PA-SAD but did not differ between PRISm and PA-SAD (P > 0.05). (D) The mean percentage of low-attenuation areas with attenuation less than −950 Hounsfield units (LAA < −950%) was higher in AFO (11%) compared with all other groups (4.6–4.7%) (P < 0.05 for all). There was no difference in LAA < −950% among PRISm, PA-SAD, and normal lung function. AWT = airway wall thickness; Pi10 = theoretical internal luminal perimeter of 10 mm of the whole bronchial tree.

Figure 2.

(A and B) Boxplots demonstrating (A) airway wall thickness compared with a theoretical internal luminal perimeter of 10 mm of the whole bronchial tree (AWT-Pi10, mm) and (B) luminal diameter in patients with airflow obstruction (AFO) or preserved ratio impaired spirometry (PRISm) stratified according to the coexistence of oscillometry-defined small airway dysfunction (SAD). (A) Mean AWT-Pi10 (mm) was greater in subjects with AFO and SAD compared with those with AFO alone (4.36 ± 0.70 vs. 3.67 ± 0.54). Similarly, AWT-Pi10 (mm) was greater in PRISm with SAD compared with PRISm alone (4.19 ± 0.71 vs. 3.83 ± 0.64). (B) Mean airway luminal diameter (mm) was smaller in subjects with AFO and SAD compared with those with AFO alone (2.91 ± 0.25 vs. 3.13 ± 0.25). In PRISm with SAD, luminal diameter was also smaller compared with PRISm alone (3.01 ± 0.30 vs. 3.16 ± 0.31). ***P < 0.001.

Figure 3.

Heatmap showing the correlations between computed tomography markers of airway disease and emphysema with lung function markers from spirometry and oscillometry. Oscillometry-derived markers are not correlated with emphysema markers, such as the percentage of low attenuation or mean lung attenuation in HU, but instead correlate with airway thickening and luminal diameter markers. AWT-Pi10 = airway wall thickness compared with a theoretical internal luminal perimeter of 10 mm of the whole bronchial tree; AX = area under the reactance curve; FDR = frequency dependence of resistance; Fres = resonant frequency; HU = Hounsfield units; LAA = low-attenuation area; R5 = airway resistance measured at 5 Hz; X5 = airway reactance measured at 5 Hz.