Small Airway Wall Thickening Assessed by Computerized Tomography Is Associated With Low Lung Function in Chinese Carbon Black Packers

Abstract

Nanoscale carbon black as virtually pure elemental carbon can deposit deep in the lungs and cause pulmonary injury. Airway remodeling assessed using computed tomography (CT) correlates well with spirometry in patients with obstructive lung diseases. Structural airway changes caused by carbon black exposure remain unknown. Wall and lumen areas of sixth and ninth generations of airways in 4 lobes were quantified using end-inhalation CT scans in 58 current carbon black packers (CBPs) and 95 non-CBPs. Carbon content in airway macrophage (CCAM) in sputum was quantified to assess the dose-response. Environmental monitoring and CCAM showed a much higher level of elemental carbon exposure in CBPs, which was associated with higher wall area and lower lumen area with no change in total airway area for either airway generation. This suggested small airway wall thickening is a major feature of airway remodeling in CBPs. When compared with wall or lumen areas, wall area percent (WA%) was not affected by subject characteristics or lobar location and had greater measurement reproducibility. The effect of carbon black exposure status on WA% did not differ by lobes. CCAM was associated with WA% in a dose-dependent manner. CBPs had lower FEV1 (forced expiratory volume in 1 s) than non-CBPs and mediation analysis identified that a large portion (41-72%) of the FEV1 reduction associated with carbon black exposure could be explained by WA%. Small airway wall thickening as a major imaging change detected by CT may underlie the pathology of lung function impairment caused by carbon black exposure.

Keywords: airway wall thickening; carbon black; carbon content in airway macrophage; dose-response; mediation effect.

Figure 1.

Airway wall thickening mediating carbon black exposure induced lung function impairment. The study design, variables measured, and overall analytic plan all followed a framework for mediation analysis. In mediation analysis (A), the c coefficient denotes the direct effect of carbon black exposure on spirometry, without controlling for airway wall thickening (eg, WA%, mediator). The c' coefficient denotes the direct effect of carbon black exposure on spirometry, controlling for airway wall thickening (mediator). The proportion mediated is equal to delta c (ie, c–c') divided by c. We took a permutation-based method to assess whether the proportion mediated was statistically significant or not (B). The relationship between spirometry and the vector of independent variables was permuted for 500 times. Each permutated database allowed the association analysis of spirometry with carbon black exposure and other covariates without and with including wall area percent (WA%) to calculate the c and c'. Permutation was conducted for 500 times to generate the distribution of c–c' under null hypothesis of no mediation. Value of c-c' calculated using observed data (−0.15) for associations between carbon black exposure and forced expiratory volume in 1 s (FEV1) was compared with the distribution generated by permutation and Pperm was calculated as the number of permuted databases generating a c–c' that is smaller than observed value (n = 0 for FEV1) divided by 500. Thus, Pperm was < 0.002 in this case.

Figure 2.

Representative photos of macrophages containing varied amount of carbon particles (A–D) and carbon content in airway macrophage (CCAM) levels by carbon black exposure and smoking status (E). Four well-stained macrophages were identified from a sputum sample from a female carbon black packer (CBP) who was never smoker and had a carbon black working history of 8 years. The values of carbon black area in nucleus negative cytoplasm area were 1.34, 10.40, 13.61, and 24.86 μm² for macrophages A–D, respectively. The values of CCAM were 0.34%, 2.47%, 2.76%, and 4.46% for macrophages A–D, respectively. The median levels of CCAM were 1.4 and 10.8 for 61 non-CBPs and 35 CBPs, respectively. The ps = .75 and .60 between ever (blue, left) and never (red, right) smokers in CBPs and non-CBPs, respectively. Sample sizes were 26, 9, 28, and 33 for CBP ever smokers, CBP never smokers, non-CBP ever smokers, and non-CBP never smokers, respectively. The 5 horizontal bars from bottom to top represent the minimum, first quartile, median, third quartile, and maximum. Symbols (eg, ○ and +) represent means.

Figure 3.

Scatter plots and correlation between carbon content in airway macrophage (CCAM) and average wall area percent (WA%). A total of 61 non-carbon black packer (CBPs) and 35 CBPs had CCAM available. The scatter plots between CCAM with and without natural log transformation and WA% for sixth and ninth generations of airways suggested a dose-dependent linear relationship. Spearman coefficients for correlations of CCAM with sixth WA% and ninth WA% in 96 subjects were 0.28 (p < .0001) and 0.20 (p < .0001), respectively. Pearson coefficients for correlations of natural log transformed CCAM with sixth WA% and ninth WA% in 96 subjects were 0.28 (p < .0001) and 0.26 (p < .0001), respectively. Spearman coefficients for correlations of CCAM with sixth WA% and ninth WA% in 35 CBPs were 0.26 (p = .0016) and 0.16 (p = .064), respectively.

Figure 4.

Scatter plots and correlation between average wall area percent (WA%) and spirometry. Significant inverse associations were identified between average WA% at sixth and ninth generations of airways and spirometry measurements (eg, forced expiratory volume in 1 s [FEV1], forced vital capacity [FVC], and FEV/FVC). Pearson coefficients for correlations of sixth WA% with FEV1, FVC, and FEV1/FVC ratio were −0.19 (p = .018), −0.12 (p = .13), and −0.23 (p = .005), respectively. Pearson coefficients for correlations of ninth WA% with FEV1, FVC, and FEV1/FVC ratio were −0.16 (p = .047), −0.07 (p = .38), and −0.26 (p = .001), respectively.