Abstract
Rationale and Objectives
There are limited data on, and controversies regarding gender differences in the airway dimensions of smokers. Multi-detector CT (MDCT) images were analyzed to examine whether gender could explain differences in airway dimensions of anatomically matched airways in smokers.
Materials and Methods
We used VIDA imaging software to analyze MDCT scans from 2047 smokers (M:F, 1021:1026) from the COPDGene® cohort. The airway dimensions were analyzed from segmental to subsubsegmental bronchi. We compared the differences of luminal area, inner diameter, wall thickness, wall area percentage (WA%) for each airway between men and women, and multiple linear regression including covariates (age, gender, body sizes, and other relevant confounding factors) was used to determine the predictors of each airway dimensions.
Results
Lumen area, internal diameter and wall thickness were smaller for women than men in all measured airway (18.4 vs 22.5 mm2 for segmental bronchial lumen area, 10.4 vs 12.5 mm2 for subsegmental bronchi, 6.5 vs 7.7 mm2 for subsubsegmental bronchi, respectively p < 0.001). However, women had greater WA% in subsegmental and subsubsegmental bronchi. In multivariate regression, gender remained one of the most significant predictors of WA%, lumen area, inner diameter and wall thickness.
Conclusion
Women smokers have higher WA%, but lower luminal area, internal diameter and airway thickness in anatomically matched airways as measured by CT scan than do male smokers. This difference may explain, in part, gender differences in the prevalence of COPD and airflow limitation.
Keywords: Airway dimensions, CT scan, Gender differences, Smoker
INTRODUCTION
Smoking is a major risk factor for chronic obstructive pulmonary disease (COPD) and airflow obstruction. However, only a minority of smokers develop COPD, and the relationship between smoking history and the severity of airflow obstruction is weak (1). Thus, there is a new appreciation that COPD may be a heterogeneous disorder of smoking with many phenotypes (2). Some of the factors that are associated with the severity of airflow obstruction include: age, height, race, gender, genetic susceptibility, air pollution, and airway dimensions (3–7). Within the past several decades, there has also been a demographic shift in gender distribution of individuals with COPD. In 2000 there were more women diagnosed with COPD than men in the United States (8). Some have postulated that women may be more susceptible to the damaging effects of smoking and may be at greater risk of smoking-induced lung function impairment (4, 9–11).
Besides spirometry, chest CT has recently been used as a valuable tool to assess lung damage from smoking. Advances in CT imaging have permitted more detailed analysis of airway dimensions (12). It has been suggested that these CT measurements has potential power to represent histological dimension changes in the airway (13, 14). Although CT measurements of airway dimensions are predominantly of medium-sized airways, they could be representative of the degree of remodeling in small airways determined by pathology (13). Aysola et al. (14) reported that the airway thickness on endobronchial biopsy samples from individuals with asthma and healthy subjects correlated with wall area percentage (WA%).
Histologically, Martinez et al. (15) reported that women exhibited smaller airway lumens with disproportionately thicker airway walls than men in patients with severe COPD. Three other publications (7, 16, 17) that reported sex differences associated with CT airway thickness found that female smokers did not show increased wall thickness compared to men; however, a recent publication found that the square root of the wall thickness of a hypothetical airway of internal perimeter of 10 mm (SQRTWA@pi10) was higher in men than women (16).
None of these studies has reported gender difference of anatomically matched, specified airway wall. Furthermore, most studies used only a single population of subjects for airway measurement. To overcome these limitations and to evaluate whether there are gender differences in airway dimensions even when including confounding variables, we used the COPDGene® (the Genetic epidemiology of COPD) cohort (http://www.copdgene.org/) (18) to determine whether gender could explain the differences in airway dimensions of anatomically matched airways in smokers.
MATERIALS AND METHODS
Study Populations
The COPDGene® Study is an ongoing multicenter investigation of the genetic epidemiology of smoking-related lung disease (18). The first 2,047 smokers with quantitative CT data from the COPDGene® cohort were included in this study. All subjects were studied after obtaining the consent of study under protocols approved by local Institutional Review Board (IRB) and with guidelines recommended by the National Institutes of Health. Subjects were men and women; non-Hispanic whites or African-Americans aged 45 to 80 years with a smoking history of at least 10 pack years.
Previously proposed exclusion criteria (18) were applied in the cohort (18): exclusion criteria are a pregnant woman, a history of other lung disease except asthma (e.g., pulmonary fibrosis, extensive bronchiectasis, cystic fibrosis), previous surgical excision of at least one lung lobe (or lung volume reduction procedure), active cancer under treatment, suspected lung cancer (large or highly suspicious lung mass), metal in the chest, recent exacerbation of COPD treated with antibiotics or steroids, recent eye surgery, MI, other cardiac hospitalization, recent chest or abdominal surgery, inability to use albuterol, multiple self-described racial categories, history of chest radiation therapy, and first- or second-degree relative already enrolled in the study. Smokers who have an unclassified pattern by GOLD (Global initiative for chronic Obstructive Lung Disease) criteria on spirometry, denoted as GOLD U (normal FEV1/FVC but reduced FEV1) and GOLD 0 (smokers with normal spirometry) are eligible for the study. Each subject underwent a spirometry and multi-detector CT (MDCT). The COPDGene® cohort includes nineteen clinical centers in the United States (18) (see center and investigator list in Acknowledgments).
Quantitative CT Analysis
Analysis of COPDGene cohort using VIDA software
In COPDGene® study, all MDCT (at least 16 detector channels) of the chest used a tube potential of 120 kVp and an effective mAs of 200 (Supplemental Tables S1-A, B and C) (18). Submillimeter near isotropic MDCT scans without contrast were acquired at end inspiration. The images were reconstructed with slice thicknesses of 0.625, 0.75 or 0.9 mm depending on the CT manufacturer (General Electric Medical Systems, Siemens and Philips) (18). The optimal reconstruction kernel for a given model of CT scanner for the VIDA software program was used to segment the lungs, lobes and airway tree. The image matrix size was 512 × 512 pixels, and the pixel sizes ranged from x: 0.55 to 0.78 mm, y: 0.55 to 0.78. Other detailed CT protocols were the same with the previous report (18).
Airway dimensions were measured using automated, quantitative software that was designed to label and quantify the bronchial tree (Pulmonary Workstation+ VIDA Diagnostics; Iowa City, IA. www.vidadiagnostics.com, Supplemental Figure 1) (19, 20). These airways were as follows: right upper apical segmental, subsegmental and subsubsegmental brochi, right middle lateral segmental, subsegmental and subsubsegmental bronchi, right posterior basal segmental, subsegmental and subsubsegmental bronchi, left upper apical segmental, subsegmental and subsubsegmental bronchi, left superior lingular segmental, subsegmental and subsubsegmental bronchi, and left posterior basal segmental, subsegmental and subsubsegmental bronchi.
These airway indices were measured from the centerline to the airway edge in each slice of the 3D image set. Reported airway dimensions represented the average of all the measurements collected along the middle third of each individual airway segment. For each individual, the segmental, subsegmental and subsubsegmental airway data were averaged to provide a mean value for each level of branching. Structural measurements of airway dimensions included the lumen area (Ai), inner diameter, airway wall thickness, wall area percentage (WA%) and SQRTWA@pi10 in each anatomically matched airway. The perimeters of the airway lumen and of the adventitia subtended two areas: Ai (luminal area) and Ao (total area). WA% was calculated as (Ao – Ai)/Ao × 100. SQRTWA@pi10 was calculated for each subject by fitting a linear relationship between Pi and SQRTWA of each measured bronchus (16, 21).
For determining the extent of emphysema, quantitative densitometric analysis was performed with VIDA and areas of CT emphysema were defined as low attenuation areas (LAA) [<–950 Hounsfield units, HU]. Then, the percentage of LAA (LAA% −950HU) was determined for the entire lung. Region growing of airway tree was performed by research assistants under the training and supervision of the Imaging core of the COPDGene® study (list in Acknowledgements). The stability of CT measurements for each scanner is monitored by monthly scanning using a custom COPDGene phantom (18).
Statistical Analysis
Gender differences were evaluated using t-tests for continuous variables and χ2 tests for categorical variables. Data that were not normally distributed (e.g., LAA% −950 HU and packs years of smoking (PYs)) were also analyzed after log transformation. Regression analysis was used to determine predictors of WA%. Multivariate analysis was performed using linear regression models for WA% adjusted for subject’s age, sex, height, weight, PYs, race, smoking status (current/former smoker), LAA% −950HU and total lung capacity (TLC% predicted) to adjust demographic and body size differences, and confounding factors that could affect WA%. Clinical centers and CT scanner types were also included as a variable to adjust those differences in multicenter COPDGene® cohort. Correlations between lung function (FEV1% pred.) and airway parameters were determined using parametric testing methods with Pearson correlation coefficients. P values less than 0.05 were considered statistically significant. Statistical software (SPSS, version 17.0; Chicago, IL) was used for analysis.
RESULTS
Demographics, smoking history and lung function for the 2047 COPDGene® subjects who were included in the study are shown in Table 1. Subjects were predominantly White, but there were no ethnic differences between genders (Caucasian 78.0% in male, 77.3% in female, p = 0.706). Subjects had a heavy smoking history and PYs was higher in males (M:F, 49.9: 42.1 PYs, p < 0.01). Current smokers were more frequently male (M:F, 42.8% : 36.6%, p = 0.004). Height and body weight were smaller in women. There were no significant differences in lung function between genders and mean FEV1% (pred.) results (M:F, 72.5%: 73.9%) were consistent with GOLD stage II disease (812 smokers without evidence of airway obstruction, 146 smokers with GOLD-I, 486 smokers with GOLD-II, 294 smokers with GOLD-III, 158 smokers with GOLD-IV, and 151 smokers with GOLD U).
Table 1.
Subject characteristics*
| Men (n = 1021) | Women (n = 1026) | P-value | |
|---|---|---|---|
| Age (yr) | 61.7 ± 9.3 | 61.7 ± 9.1 | 0.968 |
| Caucasian (%) | 78.0 | 77.3 | 0.706 |
| Pack years smoking | 49.9 ± 29.3 | 42.1 ± 24.5 | <0.001‡ |
| Current smoker (%) | 42.8 | 36.6 | 0.004† |
| Height, cm | 176.4 ± 72 | 162.9 ± 6.4 | <0.001‡ |
| Body weight, Kg | 88.4 ± 18.5 | 75.9 ± 18.6 | <0.001‡ |
| FEV1%pred(postBD) | 72.5 ± 27.9 | 73.9 ± 26.7 | 0.229 |
| GOLD 0§ | 98.4 ± 11.9(n, 389) | 97.6 ± 12.0(n, 423) | 0.339 |
| GOLD 1 | 91.1 ± 9.4(n, 82) | 92.3 ± 10.4(n, 64) | 0.464 |
| GOLD 2 | 64.6 ± 8.6(n, 245) | 63.4 ± 8.5(n, 241) | 0.108 |
| GOLD 3 | 39.1 ± 5.8(n, 150) | 40.2 ± 5.7(n, 144) | 0.116 |
| GOLD 4 | 22.0 ± 4.5(n, 91) | 21.9 ± 4.7(n, 67) | 0.916 |
| GOLD U¶ | 70.7 ± 6.6(n, 64) | 70.2 ± 8.0(n, 87) | 0.675 |
Data are reported as mean ± SD unless otherwise indicated.
P-values are indicated as follows: ******p < 0.05.
p < 0.001.
GOLD 0
GOLD U denote smokers with normal spirometry, and smokers with normal FEV1/FVC but reduced FEV1(%pred), respectively.
Most airway measurements (inner diameter, wall thickness and lumen area) were lower in women compared to men (Table 2 and Supplemental Figs. 2A and 2B). The numbers of obtainable measurements were slightly decreased as the bronchial branches go more distal (n = 2043 for segmental bronchi, 2040 for subsegmental, 2033 for subsubsegmental). However, women had higher WA% in the subsegmental and subsubsegmental bronchi. In subgroup analyses using subjects with or without airflow obstruction, gender differences of airway dimensions were consistent (Supplemental Table S2-A and B). SQRTWA@pi10 was not significantly different between men and women (Supplemental Table S3).
Table 2.
Mean CT airway parameters at each level of bronchial branching by gender*
| Site of airway | Airway parameter† | Men | Women | P-value |
|---|---|---|---|---|
| Segmental bronchi | Inner diameter | 5.3 ± 0.7 | 4.8 ± 0.6 | <0.001 |
| (M:F, 1019:1024) | Lumen area (mm2) | 22.5 ± 6.5 | 18.4 ± 4.8 | <0.001 |
| Wall Sickness | 1J ± 0.2 | 1.5 ± 0.2 | <0.001 | |
| WA% | 61.2 ± 3.4 | 61.3 ± 3.2 | 0.411 | |
| Subsegmental bronchi | Inner diameter | 3.9 ± 0.5 | 3.6 ± 0.5 | <0.001 |
| (M:F, 1017:1023) | Lumen area (mm2) | 12.5 ± 3.4 | 10.4 ± 2.8 | <0.001 |
| Wall Sickness | 1.5 ± 0.2 | 1.4 ± 0.2 | <0.001 | |
| WA% | 63.9±2.7 | 64.8 ± 2.6 | <0.001 | |
| Subsubsegmental bronchi | Inner diameter | 3.1 ± 0.4 | 2.8 ± 0.3 | <0.001 |
| (M:F, 1014:1019) | Lumen area (mm2) | 7 J ± 1.9 | 6.5 ± 1.6 | <0.001 |
| Wall Sickness | 1.5 ± 0.2 | 1.4 ± 0.2 | <0.001 | |
| WA% | 66.7 ± 2.1 | 67.8 ± 1.9 | <0.001 |
Data are reported as mean ± SD (mm) unless otherwise indicated. Histograms of segmental and subsegmental airway dimensions are shown in Supplemental Figures 2A and 2B.
Airway parameter: WA% = percentage of wall area.
Emphysema (LAA% < −950 HU) was more extensive and CT measured lung volume (TLC% predicted) was lower in men than women (Supplemental Table S3). Univariate analysis was used to determine which factors might be associated with WA% for different airways (Supplemental Table S4). There were significant associations between WA% and most variables (age, gender, race, pack-years, smoking status, height, weight, emphysema score, TLC, study center, and scanner type). Male gender (t ratios −7.8, −12.2), height (t ratios −10.4, −13.3) and LAA% (t ratios 9.1, 7.9) were stronger predictors than other variables in subsegmental and subsubsegmental bronchi while body weight and TLC% were more powerful in segmental bronchi compared to other variables.
In multivariate analysis including all of these variables (Table 3: shown for several key variables, and Supplemental Table S5 shown for all variables), PYs, smoking status (current smoker), height, weight and TLC% showed consistent and significant associations with WA% from all airways, from segmental to subsubsegmental bronchi. Male gender was negatively associated with subsegmental and subsubsegmental WA% (t = −3.47, −6.9; p = 0.001, <0.001, repectively) while there was no significant associations between gender and segmental WA%.
Table 3.
Multivariate regression analysis of predictors of generational WA%
| Site | Variables | Estimates | Std. Error | t | Adjusted R Square |
P-value |
|---|---|---|---|---|---|---|
| Segmental Bronchi | Male gender | 0.162 | 0.187 | 0.867 | 0.207 | 0.386 |
| PYs | 0.012 | 0.003 | 4.556 | <0.001 | ||
| Current smoker | 0.921 | 0.161 | 5.737 | <0.001 | ||
| Height, cm | −0.084 | 0.01 | −8.291 | <0.001 | ||
| Weight, kg | 0.041 | 0.004 | 10.269 | <0.001 | ||
| LAA% −950HU | 0.072 | 0.008 | 9.149 | <0.001 | ||
| TLC% | −0.031 | 0.005 | −6.511 | <0.001 | ||
| Subsegmental bronchi | Male gender | −0.514 | 0.148 | −3.47 | 0.247 | 0.001 |
| PYs | 0.008 | 0.002 | 4.16 | <0.001 | ||
| Current smoker | 0.747 | 0.127 | 5.877 | <0.001 | ||
| Height, cm | −0.072 | 0.008 | −8.994 | <0.001 | ||
| Weight, kg | 0.018 | 0.003 | 5.62 | <0.001 | ||
| LAA% −950HU | 0.085 | 0.006 | 13. 692 | <0.001 | ||
| TLC% | −0.031 | 0.005 | −9.176 | <0.001 | ||
| Subsubsegmental bronchi | Male gender | −0.794 | 0.115 | −6.928 | 0.268 | <0.001 |
| PYs | 0.007 | 0.002 | 4.43 | <0.001 | ||
| Current smoker | 0.586 | 0.098 | 5.951 | <0.001 | ||
| Height, cm | −0.051 | 0.006 | −8.174 | <0.001 | ||
| Weight, kg | 0.006 | 0.002 | 2.429 | 0.015 | ||
| LAA% −950HU | 0.066 | 0.005 | 13. 647 | <0.001 | ||
| TLC% | −0.031 | 0.003 | −10.544 | <0.001 |
Among the above variables, height, LAA% and TLC% were more powerful predictors than other demographic predictors in the subsegmental and subsubsegmental paths. In multivariate analysis for other airway parameters such as lumen area and wall thickness (Table 4: shown for several key variables, Supplemental Table S6, S7-A and B), gender was one of the significant and powerful determinants for each quantitative CT parameter. However, gender was not a significant predictor of SQRTWA@pi10 in the multivariate analysis. WA%, lumen area and SQRTWA@pi10 were significantly correlated with FEV1% predicted (Table 5).
Table 4.
Multivariate regression analysis of predictors of luminal area of subsegmental bronchi, wall thickness of subsegmental bronchi and SQRTWA@pi10
| Parameters | Variables | Estimates | Std. Error | t | Adjusted R2 | Sig. |
|---|---|---|---|---|---|---|
| Lumen area* | Male gender | 1.307 | 0.l82 | 7.177 | 0.247 | <0.001 |
| PYs | −0.009 | 0.002 | −3.542 | <0.001 | ||
| Current smoker | −0.736 | 0.l56 | −4.702 | <0.001 | ||
| Height, cm | 0.084 | 0.0l | 8.527 | <0.001 | ||
| Weight, kg | 0.002 | 0.004 | 0.387 | 0.699 | ||
| LAA% −950HU | −0.092 | 0.008 | −12.075 | <0.001 | ||
| TLC% | 0.029 | 0.005 | 6.247 | <0.001 | ||
| Wall thickness* | Male gender | 0.071 | 0.0l | 7.345 | 0.33 | <0.001 |
| PYs | <0.001 | 0 | 0.683 | 0.495 | ||
| Current smoker | −0.0l4 | 0.008 | −1.658 | 0.097 | ||
| Height, cm | <0.001 | 0.00l | −0.737 | 0.46l | ||
| Weight, kg | 0.003 | 0 | 16.335 | <0.001 | ||
| LAA% −950HU | <0.001 | 0 | 0.884 | 0.377 | ||
| TLC% | 0.00l | 0 | −1.918 | 0.055 | ||
| SQRTWA@pi10 | Male gender | 0.002 | 0.007 | 0.253 | 0.197 | 0.8 |
| PYs | 0.00l | 0 | 2.l5l | 0.032 | ||
| Current smoker | 0.03 | 0.006 | 4.654 | <0.001 | ||
| Height, cm | −0.002 | 0 | −4.797 | <0.001 | ||
| Weight, kg | 0.00l | 0 | 8.749 | <0.001 | ||
| LAA% −950HU | 0.002 | 0 | 6.795 | <0.001 | ||
| TLC% | −0.002 | 0 | −9.4l5 | <0.001 |
Lumen area, Wall thickness and WA% in subsegmental bronchi. The results in other airway sites, segmental and subsubsegmental bronchi (Supplemental Tables S7-A and B), were similar as above.
Table 5.
Correlations between airway dimensions and FEVl (% predicted)
| Airway parameters† | r | P-value |
|---|---|---|
| SQRTWA@pi10 | −0.301 | <0.001 |
| Segmental WA% | −0.502 | <0.001 |
| Subsegmental WA% | −0.557 | <0.001 |
| Subsubsegmental WA% | −0.510 | <0.001 |
| Segmental Ai | 0.395 | <0.001 |
| Subsegmental Ai | 0.465 | <0.001 |
| Subsubsegmental Ai | 0.393 | <0.001 |
Airway parameters: SQRTWA@pi10 = square root of the wall area at a airway internal perimeter of 10 mm. WA% (or Ai) = mean wall area percentage (or lumen area).
DISCUSSION
Computed tomography is becoming a useful, non-invasive tool to evaluate the airway dimensions. There are several different investigational methods used to express the morphologic characteristics of airway wall. These include the two most frequently used metrics: WA% and SQRTWA@Pi10 (21–23). It should be noted that these two metrics are not directly measured, but are derived from other airway measurements. Directly measured metrics include luminal area, inner diameter and wall thickness. These computational differences have led to different investigators reporting apparently paradoxical conclusions regarding gender differences and have led to confusion in interpretation of CT derived airway measurement. For example, WA% is a deceptive measure of wall thickness because as airways become smaller, the WA% becomes larger (24).
Thus, WA% is affected not only by airway thickness, but also by airway size. SQRTWA@Pi10 is a useful method to correct for differences in airway size; however, the concept of a hypothetical airway is less relevant when one can measure actual airways that have been anatomically matched. Using the SQRTWA@Pi10 also discounts the importance of airway size on airflow. We speculated that this is why WA% and luminal area had better correlation with FEV1 (% predicted) than SQRTWA@Pi10 (Table 5).
To our knowledge, this report is the largest investigation of airway dimensions measured by MDCT and the only report of gender differences in airway dimensions classified according to bronchial branching order. A novel finding is that in anatomically matched sites, especially in distal airways such as subsegmental and subsubsegmental bronchi, female smokers have higher WA% compared to male smokers. However, they have lower luminal area, airway thickness, and internal diameter of airway in anatomically matched airways than do male smokers. The significance of reduced luminal area in women is particularly important to physiology because the smaller size of women’s lungs is associated with lower flow rates (25).
Furthermore, airflow limitation in COPD is more closely related to the dimensions of the distal (small) airways than proximal (large) airways (23). The diameters of subsubsegmental bronchi in our study were around 3 mm, which is thought to be more representative of airflow limitation (26). Thus, the direct measurement of anatomically matched airway lumen also has an important physiologic relevance to airflow. The smaller lumen area and the higher WA% of these distal airways in women could explain why women have a higher prevalence of COPD and may also explain gender differences in the presentation and pathophysiology of airflow obstruction and COPD.
SQRTWA@Pi10 is a hypothetical airway parameter that is obtained by fitting a linear relationship between Pi and SQRTWA (21). Other studies (7, 16, 17) have come to different conclusions from our results; namely that airways are thicker in men compared to women in terms of SQRTWA@Pi10. In our study, there were no significant differences of SQRTWA@Pi10 between genders. In the subgroup analysis according to GOLD stage, there was significant difference of SQRTWA@Pi10 between genders (men : women, 3.84 : 3.80 mm, p = 0.02) only among severe COPD (GOLD 3 and 4) like the other previous report (7). But, the differences of SQRTWA@Pi10 between genders are very small. In the other studies (7, 16, 17), the differences of SQRTWA@Pi10 between genders (around 0.2–0.3mm) were also very small.
We postulated that these small differences could be easily obscured by other hidden confounding factors such as different airway measurement algorithm in each study. In contrast to SQRTWA@Pi10, women had the higher WA% in the subsegmental and subsubsegmental paths through all GOLD stages. Thus, in regarding whether airways are thicker in women compared to men, it is important to consider which definitions of airway dimensions are reported. However, further study will be needed to clarify these discrepancies and its contribution to clinical relevance, and to evaluate which airway parameter could be more important to clinical settings.
The other major difference between this study and the recently published studies of airway measurements (7, 16, 17) is the methodology for determining airway wall thickness. Most publications have used the Full-Width-At-Half-Maximum (FWHM) method to measure the dimensions. We used an optimal surface algorithm (VIDA) to determine airway boundaries. The results from VIDA have showed better subpixel accuracy for the inner border and equivalent results for the outer wall border compared with those of the FWHM method (20). The segmentation algorithm of VIDA retrieves a significantly higher count of airway branches compared with a commonly used region growing segmentation algorithm (20). However, the numbers of obtainable measurements were decreased as the branches go more distal. This suggests that some difficulties including reproducibility are still remained to measure airway dimensions especially in the small airways with the VIDA software even though this is a more updated and automatically operated software that has been validated previously (28).
Also, long-term reproducibility should be validated in a longitudinal future study using VIDA. But, a strength of this study includes the large sample of airway measurements with averaging data for each generational path. The problem of reproducibility of the measurements could be weakened, at least, to some extent by averaging the values of each different airway from a large number of subjects. Furthermore, parallel imaging analyses of LAA% were done using Airway Inspector (www.airwayinspector.org) and 3D Slicer (http://www.slicer.org/) for all subjects and, airway dimensions were also measured using 3D slicer for more than 80 subjects. The results of gender differences from 3D slicer were similar to VIDA’s (data not shown).
A secondary finding in this study was that age, smoking (status and amount), body sizes (height and weight), emphysema and other various factors affect airway measurements. Smoking status, body sizes and TLC% showed significant associations with WA% from all the airways. Gender effects for WA% were present in the subsegmental and subsubsegmental paths, not in the segmental bronchi. Among the variables for WA%, height, LAA% and TLC% were more powerful predictors than other demographic predictors in the subsegmental and subsubsegmental paths. For other airway parameters, gender is one of the powerful predictors for luminal area and wall thickness, but not for SQRTWA@Pi10. This suggests that each variable could affect different airway metrics with different intensity and different location, and this might be associated with the heterogeneity of COPD and the importance of airway measurements in anatomically matched sites.
There were small gender differences (around 1%) in WA% that could affect the small physiologic relevance. However, the small changes in each variable should be considered to better understand the heterogeneity of COPD because the factors of airflow obstruction and COPD are multifactorial. Additionally, gender differences in WA% were exaggerated in current smokers with COPD (Supplemental Table S8). This could suggest that women’s airway may be more susceptible than men’s to the airway damaging effects of current smoking. However, to clarify these heterogeneous relationships between each airway parameter and other variables, and smoking effects in gender, a longitudinal study is needed in the future.
We found other sources of variability in airway wall measurements including clinical center, CT scanner type, and location of airways. However, the magnitudes of gender effect for luminal area (t ratio, around 7.2, Table 4) and wall thickness (t-ratio, around 7.2, Table 4) in all airway paths were higher than those of CT scanner type (t-ratio, around 1.0, data not shown) or clinic center (t-ratio, around 2.0, data not shown). The magnitudes of gender effect for WA% in subsegmental and subsubsegmental paths were also higher than they were for scanner or center. This suggests gender differences are consistent irrespective of scanner type or clinic center. However, center or scanner type could affect the quantitative measurements as a noise to some extent.
These findings indicate that gender is just one factor for airway wall, that a complex background of airway dimensions exists and that failure to take into account other clinical variables may weaken any observed differences in CT derived airway wall measurements. The most likely explanations for gender differences in airways are biological and environmental determinants (30). Several studies suggest that genetic interactions may be important to the gender differences and CT phenotypes associated with COPD (4, 31, 32). However, the precise mechanism and determinants of gender differences on airway dimension remains unknown. This study also suggests that other causes of variability in airway measurements need further investigation.
The limitations of this study are similar to those of the parent study (COPDGene®). First, it is cross-sectional and may not account for changes in airway dimension over time. Second, nonsmokers were not included. Thus, we cannot fully evaluate the effects of smoking per se on airway dimensions. Additionally, a fundamental limitation of airway measurements is the spatial resolution (voxel size: x, y, z = slice thickness) of the underlying CT image data. For example, for a CT acquisition field-of-view of 35 cm, typical for the COPDGene® cohort, the 512 × 512 image size translates to an x, y pixel dimension of 0.68 mm. Slice thickness (z) ranged from 0.625 to 0.9 mm depending on the CT scanner manufacturer. Therefore, to have two pixels on a feature of interest (e.g., an airway lumen or airway wall) as suggested from Nyquist sampling theory would require a feature of approximately 1.37 mm size or greater (33).
Airway wall thickness may be near the spatial resolution limit in subsegmental and subsubsegmental airways in the COPDGene cohort. This requires further study and comparison to CT phantom results. Note that larger field-of-view dimensions and larger slice thicknesses will further decrease spatial resolution. These technical issues may explain, at least in part, the variation of airway dimensions by clinical center and scanner type. Last, we used a convenience cohort from COPDGene cohort that was obtained to look for genetic factors in COPD. The cohort is a heterogeneous mix of individuals with varied smoking histories and a range of airflow obstruction. It was not ideally recruited to answer the question about gender differences. More studies will be needed only for gender differences of airway dimensions to confirm these differences.
In conclusion, female smokers have disproportionately higher WA%, but lower luminal area and airway thickness in anatomically matched sites, subsegmental and subsubsegmental bronchi as measured by CT scan than do male smokers. This difference may explain, in part, gender differences in the heterogeneity of COPD and airflow obstruction. Awareness of the gender difference in airway dimensions should be considered in future investigations of airway related disease.
Supplementary Material
ACKNOWLEDGMENTS
The authors wish to thank COPDGene® investigators (below) and Christina Schnell—including her invaluable secretarial support—for data collection and their support of writing this manuscript.
Funding Support: This study was supported by National Heart, Lung and Blood Institute (NHLBI RO1HL 095432, U01 HL089856, U01 HL089897); UL1 RR025780 from NCRR/HIH; and K25HL104085 (SJE).
ABBREVIATIONS
- COPD
chronic obstructive pulmonary disease
- FEV1% pred.
% Predicted Forced Expiratory Volume in 1 Second
- LAA%
percentage of low attenuation areas
- SQRTWA@pi10
square root of the wall area at a airway internal perimeter of 10 mm
- WA%
percentage of wall area
Footnotes
DECLARATION OF INTEREST The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
COPDGene study participating centers and investigators are as follows:
Ann Arbor VA (AVA): Jeffrey Curtis, MD (PI), Ella Kazerooni, MD(RAD)
Baylor College of Medicine (BAY), Houston, TX: Nicola Hanania, MD, MS (PI), Philip Alapat, MD, Venkata Bandi, MD, Kalpalatha Guntupalli, MD, Elizabeth Guy, MD, William
Lunn, MD, Antara Mallampalli, MD, Charles Trinh, MD(RAD), Mustafa Atik, MD
Brigham and Women’s Hospita (BWH)l, Boston, MA: Dawn DeMeo, MD(Co-PI), Craig Hersh, MD(Co-PI), Francine Jacobson, MD, MPH (RAD)
Columbia University (COL), New York, NY: R. Graham Barr, MD, DrPH (PI), Byron Thomashow, MD, John Austin, MD(RAD)
Duke University Medical Center (DUK), Durham, NC: Neil MacIntyre, Jr., MD (PI), Lacey Washington, MD (RAD), H Page McAdams, MD
Fallon Clinic, Worcester (FAL), MA: Richard Rosiello, MD (PI), Timothy Bresnahan, MD
Health Partners Research Foundation (HPR), Minneapolis, MN: Charlene McEvoy, MD, MPH (PI), Joseph Tashjian, MD(RAD)
Johns Hopkins University (JHU), Baltimore, MD: Robert Wise, MD (PI), Nadia Hansel, MD, MPH, Robert Brown, MD (RAD)
Los Angeles Biomedical Research Institute at Harbor UCLA Medical Center (UMC), Los Angeles, CA: Richard Casaburi, MD (PI), Janos Porszasz, MD, PhD, Hans Fischer, MD, PhD (RAD), Matt Budoff, MD
Michael E. DeBakey VAMC, Houston (HVA), TX: Amir Sharafkhaneh, MD (PI)
Minneapolis VA (MVA): Dennis Niewoehner, MD (PI), Tadashi Allen, MD (RAD), Kathryn Rice, MD
Morehouse School of Medicine (MSM), Atlanta, GA:Marilyn Foreman, MD, MS (PI), GloriaWestney, MD, MS, Eugene Berkowitz, MD, PhD
National Jewish Health (NJH), Denver, CO: Russell Bowler, MD, PhD (PI), Adam Friedlander, MD, Eleonora Meoni, MD
Temple University (TEM), Philadelphia, PA: Gerard Criner, MD (PI), Victor Kim, MD, Nathaniel Marchetti, DO, Aditi Satti, MD, A. JamesMamary, MD, Robert Steiner, MD (RAD), Chandra Dass, MD (RAD)
University of Alabama, Birmingham (UAB), AL: William Bailey, MD (PI), Mark Dransfield, MD (Co-PI), Lynn Gerald, PhD, MSPH, Hrudaya Nath, MD (RAD)
University of California, San Diego (USD), CA: Joe Ramsdell, MD (PI), Paul Ferguson, MS, RCP, Paul Friedman, MD (RAD)
University of Iowa, Iowa City (UIA), IA: Geoffrey McLennan, MD, PhD (PI), Edwin JR van Beek, MD, PhD (RAD)
University of Michigan, Ann Arbor (HAR), MI: Fernando Martinez, MD (PI), MeiLan Han, MD, Deborah Thompson, PhD, Ella Kazerooni, MD (RAD)
University of Minnesota, Minneapolis, MN (UMN): Christine Wendt, MD (PI), Tadashi Allen, MD (RAD)
University of Pittsburgh, Pittsburgh (PIT), PA: Frank Sciurba, MD (PI), JoelWeissfeld, MD, MPH, Carl Fuhrman, MD(RAD), Jessica Bon, MD
University of Texas Health Science Center at San Antonio, San Antonio, TX (TXS): Antonio Anzueto, MD (PI), Sandra Adams, MD, Carlos Orozco, MD, C. Santiago Restrepo, MD (RAD), Amy Mumbower, MD (RAD)
Administrative Core: James Crapo, MD (PI), Edwin Silverman, MD, PhD (PI), Barry Make, MD, Elizabeth Regan, MD, Jonathan Samet, MD, Amy Willis, MA, Douglas Stinson
Genetic Analysis Core: Terri Beaty, PhD, Barbara Klanderman, PhD, Nan Laird, PhD, Christoph Lange, PhD, Iuliana Ionita, Stephanie Santorico, PhD, Edwin Silverman, MD, PhD
Imaging Core: David Lynch, MD, Joyce Schroeder, MD, John Newell, Jr., MD, John Reilly, MD, Harvey Coxson, PhD, Philip Judy, PhD, Eric Hoffman, PhD, Raul San Jose Estepar, PhD, George Washko, MD, Rebecca Leek, Jordan Zach, Alex Kluiber, Anastasia Rodionova, Tanya Mann
PFT QA Core: Robert Crapo, MD, Robert Jensen, PhD Biological Repository, Johns Hopkins University, Baltimore, MD: Homayoon Farzadegan, PhD
Data Coordinating Center and Biostatistics, National Jewish Health, Denver, CO: James Murphy, PhD, Douglas Everett, PhD, Carla Wilson
Epidemiology Core, University of Colorado School of Public Health, Denver, CO: John Hokanson, MPH, PhD
REFERENCES
- 1.Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. NHLBI/WHO global initiative for chronic obstructive lung disease (gold) workshop summary. Am J Respir Crit Care Med. 2001;163:1256–1276. doi: 10.1164/ajrccm.163.5.2101039. [DOI] [PubMed] [Google Scholar]
- 2.Friedlander AL, Lynch D, Dyar LA, Bowler RP. Phenotypes of chronic obstructive pulmonary disease. COPD. 2007;4:355–384. doi: 10.1080/15412550701629663. [DOI] [PubMed] [Google Scholar]
- 3.Rabe KF, Hurd S, Anzueto A, Barnes PJ, Buist SA, Calverley P, Fukuchi Y, Jenkins C, Rodriguez-Roisin R, van Weel C, Zielinski J. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: Gold executive summary. Am J Respir Crit Care Med. 2007;176:532–555. doi: 10.1164/rccm.200703-456SO. [DOI] [PubMed] [Google Scholar]
- 4.Silverman EK, Weiss ST, Drazen JM, Chapman HA, Carey V, Campbell EJ, Denish P, Silverman RA, Celedon JC, Reilly JJ, Ginns LC, Speizer FE. Gender-related differences in severe, early-onset chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2000;162:2152–2158. doi: 10.1164/ajrccm.162.6.2003112. [DOI] [PubMed] [Google Scholar]
- 5.Orozco-Levi M, Garcia-Aymerich J, Villar J, Ramirez-Sarmiento A, Anto JM, Gea J. Wood smoke exposure and risk of chronic obstructive pulmonary disease. Eur Respir J. 2006;27:542–546. doi: 10.1183/09031936.06.00052705. [DOI] [PubMed] [Google Scholar]
- 6.Liou TG, Kanner RE. Spirometry. Clinic Rev Allerg Immunol. 2009;37:137–152. doi: 10.1007/s12016-009-8128-z. [DOI] [PubMed] [Google Scholar]
- 7.Kim WJ, Silverman EK, Hoffman E, Criner GJ, Mosenifar Z, Sciurba FC, Make BJ, Carey V, Estepar RS, Diaz A, Reilly JJ, Martinez FJ, Washko GR. Ct metrics of airway disease and emphysema in severe COPD. Chest. 2009;136:396–404. doi: 10.1378/chest.08-2858. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Mannino DM, Homa DM, Akinbami LJ, Ford ES, Redd SC. Chronic obstructive pulmonary disease surveillance—United states, 1971–2000. MMWR Surveill Summ. 2002;51:1–16. [PubMed] [Google Scholar]
- 9.Dransfield MT, Davis JJ, Gerald LB, Bailey WC. Racial and gender differences in susceptibility to tobacco smoke among patients with chronic obstructive pulmonary disease. Respir Med. 2006;100:1110–1116. doi: 10.1016/j.rmed.2005.09.019. [DOI] [PubMed] [Google Scholar]
- 10.de Torres JP, Casanova C, Hernandez C, Abreu J, Aguirre-Jaime A, Celli BR. Gender and copd in patients attending a pulmonary clinic. Chest. 2005;128:2012–2016. doi: 10.1378/chest.128.4.2012. [DOI] [PubMed] [Google Scholar]
- 11.Chapman KR. Chronic obstructive pulmonary disease: Are women more susceptible than men? Clin Chest Med. 2004;25:331–341. doi: 10.1016/j.ccm.2004.01.003. [DOI] [PubMed] [Google Scholar]
- 12.Coxson HO. Quantitative computed tomography assessment of airway wall dimensions: Current status and potential applications for phenotyping chronic obstructive pulmonary disease. Proc Am Thorac Soc. 2008;5:940–945. doi: 10.1513/pats.200806-057QC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Nakano Y, Wong JC, de Jong PA, Nagao T, Coxson HO, Elliott WM, Hogg JC, Pare PD. The prediction of small airway dimensions using computed tomography. Am J Respir Crit Care Med. 2005;171:142–146. doi: 10.1164/rccm.200407-874OC. [DOI] [PubMed] [Google Scholar]
- 14.Aysola RS, Hoffman EA, Gierada D, Wenzel S, Cook-Granroth J, Tarsi J, Zheng J, Schechtman KB, Ramkumar TP, Cochran R, Xueping E, Christie C, Newell J, Fain S, Altes TA, Castro M. Airway remodeling measured by multidetector ct is increased in severe asthma and correlates with pathology. Chest. 2008;134:1183–1191. doi: 10.1378/chest.07-2779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Martinez FJ, Curtis JL, Sciurba F, Mumford J, Giardino ND, Weinmann G, Kazerooni E, Murray S, Criner GJ, Sin DD, Hogg J, Ries AL, Han M, Fishman AP, Make B, Hoffman EA, Mohsenifar Z, Wise R. Sex differences in severe pulmonary emphysema. Am J Respir Crit Care Med. 2007;176:243–252. doi: 10.1164/rccm.200606-828OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Grydeland TB, Dirksen A, Coxson HO, Pillai SG, Sharma S, Eide GE, Gulsvik A, Bakke PS. Quantitative computed tomography: Emphysema and airway wall thickness by sex, age and smoking. Eur Respir J. 2009;34:858–865. doi: 10.1183/09031936.00167908. [DOI] [PubMed] [Google Scholar]
- 17.Camp PG, Coxson HO, Levy RD, Pillai SG, Anderson W, Vestbo J, Kennedy SM, Silverman EK, Lomas DA, Pare PD. Sex differences in emphysema and airway disease in smokers. Chest. 2009;136:1480–1488. doi: 10.1378/chest.09-0676. [DOI] [PubMed] [Google Scholar]
- 18.Regan EA, Hokanson JE, Murphy JR, Make B, Lynch DA, Beaty TH, Curran-Everett D, Silverman EK, Crapo JD. Genetic epidemiology of copd (copdgene) study design. COPD. 2010;7:32–43. doi: 10.3109/15412550903499522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Palagyi K, Tschirren J, Sonka M. Quantitative analysis of intrathoracic airway trees: Methods and validation. Inf Process Med Imaging. 2003;18:222–233. doi: 10.1007/978-3-540-45087-0_19. [DOI] [PubMed] [Google Scholar]
- 20.Tschirren J, Hoffman EA, McLennan G, Sonka M. Segmentation and quantitative analysis of intrathoracic airway trees from computed tomography images. Proc Am Thorac Soc. 2005;Vol 2:484–487. doi: 10.1513/pats.200507-078DS. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Patel BD, Coxson HO, Pillai SG, Agusti AG, Calverley PM, Donner CF, Make BJ, Muller NL, Rennard SI, Vestbo J, Wouters EF, Hiorns MP, Nakano Y, Camp PG, Nasute Fauerbach PV, Screaton NJ, Campbell EJ, Anderson WH, Pare PD, Levy RD, Lake SL, Silverman EK, Lomas DA. Airway wall thickening and emphysema show independent familial aggregation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2008;178:500–505. doi: 10.1164/rccm.200801-059OC. [DOI] [PubMed] [Google Scholar]
- 22.Nakano Y, Muro S, Sakai H, Hirai T, Chin K, Tsukino M, Nishimura K, Itoh H, Pare PD, Hogg JC, Mishima M. Computed tomographic measurements of airway dimensions and emphysema in smokers. Correlation with lung function. Am J Respir Crit Care Med. 2000;162:1102–1108. doi: 10.1164/ajrccm.162.3.9907120. [DOI] [PubMed] [Google Scholar]
- 23.Hasegawa M, Nasuhara Y, Onodera Y, Makita H, Nagai K, Fuke S, Ito Y, Betsuyaku T, Nishimura M. Airflow limitation and airway dimensions in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2006;173:1309–1315. doi: 10.1164/rccm.200601-037OC. [DOI] [PubMed] [Google Scholar]
- 24.Coxson HO, Mayo J, Lam S, Santyr G, Parraga G, Sin DD. New and current clinical imaging techniques to study chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2009;180:588–597. doi: 10.1164/rccm.200901-0159PP. [DOI] [PubMed] [Google Scholar]
- 25.Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general U.S. Population. Am J Respir Crit Care Med. 1999;159:179–187. doi: 10.1164/ajrccm.159.1.9712108. [DOI] [PubMed] [Google Scholar]
- 26.Hogg JC, Macklem PT, Thurlbeck WM. Site and nature of airway obstruction in chronic obstructive lung disease. N Engl J Med. 1968;278:1355–1360. doi: 10.1056/NEJM196806202782501. [DOI] [PubMed] [Google Scholar]
- 27.Estepar RS, Washko GG, Silverman EK, Reilly JJ, Kikinis R, Westin CF. Accurate airway wall estimation using phase congruency. Med Image Comput Comput Assist Interv Int Conf Med Image Comput Comput Assist Interv. 2006;9:125–134. doi: 10.1007/11866763_16. [DOI] [PubMed] [Google Scholar]
- 28.Hu S, Hoffman EA, Reinhardt JM. Automatic lung segmentation for accurate quantitation of volumetric x-ray ct images. IEEE Trans Med Imaging. 2001;20:490–498. doi: 10.1109/42.929615. [DOI] [PubMed] [Google Scholar]
- 29.Tschirren J, McLennan G, Palagyi K, Hoffman EA, Sonka M. Matching and anatomical labeling of human airway tree. IEEE Trans Med Imaging. 2005;24:1540–1547. doi: 10.1109/TMI.2005.857653. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Becklake MR, Kauffmann F. Gender differences in airway behaviour over the human life span. Thorax. 1999;54:1119–1138. doi: 10.1136/thx.54.12.1119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.DeMeo DL, Hersh CP, Hoffman EA, Litonjua AA, Lazarus R, Sparrow D, Benditt JO, Criner G, Make B, Martinez FJ, Scanlon PD, Sciurba FC, Utz JP, Reilly JJ, Silverman EK. Genetic determinants of emphysema distribution in the national emphysema treatment trial. Am J Respir Crit Care Med. 2007;176:42–48. doi: 10.1164/rccm.200612-1797OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Hizawa N, Makita H, Nasuhara Y, Hasegawa M, Nagai K, Ito Y, Betsuyaku T, Konno S, Nishimura M. Functional single nucleotide polymorphisms of the ccl5 gene and nonemphysematous phenotype in COPD patients. Eur Respir J. 2008;32:372–378. doi: 10.1183/09031936.00115307. [DOI] [PubMed] [Google Scholar]
- 33.Shannon CE. Communication in the presence of noise. Proc IRE. 1949;37:10–21. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.