Abstract
Rationale: There is limited knowledge about the relationship between respiratory symptoms and quantitative high-resolution computed tomography measures of emphysema and airway wall thickness.
Objectives: To describe the ability of these measures of emphysema and airway wall thickness to predict respiratory symptoms in subjects with and without chronic obstructive pulmonary disease (COPD).
Methods: We included 463 subjects with chronic obstructive pulmonary disease (COPD) (65% men) and 488 subjects without COPD (53% men). All subjects were current or ex-smokers older than 40 years. They underwent spirometry and high-resolution computed tomography examination, and completed an American Thoracic Society questionnaire on respiratory symptoms.
Measurements and Main Results: Median (25th percentile, 75th percentile) percent low-attenuation areas less than −950 Hounsfield units (%LAA) was 7.0 (2.2, 17.8) in subjects with COPD and 0.5 (0.2, 1.3) in subjects without COPD. Mean (SD) standardized airway wall thickness (AWT) at an internal perimeter of 10 mm (AWT-Pi10) was 4.94 (0.33) mm in subjects with COPD and 4.77 (0.29) in subjects without COPD. Both %LAA and AWT-Pi10 were independently and significantly related to the level of dyspnea among subjects with COPD, even after adjustments for percent predicted FEV1. AWT-Pi10 was significantly related to cough and wheezing in subjects with COPD, and to wheezing in subjects without COPD. Odds ratios (95% confidence intervals) for increased dyspnea in subjects with COPD and in subjects without COPD were 1.9 (1.5–2.3) and 1.9 (0.6–6.6) per 10% increase in %LAA, and 1.07 (1.01–1.14) and 1.11 (0.99–1.24) per 0.1-mm increase in AWT-Pi10, respectively.
Conclusions: Quantitative computed tomography assessment of the lung parenchyma and airways may be used to explain the presence of respiratory symptoms beyond the information offered by spirometry.
There is limited knowledge about the association between respiratory symptoms and quantitative computed tomography (CT) measures of emphysema and airway wall thickness.
Quantitative CT measures of emphysema and airways may be used to explain the presence of respiratory symptoms in subjects with chronic obstructive pulmonary disease beyond the information available from spirometry alone. Level of dyspnea was predicted by quantitative CT measures of both emphysema and airway wall thickness, whereas the presence of coughing and wheezing was best predicted by airway wall thickness. There were only small sex differences.
Quantitative computed tomography (CT) has become a popular technique to assess the pathological changes in lung structure associated with COPD, and to separate the various subtypes of COPD according to the contribution of airway disease and emphysema (3–6). Investigators have examined the relationship between spirometry and quantitative CT measures of emphysema and airway wall thickness (5, 7–11). They found significant correlations, but spirometry could not explain the whole variation of the quantitative CT measures.
Only a few studies have previously examined the relationship between quantitative CT measures of emphysema and airway wall thickness, and respiratory symptoms. In a study of 51 patients with COPD, Camiciottoli and colleagues (12) found that the degree of emphysema measured by CT was associated with the level of dyspnea. In a larger study of 1,159 subjects, Patel and colleagues (5) observed that CT-derived airway wall thickness was greater in subjects with cough and phlegm as compared with those without. However, no adjustment was made for FEV1 in that study.
Men and women report respiratory symptoms differently (13), but neither of two studies examined whether the relationships between respiratory symptoms and CT differed by sex (5, 12). Hence, knowledge is limited regarding CT-measured indices of emphysema and airway wall thickness and their ability to predict respiratory symptoms in subjects with COPD, and it is unknown whether these CT variables may be used to predict respiratory symptoms in subjects without COPD.
The objective of the present study was to describe the independent relationship between respiratory symptoms of COPD and quantitative high-resolution computed tomography (HRCT) measures of emphysema (percent low-attenuation areas less than −950 Hounsfield units, %LAA) and airway wall thickness (AWT) at an internal perimeter of 10 mm (AWT-Pi10). Furthermore we wanted to assess whether these relationships varied between subjects with and without COPD, and between sexes. We hypothesized that there would be a significant relationship between the CT-measured degree of emphysema and the level of dyspnea, and between airway wall thickness and symptoms of chronic bronchitis such as cough, phlegm, and wheezing. Some of the results of this study have been previously reported in the form of an abstract (14, 15).
The study was conducted between January 2003 and January 2005 in Bergen, Norway. All subjects gave informed consent, and the local ethics committee and review board approved the study. The subjects included in the current study were recruited from a hospital patient register and from two general population studies, and all were participants in the GenKOLS (Genetic COPD Study) (16, 17). GenKOLS was a genetic association study aimed at examining single-nucleotide polymorphism and haplotype differences between smokers with and without COPD. The subjects of the current study constituted the approximate half of the GenKOLS population (951 of 1,909) that received an optional high resolution computed tomography (HRCT) scan (18). A total of 463 subjects with COPD and 488 subjects without COPD were included. Enrollment criteria were as follows: (1) white; (2) age above 40 years; (3) current or former smoker with at least 2.5 pack-years of smoking history; and (4) no α1-antitrypsin deficiency. A complete list of inclusion and exclusion criteria is given elsewhere (18).
All subjects completed an American Thoracic Society questionnaire (19), including the Medical Research Council Dyspnea Scale (MRCDS) (20), three questions on coughing (morning cough, chronic cough, and phlegm cough), and wheezing attack. The exact wording of the questions is given in the online supplement (Table E1). The minority (<5%) that did not answer, or answered “Don't know,” on a respiratory symptom question, was coded as having answered “No” to that particular question.
Subjects were examined by spirometry according to American Thoracic Society standards, using a Vitalograph 2160 spirometer (Vitalograph Ltd, Maids Moreton, UK), before and after bronchodilation with 0.4 mg of salbutamol. Subjects were assessed at least 6 weeks after any respiratory infection, but were not asked to withhold regular medication. Norwegian reference values for FEV1 and FVC were used (21). Subjects with COPD had a postbronchodilator FEV1/FVC ratio less than 0.70 and FEV1 less than 80% of predicted (2). Subjects without COPD had a postbronchodilator FEV1/FVC ratio greater than 0.70 and FEV1 greater than 80% of predicted.
The HRCT scans were acquired with a GE LightSpeed Ultra CT scanner (120 kVp, 200 mA; GE Healthcare, Piscataway, NJ), at suspended full inspiration (apex to base) using a 1-mm slice thickness at 20-mm intervals (18). The CT scans were reconstructed using both a low spatial frequency reconstruction algorithm (“standard”) for density measurements, and a high spatial frequency algorithm (“bone”) for airway measurements. CT-derived total lung volume was measured, and a proxy for inflation level was obtained by dividing that volume by the predicted total lung capacity (22). This proxy was used as a separate adjustment factor in the logistic regression models, but was not part of the calculation of the extent of emphysema or airway wall thickness. The extent of emphysema was assessed primarily on the basis of the percentage of lung voxels with X-ray attenuation values less than −950 Hounsfield units (HU) (percentage low-attenuation areas, %LAA). The cutoff of −950 HU has been shown to be appropriate for this CT acquisition technique (23), but for the sake of completeness analyses were also performed using other published cutoffs (−910 and −856 HU). Airways cut in cross-section (short-to-long axis greater than two thirds) were identified on the CT scans and measured using the full-width-at-half-maximum algorithm (4). To reduce the technical errors associated with small airways, we included only airways with an internal perimeter greater than 6 mm. To avoid potential bias from different distribution of airway sizes between subjects, a standardized measure for airway wall thickness (AWT-Pi10) was derived for each subject by plotting the square root of the airway wall area against the internal perimeter of each measured airway (18). The resulting regression line was used to calculate the square root of the wall area for a “theoretical airway” with an internal perimeter of 10 mm (AWT-Pi10) (5). More details regarding the HRCT methodology used in this study can be found elsewhere (18).
Proportions were tested using Pearson's chi square test (24), means using one-way analysis of variance, and medians using the Kruskal-Wallis test (25). Means are reported with standard deviations. The associations of the main explanatory variables (%LAA and AWT-Pi10) to the respiratory symptoms were examined with multiple logistic regression analyses. For the MRCDS a multiple ordinal logistic regression model was used (26). Results are reported in terms of odds ratios. The analyses were performed separately for subjects with COPD and subjects without COPD, and adjustments were made for sex, age, pack-years, current smoking status, and inflation level (CT-measured lung volume as a fraction of the predicted total lung capacity) in the standard model (Table 3). We also added adjustments for percent predicted FEV1, single-breath diffusing capacity of carbon monoxide (DlCO), body mass index (BMI), and comorbid heart disease in alternative models. Sex × %LAA and sex × AWT-Pi10 interactions were also added separately to the standard model. A significance level of 0.05 was applied. All analyses were performed with Stata 10 (StataCorp LP, College Station, TX).
Table 1 shows the characteristics of the subjects with COPD and subjects without COPD, by sex. There was a higher percentage of men among the subjects with COPD than among the subjects without COPD. The subjects with COPD were older, were more likely to be current smokers, and had a larger total smoking consumption than the subjects without COPD. By design, no subjects in Global Initiative for Chronic Obstructive Lung Disease (GOLD) stage I were included in the study (16, 17). Almost 60% of the subjects with COPD were in GOLD stage II, and approximately one quarter were in GOLD stage III. There was no sex difference in the distribution of GOLD stages. About 15% of the subjects without COPD reported at least dyspnea grade 2 in the MRCDS, whereas the corresponding figure among the subjects with COPD was 70%. The cough, phlegm, and wheezing symptoms were two to five times more frequent in those with COPD compared with those without COPD. In both groups dyspnea grade 2 or higher and wheezing were reported more often by women than men, whereas phlegm was more frequently reported by men.
Subjects with COPD | Subjects without COPD | |||||||
|---|---|---|---|---|---|---|---|---|
| Men (n = 299) | Women (n = 164) | Men (n = 260) | Women (n = 228) | |||||
| Age (yr), mean (SD) | 65.2 (9.4) | 62.6 (9.0) | 56.2 (9.7) | 54.7 (9.1) | ||||
| Current smokers, % | 46.82 | 53.66 | 37.31 | 46.93 | ||||
| Pack-years, median (25th, 75th percentiles) | 30.6 (20.8, 43.0) | 22.5 (16.4, 32.1) | 18.0 (10.3, 27.1) | 14.2 (8.0, 23.3) | ||||
| PB FEV1%pred, mean (SD) | 52.6 (17.4) | 53.1 (16.1) | 94.2 (8.6) | 95.8 (9.3) | ||||
| DlCO, mean (SD) | 6.1 (1.9) | 4.8 (1.3) | 8.7 (1.7) | 6.3 (1.2) | ||||
| GOLD category, % | ||||||||
| Stage II | 58.2 | 62.8 | — | — | ||||
| Stage III | 28.1 | 28.1 | — | — | ||||
| Stage IV | 13.7 | 9.2 | — | — | ||||
| MRCDS, % | ||||||||
| Grade 1 | 31.4 | 26.3 | 89.6 | 80.9 | ||||
| Grade 2 | 33.1 | 35.0 | 9.3 | 15.6 | ||||
| Grade 3 | 20.0 | 21.3 | 1.2 | 3.1 | ||||
| Grade 4 | 10.3 | 9.4 | 0.0 | 0.0 | ||||
| Grade 5 | 5.2 | 8.1 | 0.0 | 0.4 | ||||
| Respiratory symptoms (% yes) | ||||||||
| Morning cough | 54.9 | 51.2 | 20.8 | 19.3 | ||||
| Chronic cough | 47.5 | 50.6 | 15.4 | 15.8 | ||||
| Phlegm cough | 65.6 | 62.8 | 30.8 | 24.1 | ||||
| Wheezing attacks | 34.5 | 45.7 | 6.9 | 9.7 | ||||
Subjects with COPD had significantly more emphysema (P < 0.0001) and significantly greater airway wall thickness (P < 0.0001) than did subjects without COPD (Table 2). We also performed qualitative/semiquantitative analyses on the material (data not shown). We found a good correlation between %LAA and the visually assessed semiquantitative CT data on extent of emphysema in the whole sample (R2 = 0.51) and among subjects with COPD (R2 = 0.51), but only a weak correlation among subjects without COPD (R2 = 0.02).
%LAA | AWT-Pi10 | |||||||
|---|---|---|---|---|---|---|---|---|
| Subjects with COPD (n = 462) (Median) | Subjects without COPD (n = 485) (Median) | Subjects with COPD (n = 438) (Mean) | Subjects without COPD (n = 455) (Mean) | |||||
| All subjects | 7.0 | 0.5 | 4.94 | 4.77 | ||||
| 25th, 75th percentiles (%LAA); SD (AWT-Pi10) | (2.2, 17.8) | (0.2, 1.3) | (0.33) | (0.29) | ||||
| MRCDS | ||||||||
| 1 | 3.1 | 0.5 | 4.93 | 4.76 | ||||
| 2 | 6.7 | 0.7 | 4.96 | 4.78 | ||||
| 3 | 12.2 | 0.6 | 4.94 | 5.00 | ||||
| 4 | 18.2 | — | 4.96 | — | ||||
| 5 | 23.3 | 0.1 | 4.90 | 4.44 | ||||
| P value | <0.001 | 0.178 | 0.892 | 0.086 | ||||
| Morning cough | ||||||||
| Yes | 8.0 | 0.7 | 4.98 | 4.81 | ||||
| No | 6.3 | 0.5 | 4.90 | 4.75 | ||||
| P value | 0.072 | 0.286 | 0.014 | 0.117 | ||||
| Chronic cough | ||||||||
| Yes | 7.8 | 0.7 | 4.98 | 4.80 | ||||
| No | 6.5 | 0.5 | 4.90 | 4.76 | ||||
| P value | 0.109 | 0.108 | 0.011 | 0.291 | ||||
| Phlegm cough | ||||||||
| Yes | 8.7 | 0.5 | 4.95 | 4.79 | ||||
| No | 4.9 | 0.5 | 4.93 | 4.76 | ||||
| P value | 0.003 | 0.985 | 0.598 | 0.273 | ||||
| Wheezing attacks | ||||||||
| Yes | 8.0 | 0.9 | 4.96 | 4.87 | ||||
| No | 6.5 | 0.5 | 4.93 | 4.76 | ||||
| P value | 0.102 | 0.032 | 0.292 | 0.022 | ||||
Table 2 shows the crude relationships between each of the CT indices and the respiratory symptoms. The MRCDS increased with increasing level of %LAA in subjects with COPD, whereas airway wall thickness did not affect the degree of dyspnea. We found more emphysema (%LAA) and thicker airway walls (AWT-Pi10) in the subjects who answered yes to the cough and wheezing attack questions. These differences reached the level of significance for phlegm cough versus emphysema in subjects with COPD, for all cough symptoms versus airway wall thickness in subjects with COPD, and for wheezing versus both emphysema and airway wall thickness in subjects without COPD.
After adjusting for sex, age, smoking, and level of inflation we observed a significant relationship between increasing %LAA and increasing dyspnea level in the COPD group (Table 3). The ordinal multiple regression analysis showed that the odds for achieving a dyspnea score greater than or equal to a chosen level on the MRCDS was increased by an estimated factor of 1.87 if the %LAA was increased by 10 percentage points, regardless of the chosen MRCD level, ceteris paribus. An adjusted odds ratio of similar size was noted in those without COPD, but the confidence interval was wide and included unity (Table 3). None of the other respiratory symptoms varied significantly with level of emphysema in the multivariate model, neither in subjects with COPD nor in subjects without COPD.
%LAA per 10% | AWT-Pi10 per 0.1 mm | |||||
|---|---|---|---|---|---|---|
| Dependent Variable | Subjects with COPD (n = 462) [OR (95% CI)] | Subjects without COPD (n = 485) [OR (95% CI)] | Subjects with COPD (n = 438) [OR (95% CI)] | Subjects without COPD (n = 455) [OR (95% CI)] | ||
| MRCDS 1–5 (ordinal) | 1.87 (1.52 to 2.31) | 1.93 (0.56 to 6.66) | 1.07 (1.01 to 1.14) | 1.11 (0.99 to 1.24) | ||
| Morning cough | 1.08 (0.86 to 1.35) | 0.98 (0.26 to 3.73) | 1.09 (1.02 to 1.16) | 1.06 (0.96 to 1.16) | ||
| Chronic cough | 1.19 (0.95 to 1.49) | 1.01 (0.26 to 4.03) | 1.12 (1.04 to 1.20) | 1.03 (0.93 to 1.14) | ||
| Phlegm cough | 1.15 (0.90 to 1.47) | 1.45 (0.54 to 3.89) | 1.03 (0.96 to 1.11) | 1.00 (0.92 to 1.09) | ||
| Wheezing attacks | 1.10 (0.87 to 1.39) | 1.97 (0.63 to 6.19) | 1.08 (1.01 to 1.15) | 1.21 (1.07 to 1.37) | ||
Increasing airway wall thickness was significantly associated with impaired dyspnea level and presence of morning and chronic cough symptoms among the patients with COPD in the multiple logistic regression analyses (Table 3). In both subjects with COPD and subjects without COPD increasing airway wall thickness was significantly related to wheezing in the multiple logistic regression analyses.
We also performed the same analyses as shown in Table 3 without inflation level as an adjustment factor of the relationships between %LAA and AWT-Pi10, and the respiratory symptoms (data not shown). Without inflation level in the model, the explained variation (pseudo-R2) decreased by 0–3% in all models as compared with including the inflation level, but the relationships between %LAA and cough and wheezing symptoms actually gained significance among subjects with COPD.
Separate analyses on the relationship between percent predicted FEV1 and %LAA and AWT-Pi10 showed an inverse relationship between %LAA and percent predicted FEV1 in subjects with COPD, and an inverse relationship between AWT-Pi10 and percent predicted FEV1 in subjects without COPD. When we added percent predicted FEV1 to the models presented in Table 3, the adjusted odds ratios for degree of dyspnea by %LAA and by AWT-Pi10 were only slightly reduced and still statistically significant among subjects with COPD (Figures 1A and 1B). Similarly, the associations between airway wall thickness and symptoms of coughing and wheezing were not notably changed by the additional adjustments for percent predicted FEV1 (data not shown). Additional analysis adding DlCO, BMI, and comorbidity in terms of cardiovascular disease to the model did not alter the relationships described previously (data not shown).
Figure 1. (A) Adjusted odds ratio with 95% confidence interval for increased dyspnea (Medical Research Council Dyspnea Scale, MRCDS) per 10% increase in percentage low-attenuation areas (%LAA) before (solid circles) and after (open circles) adjustment for percentage of predicted FEV1 (FEV1%pred). The multiple ordinal logistic regression analyses were run separately for 462 subjects with chronic obstructive pulmonary disease (COPD) and 485 subjects without COPD. Adjustments were made for sex, age (40 to 86 yr), pack-years (2.5 to 119), current smoking (Ex-, <10, 10–20, and >20), and inflation level (0.2 to 1.9) (B) Adjusted odds ratio with 95% confidence interval for increased dyspnea (MRCDS) per 0.1-mm increase in airway wall thickness at an internal perimeter of 10 mm (AWT-Pi10) before (solid circles) and after (open circles) adjustment for FEV1%pred. The multiple ordinal logistic regression analyses were run separately for 438 subjects with COPD and 455 subjects without COPD. Adjustments were made for sex, age (40 to 86 yr), pack-years (2.5 to 119), current smoking (Ex-, <10, 10–20, and >20), and inflation level (0.2 to 1.9).
[More] [Minimize]We also added both %LAA and AWT-Pi10 into the same model as in Table 3, with and without additional adjustment for percent predicted FEV1. Both variables were significantly and independently related to the level of dyspnea in subjects with COPD, whereas the relationship between AWT-Pi10 and MRCDS reached borderline significance in subjects without COPD. The odds ratio (95% confidence interval) for increased dyspnea in subjects with COPD in this combined model was 1.96 (1.58 to 2.44) per 10% increase in %LAA and 1.12 (1.05 to 1.19) per 0.1-mm increase in AWT-Pi10.
Interaction analysis by sex showed that the relationship of %LAA to phlegm cough differed significantly between male and female subjects with COPD. The adjusted odds ratios were 1.46 (1.04 to 2.04) in male subjects with COPD and 0.82 (0.54 to 1.25) in female subjects with COPD, after adjusting for age, pack-years, current smoking, and inflation level. We observed no other significant interactions between sex and the two quantitative HRCT indices.
We also performed the analyses on the whole sample (subjects with COPD and subjects without COPD combined), using the same model as in Table 3, but with additional adjustment for COPD/non-COPD status (see Table E2 in the online supplement). These analyses revealed similar relationships between CT parameters and respiratory symptoms as we found among subjects with COPD (Table 3). Interaction analyses on the whole sample revealed no interactions between COPD/non-COPD status and %LAA or AWT-Pi10, respectively.
Regarding the %LAA cutoff and the relation to dyspnea, both −910 and −856 HU yielded similar results as −950 HU among subjects with COPD, but −910 HU was weaker and −856 HU was not significant. There was no difference between cutoffs among subjects without COPD regarding dyspnea. Regarding cough and wheezing symptoms, there were no significant differences between cutoffs among either subjects with COPD or subjects without COPD.
This is the first study to investigate the association of CT-measured indices of both emphysema and airway wall thickness to respiratory symptoms, after extensive adjustment for potential confounders. %LAA and AWT-Pi10 were independently and significantly related to dyspnea among subjects with COPD, even after allowing for percent predicted FEV1. AWT-Pi10 was significantly related to morning cough, chronic cough, and wheezing attacks in subjects with COPD and to wheezing attacks in subjects without COPD. Interaction analyses revealed that %LAA was significantly associated with phlegm cough in male subjects with COPD, but not in female subjects with COPD.
Although the difference in %LAA between subjects with COPD and subjects without COPD was large and significant, the difference in AWT-Pi10 between subjects with COPD and subjects without COPD was much smaller. However, the difference was still significant (P < 0.0001), and both measures are clearly related to COPD. The sizes of these differences have been discussed previously (18).
That increasing emphysema measured as %LAA was associated with increasing dyspnea is in line with a previous study (12) of 51 patients with COPD. However, that study included only five women, and no adjustments for airway wall thickness or lung function were made. We have extended this knowledge by showing a relationship between %LAA and dyspnea in both male and female subjects with COPD irrespective of percent predicted FEV1.
The significant relationship between airway wall thickness and dyspnea shown in the multiple ordinal logistic regression analyses (Table 3) has not previously been explored, and it was significant only among subjects with COPD. It is interesting to note that the significant differences became apparent only after adjustments were made, and not in the crude values shown in Table 2. Furthermore, there was no significant interaction between COPD/non-COPD status and AWT-Pi10 in the effect on dyspnea when analyzing the whole material (subjects with COPD and subjects without COPD together). The significant relationship between dyspnea and AWT-Pi10 among subjects with COPD, and the lack of significance among subjects without COPD, must therefore be interpreted with caution.
There may be several explanations for the independent association of these CT indices with the level of dyspnea. In emphysematous patients there is a reduced alveolar surface over which gas exchange may occur, and a reduced pulmonary capillary blood volume (27). Emphysema results in static and dynamic hyperinflation, which implies that volume changes take place over a suboptimal part of the pressure–volume curve, and that the respiratory muscles may not operate on an optimal part of the length–tension relationship. The work of breathing is increased and dynamic hyperinflation has been shown to be related to dyspnea (28). Loss of elastic fibers may be seen both in emphysema and in airway wall remodeling, causing increased lung and airway compliance (28). Increased airway wall thickness may reflect systemic inflammation and impaired muscle mass, causing exhaustion and breathlessness on light exertion (29).
Cardiac comorbidity may also cause dyspnea (30), but was partly adjusted for by adding self-reported heart disease to the models without affecting the overall findings.
The relationships between level of FEV1 and dyspnea, and between DlCO and dyspnea, have previously been shown to be stronger in men than women (31, 32), both in patients with COPD and in community studies. In this study, the relationship between the CT indices and dyspnea did not differ by sex. This difference could be due to the fact that whereas FEV1 and DlCO are physiological variables, %LAA and AWT-Pi10 are anatomical or structural measures. If so, this implies that previously observed sex differences in perception of dyspnea are not related to the anatomical structure of the lung. It has been speculated that the sex interactions in dyspnea are partly driven by premenopausal women being more sensitive to changes in the bronchial mucosa during the menstrual cycle (13). However, in our study most of the women were likely postmenopausal as 78% were above the age of 50 years, and excluding those below the age of 50 years did not alter the results.
%LAA was associated with phlegm cough in subjects with COPD and with wheezing in subjects without COPD in the bivariate analyses. After adjustments for sex, age, pack-years, current smoking status, and inflation level, these relationships lost statistical significance. However, interaction analyses showed that %LAA was significantly associated with phlegm cough in male subjects with COPD, but not in female subjects with COPD. It has previously been speculated that women tend to underreport phlegm partly for social reasons (13, 33). Another explanation might be that women are more prone to have a COPD phenotype with less phlegm.
The observed bivariate relationships between AWT-Pi10 and morning cough, chronic cough, and wheezing attack remained significant after extensive adjustment for confounders, including FEV1, BMI, and comorbid heart disease, and the relationship with wheezing gained significance among subjects with COPD. This finding is in line with findings by Patel and colleagues (5). However, they only reported crude AWT-Pi10 levels in those with and without respiratory symptoms, and did not allow for age, sex, smoking history, lung function, or inflation level. As we have shown (18), there are considerable differences between men and women regarding both airway size and thickness (AWT-Pi10), and this warrants sex adjustment or separate analyses between sexes. The greater airway wall thickness in symptomatic individuals may reflect increased airway inflammation, increased density of goblet cells, and more bronchial mucus (34–37), which may all cause cough, phlegm, and wheezing.
Community studies have shown that a large number of healthy subjects, especially smokers, experience persistent respiratory symptoms (38). In the current study, significant associations between respiratory symptoms and CT indices were found among subjects with COPD. The odds ratios were similar among subjects without COPD, but did not reach statistical significance except for AWT-Pi10 versus attacks of wheezing. This could indicate that the same relationships are true for subjects without COPD, but that we either did not have enough power to prove them, or that the noise of the measurements obscured a relatively weak signal among the subjects without COPD. However, %LAA varied between 0.1 and 2.8 (10th and 90th percentiles) in all subjects without COPD compared with a median %LAA of 3.1 in subjects with COPD without dyspnea. Similarly, for all symptomatic subjects without COPD, the mean AWT-Pi10 was lower than the mean AWT-Pi10 in asymptomatic subjects with COPD. This may imply that most of the respiratory symptoms seen in healthy smokers are due to factors not visible on CT, and that symptoms precede the physiological and structural changes that can be assessed by pulmonary function tests and HRCT. Longitudinal studies will be necessary to examine whether the combination of respiratory symptoms in healthy smokers and small changes in %LAA and AWT-Pi10 can predict the development of COPD.
This study is a large single-center study, which allows for extensive adjustment for important confounders. All subjects without COPD were sampled from a community study with high response rates (39–41). Furthermore, a survey of nonresponders has shown that the study samples are representative of the population at large with respect to sex, age, and smoking (39). Finally, all CT scans were performed with the same scanner.
There are some limitations to this study. First, the respiratory symptoms were recorded only once. Especially in subjects without COPD the respiratory symptoms may be transient (42), which could weaken the observed relationships between the CT indices and the symptoms. A longitudinal study would enable a distinction between persistent and transient symptoms and allow an assessment of how sensitive the CT measurements are to changes in respiratory symptoms. Second, even though subjects with a total smoking consumption as low as 2.5 pack-years were included in the study, we did not include never-smokers. Third, the study did not include GOLD stage I. However, GOLD stages II–IV are those regarded as clinically important (43). Fourth, the CT scans were not spirometrically gated, but it has been shown that the repeatability of quantitative CT tests is high and not likely to improve by using spirometric gating (44). Furthermore, all analyses were adjusted for inflation level (18). Fifth, we used a slice/gap CT technique, instead of the newer volumetric approach, and this could have weakened the relationships between AWT-Pi10 and the respiratory symptoms. As we were not able to determine the generations of the airways measured, the generations might have differed between those with small and large lungs. But adjustments for lung volume, lumen size, or number of airways measured did not alter the relationships observed in Table 3. Finally, the airway wall thickness was estimated using the CT technique of full width at half maximum in larger airways, whereas COPD is known as a small-airway disease (34, 45). There are numerous CT techniques available to measure airways and although it can be argued that some techniques are better than others no single algorithm has risen to be the ultimate answer for airway wall dimensions (46). Acknowledging this limitation, our study is based on previous studies that have shown that the inflammatory response seen in small airways can also be seen in large airways (47), and that CT estimates of airway wall thickness are correlated with histological measurements of small-airway wall dimensions (4).
Our study supports the clinical experience that FEV1 is far from explaining the whole picture of COPD, and shows that quantitative CT assessment of both the lung parenchyma and the airways may be used to explain the presence of respiratory symptoms beyond the information offered by spirometry. Although %LAA was best at predicting dyspnea and AWT-Pi10 was associated with dyspnea, cough, and wheezing, both variables showed these associations primarily among the subjects with COPD. Future studies should be longitudinal in design, and should explore whether the relatively small changes in the quantitative CT measures among symptomatic subjects without COPD could predict the development of COPD.
The authors thank Claudine Storness-Bliss for valuable help with the quantitative analysis and Anh-Toan Tran for development and maintenance of the analysis software and database.
| 1. | Lopez AD, Shibuya K, Rao C, Mathers CD, Hansell AL, Held LS, Schmid V, Buist S. Chronic obstructive pulmonary disease: current burden and future projections. Eur Respir J 2006;27:397–412. |
| 2. | Rabe KF, Hurd S, Anzueto A, Barnes PJ, Buist SA, Calverley P, Fukuchi Y, Jenkins C, Rodriguez-Roisin R, van Weel C, et al. 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. |
| 3. | de Jong PA, Long FR, Wong JC, Merkus PJ, Tiddens HA, Hogg JC, Coxson HO. Computed tomographic estimation of lung dimensions throughout the growth period. Eur Respir J 2006;27:261–267. |
| 4. | Nakano Y, Wong JC, de Jong PA, Buzatu L, 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. |
| 5. | Patel BD, Coxson HO, Pillai SG, Agusti AG, Calverley PM, Donner CF, Make BJ, Muller NL, Rennard SI, Vestbo J, et al. Airway wall thickening and emphysema show independent familial aggregation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2008;178:500–505. |
| 6. | Dirksen A. Is CT a new research tool for COPD? Clin Respir J 2008;2:76–83. |
| 7. | Martinez FJ, Curtis JL, Sciurba F, Mumford J, Giardino ND, Weinmann G, Kazerooni E, Murray S, Criner GJ, Sin DD, et al. Sex differences in severe pulmonary emphysema. Am J Respir Crit Care Med 2007;176:243–252. |
| 8. | Dransfield MT, Washko GR, Foreman MG, Estepar RS, Reilly J, Bailey WC. Gender differences in the severity of CT emphysema in COPD. Chest 2007;132:464–470. |
| 9. | Gould GA, Redpath AT, Ryan M, Warren PM, Best JJ, Flenley DC, MacNee W. Lung CT density correlates with measurements of airflow limitation and the diffusing capacity. Eur Respir J 1991;4:141–146. |
| 10. | Nakano Y, Muro S, Sakai H, Hirai T, Chin K, Tsukino M, Nishimura K, Itoh H, Pare PD, Hogg JC, et al. Computed tomographic measurements of airway dimensions and emphysema in smokers: correlation with lung function. Am J Respir Crit Care Med 2000;162:1102–1108. |
| 11. | Washko GR, Criner GJ, Mohsenifar Z, Sciurba FC, Sharafkhaneh A, Make BJ, Hoffman EA, Reilly JJ. Computed tomographic–based quantification of emphysema and correlation to pulmonary function and mechanics. COPD 2008;5:177–186. |
| 12. | Camiciottoli G, Bartolucci M, Maluccio NM, Moroni C, Mascalchi M, Giuntini C, Pistolesi M. Spirometrically gated high-resolution CT findings in COPD: Lung attenuation vs lung function and dyspnea severity. Chest 2006;129:558–564. |
| 13. | Becklake MR, Kauffmann F. Gender differences in airway behaviour over the human life span. Thorax 1999;54:1119–1138. |
| 14. | Grydeland TB, Coxson HO, Dirksen A, Sharma S, Pillai SG, Anderson W, Gulsvik A, Bakke PS. Indications of airway disease and emphysema estimated by HRCT: relation to respiratory symptoms [abstract]. Poster presentation at ATS 2008—Toronto International Conference, May 16–21, 2008. Available from: http://www.abstracts2view.com/ats08/view.php?nu=ATS08L_4387. |
| 15. | Grydeland TB, Coxson HO, Dirksen A, Sharma S, Pillai SG, Anderson W, Gulsvik A, Bakke PS. Gender differences in the relation between respiratory symptoms and emphysema estimated by HRCT [abstract]. Oral presentation at the ERS Annual Congress 2008—Berlin, October 4–8, 2008. Available from: http://www.ersnet.org/learning_resources_player/abstract_print_08/files/42.pdf. |
| 16. | Sørheim IC, Johannessen A, Grydeland TB, Omenaas ER, Gulsvik A, Bakke PS. Case–control studies on risk factors for COPD: how does the sampling of the cases and controls affect the results? The Clinical Respiratory Journal 2009 [Published ahead of print 8 Jul 2009 as DOI: 10.1111/j.1752-699X.2009.00154.x. |
| 17. | Zhu G, Warren L, Aponte J, Gulsvik A, Bakke P, Anderson WH, Lomas DA, Silverman EK, Pillai SG. The serpine2 gene is associated with chronic obstructive pulmonary disease in two large populations. Am J Respir Crit Care Med 2007;176:167–173. |
| 18. | 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. |
| 19. | Ferris BG. Epidemiology Standardization Project (American Thoracic Society). Am Rev Respir Dis 1978;118:1–120. |
| 20. | Bestall JC, Paul EA, Garrod R, Garnham R, Jones PW, Wedzicha JA. Usefulness of the medical research council (MRC) dyspnoea scale as a measure of disability in patients with chronic obstructive pulmonary disease. Thorax 1999;54:581–586. |
| 21. | Gulsvik A, Tosteson T, Bakke P, Humerfelt S, Weiss ST, Speizer FE. Expiratory and inspiratory forced vital capacity and one-second forced volume in asymptomatic never-smokers in Norway. Clin Physiol 2001;21:648–660. |
| 22. | Miller MR, Crapo R, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A, Enright P, van der Grinten CP, Gustafsson P, et al. General considerations for lung function testing. Eur Respir J 2005;26:153–161. |
| 23. | Gevenois PA, de Maertelaer V, De Vuyst P, Zanen J, Yernault JC. Comparison of computed density and macroscopic morphometry in pulmonary emphysema. Am J Respir Crit Care Med 1995;152:653–657. |
| 24. | Lydersen S, Fagerland MW, Laake P. Recommended tests for association in 2 × 2 tables. Stat Med 2009;28:1159–1175. |
| 25. | Kruskal W, Wallis WA. Use of ranks in one-criterion variance analysis. J Am Stat Assoc 1952;47:583–621. |
| 26. | Kleinbaum DG, Klein M. Logistic regression, a self-learning text. Berlin: Springer; 2005. |
| 27. | Thurlbeck WM, Muller NL. Emphysema: definition, imaging, and quantification. AJR Am J Roentgenol 1994;163:1017–1025. |
| 28. | O'Donnell DE, Laveneziana P. Dyspnea and activity limitation in COPD: mechanical factors. COPD 2007;4:225–236. |
| 29. | Wouters EF. Local and systemic inflammation in chronic obstructive pulmonary disease. Proc Am Thorac Soc 2005;2:26–33. |
| 30. | Iwamoto H, Yokoyama A, Kitahara Y, Ishikawa N, Haruta Y, Yamane K, Hattori N, Hara H, Kohno N. Airflow limitation in smokers is associated with subclinical atherosclerosis. Am J Respir Crit Care Med 2009;179:35–40. |
| 31. | Gulsvik A. Prevalence of respiratory symptoms in the city of Oslo. Scand J Respir Dis 1979;60:275–285. |
| 32. | Welle I, Eide GE, Bakke PS, Gulsvik A. The single-breath transfer factor for carbon monoxide and respiratory symptoms in a Norwegian community sample. Eur Respir J 1999;14:1320–1325. |
| 33. | Gulsvik A. Prevalence and manifestations of obstructive lung disease in the city of Oslo. Scand J Respir Dis 1979;60:286–296. |
| 34. | Hogg JC. Pathophysiology of airflow limitation in chronic obstructive pulmonary disease. Lancet 2004;364:709–721. |
| 35. | James AL, Wenzel S. Clinical relevance of airway remodelling in airway diseases. Eur Respir J 2007;30:134–155. |
| 36. | McParland BE, Macklem PT, Pare PD. Airway wall remodeling: friend or foe? J Appl Physiol 2003;95:426–434. |
| 37. | Saetta M, Turato G, Baraldo S, Zanin A, Braccioni F, Mapp CE, Maestrelli P, Cavallesco G, Papi A, Fabbri LM. Goblet cell hyperplasia and epithelial inflammation in peripheral airways of smokers with both symptoms of chronic bronchitis and chronic airflow limitation. Am J Respir Crit Care Med 2000;161:1016–1021. |
| 38. | Eagan TM, Bakke PS, Eide GE, Gulsvik A. Incidence of asthma and respiratory symptoms by sex, age and smoking in a community study. Eur Respir J 2002;19:599–605. |
| 39. | Eagan TM, Eide GE, Gulsvik A, Bakke PS. Nonresponse in a community cohort study: predictors and consequences for exposure–disease associations. J Clin Epidemiol 2002;55:775–781. |
| 40. | Bakke P, Gulsvik A, Lilleng P, Overa O, Hanoa R, Eide GE. Postal survey on airborne occupational exposure and respiratory disorders in Norway: causes and consequences of non-response. J Epidemiol Community Health 1990;44:316–320. |
| 41. | Brogger J, Bakke P, Eide GE, Gulsvik A. Contribution of follow-up of nonresponders to prevalence and risk estimates: a Norwegian respiratory health survey. Am J Epidemiol 2003;157:558–566. |
| 42. | Eagan TM, Gulsvik A, Eide GE, Bakke PS. Remission of respiratory symptoms by smoking and occupational exposure in a cohort study. Eur Respir J 2004;23:589–594. |
| 43. | Buist AS, McBurnie MA, Vollmer WM, Gillespie S, Burney P, Mannino DM, Menezes AM, Sullivan SD, Lee TA, Weiss KB, et al. International variation in the prevalence of COPD (the Bold Study): a population-based prevalence study. Lancet 2007;370:741–750. |
| 44. | Gierada DS, Yusen RD, Pilgram TK, Crouch L, Slone RM, Bae KT, Lefrak SS, Cooper JD. Repeatability of quantitative ct indexes of emphysema in patients evaluated for lung volume reduction surgery. Radiology 2001;220:448–454. |
| 45. | Hogg JC, Chu F, Utokaparch S, Woods R, Elliott WM, Buzatu L, Cherniack RM, Rogers RM, Sciurba FC, Coxson HO, et al. The nature of small-airway obstruction in chronic obstructive pulmonary disease. N Engl J Med 2004;350:2645–2653. |
| 46. | Coxson HO. Quantitative chest tomography in COPD research: chairman's summary. Proc Am Thorac Soc 2008;5:874–877. |
| 47. | Tiddens HA, Pare PD, Hogg JC, Hop WC, Lambert R, de Jongste JC. Cartilaginous airway dimensions and airflow obstruction in human lungs. Am J Respir Crit Care Med 1995;152:260–266. |
