Results

Patient characteristics, lung function data, and current medications are shown in table 1. As expected, FEV₁ was lower in patients with NFA and moderate asthma than in those with mild asthma and normal controls. The mean (SD) FEV₁ as a percentage of predicted was 68 (20)% for patients in group 1, 73 (12)% for group 2, 102 (12)% for group 3, and 103 (12)% for group 4. There were no statistically significant differences in FEV₁between either groups 1 and 2, or groups 3 and 4. However, both groups 1 and 2 had a significantly lower FEV₁ than either group 3 or 4 (p < 0.0001). Patients in groups 1 and 2 had a higher RV and FRC than those in group 3 (p<0.05; table 1).

The results of HRCT scan measurements are presented in table 2. Figure1 shows representative images from three different subjects. There were no statistically significant differences between the groups in the mean number of bronchi measured per patient (table 2). The mean T/D ratios and WA% were 0.27 (0.05) and 78.0 (9.2)% for patients in group 1, 0.27 (0.05) and 78.8 (9.2)% for those in group 2, 0.25 (0.04) and 74.2 (7.5)% for patients in group 3, and 0.23 (0.04) and 70.9 (8.2)% for the control subjects in group 4 (table 2). There was no statistically significant difference between patients with NFA (group 1) and moderate asthma (group 2) in T/D or WA%. Both groups 1 and 2 had higher T/D and WA% than either group 3 or 4 (p<0.001). There were small but significant differences between the patients with mild asthma (group 3) and normal controls (group 4) in T/D and WA% (p<0.03). The differences (95% CI) between the groups in WA% were 7.1% (95% CI 0 to 14.4) for groups 1 and 4, 3.8% (95% CI –3.4 to 10) for groups 1 and 3, and 3.3% (95% CI –4.4 to 10) for groups 3 and 4. We repeated the statistical analysis after getting the mean T/D ratio and the mean WA% for each subject. Groups 1 and 2 still had significantly higher T/D and WA% than either group 3 (p<0.02) or group 4 (p<0.002). However, the difference between groups 3 and 4 did not reach statistical significance (p = 0.1). We then divided the bronchi into large airways with L >2 mm and small airways with L ⩽2 mm. This subgroup analysis revealed that the differences in T/D and WA% between the groups was similar for both large and small airways (table 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Representative HRCT images of (A) a patient with a history of near fatal asthma, (B) a patient with mild asthma, and (C) a control subject.

Discussion

We chose a modified HRCT technique to assess airway dimensions in patients with asthma. The T/D ratio and WA% were higher in asthma patients with different degrees of severity compared with normal controls. There was no difference in thickening of the airway wall between patients with a history of NFA and patients with moderate asthma who had a similar degree of airway obstruction; however, both these groups had more airway wall thickening than patients with mild asthma. This airway wall thickening was observed in both large and small airways.

Increased airway wall thickness has been recognised as a feature of asthma for almost a century.19 Studies of histological specimens obtained from necroscopic studies3 ,4 and bronchial biopsy specimens20 showed that this airway wall thickening is secondary to chronic inflammatory changes. The importance of airway wall thickening in the mechanism of airway narrowing has been discussed.5 In summary, the same degree of bronchial smooth muscle contraction in an airway with a thickened wall will result in a greater degree of narrowing than an airway of normal thickness. The measurement of airway dimensions using HRCT scanning has been performed in experimental models10 and human subjects.18 These studies were done using an HRCT protocol similar to that used here. However, in the previous studies the HRCT images were digitised and the internal and external perimeters were outlined in order to measure bronchial cross sectional areas. Using this technique Okazawa and colleagues18 showed that smaller airways (1.5–6 mm luminal diameter) of asthma patients were significantly thickened compared with normal controls. Boulet and colleagues12 found no difference in the T/D ratio of the bronchus intermedius between asthma patients and normal controls but smaller bronchi were not assessed. We measured airway dimensions of segmental and subsegmental bronchi directly from magnified HRCT images and calculated the cross sectional areas of the airways. We used both the T/D ratio as used by Boulet and colleagues12 and WA% as used by Okazawa and colleagues18 to assess thickening of the airway wall. Our primary measurement was the T/D ratio as T and D were directly measured from the HRCT images. However, we also assessed the calculated WA and WA% to compare our results with previous studies.18 Okazawa and colleagues18found that the mean WA% in asthma patients was 84 (3)% for small airways with luminal diameters of <2 mm and 58 (5)% for large airways with a luminal diameter of >6 mm. Our results showed that the mean WA% for all airways assessed was 78 (9.2)% for patients with NFA, 78.8 (9.2)% for those with moderate asthma, 74.2 (7.5)% for patients with mild asthma, and 70.9 (8.2)% for normal controls. Others have suggested a more complex method of measurement21 ,22but the gradient in airway thickness from normal subjects to subjects with severe disease suggests that our method is valid. Our method is simpler and less time consuming than that of Brown and colleagues21 ,22 as the measurements are taken directly from the HRCT images and the bronchial cross sectional areas are calculated.

Studies in phantoms have shown that HRCT scanning allows accurate assessment of hollow cylinders ranging from 1 mm to 5 mm in diameter with a wall thickness ranging from 0.5 mm to 2 mm.23 ,24However, in the current study, as in the previous ones,10 ,18 CT scanning overestimated bronchial wall thickness and underestimated lumen diameter of the airways. The overestimation of bronchial wall thickness on HRCT scans presumably results from the relatively low spatial resolution leading to inaccuracy in boundary detection with volume averaging of the surrounding interstitium and the mucosal folds. However, it should also be realised that studies such as those of James et al 5 which have measured the airway wall thickness in vitro may have underestimated bronchial wall thickness because of the use of fixatives and the lack of airway wall perfusion.

Although others have claimed to measure the entire bronchial wall area, the adjacent pulmonary artery often obscures part of the wall of the bronchus. In these cases the thickness of the airway wall has been assumed to be constant in that area.18 A potential limitation of our study is that we only measured the bronchial wall considered by two observers to be representative of the average bronchial wall thickness. Furthermore, our results using a single measurement of bronchial wall thickness are similar to the results of Okazawa and colleagues who made measurements of the entire bronchial wall thickness. It should be noted, however, that, regardless of the methods used, bronchial wall thickness is probably overestimated on CT scans because of inaccuracies in boundary detection and inclusion of adjacent peribronchial interstitium.25

We chose to study patients with a history of NFA as such episodes are well recognised as being a marker for increased risk of subsequent death from asthma and this group of patients represents a population with the most severe form of asthma.6 ,7 ,26 We compared patients with NFA with a group of asthma patients who had a similar degree of airflow obstruction and asthma severity (moderate asthma) and a group of patients with mild asthma and normal baseline lung function. We found no difference in airway wall thickening which is a marker of chronic airway inflammation between patients with a history of NFA and those with moderate asthma. This is not surprising as these two groups were matched for their level of airway obstruction and the patients with NFA were studied when they were stable and taking regular anti-asthma medication. Previous studies have shown that asthma patients at risk of death from asthma and patients admitted to hospital with acute asthma are similar with respect to recognised markers of asthma severity and an adequate match for asthma severity can be obtained in case control studies of asthma deaths.27 It is interesting to note that even patients with mild asthma had thickening of the airway wall compared with normal controls. The patients with mild disease had a long history of asthma but they were never admitted to hospital and had normal baseline lung function tests at the time of the study. This supports the hypothesis that chronic inflammatory changes lead to airway wall thickening even in patients with mild disease. However, it has been shown in previous studies that the early use of inhaled corticosteroids may prevent the development of chronic airway obstruction28 ,29 and this has been emphasised in national guidelines.30 However, it must be emphasised that the difference between patients with mild asthma and normal subjects is small and the clinical significance of this difference is not known. In addition, the statistical significance between mild asthmatic subjects and normal controls was lost when the analysis was performed based on individual subjects after recalculating a mean from each subject rather than taking the mean of all subjects.

The predominant site of airway narrowing in asthma is not clear. However, thickening of the airway wall in patients with fatal asthma was observed in both large cartilaginous and small membranous bronchi.5 ,31 This thickening occurred predominantly in the small airways of patients with non-fatal asthma.31There was no difference in the thickness of the airway wall at the level of the bronchus intermedius between patients with asthma and normal controls.12 In our study we assessed the bronchi at both segmental and subsegmental levels. Unlike Boulet and colleagues we did not measure airway wall thickness of the lobar or main bronchi. We found thickening of the airway wall in both large airways (L >2 mm) and small airways (L ⩽2 mm) of asthma patients compared with normal controls. Our findings are similar to those of Okazawa and colleagues18 who found that airways smaller than 6 mm in luminal diameter of asthma patients were significantly thickened compared with normal controls.

We believe the methodology described in this study would be useful in early intervention studies to see whether more aggressive and earlier anti-inflammatory therapy could prevent some of the changes in the airway wall described here. Although the technique is time consuming and relatively expensive, it offers the opportunity for repeated measurements. The exposure to radiation is a possible concern but the use of the methodology described here would give radiation exposure equivalent to six chest radiographs per assessment.

In conclusion, we have shown with the use of HRCT scanning that patients with asthma of different degrees of severity have more airway wall thickening than normal controls. Furthermore, patients with more severe asthma have more airway wall thickening than those with mild asthma.