| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Vermont Lung Center, University of Vermont College of Medicine; and 2Department of Radiology, University of Vermont, Fletcher Allen Health Care, Burlington, Vermont
Submitted 14 May 2007 ; accepted in final form 4 October 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
respiratory system impedance; frequency dependence; lung attenuation; airways hyperresponsiveness; supine; deep inflation
20 Hz, accompanied by variable effects on the frequency dependence of respiratory system elastance (Ers) (22, 30, 46). While heterogeneity is clearly a feature of bronchoconstriction in humans, the extent to which heterogeneity distinguishes healthy subjects from those with asthma, independent of the degree of overall airway constriction, remains unclear. The study by King and colleagues (24) addressed this question by comparing the heterogeneity of airway narrowing between healthy and asthmatic subjects and found that heterogeneity was greater in asthmatic subjects for airways > 2 mm in diameter but not for smaller airways. Yet, imaging studies using PET (45, 47), and modeling of lung impedance (6, 21, 30, 46), implicate narrowing of small, peripheral airways as a key feature of asthma. These airways cannot be visualized directly by CT. However, the consequences of such peripheral airway narrowing include air trapping and hyperinflation, together with the reduced blood flow that occurs as a result of local hypoxic vasoconstriction (19). These factors all cause decreases in the CT attenuation of the lung parenchyma. The relative degree and heterogeneity of CT attenuation changes would therefore reflect the degree and heterogeneity of peripheral airway narrowing.
With this rationale in mind, we compared the heterogeneity of peripheral airway narrowing in healthy subjects and those with mild asthma at similar degrees of bronchoconstriction by measuring changes in CT attenuation of the lung parenchyma and changes in Rrs and Ers using the FOT. Because our study design necessarily required subjects to be supine, our findings relate specifically to a comparison of structural and functional heterogeneity between healthy and asthmatic subjects in the supine position.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Subjects. We recruited adult subjects with asthma and nonasthmatic, healthy control adult subjects. On the basis of the expected variability of Rrs (22, 23), and Hounsfield unit (HU) parenchymal density (18), we estimated that a sample size of 12 subjects per group would be needed to detect changes of 50% in Rrs and 10% in HU density with 80% power. We defined asthmatic subjects based on National Institutes of Health guidelines (35) with the following characteristics: no history of cardiopulmonary disease other than asthma, no smoking within the last 3 years and less than a total of 5 pack-years, and positive methacholine challenge defined as a provocative concentration of methacholine causing a 20% fall in the forced expiratory volume in 1 s (FEV1) (PC20) < 8 mg/ml. Nonasthmatic control subjects met the same criteria except they had no history of asthma and had a PC20 > 16 mg/ml. Subjects were excluded from testing if they had had an upper respiratory infection within the last 4 wk. Subjects with asthma had to withhold any short-acting bronchodilators for 8 h and any long-acting bronchodilators for 24 h before any testing. Female participants had to prove they were not pregnant by urine pregnancy testing. All subjects provided written, informed consent, and the Institutional Review Board of the University of Vermont approved the study.
All subjects underwent a first visit during which we obtained baseline history and performed a physical examination. We also measured spirometry and performed a methacholine challenge using the five deep breath method (3), all while subjects were seated. Qualified participants then returned for a second visit during which we performed concurrent measurement of respiratory system impedance (Zrs) and lung imaging by CT as outlined in Fig. 1.
|
|
Subjects were instructed to suspend their breathing at end-tidal exhalation before commencement of measurement and to remain relaxed while the V perturbations were delivered. In addition to nose clips, we had all subjects wear a custom-made, firmly fitted mask to support their cheeks to minimize upper airway shunting (13). We performed a number of test runs in each subject to allow them to practice relaxation during application of the perturbations in V. Inadequate relaxation manifested as a noticeable increasing or decreasing trend in P over the 16-s measurement period, and we aborted such trials. At least two satisfactory measurements of Zrs were made for each subject under each experimental condition.
CT imaging.
CT scans of the chest were performed during suspended breathing at end-tidal exhalation to correspond to the conditions under which Zrs was measured. We confirmed stable end-tidal lung volume during the scans by monitoring subject breathing through a spirometer (25). Subjects were able to practice the breath-hold maneuver until they were comfortable performing it. Only when a proper breath-hold was observed did we perform the CT scan using a 16-slice helical scanner (Philips MX8000). Scans were centered 4 cm below the main carina and were focused on the right lower lobe. We chose this region for a number of reasons. First, the right side was selected to minimize motion artifact from cardiac activity. Second, the lower lobe was selected to focus on a dependent region that would optimize detection of changes in lung attenuation from air trapping or altered blood flow. Third, we only looked at the lower lobe, as opposed to the whole lung, to minimize any changes in lung attenuation from gravitational effects, which can be substantial in subjects with asthma (8). Fourth, because we purposely designed a very low dose radiation protocol to limit exposure, we reconstructed the images to a small field of view to maximize resolution and optimize visualization of density variations in the parenchyma. We obtained 16 slices of 1.5-mm thickness each, with a gantry rotation time of 0.75 s, for a total scan width of 24 mm. The x-ray parameters used were 120 kVp and 40 mA. By design, this acquisition technique was not sufficient for reliable measurement of airway dimensions. Instead, we inferred the extent and variability of peripheral airway narrowing on the basis of attenuation measurements of the lung parenchyma (see below). One asthmatic participant was unable to tolerate the FOT when supine, so CT scanning was not performed on this subject.
Protocol. The objective of the protocol (Fig. 1) was to obtain concurrent measurements of Zrs and CT lung attenuation before and after inhalation of methacholine, as well as to determine the effect of deep inhalation (DI) and inhaled albuterol on Zrs. Since we used CT data as one measure of heterogeneity, all subjects were supine for administration of methacholine and the use of the FOT. First, subjects performed spirometry while seated, to confirm stability of lung function compared with baseline testing. If lung function was within 10% of baseline measurements, subjects then underwent the FOT while sitting. This usually involved a number of trials to allow the subjects to become comfortable and successful with the technique. Subjects then lay down on the CT gantry table and were carefully positioned with their arms over their heads, where they remained for all subsequent FOT measurements and CT imaging. We next obtained baseline supine measurements of Zrs, and then positioned the gantry inside the CT scanner, without the subjects changing body position, to obtain baseline CT imaging. The gantry was then moved back out of the scanner, and spirometry was performed to assess baseline lung function in the supine position.
We then had each subject inhale five deep breaths of nebulized methacholine while they remained supine. We used the individual PC20 concentration of methacholine, as determined during the screening day, for each of the asthmatic subjects, and 16 mg/ml for each of the control subjects. Because we were interested in comparing the responses to methacholine in asthmatic and control subjects, we needed to select a dose of methacholine to which the control subjects would respond. Knowing that healthy subjects have increased responsiveness when supine (42), we chose the highest concentration of methacholine to which the control subjects had not responded when upright (16 mg/ml). Preliminary work showed that, while supine, these control subjects indeed reacted to 16 mg/ml of methacholine with a similar mean change in FEV1 (see RESULTS) as the asthmatics had when they inhaled their upright PC20 dose of methacholine while supine. This serendipitous observation led to our choosing the 16 mg/ml dose for all control subjects, thus allowing us to compare the CT and FOT responses when both groups were bronchoconstricted to similar degrees.
Two minutes after the methacholine administration, we repeated the FOT measurements, immediately followed by a second CT scan of the same anatomic zone as the baseline scans. This second CT scan took place 3–4 min after completion of the methacholine inhalation. We next repeated the FOT after the subjects took three consecutive DIs to total lung capacity (TLC) and then had the subjects perform spirometry again. These spirometric measurements took place 5 min after completion of the methacholine challenge. We then administered two inhalations of albuterol (180 mcg) via a metered dose inhaler through a spacer, again while subjects remained supine, and performed the FOT 5 min later. Finally, subjects sat upright, and a final set of spirometry measurements were made to determine that lung function had returned to within 10% of baseline before discharge from the CT suite.
Data Analysis
Spirometry. Spirometry values were measured in absolute terms and converted to a percentage of the predicted values based on data from Hankinson and colleagues (20).
Zrs by FOT.
Calculation of Zrs from the FOT measurements of P and V was complicated by the fact that many subjects experienced difficulty remaining relaxed with an open glottis throughout the entire 16-s oscillation period. Furthermore, subjects differed in where they were most relaxed during the measurement period; some subjects settled down quickly and produced their best data early on, while others took a few seconds to relax properly and so produced their best Zrs measurements toward the end of the 16-s measurement period. We therefore searched each 16-s data set for the 8 contiguous seconds during which the subjects were most relaxed. This was achieved by first smoothing both P and V signals with a 1-s running mean and reducing the sampling rate from 256 to 128 Hz. Next, Zrs was calculated within 8-s windows spaced 1 s apart across the 16-s data set, and the Zrs with the highest mean coherence across all frequencies was retained. Within each 8-s window, we calculated the ensemble averages of P(f)V*(f) and V(f)V*(f) using 2-s windows overlapping by 50%, where P(f) and V(f) are the fast Fourier transforms of P and V, respectively, * denotes complex conjugation, and f is frequency (in Hz). Zrs was then calculated as –iP(f)V*(f)/2
V(f)V*(f), where i is the positive square root of –1. We initially sought to fit Zrs to the constant-phase model proposed by Hantos and colleagues (21). However, this model did not fit the imaginary part of Zrs well. Therefore, instead, we empirically fit a hyperbolic curve only to the real part of Zrs, as
 | (1) |
are constants and f is frequency. Curve fitting was achieved by minimizing the squared residuals weighting each data point inversely to its coherence. Thus we used all the data, but allowed the more reliable points to have a stronger role in generating the fitted curve (Eq. 1). We then extracted from the fitted curve the values of Rrs at 1 Hz and 20 Hz, denoted R1 and R20, respectively, and quantified the frequency dependence of Rrs as the difference R1 – R20 (10).
We calculated elastance (Ers) as a function of f as –2fXrs(f), where Xrs is the imaginary part of impedance. Because the Ers data were more highly variable than the Rrs data, consistent curve fits were not reliable. Instead, we plotted Ers vs. frequency and categorized the pattern according to the type A and type B designations given by Kacska and colleagues (22). Specifically, if the baseline, sitting data became more negative with frequency, then this pattern was defined as type A, and if the baseline, sitting data became more positive with frequency, then this pattern was defined as type B. The FOT data from one asthmatic participant were unacceptable because of low coherence and were excluded from analysis. Another asthmatic participant was unable to tolerate the FOT while supine. The FOT data from one control subject at baseline and another following methacholine were also unacceptable due to low coherence and were excluded from analysis.
Air-space dimensions based on CT imaging.
Once we isolated the region of interest in the right lower lobe, we registered it to the same lung region pre- and post-methacholine in each subject by visually matching anatomic landmarks such as blood vessels and larger airways (Fig. 2A). We then calculated a histogram of Hounsfield units (HU) less than –300 within each region of interest (Fig. 2B, left), as this corresponds to the parenchyma (17, 36). However, because HU cannot be less than –1,000, the HU histogram can give a distorted view of changes in regional heterogeneity of lung aeration. For example, when more pixels occupy similar, low-attenuation values, then the mean HU density will fall and the variance (SD) will also fall, suggesting more homogeneity of lung density. However, this may not reflect the variance in underlying air-space dimensions, which we felt was a more physically meaningful notion. We assumed that the variation in air-space dimensions would reflect (inversely) the corresponding dimensions of the peripheral airways leading to that region. In other words, as peripheral airways narrow to different degrees, then more air may be trapped distal to them depending on the degree of narrowing. This air trapping would result in larger air spaces that would vary inversely in size according to the degree of peripheral airway narrowing. However, the pixel density of these regions may be quite similar since there is a finite limit of density distribution. Therefore, to relate changes in HU pre- and post-methacholine to changes in regional aeration of the lung in a more effective, physical manner, we constructed a simple, computational model of the lung parenchyma. We considered the lung parenchyma as a foam of identical space-filling cubic alveoli having internal side length a. The air-space volume within each alveolus, which varies with regional aeration, is thus a3. The tissue volume per alveolus, Vtis, is assumed to remain constant in the setting of air trapping. Assuming the CT image pixel dimensions are substantially greater than those of each alveolus, we can relate HU to the average air/tissue fraction within a pixel. With HU = 0 for tissue and HU = –1,000 for air, the pixel air fraction is
 | (2) |
 | (3) |
 | (4) |
|
which can be calculated from HU using Eq. 4, is thus an estimate of the mean linear air-space dimension (LAD) within each pixel. Using Eq. 4, we therefore converted the HU histograms to histograms of 
(Fig. 2B, right). The mean and SD of the histograms were taken as global measures of the size and distribution of LADs across the examined lung. We thus defined air-space heterogeneity as the SD of LAD (LAD-SD). Statistics. All demographic, spirometry, and CT attenuation results are reported as means ± SD, and PC20 is reported as geometric mean and range. All FOT data and LAD data are reported as median (25–75 interquartile range). Statistical comparisons were made using Student's t-test or Wilcoxon rank test, as appropriate, and associations were measured with Spearman rank correlation. We did not formally analyze the Ers by statistical methods because of extreme variability in response of this parameter. Instead, we made only descriptive comparisons. To quantify the effects of methacholine, we compared supine lung function after methacholine to supine baseline measurements. Likewise, to determine the effects of DI and albuterol, we compared lung function after DI and albuterol to post-methacholine measurements. Finally, we measured post-albuterol spirometry when upright again to assess any change from baseline upright values and thus ensure sufficient return of lung function to normal before discharge from the protocol. All P values are two-sided, and values < 0.05 were considered statistically significant. All statistical analyses were performed using a statistical software package (JMP; SAS Institute, Cary, NC).
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
On lying down, both subject groups experienced significant falls in FEV1, but these changes were not significantly different between the two groups (Table 2).
|
At baseline, sitting, R1, R20, and R1 – R20 were similar between asthmatic and control subjects (Table 1), and none of these parameters changed significantly on lying down (Table 2). Following methacholine inhalation, both groups had significant increases in R1, but only the control subjects had a significant increase in R20 (Table 2). For both groups combined, the change in R1 after methacholine correlated with the change in FEV1 after methacholine (Spearman rho = –0.38, P = 0.03). Both groups significantly increased R1 – R20 after methacholine, indicating a substantial increase in frequency dependence, and this also correlated with the change in FEV1 after methacholine (Spearman rho = –0.38, P = 0.03) for both groups combined. The changes in R1 and R20 were not significant after DI in either group, but the changes in both parameters, as well as R1 – R20, were significant after albuterol in both groups. There were no significant differences in any of the responses comparing asthmatic and nonasthmatic subjects. These results are visually apparent in the plots of Rrs vs. frequency shown in Fig. 3.
|
FEV1 after methacholine, PC20, or any of the CT or LAD parameters (data not shown).
|
|
Baseline CT attenuation (Table 1) and changes in attenuation (Table 3) were very similar between groups. Methacholine inhalation caused the HU histograms of both control and asthmatic subjects to shift to the left, (Fig. 2B, left), with a decrease in mean HU (HU), indicative of increased air trapping or regional hypoperfusion due to local airway narrowing or closure. The HU histograms also became narrower, as seen by a decrease in the SD of HU (HU-SD) in both groups. The values of mean LAD (LAD-mean) in asthmatic and control subjects were comparable at baseline, and increased significantly after methacholine, indicative of an increase in mean air-space dimension (Fig. 2B, right). However, unlike the distribution of HU, the SD of LAD (LAD-SD), which was comparable between the groups at baseline, increased significantly after methacholine, indicative of an increased variability in air-space dimensions. The drop in FEV1, but not the change in R1, R20, or R1 – R20, in response to methacholine correlated significantly with the change in HU, LAD, and LAD-SD in asthmatic but not control subjects (Spearman rho = –0.68, P = 0.003; rho = –0.70, P = 0.002; rho = –0.64, P = 0.006, respectively). There were no significant correlations between baseline values or changes in LAD-SD and R1 – R20.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
20%). There was a weak correlation among all subjects between the overall degree of response as measured by FEV1 and the change in heterogeneity as measured by R1 – R20. Likewise, there was a correlation between the change in FEV1 and the change in heterogeneity as measured by LAD-SD, but this occurred in the asthmatic subjects only. However, for all subjects, there was no correlation between the two different measures of heterogeneity themselves, R1 – R20 and the LAD-SD. In addition, both asthmatic and control subjects displayed both A and B types of Ers patterns. Overall, while the asthmatic subjects clearly responded to a lower dose of methacholine compared with the control subjects, measures of heterogeneity did not distinguish between the two groups. These results demonstrate that CT imaging and the FOT reveal independent measures of heterogeneity, which are associated with airway narrowing but are not unique to mild asthma. A key component of our analysis is the use of CT to infer information about heterogeneity from the spatial distribution of HU values within the lung parenchyma. A number of other CT studies have measured changes in parenchymal x-ray densities following bronchoconstriction or anti-inflammatory therapy (9, 18, 19, 33, 36, 48), but none related these changes to heterogeneity of air-space size and corresponding peripheral airway narrowing. CT can also be used to study changes in the large airways (24, 27, 37). Indeed, heterogeneous reductions in the radii of large airways following bronchial challenge have been directly visualized in humans (11, 12, 24). Furthermore, King and colleagues (24) found greater heterogeneity among airways > 2 mm in diameter in asthmatic compared with normal control subjects. Also, Brown and colleagues (11) found that large airway dimensions were closely correlated with FEV1/FVC and measures of airways hyperresponsiveness. In the present study, however, we focused on the changes in parenchymal lung density because the pattern and extent of peripheral airway narrowing are key components in determining overall lung mechanics (28, 29, 46). In addition, CT information about the lung periphery in asthma has not yet been compared with assessments of lung function beyond spirometry. Assessment of parenchymal lung density can also be made with a smaller x-ray dose than that required for measurement of airway dimensions.
The analysis of the CT data was based on the hypothetical concept of the parenchyma consisting of a collection of equivalent air spaces each having a particular linear dimension a. This was necessary here because the width of the HU distribution is highly dependent on its mean; as the HU-mean decreases, the leftward spread of the histogram becomes progressively more confined by the hard lower limit of –1,000. This explains why HU-SD decreased following methacholine (Table 3). To avoid this effect, we transformed the HUs to a quantity related to a linear air-space dimension, a, which was now a function whose variability should no longer be directly related to its mean. Our results showed that the variability of air-space size, as measured by the SD of LAD, increased following methacholine (Table 3). We interpret this to mean that the airway narrowing that occurred proximal to these air spaces was, likewise, more heterogeneous following methacholine. Nevertheless, it is important to realize that a is a purely hypothetical quantity that merely serves to relate HUs to an equivalent effective air-space size, so it cannot be related to any particular parenchymal structure.
We used the FOT to assess the increase in heterogeneity in terms of the negative frequency dependence of Rrs below 20 Hz, which is thought to arise primarily from heterogeneous airway narrowing (28). While tissue resistance likely accounts for most of the frequency dependence of resistance and elastance at baseline, several studies have shown that airway inhomogeneities account for the majority of the frequency-dependent changes in these parameters during bronchoconstriction (6, 21, 30, 46). The changes we saw in Rrs occurred predominately at low frequencies, indicating their location in the lung periphery (Fig. 3). Interestingly, only the changes in R1, not R20, correlated with the changes in FEV1 after methacholine, suggesting that the common measure FEV1 reflects peripheral more than central airway resistance. Our findings of peripheral lung involvement are similar to those of Tgavalekos and colleagues (45), who obtained PET images and measured Zrs following challenge with inhaled methacholine in asthmatic subjects. Using an anatomically based computational model of the lung, they found that both data sets could be explained in a consistent manner only if the model included constriction of airways < 2.4 mm in diameter. Using a different computational approach, Venegas and colleagues (47) showed that the pattern of poorly ventilated areas seen in PET images can be attributed to self-organized clustering of highly constricted small airways. The results of the present study thus suggest that CT imaging also provides a view of regional heterogeneity in bronchoconstriction occurring at the small airway level.
To determine whether other aspects of Zrs could distinguish asthmatic subjects from controls, we separated the baseline, sitting Ers data into two groups (Fig. 4) as suggested by Kaczka and colleagues (22). The group categorization was made on the basis of subjective judgment and is therefore somewhat arbitrary, but we can see an overall difference between the type A subjects, with their tendency for baseline, sitting Ers to decrease at higher f (Fig. 4, left), and the type B subjects, in whom baseline, sitting Ers remains elevated (Fig. 4, right). We found similar patterns of Ers in both healthy and asthmatic subjects when Rrs and Ers of the lung periphery were studied by applying the FOT through a wedged bronchoscope (23). In the present study, type B subjects, both control and asthmatic, were generally characterized by higher baseline values of R1, R20, and R1 – R20, likely reflecting overall increased airway resistance (Fig. 5). This finding is in accord with the explanation for the type B pattern offered by Kaczka and colleagues (22). However, we report two additional features. First, the type A asthmatic subjects seemed to be more reactive than the type A controls or any of the type B subjects, whether asthmatic or control. Perhaps the lower baseline peripheral resistance in type A asthmatic subjects allowed a broader and deeper deposition of methacholine during inhalation. Second, the type B subjects had more highly variable responses to sitting, methacholine, DI, and albuterol than the type A subjects. We believe this behavior may reflect the more complex interaction between peripheral airway resistance, tissue viscoelasticity, and central airway shunting that computational modeling suggests are involved in the type B subjects (22). Our finding of these patterns in healthy subjects as well as in asthmatic subjects suggests that these fundamental elements of lung mechanics are operative in all humans.
Despite the similarities in response between the control and asthmatic subjects in our study, the asthmatic subjects only had mild disease, so we cannot say that enhanced heterogeneity is not a hallmark feature of subjects with more severe disease or in a state of exacerbation. In fact, Lutchen and colleagues (28) have shown the severity of heterogeneity parallels the baseline severity of asthma. Nevertheless, enhanced heterogeneity of bronchoconstriction does not appear to be a characteristic feature of mild asthma compared with control subjects when supine.
An interesting consequence of our study design, necessitated by the practicalities of obtaining CT images, is that we focused on the nature of bronchoconstriction in the supine position. Most studies of human airway responsiveness have been performed vertically, yet responsiveness in the horizontal position is of prime concern in nocturnal asthma (5, 44). A dramatic rise in responsiveness to methacholine has been demonstrated in normal subjects when they become supine (14, 31, 40, 42), and we also found that the nonasthmatic subjects in the present study behaved like subjects with asthma when they lay down and were exposed to methacholine (Table 2). What we did not observe was that there was any change in heterogeneity, at least as measured by R1 – R20, in moving from the upright to the supine position (Table 2). This would imply that heterogeneity of airway narrowing is more dependent on the state of activation of airway smooth muscle and less dependent on the effects of posture; i.e., lung volume. However, the effects of supine posture on reduced lung volume and airway-parenchymal interdependence (1, 7, 14, 31) are strongly suggested by the lack of a bronchodilatory effect of DI in both the control and asthmatic subjects (Table 2). These results are reminiscent of those of Skloot and colleagues (43), who found that preventing healthy subjects from taking a DI for a period of time reduced the bronchodilatory effectiveness of a subsequent DI. Lutchen and colleagues (28) observed a similar phenomenon in subjects with severe asthma, although not in mildly asthmatic or normal control subjects. Thus, in lying supine for a period of time in the CT scanner, our subjects may have experienced both reduced forces of airway-parenchymal interdependence and a period of limited periodic stretch of their airway smooth muscle that caused their airway smooth muscle to become "frozen" in a constricted state (16, 28). This could explain why we were only able to relieve their bronchoconstriction pharmacologically (Table 2). Taken together, these results emphasize how both the state of activation of airway smooth muscle, as well as its load, are critically important in determining airways responsiveness (40).
The results of our study must be considered in the context of a number of practical limitations we faced. For example, our measurements of Rrs included a contribution from the chest wall, which may have changed relative to that of the lung with changes in posture. However, Nagels and colleagues (34) showed that the contribution of the chest wall to total pulmonary resistance in the supine position is small. Our findings may also be limited because we restricted out analysis of CT data to one region of the lung only, and heterogeneity of lung attenuation may have occurred to different extents in different lung regions. In addition, because we designed the CT protocol to detect changes in lung attenuation only and not to detect changes in airway dimensions, we do not know whether heterogeneity as determined by the FOT might have correlated with heterogeneity of larger airway size. Another limitation is that although all CT and FOT measurements were made at end expiration, we did not measure absolute lung volume. Consequently, we cannot be sure that air trapping secondary to bronchoconstriction did not elevate lung volume and so contribute to the increased air-space size that we detected following methacholine administration. We did not track lung volume directly, but we did track FVC before and after methacholine. Assuming TLC did not change, a fall in FVC would reflect the extent of elevation of residual volume and hence air trapping. We found that some degree of air trapping did occur in both asthmatic and control subjects, because FVC fell 5% and 10%, respectively (P = 0.18). Nevertheless, our analysis indicates that air-space heterogeneity also increased in both groups, which is not what would be expected from an increase in lung volume (4). A further limitation imposed by our experimental design was that the DIs performed following the supine methacholine challenge may have influenced the subsequent measurements. We attempted to minimize this effect by waiting 2 min (26) after methacholine administration before acquiring FOT and CT data. DIs may have also affected the methacholine response as measured by FEV1 (2), but since performing the FEV1 itself involves a DI, we thought the effect of any preceding DIs would be small.
In conclusion, we have demonstrated a novel method of assessing parenchymal heterogeneity by CT imaging and used the technique to show that heterogeneity of bronchoconstriction as measured by CT and the FOT increases to similar degrees in healthy and mildly asthmatic subjects when they are constricted to equivalent levels in the supine position. Heterogeneity of bronchoconstriction thus does not appear to be a function of mild asthma per se but rather is associated with bronchoconstriction in general. Indeed, under the conditions of this study, the behavior of healthy and asthmatic subjects was strikingly similar. Yet, since the asthmatic subjects were still distinguished by their hyperresponsiveness to methacholine, important differences in airway and parenchymal function must be operative in determining the unique physiological phenotype of asthma.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. W. Kaczka, R. H. Brown, and W. Mitzner Assessment of heterogeneous airway constriction in dogs: a structure-function analysis J Appl Physiol, February 1, 2009; 106(2): 520 - 530. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Lahzami and G. G. King Targeting small airways in asthma: the new challenge of inhaled corticosteroid treatment Eur. Respir. J., June 1, 2008; 31(6): 1145 - 1147. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
).