Airway closure on imaging relates to airway hyperresponsiveness and ...

J Appl Physiol 113: 958 –966, 2012. First published July 26, 2012; doi:10.1152/japplphysiol.01618.2011.

HIGHLIGHTED TOPIC

Imaging Lung Physiology

Airway closure on imaging relates to airway hyperresponsiveness and peripheral airway disease in asthma Catherine E. Farrow,1,2 Cheryl M. Salome,1,2,3 Benjamin E. Harris,1,2,4 Dale L. Bailey,2,5 Elizabeth Bailey,5 Norbert Berend,1,2,3 Iven H. Young,6 and Gregory G. King1,2,3,4 1

Woolcock Institute of Medical Research, Glebe, New South Wales, Australia; 2The University of Sydney, Sydney, New South Wales, Australia; 3Cooperative Research Centre for Asthma and Airways, Glebe, New South Wales, Australia; 4Department of Respiratory Medicine, Royal North Shore Hospital, St. Leonards, New South Wales, Australia; 5Department of Nuclear Medicine, Royal North Shore Hospital, St. Leonards, New South Wales, Australia; and 6Department of Respiratory Medicine, Royal Prince Alfred Hospital, Camperdown, New South Wales, Australia Submitted 28 December 2011; accepted in final form 19 July 2012

Farrow CE, Salome CM, Harris BE, Bailey DL, Bailey E, Berend N, Young IH, King GG. Airway closure on imaging relates to airway hyperresponsiveness and peripheral airway disease in asthma. J Appl Physiol 113: 958 –966, 2012. First published July 26, 2012; doi:10.1152/japplphysiol.01618.2011.—The regional pattern and extent of airway closure measured by three-dimensional ventilation imaging may relate to airway hyperresponsiveness (AHR) and peripheral airways disease in asthmatic subjects. We hypothesized that asthmatic airways are predisposed to closure during bronchoconstriction in the presence of ventilation heterogeneity and AHR. Fourteen asthmatic subjects (6 women) underwent combined ventilation single photon emission computed tomography/computed tomography scans before and after methacholine challenge. Regional airway closure was determined by complete loss of ventilation following methacholine challenge. Peripheral airway disease was measured by multiple-breath nitrogen washout from which Scond (index of peripheral conductive airway abnormality) was derived. Relationships between airway closure and lung function were examined by multiple-linear regression. Forced expiratory volume in 1 s was 87.5 ⫾ 15.8% predicted, and seven subjects had AHR. Methacholine challenge decreased forced expiratory volume in 1 s by 23 ⫾ 5% and increased nonventilated volume from 16 ⫾ 4 to 29 ⫾ 13% of computed tomography lung volume. The increase in airway closure measured by nonventilated volume correlated independently with both Scond (partial R2 ⫽ 0.22) and with AHR (partial R2 ⫽ 0.38). The extent of airway closure induced by methacholine inhalation in asthmatic subjects is greater with increasing peripheral airways disease, as measured by ventilation heterogeneity, and with worse AHR. asthma; airway closure; airway hyperresponsiveness; single photon emission computed tomography; ventilation heterogeneity ASTHMA IS CHARACTERIZED BY intermittent airway narrowing and airway closure, which cause typical symptoms of wheeze, breathlessness, and chest tightness. Airway closure occurs during bronchial challenge [measured by a decrease in forced vital capacity (FVC) (5, 9, 34) and an increase in closing capacity (CC) (43)] and is associated with more severe asthma in terms of greater risk of exacerbations (9, 14, 34). Airway hyperresponsiveness (AHR) in asthma is also associated with more severe and poorly controlled disease, greater risk of

Address for reprint requests and other correspondence: C. E. Farrow, Woolcock Institute of Medical Research, P. O. Box M77, Missenden Rd Post Office, Camperdown NSW 2050, Australia (e-mail: [email protected]). 958

hospitalization (32), and lung function decline (35). AHR is characterized by loss of the maximal response plateau, which implies that asthmatic airways may narrow to complete closure (26). Despite important associations between excessive airway closure and poor clinical outcomes, the mechanisms behind closure are poorly understood, and the relationship between airway closure and AHR is unclear. Airway closure, measured by decrease in FVC during bronchoconstriction, is associated with AHR independently of the extent of airway narrowing (5). However, spirometry is a global measurement and gives no information on the regional effects of closure. Indeed, a previous imaging study showed that, at baseline, the difference between asthmatic subjects and healthy controls is not in the volume of absent ventilation, but in its patchy distribution and the loss of its relationship with age (22). Furthermore, there are no published imaging studies on the association between either the magnitude or distribution of ventilation loss during bronchoconstriction and AHR. Airway closure causes ventilation defects that are measurable in three-dimensional ventilation images, such as single photon emission computed tomography (SPECT), positron emission tomography, high-resolution computed tomography (CT), and magnetic resonance imaging (MRI) with hyperpolarized He-3 (13, 22, 33). The ventilation defects that are apparent on ventilation imaging, and could be caused by airway closure and/or severe regional narrowing, are clearly identifiable regions or clusters at baseline, with new regions developing with bronchoconstriction in asthmatic subjects. Recent results from computational modeling of asthmatic lungs suggest that the basis of regional airway closure is widespread disease of small airways (38). Ventilation distribution in asthmatic subjects is characteristically more uneven and heterogeneous (4), which, in turn, causes airways to be unstable during bronchoconstriction. This results in patchy narrowing (20) and widespread airway closure (37). However, these theories have not been confirmed by experimental data to determine if there is any association between markers of peripheral airway dysfunction and the development of widespread airway closure during bronchoconstriction. We hypothesize that abnormal peripheral airway function at baseline, which increases ventilation heterogeneity, is associated with more extensive airway closure during bronchocon-

8750-7587/12 Copyright © 2012 the American Physiological Society

http://www.jappl.org

Imaging Airway Closure in Asthma

striction, leading to greater loss of ventilation. Furthermore, we hypothesize that the loss of ventilation during bronchoconstriction is increased in subjects with more severe AHR. The aim of the present study was to determine whether the extent of ventilation loss measured by ventilation single photon emission computed tomography (VSPECT)/CT during bronchoconstriction [standardized to 20% fall in forced expiratory volume in 1 s (FEV1)] was related to the severity of AHR to methacholine and to the severity of peripheral airway disease, measured by multiple-breath nitrogen washout (MBNW), in asthmatic subjects. Some of the results of this work have been previously reported in abstract form (42). METHODS

Subjects. Asthmatic subjects were recruited from the Department of Respiratory Medicine, Royal North Shore Hospital, the Woolcock Institute of Medical Research, and by local advertising. Asthma was defined as a respiratory physician diagnosis and a history of asthma symptoms and medication use within the last 6 mo. Exclusion criteria were current smoking and any history of smoking ⱖ10 pack·yr; FEV1 ⬍ 60% predicted or 1.4 liters before bronchial challenge testing; and respiratory tract infection within the last 4 wk. All subjects provided written, informed consent, and the study was approved by the local Human Research Ethics Committee [protocol no. 0512–232M (SP)]. Study design. The study design is depicted in Fig. 1. Subjects attended for two visits, 2–7 days apart, and were instructed to withhold short-acting bronchodilators for at least 8 h and long-acting bronchodilators for at least 24 h before each visit. At the initial visit, subjects completed the Asthma Control Questionnaire (ACQ) (16), underwent baseline lung function, including a MBNW test, followed by a VSPECT/CT scan, and then completed a methacholine challenge test. At the second visit, an abbreviated methacholine challenge test was administered. The second VSPECT scan was immediately acquired after the FEV1 decreased by 20% of the value recorded at the baseline scan. If baseline FEV1 at the second visit differed by ⬎10% compared with the baseline VSPECT/CT study, the subject did not proceed to scanning. Atopy was determined by skin prick testing. Lung function testing. Spirometry and lung volumes were measured in accordance with European Respiratory Society/American Thoracic Society guidelines (29) using a Vmax 22 Autobox (Sensormedics, Yorba Linda, CA) and Medgraphics CPFS/D USB spirometer (Medgraphics, St. Paul, MN). MBNW was performed as previously

Study Design Visit 1 Baseline • Questionnaire • Respiratory Function Tests • MBNW • VSPECT/CT • Methacholine Challenge

Visit 2 Methacholine Challenge • Spirometry • Modified Methacholine Challenge # • VSPECT

Fig. 1. Study design. The study was conducted over 2 visits that were a minimum of 48 h apart. #The modified methacholine challenge was an abbreviated challenge, starting with the penultimate dose of the initial challenge, to achieve a 20% decrease in forced expiratory volume in 1 s (FEV1) from baseline. MBNW, multiple-breath nitrogen washout; VSPECT/CT, ventilation single photon emission computed tomography.

•

Farrow CE et al.

959

described (8) from which lung clearance index (LCI), Scond and Sacin (measures of ventilation heterogeneity in peripheral conducting and diffusion dependent airways, respectively) were derived (8, 39). MBNW test and analysis technique. Briefly, after establishment of stable tidal breathing, subjects inhaled 1- to 1.3-liter tidal breaths at a rate of 8 –12 breaths/min until expired nitrogen concentration was ⬍2%. The phase III slopes of each washout breath were normalized by dividing by the mean expired nitrogen concentration. The expired volumes were normalized for lung size by dividing the cumulative expired volume, by the functional residual capacity (FRC) derived from the MBNW. FRC was determined by measurement of the total cumulative expired nitrogen at the end of the washout. Thus one lung turnover represents a cumulative expired volume equal to FRC. The normalized phase III slopes were then plotted against lung turnover. Scond was determined by the slope of the linear regression between turnovers 1.5 and 6. Sacin was determined as the phase III slope of the first breath, minus the component attributed to conductive heterogeneity, based on the Scond value (lung turnover value of the first breath ⫻ Scond) (8). The LCI was determined as the number of lung turnovers at which mean expired nitrogen concentration reaches 4% of the baseline value and is a global index. Scond is an index of ventilation heterogeneity in convection-dependent airways, while Sacin is an index in diffusion-convection dependent airways (8, 39). Methacholine challenge protocol. Methacholine challenge was performed with cumulative doubling doses administered via a Koko dosimeter (PDS Instrumentation, Louisville, KY), ranging from 0.06 to 200 ␮mol, with spirometry being measured 60 s after methacholine administration. At the second visit, an abbreviated challenge was performed, starting with the penultimate dose of the initial challenge, to achieve a 20% decrease in FEV1. Pulse oximetry was monitored during methacholine challenge testing (BCI Fingerprint Pulse Oximeter, Waukesha, WI). AHR was defined as a provocative dose that causes 20% decline in FEV1 of ⱕ4 ␮mol. The severity of AHR was determined by the dose response ratio (%fall in FEV1/final dose ⫹ 3), which is a continuous measure that allows inclusion of data in which ⬍20% decrease in FEV1 occurred. The constant “3” is added to avoid negative values in completely unresponsive subjects (30), which then allows log transformation for statistical analysis. Lung imaging. VSPECT/CT was performed as previously described using Technegas as the ventilation agent (Cyclomedica Australia, Sydney, Australia) (11, 21). Technegas is a 100-nm radiolabeled carbon particle whose inhaled deposition has a similar distribution to a gas, but remains unchanged for 20 min or more (1, 23). This allows inhalation in the upright, seated posture with subsequent supine scanning. One-liter volumes of Technegas were slowly inhaled from FRC (⬃30 – 60 l/min), requiring two to four inhalations to achieve radioactivity of ⬃1.5 kilocounts/s required for imaging. The bag-inbox device allowed delivery of a 1-liter volume from a set preinspiratory lung volume (in this case FRC), with respiratory flow rates being displayed in real time to ensure the subject did not go over the 1-liter inhalation (Fig. 2). VSPECT images were then acquired, immediately followed by a low-dose CT scan. Subjects remained stationary on the same gantry, since the two scanners were juxtaposed. VSPECT and CT were both acquired during resting tidal breathing. Total dose for the entire study was ⬍4.9 mSv per subject. The SPECT/CT machine was purpose-built, incorporating a singleslice, low-energy, high-resolution collimator scanner (Picker PQ5000 CT) and dual-detector variable-angle hybrid SPECT/CT system (Phillips SKYLight) (12). All SPECT studies were acquired using a 128 ⫻ 128 matrix, at 15 s per stop with 3° steps over 360°. Low-dose CT was performed without contrast, with a low-dose protocol (30 mA, 120 kVp, pitch 1.5, slice thickness 4 mm). Image reconstruction and registration. The ventilation study was reconstructed using ordered subset expectation maximization, with 128 slices and a Butterworth filter with a cut-off 0.8 and order 9. CT images are reconstructed using a 512 ⫻ 512 matrix with a Siemens

J Appl Physiol • doi:10.1152/japplphysiol.01618.2011 • www.jappl.org

960


•

Farrow CE et al.

c Patient Mouth

c b

Fig. 2. Technegas inhalation circuit. This is a one-way breathing circuit for subjects to inhale Technegas in 1-liter breaths, where breaths are monitored by the attached laptop to ensure correct volume. a, Pneumotach; b, four-way valve; c, one-way valve.

Filter Technegas

Room Air Computer

a

algorithm and transferred to a HERMES (Nuclear Diagnostics, Stockholm, Sweden) workstation in DICOM format. Registration of the SPECT ventilation study to CT scan was performed using the known, fixed spatial relationship between the SPECT and CT scanners. Additional three-dimensional registration was further performed using a HERMES Multimodality software package (Nuclear Diagnostics), which used a mutual information algorithm (12). Image analysis. Airway closure was determined from the registered VSPECT and CT scans, as shown in Fig. 3AA. A CT mask of less than ⫺500 Hounsfield units was used to define the lung outline, which was then applied to the VSPECT image. The CT lung outline was the FRC measured from CT image data, as the total number of voxels ⫻ voxel volume. Any voxels in the VSPECT image dataset lying outside the lung outlines were set to zero and thus removed from further analysis. This eliminated localized spots of Technegas activity that lay outside the lung, such as the pharynx, esophagus, and stomach, which is a common occurrence in clinical VSPECT imaging. Quantification of nuclear ventilation scans requires removal of noise due to photon scatter. This is achieved using a “threshold”, which is a single value derived from the distribution of Technegas activity. A threshold value is calculated individually from each scan, allowing normalization across all VSPECT scans. Deposition in locations of turbulence (typically in very proximal airways) causes “hot spots”, i.e., localized areas of high activity, which is another common artifact in nuclear medicine ventilation imaging. Neither scatter nor hot spots accurately represent ventilation and are extreme values in the frequency distribution of Technegas activity. The image analysis algorithm must, therefore, adapt to the wide variation in Technegas distributions for accurate volume measurements. Traditional threshold methods are based on the hot spot (20), which are inadequate for very heterogeneous ventilation that occurs in asthma (3). Therefore, an adaptive threshold was used to calculate ventilated lung volume and eliminate noise due to scatter in the VSPECT image. This was done using the following steps on the VSPECT data set after applying the CT mask (see Fig. 3B): 1) voxels with activity values between 5 and 80% of maximum were identified; 2) the mean ventilation value from this data set was calculated; 3) 50% of this mean value was used as the threshold; and 4) anything less than this value in the original CT-masked data set was considered to be

nonventilated lung and set to zero. In using this method, extreme values (scatter and hotspots) had negligible effect on the threshold value, so that it was then based on the vast majority of ventilated lung. The threshold method was verified by comparing baseline ventilated volumes calculated from this method against FRCs derived from MBNW. Threshold values were typically 15–30% of peak values, depending on the skewness of the distribution of ventilation values. Lung volume was defined from the CT, while ventilated volume was defined from VSPECT. Airway closure was calculated as the nonventilated volume (NVV) on VSPECT (a measurement of airway closure and/or severe regional narrowing) and reported as a proportion of total lung volume determined from the CT (Fig. 4). The topography of airway closure was defined in relation to its craniocaudal distribution by dividing the lung into thirds by axial length based on CT slice. NVV in each lung third was described as a percentage of the CT volume of that third to account for differences in regional volume. The ratio of closure (NVV) in the middle to lower thirds was used as a measure of distribution. A ratio ⬎1 indicates greater closure in the middle third. Statistical analyses. Comparisons between closure at baseline and postmethacholine were examined using paired t-tests, significance P ⬍ 0.05. Data are presented as means ⫾ SD, unless otherwise stated. The relationship between the ratio of middle to lower third closure and AHR was examined using the Spearman rank correlation. The relationships between the extent of airway closure and lung function parameters (spirometry, ventilation heterogeneity, and AHR), were examined using linear regression analyses. For an expected Pearson correlation coefficient of 0.65, we required 15 subjects for 80% power at the 0.05 level of significance (8). Data were analyzed using Analyse-It for Excel (Analyse-It Software, Leeds, UK). RESULTS

Seventeen asthmatic subjects were recruited (age range 19 – 67 yr), but three subjects were subsequently excluded at the second visit as their baseline FEV1 differed from the initial visit by ⬎10%. Characteristics for the remaining 14 subjects are shown in Table 1. There was a wide range of asthma severity, as determined by the wide range of AHR and spirometry values. Seven of fourteen subjects had AHR. At the second



•

Farrow CE et al.

961

Fig. 3. A: imaging analysis technique: VSPECT ventilation (A); VSPECT ventilation with CT mask overlaid (B); VSPECT ventilation with pixels outside CT mask set to zero (C); CT mask of lung (D); fused VSPECT and CT (E). VSPECT images are registered to the CT and then masked using CT lung mask. Histogram of pixel intensity is created from masked VSPECT data. B: histogram of pixel intensity. A histogram of pixels per count taken from a subject’s VSPECT scan after processing shows key values for image analysis. 5%, 5% of maximum pixel value; 50% mMean, 50% of modified mean; mMean, modified mean (the mean of the pixels from 5– 80% of maximum pixel value); 80%, 80% of maximum pixel value.

visit, FEV1 was 88 ⫾ 14% predicted, which was not significantly different from the first visit (P ⱖ 0.05). Measures of ventilation heterogeneity (Scond, Sacin, and LCI) obtained from MBNW were similar to previously published data in asthmatic subjects (8). Nine of fourteen subjects were taking regular inhaled corticosteroids, and eight of fourteen subjects were also taking long acting ␤-agonist. Asthma was well to moderately controlled in the majority with ACQ scores ranging from 0.3 to 1.7.

The volume of the entire lung calculated from CT acquired at FRC was 2.96 ⫾ 0.84 liters and correlated with plethysmographic FRC of 3.13 ⫾ 0.94 liters (r ⫽ 0.84, P ⫽ 0.0002). Figure 5A shows the comparison using Bland and Altman plots, which indicate that there is no systematic difference in measurement between the two methods. The mean ventilated volume measured from VSPECT was 2.46 ⫾ 0.61 liters and correlated with mean nitrogen washout FRC of 2.47 ⫾ 0.63 liters (r ⫽ 0.76, P ⫽ 0.0015), but not with plethysmographic


962


•

Farrow CE et al.

Fig. 4. VSPECT and CT fusion images. Baseline images (A, C, and E) and corresponding images after methacholine challenge (B, D, and F) are shown. A and B are coronal slices, and images C–F are axial images. Regional ventilation is represented by the color bar on the left, with the greatest ventilation represented by red, and the lowest by purple. Absent ventilation due to airway closure and narrowing is represented by no color, seen in B and D, while images E and F show airway closure in pink, as determined by image analyses described in the text. After methacholine challenge, there is an increase in nonventilation, seen as a reduction in colored areas in D compared with C, which has been measured by the increase in pink areas in F compared with E.

FRC. Both VSPECT volume and nitrogen washout FRC are dependent on airways that are open to the mouth during tidal breathing. Figure 5B shows that there were no systematic differences in measurement between the two methods. The baseline NVV representing airway closure was 16 ⫾ 4% of the total volume. At baseline, airway closure did not correlate with FVC (%predicted), ventilation heterogeneity, or AHR. Methacholine challenge caused significant reduction in spirometric function and an increase in airway closure on imaging, as seen in fused images (Fig. 4). The mean methacholine dose given at the second visit was 6.5 ⫾ 9 ␮mol. The mean decrease in FEV1 was 22.6 ⫾ 5% of baseline (range 17–35%, P ⬍ 0.0001) and in FVC was 15 ⫾ 6% (range 7–28%, P ⬍ 0.0001). The FEV1-to-FVC ratio decreased from 73.2 ⫾ 8.9 to 66.2 ⫾ 8.6% (P ⫽ 0.001), and oxygen saturation, measured by Table 1. Subject characteristics from visit 1 Sex (M/F) Age, yr Atopic (Y/N) BMI, kg/m2 FEV1, %predicted FVC, %predicted FEV1/FVC, % DRR mean (range), %decrease FEV1/␮mol Scond, liter Sacin, liter LCI (lung turnovers), no. Asthma control score

8/6 40.9 ⫾ 18.8 14/0 25.5 ⫾ 4.1 87.6 ⫾ 15.8 99.4 ⫾ 10.5 73.3 ⫾ 8.9 27.7 (3.1–99.9) 0.068 ⫾ 0.04 0.193 ⫾ 0.06 10.3 ⫾ 2.25 0.78 ⫾ 0.53

Values are means ⫾ SD. M, male; F, female; Y, yes; N, no; BMI, body mass index; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; DRR, dose response ratio; Scond and Sacin, measures of ventilation heterogeneity in peripheral conducting and diffusion dependent airways, respectively; LCI, lung clearance index.

pulse oximetry, was unchanged (P ⱖ 0.05). After methacholine challenge, mean ventilated volume from VSPECT was 2.07 ⫾ 0.53 liters, and airway closure increased from 16 ⫾ 4% at baseline to 29 ⫾ 13% of total lung volume (P ⫽ 0.002). The change in airway closure measured by NVV, induced by methacholine challenge, correlated with AHR (r ⫽ 0.56, P ⫽ 0.04) (Fig. 6A) and with the three indexes of ventilation heterogeneity at baseline (Scond: r ⫽ 0.58, P ⫽ 0.028, Fig. 6B; Sacin: r ⫽ 0.60, P ⫽ 0.025; LCI: r ⫽ 0.69, P ⫽ 0.007). However, this change in NVV did not correlate with changes in FEV1 or FVC (r ⫽ 0.05, P ⫽ 0.85 and r ⫽ 0.10, P ⫽ 0.74, respectively). In a multivariate analysis, the change in NVV induced by methacholine correlated independently with AHR and with Scond (total R2 ⫽ 0.60, P ⫽ 0.007). The partial R2 values were 0.22 for Scond and 0.38 for AHR. Airway closure measured by NVV at baseline was greater in the lower compared with the middle third of the lung (23 ⫾ 11 and 9 ⫾ 2.6% of regional lug volume, P ⬍ 0.0001) (Fig. 7). The extent of NVV at baseline in the upper third (23 ⫾ 6.2%) was greater than in the middle third (P ⬍ 0.0001), but equal to that in the lower third. After methacholine challenge, NVV in the lower third (48 ⫾ 23.6%) was greater than NVV in both the middle (21 ⫾ 13.5%, P ⫽ 0.0003) and upper thirds (26 ⫾ 18%, P ⫽ 0.02). Therefore a gravitational gradient occurred after induced airway closure, with no increase in NVV in the upper third (3 ⫾ 18%), but increased NVV in the middle (12 ⫾ 14%) and the lower thirds (25 ⫾ 22%). The ratio of NVV in the middle to lower thirds, either at baseline or after methacholine challenge, did not correlate with age, AHR, spirometry, indexes of ventilation heterogeneity measured by MBNW, body mass index, or asthma control score.



A

•

963

Farrow CE et al.

B 1.20

MBNW – SPECT Lung Volume (Litres)

Plethysmographic – CT Lung Volume (Litres)

1.50 1.00 0.50 0 -0.50 -1.00 -1.50

0.80 0.20 0 -0.40 -0.80 -1.20

Average Plethysmographic & CT Lung Volumes (Litres)

Average MBNW & SPECT Lung Volumes (Litres)

Fig. 5. A: Bland and Altman plot of the difference between plethysmographic functional residual capacity (FRC) and CT measured lung volume. B: Bland and Altman plot of the difference between MBNW FRC and single-photon emission computed tomography (SPECT) ventilated volume.

Although AHR and ventilation heterogeneity at baseline both correlated with the change in airway closure from VSPECT after challenge, they were not correlated with each other (AHR vs. Scond: r ⫽ 0.13, P ⫽ 0.66; AHR vs. Sacin: r ⫽ 0.29, P ⫽ 0.32; AHR vs. LCI: r ⫽ 0.38, P ⫽ 0.18). AHR was also unrelated to residual volume/total lung capacity, FRC/total lung capacity, and FEV1/FVC. The ACQ score did not correlate with dose-response ratio, VSPECT airway closure at baseline, or the change induced by methacholine challenge. DISCUSSION

In this study of asthmatic subjects, the extent of airway closure induced by methacholine challenge was related to the severity of ventilation heterogeneity at baseline. Furthermore,

B 60

Change in Airway Closure (% of lung volume)

Change in Airway Closure (% of lung volume)

A

we found that subjects with more severe AHR had the greatest NVV, despite a similar level of bronchoconstriction to those subjects with mild or no AHR. Additionally, there was greater NVV in the lower third of the lung, but its distribution, in relation to middle third or lower third predominance, was not associated with AHR, ventilation heterogeneity, or asthma severity. The significant relationship demonstrated between induced NVV, due to severe airway narrowing and closure measured by VSPECT/CT, and AHR extends the observations of Chapman et al. (5), as well as the association between airway closure and severe asthma (9, 14). In addition, the correlation between NVV and Scond suggests that peripheral airway dysfunction may predispose to severe airway narrowing and closure during bronchoconstriction.

50 40 30 20 10 0 -10 -20 -30 0.4

0.9

1.4

1.9

Dose response ratio (log( %Fall FEV1/ µmol methacholine +3))

60 50 40 30 20 10 0 -10 -20 -30 -1.8

-1.3

-0.8

log Baseline Scond (L-1)

Fig. 6. Univariate relationships between the change in airway closure measured by VSPECT/CT and airway hyperresponsiveness measured by DRR (r ⫽ 0.56, P ⫽ 0.04; A) and conductive airway ventilation heterogeneity (Scond) at baseline (r ⫽ 0.60, P ⫽ 0.02; B). J Appl Physiol • doi:10.1152/japplphysiol.01618.2011 • www.jappl.org

964


Upper 3rd

Middle 3rd

Lower 3rd

Closure (% of CT volume) Fig. 7. The distribution of airway closure in gravitational lung thirds at baseline and after methacholine challenge, as a percentage of CT lung volume in that third. , Baseline. Œ, Postmethacholine. Closure is least in the middle third of the lungs, both before and after methacholine challenge.

The correlation between the increase in NVV and heterogeneity at baseline supports a previously published model that suggests regionally clustered airway closure results from ventilation heterogeneity (38). In this model, ASM activation leads to widespread airway narrowing, which, in a heterogeneous airway tree, causes large regions of airways to become unstable, leading to airway closure. It has been suggested that this regional instability is secondary to the interaction of forces between airways in close proximity. Consequently, narrowing of airways is highly interdependent. Further modeling suggests that closure is dependent on widespread heterogeneous peripheral airway narrowing with severe narrowing or closure in more proximal airways (37). The development of nonventilated regions due to severe airway narrowing and closure during bronchoconstriction (Fig. 4) has been observed in previous ventilation imaging studies in which ventilation was measured in a variety of ways (7, 31, 33, 36). In the present study, the concurrent low-dose CT scan was used to define the anatomical outline of the lung, and VSPECT was used to define ventilation, allowing objective measurement of NVV in relation to CT lung volume. This was an improvement of a previously validated method, where VSPECT closure correlated closely with CC in healthy subjects (21). Although the low-dose CT did not allow further density measures for heterogeneity, FRC measured by CT and by plethysmography were similar. Since FRC was likely reduced in supine posture during CT (27), it was surprising that CTmeasured FRC was not less than FRC measured by body plethysmography. This may be due to overestimation of CT lung volume due to the inclusion of blood vessels and to blurring of the lung borders due to breathing artifact. VSPECT measures only lung regions that communicate with the mouth, and it is reassuring that it is similar to MBNW FRC. Nevertheless, some errors in registration of the VSPECT and CT images may have occurred, which would have affected the measurement of NVV. The misregistrations would have unpredictable effects on the measurements of NVV, but are unlikely to systematically cause a spurious relationship between NVV and Scond and AHR.

•

Farrow CE et al.

The contribution of imaging to understanding airway closure and AHR is its topographical localization. In the present study, subjects who have worse AHR had a larger volume of lung that closed at similar falls in FEV1 of ⬃20%. Baseline closure, however, did not correlate with AHR. We speculate that the lung regions that closed may be responsible for driving AHR, whereas airways already closed at baseline may already be near their maximal response and may not be capable of any further augmentation with bronchoconstriction. This topographical construct is consistent with the known dysfunction of peripheral airways in asthma (5, 8, 15, 17, 18, 41) and its association with AHR (5, 8). Closure occurs both peripherally and centrally (8, 10) and could be due to worse inflammation, edema, mucous production, surfactant abnormalities, airway wall thickening, and changes in wall compliance (19, 40). By subjective assessment, airway closure occurs at the same location on different days (6). Confirmation by objective analyses and reduction by anti-inflammatory treatment, in parallel with improvements in AHR, would be needed to determine whether closure was due to localized rather than diffuse or transient phenomena, such as mucous or surfactant abnormalities. This would have implications for inhaled treatment with respect to targeting aerosol treatment to specific locations. Airway closure measured by VSPECT and by FVC did not correlate in the present study and is likely to represent different aspects of function. Similarly, closure measured by VSPECT correlates with single-breath CC in normal subjects (21), but does not correlate in asthmatic subjects (22). In a hyperpolarized He-3 MRI study, methacholine-induced closure measured subjectively as the average number of defects per MRI slice (6), correlated with spirometry when pre- and postmethacholine data were pooled. There were similar findings in the present study between pooled data of FVC %predicted and airway closure (r ⫽ ⫺0.47, P ⫽ 0.01), indicating that, as expected, bronchoconstriction decreased FVC and increased the extent of airway closure on VSPECT. However, there were no such relationships in either premethacholine or postmethacholine data analyzed separately, nor in the changes. Thus pooling data of two different physiological states may be of dubious significance. The absence of relationship between FVC and VSPECT measured closure may also be because Technegas was inhaled as a 1-liter bolus from FRC and because of volume history. The VSPECT imaging, therefore, reflects the extent of closure at FRC on inspiration (so that a comparison with MBNW can be made), whereas changes in FVC are due to airways closing during forced expiration, usually below FRC. Thus the differences in physiological conditions in which spirometry and VSPECT imaging occur make any correlations between changes in FVC and imaging unlikely. The spatial resolution of the SPECT images is ⬃15 mm3 and could affect these relationships. We were only able to detect closure that affected lung regions of ⬃4 –5 ml in volume. Smaller subresolution areas affected by closure were probably present and would contribute to the reduction in FVC. Finally, NVV may be due to closure and severe narrowing, and the effects of these two phenomena on NVV may differ to their effects on FVC, leading to a disassociation. We hypothesized that the responsive airways that close occurred anywhere in lung, rather than being predominantly basal (21, 22), and that the distribution of closure (described in gravitational thirds) was related to AHR. Our hypothesis was



not supported, and we observed greater airway closure in the lower third of the lung. Therefore, although closure may be patchy in asthmatic subjects, it still follows a vertical distribution, suggesting that the gradient of elastic recoil operates in asthma as it does in normal subjects, which is in keeping with early distribution studies (3). Nevertheless, there could also be greater inflammation and remodeling in the lower third, similar to chronic obstructive pulmonary disease (2). The NVV in the upper third of the lung was greater than in the middle third. The difference was unlikely to be due to more airway closure in the upper third, but due to high lung recoil pressures, such that alveoli are closer to full expansion at FRC (28). Thus a 1-liter breath would minimally distend alveoli, resulting in very low specific ventilation and minimal Technegas deposition, perhaps below the lower limit of detection of our technique. Technegas was inhaled in the upright posture, but VSPECT were acquired in the supine posture. Although the change in posture after Technegas inhalation may have affected the measured NVV, due to changes in regional lung inflation (27), this effect was likely to be small, as suggested by previous validation (21). If anything, differences in posture may have weakened the associations between closure and AHR and ventilation heterogeneity. Although Technegas was inhaled slowly, some nonventilated regions on VSPECT may be due to severe airway narrowing rather than true closure. The nitrogen washout phase 3 slopes, however, represent differences in regional time constants. Therefore, some of the relationship between Scond (derived from the nitrogen washout phase 3) and airway closure by VSPECT may be related to severe regional airway narrowing, resulting in very long local time constants. The strong relationship between CC and SPECT airway closure, however (21), suggests that the relationship is driven predominantly by closure rather than narrowing. Four subjects did not have AHR, as defined by a provocative dose that causes 20% decline in FEV1 ⬍ 4 ␮mol. Although it is possible they may not have asthma (24), their lack of AHR was likely due to continuing inhaled corticosteroid treatment (25). In summary, airway closure in asthmatic subjects induced by bronchoconstriction is related independently both to AHR and to ventilation heterogeneity. These data provide support for computational model predictions that suggest that airways tend to close rather than narrow when peripheral airway disease is worse, and that this closure is greater with more severe AHR. These findings have potential implications for the underlying mechanisms of AHR, in particular whether the location of induced closure is reproducible in asthma and if it is affected by anti-inflammatory treatment. Further studies should examine these questions and also compare asthmatic imaging data with nonasthmatic normal subjects. ACKNOWLEDGMENTS The authors acknowledge the support from the volunteers who participated in this study, and Robin Schoeffel who assisted with subject visits. They also thank all of the staff in the Departments of Respiratory Medicine and Nuclear Medicine at Royal North Shore Hospital, Sydney, and the Airway Physiology and Imaging Group at the Woolcock Institute of Medical Research for assistance. GRANTS This study was suppored by National Health and Medical Research Council of Australia Project Grant 457346, The Barbara Dunn Trust Fund, and The Cooperative Research Centre for Asthma (Project 2.1).

•

Farrow CE et al.

965

DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the author(s). AUTHOR CONTRIBUTIONS Author contributions: C.E.F., C.M.S., B.E.H., D.L.B., N.B., I.H.Y., and G.G.K. conception and design of research; C.E.F., D.L.B., and E.B. performed experiments; C.E.F., B.E.H., D.L.B., E.B., and G.G.K. analyzed data; C.E.F., C.M.S., B.E.H., D.L.B., E.B., N.B., and G.G.K. interpreted results of experiments; C.E.F. and G.G.K. prepared figures; C.E.F., E.B., and G.G.K. drafted manuscript; C.E.F., C.M.S., N.B., I.H.Y., and G.G.K. edited and revised manuscript; C.E.F., C.M.S., B.E.H., D.L.B., E.B., N.B., I.H.Y., and G.G.K. approved final version of manuscript. REFERENCES 1. Amis TC, Crawford ABH, Davison A, Engel LA. Distribution of inhaled 99mTechnetium labelled ultrafine carbon particle aerosol (Technegas) in human lungs. Eur Respir J 3: 679 –685, 1990. 2. Berend N. Lobar distribution of bronchiolar inflammation in emphysema. Am Rev Respir Dis 124: 218 –220, 1981. 3. Berend N, Salome CM, King GG. Mechanisms of airway hyperresponsiveness in asthma. Respirology 13: 624 –631, 2008. 4. Brown R, Woolcock AJ, Vincent NJ, Macklem PT. Physiological effects of experimental airway obstruction with beads. J Appl Physiol 27: 328 –335, 1969. 5. Chapman DG, Berend N, King GG, Salome CM. Increased airway closure is a determinant of airway hyperresponsiveness. Eur Respir J 32: 1563–1569, 2008. 6. de Lange EE, Altes TA, Patrie JT, Parmar J, Brookeman JR, Mugler JP, Platts-Mills TAE. The variability of regional airflow obstruction within the lungs of patients with asthma: assessment with hyperpolarized helium-3 magnetic resonance imaging. J Allergy Clin Immunol 119: 1072–1078, 2007. 7. de Lange EE, Altes TA, Patrie JT, Gaare JD, Knake JJ, Mugler JP III, Platts-Mills TA. Evaluation of asthma with hyperpolarized helium-3 MRI: correlation with clinical severity and spirometry. Chest 130: 1055– 1062, 2006. 8. Downie S, Salome C, Verbanck S, Thompson B, Berend N, King G. Ventilation heterogeneity is a major determinant of airway hyperresponsiveness in asthma, independent of airway inflammation. Thorax 62: 684 –689, 2007. 9. Gibbons WJ, Sharma A, Lougheed D, Macklem PT. Detection of excessive bronchoconstriction in asthma. Am J Respir Crit Care Med 153: 582–589, 1996. 10. Gustafsson PM, Ljungberg HK, Kjellman B. Peripheral airway involvement in asthma assessed by single-breath SF6 and He washout. Eur Respir J 21: 1033–1039, 2003. 11. Harris B, Bailey D, Miles S, Bailey E, Rogers K, Roach P, Thomas P, Hensley M, King G. Objective analysis of tomographic ventilationperfusion scintigraphy in pulmonary embolism. Am J Respir Crit Care Med 175: 1173–1190, 2007. 12. Harris B, Bailey D, Roach P, Bailey E, King G. Fusion imaging of computed tomographic pulmonary angiography and SPECT ventilation/ perfusion scintigraphy: initial experience and potential benefit. Eur Respir J 34: 135–142, 2007. 13. Harris RS, Winkler T, Tgavalekos N, Musch G, Melo MFV, Schroeder T, Chang Y, Venegas JG. Regional pulmonary perfusion, inflation, and ventilation defects in bronchoconstricted patients with asthma. Am J Respir Crit Care Med 174: 245–253, 2006. 14. in’t Veen JC, Beekman AJ, Bel EH, Sterk PJ. Recurrent exacerbations in severe asthma are associated with enhanced airway closure during stable episodes. Am J Respir Crit Care Med 161: 1902–1906, 2000. 15. Irvin CG, Bates JHT. Physiologic dysfunction of the asthmatic lung: what’s going on down there, anyway? Proc Am Thorac Soc 6: 306 –311, 2009. 16. Juniper E, O’Byrne P, Guyatt G, Ferrie P, King D. Development and validation of a questionnaire to measure asthma control. Eur Respir J 14: 902–907, 1999. 17. Kaminsky D, Wenzel S, Carcano C, Gurka D, Feldsien D, Irvin C. Hyperpnea-induced changes in parenchymal lung mechanics in normal subjects and in asthmatics. Am J Respir Crit Care Med 155: 1260 –1266, 1997.


966


18. Kaminsky DA, Bates JH, Irvin CG. Effects of cool, dry air stimulation on peripheral lung mechanics in asthma. Am J Respir Crit Care Med 162: 179 –186, 2000. 19. King G, Paré P, Seow C. The mechanics of exaggerated airway narrowing in asthma; the role of smooth muscle. Respir Physiol 118: 1–13, 1999. 20. King GG, Carroll JD, Muller NL, Whittall KP, Gao M, Nakano Y, Pare PD. Heterogeneity of narrowing in normal and asthmatic airways measured by HRCT. Eur Respir J 24: 211–218, 2004. 21. King GG, Eberl S, Salome CM, Meikle SR, Woolcock AJ. Airway closure measured by a Technegas bolus and SPECT. Am J Respir Crit Care Med 155: 682–688, 1997. 22. King GG, Eberl S, Salome CM, Young IH, Woolcock AJ. Differences in airway closure between normals and asthmatics measured by SPECT and Technegas. Am J Respir Crit Care Med 158: 1900 –1908, 1998. 23. Lemb M, Oei TH, Eifert H, Gunther B. Technegas: a study of particle structure, size and distribution. Eur J Nucl Med 20: 576 –579, 1993. 24. Luks VP, Vandemheen KL, Aaron SD. Confirmation of asthma in an era of overdiagnosis. Eur Respir J 36: 255–260, 2010. 25. Lundbäck B, Rönmark E, Lindberg A, Jonsson AC, Larsson LG, James M. Asthma control over 3 years in a real-life study. Respir Med 103: 348 –355, 2009. 26. Macklem P. Factors determining bronchial smooth muscle shortening. Am Rev Respir Dis 143: S47–S48, 1991. 27. Meinero M, Coletta G, Dutto L, Milanese M, Nova G, Sciolla A, Pellegrino R, Brusasco V. Mechanical response to methacholine and deep inspiration in supine men. J Appl Physiol 102: 269 –275, 2007. 28. Milic-Emili J, Henderson JA, Dolovich MB, Trop D, Kaneko K. Regional distribution of inspired gas in the lung. J Appl Physiol 21: 749 –759, 1966. 29. Miller MR, Crapo R, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A, Enright P, Grinten CPM, Gustafsson P, Jensen R, Johnson DC, MacIntyre N, McKay R, Navajas D, Pedersen OF, Pellegrino R, Viegi G, Wanger J. General considerations for lung function testing. Eur Respir J 26: 153–161, 2005. 30. O’Connor G, Sparrow D, Taylor D, Segal M, Weiss S. Analysis of dose response curves to methacholine. An approach suitable for population studies. Am Rev Respir Dis 136: 1412–1417, 1987. 31. Pellegrino R, Biggi A, Papaleo A, Camuzzini G, Rodarte JR, Brusasco V. Regional expiratory flow limitation studied with Technegas in asthma. J Appl Physiol 91: 2190 –2198, 2001. 32. Rasmussen F, Taylor DR, Flannery EM, Cowan JO, Greene JM, Herbison GP, Sears MR. Risk factors for hospital admission for asthma from childhood to young adulthood: a longitudinal population study. J Allergy Clin Immunol 110: 220 –227, 2002. 33. Samee S, Altes T, Powers P, de Lange EE, Knight-Scott J, Rakes G, Mugler JP 3rd,k Ciambotti JM, Alford BA, Brookeman JR, Platts-

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

•

Farrow CE et al.

Mills TA. Imaging the lungs in asthmatic patients by using hyperpolarized helium-3 magnetic resonance: assessment of response to methacholine and exercise challenge[see comment]. J Allergy Clin Immunol 111: 1205– 1211, 2003. Sorkness RL, Bleecker ER, Busse WW, Calhoun WJ, Castro M, Chung KF, Curran-Everett D, Erzurum SC, Gaston BM, Israel E, Jarjour NN, Moore WC, Peters SP, Teague WG, Wenzel SE, for the National Heart Lung, and Blood Institute Severe Asthma Research Program. Lung function in adults with stable but severe asthma: air trapping and incomplete reversal of obstruction with bronchodilation. J Appl Physiol 104: 394 –403, 2008. ten Brinke A, van Dissel JT, Sterk PJ, Zwinderman AH, Rabe KF, Bel EH. Persistent airflow limitation in adult-onset nonatopic asthma is associated with serologic evidence of Chlamydia pneumoniae infection. J Allergy Clin Immunol 107: 449 –454, 2001. Tgavalekos NT, Musch G, Harris RS, Vidal Melo MF, Winkler T, Schroeder T, Callahan R, Lutchen KR, Venegas JG. Relationship between airway narrowing, patchy ventilation and lung mechanics in asthmatics. Eur Respir J 29: 1174 –1181, 2007. Tgavalekos NT, Tawhai M, Harris RS, Mush G, Vidal-Melo M, Venegas JG, Lutchen KR. Identifying airways responsible for heterogeneous ventilation and mechanical dysfunction in asthma: an image functional modeling approach. J Appl Physiol 99: 2388 –2397, 2005. Venegas JG, Winkler T, Musch G, Vidal Melo MF, Layfield D, Tgavalekos N, Fischman AJ, Callahan RJ, Bellani G, Scott Harris R. Self-organized patchiness in asthma as a prelude to catastrophic shifts. Nature 434: 777–782, 2005. Verbanck S, Schuermans D, Van Muylem A, Paiva M, Noppen M, Vincken W. Ventilation distribution during histamine provocation. J Appl Physiol 83: 1907–1916, 1997. Wagers S, Lundblad LKA, Ekman M, Irvin CG, Bates JHT. The allergic mouse model of asthma: normal smooth muscle in an abnormal lung? J Appl Physiol 96: 2019 –2027, 2004. Wagner EM, Liu MC, Weinmann GG, Permutt S, Bleecker ER. Peripheral lung resistance in normal and asthmatic subjects. Am Rev Respir Dis 141: 584 –588, 1990. Walsh C, Harris B, Bailey D, Salome C, Berend N, Downie S, King G. Ventilation heterogeneity measured by single photon emission computed tomography/computed tomography (SPECT/CT) and multiple breath nitrogen washout (MBNW) in asthma (Abstract). Respirology 13, suppl S2: A1, 2008. Zeidler MR, Goldin JG, Kleerup EC, Kim HJ, Truong DA, Gjertson DW, Kennedy NJ, Newman KB, Tashkin DP, Silverman JM, Corren J. Small airways response to naturalistic cat allergen exposure in subjects with asthma. J Allergy Clin Immunol 118: 1075–1081, 2006.
