We studied both central conducting airway response and changes in the distribution of regional ventilation induced by inhaled histamine in healthy anesthetized and mechanically ventilated rabbit using a novel xenon-enhanced synchrotron radiation computed tomography (CT) imaging technique, K-edge subtraction imaging (KES). Images of specific ventilation were obtained using serial KES during xenon washin, in three axial lung slices, at baseline and twice after inhalation of histamine aerosol (50 or 125 mg/ml) in two groups of animals (n = 6 each). Histamine inhalation caused large clustered areas of poor ventilation, characterized by a drop in average specific ventilation (sV̇m), but an increase in sV̇m in the remaining lung zones indicating ventilation redistribution. Ventilation heterogeneity, estimated as coefficient of variation (CV) of sV̇m significantly increased following histamine inhalation. The area of ventilation defects and CV were significantly larger with the higher histamine dose. In conducting airways, histamine inhalation caused a heterogeneous airway response combining narrowing and dilatation in individual airways of different generations, with the probability for constriction increasing peripherally. This finding provides further in vivo evidence that airway reactivity in response to inhaled histamine is complex and that airway response may vary substantially with location within the bronchial tree.
- X-ray computed
asthma is characterized by heterogeneous airway constriction that leads to nonuniform alveolar ventilation. Several studies using high-resolution computed tomography (CT) in both human (5) and animal models (4) have established the heterogeneity of individual airway response to constricting stimuli. Ventilation heterogeneity, on the other hand, has been documented using different imaging techniques such as scintigraphy (32), single-photon emission tomography (20), nuclear magnetic resonance imaging following inhalation of hyperpolarized helium (9), deposition of fluorescent microsphere aerosol (30), and positron emission tomography in both human and animal models (38, 39). None of these imaging techniques allow simultaneous high-resolution imaging of both central airway geometry and the quantitative distribution of regional ventilation, limiting the ability to study structure-function relationships. Xenon-enhanced CT, on the other hand, allows the demonstration of both central airway structure and the measurement of regional lung ventilation in large animal models (13). More recently, paradoxical airway responses combining distal constriction and proximal dilatation in trachea and the main bronchi have been demonstrated in human asthmatic patients in response to inhaled methacholine (21). Computational modeling suggests that complex airway behavior during provocation, combining constriction and dilatation within the central airways, associated with severe regional ventilation heterogeneity with large areas of contiguous poorly ventilated areas can be simulated (41).
An advantage of the K-edge subtraction imaging (KES) technique with synchrotron radiation CT (SRCT) is that it allows simultaneous imaging of central conducting airway dimensions and quantitative measurement of regional ventilation in experimental models of induced airway obstruction. Using KES and SRCT, we previously found in healthy rabbit that the distribution of ventilation is patchy after histamine provocation (26) and that there are substantial differences in the kinetics of response in central and peripheral airways to inhaled histamine (3). In the latter study, large central airways showed slower response to histamine inhalation and were maximally constricted only when small peripheral airways had significantly recovered. However, quantitative measurements of regional lung ventilation were not performed.
The objectives of the present study were 1) to quantitatively assess changes in central airway cross-sectional area following histamine inhalation and 2) to measure the effect of histamine inhalation on regional ventilation, on its heterogeneity, and on the area of poorly ventilated lung zones.
Animal care and procedures of the experiment were in accordance with the “Guidelines for the Care and Use of Animals” provided by the American Physiological Society and were approved by the local institutional authorities. The experiments were performed on 12 male New Zealand rabbits (2.00–2.50 kg; Elevage Scientifique des Dombes, Chatillon sur Chalaronne, France) in two groups to test the effect of histamine dose. A catheter (22 G) was inserted in the marginal ear vein (Cathlon IV; Ethicon, Rome, Italy) under local anesthesia using 5% topical lidocaine (Emla, Astra-Zeneca, France). Anesthesia was induced by intravenous injection of thiopental sodium (25 mg/kg iv, Nesdonal; Rhone-Poulenc-Rohrer, Paris, France). The animal was tracheostomized, and an endotracheal tube (no. 3; Portex, Berck sur Mer, France) was inserted and secured with a gas-tight seal. A catheter (22 G) was inserted into the left carotid artery for blood pressure monitoring and arterial blood sampling for blood gas measurements (ABL70; Radiometer, Copenhagen, Denmark). The lower extremities of the animal were wrapped with bandages, and the animal was immobilized in a homemade cylindrical polyvinyl chloride holder in a vertical position. The chest wall and diaphragmatic motion were free of constraint, because the rabbit was maintained in a vertical position using a plastic attachment below the lower maxillary bone, which supported the weight of the animal. There was very little stress on the forelimbs, which were taped to keep them out of the image field. The chest and diaphragm were entirely free, and the lower limbs were securely maintained in the cylindrical holder using foam. Anesthesia was maintained with 0.4–0.6% inhaled isoflurane (Forene; Abbott, Paris, France). Paralysis was induced by pancuronium bromide (1.0 mg/h iv, Norcuron; Organon, Fresnes, France).
Respiratory gas flow was monitored using a heated pneumotachometer (Hans Rudolph, Kansas City, MO). Endotracheal pressure was monitored continuously. All monitored signals were amplified, digitized at 400 Hz (PowerLab; ADInstruments, Oxford, UK), and recorded on a computer.
K-edge subtraction imaging and synchrotron instrumentation.
A detailed experimental setup description has been published earlier (3, 11). The KES method was originally developed for human coronary angiography with iodine contrast (31, 35) and was later extended to imaging lung ventilation with stable xenon gas and perfusion with iodine as contrast agent (12, 34). A description of this method has been published previously (28). The KES imaging method allows simultaneous imaging of lung structure and function (34). Visualization and quantitative measurement of xenon within the airways is based on the property that the attenuation coefficient of xenon increases by a factor of 5.4 when the energy of the incident X-ray beam crosses the energy threshold of 34.56 keV, which is called the xenon K-edge. The change in attenuation coefficients of cortical bone and lung tissue is negligible. Imaging is performed simultaneously with two X-ray beams at two slightly different energies, above and below the xenon K-edge. The two images are subtracted to obtain a quantitative image of contrast agent concentration. The lateral resolution, determined by the detector pixel width, was 350 μm.
Details of the ventilation system were published earlier (28). The setup consisted of a custom-made apparatus including a T tube and an electromagnetic valve, used to synchronize mechanical ventilation with the image acquisition. The input gas flow was maintained constant at baseline. The tidal volume (Vt) was ∼5.5 ml/kg, with a respiratory rate that was set to achieve an arterial Pco2 close to 40 Torr at baseline, and then ventilatory settings were maintained throughout the study. Ventilation and the inhaled gas mixture were remotely controlled with electromagnetic valves. Gas flows were continuously measured and recorded using mass flow meters (Aalborg, Orangeburg, NY) and were adjusted before data acquisition. During the resting period, the animal breathed air, and during imaging, it breathed a mixture of xenon (Xe 70%) and oxygen (O2 30%). Ventilation was paused in expiration for image acquisition.
Lung mechanics, ventilation parameters, and histamine challenge.
The overall respiratory system resistance (R) was calculated using a multiple linear regression method with the following relation between physiological and ventilation parameters: (1) where Ptr (cmH2O) is the tracheal pressure, P0 (cmH2O) is the dynamic positive end-expiratory pressure, E (cmH2O/l) is respiratory system elastance, V (liter) is the lung volume, R (cmH2O·s/l) is respiratory system resistance, and aw (ml/s) is gas flow. P0, E, and R are estimated for each respiratory cycle by multiple linear regression based on selected parts of the flow signal (10).
Histamine (Sigma, St Quentin Fallavier, France) solutions of 50 and 125 mg/ml in normal saline were prepared daily and administered using an ultrasonic nebulizer (SAM LS2000; Syst'AM, Villeneuve sur Lot, France). The mass median aerodynamic diameter (MMAD) of the aerosol particles was 3.5 μm with a geometrical standard deviation of 2.0, as determined by laser optical diffraction according to the manufacturer. Histamine solution in normal saline was administered continuously for 2 min, with delivered histamine doses of 180 and 440 cmH2O·s/l inspired air, for each histamine concentration, respectively.
Data acquisition protocol.
The experimental setup was placed in a radioprotective hutch that was closed at all times during image acquisition. Image acquisition and mechanical ventilation were remotely controlled from a computer in an adjacent control room. Animal position was checked from an anteroposterior thoracic reference image. On the basis of the reference image, three different axial slices were selected approximately in the third (apical), fifth (middle), and eighth (caudal) thoracic vertebral locations for tomographic image sequences. Cardiac gating was not used; however, it has been verified that cardiac motion causes only minor artifacts on the sequential CT images (28).
Figure 1 summarizes the experimental protocol. Two separate groups of animals (n = 6) were used for the two histamine doses. Each rabbit served as its own control. Imaging was performed in each slice at baseline and at two time intervals after histamine administration, posthistamine 1 and posthistamine 2, respectively. The imaging sequence started with a switch of the inhaled gas from air to the Xe-O2 mixture. After five respiratory cycles, ventilation was stopped upon expiration for 3 s, and CT imaging was performed during apnea, after which ventilation was resumed. The cycle was repeated 12 times to follow xenon washin. At the end of the sequence, inspired gas was switched back to air. Imaging was performed successively in the three lung slices, with time allowed between sequences for complete washout of residual xenon. The duration of each image sequence acquisition was ∼2.5 min. After acquisition of imaging data in the three slices, access to the hutch was allowed for arterial blood gas measurements at baseline and at posthistamine 1 and posthistamine 2.
Images were processed using the MatLab programming package (Mathworks, Novi, MI). For the calculation of specific ventilation, lung tissue within the monochromatic CT images was selected by segmentation based on tissue density thresholds. In this fashion, major blood vessels and airways were excluded from the analyses. In the one-compartment model, where the slight diffusion of xenon to pulmonary circulation is considered negligible given the short acquisition time, the xenon density for a given region of interest in the KES images increases in the washin sequence following monoexponential dynamics (28).
Specific ventilation (s) was estimated from a monoexponential model using (2) where d(t) is xenon density and dA is the asymptotic density; τ = 1/s and is the time constant of xenon washin. The arrival time t0 includes the transfer time from the gas valve to the trachea and the transfer time from the trachea to the entrance of the acini.
sV̇ was calculated for each voxel of the CT image to produce the sV̇ images (25, 36). The exponential fit was performed using multidimensional unconstrained nonlinear minimization (Nelder-Mead), and the goodness of fit was assessed by examination of the normalized summed squared residuals (SSR). To minimize the effects of image registration errors, images were smoothed before the model fit with a 5 × 5-pixel moving average window to produce 1.75 × 1.75-mm-resolution smoothed sV̇ images, where the effects of small differences (up to 2 pixels) in the animal position between successive images are eliminated.
For each sV̇ image, a histogram of sV̇ was calculated. At baseline, the distributions were unimodal. After histamine provocation, a broad, diffuse distribution appeared below the main distribution. Because of the fractal structure of the airways, the statistical distributions of regional ventilation at baseline and in the well-ventilated zones after histamine provocation are expected to be log normal (24). Therefore, this functional form was used for fitting the histograms of sV̇, using Nelder-Mead minimization. Goodness of fit was assessed on the basis of the SSR. The probability density function of the log-normal distribution for a random variable x is given by (23) (3)
This function was fit to the histogram of sV̇ for each rabbit and imaging slice (28). Poorly ventilated zones were defined as areas where sV̇ was less than the median of the distribution minus two standard deviations, (μ − 2σ). On the basis of this criterion, respective means, areas, and heterogeneity values defined as the coefficient of variation (CV) of sV̇ were calculated for well- and poorly ventilated lung zones for each slice and condition separately.
To account for small differences in minute ventilation between animals, we used a normalization term, V̇e /V̇em, where V̇e is the minute ventilation and V̇em is the average minute ventilation of the whole group, during each image slice acquisition (28). The sV̇ data presented in Table 2 and Fig. 3 were normalized in this fashion. Also, the fraction of minute ventilation going to the total of the three slices as well as to the well- and poorly ventilated zones, termed V̇/V̇e, was calculated (Table 2), where V̇ is the sum of absolute ventilation in all three slices: (4) where Vslice is the volume of imaged lung, computed as the lung surface area times the image slice thickness of 0.7 mm.
The cross-sectional area of central airways was measured by fitting an ellipse to the observed airway cross sections using monochromatic images. Because the airways other than the trachea were not always perpendicular to the imaging axis, the small axis of the ellipse was measured as the inner radius of the airway in the middle and caudal slices and used for the calculation of airway surface area (3).
We used two-way repeated-measures ANOVA with the Student-Newman-Keuls pairwise multiple comparison procedure to test for differences in airway cross section and specific ventilation parameters between experimental states and dose groups. Student's paired t-test was used for the comparison of blood gas, minute ventilation, and respiratory system resistance data. P < 0.05 was considered significant.
Minute ventilation, tidal volume, tracheal pressure, and arterial blood gas data are summarized in Table 1 together with the respiratory system resistance. Average time after histamine aerosol administration was 14 ± 3 min in the early set of image data (posthistamine 1) and 31 ± 6 in the late set (posthistamine 2). Histamine aerosol administration caused a significant rise in R at both 50 and 125 mg/ml. With the higher histamine dose, the rise in R was accompanied by a significant drop in arterial Po2 and combined respiratory and metabolic acidosis, the respiratory component of which partially recovered in the later set of image data (posthistamine 2).
Examples of sV̇ images in one animal are shown in Fig. 2. Histamine inhalation caused mainly clustered defects and a broadened regional ventilation distribution, based on visual examination of the ventilation images and histograms. At baseline, specific ventilation averaged over the whole ventilation image (sV̇m) in the middle and caudal slices was higher than in the apical image slice in both dose groups. There also was a positive ventral-to-dorsal gradient of sV̇m in the caudal slice. Both of these observations are consistent with our previous findings (28). After histamine inhalation, sV̇m significantly increased compared with baseline and remained higher when the later set of images was acquired. Figure 3A shows the relative change in sV̇m averaged over all three image slices at posthistamine 1 and posthistamine 2. When the well-ventilated lung zones were considered separately from the poorly ventilated zones (Table 2), it appeared that the overall rise in sV̇m was due to the relative increase in the former zones only (Fig. 3B). This rise was not different between the two dose groups. In the remaining zones, defined as poorly ventilated, sV̇m dropped compared with the overall sV̇m at baseline in both posthistamine 1 and posthistamine 2 (Fig. 3C). When sV̇m was calculated at baseline within the same regions as those that became poorly ventilated at posthistamine 1, these areas were found to be hyperventilated compared with the regions outside. The ratio of sV̇m within to the sV̇m outside of these zones was 1.09 ± 0.14 (range: 0.71 to 1.44), which was significantly larger than unity (P < 0.005). However, correlation between this ratio at baseline and the relative drop in sV̇ within these zones at posthistamine 1 was not significant (r = −0.35, P = 0.051).
Area of poorly ventilated lung zones.
At baseline, ∼6% of the total area of ventilated zones fell into the poorly ventilated category (Table 2) without significant difference between the 50 and 125 mg/ml groups. After histamine aerosol administration, this proportion significantly increased (Fig. 4) in both dose groups, but the rise was about twice as large with the larger histamine dose at posthistamine 1. The percentage of poorly ventilated zones remained significantly higher than baseline at posthistamine 2 and was not significantly different between the dose groups.
Estimated as the CV of sV̇, the heterogeneity of the sV̇ image significantly increased at posthistamine 1 (Fig. 5A and Table 2). The rise in heterogeneity was significantly larger in the 125 mg/ml group than in the 50 mg/ml group at posthistamine 1. Estimation of heterogeneity within the well-ventilated and poorly ventilated regions showed that the increase in ventilation heterogeneity was due to the constricted lung zones only (Fig. 5, B and C, and Table 2) and that in these zones, ventilation remained significantly more heterogeneous than baseline at posthistamine 2. On the contrary, ventilation became more even in the well-ventilated zones following histamine administration in both dose groups without a significant difference.
Central airway cross section.
Central airway cross-sectional data are summarized in Table 3. Studied airway cross-sectional areas at baseline varied from 16.5 to 3.3 mm2, corresponding to diameters from 4.6 to 2.1 mm and to airway generations 1 to about 7 (29). The trachea was imaged in the apical slice in all animals. As the image plane was lowered in the caudal direction, central airway cross section rapidly decreased given the monopodial morphology of the bronchial tree in rabbit, except in the tracheal slice. Central airways of different caliber could be analyzed, and the variations in their cross section in response to histamine were compared. Figure 6 illustrates the relation between individual airway sizes at baseline and at posthistamine 1. After histamine inhalation, central airway response was heterogeneous. Both constriction and dilatation were observed, with the probability of constriction increasing in central airways imaged in the caudal slice. The average responses in central airways in each image slice are summarized in Table 3.
The objective of this investigation was to measure response to histamine aerosol administration simultaneously in central and small peripheral airways by using SRCT imaging. The major finding of this study is that histamine aerosol causes a heterogeneous airway response combining narrowing and dilatation in individual airways, with the probability for constriction increasing peripherally. This finding provides further in vivo evidence that airway reactivity in response to inhaled histamine is complex and that the distribution of airway response within the bronchial tree may vary substantially with caliber.
In this study, we found that histamine inhalation caused mainly clustered defects and a broadened distribution of regional ventilation. These observations are consistent with previous observations in both human and animal models made using different imaging techniques. Vidal Melo et al. (40), using positron emission tomography (PET) in sheep, found that after methacholine-induced bronchoconstriction, areas with poor regional ventilation were topographically distributed in large and contiguous regions of the lung, which were adjacent to regions with nearly complete washout. In the present study, however, the adjacent well-ventilated regions had increased specific ventilation with a narrower distribution, which suggests redistribution of ventilation to these zones. These were the findings despite the fact that overall minute ventilation was slightly lower following histamine administration, particularly within 15 min after the administration of histamine (posthistamine 1) interval. Whether this redistribution was solely due to airway closure or near closure in the poorly ventilated zones or whether peripheral airway dilatation within the adjacent well-ventilated zones is involved, as well, cannot be quantitatively assessed on the basis of the present data, since lung sampling was partial and based on three image slices. Further study requiring three-dimensional (3-D) assessment of regional ventilation with KES imaging is needed to assess this point.
With the use of PET alone, airway dimensions cannot be measured simultaneously to regional ventilation imaging; it is therefore difficult to determine which airway generation involvement explains the observed topographic distribution of poorly ventilated zones. Tgavalekos et al. (37) used a combination of PET imaging of regional ventilation in asthmatic patients provoked with inhaled methacholine and oscillatory mechanical measures, with a 3-D airway tree model to examine the role of large vs. small airways in the simultaneous deterioration of mechanical function and ventilation. They found that matching both the degree of mechanical impairment and the size and location of the PET ventilation defects requires either constriction of airways <2.4 mm alone or a simultaneous constriction of small and large airways, but not just large airways alone.
In this study, we found that the response in larger caliber central airways was highly heterogeneous, combining constriction or dilatation of individual airways. This was particularly obvious in the middle image slice (Fig. 6). The probability for constriction increased in caudal central airways and in the peripheral airways, causing alterations in specific ventilation. Although with KES imaging the small peripheral airways could not be directly visualized, when combining both lung mechanics data and imaging of central conducting airways, we (3) previously observed that peripheral defects in ventilation might be present immediately after histamine inhalation, without any constriction in central conducting airways. With time following histamine administration, constriction in central airways appeared progressively with a larger magnitude in caudal central airways. We could not measure both regional ventilation and its kinetics in the present study due to time constraints; however, the combined constriction of individual small-caliber central airways and peripheral airways is consistent with our previous data. In that study, however, the central airways were imaged upon inspiration at 70% total lung capacity, which may have prevented the observation of central airway dilatation, and the selected image planes were in a lower vertical position with reference to the lung apex compared with the axial slices in the present study.
Paradoxical individual airway dilatations in response to histamine, a spasmogen, are difficult to explain. It should be pointed out that in the present study, all images were acquired at end expiration at atmospheric tracheal pressure; therefore, the observed changes in airway cross section cannot be simply due to variations in transbronchial pressure induced by the experimental setup. In vivo data on regional central airway response to exogenous stimuli are limited. However, paradoxical airway response has previously been demonstrated in dog, using tantalum bronchography (32). In this model, the authors found that although methacholine caused progressively increased airway narrowing from the trachea to small distal airways, PGF2α caused airway dilatation in airway generations 1 to 4. In contrast, airway constriction was observed in generations 5 and 6. On the other hand, Brown et al. (4), in their study of individual canine airway responsiveness to aerosolized histamine and methacholine in dog, did not find a paradoxical response in central airways >10 mm, suggesting that such complex airway behavior in response to histamine might be species dependent. In healthy human subjects, Brown et al. (5) found that some airways showed a paradoxical dilatation in response to inhaled methacholine, although it is not clear whether this paradoxical response was more frequent in larger conducting airways. A more recent high-resolution CT study, Kotaru et al. (21) demonstrated airway dilatation in asthmatic subjects in response to methacholine provocation can be complex, with the probability of constriction decreasing proximally in the trachea and main bronchi, where the average response was a small dilatation. Conversely, the frequency and magnitude of constriction increased in lobar and segmental bronchi, similar to the findings of this study.
One hypothesis to explain differences between central and peripheral airway responses to inhaled histamine is the role of the density distribution of histamine receptor subtypes in the rabbit bronchial tree. In vitro studies with antagonists and agonists indicate that histamine can produce relaxation of rabbit trachea by activation of histamine H2 receptors (19, 7, 17). A similar atypical relaxant effect of histamine has been observed in vitro in precontracted cat trachea (7). However, differences in histamine receptor density distribution can hardly explain the paradoxical dilatation of individual central airways within a given bronchial generation.
Whether paradoxical airway dilatations may be due to a compensatory reflex response is not clear. In their investigation of individual airway responsiveness to aerosolized histamine and methacholine using high-resolution CT, Brown et al. (4) found a substantial degree of baseline tone in canine airways. In their study, individual airways had a mean size of ∼50% of their fully atropine-relaxed state. In the present study, the animals were not vagotomized, and changes in the distribution of airway tone following histamine administration may play a significant role, since a reduction in airway tone in large-caliber central airways could potentially explain a paradoxical dilatation. Interestingly, central stimulation via an H1 histamine receptor-mediated mechanism reduces tracheal tone in rabbit (16). Vagotomy produces a small but significant increase in lung conductance in anesthetized and paralyzed rabbit. In this species, vagal constrictor fibers to the lungs act chiefly on the larger conducting airways (18), although baseline airway tone is known to be less prominent in rabbit than in dog (27). In this study, we found that by increasing the dose of inhaled histamine, it is essentially the area of poorly ventilated zones, the drop in specific ventilation as well as the heterogeneity of ventilation that increases within these zones. We did not find a clear relation between histamine dose and the complex central airway behavior. This may suggest that such central airway behavior may not be a direct consequence of histamine action on central airways, but rather the result of compensatory changes in central airway tone. Nevertheless, it is again difficult to explain how heterogeneity of individual airway response can be induced by reflex changes in central airway tone.
The mechanisms involved in the topographical heterogeneity of peripheral ventilation following histamine inhalation are not known. We found a slight but significant hyperventilation at baseline in the lung zones that became poorly ventilated following histamine administration. This finding suggests that increased airway constriction in these zones is likely to be due to a larger convective delivery of histamine aerosol particles. A similar observation has been made using PET measurements in sheep (39), where the ratio of ventilation in tracer-retaining regions to ventilation outside of these regions at baseline showed a strong correlation with the degree of hypoventilation resulting from methacholine aerosol administration. The lack of a significant statistical correlation between the drop in regional ventilation within the poorly ventilated zones and the ratio of sV̇m within to the sV̇m outside of poorly ventilated zones at baseline in this study was possibly due to partial sampling of the lung.
Anafi and Wilson (1) studied the heterogeneity of peripheral airway constriction by using a computational model describing a feedback between flow and airway resistance mediated by parenchymal interdependence and the mechanics of activated smooth muscle. Constricted terminal airways were predicted to have two stable states: one effectively open and one nearly closed. The authors argued that the heterogeneity of whole lung constriction is a consequence of this behavior. Airways are partitioned between the two states in constricted lung to accommodate total flow. Our finding that increased histamine dose caused an increase in both the area of poorly ventilated zones and the heterogeneity of ventilation within these zones may be a consequence of increased airway smooth muscle activation in peripheral airways, increasing the probability of near closure in a given individual airway. Venegas et al. (38), using an integrative mathematical model, have shown that even in a symmetrical airway tree subjected to uniform airway activation, severe ventilation heterogeneity with large contiguous areas of poor ventilation can appear. Factors such as the heterogeneous distribution of the bronchoconstricting agonist receptors or uneven distribution of the constricting agonist itself, structural differences between airways of different sizes, and differences in smooth muscle reactivity between individual airways all may be theoretically involved, as well. The asymmetry of the bronchial branching pattern tends to increase the heterogeneity of acinar particle deposition regardless of breathing pattern, based on computational modeling data (14). The degree of asymmetry is presumably larger in monopodial than in dichotomous airway trees (15) as in rabbit and may have contributed to a less homogenous distal aerosol deposition leading to heterogeneous peripheral airway constriction.
Mechanisms other than isolated airway functional or structural properties may contribute to complex airway behavior. Using a mathematical model that integrates dynamic interactions among serial and parallel pathways, such as airway-parenchymal interdependence and tidal expansion, Winkler et al. (41) recently demonstrated that the apparently paradoxical individual airway behavior combining constriction and dilatation in response to an agonist can be reproduced in silico. The authors found a combination of constriction and dilatation in all airway generations in their model, but the degree of narrowing was much larger in peripheral than in central airways. Also, the change of the average airway size after the emergence of ventilation defects was very small in central airways, whereas the size of the ventilation defect dramatically changed with increasing smooth muscle activation (41). The explanation proposed by the authors is that airway/parenchymal interdependence applies not only to local interactions but also to regional interactions along central-to-peripheral and interregional pathways. These model predictions are in agreement with the present experimental data that confirm the possibility that paradoxical individual central airway responses can coexist with severe patchy peripheral airway obstruction in response to an inhaled constricting agonist.
Technical advantages and shortcomings of the KES methodology have been extensively discussed elsewhere (2, 28). Xenon is a dense gas, which may have an effect on the xenon wash-in time constants and therefore on the calculations of sV̇ (8). We used inhaled isoflurane to maintain anesthesia. Isoflurane and other inhalational anesthetics seem to protect against induced bronchoconstriction and may affect baseline airway tone (22). Since isoflurane was administered at the same inhaled concentration from the beginning and throughout each experiment, we are assuming that comparison with baseline airway size accounts for any effect of the anesthetic. Because of the time necessary for obtaining images of regional ventilation, the number of the assessed lung slices was limited to three, and data acquisition was performed at two time intervals following histamine administration. We are assuming that the regional ventilation data from these three slices represent the lung as a whole, which is a known sampling bias. The three slices were imaged in the same order in all animals at both time intervals, and for the same slice, data were acquired at equal time intervals between animals. The recovery from the effects of histamine may be rapid, particularly in the early phase, and this may cause systematic within-slice differences; however, changes in lung mechanics during data acquisition were small during the posthistamine 1 interval, and the effect of histamine was quasi-stable in the posthistamine 2 interval. Despite this, we avoided between-slice comparisons after histamine administration. The cutoff defining poorly ventilated lung zones was determined independently for each image slice. This choice was motivated by the fact that using a fixed cutoff based on the baseline distribution of sV̇ led in many cases to the underestimation of the area of poorly ventilated zones. This was due to the fact that sV̇ increased in the imaged slices compared with baseline following histamine inhalation, whereas examination of the sV̇ images revealed the presence of patchy zones of clearly decreased specific ventilation compared with the rest of the image. A likely explanation is that is that following histamine inhalation, areas of poor specific ventilation due to airway closures or near closures exist outside of the imaged slices, which cause redistribution of ventilation, thereby increasing the average sV̇ to the imaged slice on the absolute scale, even though areas within the slice remained clearly less ventilated than the rest of the sV̇ image. Therefore, we used a statistical definition of poorly ventilated zones, which accounts for such changes in the overall sV̇ of the slice. Reliable analysis of central airway diameter requires that airways be located, matched, and measured at the same anatomic location. Changes in the image due to displacement of the animal versus the imaging plane are a potential source of error. In the present study, changes in lung surface area, indicative of changes in lung volume following histamine administration, were small and statistically nonsignificant. The reproducibility of central airway cross section measurements and the quality of anatomic matching were verified in a separate set of five animals at baseline and following methacholine inhalation for a different study, where 30 contiguous image slices were acquired at the midthoracic level with a 0.5-mm vertical step. Airway cross section measurements were found to be reproducible, and no significant bias due to image registration artifact was evidenced.
In summary, we have used a novel imaging technique to measure both central conducting airway response and the changes in regional specific ventilation simultaneously in response to a constricting agonist. We found that histamine aerosol causes a heterogeneous airway response combining narrowing and dilatation in individual airways, with the probability for constriction increasing peripherally. This finding provides further in vivo evidence that airway reactivity in response to inhaled histamine is complex and that airway response may vary substantially with location within the bronchial tree. We found that increasing the dose of administered histamine increased the area of poorly ventilated zones as well as the heterogeneity of ventilation. However, no clear relation between inhaled histamine dose and central airway response was found. These findings provide further insight into the local differences in airway response, which is significant in advancing the current understanding of asthma pathophysiology and inhaled drug delivery in this disease.
This work was supported in part by Academy of Finland Projects 50656 and 126747, the Tampere Tuberculosis foundation, the Instrumentarium Foundation for Science, the Ida Montin Foundation, and the European Synchrotron Radiation Facility.
We gratefully acknowledge Christian Nemoz, Gilles Berruyer, Thierry Brochard, Sylvie Monfraix, and Dominique Dallery for technical assistance.
- Copyright © 2009 the American Physiological Society