INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Inhalational challenges are commonly used to measure airway reactivity in asthmatic and normal subjects. The two agents commonly used for such challenges, methacholine and histamine, are thought to have their primary effects on different locations of the airway tree. It is generally believed that the major effect of cholinergic agonists is in larger airways and that of histamine is in smaller airways. However, the evidence for this belief is quite indirect, arising by extrapolation from both in vitro (1) and in vivo studies (2, 3). In vivo evidence has often required the assumption that changes in dead space reflect changes in large airways while changes in respiratory resistance (with no change in dead space) reflect changes in small airways (2, 3). Other indirect methods using relative changes in dynamic compliance and lung resistance have also been used to assess the relative contribution of large and small airways to agonist challenge (4). The conclusions from such studies are based on the assumption that changes in lung compliance or lung tissue resistance result entirely from constriction of alveolar ducts or other peripheral smooth muscle. However, this assumption has been invalidated by more recent studies that have clearly demonstrated that major changes in lung compliance can occur with constriction limited to the conducting airways (2, 7). Other more direct methods using radiographs of tantalum dusted airways have been used to study the serial distribution of vagal (8) and Mch stimulation (9). Both of these stimulations were shown to cause a greater muscle contraction in the smaller airways, thus challenging the notion that Mch has a preferential effect in the large airways. More recently, using CT imaging of airways in human subjects, Okazawa and coworkers (10) have similarly demonstrated that the responsiveness to Mch challenge is greater in small airways than in large airways.

In this study we have systematically investigated the serial distribution of airway responsiveness following stimulation with histamine or methacholine by directly measuring airway area in vivo with high resolution computed tomography (HRCT). Our results show similar responsiveness to each agonist over all airway sizes measured, with a slightly increased responsiveness to Mch compared to histamine throughout the airway tree.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our study protocol was approved by the Johns Hopkins Animal Care and Use Committee. Ten studies were performed on five dogs weighing approximately 20 kg. The dogs were anesthetized with thiopental (15 mg/kg induction dose followed by 10 mg/kg/h intravenous maintenance dose). After induction of anesthesia, the dogs were paralyzed with 0.5 mg/kg of succinylcholine with occasional supplemental doses as required to ensure no respiratory motion during imaging. Following tracheal intubation with a 8.5-mm ID endotracheal tube, the dogs were placed supine and their lungs were ventilated with 100% oxygen with a volume-cycled ventilator (Harvard Apparatus, Millus, MA) at a tidal volume of 15 ml/kg and a rate of 18 breaths/min. A stable depth of anesthesia was maintained by monitoring heart rate changes and eyelash reflex. Dynamic respiratory system compliance was measured as the tidal volume divided by the change in airway pressure during normal ventilation at functional residual capacity (FRC).

Imaging of Airways

High resolution computed tomography (HRCT) scans were obtained with a Somatom Plus Scanner (Siemens, Iselin, NJ) using a 1-s scan time, 137 kVp, and 220 mA as previously described (11). The images were reconstructed as a 256 × 256 matrix using a maximum zoom of 4.0 (12 cm field of view). Twenty-five to 50 contiguous sections were obtained, starting at the carina and moving caudally using a 1-mm table feed and a 2-mm slice thickness. The dogs were apneic with a constant controlled airway pressure for the duration of the scans (approximately 2 min).

Images were reconstructed with the use of a high spatial frequency (resolution) algorithm that enhanced edge detection, at a window level of 450 Hounsfield units (HU) and a window width of 1,350 HU. All airways visualized approximately perpendicular to the scan plane (long to short axis ratio < 1.5:1) were measured. For repeated airway measurements in a given dog within each experimental protocol and across experiments on different days, adjacent anatomic landmarks, such as airway or vascular branching points, were defined and the airways were matched by these adjacent landmarks and measured.

Measurement of Lung Volume

Standard resolution computed tomography scans were obtained with a Somatom Plus Scanner (Siemens) using a 1-s scan time, 137 kVp, and 210 mA. The images were reconstructed as a 256 × 256 matrix using a maximum zoom of 2.0. Twenty-five to 50 contiguous sections were obtained, starting at the apex of the lungs and moving caudally to the bases using an 8-mm table feed and an 8-mm slice thickness. Images were reconstructed with the use of a standard lung algorithm at the same window settings described above, and the area of the lung on each CT scan was defined as the area within the pleural border excluding the heart and diaphragm. The total lung volume was then calculated as the sum of the lung areas of each CT slice times the slice thickness. Measurements of lung volume were made at control, and after 1, 50, and 500 mg/ml concentrations of agonist challenge.

Image Analysis

The HRCT images were analyzed using the airway analysis module of the Volumetric Image and Display Analysis (VIDA) image analysis software package (Department of Radiology, Division of Physiologic Imaging, University of Iowa, Iowa City, IA) as previously described and validated (12, 13).

We also attempted to measure airway wall area in each airway. Because adjacent vascular structures make the outer boundary of airways difficult to define continuously, we first measured mean wall thickness in each airway. This was done by drawing three lines through the airway wall. The program automatically displayed a histogram of the pixel intensity along each line. The inflection points of increased intensity along the line that represent the inner and outer edges of the airway wall were selected. These points were identified as those that changed approximately 10% of the intensity range between the center of the wall and the lumen or parenchyma, respectively. Since the software we use is capable of measuring fractions of a pixel when the rays are drawn oblique to the side of a pixel, the measurements of wall thickness were not necessarily quantized to multiples of the pixel dimension. The three lines were averaged, and from the measured lumenal area and average wall thickness, we could then calculate the wall area and total airway area (Ao) (14) by simply assuming that the average measured wall thickness was uniform around the airway. The total airway area, comprising the area inside the outer airway perimeter, then equals: [T+(Ai/)]², where Ai = lumenal area and T = wall thickness. A few of the walls of the smallest airways could not be measured with this approach.

Protocol

Each dog served as its own control. The dogs were anesthetized and ventilated as described above and baseline scans were obtained. On separate days the dogs received cumulative inhalation challenges of either aerosol methacholine (Sigma Chemical, St. Louis, MO) or histamine (Sigma Chemical), in concentrations of 1, 10, 50, 100, and 500 mg/ml. Five breaths at each dose were administered to a peak airway pressure of 15 cm H₂O, held for 1 s, and then released to atmospheric pressure. Aerosol challenges were administered by a Hudson 3000 nebulizer (Hudson, Temecula, CA) driven by compressed oxygen at 10 L/ min. Under test parameter conditions with an operating pressure of 50 psi and an atomization flow rate of 10 L/min, the nebulizer produces particles of mass median diameter of 3.1 µm with a geometric standard deviation of 3.2. Given such a relatively disperse particle size distribution, good central and peripheral distribution should have been achieved (15). Approximately 1 ml of solution was administered per challenge. After completion of the 500 ml/mg scans, the dogs received 0.2 mg/kg atropine IV, a dose previously shown to completely block vagal tone in the dog (16), and the HRCT scans were repeated.

Analysis

The mean airway areas after atropine for the methacholine and histamine days were compared by paired t test. The completely relaxed airway after atropine was defined as 100% (relaxed state), and the airway lumenal area were expressed as a percent of the relaxed area. Because we found this fully relaxed airway area to be unchanged on the different experimental days, it thus serves as a reference on which to normalize the endogenous and exogenous airway contractions. Baseline airway area variability within dogs on the methacholine and histamine study days were analyzed by generalized analysis of variance, with airway area as a percent of maximum the dependent variable, and individual dogs, the multiple airways per dog, and the aerosol challenge the independent variables. Corrections were made for multiple pairwise comparisons of means.

Furthermore, to determine how airway responses varied with airway size, airways were divided into three size categories based on their size in the relaxed state: small (3-6 mm in diameter), medium (6-10 mm in diameter), and large (> 10 mm diameter).

Airway responses to methacholine and histamine were analyzed by generalized analysis of variance, with the percent of relaxed airway area the dependent variable, and individual dogs, the multiple airways per dog, and the aerosol challenge the independent variables. Corrections were made for multiple pairwise comparisons of means. The effect of airway size on the responsiveness was assessed by separately repeating the generalized analysis of variance for each size category.

Changes in airway wall area as a percent of relaxed airway wall area for each airway size category for Mch and histamine were separately analyzed by one way analysis of variance, with airway wall area (as a percent of relaxed airway area) the dependent variable and aerosol dose the independent variables. Corrections were made for multiple pairwise comparisons of means.

Changes in lung volume and compliance with methacholine and histamine were separately analyzed by two-way analysis of variance with lung volume the dependent variable and individual dogs and the aerosol challenge the independent variables and lung compliance the dependent variable and individual dogs and the aerosol challenge the independent variables; linear regression models were also analyzed for the lung volume and compliance responses to methacholine and histamine. The significance of the slopes of the regression lines being different from zero were tested. Significance was considered to be p < 0.05.

Based on preliminary data, a power calculation was performed. In each animal we are able to identify and measure at least 12 airways in cross-section. With an  = 0.05, this sample size was determined to provide > 99% power to detect a difference of 10% between the means of the airway caliber in five dogs after histamine compared to methacholine aerosol challenge.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In each dog, 13-14 airways (range 1.1 to 16.9 mm in diameter at baseline) were identified and measured under all conditions. Relaxed airway areas after atropine on the different methacholine and histamine challenge days did not differ significantly in each dog (p = 0.23). Therefore, the largest airway area measured after atropine administration on either day was defined as 100% (relaxed state) for all subsequent calculations.

Baseline airway size showed enormous variability compared with the relaxed condition within and between dogs (Table 1), varying between 5% and 100% of the maximally relaxed size in individual airways. The mean airway area as a percent of the relaxed state was slightly but significantly different on the methacholine day (50 ± 3%) compared with the histamine day (56 ± 3%), controlling for the individual dogs and for the multiple airway measurements per dog (p = 0.003). Figure 1 shows baseline size as a function of the relaxed airway area for all airways in each dog. Baseline airway size was significantly correlated with relaxed airway size (r = 0.57, p < 0.001), although it is clear from Figure 1 that there is considerable variability for any given range of airway sizes.

View larger version (16K): [in this window] [in a new window]   Figure 1.   The relationship between individual baseline airway size (as a percentage of maximum) and the relaxed size. The linear regression line shows the tendency for larger airways to be less contracted in the baseline state (r = 0.57, p < 0.001 for the slope different from zero).

The airway lumenal area (Ai) as a percent of maximum in each dog decreased with increasing dose after both methacholine and histamine aerosol challenges. Figure 2 shows the mean dose response curve averaged across all dogs. At all doses greater than 1 mg/ml, this mean area was slightly smaller after methacholine than after histamine (p < 0.001), with constriction at maximal dose being 15 ± 1% and 22 ± 1% of the relaxed airway size for methacholine and histamine, respectively. Similar results were also observed in each of the individual dogs. The mean response of airways separated into the three different size categories is shown in Figure 3. There were no significant differences in the degree of contraction between any of the three groups at any dose (p = 0.50). To assess further whether the extent of constriction varied with airway size, we plotted airway size at the highest histamine or methacholine dose (500 mg/ml) as a function of the relaxed size (Figure 4). Separate linear regression lines are drawn for histamine and methacholine. At this maximal dose there was no correlation between relaxed size and degree of constriction for either methacholine (p = 0.10) or histamine (p = 0.48). A similar lack of correlation with airway size was also found with each of the smaller doses (data not shown).

View larger version (12K): [in this window] [in a new window]   Figure 2.   Dose-response curves plotting mean airway lumenal area (Ai) as a percent of maximum during histamine (dashed line) and Mch (solid line) challenges averaged across all dogs. At all doses greater than 1 mg/ml, this mean area was slightly smaller after methacholine than histamine challenges (p < 0.001).

View larger version (16K): [in this window] [in a new window]   Figure 3.   Dose-response curves plotting mean airway lumenal area (Ai) as a percent of maximum during histamine (dashed line) and Mch (solid line) challenges averaged across all dogs, separated into three airway size categories.

View larger version (19K): [in this window] [in a new window]   Figure 4.   The relationship between individual airway lumenal size (as a percent of maximum) after challenge with 500 mg/ml of methacholine (open squares) or histamine (closed diamonds) and their fully relaxed size. There was no correlation between the degree of constriction and airway with either Mch (p = 0.10) or histamine (p = 0.48).

Figure 5 shows the individual baseline airway wall area fraction (of total airway area) as a function of airway size for all dogs. Data are presented from baseline measurements preceding the histamine (closed diamonds) and Mch (open squares) challenges. The mean wall area fraction varies from about 0.35 in the larger airways to about 0.85 in the smaller airways. Figure 6 shows the mean percentage changes in wall area with increasing doses of either Mch or histamine. The calculated wall area is seen to decrease by 40% at the highest doses of either agonist.

View larger version (20K): [in this window] [in a new window]   Figure 5.   Individual baseline airway wall area fraction (of total airway area) as a function of airway size for all dogs. Data are presented from baseline measurements preceding the Mch (open squares) and histamine (closed diamonds) challenges. See text for discussion of limitations regarding the quantitative accuracy of this measurement.

View larger version (16K): [in this window] [in a new window]   Figure 6.   Mean percentage changes in calculated wall area with increasing doses of either Mch (solid line) or histamine (dashed line). See text for discussion of limitations regarding the quantitative accuracy of this measurement.

Figure 7 shows dynamic compliance as a function of dose of Mch and histamine. As was observed with the airway area responses, there was a slightly greater effect of Mch at the highest doses (p < 0.003). Lung volume at FRC also decreased significantly (p < 0.001) in response to each increasing concentration of Mch and histamine, but these changes were not significantly different between the two agonists (p = 0.23). At doses of 1, 50, and 500 mg/ml lung volume decreased to 91 ± 2%, 79 ± 3%, and 74 ± 1% for Mch and 94 ± 1%, 84 ± 3%, and 76 ± 1% for histamine, respectively.

View larger version (15K): [in this window] [in a new window]   Figure 7.   Dose-response curves plotting mean dynamic compliance during histamine (dashed line) and Mch (solid line) challenges. There was a slightly greater effect of Mch at the highest doses.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our results show that individual canine conducting airways (greater than 1.1 mm in diameter at baseline) show a slightly greater contraction in response to methacholine compared with histamine. This result was independent of airway size over the range of airways studied. Similarly, the lung compliance also showed a slightly greater response to methacholine. If changes in compliance are assumed to reflect primarily contraction of smaller airways (less than 2 mm in diameter), these results thus demonstrate a fairly consistent and uniform response throughout the airway tree for each agonist. There was no indication that either agonist showed preferential sensitivity to central or peripheral airways. Although these results seem to conflict with the common notion that histamine acts peripherally and Mch (or vagal tone) more centrally, many studies either show no differences in responsiveness for these two agonists or results that can be interpreted that way.

Although the use of dynamic compliance to assess very peripheral airway constriction has been common (4), there is strong evidence that larger conducting airways can also have major effects on lung compliance (2, 7). Indeed, in a preparation where there was no constriction of any airways smaller than 0.5 mm, constriction of the conducting airways caused substantial changes in lung compliance (7). Although we believe that in the present study, the aerosol reached all airways, we have no direct evidence of its peripheral deposition. However, we could in fact explain our results even if there were no agonist deposition in the periphery. If the changes in compliance resulted from a conducting airway tree constriction that stiffened the lung, then we would expect to see changes in compliance mirroring the changes in airway constriction. This is precisely what we have seen, with the Mch causing a slightly greater response in both conducting airways and lung compliance. To the extent that there was some constriction in the most peripheral airways, this should have increased the compliance responses.

These experimental findings contrast with studies in dogs and humans (2, 3) that have assessed the small airway contribution from the measurement of resistance and the large airway response from that of dead space. In both these studies, it was concluded that histamine has its predominant effect on small airways, whereas methacholine (or vagal stimulation) has its predominant effect on larger airways. However, this indirect method of estimating airway area from measurements of dead space has several interpretative problems that are well discussed in those papers. Since both airway resistance and dead space must both be related to airway size, separation of the two effects requires several assumptions.

A vagal reflex component to the airway response to histamine has been suggested by several studies, but the literature is not consistent on this matter. DeKock and coworkers (17) clearly demonstrated that the response to histamine could be blocked by vagotomy, whereas Loring and colleagues showed no effect of vagal cooling or vagotomy (18). In our present study, we clearly found a substantial degree of baseline tone that could be abolished by atropine, but we did not study the effect of atropine on response to histamine. There are several relevant studies directly comparing histamine and Mch responsiveness in normal subjects. Sterk and colleagues (19) showed a similar degree of airway responsiveness in human subjects with the same concentrations of either Mch or histamine aerosol. In their study, they only measured spirometric indices, so they could not comment on the serial distribution of airway responsiveness. Robatto and associates (5) also found no differences in airway resistance, tissue resistance (an indirect measure of peripheral airways), or lung compliance, in response to histamine or Mch. As with our present study, their animals were not atropinized. Perhaps the inability to demonstrate slightly greater sensitivity to Mch in these two studies reflects the gross averaging inherent in global measurements of pulmonary function. There are also discrepancies with findings in several other studies related to the serial distribution of responsiveness. Radiographic imaging in dogs (8, 9, 20) and CT imaging on humans (10) have reported that Mch challenge causes increasing responsiveness in the smaller airways at least down to 1 mm. Although we did not find this slightly increased responsiveness in the smaller airways, neither result supports the simple paradigm of histamine affecting small airways and Mch affecting large airways, since the responsiveness to Mch is never seen to be decreasing in the smaller airways.

In our present study we normalized each airway response to the size measured in the completely relaxed state (at the end of the experiment) after a dose of intravenous atropine previously shown to completely block vagal tone (21, 22). This was essential to do because, as was found in the present and other studies (23), canine airways normally have a substantial degree of baseline tone. In the current study, prior to giving atropine, individual airways at baseline ranged in size from 5% to 100% of the relaxed state, with a mean size of about 50% of the maximally relaxed size (Table 1). This variability in baseline size among airways within an animal's lungs is also consistent with previous work in dogs, where individual airways were studied over the course of 1 yr (24). In the present study the variability also manifested itself by finding that the mean airway area at baseline was slightly but significantly different on the methacholine and the histamine challenge days even within an individual dog. In the conventional measurements of airway responsiveness, this normal baseline airway tone of the airway must play a major role. To determine the possible effect of this baseline tone on airway responsiveness, we divided the airway into two groups: "high tone" and "low tone" groups. The high tone group included airways with lumenal area less than 50% of the relaxed state. For the high tone airways, the mean airway areas were 30 ± 1% and 35 ± 2% on the methacholine and histamine days, respectively (p = 0.11). For the low tone airways, the mean airway areas were 78 ± 3% and 79 ± 3% on the methacholine and histamine days, respectively (p = 0.85). Then we reanalyzed the dose response curves separately for the high and the low tone groups of airways. We found that methacholine caused a slightly greater constriction in both the initial high tone and initial low tone groups, indicating that initial baseline tone was not responsible for the increased sensitivity to Mch.

Although our present data show that the airway can be decreased to about 20% of its maximal size at the largest doses of the two agonists used, measurements of airway reactivity measured from the baseline condition without atropine would show a maximal area constriction to only about 40% of baseline. This level of airway constriction (to 40% of baseline area) would be expected to increase the airway resistance (Raw) about sixfold. This is in the range of constriction at maximal doses generally reported in several species, including man (25).

We also found that the dose-response curves never showed a flat plateau, so that the level of maximal airway constriction cannot be determined from this study. In our protocol we gave the highest concentration of agonist that we could maintain in solution to nebulize. This concentration of 500 mg/ml is an order of magnitude greater than the largest doses normally given to human subjects. There are several studies of in vivo airway responsiveness in animal models that present variable results, particularly whether the dose response curve shows a plateau at maximal dose. In studies in dogs, the presence of a maximal plateau depends on the variable being measured and the experimental protocol (5, 9, 26, 27, 30). The lack of a plateau in the present study suggests that were we able to deliver higher doses, we may have observed airway closure. The slightly greater airway narrowing observed with Mch compared with histamine thus could simply result from an increased sensitivity to Mch.

Lung volume at baseline (FRC) was not different on the methacholine and histamine challenge days, nor were the changes in lung volume to methacholine and histamine challenges. Decreases in lung volume with agonist challenge in dogs are commonly reported in the literature (2). However, the decrease in lung volume was small in comparison to the decrease in airway area at the same dose. On average, airway area decreased by 22% after 1 mg/ml of methacholine and histamine, but the decrease in lung volume was only 7%. After 500 mg/ml of agonist agent, lung volume was 75% of control, while airway area was only 18% of the relaxed state. Since the variation of airway area with lung volume can be modeled as (VL)^2/3 (11), the observed decrease in lung volume could cause a maximum decrease in airway area to about 83% of control (i.e., a 17% decrease). Thus the approximately 25% decrease in lung volume at the maximal doses is clearly not large enough to account for the approximately 80% decrease in airway area.

Finally, the measurement and results of airway wall area warrant some discussion. We were at first surprised by our findings of a progressive decrease in airway wall area with increasing doses of both histamine and methacholine (Figure 6). The only physiologic mechanism that could possible account for this wall thinning is that smooth muscle constriction progressively squeezes out blood volume or interstitial fluid from the airway wall. However, this seemed unlikely because even with extreme engorgement the total amount of blood volume in the airway wall comprises less than 25% of the wall area (31), and it is only the small fraction of that vascular volume interior tot he contracting smooth muscle that would be under compression. In addition, although little is known about the amount of free interstitial water in the airways, it likewise seems very unlikely that the amount inside the smooth muscle could comprise 40% of the airway wall. Thus, to explain this observation we need to consider the possibility of a systematic error in the measurement of airway wall thickness. This problem of measurement error in small airways has been evaluated in a recent study by Wood and coworkers (32). In this work it was shown that as a solid structure gets smaller, CT measurement of the distance across that structure leads to a progressive overestimation. The magnitude of this overestimation could be as much as 100% for a true distance of 1 mm. Since the normal wall thickness of a relaxed 5 mm human airway is on the order of 0.5 mm (33), this overestimation cannot be ignored. Histologic measurements of the airway wall area in normal humans as a fraction of total airway area range from 10-20%, and up to 30% in asthmatics (33). In noncartilagenous airways of sheep, the fractional wall area is about 30% at FRC (31). The values obtained by CT in our study range from 30 to 90%, and this range is comparable to that found in other CT studies (10, 34). Thus, is it clear that the CT estimates of airway wall thickness of large airways can be 2-3 times greater than the actual wall thickness, and in small airways the estimates can be 3-5 times greater. The situation in a contracted airway is more complex, because airway contraction to a smaller airway size necessarily leads to a thickening of the wall; thus, the overestimation error in the measurement of airway wall thickness will decrease with contraction. Since our calculation of wall area is based on the wall thickness measurement, this effect could easily lead to the apparent decrease in wall area observed shown in Figure 5. We believe this measurement artifact also accounts for the similar finding by McNamara and associates (34) of a decrease in wall area following methacholine challenge in isolated dog lungs and by Okazawa and associates in normal humans (10). This substantial overestimation of wall thickness is inherent in the CT technology, and therefore quantitative measurements of airway wall thickness or area with this method must be interpreted cautiously.

In summary, we have shown that in the intact dog, there was a slightly greater responsiveness to Mch in all conducting airways that can be accurately measured with HRCT (i.e., those greater than  2 mm). This greater responsiveness was also present in those airways that are responsible for the decreased Cdyn seen with Mch challenge (i.e., those less than  2 mm). We thus conclude that in vivo aerosol challenges with histamine or methacholine do not lead to preferential constriction of large or small airways.