INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The mechanisms that cause airflow limitation in exercise- or hyperpnea-induced bronchospasm (HIB) are poorly understood. Current evidence indicates that the stimulus of airway cooling and/or drying is translated into airflow limitation via the generation of bronchoactive mediators, neural reflexes or hyperemia and edema that result in a net narrowing of the airway (1, 2). In addition, airway instability may occur, which could lead to airway closure and derecruitment of some airways, processes that would be expected to contribute to the gas trapping and hyperinflation observed in HIB (3). Thus, airflow limitation in HIB may be due to components of both airway narrowing and derecruitment.

We have previously demonstrated that peripheral airway resistance (Rp) rises more in asthmatics than in normal subjects after segmental challenge with cool, dry air using a wedged bronchoscope technique (4). Insofar as this challenge model mimics events occurring in the lung periphery during HIB, our data directly implicate changes in the peripheral small airways in the pathogenesis of HIB. The current study extends our previous findings by measuring the pressure decay within the segment after the cessation of airflow, similar to the analysis in dogs made by Smith and colleagues (5). These investigators found that the pressure decay within the wedged segment was characterized by an initial, abrupt fall in pressure, followed by a more gradual decline in pressure. The initial fall was assumed to be due to resistance in the airways immediately distal to the bronchoscope but proximal to the collateral pathways, whereas the more gradual fall was thought to be due to the resistance of the collateral pathways. The purpose of the current study was to determine, first, whether pressure decay within the wedged segment in humans behaves similarly and, second, how asthmatic subjects with HIB might be different from normal subjects. Specifically, we sought to understand better the mechanisms involved in generating the higher Rp of asthmatics at baseline compared with normal subjects, and the increase in Rp in asthmatics with HIB after cool, dry air stimulation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Seven asthmatic subjects with documented HIB and seven normal subjects without asthma gave informed consent and were studied in accordance with the guidelines of the Institutional Review Board. The current study included five of the original asthmatic subjects and six of the original normal subjects studied previously (4). All subjects were characterized by clinical history, physical examination, spirometry, and challenges with methacholine, exercise, and hyperpnea, as previously described (4).

Bronchoscopy with Cool, Dry Air Challenge

The protocol for bronchoscopy with segmental cool, dry air challenge has been previously described in detail (4). In brief, after standard preparation for bronchoscopy, an esophageal balloon was passed via one naris for monitoring of esophageal pressure in order to confirm that all measurements of Rp in any given subject were made at equivalent lung volume (FRC). Then, after placement of the bronchoscope in a wedge position in the anterior segment of the right lower lobe, a 5 Fr. double-lumen catheter (Baxter Healthcare Corp., Irvine, CA) was snugly inserted via the instrument channel of the bronchoscope and positioned just beyond the bronchoscope tip. We passed fully saturated 5% CO₂/air through one lumen of the catheter, regulating flow with a mass flowmeter (Sierra Instruments, Inc., Monterey, CA) and temperature and humidity with a heated respiratory humidifier (Bird Life Design, Carrollton, TX; Fisher & Paykel, Auckland, New Zealand).

Pressure at the tip of the bronchoscope (Pb) was measured through the second lumen of the catheter using a differential pressure transducer (Validyne Engineering Corp., Northbridge, CA). Continuous recordings of Pb, esophageal pressure, and flow were made on a chart recorder (Gilson Medical Electronics, Middleton, WI) as well as captured digitally with a computer data acquisition program (AcqKnowledge; Biopac Systems Inc., Goleta, CA). Subjects were periodically instructed to stop breathing at FRC, which was confirmed by constant esophageal pressure. At that point, surrounding alveolar pressure was zero, and a steady-state pressure was achieved between Pb and segmental pressure (Ps) (6).

After each steady-state pressure measurement, airflow was stopped during the breathhold and the subsequent decline in pressure was measured (Figure 1). Each stop-flow maneuver lasted approximately 5 s, or until an apparent plateau occurred or the subject could no longer maintain a breathhold at FRC. This maneuver allowed us to analyze the pattern of decay in pressure, and thus partition Rp into the resistances of the large airways proximal to the wedged segment and the small airways serving as the pathways for collateral flow out of the segment (6). The final plateau pressure (Pp), or "back pressure," obtained under conditions of no flow, was also recorded and taken as reflective of closure or extreme narrowing of either the relatively large airways proximal to the wedged segment, or the small collateral pathways (6, 7).

View larger version (23K): [in this window] [in a new window]   Figure 1.   Method of stop-flow analysis. Pressure at the tip of the bronchoscope (Pb) is recorded continuously over time while a constant flow of 5% CO2/air is instilled into the wedged segment (hatched bar). During quiet breathing, a steady, rhythmic oscillation in pressure between inspiration and expiration is seen (solid line with arrowheads). The subject is instructed to breathhold at the end of quiet expiration (FRC). The value of Pb that stabilizes at this point (upper horizontal dashed line) is that used to calculate Rp. The flow of instilled air is then abruptly stopped (while the subject is still at FRC), and Pb decays as air exits the segment via collateral channels to a steady-state plateau pressure (Pp) (lower horizontal dashed line). The subject then resumes breathing.

The local, segmental challenge protocol consisted of three phases (Figure 2). In the Pressure-Flow Phase, baseline Rp was first measured with a flow rate of 100 ml/min of 37° C saturated 5% CO₂/air. These gas conditions were chosen to simulate the natural environment in the lung periphery (8). Increasing flows of gas from 100 to 1,000 ml/min were then administered in increments of 100 or 200 ml/ min, with Pb recorded at each level of flow after Pb had stabilized (usually within 30 to 60 s). Flow was stopped at each level to allow recording of the pressure decay and final plateau pressure. Two to three recordings of Pb and stop-flow were made at each flow rate over a period of 2 to 3 min, and the mean values of Pb and Pp were determined for each level of flow. After the highest flow rate achieved, which was either 1,000 ml/min or an arbitrary Pb limit of 40 cm H₂O (9), flow was returned to 100 ml/min to reestablish baseline Rp. The entire Pressure-Flow Phase took 10 to 20 min. The Pressure-Flow Phase was performed to determine the specific pressure-flow characteristics of each subject and to establish the maximal flow rate to use for each individual subject during the subsequent high-flow cool, dry air challenge.

View larger version (10K): [in this window] [in a new window]   Figure 2.   Study protocol. The three phases of the peripheral lung challenge are depicted. Stop-flow maneuvers were performed at each level of flow during the Pressure-Flow Phase, during the last minute of the Challenge Phase, and at 1, 5, and 10 min into the Recovery Phase. Warm, humid = 37° C, saturated; cool, dry = ~ 22° C, dry.

After the reestablishment of baseline Rp, the Challenge Phase was instituted by passing dry, tank-source 5% CO₂/air through 50 feet of copper tubing surrounded by frozen ethylene glycol; such gas exited the catheter in the wedged segment at a cool temperature of 22° C. The transition from baseline conditions (37° C, saturated gas at 100 ml/min) to the high flow, cool, dry challenge conditions took approximately 30 to 60 s. This cool, dry gas was then instilled for 5 min at the maximal flow rate determined previously in the Pressure-Flow Phase. Stop-flow maneuvers were performed in the last minute of the Challenge Phase. Immediately after the challenge, the Recovery Phase was initiated, as airflow was returned to baseline conditions (100 ml/ min; 37° C, saturated), and Rp was measured over the remaining 5 to 15 min. Stop-flow maneuvers were performed in duplicate at 1, 5, and 10 min into the Recovery Phase.

Data Analysis

Rp was calculated as Pb divided by the flow rate of the insufflating gas (6). Because all of the airways leading from the wedged bronchoscope, across the collateral pathways and back out the airways of the adjoining segments are exposed to steady-state airflow, Rp represents the combined flow resistance of these airways.

The Pb data during breathhold were analyzed according to the method of Smith and colleagues (5). First, we determined the resistance of the relatively large airways leading from the tip of the bronchoscope into the wedged segment. During the period of constant flow, the pressure Ps at the distal end of these airways is less than Pb by an amount equal to the resistive pressure drop across the airways. When flow is stopped this pressure drop is obliterated and Pb becomes immediately equal to Ps. Consequently, the resistive pressure drop Pb  Ps prior to stopping flow appears as an immediate drop in Pb when flow is stopped. To determine Ps, we fit a line to the logarithm of Pb between ~ 0.2 s and the point at which the logarithm of Pb began to visually deviate from a straight line (usually 0.4 to 1 s), (Figure 3), thereby restricting the analysis to the initial monoexponential decay of pressure. The point at which this line was back-extrapolated to cross the pressure axis at the instant flow was stopped was taken as Ps. The pressure drop across the airways leading to the distal lung segment was therefore the difference Pb  Ps. The ~ 0.2 s time point was chosen as the beginning of the line fit because any initial pressure drop Pb  Ps should have been completed by this time (5).

View larger version (20K): [in this window] [in a new window]   Figure 3.   Mathematical analysis of decay curve. On the left (A) is shown the initial decay in pressure at the tip of the bronchoscope (Pb) following the cessation of flow. On the right (B), the plot is log-transformed in the ordinate and the first linear portion of decay after the initial 0.2 s is identified (dashed vertical lines). The equation of the best- fit line through this region allows calculation of the Y-intercept at time zero, or the segmental pressure (Ps) prior to stop-flow. Pb  Ps is due to resistance across the airways between the wedged bronchoscope and the subtended segment. In this example, the equation of the line is logPb = 0.62T + 1.43. Measured Pb at time zero = 27.5 cm H2O, and calculated Ps at time zero is also 27.5 cm H2O. Thus, there was no measurable pressure drop across the airways proximal to the wedged segment. This latter finding was true in all subjects during all phases of the protocol. Rp is calculated as Pb/flow, and Cp is calculated as /Rp, where the time constant () is calculated from the initial linear portion of log Pb.

Next, we measured the time constant () of the first part of the Pb decay curve from the slope of its logarithm versus time. We used only this early part of the curve because the entire decay curve was clearly not monoexponential, presumably because of progressive increases in collateral resistance as lung volume decreased (10). We reasoned, however, that the early part of the curve would reflect the Rp determined from the prior constant flow as the segment emptied passively through the collateral pathways. The peripheral compliance of the segment (Cp) was then calculated as /Rp (7).

The descriptive factors of age, height, and percent predicted FEV₁ are reported as means ± SD, whereas measured changes in the outcome variable FEV₁ after exercise and hyperpnea are expressed as means ± SEM and medians with interquartile range (IQR), respectively. Fischer's exact test was used to compare categorical data. PC₂₀ values were log-transformed for statistical comparisons and are expressed as geometric mean and range. Comparisons of FEV₁ and PC₂₀ FEV₁ between normal subjects and asthmatics were made with unpaired t-tests, whereas changes in FEV₁ before and after exercise were made with paired t-tests, and changes in FEV₁ after HP were made using Wilcoxon's rank sum test. Values of Rp, Pp, Cp, and  are reported as least-squares means ± SEM, and the significance of differences between and within groups across time for these variables was determined by a repeated-measures ANOVA allowing for unequal compound symmetric variance matrices (11, 12). Correlations between Rp, Cp, Pp, and PC₂₀ FEV₁, the fall in FEV₁ after exercise (FEV₁-EX) and the fall in FEV₁ after hyperpnea (FEV₁-HP) in asthmatics were determined using Spearman's rank-order test. Two-tailed p values  0.05 were considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There were no significant differences in age, sex, and height between the normal and asthmatic subjects. Asthmatics had significantly lower FEV₁ and greater drops in FEV₁ after challenges with methacholine, exercise, and hyperpnea (Table 1).

Peripheral Resistance

The asthmatics had a statistically higher Rp at baseline when compared with the normal subjects (Table 2), and Rp decreased with increasing flow in both groups (Figure 4). Only the Rp of asthmatics increased significantly over baseline after cool, dry air challenge (Figure 5), as reported in our previous study for a subgroup of the same subjects (4). Also, similar to our previous study, Rp correlated with all measures of airway hyperresponsiveness (methacholine, exercise, and hyperpnea) (data not shown).

View larger version (28K): [in this window] [in a new window]   Figure 4.   Changes in peripheral resistance (Rp), plateau pressure (Pp), and peripheral compliance (Cp) with increasing flow of warm, humid air in normal and asthmatic subjects. Notice that Rp falls, but Pp does not change consistently in either group with increasing flow. Overall, Cp also does not appear to change consistently with increasing flow in either group, but some members of each group appear to have slight increases in Cp with flow. The scale of Cp in asthmatics has been changed to allow better visualization of the data in this group.

View larger version (26K): [in this window] [in a new window]   Figure 5.   Changes in peripheral resistance (Rp) after cool, dry air segmental challenge in normal and in asthmatic subjects. Notice the significant rise in Rp in the asthmatics at the 10 min postchallenge time point. Rp remains significantly higher in asthmatics than in normal subjects at baseline and during recovery. Horizontal lines = least-squares mean values; *p < 0.01 versus Normals; p < 0.01 versus Baseline.

Pattern of Pressure Decay after Stop-Flow

The fit of all Pb decay curves to a single exponential function was poor, as evidenced by curvilinear shapes in a semilog plot. During all phases of the protocol and in all subjects, we were unable to demonstrate any measurable initial drops in Pb (Pb  Ps), indicating that the pressure drop across the large proximal airways was negligible. An example of such an analysis from an asthmatic subject is shown in Figure 3. Because the airways leading from the bronchoscope to the obstructed segment did not contribute measurably to Rp, then Rp appeared to be determined solely by the resistance of the distal, collateral pathways.

Plateau Pressure

Asthmatics exhibited a consistently higher Pp at baseline when compared with normal subjects (Table 2). However, Pp did not significantly change from its baseline value either with increasing flow (Figure 4) or after cool, dry air challenge (Figure 6) in either group. Plateau pressures were significantly correlated with baseline Rp in both normal subjects (Spearman's Rho = 0.81, p < 0.01) and asthmatics (Spearman's Rho = 0.96, p < 0.01) (Figure 7). In addition, Pp significantly correlated with all measures of airway hyperresponsiveness in asthmatics (methacholine PC₂₀ FEV₁, FEV₁-EX, and FEV₁-HP) (Table 3).

View larger version (24K): [in this window] [in a new window]   Figure 6.   Changes in plateau pressure (Pp) after cool, dry air segmental challenge in normal and in asthmatic subjects. Notice the lack of change in Pp compared with that at baseline after the challenge in either group. Pp remains significantly greater in asthmatics than in normal subjects at baseline and during recovery. Horizontal lines = least-squares mean values; *p < 0.01 versus Normals.

View larger version (12K): [in this window] [in a new window]   Figure 7.   The correlation between baseline peripheral resistance (Rp) and plateau pressure (Pp) in normal and in asthmatic subjects. Spearman's rho = 0.96 for Asthmatics and 0.81 for Normals.

Peripheral Compliance and Time Constant

Baseline Cp was significantly lower in the asthmatics, being about 35% that of the normal subjects (Table 2), and remained lower than the Cp of normal subjects at all time points postchallenge (Figure 8). As flow was increased during the Pressure-Flow Phase, Cp did not consistently change in either group, although some members in each group appeared to have slight increases in Cp with increasing flow (Figure 4). After challenge, Cp did not change significantly from baseline within either group at the 1, 5, or 10 min time points (Figure 8). Likewise, Cp during high flow of warm, humid air during Phase I of the challenge protocol did not differ significantly from Cp during the cool, dry air challenge at the same high flow rate in either group. Although Rp correlated with all measures of airway hyperresponsiveness (methacholine, exercise, and hyperpnea), as in our previous study (4), Cp did not correlate with any of these measures.

View larger version (23K): [in this window] [in a new window]   Figure 8.   Changes in peripheral compliance (Cp) after cool, dry air segmental challenge in normal and in asthmatic subjects. Notice the lack of change in Cp compared with that at baseline after the challenge in either group. Cp remains lower in asthmatics than in normal subjects at each time point, but is significantly different only at baseline (p = 0.05) and at 1 min postchallenge (p = 0.04). (p = 0.06 at 5 min postchallenge and p = 0.10 at 10 min postchallenge.)

Baseline  was greater in the asthmatics at baseline, but this difference did not achieve statistical significance (Table 2);  fell significantly below baseline at 1 min postchallenge in the asthmatics (from 3.53 ± 1.38 to 1.66 ± 1.43 s, p < 0.01), but was not significantly different at 5 or 10 min postchallenge. In the normal subjects,  did not significantly change from baseline at any time point postchallenge.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The findings of this study can be summarized as follows. In all subjects, Rp is attributable only to the collateral pathways, not to the larger airways that lie between the wedged bronchoscope and the distal segment. At baseline, the Rp of asthmatics is significantly higher than that of normal subjects, falls during the application of increasing flows of warm, humid air and during the cool, dry air challenge, and rises over baseline after challenge. The Pp of asthmatics is greater than that of normal subjects at baseline, but it does not change with increasing flows of warm, humid air, or during or after cool, dry air challenge. In addition, the Pp of both normal subjects and asthmatics correlates closely with baseline Rp, and, in asthmatics, both Rp and Pp are significantly correlated with PC₂₀, FEV₁-EX, and FEV₁-HP. The Cp of asthmatics is lower at baseline than that of normal subjects, but it does not change significantly from baseline either during increasing flow or after challenge. Taken together, these results demonstrate that asthmatics at baseline have a higher Rp, higher minimal closing pressure (higher Pp), and lower Cp than do normal subjects. However, neither Pp nor Cp are altered by segmental challenge with cool, dry air, despite increases in Rp. These findings suggest that asthma is associated with physiologic changes at the level of both the collateral pathways and the lung parenchyma, but it is only the collateral pathways that respond to local, cool dry air stimulation by narrowing or closing.

Methodologic Considerations

Before further discussing the implications of these findings, it is important to consider various methodologic issues. Many of these details were discussed in our previous report (4) in which we addressed the use of unidirectional flow, bypassing of the upper airway, and the use of 5% CO₂/air as the challenge gas. We also paid careful attention to pressure waveforms to assure a secure, airtight wedge and patent catheter, as determined by observing clear cardiogenic artifacts and pressure fluctuations with breathing.

In the current report, it is also important to consider the stop-flow maneuver used to measure  and thus calculate Cp. Concerns have been raised (13) regarding the use of this method because of the potential inability for subjects to hold their breath long enough to measure true plateau pressures, and because of the potential variability and inconsistent reproducibility of the measurements. We found that nearly all subjects reached an apparent plateau pressure within the limits of their breathhold time, defined visually as a pressure change of less than 0.5 cm H₂O over approximately 1 s. The reproducibility of the plateau pressures was usually within 2 cm H₂O in normal subjects and 5 cm H₂O in asthmatics; we took at least two measurements at each level of flow and averaged the results.

How do our results compare with others using this technique in humans? In the study by Terry and colleagues (14), Cp was found to be 1.9 × 10³ L/cm H₂O in young, supine normal subjects compared with our value of 0.70 × 10³ L/cm H₂O. These investigators did not state what flow rate was used or in which lobe this measurement was made. Inners and colleagues (15) found a Cp of 5.6 × 10³ L/cm H₂O in the right upper lobe of supine normal subjects, but also did not specify the flow at which this measurement was made. The compliance we calculated in our supine normal subjects is clearly lower than that found in these previous studies, but we measured compliance using a single flow and only in the right lower lobe. Our data seem reasonable given that a similar degree of variability is found in Rp between the right upper and right middle lobes in humans (14). Interestingly, our data are also in line with those of Woolcock and Macklem (7), who measured segmental compliance in cadaveric human lungs using a bolus injection technique, and found it to vary between 0.05 and 1.82 × 10³ L/cm H₂O. Finally, if we consider that the wedged bronchoscope subtends approximately 1% of total lung volume, and total respiratory system compliance is on the order of 0.1 L/cm H₂O, then we would expect Cp to be approximately 1% of 0.1, similar to our value of 0.7 × 10³ L/cm H₂O.

Stop-Flow Analysis

After stop-flow, an abrupt initial fall in pressure is thought to be indicative of the resistive drop across the large, immediately subtended, airways just distal to the tip of the bronchoscope, whereas the subsequent slower decay reflects the dissipation of pressure through the small, peripheral airways that comprise the collateral channels (6). This partitioning of resistance, however, is also subject to debate (13). Previous studies have suggested that increases in large airways resistance could be detected after methacholine in dogs (5, 6), or in human subjects with emphysema (14). In contrast, a subsequent study by Ludwig and colleagues (10) was unable to confirm significant conducting airway resistance in dogs. We also were unable to demonstrate a measurable initial drop in Pb (Pb  Ps) when flow was stopped, in contrast to Smith and colleagues (5). There are several likely reasons that might explain the discrepancy among these studies. The study of Smith and colleagues (5) involved the direct instillation of methacholine into the airways, whereas the study of Ludwig and colleagues (10) involved nebulization of methacholine from the tip of the bronchoscope. Direct instillation of a bronchoconstricting stimulus would be expected to act locally and cause proximal airway narrowing, but distal nebulization of medication may have a more peripheral effect. A surprising finding in the current study is that the instillation of cool, dry air appears to have distal effects, despite being a local stimulus, as the instilled air would have been expected to reach BTPS conditions by the time it arrived at the distal regions of the wedged segment. As discussed in previous reports (3, 4), these distal effects may arise on the basis of mediator release, neurogenic stimulation, or perhaps surfactant dysfunction.

Plateau Pressure

The final plateau pressure obtained after cessation of airflow (Pp) is assumed to occur because of airway closure (7). We reasoned that if the bronchoscope were not wedged, then, after cessation of airflow, Pb should decay rapidly to atmospheric pressure, i.e., zero. Likewise, if the wedged segment were in complete communication with the surrounding segments, Pb would also decay to zero. The fact that Pb did not decay to zero must therefore mean that air must have been trapped in the isolated segment, unable to communicate with surrounding units via collateral channels. As suggested by Menkes and Traystman (6), this situation would occur if all the small airways serving the wedged segment closed. Alternatively, Woolcock and Macklem (7) have offered that closure of the airways between the wedged bronchoscope and the collateral pathways could also account for this elevated "back pressure." However, we were unable to measure any significant pressure drop and hence resistive component in these airways; therefore, we believe that Pp was determined by the degree of closure of the distal small airways that serve as the collateral pathways for the obstructed segment.

It is important to point out that Pp is only sensitive to the airways with the lowest, minimal closing pressure, and therefore measurement of Pp does not allow us to determine whether some airway closure occurs at closing pressures higher than the minimal closing pressure. Nevertheless, Pp was higher in our subjects with mild asthma and correlated with other measures of airways responsiveness.

Peripheral Compliance

The physical meaning of Cp is less clear. Menkes and Macklem (16) were uncertain whether this index of compliance related to recruitment of lung units or to changes in the intrinsic elastic properties of the tissue in the challenged segment. Changes in elastic properties can occur via changes in airway smooth muscle tone or changes in surface tension (6). Changes in dynamic lung compliance after direct large airway smooth muscle constriction have been shown recently in sheep (17), and it has been suggested that contractile elements in the lung parenchyma may also be stimulated and contribute to alterations in parenchymal compliance (18). The mechanical effects of HP might change surface tension properties and thereby also change compliance (19). Therefore, the lower Cp we observed in asthmatic subjects may represent an increased intrinsic stiffness of the tissue in the isolated airway segment during emptying.

However, we know that the expiratory pressure-volume curve of the whole lung of subjects with mild asthma and HIB has the same slope as that of normal subjects, indicating relatively little change in intrinsic elastance (3). This suggests that decreases in Cp in asthmatics might arise because of the known recruitment and derecruitment of lung units that occurs with lung inflation and deflation (20). Recruitment of lung units during inflation is evidenced by the abrupt increase in slope of the inflation limb of the pressure-volume curve seen after whole-lung HP in asthmatic subjects (3). In the present study, application of increasing flows of warm, humid air appeared to be associated with slight increases in Cp in some members of both the control and asthmatic groups (Figure 4), consistent with the concept of lung unit recruitment and derecruitment.

In addition to derecruitment, the lower Cp of asthmatics may reflect the effects of interdependence between the wedged segment and the surrounding lung parenchyma, as explained by Woolcock and Macklem (7). Inhomogeneous inflation and deflation of the wedged segment in relation to the surrounding lung parenchyma may lead to increased interdependence of forces between these regions, resulting in a fall in the effective compliance of the wedged segment. Presumably, increased interdependence between the distal acinar units within the wedged segment itself could also contribute to the decreased Cp measured in this study. The fact that Cp does not change acutely with direct stimulation by cool, dry air, but is reduced in the asthmatics at baseline, suggests that Cp may be determined by the more chronic effects of ongoing inflammation in asthma.

Modeling of Events in the Lung Periphery

To better understand the anatomic relationships of the components of the wedged segment, we constructed a conceptual model based on the data. First, because our data analysis involved only the initial, monoexponential part of the Pb decay curve, we chose to model our findings according to a linear, single compartment represented as the pipe and balloon configuration shown in Figure 9A. The pipe represents the airways immediately distal to the wedged bronchoscope, and leads to the multiple acinar units depicted by the balloon that comprises the parenchymal segment. The collateral pathways are represented by the smaller pipe coming off the main conduit at a distal point in the wedged segment. The respiratory bronchioles and alveolar ducts comprise the collateral pathways (13), and are thus still proximal to the acinar units of the parenchyma.

View larger version (20K): [in this window] [in a new window]   Figure 9.   Modeling of wedged airway segment. (A) A linear single-compartment model of the wedged airway segment. The bronchoscope is wedged into the airway (pipe) subtending a distal lung segment (balloon). Airflow () from the wedged bronchoscope proceeds down the airway and fills the distal segment until the pressure within the segment (Ps) comes into equilibrium with the pressure at the tip of the bronchoscope (Pb). At this point, will continue out of the collateral channel because of the pressure difference across it, equal to Ps minus atmospheric pressure at the airway opening (Pao). The ratio Pb/ during a breathhold at FRC equals the resistance of the model, which comprises the resistance of the large airway between the bronchoscope and the distal segment (Raw), together with the resistance of the collateral channel, Rp. The elastance and volume of the model is represented by the effective compliance of the wedged segment (Cp). ( B) A nonlinear model of the wedged airway segment. In comparison to the linear model, the nonlinear model contains two or more collateral pathways, which can narrow or close, and the single balloon of the distal parenchyma is replaced with several acinar units, each composed of small airways (respiratory bronchioles and alveolar ducts) leading to alveolar units. Notice that the collateral channels and small airways leading to the distal alveolar units could be narrowed or closed to variable extents, leading to changes in Rp in the case of the collateral channels and to changes in Cp in the case of the acinar units.

This single compartment model allows us to explain two separate findings of this study, namely, the baseline differences between asthmatics and nonasthmatics, and the events occurring after cool, dry air stimulation. At baseline, the model suggests that the higher Rp of asthmatics is due to a narrower collateral channel, whereas the lower Cp is due to increased intrinsic tissue stiffness within the parenchymal segment and/or fewer lung units within the segment available for ventilation. After stimulation with cool, dry air, the model indicates that the increase in Rp is due to collateral pathway narrowing, but the lack of change in Cp indicates no change in either intrinsic tissue stiffness or the degree of lung unit recruitment within the parenchymal segment.

A limitation to this model is that if it is linear (that is, if Rp and Cp remain fixed in any individual), the model predicts that Pp should decay to zero once flow is stopped. This was not the finding in either the asthmatic or the nonasthmatic groups. Also, a single exponential described only the initial part of the Pb decay curve but not the entire curve. Thus, a linear single-compartment model does not adequately explain our data, suggesting that the segment might be better modeled either as a collection of multiple compartments or as a single nonlinear compartment (10). Data from the dog study of Ludwig and colleagues (10) would suggest the presence of a single nonlinear compartment in mammals. Indeed, our finding that Pp asymptotes toward a finite plateau is indicative of nonlinear behavior because such behavior suggests that Rp increases dramatically as Pp decreases, eventually becoming infinite. Interestingly, in the dog study of Ludwig and colleagues (10), complete closure of collateral channels did not appear to occur even though Rp obviously did vary markedly with pressure. Thus, our data reveal a non-linear characteristic not found in the dog and therefore perhaps unique to the human lung.

Although a nonlinear model is needed to better predict the experimentally observed data, the interpretations of Rp, Pp, and Cp are essentially the same. We consider such a nonlinear model as one in which there is more than one collateral channel exiting a given lung segment, and the distal parenchyma is represented by many individual acinar units (Figure 9B). On the basis of this view, the higher Rp of asthmatics compared with that of nonasthmatics at baseline implies either narrower or fewer collateral channels. Furthermore, the last of these channels to close off during segment-emptying must do so at higher segmental pressure in the asthmatics, explaining their higher Pp. Finally, the lower Cp of asthmatics suggests that fewer lung units within the distal lung parenchyma are available for ventilation (derecruitment), although changes in the intrinsic elastance or interdependence among these units may also contribute to the lower effective Cp.

This nonlinear model is compatible with the known anatomic relationships between the conducting airways, collateral channels, and distal acinar units. The distinct anatomic locations of the components that make up the Rp (the collateral channels) and the Cp (the parenchymal units) would explain the lack of correlation between Rp and Cp at baseline. This model would also explain the lack of change in Cp to cool, dry air because the stimulus would not reach the parenchymal units as airflow was shunted away through the collateral channels. However, Rp and Pp are correlated at baseline (Figure 7), which would be expected if they were both related to the same anatomic site, in this case, the collateral channels. Of note, our data suggest that the collateral pathways must originate proximal to the distal acinar units, and not be a component of them, because if there were narrowing or closure of small acinar airways in response to cool, dry air, then we would have expected to see a fall in Cp directly associated with the rise in Rp.

Implications for Clinical Asthma

The present study reveals important differences in peripheral lung mechanics between asthmatic and nonasthmatic subjects, and emphasizes the concept that lung mechanics are altered in a complex way even in subjects with very mild asthma who have normal spirometry. With regard to HIB, the results of our study may only apply to the special situation of an isolated, wedged segment in which only the collateral pathways are stimulated with high-flow, cool, dry air. However, during real-life hyperpnea, airflow will occur within the respiratory bronchioles and alveolar ducts leading to the distal alveolar units, and perhaps similar effects will be seen at the level of these airways as occurred at the level of the collateral pathways in our study. Indeed, we know from a previous study that the lung periphery is involved in the reaction after whole-lung HP challenge (3).

The strong correlation between baseline Rp and Pp in both normal subjects and asthmatics is supportive of the view that Rp is due in part to the degree of collateral airway closure. Such closure could lead to the heterogeneous distribution of recruited and derecruited lung units seen in asthma (21). Ironically, this phenomenon may be viewed as protective, in the sense that areas of derecruited or microatelectatic lung could exert radial traction on recruited and open areas, thereby utilizing the mechanism of interdependence to stabilize these areas against closure and collapse (22). Consequently, if elastic recoil forces and interdependence were abruptly diminished, as has been hypothesized to occur in unstable asthma (23), then further and more severe airway narrowing and derecruitment might occur. In support of this contention, we have shown that whole lung interdependence in nocturnal asthmatics is preserved during the wake state but is lost during sleep (24).

Finally, loss of lung units subtended by a common airway in asthma might contribute to the phenomenon of airway hyperresponsiveness. The fewer remaining open lung units could be exposed to a proportionally higher dose of any bronchoconstricting stimulus, resulting in enhancement of airway hyperresponsiveness. Indeed, the loss of lung units from airway closure in response to a bronchoconstricting stimulus has been postulated to be an index of excessive bronchoconstriction and asthma severity (25).

In conclusion, previous work in humans using the wedged bronchoscope technique has yielded important information on peripheral airway resistance. This study is unique in that we have demonstrated how information about segmental compliance and closing pressure can be derived from an analysis of pressure decay after cessation of airflow into the segment. We have demonstrated fundamental differences in peripheral lung physiology in asthmatic and normal subjects and in their responses to cool, dry air. Asthma is characterized by changes in the lung periphery that result in higher closing pressures and lower peripheral compliance compared with the nonasthmatic condition. After cool, dry air stimulation of a wedged segment, the mechanical response in asthmatics appears to be solely limited to the collateral channels, which narrow or close. There is no evidence that the larger airways proximal to the collateral channels, nor the distal acinar units beyond the collateral channels, are affected. The behavior of the wedged lung segment after stop-flow is best described by a nonlinear model, similar to previous studies of the wedged lung segment in dogs (10). We believe this type of analysis will be useful for investigating further the mechanisms involved in HIB and other phenotypes of asthma, as well as more general mechanisms leading to altered parenchymal pathophysiology in other lung diseases.