INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Airway narrowing following a period of exercise or hyperpnea in subjects with asthma may be caused by many factors, including airway smooth muscle constriction, airway edema, mucus secretion, and reactive hyperemia of the airway wall (1). Reactive hyperemia refers to the rapid increase in bronchial blood flow within the airway wall in response to the cooling and drying of the airway that occurs during the hyperpnea of exercise (2). Such an increase in blood flow could lead to vascular engorgement and perivascular edema, both contributing to direct narrowing of the airway (2). A reactive hyperemia of the bronchial circulation in hyperpnea-induced bronchospasm (HIB) is suggested by several lines of evidence. First, the airways of subjects with asthma rewarm faster than those of normal subjects following a bout of exercise or hyperpnea (3, 4). Second, subjects with asthma have increased vascularity of their airway walls (5) and increased bronchial blood flow at baseline (8) compared with normal subjects. Third, maneuvers that alter bronchial blood flow are associated with corresponding predicted changes in airway resistance (2, 9, 10). However, some authors question the significance of the effect of vascular engorgement on airflow (11), and the reactive hyperemia hypothesis remains unproven because no direct measurements of bronchial blood flow following hyperpnea in asthmatics have been made.

Whereas the bronchial circulation is the source of heat and water for the conducting airways of the lung, the pulmonary circulation is the principal source for the lung parenchyma, including the peripheral airways. This peripheral lung region is also involved in the response of subjects with asthma to exercise or hyperpnea. In a dog model of airflow-induced bronchospasm, Freed and colleagues have shown that the resistance in the peripheral airways increases following cool, dry air stimulation (12). In humans, peripheral airway resistance is higher in subjects with asthma than in normal subjects (13), and increases in response to bradykinin (14) and histamine (15), implicating the direct involvement of these small airways in the pathophysiology of asthma. Recently, we have shown that peripheral airway resistance in subjects with asthma with exercise- and hyperpnea-induced bronchospasm (HIB) increases in response to directly instilled cool, dry air (16). Furthermore, this resistance and its response to cool, dry air localizes specifically to the peripheral airways comprising the collateral ventilatory pathways (17), thus implicating their involvement in HIB. Other studies have also demonstrated small airway abnormalities in asthma and exercise (18, 19). Finally, in addition to small airway involvement, hyperpnea also results in changes in lung tissue mechanics that suggest the involvement of the lung parenchyma in exercise (20) or hyperpnea-induced bronchospasm (21). Thus, the lung periphery, supplied by the pulmonary circulation, is involved in the pathophysiology of HIB.

These observations suggest that just as alterations in bronchial blood flow have been implicated in hyperpnea-induced airflow limitation within the large, conducting airways, so the pulmonary circulation might be involved in the response of the lung periphery to hyperpnea. Although it would be unlikely for the pulmonary circulation to respond to thermodynamic fluxes in the proximal airways by virtue of its deep location within the lung, it is possible that primary changes in bronchial blood flow could cause secondary changes in pulmonary blood flow at the level of the peripheral airways because of the extensive anastomoses of the two circulations in this region (22, 23). Indeed, many of the maneuvers described to alter the response to hyperpnea via changes in bronchial blood flow (including saline loading [10] or application of shock trousers [9]) could also alter pulmonary blood flow. Alternatively, the chemical mediators or neural events involved in the bronchospastic response to exercise might affect the pulmonary circulation directly. For example, the application of vasodilators into the airways of sheep has been shown to increase peripheral lung resistance, which includes a component of tissue resistance (24). Increased blood flow through the pulmonary circulation could affect airflow in the peripheral airways by causing airway narrowing from capillary engorgement, airway wall edema, or direct mechanical compression (23). Likewise, parenchymal lung mechanics may be altered by increased pulmonary capillary blood flow leading to capillary leak and interstitial edema (23).

Based on the known changes in peripheral lung mechanics that occur in HIB and the vascular supply of this region by the pulmonary circulation, we hypothesized that pulmonary capillary blood flow would increase in response to hyperpnea in subjects with asthma with HIB, and such increases would be accompanied by increases in pulmonary capillary blood volume (VC). To test this hypothesis, we measured the VC component of the DL_CO before and after hyperpnea in subjects with asthma with HIB and in control subjects without asthma.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Based on published data regarding changes in VC and its measurement variability (25) we determined that a sample size of 20-25 subjects would be necessary for the study to have adequate statistical power (1   = 0.8) to detect a 15-20% change in VC at an  level of 0.05. Such a change appears biologically relevant because this degree of change has been associated with exercise (26) and with the increase in airway resistance in subjects with asthma with nocturnal symptoms (28).

Subjects with exercise-induced asthma and control subjects without asthma were recruited for participation in this study. All subjects gave informed consent and were studied in accordance with the guidelines of the Institutional Review Board of the University of Vermont College of Medicine. All subjects were characterized by clinical history, physical examination, spirometry, and challenges with methacholine and hyperpnea. Control subjects were defined as subjects who had no clinical history of asthma or respiratory disease, a normal physical examination, and spirometry within normal limits. Control subjects also demonstrated a provocative concentration of methacholine that caused a 20% fall in FEV₁ (PC₂₀) of > 8 mg/ml (29). Subjects with asthma were required to meet the ACCP-ATS definition of asthma (30) and to have a normal physical examination, with a resting FEV₁  60% of predicted after withholding medication. Subjects with asthma were also required to demonstrate a PC₂₀  8 mg/ml (29) and a  10% fall in FEV₁ following a hyperpnea challenge (31). All subjects with asthma were on inhaled bronchodilators alone as chronic therapy, and withheld their medication for 12 h prior to any visit. None of the control subjects or subjects with asthma were smokers, and none of them had had any respiratory illness within the previous 4 wk.

Protocol

All subjects were studied on each of 3 d, with no less than 24 h between study days. The following specific procedures were performed.

Spirometry

Spirometry was performed following ATS guidelines (32) on Day 1 by an automated spirometer using a dry-seal system (Collins Medical, Braintree, MA) to measure FVC and FEV₁; the best of three efforts was recorded, and expressed as percent predicted (33).

Methacholine Challenge

Methacholine challenge testing using the five deep breath technique was also performed on Day 1 according to published guidelines (29). The PC₂₀ FEV₁ was determined by linear extrapolation of the dose- response constructed from a semilog plot of the observed data.

Hyperpnea Challenge

Subjects returned on Day 2 to document HIB. Following the technique of Eliasson and coworkers (31) each subject performed hyperpnea with cool, dry gas for 5 min at a ventilatory target of 80% of their FEV₁ × 35. A fall in FEV₁  10% from baseline following the hyperpnea challenge was considered a positive response according to the criteria established by Eliasson and colleagues (31).

DL_CO

On Day 3, subjects returned for measurement of DL_CO before and after a hyperpnea challenge identical to the challenge that occurred on Day 2 (Figure 1). Measurements of DL_CO were made in duplicate using the single-breath technique following published guidelines (34). Two features of the measurement were used to protect against the technical difficulties associated with measurement of DL_CO in the presence of bronchospasm (35). First, we used the modified Jones- Meade breath-hold time (36), which was found in subjects with asthma to be as accurate as the more precise three-equation method (35). Second, we used a real-time methane gas tracer to precisely follow and select deadspace washout volume and alveolar gas sample volume, improving accuracy (37). End-expiratory oxygen fraction (FE_O2) was measured with an oxygen sensor (Oxychek, Critikon, Tampa, FL) placed immediately distal to the mouthpiece.

View larger version (20K): [in this window] [in a new window]   Figure 1.   Diffusion measurement protocol. The diffusion measurement protocol consisted of a series of measurements made before and after a 5 min hyperpnea challenge. During baseline, spirometry (spiro) was first measured, and then the diffusing capacity of the lung for carbon monoxide (DLCO) was measured in duplicate with each of a standard and hyperoxic test gas. At least 4 min was allowed to pass between successive measurements. During the posthyperpnea period, spirometry was performed at 1 and 3 min posthyperpnea, and then the same series of DLCO measurements made during the baseline period was repeated. To confirm continued airflow limitation at the end of the DLCO measurements, spirometry was repeated following the final DLCO measurement, approximately 22 min posthyperpnea.

Based on the protocol by Morrison and colleagues (27) for calculating VC, we repeated the DL_CO measurement at only one higher oxygen tension (89%), which Morrison and colleagues found to be as accurate as using three or more oxygen tensions. We followed this simplified protocol because we wished to perform all measurements in duplicate, but we were constrained by the limited time frame of the posthyperpnea period. All acceptable values of DL_CO, alveolar volume (VA), and FE_O2 were entered into a computer spreadsheet for later data analysis using the mean of the repeated measurements.

Following the baseline measurements of DL_CO, subjects performed an identical period of hyperpnea as on Day 2. Immediately following hyperpnea, spirometry was recorded at 1 and 3 min postchallenge to document airflow, and then the sequence of DL_CO measurements with two different test gases, made in duplicate, was repeated. A final spirometric measurement was performed at the end of the DL_CO testing, approximately 22 min following the cessation of the hyperpnea challenge.

Two control experiments were also performed with regard to measurement of VC. First, to test the sensitivity and validity of the testing procedure for detecting increases in VC, three normal subjects and three subjects with asthma performed the two-gas DL_CO maneuver in both the standing and supine positions, in random sequence, during which we expected VC to rise while the subject was supine (25). Second, to assess for a testing sequence bias, two normal subjects performed the two test-gas DL_CO maneuver on separate occasions, once using the standard test gas before the hyperoxic test gas (as used in the study protocol), and once using the hyperoxic test gas before the standard test-gas. The variability in measuring DL_CO by altering the sequence of test gases was expressed as the coefficient of variation and compared between these two subjects and the study subjects as a whole.

Data Analysis

All DL_CO measurements were corrected for CO back pressure, the "anemia effect" of carboxyhemoglobin and for hemoglobin using published guidelines (34).

The calculations of VC and D_m were made according to the method of Roughton and Forster (38) using the formula 1/DL_CO = 1/D_m + 1/ VC, where  is the reaction constant of hemoglobin and oxygen, and D_m is the intrinsic diffusing capacity of the alveolar-capillary membrane. Assuming a ratio of  for the permeability of the red blood cell membrane to that of the red blood cell interior,  was determined from 1/ = 0.33 + 0.0057Pcap_O2, where Pcap_O2 is the partial pressure of oxygen in the pulmonary capillary. PcapO₂ was estimated as end-expiratory partial pressure of oxygen (PE_O2) 10, where PE_O2 = barometric pressure  47 × FE_O2. Once  was calculated, 1/DL_CO was plotted against 1/, from which VC was calculated as the reciprocal of the slope. D_m was also calculated as the reciprocal of the y-intercept.

All subject characteristic data are expressed as mean ± SD, except for the PC₂₀, which is expressed as geometric mean and range. The calculated values of VC are expressed as means ± SEM. Due to the nonnormal distribution of D_m, log-transformed values of D_m were used for comparisons and D_m is expressed as geometric mean and range. Comparisons between subjects with asthma and normal subjects were made with an unpaired Student's t test, and within groups before and after hyperpnea by a paired Student's t test. Correlations among data were analyzed by analysis of covariance with calculation of the Pearson coefficient of determination. Two-tailed p values < 0.05 are considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject characteristics are shown in Table 1. The 10 control subjects and 13 subjects with asthma were comparable in age, height, and sex distribution, and FEV₁ as a percent of predicted, but the subjects with asthma exhibited mild airflow limitation at baseline as seen by a lower FEV₁/FVC ratio. Following hyperpnea, the subjects with asthma had significant falls in FEV₁ and FVC, and these changes were significantly different than the corresponding changes in the control subjects. Although the control subjects also had significant falls in FEV₁ and FVC from baseline, these changes were small and considered not clinically significant.

The DL_CO control experiments revealed that the test procedure was able to measure a mean increase of 34 ± 8% (mean ± SE) in VC in the supine compared with the standing position. The smallest change detected was 9%. These data are comparable to those of Lewis and colleagues (25), who found a mean increase in VC of 46% in moving from the sitting to recumbent position. In addition, we found no significant difference in the variability of measuring DL_CO whether measured by performing the DL_CO maneuver first with the standard test gas or first with the hyperoxic test gas (coefficient of variation = 0.04 for all study subjects versus 0.02 for the two control subjects, p = 0.6).

The baseline DL_CO data (Table 2) reveal that the subjects with asthma had a slightly higher DL_CO than the normal subjects, although not statistically significant, with a corresponding slightly higher VC and statistically greater D_m. Following hyperpnea challenge (Figure 2), overall DL_CO did not change in either group, nor were there significant changes in VA. Likewise, VC did not change significantly in either group (Table 3 and Figure 3), despite continued airflow limitation in the subjects with asthma at the end of DL_CO testing as documented by the mean change in FEV₁ of 11% at 22 min. There was no change in VC whether expressed as VC/VA, VC/height, or VC/ DL_CO predicted (Table 3), and there was also no change in D_m either alone (Table 3) or when corrected for these size-normalizing factors (data not shown).

View larger version (29K): [in this window] [in a new window]   Figure 2.   Changes in DLCO (A) and alveolar volume (VA) (B) in control subjects and subjects with asthma pre- and posthyperpnea challenge. Short horizontal lines represent mean values.

View larger version (26K): [in this window] [in a new window]   Figure 3.   Changes in pulmonary capillary blood volume (VC) and airflow (%FEV1) in control subjects and subjects with asthma pre- and posthyperpnea challenge. Short horizontal lines represent mean values. During the posthyperpnea period, spirometry was repeated at 1, 3, and approximately 22 min, with percent changes in FEV1 from baseline values depicted for each time point in the rectangular boxes.

Analysis of covariance revealed that VC varied slightly but significantly with FEV₁ in subjects with asthma (r² = 0.36, p = 0.02), but D_m did not. Neither VC nor D_m was associated with FEV₁/FVC. D_m was associated with VA in subjects with asthma (r² = 0.53, p = 0.01), but VC was not. There were no associations between VC or D_m and PC₂₀ or FEV₁ following hyperpnea.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of this study indicate that no change in VC occurs in subjects with asthma with HIB following a bout of hyperpnea. Insofar as VC reflects pulmonary capillary blood flow, this finding does not support the concept that pulmonary capillary blood flow is altered in response to hyperpnea in subjects with asthma with HIB.

Three important issues must be considered before accepting the results of this study. First, bronchospasm persisted throughout the time of measurement (Figure 3), making it unlikely that we simply missed the response. Second, despite the development of bronchospasm in the subjects with asthma, the DL_CO technique was a valid method by which to calculate VC because appropriate modifications of the technique were employed. Third, the measurement technique had sufficient sensitivity (based on the control experiments) and the sample size had sufficient power to detect any significant changes in VC, as defined previously.

We assumed that the falls in FEV₁ seen in our subjects were associated, at least in part, with changes in peripheral airway resistance and lung mechanics. Although we did not measure these parameters directly, two lines of evidence support this contention. First, the current data revealed falls in FVC in addition to FEV₁ following hyperpnea, implicating airtrapping due to distal airway narrowing or closure (39). Second, we have previously shown that an identical hyperpnea stimulus also results in increases in upstream resistance and pressure- volume hysteresis, further implicating involvement of the small airways and lung parenchyma (21).

How can we explain the lack of change in VC in light of ongoing bronchospasm? First, as previously stated, it would be unlikely for the pulmonary circulation to respond in the same way to heat and water fluxes within the conducting airways as has been postulated for the bronchial circulation. While the heat flux caused by cool, inspired gas may penetrate very deeply into the lung at high minute ventilation (4), it is not known whether it will reach the level of the peripheral airways or lung parenchyma. In addition, it would be unlikely that hyperemia would occur in response to airway drying at this level, as there is sufficient airway wall lining fluid available to rehumidify the gas within the first 12 generations of airways (40).

Second, VC may not have changed because bronchial blood flow did not change. Increases in bronchial blood flow to the intrapulmonary airways would be expected to lead to increases in VC because of precapillary anastomoses (22) of the two circulatory beds within the region of the respiratory bronchioles and alveolar ducts. However, the exact site of such anastomoses is controversial (22). Our protocol was not designed to detect changes in bronchial blood flow and consequently cannot address the bronchial circulation directly.

Third, it is possible that changes in bronchial blood flow did occur, but were not associated with any changes in the pulmonary circulation. However, the data supporting a reactive hyperemia of the bronchial circulation are circumstantial. The most direct piece of evidence has been the observation that the airways of subjects with asthma rewarm more rapidly than those of normal subjects following a period of airway cooling induced by exercise or hyperpnea (3, 4). Although this has been postulated to be due to the need both to rewarm and rehumidify the airway mucosa, some authors have questioned whether a true hyperemia actually occurs, or whether airway temperatures rewarm faster simply due to ongoing obstruction and increased residence time of gas in airways of subjects with asthma (40).

Perhaps small changes in pulmonary blood flow do occur and were beyond the detection limit of our technique, in which case hyperemia may be one of many factors that thickens the airway wall and therefore enhances the airway narrowing effect of a given degree of airway smooth muscle constriction (41). Other factors affecting peripheral lung mechanics may also be involved, such as peripheral airway smooth muscle constriction, myofibroblast activation, edema formation, and airway narrowing or closure from surfactant dysfunction (21). How a proximal stimulus such as airway cooling and drying might cause such distal responses is unknown, but may be related to mediator production, neural stimulation, or possibly mechanical factors (21).

One interesting finding of this study was that D_m was higher at baseline in these subjects with mild asthma than in normal subjects, perhaps explaining the slightly higher DL_CO found in the subjects with asthma. The higher DL_CO often seen in stable subjects with asthma is usually assumed to be due to a higher VC, possibly from increased negative intrathoracic pressures (42) or enhanced apical perfusion in asthma (43). Indeed, changes in VC have been associated with changes in pulmonary function in asthma, as VC has been found to increase overnight in subjects with asthma with worsening nocturnal symptoms associated with falls in FEV₁ (28).

In this study, though, D_m, not VC, was significantly higher in the subjects with asthma. Unfortunately, this finding is not as reliable a measurement as VC because of the greater dependence of D_m on the chosen value for red blood cell membrane permeability (assumed to be  here) (38). Nevertheless, one mechanism proposed to explain an increase in D_m is hyperinflation, as Pecora and colleagues (44) found an increased D_m in children with severe asthma with hyperinflation. Our data support such a finding insofar as D_m might be related to overall lung size, because there was a significant association between D_m and VA in the subjects with asthma. However, although lung volumes were not measured, the subjects with asthma in our study were not significantly hyperinflated as assessed by their normal FVC at baseline, which we would have expected to be reduced in the setting of significant hyperinflation. Another mechanism that may explain an increase in D_m is capillary recruitment, such as might occur with exercise (26) or in microgravity (45). We hypothesize that capillary recruitment and remodeling of the pulmonary vasculature may occur in asthma due to chronic inflammation in the lung periphery. Theoretically, recruitment and remodeling could lead to a larger change in D_m than in VC depending on the relative sizes, shapes, and thickness of the new capillaries (26). Further studies looking at the components of VC and D_m in patients with various severities of asthma and under different treatments would be needed to elucidate this finding.

In summary, this study has shown that despite the development of HIB, subjects with asthma do not have an increase in their pulmonary capillary blood volume, and therefore alterations in the pulmonary circulation in response to hyperpnea are unlikely. Other mechanisms resulting in increased peripheral airway resistance and altered parenchymal mechanics must be involved.