INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Exercise-induced asthma (EIA) is characterized by transient airway obstruction that usually occurs after intrathoracic bronchi are exposed to cool dry air. Anderson and coworkers (1) suggested that hyperventilation increases airway surface fluid (ASF) osmolality, stimulating local mediator release and initiating airway narrowing via smooth muscle constriction. However, because of the technical difficulties associated with recovering ASF from bronchial mucosal surfaces, direct measurements of exercise- or hyperventilation-induced changes in human ASF have not been made. Rather, water flux and ASF osmolality after hyperventilation or exercise have been estimated using indirect measurements with contradictory results.

Gilbert and coworkers (2) recorded water loss from normal and asthmatic subjects breathing frigid air during exercise. These data were used to calculate hyperpnea-induced alterations in ASF osmolality throughout the tracheobronchial tree, and indicated that hyperventilation under these conditions did not increase ASF osmolality. Anderson and coworkers (3) reevaluated these data using a model that accounted for local changes in ASF volume, and concluded that the water lost under the experimental conditions described would result in ASF hyperosmolality. However, if ASF lost during and immediately after hyperventilation was rapidly replaced from intra- and extravascular sources (4), then measurements of airway water loss would not accurately reflect changes in either paracellular or transepithelial water flux. Thus, calculations of ASF losses used to estimate the strength of an osmotic stimulus would be misleading. As a result, the role of ASF hyperosmolality in the development of EIA remains in question. Thus, we used a canine model of EIA (7, 8) to (1) document the transient effects of dry air hyperventilation on ASF volume and osmolality, and (2) determine if the observed ASF kinetics are consistent with the hypothesis that hyperpnea-induced changes in ASF osmolality initiate bronchoconstriction.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Techniques

All animals were handled and maintained in accordance with the Policy and Procedures Manual published by the Johns Hopkins University School of Hygiene and Public Health's Institutional Animal Care and Use Committee.

Measurement of peripheral airway resistance (Rp). Male mongrel dogs were anesthetized with sodium thiopental (25 mg/kg, intravenously). A continuous thiopental infusion (4 to 6 mg/kg/h) supplemented with fentanyl citrate (25 to 50 µg, intravenously) every 15 to 20 min was used to maintain anesthesia. Dogs were intubated and mechanically ventilated with room air (17 ml/kg); end-expiratory CO₂ was monitored with a CO₂ analyzer (Beckman LB-2; Beckman Inc., Anaheim, CA) and maintained around 4.5% by adjusting respirator frequency. The depth of anesthesia was assessed by heart rate, blood pressure, canthal reflex, and the presence of spontaneous movement or breathing. Heart rate and blood pressure were recorded via a noninvasive monitor (Datascope Accutorr 1A; Datascope Corp., Paramus, NJ), and rectal temperature was recorded using a telethermometer (Yellow Springs Instrument Co., Yellow Springs, OH). A fiberoptic bronchoscope (5.5 mm outer diameter [o.d.], Olympus BF Type P10; Olympus Corp. of America, New Hyde Park, NY) was inserted through an airtight portal of the endotracheal tube and gently wedged into a sublobar bronchus. A polyethylene 90 catheter (interior diameter [i.d.] = 1.19 mm, o.d. = 1.7 mm) was threaded through the suction port of the bronchoscope, connected to a pressure transducer (Statham; Gould Inc., Oxnard, CA), and used to measure airway pressure at the tip of the bronchoscope (P_b). Compressed, dry, room-temperature 5% CO₂ in air was delivered around the catheter at a rate of 200 ml/min, into the wedged sublobar segment. R_p was measured by interrupting the ventilator at functional residual capacity of the lungs. Under conditions of constant flow, P_b decays to a plateau pressure greater than the alveolar pressure (Palv) in the surrounding unobstructed lung (atmospheric) so that R_p = (P_b  Palv)/200 ml/min.

Airflow challenge.

Dry air challenge (DAC). Insufflation of dry, room-temperature 5% CO₂ in air was increased from 200 to 2,000 ml/min for 5 min. Airflow was then returned to 200 ml/min and R_p was monitored over the next 17 min.

Warm humidified air challenge (WAC). Insufflation of warm humidified 5% CO₂ in air was increased from 200 to 2,000 ml/min for 5 min. Dry, room-temperature (23° C) 5% CO₂ in air was bubbled through a Plexiglas canister containing a heat exchanger submersed in warm (42° C) distilled H₂O, and delivered to the suction port of the bronchoscope via a heated, water-jacketed tube, and an inline water-jacketed water trap. The temperature and relative humidity (RH) of air leaving the tip of a bronchoscope wedged in an airway during a 5-min 2,000 ml/min WAC was 31.2 ± 0.2° C and 96 ± 1.2%, respectively.

Measurement of peripheral ASF osmolality. A preweighed filter paper disc (diameter: 6.3 ± 0.04 mm, n = 5; dry weight: 3.0 ± 0.04 mg; saturated weight: 14.6 ± 1.08 mg; n = 20) was folded to form a ~ 2 × 6 mm pledget. The pledget was passed through the port of a bronchoscope with biopsy forceps and placed on the airway mucosa ~ 2 cm from the tip of the scope. After 2 min the forceps and pledget were removed, the pledget was immediately placed in a preweighed, airtight vial, weighed (Model A200S; Sartorius, Bohemia, NY), unfolded, and placed in a vapor pressure osmometer (Model 5500; Wescor, Logan, UT) for analysis. The difference in pledget weight ( mg) was used to estimate the volume of ASF sample (µl) recovered on the pledget. The time required to remove the pledget from the airway and place it in the osmometer was recorded as "handling time," which averaged 39 ± 0.7 s (n = 24). A control filter paper disc containing 10 µl of a known standard solution (290 mmol/kg) was analyzed and recorded just prior to obtaining the ASF sample. Another control disc containing an amount of standard solution equal to the recovered ASF sample was placed in a preweighed, airtight vial for a time equal to the handling time and was analyzed within minutes of obtaining the ASF sample. The ASF osmolality corrected for volume and handling time was then calculated: ASF osmolality = ASF sample × (standard sample/adjusted standard sample). Uncorrected, volume and handling time accounted for a 52.96 ± 5.8 mmol/kg (n = 24) increase in canine ASF osmolality when laboratory conditions were 23.9 + 0.3° C and 37 ± 2.7% RH. During the course of the study, room temperature (T_a) and RH ranged from 20 to 26° C (22 ± 0.3° C) and from 25 to 65% (41 ± 2%, n = 29), respectively. However, T_a and RH remained relatively constant throughout an experiment.

Location of the filter paper pledget in a sublobar segment. The distance (d) of the filter paper pledget from the tip of the bronchoscope was determined by measuring the movement of the biopsy forceps as it was passed through the port of the bronchoscope. The size of the airway in which the pledget was placed was estimated using d (in mm) in conjunction with the following published regression equations (6) to solve for airway diameter (D_aw) in the right upper lobe: D_aw = (d39.418)/6.064; right middle lobe: D_aw = (d33.255)/4.842; right lower lobe: D_aw = (d43.952)/8.911; left upper lobe: D_aw = (d48.898)/10.676; left middle lobe: D_aw = (d35.433)/6.205; left lower lobe: D_aw = (d46.2)/10.124.

Experimental Protocols

Effect of the filter paper pledget: measurement of ASF osmolality in a model airway. A 22-mm-long Tygon tube attached to an 8-mm-diameter plastic T-tube served as a model airway. The model airway was submerged in a 38° C water bath with the vertical portion of the T-tube (plugged with a rubber stopper) and the distal end of the tygon tube resting above the water line. The length of bronchoscope that normally rests inside a dog was placed in the water bath and its tip was held in the proximal end of the horizontal portion of the T-tube with a Silastic seal. The rubber stopper was then removed from the T-tube and a 100-µl droplet of 290 mmol/kg NaCl standard solution was placed on the bottom of the artificial airway about 2 to 3 cm away from the tip of the bronchoscope. Two types of experiments were done: In the first series of experiments, the tube was ventilated via the bronchoscope with either 200 or 2,000 ml/min 5% CO₂ in air. After the exposure ended, the pledget was passed through the bronchoscope with biopsy forceps, dipped in the 100-µl standard sample, immediately removed, and handled as previously described. A micropipet was used to recover 10 µl of the remaining sample for comparison with the pledget sample. Differences in osmolality of these two samples would reflect the effect of sample recovery using a pledget. In the second series of experiments the pledget was placed in the 100-µl standard during exposure to 200 or 2,000 ml/min dry air, and was compared with a 10-µl micropipet sample of the remaining fluid. This latter series of trials reflected differences due to recovering the sample on filter paper, and the additional effect of wicking that may occur during the exposure.

Rp, ASF volume, and ASF osmolality during baseline ventilation. Tracheal ASF and plasma samples (n = 6 dogs) were collected and analyzed for comparison with bronchial ASF samples. To determine the effect of bronchoscopy on ASF, and to demonstrate the stability of ASF volume and osmolality measurements, a bronchoscope was wedged in four dogs (mass ± SEM = 18.2 ± 0.9 kg), and R_p, ASF volume, and ASF osmolality were recorded seven times during a 35-min period of sublobar ventilation (200 ml/min) with dry air.

To determine the effect of conditioned air on baseline ASF and R_p, R_p, and ASF osmolality and volume were obtained in 12 dogs (17.6 ± 0.8 kg) during ventilation with dry, room-temperature air. The gas used for peripheral airway ventilation was switched from dry air to warm humidified air, and baseline values were again recorded.

Rp, ASF volume, and ASF osmolality before and after DAC. After baseline R_p was recorded in seven dogs (21.1 ± 1 kg) using dry, room-temperature air, DAC was completed, and R_p was recorded at 3, 7, 12, and 17 min postchallenge. ASF volume and osmolality were measured before challenge and between 0-2, 4-6, 9-11, and 14-16 min after challenge.

Rp, ASF volume, and ASF osmolality before and after WAC. After recording baseline R_p in five dogs (22.9 ± 1.5 kg) using dry, room-temperature air, WAC was performed and R_p was recorded at 4, 8, 12, and 17 min postchallenge. ALF was sampled for 2 min before and at 1-3, 5-7, 9-11, and 14-16 min after WAC. Placement of the pledget after WAC was delayed 1 min in order to dry the bronchoscope before passing the pledget through its port.

ASF volume and osmolality during DAC and the effect of conditioned air on post-DAC Rp. Before DAC was done in seven dogs (17.6 ± 0.8 kg), baseline values for R_p, ASF osmolality, and ASF volume were recorded during ventilation with warm humidified air. Unlike the preceding experiments, a pledget was placed on the bronchial mucosa during the 5-min DAC to measure ALF volume and osmolality. After DAC, R_p was recorded at 30 s, and 5, 10, 14, and 18 min after the DAC, while ventilating the wedged segment with dry, room-temperature gas. ASF samples were obtained at 2-4, 7-9, 11-13, and 15-17 min post-DAC. A similar experiment was repeated using six dogs, but this time warm humidified air was used to ventilate the peripheral airways during the postchallenge period, and the port of the bronchoscope was dried immediately before sampling.

Additional ASF samples (n = 15 lobes, 9 dogs) were collected during DAC, and these data were combined with those from the preceding study for correlation analysis (n = 27).

Statistical Analyses

In vivo data were analyzed using the Friedman repeated measures analysis of variance (ANOVA). The Student-Newman-Keuls test applied to ranks was used to compare individual treatment means. Dunnett's test was used to compare individual treatment means to a baseline value. Either a Mann-Whitney U test or a Kruskall-Wallis one-way ANOVA was used to compare measurements of sample weight and osmolality made using the model airway. Spearman's rank correlation (r_s) was used to test for correlations between changes in R_p (R_pmax) and changes in ASF volume ( volume) and osmolality ( osmolality) that occur immediately after hyperventilation with either dry or warm humidified air. All values were expressed as the mean ± SEM. Statistical significance in all cases was judged at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of the Filter Paper Pledget on the Measurement of ASF Osmolality

During 200 ml/min ventilation of the model airway, the osmolality of the fluid recovered by pledget was consistently higher than that recovered by pipette, regardless of whether the pledget was placed in the airway during or after the exposure, and whether warm wet air or cool dry air was used for ventilation (Figure 1A). Additionally, pipette (p = 0.001) and pledget (p = 0.029) sample osmolalities collected during the 5-min 200 ml/min low-flow exposure were greater than similar samples collected after an identical exposure.

View larger version (20K): [in this window] [in a new window]   Figure 1.   Osmolality (A) and volume (B) of standard fluid samples (290 mOsm/kg) recovered from a model airway either by pipette or filter paper pledget (see METHODS). Standard fluid samples were exposed to 5% CO2 in air for either 2 min at 200 ml/min (200/2), for 5 min at 200 ml/min (200/5), or for 5 min at 2,000 ml/min (2,000/5). Samples were collected either after exposure to dry air (closed symbols with dot), during exposure to dry air (closed symbols), or after exposure to wet air (open symbols). Values represent mean ± SEM. *p < 0.05, **p < 0.01 comparing pledget to pipette; +p < 0.05, p < 0.01 comparing wet to dry; p < 0.01 comparing 200 ml/min to 2,000 ml/min; §p < 0.05 comparing "during" to "after"; xp < 0.01 comparing 2 min to 5 min. 200/2 dry: n = 8; 200/2 wet: n = 9; 200/5: n = 7; 2,000/5 wet and dry: n = 8.

Sampling technique did not affect the osmolality of samples measured either during (p = 0.083) or after (p = 0.442) a 5-min exposure when flow rate was increased to 2,000 ml/min. There was no difference between pipette (p = 0.130) and pledget (p = 0.279) samples collected during and after this exposure (Figure 1A).

Finally, sampling technique did not affect the osmolality of fluid collected after a 5-min exposure to warm humidified air delivered at 2,000 ml/min (p = 0.710). Samples obtained by either pipette (p < 0.001) or pledget (p < 0.001) indicated that warm humid air reduced osmolality during a 2,000 ml/min exposure compared with dry air (Figure 1A).

In general, the volume of fluid on the pledget did not vary with sampling technique, length of exposure, condition of the air, or flow rate (Figure 1B).

Location of the Filter Paper Pledget in the Canine Lung Periphery

Distance of the pledget from the tip of the bronchoscope was measured in 26 experiments, and averaged 2.0 ± 0.1 cm (range, 0.5 to 3.1 cm). Based on the regression equations previously described, pledgets recovered ASF from bronchi with an estimated diameter of 2.6 ± 0.2 mm (range, 0.9 to 4.9 mm).

R_p, ASF Volume, and ASF Osmolality During Baseline Ventilation

Baseline R_p (Figure 2A: p = 0.279), ASF volume (Figure 2B: p = 0.395), and ASF osmolality (Figure 2C: p = 0.938) remained stable throughout the 35-min protocol (n = 5 lobes; four dogs). Switching from cool dry to warm humidified air did not affect either baseline R_p (Figure 3A: p = 0.974) or ASF osmolality (Figure 3C: p = 0.181), but increased ASF volume (Figure 3B: p < 0.001) (n = 12 lobes; seven dogs).

View larger version (20K): [in this window] [in a new window]   Figure 2.   (A) Rp, (B) ASF volume, and (C ) ASF osmolality during baseline ventilation with dry air. Vertical bar marks the time period during which hyperventilation challenge would normally be done. Values represent mean ± SEM (n = 5 lobes; four dogs).

View larger version (18K): [in this window] [in a new window]   Figure 3.   (A) Rp, (B) ASF volume, and (C ) ASF osmolality during baseline ventilation with either dry air (open circles) or warm humidified air (open squares). Values represent mean ± SEM (n = 12 lobes; seven dogs). *p < 0.05 compared with dry air baseline.

R_p, ASF Volume, and ASF Osmolality Before and After DAC

R_p increased at 3, 7, 12, and 17 min after DAC (Figure 4A: p < 0.001, n = 7 lobes; seven dogs). ASF volume 0-2 min after DAC was similar to the baseline value, and increased above baseline at 4-6, 9-11, and 14-16 min after DAC (Figure 4B: p < 0.001, n = 8 lobes; seven dogs). ASF osmolality increased from 320 ± 10 mmol/kg before DAC to 360 ± 11 mmol/kg 4 to 6 min after DAC (Figure 4C: p = 0.037, n = 8 lobes; seven dogs). Note that the different numbers of lobes/dogs per experiment reflect technical problems that interfered with our ability to make a measurement.

View larger version (21K): [in this window] [in a new window]   Figure 4.   (A) Rp, (B) ASF volume, and (C ) ASF osmolality before and after hyperventilation with dry air. Vertical bar marks the time period during which hyperventilation challenge was done. Values represent mean ± SEM [n = 7 (A) or 8 (B and C ) lobes; seven dogs]. *p < 0.05 compared with baseline.

The mean tracheal ASF sample volume was 4 ± 0.8 µl and tracheal ASF osmolality averaged 322 ± 11 mmol/kg. Osmolality of plasma samples averaged 292 ± 1 mmol/kg (n = six dogs).

R_p, ASF Volume, and ASF Osmolality Before and After WAC

R_p increased at 4 and 8 min after WAC (Figure 5A: p = 0.007, n = 8 lobes; five dogs). ASF volume increased above baseline at 1-3, 5-7, and 9-11 min after the WAC (Figure 5B: p < 0.001). ASF osmolality decreased from 340 ± 5 mmol/kg before to 248 ± 22 mmol/kg 1 to 3 min after WAC (Figure 5C: p = 0.004). ASL osmolality returned to baseline 5 to 7 min after WAC.

View larger version (18K): [in this window] [in a new window]   Figure 5.   Rp, ASF volume, and ASF osmolality before and after hyperventilation with warm humidified air. Vertical bar marks the time period during which hyperventilation challenge was done. Values represent mean ± SEM (n = 8 lobes; five dogs). *p < 0.05 compared with baseline.

WAC versus DAC: Comparisons and Correlations Between R_p, ASF Volume, and ASF Osmolality

Challenge condition affected R_pmax,  volume, and  osmolality (p < 0.001). Although DAC increased R_p by 132 ± 41% (Figure 4A) compared with only a 45 ± 18% increase after WAC (Figure 5A), R_pmax was different only at 12 min postchallenge (p < 0.05).  Volume was less (p < 0.05) 0 to 2 min after DAC when compared with 1 to 3 min after WAC (Figures 4B and 5B).  Osmolality 0 to 7 min after DAC was greater (p < 0.05) than after WAC (Figures 4C and 5C). However,  osmolality was not correlated with R_pmax after either DAC (r_s = 0.036, p = 0.905, n = 7 lobes) or WAC (r_s = 0.167, p = 0.662, n = 8 lobes). A statistically significant relationship between  osmolality and R_pmax was seen when the DAC and WAC data were combined (r_s = 0.514, p = 0.048, n = 15 lobes) (Figure 6A). Finally,  volume was not correlated with R_pmax after either DAC (r_s = 0.214, n = 7, p = 0.602) or WAC (r_s = 0.548, n = 8, p = 0.139).  Volume and R_pmax were not correlated even after combining the DAC and WAC data (r_s = 0.321, n = 15, p = 0.235) (Figure 6B).

View larger version (12K): [in this window] [in a new window]   Figure 6.   Correlations between (A)  osmolality and Rp, and (B)  volume and Rp before and after hyperventilation with either dry air or warm humidified air. (A)  Osmolality versus Rpmax: DAC =  (rs = 0.036, p = 0.905, n = 7 lobes); WAC =  (rs = 0.167, p = 0.662, n = 8 lobes);  +  (rs = 0.514, p = 0.048, n = 15 lobes). (B)  Volume versus Rpmax: DAC =  (rs = 0.214, n = 7, p = 0.602); WAC =  (rs = 0.548, n = 8, p = 0.139);  +  (rs = 0.321, n = 15, p = 0.235).

ASF Volume and Osmolality During DAC and the Effect of Conditioned Air on Post-DAC R_p

R_pmax, ASF volume, and ASF osmolality after the DAC did not differ between the cool dry air and warm humidified air post-DAC insufflation protocols, and were combined for further analysis. R_p was increased at 0.5, 5, 10, 14, and 18 min after DAC (Figure 7A: p < 0.001, n = 12). ASF volume was decreased during and at 2 to 4 min after the DAC (Figure 7B: p < 0.001). Baseline ASF osmolality increased from 324 ± 7 to 464 ± 45 mmol/kg during DAC (Figure 7C: p < 0.001). Unlike data from the preceding experiments, ASF osmolality was not significantly elevated at any time after the DAC.

View larger version (20K): [in this window] [in a new window]   Figure 7.   (A) Rp, (B) ASF volume, and (C ) ASF osmolality during and after hyperventilation with dry air. Vertical bar marks the time period during which DAC was done. Values represent mean ± SEM (n = 12 lobes; seven dogs). *p < 0.05 compared with warm humidified air baseline.

A statistically significant relationship existed between  osmolality and R_pmax (r_s = 0.727, p = 0.009, n = 12) (Figure 8A). As before,  volume was not correlated with R_pmax (r_s = 0.474, p = 0.132) (Figure 8B). An increase in sample size confirmed the significant relationship between  osmolality and R_pmax (r_s = 0.499, p = 0.008, n = 27) (Figure 8C), and suggested that an inverse relationship exists between  volume and R_pmax (r_s = 0.355, p = 0.069) (Figure 8D).

View larger version (21K): [in this window] [in a new window]   Figure 8.   Correlation of either  osmolality or  volume recorded during DAC with Rpmax recorded after the DAC. (A)  Osmolality versus Rpmax: (rs = 0.727, p = 0.009, n = 12) and (B)  volume versus Rpmax: (rs = 0.474, p = 0.132, n = 12). After increasing sample size from 12 to 27: (C )  osmolality versus Rpmax: (rs = 0.499, p = 0.008, n = 27) and (D)  volume versus Rpmax: (rs = 0.355 p = 0.069, n = 27).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ASF kinetics, as characterized by transient changes in ASF volume and osmolality, have never been documented in vivo before, during, and after hyperventilation of the peripheral lung. In doing so, we used biopsy forceps and a filter paper pledget to recover ASF from canine bronchi averaging 2.6 mm in diameter. ASF recovered in this way represents at best a semiquantitative index of local ASF availability and osmolality. Erjefält and Persson (9) reported that filter paper discs used to sample tracheal surface fluid in guinea pigs disturbed epithelial-barrier function, sampled subepithelial fluid and solutes including macromolecules, and concentrated a plasma tracer molecule. Despite these problems, they concluded that inflammation-associated changes in the content of tracheal fluid were readily detectable using this technique.

We used a model airway ventilated with a range of flow rates and exposure times to determine if our method for collecting canine ASF provided realistic and reproducible values of ASF osmolality. For baseline flow conditions we found small but significant increases in sample osmolality attributable to the pledget, but nothing approaching the 4- to 5-fold increase in plasma tracer concentration reported by Erjefält and Persson (9). Increasing dry air flow to 2,000 ml/min markedly increased fluid sample osmolality as expected, but in contrast to baseline flow rates, sampling technique had no effect on the values measured. Using 2,000 ml/min warm humidified air eliminated this increase in osmolality regardless of the recovery technique used (Figure 1A). Except for the first trial using 200 ml/min dry air, the volumes recovered from all samples were of similar size (Figure 1B). Thus, we are confident that we can accurately measure changes in ASF osmolality with this technique.

We used a bronchoscope to measure R_p in vivo and to collect tracheal and bronchial ASF samples from anesthetized dogs. Tracheal and bronchial ASF osmolality were similar, and both were consistently higher than plasma osmolality. Our values for tracheal ASF osmolality are comparable with those reported by others (10, 11), but our values for bronchial osmolality (Figure 2C) are higher than previously reported (10). At first we believed that our use of dry air to measure baseline R_p accounted for the relatively high bronchial ASF osmolalities reported in this study. However, in a second experiment in which baseline R_p was recorded during sequential insufflation of dry and then conditioned air (Figure 3A), we found no detectable effect on baseline ASF osmolality (Figure 3C) despite an increase in ASF volume (Figure 3B). Samples recovered by pipette from a model airway ventilated for 2 min with 200 ml/min revealed that osmolality increased 8 mmol/kg when cool dry air (309 ± 2.4 mmol/kg) was used instead of warm humid air (300 ± 1.6 mmol/kg, p = 0.012) (Figure 1A). Thus, the 7 mmol/kg difference in ASF osmolality recorded in vivo after insufflation of cool dry (331 ± 4.9 mmol/kg) and warm humid air (324 ± 4.6 mmol/kg, p = 0.181) may be real, but lacks a physiological consequence with respect to R_p.

Hyperventilation with dry air increases R_p, ASF volume, and ASF osmolality in dogs (Figure 4). ASF osmolality increased by ~ 40 mmol/kg (Figure 4C). This change is considerably smaller than that predicted from measurements of pulmonary water loss made during exercise in humans (3). ASF volume increased within 4 to 6 min after DAC, suggesting an increased availability of ASF during the development of bronchial obstruction (Figure 4B). This is consistent with our hypothesis that airway water lost during hyperventilation is rapidly replaced from intra- and extravascular sources (4), making airway water loss an inaccurate measure of changes in the local osmotic environment. The increase in ASF availability may result from either a hyperpnea- or hyperosmotic-induced increase in mucociliary clearance (12, 13) or an osmotic-induced movement of water (14) through or around the bronchial epithelium. The fact that the increase in ASF volume (Figure 4B) was observed while ASF hyperosmolality (Figure 4C) was maintained suggests that enhanced mucociliary clearance may transport hypertonic ASF proximally from more distal locations. It is likely that isotonic fluid also moves down an osmotic gradient from the bronchial vasculature to the airway lumen, but this mechanism for increasing ASF volume would tend to decrease osmolality. In additon, bronchovascular leakage occurs immediately after DAC (4), and may provide fluid for the maintenance of mucosal hydration and mucociliary clearance either during or after hyperpnea. If this is correct, ASF osmolality would be markedly greater during hyperventilation, prior to the initiation of any postchallenge compensatory mechanisms.

A very different picture emerges when hyperventilation is done with warm humidified air (Figure 5). As seen in this (Figure 5A) and other studies (4, 6, 8), hyperventilation with conditioned air has a relatively small effect on R_p. It is surprising that this attenuated response to hyperpnea is accompanied by a marked increase in ASF volume and a concomitant reduction in ASF osmolality (Figure 5). This increase in volume and decrease in osmolality are not seen in samples recovered from our model airway after hyperventilation with warm humid air (Figure 1). This indicates that the bronchoscope was effectively dried before inserting a pledget, and that water was not sprayed into the airway during the WAC. Given these facts, the transient changes in ASF volume (Figure 5B) and osmolality (Figure 5C) observed after WAC suggest that under these conditions the peripheral airways either secrete hypoosmotic fluid or secrete isotonic fluid and absorb salt. A rapid compensatory absorption of water then occurs within minutes after the challenge stops.

The  ASF volume and osmolality determined from filter paper pledgets recovered after either DAC or WAC weakly correlated with RPmax. A poor but significant relationship between  osmolality and R_pmax was seen after the DAC and WAC data were combined (Figure 6A), which allows us to evaluate this relationship across two extremes of a continuous variable, i.e., evaporative water loss. No relationship was seen between  ASF volume and R_pmax (Figure 6B). Thus, postchallenge changes in ASF osmolality and volume appear to be relatively poor predictors of the magnitude of hyperventilation-induced bronchoconstriction (HIB). There are several possibilities to explain this: First, ASF samples obtained after hyperventilation may not accurately reflect hyperventilation-induced changes in volume and osmolality. However, the fact that we can make repeatable measurements of ASF osmolality in vivo that are consistent with those previously reported by others (10, 11) suggests that this is not the case. Second, we sampled airways with an estimated average diameter of 2.6 mm, whereas the site of R_p is believed to be at the level of the respiratory bronchioles and alveolar ducts (17). Thus, postchallenge changes in ASF osmolality in 2.6-mm airways may not directly stimulate peripheral airway narrowing. Finally, it is possible that changes in the volume or osmolality of any airway fluid compartment do not initiate airway obstruction.

To test the hypothesis that changes in either the ASF or the intracellular or subepithelial osmotic environment during hyperventilation stimulate airway narrowing, we measured ASF osmolality during hyperventilation. During DAC, ASF osmolality increased by ~ 130 mmol/kg (Figure 7C). However, in contrast to our post-DAC pledget studies (Figure 4), ASF volume and osmolality decreased after the DAC. It is possible that placement of the pledget during DAC stimulated fluid movement across the bronchial mucosa (9), but if this was the case, similar results should have been recorded during our post-DAC experiments. Thus, it is difficult to account for this discrepancy other than to conclude that the presence of the pledget during the DAC protects the underlying mucosa from desiccation.

Finally, the  ASF osmolality determined from filter paper pledgets recovered during DAC was strongly correlated with R_pmax (Figure 8A). Increasing sample size confirms this correlation (Figure 8C), and reveals an inverse relationship between  ASF volume and R_pmax (Figure 8D). Thus, measurements of ASF osmolality and volume made during DAC are better predictors of the magnitude of HIB than values obtained after the DAC. There are several possibilities that can explain how changes in ASF osmolality in central airways can reflect and/or affect the resistance of peripheral airways: Changes in ASF osmolality in 2.6-mm airways may (1) reflect smaller but qualitatively similar changes in the lung periphery, (2) stimulate local mediator release which modulates R_p via the bronchial circulation, or (3) increase R_p by stimulating an axon reflex.

In summary, hyperventilation with dry air increased ASF osmolality between 0-5 min after the DAC. ASF volume did not change initially, but increased between 2-10 min after challenge. During hyperventilation ASF osmolality increased markedly when compared with values recorded after the DAC. These observations implicate a postchallenge compensatory mechanism in which ASF accumulates in the airway lumen during the initial development of bronchial obstruction, and provides fluid for the maintenance of mucosal hydration and mucociliary clearance. Rapid fluid replacement probably accounts for the relatively small hyperpnea-induced increase in ASF osmolality observed in this study, compared with predictions based on models that do not consider this variable (3). In contrast to DAC, hyperventilation with warm humidified air increases ASF volume and decreases osmolality. In this case, the excess ASF that temporarily accumulates during the WAC and is then reabsorbed after the challenge may contribute to the inhibition of bronchoconstriction.  ASF volume and osmolality collected after challenge were poorly correlated with hyperventilation-induced changes in R_p. In contrast to the postchallenge data,  ASF osmolality during DAC was highly correlated with R_pmax. The fact that hyperventilation increases ASF osmolality during exposure and increases R_p after hyperventilation ends supports the hypothesis that acute changes in ASF osmolality during DAC initiate airway obstruction.