INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: adaptive response; persistent response

The use of repetitive daily exposure to assess the cumulative effects of ozone (O₃) has been a staple of clinical toxicology (1). Although virtually all these studies relied at least in part on spirometry to plot the time course of the functional response, none was specifically designed to assess small airways. The latter have long been a focus of clinical interest because of their likely role in the pathogenesis of chronic lung disease (7). Earlier experience in this laboratory suggested that O₃- induced changes in small airway function following acute exposure differed in magnitude and duration from those involving the larger central airways and the volume capacity of the lung. Whereas recovery of FVC and FEV₁ began within 30 min of the end of exposure and was complete within 24 h, small airway constriction, as measured by the reduction in isovolumetric (isoV). FEF_25-75, max50, and max75, continued to increase during the first 30 min postexposure and showed little or no evidence of remission 24 h later (8). Also, whereas the changes in FVC and FEV₁ were unimodal in distribution and skewed toward minimal responsiveness, the changes in (isoV) FEF_25-75 showed a wider distribution that may have constitute more than one mode (8). At 24 h postexposure, the reduction in (isoV) FEF_25-75 correlated significantly with the increase in the fibrinogen level in bronchoalveolar lavage fluid (BALF), used as a marker of plasma fluid exudation (10).

In testing the hypothesis that repetitive exposure to the same inhaled dose of O₃ over a 4-d period, as might occur during an episode of photochemical smog, would affect small airways disproportionately, we compared the results among both spirometric and ventilatory variables (respiratory frequency [f] and tidal volume [VT]) for evidence of adaptive and persistent responses. As is customary, we have defined an adaptive response as one characterized by a diminution in the daily effect of O₃ with repeated exposure and a persistent response as one characterized by either a depression of the baseline (spirometry) or a mean change in ventilatory pattern (f, VT) in the days following the initial exposure.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject Selection

We studied eight healthy young adults (five white males, two white and one black female), 25 to 31 yr of age. Four subjects (Numbers 1, 4, 5, and 6, Table 1) had participated in O₃ studies one or more years earlier. None of the subjects smoked. We defined nonsmoking as equivalent to a lifetime total of less than 3 pack-years plus abstinence from smoking for at least 1 yr prior to the study. All underwent a screening procedure consisting of a medical history, physical examination, spirometry and electrocardiogram (ECG) (12 leads) during rest, and submaximal exercise on a treadmill (Model 20-55; Quinton, Inc., Seattle, WA). Minute ventilation (E) was measured throughout the exercise test. Subjects were excluded from the study if they had a history of chronic respiratory or cardiovascular disease, upper respiratory infection during the past 6 wk, FVC, or FEV₁ values  80% predicted, or showed an apparent inability to sustain moderately heavy exercise for at least 30 min. Information on the women's menstrual cycle was recorded but not considered during scheduling based on our previous findings that the cycle did not affect O₃ responsiveness (9). The study and consent forms were approved by the Committee on Human Volunteers of the Johns Hopkins School of Hygiene and Public Health; informed consent was obtained from each subject.

Experimental Procedure

The study required 11 visits by each participant: once for the screening procedure and five times for each of the two regimes, O₃ and filtered air (FA), that is, four consecutive days of chamber exposure followed 24 h later by bronchoscopy. The sequence of the two regimes was randomized. Participants in previous O₃ studies were exempted from the screening procedure. Three or more weeks separated the O₃ and FA regimes. All subjects who began the study completed it. A single exceedance of the 1 h ambient O₃ standard occurred during the week that subject 5 was administered FA; he was cautioned to minimize outdoor activities. Prior to exposure, subjects were instructed to withhold vitamin supplements for 3 d and avoid caffeine-containing beverages on the morning of exposure. The exposures lasted 130 min. The first 120 min consisted of alternating 30-min periods of rest and exercise. The exposures ended with a final 10-min rest period.

The exercise level was set by controlling the speed and inclination of the treadmill. The level selected for each subject was based on the results of the screening submaximal exercise test. Our objective was to induce a level of E equivalent to about 8 × FVC, thereby normalizing exposure to an index of lung size. E was measured during the 8th, 18th, and 28th min of each exercise period with the subject wearing a nose clip and breathing through a one-way valve connected to a dry-gas meter. Respiratory frequency (f) was counted visually from the dry-gas meter gauge; average tidal volume (VT) was calculated from E and f. Chamber temperature was recorded during the procedure and used to convert E to BTPS. If necessary, the work load was adjusted to maintain the desired E. The procedure was intended to control E and allow adjustments in VT and f.

Heart rate was monitored electronically throughout exposure (Model Exersentry 3A; Computer Instruments Corp., Monroeville, PA), and arterial oxyhemoglobin saturation (Sa_O2) was monitored with pulse oximetry (Nellcor-200; Nellcor, Hayward, CA) during the rest period only.

FEF_25-75, max50, and max75 were used to assess changes in small airway caliber. All values following exposure were adjusted for any changes in FVC that occurred, so that the before-after comparisons were isovolumetric. Our method of adjusting volume from the spirometric tracing and the assumptions underlying the method were described earlier (8, 11).

Subjects performed spirometry three times: before entering the exposure chamber (control), just before leaving the chamber (endexposure), and 25 min after leaving the chamber (25 min postexposure); all postexposure values reported are means of measurements made at endexposure and 25 min postexposure. On Day 5, the subjects underwent bronchoscopy for the measurement of peripheral airway resistance (Rp) and for bronchoalveolar lavage. Spirometry was measured in the bronchoscopy suite before and after bronchoscopy.

Spirograms were obtained in triplicate in accordance with the recommendations of the American Thoracic Society. The measurements were made with an 8-L water-sealed spirometer (Warren Collins, Inc., Braintree, MA) provided with a potentiometer connected by an analog digital converter to an IBM PC. The spirometer was calibrated daily and temperature was recorded before each set of measurements. The spirograms were reviewed visually for quality; any marred by coughing were eliminated. All spirometric variables were temperature corrected automatically and calculated using a software program developed locally. The forced expiratory effort producing the highest sum of FVC + FEV₁ is reported. For the maximum flow-volume curves, the volume signal from the potentiometer was differentiated electronically to give flow; both signals were recorded using the same software.

Prior to bronchoscopy on Day 5, subjects inhaled 5 ml of aerosolized 4% lidocaine (DeVilbiss nebulizer no. 646, Somerset, PA) and had a balloon-tipped catheter passed by nose to the lower esophagus to measure local pressure; the latter was used as a surrogate for pleural pressure. Subjects were than placed supine and trained to breath-hold at their functional residual capacity (FRC) after three deep breaths; reproducibility of FRC during breath-holding was checked by esophageal pressure. All subjects were premedicated with 100 µg of fentanyl given intravenously. Additional topical anesthesia was achieved by instilling 2% lidocaine through the bronchoscope. Four randomly chosen subjects also received 0.6 mg of atropine intravenously. Before the bronchoscope was wedged, the working channel was cleaned with a brush and a #5 FR double-lumen catheter was inserted. The bronchoscope was wedged into the anterior segment of the right upper lobe, with care taken to return to the same site on each subject. Five percent CO₂ in air was administered at a rate determined by a mass flow controller (Sierra, Carmel Valley, CA) through one lumen of the catheter (B) while pressure at the tip of the bronchoscope (PB) was measured through the other lumen. B was increased incrementally from 200 ml/min to 500 ml/min or until PB reached 16 cm H₂O, then decreased to 100 ml/min and finally returned to 200 ml/ min. Each flow was maintained until PB had stabilized (approximately 5 to 15 s); the subject was then instructed to take three deep breaths and breath-hold at FRC after the third breath, while the steady-state PB was recorded. The breath-holding maneuver was repeated at least twice at each B. Rp was calculated as the average PB/B at each static flow and averaged over five flows. The measurement has also been referred to as collateral resistance (R_coll) (12, 13).

After Rp was measured, the bronchoscope was withdrawn to the trachea, additional 2% topical lidocaine was given topically, if needed, and the bronchoscope was repositioned in the right middle lobe. The lobe was lavaged with five aliquots (20 ml each) of normal saline warmed to 37° C. The returns were combined and kept on ice until the end of the procedure, when small amounts were removed for determination of cell counts, cell viability, and differential staining. The samples were then centrifuged at 600 × g for 15 min at 20° C and the supernatants subdivided for measurement of albumin, fibrinogen, and kinins.

The techniques for the differential cell counts and biochemical assays were described previously (10). The kinin assay does not distinguish between bradykinin and lysylbradykinin on a molar basis.

Environmental Chamber

The details of the exposure system were described previously (10). A free-standing chamber measuring 2.7 × 2.1 × 2.4 m, with 10-cm-thick insulated sides, floor, and ceiling and lined on the interior with anodized aluminum, was used for exposure. The chamber could accommodate two subjects and was furnished with a spirometer, chairs, and a treadmill controlled from outside (Model 20-55, Quinton, Inc., Seattle, WA). Outdoor air was supplied to the chamber air-purifier through the central air-conditioning system that filtered coarse particulate matter and controlled air temperature and relative humidity thermostatically. The chamber flow rate was 5.6 m³/min (23.7 air changes/h, one-pass design). Static pressure within the chamber was maintained slightly below (0.1 cm H₂O) that of the surrounding laboratory by adjustment of the supply damper. The mean chamber temperature and relative humidity were 21.4 ± 0.2 (SD)° C and 46.3 ± 3.5% during FA treatment, and 21.4 ± 0.2 and 43.9 ± 2.4% during O₃ treatment.

Ozone was generated by electrical arcing (Model G1-L Ozone Generator, PCI Ozone Corp., West Caldwell, NJ) of 100% oxygen and mixed with purified air entering the chamber. The concentration of O₃ in the chamber was monitored with an ultraviolet photometer (Dasibi 1003 AH; Glendale, CA) at two sites near the breathing zones for the treadmill and chairs. All tubing to the monitor was Teflon; the monitoring line valves were stainless steel. O₃ monitoring was continuous during both FA and O₃ exposures and was recorded every minute. The concentration per exposure was based on the mean of five evenly spaced measurements made every 20 min after the first 5 min. The mean (± SD) concentration during O₃ exposure was 0.254 ± 0.001 ppm, and during FA exposure was 0.003 ± 0.001 ppm.

Statistical Analysis

We used the principal components statistical method (14) to group the three isovolumetric spirometric measures of small airway function (FEF_25-75, max50, max75), thereby forming a single value for each subject. The symbol used in the text to designate the value is SAW_grp. Our rationale for selecting this method along figures showing the results for the individual components of SAW_grp is presented in the Appendix .

Analysis of longitudinal data. Because repeated measures on the same subject violate the assumption of independence in linear regression, we used a "mixed effects" (combined fixed and random effects) regression model to analyze the longitudinal data on spirometric and ventilatory variables:

Outcome = ₀ + ₁*O₃ + ₂*day + ₃*O₃* day + b₁* subject

where ₀ was the intercept, the explanatory variables included O₃ treatment modeled as a binary variable, day of exposure as a categorical variable, O₃*day as an interaction, and b₁* subject, the study population, as the random effect (15). The analyses were performed on untransformed data and included the following: differences between FA and O₃ treatments: a, overall regression coefficients; b, paired points; differences within each treatment: c, daily pre- to postexposure changes to assess adaptive responses; d, day-to-day changes in baseline (spirometry) or mean value (f and VT during exercise) to assess persistent responses. Significant changes in the components of the model are cited in the figures. Significance was taken as p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

During exercise, overall E averaged 38.6 ± 2.5 (SEM) L/min on FA and 39.1 ± 2.1 L/min on O₃. The mean values on individual days ranged between 38.1 ± 2.8 L/min and 39.1 ± 2.6 L/min on FA, and between 38.1 ± 2.1 L/min and 39.8 ± 2.2 L/min on O₃. No significant trends were noted over the 4-d periods. The overall targeted E was 39.5 ± 2.3 L/min.

Spirometry

The coefficients of variation for all spirometric variables including the small airway group (SAW_grp) are shown in Table 2; the values were based on FA breathing. Similar gradients in measurement error have been reported before (11, 16). The coefficient of variation was highest among the small airway variables, tending to increase as lung volume and flow rate fell. Such large coefficients clearly limit the utility of these measurements for detecting functional change.

The time course of the adaptive response differed slightly among spirometric variables (Figure 1). The maximal mean reductions in FVC and FEV₁ occurred on Day 2 of exposure to O₃; 7.7 ± 4.5% (SEM) (p < 0.01, model components b, c, and d) and 9.1 ± 5.7% (p < 0.01, model components b, c, and d), respectively. By Day 4, these daily changes were negligible. The maximal mean reduction in the SAW_grp occurred on Day 1 of exposure, 6.7% (p < 0.05, component b), and thereafter dissipated rapidly.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Adaptive response to repeated O3 exposure. Mean, SEM. Components of the mixed effects statistical model include (1) the differences between filtered air and ozone: a, overall; b, paired points; and (2) the differences within filtered air and ozone: c, day to day; d, within day. *p < 0.05, **p < 0.01; FA = filtered air; O3 = ozone; D = day.

The only spirometric variable to show a persistent response with repetitive O₃ exposure was the SAW_grp (Figure 2). By Day 2, the SAW_grp baseline was depressed by 6.9 ± 3.5% (p = 0.06, component c) and remained depressed by slightly more than 8% (p < 0.05, component c) over the next 3 d. This persistent effect exceeded in magnitude the maximal pre- to postexposure reduction in the same variable seen on Day 1 (Figure 1).

View larger version (17K): [in this window] [in a new window]   Figure 2.   Persistent effect of repeated O3 exposure on small airway function (SAWgrp). Mean, SEM. See legend to Figure 1 for description of components of the statistical model. FVC and FEV1 baselines were not significantly depressed by O3 (model component c). The small but significant differences between FA and O3 for both FVC and FEV1 (model components a and b) were due in great measure to the unexplained upward shift in their baselines over the course of FA breathing. *p < 0.05; **p < 0.01; FA = filtered air; O3 = ozone; D = day.

Breathing Pattern

To test for evidence of an adaptive response affecting the pattern of breathing, we compared the magnitude of within-day changes in f and VT across the 4 d of exposure. The changes occurred between the 10th and 60th minutes of exercise, that is, between the first and last of six measurements made during exercise (Figure 3).

View larger version (23K): [in this window] [in a new window]   Figure 3.   Adaptive changes in respiratory frequency (f) and tidal volume (VT) during exercise. Mean, SEM. See legend to Figure 1 for description of components of the statistical model. Mean ventilation (L/min) was maintained at about 8 × FVC per subject. The values shown are the first and last of six measurements made at 10 min intervals during 1 h of intermittent exercise. The SEMs of these mean within-day changes on Days 1-4 were as follows: for f on FA (1.0, 0.9, 1.0, and 0.8 BPM) and on O3 (1.4, 2.7, 1.7, and 1.4 BPM); for VT, the SEMs were 0.1 L on both treatments, all days. *p < 0.05; **p < 0.01; BPM = breaths/min; L = liters.

The maximal mean increase in f during O₃ exposure occurred on Day 2, +6.6 ± 2.7 (SEM) breaths/min (BPM) or +30% (p < 0.01, component d). By Day 4, the change was indistinguishable from that seen with FA. (Respiratory frequency tended to increase slightly over the course of exercise on FA. On Day 4, this increase was similar for FA and O₃ although absolute f was higher with O₃.) The maximal mean reduction in VT during O₃ exposure occurred on Day 3, 0.36 ± 0.09 L or 19% (p < 0.01, component d). Evidence of an adaptive response appeared the next day: 0.22 ± 0.08 L or 11% (p < 0.05, component d).

To test for evidence of persistent changes in breathing pattern attributable to O₃, we compared the mean values for f and VT on Day 1 (based on all six measurements) against the means on the remaining 3 d (Figure 4). In effect, the pattern of breathing on the first day of exposure constituted the "baseline." (The means on Day 1 were similar for FA and O₃. Either could have served as baseline and yielded the same outcome. We selected the O₃ values to conform to the analysis carried out on the spirometric variables, as in Figure 2.)

View larger version (20K): [in this window] [in a new window]   Figure 4.   Persistent effect of repeated O3 exposure on respiratory frequency (f ) and tidal volume (VT) during exercise. Mean, SEM. See legend to Figure 1 for description of components of the statistical model. A more rapid, shallow pattern of breathing was established by Day 2. Each point represents the mean of all six measurements made that day. *p < 0.05; **p < 0.01; FA = filtered air; O3 = ozone; D = day; BPM = breaths/min; L = liters.

On Day 2 of O₃ exposure, mean f was elevated by +4.7 ± 2.7 BPM or +22% (p < 0.01, component c), then gradually tapered off to +3.2 ± 1.4 BPM or +15% by Day 4 (p = 0.06, component c). Mean VT was reduced by 0.2 ± 0.1 L or 14% (p < 0.01, component c) on Day 2 and remained at essentially this level thereafter. This persistent pattern of more rapid, shallow breathing is foreshadowed in Figure 3: note that beginning with Day 2, the first of the six measurements of f and VT were at new levels compared with Day 1, further evidence of "carryover" effects on the control of breathing.

BALF Content

The total cell count (Day 5) tended to be higher after O₃ than FA, but the difference was not significant (p = 0.26 [Table 3]). The neutrophilic (polymorphonuclear, PMN) count was significantly higher after O₃: + 61% (p = 0.04). None of the cell types, expressed as percentages of total cells, differed significantly in the two circumstances.

The mean levels of albumin, fibrinogen, and kinins were not significantly different after O₃ and FA. Nonetheless, these findings were notable in two respects: O₃-FA differences in all three variables were strongly interdependent within subjects and varied widely in magnitude among subjects. Thus, a, the correlation coefficients (r) for the O₃-FA differences among the three variables, ranged between 0.84 and 0.93 (p < 0.01); b, the percentage increases in albumin, fibrinogen, and kinins in the two subjects with the most marked exudative responses, exceeded the mean changes for the group, respectively, by 10-fold, 15-fold, and 2-fold (Subject 2), and by 45-fold, 17-fold, and 7-fold (Subject 4).

Correlations between the changes in exudative markers and PMN counts were negligible (r = 0.05 to 0.20). Correlations between the changes in both exudative markers of inflammation and PMN counts and the level of depression of the small airway baseline (SAW_factor1; see Appendix for derivation) on Day 5 were inverse and weak (r = 0.30 to 0.46); the highest value, 0.46 (p = 0.26), was found between the changes in fibrinogen and SAW_factor1). Correlations between the changes in PMN counts and changes in both f and VT on Day 5 were moderately strong (r = 0.52 and 0.65, respectively), but appeared paradoxical, that is, as PMN influx increased, the degree of rapid shallow breathing decreased. A similar paradoxical association between the degree of neutrophilia and the decline in FEV₁ has been seen following acute O₃ exposure (17, 18).

Peripheral Airways Resistance

Rp, measured in tandem with bronchoalveolar lavage (BAL), appeared unaffected by repetitive O₃ exposure. Atropine was withheld in four subjects prior to bronchoscopy with no apparent effect on the outcome. The respective Rp values in cm H₂O/ml/min (mean, range) after O₃ and Fa were as follows: with atropine pretreatment, 0.027 (0.002 to 0.066) versus 0.027 (0.003 to 0.063); with atropine withheld, 0.014 (0.003 to 0.038) versus 0.023 (0.003 to 0.069).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Functional Response

Repetitive exposure to 0.25 ppm O₃ over a 4-d period elicited two patterns of response, adaptive and persistent. All spirometric and ventilatory variables underwent adaptive changes. The changes varied in day of onset. The SAW_grp, a composite of (isoV) FEF_25-75, max50, and max75, was first to adapt (Day 1), followed 1 d later by FVC, FEV₁, and by f, and VT on the last day of exposure (Figures 1 and 3).

Adaptive responses to O₃ involving FVC, FEV₁, symptoms of respiratory irritation, bronchial reactivity, exercise performance, and, more recently, a number of inflammatory markers are well documented (1, 4). Functional adaptation, as measured by FEV₁, has been seen despite cellular and biochemical evidence in BALF of ongoing inflammation, and of neutrophil infiltration in bronchial mucosa (6). To our knowledge, volume-corrected spirometric measures or other indices of small airway function have not been examined before in a similar way. Evidence of an adaptive change in breathing pattern with repetitive O₃ exposure, particularly of f, was first described by Foxcroft and Adams (19, Figure 2). (That acute exposure elicits progressive tachypnea during exercise at concentrations at or above 0.20 to 0.25 ppm is also well documented [1, Table 7-1].) The onset of an adaptive response appears to be influenced by both the level of exposure and sensitivity of the individual to O₃. More intense exposure and increased sensitivity are likely to delay the onset (1, 20).

The small airway composite measure alone among spirometric variables displayed a persistent preexposure baseline depression, beginning with Day 2 (Figure 2). This dichotomy between the SAW_grp and both FVC and FEV₁ accords with earlier results seen following acute O₃ exposure, namely, that FVC and FEV₁ recovered more rapidly than (isoV) max (9, 10). Persistent depression of the FVC and FEV₁ baselines have generally been uncommon or marginal during repetitive O₃ exposure at concentrations  0.4 ppm (2, 3, 5, 20, 22).

Our method of adjusting the three components of the SAW_grp for O₃-induced changes in FVC (thereby solving for intrinsic changes in small airway caliber) relies on two assumptions, namely, that O₃ exposure has little or no effect on either residual volume (RV) or lung elastic recoil. To our knowledge, neither assumption has been tested experimentally with repetitive exposure. Hazucha and coworkers (23) concluded that the reduction in FVC following acute O₃ exposure (0.5 ppm × 2 h) was not attributable to the small increase in RV (+11%, p < 0.05) or the tendency toward an increase in elastic recoiling force they observed. Beckett and coworkers (0.4 ppm × 2 h) (24) and our group (0.35 ppm × 130 min) (8, 9) found no significant change in RV following acute exposure. It is to be noted that an increase in RV, as might occur if partially obstructed small airways closed prematurely during forced expiration, would act to diminish the calculated reduction in isovolumetric flow. Similarly, an increase in elastic recoil would, if transmitted to the outer walls of intrapulmonary airways, act as a distending force to preserve maximal flow at low lung volumes (25).

Although we did not determine the time required for the small airway dysfunction to remit, we suspect it is tied to complete remission of the inflammatory process, as discussed in the next section. Whether the dysfunction may serve as a forerunner of more permanent loss is an open question. Permanent loss might follow in the event that O₃ exposure is repeated over extended periods of time or individuals are unusually reactive to O₃, as Subjects 2 and 4 appeared to be. Recently, Künzli and coworkers (26) reported finding a significant correlation between estimated lifetime exposure to ambient O₃ among college freshmen (lifetime residents of California) and impaired small airway function. The latter was assessed spirometrically, based on FEF_25-75% and FEF_75%; FVC and FEV₁ were unaffected. Additional epidemiological and field studies to test possible intermediate and long-term effects of photochemical air pollution on small airways appear indicated.

Inflammatory Response

In our study of the acute effects of O₃, the levels of albumin, fibrinogen, kinins, and PMN (both absolute number and percentage of total cell count) were significantly elevated 24 h postexposure (10). In the present study, only the absolute number of PMN remained elevated by Day 5. In the earlier study, the changes in fibrinogen, a marker of plasma transudation, and (isoV) FEV_25-75 were inversely and strongly correlated (r = 0.88; p = 0.001; n = 8); the spirometric measurement had also been made 1 d postexposure and, in effect, represented a "persistent" change of at least 24 h. In the present study, the changes in fibrinogen correlated inversely but weakly with the degree to which the small airway baseline, expressed as SAW_factor1, was depressed (r = 0.46; p = 0.26; n = 8). Together, these findings are consistent with a general waning of the exudative component of the inflammatory response following repetitive O₃ exposure, as reported by other investigators (4). The exceptionally high levels of exudative markers in two subjects (Numbers 2 and 4) also conform to previous evidence of wide differences in inflammatory activity among healthy subjects following repetitive exposure (5, Figure 3).

Mechanisms

We propose that both the adaptive and persistent functional responses were byproducts of inflammatory changes set in motion on the first day of exposure. The former reflects the protective action of inflammation and the latter reflects the functional inefficienciesor costsassociated with that protection. We use the term "inflammation" broadly to include increased vascular permeability, cellular influx, release of mediators, and any responses these primary events may entrain.

Although the acute inflammatory response is emblematic of tissue injury, it also is viewed as an adjustment that helps confine injury and defends against further environmental challenge (27). Plasma exudate contains elements that can interrupt the cascade of chemical reactions associated with lipid peroxidation (28), scavenge reactive oxygen species (albumin, a traditional marker of transudation, is, itself, an effective antioxidant [29]), limit tissue injury, and stimulate repair and regeneration (31). Mucorrhea and increased airway smooth muscle tone, frequent accompaniments of inflammation, may interfere mechanically with local gas flow and additional O₃ transport to inflamed sites. That the epithelial lining fluid, as a result of inflammatory changes in volume and composition, is likely to be a more effective barrier against oxidative stress is implicit in the O₃ transport model (32, Figures 2 and 4). Several of these factors have indeed been cited as promoters of functional and symptomatic adaptation to O₃.

Admittedly, the generally weak correlations between inflammatory and persistent functional changes provide little support for our proposition that the latter reflect the inefficiencies of an otherwise protective response. Nonetheless, we do not regard them as voiding the proposition. Although useful diagnostically, bronchoalveolar lavage is an invasive procedure that cannot provide an adequate accounting of a process as complex and nonlinear in its dynamics as inflammation. Estimates of the interdependencies between inflammatory and functional responses are therefore subject to considerable uncertainty. Insofar as functional and inflammatory responses follow different time courses, the apparent strength of their interdependency will be sensitive to the timing of each measurement.

Tachypnea

Stimulation of afferent vagal fibers (bronchial C-fibers) is considered the chief basis for the rapid, shallow breathing (during exercise), chest discomfort, and reduced inspiratory capacity associated with acute O₃ exposure (23, 33, 34). It is therefore difficult to reconcile the persistent tachypnea seen during repetitive exposure with the absence of analogous changes in FVC and FEV₁, unless another mechanism keyed solely to the control of breathing is invoked. A plausible candidate for this role is a change in the viscoelastic properties of the distal lung, triggered by the tissue injury and inflammation that accompany the first day's exposure to O₃. The rationale for this potential mechanism was developed decades ago to explain breathlessness during exercise.

The pattern of breathing used to achieve a specific E is highly variable (35). Campbell (36) postulated that information about respiratory muscular and mechanical behavior is processed unconsciously by the cortex, which exerts "unconscious control" of the breathing pattern. If, as a consequence of O₃-induced injury and inflammation, tissue viscance (Vti) were increased out of proportion to any narrowing of the large conducting airways, more rapid, shallow breathing would likely follow, especially with exercise (37, 38). This adjustment in breathing pattern should mitigate respiratory discomfort or breathlessness arising from the altered mechanical behavior of the distal lung.

The premise about the O₃-induced change in Vti is untested. Certainly, constriction of large, central airways, as reflected in an increase in airway resistance (SRaw or Raw), is not a prominent feature of the response to acute exposure (17, 23, 34) and should not exact a high cost in work during tachypnea. Moreover, with repetitive exposure, SRaw adapts readily and has shown little or no baseline elevation (5, 39). In our study, the underlying adjustment of the respiratory pattern did not occur until Day 2 when, presumably, tissue injury and inflammatory changes were more fully established. The proposed change in Vti does not imply a parallel change in Rp, which was unaffected. The two measures may be governed by anatomically distinct contractile elements, as evidenced by their dissimilar responses to histamine and leukotriene C₄ aerosols (13).

Peripheral Airway Resistance

As noted above, Rp was unaffected by repetitive O₃ exposure notwithstanding the evidence of persistent small airway narrowing and inflammation of the peripheral lung. This dichotomy between collateral and small airway responses was also seen after acute O₃ exposure (10). Following the acute exposure, all subjects prior to bronchoscopy received atropine, a cholinergic antagonist, and lidocaine, a topical anesthetic. In the present study, atropine was withheld from half of the subjects without affecting the outcome. It would appear that either the dose of O₃ to the airways responsible for Rp was negligible relative to the dose to small airways, the former airways were highly resistant to O₃, or that lidocaine premedication abolished any increase in Rp that might have occurred. We favor the first of these possibilities. Consistent with our findings is the recent observation that in anesthetized, mechanically ventilated dogs, 0.2 ppm O₃ administered for 6 h by endotracheal tube had no effect on Rp unless the animals were pretreated with probenecid, an inhibitor of endogenous antioxidant transport. In contrast, pulmonary resistance increased in both untreated and pretreated animals (40).

Collateral channels having the structural characteristics of respiratory bronchioles and alveolar ducts appear to be chiefly responsible for Rp (12). They represent diffusion pathways in normal lungs. Small airways, which dominate maximal expiratory flow at intermediate to low lung volumes, are likely to include subsegmental bronchi and membranous bronchioles. Opposing views have been expressed about whether collateral channels have a significant role in gas exchange in the normal lung (12, 41). Our results are consistent with their having a minor role, especially among healthy young adults.

In summary, we observed two patterns of functional response among healthy, young adults exposed repetitively to O₃, adaptive and persistent. All spirometric variables, as well as f and VT, underwent adaptive changes. Spirometric evidence of persistent change was confined to small airways; rapid, shallow breathing during exercise also persisted. Neutrophilia in BALF was evident 1 d following the end of O₃ exposure. We suggest that both types of functional response are linked causally to inflammation. The adaptive component is attributable at least in part to a reduction in local tissue dose during repetitive exposure that is likely to result from the biochemical, mechanical, and morphological changes set in motion by inflammation. The persistent component represents the inefficiencies incurred through inflammation. Whether the persistent small airway dysfunction is a forerunner of more permanent change in the event that oxidant stress is extended over lengthy periods of time is unclear.