INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE EFFECTS OF OZONE, one of the national ambient air-quality "criteria" air pollutants, have been studied extensively in both humans and animals. The respiratory system has been the primary focus of these investigations, since it is the first and the principal site of action of this gas. Exposure of intermittently exercising volunteers to ozone elicits both symptomatic and functional responses, with a large range of interindividual variability. Typical symptoms include shortness of breath, substernal soreness, pain on deep inspiration, and cough. The overall severity of symptoms correlates with average spirometric changes [forced vital capacity (FVC) and forced expiratory volume in 1 s (FEV₁)], which reflect the individual's inability or unwillingness to take a full inspiration rather than a change in lung mechanics (11, 15). Subjects with the greatest spirometric decrements almost invariably complain of pain on deep inspiration. These observations suggest that nociceptive mechanisms might be involved in ozone-induced inhibition of maximal inspiration (11, 15). If this were so, pain suppression or inhibition by analgesics should acutely reverse ozone-induced lung functional impairment (7, 13). Moreover, if activity of the endogenous opioid system were important in modulating the respiratory response to ozone, blockade of this system should result in greater decrements in pulmonary function. In previous work (15), we have postulated that inhibition of maximal inspiration involves bronchial C fibers and is reflex in origin. Studies by Schelegle and colleagues (21) in dogs confirmed experimentally that afferent vagal C fibers play a major role in initiating reflex effects and modulating respiratory responses to ozone.

Our primary hypothesis is that the reduction in total lung capacity and vital capacity provoked by ozone exposure is related to inhibition of inspiration caused by nociception on deep inspiration. A related hypothesis is that individuals with little response to ozone have elevated baseline plasma concentrations of -endorphins or endogenously released -endorphins triggered by ozone exposure (with exercise), or both, compared with strong ozone responders. Sufentanil and naloxone were employed in this study to manipulate nociceptive neural pathways and the endogenous analgesic system, respectively. Finally, we wished to explore the possibility that intrinsic individual (cutaneous) pain sensitivity might be predictive of ozone responsiveness as assessed by standard spirometric tests.

	METHODS

Top Abstract Introduction Methods Results Discussion References

This study was approved by The Committee on the Protection of the Rights of Human Subjects of the University of North Carolina at Chapel Hill. Male and female subjects were nonsmokers, 18-59 yr of age. All subjects were in good general health, with normal values in peripheral complete blood count and SMA-20, and without cardiac or pulmonary disease. Subjects were recruited from the general population by newspaper advertising. All subjects were given training in spirometry and treadmill exercise. During treadmill exercise training, the exercise load necessary to cause 17.5 l/m² body surface area (BSA) minute ventilation was determined. Because relatively few older individuals exhibit vigorous spirometric responses to ozone exposure, subjects were age stratified and classified as young if <35 yr of age and as middle-aged if 35 yr of age.

The subjects were classified as "strong" or "weak" responders to ozone inhalation 2 wk before entry into this study based on the measured decline in FEV₁ induced by a 1.5-h exposure to 0.42 parts/million (ppm) ozone with 20-min exercise periods alternating with 10-min rest periods. Strong responders to ozone were defined as those who experienced a >15% decline from baseline in FEV₁. Weak responders were defined as those who experienced no more than a 5% decline in FEV₁ with ozone exposure. Subjects that did not meet either of these spirometric criteria were excluded from this study. Overall, there were 20 weak and 42 strong responders to ozone. As anticipated, young subjects (<35 yr of age) were more likely to be strong responders (79%) than were older subjects (33%).

All subjects were allowed a liquid breakfast and had vital signs measured on arrival. Telemetry electrocardiographic electrodes were placed for continuous cardiac monitoring, and a preexposure symptom questionnaire was completed. Female subjects underwent a urine pregnancy test. Cutaneous pain thresholds and tolerance to heat and cold were quantified by the Marstock method (22), which involves presenting thermal stimuli to the subject with a thermal probe on the volar forearm. Pain threshold and tolerance were determined by the method of limits. Threshold was defined as the temperature at which an individual notes a stimulus to be either hot or cold. Tolerance was defined as the highest or lowest temperature an individual was willing to endure. Stimulus order and intensity were randomized under computer control. To diminish subject bias, some stimuli were presented twice.

An intravenous catheter was inserted in an antecubital vein, and a 10-ml blood sample was obtained at appropriate intervals. Plasma -endorphin levels were determined by radioimmunoassay performed by Roche Biomedical Laboratories, Burlington, NC.

Baseline spirometry was then performed. Spirometry included routine forced expiratory tests done with a Cardio Pulmonary Instrument rolling-seal spirometer or a Medical Graphics model 1085 system.

All exposures were conducted in a 4 × 6 × 3.2-m stainless steel chamber maintained at 21°C and 40% relative humidity. Chamber atmosphere was maintained by continuous reconditioning and recirculation of the chamber air through high-efficiency particle filters. During ozone exposures, nitric oxide, nitrogen dioxide, and sulfur dioxide concentrations were typically <0.02, 0.005, and 0.005 ppm, respectively. A detailed description of the exposure chambers has been published (23).

Each subject underwent two 2-h exposures, separated by at least 1 wk. By design, ~25% of subjects (randomly chosen) were exposed to air during both exposure days as a formal control group. The other 75% of subjects were exposed to 0.40 ppm of ozone during both exposure days. The investigators collecting data were blinded to the type of exposure.

During each exposure period, exercise was performed on an exercise treadmill by using speeds of 3.2-3.7 miles/h (equivalent to a moderate-to-brisk walk), at a grade of 0-5%. The amount of exercise necessary to elicit 17.5 l/m² BSA minute ventilation had been determined for each subject during his or her initial training before the subject was screened for ozone responsiveness. Each exposure was divided into 15-min segments of exercise alternating with 15-min rest. Minute ventilation was monitored during the 3rd and 12th minute of the first exercise segment and during the 12th minute of the subsequent exercise segments, and the level of exercise was adjusted to maintain target ventilation if necessary. Halfway through each exposure spirometry was performed, and another blood sample was obtained.

After the last exercise segment, a 10-min recovery period was allowed, during which a symptom questionnaire was completed. Spirometry was then performed and a blood sample obtained. Weak responders were then given either 0.15 mg/kg of naloxone or normal saline intravenously. Strong responders were given either sufentanil or normal saline. Because the initial dose of 0.25 µg/kg of sufentanil resulted in nausea in some subjects (n = 4), we reduced the dose for all subsequent strong responders to 0.2 µg/kg. Drug assignment within the weak and strong subgroups was random, all drugs were diluted to a final volume of 20 ml, and the investigators performing spirometry and pain threshold testing were blinded to the drug.

Less than 5 min after drug administration, a spirometry was performed, blood sample was obtained, and a symptom questionnaire was completed. Subjects were observed for 1-2 h until they met standard postsedation discharge criteria. Due to the administration of sufentanil in this study, a taxi ride home was provided for all subjects as a safety precaution.

The SAS PROC MEANS statistical program was used to calculate the descriptive statistics for the variables of interest. For each subject and each of the four exposure conditions (air-saline; air-naloxone; ozone-saline; and ozone-naloxone), we calculated the differences between the measurements acquired at midexposure (when obtained), postexposure, and postdrug period and at the preexposure (baseline) period, respectively (i.e., first-order differences). Subsequently, we calculated the average differences by exposure condition and sex groups. For each subject and the same exposure atmosphere, we also calculated the differences between the respective saline and drug baseline, midexposure (when obtained), postexposure, and postdrug measurements (i.e., second-order differences). Subsequently, we calculated the group averages by sex and responder category. Such analysis was applied to each of the outcome variables of interest: the decrements in FVC and FEV₁ (expressed as %baseline values); plasma -endorphin levels; and the symptoms of chest pain, cough, and pain on deep inspiration. For the analyses of pain threshold and tolerance (Marstock test-derived variables), the plasma -endorphin levels, and FEV₁, Spearman rank correlation (r_s) calculations were performed by using the SAS PROC FREQ. Finally, we used a general linear-model method (SAS PROC GLM) to model the second-order differences for each outcome variable of interest, as stratified for sex, age (young vs. middle-aged), and exposure atmosphere (air vs. ozone). These analyses evaluated interactions of such factors with those for drug vs. saline.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Table 1 presents the average physical characteristics of the subjects and their baseline (spirometric) lung function, expressed in terms of percent of predicted values, stratified by age, sex, and ozone responsiveness. Although the weak responders in both the young and middle-aged group had a greater mean FVC and FEV₁ (%predicted) than the strong responders, these differences were not statistically significant. All the expected strong responders (as classified by the screening exposure) remained strong responders in the study. However, some of the weak responders from the screening phase showed FEV₁ decrements >5% in the study. Because they were still classified as weak responders, this accounts for the postozone FEV₁ of 5.4% and 6.3% in the saline and naloxone groups, respectively. The number of subjects for each category and the combination of exposure atmosphere-drug are given in Table 2.

The subjective symptoms reported after air exposures (and exercise) were minimal and not significantly different from preexposure symptoms. Administration of either saline or study drugs completely relieved these few symptoms. The subjective symptoms induced by ozone exposure are presented in Fig. 1. The symptoms were much less severe in weak (Fig. 1B) than in strong (Fig. 1A) responders. In the group of weak ozone responders, none of the symptoms were significantly different from air exposure and were minimally affected by administration of saline or naloxone. As expected, the strong responders generally experienced more intense symptoms from ozone exposure. Although administration of saline (ozone-saline day) provided some relief, the intensity of all symptoms was still statistically significantly elevated (P < 0.05) compared with respective air-exposure conditions (not graphed) that did not induce any symptoms. Administration of sufentanil (ozone-sufentanil day) had, however, significant effects on all symptoms. The drug completely abolished chest pain, and the discomfort from all other symptoms (shortness of breath, chest tightness, cough, pain on deep inspiration, and throat irritation), although not completely abolished, was statistically significantly reduced.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   A: average symptom scores for 6 subjective symptoms of strong responders exposed to ozone. By our definition, score of 1 reflects mild, i.e., noticeable but not annoying symptoms; 2 reflects moderate, i.e., noticeable and annoying symptoms. In each cluster of 4 bars, open bar represents score reported immediately after exposure and bar with horizontal lines represents score reported after subsequent administration of saline on ozone-saline exposure day; medium-gray bar represents score reported immediately after exposure and solid black bar the score reported after subsequent administration of sufentanil on the ozone-sufentanil exposure day. B: average symptom scores for weak responders to ozone. Bar fill type represents same conditions as for strong responders except for the solid bar, which for this group is the score reported after administration of naloxone on ozone-naloxone exposure day. Insp, inspiration. P values for comparisons indicated by brackets are as follows: * P < 0.05, ** P < 0.01, *** P < 0.001.

Figure 2 shows FEV₁ data from air and ozone exposure and postdrug recovery in strong responders to ozone. Exposure to air did not cause any significant changes in FEV₁. There was no sex difference in ozone responsiveness with our exposure protocol. Ozone exposure caused a significant decline in pulmonary function, similar for both males and females, at respective exposure/drug conditions (P < 0.001). The ozone-exposed strong responders who then received saline showed minor and nonsignificant improvement, consistent with spontaneous recovery after cessation of ozone exposure. However, the ozone-exposed subjects who received sufentanil showed dramatic improvement in FEV₁ (P = 0.001 compared with the saline group). Even with such a dramatic improvement, however, the ozone-sufentanil groups (men and women) did not quite return to baseline FEV₁, and the respective residual mean effects were still significant (P < 0.001). A similar pattern of response was observed for FVC. The mean FEV₁ differences between measurement periods with respective statistics are tabulated in Table 3.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Plot of mean forced expiratory volume in 1 s (FEV1) (%baseline) vs. duration of exposure and postexposure treatment for strong responders. Air-exposed cohorts (dashed lines: , saline; *, sufentanil) represent pooled data n = 8 (4 men and 4 women). Ozone-exposed cohorts are denoted by  and  for men (n = 14), and  and  for women (n = 15), respectively. Solid lines, cohorts who received sufentanil; dotted lines, cohorts who received saline. Arrow denotes time of drug administration (~2.1 h). Vertical bars associated with symbols are 1-sided SE. ppm, Parts/million; admin, administration. P values for various comparisons are reported in Table 3.

View this table: [in this window] [in a new window]   Table 3.   Mean and P values for 1st- and 2nd-order differences for FEV1 (%baseline) and 2nd-order differences for sex, age, and atmosphere among 3 test periods of ozone-exposure day

Figure 3 is a plot of FEV₁ decrements (%baseline) from air and ozone exposures and postdrug recovery in weak responders. Air exposure did not significantly change FEV₁. Ozone exposure had similar effects on both men and women. The FEV₁ decrements in both the ozone-saline (5.4%) and ozone-naloxone (6.3%) groups were significant (P < 0.001). Administration of saline or naloxone (0.15 mg/kg) changed the response little, and mean FEV₁ remained significantly below baseline (P < 0.001). However, naloxone administration did not cause a further decline in FEV₁. A similar pattern of response was seen for FVC. The mean differences for FEV₁ (%) and respective P values for various exposure conditions are reported in Table 3.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Plot of mean FEV1 (%baseline) vs. duration of exposure for weak responders. Air-exposed cohorts (dashed lines: , saline; *, naloxone) represent pooled data [n = 5 (2 men and 3 women)]. Ozone-exposed cohorts are denoted by  and  for men (n = 5), and  and  for women (n = 10), respectively. Solid lines, cohorts who received naloxone; dotted lines, cohorts who received saline. Arrow denotes time of drug administration (~2.1 h). Vertical bars associated with symbols are 1-sided SE. P values for various comparisons are reported in Table 3.

The overall baseline mean -endorphin level for the weak responders was higher (1.98 pmol/l) than that for the strong responders (1.40 pmol/l); this difference, although small, was statistically significant (P < 0.02). Except for the air-saline condition in the weak responders, regardless of the exposure atmosphere, the plasma -endorphin level invariably increased both at the end of exposure and after administration of a drug. Only the sufentanil group, however, showed a statistically significant increase (P < 0.05), as shown in Table 4. Determination of the strength of association between -endorphin level and ozone responsiveness (FEV₁), between various exposure groups and sessions when using the Spearman rank-correlation test, showed physiologically inconsistent results.

View this table: [in this window] [in a new window]   Table 4.   Plasma -endorphin levels for strong and weak responders by exposure group and test period

In the weak responders, the two key correlations of interest were the baseline (preexposure) plasma -endorphin level with 1) the postozone  baseline decrement (%baseline) in FEV₁ (r_s = 0.713; P < 0.01), and 2) with the postozone  postdrug decrement (%baseline) in FEV₁ (r_s = 0.572; P < 0.05). Respective correlations for the strong responders were 1) r_s = 0.317, not significant; and 2) r_s = 0.455, P < 0.05. Although the above changes are in the expected direction and internally consistent, the weaker association for the strong responders than that observed for the weak responders does not suggest that -endorphin levels modulate significantly the response to ozone.

Figure 4 illustrates warmth threshold, heat pain tolerance, and cool threshold plotted against the percentage decline that the individual subject had in FEV₁. The n = 16 subjects for Fig. 4, because the thermal testing apparatus was unavailable for part of the study and not all of the subjects opted to participate in this test. The correlations between ozone responsiveness and cutaneous sensitivity to heat or cold perception and pain are weak and not significant.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   A: a plot of warm threshold () and heat pain tolerance temperatures (), expressed in °C vs. %decrement in FEV1 (%baseline) induced by ozone exposure. Dotted line represents best fit line through the heat tolerance data points (n = 16 subjects; r = 0.228). Dashed line represents best fit line through the warm threshold data points (n = 16 subjects; r= 0.260). B: a plot of cool threshold temperatures () expressed in °C vs. %decrement in FEV1 induced by ozone exposure. Dashed line represents best fit line through the cool threshold data points (n = 16 subjects; r = 0.184).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

To facilitate the conduct and interpretation of the study, the subjects were preselected for their responsiveness to ozone. Only weak and strong responders were accepted for the naloxone and sufentanil arms of the study, respectively. Because the preselection of the subjects was based on spirometry (FEV₁) with a defined range of percent decrement from baseline for each class of responders, no relevant comparisons could be made with distributions of ozone response reported in the literature. In our study, subjects who were sham exposed (to air) had no significant postexposure symptoms or spirometric changes. Administration of either naloxone (n = 5) or sufentanil (n = 11) to these individuals caused no change in lung function. In the ozone-exposed group, within each class of responders, the mean postexposure spirometric changes in the cohorts who received saline or drug, respectively, were similar. This shows that despite an unequal size (by design) of these cohorts, the postozone spirometric data obtained in the saline-treated cohorts were, in terms of average magnitude of changes, a valid control for the respective drug-treated cohort.

Ozone inhalation activates bronchial C fibers. This has been shown in dogs by Coleridge et al. (8) using single vagal fiber recording and by Schlegele et al. (21) using vagal cooling to 7°C as a means of preventing conduction in myelinated nerves while allowing conduction in nonmyelinated nerves (C fibers). Ozone exposure has also been shown to deplete neuropeptides synthesized and released from C fibers in human airway epithelium (19) and to cause increased levels of C-fiber-associated tachykinin, substance P, in bronchial lavage in humans (14).

C-fiber function is downregulated by activation of opioid receptors (5). Therefore, we have explored the role of C fibers in the genesis of several ozone-exposure-associated phenomena in humans (decreased vital capacity and FEV₁, inspiratory substernal discomfort, and cough) by administering a rapid-acting lipophilic opioid agonist sufentanil to 31 adult volunteers who were preselected as strong responders to ozone. Most of these, as expected (20), were young individuals (n = 29). Ozone-induced chest and substernal soreness was completely relieved by the dose of sufentanil we administered. Although pain might be elicited by activation of A--fibers (myelinated) as well as C fibers (nonmyelinated), a differential stimulation of these afferents showed that, in humans, morphine (opioid agonist) acts on sensory C fibers and not on A--fibers (10). Thus our data demonstrate that the polysynaptic nociceptive pathway modulated by opioids is a dominant system responsible for postozone chest pain. Our observations do not allow us to determine the precise site(s) of action of sufentanil along the nociceptive pathway, although animal studies point to the peripheral nerve terminals as the primary site of action (4).

Although highly significant, reversal of spirometric decrements was not complete in any of the young individuals. Nor was chest pain on deep inspiration completely ablated, although chest pain during tidal breathing disappeared. These findings are most easily explained by assuming that the administered dose of 0.2 µg/kg sufentanil (0.25 µg/kg in a few early subjects) was inadequate to completely block opioid-sensitive nociceptive pathways. However, Bailey et al. (3) studied a variety of end points (including cutaneous pain threshold to electrical stimulation) after administration of 0.1, 0.2, or 0.4 µg/kg iv sufentanil, failed to find a significant dose-response relationship for their several end points over this dose range (although this failure might have been due to a lack of statistical power). Nevertheless, it seems appropriate to briefly consider alternative explanations. It is possible that some fraction of the decline in function and of the symptom of pain on deep inspiration is mediated by vagal afferents that are not opioid modulated. Also, ozone inhalation is well known to result in airways inflammation and bronchial hyperreactivity in young adults, which would remain unchanged immediately after sufentanil administration and could be responsible for persistent functional decrements (24).

The relatively modest suppression of ozone-induced cough by intravenous sufentanil was surprising. The pathophysiology of the cough reflex is complex, and on the afferent limb certainly involves intraepithelial polymodal receptors categorized as rapidly adapting inflation/deflation receptors (RAR), the impulses of which are conducted through small myelinated afferent fibers in the vagus (1).

The precise role of C fibers in the genesis of cough is debated (18). Widdicombe (25) has synthesized a large body of evidence with the suggestion that cross talk between the bronchial C-fiber and RAR systems occurs at several levels. In the airway epithelium itself, where tachykinins released from C fibers in close proximity to RAR cause activation of the latter via neurokinin NK₂ receptors (2), C-fiber activation promotes cough. However, in the central nervous system, C-fiber impulses, rather than provoking the cough reflex, actually cause gating of RAR impulses, thus inhibiting cough provoked by peripheral stimulation of RAR. Thus, if opioid receptors are confined to the C-fiber system, their activation by sufentanil could have opposing effects on cough provoked by RAR activation.

In animals, ozone inhalation activates both RAR and pulmonary stretch receptors (8, 16). The single vagal afferent fiber studies of Coleridge et al. (8) in ozone-exposed dogs were interpreted as showing specific activation of bronchial C fibers, not modulated by the respiratory cycle. Although 10 of 15 RAR fibers identified by them developed increased (inspiratory) firing rates during ozone exposure, this was preceded by a change in lung mechanics (decreased dynamic lung compliance). The authors considered this activation of RAR by ozone to be secondary, because it could be acutely abolished by a high-volume lung inflation that also temporarily restored dynamic lung compliance. Although these results in anesthetized dogs may not be applicable to awake humans, it is possible that ozone-exposed humans develop increased RAR activity (especially on attempted deep inspiration) related to subtle changes in lung mechanics (15) as well as to the previously discussed peripheral cross talk with C fibers. Thus suppression of C-fiber activation by opioid does not abolish ozone-induced cough.

Therefore, we suggest that ozone exposure in humans causes multiple effects on the airways. Stimulation and/or sensitization of opioid-modulated fibers [possibly by PGE₂ (9)] is largely responsible for the sensations of substernal pain and irritation and for most of the lung-function decrements (either reflexly or volitionally). This contention is supported by recent findings, in which this eicosanoid was found to stimulate rat pulmonary C fibers but not RAR or pulmonary stretch receptors (17). However, inflammatory airway changes might also contribute to lung function decrements. Stimulation of RAR leads to cough, although this stimulation may be indirect and attributable to subtle changes in lung mechanics.

The lack of any effect of a large dose of intravenous naloxone on weak ozone responders indicates that the small ozone response in these individuals cannot be attributed to high levels of endogenous opioids (endorphins). Endorphins are potent endogenous polypeptide opioid-receptor ligands, which have been shown to play a significant role in downregulation of pain. In our study, we did not find any meaningful changes in plasma -endorphin levels, although in seven of eight exposure conditions [including both sham (air) and ozone exposures], the postexposure levels were elevated. Because we did not find any consistent response that could be related to the exposure atmosphere, the postexposure increase was most likely due to exercise, which is known to induce a release of -endorphin (12). Further increase in plasma -endorphin levels after drug administration, even in those subjects who did not receive sufentanil, may have been due to stress associated with blood-sample drawing. Because this change was observed after both sham (air) and ozone exposure, it appears that the rise in plasma -endorphin level, particularly after sufentanil, is independent of exposure atmosphere. Although our weak responders had a slightly higher mean baseline plasma concentration of -endorphins than did strong responders, the level was apparently not high enough to have a significant impact on ozone response, since administration of naloxone did not provoke further impairment of pulmonary function or emergence of symptoms. Whether the marked interindividual differences in response of healthy adults to ozone inhalation can be attributed to differences in bronchial C-fiber anatomy and physiology or to differences in the chemical events after ozone uptake and reaction with substrates in the airway-lining liquid and epithelium is unknown at the present time.

We explored the possibility that the airway nociceptive system response to inhaled ozone might be paralleled by sensitivity to stimuli that cause somatic (cutaneous) pain. Although the findings were not significant, we noted a negative trend for the relationship between the initial preexposure heat pain tolerance and the magnitude of decline (as a %baseline) in postexposure FEV₁. The same trend was observed for heat pain threshold. Individuals with a lower cold threshold had smaller decrements in FEV₁ than did individuals who perceived cold at higher temperatures. The observed trends suggest that further exploration of nociceptive sensitivity as a predictor of ozone response may be of interest.

In summary, the results from this study demonstrate that ozone-induced decrements in spirometric lung function in healthy adults are principally neural in origin. The primary neural pathway involves opioid-modulated sensory C fibers. Administration of the potent µ-receptor agonist sufentanil immediately relieved ozone-induced symptoms and dramatically improved pulmonary function. Although the nociceptive (opioid-inhibitable) system mediates much of the immediate pulmonary response, other mechanisms, such as airway hyperresponsiveness and inflammation, are possible contributing factors. Endogenous opioids, however, do not seem to have a major modulating role in determining ozone responsiveness. Ozone exposure does not stimulate the release or production of -endorphins, individual sensitivity to ozone is not related to the circulating levels of -endorphin, and high-dose naloxone treatment failed to elicit (occult) ozone sensitivity in weak responders. Cutaneous pain sensitivity, as assessed by the Marstock technique, is not predictive of lung function response. However, exploration of airway mucosal nociceptive sensitivity as a marker of ozone responsiveness may be useful.