Ground-level ozone and particulate matter (PM) are the primary ambient air pollutants of smog. Exposure to ozone causes airway irritation, wheezing, coughing, pain upon inspiration, and breathing difficulties, and repeated exposure may impair lung growth and cause permanent lung damage [1]. Exposure to PM has been repeatedly associated with increased mortality, although how PM causes death is not well understood [2]. For both ozone and PM, health risks rise as exposure rises.

Student athletes often exercise vigorously in smoggy outdoor environments, and athletic practices frequently occur in the mid- to late-afternoon when diurnal ozone levels are highest. Vigorous exercise increases minute ventilation and inspiratory flow rate, intensifying exposure of the distal lung to airborne pollutants. Although several studies have documented short- and long-term effects of ozone related to exposure during exercise [3-15], few have examined children or young adults, and little is known about effects of repeated exposures to ambient ozone or PM among student athletes.

Breath condensate contains a number of constituents derived from the respiratory surfaces that hold promise as indices of inflammation. The pH of exhaled breath condensate (EBC) has been proposed as a biomarker of inflammation reflecting the acid-base balance of the airways, which is regulated primarily by the proton-buffering effects of ammonium derived from the oral cavity and airways [16]. Several studies suggest that pH of exhaled breath condensate is low in various inflammatory lung diseases including poorly controlled asthma, COPD, cystic fibrosis, and acute lung injury [17-27].

Studies reporting negative correlations between EBC pH and pro-inflammatory cytokines in the airways [17,18] further suggest that EBC pH has promise as a biomarker of the early airway inflammation preceding frank symptoms or lung function impairment. Because it is relatively easy to collect and measure, breath pH has potential as an effective way to assess inflammatory effects of air pollution in high risk populations. Two studies have explored EBC constituents as biomarkers of respiratory morbidity from ozone exposure, including one that examined breath pH, finding no significant change in breath pH after ozone exposure [5,19].

The aim of the present study was to test the hypothesis that vigorous outdoor activity in student athletes exposed to summer air pollution might induce lung inflammation and thereby reduce breath pH. To accomplish this, we collected EBC pre- and post-exercise on ten days during peak smog season in 16 high school athletes engaged in daily long-distance running in a down-wind suburb of Atlanta.

Mean age of participants (n = 16) was 14.9 years, 56% were male, and 69% were white (Table 1). Thirteen of 16 (81%) participants were distance runners and three were sprinters. Two of 16 (13%) reported having asthma, one of whom participated in only the first three days of the study. None were smokers. From 16 participants, we collected 144 pre-exercise and 146 post-exercise breath samples.

Mean (± SD) ambient 1-hour maximum ozone concentration was 71 (18) ppb. Median (interquartile range, IQR) was 61 (54–67) ppb. Mean (± SD) PM_2.5 measured at 5 pm was 27.2 (11.9) μg/m³. Median (IQR) was 23.2 (21.7 – 34.7) μg/m³. Four of 10 study days were air quality alert days, three of these triggered by high ozone levels and one by high PM levels. On three study days ozone concentrations exceeded health-based standards.

In this study, we sought to examine the association of exhaled breath pH with ambient ozone and particulate matter concentrations among 16 adolescent athletes exposed to these air pollutants during vigorous outdoor exercise. Although there was considerable variation in air quality over the 10 study days, we did not observe a relationship between air pollution levels and post-exercise breath pH. Overall, post-exercise breath pH was indistinguishable from pre-exercise breath pH, suggesting no acute effect of air pollution exposure during vigorous outdoor exercise on breath pH compared to resting values. This finding is similar to that of Corradi et al [5], who found no change in EBC pH after short-term exposure to ozone during light exercise and is the only study apart from ours that examined the association between air pollution exposure during exercise and EBC pH.

Although we did not observe an acute effect of outdoor exercise, we found that both resting and post-exercise breath pH values were intermittently lower than expected in a majority of subjects. Although airway pH in healthy individuals is believed to be slightly alkaline with pH of 7–8 [21,22], we observed in a majority of subjects intermittent breath pH values comparable to levels seen in severe inflammatory conditions such as acute asthma and sickle cell anemia. Day-to-day variability in breath pH was common and similar in magnitude to that seen in children with persistent asthma. We observed much greater within-subject variation in breath pH than previously reported [17,22-25]. The intermittently low breath pH values we observed were not correlated with subject, study day, or assay date, and resampling of banked samples suggested that they could not be attributed to random laboratory error.

Potential explanations for the low breath pH values that we observed include the possibility that salivary contamination (for instance, from gastroesophageal reflux) or other artifact of sample collection (such as recent eating or drinking) artificially lowered breath pH. From a mechanistic standpoint, it may be possible that routine vigorous exercise and its concurrent effect on systemic lactate production influenced airway pH. Because EBC samples were collected outdoors, it is plausible that inhaled atmospheric sulfates reduced EBC pH. Alternatively, it is possible that these otherwise healthy athletes experienced significant intermittent airway acidification events that were accurately reflected by the observed low breath pH values. We address each of these possibilities in turn.

Our use of a one-way valve system with saliva trap during EBC collection makes it unlikely that salivary contamination was responsible for the low pH values [21], and a previous study of the potential for salivary contamination using the same collection system showed no such contamination [22]. Although we cannot rule out that the athletes had eaten in the hour prior to collection of the resting breath samples, it is unlikely because samples were collected immediately after the students were dismissed from final classes of the day. Furthermore, the subjects did not eat during running, so we are confident that the low post-exercise breath pH values are not attributable to having eaten. Exposure to ingested liquids was restricted to water and non-carbonated sports drinks. Although it is plausible that sports drink consumption could contribute to acidic breath pH, three observations make us doubt that this is the case: (1) the principal investigator was present during each breath sampling period and observed only a small number of students consuming sports beverages, far fewer than the 75% with acidic breath pH values; (2) presumably more sports drinks would have been consumed after vigorous exercise, but acidic breath pH values were observed with similar frequency before and after exercise; (3) in a recent re-analysis of the breath pH samples, we identified citric acid – a common ingredient in sports drinks – in only six breath samples, suggesting that few athletes consumed sports drinks. Moreover, pH was normal in the breath samples containing citric acid.

Another possible explanation for the observed low pH values is the potential effect of gastroesophageal reflux on breath pH. It is plausible that this acid reflux is stimulated by vigorous exercise and may contribute to intermittent breath pH acidification events via salivary contamination. If this were the case, however, we would expect more frequent breath acidification events following vigorous exercise. On the contrary, we saw similar frequencies of acidification events before and after exercise. It is conceivable that reflux may have been stimulated by physical activity prior to practice, for instance if athletes were rushing from class to the practice field. We have no evidence to confirm or refute this possibility other than anecdotal recollection suggesting that this level of pre-practice activity was unlikely to have occurred often enough to account for the frequency with which acidic pre-practice breath pH values were observed. Also, as noted above, salivary contamination is unlikely due to the design of the Rtube^® breath collection system. Nonetheless, acid reflux as a contributor to the observed low breath pH values cannot be ruled out definitively.

Because (EBC) samples were collected outdoors, we explored the possibility that the low breath pH values we observed were artifacts of outdoor sample collection (rather than accurate reflections of airway pH). Outdoor sample collection could conceivably lower condensate pH if inhaled ambient sulfate particles did not deposit in the respiratory tract but were rather collected in the breath condensate. To examine this possibility, we constructed a simple mathematical model to estimate the maximum amount of sulfate expected in the condensate samples based on the concentration of sulfate and other relevant atmospheric ions in the ambient air (see Additional file 1). We also estimated the predicted breath pH and compared it to measured breath pH. These comparisons were conducted on a subsample of five runners on two study days, one chosen to represent a day with high ambient sulfate levels and the other a day with low ambient sulfate levels. The results of this modeling (see Additional file 1, Table 3) suggest that inhaled ambient sulfate particles may potentially lower breath pH by a small amount, but this effect is minimal and unlikely to explain entirely the surprisingly low breath pH values we observed in the student athletes.

Why we observed low and highly variable breath pH values in this group of otherwise healthy athletes is unclear. One possibility is that the routine vigorous exercise reported by these athletes caused systemic lactic acid production that altered resting airway pH. The mechanism by which the respiratory and systemic acid-base buffering systems may interact to cause this effect is poorly understood and requires more study. Another potential explanation is that long-term exposure to outdoor air pollution – such as may be experienced by children who routinely exercise outside during polluted summers – may trigger intermittent endogenous airway acidification events indicative of pollution-related lung inflammation. This idea is broadly consistent with results of several studies that suggest that long-term exposure to urban air pollution may cause lung damage or lung growth impairment in children [4,26-28] and could partially explain why we did not see similar values among a comparison group of healthy but relatively sedentary adults. Another possibility is that low pH values representing endogenous intermittent airway acidification events are normal rather than pathologic phenomena. Supporting this speculation are several recent studies that measured breath pH among healthy controls (Table 3) and found surprisingly low breath pH in a subset of subjects [25,29], including one large population-based study that found breath pH values as low as 4.4 in healthy children [30]. The possibility that intermittent low breath pH values reflect normal biologic processes calls into question the utility of breath pH measurement – especially one-time breath pH measurement – as a reliable biomarker of lung inflammation or respiratory health effects.

This analysis has several limitations. The small number of subjects (n = 16) may have limited statistical power to detect relationships between ambient air pollutants and post-exercise breath pH. We had no data from personal monitoring of air pollutant exposure or on long-term exercise patterns or residential histories of the subjects and, thus, were unable to examine if ambient air pollution or low breath pH values were associated with cumulative exposure to polluted environments. That we did not see similar low breath pH values in a group of healthy adults with breath samples taken outdoors suggests that the low breath pH values observed among the student athletes may be attributable to characteristics of their lungs that we were unable to measure in this study. Because characteristics of our study subjects may have varied from those of student athletes in general, we caution against generalizing our results.

In summary, we found no acute effect on breath pH from vigorous outdoor exercise on days with significant ozone and PM pollution in this group of 16 student athletes. We did, however, observe highly variable and surprisingly low breath pH values in a majority of these student athletes. These intermittent, low breath pH values are unexplained and may represent endogenous airway acidification. Additional studies with greater statistical power are necessary to rigorously examine the potential effect of outdoor exercise in polluted environments on the lungs of young athletes. However, before additional studies propose to use the pH of exhaled breath condensate as a biomarker of airway inflammation or respiratory morbidity, it will be necessary to determine if the low breath pH values we and others have observed among healthy individuals are indicative of respiratory pathology or normal biologic phenomena.

Akaike's Information Criterion (AIC), body mass index (BMI), Centers for Disease Control and Prevention (CDC), exhaled breath condensate (EBC), environmental tobacco smoke (ETS), forced expiratory volume in 1 second (FEV₁), interquartile range (IQR), National Health and Nutrition Examination Survey (NHANES), not reported (NR), particulate matter (PM), particulate matter ≤2.5 microns in diameter (PM_2.5), parts per billion (ppb), standard deviation (SD), standard error (SE), standard error of the mean (SEM)

The author(s) declare that they have no competing interests.

JF conceived of, designed, and coordinated the study, collected and analyzed data, and prepared the manuscript. CG performed statistical analysis and provided statistical expertise. RG participated in study design, data collection, data analysis, and manuscript preparation. DV participated in the design of the study, data collection, and data analysis. EH carried out laboratory analyses. WGT participated in study design, coordination, data collection, data analysis, and manuscript preparation. All authors read and approved the final manuscript.