texas commission on environmental quality€¦ · web viewozone is a secondary pollutant, and it is...

Dose response curves derived from clinical ozone exposures can inform public policy

Authors: Sabine S. Lange, Ge Tao, Lorenz R. Rhomberg, Julie E. Goodman, Michael L. Dourson, Michael E. Honeycutt

Abstract

Context: Ozone is one of six criteria air pollutants for which regulations are set by the US Environmental Protection Agency using the National Ambient Air Quality Standards (NAAQS). The ozone NAAQS level is currently set at 75 parts per billion (ppb).

Objectives: We used data from human clinical studies to inform policy decisions about a protective ozone level.

Methods: We plotted mean forced expiratory volume (FEV1) response versus total inhaled ozone dose (calculated from ozone concentration, duration of exposure and ventilation rate) from clinical studies of 1-8 hour durations. This produced two distinct sigmoidally-shaped curves, for the shorter-duration exposures (≤ 3 hours) and the longer-duration exposures (6-8 hours). The initial plot used data from healthy young adults; additional analyses incorporated data from children and asthmatics, but results did not differ.

Results: There were clear thresholds of effect, which were consistent with the known ozone mode of action. We estimated typical ozone doses at ambient ozone concentrations of 75, 70 or 65 ppb (8-hour maximum average). Doses were similar at these three concentrations, and almost all doses were below those associated with a 5% FEV1 decrement, even when different exposure times and ventilation rates were assessed.

Conclusions: This type of analysis can determine thresholds of ozone toxicity, which can be crucial for choosing an ambient ozone concentration that is protective of human health. Setting the ozone NAAQS at 65-70 ppb will have a marginal impact, if any, on ozone doses and lung function, but could have significant societal and economic implications.

Ozone (O3) is a chemical that is produced both in the stratosphere, where it forms the protective “ozone layer,” and in the trophosphere (ground level), where it is an air pollutant (commonly known as smog). Ozone is a secondary pollutant, and it is formed when oxides of nitrogen (NOx) or volatile organic compounds (VOCs) react with ultraviolet radiation produced by the sun. The chemistry of ozone formation and removal is complex, and therefore the control strategies for ground-level ozone need to be sophisticated and carefully considered. Ozone is a highly reactive chemical, and because of this, it reacts with indoor structures such as ventilation systems, household items and even human skin and hair, which all scavenge it from the air (Weschler, 2000, Weschler, 2015). This means that indoor ozone concentrations are low and the majority of ozone exposure occurs when people are outdoors (Lee et al., 2004, Sarnat et al., 2005, Sarnat et al., 2001).

Ozone is one of the six criteria air pollutants that are regulated by the United States Environmental Protection Agency (EPA) using the National Ambient Air Quality Standards (NAAQS). The NAAQS standards have four elements: the indicator (the chemical that is being measured – in this case, for photochemical oxidants, ozone is the indicator), the averaging time (in the case of ozone, there have historically been two averaging times – a 1-hour maximum and the current maximum daily 8-hour average), a form (the current form of the ozone standard is the 4 th highest level, averaged over three years), and a level (the concentration of the pollutant). From 1979-1997, the level of the ozone NAAQS was set at 0.12 parts per million (ppm) for a 1-hour daily maximum average. In 1997, it was set at 0.08 ppm (rounding conventions made this 0.084 ppm) for a maximum daily 8-hour average. In 2008, the level of the standard was decreased to 0.075 ppm with the same averaging time, and in 2014 the EPA proposed to decrease the level of the standard to somewhere within the range of 0.065-0.070 ppm (USEPA, 2014b). As the level of a NAAQS decreases, both the uncertainty in the data (USEPA, 2014b) and the cost of attaining the standard (USEPA, 2014c) increase.

Setting the ozone standard involves an evaluation of several evidence streams, including animal toxicology, epidemiology and human clinical studies. For choosing the level of the ozone standard, EPA places the most confidence on the human clinical studies (USEPA, 2014b). However, there are several aspects of the human clinical studies that make them difficult to translate into the elements of the NAAQS (Goodman et al., 2015). One of these aspects is how the health endpoint being measured is related to the ozone mode of action (MOA, discussed more below). Another aspect is the dose, which comprises three elements – duration of exposure, concentration of exposure, and ventilation rate. In order to relate the findings from these studies to the general population, all of these aspects need to be considered, and put in the context of what would be expected for real-world exposures.

The MOA of ozone centers on its action in the respiratory tract. When ozone is inhaled into the respiratory tract, it encounters the extracellular lining fluid (ELF), which contains high amounts of antioxidants, particularly uric acid, ascorbic acid and glutathione (Mudway et al., 1999, Mudway et al., 1996, Mudway and Kelly, 1998). Because ozone is an oxidant, it can be titrated by reaction with the antioxidants. Ozone is so reactive, and there are so many potential reaction partners in the ELF, that it may never encounter the cellular layer (Pryor, 1992). However, the secondary reaction products of ozone may do so. Several groups have measured the loss and replenishment of antioxidants in nasal and bronchoalveolar lavage samples. In the bronchoalveolar area, glutathione and ascorbic acid play larger roles in scavenging ozone (Mudway et al., 1996). In nasal samples, uric acid is the primary antioxidant, and it decreases by ~ 30% after exposure to 0.2 ppm ozone (with alternating 15 min periods of exercise and rest) for 1 hour. However, there was no further decrease in uric acid after the second hour of ozone exposure, suggesting that it was being replenished during that time (Mudway et al., 1999). This implies that if exposure to ozone is gradual, then a higher total dose may be tolerated because of replenishment of the antioxidants in the ELF.

It is known that when the respiratory tract is exposed to ozone, the ozone reacts with antioxidants and is inactivated. Larger amounts of ozone can overwhelm the antioxidant defenses, allowing secondary reaction products to form and incite a variety of health effects. This includes inflammation, neuronal responses (which lead to spirometric responses), and increased epithelial permeability and bronchial smooth muscle reactivity (Mudway and Kelly, 2004, Hazucha et al., 1989, Passannante et al., 1998, Kehrl et al., 1999, Kehrl et al., 1987). Of these respiratory effects of ozone, spirometric responses, particularly forced expiratory volume in 1 second (FEV1), and inflammation (measured by neutrophil infiltration) are the best characterized. Dose-response relationships have been described for both endpoints (McDonnell et al., 2012, Schelegle et al., 2012, Mudway and Kelly, 2004), and the ozone-FEV1 dose-response models by McDonnell et al (2012) were used in the most recent ozone NAAQS risk assessment (USEPA, 2014a).

Another consideration when setting a level for the ozone standard is sensitive subpopulations. To ensure that these populations are protected by the standard, they need to be evaluated in dose-response relationships. Experimental ozone exposure studies have been conducted with children, asthmatics, and elderly adults (McDonnell et al., 1985, Horstman et al., 1995, Jenkins et al., 1999, Kehrl et al., 1999, Linn et al., 1994, Peden et al., 1997, Holz et al., 1999, Weymer et al., 1994).

Although several groups have derived dose-response relationships, outside of modeling done by the EPA, these dose-response curves have not been utilized as tools to inform the choice of a protective ambient ozone concentration. They have also not considered potentially sensitive subpopulations. To translate the ozone dose that is used in these clinical experiments to a level that can be used in setting a standard that protects the general population requires consideration of the three elements of dose: concentration, duration and ventilation rate. Ventilation rate in particular is an important factor that is not an explicit part of the NAAQS elements (concentration and duration are represented by level and averaging time, respectively), and so it must be considered before drawing conclusions about where the ozone level should be set. In order to do this, known ventilation rates of populations must be investigated.

In this manuscript, we describe the development of a dose-response curve between ozone and FEV1, using data from healthy adults, children and asthmatics. Threshold doses of the FEV 1 response were derived and compared to ozone doses that would be expected given known population exposure scenarios. These threshold doses were also applied, using known ventilation rates and exposure times, to estimate the concentrations at which ozone could cause certain FEV1 responses. This creates a tool that can be used by decision makers, such as the EPA Administrator, to directly apply the results from human clinical studies to the general population, including sensitive subpopulations.

Study selection and data extraction.

We assessed ozone dose-response using both the group mean FEV1 response, as well as individual responses using data from published studies. For the group mean response, we used data from 13 studies for 6-8-hour (long) exposures, and data from 4 studies for ≤ 3 hour (short) exposures (summarized in Table 1). To find relevant studies for this analysis, we started with those used in the dose-response analysis of Schelegle et al. (2012), which included all of the studies conducted at the EPA exposure chambers at the University of North Carolina Chapel Hill campus, and at the Human Performance Laboratory at the University of California Davis campus. We excluded three studies cited in that paper (Kehrl et al., 1987, Koren et al., 1989, McDonnell, 1989) because of insufficient information available in the papers. Two long-exposure studies were added to our analysis – (Kim et al., 2011) because this was a seminal study used in the EPA’s ozone assessment (USEPA, 2013), and (Hazucha et al., 1992), because one of the papers included in the Schelegle group of studies (Adams, 2006b) was intended to repeat the Hazucha study. The short-exposure study by (Folinsbee et al., 1978) was included to increase the number of short-exposure data points, and because it used varying exercise levels during exposure.

We obtained individual FEV1 response data (864 data points from 541 individuals) from the EPA dataset provided by WF McDonnell, which has been used in previous dose-response modeling (McDonnell et al., 2007). Some of this data comes from studies that are included in our group mean analysis (Folinsbee et al., 1988, Horstman et al., 1995, Horstman et al., 1990, McDonnell et al., 1991, McDonnell et al., 1983, Seal et al., 1993). These data were compiled from a series of 15 controlled human exposure studies that were conducted using similar methods at the US EPA Human Research Facility in Chapel Hill, North Carolina, over the period from 1980-1993. The study subjects were healthy, non-smoking, Caucasian, male volunteers, aged 18-35. The characteristics of volunteers and ozone exposures are summarized in Table 2. The physical characteristics of the volunteers are similar in long-exposure studies and short-exposure studies. Some individual subjects participated in both types of studies. The ozone exposure ranged from 0.08-0.24 ppm for long-exposure studies and from 0.12-0.40 ppm for short-exposure studies. The ozone concentration was held constant throughout each experiment. Minute ventilation was measured during each exercise period. For periods of rest, minute ventilation was assumed to be 5 L/(min × m2).

We also examined the effects of ozone on various subpopulations. To obtain studies that assessed the response of asthmatics to ozone, we conducted a literature search in PubMed in January 2015 using the search terms: “ozone," "forced expiratory volume," and "asthma." We included those studies with: i) ozone-only single exposures; ii) measured ventilation rate or ventilation rate/body surface area (BSA) and the BSA of the study subjects; iii) quantified FEV1

responses to ozone; iv) asthmatic subjects; v) exposure times of ≤ 3 hours or 6-8 hours (to fit with the healthy young adult dose-response curves); vi) original research ( i.e., not meta-analyses or reviews); and vii) text in English. Our initial search yielded 87 studies; 12 studies met our inclusion criteria and contained data that could be plotted on the dose-response curves (Table 1). Data from healthy adult controls in these studies were included in the healthy adult dose-response curve. To obtain studies that assessed the response of children to ozone, we conducted a literature search in PubMed in January 2015 using the search terms: "ozone," "forced expiratory volume," and "children." We used the same inclusion criteria as with the asthmatics (except that the subjects had to be children instead of asthmatic). Our initial search yielded 58 studies; only 1 met our inclusion criteria (McDonnell et al.,

1985). There weren’t enough data points to derive a dose-response curve based on these subpopulations, but they were plotted with the young healthy adult dose-response curve to determine whether or not they fell within the normal range of those curves.

We analyzed the mean and standard deviation (SD) of the percent FEV1 decrement. The FEV1

data were presented in different ways in different studies, so different calculations were used to collect the appropriate data:

If percent decrements were reported, then they were used; some studies did not report the variance and we were not able to calculate it.

If only volume in liters was reported, then mean percent decrement of FEV1 was obtained by dividing the final volume by the initial volume (then subtracting 1 and multiplying by 100 to get the percent). This did not allow calculation of an SD.

If volume only was reported, but individual data were provided, we calculated the percent decrement for each person, then calculated the mean and SD.

If the change in volume and the initial volume were reported, we calculated the mean percent decrement in FEV1 by dividing the change in volume by the initial volume. If an SD for the change in volume was not reported, we calculated the percent change SD by dividing the volume change by the final volume.

For ventilation rates, often only the exercise ventilation rate was provided. When this was the case, we assumed a resting ventilation rate based on what was measured in (Adams, 2006a)) (11.6 L/min for 6.6 hour exposures with 10 minute rest periods) or in (Adams, 2006b)) (11.4 L/min for 8 hour exposures with 30 minute rest periods). For the shorter (≤ 3 hours) exposures, a default resting ventilation rate of 11 L/min was used. When resting ventilation rate was required to calculate total dose, the resting ventilation rate we used is noted in Table 1.

For the intermediate time-point analysis, we used the time-course curves reported in nine of the long exposure studies (marked in Table 1). These studies provided graphs or individual data of FEV1

decrements at different time points within the experiment. We estimated the FEV1 decrements from the graphs, and calculated the dose at each time point. These points were plotted based on the exposure type (square wave versus triangular wave or “stepped” exposures), or according to exposure time (≤ 3 hours, > 5 hours).

Dose-response curves and calculation of threshold doses.

We fitted sigmoid dose-response models to individual data of long exposure, individual data of short exposure, group mean data of long exposure and group mean data of short exposure. All data points were plotted, including the FEV1 responses to filtered air exposure. The fitted models were in the following form,

)}/dose Totalln({exp1FEV%

501

,

where 1FEV% is the percent change in FEV1 defined as pre 1pre 1post 1 FEV/)FEVFEV(%100 , is the top plateau of FEV1 decrements at minimal dose, is the bottom plateau of FEV1 decrements at high

dose, is the slope parameter that defines the steepness of the curve, and 50 is the dose at which the response is halved. The dose-response modeling were performed using SAS version 9.3 (SAS Institute Inc., Cary, NC, USA) software package. After obtaining a unique fitted dose response function for each dataset, we calculated threshold doses corresponding to FEV1 decrements of 0%, 5%, 10% and 15% by inversing the fitted dose response functions. The inversed dose response functions were in the form of

ˆ/ˆFEV%

FEV%ˆlnexpˆdose Total

1

150

,

where ˆ ,ˆ ,ˆ and 50̂ are estimated values of , , and 50 from model fitting based on each

dataset, and 1FEV% was assigned values 0%, 5%, 10% and 15%.

The 95% confidence intervals of the threshold doses were obtained by the profile likelihood method as recommended by (Crump and Howe, 1983) using the NLMIXED procedure in SAS (Wheeler, 2005).

Using the threshold doses, with the ventilation rates from Table 4, we made a matrix of time and activity, and calculated the concentration of ozone at the different exposure times and exercise levels where the FEV1 threshold was crossed.

Dose-response curves with subtraction of filtered air responses.

Another analysis was conducted using the individual-response dataset with subtraction of an individual’s filtered air response from their ozone response. Not all of the individuals exposed to ozone were also exposed to filtered air. 340 observations from short exposure experiments and 10 observations from long exposure experiments do not have matched filtered air responses, so the differences could not be calculated, and the dose response curve for short exposures could not be fitted due to lack of data at the 1000-2000 ppm x L doses. Therefore, for observations without a filtered air response, we used the median of available filtered air responses of long exposure experiments or short exposure experiments, according to the observation's exposure duration, as its

filtered air response. Sigmoidal curves were then fit to this data, and threshold doses were calculated as above.

Calculation of individual FEV1 decrements of ≥ 10%.

Using the individual-response dataset, all FEV1 decrements of ≥ 10% were categorized according to the dose at which they occurred, and were organized into dose intervals of 250 ppm x L. The percent of all FEV1 responses in that dose interval that were decrements of ≥ 10% was plotted against the dose interval. This was done separately for individuals exposed to ozone for ≤ 3 hours or for 6-8 hours, and 3rd order polynomial trend lines were fit to the data points (R2 values for the curves were > 0.97).

Analysis of consistency of FEV1 responses.

We also analyzed whether an individual could consistently be categorized as sensitive. This analysis used the EPA data set, and individuals who were exposed more than once, either to filtered air or to ozone at 80, 100 or 120 ppb. Very little data were available for repeated exposures ( i.e., an individual exposed on two separate occasions to the same dose of ozone or filtered air), so we worked mostly with cases in which a single individual was exposed to different ozone concentrations on different occasions. The data were all from longer, 6-8 hour exposures to ozone or filtered air, because the individuals exposed for shorter periods of time rarely were exposed to any dose more than once. We did pair-wise comparisons using individuals’ FEV1 decrements caused by their exposures to filtered air and to ozone. We evaluated six pair-wise comparisons, and Pearson’s correlation coefficients and linear regression were applied to the relationship between each dataset.

Ambient ozone concentrations and population exercise duration and ventilation rates.

Ambient ozone concentrations were obtained from the Texas Commission on Environmental Quality (TCEQ) Texas Air Monitoring Information System (TAMIS). Hourly data were obtained from days in 2014 that had 8 hour maximum averages of 75 ppb, 70 ppb or 65 ppb ozone (10 days were used at each concentration). This hourly data was used to calculate the average 1 hour maximum, the maximum averages of 2-14 hour time periods, and the daily average ozone concentration on days with those 8 hour maximum averages. These maximum and average concentrations are displayed in Table 5.

http://tamis.tceq.texas.gov/tamis/

Measured exercise duration and ventilation rates were obtained from several sources. US EPA (1994) provided occupational 8-hour and non-occupational 24-hour ventilation rates used for risk assessments. (Zuurbier et al., 2009) and (Samet et al., 1993) provided ventilation rates measured by correlation with heart rate, and Samet also included diary data of ventilation rate and duration for activities in which the individuals in the study engaged. (USEPA, 2009) used diary data from the National Human Activity Pattern Survey, and calculated ventilation rates for physical exertion of varying degrees (at rest, sedentary, and light, moderate or high intensity exercise). This was then applied using probability distributions to the individuals from the National Health and Nutrition Examination Survey. This produced theoretical exercise ventilation rates and durations of activity for 24-hour periods for the people examined in the survey. From this analysis, exercise ventilation rates and durations were estimated for different age groups, genders and activity levels. There was no consideration of fatigue in the ventilation rates and durations [that is, that a person can only sustain certain exercise ventilation rates for certain amounts of time; (Isaacs et al., 2008)]. This means that the exercise ventilation rates and durations could be over-estimated. We used the mean (with the 10 th

and 90th percentile) ventilation rates and durations for the different activities and age groups (we used data averaged from males and females, and data from children 6-11 years old and adults 21-31 years old).

Results

Dose-response curves with total inhaled ozone dose and FEV1.

In order to understand the dose-response relationship between ozone and FEV1, we plotted total inhaled dose of ozone [which is calculated by multiplying ozone concentration in parts per million (ppm), duration of exposure in minutes, and ventilation rate in liters/minute (L/min)] versus FEV1

decrement. Ventilation rate is dictated by exercise, so these experiments were generally conducted while the participants were exercising to achieve significant doses of ozone. Because many of the references that reported population ventilation rates did not normalize the rates to body surface area (BSA), we did not divide the ventilation rate by BSA in our main analysis. We used both the EPA dataset from 541 individuals (McDonnell et al., 2007), as well as the mean FEV1 decrement reported in 16 clinical experiments (see Table 1). We divided the data into two exposure categories, ≤ 3 hours and 6-8 hours, to determine whether the time it takes to achieve a specific dose affects the dose-response curve. For the main analysis, the zero ozone dose (called filtered air in the experiments) was also plotted on the graphs, but lung function at the zero dose was not subtracted from the ozone response data, as some investigators have done. Rather, all changes were determined in relation to the pre-exposed FEV1 value, which is the experimental control. We did an additional comparative analysis of the individual data with the filtered air response subtracted from the ozone response (see below).

For the curves based on individual data points, there were two distinct sigmoidal-shaped curves for the short and long exposure times (regression analysis, p < 0.0001, Figure 1A, B). The difference between these curves might be attributable to the upregulation or replenishment of antioxidant molecules in the respiratory tract with longer exposure, which could mitigate the physiological response to ozone (Mudway et al., 1999). Similarly, when the group mean response short- and long-exposure data were plotted, two different curves were observed (p < 0.001, Figure 1C). The long-exposure curve was sigmoidal, and when the individual and group mean long exposure responses were plotted together the curves were almost identical. (Figure 1D). The group mean short-exposure curve demonstrated an apparent linear relationship. Because the short-exposure individual data curve was sigmoidal, and the ozone MOA supports a sigmoidal-type response, it is likely that there are not enough group-mean short-exposure data points to demonstrate the true sigmoidal relationship. To confirm that the group mean response to the ≤ 3 hour exposures was indeed sigmoidal, we combined the early time points (≤ 3 hours) from nine of the long exposure studies, with the short exposure group mean data (as shown in Figure 1C), and we found that these data together demonstrated a sigmoidal relationship (Figure 2A). Because these early time points from the long exposure studies had to be estimated from graphs in papers, they were not included in the main analysis.

To confirm the finding that FEV1 responses to ozone exposure are greater if the dose is attained over a shorter time period, we evaluated the intermediate time points from nine of the studies. We estimated the FEV1 responses from the graphs in the papers, and calculated the doses for exposure times of ≤ 3 hours or > 5 hours. We found that the responses to the shorter exposure times were greater than the responses to the longer exposure times at equal total doses (Figure 2B).

The ozone-FEV1 response shows considerable inter-individual variation, which is demonstrated by the wide array of individual responses (Figure 1A). This results in some uncertainty in the mean curve, and this uncertainty increases with increasing dose. The bearing of this variation on assessment of individual differences in sensitivity is discussed further below.

Time-course and exposure type for dose-response curves.

Many studies have considered different exposure scenarios, with the intention of mimicking observed diurnal variations in ozone concentration. The most common exposure scenario is the “square wave” schedule, in which the subjects are exposed to a single concentration of ozone for the selected period of time (see Table 1). This has been compared with a stepped or ramped “triangular” schedule, in which the subjects are exposed to increasing and then decreasing levels of ozone, which is consistent with ozone diurnal variations, and which in total equals the dose of the square wave [e.g., (Adams, 2003a, Adams, 2006a)]. The triangular exposure has been shown to cause faster and greater FEV1 decrements in the middle of the exposure, but by the end of the exposure the square and triangular FEV1 decrements are typically very similar [e.g., Figure 1 from (Adams, 2006a)]. Using group mean data from nine of the long-exposure studies, we plotted the dose and FEV 1 decrement from graph-estimated intermediate time points. The FEV1 decrements from both the triangular and square exposure types are interspersed and together demonstrate a clear dose-response relationship (Figure 2C). This shows that the FEV1 response is largely dependent on the total ozone dose that is inhaled, regardless of the exposure profile.

Threshold doses derived from dose-response curves.

Using the best-fitting dose-response curves from Figure 1A-C, we calculated the doses at which five different mean FEV1 decrements (0%, 5%, 10%, 15% and 20%), would be expected to occur. The results are summarized in Table 3. We evaluated these specific decrements in part because the EPA considers the range of 10-20% FEV1 decrements to be moderate (USEPA, 1989), and has stated that a decrement of 10% might be adverse in a person with a pre-existing lung disease (USEPA, 2014a). We calculated these for the group mean and individual analyses, and for short-exposure and long-exposure curves, separately. Note that doses were not derived for a 15% or 20% FEV1 decrement from the long-exposure curve, because the curve plateaued at a smaller decrement. Similarly, no 20% FEV1 decrement dose was derived for the short-exposure curve, because that curve was never at or below 20%. Figure 1B also shows the presence of threshold doses at which no FEV1 response would be expected to occur (the 0% FEV1 threshold), which occurs at doses somewhere below 500-750 ppm x L. Such a threshold is consistent with the known ozone mode of action, in which antioxidants scavenge ozone in the epithelial lining fluid and prevent it from reacting and causing damage in the respiratory tract (USEPA, 2013). Other groups that have modeled ozone-FEV1 dose-response curves have also shown evidence of thresholds or doses of onset (McDonnell 2012, Schelegle 2012).

We also estimated the percentage of individuals exposed to a given dose interval (the intervals were 250 ppm x L wide) who experienced an FEV1 decrement ≥ 10% (Figure 3A). For short exposures (≤ 3 hours), doses of 750-1000 ppm x L caused a more than five-fold increase in the number of people experiencing a ≥ 10% FEV1 decrement, compared to doses of 500-750 ppm x L. For longer 6-8 hour exposures, a dose in the range of 1250-1500 ppm x L caused ~33% of people to experience an FEV1 decrement of ≥ 10%, whereas the range of 750-1000 ppm x L led to ~17% of people responding with an FEV1 decrement of ≥ 10%. The 5% FEV1 group mean threshold derived above (Table 3, individual data thresholds: 740 ppm x L for short exposure; 954 ppm x L for long exposure) correspond to the dose ranges that provide more protection against individuals experiencing ≥ 10% FEV1 decrements (500-750 ppm x L for short exposures; 750-1000 ppm x L for long exposures). This shows that using 5% FEV1 decrement thresholds would prevent not only group mean decrements of ≥ 10%, but would also prevent a significant number of more responsive individuals from experiencing a ≥ 10% FEV1 decrement.

Ozone-FEV1 dose-response using subtraction of filtered air responses.

Other dose-response analyses have considered the effects of ozone on FEV1 response using the filtered air response subtracted from the ozone response. The purpose of this calculation is to control for an individuals’ response to the exercise in the exposure protocol. We used the individual data and created an ozone-FEV1 dose-response using matched filtered air responses subtracted from ozone responses. Many of the short exposures, and a few of the long exposures, did not have matching filtered air responses, and so to derive the dose-response curves using that data we calculated the median filtered air response for the long and short exposures and subtracted this from the ozone exposures. It is not clear how the authors of other dose-response models dealt with this [eg. (McDonnell et al., 2012)]. Using this method, we produced dose-response curves that were sigmoidal in shape (Figure 3B), and the FEV1 thresholds were slightly lower than when the filtered air was not subtracted, although they were within the 95% confidence intervals (Table 3).The curves were quite similar, particularly the short-exposure curves, and were within the 95% confidence intervals of one another (Figure 3B). These curves still include a similar, distinct threshold below which no FEV1 response was observed. Because the filtered air subtraction dataset had to be

interpolated due to lack of data, we completed the remaining analysis with the dose-response and thresholds derived from the dataset that did not subtract the filtered air responses.

Potential sensitive subpopulations.

Our initial spirometric ozone dose-response analyses were conducted using data from healthy young adults (18-35 years old). However, various potentially sensitive subpopulations have also been experimentally exposed to ozone, including children, the elderly and asthmatics (Table 1).

Few experimental ozone exposure studies have been conducted with children (Avol et al., 1987, McDonnell et al., 1985), and children were exposed to ozone alone in only one study (McDonnell et al., 1985). This study exposed 22 boys between the ages of 8 and 11 to filtered air or 120 ppb ozone for 2.5 hours while they engaged in intermittent, vigorous exercise. We plotted individual data for these boys together with the adult individual short-exposure response curve. In this case, to compare adults with children, it is appropriate to scale the ventilation rate to body surface area, so we used a dose-response curve with the ventilation rate divided by BSA (Figure 4A). This plot shows that the 95% confidence interval (CI) for the children’s FEV1 responses largely overlaps the 95% CI for the curve derived from adult data. In addition, we compared the expected FEV 1 response in the adults at the same dose (304.6 ppm x L/m2 BSA) that caused the observed FEV1 response in children. The adult expected FEV1 response of -2.3% was similar to the observed children’s -3.8% FEV1 response (p = 0.51). The group mean response for the children was also consistent with the dose-response plotted for the adult group mean (Figure 4B, discussed more below). The lack of difference between the child and adult spirometric responses is consistent with the conclusions of (McDonnell et al., 1985).

Adults in varying age groups have also been exposed to ozone.(Hazucha et al., 2003) studied the effects of ozone exposure on men and women in two age groups, younger than 35 and older than 35. They found that with increasing age there was a decreasing FEV1 response to the same dose of ozone. We did not plot this on our curve because of insufficient dose information published in the paper. This result is consistent with the conclusion from dose-response modeling that spirometric responses decrease with age in adults (McDonnell et al., 2007).

Asthmatics are a sensitive population with regard to ozone exposure. Many studies have investigated the response of mild asthmatics to ozone exposure. More severe asthmatics have not been studied, likely because of exercise-induced bronchoconstriction (Parsons et al., 2013), which would render them incapable of reaching the ventilation rate for the required dose. Mild asthmatics have been exposed for both short and long times and spirometric responses and other biomarkers have been measured. We plotted the adult asthmatic group mean FEV1 response with the short- and long-exposure curves for healthy adults and found that for both exposures, the asthmatic responses were similar to those of healthy young adults (Figure 4 A, B). Overall, this suggests that adult asthmatics do not demonstrate increased spirometric responses to ozone, which is consistent with the conclusions reported from many studies (Linn et al., 1994, Balmes et al., 1997, Koenig et al., 1987, Koenig et al., 1985, Stenfors et al., 2002, Nightingale et al., 1999, Basha et al., 1994). There are also data investigating the effects of short-term ozone exposure on healthy and asthmatics adolescents (aged 11 to 18 years old). These studies were conducted with quite low doses, making it difficult to derive a dose-response relationship. Generally however, the responses between the healthy and asthmatic adolescents were similar (Figure 4B). These data all show that healthy children and asthmatics have similar FEV1 responses to ozone as healthy young adults.

Sensitive individuals within the healthy population.

Another way to investigate populations sensitive to ozone is to assess the more responsive people amid the population of healthy adult individuals (McDonnell et al., 2012, USEPA, 2014a). In an effort to examine this, other scientists have investigated ozone dose-response results in a different way than is presented here. They have plotted specifically (and only) those people who have FEV 1

decrements (after subtraction of filtered air response) that are greater than 10%, 15% or 20%. These people have been considered as sensitive amongst the healthy population. However, there are several problems with this method. The first is that it excludes the majority of the data: for example, in the individual data set used in this study, 75% of the short ozone exposure FEV1 decrements and 83% of the long ozone exposure FEV1 decrements are < 10%, so restricting the analysis would discard 75-83% of the dataset. Another problem is that these large FEV1 decrements can be attained just with filtered air exposure (this important observation is lost when filtered air responses are subtracted from ozone responses). As can be seen in the individual data curve (Figure 1A), FEV1

decrements > 10% have been observed just in response to filtered air, which demonstrates the natural heterogeneity of response that occurs during this type of experiment.

An additional concern about labeling individuals with a ≥ 10% FEV1 decrement as sensitive is that it assumes that the individuals are consistently sensitive to ozone. To determine whether this assumption holds true, we used the individual data for subjects that had been exposed to different ozone levels in separate tests and measured the correlation across individuals between their FEV 1

responses after different long exposures. We found that there is no correlation between an individuals’ FEV1 response to ozone and their FEV1 response to filtered air (Figure 5A-C). This is unsurprising, as one would expect that different mechanisms would lead to decrements from filtered air (presumably caused by exercise) versus decrements from ozone. When comparing an individual's magnitude of response to different ozone exposures in a single clinical experiment, we found that there was a statistically significant correlation between the magnitudes of an individual's response from one ozone exposure to another (Figure 5 D-F). However, because the correlation coefficients ranged from 0.22 to 0.55 (depending on the doses that were compared) the association was not very strong. Based on these data, we conclude that focusing only on those individuals with a particular FEV1 decrement is not reflective of the entire population.

Ozone doses at which non-spirometric responses occur.

Many clinical endpoints have been measured in ozone human exposure studies, including those relevant to inflammation, epithelial barrier function, airway hyper-responsiveness and immune system modification. These endpoints typically represent biomarkers for the activation of physiological responses, and are not necessarily themselves adverse physiological responses ((ATS), 2000, Goodman et al., 2010). Dose-response modeling is often not possible for these datasets, owing to the lack of consistent measurement methods between different studies and the difficulty in quantifying some of these markers. The exception is neutrophil influx in bronchoalveolar lavage samples, which can be quantified in a dose-response curve. This was done by (Mudway and Kelly, 2004,) and we refer to their work here. The only caveat to (Mudway and Kelly, 2004) is that we found that their doses are off by a factor of 1,000 because of a mistake in converting m3 to liters. Because of this, we divided all of their final numbers by 1,000.

The response of many inflammatory biomarkers to ozone has been examined, including increases in inflammatory cytokines such as IL-8 and TNF, as well as influx of inflammatory cells,

particularly neutrophils. These have been measured in lavage samples from the nasal cavity and the upper and lower respiratory tracts, and in bronchial biopsies. Mudway and Kelly’s (2004) dose-response analysis of neutrophils in bronchoalveolar lavage samples (the only sample type that had the consistency of data required for a quantitative dose-response analysis) found that in the 6 hours after ozone exposure, there was a threshold ozone dose of 645 μg/m2 (658 ppm x L, assuming an average BSA of 2) to cause significant neutrophil influx, and at 18 - 24 hours after ozone exposure, there was a threshold ozone dose of 810 μg/m2 (827 ppm x L) to cause significant neutrophil influx. These threshold doses are similar to the thresholds presented here for FEV1 response (Table 3).

Real-world time and exercise exposure data.

The threshold doses in Table 3 take into account exposure duration, ventilation rate, and ozone concentration. In contrast, the ozone NAAQS is only an 8-hour maximum average concentration and do not consider ventilation rate. Differences in exposure duration also require consideration of replenishment of detoxifying antioxidants, as evidenced by the different dose-response curves for ≤ 3 hour and 6-8 hour exposures. Incorporating the production of antioxidants into the model at even longer exposure durations would raise the values of the FEV1 decrement thresholds. We considered reasonable, real-world exposure durations and ventilation rates, and combined them with measured ozone concentrations to determine whether the resultant doses would be expected to cause significant FEV1 decrements.

Guidance documents as well as published information exist about the duration and ventilation rates of people exercising in the general population. Because a person has to be exercising at moderate to vigorous intensity to achieve a significant dose of ozone at current US ambient concentrations [(Adams, 2006a, Schelegle et al., 2009), etc.], we used information about exercising populations in this analysis. While we did not include exposure location in our analysis, we note that ozone is primarily an outdoor pollutant (Sarnat et al., 2005, Sarnat et al., 2001), so these exposure scenarios assume that the person is exercising outdoors. We combined the information about exercise duration and ventilation rate with actual ambient ozone concentrations to calculate the ozone doses that people are expected to receive while exercising outdoors. These doses were compared to the dose thresholds from Table 3 to determine whether these exposure scenarios are likely to cause the designated FEV1 decrements.

To determine exposure times and ventilation rates, we used several different EPA guidance documents (USEPA, 1994, USEPA, 2009), as well as experimental observations made by (Samet et al., 1993) and (Zuurbier et al., 2009). These values are shown in Table 4. Ozone concentrations are diurnally variable, and we used the TCEQ TAMIS database to determine what ozone concentrations at different averaging times would be expected on days that just meet the current NAAQS standard level (75 ppb maximum eight-hour average), or the potential alternative NAAQS standard levels (70 and 65 ppb maximum eight-hour averages). Monitoring data from 10 days where the maximum eight-hour average was 75, 70 or 65 ppb were used to evaluate other maximum averaging times for 1-14 hours or the 24-hour average (Table 5).

Protective ambient ozone concentrations for different thresholds and endpoints.

http://tamis.tceq.texas.gov/tamis/

The exposure ventilation rate for a certain activity, the time spent at that activity, and the appropriate ozone concentration for that duration and standard level were combined to produce an expected dose (in ppm x L). The following is an example of how that was calculated:

From Table 4: Sedentary child ventilation rate = 4.8 L/min; duration = 13.7 hrs

From Table 5: 14 hr max average (matching the 13.7 hr duration) = 64.7 ppb (at a 75 ppb max 8 hour average level)

Total Inhaled Dose at 75 ppb = 4.8 L/min x 13.7 hrs x 60 min/hr x 0.0647 ppm = 255 ppm x L

We plotted the ozone doses associated with each activity as well as the 5%, 10% and 15% FEV 1

decrement thresholds from the individual dose response data analysis (Table 3). We found that all of these activities are associated with doses below any of the thresholds, regardless of which NAAQS level was used (Figure 6). Changing the standard level made very little difference in expected dose. Most of the doses were so low that we wouldn’t expect the more sensitive responders to experience decrements of ≥ 10% (compare to doses in Figure 3). Only the 24-hour non-occupational exposure at the 75 and 70 ppb standard levels was over even the 5% group mean FEV 1 decrement threshold (but was still well below the 10% FEV1 threshold that the EPA considers is needed to protect sensitive individuals). This non-occupational scenario assumes a 24-hour outdoor exposure, eight hours of which are spent doing manual labor (USEPA, 1994). We also demonstrated that longer exposure times cause smaller decrements at the same total dose (Figure 1, Figure 2), and therefore, using the 6-8-hour experimental exposure threshold over-estimates the response to a 24-hour exposure.

In a separate analysis using these data, we combined duration of exposure from 1-24 hours, and the individual data short- and long-exposure dose-response thresholds with various relevant ventilation rates to calculate at what dose of ozone (for that duration and ventilation rate) a given FEV1 mean response would be predicted to occur. For times less than or equal to 4 hours, we used the short-exposure dose-response thresholds, and for times greater than 4 hours, we used the long-exposure dose-response thresholds. We produced several matrices to reflect different FEV 1

decrements (Tables 6-8). From this we see that an adult worker (or a child exercising at moderate intensity) would have to be exposed to 90 or 147 ppb ozone for 8 hours to experience FEV 1

decrements of 5% or 10%, respectively. This tool can be used by policy makers to choose ozone concentrations that are protective of the population.

Discussion

In this study, we use ozone dose-response relationships to inform the choice of a health-protective ozone level. To do this, we considered several different aspects of ozone exposure and response, including: total group mean and individual-level responses to ozone; thresholds of ozone response; different types of ozone exposure; responses of subpopulations to ozone; and different health endpoints. We applied our final analysis to exposure times and ventilation rates that are relevant to the general population to produce a table of protective ozone concentrations.

To be confident in the results produced by any model, it is important that it be corroborated by other evidence streams. In this case, we used the known ozone MOA to judge our dose-response model. The presence of threshold doses at which no FEV1 response would be expected to occur (the 0% FEV1 threshold) is consistent with the known ozone MOA, where antioxidants scavenge ozone in the ELF and prevent it from reacting and causing damage in the respiratory tract. Using the short exposure times, the dose for a 0% FEV1 response was either very low (19 ppm x L) or could not be calculated because the dose-response curve was always below zero. However, the shape of the individual-response curve shows that the curve does not deviate significantly from zero until a dose of almost 500 ppm x L. This means that the short-exposure responses are consistent with a threshold of response to ozone exposure. Other groups who have investigated dose-response curves for ozone exposure have also shown evidence of thresholds or doses of onset (McDonnell et al., 2012, Schelegle et al., 2012).

The results presented here only model the physiological response of one health endpoint: FEV 1. Although this endpoint has been very well characterized and there is guidance for FEV1 adverse effects, we also discussed briefly non-spirometric endpoints. While it is important to quantify non-spirometric respiratory responses to ozone, there are several caveats when interpreting these data. One is that the different responses do not tend to correlate with one another or with the spirometric responses, indicating that these responses are mediated by separate pathways (Aris et al., 1993, Schelegle et al., 1991, Kehrl et al., 1987, Que et al., 2011). This means that a single individual that has a heightened reaction within one aspect of the ozone response will not necessarily have a heightened reaction within another ozone response. Therefore, there are likely few entirely sensitive individuals to ozone exposure, and so additional cautions do not need to be put in place for these theoretically hyper-sensitive people. Another caveat is that while these are known responses to ozone, it is difficult to judge whether they are adverse responses (Goodman et al., 2014). Therefore, these respiratory responses to ozone are useful for elucidating MOA, but are not appropriate for setting a protective ozone level.

As mentioned above, there is published information regarding the adverse effect levels of FEV1. It is important to consider the adverse effect level, because the public does not need to be protected from homeostatic responses, just those that are adverse. For FEV1, the American Thoracic Society (ATS) suggests that a significant decrease in FEV1, combined with significant symptoms, should be considered as adverse ((ATS), 2000). In the studies used for our analysis, the lowest dose at which a long-term exposure met this criteria was 912 ppm x L [from (Adams, 2002)], and for a short-term exposure the lowest dose was 608 ppm x L [from (McDonnell et al., 1983)]. However, because an individuals’ FEV1 response and their symptoms correlate poorly (Schelegle et al., 2009, McDonnell et al., 1999, Frampton et al., 1997), this threshold for adversity is difficult to apply to ozone response.

Alternatively, the ATS has suggested that two-point changes in FEV1 of >12% may be clinically significant (Pellegrino et al., 2005). In its most recent ozone standard review, the EPA uses two FEV1

thresholds of adverse effects: a 10% decrement for populations with respiratory disease, and a 15% decrement for healthy populations (USEPA, 2014a).

Although children and asthmatics have been characterized as at-risk populations for ozone exposure (USEPA, 2014b), papers have been published that suggest that these groups do not have an increased spirometric response compared to healthy young adults (McDonnell et al., 1985, Linn et al., 1994, Balmes et al., 1997, Koenig et al., 1985, Koenig et al., 1987, Stenfors et al., 2002, Nightingale et al., 1999, Basha et al., 1994). To investigate this discrepancy, and to determine whether or not the dose-response relationship that we derived could be applied to these populations, we plotted the dose-response of children and asthmatics on our healthy young adult curves. Both children and asthmatics had responses that were consistent with healthy young adults (Figure 3), and can be modeled using the healthy young adult dose-response.

Several other papers have been published that describe the relationship between ozone dose and different health endpoints, particularly the FEV1 response (e.g. (McDonnell et al., 2007, Schelegle et al., 2009, McDonnell et al., 2012). The McDonnell 2012 model was used to derive the MSS model on which the EPA relies to model FEV1 decrements in its most recent ozone health risk and exposure assessment (USEPA, 2014a). Our work differs from these papers and from the MSS model in several important ways. One way is that we do not subtract the individuals’ filtered air response, but rather include it as a zero dose in the dose-response curve. Including a zero-dose is a common practice in modeling dose-response and allows the inclusion of the inter-individual variability that occurs just in response to the study protocol. We also demonstrated that there is no correlation between an individual’s filtered air response and their ozone response (Figure 5), and so it may be inappropriate to subtract filtered air because the responses represent different biological mechanisms. In addition, the ozone exposure study design itself may preclude the subtraction of the filtered air response, because sometimes the exposures to filtered air vs. ozone are weeks or months apart (Schelegle et al., 2009), when other biological or environmental factors could be affecting the differences seen in a person’s FEV1 response; and also, many experiments did not expose participants to both ozone and filtered air, so there is no filtered air response to subtract from the ozone response for that individual. All together, the most valid way to use the filtered air response is to graph it as a zero ozone dose in the ozone-FEV1 dose-response.

Another difference between this analysis and other ozone-FEV1 dose-response analyses is that we consider all responses to ozone, not just those responses that are much greater than the mean response. By including the entire dataset, we are provided with a model that is more likely to represent a general population response. When other models use only the “sensitive” responders in a population, it ignores the fact that they may not consistently be responders, as our analysis shows (Figure 5). In addition, because these analyses are being used to inform the ozone NAAQS, it is important to consider the language of the Clean Air Act when considering how to model the data. As is stated in the ozone proposed rule (USEPA, 2014b):

The legislative history of section 109 indicates that the primary standard is to be set at “the maximum permissible ambient air level…which will protect the health of any [sensitive] group of the population,” and that, for this purpose, “reference should be made to a representative sample of persons comprising the sensitive group rather than to a single person in such a group.” S. Rep. No. 91-1196, 91st Cong,. 2d Sess. 10 (1970)

Therefore, we also consider the data from different sensitive subpopulations that have been experimentally exposed to ozone, to determine whether or not the healthy young adult dose-response also adequately represents these populations. Finally, unlike the other papers that have published dose-response models, we use real-world ventilation rates and exposure times to allow this information to be applied to the general population, which can aid policy makers in making decisions.

There are uncertainties in this analysis. One uncertainty is in the sample size – there are several hundred (mostly young and healthy) individuals used in this analysis, compared to a general American population of greater than 300 million. The large number of studies that have investigated the response to ozone mitigates this uncertainty somewhat, but it still must be considered when generalizing the results. In addition, we use the group mean response, not a lower CI, as the threshold against which the population should be protected; however, we did calculate a lower CI thresholds and it could be used. This may be offset by the conservative assumptions made in calculating expected population dose, including that the person was exercising outdoors (and therefore was exposed to ambient levels of ozone), that they were exposed to the maximum ozone average that day, and that the exposure day only just met the level of the current or alternative standards (the form of the standard dictates that there only be 4 days per year that meet or exceed that number – the other 361 days must be lower). Another uncertainty is that while there are some subpopulations considered in this analysis, there are others that were not included because of a lack of data. This includes more severe asthmatics, as well as those with other respiratory diseases. However, the ventilation rate component of the ozone dose eases this concern to some degree. In order to get a high enough dose of ozone to cause an effect, a person has to be inhaling at exercise-level ventilation rates. A person with a significant respiratory disease is not likely able to achieve such ventilation rates.

There are two ultimate products of this analysis: i) the doses of ozone the population would be expected to achieve, given known ozone concentrations and exercise ventilation rates and durations; and ii) the ozone concentrations that would cause specific FEV1 decrements, assuming different exposure times and ventilation rates. It is very important to make realistic assumptions about all three of the components of ozone inhaled dose, including the ambient ozone concentration. This is because even if a standard is set at 75 ppb, the form of the standard dictates that this cannot be exceeded more than 3 times a year. So if an area is in attainment of the current standard, an 8-hour concentration of 75 ppb or above would only happen in that area maximally 4 days per year, leaving the other 361 days with levels lower than that.

Conclusions

In conclusion, we report a unique analysis that uses controlled human exposures, incorporates ozone exposures with group mean and individual FEV1 responses, and considers other subpopulations to produce dose-response curves that exhibit thresholds in the ozone dose response. These thresholds are consistent with the known ozone MOA. Threshold doses derived from these dose-response curves can be compared to expected doses of people in the population at ozone concentrations that correspond to the current or alternative ozone NAAQS levels. This analysis indicates that changing the 8 hour maximum ozone concentration between 75, 70 and 65 ppb makes little difference in the doses that the population would be expected to experience, and all of these doses are below thresholds for adverse effects of ozone exposure as defined by the EPA and others.

We also used the different threshold doses to make a matrix of time, ventilation rate and ozone concentrations at which a given FEV1 decrement would be expected to occur. This provides a tool that uses the controlled ozone human clinical data in a format that allows policy makers to decide on a level for the ozone NAAQS that, when combined with the ozone NAAQS form and averaging time, will protect public health with an adequate margin of safety. The policy decisions that need to be made to use this tool are: Who are we protecting? And what FEV1 decrement are we protecting them from?

Acknowledgments

The authors would like to thank Dr. Stephanie Shirley of TCEQ for her discussions and suggestions for this work.

Declaration of InterestThe authors SSL and MEH are employed by the Texas Commission on Environmental Quality (TCEQ), a State environmental regulatory agency, and report no declarations of interest. GT, LRR, and JEG are employed by Gradient, an environmental consulting firm. Their contributions to this manuscript were completed during the normal course of business, with funding to Gradient provided by TCEQ. This work represents the individual professional views of the authors and not necessarily the views of TCEQ. The authors…

Study nExposure Characteristics

Type Pattern Duration (hrs)

Ozone Conc. (ppb)

Ventilation (L/min)

Healthy Adult Subjects(Adams, 2000)* 15M

15FFacemask Square wave 6.6 FA, 120 Total – 25,

29, 33(Adams, 2002)* 15M

15FChamber, Facemask

Square wave 6.6 FA, 40, 80, 120

Total – 29

(Adams, 2003a)*

15M15F

Chamber, Facemask

Square waveTriangle-step

6.6 FA, 80 Total - 30

(Adams, 2003b) 15M15F

Chamber Square wave 6.6 80 Total - 30

(Adams, 2006a)*

15M15F

Chamber Square waveTriangle-step

6.6 FA, 40, 60, 80

Ex – 36Rest – 11.6

(Adams, 2006b)*

15M15F

Chamber Square wave 8 FA, 120 Ex – 36Rest – 11.4

(Adams and Ollison, 1997)*

12M Facemask Square waveTriangle - ramp

6 FA, 120, 80 Total – 33, 19

(Folinsbee et al., 1988)

10M Chamber Square wave 6.6 FA, 120 Ex – 40, 42Rest – 11.6

(Hazucha et al., 1992)*

23M Chamber Square waveTriangle-ramp

8 FA, 120 Ex – 39Rest – 11.4

(Horstman et al., 1990)*

22M Chamber Square wave 6.6 FA, 80, 100, 120

Ex – 39Rest – 11.6

(Kim et al., 2011)

27M32 F

Chamber Square wave 6.6 FA, 60 Ex – 37Rest – 11.6

(McDonnell et al., 1991)

38M10M

Chamber Square wave 6.6 FA, 80FA, 80, 100

Ex – 40Rest – 11.6

(Schelegle et al., 2009)*

16M15F

Chamber Triangle - step 6.6 FA, 63, 72, 81, 88

Total – 33

(Adams, 2003b) 15M15F

ChamberFacemask

Square wave 2 300 Total – 38

(Folinsbee et al., 1978)

10M10M10M10M

Chamber Square wave 2 FA100300500

Ex – 10, 30, 50, 70

Rest – 10


20M22M20M21M21M29M

Chamber Square wave 2.5 FA120180240300400

Ex – 64-68Rest – 11

(Seal et al., 1993)

17M15M15M16M15M15M

Chamber Square wave 2.33 FA120180240300400

Ex – 43Rest – 11

Healthy Child Subjects


C - 22M Chamber Square wave 2.5 FA, 120 Ex – 39Rest – 10

Asthmatic Subjects(Horstman et al.,

1995)H - 9MA – 7M,

10F

Chamber Square wave 7.6 FA, 160 Ex – 26-30Rest – 11.4

(Jenkins et al., 1999)

A – 9 M, 2 F


(Kehrl et al., 1999)

A – 4M, 5F

Chamber Square wave 7.6 FA, 160 Ex – 25Rest – 11

(Linn et al., 1994)

H – 8M, 7F

A – 13M, 17F

Chamber Square wave 6.5 FA, 120 Ex – 28-31Rest – 11.6

(Peden et al., 1997)

A – 8M Chamber Square wave 7.6 FA, 160 Ex – 25Rest – 11.4

(Holz et al., 1999)

H – 10M, 15F

A – 5M, 10F

Facemask Square wave 3 FA, 125, 250

Ex – 26Rest – 6.8

(Jenkins et al., 1999)

A – 9M, 2F


(Weymer et al., 1994)

A – 12M, 9F

6M, F

Chamber Square wave 1 FA, 100, 250

400

Ex – 27Rest – 10

Adolescent Subjects(Koenig et al.,

1985)H – 4M,

6FA – 4M,

6F

Mouthpiece Square wave 1 FA, 120 Rest - 8

(Koenig et al., 1987)

H – 3M, 7F

A – 4M, 6F

H – 3M, 7F

A – 7M, 3F

Mouthpiece Square wave 0.67 FA, 120

FA, 179

Ex – 32 – 41Rest – 8 - 9


H – 5M, 7F

A – 9M, 3F

Mouthpiece Square wave 1 FA, 120 Ex – 33 – 35Rest – 9


A – 8M, 5F

Mouthpiece Square wave 1 122 Ex – 29Rest - 8

Notes: C – children; H – healthy individuals; A – asthmatic individuals; FA – filtered air; asterisk (*) – studies used for intermediate time point analyses; Total – total ventilation was provided; Ex – exercise ventilation rate; Rest – resting ventilation rate; Square wave – ozone concentration remains constant; Triangle-step – ozone concentration increases and decreases in a stepwise fashion and the average is equal to the given concentration; Triangle-ramp - ozone concentration increases and decreases in a continuous fashion and the average is equal to the given concentration.

Table 2: Characteristi

cs of Volunteers

and Exposures from EPA dataset


Long Exposure Studies

(n = 115)

Short Exposure Studies

(n = 453)

All Studies

(n = 541)

Age (SD)

25.5 (4.2)

23.9 (3.7)

24.5 (4.0)

Height (SD)

180.9 (7.3)

180.4 (7.0)

180.6 (7.1)

BSA (SD)

1.98 (0.16)

1.94 (0.14)

1.96 (0.15)

BMI (SD)

23.7 (3.1)

23.1 (2.6)

23.3 (2.8)

Weight (SD)

77.8 (11.8)

75.5 (10.1)

76.2 (10.8)

Ozone concentration range (ppb)

80-240

120-400

80-400

Table 3: Total doses to produce a mean FEV1 responseMean % ΔFEV1 Long Exposure Dose (ppm x L) Short Exposure Dose (ppm x L)

Group Mean Dose Response (95% CI)0 659 (484, 796) 19 (NC, 331)

- 5 972 (908, 1040) 619 (480, 749)- 10 1481 (1287, NC) 1000 (848, 1157)- 15 NC 1448 (1256, 1720)- 20 NC NC

Individual Dose Response (95% CI)0 608.5 (NC, 794) NC

- 5 954 (846, 1058) 740 (663, 812)- 10 1554 (1240, NC) 927 (868, 998)- 15 NC 1467 (1157, NC)- 20 NC NC

Individual Dose Response, Filtered Air Subtraction (95% CI)0 NC NC

- 5 898 (702, 1033) 738 (639, 822)- 10 1271 (1080, NC) 913 (848, 989)- 15 NC 1377 (1084, NC)- 20 NC NC

Notes: NC – Could not be calculated

Source Population Exercise Intensity Ventilation Rate (L/min)* Duration (hrs)*

(USEPA, 2009)

Children (6 - < 11 years old)

Sedentary 4.8 (3.7-6) 13.7 (13-15)

Light 11.3 (9.2-14) 7.4 (5.5-9.6)

Moderate 21.6 (17-26.8) 2.6 (0.9-4.1)

High 41.5 (31.4-53.5) 0.3 (0.02-0.9)

Adult (21 - < 31 years old)

Sedentary 5.3 (3.6-5.9) 12.5 (11.2-13.8)

Light 11.8 (9.2-14.9) 6.3 (3.8-9.7)

Moderate 26.1 (18.8-34.4) 5 (1.8-7.6)

High 49.8 (34.6-67.2) 0.3 (0.05-0.6)

(USEPA, 1994)

Non-occupational24 hr Ventilation with 8 hrs Manual labor

14 24

Occupational Manual labor 22 8

(Zuurbier et al., 2009)

Adult Bicycle commute 23.5 (11-47.7) 2

(Samet et al., 1993)

Child Outdoor play 16 (12.1-17.4) 1.9

Child Bicycling 27.1 (16.7-34.8) 2.1

Adult Vigorous bicycling 65 (40.8-87.8) 0.8

Notes: (*) - Mean ventilation rates and times, and where available, the 10 th and 90th percentiles in parentheses

Table 5: Ozone concentrations on days with maximum 8 hr concentrations of 75, 70 or 65 ppb.

Concentration Metric

75 ppb Days (ppb)mean (SD)



1-hr max 85.8 (3.5) 77.4 (5.7) 72.4 (4.7)

2-hr max average 84.2 (3.2) 76.2 (5.2) 71.3 (4.1)

3-hr max average 82.8 (2.7) 75.3 (4.8) 70.3 (3.7)

4-hr max average 80.8 (2.2) 74.3 (4.2) 69.6 (3.4)

5-hr max average 79.4 (1.7) 73.3 (3.2) 68.6 (2.6)

6-hr max average 78.2 (1.3) 72.4 (2.1) 67.6 (1.9)

7-hr max average 76.8 (0.9) 71.4 (1.4) 66.6 (1.1)

8-hr max average 75.4 (0.6) 70.2 (0.7) 65.6 (0.8)

9-hr max average 72.6 (4.5) 69 (0.7) 64.5 (1.2)

10-hr max average 71.5 (3.2) 97.6 (1.6) 63.2 (2.0)

11-hr max average 70.4 (2.2) 66.2 (2.4) 61.9 (2.9)

12-hr max average 69.2 (1.8) 64.9 (3.2) 60.6 (3.5)

13-hr max average 68 (1.8) 63.8 (4.0) 59.2 (4.2)

14-hr max average 66.9 (2.1) 62.7 (4.7) 57.9 (4.8)

24-hr average 52.2 (5.5) 51.4 (8.5) 46.2 (6.5)Notes: provided are the mean maximum averages using different time metrics from 10 days with eight-hour maximum averages of 75, 70 or 65 ppb (standard deviation in parentheses). Shaded is the measured eight-hour maximum average, the form of the current standard.

FEV1 Decrement = 5% Ozone Concentration (ppb)Source Population & Exercise VE (L/min) 1 hr 2 hrs 3 hrs 4 hrs 5 hrs 6 hrs 7 hrs 8 hrs 12 hrs

EPA Sedentary Child 5 2569 1285 856 642 663 552 473 414 276EPA Sedentary Adult 5 2327 1164 776 582 600 500 429 375 250EPA Light Int Child 11 1121 561 374 280 289 241 206 181 120EPA Light Int Adult 12 1028 514 343 257 265 221 189 166 110

Samet Child Outdoor Play 16 771 385 257 193 199 166 142 124 83EPA Mod Int Child 22 561 280 187 140 145 120 103 90 60

TCEQ Adult Worker (8 hr) 22 561 280 187 140 145 120 103 90 60Zuurbier Adult Bicycle Commute 24 525 262 175 131 135 113 97 85 56

EPA Mod Int Adult 26 474 237 158 119 122 102 87 76 51Samet Child Bicycling 27 457 228 152 114 118 98 84 74 49

EPA High Int Child 42 294 147 98 73 76 63 54 47 32EPA High Int Adult 50 247 123 82 62 64 53 45 40 27

Samet Adult Male Bicycling 65 190 95 63 47 49 41 35 31 20

TAMIS Max hourly [O3] with 75 ppb Stda 86 84 83 81 79 78 77 75 69



Table 6. Concentration of ozone at which a population would be expected to experience an FEV1

decrement of 5%, given different exposure durations and ventilation rates (VE – ie. exercise levels)

Notes: For times ≤ 4 hours, the short exposure dose-response curve was used and for times > 4 hours, the long exposure dose-response curve was used; threshold doses were derived using individual data with no subtraction of filtered air responses; ozone concentrations that are greyed-out are those that occur at times greater than the maximum duration associated with that ventilation rate, as shown by (USEPA, 2009); bolded, outlined ozone concentrations are those associated with matched ventilation rates and durations (see Table 4); Highlighted are the 8 hour ozone concentrations, which can be directly compared to the 8 hour maximum average ozone standard of 75, 70 or 65 ppb.a The ozone concentrations at each ventilation rate and duration can be compared to the ozone concentrations in the bottom three rows of the table (also see Table 5). These three rows are the maximum average ozone concentrations for the exposure duration specified in the column headings, on days where the 8 hour maximum ozone concentration would be 75, 70 or 65 ppb.

FEV1 Decrement = 10% Ozone Concentration (ppb)Source Population & Exercise VE (L/min) 1 hr 2 hrs 3 hrs 4 hrs 5 hrs 6 hrs 7 hrs 8 hrs 12 hrs

EPA Sedentary Child 5 3219 1609 1073 805 1078 899 770 674 449EPA Sedentary Adult 5 2915 1458 972 729 977 814 698 610 407EPA Light Int Child 11 1405 702 468 351 471 392 336 294 196EPA Light Int Adult 12 1288 644 429 322 431 359 308 270 180

Samet Child Outdoor Play 16 966 483 322 241 324 270 231 202 135EPA Mod Int Child 22 702 351 234 176 235 196 168 147 98

TCEQ Adult Worker (8 hr) 22 702 351 234 176 235 196 168 147 98Zuurbier Adult Bicycle Commute 24 657 329 219 164 220 184 157 138 92

EPA Mod Int Adult 26 594 297 198 149 199 166 142 124 83Samet Child Bicycling 27 572 286 191 143 192 160 137 120 80

EPA High Int Child 42 368 184 123 92 123 103 88 77 51EPA High Int Adult 50 309 155 103 77 104 86 74 65 43

Samet Adult Male Bicycling 65 238 119 79 59 80 66 57 50 33




Table 7. Concentration of ozone at which a population would be expected to experience an FEV1

decrement of 10%, given different exposure times and ventilation rates (VE – ie. exercise levels)

Notes: For times ≤ 4 hours, the short exposure dose-response curve was used and for times > 4 hours, the long exposure dose-response curve was used; threshold doses were derived using individual data with no subtraction of filtered air responses; ozone concentrations that are greyed-out are those that occur at times greater than the maximum duration associated with that ventilation rate, as shown by (USEPA, 2009); bolded, outlined ozone concentrations are those associated with matched ventilation rates and durations (see Table 4); Highlighted are the 8 hour ozone concentrations, which can be directly compared to the 8 hour maximum average ozone standard of 75, 70 or 65 ppb.a The ozone concentrations at each ventilation rate and duration can be compared to the ozone concentrations in the bottom three rows of the table (also see Table 5). These three rows are the maximum average ozone concentrations for the exposure duration specified in the column headings, on days where the 8 hour maximum ozone concentration would be 75, 70 or 65 ppb.

Figure Legends

Figure 1. Ozone-FEV1 dose-response curves. (A) Total inhaled ozone dose (in ppm x L) versus mean percent change in forced expiratory volume in 1 second (FEV1) of healthy young adults from the US EPA individual dataset exposed to ozone for ≤ 3 hours (short exposure, red diamonds and red trend line) or 6-8 hours (long exposure, blue diamonds and blue trend line) while exercising, with the equations associated with each curve below the graph; (B) The trend lines from A with 95% confidence intervals (dashed lines); (C) Plot as in A, using group mean data from 16 clinical experiments with the equations associated with each curve below the graph; (D) Plots of the individual data curves from (B) and the group mean curves from (C). (*) p < 0.01.

Figure 2. Intermediate time points of ozone-FEV1 dose-response data. (A) Total inhaled ozone dose (in ppm x L) versus mean percent change in forced expiratory volume in 1 second (FEV 1) of intermediate time points (using times of ≤ 3 hours) from long ozone exposures experiments (short-int, red squares), and of the group mean changes from short ozone exposure experiments (short, red squares); also plotted are the 3rd order polynomial trend lines for each dataset; (B) Total inhaled ozone dose (in ppm x L) versus mean percent change in FEV1 of intermediate time points from long ozone exposure experiments, separated into shorter ≤ 3 hour exposures (short int, red diamonds) and longer > 5 hour exposures (long int, blue squares); also plotted are the 3 rd order polynomial trend lines for each dataset; (C) Total inhaled dose (in ppm x L) versus mean percent change in FEV1 of intermediate time points taken during the ozone exposures from 9 long-exposure clinical experiments, which used either square wave (blue squares) or stepped or ramped triangular wave (red triangle) exposures.

Figure 3. Individuals experiencing ≥ 10% FEV1 decrements and dose-response with subtraction of filtered air. (A) Percent of individuals who experienced FEV1 decrements greater than 10% at ozone dose increments of 250 ppm x L during short (≤ 3 hour; red squares) and long (6-8 hour; blue diamonds) exposures to ozone. (B) Trendlines of total inhaled ozone dose (in ppm x L) versus mean percent change in FEV1 of healthy young adults from the US EPA individual dataset exposed to ozone for ≤ 3 hours (short exposure) or 6-8 hours (long exposure) while exercising, with subtraction of filtered air responses from ozone responses for long (yellow line) and short (green line) exposures, as well as responses with filtered air not subtracted for long (blue line) and short (red line) exposures.

Figure 4. Ozone-FEV1 dose-response curves that include subpopulations. (A) Trend line of total inhaled dose (in ppm x L) versus mean percent change in forced expiratory volume in 1 second (FEV1) of healthy young adults exposed for ≤ 3 hours to ozone with 95% confidence intervals, plotted with the individual exposure data from healthy children aged 8-11 exposed to ozone (green diamonds) – marked is the 95% confidence interval of the child ozone exposure data; (B) Total inhaled dose versus group mean percent change in FEV1 of healthy young adults (dark red diamonds and trend line) exposed for ≤ 3 hours to ozone with 95% confidence intervals, and also the group mean responses of healthy children aged 8-11 exposed to ozone (grey circles), adult asthmatics (grey squares), healthy adolescents (grey upward triangles), and asthmatic adolescents (grey downward triangles); (C) Total inhaled ozone dose versus mean percent change in FEV1 of healthy

young adults (blue diamonds and trend lines) exposed for six to eight-hours to ozone and adult asthmatics (yellow diamonds).

Figure 5. Correlation between FEV1 responses of individuals to different exposures. FEV1

responses of individuals exposed to 80 ppb ozone or to filtered air (A), exposed to 100 ppb ozone or to filtered air (B), exposed to 120 ppb ozone or to filtered air (C), exposed to 100 ppb ozone or to 80 ppb ozone (D), exposed to 120 ppb ozone or to 80 ppb ozone (E), or exposed to 120 ppb ozone or to 100 ppb ozone (F); correlation coefficients and p-values are provided for each comparison.

Figure 6. Modeled ozone doses compared to ozone dose-response thresholds. Total mean (± SD) ozone dose associated with different exposure scenarios, calculated by multiplying the exposure duration and ventilation rate for each activity (from Table 4), with the ozone concentration for the appropriate duration that would occur if the ambient 8-hour maximum average ozone concentration were 75, 70 or 65 ppb (from Table 5). These are separated into short (≤ 4 hour; A) and longer (> 4 hour; B) exposure times. The threshold ozone dose at which group mean FEV1 decrements of 5%, 10% or 15% (15% threshold available for short exposure only) would be expected to occur is also plotted. Standard deviation based on variability in ozone concentrations, not on variability in exposure duration or ventilation rates.

Figure 1D

Figure 2

Figure 5

Figure 6 A

Threshold O3 dose for 15% FEV1

ReferencesATS. (2000). What constitutes an adverse health effect of air pollution? Am J Respir Crit Care Med

161:665-673.

Adams WC. (2000). Ozone dose-response effects of varied equivalent minute ventilation rates. J Expo Anal Environ Epidemiol 10:217-26.

Adams WC. (2002). Comparison of chamber and face-mask 6.6-hour exposures to ozone on pulmonary function and symptoms responses. Inhal Toxicol 14:745-64.

Adams WC. (2003a). Comparison of chamber and face mask 6.6-hour exposure to 0.08 ppm ozone via square-wave and triangular profiles on pulmonary responses. Inhal Toxicol 15:265-81.

Adams WC. (2003b). Relation of pulmonary responses induced by 6.6-h exposures to 0.08 ppm ozone and 2-h exposures to 0.30 ppm ozone via chamber and face-mask inhalation. Inhal Toxicol 15:745-59.

Adams WC. (2006a). Comparison of chamber 6.6-h exposures to 0.04-0.08 PPM ozone via square-wave and triangular profiles on pulmonary responses. Inhal Toxicol 18:127-36.

Adams WC. (2006b). Human pulmonary responses with 30-minute time intervals of exercise and rest when exposed for 8 hours to 0.12 ppm ozone via square-wave and acute triangular profiles. Inhal Toxicol 18:413-22.

Adams WC & Ollison WM. (1997). Effects of prolonged simulated ambient ozone dosing patterns on human pulmonary function and symptomatology. June 8-13, 97-MP9.02 1-16.

Aris RM, Christian D, Hearne PQ, Kerr K, Finkbeiner WE & Balmes JR. (1993). Ozone-induced airway inflammation in human subjects as determined by airway lavage and biopsy. Am Rev Respir Dis 148:1363-72.

Avol EL, Linn WS, Shamoo DA, Spier CE, Valencia LM, Venet TG, Trim SC & Hackney JD. (1987). Short-term respiratory effects of photochemical oxidant exposure in exercising children. JAPCA 37:158-62.

Balmes JR, Aris RM, Chen LL, Scannell C, Tager IB, Finkbeiner W, Christian D, Kelly T, Hearne PQ, Ferrando R & Welch B. (1997). Effects of ozone on normal and potentially sensitive human subjects. Part I: Airway inflammation and responsiveness to ozone in normal and asthmatic subjects. Res Rep Health Eff Inst 1-37; discussion 81-99.

Basha MA, Gross KB, Gwizdala CJ, Haidar AH & Popovich J, Jr. (1994). Bronchoalveolar lavage neutrophilia in asthmatic and healthy volunteers after controlled exposure to ozone and filtered purified air. Chest 106:1757-65.

Crump KS & Howe R 1983. A review of methods for calculating confidence in low dose extrapolation. In: Clayton, Krewski & Munroe (eds.) Toxicological Risk Assessment. Boca Raton, FL: CRC Press.

Folinsbee L, Drinkwater BL, Bedi JF & Horvath SM 1978. The influence of exercise on the pulmonary function changes due to low concentrations of ozone. In: Folinsbee, L., Wagner, J. A., Borgia, J. F., Drinkwater, B. L., Gliner, J. A. & Bedi, J. F. (eds.) Environmental Stress. New York.

Folinsbee LJ, Mcdonnell WF & Horstman DH. (1988). Pulmonary function and symptom responses after 6.6-hour exposure to 0.12 ppm ozone with moderate exercise. JAPCA 38:28-35.

Frampton MW, Balmes JR, Cox C, Krein PM, Speers DM, Tsai Y & Utell MJ. (1997). Effects of ozone on normal and potentially sensitive human subjects. Part III: Mediators of inflammation in bronchoalveolar lavage fluid from nonsmokers, smokers, and asthmatic subjects exposed to ozone: a collaborative study. Res Rep Health Eff Inst 73-9; discussion 81-99.

Goodman JE, Dodge DG & Bailey LA. (2010). A framework for assessing causality and adverse effects in humans with a case study of sulfur dioxide. Regul Toxicol Pharmacol 58:308-22.

Goodman JE, Prueitt RL, Chandalia J & Sax SN. (2014). Evaluation of adverse human lung function effects in controlled ozone exposure studies. J Appl Toxicol 34:516-24.

Goodman JE, Sax SN, Lange S & Rhomberg LR. (2015). Are the elements of the proposed ozone National Ambient Air Quality Standards informed by the best available science? Regul Toxicol Pharmacol 72:134-140.

Hazucha MJ, Bates DV & Bromberg PA. (1989). Mechanism of action of ozone on the human lung. J Appl Physiol (1985) 67:1535-41.

Hazucha MJ, Folinsbee LJ & Bromberg PA. (2003). Distribution and reproducibility of spirometric response to ozone by gender and age. J Appl Physiol (1985) 95:1917-25.

Hazucha MJ, Folinsbee LJ & Seal E, Jr. (1992). Effects of steady-state and variable ozone concentration profiles on pulmonary function. Am Rev Respir Dis 146:1487-93.

Holz O, Jorres RA, Timm P, Mucke M, Richter K, Koschyk S & Magnussen H. (1999). Ozone-induced airway inflammatory changes differ between individuals and are reproducible. Am J Respir Crit Care Med 159:776-84.

Horstman DH, Ball BA, Brown J, Gerrity T & Folinsbee LJ. (1995). Comparison of pulmonary responses of asthmatic and nonasthmatic subjects performing light exercise while exposed to a low level of ozone. Toxicol Ind Health 11:369-85.

Horstman DH, Folinsbee LJ, Ives PJ, Abdul-Salaam S & Mcdonnell WF. (1990). Ozone concentration and pulmonary response relationships for 6.6-hour exposures with five hours of moderate exercise to 0.08, 0.10, and 0.12 ppm. Am Rev Respir Dis 142:1158-63.

Isaacs K, Glen G, Mccurdy T & Smith L. (2008). Modeling energy expenditure and oxygen consumption in human exposure models: accounting for fatigue and EPOC. J Expo Sci Environ Epidemiol 18:289-98.

Jenkins HS, Devalia JL, Mister RL, Bevan AM, Rusznak C & Davies RJ. (1999). The effect of exposure to ozone and nitrogen dioxide on the airway response of atopic asthmatics to inhaled allergen: dose- and time-dependent effects. Am J Respir Crit Care Med 160:33-9.

Kehrl HR, Peden DB, Ball B, Folinsbee LJ & Horstman D. (1999). Increased specific airway reactivity of persons with mild allergic asthma after 7.6 hours of exposure to 0.16 ppm ozone. J Allergy Clin Immunol 104:1198-204.

Kehrl HR, Vincent LM, Kowalsky RJ, Horstman DH, O'neil JJ, Mccartney WH & Bromberg PA. (1987). Ozone exposure increases respiratory epithelial permeability in humans. Am Rev Respir Dis 135:1124-8.

Kim CS, Alexis NE, Rappold AG, Kehrl H, Hazucha MJ, Lay JC, Schmitt MT, Case M, Devlin RB, Peden DB & Diaz-Sanchez D. (2011). Lung function and inflammatory responses in healthy young adults exposed to 0.06 ppm ozone for 6.6 hours. Am J Respir Crit Care Med 183:1215-21.

Koenig JQ, Covert DS, Hanley QS, Van Belle G & Pierson WE. (1990). Prior exposure to ozone potentiates subsequent response to sulfur dioxide in adolescent asthmatic subjects. Am Rev Respir Dis 141:377-80.

Koenig JQ, Covert DS, Marshall SG, Van Belle G & Pierson WE. (1987). The effects of ozone and nitrogen dioxide on pulmonary function in healthy and in asthmatic adolescents. Am Rev Respir Dis 136:1152-7.

Koenig JQ, Covert DS, Morgan MS, Horike M, Horike N, Marshall SG & Pierson WE. (1985). Acute effects of 0.12 ppm ozone or 0.12 ppm nitrogen dioxide on pulmonary function in healthy and asthmatic adolescents. Am Rev Respir Dis 132:648-51.

Koenig JQ, Covert DS, Smith MS, Van Belle G & Pierson WE. (1988). The pulmonary effects of ozone and nitrogen dioxide alone and combined in healthy and asthmatic adolescent subjects. Toxicol Ind Health 4:521-32.

Koren HS, Devlin RB, Graham DE, Mann R, Mcgee MP, Horstman DH, Kozumbo WJ, Becker S, House DE, Mcdonnell WF & Et Al. (1989). Ozone-induced inflammation in the lower airways of human subjects. Am Rev Respir Dis 139:407-15.

Lee K, Parkhurst WJ, Xue J, Ozkaynak AH, Neuberg D & Spengler JD. (2004). Outdoor/Indoor/Personal ozone exposures of children in Nashville, Tennessee. J Air Waste Manag Assoc 54:352-9.

Linn WS, Shamoo DA, Anderson KR, Peng RC, Avol EL & Hackney JD. (1994). Effects of prolonged, repeated exposure to ozone, sulfuric acid, and their combination in healthy and asthmatic volunteers. Am J Respir Crit Care Med 150:431-40.

Mcdonnell WF 1989. Individual variability in the magnitude of acute respiratory response to ozone exposure. In: Frank (ed.) Susceptibility to Inhaled Pollutants. Philadelphia: American Society for Testing and Materials.

Mcdonnell WF, 3rd, Chapman RS, Leigh MW, Strope GL & Collier AM. (1985). Respiratory responses of vigorously exercising children to 0.12 ppm ozone exposure. Am Rev Respir Dis 132:875-9.

Mcdonnell WF, Horstman DH, Hazucha MJ, Seal E, Jr., Haak ED, Salaam SA & House DE. (1983). Pulmonary effects of ozone exposure during exercise: dose-response characteristics. J Appl Physiol Respir Environ Exerc Physiol 54:1345-52.

Mcdonnell WF, Kehrl HR, Abdul-Salaam S, Ives PJ, Folinsbee LJ, Devlin RB, O'neil JJ & Horstman DH. (1991). Respiratory response of humans exposed to low levels of ozone for 6.6 hours. Arch Environ Health 46:145-50.

Mcdonnell WF, Stewart PW & Smith MV. (2007). The temporal dynamics of ozone-induced FEV1 changes in humans: an exposure-response model. Inhal Toxicol 19:483-94.

Mcdonnell WF, Stewart PW, Smith MV, Kim CS & Schelegle ES. (2012). Prediction of lung function response for populations exposed to a wide range of ozone conditions. Inhal Toxicol 24:619-33.

Mcdonnell WF, Stewart PW, Smith MV, Pan WK & Pan J. (1999). Ozone-induced respiratory symptoms: exposure-response models and association with lung function. Eur Respir J 14:845-53.

Mudway IS, Blomberg A, Frew AJ, Holgate ST, Sandstrom T & Kelly FJ. (1999). Antioxidant consumption and repletion kinetics in nasal lavage fluid following exposure of healthy human volunteers to ozone. Eur Respir J 13:1429-38.

Mudway IS, Housley D, Eccles R, Richards RJ, Datta AK, Tetley TD & Kelly FJ. (1996). Differential depletion of human respiratory tract antioxidants in response to ozone challenge. Free Radic Res 25:499-513.

Mudway IS & Kelly FJ. (1998). Modeling the interactions of ozone with pulmonary epithelial lining fluid antioxidants. Toxicol Appl Pharmacol 148:91-100.

Mudway IS & Kelly FJ. (2004). An investigation of inhaled ozone dose and the magnitude of airway inflammation in healthy adults. Am J Respir Crit Care Med 169:1089-95.

Nightingale JA, Rogers DF & Barnes PJ. (1999). Effect of inhaled ozone on exhaled nitric oxide, pulmonary function, and induced sputum in normal and asthmatic subjects. Thorax 54:1061-9.

Parsons JP, Hallstrand TS, Mastronarde JG, Kaminsky DA, Rundell KW, Hull JH, Storms WW, Weiler JM, Cheek FM, Wilson KC & Anderson SD. (2013). An official American Thoracic Society clinical practice guideline: exercise-induced bronchoconstriction. Am J Respir Crit Care Med 187:1016-27.

Passannante AN, Hazucha MJ, Bromberg PA, Seal E, Folinsbee L & Koch G. (1998). Nociceptive mechanisms modulate ozone-induced human lung function decrements. J Appl Physiol (1985) 85:1863-70.

Peden DB, Boehlecke B, Horstman D & Devlin R. (1997). Prolonged acute exposure to 0.16 ppm ozone induces eosinophilic airway inflammation in asthmatic subjects with allergies. J Allergy Clin Immunol 100:802-8.

Pellegrino R, Viegi G, Brusasco V, Crapo RO, Burgos F, Casaburi R, Coates A, Van Der Grinten CP, Gustafsson P, Hankinson J, Jensen R, Johnson DC, Macintyre N, Mckay R, Miller MR, Navajas D, Pedersen OF & Wanger J. (2005). Interpretative strategies for lung function tests. Eur Respir J 26:948-68.

Pryor WA. (1992). How far does ozone penetrate into the pulmonary air/tissue boundary before it reacts? Free Radic Biol Med 12:83-8.

Que LG, Stiles JV, Sundy JS & Foster WM. (2011). Pulmonary function, bronchial reactivity, and epithelial permeability are response phenotypes to ozone and develop differentially in healthy humans. J Appl Physiol (1985) 111:679-87.

Samet JM, Lambert WE, James DS, Mermier CM & Chick TW. (1993). Assessment of heart rate as a predictor of ventilation. Res Rep Health Eff Inst 19-55; discussion 57-69.

Sarnat JA, Brown KW, Schwartz J, Coull BA & Koutrakis P. (2005). Ambient gas concentrations and personal particulate matter exposures: implications for studying the health effects of particles. Epidemiology 16:385-95.

Sarnat JA, Schwartz J, Catalano PJ & Suh HH. (2001). Gaseous pollutants in particulate matter epidemiology: confounders or surrogates? Environ Health Perspect 109:1053-61.

Schelegle ES, Adams WC, Walby WF & Marion MS. (2012). Modelling of individual subject ozone exposure response kinetics. Inhal Toxicol 24:401-15.

Schelegle ES, Morales CA, Walby WF, Marion S & Allen RP. (2009). 6.6-hour inhalation of ozone concentrations from 60 to 87 parts per billion in healthy humans. Am J Respir Crit Care Med 180:265-72.

Schelegle ES, Siefkin AD & Mcdonald RJ. (1991). Time course of ozone-induced neutrophilia in normal humans. Am Rev Respir Dis 143:1353-8.

Seal E, Jr., Mcdonnell WF, House DE, Salaam SA, Dewitt PJ, Butler SO, Green J & Raggio L. (1993). The pulmonary response of white and black adults to six concentrations of ozone. Am Rev Respir Dis 147:804-10.

Stenfors N, Pourazar J, Blomberg A, Krishna MT, Mudway I, Helleday R, Kelly FJ, Frew AJ & Sandstrom T. (2002). Effect of ozone on bronchial mucosal inflammation in asthmatic and healthy subjects. Respir Med 96:352-8.

USEPA. (1989). Review of the National Ambient Air Quality Standard for Ozone. Assessment of Scientific and Technical Information. EPA-450/2-92:Research Triangle Park, NC.

USEPA. (1994). Methods for derivation of inhalation reference concentrations and application of inhalation dosimetry. EPA/600/8-90/066F, Washington DC.

USEPA. (2009). Metabolically derived human ventilation rates: a revised approach based upon oxygen consumption rates. National Center for Environmental Assessment (NCEA), EPA/600/R-06/129F, May, 153 p.

USEPA. (2013). Integrated Science Assessment for Ozone and Related Photochemical Oxidants (Final). National Center for Environmental Assessment (NCEA), EPA/600/R–10/076F, February, 1251 p.

USEPA. (2014a). Health Risk and Exposure Assessment for Ozone (Final Report). EPA-452/R-14-004a, August, 502 p.

USEPA. (2014b). National ambient air quality standards for ozone (Proposed rule). 40 CFR Parts 50, 51, 52, 53 and 58.

USEPA. (2014c). Regulatory Impact Analysis of the Proposed Revisions to the National Ambient Air Quality Standards for Ground-Level Ozone. EPA-452/P-14-006, November, 575 p.

Weschler CJ. (2000). Ozone in indoor environments: concentration and chemistry. Indoor Air 10:269-88.

Weschler CJ. (2015). Roles of the human occupant in indoor chemistry. Indoor Air.

Weymer AR, Gong H, Jr., Lyness A & Linn WS. (1994). Pre-exposure to ozone does not enhance or produce exercise-induced asthma. Am J Respir Crit Care Med 149:1413-9.

Wheeler MW. (2005). Benchmark dose estimation using SAS. SUGI 30 Proceedings 201-230.

Zuurbier M, Hoek G, Van Den Hazel P & Brunekreef B. (2009). Minute ventilation of cyclists, car and bus passengers: an experimental study. Environ Health 8:48.

texas commission on environmental quality€¦ · web viewozone is a secondary pollutant, and it is...

Documents