Appendix C

Assessment of Eight Controlled Human Exposure Studies

The U.S. Environmental Protection Agency (EPA) provided eight CHIE studies to the committee for its consideration.¹ The studies were selected from among the CHIE studies at the EPA Human Studies Facility that were active at some point in time from January 2009 to October 2016. See Table C-1. The eight studies are:

The committee used the following considerations:

1. Research Question:

• Is the research question well focused?
• Is there a clear hypothesis that is testable?

2. Background, Gap Analysis, and Rationale

• Range and variation of pollutant exposures in the United States and perhaps elsewhere.
• Relevance of the condition chosen for study in the general population (e.g., age, disease, etc.).
• Current air quality standard and the relevance for future reviews of the standard.

__________________

¹ An application for institutional review board approval and a consent to participate in a research study for each CHIE study was provided by EPA on November 19, 2014. Publications resulting from those eight studies are Devlin et al. 2014 and Ghio et al. 2012 for XCON; Tong et al. 2012, 2015 for OMEGACON; Madden et al. 2014; Stiegel et al. 2015; Stiegel et al. 2016 for DEPOZ.

• Critical toxicologic pathways and evidence of perturbations.
• Document citations that support a need for this research (i.e., strategic plan, consensus statements, etc.).
• Remaining information needs after current and forthcoming epidemiologic and animal/in vitro toxicologic data are considered.
• Additional knowledge and/or level of certainty that this controlled human exposure study would provide.
• Research goal in the regulatory context of providing public health protection, including the identification and protection of sensitive subpopulations.
• Research goal regarding the
  • Relationship between physiologic function and pollutant exposure,
  • Biologic plausibility and/or mechanisms of air pollution health effects,
  • Interpretation of effects observed in toxicologic or epidemiologic studies, and
  • Other: Statistical analysis plan in detail at proposal stage.

3. Study Design

   Does the design of the study adequately reflect the information uncertainty being addressed?

4. Subject Selection

   Is the health status of the study subjects reflective of the sensitive subgroups that the relevant air pollution standard is intended to protect?

5. Experimental Methods

   Is the study method appropriate?

6. Exposure Protocol

   Was the choice of the study’s exposure concentration and duration appropriate?

7. End points
   • Was the choice of study end points appropriate for the experimental question?
   • Were the time points for measurement appropriate?

8. Analysis

   Was the statistical analysis appropriate?

9. Interpretation and Generalizability of the Findings (expected or reported)
   • What are the strengths and limitations of the study?
   • What are the major findings of the study?
   • What are the remaining uncertainties?
   • Describe how the findings have or may contribute to the following aspects:
   • Quantification of the relationship between physiologic function and pollutant exposure,
   • Understanding biologic plausibility of effects of concern,
   • Increased ability to interpret effects observed in toxicologic or epidemiologic studies, and
   • Other: Contributions of publications out of the study.

TABLE C-1 Controlled Inhalation Exposure Studies in EPA Human Studies Facility^a

^aThe studies were active at some point from January 2009 to October 2016.

^b”UNC” means the study was conducted in the EPA Human Studies Facility but initiated or led by UNC investigators, many with EPA collaboration.

Source: EPA, unpublished material, submitted November 2, 2016.

CARDIOPULMONARY RESPONSES TO EXPOSURE TO OZONE AND DIESEL EXHAUST WITH MODERATE EXERCISE IN HEALTHY ADULTS (DEPOZ)

Is the research question well focused?

This study focused on the joint effect on lung function of diesel-engine exhaust (DE) exposure and ozone (O₃), and on the effect of prior exposures to DE, O₃, and DE+O₃ on the subsequent response to O₃ alone. These effects were measured by sampling forced vital capacity (FVC) and forced expiratory volume for 1 second (FEV₁). In addition to samples of FVC and FEV₁, the protocol called for monitoring of heart rate variability and blood pressure, and collection of samples of blood, saliva, and urine, and included information on cardiovascular function. However, there was no clear indication of how these data would be statistically analyzed or otherwise used to study cardiovascular responses to O₃ and DE. The published report from this study (Madden et al., 2014) does not report on cardiovascular effects.

Is there a clear hypothesis that is testable?

There were three specific hypotheses tested by this study: It was hypothesized that an exposure to DE with O₃ (day 1) or DE exposure (day 1) given prior to O₃ exposure (day 2) would not induce a significant decrement in lung function in healthy young adults as a group relative to O₃ alone; that an O₃ exposure (on day 2) after the DE and O₃ coexposure (day 1) would cause significant cardiopulmonary responses in healthy young adults; and that 2 consecutive days of O₃ exposure would affect cardiovascular responses. The first two hypotheses involving pulmonary responses could be tested using the data collected on FVC and FEV₁.

The hypothesis concerned with cardiovascular responses is not as specific as those involving lung function and the protocol does not state how this hypothesis would be tested. The statistical power calculation is based only pulmonary function. Thus, it is not clear that the hypothesis involving cardiovascular responses was testable.

Range and variation of pollutant exposures in the United States and perhaps elsewhere.

The proposal does not contain a review of pollution levels in various locations. In the information provided to participants, some largely anecdotal information regarding typical DE levels in various situations are provided (truck drivers, miners, near busy intersections), but complete references are not provided.

Relevance of the condition chosen for study in the general population (e.g., age, disease, etc.).

This study focused on the joint effect on lung function of DE and O₃ exposure. People of all ages are exposed to these contaminants.

Current air quality standard and the relevance for future reviews of the standard.

Air quality standards related to DE and O₃ are not discussed in the proposal. Also, there is no description of how the results of this study could affect air quality standards, other than a general statement that “[t]he data obtained from this study will contribute to the overall assessment of air pollution effects in the U.S. and thereby may influence future health policy.”

Critical toxicologic pathways and evidence of perturbations

For DE exposures, evidence is presented of adverse cardiopulmonary effects including premature mortality, cardiopulmonary problems including infections, exacerbation of asthma symptoms, and heart attacks. For O₃ exposures, evidence is presented for decrements of lung function and an influx of neutrophils and other markers of inflammation. Studies are cited that indicate persons with a glutathione-S-transferase Ml (GSTM1) null genotype may be particularly susceptible to the effects of O₃ exposure on lung function.

Document citations that support a need for this research (i.e., strategic plan, consensus statements, etc.).

There are no documents cited that state a need for this research. It is stated without citation that it is not known whether coexposure to both O₃ and DE, as would occur when exposed to polluted ambient air, can induce additive or synergistic effects, and also whether exposure to DE, or DE with O₃, can alter a subsequent exposure to O₃.

Remaining information needs after current and forthcoming epidemiologic and animal/in vitro toxicologic data are considered.

This study only investigated responses to O₃ and DE under very specific and not very realistic conditions. There still remain questions regarding responses that would be expected under variable and more realistic conditions.

Additional knowledge and/or level of certainty that this controlled human exposure study would provide.

This study could reinforce the indications that exposure to O₃ can have reversible effects on pulmonary function. It could also provide information on whether the combination of DE and O₃ exposure would have a greater than additive effect and provide some information on the level of such an effect.

Research goal in the regulatory context of providing public health protection, including the identification and protection of sensitive subpopulations.

The relationship between the results of this study and how the results might be used in setting regulations is not discussed directly. The study does address whether persons with a GSTM1 null genotype may be a sensitive subpopulation that is particularly susceptible to the effects of O₃ and DE on lung function.

Research goal regarding the...
Relationship between physiologic function and pollutant exposure

The research goals including comparing short-term lung function resulting from simultaneous exposure to O₃ and DE to that resulting from separate exposures to these substances and also to compare the effect of prior exposures to O₃ and DE to lung function after exposure to O₃ only.

Biologic plausibility and or mechanisms of air pollution health effects

One research goal was to elucidate the effect of the GSTM1 null genotype on the effects of O₃ exposure on lung function.

Interpretation of effects observed in toxicologic or epidemiologic studies

It had been observed in a mouse model that both DE and O₃ were needed to increase lung resistance, which was not observed with individual pollutant exposures (Madden et al. 2014). DEPOZ investigated whether similar responses would be present in humans.

Does the design of the study adequately reflect the information uncertainty being addressed?

The goals of the study involved clarification of short-term effects of exposure regimens, as opposed to possible long-term effects. The study design was appropriate for this limited goal.

Is the health status of the study subjects reflective of the sensitive subgroups that the relevant air pollution standard is intended to protect?

The subjects in this study were selected from among healthy relatively young adults (men and women between the ages of 18 and 55 years). Therefore, their responses will possibly not be reflective of responses that might occur in the elderly or the very young or in those who already have compromised health.

Is the study method appropriate?

The study method is appropriate for testing the narrow hypotheses being tested.

Was the choice of the study’s exposure concentration and duration appropriate?

The rationale for the exposure concentration and duration are not clearly stated in the proposal. There is no clear description of the rationale for the target exposures selected (0.3 ppm O₃, and 300 µg/m³ DE). Several studies are cited that exposed volunteers to 300 µg/m³ DE. A number of studies involving O₃ exposures are cited but this information is not tied to the exposure level of 0.3 ppm used in this study. Apparently this team has experience with exposing O₃ to volunteers at levels around 0.3 ppm. But this experience is not clearly stated in the application or referred to in support the 0.3 ppm exposure level.

Also, there is no mention of the type of diesel engine that would be used to generate the exhaust to which the volunteers would be exposed to. Diesel technology has undergone important changes in recent years

and DE emissions have changed both quantitatively and qualitatively (Hesterberg et al., 2011 McClellan et al. 2012, Khalek et al. 2011, 2015). Also, diesel emissions are composed of gases and particulate of various sizes. The protocol does not state what fraction of this complex mixture will be used to quantify DE. For a full characterization of DE exposure, information on the diesel engine to be used to generate the DE should have been provided in the proposal along with specific information on how DE is to be measured. The manufacturer and model of the diesel engine used to generate the DE was provided in the published paper (but not in the protocol), although no characterization of the emissions from this engine were provided (Madden et al., 2014).

Was the choice of study end points appropriate for the experimental question?

The principle end points were indicators of lung function as measured by FEV₁ and FVC. These end points were sufficient to answer questions about effects on lung function. However, they provided little direct information on effects on cardiovascular function.

Were the time points for measurement appropriate?

FVC₁ and FEV were measured immediately following the completion and hourly thereafter for 4 hours, and again 20 hours following the completion of exposure. This was sufficient to indicate the trend in rebound in FVC₁ measurements, and FVC₁ measurements had returned to baseline or nearly so after 20 hours (Madden et al., 2014).

Was the statistical analysis appropriate?

The data obtained in the study were analyzed using a repeated measures analysis of variance (ANOVA) parametric test. This type of test is appropriate for a randomized crossover design like that used in this study. The statistical power calculation indicated that 14 subjects would provide 80% power for detecting a 10% decrease in FEV₁ from O₃ exposure. This calculation seems appropriate. However, this power calculation applies only to the FEV₁ data and does not provide information on the power from analysis of cardiovascular data.

What are the strengths and limitations of the study?

The study design and group size was adequate to study the limited hypotheses of the study. Like most studies of this type it involved a limited number of healthy individuals exposed to a very small range of exposure conditions. Consequently, it would be problematic to use the results from this study to predict responses in the population at large.

What are the major findings of the study?

Results from this study suggest that altered respiratory responses to the combination of O₃ and DE exposure can occur in a greater than additive manner, and O₃-induced lung function decrements can be greater with a prior exposure to DE compared to a prior exposure to filtered air (Madden et al. 2014).

What are the remaining uncertainties?

Like all studies of limited and prescribed exposures to volunteers, many questions remain. The study exposed healthy volunteers between the ages of 18 and 55. Effects upon the very young and old and upon those with existing health conditions were not investigated. The study involved a limited number of exposure conditions, none of which are typical of real-world exposures, so using the results of the study to predict responses in real-world situations would be problematic.

Quantification of the relationship between physiologic function and pollutant exposure.

The study provided quantitative information on the physiologic response to exposure to O₃, DE, and O₃+DE, and response to O₃ following earlier exposures O₃, DE, and O₃+DE, in a few healthy volunteers, under specific exposure conditions. It could be problematic to extrapolate these quantitative responses to more general situations.

Understanding biologic plausibility of effects of concern.

The results of this study reinforced the biologic plausibility of effects upon lung function from exposure to O₃ and O₃+DE.

Increased ability to interpret effects observed in toxicologic or epidemiologic studies.

The results from this study could possibly provide reinforcement to interpretations of effects observed other studies.

EFFECTS OF SEQUENTIAL EXPOSURE TO NITROGEN DIOXIDE AND OZONE IN HEALTHY ADULT HUMAN VOLUNTEERS (ENDZONE)

Is the research question well focused? Yes.

Is there a clear hypothesis that is testable? Yes.

Range and variation of pollutant exposures in the United States and perhaps elsewhere. Good.

Relevance of the condition chosen for study in the general population (e.g., age, disease, etc). Appropriate.

Current air quality standard and the relevance for future reviews of the standard. Good.

Critical toxicologic pathways and evidence of perturbations.

Adequate, in consideration of the paucity of relevant prior published research.

Document citations that support a need for this research (i.e., strategic plan, consensus statements, etc.). Good.

Remaining information needs after current and forthcoming epidemiologic and animal/in vitro toxicologic data are considered. Good.

Additional knowledge and/or level of certainty that this controlled human exposure study would provide.

Yes, it would generate useful and interesting additional knowledge. However, it would be inadequate, with regard to level of certainty about the real-world effects of sequential inhalation exposures to O₃ and NO₂.

Research goal in the regulatory context of providing public health protection, including the identification and protection of sensitive subpopulations.

Not applicable insofar as known sensitive subpopulations are not being studied.

Research goal regarding the...
Relationship between physiologic function and pollutant exposure.

Limited, insofar as only one concentration and one exposure duration are specified.

Biologic plausibility and or mechanisms of air pollution health effects.

It could provide valuable increments on the biologic mechanisms contributing to the effects of inhaled O₃ and NO₂ alone, and especially on any additional responses resulting from sequential exposures.

Interpretation of effects observed in toxicologic or epidemiologic studies

1. New interpretations of the toxicologic effects of inhaled O₃ are unlikely, since these effects of short-term exposures to O₃ alone are already well studied and described;
2. New interpretations of the toxicologic effects of inhaled NO₂ are possible, but also unlikely, since there has been virtually no consistency in reported effects of short-term inhalation exposures to NO₂ alone in previous studies with single concentrations in the same range;
3. New interpretations of the acute epidemiologic associations of inhaled O₃ are unlikely, since these associations are already well studied and described;
4. New interpretations of any acute epidemiologic effects of inhaled NO₂ alone are unlikely given that any previously reported associations of NO₂ with either acute or chronic inhalation exposures are as, or more, likely to be due to copollutants; and
5. New interpretations of the toxicologic effects of sequential inhalation exposures to O₃ and NO₂ are unknowable in advance, but could be interesting and important if they occur.

Other.

Direct coherence of physiologic responses to inhaled O₃ and NO₂ from toxicologic and epidemiologic studies, and their concentration–response relationships, as they may affect the selection of short-term National Ambient Air Quality Standards (NAAQS), is unlikely, because controlled human exposures are (a) almost always at constant concentrations for relatively short time intervals, while ambient air exposures vary during the day, extend over longer time intervals, and vary with location within a community; (b) almost always to O₃ or NO₂ alone, while ambient air exposures always include other gaseous and particulate matter (PM) components that can produce the same or similar responses; and (c) O₃ and NO₂ are index pollutants for the NAAQS for photochemical oxidants and NO_x, respectively. Epidemiologic responses may be due, in part, to peroxides and/or nitric acid in the ambient air as well as O₃ and NO₂.

Does the design of the study adequately reflect the information uncertainty being addressed?

No, with respect to the temporal sequences of the exposures. For a study of the physiologic effects of sequential exposures of inhalation exposures to O₃ and NO₂, it was disappointing that there was no justification provided for the temporal sequences. As noted in the submission, the real-world sequential exposures involve peak morning exposures to NO₂ followed by peak early afternoon exposures to O₃. Why then select the second of the sequential 2-hour controlled inhalation exposures 24 hours later? The premise of the statement that the effects of the ambient air exposures to O₃ or NO₂ may be severe and delayed seem odd in view of the absence of literature documenting severe effects.

Is the health status of the study subjects reflective of the sensitive subgroups that the relevant air pollution standard is intended to protect?

No, but the population to be studied is the correct one for this exploratory study of sequential exposures.

Is the study method appropriate? Yes.

Was the choice of the study’s exposure concentration(s) and duration appropriate?

They were reasonable.

Was the choice of study end points appropriate for the experimental question? Yes.

Were the time points for measurement appropriate?

Yes, but measurements at additional time points could have been very informative.

Was the statistical analysis appropriate? Yes.

What are the strengths and limitations of the study?

Summarized above.

EPIGENETIC EFFECT MODIFICATIONS WITH OZONE EXPOSURE ON HEALTHY VOLUNTEERS (GEMINOZ)

Is the research question well focused?

The purpose of this protocol is to assess whether epigenetic factors in (1) healthy individuals or (2) individuals with the same genetic makeup (i.e., identical twins) make a person more or less responsive to inflammation following ozone exposures. Yes, the research question is well focused.

Is there a clear hypothesis that is testable? Yes there is a clear hypothesis that is testable.

Range and variation of pollutant exposures in the Untied States and perhaps elsewhere.

The proposal does not contain a review of pollution levels in various locations. In the information provided to participants on air pollution and heart and pulmonary disease it does not state it.

Relevance of the condition chosen for study in the general population (e.g., age, disease, etc.).

Yes. This study focused on individuals 18-40 years of age, healthy, nonsmoking, and from the mid-Atlantic twins registry and healthy age-similar controls.

Document citations that support a need for this research (i.e., strategic plan, consensus statements, etc.).

Yes. This is a study of exposure to ozone and clean air, and exposure to ozone and its detrimental effects on the pulmonary system are well documented. Responses of primary interest are pulmonary: lung function, lung inflammation, and epigenetic changes as evaluated in bronchoalveolar lavage.

Relationship between physiologic function and pollutant exposure.

The research goals include comparing short-term lung function. Epigenetic responses obtained from the alveolar cell after pollutant exposure. The hypothesis is that pollutant exposure will modulate the epigenome and this will trigger gene expression that could result in proteins produced that alter pulmonary function. The study will be done in normal controls, and in monozygotic (MZ) twins to evaluate the biologic changes (epigenome) after pollutant exposure.

Biologic plausibility and or mechanisms of air pollution health effects

That there is biologic plausibility response to ozone in the lung may depend on the epigenome and this crossover study in MZ twins may confirm that observation or hypothesis and will lead to new research and possible risk factor identification of ozone and lung injury.

Does the design of the study adequately reflect the information uncertainty being addressed?

The goals of the study involved clarification of short-term effects of lung inflammation in MZ twins exposed in a crossover design to either clean air, then ozone, or in reverse. The study subjects are between the ages of 18 and 40 years and there are no significant health concerns. The study design was appropriate for this limited goal.

Is the health status of the study subjects reflective of the sensitive subgroups that the relevant air pollution standard is intended to protect?

The subjects in this study were selected from among MZ twin healthy adults (men and women between the ages of 18 and 40 years).

Is the study method appropriate?

The study method is appropriate for testing the narrow hypotheses being tested.

Was the choice of the study’s exposure concentration and duration appropriate?

The rationale for the exposure concentration and duration are similar to their published protocols from this group. The subjects will all be exposed to clear air then will be randomized to ozone or clean air, then after 13 days there will be a crossover. Responses of interest include lung inflammation, lung function, and epigenetic changes evaluated by bronchopulmonary lavage. The end point was chosen because other work has shown that epithelial cells lining the airways are the first target of ozone and respond by making proinflammatory cytokines such as IL-6 and IL-8. Epigenetic changes are dependent on tissue type, and airway epithelial cells can be obtained from brush biopsies during bronchoscopy and assayed for epigenetic changes. The study’s aim will determine if baseline epigenetic profiles between subjects are associated with responsiveness to ozone and if ozone exposure itself causes acute changes in a subject’s epigenome.

Was the statistical analysis appropriate?

Yes. The analysis will estimate the within-twin associations for twin subjects and overall associations for nontwin subjects between ozone response and gene-specific methylation. They propose to take a pathway approach where they estimate associations between inflammatory pathway–specific combinations of gene-specific methylation scores most associated with ozone responsiveness. Specifically they will look at the percent change of FEV₁, percent PMNs (type of white blood cell), IL-8, and PG-2 present in the lung between baseline and exposure conditions. For twins, the outcomes will be contrasted with those of the corresponding twin.

What are the strengths and limitations of the study?

This study’s goal is to detect within-twin correlation for twin subjects and overall associations for nontwin subjects between methylation and ozone susceptibility, and all power calculations were made under the assumption of two-sided tests at a significance level, global at the pathway level, of 0.05 response, combined with a conservative Bonferroni correlation for multiple tests across genes within a pathway. For the inflammatory pathway of 20 genes for n=50 twin pairs or 50 nontwin subjects, p = 0.025, 80% power, to detect a correlation of 0.43 between ozone, measured continuously, and methylation. The power calculations were made under the assumption of two-sided tests, global at the pathway level.

Limitations of this study are the MZ twins, in which the environmental exposure is the only difference between them, so the association is probably going to be as good as it could get. This study will be very challenging to duplicate in the general population; however, these results if definitive could lead the way to new treatments to reduce the toxicities of air pollution.

The study design and group size was adequate to study the limited hypotheses of the study. Like most studies of this type it involved a limited number of healthy individuals exposed to a very small range of exposure conditions. Consequently, it would be problematic to use the results from this study to predict responses in the population at large.

MECHANISMS BY WHICH AIR POLLUTION PARTICLES EXACERBATE ASTHMA IN OLDER ADULTS WITH MILD ASTHMA (KINGCON)

Is the research question well focused? Yes.

Is there a clear hypothesis that is testable? Yes.

Range and variation of pollutant exposures in the United States and perhaps elsewhere. Adequately described.

Relevance of the condition chosen for study in the general population (e.g., age, disease, etc.). Appropriate importance.

Current air quality standard and the relevance for future reviews of the standard.

Current air quality standards and the relevance to future studies. The citations state that particulate matter is an important part of air pollution and concentrations of exposure in the study are cited. But the standard is neither stated nor reviewed other than stating the exposure in the study will not exceed the total exposure encountered in 24 hours on a typical urban smoggy day. Future reviews of the standard are not addressed.

Critical toxicologic pathways and evidence of perturbations. Discussion appropriate.

Document citations that support a need for this research (i.e., strategic plan, consensus statements, etc.). Acceptable, extensive citations.

Remaining information needs after current and forthcoming epidemiologic and animal/in vitro toxicologic data are considered.

There is very limited discussion of remaining information needs after current study. Forthcoming studies are not addressed except for the storage of blood RNA/DNA sample.

Additional knowledge and/or level of certainty that this controlled human exposure study would provide.

Yes, further details on pulmonary function of aging asthmatics will be determined but the level of certainty is limited.

Research goal in the regulatory context of providing public health protection, including the identification and protection of sensitive subpopulations.

Nothing is addressed. Although standards for environmental exposures are not cited the implication of these studies has great importance for standard setting. This presentation to the Institutional Reveiw Board (IRB) would be clearer if the principles and background were presented in the study submission.

Does the design of the study adequately reflect the information uncertainty being addressed? Include consideration of the outcome measures (effects which may occur readily and be mild) relative to the key health effects of concern for the studied air pollutant (which may be severe and delayed).

The study design reflects information uncertainty. This is a study of short-term exposure addressing acute phase reactants, acute oxidative stress, and immediate physiologic responses. The outcome measures will need evaluation with regard to long-term potential toxic effects from single exposures, cumulative effects, chronic manifestations, and specific interactions of exposures. The effect of age and genetic susceptibility will be better addressed. The study of specific inhalants for those with more consequential and different airway diseases may lead to more specific and refined study of current air quality standards. There are few long-term or broad hypotheses for future work offered in the submission.

Is the health status of the study subjects reflective of the sensitive subgroups that the relevant air pollution standard is intended to protect?

The specificity of age, pulmonary function, chronic diseases, and drug or medication use are well presented. There is confusion between Exclusion Criterion 1—stating severe asthma and Exclusion Criterion 4—stating moderate to severe asthma, although elsewhere under subjects it is stated only patients with mild

asthma are included. These restrictive entry criteria lay the groundwork for future studies to define populations at other levels of risk.

Is the study method appropriate?

Yes, the methods are rigorous and appropriate.

Was the choice of the study’s exposure concentration and duration appropriate?

The authors do not offer current air quality standard data, suggesting only the maximum exposure is comparable to a smoggy urban day. In addition, they suggest neither the risk of >600 µg/m³ particulate mass levels nor the reason for a 10-minute delay prior to termination of the study. It is not clear whether since the concentration may exceed 600 µg/m³ (by an unknown amount) for 16 minutes prior to study termination that the risk is adequately addressed.

Was the choice of study end points appropriate for the experimental question? Yes.

Were the time points for measurement appropriate?

These end points are chosen for immediate evaluation and evaluation at 24 hours. This study exposure is not expected to have long-term effects and was not designed to assess that risk or effect.

Was the statistical analysis appropriate? Yes.

Interpretation and generalizability of the findings (expected or reported)

There is one spelling error— “immunodeficiency” and one misleading abbreviation—the term “chronic respiratory diseases” stands alone—but it must include chronic asthma (mild). It should be rewritten as “chronic respiratory diseases including moderate to severe asthma.”

CARDIO-PROTECTIVE EFFECTS OF OMEGA-3 FATTY ACIDS SUPPLEMENTATION IN HEALTHY OLDER SUBJECTS EXPOSED TO AIR POLLUTION PARTICLES (OMEGACON)

Is the research question well focused?

The purpose of this study was to examine the health effects of fine and ultrafine particulate matter (PM<2.5) exposure on the cardiovascular system and examine whether omega-3 fatty acids supplementation pretreatment would attenuate the adverse cardiovascular effects. A strong positive outcome would provide support for the EPA to advocate a prevention strategy. Overall, the research question is well focused as laboratory data support the role of omega-3 fatty acids as modulating a toxic response.

Is there a clear hypothesis that is testable?

There is a clear hypothesis that fine and ultrafine ambient PM (PM< 2.5) exposures alter the outcome of adverse cardiac events, especially on the cardiac autonomic function and systemic inflammation. Another goal was to evaluate the efficacy of omega-3 fatty acids as protection against the cardiovascular (CV) effects of PM exposure. It is known that omega-3 polyunsaturated fatty acids have several potential cardioprotective effects including antiarrhythmic, antithrombotic, anti-inflammatory, and lowering lipid levels.

Range and variation of pollutant exposures in the United States and perhaps elsewhere.

The proposal does not contain a review of pollution levels in various locations in the information provided to participants on air pollution and heart and pulmonary disease.

Relevance of the condition chosen for study in the general population (e.g., age, disease, etc.).

This study focused on middle aged individuals 50-75 years of age. The age is appropriate for heart disease, and both males and females were studied. O₂ saturation of >94% was used, which is normal O₂ saturation.

Current air quality standard and relevance for future review of standard N/A

Critical toxicologic pathways and evidence of perturbations

For PM exposures, evidence is presented of adverse cardiopulmonary effects including premature mortality, cardiopulmonary problems including infections, exacerbation of asthma symptoms, and heart attacks. For PM exposures, evidence is presented for decrements of lung function and an influx of neutrophils and other markers of inflammation. Studies are cited that indicate persons with a GSTM1 null genotype may be particularly susceptible to the effects of PM exposure on lung function.

Document citations that support a need for this research (i.e., strategic plan, consensus statements, etc.).

The background documentation cites literature on PM exposure, air pollution, and elevated risk of cardiovascular diseases. Also, epidemiologic studies that show that ASCVD is higher in areas with high air pollution. Similarly there are data cited that omega-3 fatty acids reduce heart disease, arrhythmias, sudden death, and sudden cardiac death. Data support that air pollutants change the autonomic function in the heart, and this may be mediated by oxidative stress. The respiratory effects of PM are well known, but the CV ones are not. The goals of this study were to determine if fish oils/omega-3 fatty acids would reduce or mitigate respiratory and cardiovascular effects of PM.

Remaining information needs after current and forthcoming epidemiologic and animal/in vitro toxicologic data are considered.

This study only investigated responses to PM under very specific and not very realistic conditions. After 4 weeks of fish or olive oil supplementation, subjects were exposed to PM, then end points were neutrophil counts, lung function, inflammation markers, heart rate variability, brachial artery diameter (flow mediated dilation [FMD]), change in blood vasoactivators, and coagulation factors. Secondary end points included pulmonary function tests and endothelial cell function, and peripheral venous blood markers. Exploratory end points included inflammatory cytokines and GSTM1 allele variants. There still remain questions regarding responses that would be expected under variable conditions, and after longer-term exposure.

Additional knowledge and/or level of certainty that this controlled human exposure study would provide.

This study could reinforce that the pulmonary and cardio vascular effects of acute PM exposure can be modified with fatty acids (FAs) or olive oil (OO), if study durations were longer and a dose response was evaluated.

Research goal in the regulatory context of providing public health protection, including identification and protection of sensitive populations.

Good. The study focused on an at-risk, sensitive population for heart disease with the goal of determining if FA or OO supplementation would prevent PM-induced stress on the cardiovascular system.

Research Goal regarding the...

Relationship between physiologic function and pollutant exposure.

The research goals included comparing short-term lung function, and changes in cardiovascular function (heart rate variation and pulmonary function, FMD) with and without fish oil or olive oil supplementation. Also, second goal older subjects with GSTM1 genotype will have a lower risk of CV events than

GSTM1 null genotype when exposed to PM and have worse outcomes than GSTM1 positive genotypes and could benefit more from fish oil.

Biologic plausibility and/or mechanisms of air pollution health effects.

There is biologic plausibility that fish oils can reduce the CV effects and inflammatory pulmonary effects of PM. In addition, the GSTM1-null variant subjects have worse outcomes after PM, and may have a more significant reduction in their response with fish oil or olive oil.

Interpretation of effects observed in toxicologic or epidemiologic studies.

It had been observed in clinical studies that PM affects pulmonary and cardiovascular function and fish oils/olive oil reduce CV events through a number of pathways including antioxidation and anti-inflammatory pathways. This information provided the rationale for these studies.

Does the design of the study adequately reflect the information uncertainty being addressed? The goals of the study involved clarification of short-term effects of exposure regimens, as opposed to possible long-term effects. The study design was appropriate for this limited goal.

Is the health status of the study subjects reflective of the sensitive subgroups that the relevant air pollution standard is intended to protect?

The subjects in this study were selected from among healthy adults (men and women between the ages of 50 and 75 years). Therefore, their responses will be reflective of responses that might occur in the elderly who already have compromised health function that could include cardiovascular or pulmonary diseases.

Is the study method appropriate?

The study method is appropriate for the narrow hypotheses being tested.

Was the choice of the study’s exposure concentration and duration appropriate?

The rationale for the exposure concentration and duration are not clearly stated in the proposal. The study subjects were exposed to clear air for 2 hours on one day and then air with fine and ultrafine PM (PM< 2.5) for 2 hours on the second day in an exposure chamber after the supplementation. There is no clear description of the rationale for the target exposures selected, that of concentrated airborne particulate matter (CAP) (mean 253±16 µg/m³) for 2 hours. Several studies are cited that exposed for epidemiology literature and animal studies; however, this exposure was not derived from literature. Apparently this team has experience with exposing to volunteers at levels around MEAN 253±16 µg/m³ of fine particulate matter. A study by Romieu et al. (2005) studied omega-3 fatty acid to prevent heart disease associated with particulate matter. But this experience is not clearly stated in the application or referred to in support of that exposure level.

Was the choice of study end points appropriate for the experimental question? Yes.

Were the time points for measurement appropriate? Yes, but measurements at additional time points could have been informative.

Was the statistical analysis appropriate?

The data obtained in the study were analyzed using ANOVA parametric test for continuous variables, and rank sum tests for noncontinuous variables to compare the effects of fish oil and olive oil, GSTM1 positive and null genotypes, and pre- and postexposure. This type of test is appropriate for a randomized double-blinded study like that used in this study. The statistical power calculation indicated that 30 subjects (15 per group) would provide 80% power for detecting a 0.13 unit change in brachial artery ultrasound (BAU) diameter. This calculation seems appropriate. However, this power calculation applies only to the BAU diameter and does not provide information on the power from the other end points in the study.

What are the strengths and limitations of the study?

Summarized above. The study design and group size were adequate to study the limited hypotheses of the study. Like most studies of this type it involved a limited number of healthy individuals exposed to a very small range of exposure conditions. Consequently, results of this study should be interpreted cautiously and not necessarily used to predict responses in the population at large.

What are the major findings of the study?

There were two publications (Tong et al., 2012, 2015).

1. Heart rate variability (HRV) was significantly elevated in the OO group immediately after PM exposure whereas the omega-3 groups showed no significant change. The response was quite clear. QT interval as prolonged by OO but not fish oil (FO) but the actual response was quite modest. The most striking finding was that pollution exposure in the OO group markedly elevated very low density lipoprotein (VLDL) and triglycerides whereas FO effectively lowered both. Thus, FO blunted the negative heart rate variability and QT prolongation caused by exposure to particulate pollutants.
2. This study published in 2015 was from the same protocol but with a different outcome. The key finding was that dietary supplementation with OO but not FO clearly blunted the negative impact of particulates of FMD. In addition, the OO group exhibited increased levels of the fibrinolysis marker tPA after particulate exposure. The results from all the other assays were largely inconclusive when comparing the groups.

What are the remaining uncertainties?

Like all studies of limited and prescribed exposures to volunteers, many questions remain. The study exposed healthy volunteers between the ages of 50 and 75 years. Effects upon the very young and old and upon those with existing health conditions were not investigated. The study involved a fixed exposure, none of which are typical of real-world exposures, so using the results of the study to predict responses in real-world situations would be problematic.

Describe how the findings have or may contribute to the following aspects:
Quantification of the relationship between physiologic function and pollutant exposure.

The study provided quantitative information on the cardiovascular and pulmonary and inflammatory responses to fine particulate matter, PM < 2.5µm, in individuals aged 50 to 75 years under specific exposure conditions. It could be problematic to extrapolate these quantitative responses to more general situations.

Understanding biologic plausibility of effects of concern.

The results of this study reinforced the biologic plausibility of effects upon particulate matter mean 253±16 µg/m³ and the cardiovascular, thrombolytic systems observed in toxicologic or epidemiologic studies.

Increased ability to interpret effects observed in toxicologic or epidemiologic studies. Not clear.

Other: Contributions of publications out of the study. None.

OMEGACON Publication Reviews

Purpose of Study

To examine the health effects of fine and ultrafine particulate matter (PM<2.5) exposure on the cardiovascular system and examine whether omega-3 fatty acid supplementation pretreatment would attenuate the adverse cardiovascular effects. A strong positive outcome would provide support for the EPA to advocate a prevention strategy.

Randomized, double-blind study.

The subjects were exposed to clean air on the first day then concentrated ambient fine and ultrafine PM on the second day after the dietary supplementation. Possible health effects of acute 2-hour exposure to air pollution particles include chest pain, mild dyspnea, headache, cough/wheeze, various adverse cardiovascular effects such as increase in heart rate, and decreases in HRV.

Two publications in Environmental Heath Perspectives resulted from this study.

Tong et al. 2012. Omega-3 Fatty Acid Supplementation Appears to Attenuate Particulate Air Pollution–Induced Cardiac Effects and Lipid Changes in Healthy Middle-Aged Adults.

Objective:

Study effects and protective potential of omega-3 fish oil (FO) versus olive oil (OO) control on cardiac function and lipid pretreatment in response to acute air pollution exposure.

Design:

29 patients, 16 FO, 13 OO-3 g/day oil for 28 days. Patients first exposed to filtered air, then next day to 30-fold concentrated ambient air (300 µg/m³ particulate) for 2 hours. Ambulatory EKG measurements for heart rate variability (HRV), QT changes, ventricular polarization. Venous blood measurements for plasma fatty acids, triglycerides, VLDL, HDL.

Results:

The plasma fatty acid composition reflected the diet with the FO patients having substantially elevated EPA (6X), and DPA (2X)-docosa-pentaenoic acid and somewhat lower omega-6 arachidonic acid.

Cardiac results:

HRV was significantly elevated in the OO group immediately after PM exposure whereas the omega-3 group showed no significant change. The response was quite clear. QT interval was prolonged by OO but not FO but the actual response was quite modest. The most striking finding was that pollution exposure in the OO group markedly elevated VLDL and triglycerides, whereas FO effectively lowered both. Thus, FO blunted the negative heart rate variability and QT prolongation caused by exposure to particulate pollutants.

Discussion:

The discussion evoked many literature citations and implications that the current data responses implied autonomic nervous system changes—but there were no actual data to prove that point. In fact the effects of FO on QT were quite modest, and the pollution elevation of triglycerides and VLDL was consistent with the previous literature. The authors’ suggestion that the particulate-exposed cardiac change was a potential mechanism of atherosclerosis is unwarranted. Furthermore, it is not surprising that omega-3 fatty acids reduce the pollution elevated triglycerides. Triglyceride lowering has firmly been shown with FO in many studies. The negative cardiac effects of pollution and the positive triglyceride effect of FO were previously established.

Tong et al. 2015. Dietary Supplementation with Olive Oil or Fish Oil and Vascular Effects of Concentrated Ambient Particulate Matter Exposure in Human Volunteers.

The publication evaluates the potential beneficial vascular effect of 4-week FO or OO dietary supplementation in 42 patients exposed to 2 hours of PM_2.5 particulate pollution.

The primary physiologic measurements were reactive hyperemia measured as flow mediated dilation (FMD), brachial artery diameter, and blood pressure. ELISA assays of venous blood measured numerous coagulation factors, lipid and vascular markers.

Results:

The key finding was that dietary supplementation with OO but not FO clearly blunted the negative impact of particulates of FMD. In addition, the OO group exhibited increased levels of the fibrinolysis marker tPA after particulate exposure. The results from all the other assays were largely inconclusive when comparing the groups.

Discussion:

The elaborate discussion of the literature tried unsuccessfully to implicate endothelial dysfunction. Indeed, two major, well-documented mediators of endothelia-vascular changes, nitric oxide and prostacyclin, were not measured.

Interestingly, there was no mention of the 2012 paper in the discussion. In that report the same authors found that OO supplementation had the negative impact by elevating the heart rate variability and prolonged QT interval caused by acute pollution exposure, i.e., omega-3 was protective. On the other hand, the 2015 paper proposed that OO was protective against the reactive hyperemia produced by pollutants and not FO. The differing effects of FO of various cardiovascular responses to pollution do not warrant, at this point, advocating the protective potential of FO.

In both papers the authors pointed out “conclusions derived from the small numbers of participants included in the study may not be applicable to the population as a whole. Furthermore, the modest sample size and the number of secondary endpoints measured could inflate the significance of the findings.”

Opinion:

The authors are quite correct about the limitations of such a small study. In addition, acute 2-hour exposure is insufficient to project about the anticipated population exposure, which is chronic. These papers were published in specialized journals. Key findings were predictable. The limitations and inconsistencies (especially in the blood marker studies) in the data would suggest publication in a rigorous cardiovascular medical journal seems highly unlikely.

Overall:

The justification of such human experiments considering the obvious limitations in design is unclear. The negative effects of the pollution particulates are well documented. Any safety risk in small, limited trials seems unwarranted. Overall, these studies were hypothesis generating and large trials with greater exposure times would be required to determine clinical utility.

THE INTERACTION OF SOCIAL FACTORS WITH AIR POLLUTION (SOZIAL)

Is the research question well focused? Yes.

Is there a clear hypothesis that is testable? Yes.

Range and variation of pollutant exposures in the United States and perhaps elsewhere.
Usual and experimental levels of ozone are clearly described.

Relevance of the condition chosen for study in the general population (e.g., age, disease, etc.). Eligibility of only younger individuals (18-33 years) reduces the likelihood of adverse reactions, but may not be relevant to older individuals who are an important target population to consider for these air pollution exposures.

Current air quality standard and the relevance for future reviews of the standard.

The current standard is less than 300 ppb ozone and this protocol follows that standard.

Critical toxicologic pathways and evidence of perturbations.

The concept and physiologic role of stress pathways and allostasis are well described.

Document citations that support a need for this research (i.e., strategic plan, consensus statements, etc.).

Appropriate literature cited.

Remaining information needs after current and forthcoming epidemiologic and animal/in vitro toxicologic data are considered.

No comment.

Additional knowledge and/or level of certainty that this controlled human exposure study would provide.

The study should result in important new information that takes into account the characteristics of exposed populations and not just the physiologic or toxicologic responses.

Research goal in the regulatory context of providing public health protection, including the identification and protection of sensitive subpopulations.

Although this study may provide information on persons with different perceived levels of stress, it does not really allow recommendations to be delivered to any particular population subgroup (e.g., disadvantaged populations of low socioeconomic position).

Research goal regarding the...

Relationship between physiologic function and pollutant exposure. Yes.

Biologic plausibility and or mechanisms of air pollution health effects. Yes.

Interpretation of effects observed in toxicologic or epidemiologic studies. Partially.

Does the design of the study adequately reflect the information uncertainty being addressed? Yes, except for some lack of clarity in their randomization procedure. There were two arms described: one with low PPS and the other with high PPS. However, it was not clear whether the randomization to ozone exposure or to clean air (placebo) was carried out within arms or strata of PPS or for all volunteers together. It should have been within strata.

Is the health status of the study subjects reflective of the sensitive subgroups that the relevant air pollution standard is intended to protect?

As above, it includes persons ages 18-33 years and may miss effects likely in older groups with more compromised lung function or ability to deal with stress. However, the likelihood of adverse reactions in an older group is likely increased.

Is the study method appropriate? Yes.

Was the choice of the study’s exposure concentration and duration appropriate? Yes.

Was the choice of study end points appropriate for the experimental question? Yes.

Were the time points for measurement appropriate? Yes.

Was the statistical analysis appropriate?

The planned analysis seems appropriate, but there were no actual analyses to review.

What are the strengths and limitations of the study?

As stated above the restriction to younger ages limits the generalizability to the full population age range.

EFFECTS OF WOOD SMOKE PARTICLES ON INFLUENZA-INDUCED NASAL INFLAMMATION IN NORMAL VOLUNTEERS (WOODSIE)

Is the research question well focused?

The research question is well focused.

Is there a clear hypothesis that is testable?

There is a clear and testable hypothesis: exposure to wood smoke particles enhances influenza virus–induced granulocyte and NK cell activation via hyaluronic acid–mediated effects on IFNg production.

Range and variation of pollutant exposures in the United States and perhaps elsewhere.

Wood smoke is an increasingly important source of ambient particulate matter in the United States. Outdoor sources include landscape fires (67,774 wildfires involving 9,326,238 acres in the United States in 2012) while indoor sources include the use of wood for indoor heating or recreation/ambience, structural fires, and intrusion of outdoor source wood smoke.

Wood smoke is also an international health concern with 2 billion people using biomass (including wood) for indoor heating and cooking. In some areas, indoor-generated smoke is an important contributor to outdoor particulate air pollution.

Relevance of the condition chosen for study in the general population (e.g. age, disease, etc.).

The condition chosen for study, influenza virus infection, is highly relevant because influenza infections are an important cause of morbidity and mortality in the United States and worldwide. The finding that wood smoke alters influenza infection would have broad public health implications. The physiologic and biomarker changes shown in this study could potentially be used for population-based studies.

Current air quality standard and the relevance for future reviews of the standard.

The current primary air quality standard for PM_2.5 is 12 µg/m³ as an annual mean, averaged over 3 years and a 24-hour standard of 35 µg/m³ (98th percentile averaged over 3 years). The current primary air quality standard for PM₁₀ is 150 µg/m³ which is a 24-hour averaging time not to be exceeded more than once per year on average over 3 years.

The concentration used in this study is 500 µg/m³ for 2 hours’ duration, and is justified based on the measurement of similar levels of particulate matter sometimes seen in homes using wood as their main energy source.

Critical toxicologic pathways and evidence of perturbations.

A strength of the study is the focus on critical pathways for viral host defense and seeking evidence of perturbations with wood smoke exposure.

Document citations that support a need for this research (i.e. strategic plan, consensus statements, etc.).

Remaining information needs after current and forthcoming epidemiologic and animal/in vitro toxicologic data are considered.

The IRB submission does not review epidemiologic and animal/in vitro toxicologic data (this is not normally required).

Additional knowledge and/or level of certainty that this controlled human exposure study would provide.

Pathways of toxicity and biologic plausibility.

Research goal in the regulatory context of providing public health protection, including the identification and protection of sensitive subpopulations.

Wood smoke exposure studies might help identify sensitive subgroups.

Does the design of the study adequately reflect the information uncertainty being addressed?

The study design adequately reflects the information uncertainty being addressed. There is a clean air comparator allowing each subject to serve as his/her own control. A strength of the design is the inclusion of repeated time points, and sampling of nasal lavage fluids, nasal biopsy material, and blood.

Is the health status of the study subjects reflective of the sensitive subgroups that the relevant air pollution standard is intended to project?

The study is approved for 40 subjects, with 14 enrolled prior to the renewal request. Healthy nonsmoking adults age 18-40 are the study population. The air pollution standard applies to this group of individuals. The standard also extends to sensitive subgroups including children, the elderly, and people with respiratory diseases. It is possible that differences in immune function in these groups would result in differences in host response.

Is the study method appropriate?

The study methods are appropriate, with use of standardized dosing with live attenuated influenza virus. The range of inflammatory end points includes mediators and cell-based assays.

Was the choice of the study’s exposure concentration and duration appropriate?

The exposure protocol was based on the study of Ghio et al. in which healthy adult volunteers were exposed to 500 µg/m³ wood smoke particles (WSPs) for 2 hours with exercise. The investigator indicates the exposure is lower than that for forest firefighters, those living in areas near forest fires or agricultural burning, or people living in developing countries using biomass for cooking.

Was the choice of the study end points appropriate for the experimental question?

The choice of end points is appropriate for the pathways under investigation. A strength of the study is the partnering with other laboratories, leveraging the study results to address additional end points (granulocyte activation and lipid mediator activation) that are highly relevant to the toxicologic pathways involved in this response.

Were the time points for measurement appropriate?

Time points are appropriate (0, 1, 2, 7-10, and 21-28 days), covering the anticipated duration of infection and assessing acute and subacute time points.

Was the statistical analysis appropriate?

A sample size justification is provided and is acceptable. It is based on the area under the curve for nasal lavage fluid gamma interferon based on their previously published study of diesel exposure. Absent previous wood smoke exposure data, this seems a reasonable surrogate.

What are the strengths and limitations of the study?

Strengths of the study are the public health importance of the research question, the experimental design with a robust array of end points and time points.

Describe how the findings have or may contribute to the following aspects:
Quantification of the relationship between physiologic function and pollutant exposure.

Physiologic function (and biomarkers) are evaluated.

Understanding biologic plausibility of effects of concern.

The investigators have examined these pathways looking at other toxicants (e.g., tobacco smoke).

Increased ability to interpret effects observed in toxicologic or epidemiologic studies.

Will provide biologic plausibility as well as indicate end points for use in epidemiologic studies. Will also assist in interpretation of findings from animal/toxicologic studies.

PHYSIOLOGICAL CHANGES IN ADULTS WITH METABOLIC SYNDROME EXPOSED TO CONCENTRATED ULTRAFINE CHAPEL HILL AIR PARTICLES (XCON)

Is the research question well focused?

The investigators state that the research question is to examine the acute health effects of concentrated ambient ultrafine (UF) particulate matter (PM) exposure in patients with metabolic syndrome. The outcomes they are studying are not, however, clinical health effects, but rather biologic and physiologic responses that may, if associated with the exposure, lend biologic plausibility to observational studies finding associations of UF with clinical outcomes.

This particular subgroup (a subgroup with metabolic syndrome) is at risk of developing cardiovascular disease (CVD) or type 2 diabetes mellitus (DM), independent of PM exposure. Clinical CVD and DM have been shown to increase susceptibility to clinical health effects of ambient particle pollution. Some (but not all) studies suggest that metabolic syndrome and its components may increase susceptibility to

inflammatory or physiologic effects of pollution. The purpose of this study, as stated in the IRB application,² is

The IRB application for XCON also states, “The results from this human study will provide information data that assist the EPA in determining whether or not to retain the current standard on PM” (p. 15).

EPA’s 2011 policy assessment for the review of the PM NAAQS indicates that the agency considered “whether there is support to consider standards with a different size fraction and/or distinct standards focused on regulating a specific PM_2.5 component or group of components associated with any source categories of fine particles” (EPA 2011, pp. 1-16). However, EPA staff concluded that there is insufficient information at this time to consider a separate indicator for ultrafine particles.

Also, the potential role of ultrafine particles in causing adverse health effects, relative to PM_2.5, may inform the development of mitigation approaches. EPA’s Clean Air Research Multi-Year Plan 2008-2012 states (EPA 2008):

Is there a clear hypothesis that is testable?

The hypothesis is that UF particular matter (PM) exposure in this population will result in changes in endothelial response as assessed by flow-mediated dilation of the brachial artery and various electrophysiologic outcomes (e.g., heart rate variability) and blood end points.

The study used a design of two exposure treatments x two sequences x two periods. Each exposure was separated by at least 2 weeks. One consideration regarding the testability of the hypothesis is whether the wait period between exposure treatments was sufficient to ensure washout of the previous exposure, and yet short enough to minimize the potential effect of confounders. To avoid the potential carryover effect, a panel for the Food and Drug Administration (FDA) has recommended that the 2 x 2 crossover design not be used in drug evaluation (Kuehl, 1994). Although it is not considered to be a critical shortcoming for this study, it is important that the choice of the washout period not be routine, but instead be based on a consideration of the specific aspects of a particular experiment.

Range and variation of pollutant exposures in the United States and perhaps elsewhere.

The proposal does not contain a review of UF levels in various locations. It mentions a number of times that the 600,000 particles/cc is similar to what an individual might experience driving in a heavy traffic highway in Los Angeles, but no other range or variation information is discussed.

__________________

² Application for IRB approval of human subjects research for XCON. EPA, unpublished material, November 19, 2014.

Relevance of the condition chosen for study in the general population (e.g., age, disease, etc.).

Appropriate. This study is focused on evaluating 34 subjects with metabolic syndrome between the ages of 25 and 70 years. This age range is a good reflection of the population with metabolic syndrome; however, it usually has a higher prevalence after the age of 50 years (44% in this study). Also it was reported that about 24% of the American adult population meets the definition of metabolic syndrome, which supports selected target subgroups (non-negligible size of susceptible subgroups).

Current air quality standard and the relevance for future reviews of the standard.

Air quality standards related to UF PM exposure and lung function are not discussed in the proposal. Also, there is no description of how the results of this study could affect air quality standards. The current standards are based on PM_2.5 or PM₁₀. Since the EPA is asking whether there should be a standard based on UF, this study can provide an answer to the necessity of future EPA standard for UF PM but not to the standard. The EPA should clarify whether it is conducting a study to clarify where efforts for regulation should focus, in terms of regional sources, even if the regulation may not directly relate to standard setting.

Critical toxicologic pathways and evidence of perturbations.

The background/rationale for the toxicologic pathways and evidence of perturbations for UF PM at current or below EPA air quality standards and cardiovascular disease, both morbidity and mortality are provided. There is a good description of why the particles cause organ toxicity and the leading hypothesis that very small particle size can move deep into the lung tissue with inspiration, and cause greater tissue damage or move more easily into the circulation. Identification of subsets in our U.S. population that are at higher risk for toxicity from PM (the elderly, individuals with cardiopulmonary diseases, and people with diabetes) is discussed, as well as the biologic underpinnings of potential susceptibility of people with metabolic syndrome to pollution.

Document citations that support a need for this research (i.e., strategic plan, consensus statements, etc.).

Good. The document supports a need for this research by demonstrating it is part of the EPA strategic plan, and by statistics about the increasing prevalence of metabolic syndrome in the United States, how this syndrome is associated with increased CV disease, and endothelial dysfunction manifesting in a proinflammatory and prothrombotic state due to increased production of inflammatory cytokines and C reactive protein. Since endothelial dysfunction and inflammation are toxicities of PM of air pollution, the rationale to support this research is acceptable. Previously, the investigators have evaluated CAP effects on endothelial function in subjects with diabetes. Evaluating the UF PM effects on endothelial response and blood end points in individuals with metabolic syndrome is supported adequately by citations.

Remaining information needs after current and forthcoming epidemiologic and animal/in vitro toxicologic data are considered.

The XCON IRB application³ indicated that:

__________________

³ Application for IRB approval of human subjects research for XCON. EPA, unpublished material, November 19, 2014.

Additional knowledge and/or level of certainty that this controlled human exposure study would provide.

This is not clear, as the purpose of this study is to evaluate the acute health effects (biologic and physiologic) of UF PM exposure in patients with metabolic syndrome (MS). This study has a specific focus, as it is only planning to study 30 subjects with MS, and the end point changes in endothelial response assessed by FMD of BA and HRV and blood end points. This study is part of the charter of the EPA to study susceptible populations and determine health risks with ambient levels UF PM.

If the investigators find associations of UF PM with the subclinical outcomes they are measuring, this may lend biologic plausibility to observational studies linking UF with clinical health outcomes, and may inform regulation if it is feasible for the EPA to regulate UF, which fall off rapidly as distance from traffic increases. If they do not find associations, this may be because (1) the population is too small; (2) the UF specific to Chapel Hill are not toxic (limited generalizability); or (3) UF does not cause the effects that they are measuring.

The conclusions reported in the publication are as follows:

• Exposure to concentrated ambient ultrafine particles (UCAPS) does not cause changes in brachial artery diameter or blood pressure.
• Exposure to UCAPS causes changes in cardiac repolarization and heart rate variability.
• Exposure to UCAPS causes changes in vascular markers of inflammation and fibrinolysis.

In comparison to PM_2.5, whether or not the findings above are related to the small size or number of UF is not fully discussed.

Research goal in the regulatory context of providing public health protection, including the identification and protection of sensitive subpopulations.

As mentioned above, results of this study will inform considerations of whether a UF-specific standard is warranted. This study is directed to protection of a sensitive subpopulations—individuals with metabolic syndrome.

Research goal regarding the...

Relationship between physiologic function and pollutant exposure.

The toxicologic data information on how UF PM may influence biologic and physiologic outcomes and the endothelium is described. Since individuals with MS have high risk of CVD, the need to determine how UF exposure affects this subpopulation is important.

Biologic plausibility and or mechanisms of air pollution health effects.

Previous data from earlier studies are discussed to demonstrate the toxicity in the lung and endothelial cells from UF PM, and the association with CVD. Also, the discussion of subpopulations that are at increased risk for CVD from air pollution is given.

Interpretation of effects observed in toxicologic or epidemiologic studies.

The background epidemiology and genetic epidemiology to support performing this study are adequate.

Other: Statistical analysis plan in detail at proposal stage.

The substantial involvement of a statistician at the proposal stage is critical in ensuring the appropriate design of experimental questions and statistical questions according to the research questions or goals.

Specifically at the proposal stage, a simulation and a pilot test are very useful for planning and setting experimental design. A few scenarios are usually possible based on previous studies or known facts.

Study Design

Does the design of the study adequately reflect the information uncertainty being addressed?

This is a double-blinded study in which each participant will be exposed to filtered clean air and air containing concentrated UF particles in randomized order. Through both blinding and randomization this study could avoid bias in testing.

The study has a two-treatment, two-sequence, and two-period crossover design, where the treatment effects are intended not to be confounded with the effects of sequences or periods. It was not reported whether baseline observations were taken prior to any treatment to assess potential carryover effects.

The study design also included repeated measurements over a 24-hour time span. This enables the investigation if the UF PM effect is immediate, delayed, or remained within a day. Thus, the proposed study design was reasonable but further investigation on time-related effect of UF was not done.

Is the health status of the study subjects reflective of the sensitive subgroups that the relevant air pollution standard is intended to protect?

Metabolic syndrome and its components are risk factors for CVD and type-2 DM. Those study subject characteristics are considered to be reflective, albeit indirectly, of sensitive subgroups. If there were an interest in directly studying subjects known to be at higher risk of clinical cardiovascular outcomes with acute air pollution exposures to these particles (e.g., those with overt diabetes or clinical cardiovascular disease), other study designs, such as community-based, repeated-measures, observational studies, would be more appropriate.

The subjects in this study were selected for metabolic syndrome in men and women between the ages of 25 and 70 years. A potential limitation of the study may be that their responses will possibly not be reflective of responses that might occur in the elderly who might have compromised health function.

The subject selection process was well established overall in the proposal through physical screening and informed-consent form. However, Ghio et al. presented a case of a 58-year-old woman who showed an increased risk for supraventricular arrhythmia and was later hospitalized overnight for observation. This event illustrates that physical screening and the informed-consent form cannot ensure there is zero risk of an adverse event occurring during a controlled exposure study.

Is the study method appropriate?

The study method is appropriate for testing the narrow hypotheses being tested:

• - the order of the clean air and UFPM exposures was randomized,
• - a double-blinded randomization was used to avoid bias,
• - the health end points of interest were measured three times at the proper time, and
• - measurement time was consistent to avoid variations within a day.

However, the maximum of the UF PM concentration was not controlled precisely, and a reasonable consideration of dropout of subjects was missing.

Was the choice of the study’s exposure concentration and duration appropriate?

The rationale for the exposure concentration and duration are clearly stated in the proposal. The study subjects will be exposed to ultrafine particles concentrated from Chapel Hill air, typical concentrations range from 4,000 to 12,000 particles/cc, an instrument will concentrate the particles so that it is anticipat-

ed that the study subjects will be exposed to 11,000-330,000 UFP/cc on average and will establish a maximum of 600,000 particles/cc which is less than one would be exposed to driving a heavily traveled highway in Los Angeles.

This study was arranged as a 2-hour exposure duration, which seemed to follow previous study designs. The rationale for the 2-hour duration was not provided. Also, the subjects were sitting at rest with no activities during the 2-hour exposure time in this study. There could be differences in real exposure level with/without activity.

Was the choice of study end points appropriate for the experimental question?

Appropriate. The study outcomes are flow-mediated dilation (brachial artery ultrasound) and heart rate variability, peripheral venous blood samples, specific and nonspecific immune responses (cytokines and C-reactive protein), coagulation factors (von Willibrand factor, factor IX, fibrinogen and thrombin), vasoactive factors (endothelina, catecholamine), and soluble components of PM (transition metals).

Were the time points for measurement appropriate?

Appropriate. Three measurements, before exposure and 1 and 20 hours after the exposure. This would likely be sufficient to indicate the trend in rebound cardiovascular, inflammatory, and endothelial end points.

Was the statistical analysis appropriate?

Appropriate. The primary outcome will compare preexposure flow-mediated dilation to mean flow-mediated dilation 1 and 20 hours after the exposure. A second analysis will compare mean preexposure heart rate variability measures to the mean measures taken 1 and 20 hours after exposure. Both analyses will use an F-test to control for type I error, with flow-mediated dilation and HRV as separate dependent variables and exposure and time from exposure as independent variables. Although there are only 30 subjects to be recruited, they plan to use a repeated-measures ANOVA, and paired T-tests will be applied to the blood end points and Holter monitor analysis. With such a small sample size, nonparametric statistics might be more appropriate.

The sample size estimate for the research question is very low, 30; however, the investigators state that they expect to have 80% power to detect a difference of 3% in flow-mediated dilation assuming a SD of 4%. An N of 16 would give 80% power, p <0.05; however, they are recruiting 30 subjects to allow for subject variability.

The analysis in the publication did not justify the assumption of a normal distribution for the ratio of two response measurements. Let Y1, Y2, and Y3 represent the response measured at time points 1 (baseline), 2, and 3, respectively. In the publication both ratios of Y2/Y1 and Y3/Y1 were assumed to have normal distributions. In general, the ratio of two responses, Y2/Y1, is not normal even if both Y1 and Y2 are normally distributed. As described in the proposal appendix, the study analysis should start with postulating the distribution of the responses or end points. There are a few methods available for handling a ratio of two correlated normal variables. As an alternative, a lognormal distribution can be assumed for the health end points, if applicable.

The publication also used a genetic factor from which the whole population (N=19) was compared to its subset, GSTM1 null population (N=9). For this small sample size of 9, nonparametric test methods are recommended.

As indicated in the publication, the study examined multiple responses from individual study subjects. Due to small sample size, the correlations among the multiple responses were not considered in the analysis.

What are the strengths and limitations of the study?

The strengths of this study can be regarded as threefold:

1. This study can generate a profile of a susceptible and/or sensitive subpopulation (people with metabolic syndrome), which can be used for future studies;
2. This study has information on the UF PM distribution in terms of number of particles; and
3. This study may separate UF PM effect from PM_2.5 effect.

Like most studies of this type, it involved a limited number of subjects. There was a lack of a rigorous statistical analysis of the uncertainties in estimates due to relatively small sample size (N=34 in total) and very small sample size for GSTM1 null population (N=9). However, if this subpopulation has a significant response to this exposure based on the outcome variables, additional studies will be warranted, and in addition, the EPA will have fulfilled its mission to discover subpopulations at risk for other diseases from ultrafine air pollutants.

What are the major findings of the study?

They reported in their publication as follows:

• Exposure to UF does not cause changes in brachial artery diameter or blood pressure.
• Exposure to UF causes changes in cardiac repolarization and heart rate variability.
• Exposure to UF causes changes in vascular markers of inflammation and fibrinolysis.
• The results suggest that UF may affect some biologic pathways by oxidative processes in which GSTM1 (an enzyme) may play a role.

The lack of BAD or BP responses may reflect the variability of the measure and lack of power, the specific content of these UFs (generalizability), subject susceptibility, or differences in the biologic effects of UF, compared to PM_2.5 in Chapel Hill.

What are the remaining uncertainties?

Like all studies of limited and prescribed exposures to volunteers, many questions remain. The study volunteers were between the ages of 25 and 70 years. Effects upon the very young and old were not investigated. Given that metabolic syndrome is more prevalent in individuals over the age of 50 years (44%), it might have been worthwhile to increase the enrollment of that age group. On the other hand, people younger than 70 often have metabolic syndrome without established CVD, whereas those older than 70 may be more likely to have overt clinical disease, which may be a contraindication for enrolling them.

Three results remain uncertain: (a) the results from GSTM1 null (N=9) due to small sample size, (b) UF effect with particle number or particle mass needs further investigation, and (c) UF effect and PM_2.5 effect are not comparable due to the use of different units of measurement (that is, particle number and particle mass).

Describe how the findings have or may contribute to the following aspects:
Quantification of the relationship between physiologic function and pollutant exposure.

The main measure for UF-induced change in responses was percent change in relative ratio to the individual baselines. This study quantified the changes in end points after exposure to UF, not the relationship between UF and end points.

Understanding biologic plausibility of effects of concern.

The results of this study, when completed, may determine if individuals with metabolic syndrome when exposed to UF PM have changes in the CV function, inflammation, and endothelial function and thrombosis that may over time increase their risk of development of cardiovascular disease.

Increased ability to interpret effects observed in toxicologic or epidemiologic studies.

The results from this study could possibly provide reinforcement to interpretations of effects observed in other studies and define a new at-risk subpopulation, individuals with metabolic syndrome.

Other: Contributions of publications out of the study.

There are two publications related to this study, Devlin et al. (2014) and Ghio et al. (2012). Devlin reported significant changes in HRV caused by UF PM with no significant changes in FMD and blood end points.

Ghio presented a case of a 58-year-old woman, who volunteered to participate in a controlled exposure to concentrated ambient particles and showed an increased risk for supraventricular arrhythmia. The authors claimed it the first case of cardiovascular disease after exposure to elevated concentrations of any air pollutant, which was not supported by Langrish et al. (2014). Langrish reported no changes from his controlled exposure experiment and indicated the atrial fibrillation (AF) as the most common cardiac arrhythmia in the general population and associated with increasing age, hypertension, and cardiac dysfunction. The authors suggest that in the case Ghio reported it is more likely that the investigators simply witnessed an asymptomatic episode of AF in a patient at increased arrhythmic risk due to coexistent hypertension, age, and frequent atrial ectopy, and the occurrence of AF in the exposure chamber is likely to have been coincidence and simply due to chance.

According to the Office of Inspector General report: “NHEERL management met and determined that no screening could have feasibly been done to have predicted this issue. The XCON consent form already warned study participants not to participate if they had cardiovascular disease including coronary artery disease, heart failure, or rhythm disturbances” (EPA 2014, page 27).

One other thing to note: antioxidant genes that are highly prevalent have been shown in a number of studies to increase risk of a variety of subclinical outcomes associated with pollution exposures. The 2014 published report considers GSTM1 null as a source of susceptibility to UF.

REFERENCES

Devlin, R.B., C.B. Smith, M.T. Schmitt, A.G. Rappold, A. Hinderliter, D. Graff, and M.S. Carraway. 2014. Controlled exposure of humans with metabolic syndrome to concentrated ultrafine ambient particulate matter causes cardiovascular effects. Toxicol. Sci. 140(1):61-72.

EPA (U.S. Environmental Protection Agency). 2008. Clean Air Research Multi-Year Plan 2008-2012. EPA 620/R-08/001.Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC. June 2008. https://permanent.access.gpo.gov/lps106873/Air-MYP-narrative-final.pdf [accessed March 3, 2017].

EPA. 2011. Policy Assessment for the Review of the Particulate Matter National Ambient Air Quality Standards. EPA/452/R-11-003. Office of Air Quality Planning and Standards, EPA, Research Triangle Park, NC. April 2011 [online]. Available: https://www3.epa.gov/ttn/naaqs/standards/pm/data/20110419pmpafinal.pdf [accessed March 3, 2017].

EPA. 2014. Improvements to EPA Policies and Guidance Could Enhance Protection of Human Study Subjects. Report No. 14-P-0154. Office of Inspector General. March 31, 2014 [online]. Available: https://www.epa.gov/sites/production/files/2015-09/documents/20140331-14-p-0154.pdf [accessed May 24, 2016].

Ghio, A.J., M. Bassett, T. Montilla, E.H. Chung, C.B. Smith, W.E. Cascio, and M.S. Carraway. 2012. Case report: Sypraventricular arrhythmia after exposure to concentrated ambient air pollution particles. Environ. Health Perspect. 120(2):275-277.

Hesterberg, T.W., C.M. Long, S.N. Sax, C.A. Lapin, R.O. McClellan, W.B. Bunn, and P.A. Valberg. 2011. Particulate matter in new technology diesel exhaust (NTDE) is quantitatively and qualitatively very different from that found in traditional diesel exhaust (TDE). J. Air Waste Manage. Assoc. 61(9):894-913.

Khalek, I.A., T.L. Bougher, P.M. Merritt, and b. Zielinska. 2011. Regulated and unregulated emissions from highway heavy-duty diesel engines complying with U.S. Environmental Protection Agency 2007 emissions standards. J. Air Waste Manag. Assoc. 61(4):427-442.

Khalek, I.A., M.G. Blanks, P.M. Merritt, and B. Zielinska. 2015. Regulated and unregulated emissions from modern 2010 emissions-compliant heavy-duty on-highway diesel engines. J. Air Waste Manag. Assoc. 65(8):987-1001.

Kuehl, R.O. 1994. Statistical Principles of Research Design and Analysis. Belmont, CA: Duxbury Press.

Langrish, J.P., S.J. Watts, A.J. Hunter, A.S.V. Shah, J.A. Bosson, J. Unosson, S. Barath, M. Lundbäck, F.R. Cassee, K. Donaldson, T. Sandström, A. Blomberg, D.E. Newby, and N.L. Mills. 2014. Controlled exposures to air pollutants and risk of cardiac arrhythmia. Environ. Health Perspect. 122(7):747-753.

Madden, M.C., T. Stevens, M. Case, M. Schmitt, D. Diaz-Sanchez, M. Bassett, T.S. Montilla, J. Berntsen, and R.B. Devlin. 2014. Diesel exhaust modulates ozone-induced lung function decrements in healthy human volunteers. Part. Fibre Toxicol. 11:37.

McClellan, R.O., T.W. Hesterberg and J.C. Wall. 2012. Evaluation of carcinogenic hazard of diesel engine exhaust needs to consider revolutionary changes in diesel technology. Regul. Toxicol. Pharmacol. 63(2): 225-258.

Romieu, I., M.M. Tellez-Rojo, M. Lazo, A. Manzano-Patino, M. Cortez-Lugo, P. Julien, M.C. Bélanger, M. Hernandez-Avila, and F. Holguin. 2005. Omega-3 fatty acid prevents heart rate variability reductions associated with particulate matter. Am. J. Respir. Crit. Care Med. 172(12):1534-1540.

Stiegel, M.A., J.D. Pleil, J.R. Sobus, M.K. Morgan, and M.C. Madden. 2015. Analysis of inflammatory cytokines in human blood, breath condensate, and urine using a multiplex immunoassay platform. Biomarkers 20(1):35-46.

Stiegel, M.A., J.D. Pleil, J.R. Sobu, and M.C. Madden. 2016. Inflammatory cytokines and white blood cell counts response to environmental levels of diesel exhaust and ozone Inhalation exposures. PLoS One 11(4):e0152458.

Tong, H., A.G. Rappold, D. Diaz-Sanchez, S.E. Steck, J. Berntsen, W.E. Cascio, R.B. Devlin, and J.M. Samet. 2012. Omega-3 fatty acid supplementation appears to attenuate particulate air pollution–induced cardiac effects and lipid changes in healthy middle-aged adults. Environ. Health Perspect. 120(7):952-957.

Tong, H., A.G. Rappold, N. Caughney, A.L. Hinderliter, M. Bessett, T. Montilla, M.W. Case, J. Berntsen, P.A. Bromberg, W.E. Cascio, D. Diaz-Sanchez, R.B. Devlin, and J.M. Samet. 2015. Dietary supplementation with olive oil or fish oil and vascular effects of concentrated ambient particular matter exposure in human volunteers. Environ. Health Perspect. 123(11):1173-1179.

WHO (World Health Organization). 2015. Residential Heating with Wood and Coal: Health Impacts and Policy Options in Europe and North America [online]. Available: http://apps.who.int/iris/bitstream/10665/153671/1/ResidentialHeatWoodCoal.pdf [accessed March 3, 2017].

This page intentionally left blank.