Evaluating Potential Response-Modifying Factors for Associations between Ozone and Health Outcomes: A Weight-of-Evidence Approach

Background: Epidemiologic and experimental studies have reported a variety of health effects in response to ozone (O3) exposure, and some have indicated that certain populations may be at increased or decreased risk of O3-related health effects. Objectives: We sought to identify potential response-modifying factors to determine whether specific groups of the population or life stages are at increased or decreased risk of O3-related health effects using a weight-of-evidence approach. Methods: Epidemiologic, experimental, and exposure science studies of potential factors that may modify the relationship between O3 and health effects were identified in U.S. Environmental Protection Agency’s 2013 Integrated Science Assessment for Ozone and Related Photochemical Oxidants. Scientific evidence from studies that examined factors that may influence risk were integrated across disciplines to evaluate consistency, coherence, and biological plausibility of effects. The factors identified were then classified using a weight-of-evidence approach to conclude whether a specific factor modified the response of a population or life stage, resulting in an increased or decreased risk of O3-related health effects. Discussion: We found “adequate” evidence that populations with certain genotypes, preexisting asthma, or reduced intake of certain nutrients, as well as different life stages or outdoor workers, are at increased risk of O3-related health effects. In addition, we identified other factors (i.e., sex, socioeconomic status, and obesity) for which there was “suggestive” evidence that they may increase the risk of O3-related health effects. Conclusions: Using a weight-of-evidence approach, we identified a diverse group of factors that should be considered when characterizing the overall risk of health effects associated with exposures to ambient O3. Citation: Vinikoor-Imler LC, Owens EO, Nichols JL, Ross M, Brown JS, Sacks JD. 2014. Evaluating potential response-modifying factors for associations between ozone and health outcomes: a weight-of-evidence approach. Environ Health Perspect 122:1166–1176; http://dx.doi.org/10.1289/ehp.1307541


Literature search strategy
The References identified through the multipronged search strategy were screened by title and abstract by scientists at the EPA. Non-English language papers were excluded. Those references that were potentially relevant after reading the title were "considered" for inclusion in the ISA and were added to the Health and Environmental Research Online (HERO) database developed by EPA; which is available to the public.
Only those studies that have undergone scientific peer review and have been published or accepted for publication and published reports that have undergone peer review were considered for inclusion. All relevant epidemiologic, controlled human exposure, toxicological, ecological, and welfare effects studies published since the last O 3 review were considered, including those related to exposure-response relationships, mode(s) of action (MOA), and response modifying factors that may increase or decrease the risk of an O 3 -related health effect in specific populations and lifestages. Studies and data analyses on atmospheric chemistry, air quality and emissions, environmental fate and transport, dosimetry, toxicokinetics and exposure were also considered for inclusion. This large global search identified approximately 22,000 studies that examined both health and ecological effects and O 3 exposure.

Study selection and evaluation of individual study quality
After the literature search was conducted the selection of studies considered for inclusion was based on the extent to which the study is informative and policy-relevant. This evaluation was performed by scientists at the EPA for studies of health, ecological, and welfare effects; however for this paper we focus on the identification and evaluation of health effects studies.
In general, in assessing the scientific quality of health effects studies, the following considerations were taken into account. § Were study design, study groups, methods, data, and results clearly presented to allow for study evaluation? § Were the study site(s), study populations, subjects, or organism models adequately selected, and are they sufficiently well -defined to allow for meaningful comparisons between study or exposure groups? § Are the air quality data, exposure, or dose metrics of adequate quality and sufficiently representative of information regarding ambient conditions? § Are the health effect measurements meaningful, valid and reliable? § Were likely covariates or modifying factors adequately controlled or taken into account in the study design and statistical analysis? § Do the analytical methods provide adequate sensitivity and precision to support conclusions? § Were the statistical analyses appropriate, properly performed, and properly interpreted?
These criteria provide benchmarks for evaluating various studies and for focusing on the policyrelevant studies in assessing the body of health effects evidence. Of most relevance for inclusion are studies that provide useful qualitative or quantitative information on O 3 exposure-effect or exposure-response relationships at doses or concentrations relevant to ambient conditions that can inform decisions on whether to retain or revise the standards. Therefore, concentrations above 2 ppm were excluded from the review.
The results from the large global search were reduced using exclusion criteria (e.g. non-English language and not related to ambient air, such as disinfection byproducts) and targeted searches for key health endpoints to 4,057 references that were considered for inclusion in the O 3 ISA. A total of 2,270 references deemed by EPA scientists to be of high quality, based on the above considerations, was included in the final document.

Evaluation of scientific evidence and the causal framew ork
To aid judgment in interpreting scientific results, various "aspects" of causality have been discussed by many philosophers and scientists. The "aspects" to judging causality developed by Sir Austin Bradford Hill (Hill 1965) formed the basis for EPA's causal determination framework, but was modified to encompass a broader array of data (Table S1), i.e., epidemiologic, controlled human exposure, ecological, and animal toxicological studies, as well as in vitro data, and to be more consistent with EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA 2009;U.S. EPA 2005). Additionally this framework was developed to be specific to examining causality for health and welfare effects for pollutant exposures. Although these aspects provide a framework for assessing the evidence, they do not lend themselves to being considered in terms of simple formulas or fixed rules of the evidence necessary to lead to conclusions about causality (Hill 1965). Rather, these aspects provide a framework for systematic appraisal of the body of evidence, informed by peer and public comment and advice, which includes weighing alternative views on controversial issues. In addition, it is important to note that the aspects presented in Table S1 cannot be used as a strict checklist, but rather to determine the weight of the evidence for inferring causality. In particular, not meeting one or more of the principles does not automatically preclude a determination of causality [see discussion in (CDC 2004)]. Building off these aspects used to judge causality the US EPA developed a causal framework to draw conclusions regarding the causal relationship between relevant pollutant exposures and health or environmental effects as discussed in the O 3 ISA.
This weight of evidence approach is detailed in Table S2. It is with these aspects in judging causality in mind that we modified the causality framework detailed in Table S2 to encompass examining response modifying factors, which is used to draw conclusions regarding whether a specific factor increases or decreases the risk of an air pollutant (i.e. O 3 )-related health effect.

Consistency of the observed association
An inference of causality is strengthened when a pattern of elevated risks is observed across several independent studies. The reproducibility of findings constitutes one of the strongest arguments for causality. If there are discordant results among investigations, possible reasons such as differences in exposure, confounding factors, and the power of the study are considered.

Coherence
An inference of causality from one line of evidence (e.g., epidemiologic, clinical, or animal studies) may be strengthened by other lines of evidence that support a cause-and-effect interpretation of the association. Evidence on ecological or welfare effects may be drawn from a variety of experimental approaches (e.g., greenhouse, laboratory, and field) and subdisciplines of ecology (e.g., community ecology, biogeochemistry, and paleontological/historical reconstructions). The coherence of evidence from various fields greatly adds to the strength of an inference of causality. In addition, there may be coherence in demonstrating effects across multiple study designs or related health endpoints within one scientific line of evidence.
An inference of causality tends to be strengthened by consistency with data from experimental studies or other sources demonstrating plausible biological mechanisms. A proposed mechanistic linking between an effect and exposure to the agent is an important source of support for causality, especially when data establishing the existence and functioning of those mechanistic links are available.
Biological gradient (exposure-response relationship) A well-characterized exposure-response relationship (e.g., increasing effects associated with greater exposure) strongly suggests cause and effect, especially when such relationships are also observed for duration of exposure (e.g., increasing effects observed following longer exposure times).

Strength of the observed association
The finding of large, precise risks increases confidence that the association is not likely due to chance, bias, or other factors. However, it is noted that a small magnitude in an effect estimate may represent a substantial effect in a population.

Experimental evidence
Strong evidence for causality can be provided through "natural experiments" when a change in exposure is found to result in a change in occurrence or frequency of health or welfare effects.

Temporal relationship of the observed association
Evidence of a temporal sequence between the introduction of an agent, and appearance of the effect, constitutes another argument in favor of causality.

Specificity of the observed association
Evidence linking a specific outcome to an exposure can provide a strong argument for causation. However, it must be recognized that rarely, if ever, does exposure to a pollutant invariably predict the occurrence of an outcome, and that a given outcome may have multiple causes.

Analogy
Structure activity relationships and information on the agent's structural analogs can provide insight into whether an association is causal. Similarly, information on mode of action for a chemical, as one of many structural analogs, can inform decisions regarding likely causality.

Health effects Causal relationship
Evidence is sufficient to conclude that there is a causal relationship with relevant pollutant exposures (i.e., doses or exposures generally within one to two orders of magnitude of current levels). That is, the pollutant has been shown to result in health effects in studies in which chance, bias, and confounding could be ruled out with reasonable confidence. For example: a) controlled human exposure studies that demonstrate consistent effects; or b) observational studies that cannot be explained by plausible alternatives or are supported by other lines of evidence (e.g., animal studies or mode of action information). Evidence includes multiple high-quality studies.

Likely to be a causal relationship
Evidence is sufficient to conclude that a causal relationship is likely to exist with relevant pollutant exposures, but important uncertainties remain. That is, the pollutant has been shown to result in health effects in studies in which chance and bias can be ruled out with reasonable confidence but potential issues remain. For example: a) observational studies show an association, but copollutant exposures are difficult to address and/or other lines of evidence (controlled human exposure, animal, or mode of action information) are limited or inconsistent; or b) animal toxicological evidence from multiple studies from different laboratories that demonstrate effects, but limited or no human data are available. Evidence generally includes multiple high-quality studies.

Suggestive of a causal relationship
Evidence is suggestive of a causal relationship with relevant pollutant exposures, but is limited. For example, (a) at least one high-quality epidemiologic study shows an association with a given health outcome but the results of other studies are inconsistent; or (b) a well-conducted toxicological study, such as those conducted in the National Toxicology Program (NTP), shows effects in animal species, Inadequate to infer a Evidence is inadequate to determine that a causal relationship exists with relevant causal relationship pollutant exposures. The available studies are of insufficient quantity, quality, consistency, or statistical power to permit a conclusion regarding the presence or absence of an effect.

Not likely to be a
Evidence is suggestive of no causal relationship with relevant pollutant exposures.

causal relationship
Several adequate studies, covering the full range of levels of exposure that human beings are known to encounter and considering at-risk populations, are mutually consistent in not showing an effect at any level of exposure.