Looking Across the Research Topics
In the preceding chapter, the committee provided its review of progress on research topics 1-10.1 This new information will help to reduce uncertainties in the framework for assessing the public health risks from emissions of airborne particles and their gaseous precursors. The committee’s research portfolio is for research to investigate particle toxicity, assess human exposures, examine biological mechanisms, identify particle sources, and develop tools needed to formulate effective control strategies. Research findings would constitute the scientific basis for assessing the burden of disease associated with specific particle categories and for evaluating the potential effectiveness of various control strategies for protecting public health. The committee has given similar emphasis to the 10 topics, recognizing that all need to be addressed to have an evidence-grounded approach to controlling particulate air pollution.
As the context for particulate matter (PM) research has evolved, five cross-cutting issues have emerged: (1) an increasing number of adverse health outcomes associated with PM and the related susceptible subpopulations; (2) particle toxicity in relation to different particle characteristics and emission-source types; (3) increasing emphasis on exposure-dose-response relationships; (4) consideration of particle health effects within the broader context of the myriad other pollutants present in the ambient air; and (5) consideration of the implications for setting and implementing the PM NAAQS.
HEALTH OUTCOMES AND SUSCEPTIBLE SUBPOPULATIONS
Research results under the topics of outdoor measures versus actual human exposures (topic 1), dosimetry (topic 6), combined effects of PM and gaseous pollutants (topic 7), susceptible subpopulations (topic 8), and mechanisms of injury (topic 9) indicate a broadening scope of health concerns since the committee’s 1998 report. At that time, emphasis was largely placed on total morbidity and mortality from respiratory causes, such as exacerbation of chronic respiratory diseases, including chronic obstructive pulmonary disease (COPD) and asthma, and the respiratory health of children. Subsequently, the list of particle-related health outcomes was broadened and now includes several adverse cardiac outcomes, such as changes in heart rate variability, cardiac arrhythmias, ischemic events, and congestive heart failure, as well as reproductive outcomes. Although findings on several of these outcomes remain preliminary and inconsistent, the interest in investigating these outcomes and exploring new ones has grown.
Individuals with chronic lung disease have long been considered to be at increased risk for adverse effects of air pollution, based on compromised physiological reserve capacity. Dosimetry studies show that such persons have enhanced deposition of particles in the central lung, possibly contributing to increased risk. The expanded scope of health studies now includes investigation of other potentially susceptible subpopulations, such as individuals with coronary heart disease or diabetes.
New studies were conducted in many U.S. cities to better understand the relationship between outdoor measures of PM and actual human exposures. Panels of susceptible subpopulations were investigated, including people with COPD or coronary heart disease, older adults, children, and people with asthma. These studies found that there were varying degrees of association between personal exposures and ambient concentrations for the measured individuals, with almost half of the associations being nonsignificant (see Chapter 3). More important, studies conducted in the eastern United States showed little difference in PM2.5 exposures among the different investigated cohorts, despite their differing time-activity patterns. Future research studies should be conducted to investigate populations at high risk residing near source-dominated environments. Progress in air quality model development and testing (topic 4) and characterization of emission sources (topic 3) will facilitate greater accuracy in the identification of populations exposed to high PM concentrations.
ASSESSING HAZARDOUS PARTICULATE MATTER COMPONENTS
Ambient particles contain a large spectrum of individual compounds. Research findings from the Supersites Program and other atmospheric characterization studies have elegantly demonstrated the complexity of ambient particle characteristics. Research to assess hazardous PM components (topic 5) seeks to understand the comparative toxicity of particles in relation to their specific characteristics (for example, size or composition). Such information is helpful for the development of effective controls on emission sources to protect public health.
The current National Ambient Air Quality Standards (NAAQS) for PM are based on size and mass and assume that all particles have the same toxicity per unit mass irrespective of chemical composition. In the committee’s judgment, that assumption greatly oversimplifies complex biological phenomena that are influenced by PM and other pollutants. There are numerous physical and chemical characteristics of particles that are potentially relevant to their toxicity; however, to date, there is little information on the relationship between health outcomes and specific particle properties or source types.
Research to date has provided some new insights concerning particle characteristics and toxicity. For example, as discussed in Chapter 3, there are studies suggesting that health impacts of sulfate per se may not be proportional to their contribution to ambient PM mass. From the regulatory point of view, that is an important finding, because ammonium sulfate represents a significant fraction of PM, especially in the eastern United States, where it is the dominant component of secondary PM and is largely attributed to a small range of source types (for example, coal combustion). The toxicity of a range of particle components and sizes will need to be explored across the relevant health outcomes. Investigating the rest of PM components will be a challenging task, considering the complexity of PM as supported by numerous PM characterization studies (see Chapters 1 and 3); however, it is imperative that more progress be made in this area.
Without sufficiently compelling findings on the assessment of hazardous PM components, the committee’s research agenda could stall, and the possibility of standards and control strategies that go beyond the current mass-based approach would be delayed. Further research on emission characterization, development and testing of air quality models, and exposure of individuals at greatest risk needs to be linked to and redirected by advances in research on the assessment of hazardous PM components. A
better understanding of particle toxicity will enable atmospheric scientists to focus their air quality models on particle types most critical to public health. An improved understanding of PM toxicity will help to develop air quality monitoring networks to support epidemiological studies investigating PM toxic components. New information on hazardous PM components would help to guide efforts to better understand the deposition and fate of relevant particles (topic 6). Research on biological mechanisms (topic 9) and identification of susceptible populations (topic 8) could be better enhanced by research accomplishments on topic 5. Findings from research on topic 5 will guide mechanistic studies, topic 9, to focus on specific toxic components or size fractions of PM. Integration of research across topics 5 and 9 should be further enhanced.
The slow pace of research on assessing the hazardous components of PM may reflect not just the difficulty of the scientific questions but also the limitations of the investigator-initiated, hypothesis-driven approach to carrying out systematic screening across the matrix of particle characteristics and health outcomes that is the foundation for topic 5. A large array of questions for investigation is defined when the diversity of possibly relevant particle characteristics is crossed with the broad range of potential health effects. This array has not been screened to identify the most plausible pairings of characteristics and effects. Instead, most investigators have pursued specific hypotheses, often with some justification, but leaving plausible alternatives unexplored and not attempting to systematically cover the full range of potential characteristics of interest. A much more systematic approach may need to be taken for future research to assess the hazardous components of this complex mixture called PM. New research and research management strategies may also be needed, as discussed in Chapter 6.
The committee considered the kind of evidence that would be informative for assessment of hazardous PM components. As emphasized previously, this topic does not have the simplifying and unrealistic objective of identifying single characteristics of particles that determine toxicity. Rather, the emphasis is on assessing the comparative toxicity, including exposure-dose-response relationships, of particles of differing characteristics and from different sources. Epidemiology and toxicology are two relevant and complementary research approaches for pursuing that objective. For either discipline, the ideal data might take the form given in Figure 4-1, which illustrates a comparative assessment of risk for particles of different compositions. Efforts to compare different types of PM are complicated by the possibility that the exposure-dose-response relationships for some types of PM and health effects may not be linear.
Epidemiologists approach the topic of assessing hazardous PM components by carrying out research across locations having particles with different characteristics or across time periods over which particle characteristics might differ. Multivariable models are the fundamental tool for attempting to gauge the comparative effects of exposures to different particle mixes. Additionally, in several epidemiological studies, data on particle composition have been used to develop source-surrogate exposure indicators, and in a larger number of studies, exposure to one source, onroad vehicle exhaust, has been studied. The category of source-oriented studies can provide data on relative toxicity. Toxicologists have compared the toxicity of particles having differing characteristics in the same bioassay system, but there is no standardization of such assays or of the doses at which they are applied. Findings may not be readily compared across studies as a result. Integrated epidemiological and toxicological approaches will be needed so that hypotheses can be mutually tested and findings cross-validated based on human observational and animal experimental data.
INCREASING EMPHASIS ON EXPOSURE-DOSE-RESPONSE RELATIONSHIPS
In the committee’s view, emphasis should be shifted from research directed primarily at the question of whether particles are causing particular health effects to that of characterizing exposure-response or dose-response relationship—that is, what is the form of the quantitative relationship between exposure and risk for an outcome? This quantitative understanding can guide decisionmaking, offering a foundation for estimating the burden of morbidity and mortality caused by particles and comparing the benefits of alternative scenarios of air quality management. The design of research directed at characterizing dose-response relationships may differ from that directed at hazard identification. Information on dose-response relationships is particularly needed across a range of particle exposures relevant to those received by people at contemporary concentrations.
Knowledge of exposure-dose-response relationships is important to establishing the NAAQS for PM and implementing effective control strategies. Exposure-dose-response was viewed as the umbrella issue that included health-related research topics. During the past 6 years, research on these topics focused mainly on examining the causal association between PM exposure and increases in risks for adverse health effects. Very few of the studies using in vitro or laboratory animal models have included more than one exposure or dose level, limiting progress in characterizing exposure-response relationships. However, the research performed to date provides essential background information for conducting the next generation of research on population-based exposure-dose-response models, which are needed to evaluate the effectiveness of implementation plans.
The incorporation of approaches that will allow examination of exposure-dose-response relationships in in vitro and laboratory animal experiments will be particularly critical in further research evaluating the comparative toxicity (potency) of PM components. The strongest information base on exposure-response relationships for PM and other pollutants comes from some of the largescale epidemiological studies that have examined several PM metrics, such as PM10 or PM2.5, and reported the results as an increase in adverse effects per 10 micrograms per cubic meter (µg/m3) of the metric. The linear relationship between exposure and response has been supported by some detailed evaluations that have generally supported a monotonic increase in response associated with increased exposure. It is anticipated that future epidemiological studies can build on the toxicological findings and test the hypothesis that some specific PM components have potencies different from those observed for PM10, PM2.5,
or particles with aerodynamic diameters between 10 and 2.5 μm. In addition, epidemiological studies can benefit from future exposure studies of susceptible individuals to particles and other pollutants. Also, exposure studies will be necessary to inform epidemiological investigations about relationships among personal exposures to hazardous PM components and ambient concentrations for susceptible subpopulations and the general public.
MIXTURES AND COPOLLUTANTS
For some time, investigators have recognized that surrogates for “dirty air” have been derived from assessing one pollutant at a time, and possible health effects attributed to a single pollutant have often been used in part to make regulatory decisions. Over the past several years, scientists and regulators have been concerned about human exposure to air pollutant mixtures in settings other than specific occupational settings (topics 1 and 2). In addition, the recognition of a broader range of health impacts as being putatively related to ambient pollution raises the possibility that single pollutants might be acting together to increase risk through additive or synergistic mechanisms (topic 7).
Research findings on the combined effects of particles and gaseous copollutants (topic 7), susceptible populations (topic 8), mechanisms of injury (topic 9), and human exposures (topic 1) have implications that extend beyond PM alone. The finding of synergism between PM and other criteria pollutants, especially ozone, would have implications for the form of the NAAQS, which assumes independence of effects. For instance, clear evidence of synergism among the various pollutants for which NAAQS are set would provide a rationale for more integrated standards reflecting realistic atmospheric mixture exposures to populations at risk and the potential for overall mixture toxicity. A more detailed discussion on this topic is presented in Chapter 6.
There is overlap among some mechanisms and targets, such as activation of inflammatory cascades, that are considered to underlie the adverse effects of PM and certain gaseous pollutants (topic 9). Oxidative injury is another mechanism thought to have a role in lung injury by PM, ozone, and nitrogen oxides (topic 9). Thus, the insights gained from investigations of mechanisms of respiratory injury by PM may have wider applicability, although some pollutants have specific targets, for example, carbon monoxide and hemoglobin.
Recent reports on proximity to traffic and risks to health further support research directed more broadly at air pollution. Studies on the adverse health effects of air pollution on children with asthma and other populations residing in relatively close proximity to highly trafficked roadways underline the importance of identifying susceptible populations on the basis of exposure to high concentrations of air pollution (topics 1, 2, 3, and 4) and increased responsiveness to air pollution toxicity (topics 8 and 10). Additional studies on the exposure-response patterns of older subpopulations and patients with diabetes who appear to be susceptible to multiple pollutants further underline the importance of mixtures. To date, limited human health effects research has been conducted to apportion the effects of single pollutants coexisting with other pollutants in complex mixtures (topic 7). Similarly, only a few animal inhalation studies have used more than two pollutants in sequence or simultaneously, for example, concentrated ambient particles (CAPs) with carbon monoxide or ozone (topic 9). However, considering the complexity of exposures and the variety of potential responses to air pollutant mixtures, the paradigm of assigning effects to single pollutants should change.
Important remaining research needs are those that could affect the four elements of the NAAQS for PM, namely, the indicator, the averaging time, the numerical level of the indicator, and the statistical form of the standard. For control purposes, sources that should have priority for reduction need to be identified. Exposure assessment, emissions characterization, development and testing of air quality models, assessment of hazardous PM components, combined effects of PM and gaseous pollutants, and mechanisms of injury (topics 2, 3, 4, 5, 7, and 9) are particularly relevant. The current and all previous versions of the NAAQS for PM have been based primarily on human epidemiological data with support of data from other lines of research. The committee’s review of current and ongoing research suggests that epidemiological evidence will remain central in setting the NAAQS, but enhanced mechanistic understanding could guide the choice of the indicator and averaging time.
More research is necessary for emissions characterization and air quality model testing and development to meet implementation deadlines (see Chapter 5). Measurements of source emission rates for PM and precursor gases as well as evaluation of models are needed, because states are in the process of developing state implementation plans for attaining PM NAAQS. To expedite progress on those topics, EPA should provide more guidance, leadership, and coordination among the groups carrying out this work, particularly those conducting emissions characterization (see Chapter 6).
Results of epidemiological research could provide greater information about those components of PM and air pollution that can most impact health. Future state implementation plans (SIPs), which are plans for reducing emissions so that an area can come into compliance with NAAQS, could therefore benefit from new and existing research in this area. Such information would enable SIPs to prioritize emission-control efforts starting with those sources contributing the greatest potential health impacts.