Chapter 29: Air Pollution

The second half of the 20th century was marked by remarkable changes in how the public viewed its relationship to the environment. Until that time, national pride and prosperity were often depicted as an expanse of urban factories with smokestacks belching opaque dark clouds of industrial effluent into a neutral blue sky. But the price of that unchecked human progress through the first half of the century led to several air pollution catastrophes highlighting the profoundly detrimental impact that reckless prosperity could have on the environment. These images of “modern” life gradually gave rise to public outcry for governmental action to protect air quality and public health—a challenge to industry that had been focused on economic growth alone. The ensuing 50 years of regulatory legislation in the United States and Western Europe along with cost-efficient innovations by the private sector have remade this industrial image in most technologically developed nations.

Ironically, as regulatory control measures began to reduce emissions from stationary industrial sources of air pollution, highways to “open spaces” and urban flight took many people to the suburbs with its cleaner air and safe, comfortable lifestyle. Meanwhile, the developing world saw little of this growth and what has grown is frequently cast-off old technology and variants of exploitation by the Western corporations attracted by abundant resources, a cheap workforce, and few regulations of constraints. This situation has persisted into the 21st century but is evolving with broader globalization of improved technology and communication.

The change in land use and demography in the United States in the 1950s and 1960s altered the national character and distribution of air pollution. The commute from suburban home to city workplace back to suburban home led increasingly to congested thoroughfares, whose emissions contributed to a photochemical cauldron of oxidant air pollution around expanding metro–suburban areas. Moreover, postwar population growth and rising expectations for a better (peace-time) standard of living led to unrestrained consumption, including inexpensive gasoline for commuting and recreation.

Today, the search is worldwide for cheap oil to fuel transportation and goods movement as easily accessible North American oil resources have been greatly diminished. International sources are increasingly prominent. For nontransportation sectors of industry and electricity generation, low cost and domestic availability have brought coal to the top of the energy pyramid. Other energy sources, natural gas and biomass in its varied forms, are being intensively explored to meet the energy demands of modern life. Most notable as a potential major resource is natural gas, which is being recovered through less conventional methods (eg, fracking) which in 2012 is thought by some to be the next revolution in energy resources. Decades to come will see an evolution in our energy portfolio driven by cost and access, environmental impacts including climate change, and technological innovation. Nevertheless, so long as organically derived fuel is combusted to derive energy, its potential for impact on air quality and on public health and the environment will remain.

While great strides have been made in balancing regulation and technology to reduce emissions from stationary and mobile sources, unsatisfactory air quality continues to impose a risk to public and environmental health even in highly developed countries throughout the world. More than half the US population resides in counties that are not in compliance with current National Ambient Air Quality Standards (NAAQS) (Fig. 29-1). These noncompliant areas correspond well with where people live and work and reflect the redistribution of population growth from major urban and industrial centers. Air pollutants can also be transported across many miles such that previously pristine, rural areas have measurable pollution above their historic background levels. Indeed, as the developing world grows industrially, air pollution now is intercontinental with transport through the atmosphere via pathways close to the earth’s surface as well as upper atmosphere. Emissions of precursors of ozone can come from hundreds of miles away and contribute to local pollution levels; likewise other emissions such as mercury from coal-fired power plants in Asia can account for the majority of US deposition on land and in streams many thousands of miles from the original sources (Hemispheric Transport of Air Pollution, 2010).

While episodes of extreme air pollution are rare in the Western world today, occasionally, unusual meteorological stagnations exaggerate typical air pollution patterns in both the levels of pollutants and their expanse (eg, 1995 in the United Kingdom and Western Europe). Likewise, in specific locales prone to stagnations due to topography (eg, Utah Valley through the 1980s–1990s) or where there exists a single dominant source (eg, port areas, transport terminals, or smelters), repeated episodic or even patterned poor air quality can impact public health. From a broader public community perspective, typical exposures are characterized by prolonged periods of relatively low, cyclic levels of complex mixtures of photochemically transformed industrial and mobile emissions with occasional moderate excursions due to weather patterns. As pollution now extends even into remote and wilderness areas, significant damage to flora and crops can also occur.

The situation in many developing countries experiencing rapid population growth, industrialization, and economic expansion is reminiscent of the urban, industrial pollution in the United States and Europe more than half a century ago. The lessons of Western industrial nations and the consequences of inadequate air pollution control in some countries have not necessarily outweighed economic drivers or public health concerns. For example, despite being a modern economic center, two marathoners died in Hong Kong in February 2006 while running on a weekend during a protracted period of severe air pollution. Just as in US history, the urban air in rapidly industrialized nations simply reflects poor air quality management and the desire to prosper economically. Air pollutant levels in some of the developing urban centers can run 10 or more times those of the United States today, and their expanding worldwide distribution contributes substantially to the global burden of many air pollutants such as mercury, particulate matter (PM), sulfate/SO₂, and major global warming pollutants such as black carbon, CO₂, and methane. A willingness to balance worldwide economic growth and industrialization using lessons learned from the misadventures of the Second Industrial Revolution and the Age of the Automobile will determine the impact of the “new” global economy on the health of both the earth and its inhabitants.

Other issues facing many parts of the developing world tie closely to domestic culture and economy, as well as to the level of technological sophistication. Prime among these problems is exposure to carbon and soot from combustion of biomass in cooking and heating in domestic stoves. Approximately three billion people worldwide use biomass for home cooking in households with little ventilation. The World Health Organization (WHO) estimates two million deaths per year result from these exposures, especially women who are exposed day in and day out over many years, often with their infant children by their sides. Understanding the intersection of technological as well as socioeconomic and political challenges will be at the core of any resolution to these issues.

The gathering of scientific information regarding the impacts of air pollution on human health began with general indices of haziness or the darkness of a spot on a filter. The interest in learning causation drove science to focus more on individual pollutants to better understand the linkages with health and to aid in mitigation of risk. In turn, this knowledge was used both to develop public health standards and to establish regulatory controls. Despite the obvious complexities of urban air pollution, it was felt that single pollutant regulations were the best path to success, and, indeed, great strides have been made in that regard. However, research agendas are evolving to include the study of multipollutant interaction and transformation of individual pollutant components within atmospheric systems—a truer reflection of real-world exposure complexities. These multipollutant approaches entail sophisticated statistical models and empirical study designs, all with the goal of a more complete systems-based understanding of air pollution and risk to maximize benefits while minimizing costs and unintended consequences.

This chapter will present an overview of the current state of knowledge regarding air pollution and its ensuing impacts on human health, and the role of toxicology and translational biology in unraveling the complexities to allow for informed assessment of risk. As such, the contextual complexities of regulatory decisions and risk assessment will also be addressed. The intent is to provide the reader with both a fundamental knowledge and appreciation of the nature of the problem, how traditional and modern-day approaches aid in this understanding, as well as a sense of the uncertainties in need of future investigation to resolve issues even before they arise.

A Brief History of Air Pollution and Its Regulation

For most of history, air pollution has been a problem of microenvironments and domestic congestion. The smoky fires of early cave and hut dwellers choked the air inside their homes, and even when the emissions were vented outdoors, they simply combined with those of the neighbors to settle around the village on damp cold nights. With urbanization and a concomitant decrease in forest wood as a source of fuel to heat and cook, the need for energy led to the burning of easily accessible, dirty coal and the ambient release of sulfurous, sooty smoke. Industrialization brought kilns to make quicklime for construction and metal smelters needed for the development of progressive “modern” cities, only to push smoke and chemical emissions into the air. Unfortunately, the city dwellers who worked near these industries had to endure the bad air, while those of wealth frequently had country homes to which they could escape. The poor quality of urban air has been captured historically by many writers including Charles Dickens, who noted London’s fogs in his writings, all the way back to the ancients such as Seneca, the Roman philosopher, who in ad 61 wrote: “As soon as I had gotten out of the heavy air of Rome, and from the stink of the chimneys thereof, which being stirred, poured forth whatever pestilential vapors and soot they had enclosed in them, I felt an alteration to my disposition” (emphasis added: Miller and Miller, 1993).

Historically, efforts to regulate air pollution have, like today, competed with national, regional, and industrial economies, and, as a result, they have evolved slowly. Early on, in the time of Greece and Rome, individual civil suits could be levied against local polluters, although these were of marginal success. Beginning in the 13th century, community-based outcries received some recognition by governing officials; one example was the banning of “sea coal” from lime kilns and domestic heaters in London by Edward I. Enforcement, however, was not effective and the populace largely resigned itself to polluted air as part of urban life. By the 17th century, England, in the middle of several decades that some refer to as “the little ice age,” depleted its forest wood resources, which only increased reliance on sea coal for domestic heating. Meanwhile, Percival Pott’s discovery that chimney soot was related to the incidence of scrotal cancer in chimney sweeps led to some in the health community to recommend: “Fly the city, shun its turbid air; breathe not the chaos of eternal smoke…” (Brimblecombe, 1999)—essentially the same advice advanced by Seneca 1600 years earlier. In the late 18th century, the industrial revolution, which was powered by the burning of “cleaner” mined coal, added a second dimension to urban air pollution. These emissions were more acidic and hung in the air longer than the fluffy soot of the cheaper sea coal. Continued soiling of buildings and damage to nearby crops brought community boards to address sanitary reforms to cut the worse of the pollution peaks and episodes, but any gains were soon offset by growth. By the end of the 19th century and into the early 20th century, power plants were being built to provide energy for factories and eventually to light homes in the Western world. Steel mills and other industries proliferated along riverbanks and lakeshores, oil refineries rose in port cities, and near oil fields smelters roasted and refined metals in areas near large mineral deposits.

By 1925, air pollution was common to all industrialized nations. However, people slowly grew less tolerant of the nuisance of acidic-soot corrosion of all exposed surfaces and the general discomfort that came with smoky air—this acidic, sooty form of air pollution was termed reducing air pollution because of its reducing chemical nature. Public surveys were initiated—in Salt Lake City in 1926, New York City in 1937, and Leicester, Great Britain, in 1939—to bring political attention to the problem and promote the implementation of controls (Miller and Miller, 1993). However, it was the cumulative impact of the great air pollution disasters in the Meuse Valley, Belgium, in 1930; Donora, Pennsylvania, in 1948; and the great London fog of 1952 that indicted air pollution as a health risk. In the United States, California was already leading the way with passage of the Air Pollution Control Act of 1947 to regulate the discharge of opaque smokes. Visibility problems in Pittsburgh during the 1940s had also prompted efforts to control smoke from local industries. At the national level, however, it was the initiative of President Truman that provided the federal impetus to deal with air pollution. His early efforts culminated in congressional passage of a series of acts starting with the Air Pollution Control Act of 1955 under the Eisenhower administration.

The prosperity of the late 1950s and the accompanying suburban sprawl introduced the third and perhaps most chemically complex dimension of air pollution. The term smog, though originally coined to describe the mixture of smoke and fog that hung over large cities such as London, was curiously adopted for the eye-irritating photochemical reaction products of auto exhaust that blanketed cities such as Los Angeles and Mexico City. Early federal legislation addressing stationary sources was soon expanded to include automobile-derived pollutants in the United States (the Clean Air Act [CAA] of 1963, amended in 1967, and the Motor Vehicle Air Pollution Control Act of 1965). The landmark CAA Amendments of 1970 evolved from the early legislation, and, despite being only an amendment, it was revolutionary. It recognized the problem of air pollution as a national issue and set forth a plan to control it. The Act charged the newly instituted US Environmental Protection Agency (USEPA) with the responsibility to protect the public from the hazards of polluted outdoor air. Seven “criteria” air pollutants (ozone [O₃], sulfur dioxide [SO₂], PM, nitrogen dioxide [NO₂], carbon monoxide [CO], lead [Pb], and total hydrocarbons—the last now dropped from the list, leaving six criteria pollutants) were specified as significant health hazards in need of individual NAAQS. These NAAQS were mandated for review every five years as to the adequacy of the existent standard to protect human health (Table 29-1), although adherence to this schedule has generally not been achieved. The explosion in the literature databases for the criteria pollutants and the extensive review and commentary process has typically led to delays in completing the process on schedule. In 2006, however, this process was streamlined through the creation of an Integrated Science Assessment (ISA) to replace each Criteria Document where the findings since the last assessment could be integrated with older studies (summarized with the foundation literature confined to tables and an appendix). This ISA is combined with a Risk and Exposure Assessment into a Policy Assessment to develop a range of proposed standards based on risk analyses. The US EPA Administrator considers proposals from both “evidence”- and “risk”-based assessments to establish policy and set the NAAQS (www.epa.gov/ttn/naaqs/naaqs_process_report_march2006.pdf). With regard to the Primary NAAQS, only health criteria can be considered in the development of the standard, including safety considerations for potentially susceptible groups. The Secondary NAAQS consider agricultural and structural welfare. Economic impacts are not to be involved in the standard setting process itself—only in assessing the cost of the implementation procedures. Other hazardous air pollutants (HAPs), of which there were eight listed in 1970, were to undergo health assessments to establish emission controls—today there are 189 (see below). The CAA of 1970 was by far the most far-reaching legislation to control air pollution to date.

The accidental release of 30 tons of methyl isocyanate vapor into the air of the shanty village of Bhopal, India, on December 3, 1984, killed an estimated 3000 people within hours of the release, with several thousand delayed deaths and 200,000 injured or permanently impaired. The tragedy shocked the world, and raised the issue of HAPs in the United States to a new level of concern. While such a disaster has never struck the United States, accidental industrial releases or spills of toxic chemicals are surprisingly common, with 4375 cases recorded between 1980 and 1987, inflicting 11,341 injuries and 309 deaths (Waxman, 1994). The Toxic Release Inventory (TRI) of 650 chemicals reported annually by industry is required by EPA and suggests a decrease in releases (~9% in 10 years across land, water, and air), but these data are self-reported and do not include most of the 80,000 low-volume chemicals in use today.

The HAPs, which had been the stepsister of the criteria pollutants for more than a decade after the passage of the 1970 CAA, have since garnered more public and policy attention. There is concern not only for the acute effects of accidental releases of fugitive or secondary chemicals—such as phosgene, benzene, butadiene, and dioxin, into the air of populated industrial centers—but also for potential chronic health effects, with cancer often being the focus of attention. The slow progress of regulatory decisions on HAPs (only eight between 1970 and 1990) led to a mandated acceleration of the process under the CAA amendment of 1990. Section 112(b) currently lists 189 chemicals or classes of chemicals for which special standards and risk assessments are required. The chemicals listed are those of greatest concern on the basis of toxicity (including cancer) and estimated release volumes. Currently, there is a list of 33 HAPs from the list of 189 that are deemed to be of greatest concern, so-called a list of the “dirty 30.” Emissions of HAPs are mandated for control to the maximal achievable control technology (MACT), and any residual health risk after MACT is to be considered in a separate quantitative risk assessment. The database for this process utilizes existing knowledge or, if necessary, mandates further research by the emitter. While many of these chemicals are now better controlled than in the past, residual risk estimates are yet to be completed for many HAPs. The database from which these assessments are made is called the Integrated Risk Information System (IRIS, www.epa.gov/iris/index.html) and currently contains 550 chemicals that have health data.

Emissions from motor vehicles include a number of HAPs (eg, benzene, butadiene) as well as NAAQS pollutants (eg, CO, NO_x, PM), but are addressed primarily under the CAA Title II, Emission Standards for Mobile Sources. The reduction of emissions from mobile sources is complex and involves both fuel formulation and engine/vehicle reengineering. Refinements in combustion engineering are ongoing making use of advanced computerized ignition and timing, but fuel properties have drawn most recent attention for improvement. For example, to reduce wintertime CO, several oxygenates (including ethers and alcohols) have been formulated into fuels to both reduce cold-start emissions and enhance overall combustion. Perhaps the most prominent of the ethers is methyl tertiary-butyl ether (MTBE), which became a controversial additive in the early 1990s, arising in part from odor and reports of asthma-like reactions by some individuals during auto refueling. However, the controversy took an unexpected twist; MTBE has been removed from fuel, not because of health concerns associated with airborne exposure but rather due to leakage from service-station storage tanks into groundwater. Ironically, this prescribed remedy for an air problem evolved into an unexpected problem: groundwater contamination. This example illustrates the broad complexity of all pollution control measures in that unappreciated factors can transcend engineering or other targeted features if a broader systematic assessment is not undertaken. Nevertheless, fuel additives have been developed to boost octane ratings of fuels and/or improve engine performance and combustion (eg, organic oxygenates, methylcyclopentadienyl manganese tricarbonyl [MMT], platinum compounds for diesel). These additives are reviewed under Title II because of concerns regarding the potential changes in combustion product reactivity or the introduction of metals into the environment, reminiscent of use of lead in fuels from the 1930s to the 1970s, when lead fuel additives were banned in the United States.

Internationally, the magnitude and control of air pollution sources vary considerably, especially among developing nations, which often forgo concerns for health and welfare because of cost and the desire to achieve prosperity. Fig. 29-2 illustrates the international variation in air pollution–related mortality (outdoor and indoor) based on economic groupings. It is clear that there are wide differences reflecting economic imbalances—particularly prominent are the indoor particulate levels in developing nations where biomass combustion is used for heating and cooking. These regions also contain many of the megacities of the world with major air pollution problems. The political upheaval in Eastern Europe since 1990 has revealed the consequences of decades of uncontrolled industrial air pollution. While vast improvements are now evident in this area, as industries are being modernized and emissions controlled, many Asian, African, and South American cities have virtually unchecked air pollution. Some nations as well as the WHO have adopted air quality standards as a rational basis for guiding control measures, but the lack of binding regulations and/or economic fortune impedes significant controls and improvements (Lipfert, 1994). In addition to local socioeconomic and political concerns, emissions of air pollutants have spawned problems of “international pollution.” Long-range transport of polluted air masses from one country to another has been anticipated as a global issue for several years (Reuther, 2000). This issue was the subject of some controversy between Canada and the United States in the late 1980s and into the 1990s as a result of the air mass transport of acid sulfates from industrial centers of the Midwestern United States to southern Ontario. However, reduction in SO₂ emissions has somewhat relieved the tension over the last several years (Fig. 29-3). Improvement in NO_x, also a product of stationary source fossil fuel burning, is also apparent. Importantly, growing prosperity in China has led to enormous growth in its power sector and coal combustion resulting in transoceanic transport of airborne mercury such that transported mercury now accounts for the majority of mercury deposition in the waters and land in the United States (Seigneur et al., 2004).

Risk assessment is a formalized process, originally described in the landmark 1983 National Research Council report, whereby toxicity, exposure, and dose-dependent outcome data can be systematically integrated to estimate risk to a population. Fig. 29-4 provides a modified version of the paradigm of the NAS (NRC, 1983), incorporating recent interest in providing evidence of “accountability” that the regulations indeed do have impacts. The health database for any air pollutant may comprise data from animal toxicology, controlled human studies, and/or epidemiology. But, because each of these research approaches has inherent strengths and limitations, an appropriate assessment of an air pollutant requires the careful integration and interpretation of data from all three methodologies. Thus, one should be aware of the attributes of each (Table 29-2).

Epidemiological studies reveal associations between exposure to a pollutant(s) and the health effect(s) in the community or population of interest. Because data are garnered directly under real-world exposure conditions and often involve large numbers of people, the data are of direct utility to regulators assessing pollutant impacts. With proper design and analysis, studies can explore either acute or long-term exposures and theoretically can examine patterns in mortality and morbidity, both acute and chronic, especially if these responses appear disproportionately in population subsets (ie, sensitive groups). Why, then, is this approach to the study of air pollution not the exclusive choice of regulators in decision making? The problem is that it is difficult to control confounding personal variables in the population. Factors such as genetic diversity and lifestyle differences among individuals, and population mobility are difficult to control. Perhaps most problematic is the lack of adequate exposure data—especially on a personal basis. Exposure assessment is often one of the major weaknesses of an epidemiological study, not only because of the difficulties of assessing exposure (as a measure of dose) to the pollutant of interest but also because it is difficult to segregate a single pollutant from correlated copollutants and other environmental influences, such as meteorology. Thus, only associations, and not causality, can be drawn between the broad-based exposure data and effects observed in epidemiology studies. Causal relationships are sometimes inferred in the presence of strong statistical significance, but such determinations are likely to be criticized. However, recent advances in exposure estimation and study design and analysis (eg, time series) have allowed epidemiologists to examine relationships with greater confidence and specificity. These models limit the impact of covariates and longer time-based influences and thus allow epidemiologists to tease out effects of short-term pollution not accessible formerly (Schwartz, 1991). Similarly, newer approaches that employ field studies—sometimes called panel studies—incorporate time-series design and multiple regression analyses of more focused and complete exposure data (ideally personal) and targeted clinical end points in the exposed population under study. The end points often derive from empirical human and animal studies and therefore have a priori conceptual ties. The advent of new genetic approaches for characterizing polymorphisms of potentially influential traits (eg, glutathione S-transferase M1 [GSTM1]; Tujague et al., 2006) opens the genomic door for assessing gene–environment interactions. These factors may well underlie a significant portion of the human susceptibility to air pollution. Novel approaches such as this are evident in the most recent studies of PM air pollution (see below).

Studies that involve controlled human exposures have been used extensively to evaluate the criteria air pollutants regulated by the USEPA. Because most people are exposed to these pollutants in their daily lives, human volunteers can be ethically exposed to them in a highly controlled fashion (with the exception of Pb, which has cumulative and irreversible effects). Exposures are conducted in a controlled environment (usually in a chamber or with a mask) and are generally of short or limited repeat durations, given assurances that all responses are reversible. Clearly, data of this type are very valuable in assessing potential human risk, since they are derived from the species of concern and are rooted in well-established clinical knowledge and experience.

Suspected “susceptible or sensitive” individuals (see below) representing potential higher-risk groups can also be studied to better understand the breadth of response in the exposed public. However, clinical studies have several practical limitations. Ethical issues are involved in every aspect of a clinical test; potentially irreversible effects and carcinogenicity are also always of concern, along with the definition of an acceptable level of hyperresponsiveness in so-called sensitive individuals who volunteered to participate in the study. Likewise for any test subject, there are obvious restrictions on the invasiveness of biological procedures, although sophistication in medical technology has made accessible a large array of molecular biomarkers from peripheral blood and nasal, bronchial, and alveolar lavage fluids as well as biopsied cells from airway segments (Devlin et al., 1991; Salvi et al., 1999). With the advent of cutting-edge genomic and proteomic high-throughput technologies, new tools are emerging to identify responses in humans and the relationships to innate susceptibility. Obviously, the issue of cost, the limited numbers of subjects that can be practically evaluated, and the inability to address chronic exposure issues remain constraints on human testing. Where partnerships with animal toxicology studies have been established, studies in laboratory animal species can sometimes provide ethical justification for at least limited direct human exposure for addressing critical questions. Analogously, in vitro studies in both human and animal cell and tissue systems, often augmented with similar genomic tools, allow the elucidation of mechanisms of toxicity. These basic biological responses inform extrapolation models that link animal data to humans, and they support the feasibility and preclude risks that may reside in human studies with some toxic air pollutants (see below).

Animal toxicology is used to predict or corroborate, through plausible mechanisms, suspected effects in humans. In the absence of human data, animal toxicology constitutes the essential first step of risk assessment: hazard identification. Animal toxicology is often required before any controlled human exposure can be conducted. It is particularly useful in elucidating pathogenic mechanisms involved in toxic injury or disease, providing basic knowledge that is critical to extrapolating databases across species, estimating uncertainties, and determining the relevance of information to humans. Knowledge of the toxic mechanism(s) provides the underpinnings to the “plausibility” of findings in the human context and, under carefully defined and highly controlled circumstances, may allow quantitative estimates of risk to human populations. Animal toxicology studies have been used to investigate all of the criteria air pollutants and many of the HAPs as well. The strength of this discipline is that it can involve methods that are not practical in human studies and can provide more rapid turnaround of essential toxicity data under diverse exposure concentrations and durations. The minimization of uncontrolled variables (eg, genetic and environmental) may be the greatest strength of the animal bioassay.

The clear limitation of animal studies in human risk assessment lies in the unknowns that weaken the extrapolation of findings in animals to the day-to-day human life scenario. Ideally, a test animal is selected with knowledge that it responds in a manner similar to that of the human (homology). Qualitative extrapolation of homologous effects is not unusual with many toxic inhalants, but quantitative extrapolation is frequently clouded by uncertainties of the relative sensitivity of the animal or specific target tissue compared with that of the human. Uncertainties about the target tissue dose also loom large, constituting the first obstacle to quantitative extrapolation (see below). With respect to the target tissue dose, however, most animal toxicologists make every effort to keep exposure concentrations at 5- to 10-fold that of the anticipated human exposure until appropriate dosimetric data can be ascertained. An often overlooked issue is that the dose to the target (lung region) for the test animal is frequently less than that of the human under similar exposure conditions—especially when exposures are conducted during dormancy for the animals (Wichers et al., 2006). Most human inhalation studies involve exercise exposure paradigms where increased ventilation augments dose. Thus, in the absence of exercise, higher doses may be needed to achieve a group response among a limited pool of genetically similar animals (maybe 6–10) to represent a large population effect. However, it must be appreciated that mechanisms may well differ at different dose levels and some responses may be misleading if assessed only at the higher dose levels. Despite these limitations, however, animal studies have provided the largest database on a wide range of air toxicants and have proven utility in predicting human adverse responses to chemicals. New high-throughput technologies offer great promise in shortening the toxicity assessment process and reducing costs of both the test process and the consumption of animals, but it remains to be seen how this new technology aids in the quantitative assessment of toxic inhalants.

To be effective, any health assessment should consider the strengths and weaknesses of the approaches selected to estimate actual toxic risk. In the larger picture, other scientific disciplines can be highly valuable to a more accurate assessment of the impact of air pollution on society. The atmospheric sciences, particularly the chemical and physical sciences, provide insights into exposure by better characterizing what is in the air. Better pollutant characterization, linked to exposure assessment, can only strengthen epidemiological outcome associations. Similarly, better air characterization and dose estimates support toxicity evaluations based on biological test systems. In the latter case, the uncertainties of dose relevancy can be reduced and linkages to defined pollutant physicochemical attributes and interactions can be pursued. Recreating realistic exposure environments to the extent possible is invaluable to developing models to estimate human risk.

Lastly, studies of botanical responses to air pollutants are now appreciated more than ever. Not only are commercial and native vegetation affected by pollution but also some plant species are being exploited as sensitive “sentinels,” warning of the impacts of pollution on both human and environmental receptors. When considered collectively, economists can inform regulators and the public at large of the cumulative impact and adversity of pollution on our quality and standard of living (Maddison and Pierce, 1999). Interestingly, some basic mechanisms (eg, the involvement of antioxidants) between plants and animals have remarkable parallels.

Animal-to-Human Extrapolation: Issues and Mitigating Factors

The value of animal toxicology in inhalation studies is highly dependent on the ability to extrapolate or relate empirical findings to real-world human scenarios. Several factors of study design play into the process of extrapolation (eg, exposure concentration, duration, and patterns), but most important is the selection of the animal species that will serve as the toxicological model. Therefore, this selection should involve more than considerations of cost and convenience. Whenever possible, effects that are homologous and involve the same mode of action between the study species and the human should guide the decision of the most appropriate test species. For example, if upper airway irritant responses such as bronchoconstriction are anticipated (eg, SO₂ or formaldehyde), then the guinea pig, with its human-like reactive airway reflexes, should be selected over the rat, which is not particularly responsive in this regard. However, if the underlying molecular events in tissue remodeling are of interest, then rat or mouse might better serve as a model because of cellular mechanistic parallels with human tissue responses. In part, the availability of probes to aid in such studies could factor into the selection of the rodent strain as well. Additionally, as inbred strains of rats and mice differ in their neutrophilic responsiveness to deep lung inhalants (eg, O₃), contrasting responsiveness or the mode of action may be more revealing and support one strain over another in its selection for study (Costa et al., 1985). Other innate differences in sensitivity among species may also relate to differences in lung structure, regional distribution of cell types having varied metabolism, genetic polymorphisms, or antioxidant defenses. Genetic control of responses such as inflammation have been identified in mice where a specific quantitative trait locus (Inf2) on chromosome 17 exhibits a more robust neutrophilic response to O₃ (Bauer and Kleeberger, 2010). Thus, when a test animal strain is selected where such nuances are unclear or unknown, the replication of responses in multiple species builds confidence in the finding as being homologous or having species-conserved modes of action that are relevant to the human and bear careful consideration in a risk assessment.

An essential, but often overlooked, part of response extrapolation from species to species is knowledge of the relative dosimetry of the pollutant along the respiratory tract. Significant advances in studies of the distribution of gaseous and particulate pollutants have been made through the use of empirical and mathematical models, the latter of which incorporate parameters of respiratory anatomy and physiology, fluid dynamics, and physical chemistry into predictions of deposition and retention. Empirical models combined with theoretical models aid in relating animal toxicity data to humans and help refine the study of injury mechanisms due to better estimates of the target dose. Fig. 29-5 illustrates the application of such an approach to the reactive gas O₃ and insoluble 0.6-μm spherical particles, respectively, as each is distributed along the respiratory tract of humans and rats. Anatomic differences between the species clearly affect the deposition of both gases and particles, but the relative distributions are quite similar. This is not surprising if one argues teleologically that the lungs of each species evolved with similar functional drivers (ie, O₂–CO₂ exchange, blood acid/base balance), mechanical impediments, and environmental stresses. One needs only a cursory review of the comparative lung physiology literature to appreciate the allometric consistency of the mammalian respiratory tract to meet the challenge of breathing air. This design coherency has provided the fundamental rationale for the use of animal models for the study of air pollutants. With knowledge of ventilation parameters, one should be able to compute nominal tissue doses based on lung surface area using available published data.

What is an Adverse Health Effect?

Establishing criteria to define an “adverse effect” of air pollution is difficult. When relating a health effect to an air pollutant, a response must be appreciated at two levels—that of the individual and that of the population. Clearly, an effect on an individual can be beyond an acceptable limit potentially putting that person’s overall health in jeopardy, but this response may be lost in an index reflecting a population-based response. The risk to a population reflects the averaging of individual responses or risks and may be measured as a shift in the normal distribution of some index of response for that population. Hence, on average, the entire population may be judged to be at some enhanced risk. These two forms of risk are clearly related, but most often in practice, the population risk is considered most appropriate from a public health perspective. It is also generally most credibly quantifiable. However, subpopulations that carry some related risk factor may represent a more limited group with unusual risk that too may be quantifiable. Such a group is often called a susceptible subpopulation.

Defining an air pollutant effect as “adverse” within the range of effects that may result from exposure is not always straightforward. Clearly, in humans, some effects would pass uncontested as adverse, for example, death, acute life-threatening dysfunction or disease, irreversible impairments, and pain. In animal models, pathology has traditionally been the hallmark of an adverse effect. In either humans or animals, however, other effects that reflect minor and temporary dysfunctions or discomfort could be argued as not warranting significant or costly concern, especially if the effects are minor and transient with no long-term untoward outcomes. This vein of thought would simply attribute these effects to be within normal physiologic ranges and are readily compensated within functional or biochemical reserve. Thus, if one is to try to assess the impacts of air pollution on health, it is desirable that there exist some objective criteria to define what is indeed adverse based on the nature and the magnitude of the effect under evaluation. Moreover, distinguishing an air pollution effect from other adverse stimuli or disease processes can be complex and fraught with confounding factors, such as smoking and negative lifestyle factors.

Not surprisingly, air pollution impacts have long been focused on the lung as the primary target organ. With regard to acute effects in humans, the American Thoracic Society in 1985 issued a position paper that attempted to define an adverse effect related to air pollution in terms of its impact on lung function (where, eg, a 10% or more drop in forced vital capacity [FVC] was deemed adverse); at that time, lung function was the primary human health effect that could be studied under controlled inhalation conditions. The definition of “adverse” was revised in 2000, because of the many advances in clinical medicine and empirical health sciences, to include seven broad areas: biomarkers, quality of life, physiologic impacts, symptoms, clinical outcomes, mortality, and population health versus individual risk (American Thoracic Society, 2000). The summary conclusion states that caution should be exercised in evaluating the many new biomarkers of effect (especially cell and molecular markers), as there is need for validation that small changes in these markers represent a progression along a course to disease or permanent impairment. Since that time, further advances in molecular markers and systems assessment of toxic pathways have opened a new dimension of evaluation of health outcomes where their detailed pathogenesis can be examined. Of course, many of the same questions arise as to the criteria to define when a pathogenic process is qualitatively or quantitatively adverse, but clearly the ability to measure effects and understand their nature and relationship to chronic disease is advancing rapidly.

Susceptibility

A common thread through all of these subject areas is the influential role of susceptibility, which can take the form of hyperresponsiveness or loss of reserve. What is a minor reversible effect in the majority of individuals may be a dysfunction that cannot be reversed or compensated in certain individuals (Fig. 29-6). Obvious examples would be cardiopulmonary-compromised individuals who function with little or no reserve. In the end, however, the 2000 ATS statement acknowledges the limits of definitions and the importance of value judgment in the final assessment. Implied in this position is that a loss in quality of life due to air pollution as well as enduring its associated effects may also be designated as adverse. As science continues to advance, especially in the realm of molecular biology where small signals can be detected that may forecast an adverse effect or otherwise may identify individuals or groups at risk, the definition of adverse will certainly need reexamination. That same sensitivity in measurement that serves to predict an adverse effect must be separated from signals that are essential for homeostasis and the maintenance of life. Clearly, dissecting and defining these phenomena will have implications not only for assessing clinical adversity but also for predictive toxicology.

In actuality, there is no widely accepted definition for a “susceptible” individual and quite frequently the term is used interchangeably with “vulnerable.” The USEPA, however, separates the two terms and utilizes “vulnerability” to refer to extrinsic (ie, nonbiological) factors (eg, an increased exposure to ambient air pollutants because one’s school is located adjacent to a high-traffic-volume roadway), whereas “susceptibility” refers to intrinsic biological factors such as genetics, age, or preexisting disease. Regardless of the precise definition, a susceptible subpopulation encompasses both intrinsic and extrinsic factors that increase the response to air pollution. As demonstrated in both toxicological and epidemiological studies, a combination of factors influences the incidence and magnitude of adverse responses of susceptible subpopulations to inhaled pollutants (Table 29-3). For example, over the last 20 years, hundreds of publications have documented association between ambient PM levels and increases in cardiopulmonary morbidity and mortality. Age has been consistently shown to influence health outcomes, and the association between ambient air pollution and increases in hospital admissions for asthma is driven by the response seen in children. Conversely, the association between episodic exposure to PM and cardiac mortality is strongest in those over 65 years of age.

Susceptible subpopulations that show exaggerated responsiveness to pollutants merit special mention. The existence of sensitive individuals and groups is well accepted among those who conduct air pollution health assessments, but relatively little is actually known about the host traits that render certain individuals sensitive to inhaled particles and gases. This appreciation for sensitive populations is specifically noted in the CAA, where protection of these groups is mandated in the setting of the NAAQS in the United States. There are some definable subgroups that are considered inherently more susceptible, including children, the elderly, and those with a preexisting disease (eg, asthma, cardiovascular disease, lung disease). The importance of susceptibility in air pollutant responses is gaining more and more attention as test subject responses that were once considered “outliers” in a controlled chamber study may well be evidence of unusual responsiveness.

In some cases, susceptibility may simply reflect differences in dosimetry. Children spend more time outdoors than adults, are more active, and have basal ventilation rates that exceed adults on a volume to body weight ratio, and therefore may experience overall greater dose to the lungs. Certainly, rapidly growing tissues may also factor in and may have contributed, perhaps with dosimetry factors, to epidemiology findings that polluted urban air retards lung growth (Gauderman et al., 2004). Similarly, adult humans and animals with obstructive airway disease, for example, may have “hotspots” of particle deposition in the airways that exceed normal local tissue doses many-fold (Kim and Kang, 1997; Sweeney et al., 1995). Another often underappreciated aspect of susceptibility relates to the loss of functional reserve or compensation due to age or disease, perhaps altering a response threshold or impairing recovery and the reestablishment of homeostasis. Air pollution– induced changes in physiologic measurements such as blood pressure or heart rate may be easily restored in a young healthy individual but not in a more fragile older person with cardiovascular disease.

As already noted, there is also growing interest in gene–environment interactions where genetic differences can influence responsiveness. For example, because the variability in response to ozone was much greater among human subjects than within subjects exposed to ozone in a controlled clinical exposure setting, research in the 1990s suggested that genetic factors influence the response to ozone (McDonnell, 1991). This observation was preceded by an animal study in which significant interstrain differences were found in inbred mice exposed to high levels of ozone (Goldstein et al., 1973). Since these early studies, human (controlled exposure and epidemiology studies) and animal studies have clearly shown that the response to an inhaled pollutant is dependent on genetics. In humans, GSTM1 polymorphisms affect ~50% of the population and, thus, any defect in gene function may undermine antioxidant defenses in those individuals (McCunney, 2005). A number of studies have shown that the null (deletion) GSTM1 genotype, in comparison to the normal wild-type allele, is associated with a greater response to ozone and that the observed pulmonary function changes are even larger if other gene variants involved in oxidative stress are present. Unfortunately, the majority of human studies have examined the contribution of only a handful of gene variants at a time. Current state-of-the-art techniques, however, will now permit the study of the individual and interactive role of numerous (and someday all) gene polymorphisms in the response to inhaled pollutants. Animal studies have advanced from classical quantitative linkage analysis studies of single sensitive and resistant inbred mouse strains to haplotype mapping studies that take advantage of the complete sequencing or deep SNP analysis of multiple inbred mouse strains. Similarly, in human studies, we now have the capability to perform genome-wide association (GWA) studies that link up to one million SNPs or haplotypes with a given phenotype/disease in large (ie, tens of thousands) population studies. Not surprisingly, these GWA studies have determined that individual gene polymorphisms contribute to only a small fraction of the variability in response or disease. Thus, at present, genetic susceptibility, as well as overall susceptibility, to inhaled pollutants is a complex concept that bears significantly on any response. Individual variations in response to ambient air pollution may reflect differences in dose, sensitivity, and/or compensation. As a result, outcomes among individuals may be similar, but the balance of contributing factors that underlie the individual response may be multifactorial.

The study of adverse effects in human volunteers undergoing controlled exposure to air pollutants is ethically limited, but it is even more so for compromised human subjects. Such studies of susceptibility are confined to subjects of only modest suspected risk such as mild asthmatics, healthy older adults, and asymptomatic cardiopulmonary subjects. Inroads have nevertheless been made in recent years because of more thorough prestudy assessments of potential risk factors, allowing researchers to design studies that need not carry undue risk of serious complications. New molecular methods also allow more sensitive assessment of putative biomarkers of response. Additionally, the development of more appropriate animal models of disease or dysfunction (eg, obese or diabetic mouse models) provides a useful adjunct to explore susceptibility factors more extensively prior to study in humans. Hence, coordinated studies in animals and human subjects are proving to be useful to investigate specific questions such as the roles of preexisting disease, diet (eg, antioxidant content), exercise (as it relates to dosimetry), age, gender, and race. The goal of such susceptibility studies is to elucidate patterns, common factors, or pathways that may inform potential intervention or mitigation strategies as well as basic information to reduce the uncertainties regarding risk factors (Kodavanti et al., 1998).

To augment genome studies in humans, recent advances in molecular biology allow the assessment of phenotypic traits in animal models tied to identifiable genes that are homologous to humans. Natural variants in mouse genetics and specially bioengineered transgenic and knockout strains (and in some cases rats) are now widely used to explore the mechanistic basis of phenotypic responses to air pollutants. These new biological tools hold great promise in better understanding responses and establishing gene–environment interactions that may underlie variation in human responsiveness. Transgenic strains can be devised to express desired gene-linked traits using genes inserted from humans as well as other animal models, while knockout models with specific genes removed or silenced can be made devoid of specific traits to isolate responsiveness mechanisms to a toxic challenge. These engineered animal models add to the availability of natural mutants that have been inbred historically to “fix” a desired genotype with a specific phenotype expression (Glasser and Nogee, 2006; Shapiro, 2007). Current technology can also target specific genes for isolated expression in the lung (eg, if linked to surfactant protein C), and in some cases controlled by genes that an investigator can switch on or off using a pharmacologic or chemical prechallenge. These advances allow the dissection of underlying mechanisms under very controlled scenarios and avoid the problems of having a gene be inappropriately active or inactive through all life stages or throughout all body tissues.

To date, the emphasis of studies using these genetically modified animal models has been on mechanisms associated with disease pathogenesis (Suga et al., 2000; Yoshida and Whitsett, 2006). Among the most popular uses of knockout and transgenic mice has been in the study of inflammatory cytokines and associated products in asthma, where the expression of specific mediators is thought to be under the control of single genes. Analogous models derived to exhibit a desired pathology or disease stemming from a genetic defect—for example, involving lung structure or growth (eg, cystic fibrosis or emphysema)—may serve as a surrogate of the human condition (eg, O’Donnell et al., 1999). These models are not limited to the lung but include other risk factors to study such provocative associations between air pollution and atherosclerosis and cardiac disease (eg, ApoE^−/−; Sun et al., 2005; Chen et al., 2010).

The use of genetically modified animal models in air pollution research has generally lagged behind that of basic science and toxicology in general. The reasons for this are likely many, including practicality and expenses, but may relate to the difficulties in incorporating such data into conventional risk assessment paradigms. However, with recent interest in susceptible groups, there has been a definitive upswing in the use of pharmacologically or naturally altered, as well as bioengineered, animal models to more closely link mechanistic profiles to basic human biology. Ozone has been a common test pollutant in these new studies, since more is known about O₃ and its effects in humans than about any other air pollutant. Frequently, these studies address aspects of inflammation and antioxidant capacity relative to challenge by O₃ and other oxidants (Kleeberger et al., 2000; Shore et al., 2009). But with the current interest in PM health effects, these and other models are being redirected. Examples include strain differences and acid-coated PM (Ohtsuka et al., 2000), and hypertransferrinemic mice and metal-rich PM (Ghio et al., 2000). Among rats, the spontaneously hypertensive rat (SHR) and its variants have gained considerable popularity since hypertension is thought to be a human risk factor and because this rat model resembles human hypertension with serum oxidant imbalances, heart disease, as well as its sensitivity to lung injury from inhalants (Kodavanti et al., 2005). The curious are directed to the rapidly evolving literature in this area of research.

Air Pollution: Sources and Personal Exposure

Six major air pollutants (PM, O₃, NO_x, SO₂, CO, and Pb) are considered ubiquitous to industrialized communities and are thought to carry the greatest risk to human and environmental health. With the exception of O₃, these pollutants are emitted by anthropogenic combustion processes along with the myriad of special chemical compounds (mostly volatile organic compounds [VOCs]) considered under the category of HAPs. There are many natural sources of air pollutants as well (eg, volcanoes, wildfires, windblown dust, natural biogenic vapors) but it is the anthropogenic sources that emit pollutants that concentrate where people live that raise concerns about their potential health impacts. These factors do not dismiss the significance of potential risks posed by the natural emissions but put focus on the potential for human exposure and risk.

Great strides have been made in reducing the emissions of some pollutants. Airborne Pb is down >98% in the United States since it was removed from motor vehicle gasoline in 1973. Similar measures were taken throughout Europe in the 1970s and 1980s. The United Nations Environment Program has been actively advising all countries in the world to use only unleaded gasoline and currently only a small number of countries have not restricted the use of Pb in gasoline. What remains in ambient air emanates from a few isolated stationary sources and re-entrained Pb in road dust. Of the other five criteria pollutants, overall emissions have been cut 63% since 1970, while at the same time the US Gross National Product (GNP) has increased 204%. Energy consumption has increased 25%, and vehicle miles driven have increased 168% (National Emissions Inventory [NEI] Air Pollutant Emissions Trends, 2011 {http://www.epa.gov/ttnchie1/trends/}]). The changing profile of pollutant emissions since 1900 is reflected in Fig. 29-7. Since 2001, air quality has continued to improve with reductions in ambient SO₂ being largest (−50%) (http://epa.gov/airtrends/2011/report/ozone.pdf). O₃, which is formed in the air from emitted precursors of both natural and man-made sources, has changed the least (~−13%). Obviously, for any specific locality, air quality can vary depending on the emission profiles of local sources, geographic topography, and meteorology. In the vicinity of a smelter, for example, SO₂, metals, and/or PM may dominate the pollutant profile, while a refinery air shed might be dominated by VOCs and other carbonaceous products. In suburban areas, where the automobile is the main source of pollution, CO, VOCs, and NO₂ would prevail along with their primary photochemical product, O₃. NO_x releases (including NO₂) by stationary sources also contribute to the local O₃ levels along with that derived secondarily from auto emissions. In all, in 2010 about 124 million people in the United States lived in counties that have violations of the NAAQS designed to minimize risks to the criteria pollutants. However, it is clear that with continued population growth and demands on energy and transportation, efficiencies continue to reduce pollutant emissions substantially—with CO₂ reductions being the next great challenge (Fig. 29-8).

Assessing exposure to an air pollutant has long rested on observational measures of what is in the air. Indeed, much of epidemiology as will be discussed below rests on the principle that what is in the air reflects what an individual is actually exposed to and ultimately inhales. Exposure science is advancing rapidly and now utilizes approaches that range from novel statistical treatments of traditional exposure metrics to sophisticated models that systematically involve aerodynamic and microenvironmental characteristics to estimate or predict exposures to individuals or populations. The treatment of this subject is well beyond the scope of this chapter (Zou et al., 2009), but the concept is critical to any toxicological evaluation of air pollutants. It also bears heavily on the appropriate approach to conducting inhalation studies in animal models to study air pollutants. Typically, inhalation studies of animals involve square-wave exposure patterns, although it is well appreciated that human exposures vary spatially and temporally. Moreover, animal studies are generally done during the workday (daytime), which is the nocturnal sleep time of most rodent models. In the latter case, one might expect different ventilation patterns in sleeping rodents as compared with those in humans in their daytime activities and perhaps inverted diurnal biochemical activities associated with feeding cycles. Such differences may not have significant impacts on animal toxicological results but the potential for differences needs to be appreciated and respected in translational or interpretational activities.

Indoor versus Outdoor

People in the United States (and in most industrialized nations) spend in excess of 80% of their time indoors while at work, school, and home or between these places in an automobile (Robinson and Nelson, 1998). Generally, the time spent indoors is disproportionately higher for adults, who have relatively less time to participate in outdoor activities, especially during the day, when outdoor pollutants are usually at their highest levels. The increased outdoor activity pattern for children is changing and potentially has consequences such as the increased risk of obesity and other health/growth parameters. Perhaps the most significant risk factor in indoor air for children is the presence of asthma. With nearly 10% of children with asthma, the risk of exposure to pollutants and allergens in concentrated form is particularly great.

As to the issue of total exposure, children and outdoor workers are thought to be more likely to encounter outdoor air pollution at its worst; in fact, the relatively high physical activity levels of these subgroups leads to larger doses of any given pollutant being delivered to the lungs. Thus, while it is important to characterize and track pollution levels in outdoor air, a realistic exposure estimate is perhaps best achieved when some indoor/outdoor proportionality can be included as well as some indicator of physical activity. From these data, it would be possible to estimate lung dose, although uncertainties can increase with each increment of personal detail. Defining personal exposure can be extremely difficult, as personal monitoring is tedious and expensive, and can sometimes be confounded by other contributions to the indicator being monitored. Hence, exposure measures are typically drawn from ambient measurements or derived from models developed from studies of groups of people carefully characterized across personal exposure modifiers—exercise, personal lifestyles, etc.

The indoor environment has gained appreciation as a major contributor to total personal exposure. The energy crisis of the 1970s spurred efforts to increase home and building insulation, reduce infiltration of outside air, and minimize energy consumption. At the same time, indoor sources of air contaminants have been on the rise from household products and furnishings, which—when combined with poorly ventilated heating systems and overall reductions in air exchange rates—give rise to potentially unhealthy indoor air environments. As people begin to notice patterns of odors, microbiological growth, and even ill health, measures of indoor air become a significant part of environmental risk assessment. Personal exposure has, therefore, come to include the myriad of potential sources, both outdoors and indoors.

It is clear now that indoor air can at times be more complex than outdoor air, a point often raised in challenging the legitimacy of studies that rely solely on outdoor monitoring data. Nevertheless, the national monitoring network for the criteria pollutants has been shown to reflect human exposure reasonably well for some pollutants, especially those that are nonreactive. Indeed, outdoor air permeates the indoor environment in spite of the reduced air exchange in most buildings. However, many variables determine how well components of the outdoor air infiltrate. The current evidence suggests that the average insulated home has about one air change per hour, resulting in indoor concentrations of pollutants that range from 30% to 80% of those outdoors. For nonreactive gases (eg, CO), there could likely be nearly a 1:1 indoor/outdoor ratio in the absence of a “sink” for that gas; the ratio for fine PM (<2.5 μm) could also be fairly high (~0.4–0.7), since these particles can easily penetrate through cracks and open spaces. In contrast, the indoor/outdoor ratio of O₃ would likely be low (<0.3) because of its reactivity. Obviously, household differences in the use of window ventilation and air conditioning can also be important variables. Not to be overlooked, of course, are independent indoor sources of contamination where the ratio of an indoor pollutant to that outdoors can even exceed 1 (eg, NO₂). Particles and gases from tobacco smoke, unvented space heaters, and poorly vented fireplaces and wood stoves as well as fresh paint and cleaning agents can be significant indoor sources. Attention is also being directed toward the many and varied insidious sources of indoor contaminants: certain soils and construction masonry emit radon, gas cooking appliances emit NO_x, sidestream tobacco smoke emits PM, CO, and a host of carcinogenic polyaromatics, and carpets, furnishings, dry-cleaned clothes, and household air fresheners can emit a wide range of VOCs. Responses appear also to be affected by ambient comfort factors such as temperature and humidity. Additionally, some of these chemicals can even interact with one another as has been found to occur with O₃ diffused indoors reacting with VOCs emitted from household cleaners to form small particles. The complexity of these multiple sources underscores the importance of appreciating the total exposure scenario if we are to understand the nature of air pollution and its potential effects on human health (Fig. 29-9).

There remain two broadly defined illnesses that were identified in the 1980s that are largely unique to the indoor building environment (Brooks and Davis, 1992). The first is “sick building syndrome” (SBS), which is a collection of ailments defined by a set of persistent symptoms enduring at least two weeks (Table 29-4) and appears to occur in at least 20% of those exposed. The pattern is complaints by the building inhabitants, typically of unknown specific etiology sometimes even with ambient measurements, but is relieved sometime after an affected individual leaves the offending building (Hayes et al., 1995). Frequently but not always, this syndrome occurs in new, poorly ventilated, or recently refurbished office buildings. The suspected causes include combustion products, cleaning chemicals, biological emissions from mold, and vapor emissions from furnishings frequently exacerbated by discomfort. The perception of irritancy to the eyes, nose, and throat ranks among the predominant symptoms that can become intolerable with repeated exposures. Controlled clinical studies have shown concentration- and duration-dependent worsening of sensory discomfort after exposure to a complex mixture of 22 VOCs commonly found in the indoor environment (Molhave et al., 1986). The many factors contributing to such responses are poorly understood but include various host susceptibility factors such as personal stress and fatigue, diet, and alcohol use. Current biomarkers of response used in the laboratory include sensory irritancy to the eyes in volunteer test subjects and sometimes in animals. However, in general, animal studies using standard measures of sensory irritation or other corporal end points have had limited success in assessing SBS and related syndromes. Human perception factors are believed to be critical to the response patterns.

The second syndrome (building-related illnesses) is a group of illnesses that, in contrast to the SBS, consists of well-documented conditions with defined diagnostic criteria and generally recognizable etiology. These illnesses typically call for conventional medical treatment strategies, because simply exiting the building where the illness was contracted may not readily reverse the symptoms. Several biocontaminant-related illnesses (eg, legionnaires’ disease, hypersensitivity pneumonitis, humidifier fever) fall into this group, as do allergies to animal dander, dust mites, and cockroaches. Medical treatment and mitigation of exposure (source elimination or personal protection) are generally needed to abate symptoms. Some typical outdoor pollutants can also be problematic indoors—most notably, CO from poorly vented heaters. Indoor exposures to NO₂ can pose problems as well, especially in people who may be sensitive—such as asthmatics. At times, however, when the concentrations of CO, NO₂, and many VOCs (passively emitted from new furniture or rugs, or from molds in the ventilation system) result in less discernible or definable signs and symptoms, the responses may be misdiagnosed as SBS, which can complicate assessment of the situation.

Many inhalants, such as NO₂ and trichloroethylene (a VOC common to the indoor air arising from chlorinated water or dry-cleaned clothes), have been shown in animal toxicology studies to suppress immune defenses and allow opportunistic pathogens to proliferate in the lung. The involvement of immunologic suppression is a particularly controversial, yet important attribute of indoor pollution because of its insidious nature and its implications for all building-related illnesses. This problem is further complicated in that complex indoor environments comprising of chemicals and biologicals (dust mites, fungi, molds, etc) may also lead to unexpected interactions that remain virtually unstudied and thus are underappreciated in the assessment of indoor pollution.

Recently, there have been incidents that highlight indoor air pollution. Post Hurricane Katrina (2005) deployment of trailers where people lived for extended periods resulted in exposure to formaldehyde from urea foam insulation with a myriad of effects on inhabitants—notably children. And more recently, wallboard of Chinese origin, which had sulfate and other chemical components on the sealant surface paper that vaporized into the air, resulted in a wide array of symptoms among inhabitants as well as corrosion of some materials. Clearly, the indoor environment is not to be underestimated as to its potential for impact on human health.

Indoor Air in the Developing World

Pollution of the indoor environment in the developing world is a major issue; in some parts of the world this issue actually overrides outdoor air pollution. In the more urbanized environments, there is indoor infiltration of outdoor pollutants unhampered due to poor housing stock. Superimposed on the infiltration of ambient air pollutants are indoor emissions from cooking practices, the cultural use of incense, tobacco, and various other substances, such as perfumes. In less developed communities, unvented or poorly vented cookstoves that burn biomass are used much as they have been for centuries. The emissions from these cookstoves accumulate indoors exposing families, and especially women and children, to soot and account for two million deaths around the world (Fig. 29-10). Chronic lung diseases, such as bronchitis, emphysema, and cancer, are major killers of exposed women while children suffer from bronchitis and various other infectious lung diseases. The cookstove issue is the focus of a UN Foundation called the Global Alliance for Clean Cookstoves (http://www.unfoundation.org/assets/pdf/global-alliance-for-clean-cookstoves-factsheet.pdf), which is a multinational effort to replace existing dirty stoves with clean culturally practical alternative cooking stoves and devices. The black carbon (soot) emitted from the dirty stoves is also believed to be a major contributor to global climate change.

The Evolving Profile of Outdoor Air Pollution

Classically, ambient air pollution was distinguished on the basis of the chemical redox nature of its primary components. Dickens’s 18th-century “London’s particular” referred to a chilled, acrid fog resulting from the accumulation of SO₂ and smoke from incomplete combustion during frequent meteorological inversions. This acidic mix reacts with surfaces, corroding metal and eroding masonry, characteristic of reductive chemistry. These reducing-type atmospheres have long been associated with smelting and related combustion-based industries (as along the Meuse River in 1930 and Donora, Pennsylvania, in 1948) as well as with large, urban centers (eg, London in 1952, New York in 1962) burning coal or other dirty fossil fuels for energy. In contrast, Los Angeles has always had a characteristically “oxidant-type” pollution consisting of NO_x and other secondary photochemical oxidants, such as O₃, aldehydes, and electron-hungry hydrocarbon radicals. By virtue of its regional topography and summer meteorology, Los Angeles is particularly prone to trap traffic emissions, while other areas, such as Atlanta, are simply plagued by stagnant, humid summer air.

The classical types of air pollution were implicitly seasonal. Reducing-type air pollution occurred during winter periods of oil and coal combustion for heating and power coupled with meteorological inversions, while the oxidant atmospheres occurred during the warmer months of spring and summer, when sunlight is most intense and can catalyze reactions among the constituents of auto exhaust. Today the urban distinctions between reducing and oxidant smogs have become largely an academic or historic exercise. Most US urban areas have virtually eliminated smoky, sulfurous emissions (through regulation and a changing national industrial profile) and now with suburban sprawl have a proliferation of automobiles that contribute tons of oxidant precursors into the ambient air. However, in the Midwest there remains still relatively heavy industrial activity that continues to emit SO₂, albeit much less than decades ago, but whose emissions are targeted for further reduction. These emissions continue to have important regional implications as they undergo complex cloud chemistry to form sulfate and impact visibility (regional haze) as well as contribute significantly to PM mass, the primary metric linked to human health effects (Zeger et al., 2008). The air above metropolitan areas in the eastern half of the United States is currently composed of regional reducing pollutants that distribute and blanket a large segment of the country as well as temporally and spatially varied local source pollutants (HAPs) and secondary oxidant pollutants. Increasingly, oxidants and sulfates formed high in the atmosphere are transported across regions to other locales. In the East, sulfates may still predominate over nitrates in the air, in contrast to the Southwest United States, but nitrates occupy a higher ratio of the pollutant mix than ever due in large part to the growth in traffic. No longer is the Northeastern air pollution simply a sulfur-based problem. Adding further to the complexity of the ambient air profile is that not only has the composition of the urban haze changed but also the ambient chemistry has affected its temporal patterns. Long extended periods of O₃ (8–12 hours) now prevail over wide areas such as southern California and the US Southeast and even along the Northeast corridor. The prototypic spike pattern of an hour or two depicted for Los Angeles in the 1970s, and assumed to be the norm for photochemical processes, has changed. In part, the chemical mixes are different as is the distribution of the sources (motor vehicle traffic). As climate change progresses, there is expectation that the underlying chemistry will change further altering these patterns because the assumptions and data that led to current O₃ control measures will be undermined.

Outside the United States, many international megacities remain plagued by the classic reducing and oxidant forms of air pollution. For example, uncontrolled industrial and coal-fired power plant emissions surrounding cities such as Beijing and the northern sectors of Mexico City are dominated by sulfurous, particulate emissions, whereas southern Mexico City, Santiago, and Tokyo have substantially (although not exclusively) automobile-associated oxidant smogs. Impacts on health, visibility, and general welfare are clear and are bringing ever increasing public concern in an all too familiar scenario played out in the United States in the mid-20th century. Urban air pollution is a worldwide problem, where the estimate of people exposed to O₃ at potentially harmful levels exceeds 480 million (Schwela, 1996), with a recent estimate of attributable yearly mortality at 700,000 (Anenberg et al., 2010). The WHO estimates of PM-related mortality are at 1.3 million per year (WHO, 2009).

Outdoor Air Pollution

Acute and Episodic Exposures

A number of air pollution incidents have been documented where pollutant concentrations rose to levels that are clearly hazardous to human health. Where a single chemical has been accidentally released (eg, methyl isocyanate in Bhopal, India), establishing the relationship between cause and ill effect is straightforward. However, most air pollution situations involve complex atmospheres, and establishing a specific cause other than the air pollution incident itself can be difficult. Three acute episodes of community air pollution are considered classic (Meuse Valley, Belgium—1930; Donora, Pennsylvania—1948; and London, United Kingdom—1952). In each event, community inhabitants were clearly affected adversely; hospitalizations were concomitant with an elevated mortality rate. Likewise, each involved a meteorological inversion (cold air capped above by a blanket of warm air, with little or no vertical air mixing) that prevailed for three to five days, during which time the concentration of an array of combustion pollutants rose well above the normal levels for these already heavily polluted areas. No actual measurements of pollution were made in the Meuse Valley and Donora, but crude daily averages of smoke and SO₂ of the London smog incident were recorded—estimated at 4.5 mg/m³ and 1.34 ppm, respectively, on the worst day. Brief (on the order of hours) peaks of these pollutants certainly were much higher. During the Meuse Valley episode, 65 people died, while in Donora the number was 20. These deaths are considered “excess” deaths when compared with normal mortality rates for that time of year that were likely already affected by the background air pollution.

The “London smog” of 1952 was a true disaster with an estimated 4000 excess deaths during the event itself with perhaps thousands more in the weeks to follow. Hospital admissions increased dramatically, mainly among the elderly and those with preexisting cardiac and/or respiratory disease. Even otherwise healthy pedestrians, their visibility limited to as little as 3 ft, covered their noses and mouths in an attempt to minimize their exposure to the acrid air. Many reports of sudden death were reported among workers commuting on bicycle or on foot with symptoms described as “choking.” People with preexisting cardiac and lung problems were particularly affected. It is ironic that 16 years earlier, shortly after the Meuse Valley episode, Firkert (1936) predicted that 3200 deaths would occur should a Meuse-like smog strike London. Although the London incident brought the issue of air pollution to the public consciousness, additional episodes occurred with the 1956 and 1962 incidents being among the most notable. But none were of the magnitude seen in 1952. Interestingly, in December 1991, London experienced a winter inversion smog alert, but one with a very different modern pollutant profile: NO₂ at 423 ppb, black smoke at 148 μg/m³, and SO₂ at 72 ppb (five, four times, and twice the seasonal average, respectively). The differences between this episode and previous episodes were that the air pollution was neither of the magnitude of earlier events nor completely dominated by coal emissions as traffic was at this point a major source of oxidant emissions (Anderson, 1999). Overall increases in mortality and hospital admissions were not so apparent, but regional comparisons suggested impacts among the elderly and cardiopulmonary-impaired (mortality: ↑ 14% cardiovascular and ↑ 22% respiratory; ↑ 43% for respiratory admissions). Susceptible individuals are always the first to respond to these pollution episodes, but the impacts eventually are widespread.

During the latter half of the 20th century, other cities experienced notable air pollution events; among these were New York City, Steubenville, Ohio; Pittsburgh, Pennsylvania; Athens, Greece; and major regions of Western Europe from the Netherlands to the Ruhr Valley of Germany. Major episodes of anthropogenic air pollution continue to decrease in the developed world, in both frequency and intensity, but “natural” events such as the vegetation/forest fires of Indonesia of 1997 and 1998, greater Moscow in 2009, and northern California in 2009 and 2010 all saw significant impacts on regional air quality and public health. The impacts of these fires, like that of the US Northwest in 2003, were tracked as their plumes dispersed across the country.

So much had the air cleared in much of the United States from the late 1960s into the 1970s that many thought that the problem of industrial pollution was on its way to full resolution. Smoke from factory chimneys was less visible and the air epidemiology of the times showed little health impact. However, in the 1980s, it was realized that smoky skies were largely due to carbonaceous material and that more efficient burning had eliminated much of what could be seen in the air. But the sulfur in fuel emitted as sulfur dioxide reacted in clouds to form acidic sulfate and traveled great distances depositing and acidifying lakes and streams and at the same time painting the leaves and the ground resulting in defoliation of forests (Calvert et al., 1985). There were also new studies showing that acid haze with O₃ was associated with emergency room visits among susceptible populations—for example, asthmatics (Bates and Sizto, 1987). A series of studies showed acute effects of regional summer haze in areas of central and Northeastern North America. These two- to three-day episodes of haze were typified by increases in O₃ and acid and neutralized (ammonium) sulfates, characteristic of the new generation of pollution in many US urban areas. Interestingly, the apparent combined temporal or sequential patterns of O₃ and sulfate were associated with the health effects, and neither constituent seemed to be acting alone. Similar results were reported for the upstate New York area as well (Thurston et al., 1992), but it seemed the acidity as [H⁺] of the haze was the active toxicant. Studies of children at summer camps, where children are active and outdoors most of the day, had reported decrements in daily measured pulmonary function on days of haze when both O₃ and acidity levels were elevated. The responses appeared to exceed those predicted by chamber studies of O₃ alone, suggesting dose differences or interactions between haze components (Lippmann, 1989). Clinical studies in adolescent asthmatics and animal studies lent further support to the belief that H⁺ can affect airway function, particularly in the presence of O₃. Studies in the South and Southwest similarly found effects in young asthmatics and diminished performance in athletes, but these appeared to relate more specifically to O₃, since sulfate was less prominent in the ambient air of these regions. Yet these epidemiological and field studies showing associations with “acid” haze could not be replicated in the laboratory with either animal or human exposure studies. There could be any number of reasons for this lack of coherence but the exposure concentrations were often orders of magnitude more than in the ambient air. When refined acid measurements in ambient air were used, the consistency of the findings could not be reliably replicated. With no apparent impact on mortality in epidemiological studies, the focus of air pollution toxicology focused more and more on O₃ where clinical and animal exposure studies easily could measure impacts on lung function. It was felt that O₃ was the likely culprit and its ubiquity in urban air argued for greater and greater attention. Indeed, many health effects could be ascribed to oxidant/irritant properties of O₃.

Of the many air pollution studies over the last 25 years, none have had more impact on the perception of pollutant risk and the direction of research today than a series of epidemiological studies that showed an association between PM mass concentration (as measured by the regulatory monitoring network in the United States) and daily mortality. These studies departed from conventional epidemiological methods used for air pollution studies in the United States and utilized novel time-series analyses that could more easily blunt the impact of weather, smoking, and other variables that might obscure patterns in health variables linked to the air monitoring data. These studies showed significant and consistent associations between health outcomes of ambient PM at levels previously thought to be safe. Prior to this period, measurable effects of PM and SO₂ were not easily detected below the 24-hour means for smoke and SO₂—250 μg/m³ and 0.19 ppm, respectively. In fact, the new findings showed effects, as evidenced by increases in mortality and morbidity rates at or below contemporaneous NAAQS levels (1987: 50 μg/m³, annual mean; 150 μg/m³ daily maximum for PM of diameter <10 μm [PM₁₀]). Time-series analyses are based on Poisson regression modeling to distinguish changes in daily death counts (or hospital admissions) associated with short-term changes in air pollution. The studies initially found effects with cruder measures of PM—total suspended particulate material (TSP), which includes virtually all particles <100 μm in mass median aerodynamic diameter (MMAD—a median particle size normalizing the particle to unit density and spherical shape for aerodynamic comparison). These studies were followed with stronger associations with particles considered inhalable—PM₁₀ (an MMAD at which PM is aerodynamically separated at an initial 50% efficiency at 10 μm and increasingly at smaller sizes). Beginning in the mid-1990s, even stronger associations have been found with an analogous but fully and deeply respirable particle—PM_2.5. As the PM diameter gets smaller, it better represents anthropogenic sources of pollution. The statistical methodology applied in these time-series studies had an advantage over conventional regression analyses in that it could detect short-term trends and minimized the effects of other pollutants and potential confounders with longer time constants (Schwartz, 1991; reviewed by Pope and Dockery, 2006).

In contrast to the three epidemiological studies used by the US EPA to develop the 1987 PM₁₀ NAAQS, there were more than 30 such studies for the 1997 revision and the promulgation of the new index, PM_2.5 NAAQS ([15 μg/m³, annual mean; 65 μg/m³ daily maximum for PM of diameter <2.5 μm [PM_2.5]). These studies showed a significant health impact of PM, linked to mass and not necessarily sulfate or any other constituent. In 2006, the PM NAAQS was revised keeping the annual PM_2.5 NAAQS at 15 μg/m³, with a tightening of the daily maximum to 35 μg/m³. The NAAQS for PM is reviewed about every five years depending on legal or procedural issues as data accrue; an update on the primary (human) standards for PM_2.5 (fine PM) and PM_2.5-10 (coarse PM) is expected in the 2012 time frame. The PM NAAQS is specifically noted since of all the NAAQS, it has the greatest impact on human health of any US regulation in terms of both cost and benefit. As a result, the health science (epidemiology, clinical human exposure and animal toxicology studies) is scrutinized more than any other US regulatory decision.

PM stands as the preeminent air pollutant because of its health impact as well as the pollutant that opened the door to unsuspected targets of injury. The major health outcome revealed in the study of PM has been the involvement of the cardiovascular system as a prime target for adverse impact, and not the lung as historically assumed. While the heart, as part of the cardiopulmonary system, has always held an indirect role in health impacts or disease from air pollution, both epidemiological and toxicological studies now point to major cardiac involvement in PM-associated mortality. Current thinking is that cardiac-mediated effects are more germane to the PM-associated mortality findings than pulmonary effects (Brook et al., 2010). Not surprisingly, effects are most apparent in subpopulations already compromised by cardiopulmonary and perhaps vascular diseases (eg, diabetes). There exists no one generally accepted mechanism or mode of action to account for these findings (Costa, 2000; Brook et al., 2010). However, several pathways have been proposed that attempt to link exposure and cardiac effects that may or may not include pulmonary mediation. These potential mechanisms are illustrated in Fig. 29-11.

From a toxicological perspective, it has been difficult to accept that the association between PM and health outcomes is not somehow linked to particle composition rather than mass alone. The linkage to mass and not composition is counterintuitive, especially in light of the fact that all PM is not constitutively identical. The actual “biochemical lesion” caused by PM is generally thought to involve oxidant mechanisms (generation of reactive oxygen and perhaps nitrogen species) by constituents or attributes (eg, reactive surface area) of the particles at the cell or molecular level. Initially, there was a credibility hurdle (“biologically plausibility”) that day-to-day fluctuations in the mass concentration of ~10 μg/m³ airborne PM would result in an increase of about 0.6% to 1% (excess) mortality. However, there now exist several plausible hypotheses that have some degree of empirical support: metals, organics, nanosize with reactive surfaces, and constitutive oxidants to name those most prominent.

The experience with PM has expanded the thinking of the possible extrapulmonary effects of other air pollutants. Potential links to cardiovascular toxicity of several of the NAAQS pollutants have been demonstrated, although not with the consistency of PM. Moreover, there are new data showing pollutant-related impacts on birth outcomes, postnatal development, and neurologic health. In the latter case, it is unclear if these effects are related to vascular impacts such as those caused by PM or to other systemic oxidant intermediates.

Long-Term Exposures

Epidemiological studies of the chronic effects of air pollution are difficult to conduct by the very nature of the goal: outcomes associated with long-term exposures. Looking back in time with retrospective, cross-sectional studies was the common approach, but those studies frequently were confounded with unknown variables and inadequate historical exposure data. A good example of the problem of confounding is cigarette smoking. Without extensive information on both active and passive smoking, the ability to discern the impact of an air pollution disease outcome such as chronic bronchitis and emphysema would be greatly impaired. Prospective studies, on the other hand, have the advantage of more precise control of confounding variables, such as the tracking of urinary cotinine as an index of tobacco smoke exposure, but they can be very expensive and require substantial time and dedication on the part of both the investigators and the study population. Depending on the study size and design, exposure assessments can be complex, and the loss of subjects due to dropout is sometimes unpredictable.

Despite these deficiencies and problems, there are both retrospective and prospective epidemiology studies that have tackled the issue of long-term air pollution health effects. In general, these studies have suggested a positive association between urban pollution and progressive pulmonary impairments. On the one hand, cross-sectional studies in the Los Angeles Air Basin have found evidence of accelerated “aging-like” loss of lung function in people living for extended periods in regions of high oxidant pollution, when compared with that in people living in areas where sea air circulation lowers overall pollutant concentrations (Detels et al., 1991). Similarly, exposure to SO₂ and PM in the Netherlands over a 12-year period was shown prospectively to gradually impair lung function (Van De Lende et al., 1981). And even rural areas in western Pennsylvania, which are swept by reducing-type pollutants transported from Midwestern industrial centers, have been shown to have a higher incidence of respiratory symptoms as determined from a questionnaire-based design (Schenker et al., 1983). While a negative impact of any specific pollutant followed in these studies is difficult to dissect, the message that long-term exposure to air pollution contributes to deterioration of lung health seems clear.

Among the most detailed prospective epidemiological studies of chronic health effects across a range of levels of air pollution has been the so-called Harvard Six Cities Study begun in the early 1970s. The cities were chosen to represent a range of air quality based on SO₂ and PM—it was designed as a study of “acid” air pollution. Early on in the study, there was great dependence on routine regional air monitoring data, but over time the investigators themselves conducted the air analyses of exposure microenvironments (local and indoor). The initial design of these studies included the gathering of parental questionnaire data (including approximately 20,000 people) about the prevalence of respiratory problems in school children and has continued over 30 years along with periodic assessments of pulmonary function. When compared across cities, [H⁺] (measured in four of the six cities) was correlated (Fig. 29-12D) better than was sulfate with the prevalence of bronchitis in children aged 10 to 12 (Speizer, 1989). However, as the assessment program evolved, more detailed study revealed mortality associations with PM as represented in Fig. 29-12A-C. The role of [H⁺] in relation with acute mortality was less convincing than that associated with the sulfate or PM_2.5 (sulfates co-associate with fine PM in the atmosphere) (Dockery et al., 1993). More importantly with regard to long-term health, this study showed very significant effects of PM on the life spans of people living in Steubenville, Ohio—the dirtiest of the industrial centers. Over a 15-year period, the average human life span was reduced by about two years due to PM exposure. These findings were corroborated by a prospective cohort-based mortality study using a database collected on over 500,000 people by the American Cancer Society (ACS) across 151 cities from 1982 to 1989 (Pope et al., 1995). The study showed a 15% to 17% increased mortality risk over seven years due to PM, about equivalent to the risk of smoking over that period. Subsequently, Pope et al. (2002) reported results of a 10-year and again an 18-year follow-up study of the ACS cohort (Pope et al., 2011). These updated analyses, which included gaseous copollutant and new fine particle measurements, and more comprehensive personal information on the enrollees, confirmed the linkage of health effects, including an increased risk of lung cancer, with PM_2.5 exposures. These findings reinforce the concerns for potential chronic health impacts of PM and the heightened risk of premature death from lifelong air pollution exposure. There is currently a 10-year study (results expected ~2015) to assess the impact of air pollution, with emphasis on PM, on atherosclerosis in humans as part of the Multi-Ethnic Study of Atherosclerosis Air Pollution (MESA AIR; http://depts.washington.edu/mesaair/).

Classic Reducing-Type Air Pollution

The acute air pollution episodes of the 20th century showed that high concentrations of the reducing-type air pollution, characterized by SO₂ and smoke, are capable of producing dramatic human health effects. Empirical studies in human subjects and animals have long stressed the irritancy of SO₂ and its role in these incidents, while the full potential for interactions among the copollutants in the smoky, sulfurous mix has a mixed record of replication in the human exposure laboratory. Because of its history, SO₂ is often thought of as the prototypic air pollutant of the industrial age, and as such it has a detailed historical literature base. It is an irritant gas that has a toxicology of its own and, through atmospheric reactions, can transform photochemically into sulfites or sulfates within a secondarily irritant particle. The impact of SO₂ emissions on irritancy, the formation of secondary particles that affect visibility, and its damaging acidification of the environment have been the focus of much regulation for coal-fired power plants since 1970. Levels are greatly reduced (>60%) in the United States and continue to fall under new cross-state regulations.

Sulfur Dioxide

General Toxicology

Sulfur dioxide is a water-soluble irritant gas. As such, it is absorbed predominantly in the upper airways. It is a sensory irritant and can stimulate bronchoconstriction and mucus secretion in a number of species, including humans. Early studies with relatively high exposure concentrations of SO₂ showed airway cellular injury and subsequent proliferation of mucus-secreting goblet cells. This attribute of SO₂ has led to its use (>250 ppm) in the production of laboratory animal models of bronchitis and airway injury (Kodavanti et al., 2000). At much lower concentrations (<1 ppm), such as might be encountered in the polluted ambient air of industrialized areas, long-term residents experience a higher incidence of bronchitis. In fact, prior to the breakup of the Soviet block, many eastern European cities were renowned for widespread public affliction with bronchitis; 20 years later, the prevalence of bronchitis was greatly reduced (von Mutius et al., 1994). While other factors (diet, access to health care, other pollutants) may well have been involved in this reversal, reductions in ambient smoke and SO₂ are generally thought to be the most important.

The concentrations of SO₂ likely to be encountered in the United States are lower still—on average, considerably less than 0.01 ppm. Rats exposed for 70 to 170 hours to 0.1, 1.0, and 20 ppm exhibited reduced clearance of inert particles, while dogs exposed to 1 ppm for a year had slowed tracheal mucociliary transport. Analogously, mouse models show that mice exposed to 0.1 ppm were more susceptible to bacterial infections. The fact that the low-concentration exposures showed marked effects when extended over longer periods is consistent with the epidemiological associations between SO₂ exposure and bronchitis. The evidence is not clear, however, as some studies show no overt long-term pulmonary pathology. Guinea pigs and monkeys, for example, showed no effect on lung function or pathology after a year of continuous exposure to concentrations of 0.1 to 5 ppm SO₂ (Alarie et al., 1970, 1972).

The penetration of SO₂ into the lungs is greater during mouth as opposed to nose breathing. An increase in the airflow during deep rapid breathing augments penetration of the gas into the deeper lung. As a result, persons exercising would inhale more SO₂ and, as noted with asthmatics, are likely to experience greater irritation. Once deposited along the airway, SO₂ dissolves into surface lining fluid as sulfite or bisulfite and is readily distributed throughout the body. It is thought that the sulfite interacts with sensory receptors in the airways to initiate local and centrally mediated bronchoconstriction.

Pulmonary Function Effects

The basic pulmonary response to inhaled SO₂ is mild bronchoconstriction, which is reflected as a measurable increase in airflow resistance due to narrowing of the airways. Concentration-related increases in resistance have been observed in guinea pigs, dogs, and cats as well as humans. Exposure of isolated segments of the nose or airways of dogs and guinea pigs appeared to alter resistance in a manner consistent with receptor-mediated sensory stimulation. Airflow resistance increased more when the gas was introduced through a tracheal cannula than via the nose, since nasal scrubbing of the water-soluble gas was bypassed. Isolated nasal exposures increased airflow resistance through the nose largely as a result of mucosal swelling, but the irritant effect appeared to signal the more distal airways as well. Direct exposure of the trachea had a more dramatic effect on airflow resistance. Exposure of the intact nose also induced some response, consistent with the existence of a nasal neural network being involved in bronchoconstriction (Frank and Speizer, 1965; Nadel et al., 1965). Intravenous injection of atropine (a parasympathetic receptor blocker) or cooling of the cervical vagal nerves abolishes bronchoconstriction in the cat model; rewarming of the nerve reestablishes the response. The rapidity of the response and its reversal emphasize the parasympathetic tonal change in airway smooth muscle. Studies in human subjects have confirmed the predominance of parasympathetic mediation, but histamine from inflammatory cells may play a secondary role in the bronchoconstrictive responses of asthmatics.

Human subjects exposed to 1, 5, or 13 ppm SO₂ for just 10 minutes exhibit a rapid bronchoconstrictive response, with 1 to 3 ppm being a threshold for most if exercise is involved. Exposures to 0.25 to 1 ppm for a few hours can induce bronchoconstriction in adult and adolescent subjects with clinically defined mild asthma (Sheppard et al., 1981; Koenig et al., 1981). Findings such as these (responses <0.5 ppm) have raised concerns about potential adverse effects in this sensitive subpopulation when it is exposed to peaks of SO₂ that are known to occur near point sources.

Chronic Effects

Only a few long-term studies have been conducted with SO₂ at levels approaching those found in ambient air. Alarie et al. (1970) exposed guinea pigs to 0.1, 1.0, or 5.7 ppm SO₂ continuously for a year without adverse impact on lung mechanics. Similarly, monkeys exhibited no alteration in pulmonary function when exposed continuously for 78 weeks to 0.1, 0.6, and 1.3 ppm SO₂ (Alarie et al., 1972). Even in the presence of 0.1 mg/m³ sulfuric acid, dogs exposed 16 hours a day for 18 months to 0.5 ppm SO₂ showed no impairment in pulmonary function (Vaughan et al., 1969). Higher levels of SO₂ for protracted periods of time (dogs to 5 ppm for 225 days [Lewis et al., 1969]; rats to 350 ppm for 30 days [Reid, 1963]) have been shown to alter airway mucus secretion, goblet cell topography, or lung function, but these results are of little relevance to typical SO₂ levels in ambient air.

Sulfuric Acid and Related Sulfates

The conversion of SO₂ to sulfate is favored in the environment with subsequent ammonia neutralization to ammonium sulfate [(NH₄)₂SO₄] or ammonium bisulfate [NH₄HSO₄]. During oil and coal combustion or the smelting of metal ores, sulfuric acid condenses downstream of the combustion processes with available metal ions and water vapor to form submicrometer sulfuric acid fume and sulfated fly ash. Sulfur dioxide continues to oxidize to sulfate within dispersing smokestack plumes, which can be augmented by the presence of free soluble or partially coordinated transition metals such as iron, manganese, and vanadium within the effluent ash. When coal is burned, the acid may adsorb to the surface or solubilize in ultrafine (<0.1 μm) metal oxide particles during emission. Photochemical reactions also promote acid sulfate formation via both metal-dependent and independent mechanisms, but studies have shown that most of the oxidation of SO₂ occurs within plumes as they disperse in the atmosphere. Stack emissions may undergo long-range transport to areas distant from the emission source, allowing considerable time for sunlight-driven chemistry. Although the fine particle sulfates may exist as fine sulfuric acid (the primary source of free H⁺), partially or fully neutralized forms of sulfate predominate due to the abundance of natural atmospheric ammonia. As fine PM sulfates are transported long distances, they may contribute to regional summer haze and pose a health hazard to certain groups such as asthmatics (Koenig et al., 1989), and may also stress the general environment as acid rain (Calvert et al., 1985) (Fig. 29-13). The CAA and subsequent regulations for the reduction of sulfur emissions (eg, the 1990 Acid Rain Program and the 2012 Cross State Air Pollution Rule) continue to reduce the environmental acidification and visibility problems associated with transport and deposition of acidic products formed from stack emissions in North America.

General Toxicology

Sulfuric acid irritates by virtue of its ability to protonate (H⁺) receptor ligands and other biomolecules. This action can either directly damage membranes or activate sensory reflexes that initiate inflammation. Ammonia, which exists in free air at about 25 ppb and in much higher concentrations within the mammalian naso-oropharynx (in the human up to 350 ppm), is capable of neutralizing most of the irritant acidic sulfates (Utell et al., 1989). Neutralization can also be quite efficient in standard whole-body animal exposure studies, in which excreta and bacteria in the chambers interact, giving rise to in-chamber ammonia concentrations up to 1100 ppb—more than enough to fully neutralize neat sulfuric acid up to several mg/m³ (Higuchi and Davies, 1993).

Interestingly, there is considerable species variability in sensitivity to sulfuric acid, with guinea pigs being quite responsive to acid sulfates, in contrast to rats, which seem generally resistant. The reasons for this difference relate to sensory fiber network density in the airways, and probably not on differences in neutralization by ammonia in the airway. The sensitivity of healthy humans appears to fall somewhere in between, with asthmatic humans being perhaps best modeled by the guinea pig. Overall, however, the collective data involving animals and humans are remarkably coherent (Amdur, 1989).

Unlike other irritants, such as O₃ (see below), inhaled sulfuric acid does not appear to stimulate a classic neutrophilic lung inflammation. Rather, eicosanoid homeostasis appears to be disturbed resulting in macrophage dysfunction and altered host defense.

Pulmonary Function Effects

Sulfuric acid produces an increase in flow resistance in guinea pigs due to reflex airway narrowing, or bronchoconstriction, which impedes the flow of air into and out of the lungs. This response might be thought of as a defensive measure to limit the inhalation of air containing noxious gases, but this explanation may be more teleological than fact. The magnitude of the response is related to both acid concentration and particle size (Amdur et al., 1978). Early studies indicated that as particle size was reduced from 7 μm to the submicrometer range, the concentration of sulfuric acid necessary to induce a response and the time to the onset of the response fell significantly. With large particles, even the sensitive guinea pig was able to withstand an exceedingly high (30 mg/m³) challenge with little change in pulmonary resistance, in contrast to the <1 mg/m³ challenge needed with the 0.3-μm particles (Amdur et al., 1978). Human asthmatics exposed to 2 mg/m³ of acid fog (10 μm) for one hour, a very high concentration for an asthmatic, experienced variable respiratory symptoms suggesting irritation, but no changes in spirometry were elicited (Hackney et al., 1989). The apparent reason for this PM size–based differential is probably the scrubbing of large particles in the nose, while small particles are able to penetrate deep into the lung. The thicker mucus blanket of the nose may also blunt (by dilution or neutralization by mucus buffers) much of the irritancy of the deposited acid. In contrast, the less shielded distal airway tissues, with higher receptor density, would be expected to be more sensitive to the acid particles reaching that area (Costa and Schlegele, 1998). Regional sensitivity and the longer residence times of a deposited particle relative to SO₂ gas are reflected in the relatively protracted recovery times observed in acid-exposed guinea pigs compared with those in animals exposed to SO₂ alone.

Asthmatics appear to be somewhat more sensitive to the bronchoconstrictive effects of sulfuric acid than are healthy individuals, but published studies have been inconsistent in this finding (Koenig et al., 1989; Utell et al., 1984). Asthma generally is characterized by hyperresponsive airways, so their tendency to constrict at low acid concentrations would be expected, just as asthmatic airways are sensitive to nonspecific airway smooth muscle agonists (eg, carbachol, histamine, exercise). The variability may well relate to differences in the degree of impairment or underlying inflammation in the subjects, but this hypothesis remains to be confirmed. Airway hyperreactivity has been observed as an acute response in guinea pigs two hours after a one-hour exposure to 200 μg/m³ sulfuric acid and appears to be associated with pulmonary inflammation. Likewise in rabbits, increased airway reactivity was associated with arachidonate metabolites, products of both epithelial and inflammatory cells. The general correlation between airway responsiveness and inflammation that appears to be important in grading asthma severity and risk of negative clinical outcomes may also be predictive of responses to environmental stimuli.

Effects on Mucociliary Clearance and Macrophage Function

Sulfuric acid alters the clearance of particles from the lung. Using insoluble, radioactively labeled ferric oxide particles as a probe, as little as a single one-hour exposure in donkeys, rabbits, and human subjects can slow clearance. Mucus clearance appears to vary directly with the acidity ([H⁺]) of the acid sulfate, with sulfuric acid having the greatest effect and ammonium sulfate the smallest (Schlesinger, 1984). Curiously, there appears to be a biphasic response to acid. In general, brief, single exposures of <250 μg/m³ accelerate clearance, while high concentrations of >1000 μg/m³ clearly depress clearance. With repeated daily exposures to low levels, there appears to be a cumulative (concentration times duration) dose-related depression of clearance. Longer-term exposure of rabbits to low-level acid also results in hyperplasia of airway mucosecretory cells (Gearhart and Schlesinger, 1989).

Collectively, there seems to be coherence in the data to rank sulfate irritancy: sulfuric acid > ammonium bisulfate > ammonium sulfate. Acidity [H⁺] appears to be the primary driver on most respiratory effects attributable to the acid sulfates even at the level of pulmonary macrophages. Lavaged rabbit macrophage phagocytosis was affected more after a single exposure to 500 μg/m³ sulfuric acid than after exposure to 2000 μg/m³ ammonium bisulfate (Schlesinger et al., 1990). Nevertheless, in the complexity of summer haze, it remains unclear whether the bioactive form of [H⁺] is more appropriately assayed as free ion concentration (as pH) or as total available ion concentration (titratable H⁺).

Chronic Effects

Not surprisingly, sulfuric acid induces qualitatively similar effects in the airways as found at high concentrations of SO₂. As a fine aerosol, sulfuric acid deposits deeper along the respiratory tract, and its high specific acidity imparts greater effect on various cells (eg, phagocytes and epithelial cells). Thus, a primary concern with regard to chronic inhalation of acidic aerosols is the potential for bronchitis, since this has been a problem in occupational settings in which employees are exposed to sulfuric acid mists (eg, battery plants). Early studies in the donkey (later confirmed in a rabbit model) have provided fundamental data on this issue. The depression of clearance observed in donkeys exposed repeatedly (100 μg/m³ one hour per day for six months) raises concerns that a similar response (potentially contributing to chronic bronchitis) can occur in humans. Studies with cigarette smoke showed parallel responses.

Studies conducted with sulfuric acid in the rabbit are in general agreement with the early findings in the donkey (Schlesinger, 1984). The initial early stimulation of clearance with subsequent depression has been shown to occur over 12 months with as little as two hours per day at 125 μg/m³ sulfuric acid (Schlesinger et al., 1992). Related studies also have demonstrated that the airways of exposed animals become progressively more sensitive to challenge with acetylcholine, show a progressive decrease in diameter, and experience an increase in the number of secretory cells, especially in the smaller airways. These studies have expanded our knowledge of the biological response and its exposure-based relationship to sulfuric acid. It seems reasonable to postulate that chronic daily exposure of humans to ~100 μg/m³ sulfuric acid may lead to impaired clearance and mild chronic bronchitis. The possibility that chronic irritancy may elicit bronchitis-like disease in susceptible individuals (perhaps over a lifetime or in children because of dose differences) appears to be reasonable.

Particulate Matter

PM was referred to as “soot” in the “reducing-type” air pollution of the classic episodes. The major constituents of this soot consisted of incompletely burned carbonaceous materials, acid sulfates, various metals, and silicates associated with the solid nature of the fuel. Metals were abundant, with a considerable amount of zinc in the form of zinc sulfate—as reported from postepisode analyses of the Donora PM. Soot is indicative of poorly (inefficiently) combusted fuel. Over time, combustion technology and improved fuels provided gains by increasing efficiency and minimizing gross soot emissions. Improvements in combustion methods simultaneously reduced the size of emitted particles and overall less mass. A side benefit of the smaller particles was the reduction in light diffraction through the emissions and hence a less visible plume. As such, much of the early cleanup was largely achieved through technological improvements. Carbon in fuel that is not fully oxidized to CO₂ persists in elemental form or as products of incomplete combustion—organic carbon. Oxidized S becomes sulfate and many of the metals appear as oxides. As noted above, sulfate was long suspected as the culprit of most health impacts associated with stationary sources, but this relationship is less discernable in contemporary particle epidemiology and toxicology.

In the last 20 years PM has reemerged as the dominant issue in the air pollution community, overtaking O₃ as the pressing air pollution health issue. The reason for this shift was the emergence of epidemiology data consistently showing increased mortality, an adverse effect of greatest import with the major impact on the cost/benefit analyses. As already noted, the collective studies showed health impacts (both morbidity and mortality) at levels of PM thought to be “safe” with no apparent threshold (reviewed in Pope and Dockery, 2006). Over time, science advances and reanalyses have confirmed the initial mortality findings. Several hypotheses that lend “biological plausibility” to the findings have emerged along with several “ancillary” impacts of PM on health not previously realized. The impacts of PM on health appear to be ever expanding, beyond lung effects, perhaps with even greater cardiovascular effects and an ever growing body of data suggesting impacts on neurologic and reproductive health, as well as growth and development.

PM in the atmosphere can be solid, liquid, or a combination of both with a mélange of organic, inorganic, and biological compounds. The compositional matrix of PM can vary significantly depending on the emission source and secondary transformations, many of which involve gas to particle conversions. Long-range transport of emissions or transformation products can contribute significantly to the regional matrix of PM, particularly in the case of sulfate of East Coast PM_2.5. Particles of larger size tend to have more local sources, the reason being that they are formed from dispersed dust and attrition of materials. Being of larger size, they tend to “fall out” or settle from the air due to gravity (although winds can in fact carry these particles great distances—eg, Sahara desert particles have been found on the US East Coast). Particles in the range of 10 to 2.5 μm (PM_10-2.5—coarse PM) are highly inhalable by humans. In the urban setting there is considerable spatial and temporal heterogeneity of coarse PM while PM_2.5 appears more homogenous throughout a regional environment. The size designation of fine and coarse PM is based on their relative respirability—those in the range of PM₁₀ are inhalable into the larger thoracic airways while the PM_2.5 is inhalable into the deeper reaches (gas exchange areas) of the lung (see Chap. 15).

The large epidemiological database contending that PM elicits both short- and long-term health effects is largely founded on data from monitoring networks used for PM regulation. As such, the PM-associated effects appear to be dependent on the gravimetric measures that these networks yield. From this, it might then be argued that the effects are not influenced by particle composition (eg, inorganic and organic components). Further, because the findings strengthen with decreasing particle size (eg, total suspended particulate mass [pre-1978 index], PM₁₀, PM_2.5), it is the mass concentration of PM, despite compositional complexity and variation, that is the index of choice. However, toxicologists argue that a mass-based relationship contradicts the basic tenets of conventional air pollution toxicology, which is rooted in the concept of chemical-specific toxicities. A number of hypotheses that draw on various physical and chemical attributes of PM have also been offered in search of a “biologically plausible” explanation for the reported epidemiological observations. However, no one constituent has been identified as singularly determinant of the spectrum of health impacts. The smallest particles, which derive from anthropogenic combustion activities, appear to drive many of the health effects of particles. Prominently included among the causative hypotheses that tie toxicity to particle characteristics are metals, organics, acidity, size distribution (focusing on the unique bioactivity of ultrafine PM—PM <0.1 μm), PM oxidant activity or reactivity, and the presence of potentially toxic or allergenic biologicals. However, at present, there remains insufficient understanding of the relative importance of these theories to choose one over another, especially if the spectrum of heath effects is included. As a result, there seems to be insufficient reason to unseat the PM mass-based correlation with health outcomes as the driver of regulatory policies. Although the laboratory animal and human toxicological database is growing rapidly with regard to the issue of causation, much remains to be learned before new regulatory indices can be considered.

From research directed initially toward potential occupational hazards, it is known that several metals and silicates that make up much of the inorganic phase of PM can be cytotoxic to lung cells. Organic constituents, as well, can induce toxicity either directly or via metabolism product—some of which are genotoxic. Other PM attributes may also come into play. Studies focusing on very small, ultrafine particles suggest that although these particles are low in mass, they are high in number and thus provide substantial reactive particle surface to interact with biological substances. Less is known about the role of biologically derived materials, such as endotoxin, glucans, plant glycoproteins, and bioallergen fragments, which may elicit rudimentary inflammatory responses in the lung. The involvement of biologicals may be greatest in agricultural and indoor exposure environments. There has been considerable interest in PM–copollutant interactions, but our knowledge in this area is somewhat limited and has spurred toxicologists to better access the experience and knowledge of atmospheric scientists as to the atmospheric interactions that are most relevant to biological outcomes.

Metals

There have been many standard acute and subchronic rodent inhalation studies with specific metal compounds, often as oxides, chlorides, or sulfates. These exposure studies relate most appropriately to occupational exposures. The varied systemic toxicities of metal compounds are presented in detail elsewhere in this book; however, it should be appreciated that the effects of metals delivered by inhalation may differ from their impacts when administered by other routes. Metals may arise from natural as well as anthropogenic activities, and as a result metals are a common constituent in ambient PM. The metal profiles among regions differ appreciably in concentration and type and they also differ by the size mode of PM. Coarse PM (2.5–10 μm) arises largely from natural sources and thus has prominent earthen metals such as iron, sodium, silica, and magnesium—usually in oxide forms. Combustion-derived metals reflect the fuel source. For example, oil may have vanadium, nickel, and perhaps zinc and iron, while coal may have zinc and selenium. Their chemical forms vary from water-soluble salts to oxide and phosphate forms. Other metals are emitted from vehicles burning fuels to which metal compounds were added to alter functionality (eg, lead, manganese, platinum) or as engine wear and catalyst by-products. Similarly, metals may also derive from brake (copper, iron), tire (zinc), and dispersed road (earthen silicates) wear. Metals have many biological properties, some essential to life while others being directly toxic to cells or act indirectly in a pro-oxidant toxic fashion. Thus, metals have garnered considerable interest regarding their role in PM toxicity (Costa and Dreher, 1997).

Metal compounds can be separated nominally by physicochemical characteristics: those that are essentially water-insoluble (eg, metal oxides and hydroxides such as those that might be released from high-temperature combustion sources or derived from the geocrustal matrix) and those that are soluble or somewhat soluble in water (often chlorides or sulfates such as those that might form under acidic conditions in a smoke plume or leach from acid-hydrated silicate particles in the atmosphere). Solubility appears to play a role in the toxicity of many inhaled metals by enhancing metal bioavailability (eg, nickel from nickel chloride vs nickel oxide), but insolubility can also be a critical factor in determining toxicity by increasing pulmonary residence time within the lung (eg, insoluble cadmium oxide vs soluble cadmium chloride). Moreover, some metals, either in their soluble forms or when partially coordinated on the surface of silicate or bioorganic materials, can promote electron transfer to form reactive oxidants. Complexes with particulate organic material in a partially hydrated form (as might be promoted by the presence of sea salt) have been shown to interact with poorly soluble metals to free coordination sites that again are pro-oxidant (Kieber et al., 2005). Thus, caution is warranted in assessing inhaled PM-associated metals, as both their chemical and physical attributes and their interaction with cocontaminants in PM may influence their apparent toxicity. Simply measuring total metal mass to estimate effects in the lung can be misleading.

Perhaps the most compelling evidence for a role of soluble metals in PM toxicity derives from a series of studies on PM sampled in Utah Valley where a large open hearth steel mill went from full production to a one-year hiatus due to a strike and then resumed its open hearth operations. These studies related the metal content of PM sampled in the area to plant operation showing a reduction in total PM as well as a reduction of >90% PM metal content during the strike. Studies of hospital admissions for a variety of lung ailments and death rates showed reductions during the strike and toxicological studies where PM extractions from the operational and shutdown periods were tested in human and animal lungs, as well as on human lung epithelial cells, all correlated with plant operations, PM metal content, and the findings in parallel population studies (Pope, 1996; Dye et al., 2001; Ghio, 2004).

Gas–Particle Interactions

As already noted, these gas–particle interactions can be extremely complex involving multiple components of the particles, gases/vapors, and sunlight. However, more than 40 years ago, generic binary interactions between particles and gases in the absence of light were shown to alter the toxicity of either the particle or the gas acting alone. The guinea pig bronchoconstriction model used for many years by Amdur and associates showed early on that SO₂ can interact with hydrated metal salts to potentiate particle irritancy. The mechanism(s) behind this interaction has yet to be fully discerned, but it appears to involve solubility of SO₂ in a hydrated aerosol and the ability of the metal to catalyze the oxidation of the dissolved SO₂ to sulfate. In the case of sodium chloride aerosol, potentiation appeared to be governed primarily by the solubility of SO₂ in the salt droplet and enhanced respiratory penetration. The metal salts of manganese, iron, and vanadium, on the other hand, catalyzed the formation of sulfate. Studies in humans have been less revealing about such interactions, but this database affirms the need to consider the complexity of the atmospheric challenge in estimating biological outcomes.

Complex chemistry also occurs within the effluent of the combustion source. Using a laboratory-scale furnace, the emission mix of sulfuric acid and metal oxide particles common to metal smelting and coal combustion was used to explore potential plume interactions that might impact respiratory irritancy (Amdur et al., 1986). These emitted metal oxide particles, once aged and cooled, were a mixture of singlet and agglomerated ultrafine particles that would be expected to distribute throughout the lung on inhalation. Exposures in guinea pigs of 30 to 60 μg/m³ sulfuric acid combined with ultrafine zinc oxide produced progressive decreases in DL_CO, total lung capacity, and vital capacity and increases in cells, protein, and a variety of enzymes in lavage fluid that were not completely resolved 96 hours after exposure (Amdur, 1989). It is unclear whether the acid was on the surface of the particles or made the metal more soluble, but the combination was clearly more toxic than acid alone. These effects greatly exceeded the changes in airway resistance found with relatively simple, binary mixture of SO₂ and water-soluble metal salts.

Combustion studies using different coals again emphasized the significance of surface-associated acidic S-compounds. Ultrafine combustion particles from Illinois No. 6 coal had a layer of sulfuric acid adsorbed on the surfaces resulting in greater effects in guinea pigs than the more alkaline Montana lignite. Despite the greater sulfur content of the Montana coal, this emission ultrafine particle had neutralized the irritating sulfate (Chen et al., 1990). Similar studies using inert carbon black appear consistent with its role as carrier for reactive gases such as O₃ and various aldehydes to enhance delivery of toxic materials to the deep lung (Jakab, 1992). The result of the latter study was enhanced infectivity when the carbon-gas preexposed test animals were subsequently exposed to pathologic bacteria.

Similar interactions may result from gaseous pollutants that impair the clearance of particles from the lung or otherwise alter their metabolism. Studies by Laskin at New York University in the 1960s showed an intriguing interaction of SO₂ and benzo(a)pyrene. It was thought that impaired clearance and greater residence time in the lung led to the enhanced probability of carcinogenic expression of the particle. Similarly, rats exposed to an urban eight-hour daily profile of O₃ for six weeks, followed by a five-hour exposure to asbestos, were found to retain three times as many fibers as did the controls 30 days postexposure. The fibers were deposited in the distal airways and penetrated more deeply into airway tissues making them less accessible to phagocytic removal (Pinkerton et al., 1989). These studies, together with those focusing on irritancy and infectivity, raise the prospect that realistic exposure scenarios of gaseous and particulate pollutants can interact through either chemical or physiologic mechanisms to enhance health risks of complex polluted atmospheres.

Ultrafine Carbonaceous Matter

Ultrafine carbon particles (often called black carbon) typically result from high-temperature pyrolysis or as the product of atmospheric transformation involving organic vapors and sunlight. The size of these particles allows them to slip between gas molecules moving primarily by diffusion and principles of Brownian motion. Agglomeration on surfaces or other particles in the air is their primary mode of dissipation. When concentrations exceed ~1 million/cm³, they rapidly agglomerate with each other to form larger clumps or chains of ultrafine particles. As an air pollutant, therefore, elemental carbon particles generally do not exist as singlets except near their emission points—for example, traffic or other high-temperature sources. Fine PM consists in part of agglomerates of carbonaceous organic material that if partially oxidized may be somewhat soluble in water. Some organic materials, which exist in the vapor form, condense on the ultrafine carbon (eg diesel PM). Estimates of the carbonaceous (including organic) content of ambient fine PM vary considerably but are nominally considered to be about 10% to 60% of the total mass depending on the urban or regional area. The sources of organic carbon are varied and include the combustion products of natural smoke (eg, forest fires), engine exhaust, and stationary sources as well as transformed condensates from VOCs in the air. Elemental carbon in diesel PM frequently combines into long ultrafine chains, with variably complex organics associated with its surface depending on the combustion conditions. It has been estimated that diesel contributes about 7% of the fine urban PM emissions, which, when expressed as an annual US average, is about 2 μg/m³ (USEPA, 1993)—but focus estimates (urban canyons) vary widely. The higher use of diesel fuel in Europe and areas of concentrated trucking in the United States has led to estimates as high as 30% of ambient fine PM mass. Elemental carbon has traditionally been associated with traffic, but it is a better marker of older vehicles as new engine technology for diesels, designed to meet new emission standards in 2007 and 2010, emits substantially (>98%) less black carbon. Automobiles release even less. The expected result over the next 20 years, with the turnover of truck fleets in particular, is a great reduction in ambient levels of elemental carbon.

Recent epidemiological and field studies with PM have been focusing on black carbon as an index of traffic and have reported strong associations between traffic (through the carbon indicator) and health effects. Most notably the observed adverse effects have been cardiovascular, thus giving some credence to hypotheses that ultrafine particles somehow find their way into the systemic circulation (see below) or trigger systemic inflammation that links to the cardiovascular effects (Peters, 2006; Tong et al., 2010). The penetration of ultrafine particles into the circulation under experimental conditions appears to be composition dependent (Kreyling et al., 2009).

Diesel particles vary widely in the ratio of organic and elemental carbonaceous materials, which in empirical studies has been shown to influence toxic outcomes, such as to their inflammatory and carcinogenic potential (Singh et al., 2004). Some diesel particles also appear to have adjuvant activity when tested with bioallergens in both animals and humans (Diaz-Sanchez et al., 1999). When reacted in vitro with O₃, there appears to be an enhancement of lung inflammation relative to the diesel or O₃ alone (Madden et al., 2000). However, it is important to realize that diesel particles should not be equated with whole diesel exhaust that also contains significant amounts of gaseous pollutants: NO_x, CO, and SO_x as well as various VOCs and carbonyl irritants. Exposure to diluted diesel exhaust in humans reveals that the exhaust mix is inflammogenic and to a degree cytotoxic to airway cells (Ghio et al., 2012). The use of diesel particles alone in toxicology studies does not seem to display similar toxicity, thus underscoring the potential importance of interactions among air pollutants as a critical consideration in air pollution toxicity (Mauderly and Samet, 2009).

Elemental carbon itself is generally considered to be of low toxicity, although long-term, high-concentration exposure conditions in rats can lead to lung “overload” where there is evidence of lung damage and carcinogenicity (addressed below). In the environment, carbon has the potential to act as a carrier of certain irritant gases as was noted earlier. However, carbon in the ultrafine mode (<0.1 μm) has been suggested to be more toxic than the fine mode (2.5 μm) form, perhaps due to enhanced surface reactivity or tissue penetration (Oberdorster et al., 2002; Donaldson et al., 1998). Size is not the only factor as it appears that composition of the ultrafine particle also contributes to its effects and behavior (Kreyling et al., 2009). Ultrafine particles in the environment exist in extremely high numbers but contribute negligibly to mass. Recent commercial introduction of “engineered” nanoparticles brings many of the same concerns as ultrafines by virtue of their similar sizes. Additionally, being “engineered” particles, they may possess design features that “natural” combustion ultrafine (or nano) particles do not.

Chronic Effects and Cancer

The role of air pollution in human lung cancer is difficult to assess because the vast majority of respiratory cancers result from cigarette smoking. The ACS study noted above (Pope et al., 2011) showed a significant linkage between PM and lung cancer, but in general it has been difficult to show these effects for the many HAP compounds that occur as urban air pollutants and are also thought to be carcinogenic. However, most of the HAPs and even fewer (about 10%) of the more than 2800 compounds that have been identified in the air have been assayed for carcinogenic potency. Fig. 29-14 gives estimates of the relative contributions of various chemicals to the lung cancer rate that are not associated with cigarette smoking, which, for outdoor air, is estimated to be about 2000 cases per year (Lewtas, 1993). This compares with about 2000 cases per year for passive environmental tobacco smoke and >100,000 cases per year for smokers. VOCs and nitrogen-containing and halogenated organics account for most of the compounds that have been studied with animal and genotoxicity bioassays. Most of these compounds are derived from combustion sources ranging from tobacco to power plants to incinerators to motor vehicles. Other potential carcinogens arise from mobile sources (including diesel) as products of incomplete combustion as well as their atmospheric transformation products. Fugitive or accidental chemical releases also figure into the many chemicals in ambient air. The National Scale Air Toxics Assessment (NATA) is an assessment of the cancer and noncancer risks associated with ambient HAPs (www.epa.gov/nata2005/). The profile of outdoor carcinogens contrasts with that of indoor air, where the sources are thought to derive largely from environmental tobacco smoke and radon, with some contribution from off-gassed organics (eg, adhesives, carpet polymers, cleaning agents). Human exposure to airborne toxicants is highly complex compositionally as well as in its temporal and spatial heterogeneity.

The lung cancer risk of any individual is some function of the carcinogenic nature of the substance, the amount of material deposited in the lungs, which is itself a function of the concentration in the ambient air, the physical and chemical properties of the inhalant that may determine deposition efficiency, and the cumulative volume of air inhaled. Of course, the innate susceptibility of the individual (including genotype and environmental factors such as diet, etc) is also likely to be important. A significant body of data suggests that the majority of lung cancer risk from ambient air pollution lies within the PM fraction. Among the many potent chemicals are the polycyclic organic chemicals, along with a group of less-volatile organics sometimes referred to as “semivolatiles” (including nitroaromatics). These persistent organics associate with the PM matrix and thus could have a prolonged residence time at deposition sites within the respiratory tract. Genetic bioassays have revealed the potent mutagenicity, and presumably carcinogenicity, of various chemical fractions of ambient aerosols (Lewtas, 2007). Some of these compounds require metabolic transformation to activate their potency while others may be detoxified by their metabolism. Not to be forgotten, although not a feature of this chapter, are carcinogenic vapors such as benzene that are inhaled but have target tissues away from the lung—in the case of benzene, bone marrow leukemia.

The cells lining the respiratory tract turn over relatively quickly, since they interface with the ambient environment with every breath. Conceptually, their DNA would thus be vulnerable to carcinogenic or oxidant-induced replication errors that, when fixed as mutations, could give rise to tumors. Copollutants, such as irritant gases, that initiate inflammation may promote carcinogenic activity by damaging cells and further enhancing their turnover. For example, there is experimental evidence that benzo(a)pyrene inhaled by rats whose respiratory tracts have been chronically exposed to SO₂ are prone to bronchogenic carcinomas. Likewise, epidermoid carcinomas were produced in mice that inhaled ozonized gasoline vapors, containing many reactive organic products, but only if these mice had been previously infected with influenza virus and presumably had inflamed lungs. Many believe that the so-called rural–urban gradient of lung cancer, apparent even when corrected for cigarette smoking, is a product of such complex interactions. Thus, while the phenomenon of environmental lung cancer remains poorly understood, there is general sentiment for the early opinion expressed by Kotin and Falk in 1963: “Chemical, physical and biological data unite to form a constellation that strongly implicates the atmosphere as one dominant factor in the pathogenesis of lung cancer.” At the time of this statement, however, the role of tobacco smoke was not widely appreciated.

To understand the role of inhaled particles to non-cigarette-smoke-related lung cancer, chronic exposure studies have been conducted with a number of particles ranging from titanium dioxide and carbon to diesel exhaust and coal fly ash aerosol. Of these substances, diesel exhaust has been the most extensively studied (reviewed by Cohen and Nikula, 1999; Ghio et al., 2012). The primary concern with diesel has been the suspicion that it can induce lung cancer (thus its IARC classification as a Group 1 human carcinogen). However, the evidence from over 40 occupational studies (primarily railway yard, truck, and bus workers) implicating diesel exhaust as a mild carcinogen continues to be debated because of a variety of confounding issues (Gamble, 2010). Taking the empirical data alone, however, carcinogenicity is suggested by several chronic exposure studies in animals and in vitro data indicating mutagenicity in Salmonella bacteria and enhanced sister chromatid exchange rates in Chinese hamster ovary cells. (The latter genotoxic effects have been linked to the nitroarenes associated with the diesel PM.) Rodent studies, unfortunately, have not fully resolved the question of human carcinogenic risk because of the overload needed to yield positive results, and the tumors develop only in the rat. At high concentrations of diesel PM (3.5 and 7 mg/m³), normal mucociliary clearance in the rodent models becomes overwhelmed, resulting in a progressive buildup of particles in the lungs. By 12 months in the rat, clearance irreversibly decreases to cessation with concomitant inflammation, oxidant generation, epithelial hyperplasia, and fibrosis. Rats seem to react more to this circumstance. Particle agglomerates within the alveolar lumen become the focus of inflammation, injury, and the eventual development of adenosarcomas and squamous cell carcinomas. At lower concentrations, where the particle buildup does not occur, tumors do not develop.

“Overload” is a common finding at the highest exposure concentrations of chronic particle inhalation bioassays. While this phenomenon is not likely in humans exposed to ambient PM, it is a hypothesis worthy of consideration in a discussion of particle health effects, as it relates to the interpretation of toxicological data relevant to product use and industrial exposures. Several poorly soluble particles (PSPs) have induced lung tumors in chronic rat bioassays under conditions of overload; tumors have not developed under similar conditions in mice and hamsters. Among these particles are titanium dioxide, carbon black, toner dust, talc, and diesel emission; the potential for tumors is especially marked when the particles are in the ultrafine mode. In the rat, the timecourse and pattern of accumulation, chronic inflammation, epithelial hyperplasia, and tumorigenesis are essentially the same for all of the particles. In contrast, the degree of active inflammation in the mouse and hamster under similar overload conditions appears less intense, and thus is an important distinction among the species that relates to their relative sensitivities. On one hand, such exposures do not typically demonstrate classic in vivo genotoxicity, although bulky adducts and other modifications to DNA have been described after in vivo exposure to carbon black and diesel exhaust (reviewed by Schins and Knaapen, 2007). Also, mutation of the hprt (hypoxanthine guanine phosphorybosyl transferase) gene was found in rat epithelial cells cultured with bronchoalveolar lavage from chronic carbon black– and titanium dioxide–treated rats (Driscoll et al., 1997). Even though these few studies have demonstrated inert particle-induced modifications to DNA, generalizing this finding to overload cancers, not surprisingly, remains controversial. Typically, the pathways leading to lung cancer by relatively inert and PSPs (eg, carbon black and diesel exhaust particles) are thought to be secondary and involve pathways such as oxidative stress rather than a primary carcinogenesis mechanism such as DNA mutations (Greim et al., 2001). The closest analogy in humans would be coal miners who do not appear to have an enhanced risk of lung cancer except when smoking is not involved.

The issue, then, is whether rat bioassay cancer data under conditions of overload are relevant to risk assessment. A review by an expert panel (ILSI, 2000) concluded that rats, while apparently unique in this response, may represent a sensitive subgroup, and that tumorigenesis data from the rat bioassay under conditions of overload cannot be summarily dismissed as not relevant to the consideration of cancer risk in humans. However, the data should be interpreted and weighed in the context of lower concentrations and the tumor incidence and pathology found therein.

Photochemical Air Pollution

Photochemical air pollution (notably O₃) remains the most elusive of the criteria pollutants to bring under control. It arises secondarily from a series of complex reactions in the troposphere activated by the ultraviolet (UV) spectrum of sunlight. In addition to O₃, it comprises a mixture of nitric oxides (NO_x), aldehydes, peroxyacetyl nitrates (PAN), and a myriad of aromatics and alkenes along with analog reactive radicals. If SO₂ is present, sulfates may also be formed and, collectively, they yield “summer haze.” Likewise, the complex chemistry can generate organic PM, nitric acid vapor, and various condensates. Attempts to mitigate photochemical smog (especially O₃) by controlling hydrocarbon and/or NO_x emissions have proven difficult due to the complex stoichiometry of atmospheric photochemistry. Some progress has been made in controlling peak values in the United States (regulated by the pre-1997 one-hour NAAQS), but the longer-time-frame eight-hour NAAQS has seen generally less progress. Because of a particularly strong regulatory effort in Southern California, however, the average number of high ozone days per year dropped by 33% between 2000 and 2010 (State of the Air, 2012) in the Los Angeles–Long Beach–Riverside metropolitan area, the most polluted part of the United States, in terms of ambient ozone levels.

From the point of view of the toxicology of photochemical air pollutant gases, O₃ is by far the toxicant of greatest concern. It is highly reactive and more toxic than NO_x, and because its generation is fueled through cyclic hydrocarbon radicals, it reaches greater concentrations than the hydrocarbon radical intermediates. The gaseous hydrocarbons, so integral to the chemistry, are no longer listed collectively as a criteria pollutant in the United States since they do not have a strong health-based driver. In general, the concentrations of the hydrocarbon precursors in ambient air do not reach levels high enough to produce acute toxicity, although some individual compounds fall under the HAPs rule. Rather, the importance of these hydrocarbons stems from their role in the chain of photochemical reactions.

Although O₃ is of toxicological importance in the troposphere, in the stratosphere it plays a critical protective role. About 10 to 50 km above the earth’s surface, UV light directly splits molecular O₂ into atomic O^•, which then combines with O₂ to form O₃. The O₃ also dissociates back but much more slowly. The result is an accumulation of O₃ to several ppm within a relatively thin strip of the stratosphere forming an effective “permanent” barrier by absorbing the short-wavelength UV in the chemical process. This barrier had in recent years been threatened by various anthropogenic emissions (Cl₂ gas and certain chlorofluorocarbons) that enhance O₃ degradation (creation of an “O₃ hole”), but recent restrictions on the use of these degrading chemicals seem to have been effective in reversing this process. The benefits are believed to be a reduction of excess UV light infiltration to the earth’s surface and reduced skin cancer risk.

This protective issue is quite different in the troposphere, where accumulation of O₃ serves no known purpose and poses a threat to the respiratory tract. Near the earth’s surface, NO₂ arising from combustion processes efficiently absorbs longer-wavelength UV light, from which a free O atom is cleaved, initiating the following simplified series of reactions:

This process is inherently cyclic, with NO₂ regenerated by the reaction of the NO^• and O₃. In the absence of unsaturated hydrocarbons (olefins and substituted aromatics) arising from fuel vaporization or combustion, as well as biogenic terpenes, this series of reactions would approach a steady state with little buildup of O₃. The free electrons of the double bonds of unsaturated hydrocarbons are attacked by free atomic O^•, resulting in oxidized compounds and radicals that react further with NO^• to produce more NO₂. Thus, the balance of the reactions sequence shown in Eqs. (29-1) to (29-3) is tipped to the right, leading to buildup of O₃. This reaction is particularly favored when the sun’s intensity is greatest at midday, utilizing the NO₂ provided by morning rush-hour traffic. Carbonyl compounds (especially short-chained aldehydes) are also by-products of these reactions. Formaldehyde and acrolein account for about 50% and 5%, respectively, of the total aldehyde content in urban atmospheres. Peroxyacetyl nitrate (CH₃COONO₂), often referred to as PAN, and its homologs also arise in urban air, most likely from the reaction of the peroxyacyl radicals with NO₂. In the evening when sun intensity wanes, the second rush-hour peak of NO₂ shifts the balance back by reacting with O₃ bringing the daytime peak concentration slowly down. In assessing the excess adverse effects of O₃, it is important to determine the background level of ambient O₃. Interestingly, background levels of ozone in North America are known to peak as high as 0.14 ppm (one hour), particularly in high elevation regions such as Wyoming, although the contribution of these transient one-hour levels of ozone to public health is unclear. This high altitude peak concentration is thought to be due to a combination of high VOC levels, NO_x, and the additional UV light reflected off snow cover. However, other factors such as season, thickness of boundary layers, vertical intrusions from the stratosphere, and long-range transport also play roles in determining background levels of O₃ in North America.

Short-Term Exposures to Smog

In the 1950s to 1960s, early air pollution toxicologists were challenged by the complexity of photochemical air pollution both to ascertain its potential to adversely affect human health and to determine what about these atmospheres was responsible for the effects. The focus was immediately placed on O₃ because its toxicity was found to be very high even at low ppm concentrations. Concerns that the complex atmosphere was even more hazardous led to a number of studies with actual (outdoor-derived) or synthetic (photolyzed laboratory-prepared atmospheres) smog in an attempt to assess the potency of the actual pollution mix. When human subjects were exposed to real-world photochemical air pollution (Los Angeles ambient air pumped into a laboratory exposure chamber), they experienced changes in lung function similar to those described in controlled clinical studies of O₃ alone (ie, reduction in spirometric lung volumes; see below), thus supporting the view that this oxidant was the pollutant of primary concern. Acute animal studies using synthetic atmospheres (usually irradiated auto exhaust) provided supportive evidence indicating deep lung damage, primarily within the small airway and proximal alveolar epithelium.

Thus, O₃ appeared to be the prime toxicant in many of these early studies; there was some evidence that other copollutants were involved in the effects observed with smog. When guinea pigs were exposed to irradiated auto exhaust, airway resistance increased quickly, in contrast to the pattern of O₃ alone, where less effect is seen on resistance than on breathing rate. This indicated that a more soluble irritant(s) probably was active, presumably reactive aldehydes. Thus, the array of effects of a complex atmosphere may be more diverse than would be predicted if it were assumed that O₃ alone was responsible. Interestingly, the focus over time has been almost exclusively on O₃, perhaps with the emphasis coming from the regulatory perspective tied to this single pollutant.

Chronic Exposures to Smog

Epidemiological studies in human populations as well as empirical studies in laboratory animals have attempted to link degenerative lung disease with chronic exposure to photochemical air pollution. Cross-sectional and prospective field studies have suggested an accelerated loss of lung function in people living in areas of high pollution. However, as with many studies of this type, there were problems with confounding factors (meteorology, imprecise exposure assessment, and population variables). Studies have been conducted in children living in modern-day Mexico City, which has oxidant and PM levels far in excess of any city in the United States. These studies have focused on the nasal epithelium as an exposure surrogate for pulmonary tissues, using biopsy and lavage methodologies to assess damage. Dramatic effects were found in exposed children, consisting of severe epithelial damage and metaplasia as well as permanent remodeling of the nasal epithelium. When children migrated into Mexico City from cleaner, nonurban regions, even more severe damage was observed, suggesting that the tissue remodeling in the permanent residents imparted some degree of incomplete adaptation. Because the children were of middle-class origin, these observations were less likely confounded by socioeconomic variables (Calderon-Garcidueñas et al., 1992). Changes in lung hyperinflation as estimated by x-ray tomography and impaired lung function have been reported among similarly exposed children living in Mexico City (Calderon-Garcidueñas et al., 2003) raising further concerns of long-term effects. Recently, a more mechanistic panel study demonstrated that the epithelial cell damage in the nasal cavity of Mexico City children was inversely correlated with glutathione peroxidase, a marker of oxidative stress (Hernández-Escobar et al., 2009).

A now classic synthetic smog study in animals was undertaken at the Cincinnati EPA laboratory in the mid-1960s in an attempt to address the potential for long-term lung disease. Beagle dogs were exposed on a daily basis (16 hours) for nearly six years, followed by a clean air recovery period of about three years (Lewis et al., 1974). A series of physiologic measurements were made on the dogs after the exposure, and after their three-year recovery. They were then moved to the College of Veterinary Medicine at the University of California at Davis. The lungs of the dogs then underwent extensive morphologic examination to correlate with the physiology. The dogs had been exposed to nonirradiated auto exhaust (group 1), irradiated auto exhaust (group 2), SO₂ plus sulfuric acid (group 3), the two types of exhaust plus the sulfur mixture (groups 4 and 5), and a high and a low level of NO_x (groups 6 and 7). The irradiated exhaust contained oxidant (measured as O₃) at about 0.2 ppm and NO₂ at about 0.9 ppm. The raw exhaust contained minimal concentrations of these materials and about 1.5 ppm NO. Both forms of exhaust also contained about 100 ppm CO. The control group did not show time-related lung function changes, but all the exposure groups had functional abnormalities, most of which persisted or worsened over the three-year recovery period in clean air. Enlargement of airspaces and loss of interalveolar septa in proximal acinar regions were most severe in dogs that were exposed to NO_x and SO_x with irradiated exhaust (Hyde et al., 1978). These studies described a morphologic lesion that was degenerative and progressive in nature, not unlike that of chronic obstructive pulmonary disease (COPD)—a condition most often associated with lifelong tobacco smoking.

Long-term “sentinel animal” studies in polluted cities have been attempted whereby the animals live in the same highly polluted air to which people are exposed. Such approaches have had a troubled past, but newer studies appear to be better controlled for the problems of infection, inappropriate animal care (eg, heat), and variable exposure atmospheres. One such study, conducted in rats exposed for six months to the air of São Paulo, Brazil, found considerable airway damage, lung function alterations, and altered mucus rheology (Saldiva et al., 1992). The concentrations of O₃ and PM in São Paulo frequently exceed daily maximum values (in the summer months of February and March) of 0.3 ppm and 75 μg/m³, respectively. This collage of effects is not unlike a composite of injury one might suspect from a mixed atmosphere of oxidants and acid PM in controlled laboratory animal studies. Strangely enough, however, a seven-week study in rats exposed to the polluted Mexico City air, which had induced significant lesions in children, did not reveal any nasal or lung histopathology in F-344 rats (Moss et al., 2001). While there is no clear reason for the apparent differences in the findings, it is important to appreciate that all sentinel studies have elements of exposure that may be uncontrolled and hence at times can yield conflicting findings.

Ozone

General Toxicology

Ozone is the primary oxidant of concern in photochemical smog because of its inherent bioreactivity and its concentration relative to other reactive species. Although it depends on the meteorological conditions of a given year, approximately 129 million Americans lived in areas not in absolute compliance with the 8-hour NAAQS in 2010. Progress has been made in reducing ambient O₃ levels and Los Angeles, for example, has not exceeded the Stage 1 smog alert (one-hour peak of 0.20 ppm or more) since 1998 and dropped the number of exceedances of the 0.075 ppm NAAQS in half from 1990 to 2011. Unlike SO₂ and the reducing-type pollution profile discussed above, current mitigation strategies for O₃ have been only marginally successful despite significant reductions in individual automobile emissions. These reductions have been offset by population growth, which brings with it additional vehicles and vehicle miles driven. With suburban sprawl and the downwind transport of air masses from populated areas to more rural environments, the geographic distribution of those exposed has also expanded, as has the temporal profile of individual exposure. In other words, ambient O₃ exposures are no longer stereotyped as brief one- to two-hour peaks. Instead, there is more typically a prolonged period of exposure of six hours or more at or near the NAAQS level. This important change in the exposure profile to O₃ has given rise to concerns that cumulative damage over an exposure of several hours may be more significant than brief pulse-like exposures, and that, as a result, many more people are at risk than was previously thought. With the 1997 revision of the US O₃ NAAQS to include an eight-hour daily average of 0.08 ppm, more cities and suburban areas find themselves in violation of the standard. If indeed the damage to the lung is cumulative over an eight-hour time period, then people in areas of compliance but with exposure levels near the standard may be affected. The American Lung Association estimates that nearly 50% of US inhabitants live in counties that are not in O₃ compliance (State of the Air, 2012). Perhaps of greater significance is that of those who might be considered susceptible due to age and/or preexistent cardiopulmonary impairments, 80% to 90% live in those areas that fail to comply with the present O₃ NAAQS.

Ozone induces a variety of effects in humans and experimental animals at concentrations that occur in many urban areas (reviewed by Lippmann, 1989; U.S. EPA, 2004). These effects include morphologic, functional, immunologic, and biochemical alterations. Because of its low water solubility, a substantial portion of inhaled O₃ penetrates deep into the lung, but its reactivity is such that about 17% and 40% is scrubbed by the nasopharynx of resting rats and humans, respectively (Hatch et al., 1994; Gerrity et al., 1988). The reason for the higher degree of scrubbing in humans is unclear, but the finding is reproducible. Moreover, the mouth as well as the nose appears to scrub. Nevertheless, regardless of species, the region of the lung that is predicted to have the greatest O₃ deposition (dose per surface area) is the centriacinar region, from the terminal bronchioles to the alveolar ducts, also referred to as the proximal alveolar ductal region (Overton and Miller, 1987). Because O₃ penetration increases with increased tidal volume and flow rate, exercise increases the dose to the target area. Using ¹⁸O₃ (a nonradioactive isotope of oxygen), Hatch and coworkers have shown that the dose to the distal lung and the degree of damage to the lung as determined by leakage of plasma protein into the alveolar space (as collected by bronchoalveolar lavage) in exercising human subjects exposed to 0.4 ppm for two hours with intermittent periods of 15 minutes of exercise (threefold normal ventilation on average) are similar to those in resting rats exposed for the same length of time to 2.0 ppm. Thus, it is important to consider the role of exercise-associated dosimetry in a study of O₃ or any inhalant before making cross-study comparisons, especially if that comparison is across species.

Animal studies indicate that the acute morphologic responses to O₃ involve epithelial lining cells along the entire respiratory tract. The pattern of injury parallels the dosimetry profile, with the majority of damage occurring in the distal lung. Along the conducting airways, ciliated cells appear to be most sensitive to O₃, while Clara cells and mucus-secreting cells are the least sensitive. Studies in the rat nose indicate that O₃ also is an effective mucus secretagogue. In the alveolar region, type 1 epithelia are very sensitive to O₃ relative to type 2 cells, which have an active metabolic machinery. The more resistant type 2 cells serve as the stem cells for the replacement of type 1 cells. Ultrastructural damage to alveolar epithelia can be observed in rats after a few hours at 0.2 ppm, but sloughing of cells in the distal airway generally requires concentrations above 0.8 ppm. Recovery occurs within a few days, and there appears to be no residual pathology. Hence, from animal studies, it would appear that a single exposure to O₃ at a relatively low concentration is not likely to cause permanent damage. On a gross level, when a bronchoscope is used to peer into the human bronchus after O₃ exposure, the airways appear “sunburned” and, as with mild skin sunburn, recovery is typical. What is uncertain is the impact of repeated “sunburning” of the airways and lung.

Studies have shown particular sensitivity to O₃-induced inflammation in neonatal rats that decreased with age into adulthood. However, enhanced sensitivity to O₃ again appears in older rats as evidenced by interstitial edema that is not seen in younger adult animals. Older rats have lower ascorbate in the lung and this diminishment of antioxidant capacity may be in part responsible. Elderly humans, however, seem less responsive to the lung function impacts of high ambient O₃. The reasons for this are unclear, but caution in linking lung function and inflammation in the O₃ response is warranted.

The mechanisms by which O₃ causes injury have been studied using cellular as well as cell-free systems. As a powerful oxidant, O₃ seeks to extract electrons from other molecules. The surface fluid lining the respiratory tract and the cell membranes that underlie the lining fluid contain a significant quantity of polyunsaturated fatty acids (PUFA), either free or as part of the lipoprotein structures of the cell. The double bonds within these fatty acids have a labile, unpaired electron that is easily attacked by O₃ to form ozonides that progress through a less stable zwitterion or trioxolane (depending on the presence of water); these ultimately recombine or decompose to lipohydroperoxides, aldehydes, and hydrogen peroxide. These pathways are thought to initiate propagation of lipid radicals and auto-oxidation of cell membranes and macromolecules (Fig. 29-15).

Evidence of free radical–related damage in vivo includes detection of exhaled breath pentane and ethane and tissue measurements of diene conjugates. Damage to the air–blood interface disrupts its barrier function and promotes inflammation. Inflammatory cytokines (eg, interleukins 6, 8, and others, TNF) are released from epithelial cells and macrophages that mediate early responses and initiate repair. This inflammatory process is generally transient, but it may also interact with neural irritant responses to affect lung function acutely. The latter response may have implications for those with preexistent inflammation or disease.

Pulmonary Function Effects

Exposure to O₃ produces a variety of pulmonary function changes and human subjects appear to be more sensitive than animal models. Early studies showed that exercising human subjects exposed for two to three hours to 0.12 to 0.4 ppm O₃ experience reversible concentration-related decrements in forced exhaled volumes (FVC and forced expiratory volume in one second [FEV₁]) (McDonnell et al., 1983). Because of the concern that prolonged periods of exposure (six–eight hours) may lead to cumulative effects, similar protocols with lower exposure and exercise levels were extended up to 6.6 hours. In these studies, exposures to 0.12, 0.10, and 0.08 ppm induced progressive lung function impairment during the course of the exposure (Horstman et al., 1989). The pattern of response was a function of exposure time such that functional changes were not detectable at 1 or 2 hours but reached significance by four to six hours. Decrements in FEV₁ after 6.6 hours at 0.12 ppm averaged 13.6% and were comparable to that observed after a two-hour exposure to 0.22 ppm with much heavier exercise. These studies were then followed up with exposures to even lower O₃ concentrations. The latter controlled human exposure studies demonstrated that multihour exposures to 0.06 and 0.08 ppm O₃ (with exercise) produced small but statistically significant mean decreases in FEV₁ (in the range of 3%–4%) as well as increases in respiratory symptoms and inflammation. It is noteworthy that in these and other controlled human exposure studies with O₃, considerable interindividual differences in response were observed, thus confirming the many animal studies showing that the response to ozone is dependent on many innate factors such as age, diet, genetics, preexisting disease, and gender.

It is not clear what mechanisms underlie the altered lung function (in terms of changes in FEV₁) produced by O₃. Although chest pain/discomfort is thought to contribute to O₃-induced decreases in an effort-dependent lung function maneuver such as FEV₁, there is also evidence that the decrements in lung function are vagally mediated, and that the response can be abrogated by analgesics, such as ibuprofen and opiates, which also reduce pain and inflammation (see below). Thus, pain reflexes involving C-fiber networks are thought to be important in the reduction in forced expiratory volumes. On the other hand, animal studies show a prominent role for vagal reflexes in altered airway reactivity and bronchoconstriction. There is also evidence to implicate vagal reflexes in cardiac as well as thermal regulation, at least in rodents.

Airway responsiveness to specific (eg, allergen) and nonspecific (eg, cold air, inhaled methacholine) bronchoconstriction is another commonly used test of the pulmonary response to inhaled pollutants such as O₃. These types of tests are very important because airway hyperresponsiveness is a central feature of asthma and asthmatics are a sizeable subpopulation (7%–9% of the total population in the United States) that may be particularly sensitive to the adverse respiratory effects of inhaled pollutants. A number of animal models have demonstrated that O₃ induces airway hyperresponsiveness and that innate factors such as genetics, diet, and obesity can modify this effect. For the most part, many of these studies have utilized O₃ exposure concentrations that are far above ambient levels. Importantly, controlled human exposure studies have not only verified the findings of the high-dose animal studies but also demonstrated adverse effects at concentrations relevant to ambient O₃ levels. Airway hyperreactivity to nonspecific agonists in humans has been reported after acute exposure to as little as 0.08 ppm O₃. It is widely thought that hyperreactive airways may be a marker of predisposition to other pollutants such as sulfuric acid or aeroallergens, but such evidence is limited.

Inflammation of the Lung and Host Defense

The mechanism by which O₃ produces decrements in pulmonary function is not fully understood and its link to airway inflammation is unclear. For example, O₃-induced lung dysfunction does not appear to be enhanced in asthmatics whose lungs are generally in a state of inflammation. Asthmatics are, however, clearly more sensitive to the bronchoconstrictive effects of both SO₂ and sulfuric acid aerosols. Other factors that play a role in the pulmonary response to O₃ include age and obesity. Animal studies have demonstrated that O₃-induced lung injury and inflammation is greater in neonatal compared with that in adult rabbits, rats, and mice (Vancza et al., 2009). In addition to age as a susceptibility factor, a series of studies has shown that obesity predisposes mice to the adverse effects of O₃. Because obesity is associated with ongoing systemic inflammation, it suggests that a preexisting physiologic or disease state enhances the adverse effects of O₃.

Both human and animal studies have demonstrated that sensitivity to O₃ appears to have a genetic component as well. Studies in inbred stains of mice have shown that O₃-induced pulmonary neutrophilia, airway hyperresponsiveness, and permeability are significantly affected by SNPS in a single gene linked to the Toll-like receptor 4 (Tlr4) locus that has been associated with endotoxin sensitivity (Kleeberger et al., 2000). Further research has determined that the involvement of an innate immunity factor such as Tlr4 in the multifaceted response to O₃ involves the release of hyaluronan fragments that appear to directly activate the Tlr4 cascade in mice (Li et al., 2011). The role of genetic polymorphisms in the pulmonary response to O₃ is complicated, however, and many other pathways including TNF, IL-1, Nrf-2, and NF-κB can modulate the response to O₃ in animal models (Bauer and Kleeberger, 2010).

The growth of genomic and proteomic technologies has made it possible to begin studies in humans to define genetic linkages to O₃ sensitivity. These studies have typically focused on genes involved in oxidative stress pathways. For example, again using Mexico City as a backdrop, it was found that a polymorphism in GSTM1 (as noted above with regard to PM) appears to convey O₃ sensitivity as well (Romieu et al., 2004). Similarly, functional polymorphisms in the catalase and the myeloperoxidase genes have been associated with enhanced response to ambient ozone. More detailed controlled O₃ chamber exposures have confirmed that genetic polymorphisms in these and other genes, including NAD quinone oxidoreductase (NQO), GSTP1, heme oxygenase-1, and TNF, are associated with responsiveness to ozone and that gene–gene interactions are involved.

Because of the strong epidemiological evidence that O₃ can exacerbate asthma (eg, increased hospital admissions) and the inability of epidemiology studies to demonstrate causality, investigators have utilized chamber studies to test the hypothesis that O₃ enhances the immune response to allergens. The potential for O₃ to influence allergic sensitization or challenge–responses has received a significant amount of investigation in animal studies. Enhanced response to allergen challenge has been noted in a number of animal species exposed to ozone. Biochemical, cellular, and pulmonary function parameters of allergic response have all been observed to increase in sensitized animals exposed to O₃ prior to antigen challenge. Ozone’s effect on allergic asthma also occurs at the sensitization stage in which repeated exposure to O₃ before or during the sensitization process exacerbates a range of allergic end points. Thus, in general, animal studies have shown the ability of O₃ to enhance the allergic process under different conditions, but evidence of this in humans is less strong. Controlled studies of heightened antigen responsiveness in allergic subjects have only been suggestive, with enhancement of allergic rhinitis after 0.5 ppm for four hours. However, diary studies of asthmatic nurses reported worsened allergy symptoms, as well as durations thereof, at concentrations of O₃ near the NAAQS (Schwartz, 1992). Similarly, eosinophils and IL-4, central components of allergic asthma, were increased in some but not all human chamber studies of O₃ exposure (U.S. EPA ISA, 2012).

Early research showed that exposure to O₃ before a challenge with aerosols of infectious agents produces a higher incidence of infection than is seen in control animals (Coffin and Blommer, 1967). Studies have demonstrated that this type of effect can be a direct result of altered phagocytosis by macrophages in O₃-exposed mice previously infected with an aerosol of Streptococcus (group C) bacteria (Gilmour et al., 1993). The host resistance model has shown responsiveness to an exposure as low as 0.08 ppm for three hours. The susceptibility of mice and hamsters to Klebsiella pneumoniae aerosol is also increased by prior exposure to O₃. In the rat, altered microbe-killing may relate to membrane or receptor damage in macrophages, thus impairing the production of bactericidal agents such as superoxide anions. The rat appears less susceptible than mice because it has a more vigorous PMN response to bacteria than do mice, which seems to compensate for macrophage impairments. This is yet another example of where susceptibility lies more in the inability to compensate than in the initial responsiveness to a given challenge. It is not clear whether O₃’s impairment of bacterial infections in the lung also extends to viral infections. While there is epidemiological evidence of ambient O₃ being associated with increased hospital admissions for viral and bacterial infections, animal studies have shown mixed results for O₃’s effect on viral infections, although some of the contradictory findings may be a result of experimental design differences (eg, O₃ exposure occurring before or after viral infection).

Chronic Effects

Morphometric studies of the centriacinar region of rats exposed for 12 hours per day for six weeks to 0.12 or 0.25 ppm O₃ have shown hyperplasia and hypertrophy of type 1 alveolar cells (smaller and thicker cells) coupled with damage and alterations in ciliated and Clara cell populations in small airways (Barry et al., 1988). A collective cross-protocol analysis of type 1 cell hypertrophy was conducted; type 1 cell thickness appeared to be linearly related to the O₃ C × T (Chang et al., 1991, 1992). This finding suggested that over a season, the impact of O₃ in the distal lung may be cumulative and perhaps more importantly may be without threshold. The biological significance of this change is unclear—it may be part of a compensatory response to “thicken” that part of the alveolar duct junction that receives the greatest dose and is most affected. This response may be protective since the thickened cells were also smaller, offering therefore a smaller exposure surface to the incoming O₃. This changed morphology could limit membrane damage. When returned to clean air, most of the epithelial morphologic changes regressed, but there was evidence of residual interstitial remodeling below the epithelium in the alveolar duct region. Examination of autopsied lung specimens from young smokers shows many analogous tissue lesions that come to be described as the “smoldering” precursor of emphysema. A long series of studies at the University of California, Davis, has examined the structural and functional changes produced in nonhuman primates exposed chronically to O₃. Because of the greater similarity of monkey respiratory tract to that of humans, the burden of extrapolation from rodent study results to risk assessment for humans is lessened, and these studies have proven important in establishing the causality of many of the long-term effects of O₃ observed in epidemiology studies. Episodic exposure of adult monkeys to 0.25 ppm O₃ for 18 months produced physiologic, cellular, and biochemical changes in the lung. Exposure to as little as 0.2 ppm O₃ for up to 90 days was found to produce cellular changes throughout the respiratory tract. Nasal lesions were accompanied by changes in cell populations in the respiratory bronchioles. In the second phase of these studies, infant rhesus monkeys were cyclically exposed to 0.5 ppm O₃ for eight hours per day for five days followed by several days of exposure to air. While the study design did not permit a comparison of the responsiveness of infant versus adult monkeys, the infant monkeys showed evidence of structural changes that parallel epidemiology studies linking lung growth and development with exposure to air pollution. The infant monkeys also showed a significant interaction between sensitization to house dust mite antigen and ozone. While antigen or ozone alone produced small changes in baseline airway resistance and airway responsiveness, antigen plus ozone produced more than additive effects on both parameters.

Studies involving episodic exposures of rats and monkeys using a pattern of alternating months of O₃ (0.25 ppm) for 18 months indicate that there may be carryover effects, notably thickening of interstitial fibrous matrix (Tyler et al., 1988, 1991). These interstitial changes were quantitatively similar regardless of the twofold difference in the cumulative exposure dose (ie, C × T). This would imply that a pattern of exposure resembling seasonal O₃ patterns might result in more serious lesions than predicted by dose alone—indeed more than would have occurred had the exposure been continuous. Hence, the concept of “more dose = more effect” may not hold in chronic episodic scenarios for O₃, as it appears to do with uninterrupted exposures. The number of episodes experienced may well be more significant to long-term outcomes than total dose—a phenomenon not unlike that of repeated sunburning and deterioration of the skin.

Studies of lung function in rodents exposed chronically to O₃ have been conducted, but have yielded mixed results at relevant exposure concentrations. Generally, the dysfunction is reflective of stiffened or fibrotic lungs, particularly at higher concentrations. There have been two prominent chronic O₃ studies—the EPA 18-month chronic study fashioned after a realistic urban exposure profile with a peak exposure of 0.25 ppm (Chang et al., 1992; Costa et al., 1995) and the National Toxicology Program (NTP)–Health Effects Institute (HEI Report, 1994/1995) study of 0.125 to 1 ppm for 20 months (square wave; six hours per day; five days per week). From an environmental relevance perspective, the C × T doses for these studies were similar, but the urban profile study produced evidence for centriacinar interstitial fibrosis suggesting a possible influence of the exposure pattern. There was no general biochemical evidence for fibrosis (Last et al., 1994), as reported in monkey and rat studies that had been conducted at higher O₃ concentrations. If one attempts to compare these results with the Cincinnati beagle study, one finds that the synthetic smog atmosphere showed degenerative and not fibrotic lung lesions. However, it should be noted that the air pollutant mixture used in the beagle study both was more complex and involved considerably higher concentrations than more recent studies.

The ability of O₃ to induce tolerance to itself is a curious phenomenon that has implications for both episodic and chronic exposures. Classic O₃ tolerance takes the form of protection against a high or even lethal dose in animals that received a very low initial challenge or challenges several days before. This term, tolerance, is sometimes used to describe “adaptation” or acclimatization over time to near-ambient levels of O₃, and, as such, has led to some confusion. However, with regard to “adaptation” to O₃, the process begins during and immediately after the initial exposure and progresses to completion in at most two to four days. This adaptive phenomenon has been well established in humans with regard to lung function and has been correlated with several inflammatory end points (Devlin et al., 1993). Lavage lactate dehydrogenase (LDH; a marker of cell injury) and elastase (enzymatically active against lung matrix material), interestingly, do not appear to adapt in humans based on the Devlin study. An analogous pattern of adaptation of functional and biochemical end points (including LDH and elastase) in rodents also takes place with repeated exposures up to a week. But to date, the linkages between acute, adaptive, and long-term process remain unclear, since over longer periods of exposure both morphologic and functional effects do appear to develop. The precise mechanism for O₃ adaptation is not known and several theories abound, including changes in cell profiles, lung surface fluids, and induced antioxidants. Few studies have tackled the problem but in rats the adaptation of the neutrophilic response appears to be related to the induction of an endogenous acute-phase response (McKinney et al., 1998). On the other hand, adaptation to lung function changes in rats after chronic exposure appears linked to lung antioxidants such as ascorbic acid (Wiester et al., 2000). The significance of this finding in humans is uncertain because ascorbic acid is not endogenously synthesized as it is in the rat. However, self-administration of ascorbate has been shown to reduce O₃-induced lung function decrements in adults (Mudway et al., 1999), and in children, supplementation with ascorbate and α-tocopherol lessened nasal inflammatory responses to O₃ in Mexico City (Sienra-Monge et al., 2004). Despite these interesting findings, it remains unclear if antioxidant supplements can protect humans from long-term O₃ effects given the many mechanisms that may be involved in the various responses. How these interplay with long-term adaptation and the likelihood of degenerative disease is unclear.

Ozone Interactions with Copollutants

An approach simplifying the complexity of synthetic smog studies, yet addressing the issue of pollutant interactions involves the exposure of laboratory animals or humans to binary or more complex synthetic mixtures of pollutants that occur together in ambient air. The most frequent combination involves interactions of O₃ and NO₂ or O₃ and PM (eg, sulfuric acid or diesel particles). Not surprisingly, study design adds a level of complexity in interpretation such that evidence exists supporting either augmentation or antagonism of lung function impairments, lung pathology, and other indices of injury. This apparent conflict in the findings only emphasizes the need to carefully consider the myriad of factors that might affect studies involving multiple determinants and the nature of the exposure that is most relevant to reality.

When O₃ and NO₂ (1 and 14 ppm, respectively) were administered to rats from a premixed retention chamber, the resulting damage evident in bronchoalveolar lavage exceeded that of either toxicant alone, regardless of the temporal sequence of exposure (Gelzleichter et al., 1992). Biochemical and histological indices of fibrogenesis also were increased in related studies (Last et al., 1994). In retrospect, it was hypothesized that the two oxidants formed relatively stable intermediate nitrogen radicals that were more toxic than either gas alone. At lower, more realistic concentrations (0.3 ppm O₃ and 3.0 ppm NO₂), where this reaction would not be favored, the impact of these irritants on rabbits was only additive (Schlesinger et al., 1991). This contrast in response serves to illustrate that the tenets of dose dependency that hold for any single-toxicant response may be of equal or more importance when two or more pollutants coexist and have the potential to interact.

Studies of O₃ mixed with acid aerosols also have shown enhanced or antagonistic responses that were time-dependent during the period of exposure. On the one hand, as noted above, field studies of children in camps and studies of asthma admissions in the Northeast and in Canada suggested an interaction of acid and O₃ underlying responses to summer haze. Yet, in an experimental setting with rabbits exposed over an extended period, there was exposure duration–specific evidence of enhanced as well as antagonized secretory cell responses with combined O₃ (0.1 ppm) and sulfuric acid (125 μg/m³) over the course of the one-year exposure (Schlesinger et al., 1992). As the number of interacting variables increases, so does the difficulty in interpretation. Studies of complex atmospheres involving acid-coated carbon combined with O₃ at near-ambient levels also show varied evidence of interaction on lung function and macrophage receptor activity (Kleinman et al., 1999). As such, the platform of any multicomponent study is its statistical design and the ability to either separate or determine the nature of the interacting variables. However, it is indeed the complex mixture to which people are exposed that we wish to evaluate. Creative approaches to understanding mixture responses are a likely part of the new agenda that toxicologists will need to address in the future (Mauderly, 2006).

Nitrogen Dioxide

General Toxicology

Nitrogen dioxide, like O₃, is a deep lung irritant that can produce pulmonary edema if it is inhaled at high concentrations. It is a much less potent irritant and oxidant than O₃, but NO₂ can pose clear toxicological problems. Potential life-threatening exposure is a real-world problem for farmers, as near-lethal high levels of NO₂ can be liberated from fermenting fresh silage. Being heavier than air, the generated NO₂ and CO₂ displace air and oxygen at the base of silo and diffuse into closed spaces where workers can inadvertently get exposed to very high concentrations perhaps with depleted oxygen. Typically, shortness of breath rapidly ensues with exposures nearing 75 to 100 ppm NO₂, with delayed edema and symptoms of pulmonary damage. Not surprisingly, the symptoms are collectively termed “silo-filler’s disease.” Nitrogen dioxide is also an important indoor pollutant, especially in homes with unventilated gas stoves or kerosene heaters (Spengler and Sexton, 1983) or, in developing countries, with the unvented burning of biomass fuels (Kumie et al., 2009). Under such circumstances, very young children and their mothers who spend considerable time indoors may be especially at risk. In general, indoor environments with NO₂ sources achieve concentrations far in excess of those observed outdoors at central monitoring sites. Peak levels of NO₂ near high-traffic-volume roadways can be several-fold higher than away from roadways and are thus similar to peak levels observed indoors with unvented stoves (ie, 0.2–0.5 ppm). Interestingly, protocols that simulate an urban (rush-hour) or household (cooking) patterns of two daily peaks superimposed on a low continuous background concentration have elicited effects in experimental animals that continuous exposure to NO₂ did not evoke, suggesting an important dependency on exposure profile. Among the common air pollutants, empirical studies of NO₂ have frequently shown greater effects with higher peak concentrations at equivalent C × T steady-state concentrations.

Although the distal lung lesions produced by acute NO₂ are similar among species, there exist differences in species sensitivity. Where direct comparison is possible, guinea pigs, hamsters, and monkeys appear more sensitive than rats, although comparative dosimetry information might explain some of this difference. As in the case of O₃, theoretical dosimetry studies indicate that NO₂ is deposited along the length of the respiratory tree, with preferential deposition being in the distal airways. Not surprisingly, the pattern of damage to the respiratory tract reflects this profile: damage is most apparent in the terminal bronchioles, just a bit more proximal in the airway than is seen with O₃. At high concentrations, the alveolar ducts and alveoli are also affected, with type 1 cells again showing their sensitivity to oxidant challenge. In the airways of these animals there is also damage to epithelial cells in the bronchioles, notably with loss of ciliated cells, as well as a loss of secretory granules in Clara cells. Thus, the pattern of injury of NO₂ is quite similar to that of O₃, although its potency is about an order of magnitude lower.

Pulmonary Function Effects

Exposure of normal human subjects to concentrations of <4 ppm NO₂ for up to three hours produces no consistent effects on spirometry. However, a study has shown slightly enhanced airway reactivity with 1.5 to 2.0 ppm. Interestingly, ascorbic acid pretreatment of human subjects appeared to protect them from this hyperreactivity (Mohsenin, 1987). Whether asthmatics have a particular sensitivity to NO₂ is a controversial issue. A number of factors appear to be involved (eg, exercise, inherent sensitivity of the asthmatic subject, exposure method). Some studies have reported effects in some individuals at 0.2 ppm, which approximates an indoor level in a household with an unvented gas stove. Recent meta-analyses, which have incorporated the findings of many studies to achieve a weight-of-evidence perspective, support an effect of NO₂ on asthmatics, but it is not clear if the small but statistically significant changes in airway hyperresponsiveness in asthmatics exposed to NO₂ below 0.6 ppm are adverse. As for an appropriate animal model, only very high concentrations (10 ppm NO₂) invoke an irritancy response in guinea pigs (tachypnea); these levels are well above those a person probably would encounter in everyday life. However, NO₂ has been found to be associated with mortality in some time-series studies of air pollution attempting to tease out specific pollutant effects (focusing mainly on PM) (Gold et al., 2000; Samoli et al., 2006). These epidemiology studies have demonstrated that each 20 to 30 ppb increase in NO₂ results in a small but significant increase in acute cardiovascular and respiratory mortality. Similarly, NO₂ has been associated with increased hospital admissions and emergency room visits, but these epidemiology studies are confounded by the coexposure to multiple pollutants. Thus, as stressed elsewhere, it is important that epidemiology studies results be interpreted with caution when considering causality and that controlled human exposure and animal studies be utilized, within ethical limits, to confirm the biological plausibility of the effect of NO₂, or any other pollutant, on the cardiopulmonary system. This caution obviously extends to many recent studies that have consistently linked adverse health outcomes with living near high-traffic-volume roadways. Because of the many constituents present in the exhaust of both gas and diesel combustion engines, as well as their ever changing chemistry, it is unclear from these studies whether NO₂ is acting as a marker or surrogate for vehicular traffic rather than as an indication of a specific NO₂ effect. These studies have found correlates with cardiovascular deaths, which have raised new questions of the mechanisms by which pollutants might affect health in susceptible subgroups (Rosenlund et al., 2006).

Inflammation of the Lung and Host Defense

Unlike O₃, NO₂ does not induce significant neutrophilic inflammation in humans at exposure concentrations encountered in the ambient outdoor environment. There is some evidence for bronchial inflammation after four to six hours at 2.0 ppm, which approximates the highest transient peak indoor levels of this oxidant. Exposures at 2.0 to 5.0 ppm have been shown to affect T lymphocytes, particularly CD8⁺ cells and natural killer cells that function in host defenses against viruses. Although these concentrations may be high, epidemiological studies variably show effects of NO₂ on respiratory infection rates in children, especially in indoor environments. Animal models, by contrast, have for years shown associations between NO₂ and bacterial infection (Gardner, 1984). As noted for other effects, the incidence of infection in exposed models appears to be governed more by the peak exposure concentration than by exposure duration. The effects are ascribed to suppression of macrophage function and clearance from the lung, in the form of suppressed bactericidal and/or motility functions of macrophages from rabbits exposed to 0.3 ppm for three days (Schlesinger, 1987). Similar effects have been reported in humans exposed to 0.10 ppm for 6.6 hours (Devlin et al., 1991).

Toxicological studies of the interaction of NO₂ with viruses also suggest enhanced infectivity. Squirrel monkeys infected with nonlethal levels of A/PR-8 influenza virus and then exposed continuously to 5 or 10 ppm NO₂ suffered high mortality rates; 6/6 monkeys exposed to 10 ppm died within three days, while only 1/3 exposed to 5 ppm died (Henry et al., 1970). Other experiments suggest that exposure of squirrel monkeys for five months to 5 ppm NO₂ depresses the formation of protective antibodies against the A/PR-8 influenza virus. Controlled human studies with virus challenges, however, have been inconclusive, perhaps because of low subject numbers. One study showed decreased virus inactivation by alveolar macrophages recovered from four of nine subjects when cultured and exposed for 3.5 hours to 0.6 ppm NO₂ in vitro. The responsive macrophages produced interleukin-1, a known cytokine modulator of immune cell function (Frampton et al., 1989). Thus, the potential for augmented risk of viral infection associated with NO₂ exposure remains unclear and suggests a role for underlying host susceptibility. This concern would be greatest for children, especially during seasonal use of unvented gas heaters, who have less mature pulmonary immune function.

Chronic Effects

Concern about the chronic effects of NO₂ stems from early observations that a continuous 30-day, 30-ppm exposure produces emphysema in hamsters. Whether the result of this exposure scenario relates to cyclic human exposures at 1/100th that level is unclear. On the one hand, an 18-month study in rats exposed to an urban pattern of NO₂ in which a daily background of 0.5 ppm peaked at 1.5 ppm for 4 hours each day showed little ultrastructural damage to the distal lung (Chang et al., 1988). On the other hand, mice exposed for a year to a base level of 0.2 ppm NO₂ with a one-hour spike of 0.8 ppm twice a day, five days per week (Miller et al., 1987) displayed effects that differed between the base-only and peak-only exposure groups. The base level produced no effects, while the overlaid peaks induced slight functional impairment and augmented susceptibility to bacterial infection. Early studies (Ehrlich and Henry, 1968) showed that clearance of bacteria from the lungs is suppressed with 0.5 ppm NO₂ through 12 months of exposure. Interestingly, studies with a similar double diurnal peak design for NO₂, with NO used as a negative control, showed more pronounced effects of NO on alveolar septal remodeling than did NO₂ (Mercer et al., 1995). These and similar studies utilizing varied peak-plus-baseline NO_x exposures indicate the importance of exposure profile for at least this pollutant.

Other Oxidants

While a number of other reactive oxidants have been identified in photochemical smog, most are short-lived because of their reaction with copollutants. PAN, which is thought to be responsible for much of the eye-stinging activity of smog, is known to exist in smog situations. It is more soluble and reactive than O₃, and hence rapidly decomposes in mucous membranes before it can penetrate into the respiratory tract. The cornea is a sensitive target and is prominent in the burning/stinging discomfort often associated with oxidant smogs. A few studies with high levels of PAN have shown that it can cause lung damage and have mutagenic activity in bacteria, but it is not likely that these scenarios are relevant to ambient levels of PAN.

Aldehydes

Carbonyl compounds, notably short-chained (2-4 C) aldehydes, are common photo-oxidation products of unsaturated hydrocarbons. Two aldehydes are of major interest by virtue of their concentrations and irritancy: formaldehyde (HCHO) and acrolein (H₂C=CHCHO). They contribute to the odor as well as eye and sensory effects of smog. Formaldehyde accounts for about 50% of the estimated total aldehydes in polluted air, while acrolein, the more irritating of the two, accounts for about 5% of the total. Acetaldehyde (C₃HCHO) and many other longer-chain aldehydes make up the remainder, but they are not as intrinsically irritating, exist at low concentrations, and have less solubility in airway fluids. Formaldehyde and particularly acrolein are also found in mainstream tobacco smoke (~90 and ~8 ppm, respectively, per puff) and are likely to be found at lower levels in sidestream smoke as well. Formaldehyde is also an important indoor air pollutant and can often achieve higher concentrations indoors than outdoors due to off-gassing by new upholstery or other furnishings. In the United States, the formaldehyde off-gassing issue was highlighted in the controversy surrounding the trailer homes provided to hurricane-displaced home owners in Katrina-New Orleans-2005.

Empirical studies have shown that formaldehyde and acrolein are competitive agonists for similar irritant receptors in the airways. Thus, irritation may be related not to “total aldehyde” concentration but to specific ratios of acrolein and formaldehyde. Their relative difference in solubility, with formaldehyde being somewhat more water-soluble and thus having more nasopharyngeal uptake, may distort this relationship under certain exposure conditions (eg, exercise). On the other hand, acrolein is very reactive and may interact easily with many tissue macromolecules and, for example, can form DNA adducts.

Formaldehyde

Formaldehyde is a primary sensory irritant. Because it is very soluble in water, it is absorbed in mucous membranes in the nose, upper respiratory tract, and eyes. The dose–response curve for formaldehyde is steep: 0.5 to 1 ppm yields a detectable odor, 2 to 3 ppm produces mild irritation, and 4 to 5 ppm is intolerable to most people. Formaldehyde is thought to act via sensory C-fibers that signal locally as well as through the trigeminal nerve to reflexively induce bronchoconstriction through the vagus nerve. In guinea pigs, a one-hour exposure to about 0.3 ppm of formaldehyde induces an increase in airflow resistance accompanied by a smaller decrease in compliance (Amdur, 1960). Respiratory frequency and minute volume also decreased, but these changes were not statistically significant until >10 ppm. The no observed effect level (NOEL) using these lung function criteria is about 0.05 ppm. The general pattern of the irritant response and its rapid recovery is similar to that produced by higher concentrations of SO₂. Like SO₂, breathing through a tracheal cannula to bypass nasal scrubbing greatly augments the irritant response, indicating that deep lung irritant receptors can also be activated by this vapor.

The irritancy of inhaled formaldehyde vapor, again like SO₂, has been shown to be potentiated by water-soluble salt aerosols. Irritancy appears to be augmented in proportion to the aerosol concentration, but the potentiation could not be accounted for by a simple aerosol “carrier” effect (Amdur, 1960). Moreover, reversal of bronchoconstriction was slower than had been observed with SO₂. Thus, it appeared that the vapor-aerosol itself constituted a new irritant species, the product of a chemical transformation of formaldehyde—perhaps methylene hydroxide. In addition to interactions with water-soluble particles, formaldehyde has been shown to interact with carbon-based particles (Jakab, 1992) to augment bacterial infectivity in a murine model. In this case, the potentiation appears to correlate with the surface carrying capacity of the inhaled particle.

Two aspects of formaldehyde toxicology have brought it from relative obscurity to the forefront of attention in recent years. One is its near ubiquitous presence in indoor atmospheres as an off-gassed product of construction materials such as plywood, furniture, or improperly polymerized urea-formaldehyde foam insulation (Spengler and Sexton, 1983). Complaints of formaldehyde irritation in industry have been reported at 50 ppb (Horvath et al., 1988). In studies relating household formaldehyde to chronic effects, children were found to have significantly lower peak expiratory flow rates (about 22% in homes with 60 ppb) than did unexposed children and asthmatic children were affected below 50 ppb. Thus, this irritant vapor can cause respiratory effects, and perhaps act as an allergen, at commonly experienced exposure levels (Krzyzanowski et al., 1990). Also, there is epidemiological evidence, sometimes inconsistent, that formaldehyde is associated with asthma (McGwin, 2010) and lower respiratory tract infections in children (Roda et al., 2011).

A longtime concern regarding formaldehyde has been its potential carcinogenicity. In 2004, IARC concluded, based on a thorough review of the published data, that formaldehyde was a probable human carcinogen; IARC deemed that there was “sufficient” epidemiological evidence that formaldehyde causes nasopharyngeal cancer, “strong but not sufficient” evidence of leukemia, and limited evidence of sinonasal cancer, along with “sufficient” evidence that formaldehyde causes nasal cancer in animals (Cogliano, 2005). Nasal cancer had been induced empirically with formaldehyde vapor in a two-year study where rats were exposed to 2, 6, or 14 ppm six hours per day, five days per week. The incidence of nasal squamous cell carcinomas was zero in the control and 2-ppm groups, 1% in the 6-ppm group, and 44% in the 14-ppm group. Mice were much less sensitive; only one carcinoma was seen at 14 ppm. The detection of DNA adducts in the two species paralleled the difference in the incidence of tumors as well as regional dosimetry. Formaldehyde, with its large and diverse database and potential public health impact, has remained the focus of considerable debate among modelers and risk assessors. The arguments behind this debate crosses both cancer and noncancer considerations and is beyond that which can be discussed in this chapter. The reader is encouraged to sample the recent literature (eg, Conolly et al., 2004; Nielsen et al., 2010).

Acrolein

Because acrolein is an unsaturated aldehyde, it is more reactive than formaldehyde. It penetrates a bit deeper into the airways and may not have the same degree of sensory irritancy but it may cause more damage. Concentrations below 1 ppm cause irritation of the eyes and the mucous membranes of the respiratory tract. Exposure of guinea pigs to <0.6 ppm reversibly increased pulmonary flow resistance and tidal volume and decreased respiratory frequency (Murphy et al., 1963). With irritants of this type, flow resistance is increased by concentrations below those that cause the classic decrease in frequency seen with sensory irritants. This suggests that increases in flow resistance would be produced by far lower concentrations of acrolein than were tested. The mechanism of increased resistance appears to be mediated through both a local C-fiber and centrally mediated cholinergic reflexes (Bessac and Jordt, 2010). Ablation of the C-fiber network and atropine (muscarinic blocker) block this response.

Exposures of rats to 0.4, 1.4, or 4.0 ppm for six hours per day, five days per week for 13 weeks resulted in paradoxical effects on lung function (Costa et al., 1986). The lowest concentration resulted in hyperinflation of the lung with an apparent reduction in small-airway flow resistance, while the highest concentration resulted in airway injury and peribronchial inflammation and fibrosis. The intermediate concentration was functionally not different from the control, although airway pathology was evident. It appears that the high-concentration response reflected the cumulative irritant injury and remodeling as a result of the repeated acrolein, while the low-concentration group had little overt damage and appeared to have slightly stiffened airways, perhaps a result of the protein cross-linking action of acrolein. The pathology in these rats contrasts with that found in formaldehyde studies of similar duration where more upper airway involvement was observed. Ambient exposure to acrolein probably would be about 10% to 20% of the low concentration used in the subchronic study discussed above. Because of a lack of sufficient human data, these animal data were used to perform a risk assessment analysis that showed an excess risk to humans for the adverse effect of ambient levels of acrolein on pulmonary function (Woodruff et al., 2007). However, ambient concentrations of acrolein are well below those found in mainstream tobacco smoke and the occupational exposure levels. As reviewed elsewhere (Bein and Leikauf, 2011), acrolein produces a wide range of adverse pulmonary effects including lung cancer that may be linked mechanistically to p53 (a tumor suppressor gene) DNA adducts and mutations.

Thus, as a class the aldehydes can be very irritating and may constitute a significant fraction of the discomfort and sensation experienced during an oxidant pollution episode, especially in mixed atmospheres containing particles.

Carbon Monoxide

Carbon monoxide is classed toxicologically as a chemical asphyxiant because its toxic action stems from its formation of carboxyhemoglobin, preventing oxygenation of the blood for systemic transport. The fundamental toxicology of CO and the physiologic factors that determine the level of carboxyhemoglobin attained in the blood at various atmospheric concentrations of carbon monoxide are detailed in Chap. 11.

The normal concentration of carboxyhemoglobin (COHb) in the blood of nonsmokers is about 0.5%. This is attributed to endogenous production of CO from heme catabolism. Blood COHb is a function of the concentration in air, the length of exposure, and the ventilation of the exposed individual. Uptake is said to be ventilation-limited, implying that virtually all the CO inspired in a breath is absorbed and bound to the available hemoglobin. Thus, continuous exposure of human subjects to 30 ppm CO leads to an equilibrium value of 5% COHb. The Haldane equation is used to compute the COHb equilibrium under a given exposure situation. The equilibrium values generally are reached after eight hours or more of exposure, but the time required to reach equilibrium can be shortened by physical activity.

Carbon monoxide emissions from automobiles have progressively decreased since the 1970s. Increased combustion efficiency and catalytic converters have reduced emissions at the tailpipe by more than 90%. However, motor vehicles still account for two thirds of urban CO. Depending on the location in a community, traffic density, vehicle types and age, and the urban structure, CO concentrations can vary widely. Concentrations predicted inside passenger compartments of motor vehicles in downtown traffic were almost three times those for central urban areas and five times those expected in residential areas. Occupants of vehicles traveling on expressways had CO exposures somewhere between those in central urban areas and those in downtown traffic and concentrations in underground garages, tunnels, and buildings over highways can still reach high levels. But it must be kept in mind that there are many sources of CO—anywhere combustion is ongoing. Certainly both main and sidestream tobacco smoke are a source. Home heating systems and mobile auxiliary heating sources emit CO that when used inappropriately (ie, unventilated) can create life-threatening circumstances.

No overt clinical human health effects have been demonstrated for COHb levels below 2%, while levels above 40% cause fatal asphyxiation. However, there is evidence that in some people COHb in the range of 2.0% to 2.4% can elicit acute cardiovascular effects. A 90-minute exposure to about 50 ppm CO would likely result in COHb levels of ~2.5%, with a resultant impairment of time-interval discrimination even in healthy subjects; at approximately 5% COHb, there is an impairment of other psychomotor faculties. At 5% COHb in nonsmokers (the median COHb value for smokers is about 5%), however, maximal exercise duration and maximal oxygen consumption are reduced (Aronow, 1981). Clear cardiovascular changes also may be produced by exposures sufficient to yield COHb in excess of 5%. These include increased cardiac output, arteriovenous oxygen difference, and coronary blood flow in patients without coronary disease. Decreased coronary sinus blood PO₂ occurs in patients with coronary heart disease, and this would impair oxidative metabolism of the myocardium. In the early 1990s, a series of studies in subjects with cardiovascular disease was conducted in several laboratories under the sponsorship of the Health Effects Institute (HEI) to determine the potential for angina pectoris when they exercised moderately with COHb levels in the range of 2% to 6% (Allred et al., 1989). The results of these studies indicate that premature angina can occur under these conditions but that the potential for the induction of ventricular arrhythmias remains uncertain.

Carbon monoxide has been implicated as a potentially important copollutant of PM with potential human health effects at low ambient levels. Independent CO effects have been reported in some epidemiological studies of ambient air pollution, but its singular role versus indicator role (for traffic) continues to be debated. The fact that effects noted in these studies are cardiovascular and involve angina and effects on EKG patterns lends credence to a role for CO in epidemiological health outcomes. Other data from epidemiological studies have shown a linkage between CO and some general CNS indices as well as birth outcomes. The sensitivity of newer analysis methods for these population studies is revealing previously unappreciated health effects. Among these are preterm births, cardiac birth defects, and infant mortality in the postneonatal period. Moreover, toxicological studies with animals also suggest potential developmental risks.

Reducing CO emissions is a major goal because of its high hazard—a colorless and odorless nonirritant—so technical combustion engineering advances are being pursued as well as changes in fuel formulations. The introduction of gasoline oxygenates such as MTBE in the 1990s and other ether derivatives constitutes attempts to enhance fuel combustion and reduce CO emissions. The experience with MTBE (as noted earlier) has led to such formulations being approached with some caution with greater attention being placed on assessments of potential unexpected consequences.

Hazardous Air Pollutants

HAPs (so-called air toxics) represent an inclusive classification for air pollutants of anthropogenic origin that are generally of measurable quantity in the air, and are not covered in the Criteria Pollutant list. They are covered under Section 112 of the CAA. Selected regulatory issues of the HAPs were discussed above. Most exposure estimates for these pollutants are derived from emission inventories that are modeled into the National Air Toxics Assessment (www.epa.gov/ttn/atw/nata/sitemap.html). The most recent NATA assessment (2005) used 2005 data on 177 of the 188 HAPs, plus diesel PM. The diverse nature of even 33 of the 188 HAPs (the so-called dirty 30 noted above) complicates a general discussion of their toxicology because the group includes various classes of organic chemicals (by structure, eg, acrolein, benzene), minerals (eg, asbestos), polycyclic hydrocarbon particulate material (eg, benzo(a)pyrene), and various metals and metal compounds (eg, mercury, beryllium compounds) and pesticides (eg, carbaryl, parathion).

The focus to date on the HAPs has been on their potential carcinogenicity, as shown in chronic bioassays, mutagenicity tests in bacterial systems, structure–activity relationships, or—in a few special cases (eg, benzene, asbestos)—their known carcinogenicity in humans. These cancers need not be, and generally are not, pulmonary. Noncancer issues frequently relate to direct lung toxicants that, on fugitive emissions or accidental release, might risk those with preexisting diseases (eg, asthma) or might lead to chronic lung disease. The assessment of noncancer risk by air toxics to any organ system is based on the computation of long-term risk reference exposure concentrations (RfCs) to which individuals may be exposed over a lifetime without adverse, irreversible injury. This approach to HAP assessment is discussed in detail by Jarabek and Segal (1994). An analogous short-term RfC method has been developed for exposures up to 30 days. Of the noncancer risks deemed highest in this analysis, acrolein ranks at the top based on emission inventories, the potential for exposure, and its inherent irritancy. Discussion of this topic and an approach to mixed exposures to noncancer HAPs has been reviewed (Costa, 2004).

Accidental versus “Fence-Line” Exposures

The relationship between the effects associated with an accidental release of a large quantity of a volatile chemical into the air from a point source such as a chemical plant and the effects associated with a chronic low-level exposure over many years or a lifetime is not clear. With regard to cancer, which defaults to a linearized model of dose and effect (though some alternative models can be used if there are appropriate data), the issue is fairly straightforward. Any exposure must be minimized if not eliminated if cancer risk is to be kept as close to zero as possible. With noncancer risks, the roles of nonspecific or specific host defenses, thresholds of response, and repair and recovery after exposure complicate the assessment of risk. In large part, the issue here relates to C × T. Can we better relate disease or injury to cumulative dose or peak concentration for protracted exposures? Is there an exposure peak beyond which a cumulative approach fails (ie, the effect is concentration-driven), or is concentration always the dominant determinant? Many of these questions have yet to be answered, not to mention their specificity with regard to individual compounds and tissues affected.

Methyl isocyanate provides a contrast between the effects of a large accidental release and those produced by cyclic or continuous small fugitive vapor releases. The reactive nature of methyl isocyanate with aqueous environments is of such magnitude that on inspiration, almost immediate mucous tissue corrosion can be perceived. The vapor undergoes hydrolysis within the mucous lining of the airways to generate hydrocyanic acid, which destroys the airway epithelium and causes acute bronchoconstriction and edema. The damage is immediately life-threatening at concentrations above 50 ppm; at 10 ppm, it is damaging in minutes. These concentrations are in the range of the dense vapor cloud that for several hours enshrouded the village of Bhopal bordering the Union Carbide pesticide plant. Studies in guinea pigs showed the immediate irritancy of this isocyanate, which in just a few minutes also resulted in significant pathology (Alarie et al., 1987). Rats exposed to 10 or 30 ppm for two hours also showed severe airway and parenchymal damage, which did not resolve in surviving rats; transient effects were seen at 3 ppm. Even six months after exposure, the airway and lung damage remained, having evolved into patchy, mostly peribronchial fibrosis with associated functional impairments (Stevens et al., 1987). There was also cardiac involvement secondary to the damage to the pulmonary parenchyma and arterial bed. As a result, there was pulmonary hypertension and right-sided heart hypertrophy. This same spectrum of health effects has resulted in disability and deaths of thousands of initial survivors since the incident.

In the United States, methyl isocyanate has been measured in Katawba Valley, Texas, as a result of small but virtually continual fugitive releases of the vapor into the community air (“fence-line”) from an adjoining region with several chemical plants. While these levels of methyl isocyanate are not sufficient to cause the damage seen in Bhopal, there is concern that low-level exposure over many years may have more diffuse, chronic effects. Residents complain of odors and a higher frequency of respiratory disorders, but clear evidence of injury or disease is lacking.

Phosgene is best known for its use as a war gas, but it is also a common intermediate reactant used in the chemical industry, particularly in pesticide formulation. It is also a constituent of photochemical smog. Because of its direct pulmonary reactivity, it lends itself to use as a model pulmonary toxicant for studies addressing C × T relationships. These studies suggest that there may be a threshold below which compensatory and other bodily defenses (eg, antioxidants) may be able to cope with long-term low-level exposure (tolerance). For phosgene, this appears to be at or below the current threshold limit value of 0.1 ppm for eight hours. At higher concentrations, however, concentration appears to be the primary determinant of injury or disease regardless of duration. Thus, even though there is some adaptation with time, there continues to be a concentration-driven response that exceeds that predicted by C × T. This relationship appears to be different from that of O₃ at ambient levels, which can be approximated acutely by the C × T paradigm.

The pollutants in the atmosphere of any community vary considerably in space and time, and are charged by the varied output from a wide range of sources, only to be transformed stoichiometrically by a patterned intensity of sunlight. These complexities have been daunting to toxicologists for decades and have prompted scientists and risk assessors to adopt a reductionist approach—nominally one pollutant at a time. While the reductionist approach has been very successful in diminishing pollutants of primary concern and improving public health, it is widely thought that there are likely chemical and physiologic interactions between and among pollutants that are of public health consequence and have not been appreciated. Indeed, interactions are more likely to be apparent as exposure concentrations decrease, and while we might expect synergistic interactions as found with some particles and SO₂, it may be that antagonisms are prevalent. As the science and the statistical methods to look at interactions have improved, there is renewed interest in multipollutant toxicology.

Recent publications are revisiting this issue (Mauderly and Samet, 2009; NRC Report on Air Quality, 2004), where scientists from all sectors are trying to determine if there are approaches to studying interactions (Dominici et al., 2010; Vedal and Kaufman, 2011), synthesizing or modeling findings (Johns et al., 2012), and then translating models into practical approaches to reducing the most important mixtures and components and thereby health risk (Fann et al., 2011). There is yet to be any consensus on any of these aspects of multipollutant science other than this is the appropriate time to reinitiate the discussion. Early toxicology studies used complex mixtures, some of which were photochemically altered (discussed earlier), and epidemiology found that different mixtures appeared to vary in their health impacts, only to abandon the task. But the revival of the PM story, and that it is itself a mixture, and that progress has been made in understanding its toxicity and some of the key attributes that drive the PM mixture, suggest that the challenge to reengage the issue is reasonable and likely marks the next dimension of air pollution toxicology. While single pollutant approaches to air pollution research, health assessments, and ambient air quality standard setting have been successful in reducing air pollution over the past few decades, there is a clear need for parallel efforts within both the scientific and the regulatory/policy communities to advance methods for evaluating and managing the effects of air pollution in a multipollutant manner.

In writing this textbook chapter on air pollution toxicology, the authors’ goal has been to relate empirical studies in animals to phenomena known to occur in humans through epidemiological or controlled clinical study. The breadth and complexity of the problem of air pollution—from the development of credible databases to supporting regulatory action and decision making—has been the theme throughout. The classic and still most important air pollutants provide a foundation for understanding and appreciating the nuances of the issues and strategies for air pollution control and protection of public health. The key role of the toxicologist is to develop sensitive methods to assay responses to low pollutant concentrations, apply these methods to relevant exposure scenarios and test species, and develop paradigms to relate empirical toxicological data to real life through an understanding of mechanism. Last, the toxicologist must continually integrate laboratory data with those of epidemiology and clinical study to ensure their maximum utility.