Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 1 Are All Air Pollution Particles Equal? – How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health George Thurston, Yewande Awe, Bart Ostro, Ernesto Sánchez-Triana © 2021 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 202-473-1000 www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemina- tion of its knowledge, this work may be reproduced, in whole or in part, for noncommercial pur- poses as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Cover photo Tatiana Grozetskaya / Shutterstock Sample credit for a non-WB image © Ami Vitale / Panos Pictures. Used with the permission of Ami Vitale / Panos Pictures. Further permission required for reuse. Sample credit for a WB image © Curt Carnemark / World Bank. Further permission required for reuse. Report design Spaeth Hill Are All Air Pollution Particles Equal? – How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health George Thurston, Yewande Awe, Bart Ostro, Ernesto Sánchez-Triana Abbreviations Al aluminum OC organic carbon ACS American Cancer Society Pb lead BC black carbon PM Particulate Matter—A mixture of solid particles and liquid droplets Br bromine found In the air Ca calcium PM2.5 Particulate Matter equal to or less than 2.5 microns in diameter Cl chlorine PM10 Particulate Matter equal to or Cu copper less than 10 microns in diameter CVD cardiovascular disease RSD respiratory disease EC elemental carbon S sulfur EDV emergency department visit Se selenium Fe iron Si silicon IHD ischemic heart disease SO42- sulfate K potassium Ti titanium LMICs low- and middle-income V vanadium countries Zn zinc LUR land use regression Mg magnesium Mn manganese µg/m3 micrograms per cubic meter Na sodium NH4+ ammonium Ni nickel NO2 nitrogen dioxide NO3- nitrate Table of Contents PREFACE 9 ABOUT THIS WORK 11 ACKNOWLEDGMENTS 13 ABOUT THE AUTHORS 15 I. EXECUTIVE SUMMARY 19 II. CONTEXTUAL BACKGROUND AND OBJECTIVE 23 III. INTRODUCTION 25 IV. CHALLENGES AND UNCERTAINTIES 29 V. METHODS 33 VI. SHORT-TERM EXPOSURE 35  ssociations of Short-Term Exposure to PM2.5 A. A Source-Related Components, and/or Their Constituent Tracers, with Daily Mortality 35 Short-Term Exposure to PM2.5 Sources and B.  Constituents and Their Associations with Daily Morbidity 39 C. Summary of Short-Term Studies 42 VII. LONG-TERM EXPOSURE TO PM2.5 SOURCES AND CONSTITUENTS AND THEIR ASSOCIATIONS WITH MORTALITY 47 VIII. SUMMARY AND CONCLUSIONS 55 APPENDIX 58 REFERENCES 69 Tables TABLE 1 ASSOCIATIONS OF ADVERSE HEALTH EFFECTS WITH ACUTE EXPOSURES TO PM2.5 CONSTITUENTS 58 TABLE 2 ASSOCIATIONS OF ADVERSE HEALTH EFFECTS WITH ACUTE EXPOSURES TO PM2.5 SOURCE COMPONENTS 62 TABLE 3 ASSOCIATIONS OF LONG-TERM EXPOSURE TO PM2.5 CONSTITUENTS WITH MORTALITY 64 TABLE 4 ASSOCIATIONS OF LONG-TERM EXPOSURE TO PM2.5 SOURCES WITH MORTALITY 66 World Bank Group 8 9 Preface Improving the air we breathe is quite literally an endeavor of global proportions, impor- tance, and urgency. Across the world, outdoor air pollution kills an estimated 4.2 million people a year and harms millions more, makes cities less livable, hampers economic competitiveness, harms ecosystems, and disproportionally affects the most vulnerable. Decades of global studies have shown that air pollution made of inhalable, microscopic particles known as fine particulate matter, or PM2.5, is especially harmful to humans. Most studies on the health impacts of the particles have focused on their mass, with the as- sumption that particles of the same mass are equally harmful, regardless of their source. Yet PM2.5 can come from a variety of sources - such as vehicle exhaust, road dust, indus- trial smokestacks, and biomass burning - and can be made of different chemical constit- uents. Evolving evidence shows that the danger to humans can actually vary depending on the source and chemical composition of PM2.5. Furthermore, particles in low- and middle-income countries usually have very different sources and compositions from par- ticles in high-income countries. Therefore, health effects per unit mass of PM2.5 are likely different in low- and middle-income countries than those in high-income countries, which form the basis of present global and regional assessments of health impacts. This report offers a comprehensive review of the health effects of short- and long-term exposure to the different sources and constituents of outdoor PM2.5 air pollution. The report finds that PM2.5 mass and chemical constituents from fossil fuel combustion are some of the most toxic and damaging to health. The health risk is particularly great from coal combustion- and traffic-related PM2.5 particles, which are most consistently associated with cardiovascular mortality, especially from heart attacks. These findings are supported by links between hospitalizations and mortality and chemical constituents, notably sulfate and elemental carbon, that are important markers of coal burning and diesel-fueled traffic respectively. It is envisaged that these important findings will bring a new focus to the fight against air pollution in developing countries. Because reducing PM2.5 emissions from fossil fuel combustion, such as sulfur-emitting coal-fired power plants and diesel vehicles, can produce the most significant health benefits, countries’ air pollution control efforts should prioritize these sources. Such efforts could also help to mitigate climate change, as these same fossil fuel combustion sources contribute significantly to global climate pollution. World Bank Group 10 11 About This Work The pioneering analytical work in this report builds on a growing body of evidence that the World Bank is building to inform more effective and efficient pollution management interventions and harness the transition to a circular economy in low- and middle-income countries. This growing body of evidence focuses on strengthening the knowledge base that will prompt dedicated action to tackle the forms of pollution that cause the most significant health and social costs in low-and middle-income countries. It also advances inter-disciplinary approaches to assess the linkages between pollution management and circular economy and the World Bank’s dual goals of eradicating poverty and promoting shared prosperity. Recent contributions from this body of work include: (i) the monetary valuation of the global cost of mortality and morbidity caused by exposure to ambient fine particulate matter air pollution; (ii) bolstering the case for establishment and strengthening of ground-level air quality monitoring networks in low- and middle-income countries; (iii) the development of a systematic framework to support analysis of health impacts from land-based pollution; and (iv) economic and financial instruments to support the transition to circular economy. World Bank Group 12 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 13 Acknowledgments This report was prepared by a team led by Yewande Awe, with the core team consisting of George Thurston (New York University School of Medicine), Bart Ostro (University of California, Davis), and Ernesto Sánchez-Triana (World Bank). Dr. Thurston and Dr. Ostro prepared the background report. The contribution of Dr. Chris Lim (Yale University) to the preparation of this report is appreciated. The team would like to acknowledge, with thanks, the valuable advice and inputs of the peer reviewers: Jiang Ru, Martin Ochoa, Tamer Rabie, and Stephen Dorey. This report also benefitted from comments and other inputs provided by the following colleagues: Fernando Loayza, Marcelo Bortman, Momoe Kanada, and Hocine Chalal The contributions of Phil Dickerson (US Environmental Protection Agency AirNow program) and Joanne Green (Ricardo Energy, formerly of The Clean Air Institute) are also appreciated. This report is a product of the Environmental, Natural Resources and Blue Economy Global Practice of the World Bank. This work was conducted under the supervision of Juergen Voegele (Vice President, GGSVP), Karin Kemper (Global Director, SENDR), Julia Bucknall (Global Director, SESD2), Benoit Bosquet (Regional Director, SEADR), Benoit Blarel (Lead Environment Specialist, SENDR), Iain Shuker (Acting Practice Manager, SAEE2), and Christian Albert Peter (Practice Manager, SENGL). The financial support of the World Bank-administered Pollution Management and Environmental Health multi-donor trust fund in the preparation of this report is gratefully acknowledged. World Bank Group 14 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 15 About the Authors George Thurston’s research has focused on the human health effects of air pollution, especially regarding fine particulate matter (PM2.5). It has involved studies of air pollution exposure and their health effects in panels of individual human subjects, as well as large cohorts. In 1987, he published the first research documenting the association between PM2.5 mass and its components with mortality. Recent awards include the 2012 “Haagen Smit Prize” from Atmospheric Environment. Dr. Thurston has also been a leader in bringing scientists and physicians together to reach consensus on key issues. He led the recent American Thoracic Society (ATS) and European Respiratory Society (ERS) Statement on “What Constitutes an Adverse Effect of Air Pollution” (2017), and received the 2017 American Thoracic Society’s Public Service Award. He has also been a pioneer researching the human health co-benefits of climate change mitigation, first publishing in Science on the issue in 2001. Yewande Awe is a Senior Environmental Engineer at the World Bank where she also holds a leadership position in the Bank’s Air Quality Management Community of Practice. She has led and been team member, in the preparation and supervision of policy-based programs, investment projects, technical assistance operations, and analytical activities in several countries. She has worked on projects in Antigua and Barbuda, Argentina, Bolivia, Bosnia and Herzegovina, Kosovo and North Macedonia, Colombia, Dominica, El Salvador, Grenada, Guatemala, Jamaica, Mexico, Nigeria, Peru, Poland, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, and Yemen. She has authored several publications, in peer reviewed scientific journals, on environmental engineering, environ- mental health and emerging strategies for environmental quality monitoring in low- and middle-income countries. She was Scientific Adviser for the First Global Conference on Air Pollution and Health, and a member of the Report Steering Committee for the Lancet Commission on Pollution and Health. She sits on the Board of Directors for Pearson College UWC in Canada, a pre-university school in a movement of 18 global schools dedicated to uniting cultures and countries around the world through education. Awe is involved with Women in Technology International, a global organization dedicated to empowering innovators and building inclusive cultures. Nominated by her colleagues, she serves as Respectful Work- place Advisor at the World Bank. Awe holds a PhD in Environmental Engineering from Imperial College London, United Kingdom. Bart Ostro was the Chief of the Air Pollution Epidemiology Section, California Environ- mental Protection Agency (CalEPA) for almost 25 years and is now a Research Professor at the University of California at Davis. His primary responsibilities at CalEPA included formulating the Agency’s recommendations for state air pollution standards and inves- tigating the health effects of air pollution and climate change. Dr. Ostro was a co-author of the U.S. EPA Regulatory Impact Analysis that was a basis for the federal ban of lead in gasoline. He chaired the U.S. EPA’s Science Advisory Board Clean Air Council, and served World Bank Group 16 on the National Academy of Sciences’ Committee on Estimating the Benefits of Proposed Air Pollution Regulations. He has published over 100 scientific papers and has consulted for several federal and international institutions and governments. He was honored with Special Achievement Awards from the U.S. EPA and the California Air Resources Board. In 2005 he was given the American Lung Association of California’s Clean Air Award, presented to the individual who “has made the most significant contribution to the cause of cleaner air in California.” Ernesto Sánchez-Triana is the Global Lead for Pollution Management and Circular Econ- omy for the World Bank. He has worked on projects in numerous countries, including, among others, Afghanistan, Argentina, Bangladesh, Bhutan, Bolivia, Brazil, Ecuador, India, the Lao People’s Democratic Republic, Mexico, Pakistan, Panama, Paraguay, and Peru. Before joining the World Bank, he taught at Colombia’s National University and worked for the Inter-American Development Bank. He also served as Director of Environmental Policy at Colombia’s National Department of Planning and as President of the Board of Directors of the Cundinamarca Environmental Protection Agency. Dr. Sánchez-Triana has led the preparation of numerous policy-based programs, invest- ment projects, technical assistance operations, and analytical works. He holds an engineer degree from Universidad de Los Andes (Colombia), and two master of science degrees and a PhD from Stanford University. He has authored numerous publications on air quality management, air quality monitoring, circular economy, cleaner production, climate change, country environmental analyses, energy efficiency, energy policy, environmental economics, environmental sustainability, environmental health, environmental impact assessment, environmental management, industrial sustainable growth, green growth, land-based pollution, organizational learning, policy strategic environmental assessments, sustainable development, and water resources management. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 17 Executive Summary 18 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 19 I. Executive Summary • Global studies of the health impacts of fine particulate matter (PM2.5) have been based on particle mass, but there is evolving evidence indicating that adverse health effects can vary depending on the source and composition of PM2.5. This complicates the equitoxicity assumption commonly made regarding particle air pollution while raising the question as to which types of PM2.5 are most toxic. • This report provides a comprehensive review of health effects of short- and long-term exposures to source-related components and trace constituents (specific markers of the individual sources) of fine particulate matter (PM2.5) in outdoor air pollution. The analysis evaluates the data and methods used, as well as the uncertainties in the underlying epidemiological studies, based on the relevant published literature.1 • Most deaths attributed to outdoor PM2.5 air pollution (for example, by the Global Burden of Disease study) are caused by cardiovascular disease (CVD). The current ev- idence summarized in this report indicates that trace constituents from PM2.5 and PM2.5 mass from fossil-fuel combustion are among the greatest contributors to PM2.5 toxicity. Notably, PM2.5 from fossil-fuel combustion poses a larger cardiovascular disease risk per unit mass of PM2.5 than soil or biomass particles. • Of the fossil-fuel combustion particles, coal- and traffic-related PM2.5 were found to be most consistently associated with cardiovascular mortality, especially due to ischemic heart disease (heart attacks) in both the short- and long-term exposure studies that were reviewed. • The importance of coal combustion and traffic-emission sources to the overall associations of PM2.5 mass with mortality and hospital admissions is supported by the health risk associations that are also observed with their source-specific trace constituents. For example, particulate selenium and sulfur are important markers for coal combustion. Particulate sulfur compounds in the environment are usually present as ammonium sulfate, ammonium bisulfate, or sulfuric acid. Similarly, elemental carbon (EC) is a strong marker for traffic, especially for diesel-fueled vehicle emissions. Based on the results of several recent epidemiological and toxicological studies, some researchers reported results indicating that particulate sulfur is among the most, if not the most, important constituent of PM2.5 associated with adverse health effects such as additional hospital admissions and mortality. • The long-term studies reviewed in this report indicate that the cardiovascular disease risks of sulfate, EC, and PM2.5 from coal combustion are larger than that from PM2.5 mass, in general. Furthermore, a fairly consistent association between sulfate and cardiovascular mortality is indicated. 1 This report covers literature published through October 2018.  Executive Summary 20 • The apparently central role of particulate sulfur in these PM2.5-health associations may not be due to the direct adverse effects of sulfur but more likely to its usual occurrence in intimate mixtures with other compounds concentrated in fossil-fuel combustion particles, including transition metals that can cause systemic oxidative stress in the human body. Oxidative stress is a known predictor of cardiovascular disease. The acidic nature of particulate sulfur compounds in the environment (for example, sulfuric acid) has been shown to make such transition metals more bioavailable, enhancing the potential of sulfur-containing particles to cause oxidative stress and serious adverse health effects throughout the human body. • However, it is possible that since particulate sulfur is more homogeneous spatially (that is, more equally dispersed over a wider area) than constituents of primary PM2.5, such as EC, measurement of exposure to particulate sulfur might be more accurately obtained by a limited set of monitors (or a single monitor) in a metropolitan area than more spatially variable constituents of PM2.5. Such greater accuracy of measurement would lead, in general, to a greater likelihood of observing an association. • PM2.5 mass, its source-specific mass components, and its elemental constituents are less consistently associated with respiratory disease (RSD) outcomes than with cardio- vascular disease outcomes in general, except for short-term exposures to PM2.5 from biomass combustion. • In general, weaker associations were found between mortality and concentrations of PM2.5 from biomass combustion or crustal sources (for example, windblown soil). These two sources are extremely varied spatially and temporally, especially for esti- mating long-term exposure. A resulting population exposure misclassification could contribute to the lack of detection of their health associations. In addition, there are fewer US metropolitan areas studied that have significant concentrations of biomass and soil sources, which may limit the ability of national-scale studies to identify risks of these two sources of PM2.5. On the other hand, emissions from traffic and their markers (for example, EC) are also quite spatially variable within a metropolitan area (for example, near versus far from roadways), causing exposure misclassification, but still show consistent associations with adverse health effects across studies. • Particles in low- and middle-income countries (LMICs) usually have very different sources and compositions from particles in high-income countries. Therefore, health effects per unit mass of PM2.5 are likely different in LMICs than those in high-income countries, which form the basis of present global and regional assessments of health impacts. • Overall, this report’s key finding from the review of currently available acute (or short-term) and long-term studies suggest that reductions in the emissions of PM2.5 from fossil-fuel combustion, such as sulfur-emitting coal-fired power plants and diesel vehicles, can produce the most significant health benefits per microgram per cubic meter (µg/m3) of PM2.5 reduced. In addition, these PM2.5 emissions reductions would provide additional climate mitigation benefits, since these same fossil-fuel combustion sources contribute significantly to global climate pollution, most notably in the form of CO2 and elemental carbon emissions. • The relative lack of studies of the health effects of PM2.5 constituents in the published literature for LMICs supports the need for these countries to begin measuring the Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 21 concentrations of the source-marker constituents of their particulate mixes (for exam- ple, arsenic for coal combustion and vanadium for oil combustion). Particulate sulfur and EC, which can be measured continuously, are especially important combustion markers to monitor more widely than at present. Such efforts will help determine the contribution of specific sources to PM2.5 mass and help prioritize subsequent pollution-control efforts. • There is also a great need to conduct short- and long-term epidemiological studies of mortality and hospitalization in LMICs, where biomass burning for cooking and heating is a significant source of PM2.5 but a less-well quantified human health risk relative to fossil-fuel-combustion PM2.5. In order to develop economically efficient abatement strategies that maximize human health benefits per ton of PM2.5 emissions reduced, it is essential to (i) measure the concentration of PM2.5 from the specific source (in this case, biomass burning), or of a marker of the source (for example levoglucosan), and (ii) determine the specific toxicities of the various PM2.5 contributions from specific sources in LMICs. • Ambient air pollution–control efforts in LMICs need to account for both the contrib- uting sources of PM2.5 and the toxicity of the PM2.5 from each source category. The strongest evidence to date of increased risk of cardiovascular disease health effects per unit mass of PM2.5 is found from coal combustion and diesel-fueled traffic. Therefore, future mitigation efforts that focus on monitoring and regulating the constituents of PM2.5 from these sources should be prioritized. Reducing pollution from these sources can be expected to return greater cardiovascular health benefits per unit mass of PM2.5 reduced than if PM2.5 mass continues to be addressed equally, irrespective of source and composition. “The strongest evidence to date of increased risk of cardiovascular disease health effects per unit mass of PM2.5 is found from coal combustion and diesel-fueled traffic. Therefore, future mitigation efforts that focus on monitoring and regulating the constituents of PM2.5 from these sources should be prioritized.” Contextual Background and Objective 22 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 23 II. Contextual Background and Objective This report was produced as part of a program of analytical work conducted under the framework of the World Bank’s multi-donor funded Pollution Management and Environ- mental Health (PMEH) Program. The overall aim of the work is to develop knowledge and guidance for the improvement of air quality monitoring and the estimation of health risks and effects of ambient air pollution in LMICs. One component of the analytical work focuses on improving the estimation of health impacts of ambient air pollution in LMICs. In this context, the analytical work aims to accumulate new evidence related to health impact estimation, particularly in LMICs. While exposure to fine particulate matter mass (PM2.5) has been associated with a wide range of adverse human health effects at multiple locations around the globe, the absolute size of the effect has varied from place to place and over time, presumably because the source and nature of the particles is known to also vary over time and place. This report therefore aims to enhance the understanding of the health ef- fects associated with specific constituents and sources of fine particles (PM2.5). Introduction 24 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 25 III. Introduction Over the past few decades, epidemiological studies have reported consistent asso- ciations2 between both short- and long-term exposures to fine particulate matter air pollution—that is, the mass concentration of particles ≤ 2.5 micrometers in diameter (PM2.5)—and mortality. In the air pollution studies, short-term refers to exposures of a day or multiple days while long-term refers to exposures of a year or more. Studies of short- term PM2.5 exposures to constituents and mortality have been reported for the European Union, North America, South America, the West Pacific, and Asia (Achilleos et al. 2017; Chen et al. 2017) while long-term exposure and mortality studies have been published mostly in North America and Western Europe (Atkinson et al. 2014; Brook et al. 2010; Chen et al. 2017; US EPA 2009). It is important to recognize, however, that ambient PM2.5 mass consists of a heteroge- neous mixture of solid and liquid particles generated by many sources (for example, cars and trucks, fossil-fuel combustion, industrial emissions, biomass, windblown dust), and, until recently, there has been only limited evidence as to which trace constituents or sources of PM2.5 mass are associated with the greatest risks. In this context, the term “constituents” refers to specific elements and compounds that are contained in PM2.5, while “components” are groupings of PM2.5 constituents that are of the same origin, es- pecially “source-related components” which are groupings of constituents that comprise the PM2.5 from a specific source category, such as coal combustion. Constituents that are fairly unique to a source component (for example, vanadium for residual oil combustion and arsenic for coal combustion particles) are termed “tracers” for those source catego- ries, as their presence indicates a PM2.5 mass contribution from the source in which they are enriched, relative to PM2.5 mass from other sources. The fact that the mix of PM2.5 mass source components and their respective constituent contributions to the mass vary over time and space also likely means that the toxicity of PM2.5 mass varies from place to place and over time as well. But this aspect of PM2.5 has been largely unaddressed by past PM2.5 mass studies that have most often treated all PM2.5 as the same, irrespective of the wide PM2.5 mass composition and source variations across places and time. The US National Research Council has highlighted the importance of investigating characteristics and constituents of particulate matter (PM) that contribute to its toxicity (National Research Council 2004). Indeed, information regarding the relative toxicity of PM2.5 constituents and/or source components would have important implications for establishing ambient air quality standards, targeting pollution-control strategies to 2 H  ere, the term “association” is defined as a statistical relationship between two or more events, characteris- tics, or other variables—for example, an association between exposure to X and a health effect, Y. A constit- uent X may be the toxic agent but is more likely serving as a marker for toxic pollution from a source and is not itself the specific causal agent. In this case, the constituent is referred to as being health-associated. Association, per se, does not imply causality but is a key element in the assessment of causality (Hill 1965) Introduction 26 ensure improved cost efficiency, and enhancing the accuracy of health impact assess- ments by focusing efforts on the most toxic constituents/components. Routine collection of speciated PM2.5 data (that is, when the mass sample is analyzed for its compositional constituents, such as metals and ions; for example, see Spengler and Thurston [1983]) in the US and some other countries now provides opportunities to examine relationships between specific PM constituents and both morbidity and mortality. In addition, several of the constituents are known to be tracers for specific PM2.5 sources, which enables an indication of their relative source-specific PM2.5 health impacts. Moreover, the constituents can be grouped together to quantitatively estimate source-specific PM2.5-component con- tributions to PM2.5 mass using source apportionment methods, providing an even more direct assessment of source-specific health impacts (Thurston et al. 2005).3 Below, the available epidemiological studies using the above-noted PM2.5-air-pollution source and constituent exposure metrics have been reviewed to determine these various metrics’ respective impacts on adverse health outcomes. In this discussion, these PM2.5 health studies are broken into two categories: (1) short-term acute studies (for example, considering risks of daily exposures to air pollution), and (2) long-term chronic studies (for example, considering risks of annual average exposures). The focus of the review is on the most severe adverse health effects that have been associated with PM2.5 mass: mortality and hospital admissions. These evaluations will allow insight into which source-specific PM2.5 components (or their trace constituent markers) may be driving the associations found in studies of long- and short-term PM2.5 exposure. These source components, therefore, are those that should become priorities for pollution-control strategies. “The fact that the mix of PM2.5 mass source components and their respective constituent contributions to the mass vary over time and space also likely means that the toxicity of PM2.5 mass varies from place to place and over time as well.” 3 S  ource apportionment is a statistical process by which the total PM2.5 mass is separated into components that are associated with a specific source or source class based on the intrinsic characteristics of PM from that source or source class. For example, before lead was removed from gasoline, it was a very reliable tracer of automobile combustion PM2.5 (Cooper and Watson 1980; US EPA 1984). This then allows these esti- mates of source-specific PM2.5 mass to also be considered in PM2.5 health effects analyses. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 27 Challenges and Uncertainties 28 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 29 IV. Challenges and Uncertainties There are many challenges that contribute to the difficulty of determining the relative harmfulness of the various constituents and sources of PM2.5. First, the risk associated with an individual constituent or source may be confounded by correlated (measured or unmeasured) co-pollutants4 or other PM2.5 sources. There are often strong correlations among the particle constituents and co-pollutants emitted from the same sources, complicating the separation of their respective roles in associations with health impacts. Strong statistical intercorrelations can reflect (1) a consistent chemical signature of multi- ple pollutants associated with PM emitted from individual major source categories, and/ or (2) the similarity of certain pollutant constituents derived from different sources, such as gasoline and diesel vehicles. For example, a high correlation (of between r = 0.5 and 0.8) among elemental constituents was reported for study participants located within 10 kilometers of a monitor in California (Ostro et al. 2010). Importantly, this concern can be addressed by the consideration of PM2.5 source components (comprising their respective source-specific groups of constituents) versus considering individual elemental constitu- ents (that is, ignoring their multiple sources). Second, there may be significant constituent- and source-specific PM2.5 differences in the ultimate population exposures, making it difficult to determine the underlying differences in toxicity. Specifically, different PM2.5 constituents and source components will likely vary in terms of the amount of time a person is exposed to each of them, since some are more spatially homogenous than others. In addition, other aspects of exposure (for example, indoor versus outdoor, exercising versus resting, proximity to source) will impact the risks associated with a given source in any given study. Thus, even if two components were equally harmful on a per-unit basis, epidemiological studies might still indicate differenc- es in harm, due to their exposure differences (for example, in spatial variability). Third, comparing the health effects of PM2.5 source-related components over multiple cit- ies may be difficult since the constituents that make up the source can differ by location across studies. However, in practice, this has not been found to be a major issue, as the source-specific characteristics are relatively stable even though their contributions to the total mass can vary greatly (Thurston et al. 2011). 4 C  o-pollutants are defined here as other, usually gaseous, air pollutants (for example, nitrogen oxides, NOx) that may be emitted by the same sources (for example, motor vehicles) as specific PM2.5 constituents (for ex- ample, elemental black carbon from diesel vehicles). As a result, separating their individual health effects can be difficult. Similarly, co-exposure to other PM2.5 sources can complicate the attribution of effects to a single source. Thus, many studies are needed to discern their respective associations, as considered in this report. Challenges and Uncertainties 30 Fourth, the extent of association of a given tracer constituent or source-related compo- nent with a health effect outcome (for example, daily hospital admission counts) could be due to differences in the amount of their respective monitoring (for example, the number of stations, relative to their individual variability over time and space). Thus, differential misclassification of exposure among the source-component contributions (and/or their tracer constituent levels) can impact the subsequent risk estimates. Finally, for some sources, such as windblown soil and dust, the health effects may be more associated with particles larger than 2.5 micrometers in aerodynamic diameter. These health effects can include effects on the upper airways, such as the exacerbation of asthma, chronic bronchitis, or emphysema. However, despite these concerns, it has been found that the overall uncertainty contribut- ed by variability in the method of source apportionment of PM2.5 mass is small compared to the overall PM2.5-health-association uncertainty (Thurston et al. 2005). As a result of this consistency, certain clear patterns are discernible, and some reasonable conclusions can be made by reviewing the available studies. Below, a summary and review of the epi- demiological studies of health effects associations of both short- and long-term exposure to PM2.5 source-specific components (or their constituent tracers) is provided. The review is followed by the Summary and Conclusions section that discusses the overarching issues in these PM2.5 source and constituent associations, caveats, and overall findings. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 31 Methods 32 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 33 V. Methods The various studies of short-term exposure considered in this analysis are summarized in tables 1 and 2. In order to give an indication of the strength of the health risk found with the various PM2.5 constituents or source-specific components, the following designations were used in the tables: 0 if they were considered but were not statistically significant even at the p (or p-value) = 0.10 level; 0.5 if they were significant at the p = 0.10 level; and 1.0 if they were significant at the p = 0.05 level. The p-value indicates the likelihood of incorrectly rejecting the hypothesis of no observed association. If, over all studies in a category, a strength of association was found to be present 50 percent of the time or more (for example, ≥ 5 significance index in 10 cases), then the constituent/source was considered to be sufficiently consistent in its association with that health outcome. This provided a normalized index, allowing direct comparisons of the strength of associations across the various health outcomes and PM2.5 components. Tables 3 and 4 summarize the studies of associations of long-term exposure to PM2.5 constituents and/or source-specific PM2.5 with mortality only. There were too few long-term exposure studies published to date of morbidity outcomes with trace constituents or source-specific PM2.5 to provide the power to make meaningful health risk comparisons. The overall findings of each study considered are summarized below. Based on these review findings, as well as the overall semiquantitative evaluations of the studies in tables 1 through 4, conclusions are drawn regarding the sources and constituents most related to mortality and morbidity. Methods 34 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 35 VI. Short-Term Exposure Associations of Short-Term Exposure to PM2.5 A.  Source-Related Components, and/or Their Constituent Tracers, with Daily Mortality The first study identified that evaluated the acute (day-to-day) risks of PM2.5 constituents or source-specific PM2.5 with mortality used the Harvard Six Cities Study’s cohort data to evaluate the source-specific PM2.5 using a single trace element as a marker (or “target”) of that source across six US cities (Laden et al. 2000). The PM2.5 sources and their ele- mental tracers identified in all six cities were: crustal (silicon, Si, as tracer), mobile sources (traffic) emissions (lead, Pb, as tracer), and coal combustion (selenium, Se, as tracer). Of these, mobile sources and coal were statistically significant overall, while the crustal com- ponent (primarily from windblown soil) was consistently not associated overall with daily mortality or in any of the individual six cities considered. Overall, the authors concluded that combustion particles in the fine fraction from mobile and coal combustion sources, but not fine crustal particles, were associated with increased mortality. As part of a US Environmental Protection Agency (EPA) Workshop on Source Apportion- ment of PM Health Effects (Thurston et al. 2005), Mar et al. (2006) and Ito et al. (2006) published multianalyses of the impacts on daily mortality from exposure to PM2.5 sources in Phoenix, Arizona, and Washington, DC, respectively. In Phoenix, the sources with the largest cardiovascular disease (CVD) mortality risk size across analyses were secondary particulate matter5 (resulting from sulfate, SO42-) and traffic. For total mortality, the associations were weaker. Sea salt was also found to be associated with both total and cardiovascular mortality, but at a less biologically plausible five-day lag after exposure. Fine particle soil and biomass burning factors were not associated with increased risks in Phoenix. For Washington, DC, Ito et al. (2006) reported that the largest (and most significant) percent excess deaths per increment of PM2.5 sources for total mortality was for SO42-. Primary coal-related PM2.5 (identified by three workshop teams) was similarly significantly associated with total mortality with the same three-day lag as sulfate. Risk estimates for traffic-related PM2.5, while significant in some cases, were more variable. Soil-related PM showed smaller risks. Overall, these two cities both gave support for associations of daily mortality with the SO42- constituent, which is a result of fossil-fuel combustion. Of the sources considered by the various analyses, coal burning was most consistently associated with daily mortality. Ostro et al. (2007) and Ostro et al. (2008) published studies addressing daily data from 2000 to 2003 on mortality (all-cause, cardiovascular, respiratory, and mortality age > 65 years) and PM2.5 mass and its constituents in California. The constituents considered 5 S  econdary particulate matter is defined here as particles not directly emitted from a pollution source (for example, a power plant stack) as particles but instead formed in the atmosphere from gaseous air pollutants, such as sulfates, which are formed from sulfur dioxide gas. Sulfate and nitrate are often considered sources since they are not directly emitted, but they also could be considered constituents. Short-Term Exposure 36 included elemental and organic carbon (EC and OC), nitrates, sulfates, and various metals. PM2.5 mass and several constituents were found to be associated with multiple mortality categories, especially CVD deaths. For example, for a three-day lag, CVD deaths increased by 1.6 percent, 2.1 percent, 1.6 percent, and 1.5 percent for PM2.5, EC, OC, and nitrates, respectively. The authors concluded that their findings supported the hypothesis that combustion-associated pollutants are particularly important to PM2.5-associated mortality. Separate models were run after stratification by gender, race/ethnicity (White, Hispanic, Black), and education (high school graduation or not). After stratification, daily counts of cardiovascular mortality were associated with PM2.5 and several of its species including EC, OC, nitrates, sulfates, potassium, copper, and iron. For many of these spe- cies, there were significantly higher risk estimates among those with lower educational attainment and Hispanic individuals. Cakmak et al. (2009) evaluated the association between several elements of PM2.5 and mortality in the general population of Santiago, Chile, during the period 1998 to 2006. The strongest individual constituent mortality association was seen with elemental carbon. Using factor analysis, a group of elements consistent with a mobile combustion source (carbon monoxide, nitrogen dioxide, and elemental and organic carbon) was significantly associated with total mortality. Soil-sourced particles had a weaker but statistically significant association with mortality. Of the many sources examined, the authors concluded that particulate air pollution from motor vehicle exhaust had the greatest observed risk on mortality in this Chilean city. Zhou et al. (2010) used time-series analyses to determine the associations between daily fine-PM2.5 mass and its constituents with daily all-cause, CVD, and RSD mortality in two US cities: Seattle, Washington, and Detroit, Michigan. Total and CVD mortality in Detroit were most positively associated with warm-season PM2.5 secondary aerosols and traffic markers but negatively with soil PM2.5. In Seattle, total and CVD mortality were most closely associated with PM2.5 from traffic and other combustion sources, including residu- al oil and wood burning, during the cold season. Ito et al. (2011) examined the role of PM2.5 and its key chemical constituents on both CVD hospitalizations and mortality in New York City for the years 2000–2006. Coal combustion–related trace constituents (for example, selenium) were associated with CVD mortality in summer and CVD hospitalizations in winter, whereas elemental carbon and NO2 showed associations with these outcomes in both seasons. Ostro et al. (2011) examined the risks of various source-specific PM2.5 components and/ or their tracer constituents on daily mortality for years 2003 through 2007 in Barcelo- na, Spain. Source apportionment for PM2.5 and PM10 identified eight different sources. Case-crossover regression analysis6 was used to estimate the risks of each source - spe- cific component. There were significant associations found between road dust, minerals, fuel oil combustion, vehicle exhaust, and traffic with all-cause mortality and PM2.5. For cardiovascular mortality, associations were also observed for the sulfate/organics PM2.5 constituents. In some cases, relative risks were higher for specific sources than for total PM2.5 mass. 6  ase-crossover analysis is an alternative to time-series analyses in which a case’s (individual’s) exposure C immediately prior to or during the case-defining event (for example, a hospital admission) are compared with that same person’s exposure at other “reference” times. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 37 Son et al. (2012) investigated the daily association between daily mortality and PM2.5 mass and its constituents in Seoul, Republic of Korea, from August 2008 through Oc- tober 2009. Magnesium (Mg) concentration was associated with total mortality on the following day. Several constituents that were among the largest contributors to PM2.5 total mass—nitrate (NO3-), SO42-, and ammonium (NH4+)—were moderately associated with same-day cardiovascular mortality. Other constituents having smaller mass contributions (Mg and chlorine [Cl]) exhibited moderate associations with respiratory mortality on the following day. Cao et al. (2012) examined the short-term association between PM2.5 constituents and daily mortality in Xi’an, a heavily polluted Chinese city, from 2004 to 2008. After adjust- ment for PM2.5 mass, they found significant positive associations of total, cardiovascular, or respiratory mortality with OC, EC, ammonium, nitrate, chlorine, and nickel for at least one of the multiple lag days that were examined. Nitrate demonstrated stronger associa- tions with total and cardiovascular mortality than PM2.5 mass. The authors concluded that their findings added support to previously reported evidence of PM2.5-related health risks in China, suggesting that combustion-associated air pollutants are particularly important. Huang et al. (2012) also evaluated the association of daily PM2.5 and constituents with all- cause and cause-specific mortality data (2004–2008) in inland Xi’an, China. Secondary pollutants (sulfate and ammonium), combustion-related constituents (elemental carbon, sulfur, chlorine), and transition metals (chromium, lead, nickel, and zinc) appeared most responsible for increased risk from PM2.5 exposure, particularly in the cold months. In Xi’an, water-soluble and carbonaceous aerosols from coal combustion and vehicle emis- sions were found to be the dominant species in PM2.5 in the cold months, when health risk associations were the strongest in this study. The authors concluded that they observed a significant mortality association with exposure to PM2.5 and species originating from fossil-fuel combustion, vehicle emission, and industry process sources. Sacks et al. (2012) applied various regression models developed in previous multicity time-series studies of air pollution and mortality to data from Philadelphia, Pennsylvania (May 1992–September 1995). Single-pollutant analyses examined daily cardiovascular mortality associations with PM2.5 mass and constituents and gaseous pollutant data. Principal component analysis yielded source-related components for species associated with traffic, crustal material, residual oil, and coal. The traffic and crustal (that is, soil) source-specific components showed consistently positive associations with mortality) in the warm season, while the coal combustion factor showed consistently positive associa- tions with variations in daily mortality in the cold season. Krall et al. (2013) estimated short-term associations between daily mortality and PM2.5 constituents across 72 urban US communities from 2000 to 2005. Using US Environ- mental Protection Agency (EPA) Chemical Speciation Network data, they analyzed seven constituents that together composed 79–85 percent of PM2.5 mass: organic carbon matter (OCM), elemental carbon (EC), silicon, sodium ion, nitrate, ammonium, and sulfate. Results indicated that OCM, EC, silicon, and sodium ion were associated with estimated increases in mortality. Unlike PM2.5 mass associations with mortality in past studies, the Short-Term Exposure 38 authors did not find evidence that chemical constituent mortality risks varied by season or region across the nation, indicating that health risk associations with constituents are more stable than with PM2.5 mass. This result is consistent with the hypothesis that PM2.5 composition drives its toxicity, rather than mass, per se. Dai et al. (2014) estimated the risks of PM2.5 species on mortality and how air infiltration rates into buildings may modify the association. PM2.5 data for 75 mostly eastern and midwestern US cities from 2000 to 2006. These results were used in an analysis to determine which PM2.5 source-specific constituents contributed the most to the associa- tion of PM2.5 mass with mortality. Silicon (predominantly from windblown soil) and sulfur (predominantly from fossil-fuel combustion) were associated with both all-cause and CVD mortality, whereas only sulfur was related to respiratory deaths. Indeed, the authors observed that the risk of PM2.5 mass on all-cause and respiratory mortality was increased by an increased presence of sulfur in the PM2.5. Coal combustion emissions are a major contributor to sulfur in the US, which suggests that species derived from coal combustion might have greater toxicity, especially due to respiratory diseases (RSDs). Heo et al. (2014) evaluated the association of daily mortality counts in Seoul, Korea, with PM2.5 chemical speciation data collected every three days as well as nine sources of PM2.5 mass resolved through source apportionment using data from March 2003 through November 2007. Organic carbon, elemental carbon, and lead were associated with mor- tality outcomes when using multipollutant models adjusted for other chemical species levels. Source-specific PM2.5, including gasoline- and diesel-powered vehicles, as well as biomass burning, were associated with respiratory mortality and cardiovascular mortality, respectively. There were moderate associations of industry and roadway emissions with cardiovascular mortality. Li et al. (2015) investigated the association of both daily mortality and morbidity with particulate matter and its constituents in data collected in Beijing, China, from January 2005 through December 2009. Non-accident mortality was associated with potassium (K), sulfate (SO42-), calcium (Ca), and nitrates (NO3-). Similar findings were found for respiratory mortality, and sulfate was significantly associated with cardiovascular disease mortality. The authors concluded that combustion-related products, traffic sources, vegetative burning, and crustal component and resuspended road dust may play a key role in the associations between air pollution and public health in Beijing. Basagana et al. (2015) considered five cities in southern Europe: three cities in Spain, namely Barcelona (2003–2010), Madrid (2007–2008), and Huelva (2003–2010); and two cities in Italy, namely Rome (2005–2007) and Bologna (2011–2013). A case-cross- over design was used to link (i) cardiovascular and respiratory hospital admissions; (ii) total mortality, cardiovascular mortality and respiratory mortality, with (iii) 16 PM2.5 constituents. Most of the elemental constituents studied—namely EC (elemental carbon), SO42- (sulfate), Si (silicon), Ca (calcium), Fe (iron), Zn (zinc), Cu (copper), Ti (titanium), Mn (manganese), V (vanadium), and Ni (nickel)—were associated with increased changes in cardiovascular and/or respiratory hospitalizations. For mortality, although positive associations were identified (between both Fe and Ti and total mortality; between EC, Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 39 Mn, and Mg (magnesium) and cardiovascular mortality; and between NO3- and respiratory mortality), the patterns across outcomes were less clear. Given the noted variations in health associations for the different PM2.5 constituents, the authors recommended that more studies should include a focus on PM speciation associations with health. Kim et al. (2012) analyzed daily concentrations of PM2.5 constituents measured over five years in order to investigate associations with daily mortality. The mortality relative risks (RRs) estimated for EC and OC were larger than those for sulfate and nitrate. They gener- ally did not observe associations between constituents and cardiovascular or respiratory mortality but did find associations with ischemic heart disease mortality. In addition, there were PM2.5 constituent associations with cancer mortality, particularly by EC and OC. The authors concluded that PM2.5 constituents originating from combustion-related sources were more strongly associated with daily mortality than were secondary inorgan- ic aerosols. Short-Term Exposure to PM2.5 Sources and B.  Constituents and Their Associations with Daily Morbidity Sarnat et al. (2008) analyzed the association of emergency department (ED) visits (in Atlanta, Georgia, from November 1998 to December 2002) for combined CVD and respi- ratory causes with speciated PM2.5 data. Source-specific PM2.5 components were estimat- ed using several PM2.5 source apportionment approaches. All the methods indicated that PM2.5 from mobile sources showed the most consistent associations with both cardiovas- cular and respiratory EDVs. The same observation was true for PM2.5 from biomass burn- ing. In addition, there were consistent associations between sulfate-rich secondary PM2.5 and respiratory visits. Despite differences between the source apportionment methods, the consistent findings across methods confirmed that modeled source-apportioned data produce robust estimates of acute health risk, consistent with the findings by Thurston et al. (2005) and Thurston et al. (2011). Bell et al. (2009) linked two national datasets: (1) A US EPA database used to form long- term average concentrations of PM2.5 chemical constituents across the US for 2000–2005, and (2) daily cardiovascular and respiratory Medicare hospitalizations for persons 65 years or older in 106 US counties from 1999 through 2005. It was found that communities with a generally higher PM2.5 content of nickel, vanadium, and elemental carbon yielded a higher risk of hospitalizations associated with short-term exposure to PM2.5. Ito et al. (2011), as discussed above, examined the role of PM2.5 and its key chemical constituents on CVD hospitalizations and mortality in New York City for the years 2000–2006. Coal combustion–related constituents (for example, selenium) were asso- ciated with CVD hospitalizations in winter, whereas elemental carbon and NO2 (both major products of traffic-related emissions in urban areas) showed associations with CVD hospitalizations in both seasons. The authors noted that the PM2.5 constituents associated Short-Term Exposure 40 with the regional transported pollutants from coal combustion (Se and SO42-) exhibited the season-specific association pattern most similar to that of PM2.5 mass, suggesting that they are major contributors to the risks associated with PM2.5 mass. Kim et al. (2012) explored the associations of hospital admissions with daily PM2.5 chem- ical constituents for five years in the Denver Aerosol Sources and Health (DASH) study. Daily sulfate and nitrate constituents were more highly correlated with PM2.5 mass over time than were EC and OC. Estimated short-term associations of PM2.5 chemical constitu- ents, especially EC and OC, were immediate (that is, same day) for cardiovascular admis- sions. By contrast, estimated short-term associations of sulfate and nitrate were delayed in time for cardiac and respiratory admissions. Of all the subcategories of cardiovascular disease admissions, ischemic heart disease admissions showed immediate risks for EC and OC that paralleled the pattern observed for total cardiovascular disease admissions. Levy et al. (2012) performed a literature review and conducted a multisite time-series analysis of hospital admissions and exposure to PM2.5 constituents (elemental carbon, organic carbon matter, sulfate, and nitrate) in a population of 12 million US Medicare en- rollees for the period 2000–2008. The analysis found statistically significant associations between short-term changes in elemental carbon and cardiovascular hospital admissions. However, the analyses were limited to those constituents that dominate PM2.5 mass and are significantly correlated with PM2.5 (that is, EC, nitrate, sulfate, and OCM), limiting the usefulness of this study for evaluating source- or constituent-specific impacts. Mostofsky et al. (2012) analyzed data on 18 PM2.5 constituents and data from 1,060 patients admitted to a Boston, Massachusetts, medical center with ischemic stroke in 2003–2008. The authors utilized several options for modeling the association between constituents and health outcomes that account for the impact of PM2.5. The different methods yielded relative rankings of the association between constituents and ischemic stroke that were fairly consistent across models. Although many trace constituents (such as Se, Ni, Br, Pb) represent negligible contributions to total PM2.5, Mostofsky et al. detect- ed associations between several elemental constituents and stroke risk. This was thought to be because constituents with a small mass may nonetheless have high toxicity on their own, or (more likely) they are correlated with (that is, markers of) other co-occurring PM2.5 constituents or source-specific mixtures. Models considering constituent concentra- tions, after adjusting the model for the risks of PM2.5 mass (for example, simultaneously including PM2.5 mass as a model variable), indicated that higher levels of black carbon (BC), Ni, and V were associated with higher stroke risk. However, the authors noted that it was unclear whether adjustment for PM2.5 mass correctly accounts for confounding, or whether it inadvertently overadjusts for (that is, confounds the risks of) constituents that are highly correlated with PM2.5 mass, but that are also inherently toxic. In models in which the overall association of PM2.5 mass was not adjusted for in the model, the strongest associations with stroke risk were seen for black carbon (BC), nickel (Ni), and sulfur (S). Bell et al. (2013) examined pollutant sources for associations with risk of hospital admis- sions for cardiovascular and respiratory causes. The data included over 330,000 Medicare participants in four counties in Connecticut and Massachusetts from August 2000 to Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 41 February 2004. Source apportionment was used to estimate daily PM2.5 contributions from sources (traffic, road dust, oil combustion, and sea salt, as well as a regional source representing coal combustion and other sources). Associations of daily PM2.5 constituents and sources with cardiovascular and respiratory hospitalizations were estimated with time-series analyses. PM2.5 and PM2.5 road dust were associated with cardiovascular hospi- talizations, as were several PM2.5 constituents: calcium, black carbon (BC), vanadium, and zinc. PM2.5 from road dust and sea salt was associated with respiratory hospitalizations, as were the constituents aluminum, calcium, chlorine, black carbon, nickel, silicon, titanium, and vanadium. In the study area, motor vehicle emissions were a main contributor to ambient Zn and BC, road dust to Si and Al, oil combustion to V and Ni, sea salt to Cl, and regional sources to S. The authors concluded that, in the studied counties, the most harmful particles, based on their estimated risks, included: black carbon, calcium, and PM2.5 from the road dust source. As discussed above, Li et al. (2015) investigated the association of both daily mortality and morbidity with particulate matter and/or its constituents in Beijing, China, from 2005 through 2009. For morbidity, sulfate was associated with respiratory diseases, especially during the warm season. Jones et al. (2015) conducted a case-crossover analysis using daily average ambient PM2.5 total mass and species estimates derived from the Community Multiscale Air Quality (CMAQ) model and available ambient observations. They examined the association be- tween the chemical constituents of PM (including elemental and organic carbon, sulfate, nitrate, ammonium, and other remaining mass) and respiratory hospitalizations in New York State. Modification of the PM2.5-hospitalization association was assessed via models stratified by mass fractions of both primary and secondary PM constituents. The authors concluded that ambient concentrations of PM2.5 and secondary aerosols (including sul- fate, ammonium, and nitrate) were positively associated with respiratory hospitalizations in New York State. Ostro et al. (2016) used city-specific PM2.5 mass source apportionment in eight major metropolitan areas in California from 2005 to 2009 to examine the associations of source-specific PM2.5 exposures from vehicular emissions, biomass burning, soil, and secondary nitrate and sulfate sources with emergency department visits (EDVs) for cardiovascular and respiratory diseases. Using a case-crossover analysis, they observed associations of vehicular emissions with all cardiovascular EDVs. Vehicular emissions, biomass burning, and soil sources were associated with all respiratory EDVs and with EDVs for asthma. The soil source, which includes resuspended road dust, generated the highest risk estimate for asthma. Overall, the authors concluded that their results provide additional evidence of the public health consequences of exposure to specific sources of PM2.5 and indicate that some source-specific components of PM2.5 may pose higher risks (per µg/m3) than the overall PM2.5 mass. Short-Term Exposure 42 C. Summary of Short-Term Studies Tables 1 and 2 summarize the findings reviewed above regarding the associations be- tween health outcomes and short-term exposures to PM2.5 constituents and source-spe- cific components. From these tables, it is observed that traffic-related PM2.5 is among the most consistently associated source components. Acute exposure to traffic PM2.5 shows consistent associations with CVD mortality (significance scoring of 4 in six studies), as well as with CVD hospital admissions (significance score of 5 in six CVD hospital admis- sions studies). Similarly, looking at the traffic-related PM2.5 constituent associations with acute health outcomes, elemental carbon (EC, a marker for traffic-based combustion) is consistently associated with CVD mortality and morbidity (significance score of 8 in 12 studies, and significance score of 6 in seven CVD hospital admissions studies). However, copper (Cu), which is often cited as a tracer for traffic emissions from metal brake wear, is not found to be consistently associated with CVD mortality or hospital admissions (though it is moderately associated with respiratory hospital admissions). Roadway-re- lated PM2.5 also appears to be generally associated with CVD mortality and more moder- ately with respiratory disease (RSD) hospital admissions. This observation lends further support to the importance of the traffic source as reported for PM2.5 mass associations with mortality and morbidity. Especially revealing is the fact that traffic PM2.5 and EC (primarily emitted by diesel vehicles in urban areas) are both related to CVD mortality and hospital admissions, but Cu is not. This suggests that it is the diesel-powered vehicles that are driving these traffic-related adverse risk associations, not the gasoline-powered vehicle emissions (which also cause Cu emissions from brake wear but not nearly as much EC as diesel vehicles). These findings about EC are supported by the recent me- ta-analysis by Achilleos et al. (2017), which found EC strongly associated with all-cause and cardiovascular mortality. Its correlate, black carbon, was reported to be associated with respiratory mortality. Acute exposure to PM2.5 from coal combustion is also especially associated with CVD mortality (significance score of 2 in three studies considering that exposure and out- come). In addition, the associations of trace element constituents of coal combustion (that is, Se and S) with these health outcomes are reasonably consistent with the con- clusion about the importance of coal combustion-related PM2.5. Both Se and S, which are significantly generated from coal burning are also often correlated with CVD mortality. Thus, coal-burning PM2.5, when present, appears to be an important driver of the acute CVD toxicity of PM2.5 mass. The meta-analysis of Achilleos et al. (2017) reported associa- tions between sulfate and all-cause mortality and more modest associations with cardio- vascular mortality. Some of their sulfate measures used by the authors were converted from sulfur, which could have added some uncertainty to the estimates. Acute exposures to residual oil combustion-derived PM2.5 were also found to be con- sistently associated with CVD mortality (significance score of 3 for four studies) and hospital admissions (2 for four), across the available studies where it was considered. In support of these findings, vanadium, a known oil combustion tracer, was also associated with total mortality (2 of four), CVD hospital admissions (2 of four), and RSD admissions Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 43 (3 of three). There is also evidence of a vanadium effect report in Achilleos et al. (2017). Acute exposure to PM2.5 derived from biomass combustion is also identified as associated with CVD mortality (score of 3 for three studies available for review). However, potassium (K), a common marker for biomass burning (for example, wood burning), was not consis- tently associated with any outcome in those studies. However, K can come from various sources, including windblown dust, and the lack of findings may be due in part to total K being a less-specific marker of biomass than water-soluble K or levoglucosan. In contrast, the meta-analysis by Achilleos et al. (2017) reports evidence of risk from K exposure with all-cause, cardiovascular, and respiratory mortality. Organic carbon (OCM) was found to be consistently associated with total mortality and with CVD hospital admissions, and OCM is a contributor in the source apportionment analysis to the biomass source, but it is but one source of many for OCM. The overall results suggest that the estimated effect of biomass is not robust across various data and methods used in studies that examine this source, and the biomass PM2.5 source association with acute mortality and morbidity is only moderately supported by constituent and source-specific analyses from developed countries to date. However, given the abundant and consistent use of biomass for heating and cooking in LMICs, it constitutes a major area of future research. Source-specific PM identified as “industrial” (for example, from steel mills) was also found to be moderately associated with CVD mortality (2.5 significance score for five studies) and hospital admissions (2 for three studies). However, it was difficult to discern support- ive associations with constituents commonly associated with industrial emissions. For example, lead (Pb), emitted by lead-smelting operations, was not found to be consistently associated with CVD outcomes. However, manganese (Mn), which is emitted by the steel industry, was associated with acute CVD mortality and RSD admissions. Thus, the results for industry sources and constituents were mixed, and not as compelling as combustion categories, except for the steel industry. Soil PM2.5 source impacts were associated with CVD mortality (4 for six studies) but not RSD mortality. However, soil was not associated with CV or RSD hospital admissions. Similarly, Si (a key constituent of soil) was also significantly associated with CVD mor- tality but not RSD mortality or any hospital admissions category. In contrast, the me- ta-analysis by Achilleos et al. (2017) reports associations between markers of soil (silicon and calcium) with both all-cause and cardiovascular mortality. This indicates a lack of risk-estimate robustness of soil-related PM2.5 mass associations across the various study data and methods. Neither CVD nor RSD mortality was found to be associated with acute exposure to salt PM2.5 (for example, resulting from ocean sea spray) or with salt’s key tracers, sodium (Na) or Cl. However, an association was found with RSD hospital admissions, while its key constituents, Na and Cl were also found to be associated with increased RSD morbidity. Overall, the collective evidence from tables 1 and 2, consistent with the above discussions of the individual studies, indicates that the particles from combustion, and especially fossil-fuel combustion, are the most consistent major driving factors in the acute PM2.5 Short-Term Exposure 44 exposure associations found with mortality and morbidity. Of the fossil-fuel sources, coal combustion and its secondary PM2.5 products (for example, sulfates), when present, were an important driver of the PM2.5-CVD association. Acute exposures to residual oil com- bustion–derived PM2.5 was also found to be consistently associated with CVD mortality and hospital admissions. Of the traffic sources, acute exposure to diesel-combustion PM2.5 were suggested as the traffic component most associated with adverse health risks. Biomass and soil PM2.5 were also found to be associated, but less consistently, with both mortality and morbidity. However, there are fewer studies that include these two sources, and their results were less robust across study design and methods than most of the other sources. “Particles from combustion, and especially fossil-fuel combustion, are the most consistent major driving factors in the acute PM2.5 exposure associations found with mortality and morbidity.” Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 45 Long-Term Exposure to PM2.5 Sources and Constituents and Their Associations with Mortality 46 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 47 VII. Long-Term Exposure to PM2.5 Sources and Constituents and Their Associations with Mortality For long-term exposure studies, search terms from the bibliographic source PubMed including “long-term exposure” or “chronic exposure” with “species,” “components,” or “constituents” were used in the development of table 3. The first published efforts to examine the associations of long-term exposure to PM2.5 mass, constituents and sources with mortality was an ecological cross-sectional study7 of 1980 mortality rates in 98 US metropolitan areas (Ozkaynak and Thurston 1987). Data on PM2.5, its trace element constituents, and sulfates were obtained from the US EPA Inhalable Particle Monitoring Network (IPMN) that operated from 1979 to 1982. Multiple regression was used to control for metropolitan area-wide population characteristics, such as median age, percent above 65 years of age or older, population density, percent in poverty, and percent with a college degree. Although this was an ecological population study, it gave similar PM2.5 mass-mortality associations as later cohort studies with individual-level characteristics data (for example, Pope et al. [2002]). In addition to fine particles, sulfate (the only constituent individually studied in this analysis) was associated with mortality rates (0.7 percent increase per µg/m3, PM2.5, 95 percent confidence interval (CI) = 0.2–1.28; and 2.2 percent per µg/m3 SO42-, 95 percent CI = 1.2–3.0). An additional analysis was conducted of the association of mortality with source-specific PM2.5 components, as derived from the trace element constituents (Thurston et al. 1984), including soil (associated with Si, Fe), auto emissions (Pb and Br), residual oil combustion (V and Ni), metals (Zn, Mn, Fe), and coal combustion (Se, S). Of these source-specific PM2.5 components, coal combustion had the most significant impact on mortality (Thurston et al. 1987), which was also later confirmed by the subsequent source-specific PM2.5 analysis of the American Cancer Society (ACS) cohort (Thurston et al. 2016). 7  nlike a cohort study that considers individual-level deaths and characteristics over time, an ecological U cross-sectional mortality study is one that considers mortality rates of entire populations (for example, of cit- ies), as well as their population characteristics (for example, percent > 65 years old in a city), and compares them cross-sectionally (that is, across populations in different locations). 8 A “95 percent CI” refers to the 95th-percentile confidence interval of the estimate provided.  Long-Term Exposure to PM2.5 Sources and Constituents and Their Associations with Mortality 48 Dockery et al. (1993) used the cohort of the Harvard Six Cities Study to examine the risks of PM2.5, PM10, and sulfate on mortality. The data included a 14- to 16-year-long follow-up of approximately 8,100 adults from cities in the eastern and midwestern US. A Cox pro- portional hazards regression model was applied. Among the results, the analysis indicat- ed statistically significant associations between mortality and both PM2.5 and sulfate (the only constituent examined). Converting the results indicated a 2.9 percent (95 percent CI = 1.0, 4.6) change in all-cause mortality per µg/m3 of sulfate. This compared to a 1.2 percent (95 percent CI = 0.4, 2.1) change in all-cause mortality per µg/m3 of PM2.5. These cohort results were confirmatory to the previously published cross-sectional results (for example, Ozkaynak and Thurston 1987). However, in these studies, sulfates were highly correlated with PM2.5, so it is difficult to disentangle the two risk estimates. A subsequent reanalysis of the Harvard Six Cities Study data included examination of cardiopulmonary mortality with a risk increase of 3.5 percent (95 percent CI = 0.7, 6.1) per µg/m3 of sulfates (Krewski et al. 2000). Pope et al. (1995), Krewski et al. (2000), and Pope et al. (2002) examined survival among 550,000 nationwide participants from the American Cancer Society Prevention Study II cohort. The risks of long-term exposure to sulfates were based on the IPMN and other EPA monitoring data. The study examined individuals that were followed from 1982 to 1998. For all-cause mortality, the change in mortality per µg/m3 of sulfates was 1.3 percent (95 percent CI = 1.1, 1.5), versus risks for PM2.5 of 0.6 percent (95 percent CI = 0.4, 0.9). Sulfate risks on cardiopulmonary mortality were estimated to be 1.8 percent increase in mortality per µg/m3 (95 percent CI = 1.2, 2.4). The authors concluded that long-term exposure to combustion-related fine particulate air pollution is an important environmen- tal risk factor for cardiopulmonary and lung cancer mortality (Pope et al. 2002). More recently, several studies have evaluated the possible associations of long-term exposures to PM2.5 constituents and/or sources with mortality. Smith et al. (2009) used data from the American Cancer Society Prevention Study II cohort to analyze the risks of long-term exposures to PM2.5, sulfate, and black carbon. Sulfate data were similar to that used by Krewski et al. (2000) and a three-year average (2003–2005) of elemental carbon data was developed from data made available by research funded by the Health Effects Institute (Thurston et al. 2011). In this work, there was a subcohort of 350,000 partici- pants from 66 metropolitan areas followed from 1982 to 2000 for which complete data were available. Mortality risks were estimated with models for independent pollutants and various combinations of co-pollutants. The researchers tested two-way linear inter- actions between all pollutants. A Cox proportional hazards model was used to assess the all-cause and cardiopulmonary mortality risk estimates. For all-cause mortality, the risk estimates per µg/m3 were 0.6 percent (95 percent CI = 0.2, 1.0) for PM2.5, 1.1 percent (95 percent CI = 0.8, 1.4) for sulfate and 5.5 percent (95 percent CI = 0.74, 10.51) for EC. The respective constituent risks for cardiopulmonary mortality were 1.3 percent (95 percent CI = 0.8, 1.8), 1.6 percent (95 percent CI = 1.03 to 2.08) and 10.6 percent (95 percent CI = 2.9, 18.9). The large and significant risk estimates for both sulfate and elemental carbon in this analysis again pointed toward fossil-fuel combustion products as a major contributor to the PM2.5 associations with mortality. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 49 Ostro et al. (2010) examined long-term exposure among 102,000 California women teachers and administrators from the California Teachers Study (CTS) who were followed from 2001 through 2007. To reduce exposure misclassification9, the sample was restricted to those living within 30 kilometers of a fixed-site monitor in the US Environmental Protection Agency’s Species Trend Network. This resulted in a study population of just under 45,000 women. PM2.5 constituents examined included EC, OC, sulfate, nitrate, iron, potassium, silicon, and zinc. A Cox proportional hazards model was used to derive the risk estimates. For cardiopulmonary mortality, among the major constituents of PM2.5, statistically significant risk estimates were observed with sulfate and silicon, along with PM2.5 mass. For IHD, the largest risks were observed for EC and sulfate, consistent with the ACS study results, although estimates were fairly similar among the constituents, except for silicon and OC which had somewhat lower estimates. For comparative purposes, the CTS IHD results were converted into risks per one µg/ m3. For cardiopulmonary mortality, the risk of one µg/m3 of sulfates was 3.4 times that of PM2.5, while EC was 6.1 times that of PM2.5. For IHD, the ratio of risks relative to PM2.5 was 4.0 and 13.2 for sulfates and EC, respectively (Ostro et al. 2010). The same data set (CTS) was examined using exposure data at the residential level provided by a chemical transport model that computed pollutant concentrations from over 900 sources in California (Ostro et al. 2015). Besides particle mass, monthly con- centrations of 11 species and eight sources of primary particles were generated at 4-kilo- meter grids throughout the state. Both chemical constituents and sources of PM2.5 were examined, and the focus was on cardiovascular and IHD-related mortality. Specifically, the following PM2.5 constituents were examined: copper, iron, manganese, nitrate, elemental carbon (EC), organic carbon, “other” species (that is, mineral dusts and constituents not measured), “other” metals (those besides Cu, Fe, and Mn that were explicitly resolved), and secondary organic aerosols. Sulfates were not measured. The sources of primary aerosols were not estimated via PM2.5 source apportionment methods but instead by the University of California, Davis/California Institute of Technology (UCD/CIT) chemical transport model from calculated meteorological fields and emissions estimates for different sources to predict source-specific airborne particulate matter concentrations. These PM2.5 sources included on- and off-road gasoline, on- and off-road diesel, wood smoke, meat cooking, high-sulfur fuel combustion (including distillate oil, marine vessel fuel, aircraft jet fuel, and liquid and solid waste fuels), and “other anthropogenic.” It is important to note that most electricity in California is produced from natural gas, with little or no use of coal or oil. There were many statistically significant associations between the constituents and IHD mortality in the CTS cohort, the outcome of interest in the study. The highest mortality risk was associated with EC, considered to be a tracer for diesel exhaust. Among the sources, there were statistically significant associations between IHD and all four of the vehicle sources, meat cooking, and high-sulfur-content fuel combustion in California. High-sulfur-content fuel combustion includes emissions from distillate oil from industrial boilers, marine vessel fuel, aircraft jet fuel, and liquid and solid waste fuels. Thus, overall, this study indicates that exposures to traffic (EC) and sulfur combustion are among the 9  xposure misclassification occurs when the exposure assigned to an individual or population is over- or E under-estimated. Long-Term Exposure to PM2.5 Sources and Constituents and Their Associations with Mortality 50 most important sources of mortality. To obtain an approximation of the relative toxicity, the risks of PM2.5, EC, and sulfur combustion were converted to a common denominator of one µg/m3. The percent increase in mortality is 1.7 percent (95 percent CI = 0.8, 2.7) for PM2.5, 16.4 percent (95 percent CI = 6.1, 26.9) for EC, and 19.2 percent (95 percent CI = 5.0, 30.6) for sulfur combustion. Thus, on a per µg/m3 basis, the results are suggestive of greater toxicity for motor vehicle and sulfur-emitting sources than for overall PM2.5 in general. Note that some caution should be exercised in interpreting these figures, given the inherent uncertainty in measuring the PM constituents and estimating PM2.5 source contributions. Gan et al. (2011) examined the risk of coronary heart disease mortality in a cohort of 450,000 individuals in Vancouver, Canada, during the period of 1999 through 2002. Five- year averages of both black carbon (subsequently converted to EC) and PM2.5 developed from land use regressions were used as exposures. The risk estimate for a one µg/m3 change in EC was 7.7 percent (95 percent CI = 3.9, 6.5) versus, for PM2.5, 0.6 percent (95 percent CI = -1.3, 3.1). Vedal et al. (2013) used data from the Women’s Health Initiative–Observational Study (WHI-OS), a cohort of about 90,000 women from 45 US cities across the country, to investigate the risks of constituents on total cardiovascular mortality and several sub- classes. Also, first events of cardiovascular disease were examined. Besides PM2.5, data on EC, OC, sulfur, and silicon were obtained from the US speciation dataset and the Interagency Monitoring of Protected Visual Environments (IMPROVE) network. The au- thors’ source apportionment analysis indicated that these constituents were, respectively, markers of local combustion including traffic, primary gasoline and biomass combustion and secondary OC formation, secondary sulfate formation, and crustal/soil. The cohort was followed from 1995 to 2005, while the pollution measurements were for 2009. In the basic analysis across metropolitan areas, OC was most consistently associated (P < 0.05) with cardiovascular death and many of its subclasses including atherosclerotic and cerebrovascular deaths. PM2.5, EC, and sulfur were more modestly associated (P < 0.06) with atherosclerotic deaths and silicon was associated (P < 0.05) with deaths from possible coronary heart disease. The risk estimates suggest much larger risks per µg/m3 for these constituents than for PM2.5 mass as a whole. Regarding cardiovascular events, sulfur was associated with total cardiovascular events (MI, coronary revascularization, stroke, atherosclerotic cardiac disease death, possible CVD death, and cerebrovascular death), stroke and stroke death, and myocardial infarction. Of the other constituents, OC was associated with stroke and stroke death. The Health Effects Institute review commit- tee noted the difficulties in measuring EC at a small spatial scale. Wang et al. (2014) conducted an analysis of 320,000 individuals from 19 European cohorts in 12 countries as part of the ESCAPE study. The cohorts began from 1985 to 2000 and were followed for approximately 11 years. Estimates of PM2.5 constituents were developed using land use regression (LUR) models that provided modest fits of the data: for sulfur and silicon, the average correlation between the LUR estimates and air monitor measurements across the cohorts was 0.37 and 0.49, respectively. Eight constituents Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 51 representing major sources in Europe were examined including copper, iron, potassium, nickel, sulfur, silicon, vanadium, and zinc. Four models with increasing control for con- founders were run. Basically, no associations with cardiovascular mortality were observed for any of the constituents, but sulfur and silicon exhibited the highest risk estimates. Chung et al. (2015) examined 12.5 million Medicare enrollees residing in the eastern Unit- ed States. A seven-year average from 2000 to 2006 pollution exposures was developed for six major constituents of PM2.5 taken from the US EPA chemical speciation network. The constituents included elemental carbon (EC), organic carbon matter (OCM), sulfate (SO42-), silicon (Si), nitrate (NO3-), and sodium (Na). Rather than using a traditional Cox regression, the authors used a novel Bayesian hierarchical regression model to determine whether changes in monthly mortality rates were associated with monthly changes in the concentrations of the PM2.5 constituent. The regression model for a given constituent controlled for other constituents by including a term for the annual average of PM2.5. The authors also examined whether the risk of the seven-year average of PM2.5 on mortality was modified by the concentrations of the constituents. In their basic model, they found that EC, Si, and nitrates were associated with mortality. They also found that PM2.5 mass that contained more sulfates demonstrated significantly greater risk estimates. Kioumourtzoglou et al. (2015) attempted to determine whether PM composition mod- ified the risks on mortality. Their sample included 19 million Medicare enrollees from 81 US cities from around the country for the years 2000 to 2010. Source clusters were developed using EPA monitoring data for 24 species. Most metropolitan areas relied on a single monitor to measure species concentrations. After controlling for long-term trends and factors varying across cities, the authors examined whether within-city year-to-year changes in the constituent concentrations were associated with corresponding changes in mortality. Overall, the authors concluded that the strongest positive modifications of the association between PM2.5 and mortality were from (i) sulfates in the Southeast, (ii) metals associated with ship emissions in harbor cities, and (iii) industrial emissions and elemental carbon in several parts of the country. Another recent study examined the risk of PM2.5 source-related components and their tracer constituents using a subset of the national American Cancer Society’s Cancer Prevention Study II cohort (Thurston et al. 2016). A total of 446,000 adults in 100 US metropolitan areas were followed from 1982 to 2004. In this case, the average of con- centrations from the available US EPA speciation monitors in each city were used as a measure of metropolitan area-wide concentrations of the pollutants. Sixteen constituents and eight sources were examined. Four different Cox regression models focusing on IHD mortality were employed, where additional controls for possible confounding factors were progressively introduced. Among the constituents that contributed significantly to PM2.5 mass, EC (a tracer for traffic) and sulfur (a tracer for combustion from coal and residual oil) were statistically significant in most models, while organic carbon and silicon were not. Of note, the risks relating to the coal-related mass component were not altered when other sources were controlled for in models with two sources. PM2.5 from both wind-blown soil and biomass combustion was not associated with IHD mortality. Based Long-Term Exposure to PM2.5 Sources and Constituents and Their Associations with Mortality 52 on the model including random effects but no population-based contextual variables, mortality risk estimates for PM2.5 mass, coal combustion, and EC were 1.6 percent (95 percent CI = 0.6, 2.5), 4.6 percent (95 percent CI = 1.5, 9.1), and 11.5 percent (95 percent CI = 0, 22.6), respectively, on a µg/m3 basis. In a prior analysis of sources of PM2.5 and other health endpoints using these same data, Thurston et al. (2013) also reported associations between all-cause mortality and both traffic and coal sources, between respiratory mortality and both traffic and soil, and between lung cancer and coal. Using these studies of long-term exposure (table 3), it is possible to more appropriately assess the source risk estimates on a comparable “per µg/m3 of PM2.5 mass” basis. Each of the constituents (for example, EC and SO42-) is a general indicator of PM2.5 source con- tributions (for example, diesel traffic and fossil-fuel combustion, respectively). However, these associated source-specific PM2.5 constituents actually include more PM2.5 mass than the single constituent alone. Fortunately, there are estimates of the mass associated with these key source tracers for the US (Thurston et al. 2011). Based on that research, EC represents 15.3 percent of the total PM2.5 mass associated with traffic sources. Therefore, multiplying the EC concentration by 1/.153 (= 6.54) yields an estimate of the associated PM2.5 mass in µg/m3. Thus, the reported risk per µg/m3 of EC needs to be divided by 6.54 to give an estimate of the impact per µg/m3 on a PM2.5 mass basis that is comparable to the overall PM2.5 mass risk per µg/m3. Based on the studies of Gan et al. (2011), Ostro et al. (2015), and Thurston et al. (2016), the EC risk estimates are estimated as 13, 10, and 6.9 times that of overall PM2.5 mass, respectively. Using the mean of these ratios (10) and adjusting for the actual PM2.5 mass associated with EC, the EC associated mass risk is estimated at approximately 1.5 times that of PM2.5 mass, in general. In order to determine the sulfate-specific mortality risks, similar associated mass calcula- tions were conducted to take into account the ammonium ion that is also associated with measured sulfate mass. As summarized in table 4, the available studies report a mean sulfate risk of 2.7 times that of PM2.5 mass. Based on the ratio of ammonium sulfate to sulfate molecular weights (132/96 = 0.73), sulfate makes up about 73 percent of the total sulfate-associated PM2.5 mass. Therefore, after adjustment for total mass, the sulfate-as- sociated PM2.5 risk estimate (that is, per µg/m3 PM2.5 mass of sulfate) is about 2.7 * 0.73 = 2.0 times the mortality risk of overall PM2.5 per µg/m3 mass. When similarly considering the source-specific mass impacts, the highest ischemic heart disease mortality risks per µg/m3 found by Thurston et al. (2016) was for coal-combustion-specific PM2.5, which was approximately five times the risk per µg/m3 as found for ambient PM2.5 mass, in general. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 53 Summary and Conclusions 54 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 55 VIII. Summary and Conclusions Taken together, the available short- and long-term exposure studies suggest the following: • This is the first PM2.5-health-effects review that, to the authors’ knowledge, collectively examines both short- and long-term exposures to fine particle constituents and sources, taking into account the data and methods used as well as the uncertainties in the underlying epidemiological studies. • The current evidence summarized in this report indicates that trace constituents from PM2.5 and PM2.5 mass from fossil-fuel combustion are among the greatest contribu- tors to PM2.5 toxicity. The results of both the short- and long-term exposure studies strongly suggest that the PM2.5 mass component derived from fossil-fuel combustion is of particular concern for public health. The health risk is particularly great from coal combustion- and traffic-related PM2.5. For the latter, the many associations with elemental carbon (EC) from soot exposure indicate that, of traffic-related sources, diesel-fueled vehicles are particularly associated with adverse human health impacts. These two sources, coal- and traffic-related PM2.5, are most consistently associated with cardiovascular mortality, especially ischemic heart disease (heart attacks) in both the short- and long-term exposure studies. • The importance of coal-combustion and traffic-emission sources to the overall PM2.5 mass associations with mortality and hospital admissions is supported by the health- risk associations that are also observed with their source-specific trace constituent markers. For example, particulate selenium and sulfur are important markers for coal and fossil-fuel combustion, while elemental carbon is a strong marker for traffic, and especially for diesel-fueled vehicle, emissions. Based on prior epidemiological and toxicological studies, Lippmann at al. (2013) attempted to draw conclusions about the importance of the different PM2.5 constituents, finding that sulfate (a marker primarily of coal and residual-fuel-oil combustion) was clearly among the most, if not the most, important adverse health-associated constituents of PM2.5. • As summarized in tables 3 and 4, after adjusting for per µg/m3 mass equivalence, the long-term studies indicate that the cardiovascular disease risks of sulfate, EC, and PM2.5 mass from coal combustion are larger than that of PM2.5 mass, in general. • This review of the studies indicates a fairly consistent association between sulfate and cardiovascular mortality. This may well be because sulfates do not occur as pure chemicals in the air but in intimate mixtures with other compounds, including transition Summary and Conclusions 56 metals. Thus, the acidic nature of sulfur compounds in the environment (for example, as ammonium sulfate, ammonium bisulfate, or sulfuric acid) can make transition metals more bioavailable, increasing those particles’ potential to cause oxidative stress and serious adverse health effects in the human body (for example, see COMEAP [2009] and Fang et al. [2017]). • However, there is a caveat concerning the potential confounding effects of exposure characteristics on the reported sulfate associations. Specifically, it is possible that since sulfate is more homogeneous spatially (that is, more equally dispersed over a wider area) than many other PM2.5 constituents and source-related components, exposures to sulfate might well be measured more accurately by a small set of monitors (or a single monitor) in a metropolitan area than some of the other, more spatially variable, constit- uents or source-related components. Such superior measurement of exposures leads, in general, to a greater likelihood of observing an association. • In contrast, PM2.5 mass, its constituents, and its source-related components are less consistently associated with respiratory disease mortality, except for short-term expo- sures to PM2.5 from biomass burning. • In general, weaker associations were most often found between mortality and con- centrations of PM2.5 from biomass combustion or crustal sources. However, several short-term exposure studies report associations between biomass and mortality. These two sources are extremely varied spatially and temporally, especially for estimating long-term exposure. This could contribute to the lack of detection of their health associations. In addition, there are fewer US metropolitan areas studied that have significant concentrations of biomass and soil sources, which may limit the ability of national-scale studies to identify risks of these two sources of PM2.5. On the other hand, traffic emissions and their markers (for example, EC), which are often associated with mortality, are also quite variable spatially within a metropolitan area, but they still show consistent associations with adverse health. • Overall, this report’s key findings from the review of currently available acute (or short-term) and long-term studies suggest that reductions in the emissions of PM2.5 from fossil-fuel combustion, such as sulfur-emitting coal-fired power plants and diesel vehicles, can produce the most significant health benefits per µg/m3 of PM2.5 reduced. In addition, these PM2.5 emissions reductions would provide additional climate-mitigation benefits, since these same fossil-fuel combustion sources contribute significantly to climate change, most notably in the form of CO2 and elemental carbon emissions. • The relative dearth of PM2.5-constituent health effects studies found for LMICs in this review supports the need for LMICs in particular to begin measuring the concentrations of the constituents of their particulate mixes. Particulate S and EC, which can be measured continuously (McMurry 2000), are especially important combustion markers to monitor more widely. This will help determine the contribution of specific sources to PM2.5 mass and would help prioritize subsequent pollution-control efforts. To date, results indicate a need to focus on reducing PM2.5 pollution from fossil-fuel combustion, especially coal-burning and diesel-vehicle emissions. • There is an especially great need to conduct short- and long-term cohort epidemio- logical studies of mortality and hospitalization in LMICs, where biomass burning for cooking and heating is a significant source of PM2.5 but of uncertain human health risk Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 57 relative to fossil-fuel combustion particulate matter. In order to develop economically efficient abatement strategies that maximize human health benefits per ton of PM2.5 emissions reduced, it is essential to (i) measure the concentration of PM2.5 from the specific source (in this case biomass burning), or of a marker of the source (for example levoglucosan), and (ii) determine the specific toxicities of source-specific contributions to PM2.5 in LMICs. • Efforts to control ambient air pollution in LMICs need to account for both the con- tributing sources of PM2.5 and the toxicity of the PM2.5 from each source category. The strongest evidence to date of increased risk of CVD health effects per unit mass of PM2.5 is found from coal combustion and diesel-fueled traffic. Therefore, future mitiga- tion efforts that focus on monitoring and regulating the constituents of PM2.5 from these sources should be prioritized. Reducing pollution from these sources can be expected to return greater cardiovascular health benefits per unit mass of PM2.5 reduced than if PM2.5 mass continues to be addressed equally, irrespective of source and composition. “The current evidence summarized in this report indicates that trace constituents from PM2.5 and PM2.5 mass from fossil- fuel combustion are among the greatest contributors to PM2.5 toxicity. The results of both the short and long-term exposure studies strongly suggest that the PM2.5 mass component derived from fossil- fuel combustion is of particular concern for public health. The health risk is particularly great from coal combustion and traffic-related PM2.5.” Appendix 58 Appendix TABLE 1 Associations of Adverse Health Effects with Acute Exposures to PM2.5 Constituents MORTALITY Authors Year Location City Outcomes PM OCM EC BC TC Nitrate Nitrite Sulfate Sulfite NH4+ Al As B 1 Laden 2000 US 6 cities All 0 0 2 Ostro 2007 US 6 California counties All 1 0.5 0.5 0.5 1 0 CVD 1 0.5 1 0.5 1 0 RSD 1 0 0 0 0 0 3 Ostro 2008 US 6 California counties CVD 1 0 1 0.5 1 4 Cakmak 2009 Chile Santiago All 1 1 1 0 CVD 1 1 1 0 RSD 1 1 1 0 5 Zhou 2010 US Detroit and Seattle All 1 1 1 CVD 1 1 0 RSD 0 0 0 6 Ito 2011 US New York City CVD 1 1 1 1 1 7 Son 2012 Korea Seoul All 0 0 0 0 0 0 CVD 0 0 0 0.5 0.5 1 RSD 0 0 0 0 0 0 8 Cao 2012 China Xi’an All 1 1 1 1 1 1 1 CVD 1 1 1 1 1 1 0 RSD 1 1 1 1 1 1 0 9 Huang 2012 China Xi’an All 1 0 1 0 0 0 CVD 1 0 1 0 0 0 RSD 1 0 0 0 0 0 10 Sacks 2012 US Philadelphia CVD 0 11 Krall 2013 US 72 cities All 1 1 1 0 0 0 12 Dai 2014 US 75 cities All 0 0 0 0 CVD 0 0 0 0 RSD 1 0 0 0 13 Heo 2014 Korea Seoul All 0 1 0 0 0 CVD 1 1 0 0 0 RSD 0 0 1 0 0 14 Li 2014 China Beijing All 1 1 0 1 0 CVD 1 1 0 1 0 RSD 1 1 0 1 0 15 Basagana 2015 Europe 5 cities CVD 1 0 1 0 0 1 RSD 1 0.5 1 0.5 0 0.5 16 Kim 2015 US Denver All 0 1 1 0 0 CVD 0 0 0 0 0 RSD 0 0 0 0 0 Fraction ALL cause sig 7.0/11 5.5/9 6.5/10 2.5/8 3.0/8 1.0/5 1.0/5 1.0 Fraction CVD sig 9.0/13 4.5/11 8.0/12 4.5/10 6.5/10 2.0/4 0.0/4 1.0 Fraction RSD sig 7.0/11 2.5/9 4.0/10 2.0/8 2.5/8 1.0/4 0.0/4 1.0 Fraction any category significant 10./13 7./12 10./13 5./11 6.5./11 2./5 1./5 3. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 59 H4 Al As Br Ca Cl Cr Co Cu F Fe K Mg Mn Na Ni Pb S Se Si Ti V Zn 0 0 0 0 0 0 1 0 0 1 1 1 0 0 1 0 0 0 1 1 1 1 1 1 0 1 1 1 1 1 1 0 0 0 1 1 1 1 1 0 0 0.5 0 1 1 1 0 0 0 0 1 0 0 0 0 0 0 0 1 0.5 0.5 0 0 0 1 0.5 0.5 0.5 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 1 0 1 0 0 1 0 0 1 1 0 1 0 0 0 0 0 0 0 0 0 1 1 0 1 1 0 1 0 0 0 1 0 0 0 0 0 0 0 0.5 0 0.5 0 1 1 1 0 0 1 1 0 1 1 1 0 1 1 0 1 1 1 1 1 0 0 1 1 0 1 0 1 0 0 1 1 0 1 0 1 0 0 1 0 1 0 1 1 0 0 0.5 0 0 1 1 0.5 0 0 1 1 0 1 0 0 0 0 0 0.5 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 0 0 0 0 0 0 0 0 1 1 1 1 0 1 0 1 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1 0 0 1 1 /5 1.0/5 1.0/5 3.0/6 3.0/7 1.0/5 2.0/5 4.0/9 1.0/3 1.0/3 1.0/6 3.0/7 2.0/5 5.5/7 4.0/6 1.0/1 2.0/4 2.0/7 /4 0.0/4 1.0/5 0.0/7 3.0/7 2.0/7 2.5/7 4.5/11 1.0/4 2.0/3 1.0/7 2.0/9 1.0/4 4.0/8 2.0/3 2.5/7 1.5/3 1.0/6 5.0/10 /4 0.0/4 1.0/4 0.0/6 2.5/6 1.0/5 1.0/5 3.0/9 1.5/4 0.0/3 1.0/6 2.0/7 2.0/4 1.5/6 0.0/5 1.0/2 1.5/4 1.5/7 5 1./5 3./6 3./8 4./8 2./8 4.5/8 5.5/12 2./4 2./3 3./8 3./8 3./4 6.5/8 2./4 4.5/8 1.5/3 2./6 5./11 Appendix 60 TABLE 1 (CONTINUED) Associations of Adverse Health Effects with Acute Exposures to PM2.5 Constituents HOSPITALIZATIONS Authors Year Location City Outcomes PM OCM EC BC TC Nitrate Nitrite Sulfate Sulfite NH4+ Al As B 1 Sarnat 2008 US Atlanta CVD 1 1 1 0 0 RSD 0 0 0 0 1 2 Bell 2009 US 106 counties CVD 0 1 0 0 0 0 0 RSD 0 1 0 0 0 0 0 3 Ito 2011 US New York City CVD 1 1 1 1 1 4 Lall 2011 US New York City CVD 0 RSD 0 5 Kim 2012 US Denver CVD 0.5 1 1 0 0 RSD 0 0 0 0 0 6 Levy 2012 US CVD 0 1 0 0 RSD 0 0 0 0 Mostof- 7 2012 US Boston Ischemic 1 1 0 sky 8 Bell 2013 US 4 counties CVD 1 1 0 0 RSD 0 1 1 9 Li 2014 China Beijing CVD 1 0 0 0 0 RSD 1 0 0 1 0 10 Jones 2015 US New York State RSD 1 1 0 1 1 1 11 Ostro 2016 US 8 metros CVD 0 0 1 0 0 RSD 1 1 1 1 0 Fraction CVD sig 4.5/6 3.0/6 6.0/7 1.0/7 1.0/7 0.0/2 1.5 Fraction RSD sig 3.0/7 1.0/6 2.0/7 2.0/7 3.0/7 1.0/3 0. Fraction any category significant 7.5/9 5./7 6./8 3./8 4./8 1./3 1.5 Source: Compiled by the authors of this report based on the works of the authors indicated in the table. Note: CVD = cardiovascular disease, OCM = organic carbon matter, PM = particulate matter, RSD = respiratory disease, sig = statistically significant, TC = total carbon. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 61 H4 Al As Br Ca Cl Cr Co Cu F Fe K Mg Mn Na Ni Pb S Se Si Ti V Zn 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 1 1 1 1 1 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0.5 0 0.5 0 0 0 0 0 0 0.5 1 0 0 0 0 0 1 0 1 0 1 1 0 0 1 1 1 0 0 0 0 0 1 1 0 1 1 0 1 1 0 0 1 0 0 0 0 0 0 1 1 1 1 1 0 0 1 0 1 1 /2 1.5/2 1.0/3 0.0/3 0.0/2 0.0/2 2.0/5 0.0/2 2./4 2.5/4 0.0/3 1.0/6 0.0/3 2.0/4 3.0/5 /3 0.0/1 1.0/3 2.0/3 1.0/2 1.0/2 1.0/5 2.0/2 1./2 1.0/3 0.0/3 2.0/5 1.0/3 3.0/3 1.0/4 3 1.5/3 1./4 2./3 1./3 1./3 2./6 2./3 2./4 2.5/5 0.5/4 3./7 1./4 3./5 3./6 Appendix 62 TABLE 2 Associations of Adverse Health Effects with Acute Exposures to PM2.5 Source Components MORTALITY Authors Year Location City Outcomes Residual Coal Traffic Biomass Soil Roadway Salt Industry 2nd 2nd Oil Nitrate Sulfate 1 Laden 2000 US 6 cities All 0 1 1 0 0 0 2 Mar 2006 US Phoenix All 0 0 0 1 1 0 CVD 1 1 0 1 1 1 3 Ito 2006 US Washington, DC All 0 1 1 0 1 0 0 0 1 CVD 1 1 1 1 1 0 0 1 1 4 Cakmak 2009 Chile Santiago All 1 1 1 1 CVD 1 1 1 1 RSD 1 1 1 1 5 Ostro 2011 Spain Barcelona All 1 1 0 1 0 0 0 0 CVD 1 1 1 1 0 0 1 1 6 Sacks 2012 US Philadelphia CVD 0 1 0 1 0 7 Heo 2014 Korea Seoul All 0 0 0 0 0 0 0 0 CVD 0 0 1 0 0.5 0 0.5 0 RSD 0 1 0 0.5 0 0 0 0 Fraction ALL cause sig 2./4 2./3 4./6 0/3 2.0/6 2./3 0/5 2.0/6 1./4 Fraction CVD sig 3./4 2./3 4./6 3./3 4.0/6 1.5/3 1./4 2.5/5 3./4 Fraction RSD sig 0/1 1./1 0/1 0.5/1 0/1 0/1 0/1 0/1 Fraction ANY sig 3./5 3./4 6./7 3./3 4.5/7 2.5/3 1./5 2.5/6 3./4 Source: Compiled by the authors of this report based on the works of the authors indicated in the table. Note: CVD = cardiovascular disease, RSD = respiratory disease, sig = statistically significant. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 63 TABLE 2 (CONTINUED) Associations of Adverse Health Effects with Acute Exposures to PM2.5 Source Components HOSPITALIZATIONS Authors Year Location City Outcomes Residual Coal Traffic Biomass Soil Roadway Salt Industry 2nd 2nd Oil Nitrate Sulfate 1 Sarnat 2008 US Atlanta CVD 1 1 0 0 1 0 0 RSD 0 0 0 0 0 0 1 2 Lall 2011 US New York City CVD 0 1 0 0 0 RSD 0 0 0 1 0 3 Bell 2013 US 4 counties CVD 0 0 1 0 0 RSD 0 0 1 1 0 4 Kioumourt- zoglou 2014 US Boston CVD 1 1 0 0 0 1 5 Ostro 2016 US 8 metros CVD 1 0 0.5 0 0 RSD 1 1 0.5 1 0 6 Krall et al. 2017 US 63 counties CVD 1 1 0 1 1 1 Fraction CVD sig 2./4 5./6 0.5/5 1./3 1./3 2./3 2./6 Fraction RSD sig 0/2 1./4 0.5/4 1./2 1./1 1./2 1./4 Fraction ANY sig 2./4 5./6 0.5/5 1./3 2./3 3./3 3./6 Source: Compiled by the authors of this report based on the works of the authors indicated in the table. Note: CVD = cardiovascular disease, RSD = respiratory disease, sig = statistically significant. Appendix 64 TABLE 3 Associations of Long-Term Exposure to PM2.5 Constituents with Mortality First Author Years Population Exposures Statistical Significant Species Comment Mortality PM2.5 O (year) Method and Sources Outcomes Ozkaynak (1987) 1980 US IPMN Multiple Sulfate Sulfate effect on all-cause about 3x All 1 regression PM2.5 Dockery (1993), 1977–1991 Six eastern and Six-City Data Cox PH Sulfate Sulfate effect on all-cause is 2.4x All 1 Krewski midwestern PM2.5 CP 1 cities LC 0 Pope (1995), 1982–1989 American Can- EPA and Cox PH Sulfate Sulfate effect on all-cause is about All 1 Krewski (2000) cer Society IPMN 2x PM2.5 CP 1 CV 1 RD 0 LC 0 Smith (2009) 1982–2000 American Can- EPA and Cox PH Sulfate, EC Sulfate effect on all-cause 2x PM2.5, All cer Society IPMN EC effect 9x PM2.5.; sulfate effect on CP CP 1.2x PM2.5, EC 10x PM2.5 Ostro (2011) 2001–2007 California EPA Specia- Cox PH EC, silicon, sulfate EC effect on CP 6x PM2.5; sulfate All 0 Teachers Cohort tion Network effect CP 3.4x PM.2.5 CP 1 IHD 1 RD 0 Gan (2011) 1999–2002 Vancouver, Land use Cox PH EC (converted from BC) Only component examined; EC effect CV 0 Canada regression on coronary heart disease about 13x PM2.5 Vedal (2013) 1995–2005 Women’s EPA Specia- Cox PH OC, EC, sulfur OC strongest association with CV CV 0 Health Initiative tion Network mortality; sulfur with first CV event STR 0 ART 0 PCHD 0 FCE 1 Wang (2014) 1985–2012? Europe Land use Cox PH No associations Sulfur and silicon produce highest CVD 0 regression risk estimates Ostro (2015) 2001–2007 California Chemical Cox PH EC, nitrate, SOA; gas and EC highest risk or IHD among com- All 0 Teachers Cohort transport diesel engines, sulfur fuel ponents about 10x PM2.5; no major CV 0.5 model combustion sulfate sources in California IHD 1 RD 0 Chung (2015) 2000–2006 1.25 Medicare EPA Specia- Bayesian EC, nitrate, silicon, sulfate Basic model examined monthly All 1 enrollees, east- tion Network hierarchical changes in exposure; sulfate and Na ern US significantly modified effect of PM2.5 Kioumourtzo- 2000–2010 US EPA Specia- Extended Cox Sulfate, EC vanadium, Clusters of fuel oil combustion, power All 1 glou (2015) tion Network PH nickel plant emissions, EC, and metals have significant regional impacts Thurston (2016) 1982–2004 American Can- EPA Specia- Cox PH EC (traffic tracer) and For IHD, effect of EC and coal is 11.5% IHD 1 cer Society tion Network sulfur (coal and oil com- and 4.6% per ug/m3 versus 1.6% for bustion tracer) PM2.5 (10x and 2.8x PM2.5) Fraction all sig 5/7 Fraction CVD or CP or STR Sig 4.5/8 Fraction RSD sig 0/2 Fraction IHD sig 3/3 Fraction any category significant 9/11 Source: Compiled by the authors of this report based on the works of the authors indicated in the table. Note:  ART = atherosclerotic, BC = black carbon, CP = cardiopulmonary, CV = cardiovascular, CVD = cardiovascular disease, EC = elemental carbon, EPA = United States Environ- mental Protection Agency, FCE = first cardiac event, IHD = ischemic heart disease, IPMN = United States Environmental Protection Agency’s Inhalable Particle Monitoring Network, LC = lung cancer, OC = organic carbon, OCM = organic carbon matter, PCHD = possible coronary heart disease, PH = proportional hazards, RD = respiratory, RSD = respiratory disease, Sec Org = secondary organics, sig = statistically significant, SOA = secondary organic aerosol, STR = stroke. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 65 Mortality PM2.5 OCM EC Nitrate Sulfate Sec Org As, Cl, Se Ca Cu Fe K Mg Mn Na Ni Pb S Si V Zn Outcomes All 1 1 1 All 1 1 CP 1 1 LC 0 0 All 1 1 CP 1 1 CV 1 1 RD 0 0 LC 0 1 All 1 1 CP 1 1 All 0 0 0 0 0 0 0 0 0 CP 1 0 0 1 1 0 0 1 0.5 IHD 1 1 1 1 1 1 1 1 1 RD 0 0 0 0 0 0 0 0 0 CV 0 1 CV 0 1 0 0 0 STR 0 1 0 0 0 ART 0 1 0.5 0.5 0 PCHD 0 0 0 0 0 FCE 1 0 0 1 1 CVD 0 0 0 0 0 0 0 0 0 All 0 0 0 0 0 0 0 0 CV 0.5 0 0 1 0.5 0 0 0 IHD 1 0.5 1 1 1 1 0 0 RD 0 0 0 0 0 0 0 0 All 1 0 1 1 1 1 All 1 1 1 1 1 1 1 1 IHD 1 0 1 1 0 1 0 0 0 0 1 1 0 1 0 5/7 0/4 8/8 6/7 0/2 0/1 1/1 0/1 1/2 2/2 4.5/8 1/2 3/5 4/4 0/2 0/2 1/1 1/2 1/3 0.5/2 0/2 0/2 0/2 0/2 0/2 0/1 0/0 0/1 0/1 0/1 3/3 3/3 3.5/4 2/2 2/3 1/2 1/1 1/2 2/2 1/2 9/11 3/5 8/8 8/8 2/4 1/3 2/3 2/4 2/3 2/4 Appendix 66 TABLE 4 Associations of Long-Term Exposure to PM2.5 Sources with Mortality First Author Years Population Exposures Statistical Significant Species Comment Mortality Traffic (year) Method and Sources Outcomes Ozkaynak (1987) 1980 US IPMN Multiple Sulfate Sulfate effect on all-cause about 3x All regression PM2.5 Dockery (1993), 1977–1991 Six eastern and Six-City Cox PH Sulfate Sulfate effect on all-cause is 2.4x All Krewski (2000) midwestern Data PM2.5 CP cities LC Pope (1995), 1982–1989 American Can- EPA and Cox PH Sulfate Sulfate effect on all-cause is about All Krewski (2000) cer Society IPMN 2x PM2.5 CP CV RD LC Smith (2009) 1982–2000 American Can- EPA and Cox PH Sulfate, EC Sulfate effect on all-cause 2x PM2.5, All cer Society IPMN EC effect 9x PM2.5.; sulfate effect on CP CP 1.2x PM2.5, EC 10x PM2.5 Ostro (2011) 2001–2007 California EPA Cox PH EC, silicon, sulfate EC effect on CP 6x PM2.5; sulfate All Teachers Speciation effect CP 3.4x PM.2.5 CP Cohort Network IHD RD Gan (2011) 1999–2002 Vancouver, Land use Cox PH EC (converted from BC) Only component examined; EC CV Canada regression effect on coronary heart disease about 13x PM2.5 Vedal (2013) 1995–2005 Women’s EPA Cox PH OC, EC, sulfur OC strongest association with CV CV Health Initiative Speciation mortality; sulfur with STR Network first CV event ART PCHD FCE Wang (2014) 1985–2012 Europe Land use Cox PH No associations sulfur and silicon produce highest CVD regression risk estimates Ostro (2015) 2001–2007 California Chemical Cox PH EC, nitrate, SOA; gas EC highest risk or IHD among com- All 0 Teachers transport and diesel engines, ponents about 10x PM2.5; CV 0 Cohort model sulfur fuel combustion no major sulfate sources in Cali- IHD 1 fornia RD 0 Chung (2015) 2000–2006 1.25 Medicare EPA Bayesian EC, nitrate, silicon, Basic model examined monthly All enrollees, Speciation hierarchical sulfate changes in exposure; eastern US Network sulfate and Na significantly modified effect of PM2.5 Kioumourtzo- 2000–2010 US EPA Extended Cox Sulfate, EC vanadium, Clusters of fuel oil combustion, All glou (2015) Speciation PH nickel power plant emissions, Network EC, and metals have significant regional impacts Thurston (2016) 1982–2004 American Can- EPA Cox PH EC (traffic tracer) and For IHD, effect of EC and coal is IHD 0 cer Society Speciation sulfur (coal and oil com- 11.5% and 4.6% per µg/m3 Network bustion tracer) versus 1.6% for PM2.5 (10x and 2.8x PM2.5) Fraction all sign Fraction CVD or CP sig Fraction RSD sig Fraction IHD sig Fraction any category significant ww Source: Compiled by the authors of this report based on the works of the authors indicated in the table. Note:  ART = atherosclerotic, BC = black carbon, CP = cardiopulmonary, CV = cardiovascular, CVD = cardiovascular disease, EC = elemental carbon, EPA = United States Environ- mental Protection Agency, FCE = first cardiac event, IHD = ischemic heart disease, IPMN = United States Environmental Protection Agency’s Inhalable Particle Monitoring Network, LC = lung cancer, OC = organic carbon, PCHD = possible coronary heart disease PH = proportional hazards RD = respiratory, RSD = respiratory disease, sig = statistically significant, SOA = secondary organic aerosol, STR = stroke. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 67 ortality Traffic Powerplant Biomass Crustal Industrial Fuel oil utcomes /coal Process combustion ll ll P C ll P V D C ll P ll P HD D V V TR RT CHD CE VD ll 0 0 1 V 0 0 1 HD 1 0 1 D 0 0 0 ll ll 1 0 0 1 1 HD 0 1 0 0 1 0 References 68 Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 69 References Achilleos, Souzana, Marianthi-Anna Kioumourtzoglou, Chih-Da Wu, Joel D. Schwartz, Petros Koutrakis, and Stefania I. Papatheodorou. 2017. “Acute Effects of Fine Particulate Matter Constituents on Mortali- ty: A Systematic Review and Meta-Regression Analysis.” Environment International 109 (September): 89–100. Atkinson, R. W., S. Kang, H. R. Anderson, C. Mills, and H. Walton. 2014. “Epidemiological Time Series Stud- ies of PM2.5 and Daily Mortality and Hospital Admissions: A Systematic Review and Meta-Analysis.” Thorax 69: 660–65. Basagaña, Xavier, Bénédicte Jacquemin, Angeliki Karanasiou, Bart Ostro, Xavier Querol, David Agis, Ester Alessandrini, Juan Alguacil, Begoña Artiñano, Maria Catrambone, Jesús D. de la Rosa, Julio Díaz, An- nunziata Faustini, Silvia Ferrari, Francesco Forastiere, Klea Katsouyanni, Cristina Linares, Cinzia Per- rino, Andrea Ranzi, Isabella Ricciardelli, Evangelia Samoli, Stefano Zauli-Sajanio, Jordi Sunyer, and Massimo Stafoggia. 2015. “Short-Term Effects of Particulate Matter Constituents on Daily Hospital- izations and Mortality in Five South-European Cities: Results from the MED-PARTICLES Project.” Environment International 75: 151–58. Bell, Michelle L., Keita Ebisu, Brian P. Leaderer, Janneane F. Gent, Hyung Joo Lee, Petros Koutrakis, Yun Wang, Francesca Dominici, and Roger D. Peng. 2013. “Associations of PM2.5 Constituents and Sources with Hospital Admissions: Analysis of Four Counties in Connecticut and Massachusetts (USA) for Persons ≥ 65 Years of Age.” Environmental Health Perspectives 122 (2): 138–44. Bell, Michelle L., Keita Ebisu, Roger D. Peng, Jonathan M. Samet, and Francesca Dominici. 2009. “Hospital Admissions and Chemical Composition of Fine Particle Air Pollution.” American Journal of Respiratory and Critical Care Medicine 179 (12): 1115–20. Brook R. D., Sanjay Rajagopalan, C. Arden Pope, Jeffrey R. Brook, Aruni Bhatnagar, Ana V. Diez-Roux, Fernan- do Holguin, Yuling Hong, Russell V. Luepker, Murray A. Mittleman, Annette Peters, David Siscovick, Sidney C. Smith, Laurie Whitsel, and Joel D. Kaufman. 2010. “Particulate Matter Air Pollution and Cardiovascular Disease. An Update to the Scientific Statement from the American Heart Association.” Circulation 121 (21): 2331–78. Cakmak, Sabit, Robert E. Dales, and Claudia Blanco Vida. 2009. “Components of Particulate Air Pollution and Mortality in Chile.” International Journal of Occupational and Environmental Health 15 (2): 152–58. Cao, Junji, Hongmei Xu, Qun Xu, Bingheng Chen, and Haidong Kan. 2012. “Fine Particulate Matter Consti- tuents and Cardiopulmonary Mortality in a Heavily Polluted Chinese City.” Environmental Health Pers- pectives 120 (3): 373–78. Chen, Renjie, Peng Yin, Xia Meng, Cong Liu, Lijun Wang, Xiaohui Xu, Jennifer A. Ross, Lap A. Tse, Zhuohui Zhao, Haidong Kan, and Maigeng Zhou. 2017. “Fine Particulate Air Pollution and Daily Mortality. A Nationwide Analysis in 272 Chinese Cities.” American Journal of Respiratory and Critical Care Medicine 196 (1): 73–81. Chung, Yeonseung, Francesca Dominici, Yun Wang, Brent A. Coull, and Michelle L. Bell. 2015. “Associations between Long-Term Exposure to Chemical Constituents of Fine Particulate Matter (PM2.5) and Mor- tality in Medicare Enrollees in the Eastern United States.” Environmental Health Perspectives 123: 467– 74. References 70 COMEAP (Committee on the Medical Effects of Air Pollutants). 2009. Long-Term Exposure to Air Pollution: Effect on Mortality. A Report by the Committee on the Medical Effects of Air Pollutants. London, UK. Cooper, John A., and John G. Watson. 1980. “Receptor Oriented Methods of Air Particulate Source Appor- tionment.” Journal of the Air Pollution Control Association 30 (10): 1116–25. Dai, Lingzhen, Antonella Zanobetti, Petros Koutrakis, and Joel D. Schwartz. 2014. “Associations of Fine Par- ticulate Matter Species with Mortality in the United States: A Multicity Time-Series Analysis.” Envi- ronmental Health Perspectives 122 (8): 837–42. Dockery, Douglas W., C. Arden Pope, Xiping Xu, John D. Spengler, James H. Ware, Martha E. Fay, Benjamin G. Ferris Jr., and Frank E. Speizer. 1993. “An Association between Air Pollution and Mortality in six U.S. cities.” New England Journal of Medicine 329: 1753–59. Fang, Ting, Hongyu Guo, Linghan Zeng, Vishal Verma, Athanasios Nenes, and Rodney J. Weber. 2017. “Highly Acidic Ambient Particles, Soluble Metals, and Oxidative Potential: A Link between Sulfate and Aerosol Toxicity.” Environ Sci Technol 51 (5): 2611–20. Gan, Wen Qi, Mieke Koehoorn, Hugh W. Davies, Paul A. Demers, Lillian Tamburic, and Michael Brauer. 2011. “Long-Term Exposure to Traffic-Related Air Pollution and the Risk of Coronary Heart Disease Hospi- talization and Mortality.” Environmental Health Perspectives 119: 501–07. Heo, Jongbae, James Schauer, Okhee Yi, Domyung Paek, Ho Kim, and Seung-Muk Yi. 2014. “Fine Particle Air Pollution and Mortality.” Epidemiology 25 (3): 379–88. Hill, Austin Bradford. 1965. “The Environment and Disease: Association or Causation?” Proceedings of the Royal Society of Medicine 58: 295–300. Huang, Wei, Junji Cao, Yebin Tao, Lingzhen Dai, Shou-En Lu, Bin Hou, Zheng Wang, and Tong Zhu. 2012. “Seasonal Variation of Chemical Species Associated with Short-Term Mortality Effects of PM2.5 in Xi’an, a Central City in China.” American Journal of Epidemiology 175 (6): 556–66. Ito, K., W. F. Christensen, D. J. Eatough, R. C. Henry, E. Kim, F. Laden, R. Lall, T. V. Larson, L. Neas, P. K. Hopke, and G. D. Thurston. 2006. “PM Source Apportionment and Health Effects: 2. An Investigation of Intermethod Variability in Associations between Source-Apportioned Fine Particle Mass and Daily Mortality in Washington, DC.” Journal of Exposure Science & Environmental Epidemiology 16 (4): 300–10. Ito, Kazuhiko, William F. Christensen, Delbert J. Eatough, Ronald C. Henry, Eugene Kim, Francine Laden, Ramona Lall, Timothy V. Larson, Lucas Neas, Philip K. Hopke, and George D. Thurston. 2011. “Fine Particulate Matter Constituents Associated with Cardiovascular Hospitalizations and Mortality in New York City.” Environmental Health Perspectives 119 (4): 467–73. Jones, Rena R., Christian Hogrefe, Edward F. Fitzgerald, Syni-An Hwang, Halûk Özkaynak, Valerie C. Garcia, and Shao Lin. 2015. “Respiratory Hospitalizations in Association with Fine PM and Its Components in New York State.” Journal of the Air & Waste Management Association 65 (5): 559–69. Kim, Sun-Young, Jennifer L. Peel, Michael P. Hannigan, Steven J. Dutton, Lianne Sheppard, Maggie L. Clark, and Sverre Vedal. 2012.” The Temporal Lag Structure of Short-Term Associations of Fine Particulate Matter Chemical Constituents and Cardiovascular and Respiratory Hospitalizations.” Environmental Health Perspectives 120 (8): 1094–99. Kioumourtzoglou, Marianthi-Anna, Brent A. Coull, Francesca Dominici, Petros Koutrakis, Joel Schwartz, and Helen Suh. 2014. “The Impact of Source Contribution Uncertainty on the Effects of Source-Specific Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 71 PM2.5 on Hospital Admissions: A Case Study in Boston, MA.” Journal of Exposure Science and Environ- mental Epidemiology 24 (4): 365–71. Kioumourtzoglou, Marianthi-Anna, Elena Austin, Petros Koutrakis, Francesca Dominici, Joel Schwartz, and Antonella Zanobetti. 2015. “PM2.5 and Survival among Older Adults: Effect Modification by Particulate Composition.” Epidemiology 26 (3): 321–27. Krall, Jenna R., G. Brooke Anderson, Francesca Dominici, Michelle L. Bell, and Roger D. Peng. 2013. “Short- term Exposure to Particulate Matter Constituents and Mortality in a National Study of U.S. Urban Communities.” Environmental Health Perspectives 121 (10): 1148–53. Krall, Jenna R., Amber J. Hackstadt, Roger D. Peng. 2017. “A hierarchical modeling approach to estimate re- gional acute health effects of particulate matter sources.” Statistics in Medicine 36 (9): 1461-75. Krewski, Daniel, Richard Burnett, Mark Goldberg, B. Kristin Hoover, Jack Siemiatycki, Michael Jerrett, Mi- chal Abrahamowicz, and Warren White. 2000. Reanalysis of the Harvard Six Cities Study and the Ameri- can Cancer Society Study of Particulate Air Pollution and Mortality: A Special Report of the Institute’s Particle Epidemiology Reanalysis Project. Cambridge, MA: Health Effects Institute. Laden, F., L. M. Neas, D. W. Dockery, and J. Schwartz. 2000. “Association of Fine Particulate Matter from Different Sources with Daily Mortality in Six U.S. Cities.” Environmental Health Perspectives 108 (10): 941–47. Lall, Ramona, Kazuhiko Ito, and George D. Thurston. 2011. “Distributed Lag Analyses of Daily Hospital Ad- missions and Source-Apportioned Fine Particle Air Pollution.” Environmental Health Perspectives 119 (4): 455–60. Levy, Jonathan I., David Diez, Yiping Dou, Christopher D. Barr, and Francesca Dominici. 2012. “A Meta-Anal- ysis and Multisite Time-Series Analysis of the Differential Toxicity of Major Fine Particulate Matter Constituents.” American Journal of Epidemiology 175 (11): 1091–99. Li, Pei, Jinyuan Xin, Yuesi Wang, Guoxing Li, Xiaochuan Pan, Shigong Wang, Mengtian Cheng, Tx Wen, Guangcheng Wang, and Zirui Liu. 2015. “Association between Particulate Matter and Its Chemical Constituents of Urban Air Pollution and Daily Mortality or Morbidity in Beijing City.” Environmental Science and Pollution Research 22: 358–68. Lippmann, Morton, Lung Chi Chen, Terry Gordon, Kazuhiko Ito, and George D. Thurston. 2013. National Particle Component Toxicity (NPACT) Initiative: “Integrated Epidemiologic and Toxicologic Studies of the Health Effects of Particulate Matter Components.” Research Report, Health Effects Institute 177: 5–13. Lippmann, Morton. 2014. “Toxicological and Epidemiological Studies of Cardiovascular Effects of Ambient Air Fine Particulate Matter (PM2.5) and Its Chemical Components: Coherence and Public Health Im- plications.” Critical Reviews of Toxicology 44: 299–347. Mar, Therese F., Kazuhiko Ito, Jane Q. Koenig, Timothy V. Larson, Delbert J. Eatough, Ronald C. Henry, Eu- gene Kim, Francine Laden, Ramona Lall, Lucas Neas, Matthias Stölzel, Pentti Paatero, Philip K. Hopke, and George D. Thurston. 2006. “PM Source Apportionment and Health Effects. 3. Investigation of Inter-Method Variations in Associations between Estimated Source Contributions of PM2.5 and Daily Mortality in Phoenix, AZ.” Journal of Exposure Science & Environmental Epidemiology 16 (4): 311–20. McMurry, Peter H. 2000. “A Review of Atmospheric Aerosol Measurements.” Atmospheric Environment 34 (12–14): 1959–99. References 72 Mostofsky, Elizabeth, Joel Schwartz, Brent A. Coull, Petros Koutrakis, Gregory A. Wellenius, Helen H. Suh, Diane R. Gold, and Murray A. Mittleman. 2012. “Modeling the Association between Particle Constitu- ents of Air Pollution and Health Outcomes.” American Journal of Epidemiology 176 (4): 317–26. National Research Council. 2004. Research Priorities for Airborne Particulate Matter. IV. Continuing Re- search Progress. Washington, DC: National Academies Press. Ostro, Bart, Wen-Ying Feng, Rachel Broadwin, Shelley Green, and Michael Lipsett. 2007. “The Effects of Components of Fine Particulate Air Pollution on Mortality in California: Results from CALFINE.” Environmental Health Perspectives 115 (1): 13–19. Ostro, Bart D., W.-Y. Feng, R. Broadwin, B. J. Malig, R. S. Green, and M. J. Lipsett. 2008. “The Impact of Components of Fine Particulate Matter on Cardiovascular Mortality in Susceptible Subpopulations.” Occupational and Environmental Medicine 65 (11): 750–56. Ostro, Bart, Michael Lipsett, Peggy Reynolds, Debbie Goldberg, Andrew Hertz, Cynthia Garcia, Katherine D. Henderson, and Leslie Bernstein. 2010. “Long-Term Exposure to Constituents of Fine Particulate Air Pollution and Mortality: Results from the California Teachers Study.” Environmental Health Perspec- tives 118: 363–69. Ostro, Bart, Aurelio Tobias, Xavier Querol, Andrés Alastuey, Fulvio Amato, Jorge Pey, Noemí Pérez, and Jordi Sunyer. 2011. “The Effects of Particulate Matter Sources on Daily Mortality: A Case-Crossover Study of Barcelona, Spain.” Environmental Health Perspectives 119 (12): 1781–87. Ostro, Bart, Brian Malig, Sina Hasheminassab, Kimberly Berger, Emily Chang, and Constantinos Sioutas. 2016. “Associations of Source-Specific Fine Particulate Matter with Emergency Department Visits in California.” American Journal of Epidemiology 184 (6): 450–59. Ostro, Bart, Jianlin Hu, Debbie Goldberg, Peggy Reynolds, Andrew Hertz, Leslie Bernstein, Michael J. Kl- eeman. 2015. “Associations of Mortality with Long-Term Exposures to Fine and Ultrafine Particles, Species and Sources: Results from the California Teachers Study Cohort.” Environmental Health Per- spectives 123 (6): 549–56. Ozkaynak, Halûk, and George D. Thurston. 1987. “Associations between 1980 U.S. Mortality Rates and Alter- native Measures of Airborne Particle Concentration.” Risk Analysis 7: 449–61. Pachon, Jorge, Rodney Weber, Xiaolu Zhang, James Mulholland, and Armistead Russell. 2013. “Revising the Use of Potassium (K) in the Source Apportionment of PM2.5.” Atmospheric Pollution Research 4: 14–21. Pope, C. Arden 3rd, Michael J. Thun, Mohan M. Namboodiri, Douglas W. Dockery, John S. Evans, Frank E. Speizer, Clark W. Heath Jr. 1995. “Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults.” American Journal of Respiratory and Critical Care Medicine 151 (12): 669-74. Pope, C. Arden 3rd, Richard T. Burnett, Michael J. Thun, Eugenia E. Calle, Daniel Krewski, Kazuhiko Ito, and George D. Thurston. 2002. “Lung Cancer, Cardiopulmonary Mortality and Long-Term Exposure to Fine Particulate Air Pollution.” Journal of the American Medical Association 287: 1132–41. Sacks, Jason D., Kazuhiko Ito, William E. Wilson, and Lucas M. Neas. 2012. “Impact of Covariate Models on the Assessment of the Air Pollution-Mortality Association in a Single- and Multipollutant Context.” American Journal of Epidemiology 176 (7): 622–34. Sarnat, Jeremy A., Amit Marmur, Mitchel Klein, Eugene Kim, Armistead G. Russell, Stefanie E. Sarnat, James A. Mulholland, Philip K. Hopke, and Paige E. Tolbert. 2008. “Fine Particle Sources and Cardiorespi- Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 73 ratory Morbidity: An Application of Chemical Mass Balance and Factor Analytical Source Apportion- ment Methods.” Environmental Health Perspectives 116 (4): 459–66. Smith, Kirk R., Michael Jerrett, H. Ross Anderson, Richard T. Burnett, Vicki Stone, Richard Derwent, Richard W. Atkinson, Aaron Cohen, Seth B. Shonkoff, Daniel Krewski, C. Arden Pope III, Michael J. Thun, and George Thurston. 2009. “Public Health Benefits of Strategies to Reduce Greenhouse-Gas Emissions: Health Implications of Short-Lived Greenhouse Pollutants.” Lancet 374: 2091–103. Son, Ji-Young, Jong-Tae Lee, Ki-Hyun Kim, Kweon Jung, and Michelle L. Bell. 2012. “Characterization of Fine Particulate Matter and Associations between Particulate Chemical Constituents and Mortality in Seoul, Korea.” Environmental Health Perspectives 120 (6): 872–78. Spengler, J., and G. Thurston. 1983. “Mass and Elemental Composition of Fine and Coarse particles in 6 Unit- ed States Cities.” Journal of the Air Pollution Control Associations 33: 1162–71. Thurston, George D., Haluk Ozkaynak, and Arthur Schatz. 1984. “A Chemical Characterization and Source Apportionment of the IP Network Fine Particle Data.” Proceedings of the 77th Annual Meeting of the Air Pollution Control Association. San Francisco, California. Thurston, George D., Kazuhiko Ito, Therese Mar, William F. Christensen, Delbert J. Eatough, Ronald C. Hen- ry, Eugene Kim, Francine Laden, Ramona Lall, Timothy V. Larson, Hao Liu, Lucas Neas, Joseph Pinto, Matthias Stölzel, Helen Suh, and Philip K. Hopke. 2005. “Workgroup Report: Workshop on Source Apportionment of Particulate Matter Health Effects—Intercomparison of Results and Implications.” Environmental Health Perspectives 113 (12): 1768–74. Thurston, George D., Kazuhiko Ito, and Ramona Lall. 2011. “A Source apportionment of U.S. Fine Particulate Matter Air Pollution.” Atmospheric Environment 45: 3924–36. Thurston, George D. Kazuhiko Ito, Ramona Lall, Richard T. Burnett, Michelle C. Turner, Daniel Krewski, Yu- anli Shi, Michael Jerrett, Susan M. Gapstur, W. Ryan Diver, and C. Arden Pope III. 2013. “NPACT Study 4. Mortality and Long-Term Exposure to PM2.5 and Its Components in the American Cancer Society’s Cancer Prevention Study II Cohort.” National Particle Component Toxicity (NPACT) Initiative: Integrat- ed Epidemiologic and Toxicologic Studies of the Health Effects of Particulate Matter Components. Research Report 177. Boston, MA: Health Effects Institute. Thurston, George D., Richard T. Burnett, Michelle C. Turner, Yuanli Shi, Daniel Krewski, Ramona Lall, Ka- zuhiko Ito, Michael Jerrett, Susan M. Gapstur, W. Ryan Diver, and C. Arden Pope III. 2016. “Ischemic Heart Disease Mortality and Long-Term Exposure to Source-Related Components of U.S. Fine Parti- cle Air Pollution.” Environmental Health Perspectives 124: 785–94. US EPA (US Environmental Protection Agency). 1984. Receptor Model Technical Series, Volume V: Source Apportionment Techniques and Considerations in Combining Their Use. Research Triangle Park, NC, EPA-450/4-84-020. US EPA (US Environmental Protection Agency). 2009. Integrated Science Assessment for Particulate Matter (Final Report). Washington, DC, EPA/600/R-08/139F. Vedal, Sverre, Matthew J. Campen, Jacob D. McDonald, Joel D. Kaufman, Timothy V. Larson, Paul D. Samp- son, Lianne Sheppard, Christopher D. Simpson, and Adam A. Szpiro. 2013. “Section 1. NPACT Epide- miologic Study of Components of Fine Particulate Matter and Cardiovascular Disease in the MESA and WHI-OS Cohorts.” National Particle Component Toxicity (NPACT) Initiative Report on Cardiovascu- lar Effects. Research Report 178. Boston, MA: Health Effects Institute. References 74 Wang, Meng, Rob Beelen, Massimo Stafoggia, Ole Raaschou-Nielsen, Zorana Jovanovic Andersen, Barbara Hoffmann, Paul Fischer, Danny Houthuijs, Mark Nieuwenhuijsen, Gudrun Weinmayr, Paolo Vineis, Wei W. Xun, Konstantina Dimakopoulou, Evangelia Samoli, Tiina Laatikainen, Timo Lanki, Anu W.Turunen, Bente Oftedal, Per Schwarze, Geir Aamodt, Johanna Penell, Ulf De Faire, Michal Ko- rek, Karin Leander, Göran Pershagen, Nancy L. Pedersen, Claes-Göran Östenson, Laura Fratiglioni, Kirsten Thorup Eriksen, Mette Sørensen, Anne Tjønneland, Bas Bueno-de-Mesquita, Marloes Eeftens, Michiel L. Bots, Kees Meliefste, Ursula Krämer, Joachim Heinrich, Dorothea Sugiri, Timothy Key, Kees de Hoogh, Kathrin Wolf, Annette Peters, Josef Cyrys, Andrea Jaensch, Hans Concin, Gabriele Nagel, Ming-Yi Tsai, Harish Phuleria, Alex Ineichen, Nino Künzli, Nicole Probst-Hensch, Emmanuel Schaff- ner, Alice Vilier, Françoise Clavel-Chapelon, Christophe Declerq, Fulvio Ricceri, Carlotta Sacerdote, Alessandro Marcon, Claudia Galassi, Enrica Migliore, Andrea Ranzi, Giulia Cesaroni, Chiara Badaloni, Francesco Forastiere, Michail Katsoulis, Antonia Trichopoulou, Menno Keuken, Aleksandra Jedynska, Ingeborg M. Kooter, Jaakko Kukkonen, Ranjeet S. Sokhi, Bert Brunekreef, Klea Katsouyanni, and Ge- rard Hoeka. 2014. “Long Term Exposure to Elemental Constituents of Particulate Matter and Cardio- vascular Mortality in 19 European Cohorts: Results from the ESCAPE and TRANSPHORM Projects.” Environment International 66: 97–106. Are All Air Pollution Particles Equal? How Constituents and Sources of Fine Air Pollution Particles (PM2.5) Affect Health 75 WWW.WORLDBANK.ORG