Particulate Matter (PM) Epidemiology and Cardiopulmonary Disease

RoseyFields evaluates the human epidemiological evidence for particulate matter as follows. This summary catalogs the populations, exposure metrics, and health outcomes examined in the published epidemiological literature on ambient particulate matter (PM), characterizes the strength and consistency of the evidence for each outcome, and describes how RoseyFields integrates this evidence with the toxicological record to support general and specific causation opinions in litigation and regulatory matters. Because PM is a ubiquitous ambient exposure rather than a workplace or point-source contaminant, this file necessarily draws more heavily on general-population cohort and time-series literature than RoseyFields’ single-chemical occupational files, and its exposure assessment methods differ accordingly (see Section 1.6).

1. Overview of Study Populations and Data Sources

The epidemiological record for particulate matter is the largest and most methodologically diverse of any pollutant RoseyFields evaluates, spanning multiple study designs across six continents and more than three decades of continuous research.

1.1 Prospective General-Population Mortality Cohorts

The foundational studies in this literature are prospective cohort mortality studies that enrolled general-population adults and followed them for one to several decades while tracking ambient PM exposure at their residence. The Harvard Six Cities Study enrolled over 8,000 adults across six U.S. cities beginning in the 1970s and has been followed, extended, and reanalyzed repeatedly over more than 35 years (Dockery et al., 1993; Krewski et al., 2000; Laden et al., 2006; Lepeule et al., 2012). The American Cancer Society Cancer Prevention Study II (ACS CPS-II) cohort, encompassing more than 500,000 adults across approximately 150 U.S. metropolitan areas, produced the other foundational long-term mortality analyses (Pope et al., 1995, 2002, 2004). Both cohorts were the subject of an independent, multi-year reanalysis commissioned by the Health Effects Institute in response to concerns about the studies’ pivotal regulatory role (Krewski et al., 2000), which confirmed the original investigators’ central findings using independently audited data and alternative statistical approaches.

1.2 Large-Scale Administrative/Claims-Based Cohorts

More recently, national health insurance claims databases have enabled construction of cohorts orders of magnitude larger than the original Six Cities and ACS studies. A study of the entire U.S. Medicare population (more than 60 million person-years) linked individual-level enrollment and mortality records to high-resolution modeled PM2.5 exposure at the ZIP-code level (Di et al., 2017), with cause-specific extensions examining respiratory, cardiovascular, and cancer mortality separately (Pun et al., 2017). This administrative-cohort design trades the detailed individual risk factor data (smoking, occupation, diet) available in the Six Cities and ACS cohorts for vastly greater sample size and statistical power, particularly useful for examining effects at the lower exposure concentrations now prevalent in the United States.

1.3 Daily Time-Series (Short-Term Exposure) Studies

A methodologically distinct body of evidence examines short-term (daily to weekly) fluctuations in ambient PM concentration against same-day or lagged daily counts of mortality or hospital admission in a given city or region, using each city as its own control across time and thereby inherently controlling for any fixed individual-level or area level confounder. The National Morbidity, Mortality, and Air Pollution Study (NMMAPS), covering the 20 largest U.S. cities, is the principal U.S. multi-city time-series analysis (Samet et al., 2000). This design is complementary to, and provides an important cross-check on, the long-term cohort mortality studies described above, because it is subject to an entirely different set of potential biases (see Section 5).

1.4 Women’s Health and Sex-Specific Cohorts

A cohort of 65,893 postmenopausal women enrolled in the Women’s Health Initiative Observational Study, all free of cardiovascular disease at baseline, was followed for incident (not merely fatal) cardiovascular events in relation to residential PM2.5 exposure (Miller et al., 2007), providing an important extension of the mortality-only findings from the general-population cohorts to incident, nonfatal cardiovascular disease.

1.5 Children’s and Birth Cohorts

The Southern California Children’s Health Study prospectively followed more than 1,700 children across 12 communities with a wide range of ambient PM, ozone, and nitrogen dioxide exposure, measuring lung function annually from age 10 to 18 (Gauderman et al., 2004), and later documented improvement in childhood lung function growth coincident with a documented reduction in regional air pollution over time (Gauderman et al., 2015). A pooled analysis of five European birth cohorts under the ESCAPE project examined childhood lung function at age 6–8 in relation to standardized residential air pollution exposure estimates (Gehring et al., 2013). Separately, a substantial and continually growing body of birth-cohort literature has examined maternal PM exposure during pregnancy against birth weight and gestational-age outcomes, synthesized in systematic reviews and meta-analyses (Stieb et al., 2012; Sun et al., 2016).

1.6 Exposure Assessment: A Structural Difference from Single-Chemical Epidemiology

Unlike the single-chemical files in RoseyFields’ library (e.g., carbon tetrachloride, ethylene oxide), PM epidemiology does not generally rely on serum, urine, or breath biomarkers of individual exposure, because ambient PM is a complex, heterogeneous physical mixture rather than a single measurable analyte with a validated internal-dose biomarker. Instead, exposure assignment in this literature has evolved from simple nearest-monitor concentration data (used in the original Six Cities and ACS analyses) to increasingly sophisticated modeled exposure surfaces — land-use regression, chemical transport models, and, most recently, satellite-derived aerosol optical depth calibrated against ground monitors — that allow individual residential exposure estimates even for cohort members far from a fixed monitoring station (as used in the Medicare, WHI, and ESCAPE analyses). RoseyFields treats this progression in exposure-assessment sophistication as a material factor improving confidence in the more recent literature relative to the earliest monitor-based studies, and addresses its residual uncertainty in Section 5.

1.7 Value of Using Multiple, Diverse Populations

RoseyFields’ methodology places particular weight on the fact that the mortality-cohort, administrative-cohort, time series, women’s-health, and children’s/birth-cohort literatures each have different, largely non-overlapping sources of potential bias — individual confounding and healthy-volunteer effects in the enrolled cohorts; area-level confounding and lack of individual risk-factor data in the administrative and time-series studies; and different exposure-assessment methods across all of them. A finding that recurs consistently across all of these designs, across multiple decades, and across at least three continents — as is the case for all-cause and cardiopulmonary mortality — is accordingly among the most robust associations in all of environmental epidemiology. Conversely, outcomes supported by only one or two of these population types, or with materially inconsistent results across designs (as with diabetes, discussed in Section 2.4), warrant a correspondingly more circumspect evidentiary characterization.

2. Health Outcomes Investigated in Humans, by Body System

The following organizes the principal outcomes identified in the peer-reviewed human literature, grouped by body system, and classified into three evidentiary tiers: (A) strong, consistent, quantitatively modeled evidence; (B) suggestive but inconsistent evidence; and (C) outcomes investigated with null or inadequate findings. Given the size of the PM literature, this section focuses on the outcomes with the deepest and most frequently litigated evidentiary record rather than attempting an exhaustive catalog of every published association.

2.1 Cardiovascular System

Tier A — Strong, Consistent Evidence: Cardiopulmonary and Cardiovascular Mortality

Long-term PM2.5 exposure is associated with increased cardiopulmonary and cardiovascular mortality with a consistency and quantitative reproducibility that is unusual in environmental epidemiology. The original Six Cities analysis found a statistically significant association between long-term PM2.5 exposure and both all-cause and cardiopulmonary mortality (Dockery et al., 1993); the ACS cohort independently confirmed this association at national scale (Pope et al., 1995) and subsequently documented a specific, biologically coherent cardiovascular mortality gradient with proposed mechanistic pathways (Pope et al., 2002, 2004). The independent Health Effects Institute reanalysis of both cohorts confirmed the core findings (Krewski et al., 2000), and extended Six Cities follow-up through 2009 found a persistent, approximately linear concentration-response relationship for cardiovascular mortality down to concentrations as low as 8 µg/m3 (Lepeule et al., 2012). At a scale of tens of millions of subjects, the Medicare cohort confirmed elevated cardiovascular and respiratory mortality risk with no evidence of a population-level threshold (Di et al., 2017; Pun et al., 2017).

Tier A — Strong, Consistent Evidence: Incident (Nonfatal) Cardiovascular Events

The Women’s Health Initiative cohort extended the mortality-only findings above to incident, nonfatal cardiovascular events in a cardiovascular-disease-free cohort at enrollment, finding a statistically significant increase in first cardiovascular events (myocardial infarction, coronary revascularization, and cerebrovascular events combined) per unit increase in long-term PM2.5 exposure (Miller et al., 2007), providing important confirmation that the mortality associations above are not solely attributable to accelerated death among those already critically ill.

Tier B — Suggestive but Inconsistent Evidence: Stroke, Hypertension, and Arrhythmia

Findings for cerebrovascular disease (stroke) specifically, as opposed to the broader cardiovascular composite endpoints above, are less consistent: several cohort studies in North America, Europe, and Asia report positive associations between long-term PM exposure and incident stroke or cerebrovascular mortality, but a dedicated etiological sub-analysis within the Women’s Health Initiative cohort found the evidence linking long-term PM exposure specifically to incident stroke to be less clear than the evidence for cardiovascular disease overall. Findings for incident hypertension and cardiac arrhythmia are similarly positive in direction across most, but not all, of the studies that have examined these more specific endpoints, and are properly classified as suggestive rather than firmly established.

2.2 Respiratory System

Tier A — Strong, Consistent Evidence: Impaired Childhood Lung Function Development

Long-term exposure to PM and related traffic-associated pollutants during childhood is associated with clinically and statistically significant deficits in lung function growth. The Southern California Children’s Health Study found that children in the highest-PM communities had substantially reduced growth in forced expiratory volume (FEV1) over an eight-year follow-up compared to children in the lowest-PM communities, with an effect size comparable to that reported for exposure to maternal smoking (Gauderman et al., 2004). A subsequent analysis of three sequential cohorts from the same study, spanning a period of substantial regional air-quality improvement, found a corresponding improvement in childhood lung function growth as ambient PM and other pollutants declined over time — a finding of particular importance because it demonstrates a within-population, exposure-tracking change rather than a cross sectional association alone (Gauderman et al., 2015). The independently conducted, standardized European ESCAPE pooled analysis of five birth cohorts corroborated reduced childhood lung function in association with air pollution exposure using an entirely separate population and exposure-assessment protocol (Gehring et al., 2013).

Tier A — Strong, Consistent Evidence: Respiratory Mortality and Lung Cancer

Long-term PM2.5 exposure is associated with increased lung cancer mortality in both of the foundational U.S. cohorts (Pope et al., 2002; Lepeule et al., 2012) and in the more recent, much larger Medicare-based analysis for respiratory mortality broadly, with a distinct cause-specific decomposition showing consistent cardiovascular and respiratory mortality associations across nearly the entire studied concentration range (Pun et al., 2017); the same Medicare analysis, notably, did not find PM2.5 associated with lung cancer mortality specifically, illustrating that even within

Tier A outcomes, effect estimates for narrower cause-specific endpoints can differ across studies of different design and population, which RoseyFields addresses further in Section 5.

Tier B — Suggestive but Inconsistent Evidence: Asthma Incidence

Evidence linking long-term PM exposure to new-onset (incident) childhood asthma, as distinct from exacerbation of pre-existing asthma, is directionally positive across most cohort studies but less uniformly consistent in magnitude and statistical significance than the lung-function-growth findings above, and RoseyFields classifies asthma incidence as a suggestive, rather than firmly established, outcome.

2.3 All-Cause Mortality and Life Expectancy

Tier A — Strong, Consistent Evidence

All-cause mortality findings parallel the cardiopulmonary-specific findings in Section 2.1 and are supported by the same body of foundational and confirmatory cohort literature (Dockery et al., 1993; Pope et al., 1995; Krewski et al., 2000; Laden et al., 2006; Lepeule et al., 2012; Di et al., 2017), as well as by the independent short-term time-series literature, which found a consistent, statistically significant association between day-to-day fluctuation in PM concentration and same-week mortality across the 20 largest U.S. cities using a design inherently robust to fixed individual- and area-level confounding (Samet et al., 2000). The convergence of the long-term cohort and short-term time-series literatures — two entirely different study designs with different bias structures, both showing a consistent positive association — is a principal reason this outcome sits at the top of RoseyFields’ evidentiary hierarchy for PM.

2.4 Endocrine/Metabolic System

Tier B/C — Inconsistent Evidence: Type 2 Diabetes

The literature examining long-term PM exposure and incident type 2 diabetes is a clear illustration of a genuinely inconsistent evidentiary picture. Several cohort studies have reported statistically significant positive associations (Chen et al., 2013, in a large Ontario administrative cohort; Hansen et al., 2016, in a Danish cohort), while others have found no significant association (Weinmayr et al., 2015; Puett et al., 2011), and at least two have reported a non significant inverse association (Renzi et al., 2018). RoseyFields classifies this outcome as suggestive at best (Tier B, bordering on Tier C), reflecting a genuine, unresolved split in the primary literature rather than a body of evidence that merely has wide confidence intervals around a consistently positive point estimate.

2.5 Neurological System

Tier B — Suggestive but Inconsistent Evidence: Dementia and Cognitive Decline

A large Ontario population-based cohort study (approximately 2.1 million adults) found a statistically significant association between long-term residential PM2.5 exposure and incident dementia diagnosis, using validated health administrative-data case ascertainment (Chen et al., 2017). This finding is broadly, though not uniformly, corroborated by a growing number of subsequent cohort studies in Europe, Asia, and North America using varied dementia/cognitive-decline case definitions and exposure-assessment methods, several of which are summarized in independent systematic reviews concluding that the evidence, while suggestive of an association, remains insufficiently mature — particularly with respect to reverse causation, differential diagnostic ascertainment, and residual confounding by cardiovascular risk factors that independently affect both dementia risk and, potentially, residential location — to be classified as an established causal outcome at this time.

2.6 Reproductive and Developmental Outcomes

Tier B — Suggestive, Reasonably Consistent Direction but Methodologically Heterogeneous

A large and rapidly growing literature has examined maternal PM2.5 exposure during pregnancy in relation to birth weight, preterm birth, and related outcomes. A systematic review and meta-analysis of ambient air pollution and birth outcomes found a consistent, direction-stable association between PM2.5 exposure and reduced birth weight and increased preterm birth risk across the pooled literature (Stieb et al., 2012), a finding corroborated by a subsequent, larger meta-analysis focused specifically on PM2.5 and birth weight (Sun et al., 2016). RoseyFields classifies this body of evidence as suggestive rather than Tier A, notwithstanding the consistent direction of effect across a large number of studies, because of substantial heterogeneity across studies in exposure window definition (whole-pregnancy versus trimester-specific versus critical-window models), outcome ascertainment, and the degree of control for important co occurring exposures (maternal smoking, socioeconomic status, other ambient co-pollutants such as ozone and nitrogen dioxide), such that the pooled meta-analytic point estimates, while directionally consistent, vary more across the underlying primary literature than the mortality findings in Sections 2.1 and 2.3.

2.7 Other Outcomes

Tier C — Null or Inadequate Evidence

A number of other outcomes — including chronic kidney disease, site-specific cancers other than lung, and specific neurodevelopmental diagnoses in children (e.g., autism spectrum disorder, attention-deficit/hyperactivity disorder) — have been examined in a smaller number of studies with results that are, at present, inconsistent, based on limited replication, or otherwise inadequate to support even a suggestive tiering. RoseyFields treats these as evidence gaps warranting monitoring of the emerging literature rather than either an affirmative or a null classification.

3. Summary Table of Evidence Tiers

4. Detailed Review of Strong-Evidence Outcomes

4.1 All-Cause and Cardiopulmonary Mortality

Exposure Metrics: The foundational cohort studies assigned exposure using the ambient monitor located in or nearest to each subject’s city or metropolitan area, generally averaged over multiple years to approximate chronic exposure (Dockery et al., 1993; Pope et al., 1995); later analyses of the same cohorts, and the newer Medicare-based analysis, used increasingly refined modeled exposure surfaces incorporating satellite data, land-use variables, and chemical transport modeling to assign residence-specific rather than city-average exposure (Di et al., 2017). Effect Sizes/Direction: Across these studies, each 10 µg/m3 increase in long-term PM2.5 exposure is associated with an approximately 6–14% increase in all-cause mortality and a somewhat larger relative increase in cardiopulmonary specific mortality, with the extended Six Cities follow-up finding the concentration-response relationship to be approximately linear down to 8 µg/m3 with no evidence of a population threshold (Lepeule et al., 2012), a finding corroborated at much larger scale by the Medicare cohort (Di et al., 2017). Consistency: This is the most consistently replicated quantitative exposure-response relationship in all of PM epidemiology, recurring across at least four independent cohort designs (Six Cities, ACS, Medicare, and Women’s Health Initiative), an independent reanalysis (Krewski et al., 2000), and an entirely separate short-term time-series literature (Samet et al., 2000), spanning nearly 40 years of U.S. data collection under substantially declining ambient PM concentrations throughout.

4.2 Childhood Lung Function Development

Exposure Metrics: The Children’s Health Study assigned exposure using fixed community-level monitors representing each of the 12 study communities, selected specifically to span a wide gradient of ambient PM, ozone, acid vapor, and nitrogen dioxide concentrations (Gauderman et al., 2004), while the later analysis leveraged a documented, region wide decline in ambient concentrations over calendar time to examine within-population change (Gauderman et al., 2015); the European ESCAPE pooled analysis used a standardized land-use regression protocol applied uniformly across birth cohorts in four countries (Gehring et al., 2013). Effect Sizes/Direction: Children in the highest-PM communities exhibited substantially reduced 8-year growth in FEV1 relative to children in the lowest-PM communities, with the risk of clinically low lung function at age 18 several-fold higher in the highest- versus lowest exposure community (Gauderman et al., 2004); the follow-up analysis found lung function growth measurably improved in the more recent cohort cohorts that experienced lower ambient pollution during the same developmental window (Gauderman et al., 2015). Consistency: The direction and approximate magnitude of effect is corroborated in an entirely independent population and exposure-assessment framework (Gehring et al., 2013), and the exposure tracking design of the improved-air-quality analysis (Gauderman et al., 2015) provides a form of natural-experiment confirmation that is considered particularly persuasive evidence of a causal, rather than purely confounded, relationship.

4.3 Short-Term (Daily) Mortality Association

Exposure Metrics: NMMAPS and related time-series analyses use routinely collected daily ambient PM monitoring data for a given city, matched to daily mortality counts for the same city, with statistical adjustment for daily weather variables, day-of-week, and long-term/seasonal trends (Samet et al., 2000). Effect Sizes/Direction: A statistically significant, modest increase in same-day and lagged mortality was associated with each unit increase in daily PM concentration, pooled across the 20 largest U.S. cities. Consistency: Because this design uses each city as its own control across time, it is inherently robust to any fixed individual-level confounder (e.g., smoking status, socioeconomic status) or fixed area-level confounder (e.g., regional industrial composition) that does not vary from day to day — a fundamentally different set of strengths and weaknesses than the long-term cohort designs in Section 4.1 — such that the convergence of both literatures on a consistent, positive PM-mortality association substantially strengthens overall causal confidence.

5. Study Quality Considerations

RoseyFields’ confidence in any individual PM epidemiological finding is calibrated against the following recurring quality considerations, several of which are structurally different from those affecting RoseyFields’ single-chemical occupational epidemiology files.

5.1 Exposure Misclassification

Because PM exposure assignment relies on modeled or monitored ambient concentration rather than a personal or biological measurement, some degree of exposure misclassification is inherent to every study design in this literature, and the degree of misclassification has generally declined over time as exposure-assessment methods have advanced from simple nearest-city-monitor assignment (the original Six Cities and ACS analyses) to individual residence-level modeled surfaces incorporating satellite and land-use data (the Medicare, WHI, and ESCAPE analyses). Non differential exposure misclassification of this kind would generally be expected to bias effect estimates toward the null, meaning the true effect could plausibly be somewhat larger than reported in the earliest studies; this is one reason RoseyFields places particular weight on the more recent, more refined-exposure-assessment literature when it is available, without discounting the qualitative direction of the earlier foundational studies.

5.2 Co-Pollutant Confounding

Ambient PM rarely occurs in isolation; it is correlated, to varying degrees depending on location and source mix, with other ambient pollutants including ozone, nitrogen dioxide, sulfur dioxide, and traffic-related pollutants more broadly. Most of the key studies discussed above (e.g., the Children’s Health Study; Gauderman et al., 2004) explicitly modeled multiple co-pollutants and generally found PM-specific effects that were robust to co-pollutant adjustment, but RoseyFields recognizes that fully disentangling the independent contribution of PM from the broader ambient pollution mixture remains an area of ongoing scientific refinement, particularly for outcomes (such as birth weight and preterm birth; Section 2.6) where co-pollutant adjustment has been less uniformly applied across the underlying primary studies pooled in the meta-analyses.

5.3 Confounding by Individual and Area-Level Socioeconomic Factors

Because ambient PM concentration is not randomly distributed across a population — it is systematically correlated with urbanicity, proximity to industrial or traffic sources, and, in many settings, area-level socioeconomic status — unmeasured or incompletely measured confounding by individual smoking status, occupational exposure, diet, and socioeconomic factors is a genuine concern for the cohort literature. The original Six Cities and ACS analyses addressed this directly by collecting individual-level data on smoking, education, and body mass index (Dockery et al., 1993; Pope et al., 1995), and the Women’s Health Initiative analysis similarly adjusted for an extensive individual risk factor panel (Miller et al., 2007); the independent Health Effects Institute reanalysis specifically tested, and did not find material evidence for, confounding as an alternative explanation for the core mortality findings (Krewski et al., 2000). Confounding by area-level socioeconomic factors is comparatively harder to fully exclude in the large administrative cohort (Medicare) and time-series designs, which generally have less-detailed individual risk-factor data, though the internal, within-city time-series design in Section 4.3 is structurally robust to any fixed area-level confounder.

5.4 Reverse Causation

Reverse causation is not a material concern for the childhood lung-function and mortality literature (Sections 4.1–4.2), given the prospective cohort designs and biologically implausible reverse pathway (disease does not cause ambient PM levels). It is a more relevant consideration for the dementia/cognitive-decline literature (Section 2.5), where an alternative explanation — that early, undiagnosed cognitive impairment could plausibly influence where and how a person lives, or that shared cardiovascular risk factors independently drive both residential exposure patterns and dementia risk — has not been fully excluded across the existing literature, which is one of the principal reasons RoseyFields classifies this outcome as suggestive rather than established.

5.5 The Complementary Role of Time-Series and Cohort Designs

As discussed in Section 4.3, the daily time-series literature and the long-term cohort literature have essentially non overlapping strengths and weaknesses — the former is robust to fixed confounding but cannot capture the cumulative health effects of sustained chronic exposure, while the latter captures chronic exposure effects but requires more extensive individual-level confounder control. RoseyFields treats the consistent convergence of both literatures on a positive PM-mortality association as one of the strongest possible forms of methodological triangulation available in observational epidemiology, and applies the same triangulation logic when evaluating the relative confidence warranted for outcomes (e.g., diabetes, dementia) supported by only one study design or population type.

6. Independent Expert Panel and Agency Weight-of-Evidence Conclusions

The International Agency for Research on Cancer’s Monograph Working Group concluded that outdoor air pollution, and particulate matter as a major component of that mixture, causes lung cancer in humans and classified both outdoor air pollution and particulate matter specifically in Group 1 (carcinogenic to humans), a conclusion published in the peer-reviewed literature by the Working Group (Loomis et al., 2013) and based on the convergence of the cohort mortality literature summarized in Sections 2.2 and 4.1, supporting mechanistic and animal toxicological evidence, and the biological plausibility of particle-induced pulmonary carcinogenesis.

Beyond the cancer-specific classification, RoseyFields’ tiered evidentiary framework in Section 3 is broadly consistent with the causal-determination framework used by expert scientific review panels internationally in evaluating this literature: a ‘causal’ determination is generally reserved for outcomes (all-cause and cardiovascular mortality, respiratory effects including impaired lung development) supported by consistent findings across multiple study designs and populations with a coherent, biologically plausible exposure-response relationship, corresponding to RoseyFields’ own Tier A; a ‘likely to be causal’ or ‘suggestive’ determination is generally applied to outcomes (reproductive and developmental effects, nervous system effects, metabolic effects) with a positive but less uniformly consistent evidentiary record, corresponding to RoseyFields’ own Tier B; and ‘inadequate’ determinations are reserved for outcomes with a limited, inconsistent, or contradictory evidentiary record, corresponding to RoseyFields’ own Tier C. RoseyFields’ independent classification was derived directly from the primary peer-reviewed literature summarized above and was not adopted wholesale from any single regulatory determination, though it is gratifying that the two approaches converge closely for nearly every outcome addressed in this summary.

7. Use of Epidemiological Evidence to Support Causation Opinions

7.1 General Causation

RoseyFields’ general causation opinions integrate the epidemiological tiering in Section 3 with the toxicological and mechanistic record for particulate matter (addressed in RoseyFields’ companion toxicology methodology materials for this pollutant). For Tier A outcomes — all-cause and cardiopulmonary mortality, incident cardiovascular events, and impaired childhood lung function development — RoseyFields’ opinion is that the totality of the evidence, spanning multiple independent cohort designs, an independent reanalysis, and a structurally distinct short-term time-series literature, is sufficient to support general causation for long-term ambient PM exposure at concentrations within the range studied in the foundational cohorts. For Tier B outcomes (stroke and other specific cardiovascular endpoints, dementia and cognitive decline, adverse birth outcomes, childhood asthma incidence), RoseyFields’ opinion is that the human evidence is suggestive and biologically plausible but not, standing alone, independently sufficient to establish general causation with the same confidence as the Tier A outcomes, and RoseyFields’ opinions in matters involving these outcomes are correspondingly more circumspect and more heavily dependent on the case-specific facts and the state of the then-current literature. For Tier C outcomes and the genuinely inconsistent diabetes literature (Section 2.4), RoseyFields does not offer an opinion that the epidemiological evidence, standing alone, establishes general causation.

7.2 Specific Causation

For specific causation opinions in individual matters, RoseyFields’ epidemiologists and toxicologists layer the general causation framework above against case-specific facts, which for an ambient pollutant like PM presents distinctive challenges relative to RoseyFields’ occupational chemical files: reconstruction of an individual’s residential and occupational exposure history (using the same categories of monitoring, land-use regression, or satellite-derived modeled exposure data that underlie the cohort studies) is necessarily an estimate of ambient concentration at the person’s residence or workplace, rather than a direct personal exposure measurement, and must further account for time spent indoors versus outdoors, use of air filtration, and other factors affecting personal exposure relative to ambient concentration; the specific health outcome(s) alleged and their consistency with the Tier A/B/C framework above; the magnitude and duration of the individual’s estimated cumulative exposure relative to the exposure ranges studied in the relevant cohort(s); latency and biological plausibility of the alleged disease process; and the presence of alternative or competing causes, including smoking, occupational exposures, genetic and family history factors, and other individual cardiovascular, metabolic, or respiratory risk factors. Given the ubiquity of ambient PM exposure and the population level (rather than point-source) nature of most of the underlying epidemiology, RoseyFields’ specific causation opinions for PM-related matters often place particular emphasis on quantifying the incremental exposure attributable to a specific alleged source (e.g., a nearby facility or activity) against regional background ambient PM levels, since general causation for ambient PM as a category does not by itself establish that a specific, discrete source made a material contribution to an individual’s cumulative exposure or disease risk.

8. Bibliography

(Author-year citations above correspond to the following peer-reviewed epidemiological sources and independent expert/agency determinations, listed alphabetically.)

Chen, H; Burnett, RT; Kwong, JC; et al. (2013) Risk of incident diabetes in relation to long-term exposure to fine particulate matter in Ontario, Canada. Environ Health Perspect 121:804–810.

Chen, H; Kwong, JC; Copes, R; et al. (2017) Exposure to ambient air pollution and the incidence of dementia: a population-based cohort study. Environ Int 108:271–277.

Di, Q; Wang, Y; Zanobetti, A; et al. (2017) Air pollution and mortality in the Medicare population. N Engl J Med 376:2513–2522.

Dockery, DW; Pope, CA III; Xu, X; et al. (1993) An association between air pollution and mortality in six U.S. cities. N Engl J Med 329:1753–1759.

Gauderman, WJ; Avol, E; Gilliland, F; et al. (2004) The effect of air pollution on lung development from 10 to 18 years of age. N Engl J Med 351:1057–1067.

Gauderman, WJ; Urman, R; Avol, E; et al. (2015) Association of improved air quality with lung development in children. N Engl J Med 372:905–913.

Gehring, U; Gruzieva, O; Agius, RM; et al. (2013) Air pollution exposure and lung function in children: the ESCAPE project. Environ Health Perspect 121:1357–1364.

Hansen, AB; Ravnskjær, L; Loft, S; et al. (2016) Long-term exposure to fine particulate matter and incidence of diabetes in the Danish Nurse Cohort. Environ Int 91:243–250.

Krewski, D; Burnett, RT; Goldberg, MS; et al. (2000) Reanalysis of the Harvard Six Cities Study and the American 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; Schwartz, J; Speizer, FE; Dockery, DW. (2006) Reduction in fine particulate air pollution and mortality: extended follow-up of the Harvard Six Cities study. Am J Respir Crit Care Med 173:667–672.

Lepeule, J; Laden, F; Dockery, D; Schwartz, J. (2012) Chronic exposure to fine particles and mortality: an extended follow-up of the Harvard Six Cities study from 1974 to 2009. Environ Health Perspect 120:965–970.

Loomis, D; Grosse, Y; Lauby-Secretan, B; et al. (2013) The carcinogenicity of outdoor air pollution. Lancet Oncol 14:1262–1263.

Miller, KA; Siscovick, DS; Sheppard, L; et al. (2007) Long-term exposure to air pollution and incidence of cardiovascular events in women. N Engl J Med 356:447–458.

Pope, CA III; Thun, MJ; Namboodiri, MM; et al. (1995) Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am J Respir Crit Care Med 151:669–674.

Pope, CA III; Burnett, RT; Thun, MJ; et al. (2002) Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 287:1132–1141.

Pope, CA III; Burnett, RT; Thurston, GD; et al. (2004) Cardiovascular mortality and long-term exposure to particulate air pollution: epidemiological evidence of general pathophysiological pathways of disease. Circulation 109:71–77.

Puett, RC; Hart, JE; Schwartz, J; et al. (2011) Are particulate matter exposures associated with risk of type 2 diabetes? Environ Health Perspect 119:384–389.

Pun, VC; Kazemiparkouhi, F; Manjourides, J; Suh, HH. (2017) Long-term PM2.5 exposure and respiratory, cancer, and cardiovascular mortality in older US adults. Am J Epidemiol 186:961–969.

Renzi, M; Cerza, F; Gariazzo, C; et al. (2018) Air pollution and occurrence of type 2 diabetes in a large cohort study. Environ Int 112:68–76.

Samet, JM; Dominici, F; Curriero, FC; Coursac, I; Zeger, SL. (2000) Fine particulate air pollution and mortality in 20 U.S. cities, 1987–1994. N Engl J Med 343:1742–1749.

Stieb, DM; Chen, L; Eshoul, M; Judek, S. (2012) Ambient air pollution, birth weight and preterm birth: a systematic review and meta-analysis. Environ Res 117:100–111.

Sun, X; Luo, X; Zhao, C; et al. (2016) The associations between birth weight and exposure to fine particulate matter (PM2.5) and its chemical constituents during pregnancy: a meta-analysis. Environ Pollut 211:38–47.

Weinmayr, G; Hennig, F; Fuks, K; et al. (2015) Long-term exposure to fine particulate matter and incidence of type 2 diabetes mellitus in a cohort study: effects of total and traffic-specific air pollution. Environ Health 14:53.

This epidemiology summary is prepared by RoseyFields for internal and client reference and reflects the peer-reviewed human epidemiological literature on particulate matter as summarized above. Application to any specific individual, site, or matter requires additional case-specific exposure reconstruction and medical/differential-diagnosis review, and is documented separately.