Final Report: Spatial Investigation of Sources, Composition, and Long-Term Health Effects of Coarse Particulate Matter (PM10-2.5) in the Multi-Ethnic Study of Atherosclerosis (MESA) Cohort

EPA Grant Number: R833741
Title: Spatial Investigation of Sources, Composition, and Long-Term Health Effects of Coarse Particulate Matter (PM10-2.5) in the Multi-Ethnic Study of Atherosclerosis (MESA) Cohort
Investigators: Larson, Timothy V. , Adar, Sara D. , Barr, R. Graham , Burke, Gregory L. , Daviglus, Martha , Jacobs, David , Kaufman, Joel D. , Sheppard, Lianne (Elizabeth) A. , Simpson, Chris , Szpiro, Adam
Institution: University of Washington , Columbia University in the City of New York , Northwestern University , University of Minnesota , Wake Forest University
EPA Project Officer: Chung, Serena
Project Period: March 1, 2008 through February 28, 2013
Project Amount: $1,199,217
RFA: Sources, Composition, and Health Effects of Coarse Particulate Matter (2006) RFA Text |  Recipients Lists
Research Category: Air , Air Quality and Air Toxics , Particulate Matter

Objective:

Although there is ample evidence that fine airborne particles (median diameter ≤ 2.5 μm) are detrimental to human health, the importance of larger, coarse particles (2.5-10 μm) is less clear. Uncertainties are especially evident for chronic health effects, due to an incomplete understanding of spatial differences in coarse mass and composition. Therefore, the objectives of this project were to characterize the spatial variability of coarse particles from natural and anthropogenic sources and examine their associations with cardiovascular and respiratory disease.

Summary/Accomplishments (Outputs/Outcomes):

This project leveraged the large and well-characterized Multi-Ethnic Study of Atherosclerosis (MESA) cohort to provide important and novel data regarding AmericanÂ’s' long-term exposures to coarse particulate matter (PM) and related health implications. MESA is a prospective cohort study of 6,814 white, black, Hispanic, and Chinese participants without clinical cardiovascular disease residing in six American communities (Baltimore, MD; Chicago, IL; Forsyth County, NC; Los Angeles County, CA; Northern Manhattan, NY; and St. Paul, MN). This work both leverages and extends the EPA-funded MESA Air study, which is designed to assess long-term health effects of fine but not coarse particles. As part of this grant, we have:

  1. Conducted spatially rich air pollution monitoring, collecting 235 samples from 175 locations with varying local characteristics in 3 American metropolitan areas (Chicago, IL, St. Paul, MN, and Winston Salem, NC) which included urban and rural locations.
  2. Performed source-apportionment analysis to identify key chemical tracers of brake wear, tire wear, windblown soil, and fertilized soil across all sampled locations.
  3. Developed site-specific spatial prediction models of long-term concentrations of coarse mass, endotoxin, and chemical tracers of brake wear, tire wear, windblown soil, and fertilized soil using geographic predictors such as land use, vegetation, and roadway characteristics allowing us to generate individual-level estimates of exposure for approximately 3,000 participants in the MESA Coarse cities.
  4. Examined associations between individual-level estimates of coarse mass and chemical tracers of brake wear, tire wear, windblown soil, and fertilized soil with subclinical markers of respiratory and cardiovascular disease as well as clinical cardiovascular events. 

Collectively, the products of this grant advance our existing knowledge by providing a better understanding of the spatial patterns of coarse particles across metropolitan areas containing urban and rural environments as well as characterizing the associated human health effects of total mass and specific sources over long-term time scales. Overall, our epidemiological findings suggest relationships between long-term coarse particles and changes in lung structure and function over time, but not the progression of subclinical atherosclerosis or incident cardiovascular clinical events. Associations were observed to vary by source type and geographical location suggesting the importance of evaluating health with more detailed indicators of exposure than just coarse mass. A more detailed summary of our major contributions/ findings and their significance is presented below.

Created a resource of spatially rich, chemically speciated coarse particle samples in 3 cities

Between January and September 2009, we collected 235 two-week samples of collocated PM10 and PM2.5  measurements during summer and winter snapshot campaigns in each of the 3 MESA Coarse cities. Monitoring locations were selected to cover a range of geographic characteristics across urban and rural areas of three metropolitan areas. After rigorous quality control procedures, we had 207 (88%) and 195 (83%) valid PM2.5 and PM10 mass measurements resulting in 191 (81%) valid PM10-2.5 sample pairs from 121 unique locations. All samples were analyzed for mass, light absorbing carbon, elemental species by X-ray fluorescence (valid n=190), and endotoxin (valid n=169). Measured concentrations were generally precise (RPD=18%) and measurements collocated with regulatory monitors demonstrated consistency with long-term averages. With each seasonal snapshot campaign including 30-40 diverse locations measured simultaneously, these data have not only been instrumental to this grant but also have served and continue to serve as an excellent resource for other research projects. 

Documented substantial variability in concentrations within cities

PM10-2.5 mass concentrations were generally found to be consistent across city and season. Average concentrations (+SD) across both seasons were 5.7 ± 2.0, 5.3 ± 3.3 and 3.6 ± 1.4 µg m-3 in Chicago, St. Paul and Winston-Salem, respectively. Between-city variability was generally smaller than within-city variability and within-city variability was predominantly driven by differences by location rather than time as we saw only limited evidence of seasonal differences in St. Paul (averaging 3.3 and 6.7 µg m-3 in winter and summer, respectively) but not the other two cities. Chemical species also demonstrated good variability within city as well as between cities. For example, higher concentrations of chemicals known to be generated by anthropogenic activities such as copper and zinc were found in Chicago with moderate levels in St. Paul and lower levels in Winston-Salem. Interestingly, very low concentrations of endotoxin were documented during all but one sampling round in St. Paul. For all other time periods nearly 90% of all samples were less than 0.1 EU/m3, a level generally believed to be associated with no health effects.

Characterized key sources of coarse particles and identified indicator species

We utilized a modified positive matrix factorization approach to characterize source profiles for coarse particles. This approach was selected to allow for the use of universal prior source profile information in addition to localized profiles by region for different crustal compositions (soil). These models successfully identified break wear, tire wear, crustal, and fertilized soil features along with chemical tracers as indicators for each of these sources (copper, zinc, silicon, and phosphorus, respectively).  

Fine-scale spatial variability of coarse mass and components predicted in three American cities

In order to characterize long-term exposures to coarse PM, we created prediction models for long-term concentrations of coarse mass and selected chemical indicators of natural and anthropogenic sources in three American cities. These models were developed for PM10-2.5 mass, copper, zinc, phosphorus, silicon, and endotoxin using land use regression and universal kriging approaches. Geographic data including information on traffic, industrial emissions, impervious surfaces, land use, population density, and vegetation, were explored as potential predictors. This work represents a significant contribution of the MESA Coarse study as these are among the first models to predict spatially patterning of coarse mass and components, and the first in the United States to be based on intensive spatial sampling. Furthermore, by including data from urban and rural areas, we were able to characterize both anthropogenic and natural sources of coarse particles.

Overall, our site-specific models demonstrated good predictive performance for coarse mass, explaining between 40% and 70% of the variability under 10-fold cross-validation. These models also performed well for our four indicator species of interest with cross-validated R2s of 0.5-0.9 for copper, 0.4-0.7 for zinc, 0.5-0.8 for phosphorus and 0.5-0.9 for silicon. Though our models for endotoxin failed to perform very well with low CV R2(0.3-0.4) and high RMSEs, this poor performance is most likely due to the low levels measured in our metropolitan areas rather than a problem with the modeling approach.

Individual-level concentrations estimated for all participants in the three MESA Coarse cities

Using our spatial prediction models, we predicted long-term concentrations of coarse mass, copper, zinc, phosphorus, and silicon for participantsÂ’' residential addresses within the MESA Coarse regions. Although we initially proposed to include endotoxin, it ultimately was not included due to low levels and poor predictive performance of our spatial models. Clear differences in the spatial distributions of coarse mass and components were observed (see Figures 1-3), facilitating the differentiation of various sources in our health analyses.

Figure 1. Distribution of coarse particulate mass by location (µg/m3)


Figure 2. Distribution of copper (an indicator of break wear) in coarse particles by location (ng/m3) 

Figure 3. Distribution of phosphorus (an indicator of fertilized soil) in coarse particles by location (ng/m3)

Long-term exposures to coarse particulate matter suggestively associated with the progression of subclinical respiratory endpoints

As originally proposed, we examined relationships between long-term exposures to coarse particles and components with the progression of sub-clinical chronic obstructive pulmonary disease using repeated measurements of lung density. Total percent emphysema and the ratio of percent emphysema in the upper and lower lobes of the lung were measured based on regions < -910 Hounsfield Units via cardiac computed tomography (CT) at each examination. In addition, we expanded our original scope to include analyses of lung function measures (i.e., expiratory volume in 1 second [FEV1] and forced vital capacity [FVC]) collected by spirometry in a subset of the cohort. For CT endpoints, changes were examined over a 3.5-year average follow-up period, whereas spirometry was for 5-year average follow-up. For all endpoints, exposures were estimated based on the MESA Coarse predictions in St. Paul, Winston Salem, and Chicago as well as nearest monitor estimates from local AQS stations in all six MESA cities.

Among the participants in the MESA coarse cities, we found no evidence of associations between PM10-2.5 mass or components with change in total percent emphysema over time after controlling for individual risk factors and metropolitan area. A 0.12 point/yr increase in the change of the ratio of the upper to lower lobe ratio of percent emphysema (95% CI: 0.003 to 0.23 per yr) was found per 4 ng/m3 of copper, however. This suggests a shift towards centrilobular emphysema, the most common presentation of emphysema among smokers. This associations was robust to control for PM2.5 and oxides of nitrogen (NOx), though was weakened by control for light absorbing carbon (LAC). No associations were observed with the other pollutants nor were nearest monitor levels associated with either emphysema endpoint among the complete MESA cohort.

Among the smaller subset of 1,624 individuals with spirometry and MESA coarse estimates, higher copper concentrations also were associated with a faster rate of decline in FEV1 and FVC over time. A 4 ng/m3 higher level of copper was associated with 10.4 mL/yr (95% CI: -3.8 to 17.0 mL/yr) and 9.7 mL/yr (95% CI: -0.9 to -18.5 mL/yr) steeper reductions in FEVand FVC over time, respectively, after control for individual risk factors, study site, and PM10-2.5 mass concentrations. These associations were linear (Figure 4) and robust to control for PM2.5, NOx, and LAC. No associations were found for any other pollutants or using coarse mass concentrations from the nearest AQS monitor.

Text Box: Change in FVC 
Over Time (mL/y)
 Text Box: Change in FEV1 
Over Time (mL/y)

Figure 4. Higher copper concentrations associated with faster declines in lung function over time, adjusted for risk factors, study site, PM10-2.5 mass, PM2.5, and indicators of fresh traffic combustion

Note: Models controlled for age, gender, race/ethnicity, socio-economic status, body size, hay fever, workplace exposures, active and passive smoking, site, and PM10-2.5 mass, PM2.5, NOx, NO2, and LAC.

Overall, our findings suggest relationships between long-term coarse copper concentrations and changes in lung structure and function over time. The magnitude of the association observed for lung function was similar to what would be observed with a 2 to 3 year increase in age. Shifts in the distribution of emphysema in the lung associated with copper concentrations also were similar to the presentation most commonly found among smokers. As associations were limited to coarse copper concentrations, independent of PM2.5 and some indicators of traffic combustion emissions, we hypothesize that brake wear or other sources of copper may have important respiratory impacts.

No clear associations observed for long-term exposures to coarse particulate matter with subclinical atherosclerosis or clinical cardiovascular events

As proposed in the grant, associations between long-term exposures to coarse particles and the progression of atherosclerosis were explored using measures of intima-medial thickness (IMT) of the common carotid artery via ultrasound and coronary artery calcium (CAC) by cardiac CT. Extending our original proposal, we also examined changes in ankle brachial index (ABI).  Follow-up periods for these outcomes ranged from 2 to 7 years. Incident cardiovascular clinical events were also examined as originally proposed.

Overall, we did not find consistent or statistically significant evidence of an association between coarse particles and the various measures of atherosclerosis using MESA coarse exposure estimates nor were associations identified among participants in all cities using nearest AQS monitor data. Similarly, our estimated hazard ratios did not support positive associations with MESA coarse exposure estimates and no associations were observed within the whole MESA cohort based on nearest AQS monitor data. 

Long-term exposures to coarse particulate matter suggestive associations with other subclinical markers of disease

In addition to our primary analyses outlined in the grant proposal, we also explored associations between long-term exposures to coarse particles and their components with numerous additional subclinical endpoints (i.e., ABI, pulmonary function, inflammation, coagulation, retinal diameters, right and left ventricular mass and function, incident hypertension, and incident respiratory clinical events) to create a more comprehensive mechanistic picture. While the observed associations varied by endpoint and species, we found consistently positive associations with a summary inflammation score based on interleukin-6, C-reactive protein, fibrinogen, and total homocysteine in a cross-sectional analysis of baseline blood samples and MESA Coarse estimates. These associations were positive for mass and all chemical components, reaching statistical significance for zinc. Positive associations were similarly found with a summary coagulation score based on factor VIIIc, d-dimer, and plasmin-antiplasmin complex for coarse mass, copper, and silicon. No associations were found for zinc and phosphorus. Suggestive positive associations also were identified in a cross-sectional analysis with right ventricular end-diastolic mass, a possible indicator for pulmonary hypertension, for all exposures except zinc among a subset of participants (n=1,763) with cardiac magnetic resonance imaging data and MESA Coarse estimates.

Suggestive evidence of different susceptibility with age and metropolitan area

For all health analyses, we explored susceptibility by age, gender, race/ethnicity, obesity, study site, and an indicator for urban vs. rural location. In general, we did not observe strong evidence of effect modification for most analyses though some evidence of effect modification was found in different analyses. The most consistent characteristic associated with enhanced risk appeared to be advanced age with evidence from analyses with both respiratory and cardiovascular endpoints. While a crude indicator for urban/rural locations did not seem to modify associations, there was evidence of larger associations in Winston Salem as compared to St. Paul and Chicago in several analyses. This perhaps suggests differences in composition by region or different characteristics of the study population.

Conclusions:

The MESA Coarse study has helped to further existing knowledge in two critical ways. First, this work substantially adds to our understanding of the spatial patterns of coarse particles from natural and anthropogenic sources in three American metropolitan areas. The generation of spatially intensive monitoring data and spatial prediction models for total coarse particulate mass and indicators of several sources represents an important contribution by creating a resource for future studies and allowing for individual-level estimates of exposure in MESA. Another key contribution is the investigation of numerous subclinical and clinical health endpoints with long-term exposures to coarse particulate matter and its components. Ultimately, our findings indicate the presence of stronger associations between coarse particles and respiratory, rather than cardiovascular, endpoints and suggest that associations can vary by source type and geographical location. After the completion of this grant, we remain committed to actively disseminating our research at international conferences and in peer-reviewed journals.


Journal Articles on this Report : 6 Displayed | Download in RIS Format

Other project views: All 16 publications 7 publications in selected types All 6 journal articles
Type Citation Project Document Sources
Journal Article Adar SD, Kaufman JD, Diez-Roux AV, Hoffman EA, D'Souza J, Stukovsky KH, Rich SS, Rotter JI, Guo X, Raffel LJ, Sampson PD, Oron AP, Raghunathan T, Barr RG. Air pollution and percent emphysema identified by computed tomography in the Multi-Ethnic Study of Atherosclerosis. Environmental Health Perspectives 2015;123(2):144-151. R833741 (Final)
R831697 (Final)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: EHP-Full Text PDF
  • Abstract: EHP-Abstract & Full Text HTML
  • Journal Article Adar SD, D'Souza J, Mendelsohn-Victor K, Jacobs DR, Cushman M, Sheppard L, Thorne PS, Burke GL, Daviglus ML, Szpiro AA, Diez Roux AV, Kaufman JD, Larson TV. Markers of inflammation and coagulation after long-term exposure to coarse particulate matter: a cross-sectional analysis from the Multi-Ethnic Study of Atherosclerosis. Environmental Health Perspectives 2015;123(6):541-548. R833741 (Final)
    R831697 (Final)
    R834796 (2014)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: EHP-Full Text PDF
  • Abstract: EHP-Abstract & Full Text HTML
  • Journal Article Sturtz TM, Adar SD, Gould T, Larson TV. Constrained source apportionment of coarse particulate matter and selected trace elements in three cities from the Multi-Ethnic Study of Atherosclerosis. Atmospheric Environment 2014;84:65-77. R833741 (Final)
    R831697 (2013)
    R831697 (Final)
  • Full-text: ScienceDirect-Full Text HTML
    Exit
  • Abstract: ScienceDirect-Abstract
    Exit
  • Other: ScienceDirect-Full Text PDF
    Exit
  • Journal Article Wang M, Keller JP, Adar SD, Kim S-Y, Larson TV, Olives C, Sampson PD, Sheppard L, Szpiro AA, Vedal S, Kaufman JD. Development of long-term spatiotemporal models for ambient ozone in six metropolitan regions of the United States: the MESA Air Study. Atmospheric Environment 2015;123(A):79-87. R833741 (Final)
    R831697 (Final)
    R834796 (2016)
    R834796 (Final)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: ScienceDirect-Full Text HTML
    Exit
  • Abstract: ScienceDirect-Abstract
    Exit
  • Other: ScienceDirect-Full Text PDF
    Exit
  • Journal Article Wang M, Sampson PD, Hu J, Kleeman M, Keller JP, Olives C, Szpiro AA, Vedal S, Kaufman JD. Combining land-use regression and chemical transport modeling in a spatiotemporal geostatistical model for ozone and PM2.5. Environmental Science & Technology 2016;50(10):5111-5118. R833741 (Final)
    R831697 (Final)
    R833864 (Final)
    R834796 (2016)
    R834796 (Final)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: ACS-Full Text HTML
    Exit
  • Abstract: ACS-Abstract
    Exit
  • Other: ACS-Full Text PDF
    Exit
  • Journal Article Zhang K, Larson TV, Gassett A, Szpiro AA, Daviglus M, Burke GL, Kaufman JD, Adar SD. Characterizing spatial patterns of airborne coarse particulate (PM10-2.5) mass and chemical components in three cities: the Multi-Ethnic Study of Atherosclerosis. Environmental Health Perspectives 2014;122(8):823-830. R833741 (Final)
    R831697 (Final)
  • Full-text from PubMed
  • Abstract from PubMed
  • Associated PubMed link
  • Full-text: EHP-Full Text PDF
  • Abstract: EHP-Abstract & Full Text HTML
  • Supplemental Keywords:

    RFA, Scientific Discipline, Air, particulate matter, Health Risk Assessment, Biology, atmospheric particulate matter, sensitive populations, atmospheric particles, cardiopulmonary responses, human health effects, bioavailability, cardiovascular vulnerability, sensitive subgroups, cardiotoxicity, exposure assessment

    Progress and Final Reports:

    Original Abstract
  • 2008 Progress Report
  • 2009 Progress Report
  • 2010 Progress Report
  • 2011 Progress Report