Grantee Research Project Results
2011 Progress 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: R833741Title: 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. , Kaufman, Joel D. , Sheppard, Lianne (Elizabeth) A. , Simpson, Chris , Barr, R. Graham , Burke, Gregory L. , Jacobs, David , Daviglus, Martha , Adar, Sara D.
Current Investigators: Larson, Timothy V. , Sheppard, Lianne (Elizabeth) A. , Kaufman, Joel D. , Adar, Sara D. , Barr, R. Graham , Burke, Gregory L. , Szpiro, Adam , Simpson, Chris , Jacobs, David , Daviglus, Martha
Institution: University of Washington , Columbia University in the City of New York , Northwestern University , Wake Forest University , University of Minnesota
Current 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 Period Covered by this Report: March 1, 2011 through February 29,2012
Project Amount: $1,199,217
RFA: Sources, Composition, and Health Effects of Coarse Particulate Matter (2006) RFA Text | Recipients Lists
Research Category: Particulate Matter , Air Quality and Air Toxics , Air
Objective:
The primary objectives of this project are to: 1) characterize the within-city spatial variability of PM10-2.5 from natural and anthropogenic sources using a spatially intensive approach; and 2) examine associations of PM10-2.5 with cardiovascular and respiratory disease. We hypothesize that geographically varying parameters such as proximity to roadway, land use, and vegetative indices will be predictive of total PM10-2.5 mass and source-specific component concentrations measured at different locations. We further hypothesize that increased long-term exposures to PM10-2.5, especially to traffic-associated PM10-2.5, are associated with increased clinical cardiovascular events, as well as a more rapid sub-clinical progression of atherosclerosis and emphysema.
Progress Summary:
Over the past reporting period, we have been productive and are making good progress towards completing our two proposed aims. For Specific Aim 1, we have accomplished the following:
· Performed positive matrix factorization with and without constraints to characterize PM10-2.5 source profiles in each of the three MESA Coarse cities.
· Identified individual chemicals to serve as indicator species for traffic break wear, traffic tire wear, agriculture, and soil sources of PM10-2.5 and validated these choices against source-specific geographic features.
· Expanded geographic covariate list for spatial prediction modeling to include truck routes, Caline traffic model predictions, and new land use data.
· Developed and implemented systematic procedures and programs to conduct spatial prediction modeling for PM10-2.5 mass and components in all cities.
· Submitted abstracts on our selection procedures for indicator components and spatial modeling of components to an international conference.
· Prepared written manuscripts of the PM10-2.5 mass and components prediction models to be submitted for scientific publication.
With the completion of our spatial prediction models under Specific Aim 1, we have begun to work intensively on Specific Aim 2 and have successfully completed the following tasks during this past reporting period:
· Generated estimates of long-term PM10-2.5 concentrations for the complete MESA Coarse cohort weighted by residential history.
· Examined associations between PM10-2.5 mass concentrations and markers of systemic inflammation.
· Investigated the relation between PM10-2.5 mass concentrations and indicators of right ventricular mass and function.
· Prepared written manuscripts on associations with systemic inflammation and right ventricular mass and function to be submitted for scientific publication.
Preliminary Results:
Between September 2008 and September 2009, we collected nearly 200 valid PM10-2.5 samples from approximately 180 unique locations. Average PM10-2.5 concentrations (+SD) across seasons were 5.7 ± 2.0, 5.1 ± 3.8 and 3.7 ± 1.5 µg m-3 in Chicago, St. Paul and Winston-Salem, respectively. As shown in Figure 1, between-city variability was generally smaller than within-city variability. Within-city variability predominantly represented spatial contrasts as we only saw evidence of seasonal differences in St Paul (averaging 3.2 and 6.7 µg m-3 in winter and summer, respectively) but not the other two cities.
Figure 1: Measured concentrations of PM10-2.5 by city and season.
Modeling of PM10-2.5 mass concentrations was conducted using land use regression and universal kriging approaches. Geographic information including data on traffic, industrial emissions, impervious surfaces, land use, population density, and vegetation were explored in our initial models. During the past year we added new variables including truck routes, updated land use information, and Caline traffic model estimates, which improved our model performance. Our city- and season-specific models explain between 50 and 90% of the variability in PM10-2.5 mass under 10-fold cross-validation. For all cities, season-specific models performed better than pooled models, which combined data from the two seasons. Interestingly, there was no evidence of residual spatial correlation in any of our models after control for geographic covariates (Table 1).
Table 1. PM10-2.5 mass models predict concentrations with little evidence of residual spatial correlation
| City | Season | LUR | UK | ||
|---|---|---|---|---|---|
| R2 | RMSE | R2 | RMSE | ||
| Chicago, IL | Winter | 0.88 | 0.67 | 0.88 | 0.67 |
| Summer | 0.68 | 1.19 | 0.68 | 1.19 | |
| Pool | 0.68 | 1.14 | 0.65 | 1.20 | |
| St. Paul, MN | Winter | 0.66 | 2.03 | 0.65 | 2.05 |
| Summer | 0.64 | 1.99 | 0.63 | 2.02 | |
| Pool | 0.57 | 2.48 | 0.58 | 2.46 | |
| Winston-Salem, NC | Winter | 0.48 | 0.88 | 0.48 | 0.97 |
| Summer | 0.76 | 0.85 | 0.76 | 0.85 | |
| Pool | 0.32 | 1.21 | 0.33 | 1.20 | |
Concentrations of PM10-2.5 species also varied with city, season and species type. As with PM10-2.5, within-city variations were larger than between-city variations for most species. Examining these data using constrained and unconstrained positive matrix factorization, seven unique profiles were indentified (break wear, tire wear, agricultural activities, and dust specific to Chicago, Winston Salem in the winter, and St Paul in the summer). Cu, Zn, P, and Si were identified as indicators of brake wear, tire wear, agricultural fertilizer, and general resuspended dust, respectively, and were compared to geographic features relevant to each source such as traffic density and spatially resolved agricultural tillage and fertilizer application data.
Spatial prediction models also performed well for these species with CV R2 of 0.5-0.9 for copper, 0.4-0.9 for zinc, 0.5-0.6 for phosphorus and 0.5-0.8 for silicon. Again, there was no evidence of residual spatial structure after control for geographic predictors. We also noted clear differences in the spatial distribution of PM10-2.5 mass and components, which should allow for separation in our health analyses.
Based on the above exposure assessment, we have conducted three epidemiological analyses to examine associations between PM10-2.5 mass and 1) inflammatory markers in the blood, 2) retinal arteriolar diameters, and 3) right ventricular mass and function. In these preliminary models, we found evidence that higher long-term exposure to PM10-2.5 mass is associated with higher levels of C-reactive protein (6.2% per 6.2 mg/m3, 95% CI: 0.1 to 12.7%) and fibrinogen (1.1% per 6.2 mg/m3, 95% CI: 0.02 to 2.3%) after control for risk factors, study site, and PM2.5 concentrations. Suggestive but non-significant associations were found for retinal arteriolar diameters and right ventricular mass.
Future Activities:
The next steps for this project are to publish our exposure findings and complete our epidemiological analyses. Draft manuscripts on the spatial modeling of PM10-2.5 mass and key components have been written and are currently under review with co-authors. A third paper on our PMF models is under preparation. We are concurrently focusing our main efforts on completing the planned epidemiologic analyses of the relations between PM10-2.5 mass and key components with the progression of atherosclerosis, lung density (a measure of emphysema), and clinical events. We will also continue to explore with secondary endpoints and publish these findings. This will include inflammation in the blood and right ventricular mass/function for which written manuscripts are ready to be circulated co-authors.
Journal Articles:
No journal articles submitted with this report: View all 19 publications for this projectSupplemental Keywords:
RFA, Air, Scientific Discipline, Health Risk Assessment, particulate matter, Biology, bioavailability, sensitive populations, sensitive subgroups, atmospheric particles, cardiopulmonary responses, cardiovascular vulnerability, cardiotoxicity, atmospheric particulate matter, chemical speciation sampling, human health effectsProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.