Particulate Matter Source Types Associated with Cardiovascular Hospitalizations and Mortality

KAZUHIKO ITO (1), Robert Mathes (2), Zev Ross (3), George Thurston (1), and Thomas Matte (2)

(1) New York University School of Medicine, Tuxedo, NY; (2) New York City Department of Health and Mental Hygiene, New York, NY; (3) ZevRoss Spatial Analysis, Ithaca, NY.

     Abstract Number: 403
     Last modified: November 9, 2009

Abstract
Background:
Recent multi-city time-series studies have indicated that both mortality and hospitalizations were associated with particulate matter (PM). While the PM associations were reported to be strongest in the Northeastern cities for both mortality and hospitalizations, the associations for mortality were different: stronger in summer for non-accidental mortality and in winter for the elderly hospitalizations for cardiovascular diseases (CVD) and respiratory illness. The interpretation of the apparent difference in the seasonal pattern of associations between the two outcomes is complicated by the fact that these outcomes were not evaluated for consistent disease categories or age groups. We investigated this issue in New York City (NYC), where PM originates from both regional and local combustion sources.
Methods:
Daily counts of deaths and emergency hospitalizations available at the NYC Department of Health and Mental Hygiene were aggregated for cardiovascular diseases for age 39 years and older during the years 2000-2006. Data for PM10, PM2.5 and its chemical components, nitrogen dioxide (NO$_2), ozone (O$_3), carbon monoxide (CO), and sulfur dioxide (SO$_2) were retrieved from US Environmental Protection Agency. To examine the relative importance of PM size, we also estimated thoracic coarse particles (PM$_(10-2.5)) mass concentrations by subtracting the daily city-average PM$_(2.5) from city-average PM$_(10) (for days when the former was equal to or smaller than the latter), for years 2000-2003 on days when both PM$_(10) and PM$_(2.5) data were available. We estimated excess risks for these multiple pollutants using Poisson time-series model adjusting for temporal trends/seasonal cycles, immediate and delayed temperature effects, and day-of-week. Risks were estimated for inter-quartile-range increases at lag 0 through 3 days, and for warm (April through September) and cold (October through March) seasons.
Results
The CVD mortality series exhibit strong seasonal trends, while the CVD hospitalization series show a stronger influence of day-of-week pattern. In the analysis of PM$_(2.5) and gaseous pollutants for years 2000-2006, PM$_(2.5) was most strongly associated with CVD mortality, particularly in the warm season, whereas for CVD hospitalizations, NO$_2 was most strongly associated in both seasons. In the analysis of size-fractionated PM, PM$_(10), PM$_(2.5), and computed PM$_(10-2.5), all showed significant positive associations with CVD mortality, especially in the warm season, in the range of 1 to 2 percent excess risk per 10 micro-gram/m$^3 increase, while the associations with CVD hospitalizations were stronger in the cold season. Among the PM$_(2.5) chemical components, the species originated from both regional secondary aerosols (e.g., sulfate) and local combustion (e.g., elemental carbon) were associated with CVD mortality in the warm season, whereas for CVD hospitalizations, those originated from local sources (e.g., elemental carbon, Zn, Si) showed associations in the cold season.
Discussion
The associations between PM and CVD mortality and hospitalizations showed different seasonal patterns and source types. Apparent modification by season and source type of the relations of PM to CVD mortality and hospitalizations merits further study to examine potential differences in at-risk subpopulations, relationship of ambient measurements to population exposures, and interactions of PM with seasonal factors such as weather and communicable disease.