Coarse and fine particles and daily mortality in the Coachella Valley, California: a follow-up study BART D. OSTRO, a RACHEL BROADWIN a AND MICHAEL J. LIPSETT a,b a California Office of Environmental Health Hazard Assessment, California b University of California, San Francisco, California Many epidemiological studies provide evidence of an association between ambient particles, measured as PM10, and daily mortality. Most of these studies have been conducted in urban areas where PM10 is highly correlated with and dominated by fine particles less than 2.5 m in diameter (PM2.5). Fewer studies have investigated impacts associated with the fraction of coarse mode particles (between 2.5 and 10 m in diameter). In a previous study using data from 1989 through 1992 in the Coachella Valley, a desert resort and retirement area east of Los Angeles, we reported associations between PM10 and several different measures of mortality [Ostro B.D., Hurley S., and Lipsett M.J. Air pollution and daily mortality in the Coachella Valley, California: a study of PM10 dominated by coarse particles. Environ. Res. 1999: 81: 231±238]. In this arid environment, coarse particles of geologic origin are highly correlated with and comprise approximately 60% of PM10, increasing to >90% during wind events. This study was intended to repeat the earlier investigation using 10 years (1989±1998) of daily data on mortality and PM10. The last 2.5 years of data also included daily measures of PM2.5, allowing examination of size-specific impacts. To ensure adequate statistical power, we attempted to develop predictive models for both fine and coarse particles to use in analyses of the full 10-year period. An acceptable fit was found only for coarse particles, which were found to be a cubic function of PM10 ( R 2 =0.95). Outcome variables included several measures of daily mortality, including all-cause (minus accidents and homicides), cardiovascular and respiratory mortality. Multivariate Poisson regression analyses using generalized additive models were employed to explain the variation in these endpoints, controlling for temperature, humidity, day of the week, season, and time, using locally weighted smoothing techniques. Pollution lags of up to 4 days were examined. Several pollutants were associated with all-cause mortality, including PM2.5, carbon monoxide and nitrogen dioxide. More consistent results were found for cardiovascular-specific mortality, for which associations were found for coarse particles (RR=1.02; 95% C.I., 1.01±1.04), PM10 (RR=1.03; 95% C.I., 1.01±1.05). None of the pollutants was associated with respiratory-specific mortality. Ozone was not associated with any of the mortality outcomes. These findings are generally consistent with those we previously reported for the Coachella Valley for the period 1989±1992, demonstrating associations between several measures of particulate matter and daily mortality in an environment in which particulate concentrations are dominated by the coarse fraction. Journal of Exposure Analysis and Environmental Epidemiology (2000) 10, 412± 419. Keywords: air pollution, coarse particles, epidemiology, mortality, particulate matter, time-series. Introduction In most urban areas of the United States, ambient PM10 (particles less than 10 m in diameter) is dominated by the fine fraction less than 2.5 m in diameter (PM2.5 or fine particles). PM10 has been repeatedly associated with a wide range of adverse health outcomes, including mortality, and it has been proposed that PM10-associated toxicity is attributable principally to the fine fraction (U.S. EPA, 1996). However, the potential impacts of the coarse fraction (PM10 minus PM2.5) have been less well documented. Specifically, there have been few studies examining whether exposures to ambient particulate matter originating primarily from geologic sources (e.g., soil and sand) rather than combustion are associated with adverse health outcomes. Coachella Valley is roughly 100 miles east of Los Angeles and includes the cities of Palm Springs at the northwestern end of the populated corridor in the Valley and Indio towards the southern end. The Valley is a desert area with hot summers and mild winters, and is bordered by mountains on the north, east and west and by the Salton Sea on the south. Geologic particles comprise a significant percentage of the annual average particulate mass (approxi- mately 60% of PM10), increasing to >90% during wind episodes that occur throughout the year. In a previous study using data from 1989 through 1992, we reported associations between daily concentrations of PM10 and several different measures of mortality in the Valley (Ostro et al., 1999). The current study was intended to repeat the earlier investigation using 10 years (1989± 1998) of daily data on mortality and PM10. In addition, 1. Address all correspondence to: Dr. Bart D. Ostro, PhD, Chief, Air Pollution Epidemiology Unit, California Office of Environmental Health Hazard Assessment, 1515 Clay St., 16th Floor, Oakland, CA 94612. Tel.: +1-510-622-3157. Fax: +1-510-622-3210. E-mail: bostro@oehha.ca.gov Received 4 February 2000; accepted 17 May 2000. Journal of Exposure Analysis and Environmental Epidemiology (2000) 10, 412±419 # 2000 Nature America, Inc. All rights reserved 1053-4245/00/$15.00 www.nature.com/jea