Air Quality and Health and Welfare

2.3 Particulate Matter

2.3.2 The Particle Epidemiology Reanalysis Project
The Harvard Six City Study and the American Cancer Society study examined the long-term effects of exposure to particulate air pollution on mortality. The Harvard Six-Cities Study by Dr. Douglas Dockery of the Harvard School of Public Health, and others, was published in December 1993 in the New England Journal of Medicine. Researchers followed the health of more than 8,000 people in six small cities that fell along a gradient of air pollution concentrations for a period of 14 to 16 years. As particle concentrations increased, there was an almost directly proportional increase in the death rate in the residents studied. Residents of the most polluted city in the study, Steubenville, Ohio, had a 26 percent increased risk of premature mortality, compared to the residents of the cleanest city studied, Portage, Wisconsin. According to study authors, this translates into a shortened life expectancy of one to two years for residents of Steubenville compared to residents of Portage.

The March 1995 American Cancer Society study, by Dr. Arden Pope of Brigham Young University, and others, found an association between chronic exposure to fine particle air pollution and premature death in a study group of over half a million people in 151 cities. Sulfate pollution was also associated with early death. The study reported strong associations between sulfates and fine particles and death by cardio-pulmonary causes.

These original studies used statistical techniques to adjust for age, and to control for the effects of smoking, diet, and occupational exposure.

Dr. Daniel Krewski of the University of Ottawa and his associates conducted the newly released reanalysis of these two studies for the Health Effects Institute. First, the HEI - funded researchers undertook a reanalysis of the original studies and a quality audit of the underlying data. Second, researchers performed an extensive sensitivity analysis using alternative statistical methods, and considering the role of 20 potential confounders such as other pollutants, climate, and socio-economic factors on study results.

The reanalysis by independent investigators validates the original studies.