A statistical study of the macroepidemiology of air pollution and total mortality

Abstract

A statistical analysis of spatial patterns of 1980 US urban total mortality (all causes) was performed, evaluating demographic, socioeconomic and air pollution factors as predictors. Specific mortality predictors included cigarette smoking, drinking water hardness, heating fuel use, and 1978-1982 annual concentrations of the following air pollutants: ozone, carbon monoxide, sulfate aerosol, particulate concentrations of lead, iron, cadmium, manganese, vanadium, as well as total and fine particle mass concentrations from the inhalable particulate network (dichotomous samplers). In addition, estimates of sulfur dioxide, oxides of nitrogen, and sulfate aerosol were made for each city using the ASTRAP long-range transport diffusion model, and entered into the analysis as independent variables. Because the number of cities with valid air quality and water hardness data varied considerably by pollutant, it was necessary to consider several different data sets, ranging from 48 to 952 cities. The relatively strong associations (ca. 5--10%) shown for 1980 pollution with 1980 total mortality are generally not confirmed by independent studies, for example, in Europe. In addition, the US studies did not find those pollutants with known adverse health effects at the concentrations in question (such as ozone or CO) to be associated with mortality. The question of causality vs. circumstantial association must therefore be regarded as still unresolved. 59 refs., 20 figs., 40 tabs.

Keywords

Affiliated Institutions

• Brookhaven National Laboratory US
• Stony Brook University US

Related Publications