The National Environmental Respiratory Center Program was initiated as an experiment to explore strategies for identifying the components of complex air pollution mixtures that cause health effects associated statistically with air pollution. A strategy involving multivariate analysis of a composition-concentration-response database was adopted. A novel database was created by exposing rodents daily for up to six months to one of four combustion-related mixtures and measuring respiratory, cardiovascular and general toxicological responses after one week or six months of exposure. The mixtures included multiple concentrations of diesel and gasoline engine exhaust, hardwood smoke and simulated downwind coal combustion emissions. After reporting the biological effects of each mixture and comparing effects among them, 47 significant effects were selected for multiple additive regression tree analysis to identify putative causal components. Although the four mixtures provided a database marginally sufficient for the analysis, the results suggested the putative causes of 19 significant effects with acceptable confidence. This article describes and critiques the Program and its strategy. The integrated results are presented in two accompanying papers, and mixture-specific results were presented in preceding papers, which are cited. The experiment demonstrated the potential utility of the general approach and identified certain cause-effect relationships for confirmatory studies. A follow-up study provided support for causation by the components implicated for one of those relationships. The advantages and disadvantages of the Program's management and funding strategies are discussed.
Keywords: Air pollution; coal emissions; diesel exhaust; gasoline exhaust; mixtures; multipollutant; regression tree; wood smoke.