Validation and calibration of a model used to reconstruct historical exposure to polycyclic aromatic hydrocarbons for use in epidemiologic studies

Objectives: We previously developed a historical reconstruction model to estimate exposure to airborne polycyclic aromatic hydrocarbons (PAHs) from traffic back to 1960 for use in case-control studies of breast cancer risk. Here we report the results of four exercises to validate and calibrate the model. Methods: Model predictions of benzo[a]pyrene (BaP) concentration in soil and carpet dust were tested against measurements collected at subjects' homes at interview. In addition, predictions of air intake of BaP were compared with blood PAH-DNA adducts. These same soil, carpet, and blood measurements were used for model optimization. In a separate test of the meteorological dispersion part of the model, predictions of hourly concentrations of carbon monoxide from traffic were compared with data collected at a U.S. Environmental Protection Agency monitoring station. Results: The data for soil, PAH-DNA adducts, and carbon monoxide concentrations were all consistent with model predictions. The carpet dust data were inconsistent, suggesting possible spatial confounding with PAH-containing contamination tracked in from outdoors or unmodeled cooking sources. BaP was found proportional to other PAHs in our soil and dust data, making it reasonable to use BaP historical data as a surrogate for other PAHs. Road intersections contributed 40-80% of both total emissions and average exposures, suggesting that the repertoire of simple markers of exposure, such as traffic counts and/or distance to nearest road, needs to be expanded to include distance to nearest intersection.