Development and performance evaluation of a GIS-based metric to assess exposure to airborne pollutant emissions from industrial sources

Abstract

Introduction Few studies have investigated the effect of airborne exposure to dioxins and cadmium (Group 1 carcinogens by the International Agency for Research on Cancer) on breast cancer risk and overall results are inconclusive. Inconsistency across results may be explained by methodological limitations, including lack of historical measurements and residential history information. The multiplicity of exposure sources and the latency between exposure and cancer occurrence represent major challenges and require to precisely characterize the spatial-temporal variability of exposures over large areas and long time-periods. To overcome these limitations, this study aimed to develop and assess the performance of an exposure metric based on a Geographic Information System (GIS) through comparison with a validated dispersion model to estimate historical (1990–2008) industrial dioxin and cadmium exposures. Methods We carried out a detailed retrospective inventory of dioxin and cadmium emitting sources from 1990 to 2008, through national databases and contact with facility operators. A careful collection of technical characteristics of sources and activity rate allowed us to estimate annual dioxin and cadmium emissions of each source using default emissions factors from the literature. The location of each facility was precisely geocoded and together with the emission estimates used as input data for the GIS-based metric. Based on the review of the literature, we identified relevant parameters to be included into the GIS-based metric: emissions’ intensity and location, subject's residence-to-source distance, wind direction and speed, exhaust smoke velocity and stack height. To identify the most relevant combination of parameters, we compared agreement of categorical dioxin exposure classification of study subjects, between the GIS-based metric and a referent dispersion model (SIRANE) in three selected areas (rural, urban and urban-costal) and for three distinct years (9 scenarios). The agreement was estimated by calculating weighted kappa statistics (wκ) and coefficients of determination (R2). During the calibration phase, we identified the combination of parameters that provided the best agreement with dispersion model results across the nine scenarios. The performance of the final GIS-based metric equation was tested for a new set of subjects’ localisation (n = 450) and for the estimation of cadmium exposure. Results Between 1990 and 2008, we inventoried and estimated emissions of respectively 2620 and 2700 sources of dioxins and cadmium, respectively. Over this period, 82% of sources were positioned at the stack, 13% at the centroid of the building and 5% at the parcel. The agreement between the GIS-based metric and the dispersion model for dioxin exposure varied from “substantial” to “almost perfect”: median wκ=0.78 (1st quintile = 0.72, 3rd quintile = 0.82) and median R2 = 0.82 (1st quintile = 0.71, 3rd quintile = 0.87). We observed similar performances for cadmium. The final metric combined residential distance to facilities, wind direction and proportion of the year blown and technical parameters of the facilities. Weighted kappa were systematically below 0.55 when no meteorological parameters were integrated into the GIS-based metric. Conclusions We developed and evaluated a GIS-based metric in order to estimate the retrospective airborne dioxin exposure of participants of a cohort-nested case-control study. This combination of parameters showed reliable estimates in comparison to an atmospheric dispersion model across different scenarios. This metric was used to estimate historical dioxin exposure in an epidemiological study on breast cancer risk. The GIS-based metric also provided reliable estimates for cadmium exposure from industrial sources and may be able to assess exposure to other air pollutants with similar properties and behavior than dioxins and cadmium (i.e. heavy metals, PM10, etc.), in particular when monitoring data are lacking.