Exposure apportionment: ranking food items by their contribution to dietary exposure.

This paper identifies and ranks food items by estimating their contribution to the dietary exposure of the US population and 19 subpopulation groups. Contributions to dietary exposures to arsenic, cadmium, chromium, lead, nickel, benzene, chlorpyrifos, and diazinon are estimated using either the Dietary Exposure Potential Model (DEPM) approach, the National Human Exposure Assessment Survey Arizona (NHEXAS-AZ) approach or the combination of the two. The DEPM is a computer model that uses several national databases of food consumption and residue concentrations for estimating dietary. The DEPM approach ranks the contribution of food items to the total dietary exposure using two methods, the direct method that ranks contributions by population exposure magnitude and the weighted method that ranks by subpopulation exposure magnitude. The DEPM approach identifies highly exposed subpopulations and a relatively small number of food items contributing the most to dietary exposure. The NHEXAS-AZ approach uses the NHEXAS-AZ database containing food consumption data for each subject and chemical residues of a composite of food items consumed by each subject in 1 day during the sampling week. These data are then modeled to obtain estimates of dietary exposure to chemical residues. The third approach uses the NHEXAS-AZ consumption data with residue values from the national residue database. This approach also estimates percent contributions to exposure of each ranked food item for the Arizona population. Dietary exposures estimated using the three approaches are compared. The DEPM results indicate groups with highest dietary exposures include Nonnursing Infants, Children 1-6, Hispanic, Non-Hispanic White, Western, Northeast and Poverty 0-130%. The use of the Combined National Residue Database (CNRD) identifies 43 food items as primary contributors to total dietary exposure; they contribute a minimum of 68% of the total dietary exposure to each of the eight chemical residues. The percent contribution of ranked food items estimated using the NHEXAS samples is smaller than those obtained from the western US population via the DEPM. This indicates differences in consumption characteristics of the two groups with respect to the ranked food items. Six of 15 food items consumed by the NHEXAS-AZ subjects per day are ranked food items contributing between 56% and 70% of the estimated NHEXAS-AZ dietary exposure to each of the eight chemical residues. The difference between total dietary exposure estimates from the DEPM and NHEXAS-AZ approaches varies by chemical residue and is attributable to differences in sampling and analytical methods, and geographic areas represented by the data. Most metal exposures estimated using the NHEXAS consumption data with the CNRD have lower values than those estimated via the other approaches, possibly because the NHEXAS-AZ residue values are higher than the CNRD values. In addition, exposure estimates are seemingly affected by the difference in demographic characteristics and factors that affect types and amounts of food consumed. Efficient control strategies for reducing dietary exposure to chemical residues may be designed by focusing on the relatively small number of food items having similar ingredients that contribute substantively to the total ingestion exposure.