Measured versus simulated dietary pesticide intakes in children

Abstract

Children's dietary pesticide intakes can be measured directly through duplicate diet samples, but instead are frequently estimated using national residue data. We compared organophosphorus and pyrethroid pesticide intakes from conventional fruit, fruit juices and vegetables measured for two cohorts of children aged 3–11 years in the Children's Pesticide Exposure Study (CPES) with corresponding intakes simulated using CPES consumption and body weight data and residue data from the US Pesticide Data Program (PDP). We calculated daily measured pesticide intakes by multiplying grams eaten with measured concentrations and dividing by body weight. For the simulated intakes we combined the CPES consumption and PDP residue data, randomly sampling the PDP data 500 times in order to create distributions of daily intakes for each cohort, including 95% uncertainty intervals for each percentile. In all cases, the measured medians fell below the lower uncertainty bounds of the simulated medians, reflecting the lower detection limits of CPES versus PDP and the high number of non-detects in each. Upper percentile measured intakes were generally lower as well, except for higher measured intakes of phosalone from watermelon. This work shows that using PDP data could generate probabilistic estimates of dietary pesticide intakes that do not differ appreciably from measured intakes except in some cases.

Keywords:

• children
• Children's Pesticide Exposure Study (CPES)
• diet survey
• Pesticide Data Program (PDP)
• pesticides
• risk assessment

Acknowledgements

This paper was developed under the support of the Science to Achieve Results (STAR) research assistance programme (R-832244 and R-829364), awarded by the US Environmental Protection Agency (USEPA). The content of this paper was not formally reviewed by the USEPA, and the views expressed in this document are solely the authors’. The USEPA does not endorse any products or commercial services mentioned in this paper. The authors thank Steve Nako and David Miller of the USEPA for helpful comments; and Nancy Diao at Harvard School of Public Health for her assistance in running the CPES-WA models on the cluster computers. The authors disclose no potential conflicts of interest regarding this paper.