Performance evaluation of a beta version of USEPA’s SHEDS-HT chemical exposure model

Abstract

The United States Environmental Protection Agency (USEPA) is in the process of developing a high throughput (HT) model for exposure-based prioritization to inform toxicity testing and chemical risk assessment within its ExpoCast program. This mechanistic modeling approach is adapted from the prior Stochastic Human Exposure and Dose Simulation (SHEDS) framework, which was implemented in SAS. SHEDS-HT, written in R, reduces user burden and input demands by linking chemicals to particular exposure scenarios based on consumer product categories or food groups. SHEDS-HT rapidly generates probabilistic population distribution estimates of pathway-specific exposures, overall exposures, and intake doses. These predictions are based on specific exposure scenarios, a fugacity-based indoor environmental media model, and information from human activity databases. The predictive ability of SHEDS-HT is best evaluated by comparison to available biomarker measurements or to alternative predictions from prior measurement-intensive exposure studies. A key example of the latter is the USEPA’s Children’s Total Exposure to Persistent Pesticides and Other Persistent Organic Pollutants (CTEPP) study, which was conducted in Ohio and North Carolina. CTEPP provides both environmental and corresponding urinary biomarker data for a relatively large sample of commonly encountered commercial chemicals. The National Health and Nutrition Examination Survey (NHANES) also provides representative urinary biomarker data for some of the same compounds on a whole US population basis. However, biomarker data can reflect exposure to both a parent compound and its metabolite(s). Absence of environmental measurements for most biomarkers greatly reduces the number of compounds for which complete accounting of inputs and outputs can be conducted. In this evaluation of SHEDS-HT, focus is placed on 2,4-dichlorophenoxyacetic acid (2,4-D) and three parent compounds of the urinary biomarker 3,5,6-trichloro-2-pyridinol (TCPy), chlorpyrifos, chlorpyrifos methyl, and triclopyr.