Issues in setting health-based cleanup levels for arsenic in soil.

Health risk assessments often do not take into account the unique aspects of evaluating exposures to arsenic in soil. For example, risks from ingestion of arsenic in soil are often based on toxicity factors derived from studies of arsenic (soluble arsenate or arsenite) in drinking water. However, the toxicity of arsenic in drinking water cannot be directly extrapolated to toxicity of soil arsenic because of differences in chemical form, bioavailability, and excretion kinetics. Because of the differences between soil arsenic and water arsenic, we conclude that risks from arsenic in soil are lower than what would be calculated using default toxicity values for arsenic in drinking water. Site-specific risk assessments for arsenic in soil can be improved by characterizing the form of arsenic in soil, by conducting animal feeding or in vitro bioavailability studies using site soils, and by conducting studies to evaluate the relationship between urinary arsenic and soil arsenic levels. Such data could be used to more accurately measure the contribution that soil arsenic makes to total intake of arsenic. Available data suggest that arsenic usually makes a small contribution to this total.