Predicting oral relative bioavailability of arsenic in soil from in vitro bioaccessibility

ABSTRACT

Several investigations have been conducted to develop in vitro bioaccessibility (IVBA) assays that reliably predict in vivo oral relative bioavailability (RBA) of arsenic (As). This study describes a meta-regression model relating soil As RBA and IVBA that is based upon data combined from previous investigations that examined the relationship between As IVBA and RBA when IVBA was determined using an extraction of soil in 0.4 M glycine at pH 1.5. Data used to develop the model included paired IVBA and RBA estimates for 83 soils from various types of sites such as mining, smelting, and pesticide or herbicide application. The following linear regression model accounted for 87% of the observed variance in RBA (R² = .87): RBA(%) = 0.79 × IVBA(%) + 3.0. This regression model is more robust than previously reported models because it includes a larger number of soil samples, and also accounts for variability in RBA and IVBA measurements made on samples collected from sites contaminated with different As sources and conducted in different labs that have utilized different experimental models for estimating RBA.

Acknowledgments

Portions of this work were funded by the U.S. Environmental Protection Agency, Office of Superfund Remediation and Technology Innovation (OSRTI), under General Services Administration contract GS 00F 0019L and U.S. Environmental Protection Agency Contract EP-W-09-27, and by U.S. Department of Defense Environmental Security Technology Certification Program grant ESTCP ER-0916. The authors gratefully acknowledge the contributions of Penny Hunter (Environmental Resources Management), William Thayer (SRC, Inc.), Lynn Woodbury (CDM Smith), and the U.S. EPA OSRTI Technical Review Workgroup Bioavailability Committee.