Abstract:Previous studies have established in vivo-in vitro correlations (IVIVC) between arsenic (As) relative bioavailability (RBA) and bioaccessibility in contaminated soils. However, their ability to predict As-RBA in soils outside the models is unclear. In this study, As bioaccessibility and As-RBA in 12 As-contaminated soils (22.2-4172 mg kg(-1) As) were measured using five assays (SBRC, IVG, DIN, PBET, and UBM) and a mouse blood model. Arsenic RBA in the soils ranged from 6.38 ± 2.80% to 73.1 ± 17.7% with soils containing higher extractable Fe showing lower values. Arsenic bioaccessibility varied within and between assays. Arsenic bioaccessibility was used as input values into established IVIVC to predict As-RBA in soils. There were significant differences between predicted and measured As-RBA for the 12 soils, illustrating the inability of established IVIVC to predict As-RBA in those contaminated soils. Therefore, a new IVIVC was established by correlating measured As-RBA and As bioaccessibility for the 12 soils. The strength of the predictive models varied from r(2) = 0.50 for PBET to r(2) = 0.83 for IVG, with IVG assay providing the best prediction of As-RBA. When IVIVC were compared to those of Juhasz et al. (2014a), slopes of the relationships were significantly higher possibly due to different As-RBA measurements. Our research showed that IVG has potential to measure As bioavailability in contaminated soils from China though UBM and SBRC assays were also suitable. More research is needed to verify their suitability to predict As-RBA in soils for refining health risk assessment.

In Vitro Bioaccessibility and in Vivo Relative Bioavailability in 12 Contaminated Soils: Method Comparison and Method Development