Hand-to-mouth activity is an important pathway for children's exposure to contaminated soils, which is often co-contaminated by Pb and As in mining and smelting sites. To reduce soil Pb risk to humans by oral exposure, phosphate amendments have been used to reduce Pb relative bioavailability (RBA), but its efficiency has not been investigated using validated in vitro assays nor its influence on As-RBA. Here, 5 contaminated soils (A-E) were amended with 0.5% phosphoric acid (PA) to study its effect on Pb- and As RBA using a newly-developed mouse kidney model and bioaccessibility using 4 in vitro assays including UBM, SBRC, IVG, and PBET. Based on the mouse kidney model, Pb-RBA in PA-amended soils decreased from 14.2-62.5% to 10.1-29.8%. In contrast, As-RBA decreased from 26.5% to 15.9% in soil B but increased from 27.5 to 41.2% in soil D, with changes being insignificant in 3 other soils (35.8-58.8 to 28.1-61.1%). When assessing Pb bioaccessibility in PA-amended soils, decreased bioaccessibility were found using PBET and SBRC. For As, its bioaccessibility increased in PA-amended soils, inconsistent with in vivo data. Our results shed light on the importance of method selection to assess risk in Pb- and As-contaminated soils amended with phosphate. (C) 2017 Elsevier B.V. All rights reserved.