Assessment of oral bioaccessibility of arsenic in playground soil in Madrid (Spain): a three-method comparison and implications for risk assessment.

Three methodologies to assess As bioaccessibility were evaluated using playground soil collected from 16 playgrounds in Madrid, Spain: two (Simplified Bioaccessibility Extraction Test: SBET, and hydrochloric acid-extraction: HCl) assess gastric-only bioaccessibility and the third (Physiologically Based Extraction Test: PBET) evaluates mouth-gastric-intestinal bioaccessibility. Aqua regia-extractable (pseudo total) As contents, which are routinely employed in risk assessments, were used as the reference to establish the following percentages of bioaccessibility: SBET-63.1; HCl-51.8; PBET-41.6, the highest values associated with the gastric-only extractions. For Madrid playground soils--characterised by a very uniform, weakly alkaline pH, and low Fe oxide and organic matter contents--the statistical analysis of the results indicates that, in contrast with other studies, the highest percentage of As in the samples was bound to carbonates and/or present as calcium arsenate. As opposed to the As bound to Fe oxides, this As is readily released in the gastric environment as the carbonate matrix is decomposed and calcium arsenate is dissolved, but some of it is subsequently sequestered in unavailable forms as the pH is raised to 5.5 to mimic intestinal conditions. The HCl extraction can be used as a simple and reliable (i.e. low residual standard error) proxy for the more expensive, time consuming, and error-prone PBET methodology. The HCl method would essentially halve the estimate of carcinogenic risk for children playing in Madrid playground soils, providing a more representative value of associated risk than the pseudo-total concentrations used at present.