The aim of the present study was to determine the internal dose in humans after the ingestion of soil highly contaminated with uranium. Therefore, an in vitro solubility assay was performed to estimate the bioaccessibility of uranium for two types of soil. Based on the results, the corresponding bioavailabilities were assessed by using a recently published method. Finally, these bioavailability data were used together with the biokinetic model of uranium to assess the internal doses for a hypothetical but realistic scenario characterized by a daily ingestion of 10 mg of soil over 1 year. The investigated soil samples were from two former uranium mining sites of Germany with (238)U concentrations of about 460 and 550 mg/kg. For these soils, the bioavailabilities of (238)U were quantified as 0.18 and 0.28 % (geometric mean) with 2.5th percentiles of 0.02 and 0.03 % and 97.5th percentiles of 1.48 and 2.34 %, respectively. The corresponding calculated annual committed effective doses for the assumed scenario were 0.4 and 0.6 µSv (GM) with 2.5th percentiles of 0.2 and 0.3 µSv and 97.5th percentiles of 1.6 and 3.0 µSv, respectively. These annual committed effective doses are similar to those from natural uranium intake by food and drinking water, which is estimated to be 0.5 µSv. Based on the present experimental data and the selected ingestion scenario, the investigated soils-although highly contaminated with uranium-are not expected to pose any major health risk to humans related to radiation.

The aim of the present study was to improve the estimation of soil-derived uranium absorption in humans. For this purpose, an in vitro solubility assay was combined with a human study by using a specific edible soil low in uranium. The mean bioaccessibility of the soil-derived uranium, determined by the solubility assay in artificial gastrointestinal fluid, was found to be 7.7% with a standard deviation of 0.2%. The corresponding bioavailability of the soil-derived uranium in humans was assumed to be log-normal distributed with a geometric mean of 0.04% and a 95% confidence interval ranging from 0.0049% to 0.34%. Both results were used to calculate a factor, denoted as fA(sol), which describes the relation between the bioaccessibility and the bioavailability of soil-derived uranium. The geometric mean of fA(sol) was determined to be 0.53% with a 95% confidence interval ranging from 0.06% to 4.43%. Based on fA(sol), it is possible to estimate more realistic values on the bioavailability of uranium for highly uranium-contaminated soils in humans by just performing the applied solubility assay. The results of this study can be further used to obtain more reliable results on the internal dose assessment of ingested highly uranium-contaminated soils.

In the text, at page 14725, right column, line 19 and 20, the units are incorrectly given (μg/day instead of ng/day). The amended sentence is "Thereby, the daily urinary excretion values of 238 U over 3 days before healing soil ingestion were 3.9 ng/day ±1.5 ng/day, 8.1 ng/day ±0.6 ng/day and 9.0 ng/day ±1.2 ng/day (mean ± SD)."