Equation Reliability of Soil Ingestion Estimates in Mass‐Balance Soil Ingestion Studies

Abstract

Exposure to chemicals from ingestion of contaminated soil may be an important pathway with potential health consequences for children. A key parameter used in assessing this exposure is the quantity of soil ingested, with estimates based on four short longitudinal mass-balance soil ingestion studies among children. The estimates use trace elements in the soil with low bioavailability that are minimally present in food. Soil ingestion corresponds to the excess trace element amounts excreted, after subtracting trace element amounts ingested from food and medications, expressed as an equivalent quantity of soil. The short duration of mass-balance studies, different concentrations of trace elements in food and soil, and potential for trace elements to be ingested from other nonsoil, nonfood sources contribute to variability and bias in the estimates. We develop a stochastic model for a soil ingestion estimator based on a trace element that accounts for critical features of the mass-balance equation. Using results from four mass-balance soil ingestion studies, we estimate the accuracy of soil ingestion estimators for different trace elements, and identify subjects where the difference between Al and Si estimates is larger (>3 RMSE) than expected. Such large differences occur in fewer than 12% of subjects in each of the four studies. We recommend the use of such criteria to flag and exclude subjects from soil ingestion analyses.

Meta‐Analysis of Mass‐Balance Studies of Soil Ingestion in Children

“…We use the RMSE estimates of accuracy from Stanek et al . ( 35 ) to construct a weighted least squares (WLS) estimate of soil ingestion based on Al and Si to screen for possible source error prior to conducting the analysis. A subject's estimate of soil ingestion was considered to have an elevated chance for bias if the trace element soil ingestion estimate of Al or Si differed by more than three standard deviations from the WLS mean (constructed from the Al and Si soil ingestion estimates for the subject).…”

“…We used data from each soil ingestion study in the mass‐balance soil ingestion equation to evaluate the accuracy of soil ingestion estimators based on different trace elements. ( 35 )…”

“…In order to evaluate and compare the reliability of trace elements for soil ingestion estimation, and identify possible source errors for individual trace elements, we developed a model for soil ingestion in mass‐balance studies. ( 35 ) We briefly review the results of this development and their application to data used in the meta‐analysis. Under an idealized setting, there is no trace element bioavailability, no sample loss, no measurement error, and no missed input or output samples in the soil ingestion study.…”

“…We used data from each soil ingestion study in the massbalance soil ingestion equation to evaluate the accuracy of soil ingestion estimators based on different trace elements. (35) A third consideration was identification of subject-periods where there was evidence of large trace element ingestion from non-(food, beverage, medication, soil) sources. Since the quantity of trace element from such sources is not known, these biasing effects can only be observed indirectly by comparing trace element amounts measured in fecal samples relative to the expected amount of trace element in the fecal sample based on measured intake and estimated soil ingestion (based on other trace elements).…”

Soil ingestion rates for children under 3 years old in Taiwan

“…A number of previous studies have shown various limitations and uncertainties for the tracer element methodology including study duration and location, representativeness of soil tracer element concentrations, adjustments made for tracer element concentrations of non-soil substances, gastrointestinal absorption and transit times, weight of feces, and recovery rates of different tracer elements. [1][2][3][4][5][6][7][8][9][10][11][12][13] The tracer should be abundant and uniform in soils and low in the diet or food. Moreover, it assumed that children primarily ingested soil from their house, and the soil samples could be representative of the entire house.…”