SOIL INGESTION ESTIMATES FOR CHILDREN RESIDING ON A SUPERFUND SITE

Soil ingestion estimates were obtained from a stratified, simple rando m sample of 64 children aged 1-4 years residing on a superfund site in Montana. The study was conducted during the month of September for 7 consecutive days. The study utilized a mass-balance methodology in whi ch eight naturally occurring soil tracers (Al, Si, Ti, Ce, Nd, La, Y, and Zr) believed to be poorly absorbed by the gastrointestinal tract w ere employed to provide soil ingestion estimates. Food and fecal sampl es were analyzed on a daily basis. Soil/dust samples representative of where the children played during the study period were obtained. Very high compliance among the participants was maintained throughout the study. The identical methodology employed in the childrens' study was used in an adult study of tracer recovery in 10 subjects over 28 days of observation to provide validation that soil ingestion over the rang e of 20 to 500 mg/day could be detected. Soil ingestion was estimated by each soil tracer via traditional methods as well as by an improved approach using five trace elements (Al, Si, Ti, Y, and Zr), called the Best Tracer Method (BTM), which substantially corrects for error due to misalignment of tracer input and output as well as error occurring from ingestion of tracers from nonfood, nonsoil sources, while being i nsensitive to the particle size of the soil/dust ingested. According t o the BTM, the median soil ingestion was less than 1 mg/day while the upper 95% was 160 mg/day. No significant age (1 year vs 2, vs 3) or se x-related differences in soil ingestion were observed. These estimates are lower than estimates observed in another study in Massachusetts d uring September and October. Significant methodological improvements i n this study as compared to previously conducted soil ingestion studie s include the selection of a representative sample of children, longer study duration, inclusion of dietary recommendations to reduce food t racer input and variability, use of the BTM, and a stronger adult vali dation study with respect to number of subjects, and duration and rang e of possible soil ingestion rates. Despite these methodological impro vements, evidence exists that this study displays a net residual negat ive error, suggesting that the above estimates are below the true soil ingestion. The magnitude of this residual negative error cannot be qu antified with the BTM but is likely to not affect the median by more t han 40 mg/day, while the impact of such an error on the upper end of t he distribution is more uncertain. (C) 1997 Academic Press.