Meta-analysis of soil and dust ingestion studies

The ingestion of soil and dust by children and adults is a potential source of exposure to environmental contaminants. To advance beyond the simple averaging of estimates used in the U.S. EPA's Exposure Factors Handbook (EFH), we describe a novel meta-analysis of all available studies that provided soil or dust ingestion estimates for children or adults conducted in the United States and Canada. Using meta-analytic techniques, we estimate the mean total soil plus dust ingestion rates and confidence intervals (CIs) for eleven age groups (0 - <1 month (m), 1 - <3 m, 3 - <6 m, 6 - <12 m, 1 - <2 years (y), 2 - <3 y, 3 - <6 y, 6 - <11 y, 11 - <16 y, 16 - <21 y, and 21+ y). These age groups were selected for consistency with the EFH update to Chapter 5 and the U.S. EPA's Age Grouping Guidance. For each age group, we calculated best estimates for the three main types of ingestion studies: tracer studies based on the aluminum tracer, biokinetic studies, and activity pattern (modeling) studies, as well as overall estimates for all three study types combined. Our meta-analysis combined study estimates using the alternative statistical approaches of the fixed effect method (inverse variance method, "I-V") and two random effects methods, DerSimonian and Laird's method of moments ("DSL") and the restricted maximum likelihood method ("MIXED"). For each approach, the mean total soil plus dust ingestion rate estimates for each study type generally aligned well with the EFH, ranging from 36 to 68 mg/day for infants, 56-72 mg/day for young children, and 12-32 mg/day for adolescents and adults. When all three study types were combined, the upper bounds of the 95% CI were generally the lowest for the I-V method and the highest for the MIXED method. The estimates produced here can be used for stochastic risk assessments and provide a better estimate of soil and dust ingestion rates across age groups.