Exposure Through Soil and Dust Ingestion - Dealing with Contaminated Sites: From Theory Towards Practical Application - page 278

Human Exposure and Dose Simulation) model. Soil ingestion rate estimates were
derived using data from Calabrese's Amherst and Anaconda studies (by Stanek and
Calabrese ( 2000 ) and Stanek et al. ( 2001a )). Data statistics from both of these stud-
ies were used to fit distributions of soil/dust ingestion rates (Stanek et al. 2001b ).
The statistical distributions generated for variability and uncertainty distributions
relied upon two tracers only, aluminium and silicon, in estimating the parameters
of the lognormal variability and uncertainty distributions. Using Monte-Carlo sam-
pling, values from the fitted distribution were sampled. The sampled values were
separated into those values under 500 mg/d and values that exceeded 500 mg/d. The
model assumes that soil ingestion values that exceed 500 mg/d are representative for
pica behaviour. The soil ingestion rate distribution for non-pica behaviour children
and children exhibiting pica behaviour are shown in Table 6.5 .
The age-specific default soil/dust ingestion rates recommended for use in the
Integrated Exposure Uptake Biokinetic (IEUBK) model for lead in children range
from 85 to 135 mg/d and are illustrated in Table 6.6 . These values are considered
representative of arithmetic mean daily intake rates. The default soil and dust inges-
tion rates are based on the tracer studies of Binder et al. ( 1986 ), Clausing et al.
( 1987 ), Calabrese and Stanek ( 1992 ) Calabrese et al. ( 1989 ), Van Wijnen et al.
( 1990 ), and Davis et al. ( 1990 ).
Van Holderbeke et al. ( 2008 ) conducted an analysis of soil ingestion rates from
tracer studies, hand-loading studies, Biokinetic model comparison and empirical
relations between contaminant levels in the environment and Biomonitoring data.
The authors selected tracer studies fulfilling the following requirements:

• data were corrected for intake of tracers from sources other than soil;

• only the most recently published re-analysed results were selected (published and

re-analysed by the same authors);

• data on titanium were left out, because this was considered the least reliable

tracer.

Table 6.5 The soil ingestion rate distribution for non-pica behaviour children and children
exhibiting pica behaviour as input for the SHEDS-WOOD model

Variability
distribution

Arithmetic
mean

St.
dev. Scenario Median P25 P75 P95 P99 Resources

Typical
child

Lognormal
(31.4) <
500 mg/d

60.6 80.5 29.8 11.9 73.4 236 402 Stanek and

Calabrese
( 2000 );
Stanek
et al.
( 2001a )

Pica child Lognormal

(31.4) >
500 mg/d

962 758 735 590 1046 2130 3852 ATSDR

( 2001 )

Zartarian et al. ( 2005 )