NERL Research Abstract

EPA's National Exposure Research Laboratory
GPRA Goal 8 - Sound Science

Significant Research Findings

First-Generation Residential Pesticide Exposure Model

Scientific Problem and The Food Quality Protection Act (FQPA) of 1996 requires EPA to

consider aggregate human exposure to a pesticide, especially for infants
and children, when setting regulatory limits on that pesticide's usage.
Aggregate exposure refers to the total exposure of humans to the
pesticide from all its uses and through various pathways and routes.
These include dietary ingestion of pesticide residues in food and
drinking water, inhalation of air containing pesticides, dermal contact
with surfaces containing residues (indoors and residential lawns), and
non-dietary ingestion of residues from hand- or object-to-mouth
activities. Estimation of the population's exposure from dermal and
non-dietary pathways is difficult because many difficult-to-measure
factors (such as the frequency and subsequent transfer from contacting
surfaces, putting fingers and objects into one's mouth, and washing
one's hands) influence a person's exposure. This results in a wide range
of variability of exposures to individuals within a population. Hence, a
probabilistic model that predicts the range and distribution of personal
exposures and doses within a population was developed in this project.
The model is called the Stochastic Human Exposure and Dose
Simulation Model for Pesticides, or SHEDS-Pesticides. Following is a
more detailed summary of work and accomplishments to date.

The study's primary objective is to develop a computerized modeling
framework for conducting probabilistic assessments of aggregate
pesticide exposure and dose. This framework is demonstrated by
performing an example assessment for the organophosphorus
insecticide chlorpyrifos. Chlorpyrifos is widely used both in homes and
on food crops. The assessment focused on children's exposure and
absorption via the pathways of dermal (skin) contact with chlorpyrifos-
contaminated surfaces and also non-dietary ingestion caused by transfer
of chlorpyrifos from hands or objects to the mouth. Some of the
relevant factors that affect exposures by these pathways include
amounts of chlorpyrifos per unit area on carpets and other household
surfaces and lawns, the frequency of contacting such surfaces, the
amounts of time spent in various places in a residence, and the fraction
transferred to the mouth when fingers or objects are placed in the

Research
Approach

-------
mouth. Distributions of values for each factor were estimated from
published studies where available, or from the judgment of researchers.
Data from a national survey of how and where people spend time were
combined with data from analysis of videotapes of children playing and
incorporated into the model. The model also included factors to
simulate how much of the pesticide actually entered the child's blood
(i.e., the "dose"), and how much of its metabolite was eliminated each
day in the child's urine. It is important to model the urinary levels in
order to compare the model outputs with other published data on
amounts of the metabolite found in urine. This type of comparison to
so-called "biomarkers" is one of the best available ways to judge the
validity of an exposure model's results. When all the needed data were
set as inputs to the model, the computer program performed simulations
for individuals. For each person, values were randomly sampled from
the distributions of factor values. These values were used in equations
in the model to compute 24-hour profiles of exposure and dose. The
process was repeated for 1,500 individuals. This yielded probabilistic
distributions of exposure and dose for a selected population of children.
Several such sets of distributions were produced, corresponding to
different residential uses of chlorpyrifos (crack-and-crevice spraying or
broadcast application), different periods of time after its usage, and
different age groups of children.

Results and	Results from this modeling study indicate that the median daily urinary

Implications	eliminated TCP could range from about one microgram per liter to

many hundreds of micrograms per liter, depending on the type of
application and time since the pesticide was applied. Broadcast
applications of chlorpyrifos, which have been phased out, resulted in
estimated exposures and doses that were about 10 to 100 times higher
than those from crack-and-crevice applications because of differences
in resulting surface residues. Non-dietary ingestion resulted in greater
doses than dermal absorption for acute (<1 day) and short-term (1-7
days) post-broadcast application scenarios; however, dermal absorption
was the major contributor to dose for the other scenarios. Younger
children (0-4 years old) had higher dermal exposures, non-dietary
ingestion, and blood and urine concentrations of the metabolite than did
older children (5-9 years old) because of differences in assumed activity
patterns. Contact with smooth surfaces caused higher exposure than
contact with textured surfaces such as carpets, or with lawn surfaces -
an effect due to greater transfer efficiency from smooth surfaces. Model
results of metabolite in urine are of the same order of magnitude as
published data from other studies, indicating that the model results are
realistic.

Implications from these findings are that (1) broadcast application of
indoor insecticides can result in much higher exposures than those from
crack-and-crevice applications, and that (2) incidental residential

-------
contact with and ingestion of pesticide residues can be a significant
contributor to exposure and dose. This reinforces the wisdom of
considering aggregate exposure from all pathways, as called for by
FQPA.

The development of this first-generation model is still at an early stage,
with many assumptions built into the model. Nevertheless, these
preliminary results help researchers to better understand the events and
factors that lead to pesticide exposure. As development progresses, the
model will help point out areas of greatest uncertainty and those
needing more research.

This research project directly supports ORD's research to improve the
scientific foundation of human health risk assessment under the
Government Performance and Results Act (GPRA) Goal 8 ("Sound
Science, Improved Understanding of Environmental Risk and Greater
Innovation to Address Environmental Problems"), Objective 2
("Research for Human Health Risk Assessment"), Sub-Objective 1
("Human Health Research"). The results of this project address GPRA
annual performance goal (APG) 02 ("by 1999 Produce First Generation
Exposure Models Describing Residential Exposure to Pesticides"),
annual performance measure (APM) 970 (" First Generation Residential
Exposure Model"). Although this work directly supports a Goal 8 APM,
the results also will be valuable to reduce uncertainties in exposure
assessement under GPRA Goal 3 (Safe Food, Research to Support New
Regulatory Requirements Under FQPA).

The SHEDS-Pesticide modeling project was conducted primarily by a
team of NERL staff scientists. EPA's Office of Pesticide Programs
cooperated and provided technical input. Contractor assistance in
writing computer programs was provided by ManTech Environmental
Technology, Inc. and independent consultant Dr. Jianping Xue.

This research has been presented in several conference presentations,
and in the following manuscript:

Zartarian, V.G., Ozkaynak, H., Burke, J.M., Zufall, M.J., Rigas, M.L., and Furtaw,
E.J.,Jr. (1999). "A Modeling Framework for Estimating Children's
Residential Exposure and Dose to Chlorpyrifos via Dermal Residue Contact
and Non-Dietary Ingestion." Submitted to Environmental Health
Perspectives.

Future Research	The SHEDS-Pesticides model described herein is part of a larger on-

going effort to develop models to improve our understanding of the
routes, pathways, and factors that contribute to pollutant exposure and
dose, especially to infants and children. As a first-generation exposure-
to-dose model, it will be improved in future work. More and better data
are needed for some of the exposure factor distributions. Additional
work is being conducted to incorporate into the model the dietary
ingestion and inhalation exposure routes. Also, work is being done to
refine the dose aspects of the model using the dose estimating exposure

Research Collaboration
and Publications

-------
models (DEEM) being developed in NERL. The DEEM research will
help SHEDS to better simulate the body's uptake, metabolism, and
elimination of pesticides.

Questions and inquiries on NERL's human exposure and dose modeling
research can be directed to:

Haluk Ozkaynak, Ph.D.

US EPA

National Exposure Research Laboratory (MD-56)

Research Triangle Park, NC 27711
Phone: 919/541-5172
E-mail: ozkaynak.haluk@epa.gov

-------