Strategies for assessing children’s organophosphorus pesticide exposures in agricultural communities

Abstract

Children can be exposed to pesticides from multiple sources and through multiple pathways. In addition to the standard pathways of diet, drinking water and residential pesticide use, children in agricultural communities can be exposed to pesticides used in agricultural production. A research program on children and pesticides was established at the University of Washington (UW) in 1991 and has focused on two major exposure pathway issues: residential proximity to pesticide-treated farmland and transfer of pesticides from the workplace to the home (paraoccupational or take-home exposure). The UW program selected preschool children of agricultural producers and farm workers in the tree fruit region of Washington state as a population that was likely to have elevated exposures from these pathways. The organophosphorus (OP) pesticides were selected as a common class of chemicals for analysis so that issues of aggregate exposure and cumulative risk could be addressed. This paper provides an overview of key findings of our research group over the past 8 years and describes current studies in this field. Soil and housedust concentrations of OP pesticides were elevated in homes of agricultural families (household members engaged in agricultural production) when compared to non-agricultural reference homes in the same community. Dialkyl phosphate metabolites of OP pesticides measured in children's urine were also elevated for agricultural children when compared to reference children and when compared to children in the Seattle metropolitan area. Proximity to farmland was associated with increased OP pesticide concentrations in housedust and OP pesticide metabolites in urine. Current studies include a community-based intervention to reduce parental transfer of pesticides from the workplace, and a systematic investigation of the role of agricultural spray drift in children's exposure to pesticides.

Full text

Strategies for assessing children's organophosphorus pesticide exposures
in agricultural communities
RICHARD A. FENSKE, CHENSHENG LU, NANCY J. SIMCOX, CARRIE LOEWENHERZ,
JENNIFER TOUCHSTONE, THOMAS F. MOATE, EMILY H. ALLEN AND JOHN C. KISSEL
Department of Environmental Health and Pacific Northwest Agricultural Safety and Health Center, School of Public Health and Community Medicine,
University of Washington, Seattle, Washington 98195
Children can be exposed to pesticides from multiple sources and through multiple pathways. In addition to the standard pathways of diet, drinking water and
residential pesticide use, children in agricultural communities can be exposed to pesticides used in agricultural production. A research program on children and
pesticides was established at the University of Washington (UW ) in 1991 and has focused on two major exposure pathway issues: residential proximity to
pesticide - treated farmland and transfer of pesticides from the workplace to the home ( paraoccupational or take - home exposure ) . The UW program selected
preschool children of agricultural producers and farm workers in the tree fruit region of Washington state as a population that was likely to have elevated
exposures from these pathways. The organophosphorus ( OP) pesticides were selected as a common class of chemicals for analysis so that issues of aggregate
exposure and cumulative risk could be addressed. This paper provides an overview of key findings of our research group over the past 8 years and describes
current studies in this field. Soil and housedust concentrations of OP pesticides were elevated in homes of agricultural families ( household members engaged
in agricultural production) when compared to non- agricultural reference homes in the same community. Dialkyl phosphate metabolites of OP pesticides
measured in children's urine were also elevated for agricultural children when compared to reference children and when compared to children in the Seattle
metropolitan area. Proximity to farmland was associated with increased OP pesticide concentrations in housedust and OP pesticide metabolites in urine.
Current studies include a community - based intervention to reduce parental transfer of pesticides from the workplace, and a systematic investigation of the role
of agricultural spray drift in children's exposure to pesticides. Journal of Exposure Analysis and Environmental Epidemiology (2000 ) 10, 662± 671.
Keywords: agriculture, children, exposure assessment, housedust, organophosphorus pesticides, urinary metabolites
.
Introduction
The potential health risks associated with exposure of
children to pesticides have received increased attention
with the passage of the Food Quality Protection Act of
1996 and with new federal emphasis on children's health
(U.S. Environmental Protection Agency, 1997) . Current
efforts to characterize such exposures within a risk
assessment framework have focused primarily on expo-
sures of the general population. Definitions of specific
subpopulations are currently restricted to age categories
(e.g., children and adults) and occupation (e.g., pesticide
handlers, agricultural reentry workers) . However, children
in farming communities can also be viewed as a definable
subpopulation because it is known that pesticide use
patterns in such communities differ from those in
metropolitan and other rural areas. The U.S. Environ-
mental Protection Agency ( EPA ) recognizes that these
children are likely to have a different exposure profile
than those in the general population, and has devoted
resources to promote research on such populations,
including the studies reported here (U.S. Environmental
Protection Agency, 1999). There is a clear need to
develop appropriate methods to evaluate exposures in
these children.
Research on children and pesticides began at the
University of Washington (UW) in 1991. The UW program
was developed within a public health research model that
employs epidemiological and surveillance approaches to
risk characterization. The UW program has been guided by
four principles: (1) focus research on high-risk popula-
tions; (2) evaluate exposure to a common class of
chemicals; (3) measure multiple exposure pathways; and
(4 ) identify opportunities for community-based inter-
vention.
Abbreviations: DEP, diethyl phosphate; DETP, diethylthio phosphate;
DEDTP, diethyldithio phosphate; DMP, dimethyl phosphate; DMTP,
dimethylthio phosphate; DMDTP, dimethyldithio phosphate; EPA, U.S.
Environmental Protection Agency; GPS, global positioning system;
LIDAR, light detection and ranging; OP, organophosphorus; UW,
University of Washington; WIC, Women, Infants, and Children program;
g, gram ; l, liter; ml, milliliter; ng, nanogram; g, microgram
Address all correspondence to: Dr. Richard A. Fenske, University of
Washington, Box 357234, Seattle, WA 98195. Tel.: + 1 - 206- 616- 1958.
Fax: + 1 - 206 - 616 - 2687. E-mail: rfenske@u.washington.edu
Received 3 December 1999; accepted 17 July 2000.
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10, 662 ± 671
# 2000 Nature America, Inc. All rights reserved 1053-4245/00/$15.00
www.nature.com/jea

High- risk populations can be defined as groups which
are either more highly exposed to an environmental agent or
more susceptible to its effects (Ashford et al., 1990 ).
Preschool children of agricultural producers and farm
workers appear to meet both of these criteria; children are
considered potentially vulnerable to environmental agents
due to developing organ systems, and children in farm
communities are likely to receive greater pesticide expo-
sures than other children.
The organophosphorus (OP ) pesticides were selected as
an important class of chemicals that could be evaluated
simultaneously. The integration of such exposures via
common biomarkers is also possible. They have since been
designated as a class of compounds that act by a common
mechanism (Mileson et al., 1998), and have been selected
by EPA for cumulative risk evaluation (International Life
Sciences Institute, 1999; U.S. Environmental Protection
Agency, 1999).
The OP pesticides are widely used in agriculture, but are
also used in residential settings for pest control. In outdoor
settings, OP pesticides are relatively non- persistent as they
are degraded by natural and microbiological actions.
However, when used indoors or as a part of structural
treatments, these compounds can remain stable for extended
periods of time (i.e., months to years) . EPA currently
considers children's exposure to be characterized by three
pathways: diet, drinking water, and residential use. In our
studies, we have hypothesized that children in agricultural
communities receive additional exposures from living near
agricultural pesticide use and because of their parents'
occupation. We have designated these additional exposure
sources as the ``farm proximity pathway'' and the ``parental
take- home pathway'' (Figure 1). Thus, children in
agricultural communities have a more complex aggregate
exposure profile than other children, as well as the potential
for exposure to multiple OP compounds.
Methods
Sampling Frame Ð Study Design
A major challenge in the conduct of population- based
exposure assessment studies is the definition of a sampling
frame. Most sampling frames are geographical and are often
based on existing database definitions; e.g., census data or
political jurisdictions such as states or counties. Ideally, a
probabilistic sample can be drawn from a well- defined
sampling frame so that results can be generalized to the
entire population within the frame. The current National
Human Exposure Assessment Survey is a good example of
such an approach (Callahan et al., 1995 ). However, the
definition of ``agricultural communities'' is more proble-
matic. Such communities are widely dispersed and may not
conform to census or political boundaries. Also, traditional
methods of access to families may not be feasible in such
communities. Multiple families may live in residences
designed for a single family, and telephone -based sampling
methods may miss a significant fraction of the population.
In Washington state's agricultural regions, the primary
language is Spanish, so bilingual capabilities are essential.
Finally, age is an important consideration for a sampling
frame directed toward children. Recruitment of children
within a relatively narrow age range ( e.g., preschool) can
Figure 1. Pesticide exposure pathways for children in agricultural
communities. The three traditional pathways of drinking water, diet,
and residential use are shown on the left of the figure. Two additional
pathways relevant to farm children Ð farm proximity and parental
take-home Ð are shown on the right.
Figure 2. Pesticide use patterns in Washington state apple orchards in
1998. EPA placed new restrictions on azinphos methyl and methyl
parathion use in 1999. Carbaryl, a carbamate insecticide (cholinester-
ase inhibitor), is used as a chemical thinning agent.
Strategies for assessing children's pesticide exposures Fenske et al.
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(6) Part 2 663

be daunting with traditional methods; i.e., the fraction of the
population with young children may be relatively small,
requiring initial screening of a large sample.
The study area selected for our studies centered around
the city of Wenatchee, Washington, considered the heart of
apple cultivation in the region. The region consists of an
urban zone along the Columbia River, with orchards
extending into the surrounding mountain canyons as well
as upriver, and newer residential development interspersed
with farmland. This entire region Ð some 150 square miles
Ð was considered the ``agricultural community'' for this
study. Orchard management in the area includes periodic
application of several OP pesticides, along with numerous
other crop protection chemicals. Figure 2 illustrates the
pattern of OP pesticide use in the region's apple orchards
across the growing season. Azinphos methyl, chlorpyrifos,
and diazinon are commonly used in the early spring and
throughout most of the season. Phosmet, malathion, methyl
parathion, and carbaryl (a carbamate cholinesterase in-
hibitor) are used from mid -May through August.
Subject recruitment took place through community
organizations, including social service agencies and produ-
cer-operated cooperatives. This approach allowed us to
quickly identify families with young children. In our 1998
studies, we attempted probability-based sampling based on
census tract data, but this approach required a randomized
door-to- door contact because most of the population did
not have telephone service. We also found that families were
Table 1. Organophosphate pesticides registered in Washington state in 1998
a
.
Active ingredient Alkyl ester structure
b
Dialkyl phosphate metabolites Specific OP metabolites
Acephate dimethyl (S) ±
c
Azinphos methyl dimethyl DMP, DMTP, DMDTP
Bensulide di-2-propyl ±
Chlorpyrifos diethyl DEP, DETP 3,5,6-trichloropyridinol
Coumaphos diethyl DEP, DETP
Diazinon diethyl DEP, DETP under development
d
Dichlorvos dimethyl DMP
Dicrotophos dimethyl DMP
Dimethoate dimethyl DMP, DMTP, DMDTP
Disulfoton diethyl DEP, DETP, DEDTP
Ethion diethyl DEP, DETP, DEDTP
Ethoprop dipropyl (S,S) ±
Ethyl parathion
e
diethyl DEP, DETP p-nitrophenol
Fenamiphos ethyl, 2-propyl (NH) ±
Fenthion dimethyl DMP, DMTP
Fonofos ethyl, (ethylphosphono) ±
Isofenphos ethyl, 2-propyl (NH) ±
Malathion dimethyl DMP, DMTP, DMDTP under development
Methamidophos dimethyl (S) ±
Methidathion dimethyl DMP, DMTP, DMDTP
Methyl parathion dimethyl DMP, DMTP p-nitrophenol
Naled dimethyl DMP
Phorate diethyl DEP, DETP, DEDTP
Phosmet dimethyl DMP, DMTP, DMDTP
Propetamphos ethyl, methyl (NH) ±
Sulfotepp diethyl DEP, DETP
Temephos dimethyl DMP, DMTP
Terbufos diethyl DEP, DETP, DEDTP
Tetrachlorvinphos dimethyl DMP
Trichlorfon methyl (C) ±
a
Source: Pesticide Information Center On-Line Databases, Washington State University, updated January 1999 E-mail: {http://picol.cahe.wsu.edu/plirs/
pl-logscreen.html}.
b
(S) means a sulfur substitution for oxygen in the ester; (NH) means amino substitution for oxygen in the ester; (C) means carbon in lieu of ester.
c
Compound would not be detected in normal dialkyl phosphate analysis.
d
Under development at Centers for Disease Control and Prevention (D. Barr, personal communication).
e
No longer registered for use in Washington state, but included as an analyte in WA studies.
Fenske et al. Strategies for assessing children's pesticide exposures
664 Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(6) Part 2

wary of strangers approaching their doors and were often
unreceptive to our request for participation. This method
was ultimately abandoned due to prohibitive cost. Instead,
we chose to work with the local WIC (Women, Infants, and
Children) clinic and recruited from its clients. This
approach assured us access to families with young children
and provided a non- threatening means of enrolling
participants.
The 1992 and 1995 studies divided households into two
groups based on proximity to farmland and parental
occupation. ``Agricultural'' families were defined as
households that included at least one adult working on a
farm. Adult workers were further classified as pesticide
applicators and farm workers in the 1995 study. None of
the pesticide applicators in these studies conducted this
activity full time; i.e., they were not commercial
applicators. Rather, they were responsible for periodic
treatment of crops as part of farm management. A smaller
``reference'' family population was also recruited. These
families had no household members working on a farm
and lived more than one quarter of a mile (about 400 m)
from the farmland. Airblast applications, which are
common in tree fruit orchards, are known to produce
measurable drift up to a distance of 200 ft or 60 m (Fox
et al., 1993) .
Children up to 6 years of age were recruited from these
families. Often, more than one child per family would
participate in the study. Samples from all participants were
analyzed, but for statistical purposes only, a single
``focus'' child from each household was used for most
data analyses to remove within- household dependence.
The focus child for a household was selected randomly
from children with complete urine samples and creatinine
measurements.
Environmental Sampling of OP Pesticides
In 1998, 30 OP pesticides were registered for use in
Washington state (Table 1). The need to measure
multiple OP pesticides in environmental media required
the development of new analytical methods by our
laboratory. While some laboratories offer screens for a
broad panel of pesticides, the accuracy and the precision
of such analyses are usually limited for specific
compounds. When we began this work in 1991, there
were no laboratories prepared to conduct multiple OP
residue analysis in media other than food, and even
acquiring appropriate standards was problematic. Our
1992 and 1995 studies focused on four OP pesticides used
in Washington state orchards Ð azinphos methyl,
phosmet, chlorpyrifos, and ethyl parathion Ð and
included soil and housedust sampling ( Simcox et al.,
1995; Loewenherz et al., 1997). The 1998 studies
expanded the list of target compounds to include diazinon,
dichlorvos, malathion, methyl parathion, methidathion,
mevinphos, ethoprop, phorat, dimethoate, and terbufos,
and sample media were expanded to include 24- h indoor
air, indoor and outdoor surface wipes, and drinking water
(Table 2 ). Duplicate 1-day diet samples were analyzed
by the Food and Environmental Quality Laboratory at
Washington State University. As Table 1 indicates,
however, more than half of the OP pesticides registered
in Washington State still fell outside these analytical
capabilities.
Dialkyl Phosphate Metabolite Analysis
The challenge of evaluating exposure to multiple OP
compounds extends to biological monitoring. Of the 30
pesticides listed in Table 1, only five have urinary
metabolites that can be considered compound- specific
Table 2. Analytes and sample matrices for 1992, 1995, and 1998 studies.
House dust Outdoor soil Indoor 24-h air Child hand wipe Toy wipe 24-h food Child urine
Azinphos methyl 1992, 1995, 1998 1992, 1998 1998 1998 1998 1998
Chlorpyrifos 1992, 1995, 1998 1992, 1998 1998 1998 1998 1998
Diazinon 1998 1998 1998 1998 1998 1998
Dichlorvos 1998 1998 1998 1998 1998 1998
Dimethoate 1998
Ethoprop 1998
Ethyl parathion 1992, 1995 1992 1998
Malathion 1998
Methyl parathion 1998
Methidathion 1998
Mevinphos 1998
Phorat 1998
Phosmet 1992, 1995, 1998 1992, 1998 1998 1998 1998 1998
Terbufos 1998
Dialkyl phosphates 1998 1995, 1998
Strategies for assessing children's pesticide exposures Fenske et al.
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(6) Part 2 665

(D. Barr, Centers for Disease Control, personal commu-
nication). The lack of specific metabolites for OP
pesticides has made the use of the more generic dialkyl
phosphate analysis attractive. The dialkyl phosphate
method was first developed for occupational exposure
assessments (Shafik et al., 1973) and allowed resolution
of dimethyl and diethyl compounds. Six metabolic
products are normally measured by gas chromatography
following derivatization: dimethyl phosphate (DMP) ,
dimethylthio phosphate (DMTP ), dimethyldithio phos-
phate (DMDTP) , diethyl phosphate ( DEP ), diethylthio
phosphate (DETP ), and diethyldithio phosphate
(DEDTP ). Figure 3 illustrates the possible metabolic
breakdown products for OP pesticides. OP compounds
with a sulfur atom ( thio) at the double bond position or
linked to the leaving group can appear in urine as either a
non- sulfonated metabolite ( DMP or DEP) or as a single
sulfonated metabolite ( DMTP or DETP). OP compounds
with sulfur atoms in both positions can produce all three
of the dialkyl phosphate metabolites. Not pictured in
Figure 3 are the non-sulfonated OP compounds that can
produce only DMP or DEP.
It is important to note, however, that even this more
generic assay does not necessarily capture all OP
compounds. As indicated in Table 1, 8 of 30 OP
pesticides have unusual alkyl ester structures, and in the
case of trichlorfon, a single carbon is present in lieu of an
alkyl ester. Metabolites from these compounds will
therefore differ from the six dialkyl phosphates mentioned
above. Two laboratories have published methods for the
dialkyl phosphate assay recently (Aprea et al., 1996;
Moate et al., 1999 ), and population surveys of these
Figure 3. Generic chemical structures of OP pesticides and dialkyl phosphate metabolites. Monothio compounds can produce the first two
metabolites. Dithio compounds can produce all three metabolites.
Fenske et al. Strategies for assessing children's pesticide exposures
666 Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(6) Part 2

metabolites are underway in some countries (Aprea et al.,
1996) . The limits of detection for the dialkyl phosphates
are now in the range of 1± 10 g/ l (1 ±10 ppb), and
these methods demonstrate good accuracy and precision.
The method described by Moate et al. ( 1999) has been
used since 1995 for our studies.
Results
The 1992 studies included soil and housedust sampling of
48 agricultural families and 11 reference families (Simcox
et al., 1995) . Figure 4 provides median values for four OP
pesticides in housedust and soil. These data indicated that
housedust concentrations were significantly higher than soil
concentrations for all compounds, and that azinphos methyl
and phosmet, both dimethyl compounds, were found at the
highest concentrations in housedust. These findings,
coupled with knowledge that these children spent most of
their time indoors, led to the conclusion that housedust
concentration was the most useful indicator of exposure
potential for this population. Figure 5 compares the OP
pesticide housedust concentrations for agricultural and
reference families, demonstrating that children in agricul-
tural households had higher exposure potential than did
children in reference families for all four OP compounds
measured. These findings confirmed the primary hypothesis
of the study and led to further studies that incorporated
biological monitoring.
The 1995 studies included housedust sampling in 76
homes and collection of urine samples from 109 children
child urine sampling ( Lu et al., 2000). An initial report
of this study focused on 48 applicator families and 14
reference families and presented DMTP concentrations
for children in these households ( Loewenherz et al.,
1997) . Azinphos methyl and phosmet were combined to
Figure 4. Comparison of median OP pesticide concentrations (ng/g)
in housedust and outdoor soil. Samples were collected from homes of
48 agricultural families living in an agricultural community in central
Washington state. Data are from Simcox et al. (1995).
Figure 5. Comparison of median OP pesticide concentrations (ng/g)
in housedust between 48 agricultural and 14 reference families in an
agricultural community in central Washington state. Data are from
Simcox et al. (1995).
Figure 6. Median dimethyl OP pesticide concentrations (ng/g) in
housedust of 62 agricultural and 14 reference families; grouped by
proximity to pesticide-treated farmland. Data are from Lu et al.
(2000).
Figure 7. Median dimethyl OP pesticide metabolite levels (g/ml) in
urine of 62 agricultural and 14 reference children; grouped by
proximity to pesticide-treated farmland. Data are from Lu et al.
(2000).
Strategies for assessing children's pesticide exposures Fenske et al.
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(6) Part 2 667

produce a single dimethyl OP pesticide metric for
evaluation of housedust concentrations. Figure 6 indicates
a fivefold difference in housedust concentrations for
agricultural and reference families. This graph also
illustrates the association of housedust concentrations
with proximity to farmland (Lu et al., 2000 ). Figure 7
displays median values for OP urinary metabolites ( sum
of DMTP and DMDTP ). The patterns for metabolite
concentrations were very similar to those for housedust
concentrations. The difference between children from
agricultural and reference families was about four- to
fivefold, and metabolite concentration decreased with
increasing distance from farmland ( Lu et al., 2000) .
Although not readily discernible in Figures 6 and 7,
concentrations for the agricultural population living more
than 1 /4 mile from farmland were higher than those of
the reference population for both housedust and urine
samples, indicating a contribution from parental take-
home exposure.
The 1998 studies included biweekly urine sample
collection from 44 Wenatchee children for 1 year, cross-
sectional biomonitoring studies among 96 children in two
Seattle metropolitan area communities, and a pilot
multipathway exposure analysis in 13 homes. Preliminary
results demonstrate that OP pesticide exposure continues
year round for most children in the agricultural commu-
nity, and that there is significant temporal variation in
metabolite concentrations; i.e., levels were higher during
the spring spraying season. We have also compared the
DMTP concentrations reported in our 1995 studies
(Loewenherz et al., 1997 ) with those measured in
Seattle (Figure 8 ) and found that concentrations from
Seattle children appear to be similar to those of the
Wenatchee reference population.
The studies reported here have also spawned an
attempt to estimate children's doses to OP pesticides.
Current risk assessment methods in regulatory agencies
such as EPA do not often use data from biological
monitoring. Instead, environmental concentration data are
linked to behavioral data and route -specific absorption
factors to produce estimates of internal dose (U.S.
Environmental Protection Agency, 1998). Biologically
based dose estimates can be generated from urinary
metabolite data if several assumptions related to pharma-
cokinetics are adopted. Our initial attempt to generate
such estimates indicated that a substantial fraction of the
children in this study exceeds current EPA reference
doses (RfDs ) and World Health Organization acceptable
daily intake (ADI) values (Fenske et al., 2000 ). We have
also used the soil and housedust data in traditional
exposure models to predict urinary metabolite levels and
have found that these models underpredict the concentra-
tions observed in these children ( Kissel et al., 1999) .
Discussion
Study Design Considerations
Field study design provides the framework for sampling
strategies, data analysis, and hypothesis testing. As in
epidemiologic studies, field studies must adapt to
existing characteristics of the study population and
environment. In the case of pesticide exposures in
agricultural communities, study designs must be modified
to account for regional pesticide use patterns, workforce
and labor hiring patterns, seasonal factors ( e.g., weather,
length of growing season ), and population characteristics
(e.g., accessibility via telephone, language, parenting
practices).
Several design features in these studies proved helpful
in drawing meaningful conclusions from our data. First,
the inclusion of a well -defined reference population
allowed useful comparisons across groups. Our two
criteria for the reference population Ð no adult house-
hold members working in field agricultural production
and residence distant from treated farmland Ð were tied
explicitly to the hypothesized exposure pathways: take -
home exposure and spray drift. Although sample sizes in
these studies were relatively small and exposure measure-
ments variable, we were still able to find significant
differences across these groups, indicating that these
pathways are important components of children's ex-
posures in agricultural communities. Second, the choice
of a ``reference'' population ( a comparison population
within the same rural community ) rather than a
traditional ``control'' population ( e.g., children from the
Seattle metropolitan area ) allowed examination of path-
ways specific to crop production rather than a broader
Figure 8. Comparison of median DMTP levels in urine (g/ml) of 96
children from the Seattle metropolitan area (children A), non-
agricultural (children B), and agricultural (children C) families living
in an agricultural community in central Washington state. Data for
children A are from Knutson (1998). Data for children B and C are
from Lu et al. (2000).
Fenske et al. Strategies for assessing children's pesticide exposures
668 Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(6) Part 2

urban±rural comparison. We have subsequently measured
urinary metabolites in urban environments, as indicated in
Figure 8.
Third, detailed documentation of residential proximity to
treated farmland allowed creation of four proximity
categories in these studies. Had we used only two distance
categories, our findings would have been less striking.
Proximity in these studies was determined by self-reports of
participants and observations by research staff. In current
studies, we are using global positioning system (GPS )
technology to document these distances. This more
objective measure of distance should decrease misclassifi-
cation. Still, an accurate definition of ``treated farmland''
can require extensive investigation. Farm workers may not
know the spraying schedules on nearby property, so it may
be necessary to interview farm owners or local cooperative
extension agents to document pesticide use patterns.
Environmental Concentration Measurements
Soil measurements provided little information on potential
OP pesticide exposures for children in these environments,
but housedust measurements proved to serve as an index of
indoor environmental contamination of agricultural pesti-
cides. The findings of these studies confirm the theory that
relatively non-persistent chemicals such as the OP
pesticides can be stable in residences. Since the children
in these studies also spent most of their waking hours
indoors, the pesticides in housedust also represent a
potential exposure source. We were unable, however, to
demonstrate a strong relationship between housedust
concentrations and biological levels in this population.
This finding is probably due to several factors. First, the
complexity inherent in children's exposures Ð intermittent
contact with surfaces, variable hand- to-mouth behaviors
Ð would tend to decrease an association of this kind.
Also, there is relatively high variability in the biologic
measures employed (spot urine sampling). Finally, the
instrument used for dust sampling ( HVS- 3) tends to
remove particles from deep carpet as well as from surfaces.
The concentrations found may not represent chemical
available to children during normal residential activity. On
the other hand, the HVS -3 does provide a systematic and
reproducible method for dust collection. Had we simply
collected vacuum cleaner bags from families, as has been
proposed recently for some studies, it is unlikely that we
would have observed a strong association between
concentrations and distance from treated farmland. House-
dust sampling methods should conform to standard
scientific criteria such as accuracy, precision, and systema-
tic, protocol -based procedures.
Biological Measurements
The use of the generic dialkyl phosphate metabolites for
children's exposure has both strengths and limitations.
This method captures most OP pesticide exposures and
allows source and pathway identification for dimethyl
and diethyl compounds. This distinction can be very
useful if it is known, for example, that a family uses
diethyl chemicals for residential pest control. Also, it is
important to recognize that unique metabolites have not
been identified for most OP pesticides, so a multi-
metabolite screen (analogous to a multiresidue analysis
in food ) is not currently feasible. The generic assay
overcomes this problem.
The primary limitation of this method, of course, is a lack
of specificity. New compound -specific assays have been
developed by CDC researchers for such OP pesticides as
malathion and diazinon, and samples from our current
studies will be analyzed for both unique metabolites and
dialkyl phosphate metabolites. Biological measures in
populations such as the one discussed in this paper can
provide important baseline information for evaluating the
impact of new policies associated with pesticide use
reduction.
Recent Studies of OP Pesticide Exposure
Two papers have been published recently that have
examined children's exposure to pesticides in Latin
America. The first, an ecological study of Yaqui Indian
children in Mexico, found physiological and neurological
deficits in preschool children who were presumably
exposed to pesticides (Guillette et al., 1998) . The study
design did not include measurements of pesticide concen-
trations in the children's environment, nor did it include
biological monitoring, so the attribution of the observed
effects to pesticides remains speculative. A carefully
conducted exposure assessment in conjunction with health
outcome testing in this population would clarify the etiology
of these effects.
The second study, conducted in rural El Salvador,
evaluated OP pesticide exposure in children 8 ±17 years
old (Azaroff, 1999). The study employed a dialkyl
phosphate assay that yielded qualitative data, resulting in
a classification scheme of no, low, or high metabolite
concentrations. These categorical data were compared to
self- reported occupational activities and residential pesti-
cide use through logistic regression analysis. The study
found a significant association between adult family
member and children OP pesticide metabolite concentra-
tions, but the other statistical analysis were confounded by
the pooling of adult and child data. These data were clearly
not independent, and so should have been separated for
these analyses.
Ongoing Studies of OP Pesticide Exposure
Several other OP pesticide exposure studies have been
initiated that include children in agricultural communities
(Table 3 ). The most common target pesticides are
Strategies for assessing children's pesticide exposures Fenske et al.
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(6) Part 2 669

chlorpyrifos and diazinon, although in some cases, OP
pesticides as a class are under investigation, and a dialkyl
phosphate metabolite method is proposed. Several of these
studies hope to link biologic measures to near-term
outcomes such as cholinesterase inhibition. It remains to
be seen whether the exposure levels currently detected in
child populations are sufficiently high to produce measur-
able changes in such enzymes or in other biomarkers of
effect. Perhaps the most ambitious among these studies is
the University of California's prospective evaluation of a
group of women and their offspring in an agricultural
community. Women are being enrolled at time of
pregnancy and their children followed through age 2
(Eskenazi et al., 1999 ).
The UW studies reviewed here have provided some new
insights into the extent of OP pesticide exposure in
agricultural communities. It seems clear that children who
live near treated farmland or with parents working in
agriculture can have higher exposures than other children
in the same community, and therefore constitute a high -
risk population based on the definition provided at the
outset of this paper. The two additional exposure pathways
for these agricultural children Ð farmland proximity and
parental take- home Ð are the focus of evaluation and
intervention studies within the UW program. The relation-
ship between pesticide drift and children's exposure is
being assessed with novel technologies. Light detection
and ranging ( LIDAR) technology is being adapted to
characterize both sources and downwind features of spray
drift. Data from this system should provide a spatial ±
temporal exposure profile that can be linked with child
activity. Child activity, in turn, will be monitored through
adaptation of existing GPS technology, documenting the
movement of children over time. Biological monitoring
will be used to evaluate the accuracy of exposure models
derived from these technologies.
In the case of parental transfer of pesticides from
workplace to home, the UW program is working with
researchers at the Fred Hutchinson Cancer Research Center
to develop a community -based intervention to break the
take- home exposure pathway. Twenty- four communities in
the lower Yakima Valley of Washington state have been
enrolled in this study. Adult and child urine samples, as well
as housedust and vehicle dust samples, were collected from
about 200 households in 1999 to provide baseline data for
the intervention. Community-based intervention activities
are being conducted for 2 years, followed by another round
of biological and environmental sampling. This study
design provides the opportunity to evaluate the effectiveness
of the intervention with objective measures of exposure.
Further studies that employ both environmental sampling
and biological monitoring as measures of OP pesticide
exposure are needed to characterize potential health risks for
children in agricultural communities. Improved cost-
effective methods for the analysis of environmental and
biological samples are an important component of such
Table 3. Current biological monitoring studies of children's OP pesticide exposure studies in the US.
Institution Location Population Target OP pesticides Data available
(years)
University of Minnesota Minneapolis ± St. Paul
and rural counties
Urban ± rural comparison of
3 ± 12-year-old children
malathion, chlorpyrifos 1 ± 2
University of Minnesota Minneapolis SES comparison of K-5
school children
malathion, chlorpyrifos 2 ± 3
University of Arizona Yuma County, AZ Farm worker children Chlorpyrifos, diazinon 1 ± 2
University of Washington Chelan-Douglas and
King Counties, WA
Clinic-based urban ± rural
comparison
OP pesticides (dialkyl phosphates) 1 ± 2
University of Washington Yakima County, WA Children in proximity to
farmland
OP pesticides (dialkyl phosphates) 3 ± 5
University of Washington Yakima County, WA Farm worker children in
28 communities
OP pesticides (dialkyl phosphates) 3 ± 5
University of California,
Berkeley
Monterey County, CA Clinic-based enrollment during
prenatal care
OP pesticides (dialkyl phosphates) 3 ± 5
Oregon Health Sciences
University
Oregon Farm worker children OP pesticides (dialkyl phosphates) 3 ± 5
National Cancer Institute Iowa, North Carolina Subpopulation of farm families
from Ag Health Study
Extensive pesticide screen 3 ±5
ATSDR Six states Intervention design+cohort
investigation (children)
Methyl parathion 3 ± 5
ATSDR Texas border area Epidemiological study of
neural tube defects
Methyl parathion, chlorpyrifos, malathion 3 ± 5
ATSDR US ± Mexico border Comparison of high and low
potential exposure (children)
OP pesticides (dialkyl phosphates) 3 ± 5
EPA/ORD California and North
Carolina
Clinic-based high-risk pediatric
population
OP pesticides (dialkyl phosphates) 3 ± 5
Fenske et al. Strategies for assessing children's pesticide exposures
670 Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(6) Part 2

studies. Risk assessments aimed at children and pesticides
need to account for the additional pathways that exist in
agricultural communities. One or more well-defined
prospective community- based studies of children and
pesticides would provide the most meaningful new
information in this area.
Acknowledgments
This work was supported by funding from the U.S.
Environmental Protection Agency (cooperative agreement
no. R819186 -01) , the Association of Schools of Public
Health ( cooperative agreement no. S147-14/16 ), the U.S.
Environmental Protection Agency STAR grant program
(grant no. 916001537 ), and the National Institute for
Occupational Safety and Health (Pacific Northwest Agri-
cultural Safety and Health Center cooperative agreement no.
U07/ CCU012926 ). The authors thank Kathy Yuknavage,
I- Chwen Lee, Garland Bellamy, and David Kalman for
their assistance in this work.
References
Aprea C., Sciarra G., and Lunghini L. Analytical method for the
determination of alkyl phosphates in subjects occupationally exposed
to organophosphorus pesticides and in the general population. J Anal
Toxicol 1996: 20: 559 ± 563.
Ashford N.A., Spadafor C.J., Hattis D.B., and Caldart C.C. Monitoring the
Worker for Exposure and Disease: Scientific, Legal, and Ethical
Considerations in the Use of Biomarkers. The Johns Hopkins
University Press, Baltimore, MD, 1990.
Azaroff L.S. Biomarkers of exposure to organophosphorous insecticides
among farmers' families in rural El Salvador: factors associated with
exposure. Environ Res 1999: 80: 138 ± 147.
Callahan M.A., Clickner R.P., Whitmore R.W., Kalton G., and Sexton K.
Overview of important design issues for a National Human Exposure
Assessment Survey. J Expos Anal Environ Epidemiol 1995: 5 ( 3 ) :
257 ± 282.
Eskenazi B., Bradman A., and Castorina R. Exposures of children to
organophosphate pesticides and their potential adverse health effects.
Environ Health Perspect 1999: 107: 409 ± 419.
Fenske R.A., Kissel J.C., Lu C., Kalman D.A., Simcox N.J., Allen E.H.,
and Keifer M.C. Biologically based analysis of organophosphorus
pesticide exposures in children in agricultural communities. Environ
Health Perspect 2000: 108: 515 ± 520.
Fox R.D., Reichard D.L., Brazee R.D., Krause C.R., and Hall F.R.
Downwind residues from spraying a semi - dwarf apple orchard. Trans
ASAE 1993: 36: 330 ± 334.
Guillette E.A., Meza M.M., Aquilar M.G., Soto A.D., and Garcia I.E. An
anthropological approach to the evaluation of preschool children
exposed to pesticides in Mexico. Environ Health Perspect 1998: 106:
347 ± 353.
International Life Sciences Institute. A Framework for Cumulative Risk
Assessment. ILSI Risk Science Institute, Washington, DC, 1999.
Kissel J., Kedan G., Wawrukiewicz A., and Fenske R.A. Attempted
reconciliation of predicted children's exposure to organophosphate
pesticides with measured metabolite levels in urine. Epidemiology
1999: 10 ( suppl) : S110 ( Abstract 305 O ) .
Loewenherz C., Fenske R.A., Simcox N.J., Bellamy G., and Kalman D.A.
Biological monitoring of organophosphorus pesticide exposure among
children of agricultural workers in central Washington State. Environ
Health Perspect 1997: 105: 1344 ± 1353.
Lu C., Fenske R.A., Simcox N.J., and Kalman D. Pesticide exposure of
children in agricultural community: evidence of household proximity to
farmland and take home exposure pathways. Environ Res 2000: 84 (3).
Mileson B.E., Chambers J.E., Chen W.L., et al. Common mechanisms of
toxicity: a case study of organophosphorus pesticides. Toxicol Sci
1998: 41: 8 ± 20.
Moate T., Lu C., Fenske R.A., Hahne R., and Kalman D.A. Improved
cleanup and determination of dialkyl phosphates in the urine of
children exposed to organophosphorus insecticides. J Anal Toxicol
1999: 23: 230 ± 236.
Shafik T., Bradway D.E., Enos H.F., and Uyobs A.R. Human exposure to
organophosphorus pesticides: a modified procedure for the gas ± liquid
chromatographic analysis of alkyl phosphate metabolites in urine. J
Agric Food Chem 1973: 21: 625 ± 629.
Simcox N.J., Fenske R.A., Wolz S.A., Lee I.C., and Kalman D.A.
Pesticides in household dust and soil: exposure pathways for children
of agricultural families. Environ Health Perspect 1995: 103: 1126±
1134.
U.S. Environmental Protection Agency. 1996 Food Quality Protection Act
Implementation Plan. Office of Prevention, Pesticides and Toxic
Substances, Washington, DC, 1997.
U.S. Environmental Protection Agency. Hazard assessment of the
organophosphates. Report of the Hazard Identification Assessment
Review Committee, Health Effects Division, Office of Pesticide
Programs, Washington, DC, 1998.
U.S. Environmental Protection Agency. Implementing the Food Quality
Protection Act: Progress Report. EPA 735 - R - 99001. Office of
Prevention, Pesticides and Toxic Substances, Washington, DC, 1999.
Strategies for assessing children's pesticide exposures Fenske et al.
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(6) Part 2 671