Content uploaded by Dana B Barr
Author content
All content in this area was uploaded by Dana B Barr on Jan 10, 2014
Content may be subject to copyright.
Nondietary ingestion of pesticides by children in an agricultural
community on the US/Mexico border: Preliminary results
STUART L. SHALAT,
a
KIRBY C. DONNELLY,
b
NATALIE C.G. FREEMAN,
a
JAMES A. CALVIN,
b
SOWMYA RAMESH,
b
MARTA JIMENEZ,
a
KATHLEEN BLACK,
a
CATRIONA COUTINHO,
b
LARRY L. NEEDHA M,
c
DANA B. BARR
c
AND JUAN RAMIREZ
d
a
Environmental and Occupational Health Sciences Institute (a jointly sponsored institute of Rutgers the State University of New Jersey and the University of
Medicine and Dentistry of New Jersey ), Robert Wood Johnson Medical School, Piscataway, New Jersey, USA
b
Center for Environmental and Rural Health, Texas A&M University, College Station, Texas, USA
c
National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, Georgia, USA
d
TDI -Brooks International, College Station, Texas, USA
An environmental measurement and correlation study of nondietary ingestion of pesticides was carried out in a colonia in south Texas. The purpose of the study
was to evaluate young children’s exposure to environmental levels of organophosphate ( OP ) pesticides in the household. Samples were collected to measure
levels of OP pesticides in housedust and on children’s hands. These, in turn, were compared to levels of OP pesticide metabolites in urine. A total of 52
children, 25 boys and 27 girls, participated in the spring and summer of 2000. The children were 7 – 53 months of age at the time of recruitment. Univariate and
multivariate regression analyses were carried out using SAS statistical software. Seventy - six percent of housedust samples and 50% of hand rinse samples
contained OP pesticides. All urine samples had at least one metabolite and over 95% had at least two metabolites above the limit of detection ( LOD). Total OP
loadings in the housedust ranged from nondetectable ( nd ) to 78.03 nmol / 100 cm
2
( mean=0.15 nmol / 100 cm
2
; median = 0.07 nmol / 100 cm
2
); total OP
loadings on the children’s hands ranged from nd to 13.40 nmol /100 cm
2
( mean=1.21 nmol / 100 cm
2
; median=1.41 nmol / 100 cm
2
), and creatinine corrected
urinary levels ( nmol / mol creatinine ) of total OP metabolites ranged from 3.2 to 257 nmol / mol creatinine ( mean = 42.6; median 27.4 nmol/mol creatinine).
Urinary metabolites were inversely associated with the age of the child ( in months ) with the parameter estimate ( pe ) = 2.11, P =0.0070, and 95% confidence
interval 3.60 to 0.61. The multivariate analysis observed a weak association between concentrations of OP pesticides in housedust, loadings in housedust,
and concentration on hands, hand surface area, and urinary levels of OP metabolites. However, hand loadings of OP pesticides were more strongly associated
( r
2
= 0.28; P= 0.0156 ) with urinary levels of OP metabolites ( pe = 6.39; 95% CI 0.98 – 11.80). This study’s preliminary findings suggest that surface loadings
of pesticides, on hands, are more highly correlated with urinary bioassays and, therefore, may be more useful for estimation of exposure in epidemiologic
studies than levels of pesticides in housedust.
Journal of Exposure Analysis and Environmental Epidemiology ( 2003 ) 13, 42 – 50 doi:10.1038/sj.jea.7500249
Keywords: border, children, exposure, Hispanic, organophosphate pesticides
.
Introduction
In agricultural comm unities, the potential for exposure to
pesticides is greater than for the general population (Shalat
et al., 2002). Numerous studies have shown that children
are exposed to environmental chemicals, particularly pes-
ticides, through different mechanisms and often in greater
quantities than adults ( Fenske et al., 1990, 2000; Simcox
et al., 1995; Bradman et al., 1997; Loewenherz et al., 1997;
Lu and Fenske, 1999; Lu et al., 2000 ). The question of how
vulnerable children are to the potential toxic effects of
pesticides is clearly dependent upon the dose they receive.
A major route of exposure that has long been appreciated is
nondietary ingestion of pesticides through the mouthing
behaviors of children. These exposures result from hand
contact with floors, surfaces, and soil contaminated with
pesticides and the subsequent ingestion, when hands
inevitably end up in the child’s mouth. These activities
may also result in secondary contamination of food items
with p esticides that have gotten on children’s hands and/or
food handling or preparation areas in the home. In addition,
they occur when children’s toys or other objects become
contaminated with pesticides and the child puts those
objects in their mouth. Regular use of pesticides to control
household pests is widespread ( Savage et al., 1981; Davis
et al., 1992). The US EPA Nonoccupational Pesticide
Exposure Study found elevated levels of pesticide residues,
from household use, in nearly all houses sampled ( White-
more et al., 1994). Estimates of soil ingestion by children
1. Address all correspondence to: Dr. Stuart L. Shalat, EOHSI, 170 Fre-
linghuysen Road, Piscataway, NJ 08854, USA. Tel.: + 1 - 732 - 445 - 1295.
Fax: + 1 - 732 - 445 - 0116. E - mail: shalat@eohsi.rutgers.edu
Received 19 August 2002.
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13, 42 – 50
# 2003 Nature Publishing Group All rights reserved 1053-4245/03/$25.00
www.nature.com/jea
range from 40 to 100 mg/day ( van Wijnen et al., 1990 ).
While dust ingestion has not been adequately assessed, it is
assumed that most soil and dust ingestion will come directly
from hand - to - mouth activities or object-to-mouth activ-
ities. The greatest freque ncy is assumed to occur among
toddlers, slightly lower in infants, and the least among school
age children. However, detailed models are scarce and data
to validate them are limited.
Recent studies have examined in detail the exposure and
uptake of pesticides in children who reside in agricultural
communities (Loewenherz et al., 1997; Fenske et al., 2000;
Lu et al., 2000; O’Rourke et al., 2000; Sumner and Langley,
2000; McCauley et al., 2001). One of the findings of these
studies is that children who live in agricultural communities
had five times higher pesticide metabolites in their urine
than children who resided in nonagricultural communities
(Lu et al., 2000 ). In particular, they examined children who
resided near orchards and found both environmental as well
as urinary levels of pesticides to be statistically significantly
higher in these children. When compared to EPA chronic
dietary reference doses, 56% of those whose parents worked
in agricultural settings exceeded these levels (Fenske et al.,
2000).
Biological markers have been previously used to assess
exposure to pesticides. Urine samples have been used to
examine exposure to a suite of pesticides, through measure-
ment of pesticides or their metabolites (Griffith and Duncan,
1985; McCurdy et al., 1994; Hill et al., 1995; Fenske et al.,
2000; O’Rourke et al., 2000). In the most recent NHANES
III report, 82% of individuals had measurable (>1 g/l)
trichloropyridinal (TCPY), the primary metabolite of
chlorpyrifos, in their urine ( Hill et al., 1995 ). Urinary
pesticide metabolites have been found to correlate well with
erythrocyte acetyl cholinesterase ( McCurdy et al., 1994).
However, elevated levels of pesticide metabolites have been
found in urine following organophosphate (OP) pesticide
exposure, even when cholinesterase inhibition was not
found in blood serum (Richter et al., 1992 ).
A study of a border population in Yuma county, AZ,
observed OP pesticide metabolites in the urine of 33% of
children 6 years of age or under (O’Rourke et al., 2000).
Another study observed a very large and statistically
significant difference between levels of exposure to OP
pesticides in children of agricultural families and a
nonagricultural comparison group (Lu et al., 2000 ). That
study examined soil, housedust, and urine samples. While
not surpr ising that children in these communities were
highly exposed, the fact that 67% of the urine samples from
the children had detectable levels (limit of detection, LOD,
1–10 g/l) of dialkyl phosphates should be viewed as
alarming. Mean and median levels for these metabolites
were 20 and 5.0 g/ l, respectively. When compa ring the
percentage of children with detectable levels of pesticide, it
is of course important to take into account the study’s level
of detection. The level of detection in the O’Rourke et al.
(2000) study was 25 g / l, potentially explaining the
difference in the proportion in these two studies.
The purpose of this study was to examine an important
aspect of the relationship between environmen tal exposures
to OP pesticides in infants and young children who live in
border agricultural communities ( colonias) and dose levels
of OP pesticides as measured by six of their metabolites in
urine. Environmental media ( soil and dust ) were utilized to
characterize the environmental exposures. The removable
quantities of pesticides on floor surfaces and children’s
hands, known as loadings, were examined as well. By
providing an alternative model for estimating pesticide
exposure, a better understanding may b e achieved of the
possible risks pesticides may represent to children’s health.
Methods
A 3 - yea r environmental measurement and correlation study
was conducted in the mid-Rio Grande Valley. Public
meetings were held with several community groups, in
order to identify communities with which to partner this
study. Because of their history of activities in this
community and their willingness to take an active role in
the planning and conducting of the study, the Sisters of
Mercy, a religious order, was the primary community study
partner. The sisters make broad use of promotoras for lay
health education in the community. For this reason, we
chose to train the promotoras and employ them for
contacting households, and administering questionnaires
and sample collection.
As a result of the community meet ings, a small colonia of
approximately 5000 residents, located on the US /Mexico
border, was selected for the study. More than 98% of the
residents are Mexican/American. A census was conducted
utilizing the promotoras, to identify all households having
children between 6 and 48 mont hs of age. Households were
selected for possible participa tion on the basis of the census
results. A total of 920 residences were identified; of these,
870 were occupied at the time of the census (as determined
by direct observation and discussions with neighbors ). Six
hundred forty-three of these were contacted, in-person, by
study personnel, with 91 households having at least one
child under 3 years of age. Initial ly, only homes with two or
more children were invited to participate.
The study protocol, questionnaires, and letter of consent
were all reviewed and approved by UMDNJ-RWJMS
Institutional Review Board (IRB no. 2708). The interviews
and sampling were timed to coincide with one of the two
growing seasons in the area. The initial round began in the
spring of 2000. Households were contacted and invited to
participate in the study. An appointment was scheduled for
the administration of the questionnaire, collection of
Pesticides and children on the US/Mexico border Shalat et al.
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(1) 43
environmental samples, and videotaping of the child. This
was carried out on 1 day and urine samples were collected
the following morning. Every attempt was made to
videotape and obtain hand rinses of children residing in
the same household on consecutive days. After a complete
description of the study was provided, those who agreed to
participate were asked to sign a letter of informed consent.
To those parents who agreed to participate, a baseline
questionnaire, which included medical and occupational
information, time /activity questions, questions on hand-
to- mouth activities, diet, and residential pesticide use, was
administered in Spanish — the primary language of the
residents of the commu nity.
Because of the complexity of sources and conditions that
effect environmental exposure to pesticides, and to better
characterize the residential exposures of children from this
agricultural area, four exposure metrics were used in this
study. The metrics included: (1) pesticide concentration in
soil; ( 2 ) pesticides in housedust; (3 ) pesticides on children’s
hands; and ( 4) urinary levels of pesticide metabolites.
Metrics 1, 2, and 3 were assessed in combination with direct
observation of hand/mouth activity and parental question-
naire and compared to actual bioassay measurements in
metric 4. Environmental samples of housedust were taken
from the floor inside the home, near the main en trance,
which in most instances was also a major play area for the
children. Dust was collected by a wipe sample from a square
meter of floor. The samples were wrapped in aluminum foil,
then placed into a polyethylene coll ection bag, and shipped
to the laboratory for analysis. Hand rinses utilizing 225 ml
of isopropyl alcohol were obtained from the child’s hands.
Tracings of the child’s hands were also obtained in order to
provide an estimate of surface area. This was used in
concert with the pesticide levels measured in the alcohol
rinses to compute pesticide loading on the child’s hands.
Urine samples were used to assess exposure to pesticides
through the measurement of pesticide metabolites.
Soil samples were collected from the front yard of one
house from each block of the neighborhood. This was done
in order to reflect levels of pesticide drift from the nearby
farm fields. One sample was collected from each of these
houses. Surface soils were collected with a precleaned
aluminum foil-wrapped, stainless steel trow el. The area to
be sampled was first cleaned of any surface debri s or
vegetation. The trowel was then used to remove soil to a
depth of approxi mately 4–15 cm. The soil was placed in a
precleaned I- Chem
1
jar (I -Chem, New Castle, DE ) with
a Teflon-lined lid, labeled, and stored in a ziplock bag.
The housedust floor sample was obtained as close as
possible to the front door, on tile or linoleum. One sample
was collected from each house participating in the study.
The area that was sampled was marked and was
approximately 1 m
2
. The exact area sampled was measured
and recorded in the logbook. While wearing gloves, the
technician removed a prepared glass fiber filter cloth out of
the aluminum foil. The cloth was then wet with 30 ml of
pesticide grade isopropyl alcohol. The area of the floor to be
sampled was then cleaned with the cloth in the following
way: starting at the top of the area and going from left to
right, the cloth was passed in one straight line across the
very top of the area to be cleaned. Next, going from right to
left, the cloth was passed in one straight line back across the
area, directly below the first swipe. This process was
continued until the entire area was cleaned. The fiber cloth
was then folded and returned to the aluminum foil.
Prior to the commencement of videotaping, the child’s
hands were rinsed as described below and the rinse
discarded. The child was then videotaped by a technician
while the child resumed his/ her normal activities over the
subsequent 4 h. One hand rinse was collected for analysis
after the completion of the videotaping. Hand rinse samples
were collected by washing the child’s hands in 225 ml of
reagent-grade isopropanol. The hands were initially rinsed
in 150 ml of isopropanol in a clean ziploc bag. This was
carried out by having the child place both hands in the bag
and having the trained technician gently agitate the bag from
the outside for approximately 30 s. After collection, the
handwash sample was transferred to a clean pesticide -grade
I- Chem
1
250-ml amber glass jar. The ziploc bag was then
washed one time with the remaining 75 ml of isopropanol to
remove residues from the ziploc bag. The rinse was then
added to the sample in the amber jar. The total volume of the
isopropanol used for each rinse was 225 ml.
Pesticide extraction and analysis of all housedust,
hand rinses, and soils were performed at a laboratory
(TDI -Brooks International) located in College Station, TX.
TDI-Brooks International has served as a contract labora-
tory on trace organics and metals analysis for the US Fish
and Wildlife Services, Texas A&M University, and various
government agencies, and has consistently been one of the
top performers on the annual NIST / NOAA/ EPA trace
organic intercalibration exercises. Analysis was carried out
for the presence of OP pesticides. The following pesticides
were selected for analysis based upon information obtained
from the local office of the Texas Agricultural Extension
Service (TAEX): azinphos -methyl, chlorpyrifos, demoton
O, demoton S, diazinon, disulfoton, ethion, fenithrothion,
fonofos, malathion, ethyl parathion, and methyl parat hion.
Extraction of dust filters employed an automated
extraction apparatus, Dionex ASE 200 Accelerated Solvent
Extractor (Dion ex, Sunnyvale, CA) ( Richter et al., 1994,
2001; EPA, 1997; Schantz et al., 1997; Ezzel, 1998;
Zuloaga et al., 2000 ). This was used to extract organic
analytes from 15 g of dried sediment and dust filter samples.
Triphenyl phosphate was used as a surrogate and was added
to the samples before extraction and used to assess the
extraction and concentration efficiency of the procedure.
The surrogate compound was resolved from — but eluted
Shalat et al. Pesticides and children on the US/Mexico border
44 Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(1)
in close proximity to — the analytes of interest. The
extractions were performed with dichlorom ethane:acetone
(50:50) inside stainless steel extraction cells held at
elevated temperature ( 1008 C ) and solvent pressure ( 2000
psi). The analytes dissolved in the solvent were transferred
from the heated extraction cells to a 60 - ml glass collection
vial. Extracts were then concentrated to a final volume of
1 ml ( hexane:acetone 5 0:50 ) using an evaporative solvent
reduction apparatus (Zymark TurboVap II; Zymark, Hop-
kinton, MA) (Patnaik, 1997; Wade et al., 1986; EPA, 199 7a;
Zuloaga et al., 2000; Richter et al., 2001; Schantz et al.,
1997; Richter et al., 1994; Ezzel, 1998).
Extraction of hand rinses was carried out as follows. The
surrogate compound (triphenyl phosphate) was added to the
dermal rinse samples (collected in 100% isopropyl alcohol)
prior to concentration and was used to assess the efficiency
of the procedure. The surrogate compound was resolved
from — but eluted in close proximity to — the analytes of
interest. Extracts were concentrated to a final volume of
1 ml of hexane:acetone (50:50) using the Zymark TurboVap
II (Zymark ) set at an initial temperature of 708C.
A capillary gas chromatographic method was used to
determine the concentration of OP compounds. Fused silica,
open-t ubular columns were used in this method as they
offered improved resolution, selec tivity, increased sensitiv-
ity, and decreased time over packed columns. Quantification
of pesticides was carried out with a HP5890 gas chro-
matograph (GC; Hewle tt - Packard, Palo Alto, CA) with a
nitrogen phosphorous detector ( GC - NPD ) (Richter et al.,
1994, 2001; EPA, 1997 ). The internal standard solution of
1-bromo- 2-nitrobenzene was added to the sample extracts
prior to instrument analysis for the determination.
The GC- NPD was temperature-programmed and oper-
ated in the split with a DB-5 ( 30 m, 0.25 mm ID, and
0.25 m film thickness; J&W Scientific, Folsom, CA).
Carrier flow was by conventional pressure control. The
autosampler is capable of making 1– 5 l of injections.
Single high-resolution capillary colum n and NPD were
used. The data acquisition system was by HP Chemstation
software, capable of acquiring and processing GC data.
Instruments were calibrated using commercially purchased
standards. Calibration solutions for pesticides were prepared
at five concentrations, ranging from 0.5 to 10 g/ml, by
diluting a commerci ally available solution containing the
analytes of interest. A cali bration curve was established by
analyzing each of the five calibration standards (0.5, 1.0, 5.0,
8.0, and 10.0 g/ml), and fitting the data to a linear equation.
Urine collection is a relatively easy and noninvasive
method for biological monitoring. Urine samples were used
to examine exposure to a suite of OP pesticides through
measurement of nonspecific dialkyl phosphate metabolites
of the pesticide. A complete void was collected on the
morning following the collection of the housedust and hand
rinse samples, so as to better represent the exposure period.
Collection of urine samples from children who are not
toilet-trained has been considered problematic and, thus,
2-year -olds are typically not used in studies where urine
sampling is conducted. Because of difficulties in extracting
the urine from the gel formed in commercially available
disposable diapers, these are generally not considered useful
for urine collection. Some investigators have utilized cotton
gauze for this purpose (Hu et al., 2000). A similar
methodology was employed; however, in place of cotton
gauze, cotton terry cloth diaper inserts (Organic Diaper
Doublers; Ecobaby Organics, El Cajon, CA) were em-
ployed. These are larger, more absorbent, and, in tests we
conducted, exhibited no breakthrough with up to 500 ml of
liquid. Prior to use, each insert was washed once in a mild
detergent and rinsed five times, with the final rinse
containing vinegar. Diapers were dried following washing.
The inserts were provided to the parents along with a
disposable outer covering (water -proof laminate), pur-
chased from the same source. Children who were toilet-
trained had their urine sample collected either directly from
lined ‘‘potty chairs’’ or they were allowed to urinate directly
into a cup. The urine samples were collected the morning
after the hand rinse and placed in a plastic bag and stored on
ice. Diaper inserts were similarly stored and sent to our
laboratory at TAMU for extraction. Inser ts were extracted
with a needleless 20-ml syringe. Uri ne was recovered from
the diaper insert by placing the syringe firmly on the insert
and withdrawing the plunger. After collection, samples were
shipped, on dry ice, to the Centers for Disease Control and
Prevention ( CDC ) in Atlanta.
The urine samples were analyzed for the presence
metabolites of OP pesticides. All OPs have the same
general structure and mode of toxicity. They are composed
of a phosphate ( or phosphorothioate or phosphorodithioate )
moiety — which, in most cases, is O,O- dialkyl -substi-
tuted, where the alkyl groups are either dimethyl or diethyl
— and an organic group, which is specific to each pesticide.
For instance, chlorpyrifos is composed of an O,O- diethyl
phosphorothioate to which a 3,5,6-trichloropyridinyl group
is attached. After a given OP pesticide enters the body, it is
usually metabolized by enzymatic hydrolysis to form one or
more of six common dialkyl phosphate metabolites and the
more specific ‘‘organic’’ moiety. All of these metabolites are
excreted in urine. However, only 28 of 39 commercially
used OP pesticides metabolize to these six metabolites. At
the CDC, the common alkyl phosphate metabolites are
measured via codistillation of 4 ml of urine, chemical
derivatization of the metabolites to chloropropyl phosphate
esters, and analysis using gas chromatography tandem mass
spectrometry (GC-MS/MS) with quantification by the
isotope dilution technique ( Bravo et al., 2002). Matrix-
based pools were used for quality control.
The use of the stable isotope of each of these metabolites
allows for the highest degree of accuracy and precision. The
Pesticides and children on the US/Mexico border Shalat et al.
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(1) 45
LODs of the method are in the low to mid picograms-per-
milliliter range (parts per trillion ) with coefficients of
variation (CV ) of about 10 – 15% for most of the analytes.
These six metabolites are dimethylphosphate (DMP),
dimethylthiophosphate ( DMTP ), dimethyldithiophosphate
(DMDTP), diethylphosphate ( DEP ), diethylthiophosphate
(DETP), and diethyldithiophosph ate ( DEDTP). O,O -
dimethyl pesticides can metabolize to DMP, DMTP, and/
or DMDTP. For example, methyl malathion, which is an
O,O-dimethyl phosphorothioate, can metabolize to DMP
and DMTP. O,O -diethyl pesticides can metabolize to DEP,
DETP, and/or DMDTP. For example, chlorpyrifos can
metabolize to form DEP and DETP. The LOD and the
general information on metabolites analyzed at CDC are
presented in Table 1. All laboratory procedures were
performed in accordance with the Clinical Laboratory
Improvement Act of 19 88.
Statistical analysis was carried out on the results of the
environmental and bioassay sampling. Both univariate and
multivariate analyses were carried out by utilizing SAS
statistical software (version 8.0). For envir onmental
samples, the recorded value, even if less than the method
detection limit ( MDL), was used in the calculations. All
urine samples under the LOD were assumed to be 0.00. The
statistics computed included the mean, median, range,
distribution, and standard error of the environmental and
bioassay samples. Multiple linear regression analyses were
computed, as well as correlation matrices. This analysis was
used to examine the potential association between the
environmental metrics ( housedust and hand loading of
pesticides) and the urinary bioassay ( pesticide metabolites)
as the outcome variable. Total OP levels were compared for
these three metrics. Other variables included gen der, age
(months), and hand surface area.
Table 1. CDC reference OP pesticide analysis summary.
Analyte Parent compound( s ) Analytical LOD
a
Reference range
b,c
g/l (g/g)
DMP Any dimethyl OP ( e.g., methyl parathion,
azinphos methyl )
1.0 Median = 1.8 ( 1.5 ); 95th = 12 ( 18 )
DMTP Any dimethyl OP with one or two sulfurs
( e.g., methyl parathion, chlorpyrifos methyl )
1.0 Median = 4.8 ( 4.0 ); 95th = 67 ( 59 )
DMDTP Any dimethyl OP with two sulfurs
( e.g., azinphos methyl, dimethoate)
0.50 Median = 0.55 ( 0.39 ); 95th = 18 ( 18 )
DEP Any diethyl OP ( e.g., fonofos, chlopyrifos ) 1.0 Median = 4.5 ( 3.5 ); 95th = 25 ( 34 )
DETP Any diethyl OP with one or two sulfurs
( e.g., chlorpyrifos, parathion)
0.50 Median = 1.2 ( 1.1 ); 95th = 23 ( 23 )
DEDTP Any diethyl OP with two sulfurs
( e.g., fonofos, terbufos )
0.50 Median = nd ( 0.05 ); 95th = 3.1 ( 2.3 )
a
Calculated for each study. Results nearer to the LOD are subject to greater uncertainty. The LOD is determined for the entire measurement system, not just
the instrument.
b
CDC, unpublished data, NHANES III ( 1988 – 1994 ).
c
95th = 95th percentile; ND means not detected above the LOD; values are expressed in micrograms per liter urine and micrograms per gram creatinine in
parentheses.
Table 2. Results of the environmental sampling for OP pesticides ( nmol ).
Pesticide
Housedust ( n = 29 ) Hand rinses ( n =41)
Mean Median SD Mean Median SD
Demeton O 1.33 0.00 3.67 0.08 0.08 0.42
Demeton S 2.22 0.62 6.56 0.28 0.00 1.05
Fonofos 0.08 0.04 0.11 0.03 0.00 0.11
Diazinon 4.03 0.20 16.31 0.11 0.00 0.31
Disulfoton 0.17 0.11 0.24 0.10 0.00 0.14
Parathion methyl 0.49 0.15 0.64 0.50 0.00 1.66
Fenithrothion 0.36 0.11 0.77 0.01 0.00 0.06
Malathion 0.06 0.00 0.11 0.07 0.00 0.23
Chlorpyrifos 0.87 0.06 3.11 0.08 0.00 0.27
Parathion ethyl 0.00 0.00 0.00 0.13 0.00 0.36
Ethion 0.34 0.13 0.56 0.02 0.00 0.06
Azonphos methyl 0.94 0.16 1.61 0.11 0.00 0.34
Total OP ( levels ) 10.88 5.47 17.24 1.43 0.30 2.86
Total OP ( loadings ) ( nmol / 100 cm
2
) 0.15 0.07 0.23 1.33 0.27 2.69
Shalat et al. Pesticides and children on the US/Mexico border
46 Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(1)
Results
The total number of households enrolled was 29 (target 30)
and the total number of children enrolled was 52, composed
of 25 boys and 27 girls (target 30 boys and 30 girls). The
children were 7– 53 months of age at time of testing in the
spring/summer of 2000.
The fathers or adult male members of the household
worked in a variety of occupations: truck driver ( 13.3% ),
carpenter (13.3%), electrician (13.3%), construction work-
er (10%), and farmworker ( 13.3% ) were the most
common. None of the men worked directly in mixing or
loading of pesticides, with one driving farm equipment.
Only three women in the study households (10%) worked
outside the home, with one working in a farm -related
activity (packer).
Indoor use of pesticides within the last 6 months was
reported by 82.8% of the families. One - third of respondents
did not know what pesticide was used. The most frequently
reported pesticides were Raid (six ) and Green Light (four).
Also reported were Combat, Ray, Ray Max, Roachbox,
Baygon, Hotshot / Gis, Tat, D - Con, and Max (one each ).
Parents were asked where they used pesticides. Most of the
families reported using the pesticides in the kitchen. More
than half of the families reported using pesticides in other
rooms as well.
The most common areas in rooms where the pesticides
were used were the floor (48.3%), cupboards where dishes
were stored ( 41.4% ), and cabinets used for storage ( 31% ).
All applicati ons were performed by the resident, and none
done by professional exterminators. Most of the users
reported using pesticides three to six times in the last
6 months (61.1%). Most reported that they used pesticides
on an ‘‘as needed’’ basis (87%).
Outdoor use of pesticides was reported by 58.6% of the
families, although no specific products were reported.
Outdoor use was usually two to three times in the last
6 months (45.5% of users ). Treatment of lawns was
infrequently reported (10.3%). Only about 30% of the
homes has lawns. The majority of participants reported that
they purchased pesticides for their house and yard at the
supermarket (76% ). Pesticides were also purchased at
hardware stores ( 10.1% ) and farm supply stor es (10.1%).
We also examined whether children might be exposed to
pesticides through the pets or other residential animals.
Forty-two percent of families reported having at least one
dog. Flea collars were used by half of the families with dogs.
Other commonly owned animals were birds (25.9%) and
chickens ( 11.1%).
Analysis of housedust and ha nd rinses was conducted for
OP pesticides. The pesticides that were detected, either in
housedust or hand rinses were: azinphos-methyl, chlorpyr-
ifos, demoton O, demoton S, diazinon, ethion, fenithrothion,
ethyl parathion, and methyl parathion. Actual LODs for total
quantity of sample were between 1 and 10 pg. MDL for OPs
was 0.5 g/g for housedust and 0.5 g/ml for hand rinses.
The means, medians, and standard deviations for the
individual OP pesticides detected in the housedust and
hand rinses are presented in Table 2. A univariate statistical
analysis was carried out on the levels of pesticide in the
housedust, hand rinse, and urine samples. Approximately
three-quarters ( 76% ) of the housedust samples and half of
the hand rinse samples and none of the soil samples
contained OP pesticides. In addition to concentrations, OP
loadings were calculated for floor surfaces and children’s
hands. These were computed by multiplying the detected
concentration by the quantity of sample and dividing by the
measured surface area.
The results of the CDC analysis on six urinary
metabolites of OP pesticides for 41 urine samples are
presented in Table 3 and the distribution of levels of total OP
pesticide metabolites is presented in Figure 1. The most
commonly detected metabolites were DMP and DETP,
which were found in 100% and 95.7% of the subjects,
respectively. Additionally, DMP had the highest mean and
median levels of metabolite detected. This metabolite is
consistent with the pesticides methyl parathion and
azinphos-methyl, both of which were detected in the hand
rinses and housedust samples. Either methyl parathion or
Table 3. Results of OP metabolite in urine testing ( nmol / mol
creatinine corrected ); N =41.
Analyte Mean
a
Median
a
Range % Detects
DMP 20.0 8.8 nd – 143.5 95.7
DMTP 5.0 0.0 nd – 79.6 19.6
DMDTP 0.04 0.0 nd – 1.5 4.4
DEP 9.8 1.8 nd – 64.8 56.5
DETP 7.6 4.6 0.9 – 40.6 100.0
DEDTP 0.6 0.0 nd– 12.8 26.1
Total OPs 42.6 27.4 3.2– 257.0 100.0
nd = nondetect.
a
Nondetects were assumed as 0.00 for purposes of computing means and
medians.
Figure 1. Distribution of levels ( nmol / mol creatinine ) of urinary
metabolites for total OP pesticides in 41 children residing in Rio
Bravo, TX.
Pesticides and children on the US/Mexico border Shalat et al.
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(1) 47
azinphos methyl was detected in approximately 83% of the
dust samples and 24% of the hand rinses. Total OP loadings
in the housedust ranged from nondetectable (nd ) to 78.03
nmol/100 cm
2
(mean = 0.15 nmol/100 cm
2
; median = 0.07
nmol/100 cm
2
); total OP loadings on the children’s hands
ranged from nd to13.40 nmol/ 100 cm
2
(mean=1.21 nmol/
100 cm
2
; median = 1.41 nmol/100 cm
2
), and creatinine
corrected urinary levels (nmol/mol creatinine ) of total OP
metabolites ranged from 3.2 to 257 ( mean = 42.6; median
27.4 nmol / mol creatinine).
Multivariate linear regression analysis was carried out to
evaluate the association of pesticide levels and loadings of
total OPs in housedust and on hands with total urinary OP
metabolite levels. Age in months and gender ( male =1,
female=0) were also included in the model , as well as
hand surface area. The latter term was included as an
additional surrogate of physical development in addition to
age, as age and growth can be effected by premature birth.
Only those chil dren with all environmental variables were
included in the analysis. A total of 41 children were
included in the multivariate analyses ( 19 boys and 22
girls). Separate analyses were carried out using levels and
loadings. Models that included both were somewhat
unreliable in their parameter estimates because of the high
degree of colinearity between levels and loadings both for
housedust and hand rinses. For the former, the correlation
coefficient was 0.99 and for the latter 0.90. In the final
model ( Table 4), age was inversely associated with urinary
levels of OP metabolites parameter estimate, with total OP
levels in urine declining by 2.1 nmol / mol creatinine/
month of age (95% CI 3.60 to 0.61 ). Gender was not
statistically signifi cantly predictive for OP metabolite level.
The overall fit of the complete model with regard to
housedust and hand pesticide loadings just missed
statistical significance at the 0.05 level (r
2
=0.26;
P=0.0758 ). When housedu st was dropped from the model,
the fit improved (r
2
=0.28; P= 0.015 6 ). Individual terms
included in the model were hand surface area ( pe= 1.28,
95% CI 0.28 – 2.28) and the hand loading term (pe = 6.39,
95% CI 0.98– 11.80). Levels of correlation between
dependent variables did not exhibit a high degree of
colinearity. The highest correlation coefficient observed
was between age (months) and hand surface area
(r= 0.6581 ). Somewhat surprisingly, hand surface area
and hand loadings of OPs were not highly correlated at all
(r = 0.0300 ). Separate analys es utilizing transforms to
adjust the data to compensate for lack of normality in the
distribution of the dependent and independent variables did
not imp rove the fit of the model.
Discussion
Seventy-six percent of the homes sampled contained
detectable OP pesticides, while half the rinses from
children’s hands had detectable levels of OP pesticides.
The soil samples contained no OP pesticides above the
MDL. The absence of OP pesticides in soils was somewhat
surprising, but may be the result of the unusually warm
weather in the area durin g the first round of sampling in the
spring/summer of 2000. These hot, sunny conditions may
have increased the rate at which OP pesticides degraded
outdoors, while these same compounds may not have
degraded as readily in dust samples from within the
dwellings and, therefore, out of direct sunlight. At the same
time, detectable levels of OP metabolites were present in the
urine samples from all the children tested in the first round
of this study.
In the present study, detectable levels of OP metabolites
were observed in urine samples from all of the children
tested. However, only three out of four of the homes had
detectable levels of OPs in the housedust. It also is worth
noting that over 21% of the urine samples had OP
metabolite levels greater than 50 nmol /mol creatinine, with
9.8% over 100 nmol /mol creatinine. This suggests that the
cross-sectional environmental sampling conducted in this
study was not able to identify all of the children’s sources of
exposure. Even though many of the current OP compounds
are readily metabolized and excreted, a detailed under-
standing of the relationship between observed environ-
mental levels of pesticides and urinary metabolites clearly
requires more study.
Another interesting factor about the population in the
current study is that the level of OP pesticide metabolites
found in the urine samples of these children was relatively
high, given that these were not ‘‘farm’’ households. When
compared to NHANES (1999) data for the general
population, which are available for 6- to 59-year-olds,
the levels of the six metabolites in these infants and
young children ranged from approximately 3 1/2 to
almost 13 times that observed in NHANES ( CDC, 1992).
Since our findings suggest that younger children have
higher levels of OP metabolites in their urine than older
children, and the NHANES populatio n is considerably
older than ours, perhaps this finding should not be
considered surpr ising.
The preliminary data presented found virtually no
correlation between loadings of OPs in housedust on floor
surfaces and levels of OP metabolites measured in urines.
Table 4. Multivariate regression analysis — final model.
Parameter Estimate Pr>|t| 95% Confidence limits
Age ( months ) 2.11 0.0070 3.60 to 0.61
Gender ( male = 1,
female = 0 )
14.82 0.3101 44.02 to 14.38
Hand area ( cm
2
) 1.27 0.49 0.28 to 2.28
Hand load ( ng / cm
2
) 6.39 0.0219 0.98 to 11.80
Shalat et al. Pesticides and children on the US/Mexico border
48 Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(1)
This may be explained by the fact that exposures of these
very young children are likely to be multifactorial, including
ingestion of soil and dietary ingestion. In our limited study
of soils in this area, we were, however, surprised by the fact
that in general, soils had lower levels and fewer detectable
pesticides than housedust. Environmental factors including
the high spring and summer ambient temperatures and
intense ultraviolet levels from sunlight may have effected
these findings. The most unexpected factor in the analysis,
and therefore one that should be viewed guardedly, is the
apparent importance of hand surface area as a separate
predictor of dose. While it cannot be excluded that this is
solely a chance finding, it is possible that hand size is acting
as a surrogate for some factor or factors associated with the
child’s developmental stage. This will be evaluated in the
later rounds of testing in the current study.
The study observed an apparent association between
loadings of OPs on children’s hands and levels of urinary
metabolites of OPs. The current model explains about one -
quarter of the association at best. Clearly, child behavior is a
critical factor in the relationship between hand loadings and
nondietary ingestion of OPs. Currently, quantitative analysis
of hand -to -mouth and object -to-mouth behaviors of these
children is proceeding. This quantitative behavioral analysis
has the potential for providing important information on the
effect modification of an individual child’s behavior on
nondietary ingestion and, therefore, on dose. It is also likely
that some general characteristics associated with age - and
gender- speci fic behavior patterns will clarify this associa-
tion. Given the broad range of interindividual variation in
the frequencies of these be haviors that have been observed
in the preliminary analysis of the data, more data will be
required to develop meaningful models of exposure for
epidemiologic purposes. It is only through a better under-
standing of the interaction between children and their
environment that accurate dose estimations will become
possible.
One conclusion that clearly should emerge from this
preliminary evaluation is that young children, in these
border communities, have elevated levels of OPs in their
urine. We also observed a monotoni c decline in urinary OP
levels from 6 months of age, suggesting need for study of
even younger infants. Additionally, given the youthful
demographics of US/Mexico border communities and the
high birth rate, the development of a more comprehensive
understanding of the human health risks these environ-
mental contaminants present is essential. Little is known
about the potential for chronic health hazards long-term
exposure to these chemicals may represent. For these
reasons, the importance of conti nued study of environ-
mental pesticide exposure and its possible role in the
etiology of chronic illness among infants and children in
communities along the US / Mexico border must not be
overlooked.
Acknowledgments
This project was supported by EPA STAR grant no.
R827440 and NIEHS grants to the Center for Environ-
mental Health Sciences in Piscataway, NJ (P30 - ES -
05022), and the Center for Environmental and Rural Health
at Texas A&M University (P30-ES-09106 ). This study
would not have been possible without the cooperation and
assistance of the Sisters of Mercy of Laredo, TX.
References
Bradman M.A., Hearnly M.E., Draper W., Seidel S., Teran S., Wakeham D.,
and Neutra R. Pesticide exposures to children from California’s Central
Valley: results of a pilot study. J Expos Anal Environ Epidemiol 1997:
7: 217 – 234.
Bravo R., Driskell W.J., Whitehead R.D. Jr., Needham L.L., and Barr D.B.
Quantification of dialkyl phosphate metabolites of organophosphate
pesticides in human urine using GC - MS / MS with isotopic internal
standards. J Anal Toxicol 2002: 26: 245 – 252.
Centers for Disease Control and Prevention ( CDC ). National Report on
Human Exposure to Environmental Chemicals [ Online ]. Available at
http://www.cdc.gov/nceh/dlsreport, 2001.
Davis JR., Brownson RC., and Garcia R. Family pesticide use in the home,
garden, orchard, and yard. Arch Environ Contam Toxicol 1992: 22:
260 – 266.
Environmental Protection Agency ( EPA ). Method 3545: Pressurized Fluid
Extraction ( PFE ). Test Methods for Evaluating Solid Waste, Physical /
Chemical Methods, SW-846 [Version 2 ( December 1997 ), Integrated
Manual through Update III]. Washington, DC, U.S. Environmental
Protection Agency, 1997a.
Environmental Protection Agency ( EPA ). Method 8141A: Organopho-
sphorus Compounds by Gas Chromatography: Capillary Column
Technique. Test Methods for Evaluating Solid Waste, Physical /
Chemical Methods, SW-846 [ Version 2 ( December 1997 ), Integrated
Manual through Update III ]. Washington, DC, U.S. Environmental
Protection Agency, 1997b.
Ezzel J. The Use of SW - 846 Method 3545 for Automated Extraction of
Environmental Samples. American Environmental Laboratory, Inter-
national Scientific Communication (ISC), Shelton, CT, January 1998,
pp. 24 – 26.
Fenske R.A., Black K.G., Elkner K.P., Lee C.L., Methner M.M., and Soto
R. Potential exposure and health risks of infants following indoor
residential pesticide application. Am J Public Health 1990: 80: 689 –
693.
Fenske R.A., Kissel J.C., Lu C., Kalman D.A., Simcox N.J., Allen E.H.,
and Keifer M.C. Biologically based pesticide dose estimates for
children in an agricultural community. Environ Health Perspect 2000:
108: 515 – 520.
Griffith J.G., and Duncan R.C. Dialkyl phosphate residue values in the
urine of Florida citrus fieldworkers compared to the National Health
and Nutrition Examination Survey ( NHANES ) sample. Bull Environ
Contam Toxicol 1985: 34: 210 – 215.
Hill R.H., Head S.L., Baker S., Shealy D.B., Bailey S.L., Williams C.G.,
Sampson E.J., and Needham L.L. Pesticide residues in urine of adults
living in the US. Environ Res 1995: 71: 99 – 108.
Hu Y.A., Barr D.B., Akland G., Melnyk L., Needham L., Pellizzari E.D.,
Raymer J.H., and Roberds J.M. Collecting urine samples from young
children using cotton gauze for pesticide studies. J Expos Anal Environ
Epidemiol 2000: 10: 703 – 709.
Loewenherz C., Fenske R.A., Simcox N.J., Bellamy G., and Kalman D.
Biological monitoring of organo - phosphorous pesticide exposure
Pesticides and children on the US/Mexico border Shalat et al.
Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(1) 49
among children of agricultural workers in Central Washington State.
Environ Health Perspect 1997: 105: 1344 – 1353.
Lu C., and Fenske R.A. Dermal transfer of chlorpyrifos residues from
residential surfaces: comparison of hand press, hand drag, wipe, and
polyurethane foam roller measurements after broadcast and aerosol
pesticide applications. Environ Health Perspect 1999: 107: 463 – 467.
Lu C., Fenske R.A., Simcox N.J., and Kalman D. Pesticide exposure of
children in an agricultural c ommunity: evidence of household
proximity to farmland and take home exposure pathways. Environ
Res 2000: 84: 290 – 302.
McCauley L.A., Lasrev M.R., Higgins G., Rothlein J., Muniz J., Ebbert
C., and Phillips J. Work characteristics and pesticide exposures among
migrant agricultural families: a community - based research approach.
Environ Health Perspect 2001: 109: 533 – 538.
McCurdy S.A., Hansen M.E., Weisskopf C.P., Lopez R.L., Schneider F.,
Sanborn J.R., Krieger R.I., Wilson R.W., and Goldsmith D.F. As-
sessment of azinphosmethyl exposure in California peach harvest
workers. Arch Environ Health 1994: 49: 289 – 296.
O’Rourke M.K., Lizardi P.S., Rogan S.P., Freeman N.C., Aguirre A., Saint
C.G. Pesticide exposure and creatinine variation among young
children. J Expos Anal Environ Epidemiol 2000: 10: 673 – 681.
Patnaik P. Handbook of Environmental Analysis: Chemical Pollutants in
Air, Water, Soil, and Solid Wastes. Lewis / CRC Press, Boca Raton, FL,
1997, pp. 215 – 220.
Richter E.D., Kowalski M., Leventhal A., Graver F., Marzouk J., Brenner
S., Shkolnik I., Lerman S., Zahavi H., and Bashavi A. Illness and
excretion of organophosphate metabolites four months after household
pest extermination. Arch Environ Health 1992: 47: 135 – 138.
Richter B., Ezzel J., and Felix D. Single Laboratory Method Validation
Report: Extraction of Organophosphorus Pesticides, Chlorinated
Herbicides, and Polychlorinated Biphenyls Using Accelerated Solvent
Extraction ( ASE ) with Analytical Validation by GC / NPD and GC /
ECD, Document 101124. Dionex, Sunnyvale, CA, 1994.
Richter B., Hoefler F., and Linkerhaeger M. Determining Organopho-
sphorus Pesticides in Foods Using Accelerated Solvent Extraction with
Large Sample Sizes, Volume 19. LCGC, Eugene, OR, April 2001,
pp. 408 – 412 ( No. 4 ).
Savage E.P., Keefe T.J., Wheeler H.W., Mounce L., Halwic L., Applehans
F., Goes E., Goes T., Mihlan G., Rench J., and Taylor D.K. Household
pesticide usage in the United States. Arch Environ Health 1981: 36:
304 – 309.
Schantz M., Nichols J.J., and Wise S.A. Evaluation of pressurized fluid
extractio n for the extraction of environmental matrix reference
material. Anal Chem 1997: 69: 4210 – 4219.
Shalat S.L., Robson M.G., and Mohr S.N. Agricultural workers. In:
Rosenstock L., Cullen M.R., Brodkin C.D., and Redlich C. (Eds ),
Textbook of Clinical Occupational and Environmental Medicine, 2nd
edn. Saunders, Philadelphia, PA, 2003 ( In press).
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.
Sumner D., and Langley R. Pediatric pesticide poisoning in the Carolinas:
an evaluation of the trends and proposal to reduce the incidence. Vet
Hum Toxicol 2000: 42: 101 – 103.
van Wijnen J.H., Clausing P., and Brunekreef B. Estimated soil ingestion
by children. Environ Res 1990: 51: 147 – 162.
Wade T.L., Brooks J.M., Kennicutt M.C. II, McDonald T.J., Sericano J.L.,
and Jackson T.J. GERG trace Organics contaminant analytical
techniques. In: Lauenstein G.G., and Cantillo A.Y. ( Eds ), Sampling
and Analytical Methods of the NOAA National Status and Trends
Program National Benthic Surveillance and Mussel Watch Projects
1984 – 1992: Vols. I –IV. Technical Memorandum. NOS ORCA 71,
1993. NOAA / NOS / ORCA, Silver Spring, MD, 1986, pp. 121 – 128.
Whitemore R.W., Immerman F.W., Camann D.E., Bond A.E., Lewis R.G.,
and Schaum J.L. Non - occupational exposures to pesticides for resi-
dents of two US cities. Arch Environ Contam Toxicol 1994: 26: 47 – 59.
Zuloaga O., Etxebarria N., Fernandez L.A., and Madariaga J.M. Op-
timization and comparison of MAE, ASE and Soxhlet extraction for
the determination of HCH isomers in soil samples. Fresenius’ J Anal
Chem 2000: 367: 733 – 737.
Shalat et al. Pesticides and children on the US/Mexico border
50 Journal of Exposure Analysis and Environmental Epidemiology (2003) 13(1)
