Evaluation and Application of Methods for
Estimating Children' s Exposure to
Persistent Organic Pollutants in
Multiple Media
by
Jane C. Chuang, Christopher Lyu, Ying-Liang Chou, Patrick J. Callahan,
Marcia Nishioka, Kimberlea Andrews, Mary A. Pollard, Laura Brackney,
Charles Hines, Dave B. Davis, and Ronald Menton
Battelle
505 King Avenue
Columbus,Ohio 43201-2693
Volume I: Final Report
Contract Number 6S-D4-0023
Work Assignment 3-06
Project Officer
Nancy K. Wilson
Human Exposure and Atmospheric Sciences Division
National Exposure Research Laboratory
Research Triangle Park, North Carolina 27711
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
EPA DISCLAIMER
The information in this document has been funded wholly or in part by the United States
Environmental Protection Agency under EPA Contract Number 68-D4-0023 to Battelle
Memorial Institute. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
BATTELLE DISCLAIMER
Battelle does not engage in research for advertising, sales promotion, or endorsement of
our clients' interests including raising investment capital or recommending investment decisions,
or other publicity purposes, or for any use in litigation.
Battelle endeavors at all times to produce work of the highest quality, consistent with our
contract commitments. However, because of the research and/or experimental nature of this
work the client undertakes the sole responsibility for the consequences of any use, misuse, or
inability to use, any information, apparatus, process or result obtained from Battelle, and
Battelle, its employees, officers, or Trustees have no legal liability for the accuracy, adequacy, or
efficacy thereof.
ii
-------
Foreword
The mission of the National Exposure Research Laboratory (NERL) is to provide scientific
understanding, information and assessment tools that will quantify and reduce the uncertainty in EPA's
exposure and risk assessments for environmental stressors. These stressors include chemicals,
biologicals, radiation, and changes in climate, land use, and water use. The Laboratory' s primary
function is to measure, characterize, and predict human and ecological exposure to pollutants. Exposure
assessments are integral elements in the risk assessment process used to identify populations and
ecological resources at risk. The EPA relies increasingly on the results of quantitative risk assessments
to support regulations, particularly of chemicals in the environment. In addition, decisions on research
priorities are influenced increasingly by comparative risk assessment analysis. The utility of the risk-
based approach, however, depends on accurate exposure information. Thus, the mission of NERL is to
enhance the Agency's capability for evaluating exposure of both humans and ecosystems from a holistic
perspective.
The National Exposure Research Laboratory focuses on four major research areas: predictive
exposure modeling, exposure assessment, monitoring methods, and environmental characterization.
Underlying the entire research and technical support program of the NERL is its continuing development
of state-of-the-art modeling, monitoring, and quality assurance methods to assure the conduct of
defensible exposure assessments with known certainty. The research program supports its traditional
clients — Regional Offices, Regulatory Program Offices, ORD Offices, and Research Committees — and
ORD' s Core Research Program in the areas of health risk assessment, ecological risk assessment, and
risk reduction.
Human exposure to multimedia contaminants, including persistent organic pollutants (POP) is an
area of concern to EPA because of the possible mutagenicity and carcinogenicity of these compounds.
These compounds originate from industrial processes and combustion and are present in a variety of
microenvironments. The efforts described in this report provide an important contribution to our ability
to measure and evaluate human exposure to pollutants.
Gary J. Foley
Director
National Exposure Research Laboratory
iii
-------
Abstract
The objectives of this work assignment were to evaluate field methods for determining
children's exposure to selected persistent organic pollutants (POP), including polycyclic
aromatic hydrocarbons and other semi-volatile organic compounds (SVOC), to use these
methods to estimate the ranges of the potential exposures through air, dust, and food, and to
estimate the ranges of total exposures for a small set of children from low-income and middle-
income families.
The study was carried out in two phases to accomplish the above objectives. In Phase 1,
we evaluated the sampling and analysis methods for monitoring target POP in multimedia
samples and conducted a field study at nine daycare facilities. The target POP consisted of
multiple compound classes: polycyclic aromatic hydrocarbons (PAH), polychlorinated biphenyl
(PCB), phthalate esters (PE), phenols (Ph), organochlorine (OC) pesticides, organophosphate
(OP) pesticides, and a herbicide acid (HA). The sample media included air, dust, soil, duplicate
diet food, dermal wipes and urine. We applied the field methods to a field study of nine daycare
facilities to monitor target POP in multiple environmental media and food. Nine daycare centers
(four Head Start and five private centers) in North Carolina were recruited for the study.
Samples (air, dust, soil, and food) were collected from these daycare centers and analyzed for
target POP. In Phase 2, nine children were recruited as subjects from two of the nine daycare
centers. The multimedia samples (air, dust, soil, food, and wipes) were collected from the
daycare centers and from the subjects' homes. These samples were analyzed for target POP. In
addition, urine samples were collected from the subjects. The urine samples were analyzed for
selected POP and urinary POP metabolites.
This report was submitted in fulfillment of Work Assignments 2-01 and 3-06, Contract
68-D4-0023, to Battelle under the sponsorship of the United States Environmental Protection
Agency. This report covers the period from January 1,1997 to May 12, 1998, and work was
completed as of May 12, 1998.
iv
-------
Contents
Forward iii
Abstract iv
Figures viii
Tables ix
Abbreviations xii
Acknowledgment xiii
Chapter 1 Introduction 1
Chapter 2 Conclusions 5
Chapter 3 Recommendations 9
Chapter 4 Experimental Procedures 15
Method Validation for Multimedia Samples 15
Air Samples 15
Dust and Soil Samples 17
Dermal Wipe Samples 18
Urine Samples 19
Food Samples 20
Phase 1 Field Study of Nine Daycare Centers 22
Selection of Daycare Centers 23
Field Monitoring Activities 24
Sampling Methods 28
Analytical Methods 29
Data Analysis 32
Phase 2 Field Study in Two Daycare Centers and Nine Homes ........... 36
Selection of Homes 36
Field Monitoring Activities 37
Sampling Methods 41
Analytical Methods .42
Statistical Analysis 42
Estimates of Daily Persistent Organic Pollutants Doses 44
v
-------
Contents (Continued)
Chapter 5 Results and Discussion 46
Method Validation for Multimedia Samples 46
Air Samples 46
Dust and Soil Samples 50
Dermal Wipe Samples 55
Urine Samples 55
Food Samples 57
Phase 1 Field Study 60
Recruiting of Day Care Centers 60
Field Activities 61
Concentrations of Persistent Organic Pollutants in 62
Multimedia Samples
Data Analysis for Phase 1 Study 70
Summary Statistics 70
Correlation Between Sample Media 72
Correlation Between Compound Classes 74
Estimates of Daily Persistent Organic Pollutant Exposures 76
Quality Control Data for Phase 1 Study 79
Phase 2 Field Study 81
Recruiting of Subjects 81
Field Activity 82
Concentrations of Persistent Organic Pollutants in 83
Multimedia Studies
Summary Statistics 91
Correlation Between Sample Media 93
Correlation Between Compound Classes 93
Daily Potential Doses of Persistent Organic Pollutants 96
Concentration Profiles of Urinary Metabolites 100
Analysis of Variance 104
Regression Models 108
Quality Control Data for Phase 2 Study 113
References 114
vi
-------
Contents (Continued)
Appendices
A Analytical Procedures for Duplicate Diet Food Samples A-l
B Phase I Field Study B-l
C Phase 2 Field Study C-l
D Measured Target Persistent Organic Pollutant Concentrations in Air,
Dust, Soil, and Food Samples in Phase 1 Study D-l
E Summary Statics for Measured Target Persistent Organic Pollutant
Concentrations in Air, Dust, and Food Samples in Phase 1 Study E-l
F Spearman and Pearson Correlation Coefficients of Persistent Organic
Pollutants Between Sample Media F-l
G Spearman and Pearson Correlation Coefficients in Multimedia
Samples Between Compound Classes G-l
H Estimated Children's Daily Persistent Organic Pollutants Exposure Levels
in Phase 1 Daycare Centers from Inhalation, Nondietary Ingestion, and
Ingestion Pathways H-l
I Summary Statistics for Estimated Children's Daily Persistent Organic
Pollutant Exposure Levels in Phase 1 Daycare Centers 1-1
J Summary of Recoveries Data of the Spiked Persistent Organic Pollutants
in Phase 1 Multimedia Samples J-l
K Results of Duplicate Air, Dust, and Food Samples K-l
L Levels of Target Persistent Organic Pollutants Found in Phase 1 Field Blanks . L-l
M Target Persistent Organic Pollutants Concentrations in Phase 2
Multimedia Samples M-l
N Summary Statistics for Phase 2 Data in Multimedia Samples N-l
O Spearman and Pearson Correlation Coefficients Between Sample
Media for Persistent Organic Pollutants 0-1
P Spearman and Pearson Correlation Coefficients Between Compound
Classes in Multimedia Samples P-l
Q Summary of Analysis of Variance Results Q-l
R Regression Results of Urinary Metabolites Versus Environmental Sample
Media and Total Daily Potential Persistent Organic Pollutant Doses R-l
S Summary of Recovery Data of the Spiked Persistent Organic Pollutants
in Phase 2 Multimedia Samples S-l
T Levels of Target Persistent Organic Pollutants Found in Phase 2 Field Blanks . T-l
vii
-------
Figures
Figure 4.1. Locations of ten daycare centers in the Phase 1 Field Study 25
Figure 4.2. Loc ations of nine homes in the Phase 2 Field Study 39
Figure 4.3. Sampling schedule for the Phase 2 Field Study 40
Figure 5.1. Distributions of average daily exposure of persistent
organic pollutants from Phase 1 daycare centers 78
Figure 5.2. Average daily persistent organic pollutant exposures in
Head Start centers and other centers 80
Figure 5.3. Distributions of average daily potential dose of persistent
organic pollutants from homes for nine subjects 97
Figure 5.4. Distributions of average potential daily dose of persistent
organic pollutants from daycare centers for nine subjects 98
Figure 5.5. Distributions of average daily potential dose of persistent organic
pollutants from homes and daycare centers for nine subjects 99
Figure 5.6. Average potential daily persistent organic pollutants doses from
at-home exposures in low- and middle-income subjects 101
Figure 5.7. Average potential daily persistent organic pollutants doses from
at-center exposures in low- and middle-income subjects 102
Figure 5.8. Average potential daily persistent organic pollutant doses from
at-home and at-center exposures in low- and middle-income subjects 103
Figure 5.9. Average concentrations of 2,4-D and 3,5,6-TCP in subjects'
urine samples 105
Figure 5.10. Average concentrations of 1-naphthol, 2-naphthol, and
pentachlorophenol in subjects' urine samples 106
viii
-------
Tables
Table 3.1. Hypotheses to be Tested in Proposed Large-scale Study 10
Table 3.2. Minimal Sample Sizes Required to Detect a Difference in Persistent
Organic Pollutant Exposures Between Two Groups of Children
Based on a Two Sample t-test Conducted at the Five Percent
Significance Level 13
Table 3.3. Proposed Distribution of Children for the Large-Scale Study 14
Table 4.1. Summary of Daycare Centers in Phase 1 Study 26
Table 4.2. Field Monitoring Activities in Each Daycare Center 27
Table 4.3. Summary of Analytical Methods for Each Type of Samples
from Phase 1 Study 30
Table 4.4. Estimated Detection Limits of Target Persistent Organic Pollutants
in Multimedia Samples 33
Table 4.5 Summary of Subjects/Homes in Phase 2 Study 38
Table 5.1. Recovery of Target Persistent Organic Pollutants from
Filter/XAD-2 Samples 47
Table 5.2. Recovery of Target Persistent Organic Pollutants from Filter/PUF Samples .. 48
Table 5.3. Recovery of Non-Coplanar Polychlorinated Biphenyls from
Filter/XAD-2 Samples 49
Table 5.4. Recovery of 2,4-D and Pentachlorophenol from Filter/PUF Samples 51
Table 5.5. Recovery of Target Persistent Organic Pollutants from Spiked
House Dust Samples 52
Table 5.6. Recovery of Target Persistent Organic Pollutants from Spiked
Soil Samples 53
ix
-------
Tables (Continued)
Table 5.7. Recovery of Target Persistent Organic Pollutants from Spiked
Dermal Wipe Samples 56
Table 5.8. Recovery of Target Analytes from Spiked Urine Samples 57
Table 5.9. Recovery of 3.5,6-TCP from Spiked Urine Samples 58
Table 5.10 Recovery of Target Persistent Organic Pollutants from Spiked
Solid Food Samples 59
Table 5.11. Summary of Recruiting Daycare Centers for Phase 1 Study 61
Table 5.12. Summary of Target Persistent Organic Pollutants Determined
in Multimedia Samples 63
Table 5.13. Summary of Target Persistent Organic Pollutants in Air Samples from
Phase 1 Day Care Centers 64
Table 5.14. Summary of Target Persistent Organic Pollutants in Dust and Soil
Samples From Phase 1 Day Care Centers 65
Table 5.15. Summary of Target Persistent Organic Pollutants in Liquid and Solid
Food Samples from Phase 1 Daycare Centers 66
Table 5.16. Floor Dust Loadings from Phase 1 Day Care Centers 68
Table 5.17. Summary of Pairs of Sample Media with Significant Correlation
Coefficients for Target Compound Classes 73
Table 5.18. Summary of Pairs of Compound Classes with Significant
Correlation Coefficients for Each Sample Medium 75
Table 5.19. Correlation Coefficients Between Floor Dust (HVS3) and
Floor Dust (Bag) 77
x
-------
Tables (Continued)
Table 5.20. Summary of Recruiting Children from Daycare Centers D03
and D09 for Phase 2 Study 82
Table 5.21. Summary of Target Persistent Organic Pollutants in Air Samples
from Phase 2 Study 84
Table 5.22. Summary of Target Persistent Organic Pollutants in Dust and Soil
Samples From Phase 2 Study 85
Table 5.23. Summary of Target Persistent Organic Pollutants in Liquid and Solid
Food Samples from Phase 2 Study 86
Table 5.24. Summary of Target Persistent Organic Pollutants in Dermal Wipe
Samples From Phase 2 Study 87
Table 5.25. Floor Dust Loadings from Phase 2 Daycare Centers and Households 89
Table 5.26. Summary of Pairs of Sample Media with Significant Correlation
Coefficients for Target Compound Classes 94
Table 5.27. Summary of Pairs of Compound Classes with Significant
Correlation Coefficients for Each Sample Medium 95
Table 5.28. Summary of Anova on the Effect of Sampling Location: Daycare
Centers Versus Homes 107
Tabic 5.29. Summary of Anova on the Effect of Family Income: Low-income
Versus Middle-income 109
Table 5.30. Summary of Anova on the Effect of Family Income and Sampling
Location: Head Start Center and Low-Income Versus Regular Center
and Middle-Income 110
Table 5.31. Summary of Regression Model Results 112
xi
-------
Acronyms and Abbreviations
ANOVA
analysis of variance
BDL
below detection limit
DCM
dichloromethane
ECD
electron capture detection
EE
diethylether
El
electron impact
EL1SA
enzymed linked immunosorbent assay
ETS
environmental tobacco smoke
GC/MS
gas chromatography/mass spectrometry
GPC
gel permeation chromatography
HA
herbicide acid
HC1
hydrochloric acid
HP
Hewlett Packard
HVS3
High Volume Small Surface Sampler
MTBSTFA
N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide
N
sample size
NaCl
sodium chloride
N-BDL
number of samples below detection limit
NERL
National Exposure Research Laboratory
OC
organochlorine
OMB
Office of Management and Budget
OP
organophosphate
PAH
polycyclic aromatic hydrocarbons
PCB
polychlorinated biphenyls
PCP
pentachlorophenol
PE
phthalate esters
Ph
phenols
POP
persistent organic pollutants
PUF
polyurethane foam
QA
quality assurance
SIM
selected ion monitoring
SPE
solid phase extraction
SRS
analytical surrogate recovery standard
SVOC
semi-volatile organic compounds
3,5,6-TCP
3,5,6-trichloro-2-pyridinol
URG
University Research Glass
WA
Work Assignment
xii
-------
Acknowledgment
We thank Dr. Nancy K. Wilson of the U.S. EPA for her invaluable advice and participation
during this investigation.
xiii
-------
Chapter 1
Introduction
In 1992, the National Academy of Sciences identified polycyclic aromatic hydrocarbons
(PAH) and related semi-volatile organic compounds (SVOC) as among the compound classes of
highest priority for exposure research. Many of these compounds are persistent organic
pollutants (POP) and are possible endocrine disrupters or may have other quasi-hormonal or
reproductive effects. Many POP are sufficiently volatile or soluble to evaporate and condense, or
move otherwise through environmental media, including air, water, and soil. The POP can also
enter indoor microenvironments through outdoor air intrusion or through track-in and other
means of transport (1,3). Additionally, there are many potential sources of POP indoors,
including environmental tobacco smoke (ETS), consumer products, building materials, and home
chemicals.
Young children have immature organs, immune systems, and nervous systems that make
them more susceptible than adults to the potential health effects of toxic chemicals such as POP.
Their major sources of exposure can be expected to be found indoors, because young children
spend a lot of their time indoors (e.g., homes, daycare centers). Children can be exposed to POP
by inhaling contaminated air, ingesting tainted food, ingesting nondietary substances, or
absorbing pollutants through the skin from contaminated media. Exposures resulting from
ingestion of dust and soil are believed to be important for young children, because they play on
the floor or ground, and they tend to put many dirty food and non-food items into their mouths
(4). Indoor carpets, furnishings, and dust can serve as reservoirs for POP, and thus are potential
sources for human exposure to these compounds.
1
-------
In previous work, methods to measure and estimate the exposures of children in low-
income families to PAH were evaluated (5,6). However, such methods are incompletely
developed and evaluated in field situations for estimating total exposure to POP, which include
compound classes such as chlorinated organic compounds, substituted phenols, and phthalate
esters. Little information is available in the literature regarding the determination of urinary
metabolites of these compound classes.
Thus, small-scale POP exposure studies were carried out to obtain information for
designing future large-scale POP exposure studies on children. We evaluated field methods for
determining children's exposure to selected POP in multiple compound classes through multiple
environmental media and conducted two field studies in selected daycare centers and in subjects'
homes. We also validated an analytical method for determining 3,5,6-trichloro-2-pyridinol
(3,5,6-TCP), a urinary metabolite of chlorpyrifos, to determine 3,5,6-TCP in collected urine
samples, and we conducted statistical analysis of collected data.
The objectives were to evaluate field methods for determining children's exposure to
selected POP, including PAH and other SVOC, to use these methods for estimating the ranges of
the potential exposures through inhalation, ingestion, and dermal contact, and to estimate the
ranges of total exposures for a small set of children from low-income and middle-income
families.
In the first phase of this work, sampling and analysis methods for POP in multiple
environmental matrices: air, dust, soil, dermal wipes, and food were evaluated and validated.
Nine daycare centers located in and near Durham, North Carolina were recruited, and the
informed consent forms were signed by the directors of the daycare centers prior to the start of
any field sampling activities. Multimedia environmental samples (air, dust, and soil) and
duplicate diet food samples were collected from these daycare centers. The target POP included
multiple compound classes: PAH, phthalate esters (PE), phenols (Ph), polychlorinated biphenyls
(PCB), organochlorine (OC) pesticides, organophosphate (OP) pesticides, and a herbicide acid
(HA), 2,4-dichlorophenoxyacetic acid (2,4-D). In the second phase, nine children from two of
2
-------
the daycare centers were recruited. These children were between the ages of two and five and
spend at least 25 hr/week in daycare. Informed consent was obtained from each of the directors
of the daycare centers and from the subjects' parent(s) prior to any field monitoring activities.
Multimedia environmental samples were collected from the two daycare centers and the nine
subjects' homes. In addition, duplicate diet food, urine, and dermal wipe samples from the nine
subjects were collected both at their homes and at the daycare centers.
The specific tasks to accomplish the objectives of the two related studies were:
1. Prepare a work plan and quality assurance (QA) narrative and submit it for EPA
review and approval.
2. Prepare informed consent forms, questionnaires, activity diaries, and study
protocols. These were based on those developed previously under Cooperative
Agreement CR822073 and modified appropriately for this study.
3. Obtain Human Subjects Committee approval (IRB approval).
4. Validate the sampling and analysis methods.
5. Find and recruit nine daycare centers for participation in the Phase 1 study.
6. Collect environmental samples from indoor and outdoor air, carpet dust,
playground soil, and food at the nine daycare centers. Analyze the samples for the
target POP.
7. Recruit a total of nine child subjects for the Phase 2 study from two of the Phase 1
daycare centers: four subjects from a Head Start center and five subjects from
another daycare center.
8. Conduct a study of the total POP exposure of the children recruited above, having
separate samples of each matrix per child from daycare and home environments.
9. Analyze the archived sample extracts of air, dust, and soil, obtained from the PAH
exposure study under CR822073, for chlorpyrifos.
10. Analyze urine samples in the Phase 2 and in the PAH exposure studies for
chlorpyrifos.
-------
11. Conduct statistical analysis of the data and interpret the results.
12. Prepare this final report on the results of the studies in EPA ORD format.
4
-------
Chapter 2
Conclusions
The sampling and analysis methods for persistent organic pollutants (POP) in multimedia
samples were evaluated and validated. We used quartz fiber filter/XAD-2 as a sampling module
for collection of POP for polycyclic aromatic hydrocarbons (PAH), phthalate esters (PE),
organochlorine (OC), and organophosphate (OP) pesticides, polychlorinated biphenyls (PCB)
and substituted phenols (Ph) in air, and Teflon-coated glass filter/PUF for collection of a
herbicide acid (HA), 2,4-D. One analytical method — for air samples collected on filter/XAD-2
~ was validated for determining POP in all compound classes except 2,4-D. A separate
analytical method was validated for the determination of 2,4-D in air samples collected on
filter/PUF. The recoveries of target POP in all compound classes in air samples were, in general,
greater than 70%. The overall method precision for the two methods, for all target POP except
PE, was within ± 17% for the air samples. The greater variation observed in PE measurements
was from the high background levels of PE. Target POP were also stable in the respective
sampling modules over the 20-day storage interval.
Three analytical methods were validated to determine target POP in dust/soil. One
method was for compound classes PAH, PE, OC, OP, and PCB, one for Ph, and one for HA.
The POP recoveries in dust/soil were greater than 80% for most compound classes. As in the
case of the air samples, greater precisions were observed in the measurement of PE in dust/soil.
5
-------
An analytical method was validated to determine PAH, OC, OP, and PCB in dermal wipe
samples. The overall method precision was ±30% for most target POP. Quantitative recoveries
(>70%) of the spiked POP were achieved.
One analytical method was evaluated and validated for the determination of hydroxy
PAH, pentachlorophenol (PCP), and 2,4-D in urine samples. The recoveries ranged from 75 to
110% and the precision was within ± 45%. A different method was used for the determination of
3,5,6-TCP in urine samples. The recoveries of 3,5,6-TCP were greater than 90%, with precision
less than ± 5%.
An analytical method was developed for analyzing POP for all compound classes in
foods. Quantitative recoveries (>50%) were obtained for most target POP, and precision was in
general within ± 22%. The low recoveries for bisphenol-A and 2,4-D were probably due to
losses through the gel permeation chromatography (GPC) or solid phase extraction (SPE)
cleanup steps. The analytical method used for the food samples in this study needs to be revised
to provide better accuracy and precision for compound classes such as Ph and HA.
In the Phase 1 field study, target POP concentrations in indoor air were, in general, higher
than those in outdoor air. Some target compounds in the PAH, OC, and PCB classes were not
detected in the air samples. Target POP concentrations in floor dust were generally higher than
those in the playground soil. Most OC compounds and PCB were not found in the playground
soil. Concentrations of POP in solid foods were usually higher than those in liquid foods. None
of the target PCB were found in Phase 1 food samples and only a few PCB were detected in
Phase 2 food samples. The most abundant POP in all sample media were phthalate esters in
both Phase 1 and 2 samples.
Differences in POP concentrations in multimedia samples between Head Start daycare
centers and private daycare centers were small, with the exception of PAH in playground soil.
Average concentrations of PAH in the playground soil of Head Start centers were significantly
higher than those at the private centers. The difference in these averages was primarily due to
the extremely high levels of PAH found in the playground soil of center D01, where the PAH
6
-------
levels were the highest among all Phase 1 and 2 dust and soil samples. Such high PAH levels
could have come from local contamination sources. The estimated daily POP exposures from
daycare centers were similar between Head Start centers and private daycare centers.
In the Phase 2 field study, POP concentration profiles in air, dust, soil, and food were
generally similar to those seen in the Phase 1 samples. With few exceptions, differences in POP
concentrations in these multiple sample media were small between daycare centers and homes, as
well as between low-income and middle-income families. In addition, dermal wipe and urine
samples were collected from each subject both at home and at the center. Differences in POP
levels in wipes and urinary metabolites between these groups (home vs. daycare and low-income
vs. middle-income) were also small.
The correlation of POP concentrations in different sample media was investigated
separately for Phase 1 and Phase 2 data. Despite the small sample size, direct relationships
between indoor air and floor dust for target POP in PAH, OC, and OP were seen in both Phase 1
and Phase 2 samples. Since collection of floor dust samples required less effort as opposed to air
samples, it is feasible to collect floor dust samples for screening POP in future large-scale
exposure studies. The correlation of POP in each compound class with each other compound
class was also investigated separately for Phase 1 and 2 data. A direct relationship between B2
PAH (probable human carcinogens classified by U.S. EPA Integrated Risk Information System)
and total PAH was observed for dust, soil, and solid food. Since B2 PAH and total PAH are in
the same compound class, one expects to see a strong relationship. A strong relationship
between B2 PAH and total PAH in dust and soil was also observed in our previous PAH
exposure study (5,6). There were a few significant correlations between compound classes for
target POP, but they were not consistent between the Phase 1 and Phase 2 data, for which the
sample size was small.
The relative importance of the exposure pathways (inhalation, nondictary ingestion,
dietary ingestion) from the exposures at the daycare centers in Phase 1 data was similar to those
for the Phase 2 data. The differences in the distributions of exposure pathways from at-home
7
-------
exposures and at-center exposures in Phase 2 data were small. The relative rank importance of
the exposure pathways for the children's total daily potential doses of POP was different among
the compound classes. For total daily potential doses of total target PAH, organochlorinc and
organophosphate pesticides, and PCB, the relative importance of exposure pathways was
inhalation > dietary > nondietary; for total daily potential doses of phthalate esters, substituted
phenols, and 2,4-D, it was dietary > nondietary > inhalation. Nondietary ingestion was the most
important pathway for the total daily potential dose of B2 PAH.
Differences in POP concentrations in most multiple sample media between the groups of
at-home versus at-center and low-income versus middle-income were not statistically significant.
Significantly higher concentrations of PCB in air as well as total PAH and total phenols in foods
were obtained in the samples collected at daycare centers as opposed to homes. Lower
concentrations of organophosphate pesticides in outdoor air were seen in the samples from
daycare centers. Low-income subjects had higher levels of phthalate esters and organochlorine
pesticides and lower levels of total PAH in their wipe samples. Higher concentrations of
hydroxy-PAH and 2,4-D, but lower concentrations of pentachlorophenol (PCP) and 3,5,6-TCP
were found in urine samples from low-income subjects. With one exception PCP, there were no
strong positive relationships between the urinary metabolites and POP in multimedia samples, as
well as urinary metabolites and total daily potential doses of POP. Concentrations of PCP in the
children's urine samples were significantly related to the subjects' total daily dose of PCP.
8
-------
Chapter 3
Recommendations
We recommend that a large-scale field study of children's exposure to persistent organic
pollutants (POP) be conducted, using the methods evaluated in this study. The results of a
previous study (5,6) demonstrated that children's potential daily doses (ng/kg/day) of B2-PAH
(probable human carcinogens) were higher than those of adults in each of 24 low-income
households. Among those children of low-income families, inner city children appeared to have
higher polycyclic aromatic hydrocarbons (PAH) exposure than rural children. Both the dietary
and nondietary ingestion pathways were important for children's exposure to B2-PAH. Results
from the current studies have also shown that dietary ingestion is the most important exposure
pathway for phthalate esters (PE) and phenols (Ph). However, inhalation appears to be an
important pathway for children's exposure to other POP, including total PAH, polychlorinated
biphenyls (PCB), organochlorine (OC), and organophosphate (OP) pesticides. Children's
exposures to POP from the day care centers they attend appears to be as important as exposure
from their homes.
An important question is whether children in low-income families could have higher
exposure to some POP than children in higher-income families. These higher exposures could
derive from the location of their homes, from the location of the day care centers they attend,
from the presence of environmental tobacco smoke (ETS), or from other causes. A general
outline for conducting a study to explore these issues follows.
The main objective of the proposed study that we propose would be to obtain additional
data on children's exposures to POP and to determine whether low-income children are at greater
9
-------
POP exposure. A secondary objective is to examine the relative importance of the exposure
pathways for POP in various compound classes. A third objective is to examine the effectiveness
of utilizing less expensive screening tools for monitoring POP exposure. The specific
hypotheses to be tested are summarized in Table 3.1.
TABLE 3.1. HYPOTHESES TO BE TESTED IN PROPOSED LARGE-SCALE STUDY
Objective
Determine children's POP exposure profiles
and investigate impact of SES on children's
exposure to POP
Investigate impact of sources by media on
POP exposure
Investigate impact of sources in homes and in
day care centers on POP exposure
Investigate contributions of pathways to total
POP exposure
Investigate the association between screening
methods versus conventional methods
Investigate the association among various
estimates of POP exposure
Hypothesis
Exposure to POP of low-income children is/is
not higher than those of higher-income
children
Exposure from a source(s) is/is not different
in air, soil, dust, and food for different
compound classes
Exposure to POP at home is/is not equal to
those at day care centers
Exposure from different pathways is/is not
equal and distributions are/are not equal for
different compound classes
Exposure measured by screening methods
(ELISA and PAH monitor) does/does not
correlate well with conventional methods
Exposure estimates from floor dust measured
by HVS3 do/do not correlate well with those
measured using vacuum bags.
Exposure measured by multiple
environmental media does/does not correlate
well with biological markers.
Exposures from various media do/do not
correlate with each other
10
-------
Sample size calculations were conducted to determine the minimum sample size required
to statistically distinguish between POP exposures of children in low- and middle-income
families. Sample sizes were calculated for the first hypothesis, which is the primary objective of
the Battclle-proposcd study. To support the sample size calculations, data on childhood POP
exposures from the previous study under Cooperative Agreement CR822073 and the current
study Phases 1 and 2 were first reviewed. The review assessed the distribution and variability of
POP concentrations in floor dust, indoor and outdoor air, solid and liquid food, and soil for seven
target POP, namely, B2 PAH, OP, PE, Ph, diazinon, chlorpyrifos, and bisphenol-A. In addition,
the distribution of OH-PAH in urine was reviewed. It was found that
1. POP concentrations tend to be lognormally distributed.
2. While there are differences in the variability of POP concentrations among the six
media and in urine, the standard deviations of log-transformed (hi) POP
concentrations generally range from 0.50 to 2.0.
3. Differences in geometric mean POP concentrations between low-income and
higher-income families as well as between daycare centers and homes generally
range from 0 to 500%, between city and rural areas, they range from 0 to 150%, and
between smoker and non-smoker houses, they range from 0 to 250% (B2-PAH
only), depending on the compound and media.
Based on the analysis of the historical and current data the calculations were performed
with the following assumptions:
1. A two-sample t-test is conducted at the 5 percent significance level on In-
transformed POP to compare the POP exposures in the following groups of
children:
• low-income families versus middle-income families,
• at daycare centers versus at home,
• inner city versus rural areas,
• smokers' homes versus nonsmoker's homes.
11
-------
The comparison of POP exposures in smoker versus non-smoker houses was
performed only on B2 PAH, because only these two groups have data from the PAH
exposure studies.
2. Sample sizes were computed that provide 80 or 90 percent power for detecting a
significant difference between the two groups when the actual percent difference
ranges from 10 to 200 percent. (The power represents the level of confidence
desired to detect a specified difference between the two groups. An experiment
designed to have 90 percent power for detecting a specified difference will be more
sensitive than one designed to have 80 percent.)
3. Sample sizes were computed assuming that the standard deviation of ln-transformed
POP concentrations is either 0.5,1,1.5, or 2.
Table 3.2 summarizes the estimated sample sizes required to detect specific differences
between any two groups of children, i.e. low- and middle-income children, at-center and at-home
children, inner-city and rural children, or children from smoker's and nonsmoker's homes. For
example, to detect a difference between two groups of children if the standard deviation of
ln-transformed POP is 1.0 and the actual percent difference between the two groups is 100 percent,
roughly 34 children are needed in each group to give 80% power. If the standard deviation of
ln-transformed POP is 1.5, the number of children increases to approximately 75 per group.
For all seven target POP concentrations and all six media (floor dust, indoor and outdoor
air, solid and liquid food, and soil), the median percent difference in POP concentrations between
the two groups of children was 60 percent, and the median standard deviation in ln-transformed
POP concentrations was 1.0. As shown in Table 3.2, if the standard deviation in ln-transformed
POP concentrations is 1.0 and the actual percent difference between the two groups of children is
50 percent, then a sample size of approximately 100 children per group will provide 80 percent
power for detecting a statistical difference in POP exposures between any two groups of children.
Under the same conditions, a sample size of approximately 130 children per group provides 90
percent confidence that a statistical difference will be detected between the two groups. To allow
for missing samples and other data issues, a sample size of 120 to 160 children per group is
recommended.
12
-------
We could recruit 120 children each from low-income and middle-income families
(targeted at 80 % power) for the large-scale study. We could also recruit children who attend
daycare centers and monitor these subjects at their homes and at the centers they attend. We
could recruit at Head Start centers to target low-income families, and at private centers for
middle-income families. Ideally, the 120 children in each group (low-and middle-income)
should be evenly distributed in each subgroup, i.e., inner city versus rural, and smokers' versus
nonsmokers' homes, in order to provide the same power. The distribution of children we
proposed for the large-scale study is summarized in Table 3.3.
TABLE 3.2. MINIMAL SAMPLE SIZES REQUIRED TO DETECT A DIFFERENCE
IN PERSISTENT ORGANIC POLLUTANT EXPOSURES BETWEEN
TWO GROUPS OF CHILDREN BASED ON A TWO SAMPLE T-TEST
CONDUCTED AT THE FIVE PERCENT SIGNIFICANCE LEVEL
Percent Difference Between Geometric Mean POP Exposures for Two
Power Standard Groups of Children"'
Deviation00
10%
25%
50%
100%
150%
200%
0.5
433
80
25
11
<10
< 10
1
1729
316
97
34
20
15
80%
1.5
3889
710
216
75
43
30
2
6913
1,262
383
132
76
53
0.5
579
107
33
12
< 10
< 10
1
2314
423
129
45
26
19
90%
1.5
5206
951
289
99
57
40
2
9255
1,689
512
176
101
71
(a) Standard deviation of ln-transformed POP concentrations.
(b) Two groups of children could be low- and middle-income children, at-center and at-
home children, inner city and rural, or children from smokers and nonsmokers homes.
13
-------
TABLE 3.3. PROPOSED DISTRIBUTION OF CHILDREN FOR THE LARGE-
SCALE STUDY
Head Start Centers/Low-Income
Private Centers/Middle-Income
Inner City
Rural Areas
Inner City
Rural Areas
Smokers
30
30
30
30
Nonsmokers
30
30
30
30
The large-scale POP exposure study could be conducted over a period of three years. In year one,
the study design, questionnaires, consent forms and protocols would be prepared; the required approvals
from the Office of Management and Budget (OMB), the Human Subjects Committee at Battelle, and the
U.S. EPA would be obtained. In the second year, the field sampling and analysis activities would be
performed and continued through the first quarter of the third year. Data analysis would be conducted in
the remainder of the third year.
The overall technical approach would be as follows:
Establish a study design
• Establish questionnaires, consent forms, and protocols
Obtain required approval from Office of Management and Budget (OMB) and from
Human Subjects Committee at Battelle and U.S. EPA
• Recruit eligible day care centers and households
• Conduct field sampling
• Analyze collected multi-media samples
• Conduct data analysis
Prepare final report.
14
-------
Chapter 4
Experimental Procedures
Method Validation for Multimedia Samples
Air Samples
Three sets of method validation tests were carried out using the standard 4-L/min
University Research Glassware (URG) air sampler. Each sampler consisted of an inlet jet
equipped with an impactor for 10 jim particle collection, a filter (quartz fiber or Teflon-coated
glass) and a glass cartridge packed with either XAD-2 resin or polyurethane foam (PUF) plug.
These tests consisted of measurements on (1) sampling retention efficiency and storage
stability of persistent organic pollutants (POP) on XAD-2 and PUF, (2) sampling retention
efficiency of non-coplanar polychlorinated biphenyls (PCB) on XAD-2, and (3) sampling
retention efficiency of pentachlorophenol (PCP) and 2,4-D on a combined Tcflon-coated glass
fiber filter/PUF cartridge. For the first set of tests, four sampling units, each consisting of
three quartz fiber filter/XAD-2 sampling modules, were placed in a clean laboratory room at
Battelle. Prior to sampling, six of the twelve XAD-2 traps were spiked with known amounts
of the representative POP in multiple compound classes, which are: polycyclic aromatic
hydrocarbons (PAH), coplanar PCB, phthalate esters (PE), phenols (Ph), organochlorine (OC)
and organophosphate (OP) pesticides. The remaining six traps were not spiked with POP, nor
were the filters spiked with any target POP. Spiking was performed by injecting 100 jiL of
the spiking solution (100 ng/compound) into the center of the XAD-2 resin bed at about 1-cm
depth, followed by 5 min of air-drying. Indoor air was sampled at a flow rate of 4 L/min for
48 hours with the 12 samplers. These experiments were repeated under the same conditions
except that the quartz fiber filter/XAD-2 sampling module was replaced with a quartz fiber
filter/PUF sampling module.
15
-------
After sampling, three of the spiked and three of the non-spiked filter/XAD-2 samples
were extracted with dichloromcthanc (DCM) by the Soxhlet technique; filter and sorbent were
combined for extraction. The filter/PUF samples were extracted with 10 % diethylether (EE)
in hexane instead of DCM. Four of the remaining samples, two spiked and two non-spiked,
were stored at 4 C in the dark for 10 days before extraction. The remaining two samples, one
spiked and one non-spiked, were stored in the same way for 20 days before extraction. The
sample extracts were concentrated to 1 mL by Kuderna-Danish evaporation and analyzed by
gas chromatography/ mass spectrometry (GC/MS) for the target POP. To improve the
chromatographic peak shape for PCP, the sample extracts were methylated using
diazomethane and then reanalyzed by GC/MS for PCP.
For the second experiment, known amounts of 13 non-coplanar PCB (100
ng/compound), were spiked into the XAD-2 portion of two quartz filter/XAD-2 sampling
modules. Air sampling, at 4 L/min for 48 hours was carried out with these two spiked
modules and one non-spiked filter/XAD-2 trap. Samples were extracted with DCM
immediately after sampling and analyzed by GC/MS.
In the third set of experiments, known amounts of PCP and 2,4-D (100 ng/compound)
were spiked onto the filters of three Teflon-coated glass filter/PUF sampling cartridges. The
filters were spiked in this experiment because prior studies have shown that 2,4-D is retained
exclusively on the filter except under conditions of very high humidity. The retention
characteristics of PCP with this sampling module were not known. Two non-spiked sampling
modules were used as blanks. Air sampling at 4 L/min was conducted with these five
samplers for 48 hours. Samples were stored for 10 days prior to extraction. Filters and PUF
were extracted and analyzed separately using methods previously developed for 2,4-D, with
minor modification for retention of PCP in the hexane partition step (2). This modification
included removal, via Turbo-vap evaporation, of the acetonitrile from the extraction solvent
prior to the hexane partition cleanup step. Analyses were carried out using a GC/electron
capture detection (ECD) method.
16
-------
Dust And Soil Samples
Three sets of method validation tests were conducted to determine POP in all compound
classes of interest except Ph and HA, in dust and soil. These experiments consisted of
(1) sonication extraction with hexane with no cleanup, (2) sonication extraction with DCM
followed by silica SPE cleanup, and (3) sonication extraction with 10 % EE in hexanc followed
by Florisil SPE cleanup.
In the first set of experiments, nine aliquots of a house dust sample (0.5g each) and nine
aliquots of a pathway soil sample (lg each) were used. Three aliquots of the dust sample were
spiked with low levels of target POP (100 ng/compound); three other aliquots were spiked with
higher levels of the POP (500 ng/compound); and the remaining three aliquots were non-spiked
samples. The soil sample aliquots were treated the same way as the dust sample aliquots.
Following spiking, the samples were air-dried in a hood for approximately 15 minutes. The
spiked and non-spiked samples were extracted by sonication with 10 mL of hexane for
15 minutes and this extraction step was repeated with another 10 mL of hexane. The hexane
extracts were concentrated and analyzed by GC/MS for target POP.
For the second set of experiments, DCM was used as the extraction solvent and only the
low-level spiking experiments were conducted. After extraction, the sample extract was solvent
exchanged into hexane and a silica SPE column was conditioned with 10% methanol in hexane,
followed by 100% hexane. The sample extract was then applied to a silica SPE column. The
sample extract was eluted with 8 mL of 40% hexane in DCM and followed by 8 mL of 100%
DCM. Both fractions were concentrated and analyzed by GC/MS for target POP.
For the third set of experiments, the same house dust and pathway soil samples were used
and low level (100 ng/compound) and high level (1000 ng/compound) spiking experiments were
carried out. The extraction solvent was changed to 10% EE in hexane. After extraction, a
Florisil SPE column was conditioned with 50% EE in hexane, followed by 100% hexane. The
sample extract was then applied to the column and eluted with 12 mL of 15% EE in hexane.
This fraction was concentrated and analyzed by GC/MS for target POP.
17
-------
Additional spike recovery experiments for Ph were carried out using the methods
previously developed with minor modification (7). The spiked and non-spiked dust samples
were sequentially extracted with 5 % acetic acid in methanol, DCM, and 5 % acetic acid in
distilled water. The resulting mixture was then partitioned with distilled water and the DCM
extract was concentrated and divided into two portions. One portion was methylated with
diazomethane and analyzed for PCP and the other portion was not methylated and was analyzed
for other target phenols.
Dermal Wipe Samples
The dermal wipes (Johnson & Johnson Sof-Wick) were precleaned prior to use. The
wipes were cut into half along the fold using clean scissors and extracted with DCM overnight.
After extraction, the wipes were dried in a stainless canister with nitrogen under vacuum at 60 C
for at least 12 hours. After drying the wipes were stored in a clean jar. Prior to field use, an
aliquot (2 ml) of 50% isopropanol in distilled water was dispensed onto the entire surface of the
wipe. The wet wipe was folded in half and placed in a clean jar ready for field use.
Six clean wipes were used in the spike recovery test. Six wipe samples were obtained
from a child subject by wiping both forearms and palms during a three-day period. Four of these
dermal wipe samples were spiked with known amounts of target POP including PAH, PCB, OC,
and OP pesticides (20 ng/compound) and the remaining two samples were not spiked with any
POP. Both spiked and non-spiked samples were extracted with 10% ethylether (EE) in hexane
by the Soxhlet technique. Two of the spiked and one non-spiked sample extracts were
partitioned with distilled water before concentration and the remaining spiked and non-spiked
sample extracts were concentrated to 1 mL without liquid-liquid partition. The concentrated
sample extracts were then subjected to the Florisil SPE column cleanup step. The Florisil
column was conditioned with 50 % EE in hexane and followed by 100 % hexane. The sample
extract was then applied to the column and eluted with 12 mL of 15 % EE in hexane as the first
fraction and followed by 12 mL of 50 % EE in hexane as the second fraction. Both fractions
were concentrated to 1 mL and analyzed by GC/MS for spiked POP. A second spike recovery
-------
study was conducted by spiking low levels (5 ng/compound) of the target POP onto the clean
wipe samples. Five spiked samples and one non-spiked samples were extracted with 15% EE in
hexane. Two of the spiked and the non-spiked samples were cleaned up by a Florisil SPE
column with the conditions described above. The other three spiked samples were eluted with
24 mL of 15% EE in hexane. The fractions were concentrated and analyzed for spiked POP.
Urine Samples
Two sets of method validation tests were conducted to determine 2,4-D, PCP, hydroxy
PAH, and 3,5,6-TCP in urine samples. These experiments consisted of (1) analytical method
validation for 2,4-D, PCP, and hydroxy-PAH and (2) analytical method validation for 3,5,6-TCP.
In the first set of experiments, eight aliquots of an adult's urine sample (50 mL each)
were used. Two aliquots of the urine sample were spiked with low levels of 2,4-D, PCP, and
hydroxy-PAH (20 ng/compound); two aliquots were spiked with medium levels of the target
compounds (50 ng/compound), two aliquots were spiked with high levels of the target compound
(500 ng/compound); and the remaining two aliquots were non-spiked samples. Each urine
sample (spiked and non-spiked) was refluxed with 6 N hydrochloric acid (HC1) and DCM for
one hour.
The resulting mixture was extracted with DCM by liquid-liquid partitioning. The DCM
extracts were combined, concentrated, methylated by diazomethane, and solvent exchanged into
hexane. The hexane extract was subsequently cleaned up by a Florisil SPE column. The SPE
column was conditioned with 50 % DCM in hexane and followed by 100 % hexane. The sample
extract was then applied to the column and eluted with 18 mL of 50 % DCM in hexane. The
fraction was concentrated and analyzed by GC/MS for the target compounds.
In the second set of experiments, 12 aliquots of an adult's urine sample (1 mL each) were
used. Nine aliquots of the urine samples were spiked with 3,5,6-TCP at three levels (1 ng, 3 ng,
and 10 ng) in triplicate and the remaining three aliquots were non-spiked samples. Each urine
sample (spiked and non-spiked) was fortified with the internal standard (13C15N-TCP) and
acidified with 100 fil of concentrated HC1. Note that both 3,5,6-TCP and 13C15N-TCP were
-------
obtained from Dow Elanco (Indianapolis, IN). The mixture was then heated at 80°C for 2 hours.
After healing, an aliquot of 10% sodium chloride (NaCl) solution was added to the mixture. The
resulting extract was extracted with DCM and concentrated to approximately 1 mL. The
concentrated DCM extract was solvent exchanged into toluene. The toluene extract was
derivatized with 100 jj.1 of N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide (MTBSTFA)
at 60°C for one hour. Standard solutions containing 3,5,6-TCP and 13CISN-TCP were prepared in
toluene. The standards were derivatized with MTBSTFA as described above. The samples and
standard solutions were then analyzed by GC/MS.
Food Samples
The objective of the method evaluation for the food samples was to establish an analytical
method which would remove the bulk of the fat from the food, but allow for adequate recovery
of target POP of various compound classes. Several experiments were performed in an attempt
to develop an optimal extraction and clean-up procedure to cover all compound classes of
interests. This proved to be a major challenge due to the amount of fat which was contained in
the samples.
Food samples were extracted with three different solvents (acetone, hexane, and DCM) in
an attempt to reduce the amount of fat extracted with the analytes. The acetone removed all the
water from the sample making concentration extremely difficult. The hexane and DCM extracts
were placed in a freezer for a short time in an attempt to perform a gross removal of the fat. The
organic solvent could not be separated from the precipitated fat, and DCM was used as extraction
solvent for later experiments because of its ability to remove both polar and non-polar
compounds from the food matrix. Food samples were also extracted in the presence of salts
(sodium oxalate and zinc acetate) with the intent of degrading the fats.
Two chromatography columns, Sephadex and gel permeation were evaluated for the
removal of the fat from the food samples. The fat and POP elution profiles of the Sephadex
column were determined by residue weights and by GC/ECD analysis. The sample made of com
oil and selected POP including OC and OP pesticides was applied to a Sephadex column and
20
-------
eluted with 12x1 mL DCM and aliquots of the eluates were removed for residue weight
measurements of fat content. The fractions containing no residue were analyzed by GC/ECD for
OC and OP pesticides. The residue weight measurements were also determined for each
fraction.
Because the Sephadex column proved to be an unacceptable method for fat removal, gel
permeation chromatography (GPC) was then evaluated as a means of removing the fat from the
food matrix. Two GPC columns, Optima and SX-3 biobeads, were employed initially. The
Optima column used a slightly lower flow rate than the biobeads column (4 mL/minute vs.
5 mL/minute), had a faster run time (30 minutes vs. 60 minutes), and had a lower loading
capacity (0.5 gram vs. 1 gram). The biobeads column was chosen for evaluation tests because of
its higher loading capacity. The food samples spiked with target POP including PAH, PCB, PE,
Ph, OC, and OP and solvent spikes were extracted with DCM and fractionated by a biobeads
column. The fat and target POP elution profiles were determined by residue weights and by
GC/MS or GC/ECD analysis. The optimal GPC condition was established to separate
approximately 90 % of the fat from the target POP. The optimal GPC condition consisted of a
biobeads column with DCM as mobile phase operated at a flow rate of 5 mL/minute for 60 min.
The dump time was set at 0 to the valley between corn oil and phthalate esters from the standard
analysis, generally around 25 to 28 min. The collect time (generally 32 to 35 min.) was then set
from the end of the dump time to a total fractionation time of 60 min.
After the GPC cleanup, a Florisil SPE column was used to further remove possible
interference peaks from the sample extract. The eluting solvent systems were evaluated to
determine the optimal elution solvent for target POP. The solvent systems evaluated were: 6%
ethyl ether (EE) in hexane, 15% EE in hexane, 50% EE in hexane and 20 % acetone in hexane.
After the optimal GPC and Florisil SPE conditions were established, the spike recovery
experiments for food samples were carried out. A solid food duplicate diet sample was fortified
with known amounts of POP including PAH, PCB, PE, OC, and OP. Three aliquots of the
spiked food samples were extracted with DCM, and fractionated through GPC using the
procedures described above. The GPC fraction was concentrated to 1 mL and solvent exchanged
into hexane. The hexane extract was then applied to a Florisil SPE column and eluted with 2x6
-------
2x6 mL of 15 % EE in hexane. The collected fraction was then concentrated and analyzed by
GC/MS for target POP.
The analytical method for 2,4-D developed from the previous study (2) was modified for
the determination of 2,4-D and PCP in Phase 1 food samples. The procedures were described in
the later section of Phase 1 analytical methods. The PCP could not be recovered by this method.
Acceptable recoveries of 2,4-D were obtained but multiple SPE columns were required for each
sample because the SPE columns became plugged. Therefore, one additional evaluation test was
carried out for the determination of PCP and 2,4-D in food samples. The solid food sample was
spiked with known amounts of 2,4-D and PCP, extracted with DCM, and fractionated by GPC.
The GPC fraction was methylated with diazomethane, concentrated, and solvent exchanged to
hexane for Florisil SPE column clean-up. The eluting solvent used was 50% DCM in hexane.
Adequate recoveries of PCP and 2,4-D were obtained from this experiment. Thus, the Phase 2
food sample was extracted with DCM and fractionated by GPC. The GPC fraction was split into
two portions. Portion I was cleaned by Florisil SPE and analyzed for target PAH, PCB, OC, OP,
PE, and Ph. Portion II was methylated, cleaned by Florisil SPE, and analyzed for 2,4-D and
PCP. The analytical procedures for the food samples are described in Appendix A.
Phase 1 Field Study of Nine Daycare Centers
The Phase 1 field study was conducted in nine daycare centers which are located in urban
and rural areas, primarily in Durham County, North Carolina. Four of the daycare centers are in
the Head Start program, which serves mostly low-income clients, and five are not. Approvals
were obtained from the Human Subjects Committees at Battelle before field activities began.
This section summarizes the procedures used for selection of nine daycare centers; the
study consent forms, the questionnaires, and building characteristics observation survey; the
sampling procedures used to collect air, dust, soil, and food; and the analytical methods used to
determine POP in the collected samples. The recruiting material, consent form, pre-monitoring
questionnaire, post-monitoring questionnaire, and house/building characteristics observation
survey used in the Phase 1 study are presented in Appendix B.
-------
Selection of Daycare Centers
We used Select Phone, a national telephone database on CD-ROM developed by ProCD
Inc., to prepare the target list for recruiting the daycare centers which are not in the Head Start
program. The database contains 100 million U.S. business and residential listings on six
CD-ROM discs. We used the SIC code (child care providers) and geographic locations
(Durham, Chapel Hill, and Raleigh, North Carolina) to select the eligible daycare centers.
Battelle survey staff called every daycare center on the target list to confirm the addresses and
phone numbers, the names of daycare center directors, and obtain some general information
about the centers (age groups of the children, number of children enrolled in the daycare center,
and business hours).
A total of 202 advance letters were mailed to the directors of eligible child daycare
centers in the Durham, Chapel Hill, and Raleigh areas. The letter provided the background
information about the study and informed the center directors that they would be contacted by
our study staff. We began calling the center directors one week after the mailing of advance
letters. Most daycare center directors had to obtain approval from their board of directors or
the daycare owner in order to participate in the study. It took at least one to two weeks to get
a final response from a daycare center. We obtained participation consent from 18 daycare
centers in the Durham/Chapel Hill areas and 6 in the Raleigh area.
For daycare centers in the Head Start program, we used the Select Phone database
(using name search) to identify all Head Start daycare centers in the same geographical areas
(Durham, Chapel Hill, and Raleigh, North Carolina). Head Start program is a federally funded
child care program which provides free child care services to low income families. To be
eligible for the program services, a family's income must meet the federal income guidelines and
other criteria. About 90% of the children in a Head Start center are from low income families.
The same procedures were used to recruit the Head Start centers. Advance letters were
mailed to the Head Start directors. About one week after the mailing of advance letters,
survey staff called the Head Start directors to obtain participation consent. Initially, the Head
Start centers' directors were concerned about the potential negative effects from the study and
were reluctant to participate in the study. Many phone calls were made to different levels of
-------
people in charge of the program to persuade them to participate in the study. Final selection
of the centers was based on the numbers of children enrolled, the location of the centers, and
the type of centers (private/Head Start). Nine daycare centers (five private and four Head
Start) were selected for the Phase 1 study; one Head Start backup center and five backups for
private centers were also assigned.
Field Monitoring Activities
The Phase 1 field sampling activities were performed in March 1997. Initially nine
daycare centers were planned for the study. Figure 4.1 displays the locations of the participating
daycare centers for the Phase 1 study. We had to use one of the backup centers (D10)? because
the outdoor sampling of one daycare (D05) was disrupted. A piece of plastic had blocked the
cooling fan of the sampling module and the sampling pump overheated. Additional 24-hr
outdoor air sampling was conducted for D05. The other multimedia samples were collected from
daycare D05 as planned. A summary of the daycare centers participating in the Phase I study is
given in Table 4.1. The field monitoring activities for each daycare center are summarized in
Table 4.2.
The pre-monitoring questionnaire and the post-monitoring questionnaire used were
modified from previous EPA work. The pre-monitoring questionnaire contains the information
about the use of chemicals, and children's daily activities at the center. The post-monitoring
questionnaire collects information about the participant's feed back about the study activities. A
field data log book was also used for each daycare center to record critical information during the
sampling period and pictures were taken to document the progress of the study.
24
-------
N>
Ui
Oxfond
Henderson
401
; i158
emont
D05, D06, D09, AND D10
Hillsborqujgh
imi s
~llrham \ / ,—^
s1
D07 AND D08
D02 AND D04
OftafceMill
hMif751
Raleigh
15
PittSboro
3264te
. vary
arner
prings
10 mi
Figure 4.1. Locations of ten daycare centers in the Phase 1 Field Study.
-------
TABLE 4.1. SUMMARY OF DAYCARE CENTERS IN PHASE I STUDY
Daycare Center
Code
Number of Children
Age Group
Type
D01
118
3-4 years old
Head Start
D02
120
4-5 years old
Head Start
D03
106
4-5 years old
Head Start
D04
60
6 weeks to 12 years old
Private
D05
112
6 weeks to 12 years old
Private
D06
374
3-4 years old
Head Start
D07
65
6 weeks to 5 years old
Private
D08
200
2 months to 12 years old
Private
D09
86
6 weeks to 5 years old
Private
D10
99
2-6 years old
Private
26
-------
TABLE 4.2. FIELD MONITORING ACTIVITIES IN EACH DAYCARE CENTER
Monitoring Day
Activity
DAY -1
1.
Pick up signed consent form
2.
Collect copies of food menus
3.
Collect used vacuum bag, if available
4.
Conduct pre-monitoring interview
5.
Give instructions on food sample collection
6.
Remind the participant: next appointment time, no vacuum
inside the classroom
7.
Complete the building observation survey
8.
Set 14) indoor air monitor, marie the location on the sketch
9.
Set up outdoor air monitor, mark the location on the sketch
10.
Take indoor/outdoor temperature (at the location of the air
monitor)
11.
Take pictures of sampling activities
DAY-2
1.
Check the air flow of the air monitors
2.
Record indoor/outdoor temperature
DAY-3
1.
Unload indoor air samplers, remove air monitors
2.
Record indoor/outdoor temperature
3.
Collect dust samples, vacuum the classroom
4.
Unload outdoor air samplers, remove air monitors
5.
Collect 1 soil sample (selected classroom children's play
areas), mark the location on the sketch
6.
Conduct post-monitoring interview
7.
Collect food samples, examine the samples
8.
Present the certificate of appreciation to the participant
9.
Confirm the mailing information of the check ($50) with
the participant
10.
Take pictures of sampling activities
27
-------
Sampling Methods
Indoor and outdoor air were sampled simultaneously during a 48-hour period at each
daycare center. A 4 L/min URG sampler was used. Each sampler consisted of an inlet jet
equipped with an impactor for <10 jam particle collection, a filter (quartz fiber or Teflon-coated
glass) and a glass cartridge packed with either XAD-2 or a PUF plug. The indoor samplers were
placed in a Styrofoam sampling box equipped with a cooling fan. The sampling box was placed
in a playpen in the classroom to be sampled. A stroller net was used to cover the top of the
playpen to keep children from playing with the sampling module. A real-time PAH monitor was
also placed in the playpen to monitor particle-bound PAH indoors (8,9). The outdoor sampling
modules were placed in a dog house, purchased new for this use, outside the daycare center.
After 48-hour sampling, the sample module was sealed at both ends, wrapped with aluminum
foil, and placed in a cooler ready for transporting back to the laboratory for analysis. The initial
(Day-1) and final (Day-3) flow rate of each sampler was recorded; sampler flow rate was also
checked and recorded at the mid-point (Day-2) of the sampling period.
The floor dust samples were collected using the High Volume Small Surface Sampler
(HVS3) in designated areas in the classroom where the children's highest play activities occur.
The HVS3 unit was operated following an ASTM Standard Method and the manufacturer's
manual (10,11). The vacuum bag samples from the same classrooms in seven daycare centers
were also collected. The bag samples were not obtained from the remaining three daycare
centers because they were not available. The dust samples were collected and transferred to
clean jars, and placed in a cooler for shipping.
The playground soil samples were collected from each daycare center. According to the
procedures used in the previous EPA studies (5), the playground soil was scraped from the top
0.5-cm of soil. A minimum area of 0.093 m2 (1 ft2) was scraped with a stainless steel spatula,
and the soil was placed in a clean jar.
The duplicate diet food samples were collected from each daycare center. A training
session and a food collection protocol were provided to the daycare center staff before the field
sampling activities began. At each daycare center, two servings of all food served for the
28
-------
children was collected during the 2-day monitoring period. One was placed in a glass container
and one was placed in a plastic container.
In addition to the field samples, field blanks in all sample media were collected as quality
control samples. The field blank for air samples was a filter/XAD-2, or filter/PUF, that was
processed through field handling and shipping together with the field samples without sampling
air. The field blanks for dust, soil, and food samples were the empty containers that were used
for the respective samples that were processed through field handling and shipping.
Analytical Methods
The analytical methods used to analyze the samples (air, dust, soil, and food) collected in
Phase 1 are summarized in Table 4.3.
The filter/XAD-2 samples and field blanks were spiked with known amounts of surrogate
recovery standards (SRS), and extracted with dichloromethane (DCM) by the soxhlet technique.
The SRS included pyrene-d10, chrysene-d12, DDE-13C, DDT-I3C, 2,2',4,5,5-pentachloro-
biphenyl-I3C, and fenchlorfos. The DCM extract was concentrated by Kuderna Danish (K-D)
evaporation and divided into two portions. Portion I was analyzed by GC/MS operated in the
electron impact (EI), selected ion monitoring (SIM) mode for target PAH, PCB, PE, Ph, OC, and
OP. Three GC/MS analyses were performed for each sample extract, one for PAH and PE, one
for PCB, and one for OC and OP. Portion II was methylated and analyzed by GC/MS for
pentachlorophenol (PCP).
The filter/PUF samples were analyzed for 2,4-D. The results of the method evaluation
tests showed that 2,4-D was retained in the filter, and thus only the filter samples were analyzed
for 2,4-D. The analytical method used for 2,4-D analysis was developed and validated in a
previous study (2). The filter sample was extracted with a phosphate buffer solution in a
sonication bath, and the extract was processed through a CI8 SPE column. The target fraction
was concentrated, methylated, and analyzed by GC and ECD for 2,4-D.
29
-------
TABLE 4.3. SUMMARY OF ANALYTICAL METHODS FOR EACH TYPE OF
SAMPLES FROM PHASE 1 STUDY
Media
Analytical Method
Analyte Class
Indoor/ Ambient
Air
Dust/Soil
Food
Soxhlet extraction with DCM, GC/MS
analysis
Sonication with phosphate buffer and C18
SPE cleanup, methylation, GC/ECD analysis
Sonication with 10 % EE in hexane, Florisil
SPE cleanup, GC/MS analysis
Sonication with phosphate buffer, C18 SPE
cleanup, methylation, GC/ECD or GC/MS
analysis
Sonication with 5 % acetic acid in methanol,
DCM, and 5 % acetic acid in water,
methylation, GC/MS analysis
Hobart chopper, Brinkman Homogenization
DCM extraction, GPC, and Florisil SPE
cleanup, GC/MS analysis
Hobart chopper, Brinkman Homogenization
phosphate buffer extraction, liquid-liquid
partitionation, CI8 SPE cleanup, GC/MS
analysis
PAH, PCB, PE, Ph, OC,
OP
HA
PAH, PCB, PE, OC,OP
HA
Ph
PAH, PCB, PE, Ph, OC,
OP
HA
The dust samples were separated into coarse and fine (<150 |im) fraction and only the
fine fractions of the dust samples were subjected to subsequent analysis. The soil samples were
not separated into fine and coarse fractions. An aliquot of each dust and soil sample was spiked
with known amounts of SRS, and extracted twice each with 10 mL of 10% EE in hexane in a
sonication bath for 20 minutes. The sample extract was combined and concentrated to 1 mL. A
Florisil SPE column was conditioned with 50% EE in hexane, followed by 100% hexane. The
concentrated sample extract was then applied to the column and eluted with 12 mL of 15% EE in
hexane. This fraction was concentrated and analyzed by GC/MS for target PAH, PCB, PE, OC,
and OP.
30
-------
A second aliquot of each dust and soil sample was used for analysis of phenols. The
dust/soil samples were sequently extracted with 5% acetic acid in methanol, DCM, and 5% acetic
acid in distilled water. The mixture was washed with distilled water and the DCM extract was
concentrated. One portion of the DCM extract was methylated and analyzed for PCP and the
other portion was not methylated and was analyzed for other target phenols.
A third aliquot of each dust and soil sample was used for analysis of 2,4-D acid. The
analytical method used was developed from the previous EPA study (2). In brief, the dust/soil
sample was extracted with phosphate buffer, with cleanup by C18 SPE column, and analysis by
either GC/ECD (soil) or GC/MS (dust).
The solid and liquid food samples collected in glass containers from all daycare centers
were extracted separately and analyzed for target POP. Only one solid and one liquid food
sample collected in plastic containers from daycare D09 were prepared for target POP analysis.
The analytical procedures used for food samples are described in Appendix A. In brief, an
aliquot of the homogenized food sample was spiked with known amounts of an analytical
surrogate recovery standard (SRS), extracted with DCM, fractionated by GPC, and subjected to
final cleanup by Fluorisil SPE. The target fraction was concentrated, and analyzed for target
PAH, PCB, PE, OC, and OP.
A second aliquot of the food sample was prepared for the determination of 2,4-D using
the method for 2,4-D in dust/soil described earlier with minor modification. The food sample
was spiked with known amounts of SRS (3,4-D) and extracted with phosphate buffer. The
sample extract was partitioned with hexane to remove fat. The mixture had to be centrifuged to
separate into two layers because of the emulsion problems. The aqueous layer was then acidified
and applied to a CI8 SPE column. The SPE column was soon plugged up by the sample extract.
Typically, 4 to 5 SPE columns were used for one sample extract. The target fractions from the
SPE columns of the same sample were combined, concentrated, methylated, and analyzed by
GC/MS for 2,4-D.
The sample extracts and standard solutions were analyzed by 70 eV electron impact (EI)
GC/MS. A Hewlett-Packard (HP) GC/MS was operated in the SIM mode. Data acquisition and
31
-------
processing were performed with an Chem Station data system. The GC column was a DB-5
fused silica capillary column (60 m x 0.32 mm; 0.25 nm film thickness). Helium was used as the
GC carrier gas. Following injection, the GC column was held at 70°C for 2 min and then was
temperature-programmed to 150°C at 15°C/min and then to 290°C at 6°C/min. The MS was
operated in the SIM mode. Peaks monitored were the molecular ion peaks and their associated
characteristic fragment ion peaks. Identification of the target analytes was based on their GC
retention times relative to those of corresponding internal standards and relative abundance of the
monitored ions. Quantification of target analytes was based on comparisons of the integrated ion
current responses of the target ions to those of the corresponding internal standards using average
response factors of the target analytes generated from standard calibrations.
The detection limits of target POP were estimated according to the analytes' signal to
noise ratios from the lowest levels of standard calibrations. The estimated detection limits of
target POP expressed in units corresponding to each sample medium are summarized in
Table 4.4.
Data Analysis
The types of samples collected in the Phase 1 study were indoor air, outdoor air, carpet
dust, playground soil, liquid food, and solid food samples. Target POP included multiple
compound classes: PAH, phthalate esters (PE), phenols (Ph), polychlorinated biphenyl (PCB),
organochlorine (OC) pesticides, organophosphate (OP) pesticides and a herbicide acid (HA). A
total of 57 individual target compounds and seven sums of target compounds in the respective
compound classes were included in the analysis. The sums of target compounds in their
respective compound class were calculated in a way that treated below detection limit (BDL)
values as zeros and added up the non-BDL values.
Summary statistics (sample sizes, number of samples below detection limit, mean,
standard deviation, minimum and maximum) of the target POP and the sums of the target POP of
32
-------
TABLE 4.4. ESTIMATED DETECTION LIMITS OF TARGET PERSISTENT ORGANIC
POLLUTANTS IN MULTIMEDIA SAMPLES
Estimated Detection Limit
Analyte
Class
Air
Dust/Soil
Food
Dermal Wipe
Urine
ng/m3
ppm
ppb
ng/wipe
ng/mL
PAH
0.01-0.04
0.001
0.04
0.5
NA
PCB
0.04
0.001
0.04
0.5
NA
PE
0.04
0.001
0.04
0.5
NA
Ph
0.1
0.001
0.1
NA
0.03
OC
0.1
0.001
0.04
0.5
NA
OP
0.1
0.001
0.04
0.5
NA
2,4-D
0.1
0.001
0.2-0.5
NA
0.03
OH-PAH
NA(a)
NA
NA
NA
0.02-0.03
3,5,6-TCP
NA
NA
NA
NA
1.0
(a) NA denotes not applied.
33
-------
the seven compound classes were determined. Pearson correlation coefficients were calculated
based on natural log transformed data. Note that half of the detection limit was used for the
measurement that was BDL when calculating means, standard deviations, and the Pearson
correlation coefficient. Average of old and new carpet samples for the floor dust HVS3
measurement from D09 was used in the correlation analysis.
The estimates of POP exposure through multiple pathways, inhalation, nondietary
ingestion, and dietary ingestion were determined. The estimates of POP exposure through
inhalation were based on the time-weighted concentration, indoor and outdoor POP levels, the
amount of time spent indoors and outdoors, and the portion of 24-hr day that the child spent at
the daycare center. The time spent indoors and outdoors was based on the schedule provided by
each daycare center. The total operation time of each daycare center varied from 6 to 12 hours.
For the calculation, we assumed that the child spent the entire period at the daycare center. The
maximum exposure estimates from inhalation pathway was based on the following equation:
inh t + t 1000
i o
where
Einh = estimates of daily POP exposure through inhalation at the daycare center,
Hg/day
C; = indoor POP concentration at the daycare center, ng/m3
Ca - outdoor POP concentration at the daycare center, ng/m3
tj = child's time spent indoors at the daycare center, min
t0 = child's time spent outdoors at the daycare center, min
V = the estimated child ventilation rate, 15 m3/day
F = fraction of the 24-hr day child spent at the daycare center.
34
-------
The maximum POP exposure estimates from nondietary ingestion pathway were based on
HVS3 carpet dust and playground soil POP levels, the amount of time spent indoors and
outdoors, and the portion of the 24-hr day that child spent at the daycare center. The equation is
described as follows:
where
En =
estimates of daily POP exposure through nondietary ingestion at the daycare
center, jag/day
D, »
POP concentration in the floor dust, jig/g
Po ¦
POP concentration in the playground soil, |ig/g
ti
child's time spent indoors at the daycare center, min
to
child's time spent outdoors at the daycare center, min
M
child's estimated daily dust/soil intake, 0.1 g
F
= fraction of the 24-hr day child spent at the daycare center.
The maximum exposure estimates from dietary ingestion pathway were based on the POP
levels in solid and liquid food samples, and the amounts of the liquid and solid food samples
collected. The following equation is used for the calculation:
S*M+L* M,
p ' S I 6
Ed =
where
Ed = estimates of daily POP exposure through dietary ingestion at the daycare
center, |ig/day
Sj = POP concentration in the solid food samples, ng/kg
35
-------
Lj = POP concentration in the liquid food sample, |ig/kg
Ms = the total amount of the solid food samples collected during 48-hr, kg/2 days
Me = the total amount of the liquid food samples collected during 48-hr, kg/2 days.
We assumed a ventilation rate of 15 m3/day for children (12,13) and the dust/soil ingestion rate is
0.1 g/day for children (14,15).
Phase 2 Field Study in Two Daycare Centers and Nine Homes
The Phase 2 study was conducted in two daycare centers (D03 and D09) and in nine
homes. Nine children were recruited from the two daycare centers; four from D03 and five from
D09. Multi-media samples were collected at daycare centers and at subjects' homes during a
48-hours sampling period. Dermal wipe and urine samples were also collected from each subject
at daycare centers and at the homes. Informed consents were obtained from both the directors of
the Daycare centers and the parents of the subjects before any field activities were occurred.
This section summarizes the procedures used for selection of nine subjects, the sampling
protocols, and the analytical procedures. The recruiting materials, consent forms, pre-monitoring
questionnaire, post-monitoring questionnaire, house/building characteristics observation survey
and child activity diary used in the Phase 2 study are presented in Appendix C.
Selection of Homes
After reviewing the POP data obtained from the Phase 1 study, and consulting with the
EPA project officer, it was decided to select daycare D03 (Head Start) and D09 (private) for the
Phase 2 study. The directors from both daycare centers agreed to participate in the Phase 2
study. A study information package was developed for recruiting children from these two
daycare centers. Included in the package were a cover letter, a participation summary sheet, and
a brief home survey. Forty packages were delivered to these two daycare centers and were
distributed by the center directors to the selected classrooms (children aged two to five). During
-------
the first two weeks, only two responses were received, far fewer than the 9 subjects needed for
the study. A more aggressive recruiting campaign was carried out after the approval of each
center's director. New and simplified study flyers were developed and sent to the selected
classrooms. Battelle staff members also spent time at the daycare centers during the time when
parents pick up their children and discussed with parents the potential involvement in the study.
With this aggressive recruiting approach, eleven parents responded to our recruiting flyers. One
child from each home was involved in the study. Among these subjects, five attend daycare D03
and six attend daycare D09. All these children spend at least 25 hours per week in the daycare
centers. Therefore nine selected homes and two backup homes were scheduled for Phase 2 field
sampling activities. The backup home in the D03 group was used to replace an originally
selected home, because we were unable to contact the participant the day before field sampling
activities began.
Field Monitoring Activities
The Phase 2 field sampling activities were performed in June 1997. Table 4.5 summarizes
the information on the two daycare centers and nine homes in the Phase 2 study. Figure 4.2 shows
the locations of these nine homes and two daycare centers. The sampling activities were
completed during a period of two weeks. Figure 4.3 displays the sampling schedule for the Phase
2 study. In the first week, the field sampling activities were conducted in D03 and in the four
homes of the subjects who attend D03. The four subjects were divided into two groups based on
the classrooms they attend. Subjects HA3 and HB3 are in the same classroom (group 1) and
subjects HC3 and HD3 are in another classroom (group 2). As shown in Figure 4.3, group 1
homes were monitored during the first 48-hr period and followed by group 2 homes. Because all
the children's parents work during the day, the appointment time for setting up monitoring
activities was usually after 6:30 pm. Two field teams were involved in the Phase 2 study. The
field monitoring activities for each home are similar to the Phase 1 study (Table 4.2) except that
additional biological samples (dermal wipe and urine) were collected. The field monitoring
activities of the daycare centers were modified slightly from the Phase 1 study. The
-------
TABLE 4.5 SUMMARY OF SUBJECTS/HOMES IN PHASE 2 STUDY
Subject
Code
Daycare Center
Classroom No.
Adult Smoker
in Home
Subject's
Sex, Age
Subject's
Race
Low-Income
Family
HA3
D03
Classroom #2
None
Girl, 4
Hispanic
Yes
HB3
D03
Classroom #2
None
Boy, 4
Black
Yes
HC3
D03
Classroom #1
l
-------
u>
vo
O.xfohdii^n e n d
-------
WEEK 1
Daycare Center: D03
Day 1 Day 2 Day 3 Day 4 Day 5
Home Group 1: HA3 and HB3
Day 1 Day 2 Day 3
Home Group 2: HC3 and HD3
Day 1 Day 2
I l
WEEK 2
Daycare Center: D09
Day 1 Day 2 Day 3 Day 4 Day 5
I I I I I
Home Group 1: HE9 and HF9
Day 1 Day 2 Day 3
I I I
Home Group 2: HG9, HH9, and HI9
Day 1 Day 2 Day 3
I I !
Figure 4.3. Sampling Schedule for the Phase 2 Field Study.
40
Day 3
I
-------
three-day activities (Phase 1) were modified to five days to cover the two groups of subjects.
Note that only one indoor air sampler and one outdoor air sampler were used in D03 because of
the availability of the air samplers. The two classrooms were located across a hallway from each
other. The indoor air samplers were placed in classroom #2, and two consecutive 48-hr samples
were taken. Two 48-hr outdoor air samples were also taken during the same time period. Two
floor dust samples were collected from both classrooms and only one playground soil sample
was taken. Two sets of the food samples for both subject groups were collected from D03. The
dermal wipes and urine samples were also collected from each subject at D03. In the second
week, similar field monitoring activities were carried out in daycare D09 and in five subjects'
homes. Similarly, these five subjects were divided into two groups: HE9 and HF9 in Group 1
and HG9, HH9, and HI9 in Group 2.
Sampling Methods
The sampling methods used in the Phase 2 study for air, dust, soil, and food samples were
the same as those used in the Phase 1 study. To reduce the noise levels generated from the air
samplers, insulation materials were added to the sampling box and longer vent tubes were added
to the samplers. Dermal hand wipe and urine samples were collected in the Phase 2 study.
Training sessions for the collections of hand wipe and urine samples were conducted for the
participating daycare centers' staff members and participating parents. The sample collection
protocols (Appendix C) were also provided. For each subject, two wipe samples were collected
at the daycare center before the subject ate lunch and two wipe samples were collected at the
subject's home before the subjects ate dinner on each of the two days during the 48-hr sampling
period. For each subject, four urine samples, one each morning (the first void), and one each
evening (2-3 hours after dinner or before going to bed), were collected at the subject's home.
Two urine samples, one each afternoon (after lunch) were collected at the daycare center. The
collected wipe and urine samples were stored in clean chilled coolers at daycare centers or at
homes, then in a freezer after sample retrieval until they were sent to the laboratory for analysis.
At least one field blank in each sample medium was collected as a quality control sample.
-------
Analytical Methods
The analytical methods used for air. dust, soil, and food samples collected in the Phase 2
study were the same as those used in the Phase 1 study, except for the food samples used for the
2,4-D analysis. The GPC fraction of the food sample was split into two portions. One portion
was methylated and cleaned up by Florisil SPE for the determination of 2,4-D and PCP, and the
other portion was only cleaned up by Florisil SPE for the determination of other target POP. The
detailed analytical procedures for the analysis of the food samples are described in Appendix A.
The dermal wipe samples of each subject collected at home were combined, for
extraction, as were those collected at the daycare center. Known amounts of SRS were added to
each combined wipe sample and extracted with 10 % EE in hexane by the Soxhlet technique.
The extract was dried with sodium sulfate anhydride and concentrated to 1 mL. A Florisil SPE
column was conditioned with 50 % EE in hexane followed by 100 % hexane. The sample extract
was then applied to the column and eluted with 24 mL of 15 % EE in hexane. The collected
fraction was concentrated and analyzed by GC/MS for target POP.
Urine samples from each subject collected at the daycare center were combined, as were
those collected at home. One aliquot (30 mL) of each composite sample was used for
determining 2,4-D, PCP, and hydroxy-PAH, and another aliquot (1 mL) was used for
determining 3,5,6-TCP. The analytical procedures used were the same as those described in the
method validation section. Another aliquot (5 mL) of each composite sample was sent to Smith
Kline Beecham Clinical Laboratories (St Louis, MO) for determination of creatinine.
Statistical Analysis
Statistical analyses were conducted on the following types of samples collected in
Phase 2 field studies:
• 48-hr integrated indoor and outdoor air samples
42
-------
• House dust samples collected by HVS3
Playground soil samples
• 48-hr duplicate diet food samples
• Composite dermal wipe samples
• Composite grab urine samples.
Summary statistics (sample size, number of below detection limit, mean, standard deviation,
minimum, and maximum) were determined for each of the above sample types. Four additional
types of statistical analyses were performed: Spearman correlation analyses, Pearson correlation
analyses, analysis of variance (ANOVA), and regression models. While summary statistics and
Spearman correlation coefficients were performed on the raw data, the Pearson correlation
analyses, ANOVA, and regression analyses were carried out on natural log-transformed data.
For POP concentrations less than the detection limit, half of the detection limit was used.
Summary statistics were generated for measured POP data by compound, sample
medium, and family income. In addition, summary statistics were generated on the calculated
data of estimated daily dose by compound, exposure pathway, and family income.
Spearman and Pearson correlation coefficients were calculated on Phase 2 data to
examine the relationships between POP concentrations for different sample media and the
relationship of target POP concentration in each compound class within the sample medium.
An analysis of variance (ANOVA) model was employed to examine the effects of
sampling location (home vs. daycare center) and family income (low-income vs. middle-income)
on POP concentrations for each sample medium. The neutral log-transformed data in Phase 2
multimedia samples (air, dust, soil, food, wipe, and urine) were used for ANOVA. Urine data
calculated in two units, ng/mL and ^mole/mole, were fitted separately to the ANOVA models.
Regression models were employed to examine the relationships between urinary
metabolites and concentrations of corresponding POP in multimedia samples, as well as between
the urinary metabolites and estimated total daily POP dose levels. For each measurement unit
43
-------
(ng/mL and ^mole/mole), two different types of regression models were fitted to the data of
measured urinary metabolites. The first type of model included factors for sampling location (at
daycare center versus at home) and POP concentrations in multimedia samples. The POP data in
playground soil were not included in the model, because POP concentrations in soil were low,
and initial analyses showed that soil was not a significant factor for urinary metabolite
concentrations. The second type of model included factors for sampling location and total
estimated daily POP doses. The total estimated daily POP doses were computed as the sums of
the estimated daily potential POP doses from inhalation, nondietary ingestion, and dietary
ingestion that resulted from the exposures at homes and at daycare centers. Separate models
were fitted to selected hydroxy PAH, PCP, 2,4-D, and 3,5,6-TCP data in subjects' urine samples.
All regression models were fitted to the log-transformed data.
Estimates of Daily Potential Persistent Organic Pollutant Doses
The exposure values (ng/day) for inhalation and ingestion (dietary and nondietary) can be
converted to units of potential dose by assuming 100% absorption in the lung and stomach, and
normalizing for body mass. Various factors can be found in the literature to account for physical,
chemical, and/or physiological processes. For maximum estimates, this conversion gives upper
limits on POP available for delivery to target organs (lung or stomach). The potential daily dose
of PAH in ng/kg/day was estimated using the following equations:
D
C * t+ C * t
it o i
* — * f
W
ink
t * D. + t * P
M * 1000
* F
W
D
i i a
o
n
t. + t
D
Cf * Mf * 1000
* F
d
44
-------
where
Djnh = estimated dose through inhalation at home or at daycare center, ng/kg/day
Dn = estimated dose through nondietary exposure at home or at daycare center,
ng/kg/day
Dd = estimated dose through dietary exposure at home or at daycare center,
ng/kg/day
W = the measured body weight of the subject, kg
Q = indoor POP concentration at home or at daycare center, ng/m3
C0 = outdoor POP concentration at home or at daycare center, min
tj = subject's time spent indoors at home or at daycare center, min
t0 = subject's time spent outdoors at home or at daycare center, min
V = the estimated subject's ventilation rate, 15 mVday
Dj = PAH concentration in the floor dust at home or at daycare center, jig/g
P0 = PAH concentration in the pathway soil at home or at daycare center, fig/g
M = subject's estimated daily dust/soil intake, 0.1 g
Cf = POP concentration in the daily food samples at home or at daycare center,
Hg/kg
Mf = the daily mass of food intake at home or at daycare center, kg/day
F = Fraction of the time the subject spends at home or at daycare center.
We assumed a ventilation rate of 15 m3/day for child subjects (12,13) and that the
dust/soil ingestion rate is 0.1 g/day for the child (14,15). The body weight of each subject was
measured during the Phase 2 field study.
45
-------
Chapter 5
Results and Discussion
Method Validation for Multimedia Samples
Air Samples
The recovery data for spiked persistent organic pollutants (POP) from the first experiment
are listed in Tables 5.1 and 5.2. The results show that for the filter/XAD-2 sampling modules,
quantitative recoveries were obtained for target POP except for phthalate esters (PE), under all
storage conditions tested. This finding also demonstrated, by extension, that the XAD-2 sorbent
can retain the target POP under the 48-hr sampling conditions used. Greater than 100 percent
recoveries of PE are partly due to the difference between spiked and native air levels of these
compounds, with the amount spiked being approximately 5 times lower than that collected on the
non-spiked sampling modules. In addition, sample handling and processing also contribute to
background levels of PE. As seen in Table 5.2, quantitative recoveries of most POP were also
obtained from filter/PUF samples under the storage and sampling conditions tested here. Low
recovery of acenaphthene-d]0 from PUF was anticipated, as this sorbent cannot retain 3-ring PAH
due to their relatively high volatility (16). The low recovery of bisphenol-A was probably due to
incomplete extraction from the PUF; the extraction solvent for PUF (10% EE in hexane) has a
much lower solvent strength (0.04) than DCM (0.42). There was no evidence of loss or
degradation of the POP over the 20-day storage interval.
The retention efficiencies for the 13 non-coplanar PCB congeners on XAD-2 are listed in
Table 5.3. As seen there, recoveries were quantitative: 90-120%.
46
-------
TABLE 5.1. RECOVERY OF TARGET PERSISTENT ORGANIC POLLUTANTS
FROM FILTER/XAD-2 SAMPLES
Recovery, %(a)
Target Compound Day-0 Day-10 Day-20
PAH
Ac enaphthene-d j „
91 ±9
94 ±2
107
Pyrene-d10
85 ±6
83 ±3
92
Chrysene-d12
82 ± 1
81 ± 1
91
Benzo[k]fluoranthene-d 12
86 ±6
90 ± 3
110
OC Pesticide
Lindane
106± 18
129 ± 1
120
Aldrin
82 ± 1
90 ± 5
107
p,p'-DDE
89 ±0
85 ±2
100
Dieldrin
100 ±5
98 ±3
103
Endrin
93 ±2
112 ± 1
137
p,p'-DDT
94 ±5
120 ±4
118
OP Pesticide
Diazinon
82 ± 12
125 ±3
113
Chlorpyrifos
81 ±8
112 ± 3
134
PCB (Coplanar)
3,3', 4,4'-tetrachlorobiphenyl (77)
89 ±5
79 ± 1
85
3,3' 4,4'5-pentachlorobiphenyl (126)
83 ±3
81 ± 1
90
2,2'4,4'6,6'-hexachlorobiphenyl (155)
85 ± 1
84 ± 0
101
PE (Phthalate Ester)
Di-n-butylphthalate
96 ±49
83 ±0
113
Benzylbutylphthalate
159 ±90
83 ±29
440
Ph (Phenols)
Bisphenol-A
77 ± 13
83 ± 12
111
Pentachlorophenol
85 ±7
89 ±3
83
(a) Day-0, Day-10, and Day-20 represented 0-, 10-, and 20-day storage intervals after sampling.
Data are from triplicate Day-0 samples, duplicate Day-10 samples, and one Day-20 sample.
47
-------
TABLE 5.2. RECOVERY OF TARGET PERSISTENT ORGANIC POLLUTANTS
FROM FILTER/PUF SAMPLES
Recovery, %(a)
Target Compound Day-0 Day-10 Day-20
PAH
Acenaphthene-d10 5 ± 1 5 ± 1 5
Pyrene-d10 101 ±0 99 ±3 79
Chrysene-d|2 91 ± 1 102 ±2 90
Benzo[k]fluoranthene-d12 88 ± 4 107 ± 1 76
OC Pesticide
Lindane 99 ±10 110 ±4 104
Aldrin 70 ±27 89 ±14 82
p,p'-DDE 101 ±2 98 ±2 96
Dieldrin 101 ±2 92 ± 2 88
Endrin 131 ±7 121 ±3 120
p,p'-DDT 140 ±7 112 ±3 114
OP Pesticide
Diazinon 102 ±7 119 ±4 78
Chlorpyrifos 115 ±5 116 ±4 87
PCB (Coplanar)
3,3', 4,4'-tetrachlorobiphenyl (77) 97 ±1 85 ±2 81
3,3' 4,4'5-pentachlorobiphenyl (126) 83 ± 6 90 ± 1 88
2,2'4,4'6,6'-hexachlorobiphenyl (155) 97 ±3 100 ±2 97
PE (Phthalate Ester)
Di-n-butylphthalate 38 ± 13 129 ±17 94
Benzylbutylphthalate
Ph (Phenols)
Bisphenol-A 38 ± 6 53 ± 1 36
Pentachlorophenol 93 ± 1 98 ± 0 69
(a) Day-0, Day-10, and Day-20 represented 0-, 10-, and 20-day storage intervals after sampling.
Data are from triplicate Day-0 samples, duplicate Day-10 samples, and one Day-20 sample.
48
-------
TABLE 5.3. RECOVERY OF NON-COPLANAR POLYCHLORINATED BIPHENYLS
FROM FILTER/XAD-2 SAMPLES
Compound(a)
Recovery, % ^
2,4,4'-Trichlorobiphenyl (28)
96 ±0
2,2',5,5' -Tetrachlorobiphenyl (52)
95 ± 1
2,2',3,5'-Tetrachlorobiphenyl (44)
100 ±2
2,3',4',5-Tetrachlorobiphenyl (70)
108 ± 1
2,2' 3,5', 6-Pentachlorobiphenyl (95)
92 ±1
2,2' ,4,5,5' -Pentachlorobiphenyl (101)
112 ± 2
2,2'3,4,5'-Pentachlorobiphenyl (87)
119 ± 1
2,3,3',4',6-Pentachlorobiphenyl (110)
114 ± 2
2,3',4,4',5-Pentachlorobiphenyl (118)
111 ±2
2,3,3 '4,4'-Pentachlorobiphenyl (105)
108 ±2
2,2'4,4',5,5'-Hexachlorobiphenyl (153)
110± 1
2,2'3,4,4'51 -Hexachlorobiphenyl (138)
103 ±2
2,2 '3,4,4 ',5,5 '-Heptachlorobiphenyl (180)
119 ± 0
(a) The number in the paretheses are the PCB congener's numbers.
(b) Data are from duplicate Day-0 samples.
49
-------
Table 5.4 summarizes the results of experiments for retention and recovery of 2,4-D and
PCP from Teflon-coated glass filters and PUF. As shown there, the extraction and cleanup
methods can be applied equally well to PCP and 2,4-D, and 3,4-D serves adequately as an
analytical surrogate recovery standard (SRS) for each analyte in assessing analytical losses.
Relative to the analytical method, 2,4-D recovery from the filter following air sampling appeared
to be quantitative (83 vs 80%) despite the 3,4-D recovery being slightly lower than expected
(66 vs 73%). No 2,4-D was detected in the PUF extract. In contrast, PCP appears to have
moved from filter to PUF during the 48-hr sampling interval, with negligible amounts found on
the filter after sampling. On the basis of the 3,4-D recovery, we surmised that the PCP was
probably completely recovered on the PUF, with the loss being due to analytical losses, rather
than sampling losses. However, on the basis of a single set of samples for method validation, it
seems more prudent to approach PCP sampling as collection on XAD-2 and extraction with
DCM prior to analysis.
According to the results from the above method validation tests, we then used a quartz
fiber filter/XAD-2 as a sampling module for collection of target POP including PAH, OC, OP,
PCB, PE and Phenols in air. We used a Teflon-coated glass fiber filter/PUF to collect 2,4-D.
Dust and Soil Samples
The recovery data for spiked POP in house dust and soil samples are presented in
Table 5.5 and 5.6, respectively. Because the analytical methods for determining phenols and
2,4-D were validated from previous studies (2,7), we therefore excluded phenols and 2,4-D acid
from these experiments to simplify the sample matrix effect. Quantitative recoveries of most
target POP from house dust were obtained when hexane and 10% EE in hexane were used as the
extraction solvent. Recoveries of chlorpyrifos and phthalate esters were difficult to assess in the
low-level spike experiments because of the high levels of these compounds were present in the
nonspiked house dust. Levels of chlorpyrifos and benzylbutyl phthalate in the nonspiked dust
were approximately 20 times and over 100 times higher, respectively, than the spike level. For
the high-level spike experiments, quantitative recoveries of chlorpyrifos were obtained.
50
-------
TABLE 5.4. RECOVERY OF 2,4-D AND PENTACHLOROPHENOL FROM
FILTER/PUF SAMPLES
Recovery, %
Compound
Filter
PUF
Method Spike(a)
PCP
74 ±4
70 ±2
2,4-D
3,4-D
83 ± 14
73 ±5
80 ± 1
82 ±2
Day-10 Air Samples^
PCP
2,4-D
3,4-D
2 ± 1
80 ±5
66 ±3
49 ±3
0±0
51 ±5
(a) Known amounts of PCP, 2,4-D, and 3,4-D were spiked onto filters and PUF plugs for
method spike samples. Data were from triplicate filter and PUF samples.
(b) Known amounts of PCP and 2,4-D were spiked onto filter samples but not PUF plugs prior
air sampling, and known amounts of 3,4-D were spiked onto both filter and PUF samples
before sample extraction. Data were from triplicate filter and PUF samples.
51
-------
TABLE 5.5. RECOVERY OF TARGET PERSOSTEMT ORGANIC POLLUTANTS FROM SPIKED HOUSE DUST SAMPLES
Target Compound
Recovery, %(,)
10% EE in
10% EE in
Hexane(H)
Hexane(L)
Hexane(H)
Hexane (L)
DCM (L)
none
none
Floirsil-SPE
Floirsil-SPE
Silica-SPE
PAH
Acenaphthene-d,0
97 ±29
107 ± 11
86 ± 1
89 ±2
NA(c)
Pyrene-d,0
90 ±30
96 ±7
85 ± 1
91 ±2
NA
Chrysene-d,2
85 ±24
97 ±7
81 ±2
98 ±2
NA
Benzo[k]fluoranthene-d|2
94± 11
103 ±6
110 ±24
111 ± 1
NA
OC Pesticide
Lindane
118 ±28
114 ± 8
100 ± 1
105 ±2
NA
Aldrin
105 ± 34
105 ±7
93 ± 1
101 ±2
NA
p,p'-DDE
95 ±29
100 ± 12
87 ±0.1
95 ±2
NA
Dieldrin
140 ± 10
125 ± 23
96 ± 1
130 ±3
NA
Endrin
115 ± 20
115 ± 16
92 ±2
114 ±4
NA
p,p'-DDT
136 ±25
91 ±4
101 ± 1
107 ±5
NA
OP Pesticide
Diazinon
100 ± 29
93 ± 20
88 ±2
86 ± 10
NA
Chlorpyrifos
24 ±4
ND(b)
91 ± 1
273 ± 68
NA
PCB (Coplanar)
3,3', 4,4'-tetrachlorobiphenyl (77)
95 ±37
77 ±7
84 ± 1
94 ±2
NA
3,3' 4,4'5-pentachlorobiphenyl (126)
84 ± 18
101 ±8
83 ± 1
91 ±2
NA
2,2'4,4'6,6'-hexachlorobiphenyl (155)
86 ± 19
91 ±6
81 ±0.3
90 ±4
NA
PE (Phthalate Ester)
Di-n-butylphthalate
90 ±40
150 ±27
ND
66 ± 16
NA
Benzylbutylphthalate
ND(b)
ND(b)
ND
188± 162
NA
(a) Data were from triplicate spiked and non-spiked dust samples; H denotes high-level spike (500ng/compound or 1000 ng/compund) and L denotes low-level
spike (100 ng/compound).
(b) ND denotes negative recovery data were obtained because of high levels of these compounds present in the dust sample.
(c) NA denotes that the sample extracts were not analyzed because a precipitate formed during the concentration step after Silica SPE cleanup.
-------
TABLE 5.6. RECOVERY OF TARGET PERSISTENT ORGANIC POLLUTANTS FROM SPIKED SOIL SAMPLES
Recovery, % m
Hexane(H)
Hexane (L)
10% EE in Hexane (L)
DCM (L)
Target Compound
none
none
Floirsil-SPE
Silica-SPE
PAH
Acenaphthene-dI0
103 ± 1
110± 15
85 ±2
54 ±3
Pyrene-d10
110 ± 8
113 ± 6
94 ±5
58 ±3
Chrysene-d,2
110 ± 2
103 ±5
95 ±7
71 ±3
Benzo[k]fluoranthene-dl2
105 ±2
97 ±7
98 ±5
111 ± 10
OC Pesticide
Lindane
78 ±3
77 ±9
94 ±2
94 ±4
Aldrin
103 ± 1
100 ± 15
94 ±2
82 ±5
p,p'-DDE
110 ± 8
127 ± 22
95 ±2
64 ±3
Dieldrin
108 ±3
116 ± 19
109 ±4
88 ±3
Endrin
112 ± 2
103 ±21
102 ±5
123 ± 12
p.p'-DDT
103 ±6
69 ±7
107 ±3
108 ±8
OP Pesticide
Diazinon
59 ±4
19 ± 1
86 ±6
77 ± 2(b)
Chlorpyrifos
91 ± 17
34 ±4
81 ±8
97 ±3
PCB (Coplanar)
3,3', 4,4'-tetrachlorobiphenyl (77)
113 ± 2
113 ± 5
94 ±4
61 ±4
3,3' 4,4'5-pentachlorobiphenyl (126)
112 ±5
109 ±8
94 ±5
68 ±3
2,2'4,4'6,6'-hexachlorobiphenyl (155)
110 ± 5
110 ± 8
94 ±5
76 ±4
PE (Phthalate Ester)
Di-n-butylphthalate
107 ±5
76 ± 13
111 ±37
68 ±5
Benzylbutylphthalate
108 ±6
97 ±7
99 ±6
92 ±7
(a) Data were from triplicate spiked and non-spiked soil samples; H denotes high-level spike (500 ng/compound or 1000 ng/compound) and L denotes low-
level spike (100 ng/compound).
(b) Data were the sum of the recovery data of hexane/DCM and DCM fractions.
-------
For the same reason mentioned above, the recovery data of PE cannot be obtained because native
PE levels were still much higher than the high-level spike (1000 ng/compound). The fractions of
the dust samples generated from DCM extraction were not analyzed by GC/MS, because a
yellow participate formed during the concentration step after the silica SPE cleanup step. In
general, satisfactory recoveries for spiked POP were obtained from house dust when 10% EE in
hexane was used as the extraction solvent, followed by a Florisil SPE clean up step.
Quantitative recoveries were obtained for most POP from pathway soil for all the
experiments. At the lower spike level (100 ng/compound), the recovery of chlorpyrifos was 34%
when hexane was used and the recovery improved to greater than 81% when the more polar
solvent 10% EE in hexane was used. Satisfactory recoveries (68 to 111%) of PE from the soil
sample were obtained. This is mainly because the native levels of PE in the soil sample were in
the same range as the spike level. Quantitative recoveries (>80%) of all target POP were
observed when 10% EE in hexane was used as the extraction solvent, followed by Florisil-SPE
clean-up step.
From the results of the above experiments, we decided to use the analytical method
validated here to determine all target POP except phenols and 2,4-D for both Phase 1 and 2
dust/soil samples. This method consists of (1) sonication of the sample with 10% EE in hexane;
(2) clean-up of the extract using an SPE cartridge; and (3) analysis of the fraction by GC/MS.
An analytical method was developed from the previous study to determine all target
phenols except PCP in dust and soil (7). The spiked dust samples were analyzed for target
phenols using the above analytical method with miner modifications (as described in Chapter 4
experimental section). Quantitative recoveries were obtained from the triplicate spiked samples.
The recovery data were: 107±3 % for nonylphenols, 93±3 % for PCP, and 104±8 % for
bisphenol-A. Therefore, this modified analytical method was used to determine target phenols in
the Phase 1 and 2 samples. The analytical method developed for the 2,4-D analysis in dust and
soil matrix in a previous study (2) was used in the Phase 1 and 2 samples, thus no method
validation experiments were carried out for the 2,4-D analysis in dust and soil.
54
-------
Dermal Wipe Samples
The spike recovery data for the dermal wipe samples are summarized in Table 5.7. As
shown in Table 5.7, quantitative recoveries were obtained for most of the spike compounds. It
should be noted that the higher level spike experiments were conducted on real hand wipe
samples and the low-level spike experiments were conducted on clean wipes. The recovery data
of phenanthrene, and two phthalate esters could not be obtained because the background levels of
these compounds were much higher than the spiked level (20 ng/compound). For the same
reason, the recoveries of lindane could not be obtained in the low-level spike experiments. An
interference peak was eluted closely to aldrin, and the two peaks were not completely resolved
from each other in the low-level spiked sample; this could cause the recoveries greater than
100%. There is no difference in recoveries of the spiked POP between the two methods
evaluated with and without the liquid-liquid partitioning step as described in section 4.
Therefore, the analytical method used in the low-level spike experiments and the Phase 2 dermal
wipe samples did not include the liquid-liquid partitioning step. Note that target PAH and OP
pesticides were present in both fractions (15 % EE in hexane and 50 % EE in hexane) in the
high-level spike experiments. The recovery data shown in Table 5.7 are the sums of the data in
both fractions. Another eluting condition, 24 mL of 15 % EE in hexane was conducted in the
low-level spike experiments. As shown in this table, satisfactory recoveries were obtained in
both eluting conditions in the low level spike experiments. Therefore the method used in the
Phase 2 dermal wipe samples consisted of Soxhlet extraction the wipe sample with 10 % EE in
hexane, Florisil SPE cleanup with 24 mL of 15 % EE in hexane, and GC/MS analysis.
Urine Samples
Table 5.8 summarizes the spike recovery data of the spiked urine samples, and shows that
acceptable recoveries were obtained. The recoveries ranged from 75±10 % for 1-hydroxypyrene
to 99±13 % for 2,4-D in high-level spike experiments (10 ppb); from 76±34% for
1-hydroxypyrene to 110±1% for 6-hydroxyindeno[l,2,3-c,d]pyrene in medium-level spike
55
-------
TABLE 5.7. RECOVERY OF TARGET PERSISTENT ORGANIC POLLUTANTS FROM SPIKED DERMAL WIPE SAMPLES
Recovery, %(a)
High Level (20 ng/Compound) Low Level (5 ng/compound)
Target Compound a a
PAH
Acenaphthene
95 ± 14
114 ± 8
99 ± 11
Phenanthrene
ND(b)
ND(b)
ND(b)
Pyrene
105 ±7
105 ±21
94 ± 14
Benzo[a]pyrene
78 ±7
86 ±22
98 ± 8
OC Pesticides
Lindane
88 ±8
ND
ND
Aldrin
74 ±8
130 ± 51
134 ±22
p, p'-DDE
72 ±8
102 ±5
96 ±7
Dieldrin
93 ± 10
86 ±8
85 ±6
Endrin
112 ± 8
103 ± 30
128 ± 8
p,p'-DDT
107 ±6
117 ± 41
134 ±32
OP Pesticides
Diazinon
88 ± 10
103 ±24
91 ±9.6
Chlorpyrifos
102 ±8
98 ±25
94 ±23
PCB
114 ± 9
106 ±9
101 ±7
3,3',4,4'-tetrachlorobiphenyl (77)
95 ± 10
99 ± 1
104 ±23
2,2'5,5'-tetrachlorobiphenyl (52)
106 ± 12
122 ±4
113 ± 11
3,3',4,4',5-pentachlorobiphenyl (126)
89 ± 10
96 ±6
107 ±8
2,2',3,5',6-pentachlorobiphenyl (95)
ND
ND
ND
ND
ND
ND
PE (phthalate ester)
Di-n-butylphthalate
Benzylbutylphthalate
(a) A denotes that the Florisil SPE column was eluted with 12 mL of 15% EE in hexane and 12 mL of 50% EE in hexane, and B denotes that the Florisil SPE
column was eluted with 24 mL of 15% EE in hexane.
(b) ND denotes negative recovery data were obtained because of high levels of these compounds were present in the non-spiked wipe samples.
-------
TABLE 5.8 RECOVERY OF TARGET ANALYTES FROM SPIKED URINE SAMPLES
Recovery, %(a)
Target Compound
High Level
(10 ppb)
Medium Level
(1 ppb)
Low Level
(0.4 ppb)
3,4-D
99 ± 13
116 ±4
102 ±24
Pentachlorophenol
75 ±10
85 ±7
90 ± 3
1 -Hy droxypyrene
91 ±9
76 ±34
79 ± 12
6-Hydroxyindeno[ 1,2,3 -c,d]pyrene
88 ±2
110 ± 1
125 ±7
(a) Data were from duplicate spiked sampled.
experiments (1 ppb); and from 79 ±12% for 1-hydroxypyrene to 125 ±7% for
6-hydroxy[l,2,3-c,d]pyrene in low-level spike experiment (0.4 ppb). This analytical method
consisted of hydrolysis with 6N HC1, methylating with diazomethane, fractioning with Florisil
SPE column, and analyzing by GC/MS were used in the Phase 2 urine samples.
Table 5.9 summarizes the spike recovery data of 3,5,6-TCP in the spiked urine samples.
Quantitative recoveries (>90%) were obtained from the spiked urine samples at all three spiked
levels (1 ppb, 3 ppb, and 10 ppb). Thus, this analytical method was used for Phase 2 urine
samples to determine 3,5,6-TCP in urine samples. The analytical method consisted of hydrolysis
the urine sample with concentrated HC1, extracting the mixture with DCM, solvent exchanging
DCM to toluene, derivatizing the toluene extract with MTBSTFA, and analyzing the extract by
GC/MS.
Food Samples
The optimal analytical methods developed from the method evaluation experiments for
the determination of POP in PAH, PCB, PE, Ph, OC, OP, and 2,4-D in duplicate diet food
samples are described in Appendix A. In brief, the food sample was extracted with DCM, passed
57
-------
TABLE 5.9. RECOVERY OF 3,5,6-TCP FROM SPIKED URINE SAMPLES
Spike Level
Recovery of 3,5,6-TCP(a), %
lOppb
94 ±3.5
3 ppb
94 ± 4.2
lppb
93 ± 2.0
(a) Data were from triplicate spiked samples.
through GPC, applied to a Florisil SPE column, and analyzed for GC/MS for all target analytes
but 2,4-D and PCP. A portion of the GPC fraction was methylated, and fractionated by Florisil
SPE from 2,4-D and PCP analysis. Note that the evaluation experiments were conducted on the
solid samples only. We expected that the sample matrix of the liquid food is simpler than the
solid food. Thus, the analytical method developed for the solid food was modified slightly on
the extraction step for the liquid food (Appendix A). Table 5.10 summarizes the recoveries data
of target POP in the spiked solid food samples. In general, acceptable recoveries of the spiked
POP were obtained and slightly better recoveries for most target POP were observed in the
second set of samples. This is partly because we replaced the TurboVap concentration with
Kuderna Danish concentration, to minimize the losses of analytes in the concentration steps.
Note that recoveries of DDT were greater than 100% in all the spiked food samples. Further
tests were conducted by spiking known amounts of DDT into food sample matrix, and the results
showed that the GC responses of DDT in the food sample matrix were about twice of those in
clean solvent without the food matrix. This is probably due to the food sample matrix having
deactivated the surfaces of GC injector or column, which caused the responses of DDT to
increase. Recoveries greater than 100% were also observed for nonylphenols and PE. This is
mainly because of the high background levels of the non-spiked food samples. The low recovery
of bisphenol-A was partly due to the loss through Florisil SPE cleanup step. Several problems
were encountered when we prepared the food samples. The DCM extract needed to be filtered
prior to applying to GPC; very often the filter was plugged because of the sample matrix, and
58
-------
TABLE 5.10 RECOVERY OF TARGET PERSISTENT ORGANIC POLLUTANTS FROM
SPIKED SOLID FOOD SAMPLES
Target Compound
Recovery, %(1)
ii m
PAH
Acenaphthene 61 ± 3 88 ± 6 NA
Pyrene-d,0 54 ± 3 85 ± 3 NA
Chrysene-dl0 46 ±2 109 ±15 NA
Benzo[k]fluorantheDe-d12 45 ± 1 106 ±5 NA
OC Pesticide
Heptachlor 100 ±2 105 ±5 NA
Lindane 72 ± 4 89 ± 7 NA
Aldrin 76±9 90±4 NA
p,p'-DDE 84 ±5 87 ±3 NA
Dieldrin 93 ± 13 96 ±4 NA
Endrin 122 ± 19 83(d) NA
p,p'-DDT 153 ±20 193 ±26
OP Pesticide
Diazinon 58 ± 2 83 ± 3 NA
Chlorpyrifos 58 ± 2 101 ± 5 NA
PCB
2,2'4-trichlorobiphenyl 63 ±4 76 ±13 NA
2,2'5,5'-tetrachlorobiphenyl 62 ±2 76 ±10 NA
2,3,4,4'5-pentachlorobiphenyl 54 ±2 73 ±12 NA
2,2'3,5,6-pentachlorobiphenyl 52 ± 2 85 ± 19 NA
2,2'4,4',5,5'-hexachlorobiphenyI 56 ±4 82 ±13 NA
PE (Phthalate Ester)
Di-n-butylphthalate 83 ± 1 ND NA
Benzylbutylphthalate 90 ± 4 ND NA
Ph (Phenols)
Nonylphenol NDW ND NA
Bisphenol-A 29 ± 3 64 ± 3 NA
Pentachlorophenol NA(C) NA 85 ± 2
HA (Herbicide Acid)
2,4-D NA NA 63 ± 3
(a) Data were from triplicate spiked solid food samples; I denotes 25 g of high-fat food samples, II denotes 25 g of medium fat
food samples, and III denotes 10 g of medium fat food samples.
(b) ND denotes not determined; the recoveries were greater than 200%.
(c) NA denotes the target analytes were not analyzed in these samples.
(d) Only one measurement reported, interference peaks were observed in the other two duplicate samples.
59
-------
more than one filter was used to complete the filtration. This step could also contribute to the
loss of the spiked analytes. In order to include PE into the GPC fraction, the fat was not
completely removed from the sample extract after GPC. Approximately 90% of the fat was
removed after GPC, and even after Florisil SPE cleanup, the sample matrix was still very
complex. Due to the time and cost constraints, we did not continue further evaluation
experiments, and used this method (Appendix A) for the Phase 1 and 2 samples.
Better analytical methods could be developed by dividing the compound classes into
three groups; the first group would contain the stable POP including PAH, PE, PCB, and some
OC (i.e., chlordanes and heptachlor). We recommend use of the method developed for PAH to
determine these analytes (5, 6). The method consists of KOH digestion, liquid-liquid
partitioning, column cleanup and GC/MS. The second group would contain less stable POP
including OP, Ph, and some OC. We recommend use of the analytical method described here
(Appendix A) with minor modifications to determine these compounds. The cut-off point for the
GPC fraction would be changed to completely remove the fat and some PE. Since PE would not
be the compound of interest, complete removal of the fat may simplify the sample matrix. The
third group would include 2,4-D and PCP. The analytical method described here could be
modified to improve the recoveries of these compounds. The modifications could include the
use of acidic organic solvent for extracting solid food sample and then acidifying the liquid food
sample prior to organic solvent extraction. The cleanup steps could include GPC and Florisil
SPE.
Phase 1 Field Study
Recruiting of Daycare Centers
The results of recruiting daycare centers for the Phase 1 study are summarized in
Table 5.11. The overall response rate (private daycare and Head Start centers) was about 14%.
The most common reasons for those daycare centers/directors who refused to participate in the
study were: (1) not interested in the study; (2) short of staff, would not have time to do the
60
-------
TABLE 5.11. SUMMARY OF RECRUITING DAYCARE CENTERS FOR
PHASE 1 STUDY
Agreed to Did not Agree
Type
Location
Participate
to Participate
Bad(a) Data
Total
Private
Durham/Chapel Hill
18
97
3
118
Daycare
Centers
Raleigh
6
73
0
79
Head Start
Raleigh
1
0
0
1
Centers
Durham
1
0
0
1
Holly Springs
1
0
0
1
Cedar Grove
1
0
0
1
Gamer
1
0
0
I
Total
29
170
3
202
(a) Bad data denotes disconnected telephone number.
study; and (3) not approved by the owner or board of directors, company policy, concerns about
study findings that may damage the image of the daycare center or even cause problems for their
license. The $50 payment for participation was an effective incentive for many participants.
Nevertheless, the recruiting methods we used were effective and cost-efficient. All recruiting
activities were completed in about four weeks. Our goal was to recruit 9 daycare centers; we
qualified 29. We can apply the same methods for a future larger study in multiple geographical
areas.
Field Activities
The Phase I sampling activities were conducted in ten daycare centers (9 scheduled
centers and 1 backup center) during a period of three weeks in March 1997. In general, we
received good cooperation from each daycare center. There was no problem with the collection
61
-------
of food samples by the staff of each daycare center. The teachers from two daycare centers (D01
and D05) complained about the noise of the indoor air samplers. A backup daycare center (D10)
was sampled during the third week of the Phase 1 field study because the outdoor samplers used
in D05 were overheated. Multi-media samples were collected from these daycare centers as
planned (Table 4.2). It took approximately one hour to complete the pre-monitoring
questionnaire. The most difficult question for the daycare director to answer was the use of
chemicals. For many daycare centers, the maintenance work was done by an outside contractor
or a different division of the same organization. Many phone calls were made to trace the
information about the use of chemicals in these daycare centers. For future large scale studies,
the pre-monitoring questionnaire could be simplified to collect only key information that is
essential to the study. There was no problem in conducting the post-monitoring questionnaires;
it took approximately 15 minutes to complete the interview.
Concentrations of Persistent Organic Pollutants in Multimedia Samples
Table 5.12 summarizes the target POP in multiple compound classes determined in the
Phase 1 multimedia samples. The sums of the concentrations of target POP in each compound
class in air, dust/soil, and food are summarized in Table 5.13 through 5.15, respectively. These
tables show the sums of all target analytes in each compound class. Also reported is the sum of
selected target 4- to 6-ring PAH ranked as probable human carcinogens (B2) in the U.S. EPA
Integrated Risk Information System and designated as B2 PAH. For ease of discussion in this
chapter, the sums of target POP in each compound class were referred as the respective
compound class. The data reported in these tables were corrected for the corresponding field
blanks. Note that the reported data were not corrected for the recoveries of the spiked SRS
except for 2,4-D. The reported 2,4-D data were corrected for the recoveries of SRS (3,4-D). The
individual target POP data in indoor air, outdoor air, HVS3 floor dust, vacuum bag floor dust,
playground soil, liquid food and solid food samples are summarized in Table D-l through D-7,
respectively, in Appendix D.
62
-------
TABLE 5.12. A SUMMARY OF TARGET PERSISTENT ORGANIC POLLUTANTS
DETERMINED IN MULTIMEDIA SAMPLES
Compound Class Target Analyte
PAH Nophthalene
B (phenyl
Acenaphthylene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Cyclop enta[c,d]pyrene
Benz[a]anthracene*
Chrysene*
Benzo[b]fluoranthene*
Benzo[k]fluonnthene*
Benzo[e]pyrene
Benzo[a]pyrene*
lndenofl ,2,3-cJpyrene*
Dibenz[a,h]anthracene'
Benzo[g,h,i]perylene
Coronene
PE Dibutylphlhalate
Benzvlbutylphthalate
OP Diazinon
Chlorpyrifos
OC Lindane
Hcptachlor
Aldrin
gamma-Chlordane
alpha-Chlordane
p,p'-DDH
Dieldrin
Endrin
p.p'-DDT
PCB 2-Chlorobiphenyl
4-Chlorobiphenyl
2,6-Dichlorobiphenyl
4,4'-Dichlorobiphenyl
2,4,4-Trichlorobiphenyl
2,2'5,5 '-T etrachl orobi phenyl
2,2'3,5'-Tetrachlorabiphenyl
2,3',4',5-Tetrachlorobiphenyt
3,3',4,4-T etrachlorobiphenyl
2,2',3,5',6-Pentachlorobiphenyl
2,2',4,5,5'-Pentachlorobiphenyl
2,2',3 ,44'-Pentachlorobiphenyl
2J,3',4',6-Pcnlachlorobipheny1
2,3',4,4\5-Pentachlorobiphenyl
2,3,3',4,4-Pentachlorobiphenyl
33',4,4',5-Pentachlorobiphenyl
2,2\4,4',515'-Hexiichlarobiphenyl
2,2',3,4,4',5'-Hcxachlorobiphenyl
3,3'14,4',5,5,-Hexachlorobrphenyl
2,2',3,4,4',5,5'-Heptachlorobiphenyl
Ph Pentachlorophenol
Nonylphenol
Bisphenol-A
HA 2,4-D
• Denotes that the target PAH are ranked as possible human carcinogen (B2) by U.S. EPA's Integrated Risk Information System.
63
-------
TABLE 5.13. SUMMARY OF TARGET PERSISTENT ORGANIC POLLUTANTS IN AIR SAMPLES FROM
PHASE 1 DAY CARE CENTERS
Compound Concentration, ng/m3
Class
D01
D02
D03
D04
D05
D06
D07
D08
D09
D10
Indoor Air
PAH
497
262
690
120
180
171
116
171
204
241
B2 PAH
0.617
0.771
0.557
0.627
0.703
0.605
0.615
0.686
0.442
0.246
PE
985
359
477
229
273
230
121
448
127
144
OP
38.1
47.7
13.7
10.8
76.1
16.7
9.01
20.0
8.23
12.3
OC
81.9
36.1
59.3
358
23.7
18.7
16.5
74.4
18.0
42.1
PCB
7.32
5.71
22.9
8.72
10.8
246
18.6
11.4
54.8
258
Ph
273
236
529
211
223
153
52.8
221
86.5
81.6
HA
<0.1
0.225
<0.1
0.216
<0.1
<0.1
0.113
0.242
0.474
0.228
Outdoor Air
PAH
75.8
127
88.5
71.8
35.1
317
82.2
120
70.2
139
B2 PAH
0.317
0.423
0.248
0.517
0.153
0.705
0.313
0.288
0.196
0.331
PE
128
876
436
74
104
227
73.3
793
199
106
OP
3.18
30.0
3.60
26.2
3.07
11.9
0.755
6.35
5.46
8.47
OC
15.5
13.4
8.56
10.3
11.1
7.00
13.2
11.0
4.26
7.50
PCB
15.5
11.2
7.71
9.42
7.88
9.34
6.44
7.75
8.20
8.05
Ph
81.3
113
76.3
39.5
42.4
103
107
101
31.6
41.4
HA
<0.1
0.217
<0.1
<0.1
0.382
0.322
0.401
0.544
0.318
0.195
-------
TABLE 5.14. SUMMARY OF TARGET PERSISTENT ORGANIC POLLUTANTS IN DUST AND SOIL SAMPLES
FROM PHASE 1 DAY CARE CENTERS
Compound Class Concentration,'"' ppm
D01
D02
D03
D04
DOS
D06
D07
D08
D09-1
D09-2
D10
Floor Dust
PAH
2.82
9.98
3.07
1.08
0.588
0.600
0.436
1.22
0.388
8.20
9.42
B2PAH
0.993
4.50
1.25
0.507
0.281
0.303
0.172
0.583
0.130
4.40
5.29
PE
61.2
38.4
61.2
105
89.7
57.9
22.9
221
50.6
139
100
OP
0.799
1.35
0.623
0.433
1.19
0.975
0.205
1.55
0.096
0.742
0.840
OC
1.66
1.23
1.20
3.62
1.08
0.597
1.06
1.43
1.58
1.25
0.761
PCB
0.143
0.697
0.586
0.354
0.072
2.76
0.138
0.166
0.139
24.8
28.2
Ph
9.14
14.7
5.35
7.52
10.7
15.2
5.51
16.2
9.77
16.5
8.33
HA
0.118
0.052
0.020
0.264
0.050
0.235
0.024
0.315
0.139
0.160
0.618
Playground Soil
PAH
13.1
1.04
0.574
0.101
0.061
0.167
0.060
0.255
0.197
NA(b)
0.127
B2PAH
5.59
0.397
0.211
0.024
0.012
0.052
0.017
0.075
0.055
NA
0.042
PE
0.247
0.290
0.250
0.231
1.24
1.22
0.245
0.278
0.312
NA
0.242
OP
0.012
0.010
0.006
0.040
0.011
0.011
0.004
0.004
0.004
NA
0.008
OC
0.006
0.005
0.005
0.070
0.022
0.022
0.024
0.007
0.036
NA
0.014
PCB
0.004
0.001
0.007
0.004
0.001
0.001
0.09
0.008
0.007
NA
0.006
Ph
0.522
0.187
0.204
0.305
0.255
0.255
0.237
0.276
0.255
NA
0.136
HA
<0.001
0.003
<0.001
<0.001
<0.001
<0.001
<0.001
0.005
<0.001
NA
<0.001
(a) D09-1 denotes the floor dust sample was collected from the classroom where the air sample was taken and equipped with a new carpet (2 months old). D09-
2 denotes the floor sample taken from another classroom equipped with an old crpet (2 years old).
(b) NA denotes not applied, only one playground soil samples was collected from D09.
-------
TABLE 5.15. SUMMARY OF TARGET PERSISTENT ORGANIC POLLUTANTS IN LIQUID AND SOLID FOOD
SAMPLES FROM PHASE 1 DAYCARE CENTERS
Compound Concentration, ppb
Class
D01
D02
D03
D04
D05
D06
D07
D08
D09
D10
Liquid Food
PAH
0.86
3.12
2.76
0.81
0.53
0.60
0.57
0.53
3.31
0.68
B2 PAH
<0.04
<0.04
<0.04
0.05
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
PE
64.4
32.3
22.5
25.0
26.8
86.5
49.0
16.6
86.4
34.1
OP
<0.04
0.26
<0.04
<0.04
<0.04
0.23
0.180
<0.04
0.12
0.08
OC
2.00
1.31
1.28
1.20
<0.04
<0.04
0.24
0.090
0.82
0.31
PCB
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
Ph
3.45
10.1
<0.1
7.77
7.50
7.97
10.3
7.12
8.43
6.43
HA
0.29
0.22
0.64
0.58
0.39
0.20
2.36
1.59
2.08
1.71
Solid Food
PAH
1.87
2.56
2.70
6.19
3.15
4.32
6.00
4.01
3.31
9.15
B2 PAH
0.05
0.08
0.48
1.01
0.08
0.68
0.88
0.62
0.47
0.42
PE
191
188
199
194
342
279
387
154
253
234
OP
0.56
0.66
0.55
0.37
2.70
15.17
1.60
0.63
0.49
1.19
OC
0.40
2.26
0.34
4.32
5.46
0.25
1.15
0.48
2.81
3.04
PCB
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
Ph
28.6
30.2
68.6
55.8
55.3
50.4
49.1
36.6
44.2
57.9
HA
0.28
0.30
3.14
2.47
1.30
1.45
1.48
0.42
0.26
2.44
-------
The concentrations of PAH found in indoor air were higher than those in the
corresponding outdoor air. The B2 PAH concentrations in air only represent a small portion of
the total PAH concentrations. These PAH concentration profiles are in agreement with those
from previous studies (5,6). Higher indoor concentrations were also observed for target POP in
other compound classes including PE, OC, OP, PCB, and Ph. Similar indoor and outdoor
concentrations were measured for 2,4-D. In general, concentrations of target POP in all
compound classes were of the same order of magnitude among all day dare centers. The most
abundant air concentrations were found for target POP in the compound classes of phthalate
esters (PE), PAH, and phenols (Ph), followed by PCB, OC, OP, and 2,4-D.
Concentrations of individual target POP in each compound class (Appendix D) showed
that few target PAH including cyclopenta[c,d]pyrene, dibenzo[a,h]anthracene, and coronene
were not detected in the air samples. The individual target POP in PE, OP, and Ph were found in
most air samples. However, many target POP in OC were not detected in the air samples. Only
four of nine compounds, namely lindane, heptachlor, gamma-chlordane, and alpha-chlordane,
were found in all air samples. Similarly, a few of the target individual PCB congeners were not
detected in the air samples. The herbicide acid 2,4-D was found in 13 of 20 air samples. Note
that there were two outdoor air samples from D05. The D05-1 sample was collected under the
condition where the air pumps were overheated, and the D05-2 sample was a 24-hr sample and
collected the day after D05-1 was taken. The overheating resulted from blockage of the fan of
the outdoor sampling box by a piece of plastic bag. One of the air pumps for filter/PUF was
stopped, and the other one for filter/XAD-2 was still functioning (D05-1) when the field
technician checked the samplers on Day-3 of the monitoring period. The concentrations of
phenols in D05-1 sample were significantly higher than in D05-2 sample. This is probably due
to sampling artifacts from the heated plastic bubble wrap.
The dust samples were separated into coarse and fine fractions and only the fine fractions
(<150 |xm) were analyzed for target POP. The floor dust loadings of the daycare centers are
summarized in Table 5.16. Levels of the floor dust ranged from 1.06 to 35.8 g/m2 for the fine
dust loadings. The highest and lowest floor dust loading was observed in daycare center #8
67
-------
TABLE 5.16. FLOOR DUST LOADINGS FROM PHASE 1 DAY CARE CENTERS
EES
Day Care
Center**' Code
Dust Loading, g/m2
Age of Carpet,
years
Total
Fine Fraction
(<150 pm)
Percent of Fine
Fraction, %
D01
13
16.8
13.6
81
D02
4
6.96
3.84
55
D03
DKW
22.2
18.5
83
D04
5
16.8
11.3
67
D05
1
9.54
7.67
80
D06
1
15.9
15.1
95
D07
3
18.8
15.0
79
D08
1
36.7
35.8
92
D09-1
0.17
3.83
1.98
52
D09-2
2
2.31
1.06
46
D10
5
14.7
13.7
93
(a) D09-1 denotes the floor dust sample was collected from the classroom where the air sample
was taken, and the carpet was about 2 months old (0.17 years old). D09-2 denotes the floor
dust sample was collected from another classroom and the carpet was 2 years old.
(b) DK denotes don't know.
68
-------
(D08) and #9 (D09), respectively. The fine dust loadings accounted for 46 to 95% of the total
dust loadings.
The target POP concentrations in dust and soil (ppm) were calculated by subtracting the
amounts of target analytes in the field blank from those in the sample and dividing by the
amounts of dust analyzed. The reported concentrations were based on the dried weights, which
are corrected for moisture content. The data reported in Table 5.14 were the floor dust samples
collected with HVS3. Higher concentrations of target POP in all compound classes were
observed in the floor dust samples as opposed to the playground soil samples. With one
exception, PAH concentrations in the playground soil sample from D01 were higher than that in
the floor dust sample. The highest PAH concentrations in Phase 1 dust and soil samples were
found in the playground soil sample from D01. The concentrations of B2 PAH and total target
PAH in this soil sample were 5.59 and 13.1 ppm, respectively. It is possible that there were local
contamination sources of PAH that affected this playground soil sample. Concentrations of total
target PAH and B2 PAH in the floor dust samples ranged from 0.388 to 9.98 ppm and from
0.130 to 5.29 ppm, respectively. Concentrations of B2 PAH accounted for approximately 50%
of the total PAH concentrations. This finding is in agreement with the house dust samples
analyzed in the previous study (1, 5-7).
With few exceptions, concentrations of OC, OP, Ph and PCB were of the same order of
magnitude among the ten daycare centers. Note that higher PCB concentrations were found in
D09-2 and D10. For daycare center #9 (D09), two floor dust samples were taken from two
classrooms. The sample D09-1 was from a classroom with a new carpet (2 months old) and
D09-2 was from another classroom with an older carpet (2 years old). Higher concentrations of
the target POP in all compound classes were observed in D09-1 as opposed to D09-2. Note that
the dust loadings of these two classrooms were about the same (Table 5.16). Upon further
examination of the relationship between ages of the carpet (Table 5.16) and the POP
concentrations, there was no clear trend to be seen. Phthalate esters (PE) were the most abradant
contaminants found in the dust and soil samples. Concentrations of PE ranged from 22.9 to
221 ppm in the floor dust samples and from 0.231 to 1.24 ppm in the playground soil samples.
69
-------
Among all the compound classes monitored, the least abundant contaminant in dust/soil
was herbicide acid, 2,4-D. Concentrations of 2,4-D in the dust samples were less than 1 ppm and
not detected in most soil samples. Concentrations of individual target POP in each compound
class (Appendix D) indicated that mono- and di-chlorobiphenyls were not detected in the dust
samples, and only a few PCB congeners were detected in the playground soil samples. The
individual target POP in PAH, PE, OP, and Ph were found in most dust and soil samples. Target
OC including aldrin, DDE, dieldrin, endrin, and DDT were detected in all dust samples but not in
most soil samples. The other OC such as lindane, heptachlor, and cholordanes were found in all
dust samples and most soil samples.
The data for the liquid and solid food samples are expressed in units of ng of POP, and in
g of food sample (ppb). These POP concentrations were obtained by subtracting the amounts
(ng) of POP in the field blank from the amount of (ng) of POP in the sample, and then dividing
by the amount (g) of food used. Note that PCB were not found in any solid and liquid food
samples. The B2 PAH and OP were only detected in few liquid food samples. The most
abundant contaminants found in liquid and solid food were PE. The concentrations of PE ranged
from 16.6 to 86.4 ppb in liquid food samples and from 188 to 387 ppb in solid food samples.
The concentrations of phenols were above 10 ppb. The concentrations of target POP in other
compound classes were mostly less than 5 ppb. Note that higher concentrations of phthalate
esters were detected in both solid and liquid food samples collected with plastic containers as
opposed to those collected with glass containers. Similar concentrations of other target POP
were found in food samples collected by both types of containers. Thus, in Phase 2 study we
used the glass containers for collecting liquid and solid food samples.
Data Analysis for Phase I Study
Summary Statistics
Summary statistics by compound class and by sample media for each target POP and the
sums of target POP in each compound class are presented in Table E-l thorough E-7 in
-------
Appendix E. Each table contains sample size (N), number of samples below detection limit
(N-BDL), mean, standard deviation, minimum, and maximum values of multimedia samples.
For the target analytes below detection limit, half of the detection limit was used for the
calculation of summary statistics. Note that the target POP data in liquid and solid food samples
collected in plastic containers are not included in the summary statistics.
Tables E-l and E-2 show that the average indoor air POP concentrations are higher than
the average outdoor air POP concentrations across all compound classes, except for the herbicide
acid (2,4-D) and a few target compounds. Some of the target analytes in PAH, OC pesticides,
and PCB were not detected in the air samples.
Tables E-3 and E-4 summarize the POP data in HVS3 floor dust and vacuum bag floor
dust samples, respectively. The 2,4-D analysis was not performed in the vacuum bag samples.
The average POP levels were higher in the vacuum bag floor dust samples as opposed to the
HVS3 floor dust samples for all compounds classes except PCB. Summary statistics for
playground soil samples are presented in Table E-5. Note that most target analytes in OC
pesticides and in PCB were not found in the playground soil samples. Summary statistics for
liquid and solid food samples are presented in Table E-6 and E-7, respectively. The target PCB
were not found in any liquid and solid food samples. In general, average POP levels were higher
in the solid food samples than that in the liquid food samples across all compound classes.
Summary statistics for target POP by two groups of centers (Head Start versus private)
are presented in Table E-8 through E-l 4. Average indoor and outdoor concentrations of PAH
and PE in Head Start centers were higher than those in the private centers. The reversed trend
was observed for HA. Similar average concentrations of OP, OC, PCB, and Ph in air were found
in two groups of centers. Differences between the two groups of centers in POP in air were not
significant, and were probably related to their locations. Comparable concentrations of target
POP in floor dust collected by HVS3 in all compound classes but PCB were found in both
groups of centers. Higher average PCB concentrations in floor dust were found in the private
centers as opposed to Head Start centers. This is mainly from the high PCB levels found in
centers D09 and D10. A significant difference in the playground soil concentrations of PAH was
seen between the two groups of centers. This is because the unusually high levels of PAH found
-------
in D01 resulted in significantly higher average concentrations of PAH in the Head Start centers.
In fact, the highest PAH levels among all Phase 1 dust/soil samples were in the DO 1 playground
soil sample. Comparable average PAH concentrations in playground soil were obtained in PAH
data from both groups of centers if D01 was excluded. There could be a local contamination
sources, which affected playground soil in D01. Levels of target POP in other compound classes
were similar between the two groups of centers. Although the average food concentrations of
POP showed differences between the two groups of centers, these differences were not
significant, given the small sample size and the great day-to-day variability in the foods served
and the cooking methods.
Correlation Between Sample Media
It is of interest to know whether the levels of POP in one sample medium (e.g., dust) are
related to their levels in other sample media The correlation between the measured target POP
concentrations in different sample media was investigated. Note that initially, the vacuum bag
floor dust samples were not included, and only the HVS3 floor dust samples were used for the
correlation analysis between sample media. Spearman correlation coefficients (r) and Pearson
correlation coefficients (r) for the sums of target POP in each compound class in one sample
medium (e.g., floor dust) with those levels in another sample medium (e.g., playground soil) are
summarized in Appendix F. Note that the Spearman correlation coefficients were obtained by
ranking of the raw data and the Pearson correlation coefficients were calculated by the log-
transformed data.
Table 5.17 summarizes the pairs of sample media that had significant relationships at
least at 0.05 or lower confidence levels. Note that the results of Spearman correlation
coefficients and Pearson correlation coefficients were not identical but similar to each other.
This is mainly because Spearman correlation coefficients use the rankings of concentrations and
Pearson correlation coefficients use the log-transformed concentrations. The direct relationships
were observed between indoor air and floor dust, for target POP in PAH, OC, OP, and PCB. The
correlation between indoor air and playground soil was seen in total PAH, and PCB. A direct
72
-------
TABLE 5.17. SUMMARY OF PAIRS OF SAMPLE MEDIA WITH SIGNIFICANT
CORRELATION COEFFICIENTS FOR TARGET COMPOUND CLASSES
Compound Class(a)
Pair of Media with Significant Correlation Coefficient
Spearman Method(c)
Pearson Method(d)
B2 PAH
—
(IA, LF)**
Total PAH
(IA, FD)
(IA, PS)
(IA, PS)**
(LF,FD)
HA. SF1
(SF. PS}
(PS, FD)
(LF, FD)
fSF. PS}
OC
(IA,FD)
(IA, FD)**
OP
(IA, FD)**
(IA, FD)
PE
flA. SFV)
PCB
(IA, PS)
(IA, FD)**
PH
—
(IA, LF)
HA
—
—
(a) PE = phthalate esters, PH = phenols, OC = OC pesticide, OP = OP pesticide, and HA =
herbicide acid.
(b) LA = indoor air, OA = outdoor air, F = floor dust, LF = liquid food, SF = solid food, and
PS = playground soil. All pairs are at least significant at the 0.05 level, while pairs with
0.01 significant level are marked with **, and pairs with negative correlations are
underlined.
(c) Spearman's correlation method in raw data.
(d) Pearson's correlation method in log-transformed data.
73
-------
correlation between floor dust and pathway soil was significant at 0.05 level for total PAH.
There were significant but negative relationships observed between solid food samples and
indoor air samples for total PAH and PE. The reason for the negative relationship is not known.
Correlation Between Compound Classes
It is of interest to know whether the levels of POP in one compound class are related to
another compound class within each sample medium. Spearman and Pearson correlation
coefficients between compound classes within each sample medium are presented in
Appendix G. Table 5.18 summarizes the pairs of compound classes had correlation coefficients
significant, at least, at 0.05 or lower confidence levels. As shown in the Table 5.18, there were
quite a few significant relationships between compound classes for target POP. The direct
relationship between total target PAH and B2 PAH was expected since they all are in the same
compound classes. This relationship was observed in the Phase 1 dust/soil samples as well as
dust/soil samples from previous studies (I, 5-7). The significant correlation between B2 PAH
and total PAH was also found in outdoor air and solid food samples. There was no correlation
between B2 PAH and total PAH in indoor air samples. This was probably due to contamination
sources for the abundant 2- to 3-ring PAH which were different from those for B2 PAH in indoor
microenvironments, and produced high levels of 2- to 3-ring PAH in the indoor air.
Significant correlations between B2 PAH and other compound classes including OP,
PCB, and PE were observed in air and in dust samples. The negative relationship between B2
PAH and PCB was probably due to the unusually high indoor PCB concentrations in D06. The
direct relationships between total PAH and two other compound classes PE and Ph were found in
indoor air samples. In general, several pairs of compound classes were correlated with each
other in indoor air and floor dust samples. Fewer pairs of compound classes with direct
relationships were observed in outdoor air, playground soil, and food samples. It is possible to
monitor one compound class as a marker for other compound class for indoor samples (air, dust)
for future large-scale studies. Since this was a small data set, more data are needed to examine
this hypothesis.
-------
TABLE 5.18. SUMMARY OF PAIRS OF COMPOUND CLASSES WITH SIGNIFICANT
CORRELATION COEFFICIENTS FOR EACH SAMPLE MEDIUM
Sample Medium
Pair of Compound Classes with Significant Correlation Coefficient1^
Spearman Method^
Pearson Method(c)
Indoor Air
Outdoor Air
Floor Dust (HVS 3)
Floor Dust (bag)
Playground Soil
Liquid Food
Solid Food
(B2-PAH, OP)
(B2-PAH. PCB)**
(PE, TARGET PAH)
(PH, TARGET PAH)
(OP, PE)
(PH, PE)**
(PH,OP)
(B2-PAH, OP)
(B2-PAH, TARGET PAH)**
(B2-PAH, PCB)
(PCB, TARGET PAH)
(HA, PE)
(PH, OP)
(B2-PAH, TARGET PAH)**
(B2-PAH, PE)**
(PE, TARGET PAH)
(B2-PAH, TARGET PAH)**
(B2-PAH. OO
rOC. TARGET PAH)
(OC, TARGET PAH)**
(B2-PAH, TARGET PAH)
(HA, PH)**
(B2-PAH, PCB,)**
(PE, TARGET PAH)
(PH, TARGET PAH)
(PH, PE)**
(B2-PAH, TARGET PAH)**
(B2-PAH, TARGET PAH)**
(B2-PAH, PCB)
(HA PE)
(PH, OP)
(B2-PAH, TARGET PAH)**
(B2-PAH, PE)**
(PE, TARGET PAH)**
(B2-PAH, TARGET PAH)**
(B2-PAH.QC)**
rOC. TARGET PAH)
(PH, OP)
(OC, TARGET PAH)
(B2-PAH, TARGET PAH)
(HA, PH)**
(a) PE = phthalate esters, PH = phenols, OC = OC pesticide, OP = OP pesticide, and HA = herbicide
acid. All pairs are at least significant at the 0.05 level, while pairs with 0.01 significant level are
marked with **, and pairs with negative correlations are underlined.
(b) Spearman's correlation method in raw data.
(c) Pearson's correlation method in log-transformed data.
75
-------
Correlation coefficients were also determined for target POP found in the HVS3 floor
dust and vacuum bag floor dust samples. Table 5.19 shows correlation matrices for each
compound class in these two types of samples. Table 5.19 shows the Pearson correlation
coefficients on both raw data and log-transformed data as well as Spearman correlation
coefficients on the raw data The results were similar among the three correlation coefficients for
each target compound class. Direct relationships were observed for compound classes B2 PAH,
PAH, and PCB between these two types of samples. It is possible to use floor dust samples as an
indicator for other sample media and to use less expensive screening methods for determining
POP in the dust samples. The vacuum bag floor dust samples instead of HVS3 floor dust
samples could also be used in the screening step in future large-scale exposure studies.
Estimates of Daily Persistent Organic Pollutant Exposure
The potential daily POP exposure of child subject from each daycare center was
estimated for the inhalation (air), nondietary (dust/soil), and dietary (food) pathways as described
in Chapter 4. The POP exposures were expressed in ng/day, for the time that a child spends in
daycare, which is slightly less than half of the 24-hr period for the average child in this study (but
which may be most of the child's waking hours). The estimated daily POP exposures through
the three pathways are summarized in Appendix H. Summary statistics for the estimated daily
children's POP exposures from Phase 1 daycare centers are presented in Appendix I. Note that
the estimated daily POP dose (ng/kg/day) noimalized by each child's body weight was calculated
only for the Phase 2 data but not for the Phase 1 data, since the children were monitored
individually only in Phase 2.
Figure 5.1 displays the average distribution of the estimated daily POP exposure through
the three pathways at Phase 1 daycare centers. The results showed that most important pathway
for children's POP exposure at daycare centers was dietary ingestion except for compound class
PCB. Since none of the target PCB were found in the food samples, dietary ingestion was the
least important pathway for PCB exposure for children in these daycare centers. For total PAH
exposure, the inhalation pathway was almost as important as the dietary pathway. This is
-------
TABLE 5.19. CORRELATION COEFFICIENTS BETWEEN FLOOR DUST (HVS3) AND
FLOOR DUST (BAG)
Compound Class
Pearson Correlation
Coefficient of Raw
Data
Pearson Correlation
Coefficient of
Log-Transformed
Data
Spearman
Correlation
Coefficient of
Raw Data
B2 PAH
0.704
0.738*
0.821*
Target PAH
0.598
0.731*
0.893**
Phthalate Esters
0.444
0.602
0.643
OP Pesticides
0.0765
0.273
0.357
OC Pesticides
-0.403
-0.406
-0.429
Target PCB
0.764*
0.679*
0.536
Phenols
0.386
0.439
0.357
* —significant at 0.05 level.
**—significant at 0.01 level.
77
-------
Sum of B2-PAH, ng/day
Bisphenol-A, ng/day
0% 5%
95%
Total Target PCB, ng/day
¦Air
¦ Dust/Soil
~ Food
¦ Air
¦ Dust/Soil
~ Food
¦Air
¦ Dust/Soil
~ Food
Total Target PAH, ng/day
Total Target PE, ng/day
95%
Total Target Phenols, ng/day
6% 2%
92%
¦Air
¦ Dust/Soil
~ Food
¦ Air
¦ Dust/Soil
~ Food
¦Air
¦ Dust/Soil
~ Food
Total Target OP, ng/day
16%
81%
¦ Air
¦ Dust/Soil
~ Food
Total Target OC, ng/day
¦ Air
¦ Dust/Soil
3% ~ Food
2,4-D, ng/day
r-0%
U°/-
99%
¦ Air
¦ Dust/Soil
~ Food
Figure 5.1. Distributions of average daily exposure of persistent organic pollutants from Phase 1 daycare centers.
-------
probably because of the high levels of 2- to 3-ring PAH found in air samples. For B2 PAH, PE,
and HA, the estimated daily exposure through nondietary ingestion pathway was greater than that
through inhalation pathway. The reverse relative trend was observed for total PAH, OP and OC
pesticides, PCB, and phenols.
As shown in Appendix I, the averages of daily POP exposures in PAH, PE, O, and Ph
through air in Head Start centers were slightly higher than that in the private centers. The reverse
trend of the inhalation pathway was seen in the average daily POP exposures in OP, OC, PCB,
and HA. With one exception, differences between the two groups of centers in the average daily
POP exposures through nondietary and dietary ingestion were small for all target compound
classes. The average daily PCB exposures through nondietary ingestion were higher at the
private centers, by a factor of 11. This difference was from the higher PCB levels in the dust and
soil samples from private center D09 and D10. Figure 5.2 displays the average daily potential
POP exposures for these two groups of centers through inhalation (air), nondietary ingestion
(dust/soil), and dietary ingestion (food) pathways in Phase 1 study. Note that levels of average
daily Ph and PE exposures shown in Figure 5.1 need to be multiplied by 10 and 20, respectively,
to obtain the actual average levels. Differences in daily POP exposures through all three
pathways between Head Start centers and private centers, in general, were small. Higher average
daily B2 PAH, OC, PCB, and HA exposure levels were observed in the private centers. Higher
average OP, and total PAH exposure levels were seen in the Head Start centers.
Quality Control Data for Phase I Study
Accuracy was assessed by means of spike recoveries of analytical surrogate
recovery standards (SRS). Known amounts of perdeuterated and/or 13C-labelled POP were
spiked into each air, dust, soil, and food samples prior to sample preparation. The recovery
data of the spiked POP for each type of sample are presented in Appendix J. In general,
acceptable recoveries (>70%) for the spiked POP were obtained in most of the samples.
The recoveries of DDT-13C were greater than 100% in the food samples. As discussed
earlier in this Chapter, this was mainly because the GC responses of DDT-13C were
-------
Average Daily POP Exposures in Head Start Centers
and Regular Centers
Head Start Regular Head Start Regular OC Head Start Regular Head Start Regular HA
B2PAH B2PAH OC PCB PCS HA
Average Daily POP Exposures in Head Start Centers
and Regular Centers
I
s
a
s
B.
X
UJ
Q.
o
Q.
>i
Head Regular Head Regular Head Regular Head Regular
Start OP OP Start PAH PAH Start Ph Ph Start PE PE
Figure 5.2. Average daily persistent organic pollutant exposures in Head Start centers and
other centers.
80
-------
better than that in the standard solutions. Note that the recoveries of DDT-13C reported in
Appendix J were corrected for this factor.
The overall method precision was expressed as percent of relative standard
deviation (%RSD) in duplicate field samples. The precision data of target POP are
presented in Appendix K. Note that the duplicate air sample was not collected for 2,4-D
(HA) analysis. Only duplicate dust samples but not soil samples were processed because
the dust sample matrices are more complicated than the soil samples. The overall method
precision was within ±10% for the air and dust samples. The precision for the food sample
matrices were within ±20% for all compound classes but 2,4-D. As mentioned in earlier
method evaluation section, several problems occurred when analyzing food samples, such
as plugged columns, which could have contributed to the larger variations observed in the
duplicate samples.
The results of target POP found in the field blanks are presented in Appendix L.
None of the PCB and OC pesticides were found in the field blanks. The highest levels of
target POP found in the blanks for all sample media were the two phthalate esters (PE).
This is because phthalate esters are common environmental contaminants. Much higher
levels of PE were found in most samples. All data reported here were corrected for the
respective field blanks. In general, the field blank data demonstrated that there was no
significant contamination due to sample handling and preparation.
Phase 2 Field Study
Recruiting of Subjects
The results of recruiting children from the two daycare centers (D03 and D09) for the
Phase 2 study are summarized in Table 5.20. The overall response rate was about 17%. We
recruited 11 families, 5 low-income families and 6 middle-income families, in less than four
weeks. Based on 17% response rate, we estimated that we can recruit 6 families per daycare
center. If we used all 29 daycare centers recruited for the study, we could have recruited
81
-------
174 households for the study. Thus, it is feasible to apply the recruiting method used in this
study for a future large-scale study.
TABLE 5.20. SUMMARY OF RECRUITING CHILDREN FROM DAYCARE CENTERS
D03 AND D09 FOR PHASE 2 STUDY
D03 D09 OVERALL
Target Children 30 36 66
Agreed to Participate 5 6 11
Response Rate 17% 17% 17%
Field Activity
The Phase 2 sampling activities were conducted in two daycare centers and nine homes
during a two-week period in June 1997. In general, we received good cooperation from
participating daycare centers and families. The indoor air samplers were modified to reduce
noise levels because of the complaints received in the Phase 1 study. We did not receive any
complaints about the noise levels of the indoor air samplers for the Phase 2 study.
In general, there were no major problems in collecting multimedia samples in Phase 2
study. The outdoor air sampling at home D (HD3) was interrupted for four hours due to power
outage. Therefore, the sampling time was extended for an additional four hours. In daycare
center D09, 24-hour instead of 48-hour duplicate-diet food samples were collected because the
daycare center had special activities and did not provide lunch during the second day of field
monitoring. The children brought their own lunch on that day. Two of nine participating parents
from the low-income families complained that duplicate-diet food collection is too much burden
for them.
Two types of child activity diaries were employed in Phase 2 study, one for daycare
activities and one for home activities. Most parents thought that the diaries are well organized
82
-------
and easy to follow. Two adult participants complained that the instructions for
self-administrated food frequency questions were not clear enough for them to understand. One
common difficult question for most participants was the use of household chemicals, as we
observed in Phase 1 study.
Concentrations of Persistent Organic Pollutants in Multimedia Samples
The target analytes determined in Phase 2 samples were the same as Phase 1 samples
(Table 5.12) except that 2- and 4-chlorobiphenyl were excluded. The sums of levels of target
POP in each compound class in air, dust/soil, food, and wipe samples are summarized in
Tables 5.21 through 5.24, respectively. The reported data were corrected for the corresponding
field blanks but not corrected for the spiked SRS recoveries except for 2,4-D. The reported
concentrations of 2,4-D were corrected for the SRS (3,4-D) recoveries and for the background
levels. The individual target POP data are presented in Appendix M.
Target POP data in indoor and outdoor air samples are presented in Tables M-l and M-2
in Appendix M. Concentrations of target POP in indoor air samples were, in general, higher
than those in the corresponding outdoor air samples. Levels of target POP in the two 48-hr
indoor and outdoor air samples from each daycare center were similar. The POP concentrations
observed in the day-to-day variations were, in general, smaller than those observed in the site-to-
site variations (between daycare centers, and among daycare centers and homes). As seen in
Phase 1 air samples, with few exceptions the highest air concentrations were of PE, PAH, and
phenols, followed by OC, OP, and HA. The HA, 2,4-D was detected in few indoor and outdoor
air samples. As shown in Appendix M, some of the target OC were not detected in the air
samples. No p,p'-DDT was found in outdoor air samples. Some target PCB were not found in
the air samples. The phthalate ester concentrations in the Phase 2 field blanks were five to ten
times higher than those in the Phase 1 field blank. After blank correction, the phthalate esters
were below the detection limit in the outdoor air sample from household HI9. Since phthalate
esters are common environmental contaminants, the Phase 2 field blank could be contaminated
with phthalate esters. Consequently, the reported values of PE (corrected for the field blank)
-------
TABLE 5.21. SUMMARY OF TARGET PERSISTENT ORGANIC POLLUTANTS IN AIR SAMPLES FROM PHASE 2 STUDY
Concentrations, ng/m3(,)
Compound
Class
D03-1
D03-2
D09-1
D09-2
HA3
HB3
HC3
HD3
HE9
HF9
HG9
HH9
HI9
Indoor Air
PAH
1240
1090
206
218
400
205
585
215
429
1300
513
276
446
B2 PAH
0.666
0.570
0.553
0.653
0.964
0.550
1.00
0.650
0.496
0.589
0.657
0.642
0.656
PE
914
857
373
383
572
312
511
404
333
272
602
243
627
OP
36.1
28.1
3.09
4.77
161
16.8
6.71
36.0
3.70
147
38.1
1150
5.99
OC
61.7
50.5
7.37
8.69
60.2
28.1
17.0
6.94
10.4
20.5
186
120
19.4
PCB
42.2
33.6
29.0
38.8
17.0
11.9
11.4
4.58
8.28
4.31
13.2
64.7
15.6
Ph
227
173
240
402
194
355
310
3.43
108
107
418
206
4.71
HA
<0.1
<0.1
0.073
<0.1
<0.1
<0.01
<0.1
0.311
<0.1
<0.1
0.313
<0.1
<0.1
Outdoor Air
PAH
65.0
75.6
98.2
61.0
40.6
63.6
48.8
115
34.0
137
112
176
93.3
B2 PAH
0.666
0.499
0.567
0.483
0.766
0.412
0.514
0.543
0.409
0.522
0.500
0.468
0.417
PE
151
179
277
220
86.1
53.2
524
125
73.0
270
81.7
203
<0.04
OP
1.77
1.36
0.792
1.17
1.92
1.57
2.09
1.48
4.54
2.70
2.06
1.84
2.36
OC
3.06
1.98
1.99
1.29
2.20
1.39
1.74
0.701
0.964
0.801
5.56
1.95
1.26
PCB
4.66
3.61
5.15
7.95
3.73
1.32
2.82
1.35
0.951
2.084
1.12
2.48
0.662
Ph
1.85
0.984
10.3
9.95
1.33
3.37
7.64
2.88
5.31
7.29
1.41
2.80
3.01
HA
<0.1
<0.1
0.170
<0.1
0.069
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
(a) The first letter of the sample code denotes the location of the collected sample, D = daycare centers, H = households; the last two letter/number denotes the
unique code of the daycare center or household.
-------
TABLE 5.22. SUMMARY OF TARGET PERSISTENT ORGANIC POLLUTANTS IN DUST AND SOIL SAMPLES FROM
PHASE 2 STUDY
Concentrations, ppm(a)
Compound
Class
D03-1
D03-2
D09-1
D09-2
HA3
HB3
HC3
HD3
HE9
HF9
HG9
HH9
HI9
Floor Dust
PAH
6.07
4.66
0.778
0.452
0.556
0.540
1.19
0.737
0.591
11.1
1.64
1.40
1.21
B2 PAH
2.56
1.93
0.320
0.196
0.222
0.213
0.514
0.327
0.261
5.32
0.733
0.590
0.520
PE
0.080
7.81
8.34
6.12
6.70
6.49
1.11
10.9
3.07
16.4
4.79
5.11
9.13
OP
0.098
0.337
0.057
0.071
0.420
0.109
0.034
0.255
0.058
0.956
1.38
6.45
0.113
OC
0.608
0.977
0.035
0.043
1.14
0.196
0.078
0.327
0.053
0.349
1.14
0.632
0.194
PCB
0.148
0.613
0.182
0.134
0.085
0.105
0.027
0.056
0.036
0.256
0.271
0.099
0.310
Ph
5.24
14.4
55.9
49.1
7.71
10.6
4.02
8.29
11.4
9.96
6.88
11.6
9.31
HA
0.051
0.069
0.269
0.188
7.29
1.44
0.027
0.029
1.64
0.083
0.169
0.113
0.350
Playarea Soil
PAH
0.093
_
-------
TABLE 5.23. SUMMARY OF TARGET PERSISTENT ORGANIC POLLUTANTS IN LIQUID AND SOLID FOOD SAMPLES
FROM PHASE 2 STUDY
Concentrations, ppb(*>
Compound -
Class
D03-1
D03-2
D09-1
D09-2
HA3
HB3
HC3
HD3
HE9
HF9
HG9
HH9
HI9
Liquid Food
PAH
2.18
1.04
0.469
0.263
0.212
0.285
0.061
0.146
0.160
0.248
0.962
0.146
0.501
B2 PAH
0.450
<0.04
0.114
<0.04
<0.04
0.219
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
PE
8.40
4.25
30.3
7.57
5.97
9.14
8.97
24.2
5.70
9.58
63.1
35.2
14.1
OP
<0.04
<0.04
0.317
<0.04
0.234
0.198
<0.04
0.432
<0.04
<0.04
<0.04
<0.04
<0.04
OC
0.221
0.626
0.744
0.495
0.642
<0.04
<0.04
0.470
0.243
0.210
<0.04
0.564
0.289
PCB
<0.04
0.196
<0.04
0.175
<0.04
<0.04
0.072
<0.04
0.134
0.094
<0.04
<0.04
<0.04
Ph
0.139
4.68
4.51
2.39
<0.1
<0.1
<0.1
<0.1
0.280
<0.1
2.59
3.28
1.37
HA
2.01
0.917
1.80
1.64
0.353
0.705
3.19
3.02
1.50
0.954
0.675
1.67
1.03
Solid Food
PAH
4.92
6.36
3.63
1.59
7.10
5.60
2.13
2.68
3.41
3.68
5.09
1.98
2.94
B2PAH
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
<0.04
PE
25.7
126
103
14.1
461
176
20.5
52.5
101
17.9
96.5
17.0
145
OP
0.473
0.270
1.65
0.214
0.660
0.261
0.085
1.37
0.172
0.981
0.408
1.12
2.31
OC
0.088
0.350
<0.04
0.138
0.621
0.060
0.666
<0.04
<0.04
0.197
0.818
0.621
2.08
PCB
<0.04
<0.04
<0.04
0.220
0.380
<0.04
<0.04
0.268
0.411
0.136
0.161
<0.04
0.398
Ph
35.2
10.8
23.0
12.9
67.7
26.1
15.0
12.8
34.6
21.3
80.4
26.2
21.4
HA
<0.5
<0.5
1.45
0.920
2.20
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
(a) The first letter of the sample code denotes the location of the collected sample, D = daycare centers, H = households; the last two letter/number denotes the
unique code of the daycare center or household.
-------
TABLE 5.24. SUMMARY OF TARGET PERSISTENT ORGANIC POLLUTANTS IN DERMAL WIPE SAMPLES FROM
PHASE 2 STUDY
Concentrations, ng/wipe(a)
Compound
Class
A3
B3
C3
D3
E9
F9
G9
H9
19
Daycare Center
PAH
5.61
21.6
6.93
4.47
22.8
12.7
15.5
10.5
15.4
B2PAH
<0.5
6.29
0.60
1.16
1.13
0.51
<0.5
<0.5
0.49
PE
<0.5
1080
1350
552
344
<0.5
227
<0.5
69.6
OP
<0.5
3.12
2.35
14.6
<0.5
<0.5
<0.5
2.69
<0.5
OC
1.28
5.11
1.28
2.24
<0.5
<0.5
<0.5
<0.5
<0.5
PCB
1.61
1.49
0.79
<0.5
1.41
0.92
<0.5
0.67
<0.5
Home
PAH
0.65
12.6
5.62
1.89
14.0
29.3
4.00
14.3
13.0
B2 PAH
0.63
0.57
2.47
0.50
<0.5
3.32
<0.5
0.55
<0.5
PE
496
59.1
1060
868
<0.5
219
63.4
113
446
OP
13.7
0.90
6.43
4.10
<0.5
10.1
1.10
23.9
<0.5
OC
0.80
<0.5
10.5
<0.5
<0.5
1.27
3.55
0.62
<0.5
PCB
<0.5
0.61
0.53
1.57
<0.5
0.54
<0.5
<0.5
1.52
(a) The number in the subject code denotes the daycare center that the subject attends; 3 = D03, Head Start and 9 = D09, regular
daycare center.
-------
could be underestimated. For future studies, more field blanks are needed to establish the
background levels if phthalate esters are the target analytes.
Extremely high level of chlorpyrifos (1150 ng/m3) was observed in household HH9
indoors. The questionnaire data showed that indoor insecticides/pesticides application is
performed once every three months by a commercial contractor. Only days before the field
monitoring period, indoor application was conducted. Such a high indoor chlorpyrifos level
could result from the indoor application. However, the commercial contractor claimed that
chlorpyrifos was not in the material used.
The floor dust loadings of the Phase 2 samples are summarized in Table 5.25. Levels of
the fine dust loadings ranged from 1.92 to 8.38 g/m2 in the two daycare centers and from 0.28 to
20.2 g/m2 in the nine households. The fine dust loadings accounted for 28 to 75% of the total
dust loadings in these samples. The dust loadings of the private daycare D09 were slightly
lower than that of the Head Start daycare D03. Higher dust loadings were observed in the four
low-income homes (HA3, HB3, HC3, and HD3) as compared to those from the five middle-
income homes. A similar relationship between the dust loadings and household income was
observed in previous studies (1.5,6). The highest (20.2 g/m2) and the lowest (0.34 g/m2) fine
dust loadings were observed in HA3 (low-income) and HH9 (middle-income), respectively.
The reported POP concentrations (ppm) in dust and soil samples were corrected for the
background levels and for the moisture content. The target POP data of floor dust and
playground soil samples are given in Tables M-3 and M-4 in Appendix M. Concentrations of
target POP in all compound classes in the floor dust samples were higher than those found in the
corresponding soil samples. Most target POP in PAH, PE, and Ph were detected in the dust and
soil samples. The target POP in OP, OC, PCB, and HA were found in most dust samples but
not soil samples.
Levels of B2 PAH in the floor dust samples ranged from 0.196 to 2.56 ppm in the
daycare centers and from 0.213 to 5.32 ppm in the households. Concentrations of B2 PAH in the
soil samples were less than 0.1 ppm in the daycare centers and less than 0.3 ppm in the
households. Levels of B2 PAH found in most dust and soil samples accounted for approximately
half of the total PAH concentrations in these samples. This finding was in agreement with the
88
-------
TABLE 5.25. FLOOR DUST LOADINGS FROM PHASE 2 DAYCARE CENTERS
AND HOUSEHOLDS
Dust Loading, g/m2
Sample Code(a)
Age of Carpet,
year
Total
Fine Fraction
(<150 fim)
Percent of Fine
Fraction, %
D03-1
DKW
11.1
8.38
75
D03-2
DK
11.9
8.05
68
D09-1
0.2
3.50
1.92
55
D09-2
0.2
7.71
5.13
67
HA3
7
27.3
20.2
74
HB3
DK
4.14
1.64
40
HC3
2
4.10
2.06
50
HD3
1
27.2
18.4
68
HE9
3
3.05
1.32
43
HF9
8
1.00
0.28
28
HG9
5
1.51
0.72
48
HH9
11
0.63
0.34
54
HI9
7
0.65
0.39
60
(a) The first letter of the sample code denotes the location of the collected sample, D = Daycare
center, H = Household; the last two letter/number denotes the unique code of the daycare
center of household.
(b) DK denotes don't know.
89
-------
results of the Phase 1 samples and those from the previous studies. Concentrations of the sums
of PE ranged from 0.080 to 16.4 ppm in the dust samples and from <0.001 to 0.216 ppm in the
soil samples. Concentrations of sums of target phenols in the dust and soil samples were from
4.02 to 55.9 ppm and from 0.038 to 0.167 ppm, respectively.
Levels of target POP in OP, OC, PCB, and HA were either not detected or present at low
levels (<0.1ppm) in the soil samples. With few exceptions, concentrations of target POP in OP,
OC, and PCB were generally less than 1 ppm in the dust samples. Note that the highest
concentration (6.45 ppm) of OP, mainly from chlorpyrifos (6.44 ppm) was found in the dust
sample from HH9 where the highest indoor air concentration of this compound was observed.
Target POP data of the solid and liquid food samples are presented in Tables M-5 and
M-6 in Appendix M. Target PAH found in the liquid and solid food samples were mostly 2- to
4-ring PAH. The B2 PAH were detected in none of the solid food samples and only in two of the
liquid food samples. The levels of total target PAH ranged from 0.061 to 2.18 ppb in liquid food
samples and from 1.59 to 7.10 ppb in solid food samples. Higher concentrations of PE were
observed and ranged from 4.25 to 63.1 ppb in liquid food samples and from 14.1 to 461 ppb in
solid food samples. Levels of Ph in the solid food ranged from 10.8 to 80.4 ppb and were less
than 5 ppb in liquid food samples. Levels of OP, OC, and PCB found in most liquid and solid
food samples were less than 1 ppb. The concentrations of total target PAH in the liquid food
samples were, in general, lower than those in the solid food samples. These relative
concentration profiles were also observed in target POP in PE, OP, and Ph. Concentrations of
target POP in OC and PCB in the solid food samples were not always higher than those in the
liquid food samples.
The HA, 2,4-D, was detected in all the liquid food samples and only in three solid food
samples. Concentrations of 2,4-D ranged from 0.353 to 2.01 ppb in the liquid food samples and
from 0.92 to 2.20 ppb in the solid food samples. Note that the recoveries of PCP and 2,4-D in
the spiked solid food samples were only 33% and 20%, respectively. As noted earlier in
Table 5.10, better recoveries of PCP (85%) and 2,4-D (63%) were obtained using the same
analytical method in the method validation phase. The low recoveries of the spiked Phase 2 food
samples could be attributed to differences in the sample matrix and the sample size (50 g versus
-------
10 g). In addition, the recoveries of the spiked SRS 3,4-D were from 20 to 90% in the solid food
samples and from 17 to 44% in the liquid food samples. The reported PCP values were not
corrected for the PCP recovery obtained from the matrix spiked sample, since this recovery was
based on only one matrix spike sample, and there were no SRS available for PCP. Thus, the
reported values for PCP could be underestimated. However, the reported 2,4-D values were
corrected for the SRS (3,4-D) recoveries. The analytical method for the determination of these
compounds in food sample matrix needs to be modified for better overall method precision and
accuracy.
Target POP data of dermal wipe samples from each child subject collected at daycare
centers and at home are presented in Tables M-7 and M-8 in Appendix M. Concentrations of
target POP in PAH, OC, OP, and PCB were in the same order of magnitude among the daycare
wipe samples and the home wipe samples. The concentration ranges of PE were greater than
those observed in the above compound classes, and ranged from <0.5 to 1350 ng/wipe in
daycare-wipe samples and from <0.5 to 1060 ng/wipe in home-wipe samples. Note that high
levels (1190 ng/wipe) of PE were found in the field blank wipe sample. The reported values
were corrected for the background levels. For the household (HH9) having the highest
chlorpyrifos levels in indoor air and house dust, the highest concentration (23.9 ng/wipe) of this
compound was also found in the subject's composite wipe sample collected at home. The level
of chlorpyrifos was 2.69 ng/wipe in the composite wipe sample collected from the same subject
at the daycare center.
Summary Statistics
Summary statistics by compound class, by sample media, by sampling location, and by
family income for target POP are presented in Appendix N. Each table contains sample size,
number of samples below detection limit, mean, standard deviation, minimum, and maximum
values. Table N-l through N-10 summarize the POP data in the multimedia samples collected at
subjects' homes. Table N-l 1 through N-20 summarize the POP data in the multimedia samples
91
-------
collected at daycare centers that subjects attend. Tables N-21 through N-30 summarize the data
by groups of low- and middle-income families.
Average indoor and outdoor concentrations of OP in samples collected at homes were
higher than those collected at daycare centers. The high indoor average of OP is partly from the
high level of chlorpyrifos in the indoor air of HH9. Differences in average concentrations of
POP in other compound classes in indoor and outdoor air between these two groups of samples
(home vs. daycare) were small. Comparable concentrations of POP in all compound classes but
OP were obtained in these two groups of floor dust samples. Average concentration of OP in the
floor dust samples collected at homes was higher than in those collected at daycare centers, but
lower average concentration of Ph was seen in the dust samples from homes. Concentrations of
POP in playground soil samples were similar between these two groups of samples for all
compound classes except OC. Higher OC concentrations were seen in the playground soil
samples collected at homes. Comparable POP concentrations in liquid and solid food were
observed in these two groups of samples. Average levels of chlorpyrifos in dermal wipe samples
collected at homes were more than twice those in the wipes collected at daycare centers.
Concentrations of other POP in dermal wipes were similar in these two groups of samples.
Average indoor concentration of OP (mainly chlorpyrifos) in middle-income homes was
higher than that in the low-income homes. As mentioned before, this is from the high
chlorpyrifos in HH9. Differences between the two groups of samples (low-income vs. middle-
income) in average POP concentrations in other compound classes in air were not significant.
For floor dust samples, higher average concentrations of PAH, OP, and PCB were seen in the
middle-income families while higher concentrations of Ph and HA were seen in the low-income
families. In general, average POP concentrations in playground soil were low and the
differences in the concentrations in soil between low- and middle-income families were small.
There were differences in concentrations of POP in food between the two groups of families.
Average levels of PE in dermal wipe of the low-income subjects were higher than those of
middle-income subjects. This trend was seen in wipe samples collected at homes as well as at
daycare centers. Similar levels of other POP in wipe sample were observed in these two groups
of subjects.
92
-------
Correlatipp Between Sample Msdti
Correlation coefficients of measured target POP in different sample media were
calculated by Spearman and Pearson methods. The results of Spearman and Pearson correlation
coefficients are summarized in Table 0-1 through Table 0-11, as well as in Tables 0-12 through
0-22, respectively, in Appendix O. Table 5.26 summarizes the pairs of sample media showed
significant correlations, at least, at 0.05 or lower confidence levels.
Direct relationships were observed between indoor air and floor dust for target POP in
PAH, OC, and OP. This relationship was also seen in the Phase 1 data. A significant correlation
between indoor air and outdoor air was also observed in B2 PAH, OC, and PCB. Significant
correlations were seen between indoor air and solid food for OC and between outdoor air and
solid food for HA. There were few significant but negative relationships between liquid food
and other sample media such as floor dust and pathway soil.
Correlation Between Compound Classes
The results of Spearman and Pearson correlations between compound classes within each
sample medium are presented in Tables P-l through P-6, and in Tables P-7 through P-12,
respectively, in Appendix P. Table 5.27 summarizes the pairs of compound classes that had
correlation coefficients significant, at least, at 0.05 levels.
The positive correlation between B2 PAH and total PAH was seen in floor dust, solid
food, and playground soil. This relationship was also observed in the Phase 1 samples. Several
pairs of compound classes were correlated with each other in playground soil, which was not
seen in the phase 1 samples. This is probably related to the locations of the collected samples.
For example, direct relationships were also seen between OC and OP in indoor air, as well as
those in floor dust. Because of the variability of types of food, and methods of cooking, we do
not expect a lot of direct relationships between compound classes in the food samples. Only one
pair of compound classes with direct relationships was observed in liquid and in solid food. This
could be due to the sources of the foods, and the cooking methods used.
-------
TABLE 5.26. SUMMARY OF PAIRS OF SAMPLE MEDIA WITH SIGNIFICANT
CORRELATION COEFFICIENTS FOR TARGET COMPOUND CLASSES
Compound Class(1) Pair of Media with Significant Correlation Coefficient00
Spearman Method(c) Log-Transformed Data(d)
B2 PAH
(PS, OA)
(IA, OA)
Target PAH
(IA, FD)**
(OA,FD)
(IA, FD)**
OC
(IA, OA)
(IA, FD)**
(IA, SF)
(IA, OA)**
(IA, FD)**
(IA, SF)
OP
(IA, FD)**
(IA, FD)**
PE
(OA, $F)
PCB
(IA, OA)
(IA, OA)
PH
(PS. FD)
(LF, FD)
(LF, FD)
HA
(OA, SF)**
(LF. FD)
(LF. PS)
(OA, SF)**
(FD. LF)**
(SF, FD)
(LF. PS)
BEN
(IA, PS)
(OA, SF)**
(LF, FD)
CHL
(IA, FD)**
(IA, FD)**
BIS (SF, PS)
(a) PE = phthalate esters, PH = phenols, OC = OC pesticide, OP = OP pesticide,
HA = herbicide acid (2,4-D), BEN = benzylbutylphthalate, CHL = chlorpyrifos, and
BIS = bisphenol-A.
(b) LA = indoor air, OA = outdoor air, FD = floor dust, LF - liquid food, SF = solid food, and
PS = playground soil. All pairs are at least significant at the 0.05 level, while pairs with
0.01 significant level are marked with **, and pairs with negative correlations are
underlined.
(c) Spearman's correlation method in raw data.
(d) Pearson's correlation method in log-transformed data.
94
-------
TABLE 5.27. SUMMARY OF PAIRS OF COMPOUND CLASSES WITH SIGNIFICANT
CORRELATION COEFFICIENTS FOR EACH SAMPLE MEDIUM
Sample Medium
Pair of Compound Classes with Significant Correlation
Coefficient^
Spearman Method^
Log-Transformed Data(c)
Indoor Air
(B2-PAH, PE)
(OC, OP)**
(OC, OP)
Outdoor Air
(B2-PAH, PCB)
(B2-PAH, HA)
(PCB, PE)**
(PCB, OP)
(B2-PAH, PCB)
(PCB, PE)
CPCB. OP>
(HA, OP)
Floor Dust (HVS 3)
(B2-PAH, TARGET
PAH)**
(T32-PAH. HA)
(OC, OP)**
(B2-PAH, TARGET PAH)**
(OC, OP)**
Liquid Food
(PCB, PE)
(PCB, PE)
Solid Food
(TARGET PAH, PE)**
(TARGET PAH, PE)**
Playground Soil
(B2-PAH, TARGET
PAH)**
(OC, PE)
(PCB, OC)
(OC, PH)
(OC, HA)
(PH, HA)
(B2-PAH, TARGET PAH)**
(PCB, OC)**
(OC, PH)
(OC, HA)
(a) PE = phthalate esters, PH = phenols, OC = OC pesticide, OP - OP pesticide, and
HA = herbicide acid (2,4-D). All pairs are at least significant at the 0.05 level, while pairs
with 0.01 significant level are marked with **, and pairs with negative correlations are
underlined.
(b) Spearman's correlation method in raw data.
(c) Pearson's correlation method in log-transformed data.
95
-------
Daily Potential Doses of Persistent Organic Pollutants
The potential daily potential doses of POP were calculated for the inhalation (air),
nondietary ingestion (dust/soil), and dietary ingestion (food) pathways for nine subjects. The
daily POP potential doses were calculated from at-home and at-center exposures separately, as
well as from combined at-home and at-center exposures. The calculated daily potential doses of
POP from at-home exposure, at-center exposure, and combined total exposure, are presented in
Tables M-l 1 through M-13, respectively, in Appendix M. Summary statistics of the daily POP
potential doses are given in Appendix N. It should be noted that these potential doses reported
here are maximum external doses based on the measured concentrations and exposures as
calculated previously.
Figures 5.3 through 5.5 show the distributions of average daily POP potential doses
through the three pathways from exposures at homes, at centers, and at both locations. Similar
distributions of average POP potential doses in the three pathways were seen among exposures at
homes, and at centers for compound classes of total PAH, PE, PCB, Ph, and HA. Dietary
ingestion was the most important pathway for PE, Ph, and HA whereas inhalation was the most
important pathway for total PAH, and PCB. Different distributions of average POP potential
doses of OC, OP, and B2 PAH were observed between at-home and at-center exposures.
Inhalation was the most important pathway for average daily potential doses of OC and OP from
at-home exposures, but dietary ingestion was the most important pathway from at-center
exposures. Nondietary ingestion (at-home exposure) and dietary ingestion (at-center exposure)
were important pathways for average daily potential dose of B2 PAH.
The distributions of subjects' total daily potential doses of POP from both at-home and
at-center exposures differ among compound classes. The relative importance of exposure
pathways was inhalation > dietary > nondietary ingestion for average total daily potential doses
of total PAH, OP, OC, and PCB. For total daily potential doses of PE, Ph, and HA, the relative
importance of exposure pathways was dietary > nondietary ingestion > inhalation. The
nondietary ingestion was the most important pathway for total daily potential dose of B2 PAH,
and followed by dietary ingestion and air.
96
-------
Sum of B2-PAH, ng/kg/day
Bisphenol-A, ng/kg/day
Total Target PCB, ng/kg/day
22%
¦Air
¦ Dust/Soil
~ Food
¦Air
¦ Dust/Soil
~ Food
¦Air
¦Dust/Soil
~ Food
Total Target PAH, ng/kg/day
Total Target PE, ng/kg/day
Total Target Phenols, ng/kg/day
¦Air
¦ Dust/Soil
~ Food
¦ Air
¦ Dust/Soil
~ Food
¦ Air
¦ Dust/Soil
~ Food
Total Target OP, ng/kg/day
86%
Total Target OC, ng/kg/day
2,4-D, ng/kg/day
¦ Air
¦ Dust/Soil
~ Food
¦Air
¦ Dust/Soil
~ Food
¦Air
¦ Dust/Soil
~ Food
Figure 5.3. Distribution of average daily potential dose of persistent organic pollutants from homes for nine subjects.
-------
Sum of B2-PAH, ng/kg/day
¦ Air
¦ Dust/Soil
~ Food
Total Target PAH, ng/kg/day
Bisphenol-A, ng/kg/day
9%
20%
71%
Total Target PCB, ng/kg/day
3%
¦Air
¦ Dust/Soil
~ Food
¦Air
¦ Dust/Soil
~ Food
Total Target PE, ng/kg/day
10%
Total Target Phenols, ng/kg/day
13%
11%
76%
¦ Air
¦ Dust/Soil
~ Food
¦ Air
¦ Dust/Soil
~ Food
¦ Air
¦ Dust/Soil
~ Food
Total Target OP, ng/kg/day
20%
¦ Air
¦ Dust/Soil
~ Food
Total Target OC, ng/kg/day
¦ Air
¦ Dust/Soil
~ Food
2,4-D, ng/kg/day
|-0%
_
99%
¦ Air
¦ Dust/Soil
~ Food
Figure 5.4. Distributions of average daily potential dose of persistent organic pollutants from daycare centers for nine subjects.
-------
Sum of B2-PAH, ng/kg/day
Bisphenol-A, ng/kg/day
19%
Total Target PCB, ng/kg/day
¦ Air
¦ Dust/Soil
~ Food
¦ Air
¦ Dust/Soil
~ Food
¦ Air
¦ Dust/Soil
~ Food
Total Target PAH, ng/kg/day
22%
¦ Air
¦ Dust/Soil
~ Food
76%
Total Target PE, ng/kg/day
12%
¦ Air
¦ Dust/Soil
~ Food
87%
Total Target OP, ng/kg/day
22%
¦ Air
¦ Dust/Soil
~ Food
75%
Total Target OC, ng/kg/day
44%
53°/,
¦ Air
¦ Dust/Soil
~ Food
Total Target Phenols, ng/kg/day
16%
¦ Air
¦ Dust/Soil
~ Food
75%
2,4-D, ng/kg/day
0% 6%
¦ Air
I Dust/Soil
~ Food
94%
Figure 5.5. Distributions of average daily potential dose of persistent organic pollutants from homes and daycare centers for nine subjects.
-------
Figures 5.6 through 5.8 display the average daily potential dose of POP for the two
groups of subjects (low-income vs. middle-income) that resulted from exposures at homes, at
centers, and at both locations. Note that the potential dose levels of PE shown in these figures
need to be multiplied by a factor of 10 because of the high potential dose levels. In general, the
average daily POP potential doses for low- and middle-income subjects were within the same
order of magnitude. Slightly higher average daily potential doses of B2 PAH, total PAH, OC,
PCB, OP, and Ph were seen in the middle-income subjects from the exposures at homes. The
reverse trend was observed for B2 PAH and total PAH from the exposures at daycare centers.
The difference between these two groups of subject's average total daily potential doses of POP
was small, with slightly higher total potential doses of OP and Ph seen in the middle-income
subjects. The general trend for total average potential doses of POP in target compound classes
was PE > Ph > total PAH >HA > OP > OC > PCB > B2 PAH.
Concentration Profiles of Urinary Metabolites
The composite urine samples collected from each subject at daycare center and at home
were analyzed for target hydroxy-PAH, 2,4-D, PCP, and 3,5,6-TCP. Two different analytical
methods were employed, as described in Section 4, one for hydroxy-PAH, PCP, and 2,4-D, and
one for 3,5,6-TCP. The concentrations of individual target analyte measured in the urine
samples expressed in units of ng/mL of urine and |imole/mole of creatinine are presented in
Tables M-9 and M-10 in Appendix M. Summary statistics for the urine data across low-income
and middle-income subjects are given in Tables N-28 and N-29 in Appendix N. Among the
target analytes, 1- and 3-hydroxy benz[a]anthracene, 1- and 3-hydroxy benz[a]anthracene, and
6-hydroxy indeno[l,2,3-cd]pyrene were detected in some urine samples. The remaining target
analytes were found in all urine samples.
The most abundant hydroxy-PAH, urinary PAH metabolites, found in the subjects' urine
samples was l-naphthol. Note that 1-naphthol is an urinary metabolite of naphthalene as well as
carbaryl, a widely used commercial insecticide (17). Thus, 1-naphthol in the subjects' urine
samples could result from the exposures to both compounds. Levels of 2-naphthol in the urine
100
-------
Average Potential Daily POP Doses from Homes in
Low-and Middle-Income Subjects
Low-in come Middle- Low-income Middle- Low-income Middle- Low-income Mfddlo-
B2PAH Income OC Income OC PCB income HA income HA
B2PAH PCB
Average Potential Daily POP Doses from Homes in
Low- and Middle-Income Subjects
Low- Middle- Low- Middle- Low- Middle- Low- Middle-
income income income income income income income income
OP OP PAH PAH Ph Ph PE PE
Figure 5.6. Average potential daily persistent organic pollutants doses from at-home exposures
in low- and middle-income subjects.
101
-------
Average Potential Daily POP Doses from Daycare
Centers in Low-and Middle-Income Subjects
*5)
o
O
Q
a.
o
CL
>>
75
O
Low-income
B2PAH
Middle- Low-income Middle- Low-income Middle- Low-income MkJdl*-
inoome OC income DC PCB income PCB OP Income OP
B2PAH
Average Potential Daily POP Doses from Daycare
Centers in Low- and Middle-Income Subjects
>%
¦ft
0
(A
o
o
&
a.
>>
Low- Middle-
income income
HA HA
Middle
Income
income
Middle- Low- Middle-
income income Income
Ph PE PE
Figure 5.7. Average potential daily persistent organic pollutants doses from at-center exposures
in low- and middle-income subjects.
102
-------
Average Total Potential Daily POP Doses from Homes and
Daycare Centers in Low-and Middle-Income Subjects
Low-in coma Middle- Low-Income Middle- Low-income Middle- Low-Income Middle-
B2PAH income OC income OC PC8 income PCB OP income OP
B2PAH
fr
"ft
JL
c
Ct
s
Q
CL
o
CL
>»
75
Q
Average Total Potential Daily POP Doses from Homes and
Daycare Centers In Low- and Middle-Income Subjects
mm
Hi
¦
Low- Middle- Low- Middle- Low- Middle- Low- Middle-
income income income income income income income income
HA HA PAH PAH Ph Ph PE PE
Figure 5.8. Average potential daily persistent organic pollutant doses from at-home
and at-center exposures in low- and middle-income subjects.
103
-------
samples were lower than that of 1-naphthol, but in general were higher than the levels of most
other hydroxy-PAH.
The most abundant target analyte found in the urine samples was 3,5,6-TCP, an urinary
metabolite of chlorpyrifos. Figure 5.9 displays the average concentrations of 3,5,6-TCP and
2,4-D in the urine samples of the subjects from low- and middle-income families. The
concentrations of these analytes expressed in ^mole/mole and ng/mL are displayed in the upper
and lower portions of the figure, respectively. As shown in Figure 5.9, concentrations (ng/mL
and ^mole/mole) of 3,5,6-TCP and 2,4-D in the subjects' urine samples collected at their homes
were in the same order of magnitudes as those collected at the daycare centers. Similar average
concentrations in ng/mL of these two compounds were seen in the urine samples from low- and
middle-income subjects. But higher creatinine-corrected concentrations of 3,5,6-TCP in
^mole/mole were observed in middle-income subjects, which is consistent with the measured
exposures and estimated potential doses of chlorpyrifos. Figure 5.10 displays the average
concentrations of 1-naphthol, 2-naphthol, and PCP in the urine samples of the low- and middle-
income subjects. Levels of 1-naphthol in the subjects' urine samples collected at their homes
were higher in the low-income subjects than that in the middle-income subjects. The reversed
trend was observed in PCP in the urine samples collected at homes and at daycare centers in
these two subject groups.
Analysis of Variance f ANOVA)
ANOVA models were fitted to B2 PAH, total PAH, OP, OC, PE, Ph and HA as well as
selected individual POP. The results of ANOVA models of POP in all sample media on
sampling location (daycare center vs. home) are presented in Table Q-l in Appendix Q.
Table 5.28 summarizes the results of ANOVA with significant location effect, at least, at 0.05
level, and the respective geometric mean. As shown in Table 5.28, concentrations of OP in
outdoor air in subjects homes were significantly higher than those in the daycare centers.
Significantly higher PCB concentrations in indoor and outdoor air in daycare centers were
observed. The concentrations of total PAH, total target phenols, and bisphenol-A were
significantly higher in liquid food samples collected at daycare centers as opposed to those
collected at homes. There were no statistically significant difference in POP concentrations in
104
-------
Average Concentrations of 2,4-D and 3,5,6-TCP in
Subjects' Urine Samples
25
#>
20
C
01
ro
o
F
15
£
c
"5
4)
o
c
0
0
E
3
10
o
5
0
I Low-Income
I Middle-Income
iK
2,4-D-H 2,4-D-D 3,5,6-TCP-H 3,5,6-TCP-D
Sample Code
Average Concentrations of 2,4-D and 3,5,6-TCP in
Subjects' Urine Samples
Low-Income
Middle-Income
2,4-D-H 2,4-D-D 3,5,6-TCP-H 3,5,6-TCP-D
Sample Code
Figure 5.9. Average concentrations of 2,4-D and 3,5,6-TCP in subjects' urine samples. The
top plot shows the creatinine-corrected concentrations; the bottom plot shows
uncorrected concentrations.
105
-------
Average Concentrations of OH-PAH and PCP in
Subjects' Urine Samples
-------
TABLE 5.28. SUMMARY OF ANOVA ON THE EFFECT OF SAMPLING LOCATION:
DAYCARE CENTERS VERSUS HOMES
Target POP
Sample Medium
Geometric Mean(a)
Daycare Center
Home
Daycare Center^
vs.
Home
Total PAH
Liquid food, ppb
0.728
0.225
*
OP
Outdoor air, ng/m3
1.22
2.16
*
PCB
Indoor air, ng/m3
35.6
12.0
*
Outdoor air, ng/m3
5.12
1.60
**
Ph
Floor dust, ppm
21.3
8.50
*
Liquid food, ppb
1.63
0.216
*
Bisphenol-A
Liquid Food, ppb
0.198
0.062
*
(a) Sample size = 4 from daycare centers and 9 from homes.
(b) * and ** denote the tested effect is statistically significant at 0.05 and 0.01 level, respectively.
-------
playground soil, solid food, dermal wipe, and urine samples between these two groups of
samples.
The effect of family income on concentrations of POP in multimedia samples was also
investigated. The results of ANOVA are presented in Table Q-2 in Appendix Q. Table 5.29
summarizes the results of ANOVA having significant effect on family income. Levels of OP in
liquid food samples collected at low-income homes were higher than those collected at middle-
income homes. Higher concentrations of total target phenols were found in the liquid foods from
middle-income families. As mentioned before, the difference seen in the liquid food may be from
the types of food they consumed. In addition, significantly higher levels of PCP were found in
the urine samples of middle-income subjects. Target POP levels in air, dust, soil, dermal wipe,
and solid food between the low- and middle-income groups did not differ significantly.
The ANOVA was also conducted to investigate the differences in POP levels between two
groups: Head Start center and low-income families versus private daycare center and middle-
income families. The first group is referred to as low-income group and the second one is
referred to as middle-income group. The ANOVA results are given in Table Q-3 in Appendix Q.
Table 5.30 summarizes the results of ANOVA models showing significant difference in POP
concentrations between these two groups. The total phenols concentrations in liquid food were
higher in middle-income group as opposed to low-income group. Levels of bisphenol-A in floor
dust were also higher in the middle-income group. For dermal wipe samples, significantly higher
levels of benzylbutylphthalate and OC but lower levels of total PAH were observed in the low-
income group. Concentrations of total target hydroxy-PAH and 2,4-D in the urine samples from
the low-income subjects were higher than that from the middle-income subjects. The reversed
concentration trend was observed for PCP and 3,5,6-D in the urine samples between these two
groups.
Regression Models
The results of the estimated relationships between measured urinary metabolites and
measured POP concentrations based on the fitted models are summarized in Table R-l in
108
-------
TABLE 5.29. SUMMARY OF ANOVA ON THE EFFECT OF FAMILY INCOME:
LOW-INCOME VERSUS MIDDLE-INCOME
Geometric Mean(a) Low-Income(b)
Target POP Sample Medium vs.
Low-Income Middle-Income Middle-Income
OP
Liquid food, ppb
0.141
0.020
*
Ph
Liquid food, ppb
0.050
0.696
*
PCP
Urine, nmole/mole
0.299
0.718
**
(a) Sample size = 4 from low-income and 9 from middle-income homes.
(b) * and ** denote the tested effect is statistically significant at 0.05 and 0.01 level, respectively.
-------
TABLE 5.30. SUMMARY OF ANOVA ON THE EFFECT OF FAMILY INCOME AND SAMPLING LOCATION:
HEAD START CENTER AND LOW-INCOME VERSUS REGULAR CENTER AND MIDDLE-INCOME
Target POP
Sample Medium
Geometric Mean(a)
Head Start Center Regular Center
Low-Income Middle-Income
Head Start Center/Low-Income^
vs.
Regular Center/Middle-Income
Ph
Liquid food, ppb
0.126
1.08
*
Bisphenol-A
Floor dust, ppm
1.11
1.90
*
Total PAH
Wipe, ng/wipe
4.88
13.6
*
Benzylbutylphthalate
Wipe, ng/wipe
186
8.66
*
OC
Wipe, ng/wipe
1.33
0.420
*
OH-PAH
Urine, ng/mL
1.71
0.483
*
2,4-D
Urine, ng/mL
2.89
1.62
*
PCP
Urine, (imole/mole
0.208
0.652
**
3,5,6-TCP
Urine, (imole/mole
7.76
15.4
*
(a) Sample size = 6 from Head Start center and low-income families and 7 from regular center and middle-income families for food and
dust samples; 8 from the first group and 10 from the second group for wipe and urine samples.
(b) * and ** denote the tested effect is statistically significant at 0.05 and 0.01 level, respectively.
-------
Appendix R. Each row in the table represents a distinct analysis. The intercepts for the fitted
regression models are shown in the second column, and the slopes (estimated parameters) are
displayed in column three to six. The slope represents the estimated increase in log-transformed
urinary metabolites due to an increase of the log-transformed POP concentration in one sample
medium. For each regression model, a test was conducted to determine whether the relationship
between the urinary metabolite and the POP concentrations was different between sampling
location (at-home vs. at-center). The last column indicates if the test was statistically significant.
The regression results of the relationships between measured levels of urinary metabolites and
total daily POP potential doses, using the same format as Table R-l, are summarized in Table R-2
in Appendix R. Table 5.31 summarizes the regression models with statistically significant results.
Note that the format of Table 5.31 is similarly as described above for Table R-l except that the last
two columns show the r2 and p- value of the respective regression model.
The results of all regression models showed that there was no statistically significant
relationship between sampling location (at-home and at-center) and all measured urinary
metabolites. This is partly because POP concentrations in multiple sample media did not vary
significantly between homes and daycare centers.
As shown in Table 5.31, the measured values of PCP in urine were significantly related to
the measured PCP values in indoor air or floor dust. This positive relationship was also shown
between subject's total daily potential dose of POP and PCP level in the subject's urine. This
finding suggested that the measured PCP concentrations in indoor air or floor dust in subjects'
microenvironment are related to the PCP found in subject's urine. Thus, the measurement of
PCP in environmental samples can be an indicator for subject's exposure to PCP. Given the
small sample size, more samples are needed to establish such relationships between the PCP in
environmental media and PCP in urine.
None of the PAH concentrations in multimedia samples, as well as the total daily
potential dose of PAH were significantly related to hydroxy-PAH, urinary metabolites of PAH.
A weak relationship was observed between chlorpyrifos in multimedia samples and 3,5,6-TCP,
an urinary metabolite of chlorpyrifos, in urine samples. Subjects' total daily potential dose of
chlorpyrifos was not related significantly to 3,5,6-TCP in the subjects' urine samples. A
111
-------
TABLE 5.31. SUMMARY OF REGRESSION MODEL RESULTS
Urinary Regression Model
Metabolite Parameter Estimates(a) I
r p-value
Intercept IA OA FD FO
PCP
ng/mL -14.585 0.400** 0.261 0.008 -8.640 0.758 0.002
^imole/mole -0.690 -0.247 0.228 1.042* -1.891 0.581 0.041
Parameter Estimates Regression Model
Intercept Total Daily POP Potential dose r2 p-value
PCP
ng/mL -1.116** 0.428** 0.649 0.0004
^imole/mole -1.263** 0.426** 0.261 0.104
(a) * and ** denote that the estimated parameter (slope) is significantly different from zero at
0.05 and 0.01 level, respectively; IA = indoor air, OA = outdoor air, FD = floor dust, and
FO = liquid food and solid food.
112
-------
negative relationship between 2,4-D in urine and total daily potential dose of 2,4-D was obtained,
but the r2 value of this regression model was only 0.351 at the p value of 0.04. The absence of
significant relationships between urinary metabolites levels and the respective POP levels in a
multimedia samples as well as the metabolite levels and the POP total daily potential dose levels
may be due to the small sample size and the variation of elimination rate of urinary metabolites.
Quality Control Data for Phase 2 Study
The recovery data of the spiked POP in Phase 2 multimedia samples are summarized in
Appendix S. Acceptable recoveries for spiked POP were obtained in most of the sample media
except for food. Greater than 100% recovery (183%) was found in one air sample. This was
from the coeluting interference peak in GC/ECD analysis. Recoveries of 3,4-D were low (<50%)
in most food samples. It is possible that 3,4-D was lost during the GPC cleanup step. Low
recoveries of the spiked 13C-DDE, and nC-PCB were found in few dermal wipe samples. This
may be due to the losses through the SPE cleanup step. For the same reason, slightly lower
average recoveries of the spiked POP were found in the dust/soil samples when compared to
those obtained from Phase 1 study.
The results of field blanks in each sample medium are given in Appendix T. In general,
trace amounts of target POP were found in the field blank except for phthalate esters. The levels
of PE in the Phase 2 field blanks were higher than that in Phase 1 field blanks. Among the two
target PE, higher background levels were seen of dibutylphthalate as opposed to
benzylbutylphthalate. Overall, the field blank data showed that there was no significant
contamination due to sample handling and preparation.
113
-------
References
1. Chuang, J.C., Callahan, P.J., Menton, R.G., Gordon, S.M., Lewis, R.G., Wilson, N.K.
Monitoring methods for polycyclic aromatic hydrocarbons and their distribution in house
dust and track-in soil. Environ. Sci. Technol. 29(2):494-500,1995.
2. Nishioka, M.G., Burkholder, H.M., Brinkman, M.C., Gordon, S.M., Lewis, R.G.
Measuring transport of lawn-applied herbicide acids from turf to home: correlation of
dislodgeable 2,4-D turf residues with carpet dust and carpet surface residues. Environ.
Sci. Technol., 30, 3313-3320, 1996.
3. Wania, F. and MacKay, D. Tracking the distribution of persistent organic pollutants.
Environ Sci. Technol 30(9):390A-396A, 1996.
4. Stanek, E.J., m and Calabrese, E J. Daily estimates of soil ingestion in children.
Environ. Health Perspectives, 103(3): 276-285, 1995.
5. Chuang, J.C., Callahan, P.J., Lyu, C.W., Pennybacker, M.R. Characterization of
polycyclic aromatic hydrocarbon exposure among children of low income families from
inner city and rural areas. Final Report (Year 2) to U.S. EPA, Cooperative Agreement
CR 822073, October 1995.
6. Chuang, J.C., Callahan, P.J., and Lyu, C.W. Field method evaluation of total exposure of
children from low-income families that include smokers to polycyclic aromatic
hydrocarbons. Final Report to U.S. EPA, Contract Number 68-D4-0023, Work
Assignment No. 9, July 1996.
7. Chuang, J.C. and Kenny, D.V. Method validation for measurement of selected
semivolatile phenols in dust and soil, EPA/600/SR-97/027, March 1997.
8. Wilson, N.K., Barbour, R.K., Chuang, J.C., and Mukund, R. Evaluation of a real-time
monitor for particle-bound PAH in air, Polycyclic Aromatic Compounds, 5:167-174,
1994.
9. Mukund, R. and Chuang, J.C. Field and laboratory evaluations of a real-time PAH
analyzer, EPA/600/SR-97/034, July 1997.
114
-------
10. High Volume Small Surface Sampler HVS3: Operation Manual. Cascade Stack
Sampling Systems (CS3), Inc., Bend, Oregon, January 13, 1992.
11. Standard Practice for Collection of Dust for Chemical Analysis. Standard Method
D.5438-94, Annual Book of ASTM Standards Volume 11.03, American Society for
Testing and Materials, Philadelphia, pp. 517-523,1997.
12. Occupational and Residential Exposure Test Guidelines (1994, 1997). U.S.
Environmental Protection Agency Office of Prevention, Pesticides, and Toxic
Substances, Washington, DC, Series 875.
13. Barclays Official California Code of Regulations (Barclays Law Publishers. South San
Francisco, CA, 1996). "Level of exposure to carcinogens." Vol. 28, § 12721.
14. Lewis, R.G., Fortmann, R.C., and Camann, D.E. Evaluation of methods for monitoring
the potential exposure of small children to pesticides in the residential environment,
Arch. Environ. Contain. Toxicol. 26:37-46,1994.
15. Stanek HI, E.J. and Calabrese, E.J. Soil ingestion estimates for use in site evaluations
based on the best tracer method, Human and Ecological Risk Assessment, 1(2):133-156,
1995.
16. Chuang, J.C., Hannan, S.W., and Wilson, N.K. Field comparison of polyurethane foam
and XAD-2 resin for air sampling for polynuclear aromatic hydrocarbons. Environ. Sci.
Technol. 21(8):798-804,1987.
17. Shealy, D.B., Barr, J.R., Ashley, D.L., Patterson, Jr., D.G., Camann, D.E., and Bond,
A.E. Correlation of environmental carbaryl measurements with serum and urinary
1-naphthol measurements in a farmer applicator and his family, Environ. Health
Perspectives 105(5):510-513,1997.
115
-------
NF.RTUTP—O—S 79 TECHNICAL REPORT DATA
1. REPORT NO.
600/R-98/164a
2.
> iminimi ii1 mi
PB99-134 934
4. TITLE AND SUBTITLE
Evaluation and Application of Methods for Estimating Children's
Exposure to Persistent Organic Pollutants in Multiple Media
Volume I: Final Report
5.REPORT DATE
6.PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Jane C. Chuang et al.
8.PERFORMING ORGANIZATION
REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle Memorial Institute
505 King Avenue
Columbus, Ohio 43201
10.PROGRAM ELEMENT NO.
Projects E0608 and E0460
11. CONTRACT/GRANT NO.
Contract 68-D4-0023
12. SPONSORING AGENCY NAME AND ADDRESS
National Exposure Research Laboratory ,
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
13.TYPE OF REPORT AND PERIOD
COVERED
Research Report
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Field methods for determining children's exposure to selected persistent organic pollutants (POP), including
polycyclic aromatic hydrocarbons and other semi-volatile organic compounds (SVOC) were evaluated and applied
to estimate the ranges of potential exposures through air, dust, and food, of a small set of children from low-incomc
and middle-income families. A field study was conducted at nine day care facilities, and a second field study was
conducted which measured the total exposures to multiple compound classes - polycyclic aromatic hydrocarbons
(PAH), polychlorinated biphcnyls (PCB), phthalate esters (PE), phenols (Ph), organochlorine (OC) pesticides,
organophosphate (OP) pesticides, and a herbicide acid (HA) - of nine children selected from two of the day care
centers. Ingestion, both dietary and nondietary, was a primary route of exposure for many of the compounds, but
other routes were also important, depending on the compound class.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/ OPEN ENDED
TERMS
c.COSATI
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This
Report)
Unclassified
21.NO. OF PAGES
20. SECURITY CLASS (This Page)
Unclassified
22. PRICE
------- |