xvEPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/6OO/R-97/1 Off
November 1997
Field Evaluation of
Screening Techniques for
Poly cyclic Aromatic
Hydrocarbons,
2,4-Diphenoxyacetic
Acid, and
Pentachlorophenol in Air,
House Dust, Soil, and Total
Diet
-------
-------
EPA/600/R-97/109
November 1997
Field Evaluation of Screening Techniques for
Polycyclic Aromatic Hydrocarbons,
2,4-Diphenoxyacetic Acid, and
Pentachlorophenol in
Air, House Dust, Soil, and Total Diet
by
Jane C. Chuang, Ying-Liang Chou, Marcia Nishioka
Kimberlea Andrews, Mary Pollard, and Ronald Menton
Battelle
505 King Avenue
Columbus,Ohio 43201-2693
Contract Number 68-D4-0023
Work Assignment 1-04
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
Printed on Recycled Paper
-------
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.
11
-------
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 polycyclic aromatic
hydrocarbons 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 micro environments. The efforts described in
this report provide an important contribution to our capability to measure and evaluate human
exposure to pollutants.
Gary J. Foley
Director
National Exposure Research Laboratory
111
-------
Abstract
The objectives of this work assignment were to evaluate ELISA screening methods and
determine whether these methods indicate effectively those microenvironments where high
exposure to polycyclic aromatic hydrocarbons (PAH) or other semivolatile organic compounds
(SVOC) is likely.
Four commercially available assay kits for PAH, carcinogenic PAH (C-PAH), 2,4-D, and
pentachlorophenol (PCP) were evaluated. The testing procedures were refined based on the
evaluation results. The overall method precision and assay precision of each ELISA testing
method were determined. The dust/soil samples as well as sample extracts of air and food
samples collected from 13 low-income homes in the summer of 1995 were analyzed by PAH
and C-PAH assays. These sample extracts were also analyzed by gas chromatography/mass
spectrometry (GC/MS) to determine alkyl PAH and phthalates. The dust/soil samples from 13
low-income homes collected during the spring of 1996 were analyzed by PAH, C-PAH,
2,4-D, and PCP assays. Different aliquots of these samples were analyzed by conventional
(GC/MS) methods for PAH and by GC with electron capture detection (GC/ECD) for 2,4-D
and PCP. The ELISA data were compared with GC/MS data or GC/ECD data. For PAH
measurements, there is no strong relationship between the ELISA results and GC/MS results
when data of similar types of samples were combined from different field studies. The ELISA
data (C-PAH) and GC/MS (B2 PAH) data showed stronger relationships for dust/soil collected
from 22 NHEXAS homes. The ELISA screening for PAH can indicate the likely presence of
high levels of PAH in dust/soil samples. There is a positive but weak relationship between
GC/ECD data and ELISA data for 2,4-D and PCP.
This report was submitted in fulfillment of Work Assignment 1-04, Contract 68-D4-0023 by
Battelle under the sponsorship of the United States Environmental Protection Agency. This
report covers a period from May 1, 1996 to September 30, 1996, and work was completed as
of September 30, 1996.
IV
-------
Contents
Foreword iii
Abstract iv
Acknowledgment . . . . x
Chapter 1 Introduction .....,,.., 1
Chapter 2 Conclusions 4
Chapter 3 Recommendations 7
Chapter 4 Experimental Procedure 8
Method Evaluation of PAH and Carcinogenic
PAH (C-PAH) ELISA 8
Sample Preparation for Conventional PAH Analysis 10
GC/MS Analysis Method 10
Statistical Analysis 11
Method Evaluation of 2,4-D and PCP ELISA 13
Chapter 5 Results and Discussion . . . 16
Evaluation of ELISA for Screening PAH and C-PAH 16
Alkyl PAH in Multi-Media Samples 23
Phthalates in Multi-Media Samples 26
PAH in Dust/Soil Samples 30
Relationship Between B2 PAH, Total Target PAH,
and SVOC 34
Relationships of PAH Among Different Sample Media 38
Comparison of PAH Data from ELISA and GC/MS 40
Evaluation of ELISA for Screening 2,4-D and PCP 55
Quality Control Data . . , 63
References
71
-------
Appendices
Appendix A. Soil Screening Method Measuring PAH by Immunoassay
Appendix B. Revised Soil Screening Method Measuring PAH by Immunoassay
Appendix C. Alkyl PAH and PAH Data in Indoor and Outdoor Air Samples
Appendix D. Alkyl PAH Data in House Dust, Entryway Dust, and Pathway Soil Samples
Appendix E. Alkyl PAH Data in the Food Samples
Appendix F. Phthalates Data in Indoor and Outdoor Air Samples
Appendix G. Phthalates Data in House Dust, Entryway Dust, and Pathway Soil Samples
Appendix H. Phthalates Data in the Food Samples
Appendix I. PAH Data in House Dust, Entryway Dust, and Pathway Soil Samples
Appendix J. PAH Data in NHEXAS House Dust, Foundation Soil, and Yard Soil
Samples
Appendix K. Listing of ELISA Total PAH and C-PAH Responses and GC/MS Target
PAH, B2-PAH, and Alkylated PAH Responses
Appendix L. Distribution of Data for Dust, Soil, Food, and Air Samples
Appendix M. Summary Statistics of ELISA and GC/MS PAH Responses for Various Dust
and Soil Sample Types, Food Samples, and Air Samples
Appendix N. The Square of Correlation Coefficients for All Possible Combinations of the
Data
Appendix O. Concentration of 2,4-D in Dust and Soil
Appendix P. Concentration of Pentachlorophenol in House Dust
VI
-------
5.1
5.2
5.3
5.4
5.5
Figures
Relationships of B2PAH and total target PAH in dust/soil
samples from 13 low-income homes , : ....!.. 36
Relationship of B2PAH and total target PAH in dust/soil
samples from 22 NHEXAS homes 37
Scatter plots of ELISA total PAH versus GC/MS total target PAH 43
Scatter plots of ELISA C-PAH versus GC/MS B2PAH 44
ELISA total PAH versus GC/MS total PAH, log scale . 54
Vll
-------
Tables
4.1 Definitions of Performance Characteristics 12
5.1 ELISA results for the house dust sample 17
5.2 Comparison of sonication and shaking extraction method for ELISA 18
5.3 Overall method precision of ELISA screening for PAH and C-PAH 20
5.4 Assay Precision for the PAH and C-PAH ELISA Screening 20
5.5 Recoveries of phenanthrene and benzo[a]pyrene from dust/soil samples .... 21
5.6 Summary of alkyl PAH concentrations (hg/m3) in indoor and
outdoor air samples 24
5.7 Summary of alkyl PAH concentration (ppm) in house dust, entry way
dust, and pathway soil samples 25
5.8 Summary of alkyl PAH concentrations (ppb) in food samples 27
5.9 Summary of phthalate concentrations (ng/m3) in indoor and
outdoor air samples 28
5.10 Summary of phthalate concentrations (ppm) in house dust, entryway
dust, and pathway soil samples 29
5.11 Summary of phthalate concentrations (ppb) in food samples 31
5.12 Summary of PAH concentrations (ppm) in house dust, entryway
dust, and pathway soil samples 32
5.13 Summary of PAH concentrations (ppm) in house dust, foundation
soil, and yard soil 33
Vlll
-------
Tables (Continued)
5.14 Correlation coefficients (r) between B2PAH and total target
PAH in each sample medium j . .. 35
5.15 Correlation coefficients (r) for total target PAH in different
sample media 39
5.16 Correlation coefficients (r) for total target PAH and for B2PAH
in house dust, foundation soil, and yard soil . . 40
5.17 Number of samples analyzed by ELISA and GC/MS methods 41
5.18 Results of paired t-test for the difference between log
(PAH ELISA Response) and log (PAH GC/MS
Response) for dust and soil samples . 45
5.19 Regression analysis results for the dust and soil samples,
combination of North Carolina and NHEXAS study homes 47
5.20 Regression analysis results for the dust and soil samples,
separation of North Carolina and NHEXAS study homes 48
5.21 Frequency distribution of ELISA and GC/MS measurements on the
combination of dust (HD+ES+FDB) and soil
(PS+FSP+YSP) samples . .... 50
5.22 Frequency distribution of ELISA and GC/MS measurements on
food samples , 51
5.23 Frequency distribution of ELISA and GC/MS measurements on
air samples 52
5.24 Extraction and recovery efficiency using 2,4-D extraction
solvent (75% methanol) 56
IX
-------
Tables (Continued)
5.25 Extraction and recovery efficiency using PCP extraction solvent
(NaOH in 75% methanol) 57
5.26 Comparison of concentration of 2,4-D in house dust with different
extraction solvents 59
5.27 Comparison of concentration of PCP in house dust with different
extraction solvents 60
5.28 Precision of replicate ELISA measurements of soil and house
dust extracts 61
5.29 Comparison of 2,4-D concentrations in extracts using GC/ECD
and ELISA 64
5.30 Comparison of PCP in house dust extracts using GC/ECD
and ELISA 65
5.31 Levels of alkyl PAH and phthalates found in field blanks 66
5.32 Summary of recovery data of spiked perdeuterated PAH
in dust/soil samples 68
5.33 Summary of PAH, C-PAH, 2,4-D, and PCP ELISA
calibration data 69
-------
Acknowledgment
We thank Dr. Nancy K. Wilson of the U.S. EPA for her invaluable advice and participation
during this investigation.
XI
-------
-------
Chapter 1
Introduction
In 1992, the National Academy of Sciences identified polycyclic aromatic hydrocarbons (PAH)
and other semivolatile organic compounds (SVOC) as among the highest priorities for
exposure research, in part because these compounds are frequently constituents of fine aerosol
and some of them are mutagens and probable human carcinogens (1). Additionally, several of
the PAH and other SVOC, including phthalates, pentachlorophenol, and 2,4-D, are likely to
be endocrine disrupters or have other quasi-hormonal or reproductive effects. Therefore, it is
imperative that the identities, concentrations, and distributions of these compounds in the
environment be investigated. Determining exposure to PAH and SVOC is still a new area of
research. It is still largely unknown how they are distributed among the vapor and particulate
phases in air or the aqueous and nonaqueous phases in water. Likewise, their distributions and
levels in other media, such as food or soil are largely unknown. Because of the extensive and
costly sampling and analysis efforts that are required to obtain complete information on these
levels and distributions, it is desirable to apply fast, inexpensive screening methods to indicate
those environments and media that are most likely to be significant sources of human or
ecological exposure to PAH and SVOC.
Enzyme-linked immunosorbent assay (ELISA) techniques are currently available commercially
for analysis of water and soil for PAH and for other SVOC. For example, Ohmicron
Environmental Diagnostics, Inc., and the Immunosystems division of Millipore, Inc.,
currently market immunoassay testing kits intended for field screening applications (2-4). The
test kits from Ohmicron utilize the suspended magnetic particle competition assay format, as
1
-------
opposed to a well-coated competition assay format from Millipore. These immunoassays are
formatted to be used only for determining whether a given sample contains PAH at a
concentration above or below a set threshold value.
The objectives of this work assignment were to evaluate low-cost ELISA screening methods
and determine whether application of these methods indicates effectively those micro-
environments where high exposure to PAH and other SVOC is likely.
In this work assignment, ELISA techniques were evaluated for applicability to screening of air
particle sample extracts and food sample extracts generated from EPA Cooperative Agreement
CR822073. Simplified and cost effective sample preparation methods for dust/soil samples
were also evaluated for ELISA. Two different ELISA systems, one for total PAH and one for
carcinogenic PAH (C-PAH), were included in this study. In addition, two other ELISA
systems were evaluated for screening pentachlorophenol (PCP) and 2,4-dichlorophenoxyacetic
acid (2,4-D) in dust/soil samples.
This work assignment was carried out simultaneously with a portion of the NHEXAS Arizona
pilot study, which is being conducted jointly by the University of Arizona, Battelle, and the
Illinois Institute of Technology. Samples of dust/soil from 22 homes of the NHEXAS study
(5) and from 13 homes of low-income families in North Carolina (6,7) were tested by both
PAH and C-PAH ELISA systems. Different aliquots of these samples were analyzed
conventionally by gas chromatography/mass spectrometry for PAH. The results of the ELISA
screening and conventional measurements were compared to determine the ability of the
ELISA techniques to predict microenvironmental levels of PAH and other SVOC in house dust
and soil.
It is desirable to know whether high PAH levels in the dust/soil are indicators of high levels of
other SVOC in the same environmental media, because of the costly and extensive sampling
-------
and analysis efforts that are required to obtain complete information on the levels of pollutants
in multi-media samples. We, therefore, reanalyzed the sampled extracts of air, dust, soil, and
food generated from the EPA Cooperative Agreement (CR822073) by GC/MS for alkyl PAH
and phthalates.
The specific tasks that were planned to accomplish the study objectives are: I
(1) Evaluate two different ELISA systems, one for total PAH and one for
carcinogenic PAH, with dust and soil samples, as well as with sample extracts of
air and food samples collected under CR822073.
(2) Evaluate the ELISA systems for screening PCP and 2,4-D in dust and soil
samples.
(3) Analyze sample extracts of air, dust, soil and food (a total of 95 sample extracts)
collected under CR822073 for alkylated PAH and phthalates.
(4) Screen extracts of dust and soil samples (a total of 102 samples) from
22 NHEXAS homes and 13 low-income homes using ELISA methods.
(5) Analyze above dust and soil samples for PAH by conventional solvent extraction
and GC/MS analysis.
(6) Conduct statistical analysis of ELISA screening results and GC/MS results to
determine whether the ELISA technique is an effective screening tool for total
PAH exposure.
(7) Prepare a final report on the results of the study in EPA/ORD format.
1-04.
This final report summarizes the work conducted for this study under Work Assignment
-------
Chapter 2
Conclusions
The procedures from the commercial testing kits for PAH and C-PAH assays were revised to
provide adequate extraction efficiency of PAH from dust/soil. The overall precision of these
revised methods expressed as percent relative standard deviation (%RSD) of triplicate real-
world dust/soil samples was within ±30% for PAH ELISA and ± 25% for C-PAH ELISA.
The overall method accuracy for the PAH and C-PAH assays cannot be assessed for real-world
dust/soil samples (which contain multiple components of PAH), because the spike recovery
procedures are based on single component spiking: phenanthrene for PAH ELISA and
benz[a]pyrene (BaP) for C-PAH ELISA. The recoveries of phenanthrene and BaP from
dust/soil samples ranged from 68 to 150 percent and from 110 to 130 percent, respectively.
The sample extracts of indoor and outdoor air samples collected from 13 low-income homes in
previous studies (6,7) were analyzed by GC/MS for alkyl-PAH and phthalates. Among these
13 homes there were 9 nonsmokers' homes and 4 smokers' homes. Approximately half of the
homes were located in the inner city (5 nonsmokers and 2 smokers) and half of these homes
were located in rural areas (4 nonsmokers and 2 smokers). Levels of 2- to 3-ring alkyl PAH
in indoor air from these homes were higher than those in the corresponding outdoor air.
Similar concentrations of most 4- to 6-ring alkyl PAH were observed in indoor and outdoor air
for nonsmokers' households, whereas higher concentrations were in indoor air for smokers'
households. Higher outdoor concentrations were observed in the inner city as compared to the
rural area. The sums of alkyl PAH concentrations ranged from 369 to 3,270 ng/m3 in indoor
air and from 49.9 to 702 ng/m3 in outdoor air. With few exceptions, the relative
-------
concentrations trend for alkyl PAH found in dust/soil samples from these homes was house
dust > entryway dust > pathway soil, as was also observed for their parent PAH. The sums
of alkyl PAH concentrations in these samples ranged from 0.092 to 3.32 ppm. Concentrations
of alkyl PAH found in the 24-h food composite samples ranged from 0.866 to 15.6 ppb.
Indoor phthalate concentrations were higher than the corresponding outdoor levels. Total
- " *
target phthalate concentrations ranged from 1,160 to 5,330 ng/m3 in indoor air and from 64.2
to 1,070 ng/m3 in outdoor air. The general concentration trend for phthalates in dust/soil
samples was similar to those of PAH and alkyl PAH. Concentrations of total target phthalates
found in the 24-hr liquid and solid composite food samples ranged from 0.09 to 245 ppb.
The dust and soil samples collected from 13 low-income homes (6,7) and 22 NHEXAS
homes (5) were extracted, and analyzed by GC/MS for 19 target PAH. The B2 PAH
(probable human carcinogens), included among the target PAH are benzo[a]anthracene,
chrysene, benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[a]pyrene,
indeno[l,2,3-c,d]pyrene, and dibenz[a,h] anthracene. The levels of the sums of these B2 PAH
correlated well (correlation coefficient >0.90) with total target PAH (the sums of 19 target
PAH) in dust/soil samples collected from 13 low-income homes and 22 NHEXAS homes. The
results from GC/MS analysis showed that levels of the sums of B2 PAH account for
approximately half of the total PAH. There were positive but weak relationships of PAH
among different sample media (dust, soil, and air). Stronger relationships between dust and
soil samples collected from 22 NHEXAS homes were observed. Thus, house dust may be
used as a potential indicator for other sample media for PAH exposure. More studies are
needed to test this hypothesis.
Different aliquots of the above dust and soil samples were extracted and analyzed by PAH and
C-PAH assays. Statistical analysis results showed that PAH data in dust/soil samples
generated from ELISA and GC/MS methods are significantly different. In general, PAH
-------
ELISA responses were higher than PAH GC/MS responses. The regression analyses showed
that the linear relationship between ELISA and GC/MS measurements is not strong. This
relationship became stronger when the data from each type of samples were treated separately.
This finding suggested that the results of ELISA depends strongly on the sample matrices.
The screening performance of ELISA was evaluated based on the frequency distribution of
ELISA and GC/MS data. The results indicated that PAH and C-PAH ELISA can be used as
only a screening tool but not quantitative analytical method for total PAH and B2 PAH in
real-world dust and soil samples.
The precision for the 2,4-D assay was better than those for the PCP assay in both dust and soil
matrices. The average assay precision was within 20% for the 2,4-D assay and greater than
60% for the PCP assay. There was a positive but weak relationship between GC/ECD and the
ELISA method for 2,4-D data as well as for PCP data. Positive biases for 2,4-D and PCP in
most house dust samples were observed by ELISA as compared to GC/ECD.
-------
Chapter 3
Recommendations
The results of this study suggest that ELJSA results are matrix dependent. The performance of
ELISA screening could be improved by minimizing the matrix effect through a selective
extraction method. We recommend that a study be conducted to investigate an alternative
extraction method, supercritical fluid extraction (SFE) coupled with ELISA for estimating
PAH in dust/soil. Various SFE conditions need to be evaluated to determine the optimal SFE
condition. Different SFE conditions may be needed for different types of dust/soil samples.
The dust/soil samples would be extracted by SFE under the optimal conditions and analyzed by
ELISA. The ELISA results would be compared with GC/MS results to determine whether
SFE coupling with ELISA can provide better estimates of PAH in dust/soil.
With the PCP and 2,4-D assays, the 60% extraction efficiency seems to limit the accuracy and
precision of the method. Therefore, we recommend consideration of a solvent mixture, such
as acetonitrile/phosphate buffer which can quantitatively remove 2,4-D and PCP from
dust/soil. The compatibility of this solvent needs to be evaluated in the PCP and 2,4-D assays.
We also recommend that a study be conducted to investigate a cost-effective sample
preparation method for air and food samples for ELISA because sample preparation is the most
significant time-and cost-consuming step.
-------
Chapter 4
Experimental Procedure
Method Evaluation of PAH and
Carcinogenic PAH (C-PAH) ELISA
The PAH and carcinogenic PAH (C-PAH) assay kits were purchased from Ohmicron
Environmental Diagnostics. Initially, selected pathway soil samples were assayed for
screening PAH and C-PAH using the test kit procedures provided by Ohmicron Environmental
Diagnostics. These procedures are described in Appendix A.
Preliminary method evaluation tests were conducted using selected house dust, entryway dust
and pathway soil samples. The procedural conditions that were evaluated included the ratio of
dust/soil to solvent volume, the extraction techniques, and the ELISA diluent volume. The
test kit procedures involved extracting 10 g of soil with 20 mL of methanol by 1-min shaking.
Because 10 g of house dust may not always be available, alternative quantities of dust mass
and solvent (1 g/20 mL, 3 g/20 mL, and 9 g/20 mL) were evaluated. In these experiments, no
concentration steps were performed prior to ELISA. Two extraction methods, shaking and
sonication, were evaluated for removing PAH from the dust/soil sample matrices. The
shaking method followed the test kit procedures (Appendix A). The sonication method
consisted of two sequential 10-min extractions of the soil/dust sample by two aliquots of
10 mL of methanol. The methanol extracts were combined, filtered by quartz fiber filters and
assayed for PAH and C-PAH ELISA. For the 50 fold dilution of ELISA, 25 /*L of extract
was diluted into 1.225 mL of diluent instead of 250 /*L of extract diluted into 12.25 mL of
-------
diluent as described in the test kits. This revision provided sufficient quantities for assays and
reduced the quantities of chemical wastes generated from the assays.
The procedures used for ELISA were then modified by using a smaller sample size (1 g
instead of 10 g of sample), better extraction method (sonication instead of shaking), and small
amounts of diluents (1.225 mL instead of 12.25 mL). The revised procedures are described in
Appendix B.
Evaluation of recoveries of PAH from dust/soil samples using PAH and C-PAH ELISA was
also conducted. Aliquots of selected dust/soil samples were spiked with known amounts of
PAH and assayed for PAH and C-PAH by ELISA using the revised procedures (Appendix B).
The spiking conditions evaluated were: phenanthrene only, benzo[a]pyrene only,
phenanthrene-d,0 only, a mixture of phenanthrene and benzo[a]pyrene, and a mixture of 16
PAH.
The air and food sample extracts from the Cooperative Agreement (CR822073) were prepared
for ELISA. The air sample extracts were from 24-hr indoor and outdoor air samples, and the
food sample extracts were from 24-hr liquid and solid composite samples from meals
consumed by the study subjects. Aliquots of the air and food sample extracts were removed,
evaporated under a gentle nitrogen stream to dryness, and redissolved into methanol. This step
was required because the sample extracts were in dichloromethane which is incompatible for
ELISA. The methanol extracts were then subjected to ELISA screening according to the
revised procedures (Appendix B). Aliquots of the dust/soil samples from 22 NHEXAS homes
and 13 low-income homes were also prepared for ELISA screening according to the revised
procedures.
-------
Sample Preparation for Conventional PAH Analysis
The house dust samples from 13 low-income families were separated into coarse and fine
(< 150 ftm) fractions at Battelle and only the fine fractions were used for subsequent analysis.
The 150 nm cut-off point for fine dust was based on the ASTM procedure (10). The house
dust samples from 22 NHEXAS homes were separated into coarse and fine (< 62 /*m)
fractions by the University of Arizona staff and the fine fractions were sent to Battelle for
subsequent analyses. The 62 ^m cut-off point for fine dust from the NHEXAS study was
based on the sediments grade in the Arizona area indicating that sediments greater than 62 pm
are mostly sandy. The entryway dust, pathway soil, and foundation soil samples were not
separated into fine and coarse fractions prior to analysis. An aliquot (0.5 g) of each dust/soil
sample was spiked with known amounts of perdeuterated PAH and extracted with two 10-mL
aliquots of hexane in a sonication bath each for 20 minutes. The hexane extracts were
combined, filtered, and concentrated to 1 mL for PAH analysis (8).
GC/MS Analysis Method
The sample extracts were analyzed by GC/MS using 70-eV electron ionization (El). A
Finnigan TSQ-45 GC/MS/MS instrument, operated in the GC/MS mode, was used. Data
acquisition and processing are performed with an INCOS 2300 data system. The GC column
was a DB-5 fused silica capillary column or equivalent, and the column outlet is located in the
MS ion source. Helium is used as the GC carrier gas. Following injection, the GC column
was held at 70°C for 2 min and temperature-programmed to 290°C at 8°C/min. The MS is
operated in the selected ion monitoring (SIM) mode. Masses monitored are the molecular ions
of the 19 target PAH and their associated characteristic fragment ions. Identification of the
target compounds is based on their GC retention times of the 19 target PAH relative to those
of the internal standards phenanthrene-dIO, 9-phenylanthracene and benzo[e]pyrene-d12.
Quantification of target compounds was based on comparisons of the respective integrated ion
current responses of the target ions to those of. the corresponding internal standards using
average response factors of the target compounds generated from standard calibrations.
10
-------
Quantification of the total alkylated PAH isomers is based on the average response factors of
either the corresponding target alkylated PAH or their parent PAH.
Statistical Analysis
The following types of samples were collected in different field studies and analyzed by both
GC/MS and ELISA:
• House dust (HD), entryway soil (ES), and pathway soil (PS) samples taken from
the 13 low income homes in Raleigh/Durham, N.C. in both summer and spring
field studies.
• Floor dust (FDP) equivalent to house dust, foundation soil (FSP), and yard soil
(YSP) samples taken from 22 NHEXAS study (Arizona) homes.
• Air and 24-hr liquid and solid composite food sample extracts from the 13 low
income homes in Raleigh/Durham, N.C. in the summer field study.
Summary statistics (Sample Size, Mean, Standard Deviation, Minimum, and Maximum) for the
ELISA measurements of total PAH and C-PAH and GC/MS measurements of total PAH and B2-
PAH in dust, soil, food, and air samples were determined. The total PAH measurements from
GC/MS were the sums of the measured concentrations of all target parent PAH. The B2 PAH
measurements were the sums of the concentrations of target PAH which are B2 PAH (probable
human carcinogens). Three types of statistical analyses were performed on the data: paired t-
tests, regression analysis, and Fisher's exact test. First, paired t-tests were used to determine if
there are differences between the average PAH concentrations of the two analysis methods. Tests
were performed on data from dust and soil samples alone and on data from all samples. Both
total PAH and C-PAH measurements from ELISA and total PAH and B2 PAH measurements
from GC/MS were considered. All t-tests were performed on log-transformed data.
Regression analysis was used to examine the relationships between ELISA and GC/MS for
measuring PAH. The regression was performed both on raw data and on log-transformed data.
11
-------
The regression analyses were conducted on all combined data, and on data from each sample
medium.
To evaluate the screening performance of ELISA, we defined the performance measures of
sensitivity, specificity, positive predictive value, and negative predictive value according to the
results from both ELISA and GC/MS methods. The definitions of these performance measures
are shown in Table 4.1. Fisher's exact test was used to test whether ELISA and GC/MS
measurements are statistically independent (i.e., whether there is no predictive relationship
between these measurement methods).
Table 4.1 Definitions of Performance Characteristics
GC/MS Standard
Below
Above
Below
c
a
Above
d
b
ELISA Screening
Response
In the above table, the letter a represents the number of households which have an ELISA derived
PAH concentration above a given value (10 ppm for PAH ELISA or 2 ppm for C-PAH ELISA) and a
GC/MS derived concentration below a given value (1 ppm for total PAH or 0.5 ppm for B2 PAH).
Letters b, c, and d represent similar counts. From these counts the following performance
characteristics are calculated
Performance
Characteristic
Sensitivity
(or True Positive Rate)
Specificity
(or True Negative Rate)
Positive Predictive
Value (PPV)
Negative Predictive
Value (NPV)
Definition '" *
Probability of a household being above the PAH
standard for the sample matrix given that there is a
household with a high ELISA PAH response.
Probability of a household being below the PAH
standard for the sample matrix given that there is a
household with a low ELISA PAH response.
Probability of a household having a high ELISA PAH
response given that the observed PAH level in the
household is above the standard for the sample
matrix.
Probability of a household having a low ELISA PAH
response given that the observed PAH level in the
household is below the standard for the sample
matrix.
' Calculation
b/(a + b)
c/(c + d)
b/(b+d)
c/(a + c)
12
-------
Method Evaluation of 2,4-D and PCP ELISA
The 2,4-D and PCP ELISA test kits were purchased from Ohmicron Environmental
Diagnostics. The assay test kits included reagents for the assay plus extraction solvents and
extraction tubes. The assay test kits and extraction solvents were used as instructed; the assay
test kit extraction tubes were replaced with standard 15 mL or 50 mL centrifuge tubes, for 2
and 4 mL, or 20 mL extraction volumes, respectively.
Preliminary method evaluation tests were conducted using a moist humus soil, a dry clay soil,
and a house dust. The humus soil and clay soil represent two extremes of soils; the humus soil
has high humic acid (organic) and water content, the clay soil has high inorganic content and
little water. In general, we find equivalent extraction efficiency for 2,4-D and PCP from soil
and house dust using the standard acetonitrile: phosphate buffer extraction solvent mixture,
and thus it is not necessary to test all spike levels in all media (9). For this reason, evaluations
of recovery from spiked humus soil were most extensive and involved measurement of
recoveries using either a conventional GC/ECD analysis method or the ELISA method at 3
different spike levels. Extraction of analytes from clay soil was assessed at a single spike level
only using the ELISA assay. Recovery of analytes from house dust was assessed at a single
spike level with extracts analyzed by ELISA and GC/ECD.
The extraction conditions that were evaluated included: the solvent, the ratio of soil/dust to
solvent volume, and the extraction technique. The ELISA extraction solvent for PCP is NaOH
in 75% methanol/ 25% water and the ELISA extraction solvent for 2,4-D is 75% methanol/
25% water. Because of the similarity of analyte properties, extraction efficiency of each
analyte in each ELISA solvent was measured. The test kit procedures involved extracting 10 g
of soil with 20 mL of solvent. Alternative quantities of the soil mass and extract volume were
evaluated which are 1 g/2 mL, 1 g/4 mL and 1 g/20 mL. Two extraction techniques
investigated were shaking (test kits procedures) and sonication in a water bath. One analytical
13
-------
method was used to measure both 2,4-D and PCP in sample extracts by GC/ECD. A single
surrogate recovery standard, 3,4-D, was used to assess the recovery of both 2,4-D and PCP
through analytical procedures. The 3,4-D showed only minor cross-reactivity (5% relative to
2,4-D) in the 2,4-D ELISA assay, and so was used as an important diagnostic tool in analyses
where GC/ECD measurements were made.
Each dust sample (or soil sample) was weighed into a centrifuge tube, spiked with the
designated analytes, and sonified or shaken for 10 min. The tube was then centrifuged for 10
min to settle and compact the dust. If an ELISA assay was not performed, 1.8 mL of the
initial 2 mL solvent volume was removed for GC/ECD analysis. If an ELISA assay was
planned, triplicate aliquots of the extract were removed first for dilution into the respective
ELISA diluent; then, 1.5 mL of the initial 2 mL solvent volume was removed for the GC/ECD
analysis. For the 50 fold dilution of the 2,4-D assay, 100 /*L of extract was diluted into 5 mL
of diluent. For the 500 fold dilution of the PCP assay, 50 ^L of extract was diluted into
25 mL of diluent.
The aliquot removed for GC/ECD analysis was diluted with 20 mL of distilled/deionized
water, and the pH was adjusted to 1 with concentrated HC1. The acidified extract was applied
to a 500 mg CIS SPE cartridge that had been conditioned in sequence with 10 mL each of
methanol, distilled/deionized water, and 1:10 acetonitrile: 0.025M phosphoric acid. After
loading, the columns were air dried for 2 hours, and then eluted with two aliquots of 2 mL of
1:1 hexanerdiethyl ether. The eluate was concentrated to near dryness under a stream of dry
nitrogen; the internal standard (1 pg of 2,6-D) was added, and the volume was adjusted to
1 mL with 95:5 methyl-t-butyl ether: methanol.
The extracts and multi-point calibration standards were derivatized with ethereal diazomethane
generated in situ from carbitol and diazald in KOH. After derivatization, the extracts were
14
-------
allowed to sit at room temperature for 30 min, and then the excess diazomethane was removed
with a gentle stream of dry nitrogen.
Samples and standards were analyzed using a Hewlett-Packard 5890 GC/ECD with a 60 m
DB-5 column (0.25 mm id, 0.25 jtm film thickness). The GC temperature was programmed
as follows: 90-180 °C at 8 °C/min; 180-210 °C at 2°C/min; 210-300 °C at 20 °C/min; final
hold time of 15 min. The splitless injector was held at 250 °C. Standards were interspersed
among samples in the run order. The internal standard method of quantification was used with
linear regression analysis of concentration versus relative peak area.
15
-------
Chapter 5
Results and Discussion
Evaluation of ELISA for Screening PAH and C-PAH
The procedures (Appendix A) provided by Ohmicron for ELISA screening of PAH and
C-PAH in the soil samples were not suitable for the screening of the house dust samples. In
those procedures, 10 g of sample is required for conducting ELISA. However, less than 10 g
of house dust was collected from most households in both the NHEXAS study and the
Cooperative Agreement study. Therefore we revised the procedures by using 1 g of dust/soil
sample instead of 10 g of sample. In the original procedures, the extraction method used is
hand-shaking the dust/soil with methanol for 1 min. This shaking method provided by
Ohmicron may not effectively remove PAH from the dust/soil matrices. Extraction efficiency
tests were conducted on three aliquots (1 g, 3 g, and 10 g) of a house dust sample with the
1-min shaking method. The results are summarized in Table 5.1. The lowest ELISA
generated PAH concentrations were observed with the 10 g aliquot of the sample and the
highest PAH concentrations were observed with the 1 g aliquot of the same sample. This
result suggested that the 1-min shaking method cannot effectively remove all the PAH from the
house dust when 10 g or 3 g of the dust sample is used. We therefore evaluated and compared
two extraction methods, shaking and sonication for removing PAH from the dust and soil
matrices with 1 g of sample. The comparison of 1-min shaking and sonication results is
described in the following section. The amounts of diluent used for both PAH and C-PAH
assays were reduced to reduce the quantities of chemical wastes generated from the assays, but
the 50 fold dilution was maintained.
16
-------
Table 5.1. ELISA Results for the House Dust Sample
Concentration, ppm'
Sample Size
PAH Assay
C-PAH Assay
lg
3g
10 g
48
29
19
4.4
2.4
.1.5
• The reported concentrations were derived from the PAH and C-PAH ELISA responses.
It should be noted that the derived concentrations from PAH ELISA do not represent the true
sum of the concentrations of all PAH, and those from C-PAH assay are not the true sum of the
concentrations of all carcinogenic PAH either. This is mainly because the calibration
(inhibition) curves generated from PAH and C-PAH assays were based on phenanthrene, and
on benzo[a]pyrene, respectively. However, other PAH compounds also have cross activities
with both assays, and the ELISA derived concentrations cannot accurately reflect the cross
activities of PAH mixture in the samples.
Two extraction methods, sonication and shaking, were evaluated for the preparation of
dust/soil samples for PAH and C-PAH ELISA. For this set of experiments, only a 1 g aliquot
of the dust/soil sample was used for ELISA. The results are summarized in Table 5.2. For
the PAH assay, there is good agreement between the sonication method and the shaking
method. Using the paired t-tests with the null hypothesis that the two methods are equivalent
gives: t = 0.759, and p = 0.457. However, for the C-PAH assay the sonication method
results are in general slightly higher than the shaking method results. A paired t-test gives
t = 3.573, and p = 0.002. The mean difference between these two methods is 1.11 ppm.
This finding indicated that the sonication method is more effective in removing C-PAH from
17
-------
Table 5.2. Comparison of Sonication and Shaking Extraction Methods for ELISA
PAH Assay, ppm*
C-PAH Assay, ppm"
Stmple Codefc
A-HD-X
B-HD-X
C-HD-X
D-HD-X
E-HD-X
F-HD-X
G-HD-X
H-HD-X
1-HD-X
J-HD-X
K-HD-X
L-HD-X
M-HD-X
A-ES-X
C-ES-X
E-ES-X
H-ES-X
I-ES-X
J-ES-X
K-ES-X
L-ES-X
M-ES-X
C-PS-X
Sonication
84
76
61
36
33
133
29
49
13
28
56
46
52
29
15
NA°
28
7.7
16
6.2
7.1
6.4
NAb
Shaking
68
77
54
36
33
101
42
41
14
30
55
42
47
40
23
NAC
21
11
19
8.5
10
4.7
NAb
Sonication
6.8
5.3
5.2
3.0
3.6
8.1
3.1
6.9
1.9
3.6
5.7
7.9
3.9
6.2
2.0
6.2
3.4
2.1
NAC
0.72
1.2
1.2
1.4
Shaking
5.5
3.2
4.0
2.1
2.6
13
2.6
4.1
1.6
2.5
3.8
4.8
2.8
3.9
2.1
5.8
1.8
4.1
NAC
1.1
1.5
0.72
1.0
The reported concentrations were derived from the PAH and C-PAH ELISA responses.
A, B, etc. denote household code; HD denotes house dust; ES denotes entryway dust; PS denotes
pathway dust; and X denotes the field study conducted in the spring of 1996.
NA denotes that data are not available in the respective assay.
18
-------
the dust/soil sample matrices than the shaking method. This is probably due to the fact that
most C-PAH are 5- to 6-ring, and these PAH may not be completely removed from the
dust/soil by the shaking method. The revised procedures for ELISA are described in
Appendix B. These procedures consisted of reducing the sample size to 1 g, extracting the
dust/soil sample with the sonication method, and reducing the amount of diluent used for the
F '
ELISA.
The overall method precision for PAH and C-PAH ELISA screenings were determined.
Triplicate sets of dust/soil samples were processed for ELISA screening. The precision is
calculated by the percent relative standard deviation of each triplicate sample. The results are
summarized in Table 5.3. The overall method precision ranged from 11.9 to 28.5% for the
PAH assay and from 5.89 to 20.7% for the C-PAH assay. The precision for the C-PAH
ELISA is slightly better than that for the PAH ELISA. The precision for the assays alone, not
including the extraction step, was also determined by conducting ELISA in triplicate on
selected dust/soil sample extracts. The results are given in Table 5.4. As we expected, the
precision for the assay itself is better than the precision for the overall method because of the
exclusion of the extraction step. In summary, the overall method precision was within 30%
for the PAH assay and within 25% for the C-PAH assay.
Known amounts of phenanthrene were spiked into the soil samples and subsequently analyzed
by PAH ELISA, and known amounts of benzo[a]pyrene were spiked into different aliquots of
soil samples and analyzed by C-PAH ELISA. The recovery data are summarized in Table 5.5.
Quantitative recoveries of phenanthrene and benzo[a]pyrene were obtained from PAH ELISA
and C-PAH ELISA, respectively. It should be noted that when a mixture of 16 PAH was
spiked into the soil samples, greater than 400% recovery of phenanthrene was observed using
PAH ELISA. This is mainly because other PAH also contributed to ELISA responses. The
same observation was noted for the recovery of benzo[a]pyrene using C-PAH ELISA. The
recovery data of mixtures of PAH cannot be addressed, mainly due to the fact that the
19
-------
Table 5.3. Overall Method Precision of ELISA Screening for PAH and C-PAH
Precision,
Sample Codev
A-HD-S
A-HD-X
K-ES-S
G-PS-S
PAH Assay
11.9
24.7
23.0
28.5
C-PAH Assay
9.22
20.7
5.89
11.1
* The precision is expressed as percent relative standard deviation (RSD) of the triplicate sets of
dust/soil samples.
k A, B, etc. denote household code; HD denotes house dust; ES denotes entryway dust; PS denotes
pathway soil; and S and X denote the field study conducted in the summar of 1995 and in the spring of
1996, respectively.
RSD,% = Standard deviation X 100%
Mean
Table 5.4. Assay Precision for the PAH and C-PAH ELISA Screening
Precision, %"
Sample Code*
A-HD-X
A-ES-X
C-ES-X
A-PS-X
B-PS-X
C-PS-X
FSP-54-12945
PAH Assay
8.45
13.3
1.96
4.40
6.81
6.43
7.29
C-PAH Assay
3.94
7.69
17.3
5.24
7.64
10.9
7.12
The precision is expressed as percent relative standard deviation (RSD) of the triplicate sets of
dust/soil sample extracts.
A, B, etc. denote household code; HD denotes house dust; ES denotes entryway dust; PS denotes
pathway soil; X denotes the field sampling conducted in the spring of 1996; and FSP denotes
foundation soil sample.
RSD,% = standard deviation X 100%
Mean
20
-------
Table 5,5. Recoveries of Phenanthrene and Benzo[a]pyrene from Dust/SoU Samples
Sample Codea
PAH ELISA
J-PS-S
J-PS-S
B-ES-X
B-ES-X
B-ES-X
C-PAH ELISA
J-PS-S
J-PS-S
B-ES-S
B-ES-S
Spike Level, ppmb
5
5
2.5
2.5
0.5
(U
0.1
0.2
0.2
Recovery, %
150
130
107
115
68
130
120
110
120
A, B, etc. denote household code; ES denotes entryway dust; PS denotes pathway soil; and
S and X denote the field study conducted in the summer of 1995 and in the spring of 1996,
respectively.
Phenanthrene was spiked into each sample for PAH ELISA and benzo[a]pyrene was spiked
onto each sample for C-PAH ELISA.
21
-------
calibration of PAH ELISA was based on phenanthrene, and the calibration of C-PAH ELISA
was based on benzo[a]pyrene. Approximately 50% recovery of phenanthrene was observed
when known amounts of phenanthrene-d,0 were spiked onto the dust sample using PAH assay.
This finding suggested that the ELISA responses of phenanthrene and phenanthrene-d,0 are
different. The overall method accuracy for both PAH and C-PAH assays cannot be addressed
because neither assay can accurately determine either total PAH or carcinogenic PAH in the
sample matrices containing a mixture of PAH compounds. However, PAH and C-PAH assays
can still be used as screening tools for estimating PAH levels, but can not be used as a
quantitative method.
The PAH ELISA quantification limit is set at 2 ppb equivalent assay concentration by the
vendor. The sample extracts were spiked at 1, 0.1, and 0.05 ppb equivalent assay
concentrations of phenanthrene and assayed. As expected, the results showed no increase of
PAH ELISA response at these spiking levels. The C-PAH ELISA quantification limit is set at
0.2 ppb equivalent assay concentration. The sample extracts were spiked at 0.1, 0.05, 0.01,
and 0.005 ppb equivalent assay concentrations of benzo[a]pyrene and assayed. The results
showed no increase of C-PAH ELISA responses for all but 0.1 ppb level. At 0.1 ppb spike
level, the recovery of benzo[a]pyrene was 140%. This rinding suggested that C-PAH ELISA
quantification limits might be lower than the specified value (0.2 ppb).
The sample extracts of air and 24-hr liquid and solid composite food samples generated from
the Cooperative Agreement studies were analyzed by PAH ELISA and C-PAH ELISA. The
dust/soil samples from 13 low-income families and 22 NHEXAS homes were prepared and
analyzed by both assays. The ELISA derived concentrations were compared with the
concentrations from conventional GC/MS analysis. These comparisons are discussed in a later
section.
22
-------
Alkyl PAH in Multi-Media Samples
The sample extracts of air, dust/soil, and food generated from 13 low-income families
were analyzed by GC/MS for alkyl PAH (6,7). These samples were collected and prepared in
the summer of 1995. The concentrations of alkyl PAH found in the indoor and outdoor air
samples are summarized in Table 5.6. The alkyl PAH concentrations in indoor and outdoor
air of each household are given in Appendix C. Note that households A through G are located
in the inner city and households H through M are located in rural areas. These data provided
values expressed in ng/m3 for the alkyl PAH isomers and the sum of all alkyl PAH. The data
reported in Table 5.6, in Appendix C, and in the following sections were corrected for the
background levels in the field blank. The most abundant alkyl PAH found in air were methyl-
and C2-alkyl-naphthalene isomers. The parent compound naphthalene was also the most
abundant parent PAH in these air samples (6,7). The levels of 2- to 3- ring alkyl PAH found
in indoor air were higher than those in the corresponding outdoor air. Similar concentrations
of most 4- to 6-ring alkyl PAH were observed in both indoor and outdoor air within each
nonsmoker's household. Higher concentrations of these 4- to 6-ring alkyl PAH were found in
the indoor air as compared to the outdoor air within each smoker's household (households F,
G, K and M). In general, higher levels of alkyl PAH were observed in inner city outdoor air
as compared to the rural area outdoor air. The sum of the concentrations of alkyl PAH ranged
from 369 to 3270 ng/m3 in indoor air and from 49.9 to 702 ng/m3 in outdoor air.
The alkyl PAH concentrations measured in the house dust, entry way dust, and pathway soil
samples are summarized in Table 5,7. The alkyl PAH concentrations of each dust/soil sample
are presented in Appendix D. The alkyl PAH concentrations corrected for the background
levels in the field blank are expressed in units of ppm ftig/g). In general, the most abundant
alkyl PAH were alkyl 3-ring PAH isomers. The sum of the concentrations of the alkyl PAH
ranged from 0.584 to 3.32 ppm in house dust, from 0.218 to 1.54 ppm in entryway dust, and
from 0.092 to 1.98 ppm in pathway soil. With few exceptions, the relative concentration
23
-------
o
tu
tu
4
! i
^? c
1 ^
i
•1
s
(U
CxO
S
1
•S |
o "5
1 S
e
3
.§
1
1
1
t-«
VD
V")
CO
•o
CO
oo
a
£
r— 1
V)
F?
'hlaphthalene
I
cs
VO
o
^^"
cQ
oo
cs
f-.
cH
CO
oo
t-
V-H
VO
M
1
CO
CS
OS
CO
Os
s
co
cs
f~
T— 1
0
oo
52
o3
[naphthalene isom
1
CO
OS
VO
es
00
o
co
CO
o
*
co
<— (
cs
«— 1
CO
U
£
O
CO
• 1-4
i
T— 1
U
o
,_ i
r- 4
«n
cs
VO
co
cs
CO
cs
1— 1
cs
<3"
^
CO
S3
Iphenanthrene isoi
1
cs
T— 1
s
o
s
CO
oo
o
t»»
o
co
1-4
CO
Os
"*
Ipyrene isomers
I
u
S
0
i-H
a
0
oo
0
VO
CO
t-H
CO
0
^u
.
^f
Ipyrene isomers
7
cs
O
o\
o
1
o
s
*— (
§
o
1
o
co
1— 1
cs*
Ichrysene isomers
U
S
1-H
VO
o
C3
CO
co
oo
co
^fr
co
vq
oo
g
i
•*-(
I
1
•a
oo
^4
oo
OS
00
o
JP
rt
00
to
F2
o
o
co
cs
52
u
E
o
CO
I
s
s
g
7
r-H
U
VO
T— 1
cs
Os
OS
Tf
g
Q
s
CO
Q
cs
co
i
•s
CO
24
-------
JB a
A
^
>>
§
1
1
Q
I?
o
ffi
a
•p*
1
I
I
o
u
I
!
•s
e?
1/2
«-:
in
0)
1
s>
*
000
vo go g g g rg c3
O O O O O O O
o
CO
o'ooooooo'oo'o
»• •* -
O«Tjrvpr>JOO\ —« •» oo
G7 ^3 ' P^ ^^ "*1" ^** Cl? *^ **^ ^^
dooo'ooc>ooo~
Q
>.
I
8
ooooooooooo
O O »"* OO O ~
r
ON
8
ooooooooooo
O O CO CO OO CS —J f> «S S
O O O • - O - «-< O C> O O C>
12 S2
§ 8
SS 52 S -2
.S .52 °> S£ u u **
1 1 « II I I' » 8 S 5
•S ^S i5 H 5 SiSi &§•?,
S H S3 *^5 "^ ^* ^* ^B ^3 3J i«
iftififfii-
u
25
-------
trend for alkyl PAH is house dust > entryway dust > pathway soil. This relative
concentration trend was also observed for the parent PAH in these samples. The alkyl PAH
concentrations measured in food samples are summarized in Table 5.8. Alkyl PAH
concentrations are expressed in units of ppb (ng/g). The concentrations of alkyl PAH in each
adult's and child's food sample are presented in Appendix E. The reported concen-trations of
each food sample were corrected for the background levels in the field blank. The most
abundant alkyl PAH found in the food samples are 2- to 3-ring alkyl PAH. Concentrations of
alkyl PAH found in the adult's food samples were within the same order of magnitude as those
in the child's food samples. The sum of alkyl PAH concentrations ranged from 0.866 to 13.9
ppb in adult food samples and from 2.10 to 15.6 ppb in child food samples.
Phthalates in Multi-Media Samples
The concentrations of target phthalates found in indoor and outdoor air samples are
summarized in Table 5.9. The phthalate concentrations in individual air samples are given in
Appendix F. The reported values were corrected for the background levels found in the field
blank. In general, levels of phthalates found in the indoor air were higher than the
corresponding outdoor air. Indoor phthalate concentrations ranged from 3.07
(di-n-octylphthalate) to 3490 (bis(2-ethylhexyl)phthalate) ng/m3. The concentrations in
outdoor air ranged from 0.475 (di-n-octylphthalate) to 594 (butylbenzylphthalate) ng/m3.
Table 5.10 summarizes the background-corrected levels of phthalates found in house dust,
entryway dust, and pathway soil samples. The phthalate concentrations of individual dust/soil
samples are given in Appendix G. The general relative concentrations trend of phthalates was
similar to those of PAH and alkyl PAH: house dust > entryway dust > pathway soil. The
most abundant phthalates were either butylbenzylphthalate or bis(2-ethylhexyl)phthalate and
the least abundant phthalate was dimethylphthalate. Note that we only measured the six target
phthalate compounds because of the availability of the standards. There were other phthalates
present at significant levels in the dust/soil samples which are not reported in these tables.
26
-------
I
I
8
oo
o ~
8 8
—° o
o o
V V
o
o
CO
o
o
o\
s
CO VO
•s 1
o
V
CO •-<
ts
V
&
o — —
00 «-H
CO «D ^J- OO
—' -I o o
o
V
1
1
000
OO VO
I
o
o
V
&
3
VO
*- •* CO «S
R S
o o
V
a
ene
ph
5^1
« *« *« *» ^ *a
5 r T T T 7
^ CN »^* CS *•"* fS
-i O O U U O
O
o
V
§^p
*o
oo
o' o'
V
•o o\
O CO
V ~
CL.
7
f— (
0
27
-------
1
i
_
E
9
.£=
a
(U
E
9
!s
S
E
9
C
««-4
D
1
Compound
00
ON
vd
8
«n
vd
»n
VO
*
°>
ol
t"-»
CO
y-H
Dimethylphthalate
CO
oo
VO
VO
oo
V—(
Tl-
CS
CJ
vo
U*l
O4
Diethylphthalate
oo
"*
CO
r-t
OO
m
co
04
I/")
^J*
IT)
Di-n-butylphthalate
ON
a
oo
CO
04
Tt
SQ
co
t>-
04
ON
O
t-~
ON
o
Butylbenzylphthalat
ON
oo
ON
«n
•<*
a
vo
00
Q
ON
TJ*
CO
•a
fl
Bis(2-ethylhexyl)ph
^
^4-
O4
5
o
8
VO
•
r-»
C3
CO
VO
d
co
Di-n-octylphthalate
I***
^^
O4
3
O
g
8
04
8
'—i
o
CO
co
1
•a
M
*
"4-1
O
00
28
•g
-------
JS
CO
CO
r
I
•8
es
§
I
. o
1
§
U
I
g
•s
8?
CO
o
I
•S
CO
I
d d d d d d d
d d d d d d d
g VO OO ~4
O O C
® — o oo oo
o o o
-------
Levels of target phthalates ranged from 0.006 to 131 ppm in house dust, from 0.001 to 80.7
ppm in entryway dust, and from < 0.001 to 2.59 ppm in pathway soil.
The concentrations of phthalates found in the food samples are summarized in Table 5.11.
The concentrations of the individual food samples are presented in Appendix H. The reported
concentrations were corrected for the background levels found in the method blank. Note that
all food containers were made of plastic materials that can contribute to the amounts of
phthalates found in the food samples. The concentrations of phthalates ranged from < 0.02 to
110 ppb in adult food samples and from < 0.02 to 84.3 ppb in child food samples. In
general, levels of phthalates found in the adult food samples were higher than those in the
child food samples. There were also other nontarget phthalates present at significant amounts
in these food samples.
PAH in Dust/Soil Samples
House dust, entryway dust, and pathway dust samples were collected from 13 low-income
families during the spring of 1996 under a Cooperative Agreement study (6). Aliquots of the
samples were extracted with hexane and analyzed by GC/MS for target PAH. The PAH
results are summarized in Table 5.12. The PAH concentrations in individual samples are
presented in Appendix I. All the reported values were corrected for the background levels
found in the field blank. The sum of the concentrations of the B2 PAH ranged from 0.267 to
7.02 ppm in house dust, from 0.036 to 0.486 in entryway dust and from 0.009 to 0.701 ppm
in pathway soil. With few exceptions, the sum of the concentrations of B2 PAH accounted for
approximately half of the total target PAH concentrations. The concentration trend for most
PAH is house dust > entryway dust > pathway soil. The finding was also observed in the
dust/soil samples collected at the same households during the winter and summer seasons.
House dust, foundation soil and yard soil samples collected from 22 NHEXAS homes (5) were
extracted by hexane and analyzed by GC/MS for target PAH. The GC/MS results are
summarized in Table 5.13. The PAH concentrations in individual samples are presented in
30
-------
Table 5.11. Summary of Phthalates Concentrations (ppb) in Food Samples
Adult Subjects
Child Subjects
Compound
Dimethylphthalate
Diethylphthalate
Di-n-butylphthalate
Butylbenzylphthalate
Bis(2-ethylhexyl)phthalate
Di-n-octylphthalate
Sum of phthalates
Maximum
1.92
2.40
9.81
83.1
110
54.1
245
Minimum
0.012
0.013
0.285
1,31
2.37
<0.02
11.5
Average
0.447
1.18
3.75
24.3
49.8
10.2
89.6
Maximum
0.414
19.0
4.66
35.9
84.3
15.6
114
Minimum
0.028
<0.02
<0.02
<0.02
<0.02
0.011
0.09
Average
0.128
1.97
1.11
6.48
10.9
1.54
22.1
Appendix J. The reported values were corrected for the background levels found in the
laboratory method blank, since no field blank was available for this study. The sum of the
concentrations of the B2 PAH ranged from 0.263 to 4.30 ppm in house dust, from 0.011 to
2.92 ppm in foundation soil and from 0.007 to 1.82 ppm in yard soil. In general, the
concentrations of PAH in house dust samples were higher than those in the foundation soil and
yard soil samples. Similar PAH concentrations were found in the foundation soil and yard soil
samples. The sum of the concentrations of target PAH was greater than 1 ppm in 16 out of 22
house dust samples, but only in 2 foundation soil and 2 yard soil samples. The two households
having greater than 1 ppm PAH levels in foundation soil also had greater than 1 ppm PAH
levels in yard soil. The sum of the concentrations of B2 PAH accounted for approximately
half of the total target PAH concentrations for most dust/soil samples. This finding was also
observed in the dust/soil samples collected from the 13 low-income families.
31
-------
ra
£
t?
fe
w
t
.a
u
ew
•s
£?
t«
53
I
*-»v>\p'3'*p^'tnooo'io\Q>\Qtnfnoot-t1*-
*•* O C? O O *o ^H o oo c$ O 3" oo co r*4 c^i o
**^> ^3 ^y Cy ^y ^5 ^^ »••* ^^ ^^ ^5 O* ^2 ^y C5 ^5 C?
ooooooooooooooooo
»~l CA OJ
o in o)
000
s 1
8888888
i s l
S S 8 S S Q
ooooooooooooooooooooo
V V V
*^ CO ^4 If} ^0 OO ^d
*"* CO *~^ ^^ Vl Vl t'l
o o o o •-• o m
8O en oo in
en ^ in ON
—< O —< oi o
eJoocJoescScJoescJocsoeJcseJeScJ
o
t—
CO
O
***
00 VO
8 S
s
ooooooooooooooooooooo
8
8 S I I
C\ VO
o o
ooooooooooooooooooe>oc>
3"ino4o*-«encNt^'—it^oo4o\ooovoo4c^-oo
Q% •—j en oo o\ O4 *-* en in vp eM o ^* t^ ^3* to
s
s s a o
o\ t~ — oo ON
en ^H
OOOOO*-«OO4
I 2
c>. —<
04 0
s
o
oq
O
r- en
~* oo
OO O4
O O O
VO
in
^ 1
* *? I
g i? I
^T ^^ .^.
U
2 2
"z n
u* u
U
CO
32
-------
1
o
en
•s
•g
1
I
"S
i
tf
I
*g
CO
te.
I
1
§
U
&<
'S
t?.
cn
•o
•&
I
oopop
-
oooooooo.ooooooo
ddddddddddddddddddddd
V V V V
.
ooodooddddo'ddddddddrn-;
OOOOO
r^QSw'Sf^^O'J r-i -H -H o -H o OO
ooooooooooooooooooo— 5
^5 ^* ^5 ^5 ^5 Gy ^5 c* w^ C5 *^ ^? c^ ^5 c> C3 25 ^s ^^ ^& cs
ddddddddddddddddddddd
oooooooooooo-^oooooot-^^
S
111
•5 i
*l*
N- -S •«
r-1. A M
n N ^-^
a o
.« •&
2 05
I
•B
•
33
-------
Relationship Among B2 PAH, Total Target PAH, and SVOC
As we discussed above, the sums of B2 PAH concentrations accounted for approximately half
of the levels of total target PAH in most dust/soil samples. This relationship was further
examined in other sample media. Table 5.14 presents Pearson correlation coefficients (r)
obtained by correlating the sums of B2 PAH levels with total target PAH levels in the samples
from each sample medium. The p values shown in the parentheses indicate the statistically
significant level for the null hypothesis, i.e., that there are zero correlations between B2 PAH
and total target PAH in each sample medium. All sample media but indoor air samples tended
to give good correlations between B2 PAH and total PAH. The levels of B2 PAH correlated
well (r > 0.90) with total target PAH in dust/soil samples. Similar but weaker relationships
were observed in outdoor air (r = 0.860) and food (r = 0.670) samples. The poor
correlations between B2 PAH and total target PAH in indoor air samples could be due to the
high levels of 2- to 3-ring PAH found in indoor air that account for a majority of total target
PAH. Strong relationships between B2 PAH and total target PAH were observed in the
dust/soil samples collected from 13 low-income homes and 22 NHEXAS homes. These
correlations were also observed in the combined data set from 13 low-income homes and
22 NHEXAS homes. Figures 5.1 and 5.2 display the relationships between B2 PAH and total
target PAH in the dust/soil samples. Data shown in Figure 5.1 were from 78 dust/soil samples
collected from 13 low-income homes during the summer and the spring seasons. Data
displayed in Figure 5.2 were from 63 dust/soil samples of 22 NHEXAS homes. The data
from the NHEXAS samples showed a slightly higher correlation (r = 0.993) than that
(r » 0.976) from the samples from low-income homes. A similar linear relationship was also
observed from the combined data set. The dust and soil samples from 13 low-income homes
in the summer field study were also analyzed for alkyl PAH and phthalates. The relationship
(r=0.573) for alkyl PAH and total target PAH in these dust/soil samples was not as strong as
that for the B2 PAH and total target PAH. Similar results (r=0.589) were observed between
total target PAH and phthalates in these samples.
34
-------
Table 5.14. Correlation Coefficients (r) Between B2 PAH and Total Target PAH in
Each Sample Medium
Sample Medium*
Correlation Coefficient, i*
Indoor Air (low-income homes, N=13)
Outdoor Air (low-income homes, N=13)
Food (low-income homes, N=26)
House Dust (low-income homes, N*=26)
Entryway Dust (low-income homes, N=26)
Pathway Soil (low-income homes, N=26)
House Dust (NHEXAS homes, N=22)
Yard Soil (NHEXAS homes, N=21)
Foundation Soil (NHEXAS homes, N=20)
Dust/Soil (low-income homes, N=78)
Dust/Soil (NHEXAS homes, N=63)
Dust/Soil (combined data, N=141)
-0.146(0.6334)
0.860 (0.0002)
0.670 (0.0002)
0.978 (0.0001)
0.962 (0.0001)
0,991 (0.0001)
0.994 (0.0001)
0.999 (0.0001)
0.997 (0.0001)
0.976 (0.0001)
0.993 (0.0001)
0.973(0.0001)
Data of indoor air, outdoor air, and food are from the summer field study of 13
low-income homes; data of the house dust, entryway dust, and pathway soil are from the
summer and the spring field study of 13 low-income homes; and data from house dust,
yard soil, and foundation soil are from NEXAS study.
The corresponding P value is shown in parentheses.
35
-------
p
CD
O
•<£
q
CN
q
o
I
s,
q
CO
q
CN
!
a
o
1
.3
I
o
q
o
o
(O
q
to
q
o
radd'HVd-za jo umg
36
-------
s
o
l>-°
O
«>
O
iri
3
S
in
•*
mdd
37
-------
Relationships of PAH Among Different Sample Media
Multimedia samples were collected from low-income families for the Cooperative Agreement
Study. It is of interest to know whether the levels of PAH in dust and soil are related to their
levels in other sample media. The correlation between the measured total target PAH
concentrations in different sample media was investigated. Table 5.15 presents Pearson
correlation coefficients (r) for the PAH levels in one sample medium (e.g., house dust) with
the PAH levels in another sample medium (e.g., entryway dust). As shown in Table 5.15,
levels of total target PAH did not appear to be highly correlated in any of the different sample
media. Among all sample media, the strongest relationship was observed between house dust
and outdoor air samples. In general, there were positive but weak relationships for total target
PAH found among dust, soil, and air samples. Similar results were also obtained for B2
PAH.
Since the food samples were the 24-hr composite solid and liquid food consumed by the
subjects, as we expected there are no strong direct relationships between the food samples and
other types of samples. Similar results were obtained for the sums of B2 PAH in different
sample media.
The correlation between the measured PAH concentrations in house dust/yard soil/foundation
soil from 22 NHEXAS homes was also investigated. Table 5.16 summarizes the correlation
coefficients for total PAH and for B2 PAH among floor dust (house dust), yard soil, and
foundation soil. The correlations between PAH and B2 PAH levels in house dust/yard soil
and foundation soil/yard soil were higher than those obtained from the house dust/foundation
soil. Similar positive relationships of PAH levels found in house dust/entryway dust and
house dust/pathway soil were also observed from a previous 8-home study conducted at
Columbus, Ohio (8).
38
-------
2
v
"E
t/3
•O
6
•3
I
s o
9 I
OH rt
| 1
I u
I
fe
u
1
U
§
o
U
«?
I
*
(2
.2
1
voP
-H in
i en ts »•* en oo
IS *"" Q ^° *°
oc-f
«n vo
d d.
ene*: «nes- o en
t-- •* en o\ m vo
*-* •* cs e>j es CM
O S^ O O, O O.
•* en
en •*
rf o
0
en c^ en So
f^ ^f CM '—i
CO
O ON en
en O\ o
CN CM O
d d ~
1
3
IX,
i
u
39
-------
Table 5.16.
Correlation Coefficients (r) for Total Target PAH and for B2-PAH
in House Dust, Foundation Soil, and Yard Soil
Correlation Coefficient,
Correlation Calculated
Between Sample Media*
Total Target PAH
B2-PAH
FDP FSP
FDP YSP
FSP YSP
0.326(0.1610)
0.725 (0.0003)
0.680 (0.0014)
0.394 (0.0860)
0.776 (0.0001)
0.710 (0.0007)
" FDP denotes floor dust samples equivalent to house dust samples; FSP denotes foundation
soil samples; and YSP denotes yard soil samples.
b The corresponding P value is shown in parentheses.
In summary, there were positive but weak relationships observed for PAH found in house
dust/indoor air, house dust/outdoor air, house dust/entryway dust and house dust/pathway soil
from the 13 low-income families. A positive and relatively strong relationship was observed
for PAH found in house dust and yard soil from the 22 NHEXAS homes. Thus, PAH levels
in house dust may be used as qualitative indicators for PAH levels found in soil or air but not
food.
Comparison of PAH Data from ELISA and GC/MS
Statistical analysis was conducted on the PAH data of multimedia samples analyzed by both
GC/MS and FJLISA methods. Table 5.17 summarizes the number of samples analyzed for
PAH by study, sample type, and analysis method. Additionally, alkylated PAH GC/MS
measurements were made on dust, soil, food, and air samples from 13 low income homes in
the summer field study. A listing of all data used for the statistical analysis is given in
Appendix K. ELISA measurements that are not in the linear range of the calibration curves
are listed along with an asterisk (*) in Appendix K. The PAH GC/MS responses are the sums
of concentrations of all target parent PAH and the B2 PAH GC/MS responses are the sums of
40
-------
Table 5.17 Number of Samples Analyzed by ELISA and GC/MS Methods
PAH Analysis
Study
Summer
Field
Study
(North
Carolina)
Spring
Field
Study
(North
Carolina)
NHEXAS
Study
(Arizona)
Sample
Type*
HD
ES
PS
Food
Air
HD
ES
PS
FDP
FSP
YSP
ELISA
9
9
13
18
18
13
13
13
22
20
21
GC/MS
13
13
13
26
26
13
13
13
22
20
21
Both
9
9
13
18
18
13
13
13
22
20
21
ELISA
9
9
13
18
18
13
13
13
22
20
21
GC/MS
13
13
13
26
26
13
13
13
22
20
21
Both
9
9
• 13
18
18
13
13
13
22
20
21
a: HD = house dust; ES = entryway dust; PS = pathway soil; FDP = floor dust;
FSP = foundation soil; YSP = yard soil.
41
-------
concentrations of target B2 PAH. Note that nine house dust samples (FDP) and one yard soil
sample (YSP) of the NHEXAS study have very high (> 148 ppm) total PAH concentrations
according to the ELISA method. The upper and lower portions of Figure 5.3 display the
scatter plots of ELISA total PAH versus GC/MS total PAH in raw units and log-transformed
units, respectively. Note that all log-transformed data discussed in this report referred to
natural log-transformed data. Similarity, Figure 5.4 displays the scatter plots of ELISA C-
PAH versus GC/MS B2-PAH. As shown in the raw data plots, most of the data are
concentrated at lower PAH levels. The skewness of the data suggested performing the
Statistical analyses in a log scale.
To further describe the distribution of data, two-way frequency tables of ELISA total PAH
versus GC/MS total PAH and ELISA C-PAH versus GC/MS B2-PAH are presented for each
sample type in Appendix L. Cut-off points were chosen to best describe the spread of the
data.
Summary statistics (Sample Size, Mean, Standard Deviation, Minimum, and Maximum) for
the ELISA measurements of total PAH and C-PAH and GC/MC measurements of total PAH
and B2-PAH in dust, soil, food and air samples are displayed in Appendix M. The dust
samples include house dust (HD and FDP) and entryway dust (ES). Soil samples consist of
pathway soil (PS), foundation soil (FSP), and yard soil (YSP). Note the big differences in the
ranges of PAH concentrations that were obtained by the two analysis methods. For example,
ELISA total PAH concentrations in FDP samples range from 5.10 to 725 ppm, whereas
GC/MS total PAH concentrations range from 0.65 to 7.69 ppm. Similar differences exist in
other types of samples.
Paired t-tests
In order to achieve the normality of data, the natural log-transformation was used in
performing paired t-tests on dust/soil samples. Results of paired t-tests for the differences
42
-------
ELISA Total PAH vs. GC/MS Total PAH
750-
600-
450-
300
150
Sample Type ODD DusKHD+ES+FDP)
S S S Soil(PS+FSP+YSP)
s
D
6 8 10
GC/MS Total PAH. ppm
12
16
a.
I
I
7-
6-
5-
-2
EUSA Total PAH vs. GC/MS Total PAH, Log Scale
Sample Typn ODD Ousl(HOH-ES+FDP)
CSS Soil(PS-l-FSP-l-YSP)
&
I O a
s , s
ss
& '-
-3
2 -1O
GC/MS Tolol PAH. log(ppm)
Figure 5.3. Scatter plots of ELISA total PAH versus GC/MS total target PAH.
43
-------
ELISA C-PAH vs. GC/MS B2-PAH
SO
20
10
D
D 8
ODD
Sample Typo ODD —
888 —
— Dusl(HD+ES+FDP)
— Soll(PS+FSP+YSP)
D D
GC/MS B2-PAH. ppm
ELISA C-PAH vs. GC/MS B2-PAH, Log Scale
1 •
Sample Type D D D Dusl(HD-t-ES+FDP)
8 8 S Soll(PS-4-FSP-t-YSP)
*„ °DDD
D
D i
ODD
'SB D
D D !& t>
D D
s s D :
3
8
S
S
Bg
D 8
S
-5
-3 -2 -1
GC/MS B2-PAH. log(ppm)
Figure 5.4. Scatter plots of ELISA C-PAH versus GC/MS B2 PAH.
44
-------
between log PAH ELISA responses and log PAH GC/MS responses for dust and soil samples
are displayed in Table 5.18. For example, there were 133 samples analyzed by both methods
for Total PAH for the combination of dust and soil samples. The ratio of ELISA total PAH's
geometric mean versus GC/MS total PAH's geometric means was 18.2. That is, the
geometric mean of ELISA total PAH measurements is, on average, 18.2 times higher than the
geometric mean of GC/MS total PAH measurements across all dust and soil samples. The
geometric mean of data that follow a lognormal distribution is equal to the population's
median. As shown in Table 5.18, all test results are significant (i.e., the differences in
average PAH between ELISA and GC/MS methods are statistically significant when analyzing
dust and soil samples). Generally, PAH ELISA responses are higher than PAH GC/MS
responses. The ratio of geometric means between ELISA total PAH and GC/MS total PAH
for dust samples and for soil samples are 20.7 and 15.8, respectively. Also the ratio of
geometric means between ELISA C-PAH and GC/MS B2-PAH for the combination of dust
and soil samples, dust samples, and soil samples are 5.9, 5.8, and 6.1, respectively.
Table 5.18 Results of Paired t-test for the Difference Between Log (PAH ELISA
Response) and Log (PAH GC/MS Response) for Dust and Soil Samples
ELISA Total PAH
vs
GC/MS Total PAH
ELISA C-PAH
vs
GC/MS B2-PAH
Statistics
N
Ratio of
Geometric
Means
Geometric
Std. Error
p-value
Comb*
133
18.2
4.0
0.0001
Dust
66
20.7
3.7
0.0001
Soil
67
15.8
4.3
0.0001
Comb*
133
5.9
3.3
0.0001
Dust
66
5.8
3.1
0.0001
Soil
67
6.1
3.5
0.0001
Combination of Dust (HD+ES+FDP) and Soil (PS+FSP+YSP) Samples.
45
-------
Regression Analyses
Initially, the regression analyses were performed on all available ELISA and GC/MS data.
The results are summarized in Appendix N. There were weak relationships between ELISA
and GC/MS data for various sets of samples. Note that the air and food sample extracts were
the remainder of the extracts from the Cooperative Agreement study and had been spiked with
perdeuterated PAH. These perdeuterated PAH had cross activities with ELISA assays and
could contribute to the poor relationship between the ELISA and GC/MS data. Thus, further
analyses were focused on dust and soil samples. Table 5.19 summarizes the results of the
regression analyses on all the dust and soil samples from North Carolina and NHEXAS study
homes. The analyses were performed using both raw data and log-transformed data. The
summary includes the square of the correlation coefficient (R2), the intercept (a), and slope (P)
of the regression equation, and the p-value for the test that the slope is significantly different
from zero. As shown in Table 5.19, most of the p-values are less than 0.05 and none of the
RJ values exceeds 70%. The low correlations mean that the linear relationship between ELISA
and GC/MS measurements is not strong in some cases. Because ELISA derived concentrations
were based on the inhibition curve of single PAH (phenanthrene for PAH assay, and
benzo[a]pyrene for C-PAH assay) not a mixture of PAH, the sample matrix may have
significant effects on the ELISA results. To further examine the sample matrix effects from
different types of samples, we separated North Carolina study homes from NHEXAS study
homes and performed regression analyses on each of the HD, ES, PS, FDP, FSP, and YSP
samples. Table 5.20 shows the linear regression analysis results for the separated analyses on
each sample. The linear regression model of ELISA total PAH vs. GC/MS total PAH for the
FSP samples has an R2 of 89%. This indicates that there is a significant linear relationship
between the ELISA and GC/MS measurements of total PAH levels in the FSP samples. When
analyzing total PAH concentrations in FSP samples from the NHEXAS study homes, 89% of
the variation in GC/MS measurements can be explained by the variation in ELISA
measurements. Using the regression equation of ELISA total PAH=cc + P * GC/MS total
46
-------
Table 5.19. Regression Analysis Results for the Dust and Soil Samples, Combination of
North Carolina and NHEXAS Study Homes
Raw Data
ELISA vs GC/MS
1) TOTPAH_E*TOTPAH_G
2) CARPAH E*B2PAH G
^— ^—
Comb*
N=133
R2=4%
B— O* AO
~~Jl«*rp
P= 7.81
p=0.03f
N=133
R2 =39%
a=1.49
p=3.49
p<0.01
Dust
N=66
&Jz\%.
CC— 11. M
P=1.02
p=0.86
N=66
R2=26%
a=3.05
p=2.60
p<0.01
Soil
N=67
R2 = 13%
a=6.81
P= 10.20
p<0.01
N=67
R2=68%
Of =0.03
P=6.01
p<0.01
Log-transformed Data
Comb
N = 133
R2 =38%
a =2.79
P=0.75
p<0.01
N=133
R2=50%
a=1.36
P=0.71
p<0.01
Dust
N=66
R2=14%
a=3.24
P=0.54
p<0.01
N=66
R2=23%
a=1.49
P=0.53
p<0.01
Soil
N=67
R2=17%
a=2.09
P=0.47
p<0.01
N=67
R2=40%
a =0.93
P=0.62
p<0.01
* Combination of dust (HD+ES+FDP) and soil (PS+FSP+YSP) samples.
# p-value: the linear regression model is statistically significant at 0.05 level if the p-value is
less than 0.05.
1) Regression equation: ELISA Total PAH = o + p * GC/MS Total PAH.
2) Regression equation: ELISA C-PAH = a + p * GC/MS B2-PAH.
47
-------
Table 5.20 Regression Analysis Results for the Dust and Soil Samples, Separation of
North Carolina and NHEXAS Study Homes
Sample Type
ELISA Total PAH
vs
GC/MS Total PAH(1)
ELISA C-PAH
vs
GC/MS B2-PAH C)
HD
ES
PS
FDP
FSP
YSP
N=22
R2=46%
ec=34.08(p<0.01)*
P=5.35 (p<0.01)
N=22
R2=2%
a= 10.57 (p=0.01)
P=0.93 (p=0.56)
N=26
R2=28%
a=1.64(p=0.07)
P=2.20 (p=0.01)
N=22
R2=l%
a= 168.23 (p=0.01)
P=-7.46 OJ=0.74)
N=20
R2=89%
a=7.21 (p=0.01)
p= 18.91 (p< 0.01)
N=21
R2 = l%
-------
PAH, if the foundation soil's total PAH GC/MS response exceeds 1 ppm, then the predicted
total PAH ELISA response would be greater than 26.2 ppm. Similarity, ELISA and GC/MS
have significant linear relationships in analyzing C-PAH/B2-PAH levels in the FDP, FSP, and
YSP samples from NHEXAS study homes. The R2 values from the linear regression analyses
on ELISA C-PAH versus GC/MS B2-PAH for the FDP, FSP, and YSP samples are 72%,
95%, and 97%, respectively. Using the corresponding regression equations on Table 5,20, if
the floor dust, foundation soil, or yard soil's B2-PAH GC/MS response exceeds 1 ppm,: then
the predicted C-PAH ELISA response would be greater than 11.3 ppm, 8.5 ppm, or 7.3 ppm,
respectively.
Screening Tests
Four performance measures are used to characterize the screening performance of PAH ELISA
responses. They include sensitivity (or True Positive Rate), specificity (or True Negative
Rate), positive predictive value (PPV), and negative predictive value (NPV). Each of the
measures is defined in Table 4.1. Table 5.21 shows the frequency distribution of ELISA and
GC/MS measurements on the 2 x 2 contingency tables along with Fisher's Exact test results
and the four performance characteristic measurements for both ELISA total PAH versus
GC/MS total PAH and ELISA C-PAH versus GC/MS B2-PAH in the combination of dust and
soil samples. Tables 5.22 and Table 5.23 show the similar results in food and air samples,
respectively. As shown in these tables, most performance characteristic measurements are
greater than 70% and Fisher's Exact test results indicate a high degree of statistical dependence
between ELISA and GC/MS responses (at 0.05 level). This finding suggested that ELISA is a
good screening tool for total PAH and C-PAH. The relatively poor performance of ELISA on
C-PAH in food samples (Table 5.22) may be partly due to the spiked perdeuterated PAH cross
activities for ELISA assays. The possible errors associated with the estimates of the
performance parameters are 10%, 30%, and 25% for the dust/soil samples, food samples, and
air samples, respectively.
49
-------
Table 5.21. Frequency Distribution of ELISA and GC/MS Measurements on the
Combination of Dust (HD+ES+FDP) and Soil (PS+FSP+YSP) Samples
ELISA
Total
PAH
< lOppm
a lOppm
Total
GC/MS Total PAH
< 1 ppm £ 1 ppm
51 14
21 47
72 61
Fisher's Exact test: p* < .0001
Sensitivity (True Positive Rate): 69% (47/68)
Specificity (True Negative Rate): 78 % (51/65)
Positive Predictive Value (PPV): 77% (47/61)
Negative Predictive Value (NPV): 71% (51/72)
ELISA GC/MS B2-PAH
C-PAH
< 2ppm
£ 2 ppm
Total
< 0.5 ppm £ 0.5 ppm
62 15
19 37
81 52
Total
65
68
133
Total
77
56
133
Fisher's Exact test: p* < .0001
Sensitivity (True Positive Rate): 66% (37/56)
Specificity (True Negative Rate): 81 % (62/77)
Positive Predictive Value (PPV): 71 % (37/52)
Negative Predictive Value (NPV): 77% (62/81)
* p-value: two analyzing methods are not statistically independent at 0.05 level if the p-value
is less than 0.05.
50
-------
Table 5.22. Frequency Distribution of ELISA and GC/MS Measurements on Food
Samples
ELISA
Total
PAH
< 16000 ppb
^ 16000 ppb
Total
GC/MS
< 3000 ppb
9
2
11
Total PAH
:> 3000 ppb
2
5
7
Total
11
7 i
18
Fisher's Exact test: p* = .05
Sensitivity (True Positive Rate): 71% (5/7)
Specificity (True Negative Rate): 82% (9/11)
Positive Predictive Value (PPV): 71% (5/7)
Negative Predictive Value (NPV): 82% (9/11)
GC/MS B2-PAH
EJ-ylOrt
C-PAH
< 8000 ppb
£ 8000 ppb
Total
< 1040 ppb
6
6
12
* 1040 ppb
3
3
6
Total
9
9
18
Fisher's Exact test: p* = 1.00
Sensitivity (True Positive Rate): 33% (3/9)
Specificity (True Negative Rate): 67% (6/9)
Positive Predictive Value (PPV): 50% (3/6)
Negative Predictive Value (NPV): 50% (6/12)
* p-value: two analyzing methods are not statistically independent atX).05 level if the p-value
is less than 0.05.
51
-------
Table 5.23. Frequency Distribution of ELISA and GC/MS Measurements on Air
Samples
ELISA
Total
PAH
< 40000 ng/mL
i 40000 ng/mL
Total
< 50000
8
1
9
GC/MS Total PAH
ng/mL z 50000 ng/mL
2
7
9
Total
10
8
18
Fisher's Exact test: p* = .02
Sensitivity (True Positive Rate): 88 % (7/8)
Specificity (True Negative Rate): 80% (8/10)
Positive Predictive Value (PPV): 78% (7/9)
Negative Predictive Value (NPV): 89% (8/9)
GC/MS B2-PAH
ELISA
C-PAH <
< 6000 ng/mL
* 6000 ng/mL
Total
Fisher's Exact test: p* = .015
Sensitivity (True Positive Rate):
Specificity (True Negative Rate):
Positive Predictive Value (PPV):
Negative Predictive Value (NPV):
1040 ng/mL ^
5
2
7
80% (8/10)
63% (5/8)
73% (8/11)
71% (5/7)
1040 ng/mL
3
8
11
Total
8
10
18
* p-value: two analyzing methods are not statistically independent at 0.05 level if the p-value
is less than 0.05.
52
-------
Further investigation was done to determine whether the ELISA technique is an effective
screening tool for total PAH exposure. A linear regression model of ELISA total PAH versus
GC/MS total PAH was initially fitted to all paired 133 log-transformed data for the
combination of dust and soil samples. A plot of residuals versus GC/MS total PAH indicated
that a poor model fit and a possible lack-of-fit (11). The studentized residuals were large for
10 of the 133 samples. For these 10 samples, nine house dust samples (FDP) and one yard
soil sample (YSP) from the NHEXAS study, ELISA results were all greater than 148 ppm.
These high ELISA results may be from the sample matrix effect on ELISA measurements.
The linear regression model was then refitted to the log-transformed data in dust and soil
samples without these 10 data points. The mean square error was reduced from 7543 to 433
and the R2 was increased from 4% to 45%. The residual plot no longer indicated a lack-of-fit.
Figure 5.5 displays the regression line that was fitted to the log transformed data after
removing these 10 data points in combination with reference lines at log(0.1) ppm and log(l)
ppm of GC/MS total PAH, and log(2.5) ppm and log(12.1) ppm of ELISA total PAH. This
linear regression model shows that GC/MS measurements of total PAH levels at 0.1 ppm and
1 ppm for the dust/soil samples, correspond to ELISA measurements of 2.5 ppm and 12.1
ppm, respectively. Note that there are very few samples in the discordant blocks (e.g., large
GC/MS measurements and small ELISA measurements) but there are quite a few samples
scattering from the best fit line.
In summary, PAH and C-PAH ELISA can be used as screening tool to determine PAH levels
at a threshold level, but cannot provide quantitative measurements for PAH. The performance
of ELISA may be improved, if a representative PAH mixture for each type of sample can be
prepared and used for inhibition curves.
53
-------
DDD: Dusl(HD+ES+FDP)
S S S : Soil(PS+FSP+YSP)
: log(ELISA)=2.5+0.7*log(GC/MS)
3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0
GC/MS Tolal PAH, log(ppm)
Figure 5.5. ELISA total PAH versus GC/MS total PAH, log scale.
54
-------
Evaluation of ELISA for Screening 2,4-D and PCP
Recovery with 2,4-D Extraction Solvent
The results of all evaluation tests conducted with the 2,4-D extraction solvent (75% methanol
in water) are given in Table 5.24. As shown there, the sonication and shaking methods appear
to be equivalent in extraction efficiency; both about 60% for PCP and 85% for 2,4-D and
3,4-D at the mass:volume ratio of 1:2. For larger and smaller spike quantities of 2,4-D, the
recoveries drop significantly, 58% and 26%, respectively. For the larger amount (spiked at
2.5 ^g), solubility may be limited in the solvent mixture; for the smaller amount (spiked at
0.1 jig), there may be a larger percentage of the total spike tightly bound to active surface sites
on the dust, thus limiting extraction efficiency. The extraction ratio of 1 g: 20 mL does not
appear to significantly increase extraction efficiency; the kit-recommended ratio of
10 g: 20 mL appears to be slightly less effective than the 1 g: 2 mL ratio.
Further justification for eliminating consideration of the Ig : 20 mL extraction ratio is shown
in the ELISA results for the humus soil spikes. As shown there, inconsistent results were
obtained in these tests, that was probably because the concentrations of these extracts were at
the low end of ELISA calibration range. With the 1 g: 2 mL extraction ratio, recoveries were
reasonable (110-130 %) at spike levels equivalent to 1-2.5 /xg/g levels, but predictably low
with low spike quantities. Extraction from both clay soil and house dust was less than 50%
with this extraction solvent. The ELISA results for the spiked house dust indicate that false
positives and/or interferences or biases may occur with this matrix type.
Recovery with PCP Extraction Solvent
The same experiments described above were repeated with the NaOH-added extraction solvent
and the results are given in Table 5.25. As shown there, irrespective of spike level or
extraction method, the GC/ECD PCP recoveries average 60-65%, and are similar to the
recoveries obtained with the 75% methanol extraction solvent by GC/ECD method. The
results for 2,4-D and 3,4-D were less predictable, but appeared to indicate enhanced extraction
•''.'," 55
-------
Table 5.24. Extraction and Recovery Efficiency Using 2,4-D Extraction Solvent (75%
Methanol)
Extraction Mass:
Method Volume,
B : mL
2,4-D and
3,4-D spike
amount, UR
ELISA
dilution
factor
Recovery, ± standard deviation
%
n=3
2,4-D
3,4-D
PCP
Humus soil: GC/ECD
sonicalion
sonication
sonkalion
shaking
sonicfttion
sonicaiion
Humus soil: ELISA
sonication
sonicmion
soniciUion
sonicalion
sonicalion
sonication
sonication
sonication
sonicalion
Clay soil: ELISA
sooicalion
1:2
1:2
1:2
1:2
1 :20
10:20
1:2
1 :2
1:2
1:20
1:20
1:20
1:20
1 :20
1 :20
1 :2
2.5
0.1
1.0
1.0
1.0
1.0
2.5
1.0
0.1
2.5
1.0
0.1
2.5
1.0
0.1
1.0
NA'
NA
NA
NA
NA
NA
50
50
50
50
50
50
5
5
5
50
58 ±4
26 ±1
86 ±0
85 ±4
88 ±5
76 ±1
108 ±30
132 ±80
31 ±30*
68 ±24
195 ±195d
55SO±9600°
37 ±28
47 ±9
0 ±0"*
46 ±10
NT*
NT
84 ±0
83 ±4
79 ±2
67 ±0
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NT
NT
61 ±2 (0.1 ug)'
62 ±4(0. lug)
63 ±2(0. lug)
49 ±2(0. lug)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
House dust: GC/ECD
sonicalion
House dust: ELISA
sonicalion
1:2
1:2
1.0
1.0
NA
50
42 ±46
257 ±229*
NT
NA
NT
NA
*) NA-not applicable
b) NT-not tested
c) Spike level for PCP
d) Analysis at low end of ELISA calibration range
o) Concentration outside of ELISA assay calibration range
56
-------
Table 5.25 Extraction and Recovery efficiency Using PCP Extraction Solvent (NaOH
in 75% Methanol)
Extraction Mass:
Method Volume,
e : mL
PCP spike
amount, ug
ELISA
dilution
factor
Recovery, ± standard deviation
%
n=3
PCP
3,4-D
2,4-D
Humus soil: GC/ECD
sonication
sonication
sonication
sonication
shaking
sonication
sonication
Humus soil: ELISA
sonication
sonication
sonication
sonication
sonication
sonication
sonication
sonication
sonication
Clay soil: ELISA
sonication
1:2
1:2
1:4
1:4
1 :4
1 :20
10:20
1:2
1:2
1 :2
1 :20
1:20
1:20
1:20
1:20
1 :20
1:2
7.5
1.0
0.2
0.2
0.2
0.2
0.1
7.5
1.0
0.1
7.5
1.0
0.1
7.5
l.o
0.1
1.0
NA
NA
NA
NA
NA
NA
NA
500
500
500
500
500
500
SO
50
50
500
61 ± 1
60±0
36 ±6
55 ± 2
65 ± 1
71 ±4
59 ±3
105 ±10
126 ±100
227 ±393b
197 ±173
97 ±95b
67 ± 115C
89 ±26
101 ± 17
150 ±144"
123 ±40
NT
NT
54 ±3
61 ±3
79 ±2
94 ±13
80 ±4
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NT
NT
56 ±3 (2ug)«
65±4(2ug)
87 ± 2 (2 ug)
88 ±5(2ug)
65 ± 0 (1 ug)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
House dust: GC/ECD
sonication
House dust: ELISA
sonication
1:2
1:2
1.0
1.0
NA
500
5 ±2
175 ± 85
NT
NA
NT
NA
a Spike amount for 2,4-D
b Analysis at low-end of ELISA calibration range
c Concentration outside of ELISA calibration range
57
-------
with the larger 20 mL extraction volume. The ELISA assay of spiked humus soil samples
indicates reasonable recovery for spike amounts equivalent to 1-7.5 /*g/g PCP levels, either
with a 500 fold dilution of the 1:2 ratio extract or with a 50 fold dilution of the 1:20 ratio
extract. At the lowest spike level, equivalent to 0.1 /tg/g, recoveries were high and precision
was low. Acceptable recovery results (123%) were obtained for the spike to the clay soil.
The spikes to the house dust indicated that there may be difficulties associated with trying to
apply this assay to a matrix as complex as the dust matrix. Recovery, as indicated by
GC/ECD, was extremely low (5%), but ELISA results indicated a much higher recovery
(175%).
Comparison of Extraction Solvents
Tables 5.26 and 5.27, respectively, compare the calculated levels of 2,4-D and PCP in house
dust samples from the 13 low-income homes that are obtained with the two ELISA extraction
solvents. The measure of agreement between the two extraction solvents, the relative percent
difference (RPD) for the two GC/ECD measurements, ranged from 5 to 160% for 2,4-D and
from < 1 to 100% for PCP. As seen in these tables, 7 of 12 samples have RPD values <30%
for the 2,4-D concentrations (Table 5.26) and for the PCP concentrations (Table 5.27). In
most cases where the levels obtained with the two solvents are significantly different (i.e.,
%RPD >30%), the PCP solvent seems to extract the greater amount. These data seem to
indicate a relatively consistent extraction method for either assay. However, as discussed
above, the ELISA assay appears to have significant positive bias for the measurements of
2,4-D and PCP in the complex house dust matrix.
Precision of ELISA Analyses
The precision of the ELISA analyses, as indicated by the relative standard deviation (RSD) for
the three aliquots removed from the extract and assayed separately, is shown in Table 5.28.
As shown there, the precision of the 2,4-D assay appears superior to the PCP assay, not only
58
-------
Table 5.26. Comparison of Concentration of 2,4-D in House Dust with Different
Extraction Solvents
Household
Code
A
B
C
D
E
F
G
H
I
J
K
M
Cone, of 2,4-D in
dust with 2,4-D
extraction solvent,
ng/g
41
4310
1030
1980
1230
233
635
889
146
647
219
475
Cone, of 2,4-D in
dust with PCP
extraction solvent,
ng/e
364
2320
870
2880
1530
354
603
1040
112
547
513
377
Average 2,4-D
cone, in dust,
ng/g
203
3310
948
2430
1380
294
619
965
129
597
366
426
RPD (relative
percent difference)
between two
measurements
160 (PCP>2,4D)«
60 (2,4D>PCP)
16
37 (PCP>2,4D)
22
41 (PCP>2,4D)
5
16
26
17
80 (PCP>2,4D)
23
House dust concentration of 2,4-D resulting from PCP ELISA extraction solvent is greater
than the 2,4-D concentration resulting from the 2,4-D ELISA extraction solvent
59
-------
Table 5.27. Comparison of Concentration of PCP in House Dust with Different
Extraction Solvents
Household
Code
A
B
C
D
E
F
G
H
I
J
K
M
Cone, of PCP in
dust with 2,4-D
extraction
solvent, ng/g
70
97
93
89
78
144
>450
134
37
17
55
101
Cone, of PCP in
dust with PCP
extraction solvent,
ne/g
214
53
93
156
103
141
189
141
34
23
86
93
Average
PCP cone in
dust, ng/g
142
75
93
122
90
142
NAb
137
36
20
71
97
RPD (relative
percent difference)
between two
measurements
100 (PCP>2,4D)'
57 (2,4D>PCP)
<1
55 (PCP>2,4D)
28
2
NA
5
10
29
42 (PCP>2,4D)
8
House dust concentration of PCP resulting from PCP ELISA extraction solvent is greater
than the PCP concentration resulting from the 2,4-D ELISA extraction solvent
NA = not applicable.
60
-------
Table 5.28. Precision of Replicate ELISA Measurements of Soil and House Dust Extracts
Relative Standard Deviation for ELISA Analyses of Sample Extract,
%rsdforn=3
Household Code
A
B
C
D
E
F
G
H
I
J
K
L
M
average
Pathway Soil
31 ,:-•'•
37
14
27
22
42
11
9
18
23
19
15
10
21
2,4-D Assay
Entryway Soil
22
26
18
17
43
15
17
5
35
13
17
NT
19
21
House Dust
22
8
3; 46 '
10
21
15
7
NT"
6
18
31
23
4
18
PCP Assav
House Dust
100
109
35
130
46
43
83
47
73
47
19
NT
65
66
The %rsd=3 from triplicate samples that exceeded the linear range of the calibration
curve; samples were diluted 1:1 and reanalyzed with %rsd=46 for the triplicates
NT = not tested.
61
-------
in the simpler soil matrix, but also in the more complex dust matrix. The average RSD for
triplicate sample extracts in the 2,4-D assay was 20%, and for the PCP assay was about 60%.
2,4-D in House Dust, Entiyway Dust, and Pathway Soil
The data of 2,4-D in house dust, entryway dust, and pathway soil samples from the
13 low-income homes are given hi Appendix O. From these data there appears to be a very
general correlation between the GC and ELISA results. However, the RPD between matched
samples were generally greater than 100%. Duplicate analyses of house dust from home M
show very good agreement internally (i.e., low RPD for duplicate analyses by either GC or
ELISA), and indicate that analytical errors may be negligible compared with interferences to
the ELISA method. The 2,4-D levels, as indicated by the GC/ECD results, appear similar to
those found in other homes (9), and thus this data set may be reasonably representative of the
problems that may be encountered in applying this assay to such a complex matrix.
A significant number of the entryway dust and pathway soil samples show no appreciable
levels above detection limits of either detection method. Because of the low levels, it may not
be reasonable to draw conclusions about the accuracy of the ELISA method from these data.
PCP was detected at substantial levels in only one entryway dust sample.
The concentrations of 2,4-D determined from GC/ECD data, presented in Appendix O, are
corrected by the surrogate recovery value to provide the best estimate of the dust
concentration, and account for incomplete extraction. The surrogate recoveries ranged from
51 to 110% in house dust samples, from 46 to 94% in entryway dust samples, and from 62 to
102% in pathway soil samples. These data indicated reasonable extraction efficiency and
recovery through the analytical protocol. The ELISA data are not similarly corrected, so that
comparing concentrations on a ng/g basis may be of limited value for samples where the
surrogate recovery is low. For this reason, the more direct comparison of the GC/ECD and
ELISA detection is based on a measure of the 2,4-D concentration in the extract itself on the
62
-------
basis of ng/mL. These concentrations are shown for the soil and house dust sample extracts in
Table 5.29. The general agreement between the two techniques is most obvious in the soil
samples, where both techniques indicate very low levels. However, the estimated correlation
coefficient (r) between GC/ECD and ELJSA methods was 0.403 (p=0.17), indicating a
positive but weak relationship between these two methods.
PCP in House Dust . ....
The concentration of PCP in 12 house dust samples, as determined using ELISA and
GC/ECD, is shown in Appendix P. Approximately half of the matched samples show RPD
< 100%, indicating that ELISA may be useful for establishing trends or ranking samples by
concentration. The ELISA assay still tends to show a considerable false positive bias. The
analytical data appear to be adequate given that surrogate recoveries were generally >70%,
and initial extraction data indicated that PCP extraction could be limited to a maximum of
60-65%. As shown in Appendix P, the 2,4-D data included therein demonstrate again the
ability to simultaneously extract and analyze by GC both PCP and 2,4-D.
The direct comparison of PCP levels expressed in ng/mL in the sample extract by GC/ECD
and ELISA is summarized in Table 5.30. Again, approximately half of the extracts have an
RPD for a matched pair that is < 100%. The estimated correlation coefficient (r) for PCP
measured by GC/ECD and ELISA was only 0.311. This result indicates that there is a
positive but weak relationship between GC/ECD and ELISA methods.
Quality Control Data
The levels of alkyl PAH and phthalates found in the field blanks are summarized in
Table 5.31. The field blank air sample was a filter/XAD-2 module 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 containers used for dust/soil and food samples
63
-------
Table 5.29. Comparison of 2,4-D Concentrations in Extracts using GC/ECD and ELISA
Home
A
B
C
D
E
F
G
H
I
J
K
L
M
Pathway Soil
GC/ECD ELISA
<10 <35(9)*
<10 <35(8)
<10 <35(19)
<10 <35(17)
<10 <35(15)
<10 <35(21)
<10 <35(16)
<10 <35(14)
<10 <35(13)
<10 <35(12)
<10 <35(16)
<10 <35(12)
<10 <35(26)
Entrvwav Soil
GC/ECD
<10
<10
43
<10
<10
25
29
NT
19
18
<10
29
<10
ELISA
44
<35 (20)
<35(26)
<35 (18)
41
<35 (23)
<35 (23)
NT
<35 (27)
73
<35 (23)
<35(31)
<35 (19)
House Dust
GC/ECD
12
1160
421
414
313
117
280
311
50
296
66
NT
262
ELISA
217
1250
2950
77b
345
188
243
957"
293
1350
510
NT
1630b
Concentration is less than method detection limit; value in parentheses is the concentration
measured in the assay from the non-linear portion of the calibration curve
3,4-D spiked into sample; effective concentration of 3,4-D (due to cross-reactivity)
subtracted for estimated ELISA concentration
NT- not tested
64
-------
Table 5.30. Comparison of PCP in House Dust Extracts using GC/ECD and ELISA
Home
A
B
C
D
E
F
G
H
I
J
K
L
M
GC/EGD
41
23
39
29
49
62
86
65
12
8
20
NT*
61
ELISA
260
<35(20)*
325
35
100
40
40
740
105
55
130
NT
260
RPD
146
14
157 !
19
68
43
73
168
159
149
147
NAC
124
a) Concentration is less than method detection limit; value in parentheses is the concentration
measured in the assay from the non-linear portion of the calibration curve
b) NT-not tested
c) NA- not applicable
65
-------
Table 5.31. Levels of Alkyl PAH and Phthalates Found in Field Blanks
Total amount, ng
Compound
2-Methylnaphthalene
1-Methylnaphthalene
C2-alkylnaphthalene isomers
Cl-alkylphenanthrene isomers
C2-alkylphenanthrene isomers
Cl-alkylpyrene isomers
C2-alkylpyrene isomers
Cl-alkylchrysene isomers
C2-alkylchrysene isomers
Cl-alkylbenzo[a]pyrene isomers
Dimethylphthalate
Dielhylphathalate
Di-n-butyl phthalate
Butylbenzylphthalate
Bis(2-ethylhexyl)phthalate
Di-n-octylphthalate
Air
4.89
2.92
6.71
12.0
<1
<1
<1
<1
<1
<1
14.9
151
497
2400
1370
9.26
Dust/Soil
1.35
<1
2.30
1.77
18.7
3.14
<1
<1
<1
<1
3.20
17.5
27.3
11.2
44.7
1.80
Food
3.51
2.00
<1
<1
<1
<1
<1
<1
<1
<1
<1
14.7
46.6
207
25.3
2.34
66
-------
processed through field handling and shipping. As shown in Table 5.31, trace amounts of
alkyl PAH were found in the field blanks. The levels of phthalates found in the field blanks
were higher than those of the alkyl PAH levels. The phthalates in the filter/XAD-2 air blank
may originate partly from the XAD-2 resin and partly from the sampling cartridge. The
plastic containers used for the food samples may also contribute to the phthalates fpund in the
food field blank. Note that the reported concentrations of alkyl PAH and phthalates for
multimedia samples in this report are already corrected for the levels of respective analytes
found in the field blanks. i
Known amounts of perdeuterated PAH were spiked onto each dust/soil sample prior to sample
preparation. Table 5.32 summarizes the recovery data of the spiked PAH for each type of
sample. In general, quantitative recoveries (> 80 %) for the spiked PAH were obtained.
These data indicated that there was no significant loss of PAH through sample preparation
steps.
Respective control solutions were analyzed in conjunction with the sample extracts in every
assay run and treated the same way as the sample extracts for PAH, C-PAH, 2,4-D, and PCP
ELISA. A three point calibration curve was generated in every assay run. The goal for the
acceptance criteria for each assay is to obtain a correlation coefficient (r) greater than 0.99.
Table 5.33 summarizes the quality control data for PAH, C-PAH, 2,4-D and PCP ELISA. As
shown, the correlation coefficients were greater than 0.99 for all but two of the 33 assay runs.
In general, the results of the control solutions and proficiency samples were within 30 percent
of the specified values.
67
-------
cu
o\
-H
00
o\
o\
«« £ <> > 5
'•a
h .-3
§ S
e
2 I
II *
2 ^
cd CS
* II
-
3
68
-------
1
I
s
3
cc
I
4f
u
O >
IH
*?
g § r'•&• £ s-
--
r~ t^ o
VO C- ••* O\
f? & -• &
S. •«.- P».
£ f
CO CO CO
IT
5; S S5
VH O ' *~* *"N
"j O
S 82 «
^O ^^ *O ^i ^^ n5 V^ V}
oooooooo
•-< "H CO
\O \o m
SS 52
9>Q\9sO>9tOOiOOQS
O\O\ONOxv\OOvO\Ot
oo d d d -^ d d d
oooooooo
&, »
69
-------
f
u
&
o
»
or
ON
cs
I I §
oo c^ ^»
3 s* a
1 1 1 1
1 1 1 1 1 1
i i i i
1 1 1 1 1 1 1 1
s- s-
oooooooo o o o o o o
CL,
& |