Brominated Dioxins and Furans
in Human Adipose Tissue
Final Report
Paul H. Cramer
John S. Stanley
Karin Bauer
Randy E. Ayling
Kelly R. Thornburg
John Schwemberger
For Exposure Evaluation Division, TS-798
Office of Toxic Substances
U.S. Environmental Protection Agency
401 M Street, SW
Washington, D.C. 20460
Attn: Ms. Janet Remmers, Work Assignment Manager
Dr. Joseph J. Breen, Project Officer
EPA Contract No. 68-02-4252
MRI Project No. 8863-A(27)
Work Assignment 27
April 11, 1990
U.S. Environmental Protection Agency
Region 5, Library (Pi -ll''
77 West Jackson 3:, ./ , ,_. Tlwo
Chicago, IL 606G4-3o>j
-------
DISCLAIMER
This document has been reviewed and approved for publication by the Office of
Toxic Substances, Office of Pesticides and Toxic Substances, U.S. Environ-
mental Protection Agency. The use of trade names for commercial products does
not constitute Agency endorsement or recommendation for use.
-------
PREFACE
This report provides a summary of the analyses of the FY1987 NHATS
composited human adipose tissue samples for polybrominated dibenzo-p-dioxins
(PBDDs) and polybrominated dibenzofurans (PBDFs) completed for all sample
batches under EPA Contract No. 68-02-4252, Work Assignment 27, "Analysis of
Human Adipose Tissue for Dioxins and Furans." These adipose tissue samples
were analyzed for specific 2,3,7,8-substituted PBDDs and PBDFs according to
the analytical protocol identified in the Quality Assurance Program Plan
(QAPP) for this work assignment. This document presents the first reported
effort to determine the presence of PBDDs and PBDFs in the human population.
These data and reporting activities were generated by Midwest Research
Institute under the direction of Mr. Paul H. Cramer and Dr. John S. Stanley
for EPA's Office of Toxic Substances Field Studies Branch. Mr. Michael
McGrath and Mr. Paul Cramer were responsible for preparation of the composite
samples, Mr. Kelly Thornburg and Mr. Gene O'Donnell conducted the HRGC/HRMS
analyses, and Mr. Randy Ayling was responsible for standards preparation and
the data reduction efforts. The study design was provided by Battelle
Columbus Laboratories. Mr. John Schwemberger of the Design Development Branch
of OTS and Ms. Karin Bauer of MRI provided the statistical evaluation of the
quality control data for assessment of overall method performance.
MIDWEST RESEARCH INSTITUTE
Paul C. Constant
Program Manager
Reviewed:
Jack Balsinger
Quality Assurance Coordinator
Ap
in E. Going, Ph.D.
Hrector
Chemical Sciences Department
April 11, 1990
ii
-------
AUTHORS AND CONTRIBUTORS
This study on the determination of polybrominated dibenzo-p-dioxins
(PBDDs) and dibenzofurans (PBDFs) in human adipose tissue is one part of the
ongoing Human Adipose Tissue Survey. The work reported herein was conducted
as a collaborative effort among EPA's Office of Toxic Substances, Midwest
Research Institute, and Battelle Columbus Division.
EPA participation was within the Office of Toxic Substances (OTS)
Exposure Evaluation Division, the Field Studies Branch (Ms. Janet Remmers,
Work Assignment Manager, and Dr. Joseph Breen, Project Officer) and the Design
and Development Branch (Mr. John Schwemberger, Work Assignment Manager, and
Ms. Edith Sterrett, Project Officer). Contract support to OTS included MRI
for the conduct of the method evaluation, sample analysis, and reporting
(Mr. Paul Cramer and Dr. John Stanley, Co-work Assignment Managers, and
Mr. Paul Constant, Program Manager); Battelle Columbus Division for processing
of patient summary reports (Ms. Tamara Collins and Ms. Jan Clark) and
composite design (Ms. Barbara Leczynski); and Mr. Robert Heath for design of
the approach to the quality control program.
i i i
-------
TABLE OF CONTENTS
Preface 11
Authors and Contri butors 111
List of Figures v
List of Tables vi
Executive Summary vii
I. Introduction 1
A. Work Assignment Objectives 1
B. Organization of Report 2
II. Experimental Approach 2
A. Composite Design for the FY 1987 NHATS Samples 2
B. Laboratory Compositing Procedures 3
C. Analytical Procedures 4
D. QC for Chemical Analyses 7
III. Resul ts 16
A. Preliminary Method Evaluation 16
B. Analysis Results for the FY 1987 NHATS
Composi tes 19
IV. Quality Control 29
A. Internal Quantisation Standard Recoveries 29
B. Calibration 30
C. Spiked Internal QC Samples 36
D. Control QC Samples 36
E. Method Blanks 36
F. Tridecane Blanks 48
V. Assessment of Method Performance 48
A. Assignment of Data Values 48
B. Descriptive Statistics 48
C. Regression Analysis 51
D. Conclusions 60
VI.
References.
62
IV
-------
LIST OF FIGURES
Number Title
1 Ion plots from the analysis of sample 16309 (ACD8700265)
for 2,3,7,8-TBDF
2 Reconstructed ion chromatogram (RIC) from the analysis of
sample 16262 (ACD8700194) for tetra- through hexabromo
dioxins and furans
3 Ion plots from the analysis of sample 16269 (ACD8700194)
for tetrabromo dioxins and furans
4 Ion plots from the analysis of sample 16262 (ACD8700194)
for pentabromo dioxins and furans
5 Ion plots from the analysis of sample 16262 (ACD8700194)
for hexabromo dioxins and furans
6 Recoveries of i3C12-TBDF from the 48 NHATS composites
and 20 QC samples
7 Recoveries of i3C12-2,3,7,8-TBDF from the 48 NHATS
composites and 20 QC samples
8 Recoveries of i3C12-l,2,3,7,8-PeBDF from the 48 NHATS
composites and 20 QC samples
9 PBDD/PBDF recoveries from quality control samples (low
spike level)
10 PBDD/PBDF recoveries from quality control samples (high
spi ke 1evel)
11 Regression analysis and 9556 confidence intervals for
QC data on 2,3,7,8-TBDF
12 Regression analysis and 95% confidence intervals for
QC data on 2,3,7,8-TBDD
13 Regression analysis and 95% confidence intervals for
QC data on 1,2,3,7,8-PeBDF
14 Regression analysis and 95% confidence intervals for
QC data on 1,2,3,7,8-PeBDD
15 Regression analysis and 95% confidence intervals for
QC data on 1,2,3,4,7,8-HxBDF
16 Regression analysis and 9556 confidence intervals for
QC data on 1,2,3,4,7,8-HxBDD
17 Percent bias and 95% confidence limits for individual
PBDD and PBDF isomers
Page
22
24
25
26
27
32
33
34
45
46
53
54
55
56
57
58
61
-------
LIST OF TABLES
Number Title Page
1 Internal Standard Spiking Solutions for Chlorinated and
Brominated Species 5
2 Concentration Calibration Solutions for PBDD/PBDF 8
3 Typical Daily Sequence for PBDD/PBDF Analysis 9
4 Ions Monitored for PBDD/PBDF 10
5 HRGC/HRMS Operating Conditions for PBDD/PBDF Analysis 11
6 Quality Control Samples 12
7 Spiking Solutions for PBDDs and PBDFs 14
8 Brominated Spike Check Results 15
9 Mean RRFs for PBDDs and PBDFs From Initial HRGC/HRMS
Cal ibration 17
10 Recovery of Brominated Dioxins and Furans From Adipose
Ti ssue 18
11 Detection Limit Summary 20
12 Estimated Bromodiphenyl Ether Concentrations (pg/g) in
NHATS FY87 Composites 28
13 Index to Figures 6, 7, and 8 and Order of Sample Analysis... 31
14 Calibration Data—Relative Response Factor 35
15 Calibration Event Summary 37
16 Spike Recovery Summary for Samples Outside DQOs 39
17 Batch 1 Spiked Sample Results 40
18 Batch 2 Spiked Sample Results 41
19 Batch 3 Spiked Sample Results 42
20 Batch 4 Spiked Sample Results 43
21 Batch 5 Spiked Sample Results 44
22 Control Sample Results 47
23 Descriptive Statistics for Control Lipid Samples 49
24 Descriptive Statistics for Spiked Lipid QC Samples 50
25 Estimates from Regression of Measured vs. Spiked
Concentrations 59
vi
-------
EXECUTIVE SUMMARY
As part of the U.S. Environmental Protection Agency's efforts to
determine human exposure to potentially toxic compounds from commercial
products and environmental routes, a study has been conducted to determine
human adipose tissue levels of polybrominated dibenzo-p-dioxins (PBDDs) and
dibenzofurans (PBDFs). The study included the analysis of adipose tissues
collected from the general U.S. population in fiscal year 1987 through the
National Human Adipose Tissue Survey (NHATS). The 2,3,7,8-substituted
analytes (tetra- through hexabrominated dioxins and furans) targeted for
analysis were not detected in any of the study samples. However, the
analytical techniques and the data generated provide a basis for application
and comparison in future human exposure studies. Preliminary data were
generated for another class of compounds, polybrominated diphenyl ethers
(PBDPEs), which were tentatively identified as interferences in the analytical
procedure. This report describes the development and evaluation of the
analytical protocol and the results of the analysis efforts.
Regulations recently promulgated under Sections 4 and 8 of the Toxic
Substances Control Act (TSCA) have focused attention on the possible exposures
to halogenated (brominated and chlorinated) dibenzo-p-dioxins and dibenzo-
furans from commercial products. The purpose of the regulations is to limit
the potential releases of these compounds to the environment. While consider-
able information has been generated on the chlorinated compounds (PCDDs and
PCDFs) in commercial products and environmental and biological matrices, few
or no data have been generated regarding the incidence of the PBDDs and PBDFs.
Since no methods existed for measurement of PBDDs and PBDFs, this
study required modification and evaluation of existing sample preparation
techniques and HRGC/HRMS analysis methods that were originally developed for
determination of PCDDs and PCDFs. The preliminary method evaluation studies
focused on the development of HRGC/HRMS acquisition parameters, establishing
linearity of the calibration curves, determining method sensitivity, and
conducting the analysis of a series of spiked samples. The results of the
preliminary method evaluation studies demonstrated that the PBDDs and PBDFs
could be determined at low picogram/gram (pg/g, ppt) levels.
Following the evaluation of the analytical method, a total of
48 composite samples (prepared from 865 individual tissue specimens) and
20 quality control (QC) samples were analyzed for this study. The composite
samples represented the nine U.S. census divisions and three age groups (0-14,
15-44, and 45 plus). The targeted 2,3,7,8-substituted PBDDs and PBDFs were
not detected in any of the composite samples. Detection limits for each
congener were calculated for each sample. The average detection limits
calculated from the 48 composite samples were approximately 1 pg/g for TBDD
and TBDF, 10 pg/g for the PeBDD, PeBDF, and HxBDD, and 40 pg/g for the HxBDF.
Data on the precision and accuracy of the method for three PBDDs and
three PBDFs were generated from the analysis of the QC samples. These samples
included five unspiked and 10 spiked (five replicates at two spike levels)
adipose tissue (lipid) samples. Statistical treatment of the QC data
vii
-------
demonstrated that with the exception of 1,2,3,4,7,8-HxBDF, the analytical
method is unbiased in measuring sample concentrations. The bias determined,
though not statistically significant, is larger for the hexabromodioxin than
for the tetra- and pentabrominated compounds. For the hexabromofuran, the
method has a significant negative bias believed to be due to either the lack
of the appropriate internal quantitation standard, the difference in the HRMS
sensitivity, or both. These limitations are reflected in the higher detection
limits reported for the HxBDF. The method provides precise results for the
tetra- and pentabrominated compounds, while providing poor precision for the
hexabromo compounds.
While PBDDs and PBDFs were not detected in the study samples, there
was evidence of the presence of other brominated compounds in the adipose
tissues. Specifically, responses were noted that correspond to the
qualitative criteria for PBDPEs (hexa- through octabromo). It is recommended
that these responses should be confirmed as brominated diphenyl ethers through
additional analysis efforts in order to expand the data base on documented
human exposures.
The overall program effort has provided a basis for evaluating the
potential exposure to PBDDs and PBDFs. Although the data generated at this
time do not provide an indication of the presence of these compounds in human
adipose tissues, this study will serve as a landmark effort for future
comparison studies.
vm
-------
I. INTRODUCTION
The National Human Adipose Tissue Survey (NHATS) operated through
the U.S. Environmental Protection Agency's (EPA's) Office of Toxic Substances/
Exposure Evaluation Division (OTS/EED) has served as the Agency's primary
mechanism for monitoring human exposure to potentially toxic materials since
the early 1970s. The NHATS program, as operated through the 1970s, served
primarily as a monitor of the exposure to organochlorine pesticides and
PCBs. However, over the last several years, OTS has implemented a strategy
for a broader view of the number of toxic substances present in adipose
tissues that provide evidence of exposure either through environmental routes
or as a result of consumer products use.
Samples that were collected in the FY 1982 NHATS collection were
analyzed for a broad range of compounds that included volatile organics
(chlorinated alkanes, chlorinated aromatics, and aromatics), semivolatile
organics (organochlorine pesticides, PCBs by homolog, polynuclear aromatic
hydrocarbons, chlorinated benzenes, phthalates and phosphate esters), poly-
chlorinated dibenzo-p-dioxijis (PCDDs) and polychlorinated dibenzofurans
(PCDFs), and trace metals.1"1* This required development of procedures based
on state-of-the-art instrumentation, high-resolution gas chromatography/mass
spectrometry (HRGC/MS).
The samples collected from the FY 1984 NHATS program were analyzed
for the broad range of semivolatile organic compounds using the same prepara-
tion procedures as used for the FY 1982 samples and HRGC/MS and packed column
gas chromatography/electron capture detection (PGC/ECD) to demonstrate the
comparability of the techniques for specific organochlorine pesticides. The
broad scan effort for semivolatile organics based on the HRGC/MS procedure has
been repeated for the samples collected in the FY 1986 NHATS program, although
the emphasis on the analysis has been expanded to include the determination of
compounds that are classified under other federal regulations, such as SARA
Title III.
In addition to these specific NHATS programs, OTS has been involved
in two additional studies to determine human body burden levels of PCDDs and
PCDFs. These studies include a collaborative effort with the Veterans
Administration to determine the levels of specific 2,3,7,8-substituted con-
geners in adipose tissues collected from adult males between 1972 and 1982.
The second study focused on developing national estimates of body burdens of
these compounds using samples that were collected in the FY 1987 program.s~a
As a result of regulations promulgated during 1987 under Sections 4
and 8 of the Toxic Substances Control Act (TSCA), EPA is responsible for
ensuring that specific commercial products do not present a route for release
of halogenated (chlorinated or brominated) dibenzo-p-dioxins and dibenzofurans
to the environment. As a result of the focus of this regulation, there is a
need for developing data on the exposure of humans to these compounds.
A. Work Assignment Objectives
The overall objective for this work assignment was to provide data
on the body burden levels of polybrominated dibenzo-p-dioxins (PBDDs) and
1
-------
dibenzofurans (PBDFs) using adipose tissue specimens that were collected in
the FY 1987 NHATS collection effort. Data quality objectives for this program
effort are presented below:
Measurement Data quality objective
Initial Relative Response Relative standard deviation (< 20%
Factor (RRF) calibration for tetra isomers, < 30% for all
others)
Daily RRF Within 20* of initial RRF deter-
mined for tetra isomers, within
30% for all others
IQS recovery from samples 40%-150%
Recovery of compound from 40%-150%
spiked control samples
Data were also generated for the PCDD and PCDF residues, although the results
of this study have been provided under separate cover.7"*
B. Organization of Report
The remainder of this report provides details on the sources of the
samples, the procedures used to composite individual specimens into composites
representative of three different age groups from nine different census divi-
sions, and the sample preparation and analysis procedures (Section II). Sec-
tion III of this report provides a summary of the analytical results for this
program effort. Summaries of the specific quality control data that were gen-
erated in support of this analysis effort are presented in Section IV.
Statistical treatment of the quality control data and an assessment of the
overall method performance for PBDDs and PBDFs are presented in Section V.
References which have been cited in the report are presented in Section VI.
II. EXPERIMENTAL APPROACH
This section describes the procedures that were used for preparing
composited samples from the FY 1987 NHATS collection, the methods used for
preparing the composited samples for analysis, and the instrumental conditions
that were used to conduct the high-resolution gas chromatography/high-resolu-
tion mass spectrometry (HRGC/HRMS) analysis of the sample extracts. Addi-
tional details are presented on quality assurance/quality control practices
and criteria that were implemented to support overall data quality.
A. Composite Design for the FY 1987 NHATS Samples
The composite design for the FY 1987 NHATS samples was prepared by
Battelle Columbus Division (BCD).5 In developing the compositing design, BCD
-------
considered five criteria: maintaining similarity with the FY 1982 NHATS
composite design; maintaining equal weighting of the specimens within the
composites; specifying an increased number of samples that were pure sex
composites versus the FY 1982 study; controlling metropolitan statistical area
(MSA) effects; and providing the best range of group percentages across the
composite samples. After considering these criteria, BCD developed a study
design that consisted of 48 composited samples prepared from 865 individual
specimens that were collected from 41 MSAs.
In addition to developing the composite design, BCD provided a
design for order of analysis to be followed for preparation and analysis of
the composite samples along with quality control samples such that the ana-
lysts were blinded to the characteristics of each sample and that the samples
were sufficiently randomized to avoid any potential biases due to age, gender,
or geographic region. The FY 1987 NHATS composite design has been summarized
as a separate report.5
B. Laboratory Compositing Procedures
NHATS FY 1987 specimens from nine census divisions and three age
groups were divided into 48 composite samples as identified in the composite
design provided by BCD. BCD provided MRI with data sheets for each composite
that identified the specific individual specimens that were to be included in
each composite. Each composite consisted of 3 to 32 specimens. The composite
sample data sheets provided sufficient information (EPA ID number, package
number, sample weight, hospital code, etc.) such that the Individual specimens
could be cross-checked with the study design. The data sheets provided by BCD
were also used to record the actual laboratory compositing procedures.
Initially, the samples were grouped into composites, and any samples
of questionable weights were noted. The mass of each specimen required for a
composite was determined by dividing the targeted weight of the final com-
posite (10 g) by the number of individual samples to be included. For exam-
ple, a composite to be prepared from 20 individual specimens required 0.5 g of
each specimen. Three specimens of the available 865 specimens (identified
below) did not have the weight required by the composite design. These
results were relayed to the EPA work assignment manager (WAM) by telephone on
June 23, 1988. After consultation with the Design Development Branch, the EPA
WAM forwarded to MRI the following responses regarding the problem samples on
June 24, 1988.
Composite number Sample number
Problem
Response
ACD8700023
ACD8700032
ACD8700201
8706954 Low weight, ~ 0.1 g
need 0.5 g
8701765 No sample remaining
8703464 Low weight, - 1.3 g
need 2.0 g
Include as is
Omit
Include as is
-------
The samples 1n a composite were placed on dry ice during the
compositing procedure to ensure that thawing would not occur. An electronic
4-place balance was used to weigh the samples. The calibration of the balance
was checked before any weighing was begun and once during the sample weighings
with a Class P set of weights (laboratory-grade, tolerance 1/25,000).
To weigh the samples, a clean culture tube was labeled with the
composite number and placed on the balance and the weight tared. A sample was
removed from the storage area, the jar opened, and a portion of the frozen
adipose removed with a clean stainless steel spatula. The adipose was placed
in the culture tube and the weight recorded to three decimal places on the
compositing sheets. Additional adipose was added if necessary. A goal of
±10% of the desired weight (0.30 to 3.0 g depending on the number of specimens
specified to achieve 10 g) was attempted where possible. The weight of the
individual specimens were recorded on the data sheets provided by BCD.
The weight of the culture, beaker, and adipose tissue was rezeroed
and the next specimen in the composite was weighed. A clean spatula was used
for each specimen. This procedure was repeated for each specimen in the
composite. When the composite was completed, it was sealed and stored in a
sample freezer at -10°C. All data on the actual compositing procedures were
recorded on the data sheets provided by BCD. All data sheets were submitted
as a separate interim report to document the compositing activity.*
C. Analytical Procedures
Analytical procedures included the extraction and cleanup of the
composite tissue samples and the analysis by HRGC/HRMS. These procedures are
described below in detail.
1. Sample Preparation
a. Extraction
After compositing, the adipose tissue composites (- 10 g) were
stored at -10°C in 50-mL culture tubes sealed with aluminum foil. To begin
the sample extraction procedure, the samples were allowed to come to room
temperature and then fortified with 100 uL of the chlorinated internal
quantitation standards (IQS) spiking solution and 100 yL of the brominated IQS
spiking solution (Table 1). Ten milliliters of methylene chloride was added
and the sample homogenized for 1 min with a Tekmar Tissuemizer. The mixture
was allowed to separate and the methylene chloride was decanted through a
funnel of sodium sulfate into a 100-mL volumetric flask. The homogenization
was repeated with a fresh 10-mL portion of methylene chloride. The culture
tube was rinsed with additional methylene chloride and the remaining contents
of the tube transferred to the funnel. Finally, the funnel was rinsed with
additional methylene chloride until the volumetric was brought up to volume
(100 ml).
At this point the flask was stoppered and inverted several
times to mix the extract, and 1.0 ml was removed with a disposable pipet and
placed into a preweighed 1-dram (measured to 0.0001-g) glass vial. The
-------
Table 1. Internal Standard Spiking Solutions
for Chlorinated and Brominated Speciesa
Concentration
Compound (pg/wL)
Chlorinated Internal Quantitation Standards'*
i3C12-2,3,7,8-TCDD 5
i3C12-2,3,7,8-TCDF 5
i3C12-l,2,3,7,8-PeCDD 5
i3C12-l,2,3,7,8-PeCDF 5
i3C12-l,2,3,6,7,8-HxCDD 12.5
i3C12-l,2,3,6,7,8-HxCDF 12.5
i3C12-l,2,3,4,6,7,8-HpCDD 12.5
i3C12-l,2,3,4,6,7,8-HpCDF 12.5
i3C12-OCDD 25
Brominated Internal Quantitation Standards**
i3C12-2,3,7,8-TBDD 5
i3C12-2,3,7,8-TBDF 5
i3C12-l,2,3,7,8-PeBDF 5
Internal Recovery Standard0
i3C12-l,2,3,4-TCDD 50
i3C12-l,2,3,7,8,9-HxCDD 125
aAll internal quantitation and recovery stan-
ards were obtained as solutions from Cambridge
Isotope Laboratories, Woburn, Massachusetts.
DPrepared in isooctane. One hundred microliters
spiked. Separate solutions were used for
chlorinated and brominated species.
""Prepared in tridecane. Used for both chloro
and bromo analyses. Ten microliters were
added to each final extract.
-------
methylene chloride In the vial was reduced under nitrogen until a constant
weight (measured to 0.0001 g) of extractable lipld was obtained. The weight
of the lipid was obtained by difference and the percent lipid for the
composite calculated.
The remaining portion of the extract (99 ml) was quantitatively
transferred with a 20- to 30-mL rinse to a 500-mL round-bottomed flask. The
extract was concentrated to an oily residue (extractable Hpid) using rotary
evaporation.
b. Bulk Lipid Removal
An acidic silica gel slurry cleanup of the extract was
conducted by adding 200 ml of hexane and a Teflon-coated stirring bar to the
Hpid 1n the round-bottomed flask. Then, while stirring the extract on a mag-
netic stir plate, 100 g of 40% w/w sulfuric add-Impregnated silica gel was
slowly added to the extract. The mixture was stirred for 2 h. During the 2-h
slurry period, acid/neutral silica gel columns (4 g 40% H2SOi»/sil1ca gel, 1 g
silica gel) (E. Merck, W. Germany) were prepared. After the 2-h period, the
slurry mixture was allowed to settle and the hexane was decanted off the acid-
impregnated silica gel through a funnel of sodium sulfate Into the acid/
neutral silica gel column. Two 50-mL allquots of hexane were added to the
slurry mixture, and the mixture was stirred for 15 min each time. The rinses
were added to the silica gel column through the sodium sulfate funnel. The
eluate of the column was collected in a 500-mL Kuderna evaporation flask. An
additional 50 ml of hexane was placed onto the column when the solvent level
had reached the level of the chromatographic packing. The extract was then
reduced in volume over a steam bath, and the final volume was adjusted to
approximately 1 ml using nitrogen blow down.
c. Separation of Chemical Interferences
A layered, neutral alumina (ICN Adsorbent!en, W. Germany)
column was prepared containing 1 g sodium sulfate, 1 g neutral alumina, and
1 g sodium sulfate. The extract from the acid/neutral silica gel column was
transferred to the alumina column, followed by two 1-mL portions of hexane and
10 ml of 8% (v/v) methylene chloride in hexane. These eluents were archived.
The dioxins and furans (chloro and bromo) were eluted from the column with
15 ml of 60% (v/v) methylene chloride in hexane. The eluent was concentrated
to approximately 2 mL under a stream of nitrogen.
A disposable column of AX-21 carbon (Anderson Development
Corporation) on silica gel was prepared using approximately 1 g of the mixed
adsorbent (1 g AX-21 and 19 g silica gel) and preeluted with 4 ml of toluene,
2 ml of 75:20:5 methylene chloride/methanol/benzene, and 2 ml of 1:1 cyclo-
hexane/methylene chloride. The concentrated eluate from the alumina column
was added to the AX-21/silica gel column followed by two 1-mL hexane rinses.
The column was eluted sequentially with two 0.5-mL aliquots of hexane, 10 mL
of 1:1 cyclohexane/methylene chloride, and 5 mL of 75:20:5 methylene chlo-
ride/methanol /benzene. These eluents were combined and archived. The columns
were then turned upside down and the dioxins and furans eluted with 20 mL of
toluene. The extract was then reduced in volume to approximately 100 wL at
-------
which time 10 yL of recovery standard was added (Table 1) and the volume
further reduced to 10 yL under nitrogen. The extract was stored in a freezer
until HRGC/HRMS analysis.
2. HRGC/HRMS Analysis
All sample analyses were completed using a Kratos MS-50TC double-
focusing HRMS. Initial calibration of the GC/MS system was conducted by
making duplicate 1-iiL injections of the standards listed in Table 2. A CSS
(10 pg/yL of TBDD, TBDF to 25 pg/yL HxBDD, HxBOF) standard was analyzed on a
daily basis to ensure adherence to the initial calibration curve.
Sample analysis by HRGC/HRMS was conducted after Initial and routine
calibration criteria were met. Prior to the injection of the first sample, an
Injection of tridecane was analyzed to document system cleanliness. Correc-
tive action was taken by analyzing another tridecane blank if any evidence of
system contamination was found. A typical daily sequence of injections is
shown in Table 3.
A 1-wL aliquot of the extracts was injected into the GC/MS system,
operated under the conditions previously used to produce acceptable results
with the daily calibration standard.
Selected ion monitoring (SIM) data were acquired according to the
same acquisition and MS operating conditions previously used to determine the
relative response factors (Tables 4 and 5). Instrument performance was moni-
tored by examining and recording the peak areas for the recovery standard,
i3C12-l,2,3,4-TCDD.
D. QC for Chemical Analyses
The following QC criteria were targeted for the chemical analysis
portion of this program.
1. Instrument Performance
The instrument performance was characterized primarily by three
criteria: a mass resolution > 3,000, relative response factors (i.e.,
adherence to the initial RRFs) and instrument sensitivity.
a. Mass Calibration and Resolution
The mass spectrometer was tuned on a daily basis to yield
optimum sensitivity and peak shape using an ion peak (m/z 542.9664) from
perfluorokerosene (PFK). The resolution was visually monitored and maintained
at s 3,000 (10% valley definition) to provide adequate noise rejection while
maintaining good ion transmission. A visual check of the static resolution
was made by using the peak matching unit before and after each analysis.
-------
Table 2. Concentration Calibration Solutions for PBDD/PBDF
Concentration in calibration
Compound
2,3,7,8-TBDD
2,3,7,8-TBDF
1,2,3,7,8-PeBDD
1,2,3,7,8-PeBDF
1,2,3,4,7,8-HxBDD
1,2,3,4,7,8-HxBDF
CS1
200
200
200
200
500
500
CS2
100
100
100
100
250
250
CS3
50
50
50
50
125
125
CS4
25
25
25
25
62.5
62.5
CS5
10
10
10
10
25
25
solutions in pg/yL
CS6
5
5
5
5
12.5
12.5
CS7
2.5
2.5
2.5
2.5
6.25
6.25
CSS
1
1
1
1
2.5
2.5
Internal Quantisation Standards
i3C12-2,3,7,8-TBDD 50 50 50 50 50 50 50 50
i3C12-2,3,7,8-TBDF 50 50 50 50 50 50 50 50
i3C12-l,2,3,7,8-PeBDF 50 50 50 50 50 50 50 50
Internal Recovery Standard
i3C12-l,2,3,7,8,9-HxCDD 125 125 125 125 125 125 125 125
Note: The unlabeled and labeled brominated standard were obtained from Cambridge Isotope
Laboratories, Woburn, Massachusetts. All standards were prepared in tridecane.
8
-------
Table 3. Typical Daily Sequence for PBDD/PBDF Analysis
1. Tune and calibrate mass scale versus perfluorokerosene (PFK).
2. Inject concentration calibration solution 10 to 25 pg/yL (CS-5) solution.
3. Inject blank (tridecane).
4. Inject samples 1 through "N."
5. Inject concentration calibration solution 10 to 25 pg/yL (CS-5) solution
or other concentration calibration solutions CS1 to CS8 to bracket
observed sample concentration.
-------
Table 4. Ions Monitored for PBDD/PBDF
Descriptor ID
Bl -COBr (TBDF)
TBDF
TBDF
i3C12-TBDF
i3C12-TBDF
-COBr (TBDD)
TBDD
TBDD
i3C12-TBDD
i3C12-TBDD
i3C12-HxCDD
i3C12-HxCDD
HxBDPE
PFK lock mass
B2 -COBr (PeBDF)
PeBDF
PeBDF
i3C12-PeBDF
i3C12-PeBDF
-COBr (PeBDD)
PeBDD
PeBDD
HpBDPE
PFK lock mass
B3 -COBr (HxBDF)
HxBDF
HxBDF
-COBr (HxBDD)
HxBDD
HxBDD
OBDPE
PFK lock mass
Mass
374.784
481.697
483.695
493.737
495.735
390.779
497.692
499.690
509.732
511.730
401.856
403.853
643.530
480.970
454.693
561.606
563.604
573.646
575.644
470.688
577.601
579.599
721.441
580.963
532.603
641.514
643.512
548.598
657.509
659.507
801.349
580.963
Nominal dwell
time (s)
0.036
0.056
0.056
0.056
0.056
0.034
0.052
0.056
0.026
0.026
0.033
0.034
0.043
0.056
0.059
0.049
0.049
0.092
0.092
0.056
0.092
0.092
0.092
0.072
0.098
0.085
0.085
0.098
0.079
0.085
0.131
0.092
10
-------
Table 5. HRGC/HRMS Operating Conditions for
PBDD/PBDF Analysis
Mass spectrometer; Kratos MS50-TC
Accelerating voltage: 8,000 V maximum
Trap current: 500 yA
Electron energy: 70 eV
Electron multiplier voltage: -2,000 V
Source temperature: 280°C
Resolution: > 3,000
Overall SIM cycle time: 1 s
Gas chromatograph; Carlo-Erba MFC-500
Column coating: DB-5
Film thickness: 0.25 y
Column dimensions: 30 m x 0.253 mm
He linear velocity: - 35 cm/s
He head pressure: 0.75 kg/cm2
Injection type: Splitless
Split flow: 30 mL/min
Purge flow: 3 mL/min
Injector temperature: 300°C
Interface temperature: 290°C
Injection size: 1 yL
Initial temperature: 200°C
Initial time: 2 min
Temperature program: 5°/min to 330°C
Final hold time: 6 min
11
-------
Mass calibration of the mass spectrometer for the HRGC/MS
analysis of PBDD/PBDF was carried out on a daily basis. The magnetic field
was adjusted to pass m/z 368 at full accelerating voltage. PFK was admitted
to the MS and an accelerating voltage scan from 8,000 to 800 V was acquired by
the data system. This corresponded to an effective mass range of 369 to
843 amu. Upon completion of a successful calibration step, the ion descrip-
tors shown in Table 5 were updated to reflect the new mass calibration.
b. Relative Response Factor and Instrumental Sensitivities
As part of the initial and routine instrument performance
checks, calibration standards were analyzed and RRF values of the respective
analytes were compared to specific internal standards. The initial and
routine calibration criteria required that the precision (percent difference)
of the RRF measurements were within ±20% for the tetrabromo congeners and
within ±30% for the other compounds.
Sensitivity of the HRMS was documented through the responses
noted for the first calibration standard analyzed for each analysis day. The
method required the analysis of a low level standard (CSS) to document suffi-
cient instrumental response to support instrumental detection limits of
10 pg/yL for TBDD.
Routine checks on the instrumental sensitivity was achieved by
monitoring the response for the internal recovery standard (^C^-1,2,3,7,8,9-
HxCDD) from injection to injection and was documented in the MS log book.
2. QC Samples
Samples included for QC purposes are summarized in Table 6. Each of
the five sample batches included four quality control samples: a method
blank, an unspiked control lipid sample, and two spiked lipid samples. The
method blank was included to detect background contributions introduced by the
laboratory procedures. The unspiked and spiked controls were included as a
means of assessing method precision and bias (or accuracy).
Table 6. Quality Control Samples
Type Frequency
Method blank One per batch
Spiked control adipose tissue sample Two per batch (two different
spike levels)
Unspiked control adipose tissue sample One per batch
12
-------
a. Method Blanks
One method blank was generated with each batch of samples. A
method blank was generated by performing all steps detailed in the analytical
procedure using all reagents, standards, equipment, apparatus, glassware, and
solvents that were used for a sample analysis, but omitting the addition of
the adipose tissue. The method blank contained the same amounts of
i3C-labeled internal quantitation standards that were added to samples before
sample extraction.
b. Unspiked Control Samples
Control samples were prepared from a bulk sample of human
adipose tissue. This material was prepared by blending the tissue with
methylene chloride, drying the extract by eluting through anhydrous sodium
sulfate, and removing the methylene chloride using rotoevaporation at elevated
temperatures (80°C). The evaporation process was extended to ensure all
traces of the extraction solvent had been removed. The resulting oily matrix
(lipid) was subdivided into 10-g aliquots which were analyzed with each sample
batch.
c. Spiked Control Samples
Spiked lipid samples were prepared using a portion of the
homogenized control lipid. Sufficient spiked lipid matrix was prepared to
provide a minimum of two spiked samples per sample batch. A low and high
spike was prepared with each batch. The native spiking solution concentration
is presented in Table 7. The spiking solutions were checked for accuracy
prior to spiking the adipose composites with the native isomers. The results
of this spike check are shown in Table 8. The results in Table 8 demonstrate
that the spiking solution was prepared correctly. The differences in measured
concentration can be attributed to variance in the HRMS procedure, specif-
ically the variability in RRF values from initial calculation to the evaluat-
ing of the spike check.
The low spiked sample was prepared such that each aliquot was
fortified at concentrations equivalent to 25 pg/g for tetra- and penta-
brominated compounds and 62.5 pg/g for the hexabrominated compounds. The high
spikes were prepared such that the nominal concentrations of 50 pg/g for
tetra- and pentabrominated compounds and 125 pg/g of the hexabrominated com-
pounds were achieved.
d. Order of Analysis
Each of the QC samples and NHATS composite samples was assigned
to a specific analysis batch and a specific analysis order within a batch.
The method blank was the first sample analyzed within each batch. The lipid
unspiked control and spiked control samples were randomly assigned analysis
orders within the batch. The QC samples and NHATS composites were labeled
with barcodes such that the analyst was blinded to the exact nature of a
specific sample.
13
-------
Table 7. Spiking Solutions for
PBDDs and PBDFs
Concentration
Compound (pg/vL)
Internal Quantisation Standards
i3C12-2,3,7,8-TBDD 5
i3C12-2,3,7,8-TBDF 5
i3C12-l,2,3,7,8-PeBDF 5
Natives*1
2,3,7,8-TBDD 5
2,3,7,8-TBDF 5
1,2,3,7,8-PeBDD 5
1,2,3,7,8-PeBDF 5
1,2,3,4,7,8-HxBDD 12.5
1,2,3,4,7,8-HxBDF 12.5
aLow spiked samples were prepared by adding 50 yL
of the native solution to each 10-g control lipid
aliquot. High spiked samples were prepared by the
addition of 100 yL. All spike solutions were
prepared in isooctane.
14
-------
id
CO
•M
r~
3
CO
QJ
as
^£
u
QJ
r-
QJ
^£
•^
O.
CO
"O
(U
id
c
•^
£
i-
CQ
•
CO
(U
^™
X)
id
i—
ca
a.
ns
~y%.
i.
> *>*
O i-
(J QJ
QJ >
S- O
O
QJ QJ
cn s.
id
(^ l"^— '
>
.
$_
QJ
O
(J
Q)
Of.
^
*~°^
—1
3
cn
a.
^^
CM
^/
U
(U
.C
o
QJ
"5
to
m
u
QJ
O
U
QJ
0£
«
_1
3
^•s*
Q.
vH
j^j
u
QJ
^
U
QJ
J*
O.
to
$_
CO
5^
y_
QJ
O
U
QJ
Of.
&^
-J
3.
^s^
Cn
a.
l"~
(U
QJ —I
i— 3
^^
QJ cn
-id Q.
•f— <«MHI«
a.
^O
T3
C
3
0
a.
E
O
OO C\J LO CO
• • • •
^j-r-HT oooooouor>.
«— i t— i esj
^T CO CO f^ ^3 c^ ^^ ^^ ^j*
00 CO 00 T-I «— 1 •— 1 CT» CO CO
r-l i— 1 i— I
VO CO 00 O O O CM CO CO
CO OO CO «— 1 i— 1 CM Cft CO CT
>— 1 «— 1 «— 1
co CM ^* r^* oo o co ID co
^f ^^ ^f CM CM CO CM tO LO
CMCOO ooor^cocr*
oo oo cn T— 1 1— i o cn ot co
I— 1 »-l «— 1
f*H f~~ ( U) CO ^^ ^O *lf CO IO
If *f *t CMCMCMCMIOIO
ID CM U3 OOOOOO
CO CO OO *-t i-< O O CO P**
i—l i— 1 i— 1 i-H
CM »"H CO ^^t ^^ IO LO *""* ^T
^* ^f ^f CM CM CM CM LO ^T
LO LO
• •
f~\ t~°* f**i LO LO LO LO CM CM
ID LO ID CM CM CM CM VO VO
U_
Li_ a O
a a ca
ca ca ai
I— i— a. u. a LL. a
i i i LL. a a o o a
o o o a a ca m ca ca
nmn OQ CD QJ QJ X X
~ ~ ~ H-h-Q-CuIZ
CO
•o
1.
id
•o a»
«5
CO M-
0
C
O —1 •
•^* ^ ^J
+•> ^L
id o
•M «-H O
1— 1— <
•»-> TJ
C C M-
(O (t5 O
^3
cr • ai
i— r>. 3
id i—
c -o o
i- C >
QJ id
-M i—
c f-H id
i- C
CO 1-
QJ QJ l|_
^ 1—
•M -Q Id
id
«*- 1— O
—i c -o
a o QJ >t
i- -M C
J
>
84-> id o
3 J- ^
^™4 ^™* ^^
0 C X
cn co QJ
C 0 ^~
"O -O O i-
T3 J- O QJ
•o «» o
>> c id u
-Q id 3 O>
T3 CO C
O) 0 2
> TJ t- 0
QJ QJ .M i—
!c "oJi^ i
O J3 0
id id tn >,
1— i.
tn c (U QJ
Id 3 JC >
S I— 0
QJ U
C -C QJ
O -M • i-
4^ 1^. ^H f"
3 o cn
i— QJ i-
O — 1 «— Z
to a-g "~-
>)
QJ
o
u
2
01
cn
id
y.
QJ
^£
cno 1— II
T- Q)
.* ••o c
T- ^ J. C
o.r>. id a
;
5
CO -O i-
-0 C QJ
OJ c id M.
»-< -5
«*- >>
O l/> i- •»-
J
QJ QJ C
C i— > OJ
0 XJ 0 (J
i- id o s-
•u I— a» a
J
j^j ^^ -* ^_ C3L
i— c i— a>
id o id >
> 1- C T
O) •)-> L. +
J
3 QJ Id
QJ r— 4-> t-
x: o c QJ
1— CO 1- Q
id x>
**
15
-------
III. RESULTS
The analytical efforts conducted as part of this study included a
preliminary method evaluation as well as analysis of the 48 composite samples
and associated QC samples. This section presents the results of the prelimi-
nary method evaluation and summarizes data for the composite samples. Data
from the analysis of the quality control samples are presented in Section IV.
A. Preliminary Method Evaluation
Preliminary method evaluation studies focused on developing the
necessary HRGC/HRMS acquisition parameters, establishing linearity for the
calibration curves, determining limitations of the instrument sensitivity, and
conducting the analysis of a series of spiked samples that were taken through
the analytical procedures described in the experimental approach.
1. Evaluation of Instrument Variables and Establishing Calibration
Criteria
Initial investigations into the response relationship between the
chlorinated recovery standard (^C^-l.Z.S^.S.Q-HxCDD), the three brominated
IQS, and the six native brominated dioxins and furans were conducted by ana-
lyzing the brominated dioxin and furan calibration standards (Table 2) in
triplicate. Although all eight levels of the standards were analyzed,
response was only observed for the tetra- and pentabrominated dioxins and
furans at the CS7 level (2.5 pg) and higher. For the hexabrominated dioxin
and furan, response at 2.5 times noise was observed at the CS5 level (10 pg)
and above. The result of this effort is shown in Table 9.
The 2,3,7,8-TBDF was paired with the 13C-TBDF, the 2,3,7,8-TBDD was
paired with the i^C-TBDD, and the remaining unlabeled brominated dioxins and
furans were paired with the l3C-PeBDF for the determination of relative
response factors. The three internal quantitation standards were paired with
the recovery standard, ^C^-l^.S^.S^-HxCDD, for the calculation of rela-
tive response factors.
2. Recovery of Brominated Dioxins and Furans From Adipose Tissue—
Preliminary Evaluation Study
Six samples of the control adipose tissue (lipid) were spiked with
specified levels of brominated dioxins and furans by preparing three spike
levels in duplicate. These levels were 10, 25, and 50 pg/g for the tetra and
penta isomers; and 25, 62.5, and 125 pg/g for the hexa isomers. Two unspiked
samples were also included as well as a method blank.
The results of the extraction and analysis of the brominated species
from adipose tissue are given in Table 10. The average recoveries of the
internal standard quantitation standards ranged from 59% to 140%. These
recoveries were calculated versus the 13C-HxCDD recovery standard.
16
-------
Table 9. Mean RRFs for PBDDs and PBDFs From Initial
HRGC/HRMS Calibration
Compound
na Mean RRF RSD (%)
Calibration range
i3C-2,3,7,8-TBDF
i3C-2,3,7,8-TBDD
i3C-l,2,3,7,8-PeBDF
2,3,7,8-TBDF
2,3,7,8-TBDD
1,2,3,7,8-PeBDF
1,2,3,7,8-PeBDD
1,2,3,4,7,8-HxBDF
1,2,3,4,7,8-HxBDD
24
24
24
21
21
19
20
15
15
0.415
0.374
0.127
1.005
0.953
0.930
0.750
0.110
0.105
7.84
7.32
13.7
7.51
10.2
8.55
11.1
14.4
19.3
50
50
50
2.5-200
2.5-200
5-200
5-200
10-500
10-500
an =
Number of calibration standards for which all qualitative
criteria were met. A total of eight calibration standards
(CS1 to CSS) were analyzed in triplicate.
17
-------
4)
3
in
in
t-
4)
10
O
O.
<:
§
t_
u.
V)
c
a
i.
3
i^
c
ID
in
c
O
O
4)
a
c
I
m
,«_
o
L.
>
O
u
S
o
^~
""•
a
^~
•o
c
I
X
u-
CQ
X
m
4)
a.
^
CD
QJ
a.
^J
Q
e
s
«
u
o
Q
i-
i
u
0
^
g
i
U
o
in in
CM CM
V V
<*•"• ^^
. ii .
in in
CM CM
V V
QQ
1 ZZ 1
So
V V
*B^ "^
. ii ,
in in
V V
Q Q
1 ZZ 1
in in
V V
. ii ,
ID
in in
V V
. ii ,
O CM t^ O
in 10 r>» •—
o *~ oo 01
in CM to in
O — ^ Ol
- < x «
a. QQ
«c <
o o
— 8
& 2L
ID —
3<
in in
CM CM
V V
^^ *~t
s ii .
in in
CM CM
V V
s ii ,
00
V V
2 ii ,
o — — 2
2 -- §
o -- 2
O in r- CM
in to to r~
Soo — O
CM IO
O. Q. Q. O
4)
— L.
4>
> 4>
4> O)
— a
4, T? fe
J£ •» ^- >
a. O O
CO < <
« —
5tn^
in
CM O — CM ro
in
CM
V
in — •
CM O O O\
\o z to ^>
o
in i» 01 •—
CM CM CM •—
o
in ~* >.
TO O)O) L.
X XX 4)
O) O)O) >
Q. a a. o
4)
— L.
4) O)
« Vf fe
^ — CM >
Q. QQ
(O «
1«-
— ^ 0
^ *~ >
±V> -J
in
CM «»• r- CM
in
CN in r» o>
— fO 10 CM
O r- oo O
in in v —
O — CM O
in in in --
O to oo —
in in in •-
O 00 CM O
in » m •-
O in 01 O
in 10 1>- «»•
"
O «O «O CM
in ^* ^r o^
o in CM o
in in O) C
X XX 4>
O) O) O) >
a. an. o
4)
— L.
4>
> 4>
0 01
„, opo 2
4> 1 1 4>
Jf CM CM >
~ 1 1 «C
0. Q Q
4) —
£ ^ 4)
O*-— >
Z _!
4)
4)
V)
O
c
CO
^
c
a
in Q
01 CO
4) I-
c u.
41 O
ID h-
O. 1
CM
c •—
- 0
•£~~
E •
CD
c h-
O I
— 0
o—
4)
4- *
4) T3
•C 4-
ID
•O 'O
4) c
4- ID
ID 4-
4- >•
If) I.
LU 4)
O
• U
T3 4)
4) L.
U Q
4* §
4) X
•0 I
4- O
ID A
18
-------
The average recovery of the TBDD, TBDF, and PeBDF isomers ranged
from 105% to 110% with a relative standard deviation of less than 1% per
compound, not including sample AD-2-A. The average recovery of the native
pentabromodioxin over the medium and high spiked levels was 109% with a
relative standard deviation of 9%. The hexa isomers showed low recoveries at
all levels. It is expected that this is a function of their high molecular
weight, solubility, low volatility, and low RRF (lack of appropriate IQS).
In one sample spiked at the medium (25/62.5 pg/g) spike level,
sample AD-2-A, the recovery standard, the tetrabromodioxin and furan IQSs, and
the native tetrabromodioxins and furans were not observed. The reason for
this was unclear. The penta and hexa native isomers were observed, and
recoveries were calculated versus the pentabromofuran IQS.
B. Analysis Results for the FY1987 NHATS Composites
A total of 48 composite samples from the FY1987 NHATS specimens and
20 quality control samples were analyzed for PBDDs and PBDFs. The targeted
2,3,7,8-substituted PBDDs and PBDFs were not detected in any of the samples
except those prepared as spiked QC materials. The detection limits calculated
for the tetrabromo congeners in the samples ranged from 0.4 to 8.9 pg/g. The
detection limits for the higher brominated compounds were typically greater
than that observed for the tetrabrominated compounds. Table 11 provides a
summary of the detection limits determined for each of the targeted PBDDs and
PBDFs. Although the data are reported for selected isomers (which were the
only available standards), there were no indications of the presence of other
PBDDs or PBDFs of the same degree of bromination.
Although PBDDs and PBDFs were not detected, there were several
instances in which a response for both characteristic ions for 2,3,7,8-TBDF
were noted at greater than 2.5 times the observed background noise. In
several of these instances the ion ratios for the integrated peaks were within
the qualitative criteria specified in the QAPP for acceptance as a positive
identification. However, upon closer evaluation of the data, it was apparent
that in some cases, the integration for the characteristic ions was not
coincident, and in other cases the peak shapes indicated potential inter-
ference to the determination of the 2,3,7,8-TBDF. For these reasons these
responses were reported as not detected values, and the measured responses
were used to calculate the reported limit of detection.
Figure 1 provides an example of a possible 2,3,7,8-TBDF response
noted for a specific sample, NHATS Composite No. ACD8700265. In this sample,
three characteristic ions for TBDF have been plotted along with one charac-
teristic ion for the 13C12-2,3,7,8-TBDF internal quantitation standard. The
three characteristic ions for TBDF include m/z 482 and m/z 484 from the
molecular cluster, and m/z 375 which reflects the loss of COBr from the
molecular ion. The ion m/z 494 is characteristic of the 13C12-2,3,7,8-TBDF.
19
-------
a
Q
CQ
X
K
oo
CO
ic
0 co
. t-
oq CN oq 1-1
co c-^ co co
Q
CQ
K
00
n
ro
=3
OO
c
o
o
OJ
(U
o
O)
It)
Q
Q
CQ
co
D.
'«'
E-"
s
eo
co u>
^ fH O} t*
00 t> 00 CO
w
OJ
O
eo
co e<5 » 1-1 o os OJ «- t>
eo
*"H CO CO t*
oi w ^f cb
£3 TO. 23
»H »^\ "
Q
Q
CQ
o
s s a «.»
«>. CD « 00 ^
^ o o c> "-1
«H CO §§ rtI CO
«H »-i »H r-i
Q
CQ
t-
co"
CN o ^-i
Q
T3
Q
S
CO
t-
t»
«-H O C~ OO
00 >H CD "
O co eo
co o to
i—t i—i T~I
888
SP
-------
Q
Q
x
X
eo
Q
OQ
x
ffi
eo
S^H t~ ^< ^ —I
. «o . oo .
S
Q
CQ
eo
eo i
o «- eo co i- r- TH rH
2
"aS
£
P
i8iiii8Si
t-t^t^t~t~t~t~r-t~
GO 00 GO OO 00 00 00 00 OO
o o c^ d d d ^^ c^ c^
QoqQQooqo
21
-------
CO
I
00
CO
•»
CM
oo o\
in —-
vo CM ^
CM oo c T3
r~ N
oo ^«. m
C
o *
< £ «"
—• * m _
"*•>-«
1- P
O i. _
ro QJ 4^
^ 4-> o
«—i in > wm^
gf^ ::
-------
Although the ratios of m/z 482/484 by both peak area and peak height
meet the qualitative criteria (0.54 to 0.82) for Identification as the TBDF,
the peak shapes do not maximize at the same scan, the peaks are obviously not
integrated over an equivalent window, and there is a lack of response at m/z
375 to support confirmation. Therefore, 2,3,7,8-TBDF was determined to be not
detected in the sample, and the limit of detection was calculated at 3.8 pg/g.
Other samples for which this type of response was noted included:
MRI bar code NHATS code
16260 ACD8700069
16261 ACD8700461
16275 ACD8700416
16276 ACD8700014
16283 ACD8700309
16289 ACD8700167
16308 ACD8700121
16309 ACD8700265
16310 ACD8700372
16311 ACD8700050
Figures 2 through 5 provide examples of the total reconstructed ion chromato-
gram (RIC) and the ion plots from the analysis of tetra- through hexabromi-
nated dioxins and furans for a specific composite sample.
The data presented in these figures are typical of the responses
noted for the other composite adipose tissue samples. Although PBDDs and
PBDFs were not detected, several peaks that are shaded in Figures 3 through 5
warrant further consideration. These responses appear to arise from hexa-
bromo- (m/z 644, Figure 3), heptabromo- (m/z 722, Figure 4), and octabromo-
(m/z 802, Figure 5) diphenyl ethers. Fragment losses of two bromines from the
brominated diphenyl ethers would yield an ion species with exact masses
consistent with the corresponding brominated furans. In each case there is
also considerable response to an ion that corresponds to the fragment losses
of COBr from each of the potential brominated diphenyl ether (PBDPE)
responses.
Table 12 lists the estimated concentrations of the PBDPE in the
samples. These concentrations were determined by using the furan ion
responses and RRFs to calculate the DPE concentration when there was a coelut-
ing ion at 162 amu higher than the furan masses. These are estimates to one
significant figure only. Estimates are given for the composite samples only,
although a similar pattern of PBDPEs was observed in the control lipid and for
spikes and controls. No PBDPEs were observed in the method blanks.
23
-------
LD
^-
LD
C"
o
CM
CO
C-4
!£!
S
o
I.TI
© "» O ^
_l Utf O O
C-J OO CO
a ii
•i •— in
«^ ®
"^ S3
g
-------
• r~- co
CO CC< CO
CM coin
CM • •
$ z*
8 "-H
— — *•
0 ,
CO s
IT.' ^
O
H- CO
S CD _J
ir. = '
S
O Hi
\.r\ in ^
in es ^
TT — < H
COM*
-« —)
X CD
y. S -I
5 CO CD -
iX OO CO i
LD Q II
OD "^ 0^
8°
inz
3 «• •«:
S _J CD <•
Qi 5 • >
1^3 CD CM V CD
O O II CD <
H* •• C^ CD
CM CC O ^
2: — .Lncjo:
. CO CO
CO CO CO
r^ co in
^9- . •
^H CO CO
CO O"i
CM CO
CM -H
'
-*^
HI __"
^^^^^^^^™^™i
<•
' -C
-<=;
<
•^
"" C
3
• ID O
CO • •
ffTi *«"^ CO
u<
10 -H
r
" - ITJ
'
«
•
•
ii ii ii ii ii ii ii
TT CO CD CD — • in CO
• m ' • ^* • o»J • ^f • co • CD • *r
cnr- ooori CDCO CDCO r».en — CD «•«•
— • ro -co co««r OT v — in coco
5s UJ
IA »—
I/)
C
<=• i
CO CM 0 .
-s« ^g
-N 0?
fe •
° r-« .M
k o> c
• "si
£^£
£^ci
•^
t. i_pj i»»
H~ CO -*
^-«.l~ *
«- £ » £ S.
CO CO •—••?—,
-8S 8^1 •
-" gS^ct!
s'^a
<-go,dg
« ^
o\ t— *°
s • °*
«s ^ -^
r J200 * °
JO ^ UJ
•r-
U_
CD
l^p ^p
^5 •"
ijO OD
25
-------
26
-------
CD
CD
LO
=
-------
Table 12. Estimated Bromodiphenyl Ether
Concentrations (pg/g) in NHATS FY87
Composites
MRIID
16254
16255
16257
16258
16259
16260
16261
16262
16263
16264
16267
16268
16269
16270
16272
16274
16275
16276
16277
16279
16280
16281
16282
16283
16284
16286
16283
16299
16291
16298
16294
16296
16297
16299
16300
16301
16302
16304
16306
16308
16309
16310
16311
16313
16314
16315
16316
16317
Field ID
ACD8700247
ACD8700425
ACD8700023
ACD8700318
ACD8700112
ACD8700069
ACD8700461
ACD8700194
ACD8700390
ACD8700176
ACD8700256
ACD8700470
ACD8700283
ACD8700087
ACD8700489
ACD8700096
ACD8700416
ACD8700014
ACD8700274
ACD8700185
ACD8700292
ACD8700354
ACD8700434
ACD8700309
ACD8700238
ACD8700381
ACD8700210
ACD8700167
ACD8700078
ACD8700201
ACD8700130
ACD8700103
ACD8700158
ACD8700345
ACD8700149
ACD8700032
ACD8700041
ACD8700229
ACD8700363
ACD8700121
ACD8700265
ACD8700372
ACD8700050
ACD8700452
ACD8700327
ACD8700443
ACD8700336
ACD8700407
HxBDPE
200
ND (a)
700
300
8
200
20
NA(b)
200
400
ND
5
ND
300
700
500
600
4
2
300
200
7
10
ND
ND
300
ND
600
1000
100
900
200
20
ND
ND
10
ND
ND
9
30
500
ND
ND
ND
10
900
500
600
HpBDPE
200
100
200
200
200
90
200
NA
40
300
50
100
80
200
100
200
90
300
100
50
40
100
400
100
30
100
100
300
60
1
400
200
200
300
50
TO
100
3
200
200
2000
200
200
200
TO
30
400
100
OBDPE
ND
400
800
800
ND
200
100
NA
200
2000
ND
600
ND
600
400
400
100
8000
ND
ND
200
600
ND
ND
ND
400
400
3000
ND
ND
3000
400
ND
ND
300
ND
700
ND
100
ND
3000
ND
600
ND
ND
TO
2000
1000
(a) - ND- Not Detected.
(b) - NA- Data Not Available.
28
-------
The values presented in Table 12 are considered estimates and most likely
lower concentration estimates since: (1) the sample preparation procedures
were developed to remove gross interferences such as diphenyl ethers, and the
extent of sample cleanup is not known; and (2) the calculation of the concen-
trations were based on the recovery of the carbon-13 PBDFs which are expected
to recover at a different level. The effort to provide an estimate of the
concentration, however, provides some indication of the range of concentration
that these compounds are present.
Future sample analysis efforts for the determination of PBDPEs
should include experiments that exclude the carbon column cleanup from sample
preparation. This should provide a better indication of actual PBDPE concen-
trations in human tissues, especially if carbon-13-labeled PBDPEs are avail-
able as internal quantitation standards.
Although no PBDDs and PBDFs were detected in these analysis efforts,
further studies regarding background levels of these compounds warrant some
changes in the HRMS monitoring techniques. Specifically, these differences
are based on recent reports in the literature^'io that demonstrate that there
is considerable potential for the overlap of PBDPEs which have one bromine
more than PBDFs which elute within the same retention window. Since there are
potentially 209 PBDPE isomers, it may be possible that there is some overlap
of homologs. Although the COBr fragment provides valuable confirmational
information, its relative response to the molecular ion clusters is small and
may not be detected unless the compound is present at very high concentration
levels. Given these considerations, it may be more practical to monitor ions
which represent PBDPEs which are one and two bromines greater than the target
PBDFs. Having raised these considerations, however, the authors of this
report must stress that there was no evidence in the raw data to invalidate
the results reported.
IV. QUALITY CONTROL
This section provides a summary of the quality control data that
were generated as part of the analysis effort for the FY 1987 NHATS com-
posites. Summary data are provided for calibration efforts, internal
quantitation standard recoveries, control lipid analyses, and spiked sample
analyses.
A. Internal Quantitation Standard Recoveries
The recoveries of the IQS compounds are considered as a measure of
method performance from sample to sample.
In the total of the 68 samples, controls, spikes, and method blanks,
the i3C-2,3,7,8-TBDF IQS was out of the targeted data quality objective (DQO)
of 40% to 150% recovery in six samples. Recovery of the 13C-2,3,7,8-TBDD IQS
was out of the targeted DQO in 18 of the samples and the 13C-l,2,3,7,8-PeBDF
IQS was out of the targeted DQOs in 16 of the samples.
The recoveries of the i3C-TBDF and i3C-TBDD, which were outside the
DQOs, were always less than 40%. While the recoveries for the 13C12-PeBDF
29
-------
outside of the DQOs were both greater than 150% in some samples and less than
40% in others. These recoveries may be due in part to recovery through the
analytical procedure or as a result of pairing the IQS versus the recovery
standard.
The calculated recovery of the IQS for the brominated species were
affected by the responses observed for the recovery standard (RS, 1,2,3,7,8,9-
HxCDD). Responses for the recovery standard were often saturated in the adi-
pose tissue extracts although the daily standards consistently showed normal
response.
Previous method development work conducted to assess the recovery of
the brominated species from adipose tissue also showed a consistently high
response of the RS in the sample extract compared to the daily standard.
Reanalysis of these samples showed identical results. The cause of this
phenomenon was not determined. No reanalyses were conducted since there were
no apparent responses to the PBDDs and PBDFs. Samples for which recoveries
were less than 40% would have required recompositing and preparation of the
samples. Further method evaluation is necessary to address the difficulties
encountered in pairing of the IQS compounds versus a more appropriate recovery
standard.
Detailed plots of the recoveries of each internal quantisation stan-
dard versus an index number corresponding to the laboratory identification
number (Table 13) are presented as Figures 6 through 8.
B. Calibration
Although calibration curves were established prior to the method
evaluation study, it was necessary to conduct additional verification of the
RRF values prior to analysis of the composites.
The RRFs from these calibrations are given in Table 14 in the order
of analysis. Upon reviewing the data in Table 14 versus the initial evalua-
tion (Table 9), it is noted that there are considerable differences in the
mean RRF values for the ^C internal standards and the heptabromo compounds
between the two sets of calculations. However, the RRF values for the tetra-
and pentabrominated compounds are within ±10% of the two calibration events.
This difference is possibly due to the pairing of the standards versus the
appropriate recovery or internal quantitation standard. The variability in
the measured RRFs appears to be a function of the relative retention time
(RRT) of the standards versus the corresponding internal quantitation or
recovery standards. For example, PeBDD, PeBDF, HxBDD, and HxBDF are each
measured relative to the ^C^-PeBDF. The % RSD for the pentabrominated
compounds are less than that observed for the hexabrominated compounds. The
i3C12-PeBDF elutes coincidentally with the unlabeled PeBDF, and the % RSD of
the RRF is less than 10% across the calibration range. Likewise, the PeBDD
elutes very close to this IQS. However, both the HxBDD and HxBDF elute
several minutes from the IQS. This lower sensitivity of the HRMS analysis and
the differences in RRT lead to greater variability in the measured RRF values.
30
-------
Table 13. Index to Figures 6, 7, and 8 and
Order of Sample Analysis
Index
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
60
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
MRIID
16251
16252
16269
16310
16267
16313
16304
16283
16260
16256
16253
16262
16317
16280
16265
16279
16291
16271
16274
16315
16294
16273
16270
16272
16295
16266
16286
16263
16278
16309
16290
16287
16264
16275
16300
16311
16268
16288
16255
16306
16285
16254
16292
16299
16314
16259
16301
16297
16298
16296
16303
16276
16302
16281
16284
16306
16261
16318
16277
16282
16312
16308
16307
16257
16293
16316
16268
16289
Field ID
Method Blank
SI
ACD8700283
ACD8700372
ACD8700256
ACD8700452
ACD8700229
ACD8700309
ACD8700069
C
S2
ACD8700194
ACD8700407
ACD8700292
Method Blank
ACD8700185
ACD8700078
SI
ACD8700096
ACD8700443
ACD8700130
82
ACD8700087
ACD8700489
ACD8700103
C
ACD8700381
ACD8700390
Method Blank
ACD8700265
82
C
ACD8700176
ACD8700416
ACD8700149
ACD8700050
ACD8700470
ACD8700210
ACD8700425
ACD8700363
ffl
ACD8700247
Method Blank
ACD8700345
ACD6700327
ACD8700112
ACD8700032
ACD8700158
82
C
81
ACD8700014
ACD8700041
ACD8700354
ACD8700238
Method Blank
ACD8700461
81
ACD6700274
ACD8700434
82
ACD8700121
C
ACD8700023
ACD8700201
ACD6700336
ACDB700318
ACD6700167
Mass Spec. File
8862K30X2
8862K30X3
8862K30X4
8862K30X5
8862K30X6
8862K30X7
8862K30X8
8862K30X9
8862L01X2
8862L01X3
8862L01X4
8862L01X5
8862L01X6
8862L01X7
8862L01X8
8862L01X9
8862L01X10
8862L01X11
8862L01X12
8862L01X13
8862L01X14
8862L01X15
8862L02X2
8862L02X3
8862L02X4
8862L02X5
8862L02X6
8862L02X7
8862L02X8
8862L02X9
8862L02X10
8862L02X11
8862L02X12
8862L02X13
8862L05X2
8862L05X3
8862L05X4
8862L05X5
8862L05X6
8862L05X7
8862L05X8
8862L05X9
8862L05X10
8862L05X11
8862L05X12
8862L05X13
8862L05X14
8862L05X15
8862L06X2
8862L06X3
8862L06X4
8862L06X5
8862L06X6
8862L06X7
8862L06X8
8862L06X9
8862L06X10
8862L06X11
8862L06X12
8862L06X13
8862L06X14
8862L06X15
8862L06X16
8862L07X2
8862L07X3
8862L07X4
8862L07X5
8862L07X6
31
-------
O
U
0)
DC
U_
Q
ffi
00
I**
CO
CM
i
CM
O
CO
O
LO
in
(U
03
CO
O O
cvj C
•o X
C Ol
co -a
C
CO T—
(U
•!-> •
•r- CO
CO >
O
CL CO
oo
o o
co a.
(U
1- CO
i- i-H
o
> OJ
O r—
u .a
(U ro
a: I—
•£>
ai
$-
01
cc
#
32
-------
O
o
0)
DC
O
Q
ffi
h-
CO
I*-
00
CM
i
CM
O
CO
to
o>
CL
to
V)
o
t\l
T3 •
c o
« c
10 X
0) (U
-M -O
•r- C
«/> -I-
o
CL •
E bO
O >
O
to
(X) (!)
<: "a.
oo
en
$-
o
QJ l+_
^:
-t-> o>
u
§c
O)
i- S-
<*- (U
«*-
a. 01
a w
oo i
t— to
I CO
co o
•• $-
r^ u
«
co ic
«
CM l/>
I O)
ex -a
m o
~ J-
Q.
U-
O PO
i— l
(/)
(U O)
0) 03
> I—
O
U
(U
Qi
OJ
s_
33
-------
O
O
0>
cc
u_
Q
ffl
O)
o.::
Q. ^
N
S-
o
c 0)
H
s!
^
"-S
•^.b
"I
O
o
ai
oo
O)
34
-------
o o
CD eo co eo
a a s «> «> »
tN 00
CD CD
rH CO
- N ** rH W ^ S °°
tN O O ® 0} t* O O
od©iH^©d6c5
eo oo ic T}< o o m
rr\ OCDCDCOCDCDrHOO
g CO CN O rH O O 00
W OOOrHrHrHOOO
M oo o en rH t^ oo ^ in o
W6^ t-COlCrHOOCOOOOO
tN CO p rH rH p t> p p
U OOOrHrHrHOO'd
OTJ-TCCOOOOOSCO
COCOOrHlCCDt-CD
ro ?! CNCOOrHOOiCDOO
^ W OOOr-irHOOOO
*rs ^4t t*~ CD O rH CO CO CO ^ O
m C-rHCDOOC^rHQOCOt~
^ co tN o en en en c^ o o
O) VJ OOC3CJOOOOO
tn
c
S r^lft t-OlrHOSOO^CNCD
SS? t~rHCOICCDrHCDOOOO M
1/1 QH f) CN co p 05 en oo cp cs o
^u ocJooocJcJcio 'a
QJ CJ «
Oit~CDCOt>COtNOitN <»
TfrHiCCNCftlOtNCDT>« a>
tNtNOOOiOOE^Op •&
Hw OOOrHOOOOO g
OJ 2 ft
O' -H f"n t*~ en co t^ rH m rH c^ 5
rn OICt-COOOOOOrHrH g
J, BH u ^ N. 'P *"! °. °. ^ *"! "~1 r«
CCW OOOrHrHrHOCiO 2
^ erf *3
v^ ro CO 1C CD O) O) OO CD CO CO HH
c W rr ao-t-
O S rt tN OOrHrHrHC3e>C> QQ
J M CDtNCOCOTftTSinTfOJ „
COrHictNOStN-^OOO 2
C4 tN p p p p t> p p -g
ro " ^ OC>C>rHrHrHOC3C)
b
n
-« rHCDOJCOtN'^'^'CO1^1 'S5
,3, TfOcDOiCCOOOS"
**• y$ co tN[p IH\ en en t> p
OOOrHOOOOO C
ea
w
y» rHTfcnoocninrHoin
ro Tf, COOICCRtN'gJrHOOCD
tN tN tO OD tn O) t^ Cw C^ A
o'oo'oooo'oci *
^b
t-inmt-oo^tNin1^1 S.
OOTfC^t-C-rHC*"^ **
co rH o o oo en oo o _
oo'ooo'o'ooo £
cd
I
&
PQ
0)
00
b Q
rv, Q Q Q
Q§«pq
eo"g
tN ^
g9*
gg
||
eo" eo"
tN" tN~
6 6
CO CO
rH rH rH tN tN rH rH rH rH
K X
« W B
ft ^ 00 «
*l °°. ^ *"!
^j- O0_ OO^ t--| l>{ •«*" ^
r'^ t^ t^ CO CO CO CO
eo" co" tN" co" co" co"
CO
I
I
35
-------
Other calibration criteria that were considered were mass calibra-
tion of the HRMS instrumentation used in the demonstration of mass resolution
as greater than 3,000. These criteria were demonstrated on a daily basis
during sample analysis.
The daily analysis of a CSS calibration standard was conducted to
demonstrate the QA objectives of maintaining variability with ±20% of the
means for 2,3,7,8-TBDD and 2,3,7,8-TBDF and ±30% of the means for all other
compounds. A summary of calibration events conducted during the sample
analyses is given in Table 15. The analysis of the daily standards showed
that the RRFs of the internal quantitation standards and the hexabromo
compounds were difficult to keep within limits (±30%) which is associated with
the pairing of IQS versus RS. It is anticipated that the difference in reten-
tion time between the recovery standard (^C^-HxCDD) and the penta- and hexa-
substituted IQS standards contributes to greater variability in the measured
RRF values.
C. Spiked Internal QC Samples
Each sample batch contains two lipid samples that were fortified
with tetra- through hexabrominated PBDDs and PBDFs. These data were included
as a means to demonstrate the accuracy of the method for determining the
target compounds in adipose tissues.
The results of the analysis of the spiked control samples indicate
that the accuracy of the method met the targeted DQOs of 40% to 150% in all
but 8 out of 60 cases. These instances included only the hexa-substituted
isomers. These data were considered outside the DQOs because of incorrect ion
ratios (3 samples) and less than 40% but greater than 30% (3 samples). Two of
the data points could not be calculated because of nonrecovery. The specifics
of each case are summarized in Table 16. The results of the analysis of the
spiked control samples (spiked lipid) are given in Tables 17 through 21. The
native spike recoveries are graphically shown in Figures 9 and 10.
In general, the low recoveries experienced for the hexa-isomers may
be attributed to the limited solubility and low volatility of the hexa-
isomers. In addition, the RRFs for these compounds (calculated versus the
^C-PeBDF) are extremely small, less than 0.1, which adversely contributes to
the accuracy and precision measured for these higher brominated isomers.
D. Control QC Samples
The results of the analysis of the control samples (unspiked lipid)
are given in Table 22 which provides a summary of the results from the analy-
sis of the control lipid. PBDDs and PBDFs were not detected in any of the
control samples.
E. Method Blanks
The method blanks for batches 1 through 5 did not contain any
detectable PBDDs or PBDFs.
36
-------
>^
s-
IS
E
OO
4J
CU
LU
0
•^
id
i-
J3
^—
id
0
in
*""*
(U
r~*
^2
id
1—
c
o
•r—
U
id
cu
i-
u
Q)
^
$_
o
o
C/)
^_}
c
o>
p
o
°
en
•1~
O) «/» JC
i— >> o
tnJ (Q
C f>
(/) (d
0
t- TJ
-M i-
cd id
»- -o
J3 C
T- id
id co
cj
ttj
£
cu
id
a
cu
c
53
c
cu
u.
•
o •
^r ^^
1 (J>
CO
CO 1
"*" CO
LL. T-
o
CD O
0. CO
1 X
op^z
iv. co
co r-.
csT-sr"
.-H CO
i •
u csi
m •
.—i
LO
OO
CSJ
LO
t— 1
OO
CO
o
CO
Vs.
t— 1
CO
cy
X
o
CO
csi
i
-o
0
O)
CO
u_
1— 1
CSI
0
en
co
»— 1
CSI
00
00
8
^Xs
«— 1
Cy
X
o
CO
CSI
<£>
00
•
^5
^D
f
CO
•r-
O
O
CQ
X
z
i
00
i>T
rr1
CO
•
CSI
CSI
^^,
<— 1
LO
00
en
00
£
00
00
«— 1
o
1 —
CSI
CSJ
cy
X
o
CSJ
00
-^
$_
id
TJ
c
id
to
c
~1
l-
cu
a:
^
(£>
CO
•+.
w
"*~
U.
0
OQ
csi
csi
CSI
oo
00
T— 1
o
**^
CSI
cy
X
•— (
3
CSI
IO
CO
oo
I- 4->
O n-
•r—
>— 1 -M
CO i-
l£> 10
C
N O)
v^ 10
E E
CU =3
C E
3 -r-
-l-> X
cu id
Qi E
CO
Vs^
CSI
LO
OO
O
r-.
CO
o
•— i
oo
CO
CSI
o
^Ss,
CSI
00
X
CSI
O
esi
*O
00
•o
o
CD
CO
u_
Q£
CO
^*^
CSJ
in
oo
t ^
«— i
LO
«— 1
«— 1
OO
00
CSI
o
^^
esi
cy
X
CSI
3
CSJ
to
oo
00
•a
^
id
•o
c
«>« + (v. ^-t
CSJ »-H «3 OO
*o rv co + +
CO CO
CO CO LL. n- «r-
•i— t— CQ
CQ LL. a
u_ a cu a a
f^ f""fr CV CQ jQQ
CQ CQ 1 X X
t— 1— CO Z Z
1 1 • 1 1
co oo rv. co oo
r-« rv co rv. ps»
CO OO CSI ^ ^
csi csi t— i co co
i i i * •*
o <_> o csj esj
m m m •> •
CO
^s^
CSJ
CSJ
oo
t )
^.
CO
CSI
CSI
00
00
CSI
0
•*s^
CSJ
cy
X
CSI
o
csj
IO
oo
00
•
•a
i-
id
f"
id
co
C
^5
s-
cu
on
^
• CO
>S CO
O 4-
"* £
CO
•f 1 ^^
0
a OQ
0 CU
CQ Q.
X 1
z: co
1 •
OO iv.
rv co
TJ- CSJ
00 ^H
* 1
CM O
•« m
^f ^^
^*^ "^^
CO 00
If) CSI
oo oo
«_) O
•
-------
.-*
"S
•o
3
O
C
0
--'
in
T— 1
o>
^^
p—
C
O
•I1 '
•M
U
id
>
t-
4-*
U
CD
J_
J-
0
i/)
^j
c
C
p
o
CO
CO .C
i— >> CJ
E 'id id
id C £t
co id
C
0
+J $-
id id
i- -o
-a c
•f*" id
id M
ai
E
•r-
t—
O)
id
CO
U_
a:
a:
in
^s^.
^^
in
0
CM
8
00
00
to
o
CM
T— 1
CM
cr
X
to
3
CM
to
00
00
.
TJ
W
id
c
id
CO
c
=3
0)
0£.
in
CTl • •
• • r*H ^Q ^^
>s >e + m in
«o o co in
in to co + +
4> ^- *r-
CO CO
CO CO LL. •>— M—
•r- -^ a
m LL. a
u. a cu a a
a a a. ca CD
00 03 1 X X
(— i— oo a: :n
i i • i i
oo oo r-v oo oo
r-^ r^ co r^ r^
CO CO CM *3" ^"
CM CM «— 1 CO CO
III * *
CJ CJ CJ CM CM
*H FH IM t>_4 i—{
in
^^^
^f
CM
CJ
en
CM
t— 1
CM
00
00
to
O
CM
CO
cr
X
to
o
CM
to
CO
00
^^ ^^ •
in CM >e
^" *r co
+ + CO
CO CO
•i- 1- CO
•f—
Li_ 0
o o o
CD CD Q
1— 1— CO
1 1 X
oo co 3:
•t •* 1
^* r*^ oo
CO CO ^
CM CM ^~
•— 1 •— 1 CO
1 1 •
CJ CJ CM
^H PH i-^
in
>^
^^
CM
CO
0>
0
C\J
CM
OO
CO
to
0
CM
i— 1
^.
cr
X
to
0
CM
to
CO
CO
•
^S"5
IO
CO
1
00
•f—
u_
0
CO
X
3:
1
CO
rv.
^r
co
«
CM
1—4
in
in
CO
in
CM
0
CO
oo
^•^
0
CM
T-4
CM
cr
X
r-.
3
CM
to
§8
>4
CO
• • to
>« ^ +
co oo
co co co
+ + i-
CO CO Lu
•r-1-0
m
LL. O (U
a a a_
OO CO 1
1— h- 00
1 1 •>
oo oo r^
r*«* r**» co
CO CO CM
CM CM i— 1
\ 1 1
CJ CJ O
l-t *H *H
in
CM
CO
co
•^
S
oo
00
fn^
0
CM
i-H
CO
cr
X
o
CM
to
CO
00
38
-------
Table 16. Spike Recovery Summary for Samples Outside DQOs
Sample
Compound
Recovery
Explanation
S2—batch 2a 1,2.3,4.7.8-HxBDF
SI—batch 3 1,2,3,4,7,8-HxBDF
S2—batch 3 1,2,3,4,7,8-HxBDF
SI—batch 4 1,2,3,4,7,8-HxBDF
1,2,3,4,7,8-HxBDD
SI—batch 5 1,2,3,4,7,8-HxBDF
1,2,3,4,7,8-HxBDD
Not calculated
Not calculated
Not calculated
34%
30%
S2—batch 4 1,2,3,4,7,8-HxBDD 39%
Not detected
Not detected
Ion ratio outside criteria
Ion ratio outside criteria
Ion ratio outside criteria
Low recovery
Low recovery
Low recovery
aSl = low level spike.
•)S2 = high level spike.
39
-------
OD
CtL
-o
o>
o.
oo
o
id
CO
fO
S3
rH
02
riS
a
03
„
at
rH
03
o
•fe
0
o
% Recovery
-a
a
IS
i^S
-a
c
o
EL,
-*j
%
0
1
•a
c
o
l*<
-4J
O
1
1
"c
o
EL,
"c
fl
O
1
•e
i
e
o
U
co CN t* co ts_ ^J4
o o o rH g: co
rH rH rH rH °* rH
^ § 00 «O rH Cfl
ggggss
10 10 10 10 «? "?
c« w N N g g
^0 IO O -^" CS| rH t-
g> rH rH CO 00 1O OJ rH 00 3J
J§ O ^ 10 tO >O
o ri ri- S. & ^
X Q Q g Q Q
C *^ ^ Z *^ "^
£
re
o
re
o
o
o
cu
(U
-l->
o
o
o
o
o
o
cu
cu
cu
cu
a.
c
-a
c
3
O
Q.
oo
s-
cu
o
U
cu
a:
40
-------
I/I
£f
3
OJ
oi
01
t
(O
T3
d>
•r-
Q.
(X)
CsJ
^
'ft
l-H
"«
1
1
01
o
'bfi
M
51
Ib
o.
v~'
c
1
1
Q
1
"O
c
1
"c
Q
1
a
a>
s
no
c
I
^j
§
o
1
•o
c
§
ft
6
o
0
8SIS3S
iH O ^ 00 C- «O
O *H fm* d O3 CO
rH »H "^ i-H »H r-l
0 0 0 0 IS S
10 1O 10 «D ^ £5
eg N d cq g g
^ S
,bp o oo t^ eg oi co
gl OO CO CO rjJ S CO
31 Tf ws 10 10 Q r-
-to N •* e«i o 1-1 o
M id t> •*' ci oi 10
31 N d CN CO C- 00
P 00 C- '-«-->£?
eo 10 N. t* o «
o o C ^ ^ S
^^ C*^ ^J ^* *7 1^
S5 ^ ^ Z
K^ Q
fa Q Q Q
[id C^ MJ fc! b3 fc2
^^ ^^ ^^ ^^ >T5 fc"H
n n £ £ oo oo
t-" E-" oo oo t- t-
00 00 t^ t^ ^J* ^S*
t^ t^ CO CO CO CO
co eo «N w «N e»
(N 0>l i-l rH r-l tH
S-
C
c
*~
>
o
S-
3i "c
0 O
t— O
'o c
Qfa 10 -O
res c
QQ ^3
K -a o
JT< QJ tl—
00 ^~^
t*. , to
£ O) C
« > •<-
f 0
4J
w ro c
<-| 3
•5 o >
')
3
J
J
S
41
-------
VI
J__>
^^
3
C/>
O)
o:
O)
r—
Q.
rd
OJ
•r-
Q.
C/l
CO
-C
U
-U
id
OQ
•
i— I
QJ
J3
rO
1—
03
rH
W
0)
>
8
*
% Recovery |
i
ex
'So
Qu
Q.
TJ
c
3
1
4
T3
C
^
£
"c
o
4
5b
^Ja
a
>**'
"O
c
o
1
0
4
TJ
1
Q.
1
SSS35§
I9»>.
o o o o S K
10 10 10 »o 25 £3
T "* " "
Ifi kO
IO IO tf3 1C ^ Z;
W CSl W d g g
/^^ oo
^ rH | CO 0 (N S (N
O, ^ 10 10 10 Q i-l
2;
j—k. i^
5> iH O Oi C oi X "5
31 01 N r-l rH Q fM
fc
— o QS *^^ s~\ s*+ \ ,***.
S 25 w p o t*
o o S °^ SJ C-
gggigg
N^ ^|^
fa O Q Q
fo Q Q Q PQ PQ
g Q PQ PQ jg jg
W CQ ft, OH ao CO
t^ C"^ CO 00 t^ ^^
t* t^ CO CO CO CO
CO CO O4 O4 O1 OQ
^
•a
K
00
§
rH
V,
.0
**N
not met
e
it
s
p£S
"o,
to
c^
03
"B
OB
rH
CO
flj
c
•FH
00
•c
o
•J3
g
O
•P*
1
1
3
o
o
to
o
to
ro
•O
a>
4->
ro
3
U
ro
O
to
ro
S-
O)
o
o
QJ
to
ro
O
ro
3
O
O
ro
-o
O
-M
o
O
C
O
o
c
•r—
T3
O
C
3
O
(U
O)
^~*
(I)
O)
Q.
to
C
3
O
3
O
(U
o
o
Ol
o:
CL
OO
42
-------
CO
1 *
3
eu
oc
(L>
Q.
E
(0
OJ
Q.
00
^r
jz
u
.M
(O
CD
O
CM
Q}
^^
O3
t—
CM
CO
rH
CO
f^£
'S.
CO
>
o
<&
*
0>
o
o
£
,-^
a
"5b
ex
^-^
13
C
g
0
Amount F
T3
Amount Fov
^^ ^
•S"
S
*o
c
g
£
-w
C
9
o
1
T3
| Compoun
^*?
O3 »** GO GO — — ™
3^ W3 ^* 3£ QO ^^
2 Oi Ij 2 ^* Oi
CO
^* ^^ C^ ^*™ *S ^5
^ ^ CO CO
§ g § g S S
IO IO
CM CM CM N g g
/•** /**\
t** CQ
^0 CO •* rH rH Cft Cft
M C> t- Oi ^< S S
3 »"»»gg
/_^ C4 U3
-$*> ^< OQ rH OO rH 00
S1 CO l> CO
*0
S-
10 -4->
S c
O 0
i— O
'o c
<4- •!-
to T3
>, 03 C
* "O O
o g "-
4) 03 4-^
>• r— C
»Q 3 3
goo
S 03 «C
3 ^
2 * o>
•* s ^
«J- -r-
0 Qj
03 >, tO
1 1 =
« 0
•S, 0 TD
•S 2! i
W 13
•^ -« £
2 <^ -i->
5, 03 c
^ 3
5 o
" [ \ ^<"
qpj **^ "i*.
1 1
to o
00 U '
aj *°
"^ T3 <
1 1
M (U
3 S S
•s
cS ^
• +->
? §L
*+^ ^~
Q.
3
D
H
<
CD
Ol
O)
•r—
Q.
CO
1
>»
1J
^
D
J
U
1C
-a
43
-------
01
4->
5!
OJ
a:
OJ
T—
a.
.C
o
00
<— 1
CM
OJ
1—
£
1— 1
CO
"3
4)
'a
CO
rH
0)
J
•*' 10 co ei X X
31 csi w « co Q Q
os t* *•? *~* *~" * '•*'
IO t* °- ?O OS CO
o o S {S 21 Cl
^** N""' ^^ /"^ /^
P^l f^ f^ F^ p^
. ^^ ^» ^^ ^^ g^
bQl§gg
pH pH 00 OQ ^^ ^^
ao oo t- t- •*< ^<
t* t*- CO CO CO CO
OO CO S c
000
S ^l °
*" o c
M$ 'i >^
g M X,
rH , 10
S *-
o< ai c
5 0 "~
CO O T3
tJ OJ C
1 *" °
-S &S M-
S) w +J
S a
•a s- i
3 3
S 0 i
Q •> "° c
£ ° <•
g -= a
^ •£ Q
o OJ
oj s: ^
* z
J~. +>
3 £
%
D
D
H
<
"flj
OJ
OJ
•r-
Q.
CO
')
3
J
;
8
44
-------
RECOVERIES
Low Spike Level
140 •
120 •
100 -
RECOVERY 80
60-
40-
20
0 •
TBDF
o
•
o
TBDD PeBDF PeBDD
ISOMER
HxBDF HxBDD
Average ° Low Range
High Range
25 ppb Spike Level for TBDF/D and PeBDF/D
50 ppb Spike Level for HxBDF/D
Figure 9. PBDD/PBDF recoveries from quality control samples
(low spike level).
45
-------
RECOVERIES
High Spike Level
140r
120 -
100 •
RECOVERY 80
60-
40-
20-
0 -
•
o
TBDF TBDD PeBDF PeBDD HxBDF HxBDD
ISOMER
Average ° Low Range • High Range
50 ppb Spike Level for TBDF/D and PeBDF/D
125 ppb Spike Level for HxBDF/D
Figure 10. PBDD/PBDF recoveries from quality control samples
(high spike level).
46
-------
CO
O)
Q.
OO
CO IO
d o
s
IQ CD IO
Q Q Q Q Q
co
(M
B&4 C3
Q Q Q
g 0, PL, oo S
C^ 00 OO t- C~
C^ CO CO CO CO
Cfl i-l i-H rH i-(
47
-------
F. Tridecane Blanks
Tridecane blanks were analyzed daily to confirm that carryover from
the injection of standards was not a problem and to demonstrate a clean ana-
lytical system prior to the analysis of samples. These analyses demonstrated
no response to the PBDDs, PBDFs, or the internal quantitation standards.
V. ASSESSMENT OF METHOD PERFORMANCE
Control lipid and spiked lipid samples were included in each of the
five batches to assess the method accuracy and precision. This assessment has
been conducted through plots of measured vs. spiked concentrations, calcula-
tion of descriptive statistics, and regression analysis.
A. Assignment of Data Values
For each target chemical, 15 data points were generated. These data
points were obtained from the analysis of one control and two spiked lipid QC
samples within each of the five sample batches. The analytical results were
treated as follows in subsequent computations.
1. For observations where all qualitative identification criteria
were met and the response exceeded the limit of detection, the
reported concentration was used as the measured concentration
(i.e., positive quantifiable and trace values).
2. For data points where the response was below the limit of
detection (i.e., not detected values), one-half of the reported
limit of detection was used as the value in the regression
analysis.
3. For observations where the response exceeded the limit of
detection but all the qualitative identification criteria were
not met, the data points were treated as nonresponses and were
not included in the analysis. (For example, ion ratio criteria
not met.) This occurred in three cases for hexabromo furan—
once at the low spike level and twice at the high spike level.
B. Descriptive Statistics
Descriptive statistics for the six PBDD and PBDF compounds are pre-
sented in Tables 23 and 24, showing control and spiked lipid sample results,
respectively. No PBDDs or PBDFs were detected in the control lipid sample.
Table 23 presents the average limit of detection, the standard deviation of
the limits of detection, and the relative standard deviation based on the five
control lipid samples. The relative standard deviation, calculated as 100 x
standard deviation/mean is a measure of the variability of the five measure-
ments around their mean. The average detection limits increase as the degree
of bromination increases. The same pattern holds for the standard deviations;
however, it is less pronounced when comparing the relative standard devia-
tions. This general pattern is reflective of the relative sensitivities of
HRMS to each degree of bromination.
48
-------
Table 23. Descriptive Statistics for Control Lipid Samples
Compound
2378-TBDF
2378-TBDD
Number
of Samples
5
5
Mean LOD
(pg/g)
0.600
0.888
Standard
Deviation
of LODs
(pg/g)
0.184
0.325
Relative
Standard
Deviation
(%)
30.7
36.6
"•
12378-PeBDF
12378-PeBDD
,
123478-HxBDF
123478-HxBDD
5
5
....'.. . . . ;• ?
5
5
3.84
5.76
2.25
4.24
58.5
73.6
....? .T:..': ^. *... ''.....'. "...I r.....* ?:•... '?... '. . . ^
21.7
9.84
16.5
5.28
75.8
53.6
49
-------
Table 24. Descriptive Statistics for Spiked Lipid QC Samples
Compound
Number
of Samples
Spiked
Concentration
(pg/g)
Mean of
Measured
Concentrations
(pg/g)
Relative
Error *1
(%)
Standard
Deviation of
Measured
Concentrations
(pg/g)
Relative
Standard
Deviation
(%)
Low Spike Samples
2378-TBDF
2378-TBDD
12378-PeBDF
12378-PeBDD
5
5
5
5
25
25
25
25
25.1
26.7
21.6
24.4
0.4
6.8
-13.6
-2.4
1.67
1.24
2.77
5.05
6.7
4.6
12.8
20.7
••.•.•".•:.•••:."• • : •' J *. .• • -'i:-* '"' ' . : ,'S r V" ••. : <: x .%.•>»<; v .w w ,y. ;..../ :-: . ..••:¥:'••••' ".:./. ./>y v.-. •'•; .
123478-HxBDF*2
123478-HxBDD
4
5
62.5
62.5
38.6
36.5
-38.2
-41.6
28.6
30.2
74.1
82.7
High Spike Samples
2378-TBDF
2378-TBDD
12378-PeBDF
12378-PeBDD
123478-HxBDF*3
123478-HxBDD
5
5
5
5
3
5
50
50
50
50
125
125
49.5
51.2
53.0
55.7
93.0
107
-1.0
2.4
6.0
11.4
-25.6
-14.4
1.63
2.37
4.15
2.87
31.0
45.8
3.3
4.6
7.8
5.2
33.3
42.8
*1: Relative Error (%) = 100*(mean of measured concentrations - spike level)/spike level
*2: Ion ratio identification criteria not met for one sample; data point deleted
*3: Ion ratio identification criteria not met for two samples; data points deleted
50
-------
Table 24 presents these same statistics for the measured concentra-
tions at the low and high spike levels. In addition, the average relative
error, calculated as 100 x (mean of measured concentration-spike level)/spike
level, is shown here as a measure of relative accuracy. Measured tetra and
penta compound concentrations are relatively accurate as shown by the relative
errors which vary between -13.6% and 11.4%. On the other hand, the concentra-
tions for the hexa compounds exhibit a much larger inaccuracy with relative
errors ranging between -41.6% and -14.4%. The concentrations of these com-
pounds are consistently underestimated. As described previously in the
results section, the data for the hexabromo compounds are affected by the lack
of a corresponding carbon-13 labeled internal quantitation standard.
Similar to the pattern seen for control samples, the standard
deviations for the spiked samples tend to increase with increasing degree of
bromination. This is reflected both in the standard deviations and the rela-
tive standard deviations.
Using the relative standard deviation as a measure of imprecision at
a particular control or spike level, Table 24 shows that the precision is the
same (tetrabromo dioxins) or better for the higher spike levels than for the
lower levels for all compounds. However, based on only two spike levels,
overall conclusions as to an increasing trend in precision with increasing
concentration cannot be drawn. Overall, the precision as measured in the
spiked samples for the tetra and penta compounds is considerably better than
that for the hexa compounds.
C. Regression Analysis
Regression analysis was carried out to assess the overall accuracy
and precision of the analytical method. This analysis was performed for each
compound separately. One-half the detection limit was used for all concen-
trations below the detection limit. Each regression analysis provides an
estimate of the slope and intercept of the line that best fits the 15 data
points.
The slope of the straight line can be interpreted as a measure of
the accuracy of the method, simply by multiplying the slope by 100. If the
measurements were accurate, the line would have a slope of one. Thus, a
statistical comparison of the slope with 1 provides a test as to whether the
accuracy is significantly different from 100%. An estimate of the bias of the
method can be obtained by subtracting 100% from the accuracy.
The intercept of the line with the vertical axis provides an
estimate of the potential background contribution by the lipid samples. A
statistical comparison of this intercept to zero provides a test as to whether
the background is statistically negligible.
In addition to the slope and intercept, two basic statistics are
relevant in describing the scatter of the data points around the fitted
line. One is the proportion of variability in the data explained by the
line. This is the square of the correlation between the predicted and
observed measurements. The second statistic is the mean squared error (MSE),
51
-------
an estimate of the overall performance of the method. Mathematically, this
statistic is the average squared difference between measured and "true" con-
centrations. By subtraction, one can obtain an estimate of the overall method
precision, or regression standard deviation, as the square root of [MSE -
(Slope-1)*].
Figures 11 through 16 depict the regression lines and their 95%
confidence limits as fitted to the data by the least squares method.
Figures 11 through 16 also show 95% confidence limits for the prediction of
concentration based on the regression results. These confidence bands are
always wider than those for the mean found concentrations. Table 25 presents
the regression analysis results for the six compounds. The performance
parameters of the method for each compound are discussed below.
1. Overall Fit
The proportion of variability (third column in Table 25) in the data
that can be explained by the estimated linear relationship between measured
and spiked concentrations is high for tetra (slightly above 99%) and penta-
bromo (slightly above 96%) compounds. For the hexabromo compounds, a large
proportion of the variation remains unexplained, making the line in each case
a poor predictor of measured from spiked concentrations.
2. Accuracy
The estimated slope for each regression line is shown in column 5 of
Table 25. Multiplied by 100, these slopes provide estimates of the method
accuracy for each compound. The accuracy is close to 100% for tetras and
pentas, and for the hexa dioxin. In fact, as shown in columns 6 and 7, these
figures are not statistically different from 100% at the 5% significance
level. It should be noted that the confidence interval for the accuracy for
the hexa dioxin is wide (47% to 116%) due to the relatively high variability
in the measured concentrations (see also Figure 16). On the other hand, the
estimates for accuracy for tetras and pentas have narrow confidence intervals
associated with them. This is also reflected in Figures 11 through 14 with
tetras showing by far the best results overall.
3. Precision
As discussed earlier, the overall method precision was estimated by
the regression standard deviation. These results, in pg/g, are shown in
column 4 of Table 25. It is clear from these figures that the overall pre-
cision decreases with increasing degree of bromination, with an excellent
precision (1.30 to 4.22 pg/g) for tetras and pentas, but much poorer precision
(22.6 to 32.0 pg/g) for hexas.
52
-------
o
CD
CO
•I
r^
CM
O>
o.
§§
_ O
C
O
£ 8
0)
o
C
o
o
0)
I
u>
cu
M-
(U
-------
o
CO
a
o
en
CO
•«
r^
•«
CO
«*
CSJ
o o
co
o
II
a.
»- a>
o u
>*- 1-
id i-
(U
•«-> •>
C OsJ
i- O
a>
u
c
o
o
V
o
C II
a;
•o a)
•a 3
c c
ta
o
ts
> o
Jt ^- 4->
a- >.
U) 4?
at
c at
O <4-
T-
•o
• o
CM u
(6/6d)
punoj
54
-------
(6/6d)
punoj
55
-------
o
to
c
o
a
a
CD
cu
a.
GO
r-l Q.
C VO
O CM
IO II
T3
o- cu
u
S- J-
o a>
««- -t->
«/» n-
*« >>
i- •
cu «o
•(-> O
at ii
o
C 0)
cu a.
•o o
c
0)
u
c
o
O "2
O •
U J-
O)
>« JD
IT) E
CT> 3
id u
o
CM
Q. -r-
>> o
c i-
i
cu ^H
j-
C7>
-------
o
CM
CO
X
3C
00
ro
»
CSJ
en
co
O
O
'00
o
00
a,
O
(O
_ o
0>
o
o «
o w
Q> 1
.-* *.
Q. ^
U) •£
o
o
CM
(U .
3
O>
(B/6d)
punoj
57
-------
o
CM
O
O
O
CO
a
a
CO
x
3C
I
CO
• o>
co *-^
• O»
CM n.
»
t— ( CO
10
c •
O»-i
10
O -(->
<4- C
i- in
ii
c
o
o
CO
O T-
a> ex
•o o
Q)
O
C
o
o
jg
o •
O L.
a>
JQ
E
_ o
_ o
at
TJ
c x:
o i-
•r—
-------
in
c
o
d)
(J
c
o
•o
Ol
Q.
OO
"O
O)
IO
(O
(U
c
o
(/I
9>
4-*
1
^
|
53
U
0) o
= 55
5 "5
J
0
35
H
CO
•p.
C «
nt 2s
5 rt
£ w
UJ
c S
0°
B *^
C* CO
t'C
CO
a- o
b_
E
Z
•a
E
S
^
P
3 TO
&1
Q.
^
!~.
a5 1?
* -S;
3
If
a)"""
^
E
3
D
*•*
E
3
i
5
2
co
1
UJ
"3
1
5s
OJ s!L
C CD
'5 c
0.13
ui£
M
0)
Q.
E
CO
"o
CO O
in T-
T^ cvi
cn N.
{5 S
d d
i i
CO CO
CO N
d d
CM CO
q q
CO CO
cn cn
d d
CO CM
§ 2
d
o in
co in
^— ^
co in
cn cn
cn o
in in
u_ O
Q 0
m m
^ ^
CO CO
CM CM
*
••?:
>:::';
'••?.
.If
:%
:l:
i
f::'
rf;|
i
'"'•
:V
cn co
in cn
CO ^
CO CO
to TJ-
co c\i
i i
O CO
CM CM
CO !-•
o o
1- t
cn cn
d d
CM CO
q q
cn CM
q CM
^ ^f
in in
CO CO
cn o>
in in
PeBDF
PeBDD
co co
N N
CO CO
CM CM
^ T—
•.'.
.;
^ 1^
CO CO
CM CM
CO h~;
eo cn
i- CM
I 1
N> *~
cn co
CM T-
01 ,_;
d
S 1
d d
co in
CO CO
d d
CD O
CM CM
CM CO
in in
N CO
(D CD
CM in
CM
LL 0
Q O
m m
X X
X X
I i
CO GO
N N
CO CO
CM CM
deleted
a
5
CO
_»
sf
S 8
'5 CD
*l
•2 t
% s
5 5
S
M
II
'•s3 c
0) ^
oi
12
i- CM
59
g g
(O a)
8.1
^'t
a> =
fi S
e3 .co
s S
•B <—
S i
'•o .w
•— >»
S8
i§
2 eo
jgj
I CD"
Z £
10
8-
II
8 §
CO
2
o
_w
5
(O
o
o E
= o
co i:
kg
I o3
2 ^
c -a
« I
(1) W
1§
'"D _CO
°> 5
.E 2
•O O)
fl
.E A
E «
§€.
flj CD
in
. cd
-& w Q) o
g J= g t
-------
4. Background
Estimates of background levels for PBDD and PBDF are provided by the
intercept of each regression line and are shown in column 8 of Table 25.
Lower and upper 95% confidence limits to the background are listed in the last
two columns of Table 25. All six confidence intervals include 0, demonstrat-
ing that all the estimated backgrounds are statistically negligible. The
width of the intervals is a measure of the precision of the background esti-
mates. The width of these intervals increases drastically from the tetras and
pentas (less than 7.44 pg/g) to the hexas (above 42.0 pg/g).
D. Conclusions
On the basis of the statistical treatment of the quality control
data, one can conclude that except for 1,2,3,4,7,8-HxBDF, the analytical
method is unbiased in measuring sample concentrations. The bias, though not
statistically significant, is larger for the hexa dioxin than for tetras and
pentas. For the hexa furan, the method has a significant bias due to either
the lack of the appropriate internal quantitation standards, the difference in
the HRMS sensitivities to the higher brominated compounds, or both. Figure 17
provides a summary of the percent bias (calculated as 100 x [slope - 1]) for
each of the target analytes, along with the upper and lower 95% confidence
intervals. Overall, the method provides very precise results for tetras and
pentas, while providing poor precision for the hexa concentrations. In all
cases, no significant background contribution from the control lipid samples
was found. In addition, limit of detections increase with increasing degree
of bromination.
60
-------
fl
•8
trH
1
Q
*
10
o»
-a
M
1
I
• D —
^ i — i
^ LJ
•-C
A.
^r
I 1
D, . ., •
•
• PI •
j_j ^
•
«_^^H
•
"*
PI-H
_JBI
1 1 1 1 1 1
O
Q
m
*
* u
" 1
•* o
CO °
CM 42
u_
a
CO
u. a.
Q _
m ^
I *
1 J-^
00
r>-
^~
CO
CM
Q
O
m
9
Q.
CO
•^
r^
CO
CM ^
*- P!
o M
LJ
m
o
a.
i
CO
r*-
co
CM
O
Q
1
Hi
^
•r-
•o
c
J_
o
u-
^
4->
•^
E
•f—
OJ
u
OJ
-o
«^-
1^
c
o
o
>?
ir>
fj\
U1
T3
C
(0
C/)
CO (0
i— t-
A
fx. _» i
!-*• ^j
CO C
CM OJ
U
OJ
a.
* ^
O -H
CO
K Jj
1 ^
00 3
•^ fT\
r^ W
CO "-
CM Li.
OOoOOOOOOO
CXI T- T- CM W *fr IO
-------
VI. REFERENCES
1.
2.
3.
4.
U.S. Environmental Protection Agency. 1986. Broad scan analysis of the
FY82 National Human Adipose Tissue Survey specimens, Vol. II: Volatile
organic compounds. EPA-560/5-86-036.
U.S. Environmental Protection Agency. 1986. Broad scan analysis of the
FY82 National Human Adipose Tissue Survey specimens, Vol. Ill: Semi-
volatile organic compounds. EPA-560/5-86-037.
U.S. Environmental Protection Agency. 1986. Broad scan analysis of the
FY82 National Human Adipose Tissue Survey specimens, Vol. IV: Poly-
chlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans
(PCDFs). EPA-560/5-86-038.
U.S. Environmental Protection Agency. 1986.
FY82 National Human Adipose Tissue Survey
elements. EPA-560/5-86-039.
Broad scan analysis of the
specimens, Vol. V: Trace
5. Leczynski, B, Collins, T, Sasso, N, Clark, J. 1988. FY87 NHATS
composite design. Prepared for EPA Exposure Evaluation Division, Office
of Toxic Substances. EPA Contract No. 68-02-4294.
6. Cramer, P, Stanley, J. 1988. Analysis of adipose tissue for dioxins and
furans: Adipose tissue sample compositing. Prepared for EPA Exposure
Evaluation Division, Office of Toxic Substances. EPA Contract No.
68-02-4252.
7. Cramer, P, Ayling, R, Stanley, J. 1989. Determination of PCDDs and
PCDFs in human adipose tissue: data report batches 1 and 2, Revi-
sion 2. March 7, 1989.
8. Cramer, P, Ayling, R, Stanley, J. 1989. Determination of PCDDs and
PCDFs in human adipose tissue: data report batches 3, 4, and 5, Revi-
sion 1. March 7, 1989.
9. Donnelly, J R, Munslow, W D, Vonnahme, et al. 1989. The chemistry and
mass spectrometry of brominated dibenzo-p-dioxins and dibenzofurans.
Biomedical and Environmental Mass Spectrometry, 14, 465-474.
10. Donnelly, J R, Grange, A H, Nunn, N J, Sovocool, 6 W, Brumley, W C,
Mitchum, R K. 1989. Analysis of thermoplastic resins for brominated
dibenzofurans. Biomedical Environmental Mass Spectrometry, 181, 884-896.
62
-------
REPORT DOCUMENTATION
PAGE
1. REPORT NO.
EPA 560/5-90-005
3. Recipient's Accession No.
4. Till* and Subtitle
Determination of Polybrominated Dibenzo-£-dioxins (PBDDs) and
Dibenzofurans (PBDFs) in Human Adipose Tissue
5. Report Date
April 11, 1990
7. Autnord) P.H. Cramer, J.S. Stanley, K.M. Bauer, R.E. Aylinq,
K.R. Thornburq. J. Schwemberqer
8. Performing Organization Rept No.
9. Performing Organization Neme end Address
Midwest Research Institute
425 Volker Boulevard
Kansas City, MO 64110
10. Proiact/Task/Work Unit No.
8863-A(27)
11. Contrac«C) or Grant(G) No.
(068-02-4252
(6)
12. Sponsoring Organization Nam* and Address
Field Studies Branch, TS-798, Exposure Evaluation Division
Office of Toxic Substances, U.S. Environmental Protection Agency
401 M Street, S.W., Washington, DC 20460
13. Typo of Report & Period Covered
Final Report
19. Supplementary Notes
1C. Abstract (Umtt: 200 words)
This report describes the analytical efforts for the determination of polybrominated
dioxins (PBDDs) and furans (PBDFs) in human adipose tissues. Data on the precision and
accuracy of the method for three tetra- through hexabrominated dioxins and three tetra-
through hexabrominated furans (specific 2,3,7,8-substituted isomers) were generated from
the analysis of 5 unspiked and 10 spiked (5 replicates at 2 spike levels) adipose tissue
samples that were included with the analysis of the FY 1987 samples. In addition, data
are presented on the results of the analysis of 48 composite samples for the six specific
PBDD and PBDF compounds.
The targeted 2,3,7,8-substituted PBDDs and PBDFs were not detected 1n any of the
samples except those prepared as spiked QC materials. The detection limits calculated for
the tetrabromo congeners in the samples ranged from 0.46 to 8.9 pg/g (lipid basis). The
detection limits for the higher brominated congeners were typically greater than that
observed for the tetrabrominated compounds.
There is some evidence for the presence of other brominated compounds present in the
adipose tissue samples. Specifically, responses were noted that correspond to the quali-
tative criteria for polybrominated dlphenyl ethers (hexa through octabromo). It is recom-
mended that these responses should be confirmed as brominated dlphenyl ethers through
additional analysis efforts in order to expand the data base on documented human exposure.
17. Document Analysis i
Polybrominated dibenzo-p_-dioxins Polybrominated dibenzofurans
High resolution gas chromatography/mass spectrometry
Human adipose tissue
National Human Adipose Tissue Survey (NHATS)
b. Identifiers/Open-ended Terms
Determination
Analysis
c. COSATI Reid/Group
IB. Availability Statement
19. Security Class (This Report)
20. Security Class (TMs Page)
21. No. of P»ges
69
(See ANSt-Z39.1g)
See Imttructtom on Reverse
OPTIONAL FORM 272 (4-77)
(Formerly NTIS-35)
DopttrttVMfn Of COfffflfffft%fCO
-------
DATE DUE
U.S. Environmental Protection Agency
Region 5.Library (PL-12J)
77 West Jackson Boulevarc1, i/ih 4L"iooj
Chicago, IL 60604-3590 ^
-------
------- |