^r

-------
Chlorinated Dioxins and Furans
in the General U.S. Population:
       NHATS FY87 Results

             Final Report
                 Prepared by:

                John S. Stanley
           Midwest Research Institute
             425 Volker Boulevard
            Kansas City, MO 64110
 EPA Contract No. 68-DO-0137, Work Assignment 27

                    and

                 John Orban
                  Battelle
              Columbus Division
                505 King Ave.
             Columbus, OH 43201
            EPA Contract 68-02-4292
                   for the:

             Field Studies Branch
                    and
         Design and Development Branch
          Exposure Evaluation Division
           Office of Toxic Substances
      U.S. Environmental Protection Agency
              401 M Street, S.W.
            Washington, D.C. 20460
                             U.S. Environmental Protection Agency
                             Region 5.Library CPM?><
                             77 West Jackson D .''   .., .12Ji Floor
                             Chicago, IL  6060s-,	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. Environmental
Protection Agency.  The use of trade names or commercial products does not
constitute Agency endorsement or recommendation for use.

-------
AUTHORS AND CONTRIBUTORS

       The determination of the levels of the polychlorinated dibenzo-p-dioxins (PCDDs) and
dibenzofurans (PCDFs) in the general population of the United States described in this report
was achieved through cooperative efforts of many EPA and contract support staff. EPA staff
participating in the program included principal investigators from the Field Studies Branch
(FSB) and Design and Development Branch (DDE)  of the Exposure Evaluation Division (BED)
of the Office of Toxic Substances (OTS).   Contract support to OTS was provided by Midwest
Research Institute (MRI) under EPA Contract Nos. 68-02-4252 and 68-DO-0137, and Battelle
Columbus Division under Contract No. 68-02-4292. The roles and responsibilities of each of
these organizations and contributing authors to this report are presented below.

Midwest Research Institute (MRI)

       MRI was  responsible for  the  coordination of the collection of the FY87  NHATS
specimens, preparation of  the  NHATS composites  and  quality   control (QC)  samples,
conducting  the  HRGC/HRMS analysis of  the  composites,  reporting  the  results  and
coordination of the preparation of this  final  report.   Individuals contributing  to this  final
report included:  Dr. John Stanley,  Ms. Karin Bauer, Mr. Paul Cramer, and Dr. Jerry Flora.

Battelle Columbus Division

       Battelle was responsible for developing the FY87 NHATS specimen collection design,
creating and maintaining the data bases  on the Patient Summary Reports (PSRs), designing
the  specimen  compositing   plan  and the  statistical  methodology  for data analysis,  and
conducting the statistical analysis to develop estimates of the PCDD and PCDF residue levels
in the general  U.S. population based on demographic factors.  Individuals contributing to this
final report included:   Dr. John Orban, Dr. Robert Lordo, Dr. Al Unger, Dr. Ron Menton,
Ms. Barbara Leczynski, Ms. Tamara Collins, Ms. Pam Harford, and Ms. Claire Matthews.

Westat

       Mr. John Rogers of Westat provided technical support in processing all analysis reports
and conducting outlier analyses.

EPA Exposure Evaluation Division  (BED)

       EPA was responsible  for oversight in the development of the study  plan, managing
and coordinating  the conduct of the overall study, and reviewing, editing and finalizing this
report. Key staff included:  Ms. Janet Remmers, Mr. John Schwemberger as Work Assignment
Managers and Dr. Joseph Breen and Ms. Edith  Sterrett as Project Officers.
November 27,1991
                                         111

-------
CONTENTS
Preface 	iii
Figures 	vii
Tables	  viii
Glossary 	 xi
Executive Summary  	  1

1.0 Introduction	  7
       1.1  Background  	  7
       1.2  Emphasis on PCDDs and PCDFs 	  7
       1.3  Related studies 	  9
       1.4  Objectives of the FY87 study  	  9
       1.5  Report organization  	   12
2.0 Summary and Recommendations	   13
       2.1  Summary of FY87 NHATS results  	   13
       2.2  Comparison of FY87 NHATS with other studies  	   18
       2.3  Recommendations for future studies  	   21
3.0 Overall Data Quality 	   23
       3.1  Sampling	   23
       3.2  Compositing	   25
       3.3  Demonstrated laboratory quality control  	   26
       3.4  Statistical analysis	   27
4.0 NHATS Sample Design  	   29
       4.1  Sampling design	   29
       4.2  Sample collection procedures  	   35
       4.3  Sample collection summary   	   37
5.0 Composite Design  	   41
       5.1  Design goals and compositing criteria  	   41
       5.2  Laboratory compositing procedures	   43
       5.3  Summary of composite samples  	   45
6.0 Chemical Analysis Procedures and Quality Control Data  	   49
       6.1  Sample preparation	   49
       6.2  HRGC/HRMS analysis 	   52
       6.3  QA/QC for chemical analysis  	   59
       6.4  Synopsis of analytical  results  	   72
       6.5  Statistical analysis of the quality control data	   76
7.0 Statistical Methodology  	   85
       7.1  Statistical model  	   85
       7.2  Statistical analysis	89
8.0 Results 	93
       8.1  Data restrictions and descriptive statistics	93
       8.2  Population estimates	95
       8.3  Hypothesis testing  	98
       8.4  Estimated rates of change of selected PCDD and PCDF
             concentrations	  100

-------
CONTENTS (continued)
      8.5 Outlier detection	102
      8.6 Model validation	. . 104
9.0  Comparison of FY87 Data with FY82 and VA/EPA Data Bases	 . 107
      9.1 Overview of the analytical programs	108
      9.2 Comparison of study designs	108
      9.3 Comparison of analytical procedures 	110
      9.4 Comparison of results	115
10.0  Bibliography	143
Appendices

      A     FY87 NHATS Composite Data Listed by Analyte for Each Composite        A-l
      B     Analytical Protocol	B-l
      C     Method for Estimating Measures of Uncertainty	C-l
      D     Plots of Estimated Concentrations Versus Spiked Level with Tolerance
             Bounds for FY87 NHATS QC Samples  	D-l
      E     Supplementary Descriptive  Statistics  for FY87  NHATS  PCDDs  and
             PCDFs  	E-l
                                         VI

-------
FIGURES


Number                                                                        Page

ES-1.   National average concentration for PCDDs and PCDFs from FY87 NHATS	  4

2-1.    Estimated national  average concentrations for the target PCDDs and PCDFs
       from the FY87 NHATS 	  15

2-2.    Estimated averaged concentrations  by age  group for selected PCDDs and
       PCDFs from the FY87 NHATS  	  17

3-1.    Program steps designed to ensure overall program quality  	  24

6-1.    Example of the separation of 2,3,7,8-TCDD from other TCDD isomers on a 60-m
       DBS column  	  61

8-1.    Measured concentrations of 2,3,7,8-TCDD versus average  age with estimated
       regression lines	101

9-1.    Comparison of analytical methods for FY82, VA/EPA, and FY87 studies  	112

9-2.    Comparison of analytical standards from the FY82, VA/EPA, and FY87 studies	114

9-3. Example of profile plots  	122

9-4.    Profile plots of average 2,3,7,8-TCDD concentrations with 95%  confidence limits
       for FY82 and FY87 adipose tissue samples  	125

9-5.    Profile plots  of  average  2,3,4,7,8-PeCDF concentrations with 95%  confidence
       limits for FY82 and FY87 adipose tissue samples	126

9-6.    Profile plots  of  average 1,2,3,7,8-PeCDD concentrations with 95%  confidence
       limits for FY82 and FY87 adipose tissue samples	127

9-7.    Profile plots of average HxCDD concentrations with 95%  confidence limits for
       FY82 and FY87 adipose tissue samples 	128

9-8.    Profile plots of average 1,2,3,4,6,7,8-HpCDD concentrations with 95% confidence
       limits for FY82 and FY87 adipose tissue samples	129

9-9.    Profile plots of average OCDD concentrations with 95% confidence limits for
       FY82 and FY87 adipose tissue samples 	130
                                         vu

-------
TABLES


Number                                                                     Page

1-1.   Summary of Studies Conducted to Determine Levels of PCDDs and PCDFs in
      Human Tissues	   10

2-1.   Estimated National Average Concentrations  of PCDDs and PCDFs from  the
      FY82 and FY87 NHATS	   19

3-1.   Summary of DQOs for PCDD/PCDF Analysis Effort  	   26

4-1.   Summary of Sampling Strategy	   31

4-2.   Sample MSAs for Fiscal Year 1987 National Human Adipose Tissue Survey  	   33

4-3.   Age, Race,  and Sex  Subquotas for each NHATS  Collection Site within a
      Stratum-FY87 Design	   36

4-4.   FY87 NHATS Collection Summary  	   38

4-5.   FY87 NHATS Sample Sizes by Categories	   39

5-1.   Distribution of FY87 NHATS Composite Samples  by Census Division and Age
      Group	   46

5-2.   Demographic Makeup of FY87 NHATS Composite  Samples	   47

6-1.   Internal Standard Spiking Solution for chlorinated  Species	   50

6-2.   Concentration Calibration Solutions for PCDD/PCDF	   53

6-3.   Initial Calibration DataRelative Response Factors	   54

6-4.   Typical Daily Sequence for PCDD/PCDF Analysis  	   55

6-5.   Ions Monitored for HRGC/HRMS of PCDD/PCDF	   56

6-6.   HRGC/HRMS Operating Conditions for PCDD/PCDF Analysis 	   58

6-7.   Quality Control Samples	   62

6-8.   PCDD and PCDF Native Spiking Solution 	   64

6-9.   Control Sample Spike Levels	   65

6-10.  PCDD and PCDF Spike Check Results  	   66
                                       V1U

-------
TABLES (continued)


Number                                                                       Page

6-11.   OCDD Spike Check Results 	  66

6-12.   Results of  Performance  Audit Samples  Analyzed With the FY87 NHATS
       Composites 	  67

6-13.   Interlaboratory ComparisonSpike Check Solution 	  68

6-14.   Interlaboratory ComparisonControl Lipid Results (pg/g)  	  69

6-15.   Interlaboratory ComparisonLow Level Spiked Lipid (% Recovery)  	  70

6-16.   Interlaboratory ComparisonHigh Level Spiked Lipid (% Recovery)	  71

6-17.   Summary of QC Data for the FY87 NHATS PCDDs and PCDFs 	  73

6-18.   Regression Models Used to Analyze FY87 NHATS QC Data 	  79

6-19.   Estimated Recoveries for Each Batch and Estimated Average Recovery Over All
       Five Batches 	  81

6-20.   Predicted  Concentration (C, pg/g) and Coefficient  of Variation at Each Spike
       Level	  82

7-1.    NHATS Analysis Factors and Categories  	  86

8-1.    Frequency Counts of Data Restrictions and Data Qualifiers in the 48 Samples
       for FY87 NHATS Composite Samples 	94

8-2.    Summary of Unrestricted Data for FY87 NHATS Composite Samples  	96

8-3.    Estimated Average Concentrations (pg/g) with Relative Standard Errors (%) for
       Selected PCDDs and PCDFs from FY87 NHATS Composite Samples	97

8-4.    Significance  Levels  from  Hypothesis  Tests  for  Differences  Between
       Demographic Groups for NHATS FY87 PCDDs and PCDFs	99

8-5.    Estimated Rates  of Change of Selected Dioxin and Furan Concentrations in
       Human Adipose Tissue	103

8-6.    R-Squared Correlation  Between Predicted and  Observed Concentrations for
       FY87 Dioxins and Furans  	105
                                         rx

-------
TABLES (continued)
Number                                                                      Page

9-1.   Marginal Comparisons of FY82 and FY87 NHATS Individual Specimens Used
      for PCDD and PCDF Analysis	109

9-2.   Marginal Comparisons of FY82 AND FY87 NHATS Composite Designs	Ill

9-3.  Number  of  Composite Samples  and Types  of Statistical Comparisons  Made
      Between FY82 and FY87 NHATS Results  	117

9-4.   Comparisons of LODs and  Percent Detected for PCDDs and PCDFs Between
      NHATS  FY82 and FY87	118

9-5.   Comparisons  of  NHATS  FY82   and   FY87 Predicted  National Average
      Concentrations of PCDDs and PCDFs in Human Adipose Tissue 	120

9-6.   Comparisons of NHATS FY82 and FY87 Significance Levels from Hypothesis
      Tests for Differences  Between  Demographic Groups for Selected PCDD and
      PCDF Levels 	123

9-7.   Results of Profile Analysis Comparing FY82 and FY87 Predicted Concentrations
      of Selected PCDDs and PCDFs in Human Adipose Tissue - By Census Region .... 131

9-8.   Results of Profile Analysis Comparing FY82 and FY87 Predicted Concentrations
      of Selected PCDDs and PCDFs in Human Adipose Tissue - By Age Group	132

9-9.   Results of Profile Analysis Comparing FY82 and FY87 Predicted Concentrations
      of Selected PCDDs and PCDFs in Human Adipose Tissue - By Race Group  	133

9-10.  Results of Profile Analysis Comparing FY82 and FY87 Predicted Concentrations
      of Selected PCDDs and PCDFs in Human Adipose Tissue - By Sex  	134

9-11.  Weighted Average Concentrations  and Standard Errors (SE) of Selected PCDFs
      from FY82 and FY87 NHATS 	138

9-12.  Arithmetic Averages  (pg/g), Standard  Errors, and  Number of  Samples for
      Selected  PCDDs Obtained from the VA/EPA, FY82 NHATS (15-44 Age Group),
      and FY87 NHATS (15-44 Age group) Studies by Collection Year Category	139

9-13.  Arithmetic Averages (pg/g), Standard Errors, and  Sample  Sizes for Selected
      Analytes Obtained from the VA/EPA, FY82 NHATS (15-44  Age  Group), and
      FY87 NHATS (15-44 Age Group) Studies	141

-------
GLOSSARY
DDE         Design and Development Branch
BED         Exposure Evaluation Division
EPA         Environmental Protection Agency
FSB          Field Studies Branch
FY          Fiscal Year
HpCDD      Heptachlorodibenzo-p-dioxin
HpCDF      Heptachlorodibenzofuran
HpCDPE     Heptachlorodiphenylether
HxCDD      Hexachlorodibenzo-p-dioxin
HxCDF      Hexachlorodibenzofuran
HxCDPE     Hexachlorodiphenylether
IQS          Internal Quantitation Standard
MRI         Midwest Research Institute
MSA         Metropolitan Statistical Area
NHATS      National Human Adipose Tissue Survey
OCDD       Octachlorodibenzo-p-dioxin
OCDF       Octachlorodibenzofuran
OTS         Office of Toxic Substances
PCBs         Polychlorinated Biphenyls
PCDD       Polychlorodibenzo-p-dioxin
PCDF        Polychlorodibenzofuran
PCDF        Polychlorinated dibenzofuran
PeCDD      Pentachlorodibenzo-p-dioxin
PeCDF       Pentachlorodibenzofuran
RS           Recovery Standard
TCDD       Tetrachlorodibenzo-p-dioxin
TCDF        Tetrachlorodibenzofuran
TSCA        Toxic Substances Control Act
VA          Veterans Administration/U.S. Department of Veteran Affairs
                                         XI

-------
EXECUTIVE SUMMARY

BACKGROUND

       The National Human Monitoring Program (NHMP), operated by the United States
Environmental Protection Agency's Office of Toxic Substances (USEPA/OTS), is an ongoing
national  program to monitor the human  body burden of selected  chemicals.   The main
operative program of the NHMP is the National Human Adipose Tissue Survey (NHATS).
The NHATS is an annual survey to collect and analyze a nationwide sample of adipose tissue
specimens from autopsied cadavers and surgical patients.  The purpose of the NHATS is to
identify and quantify the prevalence and levels of selected compounds in human adipose
tissue.  The analysis results are used to establish an exposure-based chemicals list and to
estimate baseline levels and trends of the selected chemicals.

       In the past, NHATS data have been used to monitor levels  of selected organochlorine
pesticides and polychlorinated biphenyls (PCBs) in the United States population.  However, in
Fiscal Year 1982 (FY82)  the specimens collected  under NHATS were used in the Broad Scan
Analysis Study, in which composited NHATS specimens were analyzed using high resolution
gas chromatography/mass  spectrometry (HRGC/MS)  in order  to  expand the list  of target
analytes.  The Broad Scan Analysis Study identified 17 volatile organic compounds,  30 semi-
volatile organic compounds,  5 polychlorinated dibenzo-p-dioxins (dioxins or PCDDs),  and 5
polychlorinated dibenzofurans (furans  or  PCDFs).    The  Broad  Scan  Analysis  Study
demonstrated that the PCDDs and  PCDFs could be detected in the general U.S. population
across all geographic regions and age groups.  Using a similar study design, EPA analyzed the
FY87 NHATS samples for PCDDs and PCDFs.

      This report presents the objectives, methodology, and  results of  the  FY87 NHATS
samples and compares the FY87 NHATS results with results from the analysis of the FY82
NHATS samples and from a related study of PCDDs and PCDFs conducted jointly by the U.S.
Veterans Administration and EPA's Office of Toxic Substances (VA/EPA).

OBJECTIVES

      The specific objectives of the FY87 NHATS study were to:

            Identify the  PCDD  and  PCDF  isomers  (specifically  those  with  chlorine
             substitution in the 2,3,7,8-position) that are present in human adipose tissue,

            Estimate the average concentrations of PCDDs and PCDFs in the adipose tissue
             of humans in the U.S. population and various demographic subpopulations,

-------
             Determine if any of the demographic factors (geographic region, age, race, and
             sex) are associated with the  average concentrations of PCDDs and PCDFs in
             human adipose tissue, and

             Compare the estimated  average  concentration levels of PCDDs  and PCDFs
             found in the FY87 NHATS  with estimates from the FY82 NHATS  and the
             VA/EPA studies.
APPROACH
       Population estimates were based on data obtained from the  chemical analyses of 48
composite samples.   The composite samples  were prepared from 865 individual specimens
collected in  a  statistically designed survey of autopsied cadavers and surgical patients.   A
statistical design was also used in creating the composite samples. HRGC/HRMS was used to
measure the concentrations of seven PCDDs  and 10 PCDFs in the composite samples.  The
resulting data were statistically analyzed using  a model developed specifically for  the
composite samples to estimate average concentration levels in the  U.S. population and to
determine if any of  the four demographic factors (geographic region, age, race, and sex) are
associated with the average concentration levels.

       Traditionally, one of the objectives of NHATS has been to estimate the prevalence of
chemicals in the U.S. population as well as in various demographic subpopulations.  More
generally this involves characterizing the distribution (i.e., variability of body burden levels
among  individual donors.  While  these are still important objectives for  NHATS, it  was
decided  that the FY87  survey  would  focus  available resources  on estimating average
concentration levels  in the U.S. population  and characterizing differences among various
demographic subpopulations.  With the more focused objectives it became more efficient to
combine individual specimens into composite samples prior to chemical analysis.

       The principal  advantage of compositing is that it reduces the  number of samples  that
need to be chemically analyzed.  However, this does not necessarily result in a corresponding
reduction of precision for the population estimates.  With composite samples, the effects of
sampling error on the precision  of  average concentration estimates are substantially reduced
because the concentration  in a  composite  sample is  the average  of individual specimen
concentrations.

       The main disadvantage of using data  from composite samples is that the  prevalence
and distribution of chemical concentrations among individuals in the population cannot be
accurately estimated. However,  it was possible to achieve all of the study objectives for the
FY87 NHATS as outlined above by  using appropriate design criteria for combining specimens
into composites.

-------
       The  overall  quality of  results  was ensured  by  using statistical  sampling  and
 compositing  designs, implementing  quality control procedures,  validating  analytical  and
 statistical  methodologies, and  adhering  to  data quality  objectives.  Data quality objectives
 (DQOs) were established for target ion ratios, percent recovery of target analytes, and the
 presence of chemical interferences.  Only data  that met all DQOs were included in the
 statistical summaries and analyses.

 SUMMARY OF THE FY87 NHATS RESULTS

 National Weighted Average Concentrations

       The estimated national average concentrations of the target analytes  are presented
 graphically in Figure ES-1.  The 1,2,3,4,7,8- and 1,2,3,6,7,8-HxCDD isomers have been reported
 together due  to incomplete chromatographic separation.   Also, the percent of composite
 samples in which the compound was detected is listed for each target compound.  Data from
 all of the target PCDDs and three of the 10 PCDFs were analyzed using a statistical modeling
 approach.  Population estimates were  calculated  using  weights  determined  from the 1980
 census counts.  Based  on  the 95%  confidence intervals the maximum uncertainty of the
 estimated average concentrations is 12% for the six PCDDs and 26% for the three PCDFs.  The
 estimates for the remaining PCDFs were calculated using weighted averages of the measured
 composite concentrations.  Weights were determined from the 1980 U.S.  Census counts.  The
 seven remaining chemicals (PCDFs) did not meet minimum criteria to perform a more detailed
 statistical analysis.   Several  of these compounds  were not detected in a  sufficient number of
 composite samples to allow meaningful statistical analysis.

 Age Effects

       Each composite was  prepared with specimens from one of the three age groups (0-14,
 15-44  and  45+ years). The  average concentrations of the nine PCDDs and PCDFs that were
 statistically analyzed were found to increase significantly with the age of the donor.  For
 example, the average concentration of 2,3,7,8-TCDD increased from 1.98 pg/g in the youngest
age group (0-14 years) to 9.40  pg/g in the oldest age group (45+ years)  a  375% increase.
Increases  attributed to age effects in the other chemicals ranged from 24% for 2,3,7,8-TCDF to
863% for 2,3,4,7,8-PECDF.

-------
 CD <
 O)  ^
 CO .

 05  c  CO
^  2  ^


si  *
 c  c
 O  CD

B  8
2- O
    o
     3]       5]

OOi-OCM
c    co  r^  Tr
-^-
in
                                                      51  5]  Jo]

                                                           COCO
                                                      co  fv  c\i
T-  r^  T-  -  CM
                                                                      p
                                                                      o
                                                                      o
                                                                      o
                                                                      o
                                                                      p
                                                                      o
                                                                      o
                                                                       o
                                                                       o
                                                                   :    o     2     
                                 c
                                 g

                                 03
                                                                              C
                                                                              CD
                                                                              O
                                                                              C
                                                                              o

                                                                              O
                                                                       O
                                                                       o
                                                                       o


                                                                       o
         O5COO5^
         CO


TJ
C
2
O
Q.

c
O
O

Q
O
o
1
CO
jC
en"
CM




g
O
I
1
CO
i-C
en"
CM
^



g
O

i
CO
rC
co"
en
CM"
,_"
CO
N."
g
O
5
i
CD
CO
N."
en
c\T
,_"


g g
0 g
Q. U
I O
1
CO
rv."
CD"
TT
cn
CM"


ft fe
Q o
O co
H Q.
1 1
CO CO
iC K
en cn_
CM" CM
*^



LL
O
o

1
CO
rC
^"
cn"
CM"



fe
0
X
T
CO
rC
^"
en"
CM"
^_"


fe
O

1
CO
tvT
to"
en"
CM"
^_*


^
O
X
1
CO
rC
co"
qr
en"
CM"


LL
Q
0

1
O5
co"
rs."
en"
CM"
^_"


^
O
X"
1
00
rC
CO*
rr"
en"
CM"


LL
o

1
OJ
co"
(C
^
eo"
CM


                                                                CO
                                                                LL
                                                                O
                                                                O

                                                                o
                                                               2

                                                               TJ
                                                               0)
                                        to

                                        O

                                        CO
                                        _QJ

                                        75



                                        175
                                        UJ
                                                                                     CO
                                                                                     ^J

                                                                                     Q.


                                                                                     ro
                                                                                     CO

                                                                                     a>
                                                                                      o
                                                                                     O
                                                                                      0)

                                                                                      e
                                                                                      0)
                                                                                     Q.
                                                                                               <
                                                  u-
                                                                                                o
                                                  Q
                                                  U
                                                  c-

                                                  13
                                                   c
                                                   re
                                                   U2
                                                  D
                                                  Q
                                                  U
                                                                                                              c
                                                                                                              o
                                                                                                             x:
                                                                                                              re
                                                                                                              u.
                                                                                                             ^-*



                                                                                                              I
                                                                                                              O
                                                                                                              u

                                                                                                              cu
                                                                                                              tc
                                                                                                              re
                                                   re


                                                  I
                                                   re

                                                  2
                                                                                                 3
                                                                                                 bC

-------
Geographic Effects

      There were significant differences in the estimated average concentrations of 2,3,4,7,8-
PeCDF among the four geographic regions (north central, northeast, west, and south). The
estimated average concentration in the west region was 4.49 pg/g compared to  an average of
13.7 pg/g in the northeast and a national average of 9.70 pg/g.  The data combined with the
FY82 NHATS data also suggest the possibility of regional effects on the average concentrations
of 2,3,7,8-TCDD  and  1,2,3,4,6,7,8-HPCDD, but the  estimated regional  differences were not
statistically significant.

Race and Sex Effects

      The  differences  in  estimated average  concentrations between  Caucasian and  non-
caucasian donors and between male and female donors were not statistically significant for
any of the modeled compounds.

COMPARISON WITH FY82 NHATS

      The  FY87  NHATS  estimated  national average  concentrations  of  2,3,7,8-TCDD,
1,2,3,4,6,7,8-HPCDD, and OCDD  were  consistent with  the estimates established in the  FY82
NHATS.   However,  the FY87 estimates  for many of the  other PCDDs and PCDFs  were
significantly lower than the estimates  obtained in the  FY82 NHATS.  It is likely that these
differences are due in part to advances in the analytical method between FY82 and FY87.

COMPARISON WITH THE VA/EPA STUDY

      The VA/EPA study was a retrospective study which used  surplus specimens from the
NHATS  repository to  compare PCDD  and  PCDF  concentrations  in the adipose tissue  of
Vietnam  veterans with non-Vietnam veterans and civilians.  A total of 197 specimens collected
between  1971 and 1982 were analyzed individually using analytical methodology equivalent to
the FY87 NHATS study. All specimens in the VA/EPA study are within the middle age group
(15-44 years) and hence allowed comparison to the same age  group evaluated  in the FY82 and
FY87 NHATS studies.   The estimates obtained in the VA/EPA study were two to three times
higher than the FY87 NHATS  estimates.  The  average  values of the compounds across
collection years (1971 to 1987), particularly 2,3,7,8-TCDD, indicate  a  decrease in adipose tissue
residue levels.  These differences may be  due to a  decrease in exposure levels through
environmental pathways or consumer products over time or as a result  of the  differences in
storage times between the VA/EPA (up to  16 years) and FY87 NHATS (up to 2 years) studies.
Further studies or analyses are needed to resolve these differences.

-------
1.0 INTRODUCTION

1.1 BACKGROUND

       The National Human Adipose Tissue Survey (NHATS) is the main operating program of
EPA's National Human Monitoring Program (NHMP).  Under the NHATS program, human
adipose tissues are analyzed to monitor human exposure to potentially toxic compounds. The
adipose tissue specimens were collected according to statistically designed sampling plans. The
tissues  are  collected  by cooperating  pathologists  and  medical  examiners during routine
examination of tissues that have been excised during post mortem examinations or for therapeutic
reasons during  elective surgeries.  The tissues are collected based on a statistically developed
design of Metropolitan Statistical Areas (MSAs) to provide chemical residue information that can
be correlated to demographic data (geographic region, age, sex, and race).

       The NHATS program was established in the late 1960's. The chemical residues of interest
during that time frame were organochlorine pesticides and PCBs.  Recognizing the need to extend
the capabilities of the NHATS program, the Office of Toxic Substances (OTS) initiated a series of
programs in 1984 to expand the utility of the tissue repository. Foremost in these analysis efforts
was the conduct of a study termed the "Broad Scan Analysis Study." The broad scan analysis
included the determination of volatile and semivolatile organic  compounds based on methods
requiring  mass  spectrometry as  a detection device.   Included in the semivolatile organic
compounds were the poly chlorinated dibenzo-p-dioxins (PCDDs) and the dibenzofurans (PCDFs).
The data from the broad scan analysis effort demonstrated  that the PCDDs and PCDFs were
prevalent in the adipose tissues of the general U.S.  population from all census divisions and
within each age group.

1.2 EMPHASIS ON PCDDs AND PCDFs

      The ubiquity of the PCDDs and PCDFs in adipose tissues provides evidence of widespread
exposure to a class of potentially toxic compounds.  On the basis of animal studies, the U.S. EPA
considers 2,3,7,8-TCDD to be one of the most potent known carcinogens studied (U.S. EPA 1987,
U.S. EPA 1989).  Studies on the toxicity of other PCDDs and PCDFs have demonstrated the
compounds with chlorine substitutions in the 2,3,7,8-ring positions results in the greatest toxicity.
In addition, it has  been demonstrated that toxicity varies with the degree of chlorination.

      Potential for exposure to PCDDs and PCDFs exists from multiple sources including:

            Diet, particularly fish, poultry, meats,  and dairy products that are impacted by
             environmental releases of PCDDs and PDCFs.

-------
            Incinerator emissions  (municipal,  hospital and hazardous  waste incinerators,
             automobiles, etc.),

            Halogenated commercial products such as specific herbicides (2,4-D, Agent Orange)
             or wood preservatives  (pentachlorophenol), and

            Bleached paper products and the effluents  from pulp and paper bleaching
             operations.

            Combustion of halogenated aromatics, e.g., PCB transformer fires.

            Metal smelting and reclamation processes.

       The PCDDs and PCDFs are environmentally persistent, and the most toxic isomers, the
2,3,7,8-substituted compounds, bioaccumulate in the food chain. EPA has studied the prevalence
of these compounds extensively in the general  environment through the various programs
conducted as part of the National Dioxin Study, and in  support of investigations at potential
hazardous waste sites. The current studies involving the pulp and paper industry resulted from
data generated from fish residue data from.the National Dioxin Study. In addition, several states
(e.g., California, Vermont, Connecticut, and Mississippi) have conducted monitoring programs to
determine background levels of PCDDs and PCDFs in ambient air, soil, sediment, and fish in
areas that are potentially impacted by emissions from incineration sources or effluents from the
pulp and paper industry (Stanley et al. 1989a, Heil and Fitzgerald 1987, State of Connecticut 1986,
personal contacts).

       EPA has promulgated regulations to specifically address the releases of PCDDs and PCDFs
to the environment.  Under Sections 4 and  8 of the Toxic Substances Control Act (TSCA), the
manufacturers of specific halogenated products are required to determine and report the levels
of halogenated dioxins and furans in those products (U.S. EPA  1987).  EPA has also recently
developed guidelines for establishing test procedures for release of halogenated dioxins and furans
in industrial effluents (USEPA 1991), particularly from pulp and paper bleaching facilities.

       The NHATS program provides EPA  with a unique capability to monitor the impact of
changes in environmentally persistent chemicals. As an example, the NHATS program has
demonstrated a decline in the adipose tissue levels of PCBs in the general U.S. population since
the restriction and regulation of use and disposal of PCBs in the mid 1970s (Robinson et al. 1990).
The NHATS samples should also provide data on the result of efforts to reduce human exposure
to PCDDs and PCDFs.

-------
1.3 RELATED STUDIES

       Over the past 10 years, considerable attention has been given to the determination of
PCDDs and PCDF  residues in human tissues.   Many of these  studies are identified and
summarized in Table 1-1. Initial studies focused on body burden levels of individuals who were
potentially  exposed  through occupational exposures (for example, Vietnam veterans to  the
herbicide Agent Orange) (VA/EPA1991, Kang et al. 1991, CDC 1987, CDC 1988, CDC 1988a, Pirkle
et al. 1989, Schecter 1987, Kahn 1988,  Kahn et al.  1988, Kahn  et al. 1990, Gross et al. 1986),
improper handling  of contaminated wastes (Times Beach, Missouri) (Patterson et al. 1986,
Andrews et al. 1989, Patterson et al. 1987a, Patterson et al. 1989), contamination of food products
(Yusho and Yucheng) (Ryan et al. 1987), and accidental releases from chemical production
facilities (Seveso, Italy) (Fachetti et al. 1981, CDC 1988c).

       Several studies have been conducted to determine the levels of PCDDs and PCDFs in the
tissues of the general population both within the United States and internationally.  Table 1-1
provides a summary of the studies that have been reported to date on the residue levels of these
compounds. The studies conducted in the United States include a limited number of samples
from Binghamton, New York (Schecter et al. 1986); Atlanta, Georgia (Patterson et al. 1986); Salt
Lake City, Utah (Patterson et al. 1986); St. Louis, Missouri (Graham et al. 1986a, 1986b); the State
of Missouri (Patterson et al.); and the State of California (Stanley et al. 1989).  The studies
conducted in Binghamton, St. Louis, and the State of California focused on the determination of
total PCDDs and PCDFs. The other studies focused on 2,3,7,8-TCDD only.

1.4 DETECTIVES OF THE FY87 STUDY

       The specific objectives of the FY87 NHATS study were to:

             Identify the PCDD and PCDF isomers (specifically those with chlorine substitution
             in the 2,3,7,8 position) that are present in human adipose tissue,

             Estimate the average concentrations of PCDDs and PCDFs in the adipose tissue
             of humans in  the U.S. population and various demographic subpopulations,

             Determine if any of the demographic factors (geographic region, age, race, and sex)
             are associated with the average concentrations  of PCDDs and PCDFs in human
             adipose tissue, and

             Compare the estimated average concentration levels of PCDDs and PCDFs found
             in the FY87 NHATS with  estimates  from the FY82 NHATS and the VA/EPA
             studies.

-------
c
"S
P^




S
 -S
n ^J
"^ s
g
n W

a J


W
Cu
H!

UM
O
"3
& i
u x:
r! frl
es Conducted to Determil
Exposure classific
3
i2 "B,
^ S
fc> 8
I ^
G o
a 2
en
1 fg
H 8-5
"o J
S "8
(0
g
J-
d.
i
(U
0
5 ol
*3 c? '*"**'
^* S ^^ J
xg; ^ "5
^ > 3 WJ

"^ >3

0)
8 S

1 1

W M
BH BH
D D
Q D
 kl
a
J
1
2
General population comf
763 individual specimens
Civilian
Non- Vietnam veteran
Vietnam veteran


*> g <* >8




 ^
ii "o ii ij a
% o


eg 

1* 1


UH
D
O O
D B
y fc


1 1
s 60 e-
General population. Che
production work handlin
recreational residential e:
General population


3 " m
1-1



S
in g
s
CJ  TT? U5
 to "3 ^ 3 n
5 j2 *i ,&-  
i5 v  C/5 t/J < 	

                               I
                               I
                               en
                      a
 c
 o

3
O.
                      
                                  y  **

                              -S  EH
     5   2 s       ;z
                                          u u
                                           II
                                           1*  o
                                          t/5 t/5
                                         OD
                                         Q Q
                                          &
                                         
                                         p
                              |^    e
         D
         U


O D Q Q
Q D n D
                                                         o  o  o
                                                        B, B, pt,
                                                        (0  03  td
                                                      C X X J3
                                                      c  tu  v  w
                                                      0)  t>o 60 60
                                                     t>  c  e  -

                                                      >  S  E
                                                      goo

                                                     O
             10

-------















f
a
c
a
c
a

*i*
^
2
























Reference




1
(d
a
V
i
.S
 -g
ll

1 B


c
o
J3
5
vC
"S
(d
"u
3
1
S.


(A
"S.
6
^
c>
2
OJ
"5.
B c
a 0
 &
s JS
i"5
to
(0
"s
IE
a.
i
OJO

U
OO O
5?
t-H ri
(d td
S t!
c c
.c .c
**


rt.
S E
cL 
^ 
< cr
in 
SS
u u
Q Q
OD


yl_^
>~,


~
c
S
9) us
> C
B il
| |
^ 
l>-


t>. m







a






Q
c
u
O
5 i
0) O
$ &
p *JS
u 


o> q a
JG 5fi s
CL UH UH
III
UH
D
p n D
O O Q
X









If
s &>
c a |
E o <
u . .
3 .b ?
> ? D


"* m >S







18 s "






1
c
(U

"  ^
"is  tl
1?  n   -jrj u _f a

I   *   B S "*

^   I  8,11-1:

s  I"  lllJ-
                           
                              !
5     5
          mi  ii  m
                                   ss
                                   u u
   1
   
JS

a.
   ts
    c
O  E
H  ffi  S
       E^crb^

       B B B B
       "u "u u *u
       (0  (0 W
       tu u. u. u.
                                    01

                                   O
          is
                                                   0)

                                                   "5.

                                                   B
                                            T3
                                            c
a  -a
ta   ta
c   c

                                         c
                                         v

                                         >
                   11

-------
1.5  REPORT ORGANIZATION

       This report presents all pertinent information with respect to the assessment of data
quality, the statistical design, the conduct and results of the analytical procedures, the statistical
analysis procedures and extrapolations to population estimates, and the comparability of the FY87
data set to the previous PCDD and PCDF analysis programs conducted using the NHATS adipose
tissue samples.

       Section 2.0 (Summary and Recommendations) presents a synopsis of the significant
findings from the statistical analysis of the FY87 data set in comparison to the previous studies
conducted through the NHATS program. Overall data quality issues with respect to analytical
and statistical criteria are discussed in Section 3.0. The study design for collection of individual
specimens in the FY87 collection program is described in Section 4.0, and the compositing design
and procedures for preparing the composite samples are discussed in Section 5.0.  Section 6.0
describe the analytical procedures used to determine of PCDDs and PCDFs and summarizes the
supporting analytical quality control (QC) data. The statistical methodologies are described in
Section 7.0, and the  results  of the  statistical analysis are presented in Section 8.0.  Section 9.0
focuses on the comparisons of the FY87 study with the FY82 and the VA/EPA collaborative study.
Study designs, analytical and statistical methodologies, and graphical comparisons of the FY82 and
FY87 data are covered.

       Supporting information on  the individual sample data, the detailed analytical protocol,
measures of uncertainty, and plots  of QC data are presented in Appendices A through E.
                                          12

-------
2.0 SUMMARY AND RECOMMENDATIONS

       The human adipose tissue specimens analyzed in the FY87 NHATS were collected from
October 1986 through September 1987, following a statistically based survey design.  A statistical
design was also used to composite the specimens prior to chemical analysis. Section 2.1 presents
a  summary of the FY87 NHATS results.  The results include  estimated national average
concentrations of PCDDs and PCDFs in human adipose tissue and the effects of the geographic
region, age, race, and sex on the average concentrations. The statistical comparison of the FY82
and FY87 NHATS results and a descriptive comparison with the EPA/VA results are summarized
in Section 2.2.  Recommendations for future studies are presented in Section 2.3.

2.1 SUMMARY OF FY87 NHATS RESULTS

       Forty eight (48) adipose tissue samples analyzed in the FY87 NHATS were prepared using
865 individual specimens collected in a statistically designed survey of autopsied cadavers and
surgical patients.  The composite samples were analyzed for  PCDDs and PCDFs using high
resolution gas chromatography/mass spectrometry (HRGC/HRMS).   The resulting data  were
analyzed  using  a  statistical model  developed for composite samples  to estimate  average
concentration levels and to determine if any of the four demographic factors (geographic region,
age, race, and sex) are associated with the average concentration levels.

       Traditionally,  one of the objectives of NHATS has been to estimate the prevalence of
chemicals in the U.S. population as well as  in various demographic subpopulations.  More
generally this involves characterizing the distribution (i.e., variability) of body burden levels
among individual donors. While these are still important objectives for NHATS, it was decided
that the FY87 survey would focus available resources on estimating average concentration levels
in the U.S. population and characterizing differences among various demographic subpopulations.
With the more focused objectives it became more efficient to combine individual specimens into
composite samples prior to chemical analysis.

       The principal advantage of compositing is that it reduces the number of samples that need
to be chemically analyzed. However, this does not necessarily result in a corresponding reduction
of precision for the population estimates.  With composite samples  the effects of sampling error
on  the precision of  average concentration estimates are substantially  reduced  because the
concentration in a composite sample is the average of individual specimen concentrations.

       The main disadvantage of using  data from composite  samples,  however, is that the
prevalence and distribution of chemical concentrations among individuals in the population
cannot be accurately estimated. But, it was possible to achieve all of the study objectives for the
                                         13

-------
FY87 NHATS as outlined above by using appropriate design criteria for combining specimens into
composites.

       Because of criteria based on pre-determined data quality objectives (DQOs) the number
of composite samples used in the data analysis varied among the target compounds.  Several of
the compounds, particularly the PCDFs, were not detected in a sufficient number of composite
samples to perform a reliable statistical analysis.  This generally occurred for the compounds that
were detected at levels near or below the method detection limits. Interferences which precluded
accurate measurement, were encountered for two of the HxCDF isomers. Thus, only nine of the
target analytes met minimum criteria for applying the statistical model. However, summary
statistics are provided for all target chemicals. The criteria for applying the statistical model were:

       (1)    At least 30 composites must meet all DQOs for chemical analysis,  and

       (2)    The chemical must be detected in  at least 50%  of the composites meeting the
             DQOs.

       2.1.1 Population Estimates

       Figure 2-1 presents the estimated national average concentrations for the target chemicals.
Because of wide range  of concentration levels among the target chemicals, the estimates are
displayed on a log scale.  The estimates were calculated using either a statistical modeling
approach or by taking weighted averages of the composite sample concentrations. The weights
used in the weighted  average estimates  are based on 1980 U.S. Census population counts.
Further discussion of this approach is presented in Section 9.4.

       For the nine chemicals that were statistically analyzed, the error in estimating the average
concentration levels, based on 95% confidence limits, was less than 12% for the PCDDs and less
than 26% for the PCDFs. The quality of these estimates is supported by validation of both the
chemical analysis method and the statistical approach to handling the data and the performance
of the method as demonstrated from the results of the Quality Control Samples.

       The estimates for the chemicals, particularly PCDFs, that were detected in less than 50%
of the samples may  be  significantly affected by the detection limits of the analytical  method.
Whenever the chemical was not detected in the sample, it was  assumed that the concentration
was below the limit of  detection (LOD) of the analytical method.  Thus, the data value was
"censored" and no measured concentration was reported. To calculate an average concentration
using censored data, the value of LOD/2 was used in place of  a  measured concentration
whenever the chemical  was not detected.  This approach minimizes the potential error in the
estimated average concentration. But when more than 50% of the data are censored there is

                                         14

-------
 CD  <.
 cn o
 CO  ;
C  CO
O  CO
                                 co
                 o  CM  co
                             SI
                                                         3]  JQ|   51
                                                             coco
                                                    (A

                                                   JO

                                                   fo
    to      LO
 co  .b
 c   c
3   
Z8
O  Tf
T-  CVJ
1-  h-
                                                                        q
                                                                        o
                                                                        o
                                                                        o
                                                                        o
                                                                        q
                                                                        CD
                                                                        o
                                                                        o
                                                                        o
                                                                        o
                                                                        o
T3
JD
"o
Q
T3
 C

 O
 Q.

 O
O
                                                    o

                                                    o
                             1^      CO
         O
                      1      g
1
a
eo"
CM
1
ao
co"
^J"
1
00
N.
ec"
co
CM
1
ffl
00
NT
WI
1
eo

-------
much more uncertainty in the estimated average.  The maximum systematic error due to
censuring is the average LOD divided by two.

     2.1.2  Demographic Comparisons

     Statistical hypothesis tests were performed for each of the nine modeled compounds to
determine if there were statistically significant differences in average concentration levels among
individuals from different census regions (north central, northeast, west, and south), age groups
(0-14 years,  15-44 years, and 45+ years), race groups (Caucasian, non-caucasian), and sexes (male,
female). Statistical conclusions derived from  these analyses are summarized below.

     Age Effects.  Concentration of PCDDs and PCDFs increase with age. The differences in
the average concentrations among age groups was statistically significant at the 0.05 level for all
nine of the modeled chemicals. Figure 2-2 displays the average concentrations by age group for
each of the modeled chemicals.  The highest average concentration was always found in the
oldest age group (45+ years) and, except for 2,3,7,8-TCDF, the lowest was found in the youngest
age group (0-14 years).  The increases from the lower to the  upper age groups range from 24%
for 2,3,7,8-TCDF to 863% for 2,3,4,7,8-PeCDF.

     Geographic Effects.  One of the nine modeled chemicals  was found to have statistically
significant differences among the estimated average concentrations from different census regions.
The estimated average concentration of 2,3,4,7,8-PeCDF in the Western census region was 4.49
pg/g compared to an average of 13.7 pg/g in  the North East  region.  The national average was
9.7 pg/g.  No other chemical for the FY87 composites was found to have significant geographic
effects at the 0.05 level of significance.  This  geographic effect for the 2,3,4,7,8-PeCDF was also
detected in  the analysis of the data for the FY82 NHATS samples.

     Race  Effects. The difference in estimated average concentrations between Caucasian and
non-caucasian donors was  not found to be statistically  significant for any of  the modeled
compounds.

     Sex Effects. The difference in estimated average concentrations between male and female
donors was not found to be statistically significant for any of the modeled compounds.

     2.1.3  Quality Control Results

     Twenty (20) quality control (QC) samples were analyzed along with the 48 study samples.
Each of the five analysis batches contained eight to ten study samples, a method blank, an
unspiked control sample, and  two spiked samplesone at a low spike level and one at a high
                                          16

-------
Compound Age Group (
2,3,7,8-TCDD 0-14
15-44
45+
1,2,3,7,8-PeCDD 0-14
15-44
45+
1, 2,3,4,7,8/1 ,2,3,6,7,8-HxCDD 0 - 14
15-44
45+
1,2,3,7,8,9-HxCDD 0-14
15-44
45+
1,2,3,4,6,7,8-HpCDD 0-14
15-44
45+
OCDD 0-14
15-44
45+
0.
2,3,7,8-TCDF 0-14
15-44
45+
2,3,4,7,8-PeCDF 0-14
15-44
45+
1,2,3,6,7,8-HxCDF 0-14
15-44
45+
0.1
National Average
yrs.) Concentration (pg/g)
888888 1.98
4.37

8888888888! 9-10

8888888$ 3.30

888888888^ 9.33

88888888888$ 18.20

8^88888888881 23.4

8^8888888^8^88^ TO-S

88888888^$$8$88I 12-

888888888 6.13
8888888888 1"

88888888888$ 17-1

8888888888888^ 45.7

888888888888888^ 993
88888888888888888 174.

88888888888888888< 215.

88888888888888888888 592.
g8$888$8888$8888888< nso.
10 1.00 10.00 100.00 1000.00 10000.0
Concentration (pg/g)
^8888883 1Q7

^88888^ 1.45
^8888888 2

^888888^ 1.R7
888888888? a.oo

888888888888 IB.O

$888883 ion

$8888888$ 459

^888888888$ 10^
K> 1.00 10.00 100.00 1000.00 10000.0
                                   Concentration (pg/g)

Figure 2-2.  Estimated averaged concentrations by age group for selected PCDDs and PCDFs
           from the FY87 NHATS.
                                   17

-------
spike level. The QC samples were chemically analyzed to characterize the accuracy and precision
of the analytical method and to determine if there were significant batch effects. The statistical
analysis of the QC data demonstrated that batch effects were relatively small and that the method
recovery criteria were met for 97.5% of the target analytes.

2.2  COMPARISON OF FY87 NHATS WITH OTHER STUDIES

     The results from the FY87 NHATS were compared with results from two previous studies
involving specimens from the NHATS repository. Section 2.2.1 discusses the comparison with the
FY82 NHATS Broad Scan study and Section 2.2.2 discusses the comparison with a collaborative
study conducted between the Veterans Administration and the U.S. EPA (VA/EPA).

     2.2.1  Comparison with FY82 NHATS

     The study designs for FY82 and FY87 NHATS were very similar. Although there are minor
differences in the number of specimens collected and their distribution among demographic
populations, these differences do not significantly affect the comparison of results. On the other
hand, there were substantial improvements in the analytical method between the FY82 and FY87
NHATS. Additional internal quantitation standards were added at the sample extraction step and
carried throughout the analysis procedure.  These additional internal  quantitation standards
provide a better characterization of method performance and more accurate determination of the
PCDDs and PCDFs via the isotope dilution principle.

     Statistical comparisons were performed to determine significant differences for the specific
compounds between the FY82 and FY87 NHATS studies.  Differences noted in the measured
concentrations may be due to either actual changes in body burden levels or changes in the
analytical method.  In addition to comparing national  average concentrations, a more detailed
profile  analysis was performed to determine if the demographic effects are consistent between
FY82 and FY87.  The profile analysis provides meaningful results even if there are  systematic
differences in measured concentration levels.

     Table  2-1 compares the estimated national average concentration levels for selected
chemicals analyzed in FY82 and FY87. Also presented are the relative changes from FY82 to FY87.
For example, the estimated concentration of 2,3,7,8-TCDD is 9% lower for FY87 than determined
for  FY82.    Statistical comparisons were possible for the  PCDDs and only one PCDF, the
2,3,4,7,8-PeCDF. The other PCDFs could not be compared statistically because one or both of the
estimates are based on data which did not meet the minimum criteria for statistical modeling.
                                         18

-------
   Table 2-1.  Estimated National Average Concentrations of PCDDs and
                PCDFs from the FY82 and FY87 NHATS

PCDDs
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
PCDFs
2,3,7,8-TCDF
2,3,4,7,8-PeCDF
HxCDF
1,2,3,4,6,7,8-HpCDF
OCDF
National
FY82
(Pg/g)

5.88
73.6 >b
122 >
142
768

39.7
35.4 >
20.9
20.6
56.0
average
FY87
(Pg/g)

5.38
10.7
86.8
110
724

1.86
9.70
14.2
15.3
2.28
Percent
change*

-9%
-85%
-29%
-23%
-6%

-95%
-73%
-32%
-26%
-96%
                       (FY87-FY82)
                     18     FY82   :

b The FY82 estimate is significantly higher (denoted by >) than the FY87
 estimate at the 0.05 level of significance.
c No statistical comparisons were possible.
                                19

-------
      All of the chemicals had higher estimated concentrations in FY82.  However, among the
modeled chemicals, the differences in estimated concentrations between FY82 and FY87 were not
statistically significant for 2,3,7,8-TCDD, 1,2,3,4,6,7,8-HpCDD, and OCDD.  These chemicals had
estimated relative changes of less than 23%. Some of the differences between FY82 and FY87 can
be explained by changes in the analytical method.  This is evident in the fact that the detection
limits in FY82 were often higher than the estimated average concentration in FY87. For example,
the estimated detection limit for OCDF in FY82  was 19.0 pg/g. The  average concentration
measured for the FY87 samples was 2.28 and the detection limit was 1.67 pg/g. In addition to
achieveing lower detection levels, the analysis of the FY87 NHATS composites was enhanced
through  the use of additional internal quantitation standards  that provided more accurate
measurements  of  the PCDD and PCDF residue  levels.   The effect of using fewer internal
standards in the FY82 anaysis may have resulted in over or under estimates of the concentrations
of the penta-, hexa-, and hepta-chlorinated PCDDs and PCDFs.

      The profile analysis provided additional insight into the demographic effects on PCDDs and
PCDFs.  As previously discussed, the FY87 data demonstrated that there were geographic effects
on the concentrations of 2,3,4,7,8-PeCDF.  Although there was a significant difference in the
national average concentrations between FY82 and FY87,  the FY82 data confirmed the existence
of geographic effects. In both years, the highest estimated concentrations of 2,3,4,7,8-PeCDF were
found in the northeast. The profile analysis also suggested that there are possible geographic
effects on the concentrations of 2,3,7,8-TCDD and 1,2,3,4,6,7,8-HpCDD.  The profiles for FY82 and
FY87 were very similar with the lowest estimated concentrations in the west and south census
regions.

      2.2.2 Comparison with the VA/EPA Study

      The VA/EPA study  was a retrospective study which used surplus specimens from the
NHATS repository to compare PCDD and PCDF concentrations in the adipose tissue of Vietnam
veterans with non-Vietnam veterans and civilians. The study was conducted using approximately
200 adipose tissue specimens collected between 1971 and 1982 from male donors between the ages
of 17 and 46. Each specimen was analyzed for the same PCDDs and PCDFs that were measured
in the FY87 composites.

      Although there are differences between the VA/EPA study and the FY87 effort in terms of
study objectives, sampling plans, and size and composition of samples analyzed, it was expected
that the results for the middle age group (15-44) from the studies should allow comparison since
all analyses were conducted with equivalent methodology. The estimated concentrations in the
VA/EPA  were  two to  three  times higher than the corresponding  FY87 NHATS  estimates.
Although no formal statistical tests were performed on these data, the differences  are clearly
                                         20

-------
significant.  The average values for the compounds across collection year, particularly 2,3,7,8-
TCDD, indicate a decrease in adipose tissue residue levels.  These differences may be due to a
decrease in exposure levels through environmental pathways or commercial/consumer products
over time. Another possible explanation may be the result of differences in specimen storage time
between the VA/EPA (up to 16 years) and the FY87 NHATS (up to 2 years) studies.

2.3 RECOMMENDATIONS FOR FUTURE STUDIES

      Several recommendations, along with rationale, are presented for consideration for future
studies on PCDDs  and PCDFs in human adipose tissue.  Although the recommendations are
presented with respect to human adipose tissue, they also apply to considerations for use of blood
(whole, serum,  or plasma) as a means  to study the body burdens of these compounds in the
general U.S. population.

      Recommendation; Increase the number of composite samples and individual specimens
that are analyzed in future NHATS studies.

      Rationale; The FY87 NHATS produced estimates of national average concentration levels
of nine PCDDs and PCDFs and demonstrated that the concentrations of these chemicals increase
significantly with the age of the donor.  These conclusions were based on the chemical analysis
of 48 composite samples containing an average of 18 specimens each.  Although there were
sufficient data to achieve these important objectives, the small number of composites analyzed
presents significant challenges to achieving other NHATS objectives. For example, it is very
difficult to estimate prevalence  with composite  sample  data.  Prevalence is established by
estimating the percent of the population with chemical concentrations  above specified levels.
Estimating prevalence requires that individual specimens be analyzed to characterize the statistical
distribution of concentrations among individuals in the population.

      Another reason for increasing the number of specimens is to detect smaller but important
differences in average concentration levels among geographic regions, race groups, and sexes that
cannot be  identified statistically because of sampling and measurement errors.  Statistical power
calculations will be needed to establish the optimal allocation of individual versus  composite
samples to be analyzed.

      Recommendation; Conduct an analysis of  PCDDs and PCDFs in human adipose tissue
using a similar study design at intervals of one to five years.

      Rationale; One of the objectives of NHATS is to estimate time trends in the levels of toxic
chemicals.  Significant trends require data from at least three to five time periods before they can
                                         21

-------
be demonstrated.  Considering the promulgation of specific regulations to limit releases of these
compounds  from a  variety of sources (incineration emissions, pulp  and paper effluents,
halogenated commercial products, etc) it is expected that further changes in residue levels may
be observed.  Correlation of declining residue levels with the promulgation of the residue levels
over time may provide the Agency with a good measure of the efficacy of environmental
regulation.

     Recommendation; Establish and characterize QC samples that are representative of human
adipose tissue levels.

     Rationale; In order to ensure that an accurate assessment of data comparability can be
achieved from one analysis event to the next, it is essential that appropriate reference materials
are available for incorporation into the study design.  Adipose tissue pools representative of the
three age categories could be prepared in sufficient quantity to allow multiple determinations in
subsequent NHATS analysis efforts.  The QC pools could be stored under the same conditions
as the NHATS repository and could be used to address storage stability issues.
                                          22

-------
3.0 OVERALL DATA QUALITY

      The overall quality of results in the FY87 NHATS was achieved by performing several
critical activities:  planning a well-designed study, validating methods in advance, controlling
implementation, verifying assumptions, and documenting procedures.  At least one  of these
activities was involved in every step of the study.  Figure 3-1 shows the steps of the study and
illustrates how they are related.

      The steps are divided into the following five stages:

          Sampling,
          Compositing,
          Chemical analysis,
          Data management and processing, and
          Statistical Analysis.

Each  stage involved a design  or a plan, a set of controls applied  on implementation, and
verification and documentation procedures that were followed.  Subsections 3.1 to 3.5 describe,
for each stage, the specific activities  that were carried out.

3.1 SAMPLING

      A complete discussion of the  FY87 NHATS study sampling design is contained  in
Section 4.0 of this report and has been fully documented (Panebianco DL, 1986a).  The design
followed a multistage process in  which (1)  the  conterminous 48 states were divided into
geographic strata, (2)  Metropolitan Statistical Areas (MSAs)  were randomly sampled within
each  stratum, and  (3) cooperators were chosen in each MSA to provide  the  specimens.
Finally, each cooperator was given quotas, derived from the 1980 U.S. Census, on the number
of specimens to collect from each age group, sex, and race  group.

      The cooperators were given guidelines on the types of patients and cadavers which can
contribute specimens  to  the program.  The relevant  donor information was recorded on
Patient Summary Reports (PSRs).  Then, matching  bar  code  labels containing sample IDs
were affixed to each PSR and the corresponding sample vial.  Each PSR was then reviewed by
the laboratory  responsible for sample  collection to  determine  whether  the cooperators
followed the guidelines in selecting appropriate donors.
                                         23

-------
          Data Quality
          Objectives
           QC Plan
Sample Design
    Sample
   Collection
                               Composite
                                 Design
                               Compositing
   Chemical
   Analysis
                            /     Outlier
                                 Analysis
                                                     |
                                                     \
 PSR Data
               7^
         'ting  /	H
Compositing
    Data
                          Analysis Plan
              QC Data
              Analysis
             Statistical
              Analysis
90-48 SEVslanflow 111900
          Model
         Validation
                                           Results
           Figure 3-1.   Program steps designed to ensure overall program quality.
                                          24

-------
      The sampling design called for collecting 1377 individual specimens with specific quotas
for each combination of geographic region and donor age group, race, and sex. Forty one (41)
of the 47 MSAs specified in the sample design contributed a total of 956 specimens. Specimens
collected within the overall design quota for each MSA were designated as design specimens,
and the remaining specimens were designated as surplus.  There were 865 design specimens
in  the  FY87 NHATS.   As  discussed in Section 4.0, the  marginal percentages  of design
specimens  from donors in each geographic and demographic category were found to be
consistent with the sampling design.

3.2 COMPOSITING

      Composite design criteria were developed to ensure  that the data obtained  from the
composites could be interpreted for the populations of interest and that the results could be
compared with results from the  FY82 NHATS study.  Section 5.0 discusses these design
criteria.

      The  composite design criteria  specified that only specimens from the same Census
division and age group were to be composited together. Also, in FY87 special  efforts were
made to create  more composites containing specimens from  donors of the same sex than was
achieved in FY82. For example, 24% of the FY82 composites  were pure  sex (either all males or
all females) composites  compared with 64% of the composites prepared for the FY87 samples.
It was  not  always  possible  to form  composites  containing specimens from the same race
group.

      The  procedures  for preparing  the composite  samples  are  presented in  Section 5.2.
Quality assurance/quality control practices that were  implemented  during compositing were
focused on procedures  to assure that  the study design was  accurately  followed and reported
and that the individual specimens were composited in a manner that  preserved sample and
specimen integrity and avoided contribution from laboratory background.

      Battelle provided  MRI the study design, which identified  each individual specimen that
was to be added to form a single composite.  A form was provided for each composite sample
and was used  to  document the specific compositing activities.   This form included  the
identification of each individual specimen and demographic information that was  used to
confirm the identity of  each specimen. All specimens were retrieved from the repository and
grouped according to the composite design.  The specimens were examined to  determine if
sufficient tissue was present to  meet the composite  design.   The EPA work assignment
managers were  notified of potential problems, which were resolved prior to proceeding with
the physical compositing.
                                         25

-------
     As  the composite samples were  prepared, the mass of individual specimens were
recorded  on the composite forms.   All composite  forms were compiled as a  report and
submitted to EPA and Battelle.  Data from the compositing reports were keyed to  a computer
file using double data entry. The data were stored in the FOCUS database system which was
also used to perform range and logic checks.

3.3 DEMONSTRATED LABORATORY QUALITY CONTROL

     At the outset of the analysis effort for the FY87 NHATS composite samples for PCDDs
and PCDFs, specific  data quality objectives were developed and included  in  the  Quality
Assurance Project Plan.  These data quality objectives were identified for calibration  criteria
(relative response factors [RRFs]) for each analyte and internal  standard,  HRGC  column
performance, HRMS resolution, and method performance based on the recoveries of internal
quantitation standards (IQS) and compounds spiked into a spiked internal QC sample.

     Table 3-1  summarizes the performance achieved versus the specific criteria and data
quality objectives for the analysis of the FY87 NHATS composites.

              Table 3-1.  Summary of DQOs for PCDD/PCDF Analysis Effort
               Criteria
   Objective
           Actual
       RRF calibration
       TCDD isomer separation

       Mass resolution

       Internal quantitation
       standards (IQS)
       Spiked internal QC
       samples
20%
TCDD/TCDF
30% all other
<; 25% valley/-
valley
HRMS * 10,000
40%-150%
40%-150%
91% (378 of 416) of all RRF
factors within DQOs

Achieved for all analyses days

Maintained and documented all
analysis days
96% for all IQS standards
(585 of 612); 2.5% (17) of
deviation due to "C-OCDD;
1.5%  (9) due to double-spiked
sample
97.5% of all measurements
within criteria (4 of 160
measurements outside DQOs)
                                         26

-------
      A QC plan was proposed by Heath (1988) to monitor and characterize the performance
 of the analytical method.  The plan called for the analysis of a method blank, control sample,
 and two spiked samples in each of the five analysis batches.  The control and spiked samples
 were formed using a composite sample of a large number of surplus adipose tissue specimens.
 The method blank was  analyzed  first to confirm that the  sample analytes  would not be
 compromised by  laboratory  background or calibration  standard carryover, then the control
 and spiked samples were analyzed  in random order along with the eight to 10 study samples.
 The objective of  the QC plan was to  provide  data that could be used to characterize the
 method's precision and accuracy and to identify any significant batch effects.

      The statistical analysis of the QC data involved modeling the measured concentrations
 versus spike level. Estimates of background levels, batch effects, and method recoveries were
 established for all of the target chemicals. Also, statistical hypothesis tests were conducted to
 determine  if there  were significant batch  effects  or if  the estimated  recoveries  were
 significantly less than 100%.

      Batch effects were accounted for by including their effect in the estimated measurement
 precision.  Precision,  as determined by the coefficient of variation  at the control sample
 concentration level, was estimated to be 5% to 20% for the dioxin compounds and 13% to 48%
 for the furans.

 3.4 STATISTICAL ANALYSIS

      Prior to conducting the statistical analysis, a formal analysis plan was prepared to define
 the statistical objectives and  to specify the methods and assumptions  that  would be used to
 achieve these objectives.  The  primary objectives were to estimate average concentrations in
 the U.S. population and in specified marginal  populations,  and to determine if  there  are
 significant effects  associated  with geographic and demographic factors.  Section 7.0 discusses
 these objectives and describes the statistical analysis approach.

      The statistical model and analysis approach used for the FY87 NHATS were  developed
 specifically for the unique problem of analyzing composite sample data.  They were  developed
 and evaluated in a separate study by Orban and Lordo (1989) in which composite  data were
 simulated from actual specimen data obtained in the FY83 NHATS.  Their study demonstrated
 the validity of the analysis approach.  They also  developed and documented special computer
 programs to implement the analysis.

      Before  the  statistical analysis was  conducted,  the  data were classified according to
restrictions imposed by the predetermined data quality objectives (DQOs).  Only data which
                                          27

-------
met all of the DQOs were included in the analysis. This ensured that the reported results are
not affected by biases attributed to the analytical method.  Furthermore, the formal statistical
analysis  to  estimate  the population averages was only  performed  when  there were  a
minimum of 30 unrestricted composite measurements,  and the chemical was detected in  at
least 50% of the  samples.  Nine of the  target compounds met these  minimum criteria for
performing the statistical analyses.  Summary statistics  are provided for the chemicals  which
did not meet these criteria.

     After fitting the statistical model to  the data, the model assumptions were checked
through  analysis  of  residuals  (i.e.,  the  difference   between  predicted  and  measured
concentrations), probability plots, and correlations.  These analyses demonstrated  that the
model assumptions were reasonable.  Furthermore, as  discussed in Section 8.0, eight  of the
nine chemicals statistically analyzed had R-squared correlations exceeding 68%.  That is,  at
least 68% of the variability in the data could be explained by the assumed model.
                                          28

-------
4.0 NHATS SAMPLE DESIGN

4.1 SAMPLING DESIGN

      The human adipose tissue specimens analyzed in the FY87 NHATS were collected from
October 1986 through September  1987, following the NHATS sampling design.   The NHATS
program uses a statistically based survey design to obtain adipose tissue specimens  from
autopsied cadavers and  surgical  patients.  Although the NHATS target  population is the
general, noninstitutionalized U.S.  population,  the sampling population is limited to cadavers
and surgical patients due to the invasive nature of the process required to collect the adipose
specimens from living persons.

      Each year approximately 800-1200 adipose tissue specimens are collected using a multi-
stage sampling plan.  The 48 conterminous states are first stratified into geographical areas.
From the set of strata, a sample of Metropolitan Statistical Areas (MSAs) is  selected  with
probabilities proportional to MSA  population.  One or more cooperators (hospital pathologists
or medical examiners) are chosen from each  MSA and asked to  supply a  specified quota of
tissue specimens.  Each year an effort is made to retain the same MSAs and cooperators used
in previous years.

      Cooperators  are given overall quotas, as well as subquotas for supplying specimens in
categories defined by the donor's age group, race, and sex. The categories are

      Age Group:      0-14 Years, 15-44 Years, and 45+ Years;
      Race:            Caucasian and non-Caucasian; and
      Sex:             Male and  Female.

The subquotas are proportional to the 1980 U.S. census population counts for each sampling
stratum. The donors and tissue specimens are selected in a nonprobabilistic manner by the
cooperators.

      Because the  survey requires some divergence from strict probabilistic sampling, the
validity of the statistical estimates derived from the data depends on several assumptions.
First,  the concentrations  of toxic  substances in the  adipose  tissue of  cadavers  and surgical
patients  are assumed to be comparable to those  in the general  population.  Second, it is
assumed that the levels of toxic substances in urban residents are approximately the same as
in rural residents, and therefore the selection  of only urban hospitals and medical examiners
(i.e., those located  in MSAs) does not introduce any significant bias into the  estimates of
average concentration levels.  Finally, it is assumed that no systematic bias is introduced by
                                          29

-------
the fact that the cooperators are not randomly selected, and that the donors and specimens
are nonprobabilistically sampled according to pre-spedfied quotas.

     4.1.1  Selection of MSAs

     Prior to 1985, the sampling strata from which MSAs were randomly selected were the
nine U.S. Census divisions.  But in 1985 EPA wanted the ability to obtain estimates of average
concentration levels in each of  the  ten  EPA regions.  Therefore, beginning with the FY85
NHATS, sampling strata were redefined to be the seventeen geographic areas of the  country
which resulted from  the intersection of the nine Census divisions and the ten EPA regions
(Panebianco DL, 1986a).   Selecting  the  sample in this manner made it  possible to  make
comparisons with previous NHATS results and also obtain estimates for the EPA regions. The
states, Census divisions, and EPA regions  that define the seventeen  strata are shown in
Table 4-1.

     Beginning with the FY85 NHATS, a new sample  of MSAs was selected from the  set of
seventeen strata using a  controlled selection technique, known as  the  Keyfitz technique,
which maximized the probability of retaining the MSAs used in the FY84 design (Mack et al.
1984). The MSA sample selected in FY85 has served as  the base NHATS sample for FY86 and
FY87. The FY87 NHATS sampling  design contains forty-seven MSAs.  Thirty-eight of the FY87
MSAs were selected in the FY85 sample. The remaining nine FY87 MSAs are replacements.
Some of the FY85 sample MSAs were replaced because satisfactory cooperators could not be
found.

     The forty-seven MSAs included  in the FY87 NHATS design are listed by stratum in
Table 4-2.  Four MSAs (Los Angeles, Chicago,  Detroit, and New York) are listed as double
collection sites because their populations are much larger than the populations  of the other
MSAs.  While given positive probability of selection, no MSAs were selected from strata 13, 15,
and 17 because of the small population size of these strata.

     4.1.2  Collection Quotas

     Sampling within MSAs is done by quota.  A quota  of specimens is assigned  to each
sample MSA.  In addition, demographic subquotas are assigned to  each MSA so that the
specimens collected will be representative of the strata with respect to the  three demographic
factors:   age group, race, and sex.   The  subquota  for each MSA is determined by the
demographic makeup of the stratum to which the MSA belongs and is based on the 1980 U.S.
Census  population  counts.   Each  combination of age group and  sex is  proportionally
represented in the subquota.  The race categories are also proportionally represented, but the
                                         30

-------
               Table 4-1.  Summary of Sampling Strategy
Stratum
      Census
      division
 EPA
region
       State
   3

   4
   8

   9
  10
                New England
Middle Atlantic


Middle Atlantic
South Atlantic
                South Atlantic
                East South Central
                East North Central
West North Central

West South Central
West North Central
  3

  3
  5

  6
Maine
New Hampshire
Vermont
Massachusetts
Rhode Island
Connecticut

New York
New Jersey

Pennsylvania

Delaware
Maryland
District of
  Columbia
Virginia
West Virginia

North Carolina
South Carolina
Georgia
Florida

Kentucky
Tennessee
Alabama
Mississippi

Ohio
Indiana
Illinois
Michigan
Wisconsin

Minnesota

Arkansas
Louisiana
Oklahoma
Texas

Iowa
Missouri
Nebraska
Kansas
                                31

-------
                       Table 4-1 (Continued)
Stratum
      Census
      division
 EPA
region
       State
  11

  12
  13


  14

  15

  16


  17
West North Central
Mountain
  8
Mountain

Pacific

Mountain
Pacific

Mountain
  9

  10

  10
North Dakota
South Dakota

Montana
Wyoming
Colorado
Utah

Arizona
Nevada

California

Idaho

Washington
Oregon

New Mexico
                                32

-------
              Table 4-2.  Sample MSAs for Fiscal Year 1987
                National Human Adipose Tissue Survey
Stratum
    Census
   division
 EPA
region
        MSA
   1

   2
   4

   5
 New England
Middle Atlantic
Middle Atlantic


South Atlantic

South Atlantic
                  East South Central
                 East North Central
  1
  3

  4
                 West North Central
Springfield, MA
Boston, MA

Syracuse, NY(b)
New York, NY
-------
                                Table 4-2 (Continued)
        Stratum
      Census
     division
 EPA
region
MSA
           10



           11

           12


           14



           16
                          West South Central
West North Central


West North Central

    Mountain

      Pacific


      Pacific
  8

  8
  10
New Orleans, LA
 Browns ville-
Harlingen-San
Benito, TX^
Houston, TX
San Antonio, TX
Dallas, TX

Omaha, NE
St. Louis, MO
Wichita, KS 

Grand Forks, ND0"

Salt Lake City, UT
Denver, CO

San Francisco,  CA
Sacramento, CA
Los Angeles, CA(a)

Portland, OR
Spokane, WA
Olympia, WA*1
Tacoma, WAW
(a) Indicates a double collection site.   A double collection site is an MSA whose population
  relative to its stratum is so large that its proper representation in the sample requires it to
  be selected twice.

^ Indicates replacement MSA.  A replacement MSA is an MSA that was not selected in the
  FY85  probability sample, but was  chosen to  replace an FY85 sample  MSA for which a
  satisfactory cooperator could not be found.
                                         34

-------
 subquota does not specify that Caucasians and  non-Caucasians are to be proportionally
 represented within each combination of age group and sex.  The subquotas only specify the
 total number of Caucasian and non-Caucasian specimens to be collected from each MSA.

      The subquotas for  the seventeen sampling strata  are presented in Table 4-3.  A total
 quota  of twenty-seven specimens was specified  for  each MSA, except those that were
 designated as double collection sites. In those MSAs the quotas and subquotas were doubled.
 Cooperators within an MSA were assigned quotas and subquotas appropriate to that MSA.

      The total number of samples specified for the FY87 NHATS was  1377.  This is based on
 the quota of 27 specimens for each of the 47 MSAs plus 27 additional specimens for each of
 the four MSAs designated as double collection MSAs.

 4.2 SAMPLE COLLECTION PROCEDURES

      NHATS specimens are adipose tissue samples excised by  pathologists and  medical
 examiners during therapeutic or elective surgery or during postmortem examinations.  If the
 specimen was collected postmortem, the tissue was obtained from an unembalmed  cadaver
 which had been dead for less than twenty-four hours and had been kept under refrigeration
 during that time. The  death should have been caused by sudden traumatic injury,  such as
 cardiac arrest, car accident, or gunshot wound.

      The following groups were excluded from specimen collection:

          Institutionalized individuals;

          Persons known to be occupationally exposed to toxic chemicals;

          Persons who  died of pesticide poisoning; and

          Persons suffering from cachexia.

 These guidelines were  stipulated so that the levels of substances detected in the specimens
were a result of environmental exposure.

     All NHATS cooperators in the selected MSAs were provided with target quotas for
specimen collections from age, sex and race groups. The cooperators were asked to obtain at
least five grains  of tissue  from each donor.   The cooperators were  instructed to  avoid
contamination through  contact  with  disinfectants,  parafins,  plastics,  preservatives  and
solvents. After collection the specimens were placed in glass jars with Teflon lids and frozen
                                         35

-------
Table 4-3.  Age, Race, and Sex Subquotas for each NHATS Collection
               Site within a StratumFY87 Design
0-14 Yr
Stratum
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Census division
New England
Middle Atlantic
Middle Atlantic
South Atlantic
South Atlantic
East South Central
East North Central
West North Central
West South Central
West North Central
West North Central
Mountain
Mountain
Pacific
Mountain
Pacific
Mountain
EPA
region
1
2
3
3
4
4
5
5
6
7
8
8
9
9
10
10
6
No. of non-
Caucasians
2
5
3
6
6
5
4
1
6
2
2
2
4
7
1
2
7
Male
3
3
3
3
3
3
3
3
4
3
3
3
3
3
4
3
4
Female
3
3
3
3
3
3
3
3
3
3
3
3
3
3
4
3
3
15-44 Yr
Male
6
6
6
6
6
6
6
6
6
6
6
7
6
6
6
7
6
Female
6
6
6
7
6
6
6
6
6
6
6
7
6
7
6
7
7
45+ Yr
Male
4
4
4
4
4
4
4
4
4
4
4
3
4
4
3
3
3
Female
5
5
5
4
5
5
5
5
4
5
5
4
5
4
4
4
4
                             36

-------
at -10 to -20C.  These jars were packed on dry ice for overnight shipment to MRI.  Upon
receipt at MRI the specimens were checked to determine that they were intact and frozen.
The specimens were checked versus the cooperators quota, an approximate specimen weight
was determined and the specimens were transferred to storage at -20C.  As part of the check-
in process, the patient summary reports (PSRs) were cross checked for consistency with the
specimen labels and identification data.  The PSRs were forwarded to Battelle for processing.

4.3 SAMPLE COLLECTION SUMMARY

      In FY87, 956 specimens were collected  from the cooperators.   Of these, 771  were
collected in accordance with the quotas and subquotas.  The rest of the  specimens were
collected by cooperators, but went beyond what the quotas and subquotas requested.

      In designating which collected specimens to chemically analyze, EPA decided to include
some specimens  that  were collected outside  of  the  subquotas.    This  was  done in  the
anticipation of the use of the linear model,  with parameters for the effects of Census region,
age group, sex, and race group, to statistically analyze the results.  The maximum number of
specimens from a MSA remained at the original quota of twenty-seven or fifty-four.  In effect,
the MSA quotas were maintained, but the MSA subquotas were allowed to vary from what
was specified in the design.  This approach, while not ideal, was the best means available to
deal with  MSA non-response.   Of the  956 collected specimens, 865  were designated for
chemical analysis. These 865 specimens were labeled "Design" specimens.

      Table 4-4 is a summary of the collection effort for the FY87  NHATS. Table 4-4 shows
collection information  for the nine Census divisions.   In FY87,  EPA chose not  to  make
estimates for  EPA  regions.  Instead,  EPA maintained  similarity to  the  FY82  geographic
classifications in order to compare FY87 results to FY82 results.

      Table 4-5 shows the  number  of quota  specimens,  collected  specimens, and Design
specimens in each category defined by the four analysis factors in  the linear model.  Because
the number of samples in the chemical analysis  was not large  enough to obtain reliable
estimates for all nine Census divisions, the divisions were combined into four Census regions
for both the FY82  and FY87 model analyses.
                                          37

-------






b
re
I
3
en
1
u
CD
13
0
U
en
H
"^l
rM
2
So
^ "g
III
S a> C
a rt tu
2 *
'S-aS
X-i QJ ni
nj + p
111
11 8,
2 


MH OJO
o c
H 'J3 e
 re cu
QJ 4-. g
^ 9 
Si S 3 -Q
 3 ^ S
M-' ^ CU
. 2 
t
Tp-
~
"
re
H









1



^4 (
^
! QJ ,
t-* ,-~^ c/i
QJ fZ -^
^r MH 
-------
Table 4-5. FY87 NHATS Sample Sizes by Categories
Analysis
factor
Census
Region



Age



Sex


Race


Number of Number of Number of
quota collected design
Category specimens specimens specimens
Northeast
North Central
South
West
Total
0-14 years
15-44 years
45+ years
Total
Male
Female
Total
Caucasian
Non-Caucasian
Total
270
405
432
270
1,377
311
630
436
1,377
668
709
1,377
1,157
220
1,377
195
307
349
105
956
163
353
440
956
499
457
956
776
180
956
175
296
289
105
865
146
318
401
865
436
429
865
707
158
865
                     39

-------
5.0 COMPOSITE DESIGN

      The 865 design specimens in the FY87 NHATS were  assigned to composite samples
using specific composite design criteria (Leczynski et al. 1988).  The reasons for compositing
samples prior to chemical analysis were: (1) at least 10 grams of tissue were needed to meet
the limit of detection goals for the target compounds, and (2) the budget for chemical analysis
of samples could only support the analysis of 48 samples.

5.1 DESIGN GOALS AND COMPOSITING CRITERIA

      The seven goals of the FY87 composite design, listed in order of importance, were

          Create no more than 48 composite samples.
          Maintain similarity to the FY82 composite design.
          Maintain equal tissue mass of individual specimens within the composite samples.
          Specify more pure sex composite samples than in FY82.
           Control the MSA effect.
          Provide the best range of race group percentages across the composite samples.
          Maintain a constant number of specimens across all composite samples.

      Because of the constraints imposed by the sampling and compositing protocols and the
frequency of collection nonresponse, it was not possible to meet all of the design goals.  Each
of the last six goals required a different mix of individual specimens within the composite
samples.   Thus, attempts  were made  to achieve all goals across the  design to the extent
possible.  The criteria used  to design composite samples are discussed below.

      (1)   Create no more than 48 composite samples.

      This criterion was based entirely on the funds available for the chemical analysis.

      (2)   Maintain similarity to FY82 composite design.

      This specification was included to ensure that comparisons between the analysis results
from the two years could  be made.    The  design criterion  imposed by this objective is that
each composite sample had to be constructed from individual specimens collected from exactly
one Census division and  exactly one  age group.  Thus,  there were 27 distinct categories
defined by the intersection  of the nine Census divisions and  three age groups.
                                         41

-------
      (3)   Maintain equal weighing of specimens within the composite samples.

      The interpretation of the observed concentrations of the composite samples is far easier
when these concentrations can be interpreted as the arithmetic average of the concentrations
of the individual specimens.  Therefore,  this design goal  specified that  each  individual
specimen within a composite sample contribute an  equal amount of tissue to that composite
sample. This specification allows the lipid adjusted concentration of the composite sample to
be interpreted  as approximately the arithmetic average  of  the lipid adjusted  individual
specimen concentrations, with equality whenever all the specimens in the composite sample
have the same percentage of lipid material.

      (4)   Construct more pure sex composites than in FY82.

      Pure male or pure female composite samples were constructed when sufficient numbers
of specimens were available within a particular Census division/age group category. Pure sex
composite samples are samples in which all of the  individual specimens were collected from
donors of  the same sex. Such composites are needed to achieve more precise estimates of sex
effects in the population.  It was anticipated that including more pure sex composite samples
in the FY87 design will lead to smaller standard errors for the sex group estimates (Draper and
Smith 1981, pp. 52-55) than was the case in FY82.

      (5)   Control the effect of the MSAs contributing specimens to each composite.

      Controlling the  number of MSAs  contributing specimens to  composite samples is
intended to reduce the effect of the MSA on the estimated average concentrations.  This is
done  because  MSAs  are  regarded as  being  major sources  of differences  in  observed
concentrations across the nation due to their varied exposure scenarios (Panebianco 1986b).
To avoid confounding the MSA effect with any of the geographic or demographic effects, two
design criteria were identified. They were

      (5-a)      Keep  the number of MSAs represented in each composite sample consistent
                across the design (2-3 MSAs was the target number),  and

      (5-b)      Maintain approximately the same number of  pure  male  and pure female
                composite samples within a group of MSAs.

The first criterion helps to  ensure a constant variance  of measured concentrations across the
sample whenever the composite sample concentrations are averages over an equal  number of
MSAs. The second criterion  is intended to  prevent confounding a large MSA effect with the
sex effect.
                                         42

-------
      (6)   Vary  the  percentage  of  Caucasian and  non-Caucasian specimens  in  the
           composites as much as possible.

      For the  same reasons it is important  to construct pure sex composite samples, it is
important to construct pure race group composite samples. However, this  goal is dependent
on the number of non-Caucasian specimens collected in the twenty-seven Census division/age
group categories and the number of composite samples in the design.  Since the number of
non-Caucasian specimens collected in the FY87 sample was relatively small, it was decided to
provide the best possible range of race group percentages (i.e., mixes of the Caucasian and
non-Caucasian specimens within the composite samples) across the design rather than focus
on designing pure race group composite samples.

      (7)   Maintain a constant number of specimens across all composite samples.

      This goal, similar to goal 5-a above, could not be fully achieved for the FY87 composite
samples.

5.2 LABORATORY COMPOSITING PROCEDURES

      The FY87 NHATS specimens from nine census divisions and three  age groups were
divided into 48 composites, as  identified in the composite  design  provided  by Battelle-
Columbus Laboratories (Battelle).  Battelle provided MRI with the data sheets  that identified
the individual  specimens and their required weights to be included in each composite.  Each
composite consisted of from 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 Battelle   were  used as work sheets to  record  the  actual laboratory
compositing procedures.

      Initially,  the samples were grouped into composites, and any samples of questionable
weights were  noted.   Three samples  identified  below  did not appear to have the weight
required by the composite design. These results were relayed  to the EPA Work Assignment
Manager (WAM).   After consultation  with  DDE,  the  EPA WAM forwarded to MRI  the
following responses regarding the problem samples on June 24,1988.
                                         43

-------
  Composite number      Sample number          Problem              Response
     ACD8700023            8706954        Low weight, - 0.1 g    Include as is
                                           need 0.5 g
     ACD8700032            8701765        No sample            Omit
     ACD8700201            8703464        Low weight, - 1.3 g    Include as is
                                           need 2.0 g
      The specimens in a composite were placed on dry ice during the compositing procedure.
An electronic four-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 specimen jar was opened, and a portion of the
frozen adipose tissue removed with a clean stainless steel spatula.  The adipose tissue  was
placed in the culture tube and the weight recorded to three decimal  places on the compositing
sheets.  Additional adipose tissue  was added  if necessary.  A goal of  10% of the  desired
weight was attempted where possible.  The specimen jar was capped and returned to storage.
The  weights of the individual specimens were  recorded on the  data sheets  provided by
Battelle.

      The weight of the culture, beaker, and adipose tissue was rezeroed, and  the next
specimen  in the composite  was weighed.  A new spatula was used for each sample.  This
procedure was repeated for each sample  in the composite.  When the composite was  com-
pleted, it was capped and stored in a sample freezer at -10 to -20C.  All data on the actual
compositing procedures were recorded on the data sheets provided by Battelle. All data sheets
were submitted in a separate report to document the compositing activity (Cramer and Stanley
1988).
                                         44

-------
5.3  SUMMARY OF COMPOSITE SAMPLES

     The FY87 NHATS Composite  Design resulted  in  the  construction  of  48  composite
samples, using the 865 design specimens collected from 41 MSAs.  Table 5-1 shows the number
of composite samples for the 27 distinct combinations of Census division and age group. The
sex  and race group percentages of the composite samples vary across the design depending on
the  availability of specimens within specific demographic subpopulations. Table  5-2  shows the
demographic makeup of the FY87 NHATS composite samples.

     The 48 composite samples were randomly assigned to five batches. Within batches, the
composite samples were placed in random order for the chemical analysis.
                                        45

-------
Table 5-1. Distribution of FY87 NHATS Composite Samples
          by Census Division and Age Group
Age group
Census division
New England
Middle Atlantic
South Atlantic
East South Central
West South Central
East North Central
West North Central
Mountain
Pacific
Total
0-14 Years
1
1
2
1
1
2
1
1
1
11
15-44 Years
1
3
4
1
1
3
2
1
1
17
45+ Years
1
2
4
1
1
5
2
1
3
20
Total
3
6
10
3
3
10
5
3
5
48
                        46

-------

3
ex
Si
1

8
1
z
00
PL,
8
a,

2
u

a.
S
1
g
D
s
23
H








Percent
Caucasian


S-S
rj tU

-------
"2
3
J
in

                838088
5J8S
                                                    rH O
                                      IXOVT^ONOO  inM
-------
6.0 CHEMICAL ANALYSIS PROCEDURES AND QUALITY CONTROL DATA

       The 48 composite samples were prepared in the analysis laboratory for determination of
low pg/g (ppt) levels of PCDDs and PCDFs using high-resolution gas chromatography/high-
resolution mass  spectrometry  (HRGC/HRMS).  The  performance  of  the analysis effort was
demonstrated through the determinations of PCDDs and PCDFs in 20 quality control samples
(method  blanks, controls,  and spiked tissues).   Chemical  analysis  performance was also
documented through participation in an interlaboratory effort with two other laboratories
recognized for their expertise in the determination of PCDDs and PCDFs in human tissues. This
section describes the analytical methodology and presents the results (analytical and statistical)
for the quality control samples.  A detailed presentation of the analytical results for all PCDDs and
PCDFs in the FY87  NHATS  design samples and the quality  control samples is  given  in
Appendix A.

6.1 SAMPLE PREPARATION

       The preparation of the composited adipose tissue specimens for determination of PCDDs
and PCDFs required a multistep procedure, which included quantitative extraction and cleanup
through several chromatographic columns. The procedures described below were carried out for
each of the five sample batches for the FY87 NHATS study.

       6.1.1  Extraction

      After compositing, the adipose tissue samples were stored at -20C in 50-mL culture tubes
sealed with aluminum foil.  The extraction procedure  was initiated  by  allowing the samples to
come to room temperature  and then fortifying them  with 100 fiL  of an internal quantitation
standard (IQS) spiking solution (Table 6-1) containing nine carbon-13 labeled PCDDs and PCDFs.
Ten milliliters of methylene chloride was added and the sample was 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 two to three times with fresh 10-mL portions 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 and
the final volume was brought to 100 mL.

      At this point the flask was stoppered and inverted several times to mix the extract.  Next,
a 1-mL aliquot was removed with a disposable graduated pipet and placed into a preweighed
(measured to 0.0001 g) 1-dram glass vial. The methylene chloride in the vial was reduced under
                                        49

-------
          Table 6-1.  Internal Standard Spiking Solution
                    for chlorinated Species3
                                              Concentration
                   Analyte                       (pg/\iL)
  Chlorinated Internal Quantitation Standards'1
  13C12-2,3,7,8-TCDD                                      5
  13C12-2,3,7,8-TCDF                                      5
  13C]2-l,2,3,7,8-PeCDD                                   5
  13Cirl,2,3,7,8-PeCDF                                    5
  13C12-l,2,3,6,7,8-HxCDD                                12.5
  13C12-l,2,3,6,7,8-HxCDF                                 12.5
  13C12-l,2,3,4,6,7,8-HpCDD                               12.5
  13C]2-l,2,3,W,8-HpCDF                               12.5
  13C]2-OCDD                                          25
  Internal Recovery Standard*"
  13C12-1,2,3,4-TCDD                                     50
  13C12-l,2,3,7,8-HxCDD	125
" All internal quantisation and recovery standards were obtained
 as solutions from Cambridge Isotope Laboratories (Woburn,
 Massachusetts).
b Prepared in isooctane. One hundred microliters spiked.  Separate
 solutions were used for chlorinated and brominated species.
c Prepared in tridecane.  Used for both chloro and bromo analyses.
                              50

-------
flowing nitrogen until a constant weight of lipid was obtained. The weight of the lipid was
obtained by difference, and the percent lipid for the composite was calculated and recorded.

       The remaining portion of the extract (99 mL) was quantitatively transferred, followed by
a 20- to 30-mL rinse, to a 500-mL round-bottomed flask. The extract was concentrated under
vacuum to an oily residue (extractable lipids) using rotary evaporation.

       6.1.2 Bulk Lipid Removal

       Separation and concentration of the PCDDs and PCDFs from the lipids to achieve a final
volume of 10  \iL is necessary to detect pg/g concentrations.  The extractable lipids from some of
the composites was as high as 9 g of oily materials.  The bulk lipid was removed following an
acidic silica gel slurry cleanup procedure . This was accomplished by adding 200 mL of hexane
and a Teflon-coated stirring bar to the lipid in the round-bottomed flask Then, while stirring the
extract on a magnetic stir plate, 100 g of 40% w/w sulfuric acid-impregnated silica gel was slowly
added to the  extract. The mixture was stirred for 2-hours.  During the 2-hour slurry period,
acid/neutral silica gel columns (4 g 40% H2SCVsi]ica gel, 1 g silica gel) were prepared.  After the
2-hour 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.
The slurry mixture was rinsed for 15 minutes with two additional aliquots (50 mL) of hexane.
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-Danish 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 adjusted to approximately 1 mL using nitrogen blowdown.

       6.1.3 Separation of Chemical Interferences

       Separation of chemical interferences, such as pesticides, PCBs, and other chlorinated planar
aromatics is essential to avoid false positive measurements. Removal of chemical interference was
achieved using two different chromatographic cleanup systems. The first was prepared as a
layered column containing 1 g sodium sulfate, 4 g neutral alumina, and 1 g sodium sulfate. The
1 mL 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% (volume/volume, v/v) methylene
chloride in hexane. These eluents were collected  and archived.   The PCDDs and PCDFs were
eluted from the column with 15 mL of 60%  (v/v) methylene chloride in  hexane and the eluent
concentrated under a stream of nitrogen to approximately 2 mL.
                                         51

-------
       A disposable column of AX-21 on silica gel was prepared and preeluted with 4 mL of
toluene, 2 mL of 75:20:5 methylene chloride/methanol/benzene, and 2 mL of 1:1 cyclohexane/
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 chloride/methanol/benzene. These eluents were combined and archived.  The
columns were turned upside down and the PCDDs and PCDFs eluted with 20 mL of toluene.
The extract was then reduced in volume to approximately 100 \iL, then 10 |iL of recovery standard
in tridecane was added (Table 6-1) and the volume was further reduced to 10 |iL under nitrogen.
The extract was stored in a freezer pending HRGC/HRMS analysis.

6.2  HRGC/HRMS ANALYSIS

       Initial calibration of the GC/MS system was conducted by making single 1-pL injections
of the standards listed in Table  6-2. Relative response factors calculated from this calibration effort
are  shown in Table  6-3. A CS7 (2.5 to 12.5 pg/|iL) standard was analyzed on a daily basis to
ensure adherence to the initial calibration curve.  The traceability and  comparability  of the
analytical  standards has been demonstrated in a previous NHATS effort (USEPA 1986)  and
through participation in an interlaboratory comparison study (Bradley, et al. 1990).

       HRGS/HRMS analysis of the samples 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.  If any evidence of system contamination was found, corrective
action was taken by analyzing another tridecane blank or cleaning the injection system. A typical
daily sequence of injections is shown in Table 6-4. A 1-fiL aliquot of the extracts was injected into
the GC/MS system, which was operated under the conditions that previously produced acceptable
results with the daily calibration standard.

       Selected ion monitoring (SIM)  data were acquired  according to the acquisition and MS
operating conditions previously used to determine the relative response factors  (Tables 6-5 and
6-6). Instrument performance was monitored by examining and recording the peak areas  for the
recovery standard, 13C]21,2,3,4-TCDD.  If this area decreased to less than 50% of the calibration
standard, sample analyses were stopped until the problem was identified and corrected.
                                          52

-------
              Cft
              U
                               rq N
                                                               vd
                                                                                   O
                                                                                                          IQ
              U
                                                                                   g g g g tg
                                                                                                                    o
PH
O
o
8
       f
              in
              Cft
              U
                                                                                                                    ir>rq
 in

 
43
 O
43
 2
>
U
I
I
 
 
u
rq
        o
        en

        O

       1
       .s
        o
       1
        u

        
       U
              eft
              u
              co
              tn
              U
              rq
              Cft
              U
              Cft
              U
                     SSS8S
                                                                                   m

                                                                                                                    ifits
                                                                                                                    ir>rq
                                                                                                                    IT>CN
 01

1
              in
              sf
              u
                                                                                   iftioinir>c

                                              4  tC
                                              mm
                                                                                                                    U U
                                                              53

-------
             in
                                rH  CO  in  ON  O^  rH  SO
                                                                                   _       rqONrn^ooco
                                            rHsOOOOONlNSOOOOOO^cqiAoinCNlCNlOOit'
                                            CNlCNlrHrHCNlinCNlCNlOON^;Tj

I
     U
             T*I
             CO
                ON  OO  ON rH  ^^
                tN  in  oo in  CT^
                OO  ^  SO CO  rH
                                                        00  CO
                                                        rH  CO
                                                                                               in
                                                                                                       2
                 rHOCDrHOCDOOrHrHrHrHrHrHrHCDO
 (0
O


 O

'a
U
     CO
     yi
     U
     CN1
     V3
     U
     C/3
     U
             O  CO  rH  CO C5  SO
             t^  CN  *^>  ^^
                        CD  ^  tx  ON  O
                    IN  so  rq  oo  in  rj<  oo
                    SO  SO  CO  rH  CO  rH
                                                fNONQCNlCOrHCOCOinoptN
                                                                              SO  in  O  CO  rH
             a
             rH
             CNl
                     rHOrHOOOOrHrHrH
                                       00  SO  OO
                                       CN  00  tN
                                       in  rn  cq
                                                                      fc  %
                                                                      in  IN
                                                                                                   rH  rH  rH  O
                               ON  rH  CNl
                               CO  OO  CN
                               o  cs  ci  ^  tx
                     ON  Uj rH  sO  CO  O  in  tN  CO
                     ON^msotN^cNicNio
                     CDOrHOCDCDCDrHrH
                                                                       r
                                                                OOsOONOOsOsOCXIONlN^CNlsO
                                                                inrHCNrHOinsOtN-'frHCOrH
so
     in
     U
CNl  OO  O  rH  CNJ  ON  "^ CD
ONinorHCNicNiinso
CNI  sO  rH  in  in  SO  SO CO
CNlrHrHOrHOeDCD
                                            00  IN
                                            rH  CO
                                                                               oo  m
                                                            CM
                                                                                       CO  rH  Q  M
                                                                                   . .  CNI  ON  O  IN
                                                                                   in  00  IN  SO  rH
                                                                                                       rH  CO
                                            OrHOrHiIrHi-HrHrHrHrHrH
     I
     75
        PH
PH  O  8
o  o   -
U  U

li?
CO  CO^  CNl
fsf  CNt  rn"
CDD
,8-Pe
                                    PH  O
                                    Q  G
                                    U  U
                                     CM  (X

                                  K  E
                                  oi  06
                                R. t\ K.
                                tN  SO  SO
                                SO- *.
                                                PH
                                                O
                                                u
                                                            PH  G
    PH
O  R
o  y
g3
s^s
^sXT^.'>1,"'v^vrNv-'SV1-H'--iT  ^J so] so] tNv TJ^ sov
                                                    tN.  CO.  TjJ.  CO.  CO. CO. ^J. CO. CO. CO.
                                                    co  C*Q  co  c*i  cq cq co
                                                    CNf  r-f  CNf  r-T  r-? r-I" C
-------
                         Table 6-4.  Typical Daily Sequence for PCDD/PCDF Analysis
/        1.            Tune and calibrate mass scale versus perfluorokerosene (PFK).

         2.            Inject column performance/window-defining mixture.

         3.            Inject concentration calibration solution 2.5 to 12.5 pg/fiL (CS-7) solution.

         4.            Inject blank (Tridecane).

         5.            Inject samples 1 through "N."

         6.            Inject concentration calibration solution 2.5 to 12.5 pg/|iL (CS-7) solution or other
                      concentration calibration solutions  CSl  to  CSS to bracket observed  sample
                      concentration.
                                                   55

-------
Table 6-5.  Ions Monitored for HRGC/HRMS of PCDD/PCDF
Descriptor
Al








A2










A3






ID
TCDF

13C12-TCDF

TCDD

13C12-TCDD
HxCDPE
PFK (lock mass)
TCDF

TCDD

PeCDF

13C12-PeCDF
PeCDD

13C12-PeCDD
PFK (lock mass)
HpCDPE
HxCDF

PFK (lock mass)
J3C12-HxCDF
HxCDD

13C12-HxCDD
OCDPE
Mass
303.902
305.899
315.942
317.939
319.897
321.894
331.937
333.934
377.886
380.976
303.902
305.899
319.897
321.894
339.863
341.860
351.900
353.894
355.858
357.855
367.895
369.889
380.976
409.877
373.821
375.818
380.976
383.861
385.858
389.816
391.813
401.856
403.853
445.866
Nominal dwell
time (s)
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.090
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.035
0.035
0.080
0.080
0.080
0.080
0.080
0.080
0.080
0.080
0.080
0.080
                        56

-------
Table 6-5  (continued)
Descriptor ID
A4 PFK (lock mass)
HxCDD

HpCDF

13C]2-HpCDF

HpCDD

13C12-HpCDD

13C12-HpCDD

NCDPE
A5 PFK (lock mass)
OCDF

I3C]2-OCDF

OCDD

I3C]2-OCDD

DCDPE
Mass
380.976
389.816
391.813
407.782
409.779
417.822
419.819
423.777
425.774
435.817
437.814
429.768
431.765
479.856
380.976
441.743
443.740
453.783
455.780
457.738
459.735
469.779
471.776
513.846
Nominal dwell
time (s)
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.06
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.06
         57

-------
  Table 6-6. HRGC/HRMS Operating Conditions for PCDD/PCDF Analysis
Mass spectrometer (Kratos MS50-TC)

Accelerating voltage:                    8,000 V
Trap current:                           500 pA
Electron energy:                        70 eV
Electron multiplier voltage:              -1,800 V
Source temperature:                    280C
Resolution:                            ;> 10,000 (10% valley definition)
Overall SIM cycle time:                  1 s

Gas chromatograph (Carlo Erba MFC-500")

Column coating:                        DB 5
Film thickness:                         0.25 jim
Column dimensions:                    60 m x 0.25 mm ID
He linear velocity:                      ~ 25 cm/s
He head pressure:                      1.75 kg/cm2 (25  psi)
Injection type:                         Splitiess, 45 s
Split flow:                             30 mL/min
Purge flow:                            6 mL/min
Injector temperature:                    270C
Interface temperature:                  300C
Injection size:                          1-2 nL
Initial temperature:                     200C
Initial time:                            2 min
Temperature program:                  200C to 270C  at 5C/min
Second hold time:                      10 min
Second temperature ramp:               270 to 330C at 5C/min
Final hold time:                        5 min
                                 58

-------
 6.3 OA/OC FOR CHEMICAL ANALYSIS

      The QA/QC  program for this analysis effort included:  demonstration of instrumental
 performance, routine analysis of quality control samples (method blank, controls and spiked
 tissues), analysis of performance audit samples and participation in an interlaboratory effort for
 comparison of results on specific samples. Each of these QA/QC efforts and results are discussed.

      6.3.1 Instrument Performance

      Instrument performance was characterized primarily by three criteria: (1) mass resolution
 (^ 10,000) and calibration; (2) relative response factors (RRF), i.e., adherence to the initial RRFs;
 and (3) column performance as indicated by peak separation between 2,3,7,8-TCDD and other
 TCDD isomers.

      6.3.1.1  Mass Resolution and Calibration

      The mass spectrometer was tuned on a daily basis to yield optimum sensitivity and peak
 shape  using an ion m/z 380.9760) from PFK.  The  resolution  was visually monitored  and
 maintained  at  ^ 10,000 (10% valley  definition) to provide adequate noise rejection while
 maintaining good  ion  transmission.  Static-resolving power checks were performed  at the
 beginning and at the end of each 12-hour operation period. A visual check (i.e., documentation
 was not required) of the static resolution was made by using the peak matching unit before and
 after each analysis.  Corrective action was implemented whenever the resolving power did not
 meet the criteria of ^ 10,000.

      Mass calibration of the mass spectrometer for the HRGC/MS analysis of PCDD/PCDF was
 conducted on a  daily basis.  The magnetic field was adjusted to pass m/z 300 at full accelerating
 voltage. PFK was admitted to the MS, and an accelerating voltage scan from 8,000 to 4,000 v was
 acquired by the data system. This corresponded to an effective mass range of 301 to 593 amu.
 Upon completion of a successful calibration step, the five ion descriptors shown in Table 6-5 were
 updated to reflect the new mass calibration.

      6.3.1.2 Relative Response Factor

     As part of the initial and routine instrument performance checks, calibration standards were
analyzed and the responses of the respective analytes were compared to specific internal stan-
dards  to establish the RRF values. The initial and routine calibration criteria required that the
precision of the  RRF measurements be within 20% for the tetrachloro congeners and within
 30% for the other compounds.
                                         59

-------
      Sensitivity of the HRMS was documented through the responses noted for  the first
calibration standard of each analysis day. The method required the analysis of a low level
standard (CS7) to document sufficient instrumental response to support instrumental detection
limits of 1 pg/fiL for TCDD.

      Routine checks on the instrument sensitivity, which were documented in the MS log book,
was achieved by monitoring the response for the internal recovery standard (13C12-1,2,3,4-TCDD)
from injection to injection. If the response for this standard dropped by greater than 50% of the
response noted in the previous calibration standard, the analyst verified instrumental performance
by analyzing an additional calibration standard. Additional details on the initial and routine
calibration events are presented in the data reports provided to EPA (Cramer et al. 1989a, 1989b).

      6.3.1.3 TCDD Peak Separation

      The HRGC column performance was demonstrated at the start of each 12-h analysis period.
This  was accomplished by injecting 1  jiL of the column performance/window-defining check
solution and acquiring SIM data for all PCDD and  PCDF compounds.  The HRGC column
performance was determined based on the ability to resolve 2,3,7,8-TCDD from possible coeluting
TCDDs.

      The chromatographic peak separation between 2,3,7,8-TCDD and the peaks representing
any other TCDD isomers was resolved with a valley of s 25%,  where

                                valley % = (*/y)(100)

x =   measured height of the valley between the chromatographic peak corresponding to 2,3,7,8-
      TCDD and the peak of the nearest TCDD isomer

y =   peak height of 2,3,7,8-TCDD

Figure 6-1 is an example of the separation of a TCDD isomer mixture and the calculation of
isomer resolution.  The TCDD isomer  resolution was documented to range from 18% to 25%
during the analysis effort (Cramer et al. 1989a, 1989b)

      6.3.2  QC Samples

      Samples  included for QC purposes are summarized in Table 6-7.  Each of these quality
control samples are described in further detail below.  These quality control  samples were
included  with  the analysis of the FY87 samples. The order of preparation and analysis with
respect to the FY87 NHATS composites was specified in the study design.

                                         60

-------
                                                  en


                                                  
                                                  I
                                                  tfl

                                                 Q
                                                 Q
                                                 U
61

-------
                           Table 6-7. Quality Control Samples
             Type                  Frequency                 Application
   Method blank              One per batch           Assess laboratory background
                                                      contribution
   Spiked control adipose       Two per batch (two      Evaluate method performance
   tissue sample               different spike levels)     (accuracy and precision)
   Unspiked control adipose    One per batch           Evaluate method performance
   tissue sample                                       (accuracy and precision)
6.3.2.1  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 carbon-13 labeled internal quantitation standards that were added to samples before bulk lipid
cleanup.  The five method blanks analyzed with the samples did not contain PCDDs or PCDFs
with the exception of the method blank generated for Batch 1 samples.  This  method blank
contained a trace of 2,3,7,8-TCDF which was determined to be equivalent to 0.46 pg/g of tissue.
The detailed analysis results for the method blanks are presented in Appendix A.

     6.3.2.2 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 (80C). The evaporation process was extended to ensure that 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. A summary of the QC sample results
is presented in Section 6.5. A detailed presentation of this data by analyte is given in Appendix A.

     6.3.2.3 Internal 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, one low
and one high, per sample batch.  The  native spiking solution concentrations are shown in
                                          62

-------
Table 6-8. Additional OCDD was added to each sample with a 75 pg/iiL spiking solution. Low
and high spike levels in the control samples are shown in Table 6-9.

      The spiking solutions were checked for accuracy prior to spiking the adipose composite
with the native isomers. The results of this spike check are shown in Table 6-10. The spike check
results for the separate OCDD spiking solution (needed to reach the higher level for OCDD in
the adipose tissue) are given in Table 6-11. A summary of the results from the analysis of these
spiked materials is presented in Section 6.5.  The results are presented for each spiked sample in
Appendix A.

      6.3.3 Performance Audit Samples (PAS)

       Performance audit samples (PAS) were submitted for analysis before the first sample was
analyzed with batch 1 and batch 5.  These samples consisted of unlabeled PCDDs and PCDFs in
a solvent matrix. The samples were prepared by the project quality control coordinator (QCC)
and turned over to project personnel who fortified the samples with IQS and RS solutions and
submitted the prepared sample for HRGC/HRMS analysis. The performance audit samples were
prepared from a mixture of standards used for an interlaboratory effort to establish consensus
values for concentrations  (Bradley et al 1989). Results from the analysis were given directly to
the QCC. Acceptability criteria were 70% to 130% accuracy for each of the isomers in the sample.

      Table 6-12 provides a summary of the results from the analysis  of the two performance
audit samples. As presented the measurements were within the criteria specified for each analyte.

      6.3.4 Interlaboratory Comparisons

       External QC samples and solutions were submitted to two outside laboratories.  The
contacts and  laboratories  were Dr.  Donald Patterson with the Centers for Disease Control in
Atlanta and Dr. John J. Ryan  with the Health Protection Branch in Canada. Each laboratory
received  one  spike check solution  sealed in an amber ampule  and three blind control lipid
samples.  The lipid samples were spiked identically to those used with the analysis of the FY87
NHATS samples  and included an unspiked sample, a low level spike sample, and a high level
spike sample.  Data from the interlaboratory comparison are presented in Tables 6-13 through
6-16.  These data demonstrate that  although some differences were apparent in the analytical
standards and hence the results for  the analysis of the quality control samples, the data for the
respective compounds from the laboratories are generally within 30% relative percent difference.
                                          63

-------
Table 6-8. PCDD and PCDF Native Spiking Solution"
                                               Concentration
 Analyte                                         (pg/jiL)
2,3,7,8-TCDD                                         5
2,3,7,8-TCDF                                          5
1,2,3,7,8-PeCDD                                       5
1,2,3,7,8-PeCDF                                       5
2,3,4,7,8-PeCDF                                       5
1,2,3,4,7,8-HxCDD                                    12.5
1,2,3,6,7,8-HxCDD                                    12.5
1,2,3,7,8,9-HxCDD                                    12.5
1,2,3,4,7,8-HxCDF                                    12.5
1,2,3,6,7,8-HxCDF                                    12.5
1,2,3,7,8,9-HxCDF                                    12.5
2,3,4,6,7,8-HxCDF                                    12.5
1,2,3,4,6,7,8-HpCDD                                  12.5
1,2,3,4,6,7,8-HpCDF                                   12.5
1,2,3,4,7,8,9-HpCDF                                   12.5
OCDD                                              25b
OCDF                                              25
  Prepared in isooctane. This solution also contained similar
  concentrations of the available brominated dioxin and furan
,  congeners.
  The level of OCDD was adjusted based on the endogenous level of
  OCDD in adipose tissue.  An additional, separate solution of OCDD
  at 75 pg/fiL was used to achieve the higher spiking level needed.
                       64

-------
Table 6-9.  Control Sample Spike Levels
Analyte
2,3,7,8-TCDF
2,3,7,8-TCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,7,8/1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDD
1,2,3,4,6,7,8,9-OCDF
1,2,3,4,6,7,8,9-OCDD
Low spike level
(Pg/g)
10
10
10
10
10
25
25
25
25
50
25
25
25
25
50
350
High spike level
(P8/g)
50
50
50
50
50
125
125
125
125
250
125
125
125
125
250
700
                 65

-------
Table 6-10. PCDD and PCDF Spike Check Results
Analyte
2,3,7,8-TCDF
2,3,7,8-TCDD
1,2,3,7,8-PCDF
2,3,4,7,8-PCDF
1,2,3,7,8-PCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDD
OCDF
OCDD

Replicate no.
1
2
3
Average recovery (%)
Spike
level
(pgW
50
50
50
50
50
125
125
125
125
125
125
125
125
125
125
250
250
Table 6-11
Spike level
150
150
150

Spike check 1
(Pg/nL)
44.3
51.0
47.3
48.6
46.5
124.0
119.4
104.0
93.0
115.9
118.4
107.3
114.3
114.9
115.4
207.8
226.7
. OCDD
Recovery
89
102
95
97
93
99
96
83
74
93
95
86
91
92
92
83
91
Spike check 2
(pg/nL)
41.1
49.5
56.8
54.7
46.1
121.3
116.2
105.3
107.0
115.1
122.6
114.5
114.4
99.7
118.0
210.6
224.6
Recovery
82
99
114
109
92
97
93
84
86
92
98
92
92
80
94
84
90
Average
recovery
85
101
104
103
93
98
94
84
80
92
96
89
91
86
93
84
90
Spike Check Results
(pg/jiL) Amount found








142.6
154.6
138.1

Recovery %




95
103
92
97
                    66

-------
 cfl
 &
 "tn
 O

 I
 O
 U
2
ts.

 0>




%
 tn
 
tn
 1
 re
 a
 IH
 O)

P-,
3
 tn
 a
H
            NO
            O

            I
            !8
            PH
d


| 5-


% "
8
                    o
                    u
                       t
                       CO
                   O
                   O
                ~a  5 i-j
                fl *  =L
                -P  r! >>.
                      a
                      2
                              00
                                     v)
                                            00  CO tx
                                                O CO
                                                                 CO
                                                                                   o\
                                                                                       00

                                 OO
                                            OO
                                                       O O  O O  O O  O
                                                                o o

                                                                in in
                                                                               Q
                             OO OO  tx Cx  tx ^  ND
                               V   V   V   V  V   V  V
                                       " c" TH" r-n" T-H"
                                                                               rfD Q
                                                                               c^U U
                                                                               ^H-O O
 oj  2T-S  5

H.J2-3  
                                                                                               "* x
                                                                                              7  re
                                                                                                  P
                                                                               00  O  *J  w

                                                                               K  s  .8 ^

                                                                               en  B  cfi  2
                                                                                -  u  eg  
                                                                                ^  w  3  S
                                                                               i-i ,Q  tn  IH
                                                        67

-------
             ro
 O
43

i
en
 OJ
X
U



4
 o
 en
c
 ro


 I-
 o
U
S

 2
 o
,0.
 ro
CO
i

SO

 4)

3
 ro

H
       ro
            3?

            T3
             re
S
    f-i

    O

    O)
co oo CN
^ oo IN oo
o in IN ro
ro ON TJ< oo
00 sO SO sO
rO \o ON tS
O ON O CO
SO O rH
OO OO ON rH
I 1

ro ro ro ro
ro ro to ro
oo oo oo oo

s  y 
U U PH PH
H H oo oo
oo oo tv IN"
tx Cx ro Tf
ro to tN ro
!iri^)cotsl2'^'
fO v^3 ^D ^h tx CT1* LO ro ON
^o cr* c*^ c*^ ON ^j* o^ *~H co

^OlCCOrHrH^OtNO
CO CN OQ C*Q ^i^ CO C"^ ^^ LO
in
cococococorocoo
OOOOOOOOOOOOOOININ
C1 C? C1 C1 C1 C1 C1 '"^ T-*

LLi [l( P tH P
puPPPPPPp.R
ppppuuuuu
^y^yxxx'SS
^^hrs^'''-as-
iJn ,-L 4-< 4j< oq^ON^oq^oo^oq^
ONOOOOONtN'oo'tN'tN'lN"
OOlN"lN"oOso"tNSOSOsO
t^^}*SOtNT}
-------
           Table 6-14. Intel-laboratory ComparisonControl Lipid Results (pg/g)
Analyte
2,3,7,8-TCDF
2,3,7,8-TCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
1,2,3,4,7,8-
HxCDF
1,2,3,6,7,8-
HxCDF
2,3,4,6,7,8-
HxCDF
1,2,3,7,8,9-
HxCDF
1,2,3,4,7,8-
HxCDD
1,2,3,6,7,8-
HxCDD
1,2,3,7,8,9-
HxCDD
1,2,3,4,6,7,8-
HpCDF
1,2,3,4,7,8,9-
HpCDF
1,2,3,4,6,7,8-
HpCDD
1,2,3,4,6,7,8,9-
OCDF
1,2,3,4,6,7,8,9-
OCDD
Labi
n=l

1.2
8
ND(l)b
16
15
19
9
1.6
ND(1)
11
99
17
36
ND(1)
132
ND(1)
986
Lab 2
n=4
Average
1.3
10.6
ND(O.T)
27.1
26.6
31.7
14.6
3.0
ND(0.3)
20.2
131
18.5
31.5
ND(0.4)
212
NRf
1030
RSD (%)
23.1
9.4
-
7.4
9.4
8.8
13.0
20.2
-
11.9
5.3
16.8
18.1
-
7.5
-
6.9
Lab3
n=5

Average RSD (%)
1.121
9.29
0.48C
25.1
20.8
18.7
10.8
ND(9.84)
ND(0.66)
-
136.8d
21.1'
29.8
1.42
140
4.02*
1,184
-
9.7
-
7.5
4.6
-
21.4
-
-
-
6.8
8.4
6.3
-
6.9
70.6
3.7
a Avenge of two positive quantifiable or trace values (n=2).
b Not detected.  Detection limit in parenthesis.
c One trace value  (n=l).
d 1,2,3,4,7,8- and 1,2,3,6,7,8-HxCDD reported as total amount in Lab 3 results.
e Average of four  positive quantifiable or trace values (n=4).
' Not reported.
8 Average of three positive quantifiable or trace values (n=3).
                                         69

-------
        Table 6-15. Inter-laboratory ComparisonLow Level Spiked Lipid (% Recovery)
Analyte
2,3,7,8-TCDF
2,3,7,8-TCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDD
1,2,3,4,6,7,8,9-OCDF
1,2,3,4,6,7,8,9-0000
Labi
n=l

70
70
56
50
70
56
56
62
68
60
120
72
100
76
56
56
64
Lab 2
n=4
Average
119
98
91
100
110
120
109
107
95
88
128
93
103
133
260
NRC
NR
RSD (%)
3.8
4.4
6.6
5.7
8.8
13.5
12.7
11.1
5.1
8.5
3.1
3.1
5.6
9.6
7.6
-
-
Lab 3
n=5
Average
95
83
115a
99"
94b
_c
95a
99d
79
e
88a
95
114
115
89s
99a
87a
RSD (%)
7.1
23.3
8.8
17.9
39.1
-
13.6
35.2
15.1
-
32.3
20.0
14.6
4.2
11.5
14.9
13.3
a Average of four positive quantifiable or trace values (n=4).
b Average of three positive quantifiable or trace values (n=3).
c Diphenylether interference observed.
d Average of two positive quantifiable or trace values (n=2).
e 1,23,4,7,8- and 1,2,3,6,7,8-HxCDD isomers were reported as a sum.
f Not reported.
                                            70

-------
       Table 6-16. Interlaboratory ComparisonHigh Level Spiked Lipid (% Recovery)

Analyte
2,3,7,8-TCDF
2,3,7,8-TCDD
1,23,7,8-PeCDF
23,4,7,8-PeCDF
1,23,7,8-PeCDD
1,23,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,23,7,8,9-HxCDF
1,23,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCD
1,23,4,6,7,8-OCDF
1,2,3,4,7,8,9-OCDD
Labi
N=l

Lab 2
n=4

RSD
Average (%)
74
86
60
62
78
62
72
70
62
58
70
78
100
82
80
81
89
111
91
93
103
120
112
104
105
86
89
93
88
95
120
121
NRd
NR
" Diphenylether interference observed
b Average of two positive quantifiable or trace values
c 1,2,3,4,7,8- and 1,2,3,6,7,8-HxCDD isomers were repoi
2.1
10.0
6.9
6.2
9.7
15.1
13.9
17.2
5.6
9.2
7.7
8.6
6.7
21.3
13.8
-
-
(n=2).
ted as a sum.
Lab3
n=5

Average
92
92
100
89
88
a
68
78b
74
C
97
91
108
104
90
93
93"

RSD
(%)
4.7
3.4
6.2
9.2
11.5
-
13.2
0.9
13.3
-
12.7
8.8
15.2
10.0
19.3
9.0
29.2

d Not reported.
e Average of four positive quantifiable or trace values (n=4)
                                          71

-------
6.4 SYNOPSIS OF ANALYTICAL RESULTS

      The detailed results from the analysis of all samples have been submitted as separate
reports to EPA (Cramer et al. 1989a, 1989b).  Appendix A of this report provides detailed
documentation on the results of each FY87 NHATS sample and QC sample. The data in Appen-
dix A identifies: batch number, laboratory identification; NHATS sample identification; number
of specimens in the composite; census region and age group; composition of composite by sex and
race; identification of data quality as positive quantifiable (PQ), trace (TR) or not detected (ND);
measured concentrations and limits of detection; data restrictions; and IQS recoveries.

      Successful analyses were achieved for all samples except one, ACD8700425. This sample
appeared to have been fortified  with twice the specified amount of IQS.  In the remaining
samples, the range of 2,3,7,8-TCDD detected ranged from a nondetect value of 0.0689 pg/g to a
maximum of 15.1 pg/g. Positive quantifiable OCDD ranged from 136 to 1,660 pg/g.

      It should also be noted that in many of the analyses, responses to octachlorodiphenyl ethers
(OCDPE) overlapped with the responses of the  1,2,3,4,7,8-HxCDF  and the 2,3,4,6,7,8-HxCDF
isomers.  In these  cases the observed responses were quantitated but were reported as a
nondetected (ND) value.

      The recoveries of the internal quantitation standards (IQS) were within the QA data quality
objective  of 40% to 150%, with the exceptions that are identified below. As previously noted,
composite ACD8700425 appeared to have been fortified with twice the specified amount of IQS.
Data for this composite are considered suspect.  In composites ACD8700023 and ACD8700256, the
recoveries of 13C12-PeCDD (28%) and 13C,2-PeCDF (17%), respectively, were outside the data quality
objectives. Since only one out of nine IQS recoveries per sample were not in control, the samples
were not reanalyzed. The PeCDD and PeCDF data for these two composites were flagged as not
meeting the DQOs.  In batches 3 through 5, the recoveries of the carbon-13 labeled HpCDD in
three  composites and the labeled octachlorodioxin in 17 composites  and/or QC samples  were
outside the DQOs.

      The analysis results for the spiked  control QC samples  (Table 6-17) indicated that the
accuracy  of the method met the QA objectives of 40% to 150% for all compounds  with the
following exceptions: the 1,2,3,4,7,8- and 1,2,3,6,7,8-HxCDD  pair in the low level batch 2  spike
(13% recovery); OCDD in the high level batch 3 spike (39%);, 1,2,3,4,6,7,8-HpCDD in the low level
batch 5 spike (32%); and 2,3,4,7,8-PeCDF in the low level batch 4 spike (not recovered).  The
PeCDF recovery was affected by a momentary reduction in sensitivity resulting from the apparent
coelution of a high level compound interference. The remaining compounds showed recoveries
ranging from 44% to 137%.
                                         72

-------








2
^

PH
1
(H ^j
. ^ PH


D ^
u JH
'o D
bu
CQ p t
i
C/>
r-t
I
sO
1



c c
. o
3 V4H '-P "^
*C o .*" ff^
Q> 5 ^"^
o 5
U ^
^cJ C
111
a S-&
en TJ

I'g'b
O ** flJ "
&.H, o &
CJ "^ OJ
rt "3 ^H
CQ
C
o
10 w '^ ^^.
rt S S 'oC
I 8 3 IS
< S c ^
0
a3 ^3
11

^
2

111

f


c^osyrj ooino t^^jso co^in
iri^oo ^ TjJ m. m'^1^ ^oC1^
rHOO^ OSrHCO OOOOsO t-xrHSD



CJCJ*^ O^1"1 CJCJ00' rHsO'1*



CO CO in OO ^ 00 Csl rH
ifOrH 'CO'rH 'rHQs 'inOs
Os Os 00 Os *"^ Os tx 00





rQinoq OSSDC^J 2^^ ir1vO.vi?
""Joso" ^tNiri csjrHO incsios
rHrH^H OSrHin ^rHin CSCOsO



tsinin ininin rn^m inmin


csinm inmin in-^m inmin

i ,OO r- .OO r-^OO riOO
^'rHin ^rHin ^rHUI " 'rHin

* D D D
P Q U U
CJ rj o> o
H H PT ^
oo oo R. v
K tC c\ t\
ro_ co" ^ ^
M CsJ ^ "X
rH N

SO rH O
^ rH OO



|S5 



o *-!
i o oo <
rH 00





00 00 00 tx
cj o * oo
CS CO SO rH



m co in rH


in co in rn

,-* o o ^
0 rH in 
_ tu
D D
8 
-
V [*s,
CO" ^
Vj. co
r^ K
r-H

^ ro
r-5 ^



csq oo
co ^



rH
OS





oq -^
^~^ ^J
rH CO



in ^


tn ^

OS
PH
3,6,7,8-HxCD
c^
rH

in
co



Os
rH



00






SO
ol



in


in

a



73

-------
















(U
g
u
IX
3
QJ
3
H







4-*
_ 
*2 *o ' S"
^S *"H """^
o 5
u >
73 C
Is  '~"
"^ ' ifc
43 'E ^
55 TJ
3 QJ IH
O *J 4> -  -
fcri 13 ^ ^^.
^ "5* 8  '
to  2i
pa
-a .
|i|l
J5 Jj w ""-^
" o
u
^ j
,0 w
g 0
Z^
T3
Q
z
a
c
 TH * ro os' od so
T icoos - ; *{ o\ THTHCO cr ^
^ f? ^ CO" CO" ^>
N" TH" ^ ^ n

in os in
IX* 't .
in *^ os




3 rq oo
in ^ H
o' TH TH


^ CO
1 TH O
l-H t 1



oo Os" ro
C5 f^1





IN in in

m m in
P-I
8
OH
oo"

*
^
1-H~

-------











if
3
1
o
u
[Ss
t-H
SO

$














*j
'0 <4_, V '"~s
o 5
U >
-a c
IB  '~N
I'^lb
C/2 "Tl
TJ
3 ai ^
o 4S 53 x-^,
1114

c
"^ .2
fill
o
(J

J,. ,Q
&> "s
,Q JB
 
3 4-*
^


z


^^s
CO ^ ^

1




*cncq c
co; co^ co- J^ "-
^t *^ ^ TH rH
rH rH TH1






^_^
>
"a>
Tj 05
 O,
o
^ > 'a,
0  w
r^ ^Q J> j^
QJ ^3 -J N_ .* ^3
|^ b |

S '$ o u
? w g
CA flj *H
^ J " a>
tj O . ^O *<
C pT1 n> 3
S S ** h
cd s  CL>
* S .H re
o -g .a
 S3 n3 T3
"1 i 
3 a O O
c 5 oo oo
u a 3 3
^, rC re re
Z H pa PQ
id  u
75

-------
6.5 STATISTICAL ANALYSIS OF THE QUALITY CONTROL DATA

       The statistical analysis of the FY87 NHATS QC samples for the PCDD and PCDF analytical
results are summarized in Table 6-17.  The objectives of the analysis were to

            Estimate the percent recovery of the analytical method,

            Determine if there are significant differences in the analytical performance among
             the batches,

            Characterize the precision of the analytical method,

            Establish the relationship between the precision of the analytical method and the
             level of the spiked concentration, and

            Identify anomalous results that suggest potential problems in the  analytical
             measurements.

       Of the 68 samples analyzed for PCDDs and PCDFs in the FY87 study,  20  were QC
samples. Each of the five analysis batches contained one method blank, one unspiked control
sample, and two spiked samples. The sampling plan for the allocation of these QC samples has
been described by Heath (1988).  Leczynski et al. (1988) determined the assignment of the QC
samples to the analysis batches.

       Because it was agreed that population estimates would be calculated using only the data
that met the data quality objectives (DQOs), the same criteria were applied before evaluating the
QC data. The DQO criteria are:

       (1)    Internal quantitation standard (IQS) recovery must be between 40% and 150%,
       (2)    Ion ratio must be within 20% of the theoretical ratio, and
       (3)    There must be no problems with coelution or fragmented peaks.

If an analyte was not detected (ND), the measured concentration of the QC sample was computed
as one half of the detection limit (LOD/2). This same approach was used in the statistical analysis
of the field  samples.

       A descriptive summary of the QC data is presented in Section 6.5.1. In Section 6.5.2 the
statistical approach to analyzing the QC data is discussed and the results of these analyses are
presented in Section 6.5.3. Conclusions are presented in Section 6.6.4.
                                          76

-------
       6.5.1 Summary of the OC

        Table 6-17 presents the data for each chemical in the quality control samples. Table 6-17
 presents the data by chemical and spike level (pg/g) and provides information on the number of
 QC samples for which the DQOs were met, the number of positive quantifiable (PQ) and trace
 (TR) measurements, the average measured concentration, the background adjusted  recovery
 (BAR) for spiked samples, the standard deviation (SD) of the measured concentration, and the
 coefficient of variation (CV).  The background adjusted recoveries  at the low (L) and  high (H)
 spike levels were computed as

                      BAR(j) = 100%*[Avg(j) - Avg(O)] / spike level,

 where Avg(j) is the average measured concentration at spike level j  (j=L or H) and Avg(O) is the
 average measured concentration of the unspiked control sample. The coefficient of variation was
 computed as
                     CV(j) = 100%*[SD(j)/Avg(j)]/ for j= 0, L, and H.

       All background adjusted recoveries  were between 67 and 120% and the coefficients  of
 variation were generally between 2 and 20% at both the low and high spike levels.  In the control
 samples analytes at levels below the  detection limit of the analytical method generally had the
 higher coefficients of variation.

       Only one analyte was detected in the method blank  The measured concentration  of
 2,3,7,8-TCDF in the batch 1 method blank was 0.460 pg/g, or 41% of the average measured 2,3,7,8-
 TCDF concentration (1.12 pg/g)  in the two control samples that met the DQO criteria.  Because
 this indicated  a potential for a bias  affecting all batch 1 samples,  the measured 2,3,7,8-TCDF
 concentrations of the 48 study samples were compared across the five batches. This comparison
 did not reveal any statistical evidence  of a batch  effect on the study samples.  Therefore, no
 adjustments were made to the  measured 2,3,7,8-TCDF concentrations of the study or the QC
 samples in batch 1. In particular, all 2,3,7,8-TCDF QC data meeting the  original DQOs were
 included in the statistical analysis.

      Appendix D contains the plots of the  measured concentrations against the spike levels for
all study compounds.  It is evident from these plots that the relationship between measured
concentration and spiked concentration is nearly linear for all the compounds.  Because  only one
measurement met the DQO criteria for 1,2,3,4,7,8-HxCDF, a figure is not given for this chemical.
Also presented in these figures are  the  predicted concentrations  with tolerance bounds for
individual measured concentrations. The statistical methods for calculating the predicted values
and tolerance bounds are discussed in the following section.
                                         77

-------
       6.5.2 Statistical Approach to Analyzing the QC Data

              The QC data were statistically analyzed using linear regression models fitted to the
measured concentrations for each compound.  Three models, as described in Table 6-18, were
fitted to the data to determine the best fit and to test for significant batch effects.

       Initially the full batch effects (FB) model was fitted for each compound.  The FB model was
used to test for two types of batch effects:  fixed effects and proportional effects.  A fixed batch
effect is the constant amount by which the measured concentrations in the batch differ from the
average for all batches.  It is calculated using the intercepts (., i=l,...,5) of the FB model. For
example, the fixed effect of batch 1 is represented by  ]-K/ where  is the  average intercept.
Proportional batch effects are  characterized by differences in  the  batch recoveries.   The
proportional batch effect for batch 1, for example, is the difference between the recovery (slope)
for batch 1 and the average recovery for  all batches.  Using the notation  in Table 6-18, the
proportional effect (also called the recovery effect) for batch 1 is denoted by PrP-

       For those  analytes providing sufficient  data,  F-tests were performed to determine the
significance of the fixed and proportional batch effects.  In virtually all cases where batch effects
were detected, they were due to variations in the slopes from batch to batch. Therefore, a second
statistical model containing a constant intercept (i.e., a,  = <* in the full batch effects model) was
fitted to the data for each analyte. This model is  called the batch slopes (BS) model because it can
be used to test for significant differences in the  recoveries (slopes) between batches.

       The BS model was used to estimate recoveries for each batch, overall average recovery,
and predicted concentrations at  each spike level. Statistical F-tests were  performed to test for
significant background levels and batch effects. Background  levels  are indicated when the
estimated intercept is found to be significantly different from zero and batch effects are indicated
when at  least one of  the estimated batch slopes is found to be  significantly different from the
others. Predicted concentrations at each spike level were calculated from the estimated intercept
and average recovery.

       For some analytes it was not possible to fit either the FB or BS models because there were
insufficient data after applying  the DQO  restriction criteria.   In these  cases a simple linear
regression (SLR) model was used  to estimate average recovery and test for significant background
levels.  However, using the SLR  model, it is not possible to test for significant batch effects.
                                           78

-------













43
re
Q
U
a
&
z
&
>
PH
01
N
X
13
5
o

d
O)
to
P
tn
"aJ
T3
0
^

.2
"to
so
QJ
&
QJ
rv"
^JH
od
i i
4
Ol
3
re
H



























Remarks








IT
a>
O)
"e 2
C cs
5 a
fh
#1
i  AJ
,











3
o
2
u/
d MH
0 0
6 &3
.a g en
S* >X
re g fa *
43 S  fi
re < -5 o
^ S-d*
O> -iH d -M
* S s p<
+- 8r D
^ v CJ
C o "o! H
Ol U -rt W
(;) are the batch background
pes (PJ) are the batch recoveries. Wh
the average of the estimated batch re
estimated recovery from the SLR mo(
ted intercepts equals the estimated int
y. Oja QJ re
+> Trl -rt j 5
eL,1-" s ^ g
 - S-ilN
^ "S J 1 % "S
*J S TJ i O> O
^ 1 2 ir-5 g





or
'a,
^r 
^*- IH
, Ol
-4- -i-1
T a>
rtS
n^l
re
:srP,
O o
m '-'
M" ^<^^


tfi
4^
U
QJ
M-t
M-l

X
W
re
CO
^ "
rr ^
MH ^ '
UJ
d
0 T3
6^

W3
g o> __.
O O) O)
*"
SI "8 6
ra S  c^
43 > -->
re o w ^
rt W -rt (..
a) is the background level.
are the batch recoveries. When the <
the average of the estimated batch re
estimated recovery from the SLR mo<
quals the estimated intercept from th(
~""^H O) 0>
>-> a. T? x; -M
CL.V  > ^ fx,
Hi-iS
QJ g-i re 3 o>
H^ia &<
,5 en ^ o> .5





or
en
ca^
+ ^
Ol
 g
II s
re
r:F-
W^


'fll
U/
"G
O
*H
2
CA
Ol
{X

175
s~
j en
Cu rv\
 S-













a) is the background level.
s the recovery.
N^^ ^H
4^ y*\
PnCO-
^
S3 &,
s o
c '"^
w en





u"
en ^
ca. 
+ S
Ol
 g
!L
,:P&H
wSS




S"
H B?
re --
Ol
.s 
J -S
w
0> VI
f-H Q>
& K
g S5
en csj













 *H
-6
.4-t
re
,>
M-l
O
Ol

D-,
g
re
m
U
a
x
*3

1
.s
IL> C
> o
Ji V u
ill
_d C CJ
|>sa
^^ ^ P^M Q^ ( ;
S T3 X H.
aj g ^_i  N
5 ^^
* v  
^ y a ^
^-^ s "S
 -1-3
^ ^^ ?
re ^o w ^
TJH- _? ol S
m- S 8^^
S oj ^
_i- C
, II II
l-H U .I_<
nil rT'r'^
II *sJ U
.rt . ,W en

n






i

OH
to
X
bO
2

"S
re
4
a,
!0
^
^O
1-J*
o
&
C
8
MH
O
52
01
nO
S
13
3
cr1
01
K
re
a
o>
S
IH
01
Ol
1
^
d
I
43 
(y *C
-O QJ
J 'C
re u
 o
re O
T3Q
_ 
^s^
c" -g
S 1
+-J C
re c
o ^
A re
^-S
CH
re ^
3-S
,2 a.
* g
G re
>-3 (0

J3
79

-------
      Although the analysis established that there were statistically significant batch effects for
most of the analytes it was decided that there would be no adjustments made to the measured
concentrations of the study samples.  Thus, a model was developed in which the differences in
recoveries from batch to batch were treated as random effects; thus affecting the precision of the
analytical method. The model assumes that the standard deviation of the measured concentration
has two components: (1)  a component associated with the within-batch measurement error,
estimated by the mean squared error (MSB) from the BS model; and (2) a random component
associated with the  random-batch effects.   The standard deviation of  the  ith measured
concentration (t^) at spiked concentration SCj was computed as:
                            SD(C,,) =
where MSB is the mean squared error from the BS model, and SD(pj) is the sample standard
deviation of the estimated batch recoveries.  According to this model, the standard deviation
increases with the concentration of the sample; however, it is not necessarily proportional to the
concentration. When there was insufficient data to fit the BS model, the square root of the MSB
from the SLR model was used to estimate the standard deviation of the predicted concentration.

      Approximate prediction bounds on the individual measured concentrations were calculated
by adding plus or minus three times the estimated standard deviation to the estimated predicted
concentration. The probability that an individual measured concentration will fall  within the
prediction interval is approximately 99% according to asymptotic distribution theory.

      6.5.3  Results

      The results of the statistical analyses are summarized in Tables 6-19 and 6-20.  Table 6-19
contains an estimate of the average recovery, its  standard error (SB), and the estimated batch
recoveries for each analyte. When there was sufficient data, the individual batch recoveries were
estimated from the BS model,  otherwise the SLR model was used to  calculate the average
recovery.

      For each compound, a hypothesis test was performed to determine if the average recovery
was significantly different (at the 5% significance level) from 100%.  The result of this hypothesis
test is denoted by an " * " next to the estimated average recovery.  The average recoveries were
determined to be  significantly less than 100% for nine  compounds, and the  lowest average
recovery was estimated to be 66.1% for 1,2,3,6,7,8-HxCDF.  There was only one compound,
1,2,3,4,6,7,8-HpCDF, for which the average  recovery is statistically greater than 100%.

                                         80

-------
PQ
 cu
P;
 cu
 bo

 E
 cu
a
 0)
4-1
 re

_g

 cn
H

re
CO
w


v2
 cn
 cu

 cu

 O
 CU


.1


W




3
 cu
       cn

      .**

       cu

       $

      w
      ca
      cu
      V
      II
      cn
      W
              OT ,2
                 in
                CC
                CM
o
cu
                13

                ^
                                                                  8 &
                                     CN|
                      rq
                           o6o\
                            OO CN

                            sO tv
                            IN C>
                                                       INOO
                                                          \o
                                     so   rq
                                                                    00
                                                                    sO
                        odo'ocirQ   oo   IN in   rsiomcsc)
                        CX>O\tsOO   Tfl   ITlrH   OrHOOs
                      rq co CN
                                     oo
                                                  in
                                     tx   tvlft    rHOOOOCO
                                     so   tNO    CTvOOON
                        o\iNOtNC~qtvt7\rOTij           in

                        tx OO so sO   O\'&  CSl 00    ON in IN SO      ^* tx
                        O O ON IN   SO   OO O\    OO OO ON rH      rHOO
                      CO
                      00

                      O
                                    ^ CT*\ f^ s^ ^^ ^*i ^H l^

                                    -Hr-irHTl5tNit-s;rqiri
       *  *
       CO ^
                                     *  *  *       *

                                     r IsOCSlsD    rH
                                                                    CO
                      rHtNOOtN   SO^COOO    OtNCOrHCOrH
                      CT^ CT^ O^ CTv OO   ^O [^ r^* O^    O^ ^z!3 f~^ ^\ Q\ ^H
                                                            3.

                                                            PH
                                      b tin P PH P

                                 PgPP PgP
                                    o u u u , o
                     PH P

                     P P
                     U U
                     H H

                     o<
                                     IN
                     IN IN co
                        CO rH tSl
                  *^NOVOI

                  CO CfT Cff T)J I

                  rn" rn" rn" rq" <
                                                  S ocj^ c>v 06^ od od
,4,7,8-HxCDDc

,6,7,8-HxCDDc
                                               m" cr cr  co co
                                                                    CN
                                                                          cu
                                                                         9
                                                                          &
                                                                         8
                                                                          c
                                                            cu
                                                                         "S
                                                            cu
                                                            bo

                                                            2
                                                            cu

                                                            re
                                                                          re
                                                                 %
                                                                               cn
                                                                               cu
     u *
     #3$

     ^   I
      O^cl
     ** bo JJ

     1-a-g
      O 3  *

      S^-2

                                                                                         i
                                                                                         in
                                                                                         cu
                                                                                         re

                                                                                         .8
                                                                           
                                                                           j=
                re



                en
                CU
                                                                                       rt *-
                                                                                      > be
                                                                                       cn C
 c?
 ^=5

^ S
                                                                             g  g* fifi
                                                                              .12 "PH MM frt "^
                                                                              (4 *!" M
                                                                              tT3 J-< rt
                                                                                 W3
                                                                                 OJ
             firH
             re



             
             cu
                                                                         .x
                                                                          en
                                                                           ^3
                                                                                  cu
o

ent
sis6
llili
'$>
                                                              % < 
  
 re .jo

 "g
J-j *-!
 C 
 SJ cn
'C Si
UH re
^ .a
 cn TJ
 C C
                                                  81

-------
 OJ

 0)


 <3J


'&,
c/i


-6
 re
PJ

 re


 o
a
 re
c
 re
 C
 cu
o
 QJ
 o
u
 re

 'SB


 ex


-------
    The individual estimated batch recoveries are shown in the remaining columns of Table 6-19.
Also presented are the results of the hypothesis tests for differences among the batch recoveries.
Notice that there are significant batch effects for virtually all analytes.  However, because there
were no apparent patterns to the batch effects, it was decided to treat the batch effects as random.
This means that no "batch" corrections were made to the measured concentrations of the study
samples. Instead, as discussed in the previous section, the batch effects were treated as random
and included in the estimated precision of the analytical method.

    Table 6-20 contains  the model-derived predicted average  concentration and  estimated
coefficient of variation (CV) for each analyte and spike level. For each analyte, a hypothesis test
was performed to determine if the predicted concentration in the control sample was greater than
zero at the 5% significance level. For example, the predicted concentration of 2,3,7,8-TCDD in the
control samples was estimated to be 8.90 pg/g and this estimate is significantly greater than zero
at the 5% significance level.  This is consistent with the fact that 2,3,7,8-TCDD was detected in all
five control samples.  The background concentrations were determined to be significantly greater
than zero for six of the seven  PCDDs (not including 1,2,3,4,7,8/1,2,3,6,7,8-HxCDD because the
individual isomers are already represented) and for four of the nine PCDFs (1,2,3,4,7,8-HxCDF is
not included because of insufficient data).

    In general, the relative precision of measured concentrations is much better for PCDDs than
PCDFs.  At the control level, the CVs of the measured concentrations for the four PCDFs which
had significant background levels were between 13% and 48%.  All of the CVs for the PCDFs
were less than 21% in the spiked samples.  The CVs at the control level for the six PCDDs with
significant background levels were between 5% and 20%.  Although this might be explained by
the higher concentration levels of the PCDDs, the relative precision for measuring PCDDs is also
much better in the spiked samples. With the exception of 1,2,3,4,7,8-HxCDD, the CVs for PCDD
measurements are between  3% and 10% in  spiked samples while  the CVs  for the PCDF
measurements in spiked samples were between 2% and 21%.

    In general, the linear model provided  a good  fit to the measured concentrations for most
analytes. This is evident in the figures presented in Appendix D.  There does appear to be some
lack of  fit for  compounds  1,2,3,6,7,8-HxCDF and 1,2,3,4,7,8-HxCDD.   The lack of fit for
1,2,3,6,7,8-HxCDF is most probably due to the rather low measured concentration of 73 pg/g in
batch 3 at the high spike level. The measured concentrations for 1,2,3,4,7,8-HxCDD were reported
as a combined value with the 1,2,3,6,7,8-HxCDD and therefore could not be effectively evaluated
using the linear model.
                                          83

-------
6.5.4 Summary of QC Data


The results from the statistical analysis of the QC data are summarized as follows:


1.    Nine of the analytes had estimated average recoveries that were found to be less than
     100% at the 5% significance level and one analyte had an estimated average recovery
     that was  significantly greater than 100%.  The estimated average recoveries were
     between 66.1% and 124% for all analytes.

2.    There are statistically significant differences in the recoveries from batch to batch for
     most of the analytes.  However, it is recommended that no batch adjustment be made
     to the study samples.  Instead, the estimated measurement precision will account for
     the batch effects.
3.     Measurement precision, determined by the estimated coefficients of variation in the
      control samples, is generally between 5% and 20% for the PCDDs and between 13%
      and 48% for the PCDFs. For the spiked samples, the PCDDs had CVs between 3% and
      10%, and the PCDFs had CVs between 2% and 21%. These  estimated CVs include
      random batch effects.
4.     Statistically significant background levels of four PCDFs and six PCDDs were identified
      in the control samples.

5.     The relationship between measured and spiked concentrations is generally linear over
      the range of spiked concentrations.
                                      84

-------
7.0 STATISTICAL METHODOLOGY

       There were three objectives for the statistical analysis of FY87 NHATS data:

       1.     Estimate average concentration levels in the adipose tissue of individuals in the
             U.S. population and in various demographic subpopulations,

       2.     Construct confidence intervals for these averages, and

       3.     Determine if average concentration levels of chlorinated dioxins and furans in the
             U.S. population differ significantly by any of the four demographic factors
             (geographic region, age, race, and sex).

       The statistical analysis methods used in this report are based on an additive model for the
demographic effects. Previous studies of the effect of using composite samples demonstrated the
validity of the additive model. A technique of iterative weighted generalized least squares was
used to estimate model parameters. The resulting  estimates are approximately normal for large
samples.  This approximate normality was used in constructing  confidence intervals and
hypothesis tests.  The remainder of this section provides details  of the statistical model and
process as well as references to background work.

7.1 STATISTICAL MODEL

       The use of composite samples for determination of the levels of PCDDs and PCDFs created
a need to reevaluate the approach to estimate general population levels of these compounds. The
statistical models previously used to assess NHATS data for OC1 pesticides and PCBs using
individual sample data were not adequate for extrapolating the composite sample  data. Section
7.1.1 discusses the background on the development  of a new statistical model, the additive model,
which is presented in Section 7.1.2.

       7.1.1  Background

       Mack and Panebianco (1986) developed and used a "multiplicative" model to analyze the
composite  sample data from the NHATS FY82 Broad Scan Analysis Study.  In their model the
analyte concentrations in a composite sample are represented as a product of fixed and random
effects associated with geographic and demographic characteristics of individuals who contributed
specimens to the samples. Orban et al. (1987) studied this problem further and recommended the
"additive" model which assumes  additive effects of the donors' geographic and  demographic
characteristics.
                                          85

-------
       The multiplicative and additive models were later evaluated by Orban and Lordo (1989).
They showed that under certain distributional assumptions, both models produce asymptotically
unbiased estimates of average concentration levels in the  target populations.  However, the
multiplicative model requires that the sampling and measurement errors be distributed according
to lognormal distributions. No specific distributional forms are required to achieve asymptotically
unbiased estimates using the additive model. Orban and Lordo (1989) also compared the two
models using simulated composite sample data which were generated from actual specimen data
obtained in the FY83 NHATS.  Their analysis demonstrated, using actual NHATS data, that the
standard errors of the estimates from these models are nearly equal.

       Following the study, the additive model was chosen to be used in the FY87 NHATS and
all future NHATS for the following reasons: (1) under very general assumptions, the additive
model  produces  asymptotically  unbiased  estimates of average concentration levels in  the
population, and (2) the additive model  establishes a more  tractable relationship between the
distribution  of analyte  concentrations in  individuals  and  the  distribution of  measured
concentrations from the composite samples. The second reason is particularly important because
individual specimens are collected but composites are chemically analyzed.

       7.1.2  The Additive Model

       Table  7-1  lists the  categories of the four  analysis factors of interest to NHATS.  The
additive model assumes that the four analysis factors have fixed additive effects on the average
concentrations in  specimens.   This assumption  creates 48  subpopulations defined by  the
4x3x2x2 unique combinations of categories.

                   Table 7-1. NHATS Analysis Factors and Categories

              Analysis factor                   Category

              Census region                    Northeast
                                               North  Central
                                               South
                                               West
              Age                             0-14 years
                                               15-44 years
                                               45+ years
              Sex                             Male
                                               Female
              Race group                      Caucasian
                                               Noncaucasian
                                         86

-------
       In addition to the four analysis factors, there are three ancillary factors that are assumed
 to have  random  effects on NHATS data.   They are (1) sampling of MSAs,  (2) sampling of
 individuals within MSAs, and (3)  measurement of  analyte concentrations in the composite
 samples. The second factor, sampling individuals from MSAs, also includes the effects of selecting
 specimens from individual donors.  The first two ancillary factors have random effects on the
 actual concentrations in individual specimens, and the third has a random effect on the measured
 concentrations of composites.

       From the assumptions above, the actual concentration in a specimen from the i-th donor
 in MSA j, census region k, age group t, sex m, and race group n, is represented by
                                  CRk + \ + sm + *n + MSAJ
where
              \i is a constant
              CRk is the fixed effect of census region k; k = 1,...,4
              A, is the fixed effect of age group f; { = 1,2,3
              Sm is the fixed effect of sex m; m = 1,2
              Rn is the fixed effect of race group n; n = 1,2
              MSAj is the random effect of selecting MSA j; j = 1,2,...
              GJJ is the random effect of selecting individual i in MSA j.

Furthermore, to uniquely define the parameters, we let
                        =1       fi=l     m=l      n=l
       The random effects MSAjk and eijk are assumed to have independent error distributions
with mean zero and variances o^ and o2, respectively.  Also, because of evidence from previous
NHATS and other environmental studies that the variation in specimen concentrations increases
with average concentration  levels, it is  assumed that o2 = b2|i2, where (is is the  average
concentration level is subpopulation s.  For notational simplicity we let
                                          87

-------
                                      CRk  + At  + Sm + Rn
for some particular combination of indices k, {, m, and n.

       This defines the model for the actual concentration in a specimen collected in the survey.
However, the specimens are composited prior to chemical analysis.  Thus, data are obtained from
the chemical analyses of composite samples. Letting Yh represent the measured concentration of
composite h (h = 1,. . ., C), the natural additive effects of compositing imply that
where Cijs is the actual concentration in specimen i from MSA j and subpopulation s; yh is a
random measurement error; Mh is the number of specimens in composite h; and Ch(i,j,s) is equal
to 1 if specimen (i,j,s) is in composite h, and is equal to zero, otherwise.  The error distribution
of yh has mean zero and variance o^.

       At this point the notation for representing the model is rather complex.  However, the
main points can be illustrated using matrix notation. Let
                                              , Av A2, Sv
be the 8x1 vector  of fixed effects and /.A = (/i,,. . . , ju,48)' be a 48x1 vector representing the
48 subpopulation average concentrations. Then
where X is a 48x8 design matrix. Letting Y = (Y]7. . ., Yc)' be the Cxi vector of measured
composite concentrations, Orban and Lordo (1989) show that the expected value of Y is
                                          88

-------
                                 E(Y)  = ZXp  = Dp ,
where Z is a Cx48 composite design matrix.  Thus, according to the additive model, both the
actual concentrations of  the individual  specimens and the measured  concentrations of the
composite samples have expected values that are linear combinations of the additive effects of the
fixed analysis factors.

       Orban and  Lordo  (1989) also show that the variance-covariance matrix of Y is a block
diagonal matrix that depends on 0^,0^,  and o^.

7.2 STATISTICAL ANALYSIS

       This section describes the specific methods used to achieve the statistical objectives.  The
estimation methods are discussed in Section  7.2.1  and the  hypothesis testing procedures are
presented in Section 7.2.2.

       7.2.1  Estimation

       The specific  quantities estimated  for the  FY87 NHATS are  the average  analyte
concentrations in the adipose tissue of the U.S. population and the averages for each of the
marginal populations defined by the categories listed in Table 7-1. The estimates were calculated
in three steps:

       1.     The  additive  model parameters  associated with the four analysis factors were
             estimated using a  method called iterative  weighted generalized  least  squares
             (IWGLS).

       2.     Estimates of average concentration levels in all 48 subpopulations defined by the
             analysis factors were calculated from the parameter estimates.

       3.     National and marginal population estimates were obtained by taking weighted
             averages of the appropriate subpopulation estimates. Weights were proportional
             to the population counts of the 48 subpopulations from the 1980 U.S. census.
                                          89

-------
       According to the model described in Section 7.1.2, the vector of measured composite
sample concentrations, denoted by
                                   Y = (Ylt...tYJr ,
has a multivariate distribution with mean
                                     E(Y) =
and a variance-covariance matrix V. The vector
p = (n, CRr C/^, C/^, Av A
                                                    2,
is the vector of fixed effects to be estimated.

       To obtain asymptotically unbiased estimates of the fixed effects it is not necessary to make
any assumptions about the form of the distributions of  the random effects.  If the variance-
covariance matrix V were known, the method of generalized least squares  (GLS) produces
estimates of p that are unbiased  and have minimum variance among  all unbiased estimates.
Furthermore, if the errors are normally distributed, the  GLS  estimates are equivalent to the
maximum likelihood estimates. The GLS estimate of p is
                                P = (D'V-lDylD'V~lY
Unfortunately, V depends on three unknown variance components (o^, o^, and o^) as well as the
vector p. Therefore, Orban and Lordo (1989) proposed a method involving iterative weighting.
Thus, the method is called iterative weighted generalized least squares (IWGLS).
                                         90

-------
       Starting values for the fixed effect parameters and estimates of the variance components
were calculated using a maximum likelihood procedure. This step was performed using the P3V
program in the BMDP program package. The resulting estimate of V was then used in the GLS
formula to produce a revised estimate of p.  Each time the GLS formula was applied, the estimate
of V was updated.  This process was continued until certain convergence criteria were met.
Orban and Lordo (1989) discuss this method in more detail and describe special  computer
programs for implementing IWGLS.  They also provide formulas for calculating the standard
errors of the estimates.

       An estimate of the average concentration level in each of the 48 subpopulations was then
calculated by
Weighted  averages  of the appropriate subpopulation predicted  concentrations were  then
calculated to estimate marginal averages for the categories of each analysis factor.  For example,
the average concentration in the Northeast census region was estimated by the weighted average
of predicted concentrations in all subpopulations in the Northeast region.  Marginal estimates
were calculated for four census regions, three age groups, two sexes, and two race groups. The
U.S. population estimate was calculated in a similar manner. An approximate 95% confidence
interval for each estimate was calculated by adding and subtracting two times the standard error
of the estimate.

       7.2.2 Hypothesis Testing

       Hypothesis tests  were performed to  determine if average  concentration levels differ
significantly by any of the  geographic or demographic factors.  The specific hypotheses tested
were:
             H,:  CR, = CR2 = CR3 = CR4 = 0,

             H2:  A, = A2 = A3 = 0,

             H3:  Rj = R2 = 0, and

             H4:  S, = S2 = 0.
                                         91

-------
The hypothesis, H17 for example, states that there are no differences in average concentration
levels among the four census regions. The alternative hypothesis is that there is at least one pair
of regions for which the average concentrations are different. The results from these hypothesis
tests are somewhat related to the confidence intervals for averages in individual subpopulations.
Generally, if the confidence intervals for any pair of subpopulation averages do not intersect, then
the hypothesis of no differences among subpopulations is likely to be rejected.  However, the
contrapositive is not always true.

       In order  to test  these  hypotheses, it was  necessary  to  make  specific  distribution
assumptions for the random effects. It was assumed that the errors associated with sampling
MSAs, sampling individuals within MSAs, and measuring concentrations were independent and
normally distributed. The additive effect of compositing specimens suggests that the normality
assumption for sampling error is reasonable because concentrations of individuals are averaged
in the composite  sample. Statistical theory states  that averages and sums are  approximately
normally distributed. Distributional assumptions were tested for all analytes using probability
plots and residual analysis.  The model validation results are discussed later in Section 8.5.

       The likelihood ratio method was used to test hypotheses Hj through H4. According to
asymptotic theory, the log of the ratio of the likelihood functions (calculated under the full and
null hypothesis models) has approximately a chi squared distribution. The number of degrees of
freedom is equal to the number of independent parameters constrained under the null hypothesis.
Orban and Lordo (1989) wrote computer programs to perform these tests.
                                          92

-------
 8.0 RESULTS

      This section contains  the  results  of  the  statistical analysis of the FY87  NHATS  for
 PCDDs and PCDFs in human adipose tissue.  The analysis was performed on data obtained
 from 48 composite samples, each containing an  average of 18 adipose tissue specimens from
 sampled cadavers and surgical patients.

       A descriptive summary of the data is provided in Section 8.1  and the results of  the
 formal statistical analyses are presented in Sections 8.2 and 8.3. Section 8.4 provides estimated
 rates of change of selected PCDD and PCDF concentrations.  Section 8.5 describes the outlier
 detection procedures that identified potential data errors to be corrected  prior to conducting
 the statistical analysis. Finally, Section 8.6 discusses the steps  that were taken to demonstrate
 the validity of the statistical methodology applied to the FY87 NHATS data.

 8.1 DATA RESTRICTIONS AND DESCRIPTIVE STATISTICS

      Prior to conducting the statistical analysis, the data were classified according to  the
 specified data restrictions and data qualifiers. For each of the  target analytes, Table 8-1 shows
 the number of  composite samples for which the measured concentrations were restricted or
 qualified.

      Data restrictions indicate  whether specific data quality objectives (DQOs) were  met
 during chemical analysis. In the data  listings of Appendix A, the data restrictions are indicated
 by C, coelution; F, fragmented peaks;  I, ion ratio criterion not met; P, peak separation; and R,
 IQS  recovery criterion not met.   If any of  the data restrictions  were noted for  a  particular
 sample and analyte, the  measured concentration was not included in the data summaries or
 statistical analyses.  For example, 15 of the 48 composites failed at least one of the DQOs  for
 2,3,7,8-TCDF.   Thus, there are  only 33 unrestricted measurements.   Preliminary  analyses
 demonstrated that significant biases can occur if restricted measurements  are included in  the
 population estimates.

      Data qualifiers are defined in terms of the analytical method's limit of detection (LOD)
for each analyte.  The analyte is reported as not detected (ND) if the measured concentration
is below the LOD, trace  (TR) if it is between the LOD and five times the LOD, and  positive
quantifiable (PQ) if it is greater than  five times the LOD.   Measured concentrations  are
reported only for detected (i.e., TR and PQ) analytes.  Table 8-1 shows the number of PQ, TR,
and ND measurements for each- analyte among the unrestricted composites.  For example, of
the 33 unrestricted measurements of 2,3,7,8-TCDF, 32 were positive quantifiable and one was a
trace value.
                                         93

-------
 S
"E,

to
 (LI
J
PH
C/}
oo

 01
-s
 a

a
5
 (0
D
d


 S
 o
_

D
 
U
 r
U-.
a
         8


        B
        %
a
 I
 8
 o



Is








"o

j-.s
(D t

.g
e
I
>
11
C-t to









C Qj
u o
JH CD
Q; 43
a
si
. .0
c '_*
a S
D
D
a

a
C-,

>H T3
^ "
c g
o3 g)
1.1


"C .F^l
_^
01 C
" Cfl
M a

B c
_3 w
cS k


>>
c
<-
8O.TintxOVcNOOcNO t>
ON rH ON ON OO ON rH f""i QN


cONDcOr-iin.-,.txl>.ini-( ON
cOcOTj
CO OO t"x t** tx "*^ ^D VD t^ ^* CO t**
C"** C^ C<^ ^^ CO CO CO ^* CO CO CN CO^
(rTco^r^covc H I J
|i|| |
cr1 C jS 8"1
jrt O I)  w
"^  P^ -S * ^
4_> *4J V ,- ?..l^
1 ill "I
's n ' S "S 'g
^ ^2 a jg .a ex
5 .y Si 1 | ^
^ fi -s w  a
TJ  0  ^ "S
.S oj ^ S " h
u ri ^Q pj j3
JlS'lcx f|
" - u " " -
                                                                      94

-------
      Table 8-2 gives a summary of the unrestricted data for FY87 NHATS PCDDs and PCDFs.
 The average, standard  deviation, minimum,  median, and  maximum  concentrations were
 calculated from the unrestricted measurements.  However, in calculating these statistics, the
 value of LOD/2 was used in place of the measured concentration whenever the analyte was
 not detected.  In some cases, the minimum is reported as < LOD, where LOD is the smallest
 detection limit reported for samples in which the analyte was not detected.  The average LOD
 for all samples and the  percent detected are  also presented.  Detailed data summaries are
 provided in Appendix E.

      The results presented in Table 8-2 and  Appendix  E are based on simple unweighted
 averages of the measured concentrations from the composite samples.  As such  they  only
 summarize the data. Statistical conclusions and estimates of population average  concentration
 levels should only be based on the statistical analysis presented in Sections 8.3 and 8.4.

 8.2 POPULATION ESTIMATES

      Not all of the  analytes provided sufficient  data to warrant a meaningful  statistical
 analysis. Two criteria were used to determine  which analytes would be statistically analyzed.
 First, the analyte must be detected (TR or PQ)  in at least 50% of the unrestricted composites.
 This ensures  that there will be minimal  bias caused by substituting LOD/2 for the measured
 concentration whenever the analyte was not detected by the analytical method.  Also, because
 sufficient data  are  needed to  estimate model  parameters  associated  with  the  four analysis
 factors  and three  variance analytes, it was  decided that  a  minimum of 30 unrestricted
 measurements was needed to achieve sufficient precision for  the population estimates.  Thus,
 of the original 16 analytes (the pair 1,2,3,4,7,8-HxCDD  and 1,2,3,6,7,8-HxCDD is counted  as one
 analyte because they could not be separated in most  samples) there were nine that  met both
 criteria for performing statistical analyses.

     For each of the nine analytes analyzed statistically, Table 8-3 lists the  estimated average
 concentration in the entire U.S. population and in each of  the categories defined by the  four
 analysis factors. Also presented is the relative standard error  (percent) for each estimate.  The
 estimates and standard errors  are based on the additive  model analysis described in Section
 7.2. The estimates  are asymptotically unbiased  and were adjusted for population percentages
based on the 1980 U.S. Census.
                                          95

-------



















to
jy
!"
(0
C/}
^
1
B
o
U
1
tx,,

0
v*-

to
P
T3
I
C
"o
b
(0
B
3
oi
CO
2
(0
H


















E
I
*J3
"S 5
o
w
c
E e
c .S

1
c

> 




ir


d
V





oq






00
rH





S





m
"
o








m









B*
P
U

00
ex"
^
CS"


^omSg oinS M
i* t*N <^ C\1 rH ^* f'd f\j rH O rO O
iSlrHrH K CSICO S^S




ON r-j m
O tx iri O O VO' r-I t>s* O O rn CO
rn [x ^H rn rn cO
rH OO



* * * * *
**vOts,O **rH O

rH f^ in
vvvvv vvv v~




r-t 00 fO
inSKSSrH OVOC3 ON
vo T(* co o o c4 m tx o tx r^ ON
^1 in rH
^T



tx ^xf rH
CO c5 tx tx *^ C\1 00 ON tx O
rHodirioo'c4 c4 iri o ON c4 vo
rH CO rn T-I rn C)





&S8vS"8 &S^SS8





vo 
Stxr^ptocQ o385ofetx



















Q B. B. D
(b B, B. B< pPPPP
pPQPQ uP^vJv^
OH oo oox oq^ ox o& tx ox tC oo; bC
oOit>TcCtxioovcxivc?oovvp^rCvps 
of of of m" of of rJ (sf of of of y y









































P
O

s
o 1
ex  o-
^i CQ Q>
= RE
to v^ / n
60 "J B
C x-=
J3 ^ C
S -a "B
Q U O
 J2 -S
 D.
CL, ^ ^



(9 t/1
W M 5$

"S 8 -2 S
s"fi 
8ti ti o
, S  B
IT!  c
 "u w i>
1 1 D 
' Number of i
' Percent of u
' Not detected
f The minimu

96

-------
"
 Q> ^

1 2
 g U U
BT,  I
 2 c ">

la I
 S D  B.
fl?  t
oo W
      U
u
     U
     z
   w  C
    .
Analyte

Population perc
                  
                                  '

                                                      |g      ^6    I
                                                                         IS

                                                                  r-5 O   ON O   Ifj
                 >?   
                           ^,      ^^     ^,
                           1^    in1**    i 1**    T-H
                                tx      ^H       *-
                                                                             oo pr   
                P
                Q
                UJ
                y
                >
                rt
                CM"
                                   
                                   u
                                    Q>
                                   o-
                                   00
                                   tC
                                  g
                                                   Q
   00    >-T
s& t.-    cA
tx vO    CO
^T m    t>J
m^ t    "
ri" r-f    rt
         "
                                               NO
                                               "*

OO
tc
VO
                                  S
                                                                                 UH

                                                                                 8
a
                                                                            3
                                                                                          D

                                                                                          C

                                                                                          u
                                                                                             ao
                                                                                                     J1  c
                                                                                                     oo o
                                                                                                     J3 -a

                                                                                                     o> J3
pe
conc

tand
population

average

relative
                                                97

-------
      The standard errors are used to characterize the statistical uncertainty in the individual
estimates.  Uncertainty of an estimate is best expressed by calculating a confidence interval.
Approximate 95% confidence intervals are  calculated by adding plus or minus two times the
standard error to the  estimate.   For example, the  national average  concentration of 2,3,7,8-
TCDD was estimated to be 5.38  pg/g with  a  relative standard error of 6%.   Thus the
approximate 95% confidence interval for the national average is 5.38   0.65 pg/g (4.73 to 6.03
pg/g calculated as 5.38   2x0.06x5.38).

      Estimates of the  average concentrations in the population categories defined by the four
analysis  factors are presented even if the effects  of  those  factors were  not found to be
statistically  significant.   For  example, the   regional  estimates  of 2,3,7,8-TCDD  average
concentrations range from 4.54 pg/g in the West to 6.02 pg/g in the North East. However, as
will be discussed in Section 8.3, the differences among regions  for 2,3,7,8-TCDD were not
found to be statistically significant.

      The age group  estimates  in Table 8-3 suggest that the concentrations of  nearly all
analytes increase with  the age of the donor. All of the analytes occur at higher concentrations
in the oldest age group  (45+ years). However, conclusions about the age group  effects are
based on the statistical tests  discussed in the next section.

8.3 HYPOTHESIS TESTING

      Statistical hypothesis tests were conducted for each of the target analytes to determine if
there are statistically  significant  differences  in average  concentrations between  individuals
from  different geographic regions, age groups, sex groups, and race  groups.  The  tests were
based on likelihood ratio tests using the additive model analysis as  described in Section  7.2.

      Table  8-4 lists the attained significance levels for the tests associated  with the four
analysis factors. The attained significance level is the smallest significance level that will result
in rejection of the hypothesis that there are no differences between population averages. For
example, the  differences among estimated averages of 2,3,7,8-TCDD in the four census  regions
could only be considered significant at the 0.15 level of significance.  On the other hand, the
differences in age group averages are significant  at  the 0.002 level.   A 5% (0.05)  level of
significance is generally the  smallest level used to declare statistical significance.

      It is clear from  Table 8-4  that there are statistically significant differences  among the
average concentrations in the three age groups.  The  differences in concentrations between
each age group for each compound, except TCDF, were found to be significant. For  each of
the nine analytes (Table 8-3), the highest average concentration is found in the oldest age
                                           98

-------

.
o
<4H
en
OH
3
o
IH
U
0
2
a,
re
K
oo
O
g
Ol
-
rH
O>
O)
i

O)
CQ
en en
01 PH
C-> ^-\
s R
S u
X Pi
O>
S -2
D %
H en
^8
-s
0) O,
*" tx
tn OO
**-< N^.
tn P"1
OJ U-i
5 tn
0 H
PL, ^^
|
g2
0
viS
 1
o>
0)
,_)

0)
CJ
re
1
in
3
01
3
&
















2
D
3
T3
J3
CJ

w






















u
a,
3
S
o
8
rrt
c2


u
a,
3
O
IH
U
><
Ol
en



,)
^
2
CJ
o>
^



"d
M
0
a
S
en
3
en
d
o>
u









*
X
1
<



csj I-H m
tx ** o
SO CO SO
O O O






oo es co
rH CO O
OS N SO
O 0 0





 2- !i-i
O T*1 ^^
goo
00
o  
V V




0 t-x ^H
in * S
rH tN o
0 O o










u, o R
Q Q w
U U PH
H H oo
oo oo tx
tC tC Tj?
co co co"
rsf Csl tN



co in
Os Os
in K
o o






tx 00
Csl *
SO SO
O O





V V
rH rH
^*i f~~i
f~*t ^**i
O O
V V




O IX
SO >*
Os OO
0 0










-PeCDD
,8-HxCDF
oo tx
tx so
co" co"
rJ r4"
rH rH




CO
0






^
co
Os
o






rH
g
O
V




1
o










,87
7,8-HxCDD
Is. so
Tf CO"
co cxf
rq TH"
rH



8
o






tx
ON
in
0





V
^^
o
o
0
V




Os
00
cq
0










,9-HxCDD
oo
tx
co
Csf
rH



co o
os rsi
co in
o o






Ttj in
(S OS
Tt< rjn
C3 O





V P
d
o
T3
OJ
en
re
^
^J
en
5
o
43
f3
VH
n
o
2
3
_S
d
j
u



























'"5 r 3 i i
d> nj *Tj
> > >
Ol Ol 0)
f"H  ^ ^^
^ rH 0
C3 O rH
O O O
Ol Ol a>
& A Jl
- .M HH
-t-1 H- 4_(
re re re
a> 01 a>
 r y
" c d
.5 ,3 _S
'53 'S3 '53
en en en
S Si |
S 5 S
^ fl ^
^J ^3 'O
d d d
HH h ( ^15
o v ^_
99

-------
group (45+ years) and, except for TCDF, the lowest is found in the youngest age group (0-14
years).

      The only other statistically 'significant finding, at the 0.05 level of significance, was that
there are possible regional differences in the  average  concentrations of 2,3,4,7,8-PeCDF. The
average concentration in the Western census region was estimated to  be 4.49 pg/g compared
to the national average of 9.70 pg/g.   The highest concentrations of 2,3,4,7,8-PeCDF  were
found in the North East census region (13.7 pg/g). This regional effect is discussed further in
Section 9.0 in consideration of the comparison of the FY87 and FY82 broad scan effort.

      Some potential difference in the estimated average concentration between Caucasian and
non-caucasian  and males and females are noted for each of the PCDDs and PCDFs presented
in Table 8-3.  However, these differences were not statistically  significant for any of the
modeled compounds  based on the statistical  hypotesis  tests which are  summarized  in
Table 8-4.

8.4 ESTIMATED RATES OF CHANGE OF SELECTED FCDD AND FCDF
    CONCENTRATIONS

      The additive model analysis compared mean concentrations across  three age groups,
lower (0-14 years), middle (15-44 years)  and  upper (45+ years).  The analysis  showed that
significant differences  existed among mean age group concentrations for  the nine modeled
analytes.  However, the analysis did  not  quantify rates of change between ages.  Therefore, a
second set of analyses  using linear regression was performed to address this issue.

      For  each analyte, the measured concentration in each composite sample was regressed
against the mean age of all individuals whose specimens had been pooled into that composite.
Two linear regressions were performed, the  first to estimate the  average  rate of change in
concentration levels  from the lower to the middle age groups and the  second to estimate the
average rate from the middle  to the upper age  group.   These rates were  taken as the
regression slopes times ten to  convert them into rates of pg/g per decade.  The plot of the
measured  concentration of 2,3,7,8-TCDD versus average  age with estimated regression lines is
presented in Figure 8-1.

      The average ages taken over  all composites from the lower, middle, and upper age
groups were 3.0, 30.8,  and 65.0 years, respectively. Therefore, the  first  rate was the estimated
average rate  of  change per decade from ages  0-31 years, and  the  second  rate was the
estimated  average rate of change per decade  from ages  32+ years. These  rates were further
standardized by dividing them by the arithmetic average concentration in the first age group
                                         100

-------
                                                                        o
                                                                        CO
                                                                       o
                                                                        o
                                                                        in
                                                                       o
                                                                        o
                                                                        ro
                                                                        O
                                                                        e\j
                                                 i    I     i    i
\o
             CO   CM   ~H   o
             -H   -H   <-H   iI


               O/Od' uopiRaueouoo



                                     101
                                                                                 01
                                                                                 W)
                                                                                 ro

                                                                                 QJ
                                                                                 bo
                                                                                 CD
en
3

QJ
>

G

O
U
                                                                                 o
                                                                                 43
QJ
                                                                                cc
                                                                                   C
                                                                                   O

-------
within each regression set.  That is, the first rate was standardized by dividing by the mean
concentration for the lower  age group, and the second rate was standardized by dividing by
the mean concentration for  the  middle age group.  The standardized rates  of change were
then reported as rates of change per decade, relative to the mean  concentration at the
beginning of the time  interval analyzed.  Because the average ages tend to cluster around
three points, it is not possible to characterize  how the rates change over time.  The analysis
only provide estimates of the average rates in the two age intervals.

     Table  8-5 displays national  average concentrations for selected PCDDs and  PCDFs, with
the two rates of changes  expressed as pg/g per  decade and percentage  of  initial mean
concentrations per decade, respectively. For example, the national average concentration of
2,3,7,8-TCDD is 5.38 pg/g (as  reported in Table 8-3). The rate of change between the lower
and middle age groups was  estimated by the regression slope as 0.83 pg/g per decade, with a
standard error of 0.17.  The  average concentration of TCDD in the lower age group  was 2.06
pg/g (not shown), so the rate of  0.83 divided by 2.06 resulted in the standardized rate of 40%
per decade.   Similarly, the rate  of change between the middle and upper a*ge groups  was
1.52 pg/g  per  decade, with a  standard  error of  0.20.   The  average concentration of
2,3,7,8-TCDD in the middle age group was 4.33 pg/g (not shown), so the rate of 1.52 divided
by 4.33 resulted in a standardized rate of 35% per decade.

     The rate of change  of 2,3,7,8-TCDD of 1.52 pg/g decade  from the  middle to the upper
age groups  is  similar to the value of 2 ppt/decade  reported by Patterson et  al (1985). Their
study investigated adipose tissue samples from individuals of age 35-85,  whereas the present
study analyzed composite samples from individuals representing all age groups.

     All of  the rates of change shown in Table 8-5, except for the first rate for 2,3,7,8-TCDF,
were positive and highly significant (p < 0.0001).  For seven of the nine analytes shown, the
rates of change per decade increased after age 31.  However, when the rates were converted
to a percent of initial average concentration, eight of the nine analytes showed a decreased in
standardized rates per decade after age 31.

8.5  OUTLIER DETECTION

     Prior to conducting  the statistical analyses of the FY87 NHATS data, outlier detection
procedures were performed  to identify possible data entry errors and errors associated with
the analytical method (Rogers, 1989).  Outlier detection was performed on four types of data:
(1) measured concentrations of native  analytes, (2) internal quantitation standard recoveries,
(3) LODs, and (4) percent lipid values for composite and QC samples.
                                          102

-------










?uran

T3
d
re
d
"rt QJ
o 3
.is en
P
1 
O o
QJ X
-H SH
QJ 1^
Ux wW
cn ^
^ ^
o C
QJ fy
>
c *-<
% *^
nj h+H
X I
i!-s
O en
en d
QJ O
 '43
"3 re
Pi h
4J
T3 d
QJ aj
 cj
ro d
S o
3 u
en
PU
in
06
^H
2>
re
H










2
re
sti
>-<
-L
r
rq
m
QJ
bo

>H
O
c
o
D






E
re
QJ
>H
rH
en
o
OJ
bo
<
>-H
O
C
o
0




















g ^
S5 "^
^ 
re ^
c ^


^^
pa
aj S2.
-M
cd
Pi
. o
^t
< ^D
"5fc
OH



CJ
S  <
r-J CJ M)
1 s a
o C
43 o
re U
z





QJ

"re
5
<




inoorH oorqoo is.oo
encqc^j rneSrH tsienen





omes sD'*^^ oooooo
rqents. cq-^o cjinrn
cJ ci r-! cj pq TJJ o o o
Q,
rq rq rn is. * as
insocq osoo envort<
rH c4 Ifi rH rH O O H rH
rH CM en
V"H






OrqrH OrHTf f^rHOS
Tfinso e-q^ttx 'T'm-^





t^en^rH^ trTirr^-^ aT^r?"
rHtsirH rqinoo oenrn
OCSrH OtNO OOC5


en so o en esj en
OOOSrH Tj^C^ CSJOOC5
C5rH\O rHOseN OrHrH
rH rH OO '
rH





oo oo o oo
en is. rH is. oq ts is.
irioin rno-rt; rnosin
rH t^ rH rH CN
rH tS.





 Q Q HH
o y R o, * R
08 %& ^843
Dw^ooSoo QaJE
UCHO&tsvosis: UPHOO
enHoOts.vo"oOsO_ ^Hootx
.5 oo isT rj? en" tC *" y c oo c\ so]
&
re re
^^
H ^
ll
IH IH
0 O
+H >*H
'os'u<
QJ OJ
2.S,
QJ QJ
T3 T3
ro re
U  >
re re

M 1 ^* *
0 0
4- +>
d d
QJ QJ
CJ U
S-* l-l
QJ QJ
OH 0,
re re
tn en
re re
T3 -d
QJ QJ
on en
en en
QJ OJ
^ V^
OH OH
X X
 H->
re re
pi Pi
ra ^
103

-------
      Logic checks were  performed to identify  obvious  inconsistencies in the  data.   For
example, logic checks would reveal records having recorded concentrations but a data qualifier
of "not detected."   The extreme studentized deviate (ESD)  test statistic was applied to the
residuals of a simple linear regression model fit to the measured concentrations and recoveries
as another means of  detecting outliers.    Finally, the secondary  outlier  procedures were
performed  using  tests  for normality,  multivariate techniques,  and  graphical  techniques
(boxplots).

      No data problems were detected in the logic checks phase. However, a total of 62 of
1620 data items were identified as potential outliers.  Of these,  total  45 were found to  be
correct readings by the analytical laboratory.  The other 17 data items were incorrect readings
which were recalculated.  The laboratory also reported additional data changes resulting from
its own review of  the outlier analysis.  All data corrections were made to the master dataset
before proceeding with the statistical analysis.

8.6 MODEL VALIDATION

      Three types of analyses were performed to evaluate the adequacy of the additive model
for use on the  FY87  NHATS data.  All three  analyses were based on comparisons of the
observed (i.e.,  measured) and predicted  concentrations for the composite samples.  Predicted
concentrations were calculated using the  statistical analysis  approaches discussed in Section
7.2.  Residuals,  which were also used in the model validation analysis, were calculated  by
taking the differences between the  observed and predicted concentrations.

      Model validation analyses included (1) residual analysis, (2)  normal probability plots,
and  (3)  R-squared analysis.   The use of  the  Shapiro-Wilk tests  for normality was also
considered.  However, in this application, the Shapiro-Wilk test is not appropriate because the
data are correlated and variances increase with increasing concentrations.

      The residual plots confirmed the model assumption  that the variance of the measured
concentrations will increase with the average concentration.  The  plots also show that the
distribution of residuals is  symmetric.   This supports the use  of normal models  for the
sampling and measurement errors.  As discussed in Section 7.2.2 the normality assumption is
important for ensuring the validity of the  hypothesis tests.  Nearly all of the probability plots
were linear, thus supporting the normality  assumption for the errors.  Those  that were not
linear could be explained by the larger variances at the high concentration levels.
                                          104

-------
      Finally, Table 8-6 lists the R-squared correlations between the observed and predicted
composite concentrations calculated for each analyte.  R-squared can be interpreted as the
percent of the total variability in the data (observed concentrations)  that can be explained by
the model.  For example, 81% of the variation in measured composite concentrations of TCDD
can be explained  by the fixed effects  of  the  additive model.   Overall, these correlations
demonstrate excellent agreement between the data and the model. Eight of the nine analytes
produced R-squared correlations of at least 68%.  The lowest value of R-squared (46%) occurs
with 2,3,7,8-TCDF.  This can be explained by the relatively  small age effect on 2,3,7,8-TCDF
concentrations. Although there were  statistically significant age effects for all nine analytes,
the ratio of average concentrations in the lower and upper age groups was only 1.2 (2.45/1.97)
for 2,3,7,8-TCDF.  The ratios exceeded 2.5 for all the other analytes.  Thus, for each of these
analytes, the additive model accounts for a large percentage of the total variability in the data.

            Table 8-6. R-Squared Correlation3 Between Predicted and Observed
                       Concentrations for FY87 Dioxins and Furans
                             Chemical                        R2 (%)
            2,3,7,8-TCDF                                         46
            2,3,7,8-TCDD                                          81
            2,3,4,7,8-PeCDF                                     68
            1,2,3,7,8-PeCDD                                     85
            1,2,3,6,7,8-HxCDF                                    87
            1,2,3,4,7,8/1,2,3,6,7,8-HxCDD                          88
            1,2,3,7,8,9-HxCDD                                    79
            1,2,3,4,6,7,8-HpCDD                                  88
            OCDD                                              82

            a  R-squared is the  square of the Pearson  correlation coefficient. It
              represents the percent of variability in the data that is  explained
              by the predictive  model.
                                         105

-------
9.0 COMPARISON OF FY87 DATA WITH FY82 AND VA/EPA DATA BASES

       The analysis of the FY87 NHATS specimens as composites provides a reference point
for body burden levels  of  PCDDs and  PCDFs in the general  U.S. population.  The data
generated from the analysis of the FY87 NHATS specimens can be compared with other data
bases that have been developed from the analyses of samples collected in North America and
Europe. The documentation on the total effort (the compositing design, chemical analysis and
statistical treatment of data) offers a means of comparing the significance of the FY87 NHATS
data set with two other analysis programs conducted using the NHATS specimen repository.
These two studies are the FY82 NHATS broad scan analysis effort and a collaborative study
conducted between the U.S. Department of Veterans  Affairs (VA) and  EPA's Office  of Toxic
Substances (VA/EPA).  The  comparisons of these studies extends the utility of the data bases
in establishing  trends in  body burdens of PCDDs and PCDFs  and  identifies  limitations in
comparing the results to other data sets. This section gives an  overview of these programs
and  presents comparisons of  the  compositing designs  (FY82  and FY87 NHATS), analytical
procedures, results, and statistical methodologies.

       The objectives of the FY82 and FY87 NHATS  were quite different from those of the
VA/EPA study.  The FY82  and FY87 NHATS studies  were  conducted to  develop  baseline
estimates of tissue  concentrations.  The FY82  and FY87 NHATS  tissue  specimens were
obtained  from  cadavers  and  surgical  patients.    The  target   population  was   all
noninstitutionalized  U.S.  citizens  in the  conterminous  48  states.  On the  other hand,  the
primary objective of the  VA/EPA  study was to compare PCDD and PCDF levels  in Vietnam
veterans with  those found  in similar groups  of non-Vietnam veterans and civilians.  The
VA/EPA study  was a retrospective study based on archived  NHATS specimens.   Only
specimens from male donors born between 1936 and 1954 were included in the VA/EPA study.

       Despite the differences in study objectives,  it  is  possible to compare the  average
concentrations found in the  VA/EPA study and those found in the 15-44 year age  group from
the two NHATS surveys.  The donors in the VA/EPA study were all between 17 and 46 years
old at the time of their death or surgery.  In Section 8.0 it was concluded that the only factor
consistently  affecting concentrations of  dioxins and furans was  the age of  the  donor.
Additional information on the programs  are presented in  Section 9.1.   Comparisons of  the
study designs, chemical  analysis procedures,  and the significant results  are presented  in
Sections 9.2, 9.3, and 9.4, respectively.
                                        107

-------
9.1    OVERVIEW OF THE ANALYTICAL PROGRAMS

      The FY82 NHATS specimens were  analyzed as composites as part of a broad scan
analysis program conducted to  expand the utility of the NHATS  program beyond  the
monitoring of  organochlorine  pesticides  and  PCBs.   Forty-six composite samples  were
analyzed for tetra- through  octachloro PCDDs and PCDFs.  The FY82 effort was designed to
provide  body burden estimates  for the general U.S.  population  based on age, sex,  and
geographic region.

      The VA/EPA study used approximately 200 individual specimens collected from 1971
through 1982.  The specimens were from adult males with birthdates between 1936 to 1954
who potentially might have served in the Vietnam War and who possibly had been exposed
to the herbicide Agent Orange.   These  specimens  were categorized into three  groups:
Vietnam veterans,  veterans with no military  records indicating service in Vietnam,  and
civilians.  The design  of the  study was intended to determine whether there was any possible
difference  in the levels of 2,3,7,8-TCDD between groups.   The analysis program, however,
provided data on all of the 2,3,7,8-substituted chlorinated dioxin and furan congeners.  Hence,
this data base has generated a considerable amount of information that can be compared with
other data bases.

9.2  COMPARISON OF STUDY DESIGNS

      Similar sampling designs were used for collecting tissue specimens in the FY82 and
FY87 NHATS.  Both studies used a multi-staged sampling plan.  The conterminous 48 states
were divided into  strata; MSAs  were selected with probabilities proportional to size;  and
cooperators were solicited and assigned  quotas for collecting specimens.  There was a minor
difference  only in the method of stratification. Prior to the FY85 NHATS, MSAs were selected
from strata denned by U.S. Census divisions.  Beginning with the FY85 NHATS, sampling
strata were redefined to be the 17 geographic areas that resulted from the intersection of the
nine Census  divisions  and the  ten  EPA  Regions (Panebianco DL,  1986a).   A controlled
selection technique, known as the Keyfitz technique (Mack et al, 1984), was used to maximize
the probability of retaining MSAs used in previous years.  As a result, there were 47 MSAs in
the FY87 design compared to 35 in  FY82. Otherwise, the sampling designs for the FY82 and
FY87 were essentially the same.

      A total of 763 specimens were used to generate the composites for the FY82 study, and
865 specimens were included in the composites for the FY87 study.  As shown in Table  9-1,
the distributions of specimens among the various geographic and demographic subpopulations
were similar.
                                         108

-------
 Table 9-1.  Marginal Comparisons of FY82 and FY87 NHATS Individual Specimens Used
                            for PCDD and PCDF Analysis
Category
                                No. of specimens (%)
FY82
FY87
 1980 Census
population (%)
Census Region

  Northeast            166(22)
  North Central        206(27)
  South               331(43)
  West                _60(8)
   Total               763

Age Group

  0-14 years            178(23)
  15-44 years           312(41)
  45+ years            273(36)
   Total               763

Sex

  Male                412(54)
  Female               351(46)
   Total               763

Race Group

  Caucasian            632(83)
  Non-Caucasian       131(17)
   Total               763
                     175(20)
                     296(34)
                     289(33)
                     105(12)
                     865
                     146(17)
                     318(37)
                     401(46)
                     865
                     436(50)
                     429(50)
                     865
                     707(82)
                     158(18)
                     865
                             26
                             22
                             33
                             19
                             23
                             46
                             31
                             49
                             51
                             83
                             17
                                        109

-------
       The FY82 and FY87 NHATS had comparable compositing designs.  One of the design
criteria for compositing FY87 specimens was to maintain similarity  to  the FY82  design.
Table 9-2 gives a comparison of  the  marginal percentages of composites in  each of the
categories defined by the four analysis factors.  Population percentages from the 1980 Census
are also provided.

       Overall, the marginal percentages from the FY82 and FY87 NHATS agree reasonably
well with each other and with the census figures.  The only differences are the FY87 NHATS
had more "pure sex" composites (31 versus 11) than FY82, while the FY82 NHATS had more
"pure race" composites (17 versus 8) than FY87.

       Because the VA/EPA investigation was a retrospective study using surplus  specimens
from the  NHATS archives,  the method used to select  specimens was  different from the
NHATS sampling strategy.  Of the approximately 8,000 unused NHATS specimens collected
between 1971 and 1982, 528 were collected from males born between 1936 and 1954.  However,
there was sufficient  background information on only 494  donors.  Specimens from 40 donors
who served in Vietnam were  selected,  along with randomly selected specimens from 80 non-
Vietnam military veterans.  Finally, specimens from 80 civilian men were included in the study
by matching the birth year (2 years) and sample collection year  (2 years) of two  civilians
with each Vietnam veteran.  Thus, a total of about 200 specimens  from male donors between
the ages  of 17 and 46 formed the  basis for the VA/EPA study.  Of the specimens identified,
successful analysis was achieved for 197 individuals.

       Another major difference between the two NHATS and  the VA/EPA study is that
specimens in the NHATS studies were  composited prior to chemical analysis, while specimens
selected for the VA/EPA study were analyzed individually.  This difference affects the way in
which  the data are  statistically analyzed, but, as discussed in Section 9.4.2, it  does not affect
the comparison of average concentration levels.

9.3  COMPARISON OF ANALYTICAL PROCEDURES

       To compare  the  data  for the three studies, it  is  necessary to  review the analytical
procedures (see Figure 9-1).  While the analytical procedures used for the VA/EPA and FY87
efforts  were fairly comparable, the figure illustrates that the FY82  approach was considerably
different.  The changes in the analytical procedures were incorporated in VA/EPA and FY87
studies as an effort  to improve the state-of-the-art of the analytical technology between the
time frames that each study was conducted.
                                         110

-------
    Table 9-2. Marginal Comparisons of FY82 AND FY87 NHATS Composite Designs
No. of Composites (%Y
Category
Census Region
North Central
Northeast
South
West
Total
Age Group
0-14 years
15-44 years
45+ years
Total
Sex
Pure male
Mixed
Pure female
Total
Race Group
Pure Caucasian
Mixed
Pure Non-Caucasian
Total
FY82

12 (26)
9(20)
19 (41)
-L(13)
46

12 (26)
17(37)
17_(37)
46

6(55)
35
_5_(45)
46

11 (65)
29
6(35)
46
FY87

15 (31)
9(19)
16 (33)
_8_(17)
48

11 (23)
17 (35)
20 (42)
48

16 (52)
17
15 (48)
48

8 (100)
40
-Q-( 0)
48
1980 Census
population %

26
22
33
19


23
46
31


49

51


83

17

The percent estimates for sex and race groups are calculated as the total number of pure
composites within each study design.  For example, 6 of the 11 (55%) pure sex composites
in the FY82 study design were composed of males only.
                                      Ill

-------
9 -13C IQS
1 0g Composites


CO
o
2 s
* 
Oco
CO J
CM
X



Neutral
Alumina


CO
_o
2-33
Stf
X
<

CO
is
X 0
 62
| =
                              UL  t LL   ; LJ_
                              r? Q Q Q Q
                              g QQQ g
                              gooog

                              "
                                                             >
                                                             LJJ
                                                 5
                                                 QJ
                                                 5
                                                 U3

                                                 IN
                                                 oo
<
Q_
III


CO -2
2 If
w <55
"~" O)
C31 0






CO
o
2 g
0*"=
CO00
CM
X






CO
s?










6
1SB
Q. 
O CU
-90
CO
O

                     CO
=i: o
O 
CD OC
DC
X
LL u. u. LL^

r? Q Q Q Q-

Q ,9 R ,9 D
                                                 ea


                                                2?
                                                cs
                                                oo
                                                13
                                                o
                                                T3

03
Q U. Q '"
QQQ Q.
OOO 
HHO 0
666 "
co co ro ^
'"'""" CM



O
Q.
CD






'to
o
u_

 
PX-21/Glass
Fiber

6
iSfl
8" 03
SO
CO
O
                     CO

                     DC O
                     Is
                     o 
                     O oc
                     DC
                     n:
                     CO
                     2
                     cc o
                     T O
                     =t o
                     O co
                     CD CC
                     cc
         Q Q ?5
         Q Q g

         OOo
        O *-*

        HO?
                                                13

                                                CO
                            O
                   ta

                   I
                   o
                  U
112

-------
       Each  procedure  required  fortification with internal  quantisation  standards  (IQS),
 extraction, removal of bulk lipid, and separation of interferences from the PCDDs and PCDFs.
 The techniques for all three studies were essentially equivalent.  Extraction was achieved with
 methylene  chloride using a Tekmar Tissuemizer to promote thorough extraction of  lipids.
 Bulk lipid  removal for  the  FY87 and the VA/EPA studies  was conducted  using  identical
 techniques, consisting of treatment with sulfuric acid-modified silica gel slurries and further
 cleanup via a chromatographic column of the same material. Gel permeation chromatography
 was used to remove the bulk lipids in the FY82 composites.

       The separation of chemical interferences  was achieved  using Florisil (FY82 NHATS),
 acidic alumina (VA/EPA), or neutral alumina (FY87 NHATS).  Neutral alumina was  used for
 the FY87 samples rather  than acidic alumina to improve method recovery and  reduce possible
 background contributions due to hepta- and octachloro-PCDDs. Previous efforts using Florisil
 on  the  FY82  composites  had  demonstrated  poor  recovery   of  the  hexa-  through
 octachloro-congeners (USEPA 1986a).

       For the final cleanup of sample extracts, a carbon-based column was used.  However,
 as noted in Figure 9-1, three different carbon adsorbents were used between the studies. Two
 separate extracts were cleaned for the FY82  NHATS  composites.   Because recovery  of the
 higher chlorinated compounds was poor,  an  aliquot of the extract taken  through the GPC
 cleanup, but  not through Florisil chromatography, was taken through a PX-21/glass fiber
 column to determine the  hexa- through octachloro-PCDDs and PCDFs.

       The AX-21/silica gel column used for the FY87 NHATS composites did not provide the
 degree of cleanup  demonstrated with the Carbopak C/Celite used with VA/EPA.   This was
 primarily noted through the detection of  octachlorodiphenylethers that interfered with the
 determination of 1,2,3,4,7,8-HxCDF and  2,3,4,6,7,8-HxCDF in  the  FY87  composites.   The
 HRGC/HRMS conditions varied across  studies (Figure 9-1), depending on whether  analyses
 were conducted using a  mass resolution of R = 3,000 or R =  10,000.  The higher the R  value,
 the more specific the analyses and the higher the confidence in compound identification. The
 pattern or  fingerprint of the major PCDDs and  PCDFs observed in the  HRGC/HRMS
 chromatograms for human adipose tissue was consistent across all these studies.

       Two factors have the largest potential effect on data  comparability  among the three
 studies:  the  type  and  number  of IQS  and  the  consistent use  of analytical  standards
 (Figure 9-2).  Only three IQS compounds were available for the FY82 composites.  Since the
 calculation for PCDDs and PCDFs is based on an isotope dilution principle, the limitation on
the FY82 composites is the assumption that all compounds  will recover the  same as  the IQS.
Recovery data for the additional IQS compounds in the  FY87 and VA/EPA studies demonstrate
                                        113

-------
                                  FY82
                                 NHATS
EPA/VA
 FY87
NHATS
2,3,7,8-TCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDD
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDD
OCDF

 Internal Quantitation Standards
 13C12-2,3,7,8-TCDD
 13C12-2,3,7,8-TCDF
 13C12-1,2,3,7,8-PeCDD
 13C12-1,2,3,7,8-PeCDF
 ^C^-I^.S.SJ.S-HxCDD
 13C12-1,2,3,4,7,8-HxCDF
 13C12-1,2,3,4,6,7,8-HpCDD
 13C12-1,2,3,4,6,7,8-HpCDF
 13C12- OCDD

 Internal Recovery Standard

 13C12-1,2,3,4-TCDD
 13C12-1,^2,3,7,^8,9-HxCDD
                                                          Quantitative
                                                          Qualitative
 Figure 9-2.    Comparison of analytical standards from the FY82, VA/EPA, and FY87 studies.
                                    114

-------
that PCDDs and PCDFs recoveries differ depending on the degree of chlorination. For these
reasons,  only  the  data for 2,3,7,8-TCDD, 2,3,7,8-TCDF,  and OCDD are directly  comparable
between the other two studies.  Since the same sets of IQS standards were used between the
studies, the FY87 and VA/EPA studies are comparable.

       The second factor,  standard traceability, is another  serious consideration.  Figure 9-2
shows that all standards are directly comparable between the FY87 and the VA/EPA studies.
When considering the standards for the FY82 composites, only the 2,3,7,8-TCDD  standard is
directly comparable across all three studies.  The standards analyzed with the VA/EPA and
FY87 studies (including the 2,3,7,8-TCDD for FY82) were verified through  participation in
interlaboratory comparisons  and the  analysis of an NBS standard reference material.  The
results of these interlaboratory  studies support the quantitation of results reported in these
studies and also promote the comparability between human tissue data  sets generated by the
other laboratories participating in these studies.

9.4 COMPARISON OF RESULTS

       The results from the FY87 NHATS, the FY82 NHATS, and  the VA/EPA  studies are
compared in this section.   Because  the .same study design was used  for the two  NHATS
surveys,  it is possible to make a more detailed comparison of those two sets of results.  A
statistical comparison of the FY82 and FY87 results is presented in Section  9.4.1. The VA/EPA
results are compared with the FY82 and FY87 NHATS in Section 9.4.2.

       9.4.1  Statistical Comparison of FY82 and FY87 NHATS Results

       The results from the FY82 and FY87 NHATS were statistically compared to determine if
there were significant changes in average PCDD and PCDF concentration levels over the five-
year period.  However, as discussed in Section 9.3, advancements were made in the analytical
method  for  the FY87  survey.   The  most  significant change was  that  additional  internal
quantitative  standards  (IQSs)  were available  for the  penta-, hexa-,   and heptachloro-
compounds.   Therefore, only results for  the tetra- and octa- compounds  are expected to be
directly comparable. Statistical comparisons were performed on all compounds for  which data
were  generated in both years.   Comparisons were made, not only between  the predicted
national  averages,  but  also between the demographic profiles from  the  two surveys.  The
profile analysis, discussed in  detail later (9.4.1.3), examines possible changes in the differences
across demographic subpopulations.  This type of comparison is valuable even if systematic
differences in concentration levels exist that can be attributed to the analytical methodologies.
                                         115

-------
       Statistical comparisons were possible only when sufficient data were available from
both surveys.  The criteria for performing a model-based comparison were:  (1) the chemical
must be detected in at least 50% of the composite samples each year/ and (2) there must be at
least 30  analyzed composite samples  in  each year.   The  additive model,  described  in
Section 7.0, was used for the model-based comparisons between the FY82 and FY87 NHATS.
Since  the  FY82 data  were originally analyzed using a  multiplicative model (EPA  1989),  the
FY82  results  presented  in  this  comparison are different from  those previously  published
(USEPA 1990a).

       Table 9-3 shows  the type of comparison used for each of the  target chemicals.  Six
analytes met the criteria for a model-based (M) comparison.  The other four analytes, analyzed
in both years, were compared in a descriptive manner using weighted averages (WA). This
approach is discussed later.  (See 9.4.1.4.)

       As  discussed in Section 8.0, data restrictions derived from the  data quality objectives
were imposed on the FY87 data but were not applied in FY82. Thus, in Table 9-3, the number
of composites listed for  FY87 represents the number with unrestricted measurements, while
the number for FY82 is the total number of available composites.

       The comparison of FY82 and FY87  results is divided into four parts.  Section 9.4.1.1
compares the FY82 and FY87 results in terms of limits of detection (LODs) and the percent of
composite samples for which  each analyte  was detected. Estimates of the national averages
for the six analytes statistically  analyzed  using the additive  model  are compared  in
Section 9.4.1.2. The results of the profile analyses are  presented in Section 9.4.1.3. Finally, in
Section 9.4.1.4, a descriptive comparison of weighted  national averages is presented for  the
four analytes that were not statistically modeled.

       9.4.1.1  Comparison of LODs and Prevalence  Detected.   Table  9-4 compares  the
percent of composite samples  in which the analytes were detected and the average  detection
limit (LOD) for each year.  In FY82, LODs  were only  calculated when the concentration was
either  not detected (ND)  or qualified  as  a trace  (TR) value.   Thus, the  sample size  for
calculating the average  LOD in FY82  was often much less  than the number of samples
analyzed.  For example, TCDD was either not detected or found at a trace level in 13 of the 43
composite samples for the FY82 study.  The average LOD reported for the 13 samples was
7.28 pg/g.  In FY87, LODs were calculated for all composite samples.
                                         116

-------
             Table 9-3. Number of Composite Samples and Types of Statistical
                Comparisons Made Between FY82 and FY87 NHATS Results
Number of composite
samples
Analyte
2,3,7,8-TCDF
2,3,7,8-TCDD
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
HxCDFd
HxCDD"
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDD
OCDF
OCDD
FY82a
43
43
-
43
41
45
45
45
-
45
45
45
FY87"
33
36
43
39
35
9-45
39-41
27
46
42
23
32
Type of
comparison0
WA
M
-
M
M
WA
M
WA
-
M
WA
M
a  Number of available measurements. Two outliers were removed for 1,2,3,7,8-PeCDD.
b  Number of unrestricted composite sample, measurements.
c  WA = weighted averages
    M = model results and profile analysis
    -   = no data available for FY82
d  Analyte analysis results  for  specific isomers  of HxCDF and HxCDD were combined
   (summed) for comparisons between FY82 and FY87.
                                       117

-------




^
00
-d
C
ro
rq
00
fe
H
^1
H"^
J-j
Z
c
01
o>
_g
QJ
CQ

p
P
U
9
nj
yj
P
P
U
P*
JH

aj
4-1
0)
4-*
QJ
P
4-*
QJ
IH
QJ
PH
I ,
C,
ro
en
Q
O
* 1
"4-H
O
eft
r-
*H
O
en
c
ro
pi j
c2
0
U

<*
ON
QJ
3





.S
QJ 4-1
bO 
6 S
S VH"
U o









tx
oo
PH





















CNl
00
\^
PH




















T3
cu >  v
0 Q
H-d S


-^

" ol
bO >ii?
ro
1-1 P
Is
<. 1 1





4- -d
^ 4*
C 4J
QJ CJ
U QJ
IH 42
QJ QJ
PH "d



z



, ,
x?~
aj p^
ro ^~"
*""* P
< J


Z




4- "d
QJ tj
O QJ
QJ QJ
PH "d



a
Z






_QJ
>~i
"ro
C
*c


* *
^ CO . rx.
K C<1 ' l>1






m T-H <* in
O ON CO 00
CS CN! ^1 CO
0000







R tx ^ in
2 ON r-i ON
T-H




CO NO CO ON
CO CO Tj< CO






^^ OO
ON r
PH
OO
tx"
CO
^i.
1-H



oo






NO
ON
NO
O







K





^







o
tx
i i


ON
r-l





T-H
tx




in






PH
P
U
><
PH



s






bO
ON
i i
T I







g
1-1




tx







CO
ON


<*





OO
ON




in






P

U
X
J-i



^ , oq 0 o
i







o T-H tx, tx in
O <* O NO tx
T 1 I 1 * 1 1 1 CO







ON ^ o o o
oo ^ O ro O
1-H T 1




tx NO CNJ co cs
CN) ^ "^ CNl CO






rH
0 00
CT^ ' \O CT^ '
C^ ^H


IX r-H ^ 0
i fn





m ' oo o o
QN Q\ <^JH ^^




in m m m





[*T j fTj J]
P P P
U U U
OH OH OH
oq^ ON^ oq^
tx OO tx
rjn" TJ? T^H" PH U
co" co" co" P P
CN? CNf C
S

in
en 0
P <=>
O 15
1-3 o
*T3 *"
-S . N
TO tft _,
r-d QJ W
d 1  ,3
N^ at
^x a 2
2 i o 
d > o QJ
ro ^ H
CO CO QJ
4-J H^ "u -52
'co en "C QJ
O O *3 o
i-ris
83 |
"o o ^o "^
S3 S S3 J
rg rQ ,0 ro

3 3 3 "3
Z Z Z 
rt ,Q u T3









c
0

ro
OH
a
o
o
g
<4H
*^->
*^3
a
a
d
xS
d
c
1
o

QJ
S3

Q
P
(J

*T3
d
8
^s
' i '
M-H
o

QJ
O
eft
U
*r^
S
>S
e/5
^ tx
^ ?2
S p^
VH
'^ 
IS
1^
1 
tu ^

U rO
0.





















.
*d
4J
CJ
0)
4-1
o>
d
QJ
JJ
^
2
M
0)
a
en
-rH
P'"^
P
U
^^

1

lH
O
^ ^
PH
P
x
E.
Ui
^
^o
QJ
MH
O
K^
J
ro
vy

CJ
"S
5
"is
C
<


























P
O
j
QJ
bO
ro
IH
>
ro
o>

QJ
"ro
3
o
13
CJ
-d
4J
en
^"1
CO
-H
en
>H
OJ

.S
r^
i ^ j
M
1
eft
^""^
O

ftj
^-
"o
1

f
H
bO
118

-------
       There clearly was a significant improvement in the sensitivity of the analytical method
in FY87. The average detection limits in FY82 were in the range of 9 to 50 times higher than
those in FY87.  This, most likely, explains the statistically significant increase in the percent of
samples in which TCDF and TCDD were detected between FY82 and FY87.  For example,
2,3,7,8-TCDD was detected in only 74% of the samples in FY82 but was detected in 97% of the
samples in  FY87.   This difference (23%) is  statistically significant; however, the  average
detection  limit was decreased  from 7.28 pg/g to 0.291 pg/g.    The  average 2,3,7,8-TCDD
concentration in both years was estimated to be around 5.5 pg/g.

       9.4.1.2 Comparison of National Average Estimates for Modeled Analytes. Table 9-5
shows the estimated national average concentrations for FY82 and FY87 and  the estimated
difference (FY87-FY82)  in concentrations for  the  six  analytes statistically analyzed.  The
standard error  of  each  estimate and the significance level for testing that the difference is
different from zero also  are provided. The test was based on the approximate t-statistic of the
form
                                        NA87   NA82
                                 t  =  	
                                      {
where NA82 and NA87 are the FY82 and FY87 national average estimates and SE82 and SE87
their standard errors,  respectively. Approximate significance levels were calculated using the
standard normal distribution.

       Generally, levels less than 0.05 are used to indicate statistical significance.  Using this
criterion,  the average predicted concentrations of 2,3,4,7,8-PeCDF, 1,2,3,7,8-PeCDD, and
HxCDD  are  significantly  lower in  FY87 than in FY82.   However, as mentioned  earlier,
additional internal quantisation standards were used in the FY87 analytical procedures.  Thus,
we cannot conclude that the  average concentration in the U.S. population changed between
FY82 and FY87.  The differences may be due only to the  changes in the analytical method.
Data from the VA/EPA  study  also demonstrated significantly different concentrations for
2,3,4,7,8-PeCDF and  1,2,3,7,8-PeCDD  comparable  with the  FY87  levels.   No significant
differences in the average or predicted levels of 2,3,7,8-TCDD and OCDD were noted between
FY82 and FY87.  The same internal quantitation standards  were used in both years for these
specific analytes.
                                         119

-------

M-l
O
cn
d

H
C
C
cu
CJ
C
0
U
cu
bo
ro
QJ
>
< a
"3 1
C -22
0 H
'5 ^
7 0
2 S,
T3 :=3
(U ^3
& <*
CJ ^*
' 
xj *3
QJ TO
LH Ci
 |
IS. l-TH
00 I
>> c
PL, .n

13 rSP
r; P-i
so
rq U
oo Pi
"
" S
 2
< Q
a
zy
*-, CH
0

VI
C
o
w
c
(H
ro
a,
g
o
U
in
ON
O)
i *
-TO
H



13
o
43
CO
"J3
ro
4-t
en










OJ
CJ
c
a
Ol
iK
P





cu
OJD
ro
tH
o>
>
ro
"
f-
#H
O
'^U
c 1
tJ 
o>
^ *
ro
s
'^J
CO
td












cu
CJ
q
S

^
.
'S3









CQ
00
>H
UH
tx
oo
C.
^







00
oo
>H
H-I











cN
oo
>
fr .
M-(











CU
>
^Oj
)








^)
uT
x2-



iri
^A
P


(0
w
en







bb
>
<


a
s~
en,





bb
^






2
j>,
"ro
^


en I-H r-i m ^ tN
T-I o o ^ c5 o
f i-i m
o o cs o o o
V







X V
cs NO" ^^ ^^ ,  * f-^
NO NO O CO C3 IS,
o en ci C3 in T*
"~" ^ S CS Ci xS


ON
ON
rt ts ON OO vO ON
cJ in tN ^ T-H tri
' cN SO CO CO ^t
1 1 1 1 1

o" o^ NO" ,  , s-^.
m o cs NO ro s-*.
ro 00 -^ i-< Cs NO
o, ci o^ ro ro^ oo'
CN^





oo o
ro Is. ts oq
in ON O NO O r*
I-H 00 T-I (N
l-H IS.


^~ - 
ff\ t**^- J ^
C* J L^s X^^s x 'v X^V <^^X
in in o -i oo ts,
O ro' O O <* ON
s '  < cs" U
CS" CS" rl H- r-n" O




















o

'C
ro
D,
a
o
CJ
t-l
&

^
cu

c

a
(0
*^^*
-a
cu
1
a
8
cu
O JH
So aj
ro S
 ^
S P
ro . P
^ u y
- H^
ro CU "-"H
^^
J3 HH
 3 2
^5 S Cs'
"o o a oo
, 	 , O >H
to bc.23 PL,
~b fio o 13
r*i d. ' T; r*
J_L, M-i'-H tn
x-xs^*d 
gg^g
cu cu H p.,
^ "2 ^ c
ro ro j2 3}
T3 T3 ^ |
| I S 1
en en pi ^
as A u
120

-------
       9.4.1.3  Results  of Profile  Analysis.   Profile analysis  is  a multivariate statistical
 technique that is used to  compare profiles of two or more populations.  In this context,  a
 profile is a vector of estimates of subpopulation averages.  An example of a profile is the set of
 estimated average TCDD concentrations for the four geographic regions (NC, NE, S, and W)
 in the FY87 NHATS.

       Profile plots were used to make a visual assessment of the differences in the FY82 and
 FY87 results.  In order to determine what differences may exist and whether these difference
 are statistically significant, a formal profile analysis was conducted.

       Profile analysis tests  a sequence of statistical hypotheses  to determine  how two  or
 more multivariate populations may differ. The hypotheses address the following questions in
 order:

       1.      Are the profiles from the two fiscal years parallel?
       2.      Assuming the profiles are parallel, are they coincident?
       3.      Assuming the profiles are coincident, do they have equal levels?

 Two  profiles  are  said  to be  parallel if  all  pairwise  differences between  the   average
 concentrations across the levels of a demographic factor are  equal.  For  coincident profiles, all
 pairwise differences are equal to zero. Finally, coincident profiles are said to have equal levels
 if there are  no differences  among the average  concentrations  at  different level of  a
 demographic factor.  The different types of profiles are illustrated in Figure 9-3.

       Profile tests  are performed sequentially.  The hypothesis of  coincident profiles is only
 tested if the hypothesis of parallelism is not rejected. Similarly, the hypothesis of equal levels
 is only tested if the hypotheses of parallelism and coincidence are not rejected.  The third test
 combines  the FY82 and  FY87 data to test for significant effects of the demographic factors.
 This  approach described by Johnson and Wichern  (1982) was used to conduct the profile
 analysis.

       To provide the  background for the profile analysis, Table  9-6 compares the FY82 and
 FY87  significance levels from testing for  differences among demographic groups.  The additive
 model was used to perform the tests, with the assumption of normally distributed errors for
both years.  Although  these assumptions were found to be reasonable for the FY87 data, the
large  measurement errors  from the chemical analysis of FY82 composites  indicate  that the
assumption of normality may not be true for the FY82 data.  However,  because of these data
in FY82, it is difficult to verify any distributional assumptions.
                                          121

-------
Average
Concentration
                          Parallel Profiles
                    NC      NE      S
                          Census Region
W
                                                      Legend:
                                                        D=FY82
                                                        A. FY87
Average
Concentration
                        Coincident Profiles
                    NC      NE       s
                           Census Region
W
                                                      Legend:
                                                        D=FY82
                                                        A=FY87
Average
Concentration
Equal Profiles
 i  i i 	 i
Legend:
D=FY82
A=FY87

                    NC
W
            NE      S
           Census Region

Figure 9-3. Example of profile plots.

               122

-------





co 42
'55 v
CD ^
0 ^
&p
^! C_J
c ^
p^
S 
43 to
350
|8
CO "O
u CD

CO CO
O ^H

" tn

In **"


>-i 5
ft lH
a O
C o
to -g
00 ^

S 8
S3 i
i P
2 
"o >
to -P
ci cu
o oa
CO -
"C co
CO U
ft, C
 **
O Ctl
u ig
SO G
O 4-

H H







C
C
1

oo
*
oo
uu


















0,


CO
tj
QJ
>*7
^
















p>
C
X
CO
cyj


IN
oo
f^



oo








o
3
o
ol

CO
w




oo

ft






tN
00




C
BO
s
CD

^
^3
c
CO
u

IN
OO





OO
ft







"re
^
I-H m m in in en o
^* CO Os Os Csl Os CN
en so in en Os en in
o d d J d o d o
rH IN in IN in en
so in oo so en in
00 SO Os OS t^ Os
d d d d ! ! do




tN en IN * IN ^ in
en o rq en os csi os
rq so so os in rj< ^
d d d ! d d d d


c P
0 p p ;|X r
Q u u o ir ~. ^
P O) O) H 0& OS 00
U PH ft " ' tC OO" tC
H oo e P rti' IN so" _
v *\ ^ P m" en^ * P
ts* TJ^ en ^} f>i cq en '~~*
Csl CSI rH MH rH s^















.a
^
i -j

c
to
fH
Ol

o

CO
>

^i9
T3
CO
1
OJD
S -H
'5 ^
3 ^
a! in
-q '
13 
CJ ^
Oi fO
I.
^H fll
^ 1
^1
O o
t, g

~ s
Co

Co ^
cj co
C 'C
CO CO

c/5 H
B ^
123

-------
      In  FY82, measured concentrations for four  of  the six  analytes were found to be
significantly different among the four census regions.  In FY87, 2,3,4,7,8-PeCDF was the only
analyte with significant regional differences.  On the  other hand, in FY87, age effects were
much more evident.  In FY82, only 2,3,7,8-TCDD and 1,2,3,7,8-PeCDD showed significant age
effects at the 0.05  significance level,  and 2,3,4,7,8-PeCDF  and OCDD were significant at the
0.054 level.  None  of the analytes in either year showed significant differences in measured
concentrations among the different sex or race groups.

      Profiles of the six modeled analytes in FY82 and  FY87 are plotted for each of the four
analysis factors (region, age, race, and sex) in Figures 9-4 through 9-9.  The estimated average
concentrations within each analysis factor and fiscal year are connected with straight lines.
The vertical lines define approximate 95% confidence limits of the estimates.  The confidence
limits were calculated by adding 2  times the standard error of the estimated average.  The
estimated national averages with 95% confidence limits are also plotted.

      The numerical resultsestimated average concentrations and their standard errorsof the
profile analyses  are presented in Tables  9-7 through 9-10. The estimated difference  (FY87-
FY82) and their  standard error for each region are also presented in each table.  Finally, the
significance levels of sequential tests  for parallel profiles,  coincident profiles, and equal levels
are provided.

      Profile analyses were performed for five of the six modeled analytes. Total HxCDD was
not included,  because in FY82 only the total HxCDD was  measured, while  in FY87 each
individual isomer was measured.  To perform the profile  analysis on HxCDD, it would have
been necessary to  combine the individual isomer data  for FY87.  This would have required
making certain assumptions that would produce extremely conservative results.
                                          124

-------
   CD  00

   Cb  b*
O  II
 c
 (D
 ^^  O
 0)
J
U
H
  
oo
                                                                                                 X
                                                                                                 0)
                                                                                                CO
                                                                                                 o
                                                                                                 CO
                                                                                                cc
                                                                                            in
                                                                                 in
                                                                                     g
                                                                             - co   ">
                                                                                     CC
                                                                                     en

                                                                             -Si
                                                                                     0)
                                                                                     O
                                                                             _ O
                                                                                                c
                                                                                                o
                                                                                                CO
                                                                                                         8
                                                                                                         B5
                                                                                                         C
                                                                                                         O
                                                                                                        J3
     .
 C  "*
 P-
   0,
D  |
Q  S
y  
^T  3
00  g
                                                                                                         J
                                                                                                            g
                                                                                                        a  c
                                                                                                           00
                                                                                                         2  ^
         CO
                                               CO
                                  (6/6d) uoi
                                                  125

-------
            00  CO
            >  >
          DI
          0}
                                      G	^6	D
                                       X
                                   G	g	E3
                                                                  >  _
                                                                        |





                                                                      f-5
Q
U
00
  s
                           C3-
                                   -^
                                    D-
                                                                           X
                                                                           CD

                                                                           CO
                                                                           CD
                                                                           s
                                                                           rr
                                                                        IT)
                                                                        irt  03
                                                                      - CO
                                                                            c
                                                                            o

                                                                           "S1
                                                                           cc
                                                                            to
                                                                            13
                                                                            (O
                                                                            c
                                                                            (U
                                                                           o
                                                                      _ o
                                                                        z
                                                                            c
                                                                            g

                                                                            OB
th 95% confidence
concentrations
es
                                                O.



                                             S  s

                                             U
                                             OJ
                                             Cu
tissue
                                             oo
                                             r>T  ^
                                             ^  U3
                                             rj<  O
                                             m .S"
                                             csf "o
                                                re
                                             QJ f^
                                             bCoo

                                             sfc

                                             s^
                                             *-  Q
                                             O rq
                                               oo
                                             J2 >
                                             _o u,

                                             ^s
                                             oj 
                                             !72  VI


                                             11
Figure 9-5.
        o
        00
                       o
                       fO
o
*t
o
CM
                             (6/6d)
                                           126

-------
               (N  r-
               00  CO
            O  II
             C

             0)
             0^ Q
             V
a
Q
u
00
  N
         o
         o
         CO
o
o
CXJ
                      D-
                                             -df
                                                              0	=&
                                                     - 0.
                                                                                       -5
                                                     - in
                                                                                        -5
o
o
                                                                                               X
                                                                                               CD

                                                                                              CO
                                                                                               8
                                                                                               CD
                                                                                               O)
                                                            g

                                                       co   ?

                                                            DC

                                                            co

                                                       "J   co
                                                       z   C
                                                            CD
                                                            O
                                                       O
                                                                                               o

                                                                                               "to
                                            01
                                            C
                                            o>


                                            1
                                                                      en

                                                                      O
                                                                     .a
                                                                      03
                                                                                                        nj
                                             o  a,
                                             0  
                                            O  g
                                            Q  a)
                                            U  3
                                                                                                        bJDoo
                                                                                                           C
                                                                                                           CO
                                                                                                        o>

                                                                                                        "5 U-i
                                                                                                        0)
                                                                                                        2
                                                                                                        a.

                                                                                                        (L>


                                                                                                        I

                                                                                                        E
                                   (6/6d)
                                                       127

-------
            TJ
            C

            0)
fN

00



fa-


ll
Q
Q
U
                                                                      +

                                                                     If)
                                                                                         X
                                                                                         CD

                                                                                         CO
                                                                                         o
                                                                                         nj
                                                                                         EC
                                                                      IT)   CD

                                                                      *"   O>
                                                                                         o

                                                                                         'o
                                                                                         0}
                                                                                         tr

                                                                                         CO

                                                                                         CO
                                                                                         C
                                                                                         CD
                                                                                         o
                                                                                   _ o
                                                                                  ^
                                                                                          C
                                                                                          o

                                                                                         "
                                                                                                 S-l

                                                                                                 0
                                                                                   1
                                                                                                 C

                                                                                                 8
                                                                                  in



                                                                                  ^^
                                                                                  ^




                                                                                   C
                                                                                                 a
                                                                                                 is
                                                                                                 c
                                                                                                 o>
                                                                                                 u 


                                                                                                 if

                                                                                                 D
                                                                                                  > -o
                                                                                                  W (5
                                                                                                  en
                            o
                            o
                              o
                              o
                              CM
o
o
                                    (6/6d)
                                                   128

-------
                 OO
                     r-
                     CO
              T3
              C
              V
              Ql
              0)
              J
                                                                                              x.
                                                                                              Q)
                                                                                              CO
                                                        o   <*
                                                                                              8
                                                                                              CO
                                                                                              DC
a
Q
u
 ft
a
00
                                                                 w
                                                            in
                                                            >   CD
                                                                                        -co
                                                                                              O
                                                                                              cc
                                                                                              (0
                                                                                              3
                                                                                              CO
                                                                                              C
                                                                                              CD
                                                                                              0
                                                                                        _ o
                                                                                              .0

                                                                                              'cS
                                                                                              2
           0
           o
0
0
CO
0
0
CM
                                      (6/6d) uoi
                                                     129

-------
   CM   P^

   CO   CO

   >   >"

   fc.   Cb
D  II
 C
 0)
 O>D
 0)
                        II
                                                                                                   LL
U
O
                                                                                                   in
                                                                                                         X
                                                                                                         03
                                                                                                        CO
                                                                                     in  
                                                                                                                   a
                                                                                                         0}

                                                                                                         
                                                                                                        EC
                                                                                                    en
                                                                                                    C
                                                                                                    O
                                                                                                   J3
                                                                                                    re


                                                                                                  C
                                                                                                                      8
                                                                                                                   8
                                                                                                                  O
                                                                                                                  R
                                                                                                                  U

                                                                                                                   a>
              0
              o
              CO
                   0
                   o
                   CM
0
o
CO
0
o
                                             (6/6d) uoi
                                                           130

-------





w
tu
Q
U
C
CO
g
U
C-,
U
V5
U-l
O
C
_0
S3
i c
c o
era
^ fti
U 
o 5

1^

tx
>- 3
p
M 
OO Q,
"" 5
||
S.I
6 
u c
' ' .
"53
15
c
<^

sG
1
"o
s.
s
(U
tt2
f<
o-
_aj
t2









1
13
1
o>
w
J

1
"c
ra
1

"oc
"Sc
a
"aT
^
u
Oj
Jj2
5








'S


/erage
ft
T3
ra
_|
1




















"3
1

c

*FJ
j
o



Is
1



rq

w-
^
^H
PH







tx
UH







00
u.
























n
"s
0.



1
Ci.


(S~





tJ2
S



sr
52-



^



w
52-



db
5


c

'5b
a
t/5
3
g
01
U







I
<.
O i | O O



* *
S S 0 M S
O O O O O
V


o o o o o


SSli" S2o rHtxvDM OOVCO^
r-.rHO" in<^mO KvgvOg OSCNIQ^ ^SSP3
v-i <-i rH (^



 S^S3!
III T-1 II 1

^?*cint^ mtxrt^ < tNOOoooj c^c^om co o *c * i



^ M w ,_ ^

tri vd iri * omo;^ o^oo ^gSo gggo
^^^
o^ f i \o . o CN fs
co-^^co Lnc^mm oo\o\oo r-imiriin
i i CN *  oo QO r^ vo


"n "3 13 "ra ~5
t-t U Ui t-i Ui
^(ft 1*^ "tfl yi trt
uw Uc3 cjw ut3 uw
XJ^j;^ JSX^jj XXXjj XXj;^., XXX^
"""3^ 'o'o'SS "o'o'^S 'S'o'SS "o'o'SS
ZZ^ ZZt5 ZZ5 ZZ^ ZZcrt^



Q
D
88^
D Q> Q> CO
U OH CU tC
S K Ji ": g
ts * mv m tJ

-------

a.

s
o
QJ
60
<
PQ
1
a
8
H
4)
VI
.s-
<
c
(D
S
ffi
_c
w
PH
Q
U
tx,
C
ra
to
O
p
U

"u
"3
en
"o
c
.2
S
'c
c
0
u
_a>
u
1
C-,
t^
>
T3
c
ft
CO
tL.
60
C
1"
o
u
.2
"Jo
"3
c
-f
fc v^
5




ST
^_^
00
tL.


>




~?~
52-
(N
?2
u.
1





Cu
3
O
60
OJ
So
<









QJ
^
(0
c
*
o

! ! ! o i






VC
! ! ! o i




* * * *
rH f-H ^ tN. CO
O O O O* O
V V



CO
tx T-I cO ON M CO
O3 en T < COcOCO i-*CNlfi
r-i T i ri




T- \O O CO
ii I ^-* i ii CO ' ri


aro'co' ^^^ ^f^cn' ^^^-,
COLnr^ OOO3O txvOt^ *~ ' O QQ K^ O O
o'o'o i < r-i r-i o'oo iriirjio iritxco
^^-^



00 t"* O tv O ON CO
ON CO "^ CC O O CN CO C4 tx CO
rn r-i **ONlx c- 1 ON LO




S-S-S- ^^Ct rJSiS Sco"< ^OO'LO'
""^^ ^"^ dC-C-


ON m TH
rn ON in ON *"- O C4 ITi
CO ^ C3 Ifl I~H p4 CN ^f ITI CO NO *^
T-H r-t r-t r-i 1C tN ON






^^U >^^H >^^w >^^u. X>^
^ 2 >i ^T*|X ^^JX ^4^J>> ^^J^
1~'i+ '"".+ *? i + ^7.+ r7,i
' iri iri ' iri mi ir\ 10 irs ICN irN (O
 i t ^ ^t 1^3 ^r- 1-^ ^r-<^< ^r-irtf





p
Q
LU O ^
n  Q E"
S u u .
P a; ai oo
U P* 0- r-J
H oi 06 vo"  .
, K K "*, 5
tx Tj< CO CO^ I  '
co*1 ccT c>f c-^ *^
CN! CNJ" r-5* r-" O





















































_j
>
-22
8
O
TO
OJ
u
c
u
"c
.&

(0
'J3
-
S
tft
S
I
a
c
*
132

-------
0.
a
u

^
pa


<0
<

1
B-.
'o
j3
3
S

ON
O^
1



1
"ro
C

3

tA
1
"o3
I!
"c
RJ
_u
1
in



^
!

(U
u

sS
S











1
^-
bJ
2
01
ra
"S
ra


W


























ra
3
a-
ta


j_,
c

'u
C



Is
"s

"^







a.
g
1

J








^
"ra
d  ^
! ! o o




* *
C3 r-t C3 CQ ^H
co o o o in
CO O O ^ CX
o o o o o
V



in ts ** c\ co
o o o o o



CO
\O \D O CO
CO ON O IX CO TH f1* ^J ^o

ON CN CN UJ ON


O O ^
1 ,, soin coi-j* coin



^^.^ 	 	 ^

M in K ^* o *- fs. ^^ ^-v
sO^O -<*CN txsO COfN O f 	 ^
o c*^ T~* in o CM in tx os in
^ ^G,



M 15 c5
i < rH i  O fN I-H in
FH r~l tx EN.
^
5^o~ ??S^ ^-^^-N  ^i3- -~~
&C- G-SG- S8 t9^ ^ fn'
^~* ^""^  i





o CN o -^ ro ro
roc^l txt*s T^^ OCT\
rl rH OO IT!




C C C C C

VI Ul VI t/1 
5
in
o
d

(0
s
ra
CC
I
en
"s
'g

1
I
1
133

-------


X
CO

at
a
w

^
o
-3
c
(0
E

(ft
Q
U
CH
C
W
tfl
D
D
g
-o
if Selecte
u
w
C
_o

2
"c

o
U
' Predicted
r^
CO
>H
a
c
cd

00
U.
oo
c
"E
S.
o
U
1/1
c
1
CH
'o
_
"3
OS
o
0^
JU
^
E-"




UH
0

-2
1
3


S
en
(U
JJ

C
ra
_y
I
tft



^^
1
C
U
5









5
a
ra
V
re


E
1




















3
$



C
0*

C
"o
U



"3
1
CH





U-<
IS
PH






CO
PH






C<1
00
PH



























S
"e
D.



8
5
8_





E*
52.


5




t/^




be


w
52-




bb






X
Ot









Si
"ra
C

55 CN t-.

o j i o o




* *
ON rn Cj CO
CO O ^ CN Ifi
o" o o o o





r-irH Ifilri QQ* r-*tri COM
P3 rq co T*< fn tx


I-H \0 ^ O
r-t O O^ ON tN CO r** C^ OO T-t
' CO T < vO Ifi ^O CO CJ
i i t i i r- <


x x1 \ x ^~N

oo *~" CQ rs co co co DN x~v s-~^.

& SS


rg oo o
CN rjn tN tx OO ^O

o^ /^^\ 
1^ II e?I II 





CO r-<
so M co co mo
iri^o oo OO'ON MO COM
rfCO IX\O OCO COON
T-* r-t \G 00




n>"(3 aj"(3 Qi*(3 Qj"^3 oj"(5
"" "" ""  "w  "" 
S tin ^PH S P-, ,2 BH S P-,



D
D
B. Q y.
n Q D l"
y u u >7
Q 0) at oo
U CH CH tC
t1 00 CO *O -.
oo rC tC rj? D
tC TK- en- en- D
en- en- ws csf U














































1
i
o

to
c
1
1
VI
~S
~
^
3
(A
&
s
1
*
134

-------
Census Region Profiles

       Table 9-7 shows  that  between FY82 and  FY87 there are no  differences  at  the  5%
significance  level  in  the  profiles  of  2,3,7,8-TCDD,  1,2,3,4,6,7,8-HpCDD,  and   OCDD.
Furthermore, based on the combined FY82 and FY87 data, there are no significant differences
in the concentrations of  these analytes among the geographic regions  (parallel profiles).  The
profile comparisons for 2,3,7,8-TCDD and OCDD are consistent with the individual test results
(see Table 9-6), which indicate no significant geographic effects for these analytes.  There was
a  significant  geographic effect for 1,2,3,4,6,7,8-HpCDD in  FY82, but the  effect was  not
statistically confirmed in  FY87.  However, the marginal significance (0.088 significance level) of
the  combined tests  for  geographic effects and the fact  that the South and West  Census
Regions had the lowest estimated average concentrations in both FY82 and FY87 suggest the
possibility of geographic effects for 1,2,3,4,6,7,8-HpCDD.

       The  profile  analysis suggests that the  differences in 2,3,4,7,8-PeCDF and  1,2,3,7,8-
PeCDD concentrations between FY82  and FY87 are  constant across  all  geographic  regions
(parallel profiles).  This could be due to changes in the body burden levels or, as suspected,
differences in the analytical methods. The hypothesis of coincident profiles was rejected at the
5%  level  for these  two  analytes.   This could be explained by systematic differences  in
measured concentrations.  As mentioned earlier, because the profiles for 2,3,4,7,8-PeCDF and
1,2,3,7,8-PeCDD  were  not coincident, the  combined  test for  geographic  effects was  not
performed.  However, in both FY82 and FY87, the highest estimated average concentrations
were found  in  the Northeast  Census Region.   The geographic effect was found to be
statistically significant for 2,3,4,7,8-PeCDF  concentrations in  both FY82  and FY87,  but  for
1,2,3,7,8-PeCDD, it was only significant in FY82.

Age Group Profiles

       The comparison of profiles by age is presented in Table 9-8.  The profile analysis rejects
the hypothesis of parallel profiles  at the 0.05 level for four of the five analytes, the only
exception being  1,2,3,4,6,7,8-HpCDD. Even in this  case, the test was marginally significant at
the 0.10 level. In FY87, the test for age effects (Table 9-6)  was significant at the 0.05 level for
each of the modeled analytes.  They were also significant or nearly significant at the 0.05 level
for the same analytes in FY82, but the estimated average  concentrations  of 2,3,7,8-TCDD,
2,3,4,7,8-PeCDF,  and  1,2,3,7,8-PeCDD  in FY82  were   not increasing  with  age.   Also,  the
estimated concentrations  of OCDD were increasing in FY82, the rate of increase is lower than
the rate observed in FY87.  Thus, even though there were significant age effects for these four
analytes in both  FY82 and FY87, the hypothesis of parallel profiles was rejected in each case.
                                          135

-------
       The hypothesis of parallel profiles for 1,2,3,4,6,7,8-HpCDD was only marginally rejected
at the significance level of 0.10.  In both FY82 and FY87, the estimated average concentrations
increased with age group, but the differences among age groups in FY82 were not statistically
significant, possibly because  of  large measurement  errors.   Also,  the  rate of increase  of
estimated concentrations of 1,2,3,4,6,7,8-HpCDD was lower in FY82 than  in FY87.  Assuming
that the profiles of 1,2,3,4,6,7,8-HpCDD are parallel, the hypothesis of  coincident profiles is not
rejected, but that of equal concentrations among the three age groups based on the combined
data is rejected at the 0.002 significance level.

Race and Sex Profiles

       The profile analysis comparing race groups (Table 9-9) and sexes (Table 9-10) produced
results similar to those for age  groups.   The findings  are:  (1)  there are no significant
differences in the profiles of 2,3,7,8-TCDD, 1,2,3,4,6,7,8-HpCDD, and OCDD between FY82 and
FY87; (2)  the  estimated differences in the concentrations  of 2,3,4,7,8-PeCDF and 1,2,3,7,8-
PeCDD between FY82 and FY87 are statistically significant and consistent across different age
groups and sexes; and (3) there are  no significant differences in the concentrations of any of
these analytes among different age groups and sexes.

       9.4.1.4  Weighted Average Comparison of Analytes Not Statistically Modeled. Four
analytes, measured in  FY82 and FY87, but which did  not meet the criteria for statistical
modeling  in  both  years, were  compared  using  weighted averages.   First,  the  average
concentration of composites in each of the three  age groups was computed for each analyte.
For example, the 2,3,7,8-TCDF averages in FY87 were 2.03, 1.34, and 2.50 pg/g in the youngest
to oldest age groups, respectively.  Next, these  averages were weighted by the population
percentages from the 1980 census.  For 2,3,7,8-TCDF, the weighted average concentration in
FY87 was

              1.86 (pg/g) = 0.23(2.03 pg/g) + 0.46(1.34 pg/g) + 0.31(2.50 pg/g).

This type of national average estimate is likely to be more accurate than the simple average of
the composite concentrations, because there was  strong evidence from the statistical analysis
of FY87 modeled compounds that age has a  significant effect on concentrations of PCDDs and
PCDFs in human adipose tissue.  Additional calculations  were needed  to compare HxCDF
concentrations, since each of  the HxCDF isomers were individually  analyzed in FY87, while
only the total HxCDF was measured in FY82. Thus, the weighted averages of the individually
measured isomers were summed to estimate  the total HxCDF concentration in FY87.
                                         136

-------
       Table 9-11, which gives the weighted average estimates and standard errors for each of
these  compounds,  shows  large  decreases  in the  average measured concentrations  of
2,3,7,8-TCDF and OCDF from FY82 to FY87.  The average measured concentration of OCDF
was 56.0 pg/g in FY82 and only 2.28 pg/g in FY87. In addition, there were significant advances
in the HRMS methodology for analyzing the FY87 composites. For example, as presented in
Table 9-4, the average  LOD for OCDF was  19.0 pg/g in FY82 and 1.67 pg/g in FY87.  The
standard errors of the  estimated  average concentrations are also considerably larger for the
FY82 data.  These facts  suggest the differences may be  due to changes in the analytical
method.

       There also are differences in the weighted averages of HxCDF and 1,2,3,4,6,7,8-HpCDF.
The averages are 47% and 35% higher in FY82 than in FY87 for the two analytes, respectively.

       9.4.2 Comparison of FY82, FY87 NHATS and VA/EPA Study Results

       The results of the VA/EPA study, as reported  by Kang et al. (1990)  and Bauer et al.
(1990), are presented  in Table 9-12 with data from the 15 to 44 age group from the FY82 and
FY87 NHATS. No statistical  comparisons are made, because the NHATS and VA/EPA studies
had different objectives  and data collection strategies.

       Table  9-12 shows the average  concentrations of selected  analytes  from specimens
collected in three-year  periods beginning in 1971  from the VA/EPA study.   The combined
averages for  all specimens collected between 1971  and  1982 are  also  presented.  Since  the
specimens were taken from male donors born between 1936 and 1954, the ages of the donors
were between 17 and 46 years.  These averages are compared with the average concentrations
from NHATS composites containing  specimens from donors in  the 15-44  year age  group.
These composites did contain specimens from female donors; however, there has not been any
statistical evidence linking  concentrations of  these analytes with the sex  of the  donor.
Therefore, comparisons  of the average concentrations from the NHATS  and VA/EPA studies is
possible.  The standard  errors of the average concentrations and the number of specimens or
composite samples used to calculate the averages are also presented in Table 9-12.

       There  are obvious differences  between the  NHATS  and VA/EPA results.  Except for
concentrations of 1,2,3,7,8-PeCDD in FY82, the NHATS averages are considerably lower than
the corresponding averages from the VA/EPA study. For example, the average concentrations
of 2,3,7,8-TCDD from FY82 and FY87  NHATS are 6.87 and  4.33 pg/g, respectively, while  the
average concentration in the VA/EPA specimens  is 14.1  pg/g.
                                         137

-------




w
en
S2
g
w
IB
T3 <
^5
+-1 7
en "^
__ tx
 oo
C s_.
re 
% ~r<
f"^ *J
SH ^
5 
re cs
 oo
C >"
5 ft
 -
c 8
o o
U 43
a> c
bOft
s o
S u
<^
< -a
T3 a>
- 
Jl 
SPw
a> u
"^ '^-t
5: o
^~H
rH
OS
O)
3
E2



















rx
oo
>H
ft







rrT*
M-J
cy}


xrs
3-<  ofc
n^ v ZLT
y 60,3
-^ cQ
s ^ a
 C
 rac3


Si,
z










es
oo
>
ft




uT
en


.0 "56
"S o> )fe
y M3?
P^ TO
SPSS a
o> > C
 racj




Z











QJ
r^
re
<
OS ^-s
<* ^~, ^^. 0
OS CS OO OO
O -^ CO Os
S- ii d S.



VO -a OO
oo r%  '**
rQ  V
T3  "5
!' fi
S -) &H g
_**< W^J j-- *H
U'a3 S <~
S> O tn
> u a>
8 8-S 2
& rt -s S3
t? * -d >
o Sn-5 re
tl  S -d
S S 3
o > B S
o g 3 *bb
M-H _ M^ *ni
o fro ^
>H S H **
01  OJ vw
' ^D'
Z <"Z en
n _o u "n
138

-------
>
01
J3
M >~>
S S
0 bO
2
"Of")
01 vJ
i s
r\ S^j
^? ^
0 c
n-S
1
^c3
"8 >N
4-t O
0 ^
0) en
V! IJ
O) H
en T3
WJ SJ
H H
O 
-+i en
S'fr
fa
m O
Cu
en 01
"*3P
o <
% *
? *
rl 1
c m
R *  i
l
^
li
en"^
*-l IX.
Poo
fc>H
W PU
T3T3
1 S
ra TO
T3 
JH x^r**
2 ^
^2
'SBU
"Sb oi
d,W5
<;
&3
TO m
Sri
> r*
<
"-. E 1
Cj ^*
Tl hr1
QJ ^Lj
lz
li
?fe
rq
i t
ON
^01
3
TO
H
en
H^
^oo
is



I-
<3 00
2Hn

v,
1?
Cn rH
^



^rq
Is?
en o
^ 0
< ON
> ^
(^^


V,
T3 ON
5^
en tx
-< ON
>rH
V,
T3 so
2^?
en rj<
, tx
~

V
"2 
3^
^S
< ON
>rH






Ol
4-t
3?
TO
<

CO 00 OO rH
co rq <* co
Tt o rq o< o rq
rH rH



tx oo
oo NO in
NO o tx in tx tx
rH rq <*! rH
rH


oo NO
rH tX ON 00
^ O IX 00 O tx
rH ON rH ON
rH rH





o rq
NO in co -rjj
rq rn rq tx rn rq
rH OO rH OO





O ON
NO 00 ON O
rH CJ tx tx rH tx
rH in rH in

rq rq
CO ON tx ON
tx! T-H ON CO rq ON
i i rq rq rq




^S ^3
S^rq rq^rq






p
g e
u t
H oi
00 tx"
tC CO^
co" rq"
rq" rn"

oo
oo rq
o rq co
IX rH




bO bObO


sO
SK
rH rH





in
oo <* rq
in rH 00
rH




in
ON rH tx
in rn in
r t

0
O 00 ON
ON rH rq
rH




in
} rq tx
5 rn rq






P
P
U
$
^o,
oq^tx^
tCso"
"^f?
co rq"
rq"rn"
rH

CO
oq rH in
ON in ^ NO ON ON
ON rn rq in
tx



tx
^H CO NO O NO NO
rH rq rH NO NO rH
 H tX


tx 00
CO tx tX O rH tx
tx rH ON tx tx ON
rq rH rq^ T-H
rH




rq
T+I co rq o o rq
^ CO OO CO CO OO
rq rH T i
rH



tx
tx NO tx O "* tx
NO rq in co rH in
rq rq I-H
rH
co
rH sO ON CO NO ON
ON co rq o oo rq
rq Tj^rH
rH



3$& |^
^H





P
P
U
OH
X
oo
tx"
1 
rq" U
rn" O



S
o
ffi J
2 ^
s 1
"S *
S cj g
"TO O 
CO ^*. *^^
 9 ^
OH J C
% & <"
O 4-1 (H
H *JH QJ
So ^ |
S TJ  ,
0 P
^ ^2
 Ue
'B u "^
*^ ^H QJ
n w tJ
^2 g^
a *.g
OJ x-x ..
e o o
3 z c
OJ '^ en
O-, 73 C
en S
1/2 OJ O
? '-S
i-C QJ TO
H +^ i-
01 *i
csi ^ S
t N j j ^
Mi
^i|^
*- s J^ ?,
S C TO g
0 * ED
H8
TO 0 C ^
U^ 8 Sj
0) 4^ >,
Lj  V
IH en ^M
QJ O C ^
SH M-i n. i
l^ss
-0 'C c-
-O TO 01 .5
^l^s
2 s^-3
<" +- ^ 0 x-x
^0^ g, M)
+- CH en C 5Jh
0 ,- , S a.
e |^ 8a
g ^ c c&
Concentrations f
(age information
Statistics based o
Statistics based o
Arithmetic avera
Standard error.
RJ i u TJ 











P
P
U
o
43
X
ON
00^
tC
co"
Csf
t-H
U
X
+-^
M-H
O
5/3
F-H
QJ

Ol
73
Ol
3
^
 en
T3 TO
Ol Ol
N C
^1
C ^
TO "-
 T3
SnS
*rT? ^
VJ W
o 71
^
o 
S ^>
IH 'K
0 o
en CX
P
6 
'D rq
Ol 00
&,>H
03 IX,
T3 o>
TO 5J
3 X
T3 ^
"H H
> 0
1^
.a 01
^ "C
o S
>H !X
Ol Ol
? ^
g TO
I13
2P
*- 60
139

-------
      Average concentrations, with  standard errors  for additional analytes, are  provided in
Table 9-13.   The VA/EPA  results  (Bauer et  al.  1990)  are compared with  the average
concentrations for NHATS composite samples containing specimens from donors  in the 15-44
year age group.  As indicated, many of the compounds were detected in fewer than 50% of
the composites.   In those cases,  the  reported average concentrations may be  significantly
affected  by  the estimated LODs.   For  compounds  detected in more than  50%  of the
composites, the average concentrations from the NHATS studies are considerably lower than
those obtained in the VA/EPA study. The only exception is that the measured concentrations
of 2,3,4,7,8-PeCDD are much higher in the FY82 NHATS.

      Two explanations for the  differences in the levels found in these studies are possible.
First, the apparent decline in PCDD  and PCDF  concentrations reflects a decline in PCDD and
PCDF residues in the  general  environment over the same time  frame.  This  is a logical
possibility resulting from regulations promulgated and enforced since 1970 and environmental
awareness that has focused attention on releases of toxic chemicals via industrial effluents and
handling of hazardous wastes.

      The second possibility for the  differences observed between the VA/EPA and the FY87
NHATS results may be attributed to  storage stability.  Since PCDDs  and PCDFs are very
persistent, stability is affected by the  integrity of the tissue rather than the chemicals.  Some of
the specimens in the VA/EPA study had been stored  since  1971 before being analyzed in 1986
(16 years later).  The NHATS composite  samples from FY82 and FY87 were analyzed within
two to three years of the specimen collections.

      Although further evidence  is  required, the data in Table 9-12  suggest a correlation
between time in storage and measured concentrations.  However, the effect on storage time is
completely confounded with the effect of collection year. Thus, it is not possible to determine
which factor is causing the observed effect on concentration. To address the issue of storage
time versus  collection  year,  further studies will be  necessary to either directly study the
storage  effect or identify  a surrogate measure of tissue  stability.  Storage stability can be
studied  through development of quality control pools that can be stored with  the NHATS
archives and pulled for analysis with each analysis program.
                                         140

-------
         Table 9-13.  Arithmetic Averages (pg/g), Standard Errors, and Sample Sizes for Selected Analytes
    Obtained from the VA/EPA, FY82 NHATS (15-44 Age Group), and FY87 NHATS (15-44 Age Group) Studies
Analyte
2,3,7,8-TCDF

1,2,3,7,8-PeCDF


2,3,4,7,8-PeCDF


1,2,3,4,7,8-HxCDF


1,2,3,6,7,8-HxCDF


2,3,4,6,7,8-HxCDF


1,2,3,7,8,9-HxCDF


1,2,3,7,8,9-HxCDD


1,2,3,4,6,7,8-HpCDF


1,2,3,4,7,8,9-HpCDF



OCDF


VA Study3 NHATS"
1971-82 FY82
2.1C
0.177d
197*
0.6
0.062
197
23.0 39.4
1.071 4.62
197 17
21.4
1.005
197
10.9
0.581
197
3.3
0.192
197
0.33
0.021
197
18.0
0.837
197
36.5 21.1
1.822 3.39
197 16

1.4
0.082
197
3.1
0.290
197
NHATSb
FY87
1.34
0.098
9
0.244*'
0.042
16
8.71
0.543
14
7.13
0.817
5
4.63
0.376
15
0.281*
0.001
2
0.370*
0.043
14
10.7
0.298
15
15.9
1.88
10

0.726*
0.048
15
2.89*
2.07
6
a Includes all study specimens; concentrations not detected (ND) were replaced with LOD/2.
b Statistics based on composites in 15-44 yr. age group; concentrations not detected (ND) were replaced with
 LOD/2.
c Arithmetic average (pg/g).
d Standard error.
e Number of individual specimens or composites analyzed.
' * Indicates detection in fewer than 50% of the FY87 composites.
                                                  141

-------
10.0 BIBLIOGRAPHY
Andrews JS Jr,  Garrett WA, Patterson GD Jr,  Needham LL, Anderson JE, Roberts DW,
       Bagby JR, Paletta FX Jr.  1989.  2378-TCDD levels in adipose tissue of persons with no
       known exposure and in exposed persons. Chemosphere 18:499-506.

Andrews JS Jr.,  Garrett WA Jr., Patterson DGD Jr., Needham LL, Roberts  DW, Bagby  JR,
       Anderson JE, Hoffman RE,  Schramm W.   1989.   2,3,7,8-Tetrachlorodibenzo-p-dioxin
       levels in  adipose tissue  of persons with no known exposure and in exposed persons.
       Chemosphere 18:499-506.

Bauer  KM,  Stanley JS, Remmers J,  Breen JJ, Schwemberger J, Schultz B, Kang HK.   1991.
       Pattern recognition analysis of VA/EPA PCDD and PCDF data.  Chemosphere 23:971-980.

Beck H, Drob A, Kleenan WJ, Mathar W, 1990.  PCDD  and PCDF concentrations in different
       organs from infants. Chemosphere 20:903-910.

Beck H, Eckart K,  Mathar W, Wittkowski  I. 1987.  Levels of PCDDs and PCDFs in adipose
       tissue of occupationally exposed workers. Chemosphere 18:507-576.

Bradley JC et al., 1990.  Interlaboratory Testing Study  on 2,3,7,8-Substitutes  Polychlorinated
       Dibenzo-p-dioxin and  Polychlorinated  Dibenzofurion  Isomers  Standard   Solution
       Chemosphere 20:487-493.

Buckland SJ, Hannah DJ, Taucher JA, Bates M.  1990.  Levels of polychlorinated dibenzo-p-
       dioxins and dibenzofurans in the breast milk of New Zealand mothers.  Short paper
       from the 10th International  Meeting, Dioxin 90, Bayreuth FRG, Vol. 1 (Ecoinforma
       Press, Bayreuth FRG), pp. 225-226.

Centers for Disease Control Veterans Health Studies Group. 1988b. Serum 2,3,7,8-tetrachloro-
       dibenzo-p-dioxin levels in U.S. Army Vietnam-era veterans. JAMA 260:1249-1254.

Centers for Disease Control (CDC).   1987.  Serum dioxin in Vietnam-era veterans-preliminary
       report. MMWR 36:470-475.

Centers for Disease Control.  1988a.   Serum levels of 2,3,7,8-TCDD in Air Force health study
       participants-preliminary report. MMWR 37:309-311.

Centers for Disease Control.  1988c.  Preliminary report: 2,3,7,8-TCDD exposure to humans
       Seveso, Italy. MMWR 37:733-736.

Cramer PH, Stanley JS.  1988.  Analysis of Adipose Tissue for Dioxins and Furans:  Adipose
       Tissue Sample  Compositing.  Interim Report  No. 1.  Office of Toxic Substances, U.S.
       Environmental Protection Agency. Contract No. 68-02-4252, WA 27.
                                         143

-------
Cramer PH, Ayling RE, Stanley JS.  March 1989b.  Midwest Research Institute.  Determination
       of PCDD and PCDF in human adipose tissue. Data Report, Batches 3, 4, and 5.  Office
       of Toxic Substances, U.S. Environmental Protection Agency.  Contract No. 68-02-4252,
       WA27.

Cramer PH, Ayling RE, Stanley JS.  March 1989a.  Midwest Research Institute.  Determination
       of PCDD and PCDF in human adipose tissue. Data Report, Batches 1 and 2.  Office of
       Toxic  Substances,  U.S.  Environmental  Protection Agency.   Contract No. 68-02-4252,
       WA27.

Dewailly E, Tremblay-Rousseau H, Carrier G, Grouix  S, Gingras S, Boggess K, Stanley J,
       Weber JP.  1991.   PCDDs, PCDFs and PCBs in human milk of women exposed to a
       PCB fire and of women from the  general population of the province of Quebec -
       Canada. Chemosphere 23:1831-1835.

Dewailly E, Weber JP, Gingras S, Laliberte C.  1991. "Coplanar PCBs in Human  Milk in the
       Province of  Quebec Canada:  Are  They More Toxic Than Dioxin for  Breast  Fed
       Infants?" Bull. Environ. Contamin. Toxicol. 47(4), 491-8.

Draper NR, Smith H. 1981.  Applied regression analysis, Second Edition. New York: John Wiley
       and Sons.

Faccheti S, Fornari A, Montagna M.  1980. Distribution of 2,3,7,8-TCDD in the tissue of a
       person exposed to  the toxic cloud at Seveso. Adv. Mass Spectrom. 8B, 1405-1414.

Fingerhut  MA, Sweeney MH, Patterson DG  Jr, Piacitelli LA,  Morris JA,  Marlow DA,
       Hornung RW, Cameron LW, ConnaUy LB, Needham LL, Halperin WE.  1989.  Levels
       of 2,3,7,8-TCDD in the serum of U.S. chemical workers exposed to dioxin contaminated
       products: interim results.  Chemosphere 19:835-40.

Fiierst P, Meemken HA, Goebel W.  1986.  Determination of polychlorimeter dibenzo-dioxins
       and dibenzofurans in human milk.  Chemosphere 15:1977-1980.

Graham M, Hileman FD, Orth RG, Wendling JM, Wilson JD.  1986.  Chlorocarbons in adipose
       tissue  from a Missouri population. Chemosphere 15:1595-1600.

Gross ML, Lay JO, Lyon PA. 1984.  2,3,7,8-Tetrachlorodibenzo-p-dioxin levels in adipose tissue
       of Vietnam veterans. Environ Res 33:261-268.

Heath RG.  1988. Battelle Columbus Division. Review and evaluation of the quality assurance
       program for the study of dioxin and furan levels in adipose tissue of U.S. adult males.
       Final  Report.  Washington, DC:   Office of  Toxic  Substances,  U.S.  Environmental
       Protection Agency.  Contract No. 68-02-4294.

Heil GA, Fitzgerald BJ.   1987 (November 2).  Quality assurance plan for Rutland resource
       recovery facility site specific environmental assessment, Rutland, VT.  Vermont Air
       Pollution Control Division.
                                         144

-------
 Huteau B, Gonnord MF, GiUe P, Fraisse D, Bard D.  1990.  PCDD and PCDF levels in human
       adipose tissue in France. Short paper from the 10th International Meeting, Dioxin 90,
       Bayreuth FRG, Vol. 1 (Ecoinforma Press, Bayreuth FRG), pp. 243-246.

 Huteau B, Gonnord MF, Gille P, Fraisse D, and the  Interim Network for the Development of
       Environmental Epidemiology in Vietnam.  1990.  New data on PCDDs and PCDFs
       levels in South Vietnamese adipose tissues.  Short paper from  the 10th International
       Meeting,  Dioxin 90, Bayreuth FRG, Vol.  1  (Ecoinforma  Press,  Bayreuth FRG),
       pp. 239-242.

 Johnson RA, Wichern DW.  1982.   Applied  Multivariate Statistical Analysis.  Prentice Hall,
       Englewood Cliff, NJ.

 Kahn PC,  Gochfeld  M, Nygren M, Hansson  M,  Rappe C,  Velez H, Ghent-Guenther  T,
       Wilson WP.  1988.  Dioxins  and dibenzofurans in blood and adipose tissue of Agent
       Orange exposed Vietnam veterans and matched controls. JAMA  259:1661-1667.

 Kahn PC, Lewis  W, Ghent-Guenther  T, Hansson M, Rappe C.   1990.   Dibenzodioxin and
       dibenzofuran  levels in three groups of  Vietnam veterans who did not take part  in
       Agent Orange spraying.  Short paper from the 10th International Meeting, Dioxin 90,
       Bayreuth FRG, Vol. 1 (Ecoinforma Press, Bayreuth FRG), pp. 247-250.

 Kang HK, Watanabe  KK,  Breen JJ,  Remitters JR, Conomos  M, Stanley J, Flicker M.  1990.
       Dioxins and  dibenzofurans in adipose tissue of U.S. Vietnam veterans and controls.
      J Am Public Health Assoc 81(3):344-349.

 Knick PD, Hosenfield JM,  Turman KW.  1986.  Midwest Research Institute.  National Human
      Adipose Tissue  Survey operations manual.  Final Report.  Washington, DC:   Office  of
      Pesticides  and Toxic  Substances, US Environmental  Protection Agency.   Contract
      Numbers 68-02-3938 and 68-02-4252.

 Leczynski B, Collins T, Sasso N, and Clark J. 1988. Battelle Columbus Division. FY87 NHATS
      composite  design.  Final Report.  Washington, DC:  Office  of Pesticides  and  Toxic
      Substances, U.S. Environmental Protection Agency. Contract No. 68-02-4294.

 Mack GA, Leczynski  B, Chu A, and Mohadjer L.  1984.  Battelle  Columbus Division and
      Westat, Inc.   Survey design for  the national adipose tissue survey.  Draft Final Report.
      Washington,  DC:   Office of  Pesticides and Toxic Substances, U.S. Environmental
      Protection Agency.  Contract No. 68-01-6721.

Masuda Y,  Kuroki H, Haraguchi K, Nagayama J.  1986.  PCDFs and  related compounds  in
      humans from Yusho and Yucheng incidents.  Chemosphere 15:1621-1628.

Masuda Y,  Kuroki H, Haraguchi K, Nagayama J.   1985.   PCB and  PCDF congeners in the
      blood and tissue of Yusho and Yucheng patients. Environ. Health Perspectives 59:53-58.

Muto H, Shinada M,  Abe  T, Takizawa Y.   1991.   The tissue  distribution of 2,3,7,8-chlorine
      substituted dibenzo-p-dioxins in humans who  died of cancer. Life Sciences 48:1645-1657.
                                         145

-------
Neal  FA,  Olson JR,  Gasizwicz  TA,  Geiger LE.    1982.   The toxicokinetics  of 2,3,7,8-
       tetrachlorodibenzo-p-dioxin in mammalian systems.  Drug Metab Rev 13(3):355-385.

Needham LL, Allen KM, Patterson DG Jr, Andrews JS Jr.  1990. Human levels  of 2,3,7,8-
       tetrachlorodibenzo-p-dioxin  in  Missouri.   Short paper  from the  10th  International
       Meeting, Dioxin  90,  Bayreuth  FRG,  Vol.  1  (Ecoinforma  Press,  Bayreuth  FRG),
       pp. 259-262.

Needham LL, Patterson DG Jr, Alley CC, Isaacs S, Green VE, Andrews J, Sampson EJ.  1987.
       Polychlorinated dibenzo-p-dioxin and dibenzofuran levels  in  persons with  high and
       normal levels of 2,3,7,8-Tetrachlorodibenzo-p-dioxin.  Chemosphere 16:2027-2031.

Nguyen  TNP,  Hung BS,  Vu DQ,  Schecter A.   1989.  Dioxin levels  in  adipose  tissues  of
       hospitalized  women living in the south of Vietnam in 1984-85 with a brief  review of
       their clinical  histories.  Chemosphere 19:933-936.

Noren K, Lunden A.  1990.  Trend studies of poly chlorinated biphenyls, dibenzo-p-dioxins and
       dibenzofurans in human milk.  Short paper  from the 10th International  Meeting,
       Dioxin 90, Bayreuth FRG, Vol. 1 (Ecoinforma Press, Bayreuth FRG), pp. 263-266.

Nygren M, Rappe C, Lindstrom G.  1986.  Identification of 2,3,7,8-substituted polychlorinated
       dioxins  and   dibenzofurans  in  environmental  and human  samples.  In  Rappe  C,
       Choudhary  G, Keith  L (eds):   Chlorinated dioxins and dibenzofurans in perspective,
       Chelsea, MI:  Lewis Publishers, pp. 17-34.

Ogaki J, Takayama K, Miyata H, Kashimoto T.  1987. Levels of PCDDs  and PCDFs in human
       tissues and various foodstuffs in Japan. Chemosphere 16:2047-2056.

Olafsson PG, Bryan  AM, Stone W.  1988.  Polychlorinated biphenyl and  polychlorinated
       dibenzofurans in the tissues of patients with Yusho or YuChen. Total toxicity.  Bull.
       Environ. Contam. Toxicol. 41:63-70

Ono M, Wakimoto  T, Tatsukawa R, Masuda Y.  1986.  Polychlorinated dibenzo-pdioxins and
       dibenzofuranas in human adipose tissues of Japan.  Chemosphere 15:1629-1634.

Orban J, Lordo R, Menton R, and  Unger A.  1989. Battelle Columbus Division.  Statistical
       analysis plan for chlorinated/brominated dioxins  and  furans from FY87  NHATS.
       Task 2-34 Interim  Report.   Washington, DC:    Office  of  Toxic  Substances,  U.S.
       Environmental Protection Agency.  Contract No. 68-02-4294.

Panebianco  DL.  1986a. Battelle Columbus  Division.   Sample design for the fiscal  year 1987
       NHATS.  Draft  Final  Report.   Washington,  DC:   Office  of Pesticides  and  Toxic
       Substances, U.S. Environmental Protection Agency.  Contract No. 68-02-4243.

Panebianco DL. 1986b.  Battelle Columbus Division.  A review of hospital nonresponse and its
       effect on standard errors of sample estimates in NHATS.  Draft Final Report. Washington,
       DC:  Office of Pesticides and Toxic Substances,  U.S. Environmental Protection Agency.
       Contract No. 68-02-4243.
                                          146

-------
Patterson DG Jr, Needham LL, Pirkle JL, Roberts DW, Bagby JR, Garrett WA, Andrews JS Jr,
       Falk H,  Bernert JT,  Sampson EJ, Houk VN.  1988.   Correlation between serum and
       adipose  tissue levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin in 50 persons from Missouri.
       Arch Environ Toxicology 17(2): 139-143.

Patterson DG Jr, Holler JS, Smith SJ, Liddle JA, Sampson EJ, Needham LL.  1986c.  Human
       adipose  data for 2,3,7,8-TCDD in certain U.S. samples.  Chemosphere 15:2055-2060.

Patterson DG Jr, Holler JS, Lapeza DR Jr, Alexander LR, Groce DF, O'Connor RC, Smith SJ,
       Liddle JA, Needham LL.  1986b.  High-resolution gas chromatography/high-resolution
       mass  spectrometric analysis of human adipose tissue for 2,3,7,8-tetrachlorodibenzo-p-
       dioxin. Anal Chem 58:705-713.

Patterson DG  Jr,  Hoffman RE, Needham LL, Roberts DW,  Bagby JR,  Pirkle JL,  Falk H,
       Sampson EJ, Houk VN.  1986a.  Levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin in adipose
       tissue of exposed and control persons in Missouri-an interim report.  JAMA 256:2683-
       2686.

Patterson DG,  Fingerhut MA, Roberts  DW, Needham LL, Haring-Sweeney M, Marlow DA,
       Andrews JS,  Halperin  WE.    Levels  of polychlorinated  dibenzo-p-dioxins  and
       dibenzofurans in workers exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin.  Am J Ind Med
       16:135-146.

Phiet HP, Anh TK, Vu  DQ, Schecter A.   1989.   Preliminary observations on. the clinical
       histories,  polychlorinated  dibenzodioxin   and   dibenzofuran  tissue  levels   and
       2,3,7,8-TCDD toxic equivalents of potentially dioxin exposed patients living in the south
       of Vietnam.  Chemosphere 19:937-940.

Phuong NTN, Hing BS, Vu DQ, Schecter A.  1989. Dioxin levels in  adipose tissue of women
       living in the south of Vietnam in 1984-85 with a brief review of their clinical histories.
       Chemosphere 19:933-36.

Pirkle JL, Wolfe WH, Patterson DG  Jr, Needham LL, Michalek JE,  Miner JC,  Peterson MR.
       1989.  Estimates of  the half-life  of  2,3,7,8-TCDD in ranch hand veterans.  / Toxicol
       Environ Health 27, 165-171.

Rappe C, Nygren M, Lindstrom G, Hansson M. 1986. Dioxins and dibenzofurans in biological
       samples  of European origin. Chemosphere 15:1635-1639.

Robinson PE, Mack GA, Remmers J, Levy R, and Mohadjer L.  1990 Environmental Research
       53, 175-192.

Rogers J.  1989. FY87 NHATS chlorinated dioxins  and furans outlier analysis, Final Report to
       EPA's Office of Toxic Substances, Contract 68-02^1293, September 29.

Ryan JJ.  1986. Variation of dioxins and furans in human tissues. Chemosphere 15:1585-1593.

Ryan JJ,  Lizotte R, Lau  P-Y, et al.   1985.   Chlorinated dibenzo-p-dioxins and chlorinated
       dibenzofurans in Canadian human adipose tissue. Chemosphere 14:697-706.


                                          147

-------
Ryan JJ, Schecter A, Lizotte R, Sun W-f, Miller L.   1985.  Tissue distribution of dioxins and
       furans in humans from the general population.  Chemosphere 14:929-932.

Ryan JJ, Schecter A, Sun W-F, Lizotte R.  1986.  Distribution of chlorinated dibenzo-p-dioxins
       and  chlorinated dibenzofurans  in human tissues  from  the  general population.
       Chapter 1 in Chlorinated dioxins and dibenzofurans in perspective, Rappe C, Choudhary G,
       Keith L (eds), Chelsea MI: Lewis Publishers.

Ryan JJ, Schecter A, Masuda Y, Kikuchi  M.  1987.  Comparison of PCDDs and PCDFs in  the
       tissue of Yusho patients  wtih those from the general population in Japan and China.
       Chemosphere 16:2017-1025.

Schecter A, Ryan JJ, Gitiitz G.  1986.  Chlorinated  dioxin and dibenzofuran levels in human
       adipose  tissues  from exposed and control populations.  In Rappe C, Choudhary G,
       Keith LH (eds):  Chlorinated dioxins and dibenzofurans in perspective, Chelsea, MI:  Lewis
       Publishers, pp. 51-65.

Schecter A,  Ryan JJ.  1988.  Polychlorinated dibenzo-p-dioxin and  dibenzofuran levels in
       human  adipose tissues  from workers  32 years  after occupational  exposure  to
       2,3,7,8-TCDD. Chemosphere 17:915-920.

Schecter A, long HY,  Monson SJ, Gross ML, Constable J.   1989.   Adipose tissue levels of
       2,3,7,8-TCDD in Vietnamese adults living in Vietnam 1984-87.  Chemosphere 18:1057-
       1062.

Schecter A,  Constable  JD, Bangert JV, Tong H, Arghestani S, Monson  S, Gross M.   1989.
       Elevated body burdens of 2,3,7,8-tetrachlorodibenzodioxin in adipose tissue of United
       States Vietnam veterans.  Chemosphere 18:431-438.

Schecter A, Ryan JJ.  1988.  PCDD and PCDF levels in human adipose tissues from workers
       32 years after occupational exposure to 2,3,7,8-TCDD in 1953 in Germany. Chemosphere
       17:915-920.

Schecter A, Fiirst P, Fiirst F, Groekel W.  1990.  Levels  of dioxins, dibenzofurans, and other
       chlorinated xenobiotics in human  milk from the Soviet Union. Chemosphere 20:927-934.

Schecter A, Ryan JJ, Gasiewicz  TA.  1990.  Decrease in human tissue levels of dioxins and
       dibenzofurans over nine years after exposure in  one  male patient. Short paper from
       the 10th International Meeting, Dioxin 90,  Bayreuth FRG, Vol. 1 (Ecoinforma Press,
       Bayreuth FRG),  pp. 275-278.

Schecter A,  Vu D,  Tong  HY,  Monson SJ,  Gross ML, Constable  JD.   1989.  Levels of
       2,3,7,8-TCDD and 2,3,7,8-TCDF in human adipose tissue from hospitalized  persons in
       the north and south of Vietnam 1984-85.  Chemosphere 19:1001-1004.
                                          148

-------
Schecter A, Mes J,  Davies D.   1989.   Polychlorinated biphenyl  (PCB), DDT,  DDE and
       hexachlorobenzene (HCB) and PCDD/F isomer levels in various  organs  in  autopsy
       tissue from North American patients. Chemosphere 18:811-818.

Schecter A, Constable  J, long H,  Arghestani S, Gross  M.   1987.   The  use of  tissue
       measurements of 2,3,7,8-TCDD to  characterize elevated dioxin body burden of dioxin
       from Agent  Orange  in U.S. Vietnam veterans up to  20 years after exposure.
       Chemosphere 18:531-537.

Shecter  AJ, Ryan  JJ,  Constable  JD, et  al.   1990.    Partitioning of  2,3,7,8-chlorinated
       dibenzo-p-dioxins and dibenzofurans between adipose tissue  and  plasma lipid of  20
       Massachusetts Vietnam veterans. Chemosphere 20:951-958.

Schecter A, Ryan JJ, Lizotte R, Sun W-F, Miller L, Giuitz G, Bogdasarian M. 1985. Chlorinated
       dibenzodioxins and dibenzofurans in human adipose tissue from exposed and control
       New York State patients.  Chemosphere 14:933-937.

Schecter AJ, Ryan JJ, Gross M, Weerasinghe NCA,  Constable JD.  1986.  Chlorinated dioxins
       and dibenzofurans in human  tissues from Vietnam, 1983-84.  Chapter 3 in Chlorinated
       dioxins  and dibenzofurans in  perspective,  Rappe  C,  Choudhary  G, Keith L  (eds),
       Chelsea MI: Lewis Publishers.

Schecter A, Ryan JJ,  Gitiitz G.  1986.  Chlorinated dioxin and dibenzofuran levels in human
       adipose tissues from exposed and control populations. Chapter 4 in Chlorinated dioxins
       and dibenzofurans in perspective, Rappe  C, Choudhary G, Keith L  (eds), Chelsea MI:
       Lewis Publishers.

Schecter A, Pa'pke O, Ball M.  1990.  Chlorinated dioxin and dibenzofuran levels  in human
       blood samples from various countries, including Vietnam, the Soviet Union, the United
       States and Germany.  Short  paper  from the  10th International Meeting, Dioxin 90,
       Bayreuth FRG, Vol. 1 (Ecoinforma Press, Bayreuth FRG), pp. 271-274.

Schecter A, Furst P, Fiirst C, Groebel W.  1990.  Human milk dioxin and dibenzofuran levels
       and levels of  other chlorinated chemicals form various countries,  including Vietnam
       and Cambodia,  as compared to  the  Soviet  Union.   Short paper from the  10th
       International Meeting, Dioxin 90, Bayreuth FRG, Vol. 1 (Ecoinforma Press, Bayreuth
       FRG), pp. 267-270.

Schecter AJ, Ryan JJ, Constable  JD.  1986.  Chlorinated dibenzo-p-dioxin and dibenzofuran
       levels in human adipose tissue and milk samples from the north and south of Vietnam.
       Chemosphere 15:1613-1620.

Stanley JS, Cramer PH, Ayling RE, Thornburg KR, Remmers JC, Breen JJ, Schwemberger J.
       1990. Determination of the prevalence of poly chlorinated diphenyl ethers (PCDPEs)  in
       human adipose tissue samples. Chemosphere 20:981-985.
                                         149

-------
Stanley JS, Boggess K, Going JE, Mack GA, Remmers J, Breen J, Kutz FW, Carra ], Robinson P.
       1986.   Broad scan analysis  of human adipose tissue from  the  EPA FY82 NHATS
       Repository, pp. 161-179, Chapter 14 in Environmental Epidemiology, Kopfler FC, Craun
       GF (eds), Chelsea, MI:  Lewis Publishers.

Stanley JS, Bauer KM, Turman K, Boggess K, Cramer P.  1989.  Determination of body burdens
       for polychlorinated dibenzo-p-dioxins (PCDDs) and  polychlorinated  dibenzofurans
       (PCDFs) in California residents. NTIS Publication No. PB90-148289.

Stanley JS, Bauer  KM.  1989b. Chlorinated dibenzo-p-dioxin and dibenzofuran residue levels
       in food. NTIS Publication No. PB90-146523.

Stanley JS, Boggess K, Onstot J, Sack T, Remmers J, Breen J, Kutz F, Robinson P, and Mack G.
       1986.   PCDDs and PCDFs in human adipose  tissues from  the  EPA FY82 NHATS
       Repository. Chemosphere 15:1605-1612.

Stanley JS, Ayling RE, Cramer PH, Thornburg KR, Remmers JC, Breen JJ, Schwemberger J.
       1990.  Polychlorinated dibenzo-p-dioxins  and dibenzofurans in human adipose tissue
       samples from  the  continental United  States  collected  from 1972 through 1987.
       Chemosphere 20:895-901.

State of Connecticut.  1986.   An act establishing a program for monitoring dioxin emissions
       from resource recovery facilities. Public Act 86-332.

Tong HY, Monson SJ, Gross ML, Powell LBM. 1989. Elevated levels of 2,3,7,8-TCDD in tissue
       of agricultural sprayer  of herbicides.  Chemosphere 18:469-476.

USEPA.  1986a.   Broad scan analysis of  the  FY82 National  Human Adipose Tissue Survey
       specimens.     Volume   IV.     Polychlorinated  dibenzo-p-dioxins   (PCDD)  and
       polychlorinated dibenzofurans (PCDF).  EPA Publication No. EPA-560/5-86-038.

USEPA.  1986b. Analysis for  polychlorinated  dibenzo-p-dioxins and dibenzofurans in human
       adipose tissue:  method evaluation  study.   EPA Publication No. EPA-560/5-85-022.

USEPA. 1987a. USEPA.  Interim procedures for estimating risks associated  with exposures to
       mixtures  of chlorinated  dibenzo-p-dioxins  and dibenzofurans  (CDDs  and  CDFs).
       Washington DC. EPA/625/3-87/012.

USEPA.  1987.  40 CFR Parts 707 and 766.  Polyhalogenated  dibenzo-p-dioxin/dibenzofurans:
       testing and reporting requirements; final rule. Federal Register 52:21412-21452.

USEPA.  1989.  USEPA. Interim  procedures for  estimating risks associated  with exposures to
       mixtures of chlorinated dibenzo-p-dioxins and dibenzofurans (CDDs and CDFs) and
       1989 update. Washington DC.  EPA/625/3-89/016.

USEPA.   1990a.   NHATS broad scan analysis:   population  estimates from fiscal year  1982
       specimens.  EPA Publication No. EPA 560/5-90-001.
                                          150

-------
USEPA.   1990b.   Pattern  recognition analysis of  VA/EPA PCDD and PCDF  data.   EPA
       Publication No. EPA-560/5-90-006.

USEPA.  1991.  Water  programs; guidelines establishing test procedures for the analysis of
       pollutants. 40 CFR Part 136. Federal Register 56(26):5090-5122.

VA/EPA.  1990.  Dioxins and dibenzofurans in adipose tissue of U.S.  Vietnam Veterans and
       Controls. EPA Publication No. 560/5-89-002.

Van den Berg M, Van der Wielen Frans WM, Olie K, Van Boxtel CJ.  1986.  The presence of
       PCDDs and PCDFs in human breast milk from the Netherlands.  Chemosphere 15:693-
       706.

Weerasinghe NCA, Schecter AJ, Pan JC, Lapp RL, Giblin DE, Meehan JL, Hardell L, Gross ML.
       1986.  Levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) in adipose  tissue of
       U.S. Vietnam veterans seeking medical assistance.  Chemosphere 15:1787-1794.
                                         151

-------
                          APPENDIX A
FY87 NHATS COMPOSITE DATA LISTED BY ANALYTE FOR EACH COMPOSITE
                              A-l

-------























a

UJ I/I
j y
*
3 D
I/I U.

uj o

M 9
VI
O I/I
a. z
O X
uo
M
1/1 a
s
^
P* Z
>- at
u. O
U.I
Ik ^
ou
a ae
zo
M U.
I/I
^
_J













































ti.
a
u
i-

00
J^"
.
CO (D
-r-

U. 1 O
Q CM
O ~ n
1- U
i n o
GO ^ *^

r u 
- ac
c}I
CN TQ U
(S -f
C I/I Ol
3 L.
OCA
O +>
U n
c
m

i
c
01
M
c
M




























































































L.
01
a o
 u
01
0
c 
o +*

M a
c
o
fll -f
* o
flj .
O L.
VI
01
a
___ ^
0 Ol
n HI
82



-,
0 O)
Ox.
^"l
ft
^

1.
01

(0 M-
^ '^
1C *-
Q (0

*^

C 0)
^
**
8S
0) 2
a.
+rf
C 01
0 
01
Q.
01 S"
O) O
"*
3
in -
01 01
UK

C
* 
i"
a
^^

a
n
1/1
X
O (.
* 0)
(Q J3
i. B
o 5
r
u
(B
CD

r-OOaJUJf-^tor-oi^'CitD
(OCM(0(0(O(Ococ--(0(0'~oiin
r-lOiflCOr-cor-r^f-r-oor-r-'
^
r- (DO
^ O ~

O CM
t^oi^oocooin^oicMiflcofo
^O^^^PJ^^^^PJOO,



O) 0) CO Ifl ^ CO (O O)
* O co in  IB ID CM c-

do ddo o odoo





o o o

ooKoaaetaxaaaa



or-ooiaiocootuoocncM
otincMMincocooioO'-oi
^ ^


oooooomooo^mo
inOOOOO'-QOOVCMQ
CM ino(oooov(oo
" *" "

W W ^ ^ ^ CO ^ CO CO W CO W W


U (J CJ U U U U


W O) O ^ ^ O CO N CO O 00 CM ^
'-'-CM'-'-CM'-COCMCM'-COeM
r~incooocMcn  TfO(D(OOfO
^ W N ^" ^ (O (0 04 Ot (** U1 f*** 00
CS^OCO^O^*"CO^CS^CS
8888888888888
QOQOQQQQQQQQQ



^mr^oooo^cNCo^r^ooot
minifluimtDtotofocototDU)
C4C4O4MC4C4CNCNOIWMWOI






** CO P* CS O> ^
c? ^r o1) co in o
oo r* r- r* p* r^

'---



CM r~ ID
oo co t -
^
O O O "O





o in
- CM
o a a o a a:



in o ID in  in
to co CM (D i r-



o o o CM o co
8 S  5 8 2
"

CO CO CO ^ *" CS


^ ^ ^ ^ z" 2


r- in w r^. oo (0
-'-'-*-
P* O> (0 (D ^ *
00 00 O) ^ * (**
O ^ O ^ O c^
888888
g   S  S
<<<<<<


o CM t in us t-
r^ P* r^ r^ r^ r^
CM (M CM CM CM CM



CM CM CM CM CM CM


O
CM
ID
















4f
in
-



^


o







a



ID
O
Ol



O
O


CO


UJ


CM
CO
in
co

8
S
u


Ol
f^
CM
ID

CO



in
n
IO

CM
^

d





|H





ai
ID
CM












O

at



a
CO
01



O
o


CO


*


(D

CM
Ol
CM
8
CO
a
u


o
a
CM
(D

CO



CM
qi
in
















^
*
T-



CM

CO
o







a



CM
co



O
o


CM


/,


Ol

^
in
CO
8
co
a
.


^
co
CM
ID

CO



co
CM
ID

in
in

o




u.
|H




f-

co
0












in
CM
OS.



r-
0)



o
o
o


CM


z-


<
CM
^.
CO


CO
a
^


CM
co
CM
(0

CO


A-3

-------













_^
< 1-
i

_j a
a. u
CO < 00
at
iu 3 r-
1- U.
IH co (o
f- f*
CM 
L. 1 O
z a CM
o - < ii C
u. z re -H
IH TJ -M 0
U. OS C (/) 0)
O O 3 L
-i o c re
(IT n o ^
w ^ u u ^
z o F-
IH O +
t ae u re
l/l O -f
-j +
^
re

(U
+J
c
M

















X
L.
W >
a o
w O
(U
c.2
0 **
IH re
a

re -H
H O
a I
VI
u
et
^^
00,
c32


a ro
o \
-1 O)



L.
0)


fll OJ
u at

c
f

U O
Q) O
O
IH
0)
re
10
X
L
O C.
re n
o 3
re
r
u

re
oa


N
CO
(O










CO
in
CM


in
CM
CM

d






o



CO
CO
en


o
o


CO



2




in

on
O
CO
8
00
a
u



CO
oo
(O



CO




CO
CO
(O










in
en
-


p.
^

d






a



o
8


o
8
T~




u
z

"


CO

oa
CO
CM
8
00



00
CM
(O



CO




OOcoincMoo*-in
inmo-coa)coen
<0tt,<0tf,(D(0r-l0
fo o r-
O CO 
* - o



H4 I*H hH




"?-SSSoS
CM'-t-CO'.-'-i-CM


en (0 ^~ (O co
co in -  O
^ T* f^ ^ T^

d d d d d





o m o
 CM *-
ataooaaatoc



tr-omocMOfo
oooooooooooOoo


O'-OOOCMOOO
Or*c3oo'-ino
in o co * *


COCMCMCO-COCOCM


UJ O O IU
W*ZZtOZZ




r~*O(Dinr^too
^ ^" CM ^" ^ CM ^~
S2*5co8S
COCM'-CJCM'-*-'-
88888888
aaaaaaao
uuuuuuuu



oocooocncncnonci)
CMCMCMCMCMCMCMCM




cocococo^^v^




inininooocoo*ioco(oor-
oncMCMVcoooocointoocnoo
CD co co ^* (o r*" co in co CD co in in
CM co t- in
CM o o on
~ ~ - b



IH IH IH IH



 in
in coMtoincncor^cn cocn
O^"O~CM^-CO^-CMCMCM


r^ ^^(ocncoco^
a> v cncncMcn^nooen
CO CO ^--O'--^CM

o o oo'ooooo





in in in o
CM CM CM -
eraoeasoaaooaaao



r^ooocoocoinoocMencMO
coinr^'*o^iT"Oininco^o
r^or^ooocDCOOoooocovO


OOtOOOCMOOOOOO
OtOOO(OOOOOOO
r^wo* CM oo
^ *" ^ ^*

CM'-MCMCOCMCM'-COCMCO'-CO


w o ^j y u y^
tOZZZXifl^XOOZZ^Z




onooocoo'frcoinot-oof-en

woncM^cnco^incMOcMr^co
^^co'VCMtticMcor^inincM^
CO^OOCMCO^CMCOO^CO*
8888888888888
88gS88888gg8S



en o ^ w ^ co oo w O ^ co ^ in
enoOOOOOO^'"^'^'~
CMCOCOCOCOCOfOCOCOfOCOCOCO
CO(O(OCO(D(DCOCO(OCOCOCOCO



^r^nv^uiminininininin


A-4

-------


















	 	 K
< 1-
1- Z
< o
Q U
UI

0*
5 2
en 3
Q
UI 3
t- U.

en Q
21
otn
o z
K4
en x
1- O
< W
* Q
r- ui
oo H
> 4
u, z
M
ti Of
O O

t- oc
en o
M U.
~*
























































U.
a
o
i
oo
.
r-
.
n 03
 f^
CM 
U. I O
8M
- II
t- O
i n o
co  -
. +J
t- II (S
"He
(N M O
13 **
n C
< *
C-S8,
3 L.
O C 10
Q.O H
u m

p
c
ra


"
i
c
Q)
c
fr"l














































































































X
Jll
w
l/l >
a o

at
o
c 
o **
n fl
a
c
0
(fl **
+ o
a I
in
01
a:
- *,
0 O)

82



_
Q cn
O '**^


w

0)

4^ .^
ni ^
a n
j

C 01
01 H
0 -

tt) 2
a.
44
s
01
a.

8,1
< c.
u

U) C
3 0
VI -
C 0)
01 01
u 

i c
f" 3
0


0
01
1
t/l


X
L.
o s.
+* 01
n 5
C. E

o 2
(Q
J

j=
u

A
00


T
ID
in















cn
o




oo
M
^

b





a



^

f^




o
8



M





to


1-



(0
n
n
8
oo
Q
U





(0

CJ
(O
^





in




i
r~















^
en
M



cn
oo
n

O





o
a.


o

b
00



o
8



n





en


in



r-
O
i
r-
s
u





f-


a i- a. o < at
t- z a z ui
I/I UI Z UI O M >
UI O O t- " X Q
at z w z t- < 3
O t- ui < Z ui
< 03 z a ce
i- -i 3 x
< u ui < z < en
a o ae o ui or
CM u u. M a. M
ce ^
O O n ii ii ii ii
u.0
_i o u- M Q. a
a
UI
en
3

en
UI
O
O
u


A

-------




















t-
^
a

Ul  at
u. O
ii T
oo
s
1-
co

-1












































a
o
o
1-
1
00
^
-
n co
cs 
Q i O
O M
O - u
H U
1 CO 0
^
r- II <8
- a
CM (Q O
"1"

C CO O)
3 L.
o c ra
f^K
U 
a o
i O
01
o
c 
o **
11 (8
a
c
0
(fl -H
+- 0
(8 "
a L
VI
0)
or
om
rt H*
82




a o>
0 v.
-1 0)
a
^"
c.

.^
a <*
(8 
O (H
bJ *V
O

C OJ
o|
0) 2
a
4^
C 01
on
0)
a.
a
u 3
O)O
< -
(9
M C
3 O
VI -~
C O)


C
o o
uu
a
o

0)


X
L.
O (.
Si
si
o f.
(8

O
a


r-
co















O)
^




(O
8

d





a



o

^


0
in
(N



M



111
z



w


1^
CM
8
00
a
u
<



in
CM
CO






Ocoooooincnr~O(OOc)incnNoo(Moenoon<-
cjooc>fM-(oeoo(oco^tnin*enr^*-oo)int-
LncDCOf"oo(DCDiOr*>o>r*>t>*>[^t0(DCQr>-(Dt^(Qtfl[N*
*
oo n (o n in
< r- in 01 o

0-0 0 -




DC M M M MM




 cs  n ooooco 01  incni*-cf-t
 CM  <-  (Hoor^neninenOcno  c^ncnoovn
^ ^K ^



en - MioiD en  invin toen en*-
Sn d csoomMncnincnao owr-^oi
w  '-{^'''"wwwoi cscotoncM

oo o dodoooooo o o o o o




in in o in in
ooaaaoroaoaaaaaooraoaoaa



r-ocnenocaenuoocnNinocomvinioooNr-
cnS3S!coScooi8^*|5SSr25iS)SS
^


oooooinooO'tinoooocMooooooo
OOOOO  OgO^CNOOlOO'-OcoOOOO
inocaooo^ioooin ^o o
"""" ~ "


CM *~ ^ ^ CO ^ C*) C) CM CO CM O 09 C*) CO *" *"" CM CO CO CM! CM



uo o uj ujuj u o o o w u
jg^mi/i^t/i^pi^^jF^^^f^^^^^jg^pt/*^



28*28$2S3g2SIS-SS-?522*


innooNen-^oucoonr-enuiov^incjiv^
^O^OCS*~C4C*)^
000000000
ooooooooo
coooeococooooococooaoooooooooooocococooaaooo
aoaaaQaaaaaaQQQQaaooQQ
(JOOOOOOOOOOOUOOOOOOOOO
<<<<<<<<<<<<<<<<<<<<<<



(OCO(OCDCOCOCD(DCO(OCD(OCD(0(OCO(OCOCO(OCOCO




A-6

-------



















~
HZ"
aS
U \J
3~ 
to
Z i
at
U ^3 C^
1- U.
IH co (0
 CO If) 00
CO W CO C*5









N
^^
a
0




r- oo
*~ cn in
in co *
*"

en 
oo in CM in
cn *- - CM
O O O o







So a o


O  r~ o
o CM in in
O 00 CO 00

o o <- o
o r o
in o





- CO CM CM


U U




co r-  o
^* ^ w
00 ^ O ^
o oo ^ to
N CO N *-
8888
o Q a a



^ (0 00 O)
00 00 00 00
C* CM  (D en 

ooooooo o o o





o o o o

ooaaaoactaaaaQca


inocMOcot^ooocoocoinoo
oooooooor-(oroooiO(OQco

OOCMOOOOOVOOOCMOO
oo'-moootoootooo
OOV^f Ot~*O* *CM





CO^COCOCMCM'-CMeMCOWCM'-CO


O O LU LU O O
Z2(/9ZZCOZZZ7(/)CO7(/I




(Oinr^^rooaioooioovromo

iiiiiiiiiiiiii
00000000000000000000000000
aaaaaaaoaaoao
uuouuuuuuuuuuo



Soo en o
o o ^
CMCMCMCMCMCMCOCOCOCOCOCOCOCO


co^^^^f^^^^^tninmin



oo
in
r-

















co
O)
^



0
d







o


CM
in
00

o
o





CM


Z




r*
O4
i
8
00
a



^
T-
CO
(0

in



o
en
(

















.
*
(0


q.
J:
d







a


CO
00
00

o
8





CO


z-




oo
CM
in
I
00
a



CO
CO
(0

in



cn
in
r-

















CM
cn
,-


CO
CO
CM
d







o


CM


0
8





^


to




r-
r-
CM
CO
8
00
o
U



f
CO
(0

in



in
(O
(O


,_
CO

d





H4





N
^
t^










O

at


o
8

o
o





CO


2*




cn
CO
I
oo
a



in
o
(0

in


A-7

-------

















. 	 ,
< K
< O
O U
Ul
-I
Q.
Z CO
3 z
CO <
at
UJ 3
1- U.
^
co a

a <
SCO
z
(/) X
I- O

Is
a
t- UJ
co 1-
!
n
u. a
o o

a x
z o
n
t- a
co a
HH U,
























































a
a
u
a
~
r-
-
n (o
. js
CM 
1 O
CM
O "- ii
1-0
i co o
00 - 

r n M
at
CO B
- L C
CM IB O
ii C
ID -H
B -< z
CO UJ
> o
_J CO O!

 3
n -I O
K- < -1 h-
(S>Z< CO
iSo zo
H w < t
< -1 Z 0
h- < UJ CO U.
CO 0 X 0
11 O t- CO
Z Z ui UJ
M ui Z Z >- O
x M co z
o u H 3
u Q o a at
co Z ui Z UJ
am- t- z
0  
z i- z n a
o z at < i-
M M U y jiJ ^ Q
1- t- CO 9 H UJ Z
U < U UJ M CO
n a at a. a >-
DC i- a. u < a
i- z a z ui
CO UJ Z UJ O <-> >
UJ U O H >- X O
K Z M z 1- < U
O H UJ < Z uj
< u D i a a
H _l O X
<  UJ < Z < 10
a o ct o uj o
CM o u. M a. M
a \
O O n n n u n
U. O
_i o u. 11 a. at
a
UJ
CO
a
CO
UI
Q
O
u




A-8

-------




















1-
^
a
UJ  U.

UJ Q
h- Z
* 4
00
O (/I
a. z
Z IH
O X
u o
l-l
i/) a
i-

^S

CO l-c
> a
u. O

O U
a at
20
i-

M
-1











































u.
a
u
UJ
a.
00

CO

u. CM in
a -in
o * 
u i -
a. CM
i *- u
oo U
-co o
co" it ^J
- a
CM 'O
"- S
^ (Q O
DM
ii C

O -H 0)
C I/I 01
3 t-
6cm
o **
O (8

c
m
3
_
(Q
C
(U
4*
C
1-4




























































































X
0)
O 0

a
o
c 
0 -H
w flS
Of
c
0
IS -H
v U
a .-
a L
VI
0)
>-v
U 0)
C v.
82



" 
a o)
3m
2


i.
01

ns <*.
Q m
3

+*
2
gr
L. 
01 2
a.
+i
c u
01 ^*
0.
0)
a.
a
0) 3
no
< L.
(9
u> c
3 O
c 5)
U 01
u a

B C
gj
(/I
a


01
*

X
L.
O t-

ns J3
i. g
o 3
u
(TS
CO

o
0
CD







CM
V^
0
o









(O
S
o










o
8



o
in
CM



CM



z





fs.
^f
M
8
CO
a
^



^
in
CM
(0




O *~ M W
in o r- r~
r- * r- eo







CM CM CM
X^ "\ X^
at. a a a
o o o


00
CM
^
o



CM O> CO t~
o ? - 2
O O 0 O






2 i  



r~ o en en
* in CM CM



o o o o
O O O ^J
in



CM - - -



S S w ui



o> o * *

in co oo CM
N W ^ ^
co *~
8 8
00 CO 00 00
8888
< < < <



U) r** oo o)
in in in in
CM CM CM CM
(0 (O 10 (O




focor-^CT-enenMCMcneno
P5CMr-vcM(OP-iOoooo(OeM(O
oooooococncMQor**cnoooocooo







CMCMCMCMCM CMCMCMCM CMCM
x^ v^ -x^ *s^ ^^ -^ ""s^. "x^ X^ x^ "x^
aaaaoaoaao oo
ooooo oooo oo


in
ID
oo
o



co^ooco co co t^ in ts co
"ScomSSSwnSSSS
ooooo-oooooo-o






gggggggogocfgo



OcocniooocncMinoioin^1
uici(oa>OQ'*cn(OoocM(o



omooo^inoooocMO
-OOO*CMOOtDO^'O
^lOOOOtCDOOUl "TO



CO^'COCOCMCOCMOrOCOCO^"*"



O lu liJUJ O OOO



Cs"fOCM2ocM-S(N-2

ai<-O(0(OOnr-en(D(o
fcocACflr^mf^oofloooo)^^
f^1 O ^ O
OOOO
0000000000000
(JUOUOUOUUUOOO
<<<<<<<<<<<<<



S^cMCo^r^coenocM^infO
cototDODiOtocor^r^r^r*^
CMCMCMCMCMCMrMCMCMCMCMCMCM
(O(OIOIO(0(O(0(OIO(DIOCO(0




(0
q.
CO







CM
x^
a
0









in

o






a



in
00



00
oo



CM



*



(O

q.
r-
CM
8
CO
a
<



r-
r-
CM
CO

N


co
w
(0







CM
>^
Q
O









^
(0
o






i



(0
o
0)



o
o



m



UJ



CM

in
ao

8
co
a
<



en
r-
CM
co

CO


r~
(O
CO







CM
x^
0
o









f*
0
o






i



oo
CO
0)



0
o



CO



*



CO

CM
en
CM
8
ao
a
^



o
00
CM
CO
CO



1-
_
00







CM
X^
a
0









^
CO
(0
o










CM
00



O
O



CM



I/)



en

^
in

8
00
8
<



v-
co
CM
(O




0
r-
00







CM
X^
Q
O









en
in
O






Q



t~
T-
0)



O
o
o



CM



z1



CM

^.
CO

8
CO
a
u
<



CM
oo
CM
(O



A-9

-------











^
< K
< O u.
O U O
 u
UJ UJ
-1 O.
0. I
Z (/) 00
< z
(/) < r-
at
UJ 3 CO
1- u.
M U. CM U)
01 Q o -in
o z o - 
O. < UJ I *-
0. CM
i/) i - u
Z CO U
M -tO 0
< n C
u. Z  
ft) 3
a
^1
C ft)
ft) ^"
0 
01
a.
a
01 3
J'g
*4 *
(3
V) C
3 0
I/) 
C Ol
01 01
u a

c
'*" 3





f-

^
00
to
8
00
8
<


S
CM
CD





CO
r-
00








CM

O



^.
en

b





at



t*
in
00


*
r-
in




CM



X






o

CM
8
S
u



oo
00
CM
(0





r* to in r** o oo co ^ oo oo to in co CM ui ^ CM ^ CM
O in ^ o ^ r CM CM oo in ^ CM co CM co en r** o CO
oocoooencncncnenr-cnoocncncDoooocnoooo
to en co
o> '-in
CM >- b

CM CM CMCMCMCNCMCM CMCM CMCM CM
Q QMQQQQQQM o o <-> o o o
o o oooooo oo oo o


* ao co in
CM - CO CO >- 00
* in to r- co oo
*- o o o o o



^-co^ cnco^eninen csr^^ ^cMcno
oa  in W'-ocM'-en r-en* o n- oo co
CMCMCO 
-------



















^^
Zfe
< o
QU
Ul

Q.
5 "*

CO <
at
Ul D
I- U.
n
00 Q
O Z
o. 9

co
Z
M
oo x
H 0
< M
Z O
Z
a
r> u

Ui Z
u. a
o o

O X
z o
" a
CO O
11 U.



















































u.
Q
U
Ul
a.

00

^
-
CO

u. CM in
a -in
o *- 
Ul 1 -
a. CM
i i- u
00 O
-co o
H
CO II A
- tt
CM TJ
- i. C
- A O
DM
ii C
A +
D +* 01
C CO Ol
3 L
OCA
O *
U A
44

c
A
3


41
C
H










































































































c.
0)
CO >
a o
H U
at
a.
o
c 
o *
11 A
C
0
A -M
H U
A -
O L.
M
0)
a
got
2 x^
82
>^



^^
O Ol
o -.

n
>_^

C.
0)

A *
*< 
A 
a A



v
C 0)
01 +>
u -
t. 
at 2
a.
+J
C 01
u "A
t-X
01
a.
S3
o
i.
O
3 0
in 
C 01
01 01
ua

B C
ffy


Q
01
0
A


X
s.
o c.
+> 01
A ^
Q 3
43 Z
A
J


U

A
m

*- o

CM Ol
en o
*-







CM CM
^^ NS^
o a
3 3










in CM

(D 10

6 6







 




*- o
r-^ 6
 00



0 0
8 8

CM CO



CO CO

t- in



S S
CO *
80
0
r- r
8 U
^ ^




(0 r^
n ro
ID 10
^^ ^





in in






































P^
v^
a
0

CO


S3 S
2 /. 

< M Ul
z i/i a.

co -i o
-i M < in
o > * ^
c^j e
i/l Z < I/I
M < U Z O
H M < 
< -IX U
h- < Ul I/I U.
I/10X 0
n O t- CO
Z X Ul UJ
nutzx>- a
X M 03 Z
a u t- 3
ui a o a a
CO Z Ul Z Ul
D w t- )-Z
i- i . ni \/ nj ft r^
K H I/I < h- Ul Z
U < Ul Ul 11 CO
MCE a a. a >
at t- a. u < a
1- Z Q X Ul
co ui Z UJ O M >
ui o o H M x q
a z z t- 2 Q
O t- ui < X Ul
v
Q Q II II II II II
u. o
_i o u. 11 a. at
Q
Ul
3

CO
Ul
Q
O
U




A-ll

-------


















J-

0
Ul (/>
s!z
Z at
I/) U.

oi Q
f- Z
l-f <
(/)
O M
1 5
Sxx
0
01 Q
1-
< a
^5
[s. 2
00 N4
> a
u. O
U_ T*
u* ^
oo
sg
M U.
D>
3





































u.
Q
0
U
a.
00

.
L. CM in
a -in
o *- 
Ul 1 T-
a. M

00 O
-M O
v ii ra
- at
n TJ
 L C
w ra o
ii C

TJ + 01
C I/I O)
3 (.
o c ra
o Z
u ra
*
c
ra
c
01
c




















































































(-
0)
a o
M U
at
0
O v
M ra
Of

ra f
H 0

O t.
I/I
01
0 O)

82



a ai
o ^
a


C
01

(Q <*-
(0 i
Q (w
a

+*
c v
u ^
?i
01 y

+4
C 0)
ura
01
Q.

. 3
< 2
Ifl 
C 0>
SOI
at

c
i C
O O
So
(/I
o
01
ra
X
L.
O L
H 01
ra 3
c- E
O 3
fl

U
CO

OO'-noor~'-
Oinror-r-cocMf-^
cnr^^r^cooooocooo

M r* t
CM (D CM
 (M in
^



KM M M



(O (O n (0 en co M

cnm^rocor^CN'-cn
CN ^


ID * (0  *
(O cp o> in r- w
O O O O O O





o in in
aoaaaaSoa
art k^ rt kd A_ i^ ft A_
U, p^ (L ^^ (L M & &


oroo>o>oooan(0
8SSSSSSS3
^-


ooooooinoo
tnooooo^-op
M Ul O '-tO
iiiiiilii
oooocooooooooooooo
QQQQQQQQQ



S^ W CO
(0 (D (0
Ck4NCc5(O'-r*
 in >- CM oo t TO a) to en
^r-fflwvr-onencx r- r-
coencMcMCocococMn in- in t
ooooooooo o oo





o o
aooatoaoaaatoatao



OoenNinofflin^inioooesr^
S8^*i2ooSt25wSSSS
^


otinoooocMOoooooo
StNOOCDO  OooOOOO
^ us o o in *o- o



(MCOWCOCOWW'-'-CMCOCOCXW
UIUI O ^^^ ^ ^




85SSSSS?!S2S

"'"'^'SooScSS^'SSinco
liiilliliiiiil
oocnooooooooooooooaooooocooo
Qooooaaoaoooaa



^r^oowOc^^finujr^OTO^oj
(0(O(0(ot^p*p*t**ri"r^tsGOocjco
O4C4dWWWO4NWWNWdW
(0(0(DH)tOCDCD(D(DtO(0(0(O(0


^ M M M CN N CM (N CS CN 00 CO CO CO

A-12

-------



















^^
< K
t- Z
< O
_
w

5 *"

I/) 4
0!
U 2
K U.

I/I Q
25

8$
M
(/I JC
52
*Q
si
1-4
u. at
O 0
i- a
en o
HI U.
J
















































fc
U
o
UJ
a
00

r-
.
CO

u. CM in
8-"?
UJ 1 -
a CM
i *- u
*2.2
V II 19
w-? c
"52
ii C
(9 H
"O +> 0)
CM ft
O C 19
fft w
.2
Is
V
c
19
3
^
m
E
Q)
w
C
1-4
























































































X
t.
0)
(/) ^
a o

DC
0
c -
o H
1-4 (9
a
c
o
19 V
H U
(0 
a L
VI
01
a
^-*,
00.
A pjt
cSa
^s


- .
0 O)
o ^

Q
s_ '
01

19 H-
0|
C 0)
m ^
Si
a.
s*
iy ^
01
a.
0) "
O) O
< a
vi c
3 O
C O)
01 0)
u a:

c

2.
i/i
a
H

01
(9
in
X
o c.
(9 j5
(. E

19
U

19
m


en
CM
en
















O
^i



r-
 a CO
oo r^ co
CM *-


CM C~ CM
CM t~ (0
CO r- CM

odd



a o o
rt rt fl_
u, u. u>
 c^ O
CM in in
oo oo oo

O - O
d ^ d
in o


CO CM CM



IU
i/>  z




r> * O


- o r-
GO ^" (D
ON*"
888
00 00 CO
a a o
u u u


00 00 00
CM CM CM
(0 10 10


CO 05 CO




coint^oco4





co (D'-cn- incM
ixcMT-iocor-coooen^cooocMio
CM '--'-  '-


cn^cooo oocoeAincn^ co
01 CM  co inoar-inomoo r-
*COCMV^
a
o









^
co
CM

d




^
CM


O
8


^



w




r;


r-
CM
CO
8
oo
Q
3


CO
CO


in




CM
ID
oo

















(O
^


^
(D
CM

d



a

o
8

0
0


CO



2*




en


CO
^
^J-
8
oo
O
3


in
CO
CO


in


A-13

-------


















,_,
< H
I Z
< O u.
Q U Q
- U
HI 111
-1 O.
Z CO 00
3 Z "
co 5 t**
a
ui 3 n
H u.
M u. CN in
co a a -in

Q. < UI I T-
Qh CM
CO 1 v* II
Z oo U
i -no
co x r- * 
I- O - -H
< 11 W H 
I a -at
Z co ^
a - c. c
r- in M ra o
00 H '0 M
X ii C
U. Z 104'
11 -a * oi
u. at c co a
o o 3 s.
-J O C (Q
OX i O H
n 0 -H
i- a u m
CO O -H
M U. 
~* C

3
5


c
o












































































































X
L
01
CO >
0 0
H U
01
at
0

o Z
W CQ
o
M ?
H 0
a L
M
01
at
(-_ ^
0 0)

O 0)
u a



, V
a a
O -v
-i m
a


L.
0)

ID 
CO O
o *
o o
o o
oo oo
a a
u u





(0 r-
W CO
(O (O






in in






































^
^^
Q
O
CO


CO t
** z
CO 111
S" O
_i co at
Z co a.

CO _J 0
-i w < in
< CO Z <-
u > <
M _l O
t- < -1 t-
CO Z < CO
i 5 L) ZO
1- M < *
< -1 Z U
t- < UJ CO U.
CO U Z O
11 u K- m
Z Z UJ in
M UJ Z Z > (9
5n oa 2
i- 5
u o g a a:
co Z u Z ui
3 l- I-Z
n 11 UJ X UJ O. O
1- t- (/I < t- UJ 2
U < UJ UJ M CO
M a a a. a >-
a i- a. u < at
HZ a z ui
CO UJ Z UJ O 1 >
UJ U O H M X O
a 2  z H < o
O H UJ < Z UJ
< t)3 z at at
H -i a x
< il 111 < Z < CO
a a ae o uj a
CS O U. w Q. n
at ^
O O n if ii ii ii
U. O
_i o u. 11 a. at
a
UJ
CO

CO
UJ
o
o
u




A-14

-------





















a

UJ I/)
J 2
CL 3
Z Q
5 13
I/I U.

u a
3
(A
O I/I

O*>
^
u o
i/> a
51,1
H
^
oo M
>- a
u. O
n T
00

u a
ZO
M U.
t-
*!
-1











































g
U
U
a.
i
oo

^
K
CO

CM in
-in
IU 1 *-
a CM

oo u
-co o
t* ^ '^
H
co H n
- a
N- 1 c
^ A O
(9 +>
C V> 8)
3 c.
OCA
P'2
O 4*
u ra
c
n>
g
^
(Q
C
d)

c
1-1



























































































X
L
0)
CT O
fi O
01
at
c
o *

1-1 at
c
0
n M
V O
a L
i/i
Q)
QC
00,
SO)
Q.
*""


X-*,
a w
o \

Q^
^r

01

A 'm(Or-'-t
cntocMcocooor-oocncnt^cnr^cncoocn
^ ^

CM  
cn co cn

* d 




ox a




co - in - (0 * oo
vi^eoiocMcoin^cocnocoinooooiou)
oocoO'-cocnin*inenco'-ocor-o
- CM'- '-'-'-CM'-CM


to f oo r** M to
n CM CM oo (D cn oo oo CM ID
in mm in t r~ CM CD ro in CM en r-

o oo d N o -'-- '-cMod




o in o o
- CM  -



ot^ocnonooomiooooncMinoaiin
*- en to oo CM (0
r- r- r- oo oo r


ooooooinooo^tnoooocM
go "-
^
^ ^ ^" *" ^^ *~

CMCM*"^^CO^COCOCMfOCMCOCOCOCOT*



IUOU O IU IUIU (J (JO
Z22MI/)ZMZCOZZV)XZl^ZZ




S!2S**CM2cOCMS?COCM^CMCM^

t-U)WeocMen-ocDOOcot-en(0(0
4f M M v v (0 0 A O) Is* If) f^ 00 00 00 O) *~

888oo88888So
oaaaaaaaooooooooo

^WtN"OOOTO^CNCO^t1*COO) O CM ^ ifl
u)iAU)tnif)(0(0(0(0(0(D(0(Ot^'tN
CM
8
00

o
00
CM
CO

CO




CO
oo
cn


8

*









CO
-












o

at


CM
^
00


o
Q



CM




cn




cn

^
in
CO
8
00
8

^
00
CM
(0

0?




oo
v
en














P.
O




cn









O
*

r-
,_
cn


o


^

CM



O
z




CM

^
CO

i
00
0

CM
CO
CM
(O

0?


A-15

-------
















( v
11

UJ ""
-i
z c/>
^ z
at
UJ D
1- U.

co a
oz
a*
82
IH
(/) X
t- 0
< W

2
a
1** U


u. a
o o
(3 T
Z 0
i- a
in a
j












































a
Q
U
HI
a
00
.
**.
CO

a CM in
a -in
O ^
Ul 1 
a. CM
i *- it
oo O
-co o
H
CO II IB
- a
CM a
- (. C
^ (Q O
?"
IB +
TJ -H U
c in 01
3 L.
0 C IB
f-21"
U 
00
HH O
(U
Q
o 
M ID
or
c
o
IB *
H O
(B -
0 t-
U)
01
C
00,
3a
^^


. ^
Q 0)

0


L.
0)
(0 M-
a m
3

4^
 *
o 
L JZ
01 X
a.
4^
s *
u "m
L Z
0)
a.
a

01 0
u
U) C
3 O
V) 
sl

ft] ^
i c
ol
41 O
a


0)
A

m*-*'-cM CM Tt in in
inior-oiDcMioco <-(ocoo

~66*-o*-~~ ~ ^ ~ ~




O O 0
oaaoroaaaaoiocroKo:



r-OiflOCMOCor~Ooocoocom
cMininp^oooococoinr^^Of^
oooooooooooooor-(Or*ooOCD(D



O'-OOOCMOOOOOVOOOCM
or-ooo'-inooo^oooio
m o co * w o f- *t o *t v
v* ^ ^



COCMCMCO'-COCONCM'-eMCMCONCM



U CJ U U UJ tj U
cnz2a
-------


















~
HjE
Q 8
-V
UJ
_J

00
00 <
u.g
H u.

(fl Q
o z
a3
82
M
00 X
1- O
^ **
I a

a
r- ui
i
u. a:
0 0
w
H a
00 O
M U_
J


















































i
u
UJ
a.
i
oo

CO

Q CM in
a -in
O'- 
Ul 1 -
a. M
CO 0 "
-co o
^ * 
*
co ii m
- at
CM TJ
- L. C
'-wo
1,2"
A v
g v 0)
C 00 0)
OCR)
g J
u a
44
,^
44
C
A
3
*
(0

^
Q)
c




































































































X
L.
0)
t/) >

a
o
c ~
O +*
M (Q
a

Q
A 
+* O
A *
a c.
M
01
a
6^1

U Q.
^f



a oi

^
~~~,

^
01
n <*.
A *-
a A

^J

^i
C 0)
OI-H
11
Q.
S *
UA
01
&

ft
&3
0



U) 
C D)


h
jj3
s.

g
0)
1

X

o c.

(Q S

O 3
X3 2
(0
-J


U
m


o

CO
^













CO
in
an




^
an
O




O
a




i

S

o
8





M




(/)



P*


(D
CO
CO
ACD870O




(0

CO
(O






in


o

CO
"














m
f-
^



an
in
CO




a
a.




o

6
oo

O






CO




I/)



in


t>-
0
ACD87OO




^
^
CO
ID






in
































CM



1/1
H4

I/I h-
n Z
I/) UJ
> O
-IV) Of
< n UJ
z co a.
00-1 O
_j n < m
< oo z -
i J 0
1- < -1 1-
l/l Z < OO
n< U Z O

^ J Z (J
(_ < uj (/, u_
00 O X O
IM UK CO
Z Z u ui

X ^i 00 2C
Q O H* ^
ui a Q o at
vi z ui 2 ui
3>-il- HZ
iiz^sl"
PPtoSSuli
(J < UJ UJ M I/I
M a K a. K >
ot fc a. ^ o < a
HZ a z ui
I/I Ul Z UJ O M >
Sz'Sz'H'SS
O H UJ < Z UJ
< UD z a at
H -10 X
< If OJ < Z < 00
a o a o u o
CM (_> u.  a. w
a -v.
O Q ll II II u ll
u. O
-i o u. M Q. at
o
Ul
(/I
3

00
Ul
o
o
u


A-17

-------
IU I
_i;
5S
I/I U.
UJ I
*-;
in
  i x
  10
- at
u. a

u. X
O U

a at
2 o
           a
           u

           Z
         U. CO
         a   -
         O CN CN


         T7-
         CO CN
          - v-  H
         t- o
co M  n
  -    a


:ii
 II C
   ra +^
   ?-H 0)
    0)

 ^J
         o 
            It
            n


            01
                                u
                             v> >
                             a o
                             o  -H
                             11  (0
                                a:
                             H  O
                             CO  
                             a  L
                                4->
                                VI
                                01
                                a
                             u  01

                             32
                             a  o>
a n

   8


c u
0) ^H

iir
01 2
a.
c ai
01 -
u n


I*
                             <  L
                               O


                             (A  C
                             3  O
                             I/I -
                             C  0)
                             Q)  01
                             U 0!
   a


   at


   I
   n



 x
 L
 O (.

 n A


li
      Sco    (0   ^    co    (0    O   oo    CN    to    01   (D    co    o   O)    r*-    o    (D    01    o    o)    '**    *
      r-totDcocor-o>oooo(Dcor-coin*rcocooor-mr^r-





                 CO                                                                         00

                 *                                                                         ci









UOKCJMUUUUCJU         UUOUOi-iCJUUUOU








      (o    CN   en    in          in                                                       CN
ojiniouicor-cocoiDtocoiQaioicflincnNiDlDUltDCN

^coco^(O(0(0^*t(OCNincn(0^^"^oooooo(Oin




      r-                                                  *                                                                     ,
      co                                                  CN                                                                     V
      O                                                  co

      O                                                  b





                 in                                                                         o
o          OCNOOOOOO         ooooo<-inooooo
i-a--tt''-'-''--'-*-ar'-'-'-'-'-oCN'-        -'--
OQ-Oi-iOOOOOOQ.OOOOO>-iOOOOOO




or*-Oo>cnooo0>(DOOcncNinO(oin
                                                                  oo
                                                                  Q
                                                               f~
                                                               ao
                                                               O
o
o
r*
CO
a
o
o
r*
oo
o


i
r>
co
8
                                                                                               iO
r>
oo
                                                                                                           OP5r-o)ioco
                                                                                               OQQO
                                                                                               OOOO
r-
oo
r
CO
r^
00
i~
CO
r-
00
a

r~
co
a
s
o
t^
oo
Q
r~
co
a
u
    in    CN    *
r~    co    en    in
CN    r-    CN    CO


I    I    I    I
      Sco    co    co
      a    a    a
o    o    u    u
                                                            MMCNfNCNCNCNCNWCNCHMPICN
                                                                                                     CN    CN   CN    CN    CN   CN    CN
                                                                                                  ID    f-    en
                                                                                                  !    r-    r
                                                                                                  CN    CN    CN
                                                                                                  ID    ID    ID
                                                                                                             O    *-
                                                                                                             oo    co
                                                                                                             CN    CN
                                                                                                             ID    ID
                                                                                                                                                                 
                                                                                            CN

                                                                                            00

                                                                                            CN

                                                                                            CO
                                                                                                                                                        CO    CO    CO    CO
                                                                                 A-18

-------


















~
il
tj \,j
u
J
a
52
i/i <
at

t- u.
V) Q
O Z
aS
o co
oz
(/) X
$2
5
a
si
M
u. a
o o

i"
t- a
,
01

I/) >
o o
M O
01
a
o
c -
O +
M n
0
C
0
n v
H O

a tl
V)
01
IT
uo,
5nt
a



a a


_

^
0)
fl <+
^ *
^"
a n
B
+>
C 01
!i
Q.

4-*
s u
U (0
L3E
01
a.
 *
0) O
o
M C
3 O
VI 
C O)
01 01
oa

c
fll ^
ic
ol
Q

Q)
ra
i/)
X
L
0 L
A Ij
^ g
0 3

U

03
^a.co^cMr.o^m^^o
CNin^OnCOCMCOCOCOOO
f*r^r"(DlD(D(fl(OlD(Or*-(O


^.
r-

^



CM
X^
a o ooooo o t-"
3



(0 r- CM r-
t- (oioco^-cMinor^ooco
- CM CM CO N CO -


CO 00 (0
O - in o
in -  *

o co b b



in
__o_oooooo
Q.ZOO.OOOOOQ.OI->


coo^r^oinofMOcor^o
0)OoooOooaOOoOOoor-(0



OOO'-OOOCMOOOOO
o mo oo  t o t-
^ ^ ^


COi-PJCMCMCO'-COrOfMCM'-



uj u o o uj u
^^m^^^^ty>jg^gtyt^

f
*



innr-*^o(Oinr-^coo)O
^" ^" "^ N ^ ^ O4 ^ ^ W
O)oo*-Or*co^ococoino)
CO(*> r~ - in -
- oo r- o> ^
r* (O (D Is* r*


^
oo

b





M (J




* * CO
co ^> p. 9 (0
^


oo in
CM CO
CM in in

b b b



o
- o
fV ^V ft ^ ft
^ U \J ^ V
IH a. a. o a


oo ro o co in
r- - o  '-
r- oo o (D ID
^



< O O O CM
 o * 
^


CM CM CO CM CM



O ^^
Z Z 3 00 00





oo (O O t co
CM - cn co *-
CO  CM ID CM
6 CM CO -
8888
a a a a a



- CM < (0 00
ooooo
CO CO CO CO CO
(a in 
8
oo
a
o



in
CO
(O


in

A-19

-------


















^
< K
< 0
Q U
Ul
 1
Z to
to <
at
UJ 3
t- u.
n
to o
21

to
z
n
to x
H O

I Q
Q
r> UJ
CO 1-
11
_J
O ^
11
i- at
to o
M U.
















































U.
O
o
*.
1
CO
(yf
V
.
u. n
a -
U M CM
X -W
1 1 i-
00 CM
 II
r- t)
 n o
^ i -
* ^
P) il (0
- DC
"fe

(0 -M
c to o>
o c (0
Q.O 1-
o +
u 
o o
>-i O
01
OK
0
c -
O -H
11 (0

c
o
(0 V
H O
(0 -
a L.
VI
Q)
a

00,
32



Q 0)
o -^

n
s_,

3)
.^
(0 <4-
(0 
a to



Is
o -
L 
01 2
(L
50)
^_
u to
a> z

a.
*l
u
VI C
3 O
u> -
C O)
Of Q)
01 4J
o i
(U U
f\
Ifl
o
IM
Sample


X
i.
0 C.

(Q 5
(_ E
O 2
Q 2
m
j
_-
O

(0
00
*- r>

(0 W
(0 (O










U 0





M O

CM n











o o
o o



" O
r- O
^ 00
O O

8 8


C4 CO



to i/>


^ in


ID r~
n o
n *
 
Is- r-
00
o
u u
4 ^




(D r-

CO CO
ID ID





in m


































ts

g


to
M
2 5
to uj
-J 10 Of
< 11 UJ
i 2 -
en -i o
-J n < in
 <
M  1 O
t- < -1 t-
tfl Z < to
5o zo
1- M < *
< -1 Z U
1- < UJ tO U.
to U Z O
1 U H CO
Z Z UJ uj
M UJ Z Z >- C9
511 00 Z
H <
Ul Q O Q Of
to Z ui 2 ui
3 K <-Z
11 M UJ X UJ Q. O
K- 1- tO 3 t- Ul Z
O < Ul Ul H tO
M Q Q O. DC >
a. i- a. u< a
H- Z Q Z Ul
to ui Z iu O H >
Ul (J O 1- n X O
06 Z w Z f- < O
8H U < Z Ul
3 z a oc
K- -JO X
< u ui < Z < to
O O OC O Ul O
CM u u. n a. n
ac ^
O Q it 
-------



















1-
Q
Ul VI
li
(X) U.

Ul Q
h- 2
M ^
(/)
O (/)

x
o
M
I/I Q
^m
< o
Z 
r Z
00 M
bu O
ii T
06
(9 OC
ZO
M Ik
I/)
1













































U.
s
X
00
p*.
Q -
U CM CM
X -CM
T 
1 1 ^
00 CM

tx O
-CO O
H
co ii n
at
*" <9 O
,,?M
tQ 4^
?* 92.
 o>
3 t.
o c n
Q O b-
Uk W ~
O +"
u n
c
ra
a
ra
c
t.
V
C



























































































X
(.
u
0 0

at
0
c -
0 v
w 4

n
It
n
cn
X
c.
0 (.

(Q 2
^ g
0 3
r
u
n
a


o
o
(0













CM
a>
CO

00
CM
b







O
a.


o
8
^


o
in
CM


CM


111
z




CM


fs.
^
CM
8
00
a




^
in
<*4
(O





o
CO
r*








at




CM
0
CM

in
00
o
b







a
a.


r-
cn



O
o


CM


f\
rp




en


in
CM
^
8
00
Q




in
in
CM
to





Oen*Ooo*cMcneMO'-onixr
(O^cotoooocMincflvOcooenixo
(OlOoooor-OTODoocooor-ooini-co

O (0 cn co
r~ o (n in

*- CM i- i-

CM
^
O 1-4 H H l-l
3



CM (o cflintxcococnincMcooo t-
CM'- oo-(DtxinoooocnrxcM

o in oo on t ix ao o oo
(O CO (0 CM CO  CO CO CM COCO
b b CM b b b bob bb





o o o o

ooQoaaacotxoooacra
o.a.za.i-1-ii-ia.Ma.a.a.i-io.a


OcncnooocncoooencMinotoin
S5fSSS?enSS8^12|2Scoi2
^*


oooomoootinoooocM
OOOO'-OOO'CMOQIDO'-
inoiDOoo^mooin 
" "

'-'-'-CO'-COCOCMWOIfOCOCOCO'-


u c_j w u u cj o cj
jgmt/i^mym^p^t^^jgc/tjp^




W'-^N^COWM^'OW^-WW*-


COOOCNO)'-fl'O(D(DOCOt^OJ(DtD
c4^MT*(0(OO)O).r-u)rs'Ooooooo>^
O CO ^ O ^ ^" CO * CS ^ W iQ ^ O ^
888588888800000
QQQQQQQDQQQQQQQ




r^GOAO^cHCo^r^oootoc^^ui
U)iniri(DU)(0(0(0(OtO(0^r>>r4-tVi
C4MCV(SCMMC4C4C4C4C4C4C4CNM
(0(0(OtOU)(0(0(0(0(OCDU}(0(D(0





-

-------















^
< H
1- Z
^ O
Q U
Ul
j
Q.


Ul 3
I- u.
H4
CO Q
O Z
|*


CO X
1- O
1 3

Q
1** M1
00 (-
u. ce.
o o
01
20
IH
CO O
J







































u.
a
u
*
1
00

.
*
U. CO
Q -
O CM CM
X -CM
I *- 
00 CM

r- O
-co o
(0 T- 
CO ll IB
- IX
CM TJ
-U C
T?M
IB -H
c co 0)
3 .
o c m
f*~
U IB
C
IB
3
"
c
L.
01

c




















































































>*
0)
(/) >
a o

OL
o
0 -H
i-* ni
a
c
o
n -H
v u
s.
a
o



^
on
on



^
on in
* O

o co






a a
a. z



co o
co o
en o


o o
 I


CO *


UJ




in co
00
CO
CO CM
8 8
00 00
8 8
< <



CO *
CO 00
CM CM
(O IO
CO CO


oo
^
t^















on
,J



(D
CM
00

0






O
a



*
CM
00


O
O


CO



CO



t>
00
CO
8
8
<



(O
00
CM
(0
CO



o
(0















en




on


O






g



r-
in
CO


*
in


CM



*



*
o
CM
8
00
<



00
00
CM
IO
CO


CM
^
(O















o
IO



CM
^

d






a
a.



o
in
00


o
8


CM


U
2



o
10

8
8
<



on
oo
CM
(0
CO


r-
CO
(O















tn

V*


(0


d






o
a



in
oo


o
o


CO


^
2



ID
00

8
00
a
<



en
CM
(O
CO


0 r.
CM CO
(0 (O

w
(0
**





M




(O 00

 ^



,_
r>
CO

d




o

u. a
t- M



0 N
O eo


O CM
o *


* CO



a to



in r
" 0

8 8
 3
 -" Q M hH M
O
~"


(O^cn (D - in en CM in into

r^uicM *c^in^(binoocMen



- 10 t 10 in CM en
CO^inCMCMIOOO 00 t^O
in o in CM m o v CM ' o r~

O'-OOO'-O" O CMO




o in o o
f CM *- -
aaaaaooaaaaaao
1-iaQ.ZQ.O.Q.Q.i-ii-iQ.i-cl-a.



cor-ooocoocoinoocMcnrMO
cor-(Or~oooiO(OOcocoo3
-------


















^_
< H u.
1- Z Q
< O U
Q U X
UJ 1
-J oo
a.
Z 1/1 r-
5S V
at
iu 3 u. n
H u. a -
M CJCM CM
i/ia x -CM
OZ I - 
0- < ii*-
00 CM
I/I - II
Z Is- L>
M -mo
in x (0 - 
I- O - V
< M nun
I a -at
Z w u
O  L> C
r lu - nj o
oo t- Q M
U. Z M *
11 -g -M oi
u. a c 1/1 n
o o 5 c.
-i g c 
ax 9.0 H
H a u n
/i o v
M U. 
C
ra


^"
AS

Q)
c
M













































































































X
c.
0)
" 
IX
o

0 -M
1-1 n
at
c
0
 -H
v u
nj 
a t-
M
01
a
^^
00,
82



O Ol



>

L
0)

(Q **-

o IQ


+j
C 0)
8Z
I. 
01 X
a
+J
C 01
8
0)
a
gj 3
O) O
< L

M C
3 O
c ro
41 0)
uae

c
0) ^
i c
u o
0) U
a
M
O

Q)
1



X
L
O (.

(t J3
c. E
?3
2



U

fl
0}


^ r-
S*
(0
















Sin
00
r ao



v ai
r> ao
10 Is-

b b




CM
x^
o
O O O
a a -i

M

~ O l/> H
r-- O - O
-i 1/1 a
< l-l IU
4 ^
O O l/)_l O
_j i-i < in
8 8 S>5
 -  _| 0
t- < -1 t-
l/l Z < I/)
H 3 o zo
H n ^ ^
CM n < u z u
H- < IU - (9
5 HI CO Z
1- <
iu o Q a at
I/I Z iu Z iu
3 w K HZ
giz5SH
r- in tfMtuxujao
.- - HHWI
a: H a. o < at
HZ Q Z UJ
(O Is- IO LJ Z UJ O >-" >
 o uiooHi-ixa
co  a z  z i- < o
O O OHiu uj
^ ^ i/i
co n 3
(0 10
t- t- i/i
IU
a
8

in m


A-23

-------


















4
o
IU CO
-J Z
i
i/i a.
UJ Q
S3
(/)
o vi
a z
o x
u o
n
1/1 a
<
5
! Z
00 11
> a
u- O
n T
00
eg a
zo
M U.
5
-1








































u.
8
X
1
00
p."
-
*
O (N CM
X-N
1 1 *
00 M
- ^ II
Is* o
-co O
(Q < .
- 4^
* II (0

N (Q O
TJ 11
n C
(0 -H
1"?

0 C 19
IS
u to
M
ra
o
_

a o

Q
O 4^
M <0
cc
c
o
(0 
+ U
10 -
a L
VI
d)
a
U O)
O Ol
^-



Q U)
Ja
01
,^
(0 <4-
10 
a m
3


8,u
-
c.r
01 3
Q.
^
C 01
01 i
U (0
L Z

oio
< L
in c
3 O
VI -
C 0)
01 01
U ft

I c

u o
s
a
a
M
01
n
CO
X
L
O L

fl .Q
gl
(0

U
n
a

o
o
ID

CO
CM





n



Ol
6








in
CM
ae
n



O




O
in

CM



IU




CM

r-
*
i
00
a



- CO CM CM
00 CM 01 -  0 ID
 * jo (D 01 oi r
i i 8 I  i i
a a o a a a a
o o u o o o u



S*~ W CO ^
(O CO (D (O
CM CM CM CM M CM CM
(0 CO ID CO ID (0 ID






01
O)
(O












Ol
CM



00
r>
o



a



o
0
^






CO



z




00
^
(D
in
8
00
o
O



f*
(D
W
(0



w


CM
ID
CO






CM










^
s
o







01
J^



in
CM
ID

CM



VI




f*i
n
Q
f-
i
00



00
(D
(N
(O



Oi


o
n
t-






CM

Q
O









r*
-



i



CM




O
8

CO



s




^
CM
n
00
i
CO
a
u



O)
(Q
M
(D



W


'-enr>r-'-cocoint-in
Oenr-ocooiooi'-'-
coin^oocooor-mr^r-
^
r-.
d



CM CM CM CM CM

ooaaoiioouu
00000



n
ui ^ o ^ ,;
M CM  ^


c^ CM in
CO  ID 00 CM
O O O O " -


o
ooaoo5S222
ZZZZZnOOOO



inoiDintintooocsr--
|5co3|5?5eno!SoI



OOOCMOoOOOOO
(DO'-OooOOOO
in * o * o

COCOCO'--eMCOCOCMCM



ZCOZZZ3Z3COZ




^incor^oococMCDOi^
'-CMCM'-'-'-ro'-'-eM
r-.oi(DiDt*ineM
eooooi-'-r-oooiinco
o o o o o o o ^J o
o o o o o o o ^s o
oocooocooooooooocooo
8QQQQOOQQQ
oooouooau



^5 W ^ 1A (D ^" O) O ^ W
^r^r^r^r^p^r^cooooo
WWMWCM
-------
















^
z
<0
5i
IU
1

51


H U.
IH
I/) Q
O Z

jj
82
I/I X
H O
cnOT"CM^(poooi
cocncncnocncnOOOOOOO
(O(0(0(0(DCOCOCO(0(0(1)(0(D(0



coco^^^^^^f^^^tninin




T-
^
r-







U





CO
CO
^









O
o



o
in
00



o
o



CO






o

N
fv
CO
8
00
<



o
CO
(0



in




0
r-
CO







U




00

00










O
o



CM
in
CO



o
o



CM

z




1^

O

Q
8
00
8
<



^
CO
co



in




(0
,_
r-







U





O
CO
w









o
0



en
00
00



o
8



CO

z




CO

CM
in
^
8
S
u
<



CO
CO
(O



in




oo
co
in





CM

a
o









(0


w








CM
5



O
8



^>






r^

r-
CM
CO
8
00
Q
<



f
CO
co



in




(0
CO
CD







U





CM
CM
^*









O
o



0
8
^


o
0



CO

z




en

CO

^f
8
co
a
<



in
CO
(O



in


A-25

-------


















,  ,
< 1-
1- Z
< o
a u
UJ
J
a
f VI
-l
ra -H
c oo 6)
3 L.
o c 
a o
i-i O
0)
at
0
c -
o *
at
c
o
10 +>
H O
^ 
a (.
VI
0)
IX

00,
82



Q 0)
o ^
-1 O)
a


^
01

n> *

10 ^
O ID
a

+^
C 01
St:
"r
at as
Q.
4^
S5
O (9
(. Z
01
Q.
0,1
< t.
V) C
3 O
VI 
C 0,


i c
o 3
5 o
oL
i^
a
Sample


X
t.
O L
** S
ra j3
t. E
Q 3
J3 Z
10
i

U
*>
A
CD
_

(O
(O










o





o

CN












0

o



r-



0
8
^





(/)







S
CO
8
r-
oo
Q
U





ID

CO
(O





in


r~

^
(0










u





4_
CO
CM












0

o



o
0
oo



0
8


n



i/i



in



o
CDS 700
^




^.
-
CO
(O





in



































CM

Q
o

1/1


CO t-
>-i Z
00 UJ
> o
_i uo at
< l-l UJ
2! ^
CO -I O
o> 1 "
w _l 0
1- < -I 1-
00 Z < CO
5o zo
1 i-i < V
< _IZ U
1- < 111 00 U.
co u x o
IH (J h- CO
Z Z UJ w
IH U Z Z >- (9
51-1 CO Z
1- <
u o oo at
i/l Z UJ Z UJ
3 lit- t- Z
<<<-!
ooz SL
H4 1-4 111 ^ UJ & O
t- t- 
a: t- Q. o < a
t- z a z in
00 01 Z IU O i-l >
UJ 001- - X 0
a z w z i- < o
O t- UJ < Z UJ
< OD Z flC OS
i- -i 3 x
< n ill < Z < I/I
Q O 0 O HI C3f
CMOU. M a
at ^
o a n ii i 	
U. O
_i o u. <- Q. at
a
UJ
I/I
3

CO
u
a
8




A-26

-------
o

UJ I/I
O (/>

S5

8S
v> o

< Q
o-S

ss
IH U.
I/J
          O
          CJ
          X
        u. ro
        a   -
        CJ CM CM
        X   -CM
        en CM
         - V- II
        oo CJ
         -co o
        CO  II  (Q
         -    at

        lc

        -as
        ii  C
          I/)
          c 
         .o t-
       O (0
          
          01
                            X

                            01

                         a o


                            ce
                            o

                          o **
                          M  O

                          a I
                            +<
                             M
                             0)
                            ce



                          u  ai


                          82
                          a  oi

                          3S
   L

 ra 4-

o 15



 5

a.
*/
 2

 *
a.

 at
< i.
                          So.
                          0) 01
                          U Oi
                         I;
                         O C
                            a
                           03
lOr-ioiDajcor-cncocococnr^couitcococof-inr-t--




8








          CMCMCMCMCMCMCMCMCMCMCMeMeMeMCMCMeMCMCMCMCM

HoroaaaaaaaoaoaoooooaaaQ
          ooooooooooooooooooooo
          _i_i_i_i_i_j_i_i_i_i_i_i_i_i_i_i_i_i_i_i_i




en
(0


b



     O)   O)   f1^              01         (O^CO         '"CO         (D    ^                        CO



     bbbbbbbbbbbbbbbb--'--cMO-




in






Or-OcnenOooenioOOcneMinouJin^-uiioeofMc-






opooooinoooinoooocMOcooooo

to   o   o   o    o    o    ^m    o   o   o   ^r   w    o    o    (0    o    ^    o   oo   o   o   o   o
                                      CM   CM   '-
                                                                          uiui
                                                                                                                        -    -   CM   CO   CO   CM   CM
                                                                                                              U              CJ
                                                                                                             Z    S    c/)   Z
                                                                                                                                  ID   CM   ID   en
                                                                                                                                  -   CO   *-   *-
                                                                                                                                                       CM
                                      
                                      r
                                      oo
                                      in
                                      (M
                                      ID
                                           in
                                           CM
                                           1
                                           in
                                           in
                                           CM
                                           ID
                                                     88888
                       CM
                       03
                                                     r-    r-
                                                           2
               (M   CM
               ID   (0
                                      I-
                                      00
                                      o
CM
ID
                               f-
                               00
                               a
          f
          s
                                                                     uo
                                                                     <<
CM
ID
CM
03
          8
          r-
          

          S
CM
CO
          8
          r-
          s
          u
CM

(0
          8
          r
          oo
CM
ID
                                                                          888
                                                                                               a
                                                                                               o
CM
(0
CM
(0
CM
ID
                          p-
                          00
                          o
                          u
CM
ID
                          r*
                          00
                          a
CM
(O
                         8
                         r-
                         oo
                         o
CM
(0
                          r-
                          oo
                          a
                          u
                                                                                                                             ID
CM
ID
                         t   in    cs    r    *r
                         >   oo    en    in    eo
                         CM   *-    CM    eo    *

                         88888
                         r-   r-    r-    r-    r

                         a   a    a    a    o
                         u   u    u    u    u
CM
ID
CM
(0
O   -
oo   to
CM   CM
ID   ID
CM
00
CM
ID
                                                                                         CM   CM   CM    CM    CM    CM    CM
                                                                                                                                       CO   CO   CO   CO
                                                                         A-27

-------
< I-

< o
O (J

HI



li
10 <
   at
UJ 3
I- U.
HI
I/I O
o z
en x
t-o
< HI


* =
fs. U
00 t-

2
   HI
u. a
o o


25
I/I O
er
                              X

                              V
                           to >
                           a o

                           ~S
                              0
                           C -
                           O *
                           w 18
                              o

                           18 -H


                           S.S
                           O L
                              +rf
                              V)
                              01
                              K
                           O 0)
                           C v.
                           O CTI
                           o a
                           O 0)

                           S2
                           <8 *.
                           C 01
                           u ^

                           y
                           a.


                           S2
                           U (8
                           0)

                           OI
                           (/> C
                           3 O
                           I/I -
                           c a
                           0) 0)
                           O K




                           -*i
                           U 0
                           01 U

                           %
                           X


                           O L
                                                             01


                                                             N
QOOQi-i         00000000000000000
oooo              ooooooooooooooooo
-J   -I    -I    _l              _I_I_I_I_I_I_J_I_J_I_I_I_J_I_I_I_1





                     N
                     m   CD
                     o>   o

                     b   <-




M                      in    en    to   (0    i^    en         *-         n         o>    m    en   co    r-          
tt   CD    r-    00        PJ(MMO)'-00-O)OOOOCM'--COOOOOM


O   ("J    -    O        OOOOOO'-OOO'-O'OOOOCMO





                     in


iiii2eilgiiiiii;iiii








cnooocoooooocoOoof>(0(^eoo(OCDOoocoooO




OOO'-OOOMOOOOOtOOOCNOOOOOO

dooi^-ooO'-inoooi-oobcDoooooo
     O         mo         oo    *         or~o         *e><               oo
                                                   *-   *-    CM    -
                                                                              >-    N   -    *-
                                                                                                                                         M   (M    *    -   -
                                        88888888888888
                                        r~rt~t^r-r-r-f-i>t^r-r~i-r^

                                        aaaoaaoaaaQaaa
                                                                                                                   888888888
                                         oo

                                         3
                                              oo

                                              10
           oo   oo

           10   to
oo

co
O)   O)


CD   CD
a>

co
a

co
at   at

(O   (0
                                                                                                   IO
O   O

(O   CD
                                                                                                                   r
                                                                                                                   oo
                                                                                                                   o
o

co
                                                                                t^    t^
                                                                                00    CO
                                                                                o    a
                                                                                                                         CO
o

(0
                                                                      r
                                                                      00
                                                                      a
o

(0
                                                                      t^-
                                                                      00
                                                                      a
                                                                t^
                                                                CO
                                                                a
                                                                 r
                                                                 oo
                                                                 a
                                                                 r
                                                                 oo
                                                                 a
                                                      r
                                                      co
                                                      O
                                                                                                                                         CDCO
                                                                                                                                                    CO
                                                                                                                                                         CO    CO
                                                                           A- 28

-------

















^
HZ Q
8 |
U I
-1 00
5z *"
wi 3 ^
a
iu 3 u. n
t- u- a -
w U CM CM
1/1 a x -CM
oz z~ 
Q. < ii*-
O> M
(/I - V* II
z a u
M -co o
I/I X f- - 
HO - *>
< M CO II (0
5 a - at
CM 'O
t- iu "-mo
oo H ^ w
>< M C
u. Z  +
u. S e JJ o)
o o 3 c.

W Z ^L O ^
2 5 v
HOC u n
I/I O -H
M ^ -^
C
(Q

Qf
^
m
c
01
c










































































































X
.
fl)
V) >
a o
OL
o
c -
O -H
I-H (Q
a:
c
o
<0 -P
(Q 
Q L.
U)
a
00)
O W
u a


> n
a 01
o ^
-1 O)
a


t.
0)

(Q *^
V 
O "ffl



^
C 01
01 -H
o -
(. 
01 2
a.
4^
C Q)
Q) ^
0 
01
a.
en <5
4^ L*
O
VI C
3 O

5 a!

c
i c
U 0
2^
W
O
HI
V
1
1/1


X
L
O i.

(0 fi
l- E

Q 2
(Q
j


U

00


*

s








CM
a
o






in

oo

6











^
r*.
*


0

8


CM




tfl




^


(0
CO
CO
8
r*
00
U
^




(0

CO
(0






in



f.

10








CM
Q
0







CM
O

^






Q




O
d
00


o

8


CO




I/I




in


*
8
r-
00
a
y
^




f*
T-
n
(0






in































CM
x^
a
o

(/)
^H

^ g
I/) IU
_i in at
^f ^^ ^j
Z 1/1 a.

I/I -1 O
-i n < in
o > * *~
M _l O
H < -1 H
< -1 Z U
H < IU I/I U.
MUZ 0
M U H CO
2 Z IU IU
M IU Z Z > O
5n CO 2
H <
iu Q o o at
1/1 Z iu z iu
3 w H HZ
O < < w
Z Z H Z w tt
O O Z OS < H
11 M iu x ui a o
H H I/I < H IU Z
U < IU IU 11 I/I
i a at a oc >
ana. u < at
HZ a z iu
I/I u Z iu O M >
IU U O H n X O
a z n z H < u
O H U < Z IU
< o => z 
-------




















t-

o
UJ I/)
si 3
51
 or
u. o
o5
(9 O
Z 0
** U.
V)
1-4
-1








































a a
o a
OO
X X
XX
1 1
co co
r~|x~
(0 (0
CO CO
- -
CM CM CM
- -CM

x. 1 f
00 CM
- *- n
r- O
-co o
CO II (Q
- at
CM TJ
- c- c
^ (Q O
H?M
*
TJ * M
C I/I 0)
0 C Ifl
ft O 1
g-21-
o -M
o ra
**

c
At
a
^_
(Q
C
C.
c
























































































X
L
10 >
0 0
"1
O
0 -H
w CB
o:

c
o
ra -H
H 0
a L.
w
u
a

. *-*
U 0)
8D)
a



am
S^i)
r
^^

L.
0)
.^
10 4-
O (Q
3

^
S2
u -
L 
4) X

C U
u ra
<-*.
CD

a
fi. 3
5)5
< L
u
vi c
3 O
C 5)
OK
Cd -H
3 c
^ 3
|8
a


0)
0-
1
X
L
O C.
v 0)
(0 j]
L E
Si
a Z
(0

U
ra
a


o
oo
(O
co
(0
V-











cn
(0











o
etc



o


o
in
CM




M



UJ
Z



CM

r-

CM

co
3



in
CM
to












*
*
UJ
_l
1
I/I

01

O
O




(14







O)

in
CM
^
8
CO
a



in
in
CM
(0







O ^ ^ (O U) co *~ LO 01 oo (o (0 U) ^ r** r^ w co
co w co o) Is" oo o r* (0 ^ w ^ r^ co ^ co to O
to (0 oo r** (o co oo oo (0 r*** r^ r*- co (D P*" oo r* co
O) (O
* r^
*~ ^





M





IXCMV ID oen ix cocoinr-
^-Oin^incMOcococMvcocM~-oicn(oin

1- r- ^ *- ^ r- -

CM'- p- in n*- in t co
fo oicom'-oi'-co(D coor-CMCO
CM * tcocncnocovcnr- to r- 01 - 10
'-o oo'-O'-O'-oo o b -o --'





0 O
- i^
ooaoaoooooooaoooao



oo>enococn(ooocncMino(Dinin(o
S5fgSS5SS8?:SI5S3!S5So

oooomoootinoooocMOcoo
OOOO"-OOO*CMOO(OO*-OooO
inoioooovioooin t o >-
* T" ^ ^" ^ V* T~



^~* (O^COCOC^CONCOCOCOCO^^CNCO



O O UJ UJ UJ L> O O O UJ



O^^ooc^coocow^fr^uicor-cotof^
M*-^-C4'-COWCM-CON*-CN'-'-'-CO
cooi-*otO(oo^cn<0(D'*in
cM^^cptooienfxm^cocococn* ^-r^co
O C*) ^ O *^ ^ CO ^ CM ^ CM O ^ O ^ O CM ^*
QOOOOOOOOOOOOOOOOO
oooooooooooooooooo
COCOCOCOCOCOCOGOCOCOCOOOCOCOCOCOCOCO
aoaoaaaaaoaooaaoaa
oooooooooooooooooo



CMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCM







CM
CO
(0















CO

"


0
CO






a



co
CO
en

O
O




CO



3



(0

CM
en
CM
8

o



0
CO
CM
(0


CO




co
(^
^














r-
0



^.
*
V-






a



CM
t
CO

o
o




CM



*



en


in
CO
8
co
8



CO
CM
(O


CO




CO
CM
r-














r-
P.



CO
co
b






a



p-
^
01

o

^



CM



Z



^
CM
^
CO
^
8
r*
CO
8



CM
CO
CM
(0


CO


A-30

-------















^^
< H
K- Z
< 0
Qi\
\j
UJ
-1
X (/)
5 z
i/i <
a
U 3
1- U.

too
o z
2

H O
$ 
a
f*. m
525
U. Z
u. 
o o

jj5
t- a
M O
M U.
J




































a a
o a
(JO
z z
i i
oo oo
i- r-

(D (O

CO CO

M CM CM
- -CM
*-  
00 CM
- *- II
r- t)
-co o
*<
co u n
- at
" ? c
 -H
TJ ** Ot
CMS)
3 L.
OCA)
l^1"
u ra

c
(Q
i
1
01
c
l-l
















































































X
c.
01
o o

a:
o
o 
H 
at
 v
+ O
ra 
a t-
i/i
ac
* v
u ai
ft n\
8 a



*-\
O O)

-J O)
a


L.
01
ID -OOO

inr.0(Dinr.*ooooo

in
in
(O
CM
8
oo
a
<


cn
O
CO
(0


in



'-
(O
r-













^
[~.
cn



(0


^




a


o
in
00


o
o



CO

M

0
CM
CO
8
00
a
<


o

CO
(0


in



*"
^.
!*














^
r-


T~
^
00

6




o


CM
in
00


o
o



CM

*

r.
CM
o
in
o
8
CO
a
u
<


^
r*
CO
(0


in



CO
in















CO
CO
*~


ID
in

v-




o


cn
oo
oo


O
8



CO

i

00
CM
in
i
CO
a
u
<


CO
T-
co
10


in



oo
00
(O













to
r-
CM



CM
(O

^




o
^

CM
*


O
8




-------

















^
!E 
< o u u

UJ ^ II
_l co oo
Z (/) P r-
co < 
0 0

~!
0
0 V
11 
O >,
-1 0)
a


t.
01

4^ -~
(0-
a ra


C 0)
01 -M
u -
L 
0) X
a.
4^
S -
u n
c. z
01
a.
0) O

3 
VI 
C O)
U 01

HI ^
ic
fQ
y


a
0)
I
ra


X
L.
0 (.
H 01
m ja
^ g
0 3
Q Z
n



U

n
CD

^

00
(0













in
01
(0


(0
(O

^





CT
a


^_

f*.
If


o
8


CM


CO








(0
CO

co
8





(0

CO
(0






in



U)

t-
r-














in
CM


CO
^

^>





0
a.


o

6
00


0
8


CO


CO




in



O
i
oo
^f




^
t
CO
(0






in
































CN
o
a

to
n
I/) h-
i Z
5
M -1 0
H < -I h-
CO Z < I/)
M O
X H 03 Z
a u i- <
uj Q g a at
to Z u Z uj
DMH K-Z

O O Z S < H
11 n uj x ui a. o
I- H co < f- ui 2
O < UJ UJ (1 l/l
M at at a. a >
st H a o < a
HZ a z u
CO iu Z iu O n >
U UOH M X O
a z n z i- < o
8H UJ < Z IU
3 z ae a
H -JO X
< n UJ < Z < I/I
a o at oiu o
CM u u. M a. M
a \
O Q II II M II II
U. O
_i o u. t-i a. at
a
UI
CO
2

CO
UI
a
a
u




A-32

-------

















a
a
u
4C V
2 I
Q oo
UJ I/I ("
CL< 10
5 S5 Q m"
i/) u. a -
O CM CM
H- Z I *- 
M 3 1 1 *"
c/> en CM
2^ - ^ n
Z oo CJ
Z 11 -co o
 2 co" u 
MO -a
K CM 'O

Ig ;|S

00 M B * 01
x a CMS)
u. O 3 L.
J 0 C (Q
o **
So 2
K 
(/) C
11 
-1 3

_
(Q

4^
C





















































































01
a o

a
o
c 
o +-
1-1 ra
a
c
o
(9 
M O

a C
(A
a
^^.^
00.
o 01
u a.


O O)
3 w
a
*
L
01

ra H-
(S >-
0|

^
C 01
U^
*
In 
01 3

+J
C 01
u n
J. Z
0)
013
0) O
< L.
C9
M C
3 O
01 0)
u oc.
c^
1 
U O
So
a
o


,
X
t.
0 -
P 0)
"f2Scol2?



ooooooinoooinoooocMOooo
UIOOOOO'-QOOfCMOOUJO'-OooO
CM inocoooocoooin * o -
V- * ^ 1- W" ^ ^

CMCM^**^CO^COCOCMCOCMCOCOCOCO^^'CMCO




lUOO O ul UJUJ O CJCJCJ UJ
Z2Z(/)(S>2ZZZ3Z


CMeno^^ocorMcoooocM^c-incor-ooiocM
'--CM-'-CM-COCMCM'-COCM*-eMCM-'--CO

^cMCMTvvw(ococno)r**(nrN>oooooocn^*^r*>oo
0000 ooooooooOoooooo
O O O O O O O ^^ O ^ ^^ o o O O O o Q ^
SOOQOCOOOOOOOOOOOOOGOOOOOCOOOOOOOCOOOOO
OOQQQQOOOOOOOOOOOOO
UOOOUOOOUOOOUOUOOOCJO



^in^oocno^cMCO^^cocnOcM^inco^en
ininininmcocDtocococDCDco^f^f^r^^F^r**
CMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCM
COCDCOCO(OCO(0(OCO(0(0(0(0(0(OCDCD(OCOCO




CM
00

-------














^
< 1-
1- Z
< o

_,
UJ
_l
X I/I
2i
at
IU 3
h- U.
1-1
 Q
?!
(L <
jp
5 vi
u z
M
(/I X
<2
i S
t** UJ
521;
U. Z
M
00=
u
fit T
z5
H a
in o
j








































Q
Q

X
I
1
oo
[V,"
(0
a m
o -
OWN
5*C * W
?7-
O) (M

00 U
 m o
m H a
at
^ (Q 0
u !~
fl V
u +^ 0)
C co 8)
o c 
O v
u 
V
*
c
IB
d
n>
c
^
di
**
c
IH


















































































X
L.
tt\ %
2
o/
o
O -H
M at
c
0
ID ->
M O
(Q ^*
a L.
M


U O)

0 Ol
u ex
^.^


Q O)
3S


c.
(U

(Q t*>
-4-* 
Q (0
g

SO)
^
r
tl 3
a

is
tu ^
om
01
a.
a
01 3
OIO
< t.
u
3 O
| 
C O)
01 0)
ua

c
OJ -'
M 2
I1
Q
jfl
ns
(O
X
i.
0 L.
Si
JJ Z
ra
JZ
u
n
oa

rmr-.-" Q. >-" Q. >-"
0
J



co in r* in o i/)


r- O'-O)O>c<<*>CiNOO)'-oo-Moa'-eoc<5Oi
* M^ ^ ^^.^^ ,-,- ^ ^^^ ^


oo - r^ in in r-
co (0 in in in in f t M^-n 01 co M
Ooan^-(ooo (OOMcnooNcnooto ^-oo- r-




o o in o o

ooo'ooooiooooo'oo'O'ot/iaiao'c/iaa


cootr-omoiNOor-OconocoinoocNoicNO
WOCOOOOOOOOOOOOOt^COr-OOOCDlOOOOOOOOTO
^~ * T* ^ ^ ^


OOO'-OOOCMOOOOOtOOOMOOOOOO
OOOr^OOO*-lflOOOOOO(OOOOOOO
O ino oo* or-*o* * oo
^ ^ ^ ^ ^ ^



C*)^(V)C^CNCv)^B"OC004Cs4v"CSCs4COCNtN^~*COC4fr)^CO


uj uj o uujujoo cj o y
^^m'K^^jpm^^gtjn^^^y^/ifjflyi/t^^f/i^


*


wor*otf)tnr-5rcoo>Ooo(OO^c5iflOr^oor^(OCO(0(D(D(D(0(0(D(0(0(0(0(DCO(D(OCO


 *ininininmiOLnin


A-34

-------
















^^
< H
< 0
a u
!j
Z I/I
< ?
i/i <
a
UJ3
^ li
r^ u*
i/> a
o z
a*
O I/I
uz
(/} X
1- O

IQ
a
00 H
U. Z
u. a:
o o

n
vt o
_i



















































u

00
pj
B

g n
O CM" CM
X -CM
X - 
1 1  U
-i i/> a
^ ** in
i 52 "
. . ^23 S
O < I/IZ -
o o>S
- - w _l O
1- < -J H
I/I Z < I/I
>i  (9
X i-i 03 Z
1/1 i/i a u i- 3
IU O Q O t
i/l Z ui Z UJ
D 1-1 1- HZ
0  
z H z M a
o z a < H
P- U) i-l l-l UJ X UJ Q. Q
- - HHI/>
ana. u < a
HZ a z iu
ID P- (/) UJ Z UJ O 1-1 >
co o iu y o H M x o
m t a 2 1-1 z H < o
O O OHUI
o a a o a o uj a
u u cs o u. w a M
O Q II II II II II
u. O
_i u u. M a. a:
a
(OP- IU
f - W
r> ro 3
10 (0
- - i/i
IU
a
o
u

in in


A-35

-------















S
Q
UJ I/I
j Z
& <
ig
M U.

UJ Q

I/I
Ii
So

tn a
i-
< a
Z K
r- Z
00 M
> at
u. o
U. I
o u
O or
Z O
^ Urn
H-
in

J











































L_
a
o
a.
X
i
00
r-
ID
^

CO
.
CM

H

1
3
0










































a
u
i
oo
r>
ffl"

^
CO

CM CM
-O

CM
O "
co o
4*
ii a
at
o
L. C
?O
w
HI
(0 4-
** SL
M a
i.
OH

n
+3
c
m
a
^
2
^
0)
c
























































































(/)
a
H4
g



2
a





g
Q
U



a
o





2



44
Of
u
Q}
a
4*
c
u
0)
a.

0)
01
<
tfl
I/I
tJ

c
f"







>
L
O
H
n
L.
o
Jd
n



X
V
0
1
+
nj
a
c
0
0
C
VI
0)
a

*-s
O)



-^
O)
X..
01
a


t-
.2
*.
(Q

^


0)
^

^fc


V
IS
2

Q.
3
8
L
C
O
0)
0)


i
u
a
M

0)
M
i-
A)



(.

ft
g
3

U
n
a

CM
(O
(O
r-
00
0)










O)










o
a




o
8



o
in
CM



CM


UJ
Z



CM

r-

CM
8
00
Q




^
in
CM
(0





Ooir-,-ccococMtn~ocor-*-ino*
SSISSSPgJSrtS^SPSSSIC?
oncMO coco m cotoo
*^-r in oi o in CM uj
Or;'- 2" N " " UJ


at






co(or.cooocncoo^r.ininoo(ocn
oop-inenr-inooTfo>rcMcoeM*cM*ooin


- CM - COCM
b bbbb ^ .^ . ^~




oinoo oo o ooo
atatatataatataaoOKaxxxaa
MMMHia.Htl-ia.O.O.O.HlO.HIHIHia.Q.



r-OencnooocncDOOcncMinoioin^'in
S*SeOUJS8^*r?Scor??
^


ooooomoootinooooMOco
ino(oooo(oooin t o -
'"*"*"'" *" "~ *"


CM^^^CO^COCOCMCOCMCOCOCOCO^"^CM


ZZWM^COZflZZM^WZZ^3



r-CM'-'-eM'-COCMCM'-COCM'-CMM'-'-'-

incoootM-*0(0(oocor-o>(0(0'*
CNN^*^(p(00)O)is>u)is>oocoooo)^^r^
^ O C*) ^ O ^ ^* O ^ N ^ W O ^ O ^ ^5 W
o o o o o o o o o o o ^f ^?   ^^ 
^^ O O ^5 O O ^S Q O ^J O Q Q Q Q ^J Q
aaoaaaaoaaQooooaoa
ouuooooooooouoouou



SvcMCO^c^cocnOc^^inwr*-
CDG3(0(OCO(O(OP^f^r*f^ir"f*"
CMCMCMCMWCMCMCMCMCMCMCMCMCMCMCMCMCM






-
CM
r-
CM
b



HH





ao
r-










O
a




10
0
en



O
0



CO


UJ



CM
CO

in
eo

8
oo
a
<



on
r-
CM
ID

CO




en











r-
in



^
(O
^m




0




oo
CO
en



O
o



CO


s



ID

CM
cn
CM
8
oo
a
<



o
oo
CM
ID

CO



r-
10
r-
ID
r-
O



**





00
ID










O




CM
00



0
0



CM


I/I



en
V

^.
in
CO
o
o
ao
0
u



^
00
M
ID

CO



10
^
r-











00
r-



in
CM
^f




a




r-
a>



O
8



CM


0



CM

^
CO
qf
8
ao
o
u



CM
00
CM
ID

CO


A-36

-------



















<

Q U
-I
2 (/)
< ?
I/) <
0
u a
H- U.
M
cn a
21

8
< 
x a
a
j-i
U. K
OO
(3 T
5o
t- ox
(/I O
u









































o
u
ft
00
f-
to
a "*
o co"
X CM CM
i -O
00 - .
P- CM
580
CO II (Q
- a
**-! c
"55
n C
A -F
g +* u
C c/) 6)

OCA
f^1"
u n


c
A


n
c
t.
01

c
l-l
















































































01
a o

a
0
c -
0 -H
It (9
a
c
o
A -H
+ U
A -
Q L.
V)
01
a

. , V
g o
82


Q O)
O-v
3

01

A <*
ii
U TO
3

c 
01 S
a.
+
C 0)
u "A
01
a.
a
8L3
0O
< c
I/I C
3 O
I/I ...
C B)
U Q)
u a

fQ
o
a


01
i/i
>
O 1.
H 01
A 3
O 3
A

U
m
Ucor-wcooo^cococN^o
lDttcn(Ncnr-r-O-cnco
r-r-r-(0oooor-inc~rr-io

cn in (o
O ID V

CM O CM



CM CM
>^ X^
i-i o a M M
0 O





o m co -- in in 10 * -co
* o r- t ID o co ao (oao


CO (0 CO *
inr-r<-cfl ^-(oen ID
^ooioat cnoor-cocM
^bbb bb^^^



o o o
aacraaaaaoaoioi


coo^r-omOcMOcor-o
cnOoooooooQOooOoor^co


OOO*-OOOCMCOOOO
08w8S5lS08
^ ^> ^.



-Cl>WCMn'-COCON04'-


U U O U W U
3ZVISZ22MZZMZ


^ ^ *"" CH ** *~ CN ^ ^ CS
coino>
oo9ooi^iDr>>kOrooin^*f
nwww^-ocM'-'-'-co*"
888888888888
oooococooooooocooooooooo
QQDQQQQQQQQQ
*

o^(Oooo)^co^u)r*oiO
OOOOCOOOOOO)O)O)CnO)O)O
C4C4MC4MC4C4C4MC4CICO




COCOCOOfOCO^^^^^^
o in 


o
CM

CO
8
00
a


o
CO
(O



in
oo
cn
(0

CM
^f

^





hH






00
in








0
a


CM
in
00


o
o




CM


^\
2


r.

Ifl


00


*-
CO
(0



in
^
ID
r-

^.
CM

CM





i^






^
ID
CO







o
a


en
oo
00


0
8




CO


u
z


oo
CM
in

i
ao
a


CO
CO
10



in
ID
CM
(D

















00
(0


in
00
b



a
"

CM
;


o
8




*~



VI


^
r-
CM
ro
O
O
00
o


^
CO
(O



in
,-
ca
P~

^
^

u>





i^






CM
CM
^







O


O
8


o
o




CO


^
2


cn
CO

^
8
00
o


in
CO
(0



in
A-37

-------





















< H

rf O
0 U

ui
i
a.

3 Z
I/) ^
a
UJ 3
H U.
tH
i/> a
o z
aS
82
IH
(/) X
H O
x a

o
r- uj
CO H
> <
u. Z
I-H
3
z o
H at
(/I O
M U.
-1

















































a
u

^
t
CO

.
(0

u. 
a -
o o
Q.
I W N

CO - .
r M
IB (J
-n o
^ f -^
M II (0
- OC
 o
a I
in
01
at

. ~*.
U O)
C "^
o m
o a
"^


,-^
O 0)



^^

L.
01
(0 4-
+< 
fl 
o a


^
C 0)
01 -H
o 


a
C 0)
01 
u ra
(. X
0)
a
ft*
0)0
4 ^
(9
vi c
3 O
if

E C
0 
v u
a


0
0)


>N
U
o s.
V 0)
(Q 1
U e
El
(0
J
c.
u
4->

00


M
CO
(O


















co

(^
*~


M
en
CO

O





O
a.



^

f^
^


o
8


04




00


(s.
^




870O336
Q
%




IB

m
IB




in




^-
in
CO


r-
9)

M





1-1






en

co
m











o
K
1-1



o

6
a


o
8


m




00


in




8700407
Q




^
^
m
IB




in







































N
a
o

 o
_i oo at
< H U
5 2 a
*3 ?

M -1 O
1- < -1 H
oo z < 
w < 0 Z 0
5 -.5 S*
H < UJ I/I U.
00 O X O
M (J t- M
Z Z UJ ui
M UJ Z Z>- C3
5M oo Z
H <
uj a o a at
00 Z UJ Z UJ
a M i- HZ
Q < < 5
zZHZwa:
o o z a: < H
I-H ** UJ ^ UJ Q. Q
H H I/I < H UJ Z
O < U UJ M t/J
1-1 at at a. a >
oe H a. u < K
HZ a xui
I/I UJ Z UJ O w >
UJ O O >- >- X O
a z -c z H < o
O H UJ < Z IU
< u =) z a a
H -I 3 X
< II UJ < Z < (/)
a o a: o uj o

-------





















t-
o
u in
o!z
x a
(/) U.



i/i
25
x
o
l-l
i/i a
H-
 K
U. O

o5
zS
l-l U.
1-
M
J





































Q
U

2^
i
00
r-
(0
u. 
tJ CO
I CM CM
1 -O
en * ^

to CM
 v u
t- O
 co o
+rf
CO II IS
- CL

^" (Q O
?M
(Q 4*
TJ 4J U
C (/) 0)
3 (
OCX
o +
o 
s
3
i
Q)

c
M


















































































0)
00
" u
w
a:
o
c 
0 V
M n
a
c
0
ra v
M U
 
Q t.
VI
0)
at

. ~*
y n
c ^
82


0 O)
o ^
^r

t.
0)

(0 C
O *(




C fl)
gu
tS
Q.

^
C 0)
8 
s
0) O

II
U 0

C
I1
o
1-4

0)
(0
I/I
o t-
f $
ra 5
L E
?l
^^

U
10
CO
CM
S








CM
"S^
o
o
J









U)
*-










o
8



o
in
CM

CM

z



CM
r-
^
CM

CO
a
S

^
in

(0




Ocn^-tncoocM
SSSSSPSIS








CM CM CM CM CM CM CM
*** ^""^ ^^ "*x '^, ^s ^s,
ozoooaaoa
o o o o o o o










r r- *- in
O CM r> en n in in
0-0-00--





OOOOQOQQ




r-OencnOeoeniD
^tncMCMinniocn



doooonoo
S8588
" "
riT-^T-n^nm

Z Z   Z  S 

*

2S^?82coS
inc*)ooc4Oi^vo
CSN^"^tptDO)O)
^ O CO ^ O ^ ^ CO
8S888888
oooooooooooooooo
aaaaaaaa


inr^oooiO^c^co
lfltfltOU)(0(0(0(0
C4MC4CMC4CNC4C4





v^on^ + ^uo*-
SS?Sf:5SS"??!








CMCMCMCMCMCM CMCMCMCM
"\ *^ \. ^^ >s ^^ *vXi >sx ^X X^
oaooao aoaa
oooooo oooo




r-
00
00
o


CM r- r- co to
coen co i^to^cocoin
00"0'-00'----W





ooooooacoooo




OOcncMinoeointinu)
OOi-en
-------

















< H
< O
Q U
a.
Z (/)
1/1 <
t- u.
M
CO Q
O Z

^
oio
u z
to x
1- O
^ M
*=
fx tlj
Ei
u. ae
o o
f9 T
So
H
h- Of
co o
IH U.









































0
u
OL
00
*".
10
U. f
o co
Z CM CM
1 -O
O> - 

CO CM
 * II
r- O
-co o
co u re
CM'TJ
v* A ft
?"
ra +*
TJ * 01
C CO 0)
3 t.
o c ra
o +*
u ra

c
ra
3
,_
n

(.
0)

c


















































































>.
L.
0)
a o
~
_ o
c -
O -H
*H (Q
a
c
o
ra +*
** o
ra -
o c.
VI
01
a:

. *-*
U 0)
82


a 01
O v.
-i ai
a

t.
01

ra it-

ra >
Q n
o
C 0)
L 
01 X
a.
44
S "
ura
0)
a.
ii
< L
U
M C
3 O
01 
01 01
U K

C
aj +>
g ^
/* X
U O
rt) Q
a
o
M
0)
M

ra
1/1

i_
O L
H 01
ra j3
(. E
O 3

U

ra
oa

to
to
^







CM
a
o
i









10







Q
Z

CO
0>


o
o


CO




X



in
en
s
8
CO
8
<


CO
00
CM
to


CO



00
v
r"







CM
a
o









s
^






i

o
8


o
8






UJ
Z



CO
00
CO
CM
8
8
<


00
CM
10


CO



f-
q.
^







CM
a
o









(O
en
6









CM
00


O
o


CO




1/1



P

00
CO
8
CO
0
<


ID
00
CM
10


CO



in
en
(0








a
i









CM
*"






a

r-
in
00


-
in


CM




*




o
CM
8
co
O
<


00
CO

ID


CO



CO
CM
oo
(O
in

CO




M





CO
oa









o

at

o
in
oo


o
8


CM



UJ
Z



O
CM
10
8
CO
<


en
00
CM
ID


CO



CO
ai
co














*
-


CO
*"






at
t-

U)
r-
00


O
o






M
z



10
09
r-
8
00
O
<


en
CM
(O


CO



'-cococM^ooin(ocoinr-aj0co
-------




















< K

Q CJ
UJ
J
Q.

I/) <
DC
UJ 3
K U.
IH
I/I Q
0 Z

2
o 1/1
u z
IH
(A X
52
5
Q
p*. m
25
u. Z

U. Of.
o o
i5
IH
t- a
I/) O
M U_



















































O
^
^
00
-
r

10
U. *f
a 
U CO
a -
Z CM CM
i -O
o>  
- 1 ^m
to CM
- i- II
1- O
-co o
CO II (S
CM'TJ a
- u c
-wo
DM
ii C
9 +>

C IA Ol
3 L.
0 C 
Q.OI-
5 +J
u 
a o

QL
O
C 
0 -H
M (Q
a
c
o
ra v
V U
(0 -
O i.
1/1
0)

00,
82




a O)

-i ~n
a


t.
01

n <*
^ .^
(8 >-
O (Q
3

C 01
1
a

4^
c o
8
OJ
&

a
M 3
8) O
< C.
u) c
3 O
M 
if

h
Si
0) U
a

g
0)
1
i/>

X
i.
o c.

^ 5
O 3
j2 2
n


u

m
0}


CM

cn
(0








CM
^^
a
o









r-
CM

i"







a


^




o
8





CM



Vt


r-



eo
8
i





(0
n
(O





in




^

in
oo








CM
x^
a
o









in
in

T*










o
b
00



o






PI



I/I


in



s
i
r-
00
u




r-
n
(0





in





































CM
\
o
o
VI
M
I/) H
IH Z
I/) UJ
>- U
_i in a
f^4 UJ
i/> a.
^
(/) _J O
-i IH < m
w _l O
V" ^  J ^

M 3 o z o
h* ^*^ 3 ^
< _IZ U
1- < Ul CO U.
MUZ O
M Oh- CO
Z Z _ UJ UJ
IH u Z Z > O
X IH CO Z
o u t- 2
uj a o o a
w z uj z uj

ZZHi2iM
o o z a < t-
IH M iu X 1" CX Q
1- -< 1- uj Z
U < UJ UJ IH 
a t- a. o < oc
K z a z u
(/) U Z uj O IH >
g2zK28
a H u < z u
< O 3 Z 0 a
H -JO X
< ii uj < Z < VI
o o a o uj o
CM u u. M a. M
O Q ii u u u u
U. O
-i u u. IH a ae.
a
UJ
3


UJ
a
8




A-41

-------






















t-
^
o
UJ I/I
ii
I/I U.

ui Q
3
VJ
O I/)
a z
O X
U 0
l-l
i/> a
<
^S
00 M
X IX
u. a
.J
00
(3 IX
ZO
(-
-J









































Q
Q
O

00

r-
(O
a V
a -
O CO
a. -
Z CM CM
i -O
CO i- 
r- CM

ID O
-co o
eo ii ra
 IX
" L C
ii C
'U +* 01
C I/I 0)
3 i.
o c ra
Q n tmm

O R)
-H
ra

ra
c
01
c
























































































0)
I/I >
a o
KH U
01
IX
o
c 
o *
IX
c
o
ra v
H O
(8 -
a (.
VI
0)
IX


. ***.
00,
o fn
3 a



a o>
3 w
a


01
.^B
ra *
^ 
<9 *-
Q (0


S2

a.
^
CO)
O fl
OJ
a.
Sie
M c
3 O
IA -
C 0,
0) 01
oa:

E C

|cS
o

01
f
1
I/)
X
c.
O (.
H 0)
ra a
i. e

ra
c
u
A
oa

in O O en
CO CO CM CO
en O o en
CM -

cn
CM






M IX







oo *- O en
(O ^* (D oo
oo in in 


^ co CM
co r* oo
O 0 0




o

a a a o


o r~ o 01
t in M
en oo 


o o o o
in o o o
CM

CM CM *- *-


uj O O
Z Z Z faO


CM en o cnnO(O*
oocoooMoioooor--TOin(ooor--(Ocno)
- ^^ ^^M^^CM - - ,

(O <- co cn CM (0
-oO'-O'-'-O'-oO'-wcocM'-oo






aooooaaooaaoaooooo


OcooiiooocncMinocDin-o-iniDoocMr-
oaincniocnoc-r^r~cooor>^oocncnoocn


oinooovtnoooocMoaooooo
'-OOOVCMOOlOO'-OOOOOOO
(oo5o(OOOin t o >- o

CO'-COCONCOCMCOCOIOCO'-'-CMCOCOeMCM


O UJ UIUJ O OOO UJ O
Z(/>Zl/>ZZf>3Zl/IZZZ2Z3(/)Z


ococMCOOoocM^r^inco^ootocMiocn^
CM'-COeM'-COeM'-CMCM'---'-CO--eM
en'-*O(D(OOcor-o>(DiDt*incM*t
U)(DAOtts>LOr**oooocoo)^H'^^coo>uico
iiiiiiilillliliiil
Soocococococococococoeoooeococoooco
o888S||88gSS888SS



o^cNn^tt^coOTON^rifltDr^oio^cN
(0(0(0(0(0(OCD(DrVk^*r^ts*t>>^t^r^ooooaQ
NWWMWO4CSC4WC4P4WC4WNNNCS
{0(O(0(0(0(D(0(0(0(0(0(OCO(0(0(0(O(O





A-42

-------
















N

< H
K Z
28
u w
u "
-i
too
z
oo 3
oc
i- 
^
MO

%
82
t-i
K O
< M
T Q
2
Q
fs, j^j

u. Z
1-1
u. a
o o
ax
|5
K Ct
VI O
^ u












































Q
a
u
Q.
X
CO
t*
(0

Q *
O CO

X CM CM
i -O
ao - 
- 1 1-"
- ^ It
ID U
-co o
co u ra
a
CM -a
-'I
55
M C

o


















(0
*~

^
CM
(0
O





a



o
in
00



o
o




CO



oo



o
N

CM
r-
n
8
CO
a



o
CO
(0



in



o
(0
o


















2


CO

on
o





or



CM
in
00



o
o




CM



*



r
N

O

8
i
^


^
CO
ID



in



0
^


















0
r-
^

CO
en
o
O





o



en
00
CO



o
8




CO


y
*



00


CM
in
8

<


CO
CO
(0



in



en
r-
cn

















en
0
in




in
,_





a



CM
J



o
8




^



to



r


r-
CM
CO
8
i
<


*
CO
(0



in



0
8


















co
in



^
^
r.





a



o
8



o
o




CO


y
*



on


CO
*
8
oo
O



in
CO
(O



in


A-43

-------



















^

< H
Si
ou
Ul
Z co
Ii
a
Ml ^
1- U.
tH
co a
oz
a. <
^
O CO
oz
cox
t- O
x a

a
t- UJ
00 h-
U. Z
M
u. a
o o
_i
a x
Z o
t- at
CO O
JjMj ^
J


















































a
a
u
i
oo
pj
ID
a t"
a -
0. -
X CM CM
i -O
co - .
Is* M
 ^ II
CO U
 co o
n n 
- a
CM TJ
- t- c
'-MO
"21-1
ii C
M +*
^ +* 01
c co 0)
^
 M
ao (-

U ffJ
+rf
^
c

o
M
C
{.
01
f
c








































































































X
L
Q)
CO >
o o
1-1 IT!
a:
o
c 
O -H
l-l (
at
c
o
H O
(8 
a L.
VI
01



U O)
c\
52
*_

Q O)
O \
Q.
.*_^

L.
ID

- o
-1 CO OZ
< l-l UJ
Z - 3
M _i a
H < _l h-
CO Z < CO
p$y is
< -If U
t- < U CO U.
CO U X O
l-l O t- CO
Z Z UJ u
M UJ Z Z > C9
X i-* 03 Z
a u i- <
u o Q o at
CO Z UJ Z UJ
3 w 1- 1- Z
Q  i-i
z K z - a
oz a < K
M M UJ X UJ 0. Q
(- K CO < 1- UJ Z
(J < UJ UJ w CO
11 ox ac a. oc >
ce i- a u < oz
H- Z Q Z UJ
CO Ul Z CU O M >
z-2zr38
O 1- u < Z lu
< o D z a oz
< ii ui < z < co
a o oz o u o
CM u u, M a. M
OZ "~s
O Q ii u u n ii
U. 0
_i o u. ti a. oz
Q
Ul
CO
3

CO
Ul
Q
8




A-44

-------



















H
Q
UI I/I

ol 5
5 ?
(/I U. Q
O P-
U O O 00
^m Z O *
w 3 i o
I/I CM
O I/I - PI
Q.Z U
X i-i co o
 2 ii 
i/) o u. a
i- a o
< a o - c
^P ^| Q M ft
*fc 2"
< n C
P- Z It 4*
00 w TJ V ffl
>- a c t/i 9)
U-0 31-
-j o c ra
U- Z o ft u-
00 I;
eg a u n
z o +*
M U. -
1- +*
I/I C
_I 3
_
*
C
OJ




















































































OJ
o o
 u
OJ
a
o
c 
o +
l-l fl
a

M V
H U

O L
VI
HI
a
0 Ol
c ^
801
a


^k
a a
o^
J5

t.
0)

  c
3 O
VI 
C Ol
0) 0)
0
Sv
- 9
U 0
0) U
a
o
0)
I/I

X
L.
O -
H 01
a n

Q 2
ns
j

U
ra
CO
P- O O
oo en CM
en co CM







CM CM
N^ -^
Q a o
a o
J wJ







(0
00 (O *
ID (D -
O O *-




  


O P> O
* in
en oo
-

O O O
in d o
M


CM CM -


* z1 8

N en o
^" v* ^

* CM CM
CM * O
888
00 00 00
< <: <



* in P-
in in in
CM CM CM
(O (0 (0





OOOOfo(DOCMOfoom-en**
*-CO*-QinCM(0*-CMCMCO*-'-CO*-
T-^Oencoop--enors-p-cneoen

CM P-
U> in

0 b


CMCMCMCMCMCM CMCM
x^ x^ x^ N^ %^ x^ >^ x^
0 O Q Q Q w 0 Q w 0
OOOOOO 00



co oo en CM cocnoo *-
*- oo - co p-enr- CM
00 *- T- CM -CMT- M


in n r~ ID CM en
oo(Do>p-- en*- '-cococoin CM
ineooocM^tvcncn oo o co en (O oo
~O-OCM'-OO - o O '----'




in in


cnenooocn(OOOencMinoa)inin
SS$S8;:?

ooowooowinoooocMOco
S858885S88S582


-"CO'-COCOeMCOeMCOCOCOCO'-'-CM


..^coz^SSai^^.^^^*

*OCOCMCOOWCMP-inCOP-00
O









U





CO











jj


(D
O
01

O
o


CO


^

CM
CO

in
00
o
O
00
a
<



en

CM
ID



CO

0
in
CO

T_
^

b



u.
HH




P-
CM











O
at


00
CO
01

o
o


CO


*

(0


CM
01
CM
8
oo
<



0
oo
CM
ID


CO


O
CM
CO

^
P-

O



^
l-l




p-
*
CM










O


CM
00

O



CM


(/I

en


in
CO
O
O
oo
a
u
<



-
co
CM
ID





0
,_
*~

00


^



^
hH




^
f*fc
CO










o
at
IH


P>
n

o
8


CM


*

*
CM

CO
i
oo
<



CM
CO
CM
10


CO


A-45

-------
















< H

< O
O O
Ul
5 *"
to <
at
Ul D
t- U. Q
00 Q (J 00
O Z 0 

Z CM
O I/) ^H

IH no
 x - -
1- O *-
< W MIQ
I o u. at
Q O L. C
00 H ID H
U. Z  -M

n ft ? yj fij
oo 5 u
(IT rt O U>
2 5 JS K
\- at o 

o a

Q 0)
O \

ft
^
t_
01

ra iZ

 
a m
g

C 0)
u *^
01 X
a.

^
C 0)
u 'ra
L Z
Q)
(L
Q.

Oi O

8

8


CO
00
CM,
(0


CO



o
in
*






CM
Q
O








(0
^

^






Q
Z


O
8
^



0
I




^



Ul
Z


CO

00
CO
CM

oo
a



00
CM
CO






ooooooooooooooooooooo
enocoeov^r-fMO>op-oo*-incoo*cooinocMOcor-OcononinoocMoicMO
oooaoooooooooor>(orooocoOroir)or-oor-o>

-oe-oo^onooinencM-wo-iflNOeMi-co
weM<0ocM222co*oocMco2
-------



















-
< o
a o
Ul
l|

ui D
H u. a

I/I O O co
C3 Z O *
a. < i o
Z CM
o i/ - 
u z u
M CO 0
I/I X *- 
1- O *
< M ii n
z a u. cc
2 a u
a o c. c
i- ui o a o
00 H O W
> < u C

"" M r * g!
u. DC C I/I 8)
O O 3 (-
-1 OCA
w 5 
t- os u a
I/I O -H
H U. "
""* c
i
f_
A
C
i.
01
c
























































































X
01
M >
o o
H U
01
a
0
c -
o +-
H (fl
CC

It +>
H U
(9 
a (.
i/i
0)

"2


L
01
(0 *

o *n
3

5** 0)
+J
o -
L 
0) 3
a
4^
S-
o m
0)
d
a
ul o
o
i c
3 O
5,
oi a
u a

ic
I1

a
Sample

X
L.
O C.
f 0)
n 5
L 6
El
<0
-1

U
4*
n
CO

O

00
T"
^







CM
X^
Q
O







OI
CM





i


-
{J



O
8

CM


I/)





O
ACD8700




 o
-1 C/I OC
Z i/i a.

M -I O
-J M < in
< viz -
u > <
PS!., e
I/I Z < I/I
w < O Z O
< -1 Z U
K < HI I/I U.
MUZ O
n O 1- CO
Z Z UJ Ul
M Ul Z Z > (9
51-1 ca Z
i- 9
ui o g o a
I/I Z ui Z u
D M H HZ
< < 11
nK
o z at < t-
n M ui x u a. o
(- (- 01 3 t- u z
U < UJ Ul M (/I
w a ce a. a >-
ce H a. o < ce
h- z a z ui
I/I ui Z ui o n >
U CJ O - M X O
Ct Z M Z K- < U
81- UJ < Z UJ
3 z a a
i- -i 3 x
< n ui < z < i/i
Q O OC O Ul O
CM u u. M a. 11
o a i 	 i
u. o
_i u u- w a. cc
a
Ul
M

I/I
Ul
Q
8




A-47

-------



















i
U (/J
_j z
z K
< 3
co u. a
a r-
Hi O O co
^ z o *
-<. i O
CO CM
2 CO - II
z u
z IH co o
IH II (V
co a Q at
i- a TJ
< a o L c
In |5
P- Z (9 *"
oo M 'D +* 0)
> a c co a
U. O 3 L.
^ 0 C (Q
It I n n tmm
 
ZO 5
l-l U. ...
1- +>
2 S
a
_
ra
c
L
01

C
HH

















































































L
0)
CO >
a o
ac.
o

O -H
i-i ro
a:
c
o
ra -H
H O
flj -
a L
U)
01
a
U D)
82
a a
O ^x
 1 B)
a


i.
0)

n> M-
a 15
s
S3
o -
l.
V 2
a
44
Si
U (0
i- Z
01
a
O) O
< L
(9
Ul C
3 O
I/) -
C D)
01 01
U K

C
i C
" 3
88
a
f^
9)
f
A
CO
L
O L
H 01
n 5

O 3
u
n
ca

r~OOOOOOCoinOCMOCooin'-en o
co a



00 O
00 O

CM CO



in
2 Z





(O CM
*- CO
 in
r oo
CM -
8 8
00 00
8 8
< <


r en
r- r-
CM CM
ID U)




O
in
CO
"















O
(D
CO
n
en






a

oo
CO
en



o
o

CO




3





(O

CM
on
CM
8
00
a


o
00
CM
ID
CO



o
CM
CO
"















O
1-
r~
CO
CO






a

M
OO



O
o

CM




CO





en

^.
in
CO
8
ao
a
u


^
00
CM
(D
00



o
^
"















en
in
CM
O
CO






0

r.
en



0

"
CM



O
Z





^
CM
^,
CO
1
00
a
u


CM
oo
CM
(D



A-48

-------

















^^
< K
<0

U
iz
to 3
a
U 3
Hll Q
Ik U
25 ? o
Z CM

^J Z O
H WO
(SI 4^ ^- ^
VI ^ ^ ^
HO **
< H n n
z IT,"
Q 0 C C
Is* U O (Q O
oo 1- gw
>- < n C
U. Z fl -H
IH TJ V 01
u. or c to 01
OO 3 C.
-1 OCA
" I"1"
h Q U (Q
wo *
_l <->
c
I
a
L
0)
c
M













































































X
L
01
a o

a
0
O -H
HI (0
a
c
0
M U

a 
V)
Q)
(C
 
o cn
82

^^
a en
o\
ff
~
9)

(0 **
^ '"
(9 
Qffi
Ml
S

C 
Si
a.
^
s w
u "n
i. X
01
a.
 
Oi O
< L
3 O
i?

oa
i C
5E
0
1-4


ra
CO
X
O (.
*> Q
fl 2
^ e
Ei
5

U
m
m


o o o o o o o
in in en o co oo v
CM ^ ^ cn ^~ ^ oo








at x at ct






* U) CO S ^ CM 0
^ *^
CO

0> *t O O CO ^ O)
r*, in v^ f^ ^ ^ *^
r^ w ^ r< ^ ^ v




o o o o o o o


co o < r- o in o
0) O 00 00 CO 00 O
^ t"


O 'O O *- O O O
Q O O t*- O O O
o in o oo
^" V



111 lit 1 \
x z cn 2 z z z

"

in co r- * o to in
^" "!" ^ ^| ^"
e oo - o c~ oo -
CO CM CO CM - O CM
8888888
00 GO 00 00 00 00 00
Q O O O a Q Q
u o u u u u u
<<<<<<<


co n to co 01 *- co
00 00 00 00 CO O) O)
W W O* W W W 
^" ^



O UJ UJ O O



F^^OOOlOOOCDO^fCO
OCOCOWWN^OJCO^
cooin^^-co^cstOM
^ ^ ^ CO ^ O O CS CO *
8888888888
oouoouuuou


^inc^cno^-M^tDco
tn
-------


















 ,
< H
< a
a u
* . 
UJ
I/I <
Q^
UJ 3
H u. a
M a r-
1/1 Q U CO
o. 5 i b
Z CM
O I/I ^n
CJ Z O
p-l CO O
I/I X *- -
HO +*
< w u ig
5 ST:"
a o c. c
r- ui o n O
CO H TJ M
>- < n C
u. Z <0 *
M O * 01
U. K C I/I ft
o o g t.
-i o c ra
*   
H DC u n
t/l O +>
M U. <-
C
n
i
r
E
0)
c
hH





































































































X

01
i/i >
0 0
M fl*
ox
c-2
o +-
11 ra
at
c
o
(9 -H
* O
a L
M
0)
a
^^^
00,
0 O)
u a

00,
a ^
-1 0)
a

L
01
(Q *^
H 
Ifl 
Q n

Qf
C 01
8*
-
t. .C
01 3
Q.
4^
 -
om
01
a.
01 3
710
< (.
M C
3 O
I/I 
01 01
u at

c
i
53
$ u
a

a
n
Sample


X
t-
o c.
H 01
(0 Q
^ g
0 3
(0
J
J.
u

as
m


o

co
*

J

^





M





O
CN
in







O

a
M




[V
^f



o
8



(N


C/)



f*B



(O
 o
-i i/i a
< M UJ
9 >
I/I -1 O
-i M < in
< i/i z -
o>- 3
w -J O
H < -1 H
i/l Z < i/>
i < U Z 0
1- n < *
< -1 Z U
t- < UJ I/I U.
MUZ 0
11 U t- CO
Z Z UJ uj
M UJ Z Z > (3
5n oa Z
H 5
uj Q o a at
(/> Z UJ Z U
3 M 1- HZ

ooz z S < H
11 M uj x uj a o
H H !/>< H U Z
u < iu u w i/i
M a a: a. a >-
a H Q. u < a
HZ a z u
i/i ui z u o M >
U U O H M X O
a: z M z H < u
O H UJ 4 Z UJ
s
o a u t 	 a
u. o
_i uu. M a at
0
UJ
I/I
D

UJ
a
8




A-50

-------


















H
Q
UJ I/)
s!3
*. <*
< a
I/I U.

IU O
|_ 2
** 5
CO
O I/I
Q- Z
O X
uo
IN
00 Q
 K
U. O

u. X
OO
(9 QC
2 O
JHU.
(/I
IH
~*












































Q
Q
O
X
oo
r-
10

o m
Q .
U CM CM
X -CM

I I i-
00 CM
 ^m II
r*" O
 CO 0
4"1
co ii m
- K
V JM Q
'as
u C
A -H
C I/I 9)

OCA
Q.O (-
0 +
U A


C
A

A

0)
C






































































































L
0)
O O
a
sS
M A
at
c
o
A -H
M O
A 
O t-
V)
01
a
^
go,
82


a^>
O *"*-


*^

i.
0)

A 14-

A 
a A



C 0
"i
01 2
a.
+j
c. o

SA
01
a.

0)3
a) 0
< L.
(9
3
M 
c ai


5V
g
^ 3
U O
S U
S
VI
Q
HI
01
A
I/I

X
C.
o t-
A &

o 2
A
j


U
A
CD

o t to co
* in en co
UI (O ("* 00








a





o "t n
CM CO 00


00
in

CM











c- en o en
* CM in (D
en * oo en



O 0 O O

  8 8
 



CM - CO CO


z! i/> ^ z




en * o CM
--



in oo en *
CM " u en
888
00 CO 00 00
 n 




in oo o CM
in in 10 10
CM CM CM CM
(O 10 (0 (O








^t^CMTfCMUJOCMCOaD^-OOUJ*


s
in





IH





CM*- -CMOOOO'-CO -
cooo'-M iD'-roeotooiiniot
incoencM IOCOOI^CMUCMCOID
" " * *~
















(Dtoocooinocor-cocMCMO'-
Ol ^f CO CO U) ^ O CO CO *f If) ^ O ^
(DtOlOiOOcoOooi^lOOOnO^'
t v-


oooooooooooooo

88dci8ciSci8cic;8ci8
^ ^" ^ ^ ^ ^



CO-COCOCMCO'-CMCMCMCM'-COCM


bOZ2XZ^3ZOOOO^OO^OO




coaoa)inOioinooen-r-enr~
CM--^'CM- ^^^N^^T-



OfCMenr-ao^ooincoor^cott)
en<-cno(Or-Oin(OincMn
COOCMCO^OCM^COCOOCO^CO
88888888888888
OOOOCOOQCOOOOOOOOOCOCOaOOOOO
QQQQOQQQOQQQQQ




(Ot'-COCOOOWWCDOO^^^^
WWCJWMWWMWCOCOCOCOCO
(0(OCD(OC0(0(0(0(0(0(O(D(0(0





^ w co co co co ^r ^ ^ tfl in in u* in






























CM
^
a
o

(/I
M

M Z
^^ 
to iu
> o
-1 00 0
< 1-H UJ
^ ^
00 _l O
-i IH < m

M _l O
t- < _i i-
 2 o z o
K IH < *
< -1 Z U
t- < IU I/I U.
OIL) I 0
H o 1- CO
Z Z iu iu
IH iu Z Z > <3
5>H 00 Z
i- <
iu o oo at
I/I Z iu Z iu
3  (- 1- Z
< < IH
IH a
o o z a < i-

h- t- I/I < H IU Z
U < UJ IU H bO
IH a a a. a: >
a: H a. o < at
K Z O Z IU
I/I IU Z IU O M >
IU U O h- M X O
a z IH z - < o
^m ^ 9 ^
< n iu < Z < 
-------


















5
a
u en


S 3
CO U.

Hy
^
CO

& Z
x
o
CO Q
is
2
1** Z
00 11
>- ce
U. O
It T
00
ts at
Z 0
M U.
1-
co
HI
~*













































a
o
u
X
1
00
r
ID

a co
a -
U CM CM
X -CM
i i -
CO CM
* *- II
1- O
-co o
(0 t- .-
CO u (9
- at
w- . c
7i~
 -P
U + 01
c  oi
R o "
Q
U s

c
ID
i
*_
(0
c
0)
c




































































































>
(I)
00 >
a o
M U
a
0 -H
HI m
a
c
o
 -H
H O

a L
in
01
at
^^
00,
82


a 9>
o ^

^
^^>

0)
VB

4^ *^
a n
o

C 01
U -^
"r
01 3
a.

s *
0 
a.

01 3
010
< s.
(9
(A C
3 O
M 
SO)
01
u 

c
i C
oi
0)U
a


a
^^

01
i/)

X
L
O L
ra a
L E

W
i


U
n
m

OTT(aoQ^r--CMcMtt)OcMncn(0'-co(o
^incnrooncocMtotorocMvooi-'-aoro
^*
(0 r-
^ ,_





QC 1-4 l-|





^ w c^ o oi r^ oo ^ c*5 in ^
^-cooo(oco(0 r- o> to O) r- r~ in ^r
-  - - v

j^
(^

O






o


r^a>oo>(ococooinO(orocDt^Oifl^c')(oinc^^f
^ O O ^ fO O ^ CO ^ O ^ *~ 0? O f) O CO ^
888888888888888888
aooooooocooooooooocoooooooooooooooco
8QQQQQQQQQQOQQOQQO
UOOOUUOOOOOUOOUUU
<<<<<<<<<<<<<<<<<<.




inin(Oioo>o>a>oO'-  -
CMCMCMCMCMCMCMMCMCMCMCMtMntOPJCOW
(0(0(0(O(O(0(0(0(0(Dffl(0(0(0(0(O(O(0








*
co
(O













^.
s
1








CM
\
Q
o
I/)

- CO
5 v
-1 CO
< HI
1 SH
O CO -1
< CO Z
o > ^
^ I- < _l
co z <
H. 3 o
t- HI
CM < -11
CO 01
H O (-
Z Z UJ
HI uj Z Z
X HI
co o u (-
uj a o
CO Z UJ Z
3 HI t-
o <
o o z 5
^- *--/><>-
U < UJ UJ HI
M a a o. a
a i- a. o
i- 2 a
(0 CO Ul Z U O
(O LU CJ O ^ Hi
CO Q( Z HI Z ^*
O H UJ <
< u n z a
< II UJ < Z
a o a o
U CM O U. HI
< at ^
o a  u ii
u. O
-j U U. HI
Q
(0 LU
 CO
CO 3
(O
UJ
a
o
U



A-52

-------



















^
1- in
< Z.
a <
_l U.
if


M
vi x
< 2

I Q
O
r- u
> <
u. Z
IH
u. at
0 0
_l
" a
en o































i^
a
u
i-

00
^
CO (0
CM 
U. 1 O
a CM
O - ii
i m o
00 f 
r- n w
- a
CO V
-c. c
CM n o
b w
ii C

p ** u
C / ft
3 L.
o c n
Q.O H
6 V
u n

c

I
C
0)
c
M





























































in >
** u
a
o
c -
O -H
"1 IS
OL
C
o
Ifl -H
+j U
fl .-
Q C.
(A
01
a



. S-*.
gm
82



O)
Q >sv
32
*

(.

m *.

Q lo
<

a


0)
f"
i/i


^"
O L
E
J?

U
n
oo
enr~<0incMP>ooooinm(ffTr-
(DO  '-'root'-cnoooiocN'-'-oiro'-a)
p-oooot-(or-t^ooin^r>(0(o^r-t>.inr-r~(o

l *- oo
CM CM O

^ ^ ^



CM, CM N CM
**^ ^^ >HO.z>-ia.a.Zi-iaa.


iuii ^i uiuiiu,
j U ^ ^ U ^ J U ^ ^ ^ U J ^ U
oa^H oa ^ M CQ x MCO <-> x a i-> x
w a _i _IMO. M_ia -IO.M _I&M
ao.(/ioaoa.i/)ao.oi/iaoi/)aooina.
L:^SiwSS&sS&S3wiSs



i-c^cotoincD'-cocoinr^ocHCooocoinr-cNoo
C4C>4C4C4C4NC4CMCie>C4nOC*)COO
CO(D(0(D(OU)(0(0(D(D(0(0(OOD(D(OU)(0(0(D


^^"*"*" ^vvininuim

A-53

-------















<
1- .
HI
(/) >
a o
M O
0
c 
0 -H
HI (Q
a
c
5*
(Q r-
0 L
W
0)
a



o'o)
c \
82
>r


0$
o n
3 a


t.
V
 -
a n
3


o
l-l

(U
B-
a)
i/)



X
t_
o -
A a
1. E
C
u

(Q
00
O CO
O CM
r- r-






N
a
o
"^






CO
^
CM


(O CO
an in
i- O
0 0



i .2

^

J uj
oa ^
a a
O VI
t- X
UJ O
Z -1




- CM
in in
CM CM
(0 10


r- v

(D ^ ^ *" O (O
r> CD CM co en co
r* (0 to r^ (O r*






CM
Q
O
J






CO
CO (0 >- CM 00
r~ o en r to
in < ^ in


CM oo * in oo *
r oo * co - (0
O * CM - CM '-
o o o o o o



a a a a a a
a. a z a. a. a.

^
UJ < UJ
^ J UJ ^
M 00 ^ M
a. -i -i M a.
1/1 o a o a. i/i
at o at vi
M O UJ O O HI
z o z o -i z




co to in ID  co
in m co 10 r-> r-
CM CM CM CM CM CM
(0 CO (0 (0 CO (O


*- - N CM CM CM

CO
en
(0






CM
a
0
J











00
O)
o
o





^
5
j
03
a
i-
UJ
Z




oo
r-

^


O
CM
O



a
a.



UJ
N/
a
in
s
o
_i




in
oo
(0




(Oco'-enmOcM'-io
-rcoenoiin^oo)
r-r-p~r~rO(ooip-






CM CM
O a
o o
J J





CO ID V
r^ co (0 en *~ co (o
Ol CO 00 ^ ^ -CO - CO
in in * in


cooo- cncop-'-
in(0(0in*>*-on
'-'-'-^inint'-'-
ooooooooo



ooaoooaao
Q.Q.ZQ.Q.Q.ZQ.Q.

^ 
UJ ^ UJ 9 UJ
^ J ^ UJ J ^
11 00 H ^ 03 l-l
_j a. _i a >-" -j a
Si/iaoc/iaaoi/i
o a: en o a
MUJOi-iOUJOi-i
IZCJI-I3EOI




r~oM(Doocoinr-cM
oocnenenaiOOO^
CMCMCMCMCMCOCOCOCO
(OCDCDIOCOCOCOCOCO


cico**ttininin

P.
r~
r
















p-

^*


m
CO
0



a
a.



UJ
\s
a.
o
_i




oa
CO
(0


in

A-54

-------













<
K I/I
< z.
Q <
at
UJ 3
_l U.
Q.
5 a
z
1/1 <

*8
(/) X
1- 0

x a
a
r- ta
to H-
2

u. at
o o
_i
S5
i- a
I/) O
j

























L.
8
u
a.

00
1^

n
U. CM U)
a -in
o *- 
UJ I -
a. CM
co U
-n o
js ^ .^

nun
OS
N-?c
-IOO
1,1"
(8 V
D +* B
C /> 81
3 (.
o c n
Q.O K
5 t!
+
+
c
ra
3
1
0)
C































































c.
(/) >
o o
Q
c -
o +
a
Q
Q ,jj
Q i.
4^
M
01
DC



i  >,
O O)
C ""s.
82



^J
Q ^s.
32
s_^


L
Q)
Q 4

Sample


X
L.
O (.
n a

\1
n)
_i .
r
u

R)
CO
en - in to co co in
ID oo in in co t t-
oo co en oo ^ en co

CM


CM CM CM CM
Q a Q a IH
o o a o
^ ^j j ^





^_
en r- (O
CM  en
- in


 CM 8 - W
  " O 0
606600




o

 1 2 i  i 2



 
3 uj 3
-I UJ it -I IU
CO ^ M 69 
H O. -1 _l M
Q a. (/> o a o a.
Si/) a o ae /)
H * 5 Z ? Z *
SI 3 S 8  8 S




*- w co (0 in (0 *-
in in in in CD to r-
C4 N C4 C4 M M C4
(D (D (0 (0 (0 (0 tf)




i-  CM *-
CM n i^.
01 r- co




CM CM
a a
0 O
-i -i






CM
6
in


- co
co r- 
(MOW
o o o






2  i



^
u 3
^ J
M CO
85 I g
s




O CM (O
en en en
CM CM CM
ID ID ID




m * *


o o o
n en in
O * CO




CM
a
o
J






oo V
on -
^ ^


CM r- n
- i- en
(0(0'-
O O 0






2 a a
a. z



x
ui 3
iC ui -i
w * 00
a. IH
/ a. a
I/) O
S  5
 a uj




oo  in
en o o
CM co n
ID 10 ID




*f t in


r-
(O
en




CM
a
0












co
O

O






O







CONTROL




t-

CD




in


o o
n
0














O ID
ID 6



in CM
CO O
CM t
o o






o o
a a.




IU
^ UJ
IH ^
a. M
I/) O.
1/1
 *
S 3




CM CO
CO ID




in in


A-55

-------

















<
t- V)
at
Ul 3
-J a.
a
*l
I/I <

2
VI X
1- O

x a

a
r- ui
25
u. Z
M
u. a
oo
_i
M
i- at
I/I O
_j





































u.
a
u
Ul
a.
i
oo

.

.
CO
.
CM
II
|
|
g
u
































Q_
Q
u
Ul
1
00
r
CO
CM in
-in
f
t i~
CM
*- II
U
co o
+
n ra
at
TJ
L C
?o
M
ra 4*
H 0)
i/i 5)
L
S"
-H
(Q
+rf
+4
C
CO
3
C
L
01
c
*+


































































a
M



C
o
w


ns
v
fll
Q





H
8





Q
_l




(0
i












X
C.

n
c.
n




X
0)
o
u
01
a:
o

+j
c
0
!H
U

L

01
a:


In
2




ai


_



01
t*.
"n
B



O.
n

01
a
re
1/1






O
E

u
+J
ra
00

tn ^ in ID eo
ID oo in in oo
co co cn oo r**







CM CM
^ \
Q O
O O
J




r>  CM
O v- 00
* r> CM



(0 CM 10 ID
in to oo en *-
CM T- CO O O
o o o o o





a a a o a
z a a a z



^ ^
2 uj 3
_l Ul ^ U
CO ^ n 00
M a J
a a. i/) o a
o 1/1 a o
ui O M o Ul
Z -r x o z




^ CM co to in
in in in in  o a. o
at w o 1/1 at
z s 5 i- s z
O O M Ul O O
0 -1 X Z -1 O




(0 *- co oo in r
CM CM CM CM CM CM



CM CM CM CO CO CO



en CM ^ o O o
CM co r^ co cn in
cn t^ co o  co







CM CM

a O
O O
J J




* en o O
CO * f CO
(0 CM I- CM



(0 (0 en co  cn cn O O
CM CM CM CM CO CO
ID ID <0 (0 (D ID


co   *  in



p~
(O
cn
















7
CM



CO
r-
co
0





2







_l

o
u




r-
0
CO
(0


in



o
o
o
















00
r>
CO



CM
<
CM
O





a
a




UJ

n
a
V)
i
n
X




CM
CO
(0


in



o
CO
o
















0
CM
CO



ID
CO
O





2





UJ
^
h4
a.
c/i
o
_j




oo
CO
(0


in


A-56

-------
















^
1- I/I
< Z
a 5
a
Ul 3
u u.
a.




t/1 V
V) ^
1- 0
< 11

r- ui
00 I-
^B ^
u. Z
l
w
H a
CO O
M U.
-1





























U
Ul
a.
i
oo

i-
n
a CM in
o -in
o  .

Q. CM
1 - II
oo U
-co o
["* ^ -^
CO II (Q
at
CM TJ
fe S
* flJ O
"5 M
ii C
(Q +*
TJ + 01
c 1/1 8)
O C (8
Q.O h-
o n

.^
4.*
C
(Q

1
0)
c
M
































































>,
L.
A)
l/l >
a o
o:
c.2
o *>
a:

n -H
* o
n 
a s.
VI
at
a



. y*.
00.
32





Q SN.
32
,_^

u

.^
m M-
Q (Q
CT



O


0)
&
n
CO


x
3 v
(Q 2

m
j
*
c
u

m
CO

oooencnocoOcM
OOCO-CM'-'-CM'-
rO-enoiOenor

18



N CM CM
^*X ^"^ "^N,
gu a a
0 0







en i- in en o co
co en en ^ ^ ^
CM in *- CM co r-




(0 oo oo in to
* in (p oo o en in CM
^ cnot^iocof^oo
O O"OO'-OO





o
iS5221222z

^ v ^
2 Z 2
3 uj 3 "i ?
J UJ ^ _J Ul X -J
a) * IH CD ^ w oa
HH a. -i j w a
aaioaaoacna
luoN^oujoo^HUj
Z-IZUZU-IZZ



T-cxcotDinto-^ooo
ininminujcor-r-r-
OICNMCSCNMMCMW
(0(O(D(O(O(O(0(0(0



^^-^wwwww



0 O 0
CO CO O
00-















r- o) r~
- 6 -
CO CM (O




f-
10 00 00
0- C1 (D
CM - O






a o o
a. a. a.



Ul
U| ^
^ M
M -i a.
a o 
-------















<
^m W
UJ 3
-1 U.
Q.
co5
^
n
CO X
H 0
< M
i
a
r- u
CO t-
X
U. Z
n
u. ae
0 0
_!
(j x
z o
i- at
CO O
"


























u.
a
u
x
X
00
;
u. co"
Q -
U CM CM
00 CM
- ^ II
r- o
-co o
*^ ^ <*
*
co u n
- at
CM 1]
- C. C
'-MO
B~
u C
It) 4^
TJ * 0)
c co o)
3 t
0 C 19
O (0
+J
<->
c
a
a
01
1



























































>
(U
 >
a o
0)
a
o
c -~
O -H
c
o
(Q 4*
4* O
O L
+t
Vt
(D

U O)
C \
32


O)

O 0)
-i a.
N.v'


C.
0)
(0 *4-
Q (0



0
H4

Sample

x
0 -
H 01
m a
A * 1
2
ra

U

(Q
00
in en co r* o
* en in in t*
oo CD r-. r- r~







CM CM
"S^ -S^
o u u u a
0 0
~J ^



r- en
S* CM
CO *

in ID
*- *
 o

d d





000
a '--'- a
Z 0 0 O Z


 ^
< in 5
_l LJ *; _l
oa * M oa
a & co o a
O CO (X 0
i- 3 a z H
 3 S 8 


- CM co ID in
in in in in at
CM CM CM CM CM
ID ID ID ID ID




^ v - - CM

-
CO
CO









u




CM
en
CM










0
o






CONTROL


(0
ID
CM
(0




CM

<-
00









O




ID
in
in










o
o




UJ

n
a
CO



,_
r*
CM
ID




CM

-
en









U





co
CM










O
0



UJ
^
pjH
a
CO
5


CO
r-
CM
(0




CM

CO
en
00







CM

a
0
-i







CO
o

d





a


X
5
_i
oa
METHOD


CO
r
CM
(0




CO

ncnincnincncn
CMinoooinmiooo
cor~uimiDior-u3







CM CM

U O 0 O 0 U Q
o o




in r~ CM o
 co in r-e*-
in v- o m f in

CO 00
o co o
in - CM

o d o





o o 32
OO.OZOOOZ


)g )L
UJ < IU 9
UJ ^  1 ^ UJ _l
^ w oa i x oa
M _i a. _i a M
aocoaocoaa
: z  ,1 z 5 : g
oo^^iuOi^ouj
-IUXZUX-IZ


inr-OcM(Ooocoin
oooocncncnenoo
CMCMCMCMCMCMCOCO
(O(0(D(D(0(0
-------
















<
1- I/I
21
tt
UJ 3
-liL




g2

a -
U CM CM
X "CM
X " ^

00 CM
- ^ II
r- o
-co o
- 
co n n
- a
"fe
a IM
n C
18 >
?+> V
i/) a)
3 L

U IS
v
V
c
n

M

C
IHl

























































t
i/) >
a o
a
0

o *>

a

(0 
H U
IB ...
a c.
VI
u
at



(JTO
52




9^
3a


i.
01
nj M.
v 
1

Q


Of
I

X
O L
4-> at
(Q A
sj
IQ
-1
U

1 Q
0 0
_J -I -1






in to in CM en co
to en CM CM co oo
co en - - CM en




r~ v (0 in  co CM
r* to **-* at o>*
- O n coo
oooooo oo




O
SaaoooeeoQ
o.a.a.za.i-ia.z


< u < ui <
-1 IU ^ -1 UJ  _l
a ^ H ca ^ >-> OQ
w a. -i _/ 1-1 a.
aa.i/)oooa.i/ia
Sc/i a o of 1/1 o
.-*?*
3i883
^
"

-WCO(OUl(0'-C*)00
iflioinia K Q K 1/1 Q
,  5   3 * ?
3S^83SI


inr>OcMiooocoin
eoooenenenoioo
CMCMCMCMCMCMCOCO
(0(0(0(0(0(0(010

cococo^^v^fin


co ^ in
r- on *-
ID r^ r^
















r> CM
CM in CM
- o co
^



- r oo
SCM 00
in oo
o o o





a o o
a. a. a.


in
^ UJ
M X.
_i a. M
1/1 a.
V)
U X -1


f- CM 00
O - "
CO CO CO
ID ID ID

in in in


A-59

-------



















<
 o
-co o
ID * -
+*
* ii ra

CO B
- L. c
CM (0 O
u C *"
ra -H
TJ * 
C I/) 0)
Oc a
U ra
*
!
c
(Q
3
ra
c
Q)
c






























































X
01

O 0
hH U
at
ox
o
Q "~
M ra
*
c
o
(0 -H
H O
(0 -
a L.
+*
i/i
V
Of


00,
32



OJ
32




0)
ra *
a ra
3


a


01
I
n
1/1

x
0 (.
** 9
ra n
A ^
r

O

00

in o) co r-
v 01 in in
oo uJ r- r-


CO
CM
O



CM

O H
0






ID
tO CM 0)
O (0 -
co en


00 (0 00
r^ en oo
CM - *
odd




in
a a o 
z a. a. -


jg
3 iu
-1 IU ^
m \/ fc^
M a. _)
a a u) o
o i/) ac
K s 3 z
iu O M a
Z J I O


^- CM CO (D
in in in in
CM CM CM CM
ID ID ID ID



^ r. f. r-


0
r-
r-








CM

a
o











CO
in
o
d





z


^
5
i
oa

METHOD


in
ID
CM
(O



CM


-
en
oo


















CO
en
,_)



01
ID
6





2






_l
CONTRO




CM
(0



CM


- *- co  en  in
*- - en CM in oo o
co oi co oo r- in in








CM CM

a u u u a
o o






in ID oi r- o
d oi i-  
CM en co - oi


Sis s
odd d





0 0 D  2  Q
a a. z o o o z


 ^
IU  UJ 3
IU ^  1 IU X -1
X -> 00 * 11 00
Ha M _i a
a i/i a a o i/> Q
I/) O I/I K O
X O H 3 Z J3 t-
O *H IU O O H IU
-I I Z -1 O I Z


*- co oo in i~ o CM
i^ r. is. co oo en en
CM CM CM CM CM CM CM
ID 
O
CO
(0



in



en
r










U








en










O
o



Ul
^
H
a
1/1
i
1-4
I


CN*
^
CO
10



in


in
r-










u







t^
CM
CO









o
^
o




IU
V
H
ft
O


00
^
CO
(0



in

A-60

-------




















<
si
at
u D
_l U.
li
CO <

82
t/\ ^f
VI ^
1- O
< M
5
Q
f- IU
00 I-
5- <
u. Z
u. Z
O O
_l
H DC
1/1 O
_j
































u.
Q
U

^
|
00


U. CO
U CM CM
X -CM
I i *
en CM
-  n
CO U
-co o
.'~ ^
co n n
C.-TJ*
~&g
Q H
ii C
 -H
TJ +* fl)
C CO O)
OCR
u n
*>
+<
c
ra

|
OJ
c





































































X
L.
Ill
(/) >
O O
*H U
U
tf
O
i^
IH n
K

IS 
M U
ra -
Q t-

VI
4)
or

U 9)

32


0^
32
^^f

L
0)

n <*
19-
u (0
a


Q


V
1


X
L
5fe
n 2

II
n

U

A
00

in
9
oo








CM

Q
0










CO
CO
o
0






s


^

J
CO

a

t-
UJ
s




in
CM
10




^



en co r- o
en in in r-









CM CM

a a
0 0






en
* CO
CM O


CO CO CO
- co * in
CM in in o
o o o o






o o a Q
a. a. z z


^
IU <
IU ^ _1
^ i-l 00
M a -i
Si "  i
.55?
0 H 0 UJ




CM co (0 in
in in in ID
CM CM CM CM
(0 (0 (0 (O




--- CM



<-'-'-co'O>*incninencn
cn^t-encMincooinin(Oco
ooooenoocot^ininioioc^u)








CM CM CM CM N CM
"""S, "^ **V. "N^ ^"Si "*^
Q Q Q Q Q Q
o o o o o o
_l -I _) -I _J _I





r o CM *r co ^
oo co O ** r oo
v- en CM r- en 


CMC9'-* -CM CO CMin
-CM-incOOI^CMt~COtB
inioinocoio(0^enr~cM
OOOO-OOOO'-OO






Sa?s2So?s:


^ ^ 
IU < IU $ IU <
U ^ J 1U ^ ^1 ^ U J
^MCO^ MOO l-l ^ 03
-IHQ. w _i a. _i Q. i-i
Sai/iaaoMOovtaa
co a to a: 9 a coo
z*3t-az5i-z5xt-
SOt^uJOOnluo^iOUJ
-jzz-iuzaeuz-iz




S^-cocoinr-ocMiooocoin
r^r^r^oocooncncncnoo
CMCMCMCMCMCMCMCMWCMCOCO
(0(DCO(0(OCO(0(DIO(OIO(D




cMCMCMConcoco<#in



CO
r-
(O








CM

Q
O







--


CO
r~
O






s






_l
CONTRO




o
CO
ID




m



*
en
t-


















^
en


in
CO
(0
O






0
a



IU
^^
i^
a.
CO
o
n
z




CM
CO
(0




in



in
^
r


















t



in
,_






o
Q.




IU

^^
a.
1/1
1




00
CO
(0




in


A-61

-------

















1- 10
< z
a <
at
IU 3
_l U.
5 
co 3
z
(/) X
t- 0
4 t"4
T Q
2
a
r- LU
co t-
2:1
u. a
oo
5
t- a
CO O
j

























a o
a o
o u
X X
X X
1 1
co co
r- t-

to to

CO M
CM CM CM
- -CM
00 CM
r-"o "
-eo o
H
n ii ra
 at
CM 13
- L. c
'-WO
nil
n C
(9 *
TJ -H CJ
C CO O)
o c ro
ao-
a *
u *^
u 


c
n
3
O
s
L
c































































L
(U
 >
O O
HH U
a
o
0
c
0
(8 ?
H U
IS -
a (.
<^J
U)
0)
ae


. ^ ^
y oi
C *>Si*
O O)
u a


0)
Q "*s.
Ofli
Wl
-i a
>^-



t.
IS *
n 
o n
a


o
M

0)
1
CO

L

(Q J3
i. B
m

U

ra
co
05 en  oo to
 to in oo in






CM CM

Q O
0 0










in in 
en ^ co
t-  '-

o oo - r- in
oo cn co oo p*
CM CM * 10 CO
O O O O O






a o o o a
z a. a a. z

 
 IU <
j |U ^ 	 |
CD ^ >-l CO
w O. _l
a a. co o a
O CO Of O
IU O l^ O IU
Z -J I 0 Z


- CM co to in
in in in in to
CM CM CM CM CM
to to to to to



f- i- f -^ CM


*cM**incnM*toes-oin<
r~ui'-ooooiniflcnTf^r-coen*
to





CM CM CM

a o o
O 0 0










or-o CB-- to in in <>
**o r-incM cMt~t^ COCM
-'- -'-'- '-co

 en r- en * -
^looinn cn o> m i- *r to *t
^encoi**cn*'oooo^cooocn^tco
i-OOO'-'-OO'-'-OO'-'-






aooooooooooooo
aQ.Q.ZQ.Q.O.ZQ.Q.Q.ZQ.a.

 ^ ^
lu 3 iu < iu < ui
lU^^IU ^ -J ^IU_I it
^ 11 CO X MCO M^CO M
-IMO. M_IQ. _IQ.I-I _i a
go.coaa.ocoaocoa.aoi/)
co Q in at ? a ^ ? a T
Z5t-Sz5t-ZU3l-ZC3
OOi-ilUOOt-iliJOi-iOUJOt-"
U-IZS-IUZZUX-IZUZ


to^-nooint^ocMiooocoinc^cM
tDt^r-t~coooo)cno)cnOOO'-
CMCMCMCMCMCMCMCMCMCMCICOCOCO
(OtO(OtO(OtO(0(0(OtOtOfOtOtO



cMCMCMOncoco**nininin


(0
to
^



















^
*"

en
in
,_






O
a.



IU
^
n
a.
1/1
o


oo
m
to



in


A-62

-------



















<5
a <
at
-1 U.
a.
3 a
z
co <

85
(- O
< l->
x a
a
r- ui
oo t-
2
^ Q
o o
_l
M
H Of
CO O
M U_
_i


























a
a
u
X
X

CO
"-.
(O

a co
a -
U CM CM
?7~
0) 
- ^- tt
"-So
p. - 
+<
CO II 
"Si
u?M
W'H
5 ** 9L
C I/I 0)
3 L
O C 19

C

I

L.
0)
c
M


























































>

l/ >
a o
a
0

0 v
M (Q
a

r*in
CM^CO CMU) -CM*


 V h-
w 01 n oo -" x
_l w Q. M_IO. -10.11
a.i/iaaoi/iaoi/ia.
1/1 g M oc o ae 1/1
zxi5i-xz(5i-z5s
OO*iuiOO*iUIO*iO
U-JIZ-IOIZOXJ




(O^-coooinr-ocMCoooco
cor-r-r-oooocncneneno
CMCMCMCMCMCMCMCMPMCMCO
lOCOIOCOCDCDCOCOCOCOCO


CMCMCMCOCOCOCO*'



0
CO
p-







CM
a
o












en
o






^
5
_i
CO
METHOD




in
o
CO
(0


in



in
en
(D

















v-
-CM
CM


en
v-
^



a
a.






CONTROL




p-
0
CO
co


in



^
q.
00


















in
CM


00
CO




0



Ul

n
I/I

1-4
X




CM
CO
CD


in



CO
CO
p-

















CM
en
CO


CD




2




Ul
V
a.
3




co
CO
CO


in


A-63

-------



















(-
a 2
IX
UI 3
_l U.
za
t/l <

U M
az

1 O
< M
T* Q
9
a
P- UI
oo t-
u. Z
M
u. at
oo
-i

H4
i- a
I/) O
w U.
























u.
Q
O
a
i
oo
-
p*
(O
LL V
a -
(J CO
Z M CM
i -O
oo - 
P-  co
p- -
O O






2 



^
ui 5

n CD
v) a
s ^
M UI
X Z


O CM
e> en
CM CM
ID ID


CO ^



w en  co ^ in (o
in to en in CM co CM
(O (0 p- P- p- co P"








CM
*^
a
o
"*





- in - "
o in co -- - P- o
co oo in co in (0



CM en ID CM in
in co co ^ oo 10 ID
en o oo in (o (0 CM
o - o O o O -






20 o Q a o a
a. a. z a. a. a.



x
UI 5 UI
^ UI -I  UI
11 * 00 11 *
_i a. w _i a. M
O I/) O. O O V) 0-
K u> g at 1/1
z 15 s i- z 5 *
O n o UI O Hi O
0 I -1 Z 0 I -1


10 00 CO UI P- CM 00
en en o O o  *-
CM 
-------









u.
Q
U
Q.
X
1
oo

K - < ll C
u- Z ra -H
M TJ V 0)
u. a c 1/1 5)
o o 3 t-
-1 0 C 
11 O H
i- a u ra
1/1 o -M
_i 
c
ns

15
L.
01
c
11






























































L
41
I/I >
O O
M U
01
at
0
c
o
 V
H U
Si
44
in
01
a

(j!^n
82



^-^
91
Q ^s.
On
wi
-i a



L
m <*.
a 15
S



a


0)



X
L
0 t-
v 01
n n
(- e

U

(fl
CO
CM * co o in
co ^ ^ oo co
Is- r* r^ P* (D






CM CM
^ V,
a Q
0 0





V ID
* CO i-
CO CM


in co in en
I** in ^ (D t**
*- in v- en CM
O O - O O





2221



^ ^
< IU <
_l UJ * _l
CO ^ 11 CD
M 0. _l
a a. 1/1 o a
i s s 1 i




*- CM co (o in
in in in in ID
CM CM CM CM CM
(D (0 (D (0 (0


,-,-,-,- CM


(0 M r-
,- ,J. ,-
r- co co






CN
""^
a
o
^




(O
en r-
CM CO


in
i- 00 ID
en co oo
O - "





22




IU
IU ^

-1 M CL
o a. i/i
DC I/I
o o u
U -1 X




S^ CO
r- r>
CM CM CM
(0 ID ID


CM CM CM


i^ooooajfcn^co^ina)
*CM*cor-in(oenincMCOcM
(DoocnooiOfOCDr^r^r^oor^






CM CM CM CM CM
x^ v^, ^^ ^^ "^
o a a a a
0 O O 00
_j _j -j -i _j




CM CM CM ID
oo-co in co --CMOS
CM CO  CM i- CM


CM CM *- * *
n m in n eo ID to oo t- in ^
incM*-'-coweMrenenco
OCMO'-O'-'-'-OOO'-





2?2i2222



^  ^
^t u 5 u 5 u
_JIU X -I ^IU_I ^ UJ
oa^ w 03 n^oa M^:
n _i a. _i Q. I -10.11
aoLOi/iaot/io-Qoi/io.
o i/> at o ot 1/1 o at 1/1
IszSl^S^ll^S^
UJOOnllJOi-iOllJOi-iO
s-iOxxux-isux-i




ooint~OcMioaocoint^cMoo
r>ooooenenenenOOO-'-
CMCMCMCMCMCNCMCOCOOCOCO
(0(OIO(0(O(OIOCO(OIOIO(0


coconco5r
-------
















1-  <
u.Z
n
u. a
00
M
t- a
I/I O
M U.
_l


























a
8
1

oo
P.'
(0
a V
a
O CO
a -
Z CM CM
i  o
00 T- .
 1 f"
f~ CM
- ^ II
(D L)
-co o
+*
co ii m
- at
CM "O
- (- C
- 19 O
II ?*'
19 +
TJ -H 01
"?
O C 19
a o t-
O H
U <9
+*
+*
c
19
O
c
l_
01
c
n










C
0)
I/I >
0 0
*i U
1
0
C '^
0 *
11 <9
or
c
o
(9 4*
M -
a L
VI
01
a

O O)

O O)

N^




^^
O)
a \
On
wl
-i a




^
.2
19 " CM
(O CM U) (0
o o o o







o o a
a. a. a.

< Ul
-1 Ul ^
OB ^ i-l
i_i ^ j
O Q. VI O
g V, X
H  O Z
uiOnO
z -r z u



" CM CO ID
in in in in
CM CM CM CM
(O (D ID ID




T-  ^ **


ID
r-








CM
0
o













ID
t*
O
O







i

i

m
METHOD



in
ID
CM
(D




CM


too
10 co o
O) _ 


















n oo t
 (D in
<- - CM




r* r
- co 
CM - -







a a a
a. a. a.

Ul
Ul ^
^ n
-i 11 a
So. to
i/i
  IS
O O M
0 -1 Z



  CO
r- r-
CM CM CM
ID ID ID




CM CM CM


(0
(0
r-








CM
a
o













00
o
CM
o









1
_l
CO
Ul



oo
CM
(0




CO


o
t-


















o>
^
^




*
(D
^







o
a


u
V
n
a
I/I
3



in
00
CM
ID




CO


0
CO


















in
CM
^



en
in

O







o
a.




CONTROL



r
CO
CM
ID




CO


O 0
co r
CO CO








CM
a
3







^
(0
CM



en
CM CM
(D CM
o o







o a
a. Z

Ul <
X -1
M CD
i/i a
M Ul
Z Z



O CM
en en
CM CM
10 ffl




CO 


- o o o o
en oo CM co *
en  * - o








CM
a
0







O in co -co
* CM ID CO
* CM ^ ^



CM en
in r- CD r*
* r* oo - oo
" O O " O







a a o a o
a a. a. z a.

Ul <
^ u _i
w ^ 03
-i a M -i
1/1 a. a o
oo o cc
Z 5  t- z
O M O Ul O
0 Z -I Z 0



(o oo co in t^
en en O O O
CM CM CO CO CO
ID ID ID ID ID




*t ^ ^f in in


0 0
O en
CM O

















.
en ID

CM ^



r r-
01 en
r en
O O







a a
a. a.

Ul
X Ul
M 
a M
vi a.

-------















H VI
o 3
at
UJ 3
_l U.
Q.
3
i/i <

O I/I
OZ
M
(/I X
t- 0
< 11
i
o
!. m
00 h-
> <
u. Z
TING OF
R CHLORI
I/) O
M (^
_J



































^
00
O 
i O
CM
CO 0
*l
II in
U. 0
a TJ
o c. c
o m o
,,SM
 -f
TJ + 9
C 10 O)
3 L.
o c n
to t-
I
ra
+4
>^
v
c

Qf
c
0)

M










to >
28
8
DC
2
at
c
o
 -H
H O
j 
a L.
^
VI
01

O ^)
2 v^
0 01
u a




o nt
32
,^f


01
(Q *4
^ "
Q.


O


g-

1/1


X

o c.
+* u
n a
c. E
?3
"

U

ra
aa
O CM
oo co
oo oo


8



CM

a M
0
_j






^
in



CM
CO
in
d




o
 2
V
2
^
_1 U
oa ^
o a
a -i a. M _i
SO. I/I Q Q. O
I/I Q I/I Of
$ * 5 f * z
8 3  31 3 8




10 i- co oo in t-
CM CM CM CM CM CM
ID (D (D (D (D CO


CM CM CM CO CO CO


O
O
w







a:








CM
CM
CM


CO
in
in
O





S


u

IH
a
3
1-4




O
en
CM
(0


CO


0
en
in








at













CO

CO
o






y
2
^
-J
aa
METHOD




CM
en
CM
(D


^


o
r-















co

^





^.





2





CONTROL




(D
en
CM
(0


^f


0 0
in to
in *








at






co

00 CO
(O CD
CM


CO ID
0 0
_ ,_





o a
Q. 0.


UJ
* UJ

85 ft
$ *
x 3




00 CO
en o
CM CO
(D (0


* *


0 0
oo -
CM CO






CM
^
Q
O
_l



en
CO

CM




(O
* (O
CO O





 85
v
z
^
_J
ffl
METHOD
CONTROL




in r>.
O 0
CO CO
(D (O


in in


o o
in CM
ID (0








CH O






t^

r~ CM
O *
CM


*- 
in co
7. ,_





O O
& a.


Ul
^/ Ul
M ^
i/i a.
1/1
 >
"-" O
X -1




CM 00
^ ^
CO CO
(D (O


in in


A-67

-------
















<
H- I/)
< 2
Q <
Ot
IU 3
_1 u.
l|
i/l <

(J I/I
OZ
CO X
t- o

Q
r- uj
oo 1-
> <
u.2
u. at
S5
i- a
I/I O
j





































,,
U 00
O 
i O
CM
*- II
o
co o
+J
H ra
T3
C. C
o ra o
fan
II C

E-H 0
I/I 0)
o c ra
Q. O t-
u ra
+>

c
ra

ra
c.
0)
c






























































>>
01
00 >
O 0
0)
at
o
c -
0 -H
M (Q
DC
c
o
fl V
M O
ra -
a L
VI


. -.,
00,
0 O)
M-


Q o>
o o,
-i a
*

0)
ra  v
CM in o> in *- *



* * in co r> in
* r - o oo r
co CM in  co -



O
a o a a a o a
0* 0* CL Z CL fl. w

^
UJ 3 IU
U ^ J U ^
^ M at x. M
-i n a. M _i a
o a. co o a. o i/i
a 1/1 a to at
z  5 t- * z S
O O M Ul O O w
0 -i X X -1 0 X




to - co co in r- o
03 t r r co oo en
CM CM CM CM CM CM CM




CM CM CM CO CO CO CO


O
Ol
in











at










CO
en
en
d




a
z

^

_i
00
METHOD




CM
en
CM




^


O
r
^

















d
CO




00
in




O






CONTROL




(0
en
CM
(0



^


O
in
in











at





6
0)
r-



00
00
^




a



IU
^
n
Q.
(/I
5
M
z




CO
en
CM
10



^


O
ID
V

















o
CM
in



en
en
ID




a




IU

a!
1/1
o




CO
o
CO
(0






0 0
co *-
CM CO











U.


oo
CM

CO O
en
*-



co
CO



in
CM
a o


^
3
j
00
METHOD
CONTROL




S S
C*> CO
(D 03



in in


0 0
in CM
(0 (O











ac at





d d
in t
(O *



CM CM
co to





O 0



IU
 IU
M *
a. M
oo a.
l/I
 s
Z -1




CM CO
CO CO
03 03



in in


A-68

-------



















H I/)
23
a
Ul 3
-1 u.

o
2
I/) <


O Z
tH
I/I X
< 2
I S
a
fx m
cot-
2:2
si
85
M
i- a
I/) O
M Ik





























X
Z
1
00
r
(0
a m
o -
OWN
X -CX
z - 

a tv
-^ H
r- O
-o o
f
- at
N- L c
*-  0
DM
ii C
?** 9L
^1 0)
o c m
t^
u m


c
a

^
A
E
01
C




























































X
t_
01
a o
<-i U
0)
a
0
c -
w m
a
S
H U
ra 
a c.
M
0)
a


o
82






O)
O--S
32


C
0)
4 "*
W -
o m
a


0


0)
0
14



X
o t.
a a

-i .
c
u
m

cn^^cH^incocN^in
.
rSr-r--t~air-St--t---(O
















(O (0
o <- o o NO

q- q- <-(ODOMCx
^ ^ 
-------




















t- CO
< z.
a 2
CL
UJ 3
-I U.
n
i a
(/> <

U I/I
2
cox
t- o

C Q
2
a
r- uj
00 H
u. Z
M
U. K
OO
-1
25
I- K
CO O
w u.
































U
X
X
1
00
f"
(0*
a 
U CM CM

f 7 ^
OD N
- ^* M
* co o
<0.'" J
M II UJ
K
(STI
- (. C
* 
0 0
at
o
c -
O 4-
>-l fl
a

fl 4^
4- O
M -
a L.

VI
01
at


ua,
U Q.
N_^



"""
O \
32
>.



.
01
ID (or-r-(D

















^(vi^^ls.'codeoCDtfl
uicocowo^w^^w
*"C^^^"N^ (/> (/> a
^3zs3z0<3S2
QMOQMQMMOQ
-IXU-JI-IXI-JU



SCO O Is-
O) O O
C4C4C4tSCSCSCSC4CO(*)
(0(0(0(0(0(0(0(0(0(0

^^{NC^C^COO^^lfl


A-70

-------
     APPENDIX B
ANALYTICAL PROTOCOL
         B-l

-------
1.0  SAMPLE EXTRACTION

     1.1  Extraction of Adipose Tissue

          Addition  of  internal  quantitation  standards    Allow  the adipose
tissue composite to reach room temperature and then add the carbon-13 internal
quantitation spiking solution such that it delivers 500 to 2,500 pg of each of
the l3C-labeled surrogates.

          Add  10  ml of  methylene  chloride  and  homogenize  the  mixture for
approximately 1 min with a Tekmar Tissuemizer.

          Allow the mixture  to  separate  and  decant  the methylene  chloride
extract from  the  residual  solid material  using a  disposable pipette.   The
methylene  chloride is  eluted  through  a  filter  funnel  containing a  plug of
clean glass wool and 5 to 10 g of anhydrous sodium sulfate.  The dried extract
is collected in a 100-mL volumetric flask.

          Add to  the  sample a second 10-mL aliquot  of methylene chloride and
homogenized for 1 min.  The methylene  chloride is decanted,  dried, and trans-
ferred to the 100-mL volumetric flask.

          Rinse culture tube  with   at  least  two additional  aliquots  (10 mL
each) of  methylene chloride,  and transfer  the entire contents  to the filter
funnel containing  the  anhydrous  sodium sulfate.   The filter  funnel  and con-
tents are rinsed with  additional  methylene  chloride  (20  to 40 mL).  The  total
eluant from  the filter  funnel  is collected  in the 100-mL  volumetric flask.
Discard the sodium sulfate.

          Adjust the final  volume of the  extract  for each sample to 100 mL in
the volumetric flask using methylene chloride.

     1.2  Lipid Determination

          Preweigh a clean  1-dram glass vial  to the  nearest  0.0001 g using an
analytical balance tared to zero.

          Accurately  transfer 1.0 mL  of  the  final  extract  (100 mL) to the
1-dram vial.  Reduce the volume of methylene chloride from the extract using a
water bath  (50-60C)  gentle stream  of purified nitrogen  until  an oil residue
remai ns.

          Accurately weigh  the 1-dram vial  and  residue to the nearest 0.001 g
and  calculate  the weight  of  lipid  present in  the  vial  based  on difference.
Nitrogen blowdown is continued until a constant weight is achieved.

          Calculate the  percent  lipid content  of the  original sample  to the
nearest 0.1/K as shown in Equation B-l.
                                      B-3

-------
                   Lipid  content,  LC  (%)  = ULR v ,.EXT x 100*            Eq.  B-l
                                           WAT x VAL

where:    WLR  = weight of the lipid residue to the nearest 0.0001 g;

          VEXT = total volume of the extract in mi"11 iliters;

          WAT  = weight  of  the  original  adipose  tissue  composite  to  the
                 nearest 0.01 g; and

          VAT  = volume of the aliquot of the final extract in ml used  for  the
                 quantitative measure of the lipid residue  (1.0 ml).

     1.3  Extract Concentration

          Quantitatively transfer  the  remaining extract volume (99.0 ml) to  a
500-mL Erlenmeyer flask.  Rinse the volumetric  flask  with 20 to 30 ml of  addi-
tional methylene chloride to ensure quantitative transfer.

          Concentrate the extract  to an oily residue  using  rotary evaporation.


2.0  CLEANUP PROCEDURES

     2.1  Bulk Lipid Removal

          Add a total of 200 ml of n-hexane to  the spiked  lipid residue in  the
500-mL Erlenmeyer flask.  Slowly add, with stirring,  100 g  of the 40% w/w sul-
furic acid  impregnated silica gel.  Stir with a magnetic stir-plate for 2 h.

          Allow  solids  to  settle  and decant  liquid through a  powder funnel
containing  20  g of  anhydrous  sodium sulfate  and  collect  in  a  500-mL sample
bottle.   Rinse  solids with  two 50-mL portions of hexane.   Stir each rinse  for
15 min, decant, and  dry  by  elution through sodium sulfate  combining the hexane
extracts.

          After  the rinses  have  gone through  the sodium  sulfate,  rinse  the
sodium sulfate  with  an  additional  25 mL of hexane and combine with the hexane
extracts.

          Prepare  an acidic silica  and  a neutral  alumina  column as follows:
Pack  a  1 cm x  10 cm chromatographic  column  with  a glass  wool  plug,  add
approximately  25 mL  of  hexane,  add 1.0 g of silica gel, and then add 4.0 q of
40%  w/w  suIfuric acid impregnated silica  gel  and  allow to settle.  Elute  the
excess hexane  from the  column until  the solvent level reaches the top of  the
chromatographic  packing.  Verify  that the column does not  contain any  bubbles
or channels.   Pack  a second chromatographic column  (1 cm x  30 cm) with  a  glass
wool  plug,  add 4 g  of sodium  sulfate,  add 4.0 g of  neutral alumina and  allow
to  settle,  and then top with a 4-g layer of  sodium  sulfate.  Elute the column
with  10 mL  of  hexane until  the solvent level reaches  the  top of  the chromato-
graphic packing.   Inspect the column to ensure it  is free  of channels  and  air
bubbles.

                                      B-4

-------
          Quantitatively transfer the hexane extract from the Erlenmeyer  flask
to the  silica gel  column  reservoir.   Allow the hexane  extract to percolate
through the column and collect in a KD concentrator.

          Complete the elution of the  extract  from the silica gel column with
50 ml of  hexane  in the  KD  concentrator.  Concentrate  the  eluate to approxi-
mately 1.0 ml, using nitrogen blowdown as necessary.

          Note:   If the  40ft sulfuric  acid/silica  gel  in noted  to  be highly
discolored  throughout  the  length  of  the adsorbent  bed,  it is  necessary  to
repeat  the  cleaning procedure  beginning with the  acidic  silica  gel  slurry
procedure.

     2.2  Separation of Chemical Interferences

          Transfer  the  concentrate  (1.0 ml)  to  the  top  of  the  alumina
column.   Rinse the  K-0  concentrator with  two 1.0-mL portions  of  hexane and
transfer  the  rinses to  the  top  of  the alumina  column.   Elute  the alumina
column with 10 ml of 8% (v/v) methylene chloride  in hexane until the solvent
level  is  just below the  top of the sodium  sulfate.   Archive  this eluate.
Columns must  not  be allowed to reach  dryness  (i.e.,  a solvent  "head" must  be
maintained).

          Elute  the column  with 15  ml of  60% (v/v)  methylene chloride   in
hexane and  collect  this  fraction containing  the dioxins and furans in a  50-mL
culture tube.   Concentrate the fraction  to a volume of approximately 2 ml.

          Prepare  an AX-21/silica  gel  mixture by  thoroughly  mixing  1  g   of
AX-21  (100/325 mesh)  and 19  g  of  silica gel  in a  40-mL vial.   Activate  at
130C for 6 h.   Store  in a desiccator.  Cut off a clean 5-mL disposable  glass
pipet  (6  to 7 mm ID) at the 4-mL mark.   Insert a  plug  of glass wool  and push
to the 2-mL mark.   Add  1 g of the  activated AX-21/silica gel mixture followed
by another  glass  wool  plug.   Using  two glass rods, push both glass wool  plugs
simultaneously towards  the AX-21/silica gel  mixture and gently compress the
AX-21/silica  gel  plug  to a length of  4  cm.   Preelute the column with 4  ml  of
toluene followed  by 2 mL of  75:20:5  methylene chloride/methanol/benzene, and
4 ml of  1:1 cyclohexane  in methylene chloride.  The flow rate  should be less
than 0.5 mL/min.   While  the column is  still  wet,  add the concentrated eluate
from the  alumina  column  to  the  top  of the AX-21/silica gel column.  Rinse the
culture tube which contained the extract twice with 1 ml of hexane and add the
rinsates  to the  top of  the column.   Elute  the column  sequentially  with two
0.5-mL aliquots of hexane, 10 ml of 1:1  cyclohexane in methylene chloride, and
5 ml of 75:20:5  methylene  chloride/methanol/benzene.   Combine and archive the
first  three eluates.   Turn  the column upside  down  and elute  the  PCDO/PCDF
fraction with 20 ml of toluene into a 6-dram vial.

          Using  a stream of  nitrogen,  reduce  the toluene  volume  to  approxi-
mately  1  ml.   Carefully transfer  the  concentrate  into  a 1-mL  minivial and
reduce the volume to about 200 yL using  a stream of nitrogen.

          Rinse  the  concentrator tube with three washings  using  500  uL  of  1%
toluene in methylene chloride.  Concentrate to 200-500 uL and add 10 uL of the


                                      B-5

-------
tridecane  solution  containing the  internal  recovery  standard  and  store  the
sample in a refrigerator until HRGC/MS analysis.

          Prior  to  analysis,  using  a  gentle  stream  of  nitrogen  at  room
temperature, remove toluene and methylene  chloride.   Submit sample to HRGC/MS
once a stable 10-uL volume of tridecane is attained.

     2.3  HRGC/HRMS Analysis for PCDD/PCDF

          Once  initial  and routine calibration criteria  are met, the instru-
ment is ready for sample analysis.  Prior to the first sample, an injection of
tridecane will be analyzed to document system cleanliness.   If any evidence of
system  contamination  is found,  corrective action must  be  taken  and another
tridecane blank analyzed.

          Note:   Syringe  Technique    Congeners of PCDD/PCDF can carry-over
between  injections  in  the  syringes used for  HRGC/MS  analysis unless  the
syringes  are  properly cleaned between  samples.   The  following  procedure  has
been  found to  be  very  effective  for  PCDD/PCDF  removal  from contaminated
syringes and will be used  throughout these analyses.

          Rinse the syringe 10 times with isooctane.

         Fill  the  syringe with  toluene  and  sonicate syringe  and  plunger in
          toluene for 5 min and repeat at least twice.

         Rinse the  syringe  10 times  with  tridecane  and  pull up 1 uL  of clean
          tridecane.

          Syringe is ready for use.

At  no  time  should  air be introduced into the HRGC column  by using an  air plug
in  the syringe.  The  oxygen present in  the  air plug will  quickly degrade  a
nonbonded GC phase.

           Inject a 1-yL aliquot of the extract into the GC,  operated under the
conditions  previously  used to produce acceptable results with the performance
check  solution.

          Acquire SIM  data according  to the same acquisition and MS operating
conditions  previously used to determine the relative response  factors.

           Instrument  performance  will be monitored by examining and recording
the peak areas for  the recovery standard, 13C12-1,2,3,4-TCDD.   If this area
should decrease to  less than 50% of the calibration standard, sample  analyses
will be stopped until  the  problem is  found and corrected.

     2.4  HRGC/MS Analysis for PBDD/PBDF

           Procedures  for  the  analysis  of   the  brominated species  will be
similar to  the procedures  for  the  PCDD/PCDF  analysis,  with  the following
exceptions:


                                      B-6

-------
     1.   There is no column performance window-defining mix available.

     2.   Sensitivity and daily RRF check will be established by injecting the
          CSS level calibration standard.

          SIM  data will  be  acquired  using  the  same  chromatographic  and MS
operating conditions used to determine the relative response factors.   Instru-
ment performance will  be monitored by examining  and recording  the peak areas
for  the recovery  standard,  13C12-l,2,3,4,7,8-HxCDD.    If  this  area should
decrease to  less  than  50% of the  daily  calibration standard, sample analysis
will be stopped until the problem  is found and corrected.


3.0  CALCULATIONS

          In this  section, the  procedures  for the data reduction are outlined
for  the analysis  of data from the  HRGC/HRMS method  for  PCDD/PCDF  and  the
HRGC/MS method  for PBOD/PBOF.   Figure 8-1 presents a  schematic  of the qual-
itative criteria for identifying PCDDs and PCDFs.  Identical  calculations and
qualitative criteria will be  applied  to  the  data generated  for the brominated
analogs of the dioxins and furans.

     3.1  Qualitative Identification

          The  ion  current responses  for each mass  for a  particular  analyte
must be within 1 s to  attain  positive  identification of  that  analyte.   For
example, m/z 338 and m/z 340  must  have maximum peak responses that are within
1 s to be positively identified as a pentachlorodibenzofuran.

          The  ion  current intensities  for  a  particular PCDD/PCDF or PBDO/PBDF
must be > 2.5 times the  noise level (S/N > 2.5) for positive identification of
that isomer.  The  integrated ion current ratios of the analytical masses for a
particular PCDD/PCOF must fall within the ranges shown in Tables B-l and B-2.

     3.2  Quantitative Calculations

          Relative  response  factors for native  PCDD and PCDF  analytes (RRF)
are  calculated from the data obtained  during the analysis  of  concentration
calibration solutions using the following formula:



                               RRF =  STD '  IS                         Eq. B-2
                                     AIS * USTD

where:    ASTD = the sum of the areas of the integrated ion  abundances  for the
                 analyte  in  question.    For  example,  for TCDD, A^jn  would be
                 the sum of the integrated ion abundances for m/z J2o and 322;
                                      B-7

-------
           HRGC/MS-SIM Data
Quantitate Specific Isomer as per Protocol
               Response to
         Characteristic Molecular
       Ions within the Appropriate
           Homolog Retention
               Window?"-
             Characteristic
         Ion Ratios within 20%
              Theoretical ?
                Response
         Corresponds-to Specific
            Isomer Retention
                 Time?
     Report Compounds cs
      Not Detected (ND)
     Calculate Sample LOP
 Response Due to
 Coextracted Interference
Quantitate Compound
as Per Protocol
Report as Isomer Unknown
  Figure B-l.  Qualitative criteria for identifying PCDOs  and  PCDFs.
                                  B-8

-------
Table B-l.  Ion Ratios for HRGC/HRMS Analysis of PCDD/PCDF
Compound
TCDF
i3Ci2-TCDF
TCDD
i3C12-TCDD
PeCDF
*3C12-PeCDF
PeCDO
i3C12-PeCDO
HxCDF
i3Cl2-HxCDF
HxCDO
i3C12-HxCDF
HpCDF
i3Cl2-HpCDF
HpCOO
i3Cl2-HpCDO
OCDF
OCDD
i3Cl2-OCDD
Ions monitored
304/306
316/318
320/322
332/334
340/342
352/354
356/358
368/370
374/376
384/386
390/392
402/404
408/410
418/420
424/426
436/438
442/444
458/460
470/472
Theoretical ratio
0.76
0.76
0.76
0.76
1.55
1.55
1.55
1.55
1.22
0.50
1.22
1.22
1.02
0.44
1.02
1.02
0.87
0.87
0.87
Acceptable range
0.61
0.61
0.61
0.61
1.24
1.24
1.24
1.24
0.93
0.40
0.98
0.98
0.82
0.35
0.82
0.82
0.70
0.70
0.70
- 0.91
- 0.91
- 0.91
- 0.91
- 1.86
- 1.86
- 1.86
- 1.86
- 1.46
- 0.60
- 1.46
- 1.46
- 1.22
- 0.53
- 1.22
- 1.22
- 1.04
- 1.04
- 1.04
                           B-9

-------
      Table  B-2.   Ion  Ratios for HRGC/HRMS Analysis  of  PBOD/PBDF
Compound
Ions monitored     Theoretical  ratio     Acceptable range
TBDF
i3Cl2-TBDF
TBOO
i3C12-TBOD
PeBDF
i3C12-PeBOF
PeBOO
HxBOF
HxBOO
482/484
494/496
498/500
510/512
562/564
574/576
578/580
642/644
658/660
0.68
0.68
0.68
0.68
1.02
1.02
1.02
1.35
1.35
0.54 -
0.54 -
0.54 -
0.54 -
0.82 -
0.82 -
0.82 -
1.08 -
1.08 -
0.82
0.82
0.82
0.82
1.22
1.22
1.22
1.62
1.62
                                  B-10

-------
          ATP  a the sum of the area of the  integrated  1on abundances for the
                 labeled PCDO/F used as the internal quantisation standard for
                 the above analyte.  For example,  for i3C12-2,3,7,8-TCDO, ATP
                 would be the sum of the integrated ion  abundance for m/z 332
                 and 334.

                 concentration of the analyte in pg/uL;  and

          Cj$   concentration of the internal quantisation standard in pg/uL.

Table B-3 provides  the  pairing  of target  analytes to  internal  quantitation
standards for determining RRF values for PCDO, PCDF, PBDD, and PBDF compounds.

          Relative  response  factors for the internal quantitation standards
(RRF TP).   The  RRFjs  values are  calculated from data  obtained  during the
analysis of concentration calibration solutions using the following formula.
where Aj^ and Cj^ are defined as given in Eq. B-2 and

          ^RS ~ concentration of the internal recovery standard in pg/wL; and

          ARp * the sum of the areas of  the  integrated  ion abundances for the
                labeled   PCDD   (i3C12-l,2,3,4-TCDD   or   13C12-1, 2,3, 7,8,9-
                HxCDO).  For example, the i3Cl2-l,2,3,4~TCDD, ARS would be the
                sum of the integrated ion abundance for m/z 332 and 334.

The RRF  values for the  13C12-PBDD/PBDF  compounds will  be  calculated  using a
similar equation.  Refer to Table B-l for pairing of the internal quantitation
standards with the appropriate internal recovery standard.

     3.3  Concentrations of Sample Components

          Figure B-2 presents  a  schematic for quantitation of  PCDD and PCDFs
which meet the criteria specific in Section 3.  Calculate the concentration of
PCOO/Fs or PBDD/Fs in sample extracts using the formula:
                                                                       Eq. B.
where:    Csam_ie = the Tipid adjusted concentration  of  PCDD  or PCDF congener
                    in pg/g;

          A-ami_  sum of  the  integrated  ion  abundances determined  for the
              p     PCDD/PCDF in question;
                                     B-ll
                          ample   AJS  RRF  WAJ  LC

-------
    Table B-3.  Target Analyte/Internal Quantitation Standard and Internal
            Quantitation Standard/Internal Recovery Standard Pairs
  Target analyte
                                         Internal standards
      Quantitation
                                 Recovery
Chlorinated
2,3,7,8-TCOO
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCOO
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8-HpCDO
OCOF
OCDD

Brominated
2,3,7,8-TBDD
2,3,7,8-TBDF
1,2,3,7,8-PeBOO
1,2,3,7,8-PeBDF
1,2,3,4,7,8-HxBDO
1,2,3,4,7,8-HxBDF
13C12-
13C12-
-2,3,7,8-TCDD
 2,3,7,8-TCDF
 1,2,3,7,8-PeCDF
 1,2,3,7,8-PeCDF
 1,2,3,7,8-PeCDO
 1,2,3,4,7,8-HxCDF
 1,2,3,4,7,8-HxCDF
 1,2,3,4,7,8-HxCDF
 1,2,3,4,7,8-HxCDF
 1,2,3,6,7,8-HxCDD
 1,2,3,6,7,8-HxCDO
 1,2,3,6,7,8-HxCDO
 1,2,3,4,6,7,8-HpCDF
 1,2,3,4,6,7,8-HpCDF
 1,2,3,4,6,7,8-HpCDD
 OCDD
-OCDO
i3Cl2-2,3,7,8-TBDD
i3C12-2,3,7,8-TBDF
i3Cl2-l,2,3,7,8-PeBOF
3C12-l,2,3,7,8-PeBDF
3Cl2-l,2,3,7,8-PeBOF
i3Cl2-l,2,3,7,8-PeBOF
13C12-
13C12-1,2,3,4-TCDD
i3C12-l,2,3,4-TCDO
i3Cia-l,2,3,4-TCDD
3C12-1,2,3,4-TCDD
i3d 2-l,2,3,4-TCDD
i3C12-l,2,3,7,8,9-HxCDD
    z-l,2,3,7,8,9-HxCDD
  :12-l,2,3,7,8,9-HxCDD
    z-l,2,3,7,8,9-HxCDO
13C12-1,2,3,7,8,9-HxCDO
i3C12-l,2,3,7,8,9-HxCDO
3C12-l,2,3,7,8,9-HxCDD
  :I2-1,2,3,7,8,9-HxCDO
    2-1,2,3,7,8,9-HxCDD
  :12-l,2,3,7,8,9-HxCDD
i3C12-l,2,3,7,8,9-HxCDD
    2-l,2,3,7,8,9-HxCDO
                          i3C12-l,2,3,7,8,9-HxCOO
                          i3Cia-l,2,3,7,8,9-HxCDD
                          i3C12-l,2,3,7,8,9-HxCDD
                          i3C12-l,2,3,7,8,9-HxCDO
                          i3C12-l,2,3,7,8,9-HxCDO
                          i3C12-l,2,3,7,8,9-HxCDO
                                     B-12

-------
          QUANTUM ION
                        Report as Not Detected
                        Calculate Sample LOD
        HRGC/MS-SIMData
              Response
             Meets AI
            Qualitative
             Criteria ?
             Response
                 times
              S/N?
                      Calculate as per Protocol
                      Report as Trace (tr) Value
Quantitate as per Protocol
Report as Positive Quantifiable Value
 Figure  8-2.
Procedure for quantitation of PCDDs and  PCDFs
    in human adipose tissue.
                            B-13

-------
                    sum of the  integrated  ion  abundances determined  for the
                    labeled  PCDO/F used as  the  internal  quantitation standard
                    for the above analyte;

                    the amount (total  pg)  of the labeled internal quantitation
                    standard  added to  the  sample prior to extraction;

              RRF = relative  response factor of  the  above analyte relative to
                    its labeled  internal quantitation standard determined from
                    the initial  triplicate calibration;

              WAT = weight (g) of original adipose tissue sample; and

               LC = percent extractable lipid determined from Eq. 4-2.

Refer  to Table  8-1  for  pairing  of  target analytes with  the  appropriate
internal quantitation standard.

          Quantitative data  will be classified  to indicate the  intensity of
the signal response.   Qualifiers will  include:   not detected,  NO (signal-to-
noise ratio  is  less  than 2.5);  trace, TR (signal-to-noise ratio  is  greater
than or equal to 2.5 but less than 10); and positive quantifiable, PQ (signal-
to-noise ratio is greater than or equal to 10).

          Because  of  the lack  of analytical standards to  demonstrate isomer
specificity of the brominated compound, all PBOD and PBDF responses that elute
at  a  retention time of  a 2,3,7,8-substituted  isoraer will be classified  as a
maximum possible concentration (MPC) for that specific isomer.

          Recovery of internal quantitation standards.  Calculate the recovery
of  the  labeled internal  quantitation  standards measured  in  the final  extract
using the formula:


                 Internal Quant. Std. _      IS * ^RS     . 1QQ         Eq. B-5
                 Percent Recovery     ~ ARS  QIS  RRFJS



where:     Aj$ = sum  of  the  integrated   ion  abundances  determined  for  the
                 labeled PCDD/PCDF internal quantitation standard in question;

           AR5 = sum  of  the   integrated ion  abundances determined for m/z 332
                 and  m/z 334 of  13C12-1,2,3,4-TCDD  or m/z  390  and m/z 392 of
                 i3C12-l,2,3,7,8,9-HxCOD  (recovery standards);

               = amount  (pg)   of  the respective recovery standard, added to the
                 final extract;

               = amount  (pg) of the  labeled  internal  quantitation standard
                 added to the sample prior to extraction; and
                                     B-14

-------
 .           If a  response  for a specific dioxin  or furan congener is qualified
    as  a  trace,  TR, value (signal-to-noise ratio  is greater than or equal to 2.5
    but less than 10), the analyst will also provide  an estimated method detection
    limit.   This is  accomplished  by using the  observed  signal-to-noise ratio on
    either side of  the response and calculating  as given in  Eq. B-6.
                                         B-15

-------
                  APPENDIX C
METHOD FOR ESTIMATING MEASURES OF UNCERTAINTY
                      C-l

-------
       The additive model for analyzing NHATS composite sample data has been presented and
discussed in Chapter 7. This model assumes that the concentration for an individual specimen
is equal to a linear combination of fixed effects (census region, age group, sex, and race group
effects) plus the sum of two independent random effects.  The random effects are the effect due
to random selection of the  MSA and the effect due to random selection of the donor  from that
MSA.  The MSA effect is assumed to have variance o^, and the donor effect has variance a\.  In
addition, random measurement error in the observed concentration for an analyzed composite
sample is also present and assumed to have variance oj. This appendix discusses how these three
measures of uncertainty are estimated and interpreted.

       The estimate of o* is obtained from the analysis of control QC samples, detailed in Chapter
6. Control samples were unspiked samples  taken from a homogenized bulk sample of human
adipose tissue.  As a result, the variability in the predicted concentration for unspiked control
samples reflects measurement error exclusively. The estimate of or was obtained from Table 6-3
by multiplying the predicted concentration by the coefficient of variation (as a proportion) for the
control samples.  The result is equal to the estimated  standard deviation of the measured
concentration in the control samples. The control sample  concentration was nearly equal to the
average concentration of the study samples.   The estimate of or is given for each of the
compounds in Table C-l, along with the coefficient of variation (in percentage terms) relative to
the estimated national average concentration.

             As discussed in  Chapter 7,  estimating the  additive  model parameters in the
prediction process  requires an iterative weighted generalized least squares procedure.  The
procedure is  iterative  and weighted  in nature because the covariance matrix of measured
composite concentrations, used in estimating the fixed effect parameters, depends on the three
unknown variance terms given above and on the unknown fixed parameters themselves. Starting
estimates for these parameters, used to begin the iterative estimation process, are obtained  from
a mixed model analysis which assumes  constant error variances. This approximate model is fitted
using the P3V procedure in the BMDP statistical software package.  The P3V procedure gives
starting values for the  fixed effect  parameters and estimates for two surrogate variance
components. One of the surrogate variance components is due to random selection of MSAs and
the other is due to the combined effects of random donor sampling and measurement error.

                                         C-3

-------
          Table C-l. Estimates of Measurement and Sampling Standard Deviation
                  (in pg/g), and Coefficients of Variation (in percent), for
                  the PCDDs and PCDFs Analyzed in the FY87 NHATS
Standard deviation and CV (%) due to ...
Measurement
error
Analyte
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8/6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
2,3,4,7,8-PeCDF
1,2,3,6,7,8-HxCDF
,
1.77
1.84
15.8
2.78
9.52
57.5
0.613
4.42
4.49
(CVt)'
(33)
(17)
(21)
(24)
(9)
(8)
(33)
(46)
(78)
Sampling
MSAs
om
0.0
0.0
0.0
0.0
14.7
0.0
0.346
0.0
0.0
(CVJ
(0)
(0)
(0)
(0)
(13)
(0)
(18)
(0)
(0)
Sampling
individuals
o
0.0
2.93
0.0
0.0
57.3
571
0.0
0.0
0.0
(CVJ
(0)
(54)
(0)
(0)
(52)
(79)
(0)
(0)
(0)
a CVs (coefficients of variation) are with respect to the estimated national average
 concentration.
                                         C-4

-------
Orban and Lordo (1989) show that the surrogate variance components are linearly related to the
variance components o2., o^, and o. The variance component estimates are obtained by solving
a set of simultaneous linear equations.

       Table C-l lists the estimated standard deviations  and the corresponding coefficients of
variation (percent of national average) for each of the three random effects. For five of the nine
chemicals the  sampling standard deviations were estimated to be  zero.  The formulas for
estimating variance components can produce negative estimates.  These are replaced by zero.
This occurs frequently when measurement variation is large relative to the amount of sampling
variation that affects the composite concentrations.  However, it does not mean that there is no
sampling variation.  In fact, of the four chemicals with positive estimates of sampling standard
deviations, the sampling errors were found to be quite large.

       Two chemicals (1,2,3,4,6,7,8-HpCDD and 2,3,7,8-TCDF) had positive estimates of standard
deviations due to random selection of MSAs. The estimates were 13 and 18 percent of their
respective national average concentrations. Also, three chemicals (1,2,3,7,8-PeCDD, 1,2,3,4,6,7,8--
HpCDD, and OCDD) had positive estimates of standard deviations due to random sampling of
individuals.   These estimates  were between 52  and 79 percent of the  national  average
concentrations.

       The total sampling standard deviation due to both sampling of MSAs and individuals can
be estimated by
For example, the total sampling standard deviation for 1,2,3,4,6,7,8-HpCDD is estimated to be

                               59.2 = [(14.7)2 + (57.3)2f .

On a relative basis this is 54% of the national average concentration of 1,2,3,4,6,7,8-HpCDD.

      The statistical uncertainty of the variance component estimates is difficult to establish

                                         C-5

-------
because of the approximate nature of the estimation procedure and the relatively small number
of composite samples analyzed.  Also, the priorities for creating the sampling and compositing
designs were  aimed primarily at  producing  accurate estimates of the national average
concentrations and determining if there are significant demographic and geographic effects on
the average concentrations. To obtain more precise estimates of sampling standard deviations it
would be necessary to increase the number of composites or individual specimens that are
analyzed and to revise the compositing design objectives.
                                         C-6

-------
               APPENDIX D
PLOTS OF ESTIMATED CONCENTRATIONS VERSUS
 SPIKED LEVEL WITH TOLERANCE BOUNDS FOR
          FY87 NHATS QC SAMPLES
                   D-l

-------
                                               i- O
                                                  IO
                                                  o
                                                  CM
                                                _ O
                                               J- o
8588?
          (CyDd) uofjBJiueouoo
                                                               0)
                            D-3

-------
                                              -8
                                                 10
                                               _ O
                                              J-  o
g       9      8
O       o       O
CM       T-
                                                                 03
                                                                 SJ
                                                                 to
                                                                 V-l 
                                                                  o
                                                                 CJ  (0
                                                                 H CO
                                                                 9 oo
                                                                 O SH
                                                                 O ^
                                                                 U O
                                       O CO
                                      *w -O
                                         C
                                       C 9
                                       o o
                                      H ffl
                                      4J
                                       (0 0)
                                       ^ U
                                      4J 0
                                       C 10
                                       a> M
                                       u a)
                                       a rH
                                       o o
                                      U EH

                                      T3 43
                                       4) 4J
                                      4J-H
                                                                H iH
                                                                jj 0}
                                                                 (0 >
                                       O tJ
                                         (U
                                       -p ^
                                       O-H
                                                                 I
                                                                 a

                                                                 0)
                                                                 ^
                                                                 3
                        D-4

-------
                                                                           o
                                                                           10
                                                                                 5
                                                                                 5.
                                                                                 (0
                                                                       _  O
                                                                       -  O
8        S        ?       8       8       S
 01
 S
 CD
                                                                                              (Q

                                                                                              0)
                                                                                           eu  e

                                                                                           00  CO
                                                                                           Jfi
                                                                                           a, r-
                                                                                           3 00
                                                                                           o >>
u o

 u
 O CO


 a z
 o o
H CQ
.M
 (0 0)
 H 

 C 10
 0) ^
 U 0)
 fi ^H
 o o
U EH
H <-4
4J 0)
(0 >
U 
-------
                                                               u  O
                                                                  in
                                                                  O
                                                                  CM
                                                               _  O
                                                               I-  O
8
                  OO
                            (CyOd) uogaqueouoo
 I

Q


05




&
H
                                             D-6

-------
                                         ^ o
                                            in
-8
                                         _  o
                                         -  o
(D/Cd)
                                                         u
                                                         IT)

                                                         Q



                                                         2
                    D-7

-------
                      1-8
                      -8
                      -S
                      _ o
                      -8
                      -g
                      -3
         i  I  I  I I
                      _ O
                      h O
OO
O O O
I
e

-------
o  o
CM  21

 OOOOOOO
                                          o
                                          CM
                                          T

                                          o
                                          T


                                          8


                                          8
                                       -  P
                              i   i    i
                                                         -8
                                                         -8
                                       _  O
                                       - O
                                                       to
                                                       9
                                                       (0
                                                                         U CO
                                                                         x a)
                                                                         KrH
                                                                          i a
                                                                         00 
                                                                           (0
                                                                         r- w
                                                                          -to
                                                       9 eo
                                                       O >*
                                                       ab

                                                       o ^
                                                       u o
                                                                         o to
                                                                         w-O
                                                                           G
                                                                         a a
                                                                         o o
                                                                         H CQ
                                                                         4J
                                                                         0 0)
                                                                         >-( O
                                                                         4J (3
                                                                         C (0
                                                                         (U M
                                                                         U 0)
                                                       4) -P
                                                       4J-H


                                                       1 *
                                                       H i I
                                                       4J 0)
                                                       03 >
                                                       Ed d)
                                                       o-d
                                                         a)
                                                       P ^
                                                       O-H
                                                       PH CU
                                                       04 (O
 I
Q

O
M
S
t^
H
fc,
(C/Od) uopaaueouoo
                   D-9

-------
                                 ,_ o
                                 h CM
                                 -8


                                 -8
-g

_ o
                                 -g
     I    I   T   1
                                 _ O
                                 h O
                                             (0
                                             3
                                             CO
            Q  
            U CO
            x 0)
            SS^H
             i a
                                             oo co
            PO
             .CO
             61
            0
            G
                                               co
                                             o M
                                             u o
                                               M-l
                                             ^
                                             o (o
                                             M-4 -O
                                               C
                                             a 3
                                             o o
                                             H 03
                                             JJ
                                             (0 0)
                                             M U
                                             *J fi
                                             a A
                                             0) M
                                             U d)
                                             C ^H
                                             o o
                                             U E-i
                                             0)
            4-> (U
            u >
            u a>
              J
            M-l
            O T3
              0)
            4J ^
            O -H
            M a
            & to
                                             00
                                              I
                                             a

                                             0)
                                             M
(B/Od)
                D-10

-------
                                                                  CO
                                                                  8
                                                                  CM
                                                               -8
                                                               -  O
I
CM
            8
                    (B/Cd)
                                                                                CO
                                                                                 %
                                                                                CN
 I
Q

9)
^
3
Oi
H
&4
                                         D-ll

-------
                                                                                 -8
                                                                                     
                                                                                 ^  o
I     I     i
                                                       1     1     I     I     I     1
                                                                                      o
                                                                                      CO
                                                                                      o
                                                                                      CM
                                                                                      o
                                                                                      10
                                                                                 _   o
                                                                                      CO
                                      _  o
                                          CM
                                                                                 L   o
                                                                                 L   o
OOO
OO
                                                                         OO
                                                                                                        0]
                                                                                                        3
                                                                                                        00
                                                            a
                                                            a  
                                                            u  w
                                                            X  flj
                                                             i
                                                                                                           a,
                                                                                                           6
                                                                                                       ao en
                                                                                                       en
                                                            3 00
                                                            O !*


                                                            I*
                                                            U O
                                                               4-1

                                                            O CO

                                                               c
                                                            C 3
                                                            O O
                                                            H 
-------
                                 -8

                                 -8
                                 -8


                                   8


                                   9


                                   8


                                   8

                                   O
                                   I


                                 - O
               8
                                               CO
                                               CD
                                             Q e
                                             U H
 *frl
 H
  M
 T3 O
 C3 >H
 3
 O CO
                                             O O
 O (V
 
 10 d)
 JJ J
 to

 H 0)
 4J .M
 CO-H
 Ed CU
O CO

P CO
 I
O

0)
01
H
(CyOd)
                D-13

-------
                                                           o
                                                           CO
                                                           O
                                                           O
                                                         -8
                                                           O
                                                           CO
                                                         -8
                                                           a
                                                           10
                                  i   I   i
OQOQOOO
 I   I   I

O O O
                                                           O
                                                           CM
                                                         h- O
                                                         h O
                              CO
                              V
                             r4
                           fa a,

                           89
                           Q-OT
                           33

                           o> O
                            ^
                           co en
                           * a
                            ^
                           ro t*.
                           O o
                           C 4-1
                           S
                           O CO
                           a-o
                           g a
                           o P
                           o o
                           m o
                             c

                           O ^
                           H (!)

                           10 O
                           tt j
                           -P
                           c x:
                           flj 4J
                           O-H
                                                                        0) 0)
                                                                        -P J
                                                                        (0
                                                                        ST3
                                                                        H 0)
                                                                        -P ^
                                                                        CO -H
                                                                        H Cb
                           O ca
                             S3
                           4J CO
                           O M
                           rH 0)
                           ft. >
                                                                         I
                                                                        a
                  (0/t3d) UOflRQUCOUOQ
                                     D-14

-------
                                    -8


                                    -8

                                      o
                                      T-

                                    -8


                                    -8
                                    _  O
                                    ^  co
                                    __  o
                                       CM
                                    _  O
                                    k  O
   CO
   0)
  rH
P Oi
o e
u 3
Q-W
s

o O
  W
vo
(N
o o
c IN
3
O CO
CXI?
g G
O 3
                                                    O O
                                                    t-l U

                                                    C (0
                                                    O M
                                                    H 0)
                                                    P -H
                                                    (0 O
0) -P
O-H
C 
O
U r-t
  0)
o >
V 0)
JJ J
(0

H O
-p >J
CO -H
w a
O co
  53
4-) CO
O M
CO
rH


Q


0)
                 CM
(CyOd)
                  D-15

-------
                                           \- o

          8
                                                               CO
                                                               w
                                                            fa^<
                                                            Q  O4

                                                            O  RJ
                                                             I CO
                                                            91
                                                             -u
                                                            oo o
                                                             fc
                                                            r* w
                                                             H
                                                            vo rtj
                                                             *, jg
                                                            * a
                                                             %
                                                            co r
                                                             *oo
                                                            
                                                            4) 0)
                                                           P J
                                                            05
                                                            e-o
                                                           H 0)
                                                           P A:
                                                            CO -H
                                                           u a.
                                                              CO
                                                           IM
                                                            o ca
                                                              3
                                                           P CO
                                                            o ^
                                                           ^H 0)
                                                           A. >
 I
O

0)
(CyOd) UOQ&QUOOUOO
                     D-16

-------
                                                     CO
5

                              CM
                          -  o

                                                 Q a
                                                 88
                                                  I  W
                                                 00
                                                    w
                                                  -oo
                                                 CN (H
                                                 O O
                                                 a 4-1
                                                 a
                                                 o co
                                                 ftTJ
                                                 g C
                                                 O 3
                                                 U O
                                                    03
                                                 M
                                                 O 0)
                                                 *4-l U
                                                    C
                                                 C (0
                                                 O M
                                                 H fl)
                                                 P M
                                                 10 O
                                                 M Ei
                                                 J->
                                                 a js
                                                 (U -P
                                                 O-H
                                                 U ^H
                                                    4)
                                                 d  >
                                                 0)  0)
                                                 4-> >4
                                                 (0
                                                 W-H
                                                 U  (X
                                                   (A
                                                 4-1
                                                 O  CO
                                                    3
                                                 4J  CO
                                                 O  M
                                                 rH  O
                                                 ft >
            m
            iH

            Q

            0)
D-17

-------
                               ,_  o
                               h  co
                               ^  o
                               h  CM
                                -8
                                  
                               (-8

                                  8

                                  O
                                  o
                                  CO
                                _  O
                               }- o
8
CM
8
                                            vo
                                            iH
                                             I
                                            Q
                 0-18

-------
                 Nr-    ">
1    1    I    I    I    I    I     I    T    I
OO
O Q O
                                                                  OO
                                                                             ,_   o
                                                                             h   CM
                                                                              -8

                                                                              -8
                                                                                  S     f
                                                                                  8     |
                                                                                         *
                                                                                  s
                                                                              -3


                                                                              ~   CO
                                                                                  o
                                                                                  CM
                                                                              _  O
                                                                              1-  o
                                                                                                         0)
                                                                                                         0)
                                                                                 co

                                                                                                      i
                                                                                                     00 (0
                                                        rn
                                                         -
                                                        cs
                                                                                oo
                                                                             TJ  O
                                                                              C 4-1
                                                                              3
                                                                              O  CO
                                                                              O  3
                                                                              U  O
                                                                                a

                                                                              o  a)
                                                                             M-I  u
                                                                                 c
                                                                              ti  (0
                                                                              O  M
                                                                             H  
                                                                              4) 0}

                                                                              (0
                                                                              e-o
                                                                             H 0)
                                                                             ^j >^>
                                                                              05 -H
                                                                             Ed CU

                                                                             M-I
                                                                              O CO

                                                                             -P CO
                                                                              O M
                                                                             rH d)
                                                                             a. >
                                                                                                      I
                                                                                                     a
 &
H
                        (CyDd)
                                                 D-19

-------
             APPENDIX E
SUPPLEMENTARY DESCRIPTIVE STATISTICS
   FOR FY87 NHATS PCDDs AND PCDFs
                E-l

-------
SUPPLEMENTARY DESCRIPTIVE STATISTICS FOR
FY87 NHATS PCDDs and PCDFs
       Tables E-l and E-2 give the average measured concentration, standard error, and number
of unrestricted measurements for each of the marginal populations defined by the different levels
of the four analysis factors.  Composites classified as mixed race group or mixed sex group are
composites which contain specimens from both white and non-white donors or both male and
female donors, respectively.  All specimens in each composite are from the same census region
and same age group. For example, there were 36 unrestricted measurements of 2,3,7,8-TCDD and
the average concentration was  6.60 pg/g.  The average was 1.74 pg/g for the  five  composites
containing specimens from  the youngest  age  group (0 to 14 years), 4.33 pg/g from  the 12
composites in the middle age group (15 to 44 years), and 9.31 from the 19 composites in the oldest
age group (45+ years).  The standard errors are all less than 0.6 pg/g. This suggests that the
differences between the age group averages are statistically significant. However, the reported
significance levels in Chapter 8 were determined by a more rigorous statistical approach.
                                        E-3

-------





cS
Q
CH
H
1
B
u.
u
C
_o
"
3

8.
3
u
'
a.


^
J
1"
J?
o
'r
g_
|
U
"o
&
e
a
Z
TJ
C
a
1
t
w
T3
E
(0
TJ
1
cn
(X

S
00
jj
'2
.
U
2
01
^
f2




D
3

S











O
3

8
OS





"c
ra
V
^

6
Si









1
a:
(A
3
w
US



2i
a
w





-Si

u.


1
S


u,
m



5
'r^
2:
c
o
Z

"8
X
S

Ic
^

+
in



^
(2


S
o




5

en



U


M

x:
g




i
A3

l-H







5






^




1

SS



m



vO
*


m





C>

S



S


a
8
T*

a
f
8.
Population

tx in m m \c
"& (N CT* so oo m CN CN fx r** O J""1 O
oo'r-Ht-H T* rj oc -^ <-< -^ Lrir-Jfo tx * m 1Q 3 ^ * iri oo 0 ro oo
" O* ~i


H fx S^ * ffi odo Sr"J5 S^m ^^m rq^m SrSS 2mS ^^S
^- 00
85 ,n35 ^^ ^5  *?{q osr.
in^o gr-it^ StN^ S'^'* ocn1" 33"* lrir' ^^^

^
m^ S ^i moo c$ oo OSM S
O\OCT\ oo'o'ift *iirioo txo'tx inxcoo o-5in ^otritx * i so tx
ri-iTHrji-ii-<-iJsi-iOJ\C-ii-i r-<
rq ts ao "^
m cc oo o oo 91 m oo tx op
M rttX'-lr*r-l^i^fSm ^


S! I 3^ U2  88 oo K Os W S 000
r-! o in m o GO ^o r4 o ^ ^ tx ^o ^ o tx co oo c4 o ^ ^* -^ *
 ^ T*I r-t ^ i-< (SI



M in tx o o m
~ " ^
oSS v^S CoR ^^ ^H SR v^M
\O'~''~' i-iC-lCJ mCh** CN<~<(N O-^'-O v^r^O ^O*O O^^*O
i-t tx r-.

oo co in in tx
^Mr, M^^

cs m ^* o r-t S
inp ^ ^c in oo ** tx tx coin -^ ^
ir^ rn Q\ iQ t*^ oo in in tx o ^ o^ *~* o^ tx m o tx co o r^ o m M
TH tx ^ ^^ m ^ tx tx CT\ r^
r^ S. *-*
o ^ ^ oo o in c^ f^i
\do-o i-Jt-iin ^*cri r4oos ONOOJ o^ tx^ rC cos cov
tx co m c^ m m ro m m
ff? of Oj i-? Of Of M Of Of
















































vi
3
tn
S
u
T-* ^.
C |
0 %
1 "BO ^
if I
11 I
g iN
1 |1i
Illl
(0  V T3
E-4

-------






&
3
a
1





CL
3
S





?e group (years)
<

_
c2
3
c
CJ





.
UJ












i

i
*
^
s



i
c
o
Z
8
.2

IS

+
S

,
1-H
2
.,




f.
2

w
2


.2
c










c
^

in


c*
*




^



'




CO

s
*
Ov

m



e3


S


S
1 





"U

jjp
c
s
01
CL
c
o
"ra


O
G-
to CO ^ r- i-H ON
r-i o' oo o' o co in (*i **
r-*OCO O O ** OO -H M
vot.
i> Csl CO O tN ON
MOM <=<=> so  ^ 




i i 0 I i O i i O


o^o
*" O so O O IX, fsl f ' CO
CS CO TH CO

ss si R
csio'tx o o so o ri so

in l 5 0 IN SI
CNJ o in o o so IN -H oo
m o M o c- in
O CO rsl ON
o w ?5 ^ f^ so
CSJO'ON OO*~< MOC^
3$ ^ S&
HC;VO ""^ ^"^
sO OO OO ON C3
NO *-< CO O t* C>
** ** ** ~

M SO tV
co co co o *r ^
csooo oo^f cocsico

ON CN] CS O C^ CO
i-iOOO OOOO ^COtv.

8S Ss ooK)
CO Tf i CO
CO






IJj Ij
Q Q
s y y
y ^ ^
E7 oo oo

IN^ CO" **
r"T r-f co"
fst i-? c*T
o in
t-I <>J CO oo' -* t-< O O' CO
OO T-H < m

o sr> m so co
ixcoco ifi^Hso csio'rj

N g en iS S co
fsltsco ooes ootx
i-i OO vO CO CsJ O
<-HitM rsOO OOOO
co co m co
* CO tX CO tx
O CO CO U"i r* sO O ' *-H
s s2 SS
F-

rsi m

tx fsi r-i so rH co o ^ in
M
St-i CO OO t> CO
i-H in oo r-j m
tv co co in ^H vo ^

rn t^. in
O OO CO r-< O in
CM ^ tx sO t-x 1-"
CO 1-H







tU El* tU
D D Q
XXX
k v ^.


c c 2
o * ^ o
c2 < Co H
 J3 w -O
E-5

-------
50272-101
REPORT DOCUMENTATION 1- REPORT NO- 2.
PAGE EPA 560/5-91 -003
4. Title and Subtitle
Chlorinated dioxins and furans in the general United States.
Population: NHATS FY87 results
'. Author(s)
U.S. Stanley (1) and J. Orban (2)
). Performing Organization Name and Address
1) Midwest Research Institute, 425 Volker Boulevard, Kansas City, Missouri 64110
2) Battelle Columbus Division, 505 King Avenue, Columbus, Ohio 43201
2. Sponsoring Organization Name and Address
Exposure Evaluation Division TS-798, Office of Toxic Substances
U.S. Environmental Protection Agency, 401 M Street, S.W., Washington, DC 20460
J. Remmers, J. Schwemberger
3. Recipient's Accession No.
5. Report Date
November 26, 1 991
6.
8. Performing Organization RepL No.
10. Projectflask/Work Unit No.
9801 -A(27)
11. Contract(C) or Grant(G) No.
(C) 68-DO-0137, WA27
68-02-4292
(G)
13. Type of Report & Period Covered
Final Report
14.
5. Supplementary Notes
 6. Abstract (Limit: 200 words)

   Population estimates of the average levels of polychlorinated dibenzo-p-dioxins (dioxins or PCDD) and polychlorinated
   dibenzofurans (furans or PCDF) were established using 865 human adipose tissue specimens collected in Fiscal Year
   1987 through the U.S. Environmental Protection Agency's National Human Adipose Tissue Survey (NHATS). The
   specimens were composited into 48 unique samples prior to chemical analysis.  Estimates of the national average
   concentrations of levels were made among subpopulations defined by the donors' geographic adipose tissue of the
   U.S. population was 5.38 pg/g (0.32); however, the levels increase from 1.98 pg/g (0.81) in children under 14
   years of age to 9.40 pg/g (0.38) in adults over 45 years old. Significant age effects were determined for all nine of
   the compounds which were present at quantifiable levels in greater than 90% of all samples.  Statistically significant
   differences based on geographic regions were determined for estimated levels of 2,3,4,7,8-PCCDF with the highest
   levels in the northeast and the lowest levels in the west.  There were no significant differences in the estimated levels
   from different sexes or race groups for any of the target analytes.
 7. Document Analysis  a. Descriptors

   Polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans, PCDD, PCDF, 2,3,7,8-TCDD, Human Adipose
   Tissue, National Human Adipose Tissue Survey (NHATS), Fiscal Year 1987, FY 1987.

  b. Identifiers/Open-Ended Terms

   Determination, analysis, general United States population, body burden, weighted national averages.

  c. COSATI Field/Group
8. Availability Statement
19. Security Class (This Report)
20. Security Class (This Page)
21. No. of Pages
277
22. Price
;ee ANSI-Z39.18)
See Instructions on Reverse
                                                    OPTIONAL FORM 272 (4-77)
                                                    (Formerly NTIS-35)
                                                    Department of Commerce

-------