United States
           Environmental Protection
           Agency
                  Office of Research and
                  Development
                  Washington DC 20460
EPA/600/P-00/OOlB.fl
September 2000
SAB Review Draft
c/EPA
                              Review
                              Draft
                              (Do Not
Exposure and Human    C|te or
Health Reassessment ofQuote)
2,3,7,8-Tetrachlorodibenzo-
p-Dioxin (TCDD) and
Related Compounds

Part III: Integrated Summary and
Risk Characterization for 2,3,7,8-
Tetrachlorodibenzo-p-Dioxin
(TCDD) and Related Compounds
                        Notice
           This document is a preliminary draft. It has not been formally
           released by EPA and should not at this stage be construed to
           represent Agency policy. It is being circulated for comment on its
           technical accuracy and policy implications.

-------

-------
DRAFT
DO NOT CITE OR QUOTE
EPA/600/P-00/001Bg
September 2000
SAB Review Draft
www.epa.gov/ncea
          Exposure and Human Health Reassessment
        of 2,3,7,8-Tetrachlorodibenzo-/?-Dioxin (TCDD)
                     and Related Compounds
        Part III: Integrated Summary and Risk Characterization for
   2,3j7,8-Tetrachlorodibenzo-/?-Dioxm (TCDD) and Related Compounds
                                NOTICE

THIS DOCUMENT IS A PRELIMINARY DRAFT. It has not been formally released by the
U.S. Environmental Protection Agency and should not at this stage be construed to represent
Agency policy. It is being circulated for comment on its technical accuracy and policy
implications.
                  National Center for Environmental Assessment
                      Office of Research and Development
                     U.S. Environmental Protection Agency
                             Washington, DC

-------

-------
                                  DISCLAIMER
      This document is a draft for review purposes only and does not constitute U.S.
Environmental Protection Agency policy. Mention of trade names or commercial products does
not constitute endorsement or recommendation for use.
9/22/00
                                         n
DRAFT—DO NOT CITE OR QUOTE

-------

-------
                      TABLE OF CONTENTS - OVERVIEW

                  Exposure and Human Health Reassessment
                of 2,3,7,8-Tetrachlorodibenzo-/7-Dioxin (TCDD)
                            and Related Compounds

Part I: Estimating Exposure to Dioxin-Like Compounds (Draft Final)
            (EPA/600/P-00/001 Bb, Be, Bd) September 2000

Volume 1 :    Executive Summary (EPA/600/P-00/001Ba) j^^tlte^

Volume 2:    Sources of Dioxin-Like Compounds in the United States (EPA/600/P-00/001Bb)
            Chapters  1 through 13

Also inc
                               Database of Sources of Environmental Releases of Dioxin-
                               Like Compounds in the United States (Draft Final)
                               (EPA/600/P-98/Q02B) September 2000

Volume 3 :    Properties, Environmental Levels, and Background Exposures
            (EPA/600/P-OO/OOlBc)
            Chapters 1 through 6

Volume 4:    Site-Specific Assessment Procedures (EPA/600/P-00/00 IBd)
            Chapters 1 through 8                             -.-

Part II:     Health Assessment for 2,3,7,8-Tetrachlorodibenzo-Jp-dioxin (TCDD) and
            Related Compounds (Draft Final)
            (EPA/600/P-00/001Be) September 2000
      Chapter 1 .     Disposition and Pharmacokinetics
      Chapter 2.     Mechanism(s) of Actions
      Chapter 3.     Acute, Subchronic, and Chronic Toxicity
      Chapter 4.     Immunotoxicity
      Chapter 5.     Developmental and Reproductive Toxicity
      Chapter 6.     Carcinogenicity of TCDD in Animals
      Chapter 7.     Epidemiology/Human Data

    ,  Chapter 8.     Dose-Response Modeling for 2,3,7,8-TCDD
                   (SAB Review Draft, September 2000)
      Chapter 9.     Toxic Equivalency Factors (TEF) for Dioxin and Related Compounds
                   (SAB Review Draft, September 2000)

Part III:     Integrated Summary and Risk Characterization for
            2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds
            (SAB Review Draft, September 2000) (EPA/600/P-00/001Bg)
9/22/00
                                       111
                                                DRAFT—DO NOT CITE OR QUOTE

-------

-------
                                    CONTENTS
 1. INTRODUCTION	1
         1.1.  DEFINITION OF DIOXIN-LIKE COMPOUNDS . .	  3
         1.2.  TOXIC EQUIVALENCY FACTORS  	4
         1.3.  UNDERSTANDING EXPOSURE/DOSE RELATIONSHIPS FOR
             DIOXIN-LIKE COMPOUNDS	8
             1.3.1. Administered Dose	9
             1.3.2. Area Under the Curve	10
             1.3.3. Plasma or Tissue Concentrations  	12
             1.3.4. Steady-State Body Burdens	13
             1.3.5. Mechanistic Dose Metrics	14
             1.3.6. Summary	14

2. EFFECTS  SUMMARY	14
        2.1.  BIOCHEMICAL RESPONSES	16
        2.2.  ADVERSE EFFECTS IN HUMANS AND ANIMALS	19
             2.2.1. Cancer	19
                    2.2.1.1.  Epidemiologic Studies	19
                    2.2.1.2.  Animal Carcinogenicity	24
                    2.2.1.3.  Plausible Mode(s) of Carcinogenic Action	27
                    2.2.1 A.  Other Data Related to Carcinogenesis 	29
                    2.2.1.5.  Cancer Hazard Characterization  	._„	30
             2.2.2.  Reproductive and Developmental Effects	'	31
                    2.2.2.1.  Human	32
                    2.2.2.2.  Experimental Animal	."	"	34
                    2.2.2.3.  Other Data Related to Developmental and Reproductive
                           Effects	37
                    2.2.2.4.  Developmental and Reproductive Effects Hazard
                           Characterization	39
             2.2.3.  Immunotoxicity	'	40
                    2.2.3.1.  Epidemiologic Findings	40
                    2.2.3.2.  Animal Findings	41
                    2.2.3.3.  Other Data Related to Immunologic Effects	42
                    2.2.3.4.  Immunologic Effects Hazard Characterization	43
             2.2.4.  Chloracne	44
             2.2.5.  Diabetes	45
             2.2.6.  Other Effects	47
                    2.2.6.1.  Elevated GOT	47
                    2.2.6.2.  Thyroid Function	48
                   2.2.6.3.  Cardiovascular Disease	49
                    2.2.6.4.  Oxidative Stress	49

3. MECHANISMS AND MODE OF DIOXIN ACTION	 50
        3.1.  MODE VERSUS MECHANISM OF ACTION	51
9/22/00
                                        IV
DRAFT—DO NOT CITE OR QUOTE

-------
       3.2. GENERALIZED MODEL FORDIOXIN ACTION ..	52
           3.2.1. The Receptor Concept 	52
           3.2.2. A Framework to Evaluate Mode of Action	54
           3.2.3. Mechanistic Information and Mode of Action; Implications for Risk
                 Assessment	55

4. EXPOSURE CHARACTERIZATION	58
       4.1. SOURCES	59
           4.1.1. Inventory of Releases	 60
           4.1.2. General Source Observations  	63
       4.2. ENVIRONMENTAL FATE	66
       4.3. ENVIRONMENTAL MEDIA AND FOOD CONCENTRATIONS  	68
       4.4. BACKGROUND EXPOSURES  	70
           4.4.1. Tissue Levels	70
           4.4.2. Intake Estimates 	72
           4.4.3. Variability in Intake Levels	72
       4.5. POTENTIALLY HIGHLY EXPOSED POPULATIONS OR
           DEVELOPMENTAL STAGES	73

5. DOSE-RESPONSE CHARACTERIZATION	77
       5.1. DOSE METRIC(s)  	79
           5.1.1. Calculations of Effective Dose (ED)  	82
       5.2. EMPIRICAL MODELING OF INDIVIDUAL DATA SETS	83
           5.2.1. Cancer	84
                 5.2.1.1.  Estimates of Slope Factors and Risk at Current Background
                        Body Burdens Based on Human Data	89
                 5.2.1.2.  Estimates of Slope Factors and Risk at Current Background
                        Body Burdens Based on Animal Data	91
                 5.2.1.3.  Estimates of Slope Factors and Risk at Current Background
                       Body Burdens Based on a Mechanistic Model	93
           5.2.2. Noncancer Endpoints 	94
       5.3. MODE-OF-ACTION BASED DOSE-RESPONSE MODELING	95
       5.4. SUMMARY DOSE-RESPONSE CHARACTERIZATION	96

6. RISK CHARACTERIZATION 	99

GLOSSARY AND DEFINITIONS	160

REFERENCES FORRISK CHARACTERIZATION 	166
9/22/00
DRAFT—DO NOT CITE OR QUOTE

-------
                                  LIST OF TABLES

Table 1-1. The TEF scheme for I-TEQDF	124
Table 1-2. The TEF scheme for TEQDFP-WHO94	.125
Table 1-3. The TEF scheme for TEQDFP-WHO98	.126
Table 1-4. The range of the in vivo REP values for the major TEQ contributors  	127
Table 1-5. Comparison of administered dose and body burden in rats and humans  	128
Table 2-1. Effects of TCDD and related compounds in different animal species  	129
Table 2-2. Examples of margins of exposure (M-O-E)	130
Table 2-3. Summary of the combined cohort and selected industrial cohort studies with
          high exposure levels as described by IARC, 1997	,.131
Table 2-4. Tumor Incidence and Promotion Data Cited for the TEF-WHO98 for Principal
          Congeners  	132
Table 3-1. Early molecular events in response to dioxin'	133
Table 4-1. Confidence rating scheme	134
Table 4-2. Quantitative inventory of environmental releases of TEQDF-WHO98 in the
          United States  	135
Table 4-3. Preliminary indication of the potential magnitude of TEQDF-WHO9g releases
          from "unquantified" (i.e.,  Category D) sources in reference year 1995	137
Table 4-4. Sources that are currently unquantifiable 1 (i.e., Category E)	138
Table 4-5. Summary of North American CDD/CDF and PCS TEQ-WHO98 Levels in
          Environmental Media and Food	139
Table 4-6. Background serum levels in the United States 1995 -1997 	140
Table 4-7. Adult contact rates and background intakes of dioxin-like compounds	141
Table 4-8. Variability hi average daily TEQ intake as a function of age  ".	142
Table 5-1. Peak serum dioxin levels in the background population and epidemiological
          cohorts (back-calculated)	'.	".	 143
Table 5-2. Summary of Cancer Epidemiology and Bioassay Data in Dose-Response
          Calculations	,	145
Table 5-3. Doses yielding 1% excess risk (95% lower confidence bound) based upon
          2-year animal carcinogenicity studies using simple multistage (Portier et al.,
          1984) models	148
Table 5-4. Summary of All Site Cancer ED01s and Slope Factor Calculations  	149


                                 LIST OF FIGURES

Figure 1-1. Chemical structure of 2,3,7,8-TCDD and related compounds	151
Figure 2-1. Cellular mechanism for AhR action	152
Figure 4-1. Estimated CDD/CDF I-TEQ emissions to air from combustion sources in the
           United States, 1995	154
Figure 4-2. Comparison of estimates of annual I-TEQ emissions to air (grams I-TEQ/yr)
           for reference years 1987 and 1995	.155
Figure 4-3. Blood levels (I-TEQ for CDD/CDF + WHO94) versus  age of a subset of
           participants in the CDC (2000)	156
Figure 4-4. Lipid (a) and body burden (b) concentrations in a hypothetical female until
9/22/00
                                         VI
DRAFT—DO NOT CITE OR QUOTE

-------
                            LIST OF TABLES (continued)

           age 70 under four nursing scenarios: formula only, and 6-week, 6-month, and
           1 year nursing	157
Figure 5-1. Peak dioxin body burden levels in background populations and
           epidemiological cohorts	158
Figure 5-2. Comparison of lifetime average body burden and area under the curve in
           hypothetical background and occupational scenarios	159
9/22/00
vn
DRAFT—DO NOT CITE OR QUOTE

-------
             LIST OF ACRONYMS, ABBREVIATIONS, AND SYMBOLS

Ah       aryl hydrocarbon
AHF      altered heptacellular foci
AhR      aryl hydrocarbon receptor
ALK      alkaline phosphatase
ALT      alanine aminotransferase
Arnt      aryl hydrocarbon receptor nuclear translocator
AST      aspartate aminotransferase
ATSDR   Agency for Toxic Substances and Disease Registry
AUC      area under the curve
BaP       benzo[a]pyrene
BDD      brominated dibenzodioxin
BDF      polybrominated dibenzofuran
BMD     benchmark dose
BW       body weight
CDC      Centers for Disease Control and Prevention
CDD      chlorinated dibenzodioxin
CFD      chlorinated dibenzofuran
CI        confidence interval
CTL      cytotoxic T lymphocyte
CYP1A1  cytochrome P4501A1 enzyme
CYP1A2  cytochrome P4501A2 enzyme
GYP 1B1  cytochrome P4501B1 enzyme                         - -
DFP (subscript)       dioxins, furans, PCBs
DEN      diethylnitrosamine
DHT      5a-dihydrotestosterone
DNA      deoxyribonucleic acid
ED       effective dose
ED0]      effective dose at the 1% response level
EDC/VC  ethylene dichloride/vinyl chloride
EGF      epidermal growth factor
EGFR    • epidermal growth factor receptor
EPA      U.S. Environmental Protection Agency
FSH      follicle-stimulating hormone
g         gram
GD       gestation day
GGT      gamma glutamyl transferase
HAH      halogenated aromatic hydrocarbons
HCDD    hexachlorodibenzo-p-dioxin
HIF       hypoxia-inducible factor
HpCDD   heptachlorodibenzo-j>dioxin
hr        hairless
IARC     International Agency for Research on Cancer
ID        immunosuppressive dose
9/22/00
vm
DRAFT—DO NOT CITE OR QUOTE

-------
IgA       immunoglobulin A
I-P        initiation-promotion
IPCS      International Programme on Chemical Safety (WHO)
I-TEQ     international TEF scheme adopted by EPA in 1989
kd         kilodalton
kg         kilogram
L          liter
LEDoi     lower bound of the effective dose at the 1% response level
LH        luteinizing hormone
LMS      linearized multistage
LOAEL    lowest-observed adverse effect level
MOE  '    margin of exposure
mRNA     messenger ribonucleic acid
MRL      minimal risk level (ATSDR)
NAS      National Academy of Sciences
NHANES  National Health and Nutrition Examination Survey
NHATS    National Human Adipose Tissue Survey
ng         nanogram
NIOSH    National Institute for Occupational Safety and Health
NRC      National Research Council
NTP       National Toxicology Program
NOAEL    no-observed adverse effect level
NOEL     no-observed effect level
OCDD     octachlorodibenzo-/?-dioxin
pg         picogram
PAH      polycyclic aromatic hydrocarbon
PBPK     physiologically based pharmacokinetic
PBDD     polybrominated dibenzodioxin
PBDF     polybrominated dibenzofuran
PCB       polychlorinated biphenyl
PCDD     polychlorinated dibenzodioxin
PCDF     polychlorinated dibenzofuran
PCP       pentachlorophenol
PCQ      polychlorinated quaterphenyl
PeCDD    pentachlorodibenzo-/>-dioxin
PeCDF     pentachlorodibenzo-j>-furan
PK        pharmacokinetic
POTW     publicly-owned treatment works
ppt        part per trillion
PVC      polyvinyl chloride
RfD       reference dose (EPA)
RR        relative risk
SAB      U.S. EPA's Science Advisory Board
SMR      standardized mortality ratio
SRBC     sheep red blood cells
9/22/00
IX
DRAFT—DO NOT CITE OR QUOTE

-------
2,4,5-T     2,4,5-trichorophenoxyacetic acid
TBD       thyroid binding globulin
TCDD     2,3 5738-tetrachlorodibenzo-j?-dioxin
TCP       trichlorophenol
TDI       tolerable daily intake
TEF       toxic equivalency factor
TEQ       toxic equivalent
TEQ-WHO94        1994 WHO extension of the I-TEF scheme to include 13 dioxin-like PCBs
TEQ-WHO98        1998 WHO update to the previously established TEFs for dioxins, furans,
                    and dioxin-like PCBs
TPA       tetradecanoyl phorbol acetate
TNP-LPS  trinitrophenyl-lipopolysaccharide .
TSH       thyroid stimulating hormone
URL       unit risk level
WHO      World Health Organization

~          approximately
>          greater than
<          less than
s          greater than or equal to
<.          less than or equal to
jag         microgram
9/22/00
x
DRAFT—DO NOT CITE OR QUOTE

-------

-------
                         AUTHORS AND CONTRIBUTORS
William H. Farland
Director
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC

Linda S. Birnbaum
Director
Experimental Toxicology Division
National Health and Environmental Effects Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina

David H. Cleverly
Exposure Analysis and Risk Characterization Group
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC

Michael J. DeVito
Experimental Toxicology Division
National Health and Environmental Effects Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina

Matthew N. Lorber
Exposure Analysis and Risk Characterization Group
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC

Bruce D. Rodan
Medical Officer/Senior Health Scientist
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
9/22/00
                                         XI
DRAFT—DO NOT CITE OR QUOTE

-------
                      AUTHORS AND CONTRIBUTORS (continued)
John L. Schaum
Group Chief
Exposure Analysis and Risk Characterization Group
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC

Linda C. Tuxen
Special Assistant
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC

Dwain L. Winters
Director
Dioxin Policy Project
Office of Pollution Prevention and Toxics
Office of Prevention, Pesticides, and Toxic Substances
U.S. Environmental Protection Agency
Washington, DC
9/22/00
xn
DRAFT—DO NOT CITE OR QUOTE

-------
                                         1.  INTRODUCTION
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
       This document presents an integrated summary of available information related to
exposure to and possible health effects of dioxin and related compounds.  It also presents a short
risk characterization, which is a concise statement of dioxin science and the public health
implications of both general population exposures from environmental "background'^and
incremental exposures associated with proximity to sources of dioxin and related compounds.
Even though it summarizes key findings developed in the exposure and health assessment
portions (Parts I and II, respectively) of the Agency's dioxin reassessment, it is meant to be
detailed enough to stand on its own for the average reader.  Readers are encouraged to refer to the
more detailed documents for further information on the topics covered here and to see complete
literature citations.  These documents are:

    Estimating Exposure to Dioxin-like Compounds:  This document, hereafter referred to as Part
    I, the Exposure Document, is divided into four volumes: (1) Executive Summary; (2) Sources
    of Dioxin-Like Compounds in the United States; (3) Properties, Environmental Levels, and
    Background Exposures; and (4) Site-Specific Assessment Procedures.

    Health Assessment for 2,3,7,8-TCDD and Related Compounds: Thistdocument, hereafter
    referred to as Part II, the Health Document, contains two volumes with nine chapters
    covering pharmacokinetics, mechanisms of action, epidemiology, ariimal cancer and various
    noncancer effects, toxic equivalency factors (TEFs), and dose-response.

       Parts of this integrative summary and risk characterization go beyond individual chapter
findings to  reach general conclusions about the potential impacts of dioxin-like compounds on
human health. This document specifically identifies issues concerning the risks that may be
occurring in the general population at or near population background exposure levels. It
               'The term "background" exposure has been used throughout this reassessment to describe
        exposure which regularly occurs to members of the general population from exposure media
        (food, air, soil, etc.) that have dioxin concentrations within the normal background range.  Most
        (>95%) of background exposure results from the presence of minute amounts of dioxin-like
        compounds in dietary fat, primarily from the commercial food supply. The origin of this
        background exposure is from three categories of sources: naturally formed dioxins,
        anthropogenic dioxins from contemporary sources and dioxins from reservoir sources. The term
        "background exposure" as used in this document should not be interpreted as indicating the
        significance or acceptability of risk associated with such exposures.
        9/22/00
                                           1
DRAFT—DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
articulates the strengths and weaknesses of the available evidence for possible sources, exposures
and health effects, and presents assumptions made and inferences used in reaching conclusions
regarding these data.  The final risk characterization provides a synopsis of dioxin science and its
implications for characterizing hazard and risk for use by risk assessors and managers inside and
outside EPA and by the general public.

      This document (Part III) is organized as follows:

      1. Introduction - This section describes the purpose/organization of, and the process for
      developing, the report; defines dioxin-like compounds in the context of the EPA
      reassessment;  and explains the Toxic Equivalence (TEQ) concept.
      2. Effects Summary - This section summarizes the key findings of the Health Document
      and provides links to relevant aspects of exposure, mechanisms, and dose-response.
      3. Mechanisms and Mode of Dioxin Action - This section discusses the key findings on
      effects in terms of mode of action. It uses the "Mode-of-Action Framework" recently
      described by the World Health Organization (WHO) International Programme on
      Chemical Safety's (IPCS) Harmonization of Approaches to Risk Assessment Project and
      contained in the Agency's draft Guidelines for Carcinogen Risk Assessment as the basis
      for the discussions.                                        »-
      4. Exposure Summary - This section summarizes the key findings of the Exposure
      Document and links them to the effects, mechanisms, and dose-response characterization.
      5. Dose Response Summary - This section summarizes approaches to dose response
      that are found in the Health Document and provides links to relevant aspects of exposure
      and effects.
      6. Risk Characterization - This section presents conclusions based on an integration of
      the exposure, effects, mechanisms and dose response information. It also highlights key
      assumptions and uncertainties.

      The process for developing this risk characterization and companion documents has been
open and participatory. Each of the documents has been developed in collaboration with
scientists from inside and outside the Federal Government  Each document has undergone
extensive internal  and external review, including review by EPA's Science Advisory Board
(SAB). In September 1994, drafts of each document, including an earlier version of this risk
characterization, were made available for public review and comment.  This included a 150-day
comment period and 11 public meetings around the country to receive oral and written
        9/22/00
                                                   DRAFT—DO NOT CITE OR QUOTE

-------
comments. These comments, along with those of the SAB, have been considered in the drafting
of this final document. The Dose-Response Chapter of the Health Document underwent peer
review in 1997; an earlier version of this Integrated Summary and Risk Characterization
underwent development and review in 1997 and 1998, and comments have been incorporated.  In
addition, as requested by the SAB, a chapter on Toxic Equivalency has been developed and
underwent external peer review in parallel with the Integrated Summary and Risk
Characterization in July, 2000.  Review by the SAB of the Dose-Response Chapter, the Toxic
Equivalency Chapter and the Integrated Summary and Risk Characterization is the final step in
this open and participatory process of reassessment.  When complete, and following final SAB
review, the comprehensive set of background documents and this integrative summary and risk
characterization will be published as final reports and replace the previous dioxin assessments as
the scientific basis for EPA decision-making.

1.1. DEFINITION OF DIOXIN-LIKE COMPOUNDS
       As defined in Part I, this assessment addresses specific compounds in the following
chemical classes: polychlorinated dibenzo-£>-dioxins (PCDDs or CDDs), polychlorinated
dibenzofurans (PCDFs or CDFs), polybrominated dibenzo-/?-dioxins (PBDDs or BDDs),
polybrominated dibenzofurans (PBDFs or BDFs), and polychlorinated biphenyls (PCBs), and
describes this subset of chemicals as "dioxin-like." Dioxin-like refers to the fact that these
compounds have similar chemical structure, similar physical-chemical properties, and invoke a
common battery of toxic responses. Because of their hydrophobic nature and resistance towards
metabolism, these chemicals persist and bioaccumulate in fatty tissues of animals and humans.
The CDDs include 75 individual compounds; CDFs include  135 different compounds. These
individual compounds are referred to technically as congeners. Likewise, the BDDs include 75
different congeners and the BDFs include an additional 135 congeners. Only 7 of the 75
congeners of CDDs, or of BDDs, are thought to have dioxin-like toxicity; these are ones with
chlorine/bromine substitutions in, at a minimum, the 2,3,7,  and 8 positions. Only 10 of the 135
possible congeners of CDFs or of BDFs are thought to have dioxin-like toxicity; these also are
ones with substitutions in the 2,3,7, and 8 positions. This suggests that 17 individual
CDDs/CDFs, and an additional  17 BDDs/BDFs, exhibit dioxin-like toxicity. The database on
many of the brominated compounds regarding dioxin-like activity has been less extensively
evaluated, and these compounds have not been explicitly considered in this assessment.
       There are 209 PCB congeners.  Only 12 of the 209 congeners are thought to have dioxin-
like toxicity; these are PCBs with 4 or more lateral chlorines with 1 or no substitution in the
ortho  position.  These compounds are sometimes referred to  as coplanar, meaning that they can
assume a flat configuration with rings in the same plane.  Similarly configured polybrominated
9/22/00
DRAFT—DO NOT CITE OR QUOTE

-------
  1      biphenyls (PBBs) are likely to have similar properties. However, the database on these
  2      compounds with regard to dioxin-like activity has been less extensively evaluated, and these
  3      compounds have not been explicitly considered in this assessment. Mixed chlorinated and
  4      brominated congeners of dioxins, furans, and biphenyls also exist, increasing the number of
  5      compounds potentially considered dioxin-like within the definitions of this assessment. The
  6      physical/chemical properties of each congener vary according to the degree and position of
  7      chlorine and/or bromine substitution.  Very little is known about occurrence and toxicity of the
  8      mixed (chlorinated and brominated) dioxin, furan, and biphenyl congeners. Again, these
  9      compounds have not been explicitly considered in this assessment. Generally speaking, this
10      assessment focuses on the 17 CDDs/CDFs and a few of the coplanar PCBs that are frequently
11      encountered in source characterization or environmental samples.  While recognizing that other
12      "dioxin-like" compounds exist in -the chemical classes discussed above (e.g., brominated or
13      chlorinated/brominated congeners) or in other chemical classes (e.g., halogenated naphthalenes
14      or benzenes, azo- or azoxybenzenes), the evaluation of less than two dozen chlorinated congeners
15      is generally considered sufficient to characterize environmental "dioxin."
16            The chlorinated dibenzodioxins and dibenzofurans are tricyclic aromatic compounds with
17      similar physical and chemical properties. Certain of the PCBs (the so-called coplanar or mono-
18      ortho coplanar congeners) are also structurally and conformationally similar. The most widely
19      studied of this general class of compounds is 2,3,7,8-tetrachlorodibenzo-j?-dioxin (TCDD). This
20      compound, often called simply "dioxin," represents the reference compound for this class  of
21      compounds. The structure of TCDD and several related compounds is shown in Figure 1-1.
22      Although sometimes confusing, the term "dioxin" is often also used to refer to the complex
23      mixtures of TCDD and related compounds emitted from sources, or found in the environment or
24      in biological samples. It can also be used to refer to the total TCDD "equivalents" found in a
25      sample.  This concept of toxic equivalency is discussed extensively in Part II, Chapter 9, Section
26      9.4 and is summarized below.
27
28      1.2. TOXIC EQUIVALENCY FACTORS
29            CDDs, CDFs, and PCBs are commonly found as complex mixtures when detected in
30      environmental media and biological tissues, or when measured as environmental releases from
31      specific sources. Humans are likely to be exposed to variable distributions of CDDs,  CDFs, and
32      dioxin-like PCB congeners that vary by source and pathway of exposures. This complicates the
33      human health risk assessment that may be associated with exposures to variable mixtures of
34      dioxin-like compounds. In order to address this problem, the concept of toxic equivalency has
35      been considered and discussed by the scientific community, and TEFs have been developed and
36      introduced to facilitate risk assessment of exposure to these chemical mixtures.
        9/22/00
4
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
       On the most basic level, TEFs compare the potential toxicity of each dioxin-like
compound comprising the mixture to the well-studied and understood toxicity of TCDD, the
most toxic member of the group.  The background and historical perspective regarding this
procedure is described in detail in Part II, Chapter 9, Section 9.1, 9.2, and in Agency documents
(U.S. EPA 1987,1989,1991a). This procedure involves assigning individual TEFs to the
2,3,7,8-substituted CDD/CDF congeners and "dioxin-like" PCBs. To accomplish this, scientists
have reviewed the toxicological databases along with considerations of chemical structure,
persistence, and resistance to metabolism, and have agreed to ascribe specific, "order of
magnitude" TEFs for each dioxin-like congener relative to TCDD, which is assigned a TEF of
1.0. The other congeners have TEF values ranging from 1.0 to 0.00001. Thus, these TEFs are
the result of scientific judgment of a panel of experts using all of the available data and are
selected to account for uncertainties in the available data and to avoid underestimating risk. In
this sense, they can be described as "public health conservative" values.  To apply this TEF
concept, the TEF of each congener present in a mixture is multiplied by the respective mass
concentration and the products are summed to represent the 2,3,7,8-TCDD Toxic Equivalence
(TEQ) of the mixture, as determined by Equation 1-1.
TEQ = £ ^(Congener, x
                                                    ...... \Congenern x TEFn j
                                                                                 (1-1)
The TEF values for PCDDs and PCDFs were originally adopted by international convention
(U.S. EPA, 1989a). Subsequent to the development of the first international TEFs for
CDD/CDFs, these values were further reviewed and/or revised and TEFs were also developed for
PCBs (Ahlborg et al., 1994; van den Berg et al., 1998). A problem arises in that past and present
quantitative exposure and risk assessments may not have clearly identified which of three TEF
schemes was used to estimate the TEQ. This reassessment introduces a new uniform TEQ
nomenclature that clearly distinguishes between the different TEF schemes and identifies the
congener groups included in specific TEQ calculations.  The nomenclature uses the following
abbreviations to designate which TEF scheme was used in the TEQ calculation:

1.      I-TEQ refers to the International TEF scheme adopted by EPA in 1989 (U.S. EPA,
       1989a).  See Table 1-1.
2.      TEQ-WHO94 refers to the 1994 WHO extension of the I-TEF scheme to include 13
       dioxin-like PCBs (Ahlborg et al., 1994). See Table 1-2.
3.      TEQ-WHO98 refers to the 1998 WHO update to the previously established TEFs for
       dioxins, furans, and dioxin-like PCBs (van den Berg et al., 1998).  See Table 1-3.
        9/22/00
                                                   DRAFT—DO NOT CITE OR QUOTE

-------
  1             The nomenclature also uses subscripts to indicate which family of compounds is included
  2      in any specific TEQ calculation. Under this convention, the subscript D is used to designate
  3      dioxins, the subscript F to designate furans and the subscript P to designate PCBs. As an
  4      example, "TEQDF-WHO98" would be used to describe a mixture for which only dioxin and furan
  5      congeners were determined and where the TEQ was calculated using the WHO98 scheme. If
  6      PCBs had also been determined, the nomenclature would be "TEQDFP-WHO98." Note that the
  7      designations TEQDF-WHO94 and I-TEQDF are interchangeable, as the TEFs for dioxins and furans
  8      are the same in each scheme. Note also that in the current draft of this document, I-TEQ
  9      sometimes appears without the D and F subscripts. This indicates that the TEQ calculation
10      includes both dioxins and furans.
11              This reassessment recommends that the WHO98 TEF scheme be used to assign toxic
12      equivalency to complex environmental mixtures for assessment and regulatory purposes. Later
13      sections of this document describe the mode(s) of action by which dioxin-like chemicals mediate
14      biochemical and toxicological actions. These data provide the scientific basis for the TEF/TEQ
15      methodology.  In its 20-year history, the approach has evolved, and decision criteria supporting
16      the scientific judgment and expert opinion used in assigning TEFs has become more transparent.
17      Numerous states, countries, and several international organizations have evaluated and adopted
18      this approach to evaluating complex mixtures of dioxin and related compounds (Part II, Chapter
19      9, Section 9.2). It has become the accepted methodology, although the need for research to
20      explore alternative approaches is widely endorsed. Clearly, basing risk on TCDD alone or
21      assuming all chemicals are equally potent to TCDD is inappropriate oh the basis of available
22      data. Although uncertainties in the use of the TEF methodology have been identified and are
23      described later in this document and in detail in Part II, Chapter 9, Section 9.5, one must examine
24      the use of this method in the broader context of the need to evaluate the potential public health
25      impact of complex mixtures of persistent, bioaccumulative chemicals. It can be generally
26      concluded that the use of TEF methodology for evaluating complex mixtures of dioxin-like
27      compounds decreases the overall uncertainties in the risk assessment process as compared to
28      alternative approaches. Use of the latest consensus values for TEFs assures that the most recent
29      scientific information informs this "useful, interim approach" (U.S. EPA, 1989a; Kutz et al,
30      1990) to dealing with complex environmental mixtures of dioxin-like compounds. As stated by
31      the U.S. EPA Science Advisory Board (U.S. EPA, 1995), "The use of the TEFs as a basis for
32      developing an overall index of public health risk is clearly justifiable, but its practical application
33      depends on the reliability of the TEFs and the availability of representative and reliable exposure
34      data." EPA will continue to work with the international scientific community to update these
35      TEF values to assure that the most up-to-date and reliable data are used in their derivation and to
36      evaluate their use on a periodic basis.
        9/22/00
DRAFT—DO NOT CITE OR QUOTE

-------
       A chemical is assigned a TEF value based on all the available data comparing the
 chemical to either TCDD or PCB 126.  In addition, there are weighting criteria that place more
 emphasis on chronic and subchronic studies examining toxic endpoints (van den Berg et al.,
 1998). There is a broad range in the quantity and quality of the data available for individual
 congeners. For example, the TEF for PCB 126 is based on over 60 in vivo endpoints examining
 responses as diverse as enzyme induction, developmental toxicity, immunotoxicity, hepatic
 toxicity, alterations in hormones and tumor promotion, while the TEF for 3,4,4',5-
 tetrachlorobiphenyl (PCB 81) is based on in vitro CYP1A induction and QSAR calculations.
 Fortunately, PCB 81 does not significantly contribute to human TEQ exposures. There are 5
 congeners that contribute approximately 80% of the total TEQ in humans: 2,3,7,8-TCDD,
 1,2,3,7,8-PCDD, 1,2,3,6,7,8-HxCDD, 2,3,4,7,8-PCDF, and PCB 126 (See Part I, Volume 3 and
 Section 4.4.3 of this document). With the exception of 1,2,3,6,7,8-HxCDD, the TEFs for these
 chemicals are based on a number of different endpoints from multiple studies performed in
 different laboratories (Table 1-4). The TEF for 1,2,3,6,7,8-HxCDD is based on a two-year
 bioassay in which rats were exposed to a mixture of 1,2,3,6,7,8-HxCDD and 1,2,3,7,8,9-HxCDD.
 The TEFs for 2,3,4,7,8-PCDF and PCB 126 are similar to the mean REP value for all in vivo
 endpoints and are similar to their REPs for tumor promotion. The TEF for 12378-PCDD is
 based largely on its REP for tumor promotion hi rats. From these data, it is clear that the
 chemicals that contribute approximately 80% to the total human TEQ are well studied and the
 assigned TEFs provide reasonable estimates of the relative potency of these chemicals.  In
 contrast, while there are some chemicals in the TEF methodology which have minimal data sets
to reliably assess their relative potency, these chemicals do not contribute substantially to the
human blood TEQ.
       The ability of the TEF methodology to predict the biological effects of mixtures
containing dioxin-like chemicals has been evaluated in a number of experimental systems. These
studies generally demonstrate that the assumption of additivity provides a reasonable estimate of
the  dioxin-like potential of a mixture (Part II, Chapter 9, Section 9.4). In addition, there are
examples of non-additive interactions between dioxins and non-dioxins. Both greater than
additive and less than additive interactions have been observed in these studies. In general the
non-additive interactions between the dioxins and non-dioxins have been observed at doses that
are  considerably higher than present background human exposures (Part II, Chapter 9, Section
9.4).
       There are a number of natural chemicals that bind and activate the AhR and induce some
dioxin-like effects.  It has been proposed by some scientists that these chemicals contribute
significantly to the total TEQ exposures and that these exposures far out weigh those from
PCDDs, PCDFs and PCBs (Safe, 1995a). While this hypothesis is intriguing, there are several
9/22/00
DRAFT—DO NOT CITE OR QUOTE

-------
 1      limitations to these analyses (Part II, Chapter 9,Section 9.3.)-  The in vivo data on the natural
 2      AhR ligands is limited to enzyme induction and a single developmental study. Few, if any,
 3      toxicology studies demonstrating clear dioxin-like toxicities have been published. The natural
 4      AhR ligands are rapidly metabolized and result in both transient tissue concentrations and
 5      transient effects.  The natural ligands also have significant biological effects that are independent
 6      of the AhR and it is not clear as to the role of the AhR in the biological effects of these
 7      chemicals. Clearly this issue requires further research in order to better understand the relative
 8      potential health effect of dioxin and related chemicals as compared to natural AhR ligands.
 9            One of the limitations of the use of the  TEF methodology in risk assessment of complex
10      environmental mixtures is that the risk from non-dioxin-like chemicals is not evaluated in
11      concert with that of dioxin-like chemicals.  Another limition of the TEF methodology is their
12      application to non-biological samples.  The fate and distribution of PCDDs, PCDFs and PCBs
13      are not necessarily related to their TEF. Thus,  the use of the TEF for non-biological media must
14      be done cautiously. Future approaches to the assessment of environmental mixtures should focus
15      on the development of methods that will allow risks to be predicted when multiple mechanisms
16      are present from a variety of contaminants.
17
18      1.3. UNDERSTANDING EXPOSURE/DOSE RELATIONSHIPS FOR DIOXIN-LIKE
19      COMPOUNDS
20            Risk assessment requires  the scaling of exposure/dose across endpoints and across
21      species. Given the many responses to TCDD and its congeners, the selection of dose metrics for
22      use in quantitative risk assessments is a complex problem. The biochemical and toxicological
23      responses of TCDD and related chemicals are initiated by their interaction with the Ah receptor.
24      Some responses, such as enzyme induction, require short periods (minutes to hours) of Ah
25      receptor activation. Other responses, such as cancer, require prolonged (months to many years)
26      activation of this pathway. Still other responses, such as the developmental toxicities, require
27      receptor activation during specific windows of sensitivity. Because of the different mechanisms
28      involved in these diverse responses, it is unlikely that a single dose metric will be adequate for all
29      of these endpoints. A number  of studies have proposed a variety of dose metrics for a number of
30      different responses. These studies have taken different approaches ranging from simple curve
31      fitting exercises (Hurst et al., 2000; van Birgelen et al., 1996) to more complex PBPK modeling
32      approaches (Jusko et al., 1995; Andersen et al,. 1997; Kohn et al., 1993; Portier and Kohn, 1996).
33      Area under the curve (AUC) 'has been used traditionally in the drug literature as a dose metric of
34      choice when dose and time related to effects in humans are known.
35             The choice of dose metric not only considers mechanistic data but must also consider
36      pragmatic approaches as well.  The use of the dose metric plays a role in its choice. Because of
        9/22/00
8
DRAFT—DO NOT CITE OR QUOTE

-------
differences in life-span and uncertainties in the windows of sensitivity for various endpoints,
AUC may not be a useful dose metric for cross species extrapolation in the risk assessment of
dioxin and related compounds. However, AUC has been used in the analysis of human cancer
data on TCDD (Becher et al., 1998) and may be a useful dose metric when applied to accidental
or occupational exposures since cross species scaling is not required. The choice of dose metric
is also dependent upon the data available.  A number of dose metrics, such as Ah receptor
occupancy, induction of CYP1A2, and decreases in EOF receptor have been proposed based on
PBPK models (Jusko et al., 1995; Andersen et al.,1997; Kohn et al., 1993; Portier and Kohn,
1996). While these dose metrics have been useful in hypothesis testing in experimental systems,
they are not useful in animal to human extrapolations due to the difficulty in measuring these
parameters in humans. In the following section, the strengths and weaknesses of a variety of
proposed dose metrics will be presented.

1.3.1. Administered Dose
      In experimental studies, animals are administered a defined dose through a variety of
routes. A default method used by EPA (U.S. EPA, 1992; 1996) to estimate the human equivalent
dose when scaling across species is to use allometric scaling based on the following equation:
                                                          ,0.25
                          Dosehuman = Doserat (BWrat/BWhum J

where BW is the body weight in kg and Dose is the daily administered dose in rats or the scaled
human daily dose expressed as ng/kg/d. This method is thought to scale administered dose in
such a way as to result in equivalent effective doses in humans and experimental animals (U.S.
EPA, 1992), taking both pharmacokinetics and pharmacodynamics into account. Using this
equation, a dose of 1 ng TCDD/kg/d in a 0.35 kg rat would result in a scaled human dose of 0.27
ng TCDD/kg/d for a 70 kg human. If this scaling method applies to TCDD and related
chemicals, then 1 ng TCDD/kg/d in the rat should produce similar effective doses in a human
exposed to 0.27 ng TCDD/kg/d, some 3.8 times lower. Assuming similar sensitivity between
rats and humans at the tissue level, effective doses should be a function of tissue concentration.
Tissue concentrations of TCDD and related chemicals are directly related to the concentration of
TCDD in the body. The steady-state concentration of TCDD in the body, or steady-state body
burden, can be estimated in rats and humans using the following equation.
9/22/00
DRAFT—DO NOT CITE OR QUOTE

-------
  1      Steady-state body burden (ng/kg) = [Dose fag TEO/kg^half-life Tdavs^] * f
  2                                                     Ln(2)
  3
  4      where Dose is the daily administered dose, F is the fraction absorbed, and t/2 is the species-
  5      specific half-life of TCDD. In the present example, we will assume F is 50% and the species
  6      specific half-life of TCDD is 25 days for rats and 2593 days for humans.  Starting with an
  7      administered dose of 1 ng/kg/d in rats and the scaled human dose of 0.27 ng/kg/d, the steady-
  8      state body burdens are presented hi Table 1-5. The steady-state body burden of TCDD using the
  9      scaled human dose is approximately 28 times that of the steady-state body burden in the rat
10      (Table 1-5). Using the equation above to estimate equivalent steady state body burdens (i.e. 18
11      ng/kg), a human equivalent administered dose comparable to 1  ng/kg/day administered to the rat
12      was estimated at 0.0096 ng/kg/d, over 100 times less.
13             Clearly, the default scaling method results in an estimated human equivalent dose that
14      produces much greater estimated human tissue concentrations (505 ng/kg) than the rat's tissue
15      concentration (18 ng/kg). One reason for the discrepancy of the scaling method is  that the half-
16      life of TCDD in rodents and humans is much longer than is typically observed for other
17      xenobiotics (Bachmann et al., 1996).  The default scaling approach accounts for a difference of
18      3.7 tunes based on allometric considerations, yet the half-life of TCDD in humans  alone is
19      approximately 100 fold greater than in rats.  This exercise suggests that administered dose may
20      not provide a useful dose metric for cross species extrapolation even if the dose is scaled using a
21      the EPA default methodology. However, administered dose can be used to compare exposures
22      between human populations in order to describe potential human health risks, because the species
23      differences in half-life would not exist in this case.
24
25      1.3.2. Area Under the Curve
26            Area under the curve or AUC is frequently used as a dose metric for reversible responses
27      of pharmaceutical agents. Typically, these agents have half-lives on the order of minutes to
28      hours. In addition, the pharmacological actions of the drug and the length of time of the response
29      is clearly defined in both animals and humans. For example, for anesthetics, sleep-time is used
30      as the length of tune for determining the AUC. In essence, plasma concentrations  are readily
31      determined and the tune span is easily defined.  Mechanistic considerations also suggest that
32      AUC can be a useful dose metric for carcinogenesis. TCDD and related chemicals are thought to
33      induce tumors through promotional mechanisms as opposed to  acting as initiators. The
34      promotional effects of TCDD and related chemicals are associated with altered gene expression
35      resulting in alterations in growth and differentiation. This promotional process requires
36      sustained tissue  concentrations of TCDD sufficient to maintain increased gene expression. It is
        9/22/00
10
DRAFT—DO NOT CITE OR QUOTE

-------
 1      likely that AUC would be an appropriate dose metric for cancer in humans, and may also involve
 2      the incorporation of a threshold concentration (Hays et al., 1997). However, the use of AUC for
 3      species extrapolation for TCDD is more complicated. While blood or plasma concentrations of
 4      TCDD can be determined in both humans and animals, the determination of the time span for
 5      which the AUC is to be calculated is much less certain.  For some of the toxic responses to
 6      TCDD, the window of sensitivity is clearly defined in rodents and humans, such as induction of
 7      cleft palate. For other responses, such as the developmental reproductive alterations observed in
 8      male rats, the window of sensitivity has been  narrowed to exposures between gestational day 15
 9      and 20 in the rats, but the human window of sensitivity is uncertain. For carcinogenesis, the
10      length of time required to induce the response remains uncertain in both experimental animals
11      and humans. In order to apply AUC for species comparisons of the sensitivity to TCDD, one
12      must have a better understanding of the species differences in the windows of sensitivity to the
13      various biological effects of TCDD.
14            In addition,  differences in life-span also must be considered. Brody and Reid (1967)
15      proposed that the biological activity of a drug is related to its plasma concentrations. If animals
16      and humans had the same plasma concentrations for their entire lives, the human AUC would be
17      greater because humans have a longer half-life.  However, because the plasma concentrations
18      were the same, according to Brody and Reid (1967), the responses should be similar. Hence, in
19      order to use AUC for chronic toxicities, such  as cancer, a correction for the difference in life-
20      span must be applied. Typically, this involves the derivation of a lifetime average serum lipid
21      concentration (Cavg), which is calculated by dividing the AUC by the time period of exposure
22      (Aylward et al., 1996). An estimation of the average daily AUC is directly related to steady-state
23      body burdens.  Hence, once the AUC is corrected for life-span differences, these values are
24      equivalent to steady-state body burdens.
25            While AUC may not be an appropriate dose metric for animal to human extrapolations, it
26      is a useful tool for comparing populations exposed to high concentrations of dioxins over a short
27      period of time to the background population.  Becher et al. (1998) successfully used this
28      approach to examine dose response relationships for cancer in an occupationally exposed cohort.
29      One difficulty in determining AUC is the accuracy of the intake measurements.  Past exposures
30      through the diet are uncertain, although they have been estimated (Pinsky and Lorber, 1998).
31      Future exposures are thought to be decreasing, although the exact magnitude of this decrease is
32      uncertain. Hence, determination of AUC carries a number of uncertainties that must be
33      considered.
34
        9/22/00
11
DRAFT—DO NOT CITE OR QUOTE

-------
 1      1.3.3. Plasma or Tissue Concentrations
 2            Brodie and Reid (1967) have argued that the response to a drug is determined by the
 3      amount bound to its biological receptor and since the drug-receptor complex is in dynamic
 4      equilibrium with the free drug in the plasma, the biological response of a drug will be related to
 5      its plasma concentrations. There is no reason to believe that this relationship will not be true for
 6      TCDD and related chemicals. However, there are several data gaps that may prohibit the use of
 7      plasma or blood concentrations for species extrapolation.  First, few animal studies determined
 8      blood or plasma concentrations of TCDD, particularly in the subchronic, chronic and lifetime
 9      exposures. PBPK models can be used to estimate blood concentrations and should provide
10      reasonable estimates of these values.  In contrast, the human exposure data is based
11      predominately on blood, serum or plasma dioxin concentrations. One limitation of the human
12      data is that it is mostly presented on a lipid adjusted basis.  Hence in order to compare the human
13      and animal plasma or blood concentrations, one would have to first estimate the blood
14      concentrations in the animals using a PBPK model. Then either the animal data would have to
15      be expressed as a lipid basis or the human data would have to be expressed as a wet weight basis.
16      In either case, assumptions of the percent lipid in the blood would have to be applied as well as a
17      number of assumptions used in the PBPK models.
18            The use of tissue concentrations as a dose metric has been examined by van Birgelen et
19      al. (1996) and Hurst et al. (1999).  van Birgelen and coworkers (1996) presented data
20      demonstrating that target tissue concentrations provided an accurate prediction of enzyme
21      induction regardless of the exposure scenario (i.e. acute vs subchronic)." Similarly, Hurst et al.
22      (1999) presented data demonstrating that fetal  tissue concentrations of TCDD on gestation day
23      16 predicted decreases in sperm counts, delays in puberty in males, urethra-phallus distance and
24      the incidence of vaginal threads in rats prenatally exposed to TCDD on either gestational day 9 or
25      15.  These data suggest that target tissue concentrations may be a reasonable dose metric for
26      these responses.
27            While target tissue concentrations may aid in estimating risks, these data are unlikely to
28      be collected in humans hi sufficient numbers to be useful, particularly for fetal  concentrations.
29      Plasma concentrations are also a useful tool to compare exposures  in different human
30      populations. Application of plasma concentration as a dose metric for species extrapolation
31      requires some level of assumptions as described above, but reasonable comparisons could be
32      made, particularly for comparing steady-state in humans and animals.  Comparing plasma or
33      blood concentrations following acute exposures in experimental animals to  steady-state human
34      blood or plasma concentrations would not be appropriate. One limitation of the use of either
35      plasma, blood or target tissue concentrations as dose metrics is the lack of human PBPK models
36      to predict these values based on changes in intake patterns.
        9/22/00
12
DRAFT—DO NOT CITE OR QUOTE

-------
1.3.4. Steady-State Body Burdens
       Body burden is defined as the concentration of TCDD and related chemicals in the body
and is typically expressed as ng/kg body weight. In aninials, these values are calculated from
studies at or approaching steady-state and are associated with either biochemical or toxicological
responses.  In addition, these values are calculated based on either knowledge of the species-
specific half-life and the exposure or they are estimated based on the TCDD tissue concentration,
the size of the tissues and the weight of the animal. In humans the values are typically presented
as steady-state body burdens and are estimated based on an intake rate and the half-life of TCDD
in humans. Alternatively, body burdens in humans are estimated based on lipid adjusted serum
or adipose tissue TCDD  or TEQ concentrations (See Part I, Volume 3, Chapter 4).
       Steady-state body burdens provide a useful dose metric for several reasons.  First, tissue
and blood concentrations are directly related to body burdens. Thus, body burdens are surrogates
for tissue concentrations. Second, the differences in the half-life of TCDD between species is
accounted for because these body burdens are estimated at steady-state conditions. Third,
DeVito et al. (1995) have demonstrated that for some biochemical responses, chloracne and
cancer, species have similar rates of responses when dose is expressed on a body burden basis.
Finally, body burdens provide flexibility because they can be estimated based on either intake
rates or on measured tissue concentrations.
      Body burdens also have some limitations.  In order to estimate body burdens from lipid
adjusted tissue concentrations, an assumption of the percent body fat must be used.  In the
reassessment, a value of 25% has been used. It should be noted-that there are human populations
with body fat compositions less than 10% and greater than 35%. Also, when estimating the body
burden based on intake rates and half-lives, the uncertainty of these parameters should be
considered.  In the reassessment, the estimated steady-state body burden of approximately 5 ng
TEQDFP-WHO98/kg is based on measured serum concentrations from several populations in the
mid 1990's.  While measured concentrations should eliminate some of the uncertainties around
estimates using intake rates and half-life assumptions, it is likely that these measured values
represent a past history of higher exposure and we must anticipate a continued downward trend
to represent a "true" lifetime average concentration. Caution must be used when using body
burden as a dose metric for species extrapolation when comparing short-term animal studies to
steady-state human exposures.  Under experimental conditions in the animals, the relationship
between tissue concentrations and body burden may not be the same as under the steady-state
conditions.
9/22/00
13
DRAFT—DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
1.3.5. Mechanistic Dose Metrics
       Several groups have proposed a variety of dose metrics based on mechanistic
considerations, such as concentration of occupied AhR (Jusko, 1995), induced CYP1A2
(Andersen et al., 1997; Kohn et al., 1993) and reduced epidermal growth factor receptor (EGFR)
(Portier and Kohn, 1996). While these dose metrics are intellectually  appealing, it must be kept
in mind that they are still hypothesized dose metrics and require further research to demonstrate
their utility for cross-species extrapolations. In addition, these dose metrics are unlikely to be
measured in sufficient human samples to be useful.

1.3.6. Summary
       A variety of dose metrics have been proposed for estimating potential human health
effects following exposure to dioxins. Many of these dose metrics have limitations that prohibit
their use, such as tissue concentrations and the mechanistic dose metrics. Other dose metrics,
such as AUC have limited utility for species extrapolations because of our limited understanding
of the concept of physiological time. Some dose metrics can be used to compare different human
exposures, such as AUC and administered dose, but are not necessarily suitable for species
extrapolations. Other dose metrics, such as steady-state body burdens or blood concentrations
are useful dose metrics for species extrapolations because of they are directly related to tissue
concentrations, and can be estimated in both animals and humans. The use of any of these dose
metrics requires a number of assumptions discussed above. The choice of dose metric requires
an understanding of the data available and their application in the intended use of the dose
metric.  Future research efforts on the issue of dose metrics could provide better guidance in
choosing the dose metrics for dioxins and related chemicals. However, in the mean time, the use
of steady-state body burdens can provide a reasonable description of dose for use in species
extrapolations and risk assessments.
                               2. EFFECTS SUMMARY

       Since the identification of 2,3,7,8-TCDD as a chloracnegen in 1957, more than 5,000
publications have discussed its biological and toxicological properties. A large number of the
effects of dioxin and related compounds have been discussed in detail throughout the chapters in
Part II of this assessment. They illustrate the wide range of effects produced by this class of
compounds. The majority of effects have been identified in experimental animals; some have
also been identified in exposed human populations.
        9/22/00
                                           14
DRAFT—DO NOT CITE OR QUOTE

-------
 1            Cohort and case-control studies have been used to investigate hypothesized increases in
 2      malignancies among the various 2,3,7,8-TCDD-exposed populations (Fingerhut et al., 1991a, b;
 3      Steenland et al., 1999; Manz et al., 1991; Eriksson et al., 1990).  Cross-sectional studies have
 4      been conducted to evaluate the prevalence or extent of disease in living 2,3,7,8-TCDD-exposed
 5      groups (Suskind and Hertzberg, 1984; Moses et al., 1984; Lathrop et al., 1984, 1987; Roegner et
 6      al., 1991; Grubbs et al., 1995; Sweeney et al., 1989; Centers for Disease Control (CDC) Vietnam
 7      Experience Study, 1988; Webb et al., 1989; Ort and Zober, 1994). The limitations of the cross-
 8      sectional study design for evaluating hazard and risk are discussed in Part II, Chapter 7b, Section
 9      7.11. Many of the earliest studies were unable to define exposure-outcome relationships owing
10      to a variety of shortcomings, including small sample size, poor participation, short latency
11      periods, selection of inappropriate controls, and the inability to quantify exposure to
12      2,3,7,8-TCDD or to identify confounding exposures.  In more recent analyses of cohorts
13      (Fingerhut et al., 1991; Ort and Zober, 1996; Flesch-Janys et al., 1998), cross-sectional studies of
14      U.S. chemical workers (Sweeney et al., 1989), U.S. Air Force Ranch Hand personnel (Roegner et
15      al., 1991; Grubbs et al., 1995), and Missouri residents (Webb et al., 1989), serum or adipose
16      tissue levels of 2,3,7,8-TCDD were measured to evaluate 2,3,7,8-TCDD-associated effects in
17      exposed populations.  The ability to measure tissue or serum levels of 2,3,7,8-TCDD for all or a
18      large sample of the subjects confirmed exposure to 2,3,7,8-TCDD and permitted the investigators
19      to test hypothesized dose-response relationships.               '....-
20            A large number of effects of exposure to TCDD and related compounds have been
21      documented in the scientific literature. Although many effects have been demonstrated hi
22      multiple species (see Table 2-1), other effects may be specific to the species in which they are
23      measured and may have limited relevance to the human situation. Although the potential
24      species-specific responses are an important consideration for characterizing potential hazard, all
25      the observed effects of 2,3,7,8-TCDD illustrate the multiple sequelae that are possible when
26      primary impacts are at the level of signal transduction and gene transcription. Even though not
27      all observed effects may be characterized as "adverse" effects (i.e., some may be responses
28      within the normal range, adaptive' or compensatory and of unknown or neutral consequence),
29      they represent a continuum of response expected from the fundamental changes in biology
30      caused by exposure to dioxin-like compounds.  As discussed in the following sections, the dose
31      associated with this plethora of effects is best compared across species using a common
32      measurement  unit of steady-state body burden of 2,3,7,8-TCDD and other dioxin-like
33      compounds, as opposed to the level or rate of exposure/intake. These comparisons result in the
34      finding that, when animal data associated with effects at the low end of the range of experimental
35      observation (NOAELs/LOAELs/ED01s) are compared to current average human body burdens of
36      approximately 5 ng TEQDFP-WHO98/kg, relatively small margins of exposure (MOE) are
        9/22/00
15
DRAFT—DO NOT CITE OR QUOTE

-------
  1      obtained. Similarly, some human noncancer effects (e.g., developmental delay, neurobehavioral
  2      outcomes and impact on thyroid function in Dutch children) and cancer outcomes show
  3      comparatively small MOEs. This concept is illustrated in Table 2-2. This point will be
  4      discussed further in Sections 5.2 and 6.0 in this document.
  5             The effects discussed in the following sections are focused on development of an
  6      understanding of dioxin hazard and risk.  This discussion is by its nature selective of findings
  7      that inform the risk assessment process.  Readers are referred to the more comprehensive  ,
  8      chapters for further discussion of the epidemiologic and toxicologic database.
  9
10      2.1.  BIOCHEMICAL RESPONSES (Cross reference: Part II, Chapters 2,3, and 8)
11             As described later hi Section 3, mechanistic studies can reveal the biochemical pathways
12      and types of biological events that contribute to adverse effects from exposure to dioxin-like
13      compounds.  For example, much evidence indicates that 2,3,7,8-TCDD acts via an intracellular
14      protein (the aryl hydrocarbon receptor [AhR]), which is a ligand-dependent transcription factor
15      that functions in partnership with a second protein (known as the Ah receptor nuclear
16      translocator, Arnt) to alter gene expression.  In addition, receptor binding may result in release of
17      cytoplasmic proteins which, hi turn, alter the expression or activity of cell regulatory proteins
18      (e.g. increases in Src activity). Therefore, from a mechanistic standpoint, TCDD's adverse
19      effects appear likely to reflect alterations in  gene expression or protein activity that occur at an
20      inappropriate time and/or for an inappropriate length of time. Mechanistic  studies also indicate
21      that several other proteins (e.g. hif a, KB, sim, etc.) contribute to TCDD's gene regulatory effects
22      and that the response to 2,3,7,8-TCDD involves a relatively complex interplay between multiple
23      genetic and environmental factors.  This model is illustrated in Figure 2-1 (from Part II, Chapter
24      2).
25             Comparative data from animal and human cells and tissues suggest  a strong qualitative
26      similarity across species in response to dioxin-like chemicals. This further  supports the
27      applicability to humans of the generalized model  of initial events in response to dioxin exposure.
28      These biochemical and biological responses are sometimes considered adaptive, or reflective of
29      exposure to dioxin-like compounds but within normal homeostatic limits and, therefore, are often
30      not considered adverse in and of themselves. However, many of these biochemical changes are
31      potentially on a continuum of dose-response relationships which leads to adverse responses and,
32      considering the potential to shift population distributions in response, may be of concern. At this
33      time, caution must be used when describing these events as adaptive.
34             If, as we can infer from the evidence, 2,3,7,8-TCDD and other dioxin-like compounds
35      operate through these mechanisms, there are constraints on the possible models that can plausibly
36      account for dioxin's biological effects and also on the assumptions used during the risk
        9/22/00
16
DRAFT—DO NOT CITE OR QUOTE

-------
assessment process. For instance, the linear relationship expected between ligand concentration
and receptor binding may or may not be reflective of dose-response relationships for downstream
events requiring complex interactions of other regulatory proteins with the activated receptor.
Mechanistic knowledge of dioxin action may also be useful in other ways. For example,
knowledge of genetic polymorphisms that influence 2,3,7,8-TCDD responsiveness may also
allow the identification of individuals at particular risk from exposure to dioxin.  In addition,
knowledge of the biochemical pathways that are altered by dioxin-like compounds may help in
the development of drugs that can prevent dioxin's adverse effects.
       As described in Part II, Chapter 2, biochemical and genetic analyses of the mechanisms
by which dioxin modulates particular genes have revealed the outline of a novel regulatory
system whereby a chemical signal can alter cellular regulatory processes.  Future studies of
dioxin action have the potential to provide additional insights into mechanisms of mammalian
gene regulation that are of relatively broad interest.  Additional perspectives on dioxin action can
be found in several recent reviews (Birnbaum, 1994a,b; Schecter, 1994; Hankinson, 1995;
Schmidt and Bradfield, 1996; Rowlands and Gustafsson, 1997; Gasiewicz, 1997; Hahn, 1998;
Denison et al, 1998; Wilson and Safe, 1998).
       The ability of 2,3,7,8-TCDD and other dioxin-like compounds to modulate a number of
biochemical parameters in a species-, tissue-, and temporal-specific manner is well recognized.
Despite the ever-expanding list of these responses over the past 20 years and the elegant work on
the molecular mechanisms mediating some of these, there still exists a considerable gap between
our knowledge of the biochemical changes and the degree to which they are related to the more
complex biological and toxicological endpoints elicited by these chemicals. A framework for
considering these responses in a mode of action context is discussed later in this  document.
       TCDD-elicited activation of the Ah receptor has been clearly shown to mediate altered
transcription of a number of genes, including several oncogenes and those encoding growth
factors, receptors, hormones, and drug-metabolizing enzymes. Figure 2-2 provides an
illustrative list of gene products whose regulation or activity is modulated by 2,3,7,8-TCDD.
Although this list is not meant to be exhaustive, it demonstrates the range of potential dioxin
impacts on pathways with potential to lead to adverse effects.
       As discussed in Part II, Chapter 2, it is possible that the TCDD-elicited alteration of
activity of these genes may occur through a variety  of mechanisms. The transcription of some
genes may be directly regulated by the activated AhR. Other alterations in gene  expression may
be secondary to the initial biochemical events directly regulated transcriptionally by the AhR.
Some of the changes may also occur by post-transcriptional processes such as messenger
ribonucleic acid (mRNA) stabilization or altered protein phosphorylation (Gaido et al., 1992;
Matsumura, 1994).  Thus, the molecular mechanisms by which many, if not most, of the
9/22/00
17
DRAFT—DO NOT CITE OR QUOTE

-------
  1      biochemical processes discussed herein are altered by 2,3,7,8-TCDD treatment remain to be
  2     determined. Nevertheless, it is presumed, based on the cumulative evidence available, that all of
  3     these processes are mediated by the binding of 2,3,7,8-TCDD to the AhR. Although the
  4     evidence for the involvement of the AhR in all of these processes has not always been
  5     ascertained, structure-activity relationships, genetic data, and reports from the use of biological
  6     models like "knockout" mice that are lacking the AhR (AhR"'") are consistent with the
  7     involvement of the AhR as the initial step leading to many of these biochemical alterations. In
  8     fact, for every biochemical response that has been well studied, the data are consistent with the
  9     particular response being dependent on the AhR.
10           The dioxin-elicited induction of certain drug-metabolizing enzymes such as CYP1 Al,
11      CYP1A2, and CYP1B1 is clearly one of the most sensitive responses observed in a variety of
12     different animal species including humans, occurring at body burdens as low as 1-10 ng
13     TCDD/kg hi animals (see Part II, Chapter 8, Sections 8.3 and 8.4). These and other enzymes are
14     responsible for the metabolism of a variety of exogenous and endogenous compounds. Several
15     lines of experimental evidence suggest that these enzymes may be responsible for either
16     enhancing or protecting against the toxic effects of a variety of agents, including known
17     carcinogens as well as endogenous substrates such as hormones. These interactive effects are
18      dependent upon the compounds and the experimental system examined. Several reports
19      (Kadlubar et al., 1992; Esteller et al., 1997; Ambrosone et al., 1995;  Kawajiri et al., 1993)
20      provide evidence that human polymorphisms  in CYPIA1 and CYPIA2 that result in higher levels
21      of enzyme activity are associated with increased susceptibility to colorectal, endometrial, breast,
22      and lung tumors. Also, exposure of AhR-deficient ("knockout") mice to benzo[a]pyene (BaP)
23      results in no tumor response, suggesting a key role for the AhR, and  perhaps, CYPIA1 and
24      CYPIA2, in BaP carcinogenesis (Dertinger et al., 1998; Shimizu et al., 2000). Modulation of
25      these enzymes by dioxin may play a role in chemical carcinogenesis. However, the exact
26      relationship between the induction of these enzymes and any toxic endpoint observed following
27      dioxin exposure has not been clearly established.
28            In contrast to what is known about the P450 isozymes (CYP1 Al, CYP1A2, and
29      CYP1B1), there exists some evidence from experimental animal data to indicate that the
30      alteration of certain other biochemical events  might have a more direct relationship to sensitive
31      toxic responses observed following TCDD exposure. Some of these may be relevant to
32      responses observed in humans, and further work in these areas  is likely to lead to data that would
33      assist in the risk characterization process.  For example, changes in EGFR have been observed in
34      tissues from dioxin-exposed animals and humans (see Part II, Chapter 3, Section 3.5 and Chapter
35      6, Section 6.5 ). EOF and its receptor possess diverse functions relevant to cell transformation
36      and tumorigenesis, and changes  in these functions may be related to  a number of dioxin-induced
        9/22/00
18
DRAFT—DO NOT CITE OR QUOTE

-------
 1      responses including neoplastic lesions, chloracne, and a variety of reproductive and
 2      developmental effects. Likewise, the known ability of TCDD to directly or indirectly alter the
 3      levels and/or activity of other growth factors and hormones, such as estrogen, thyroid hormone,
 4      testosterone, gonadotropin-releasing hormone and their respective receptors,  as well as enzymes
 5      involved in the control of the cell cycle (Safe, 1995b), may affect growth patterns in cells/tissues,
 6      leading to adverse consequences. In fact, most of the effects that the dioxins produce at the
 7      cellular and tissue levels are due not to cell/tissue death but to altered growth patterns (Birnbaum,
 8      1994b). Many of these may occur at critical times in development and/or maturation and thus
 9      may be irreversible.
10            There does not yet exist a precise understanding of the relationships between the
11      alteration of specific biochemical processes and particular toxic responses observed in either
12      experimental animals  or humans exposed to the dioxins.  This is due predominantly to our
13      incomplete understanding of the complex and coordinated molecular, biochemical, and cellular
14      interactions that regulate tissue processes during development and under normal homeostatic
15      conditions. A further understanding of these processes and how 2,3,7,8-TCDD may interfere
16      with them remains an important goal that would greatly assist in the risk characterization process.
17      In particular, knowledge of the causal association of these responses coupled with dose-response
18      relationships may lead to a better understanding of sensitivity to various exposure levels of the
19      dioxin-like compounds. Nevertheless, it is important to recognize that many of the biochemical
20      and biological changes observed are consistent with the notion that 2,3,7,8-TCDD is a powerful
21      growth dysregulator.  This hypothesis may play a considerable role in the risk characterization
22      process by providing a focus on those processes, such as development, reproduction, immunity,
23      and carcinogenesis, that are highly dependent on coordinate growth regulation.
24
25      2.2. ADVERSE  EFFECTS IN HUMANS AND ANIMALS
26      2.2.1. Cancer (Cross Reference: Part II, Chapters 6, 7, and 8)
27      2.2.1.1. Epidenriologic Studies
28            Since the  last formal U.S. EPA review of the human database relating to the
29      carcinogenicity of TCDD and related compounds in 1988, a number of new follow-up mortality
30      studies have been completed. This body of information is described in Part II, Chapter 7a,
31      Section 7.5, of this assessment and has recently been published as part of an International Agency
32      for Research on Cancer (IARC) Monograph (1997) and the Agency for Toxic Substances and
33      Disease Registry  (ATSDR) ToxProfile (ATSDR, 1999a). Among the most important of these are
34      the studies of 5,172 U.S. chemical manufacturing workers by Fingerhut et al. (1991a) and
35      Steenland et al. (1999) from NIOSH and an independent study by Aylward et al. (1996); a study
36      of 2,479 German workers involved in the production of phenoxy herbicides and chlorophenols by
        9/22/00
19
DRAFT—DO NOT CITE OR QUOTE

-------
  1      Becher et al. (1996,1998) and by others in separate publications (Manz et al., 1991; Nagel et al.,
  2      1994; Flesch-Janys et al., 1995,1998); a study of more than 2,000 Dutch workers in two plants
  3      involved in the synthesis and formulation of phenoxy herbicides and chlorophenols (Bueno de
  4      Mesquita et al., 1993) and subsequent follow-up and expansion by Hooiveld et al., 1998); a
  5      smaller study of 247 workers involved in a chemical accident cleanup by Zober et al. (1990) and
  6      subsequent follow-up (Ott and Zober, 1996b); and an international study of more than 18,000
  7      workers exposed to phenoxy herbicides and chlorophenols by Saracci et al. (1991), with
  8      subsequent follow-up and expansion by Kogevinas et al. (1997). Although uncertainty remains
  9      in interpreting these studies because not all potential confounders have been ruled out and
10      coincident exposures to other carcinogens are likely, all provide support for an association
11      between exposure to dioxin and related compounds and increased cancer mortality.  Strong
12      inference regarding carcinogenic hazard 'often relies on the availability of studies with well
13      documented exposures. One of the strengths of these studies is that each has some exposure
14      information that permits an assessment of dose response. Some of these data have, in fact,
15      served as the basis for fitting the dose-response models in Part II, Chapter 8, Section 8.4.
16             In addition, during the development of its monograph on PCDDs/PCDFs (IARC, 1997),
17      the IARC Working Group abstracted, from the published literature, data concerning the most
18      highly exposed populations in the world. They focused their attention on the most exposed
19      subcohorts within cohorts with adequate latency.  IARC suggests that ifassociations between
20      exposure and risk are truly causal, they will become more apparent in these highly exposed
21      subcohorts with adequate latency.  Increased risk for all cancers combined and lung cancer
22      mortality were consistent findings in the occupational cohort studies.  Although the increase was
23      generally low (20%-50%), it was highest in subcohorts with presumed heaviest exposure. The
24      results of the IARC Working Group's analysis regarding all cancer and lung cancer mortality in
25      the recent studies are summarized in Table 2-3.  Observed numbers of cases, standardized
26      mortality ratios (SMRs) and 95% confidence intervals (CI)  are given for each of these two
27      findings for each study. In addition, the Working Group developed overall SMRs for the
28      combined studies. They state clearly that, although these total SMRs are low (1.4, 95% CI, 1.2-
29      1.6 for all cancers and 1.4, 95% CI, 1.1-1.7 for lung cancer), these results are unlikely to be due
30      to chance nor can confounding by cigarette smoking likely account for the increase in lung
31      cancer. Positive dose-response trends in the German studies and increased risk in the longer
32      duration U.S. subcohort and the most heavily exposed Dutch workers support this view. In the
33      opinion of these experts, increases in all cancers combined of this magnitude have rarely been
34      found in occupational cohorts. These results are also supported by significantly increased
35      mortality from lung and liver cancers subsequent to the Japanese rice oil poisoning accident
        9/22/00
20
DRAFT—DO NOT CITE OR QUOTE

-------
 1      where exposure to high levels of PCDFs and PCBs occurred (Kuratsune et al., 1988; Kuratsune,
 2      1989).
 3             While smoking as a confounder cannot be totally eliminated as a potential explanation of
 4      the occupational studies results, analyses (Fingerhut, 1991b; Ott and Zober, 1996b) conducted to
 5      date suggest that smoking is not likely to explain the entire increase in lung cancer and may even
 6      suggest synergism between occupational exposure to dioxin and smoking.  These analyses have
 7      not been deemed entirely satisfactory by some reviewers of the literature.  The question of
 8      confounding exposures, such as asbestos and other chemicals, in addition to smoking, has not
 9      been entirely ruled out and must be considered as potentially adding to the observed increases.
10      Although increases of cancer at other sites (e.g., non-Hodgkin's lymphoma, soft tissue sarcoma,
11      gastrointestinal cancer) have been reported (see Part II, Chapter 7a, Section 7.5), the data for an
12      association with exposure to dioxin-like chemicals  are less compelling, due to the limited
13      numbers of observed tumors at any specific site.
14             Some studies that have been discussed in Part II, Chapter 7a, report little or no increased
15      risk of cancer from exposure to 2,3,7,8-TCDD or its congeners. These studies generally suffer
16      from one or more deficiencies that limit their relevance to providing information that could assist
17      in determining the carcinogenic hazard of dioxins.  These deficiencies fall into the following
18      categories: little statistical power to detect an effect of exposure since the measured exposures
19      are lower than those seen in the studies cited above and more similar to that of the comparison
20      population; no measurements of in vivo exposure to 2,3,7,8-TCDD and potential for
21      misclassification of exposure; and inadequate latency or follow-up. In short, these mostly non-
22      positive studies lack one or more strengths of the cohort studies discussed above.
23             For example, substantial exposures to dioxin were also experienced by U.S. Air Force
24      Ranch Hand personnel spraying the defoliant Agent Orange during the Vietnam war!  In this
25      study, there is no statistically significant increase in all cancers in the exposed population.
26      Statistical power analysis based on the detailed dosimetry and health status data available for this
27      cohort indicates insufficient statistical power to detect an elevated all cancers risk at levels
28      consistent with the occupational dose-response data. Statistical power is the ability of a study to
29      detect a real  difference between two groups at pre-defined levels of statistical significance
30      (usually P<. 0.05) and relative risk. A relative risk for all cancers combined can be estimated for
31      the Ranch Handers by calculating the difference between their dose and that of the control group
32      (mean background of 4.25 ppt TCDD in lipid, Michalek et al., 1998), then multiplying this dose
33      increment by an estimated cancer risk slope factor for TCDD. The median AUC increment value
34      for the overall Ranch Hand group is 468 ngTCDD/kg lipid * years, and for the high dioxin group
35      the median is 2,280 ngTCDD/kg lipid * years (note from Joel Michalek, U.S. Airforce, to Bruce
36      Rodan, U.S. EPA, dated September 8, 2000).  Using the Becher et al. (1998) linear formula (RR
        9/22/00
21
DRAFT—DO NOT CITE OR QUOTE

-------
  1      =1+ 0.000016 x AUC ng-TCDD/kg lipid * Years; ~ 3 x 10'3 risk/pg/kg/day) described in
  2     Section 5.3 and Table 5-4 of this document, the estimated all cancers relative risk for the overall
  3     Ranch Hand cohort is approximately 1.01, and for the high exposure group 1.04 compared to the
  4     control population. Using formulae in Fleiss (1981)  and Cohen (1977), and assuming two-sided
  5     testing at a significance level of 5%, the study has no power to detect 1 to 4 percent increases in
  6     relative risk. Data on the overall prevalence of cancer in the comparison group (18.9%) and
  7     sample sizes (all Ranch Hand 845 v. 1224 controls; high category 241 v. 1200 controls) used in
  8     the above analysis were obtained from the 1997  Ranch Hand morbidity report
  9     Chttp://www.brooks.af.rml/AFRL/HED/hedb/afhs/.html).  The lack of a statistically significant
10     positive response in this study is consistent with the lack of power of this study to detect an
11      increase in all cancer risks, based on observations on cancer risk emerging from the analysis of
12     the more highly exposed occupational cohorts.
13           In addition, one of the earliest reported associations between exposure to dioxin-like
14     compounds in dioxin-contaminated phenoxy herbicides and increased cancer risk involved an
15      increase in soft tissue sarcomas (Hardell and Sandstrom, 1979; Eriksson et al., 1981; Hardell and
16      Eriksson, 1988; Eriksson et al., 1990). In this and other recent evaluations of the epidemiologic
17      database, many of the earlier epidemiological studies that suggested an association between
18      dioxin exposure and soft tissue sarcoma are criticized for a variety of reasons.  Arguments
19      regarding selection bias, lack of exposure or differential exposure misclassification, confounding,
20      and chance in each individual study have been presented in the scientific literature, which
21      increases uncertainty around this association.  Nonetheless, the incidence of soft tissue sarcoma
22      is elevated, but not statistically, in several of the most recent studies (Bertazzi et al., 1993;  1997,
23      1999; Fingerhut et al., 1991a; Hertzman et al., 1997;  Kogevinas et al., 1997; Lampi et al., 1992;
24     Lynge, 1998; Pesatori et al., 1999;  Saracci et al., 1999; Vineis et al., 1986). It is probable that
25      soft tissue sarcomas are not unlike  other site-specific cancers whose risks are difficult to define
26      from exposure to TCDD.
27            The accidental exposure of the population at Seveso serves as  an example of a more
28      highly exposed group where, to date, latency is considered to be inadequate. Although Bertazzi
29      and coworkers have published results of cancer mortality after 10 and 15 years of latency, results
30      are suggestive but not definitive regarding an association between exposure to TCDD and cancer
31      deaths. Results of the analysis of 20 years of follow-up have recently been accepted for
32      publication.
33            As mentioned above, both past and more recent human studies have focused on males.
34     Although males comprise all the case-control studies and the bulk of the cohort study analyses,
35      animal and mechanism studies suggest that males and females might respond differently to
36      TCDD. There are now, however, some limited data suggesting carcinogenic responses
        9/22/00
22
DRAFT—DO NOT CITE OR QUOTE

-------
associated with dioxin exposure in females.  The only reported female cohort with good TCDD
exposure surrogate information was that of Manz et al. (1991), which had a borderline
statistically significant increase in breast cancer.  Although Saracci et al. (1991) did report
reduced female breast and genital organ cancer mortality, this was based on few observed deaths
and on chlorophenoxy herbicide, rather than TCDD, exposures. In the later update and
expansion of this cohort, Kogevinas et al. (1997) provided evidence of a reversal of this deficit
and produced a borderline significant excess risk of breast cancer in females. Bertazzi et al.
(1993, 1997,1998) reported nonsignificant decreases in breast cancer and endometrial cancer in
women living in geographical areas around Seveso contaminated by dioxin. Although
Kogevinas et al. (1993) saw an increase in cancer incidence among female workers most likely
exposed to TCDD, no increase in breast cancer was observed in his small cohort.  In sum, TCDD
cancer experience for women may differ from that of men, but currently there are few data to
adequately address this question.
       Both laboratory animal data and mechanistic inferences suggest that males and females
may respond differently to the carcinogenic effects of dioxin-like chemicals.  Further data will be
needed to address this question of differential response between sexes, especially to hormonally
mediated tumors. In addition, recent studies of Brown et al. (1998) demonstrate that prenatal
exposure of rats to 2,3,7,8-TCDD enhances their sensitivity as adults to chemical carcinogenesis.
The experimental data in laboratory animals  suggest that exposure to women or perinatal
exposures may result in carcinogenic responses.  The epidemiological data examining the
association between exposure of adult women to  dioxin and cancer is limited. No
epidemiological data are available to address the  question of the potential impact of exposure to
dioxin-like compounds on childhood cancers, or the effects of perinatal exposures on the
development of cancers later in life. Presently, the epidemiological data have not adequately
addressed these issues.
       In summary, 2,3,7,8-TCDD and, by inference from more limited data, other dioxin-like
compounds are  described  as potentially multisite carcinogens in the more highly exposed human
populations that have been studied, consisting primarily of adult males. Although the
epidemiologic data are not sufficient by themselves to infer a causal association between
exposure to TCDD and other dioxin-like chemicals and increased cancer in humans (IARC,
1997; ATSDR,  1999a), this "limited" epidemiologic data base has been strengthened by
emerging data reflecting further follow-up and better exposure metrics. Although uncertainty
remains, the cancer findings in the epidemiologic literature are generally consistent with results
from studies of multiple laboratory animal species where dioxin-like compounds have clearly
been identified as multisite carcinogens and tumor promoters.  In addition, the findings of
increased risk at multiple  sites in occupationally exposed humans appear to be plausible given
9/22/00
23
DRAFT—DO NOT CITE OR QUOTE

-------
 1      what is known about mechanisms of dioxin action, and the fundamental level at which this class
 2      of compounds appears to act on gene expression and cellular regulation in target tissues. While
 3      several studies exhibit a positive trend in dose-response and have been the subject of empirical
 4      risk modeling (See Part II, Chapter 8 and Becher et al., 1998), the epidemiologic data alone
 5      provide little insight into the shape of the dose-response curve below the range of observation in
 6      these occupationally exposed populations. This issue will be further discussed in Section 5.2.1
 7      of this document.
 8
 9      2.2.1.2. Animal Carcinogenicity (Cross reference, Part II: Chapters 6 and 8)
10             An extensive database on the carcinogenicity of dioxin and related compounds in
11      laboratory studies exists and is described in detail in Part II, Chapter 6. There is adequate
12      evidence that 2,3,7,8-TCDD is a carcinogen in laboratory animals based on long-term bioassays
13      conducted in both sexes of rats and mice (U.S. EPA, 1985; Huff et al., 1991; Zeise et al., 1990;
14      IARC,  1997). All studies have produced positive results, leading to conclusions that TCDD is a
15      multistage carcinogen increasing the incidence of tumors at sites distant from the site of
16      treatment and at doses well below the maximum tolerated  dose. Since this issue was last
17      reviewed by the Agency in 1988, TCDD has been  shown to be a carcinogen in hamsters (Rao et
18      al., 1988), which are relatively resistant to the lethal effects of TCDD. Other preliminary data
19      have also shown TCDD to be a liver carcinogen in the small fish Medaka-(Johnson et al., 1992).
20      Few attempts have been made to  demonstrate the carcinogenicity of other dioxin-like
21      compounds.  Other than a mixture of two isomersof hexachlorodibenzo-/7-dioxin(HCDDs),
22      which produced liver tumors in both sexes of rats and mice (NTP, 1980) when given by the
23      gavage route, but not by the dermal route in Swiss mice (NTP, 1982a,b) and recent reports from
24      Rozman (Rozman, 1999; Rozman, 2000; Rozman et al., 2000) attributing lung cancer in female
25      rats to gavage exposures of l,2,3,4,6,7,8-heptachlorodibenzo-j>dioxin(HpCDD), neither the
26      more highly chlorinated PCDDs/PCDFs nor the coplanar PCBs have been studied in long-term
27      animal cancer bioassays. The National Toxicology Program (NTP) is currently testing the
28      relative carcinogenic potency of four dioxin-like congeners (PeCDF, PeCDD, and PCB 118 and
29      PCB 126), both alone and in combination. These data, when they are available, should add to
30      our understanding regarding the carcinogenicity of these dioxin-like congeners.
31             TCDD is characterized as a nongenotoxic carcinogen because it is negative in most
32      assays for DNA damaging potential and is a potent "promoter" and a weak initiator or
33      noninitiator in two-stage initiation-promotion (I-P) models for liver and for skin. The liver
34      response is characterized by increases in altered hepatocellular foci (AHF), which are considered
35      to be preneoplastic lesions because increases in AHFs are  associated with liver cancer in rodents.
36      The results of the multiple I-P studies enumerated in Table 6-5 in Part II, Chapter 6, Section 6.3,
        9/22/00
24
DRAFT—DO NOT CITE OR QUOTE

-------
 1     have been interpreted as showing that induction of AHFs by TCDD is dose-dependent (Maronpot
 2     et al., 1993; Teegarden et al., 1999), exposure-duration dependent (Dragan et al., 1992;
 3     Teegarden et al., 1999; Walker et al., 2000), and partially reversible after cessation of treatment
 4     (Dragan et al., 1992; Tritscher et al., 1995; Walker et al., 2000). Other studies indicate that other
 5     dioxin-like compounds have the ability to induce AHFs. These studies show that the compounds
 6     demonstrate a rank-order of potency for AHF induction that is similar to that for CYP1 Al
 7     (Flodstrom and Ahlborg, 1992; Waern et al., 1991; Schrenk et al., 1994). Non-ortho substituted,
 8     dioxin-like PCBs also induce the development of AHFs  according to their potency to induce
 9     GYP 1 Al (Hemming et al., 1995; van der Plas et al., 1999).  It is interesting to note that liver I-P
10     studies carried out in ovariectomized rats demonstrate the influence that the intact hormonal
11     system has on AHF development.  AHF are significantly reduced in the livers of ovariectomized
12     female rats (Graham et al., 1988; Lucier et al., 1991).
13            I-P studies on skin have demonstrated that TCDD is a potent tumor promoter in mouse
14     skin as well as rat liver.  Early studies demonstrated that TCDD is at least two orders of
15     magnitude more potent than the "classic" promoter tetradecanoyl phorbol acetate (TPA) (Poland
16     et al., 1982); that TCDD skin tumor promotion is AhR dependent (Poland and Knutsen,  1982);
17     that TCDD had weak or no initiating activity in the skin system (DiGiovanni et al., 1977); and
18     that TCDD's induction of drug-metabolizing enzymes is associated with both metabolic
19     activation and deactivation of initiating agents as described by Lucier et al. (1979).  More recent
20     studies show that the skin tumor promoting potencies of several dioxin-like compounds reflect
21     relative AhR binding and pharmacokinetic parameters (Hebert et al., 1990).
22 '           Although few I-P studies have demonstrated lung tumors in rats or mice, the study of
23     Clark et al. (1991) is particularly significant because of its use of ovariectomized animals. In
24     contrast to liver tumor promotion,  lung tumors were seen only in initiated (diethylnitrosamine
25     [DEN]), TCDD-treated rats. No tumors were seen in DEN only, TCDD  only, control, or
26     DEN/TCDD intact rats.  Liver tumors are ovary dependent, but ovaries appear to protect against
27     TCDD-mediated tumor promotion in rat lung.  Perhaps use of transgenic animal models will
28     allow further understanding of the complex interaction of factors associated with carcinogenesis
29     in rodents as well, presumably in humans. Several such systems are being evaluated (Eastin et
30     al., 1998; van Birgelen et al., 1999; Dunson et al., 2000).
31            The tumor promoting ability of a number of dioxin-like chemicals have been examined.
32     As discussed in Part II, Chapter 6, Section 6, 1,2,3,7,8-PCDD; 1,2,3,4,6,7,8-HpCDD, 2,3,4,7,8-
33     PCDF, 1,2,3,4,7,8-HCDF, PCB126, and PCB105 all promote the development of AHF within
34     rodent liver suggesting that they are also tumor promoters, like TCDD (For a summary of
35     positive tumor promotion studies for PCDDs and PCDFs in rats, see Part II, Chapter 6, Table 6-
36      5).  In addition, complex mixtures of dioxins and furans and commercial PCB mixtures act as
        9/22/00                                   25       DRAFT—DO NOT CITE OR QUOTE

-------
  1      promoters of liver AHF. For the dioxins, furans, and coplanar PCBs that comprise
  2      approximately 80% of the current, total dioxin/furan TEQ in human blood, are all positive in
  3      either rodent bioassays or rodent liver tumor promotion studies, or mouse skin tumor promotion
  4      studies. These data suggest that while the majority of dioxin-like congeners have not been tested
  5      for carcinogenicity in chronic rodent bioassays, it is likely that those individual congeners and
  6      mixtures of dioxin-like compounds that comprise the majority of the dioxin-like activity in
  7      human tissues are likely to be carcinogenic to rodents.
  8             van den Berg et al. (2000; their Table 1) present a summary of the data relied on by the
  9      European Centre for Environment and Health of the World Health Organization (WHO-ECEH)
10      and the International Programme on Chemical Safety (IPCS) in their joint consensus re-
11      evaluation of the TEFs for PCDDs, PCDFs, and dioxin-like PCBs for mammals.  These TEFs
12      were derived using a tiered approach in which in vivo toxicity data were given more weight than
13      in vitro data, toxicity more than biochemical endpoints, and chronic more than acute data. Table
14      2-4 summarizes the tumor incidence and promotion data that were cited in the development of
15      these TEFsDFP-WHO9g. The data presented are for those congeners that are principal contributors
16      to the background body burden of dioxin TEQs in the United States (see Part II, Chapter 4). For
17      1,2,3,7,8-PeCDF and 2,3,4,7,8-PeCDF, the TEF was used to adjust the dose from the studies of
18      Waern  et al. (1991) and for PCB 126 similar dose adjustments are included from Hemming et al.
19      (1995; their figure 4). For the comparison of TCDD to the HxCDDs, in addition to the NTP
20      studies, (U.S. EPA, 1980) the primary TCDD data points from the Kociba et al. (1978) bioassay
21      were graphed for both the original tumor count data and for the revised tumor counts from
22      Goodman and Sauer (1992). This reflects the contemporaneous performance and analysis of the
23      HxCDD and TCDD bioassays and pathology, and the recognition that the HxCDD pathology has
24      not been re-analyzed. Table 2-4 illustrates the comparability of the TCDD and other congener
25      data sets based on TEFs. This analysis also demonstrates that the development of the TEFs for
26      all of the congeners that contribute substantially to the background dioxin TEQ appropriately
27      reflect either cancer bioassay or tumor promotion data. Furthermore, when one considers the
28      impact  of current TEF values on compounds that made up the majority of the TEQ prior to 1990,
29      it is clear that more than 90% of the TEQ for either dioxins/furans or PCBs was made up of
30      compounds for which the current TEF is supported by data on relative potencies based on a
31      tumor promotion or carcinogenic endpoint.  This point is illustrated in Part II, Chapter 6, Table
32      6-10.
33
        9/22/00
26
DRAFT—DO NOT CITE OR QUOTE

-------
 1      2.2.1.3.  Plausible Mode(s) of Carcinogenic Action
 2            Several potential mechanisms for TCDD carcinogenicity are discussed in Part II, Chapter
 3      6, Section 6.4.  These include oxidative stress, indirect DNA damage, endocrine
 4      disruption/growth dysregulation/altered signal transduction, and cell replication/apoptosis
 5      leading to tumor promotion. All of these are biologically plausible as contributors to the
 6      carcinogenic process and none are mutually exclusive. Several biologically based models that
 7      encompass many of these activities are described in Part II, Chapter 8, Section 8.4. Further work
 8      will be needed to elucidate a detailed mechanistic model for any particular carcinogenic response
 9      in animals or in humans.  However, plausible modes of action with probable relevance to human
10      carcinogenicity are discussed below.
11             TCDD is a potent tumor promoter in rat and mouse liver.  In general terms it is believed
12      that cancer is likely due to the clonal expansion of damaged cells that have a heritable genetic
13      defect. Increased growth and accumulation of damage in critical genes ultimately aid in the
14      progression into tumors.  Consequently, promotion of carcinogenesis by TCDD may therefore
15      occur at several steps:  (1) Increased formation of initiated/susceptible cells through DNA
16      mutation and/or increase rate of fixation of damaged DNA into the genome; (2) Reduced loss of
17      initiated cells through a suppression of apoptosis; (3) Increase in growth rate and clonal
18      expansion of initiated cells; and (4) Accumulation of DNA damage in critical genes  resulting in
19      progression of clonally expanded cell populations into tumors. Within this framework, it is
20      hypothesised that TCDD may be acting as a tumor promoter through multiple mechanisms.
21      Primarily, the activation of the AHR leads to alteration in genes" involved in normal cell growth
22      response pathways.
23            TCDD may contribute to the formation of and accumulation of DNA damage via an
24      indirect mechanism involving the production of reactive oxygen species. These reactive oxygen
25      species may be formed as a result of autooxidation during futile metabolism of TCDD by the
26      induction of CYP1 enzymes or via the CYP1 -dependent production of estrogen metabolites
27      capable of redox cycling. The clonal expansion of these damaged cells by TCDD and related
28      chemicals is likely to occur through altered expression and activity of a number of genes
29      regulating the cell-cycle.  Activation of the AhR by TCDD results in altered expression or
30      activity in for EOF receptor, retinoblastoma protein, TGF-beta, and many others.  These proteins
31      all regulate the cell cycle and alterations of these proteins would alter cell growth properties.  The
32      contribution of these two pathways in the carcinogenic actions of TCDD remains uncertain.
33      However, Portier and colleagues have proposed a model in which the contribution of TCDD to
34      the number  of DNA damaged or initiated cells plays a significant role in its carcinogenic
35      response (Portier etal, 1996). In contrast, Conolly and Andersen, have proposed a rumor
36      promotion model based on a negative selection mechanism in which the actions of TCDD are
        9/22/00
27
DRAFT—DO NOT CITE OR QUOTE

-------
 1      focused on its ability to alter cell growth properties (Conolly and Andersen, 1997).  Descriptions
 2      of these models are provided in Part II, Chapter 8. Interestingly, the use of the model by Portier
 3      and colleagues, leads to a model that is consistent with low-dose linearity, whereas the Andersen
 4      and Conolly model predicts highly non-linear dose response relationships in the low dose region.
 5      Presently, the available data do not allow for adequate discrimination between these two models.
 6             TCDD causes a dose-related increase in thyroid follicular cell adenomas and carcinomas
 7      in rats and mice. One hypothesis for the induction of thyroid tumors involves the disruption of
 8      thyroid hormone homeostasis via the induction of the phase II enzymes UDP-
 9      glucuronosyltransferases (UGTs) (Hurley, PM, 1998; Hill et al., 1998). Dioxin-like compounds
10      induce the synthesis of UDP-glucuronosyltransferase-1 (UGT1) mRNA by an AhR-dependent
11      transcriptional mechanism (Bock et al., 1998; Nebert et al., 1990).  It is proposed that dioxin-like
12      chemicals increase the incidence of thyroid tumors through an extrathyroidal mechanism.
13      Dioxin-like chemicals induce hepatic UGT resulting in increased conjugation and elimination of
14      thyroxine (T4) and leading to reduced serum T4 concentrations. T4 production is controlled by
15      thyroid stimulating hormone (TSH) which is under negative and positive regulation from the
16      hypothalamus, pituitary, and thyroid by thyrotrophin releasing hormone (TRH), TSH itself,
17      thyroxine (T4), and triiodothyronine (T3). Consequently, the reduced serum T4 concentrations
18      would lead to a decrease in the negative feedback inhibition on the pituitary gland.  This would
19      then lead to a rise in secreted thyroid stimulating hormone and stimulation of the thyroid. The
20      persistent induction of UGT by dioxins and subsequent prolonged stimulation of the thyroid
21      would result in thyroid follicular cell hyperplasia and hypertrophy of the thyroid thereby
22      increasing the risk of progression to neoplasia.
23   '          In support of this hypothesis, Kohn et al.  modeled the effect of 2,3,7,8-TCDD on UGTS,
24      and thyroid hormones in female rats within the framework of a pharmacologically based
25      pharmacokinetic (PBPK) model (Kohn et al., 1996). This mathematical model described release
26      and uptake of thyroid hormones, metabolism, 2,3,7,8-TCDD induction of UGT1, regulation of
27      TSH release from the pituitary by T4 and feedback on  TRH and somatostatin which inhibits TSH
28      release. The model successfully reproduced the observed effects of 2,3,7,8-TCDD on serum T3,
29      T4, and TSH, and UGT1 mRNA and enzyme activity suggesting that this is a plausible
30      mechanism for an indirect role of 2,3,7,8-TCDD on the thyroid. This model is supported by the
31      more recent experimental work of Schuur and  colleagues, which demonstrated the extrathyroidal
32      effects of 2,3,7,8-TCDD on thyroid hormone turnover (Schuur et al., 1997).
33             Although this discussion illustrates that there is no defined molecular mechanism leading
34      to cancer in either liver or thyroid, it does demonstrate the concept of  "mode of action" as
35      defined in the Agency's proposed cancer guidelines  (U.S. EPA, 1996; 1999). In each case,
36      critical "key events"can be identified and measured which correlate with carcinogenicity. While
        9/22/00
28
DRAFT—DO NOT CITE OR QUOTE

-------
these relationships are still uncertain, they form plausible, testable hypotheses whose acceptance
by the scientific community is growing.
       Despite this lack of a defined mechanism at the molecular level, there is a consensus that
2,3,7,8-TCDD and related compounds are receptor-mediated carcinogens in that (1) interaction
with the AhR is a necessary early event; (2) 2,3,7,8-TCDD modifies a number of receptor and
hormone systems involved in cell growth and differentiation, such as the epidermal growth factor
receptor and estrogen receptor; and (3) sex hormones exert a profound influence on the
carcinogenic action of 2,3,7,8-TCDD.
2.2.1.4. Other Data Related to Carcinogenesis
       Despite the relatively large number of bioassays on 2,3,7,8-TCDD, the study of Kociba et
al. (1978) and those of the NTP (1982a), because of their multiple dose groups and wide dose
range, continue to be the focus of dose-response modeling efforts and of additional review.
Goodman and Sauer (1992) reported a re-evaluation of the female rat liver tumors in the Kociba
study using the latest pathology criteria for such lesions. The review confirmed only
approximately one-third of the tumors of the previous review (Squire, 1980). Although this
finding did not change the determination of carcinogenic hazard, as 2,3,7,8-TCDD induced
tumors in multiple sites in this study, it did have an effect on evaluation of dose-response and on
estimates of risk at low doses.  These issues will be discussed in a later section of this document.
       One of the more intriguing findings in the Kociba bioassay was reduced tumor incidences
of the pituitary, uterus, mammary gland, pancreas, and adrenals in exposed female rats as
compared to controls (Kociba et al., 1978). While these findings, coupled with evaluation of
epidemiologic data, have led some authors to conclude that dioxin possesses "anticarcinogenic"
activity (Kayajanian, 1997; Kayajanian, 1999), it should be noted that, in experimental studies,
with the exception of mammary gland tumors, the  decreased incidence of tumors is associated
with significant weight loss in these rats.  Examination of the data from NTP also demonstrates a
significant decrease in these tumor types when there is a concomitant weight loss in the rodents,
regardless of the chemical administered (Haseman and Johnson, 1996). As discussed later in
Section 3.2.3, under certain circumstances exposure to 2,3,7,8-TCDD may elicit beneficial
effects. For example, 2,3,7,8-TCDD protects against the subsequent carcinogenic effects of
polycyclic aromatic hydrocarbons (PAHs) in mouse skin, possibly reflecting induction of
detoxifying enzymes (Cohen et al., 1979; DiGiovanni et al., 1980).  In other situations, 2,3,7,8-
TCDD-induced changes in estrogen metabolism may alter the growth of hormone-dependent
tumor cells, producing a potential anticarcinogenic effect (Spink et al., 1990; Gierthy et al.,
1993). Because the mechanism of the decreases in the tumors is unknown, extrapolation of these
effects to humans is premature. In considering overall risk, one must take into account factors
9/22/00
29
DRAFT—DO NOT CITE OR QUOTE

-------
  1      such as the range of doses to target organs and hormonal state to obtain a complete picture of
  2      hazard and risk. Although exposure to dioxins may influence cancer response directly or
  3      indirectly, positively or negatively, it is unlikely that such data will be available to argue that
  4      dioxin exposure provides a net benefit to human health.
  5
  6      2.2.1.5.  Cancer Hazard Characterization
  7            TCDD, CDDs, CDFs, and dioxin-like PCBs are a class of well-studied compounds whose
  8      human cancer potential is supported by a large database including "limited" epidemiological
  9      support, unequivocal animal carcinogenesis, and biologic plausibility based on mode of action
10      data. In 1985, EPA classified 2,3,7,8-TCDD and related compounds as "probable" human
11      carcinogens based on the available data.  During the intervening years, the database relating to
12      the carcinogenicity of dioxin and-related compounds has grown and strengthened considerably.
13      In addition, EPA guidance for carcinogen risk assessment has evolved (U.S. EPA, 1996). Under
14      EPA's current approach, 2,3,7,8-TCDD is best characterized as a "human carcinogen."  This
15      means that, based on the weight of all of the evidence (human, animal, mode of action), 2,3,7,8-
16      TCDD meets the stringent criteria that allows EPA and the scientific community to accept a
17      causal relationship between 2,3,7,8-TCDD exposure and cancer hazard. The guidance suggests
18      that "human carcinogen" is an appropriate descriptor of carcinogenic potential when there is an
19      absence of conclusive epidemiologic evidence to clearly establish a cause-and-effect relationship
20      between human exposure and cancer, but there is compelling carcinogenicity data in animals and
21      mechanistic information in animals and humans demonstrating similar modes of carcinogenic
22      action. The "human carcinogen" descriptor is suggested for 2,3,7,8-TCDD because all of the
23      following conditions are met:
24            •   Occupational epidemiologic studies show an association between 2,3,7,8-TCDD
25                exposure and increases in cancer at all sites, in lung cancer, and perhaps at other sites,
26                but the data are insufficient on their own to demonstrate a causal association.
27            •   There is extensive carcinogenicity in both sexes of multiple species of animals at
28                multiple sites.
29            *   There is general agreement that the mode of 2,3,7,8-TCDD's carcinogenicity is AhR
30                dependent and proceeds through modification of the action of a number of receptor
31                and hormone systems involved in cell growth and differentiation, such as the
32                epidermal growth factor receptor and estrogen receptor.
33            •   The human AhR and rodent AhR are similar in structure and function and once
34                transformed, both bind to the same DNA response elements, designated DRE's.
35            •   Human and rodent tissue and organ cultures respond to TCDD and related chemicals
36                in a similar manner and at similar concentrations.
        9/22/00
30
DRAFT—DO NOT CITE OR QUOTE

-------
       Other dioxin-like compounds are characterized as "likely" human carcinogens primarily
because of the lack of epidemiological evidence associated with their carcinogenicity, although
there is a strong inference based on toxic equivalency that they would behave in humans as
2,3,7,8-TCDD does.  Each of the congeners that contributes substantially to human body burden
has been evaluated in vivo in cancer bioassays or tumor promotion assays. Each has a large data
base demonstrating AhR-mediated dioxin-like activities. Each has physico-chemical properties
which contribute to their persistence.  For each congener, the degree of certainty of carcinogenic
hazard is dependent on the available congener-specific data and its consistency with the
generalized mode of action that underpins toxicity equivalency for 2,3,7,8-TCDD and related
compounds. For the congeners most frequently encountered in human blood, milk and adipose
tissue, the data base in support of 2,3,7,8-TCDD-like carcinogenic hazard is strong; those with
weaker data supporting 2,3,7,8-TCDD-like carcinogenicity contribute relatively little to total
TEQ. Based on this logic, all complex environmental mixtures of 2,3,7,8-TCDD and dioxin-like
compounds would be characterized as "likely" carcinogens, but the degree of certainty of the
cancer hazard would be dependent on the major constituents of the mixture. For instance, the
hazard potential, although still considered "likely," would be characterized differently for a
mixture whose TEQ was dominated by OCDD as compared to one dominated by other PCDDs.

2.2.2. Reproductive and Developmental Effects
       Several sections of this reassessment (Part II, Chapter 5 and Chapter 7b) have focused on
the variety of effects that dioxin and dioxin-like agents can have-on human reproductive health
and development. Emphasis in each of these chapters has been on the discussion of the more
recent reports of the impact of dioxin-like compounds on reproduction and development.  These
have been put into context with previous reviews of the literature applicable in risk assessment
(Hatch, 1984; Sweeney, 1994; Kimmel,  1988) to develop a profile of the potential for dioxin and
dioxin-like agents to cause reproductive  or developmental toxicity, based on the available
literature.  An earlier  version of the literature review and discussion contained in Part II, Chapter
5, has been previously published (Peterson et al., 1993).
       The origin of concerns regarding a potential link between exposure to chlorinated dioxins
and adverse developmental events can be traced to early animal studies reporting increased
incidence of developmental abnormalities in rats and mice exposed early in gestation to 2,4,5-
trichlorophenoxyacetic acid (2,4,5-T) (Courtney and Moore, 1971). 2,4,5-T is a herbicide that
contains dioxin and related compounds as impurities. Its use was banned in the late 1970s, but
exposure to human populations continued as a result of past production, use, and disposal.
9/22/00
31
DRAFT—DO NOT CITE OR QUOTE

-------
  1      2.2.2.1.  Human
  2            The literature base with regard to potential human effects is detailed in Part II, Chapter
  3      7b, Section 7.13. In general, there is little epidemiological evidence that makes a direct
  4      association between exposure to TCDD or other dioxin-like compounds and effects on human
  5      reproduction or development.  One effect that may illustrate this relationship is the altered sex
  6      ratio (increased females) seen in the 6 years after the Seveso, Italy, accident (Mocarelli et al.,
  7      1996,2000). Particularly intriguing in this latest evaluation is the observation that exposure
  8      before and during puberty is linked to this sex ratio effect. Other sites have been examined for
  9      the effect of TCDD exposure on sex ratio with mixed results, but with smaller numbers of
10      offspring.  Continued evaluation of the Seveso population may provide other indications of
11      impacts  on reproduction and development but, for now, such data are very limited and further
12      research is needed.  Positive human data on developmental effects of dioxin-like compounds are
13      limited to a few studies of populations exposed to a complex mixture of potentially toxic
14      compounds (e.g., developmental studies from the Netherlands and effects of ingestion of
15      contaminated rice oil in Japan  [Yusho] and Taiwan [Yu-Cheng]).  In the latter studies, however,
16      all four manifestations of developmental toxicity (reduced viability, structural  alterations, growth
17      retardation, and functional alterations) have been observed to some degree, following exposure to
18      dioxin-like compounds as well as  other agents.  Data from the Dutch cohort of children exposed
19      to PCBs and dioxin-like compounds (Huisman et al., 1995a, b; Koopman-Esseboom et al.,
20      1994a-c; 1995a, b, 1996; Pluim et al., 1992, 1993, 1994; Weisglas-Kuperus et al., 1995; Patandin
21      et al., 1998, 1999) suggest impacts of background levels of dioxin and related compounds on
22      neurobehavioral outcomes, thyroid function, and liver enzymes: aspartate aminotransferase
23      (AST) and alanine aminotransferase (ALT). Although these effects cannot be  attributed solely to
24      dioxin and related compounds, several associations suggest that these are, in fact, likely to be
25      Ah-mediated effects. Similarly, it is highly likely that the developmental effects in human
26      infants exposed to a complex mixture of PCBs, PCDFs, and polychlorinated quaterphenyls
27      (PCQs) in the Yusho and Yu-Cheng poisoning episodes may have been caused by the combined
28      exposure to those PCS and PCDF congeners that are Ah-receptor agonists (Lii and Wong, 1984;
29      Kuratsune,  1989; Rogan, 1989). However, it is not possible to determine the relative
30      contributions of individual chemicals to the observed effects.
31            The incidents at Yusho and Yu-Cheng resulted in increased perinatal mortality and low
32      birthweight in infants born to women who had been exposed.  Rocker bottom heal was observed
33      in Yusho infants, and functional abnormalities have been reported in Yu-Cheng children.  Not all
34      the effects that were seen are attributable only to dioxin-like compounds. The similarity of
35      effects observed in human infants prenatally exposed to this complex mixture with those reported
36      in adult monkeys exposed only to TCDD suggests that at least some of the effects in the Yusho
        9/22/00
32
DRAFT—DO NOT CITE OR QUOTE

-------
and Yu-Cheng children are due to the TCDD-like congeners in the contaminated rice oil ingested
by the mothers of these children. The similar responses include a clustering of effects in organs
derived from the ectodermal germ layer, referred to as ectodermal dysplasia, including effects on
the skin, nails, and Meibomian glands; and developmental and psychomotor delay during
developmental and cognitive tests (Chen etal., 1992).  Some investigators believe that, because
all of these effects in the Yusho and Yu-Cheng cohorts do not correlate with TEQ, some of the
effects are exclusively due to nondioxin-like PCBs or a combination of all the congeners. It is
still not clear to what extent there is an association between overt maternal toxicity and
embryo/fetal toxicity in humans.
       Of particular interest is the common developmental origin (ectodermal layer) of many of
the organs and tissues that are affected in the human.  An ectodermal dysplasia syndrome has
been clearly associated with the Yusho and Yu-Cheng episodes, involving hyperpigmentation,
deformation of the fingernails and toenails, conjunctivitis, gingival hyperplasia, and
abnormalities of the teeth. An investigation of dioxin exposure and tooth development was done
in Finnish children as a result of studies of dental effects in dioxin-exposed rats, mice, and
nonhuman primates (Part II, Chapter 5, Section 5.2), and in PCB-exposed children (Rogan et al.,
1988). The Finnish investigators examined enamel hypomineralization of permanent first molars
in 6-7 year old children (Alaluusua  et al., 1996,1999). The length of time that infants breast fed
was not significantly associated with either mineralization changes or with TEQ levels in the
breast milk. However,  when the levels and length of breast feeding were combined in an overall
score, a statistically significant association was observed (r = 0.3, p = 0.003, regression analysis).
These data are discussed further in Part II,  Chapter 7b, Section 7.13. The developmental effects
that can be associated with the nervous system are also consistent with this pattern of impacts on
tissues of ectodermal origin, as the nervous system is of ectodermal origin. These data are
limited but are discussed in Part II, Chapter 7b, Section 7.13.
       Other investigations into noncancer effects of human exposure to dioxin have provided
human data on TCDD-induced changes in  circulating reproductive hormones.  This was one of
the effects judged as having a positive relationship with exposure to TCDD in Part II, Chapter
7b, Section 7.13.  Levels of reproductive hormones have been measured with respect to exposure
to 2,3,7,8-TCDD in three cross-sectional medical studies. Testosterone, luteinizing hormone
(LH), and follicle-stimulating hormone (FSH) were measured in trichlorophenol (TCP) and
2,4,5-T production workers (Egeland et al., 1994), in Army Vietnam veterans (CDC Vietnam
Experience Study, 1988), and in Air Force personnel, known as "Ranch Hands," who handled
and/or sprayed Agent Orange during the Vietnam War (Roegner et al., 1991; Grubbs et al.,
1995).  The risk of abnormally low testosterone was two to four times higher in exposed workers
with serum 2,3,7,8-TCDD levels above 20 ng/g than in unexposed referents (Egeland et al.,
9/22/00
33
DRAFT—DO NOT CITE OR QUOTE

-------
 1      1994). In both the 1987 and 1992 examinations, mean testosterone concentrations were slightly,
 2      but not significantly, higher in Ranch Hands (Thomas et al., 1990; Grubbs et al., 1995). FSH and
 3      LH concentrations were no different between the exposed and comparison groups. No
 4      significant associations were found between Vietnam experience and altered reproductive
 5      hormone levels (CDC Vietnam Experience Study, 1988). Only the NIOSH study found an
 6      association between serum 2,3,7,8-TCDD level and increases in serum LH.
 7            The findings of the NIOSH and Ranch Hand studies are plausible given the
 8      pharmacological and toxicological properties of 2,3,7,8-TCDD in animal models, which are
 9      discussed in Part II, Chapters 5 and 7. One plausible mechanism responsible for the effects of
10      dioxins may involve their ability to influence hormone receptors. The AhR, to which 2,3,7,8-
11      TCDD binds, and the hormone receptors are signaling pathways that regulate homoeostatic
12      processes. These signaling pathways are integrated at the cellular level and there is considerable
13      "cross-talk" between these pathways. For example, studies suggest that 2,3,7,8-TCDD
14      modulates the concentrations of numerous hormones and/or their receptors, including estrogen
15      (Romkes and Safe, 1988; Romkes et al., 1987), progesterone (Romkes et al., 1987),
16      glucocorticoid (Ryan et al., 1989), and thyroid hormones (Gorski and Rozman, 1987).
17            In summary, the results from both the NIOSH and Ranch Hand studies are limited by the
18      cross-sectional nature of the data and the type of clinical assessments conducted.  However, the
19      available data provide evidence that small alterations in human male reproductive hormone
20      levels are associated with serum 2,3,7,8-TCDD.
21
22      2.2.2.2.  Experimental Animal
23            The extensive experimental animal database with respect to reproductive and
24      developmental toxicity of dioxin and dioxin-related agents has been discussed in Part II, Chapter
25      5.  Dioxin exposure has been observed to result in both male and female reproductive effects, as
26      well as effects on development. These latter effects are among the most responsive health
27      endpoints to dioxin exposure (see Part II, Chapter  8, Section 8.3). In general, the prenatal and
28      developing postnatal animal is more sensitive to the effects of dioxin than is the adult. In several
29      instances (e.g., fetotoxicity in hamsters, rats, mice, and guinea pigs), the large species differences
30      seen in acute toxicity are greatly reduced when developing  animals are evaluated. Most of the
31      data reviewed are from studies of six genera of laboratory animals. Although much of the data
32      comes from animals exposed only to TCDD, more recent studies of animals exposed to mixtures
33      of PCDD/PCDF isomers provide results that are consistent with the studies of TCDD alone.
34
35      2.2.2.2.1. Developmental toxicity.  Dioxin exposure results in a wide variety of developmental
36      effects; these are observed in three different vertebrate classes and in several species within each
        9/22/00
34
DRAFT—DO NOT CITE OR QUOTE

-------
class. All four of the manifestations of developmental toxicity have been observed following
exposure to dioxin, including reduced viability, structural alterations, growth retardation, and
functional alterations.  As summarized previously (Peterson et al.,  1993), increased prenatal
mortality (rat and monkey), functional alterations in learning and sexual behavior (rat and
monkey), and changes in the development of the reproductive system (rat, hamster) occur at the
lowest exposure levels tested (see also Part II, Chapter 8, Section 8.3).
       Dioxin exposure results in reduced prenatal or postnatal viability in virtually every
species in which it has been tested. Previously, increased prenatal mortality appeared to be
observed only at exposures that also resulted in maternal toxicity.  However, the studies of Olson
and McGarrigle (1990) in the hamster and Schantz et al. (1989) in the monkey were suggestive
that this was not the case in all species. Although the data from these two studies were limited,
prenatal death was observed in cases where no maternal toxicity was evident. In the rat,
Peterson's laboratory (Bjerke et al., 1994a, b; Roman et al., 1995) reported increased prenatal
death following a single exposure to TCDD during gestation that did not cause maternal toxicity,
and Gray et al. (1995a) observed a decrease  in postnatal survival under a similar exposure
regimen. While identifying the presence or  absence of maternal toxicity may be instructive as to
the specific origin of the reduced prenatal viability, it does not alter the fact that pre- and
postnatal deaths were  observed. In either case, the Agency considers these effects as being
indicators of developmental toxicity in response to the exposure (U.S. EPA, 1991b).
       Some of the most striking findings regarding dioxin exposure relate to the effects on the
developing reproductive system in laboratory animals.  Only a single, low-level exposure to
TCDD during gestation is required to initiate these developmental alterations. Mably et al.
(1992a-c) originally reported that a single exposure of the Holtzman maternal rat to as low as
0.064 ng/kg could alter normal sexual development in the male offspring. A dose of 0.064 ng/kg
in these studies results in a maximal body burden in the maternal animal of 64 ng/kg during
critical windows in development. More recently, these findings of altered normal sexual
development have been further defined (Bjerke et al., 1994a, b; Gray  et al., 1995a; Roman et al.,
1995), as well as extended to females and another strain and species (hamster) (Gray et al.,
1995b). In general, the findings of these later studies have produced qualitatively similar results
that define a significant effect of dioxin on the developing reproductive system.
       In the developing male rat, TCDD exposure during the prenatal and lactational periods
results in delay of the onset of puberty as measured by age at preputial separation. There is a
reduction in testis weight,  sperm parameters, and sex accessory gland weights.  In the mature
male exposed during the prenatal and lactational periods, there is an alteration of normal sexual
behavior and reproductive function. Males  exposed to TCDD during gestation are
demasculinized. Feminization of male sexual behavior and a reduction in the number of
9/22/00
35
DRAFT—DO NOT CITE OR QUOTE

-------
  1      implants in females mated with exposed males have also been reported, although these effects
  2      have not been consistently found. These effects do not appear to be related to reductions in
  3      circulating androgens, which were shown in the most recent studies to be normal. Most of these
  4      effects occur in a dose-related fashion, some occurring at 0.05 (ig/kg and 0.064 ng/kg, the lowest
  5      TCDD doses tested (Mably et al., 1992c; Gray et al, 1997a).
  6             In the developing female rat, Gray and Ostby (1995) have demonstrated altered sexual
  7      differentiation in both the Long Evans and Holtzman strains. The effects observed depended on
  8      the timing of exposure.  Exposure during early organogenesis altered the cyclicity, reduced
  9      ovarian weight, and shortened the reproductive lifespan. Exposure  later in organogenesis
10      resulted in slightly lowered ovarian weight, structural alterations of the genitalia, and a slight
11      delay in puberty. However, cyclicity and fertility were not affected with the later exposure. The
12      most sensitive dose-dependent effects of TCDD in the female rat were structural alterations of
13      the genitalia that occurred at 0.20 ug TCDD/kg administered to the  dam (Gray et al., 1997b).
14             As described above, studies demonstrating adverse health effects from prenatal exposures
15      often involved a single dose administered at a discrete time  during pregnancy. The production of
16      prenatal effects at a given dose appears to require exposure during critical times in fetal
17      development. This concept is well supported by a recent report (Hurst et al., 2000) which
18      demonstrated the same incidence of adverse effects in rat pups born to dams with a single
19      exposure of 0.2 p.g TCDD/kgBW on gestation day  15 (GD 15) versus 1.0- u-g TCDD/kgBW on
20      gestation day 8 (GD 8).  Both of these experimental paradigms result in the same fetal tissue
21      concentrations and body burdens during the critical window of sensitivity. For example,
22      exposure to 0.2 \ig TCDD/kgB W on GD 15 results in 13.2 pg TCDD/g fetal tissue on GD 16;
23      exposure to 1.0 \ig TCDD/kgBW on gestation GD 8 resulted in 15.3 pg TCDD/g fetus on GD 16.
24      This study demonstrates the appropriateness of the use of body burden to describe the effects of
25      TCDD when comparing different exposure regimens. The uncertainties introduced when trying
26      to compare studies with steady-state body burdens with single-dose studies may make it difficult
27      to determine a lowest effective dose.  Application of pharmacokinetic models, described earlier
28      in Parts I and II, to estimate body burdens at the critical time of development is expected to be a
29      sound method for relating chronic background exposures to the results obtained from single-dose
30      studies.
31             Structural malformations, particularly cleft palate and hydronephrosis, occur in mice
32      administered doses of TCDD. The findings, while not representative of the most sensitive
33      developmental endpoints, indicate that exposure during the critical period of organogenesis can
34      affect the processes involved in normal tissue formation. The TCDD-sensitive events appear to
35      require the AhR. Mouse strains that produce AhRs with relatively high affinity for TCDD
36      respond to lower doses than do strains with relatively low-affinity receptors. Moreover,
        9/22/00
36
DRAFT—DO NOT CITE OR QUOTE

-------
congeners with a greater affinity for the AhR are more developmentally toxic than those with a
lower affinity. This is consistent with the rank ordering of toxic potency based on affinity for the
receptor as discussed in Part II, Chapter 9, Section 9.3.

2.2.2.2.2. Adult female reproductive toxicity.  The primary effects of TCDD on female
reproduction appear to be decreased fertility, inability to maintain pregnancy for the full
gestational period and, in the rat, decreased litter size.  In some studies of rats and of primates,
signs of ovarian dysfunction such as anovulation and suppression of the estrous cycle have been
reported (Kociba et al, 1976; Barsotti et al., 1979; Allen et al., 1979; Li et al., 1995a, b).  While
the majority of reproductive effects are associated with high-dose exposures in experimental
animals, the induction of endometriosis in primates occurs at body burdens near background
human exposures.
2.2.2.2.3. Adult male reproductive toxicity, TCDD and related compounds decrease testis and
accessory sex organ weights, cause abnormal testicular morphology, decrease spermatogenesis,
and reduce fertility when given to adult animals in doses sufficient to reduce feed intake and/or
body weight. In the testes of these different species, TCDD effects on spermatogenesis are
characterized by loss of germ cells, the appearance of degenerating spermatocytes and mature
spermatozoa within the lumens of seminiferous tubules, and a reduction in the number of tubules
containing mature spermatozoa (Allen and Lalich, 1962; Allen and Carstens, 1967; McConnell et
al., 1978; Chahoud et al., 1989). This suppression of spermatogenesis is not a highly sensitive
effect when TCDD is administered to postweanling animals, as an exposure of 1 |ig/kg/day over
a period of weeks appears to be required to produce these effects.

2.2.2.3. Other Data Related to Developmental and Reproductive Effects
2.2.2.3.1. Endometriosis. The association of dioxin with endometriosis was first reported in a
study of Rhesus monkeys that had been exposed for 4 years to dioxin in their feed and then held
for an additional 10 years (Rier et al., 1993). There was a dose-related increase in both the
incidence and severity of endometriosis in the exposed monkeys as compared to controls.
Follow-up on this group  of monkeys revealed a clear association with total TEQ. A study in
which Rhesus monkeys were exposed to PCBs  for up to 6 years failed to show any enhanced
incidence of endometriosis (Arnold et al., 1996).  However, many of these monkeys were no
longer cycling, and the time may not have been adequate to develop the response. In the TCDD
monkey study, it took 7 years before the first endometriosis was noted (Rier et al., 1993). A
recent study in Cynomolgus monkeys has shown promotion of surgically induced endometriosis
by TCDD within 1 year after surgery (Yang et al., 2000). Studies using rodent models for
9/22/00
37
DRAFT—DO NOT CITE OR QUOTE

-------
 1      surgically induced endometriosis have also shown the ability of TCDD to promote lesions in a
 2      dose-related manner (Cummings et al., 1996,1999; Johnson et al., 1997; Bruner-Tran et al.,
 3      1999).  This response takes at least 2 months to be detected (Cummings et al., 1996, 1999;
 4      Johnson et al., 1997).  Another study in mice which failed to detect dioxin promotion of
 5      surgically induced endometriosis only held the mice for only 1 month, not long enough to detect
 6      a response (Yang et al., 1997). Prenatal exposure of mice also enhanced the sensitivity of the
 7      offspring to the promotion of surgically induced endometriosis by TCDD.  The effects of TCDD
 8      in the murine model of endometriosis appear to be AhR-mediated, as demonstrated in a study in
 9      which AhR ligands were able to promote the lesions, while non-AhR ligands, including a non-
10      dioxin-like PCB, had no effect on surgically induced endometriosis. Dioxin has also been shown
11      to result in endometriosis in human endometrial tissue implanted in nude mice (Bruner-Tran et
12      al., 1999).
13             Data on the relationship of dioxins to endometriosis in people is intriguing, but
14      preliminary. Studies in the early 1990s suggested that women with higher levels of persistent
15      organochlorines were at increased risk for endometriosis (Gerhard and Runnebaum, 1992). This
16      was followed by the observation that Belgian women, who have the highest levels of dioxins in
17      their background population, had higher incidences of endometriosis than reported from other
18      populations (Koninckx et al., 1994). A study from Israel then demonstrated that there was a
19      correlation between detectable TCDD in women with surgically confirmed endometriosis, in
20      comparison to those with no endometriosis (Mayani et al., 1997). Recent studies from Belgium
21      have indicated that women with higher body burdens, based on serum TEQ determinations, are at
22      greater risk for endometriosis (Pauwels et al., 1999). No association was seen with total PCBs in
23      this study.  A small study in the United States, which did not involve surgically confirmed
24      endometriosis, saw no association between TCDD and endometriosis (Boyd et al., 1995).
25      Likewise, a study in Canada saw no association between total PCBs and endometriosis (Lebel et
26      al., 1998).  The lack of an association with total PCBs is not surprising because the rodent studies
27      have indicated that this response is AhR-mediated (Johnson et al., 1997).
28             The animal results lend biological plausibility to the epidemiology findings.
29      Endometriosis is not only an endocrine disorder, but is also associated with immune system
30      alterations (Rier et al., 1995). Dioxins are known to be potent modulators  of the animal immune
31      system, as well as affecting estrogen homeostasis. Further studies are clearly needed to provide
32      additional support to this association of endometriosis and dioxins, as well as to demonstrate
33      causality.
34
        9/22/00
38
DRAFT—DO NOT CITE OR QUOTE

-------
 1      2.2.2.3.2. Androgenic deficiency.  The effects of TCDD on the male reproductive system when
 2      exposure occurs in adulthood are believed to be due in part to an androgenic deficiency. This
 3      deficiency is characterized in adult rats by decreased plasma testosterone and
 4      5a-dihydrotestosterone (DHT) concentrations, unaltered plasma LH concentrations, and
 5      unchanged plasma clearance of androgens and LH (Moore et al., 1985, 1989; Mebus et al., 1987;
 6      Moore and Peterson, 1988; Bookstaff et al., 1990a). The cause of the androgenic deficiency was
 7      believed to be due to decreased testicular responsiveness to LH and increased pituitary
 8      responsiveness to feedback inhibition by androgens and estrogens (Moore et al., 1989,1991;
 9      Bookstaff et al., 1990a, b; Kleeman et al.,  1990). The single dose used in some of those earlier
10      studies (15 ugTCDD/kgBW) is now known to affect Leydig cells (Johnson et al., 1994).
11
12      2.2.2.4. Developmental and Reproductive Effects Hazard Characterization
13            There is limited direct evidence addressing the issues of how or at what levels humans
14      will begin to respond to dioxin-like compounds with adverse impacts on development or
15      reproductive function.  The series of published Dutch studies suggest that pre- and early postnatal
16      exposures to PCBs and other dioxin-like compounds may impact developmental milestones at
17      levels at or near current average human background exposures. Although it is unclear whether
18      these measured responses indicate a clearly adverse impact, if humans respond to TCDD
19      similarly to animals in laboratory studies, there are indications that exposures at relatively low
20      levels might cause developmental effects and at higher exposure levels might cause reproductive
21      effects. There is especially good evidence for effects on the fetus from'prenatal exposure. The
22      Yusho and Yu-Cheng poisoning incidents are clear demonstrations that dioxin-like compounds
23      can produce a variety of mild to severe developmental effects in humans that resemble the effects
24      of exposure to dioxins and dioxin-like compounds in animals.  Humans do not appear to be
25      particularly sensitive or insensitive to effects of dioxin exposure hi comparison to other animals.
26      Therefore, it is reasonable to assume  that human responsiveness would lie across the middle
27      ranges of  observed responses. This still does not address the issues surrounding the potentially
28      different responses humans (or animals) might have to the more complex and variable
29      environmental mixtures of dioxin-like compounds.
30            TCDD and related compounds have reproductive and developmental toxicity potential in
31      a broad range of wildlife, domestic, and laboratory animals. Many of the effects have been
32      shown to be TCDD dose-related. The effects on perinatal viability and male reproductive
33      development are among the most sensitive effects reported, occurring at a single prenatal
34      exposure range of as little as 0.05-0.075 (J-g/kg, resulting in calculated fetal tissue concentrations
35      of 3-4 ng/kg.  In these studies, effects were often observed at the lowest exposure level tested,
36      thus a no-observed adverse effect level (NOAEL) has not been established for several of these
        9/22/00
39
DRAFT—DO NOT CITE OR QUOTE

-------
  1      endpoints.  In general, the structure-activity results are consistent with an AhR-mediated
  2      mechanism for the developmental effects that are observed in the low dose range. The structure-
  3      activity relationship in laboratory mammals appears to be similar to that for AhR binding. This
  4      is especially the case with cleft palate in the mouse.
  5             It is assumed that the responses observed in animal studies are indicative of the potential
  6      for reproductive and developmental toxicity in humans. This is an established assumption in the
  7      risk assessment process for developmental toxicity (U.S. EPA, 1991b). It is supported by the
  8      number of animal species and strains in which effects have been observed. The limited human
  9      data are consistent with an effect following exposure to TCDD or TCDD-like agents.  In
10      addition, the phylogenetic conservation of the structure and function of the AhR also increases
11      our confidence that these effects may occur in humans.
12             Although there is evidence in experimental animals that exposure to dioxin-like
13      chemicals during development produces neurobehavioral effects, the situation in humans is more
14      complex. Studies in humans demonstrate associations between dioxin exposure and alterations
15      in neurological development. These same studies often show similar associations between
16      exposure to non-dioxin-like PCBs and these same effects. On the basis of the human studies, it
17      is possible that the alterations in neurological development are due to an interaction between the
18      dioxins and the non-dioxin-like PCBs. At present there are limited data that define the roles  of
19      the dioxins versus the non-dioxin-like PCBs in these effects on neurological development.
20             In general, the structure-activity results on dioxin-like compounds are consistent with an
21      AhR-mediated mechanism for many of the developmental effects that are observed. The
22      structure-activity relationship in laboratory mammals appears to be similar to that for AhR
23      binding. This is especially the case with cleft palate in the mouse. However, a direct
24      relationship with Ah binding is less clear for other effects, including those involving the
25      developing nervous system.
26
27      2.2.3. Immunotoxicity
28      2.2.3.1. Epidemiologic Findings
29             The available epidemiologic studies on immunologic function in humans relative to
30      exposure to 2,3,7,8-TCDD do not describe a consistent pattern of effects among the examined
31      populations. Two studies of German workers, one exposed to 2,3,7,8-TCDD and the other to
32      2,3,7,8-tetrabrominated dioxin and furan, observed dose-related increases of complements C3 or
33      C4 (Zober et al., 1992; Ott et al., 1994), while the Ranch Hands continue to exhibit elevations in
34      immunoglobulin A (IgA) (Roegner et al., 1991; Grubbs et al., 1995). Other studies of groups
35      with documented exposure to 2,3,7,8-TCDD have not examined complement components to any
36      great extent or observed significant changes in IgA. Suggestions of immunosuppression have
        9/22/00
40
DRAFT—DO NOT CITE OR QUOTE

-------
 1      been observed in a small group of exposed workers as a result of a single test (Tonn et al., 1996),
 2      providing support for a testable hypothesis to be evaluated in other exposed populations.
 3            Comprehensive evaluation of immunologic status arid function of the NIOSH, Ranch
 4      Hand, and Hamburg chemical worker cohorts found no consistent differences between exposed
 5      and unexposed groups for lymphocyte subpopulations, response to mitogen stimulation, or rates
 6      of infection (Halperin et al., 1998; Michalek et al.,"1999b; Jung et al., 1998; Ernst et al., 1998).
 7            More comprehensive evaluations of immunologic function with respect to exposure to
 8      2,3,7,8-TCDD and related compounds are necessary to assess more definitively the relationships
 9      observed in nonhuman species. Longitudinal studies of the maturing human immune system may
10      provide the greatest insight, particularly because animal studies have found significant results in
11      immature animals, and human breast milk is a source of 2,3,7,8-TCDD and other related
12      compounds.  The studies of Dutch infants described earlier provide an example of such a  study
13      design.  Additional studies of highly exposed adults may also shed light on the effects oflong-
14      term chronic exposures through elevated body burdens. Therefore, there appears to be too little
15      information to suggest definitively that 2,3,7,8-TCDD, at the levels observed, causes long-term
16      adverse effects on the immune system in adult humans.
17
18      2.2.3.2. Animal Findings
19            Cumulative evidence from a number of studies indicates that th&immune system of
20      various animal species is a target for toxicity of TCDD and structurally related compounds,
21      including other PCDDs, PCDFs, and PCBs. Both cell-mediated and humoral immune responses
22      are suppressed following TCDD exposure, suggesting that there are multiple cellular targets
23      within the immune system that are altered by TCDD.  Evidence also suggests that the immune
24      system is indirectly targeted by TCDD-induced changes in nonlymphoid tissues. TCDD
25      exposure of experimental animals results in decreased host resistance following challenge with
26      certain infectious agents, which likely result from TCDD-induced suppression of immunological
27      functions.
28            The primary antibody response to the T cell-dependent antigen, sheep red blood cells
29      (SRBCs), is the most sensitive immunological response that is consistently suppressed  in mice
30      exposed to TCDD  and related compounds.  The degree of immunosuppression is related to the
31      potency of the dioxin-like congeners.  There is remarkable agreement among several different
32      laboratories for the potency of a single acute dose of TCDD (i.e., suppression at a dose as low as
33      0.1 ug TCDD/kg with an average 50% immunosupressive dose [ID50] value of approximately
34      0.7 |j.g TCDD/kg) to suppress this response in Ah-responsive mice. Results of studies that have
35      compared the effects of acute exposure to individual PCDDs, PCDFs, and PCB congeners, which
36      differ in their binding affinity for the AhR, on this response have provided critical evidence that
        9/22/00
41
DRAFT—DO NOT CITE OR QUOTE

-------
  1      certain dioxin-like congeners are also immunosuppressive.  The degree of immunosuppression
  2      has been found to be related to potency of the dioxin-like congeners. Antibody responses to
  3      T cell-independent antigens, such as trinitrophenyl-lipopolysaccharide (TNP-LPS) and the
  4      cytotoxic T lymphocyte (CTL) response, are also suppressed by a single acute exposure to
  5      TCDD, albeit at higher doses than those that suppress the SRBC response. Although a thorough
  6      and systematic evaluation of the immunotoxicity of TCDD-like congeners in different species
  7      and for different immunological endpoints has not been performed, it can be inferred from the
  8      available data that dioxin-like congeners are immunosuppressive.
  9             Perinatal exposure of experimental animals to TCDD results in suppression of primarily
10      T cell immune functions, with evidence of suppression persisting into adulthood. In mice, the
11      effects on T cell functions appear to be related to the fact that perinatal TCDD exposure alters
12      thymic precursor stem cells in the fetal liver and bone marrow, and thymocyte differentiation in
13      the thymus. These studies suggest that perinatal development is a critical and sensitive period for
14      TCDD-induced immunotoxicity. Efforts should be made to determine the consequences of
15      perinatal exposure to TCDD and related compounds and mixtures on immune system integrity.
16
17      2.2.3.3. Other Data Related to Immunologic Effects
18             In addition to the TCDD-like congener results, studies using strains of mice that differ in
19      the expression of the AhR have provided critical evidence to support a role for Ah-mediated
20      immune suppression following exposure to dioxin-like compounds. Recent in vitro work also
21      supports a role for Ah-mediated immune suppression. Other in vivo arid in vitro data, however,
22      suggest that non-Ah-mediated mechanisms may also play some role in immunotoxicity induced
23      by dioxin-like compounds. However, more definitive evidence remains to be developed to
24      support this latter view.
25             Although the immunosuppressive potency of individual dioxin-like compounds in mice is
26      related to their structural similarity to TCDD, this pattern of suppression is observed only
27      following exposure to an individual congener. The immunotoxicity of TCDD and related
28      congeners can be modified by co-exposure to other congeners in simple binary or more complex
29      mixtures resulting in additive or antagonistic interactions. There is a need for the generation of
30      dose-response data of acute, subchronic, and chronic exposure to the individual congeners in a
31      mixture and for the mixture itself hi order to fully evaluate potential synergistic, additive, or
32      antagonistic effects of environmentally relevant mixtures.
33             Animal host resistance models that mimic human disease have been used to assess the
34      effects of TCDD on altered host susceptibility.  TCDD exposure increases susceptibility to
35      challenge with bacteria, viruses, parasites, and tumors.  Mortality is increased in TCDD-exposed
36      mice challenged with certain bacteria.  Increased parasitemia occurs in TCDD-exposed mice and
        9/22/00
42
DRAFT—DO NOT CITE OR QUOTE

-------
  1      rats challenged with parasitic infections. Low doses of TCDD also alter resistance to virus
  2      infections in rodents.  Increased susceptibility to infectious agents is an important benchmark of
  3      immunosuppression; however, the role that TCDD plays in altering immune-mediated
  4      mechanisms important in murine resistance to infectious agents remains to be elucidated. Also,
  5      because little is known about the effects that dioxin-like congeners have on host resistance, more
  6      research is recommended in this area.
  7             Studies in nonhuman primates exposed acutely, subchronically, or chronically to
  8      halogenated aromatic hydrocarbons (HAH) have revealed variable alterations in lymphocyte
  9      subpopulations, primarily T lymphocyte subsets. In three separate studies in which monkeys
10      were exposed subchronically or chronically to PCBs, the antibody response to SRBC was
11      .consistently found to be suppressed. These results in nonhuman primates are important because
12      they corroborate the extensive database of HAH-induced suppression of the antibody response to
13      SRBC in mice and thereby provide credible evidence for immunosuppression by HAHs across
14      species. In addition, these data indicate that the primary antibody response to this T cell-
15      dependent antigen is the most consistent and sensitive indicator of HAH-induced
16      immunosuppression.
17             The available database derived from well-controlled animal studies on TCDD
18      immunotoxicity can be used for the establishment of no-observed effect levels (NOEL). As the
19      antibody response to SRBCs has been shown to be dose-dependently suppressed by TCDD and
20      related dioxin-like compounds, this database is best suited for the development of dose-response
21      modeling.
22
23      2.2.3.4. Immunologic Effects Hazard Characterization
24             Accidental or occupational exposure of humans to TCDD and/or related compounds
25      variably affects a number of irnmunological parameters.  Unfortunately, the evaluation of
26      immune system integrity in humans exposed to dioxin-like compounds has provided data that is
27      inconsistent across studies.  However, the broad range of "normal" responses in humans due to
28      the large amount of variability inherent in such a heterogenous population, the limited number
29      and sensitivity of tests performed, and poor exposure characterization of the cohorts in these
30      studies compromise any conclusions about the ability of a given study to detect immune
31      alterations.  Consequently, there are insufficient clinical data from these studies to fully assess
32      human sensitivity to TCDD exposure. Nevertheless, based on the results of the extensive animal
33      work, the database is sufficient to indicate that immune effects could occur in the human
34      population from exposure to TCDD and related compounds at some dose level.  At present, it is
35      EPA's scientific judgment that TCDD and related compounds should be regarded as nonspecific
        9/22/00
43
DRAFT—DO NOT CITE OR QUOTE

-------
  1      immunosuppressants and immunotoxicants until better data to inform this judgment are available.
  2             It is interesting that a common thread in several human studies is the observed reduction
  3      in CD4+ T helper cells, albeit generally within the "normal" range, in cohorts exposed to dioxin-
  4      like compounds. Even though these reductions may not translate into clinical effects, it is
  5      important to note that these cells play an important role in regulating immune responses and that
  6      their reduction in clinical diseases is associated with immunosuppression. Another important
  7      consideration is that a primary antibody response following immunization was not evaluated in
  8      any of the human studies.  Because this immune parameter has been revealed to be the most
  9      sensitive in animal studies, it is recommended that TCDD and related compounds be judged
10      irnmunosupressive and that this parameter be included in future studies of human populations
11      exposed to TCDD and related compounds.  It is also recommended that research focused on
12      delineating the mechanism(s) underlying dioxin-induced immunotoxicity and
13      immunosuppression continue.
14
15      2.2.4. Chloracne
16             Chloracne and associated dermatologic changes are widely recognized responses to
17      TCDD and other dioxin-like compounds in humans. Along with the reproductive hormones
18      discussed above and gamma glutamyl transferase (GOT) levels, which are discussed below,
19      chloracne is one of the noncancer effects that has a strong positive association with exposure to
20      TCDD in humans (see Part II, Chapter 7b, Section 7.13). Chloracne is a severe acnelike
21      condition that develops within months of first exposure to high levels of dioxin and related
22      compounds. For many individuals, the condition disappears after discontinuation of exposure,
23      despite initial serum levels of dioxin in the thousands of parts per trillion (ppt); for others, it may
24      remain for many years. The duration of persistent chloracne is on the order of 25 years, although
25      cases of chloracne persisting over 40 years have been noted (see Part II, Chapter  7b, Section
26      7.13).
27             In general, chloracne has been observed in most incidents where substantial dioxin
28      exposure has occurred, particularly among TCP production workers and Seveso residents (see
29      Part II, Chapter 7b). The amount of exposure necessary for development of chloracne has not
30      been resolved, but studies suggest that high exposure (both high acute and long-term exposure) to
31      2,3,7,8-TCDD increases the likelihood of chloracne, as evidenced by chloracne in TCP
32      production workers and Seveso residents who have documented high serum 2,3,7,8-TCDD levels
33      (Beck et al., 1989; Fingerhut et al, 1991a; Mocarelli et al, 1991; Neuberger et al., 1991) or in
34      individuals who have a work history with long duration of exposure to 2,3,7,8-TCDD-
35      contaminated chemicals (Bond et al., 1989). In earlier studies, chloracne was considered to be a
36      "hallmark of dioxin intoxication" (Suskind, 1985). However,  only in two studies were risk
        9/22/00
44
DRAFT—DO NOT CITE OR QUOTE

-------
 1      estimates calculated for chloracne. Both were studies of different cohorts of TCP production
 2      workers (Suskind and Hertzberg, 1984; Bond et al., 1989); one group was employed in a West
 3      Virginia plant, the other in a plant in Michigan. Of the 203 West Virginia workers, 52.7%
 4      (p<0.001) were found to have clinical evidence of chloracne, and 86.3% reported a history of
 5      chloracne (pO.OOl) (Suskind and Hertzberg, 1984). None of the unexposed workers had clinical
 6      evidence or reported a history of chloracne.  Among the Michigan workers, the relative risk for
 7      cases of chloracne was highest for individuals with the longest duration of exposure (> 60
 8      months; RR = 3.5, 95% CI = 2.3-5.1), those with the highest cumulative dose of TCDD (based
 9      on duration of assignment across and within 2,3,7,8-TCDD-contaminated areas in the plant)
10      (RR = 8.0, 95% CI = 4.2-15.3), and those with the highest intensity of 2,3,7,8-TCDD exposure
11      (RR = 71.5, 95% CI = 32.1-159.2) (Bond et al., 1989).
12             Studies in multiple animal species have been effective in describing the relationship
13      between'2,3,7,8-TCDD and chloracne, particularly in rhesus monkeys (McNulty, 1977; Allen et
14      al., 1977; McConnell et al., 1978). Subsequent to exposure to 2,3,7,8-TCDD, monkeys
15      developed chloracne and swelling of the meibornian glands, modified sebaceous glands in the
16      eyelid. The histologic changes in the meibomian glands are physiologically similar to those
17      observed in human chloracne (Dunagin, 1984).
18             In summary, the evidence provided by the various studies convincingly supports what is
19      already presumed, that chloracne is a common sequel of high levels of exposure to 2,3,7,8-
20      TCDD and related compounds. More information is needed to determine the level and frequency
21      of exposure to dioxin-like compounds needed to cause chloracne, and whether personal
22      susceptibility plays a role in the etiology. Finally, it is important to recall that the absence of
23      chloracne does not imply lack of exposure (Mocarelli et al., 1991).
24
25      2.2.5. Diabetes
26             Diabetes mellitus is a heterogeneous disorder that is a consequence of alterations in the
27      number or function of pancreatic beta cells responsible for insulin secretion and carbohydrate
28      metabolism. Diabetes and fasting serum glucose levels were evaluated in more recent cross-
29      sectional medical studies because of the apparently high prevalence of diabetes and abnormal
30      glucose tolerance tests in one case report of 55 TCP workers (Pazderova-Vejlupkova et al.,
31      1981). Recent epidemiology studies, as well as early case reports, have indicated a weak
32      association between serum concentrations of dioxin and diabetes. This association was first
33      noted in the early 1990s when a decrease in glucose tolerance was seen in the NIOSH cohort.
34      This was followed by a report of an increase in diabetes in the Ranch Hand cohort (Michalek et
35      al., 1999; Longnecker and Michalek, 2000).  An increase in diabetes in other occupational
36      cohorts (Steenland et al., 1999; Vena et al.,  1998), as well as the Seveso population (Pesatori et
        9/22/00
45
DRAFT—DO NOT CITE OR QUOTE

-------
  1      al., 1998) has also been reported. There was not a significant increase in diabetes in the NIOSH
  2     mortality study, although 6 of the 10 most highly exposed workers did have diabetes (Calvert et
  3     al., 1999).  However, it is well understood that mortality studies are limited in their ability to
  4     assess risk from diabetes mellitus. The recent paper by Longnecker and Michalek (2000) found a
  5     pattern suggesting that low levels of dioxin may influence the prevalence of diabetes. However,
  6     these results did not show an exposure-response relationship. Because it is the only study of its
  7     type to have been published, additional population-based studies are warranted to validate its
  8     findings. The most recent update of the Ranch Hand study shows a 47% excess of diabetes in the
  9     most heavily exposed group of veterans (Michalek etal., 1999).
10            Most of the data suggest that the diabetes is Type II, or adult-onset diabetes, rather than
11      insulin dependent, or Type I. Aging and obesity are the key risk factors for Type II diabetes.
12     However, dioxins may shift the distribution of sensitivity, putting people at risk at younger ages
13     or with less weight. Dioxin alters lipid metabolism in multiple species, including humans
14     (Sweeney et al., 1997; Pohjanvirta and Tuomisto, 1994). Dioxin also alters glucose uptake into
15     both human and animal cells in culture (Enan and Matsumura, 1994; Olsen et al., 1994).
16     Mechanistic studies have demonstrated that dioxin affects glucose transport (Enan and
17     Matsumura, 1994), a property under the control of the hypoxia response pathway (Ouiddir et al.,
18      1999). A key regulatory protein in this pathway is the partner of the AhR, Arnt (also known as
19      HIFl-beta) (Gu et al., 2000; Taylor and Zhulin, 1999).  Activation of therAhR by dioxin may
20      compete with other pathways, such as the hypoxia-inducible factor (HIF) pathway, for Arnt
21      (Gradin, et al., 1992).  Dioxin has also been shown to downregulate the insulin growth factor
22      receptor (Liu et al., 1992). These three issues — altered lipid metabolism, altered glucose
23      transport, and alterations in the insulin signaling pathway — all provide biological plausibility to
24     the association of dioxins with diabetes.
25            A causal relationship between diabetes and dioxin has not been established, although the
26      toxicologic data are suggestive of a plausible mechanism. Many questions are yet to be
27      answered.  Does diabetes alter the pharmacokinetics of dioxin? Diabetes is known to alter the
28      metabolism of several drugs in humans (Matzke et al., 2000) and may also alter dioxin
29      metabolism and kinetics.  As adult-onset diabetes is also associated with overweight, and body
30      composition has been shown to modify the apparent half-life of dioxin, could the rate of
31      elimination of dioxins be lowered in people with diabetes, causing them to have higher body
32      burdens? This may be relevant to the background population, but is hardly likely to be an
33      explanation in highly exposed populations. Key research needs are twofold.  The first is to
34     develop an animal model in which to study the association between dioxins and  diabetes and
35      glucose perturbation. Several rodent models for Type II diabetes exist and may be utilized. The
36      second is to conduct population-based incidence studies that take into account dioxin levels as
        9/22/00
46
DRAFT—DO NOT CITE OR QUOTE

-------
  1      well as the many known factors associated with diabetes. Although diabetes may cause the
  2      underlying pathology leading to death, it is often not attributed as the cause of death, and thus
  3      limits the utility of mortality studies.
  4
  5      2.2.6. Other Effects
  6      2.2.6.1. Elevated GGT
  7             As mentioned above, there appears to be a consistent pattern of increased GGT levels
  8      among individuals exposed to 2,3,7,8-TCDD-contaminated chemicals.  Elevated levels of serum
  9      GGT have been observed within a year after exposure in Seveso children (Caramaschi et al.,
10      1981; Mocarelli et al., 1986) and 10 or more years after cessation of exposure among TCP and
11      2,4,5-T production workers (May, 1982; Martin, 1984; Moses et al., 1984; Calvert et al., 1992)
12      and among Ranch Hands (Roegner et al., 1991; Grubbs et al., 1995).  All of these groups had a
13      high likelihood of substantial exposure to 2,3,7,8-TCDD. In addition, for those studies that
14      evaluated dose-response relationships with 2,3,7,8-TCDD levels, the effect was observed only at
15      the highest levels or categories of 2,3,7,8-TCDD and, in the NIOSH study, only in workers who
16      reported drinking high levels of alcohol. In contrast, although background levels of serum
17      2,3,7,8-TCDD suggested minimal exposure to Army Vietnam veterans, GGT was increased, at
18      borderline significance, among Vietnam veterans compared to non-Vietnam veterans (CDC
19      Vietnam Experience Study, 1988).  In addition, despite the increases observed in some
20      occupational cohorts, other studies of TCP production workers from West Virginia or Missouri
21      residents measured but did not report elevations in GGT levels (Suskind and Hertzberg, 1984;
22      Webb etal, 1989).
23             In clinical practice, GGT is often measured because it is elevated in almost all
24      hepatobiliary diseases and is used as a marker for alcoholic intake (Guzelian, 1985). In
25      individuals with hepatobiliary disease, elevations in GGT are usually accompanied by increases
26      hi other hepatic enzymes, e.g., AST and ALT, and metabolites, e.g., uro- and coproporphyrms.
27      Significant increases in hepatic enzymes other than GGT and metabolic products were not
28      observed in individuals whose GGT levels were elevated 10 or more years after exposure ended,
29      suggesting that the effect may be GGT-specific. These data suggest that in the absence of
30      increases in other hepatic enzymes, elevations in GGT are associated with exposure to 2,3,7,8-
31      TCDD, particularly among individuals who were exposed to high 2,3,7,8-TCDD levels.
32             The animal data with respect to 2,3,7,8-TCDD-related effects on GGT are sparse.
33      Statistically significant changes in hepatic enzyme levels, particularly AST, ALT, and alkaline
34      phosphatase (ALK), have been observed after exposure to 2,3,7,8-TCDD in rats  and hamsters
35      (Gasiewicz et al., 1980; Kociba et al., 1978; Olson et al., 1980).  Only one study evaluated GGT
36      levels (Kociba et al., 1978). Moderate but statistically nonsignificant increases were noted in rats
        9/22/00
47
DRAFT—DO NOT CITE OR QUOTE

-------
  1      fed 0.10 jxg/kg 2,3,7,8-TCDD daily for 2 years, and no increases were observed in control
  2      animals.
  3             In summary, GOT is the only hepatic enzyme examined that was found in a number of
  4      studies to be chronically elevated in adults exposed to high levels of 2,3,7,8-TCDD.  The
  5      consistency of the findings in a number of studies suggests that the elevation may reflect a true
  6      effect of exposure, but its clinical significance is unclear. Long-term pathological consequences
  7      of elevated GGT have not been illustrated by excess mortality from liver disorders or cancer, or
  8      in excess morbidity in the available cross-sectional studies.
  9             It must be recognized that the absence of an effect in a cross-sectional study, for example,
10      liver enzymes, does not obviate the possibility that the enzyme levels may have increased
11      concurrent to the exposure but declined after cessation. The apparently transient elevations in
12      ALT levels among the Seveso children suggest that hepatic enzyme levels other than GGT may
13      react in this manner to 2,3,7,8-TCDD exposure.
14
15      2.2.6.2. Thyroid Function
16             Many effects of 2,3,7,8-TCDD exposure in animals resemble signs of thyroid dysfunction
17      or significant alterations of thyroid-related hormones.  In the few human studies that examined
18      the relationship between 2,3,7,8-TCDD exposure and hormone concentrations in adults, the
19      results are mostly equivocal  (CDC  Vietnam Experience Study, 1988; Roegner et al.,  1991;
20      Grubbs et al., 1995; Suskind and Hertzberg, 1984). However, concentrations of thyroid binding
21      globulin (TBG) appear to be positively correlated with current levels of 2,3,7,8-TCDD in the
22      BASF accident cohort (Ott et al., 1994).  Little additional information on thyroid hormone levels
23      has been reported for production workers and none for Seveso residents, two groups with
24      documented high serum 2,3,7,8-TCDD levels.
25             Thyroid hormones play important roles in the developing nervous system in all vertebrate
26      species, including humans. In fact, thyroid hormones are so important in development that in the
27      United States all infants are tested for hypothyroidism shortly after birth.  Several studies of
28      nursing infants suggest that ingestion of breast milk with a higher dioxin TEQ may alter thyroid
29      function (Pluim et al., 1993; Koopman-Esseboom et al.,  1994c; Nagayama et al., 1997).
30      These findings suggest a possible shift in the distribution of thyroid hormones, particularly T4,
31      and point out the need for collection of longitudinal data to assess the potential for long-term
32      effects associated with developmental exposures. The exact processes accounting for these
33      observations in humans are unknown, but when put in perspective of animal responses, the
34      following might apply: dioxin increases the metabolism and excretion of thyroid hormone,
35      mainly T4, in the liver.  Reduced T4 levels stimulate the pituitary to secrete more thyroid
36      stimulating hormone (TSH), which enhances thyroid hormone production. Early in the
        9/22/00
48
DRAFT—DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
disruption process, the body can overcompensate for the loss of T4, which may result in a small
excess of circulating T4 to the increased TSH. In animals given higher doses of dioxin, the body
is unable to maintain homeostasis, and TSH levels remain elevated and T4 levels decrease.
A plausible mode of action for thyroid effects is described in Section 2.2.1.3 above.

2.2.6.3. Cardiovascular Disease
       Elevated cardiovascular disease has been noted in several of the occupational cohorts
(Steenland et al, 1999; Sweeney et al, 1997; Flesch-Janys et al., 1995) and in Seveso (Pesatori et
al., 1998), as well as in the rice oil poisonings. This appears to be associated with ischemic heart
disease and in some cases with hypertension.  Recent data from the Ranch Hand study indicates
that dioxin may be a possible risk factor for the development of essential hypertension (Grubbs et
al., 1995). Elevated blood lipids have also been seen  in several cohorts.  The association of
dioxins with heart disease in people has biological plausibility given the data in animals.  First is
the key role of hypoxia in heart disease, and the potential for involvement of the activated AhR in
blocking an hypoxic response (Gradin et al., 1996; Gu et al., 2000). Dioxin has been shown to
perturb lipid metabolism in multiple laboratory species (Pohjanvirta and Tuomisto, 1994). The
heart, in fact the entire vascular system, is a clear target for the adverse effects of dioxin in fish
and birds (Hornung et al., 1999;  Cheung et  al., 1981). In mammals, dioxin has been shown to
disturb heart rhythms at high doses in guinea pigs (Gupta et al., 1973; Pchjanvirta and Tuomisto,
1994).

2.2.6.4. Oxidative Stress
       Several investigators have hypothesized that the some of the adverse effects of dioxin and
related compounds may be associated with  oxidative stress.  Induction of CYP1A isoforms has
been shown to be associated with oxidative DNA damage (Park et al.,  1996). Altered
metabolism of endogenous molecules such  as estradiol can lead to the  formation of quinones and
redox cycling. This has been hypothesized to play a role in the enhanced sensitivity of female
rats to dioxin-induced liver tumors (Tritscher et al., 1996). Lipid peroxidation, enhanced DNA
single-strand breaks, and decreased membrane fluidity have been shown in liver as well as in
extrahepatic tissues following exposure to high doses of TCDD (Stohs, 1990). A dose- and time-
dependent increase in superoxide anion is caused in peritoneal macrophages by exposure to
TCDD (Alsharif et al.,  1994). A recent report that low-dose (0.15 ng TCDD/kg/day) chronic
exposure can lead to oxidative changes in several tissues in mice (Slezak et al., 2000) suggests
that this mechanism or mode of toxicity deserves further attention.
        9/22/00
                                           49
DRAFT—DO NOT CITE OR QUOTE

-------
 1
 2                       3. MECHANISMS AND MODE OF DIOXIN ACTION
 3
 4             Mechanistic studies can reveal the biochemical pathways and types of biological and
 5      molecular events that contribute to dioxin's adverse effects (See Part II, Chapter 2, for a detailed
 6      discussion). For example, much evidence indicates that TCDD acts via an intracellular protein
 7      (the AhR), which functions as a ligand-dependent transcription factor in partnership with a
 8      second protein (Arnt). Therefore, from a mechanistic standpoint, TCDD's adverse effects appear
 9      likely to reflect alterations in gene expression that occur at an inappropriate time and/or for an
10      inappropriately long time.  Mechanistic studies also indicate that several other proteins contribute
11      to TCDD's gene regulatory effects and that the response to TCDD probably involves a relatively
12      complex interplay between multiple genetic and environmental factors. If TCDD operates
13      through such a mechanism, as all evidence indicates, then there are certain constraints on the
14      possible models that can plausibly account for TCDD's biological effects and, therefore, on the
15      assumptions used during the risk assessment process (e.g., Poland, 1996; Limbird and Taylor,
16      1998).
17             Mechanistic knowledge of dioxin action may also be useful in other ways.  For example,
18      a further understanding of the ligand specificity and structure of the AhR will likely assist in the
19      identification of other chemicals to which humans are exposed that may-^add to, synergize, or
20      block the toxicity of TCDD.  Knowledge of genetic polymorphisms that influence TCDD
21      responsiveness may also allow the identification of individuals at greater risk from exposure to
22      dioxin. In addition, knowledge of the biochemical pathways that are altered by TCDD may help
23      identify novel targets for the development of drugs that can  antagonize dioxin's adverse effects.
24             As described below, biochemical and genetic analyses of the mechanisms by which
25      dioxin may modulate particular genes have revealed the outline of a novel regulatory system
26      whereby a chemical signal can alter cellular regulatory processes.  Future studies of dioxin action
27      have the potential to provide additional insights into mechanisms of mammalian gene regulation
28      that are of a broader interest.  Additional perspectives on dioxin action can be found in several
29      recent reviews (Birnbaum, 1994a,b; Schecter, 1994; Hankinson, 1995; Schmidt and Bradfield,
30      1996; Gasiewicz, 1997; Rowlands and Gustafsson, 1997; Denison et al., 1998; Hahn, 1998;
31      Wilson and Safe, 1998).
32             Knowledge of the mode(s) of action by which the broad class of chemicals known as
33      dioxins act may facilitate the risk assessment process by contributing to the weight of the
34      evidence for hazard characterization, and by imposing bounds on the models used to describe
35      possible responses of humans resulting from exposure to mixtures of these chemicals (see
36      Sections 2 and 5 of this document). The relatively extensive database on TCDD, as well as the
        9/22/00
50
DRAFT—DO NOT CITE OR QUOTE

-------
  1      more limited database on related compounds, has been reviewed with emphasis on the role of the
  2      specific cellular receptor for TCDD and related compounds, the AhR, in the mode(s) of action.
  3      This discussion will focus on summarizing the elements of the mode(s) of dioxin action that are
  4      relevant for understanding and characterizing dioxin risk for humans. These elements include:
  5             •   Similarities between humans and other animals with regard to receptor structure and
  6                function;
  7             •   The relationship between receptor binding and toxic effects; and
  8             •   The extent to which the purported mechanism(s) or mode(s) of action might
  9                contribute to the diversity of biological responses seen in animals and, to some extent,
10                in humans.
11
12             In addition, this section will identify important and relevant knowledge gaps and
13      uncertainties in the understanding of the mechanism(s) of dioxin action, and will indicate how
14      these may affect the approach to risk characterization.
15
16      3.1.  MODE VERSUS MECHANISM OF ACTION
17             In the context of revising its Cancer Risk Assessment Guidelines, the EPA has proposed
18      giving greater emphasis to use of all of the data in hazard characterization, dose-response
19      characterization, exposure characterization, and risk characterization (U-.-S. EPA, 1996; 1999).
20      One aid to the use of more information in risk assessment has been the definition of mode versus
21      mechanism of action.  Mechanism of action is defined as the detailed molecular description of
22      key events in the induction of cancer or other health endpoints.  Mode of action refers to the
23      description of key events and processes, starting with interaction of an agent with the cell,
24      through functional and anatomical changes, resulting in cancer or other health endpoints.
25      Despite a desire to construct detailed biologically based toxicokinetic and toxicodynamic models
26      to reduce uncertainty in characterizing risk, few examples have emerged.  Use of a mode of
27      action approach recognizes that, although all of the details may not have been worked out,
28      prevailing scientific thought supports moving forward using a hypothesized mode of action
29      supported by data. This approach is consistent with advice offered by the National Academy of
30      Sciences (NAS) National Research Council (NRC) in its  report entitled, Science and Judgment in
31      Risk Assessment (NAS/NRC,' 1994). Mode of action discussions help to provide answers to the
32      questions: How does the chemical produce its effect? Are there mechanistic data to support this
33      hypothesis?  Have other modes of action been considered and rejected? In order to demonstrate
34      that a particular mode of action is operative, it is generally necessary to outline the hypothesized
35      sequence of events leading to effects, identify key events that can be measured, outline the
36      information  that is available to support the hypothesis, and discuss those data that are
        9/22/00
51
DRAFT—DO NOT CITE OR QUOTE

-------
 1      inconsistent with the hypothesis or support an alternative hypothesis. Following this, the
 2      information is weighed to determine if there is a causal relationship between key precursor events
 3      associated with the mode of action and cancer or other toxicological endpoint in animals, and
 4      ultimately if this inference can be extended to humans.
 5
 6      3.2.  GENERALIZED MODEL FOR DIOXIN ACTION
 7             Dioxin and related compounds are generally recognized to be receptor-mediated
 8      toxicants. The generalized model has evolved over the years to appear as illustrated in Table 3-1
 9      and Figure 2-1.
10
11      3.2.1. The Receptor Concept
12             One of the fundamental concepts that influences our approach to risk assessment of
13      dioxin and related compounds is the receptor concept. The idea that a drug, hormone,
14      neurotransmitter, or other chemical produces a physiological response by interacting with a
15      specific cellular target molecule, i.e., a "receptor," evolved from several observations. First,
16      many chemicals elicit responses that are restricted to specific tissues. This observation implies
17      that the responsive tissue (e.g., the adrenal cortex) contains a "receptive" component whose
18      presence is required for the physiologic effect (e.g., cortisol secretion). Second, many chemicals
19      are quite potent. For example, picomolar to nanomolar concentrations of-numerous hormones
20      and growth factors elicit biological effects. This observation suggests that the target cell contains
21      a site(s) to which the particular chemical binds with high affinity.  Third, stereoisomers of some
22      chemicals (e.g., catecholamines, opioids) differ by orders of magnitude in their ability to produce
23      the same biological response. This observation indicates that the molecular shape of the
24      chemical strongly influences its biological activity. This, in turn, implies that the binding site on
25      or hi the target cell also has a specific, three-dimensional configuration. Together, these types of
26      observations support the prediction that the biological responses to some chemicals involve
27      stereospecific, high-affinity binding of the chemicals to specific receptor sites located on or in the
28      target cell.  Many of these characteristics were noted for TCDD and related compounds.
29             The availability of compounds of high specific radioactivity has permitted quantitative
30      analyses of their binding to cellular components in vitro. To qualify as a potential  "receptor," a
31      binding site for a given chemical must satisfy several criteria: (1) the binding site must be
32      saturable, i.e., the number of binding sites per cell should be limited; (2) the binding should be
33      reversible; (3) the binding affinity measured in vitro should be consistent with the potency of the
34      chemical observed in vivo; (4) if the biological response exhibits stereospecificity, so should the
35      in vitro binding; (5) for a series of structurally related chemicals, the rank order for binding
        9/22/00
52
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
affinity should correlate with the rank order for biological potency; and (6) tissues that respond to
the chemical should contain binding sites with the appropriate properties.
       The binding of a chemical ("ligand") to its specific receptor is assumed to obey the law of
mass action; that is, it is a bimolecular, reversible interaction. The concentration of the liganded,
or occupied, receptor [RL] is a function of both the ligand concentration [L] and the receptor
concentration [R] as shown in Equation 3-1:
                                        k,
                                                                                    (3-1)
                            [L] + [R]
[RL]
       Inherent in this relationship is the fact that the fractional occupancy (i.e., [RL]/[RJ) is a
function of ligand concentration [L] and the apparent equilibrium dissociation constant KD, which
is a measure of the binding affinity of the ligand for the receptor, that is, [RL]/[RJ = [L]/(KD+
[L]), where KD = [L] [RJ/[LR] = k2/k,. Therefore, the relationship between receptor occupancy
and ligand concentration is hyperbolic. At low ligand concentrations (where [L]«KD), a small
increase in [L] produces an approximately linear increase in fractional receptor occupancy. At
high ligand concentration (where [L]»KD), the fractional occupancy of the receptor is already
very close to  1, that is, almost all receptor sites are occupied. Therefore, a small increase in [L]
is likely to produce only a slight increase in receptor occupancy. Theseassues are discussed in
regard to TCDD binding to the AhR and dose-response in Part II, Chapter 8.
       Ligand binding constitutes only one aspect of the receptor concept. By definition, a
receptor mediates a response, and the functional consequences of the ligand-receptor binding
represent an essential aspect of the receptor concept. Receptor theory attempts to quantitatively
relate ligand binding to biological responses.  The classical "occupancy" model of Clark (1933)
postulated that (1) the magnitude of the biological response is directly proportional to the fraction
of receptors occupied and (2) the response is maximal when all receptors are occupied.
However, analyses of numerous receptor-mediated effects indicate that the relationship between
receptor occupancy and biological effect is not as straightforward as Clark envisioned. In certain
cases, no response occurs even when there is some receptor occupancy.  This suggests that there
may be  a threshold phenomenon that reflects the biological "inertia" of the response (Ariens et
al, 1960). In other cases, a maximal response occurs well before all receptors are occupied, a
phenomenon  that reflects receptor "reserve" (Stephenson, 1956). Therefore, one cannot simply
assume  that the relationship between fractional receptor occupancy and biological response is
linear. Furthermore, for a ligand (such as TCDD) that elicits multiple receptor-mediated effects,
one cannot assume that the binding-response relationship for a simple effect (such as enzyme
induction) will necessarily be identical to that for a different and more complex effect (such as
        9/22/00
                                           53
     DRAFT—DO NOT CITE OR QUOTE

-------
 1      cancer). The cascades of events leading to different complex responses (e.g., altered immune
 2      response to pathogens or development of cancer) are likely to be different, and other rate-limiting
 3      events likely influence the final biological outcome resulting in different dose-response curves.
 4      Thus, even though ligand binding to the same receptor is the initial event leading to a spectrum
 5      of biological responses, ligand-binding data may not always mimic the dose-effect relationship
 6      observed for particular responses.
 7             Another level of complexity is added when one considers different chemical ligands that
 8      bind to the same receptor. Relative potencies are determined by two properties of the ligand:
 9      affinity for the receptor and capacity to confer a particular response in the receptor (e.g., a
10      particular conformational change), also called efficacy (Stephehson, 1956).  Ligands with
11      different affinities and the same degree of efficacy would be expected to produce parallel dose-
12      response curves with the same maximal response within a particular model system. However,
13      ligands of the same affinity with different efficacies may result in dose-response curves that are
14      not parallel or that differ in maximal response. Many of these issues may apply to dioxin-
15      receptor interactions. To the extent that they do occur, they may present complications to use of
16      the toxic equivalency approach, particularly for extrapolation purposes. As described previously,
17      this argues strongly for the use of all available information in setting TEFs and highlights the
18      important role that scientific judgment plays in the face of incomplete mechanistic understanding
19      to address uncertainty.                                             ~-
20
21      3.2.2. A Framework to Evaluate Mode of Action
22             EPA hi its revised proposed guidelines for carcinogen risk assessment (U.S. EPA, 1999)
23      recommends the use of a structured approach to evaluating mode of action.  This approach is
24      similar to and builds upon an approach developed within the WHO/IPCS Harmonization Project
25      (WHO, 2000). Fundamentally, the approach uses a modification of the "Hill Criteria" (Hill,
26      1965), which have been used in the field of epidemiology for many years to examine causality
27      between associations'of exposures and effects. The framework calls for a summary description
28      of the postulated mode of action, followed by the identification of key events that are thought to
29      be part of the mode of action.  These key events are then evaluated as to strength, consistency,
30      and specificity of association with the endpoint under discussion.  Dose-response relationships
31      between the precursor key events are evaluated and temporal relationships are examined to be
32      sure that "precursor" events actually precede the induction of the endpoint.  Finally, biological
33      plausibility and coherence of the data with the biology are examined and discussed. All of these
34      "criteria" are evaluated and conclusions are drawn with regard to postulated mode of action.
35             In  the case of dioxin and related compounds, elements of such an approach are found for
36      a number of effects including cancer in Part II.  Application of the framework to dioxin and
        9/22/00
54
DRAFT—DO NOT CITE OR QUOTE

-------
 1      related compounds would now stop short of evaluating the association between the chemical or
 2      complex mixture and clearly adverse effects. Instead, the approach would apply to early events,
 3      e.g., receptor binding and intermediate events such as enzyme induction or endocrine impacts.
 4      Additional data will be required to extend the framework to most effects, but several have data
 5      that would support a framework analysis. Several of these are discussed below.
 6
 7      3.2.3. Mechanistic Information and Mode of Action; Implications for Risk Assessment
 8            A substantial body of evidence from investigations using experimental animals indicates
 9      that the AhR mediates the biological effects of TCDD. The key role of the AhR in the effects of
10      dioxin and related compounds is substantiated by four lines of research: (1) structure/activity
11      relationships; (2) responsive versus nonresponsive mouse strains; (3) mutant cell lines; and (4)
12      the development of transgenic mice in which the gene for the AhR has been "knocked out"
13      Birnbaum, 1994; Fernandez-Salguero et al., 1996; Lahvis and Bradfield, 1998). Dioxin appears
14      not to cause effects in the AhR knockout mouse (Fernandez-Salguero et al.,  1996; Lahvis and
15      Bradfield, 1998). It is clear that the AhR is necessary, but not sufficient, for essentially all of the
16      well-studied responses to dioxin.  The AhR functions as a ligand-activated transcription factor,
17      controlling the expression of specific genes via interaction with defined nucleotide sequences in
18      the promoter regions. In order to control transcription, the TCDD-AhR complex interacts with
19      another protein, Arnt, to bind to the dioxin response element. This complex is also bound by
20      other nuclear coactivators, and/or corepressors, to bind to the transcriptional complex and initiate
21      transcription (Gu et al., 2000). However, Arnt has many other partners"that control hypoxia
22      response, neuronal differentiation, morphological branching, etc. (Gu et al., 2000). It is possible
23      that there are other mechanisms of how dioxin initiates its toxic effects, apart from its direct
24      transcriptional  activation of drug metabolizing genes. It may be that the adverse effects of dioxin
25      may result from competition of the ligand-activated AhR with other Arnt partners (Gradin et al.,
26      1996). The AhR, Arnt, and Arnt partners are all members of the PAS family of basic helix-loop-
27      helix proteins that function as nuclear regulatory proteins (Gu et al., 2000). The PAS proteins are
28      highly conserved, with homologous proteins being present in prokaryotes. They play key roles in
29      circadian rhythms and development. The embryolethality of Arnt knockout mice, as well as the
30      reduced fertility and viability of the AhR knockout mice (Abbott et al., 1999), point to a key role
31      of these proteins in normal physiology.
32            Another potential mechanism by which TCDD can cause effects involves the
33      protein/protein interactions of the AhR.  When not bound to a ligand, the AhR exists in a
34      multimeric protein complex, involving two molecules of heat shock protein 90 as well as other
35      proteins, including AIP/XAP2/ara9, ara3, ara6, src, rel, and Rb (Carver et al., 1998; Enan and
36      Matsumura, 1996; Puga et al., 2000a). AIP/XAP2/ara9 is a 37 kilodalton (kd) protein that is
        9/22/00
55
DRAFT—DO NOT CITE OR QUOTE

-------
  1      related to known immunophilins and involved in control of signal transduction processes. C-src
  2      has been shown to be associated with the AhR in several tissues and is a tyrosine kinase (Enan
  3      and Matsumura, 1996). Dioxin has been known to cause a rapid increase in phosphorylation
  4      upon exposure. Recent studies have shown that rel, which is a key component of the NF-kappaB
  5      complex that controls apoptosis, binds to the AhR complex (Tian et al., 1999; Puga et al.,
  6      2000b). Similarly, several investigators have demonstrated an association between the AhR and
  7      the retinoblastoma protein; this has been shown to affect cell cycling (Puga et al., 2000a).
  8             Thus, the AhR may act as a negative regulator of key regulator molecules involved in
  9      phosphorylation, cell cycling, and apoptosis in its unliganded state.  Upon binding of TCDD,
10      these other proteins are now able to exert their effects. In addition, dioxin may act by competing
11      for Arnt, thus blocking key roles of other PAS regulatory proteins. Both of these mechanisms for
12      the effects of dioxin are in addition to the direct role of the ligand-bound form of the receptor in
13      control of transcription via the well-studied mechanism of binding to a dioxin-response element
14      inDNA.
15             Although studies using human tissues are much less extensive, it appears reasonable to
16      assume that dioxin's mode of action to produce effects in humans includes receptor-mediated key
17      events.  Studies using human organs and cells in culture are consistent with this hypothesis.  A
18      receptor-based mode of action would predict that, except in cases where the concentration of
19      TCDD is already high (i.e., [TCDD]~KD), incremental exposure to TCDD will lead to some
20      increase in the fraction of AhRs occupied.  However, it cannot be assumed that an increase in
21      receptor occupancy will necessarily elicit a proportional increase in all biological response(s)
22      because numerous molecular events (e.g., cofactors, other transcription factors, genes)
23      contributing to the biological endpoint are  integrated into the overall response. That is, the final
24      biological response should be considered as an integration of a series of dose-response curves
25      with each curve dependent on the molecular dosimetry for each particular step. Dose-response
26      relationships that will be specific for each endpoint must be considered when using mathematical
27      models to estimate the risk associated with exposure to TCDD.  It remains a challenge to develop
28      models that incorporate all the complexities associated with each biological response.
29      Furthermore, the parameters for each mathematical model may only apply to a single biological
30      response within a given tissue and  species.
31             Given TCDD's widespread distribution, its persistence, and its accumulation within the
32      food chain, it is likely that most humans are exposed to some level of dioxin; thus, the population
33      at potential risk is large and genetically heterogeneous. By analogy with the findings in inbred
34      mice, polymorphisms in the AhR probably exist in humans.  Therefore, a concentration of TCDD
35      that elicits a particular response in  one individual may not do so in another. For example, studies
36      of humans exposed to dioxin following an industrial accident at Seveso, Italy, failed to reveal a
        9/22/00
56
DRAFT—DO NOT CITE OR QUOTE

-------
 1      simple and direct relationship between blood TCDD levels and the development of chloracne
 2      (Mocarelli et al., 1991). These differences in responsiveness to TCDD may reflect genetic
 3      variation either in the AhR or in some other component of the dioxin-responsive pathway.
 4      Therefore, analyses of human polymorphisms in the AhR and Arnt genes have the potential to
 5      identify genotypes associated with higher (or lower) sensitivities to dioxin-related effects. Such
 6      molecular genetic information may be useful in the future for accurately predicting the health
 7      risks posed by dioxin to humans.
 8            Complex responses (such as cancer) probably involve multiple events and multiple genes.
 9      For example, a homozygous recessive mutation at the hr (hairless) locus is required for TCDD's
10      action as a tumor promoter in mouse skin (Poland et al., 1982). Thus, the hr locus influences the
11      susceptibility of a particular tissue (in this case, skin) to a specific effect of dioxin (tumor
12      promotion).  An analogous relationship may exist for the effects of TCDD in other tissues. For
13      example, TCDD may produce porphyria cutanea tarda only in individuals with inherited
14      uroporphyrinogen decarboxylase deficiency (Doss et al., 1984). Such findings suggest that, for
15      some adverse effects of TCDD, the population at risk may be limited to individuals with a
16      particular genetic predisposition.
17            Other factors can influence an organism's susceptibility to TCDD. For example, female
18      rats are more prone to TCDD-induced liver neoplasms than are males; this phenomenon is
19      related to the hormonal status of the animals (Lucier et al., 1991). In addition, hydrocortisone
20      and TCDD synergize in producing cleft palate in mice. Retinoic acid and TCDD produce a
21      similar synergistic teratogenic effect (Couture et al., 1990). These findings indicate that, in some
22      cases, TCDD acts in combination with hormones or other chemicals to produce adverse effects.
23      Such phenomena might also occur in humans. If so, the difficulty in assessing risk is increased,
24      given the diversity among humans in hormonal status, lifestyle (e.g., smoking, diet), and
25      chemical exposure.
26            Dioxin's action as a tumor promoter and developmental toxicant presumably reflects its
27      ability to alter cell proliferation and differentiation processes. There are several plausible
28      mechanisms by which this could occur. First, TCDD might activate a gene (or genes) that is
29      directly involved in tissue proliferation.  Second, TCDD-induced changes in hormone
30      metabolism may lead to tissue proliferation (or lack thereof) and altered differentiation secondary
31      to altered secretion of a trophic hormone. Third, TCDD-induced changes in the expression of
32      growth factor or hormone receptors may alter the sensitivity of a tissue to proliferative stimuli.
33      Fourth, TCDD-induced toxicity may lead to cell death, followed by regenerative proliferation.
34      These mechanisms likely differ among tissues and periods of development, and might be
35      modulated by different genetic and environmental factors. As such, this complexity increases the
36      difficulty associated with assessing the human health risks from dioxin exposure.
        9/22/00
57
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
       Under certain circumstances, exposure to TCDD may elicit beneficial effects. For
example, TCDD protects against the subsequent carcinogenic effects of PAHs in mouse skin,
possibly reflecting induction of detoxifying enzymes (Cohen et al., 1979; DiGiovanni et al.,
1980). In other situations, TCDD-induced changes in estrogen metabolism may alter the growth
of hormone-dependent tumor cells, producing a potential anticarcinogenic effect (Spink et al.,
1990; Gierthy et al., 1993).  However, several recent studies in mice indicate that the AhR has an
important role hi the genetic damage and carcinogenesis caused by components in tobacco smoke
such as BaP through its ability to regulate CYP1A1 gene induction (Dertinger et al., 1998;
Shimizu et al., 2000). TCDD's biological effects likely reflect a complicated'interplay between
genetic and environmental factors.  These issues complicate the risk assessment process for
dioxin.
       Thus, it is clear that the robust data base on mode(s) of dioxin action related to
biochemical effects and to clearly adverse effects supports an understanding of dioxins' impact
on biological and cellular processes. This database is among the best available for xenobiotic
chemicals. The short-comings described above will stimulate additional research to further
elucidate details in this understanding of the impact of dioxins but should not detract from the
recognition that, among data available to aid hazard characterization and risk assessment, these
are remarkably consistent and useful findings.
                        4. EXPOSURE CHARACTERIZATION

       This section summarizes key findings developed in the exposure portion of the Agency's
dioxin reassessment. The findings are developed in the companion document entitled "Part I:
Estimating Exposure to Dioxin-Like Compounds." This document is divided into four volumes:
(1) Executive Summary; (2) Sources of dioxin in the United States; (3) Properties,
Environmental Levels, and Background Exposures; and (4) Site-Specific Assessment Procedures.
Readers are encouraged to examine the more detailed companion document for further
information on the topics covered here and to see complete literature citations. The
characterization discussion provides cross references to help readers find the relevant portions of
the companion document.
       This discussion is organized as follows: (1) Sources; (2) Fate; (3) Environmental Media
and Food Concentrations; (4) Background Exposures; (5) Potentially Highly Exposed
Populations; and (6) Trends. The key findings are presented in italics.
        9/22/00
                                          58
DRAFT—DO NOT CITE OR QUOTE

-------
 1      4.1. SOURCES (Cross reference: Part I, Volume 2: Sources of Dioxin-Like Compounds
 2      in the United States)
 3            The CDD/CDFs have never been intentionally produced other than on a laboratory scale
 4      basis for use in scientific analysis. Rather, they are generated as unintended by-products in trace
 5      quantities in various combustion, industrial and biological processes.  PCBs, on the other hand,
 6      were commercially produced in large quantities, but are no longer commercially produced in the
 7      United States. EPA has classified sources of dioxin-like compounds into five broad categories:
 8
 9            1.  Combustion Sources.  CDD/CDFs are formed in most combustion systems. These
10                can include waste incineration (such as municipal solid waste, sewage sludge, medical
11                waste, and hazardous wastes), burning of various fuels (such as coal, wood, and
12                petroleum products), other high temperature sources (such as cement kilns), and
13                poorly or uncontrolled combustion sources (such as forest fires, building fires, and
14                open burning of wastes). Some evidence exists that very small amounts of dioxin-like
15                PCBs are produced during combustion, but they appear to be a small fraction of the
16                total TEQs emitted.
17            2.  Metals Smelting, Refining, and Processing Sources. CDD/CDFs can be formed
18                during various types of primary and secondary metals operations including iron ore
19                sintering, steel production, and scrap metal recovery.       .7-
20            3.  Chemical Manufacturing.  CDD/CDFs can be formed as by-products from the
21                manufacture of chlorine-bleached wood pulp, chlorinated phenols (e.g.,
22                pentachlorophenol, or PCP), PCBs, phenoxy herbicides (e.g., 2,4,5-T), and
23                chlorinated aliphatic compounds (e.g., ethylene bichloride).
24            4.  Biological and Photochemical Processes. Recent studies suggest that CDD/CDFs
25                can be formed under certain environmental conditions (e.g., composting) from the
26                action of microorganisms on chlorinated phenolic compounds. Similarly,  CDD/CDFs
27                have been reported to be formed during photolysis of highly chlorinated phenols.
28            5.  Reservoir Sources. Reservoirs are materials or places that contain previously formed
29                CDD/CDFs or dioxin-like PCBs and have the potential for redistribution and
30                circulation of these compounds into the environment. Potential reservoirs include
31                soils, sediments, biota, water, and some anthropogenic materials. Reservoirs become
32                sources when they have releases to the circulating environment.
33
34            Development of national estimates of annual environmental releases to air, water and
35      land is complicated by the fact that only a few facilities in most industrial sectors have been
36      evaluated for CDD/CDF emissions. Thus an extrapolation is needed to estimate national
        9/22/00
59
DRAFT—DO NOT CITE OR QUOTE

-------
  1      emissions.  The extrapolation method involves deriving an estimate of emissions per unit of
  2      activity (i.e., an emission factor) at the tested facilities and multiplying this by the total activity
  3      level in the untested facilities.  In order to convey the level of uncertainty in both the measure of
  4      activity and the emission factor, EPA developed a qualitative confidence rating scheme.  The
  5      confidence rating scheme, presented hi Table 4-1, uses qualitative criteria to assign a high,
  6      medium, or low confidence rating to the emission factor and activity level for those source
  7      categories for which emission estimates can be reliably quantified. The overall "confidence
  8      rating" assigned to a quantified emission estimate was determined by the confidence ratings
  9      assigned to the corresponding "activity level" and "emission factor." If the lowest rating
10      assigned to either the activity level or emission factor terms is "high," then the category rating
11      assigned to the emission estimate is high (also referred to as "A"). If the lowest rating assigned
12      to either the activity level or emission factor terms is "medium," then the category rating
13      assigned to the emission estimate is medium (also referred to as "B"). If the lowest rating
14      assigned to either the activity level or emission factor terms is "low," then the category rating
15      assigned to the emission estimate is low (also referred to as "C"). For many source categories,
16      either the emission factor information or activity level information were inadequate to support
17      development of reliable quantitative release estimates for one or more media. For some of these
18      source categories, sufficient information was available to make preliminary estimates of
19      environmental releases of CDD/CDFs or dioxin-like PCBs; however, the-confidence in the
20      activity level estimates or emission factor estimates was so low that the estimates cannot be
21      included in the sum of quantified emissions from sources with confidence ratings of A, B, or C.
22      These estimates were given an overall confidence class rating of D. For other sources, some
23      information exists suggesting that they may release  dioxin-like compounds; however, the
24      available data were judged to be insufficient for developing any quantitative emission estimate.
25      These estimates were given an overall confidence class rating of E.
26
27      4.1.1. Inventory of Releases
28             This dioxin reassessment has produced an inventory of source of environmental releases
29      of dioxin-like compounds for the United States (Table 4-2).  The inventory was developed by
30      considering all sources identified in the published technical and scientific literature and by the
31      incorporation of results from numerous individual emissions test reports of individual industrial
32      and combustion source facilities. In order to be representative of the United States, data
33      generated from U.S. sources of information were always given first priority for developing
34      emission estimates.  Data from other countries were used for making estimates in only a few
35      source categories where foreign technologies were judged similar to those found in the United
36      States and the U.S. data were judged to be inadequate.  The inventory is limited to sources whose
        9/22/00
60
DRAFT—DO NOT CITE OR QUOTE

-------
 1      releases can be reliably quantified (i.e., those with confidence ratings of A, B, or C as defined
 2      above). As discussed below, this document does provide preliminary estimates of releases from
 3      Class D sources, but they are presented separately from the Inventory.
 4            The inventory presents the environmental releases in terms of two reference years: 1987
 5      and 1995.  1987 was selected primarily because little empirical data existed for making source-
 6      specific emission estimates prior to this time.  1995 represents the latest year that could
 7      reasonably be addressed within the timetable for producing the rest of this document. EPA
 8      expects to conduct periodic revisions and updates to the source inventory in the future to track
 9      changes in environmental releases over time.
10            Figure 4-1 displays the emission estimates to air for sources included in the Inventory
11      and shows how the emission factors and activity levels were combined to generate emission
12      estimates.  Figure 4-2 compares the annual mean I-TEQ emission estimates to air for the two
13      reference years (i.e., 1987 and 1995).
14            The following conclusions are made for sources of dioxin-like compounds included in the
15      Inventory:
16
17      •  ,    EPA's best estimates of releases ofCDD/CDFs to air, 'water, and land from reasonably
18            quantifiable sources were approximately 3,300 gram (g) TEQDF~WHO98 (3000 g I-TEQ)
19            in 1995 and 14,000 g TEQD^ WHO98 (12,800 g I-TEQ)in 1987.  This finding is derived
20            directly from Table 4-2.
21      •      The environmental releases ofCDD/CDFs in the United States  occur from a wide variety
22            of sources, but are dominated by releases to the air from combustion sources. The
23            current (1995) inventory indicates emissions from combustion sources are more than an
24            order of magnitude greater than emissions from the sum of emissions from all other
25            categories.  Approximately 70% of all quantifiable environmental releases  were
26            contributed by air emissions from just three  source categories in 1995: municipal waste
27            incinerators (representing 38% of total environmental releases); backyard burning of
28            refuse in barrels (representing 19% of total releases) and medical waste incinerators
29            (representing 14% of total releases).
30      •      The decrease in estimated releases ofCDD/CDFs between 1987 and 1995
31            (approximately 76%)was due primarily to reductions in air emissions from municipal
32            and medical waste incinerators, and further reductions are anticipated. For both
33            categories, these emission reductions have occurred from a combination of improved
34            combustion and emission controls and from the closing of a number of facilities. EPA's
35            regulatory programs estimate that full compliance with recently promulgated regulations
36            should result in further reductions in emissions from the 1995 levels of more than 1800
        9/22/00
61
DRAFT—DO NOT CITE OR QUOTE

-------
  1             grams I-TEQ. These reductions will occur in the following source types: municipal
  2             waste combustors, medical waste incinerators, and various facilities which burn
  3             hazardous waste (see Part I, Volume 2 for further details about these reductions). No
  4             Federal regulations are in place or currently under development for limiting dioxin
  5             emissions from backyard burning of refuse in barrels. A number of states have general
  6             restrictions on the practice of backyard trash burning.
  7      •      Insufficient data are available to comprehensively estimate point source releases of
  8             dioxin-like compounds to -water. Sound estimates of releases to water are only available
  9             for chlorine bleached pulp  and paper mills (356 g I-TEQDF or TEQDF-WHO98 for 1987 and
10             28 g I-TEQDF or TEQDF-WHO98 for 1995) and the manufacture of ethylene dichloride
11             (EDC)/vinyl chloride monomer (VCM) (<1  g I-TEQDF or TEQDF-WHO98 in 1995).  Other
12             releases to water bodies that cannot be quantified on the basis of existing data include
13             effluents from publicly-owned treatment works (POTW) and most industrial/commercial
14             sources.  EPA's Office of Water estimates that when full compliance is achieved with
15             limitations on effluent discharges of CDD/CDF from chlorine bleached pulp and paper
16             mills, annual emissions will be reduced to 5 g I-TEQDF or TEQDF-WHO98.
17      •      Based on the available information, the inventory includes only a limited set of activities
18             that result in direct environmental releases to land.  The only releases to land quantified
19             in the national inventory are land application of sewage sludge oteommercial sludge
20             products (106.5 g I-TEQDF or 79 g TEQDF-WHO98 in 1995), land application of pulp and
21             paper mill wastewater sludges (2.0 g I-TEQDF or TEQDF-WHO98"in 1995), use of 2,4-D
22             pesticides (18.4 g I-TEQDF or 28.9 g TEQDF-WHO98), and manufacturing wastes from
23             EDC/VCM (<1 g I-TEQDF  or TEQDF-WHO98). Not included in the inventory's definition
24             of an environmental release is the disposal of sludge and ashes into approved landfills.
25      •      Significant amounts of dioxin-like compounds produced annually are not considered
26             environmental releases and, therefore, are not included in  the national inventory.
27             Examples include dioxin-like compounds generated internal to a process, but destroyed
28             before release, waste streams which are disposed of in approved landfills and are
29             therefore outside the definition of annual environmental releases, and products which
30             contain dioxin-like compounds but for which environmental releases, if any, cannot be
31             estimated.
32
33             The procedures and results of the U. S. inventory may have underestimated releases from
34      contemporary sources. A number of investigators have suggested that national inventories may
35      underestimate emissions because of the possibility of unknown sources. This claim has been
36      supported with mass balance analyses suggesting that deposition exceeds emissions (Rappe,
        9/22/00
62
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
1991; Harrad et al, 1992; Bruzy and Kites, 1995). The uncertainty, however, in both the
emissions and deposition estimates for the United States prevents the use of this approach for
reliably evaluating the issue. A variety of other arguments, however, indicate that the inventory
could underestimate emissions of dioxin-like compounds:

       •  A number of sources lacked sufficient data to include in the inventory, but did
          have limited evidence indicating that these sources can emit CDD/CDFs.
          These sources are listed in Tables 4-3 and 4-4 and include various
          components of the metals industries such as electric arc furnaces and
          foundries and uncontrolled or minimally controlled combustion practices (e.g.,
          accidental fires at landfills).

       •  The possibility remains that truly unknown sources exist. Many of the sources
          that are well accepted today were only discovered in the past 10 years. For
          example, CDD/CDFs were found unexpectedly in the wastewater effluent
          from bleached pulp and paper mills in the. mid 1980s. Ore sintering is now
          listed as one of the leading sources of CDD/CDF emissions in Germany, but
          was not recognized as a source until the early 1990s.

4.1.2. General Source Observations
       For any given time period, releases from both contemporary formation sources and
reservoir sources determine the overall amount of the dioxin-like compounds that are being
released to the open and circulating environment.  Because existing information is incomplete
with regard to quantifying contributions from contemporary and reservoir sources, it is not
currently possible to estimate total magnitude of release for dioxin-like compounds into the U.S.
environment from all sources. For example, in terms of 1995 releases from reasonably
quantifiable sources, this document estimates releases of 3,300 g TEQDF-WHO9g (3,000 g I-
TEQDF ) for contemporary formation sources, and 2,900 g I-TEQDF or TEQDF-WHO9g for
reservoir sources.  In addition, there remain a number of unquantifiable and poorly quantified
sources. No quantitative release estimates can be made for agricultural burning or for most
CDD/CDF reservoirs or for any dioxin-like PCB reservoirs. The preliminary estimate of 1995
poorly characterized contemporary formation sources is 1,500 g I-TEQDF or TEQDF-WHO98.
The preliminary release estimates for contemporary formation sources and reservoir sources are
presented in Table 4-3. Table 4-4 lists all the sources that have been reported to release dioxin-
like compounds but cannot be characterized on even a preliminary basis.
        9/22/00
                                           63
DRAFT—DO NOT CITE OR QUOTE

-------
  1             Additional observations and conclusions about all sources of dioxin-like compounds are
  2      summarized below:
  3
  4      •      The contribution of dioxin-like compounds to waterways from nonpoint source reservoirs
  5             is likely to be greater than the contributions from point sources. Current data are only
  6             sufficient to support preliminary estimates of nonpoint source contributions of dioxin-like
  7             compounds to water (i.e., urban storm water runoff and rural soil erosion). These
  8             estimates suggest that, on a nationwide basis, total nonpoint releases are significantly
  9             larger than point source releases.
10      •      Current emissions ofCDD/CDFs to the U.S. environment result principally from,
11             anthropogenic activities.  Evidence that supports this finding includes matches in time of
12             rise of environmental levels with time when general industrial activity began rising
13             rapidly (see trend discussion in Part I, Volume 3, Chapter 6), lack of any identified large
14             natural sources, and observations of higher CDD/CDF body burdens in industrialized vs.
15             less industrialized countries (see discussion on human tissue levels in Section 4.4).
16      •      Although chlorine is an essential component for the formation ofCDD/CDFs in
17             combustion systems, the empirical evidence indicates that for commercial scale
18             incinerators, chlorine levels in feed are not the dominant controlling factor for rates of
19             CDD/CDF stack emissions.  Important factors which can affect the rate of CDD/CDF
20             formation include the overall combustion efficiency, post-combustion flue gas
21             temperatures and residence times, and the availability of surface" catalytic sites to support
22             CDD/CDF synthesis. Data from bench, pilot and commercial scale combustors indicate
23             that CDD/CDF formation can occur by a number of mechanisms. Some of these data,
24             primarily from laboratory and pilot scale combustors, have shown direct correlation
25             between chlorine content in fuels and rates of CDD/CDF formation. Other data,
26             primarily from commercial scale combustors, show little relation between availability of
27             chlorine  in feeds and rates of CDD/CDF formation. The conclusion that chlorine in feed
28             is not a strong determinant of CDD/CDF emissions applies to the overall population of
29             commercial scale combustors. For any individual commercial scale combustor,
30             circumstances may exist in which changes in chlorine content of feed could affect
31             CDD/CDF emissions. For uncontrolled combustion, such as open burning of household
32             waste, the chlorine content of the waste may play a more significant role in rates of
33             CDD/CDF formation and release than is observed at commercial scale combustors. The
34             full discussion on this issue is presented in Part I, Volume 2, Chapter 2, Section 2.4.
        9/22/00
64
DRAFT—DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Dioxins are present in some ball clays, but insufficient data are available to estimate
•whether environmental releases occur during the mining and use.  Recent studies in the
United States and Europe have measured dioxins (principally CDDs) in some ball clays
and other related clays. As discussed in Part I, Volume 2, Chapter 13, it is likely that
dioxin present in ball clay is of a natural origin. Ball clay is principally used in the
manufacture of ceramics which involves firing the clay in high temperature kilns. This
activity may cause some portion of the CDDs contained in the clay to be released into the
air, but emission tests have not yet been conducted which would allow characterizing
these releases.
Data are available to estimate the amounts ofCDD/CDFs contained in only a limited
number of commercial products. No systematic survey has been conducted to determine
levels of dioxin-like compounds in commercial products. The available data does,
however, allow estimates to be made of the amounts of dioxin-like compounds in
bleached pulp (40 g I-TEQDF or TEQDF-WHO98 in 1995), POTW sludge used in fertilizers
(3.5 g I-TEQDF or 2.6 g TEQDF-WHO9g in 1995), pentachlorophenol-treated wood (8,400
g I-TEQDF or 4,800 g TEQDF-WHO98 in 1995), dioxazine dyes and pigments (<1 g I-
TEQDF or TEQDF-WHO98 in 1995) and 2,4-D (18.4 g I-TEQDF or 28.9 g TEQDF-WHO98 in
1995).
No significant release of newly formed dioxin-like PCBs is occurring in the  United States.
Unlike CDD/CDFs, PCBs were intentionally manufactured in the United States in large
quantities from 1929 until production was banned in 1977. Although it has been
demonstrated that small quantities of coplanar PCBs can be produced during waste
combustion, no strong evidence exists that the dioxin-like PCBs make  a significant
contribution to TEQ releases during combustion. The occurrences of dioxin-like PCBs in
the U.S. environment most likely reflects past releases associated with PCB  production,
use, and disposal.  Further support of this finding is based on observations of reductions
since 1980s in PCBs in Great Lakes sediment and other areas.
It is unlikely that the emission rates of CDD/CDFs from known sources correlate
proportionally with general population exposures.  Although the Emissions Inventory
shows the relative contribution of various sources to total emissions, it cannot be assumed
that these sources make the same relative contributions to human exposure.  It is quite
possible that the major sources of dioxin in food (see  discussion in Part I, Volume 2,
Chapter 2, Section 2.6 indicating that the diet is the dominant exposure pathway for
humans) may not be those sources that represent the largest fractions of current total
emissions in the United States. The geographic locations of sources relative to the areas
from which much of the beef, pork, milk, and fish come is important to consider. That is,
        9/22/00
                                    65
DRAFT—DO NOT CITE OR QUOTE

-------
  1             much of the agricultural areas that produce dietary animal fats are not located near or
  2             directly downwind of the major sources of dioxin and related compounds.
  3      •      The contribution of reservoir sources to human exposure may be significant. Several
  4             factors support this finding:
  5             1)  Because the magnitude of releases from current sources of newly formed PCBs are
  6             most likely negligible, human exposure to the dioxin-like PCBs is thought to be derived
  7             almost completely from reservoir sources. Key pathways involve releases from both soils
  8             and sediments to both aquatic and terrestrial food chains. As discussed in Volume 3,
  9             Chapter 4, Section 4.4.2, one third of general population TEQDFP exposure is due to
10             PCBs. Thus, at least one third of the overall risk from dioxin-like compounds comes
11             from reservoir sources.
12             2)  CDD/CDF releases from soil via soil erosion and runoff to waterways may be
13             significant. These releases appear to be greater than releases to water from the primary
14             sources included in the inventory. CDD/CDFs in waterways can bioaccumulate in fish
15             leading to human exposure via consumption offish.  As discussed in Volume 3, Chapter
16             4, Section 4.4.2, fish consumption makes up about one fifth of the total general
17             population CDD/CDF TEQ exposure.  This suggests that a significant portion of the
18             CDD/CDF TEQ exposure  could be due to releases from the soil reservoir. It is not
19             known, however, how much of the soil erosion and runoff represents recently deposited
20             CDD/CDFs from primary  sources or longer term accumulation.  Much of the eroded soil
21             comes from tilled agricultural lands which would include a mix of CDD/CDFs from
22             various deposition times. The age of CDD/CDFs in urban runoff is less clear.
23             3) Potentially, soil reservoirs could have vapor and particulate releases which deposit on
24             plants and enter the terrestrial food chain. The magnitude of this contribution, however,
25             is unknown.
26
27      4.2. ENVIRONMENTAL FATE  (Cross reference: Part I, Volume 3, Chapter 2)
28             The estimates of environmental releases are presented above  in terms of TEQs. This is
29      done for convenience in presenting summary information and to facilitate comparisons across
30      sources. For purposes of environmental fate modeling, however, it is important to use the
31      individual CDD/CDF and PCB congeners values, rather than TEQs.  This is because the
32      physical/chemical properties of individual dioxin congeners vary and will behave differently in
33      the environment. For example, the relative mix of congeners released from a stack cannot be
34      assumed to remain constant during transport through the atmosphere and deposition to various
35      media. The full congener-specific  release rates for most sources are given in an electronic
36      database which is available as a companion to this document (Database of Sources of
        9/22/00
66
DRAFT—DO NOT CITE OR QUOTE

-------
  1      Environmental Releases of Dioxin-Like Compounds in the United States (EPA/600/P-
  2      98/002Ab). In Part I, Volume 4, site specific procedures are provided for estimating the impact
  3      of emissions on local populations and this section emphasizes that congener specific emission
  4      values should be used in modeling their environmental fate. Finally, it is important to recognize
  5      that this document does not use source release estimates to generate background population
  6      intake/risk estimates (rather these estimates are derived primarily from food levels and
  7      consumption rates).
  8            Dioxin-like compounds are -widely distributed in the environment as a result of a number
  9      of physical and biological processes.  The dioxin-like compounds are essentially insoluble in
10      water, generally classified as semivolatile, and tend to bioaccumulate in animals.  Some evidence
11      has shown that these compounds can degrade in the environment,  but in general they are
12      considered very persistent and relatively immobile in soils and sediments. These compounds are
13      transported through the atmosphere as vapors or attached to airborne particulates and can be
14      deposited on soils, plants, or other surfaces (by wet or dry deposition). The dioxin-like
15      compounds enter water bodies primarily via direct deposition from the atmosphere, or by surface
16      runoff and erosion.  From soils, these compounds can reenter the atmosphere either as
17      resuspended soil particles or as vapors. In water, they can be resuspended into the water column
18      from sediments, volatilized out of the surface waters into the atmosphere or become buried in
19      deeper sediments.  Immobile sediments appear to serve as permanent sinks for the dioxin-like
20      compounds.  Though not always considered an environmental compartment, these compounds
21      are also found in anthropogenic materials (such as PCP) and have  the potential to be released
22      from these materials into the broader environment.
23            Atmospheric transport and deposition of the dioxin-like compounds are a primary means
24      of dispersal of these compounds throughout the environment. The dioxin-like compounds can be
25      measured in wet and dry deposition in most locations including remote areas. Numerous studies
26      have shown that they are commonly found in soils throughout the  world. Industrialized countries
27      tend to show similar elevated  concentrations in soil, and detectable levels have been found in
28      nonindustrialized countries. The only satisfactory explanation available for this distribution is air
29      transport and deposition.  Finally, by analogy these compounds would be expected to behave
30      similarly to other compounds  with similar properties, and this mechanism of global distribution
31      is becoming widely accepted for a variety of persistent organic compounds.
32            The two primary pathways for the dioxin-like compounds to enter the ecological food
33      chains and human diet are air-to-plant-to-animal and water/sediment-to-fish.  Vegetation
34      receives these compounds via atmospheric deposition in the vapor and particle phases. The
35      compounds are retained on plant surfaces and bioaccumulated in the fatty tissues of animals that
36      feed on these plants. Vapor phase transfers onto vegetation have been experimentally shown to
        9/22/00
67
DRAFT—DO NOT CITE OR QUOTE

-------
  1      dominate the air-to-plant pathway for the dioxin-like compounds, particularly for the lower
  2      chlorinated congeners. In the aquatic food chain, dioxins enter water systems via direct
  3      discharge or deposition and runoff from watersheds. Fish accumulate these compounds through
  4      their direct contact with water, suspended particles, bottom sediments, and through their
  5      consumption of aquatic organisms. Although these two pathways are thought to normally
  6      dominate contribution to the commercial food supply, others can also be important.  Elevated
  7      dioxin levels in cattle resulting from animal contact with PCP-treated wood have been
  8      documented by the U.S. Department of Agriculture.  Animal feed contamination episodes have
  9      led to elevations of dioxins in poultry in the United States, milk in Germany, and meat/dairy
10      products in Belgium.
11
12      4.3.  ENVIRONMENTAL MEDIA AND FOOD CONCENTRATIONS (Cross reference:
13           Part I, Volume 3, Chapter 3)
14            Background levels of dioxin-like compounds in various environmental media including
15      food are presented in Table 4-5 in terms of means, variability and sample sizes used to support
16      the estimates. Estimates for background levels of dioxin-like compounds in environmental
17      media are based on a variety of studies conducted at different locations in North America. Of the
18      studies available for this compilation, only those conducted in locations representing
19      "background"  were selected. The amount and representativeness of the-data vary, but in general
20      these data were derived from studies that were not designed to estimate national background
21      means. The environmental media concentrations were similar to studies in Western Europe.
22      These data are the best available for comparing with site-specific values. Because of the limited
23      number of locations examined, it is not known if these estimates adequately capture the full
24      national variability. As new data are collected, these ranges are likely to be expanded and
25      refined. The limited data on dioxin-like PCBs in environmental media are summarized in this
26      document (Part I, Volume 3, Chapter 3).
27            Estimates for levels of dioxin-like compounds in food are based on data from a variety  of
28      studies conducted in North America. Beef, pork, and poultry were derived from statistically
29      based national surveys. Milk estimates were derived from a survey of a nationwide milk
30      sampling network. Dairy estimates were derived from milk fat concentrations, coupled with
31      appropriate assumptions for the amount of milk fat in dairy products.  The background egg
32      concentrations were based on an analysis of 15 egg samples collected from retail stores in 8
33      states (CA, OH, GA, NY, PA, OR, MN, WS; 2 samples/state except one in OR), where each
34      sample was a composite of 24 individual eggs (i.e. 15 samples represented 360 eggs). The fish
35      data, as discussed below, were derived from multiple studies with samples collected both
36      directly from water bodies and from retail outlets.  All fish concentrations were expressed on the
        9/22/00
68
DRAFT—DO NOT CITE OR QUOTE

-------
  1      basis of fresh weight in edible tissue. As with other environmental media, food levels found in
  2      the United States are similar to levels found in Europe.
  3             The procedure to evaluate background fish exposures emphasizes the use of both species-
  4      specific consumption rates and species-specific concentrations. EPA's National
  5      Bioaccumulation Study (U.S. EPA, 1992b) provides some species-specific information on
  6      freshwater/estuarine fish caught in the wild at various locations in the United States.  Additional
  7      species-specific data on store bought fish are available from studies conducted by the Food and
  8      Drag Administration during the mid to latter 1990s (Jensen and Bolger, 2000; Jensen et al.,
  9      2000).  An important aspect of the U.S. Food and Drug Administration (FDA) studies is that they
10      include data on store-bought catfish, tuna, shellfish, and salmon which are some of the most
11      highly consumed species. Accordingly, the data used to characterize CDD/CDF fish levels are
12      much improved over previous estimates with over 300 individual samples and good
13      representation of the most highly consumed species. However, the levels of dioxins in fish
14      remain more uncertain than the other foods. The compilation of data from different studies still
15      lacks the geographic coverage and statistical power of the other food surveys. The EPA and FDA
16      studies did not address dioxin-like PCBs, rather these are based on a much smaller data set
17      derived from the open literature.  Also, the estimates of dioxin intake resulting from fish
18      consumption do not include consumption of fish oils. Currently insufficient data are available to
19      support estimates of dioxin intake from direct fish oil consumption.    •--
20             The general population dioxin intake calculations used in this document are a function of
21      both consumption rate and dioxin concentration in food. The concentration data used in this
22      document were measured in raw foods.  Therefore, if cooking significantly alters the  dioxin
23      concentration in consumed portions it must be accounted for in estimating dioxin intake. This
24      issue has been examined  in a number of studies which measured the effects of cooking on the
25      levels of CDDs, CDFs and PCBs in foods (see Part I, Volume 3, Chapter 3, Section 3.7.5).
26      These studies have a range of results depending on food type and cooking method. Most of the
27      cooking experiments suggested that cooking reduces the total amount of dioxins in food but
28      causes relatively little change in its concentration. Although some cooking experiments have
29      shown increases and others have shown decreases in dioxin concentrations, the relative
30      prevalence of these impacts have not been established. Therefore given that most experiments
31      show little change and that others show change in both directions, the most reasonable
32      assumption that can be made from the existing data is that dioxin concentration in uncooked food
33      is a reasonable surrogate  for dioxin concentration in cooked food. Although cooking in general
34      does not reduce dioxin concentration hi food, some specific food preparation practices can be
35      adopted that can reduce dioxin intake by significantly reducing overall animal fat consumption.
        9/22/00
69
DRAFT—DO NOT CITE OR QUOTE

-------
 1      For example, carefully trimming fat from meat, removing skin from chicken and fish and
 2      avoiding cooking in animal fats should reduce both animal fat and dioxin intake.
 3            Some evidence from Europe suggests that during the 1990s a decline has occurred in
 4      concentrations of dioxins and furans in food products, particularly dairy products (see Part I,
 5      Volume 3, Chapter 6, Section 6.5).  For example, the United Kingdom's Ministry of Agriculture,
 6      Fisheries, and Food (MAFF) collected milk samples in 1990 and again from similar locations in
 7      1995. In 1990, the I-TEQDF ranged from  1.1 to 3.3 ppt, while the 1995 I-TEQDF ranged from 0.7
 8      to 1.4. In Germany, a sampling of 120 dairy products in 1994 found I-TEQDF concentrations that
 9      were 25% lower than a similar sampling program in 1990. Liem et al. (2000) reports on a
10      European cooperative study coordinated by the National Institute of Public Health and the
11      Environment in the Netherlands, and the Swedish National Food Administration. Ten countries
12      supplied data on food concentrations, food consumption patterns, and other data used to evaluate
13      exposure to dioxins in Europe.  Some of the data suggested reductions in concentrations over
14      time, but the available information was insufficient to draw general conclusions. No systematic
15      study of temporal trends in dioxin levels in food has been conducted in the United States.
16      Although not statistically based, one U.S. study examined dioxin levels in 14 preserved food
17      samples from various decades in the twentieth century (Winters et al.,  1998). It was found that
18      meat samples of the 1950s through the 1970s had concentrations that were 2-3 times higher for
19      the CDD/CDF TEQs and about 10 times higher for the PCB TEQs, as compared to current meat
20      concentrations.
21                                                             -       '
22      4.4. BACKGROUND EXPOSURES (Cross reference: Part I, Volume 3, Chapter 4)
23      4.4.1. Tissue Levels
24            The average CDD/CDF/PCB tissue, level for the general adult U.S. population appears to
25      be declining, and the best estimate of current (late 1990s) levels is 25 ppt (TEQDFP-WHO98, lipid
26      basis).
27            The tissue samples collected in North America in the late 1980s and early 1990s showed
28      an average TEQDFP-WHO98 level of about 55 pg/g lipid.  This finding is supported by a number of
29      studies which measured dioxin levels in adipose, blood, and human milk, all conducted in North
30      America.  The number of people in most of these studies, however, is relatively small and the
31      participants were not statistically selected in ways that assure their representativeness of the
32      general U.S. adult population.  One study, the,1987 NationaLHumaiLAdipose Tissue Survey
33      (NHATS), involved over 800 individuals and provided broad geographic coverage, but did not
34      address coplanar PCBs.  Similar tissue levels of these compounds have been measured in Europe
35      and Japan during similar time periods.
        9/22/00
70
DRAFT—DO NOT CITE OR QUOTE

-------
 1            Because dioxin levels in the environment have been declining since the 1970s (see trends
 2      discussion in Part I, Volume 3, Chapter 6), it is reasonable to expect that levels in food, human
 3      intake, and ultimately human tissue have also declined over this period. The changes in tissue
 4      levels are likely to lag the decline seen in environmental levels, and the changes in tissue levels
 5      cannot be assumed to occur proportionally with declines in environmental levels. CDC (2000)
 6      summarized levels of CDDs, CDFs, and PCBs in human blood collected during the time period
 7      1995 to 1997. The individuals sampled were all U.S. residents with no known exposures to
 8      dioxin other than normal background.  The blood was collected from 316 individuals in six
 9      different locations with an age range of 20 to 70 years. While the samples hi this data set were
10      not collected hi a manner that can be considered statistically representative of the national
11      population and lack wide geographic coverage, they are judged to provide a better indication of
12      current tissue levels in the United States than the earlier data.  PCBs 105,  118, and 156 are
13      missing from the blood data for the comparison populations reported by CDC (2000). These
14      congeners account for 62% of the total PCB TEQ estimated in the early 1990s. Assuming that
15      the missing congeners from the CDC study data contribute the same proportion to the total PCB
16      TEQ as in earlier data, they would increase our estimate of current body burdens by another 3.3
17      pg TEQ/g lipid for a total PCB TEQ of 5.3 pg/g lipid and a total of 25.4 pg TEQDFP-WHO98 /g
18      lipid.  A summary of the CDC (2000) data is shown in Table 4-6.
19            A portion of the CDC blood data were plotted as a function of age-. This plot, shown in
20      Figure 4-3, indicates that blood levels generally increase with age and also that the variability in
21      blood levels increase with age.
22            This finding regarding a current tissue level of 25.4 pg/g lipid TEQDFP-WHO98 is further
23      supported by the observation that this mean tissue level is consistent with our best estimate of
24      current adult intake,  i.e.,  65 pg WHO9g-TEQDFP/d. Using this intake in a one-compartment,
2 5      steady-state pharmacokinetic model yields a tissue level estimate of about 11.1 pg TEQ/g lipid
26      (assumes TEQDFP has an  effective half-life of 7.1 yr, 80% of ingested dioxin is absorbed into the
27      body, and lipid weight is 25% of the adult assumed body weight of 70 kg, or 17.5 kg). Because
28      intake rates appear to have declined in recent years and steady-state is not likely to have been
29      achieved, it is reasonable to observe higher measured tissue levels, such as the 25.4 pg TEQ/g
30      lipid that was observed, than predicted by the model.
31            Characterizing national background levels of dioxins in tissues is uncertain because the
32      current data cannot be considered statistically representative of the general population. It is also
33      complicated by the fact that tissue levels are a function of both age and birth year.  Because
34      intake levels have varied over time, the accumulation of dioxins in a person who turned 50 years
35      old in 1990 is different than in a person who turned  50 in 2000. Future studies should help
36      address these uncertainties. The National Health and Nutrition Examination Survey (NHANES)
        9/22/00
71
DRAFT—DO NOT CITE OR QUOTE

-------
  1     began a new national survey in 1999 that will measure blood levels of CDDs, CDFs, and PCBs
  2     126, 77,169, and 81 in about 1,700 people per year (see http:www.cdc.gov/nchs/nhanes.htm).
  3     The survey is conducted at 15 different locations per year and is designed to select individuals
  4     statistically representative of the civilian U.S. population in terms of age, race, and ethnicity.
  5     These new data should provide a much better basis for estimating national background tissue
  6     levels and evaluating trends than the currently available data.
  7
  8     4.4.2. Intake Estimates
  9           Adult daily intakes ofCDD/CDFs and dioxin-like PCBs are estimated to average 41 and
10     24 pg TEQDFP- WHOyg/day,  respectively, for a total intake of 65 pg/day TEQDFP- WHO98.  Daily
11     intake is estimated by combining exposure media concentrations (food, soil, air) with contact
12     rates (ingestion, inhalation). Table 4-7 summarizes the media concentrations, contact rates and
13     resulting intake estimates.
14           The intake estimate is supported by an extensive database on food consumption rates and
15     estimates of dioxin-like compounds in food (as discussed above). Pharmacokinetic (PK)
16     modeling provides further support for the intake estimates. Applying a simple steady-state PK
17     model to an adult average blood level of 25 ppt TEQDFP-WHO98 (on a lipid basis) yields a daily
18     intake of 146 pg TEQDFP-WHO98/day (assumes TEQDFP has an effective half-life of 7.1 yr, 80% of
19     ingested dioxin is absorbed into the body, and lipid weight is 25% of the:adult assumed body
20     weight of 70 kg, or 17.5 kg). This PK-modeled CDD/CDF/PCB  intake estimate is about 2.2
21      times higher than the direct intake estimate of 65 pg TEQDFP-WHO98/day. This difference is to be
22     expected with this application of a simple steady-state PK model to current average adipose
23     tissue concentrations. Current adult tissue levels reflect intakes from past exposure levels that
24     are thought to be higher than current levels (see Part I, Volume 3, Chapter 6). Because the
25     direction and magnitude of the difference in intake estimates between the two approaches are
26     understood, the PK-derived value is judged supportive of the pathway-derived estimate.  It
27     should be recognized, however, that the pathway-derived value will underestimate exposure if it
28     has failed to capture all significant exposure pathways.
29
30     4.4.3. Variability in Intake Levels
31            CDD/CDF and dioxin-like PCS intakes for the general population may extend to levels at
32     least three times higher than the mean.  Variability in general population exposure is primarily
33     the result of the differences in dietary choices that individuals make.  These are differences in
34     both quantity and types of food consumed.  An increased background exposure can result from
35     either a diet that favors consumption of foods high in dioxin content or a diet that is
36     disproportionately high in overall consumption of animal fats.
        9/22/00
72
DRAFT—DO NOT CITE OR QUOTE

-------
  1             The best data available to determine the variability of total fat consumption comes from
  2      several analyses of the Bogalusa Heart Study (Cresanta et al., 1988; Nicklas et al, 1993; Nicklas
  3      et al, 1995; Nicklas et al., 1995; Frank et al., 1986). These data show that the 95th percentile of
  4      total fat consumption is about twice the mean and the 99th percentile is approximately three
  5      times the mean. For a diet which has a broad distribution of animal fats (as does the typical U.S.
  6      diet), this same distribution can be assumed for dioxin intake.
  7             Although body burden data cannot be assumed to be perfectly representative of current
  8      intakes (because they reflect past exposures as well as current ones), they also provide some
  9      support for this finding. This is based on the observation that the  95th percentile blood level in
10      the CDC (2000) study was almost twice the mean level.
11             Intakes ofCDD/CDFs and dioxin-like PCBs are over three times higher for a young child
12      as compared to that of an adult, on a body -weight basis.  This is based on combining age-
13      specific food consumption rate and average food concentrations, as was done above for adult
14      intake estimates (see Table 4-8).
15             Only four of the 17 toxic CDD/CDF congeners and one of the 11 toxic PCBs account for
16      most of the toxicity in human tissue concentrations: 2,3,7,8-TCDD, 1,2,3,7,8-PCDD,
17      1,2,3,6,7,8-HxCDD, 2,3,4,7,8-PCDF, and PCB 126. This finding is derived directly from the
18      data described earlier on human tissue levels and is supported by intake estimations indicating
19      that these congeners are also the primary contributors to dietary dose. These five compounds
20      make up about 80% of the total WHO9g-TEQ tissue level.
21
22      4.5.   POTENTIALLY HIGHLY EXPOSED POPULATIONS OR DEVELOPMENTAL
23            STAGES (Cross reference: Part I, Volume 3, Chapter 5)
24             As discussed earlier, background  exposures to dioxin-like compounds may extend to
25      levels at least three times higher than the mean. This upper range is assumed to result from the
26      normal variability of diet and human behaviors. Exposures from local elevated sources or
27      exposures resulting from unique diets would be in addition to this background variability.  Such
28      elevated exposures may occur in small segments of the population such as individuals living near
29      discrete local  sources.  Nursing infants represent a special case: for a limited portion of their
30      lives, these individuals may have elevated exposures on a body weight basis when compared
31      with non-nursing infants and adults.
32             Dioxin contamination incidents involving the commercial food supply have occurred in
33      the  United States and other countries. For example, in the United States, contaminated ball clay
34      was used as an anti-caking agent in soybean meal and resulted in elevated dioxin levels in some
35      poultry and catfish. This incident, which occurred in  1998, involved less than 5% of the national
36      poultry production and has since been eliminated. Elevated dioxin levels have also been
        9/22/00
73
DRAFT—DO NOT CITE OR QUOTE

-------
 1      observed in a few beef and dairy animals where the contamination was associated with contact
 2      with pentachlorophenol-treated wood. Evidence of this kind of elevated exposure was not
 3      detected in the national beef survey. Consequently its occurrence is likely to be low, but it has
 4      not been determined.  These incidents may have led to small increases in dioxin exposure to the
 5      general population. However, it is unlikely that such incidents have led to disproportionate
 6      exposures to populations living near where these incidents have occurred, because in the United
 7      States, meat and dairy products are highly distributed on a national scale. If contamination
 8      events were to occur in foods that are predominantly distributed on a local or regional scale, then
 9      such events could lead to highly exposed local populations.
10            Elevated exposures associated with the workplace or industrial accidents have also been
11      documented. U.S. workers in certain segments of the chemical industry had elevated levels of
12      TCDD exposure, with some tissue measurements in the thousands of ppt TCDD. There is no
13      clear evidence that elevated exposures are currently occurring  among United States workers.
14      Documented examples of past exposures for other groups include certain Air Force personnel
15      exposed to Agent Orange during the Vietnam War and people exposed as a result of industrial
16      accidents in Europe and Asia.
17            Consumption of breast milk by nursing infants leads to higher levels of exposure and
18      higher body burdens ofdioxins during early years of life as compared with  non-nursing infants
19      (Part I, Volume 3, Chapter 5, Section 5.2).                          •--
20            Three German studies have compared dioxin levels in infants who have been breast-fed
21      with those who have been formula-fed. All have shown elevations in the concentrations of
22      dioxins in infants being breast-fed. Collectively these studies  included 99 infants and found that
23      blood levels (in units of pg TEQDF-WHO98/g lipid - i.e., dioxin-like PCBs not included) in infants
24      aged 4-12 months were generally more than 20 in nursing infants and less than 5 in formula fed
25      infants.
26            U.S. dioxin intakes from nursing were calculated using time dependent values for breast
27      milk concentrations, consumption rates and body weights. These calculations estimated an
28      intake immediately after birth of 242 pg TEQDFP-WHO9g/kg/day.  This dropped to 22 pg TEQDFP-
29      WHOgg/kg/day after 12 months of nursing. The average intake over one year of nursing was
30      calculated to be 92 pg TEQDFP-WHO98/kg/day.  The cumulative intake for a one year nursing
31      scenario represented about 12% of the total lifetime cumulative intake (see  Part I, Volume 3,
32      Chapter 5, Section 5.2 for details on these calculations).
33            The CDC (1997) reported that in 1995, 55% of all babies experience some breast feeding,
34      with about half of those breast feeding beyond 5 months. The average duration of breast feeding
35      was 28.7 weeks.  In a policy statement, the American Academy of Pediatrics (1997) stated that
36      exclusive breast feeding is ideal nutrition and sufficient to support optimal growth and
        9/22/00
74
DRAFT—DO NOT CITE OR QUOTE

-------
  1      development for 6 months after birth. They recommended that breast feeding continue for at
  2      least 12 months, and thereafter for as long as mutually desired.
  3             To better evaluate the impact of nursing on infants, changes in body burden were
  4      calculated using a one-compartment, first-order pharmacokinetic model. Changes in TEQ tissue
  5      concentration over time were modeled for a variety of nursing scenarios: formula only, 6 weeks
  6      nursing, 6 months nursing, and one year. These scenarios reasonably capture the range of current
  7      nursing practice. This modeling effort required using the intake assumptions described earlier
  8      and a variety of additional assumptions including: the fraction of the oral dose which is absorbed
  9      into the body, changes in body weight over time, and changes in body fat fraction over time.
10      Assumptions were also made about changes in the biological half-life of dioxins as a function of
11      body fat fraction. For the infant, the half-life was less than one year, and during adulthood the
12      half-life increased as the fraction-of body fat increased.  The short half-life at birth was based on
13      a study by Kreuzer et al. (1997) and the longer half-life during the later years of life, when body
14      fat fraction increased, was based on a model presented in Michalek et al. (1996). The complete
15      set of input values are listed in Part I, Volume 3, Chapter 5, Section 5.2.
16             The modeling results in terms of changes in lipid concentrations and body burdens as a
17      function of age are shown in Figure 4-4. Some key observations include:
18
19      •       For the 6 and 12 month nursing scenarios, lipid concentrations peaked at around 4 months
20             at about 46 ppt TEQDFP-WHO98.  The formula-fed infants peaked at less than 10 ppt after
21             the first year.
22      •       In all four scenarios, the lipid concentrations merged at about 10 years of age, at a
23             concentration of about 13 ppt TEQDFP-WHO9g. Lipid and body burdens declined slightly
24             from age 10 to about age 20, and then rose gradually through adulthood.  This rise was
25             due to the increase in half-life with age. At age 70, the modeled lipid and body burden
26             concentrations were  13 ppt TEQDFP-WHO98 lipid and 5 ppt TEQDFP-WHO98.whole body
27             weight.
28
29             A sensitivity analysis was performed to test the assumptions about changes in breast milk
30      concentrations during lactation and changes in half-life over time. In this analysis, breast milk
31      concentrations were held steady at 25 pg TEQDFP-WHO98/g lipid for a 6-month nursing scenario,
32      and the half-life of dioxins in the body remained steady at 7.1 years from birth until 70 years of
33      age.  With these two changes, the maximum infant lipid concentration increased from 46 to 70 pg
34      TEQDFP-WHO9g/g lipid. The major impact of a steady half-life assumption, instead of one which
35      increased with increasing body lipid fractions in the aging adult, was that the lipid concentrations
        9/22/00
75
DRAFT—DO NOT CITE OR QUOTE

-------
 1      stabilized at about 8 pg TEQDFP-WHO9g/g lipid in the adult, instead of rising to 13 pg TEQDFP-
 2      WHO98/g lipid at age 70.
 3            The above analysis indicates that the average annual infant intake resulting from one year
 4      of nursing, 92 pg TEQDFP-WHO98/kg/day, significantly exceeds the currently estimated adult
 5      intake of 1 pg TEQDFP-WHO9g/kg/day. The impact of nursing on infant body burdens, however,
 6      is much less, i.e. infant body burdens will not exceed adult body burdens by 92 times. Rather,
 7      the modeling suggests that peak infant body burdens are only about 2 times current adult body
 8      burdens (46 vs 25 pg TEQDFP-WHO9g/g lipid). The reduced body burden impacts in nursing
 9      infants (relative to the intake) is thought to be due to the rapidly expanding infant body weight
10      and lipid volume and the possibly faster elimination rate in infants. Impacts to nursing infants
11      should decline in the future if, as discussed earlier, general population exposures decline.
12            Consumption offish, meat, or dairy products containing elevated levels ofdioxins and
13      dioxin-like PCBs can lead to elevated exposures in comparison with the general population.
14      Most people eat some fish from multiple sources, both fresh and salt water. The estimated
15      dioxin concentrations in these fish and the typical rates of consumption are included in the mean
16      background calculation of exposure. People who consume large quantities offish at estimated
17      contamination levels may have elevated exposures. These kinds of exposures are addressed
18      within the estimates of variability of background and are not considered to result in highly
19      exposed populations.  If individuals obtain their fish from areas where the- concentration of
20      dioxin-like chemicals in the fish is elevated, they may constitute a highly exposed subpopulation.
21      Although this scenario seems reasonable, very little supporting data could be found for such a
22      highly exposed subpopulation in the United  States. One study measuring dioxin-like compounds
23      in the blood of sport fishers in the Great Lakes area showed elevations over mean background,
24      but within the range of normal variability. Another study  measuring 90 PCB congeners (seven of
25      which were dioxin-like PCBs, although PCB 126 was not measured) in the blood of sport fishers
26      consuming high amounts of fish caught from Lake Michigan (>26 pounds of sport- fish/yr) did,
27      however, show significant elevations of PCBs in their blood as compared to a control population
28      (individuals consuming < 6 pounds of sport  fish/yr). The  average total concentration of PCBs in
29      the blood of these sport fishers was over three times higher than that of the control population.
30      Similarly, elevated levels of coplanar PCBs  have been measured in the blood of fishers on the
31      north shore of the Gulf of the St. Lawrence River who consume large amounts of seafood.
32      Elevated CDD/CDF levels in human blood have been measured in Baltic fishermen. For further
33      details on these studies see Part I, Volume 3, Chapter 5.                           •
34             High exposures to dioxin-like compounds as a result of consuming meat and dairy
35      products would most likely occur in situations where individuals consume large quantities of
36      these foods and the level of these compounds is elevated.  Most people eat meat and dairy
        9/22/00
76
DRAFT—DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
products from multiple sources and, even if large quantities are consumed, they are not likely to
have unusually high exposures. Individuals who raise their own livestock for basic subsistence
have the potential for higher exposures if local levels of dioxin-like compounds are high. One
study in the United States showed elevated levels in chicken eggs near a contaminated soil site.
European studies at several sites have shown elevated CDD/CDF levels in milk and other animal
products near combustion sources, and some of these have also documented elevations in the
levels of dioxin-like compounds in blood from the families consuming their home products.
                            5. DOSE-RESPONSE CHARACTERIZATION
 1            Previous sections of this integrated summary have focused on characterizing the hazards
 2      of and exposure to dioxin-like compounds. In order to bring these issues together and provide an
 3      adequate characterization of risk, the relationships of exposure to dose and, ultimately, to
 4      response must be evaluated. Key questions to be asked include: (1) What can be said about the
 5      shape of the dose-response function in the observable range and what does this imply about
 6      dose-response in the range of environmental exposures? (2) What is a reasonable limit (critical
 7      dose or point of departure) at the lower end of the observable range and what risk is associated
 8      with this exposure? In addition, one can address the issue of extrapolation beyond the range of
 9      the data in light of the answers to the above questions.  Although extrapolation of risks beyond
10      the range of observation in animals and/or humans is an inherently uncertain enterprise, it is
11      recognized as an essential component of the risk assessment process (NAS/NRC, 1983). The
12      level of uncertainty is dependent on the nature (amount and scope) of the available data and on
13      the validity of the models that have been used to characterize dose-response. These form the
14      bases for scientific inference regarding individual or population risk beyond the range of current
15      observation (NAS/NRC, 1983,1994)
16            In Part II, Chapter 8, the body of literature concerning dose-response relationships of
17      TCDD is presented. This chapter addresses the important concept of selecting an appropriate
18      metric for cross-species scaling of dose and presents the results of empirical modeling for many
19      of the available data sets on TCDD exposures in humans and in animals. Although not all
20      human observations or animal experiments are amenable to dose-response modeling, more than
21      200 data sets were evaluated for shape, leading to an effective dose (ED) value expressed as a
22      percent response being presented for the endpoint being evaluated (e.g., ED01 equals an effective
23      dose for a  1% response). The analysis of dose-response relationships for TCDD, considered
24      within the context of toxic equivalency, mechanism of action, and background human exposures,
25      helps to elucidate the common ground and the boundaries of the science and science policy
        9/22/00
                                           77
DRAFT—DO NOT CITE OR QUOTE

-------
  1      components inherent in this risk characterization for the broader family of dioxin-like
  2     compounds. For instance, the dose-response relationships provide a basis to infer a point of
  3     departure for extrapolation for cancer and noncancer risk for a complex mixture of dioxin-like
  4     congeners given the assumption of toxic equivalency as discussed in Part II, Chapter 9, Section
  5     9.6.  Similarly, these relationships provide insight into the shape of the dose-response at the point
  6     of departure, which can help inform choices for extrapolation models for both TCDD and total
  7     TEQ.
  8            In evaluating the dose-response relationships for TCDD as a basis for assessing this
  9     family of compounds, both empirical dose-response modeling approaches and mode-of-action-
10     based approaches have been developed and applied (see Part II, Chapter 8, Section 8.3 and 8.4;
11      Portier et al., 1996). Empirical models have advantages and disadvantages relative to more
12     ambitious mechanism-based models.  Empirical models provide a simple mathematical model
13     that adequately describes the pattern of response for a particular data set; they can also provide
14     the means for hypothesis testing and interpolation between data points. In addition, they can
15      provide qualitative insights into underlying mechanisms.  However, the major disadvantage of
16     empirical models is their inability to quantitatively link data sets in a mechanistically meaningful
17      manner. On the other hand, mechanism-based modeling can be a  powerful tool for
18      understanding and combining information on complex biological  systems. Use of a truly
19      mechanism-based approach can, in theory, enable more reliable and scientifically sound
20      extrapolations to lower doses and between species. However, any scientific uncertainty about the
21      mechanisms that the models describe is inevitably reflected in uncertainty about the predictions
22      of the models.
23             Physiologically based pharmacokinetic (PBPK) models have been validated in the
24      observable response range for numerous compounds in both animals and humans. The
25      development of PBPK models for disposition of TCDD in animals has proceeded through
26      multiple levels of refinement, with newer models showing increasing levels of complexity by
27      incorporating data for disposition of TCDD, its molecular actions  with the AhR and other
28      proteins, as well as numerous physiological parameters (Part II, Chapter 1). These have provided
29      insights into key determinants of TCDD disposition in treated animals. The most complete
30      PBPK models give similar predictions about TCDD tissue dose metrics. The PBPK models have
31      been extended to generate predictions for early biochemical consequences of tissue dosimetry of
32      TCDD, such as induction of CYP1A1. Nevertheless, extension of these models to more complex
33      responses is more uncertain at this time. Differences in interpretation of the mechanism of action
34      lead to varying estimates of dose-dependent behavior for similar responses.  The shape of the
35      dose-response curves governing extrapolation to low doses are determined by these hypotheses
36      and assumptions.
        9/22/00
78
DRAFT—DO NOT CITE OR QUOTE

-------
  1             At this time, the knowledge of the mechanism of action of dioxin, receptor theory, and
  2      the available dose-response data do not firmly establish a scientific basis for replacing a linear
  3      procedure for estimating cancer potency.  Consideration of this same information indicates that
  4      the use of different procedures to estimate the risk of exposure for cancer and noncancer
  5      endpoints may not be appropriate.  Both the cancer and noncancer effects of dioxin appear to
  6      result from qualitatively similar modes of action. Initial steps in the process of toxicity are the
  7      same and many early events appear to be shared. Thus, the inherent potential for low dose
  8      significance of either type of effect (cancer or noncancer) should be considered equal and
  9      evaluated accordingly. In the observable range around 1% excess response, the quantitative
10      differences are relatively small. Below this response, the different mechanisms can diverge
11      rapidly. The use of predicted biochemical responses as dose metrics for toxic responses is
12      considered a potentially useful application of these models. However, greater understanding of
13      the linkages between these biochemical effects and toxic responses is needed to reduce the
14      potentially large uncertainty associated with these predictions.
15
16      5.1. DOSE METRIC(s)
17             One of the most difficult issues in risk assessment is the determination of the dose metric
18      to use for animal-to-human extrapolations. To provide significant insight into differences in
19      sensitivity among  species, an appropriate animal-to-human extrapolation-of tissue dose is
20      required. As described in Section 1.3, the most appropriate dose metric should reflect both the
21      magnitude and frequency of exposure, and should be clearly related to the toxic endpoint of
22      concern by a well-defined mechanism.  This is, however, often difficult because human
23      exposures with observable responses may be very different from highly controlled exposures in
24      animal  experiments. In addition, comparable exposures may be followed by very different
25      pharmacokinetics  (absorption, distribution, metabolism and/or elimination) in animals and
26      humans.  Finally, the sequelae of exposure in the form of a variety of responses related to age,
27      organ, and species sensitivity complicate the choice of a common dose metric. Despite these
28      complexities, relatively simple default approaches, including body surface or body weight scaling
29      of daily exposures, have often been recommended (U.S. EPA, 1992a, 1996).
30             As discussed in Section 1.3, dose can be expressed in a number of ways.  For TCDD  and
31      other dioxin-like compounds, attention has focused on the consideration of dose expressed as
32      daily intake (ng/kg/day), body burden (ng/kg), or AUC (DeVito et al, 1995; Aylward et al.,
33      1996).  The concept of physiological time (lifetime of an animal) complicates the extrapolation,
34      as the appropriate  scaling factor is uncertain for toxic endpoints.  Because body burden
35      incorporates differences between species in TCDD half-life (these differences are large between
36      rodent species and humans [See Part II, Chapter 8, Table 8.2]), this dose metric appears to be the
        9/22/00
79
DRAFT—DO NOT CITE OR QUOTE

-------
 1      most practical for this class of compounds (DeVito et al, 1995).  Average lifetime body burden
 2      is best suited for steady-state conditions, with difficulties arising when this dose metric is applied
 3      to evaluation of acute exposures, such as those occurring in the 1976 accidental exposure of
 4      some people living hi Seveso, Italy (Bertazzi and di Domenico, 1994). In cases such as this,
 5      increased body burden associated with the acute exposure event is expected to decline (half-life
 6      for TCDD is approximately 7 years) until it begins to approach a steady-state level associated
 7      with the much smaller daily background intake. However, this issue of acute exposure is not a
 8      major factor in the current analyses. In general, daily excursions in human exposure are
 9      relatively small and have minor impact on average body burden.  Instead, PBPK models suggest
10      that human body burdens increase over time and begin to approach steady-state after
11      approximately 25 years with typical background doses.  Occupational exposures represent the
12      middle ground where daily excursions during the working years can significantly exceed daily
13      background intakes for a number of years, resulting in elevated body burdens.
14            The relationship between occupational exposures and body burden, and between body
15      burden and AUC, are demonstrated in Figure 5-2. This figure graphs two hypothetical body
16      burden scenarios during the 70 year lifespan of an individual. The first is a continuation to 70
17      years of age of the background body burden scenario discussed, with caveats and assumptions, in
18      Part I, Volume 3, Chapter 5. In this scenario, an infant is breast fed for six months by a mother
19      with a background dioxin body burden level, and subsequently exposed to the average current
20      level of dioxin in the food supply (1 pg/kg/day). This background scenario leads to a 70 year
21      lifetime area under the curve (AUC) of 255 ng/kg* Y, equivalent to a lifetime average body
22      burden of 3.6 ng/kg (-255/70 years). In the second scenario, the same individual incurs an
23      additional occupational exposure between 20 and 30 years of age of 100 pg/kg/day - one hundred
24      times background - then ceasing. The buildup of dioxin body burden is evident in the peak level
25      and shark fin appearance.  AUC in this occupational scenario is 3911 ng/kg* Y, and LABB is
26      55.9 ng/kg.  Note that in the occupational scenario the peak body burden is -40 times
27      background, but the AUC and LABB are only 15 times background.
28            Table 5-1 and Figure 5-1 summarize literature on average levels of dioxin TEQs in the
29      background human population and peak levels in commonly cited epidemiological cohorts.
30      Table 5-1 collates data on tissue lipid levels (ppt lipid adjusted) in populations, principally from
31      serum, tabulating either current levels for the background population or back calculated peak
32      levels for the exposed cohorts. Figure 5-1 graphs the estimated range and central tendency of the
33      total TEQDFP body burden (ng/kg whole body), combining the range of measured 2,3,7,8-TCDD
34      values with the estimate of the background non-2,3,7,8-TCDD TEQ level from the U.S.
35      population in the late 1980s/early  1990s. TEQ levels are calculated for PCDD, PCDF, and
36      PCBs, based on TEQDFP-WHO98 values, and assume a constant 25% body fat ratio when
        9/22/00
80
DRAFT—DO NOT CITE OR QUOTE

-------
 1      converting from serum lipid ppt to ng/kg body burden.  Total TEQ values for the Hamburg
 2      cohort women were calculated by the authors, and for this cohort the TCDD graph includes non-
 3      TCDD TEQ.  Seveso values reported by Needham et al. (1999).are based on stored serum
 4      samples from subjects undergoing medical examinations contemporaneous with the exposure,
 5      and were not back-calculated.
 6            As discussed earlier, using background total body burden (TEQDFP-WHO98) as a point of
 7      comparison, these often- termed "highly exposed" populations have peak body burdens that are
 8      relatively close to general population backgrounds at the time. When compared to background
 9      body burdens of the late 1980s, many of the median values and some of the mean values fall
10      within a range of one order of magnitude (factor of 10) and all fall within a range of two orders
11      of magnitude (factor of 100).  General population backgrounds at the time are likely to have been
12      higher. As these are peak body burdens, measured at the time of the Seveso accident or back-
13      calculated to the time of last known elevated exposure, being compared to background averages,
14      average lifetime body burdens in these cohorts will be even closer to lifetime average
15      background levels. This will be important if, as demonstrated for some chronic effects in
16      animals and as assumed when relying on average body burden as a dose metric, cancer and other
17      noncancer effects are a consequence of average tissue levels over a lifetime.  Body burdens begin
18      to decline slowly soon after elevated exposure ceases. Some data in humans and animals suggest
19      that elimination half-lives for dioxin and related compounds may be dose-dependent, with high
20      doses being eliminated more rapidly than lower doses. Nonetheless, the use of an approximately
21      7-year half-life of elimination presents a reasonable approach for evaluating both back-calculated
22      and average lifetime levels, because for most cohorts the exposure is primarily to TCDD.
23            The ability to detect effects in epidemiologic studies is dependent on a sufficient
24      difference between control and exposed populations.  The relatively small difference (<10-100
25      fold) between exposed and controls in the dioxin epidemiology studies makes exposure
26      characterization in the studies a particularly serious issue. This point also strengthens the
27      importance of measured blood or tissue levels in the epidemiologic analyses, despite the
28      uncertainties associated with calculations extending the distribution of measured values to the
29      entire cohort and assumptions involved in back-calculations.
30             Characterization of the risk of exposure of humans today remains focused on the levels
31      of exposure that occur in the general population, with particular attention given to special
32      populations (see Part I, Volume 3, Chapters 4 and 5).  For evaluation of multiple endpoints and
33      considering the large differences in half-lives for TCDD across multiple species, it is generally
34      best to use body burden rather than daily intake as the dose metric for comparison unless data to
35      the contrary are presented.  Further discussion of this point, which provides the rationale for this
        9/22/00
81
DRAFT—DO NOT CITE OR QUOTE

-------
 1      science-based policy choice, is presented in Part II, Chapters 1 and 8, and is summarized in
 2      Section 1.3 of this document.
 3
 4      5.1.1. Calculations of Effective Dose (ED)
 5            Comparisons across multiple endpoints, multiple species, and multiple experimental
 6      protocols are too complicated to be made on the basis of the full dose-response curve. As
 7      discussed above, comparisons of this sort can be made by either choosing a given exposure and
 8      comparing the responses, or choosing a particular response level and comparing the associated
 9      exposures. In the analyses contained in Chapter 8, Section 8.3 and elsewhere in the
10      reassessment, comparison of responses is made using estimated exposures associated with a
11      given level of excess response or risk. To avoid large extrapolations, this common level of
12      excess risk was chosen such that for most studies the estimated exposure is in or near the range
13      of the exposures seen hi the studies being compared, with extra weight given to the human data.
14      A common metric for comparison is  the effective dose or ED, which is the exposure dose
15      resulting in an excess response over background in the studied population. EPA has suggested
16      this approach hi calculating benchmark doses (BMD) (Allen et al., 1994) and in its proposed
17      approaches to quantifying cancer risk (U.S. EPA, 1996;  U.S. EPA, 1999).  Although effective
18      dose evaluation at the 10% response  level (ED10 or lower bound on ED10 [LED10])  is somewhat
19      the norm, given the power of most chronic toxicology studies to detect an-effect, this level is
20      actually higher than those typically observed in the exposed groups in studies of TCDD impacts
21      on humans.  To illustrate, lung cancer mortality has a background lifetime risk of approximately
22      4% (smokers and nonsmokers combined), so that even a relative risk of 2.0 (2 times the
23      background lifetime risk) represents  approximately a 4% increased lifetime risk. Based upon this
24      observation and recognizing that many of the TCDD-induced endpoints studied in the laboratory
25      include 1% effect levels in the experimental range, Chapter 8 presents effective doses of 1% or
26      ED0i. The use of ED values below 10% is consistent with the Agency's guidance on the use of
27      mode of action in assessing risk, as described the proposed Cancer Risk Assessment Guidelines
28      (U.S. EPA, 1996; U.S. EPA, 1999) and in the evaluation framework discussed in Section 3.3, in
29      that the observed range for many "key events"  for TCDD extends down to or near the 1%
30      response level.  Determining the dose at which key events for dioxin toxicity begin to be seen in
31      a heterogeneous human population provides important information for decisions regarding risk
32      and safety.
33
        9/22/00
82
DRAFT—DO NOT CITE OR QUOTE

-------
 1      5.2. EMPIRICAL MODELING OF INDIVIDUAL DATA SETS
 2             As described in Chapter 8, Section 8.3, empirical models have advantages and
 3      disadvantages relative to more ambitious mechanism-based models. Empirical models provide a
 4      simple mathematical model that adequately describes the pattern of response for a particular data
 5      set and can also provide the means for hypothesis testing and interpolation between data points.
 6      In addition, they can provide qualitative insights into underlying mechanisms. However, the
 7      major disadvantage is their inability to quantitatively link data sets in a mechanistically
 8      meaningful manner. Data available  for a number of biochemical and toxicological effects of
 9      TCDD, and on the mechanism of action of this chemical, indicate that there is good qualitative
10      concordance between responses in laboratory animals and humans (see Table 2-1). In addition,
11      as described below, human data on exposure and cancer response appear to be qualitatively
12      consistent with animal-based risk estimates derived from carcinogenicity bioassays.  These and
13      other data presented throughout this reassessment would suggest that animal models are
14      generally an appropriate basis for estimating human responses to dioxin-like compounds.
15      Nevertheless, there are clearly differences in exposures and responses between animals and
16      humans, and recognition of these is essential when using animal data to estimate human risk.
17      The level of confidence in any prediction of human risk depends on the degree to which the
18      prediction is based on an accurate description of these interspecies extrapolation factors.  See
19      Chapter 8, Section 8.3, for a further  discussion of this point.           ••--
20             Almost all dioxin research data are consistent with the hypothesis that the binding of
21      TCDD to the AhR is the first step in a series of biochemical, cellular, arid tissue changes that
22      ultimately lead to toxic responses observed in both experimental animals and humans (see Part II,
23      Chapter 2, Section 2.3).  As such, an analysis of dose-response data and models should use,
24      whenever possible, information on the quantitative relationships among ligand (i.e., TCDD)
25      concentration, receptor occupancy, and biological response. However, it is clear that multiple
26      dose-response relationships are possible when considering ligand-receptor mediated events. For
27      example, dose-response relationships for relatively simple responses, such as enzyme induction,
28      may not accurately predict dose-response relationships for complex responses such as
29      developmental effects and cancer. Cell- or tissue-specific factors may determine the quantitative
30      relationship between receptor occupancy and the ultimate response. Indeed, for TCDD there are
31      much experimental data from studies using animal and human tissues to indicate that this is the
32      case.  This serves as a note of caution, as empirical data on TCDD are interpreted in the broader
33      context of complex exposures to mixtures of dioxin-like compounds as well as to non-dioxin-like
34      toxicants.
35             As for other chemical mechanisms where high biological potency is directed through the
36      specific and high-affinity interaction between chemical and critical cellular target, the
        9/22/00
83
DRAFT—DO NOT CITE OR QUOTE

-------
  1      supposition of a response threshold for receptor-mediated effects is a subject for scientific
  2      debate. The basis of this controversy has been recently summarized (Sewall and Lucier, 1995).
  3             Based on classic receptor theory, the occupancy assumption states that the magnitude of
  4      biological response is proportional to the occupancy of receptors by drug molecules.  The
  5      "typical" dose-response curve for such a receptor-mediated response is sigmoidal when plotted
  6      on a semilog graph or hyperbolic if plotted on a arithmetic plot. Implicit in this relationship is
  7      low-dose linearity (0-10% fractional response) through the origin. Although the law of mass
  8      action predicts that a single molecule of ligand can interact with a receptor, thereby inducing a
  9      response, it is also widely held that there must be some dose that is so low that receptor
10      occupancy is trivial and therefore no perceptible response is obtainable.
11             Therefore, the same receptor occupancy assumption of the classic receptor theory is
12      interpreted by different parties as support for and against the existence of a threshold. It has been
13      stated that the occupancy assumption cannot be accepted or rejected on experimental or
14      theoretical grounds (Goldstein et al., 1974). To determine the relevance of receptor interaction
15      for TCDD-mediated responses, one must consider (1) alternatives as well as limitations of the
16      occupancy theory; (2) molecular factors contributing to measured endpoints; (3) limitations of
17      experimental methods; (4) contribution of measured effect to a relevant biological/toxic
18      endpoint; and (5) background exposure.
19             Throughout this reassessment, each of these considerations has been explored within the
20      current context of the understanding of the mechanism of action of TCDD, of the methods for
21      analysis of dose-response for cancer and noncancer endpoints, and of the available data sets of
22      TCDD dose and effect for several rodent species, as well as humans who were occupationally
23      exposed to TCDD at levels exceeding the exposure of the general population.
24
25      5.2.1. Cancer
26             As described in Section 2.2.1.4, TCDD has been characterized as a human carcinogen,
27      and is a carcinogen in all species and strains of laboratory animals tested. The epidemiological
28      database for TCDD, described in detail in Part II, Chapter 7a, suggests that exposure may be
29      associated with increases in all cancers combined, in respiratory tumors and, perhaps, in soft-
30      tissue sarcoma.  Although there are sufficient data in animal cancer studies to model dose-
31      response for a number of tumor sites, as with many chemicals it is generally difficult to find
32      human data with sufficient information to model dose-response relationships. For TCDD, there
33      exist three studies of human occupational exposure with enough information to perform a
34      quantitative dose-response analysis.
35             Table 5-2 summarizes the epidemiology and bioassay studies used in the calculations of
36      the all cancer mortality ED01s/LEDOIs.  Results for three different occupational cohorts are
        9/22/00
84
DRAFT—DO NOT CITE OR QUOTE

-------
 1      tabulated: Hamburg, NIOSH, and BASF, along with the bioassay results on liver cancer in
 2      female Sprague-Dawley rats (Kociba et al., 1978). In addition to the three dose-response results
 3      analyzed in Part II, Chapter 8 (Flesch-Janys et al., 1998; Aylward et al., 1996; Ott and Zober
 4      1996a, b), two additional primary publications on these occupational cohorts are tabulated and
 5      graphed. Although these additional studies demonstrate dose-response relationships when using
 6      improved exposure metrics, neither can be used for the calculations in this assessment because of
 7      lack of an upper confidence interval on the risk provided in the original publication (Becher et
 8      al., 1998) or absence of a quantitative exposure metric (Steenland et al., 1999).
 9            Modeling cancer in humans uses slightly different approaches from those used in
10      modeling animal studies. The modeling approach used in the analysis of the human
11      epidemiology data for all cancers combined and lung cancer involves applying estimated human
12      body burden to cancer response and estimating parameters in a linear risk model for each data
13      set.  A linear risk model was used because the numbers of exposure groups available for analysis
14      were too small to support more complicated models. Because of this, no evaluation of the shape
15      of the dose-response data for the human studies was performed. Access to the raw data may
16      make it possible to use more complicated mathematical forms that allow for the evaluation of
17      shape. In the one case in which this has been done, the dose-response shape suggested a
18      response that was supralinear (dose raised to a power <1) (Becher et al., 1998).  For these studies,
19      there are several assumptions and uncertainties involved in modeling the data, including
20      extrapolation of dosage, both in back-calculation and in elimination kinetics, and the type of
21      extrapolation model employed.
22            As described in Part II, Chapter 8, Section 8.3, the data used in the analyses are from
23      Flesch-Janys et al.  (1998) for the Hamburg cohort, Aylward et al. (1996) for the NIOSH study,
24      and Ott and Zober (1996a,b) for the BASF cohort. The limited information available from these
25      studies is in the form of standardized mortality ratios (SMRs) and/or risk ratios by exposure
26      subgroups with some estimate of cumulative subgroup exposures.  Exposure subgroups were
27      defined either by number of years of exposure to  dioxin-yielding processes or by extrapolated
28      TCDD levels. No study sampled TCDD blood serum levels for more than a fraction of its
29      cohort, and these samples were generally taken decades after last known exposure. In each study,
30      serum fat or body fat levels of TCDD were back calculated using a first-order model.  The
31      assumed half-life of TCDD used in the model varied from study to study.  Aylward et al. (1996)
32      used the average TCDD levels of those sampled in an exposure subgroup to represent the entire
33      subgroup.  Flesch-Janys et al. (1998) and Ott and Zober (1996a) performed additional
34      calculations, using regression procedures with data on time spent at various occupational tasks,
35      to estimate TCDD levels for all members of their respective cohorts.  They then divided the
36      cohorts into exposure groups based on the estimated TCDD levels. The information presented  in
        9/22/00
85
DRAFT—DO NOT CITE OR QUOTE

-------
 1      the literature cited above was used to calculate estimated average TCDD dose levels in Chapter
 2      8, Section 8.3.
 3            To provide ED01 estimates for comparison in Chapter 8, Section 8.3, Poisson regression
 4      (Breslow and Day, 1987) was used to fit a linear model to the data described above. A linear
 5      model was chosen for several reasons.  Analysis of animal cancer data suggests a mixture of
 6      linear and nonlinear responses, with linear shape parameters predominating (Portier et al., 1984).
 7      Toxic responses to TCDD, both cancer and noncancer, are presumably more likely to result from
 8      multiple cellular and tissue-level perturbations and are less likely to follow linear relationships.
 9      This hypothesis was examined by empirical dose-response modeling of cancer and noncancer
10      effects of TCDD in experimental animals (Part II, Chapter 8, Section 8.3). This empirical
11      modeling exercise demonstrated that in general, the linear models provided the best fit to the
12      biochemical response data and that more complex responses were generally fit best with non-
13      linear models.  Many examples of adverse effects experienced at these low levels have too much
14      data variability to clearly distinguish on a statistical basis between dose-response curve options,
15      and whether dose-response follows linear, supra/sub-linear, power curve, or threshold kinetics.
16            Besides the issue of use of a linear model, additional important uncertainties in the human
17      epidemiological data discussed in Part II, Chapter 8, Section 8.3, include the representativeness
18      and precision of the dose estimates that were used, the choice of half-life and whether it is dose
19      dependent, and potential interactions between TCDD and smoking or other toxicants.
20      Nevertheless, with these qualifications, it is possible to apply simple empirical models to studies
21      in which exposure data for TCDD are available in human populations.
22            The analysis of these three epidemiological studies of occupationally exposed individuals
23      suggest an effect of TCDD on all cancers,  and on lung cancers in the adult human male.  The
24      ED0,s based upon average excess body burden of TCDD ranged from 5.7 ng TCDD/kg to 250 ng
25      TCDD/kg in humans. The lower bounds on these doses (based on a modeled 95% C.I.) range
26      from 3.5 ng TCDD/kg to 120 ng TCDD/kg. For the effect of TCDD on all cancers combined, the
27      human ED01s ranged from 5.7 ng/kg to 80.2 ng/kg.  The lower bounds on these doses (based on a
28      modeled 95% C.I.) range from 3.5 ng TCDD/kg to 37.5  ng TCDD/kg.  For the effect of TCDD
29      on lung cancers, the only tumor site increased in both rodents and humans, the human EDOIs
30      ranged from 36.6 ng/kg to 250 ng/kg. The lower bounds on these doses (based on a modeled
31      95% C.I.) range from 16.2 ng TCDD/kg to 120 ng TCDD/kg.  These estimates of EDOIs  are
32      compared to  animal estimates later in this discussion.
33            Both  empirical and mechanistic models were used to examine cancer dose-response in
34      animals. Portier et al. (1984) used a simple multistage model of carcinogenesis with up to two
35      mutation stages affected by exposure to model the five tumor types observed to be increased in
36      the 2-year feed study of Kociba et al. (1978, Sprague-Dawley rats) and the eight tumor types
        9/22/00
86
DRAFT—DO NOT CITE OR QUOTE

-------
 1      observed to be increased in the 2-year gavage cancer study conducted by the NTP
 2      (Osborne-Mendel rats and B6C3Fi mice, 1982a).  The findings from this analysis, which
 3      examined cancer dose-response within the range of observation, are presented in Part II, Chapter
 4      8, Table 8.3.2., which is reproduced with slight modifications as Table 5-3. All but one of the
 5      estimated ED0jS are above the lowest dose used in the experiment (approximately 1 ng
 6      TCDD/kg/day in both studies) and are thus interpolations rather than extrapolations. The
 7      exception, liver cancer in female rats from the Kociba study, is very near the lowest dose used in
 8      this study and is only a small extrapolation (from  1 ng TCDD/kg/day to 0.77 ng TCDD/kg/day).
 9      Steady-state body burden calculations were also used to derive doses for comparison across
10      species. Absorption was assumed to be 50% for the Kociba et al. (1978) study (feed experiment)
11      and 100% for the NTP study (gavage experiment). Also presented in Table 5-3 are the shapes of
12      the dose-response curves as determined by  Portier et al. (1984).
13             The predominant shape of the dose-response curve in the experimental region for these
14      animal cancer results is linear; this does not imply that a nonlinear model such as the quadratic or
15      cubic, or for that matter, a "J-shaped" model, would not fit these data. In fact, it is unlikely that
16      in any one case, a linear model or a quadratic model could be rejected statistically for these cases.
17      These studies had only three experimental dose groups, hence these shape calculations are not
18      based upon sufficient doses to guarantee a consistent estimate; they should be viewed with
19      caution. The ED01 steady-state body burdens range from a low value of->4 ng/kg based upon the
20      linear model associated with liver tumors in female rats to as high as 1,190 ng/kg based upon a
21      cubic model associated with thyroid follicular cell adenomas in"female rats. Lower bounds on
22      the steady-state body burdens in the animals range from  10 ng TCDD/kg to 224 ng/kg.  The
23      corresponding estimates of daily intake level at the ED0i  obtained from an empirical linear model
24      range from 0.8 to 43 ng TCDD/kg body weight/day depending on the tumor site, species, and sex
25      of the animals investigated. Lower confidence bounds on the estimates of daily intake level at
26      the ED01 in the animals range from 0.6 to 14 ng TCDD/kg body weight/day. In addition, using a
27      mechanistic approach to modeling, Portier and Kohn (1996) combined the biochemical response
28      model of Kohn et al. (1993) with a single initiated phenotype two-stage model of carcinogenesis
29      to estimate liver tumor incidence in female  Sprague-Dawley rats from the 2-year cancer bioassay
30      of Kociba et al. (1978). By way of comparison, the ED01 estimate obtained from this linear
31      mechanistic model was 0.15 ng TCDD/kg body weight/day based on intake, which is equivalent
32      to 2.7 ng TCDD/kg steady-state body burden. No lower bound on this modeled estimate of
33      steady-state body burden was provided.
34             As discussed in Part II, Chapter 8, Section 8.2,  different dose metrics can lead to widely
35      diverse conclusions. For example, as described in Chapter 8, Section 8.2, the ED01 intake for the
36      animal tumor sites presented above ranges from less than 1 to tens of ng/kg/day, and the lowest
        9/22/00
87
DRAFT—DO NOT CITE OR QUOTE

-------
 1      dose with an increased tumorigenic response (thyroid tumors) in a rat is 1.4 ng TCDD/kg/day
 2      (NTP, 1982a). The daily intake of dioxins in humans is estimated at approximately 1 pg
 3      TEQ/kg/day. This implies that humans are exposed to doses 1,400 times lower than the lowest
 4      tumorigenic daily dose in rat thyroid. However, 1.4 ng TCDD/kg/d in the rat leads to a steady-
 5      state body burden of approximately 25 ngTCDD/kg, assuming a half-life of TCDD of 25 days
 6      and absorption from feed of 50%2. If the body burden of dioxins in humans is approximately 20
 7      ng TEQ/kg lipid or 5 ngTEQ/kg body weight (assuming about 25% of body weight is lipid),
 8      humans are exposed to about 5 times less TCDD than the minimal carcinogenic dose for the rat.
 9      The difference between these two estimates is entirely due to the approximately 100-fold
10      difference in the half-life of TCDD between humans and rats. At least for this comparison, if
11      cancer is a function of average levels in the body, the most appropriate metric for comparison is
12      the average or steady-state body burden, since this accounts for the large differences  in animal to
13      human half-lives.
14            Comparisons of human and animal ED01s from Part II, Chapter 8, Section 8.3, for cancer
15      response on a body burden basis show approximately equal potential for the carcinogenic effects
16      of TCDD. In humans, restricting the analysis to log-linear models in Part II, Chapter 8, Section
17      8.3, resulted in cancer ED01s ranging from approximately 6 ng/kg to 250 ng/kg. This was similar
18      to the empirical modeling estimates from the animal studies, which ranged from 14 ng/kg to
19      1,190 ng/kg (most estimates were in the range from 14 to 500 ng/kg). The lower bounds on the
20      human body burdens at the ED0]s (based on a modeled 95% C.I.) range from 3.5 ng TCDD/kg to
21      120 ng TCDD/kg.  Lower bounds on the steady-state body burdens in the animals range from 10
22      ng TCDD/kg to 224 ng/kg. The estimate for the single mechanism-based model presented earlier
23      (2.7 ng/kg) was approximately 2 times lower than the lower end of the range of human ED0]
24      estimates and less than the lower bound on the LED01. The same value was approximately 5
25      times lower than the lower end of the range of animal ED01 estimates and less than 4 times less
26      than the LED01.
27            Using human and animal cancer ED01s, their lower bound estimates, and the value of 2.7
28      ng TCDD/kg from the single mechanism-based model, slope factors and comparable risk
        2 Steady-state body burden (ng/kg) = (daily dose (ng/kg/day) * (half-life)/Ln(2)) ( f), where f is the fraction absorbed
        from the exposure route (unitless) and half-life is the half-life in days.
        9/22/00
88
DRAFT—DO NOT CITE OR QUOTE

-------
estimates for a human background body burden of approximately 5 ng TEQ/kg (20 ng TEQ/kg
lipid) can be calculated using the following equations:

       Slope factor (per pg TEQ/kgBW/day) = risk at ED01 / intake (pg TEQ/kgBW/day)
associated with human equivalent steady-state body burden at EDOI where:
       Risk at ED01 = 0.01; and
       Intake (pgTEQ/kgBW/day) = [body burden at ED., fag TEO/kg)* Ln(2)] * 1/f
                                           half-life (days)
                                     (5-1)
and
       half-life = 2,593 days in humans and 25 days in rats (see Table 8.1 in Part II, Chapter 8)
       f = fraction of dose absorbed; assumed to be 80%
       Upper bound on excess risk at human background body burden = (human
       background body burden ( ng/kg))(risk at ED0])/lower bound on human
       equivalent steady-state body burden (ng/kg) at EDOI, where:
       Risk at EDOI = 0.01
                                     (5-2)
       Use of these approaches reflects methodologies being developed within the context of the
revised draft Cancer Risk Assessment Guidelines.  Slopes are estimated by a simple proportional
method at the "point of departure" (LED01) at the low end of the range of experimental
observation. As discussed below, these methods can be compared to previous approaches using
the linearized multistage (LMS) procedure to determine if the chosen approach has significantly
changed the estimation of slope.  The estimates of ED0!/LED01 represent the human-equivalent
body burden for 1% excess cancer risk based on exposure to TCDD and are assumed for
purposes of this analysis to be equal for TCDD equivalents (total TEQ). This assumption is
based on the toxic equivalency concept discussed throughout this report and in detail in Part II,
Chapter 9.  All cancer slope factors can be compared to the Agency's previous slope factor of 1.6
x 10"4 per pgTCDD/kgBW/day which is equivalent to 1.6 x 105 per mgTCDD/kgBW/day (U.S.
EPA,  1985).
5.2.1.1. Estimates of Slope Factors and Risk at Current Background Body Burdens Based on
Human Data
       Estimates of upper bound slope factors (per pg TCDD/kgBW/day) calculated from the
human ED01s presented in Part II, Chapter 8, Table 8.3.1, range from 8.6 x 10'3, if the LED01 for
all cancer deaths in the Hamburg cohort is used, to 2.5 x 10~4 if the ED01 for lung cancer deaths in
9/22/00
89
DRAFT—DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
the smaller BASF cohort is used. All of the other slope factors for all cancer deaths or lung
cancer deaths in the three cohorts would fall within this range. LED0]s for all cancer deaths span
approximately an order of magnitude and would generate slope factors in the range of 8.6 x 10"3
to 8 x 10"4. Slightly smaller slope factors are generated when LED0,s for lung cancer are used.
The largest slope factors based on LED01s come from the Hamburg cohort (8.6 x 10"3 and 1.9 x
10~3 respectively for all cancer deaths and lung cancer deaths.) There is no compelling reason to
choose one slope factor over the next from among those calculated, given that each study had
particular strengths and weaknesses (See Part II, Chapter la). Thus, a meta-analysis was
performed by combining all data sets into a single large data set and using Poisson regression
procedures detailed in Part II, Chapter 8, Section 8.3, yielding a slope factor estimate of
approximately 1 x 10"3perpgTCDD/kgBW/day. This represents EPA's most current upper
bound slope factor for estimating human cancer risk based on human data.
       These estimates compare well with the estimates of cancer slope and risk associated with
TCDD exposure in the Hamburg cohort published by Becher et al. (1998). The risk estimates of
Becher et al.(1998) were derived from data on TCDD exposure to male workers with a 0 or 10-
year latency and taking into account other factors affecting risk including choice of model,
latency, job category, dose metric, and concurrent exposures. These estimates range from 1.3 x
10~3 to 5.6 x 10~3 per pg TCDD/kgBW/day. In this analysis all excess cancers are attributed to
TCDD exposure, despite significant levels  of other dioxin-like compounds in blood
measurements of this cohort (see Table 5-1) with similar slope coefficients calculated for total
TEQ. Although risk estimates using TCDD alone in this cohort might suggest an overestimate of
risk because dose is underestimated, no evidence for this emerged from the analysis because
TCDD dominates the total TEQ in this population. In the preparation of this document, an
independent estimate of slope using the Becher models was performed consistent with the
approaches suggested in Part II, Chapter 8, Section 8.2 ( See Table 8.4 for more details). A slope
of 3 x 10"3 per pg TCDD/kgBW/day was derived. This slope represents a central estimate since
no upper bound could be calculated with the available data.
       Taking into account different sources of variation, Becher et al. (1998) suggest a range of
10~3 to 10"2 for additional lifetime cancer risk for a daily intake of 1 pg TCDD/kg BW/day. By
inference, that range could also apply to total TEQ intake. As described in Section 4.4.2, current
intakes hi the United States are estimated to be approximately 1 pg TEQDFP-WHO9g/kg BW/day.
Using Equation 5-2 and based on all cancer deaths in the three cohorts, the upper bound range of
risks estimated from current human body burdens of 5 ng TEQDFP-WHO9g/kgB W (which equates
to a serum level of approximately 20 pg/g lipid [see Table 4-7]) ranged from 1.4 x 10~2 to 1.3 x
10"3. Based on lung cancer deaths, the lower end of the upper bound on the estimates of excess
risk extended to 4 x 10"4. Using the LEDOI estimate of 30.1 ng/kg from the meta-analysis yields
        9/22/00
                                          90
DRAFT—DO NOT CITE OR QUOTE

-------
 1      an upper bound risk estimate of 1.7 x 10'3 for an average lifetime body burden estimate of 5
 2      ng/kg. Estimates using high end current or historical body burdens would be proportionately
 3      higher. The range of these estimates provides further support for the perspective on risk
 4      provided by Becher et al. (1998). Uncertainties associated with these estimates from human
 5      studies are discussed in Part II, Chapter 8, Section 8.3, and in Becher et al. (1998).
 6
 7      5.2.1.2. Estimates of Slope Factors and Risk at Current Background Body Burdens Based on
 8      Animal Data
 9            Upper bound slope factors (per pg TCDD/kgB W/day) for human cancer risk calculated
10      from lower bounds in ED01s (LED01s) for the animal cancers presented in Table 5-3 range from
11      3  x 10~3 to 1 x 10"4.  This spans a range from being 19 times greater than the previous upper
12      bound estimate on cancer slope (1.6 x 1CT4 [U.S. EPA, 1985]) to less than 50% of this value. The
13      largest slope factor is derived from the same study as the 1985 estimate; that is, the slope factor
14      derived from the female liver cancer in the Kociba et al. (1978) study continues to give the
15      largest slope factor.
16      Reconciling the Portier (1984) and EPA (1985) Slave Estimates
17            In attempting these comparisons, two issues became apparent. First, the body burden and
18      the intake at the ED0i from Portier et al. (1984) does not result in the same slope  factor as EPA
19      (U.S. EPA, 1985). Despite the use of the same study results, a slope factor of 1.8 x 10"5per pg
20      TCDD/kgB W/day results using the linearized multistage (LMS) approach in Portier et al. (1984).
21      This is a factor of approximately 10 lower than the EPA (U.S. EPA, 1985) estimate of the slope.
22      The differences are attributable to the aims of the respective calculations at the time.  Portier et
23      al. (1984) calculated "virtually safe doses" assuming that rodent and human doses scaled on a
24      mg/kg basis, and he used the original tumor counts from the study. EPA (U.S. EPA, 1985), on
25      the other hand, used (BW)2/3 to arrive at a human equivalent dose and used the pathology results
26      from a reread of the original Kociba study (U.S. EPA, 1980). In addition, EPA (U.S. EPA, 1985)
27      adjusted tumor counts for early mortality in the study.  The factor to adjust for (BW)2/3-scaling in
28      the rat is 5.8. The correction for early mortality can be accounted for with a factor of 1.6 (this is
29      the ratio of the intake values at the ED01 with and without the early mortality correction). If the
30      Portier et al. slope factor (1.8x10"5 per pg TCDD/kgB W/day) is multiplied by these two factors,
31      a slope of 1.7 x 10"4 per pg TCDD/kgB W/day is calculated.  This is essentially equivalent to the
32      EPA (U.S. EPA, 1985) estimate of 1.6 x  10"4 per pg TCDD/kgB W/day. Reconciling these issues
33      is important to ensure appropriate comparisons of slope factor estimates.
34      Calculating a Revised Estimate of Cancer Slope from Kociba et al. (1978)
35             More important is the calculation of slope factor estimates using current methods of
36      analysis that recognize the importance of the dose metric and the differences in half-life of
        9/22/00
91
DRAFT—DO NOT CITE OR QUOTE

-------
  1      dioxins in the bodies of laboratory animals and humans (see Part II, Chapter 8, Section 8.2, for
  2     detailed discussion). The major difference between the approaches used to calculate risks in the
  3     mid-1980s (Portier et al., 1984; U.S. EPA, 1985) and the current approach is the use of body
  4     burden as the dose metric for animal-to-human dose equivalence. The decision to use body
  5     burden accounts for the approximately 100-fold difference between half-lives of TCDD in
  6     humans and rats (2,593 days versus 25 days [see Part II, Chapter 8, Table 8.1]). Use of Equation
  7     5-1 results in an estimated body burden at the LED0] of 6.1 ng TEQ/kg, derived from the EPA
  8     (U.S. EPA, 1985) Kociba tumor counts. This compares favorably with the Portier estimate of 10
  9     ng TEQ/kg found in Table 5-3. The difference is entirely accounted for by the early deaths
10     adjustment by EPA (U.S. EPA, 1985). Use of these body burdens at the LED01 results in slope
11      factor estimates of 3.3 x 1O'3 per pg TCDD/kgB W/day and 4.9 x 10'3 per pg TCDD/kgBW/day for
12     the Portier at al. (1984) (10 ng/kg) and the newly derived body burden (6.1 ng/kg), respectively.
13     Again, the difference is due solely to the adjustment for early mortality, which EPA considers a
14     better estimate of upper bound lifetime risk than does the unadjusted estimate.  EPA's revised
15      slope factor (4.9 x  10"3 per pg TCDD/kgB W/day) would be 31 times greater than the slope factor
16      from 1985.
17            However, a second issue with the modeling of the Kociba data relates to the appropriate
18      tumor counts to use.  As mentioned in Section 2.2, Goodman and Sauer (1992) reported a second
1Q      re-evaluation of the female rat liver tumors in the Kociba study using the latest pathology criteria
20      for such lesions. Results of this review are discussed in more detail in Part II, Chapter 6, Section
21      6.2. The review confirmed only approximately one-third of the "tumors of the previous review
22      (U.S. EPA, 1980).  Although this finding did not change the determination of carcinogenic
23      hazard because TCDD induced tumors in multiple sites in this study, it does have an effect on
24      evaluation of dose-response and on estimates of risk. Because neither the original EPA (U.S.
25      EPA, 1985) slope factor estimate nor that of Portier et al.  (1984) reflect this reread, it is
26      important to factor these results into the estimate of the ED01 and slope factor. Using the LMS
27      procedure used by EPA in 1985 and the tumor counts as reported in Part II, Chapter 6, Table 6.2,
28      the revised slope factor is reduced by approximately 3.6-fold to yield a slope factor of 4.4 x 10'5
29      per pg TCDD/kgB W/day. However, because the original estimates used a (BW)2/3 scaling, this
30      must be adjusted to use body burden and obtain an appropriate result.  When dose is adjusted and
31      Equation 5-1  is used, an LEDm of 22.2 ng TEQ/kg and a slope factor of 1.4 x 10'3 per pg
32      TCDD/kgB W/day are derived. These results can also be obtained using EPA's Bench Mark
33      Dose (BMD) software and entering adjusted tumor counts and dose data to obtain a BMDL0]
34      from which an LED0] body burden of 22 ng/kg can be derived. This represents EPA's most
35      current upper bound slope factor for estimating human cancer risk based on animal data. It is 8.7
36      times larger than the slope factor calculated in U.S. EPA, 1985.  This number reflects the
        9/22/00
92
DRAFT—DO NOT CITE OR QUOTE

-------
 1      increase in slope factor based on use of the body burden dose metric (31 times greater) and the
 2      use of the Goodman and Sauer (1992) pathology (3.6 times less).
 3
 4      5.2.1.3. Estimates of Slope Factors and Risk at Current Background Body Burdens Based on
 5      a Mechanistic Model
 6            As discussed above, Portier and Kohn (1996) combined the biochemical response model
 7      of Kohn et al. (1993) with a single initiated-phenotype two-stage model of carcinogenesis to
 8      estimate liver tumor incidence in female Sprague-Dawley rats from the Kociba et al. (1978)
 9      bioassay.  The model is described in more detail in Part II, Chapter 8, Section 8.4. This model
10      adequately fit the tumor data, although it overestimated the observed tumor response at the
11      lowest dose in the Kociba study. The shape of the dose-response curve was approximately linear
12      and the estimated ED01 value for this model was 1.3  ng/kg/day. The corresponding body burden
13      giving a 1% increased effect was 2.7 ng/kg. The model authors believe that the use of CYP1A2
14      as a dose metric for the first mutation rate  is consistent with its role as the major TCDD-
15      inducible  estradiol hydrolase in liver and with its hypothesized role in the production of estrogen
16      metabolites leading to increased oxidative DNA damage and increased mutation (Yager and
17      Liehr, 1996; Hayes et al., 1996; Dannan et al., 1986; Roy et al., 1992). Although no lower bound
18      estimate of the ED01 is calculated, a maximum likelihood estimate of the slope factor can be
19      calculated. It is 7.1 x 10"3 per pg TCDD/kgBW/day. This estimate represents an example of the
20      type of modeling, based on key events in a mode of action for carcinogenesis, which is consistent
21      with future directions in dose-response modeling described in EPA's revised proposed cancer
22      risk assessment guidelines (U.S. EPA, 1999). Although a number of uncertainties remain
23      regarding structure and parameters of the model, the slope estimate is consistent with those
24      derived from humans and animals. More details on this model can be found in Part II, Chapter 8,
25      Section 8.4.
26            An alternative mechanistic model has been proposed (Conolly and Andersen, 1997). This
27      model was developed for focal lesion growth based upon two types of initiated cells applying the
28      negative selection mechanism for hepatic tumor promotion proposed by Jirtle et al. (Jirtle and
29      Meyer, 1991; Jirtle et al., 1991). In this model,  even though the two types of initiated cells
30      express the same biochemical marker, they respond differently to promotional stimulation in the
31      liver. The model presumes that a promotional stimulus to the liver is countered by mito-
32      inhibitory signals generated by the liver to constrain proliferation.  One set of mutated cells is
33      sensitive to this mito-inhibition while the other set of mutated cells is insensitive and responds
34      only to the promotional stimulus. The result is that, under increasing doses of the promoter, one
35      group of focal lesions is decreasing in size, and  hence, number of cells, while the other group is
36      increasing in size. Their model is different from those of Portier and Kohn (1996) in that it can
        9/22/00
93
DRAFT—DO NOT CITE OR QUOTE

-------
  1      result in U-shaped dose-response curves for the total number and mean size of observable focal
  2      lesions without using U-shaped parametric forms for the mutation rates or the birth rates.
  3      Conolly and Andersen (1997) did not apply their model to cancer risk estimation. Presently,
  4      there is insufficient experimental data to support or refute the use of either the Portier and Kohn
  5      (1996) model or that of Conolly and Andersen (1997).
  6
  7      5.2.2. Noncancer Endpoints
  8            At this point, sufficient data are not available to model noncancer endpoints in humans.
  9      Many studies are available to estimate ED0, values for noncancer endpoints in animals.
10      However, there are a number of difficulties and uncertainties that should be considered when
11      comparing the same or different endpoints across species. Some of these include differences in
12      sensitivity of endpoints, tunes of exposure, exposure routes, species and strains, use of multiple
13      or single doses, and variability between studies even for the same response. The estimated ED01s
14      may be influenced by experimental design, suggesting that caution should be used in comparing
15      values from different designs.  Estimates of ED0jS in Part II, Chapter 8 represent estimates of 1%
16      of the maximal response in the studies being evaluated. In addition, caution should be used when
17      comparing studies that extrapolate ED0]s outside the experimental range. Furthermore, it may be
18      difficult to compare values across endpoints.  For example, the human health risk for a 1%
19      change of body weight may not be equivalent to a 1% change in enzyme activity. Similarly, a
20      1 % change in response in a population for a dichotomous endpoint is different from a 1 % change
21      in a continuous endpoint.  Finally, background exposures are not often considered in these
22      calculations simply because they were not known.
23            Nevertheless, given these considerations, several general trends were observed and
24      discussed in Part II, Chapter 8.  The lowest ED0iS tended to be for biochemical effects, followed
25      by hepatic responses,  immune responses, and responses in tissue weight. An analysis of shape
26      parameters implies that many dose-response curves are consistent with linearity over the range of
27      doses tested. This analysis does not imply that the curves would be linear outside this range of
28      doses, but it does inform the choices for extrapolation. This is particularly true when body
29      burdens or exposures  at the lower end of the observed range are close to body burdens or
30      exposures of interest for humans, which is the case with dioxin-like chemicals and biochemical
31      effects.
32             Overall shape parameter data suggest that biochemical responses to TCDD are more
33      likely to be linear within the experimental dose range, while the more complex responses are
34      more likely to assume a nonlinear shape.  However, a large number (> 40%) of the more complex
35      responses have shape parameters that are more consistent with linearity than nonlinearity.
        9/22/00
94
DRAFT—DO NOT CITE OR QUOTE

-------
       The tissue weight changes seen for animals (using only data sets with good or moderate
empirical fits to the model) yielded a median ED01 at average body burdens of 510 ng/kg in the
multidose studies (range; 11 to 28000 ng/kg) and a median ED01 of 160 ng/kg (range 0.0001 to
9,700 ng/kg) in the single dose studies. Toxicity endpoints from the single dose studies resulted
in a median value at average body burdens of 4,300 ng/kg (range 1.3 to 1,000,000 ng/kg). For
tissue weight changes, 43% of the dose-response curves exhibited linear response. In contrast, the
toxicity endpoints from the single-dose studies exhibited predominantly nonlinear responses
(80%).  All multidose studies demonstrated a greater degree of linear response than did single-
dose studies, especially for tissue weight changes and toxicity endpoints (50% linear for
multidose versus 34% for single dose). In general, it is not possible to dissociate the differences
between cancer and noncancer dose-response as being due to differences in endpoint response or
simply to differences in the length of dosing and exposure. Also, a greater  percentage of the
noncancer ED01s were extrapolations below the lower range of the data (42%) than was the case
for the cancer endpoints (8% in animals and no extrapolations in humans).
       Results from the analysis of ED0]s and from examining LOAELs in additional studies
suggest that noncancer effects can occur at body burden levels in animals equal to or less than
body burdens calculated for tumor induction in animals. This is especially true when considering
biochemical changes which may be on the critical path for both noncancer and cancer effects,
such as enzyme induction or impacts on growth factors or their receptors.-While human
noncancer effects were not modeled in Part II Chapter 8, the observation of effects in the Dutch
studies (discussed in Section 2.2.2 in this document) suggest that subtle; but important,
noncancer human effects may be occurring at body burden levels equivalent to those derived for
both many biochemical and some clearly adverse effects in animals (See Table 2-2 for
examples). The use of ED0iS and LOAELs in this analysis provides a "point of departure" for a
discussion of margins of exposure for a variety of health endpoints. No one endpoint has been
chosen as the "critical effect," as is often done in reference dose calculations.  The range of
effects (biochemical, tissue or toxic responses) is presented and individual responses at the low
end of the range in each of these categories are discussed in the development of the hazard
characterization to demonstrate the potential significance of these responses in similarly exposed
humans.
5.3.  MODE-OF-ACTION BASED DOSE-RESPONSE MODELING
       As described in Chapter 8, Section 8.3, mechanism-based modeling can be a powerful
tool for understanding and combining information on complex biological systems.  Use of a truly
mechanism-based approach can, in theory, enable reliable and scientifically sound extrapolations
to lower doses and between species.  However, any scientific uncertainty about the mechanisms
9/22/00
95
DRAFT—DO NOT CITE OR QUOTE

-------
  1      that the models describe is inevitably reflected in uncertainty about the predictions of the models.
  2      The assumptions and uncertainties involved in the mechanistic modeling described in Chapter 8
  3      are discussed at length in that chapter and in cited publications.
  4             The development and continued refinement of PBPK models of the tissue dosimetry of
  5      dioxin have provided important information concerning the relationships between administered
  6      dose and dose to tissue compartments (Part II, Chapter 8, Section 8.2). Aspects of these models
  7      have been validated in the observable response range for multiple tissue compartments, species,
  8      and class of chemical.  These models will continue to provide important new information for
  9      future revisions of this health assessment document. Such information will likely include
10      unproved estimates of tissue dose for liver and other organs where toxicity has been observed,
11      improved estimates of tissue dose(s) in humans, and improved estimates of tissue dose for dioxin
12      related compounds.
13             As a part of this reassessment, the development of biologically based dose-response
14      (pharmacodynamic) models for dioxin and related compounds has lead to considerable and
15      valuable insights regarding both mechanisms of dioxin action and dose-response relationships for
16      dioxin effects.  These efforts, described in some detail in Part II, Chapter 8, Section 8.3, have
17      provided additional perspectives on traditional methods such as the linearized multistage
18      procedure for estimating cancer potency  or the uncertainty factor approach for estimating levels
19      below which noncancer effects are unlikely to occur. These methods have also provided a
20      biologically based rationale for what had been primarily statistical approaches. The development
21      of models like those in Chapter  8 allows for an iterative process" of data'development, hypotheses
22      testing and model development.
23
24      5.4. SUMMARY DOSE-RESPONSE CHARACTERIZATION
25             All humans tested contain detectable body burdens of TCDD and other dioxin-like
26      compounds that are likely to act through the same mode of action.  Receptor modeling theory
27      outlined in Chapter 8 indicates that xenobiotics which operate through receptor binding
28      mechanisms, such as dioxin, will follow a linear dose-response binding in the 1-10% receptor
29      occupancy region. This theoretical basis suggests, and this is supported by empirical findings,
30      that the proximal biochemical and transcription reactions for dioxins may also follow linear
31      dose-response kinetics, such as effects on DNA transcription and enzyme induction. More distal
32      toxic effects could be linear or sublinear/threshold depending on: 1) the toxic mechanism;
33      2) location on the dose-response curve; and 3) interactions with other processes such as
34      intracellular protein binding and co-factor induction/repression.  Empirical data provide dose-
35      response shape information down to approximately the  1% effect level for many toxic endpoints.
36      Many examples of adverse effect experienced at these low levels have too  much data variability
        9/22/00
96
DRAFT—DO NOT CITE OR QUOTE

-------
  1      to clearly distinguish on a statistical basis (goodness-of-fit) between dose-response curve options,
  2      and whether dose-response follows linear, supra/sub-linear, power curve, or threshold kinetics.
  3      Toxic effects seen only at higher doses are presumably more likely to result from multiple
  4      cellular perturbations and are thus less likely to follow linear relationships. Empirical dose-
  5      response data from cancer studies-both human epidemiological and bioassays—do not provide
  6      consistent or compelling information supportive of either threshold or supralinear models (see
  7      Tables 2-4 and 5-2) and are insufficient to move from EPA's default linear extrapolation policy
  8      in the proposed Carcinogen Assessment Guidelines (U.S. EPA, 1996; 1999). This policy is that
  9      for cancer dose-response the data are to be modeled within the observed range, and a point-of-
10      departure calculated from which a linear extrapolation to the origin is generated.  For noncancer
11      endpoints, EPA proposes using a margin of exposure approach due to the inability to determine
12      levels that are likely to be without appreciable effects of lifetime exposure to the population,
13      including susceptible subpopulations, for all adverse effects, particularly given the current level
14      of background exposure and human body burdens. Data on background levels of dioxins, furans
15      and coplanar PCBs (see Part I, Volume 3 and Section 4.4 in this document) indicate that current
16      levels in humans are already substantially along the dose-response curve. Thus, theoretical issues
17      regarding increases from zero body burden levels are moot, and assessments must consider
18      increments of dose to this background level. Margins of exposure between population levels and
19      the empirically observed (not modeled) one percent effect levels for a number of
20      biochemical/toxic endpoints are on the order of less than 1 to 2 orders of magnitude. Thus, the
21      extrapolation between observed effects and background levels  i's not large,  with any increments
22      to background further advancing along the dose-response curve through or toward the observed
23      range.  This further reduces the level of uncertainty when evaluating the significance of margins
24      of exposure. It is possible that any additional exposure above current background body burdens
25      will be additive to ongoing responses. The magnitude of the additional response will be a
26      function of the toxic equivalency of the incremental exposure.  This observation, the relatively
27      small margin of exposure for "key events" potentially on the pathway to cancer and noncancer
28      effects and the high percentage of observed linear responses suggest that a proportional model   ,
29      should be used when extrapolating beyond the range of the experimental data. Short of
30      extrapolating linearly over one to two orders of magnitude to estimate risk  probabilistically for
31      cancer and noncancer effects in the face of uncertainties described above, a simple margin-of-
32      exposure approach may be useful to decision-makers when discussing risk management goals.
33      However, this decision would have to be based upon a policy choice because this analysis does
34      not strongly support either approach.
35             Because human data for cancer dose-response analysis  were available and because of a
36      strong desire to stay within the range of responses estimated by these data,  the risk chosen for
        9/22/00
97
DRAFT—DO NOT CITE OR QUOTE

-------
 1      determining a point of departure was the 1% excess risk. Doses and exposures associated with
 2      this risk (the EDOIs) were estimated from the available data using both mechanistic and empirical
 3      models. Comparisons were made on the basis of body burdens to account for differences in
 4      half-life across the numerous species studied.
 5             In humans., restricting the analysis to log-linear models resulted in cancer ED01s ranging
 6      from 5.7 ng/kg to 250 ng/kg. This was similar to the estimates, from empirical modeling, from
 7      the animal studies which ranged from 14 ng/kg to  1,190 ng/kg (most estimates were in the range
 8      from 14 to 500 ng/kg), and 2.7 ng/kg for the single mechanism-based model. Lower bounds on
 9      these ED0, estimates were used to calculate upper bound slope factors and risk estimates for
10      average background body burdens.
11             Table 5-4 summarizes the ED01/LEDOI and slope factor calculations for the occupational
12      cohort and bioassay studies.  In addition to tabulating the results provided in Part II, Chapter 8,
13      this table includes: 1)  a further calculation of the central estimate from the Hamburg occupational
14      cohort using formulae derived from Becher et al. (1998); 2) a Poisson regression analysis of all
15      three occupational cohorts combined; and 3) benchmark dose (BMD) analyses of the Kociba rat
16      bioassay using  both daily dose and adipose  tissue concentration as the metrics. The slope factor
17      calculations are performed by linearly extrapolating the  LED01 values to the background response
18      rates, consistent with procedures outlined in the draft proposed guidelines for carcinogen risk
19      assessment (U.S. EPA, 1996). A slope factor estimate of approximately 4- x 10'3 per pg
20      TCDD/kgBW/day, based on the meta-analysis, represents EPA's most current upper bound slope
21      factor for estimating human cancer risk based on human data.  A slope factor of 1.4 x 1 o~3 per pg
22      TCDD/kgBW/day represents EPA's most current upper bound slope factor  for estimating
23      human cancer risk based on animal data. Details on the specific procedures and calculations are
24      provided in the footnotes. Additional details on the study characteristics and doserresponse data
25      and graphs are  available in Section 5.2 and  Table 5-2 . The Agency, although fully recognizing
26      the range and the public health conservative nature of the slope factors that make up the range.
27      suggests the use of 1 x 1Q-3 per pg TEO/kgBW/day as an estimator of upper bound cancer risk for
28      both background intakes and incremental intakes above  background.
29             Upper bound slope factors allow the calculation  of the high end (greater than 95%) of the
30      probability  of cancer risk in the population.  This means that there is greater than a 95%  chance
31      that cancer risks will be less than the upper  bound. Use of the ED01, rather than the LEDOI, to
32      provide more likely estimates based on the  available epidemiological and animal cancer data,
33      result in slope factors and risk estimates that are within 2-3 times of the upper bound estimates.
34      Even though there may be individuals in the population who might experience a higher cancer
35      risk on the basis of genetic factors or other determinants of cancer risk not accounted for in
36      epidemiologic data or animal studies, the vast majority of the population is expected to have less
        9/22/00
98
DRAFT—DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
risk per unit of exposure and some may have zero risk. Based on these slope factor estimates
(per pg TEQ/kgBW/day), upper bound cancer risk at average current background body burdens
(5 ng TEQ/kgBW) exceed 10"3 (1 in a thousand).  Current background body burdens reflect
higher average intakes from the past (approximately 3 pgTEQ/kgBW/day).  A very small
percentage of the population (< 1%) may experience risks that are 2-3  times higher than this
upper bound based on average intake if their individual cancer risk slope is represented by the
upper bound estimate and they are among the most highly exposed (among the top 5%) based on
dietary intake of dioxin and related compounds. This range of upper bound risk for the general
population has increased from the risk described at background exposure levels based on EPA's
draft of this reassessment (lO^-lO'3) (U.S. EPA, 1994).
       Estimates for noncancer endpoints showed much greater variability. In general, the
noncancer endpoints displayed lower ED01s for short-term exposures versus longer term
exposures, and for simple biochemical endpoints versus more complex endpoints such as tissue
weight changes or toxicity. In addition, the noncancer endpoints generally displayed higher
estimated ED01s than the cancer endpoints, with most estimates ranging from 100 ng/kg to
100,000 ng/kg.  The mechanism-based models for noncancer endpoints gave a lower range of
EDOIs (0.17 to 105 ng/kg). Although most of these estimates were based upon a single model,
the estimate from a  different model — the hepatic zonal induction model — gave an ED01 for
CYP1A2 induction  of 51 ng/kg and hence was within the same range.   .. -
       These estimates, although highly variable, suggest that any choice of body burden, as a
point of departure, above a body burden of 100 ng/kg would likely yield >1% excess risk for
some endpoint in humans, including those with clear clinical significance. Also, choosing a
point of departure below 1 ng/kg would likely be an extrapolation below the range of these data
and would likely represent a risk of <1%. Any choice in the middle range of 1 ng/kg to 100
ng/kg would be  supported  by the analyses, although the data provide the greatest support in the
range of 10 ng/kg to 50 ng/kg. This range of body burdens should also provide a useful point of
comparison when evaluating impacts of risk management on average body burdens in the general
population or on estimates of impact of incremental exposures above background on individual
body burdens at various ages.
                           6. RISK CHARACTERIZATION

       Characterizing risks from dioxin and related compounds requires the integration of
complex data sets and the use of science-based inferences regarding hazard, mode of action, dose
response, and exposure. It also requires consideration of incremental exposures in the context of
        9/22/00
                                          99
DRAFT—DO NOT CITE OR QUOTE

-------
 1      an existing background exposure that is, for the most part, independent of local sources and
 2      dominated by exposure through the food supply.  Finally, this characterization must consider
 3      risks to special populations and developmental stages (subsistence fishers, children, etc.) as well
 4      as the general population. It is important that this characterization convey the current
 5      understanding of the scientific community regarding these issues, highlight uncertainties in this
 6      understanding, and specify where assumptions or inferences have been used in the absence of
 7      data. Although characterization of risk is inherently a scientific exercise, by its nature it must go
 8      beyond empirical observations and draw conclusions in untested areas. In some cases, these
 9      conclusions are, in fact, untestable given the current capabilities in analytical chemistry,
10      toxicology, and epidemiology.  This situation should not detract from our confidence in a well
11      structured and documented characterization of risk, but should serve to confirm the importance
12      of considering risk assessment as an iterative process that benefits from evolving methods and
13      data collection.
14
15      Dioxin and related compounds can produce a wide variety of effects in animals and might
16      produce many of the same effects in humans.
17            There is adequate evidence based on all available information discussed in Parts I and II
18      of this reassessment, as well as that discussed in this Integrated Summary, to support the
19      inference that humans are likely to respond with a broad spectrum of effects from exposure to
20      dioxin and related compounds. These effects will likely range from biochemical changes at or
21      near background levels of exposure to adverse effects with increasing severity as body burdens
22      increase above background levels. Enzyme induction, changes in hormone levels, and indicators
23      of altered cellular function seen in humans and laboratory animals represent effects of unknown
24      clinical significance but that may be early indicators of toxic response. Induction of
25      activating/metabolizing enzymes at or near background levels, for instance, may be adaptive, and
26      in some cases, beneficial, or may be considered adverse. Induction may lead to more rapid
27      metabolism and elimination of potentially toxic compounds, or may lead to increases in reactive
28      intermediates and may potentiate toxic effects. Demonstrations of examples of both of these
29      situations are available in the published literature and events of this type formed the basis for a
30      biologically based model discussed in Section 5.  Subtle effects, such as the impacts on
31      neurobehavioral outcomes, thyroid function, and immune system alterations seen in the Dutch
32      children exposed to background levels of dioxin and related compounds, or changes in
33      circulating reproductive hormones in men exposed to TCDD, illustrate the types of responses
34      that support the finding of arguably adverse effects at or near background body burdens. Clearly
35      adverse effects including, perhaps, cancer may not be detectable until exposures contribute to
36      body burdens that exceed background by one or two orders of magnitude (10 or 100 times). The
        9/22/00
100
DRAFT—DO NOT CITE OR QUOTE

-------
 1      mechanistic relationships of biochemical and cellular changes seen at or near background body
 2      burden levels to production of adverse effects detectable at higher levels remain uncertain.
 3      Information on these mechanistic relationships is useful in hazard characterization and data are
 4      accumulating to suggest mode of action hypotheses for further testing.
 5             It is well known that individual species vary in their sensitivity to any particular dioxin
 6      effect.  However, the evidence available to date indicates that humans most likely fall in the
 7      middle of the range of sensitivity for individual effects among animals rather than at either
 8      extreme. In other words, evaluation of the available data suggests that humans, in general, are
 9      neither extremely sensitive nor insensitive to the individual effects of dioxin-like compounds.
10      Human data provide direct or indirect support for evaluation of likely effect levels for several of
11      the endpoints discussed in the reassessment, although the influence of variability among humans
12      remains difficult to assess.  Discussions have highlighted certain prominent, biologically
13      significant effects of TCDD and related compounds. In TCDD-exposed men, subtle changes in
14      biochemistry and physiology such as enzyme induction, altered levels of circulating reproductive
1 5      hormones, or reduced glucose tolerance and, perhaps, diabetes, have been detected in a limited
16      number of epidemiologic studies. These findings, coupled with knowledge derived from animal
17      experiments, suggest the potential for adverse impacts on human metabolism, and developmental
18      and/or reproductive biology, and, perhaps, other effects in the range of current human exposures.
19      These biochemical, cellular, and organ-level endpoints have been shown to be affected by
20      TCDD, but specific data on these endpoints do not generally exist for other congeners. Despite
21      this lack of congener-specific data, there is reason to infer that these effects may occur for all
22      dioxin-like compounds, based on the concept of toxic equivalency.
23             In this document, dioxin and related compounds are characterized as carcinogenic,
24      developmental, reproductive, immunological, and endocrinological hazards. The deduction that
25      humans are likely to respond with noncancer effects from exposure to dioxin-like compounds is
26      based on the fundamental level at which these compounds impact cellular regulation and the
27      broad range of species that have been demonstrated to respond with adverse effects.  For
28      example, because developmental toxicity following exposure to TCDD-like congeners occurs in
29      fish, birds, and mammals, it is likely to occur at some level in humans.  It is not currently
30      possible to state exactly how or at what levels individuals will respond with specific adverse
31      impacts on development or reproductive function, but analysis of the Dutch cohort data and
32      laboratory animal studies suggests that some effects may occur at or near background levels.
33      Fortunately, there have been few human cohorts identified with TCDD exposures high enough to
34      raise body burdens significantly over background levels (see Table 5-1 and Figure 5-1 in this
35      document) and when these cohorts have been examined, relatively few clinically significant
36      effects  were detected. However, the power of these studies to detect these effects remains an
        9/22/00
101
DRAFT—DO NOT CITE OR QUOTE

-------
 1      issue. The lack of sufficient exposure gradients and adequate human information and the focus
 2      of most currently available epidemiologic studies on occupationally TCDD-exposed adult males
 3      makes it difficult to evaluate the inference that noncancer effects associated with exposure to
 4      dioxin-like compounds may be occurring in humans.  It is important to note, however, that when
 5      exposures to very high levels of dioxin-like compounds have been studied, such as in the Yusho
 6      and Yu-Cheng cohorts, a spectrum of adverse effects have been detected in men, women, and
 7      children.  Some have argued that to deduce that a spectrum of noncancer effects will occur in
 8      humans in the absence of better human data overstates the science; most scientists involved in
 9      the reassessment as authors and reviewers have indicated that such inference is reasonable given
10      the weight-of-the-evidence from available data. As presented, this logical conclusion represents
11      a testable hypothesis which may be evaluated by further data collection.  EPA, its Federal
12      colleagues, and others in the general scientific community are continuing to fill critical data gaps
13      that will reduce our uncertainty regarding both hazard and risk characterization for dioxin and
14      related compounds.
15
16      Dioxin and related compounds are structurally related and elicit their effects through a
17      common mode of action.
18             The scientific community has identified and described a series of common biological
19      steps that are necessary for most, if not all, of the observed effects of dioxin and related
20      compounds in vertebrates including humans.  Binding of dioxin-like compounds to a cellular
21      protein called the aryl hydrocarbon receptor (AhR) represents the first step in a series of events
22      attributable to exposure to  dioxin-like compounds including biochemical, cellular,  and tissue-
23      level changes hi normal biological processes. Binding to the AhR appears to be necessary for all
24      well-studied effects of dioxin but is not sufficient, in and of itself, to elicit these responses.
25      There remains some uncertainty as to whether every dioxin response is AhR-mediated. Some
26      data from the use of sensitive biological tools such as AhR deficient (AhR"7") mice suggest a
27      small residual of effects from exposure to TCDD that does not allow us to rule out receptor-
28      independent alternative pathways.  However, these reported non-AhR mediated responses occur
29      at doses that are orders of magnitude higher than human exposures and require much higher
30      doses than other AhR mediated effects in animals. Thus, these non-AhR mediated mechanisms
31      are unlikely to impact any  of the assumptions made in this reassessment.  The well-documented
32      effects elicited by exposure of animals and, in some cases, humans, to 2,3,7,8-TCDD are shared
33      by other chemicals with similar structure and AhR binding characteristics.  In the past 5 years,
34      significant data has accumulated that support the concept of toxic equivalence, a concept that is
35      at the heart of risk assessment for the complex mixtures of dioxin and related compounds
36      encountered in the environment. These data have been analyzed and summarized in Part II,
        9/22/00
102
DRAFT—DO NOT CITE OR QUOTE

-------
Chapter 9. This chapter has been added to EPA's dioxin reassessment to address questions raised
by the SAB in 1995. The SAB suggested that, because the TEQ approach was a critical
component of risk assessment for dioxin and related compounds, the Agency should be explicit
in its description of the history and application of the process and go beyond reliance on the
Agency's published reference documents on the subject (U.S. EPA, 1987, 1989).
       Analyses in this document demonstrate that, although variability in the data underpinning
the scientific judgments regarding toxic equivalency exist, when data are restricted to longer
exposure and in vivo data, the empirical analysis strongly supports the judgment of experts in
setting TEF values. This is particularly true for the use of TEFs for assessing the animal cancer
endpoint, but will likely apply even more strongly to noncancer effects as additional congener-
specific data are collected.

EPA and the international scientific community have adopted toxic equivalency of dioxin
and related compounds as prudent science policy.
       Dioxin and related compounds always exist in nature as complex mixtures. As discussed
in the Exposure Document, these complex mixtures can be characterized through analytic
methods to determine concentrations of individual congeners. Dioxin and related compounds
can be quantified and biological activity of the mixture can be estimated using relative potency
values and an assumption of dose additivity. Such an approach has evolved over time to form
the basis for the use of TEQ in risk assessment for this group of compounds. Although such an
approach is dependent on critical assumptions and scientific judgment, it has been characterized
as a "useful, interim" way to deal with the complex mixture problem and has been accepted by
numerous countries and several international organizations. Alternative approaches, including
the assumption that all congeners carry the toxic equivalency of 2,3,7,8-TCDD, or that all
congeners other than 2,3,7,8-TCDD can be ignored, have been generally rejected as inadequate
for risk assessment purposes.
       Significant additional literature is now available on the subject of toxic equivalency of
dioxin and related compounds, and Part II, Chapter 9 provides the reader with a summary that is
up to date through 1999.  A recent international evaluation of all of the available data (van den
Berg et al., 1998) has reaffirmed the TEQ approach and has provided the scientific community
with the latest values for  TEFs for PCDDs, PCDFs, and dioxin-like PCBs. Consequently, we can
infer with greater confidence that humans will respond to the cumulative exposure of AhR-
mediated chemicals. This reassessment recommends that the WHO9g TEF scheme be used to
assign toxic equivalency  to complex environmental mixtures for assessment and regulatory
purposes. Future research will be needed to address remaining uncertainties inherent in the
current approach.  The WHO has suggested  that the TEQ scheme be reevaluated on a periodic
9/22/00
103
DRAFT—DO NOT CITE OR QUOTE

-------
  1      basis and that TEFs and their application to risk assessment be reanalyzed to account for
  2      emerging scientific information.
  3
  4      Complex mixtures of dioxin and related compounds are highly potent, "likely"
  5      carcinogens.
  6             A weight-of-the-evidence evaluation suggests that mixtures of dioxin and related
  7      compounds (CDDs, CDFs, and dioxin-like PCBs) are strong cancer promoters and weak direct or
  8      indirect initiators, and are likely to present a cancer hazard to humans. Because dioxin and
  9      related compounds always occur in the environment and in humans as complex mixtures of
10      individual congeners, it is appropriate that the characterization apply to the mixture. According
11      to the Agency's revised proposed guidelines for carcinogen risk assessment, the descriptor
12      "likely" is appropriate when the available tumor effects and other key data are adequate to
13      demonstrate carcinogenic potential to humans (U.S. EPA, 1999). Adequate data are recognized
14      to span a wide range.  The data for complex mixtures of dioxin and related compounds represents
15      a case that, according to the draft Guidelines, would approach the strong-evidence end of the
16      adequate-data spectrum.  Epidemiologic observations of an association between exposure and
17      cancer responses (TCDD); unequivocal positive responses in both sexes, multiple species,
18      multiple sites, and different routes in lifetime bioassays or initiation-promotion protocols or other
19      shorter-term in vivo systems such as transgenic models (TCDD plus numerous PCDDs, PCDFs,
20      dioxin-like PCBs); and mechanistic or mode-of action data that are assumed to be relevant to
21      human carcinogenicity, including, for instance, initiation-promotion studies (PCDDs, PCDFs,
22      dioxin-like PCBs) all support the description of complex mixtures of dioxin and related
23      compounds as likely human carcinogens.
24             Even though the database from cancer epidemiologic studies remains a point of scientific
25      discussion, it is the view of this reassessment that this body of evidence is supported by the
26      laboratory data indicating that TCDD probably increases cancer mortality of several types.
27      Although not all confounders were ruled out in any one study, positive associations between
28      surrogates of dioxin exposure, either length of occupational exposure or proximity to a known
29      source combined with some information based on measured blood levels, and cancer have been
30      reported. These data suggest a role for dioxin exposure to contribute to a carcinogenic response
31      but are not sufficient to confirm a causal relationship between exposure to dioxin and increased
32      cancer incidence. Available human studies alone cannot demonstrate whether a cause-and-effect
33      relationship between dioxin exposure and increased incidence of cancer exists. Therefore,
34      evaluation of cancer hazard in humans must include an evaluation of all of the available animal
35      and in vitro data as well as the data from exposed human populations.
36             As discussed earlier in Section 2.2.1.4, under EPA's current approach individual
37      congeners can also be characterized as to their carcinogenic hazard. 2,3,7,8-tetrachlorodibenzo-
        9/22/00
104
DRAFT—DO NOT CITE OR QUOTE

-------
 1     j?-dioxin (TCDD) is best characterized as "carcinogenic to humans." This means that, based on
 2     the weight of all of the evidence (human, animal, mode of action), TCDD meets the criteria that
 3     allow EPA and the scientific community to accept a causal relationship between TCDD exposure
 4     and cancer hazard. The  guidance suggests that "carcinogenic to humans" is an appropriate
 5     descriptor of human carcinogenic potential when there is an absence of conclusive epidemiologic
 6     evidence to clearly establish a cause-and-effect relationship between human exposure and cancer,
 7     but there is compelling carcinogenicity in animals and mechanistic information in animals and
 8     humans demonstrating similar modes of carcinogenic action.  The "carcinogenic to humans"
 9     descriptor is suggested for TCDD because all of the following conditions are met:
10         •    There is strong and consistent evidence from occupational epidemiologic studies for an
11             association between TCDD exposure and increases in cancer at all sites, in lung cancer
12             and, perhaps, at other sites, but the data are insufficient on their own to demonstrate a
13             causal association.
14         •    There is extensive carcinogenicity in both sexes of multiple species at multiple sites.
15         •    There is general  agreement that the mode of TCDD's carcinogenicity is AhR dependent
16             and proceeds through modification of the action of a number of receptor and hormone
17             systems involved in cell growth and differentiation, such as the epidermal growth factor
18             receptor and estrogen receptor.
19         •    The human AhR and rodent AhR are similar in structure and function and, once
20             transformed, both bind to the same DNA response elements, designated DRE's.
21         •    Human and rodent tissue and organ cultures respond to TCDD and related chemicals in a
22             similar manner and at similar concentrations.
23
24             Other individual dioxin-like compounds are characterized as "likely" human carcinogens
25     primarily because of the lack of epidemiological evidence associated with their carcinogenicity,
26     although the  inference based on toxic equivalency is strong that they would behave in humans as
27     TCDD does.  Other factors, such as the lack of congener-specific chronic bioassays, also support
28     this characterization.  For each congener, the degree of certainty is dependent on the available
29     congener-specific data and their consistency with the generalized mode of action that underpins
30     toxic equivalency for TCDD and related compounds.  On the basis of this logic, complex
31     environmental mixtures  of TCDD and dioxin-like compounds should be characterized as "likely"
32     carcinogens,  with the degree of certainty of the characterization being dependent on the
33     constituents of the mixture, when known.  For instance, the hazard potential, although "likely,"
34     would be characterized differently for a mixture whose TEQ was dominated by OCDD as
35     compared with one dominated by pentaCDF.
36             Although uncertainties remain regarding quantitative estimates of upper bound cancer
37     risk from dioxin and related compounds, efforts of this reassessment to bring more data into the
        9/22/00
105
DRAFT—DO NOT CITE OR QUOTE

-------
  1      evaluation of cancer potency have resulted in evaluation of the slope of the dose-response curve
  2      at the low end of the observed range (using the LED01) using a simple proportional (linear) model
  3      and a calculation of both upper bound risk and margin of exposure (MOE) based on human
  4      equivalent background exposures and associated body burdens. Evaluation of shape parameters
  5      (used to estimate degree of linearity or nonlinearity of dose-response within the range of
  6      observation) for biochemical effects indicates that many of these biochemical effects can be
  7      hypothesized as key events in a generalized dioxin mode-of-action model.  These analyses do not
  8      argue for significant departures from linearity below a calculated ED0] for endpoints potentially
  9      related to cancer response, extending down to at least one to two orders of magnitude lower
10      exposure.
11            Risk estimates for intakes associated with background body burdens or incremental
12      exposures based on this slope factor represent a plausible upper bound on risk based on the
13      evaluation of animal and human data. The slope factors, based on the most sensitive cancer
14      responses calculated in Section 5 for both animals and humans, fall in a range of approximately
15      1 *10"3 to 9 x 10'3 per pg/TEQ/kgBW/day. The ranges of estimates of upper bound cancer
16      potency calculated from the human and animal data analyzed in Part II, Chapter 8, Section 8.3,
17      overlap. The range above is bounded on the upper end by the estimate of slope from the
18      Hamburg cohort epidemiology study and on the lower end by the estimate from the results of the
19      meta-analysis of the three human studies and from reanalyzed Kociba study. Consequently, the
20      Agency, although fully recognizing this range and the public health conservative nature of the
21      slope factors that make up the range, suggests the use of 1 x  10'3 per pg/TEO/kgBW/day as an
22      estimator of upper bound cancer risk for both background intakes and incremental intakes above
23      background. This decision reflects the weight given to the meta-analytic estimate from the
24      human studies and the comparability of the revised estimate  from the animal data. Upper bound
25      slope factors allow the calculation of the high end (greater than 95%) of the probability of cancer
26      risk in the population. This means that there is greater than a 95% chance that cancer risks will
27      be less than the upper bound.  Use of the ED01, rather than the LEDOI, to provide more likely
28      estimates based on the available epidemiological and animal cancer data, result in slope factors
29      and risk estimates that are within 2-3 times of the upper bound estimates. Even though there may
30      be individuals in the population who might experience a higher cancer risk  on the basis of
31      genetic factors or other determinants of cancer risk not accounted for in epidemiologic data or
32      animal studies, the vast majority of the population is expected to have less risk per unit of
33      exposure and some may have zero risk.  Based on these slope factor estimates (per
34      pg/TEQ/kgBW/day), risks at average current background body burdens (5 ng TEQ/kgBW) that
35      result from average intakes of approximately 3 pgTEQ/kgBW/day in the past exceed 10"3 (1 in a
36      thousand). A very small percentage of the population (< 1%) may experience risks that are 2-3
37      tunes higher than this upper bound based on average intake if their individual cancer risk slope is
        9/22/00
106
DRAFT—DO NOT CITE OR QUOTE

-------
 1      represented by the upper bound estimate and they are among the most highly exposed (among the
 2      top 5%) based on dietary intake of dioxin and related compounds. This range of upper bound
 3      risk for the general population has increased from the risk described at background exposure
 4      levels based on EPA's draft of this reassessment (lO^-lO'3) (U.S. EPA, 1994).
 5            Despite the use of the epidemiology data to describe an upper bound on cancer risk, the
 6      Peer Panel that met in September 1993 to review an earlier draft of the cancer epidemiology
 7      chapter suggested that the epidemiology data alone were still not adequate to implicate dioxin
 8      and related compounds as "known" human carcinogens, but that the results from the human
 9      studies were largely consistent with observations from laboratory studies of dioxin-induced
10      cancer and, therefore, should not be dismissed or ignored. Other scientists, including those who
11      attended the Peer Panel meeting, felt either more or less strongly about the weight of the
12      evidence from cancer epidemiology studies, representing the range of opinion that still exists on
13      the interpretation of these studies.  Similar opinions were expressed in the comments
14      documented in the SAB's report in 1995 (U.S. EPA,1995). More recently, IARC  (1997), in its
15      reevaluation of the cancer hazard of dioxin and related compounds, found that whereas the
16      epidemiologic database for 2,3,7,8-TCDD was still "limited," the overall weight of the evidence
17      was sufficient to characterize 2,3,7,8-TCDD as a Category 1 "known" human carcinogen. Other
18      related members of the class of dioxin-like compounds were considered to have "inadequate"
19      epidemiologic data to factor into hazard categorization.  A similar classification has been
20      proposed within the context of the Department of Health and Human Services' Report on
21      Carcinogens (NTP, 2000). They too base their characterization on the broad base  of human,
22      animal, and mode-of-action information in humans and animals that supports this  conclusion.
23      Therefore, given that 2,3,7,8-TCDD is contained in complex mixtures of dioxin and related
24      compounds, and that the TEQ approach has been adopted as a reasonable approach to assessing
25      risks  of these complex mixtures, it is also reasonable to apply estimates of upper bound cancer
26      potency derived from epidemiology studies where 2,3,7,8-TCDD was associated with excess
27      cancer risk to  complex mixtures of dioxin and related compounds.
28            The current evidence suggests that both receptor binding and most early biochemical
29      events such as enzyme induction are likely to demonstrate low-dose linearity.  The mechanistic
30      relationship of these early events to the complex process of carcinogenesis  remains to be
31      established. If these findings imply low-dose linearity in biologically based cancer models under
32      development,  then the probability of cancer risk will be linearly related to exposure to TCDD at
33      low doses. Until the mechanistic relationship between early cellular responses and the
34      parameters in biologically based cancer models is better understood, the shape of the dose-
35      response curve for cancer below the range of observation can only be inferred with uncertainty.
36      Associations between exposure to dioxin and certain types of cancer have been noted in
37      occupational cohorts with average body burdens of TCDD approximately 1-3  orders of
        9/22/00
107
DRAFT—DO NOT CITE OR QUOTE

-------
  1      magnitude (10-1,000 times) higher than average TCDD body burdens in the general population.
  2     The average body burden in these occupational cohorts level is within 1-2 orders of magnitude
  3     (10-100 times) of average background body burdens in the general population in terms of TEQ
  4     (see Table 5-1 and Figure 5-1).  Thus, there is no need for large-scale low-dose extrapolations in
  5     order to evaluate background intakes and body burdens, and little if any data to suggest large
  6     departures from linearity in this somewhat narrow window between the lower end of the range of
  7     observation and the range of general-population background exposures. Nonetheless, the
  8     relationship of apparent increases in cancer mortality in these worker populations to calculations
  9     of general population risk remains a source of uncertainty.
10            TCDD has been clearly shown to increase malignant tumor incidence in laboratory
11      animals.  In addition, a number of studies analyzed in this reassessment demonstrate other
12     biological effects of dioxins related to the process of carcinogenesis.  Initial attempts to construct
13     a biologically based model for certain dioxin effects as described in this reassessment will need
14     to be continued and expanded to accommodate more of the available biology and to apply to a
15     broader range of potential health effects associated with exposure to dioxin-like compounds.
16
17     Use a "margin-of-exposure "approach to evaluate risk for noncancer and cancer endpoints.
18            The likelihood that noncancer effects may be occurring in the human population at
19      environmental exposure levels is often evaluated using a MOE approach,-The Agency has used
20     this approach for a number of years in  its assessment of the safety of pesticides. This  concept has
21      also been incorporated into the revised proposed Guidelines for Carcino'gen Risk Assessment. A
22     MOE is calculated by dividing a "point of departure" for extrapolation purposes at the low end of
23      the range of observation in human or animal studies (the human-equivalent animal lowest
24     observed adverse effect level (LOAEL), NOAEL, BMD, or effective dose [EDxx]) by the human
25      exposure or body burden level of interest. Generally speaking, when considering either
26      background exposures or incremental exposures plus background, MOEs in the range  of 100-
27      1,000 are considered adequate to rule out the likelihood of significant effects occurring in
28      humans based on sensitive animal responses or results from epidemiologic studies.  The
29      adequacy of the MOE to be protective  of health must take into account the nature of the effect at
30     the "point of departure," the slope of the  dose-response curve, the adequacy of the overall
31      database, interindividual variability in the human population, and other factors. Considering
32     MOEs based on incremental exposures alone divided by the human exposure of interest, is not
33      considered to give an accurate portrayal of the implications of that exposure unless background
34     exposures are insignificant.
35             One of the difficulties in assessing the potential health risk of dioxins is that background
36     exposures may not be insignificant when based on total TEQ.  The average levels of background
37      intake and associated body burdens of dioxin-like compounds in terms of TEQs in the general
        9/22/00
108
DRAFT—DO NOT CITE OR QUOTE

-------
 1      population are well within a factor of 100 of human-equivalent exposure levels associated with
 2      NOELS, LOAELs, BMDs, or EDOI values in laboratory animals exposed to TCDD or TCDD
 3      equivalents. In many cases, the MOE compared to background using these endpoints is a factor
 4      of 10 or less (see Tables 2-2 and 2-3). These estimates, although variable, suggest that any
 5      choice of body burden, as a point of departure, above 100 ng/kg would likely yield >1% excess
 6      risk for some endpoint in humans (see Part II, Chapter 8).  Also, choosing a point of departure
 7      below 1 ng/kg would likely be an extrapolation below the range of these data and would likely
 8      represent a risk of < 1 %. Any choice for a point of departure in the middle range of 1 ng/kg to
 9      100 ng/kg would be supported by the analyses, although the data provide the greatest support for
10      a point of departure in the range of 10 ng/kg to 50 ng/kg. This range of body burdens should also
11      provide a useful point of comparison when evaluating impacts of risk management on average
12      body burdens in the general population or on estimates of impact of incremental exposures above
13      background on individual body burdens at various ages.
14             Because of the relatively high background compared to effect levels, the Agency is not
15      recommending the derivation of a reference dose (RfD) for dioxin and related compounds.
16      Although RfDs are often useful because they represent a health risk goal below which there is
17      likely to be no appreciable risk of noncancer effects over a lifetime  of exposure, their primary use
18      is to evaluate increments of exposure from specific sources when background exposures are low
19      and insignificant. Any RfD that the Agency would recommend under the traditional approach for
20      setting an RfD is likely to be 2-3 orders of magnitude (100-1,000) below current background
21      intakes and body burdens.  Because exceeding the RfD is not a •statement of risk, discussion of an
22      RfD for an incremental exposure when the RfD has already been exceeded by average
23      background exposures is meaningless.
24             When evaluating incremental exposures associated with specific sources, knowing the
25      increment relative to background may help to understand the impact of the incremental exposure.
26      For instance, it would be misleading  to suggest that an incremental  exposure of 0.001 pg
27      TEQ/kg/day was below the RfD if "background" exposures were already at or above that level.
28      On the  other hand, as part of the total, the increment represents less than a 0.1% increase over
29      average "background," and we estimate that individuals within the 50%-95% range of exposure
30      within the population may be 2-3 times (200%-300%) higher. This has led us to suggest that
31      perhaps the best information for a decision-maker to have is: (1) a characterization of average
32      "background" exposures; (2) a characterization of the percent increase over background of
33      individuals or subpopulations of interest; and (3) a policy statement about when increases over
34      average "background" become significant for the decision. This is  not easy because one could
35      argue that, given high "background," any addition, if it is widespread, is too much. On the other
36      hand, someone else could argue that  a 10% increase in incremental exposure for a small
37      population around a specific point source would be well within the  general population exposures
        9/22/00
109
DRAFT—DO NOT CITE OR QUOTE

-------
 1      and would not constitute a disproportionate exposure or risk. In this case, the strategy might be
 2      to bring average "background" exposures down and to focus on large incremental exposures or
 3      highly susceptible populations.  This would be a strategy that would parallel the Agency's lead
 4      strategy.  Other parallel issues between dioxin-like compounds and lead are under discussion
 5      within the Agency.
 6            ATSDR (1999a) set a minimal risk level (MRL), which is defined similarly to the EPA's
 7      RfDj for dioxin and related compounds of 1.0 pg TEQ/kgBW/day. Some of the data regarding
 8      lower bounds on the ED01s from various noncancer effects call that MRL into question. WHO
 9      (2000) has set a tolerable daily intake of 1-4 pg TEQ/kgBW/day and has indicated that, although
10      current exposures in that range are "tolerable" (a risk management decision rather than a risk
11      assessment), efforts should be made to ultimately reduce intake levels.  Findings in this
12      reassessment are supportive of that recommendation.
13
14      Children's risk from exposure to dioxin and related compounds may be increased, but
15      more data are needed to fully address this issue.
16            The issue of children's risk from exposure to dioxin-like compounds has been addressed
17      hi a number of sections throughout this reassessment. Data suggest a sensitivity of response in
18      both humans and animals during the developmental period, both prenatally and postnatally.
19      However, data are limited. Because evaluation of the impacts of early exposures on both
20      children's health and health later in life is important to a complete characterization of risk,
21      collection of additional data in this area should be a high priority to reduce uncertainties in future
22      risk assessments.
23            Data from the Dutch cohort of children exposed to PCBs and dioxin-like compounds
24      suggest impacts from exposure to background levels of dioxin and related compounds prenatally
25      and, perhaps, postnatally on neurobehavioral  outcomes, thyroid function, and immune system
26      alterations.  Although these effects cannot be attributed solely to dioxin and related compounds,
27      several associations suggest that these are, in fact, likely to be Ah-mediated effects.  An
28      investigation of background dioxin exposure  and tooth development was done in Finnish
29      children as a result of studies of dental effects in dioxin-exposed rats, mice, and nonhuman
30      primates, and in PCB-exposed children.  The Finnish investigators examined enamel
31      hypomineralization of permanent first molars in 6-7 year old children.  The length of time that
32      infants breast fed was not significantly associated with either mineralization changes or with
33      TEQ levels in the breast milk. However, when the levels and  length of breast feeding were
34      combined in an overall score, a statistically significant association was observed.
35            In addition, effects have been seen where significantly elevated exposure occurred.  The
36      incidents at Yusho and Yu-Cheng resulted in increased perinatal mortality and low birthweight in
37      infants born to women who had been exposed.  Rocker bottom heal was observed in .Yusho
        9/22/00
110
DRAFT—DO NOT CITE OR QUOTE

-------
 1      infants, and functional abnormalities have been reported in Yu-Cheng children.  The similarity of
 2      effects observed in human infants prenatally exposed to the complex mixture in Yusho and
 3      Yu-Cheng with those reported in adult monkeys exposed only to TCDD suggests that at least
 4      some of the effects on children are due to the TCDD-like congeners in the contaminated rice oil
 5      ingested by the mothers of these children. The similar responses include a clustering of effects in
 6      organs derived from the ectodermal germ layer, referred to as ectodermal dysplasia, including
 7      effects on the skin, nails, and Meibomian glands; and developmental and psychomotor delay
 8      during developmental and cognitive tests. Some investigators believe that because all of these
 9      effects in the Yusho and Yu-Cheng cohorts do not correlate with TEQ, some of the effects are
10      exclusively due to nondioxin-like PCBs or a combination of all the congeners.  In addition, on
1.1      the basis of these data, it is still not clear to what extent there is an association between overt
12      maternal toxicity and embryo/fetal toxicity in humans.  Further studies in the offspring as well as
13      follow-up of the Seveso incident may shed further light on this issue.  In addition to chloracne
14      and acute responses to TCDD exposure seen in Seveso  children, elevated levels of serum GOT
15      have been observed within a year after exposure in some of the more highly exposed Seveso
16      children.  Long-term pathologic consequences of elevated GGT have not been illustrated by
17      excess mortality from liver disorders or cancer or in excess morbidity, but further follow-up is
18      needed.  It must be recognized that the absence of an effect thus far does not obviate the
19      possibility that the enzyme levels may have increased concurrent to the exposure but declined
20      after cessation. The apparently transient elevations in ALT levels among the Seveso children
21      suggest that hepatic enzyme levels other than GGT may react in this manner to 2,3,7,8-TCDD
22      exposure.  Recent studies in Seveso have also demonstrated an altered sex ratio in the second
23      generation (Mocarelli et al., 2000)
24            Impacts on thyroid hormones provide an example of an effect of elevated postnatal
25      exposure to dioxin and related compounds. Several studies of nursing infants suggest that
26      ingestion of breast milk with a higher dioxin TEQ may alter thyroid function. Thyroid hormones
27      play important roles in the developing nervous system of all vertebrate species, including
28      humans. In fact, thyroid hormones are considered so important in development that in the United
29      States all infants are tested for hypothyroidism shortly after birth.  Results from the studies
30      mentioned above suggest a possible shift in the population distribution of thyroid hormone
31      levels, particularly T4, and point out the need for collection of longitudinal data to assess the
32      potential for long-term effects associated with developmental exposures. The exact processes
33      accounting for these observations in humans are unknown, but when put in perspective of animal
34      responses, the following might apply.  Dioxin increases the metabolism and excretion of thyroid
35      hormone, mainly T4, in the liver. Reduced T4 levels stimulate the pituitary to secrete more TSH,
36      which enhances thyroid hormone production.  Early in  the disruption process, the body can
37      overcompensate for the loss of T4, which may result in a small excess of circulating T4 in
        9/22/00
111
DRAFT—DO NOT CITE OR QUOTE

-------
  1      response to the increased TSH. In animals, given higher doses of dioxin, the body is unable to
  2     maintain homeostasis, and TSH levels remain elevated and T4 levels decrease.
  3            A large number of studies in animals have addressed the question of effects of dioxin-like
  4     chemicals after in utero or lactational exposure. These have included both single-congener
  5     studies and exposures to complex mixtures.  However, the vast majority of the data are derived
  6     from studies of 2,3,7,8-TCDD, or single congeners (e.g., PCB 77) or commercial mixtures of
  7     PCBs. Exposure patterns have included single doses to the dams as well as dosing on multiple
  8     days during gestation beginning as early as the first day of gestation. These studies are discussed
  9     in detail in Part II, Chapter 5. The observed toxic effects include developmental toxicity,
10     neurobehavioral  and neurochemical alterations, endocrine effects, and developmental
11      immunotoxicity. For instance, results of this body of work suggest that 2,3,7,8-TCDD clearly
12     has the potential to produce alterations in'male reproductive function (rats, mice, hamsters), male
13     sexual behavior (rats), and female genitalia (rats, hamsters) after prenatal exposure. In addition,
14     impacts on neuromotor and cognitive behavior as well as development of the immune system
15     have been indicated in a number of studies.
16            No epidemiological  data and limited animal data are available to address the question of
17     the potential impact of exposure to dioxin-like compounds on childhood cancers or on cancers of
18      later life. Given the relative impact of nursing on body burdens (see the discussion of breast milk
19      exposures and body burdens below), direct impacts of increased early postnatal exposure on the
20      carcinogenic process are expected to be small.  This conclusion is based on the reasonable
21      assumptions that cancer risk is a function of average lifetime body burden or that, because dioxin
22      is a potent cancer promoter rather than a direct initiator of the cancer process, exposures later in
23      life might be more important than those received earlier.  However, recent studies of Brown et al.
24     (1998) suggest that prenatal exposure of rats to dioxin and related compounds may indirectly
25      enhance their sensitivity as adults to chemical carcinogenesis from other chemical carcinogens.
26      Further work is needed to evaluate this issue.
27             In addition to potential vulnerability during development, fetuses, infants, and children
28      are exposed to dioxins through several routes. The fetus is exposed in utero  to levels of dioxin
29      and related compounds that reflect the body burden of the mother. It is important to recognize
30      that it is not the individual meals a pregnant woman eats during pregnancy that might affect
31      development, but the consequence of her exposure history over her life, which has the greatest
32      impact on her body burden. Again, good nutrition, including a diet with appropriate levels of fat,
33      has consequences on dietary intake and consequent body burdens of dioxin and related
34     compounds. Nursing infants represent special cases who, for a limited portion of their lives, may
35      have elevated exposures on a body-weight basis when compared with non-nursing infants and
36      adults (see discussion). In addition to breast milk exposures, intakes of CDD/CDFs and dioxin-
37      like PCBs are more than three times higher for a young child than those of an adult, on a body-
        9/22/00
112
DRAFT—DO NOT CITE OR QUOTE

-------
 1      weight basis.  Table 4-9 in Section 4 of this document describes the variability in average intake
 2      values as a function of age using age-specific food consumption rates and average food
 3      concentrations, as was done for adult intake estimates. However, as with for the nursing infants,
 4      the differences in body burden between children and adults are expected to be much less than the
 5      differences in daily intake. Assuming that body burden is the relevant dose metric for most if not
 6      all effects, there is some assurance that these increased intake levels will have limited additional
 7      impact on risk as compared with overall lifetime exposure.
 8
 9      Background exposures to dioxin and  related compounds need to be considered when
10      evaluating both hazard and risk.
11            The term "background" exposure has been used throughout this reassessment to describe
12      exposure of the general  population, who are exposed to levels in environmental media (food, air,
13      soil, etc.) that have dioxin concentrations within the normal background range.  Adult daily
14      intakes of CDD/CDFs and dioxin-like PCBs are estimated to average 41 and 24 pg TEQDFP-
15      WHO98/day, respectively,  for a total intake of 65 pg/day TEQDFP-WHO98. On a body weight
16      basis, this corresponds to approximately 1 pg TEQDFP-WHO98/kg-day. Daily intake is estimated
17      by combining exposure media concentrations (food, soil,  air) with contact rates (ingestion,
18      inhalation). Table 4-7 summarizes the  intake rates derived by this method. The intake estimate
19      is supported by an extensive database on food consumption rates and food data. PK modeling
20      provides further support for the intake estimates. Current adult tissue levels reflect intakes from
21      past exposure levels, which are thought to be higher than  current levels".
22            CDD/CDF and dioxin-like PCS intakes for the general population may extend to levels at
23      least three times higher  than the mean.  Variability in general-population exposure is primarily a
24      result of differences in dietary choices that individuals make. These are differences in both
25      quantity and types of food consumed. A diet that is disproportionately high in animal fats will
26      result in an increased background exposure over the mean. Data on variability of fat
27      consumption indicate that the 95th percentile is about twice the mean and the 99th percentile is
28      approximately three times the mean. Additionally, a diet that substitutes meat sources that are
29      low in dioxin (i.e., beef, pork, or poultry) with sources that are high in dioxin (i.e., freshwater
30      fish) could result in elevated exposures.
31            Evidence of widespread background exposure can also be seen by examining data on
32      human tissue. These data indicate that on the average CDD/CDF tissue level for the general
33      adult United States population appears to be declining; the best estimate of current (mid to late
34      1990s) levels is 25 ppt (TEQDFP-WHO98, lipid basis). The tissue samples collected in North
35      America in the late 1980s and early 1990s showed an average TEQDFP-WHO9g level of about 55
36      pg/g lipid. This finding is supported by a number of studies, all conducted in North America,
37      that measured dioxin levels in adipose tissue, blood, and human milk.  The number of people in
        9/22/00
113
DRAFT—DO NOT CITE OR QUOTE

-------
  1      most of these studies, however, is relatively small and the participants were not statistically
  2      selected in ways that assured their representativeness of the general United States adult
  3      population. One study, the 1987 National Human Adipose Tissue Survey (NHATS), involved
  4      more than 800 individuals and provided broad geographic coverage, but did not address coplanar
  5      PCBs.  Similar tissue levels of these compounds have been measured in Europe and Japan during
  6      similar time periods.
  7             Because dioxin levels in the environment have been declining since the 1970s, it is
  8      reasonable to expect that levels in food, human intake, and ultimately human tissue have also
  9      declined over this period. The changes in tissue levels are likely to lag the decline seen in
10      environmental levels, and the changes in tissue levels cannot be assumed to occur proportionally
11      with declines in environmental levels. CDC (2000) summarized levels of CDDs, CDFs, and
12      PCBs in human blood collected during the time period 1995 to 1997. The individuals sampled
13      were all U.S. residents with no known exposures to dioxin other than normal background. The
14      blood was collected in seven different locations from 316 individuals with an age range of 20 to
15      70 years. All TEQ calculations were made assuming nondetects were equal to half the detection
16      limit. Although these samples were not collected in a manner that can be considered statistically
17      representative of the national population and lack wide geographic coverage, they are judged to
18      provide a better indication of current tissue levels in the United States than the earlier data (see
19      Table 4-6). PCBs  105,118, and 156 are missing from the blood data for-the comparison
20      populations reported by CDC (2000).  These congeners account for 62% of the total PCB TEQ
21      estimated in the early 1990s.  Assuming that the missing congeners from the CDC study data
22      contribute the same proportion to the total PCB TEQ as in earlier data, they would increase the
23      estimate of current body burdens by another 3.3 pg TEQ/g lipid for a total PCB TEQ of 5.3 pg/g
24      lipid and a total DFP TEQ of 25.4 pg/g lipid.
25             As noted, characterizing national background levels of dioxins in tissues is uncertain
26      because the current data cannot be considered statistically representative of the general
27      population. The task is also complicated by the fact that tissue levels are a function of both age
28      and birth year. Because intake levels have varied over time, the accumulation of dioxins in a
29      person who turned 50 in 1990 is different from that in a person who turned 50 in 2000. Future
30      studies should help address these uncertainties. The National Health and Nutrition Examination
31      Survey (NHANES) began a new national survey in 1999 that will measure dioxin blood levels in
32      about 1,700 people per year (see http:www.cdc.gov/nchs/nhanes.htm).  The survey is conducted
33      at 15 different locations per year and is designed to select individuals statistically representative
34      of the civilian U.S. population in terms of age, race, and ethnicity. These new data should
35      provide a much better basis than the currently available data for estimating national background
36      tissue levels and evaluating trends.
        9/22/00
114
DRAFT—DO NOT CITE OR QUOTE

-------
 1            As described above, current intake levels from food sources are estimated in this
 2      reassessment to be approximately 1 pg TEQ/kgBW/day. Certain segments of the population may
 3      be exposed to additional increments of exposure by being in proximity to point sources or
 4      because of dietary practices. These will be described below.
 5
 6      Evaluation of exposure of "special" populations and developmental stages is critical to risk
 7      characterization.
 8            As discussed above, background exposures to dioxin-like compounds may extend to
 9      levels at least three times higher than the mean. This upper range is assumed to result from the
10      normal variability of diet and human behaviors. Exposures from local elevated sources or unique
11      diets would be in addition to this background variability.  Such elevated exposures may occur in
12      small segments of the population, such as individuals living near discrete local sources, or
13      subsistence or recreational fishers. Nursing infants represent a special case where, for a limited
14      portion of their lives, these individuals may have elevated exposures on a body-weight basis
15      when compared to non-nursing infants and adults. This exposure will be discussed in a separate
16      section.
17            Dioxin contamination incidents involving the commercial food supply have occurred in
18      the United States and other countries. For example, in the United States, contaminated ball clay
19      was used as an anticaking agent in soybean meal and resulted in elevated-dioxin levels in some
20      poultry and catfish. This incident involved less than 5% of national poultry production and has
21      since been eliminated. Elevated dioxin levels have also been observed "in a few beef and dairy
22      animals where the contamination was associated with contact with pentachlorophenol-treated
23      wood. This kind of elevated exposure was not detected in the national beef survey.
24      Consequently, its occurrence is likely to be low, but it has not been determined. These incidents
25      may have led to small increases in dioxin exposure to the general population.  However, it is
26      unlikely that such incidents have led to disproportionate exposures to  populations living near
27      where these incidents have occurred, because in the United States, meat and dairy products are
28      highly distributed on a national scale. If contamination events were to occur in foods that are
29      predominantly distributed on a local or regional scale, then such events could lead to highly
30      exposed local populations.
31             Elevated exposures associated with the workplace or industrial accidents have also been
32      documented.  U.S. workers in certain segments of the chemical industry had elevated levels of
33      TCDD exposure, with some tissue measurements in the thousands of ppt TCDD.  There is no
34      clear evidence that elevated exposures are currently occurring among  U.S. workers. Documented
35      examples of past exposures for other groups include certain Air Force personnel exposed to
36      Agent Orange during the Vietnam War and people exposed as a result of industrial accidents in
37      Europe and Asia.
        9/22/00
115
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
32
33
34
35
36
37
       Consumption offish, meat, or dairy products containing elevated levels of dioxins and
 dioxin-like PCBs can lead to elevated exposures in comparison to the general population. Most
 people eat some fish from multiple sources, both fresh and salt water. The typical dioxin
 concentrations in these fish and the typical rates of consumption are included in the mean
 background calculation of exposure.  People who consume large quantities offish at typical
 contamination levels may have elevated exposures. These kinds of exposures are addressed
 within the estimates of variability of background and are not considered to result in highly
 exposed populations. If individuals obtain their fish from areas where the concentration of
 dioxin-like chemicals is elevated, they may constitute a highly exposed subpopulation. Although
 this scenario seems reasonable, very little supporting data could be found for such a highly
 exposed subpopulation in the United States. One study measuring dioxin-like compounds in
 blood of sports fishers in the Great Lakes area showed elevations over mean background, but
 within the range of normal variability. Another study measuring 90 PCB congeners, of which 7
 were dioxin-like mono-ortho PCBs (although PCB 126 was not measured), in Lake Michigan
 "sport-fish eaters" versus a control group (little or no sport fish consumption) showed a
 significant elevation in these PCBs. Significantly elevated concentrations of dioxins, furans, and
 coplanar PCBs were measured in Great Lakes fish by the Ontario Ministry of the Environment,
 although this was a study of known or suspected hot spots, with the purpose being to set
 consumption advisories.  It is not known to what extent individuals would be consuming fish at
 the high concentrations measured.  Elevated CDD/CDF levels in human blood have been
 measured  in Baltic fishermen. Similarly, elevated levels of coplanar PCBs have been measured
 in the blood of fishers on the north shore of the Gulf of the St. Lawrence River who consume
 large amounts of seafood.
       High exposures to dioxin-like chemicals as a result of consuming meat and dairy products
would most likely occur in situations where individuals consume large quantities of these foods
and the level of these compounds is elevated. Most people eat meat and dairy products from
multiple sources and, even if large quantities are consumed, they are not likely to have unusually
high exposures. Individuals who raise their own livestock for basic subsistence have the
potential for higher exposures if local levels of dioxin-like compounds are high. One study in the
United States showed elevated levels in chicken eggs near a contaminated soil site.  European
studies at several sites have shown elevated CDD/CDF levels in milk and other animal products
near combustion sources.
       In summary, in addition to general population exposure, some individuals or groups of
individuals may also be exposed to dioxin-like compounds from discrete sources or pathways
locally within their environment. Examples of these "special" exposures include contamination
incidents, occupational exposures, direct or indirect exposure to local populations from discrete
sources, or exposures to subsistence or recreational fishers.
        9/22/00
                                          116
DRAFT—DO NOT CITE OR QUOTE

-------
 1     Breast-feeding infants have higher intakes of dioxin and related compounds for a short but
 2     developmentally important part of their lives. However, the benefits of breast feeding are
 3     widely recognized to outweigh the risks.
 4            Three studies have compared dioxins in infants who have been breast-fed versus those
 5     who have been formula-fed, and all have shown elevations in the concentrations of dioxins in
 6     infants being breast-fed. Formula-fed infants had lipid-based concentrations < 5 ppt TEQDF-
 7     WHO9g whereas breast-fed infants had average lipid-based concentrations above 20 ppt TEQDF-
 8     WHO9g. The dose to the infant varies as a function of infant body weight, the concentration of
 9     dioxins in the mother's milk, and the trend of dioxins in the mother's milk to decline over time.
10     Using typical values for these parameters, dioxin intakes at birth were estimated to equal 242 pg
11     TEQDFP-WHO98/kg/day, which would drop to about 20 pg TEQDFP-WHO98/kg/day  after 12
12     months. The average dose over a year was calculated to be 92 pg TEQDFP-WHO98/kg/day.
13     Although this average annual infant dose of 92 pg TEQDFP-WHO98/kg/day exceeds the currently
14     estimated adult dose of 1 pg TEQDFp-WHO98/kg/day5 the effect on infant body burdens is
15     expected to be  less dramatic, i.e., infant body burdens will not exceed adult body burdens by 92
16     times. This is due to the rapidly expanding infant body weight and lipid volume, the decrease in
17     concentration of dioxins in the mother's milk over time, and possibly more rapid elimination in
18     infants.  A pharmacokinetic exercise comparing 6- and 12-month nursing scenarios with formula
19     feeding showed peak  infant lipid concentrations to exceed 40 ppt TEQ^p-WHOgg, compared
20     with peak lipid concentrations less than 10 ppt for the formula-fed infants and average adult lipid
21     concentrations of 25 ppt TEQDFP-WHO98. The dioxin concentrations in these two hypothetical
22     children merged at about 10 years of age, at a lipid concentration of about 13 ppt TEQDFP-
23     WHO98.
24            The American Academy of Pediatrics (1997) has made a compelling argument for the
25     diverse advantages of breast-feeding for infants, mother, families and society. These include
26     health, nutritional, immunologic, developmental, psychological, social, economic, and
27     environmental benefits. Breast milk is the point of comparison for all infant food, and the breast-
28     fed infant is the reference for evaluation of all alternative feeding methods. In addition,
29     increasing the rates of breast-feeding initiation is a national health objective  and one of the goals
30     of the United States Government's Healthy People 2010. WHO (1988) maintained that the
31     evidence did not support an alteration of WHO recommendations that promote and support
32     breast-feeding. A more recent consultation in 1998 (WHO, 2000) reiterated these conclusions.
33     Although it is important that the recommendations of these groups continue  to be  reevaluated in
34      light of emerging scientific information, the Agency does not believe that the finding contained
35      in this report provides a scientific basis for initiating such a reevaluation^ This conclusion is
36      based on the fact that stronger data have been presented that body burden, not intake, is the best
37      dose metric; that many of the noncancer effects, particularly those seen in children, are more
        9/22/00
117
DRAFT—DO NOT CITE OR QUOTE

-------
  1     strongly associated with prenatal exposure and the mother's body burden rather than postnatal
  2     exposures and breast milk levels; and that dioxin-like compounds are strong promoters of
  3     carcinogeniciry, a mode of action that depends on late-stage impacts rather than early-stage
  4     impacts on the carcinogenic process.
  5
  6     Many dioxin sources have been identified and emissions to the environment are being
  7     reduced.
  8            Current emissions of CDDs/CDFs/PCBs to the United States environment result
  9     principally from anthropogenic activities. Evidence that supports this finding includes matches
10     in time of the rise of environmental levels with rise in general industrial activity (see discussion
11     in Section 4.1), lack of any identified large natural sources and observations of higher
12     CDD/CDF/PCB body burdens in industrialized versus less industrialized countries (see
13     discussion on human tissue levels in Section 4.4).
14            The principal identified sources of environmental release may be grouped into five major
15     types: (1) combustion and incineration sources; (2) chemical manufacturing/processing sources;
16     (3) industrial/municipal processes; (4) biological and photochemical processes; and (5) reservoir
17     sources. Development of national estimates of annual environmental releases to air, water and
18     land is complicated by the fact that only a few facilities in most industrial sectors have been
19     evaluated for CDD/CDF emissions. Thus, an extrapolation is needed to^stimate national
20     emissions.  The extrapolation method involves deriving an estimate of emissions per unit of
21      activity (i.e., an emission factor) at the tested facilities and multiplying this by the total activity
22     level in the untested facilities. In order to convey the level of uncertainty in both the measure of
23     activity and the emission factor, EPA developed a qualitative confidence rating scheme.  The
24     confidence rating scheme, presented in Section 4, Table 4-1, uses qualitative criteria to assign a
25     high, medium, or low confidence rating to the emission factor and activity level for those source
26     categories for which emission estimates can be reliably quantified. The  dioxin reassessment has
27     produced an inventory of source releases for the United States (Table 4-2). The inventory is
28     limited to sources whose releases can be reliably quantified (i.e., those with confidence ratings of
29     A, B, or C as defined above). The inventory presents the environmental releases in terms of two
30     reference years: 1987 and 1995. For both of these periods, emissions from combustion and
31      incineration sources dominate total releases. EPA's best estimates of releases of CDD/CDFs to
32     air, water, and land from reasonably quantifiable sources were approximately 3,300 gram (g) (7
33     pounds) TEQDF-WHO98 in 1995 and 14,000 g (31 pounds) TEQDF-WHO98 in 1987. The decrease
34     in estimated releases of CDD/CDFs between 1987 and 1995 (approximately 76%) was due
35     primarily to reductions in air emissions from municipal and medical waste incinerators.
36            While this inventory is one of the most comprehensive and well-documented in the
37     world,  it is likely to underestimate total releases.  This underestimate is likely because: 1) a
        9/22/00                                   118      DRAFT—DO NOT CITE OR QUOTE

-------
  1      number of known sources lacked sufficient data to include in the inventory and 2) the possibility
  2      remains that truly unknown sources exist.
  3             Further reductions in environmental releases since the inventory for 1995 can be
  4      anticipated as a result of EPA regulations for waste combustion sources and pulp and paper
  5      facilities. EPA's regulatory programs estimate that, under full compliance with these regulations,
  6      an additional 1800 grams I-TEQ reduction in CDD/CDF emissions should occur.  With these
  7      anticipated emission reductions, uncontrolled burning of household waste would become the
  8      largest quantifiable source. Although the full magnitude of reservoir releases remain uncertain,
  9      their relative contribution to total annual releases be can reasonably anticipated to increase as
10      contemporary formation sources continue to decrease.
11             No significant release of newly formed dioxin-like PCBs is occurring in the United
12      States. Unlike CDD/CDFs, PCBs were intentionally manufactured in the United States in large
13      quantities from 1929 until production was banned in 1977. Although it has been demonstrated
14      that small quantities of coplanar PCBs can be produced during waste combustion, no strong
15      evidence exists that the dioxin-like PCBs-make a significant contribution to TEQ releases during
16      combustion.  The occurrences of dioxin-like PCBs in the U.S. environment most likely reflect
17      past releases associated with PCB production, use, and disposal. Further support for this finding
18      is based on observations of reductions since the 1980s in PCBs in Great Lakes sediment and
19      other areas.                                                     ,...-.
20             As described in Section 4.1, combustion appears to be the most significant process of
21      formation of CDDs/CDDFs today.  Important factors that can affect the" rate of dioxin formation
22      include the overall combustion efficiency, post-combustion flue gas temperatures and residence
23      times, and the availability of surface catalytic sites to support dioxin synthesis.  Although
24      chlorine is an essential component for the formation of CDD/CDFs in combustion systems, the
25      empirical evidence indicates that, for commercial-scale  incinerators, chlorine levels in feed are
26      not the dominant controlling factor for rates of CDD/CDF stack emissions. The conclusion that
27      chlorine in feed is not a strong determinant of dioxin emissions applies to the overall population
28      of commercial scale combustors. For any individual commercial-scale combustor, circumstances
29      may exist in which changes in chlorine content of feed could affect dioxin emissions. For
30      uncontrolled combustion, such as open burning of household waste, chlorine content of wastes
31      may play a more significant role in affecting levels of dioxin emissions than observed in
32      commercial-scale combustors.
33
34      Dioxins are widely distributed in the environment at low concentrations, primarily as a
35      result of air transport and deposition.
36             Once introduced into the environment, dioxin-like compounds are widely distributed in
37      the environment as a result of a number of physical and biological processes.  The dioxin-like
        9/22/00
119
DRAFT—DO NOT CITE OR QUOTE

-------
 1      compounds are essentially insoluble in water, generally classified as semivolatile, and tend to
 2      bioaccumulate in animals. Some evidence has shown that these compounds can degrade in the
 3      environment, but in general they are considered very persistent and relatively immobile in soils
 4      and sediments. These compounds are transported through the atmosphere, as vapors or attached
 5      to airborne particulates and can be deposited on soils, plants, or other surfaces (by wet or dry
 6      deposition). The dioxin-like compounds enter water bodies primarily via direct deposition from
 7      the atmosphere, or by surface runoff and erosion. From soils, these compounds can reenter the
 8      atmosphere either as resuspended soil particles or as vapors.  In water, they can be resuspended
 9      into the water column from sediments, volatilized out of the surface waters into the atmosphere,
10      or become buried in deeper sediments. Immobile sediments appear to serve as permanent sinks
11      for the dioxin-like  compounds. Though not always considered an environmental compartment,
12      these compounds are also found in anthropogenic materials (such as pentachlorophenol) and
13      have the potential to be released from these materials into the broader environment.
14            The two primary pathways for the dioxin-like compounds to enter the ecological food
15      chains and human diet are air-to-plant-to-animal and water/sediment-to-fish. Vegetation receives
16      these compounds via atmospheric deposition in the vapor and particle phases.  The compounds
17      are retained on plant surfaces and bioaccumulated in the fatty tissues of animals that feed on
18      these plants.  In the aquatic food chain, dioxins enter water systems via direct discharge or
19      deposition and runoff from watersheds.  Fish accumulate these compounds .through direct  contact
20      with water, suspended particles, and bottom sediments and through the consumption of aquatic
21      organisms. Although these two pathways are thought to normally dominate contribution to the
22      commercial food supply, others can also be important. Animal feed contamination episodes have
23      led to elevations of dioxins in poultry in the United States, milk in Germany, and meat/dairy
24      products in Belgium. Gaining a quantitative understanding of how dioxin moves in the
25      environment will be particularly important in understanding the relative contributions of
26      individual point sources to the food chain and assessing the effectiveness of control strategies to
27      reduce human exposure. Although the emissions inventory shows the relative  contribution of
28      various sources to  total emissions, it is unlikely that these sources make the same relative
29      contributions to human exposure.
30            It is quite possible that the major contributors of dioxin to food (see discussion in Section
31      4.4 indicating that the diet is the dominant exposure pathway for humans) may not be those
32      sources that represent the largest fractions of total emissions in the United States. The
33      geographic locations of sources relative to the areas from which much of the beef, pork, milk,
34      and fish are produced are important to consider.  Most of the agricultural areas that produce
35      dietary animal fats are not located near or directly downwind of the major sources of dioxin and
36      related compounds.
        9/22/00
120
DRAFT—DO NOT CITE OR QUOTE

-------
 1       .     The contribution of reservoir sources to human exposure is likely to be significant.
 2      Several factors support this finding.  First, human exposure to the dioxin-like PCBs is thought to
 3      be derived almost completely from reservoir sources.  Because one-third of general population
 4      TEQ exposure is due to PCBs, at least one-third of the overall risk from dioxin-like compounds
 5      comes from reservoir sources. Second, CDD/CDF releases from soil via soil erosion and runoff
 6      to waterways appear to be greater than releases to water from the primary sources included in the
 7      inventory. CDD/CDFs in waterways can bioaccumulate in fish-leading to human exposure via
 8      consumption offish. This suggests that a significant portion of the CDD/CDF TEQ exposure
 9      could be due to releases from the soil reservoir.  Finally, soil reservoirs could have vapor and
10      particulate releases that deposit on plants and enter the terrestrial food chain. The magnitude of
11      this contribution, however, is unknown. Collectively, these three factors suggest that reservoirs
12      are a significant source of current background TEQ exposure, perhaps contributing half or more
13      of the total.
14                                                              .              •           .
15      Environmental levels, emissions and human exposures have declined during recent
16      decades.
17            The most compelling supportive evidence of a general decline in environmental levels for
18      CDD/CDFs and PCBs comes from dated sediment core studies.  CDD/CDF and PCB
19      concentrations in sediments began to increase around the 1930s and continued to increase until
20      about 1970. Decreases began in 1970 and have continued to the time of the most recent sediment
21      samples (about 1990). Additionally, sediment studies in lakes located in several European
22      countries have shown similar trends.
23            It is reasonable to assume that sediment core trends should be driven by a similar trend in
24      emissions to the environment. The period of increase  generally matches the time when a variety
25      of industrial activities began rising, and the period of decline appears to correspond with growth
26      in pollution abatement. Many of these abatement efforts should have resulted in decreases in
27      dioxin emissions, i.e., elimination of most open burning, particulate controls on combustors,
28      phase out of leaded gas, and bans on PCBs, 2,4,5-T, hexachlorophene, and restrictions on use of
29      pentachlorophenol.  Also, the national source inventory of this assessment documented a
30      significant decline in emissions from the late 1980s to the mid-1990s.
31            Evidence of declines  in human exposure can be inferred from overall declines in
32      environmental levels and emissions.  Also, it is directly supported by limited data on
33      concentrations in food and human tissues (see Sections 4.3 and 4.4). Because of the lag in
34      environmental levels and body burdens, it is  anticipated that further declines in tissue
35      concentrations should occur.
36
        9/22/00
121
DRAFT—DO NOT CITE OR QUOTE

-------
  1      Risk Characterization Summary Statement
  2             233,7,8-Tetrachlorodibenzo-/7-dioxin (dioxin) is highly toxic to many animal species
  3      producing a variety of cancer and noncancer effects. Other 2,3,7,8-substituted polychlorinated
  4      dibenzo-_p-dioxins and dibenzofurans, and coplanar polychlorinated biphenyls (PCBs), exhibit
  5      similar effects albeit at different doses and with different degrees of confidence in the database.
  6      The similarities in toxicity between species  and across different dioxin congeners stem from a
  7      common mode of action via initial binding to the aryl hydrocarbon (Ah) receptor. This common
  8      mode of action is supported by consistency  in effects evident from multiple congener databases.
  9      This has led to an international scientific consensus that it is prudent science policy to use the
10      concept of toxic equivalency factors (TEFs) to sum the contributions of individual PCDD,
11      PCDF, and coplanar PCB congeners with dioxin-like activity. The databases supportive of
12      dioxin-like toxicity, both cancer and noncancer, are strongest for those congeners that are the
13      major contributors to the risk to human populations. In addressing receptor-mediated responses
14      resulting from complex mixtures of dioxin-like congeners, this assessment has provided a basis
15      for the use of integrated measures of dose, such as average body burden, as more appropriate
16      default metrics than daily intake.  The Agency recognizes, however, that the final choice of the
17      appropriate metric may depend on the endpoint under evaluation.
18             Dioxin and related compounds have been shown in multiple animal species to be
19      carcinogenic, developmental, reproductive,  immunological and endocrinological hazards, among
20      others. There is no reason to expect, in general, that humans would not be similarly affected at
21      some dose, and indeed there is a growing body of data supporting this assumption. Based upon
22      the animal data, current margins of exposure are too low, especially for more highly exposed
23      human populations.  The human database supporting this concern is less certain. Occupational
24      and accidentally exposed cohorts exposed at higher levels show correlations with exposure for a
25      number of cancer and noncancer effects, consistent with those seen in the  animal studies.
26             For cancer outcomes, the epidemiological evidence provides consistent findings of
27      statistically significant elevations  and dose-response trends for all-cancers combined and lung
28      cancer risk hi occupational cohorts, along with evidence of possible additional tissue-specific
29      cancer rate elevations. Given this substantial, yet still not definitive, epidemiological data; the
30      positive cancer bioassays  at multiple sites and in all animal species tested; and mechanistic
31      considerations common to animals and humans for dioxin carcinogenicity, EPA characterizes
32      2,3,7,8-tetrachlorodibenzo-£>-dioxin as "carcinogenic to humans." Complex mixtures of dioxin
33      and related compounds are considered highly potent, "likely" carcinogens. The calculated body
34      burdens of dioxin and dioxin-like substances leading to an estimated one percent increase (ED01)
35      in the lifetime risk of cancer all fall within a 10-fold range when comparing the occupational
36      studies, and are the same as those calculated based on the animal bioassay data. The ED01 for all-
37      cancers  combined from a metaanalysis of the three major occupational cohorts is 47
        9/22/00
122
DRAFT—DO NOT CITE OR QUOTE

-------
  1      ngTCDD/kgBW, with a lower confidence limit of 30 ngTCDD/kgBW.  By comparison, current
  2      background body burdens in the United States are approximately 5 ngTEQ/kgBW. Using 30
  3      ngTEQ/kgBW as the point of departure for the slope calculation, EPA calculates an upper bound
  4      on the lifetime risk of all cancers combined of 1 x 10"3 risk/pgTEQ/kg/day. This cancer slope
  5      factor is based on a statistical estimate of risks from occupational exposures, principally to
  6      healthy, adult, male workers, and must be coupled with a recognition that a small number of
  7      people may be both more susceptible and consume up to three times the average level of fat per
  8      day (the principal exposure pathway for dioxins in the general population). Using best available
  9      estimates of cancer risks, the upper bound on general population lifetime risk for all cancers
10      might be on the order of 1 in 1,000 or more.  Upper bound risk estimates allow the calculation of
11      the high end of the probability of cancer risk in the population. This means that there is greater
12      than a 95% chance that cancer risks will be less than the upper bound and could be as low as zero
13      in some individuals.
14            For noncancer effects, EPA generally calculates an RfD/RfC value which represents an
15      estimate (with uncertainty spanning perhaps  an order of magnitude) of a daily exposure to the
16      human population (including sensitive subgroups) that is likely to be without an appreciable risk
17      of deleterious effects during a lifetime.  The current estimated average dose to the U.S.
18      population (~1 pgTEQ/kg/day) is greater than RfD/RfC dose values that would be calculated
19      given the data reviewed in this assessment, and, therefore, RfD/RfC values would be
20      uninformative for safety assessment. EPA has chosen rather to characterize the margins-of-
21      exposure (MOE) for noncancer endpoints in order to inform risk management decisions. MOE is
22      the ratio of the human body burden to the effect level in the comparison species (ED01 or low
23      effect level), animal or human. For the most sensitive endpoints identified, MOE's range from,
24      for example, less than one for enzyme induction in mice, through 2.6-15 for enzyme induction
25      in rats, <3 for developmental effects, and 5 for endometriosis in non-human primates. In
26      evaluating MOEs, consideration should be given to  uncertainties in distinguishing between
27      adaptive biochemical changes and adverse effects, both on an individual level and as these
28      changes impact whole populations. Children's risks from dioxin and related compounds may be
29      greater than for adults, but more data are needed to fully address this issue.
30            Releases of dioxins to the environment from sources that have been characterized have
31      decreased significantly over the last decade and are  expected to continue to decrease.  Other
32      sources are still poorly characterized, and an environmental reservoir of dioxins from both man-
33      made and natural sources has been recognized. Human body burdens have also declined, but
34      their relationship to contemporary sources or reservoirs is uncertain.
35
36
        9/22/00
123
DRAFT—DO NOT CITE OR QUOTE

-------
       Table 1-1. The TEF scheme for I-TEQDFa
Dioxin (D) congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
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-HpCDD
1,2,3,4,6,7,8,9-OCDD

TEF
1.0
0.5
0.1
0.1
0.1
0.01
0.001

Furan (F) congener
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
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8,9-OCDF
TEF
0.1
0.05
0.5
0.1
0.1
0.1
0.1
0.01
0.01
0.001
"Note that the scheme does not include dioxin-like PCBs. The nomenclature for this scheme is I-TEQDF, where T
represents "International," TEQ represents the 2,3,7,8-TCDD toxic equivalence of the mixture, and the subscript DF
indicates that only dioxins (Ds) and furans (Fs) are included in the TEF scheme.
9/22/00
124
DRAFT—DO NOT CITE OR QUOTE

-------
        Table 1-2. The TEF scheme for TEQ
                                                DFP'
   -WHO,a
                                                         /94
Dioxin (D) congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
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-HpCDD
1,2,3,4,6,7,8,9-OCDD



TEF
1.0
0.5
0.1
0.1
0.1
0.01
0.001



Furan (F) congener
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
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,,6,7,8,9-OCDF

TEF
0.1
0.05
0.5
0.1
0.1
0.1
0 1
0.01
0.01
0.001

Dioxin-Iike
PCB(P)
PCB-77
PCB-126
PCB-169
PCB-105
PCB-118
PCB-123
PCB-156
PCB-157
PCB-167
PCB-114
PCB-170
PCB-180
PCB-189
TEF
0.0005
0.1
0.01
0 0001
00001
00001
00005
00005
0.00001
0.0005
0.0001
0.00001
0.0001
"The nomenclature for this TEF scheme is TEQDFP-WHO94, where TEQ represents the 2,3,7,8-TCDD toxic equivalency
of the mixture, and the subscript DFP indicates that dioxins (Ds), furans (Fs), and dioxin-like PCBs (P) are included in
the TEF scheme.  The subscript 94 following WHO displays the year changes were made to the TEF scheme.
9/22/00
125
DRAFT—DO NOT CITE OR QUOTE

-------
       Table 1-3. The TEF scheme for TEQDFP-WHO98a
Dioxin (D) congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
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-HpCDD
1,2,3,4,6,7,8,9-OCDD





TEF
1.0
1.0
0.1
0.1
0.1
0.01
0.0001





Furan (F) congener
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
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
1,2,3,4,6,7,8,9-
OCDF

TEF
0.1
0.05
0.5
0.1
0.1
0.1
0.1
0.01
0.01
0.0001


Dioxin-
like PCB
(P)
PCB-77
PCB-81
PCB-126
PCB- 169
PCB- 105
PCB-118
PCB- 123
PCB-156
PCB- 157
PCB- 167
PCB-114
PCB-189
TEF
0.0001
0.0001
0.1
0.01
0.0001
0.0001
0.0001
0.0005
0.0005
0.00001
0.0005
0.0001
"The nomenclature for this TEF scheme is TEQDFP-WHO98, where TEQ represents the 2,3,7,8-TCDD toxic
equivalency of the mixture, and the subscript DFP indicates that dioxins (Ds), furans (Fs), and dioxin-like PCBs (P)
are included in the TEF scheme. The subscript 98 following WHO displays the year changes were made to the TEF
scheme. Note that the changes to the TEFs since 1994 are as follows:

        •For 1,2,3,7,8-PeCDD, the new WHO TEF is 1 and the I-TEF is 0.5;      ..--
        •For OCDD, the new WHO TEF is 0.0001 and the I-TEF is 0.001;
        •For OCDF, the new WHO TEF is 0.0001 and the I-TEF is 0.001;
        • For PCB 77, the new TEF is 0.0001;
        •The addition of PCB 81 (i.e., 3,4,4',5-TCB); and
        •For the two di-ortho substituted HpCBs in the 1994 TEF scheme (i.e., PCBs 170 and 180), no TEFs have
        been assisned in the new WHO TEF scheme.
9/22/00
126
DRAFT—DO NOT CITE OR QUOTE

-------
      Table 1-4. The range of the in vivo REP values for the major TEQ contributors
• -
CHEMICAL


1,2,3,7,8-
PCDD
2,3,4,7,8-
PCDF
1,2,3,6,7,8-
HxCDD
PCB 126

Number
of
in vivo
endpoints
22

40

3

62

Range of
,REPs -
(mean±std)

0.16-0.9
(0.5±0.22)
0.018-4.0
(0.4±0.7)
0.015-0.16

0.0024-0.98
(0.20±0.20)
Number of
end points from
subchronic
studies
16

20

1

31

Range of
REPs
(raean±std)

0.19-0.9
(0.53±0.24)
0.018-0.6
(0.20±0.13)
0.04

0.004-0.18
(0.13±0.13)

TEF.


1

0.5

0.1

0.1

9/22/00
127
DRAFT—DO NOT CITE OR QUOTE

-------
       Table 1-5. Comparison of administered dose and body burden in rats and humans






Dose
(ng/kg/d)
Body Burden
(ng/kg)
(A)
Rat Daily
Administered
Dose/Body
Burden

1

18

(B)
Human
Scaled
Administered
Dose/Body
Burden1
0.27

505

(Q
Human
Equivalent
Administered
Dose/Body
Burden2
0.0096

18

(A/B)
Ratio of
Rat to
Human
Scaled
Dose
3.7

0.036

(A/C)
Ratio of
Rat to
Human
Equivalent
Dose
104

1

       1 Assumes administered dose scales across species as a function of BW3/4
       2 Assumes administered dose scales across species as a function of equivalent body burdens
9/22/00
128
DRAFT—DO NOT CITE OR QUOTE

-------
 OS
.2
'3
 es
 I
73
 a
 o
 o
 U
"Q
73
 ft
A
Q
U
H
H
e
§
V.
|i
§1
II
j=
c
Ic
U
IS
o
u
-8
en
£
3
O
s
a;
CS
" b.
'=£
i
B
O
§
uman
B
•4-*


;


-
-
+

-
-




+


o

*f~


+

Presence of AhR






-
+

+
+




+


+




+

Binding of TCDD:
AhR Complex to
he ORE (enhancer)


-


-
+
+

+





-




-(-


+

Enzyme induction


-


-
-
+

+
+




^




-}-


0

Acute lethality


-


-
-


-
-




-




4*




Wasting syndrome


-


-
-
+

-





-




~\~


~L.

Teratogenesis/fetal
toxicity, mortality


^


+
+








-




-{-


<

Endocrine effects






+
+


-




-


+

-(-


<

Immunotoxicity






+








+







j.

Carcinogenicity







+







-




-f-


+

Neurotoxicity






-


-
-









_)-.


+

Cllloracnegenic
effects







+




O


^


O

o


+

Porphyria


+


+
-
+

-
+

+


^


^

_J_


-f

Hepatotoxicity






• +
+







O


O

-j_




eC
1
S















^


+

_j_




Testicular atrophy







+



•-.'

^




+

•4-




Bone marrow
hypoplasia

















.

§
s
s
g
observed.
= observed to limited
not observed.
nk cells = no data.
u .1 u ra
+ + o m
9/22/00
129
DRAFT—DO NOT CITE OR QUOTE

-------
               Table 2-2.  Examples of margins of exposure (M-O-E)
Effect
CYP1A1/1A2/1B1
(Rats)
Tumors;Multiple sites
(Rats)
Endometriosis
(Rhesus Monkey)
Developmental Effects
(Humans)
Tumors; All/lung
(Humans)
ED01 or
Low Effect Level
13-74 ng/kg
(ED01)
14 -1190 ng/kg
(EDOJ)
38ng/kg
(LOEL)
Dutch background; early '90s
6-250 ng/kg
(ED01)
M-O-E 1
2.6-15
3-238
5
<3
1.2-50
      1 MOE =     EDni or Low Effect Level
               Current Human Body Burden (~ 5ng TEQDFP-WHO98ng/kg)
9/22/00
130
DRAFT—DO NOT CITE OR QUOTE

-------
       Table 2-3.  Summary of the combined cohort and selected industrial cohort studies
       with high exposure levels as described by IARC,  1997a
REFERENCE -
ALL CANCERS
Obs.
SMR
95% C.I.
LUNG CANCER
Obs.
SMR
95% C.I.
International cohort
Kogevinas et al.
(1997)b
394
1.2
1.1-1.3
127
1.2
1.0- 1.4
Industrial populations (high-exposure subcohorts)
Fingerhut et al.
(1991a)c
(USA)
Becher et al.
(1996)d
(Germany)
Hooiveld et al.
(1996)e
(Netherlands)
Ott & Zober
(1996b)f
(BASF accident)
TOTAL8
p value
114
105
51
18
[288]
1.5
[1.3]
1.5
1.9
[1.4]
1.2-1.8
[1.0-1.5]
1.1-1.9
1.1-3.0
[1.2-1.6]
<0.001
40
33 .
14
7
[94]
1.4
[1.4]
1
2.4
[1.4]
1.0-1.9
[1.0-2.0]
0.5- 1.7
1.0-5.0
[1.1-1.7]
<0.01
       1 Adapted from IARC; Table 38 (1997); Non-Hodgkin lymphoma, soft-tissue sarcoma, and gastrointestinal
        results not shown.
       b Kogevinas et al. (1997): Men and woman >20 years since first exposure. These data include the cohorts
        of Fingerhut et al. (1991a,b), Becher et al. (1996), Hooiveld et al. (1996a), the original IARC cohort
        (Saracci et al., 1991) and other cohorts.
       c Fingerhut et al. (1991a): Men ;>20 years latency and a 1 year exposure.
       d Becher et al. (1996): Men, Cohort I and II, summed (Boehringer-Ingelheirn, Bayer-Uerdingen cohorts).
       e Hooiveld et al. (1996): Men and women, Factory A.
       f Ott & Zober (1996b): Men, chloracne subgroup, a 20 years latency. Data presented for lung cancer
        are all respiratory tract cancers combined.
       8 TOTALS in square brackets are those calculated by the IARC Working Group.
9/22/00
131
DRAFT—DO NOT CITE OR QUOTE

-------
 Table 2-4. Tumor Incidence and Promotion Data Cited for the TEF-WHO98 for Principal
 Congeners
Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
2,3,4,7,8-PeCDF
1,2,3,6,7,8-HxCDD

1,2,3,7,8,9-HxCDD
PCB 126
TEF-WHO98 Tumor
Incidence/Promotion
Citation1
TEF Standard
Waernetal., 1991
Waernetal., 1991
NTP 1980;
1,2,3,6,7,8-HxCDD/
1,2,3,7,8,9-HxCDD;
1 :2 mixture;
long term bioassays,
Osbome-Mendel rats
in NTP studies,
Sprague-Dawley rats
in Kociba et al.,
1978
Hemming et al.,
1995
TEF-
WHO
98
1
1
0.5
0.1
0.1
0.1
% of Adipose
TEQDFP-WH098
Tissue Cone.2
8
15
7
10
2
33
Dose-Response Graphs: Dose adjusted to
reflect.TEF multiplier
Percent foci tissue In liver
0000 -.-...-.
0.8
0.7
g 0.6.
|- 0.5
Z 0.4
2 0.3
5
1 °i2
0.1 ,
0
c
llvertlssue occupied by GGT+vc foci
o -•• ro w .(-. 01 05
I — 1H 	 ' 	 ' 	 ' 	 ' 	 '
55 -
0
t TCDD
—i_TCOD outlier ,--°
-<,-_PeCDDx1.0 ,-''
...^..PeCDFxO.5 ,-'''
: ' 'y^L... 	 «
/••'
0 OJS 0.5 0.75 1 1.25 1.S
TCDD-equivalents (ug/kg/week)
... j, ... TCDD, Kociba orig. pathology
-- « — TCDD, Kooibarev. pathology
— o— TCDD, NTP 1982 y ..-•*
_X-HexaCDDx0.1, NTP 1980 /.....•--"'
•-•;"/''
..-••"""' ''
..-•••"" /' _-•
l^-^^ —
^-^^
0.02 0.04 0.06 0.08 0 1
TCDD-cqulvitents (ug;kg/day)
-i—TCDD
...tt.. PCB 126x0.1 ,-'''
—^.TCDD+PCB 125x0.1 ,-'''
^""*
,&?..f*f?&''
0.5 1 1.5 2
TCDC^equh-alents (u-yk-jVM*-)
1. van den Berg, et al. 2000. Human risk assessment and TEFs. Food Additives and Contaminants, Vol. 17(4):347 -
358. Hexa-CDD referenced to previous TEF reviews.
2. See Part II, Chapter 4, Tables 4-46,4-47
9/22/00
132
DRAFT—DO NOT CITE OR QUOTE

-------
                 Table 3-1. Early molecular events in response to dioxin
            Diffusion into the cell
            Binding to the AhR protein
            Dissociation from hsp90
            Active translocation from cytoplasm to nucleus
            Association with Arnt protein
            Conversion of liganded receptor to the DNA-binding form
            Binding of liganded receptor heteromer to enhancer DNA
            Enhancer activation
            Altered DNA configuration
            Histone modification
            Recruitment of additional proteins
            Nucleosome disruption
            Increased accessibility of transcriptional promoter
            Binding of transcription factors to promoter
            Enhanced mRNA and protein synthesis
       These events are discussed in detail in Part II, Chapter 2.
9/22/00
133
DRAFT—DO NOT CITE OR QUOTE

-------
       Table 4-1. Confidence rating scheme
Confidence
category
Confidence
rating
Activity level estimate
Emission factor estimate
Categories/media for which emissions can be reasonably quantified
A
B
C
High
Medium
Low
Derived from comprehensive
survey
Based on estimates of average
plant activity level and number of
plants or limited survey
Based on data judged possibly
nonrepresentative.
Derived from comprehensive survey
Derived from testing at a limited but
reasonable number of facilities
believed to be representative of
source category
Derived from testing at only a few,
possibly nonrepresentative facilities
or from similar source categories
Categories/media for which emissions cannot be reasonably quantified
D
E
Preliminary
Estimate
Not Quantified
Based on extremely limited data,
judged to be clearly
nonrepresentative.
No data.
Based on extremely limited data,
judged to be clearly
nonrepresentative.
1) Argument based on theory but no
data
2) Data indicating dioxin formation,
but not in a form that allows
developing an emission factor
9/22/00
134
DRAFT—DO NOT CITE OR QUOTE

-------
Table 4-2. Quantitative inventory of environmental releases of TEQDF-WHO98 in the
United States
Emission source category
Confidence rating"
Reference year 1995
A
B
C
Confidence rating1
Reference year 1987
A
B
C
Releases (g TEQnr-WHO,x/yr) toAlr
Waste Incineration
Municipal waste incineration
Hazardous waste incineration
Boilers/industrial furnaces
Medical waste/pathological incineration
Crematoria
Sewage sludge incineration
Tire combustion
Pulp and paper mill sludge incineratorsr
Power/Energy Generation
Vehicle fuel combustion - leadedb
- unleaded
- diesel
Wood combustion - residential
- industrial
Coal combustion - utility
Oil combustion - industrial/utility
Other High Temperature Sources
Cement kilns (hazardous waste burning)
Lightweight aggregate kilns burning hazardous waste
Cement kilns (nonhazardous waste burning)
Petroleum refining catalyst regeneration
Cigarette combustion
Carbon reactivation furnaces
Kraft recovery boilers
Minimally Controlled or Uncontrolled Combustion
Forest, brush, and straw firesd
Metallurgical Processes
Ferrous metal smelting/refining
- Sintering plants
Nonferrous metal smelting/refining
- Primary copper
- Secondary aluminum
- Secondary copper
- Secondary lead
Drum and barrel reclamation
Chemical Manufac./Processing Sources
Ethylene dichloride/vinyl chloride
Total quantified releases to airc
































1250
5.8



14.8






27.6
60.1







2.3


28

<0.5C


1.72

11.2


0.39
488
9.1

0.11

2
5.9
35.5
62.8


10.7
156.1
3.3
17.8
2.21
0.8
0.08

208




29.1
271

0.08

2705

















•-'














8877
5



6.1






26.4
50.8







2




<0.5C


1.29




0.78
2590
5.5

0.11

37.5
3.6
27.8
89.6


17.8
117.8
2.4
13.7
2.24
1
0.06

170

32.7


16.3
983

0.08

13081
9/22/00
135
DRAFT—DO NOT CITE OR QUOTE

-------
                              Table 4-2.  Quantitative inventory of environmental releases of TEQDF-WHO98 in
                              the United States (continued)
Emission source category
Confidence rating"
Reference year 1995
A
B
C
Confidence rating3
Reference year 1987
A
B
C
Releases (g TEQ/yr) to water
Chemical Manuf./Processing Sources
Bleached chemical wood pulp and paper mills
Ethylene dichloride/vinyl chloride
Total quantified releases to water1
19.5


0.43


19.93 .
356





356
Releases (g TEQ/yr) to land
Chemical Manuf./Processing Sources
Bleached chemical wood pulp and paper mill
sludge
Ethlyene dichloride/vinyl chloride
Municipal wastewater treatment sludge
Commercially marketed sewage sludge
2,4-Dichlorophenoxy acetic acid
Total quantified releases to land0
Overall quantified releases to the open and
circulating environment
1.4

76.6
2.6
28.9

0.73








110.23
2835
14.1

76.6
2.6
33.4










126.7
13564
                   Confidence Rating A = Characterization of the Source Category judged to be Adequate for Quantitative Estimation with High Confidence in the Emission Factor
                   and High Confidence in Activity Level.
                   Confidence Rating B = Characterization of the Source Category judged to be Adequate for Quantitative Estimation with Medium Confidence in the Emission
                   Factor and at least Medium Confidence in Activity Level.
                   Confidence Rating C = Characterization of the Source Category judged to be Adequate for Quantitative Estimation with Low Confidence in either the Emission
                   Factor and/or the Activity Level.

                   •A confidence rating reflects EPA's judgment as to the adequacy of information pertaining to the emission factor and activity level.
                   *Lcadcd fuel  production and the manufacture of motor vehicle engines requiring leaded fuel for highway use have been prohibited in the United States, (see Section 4.1
                   for details,)
                   TOTAL reflects only the total of the estimates made in this report.
                   'It is not known what fraction, if any, of the estimated emissions from forest fires represents a "reservoir" source.  The estimated emissions may be solely the result of
                   combustion,
                   •Congener-specific emissions data were not available; the I-TEQDF emission estimate was used as a surrogate for the TEQDr-WHOs! emission estimate.
                   'Included within estimate for Wood Combustion - Industrial.
-
                   9/22/00
136
DRAFT—DO NOT CITE OR QUOTE

-------
   Table 4-3. Preliminary indication of the potential magnitude of TEQDF-WHO98 releases
   from "unquantified" (i.e., Category D) sources in reference year 1995
Emission source category
/. Contemporary Formation Sources
Biogas Combustion
Oil Combustion-Residential
Coal Combustion - Commercial/Industrial
Coal Combustion - Residential
Asphalt Mixing Plants
Combustion of Landfill Gas
Landfill Fires
Accidental Fires (Structural)
Accidental Fires (Vehicles)
Forest and Brush Fires
Primary Magnesium Production
Coke Production
Electric Arc Ferrous Furnaces
Ferrous Foundries
Municipal Wastewater
//. Reservoir Sources
Urban Runoff
Rural Soil Erosion
Release medium
Air
Air
Air
Air
Air
Air
Air
Air
Air
Air
Air
Air
Air
Air
Water
Water
Water
Preliminary release estimate
(gWHCVTECWyr)
0.22*
6.0*
39.6'
32.0'
7'
6.6
1,050*
>20'
28.3*
208
11.4*
6'.9'
44.3a
-.- '17.5'
12
190*
2,700*
* Congener-specific emissions data were not available; the I-TEQDF emission factor was used as a surrogate for the TEQDF-
WHO98 emission estimate.
9/22/00
137
DRAFT—DO NOT CITE OR QUOTE

-------
r
                    Table 4-4. Sources that are currently unquantifiable * (i.e., Category E)
Category
Combustion sources
Metal smelting and refining
Chemical manufacturing
Biological and photochemical processes
Reservoir sources
Unquantified sources
Uncontrolled combustion of PCBs
Agricultural burning
Primary aluminum
Primary nickel
Mono- to tetrachlorophenols
Pentachlorophenol
Chlorobenzenes
Chlorobiphenyls (leaks/spills)
Dioxazine dyes and pigments
2,4-Dichlorophenoxy acetic acid
Tall oil-based liquid soaps
Composting
Air
Sediments
Water
Biota
PCP-treated wood
                 '/There exist no or insufficient data characterizing environmental releases from these sources. Therefore, it is currently not
                 possible to arrive at an estimate of annual environmental releases .
                 9/22/00
138
DRAFT—DO NOT CITE OR QUOTE

-------
Table 4-5. Summary of North American CDD/CDF and PCB TEQ-WHO98 Levels in
Environmental Media and Food (whole weight basis; concentrations provided in
parenthesis for food products are calculated at ND = 0).
Media
Urban Soil, ppt
Rural Soil, ppt
Sediment, ppt
Urban Air, pg/m3
Rural Air, pg/m3
Freshwater Fish and Shellfish, ppt
Marine Fish and Shellfish, ppt
Water, ppq
Milk, ppt
(Note: each composite for
CDD/F/PCB comprised of 40+
U.S. regional samples)
Dairy, ppt°
Eggs, ppt
(Note: each composite for CDD/F
data comprised of 24 eggs)
Beef ppt
Pork, ppt
Poultry, ppt
Vegetable Fats, ppt
CDD/CDFsa
n=171
9.4 ±11.2
Range = 2 -21
n = 292
2.5
Range = 0.1 -6
n=ll
5.3 ± 5.8
Range = <1 - 20
n=106
0.12 ±0.094
Range = 0.03 - 0.2
n=7
0.017
Range = 0.01 -0.02
n=289
1.0 (NAb)
n=158
0.26 (NAb)
n=236
0.00056 ± 0.00079 (NAb)
n=8 composites
0.031 ±0.0022 (0.031)
n = 8 composites
0.12 ±0.22 (0.12)
n=15 composites
0.081° (0.013)
n=63
0.20 ±0.12 (0.07)
Range = 0.2- 1.1
n=78
0.22 ± 0.22 (0.06)
Range = 0.12- 1.4
n=78
0.12 ±0.12 (0.072)
Range = 0.05 - 0.72
n=30
0.056 ± 0.24d (NAb)
PCBs"
NAb
NAb
n=ll
0.53 ± 0.69
0.0009
NAb
n = 1 composite of 10 samples plus 6
composites
1.2° (NAb)
n = 1 composite of 13 samples plus 5
composites
0.25° (NAb)
NAb
n = 8 composites
• 0.016(0.016)
n = 8 composites'
0.058 (0.058)
n = 18 plus 6 composites
0.10°(NAb)
n = 63
0.094 (0.094)
n = 78
0.0093 (0.006)
n = 78
0.044 (0.044)
n = 5 composites
0.037°
    Values are the arithmetic mean TEQs, in ppt, and standard deviations. Nondetects were set to one-half the limit of
    detection, except for soil and CDD/CDFs in vegetable fats for which nondetects were set to zero.
    NA = not available; Congener-specific PCB data, and data to calculate TEQ concentrations at ND = 0, are limited.
    Standard deviations could not be calculated due to limitations associated with the data (i.e., composite analyses).
    TEQ calculated by setting nondetects to zero.
    Dairy concentration calculated from milk lipid concentrations and then assuming a fat fraction for dairy.
9/22/00
139
DRAFT—DO NOT CITE OR QUOTE

-------
Table 4-6. Background serum levels in the United States 1995 - 1997

Median
Mean
95th Percentile
TEQDFT-WH098 (pg/g lipid)
18.7
22.1*
38.8
2,3,7,8-TCDD (pg/g lipid)
1.9
2.1
4.2
    * After adjusting to account for missing PCBs, the mean is 25.4 pg/g lipid.




    Source: CDC, 2000.
9/22/00
140
DRAFT—DO NOT CITE OR QUOTE

-------
   Table 4-7. Adult contact rates and background intakes of dioxin-like compounds
Exposure route
Soil ingestion
Soil dermal
Freshwater fish
and shellfish
Marine fish and
shellfish
Inhalation
Milk
Dairy
Eggs
Beef
Pork
Poultry
Vegetable fat
Water
Contact rate
50 mg/d
12g/d
5.9 g/d
9,6 g/d
13.3 m3/d
175 g/d
55 g/d
0.24 g/kg-d
0.67 g/kg-d
0.22 g/kg-d
0.49 g/kg-d
17 g/d
1.4L/d
Dioxins and furans
Concentration
TEQDF-WH098
9.4 pg/g
9.4 pg/g
1.0 pg/g
0.26 pg/g
0.12pg/m3
0.031 pg/g
0.12 pg/g
0.081 pg/g
0.20 pg/g
0.22 pg/g
0.11 pg/g
0.056 pg/g
0.0005 pg/L
Total
Intake
(pg TEQDF-
WHO«s/kg-d)
0.0067
0.0016
0.084
0.036
0.023
0.078
0.094
0.019
0.13
0.048
0.054
0.014
0.000011
0.59
(41 pg/d)
Dioxin-like PCBS
Concentration
TEQP-WHO98
NA
NA
1-2 pg/g
0.25 pg/g
NA
0.0 16 pg/g
0.058 pg/g
0.10 pg/g
0.094 pg/g
0.009 pg/g
0.044 pg/g
0.037 pg/g
NA

Intake
(pgTEQp-
WHO00/k2-d)
NA
NA
0.1
0.034
NA
0.040
0.046
0.024
0.063
0.0021
0.022
0.0090
NA
0.34
(24 pg/d)
Total
intake
(pg TEQDFP-
WHO,S /kg-d)
0.0067
0.0016
0.18
0.07
0.023
0.12
0.14
0.043
0.19
0.05
0.076
0.023
0.000011
0.93
(65 pg/d)
9/22/00
141
DRAFT—DO NOT CITE OR QUOTE

-------
    Table 4-8. Variability in average daily TEQ intake as a function of age
Age range
1-5 yr
6-11 yr
12-19 yr
Adult
Intake, mass basis
pgTEQnFP-WH008/d
54
59
64
65
Intake, body weight basis
pg TEQnFP-WH09S/kg-d
3.6
2
1.1
0.9
9/22/00
142
DRAFT—DO NOT CITE OR QUOTE

-------
 I
 JJ
 "S
  es
 .0,


 -2

  O


  o
  U


 "es
  u

 •a
 ,0
   <
  0)
  a
  o
  a

 -o
 OX)



 es





 I

 _g


 en







 ^S



 O
 in
 M
 2
 CS

 H




•*-»
s
4>
1
Q
Q s
O ^L
H =

£rH
<^ §*
o H
£


t»
M
h




Q
°?. 'S,
rf




O^ *o
H !§•

1 ^

H °"



6
•s
o
S






^
CJ
QJ
T3
C
H
2
^^
a
U




« o*
M w
S H

&
itral Tenden
(3
S
c.
&




•M
C3
*• — :
•<-> "O
S g
U H
>
^
o
J


822"
S-
w in"
•£°
i « S
g^'fi
gl-
ta -a m
H S -H


^
C
CO

1
CO
co'
CN


.£>
^~N

CJ
CO

10

CN?
a -2 "
2^
^3
>n


•°a
S-
e


CN

CN

VO
CO
CDC comparison
population, USA 1995 -
97; CDC 2000
S
2 3,4
co ^a
•£ a

If


,
ra

00 S

•n !S
O
CO

"S "
11
Background, Dioxin
Assessment, USA -1990s
























•s
2
3
_o
"«
u
C3
«



















serum


















a "^
CO OS
=3 '
S?
**5 oo
CN a














so
r--
CN
Os
OS
T3* .4,,
E |





serum

















o
C OS
co CN^
T3 CO
CJ ,
£













CO
oo
CN
OS
.„ OS
f2-
%~*
& o>
fl
CO 15
a
^ 2
T3  oj
>O CN oo
~ H
OS
oo
VO
t> a
S -2
CO ^3
c a
CN •"

OO F-
co
OS

%£
Hamburg cohort women;
Flesch-Janys et al., 1999





serum

















0
o
a o
§ CN"
a*?
O^'Tu
CO













CO
in
CN
NIOSH, Fingerhut et al.,
1991b, NTIS





serum

















a o
P
lo

-------
V2
 a
 o
 u
 


Seveso Zone B;
Landietal., 1998
•*-»
is
\o" 2
t— cu
c\ OT
*~* "cT
.M to
«! °
OH A GO
£ o .S
ra ca w
0) .0 =3


















Is
OO
f*- CO
oo c^















o
oo


Seveso Zone B, medical;
Needhametal., 1999h
•4-J
§ g
*» H
r-~  rj CO
g « — «
ils
l-l
^j* w















CN



Seveso NonABR;
Landietal., 1998







a
i


















3 co en
o >
r-<
o
CX
H
'Estimated from ATSDR 1
b CDC data scaled upward
°SD approximated from un
d Weighted average levels f
missing congeners).
ePCDD and PCDF derived
                                                                     >n
                                                                     CN"



                                                                     J
                                                                     3
                                                                     ON
                                                                     J

                                                                     I
                                                                     o
                                                                     e
                                                                      60
                                                                      c
                                                                      oo
                                                                   .S3 3
                                                                   •a S
                                                                    00 *-

                                                                   •2
                                                                    o «a
                                                               •R H
                                                                  e s-s
                                                                      c
                                                                  §•

                                                                   11
                                                                  si g
                                                                  ^ w t2
                                                                    QJ ^^
                                                                    & Q>
9/22/00
                                       144
DRAFT—DO NOT CITE OR QUOTE

-------
Table 5-2. Summary of Cancer Epidemiology and Bioassay Data in Dose-Response
Calculations
Study
Hamburg
cohort,
Becher et al.,
1998
Hamburg
cohort, from
Flesch-Janys
etal.,1998
NIOSH
cohort,
Steenland et
al., 1999
NIOSH
cohort, from
Aylward et al.,
1996
Exposure Groups
0 - 1 (ig/kg fat*Years
1-4 ^g/kg fat*Years
4-8 p.g/kg fat*Years
8-16 jig/kg fat* Years
16 - 64 u.g/kg fat*Years
64+ (J-g/kg fat*Years
1st quartile 0 -125.2
2nd quartile 125.2- 627.1
3rd quartile 627.1 -2503
4th quartile 2503+ ng/kgf*Y
Septile 1
Septile 2
Septile 3
Septile 4
Septile 5
Septile 6
Septile 7
<1 year
1 -5 years
5-15 years
> 15 years
Exposure
Lifetime Ave.
Body Burden
ngTCDD/kg
0-3.6
3.6-14
14-28.6
28.6 - 57
57 - 229
229+
1.41
2.5
6.5
101.2

27.82
103.3
184.5
554.5
AH Cancer
Deaths
Observed
(latency)
1.00 RR
1.12 (Oyr)
1.42
1.77
1.63
2.19
1.24 SMR
,1.34 (Oyr)
1.34
1.73
0.98 SMR
0.90 (15yr)
1.14
1.18
1.33
1.69
1.54
1.02 SMR
1.65 (20yr)
1.38
1.15

t
:
i
(t 	 MfliMKCIr
i *» tw eao j» win an n;a nit
w.ttto
-------
   Table 5-2. Summary of Cancer Epidemiology and Bioassay Data in Dose-Response
   Calculations (continued)
Study
BASF cohort,
from Ott and
Zober 1996
Exposure Groups
<0.1 |ig/kg bw. peak
0.1 - 0.99 (J.g/kg bw. peak
1 .0 - 1 .99 tig/kg bw. peak
2.0+ Hg/kg bw. peak
Exposure
Lifetime Ave.
Body Burden
ngTCDD/kg
4.63
51.9
200.1
2012
AH Cancer
Deaths
Observed
(latency)
0.80 SMR
1.2 (Oyr)
1.4
2.0

t4
savR
2?
2
18
16
14
12
1
08


^^
^^
^^~
/


3D UD 15D 2SDng/kg

S-D Rats,
Kociba et al,
1978;
Goodman &
Sauer, 1992
pathology
0 jig/kg/day
0.001 (ig/kg/day
0.01 (ig/kg/day
0.1 |ig/kg/day
4
135
425
2025
2/86 Tumors
1/50
9/50
18/45
g - Fraction Affected
Multistaae Model with 0.95 C
•6 Multistage T
.5 -^
A' ^-^^
.3 ^^
.2 " .S
BWL
0 0.02 0.04 0.06 0.08 0.1
dose
508/032000
1. For Flesch-Janys et al. (1998), the mean of the AUC in each exposure quartile was calculated as
the mean of the lognormal distribution when restricted to that range. Time mean concentrations Cs
were derived by dividing the mean AUCs by 63 years (derived by subtracting the mean year of birth
of the study subjects, 1929, from the date of followup, 1992). Body burden was computed by
multiplying this lipid concentration by 0.25  (assuming 25% lipid in the body) and adding 1.25 ng/kg
(mean background lipid concentration of 5 ng/kg, times 0.25). Parameters for the fitted lognormal
distribution are |j.=6.3617, a=2.2212.

2. Aylward et al., 1996, Table 5, Cavg/4, assuming 25% lipid

3. For Ott and Zober (1996), the lognormal fitting procedure described above was used to find mean
values for each group. AUCs were then calculated for each group by integrating the solution to the
first-order kinetics equation over time 39 years (the time from the 1953 accident to the 1992
followup). Using C0 as the initial concentration (i.e., that given in the article), this gives AUC =
        '391*]. The constant ke is ln(2)/(half-life). The time-mean concentration is taken to be AUC
    9/22/00
146
DRAFT—DO NOT CITE OR QUOTE

-------
   Table 5-2. Summary of Cancer Epidemiology and Bioassay Data in Dose-Response
   Calculations (continued)

divided by the age 71 years (mean age in 1954, 33 years, + 38 years from 1954 to the date of
followup 1992). Parameters for the fitted lognormal distribution are u=-1.8676, a=2.2927. The half-
life used for the calculation of ke is 7.1 years.

4.  Human equivalent assumption of 25% lipid in rats; empirical data in 22 month old female SD rats
varies from 4-33 % (Birnbaum 1983).
   9/22/00
147
DRAFT—DO NOT CITE OR QUOTE

-------
Table 5-3. Doses yielding 1% excess risk (95% lower confidence bound) based upon 2-year
animal carcinogenicity studies using simple multistage (Portier et al., 1984) models3
ED01
Tumor

Liver cancer in female rats (Kociba)
Squamous cell carcinoma of the tongue in male
rats (Kociba)
Squamous cell carcinoma of the nasal turbinates
or hard palate in male rats (Kociba)
Squamous cell carcinoma of the lung in female
rats (Kociba)
Squamous cell carcinoma of the nasal turbinates
or hard palate in female rats (Kociba)
Thyroid follicular cell adenoma in male rats
(NTP)
Thyroid follicular cell adenoma in female rats
(NTP)
Liver adenomas and carcinomas in female rats
(NTP)
Liver adenomas and carcinomas in male mice
(NTP)
Liver adenomas and carcinomas in female mice
(NTP)
Thyroid follicular cell adenomas and carcinomas
in female mice (NTP)
Subcutaneous tissue sarcomas in female mice
(NTP)
Leukemias and lymphomas in female mice
(NTP)
Shape

Linear
Linear
Cubic
Cubic
Linear
Linear
Cubic
Quadratic
Linear
Linear
Linear
Lin-Cubic
Linear
Animal intake for
1% excess risk
in ng/kg/day (95%
lower confidence
bound)
0.77 (0.57)
14.1 (5.9)
41.4(1.2)
40.4 (2.7)
5.0 (2.0)
4.0 (2.1)
33.0(3.1)
13.0(1.7)
1.3 (0.86)
15.1 (7.8)
30.1 (14.0)
43.2(14.1)
10.0 (5.4)
Steady-state body
burden
in ng/kg at ED01
(95% lower
confidence bound)
14 (10)
254 (106)
746 (22)
730 (48)
90 (36)
144 (76)
1,190(112)
469 (61)
20.6 (13.6)
239 (124)
478 (222)
686 (224)
159 (86)
1 Reprinted with slight modifications from Chapter 8, Table 8.3.2.
9/22/00
148      DRAFT—DO NOT CITE OR QUOTE

-------
      Table 5-4. Summary of All Site Cancer ED01s and Slope Factor Calculations
Study
Hamburg cohort, Becher et al., 1998
Hamburg cohort, from Flesch-Janys et al.,
1998
NIOSH cohort, from Aylward et al., 1996
BASF cohort, from Ott and Zober ,1996
Poisson regression on combined Hamburg
(Flesch-Janys et al., 1998), NIOSH (Aylward
et al., 1996), and BASF (Ott and Zober, 1996)
cohorts5
ED01/LED01J
(95% lower
bound) ng/kg
9.83
5.7 (3.5)
39.9 (23.0)
80.2 (37.5)
47.2(30.1)
Upper bound2
slope factor
risk/pg/kg/day
[3.0E-3]4
8.6 E-3
1.3 E-3 -
0.80 E-3
0.99 E-3

Sprague-Dawley rats, Kociba et al., 1978;
Goodman & Sauer , 1 992 pathology
31.9(22)6
BMD dose
38 (27.5)
BMD adipose
1.4 E-3
1.1 E-3
      See next page for footnotes.
9/22/00
149
DRAFT—DO NOT CITE OR QUOTE

-------
Table 5-4.  Summary of All Site Cancer ED01s and Slope Factor Calculations (cont.)

1. Algorithms used for Poisson dose-response calculation for dioxin exposure and cancer EDs.
Data: exposures for each exposure group - X^; number of deaths expected in each exposure group under conditions of
background exposure - EJ; number of deaths observed in each exposure group - Oj
The model assumes that the risk of death in an exposed group divided by the background risk of death (Ej) is a linear
function of exposure, i.e., Rj = Ej(l + bX,). The parameter b is the slope of the dose-response model. The observed
number deaths in group j, Oj is assumed to be distributed as a Poisson random variable with expected value. R . Under
these assumptions, the solution by maximum likelihood proceeds as follows: The likelihood L is:

                        L = n W [- Ej(l + ^,1- £,(j + /^F'M"
                            J~l
where N - the number of separate exposure groups. The maximum likelihood estimate (MLE),  b, of the parameter p
is obtained by taking the first derivative of the log-likelihood equation, setting it equal to 0 and solving for b.

                        ^ = I- EojXj + (OjXj /(l + bXj)) = 0

The asymptotic variance of the estimate is given by [- (f\nLld p2]"1, with the' observed value Oj replaced by its
expected value Ej(l+bXj):
where b is the MLE. This variance can then be used to obtain approximate 95% upper and lower bounds for b.
Lifetime incremental risk estimates per unit body burden are obtained by multiplying b by the background lifetime
cause-specific risk of death, P0. The ED01s are also calculated from b. Calculations incorporate a lifetime risk of
dying from cancer of 1 8.5%.

2. Formula for column entries: LED0i  * In2 * 1000/T,/2/fraction absorbed = Dose0i pg/kg/day. Cancer slope =
0.01/DoseOI. Assumes T1/2 = 7.1 years, or 2593 days. Assumes 80% absorption from human food supply.

3. Calculated from Becher et al., 1998:                                         ~"   .
i)      Xs Risk fd") = RiskCd) - RCO)  see Chapter 8, Section 8.2.2.
                    R(~) - R(0)
ii)      RRcrcoD) = 1 + 0.000016 * [AUC TCDD ng/kg fat*Y]
        RR(TCDD) = 1 + 0.000016 * 4 * 70 [TCDD  ng/kgBB] (25% lipid, 70 year lifespan)
        RRp-cDD) = 1 + 0.00448 * [TCDD ng/kg lifetime ave. BB]
        Note: Source Becher et al., 1998, table 8, 0 years latency, additive model. Similar results obtained from 10
        year latency model, with AUC adjusted by 60/70 years.
iii)     US Lifetime Risk of Dying from All-Sites Cancer ~1 8.5% during period of study

Calculation of Relative Risk Leading to a 1% Increase in Lifetime Risk of Cancer Mortality: Using the above
formula and lifetime rates, 0.01 = (Risk(EDOI) - 0.185)71-0.185, Risk(ED01) = 0.19315 (i.e.,  a 19.315 % lifetime risk
constitutes a 1% increase under the formula). Therefore, the Relative Risk (ED01) = 0.19315/0.185 ~ 1.044.
Calculation of ED0, from Combining Relative Risk and Slope Formula: RR(ED01) = 1.044 = 1 + 0.00448 * [TCDD
ng/kgBB]; ED0i = 9.8 ng/kg BB. Data are not available to  estimate the LED01. Dose0, = ED0iBB * ln2/T1/2/abs.
- 9.8 * 0.693/2593/0.8 = 0.0033 ng/kg/day. Cancer slope factor = 0.01/dose01 =3.0 E-3 risk/pg/kg/day

4. Based on central estimate; upper confidence limit unavailable.

5. Metaanalysis performed by combining all data sets (i.e., collections of (Xj, Oj, Ej) into a single large data set and
using procedures detailed above.

6. Modeled using EPA benchmark dose software, version  1.2, with either dose or adipose concentration as the metric.
50% absorption assumed from food pellets.  BMD = 0.00176849 ug/kg/day. BMDL = 0.00122517 ug/kg/day.
Therefore, rat LEDOI = 1.2251 x 25 x 0.5/ln2 = 22 ng/kg; human equivalent LED0, = 22 x In2 x 1000/2593/0.8 = 7.38
pg/kg7day; slope factor = 0.01/7.38 = 1.4 E-3 risk/pg/kg/day
9/22/00
150
DRAFT—DO NOT CITE OR QUOTE

-------
     Cl
       2,3,7,8-Tetrachlorodibenzo-p-dioxin
     Cl
       1,2,3,7,8-Pentachlorodibenzo-p-dioxin
         ci
         SjS'j^'.S.S'-Hexachlorobiphenyl
                                                 ci
      2,3,7,8-Tetrachlorodibenzofuran
                                                 Cl
      2,3,4,7,8-Pentachlorodibenzofuran
                                                     Cl
      3,3',4,4',5-Pentachlorobiphenyl
Figure 1-1.  Chemical structure of 2,397,8-TCDD and related compounds.
9/22/00
151        DRAFT—DO NOT CITE OR QUOTE

-------
  c; ^p
  Cf &CI
                                                        Co -activators
                     Differentiation
                         and
                     Proliferation
       Changes in protein levels
                                                       Altered gene expression
Figure 2-1. Cellular mechanism for AhR action.
TCDD, 2,3,7,8-tetrachlorodibenzo-p-dioxin; AhR, aryl hydrocarbon receptor; AIP, associated
immunophilin-like protein; hsp90, 90 kilodalton heat shock protein; p, sites of
phosphorylization; Arnt, AhR nuclear translocator protein; RB/retinoblastoma protein; NF-kB,
nuclear transcription factor; HIF, hypoxia inducible factor; DRE, dioxin-responsive element;
BTFs, basal transcription factors; TATA, DNA recognition sequence.
9/22/00
152
DRAFT—DO NOT CITE OR QUOTE

-------
 CYP1A1
 CYP1A2
 CYP1B1
 GSTYa
 GSTYb
 GSTYc
 UDP glucuronyl transferase
 QR quinone reductase/ Nmo
 Aldehyde dehydrogenase
 Ornithine decarboxylase
 Malic enzyme
 Phospholipase A2
 60kDa microsomal esterase
 Aminolevulinic acid synthetase
 Choline kinase
 EctoATPase
 Prostaglandin synthetase -2 (COX-2)
 Plasminogen activator inhibitor-2
 Urokinase plasminogen activator
 Nedd-4-like ubiquitin protein ligase
 PEPC kinase
 Terminal transferase
 Testosterone Talpha hydroxylase
    Human chorionic gonadotrophin
    Interleukin-1 beta
    Gastrin
    TNF alpha
    TGF-beta
    EGF
    Fibrinogen
    Plastin
    EGFR
    c-erbA related hormone receptor
    Estrogen receptor
    25Dx-putative progesterone receptor
    MDR-1 multidrug resistance
    Aryl hydrocarbon binding protein
    c-fos
    c-jun
    Cystatin-like protein
    MHC-Q1
    Protein kinase C
    pp60 c-src protein kinase
    p21 ras
    p27/Kipl
    bcl-2
Note: This list is not a comprehensive list of all responses known to be affected by TCDD.
Source: Sutler et al., 1992; Lai et al., 1996.
Figure 2-2. Some biochemical responses to TCDD
9/22/00
153
DRAFT—DO NOT CITE OR QUOTE

-------
                 Emission Source
               (tested/total units)
  Best Estimate of
I-TEQ Emission Factor
   (ng/kg or ng/L)
   Total Annual "Activity"
(thousand metric tons/yr or
      million L/yr)
    Municipal Solid Wast* Inclnoratlon (3S/13O)

           Backyard Barrel Burning (NA)

        Medic*! Waste Incineration (2O/2.4OO)

           Secondary Copper Smelting (2/3)

      Cement Kiln* Burning Has Waste 2/7)

                    Crematoria (ie/1,555)

  Hazardous Waato Inclnoratlon (17/162)

  On-Road Unleaded Gas Fuol Combustion (7/7)

    OtvRoad Leaded Gaa Fuel Combustion (?/?)
 The figures include sources with annual I-TEQ emission estimates greater than 5 g l-TEQ/yr in
 one or both of Reference Year 1995 and Reference Year 1987. Derivations of emission factors
 and annual "activity" estimates (e.g., kg of waste incinerated) are presented in the following
 chapters of this report. The difference in bar shading indicates the degree of confidence in
 the estimate. The set of numbers following the source categories indicates the number of
 facilities/sites for which emission test data are available versus the number of facilities/sites
 in the category. A question mark (?) indicates that the precise number of facilities/sites could
 not be estimated.
Annual I-TEQ Emission
    (g l-TEQ/yr)
                                                                      Legend
                                                                   Low Confidence
                                                                   Medium Confidence
                                                                   High Confidence
Figure 4-1.  Estimated CDD/CDF I-TEQ emissions to air from combustion sources in the United
States, 1995.
       9/22/00
                           154
              DRAFT—DO NOT CITE OR QUOTE

-------
   Municipal Solid Waste Incineration
          Backyard Barrel Burning
         Medical Waste Incineration
       Secondary Copper Smelting
    Cement Kilns Burning Haz Waste
         Residential Wood Burning
   Utility/Industrial Coal Combustion
   On-Road Diesel Fuel Combustion
     Secondary Aluminum Smelting
          Industrial Wood Burning
            Iron Ore Sinter Plants
 Cement Kilns Not Burning Haz Waste
       Sewage Sludge Incineration
           Manufacture of EDCA/C
    Utility/Industrial Oil Combustion
                    Crematoria
      Hazardous Waste Incineration
Road Unleaded Gas Fuel Combustion
in-Road Leaded Gas Fuel Combustion
                                                 10
        100
                1000
10000
                                                     1995  U1987
igure 4-2. Comparison of estimates of annual I-TEQ emissions to air (grams I-TEQ/yr) for reference years
>87 and 1995.
          9/22/00
155
DRAFT—DO NOT CITE OR QUOTE

-------
                                                           8
                                                           O5

                                                           O
                                                           h-
                                                           s
O
CD
                     o
                     LO
       o
       CO
o
CM
P!di| B/Bd 'O31
                                           s

                                           .£"
                                           [3


                                           «

                                           CM
                                                                       05
                                                                       O

                                                                       So

                                                                       Q


                                                                       O


                                                                       v


                                                                       a


                                                                       S




                                                                       «
                                                                       •s
                                                                       o
                                                                       o
                                                                       .1
                                                                       fa
                                                                           Q
                                                                           CO
                                               o
                                              •CO
9/22/00
     156
            DRAFT—DO NOT CITE OR QUOTE

-------
(a)
(b)
            50.0
                        10.0
20.0
                      • Formula
                      •1 Year Nursing
    30.0      40.0
      Age, years

	6-Week Nursing
50.0
60.0
70.0
                      	6-Month Nursing
            12.0
                        10.0
20.0
                      - Formula
                      -1 Year Nursing
    30.0     40.0
      Age, years

	6-Week Nursing
                          50.0
                                   60.0
                                       70.0
                      	 6-Month Nursing
Figure 4-4.  Lipid (a) and body burden (b) concentrations in a hypothetical female until
age 70 under four nursing scenarios: formula only, and 6-week, 6-month, and 1 year
nursing.
9/22/00
             157
                   DRAFT—DO NOT CITE OR QUOTE

-------
   10000
                                           14000
    1000
£
 ^   100

 §
•5
10
             8
            ?    1
                 1
                        S
                        0
                        on

                                                  §


     ff
1
                         2378-TCDD
                                               Est. Background Non-TCDD TEQs
    Figure 5-1. Peak dioxin body burden levels in background populations and
    epidemiological cohorts(back-calculated) (See Table 5-1).
    For the background U.S. populations (CDC; USA -1990s), the bars represent the range of total TEQ
    measured in the population. The lower shaded portion represents the variability from non-2,3,7,8-TCDD
    derived TEQs, the upper shaded portion the variability in the 2,3,7,8-TCDD. Note, that the respective bar
    sizes do not represent the total non-2,3,7,8-TCDD TEQ or 2,3,7,8-TCDD contributions, because a portion of
    each of these contributions is contained within the region between the x-axis and bottom of the bar, namely
    the minimum estimated body burden. For each of the back-calculated epidemiological cohort exposures, the
    bar was estimated based on the combination of two distributions: the 2,3,7,8-TCDD levels measured in the
    respective cohort plus the estimated range of background non-2,3,7,8-TCDD derived TEQs from the U.S.
    population.  The lower estimate is the combination of the lower 2,3,7,8-TCDD and lower non-2,3,7,8-TCDD
    TEQ contributions; the shading junction represents the variability in background U.S. population non-
    2,3,7,8-TCDD levels that have been added to this bar; the mean/median/geometric mean indicators represent
    the addition of the measured 2,3,7,8-TCDD central estimate with the mean background U.S. population non-
    2,3,7,8-TCDD TEQ level (-47.6 ppt lipid, 11.9 ng/kg body burden at 25% body fat); and the upper limit is the
    combination of the upper 2,3,7,8-TCDD and upper non-2,3,7,8-TCDD TEQs.
    9/22/00
                                            158
DRAFT—DO NOT CITE OR QUOTE

-------
     250.0 -r
     200.0
     150.0--
                                                           LABB  AUC

                                                           ng/kg ng/kg*Y

                                               Background   3.6    255

                                               Occupational 55.9   3911

                                               LABB = AUC/Yrs
  zn
  c

  =
  CD
  •?  100.0 --
  O
  CD
       0.0
                                   Area Under the Curve

                                      - Occupational

                                     = 3911ng/kg*Y
50.0 - -Burden Occup= 55.9 ng/k;
    0.0       10.0       20.0
                                       30.0       40.0

                                         Age, years
                                                     50.0       60.0       70.0
Figure 5-2. Comparison of lifetime average body burden and area under the curve in

hypothetical background and occupational scenarios.
9/22/00
                                   159       DRAFT—DO NOT CITE OR QUOTE

-------
GLOSSARY AND DEFINITIONS

Adverse Effect: A biochemical change, functional impairment, or pathologic lesion that affects
   the performance of the whole organism, or reduces an organism's ability to respond to an
   additional environmental challenge.

Area Under the Curve (AUC): Area under the concentration vs. time curve. The AUC is a
   summary measure that integrates serial assessments of a dose over the duration of the study.

Aryl hydrocarbon receptor (AhR): An intracellular protein, which is a ligand-dependent
   transcription factor that functions in partnership with a second protein, the aryl hydrocarbon
   receptor nuclear translocator (Arnt).

Aryl hydrocarbon receptor nuclear translocator (Arnt): An intracellular protein that
   functions as a transcription factor in the cell in partnership with a second protein, the aryl
   hydrocarbon receptor (the AhR).

Background Exposure: This is exposure which regularly occurs to members of the general
   population from exposure media (food, air, soil, etc.) that have dioxin concentrations within
   the normal background range. Most (>95%) of background exposure results from the
   presence of minute amounts of dioxin-like compounds in dietary fat, primarily from the
   commercial food supply. The origin of this background exposure is from three categories of
   sources: naturally formed dioxins, anthropogenic dioxins from contemporary sources and
   dioxins from reservoir sources. The term "background exposure" as used in this document
   should not be interpreted as indicating the significance or acceptability of risk associated with
   such exposures.

Benchmark Dose (BMD):  A statistical lower confidence limit on the dose that produces a
   predetermined change in response rate of an adverse effect, typically 1-10%, compared to
   background.

Body Burden: Body burden is defined as the concentration of TCDD and related chemicals in
   the body and is typically expressed as ng/kg body weight. In animals, these values are
   calculated from studies at or approaching steady-state and are associate with either
   biochemical or toxicological responses. In addition, these values are calculated based on
   either knowledge of the species-specific half-life and the exposure or they are estimated
   based on the TCDD tissue concentration, the size of the tissues and the  weight of the animal.
   In humans the values are typically presented as steady-state body burdens and are estimated
   based on an intake rate and the half-life of TCDD in humans. Alternatively, body burdens in
   humans are estimated based on lipid adjusted serum or adipose tissue TCDD or TEQ
   concentrations.

Cancer: A family of diseases affecting cell growth and differentiation, characterized by an
   abnormal, uncontrolled growth of cells.

Carcinogen: An agent capable of inducing cancer.

9/22/00                                   160       DRAFT—DO NOT CITE OR QUOTE

-------
Carcinogenesis:  The origin or production of a benign or malignant tumor. The carcinogenic
    event modifies the genome and/or other molecular control mechanisms of the target cells,
    giving rise to a population of altered cells.

Chronic Effect: An effect which occurs as a result of repeated exposures over a long period of
    time in relation to the lifetime of the organism.

Chronic Exposure:  Multiple exposures occurring over an extended period of time, or a
    significant fraction of the animal's or the individual's lifetime.

Chronic Study: A toxicity study designed to measure the (toxic) effects of chronic exposure to
    a chemical.

Chronic Toxicity: The capacity of a substance to cause adverse human health effects as a result
    of chronic exposure.

Cohort: A cohort is a group of animals of the  same species, including humans, identified by a
    common characteristic, which is studied over a period of time as part of a scientific or
    medical investigation.

Confidence Intervals (CI):  A range of values for a variable of interest, e.g., a rate, constructed
    so that this range has a specified probability of including the true value of the variable.

Confounder:  A condition or variable that is both a risk factor for disease and associated with an
    exposure of interest. This association between the exposure of interest and the confounder (a
    true risk factor for disease) may make it falsely appear that the exposure of interest is
    associated with disease.

Congener: Compounds that have similar chemical structures or belong to closely related
    chemical families

Coplanar: Descriptive term referring to the fact that multi-ringed, chemical structures can
    assume a flat configuration with rings in the same spatial plane.

Dioxin-like:  Dioxin-like is an adjective that refers to the fact that these compounds have similar
    chemical structure, similar physical-chemical properties, and invoke a common battery of
    toxic responses as does 2,3,7,8-TCDD. Because of their hydrophobic nature and resistance
    towards metabolism, these chemicals persist and bioaccumulate in fatty tissues of animals
    and humans. Certain members of the dioxin, furan and PCB family are termed "dioxin-like"
    in this reassessment.

Effective Dose (ED): The dose that corresponds to an increase, expressed as a percent response,
    in relation to expected levels of an adverse  effect can be defined as a percent increase over
    background rates or a percent increase between background and maximal rates.
9/22/00
161
DRAFT—DO NOT CITE OR QUOTE

-------
Effective Dose0i (EDM): The dose corresponding to a 1% increase in an adverse effect.
    Effective dose evaluation at the 10% response level (EDJO or lower bound on EDIO [LED10])
    is somewhat the norm, given the power of most chronic toxicology studies to detect an effect.
    In cases where the data allow evaluation at a lower effective dose level, the Agency suggests
    using the lower value. Such is the case for 2,3,7,8-TCDD.

Epidermal Growth Factor (EGF): A mitogenic polypeptide active on a variety of cell types,
    especially, but not exclusively, epithelial.

Follicle stimulating hormone (FSH):  FSH is an acidic glycoprotein secreted by the anterior
    pituitary gland. In women, follicle stimulating hormone stimulates the development of
    ovarian follicles (eggs) and stimulates the release of estrogens.  In men, follicle stimulating
    hormone stimulates the production of sperm.

Half-life: A measure of the time required to reduce to one half the original concentration of a
    specified chemical in the body

Hormone:  Control chemicals produced by tissues or organs specialized for that function and
    that exert their highly specific effects on other tissues of the body

Latency Period: The time between first exposure to an agent and manifestation or detection of a
    health effect of interest.

Ligand:  Any molecule that binds to another. In normal usage, a soluble molecule
    such as a hormone or neurotransmitter that binds to  a receptor, usually with high affinity.

Lower limit on Effective Dose01 (LED0i): The 95% lower confidence limit of the dose of a
    chemical needed to produce a 1% increase of an adverse effect in those exposed to the
    chemical, or to 1% of the maximal response, relative to control.

Lowest Observed Adverse Effect Level (LOAEL): The lowest exposure level at which there
    are statistically significant increases in frequency or severity of adverse effects between the
    exposed population and its appropriate control group.

Luteinizing Hormone (LH): A hormone that acts with the follicle stimulating hormone  (FSH)
    to stimulate sex hormone release.
Margin of Exposure (MOE): The LED10,LED01, or other point of departure divided by the
   actual or projected environmental exposure/dose of interest, expressed as a ratio.

Minimal Risk Level (MRL): An estimate of daily human exposure to a hazardous substance
   that is likely to be without appreciable risk of adverse noncancer health effects over a
   specified route and duration of exposure.

No-Observed Adverse Effect Level (NOAEL):  The highest exposure level at which there are
   no statistically significant increases in the frequency or severity of adverse effect between the
9/22/00
162
DRAFT—DO NOT CITE OR QUOTE

-------
   exposed population and its appropriate control; some effects may be produced at this level,
   but they are not considered adverse, nor precursors to adverse effects.

No-Observed Effect Level (NOEL): An exposure level at which there are no statistically
   significant increases in the frequency or severity of any effect between the exposed
   population and its appropriate control.

Pharmacokinetics: The quantitative description of the process of chemical disposition:
   absorption, distribution, metabolism, and excretion (metabolism and excretion equal
   elimination).

Physiologically Based Pharmacokinetic (PBPK) Model: Physiologically based model used to
   characterize pharmacokinetic behavior of a chemical. Available data on blood flow rates and
   metabolic and other processes which the chemical undergoes within each compartment are
   used to construct a mass-balance framework for the PBPK model.

Point of Departure: The dose-response point that marks the lower end of the range of
   observation and the beginning of a low-dose extrapolation. This point is most often the upper
   bound on an observed incidence or on an estimated incidence from a dose-response model, or
   the lower bound on dose associated with such an incidence.

Promoter: An agent that is not carcinogenic itself, but when administered after an initiator of
   carcinogenesis stimulates the clonal expansion of the initiated cell to produce  a neoplasm.

Receptor:  A molecular structure within a cell or on the cell's surface, characterized by selective
   binding of a specific substance and a specific physiologic effect that accompanies the binding
   (for example, see Aryl hydrocarbon receptor).

Receptor Site:  The portion of the receptor molecule or structure with which the compound
   (ligand) interacts.

Reference Dose (RfD):  An estimate (with uncertainty spanning perhaps an order of magnitude)
   of a daily oral exposure to the human population (including sensitive subgroups) that is likely
   to be without an appreciable risk of deleterious effects during a lifetime. It can be derived
   from a NOAEL, LOAEL, or benchmark dose, with uncertainty factors generally applied to
   reflect limitations of the data used. Generally used in EPA's noncancer health assessments.

Relative Risk (RR): The relative measure of the difference in risk between the exposed and
   unexposed populations in a cohort study. The relative risk is defined as the rate of disease
   among the exposed divided by the rate of the disease among the unexposed. A relative risk
   of 2 means that the exposed group has twice the disease risk as the unexposed group.

Reservoir Sources: Reservoirs are materials or places that contain previously formed
   CDD/CDFs or dioxin-like PCBs and have the potential for redistribution and circulation of
   these compounds into the environment.  Potential reservoirs include soils, sediments, biota,
9/22/00
163
DRAFT—DO NOT CITE OR QUOTE

-------
    water and some anthropogenic materials.  Reservoirs become sources when they have
    releases to the circulating environment.

Risk (in the context of human health): The probability of injury, disease, or death from
    exposure to a chemical agent or a mixture of chemicals. In quantitative terms, risk is
    expressed in values ranging from zero (representing the certainty that harm will not occur) to
    one (representing the certainty that harm will occur).

Slope Factor: An upper bound, generally approximating or exceeding a 95% confidence limit,
    on the increased cancer risk from a lifetime exposure to an agent.  This estimate, usually
    expressed in units of proportion (of a population) affected per mg/kg/day, is generally
    reserved for use in the low-dose region of the dose-response relationship j that is, for
    exposures corresponding to risks less than 1 in 100.

Standardized Mortality Ratio (SMR):  This is the relative measure of the difference in risk
    between the exposed and unexposed populations in a cohort study. The SMR is similar to
    the relative risk in both definition and interpretation. This measure is usually standardized to
    control for any differences in age, sex, and/or race between the exposed and reference
    populations.  It is frequently converted to a percent by multiplying the ratio by 100.

Statistical Significance: The probability that a result may be due to chance alone. By
    convention, a difference between two groups is usually considered statistically significant if
    chance could explain it only 5% of the time or less. Study design considerations may
    influence the a priori choice of a different statistical significance level.

Thyroid Stimulating Hormone (TSH): A hormone secreted by the anterior pituitary gland that
    activates certain actions in thyroid cells leading to production and release of the thyroid
    hormones (T3 and T4). T3 and T4 blood levels feedback on the hypothalmus/pituitary gland
    and decrease TSH production when T3 and T4 levels are high.

Tolerable Daily Intake (TDI): A TDI is an estimate of the amount of a contaminant in food or
    drinking water that can be ingested daily over a lifetime without a significant health risk.
    The term is used frequently in World Health Organization (WHO) health assessments.  The
    term "tolerable" is used as contaminants do not serve an intended function and as intake is
    unavoidably associated with the basic consumption of food and water. Tolerable does not
    generally connote "acceptable" or "risk free."

Toxic Equivalence (TEQ):  The toxic equivalency factor (TEF) of each dioxin-like compound
    present in a mixture multiplied by the respective mass concentration.  The products  are
    summed to represent the 2,3,7,8-TCDD Toxic Equivalence of the mixture.

Toxic Equivalency Factor (TEF):  TEFs compare the potential toxicity of each dioxin-like
    compound comprising the mixture to the well-studied and understood toxicity of 2,3,7,8-
    TCDD, the most toxic member of the group, with the TEF of 2,3,7,8-TCDD being 1. TEFs
    are the result of expert scientific judgment using all of the available data, taking into account
    uncertainties in the available data.
9/22/00
164
DRAFT—DO NOT CITE OR QUOTE

-------
Transcription:  The process of constructing a messenger RNA molecule using a DNA molecule
    as a template with resulting transfer of genetic information to the messenger RNA.

Transcription Factor:  A substance, usually a protein, that is developed within the organism,
    that is effective in the initiation, stimulation, or termination of the genetic transcription
    process.

Upper bound: A plausible upper limit to the true value of a quantity or response. This is
    usually not a true statistical confidence limit.

Weight-of-Evidence: An approach used for characterizing the extent to which the available
    data, including human, animal, and mechanism of action, support the hypothesis that an agent
    causes an adverse effect, such as cancer, in humans. The approach considers all scientific
    information, both positive and negative, in determining whether and under what conditions
    an agent may cause disease in humans.
9/22/00
165
DRAFT—DO NOT CITE OR QUOTE

-------
         REFERENCES FOR RISK CHARACTERIZATION
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Abbott, BD; Schmid, JE; Pitt, JA; et al. (1999) Adverse reproductive outcomes in the transgenic AhR-deficient
mouse. Toxicol Appl Pharmacol 155(1 ):62-70.

Abraham, K; Krowke, R; Neubert, D. (1988) Pharmacokinetics and biological activity of 2,3,7,8-tetrachlorodibenzo-
p-dioxin. 1. Dose-dependent tissue distribution and induction of hepatic ethoxyresorufin O-deethylase in rats
following a single injection. Arch Toxicol 62:359-368.

Ahlborg, VG; Becking, GC; Birnbaum, LS; et al. (1994) Toxic equivalency factors for dioxin-like PCBs.
Chemosphere28(6):1049-1067.                                                           '

Alaluusua, S; Lukinmaa, P-L; Torppa, T; et al. (1999) Developing teeth as biomarker of dioxin exposure. Lancet
353:206.

Alaluusua, S; Lukinmaa, P-L; Vartiainen, T; et al. (1996) Polychlorinated dibenzo-p-dioxins and dibenzofurans via
mother's milk may cause developmental defects in the child's teeth. Environ Toxicol Pharmacol 1:193-197.

Allen, BC; Kavlock, RJ; Kimmel, CA; et al. (1994) Dose-response assessment for developmental toxicity. II.
Comparison of generic benchmark dose estimates with no observed adverse effect levels. Fundam Appl Toxicol
23:487-495.

Allen, JR.; Lalich, JJ. (1962) Response of chickens to prolonged feeding of crude "toxic fat." Proc Soc Exp Biol
Med 109:48-51.

Allen, JR; Carstens, LA. (1967) Light and electron microscopic observations in Macaca mulatto monkeys fed toxic
fat. Am J Vet Res 28:1513-1526.

Allen, JR; Barsotti, DA; Van Miller, JP; et al.  (1977) Morphological changes in monkgys consuming a diet
containing low levels of 2,3,7,8-tetrachlorodibenzodioxin. Food Cosmet Toxicol 15:401-410.

Allen, JR.; Barsotti, DA; Lambrecht, LK; et al. (1979) Reproductive effects-of halogehated aromatic hydrocarbons
on nonhuman primates. Ann N Y Acad Sci 320:419-425.

Alsharif, NZ; Lawson, T; Stohs, SJ.  (1994) Oxidative stress induced by is mediated by the aryl hydrocarbon (Ah)
receptor complex. Toxicology 92:39-51.

American Academy of Pediatrics. (1997) Breastfeeding and the use of human milk. Pediatrics 100 (6): 1035-1039.

Ambrosone, CB; Freundenheim, JL; Graham, S; et al. (1995) Cytochrome P450IA1 and glutathione-s-transferase
(Ml) genetic polymorphisms and post-menopausal breast cancer risk. Cancer Res 55:3483-3485.

Andersen, ME; Bimbaum, LS; Barton, HA; et al. (1997) Regional hepatic CYP1A1 and CYP1A2 induction with
2,3,7,8-tetrachlorodibenzo-/7-dioxin  evaluated with a multi-compartment geometric model of hepatic zonation.
Toxicol Appl Pharmacol 144:145-155.

Ariens, EJ; van Rossum, JM; Koopman, PC. (1960) Receptor reserve and threshold phenomena. I. Theory and
experiments with autonomic drugs tested on isolated organs. Arch Int Pharmacodyn  127:459-478.

Arnold, DL; Nera, EA; Stapley, R; et al. (1996) Prevalence of endometriosis in rhesus (Macaca mulatta) monkeys
ingesting PCB (Aroclor 1254): review and evaluation. Fundam Appl Toxicol 3 l(l):42-55.

ATSDR. (1999a) Toxicological profile for chlorinated dibenzo-p-dioxins. United States Department of Health and
Human Services.
         9/22/00
                                                ,166
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
ATSDR. (1999b) Health Consultation (Exposure Investigation) Calcasieu Estuary (aka Mossville) Lake Charles,
Calcasieu Parish, Louisiana. Cerclis No. LA0002368173.  Prepared by: Exposure Investigation and Consultation
Branch, Division of Health Assessment and Consultation, Agency for Toxic Substances and Disease Registry.
November 19, 1999.

Aylward, LL; Hays, SM; Karch, NJ; et al. (1996) Relative susceptibility of animals and humans to the cancer hazard
posed by 2,3,7,8-tetrachlorodibenzo-p-dioxin using internal measures of dose. Environ Sci Technol 30:3534-3543.

Bachmann K, Pardoe D, White D. (1996) Scaling basic toxicokinetic parameters from rat to man. Environ Health
•Perspectl04(4):400-7.

Barsotti, DA; Abrahamson, LJ; Allen, JR. (1979) Hormonal alterations in female rhesus monkeys fed a diet
containing 2,3,7,8-tetrachlorodibenzo-/>-dioxin. Bull Environ Contain Toxicol 21:463-469.

Becher, H; Flesch-Janys, D; Kauppinen, T; et al. (1996) Cancer mortality in German male workers exposed to
phenoxy herbicides and dioxins. Cancer Causes Control 7:312-321.

Becher, H; Steindorf, K.; Flesch-Janys, D. (1998) Quantitative cancer risk assessment for dioxins using an
occupational cohort. Environ Health Perspect 106(2):663-670.

Beck, H; Eckart, K; Mathar, W; et al. (1989) Levels of PCDD's and PCDF's in adipose tissue of occupationally
exposed workers. Chemosphere 18:507-516.

Bertazzi, PA; di Domenico. (1994) Chemical, environmental, and health aspects of the Seveso, Italy, accident. In:
Dioxins and Health. Arnold Schecter, ed. New York: Plenum Press, pp. 587-632.

Bertazzi, PA; Pesatori, AC; Consonni, D; et al. (1993) Cancer incidence in a population accidentally exposed to
2,3,7,8-tetrachlorodibenzo-para-dioxin. Epidemiology 4(5):398-406.

Bertazzi, PA; Zocchetti,  C; Guercilena,  S; et al. (1997) Dioxin exposure and cancer risk": a 15-year mortality study
after the "Seveso Accident." Epidemiology 8(6):646-652.

Bertazzi, PA; Bemucci, I; Brambilla, G; et al. (1998) The Seveso studies on early and long-term effects of dioxin
exposure: a review. Environ Health Perspect 106(2):625-633.

Bertazzi, PA; Pesatori, AC; Consonni, D; et al. (1999) Epidemiology of long-term health effects of dioxin exposure
in the Seveso population. Organohalogen Compounds 44:337-338.

Bimbaum, LS. (1983) Distribution and excretion of 2,3,6,2',3',6'- and 2,4,5,2'4'5'-hexachlorobiphenyl in senescent
rats. Toxicol Appl Pharmacol 70:262-272.

Birnbaum, LS. (1994a) Evidence for the role of the AhR in responses to dioxin. In: Receptor-mediated biological
processes: implications for evaluating carcinogenesis. Progress in Clinical and Biological Research, vol. 387.
Spitzer, HL; Slaga, TJ; Greenlee, WF; et al., eds. New York: Wiley-Liss, Inc., pp. 139-154.

Birnbaum, LS. (1994b) The mechanism of dioxin toxicity: relationship to risk assessment. Environ Health Perspect
102 (Supplement 9): 157-167.

Bjerke, DL; Peterson, RE. (1994) Reproductive toxicity of 2,3,7,8-tetrachlorodibenzo-/>-dioxin in male rats: different
effects of in utero versus lactational exposure. Toxicol Appl Pharmacol 127:241-249.

Bjerke, DL; Sommer, RJ; Moore, RW; et al. (1994a) Effects of in utero and lactational 2,3,7,8-tetrachlorodibenzo-
p-dioxin exposure on repsonsiveness of the male rat reproductive system to testosterone stimulation in adulthood.
Toxicol Appl Pharmacol 127:250-257.
         9/22/00
                                                  167
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
44
45
46
47
48
49
 50
 51
 52
 53
 54
 55
 Bjerke, DL; Brown, TJ;  MacLusky, NJ; Hochberg, RB; et al. (1994b) Partial demasculinization and feminization of
 sex behavior in male rats by in utero and lactational exposure to 2,3,7,8-tetrachlorodibenzo-p- dioxin is not
 associated with alterations in estrogen receptor binding or volumes of sexually differentiated brain nuclei. Toxicol
 Appl Pharmacol 127(2): 258-67.

 Bock, KW; Gschaidmeier, H; Heel, H; et al. (1998) AH receptor-controlled transcriptional regulation and function
 of rat and human UDP-glucuronosyltransferase isoforms. Adv Enzyme Regul 38:207-22

 Bond, GG; McLaren, EA; Brenner, FE; et al. (1989) Incidence of chloracne among chemical workers potentially
 exposed to chlorinated dioxins. J Occup Med 31:771-774.

 Bookstaff, RC; Kamel, F; Moore, RW; et al. (1990a) Altered regulation of pituitary gonadotropin-releasing hormone
 (GnRH) receptor number and pituitary responsiveness to GnRH in 2,3,7,8-tetrachlorodibenzo-p-dioxin-treated male
 rats. Toxicol Appl Pharmacol 105:78-92.

 Bookstaff, RC; Moore, RW; Peterson, RE. (1990b) 2,3,7,8-tetrachlorodibenzo-p-dioxin increases the potency of
 androgens and estrogens as feedback inhibitors of luteinizing hormone secretion in male rats. Toxicol Appl
 Pharmacol 104:212-224.

 Boyd, JA; Clark, GC; Walmer, D; et al. (1995) Endometriosis and the environment:, biomarkers of toxin exposure.
 Conference on Endometriosis 2000, May 15-17.

 Breslow, NE; Day, NE. (1987) Statistical methods in cancer research. Volume II~The design and analysis of cohort
 studies. IARC Sci Publ 82:1-406.

 Brody, BB; Reid WD. (1967) Fed Proc. 26:1062-1070.

 Brown, NM; Manzolillo, PA; Zhang, JX; et al. (1998) Prenatal TCDD and predisposition to mammary cancer in the
 rat. Carcinogenesis 19(9):1623-1629.

 Bruner-Tran, KL; Rier, SE; Eisenberg, E; et al. (1999) The potential role of environmental toxins in the
 pathophysiology of endometriosis. Gynecol Obstet Invest Oct;48 Suppl S1:45-56.    .-

 Bruzy, L.P.;  Kites, R.A.  (1995) Estimating the atmospheric deposition of polychlorinated dibenzo-p-dioxins and
 dibenzofurans from soil.  Environ Sci Technol 29:2090-2098. >

 Bueno de Mesquita, HB; Doornbos, G; van der Kuip, DM; et al. (1993) Occupational exposure to phenoxy
 herbicides and chlorophenols and cancer mortality in the Netherlands. Am J Ind Med 23:289-300.

 Calvert, GM; Homung, RW; Sweeney, MH; et al. (1992) Hepatic and gastrointestinal effects in an occupational
 cohort exposed to 2,3,7,8-tetrachlorodibenzo-para-dioxin. JAMA 267:2209-2214.

 Calvert, GM; Willie, KK; Sweeney, MH; et al. (1996) Evaluation of serum lipid concentrations among U.S. workers
 exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Arch Environ Health 51(2): 100-107.

Calvert, GM; Sweeney, MH; Deddens, J; et al. (1999) Evaluation of diabetes mellitus, serum glucose, and thyroid
function among United States workers exposed to 2,3,7,8-tetrachlorodi-benzo-jo-dioxin. Occup Environ Med
56(4):270-276.

Caramaschi,  F; Del Caino, G; Favaretti, C; et al. (1981)  Chloracne following environmental contamination by TCDD
in Seveso, Italy. Int J Epidemiol 10:135-143.

 Carver, LA; LaPres, JJ; Jain, S; et al. (1998) Characterization of the AhR-associated protein, ARA9. J Biol Chem
273(50):33580-33587.
         9/22/00
                                                 168
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
40
41
42
43
44
45
46
47
48
49
50
 51
 52
 53
 54
 55
 56
 CDC. (1997) Vital and Health Statistics. Fertility, Family Planning, and Women's Health: New Data From the 1995
 National Survey of Family Growth. National Center for Health Statistics, Centers for Disease Control and
 Prevention, U.S. Department of Health and Human Services. Series 23, No. 19. May 1997.

 CDC. (2000) Personal communication from D. Patterson, CDC, Atlanta, GA to M. Lorber, U.S. EPA, Washington
 DC. April, 2000.

 Centers for Disease Control Vietnam Experience Study. (1988) Health status of Vietnam veterans. II. Physical
 health. JAMA 259:2708-2714.

 Chahoud, I.; Krowke, R.; Schimmel, A.; et al. (1989) Reproductive toxicity and pharmacokinetics of 2,3,7,8-
 tetrachlorodibenzo-^-dioxin. I. Effects of high doses on the fertility of male rats. Arch Toxicol 63:432-439.

 Chen, YCJ; Quo, YLL; Hsu, CC. (1992) Cognitive development of children prenatally exposed to polychlorinated
 biphenyls (Yu-Cheng children) and their siblings. J Formosan Med Assoc 91:704-707.

 Cheung, MO; Gilbert, EF; Peterson, RE. (1981) Cardiovascular teratogenicity of 2, 3, 7,
 8-tetrachlorodibenzo-/7-dioxin in the chick embryo. Toxicol Appl Pharmacol 61 (2): 197-204.

 Clark, GC; Tritscher, A; Maronpot, R; et al.  (1991) Tumor promotion by TCDD in female rats. In: Banbury Report
 35: biological basis for risk assessment of dioxin and related compounds. Gallo, M; Scheuplein, R; van Der Heijden,
 K, eds. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory; pp. 389-404.

 Clark, AJ. (1933) The mode of action of drags on cells. Baltimore: Williams and Wilkins.

 Cohen, GM; Bracken, WM; Iyer, RP; et al. (1979) Anticarcinogenic effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin
 on benzo(a)pyrene and 7,12-dimethylbenz(a)anthracene tumor initiation and its relationship to DNA binding. Cancer
 Res 39:4027-4033.

 Cohen, J. (1977) Statistical Power Analysis for the Behavioral Sciences, Revised Edition. Academic Press, New
 York.

 Conolly, RB; Andersen, ME. (1997) Hepatic foci in rats after diethylnitrosamine initiation and 2,3,7,8-
 tetrachlorodibenzo-p-dioxin promotion: evaluation of a quantitative two-cell model and of CYP 1A1/1A2 as a
 dosimeter. Toxicol Appl Pharmacol 146:281-293.

 Courtney, KD; Moore, JA.  (1971) Teratology studies with 2,4,5-T and 2,3,7,8-TCDD. Toxicol Appl Pharmacol
20:396-403.

 Couture, LA; Abbott, BD; Birnbaum, LS. (1990) A critical review of the developmental toxicity and teratogenicity
 of 2,3,7,8-tetrachlorodibenzo-p-dioxin: recent advances toward understanding the mechanism. Teratology 42:619-
 627.

 Cresanta, JL; Farris, RP; Croft, JB; Frank, GC; Berenson, GS.  (1988) Trends in fatty acid intakes of 10-year-old
 children, 1973-1982. Journal of American Dietetic Association 88:178-184.

 Cummings, AM; Metcalf, JL; Birnbaum, L. (1996)  Promotion of endometriosis by
2,3,7,8-tetrachlorodibenzo-p-dioxin in rats and mice: time-dose dependence and species  comparison. Toxicol Appl
Pharmacol 138(1):131-139.

Dannan, GA; Porubek, DJ;  Nelson, SD; et al. (1986) 17 beta-estradiol 2- and 4-hydroxylation catalyzed by rat
hepatic cytochrome P-450:  roles of individual forms, inductive effects, developmental patterns, and alterations by
gonadectomy and hormone replacement. Endocrinology 118:1952-1960.

 Davis, D; Safe, S. (1988) Immunosuppressive activities of polychlorinated dibenzofuran congeners: quantitative
 structure-activity relationships and interactive effects. Toxicol Appl Pharmacol 94:141-149.
         9/22/00
                                                 169
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
Denison, MS; Phelan, D; Elferink, CJ. (1998) The AhR signal transduction pathway. In: Toxicant-receptor
interactions. Denison, MS; Helferich, WG, eds. Bristol. PA: Taylor & Francis, pp. 3-33.

Dertinger, SD; Silverstone, AE; Gasiewicz, TA. (1998) Influence of aromatic hydrocarbon receptor-mediated events
on the genotoxicity of cigarette smoke condensate. Carcinogenesis 19:2037-2042.

DeVito, MJ; Ma, XF; Babish, JG; et al. (1994) Dose-response relationships in mice following subchronic exposure
to2,3,7,8-tetrachlorodibenzo-p-dioxin: cyplal, cypla2, estrogen-receptor, and protein-tyrosine phosphorylation.
Toxicol Appl Pharmacol 124:82-90.

DeVito, MJ; Birnbaum, LS; Farland, WH; et al. (1995) Comparisons of estimated human-body burdens of dioxinlike
chemicals and TCDD body burdens in experimentally exposed animals. Environ Health Perspect 103:820-831.

DiGiovanni, J.; Berry, DL; Gleason, GL; et al. (1980) Time-dependent inhibition by 2,3,7,8-tetrachlorodibenzo-p-
dioxin of skin tumorigenesis with polycyclic hydrocarbons. Cancer Res 40:1580-1587.

Diliberto, IT; Akubue, PI; Luebke, RW; et al. (1995) Dose-response relationships of tissue distribution and induction
of CYP1A1 and CYP1A2 en2ymatic-activities following acute exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin
(TCDD) in mice. Toxicol Appl Pharmacol 130:197-208.

Doss, M; Saver, H; von Tiepermann, R; et al. (1984) Development of chronic hepatic porphyria (porphyria cutanea
tarda) with inherited uroporphyrinogen decarboxylase deficiency under exposure to dioxin. J Biochem  16:369-373.

Dragan, YP; Xu, X; Goldsworthy, TL; et al. (1992) Characterization of the promotion of altered hepatic foci by
2,3,7,8-tetrachlorodibenzo-p-dioxin in the female rat. Carcinogenesis 13(8): 1389-1395.

DiGiovanni, J; Viaje, A; Berry, DL; et al. (1977) Tumor initiating ability of TCDD and Arochlor 1254 in the two
stage system of mouse skin Carcinogenesis. Bull Environ Contain Toxicol  18:552-557.

Dunagin, WG. (1984) Cutaneous signs of systemic toxiciry due to dioxins  and related-chemicals.  J Am Acad
Dermatol 10(4):688-700.

Dunson, DB; Haseman, JK; van Birgelen, APJM; et al. (2000) Statistical analysis of skin tumor data from Tg.AC
mouse bioassays. Toxicol Sci, in press.

Eastin, WC; Haseman, JK; Mahler, JF; et al. (1998) The National Toxicology Program evaluation of genetically
altered mice as predictive models for identifying carcinogens. Toxicol Pathol 26:461-473.

Egeland, GM; Sweeney, MH; Fingerhut,  MA; et al. (1994) Total serum testosterone and gonadotropins in workers
exposed to dioxin. Am J Epidemiol 139:272-281.

Enan, E; Matsumura, F. (1994) 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD)-induced changes in glucose
transporting activity in guinea pigs, mice, and rats in vivo and in vitro. J Biochem Toxicol 9(2):97-106.

Enan, E; Matsumura, F. (1996) Identification of c-Src as the integral component of the cytosolic AhR complex,
transducing the signal of 2,3,7,8-tetrachlorodibenzo-j7-dioxin (TCDD) through the protein phosphorylation pathway.
Biochem Pharmacol 52(10):1599-1612.

Eriksson, M; Hardell, L; Berg, NO; et al. (1981) Soft-tissue sarcomas and exposure to chemical substances: a case-
referent study.  Br J Ind Med 38:27-33.

Eriksson, M; Hardell, L; Adam, H. (1990) Exposure to dioxins as a risk factor for soft tissue sarcoma: a population-
based case-control study. J Natl Cancer Inst 82:486-490.
         9/22/00
                                                 170
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
43
44
45
46
47
48
49
 50
 51
 52
 53
 54
 55
 Ernst, M; Flesch-Janys, D; Morgenstern, I; et al. (1998) Immune cell functions in industrial workers after exposure
 to 2,3,7,8-tetrachlorodibenzo-p-dioxin: dissociation of antigen-specific T-cell responses in cultures of diluted wholi
 blood and of isolated peripheral blood mononuclear cells. Environ Health Perspect 106 Suppl 2:701-705.
 Eskenazi, B; Mocarelli, P; Warner, M; et al. (1998) Seveso women's health study: A study of the effects of TCDD
 on reproductive health. Orgaonhalogen compounds 38:219-222.

 Esteller, M; Garcia, A; Matinez-Palones, JM; et al. (1997) Germ line polymorphisms in cytochrome P450IA1
 (C4887 CYPIA1) and methylenetetrahydrofolate reductase (MTHFR) genes and endometrial cancer susceptibility.
 Carcinogenesis 18:2307-2311.

 Fernandez-Salguero, PM; Hilbert, DM; Rudikoff, S; et al. (1996)  Aryl-hydrocarbon receptor-deficient mice are
 resistant to 2,3,7,8-tetrachlorodibenzo-p-dioxin-induced toxiciry.  Toxicol Appl Pharmacol 140(1):173-179.

 Fingerhut, MA; Halperin, WE; Marlow, DA. (1991a) Cancer mortality in workers exposed to 2,3,7,8-
 tetrachlorodibenzo-p-dioxin. New Engl J Med 324:212-218.

 Fingerhut, MA; Halperin, WE; Marlow, D; et al. (1991b) Mortality among United States workers employed in the
 production of chemicals contaminated with 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Cincinnati, OH: U.S.
 Department of Health and Human Services, National Institute for Occupational Safety and Health NTIS# PB 91-
 125971.                                                              •

 Fleiss, JL. (1981) Statistical methods for rates and proportions.  John Wiley, New York.

 Flesch-Janys, D;  Steindorf, K; Gum, P; et al. (1998) Estimation of the cumulated exposure to polychlorinated
 dibenzo-p-dioxins/furans and standardized mortality ratio analysis of cancer mortality by dose in an occupationally
 exposed cohort. Environ Health Perspect 106(supplement 2):655-662.

 Flesch-Janys, D;  Becher, J; Berger, J; et al. (1999) Epidemiological investigation of breast cancer incidence in a
 cohort of female  workers with high exposure to PCDD/CDF and HCH. Organohalogen'Compounds 44:379-382.

 Flesch-Janys, D;  Berger, J; Gurn, P; et al. (1995) Exposure to polychlorinated dioxins and furans (PCDD/CDF) and
 mortality in a cohort of workers from a herbicide-producing plant in Hamburg, Federal Republic of Germany. Am J
 Epideimol 142:1165-1175.

 Flodstrom, S; Ahlborg, UG. (1992) Relative tumor promoting activity of some polychlorinated dibenzo-p-dioxin-,
 dibenzofuran-, and biphenyl congeners in female rats. Chemosphere 25:1(2): 169-172.

 Frank, G.C.; Webber, L.S.; Farris, R.P.; Berenson, G.S. (1986) Dietary databook:  quantifying dietary intakes of
 infants,  children,  and adolescents, the Bogalusa heart study, 1973-1983. National Research and Demonstration
 Center - Arteriosclerosis, Louisiana State University Medical Center, New Orleans, Louisiana.

 Gaido, KW; Maness, SC; Leonard, LS; et al. (1992) 2,3,7,8-Tetrachlorodibenzo-p-dioxin-dependent regulation of
transforming growth factors-a and P2 expression in a human keratinocyte cell line involves both transcriptional and
post-transcriptional control. J Biol Chem 267:24591-24595.

Gasiewicz, TA. (1997) Dioxins and the AhR: probes to uncover processes in neuroendocrine development.
Neurotoxicology 18:393-414.

Gasiewicz, TA; Holscher, MA; Neal, RA. (1980) The effect of total parenteral nutrition on the toxicity of 2,3,7,8-
tetrachlorodibenzo-p-dioxin in the rat. Toxicol Appl Pharmacol 54:469-488.

 Gierthy, JF; Bennett, JA; Bradley, LM; et al. (1993) Corrleation of in vitro and in vivo growth suppression of MCF-
 7 human breast cancer by 2,3,7,8-tetrachlorodibenzo-p-dioxin. Cancer Res 53:3149-3153.
         9/22/00
                                                  171
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Gerhard, I; Runnebaum, B; (1992) Grenzen der hormonsubsittution bei schadstoffbelastung und fertilitatsstorungen.
Zent Bl Gynekol 114:593-602.

Goldstein, JA; Hickman, P; Jue, DL. (1974) Experimental hepatic porphyria induced by polychlorinated biphenyls.
Toxicol Appl Pharmacol 27(2):437-448.

Goodman, DG; Sauer, RM. (1992) Hepatotoxicity and carcinogenicity in female Sprague-Dawley rats treated with
2,3,7,8-tetrachlorordibenzo-p-dioxin (TCDD): a Pathology Working Group reevaluation. Regul Toxicol Pharmacol
15:245-252.

Gorski, JR; Rozman, K. (1987) Dose-response and time course of hypothyroxinemia and hypoinsulinemia and
characterization of insulin hypersensitivity in 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)-treated rats. Toxicology
44(3):297-307.

Gradin, K; McGuire, J; Wenger, RH; et al. (1996) Functional interference between hypoxia and dioxin signal
transduction pathways: competition for recruitment of the ARNT transcription factor. Mol Cell Biol
16(10):5221-5231.

Graham, MJ; Lucier, GW; Linko, P; et al. (1988) Increases in cytochrome P-450 mediated 17 beta-estradiol
2-hydroxylase activity in rat liver microsomes after both acute administration and subchronic administration of
2,3,7,8-tetrachlorodibenzo-p-dioxin in atwo-stage hepatocarcinogenesis model. Carcinogenesis 9:1935-1941.

Gray, LE, Jr.; Kelce, WR; Monosson, E; et al. (1995a) Exposure to TCDD during development permanantly alters
reproductive function in male Long Evans rats and hamsters: reduced ejaculated and epididymal sperm numbers and
sex accessory gland weights in offspring with normal androgenic status. Toxicol Appl Pharmacol 131:108-118.

Gray, LE, Jr.; Ostby, J; Wolf, C; et al. (1995b) Functional developmental toxicity of low doses of 2,3,7,8-
tetrachlorodibenzo-/?-dioxin and a dioxin-like PCB (169) in Long Evans rats and Syrian hamsters: reproductive,
behavioral and thermoregulatory alterations. Organohalogen Compounds 25:33-38.  .„„

Gray, LE, Jr.; Ostby, JS. (1995) In utero 2,3,7,8-tetrachlorodibenzo-^-dioxin (TCDD) alters reproductive
morphology and function in female rat offspring. Toxicol Appl Pharmacol 133:285-2-94.

Gray, LE;  Ostby, JS; Kelce, WR. (1997a) A dose-response analysis of the reproductive effects of a single
gestational dose of 2,3,7,8-tetrachlorodibenzo-p-dioxin in male Long Evans Hooded rat offspring. Toxicol Appl
Pharmacol 146(1): 11-20.

Gray, LE; Wolf, C; Mann, P; Ostby, JS. (1997b) In utero exposure to low doses of 2,3,7,8-tetrachlorodibenzo-p-
dioxin alters reproductive development of female Long Evans hooded rat offspring. Toxicol Appl Pharmacol
146(2):237-44.

Grubbs, WD;  Wolfe, WH; Michalek, JE; et al. (1995) Air Force Health Study: an epidemiologic investigation of
health effects  in Air Force personnel following exposure to herbicides. Report number AL-TR-920107.

Gu, Yi-J; Hogenesch, JB; Bradfield, CA. (2000) The PAS Superfamily: Sensors of Environmental and
developmental signals. Annu Rev Pharmacol Toxicol 40:519-561.

Gupta, BN; Vos JG, Moore; JA, Zinkl; et al. (1973) Pathologic effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin in
laboratory animals. Environ Health Perspect 5:125-140.

Guzelian, PS. (1985) Clinical evaluation of liver structure and function in humans exposed to halogenated
hydrocarbons. Environ Health Perspect 60:159-164.

Hahn, ME. (1998) The aryl hydrocarbon receptor: a comparative perspective. Comp Biochem Physiol 121:23-53.
         9/22/00
                                                 172
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Halperin, W; Vogt, R; Sweeney, MH; et al. (1998) Immunological markers among workers exposed to 2,3,7,8-
tetrachlorodibenzo-p-dioxin. Occup Environ Med 55:742-749.

Hankinson, O. (1995) The aryl hydrocarbon receptor complex. Ann Rev Pharmacol Toxicol 35:307-340.

Hardell, L; Eriksson, M. (1988) The association between STSs and exposure to phenoxyacetic acids: a new case-
referent study. Cancer 62:652-656.

Hardell, L; Sandstrom, A. (1979) Case-control study: soft-tissue sarcomas and exposure to phenoxyacetic acids or
chlorophenols. Br J Cancer 39:711-717.

Harper, N; Connor, K; Steinberg, M; et al. (1994) An enzyme-linked immunosorbent assay (ELISA) specific for
antibodies to TNP-LPS detects alterations in serum immunoglobulins and isotype switching in C57BL/6 and DBA/2
mice exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin and related compounds. Toxicology 92:155-167.

Harrad, S.J.; Jones, K.C. (1992) A source inventory and budget for chlorinated dioxins and furans in the United
Kingdom environment. The Science of the Total Environment 126:89-107.

Haseman,  JK, Johnson, FM. (1996) Analysis of National Toxicology Program rodent bioassay data for
anticarcinogenic effects. MutatRes 350(1):131-141.

Hatch, M.  (1984) Reproductive effects of the dioxins. In: Public health risks of the dioxins. Lowrance, WW, ed.
California: William Kaufmann; pp. 255-275.

Hayes, CL; Spink, D; Spink, B; et al. (1996) 17-beta Estradiol hydroxylation catalyzed by human cytochrome P450
1B1. ProcNat Acad Sci 93:9776-9781.

Hays, SM; Aylward, LL; Karch, NJ; Paustenbach, DJ. (1997) The relative susceptibility of animals and humans to
the carcinogenic hazard posed by exposure to 2,3,7,8-TCDD: an analysis using standard and internal measures of
dose. Chemosphere, 34(5-7): 1507-22.

Hebert, CD; Harris, MW; Elwell, MR; et al.  (1990) Relative toxicity and tumor-promoting ability of
2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), 2,3,4,7,8-pentachlorodibenzofuran (PCDF), and
1,2,3,4,7,8-hexachlorodibenzofuran (HCDF) in hairless mice. Toxicol Appl Pharmacol 102:362-377.

Hemming, H; Eager, Y; Flodstrom, S; et al. (1995) Liver tumour promoting activity of 3,4,5,3',4'-pentachloro-
biphenyl and its interaction with 2,3,7,8-tetrachlorodibenzo-p-dioxin. Eur J Pharmacol 292:241-249.

Hertzman, C;  Teschke, K; Ostry, A; et al. (1997) Mortality and cancer incidence among sawmill workers exposed to
chlorophenate wood preservatives. Am J Publ Health 87(l):71-79.

Hill, AB. (1965) The environment and disease:  association or causation. Proc R Soc Med 58:295-300.

Hill, RN; Crisp, TM; Hurley, PM; et al. (1998) Risk assessment of thyroid follicular cell tumors. Environ Health
Perspect 106(8):447-57.

Hooiveld,  M;  Heederik, D; Bueno de Mesquita, HB. (1996) Preliminary results of the second follow-up of a Dutch
cohort occupationally exposed to phenoxy herbicides, chlorophenols, and contaminants. Organohalogen
Compounds 30:185-189.

Hooiveld,  M;  Heederik, DJJ; Kogevinas, M; et al. (1998) Second follow-up of a Dutch cohort occupationally
exposed to phenoxy herbicides, chlorophenols,  and contaminants. Am J Epidemiol 147(9):891-901.
         9/22/00
                                                 173
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
40
41
42
43
44
45
46
47
48
49
 50
 51
 52
 53
 54
 55
 Homung, MW; Spitsbergen, JM; Peterson, RE. (1999) 2,3,7,8-Tetrachlorodibenzo-/>-dioxin alters cardiovascular
 and craniofacial development and function in  sac fry of rainbow trout (Oncorhynchus mykiss)  Toxicol Sci
 47(1):40-51.

 Huff, JE; Salmon, AG; Hooper, NK; et al. (1991) Long-term carcinogenesis studies on
 2,3,7,8-tetrachlorodibenzo-p-dioxin and hexachlorodibenzo-p-dioxins. Cell Biol Toxicol 7(l):67-94.

 Huisman, M; Koopman-Esseboom, C; Lanting, CI; et al. (1995a) Neurological condition in 18-month-old children
 perinatally exposed to polychlorinated biphenyls and dioxins. Early Hum Dev 43:165-176.

 Huisman, M; Koopman-Esseboom, C; Fidler, V; et al. (1995b) Perinatal exposure to polychlorinated biphenyls and
 dioxins and its effect on neonatal neurological development. Early Hum Dev 41(2): 1 1 1-127.

 Hurley, PM (1998) Mode of carcinogenic action of pesticides inducing thyroid follicular cell tumors in rodents.
 Environ Health Perspect 106(8):437-45.

 Hurst, CH; DeVito, MJ; Setzer, RW; Birnbaum, LS. (2000) Acute administration of 2,3,7,8-tetrachlorodibenzo-
 p-dioxin  (TCDD) in pregnant Long Evans rats: association of measured tissue concentrations with developmental
 effects. Toxicol Sci 53(2):411-20.

 IARC. (1997) IARC monographs on the evaluation of carcinogenic risks to humans. Volume 69. Polychlorinated
 dibenzo-para-dioxins and polychlorinated dibenzofurans. Lyon, France:  IARC.

 Jensen, E; Bolger, PM. (2000) Exposure assessment of dioxins/furans consumed in dairy foods and fish.  Submitted
 for publication to, Food Additives and Contaminants.

 Jensen, E.; Canady, R; Bolger, PM (2000) Exposure assessment for dioxins and furans in seafood and dairy foods in
 the United States, 1998-99. Organohalogen Compounds 47:3 18-321.

 Jirtle, RL; Meyer, SA.  (1991) Liver tumor promotion: effect of phenobarbital on EOF and protein kinase C signal
 transduction and transforming growth factor-beta 1 expression. Dig Dis Sci 36:659-668.

 Jirtle, RL; Meyer, SA;  Brockenbrough, JS. (1991) Liver tumor promoter phenobarbital: a biphasic modulator of
 hepatocyte proliferation. Prog Clin Biol Res 369:209-216.

 Johnson,  RD; Tietge, JE; Botts, S. (1992) Carcinogeniciry of 2,3,7,8-TCDD to Medaka.  The Toxicologist
Johnson, L; Wilker, CE; Safe, SH; et al. (1994) 2,3,7,8-tetrachlorodibenzo-p-dioxin reduces the number, size, and
organelle content of Leydig cells in adult rat testes. Toxicology 89:49-65.

Johnson, KL; Cummings, AM; Birnbaum LS. (1997) Promotion of endometriosis in mice by polychlorinated
dibenzo-/7-dioxins, dibenzofurans, and biphenyls.  Environ Health Perspect 105(7):750-755.

Jung, D; Berg, PA; Edler, L; et al. (1998) Immunologic findings in workers formerly exposed to 2,3,7,8-
tetrachlorodibenzo-p-dioxin and its congeners. Environ Health Perspect 106(2):689-695.

Jusko, WJ. (1995) Pharmacokinetics and receptor-mediated pharmacodynamics of corticosteroids. Toxicology
102:189-196.

Kadlubar, FF; Butler, MA; Kaderlik, RK; et al. (1992) Polymorphisms for aromatic amine metabolism in humans:
relevance for human carcinogenesis. Environ Health Perspect 98:69-74.

Kawajari, K; Nakachi, K; Imai, K; et al. (1993) Germ line polymorphisms of p53 and CYPIA1 genes involved in
human lung cancer. Carcinogenesis 14(6): 1085-1089.
         9/22/00
                                                 174
DRAFT—DO NOT CITE OR QUOTE

-------
 Kayajanian, GM. (1997) Dioxin is a promoter blocker, a promoter, and a net anticarcinogen. Regul Toxicol
 Pharmacol 26(1):134-137 (Review).

 Kayajanian, GM. (1999) Dioxin is a systemic promoter blocker, II. Ecotoxicol Environ Saf 42(2): 103-109.

 Ketchum, NS; Michalek, IE; Burton JE. (1999) Serum dioxin and cancer in veterans of Operation Ranch Hand Am
 J Epidemiol 149(7):630-639.

 Kimmel, GL. (1988) Appendix C. In: A cancer risk-specific dose estimate for 2,3,7,8,-TCDD. U.S. EPA, External
 Review Draft.

 Kitchin, KT; Woods, JS. (1979) 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) effects on hepatic microsomal
 cytochrome  P-448-mediated enzyme activities. Toxicol Appl Pharmacol 47:537-546.

 Kleeman, JM; Moore, RW; Peterson, RE. (1990) Inhibition of testicular steroidogenesis in 2,3,7,8-
 tetrachlorodibenzo-p-dioxin-treated rats: evidence that the key lesion occurs prior to or during pregnenolone
 formation. Toxicol Appl Pharmacol  106:112-125.

 Kociba, RJ; Keeler, PA; Park, GN; et al. (1976) 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD): results of a 13 week
 oral toxicity study in rats. Toxicol Appl Pharmacol 35:553-574.

 Kociba, RJ; Keyes, DG; Beyer, JE; et al.  (1978) Results of a two-year chronic toxicity and oncogeniciry study of
 2,3,7,8-tetrachlorodibenzo-p-dioxin in rats. Toxicol Appl Pharmacol 46:279-303.

 Kogevinas, M; Saracci, R; Winkelmann, R; et al. (1993) Cancer incidence and mortality in women occupationally
 exposed to chlorophenoxy herbicides, chlorophenols and dioxins. Cancer Causes Control 4:547.

 Kogevinas, M; Becher, H; Benn, T; et al. (1997) Cancer mortality in workers exposed to phenoxy herbicides,
 chlorophenols, and dioxin. An expanded and updated international cohort study. Am J Epidemiol 145(12):1061-
 1075.                                                                      -'.'

Kohn, MC; Lucier, GW; Clark, GC; et al. (1993) A mechanistic model of effects of dioxin on gene expression in the
rat liver. Toxicol Appl Pharmacol  120:138-154.

Kohn, MC; Sewall, CH; Lucier, GW; et al. (1996) A mechanistic model of effects of dioxin on thyroid hormones in
the rat. Toxicol Appl Pharmacol 136:29-48.

Koninckx, PR; Braet, P; Kennedy, SH; et al. (1994) Dioxin pollution and endometriosis in Belgium. Hum Reprod
9(6): 1001-1002.

Koopman-Esseboom, C; Weisglas-Kuperus, N; de Ridder, MAJ; et al. (1995b) Effects of PCB/dioxin exposure and
feeding type on the infant's visual recognition memory. Chapter 7 in dissertation entitled: Effects of perinatal
exposure to PCBs and dioxins on early human development Erasmus Universiteit Rotterdam, pp. 107-121.

Koopman-Esseboom, C; Weisglas-Kuperus, N; de Ridder, MAJ; et al. (1996) Effects of polychlorinated
biphenyl/dioxin exposure and feeding type on the infant's mental and psychomotor development. Pediatrics 97:700-
706.

Koopman-Esseboom, C; Huisman, M; Weisglas-Kuperus, N; et al. (1994a) Dioxin and PCB levels in blood and
human milk  in relation to living areas in The Netherlands. Chemosphere 29(9-11):2327-2338.

Koopman-Esseboom, C; Morse, DC; Weisglas-Kuperus, N; et al. (1994c) Effects of dioxins and polychlorinated
biphenyls on thyroid hormone status of pregnant women and their infants. Pediatr Res 36(4):468-73.
9/22/00
175
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Koopman-Esseboom, C; Huisman, M; Touwen, BCL; et al. (1995a) Effects of PCB/dioxin exposure and feeding
type on the infant's visual recognition memory. Chapter 5 in dissertation entitled: Effects of perinatal exposure to
PCBs and dioxins on early human development. Erasmus Universiteit Rotterdam, pp. 75-86.

Koopman-Esseboom, C; Huisman, M; Weisglas-Kuperus, N; et al. (1994b) PCB and dioxin levels in plasma and
human milk of 418 Dutch women and their infants. Predictive value of PCB congener levels in maternal plasma for
fetal and infant's exposure to PCBs and dioxins. Chemosphere 28:1721-1732.

Kreuzer, PE; Csanady, GyA; Baur, C; Kessler, W; et al. (1997) 2,3,7,8-TetrachIorodibenzo-p-dioxin (TCDD) arid
congeners in infants. A toxicokinetic model of human lifetime body burden by TCDD with special emphasis on its
uptake and nutrition. Arch Toxicol 71:383-400.

Kuratsune, M; Ikeda, M; Nakamura, Y; et al. (1988) A cohort study on mortality of Yusho patients: a preliminary
report. In: Unusual occurrences as clues to cancer etiology. Miller, RW; et al., eds. Jpn Sci Soc Press: Tokyo/Taylor
& Francis, Ltd., pp. 61-68.

Kuratsune, M. (1989) Yusho, with reference to Yu-Cheng. In: Halogenated biophenyls, terphenyls, naphthalenes,
dibenzodioxins and related products. Kimbrough, RD; Jensen, AA, eds. 2nd ed. New York: Elsevier Science
Publishers; pp. 3 81 -400.

Kutz, FW; Barnes, DG; Bretthauer, EW; et al. (1990) The International Toxicity Equivalency Factor (I-TEF)
method for estimating risks associated with exposures to complex mixtures of dioxins and related compounds.
Toxicol Environ Chem 26:99-109.

Lahvis, GP; Bradfield, CA. (1998) Ahr null alleles: distinctive or different? Biochem Pharmacol 56(7):781-787.

Lai, ZW; Pineau, T; Esser, C. (1996) Identification of dioxin-responsive elements (DREs) in the 5' regions of
putative dioxin-inducible genes. Chem Biol Interact 100:97-112.

Lampi, P; Hakulinen, T; Luostarinen, T; et al. (1992) Cancer incidence following chlorophenol exposure in  a
community in southern Finland. Arch Environ Health 47(3):167-175.

Landi, MT; Consonni, D; Patterson, DG, Jr.; et al. (1998)  2,3,7,8-Tetrachlorodibenzo-p-dioxin plasma levels in
Seveso 20 years after the accident. Environ Health Perspect 106(5):273-277.

Lathrop, GD; Wolfe, WH; Albanese, RA;  et al. (1984) An epidemiologic investigation of health  effects in Air Force
personnel following exposure to herbicides. Baseline morbidity study results. Brooks Air Force Base, TX: U.S. Air
Force School of Aerospace Medicine, Aerospace Medical Division (unpublished).

Lathrop, GD; Wolfe, WH; Michalek, JE; et al. (1987) An epidemiologic investigation of health effects in Air Force
personnel following exposure to herbicides. First follow-up examination results, January 1985-September 1987.
Brooks Ah- Force Base, TX: U.S. Air Force School of Aerospace Medicine, Aerospace Medical Division
(unpublished).

Lebel, G; Dodin, S; Ayotte, P;  et al. (1998) Organochlorine exposure and the risk of endometriosis. Fertil Steril
69(2):221-228.

Li, X; Johnson, DC; Rozman,  KK. (1995a) Effects of 2,3,7,8-terrachlorodibenzo-p-dioxin (TCDD) on estrous
cyclicity and ovulation in female Sprague-Dawley rats. Toxicol Lett 78:219-222.

Li, X; Johnson, DC; Rozman, KK. (1995b) Reproductive effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in
female rats: ovulation, hormonal regulation and possible mechanism(s). Toxicol Appl Pharmacol 133:321-327.
         9/22/00
                                                  176
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
Liem, AKD; Atuma, S; Becker, W; Darnerud, PO; et al. (2000) Dietary intake of dioxin and dioxin-like PCBs by
the general population often European countries.  Results of EU-SCOOP Task 3.2.5. (Dioxins). Organohalogen
Compounds 48:13-16.

Limbird, LE; Taylor, P. (1998) Endocrine disrupters signal the need for receptor models and mechanisms to inform
policy. Cell 93:157-163.

Longnecker, MP; Michalek, JE. (2000) Serum dioxin level in relation to diabetes mellitus among Air Force veterans
with background levels of exposure. Epidemiology 11:44-48.

Liu, H; Biegel, L; Narasimhan, TR; et al. (1992) Inhibition of insulin-like growth factor-I responses in MCF-7 cells
by  2,3,7,8-tetrachlorodibenzo-p-dioxin and related compounds. Mol Cell Endocrinol 87(l-3):19-28.

Lii, YC; Wong, PN. (1984) Dermatological, medical, and laboratory findings of patients in Taiwan and their
treatments. Am J Ind Med 5:81-115.

Lucier, GW; Tritscher, A; Goldsworthy, T; et al. (1991) Ovarian hormones enhance TCDD-mediated increases in
cell proliferation and preneoplastic foci in a two stage model for rat hepatocarcinogenesis. Cancer Res 51:1391-
1397.

Lucier, GW; Lui, EMK; Lamartiniere, CA. (1979) Metabolic activation/deactivation reactions during perinatal
development. Environ Health Perspect 29:7-16.

Lynge, E. (1998) Cancer incidence in Danish phenoxy herbicide workers, 1947-1993. Environ Health Perspect
106(2): 683-688.

Mably, TA; Moore, RW; Goy, RW; et al. (1992b) In utero and lactational exposure of male rats to 2,3,7,8-
tetrachlorodibenzo-p-dioxin: 2. Effects on sexual behavior and the regulation of luteinizing hormone secretion in
adulthood. Toxicol Appl Pharmacol 114:108-117.

Mably, TA; Bjerke, DL; Moore, RW; et al. (1992c) In utero and lactational exposure of male rats to 2,3,7,8-
tetrachlorodibenzo-p-dioxin: 3. Effects on spermatogenesis and reproductive capability. Toxicol Appl Pharmacol
114:118-126.

Mably, TA; Moore, RW; Peterson, RE. (1992a) In utero and lactational exposure of male rats to 2,3,7,8-
tetrachlorodibenzo-p-dioxin: 1. Effects on androgenic status. Toxicol Appl Pharmacol 114:97-107.

Manz, A; Berger, J; Dwyer, JH; et al.  (1991) Cancer mortality among workers in chemical plant contaminated with
dioxin. Lancet 338:959-964.

Maronpot, RR; Foley, JF; Takahashi,  K; et al. (1993) Dose-response for TCDD promotion of hepatocarcinogenesis
in rats initiated with DEN: histologic, biochemical, and cell proliferation endpoints. Environ Health Perspect
101:634-642.

Martin, JV. (1984) Lipid abnormalities in workers exposed to dioxin. Br J Ind Med 41:254-256.

Matsumura, F. (1994) How important is the protein phosphorylation pathway in the toxic expression of dioxin-type
chemicals? Biochem Pharmacol 48(2):215-224.

Matzke GR; Frye, RF; Early JJ; Straka RJ; Carson SW. (2000) Evaluation of the influence of diabetes mellitus on
antipyrine metabolism and CYP1A2 and CYP2D6 activity. Pharmacotherapy. 20(2): 182-90.

May, G. (1982) Tetrachlorodibenzodioxin: a survey of subjects ten years after exposure. Br J Ind Med 39:128-135.
         9/22/00
                                                 177
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
 50
 51
 52
 53
 54
 Mayani, A; Barel, S; Soback, S; et al. (1997) Dioxin concentrations in women with endometriosis. Hum Reprod
 12:373-375.

 McConnell, EE; Moore, JA; Haseman, JK; et al. (1978) The comparative toxicity of chlorinated dibenzo-/?-dioxins
 in mice and guinea pigs. Toxicol Appl Pharmacol 44:335-356.

 McNulty, WP. (1977) Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin for rhesus monkeys: brief report. Bull
 Environ Contam Toxicol 18:108-109.

 Mebus, CA; Reddy, VR; Piper, WN. (1987) Depression of rat testicular 17-hydroxylase and 17,20-lyase after
 administration of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Biochem Pharmacol 36(5): 1727-1731.

 Michalek, J; Pirkle, J; Caudill, S; Tripathi, R; et al. (1996) Pharmacokinetics of TCDD in veterans of operation
 Ranch hand: 10 year follow-up. J of Tox and Env Epi 47: 209-220.

 Michalek, J.E; Rahe, AJ; Kulkami, P.M; Tripathi, R.C. (1998) Levels of 2,3,7,8-terrachlorodibenzo-p-dioxin in
 1,302 unexposed Air Force Vietnam-era veterans. J Exposure Anal Environ Epid 8:59-64.

 Michalek, JE; Akhtar, FZ; Kiel, JL. (1999a) Serum dioxin, insulin, fasting glucose, and sex hormone-binding
 globulin in veterans of Operation Ranch Hand. J Clin Endocrinol Metab (5): 1540-1543.

 Michalek JE; Ketchum NS; Check IJ. (1999b) Serum dioxin and immunologic response in veterans of Operation
 Ranch Hand. Am J Epidemiol 149:1038-1046.

 Mocarelli, P; Needham, LL; Marocchi, A; et al. (1991) Serum concentrations of 2,3,7,8-tetrachlorodibenzo-p-dioxin
 and test results from selected residents of Seveso, Italy. J Toxicol Environ Health 32:357-366.

 Mocarelli P; Brambilla P; Gerthoux, PM; et al. (1996) Change in sex ratio with exposure to dioxin [letter]. Lancet
 348:409.                                                                   ,.-

 Mocarelli, P; Gerthoux, PM; Ferrari, E; et al. (2000) Paternal concentrations of dioxin and sex ratio of offspring.
 Lancet, 355:1858-1863.

Mocarelli, P; Marocchi, A; Brambilla, P; et al. (1986) Clinical laboratory manifestations of exposure to dioxin in
 children. A six year study of the effects of an environmental disaster near Seveso, Italy. JAMA 256:2687-2695.

Moore, RW; Peterson, RE. (1988) Androgen catabolism and excretion in 2,3,7,8-tetrachlorodibenzo-p-dioxin-
treated rats. Biochem Pharmacol 37:560-562.

Moore, R W; Parsons, J A; Bookstaff, R C; Peterson, RE. (1989) Plasma concentrations of pituitary hormones in
2,3,7,8-tetrachlorodibenzo-p-dioxin-treated male rats. J Biochem Toxicol 4:165-172.

Moore, RW; Bookstaff, RC; Mably, RA; et al. (1991) Differential effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on
responsiveness of male rats to androgens, 17B-estradiol, luteinizing hormone, gonadotropin releasing hormone, and
progesterone. Presented at: Dioxin '91,  11th international symposium on chlorinated dioxins and related
 compounds; Research Triangle Park, NC.

 Moore, RW; Potter, CL; Theobald, HM; et al. (1985) Androgenic deficiency in male rats treated with 2,3,7,8-
 tetrachlorodibenzo-p-dioxin. Toxicol Appl Pharmacol 79:99-111.

 Moses, M; Lilis, R; Crow, KD; et al. (1984) Health status of workers with past exposure to
2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5-trichlorophenoxyacetic acid. Comparison of findings
with and without chloracne. Am J Ind Med 5:161-182.
         9/22/00
                                                  178
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
 Murray, F J; Smith, F A; Nitschke, K D; Humiston, CG; Kociba, RJ; Schwetz, BA. (1979) Three-generation
 reproduction study of rats given 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in the diet. Toxicol Appl Pharmacol
 50:241-252.

 Nagayama, J; Okamura, K; lida, T; et al. (1998) Postnatal exposure to chlorinated dioxins and related chemicals on
 thyroid hormone status in Japanese breast-fed infants.  Chemosphere 3 7(9-12): 1789-1793.

 Nagel, S; Berger, J; Flesch-Janys, D; et al. (1994) Mortality and cancer mortality in a cohort of female workers of a
 herbicide producing plant exposed to polychlorinated dibenzo-p-dioxins and furans. Inform. BiometEpidemiol
 Med. Biol.,25:32-38.

 Narasimhan, TR; Craig, A; Arellano, L; et al. (1994) Relative sensitivities of 2,3,7,8-tetrachlorodibenzo-p-dioxin-
 induced Cypla-1 and Cypla-2 gene expression and immunotoxicity in female B6C3F1 mice. Fundam Appl Toxicol
 23:598-607.

 NAS/NRC (Naitonal Academy of Sciences/National Research Council) . (1983) Risk assessment in the Federal
 Government. Washington, DC: National Academy Press.

 NAS/NRC. (1994) Science and Judgment in Risk Assessment. Washington, DC: National Academy Press.

 Nebert, DW; Petersen, DD; Fornace, AJ Jr. (1990) Cellular responses to oxidative stress: the [Ah] gene battery as a
 paradigm. Environ Health Perspect 88:13-25.

 Needham, LL; Gerthoux, PM; Patterson, DG; et al. (1999) Exposure Assessment: Serum Levels of TCDD in
 Seveso, Italy. Environ Res (A) 80:S200-S206.

 Neuberger, M; Landvoigt, W; Demt, F. (1991) Blood levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin in chemical
 workers after chloracne and in comparison groups. Int Arch Occup Environ Health 63:325-327.

Neubert, R; Color, G; Stahlmann, R; Helge, H; Neubert, D. (1992) Polyhalogenated dibenzo-p-dioxins and
 dibenzofurans and the immune system. 4. Effects of multiple-dose treatment with
2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on peripheral lymphocyte subpopulatibns of a non-human primate
 (Callithrixjacchus). Arch Toxicol 66:250-259.

Nicklas, T.A. (1995) Dietary studies of children:  The Bogalusa Heart Study experience. J Amer Dietetic Assc
95:1127-1133.

Nicklas, T.A.; Webber, L.S.; Srinivasan, S.R.; Berenson, G.S. (1993) Secular trends in dietary intakes and
cardiovascular risk factors in 10-y-old children: the Bogalusa heart study (1973-1988).  Amer J Clin Nut 57:930-
937.

Nicklas, T.A.; Johnson, C.C.; Meyers, L.; Webber, L.S.; et al. (1995) Eating patterns, nutrient intakes, and alcohol
consumption patterns  of young adults: the Bogalusa heart study.  Med Exercise Nut Health 4:316-324.

NTP (National Toxicology Program). (1980) Bioassay of a mixture of 1,2,3,6,7,8-hexachlorodibenzo-p-dioxin and
 1,2,3,7,8,9-hexachlorodibenzo-p-dioxin for possible carcinogenicity (gavage study). Tech. Rept. Ser. No. 198.
Research Triangle Park, NC: U.S. DHHS, PHS.

NTP. (1982a) Bioassay of 2,3,7,8-tetrachlorodibenzo-p-dioxin for possible carcinogenicity (gavage study). Tech.
Rept. Ser. No. 201. Research Triangle Park, NC: U.S. DHHS, PHS.

NTP. (1982b) Bioassay of 2,3,7,8-tetrachlorodibenzo-p-dioxin for possible carcinogenicity (dermal study). Tech.
Rept. Ser. No. 201.  Research Triangle Park, NC: U.S. DHHS, PHS.
         9/22/00
                                                 179
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
NTP. (2000) Report on carcinogens, ninth ed: carcinogen profiles 2000. U.S. Department of Health and Human
Services, Public Health Service, Research Triangle Park, NC.

Olsen, H; Enan, E; Matsumura, F. (1994) Regulation of glucose transport in the NIH 3T3 LI preadipocyte cell line
by TCDD. Environ Health Perspect 102(5):454-458.

Olson, JR; Holscher, MA; Neal, RA. (1980) Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in the Golden Syrian
hamster. Toxicol Appl Pharmacol 55:67-78.

Olson, JR; McGarrigle, BP. (1990) Characterization of the developmental toxicity of 2,3,7,8-TCDD in the Golden
Syrian hamster. Toxicologist 10:313.

Ott, MG; Zober, A. (1996a) Morbidity study of extruder personnel with potential exposure to brominated dioxins
and fiirans. 2. Results of clinical laboratory studies. Occup Environ Med 53:844-846.

Ott, MG; Zober, A. (1996b) Cause specific mortality and cancer incidence among employees exposed to 2,3,7,8-
TCDD after a 1953 reactor accident. Occup Environ Med 53:606-612.

Ott, MG; Messerer, P; Zober, A. (1993) Assessment of past occupational exposure to 2,3,7,8-tetrachlorodibenzo-p-
dioxin using blood lipid analyses. Int Arch Occup Environ Health 65:1-8.

Ott, MG; Zober, A; Germann, C. (1994) Laboratory results for selected target organs in 138 individuals
occupationally  exposed to TCDD. Chemosphere 29:2423-2437.

Ouiddir, A; Planes, C; Femandes, I; et al. (1999) Hypoxia upregulates activity and expression of the glucose
transporter GLUT1 in alveolar epithelial cells. Am J Respir Cell Mol Biol (6):710-718.

Park, J-YK; Shigenaga, MK; Ames, BN. (1996) Induction of cytochrome P4501AI by 2,3,7,8-tetrachlorodibenzo-p-
dioxin or indolo(3,2-b) carbazole is associated with oxidative DNA damage. Proc Nat^cad Sci 93:2322-2327.

Patandin, S; Koopman-Esseboom, C; de Ridder, MA; et al. (1998) Pediatr Res 44(4):538-545.

Patandin, S; Lanting, CI; Mulder, PG; et al. (1999) Effects of environmental exposure to polychlorinated biphenyls
and dioxins on  cognitive abilities in Dutch children at 42 months of age. J Pediatr 134(1):33-41.

Pauwels, A; Cenijn, P; Covaci, A; et al. (1999) Analysis of PCB congeners (by GC-ECD) and dioxin-like toxic
equivalence (by CALUX assay) in females with endometriosis and other fertility problems. Organohalogen
Compounds 44:408-412.

Pazderova-Vejlupkova, J; Nemcova, M; Pickova, J; et al. (1981) The development and prognosis of chronic
intoxication by tetrachlorodibenzo-p-dioxin hi man. Arch Environ Health 36:5-11.

Pesatori, AC; Zocchetti, C; Guercilena, S; et al. (1998) Dioxin exposure and non-malignant health effects: a
mortality study. Occup Environ Med 55(2): 126-131.

Pesatori, AC; Tironi, A; Consonni, A; et al. (1999) Cancer incidence in the Seveso population, 1977-1991.
Organohalogen Compounds 44:411-412.

Peterson, RE; Theobald, HM; Kimmel, GL. (1993) Developmental and reproductive toxicity of dioxins and related
compounds: cross-species comparisons.  Crit Rev Toxicol 23 (3):283-335.

Pluim, HJ; Koppe, JG; Olie, K; et al. (1992) Effects of dioxins on thyroid function in newborn babies. Letter to the
editor. Lancet 339:1303.
         9/22/00
                                                 180
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
Pluim, HJ; de Vijlder, JJM; Olie, K; et al. (1993) Effects of pre- and postnatal exposure to chlorinated dioxins and
furans on human neonatal thyroid hormone concentrations. Environ Health Perspect 101(6):504-508.

Pluim, HJ; Koppe, JG; Olie, K; et al. (1994) Clinical laboratory manifestations of exposure to background levels of
dioxins in the perinatal period. Acta Paediatr 83(6):583-587.

Pohjanvirta, R; Tuomisto, J. (1994) Short-term toxiciry of 2,3,7,8-tetrachlorodibenzo-p-dioxin in laboratory
animals: effects, mechanisms, and animal models. Pharmacol Rev 46(4):483-549.

Poland, AD. (1996) Meeting report.  Receptor-acting xenobiotics and their risk assessment.  Drug Metab Disp
24:1385-1388.

Poland, AD; Knutson, JC. (1982) 2,3,7,8-Tetrachlorodibenzo-p-dioxin and related halogenated aromatic
hydrocarbons: examination of the mechanism of toxicity. Ann Rev Pharmacol Toxicol 22:517-554.

Poland, AD; Palen, D; Glover, E. (1982) Tumor promotion by TCDD in skin of HRS/J mice. Nature
300(5889):271-273.

Portier, CJ; Kohn, MC. (199.6) A biologically-based model for the carcinogenic effects of 2,3,7,8-TCDD in female
Sprague-Dawley rats. Organohalogen Compounds 29:222-227.

Portier, C; Hoel, D; van Ryzin, J. (1984) Statistical analysis of the carcinogenesis bioassay data relating to the risks
from exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. In: Public health risks of the dioxins. Lowrance, W, ed. Los
Altos, NM: W. Kaufmann, pp. 99-120.

Portier, CJ; Sherman, CD; Kohn, M; et al. (1996) Modeling the number and size of hepatic focal lesions following
exposure to 2,3,7,8-TCDD. Toxicol Appl Pharmacol 138:20-30.

Puga, A; Barnes, SJ; Dalton, TP; et al. (2000a) Aromatic hydrocarbon receptor interaction with the retinoblastoma
protein potentiates repression of E2F-dependent transcription and cell cycle arrest.  J Biol Chem 275(4):2943-2950.

Puga, A; Barnes, SJ; Chang, C; et al. (2000b) Activation of transcription factors activator protein-1 and nuclear
factor-kappaB by 2,3,7,8-tetrachlorodibenzo-/)-dioxin. Biochem Pharmacol 59(8):997-1005.

Rao, MS; Subbarao, V; Prasad, JD; et al. (1988) Carcinogenicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in the
Syrian golden hamster. Carcinogenesis 9(9): 1677-1679.

Rappe, C. (1991) Sources of human exposure to CDDs and PCDFs. In: Biological Basis for Risk Assessment of
Dioxin and Related Compounds, Banbury Report No. 35, M. Gallo, R. Scheuplein, and K. van der Heiden (Eds),
Plainview, NY: Cold Spring Harbor Laboratory Press.

Rhile, MJ; Nagarkatti, M; Nagarkatti, PS. (1996) Role of Fas apoptosis and MHC genes in 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD)-induced immunotoxicity of T cells. Toxicology 110:153-167.

Rier, SE; Martin, DC; Bowman, RE; et al. (1993) Endometriosis in rhesus monkeys (Macaca mulatta) following
chronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Fundam Appl Toxicol 21(4):433-441.

Roegner, RH; Grubbs, WD; Lustik, MB; et al. (1991) Air Force Health Study: an epidemiologic investigation of
health effects in Air Force personnel following exposure to herbicides. Serum dioxin analysis of 1987 examination
results. NTIS# AD A-237-516 through AD A-237-524.

Rogan, W. (1989) Yu-Cheng. In: Halogenated biphenyls, terphenyls, naphthalenes, dibenzodioxins and related
products. Kimbrough, RD; Jensen, AA, eds. 2nd ed. New York: Elsevier Pub.; pp. 401-415.
         9/22/00
                                                  181
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
40
41
42
43
44
45
46
47
48
49
 50
 51
 52
 53
 Rogan, WJ; Gladen, BC; Hung, K-L; et al. (1988) Congenital poisoning by polychlorinated biphenyls and their
 contaminants in Taiwan. Science 241:334-338.

 Roman, BL; Sommer, RJ; Shinomiya, K; et al. (1995). In utero and lactational exposure of the male rat to 2,3,7,8-
 tetrachlorodibenzo-^-dioxin: Impaired prostate growth and development without inhibited androgen production.
 Toxicol Appl Pharmacol 134:241-250.

 Romkes, N; Safe, S. (1988) Comparative activities of 2,3,7,8-tetrachlorodibenzo-p-dioxin and progesterone as
 antiestrogens in the female rat uterus. Toxicol Appl Pharmacol 92:368-380.

 Romkes, N; Piskorska-Pliszynska, J; Safe, S. (1987) Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on hepatic and
 uterine estrogen receptor levels in rats. Toxicol Appl Pharmacol 87:306-314.

 Rowlands, JC; Gustafsson, J-A. (1997) Aryl hydrocarbon receptor-mediated signal transduction. Crit Rev Toxicol
 27:109-134.

 Roy, D; Bernhardt, A; Strobel, HW; et al. (1992) Catalysis of the oxidation of steroid and stilbene estrogens to
 estrogen quinone metabolites by the beta-naphthofiavone-inducible cytochrome P450 IA family Arch Biochem
 Biophys 296:450-456.

 Rozman, KK. (1999) Delayed acute toxicity of 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin (HpCDD), after oral
 administration, obeys Haber's rule of inhalation toxicology. Toxicol Sci 49:102-109.

 Rozman, KK. (2000) The role of time  in toxicology or Haber's c x ? product. Toxicol 149:35-42.

 Rozman, KK; Lebofsky, M; Pinson, DM. (2000) Anemia and lung cancer in 1,2,3,4,6,7,8-heptachlorodibenzo-p-
 dioxin (HPCDD)-treated female Sprague-Dawley rats after various single and multiple oral doses Toxicol Sci
 54(1):277.

 Ryan, RP; Sunahara, GI; Lucier, GW; et al. (1989) Decreased ligand binding to the hepatic glucocorticoid and
 epidermal growth factor receptors after 2,3,4,7,8-pentachlorodibenzofuran and 1,2,3,4,7,8-hexachlorodibenzofuran
 treatment of pregnant mice. Toxicol Appl Pharmacol 98(3):454-464.

 Safe, S. (1995a) Human dietary intake of aryl hydrocarbon (Ah) receptor agonists: mass balance estimates of
 exodioxins and endodioxins and implications for health assessment. Organohalogen Compounds 26:7-13.

 Safe, S. (1995b) Modulation of gene expression and endocrine response pathways by 2,3,7,8-tetrachlorodibenzo-p-
dioxin and related compounds. Pharmacol Ther 67(2):247-281.

 Saracci, R; Kogevinas, M; Bertazzi,  P; et al. (1991) Cancer mortality in workers exposed to chlorophenoxy
herbicides and chlorophenols. Lancet 38(3774):1027-1032.

Schantz, SL; Bowman, RE. (1989) Learning in monkeys exposed perinatally to 2,3,7,8-tetrachlorodibenzo-p-dioxin
(TCDD). Neurotoxicol Teratol 11:13-19.

Schantz, SL; Barsotti, DA; Allen, JR. (1979) Toxicological effects produced in nonhuman primates chronically
exposed to fifty parts per trillion 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Toxicol Appl Pharmacol 48(Part 2>
A180.

 Schecter, A, ed. (1994) Dioxins and health. New York: Plenum Press.

Schmidt, JV; Bradfield, CA. (1996) AhR signaling pathways. Ann Rev Cell Dev Biol 12:55-89.
         9/22/00
                                                 182
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
2-1
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Schrenk, D; Buchmann, A; Dietz, K; et al. (1994) Promotion of preneoplastic foci in rat liver with
2,3,7,8-tetrachlorodibenzo-p-dioxin, 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin and a defined mixture of 49
polychlorinateddibenzo-p-dioxins. Carcinogenesis 15:509-515.

Schuur, AG; Boekhorst, FM; Brouwer, A; Visser, TJ. (1997) Extrathyroidal effects of 2,3,7,8-tetrachlorodibenzo-p-
dioxin on thyroid hormone turnover in male Sprague-Dawley rats. Endocrinology 138(9):3727-34

Sewall,  CH; Lucier, GW. (1995) Receptor-mediated events and the evaluation of the Environmental Protection
Agency (EPA) of dioxin risks. Mutat Res 333(1-2):! 11-122 (Review).

Sewall,  CH; Lucier, GW; Tritscher, AM; et al. (1993) TCDD-mediated changes in hepatic epidermal growth factor
receptor may be a critical event in the hepatocarcinogenic action of TCDD. Carcinogenesis 14:1885-1893.

Shimizu, Y; Nakatsuru, Y; Ichinose, M; et al. (2000) Benzo[a]pyrene carcinogenicity is lost in mice lacking the aryl
hydrocarbon receptor. Proc Natl Acad Sci USA 97:779-782.

Slezak, BP; Hatch, GE; DeVito, MJ; et al. (2000) Oxidative stress in female B6C3F1 mice following acute and
subchronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Toxicol Sci, in press.

Smialowicz, RJ; Riddle, MM; Williams, WC; et al. (1994) Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)
on humoral immunity and lymphocyte subpopulations: differences between mice and rats. Toxicol Appl Pharmacol
124:248-256.

Spink, DC; Lincoln, DW, II; Dickerman, HW; et al. (1990) 2,3,7,8-Tetrachlorodibenzo-p-dioxin causes an
extensive alteration of 17p-estradiol metabolism in MCF-7 breast tumor cells. Proc Natl Acad Sci USA 87:6917-
6921.

Squire, RA. (1980) Pathologic evaluations of selected tissues from the Dow Chemical TCDD and 2,4,5-T rat
studies.  Submitted to  Carcinogen Assessment Group, U.S. Environmental Protection Agency on August 15 under
contract no. 68-01-5092.

Steenland, K; Piacitelli, L; Deddens, J; et al. (1999) Cancer, heart disease, and diabetes in workers exposed to
2,3,7,8-tetrachlorodibenzo-p-dioxin. J Natl Cancer Inst 91(9):779-786.

Stephenson, RP. (1956) A modification of receptor theory. Br J Pharmacol 11:379.

Stohs, SJ. (1990) Oxidative stress induced by 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Free Rad Biol Med
9:79-90.

Suskind, RR. (1985) Chloracne, the hallmark of dioxin intoxication. Scand J Work Environ Health 11:165-171.

Suskind, RR; Hertzberg, VS. (1984) Human health effects of 2,4,5-T and its toxic contaminants. JAMA
251:2372-2380.

Sutter, TR; Greenlee, WF.  (1992)  Classification of members of the Ah gene battery. Chemosphere 25:223-226.

Sweeney, A. (1994) Reproductive epidemiology of dioxins. In: Dioxins and health.  Schecter, A, ed. New York:
Plenum Press, pp. 549-558.

Sweeney, MH; Fingerhut, MA; Connally, LB; et al. (1989) Progress of the NIOSH cross-sectional medical study of
workers occupationally exposed to chemicals contaminated with 2,3,7,8-TCDD. Chemosphere 19:973-977.

Sweeney, MH; Calvert, GM; Egeland, GA; et al. (1997-98) Review and update of the results of the NIOSH medical
study of workers exposed to chemicals contaminated with 2,3,7,8-tetra-chlorodibenzo-/?-dioxin. Teratog Carcinog
Mutagen 17(4-5):241-247.
         9/22/00
                                                  183
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Taylor, BL; Zhulin, IB. (1999) PAS domains: internal sensors of oxygen, redox potential, and light. Microbiol Mol
Biol Rev 63(2):479-506.

Teeguarden, JG; Dragan, YP; Singh, J; et al. (1999) Quantitative analysis of dose- and time-dependent promotion of
four phenotypes of altered hepatic foci by 2,3,7,8-tetrachlorodibenzo-p-dioxin in female Sprague-Dawley rats.
ToxicolSci 51:211-223.

Thomas, V.M.; Spiro, T.G. (1995) An estimation of dioxin emissions in the United States. Toxicological and
Environ Chem  50:1-37.

Tian, Y; Ke, S; Denison, MS; et al. (1999) AhR and NF-kappaB interactions, a potential mechanism for dioxin
toxicity. J Biol Chem 274(1):510-515.

Tonn, T; Esser, C; Schneider, EM; et al. (1996) Persistence of decreased T-helper cell function in industrial workers
20 years after exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Environ Health Perspect 104:422-426.

Tritscher, AM; Goldstein, JA; Portier, CJ; et al. (1992) Dose-response relationships  for chronic exposure to 2,3,7,8-
tetrachlorodibenzo-p-dioxin in a rat-tumor promotion model: quantification and immunolocalization of CYPlAland
CYP1A2 in the liver. Cancer Res 52:3436-3442.

Tritscher, AM; Clark, GC; Sewall, C; et al. (1995) Persistence of TCDD-induced hepatic cell proliferation and
growth of enzyme altered foci after chronic exposure followed by cessation of treatment in DEN initiated female
rats. Carcinogenesis 16:2807-2811.

Tritscher, AM; Seacat, AM; Yager, JD; et al. (1996) Increased oxiditative DNA damage in livers of 2,3,7,8-
tetrachlorodibenzo-p-dioxin treated intact but not ovariectomized rats. Cancer Lett 98:219-225.

U.S. EPA. (1980) Risk assessment on (2,4,5-tetrachlorophenoxy) acetic acid [2,4,5-T], (2,4,5-trichlorophenoxy)
propionic acid, and2,3,7,8-tetrachlorodibenzo-j9-dioxin [TCDD]. Washington, DC. --

U.S. EPA. (1985) Health effects assessment document for polychlorinated dibenzo-p-dioxins. Prepared by the Office
of Health and Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, OH, for the
Office of Emergency and Remedial Response, Washington, DC. EPA/600/8-84/014F.

U.S. EPA. (1987) Interim procedures for estimating risks associated with exposures to mixtures of chlorinated
dibenzo-p-dioxins and -dibenzofurans (CDDs and CDFs). EPA/625/3-87/012.

U.S. EPA. (1989a) 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: Risk Assessment
Forum. EPA/625/3-89.016.

U.S. EPA. (1989b) Review of draft documents: a cancer risk-specific dose estimate for 2,3,7,8-TCDD.  Washington,
DC. EPA Science Advisory Board Ad Hoc Dioxin Panel.

U.S. EPA. (199la) Workshop report on toxicity equivalency factors for polychlorinated biphenyls congeners.
EPA/625/3-91/020.

U.S. EPA. (1991b) Guidelines for developmental toxicity risk assessment. Federal Register 57:22888-22938.

U.S. EPA. (1992a) Draft report: A cross species-scaling factor for carcinogen risk assessment based on equivalence
of mg/kg3/4/day. Federal Register 57(109):24152-24173.

U.S. EPA. (1992b) National study of chemical residues in fish. Washington, DC. Office of Science and
Technology. EPA/823-R-02-008.
         9/22/00
                                                 184
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
U.S. EPA. (1995) An SAB Report: A second look at dioxin. EPA-SAB-EC-95-021.

U.S. EPA. (1994) Health assessment document for 2,3,7,5-tetrachlorodibenzo-p-dioxin (TCDD) and related
compounds. External review draft. Prepared by the Office of Health and Environmental Assessment, Office of
Research and Development, Washington, DC. EPA/600/BP-92/001a, b, c. Available from NTIS, Springfield, VA
PB94-205457.

U.S. EPA. (1996) Proposed guidelines for carcinogen risk assessment. Federal Register 61:17960-18011.

U.S. EPA. (1999) Revised proposed guidelines for carcinogen risk assessment.

van Birgelen, AP; Van der Kolk, J; Fase, KM: et al. (1995) Subchronic dose-response study of 2,3,7,8-
tetrachlorodibenzo-p-dioxin in female Sprague-Dawley rats. Toxicol Appl Pharmacol 132:1-13.

van Birgelen, APJM; Diliberto, Devito, MJ; Birnbaum, LS. (1996) Tissue CYP1A1 activity relfects tissue 2,3,7,8-
tetrachlorodibenzo-p-dioxin concentrations. Organohalogen Compounds 29:439-442.

van Birgelen, APJM; Johnson, JD; Fuciarelli, AF; et al. (1999) Dose and time-response of TCDD in Tg.AC mice
after dermal and oral exposure. Dioxin '99: 19th International Symposium on Halogenated Environmental Organic
Pollutants and POPs. (ISBN 88-87772-02-9), Venice, Italy. Organohalogen Compounds 42:235-239.

van den Berg, M; Birnbaum, L; Bosveld, ATC; et al. (1998) Toxic equivalency factors (TEFs) for PCBs, PCDDs,
PCDFs for humans and wildlife. Environ Health Perspect 106(12):775-792.

van den Heuvel, JP; Clark, GC; Kohn, MC; et al. (1994) Dioxin-responsive genes: examination of dose-response
relationships using quantitative reverse transcriptase-polymerase chain reaction. Cancer Res 54:62-68.

van der Plas, SA; Haag-Gronlund, M; Scheu, G; et al. (1999) Induction of altered hepatic foci by a mixture of
dioxin-like compounds with and without 2,2',4,4',5,5'-hexachlorobiphenyl in female Sprague-Dawley rats. Toxicol
Appl Pharmacol 156:30-39.                                                  "'

Vecchi, A; Sironi, M; Canegrati, MA; et al. (1983) Immunosuppressive effects of 2-,3,7,8-tetrachlorodibenzo-p-
dioxin in strains of mice with different susceptibility to induction of aryl hydrocarbon hydroxylase. Toxicol Appl
Pharmacol 68:434-441.

Vena, J; Boffetta, P; Becher,  H; et al. (1998) Exposure to dioxin and nonneoplastic mortality in the expanded IARC
international cohort study of phenoxy herbicide and chlorophenol production workers and sprayers. Environ Health
Perspect 106 Suppl 2:645-653.

Vineis, P; Terracini, B; Ciccone, G; et al. (1986) Phenoxy herbicides and soft-tissue sarcomas in female rice
weeders: a population-based  case-referent study. Scand J Work Environ Health 13:9-17.

Vogel, C; Donat, S; Dohr, O; et al. (1997) Effect of subchronic 2,3,7,8-tetrachlorodibenzo-p-dioxin exposure on
immune system and target gene responses in mice: calculation of benchmark doses for CYP1A1 and CYP1A2
related enzyme activities. Arch Toxicol 71:372-382.

Waern, F; Flodstrom, S; Busk, L; et al. (1991) Relative liver tumour promoting activity and toxicity of some
polychlorinated dibenzo-p-dioxin- and dibenzofuran-congeners in female Sprague-Dawley rats. Pharmacol Toxicol
69:450-458.

Walker, NJ, Kim, A, Lucier,  G, Tritscher, A. (1998) The use of tissue burden as a dose metric for TCDD-inducible
presponses  in rat liver is end point-specific. Organohalogen Compounds 38:337-340.

Walker, NJ; Portier, CJ; Lax, SF; et al. (1999) Characterization of the dose-response of CYP1B1, CYP1A1, and
CYP1A2 in the liver of female Sprague-Dawley rats following chronic exposure to 2,3,7,8-tetrachlorodibenzo-p-
dioxin. Toxicol Appl Pharmacol 154:279-286.
         9/22/00
                                                 185
DRAFT—DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
37
 38
39
40
 Walker, NJ; Tritscher, AM; Sills, RC; et al. (2000) Hepatocarcinogenesis in female Sprague-Dawley rats following
 discontinuous treatment with 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol Sci, in press.

 Webb, KB; Evans, RG; Knudsen, DP; et al. (1989) Medical evaluation of subjects with known body levels of
 2,3,7,8-tetrachlorodibenzo-p-dioxin. J Toxicol Environ Health 28:183-193.

 Weisglas-Kuperus, N; Sas, TCJ; Koopman-Esseboom, C; et al. (1995) Immunologic effects of background prenatal
 and postnatal exposure to dioxins and polychlorinated biphenyls in Dutch infants. Pediatr Res 38:404-410.

 WHO. (2000) International Programme on Chemical Safety: harmonization of approaches to the assessment of
 chemicals. Fact Sheet No.8.

 Wilson, CL; Safe, S. (1998) Mechanisms of ligand-induced aryl hydrocarbon receptor-mediated biochemical and
 toxic responses. Toxicol Pathol 26:657-67 1 .

 Winters, D.L.; Anderson, S.; Lorber, M.; et al. (1998) Trends in dioxin and PCB concentrations in meat samples
 from several decades of the 20th century.  Organohalogen Compounds, Volume 38:75-78.

 Yager, JD; Liehr, JG. (1996) Molecular mechanisms of estrogen carcinogenesis. Ann Rev Pharmacol Toxicol
 36:203-232.

 Yang, JZ; Foster, WG. (1997)  Continuous exposure to 2,3,7,8-tetrachlorodibenzo-/?-dioxin inhibits the growth of
 surgically induced endometriosis in the ovariectomized mouse  treated with high dose estradiol. Toxicol Ind Health
Yang, JZ; Agarwal, S; Foster, WG. (2000)Subchronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin modulates
the pathophysiology of endometriosis in the cynomolgus monkey. Toxicol Sci. 56:374-381.

Zeise, L; Huff, JE; Salmon, AG; et al. (1990) Human risks from 2,3,7,8-tetrachlorodib.enzo-p-dioxin and
hexachlorodibenzo-p-dioxins. In: Advances in modern environmental toxicology, v. 17. Princeton, NJ: Princeton
Scientific Publishing Co., Inc; pp. 293-342.

Zober, A; Messerer,  P; Huber, P. (1990) Thirty-four-year mortality follow-up of BASF employees exposed to
2,3,7,8-TCDD after the 1953 accident. Int Arch Occup Environ Health 62:138-157.

Zober, MA; Ott, MG; Papke, O; et al. (1992) Morbidity study of extruder personnel with potential exposure to
brominated dioxins and furans. I. Results of blood monitoring and immunological tests. Br J Ind Med 49:532-544.

Zober, A; Ott, MG; Messerer, P. (1994) Morbidity follow up study of BASF employees exposed to 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) after a 1953 chemical reactor incident.  Occup Environ Med  5 1 -.479-486.
         9/22/00
                                                 186
DRAFT—DO NOT CITE OR QUOTE

-------