Environmental Protection
           Agency
                  urnce 01 nesearcn ana
                  Development
                  Washington DC 20460
EPA/600/P-00/001Ag
June 2000
External Review Draft
v>EPA
                             Review
                             Draft
                             (Do Not
Exposure and Human   C'te or
Health Reassessment of Quote)
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                                           EPA/600/P-00/001Ag
DO NOT CITE OR QUOTE                            June 2000
                                                 External 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
    2,3,7,8-TetrachIorodibenzo-/7-Dioxin (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.
                                         n

-------
                     TABLE OF CONTENTS - OVERVIEW
                 Exposure and Human Health Reassessment
               of 2,3,7,8-TetrachIorodibenzo-p-Dioxin (TCDD)
                          and Related Compounds
PART I:  ESTIMATING EXPOSURE TO DIOXIN-LIKE COMPOUNDS
       Volume 1:  Executive Summary
       Volume 2:  Sources of Dioxin-Like Compounds in the United States
                 Chapters 1 through 12
       Volume 3:  Properties, Environmental Levels, and Background Exposures
                 Chapters 1 through 6
       Volume 4:  Site-Specific Assessment Procedures
                 Chapters 1 through 8
       Addendum: Revisions since March are included as an addendum to Part I.

PART ILHEALTH ASSESSMENT FOR 2,3,7,8-TETRACHLORODIBENZO-/?-DIOXIN
       (TCDD) AND RELATED COMPOUNDS
       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
       Chapter 9.  Toxicity Equivalence Factors (TEF) for Dioxin and Related Compounds

Part III:  INTEGRATED SUMMARY AND RISK CHARACTERIZATION FOR
        2,3,7,8-TETRACHLORODIBENZO-p-DIOXIN (TCDD) AND RELATED
       COMPOUNDS
                                    in

-------
                                  CONTENTS
1. INTRODUCTION	              	1
        1.1. DEFINITION OF DIOXIN-LIKE COMPOUNDS	'..'.'.'.'.'.'.'.	3
        1.2. TOXICITY EQUIVALENCE FACTORS	4
        1.3. UNDERSTANDING EXPOSURE/DOSE RELATIONSHIPS FOR DIOXIN-
            LIKE COMPOUNDS	7

2. EFFECTS SUMMARY 	10
        2.1. BIOCHEMICAL RESPONSES	11
        2.2. ADVERSE EFFECTS IN HUMANS AND ANIMALS 	14
            2.2.1. Cancer  	14
                  2.2.1.1. Epidemiologic Studies	14
                  2.2.1.2. Animal Carcinogenicity	17
                  2.2.1.3. Other Data Related to Carcinogenesis 	•	19
                  2.2.1.4. Cancer Hazard Characterization 	20
            2.2.2. Reproductive and Developmental Effects	21
                  2.2.2.1. Human	22
                  2.2.2.2. Experimental Animal	24
                  2.2.2.3. Other Data Related to Developmental and Reproductive
                         Effects	27
                  2.2.2.4. Developmental and Reproductive Effects Hazard
                         Characterization 	29
            2.2.3. Immunotoxicity	31
                  2.2.3.1. Epidemiologic Finding	31
                  2.2.3.2. Animal Findings	31
                  2.2.3.3. Other Data Related to Immunologic Effects  	32
                  2.2.3.4. Immunologic Effects Hazard Characterization	33
            2.2.4. Chloracne  	34
            2.2.5. Diabetes	36
            2.2.6. Other Effects	37
                  2.2.6.1. Elevated GOT	37
                  2.2.6.2. Thyroid Function	38
                  2.2.6.3. Cardiovascular Disease	39
                  2.2.6.4. Oxidative Stress	40

3. MECHANISMS AND MODE OF DIOXIN ACTION  	40
        3.1.  MODE VERSUS MECHANISM OF ACTION	41
        3.2.  GENERALIZED MODEL FOR DIOXIN ACTION	42
            3.2.1. The Receptor Concept 	42
            3.2.2. A Framework to Evaluate Mode of Action	44
            3.2.3. Mechanistic Information, Mode of Action, and Risk Assessment	45
                                      IV

-------
4. EXPOSURE CHARACTERIZATION	48
       4.1. SOURCES  	49
           4.1.1. Inventory of Releases	50
           4.1.2. General Source Observations	52
       4.2. ENVIRONMENTAL FATE	54
       4.3. ENVIRONMENTAL MEDIA AND FOOD CONCENTRATIONS 	56
       4.4. BACKGROUND EXPOSURES  	57
           4.4.1. Tissue Levels	57
           4.4.2. Intake Estimates  	58
           4.4.3. Variability in Intake Levels	59
       4.5. POTENTIALLY HIGHLY EXPOSED POPULATIONS OR
           DEVELOPMENTAL STAGES	60
       4.6. ENVIRONMENTAL TRENDS	62

5. DOSE-RESPONSE CHARACTERIZATION	63
       5.1. DOSE METRIC(s) 	66
           5.1.1. Calculations of Effective Dose (ED)  	69
       5.2. EMPIRICAL MODELING OF INDIVIDUAL DATA SETS	70
           5.2.1. Cancer	71
                 5.2.1.1.  Estimates of Slope Factors and Risk at Current Background
                        Body Burdens Based on Human Data	76
                 5.2.1.2.  Estimates of Slope Factors and Risk at Current Background
                        Body Burdens Based on Animal Data	77
                 5.2.1.3.  Estimates of Slope Factors and Risk at Current Background
                        Body Burdens Based on a Mechanistic Model	78
           5.2.2. Noncancer Endpoints 	79
       5.3. MODE-OF-ACTION BASED DOSE-RESPONSE MODELING	80
       5.4. SUMMARY DOSE-RESPONSE CHARACTERIZATION	81

6. RISK CHARACTERIZATION 	82

REFERENCES FOR RISK CHARACTERIZATION 	131

-------
                                  LIST OF TABLES

Table 1-1. The TEF scheme for I-TEQDFa	107
Table 1-2. The TEF scheme for TEQDFP-WHO94a	108
Table 1-3. The TEF scheme for TEQDFP-WHO9ga	109
Table 2-1. Effects of TCDD and related compounds in different animal species  	110
Table 3-1. Early molecular events in response to dioxin	Ill
Table 4-1. Confidence rating scheme	112
Table 4-2. Quantitative inventory of environmental releases of TEQDF-WHO98 in the United
          States	113
Table 4-3. Preliminary indication of the potential magnitude of TEQDF-WHO98 releases from
          "unquantified" (i.e., Category D) sources in reference year 1995	115
Table 4-4. Unquantified sources	116
Table 4-5. Estimates of the range of typical background levels of dioxin-like compounds in
          various environmental media	117
Table 4-6. Estimates of levels of dioxin-like compounds in food 	118
Table 4-7. Background serum levels in the United States 1995-1997  	119
Table 4-8. Adult contact rates and background intakes of dioxin-like compounds	120
Table 4-9. Variability in average daily TEQ intake as a function of age	121
Table 5-1. Serum dioxin levels in the background population and epidemiological cohorts
          (back-calculated) 	122
Table 5-2. Doses yielding 1% excess risk (95% lower confidence bound) based upon 2-year
          animal carcinogenicity studies using simple multistage (Portier et. al, 1984)
          models  	124
                                 LIST OF FIGURES

Figure 1-1. Chemical structure of 2,3,7,8-TCDD and related compounds	125
Figure 2-1. Cellular mechanism for AhR action	126
Figure 2-2. Some of the genes whose expression is altered by exposure to TCDD	127
Figure 4-1. Estimated CDD/CDF I-TEQ emissions to air from combustion sources in the
           United States, 1995	128
Figure 4-2. Comparison of estimates of annual I-TEQ emissions to air (grams I-TEQ/yr) for
           reference years 1987 and 1995	129
Figure 5-1. Dioxin body  burden levels in background populations and epidemiological
           cohorts (back-calculated)	130
                                          VI

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

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

John L. Schaum
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC

Bruce D. Rodan
Senior Health Scientist
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
                                         vn

-------
                                         1.  INTRODUCTION

 1           This document presents an integrated summary of available information related to exposure
 2     to and possible health effects of dioxin and related compounds. It also presents a short risk
 3     characterization, which is a concise statement of dioxin science and the public health
 4     implications of both general population exposures from environmental "background"'and
 5     incremental exposures associated with proximity to sources of dioxin and related compounds.
 6     Even though it summarizes key findings developed in the exposure and health assessment
 7     portions (Parts I and II, respectively) of the Agency's dioxin reassessment, it is meant to be
 8     detailed enough to stand on its own for the average reader. Readers are encouraged to refer to the
 9     more detailed documents for further information on the topics covered here and to see complete
10     literature citations.  These documents are:
11
12     Estimating Exposure to Dioxin-like Compounds: This document, hereafter referred to as Part I,
13     the Exposure Document, is divided into four volumes:  (1) Executive Summary;  (2) Sources of
14     Dioxin in the United States; (3) Properties, Environmental Levels, and Background Exposures;
15     and (4) Site-Specific Assessment Procedures.
16
17     Health Assessment Document for 2,3,7,8-TCDD and Related Compounds:  This document,
18     hereafter referred to as Part II, the Health Document, contains two volumes with nine chapters
19     covering pharmacokinetics, mechanisms of action, epidemiology, animal cancer and various non-
20     cancer effects, toxiciry equivalence factors (TEFs), and dose-response.
21
22           Parts of this integrative summary and risk characterization go beyond individual chapter
23     findings to reach general conclusions about the potential impacts of dioxin-like compounds on
24     human health.  This document specifically identifies issues concerning the  risks that may be
25     occurring in the general population at or near population background exposure levels. It
26     articulates the strengths and weaknesses of the available evidence for possible sources, exposures
27     and health effects, and presents assumptions made and inferences used in reaching conclusions
28     regarding these data.  The final risk characterization provides a synopsis of dioxin science  and its
               'The term "background" exposure has been used throughout this reassessment to describe
        exposure of the general population, who are not exposed to readily identifiable point sources of
        dioxin-like compounds. Most (>95%) of this exposure results from minute amounts of dioxin-
        like compounds being present in dietary fat.
        6/8/00                                      1         DRAFT—DO NOT CITE OR QUOTE

-------
  1      implications for characterizing hazard and risk for use by risk assessors and managers inside and
  2      outside EPA and by the general public.
  3
  4           This document (Part III) is organized as follows:
  5
  6           1.  Introduction - This section describes the purpose/organization of, and the process for
  7           developing, the report; defines dioxin-like compounds in the context of the EPA re-
  8           assessment; and explains the Toxicity Equivalency (TEQ) concept.
  9           2.  Effects Summary - This section summarizes the key findings of the Health Document
 10           and provides links to relevant aspects of exposure, mechanisms, and dose-response.
 11           3.  Mechanisms and Mode of Dioxin Action - This section discusses the key findings on
 12           effects in terms of mode of action. It uses the "Mode-of-Action Framework" recently
 13           described by the WHO/IPCS Harmonization of Approaches to Risk Assessment Project and
 14           contained in the Agency's draft Guidelines for Carcinogen Risk Assessment as the basis for
 15           the discussions.
 16           4.  Exposure Summary - This section summarizes the key findings of the Exposure
 17           Document and links them to the effects, mechanisms, and dose-response characterization.
 18           5.  Dose Response Summary - This section summarizes approaches to dose response that
 19           are found in the Health Document and provides links to relevant aspects of exposure and
 20           effects.
 21           6.  Risk Characterization - This section presents conclusions based on an integration of
 22           the exposure, effects, mechanisms and dose response information.  It also highlights key
 23           assumptions and uncertainties.
 24
 25           The process for developing this risk characterization and companion documents has been
 26      open and participatory. Each of the documents has been developed in collaboration with
 27      scientists from inside and outside the Federal Government. Each document has undergone
 28      extensive internal and external review, including review by EPA's Science Advisory Board
 29      (SAB).  In September 1994, drafts of each document, including an earlier version of this risk
 30      characterization, were made available for public review and comment. This included a 150-day
 31      comment period and  11 public meetings around the country to receive oral and written
 32      comments.  These comments, along with those of the SAB, have been considered in the drafting
 33      of this final  document.  The Dose-Response Chapter of the Health Effects Document underwent
 34      peer review in 1997; an earlier version of this Integrated Summary and Risk Characterization
35      underwent development and review in 1997 and  1998, and comments have been incorporated. In

        6/8/00                                    2        DRAFT—DO NOT CITE OR QUOTE

-------
 1      addition, as requested by the SAB, a chapter on Toxicity Equivalence has been developed and
 2      will undergo review in parallel with this document. When complete, and following final SAB
 3      review, the comprehensive set of background documents and this integrative summary and risk
 4      characterization will be published as final reports and replace the previous dioxin assessments as
 5      the scientific basis for EPA decision-making.
 6
 7      1.1. DEFINITION OF DIOXIN-LIKE COMPOUNDS
 8           As defined in Part I, this assessment addresses specific compounds in the following
 9      chemical classes: polychlorinated dibenzodioxins (PCDDs or CDDs), polychlorinated
10      dibenzofurans (PCDFs or CDFs), polybrominated dibenzodioxins (PBDDs or BDDs),
11      polybrominated dibenzofurans (PBDFs or BDFs), and polychlorinated biphenyls (PCBs), and
12      describes this subset of chemicals as "dioxin-like."  Dioxin-like refers to the fact that these
13      compounds have similar chemical structure, similar physical-chemical properties, and invoke a
14      common battery of toxic responses.  Because of their hydrophobic nature and resistance towards
15      metabolism, these chemicals persist and bioaccumulate in fatty tissues of animals and humans.
16      The CDDs include 75 individual compounds; CDFs include 135 different compounds. These
17      individual  compounds are referred to technically as congeners.  Likewise, the BDDs include 75
18      different congeners and the BDFs include an additional 135 congeners. Only 7 of the 75
19      congeners  of CDDs, or of BDDs, are thought to have dioxin-like toxicity; these are ones with
20      chlorine/bromine substitutions in, at  a minimum, the 2, 3, 7, and 8 positions.  Only 10 of the 135
21      possible congeners of CDFs or of BDFs are thought to have dioxin-like toxicity; these also are
22      ones with substitutions in the 2, 3, 7, and 8 positions. This suggests that 17 individual
23      CDDs/CDFs, and an additional 17 BDDs/ BDFs,  exhibit dioxin-like toxicity. The database on
24      many of the brominated compounds regarding dioxin-like activity has been less extensively
25      evaluated,  and these compounds have not been explicitly considered in this assessment.
26           There are 209 PCB congeners.  Only 12 of the 209 congeners are thought to have dioxin-
27      like toxicity; these are PCBs with 4 or more lateral chlorines with 1 or no substitution in the
28      ortho position.  These compounds are sometimes referred to as  coplanar, meaning that they can
29      assume a flat configuration with rings in the same plane.  Similarly configured polybrominated
30      biphenyls (PBBs) are likely to have similar properties.  However, the database on these
31      compounds with regard to dioxin-like activity has been less extensively evaluated, and these
32      compounds have not been explicitly considered in this assessment. Mixed chlorinated and
33      brominated congeners of dioxins, furans, and biphenyls also exist, increasing the number of
34      compounds potentially considered dioxin-like within the definitions of this assessment.  The
35      physical/chemical properties of each congener vary according to the degree and position of
36      chlorine and/or bromine substitution. Very little is known about occurrence and toxicity of the

        6/8/00                                     3        DRAFT—DO NOT CITE OR QUOTE

-------
  1      mixed (chlorinated and brominated) dioxin, furan, and biphenyl congeners. Again, these
  2      compounds have not been explicitly considered in this assessment. Generally speaking, this
  3      assessment focuses on the 17 CDDs/CDFs and a few of the coplanar PCBs that are frequently
  4      encountered in source characterization or environmental samples. While recognizing that other
  5      "dioxin-like" compounds exist in the chemical classes discussed above (e.g., brominated or
  6      chlorinated/brominated congeners) or in other chemical classes (e.g., halogenated naphthalenes
  7      or benzenes, azo- or azoxybenzenes), the evaluation of less than two dozen chlorinated congeners
  8      is generally considered sufficient to characterize environmental "dioxin."
  9           The chlorinated dibenzodioxins and dibenzofurans are tricyclic aromatic compounds with
10      similar physical  and chemical properties. Certain of the PCBs (the so-called coplanar or mono-
11      ortho coplanar congeners) are also structurally and conformationally similar. The most widely
12      studied of this general class of compounds is 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). This
13      compound, often called simply "dioxin," represents the reference compound for this class of
14      compounds. The structure of TCDD and several related compounds is shown in Figure 1-1.
15      Although sometimes confusing, the term "dioxin" is often also used to refer to the complex
16      mixtures of TCDD and related compounds emitted from sources, or found in the environment or
17      in biological samples. It can also be used to refer to the total TCDD "equivalents" found in a
18      sample. This concept of toxicity equivalence is discussed extensively in Part II, Chapter 9, and is
19      summarized below.
20
21      1.2.  TOXICITY EQUIVALENCE FACTORS
22           CDDs, CDFs, and PCBs are commonly found as complex mixtures when detected in
23      environmental media and biological tissues, or when measured as environmental releases from
24      specific sources. Humans are likely to be exposed to variable distributions of CDDs, CDFs, and
25      dioxin-like PCB congeners that vary by source and pathway of exposures. This complicates the
26      human health risk assessment that may be associated  with exposures to variable mixtures of
27      dioxin-like compounds. In order to address this problem, the concept of toxicity equivalence has
28      been considered and discussed by the scientific community, and toxic equivalency factors (TEFs)
29      have been developed and introduced to facilitate risk assessment of exposure to these chemical
30      mixtures.
31           On the most basic level, TEFs compare the potential toxicity of each dioxin-like compound
32      comprising the mixture to the well-studied and understood toxicity of TCDD, the most toxic
33      member of the group. The background and historical  perspective regarding this procedure is
34      described in detail in Part II, Chapter 9, and in Agency documents (U.S. EPA 1987,1989,
35      1991 a).  This procedure involves assigning individual TEFs to the 2,3,7,8 substituted CDD/CDF
36      congeners and "dioxin-like" PCBs.  To accomplish this, scientists have reviewed the
        6/8/00                                     4        DRAFT—DO NOT CITE OR QUOTE

-------
 1      toxicological databases along with considerations of chemical structure, persistence, and
 2      resistance to metabolism, and have agreed to ascribe specific, "order of magnitude" TEFs for
 3      each dioxin-like congener relative to TCDD, which is assigned a TEF of 1.0. The other
 4      congeners have TEF values ranging from 1.0 to 0.00001.  Thus, these TEFs are the result of
 5      scientific judgment of a panel of experts using all of the available data and are selected to
 6      account for uncertainties in the available data and to avoid underestimating risk. In this sense,
 7      they can be described as "public health conservative" values. To apply this TEF concept, the
 8      TEF of each congener present in a mixture is multiplied by the respective mass concentration and
 9      the products are summed to represent the 2,3,7,8-TCDD Toxic Equivalence (TEQ) of the
10      mixture, as determined by Equation 1 -1.

, 2               TEQ = X ,_„ (Congener, x TEF,) + (congener! x TEF} )+	(congenern x TEFn \    (]. 1)
13      The TEF values for PCDDs and PCDFs were originally adopted by international convention
14      (U.S. EPA, 1989a). Subsequent to the development of the first international TEFs for CDD/Fs,
15      these values were further reviewed and/or revised and TEFs were also developed for PCBs
16      (Ahlborg et al., 1994; van den Berg et al, 1998). A problem arises in that past and present
17      quantitative exposure and risk assessments may not have clearly identified which of three TEF
18      schemes was used to estimate the TEQ. This reassessment introduces a new uniform TEQ
19      nomenclature that clearly distinguishes between the different TEF schemes and identifies the
20      congener groups included in specific TEQ calculations. The nomenclature uses the following
21      abbreviations to designate which TEF scheme was used in the TEQ calculation:
22
23      1.    I-TEQ refers to the International TEF scheme adopted by EPA in 1989 (U.S. EPA, 1989a).
24           See Table  1-1.
25      2.    TEQ-WHO94 refers to the 1994 World Health Organization (WHO) extension of the I-TEF
26           scheme to  include 13 dioxin-like PCBs (Ahlborg et al., 1994). See Table 1-2.
27      3.    TEQ-WHO9g refers to the 1998 WHO update to the previously established TEFs for
28           dioxins, furans, and dioxin-like PCBs (van den Berg  et al., 1998). See Table 1-3.
29
30             The nomenclature also uses subscripts to indicate which family of compounds is included
31      in any specific TEQ calculation. Under this convention, the subscript D is used to designate
32      dioxins, the subscript F to designate furans and the subscript P to designate PCBs. As an
33      example, "TEQDF-WHO98" would be used  to describe a mixture for which only dioxin and furan
34      congeners were determined and where the  TEQ was calculated using the WHO9g scheme. If
35      PCBs had also been determined, the nomenclature would be "TEQDFP-WHO9g." Note that the
36      designations TEQDF-WHO94 and I-TEQDF are interchangeable, as the TEFs for dioxins and furans

        6/8/00                                     5        DRAFT—DO NOT CITE OR QUOTE

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

-------
 1     1.3. UNDERSTANDING EXPOSURE/DOSE RELATIONSHIPS FOR DIOXIN-LIKE
 2     COMPOUNDS
 3            Dose can be expressed as a variety of metrics (e.g., daily intake, serum concentrations,
 4     steady-state body burdens, or area under the plasma concentration versus time curve [AUC]).
 5     Ideally, the best dose metric is that which is directly and clearly related to the toxicity of concern
 6     by a well-defined mechanism. In the mechanism-based cancer modeling for TCDD which will
 7     be discussed later, for instance, instantaneous values of a dose-metric, CYP1A2 or EOF receptor
 8     concentrations are used as surrogates for mutational rates and growth rates within a two-stage
 9     cancer model.  The utility of a particular metric will also depend upon the intended application
10     and the ability to accurately determine this dose metric. For example, if concentration of
11     activated Ah receptors in a target tissue was determined to be the most appropriate dose metric
12     for a particular response in laboratory animals, its utility would be questionable since we
13     presently have no means to determine these values in humans.
14            In this reassessment of the health effects of dioxins, dose is used to understand the
15     animal-to-human extrapolations, comparing human exposure as well as comparing the sensitivity
16     of different toxic responses.  Previous assessments of TCDD have used daily dose as the dose
17     metric and applied either an allometric scaling factor or an uncertainty factor for species
18     extrapolation. The present assessment uses steady-state body burdens as the dose metric of
19     choice. One reason for the change in dose metrics is that recent data demonstrate that the use of
20     either allometric scaling or uncertainty factors underestimates the species differences  in the
21     pharmacokinetic behavior of TCDD and related chemicals. This is due to persistence and
22     accumulation  of dioxins in biological systems and to the large (approximately  100-fold)
23     difference in half-lives between humans and rodents.
24            When  extrapolating across species, steady-state body burden appears to be the most
25     appropriate dose metric. The choice of body burden as the dose metric is based on scientific and
26     pragmatic approaches. As stated earlier, the best dose  metric is that which is directly and clearly
27     related to the toxicity of concern. For dioxins, there is evidence in experimental animals that
28     tissue concentrations of dioxins is an appropriate dose  metric for the developmental,
29     immunological, and biochemical effects of dioxins (Hurst et al., 2000; Van Birgelen et al., 1996;
30     Walker et al.,  1998). Comparing target tissue concentrations of dioxins between animals and
31     humans is impractical. In humans,  the tissues for which we have estimates of the concentration
32     are limited to those that may not be the target tissue of concern, such as serum, blood, or adipose
33     tissue. However, tissue concentrations are directly related to body burdens of dioxins. Therefore,
J4     steady-state body burdens can be used as surrogates for tissue concentrations.
J5

       6/8/00                                      7        DRAFT—DO NOT CITE  OR QUOTE

-------
  1             Body burdens have been estimated through two different methods. Serum, blood, or
  2      adipose tissue concentrations of dioxins are reported as pg/g lipid.  Evidence supports the
  3      assumption that TCDD and related chemicals are approximately evenly distributed throughout
  4      the body lipid. Using the tissue lipid concentrations and the assumption that TCDD is equally
  5      distributed based on lipid content, body burdens are calculated by multiplying the tissue
  6      concentration by the percent body fat composition. One potential problem for estimating body
  7      burdens is the hepatic sequestration of dioxins. In rodents, dioxins accumulate in hepatic tissue
  8      to a greater extent than predicted by lipid content. This sequestration is due to CYP1A2, which
  9      binds dioxins.  There is also evidence in humans that dioxins are sequestered in hepatic tissue.
 10      Estimating body burdens on serum, blood, or adipose tissue concentrations may underpredict true
 11      body burdens of these chemicals. This underprediction should be relatively small. As liver is
 12      approximately 5% of body weight, even a 10-fold sequestration in hepatic tissue compared to
 13      adipose tissue would result in a 50% difference in the body burden estimated using serum, blood,
 14      or adipose tissue concentrations.  In addition, the sequestration is dose-dependent, and at human
 15      background  exposures, hepatic sequestration should not be significant.
 16             A second method for determining body burdens is based on estimates of the daily intake
 17      and half-life of dioxins. Limitations on estimating body burden through this method are
 18      dependent upon the accuracy of the estimates for intake and half-life. Historically, intakes of
 19      dioxins have varied and there is some uncertainty about past exposures.  In addition, little is
 20      known about the half-life of dioxins at different life stages, although there is a relationship
 21      between fat composition and elimination of dioxins.  Finally, depending on the exposure
 22      scenario, using the half-life of TCDD for the TEQ concentrations may result in some
 23      inaccuracies. While the chemicals that contribute most to the total TEQ, such as the
 24      pentachlorodioxins and dibenzofurans and PCS 126, have similar half-lives to TCDD, other
 25      contributors  to the total TEQ have significantly different half-lives. This document uses
 26      pharmacokinetic modeling in a number of places where it is assumed that the 7-year half-life for
 27      TCDD can be applied to the TEQDFP of a mixture of dioxins, furans, and PCBs. The validity of
 28      this assumption was tested in the following way.  First, congener-specific half-lives and intake
 29      rates were identified for each of the dioxin and furan congeners with nonzero TEFs. These half-
 30      lives and intakes were input into a one-compartment, steady-state pharmacokinetic model to get
 31      congener-specific tissue concentrations. The congener-specific tissue levels were summed to get
 32      an overall TEQDF tissue value. Second, the pharmacokinetic model was run using the 7-year
 33      half-life and  total TEQDF intake to get a TEQDF tissue concentration. Both of these modeling
34      approaches yielded very similar TEQDF tissue levels. Although this exercise did not include
35      PCBs (because of lack of half-life estimates), and the congener-specific half-lives for many of the


        6/8/00                                      8        DRAFT—DO NOT CITE OR QUOTE

-------
 1     dioxins and furans have limited empirical support, it provides some assurance that this is a
 2     reasonable approach (see full discussion in Part I, Volume 3, Chapter 4).
 3            Body burdens also have an advantage as a dose metric when comparing occupational or
 4     accidental exposures to background human exposures.  In the epidemiological studies, the
 5     external exposure and the rate of this exposure are uncertain. The only accurate information we
 6     have is on serum, blood, or adipose tissue concentrations.  Because of the long biological half-
 7     life of TCDD, these tissue concentrations of dioxins are better markers of past exposures than
 8     they are of present exposures. Hence, body burdens allow for estimations of exposure in these
 9     occupational and accidentally exposed cohorts. In addition, this dose metric allows us to
10     compare these exposures with those of background human exposures.
11            The use of body burden, for many effects within species and, particularly, for cross-
12     species scaling, appears to provide a better dose metric than daily dose.  There is sufficient
13     scientific evidence to support the use of body burden as a reasonable approximation of tissue
14     concentrations. Future efforts to better understand the dose-response relationships for the effects
15     of dioxin-like chemicals should provide insight into determining better dose metrics for this class
16     of chemicals.
        6/8/00                                      9        DRAFT—DO NOT CITE OR QUOTE

-------
                                       2. EFFECTS SUMMARY

  1             Since the identification of TCDD as a chloracnegen in 1957, more than 5,000
  2      publications have discussed its biological and toxicological properties. A large number of the
  3      effects of dioxin and related compounds have been discussed in detail throughout the chapters in
  4      Part II of this assessment. They illustrate the wide range of effects produced by this class of
  5      compounds. The majority of effects have been identified in experimental animals; some have
  6      also been identified in exposed human populations.
  7             Cohort and case-control studies have been used to investigate hypothesized increases in
  8      malignancies among the various 2,3,7,8-TCDD-exposed populations (Fingerhut et al., 1991a,b;
  9      Steenland et al., 1999; Manz et al., 1991; Eriksson et al., 1990).  Cross-sectional studies have
10      been conducted to evaluate the prevalence or extent of disease in living 2,3,7,8-TCDD-exposed
11      groups (Suskind and Hertzberg, 1984; Moses et al.,  1984; Lathrop et al., 1984, 1987; Roegner et
12      al., 1991; Grubbs et al. 1995; Sweeney et al., 1989; Centers for Disease Control Vietnam
13      Experience Study,  1988; Webb et al., 1989; Ott and Zober, 1994).  The limitations of the cross-
14      sectional study design for evaluating hazard and risk is discussed in Part II, Chapter 7b. Many of
15      the earliest studies  were unable to define exposure-outcome relationships owing to a variety of
16      shortcomings,  including small sample size, poor participation, short latency periods, selection of
17      inappropriate controls, and the inability to quantify exposure to 2,3,7,8-TCDD or to identify
18      confounding exposures. In more recent analyses of cohorts (NIOSH, Hamburg) and cross-
19      sectional studies of U.S. chemical workers (Sweeney et al., 1989), U.S. Air Force Ranch Hand
20      personnel (Roegner et al., 1991; Grubbs et al., 1995), and Missouri residents (Webb et al., 1989),
21      serum or adipose tissue levels of 2,3,7,8-TCDD were measured to evaluate 2,3,7,8-TCDD-
22      associated effects in exposed populations. The ability to measure tissue or serum levels of
23      2,3,7,8-TCDD for all or a large sample of the subjects confirmed exposure to 2,3,7,8-TCDD and
24      permitted the investigators to test hypothesized dose-response relationships.
25             A large number of effects of exposure to TCDD and related compounds have been
26      documented in the  scientific literature.  Although many effects have been demonstrated in
27      multiple species (see Table 2-1), other effects may be specific to the species in which they are
28      measured and may  have limited relevance to the human situation. Although this is an important
29      consideration for characterizing potential hazard,  all observed effects may be indicative of the
30      fundamental level at that dioxin produces its biological impact and illustrate the multiple
31      sequelae that are possible  when primary impacts are at the level of signal transduction and gene
32      transcription. Even though not all observed effects may be characterized as "adverse" effects
33      (i.e., some may be adaptive and of neutral consequence), they represent a continuum of response
34      expected  from  the fundamental changes in biology caused by exposure to dioxin-like

        6/8/00                                     10       DRAFT—DO NOT CITE OR QUOTE

-------
 1      compounds. As discussed in the following sections, the dose associated with this plethora of
 2      effects is best compared across species using a common measurement unit of body burden of
 3      TCDD and other dioxin-like compounds, as opposed to the level or rate of exposure/intake.
 4            The effects discussed in the following sections are focused on development of an
 5      understanding of dioxin hazard and risk. This discussion is by its nature selective of findings
 6      that inform the risk assessment process.  Readers are referred to the more comprehensive
 7      chapters for further discussion of the epidemiologic and toxicologic database.
 8
 9      2.1. BIOCHEMICAL RESPONSES (Cross reference: Part II, Chapters 2,3, and 8)
10            As described later in Section 3, mechanistic studies can reveal the biochemical pathways
11      and types of biological  events that contribute to adverse effects from exposure to dioxin-like
12      compounds. For example, much evidence indicates that TCDD acts via an intracellular protein
13      (the aryl hydrocarbon receptor, AhR), which is a ligand-dependent transcription factor that
14      functions in partnership with a second protein (known as the Ah receptor nuclear translocator,
15      Arnt). Therefore, from a mechanistic standpoint, TCDD's adverse effects appear likely to reflect
16      alterations in gene expression that occur at an inappropriate time and/or for an inappropriate
17      length of time. Mechanistic studies also indicate that several other proteins contribute to TCDD's
18      gene regulatory effects  and that the response to TCDD probably involves a relatively complex
19      interplay between multiple genetic and environmental factors. This model is illustrated in Figure
20      2-1 (from Part II, Chapter 2).
21            Comparative data from animal and human cells and tissues suggest a strong qualitative
22      similarity across species in response to dioxin-like chemicals. This further supports the
23      applicability to humans of the generalized model of early events in response  to dioxin exposure.
24      These biochemical and biological responses are sometimes considered adaptive and are often not
25      considered adverse in and of themselves. However, many of these biochemical changes are
26      potentially on a continuum of dose-response relationships, which leads to adverse responses. At
27      this time, caution must  be used when describing these events as adaptive.
28            If, as we can infer from the evidence, TCDD and other dioxin-like compounds operate
29      through these mechanisms, there are constraints  on the possible models that can plausibly
30      account for dioxin's biological effects and also on the assumptions used during the risk
31      assessment process. Mechanistic knowledge of dioxin action may also be useful in other ways.
32      For example, a further understanding of the  ligand specificity and structure of the Ah receptor
33      will likely assist in the identification of other chemicals to which humans are exposed that may
34      either add to, synergize, or antagonize the toxicity of TCDD  and other dioxin-like compounds.
35      Knowledge of genetic polymorphisms that influence TCDD responsiveness may also allow the
36      identification of individuals at particular risk from exposure to dioxin.  In addition, knowledge of

        6/8/00                                     11        DRAFT—DO NOT  CITE OR QUOTE

-------
  1      the biochemical pathways that are altered by dioxin-like compounds may help in the
  2      development of drugs that can prevent dioxin's adverse effects.
  3             As described in Part II, Chapter 2, biochemical and genetic analyses of the mechanisms
  4      by which dioxin modulates particular genes have revealed the outline of a novel regulatory
  5      system whereby a chemical signal can alter cellular regulatory processes. Future studies of
  6      dioxin action have the potential to provide  additional insights into mechanisms of mammalian
  7      gene regulation that are of relatively broad  interest. Additional perspectives on dioxin action can
  8      be found in several recent reviews (Birnbaum, 1994a,b; Schecter, 1994; Hankinson, 1995;
  9      Schmidt and Bradfield, 1996; Rowlands and Gustafsson, 1997; Gasiewicz, 1997; Harm, 1998;
10      Denison et al., 1998; Wilson and Safe, 1998).
11             The ability of TCDD and other dioxin-like compounds to modulate a number of
12      biochemical parameters in a species-, tissue-, and temporal-specific manner is well recognized.
13      Despite the ever-expanding list of these responses over the past 20 years and the elegant work on
14      the molecular mechanisms mediating some of these, there still exists a considerable gap between
15      our knowledge of these changes and the degree to which they are related to the more complex
16      biological and toxic endpoints elicited by these chemicals. A framework for considering these
17      responses in a mode-of action context is discussed later in this document.
18             TCDD-elicited activation of the Ah receptor has been clearly shown to mediate altered
19      transcription of a number of genes, including several oncogenes and those encoding growth
20      factors, receptors, hormones, and drug-metabolizing enzymes.  Figure 2-2 provides an
21      illustrative list of gene products shown to be mediated by TCDD. Although this list is not meant
22      to be exhaustive, it demonstrates the range  of potential dioxin impacts.
23             As discussed in Volume 2, Chapter 2, it is possible that the TCDD-elicited alteration of
24      activity of these genes may occur through a variety of mechanisms, including signal transduction
25      processes. These alterations in gene activity may be secondary to other biochemical events that
26      may be directly regulated transcriptionally  by the AhR. Some of the changes may also occur by
27      post-transcriptional processes such as mRNA stabilization and altered phosphorylation (Gaido et
28      al., 1992; Matsumura, 1994).   Thus, the molecular mechanisms by which many, if not most, of
29      the biochemical processes discussed herein are altered by TCDD treatment remain to be
30      determined. Nevertheless, it is presumed, based on the cumulative evidence available, that all of
31      these processes are mediated by the binding of TCDD to the AhR. Although the evidence for the
32      involvement of the AhR in all of these processes has not always been ascertained,
33      structure-activity relationships, genetic data, and reports from the use of biological  models like
34      "knockout" mice that are lacking the Ah receptor (AhR"A) are consistent with the involvement of
35      the AhR as the initial step leading to many  of these biochemical alterations. In fact, for every


        6/8/00                                    12        DRAFT—DO NOT CITE OR QUOTE

-------
 1     biochemical response that has been well studied, the data are consistent with the particular
 2     response being dependent on the AhR.
 3            The dioxin-elicited induction of certain drug-metabolizing enzymes such as C YP1A1,
 4     CYP1A2, and CYP1B1 is clearly one of the most sensitive responses observed in a variety of
 5     different animal species including humans, occurring at body burdens as low as 1-10 ng
 6     TCDD/kg in animals (see Part II, Chapter 8).  These and other enzymes are responsible for the
 7     metabolism of a variety of exogenous and endogenous compounds. Several lines  of
 8     experimental evidence  suggest that these enzymes may be responsible for either enhancing or
 9     protecting against (depending on the compounds and experimental system used) toxic effects of a
10     variety of agents, including known  carcinogens as well as endogenous substrates such as
11     hormones. Several reports (Kadlubar et al., 1992; Esteller et al., 1997; Ambrosone et al., 1995;
12     Kawajiri et al., 1993) provide evidence that human polymorphisms in CYPIA1 and CYPIA2 that
13     result in higher levels of enzyme are associated with increased susceptibility to colorectal,
14     endometrial, breast, and lung tumors. Also, exposure of AhR-deficient ("knockout") mice to
15     benzo[a]pyene (BaP) results in no tumor response, suggesting a key role for the AhR, and
16     perhaps, CYPIA1 and CYPIA2, in BaP carcinogenesis (Dertinger et al., 1998;  Shimizu et al.,
17     2000). Modulation of these enzymes by dioxin may play a role in chemical carcinogenesis.
18     However, the exact relationship between the induction of these enzymes and any toxic endpoint
19     observed following dioxin exposure has not been clearly established.
20            As with certain of the cytochrome P450 isozymes, there does not yet exist  a precise
21     understanding of the relationships between the alteration of specific biochemical processes and
22     particular toxic responses observed in either experimental animals or humans exposed to the
23     dioxins.  This is due predominantly to our incomplete understanding of the complex and
24     coordinate molecular, biochemical, and cellular interactions that regulate tissue processes during
25     development and under normal homeostatic conditions. Nevertheless, a further understanding of
26     these processes and how TCDD may interfere with them remains an important goal that would
27     greatly assist in the risk characterization process. In particular, knowledge of the causal
28     association of these responses coupled with dose-response relationships may lead  to a better
29     understanding of sensitivity to various exposure levels of the dioxin-like compounds.
30            In contrast to what is known about the P450 isozymes, there exists some evidence from
31     experimental animal data to indicate that the alteration of certain other biochemical events might
32     have a more direct relationship to sensitive toxic responses observed following TCDD exposure.
33     Some of these may be relevant to responses observed in humans, and further work in these areas
34     is likely to lead to data that would assist in the risk characterization process. For example,
35     changes in epidermal growth factor (EOF) receptor have been observed in tissues  from
36     dioxin-exposed animals and humans (see Part II, Chapters 3 and 6 ).  EOF and its receptor

       6/8/00                                     13       DRAFT—DO NOT CITE OR QUOTE

-------
  1      possess diverse functions relevant to cell transformation and tumorigenesis, and changes in these
  2      functions may be related to a number of dioxin-induced responses including neoplastic lesions,
  3      chloracne, and a variety of reproductive and developmental effects .  Likewise, the known ability
  4      of TCDD to directly or indirectly alter the levels and/or activity of other growth factors and
  5      hormones, such as estrogen, thyroid hormone, testosterone, gonadotropin-releasing hormone and
  6      their respective receptors, as well as enzymes involved in the control of the cell cycle (Safe,
  7      1995), may affect growth patterns in cells/tissues, leading to adverse consequences.  In fact, most
  8      of the effects that the dioxins produce at the cellular and tissue levels are due not to cell/tissue
  9      death but to altered growth patterns (Bimbaum, 1994b). Many of these may occur at critical
 10      times in development and/or maturation and thus may be irreversible.
 11            From this brief discussion and that detailed in Part II, Chapters 2 and 8, it seems clear that
 12      much work needs to be done to clarify the exact sequence and interrelations of those biochemical
 13      events altered by TCDD and how and at what point they might lead to irreversible biological
 14      consequences.  Nevertheless, it is important to recognize that many of the biochemical and
 15      biological changes observed are consistent with the notion that TCDD is a powerful growth
 16      dysregulator.  This notion may play a considerable role in the risk characterization process by
 17      providing a focus on those processess, such as development, reproduction, and carcinogenesis,
 18      that are highly dependent on coordinate growth regulation.  Further understanding of these
 19      biochemical events in humans may provide useful biomarkers of exposure and responsiveness.
 20      The use  of these potential biomarkers may subsequently improve our understanding of the
 21      variation of responsiveness within an exposed population.
 22
 23      2.2.  ADVERSE EFFECTS IN HUMANS AND ANIMALS
 24      2.2.1. Cancer (Cross Reference: Volume 2, Chapters 6,7, and 8)
 25      2.2.1.1.  Epidemiologic Studies
 26            Since the last formal U.S. EPA review of the human database relating to the
 27      carcinogenicity of TCDD and related compounds in 1988, a number of new follow-up mortality
 28      studies have been completed. This body of information is described in Part II, Chapter 7, of this
 29      assessment and has recently been published as part of an I ARC Monograph (1997) and the
 30      ATSDR ToxProfile (ATSDR, 1999). Among the most important of these are the studies of 5,172
 31      U.S.  chemical manufacturing workers by Fingerhut et al. (1991a) and Steenland et al. (1999)
 32      from NIOSH and an independent study by Aylward et al. (1996); a study of 2,479 German
 33      workers  involved in the  production of phenoxy herbicides and chlorophenols by Becher et al.
 34      (1996, 1998) and by others in separate publications (Manz et al., 1991; Nagel et al., 1994;
 35      Flesch-Janys et al.,  1995, 1998); a study of more than 2,000 Dutch workers in two plants
36      involved in the synthesis and formulation of phenoxy herbicides and chlorophenols (Bueno de

        6/8/00                                    14       DRAFT—DO NOT CITE OR QUOTE

-------
 1      Mesquita et al., 1993) and subsequent follow-up and expansion by Hooiveld et al., 1998);  a
 2      smaller study of 247 workers involved in a chemical accident cleanup by Zober et al. (1990) and
 3      subsequent follow-up (Ott and Zober, 1996b); and an international study of more than 18,000
 4      workers exposed to phenoxy herbicides and chlorophenols by Saracci et al. (1991), with
 5      subsequent follow-up and expansion by Kogevinas et al. (1997).  Although uncertainty remains
 6      in interpreting these studies because not all potential confounders have been ruled out and
 7      coincident exposures to other carcinogens are likely, all provide support for an association
 8      between exposure to dioxin and related compounds and increased cancer mortality.  One of the
 9      strengths of these studies is that each has  some exposure information that permits an assessment
10      of dose response.  Some of these data have, in fact, served as the  basis for fitting the risk models
11      in Chapter 8. In addition, limited results have been presented on  the non-occupational Seveso
12      cohort (Bertazzi et al., 1993, 1997) and on women exposed to chlorophenoxy herbicides,
13      chlorophenols, and dioxins (Kogevinas et al., 1993).  Although these two studies have
14      methodologic shortcomings that are described in Chapter 7, they  provide findings, particularly
15      for exposure to women, that warrant additional follow-up.
16            Increased risk for all cancers combined was a consistent finding  in the occupational
17      cohort studies. Although the increase was generally low (20%-50%), it  was highest in
18      subcohorts with presumed heaviest exposure. Positive  dose-response trends in the German
19      studies and increased risk in the longer duration U.S. subcohort and the most heavily exposed
20      Dutch workers support this view.
21            One of the earliest reported associations between exposure to dioxin-like compounds in
22      dioxin-contaminated phenoxy herbicides and increased cancer risk involved an increase in soft
23      tissue sarcomas (Hardell and Sandstrom,  1979; Eriksson et al., 1981; Hardell and Eriksson, 1988;
24      Eriksson et al., 1990). In this and other recent evaluations of the  epidemiologic database, many
25      of the earlier epidemiological studies that suggested an association with soft tissue sarcoma are
26      criticized for a variety of reasons. Arguments regarding selection bias, differential exposure
27      misclassification, confounding, and chance in each individual study have been presented in the
28      scientific literature, which increases uncertainty around this association. Nonetheless, the
29      incidence of soft tissue sarcoma is elevated in several of the most recent studies (Bertazzi et al.,
30      1993; 1997,  1999; Fingerhut et al., 1991a; Hertzman et al., 1997; Kogevinas et al., 1997; Lampi
31      et al., 1992; Lynge, 1998; Pesatori et al., 1999; Saracci et al., 1999; Vinels et al., 1986),
32      supporting the findings from previous studies.  The fact that similar results were  obtained in
33      independent  studies of differing design and evaluating populations exposed to dioxin-like
34      compounds under varying conditions, along with the rarity of this tumor type, weighs in favor of
35      a consistent and real association.

        6/g/OO                                     15        DRAFT—DO NOT CITE OR QUOTE

-------
 1             In addition to soft tissue sarcoma, other cancer sites have been associated with exposure
 2      to dioxin.  Excess respiratory cancer was noted by Fingerhut et al. (1991a), Zober et al. (1994),
 3      and Manz et al. (1991). These results are also supported by significantly increased mortality
 4      from lung and liver cancers subsequent to the Japanese rice oil poisoning accident where
 5      exposure to high levels of PCDFs and PCBs occurred (Kuratsune et al., 1988; Kuratsune, 1989).
 6      Again, while smoking  as a confounder cannot be totally eliminated as a potential explanation of
 7      the occupational  studies results, analyses (Fingerhut, 1991b; Ott and Zober, 1996b) conducted to
 8      date suggest that smoking is not likely to explain the entire increase in lung cancer and may even
 9      suggest synergism between occupational exposure to dioxin and smoking.  These analyses have
10      not been deemed entirely satisfactory by some reviewers of the literature. The question of
11      confounding exposures, such as asbestos and other chemicals, in addition to smoking, has not
12      been entirely ruled out and must be considered as potentially adding to the observed increases.
13      Although increases of cancer at other sites (e.g., non-Hodgkin's lymphoma, stomach cancer)
14      have been reported (see Part II, Chapter la), the data for an association with exposure to
15      dioxin-like chemicals are less compelling.
16             As mentioned above, both past and more recent human studies have focused on males.
17      Although males comprise all the case-control studies and the bulk of the cohort study analyses,
18      animal and mechanism studies suggest that males and females might respond differently to
19      TCDD. There are now, however, some limited data suggesting carcinogenic responses
20      associated with dioxin  exposure in females.  The only reported female cohort with good TCDD
21      exposure surrogate information was that of Manz et  al. (1991), which had a borderline
22      statistically significant increase in breast cancer.  Although Saracci et al. (1991) did report
23      reduced female breast and genital organ cancer mortality, this was based on few observed deaths
24      and on chlorophenoxy  herbicide, rather than TCDD, exposures. In the later update and
25      expansion of this cohort Kogevinas et al. (1997) provided evidence of a reversal of this deficit
26      and produced a borderline significant excess risk of breast cancer in females.  Bertazzi et al.
27      (1993,1997,1998) reported nonsignificant deficits of breast cancer and endometrial cancer in
28      women living in  geographical areas around Seveso contaminated by dioxin. Although
29      Kogevinas et al. (1993) saw an increase in cancer incidence among female workers most likely
30      exposed to TCDD, no increase in breast cancer was  observed in his small cohort. In sum, TCDD
31      cancer experience for women may differ from that of men, but currently there are few data.
32      Because both laboratory animal data and mechanistic inferences suggest that males and females
33      may respond differently to the carcinogenic effects of dioxin-like chemicals, further data will be
34      needed to address this question of differential response between sexes, especially to hormonally
35      mediated tumors. No epidemiological data are available to address the question of the potential
36      impact of exposure to dioxin-like compounds on childhood cancers. However, recent studies of

        6/8/00                                    16        DRAFT—DO NOT CITE OR QUOTE

-------
 1      Brown et al. (1998) demonstrate that prenatal exposure of rats enhances their sensitivity as adults
 2      to chemical carcinogenesis.
 3             As discussed above and based on the analysis of the cancer epidemiology data as
 4      presented in Part II, Chapters 7 and 8, TCDD and, by inference, other dioxin-like compounds are
 5      described as potentially multisite carcinogens in more highly exposed human populations that
 6      have been studied, consisting primarily of adult males. Although uncertainty remains, the cancer
 7      findings in the epidemiologic literature are generally consistent with results from studies of
 8      laboratory animals where dioxin-like compounds have clearly been identified as multisite
 9      carcinogens. In addition, the findings of increased risk at multiple  sites appear to be plausible
10      given what is known about mechanisms of dioxin action, and the fundamental level at which it
11      appears to act in target tissues. While several studies exhibit a positive trend in dose-response
12      and have been the subject of empirical risk modeling (Becher et al., 1998), the epidemiologic
13      data alone provide little insight into the shape of the dose-response curve below the range of
14      observation in these occupationally exposed populations. This issue will be further discussed in
15      Section 5.2.1. The contribution of cancer epidemiology to overall cancer hazard and risk
16      characterization is discussed in Section 6.
17
18      2.2.1.2. Animal Carcinogenicity (Cross reference, Part II: Chapters 6 and 8)
19            An extensive database on the carcinogenicity of dioxin and  related compounds in
20      laboratory studies exists and is described in detail in Chapter 6. There is adequate evidence that
21      2,3,7,8-TCDD is a carcinogen in laboratory animals based on long-term bioassays conducted in
22      both sexes of rats and mice (U.S. EPA, 1985; Huff et al., 1991; Zeise et al., 1990; IARC, 1997).
23      All studies have produced positive results, leading to conclusions that TCDD is a multistage
24      carcinogen increasing the incidence of tumors at sites distant from the site of treatment and at
25      doses well below the maximum tolerated dose.  Since this issue was last reviewed by the Agency
26      in 1988, TCDD has been shown to be a carcinogen in hamsters (Rao et al., 1988), which are
27      relatively resistant to the lethal effects of TCDD. Other preliminary data have also shown TCDD
28      to be a liver carcinogen in the small fish Medaka (Johnson et al., 1992). Few attempts have been
29      made to demonstrate the carcinogenicity of other dioxin-like compounds.  Other than a mixture
30      of two isomers of hexachlorodibenzodioxin (HCDDs), which produced liver tumors in both
31      sexes of rats and mice (NTP,  1980) when given by the gavage route, but not by the dermal route
32      in Swiss mice (NTP, 1982a,b) and a recent report (Rozman et al., 2000) attributing lung  cancer in
33      female rats to gavage exposures of l,2,3,4,6,7,8-heptachlorodibenzo-p-dioxi(HpCDD), neither
34      the more highly chlorinated PCDDs/ PCDFs nor the co-planar PCBs have been studied in
35      long-term animal cancer bioassays. However, it is generally recognized that these compounds
36      bioaccumulate and exhibit toxicities similar to TCDD and are, therefore, also likely to be

        6/g/OO                                     17       DRAFT—DO NOT CITE OR QUOTE

-------
  1      carcinogens (U.S. EPA, 1989b).  The National Toxicology Program is currently testing the
  2      relative carcinogenic potency of four dioxin-like congeners (PeCDF, PeCDD, and PCS 118 and
  3      PCB 126), both alone and in combination.  Because no chronic animal bioassays are available on
  4      these compounds, these data, when they are available, should add significantly to our certainty
  5      regarding the carcinogenicity of these dioxin-like congeners.
  6             In addition to the demonstration of TCDD as an animal carcinogen in long-term cancer
  7      bioassays, a number of dioxin-like PCDDs and PCDFs, as well as several PCBs, have been
  8      demonstrated to be tumor promoters in two-stage (initiation-promotion) protocols in rodent liver,
  9      lung, and skin. These studies are described in some detail in Part II, Chapter 6. In that Chapter,
 10      TCDD is characterized as a nongenotoxic carcinogen because it is negative in most assays for
 11      DNA damaging potential, as a potent "promoter," and as a  weak initiator or noninitiator in two-
 12      stage initiation-promotion (I-P) models for liver and for skin.
 13             The liver response is characterized by increases in altered hepatocellular foci (AHF),
 14      which are considered to be preneoplastic lesions because increases in AHFs are associated with
 15      liver cancer in rodents.  The results of the multiple I-P studies enumerated in Figure 6-8 in Part
 16      II, Chapter 6, have been interpreted as showing that induction of AHFs by TCDD is dose-
 17      dependent (Maronpot et al., 1993; Teegarden et al., 1999), are exposure-duration dependent
 18      (Dragan et al., 1992; Teegarden et al., 1999; Walker et al., 2000), and are partially reversible
 19      after cessation of treatment (Dragan et al., 1992; Tritscher et al.,  1995; Walker et al., 2000).
 20      Other studies indicate that other dioxin-like compounds have the ability to induce AHFs.  These
 21      studies show that the compounds demonstrate a rank-order of potency for AHF  induction that is
 22      similar to that for CYP1A1 (Flodstrom and Ahlborg,  1992; Waern et al., 1991; Schrenk et al.,
 23      1994).  Non-ortho substituted, dioxin-like PCBs also  induce the development of AHFs according
 24      to their potency to induce CYP1A1 (Hemming et al.,  1995; van der Plas et al., 1999).  It is
 25      interesting to note that liver I-P studies carried out in ovariectomized rats demonstrate the
 26      influence that the intact hormonal system has on AHF development. AHF  are significantly
 27      reduced in the livers of ovariectomized female rats (Graham et al., 1988; Lucier et al.,  1991).
 28             I-P studies on skin have demonstrated that TCDD is a potent tumor promoter in mouse
 29      skin as well as rat liver. Early studies demonstrated that TCDD is at least two orders of
 30      magnitude more potent than the "classic" promoter tetradecanoyl phorbol acetate (TPA) (Poland
 31      et al., 1982); that TCDD skin tumor promotion is AhR dependent (Poland and Knutsen, 1982);
 32      that TCDD  had weak or no initiating activity in the skin system (DiGiovanni et al., 1977); and
 33      that TCDD's induction of drug-metabolizing enzymes is  associated with both metabolic
 34      activation and deactivation as described by Lucier et al. (1979). More recent studies show that
 35      the skin tumor promoting potencies of several dioxin-like compounds reflect relative AhR
36      binding and pharmacokinetic parameters (Hebert et al., 1990).

        6/8/00                                     18       DRAFT—DO NOT CITE OR QUOTE

-------
 1             Although few I-P studies have demonstrated lung tumors in rats or mice, the study of
 2     Clark et al. (1991) is particularly significant because of its use of ovariectomized animals. In
 3     contrast to liver tumor promotion, lung tumors were seen only in initiated (diethylnitrosamine
 4     [DEN]), TCDD-treated rats. No tumors were seen in DEN only, TCDD only, control, or
 5     DEN/TCDD intact rats.  Liver tumors are ovary dependent, but ovaries appear to protect against
 6     TCDD-mediated tumor promotion in rat lung. Perhaps use of transgenic animal models will
 7     allow further understanding of the complex interaction of factors associated with carcinogenesis
 8     in rodents as well, presumably in humans. Several such systems are being evaluated (Eastin et
 9     al., 1998; van Birgelen et al., 1999; Dunson et al., 2000).
10            Several potential mechanisms for TCDD carcinogenicity are discussed in Part II, Chapter
11     6. These include oxidative stress, indirect DNA damage, endocrine disruption/growth
12     dysregulation/altered signal transduction, and cell replication/apoptosis leading to tumor
13     promotion.  All of these are biologically plausible as contributors to the carcinogenic process and
14     none are mutually exclusive.  Several biologically based models that encompass many of these
15     activities are described in Part II, Chapter 8. Further work will be needed to elucidate a detailed
16     mechanistic model for any particular carcinogenic response in animals or in humans. Despite
17     this lack of a defined mechanism at the molecular level, there is a consensus that TCDD and
18     related compounds are receptor-mediated carcinogens in that (1) interaction with the AhR is a
19     necessary early event; (2) TCDD modifies a number of receptor and hormone systems involved
20     in cell growth and differentiation, such as the epidermal growth factor receptor and estrogen
21     receptor; and (3) sex hormones exert a profound influence on the carcinogenic action of TCDD.
22
23     2.2.1.3. Other Data Related to  Carcinogenesis
24            Despite the relatively large number of bioassays on TCDD, the study of Kociba et al.
25     (1978) and those of the NTP (1982a), because of their multiple dose groups and wide dose range,
26     continue to be the focus of dose-response modeling efforts and of additional review. Goodman
27     and Sauer (1992) reported a re-evaluation of the female rat liver tumors in the Kociba study using
28     the latest pathology criteria for such lesions. The review confirmed only approximately one-third
29     of the  tumors of the previous review (Squire, 1980). Although this finding did not change the
30     determination of carcinogenic hazard, as TCDD induced tumors in multiple sites in this study, it
31     did have an effect on evaluation of dose-response and on estimates of risk at low doses.  These
32     issues  will be discussed in a later section of this document.
33            One of the more intriguing findings in the Kociba bioassay was reduced tumor incidences
34     of the pituitary, uterus, mammary gland, pancreas, and adrenals in exposed female rats as
35     compared to controls (Kociba et al, 1978).  While these findings, coupled with evaluation of
36     epidemiologic data,  have led some authors to conclude that dioxin possesses "anticarcinogenic"
       6/g/OO                                     19       DRAFT—DO NOT CITE OR QUOTE

-------
  1      activity (Kayajanian, 1997; Kayajanian, 1999), it should be noted that, in experimental studies,
  2     with the exception of mammary gland tumors, the decreased incidence of tumors is associated
  3     with significant weight loss in these rats. Examination of the data from the National Toxicology
  4     Program also demonstrates a significant decrease in these tumor types when there is a
  5     concomitant weight loss in the rodents, regardless of the chemical administered (Haseman and
  6     Johnson,  1996). As discussed later in Section 3.2.3, under certain circumstances exposure to
  7     TCDD may elicit beneficial effects.  For example, TCDD protects against the subsequent
  8     carcinogenic effects of PAHs in mouse skin, possibly reflecting induction of detoxifying
  9     enzymes (Cohen et al., 1979;  DiGiovanni et al., 1980).  In other situations, TCDD-induced
 10     changes in estrogen metabolism may alter the growth of hormone-dependent tumor cells,
 11      producing a potential anticarcinogenic effect (Spink et al., 1990; Gierthy et al., 1993). Because
 12     the mechanism of the decreases in the tumors is unknown, extrapolation of these effects to
 13     humans is premature. In considering overall risk, one must take into account factors such as the
 14     range of doses to target organs and hormonal state to obtain a complete picture of hazard and
 15     risk. Although exposure to dioxins may influence cancer response directly or indirectly,
 16     positively or negatively, it is unlikely that such data will be available to argue that dioxin
 17     exposure provides a net benefit to human health.
 18
 19     2.2.1.4. Cancer Hazard Characterization
 20            TCDD, CDDs, CDFs, and dioxin-like PCBs are  a class of well-studied compounds whose
 21      human cancer potential is supported by a large database including limited epidemiological
 22     support, unequivocal animal carcinogenesis, and biologic plausibility based on mode-of-action
 23     data.  In 1985, EPA classified TCDD and related compounds as "probable" human carcinogens
 24     based on the available data. During the intervening years, the database relating to the
 25     carcinogenicity of dioxin and related compounds has grown and strengthened considerably.  In
 26     addition, EPA guidance for carcinogen risk assessment has evolved (U.S. EPA,  1996). Under
 27     EPA's current approach, TCDD is best characterized as  a "human carcinogen." This means that,
 28     based on the weight of all of the evidence (human, animal, mode of action), TCDD meets the
 29     stringent criteria that allows EPA and the scientific community to accept a causal relationship
 30     between TCDD exposure and cancer hazard. The guidance suggests that "human carcinogen" is
 31      an appropriate descriptor of carcinogenic potential when there is an absence of conclusive
 32      epidemiologic evidence to clearly establish a cause-and-effect relationship between human
 33      exposure and cancer, but there is compelling carcinogenicity data in animals and mechanistic
 34      information in animals and humans demonstrating similar modes of carcinogenic action. The
 35      "human carcinogen" descriptor is suggested for TCDD because all of the following conditions
36      are met:

        6/8/00                                    20       DRAFT—DO NOT CITE OR QUOTE

-------
 1            •   Occupational epidemiologic studies show an association between TCDD exposure
 2                and increases in cancer at all sites, in lung cancer, and perhaps at other sites, but the
 3                data are insufficient on their own to demonstrate a causal association
 4            •   There is extensive carcinogenicity in both sexes of multiple species of animals at
 5                multiple sites.
 6            •   There is general agreement that the mode of TCDD' s carcinogenicity is AhR
 7                dependent and proceeds through modification of the action of a number of receptor
 8                and hormone systems involved in cell growth and differentiation, such as the
 9                epidermal growth factor receptor and estrogen receptor.
10            •   Equivalent body burdens in animals and in human populations expressing an
11                association between exposure to TCDD and cancer, and the determination of active
12                AhR and dioxin-responsive elements in the general human population. There is no
13                reason to believe that these events would not occur in the occupational cohorts
14                studied.
15            Other dioxin-like compounds are characterized as "likely" human carcinogens primarily
16      because of the lack of epidemiological evidence associated with their carcinogenicity, although
17      the inference based on toxicity equivalence is strong that they would behave in humans as TCDD
18      does. Other factors, such as the lack of congener-specific chronic bioassays, also support this
19      characterization. For each congener, the degree of certainty is dependent on the available
20      congener-specific data and its consistency with the generalized mode of action that underpins
21      toxicity equivalence for TCDD and related compounds.  Based on this logic, all complex
22      environmental mixtures of TCDD and dioxin-like compounds would be characterized as "likely"
23      carcinogens, but the  degree of certainty  of the cancer hazard would be dependent on the major
24      constituents of the mixture.  For instance, the hazard potential, although still considered "likely,"
25      would be characterized differently for a mixture whose TEQ was dominated by OCDD as
26      compared to one dominated by other PCDDs.
27
28      2.2.2.  Reproductive and Developmental Effects
29            Several sections of this reassessment (Part II, Chapter 5, and Chapter 7b) have focused on
30      the variety of effects that dioxin and dioxin-like agents can have on human reproductive health
31      and development.  Emphasis in each of these chapters has been on the discussion of the more
32      recent reports of the  impact of dioxin-like compounds on reproduction and development. These
33      have been put into context with previous reviews of the literature applicable in risk assessment
34      (Hatch, 1984; Sweeney, 1994; Kimmel, 1988) to develop a profile of the potential for dioxin and
35      dioxin-like agents to cause reproductive or developmental toxicity, based on the available


        6/8/00                                     21       DRAFT—DO NOT CITE OR QUOTE

-------
  1      literature. An earlier version of the literature review and discussion contained in Part II, Chapter
  2     5, has been previously published (Peterson et al., 1993).
  3            The origin of concerns regarding a potential link between exposure to chlorinated dioxins
  4     and adverse developmental events can be traced to early animal studies reporting increased
  5     incidence of developmental abnormalities in rats and mice exposed early in gestation to 2,4,5-
  6     trichlorophenol (2,4,5-T) (Courtney and Moore, 1971). 2,4,5-T is a herbicide that contains
  7     dioxin and related compounds as impurities. Its use was banned in the late  1970s, but exposure
  8     to human populations continued as a result of past production, use, and disposal.
  9
 10     2.2.2.1. Human
 11             The literature base with regard to potential human effects is detailed in Part II, Chapter
 12     7b. In general, there is little epidemiological evidence that makes a direct association between
 13     exposure to TCDD or other dioxin-like compounds and effects on human reproduction or
 14     development. One effect that may illustrate this relationship is the altered sex ratio (increased
 15     females) seen in the 6 years after the Seveso, Italy, accident (Mocarelli et al., 1996, 2000).
 16     Particularly intriguing in this latest evaluation is the observation that exposure before and during
 17     puberty is linked to this sex ratio effect. Other sites have been examined for the effect of TCDD
 18     exposure on sex ratio with mixed results, but with smaller numbers of offspring.  Continued
 19     evaluation of the Seveso population may provide other indications of impacts on reproduction
 20     and development but, for now, such data are very limited and further research is needed.
 21      Positive human data on developmental effects of dioxin-like compounds are limited to a few
 22     studies of populations exposed to a complex mixture of potentially toxic compounds (e.g.,
 23     developmental studies from the Netherlands and effects of ingestion of contaminated rice oil in
 24     Japan (Yusho) and Taiwan (Yu-Cheng).  In the latter studies, however, all four manifestations of
 25     developmental toxicity (reduced viability, structural alterations, growth retardation, and
 26     functional alterations) have been observed to some degree, following exposure to dioxin-like
 27     compounds as well as other agents. Data from the Dutch cohort of children exposed to PCBs and
 28      dioxin-like compounds (Huisman et al., 1995a,b; Koopman-Esseboom et al., 1994a-c; 1995a,b;
 29      1996; Pluim et al., 1992, 1993, 1994; Weisglas-Kuperus et al., 1995; Patandin et al., 1998,
 30      1999) suggest impacts of background levels of dioxin and related compounds on neurobehavioral
 31      outcomes, thyroid function, and liver enzymes (AST and ALT). Although these effects cannot be
 32      attributed solely to dioxin and related compounds, several associations suggest that these are, in
 33      fact, likely to be Ah-mediated effects.  Similarly, it is highly likely that the developmental effects
 34      in human infants exposed to a complex mixture of PCBs, PCDFs, and poly chlorinated
35      quaterphenyls (PCQs) in the Yusho and Yu-Cheng poisoning episodes may have been caused by
36      the combined exposure to those PCB and PCDF congeners that are Ah-receptor agonists (Lii  and

        6/8/00                                     22        DRAFT—DO NOT CITE OR QUOTE

-------
 1     Wong, 1984; Kuratsune, 1989; Rogan, 1989).  However, it is not possible to determine the
 2     relative contributions of individual  chemicals to the observed effects.
 3            The incidents at Yusho and Yu-Cheng resulted in increased perinatal mortality and low
 4     birthweight in infants born to women who had been exposed. Rocker bottom heal was observed
 5     in Yusho infants, and functional abnormalities have been reported in Yu-Cheng children. Not all
 6     the effects that were seen are attributable only to dioxin-like compounds.  The similarity of
 7     effects observed in human infants prenatally exposed to this complex mixture with those reported
 8     in adult monkeys exposed only to TCDD suggests that at least some of the effects in the Yusho
 9     and Yu-Cheng children are due to the TCDD-like congeners in the contaminated rice oil ingested
10     by the mothers of these children. The similar responses include a clustering of effects in organs
11     derived from the ectodermal germ layer, referred to as ectodermal dysplasia, including effects on
12     the skin, nails, and Meibomian glands; and developmental and psychomotor delay during
13     developmental and cognitive tests (Chen et al., 1992).  Some investigators believe that, because
14     all of these effects in the Yusho and  Yu-Cheng cohorts do not correlate with TEQ, some of the
15     effects are exclusively  due to nondioxin-like PCBs or a combination of all the congeners. It is
16     still not clear to what extent there is  an association between overt maternal toxicity and
17     embryo/fetal toxicity in humans.
18            Of particular interest is the common developmental origin (ectodermal layer) of many of
19     the organs and tissues that are affected in the human. An ectodermal dysplasia syndrome has
20     been clearly associated with the Yusho and Yu-Cheng episodes, involving hyperpigmentation,
21     deformation of the fingernails and toenails, conjunctivitis, gingival hyperplasia, and
22     abnormalities of the teeth. An investigation of dioxin exposure and tooth development was done
23     in Finnish children as a result of studies of dental effects in dioxin-exposed rats, mice, and
24     nonhuman primates  (Chapter 5), and in PCB-exposed children (Rogan et al., 1988). The Finnish
25     investigators examined enamel hypomineralization of permanent first molars in 6-7 year old
26     children (Alaluusua et  al., 1996, 1999).  The length of time that infants breast fed was not
27     significantly associated with either mineralization changes or with TEQ levels in the breast milk.
28     However, when the levels and length of breast feeding were combined in an overall score, a
29     statistically significant association was observed (r = 0.3, p = 0.003, regression analysis).  These
30     data are discussed further in  Part II,  Chapter 7b.  The developmental effects that can be
31     associated with the nervous system are also consistent with this pattern of impacts on tissues of
32     ectodermal origin, as the nervous system is of ectodermal origin. These data are limited but are
33     discussed in Part II, Chapter 7b.
34            Other investigations into noncancer effects of human exposure to dioxin have provided
35     human data on TCDD-induced changes in circulating reproductive hormones.  This was one of
36     the effects judged as having a positive relationship with exposure to TCDD in Part II, Chapter

       6/8/00                                     23       DRAFT—DO NOT CITE OR QUOTE

-------
  1      7b. Levels of reproductive hormones have been measured with respect to exposure to 2,3,7,8-
  2      TCDD in three cross-sectional medical studies. Testosterone, LH, and FSH were measured in
  3      TCP and 2,4,5-T production workers (Egeland et al., 1994), in Army Vietnam veterans (Centers
  4      for Disease Control Vietnam Experience Study, 1988), and in Air Force personnel, known as
  5      "Ranch Hands," who handled and/or sprayed Agent Orange during the Vietnam War (Roegner et
  6      al., 1991; Grubbs et al., 1995). The risk of abnormally low testosterone was two to four times
  7      higher in exposed workers with serum 2,3,7,8-TCDD levels above 20 pg/g than in unexposed
  8      referents (Egeland et al., 1994). In both the 1987 and 1992 examinations, mean testosterone
  9      concentrations were slightly, but not significantly, higher in Ranch Hands (Roegner et al.,  1991;
 10      Grubbs et al., 1995). FSH and LH concentrations were no different between the exposed and
 11      comparison groups.  No significant associations were found between Vietnam experience and
 12      altered reproductive hormone levels (Centers for Disease Control Vietnam Experience Study,
 13      1988). Only the NIOSH study found an association between serum 2,3,7,8-TCDD level and
 14      increases in serum LH.
 15             The findings of the NIOSH and Ranch Hand studies are plausible given the
 16      pharmacological  and toxicological properties of 2,3,7,8-TCDD in animal models, which are
 17      discussed in Part II, Chapters 5 and 7. One plausible mechanism responsible for the effects of
 18      dioxins may involve their ability to influence hormone receptors. The AhR, to which 2,3,7,8-
 19      TCDD binds, and the hormone receptors are signaling pathways that regulate homoeostatic
 20      processes. These signaling pathways are integrated at the cellular level and there is considerable
 21      "cross-talk" between these pathways. For example, studies suggest that 2,3,7,8-TCDD
 22      modulates the concentrations of numerous hormones and/or their receptors, including estrogen
 23      (Romkes and  Safe, 1988; Romkes et al., 1987), progesterone (Romkes et al., 1987),
 24      glucocorticoid (Ryan et al., 1989),  and thyroid hormones (Gorski and Rozman, 1987).
 25             In summary, the results from both the NIOSH and Ranch Hand studies are limited by the
 26      cross-sectional nature of the data and the type of clinical assessments conducted. However, the
 27      available data provide evidence that small alterations in human male reproductive hormone
 28      levels are associated with serum 2,3,7,8-TCDD.
 29
 30      2.2.2.2. Experimental Animal
 31             The extensive experimental animal database with respect to reproductive and
 32      developmental toxicity of dioxin and dioxin-related agents has been discussed in Part II, Chapter
 33      5.  Dioxin exposure has been observed to result in both male and female reproductive effects,  as
 34      well as effects on development. These latter effects are among the most responsive health
35      endpoints to dioxin exposure (see Part II, Chapter 8). In general, the prenatal and developing
36      postnatal animal is more sensitive to the effects of dioxin than is the adult. In  several instances

        6/8/00                                     24        DRAFT—DO NOT CITE OR QUOTE

-------
 1     (e.g., fetotoxicity in hamsters, rats, mice, and guinea pigs), the large species differences seen in
 2     acute toxicity are greatly reduced when developing animals are evaluated.  Most of the data
 3     reviewed are from studies of six genera of laboratory animals. Although much of the data comes
 4     from animals exposed only to TCDD, more recent studies of animals exposed to mixtures of
 5     PCDD/PCDF isomers provide results that are consistent with the studies of TCDD alone.
 6
 7     2.2.2.2.1. Developmental toxicity. Dioxin exposure results in a wide variety of developmental
 8     effects; these are observed in three different vertebrate classes and in several species within each
 9     class. All four of the manifestations of developmental toxicity have been observed following
10     exposure to dioxin, including reduced viability, structural alterations, growth retardation, and
11     functional alterations. As summarized previously (Peterson et al., 1993), increased prenatal
12     mortality (rat and monkey), functional alterations in learning and sexual behavior (rat and
13     monkey), and changes in the development of the reproductive system (rat, hamster) occur at the
14     lowest exposure levels tested (see also Part II, Chapter 8).
15            Dioxin  exposure results in reduced prenatal or postnatal viability in virtually every
16     species in which it has been tested. Previously, increased prenatal mortality appeared to be
17     observed only at exposures that also resulted in maternal toxicity. However, the studies of Olson
18     and McGarrigle (1990) in the hamster and Schantz et al. (1989) in the monkey were suggestive
19     that this was not the case in all species.  Although the data from these two studies were limited,
20     prenatal death was observed in cases where no maternal toxicity was evident.  In the rat,
21     Peterson's laboratory (Bjerke et al., 1994a,b; Roman et al., 1995) reported increased prenatal
22     death following a single  exposure to TCDD during gestation that did not cause maternal toxicity,
23     and Gray et al.  (1995a) observed a decrease in postnatal survival under a similar exposure
24     regimen. While identifying the presence or absence of maternal toxicity may be instructive as to
25     the specific origin of the reduced prenatal viability, it does not alter the fact that pre- and
26     postnatal deaths were observed.  In either case, the Agency considers these effects as being
27     indicators of developmental toxicity in response to the exposure (U.S. EPA, 1991b).
28            Some of the most striking findings regarding dioxin exposure relate to the effects on the
29     developing reproductive system in laboratory animals. Only a single, low-level exposure to
30     TCDD during gestation is required to initiate these developmental alterations. Mably et al.
31     (1992a-c) originally reported that a single exposure of the Holtzman maternal rat to as low as
32     0.064 |ig/kg could alter normal sexual development in the male offspring. A dose of 0.064 ng/kg
33     in these studies results in a body maximal burden in the maternal animal of 64 ng/kg during
34     critical windows in development. More recently, these findings of altered normal sexual
35     development have been further defined (Bjerke et al., 1994a,b; Gray et al., 1995a; Roman et al.,

       6/g/00                                      25        DRAFT—DO NOT CITE OR QUOTE

-------
  1       1995), as well as extended to females and another strain and species (hamster) (Gray et al.,
  2      1995b). In general, the findings of these later studies have produced qualitatively similar results
  3     that define a significant effect of dioxin on the developing reproductive system.
  4            In the developing male rat, TCDD exposure during the prenatal and lactational periods
  5     results in delay of the onset of puberty as measured by age at preputial separation.  There is a
  6     reduction in testis weight, sperm parameters, and sex accessory gland weights. In the mature
  7     male exposed during the prenatal and lactational periods, there is an alteration of normal sexual
  8     behavior and reproductive function.  Males exposed to TCDD during  gestation are
  9     demasculinized.  Feminization of male sexual behavior and a reduction in the number of
 10     implants in females mated with exposed males have also been reported, although these effects
 11      have not been consistently found.  These effects do not appear to be related to reductions in
 12     circulating androgens, which were shown in the most recent studies to be normal. Most of these
 13     effects occur in a dose-related fashion, some occurring at 0.05 ng/kg and 0.064 ng/kg, the lowest
 14     TCDD doses tested (Mably et al., 1992c; Gray et al., 1997a).
 15            In the developing female rat,  Gray and Ostby (1995) have demonstrated altered sexual
 16     differentiation in both the Long Evans and Holtzman strains. The effects observed depended on
 17     the timing of exposure. Exposure during early organogenesis altered the cyclicity, reduced
 18     ovarian weight, and shortened the reproductive lifespan. Exposure  later in organogenesis
 19     resulted in slightly lowered ovarian weight, structural alterations of the genitalia, and a slight
 20     delay in puberty. However, cyclicity and fertility were not affected with the later exposure. The
 21      most sensitive dose-dependent effects of TCDD in the female rat were structural alterations of
 22     the genitalia that occurred at 0.20 ng  TCDD/kg administered to the  dam (Gray et al., 1997b).
 23            As described above, studies demonstrating adverse health effects from prenatal exposures
 24     often involved a single dose administered at a discrete time  during pregnancy.  The production of
 25     prenatal effects at a given dose appears to require exposure during critical times in fetal
 26     development. This concept is well supported by a recent report (Hurst et al., 2000) which
 27     demonstrated the same incidence of adverse effects in rat pups born to dams with a single
 28     exposure of 0.2 ng  TCDD/kgBW on gestation day 15 (GD 15) versus  1.0 ng TCDD/kgBW on
 29      gestation day 8 (GD 8). Both of these experimental paradigms result in the same fetal tissue
 30      concentrations and body burdens during the critical window of sensitivity. For example,
 31      exposure to 0.2 ng TCDD/kgBW on GD 15 results in 13.2 pg TCDD/g fetal tissue on GD 16;
 32      exposure to 1.0 jig TCDD/kgBW on gestation GD 8 resulted in 15.3 pg TCDD/g fetus on GD 16.
 33      This study demonstrates the appropriateness of the use of body burden to describe the effects of
 34      TCDD when comparing different exposure regimens. The uncertainties introduced when trying
35      to compare studies  with steady-state body burdens with single-dose studies may make it difficult
36      to determine a lowest effective dose.  Application of pharmacokinetics models, described earlier

        6/8/00                                      26       DRAFT—DO NOT CITE OR QUOTE

-------
 1      in Parts I and II, to estimate body burdens at the critical time of development is expected to be a
 2      sound method for relating chronic background exposures to the results obtained from single-dose
 3      studies.
 4            Structural malformations, particularly cleft palate and hydronephrosis, occur in mice
 5      administered doses of TCDD. The findings, while not representative of the most sensitive
 6      developmental endpoints, indicate that exposure during the critical period of organogenesis can
 7      affect the processes involved in normal tissue formation.  The TCDD-sensitive events appear to
 8      require the AhR. Mouse strains that produce AhRs with relatively high affinity for TCDD
 9      respond to lower doses than do strains with relatively low-affinity receptors. Moreover,
10      congeners with a greater affinity for the AhR are more developmentally toxic than those with a
11      lower affinity.  This is consistent with the rank ordering of toxic potency based on affinity for the
12      receptor as discussed in Part II, Chapter 9.
13
14      2.2.2.2.2. Adult female reproductive toxicity. The primary effects of TCDD on female
15      reproduction appear to be decreased fertility, inability to maintain pregnancy for the full
16      gestational period and, in the rat, decreased litter size. In some studies of rats and of primates,
17      signs of ovarian dysfunction such as anovulation and suppression of the estrous cycle have been
18      reported (Kociba et al., 1976; Barsotti et al., 1979; Allen et al., 1979; Li et al., 1995a,b).
19
20      2.2.2.2.3. Adult male reproductive toxicity. TCDD and related compounds decrease testis and
21      accessory sex organ weights, cause abnormal testicular morphology, decrease spermatogenesis,
22      and reduce fertility when given to adult animals in doses sufficient to reduce feed intake and/or
23      body weight. In the testes of these different species, TCDD effects on spermatogenesis are
24      characterized by loss of germ cells, the appearance of degenerating spermatocytes and mature
25      spermatozoa within the lumens of seminiferous tubules, and a reduction in the number of tubules
26      containing mature spermatozoa (Allen and Lalich, 1962; Allen and Carstens, 1967; McConnell et
27      al., 1978; Chahoud et al., 1989). This suppression of spermatogenesis is not a highly sensitive
28      effect when TCDD is administered to postweanling animals, as an exposure of 1 ng/kg/day over
29      a period of weeks appears to be required to produce these effects.
30
31      2.2.2.3.  Other Data Related to Developmental and Reproductive Effects
32      2.2.2.3.1. Endometriosis.  The association of dioxin with endometriosis was first reported in a
33      study of Rhesus monkeys that had been exposed for 4 years to dioxin in their feed and then held
34      for an additional 10 years (Rier et al., 1993). There was a dose-related increase in both the
35      incidence and severity of endometriosis in the exposed monkeys as compared to controls.

        6/8/00                                     27       DRAFT—DO NOT CITE OR QUOTE

-------
  1      Follow-up on this group of monkeys revealed a clear association with total TEQ. A study in
  2      which Rhesus monkeys were exposed to PCBs for up to 6 years failed to show any enhanced
  3      incidence of endometriosis (Arnold et al., 1996).  However, many of these monkeys were no
  4      longer cycling, and the time may not have been adequate to develop the response. In the TCDD
  5      monkey study, it took 7 years before the first endometriosis was noted (Rier et al., 1993).  A
  6      recent study in Cynomolgus monkeys has shown promotion of surgically induced endometriosis
  7      by TCDD within 1 year after surgery (Yang et al., in press). Studies using rodent models for
  8      surgically induced endometriosis have also shown the ability of TCDD to promote lesions in a
  9      dose-related manner (Cummings et al., 1996,1999; Johnson et al., 1997; Bruner-Tran et al.,
10      1999). This response takes at least 2 months to be detected  (Cummings et al., 1996, 1999;
11      Johnson et al.,  1997).  Another study in mice which failed to detect dioxin promotion of
12      surgically induced endometriosis only held the mice for only 1 month, not long enough to detect
13      a response (Yang et al., 1997).  Prenatal exposure to mice also enhanced the sensitivity of the
14      offspring to the promotion of surgically induced endometriosis by TCDD. The effects of TCDD
15      in the murine model of endometriosis appear to be AhR-mediated, as demonstrated in a study in
16      which AhR ligands were able to promote the lesions, while non-Ah ligands, including a non-
17      dioxin-like PCB, had no effect on surgically induced endometriosis. Dioxin has also been shown
18      to result in endometriosis in human endometrial tissue implanted in nude mice (Bruner-Tran et
19      al., 1999).
20             Data on the relationship of dioxins to endometriosis in people is intriguing, but
21      preliminary. Studies in the early 1990s suggested that women with higher levels of persistent
22      organochlorines were at increased risk for endometriosis (Gerhard and Runnebaum, 1992). This
23      was followed by the observation that Belgian women, who have the highest levels of dioxins in
24      their background population, had higher incidences of endometriosis than reported from other
25      populations (Koninckx et al., 1994).  A study from Israel then demonstrated that there was a
26      correlation between detectable TCDD in women with surgically confirmed endometriosis, in
27      comparison to those with no endometriosis (Mayani et al., 1997).  Recent studies from Belgium
28      have indicated that women with higher body burdens, based on serum TEQ determinations, are at
29      greater risk for endometriosis (Pauwels et al.,  1999).  No association was seen with total PCBs in
30      this study.  A small study in the United States, which did not involve surgically confirmed
31      endometriosis,  saw no association between TCDD and endometriosis (Boyd et al., 1995).
32      Likewise, a study in Canada saw no association between total  PCBs and endometriosis (Lebel et
33      al.. 1998).  The negative association with total PCBs is not surprising because the rodent studies
34      have indicated that this response is AhR-mediated (Johnson et al., 1997).  Preliminary results


        6/8/00                                    28      DRAFT—DO NOT CITE OR QUOTE

-------
 1      from Seveso suggest a higher incidence of endometriosis in the women from the two highly
 2     exposed zones (A and B) as compared to the background incidence in Italy (Eskanzi et al., 1998).
 3            The animal results lend biological plausibility to the epidemiology findings.
 4     Endometriosis is not only an endocrine disorder, but is also associated with immune system
 5     alterations (Rier et al., 1995). Dioxins are known to be potent modulators of the animal immune
 6     system, as well as affecting estrogen homeostasis.  Further studies are clearly needed to provide
 7     additional support to this association of endometriosis and dioxins, as well as to demonstrate
 8     causality.
 9
10     2.2.2.3.2. Androgenic deficiency. The effects of TCDD on the male reproductive system when
11      exposure occurs in adulthood are believed to be due in part to an androgenic deficiency. This
12     deficiency is characterized in adult rats by decreased plasma testosterone and DHT
13     concentrations, unaltered plasma LH concentrations, and unchanged plasma clearance of
14     androgens and LH (Moore et al., 1985, 1989; Mebus et al., 1987; Moore and Peterson, 1988;
15     Bookstaff et al., 1990a). The cause of the androgenic deficiency was believed to be due to
16     decreased testicular responsiveness to LH and increased pituitary responsiveness to feedback
17     inhibition by androgens and estrogens (Moore et al., 1989, 1991; Bookstaff et al., 1990a,b;  .
18     Kleeman et al., 1990). The single dose used in some of those earlier studies (15
19     ugTCDD/kgBW) is now known to affect Leydig cells (Johnson et al., 1994).
20
21      2.2.2 A. Developmental and Reproductive Effects Hazard Characterization
22            There is limited direct evidence addressing the issues of how or at what levels humans
23     will begin to respond to dioxin-like compounds with adverse impacts on development or
>4     reproductive function. The series of published Dutch studies suggest that pre- and early postnatal
 !5     exposures to PCBs and other dioxin-like compounds may impact developmental milestones at
 !6     levels at or near current average  human background exposures. Although it is unclear whether
 !7     these measured responses indicate a clearly adverse impact, if humans respond to TCDD
 :8     similarly to animals in laboratory studies, there are indications that exposures at relatively low
 9     levels might cause developmental effects and at higher exposure levels might cause reproductive
 0     effects. There is especially good evidence for effects on the fetus from prenatal exposure. The
 1      Yusho and Yu-Cheng poisoning incidents are clear demonstrations that dioxin-like compounds
 2     can produce a variety of mild to  severe developmental effects in humans that resemble the effects
 3     of exposure to dioxins and dioxin-like compounds in animals.  Humans do not appear to be
 4-     particularly sensitive or insensitive to effects of dioxin exposure in comparison to other animals.
 5     Therefore it is reasonable to assume that human responsiveness would lie across the middle

       6/8/00                                      29        DRAFT—DO NOT  CITE OR QUOTE

-------
  1      ranges of observed responses.  This still does not address the issues surrounding the potentially
  2      different responses humans (or animals) might have to the more complex and variable
  3      environmental mixtures of dioxin-like compounds.
  4             TCDD and related compounds have reproductive and developmental toxicity potential in
  5      a broad range of wildlife, domestic, and laboratory animals. Many of the effects have been
  6      shown to be TCDD dose-related. The effects on perinatal viability and male reproductive
  7      development are among the most sensitive effects reported, occurring at a single prenatal
  8      exposure range of as little as 0.05-0.075 fig/kg, resulting in calculated fetal tissue concentrations
  9      of 3-4 ng/kg. In these studies, effects were often observed at the lowest exposure level tested,
 10      thus a no-observed adverse effect level (NOAEL) has not been established for several of these
 11      endpoints. In general, the structure-activity results are consistent with an AhR-mediated
 12      mechanism for the developmental effects that are observed in the low dose range.  The structure-
 13      activity relationship in laboratory mammals appears to be similar to that for AhR binding. This
 14      is especially the case with cleft palate in the mouse.
 15             It is assumed that the responses observed in animal studies are indicative of the potential
 16      for reproductive and developmental toxicity in humans. This is an established assumption in the
 17      risk assessment process for developmental toxicity (U.S. EPA, 1991b).  It is supported by the
 18      number of animal species and strains in which  effects have been observed. The limited human
 19      data are consistent with an effect following exposure to TCDD or TCDD-like agents.  In
 20      addition, the phylogenetic conservation of the structure and function of the AhR also increases
 21      our confidence that these effects may occur in humans.
 22             Although there is evidence in experimental animals that exposure to  dioxin-like
 23      chemicals during development produces neurobehavioral effects, the  situation in humans is more
 24      complex. Studies in humans demonstrate associations between dioxin exposure and alterations
 25      in neurological development. These same studies often show similar associations between
 26      exposure to non-dioxin-like PCBs and these same effects. On the  basis of the human  studies, it
 27      is possible that the alterations in neurological development are due to an interaction between the
 28      dioxins and the non-dioxin-like PCBs.  At present there are limited data that define the roles of
 29      the dioxins versus the non-dioxin-like PCBs in these effects on neurological development.
 30             In general, the  structure-activity results on dioxin-like compounds are consistent with an
 31      AhR-mediated mechanism for many of the developmental effects that are observed. The
 32      structure-activity relationship in laboratory mammals appears to be similar to that for  AhR
 33      binding.  This is especially the case with cleft palate in the mouse.  However, a direct
34      relationship with Ah binding is less clear for other effects, including those involving the nervous
35      system.
36

        6/8/00                                     30        DRAFT—DO NOT CITE OR QUOTE

-------
 1      2.2.3. Immunotoxicity
 2     2.2.3.1.  Epidemiologic Finding
 3           The available epidemiologic studies on immunologic function in humans relative to
 4     exposure to 2,3,7,8-TCDD 
-------
  1      certain infectious agents, which likely result from TCDD-induced suppression of immunological
  2      functions.
  3             The primary antibody response to the T cell-dependent antigen, sheep red blood cells
  4      (SRBCs), is the most sensitive immunological response that is consistently suppressed in mice
  5      exposed to TCDD and related compounds.  The degree of immunosuppression is related to the
  6      potency of the dioxin-like congeners. There is remarkable agreement among several different
  7      laboratories for the potency of a single acute dose of TCDD (i.e., suppression at a dose as low as
  8      0.1 ng TCDD/kg with an average 50% immunosupressive dose [ID50J value of approximately 0.7
  9      //g TCDD/kg) to suppress this response in Ah-responsive mice.  Results of studies that have
 10      compared the effects of acute exposure to individual PCDDs, PCDFs, and PCB congeners, which
 11      differ in their binding affinity for the AhR, on this response have provided critical evidence that
 12      certain dioxin-like congeners are also immunosuppressive.  The degree of immunosuppression
 13      has been found to be related to potency of the dioxin-like congeners. Antibody responses to
 14      T cell-independent antigens, such as trinitrophenyl-lipopolysaccharide (TNP-LPS) and the
 15      cytotoxic T lymphocyte (CTL) response, are also suppressed by a single acute exposure to
 16      TCDD, albeit at higher doses than those that suppress the SRBC response.  Although a thorough
 17      and systematic evaluation of the immunotoxicity of TCDD-like  congeners in different species
 18      and for different immunological endpoints has not been performed, it can be inferred from the
 19      available data that dioxin-like congeners are immunosuppressive.
 20             Perinatal exposure of experimental animals to  TCDD results in suppression of primarily
 21      T cell immune functions, with evidence of suppression persisting into adulthood. In mice, the
 22      effects on T cell functions appear to be related to the fact that perinatal TCDD exposure alters
 23      thymic precursor stem cells in the fetal liver and bone marrow, and thymocyte differentiation in
 24      the thymus. These studies suggest that perinatal development is a critical and sensitive period for
 25      TCDD-induced immunotoxicity. Efforts should be made to determine the consequences of
 26      perinatal exposure to TCDD and related compounds and mixtures on immune system integrity.
 27
 28      2.2.3.3. Other Data Related to Immunologic Effects
 29             In addition to the TCDD-like congener results, studies using strains of mice that differ in
 30      the expression of the AhR have provided critical evidence to support a role for Ah-mediated
 31      immune suppression following exposure to dioxin-like compounds. Recent in vitro work also
 32      supports a role for Ah-mediated immune suppression. Other in vivo and in vitro data, however,
 33      suggest that non-Ah-mediated mechanisms may also play some  role in immunotoxicity induced
 34      by dioxin-like compounds.  However, more definitive evidence remains to be developed to
35      support this latter view.


        6/8/00                                    32       DRAFT—DO NOT CITE OR QUOTE

-------
 1             Although the immunosuppressive potency of individual dioxin-like compounds in mice is
 2     related to their structural similarity to TCDD, this pattern of suppression is observed only
 3     following exposure to an individual congener. The immunotoxicity of TCDD and related
 4     congeners can be modified by co-exposure to other congeners in simple binary or more complex
 5     mixtures resulting in additive or antagonistic interactions.  There is a need for the generation of
 6     dose-response data of acute, subchronic, and chronic exposure to the individual congeners in a
 7     mixture and for the mixture itself in order to fully evaluate potential synergistic, additive, or
 8     antagonistic effects of environmentally relevant mixtures.
 9            Animal host resistance models that mimic human disease have been used to assess the
10     effects of TCDD on altered host susceptibility.  TCDD exposure increases susceptibility to
11     challenge with bacteria, viruses, parasites, and tumors. Mortality is increased in TCDD-exposed
12     mice challenged with certain bacteria. Increased parasitemia occurs in TCDD-exposed mice and
13     rats challenged with parasitic infections.  Low doses of TCDD also alter resistance to virus
14     infections in rodents. Increased susceptibility to infectious agents is an important benchmark of
15     immunosuppression; however, the role that TCDD plays in altering immune-mediated
16     mechanisms important in murine resistance to infectious agents remains to be elucidated. Also,
17     because little is known about the effects that dioxin-like congeners have on host resistance, more
18     research is recommended in this area.
19            Studies in nonhuman primates exposed acutely, subchronically, or chronically to
20     halogenated aromatic hydrocarbons (HAH) have revealed variable alterations in lymphocyte
21     subpopulations, primarily T lymphocyte subsets. In three separate studies in which monkeys
22     were exposed subchronically or chronically to PCBs, the antibody response to SRBC was
23     consistently found to be suppressed. These results in nonhuman primates are important because
24     they corroborate the extensive database of HAH-induced suppression of the antibody response to
25     SRBC in mice and thereby provide credible evidence for immunosuppression by HAHs across
26     species. In addition, these data indicate that the primary antibody response to this T cell-
27     dependent antigen is the most consistent and sensitive indicator of HAH-induced
28     immunosuppression.
29            The available database derived from well-controlled animal studies on TCDD
30     immunotoxicity can be used for the establishment of NOELS. As the antibody response to
31     SRBCs has been shown to be dose-dependently suppressed by TCDD and related dioxin-like
32     compounds, this database is best suited for the development of dose-response modeling.
33
34     2.2.3.4. Immunologic Effects Hazard Characterization
35            Accidental or occupational exposure of humans to TCDD and/or related compounds
36     variably affects a number of immunological parameters. Unfortunately, the evaluation of

       6/8/00                                    33        DRAFT—DO NOT  CITE OR QUOTE

-------
  1      immune system integrity in humans exposed to dioxin-like compounds has provided data that is
  2     inconsistent across studies. However, the broad range of "normal" responses in humans due to
  3     the large amount of variability inherent in such a heterogenous population, the limited number
  4     and sensitivity of tests performed, and poor exposure characterization of the cohorts in these
  5     studies compromise any conclusions about the ability of a given study to detect immune
  6     alterations. Consequently, there are insufficient clinical data from these studies to fully assess
  7     human sensitivity to TCDD exposure. Nevertheless, based on the results of the extensive animal
  8     work, the database is sufficient to indicate that immune effects could occur in the human
  9     population from exposure to TCDD and related compounds at some dose level. At present, it is
 10     EPA's scientific judgment that TCDD and related compounds should be regarded as nonspecific
 11      immunosuppressants and immunotoxicants until better data to inform this judgment are
 12     available.
 13            It is interesting that a common thread in several human studies is the observed reduction
 14     in CD4+ T helper cells, albeit generally within the "normal" range, in cohorts exposed to dioxin-
 15     like compounds.  Even though these reductions may not translate into clinical effects, it is
 16     important to note that these cells play an important role in regulating immune responses and that
 17     their reduction in clinical diseases is associated with immunosuppression. Another important
 18     consideration is that a primary antibody response following immunization was not evaluated in
 19     any of the  human studies. Because this immune parameter has been revealed to be the most
 20     sensitive in animal studies, it is recommended that TCDD and related compounds be judged
 21      immunosupressive and that this parameter be included in future studies of human populations
 22     exposed to TCDD and related compounds. It is also recommended that research focused on
 23     delineating the mechanism(s) underlying dioxin-induced immunotoxicity and
 24     immunosuppression continue.
 25
 26     2.2.4. Chloracne
 27            Chloracne and associated dermatologic changes are widely recognized responses to
 28     TCDD and other dioxin-like compounds in humans. Along with the reproductive hormones
 29     discussed above and gamma glutamyl transferase (GOT) levels, which are discussed below,
 30     chloracne is one of the noncancer effects that has a strong positive association with exposure to
 31      TCDD in humans (see Part II. Chapter 7b). Chloracne is a severe  acnelike condition that
 32      develops within months of first exposure to high levels of dioxin and related compounds. For
 33      many individuals, the condition disappears after discontinuation of exposure, despite initial
 34      serum levels of dioxin in the thousands of parts per trillion; for others, it may remain for many
35      years. The duration of persistent chloracne is on the order of 25 years, although cases of
36      chloracne persisting over 40 years have been noted (see Chapter 7, Epidemiology).

        6/8/00                                     34       DRAFT—DO NOT CITE OR QUOTE

-------
 1              In general, chloracne has been observed in most incidents where substantial dioxin
 2     exposure has occurred, particularly among trichlorophenol (TCP) production workers and Seveso
 3     residents (see Part II, Chapter 7b). The amount of exposure necessary for development of
 4     chloracne has not been resolved, but studies suggest that high exposure (both high acute and
 5     long-term exposure) to 2,3,7,8-TCDD increases the likelihood of chloracne, as evidenced by
 6     chloracne in TCP production workers and Seveso residents who have documented high serum
 7     2,3,7,8-TCDD levels (Beck et al., 1989; Fingerhut et al., 1991a; Mocarelli et al., 1991;
 8     Neuberger et al., 1991) or in individuals who have a work history with long duration of exposure
 9     to 2,3,7,8-TCDD-contaminated chemicals (Bond et al., 1989). In earlier studies, chloracne was
10     considered to be a "hallmark of dioxin intoxication" (Suskind, 1985). However, only in two
11     studies were risk estimates calculated for chloracne. Both were studies of different cohorts of
12     TCP production workers (Suskind and Hertzberg, 1984; Bond et al., 1989); one group was
13     employed in a West Virginia plant, the other in a plant in Michigan. Of the 203 West Virginia
14     workers, 52.7% (p<0.001) were found to have clinical evidence of chloracne, and 86.3% reported
15     a history of chloracne (p<0.001) (Suskind and Hertzberg, 1984). None  of the unexposed workers
16     had clinical evidence or reported a history of chloracne. Among the Michigan workers, the
17     relative risk for cases of chloracne was highest for individuals with the longest duration of
18     exposure (>60 months; RR = 3.5, 95% CI = 2.3-5.1), those with the highest cumulative dose of
1.9     TCDD (based on duration of assignment across and within 2,3,7,8-TCDD-contaminated areas in
20     the plant) (RR = 8.0, 95% CI = 4.2-15.3), and those with the highest intensity of 2,3,7,8-TCDD
21     exposure (RR = 71.5, 95% CI = 32.1-159.2) (Bond etal., 1989).
22            Studies in multiple animal species have been effective in describing the relationship
23     between 2,3,7,8-TCDD and chloracne, particularly in rhesus monkeys (McNulty,  1977; Allen et
24     al., 1977; McConnell et al., 1978). Subsequent to exposure to 2,3,7,8-TCDD, monkeys
25     developed chloracne and swelling of the meibomian glands, modified sebaceous glands in the
26     eyelid. The histologic changes in the meibomian glands are physiologically similar to those
27     observed in human chloracne (Dunagin, 1984).
28            In summary, the evidence provided by the various studies convincingly supports what is
29     already presumed, that chloracne is a common sequel of high levels of exposure to 2,3,7,8-
30     TCDD and related compounds. More information is needed to determine the level and frequency
31     of exposure to dioxin-like compounds needed to cause chloracne, and whether personal
32     susceptibility  plays a role in the etiology.  Finally, it is important to recall that the absence of
33     chloracne does not imply lack of exposure (Mocarelli et al., 1991).
34
35
36
       6/8/00                                    35        DRAFT—DO NOT CITE OR QUOTE

-------
  1      2.2.5. Diabetes
  2            Diabetes mellitus is a heterogeneous disorder that is a consequence of alterations in the
  3     number or function of pancreatic beta cells responsible for insulin secretion and carbohydrate
  4     metabolism. Diabetes and fasting serum glucose levels were evaluated in more recent cross-
  5     sectional medical studies because of the apparently high prevalence of diabetes and abnormal
  6     glucose tolerance tests in one case report of 55 TCP workers (Pazderova-Vejlupkova et al.,
  7     1981). Recent epidemiology studies, as well as early case reports, have indicated a weak
  8     association between serum concentrations of dioxin and diabetes. This association was first
  9     noted in the early 1990s when a decrease in glucose tolerance was seen in the NIOSH cohort.
 10     This was followed by a report of an increase in diabetes in the Ranch Hand cohort (Michalek et
 11      al., 1999; Longnecker and Michalek, 2000).  Several reports from other occupational cohorts
 12     (Steenland et al., 1999; Vena et al.,  1998), as well as the Sevesd population (Pesatori et al., 1998)
 13     then followed. There was not a significant increase in diabetes in the NIOSH mortality study,
 14     although 6 of the  10 most highly exposed workers did have diabetes (Calvert et al., 1999).
 15     However, it is well understood that mortality studies are limited in their ability to assess risk
 16     from diabetes mellitus.  The recent paper by Longnecker and Michalek (2000) found a pattern
 17     suggesting that low levels of dioxin may influence the prevalence of diabetes. However, these
 18     results did not show an exposure-response relationship.  Because it is the only study of its type to
 19     have been published, additional population-based studies are warranted to validate its findings.
 20     The most recent update of the Ranch Hand study shows a 47% excess of diabetes in the most
 21      heavily exposed group of veterans (Michalek et al., 1999).
 22            Most of the data suggest that the diabetes is Type II, or adult-onset, diabetes, rather than
 23     insulin dependent, or Type I.  Aging and obesity are the key risk factors for Type II diabetes.
 24     However, dioxins may shift the distribution of sensitivity, putting people at risk at younger ages
 25     or with less weight. Dioxin alters lipid metabolism in multiple species, including humans
 26     (Sweeney et al., 1997; Pohjanvirta and Tuomisto, 1994).  Dioxin also alters glucose uptake into
 27      both human and animal cells in culture (Enan and Matsumura, 1994; Olsen et al., 1994).
 28      Mechanistic studies have demonstrated that dioxin affects glucose transport (Enan and
 29      Matsumura, 1994), a property under the control of the hypoxia response pathway (Ouiddir et al.,
 30      1999). A key regulatory protein in this pathway is the partner of the AhR, Arnt (also known as
 31      HIF1 -beta) (Gu et al., 2000; Taylor and Zhulin, 1999). Activation of the AhR by dioxin may
 32      compete with other pathways, such as the HIF pathway, for Arnt (Gradin, et al., 1992). Dioxin
33      has also been shown to downregulate the insulin growth factor receptor (Liu et al., 1992).  These
34      three issues — altered lipid metabolism, altered glucose transport, and alterations in the insulin
        6/8/00                                      36       DRAFT—DO NOT CITE OR QUOTE

-------
 1     signaling pathway — all provide biological plausibility to the association of dioxins with
 2     diabetes.
 3            A causal relationship between diabetes and dioxin has not been established, although the
 4     toxicologic data are suggestive of a plausible mechanism. Many questions are yet to be
 5     answered. Does diabetes alter the pharmacokinetics of dioxin?  Diabetes is known to alter the
 6     metabolism of several drugs in humans (Matzke et al., 2000) and may also alter dioxin
 7     metabolism and kinetics. As adult-onset diabetes is also associated with overweight, and body
 8     composition has been shown to modify the apparent half-life of dioxin, could the rate of
 9     elimination of dioxins be lowered in people with diabetes, causing them to have higher body
10     burdens? This may be relevant to the background population, but is hardly likely to be an
11     explanation in highly exposed populations. Key research needs are twofold. The first is to
12     develop an animal model in which to study the association between dioxins and diabetes and
13     glucose perturbation.  Several rodent models for Type II diabetes exist and may be utilized.  The
14     second is to conduct population-based incidence studies that take into account dioxin levels as
15     well as the many known factors associated with diabetes. Although diabetes may cause the
16     underlying pathology leading to death, it is often not attributed as the cause of death, and thus
17     limits the utility of mortality studies.
18
19     2.2.6. Other Effects
20     2.2.6.1. Elevated GGT
21            As mentioned above, there appears to be a consistent pattern of increased GGT levels
22     among individuals exposed to 2,3,7,8-TCDD-contaminated chemicals. Elevated levels of serum
23     GGT have been observed within a year after exposure in Seveso children (Caramaschi et al.,
24     1981; Mocarelli et al., 1986) and 10 or more years after cessation of exposure among TCP and
25     2,4,5-T production workers (May, 1982; Martin, 1984; Moses et al., 1984; Calvert et al., 1992)
26     and among Ranch Hands (Roegner et al., 1991; Grubbs et al., 1995). All of these groups had a
27     high likelihood of substantial exposure to 2,3,7,8-TCDD. In addition, for those studies that
28     evaluated dose-response relationships with 2,3,7,8-TCDD levels, the effect was observed only at
29     the highest levels or categories of 2,3,7,8-TCDD and, in the NIOSH study, only in workers who
30     reported drinking high levels of alcohol. In contrast, although background levels of serum
31     2,3,7,8-TCDD suggested minimal  exposure to Army Vietnam veterans, GGT was increased, at
32     borderline significance, among Vietnam veterans compared to non-Vietnam veterans (Centers for
33     Disease Control Vietnam Experience Study, 1988). In addition, despite the increases observed in
34     some occupational cohorts, other studies of TCP production workers from West Virginia or
       6/8/00                                     37       DRAFT—DO NOT CITE OR QUOTE

-------
  1      Missouri residents measured but did not report elevations in GOT levels (Suskind and Hertzberg,
  2      1984; Webb etal., 1989).
  3             In clinical practice, GOT is often measured because it is elevated in almost all
  4      hepatobiliary diseases and is used as a marker for alcoholic intake (Guzelian, 1985). In
  5      individuals with hepatobiliary disease, elevations in GGT are usually accompanied by increases
  6      in other hepatic enzymes, e.g., AST and ALT, and metabolites, e.g., uro- and coproporphyrins.
  7      Significant increases in hepatic enzymes other than GGT and metabolic products were not
  8      observed in individuals whose GGT levels were elevated 10 or more years after exposure ended,
  9      suggesting  that the effect may be GGT-specific.  These data suggest that in the absence of
 10      increases in other hepatic enzymes, elevations in GGT are associated with exposure to 2,3,7,8-
 11      TCDD, particularly among individuals who were exposed to high 2,3,7,8-TCDD levels.
 12             The animal data with respect to 2,3,7,8-TCDD-related effects on GGT are sparse.
 13      Statistically significant changes in hepatic enzyme levels, particularly AST, ALT, and ALK, have
 14      been observed after exposure to 2,3,7,8-TCDD in rats and hamsters (Gasiewicz et al., 1980;
 15      Kociba et al., 1978; Olson et al., 1980). Only one study evaluated GGT levels (Kociba et al.,
 16      1978). Moderate but statistically nonsignificant increases were noted in rats fed 0.10 ng/kg
 17      2,3,7,8-TCDD daily for 2 years, and no increases were observed in control animals.
 18             In summary, GGT is the only hepatic enzyme examined that was found in a number of
 19      studies to be chronically elevated in adults exposed to high levels of 2,3,7,8-TCDD. The
 20      consistency of the findings in a number of studies suggests that the elevation may reflect a true
 21      effect of exposure, but its clinical significance is unclear. Long-term pathological consequences
 22      of elevated GGT have not been illustrated by excess mortality from liver disorders or cancer, or
 23      in excess morbidity in the available cross-sectional studies.
 24             It must be recognized that the absence of an effect in a cross-sectional study, for example,
 25      liver enzymes, does not obviate the possibility that the enzyme levels may have increased
 26      concurrent to the exposure but declined after cessation. The apparently transient elevations in
 27      ALT levels among the Seveso children suggest that hepatic enzyme levels other than GGT may
 28      react in this manner to 2,3,7,8-TCDD exposure.
 29
 30      2.2.6.2. Thyroid Function
 31             Many effects of 2,3,7,8-TCDD exposure in animals resemble signs of thyroid dysfunction
 32      or significant alterations of thyroid-related hormones.  In the few human studies that examined
 33      the relationship between 2,3,7,8-TCDD exposure and hormone  concentrations in adults, the
 34      results are mostly equivocal (Centers for Disease Control Vietnam Experience Study, 1988;
35      Roegner et al., 1991; Grubbs et al., 1995;  Suskind and Hertzberg, 1984). However,
36      concentrations of thyroid binding globulin (TBG) appear to be positively correlated with current

        6/8/00                                     38      DRAFT—DO NOT CITE OR QUOTE

-------
 1      levels of 2,3,7,8-TCDD in the BASF accident cohort (Ott et al., 1994). Little additional
 2     information on thyroid hormone levels has been reported for production workers and none for
 3     Seveso residents, two groups with documented high serum 2,3,7,8-TCDD levels.
 4            Thyroid hormones play important roles in the developing nervous system in all vertebrate
 5     species, including humans. In fact, thyroid hormones are so important in development that in the
 6     United States all infants are tested for hypothyroidism shortly after birth. Several studies of
 7     nursing infants suggest that ingestion of breast milk with a higher dioxin TEQ may alter thyroid
 8     function (Pluim et al., 1993; Koopman-Esseboom et al., 1994c; Nagayama et al., 1997).
 9     These findings suggest a possible shift in the distribution of thyroid hormones, particularly T4,
10     and point out the need for collection of longitudinal data to assess the potential for long-term
11     effects associated with developmental exposures.  The exact processes accounting for these
12     observations in humans are unknown, but when put in perspective  of animal responses, the
13     following might apply: dioxin increases the metabolism and excretion of thyroid hormone,
14     mainly T4, in the liver. Reduced T4 levels stimulate the pituitary to secrete more TSH, which
15     enhances thyroid hormone production. Early in the disruption process, the body can
16     overcompensate for the loss of T4, which may result in a small excess of circulating T4 to the
17     increased TSH. In animals given higher doses of dioxin, the body  is unable to maintain
18     homeostasis, and TSH levels remain elevated and T4 levels decrease.
19
20     2.2.6.3. Cardiovascular Disease
21            Elevated cardiovascular disease has been noted in several of the occupational cohorts
22     (Steenland et al., 1999; Sweeney et al., 1997; Flesch-Janys et al., 1995) and in Seveso (Pesatori
23     et al., 1998), as well as in the rice oil poisonings. This appears to be associated with ischemic
24     heart disease and in some cases with hypertension. In fact, recent data from the Ranch Hand
25     study indicates that dioxin may be a possible risk factor for the development of essential
26     hypertension (Grubbs, et al., 1995). Elevated blood lipids have also been seen in several cohorts.
27     The association of dioxins with heart disease in people has biological plausibility given the data
28     in animals. First is the key role of hypoxia in heart disease, and the potential for involvement of
29     the activated AhR in blocking an hypoxic response (Gradin et al., 1996; Gu et al., 2000). Dioxin
30     has been shown to perturb lipid metabolism in multiple laboratory  species (Pohjanvirta and
31     Tuomisto, 1994).  The heart, in fact the entire vascular system, is a clear target for the adverse
32     effects of dioxin in fish and birds (Hornung et al., 1999; Cheung et al., 1981). In mammals,
33     dioxin has been shown to disturb heart rhythms at high doses in guinea pigs (Gupta et al., 1973;
34     Pohjanvirta and Tuomisto, 1994).
55

       6/g/oo                                      39        DRAFT—DO NOT CITE OR QUOTE

-------
  1      2.2.6.4.  Oxidative Stress
  2            Several investigators have hypothesized that the some of the adverse effects of dioxin and
  3      related compounds may be associated with oxidative stress. Induction of CYPIA isoforms has
  4      been shown to be associated with oxidative DNA damage (Park et al., 1996). Altered
  5      metabolism of endogenous molecules such as estradiol can lead to the formation of quinones and
  6      redox cycling. This has been hypothesized to play a role in the enhanced sensitivity of female
  7      rats to dioxin-induced liver tumors (Tritscher et al., 1996). Lipid peroxidation, enhanced DNA
  8      single-strand breaks, and decreased membrane fluidity have been shown in liver as well as in
  9      extrahepatic tissues following exposure to high doses of TCDD (Stohs, 1990). A dose- and time-
10      dependent increase in superoxide anion is caused in peritoneal macrophages by exposure to
11      TCDD (Alsharif et al., 1994). A recent report that low-dose (0.15 ng TCDD/kg/day) chronic
12      exposure can lead to oxidative changes in several tissues in mice (Slezak et al., 2000) suggests
13      that this mechanism or mode of toxicity deserves further attention.
14
15
                        3. MECHANISMS AND MODE OF DIOXIN ACTION

  1            Mechanistic studies can reveal the biochemical pathways and types of biological and
  2      molecular events that contribute to dioxin's adverse effects. For example, much evidence
  3      indicates that TCDD acts via an intracellular protein (the aryl hydrocarbon receptor, AhR), which
  4      functions as a ligand-dependent transcription factor in partnership with a second protein (known
  5      as the AhR nuclear translocator, Arnt). Therefore, from a mechanistic standpoint, TCDD's
  6      adverse effects appear likely to reflect alterations in gene expression that occur at an
  7      inappropriate time and/or for an inappropriately long time. Mechanistic studies also indicate that
  8      several other proteins contribute to TCDD's gene regulatory effects and that the response to
  9      TCDD probably involves a relatively complex interplay between multiple genetic and
10      environmental factors.  If TCDD operates through such a mechanism, as all evidence indicates,
11      then there are  certain constraints on the possible models that can plausibly account for TCDD's
12      biological effects and, therefore, on the assumptions used during the risk assessment process
13      (e.g., Poland, 1996; Limbird and Taylor, 1998).
14            Mechanistic knowledge of dioxin action may also be useful in other ways. For example,
15      a further understanding of the ligand specificity and structure of the AhR will likely assist in the
16      identification of other chemicals to which humans are exposed that may add to, synergize, or
17      block the toxicity of TCDD.  Knowledge of genetic polymorphisms that influence TCDD
18      responsiveness may also allow the identification of individuals at greater risk from exposure to


        6/8/00                                     40        DRAFT—DO NOT CITE OR QUOTE

-------
 1      dioxin. In addition, knowledge of the biochemical pathways that are altered by TCDD may help
 2      identify novel targets for the development of drugs that can antagonize dioxin's adverse effects.
 3            As described below, biochemical and genetic analyses of the mechanisms by which
 4      dioxin may modulate particular genes have revealed the outline of a novel regulatory system
 5      whereby a chemical signal can alter cellular regulatory processes. Future studies of dioxin action
 6      have the potential to provide additional insights into mechanisms of mammalian gene regulation
 7      that are of a broader interest. Additional perspectives on dioxin action can be found in several
 8      recent reviews (Birnbaum, 1994a,b; Schecter, 1994; Hankinson, 1995; Schmidt and Bradfield,
 9      1996; Gasiewicz, 1997; Rowlands and Gustafsson, 1997; Denison et al, 1998; Hahn, 1998;
10      Wilson and Safe, 1998).
11            Knowledge of the mode(s) of action by which the broad class of chemicals known as
12      dioxins act may facilitate the risk assessment process by imposing bounds on the models used to
13      describe possible responses of humans resulting from exposure to mixtures of these chemicals.
14      The relatively extensive database on TCDD, as well as the more limited database on related
15      compounds, has been reviewed with emphasis on the role of the specific cellular receptor for
16      TCDD and related compounds, the AhR, in the mode(s) of action. This discussion will focus on
17      summarizing the elements of the mode(s) of dioxin action that are relevant for understanding and
18      characterizing dioxin risk for humans. These elements include:
19         •  Similarities between humans and other animals with regard to receptor structure and
20            function;
21         •  The relationship between receptor binding and toxic effects; and
22         •  The extent to which the purported mechanism(s) or mode(s) of action might contribute to
23            the diversity of biological responses seen in animals and, to some extent, in humans.
24
25            In addition, this section will identify important and relevant knowledge gaps and
26      uncertainties in the understanding of the mechanism(s) of dioxin action, and will indicate how
27      these may affect the approach to risk characterization.
28
29      3.1. MODE VERSUS MECHANISM OF ACTION
30            In the context of revising its Cancer Risk Assessment Guidelines, the EPA has proposed
31      giving greater emphasis to use of all of the data in hazard characterization, dose-response
32      characterization, exposure characterization, and risk characterization (U.S. EPA, 1996). One aid
33     to the use of more information in risk assessment has been the definition of mode versus
34     mechanism of action. Mechanism of action is defined as the detailed molecular description of a
35     key event in the induction of cancer or other health endpoints. Mode of action refers to the
36     description of key events and processes, starting with interaction of an agent with the cell,

       6/8/00                                    41       DRAFT—DO NOT CITE OR QUOTE

-------
  1      through functional and anatomical changes, resulting in cancer or other health endpoints.
  2      Despite a desire to construct detailed biologically based toxicokinetic and toxicodynamic models
  3      to reduce uncertainty in characterizing risk, few examples have emerged. Use of a mode-of-
  4      action approach recognizes that, although all of the details may not have been worked out,
  5      prevailing scientific thought supports moving forward using a hypothesized mode of action
  6      supported by data. This approach is consistent with advice offered by the National Research
  7      Council in its report entitled, Science and Judgment in Risk Assessment (NAS/NRC, 1994).
  8      Mode-of-action discussions help to provide answers to the questions:  How does the chemical
  9      produce its effect? Are there mechanistic data to support this hypothesis?  Have other modes of
 10      action been considered and rejected?  In order to demonstrate that a particular mode of action is
 11      operative, it is generally necessary to outline the hypothesized sequence of events leading to
 12      effects, identify key events that can be measured, outline the information that is available to
 13      support the hypothesis, and discuss those data that are inconsistent with the hypothesis or support
 14      an alternative hypothesis.  Following this, the information is weighed to determine if there is a
 15      causal relationship between key precursor events associated with the mode of action and cancer
 16      or other toxicological endpoint.
 17      3.2.  GENERALIZED MODEL FOR DIOXIN ACTION
 18             Dioxin and related compounds are generally recognized to be receptor-mediated
 19      toxicants. The generalized model has evolved over the years to appear as illustrated in Table 3-1
 20      and Figure 2-1.
 21
 22      3.2.1. The Receptor Concept
 23             One of the fundamental concepts that influences our approach to risk  assessment of
 24      dioxin and related compounds is the receptor concept. The idea that a drug, hormone,
 25      neurotransmitter,  or other chemical produces a physiological response by interacting with a
 26      specific cellular target molecule, i.e., a "receptor," evolved from several observations. First,
 27      many chemicals elicit responses that are restricted to specific tissues. This observation implies
 28      that the responsive tissue (e.g., the adrenal cortex) contains a "receptive" component whose
 29      presence is required for the physiologic effect (e.g., cortisol secretion).  Second, many chemicals
 30      are quite potent. For example, picomolar to nanomolar concentrations of numerous hormones
 31      and growth factors elicit biological effects.  This observation suggests that the target cell contains
 32      a site(s) to which the particular chemical binds with high affinity. Third, stereoisomers of some
 33      chemicals (e.g., catecholamines, opioids) differ by orders of magnitude in their ability to produce
 34      the same biological response.  This observation indicates that the molecular shape of the
 35      chemical strongly influences its biological activity.  This, in turn, implies that the binding site on
36      or in the target cell also has a specific, three-dimensional configuration.  Together, these types of

        6/8/00                                     42        DRAFT—DO NOT CITE OR QUOTE

-------
 1     observations support the prediction that the biological responses to some chemicals involve
 2     stereospecific, high-affinity binding of the chemicals to specific receptor sites located on or in the
 3     target cell. Many of these characteristics were noted for TCDD and related compounds.
 4            The availability of compounds of high specific radioactivity has permitted quantitative
 5     analyses of their binding to cellular components in vitro.  To qualify as a potential "receptor," a
 6     binding site for a given chemical must satisfy several criteria: (1) the binding site must be
 7     saturable, i.e., the number of binding sites per cell should be limited; (2) the binding should be
 8     reversible; (3) the binding affinity measured in vitro should be consistent with the potency of the
 9     chemical observed in vivo; (4) if the biological response exhibits stereospecificity, so  should the
10     in vitro binding; (5) for a series of structurally related chemicals, the rank order for binding
11     affinity should correlate with the rank order for biological potency; and (6) tissues that respond to
12     the chemical should contain binding sites with the appropriate properties.
13            The binding of a chemical ("ligand") to its specific receptor is assumed to obey the law of
14     mass action; that is, it is a bimolecular, reversible interaction. The concentration of the liganded,
15     or occupied, receptor [RL] is a function of both the ligand concentration [L] and the receptor
16     concentration [R] as shown in Equation 3-1:
17                                               k,
18                                                v                                         (3-1)
19                                  [L] + [R]    ^—    [RL]
20                                               k2
21            Inherent in this relationship  is the fact that the fractional occupancy (i.e., [RL]/[RJ) is a
22     function of ligand concentration [L] and the apparent equilibrium dissociation constant KD, which
23     is a measure of the binding affinity  of the ligand for the receptor, that is, [RLJ/fRJ = [L]/(KD+
24     [L]), where KD = [L] [RJ/[LR] = k2/k,. Therefore, the relationship between receptor occupancy
25     and ligand concentration is hyperbolic.  At low ligand concentrations (where [L]«KD), a small
26     increase in [LJ produces an approximately linear increase in fractional receptor occupancy.  At
27     high ligand concentration (where [L]»KD), the fractional occupancy of the receptor is already
28     very close to 1, that is, almost all receptor sites are  occupied.  Therefore, a small increase in [L]
29     is likely to produce only a slight increase in receptor occupancy.  These issues are discussed in
30     regard to TCDD binding to the AhR and dose-response in Part II, Chapter 8.
31            Ligand binding constitutes only  one aspect of the receptor concept.  By definition, a
32     receptor mediates a response, and the  functional consequences of the ligand-receptor binding
33     represent an essential aspect of the receptor concept.  Receptor theory attempts to quantitatively
34     relate ligand  binding to biological responses.  The classical "occupancy" model of Clark (1933)
35     postulated that (1) the magnitude of the  biological response is directly proportional to the fraction
36     of receptors occupied and (2) the response is maximal when all receptors are occupied.
       6/8/00                                      43        DRAFT—DO NOT CITE OR QUOTE

-------
  1      However, analyses of numerous receptor-mediated effects indicate that the relationship between
  2     receptor occupancy and biological effect is not as straightforward as Clark envisioned. In certain
  3     cases, no response occurs even when there is some receptor occupancy.  This suggests that there
  4     may be a threshold phenomenon that reflects the biological "inertia" of the response (Ariens et
  5     al., 1960).  In other cases, a maximal response occurs well before all receptors are occupied, a
  6     phenomenon that reflects receptor "reserve" (Stephenson, 1956).  Therefore, one cannot simply
  7     assume that the relationship between fractional receptor occupancy and biological response is
  8     linear. Furthermore, for a ligand (such as TCDD) that elicits multiple receptor-mediated effects,
  9     one cannot assume that the binding-response relationship for a simple effect (such as enzyme
 10     induction) will necessarily be identical to that for a different and more complex effect (such as
 11      cancer). The cascades of events leading to different complex responses (e.g., altered immune
 12     response to pathogens or development of cancer) are likely to be different, and other rate-limiting
 13     events likely influence the final biological outcome resulting in different dose-response curves.
 14     Thus, even though ligand binding to the same receptor is the initial event leading to a spectrum
 15     of biological responses, ligand-binding data may not always mimic the dose-effect relationship
 16     observed for particular responses.
 17            Another level of complexity is added when  one considers different chemical ligands that
 18     bind to the same receptor.  Relative potencies are determined by two properties of the ligand:
 19     affinity for the receptor and capacity to confer a particular response in the receptor (e.g., a
 20     particular conformational change), also called efficacy (Stephenson, 1956).  Ligands with
 21      different affinities and the same degree of efficacy would be expected to produce parallel dose-
 22     response curves with the same maximal response within a particular model system. However,
 23     ligands of the same affinity with different efficacies may result in dose-response curves that are
 24     not parallel or that differ in maximal response. Many of these issues may apply to dioxin-
 25     receptor interactions. To the extent that they do occur, they may present complications to use of
 26     the toxicity equivalence approach, particularly for extrapolation purposes. As described
 27     previously, this argues strongly for the use of all available information in setting TEFs  and
 28      highlights the important role that scientific judgment plays in the face of incomplete mechanistic
 29      understanding to address uncertainty.
 30
 31      3.2.2. A Framework to Evaluate Mode of Action
 32             EPA in its revised proposed cancer guidelines (U.S. EPA, 1999) recommends the use of a
 33      structured approach to evaluating mode of action. This approach is similar to and builds upon an
 34      approach developed within the  World Health Organization's (WHO) International Programme on
35      Chemical Safety's Harmonization Project (WHO, 2000). Fundamentally, the approach uses a
36      modification of the "Hill Criteria" (Hill, 1965), which have been used in the field of

        6/8/00                                      44       DRAFT—DO NOT CITE OR QUOTE

-------
 1      epidemiology for many years to examine causality between associations of exposures and effects.
 2     The framework calls for a summary description of the postulated mode of action, followed by the
 3     identification of key events that are thought to be part of the mode of action. These key events
 4     are then evaluated as to strength, consistency, and specificity of association with the endpoint
 5     under discussion. Dose-response relationships between the precursor key events are evaluated
 6     and temporal relationships are examined to be sure that "precursor" events actually precede the
 7     induction of the endpoint. Finally, biological plausibility and coherence of the data with the
 8     biology are examined and discussed.  All of these "criteria" are evaluated and conclusions are
 9     drawn with regard to postulated mode of action.
10            In the case of dioxin and related compounds, elements of such an approach are found for
11     a number of effects including cancer in Part II.  Application of the framework to dioxin and
12     related compounds would now stop short of evaluating the association between the chemical or
13     complex mixture and clearly adverse effects.  Instead, the approach would apply to early events,
14     e.g., receptor binding and intermediate events such as enzyme induction or endocrine impacts.
15     Additional data will be required to extend the framework to most effects, but several have data
16     that would support a framework analysis.  Several of these are discussed below.
17
18     3.2.3. Mechanistic Information, Mode of Action, and Risk Assessment
19            A substantial body of evidence from investigations using experimental animals indicates
20     that the AhR mediates the biological effects of TCDD. The key role of the AhR in the effects of
21     dioxin and related compounds is substantiated by four lines of research: (1) structure/activity
22     relationships; (2) responsive versus nonresponsive mouse strains; (3) mutant cell lines; and (4)
23     the development of transgenic mice in which the gene for the AhR has been "knocked out"
24     Birnbaum, 1994; Fernandez-Salguero et al., 1996; Lahvis and Bradfield, 1998). Dioxin appears
25     not to cause effects in the AhR knockout mouse (Fernandez-Salguero et al., 1996;  Lahvis and
26     Bradfield, 1998). It is clear that the AhR is necessary, but not sufficient, for essentially all of the
27     well-studied responses to dioxin. The AhR functions as a ligand-activated transcription factor,
28     controlling the expression of specific genes via interaction with defined nucleotide sequences in
29     the promoter regions. In order to control transcription, the TCDD-AhR complex interacts with
30     another protein, Arnt, to bind to the dioxin response element. This complex is also bound by
31     other nuclear coactivators, and/or corepressors, to bind to the transcriptional complex and initiate
32     transcription (Gu et al., 2000).  However, Arnt has many other partners that control hypoxia
33     response, neuronal differentiation, morphological branching, etc. (Gu et al., 2000). It is possible
34     that there are other mechanisms of how dioxin initiates its toxic effects, apart from its direct
35     transcriptional activation of drug metabolizing genes. It may be that the adverse effects of dioxin
36     may  result from competition of the ligand-activated AhR with other Arnt partners (Gradin et al.,

       6/8/00                                     45        DRAFT—DO NOT CITE OR QUOTE

-------
  1       1996). The AhR, Arat, and Arnt partners are all members of the PAS family of basic helix-loop-
  2      helix proteins that function as nuclear regulatory proteins (Gu et al., 2000). The PAS proteins are
  3      highly conserved, with homologous proteins being present in prokaryotes.  They play key roles in
  4      circadian rhythms and development. The embryolethality of Amt knockout mice, as well as the
  5      reduced fertility and viability of the AhR knockout mice (Abbott et al., 1999), point to a key role
  6      of these proteins in normal physiology.
  7            Another potential mechanism by which TCDD can cause effects involves the
  8      protein/protein interactions of the AhR.  When not bound to a ligand, the AhR exists in a
  9      multimeric protein complex, involving two molecules of heat shock protein 90 as well as other
 10      proteins, including AIP/XAP2/ara9, ara3, ara6, src, rel, and Rb (Carver et al., 1998; Enan and
 11       Matsumura, 1996; Puga et al., 2000a). AIP/XAP2/ara9 is  a 37 kd protein that is related to known
 12      immunophilins and involved in control of signal transduction processes.  C-src has been shown
 13      to be associated with the AhR in several tissues and is a tyrosine kinase (Enan and Matsumura,
 14      1996). Dioxin has been known to cause a rapid increase in phosphorylation upon exposure.
 15      Recent studies have shown that rel, which is a key component of the NF-kappaB complex that
 16      controls apoptosis, binds to the AhR complex (Tian et al.,  1999; Puga et al., 2000b).  Similarly,
 17      several investigators have demonstrated  an association between the AhR and the retinoblastoma
 18      protein; this has been shown to affect cell cycling (Puga et al., 2000a).
 19            Thus, the AhR may act as a negative regulator of key regulator molecules involved in
 20      phosphorylation, cell cycling, and apoptosis in its unliganded state. Upon binding of TCDD,
 21       these other proteins are now able to exert their effects. In addition, dioxin may act by competing
 22      for Arnt<, thus blocking key roles of other PAS regulatory proteins. Both of these mechanisms
 23      for the effects of dioxin are in addition to the direct role of the ligand-bound form of the receptor
 24      in control of transcription via the well-studied mechanism  of binding to a dioxin-response
 25       element in DNA.
 26             Although studies using human tissues are much less extensive, it appears reasonable to
 27      assume that dioxin's mode of action to produce effects in humans includes receptor-mediated key
 28      events.  Studies using human organs and cells in culture are consistent with this hypothesis.   A
 29      receptor-based mode of action would predict that, except in cases where the concentration of
 30      TCDD is already high (i.e., [TCDD]~KD), incremental exposure to TCDD will lead to some
 31      increase in the fraction of AhRs occupied.  However, it cannot be assumed that an increase in
 32      receptor occupancy will necessarily elicit a proportional increase in all biological response(s)
 33      because numerous molecular events (e.g., cofactors, other transcription factors, genes)
 34      contributing to the biological endpoint are integrated into the overall response. That is, the final
 35      biological response should be considered as an integration of a series of dose-response curves
36      with each curve dependent on the molecular dosimetry for each particular step.  Dose-response

        6/8/00                                      46       DRAFT—DO NOT CITE OR QUOTE

-------
 1      relationships that will be specific for each endpoint must be considered when using mathematical
 2     models to estimate the risk associated with exposure to TCDD.  It remains a challenge to develop
 3     models that incorporate all the complexities associated with each biological response.
 4     Furthermore, the parameters for each mathematical model may only apply to a single biological
 5     response within a given tissue and species.
 6            Given TCDD's widespread distribution, its persistence, and its accumulation within the
 7     food chain, it is likely that most humans are exposed to some level of dioxin; thus, the population
 8     at potential risk is large and genetically heterogeneous. By analogy with the findings in inbred
 9     mice, polymorphisms in the AhR probably exist in humans.  Therefore, a concentration of TCDD
10     that elicits a particular response in one individual may not do so in another. For example, studies
11      of humans exposed to dioxin following an industrial accident at Seveso, Italy, fail to reveal a
12     simple and direct relationship between blood TCDD levels and development of chloracne
13     (Mocarelli et al., 1991). These  differences in responsiveness to TCDD may reflect genetic
14     variation either in the AhR or in some other component of the dioxin-responsive pathway.
15     Therefore, analyses of human polymorphisms in the AhR and Arnt genes have the potential to
16     identify genotypes associated with higher (or lower) sensitivities to dioxin-related effects. Such
17     molecular genetic information may be useful in the future for accurately predicting the health
18     risks dioxin poses to humans.
19            Complex responses (such as cancer) probably involve multiple events and multiple genes.
20     For example, a homozygous recessive mutation at the hr (hairless) locus is required for TCDD's
21     action as a tumor promoter in mouse skin (Poland et al., 1982). Thus, the hr locus influences the
22     susceptibility of a particular tissue (in this case, skin) to a specific effect of dioxin (tumor
23     promotion). An analogous relationship may exist for the effects of TCDD in other tissues. For
24     example, TCDD may produce porphyria cutanea tarda only in individuals with inherited
25     uroporphyrinogen decarboxylase deficiency (Doss et al., 1984).  Such findings suggest that, for
26     some adverse effects of TCDD, the population at risk may be limited to individuals with a
27     particular genetic predisposition.
28            Other factors can influence an organism's susceptibility to TCDD.  For example, female
29     rats are more prone to TCDD-induced liver neoplasms than are males; this phenomenon is
30     related to the hormonal status of the animals (Lucier et al., 1991). In addition, hydrocortisone
31      and TCDD synergize in producing cleft palate in mice. Retinoic acid and TCDD produce a
32     similar synergistic teratogenic effect (Couture et al., 1990).  These findings indicate that, in some
33     cases, TCDD acts in combination with hormones or other chemicals to produce adverse effects.
34     Such phenomena might also occur in humans.  If so, the difficulty in assessing risk is increased,
35     given the diversity among humans in hormonal status, lifestyle (e.g., smoking, diet), and
36     chemical exposure.
       6/g/00                                      47       DRAFT—DO NOT CITE OR QUOTE

-------
 1            Dioxin's action as a tumor promoter and developmental toxicant presumably reflects its
 2      ability to alter cell proliferation and differentiation processes. There are several plausible
 3      mechanisms by which this could occur.  First, TCDD might activate a gene (or genes) that is
 4      directly involved in tissue proliferation.  Second, TCDD-induced changes in hormone
 5      metabolism may lead to tissue proliferation (or lack thereof)  and altered differentiation secondary
 6      to altered secretion of a trophic hormone. Third, TCDD-induced changes in the expression of
 7      growth factor or hormone receptors may alter the sensitivity of a tissue to proliferative stimuli.
 8      Fourth, TCDD-induced toxicity may lead to cell death, followed by regenerative proliferation.
 9      These mechanisms likely differ among tissues and periods of development, and might be
10      modulated by different genetic and environmental factors. As such, this complexity increases the
11      difficulty associated with assessing the human health risks from dioxin exposure.
12            Under certain circumstances, exposure to TCDD may elicit beneficial effects.  For
13      example, TCDD protects against the subsequent carcinogenic effects of PAHs in mouse skin,
14      possibly reflecting induction of detoxifying enzymes (Cohen et al., 1979; DiGiovanni et al.,
15      1980). In other situations, TCDD-induced changes in estrogen metabolism may alter the growth
16      of hormone-dependent tumor cells, producing a potential anticarcinogenic effect (Spink et al.,
17      1990; Gierthy et al., 1993). However, several recent studies in mice indicate that the AhR has an
18      important role in the genetic damage and carcinogenesis caused by components in tobacco smoke
1.9      such as benzo[a]pyrene through its ability to regulate CYP1A1 gene induction (Dertinger et al.,
20      1998; Shimizu et al., 2000). TCDD's biological effects likely reflect a complicated interplay
21      between genetic and environmental  factors. These issues complicate the risk assessment process
22      for dioxin.
23
24
                               4. EXPOSURE CHARACTERIZATION

 1            This section summarizes key findings developed in the exposure portion of the Agency's
 2      dioxin reassessment. The findings are developed in the companion document entitled "Part I:
 3      Estimating Exposure to Dioxin-Like Compounds."  This document is divided into four volumes:
 4      (1) Executive Summary; (2) Sources of dioxin in the United States; (3) Properties,
 5      Environmental Levels, and Background Exposures; and (4) Site-Specific Assessment Procedures.
 6      Readers are encouraged to examine  the more detailed companion document for further
 7      information on the topics covered here and to see complete literature citations.  The
 8      characterization discussion provides cross references to help readers find the relevant portions of
 9      the companion document.


        6/8/00                                     48        DRAFT—DO NOT CITE OR QUOTE

-------
 1            This discussion is organized as follows: (1) Sources; (2) Fate; (3) Environmental Media
 2     and Food Concentrations; (4) Background Exposures; (5) Potentially Highly Exposed
 3     Populations; and (6) Trends.  The key findings are presented in italics.
 4
 5     4.1. SOURCES (Cross reference: Part I, Volume 2: Sources of Dioxin-Like Compounds
 6     in the U.S.)
 7            The CDD/CDFs have never been intentionally produced other than on a laboratory scale
 8     basis for use in scientific analysis. Rather, they are generated as unintended by-products in trace
 9     quantities in various combustion, industrial and biological processes. PCBs on the other hand,
10     were commercially produced in large quantities, but are no longer commercially produced in the
11     United States. EPA has classified sources of dioxin-like compounds into five broad categories:
12
13            1.  Combustion Sources. CDD/CDFs are formed in most combustion systems. These
14               can include waste incineration (such as municipal solid waste, sewage sludge, medical
15               waste, and hazardous wastes), burning of various fuels (such as coal, wood, and
16               petroleum products), other high temperature sources (such as cement kilns), and
17               poorly or uncontrolled combustion sources (such as forest fires, building fires, and
18               open burning of wastes). Some evidence exists that very small amounts of dioxin-like
19               PCBs are produced during combustion, but they appear to be a small fraction  of the
20               total TEQs emitted.
21            2.  Metals Smelting, Refining, and Processing Sources. CDD/CDFs can be formed
22               during various types of primary and secondary metals operations including iron ore
23               sintering, steel production, and scrap metal recovery.
24            3.  Chemical Manufacturing. CDD/CDFs can be formed as by-products from the
25               manufacture of chlorine-bleached wood pulp, chlorinated phenols (e.g.,
26               pentachlorophenol, or PCP), PCBs, phenoxy herbicides (e.g., 2,4,5-T), and
27               chlorinated aliphatic compounds (e.g.,  ethylene bichloride).
28            4.  Biological and Photochemical Processes. Recent studies suggest that CDD/CDFs
29               can be formed under certain environmental conditions (e.g., composting) from the
30               action of microorganisms on chlorinated phenolic compounds. Similarly,  CDD/CDFs
31               have been reported to be formed during photolysis of highly chlorinated phenols.
32            5.  Reservoir Sources. Reservoirs are materials or places that contain previously formed
33               CDD/CDFs or dioxin-like PCBs  and have the potential for redistribution and
34               circulation of these compounds into the environment. Potential reservoirs include
35               soils, sediments, biota, water, and some anthropogenic materials. Reservoirs  become
36               sources when they have releases to the circulating environment.

       6/g/OO                                     49       DRAFT—DO NOT CITE  OR QUOTE

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

        6/8/00                                      50        DRAFT—DO NOT CITE OR QUOTE

-------
 1      defined above).  Also, it is limited to sources with releases that are created essentially
 2      simultaneously with formation.  This means that the reservoir sources are not included. As
 3      discussed below, this document does provide preliminary estimates of releases from these
 4      excluded sources (i.e., reservoirs and Class D sources) but they are presented separately from the
 5      Inventory.
 6            The inventory presents the environmental releases in terms of two reference years: 1987
 7      and 1995.  1987 was selected primarily because little empirical data existed for making source-
 8      specific emission estimates.  1995 represents the latest year that could reasonably be addressed
 9      within the timetable for producing the rest of this document.  EPA expects to conduct periodic
10      revisions to the inventory in the future to track changes in environmental releases over time.
11            Figure 4-1 displays the emission estimates to air for sources included in the Inventory and
12      shows how the emission factors and activity levels were combined to generate emission
13      estimates. Figure 4-2 compares the annual mean TEQDF-WHO98 emission estimates to air for the
14      two reference years (i.e., 1987 and 1995).
15            The following conclusions are made for sources of dioxin-like compounds included in the
16      Inventory:
17
18            •   EPA's best estimates of releases ofCDD/CDFs to air, water, and land from
19                reasonably quantifiable sources were approximately 2,800 gram (g) TEQDF-WHO98 in
20                1995 and 13,500 g TEQDrWHO98 in 1987.
21            •   The decrease in estimated releases ofCDD/CDFs between 1987 and 1995
22                (approximately 80%) was due primarily to reductions in air emissions from municipal
23                and medical waste incinerators, and further reductions are anticipated. For both
24                categories, these emission reductions have occurred from a combination of improved
25                combustion and emission controls and from the closing of a number of facilities.
26                Regulations recently promulgated or under development should result in additional
'.7                reductions in emissions from major combustion sources.  Recent data, although not
18                included in the 1995 inventory,  support this trend.
>9            •   The environmental releases ofCDD/CDFs in the United States occur from a wide
50                variety of sources, but are dominated by releases to the air from combustion sources.
J1                The current (1995) inventory indicates emissions from combustion sources are more
!2                than an order of magnitude  greater than emissions from the sum of emissions from all
 !3                other categories.
 !4            .   Insufficient data are  available to comprehensively estimate point source releases of
!5                dioxin-like compounds to water.  Sound estimates of releases to water are only
i6                available for chlorine-bleached pulp and paper mills and manufacture of ethylene

       6/8/00                                     51        DRAFT—DO NOT CITE OR QUOTE

-------
  1                dichloride/vinyl chloride monomer.  Other releases to water bodies that cannot be
  2               quantified on the basis of existing data include effluents from POTWs and most
  3               industrial/commercial sources.
  4            •  Based on the available information, the inventory includes only a limited set of
  5               activities that result in direct environmental releases to land. The only releases to
  6               land quantified in the inventory are land application of sewage  sludge, and pulp and
  7               paper mill wastewater sludges.  Not included in the Inventory's definition of an
  8               environmental release is the disposal of sludges and ash into approved landfills.
  9            •  The inventory is likely to underestimate total releases. A number of investigators
 10               have suggested that national inventories may underestimate emissions due to the
 11                possibility of unknown sources.  These possibilities have been supported with mass
 12               balance analyses suggesting that deposition exceeds emissions. The uncertainty,
 13               however, in both the emissions and deposition estimates for the United States prevent
 14               the use of this approach for reliably evaluating the issue.  As explained below, this
 15               document has instead evaluated this issue by making preliminary estimates of poorly
 16               characterized sources and listing other sources that have been reported to emit dioxin-
 17               like compounds but cannot be characterized on even a preliminary basis.
 18
 19     4.1.2. General Source Observations
 20            The preliminary release estimates for contemporary formation sources and reservoir
 21      sources are presented in Table 4-3. Table 4-4 lists all the sources that have been reported to
 22     release dioxin-like compounds but cannot be characterized on even a preliminary basis.
 23            For any given time period, releases  from both contemporary formation sources and
 24     reservoir sources determine the overall amount of the dioxin-like compounds that are being
 25     released to the open and circulating environment. Because existing information is incomplete
 26     with regard to quantifying contributions from contemporary and reservoir sources, it is not
 27     currently possible to estimate the total magnitude of release for dioxin-like compounds into the
 28      U.S. environment from all sources. For example, in terms of 1995 releases from reasonably
 29      quantifiable sources, this document estimates releases of 2,800 g TEQDF-WHO98 for
 30      contemporary formation sources and 2,900 g TEQDF-WHO98 for reservoir sources. In addition,
 31      there remains a number of unquantifiable and poorly quantified sources. No quantitative release
 32      estimates can be made for agricultural burning or for most dioxin/furan reservoirs or for any
 33      dioxin-like PCB reservoirs.  The preliminary estimate of 1995 poorly characterized contemporary
 34      formation sources is 1,900 g TEQDF-WHO9g.
35             Additional observations and conclusions about all sources of dioxin-like compounds are
36      summarized below:

        6/8/00                                     52        DRAFT—DO NOT CITE OR QUOTE

-------
 1            •   The contribution ofdioxin-like compounds to waterways from nonpoint source
 2               reservoirs is likely to be greater than the contributions from point sources.  Current
 3               data are only sufficient to support preliminary estimates of nonpoint source
 4               contributions of dioxin-like compounds to water (i.e., urban storm water runoff and
 5               rural soil erosion). These estimates suggest that, on a nationwide basis, total nonpoint
 6               releases are significantly larger than point source releases.
 7            •   Current emissions ofCDD/CDFs to the U.S. environment result principally from
 8               anthropogenic activities.  Evidence that supports this finding includes matches in
 9               time of rise of environmental levels with time when general industrial activity began
10               rising rapidly (see trend discussion in Section 4.6), lack of any identified large natural
11               sources, and observations of higher CDD/CDF body burdens in industrialized vs. less
12               industrialized countries (see discussion on human  tissue levels in Section 4.4).
13            •   Although chlorine is an essential component for the formation ofCDD/CDFs in
14               combustion systems, the empirical evidence indicates that for commercial scale
15               incinerators, chlorine levels in feed are not the dominant controlling factor for rates
16               ofCDD/F stack emissions.  Important factors that  can affect the rate of dioxin
17               formation include the overall combustion efficiency, postcombustion flue gas
18               temperatures and  residence times, and the availability of surface catalytic sites to
19               support dioxin synthesis.  Data from bench, pilot, and commercial-scale combustors
20               indicate that dioxin formation can occur by a number of mechanisms.  Some of these
21               data, primarily from laboratory and pilot-scale combustors, have shown direct
22               correlation between chlorine content in fuels and rates of dioxin formation. Other
23               data, primarily from commercial-scale combustors, show little relation with
24               availability of chlorine and rates of dioxin formation. The conclusion that chlorine in
25               feed is not  a strong determinant of dioxin emissions applies to the overall population
26               of commercial-scale combustors. For any individual commercial scale combustor,
27               circumstances may exist in which changes in chlorine content of feed could affect
28               dioxin emissions.  For uncontrolled combustion, such as open burning of household
29               waste, chlorine content of wastes may play a more significant role in affecting levels
30               of dioxin emissions than observed in commercial-scale combustors.
31            •   No significant release of newly formed dioxin-like  PCBs is occurring in the United
32               States. Unlike CDD/CDFs, PCBs were intentionally manufactured in the United
33               States in large quantities from 1929 until production was banned in 1977. Although it
34               has been demonstrated that small quantities of coplanar PCBs can be produced during
35               waste combustion, no strong evidence exists that the dioxin-like PCBs make a
36               significant  contribution to TEQ releases during combustion. The occurrences of

       6/8/00                                      53        DRAFT—DO NOT CITE OR QUOTE

-------
  1                dioxin-like PCBs in the U.S. environment most likely reflects past releases associated
  2                with PCB production, use, and disposal.  Further support of this finding is based on
  3                observations of reductions since 1980s in PCBs in Great Lakes sediment and other
  4                areas.
  5             •   It is unlikely that the emission rates of CDD/CDFs from known sources correlate
  6                proportionally with general population exposures.  Although the Emissions Inventory
  7                shows the relative contribution of various sources to total emissions, it cannot be
  8                assumed that these sources make the same relative contributions to human exposure.
  9                It is quite possible that the major sources of dioxin in food (see discussion in Section
 10                2.6 indicating that the diet is the dominant exposure pathway for humans) may not be
 11                those sources that represent the largest fractions of current total emissions in the
 12                United States. The  geographic locations of sources relative to the areas from which
 13                much of the beef, pork, milk, and fish come is important to consider.  That is, much
 14                of the agricultural areas that produce dietary animal fats are not located near or
 15                directly downwind of the major sources of dioxin and related compounds.
 16             •   The contribution of reservoir sources to human exposure may be significant.  Several
 17                factors support this  finding. First, human exposure to the dioxin-like PCBs is thought
 18                to  be derived almost completely from reservoir sources. Because one-third of general
 19                population TEQ exposure is due to PCBs, at least one-third of the  overall risk from
 20                dioxin-like compounds conies from reservoir sources. Second, CDD/CDF releases
 21                from soil via soil erosion and runoff to waterways appear to be greater than releases to
 22                water from the primary sources included in the Inventory.  CDD/CDFs in waterways
 23                can bioaccumulate in fish, leading to human exposure via consumption offish.  This
 24                suggests that a significant portion of the CDD/CDF TEQ exposure could be due to
 25                releases from the soil reservoir.  Finally, soil reservoirs could have vapor and
 26                particulate releases that deposit on plants and enter the terrestrial food chain.  The
 27                magnitude of this contribution, however,  is unknown.
 28
 29      4.2. ENVIRONMENTAL FATE (Cross reference: Part  I, Volume 3, Chapter 2)
 30             Dioxin-like compounds are widely distributed in the environment as a result of a number
 31      of physical and biological processes. The dioxin-like compounds are essentially insoluble in
 32      water, generally classified as semivolatile, and tend to bioaccumulate in animals. Some evidence
 33      has shown that these compounds can degrade in the environment, but in general they are
34      considered very persistent and relatively immobile in soils and sediments. These compounds are
35      transported through the atmosphere as vapors or attached to airborne particulates and can be
36      deposited on soils, plants, or other surfaces (by wet or dry deposition). The dioxin-like

        6/8/00                                     54       DRAFT—DO NOT CITE OR QUOTE

-------
 1      compounds enter water bodies primarily via direct deposition from the atmosphere, or by surface
 2     runoff and erosion. From soils, these compounds can reenter the atmosphere either as
 3     resuspended soil particles or as vapors. In water, they can be resuspended into the water column
 4     from sediments, volatilized out of the surface waters into the atmosphere or become buried in
 5     deeper sediments. Immobile sediments appear to serve as permanent sinks for the dioxin-like
 6     compounds. Though not always considered an environmental compartment, these compounds
 7     are also found in anthropogenic materials (such as PCP) and have the potential to be released
 8     from these materials into the broader environment.
 9            Atmospheric transport and deposition of the dioxin-like compounds are a primary
10     means of dispersal of these compounds throughout the environment. The dioxin-like compounds
11     can be measured in wet and dry deposition in most locations including remote areas. Numerous
12     studies have shown that they are commonly found in soils throughout  the world. Industrialized
13     countries tend to show similar elevated concentrations in soil, and detectable levels have been
14     found in nonindustrialized countries. The only satisfactory explanation available for this
15     distribution is air transport and deposition. Finally, by analogy these compounds would be
16     expected to behave similarly to other compounds with similar properties, and this mechanism of
17     global distribution is becoming widely accepted for a variety of persistent organic compounds.
18           The two primary pathways for  the dioxin-like compounds to enter the ecological food
19     chains and human diet are air-to-plant-to-animal and water/sediment-to-fish.  Vegetation
20     receives these compounds via atmospheric deposition in the vapor and particle phases. The
21     compounds are retained on plant surfaces and bioaccumulated in the fatty tissues of animals that
22     feed on these plants. Vapor phase transfers onto vegetation have been experimentally shown to
23     dominate the air-to-plant pathway for the dioxin-like compounds, particularly  for the lower
24     chlorinated congeners. In the aquatic food chain, dioxins enter water systems  via direct
25     discharge or deposition and runoff from watersheds. Fish accumulate  these compounds through
26     their direct contact with water, suspended particles, bottom sediments, and through their
27     consumption of aquatic organisms.  Although these two pathways are thought to normally
28     dominate contribution to the commercial food supply, others can also be important. Elevated
29     dioxin levels in cattle resulting from animal contact with PCP-treated wood have been
30     documented by the U.S. Department of Agriculture. Animal feed contamination episodes have
31     led to elevations of dioxins in poultry in the United States, milk in Germany, and meat/dairy
32     products in Belgium.
33
       6/g/00                                      55        DRAFT—DO NOT CITE OR QUOTE

-------
  1      4.3.  ENVIRONMENTAL MEDIA AND FOOD CONCENTRATIONS (Cross reference:
  2           Part I, Volume 3, Chapter 3)
  3             Estimates of the range of typical background levels of dioxin-like compounds in various
  4      environmental media are presented in Table 4-5.  Estimates for background levels of dioxin-like
  5      compounds in environmental media are based on a variety of studies conducted at different
  6      locations in North America.  Of the studies available for this compilation, only those conducted
  7      in locations representing "background"  were selected. The amount and representativeness of the
  8      data vary, but in general these data were derived from studies that were not designed to estimate
  9      national background means.  The environmental media concentrations were similar to studies in
 10      Western Europe. These data are the best available for comparing with site-specific values.
 11      Because of the limited number of locations examined, it is not known if these estimates
 12      adequately capture the full national variability. As new data are collected, these ranges are likely
 13      to be expanded and refined. The limited data on dioxin-like PCBs in environmental media are
 14      summarized in this document (Part I, Volume 3, Chapter 4).
 15             Estimates for levels of dioxin-like compounds in food are based on data from a variety of
 16      studies conducted in North America. Beef, pork, and poultry were derived from statistically
 17      based national surveys. Milk estimates were derived from a survey of a nationwide milk
 18      sampling network. Dairy estimates were derived from milk fat concentrations, coupled with
 19      appropriate assumptions for the amount of milk fat in dairy products.  Egg samples were grab
 20      samples from retail stores. Fish data were collected from a combination of field and retail
 21      outlets, and all concentrations were expressed on the basis of fresh weight in edible tissue. As
 22      with environmental media, food levels found in the United States are similar to levels found in
 23      Europe.
 24
 25             The current data on levels of dioxin-like compounds in fish and eggs are limited
 26      compared with other meats and dairy products. EPA hopes to receive additional data sets over
 27      the next few months that can be incorporated into this report before it becomes final. Issues
 28      specific to fish and eggs are discussed below:
 29             •   Fish. The data set used for deriving dioxin-like compound levels in
 30                freshwater/estuarine fish are  somewhat dated because the sample collections and
 31                chemical analysis occurred in the late  1980s. Additionally, freshwater fish used in
 32                this study were all caught in the wild and may not be representative of the commercial
 33                species commonly consumed. For example, no farm-raised fish were sampled, and
 34                they represent almost all of the commercial freshwater fish consumed. Very few
35                studies were found describing levels of dioxin-like compounds in marine fish.  The
36                currently used marine fish data for dioxin-like compounds do not represent some of

        6/8/00                                     56       DRAFT—DO NOT CITE OR QUOTE

-------
 1                the most highly consumed species in the United States (e.g., tuna, cod, salmon, etc.).
 2               EPA will continue to seek new data, but new surveys are likely to be needed to
 3               improve our understanding of dioxin levels in fish.
 4           •   Eggs. EPA is currently reviewing some unpublished egg data and, if found
 5               acceptable, will incorporate them into this report before it becomes final. Based on a
 6               preliminary analysis, it does not appear that these new data will significantly change
 7               the current background TEQ estimate for eggs, but they should provide additional
 8               support and strengthen the confidence in the estimate.  Given the low egg
 9               consumption rate, total TEQ intakes also will not be significantly affected.
10
11     4.4. BACKGROUND EXPOSURES (Cross reference: Part I, Volume 3, Chapter 4)
12     4.4.1. Tissue Levels
13           The average CDD/CDF/PCB tissue level for the general adult U.S. population appears to
14     be declining, and the best estimate of current (late  1990s) levels is 25 ppt (TEQDFP-WHO9K, lipid
15     basis).
16           The tissue samples collected in North America in the late 1980s and early 1990s showed
17     an average TEQDFP-WHO9g level of about 55 pg/g lipid. This finding is supported by a number of
18     studies which measured dioxin levels in adipose, blood, and human milk, all conducted in North
19     America. The number of people in most of these studies, however, is relatively  small and the
20     participants were not statistically selected in ways that assure their representativeness of the
21     general U.S. adult population.  One study, the 1987 National Human Adipose Tissue Survey
22     (NHATS), involved over 800 individuals and provided broad geographic coverage, but did not
23     address coplanar PCBs. Similar tissue levels of these compounds have been measured in Europe
24     and Japan during similar time periods.
25           Because dioxin levels in the environment have been declining since the 1970s (see trends
26     discussion), it is reasonable to expect that levels in food, human intake,  and ultimately human
27     tissue have also declined over this period. The changes in tissue levels are likely to lag the
28     decline seen in environmental levels, and the changes in tissue levels cannot be assumed to occur
29     proportionally with declines  in environmental levels.  CDC (2000) summarized  levels of CDDs,
30     CDFs, and PCBs in human blood collected during the time period 1995 to 1997. The individuals
31     sampled were all U.S. residents with no known exposures to dioxin other than normal
32     background.  The blood was collected from 316 individuals in six different locations with an age
33     range of 20 to 70 years. All  TEQ calculations were made assuming nondetects were equal to half
34     the detection limit. While these samples were not collected in a manner that can be considered
35     statistically representative of the national population and lack wide geographic coverage, they are
36     judged to provide a better indication of current tissue  levels in the United States than the earlier

       6/8/00                                     57       DRAFT—DO NOT CITE OR QUOTE

-------
  1     data.  PCBs 105, 118, and 156 are missing from the blood data for the comparison populations
  2     reported by CDC (2000).  These congeners account for 62% of the total PCB TEQ estimated in
  3     the early 1990s. Assuming that the missing congeners from the CDC study data contribute the
  4     same proportion to the total PCB TEQ as in earlier data, they would increase our estimate of
  5     current body burdens by another 3.3 pg TEQ/g lipid for a total PCB TEQ of 5.3 pg/g lipid and a
  6     total DFP TEQ of 25.4 pg/g lipid (see Table 4-7).
  7            This finding regarding current tissue levels is further supported by the observation that
  8     this mean tissue level is consistent with our best estimate of current intake, i.e., 1 pg/kg-d in
  9     TEQDFP-WHO98. Using this intake in a one-compartment, steady-state pharmacokinetic model
 10     yields a tissue level estimate of about 16 pg TEQDFP WHO98/g lipid (assumes TEQDFP has an
 11     effective half-life of 7 yr, 80% of ingested dioxin is absorbed into the body, and lipid volume is
 12     19 L). Because intake rates appear to have declined in recent years and steady-state is not likely
 13     to have been achieved, it is reasonable to observe higher measured tissue levels than predicted by
 14     the model.
 15            Characterizing national background levels of dioxins in tissues is uncertain because the
 16     current data cannot be considered statistically representative of the general population.  It is also
 17     complicated by the fact that tissue levels are a function of both age and birth year.  Because
 18     intake levels have varied over time, the accumulation of dioxins in a person who turned 50 years
 19     old in 1990 is different than in a person who turned 50 in 2000. Future studies should help
 20     address these uncertainties. The  National Health and Nutrition Examination Survey (NHANES)
 21     began a new national survey in 1999 that will measure blood levels of CDDs, CDFs, and PCBs
 22     126, 77, 169, and 81 in about 1,700 people per year (see http:www.cdc.gov/nchs/nhanes.htm).
 23     The survey is conducted at 15 different locations per year and is designed to select individuals
 24     statistically representative of the  civilian United States population in terms of age, race, and
 25     ethnicity. These new data should provide a much better basis for estimating national background
 26     tissue levels and evaluating trends than the currently available data.
 27
 28     4.4.2. Intake Estimates
 29           Adult daily intakes ofCDD/CDFs and dioxin-like PCBs are estimated to average 45 and
 30     25 pg TEQDFr WHO9f/day, respectively, for a total intake of 70 pg/day  TEQDFP- WHO98. Daily
 31      intake is estimated by combining exposure media concentrations (food, soil,  air) with contact
 32     rates (ingestion, inhalation).  Table 4-8 summarizes the intake rates derived by this method.
 33           The intake estimate is supported by an extensive database on food consumption rates and
 34     estimates of dioxin-like compounds in food (as discussed above).  Pharmacokinetic (PK)
 35     modeling provides further support for the intake estimates.  Applying a simple steady-state PK
36     model to an adult average CDD/CDF adipose tissue level of 18.8 ppt TEQDF-WHO9g (on a lipid

        6/8/00                                    58        DRAFT—DO NOT CITE OR QUOTE

-------
 1     basis) yields a daily intake of 110 pg TEQDF-WHO9g/day.  Insufficient half-life data are available
 2     for making a similar intake estimate for the dioxin-like PCBs. This PK-modeled CDD/CDF
 3     intake estimate is about 2.5 times higher than the direct intake estimate of 45 pg  TEQDF-
 4     WHO98/day. This difference is to be expected with this application of a simple steady-state PK
 5     model to current average adipose tissue concentrations. Current adult tissue levels reflect intakes
 6     from past exposure levels that are thought to be higher than current levels (see Trends, Section
 7     2.6). Because the direction and magnitude of the difference in intake estimates between the two
 8     approaches are understood, the PK-derived value is judged supportive of the pathway-derived
 9     estimate. It should be recognized, however, that the pathway-derived value will  underestimate
10     exposure if it has failed to capture all significant exposure pathways.
11
12     4.4.3. Variability in Intake Levels
13            CDD/CDF and dioxin-like PCB intakes for .the general population may extend to levels at
14     least three times higher than the mean. Variability in general population exposure is primarily
15     the result of the differences in dietary choices that individuals make. These are differences in
16     both quantity and types of food consumed.  A diet that is disproportionately high in animal fats
17     will result in an increased background exposure over the mean.  Data on variability of fat
18     consumption indicate that the 95th percentile is about twice the mean and the 99th percentile is
19     approximately three times the mean. Additionally, a diet that substitutes meat sources that are
20     low in dioxin (i.e., beef, pork, or poultry) with sources that are high in dioxin (i.e., freshwater
21     fish) could result in exposures elevated over three times the mean.  This scenario may not
22     represent a significant change in total animal fat consumption, even though it results in an
23     increased dioxin exposure.
24            Intakes ofCDD/CDFs and dioxin-like PCBs are over three times higher for a young child
25     as compared to that of an adult, on a body weight basis. Using age-specific food consumption
26     rate and average food concentrations, as was done above for adult intake estimates, Table 4-9
27     describes the variability in average intake values as a function of age.
28            Only four of the 17 toxic CDD/CDF congeners and one of the 11 toxic PCBs account for
29     most of the toxicity in human tissue concentrations:  2,3,7,8-TCDD, 1,2,3,7,8-PCDD,
30     1,2,3,6,7,8-HxCDD, 2,3,4,7,8-PCDF, and PCB 126. This finding is derived directly from the
31     data described earlier on human tissue levels and is supported by intake estimations indicating
32     that these congeners are also the primary contributors to dietary dose.  These five compounds
33     make up more than one-half of the total TEQ tissue level.  The variability in intake levels is also
34     supported by the blood data from CDC (2000), which showed that the 95th percentile of blood
35     level estimates, presumably resulting primarily from dietary intake, was almost twice the mean
36     level.
       6/8/00                                      59       DRAFT—DO NOT CITE OR QUOTE

-------
  1      4.5.   POTENTIALLY HIGHLY EXPOSED POPULATIONS OR DEVELOPMENTAL
  2            STAGES (Cross reference: Part I, Volume 3, Chapter 6)
  3             As discussed earlier, background exposures to dioxin-like compounds may extend to
  4      levels at least three times higher than the mean. This upper range is assumed to result from the
  5      normal variability of diet and human behaviors. Exposures from local elevated sources or
  6      exposures resulting from unique diets would be in addition to this background variability. Such
  7      elevated exposures may occur in small segments of the population such as individuals living near
  8      discrete local sources, or subsistence or recreational fishers. Nursing infants represent a special
  9      case: for a limited portion of their lives, these individuals may have elevated exposures on a body
 10      weight basis when compared with nonnursing infants and adults.
 11             Dioxin contamination incidents involving the commercial food supply have occurred in
 12      the United States and other countries. For example, in the United States, contaminated ball clay
 13      was used as an anticaking agent in soybean meal and resulted in elevated dioxin levels in some
 14      poultry and catfish. This incident, which occurred in 1998, involved less than 5% of the national
 15      poultry production and has since been eliminated. Elevated dioxin levels have also been
 16      observed in a few beef and dairy animals where the contamination was associated with contact
 17      with PCP-treated wood. Evidence of this kind of elevated exposure  was not detected in the
 18      national beef survey. Consequently its occurrence is likely to be low, but  it has not been
 19      determined. These incidents may have led to small increases in dioxin exposure to the general
 20      population. However, it is unlikely that such incidents have led to disproportionate exposures to
 21      populations living near where these incidents have occurred, because in the United States, meat
 22      and dairy products are highly distributed on a national scale. If contamination events were to
 23      occur in foods that are predominantly distributed on a local or regional scale, then such events
 24      could lead to highly exposed local populations.
 25             Elevated exposures associated with the workplace or industrial accidents have also been
 26      documented.  United States workers in certain segments of the chemical industry had elevated
 27      levels of TCDD exposure, with some tissue measurements in the thousands of ppt TCDD. There
 28      is no clear evidence that elevated exposures are currently occurring among United States
 29      workers.  Documented examples of past exposures for other groups include certain Air Force
 30      personnel exposed to Agent Orange during the Vietnam War and people exposed as a result of
 31      industrial accidents in Europe and Asia.
 32             Consumption of breast milk by nursing infants leads to higher levels of exposure and
 33      higher body burdens ofdioxins during early years of life as compared with nonnursing infants.
 34      Two studies have compared dioxins in infants who have been breast-fed versus those who have
 35      been formula-fed, and both have shown elevations in the concentrations ofdioxins in infants
36      being breast-fed. One study obtained blood samples from two infants (1 breast-fed and 1

        6/8/00                                     60       DRAFT—DO NOT CITE OR QUOTE

-------
 1      formula-fed) at 11 and 25 months and the other obtained adipose tissue from 17 infants (9 breast-
 2     fed and 8 formula-fed) who had died from Sudden Infant Death Syndrome.  Both studies showed
 3     formula-fed infants having lipid-based concentrations <5 ppt TEQDF-WH09g, while breast-fed
 4     infants had average lipid-based concentrations >20 ppt TEQDF-WHO98 (maximum of 35 ppt
 5     TEQDF-WHO98).  The dose to the infant varies as a function of infant body weight, the
 6     concentration of dioxins in the mother's milk, and the trend of dioxins in the mother's milk to
 7     decline over time. Using current data on this information and PK modeling, a 12-month nursing
 8     scenario was modeled and results include:
 9
10            •   Doses at birth could exceed 200 pg TEQDFP-WHO98/kg/day, which would drop to
11               about 20 pg TEQDFP-WH09g/kg/day after 12 months.  The average dose over a year
12               was calculated to be 77 pg TEQDFP-WHO98/kg/day. These results assumed an initial
13               concentration in the mother's milk of 25 ppt TEQDFP-WHO98, which declined to about
14               6 ppt TEQDFP-WHO98 after 1 year, and an initial infant total body weight of 3.3 kg,
15               which rose to over 9 kg after 1 year.
16            •   On a mass basis, this hypothetical exposure to dioxins in breast milk over the course
17               of a year is estimated to represent about 10% of the total lifetime dose of an
18               individual to dioxins.
19            •   Infant lipid concentrations were found to peak at about 42 ppt TEQDFP-WHO9g,
20               compared with lipid concentrations of less than 10 ppt for the formula-fed infants.
21               The dioxin concentrations in these two hypothetical children merged at about 10 years
22               of age, at a lipid concentration of about 13 ppt TEQDFP-WHO98.
23            While the average annual infant dose of 77 pg TEQDFP-WHO9g/kg/day exceeds the
24     currently estimated adult dose of 1 pg TEQDFP-WHO98/kg/day, the effect on  infant body burdens
25     is expected to be less dramatic — i.e., infant body burdens will not exceed adult body burdens
26     by 77 times. This is due to the rapidly expanding infant body weight and lipid volume, the
27     decrease in concentration of dioxins in the mother's milk over time, as well  as the possibly faster
28     elimination in infants.  As noted above by both monitoring and modeling, dioxin concentrations
29     in the lipids of breast-fed infants appear to be in the range of <20 to >40 ppt TEQDFP-WHO98,
30     which compares to the 25 ppt TEQDFP-WHO9g identified as the representative current background
31     lipid concentrations in adults.
32            Consumption of unusually high amounts offish, meat, or dairy products containing
33     elevated levels of dioxins and dioxin-like PCBs can lead to elevated exposures in comparison
34     with the general population. Most people eat some fish from multiple sources, both fresh and
35     salt water. The estimated dioxin concentrations in these fish and the typical rates of consumption
36     are included in the mean background calculation of exposure. People who consume large

       6/8/00                                    61        DRAFT—DO NOT CITE OR QUOTE

-------
  1      quantities offish at estimated contamination levels may have elevated exposures. These kinds of
  2      exposures are addressed within the estimates of variability of background and are not considered
  3      to result in highly exposed populations.  If high-end consumers obtain their fish from areas where
  4      the concentration of dioxin-like chemicals in the fish is elevated, they may constitute a highly
  5      exposed subpopulation. Although this scenario seems reasonable, no supporting data could be
  6      found for such a highly exposed subpopulation in the United States.  One study measuring
  7      dioxin-like compounds in the blood of sport fishers in the Great Lakes area showed elevations
  8      over mean background, but within the range of normal variability. Elevated CDD/CDF levels in
  9      human blood have been measured in Baltic  fishermen.  Similarly elevated levels of coplanar
 10      PCBs have been measured in the blood of fishers on the north shore of the Gulf of the St.
 11      Lawrence River who consume large amounts of seafood.
 12             Similarly, high exposures to dioxin-like chemicals as a result of consuming meat and
 13      dairy products would only occur in situations where individuals consume large quantities of these
 14      foods and the level of these compounds is elevated. Most people eat meat and dairy products
 15      from multiple sources and, even if large quantities are consumed, they are not likely to  have
 16      unusually high exposures. Individuals who raise their own livestock for basic subsistence have
 17      the potential for higher exposures if local levels of dioxin-like compounds are high. One study in
 18      the United States showed elevated levels in  chicken eggs near  a contaminated soil site.  European
 19      studies at several sites have shown elevated CDD/CDF levels in milk and other animal  products
 20      near combustion sources.
 21
 22      4.6. ENVIRONMENTAL TRENDS (Cross reference: Part I, Volume 3, Chapter  6)
 23             Concentrations ofCDD/CDFs and PCBs in the United States environment were
 24  ,    consistently low before the 1930s.  Then concentrations rose steadily until about 1970.  At this
 25      time, the trend reversed and the concentrations have declined to the present.
 26             The most compelling supportive evidence of this trend for the CDD/CDFs and PCBs
 27      comes from dated sediment core studies.  Sediment concentrations in these studies are generally
 28      assumed to be an indicator of the rate of atmospheric deposition. CDD/CDF and PCB
 29      concentrations in sediments began to increase around the 1930s and continued to increase until
 30      about 1970.  Decreases began in 1970 and have continued to the time of the most recent sediment
 31      samples (about 1990).  Sediment data from 20 United States lakes and rivers from seven separate
 32      research efforts consistently support this trend. Additionally, sediment studies in lakes located in
 33      several European countries have shown similar trends.
 34             It is reasonable to assume that sediment core trends should be driven by a similar trend in
35      emissions to  the environment. The period of increase generally matches the time when a variety
36      of industrial activities began rising and the period of decline appears to correspond with growth

        6/8/00                                     62        DRAFT—DO NOT CITE OR QUOTE

-------
 1      in pollution abatement. Many of these abatement efforts should have resulted in decreases in
 2      dioxin emissions, i.e., elimination of most open-burning, particulate controls on combustors,
 3      phase out of leaded gas, and bans on PCBs, 2,4,5-T, hexachlorophene, and restrictions on the use
 4      of PCP.  Also, the national source inventory of this assessment documented a significant decline
 5      in emissions from the late 1980s to the mid-1990s.  Further evidence of a decline in CDD/CDF
 6      levels in recent years is emerging from data, primarily from Europe, showing declines in foods
 7      and human tissues.
 8            In addition to the congener-specific PCB data discussed earlier, a wealth of data on total
 9      PCBs, Aroclors, and other commercial PCB mixtures exist that also supports these trends. It is
10      reasonable to assume that the trends for dioxin-like PCBs are similar to those for PCBs as a class
11      because the predominant source of dioxin-like PCBs is their occurrence in Aroclor mixtures.
12      PCBs were intentionally manufactured in large quantities from 1929 until production was banned
13      in the United States in 1977. United States production peaked in 1970, with a volume of 39,000
14     metric tons. Further support is derived from data showing declining levels of total PCBs in Great
15     Lakes sediments and biota during the 1970s and 1980s.  These studies indicate, however, that
16     during the 1990s the decline was slowing and may have been leveling off.
17           Past human exposures to dioxins were most likely higher than current estimates.  This is
18     supported by a study that applied a non-steady-state PK model to data on background United
19     States tissue levels of 2,3,7,8-TCDD from the 1970s and 1980s. Various possible intake
20     histories (pg/kg-day over time) were tested to see which best-fit the data.  An assumption of a
21      constant dose over time resulted in a poor fit to the data.  The "best-fit" (statistically derived) to
22     the data was found when the dose, like the sediment core trends, rose through the 1960s into the
23     1970s and declined to current levels.  Some additional support for this finding comes from a
24     limited study of preserved meat samples from several decades in the 20th century. One sample
25     from before 1910 showed very low concentrations of dioxins and coplanar PCBs.  Thirteen other
26     samples, from the 1940s until the early 1980s consistently showed elevated levels of all dioxin-
27     like compounds as compared with food surveys conducted during the  1990s.
28
29
                           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
       6/8/00                                     63       DRAFT—DO NOT CITE OR QUOTE

-------
  1      dose-response in the range of environmental exposures? (2) What is a reasonable limit (critical
  2      dose or point of departure) at the lower end of the observable range and what risk is associated
  3      with this exposure?  In addition, one can address the issue of extrapolation beyond the range of
  4      the data in light of the answers to the above questions.  Although extrapolation of risks beyond
  5      the range of observation in animals and/or humans is an inherently uncertain enterprise, it is
  6      recognized as an essential component of the risk assessment process (NAS/NRC, 1983).  The
  7      level of uncertainty is dependent on the nature (amount and scope) of the available data and on
  8      the validity of the models that have been used to characterize dose-response. These form the
  9      bases for scientific inference regarding individual or population risk beyond the range of current
 10      observation ((NAS/NRC, 1983,1994)
 11             In Part II, Chapter 8, the body of literature concerning dose-response relationships of
 12      TCDD is presented. This chapter addresses the important concept of selecting an appropriate
 13      metric for cross-species scaling of dose and presents the results of empirical modeling for many
 14      of the available data sets on TCDD exposures in humans and in animals. Although not all
 15      human observations or animal experiments are amenable to dose-response modeling, more than
 16      200 data sets were evaluated for shape, leading to an effective dose (ED) value expressed as a
 17      percent response being presented for the endpoint being evaluated (e.g., ED01 equals an effective
 18      dose for a 1% response). The analysis of dose-response relationships for TCDD, considered
 19      within the context of toxicity equivalence, mechanism of action, and background human
 20      exposures, helps to elucidate the common ground and the boundaries of the science and science
 21      policy components inherent in this risk characterization for the broader family of dioxin-like
 22      compounds.  For instance, the dose-response relationships provide a basis to infer a point of
 23      departure for extrapolation for cancer and noncancer risk for a complex mixture of dioxin-like
 24      congeners given the assumption of toxicity equivalence as discussed in Part  II, Chapter 9.
 25      Similarly, these relationships provide insight into the shape of the dose-response at the point of
 26      departure, which can help inform choices for extrapolation models for both TCDD and total
 27      TEQ.
 28             In evaluating the dose-response relationships for TCDD  as a basis for assessing this
 29      family of compounds, both empirical dose-response modeling approaches and mode-of-action-
 30      based approaches have been developed and applied (see Part II,  Chapter 8; Portier et al., 1996).
 31      Empirical models have advantages and disadvantages relative to more ambitious
 32      mechanism-based models.  Empirical models provide a simple mathematical model that
 33      adequately describes the pattern of response for a particular data set; they can also provide the
34      means for hypothesis testing and interpolation between data points.  In addition, they can provide
35      qualitative insights into underlying mechanisms. However, the major disadvantage of empirical
36      models is their inability to quantitatively link data sets in a mechanistically meaningful manner.

        6/8/00                                     64       DRAFT—DO NOT CITE OR QUOTE

-------
 1     On the other hand, mechanism-based modeling can be a powerful tool for understanding and
 2     combining information on complex biological systems. Use of a truly mechanism-based
 3     approach can, in theory, enable more reliable and scientifically sound extrapolations to lower
 4     doses and between species. However, any scientific uncertainty  about the mechanisms that the
 5     models describe is inevitably reflected in uncertainty about the predictions of the models.
 6            Physiologically based pharmacokinetic (PBPK) models have been validated in the
 7     observable response range for numerous compounds in both animals and humans.  The
 8     development of PBPK models for disposition of TCDD in animals has proceeded through
 9     multiple levels  of refinement, with newer models showing increasing levels of complexity by
10     incorporating data for disposition of TCDD, its molecular actions with the AhR and other
11     proteins, as well as numerous physiological parameters (Part II, Chapter 1). These have provided
12     insights into key determinants of TCDD disposition in treated animals.  The most complete
13     PBPK models give similar predictions about TCDD tissue dose metrics. The PBPK models have
14     been extended to generate predictions for early biochemical consequences of tissue dosimetry of
15     TCDD, such as induction of CYP1A1. Nevertheless, extension of these models to more complex
16     responses is more uncertain at this time.  Differences in interpretation of the mechanism of action
 17     lead to varying  estimates of dose-dependent behavior for similar responses.  The shape of the
 8     dose-response curves governing extrapolation to low doses are determined by these hypotheses
 9     and assumptions.
 !0            At this time, the knowledge of the mechanism of action of dioxin, receptor theory, and
 11     the available dose-response data do not firmly establish a scientific basis for replacing a linear
 !2     procedure for estimating cancer potency. Consideration of this same information indicates that
 :3     the use of different procedures to estimate the risk of exposure for cancer and noncancer
 4     endpoints may not be appropriate. Both the cancer and noncancer effects of dioxin appear to
 5     result from qualitatively similar modes of action. Initial steps in the process of toxicity are the
 6     same and many early events appear to be shared. Thus, the inherent potential for low dose
 7     significance of  either type of effect (cancer or noncancer) should be considered equal and
 8     evaluated accordingly. In the observable range around 1% excess response, the quantitative
 9     differences are relatively small.  Below this response, the different mechanisms can diverge
 0     rapidly.  The use of predicted biochemical responses as dose metrics for toxic responses is
 1     considered a potentially useful application of these models.  However, greater understanding of
 2     the linkages between these biochemical  effects and toxic responses is needed to reduce the
 3     potentially large uncertainty associated with these predictions.
      6/8/00                                     65        DRAFT—DO NOT CITE OR QUOTE

-------
  1      5.1.  DOSE METRIC(s)
  2             One of the most difficult issues in risk assessment is the determination of the dose metric
  3      to use for animal-to-human extrapolations. To provide significant insight into differences in
  4      sensitivity among species, an appropriate animal-to-human extrapolation of tissue dose is
  5      required. The most appropriate dose metric should reflect both the magnitude and frequency of
  6      exposure, and should be clearly related to the toxic endpoint of concern by a well-defined
  7      mechanism. This is, however, often difficult because human exposures with observable
  8      responses may be very different from highly controlled exposures in animal experiments. In
  9      addition, comparable exposures may be followed by very different pharmacokinetics (absorption,
 10      distribution, metabolism and/or elimination) in animals and humans. Finally, the sequelae of
 11      exposure in the form of a variety of responses related to age, organ,  and species sensitivity
 12      complicate the choice of a common dose metric. Despite these complexities, relatively simple
 13      default approaches, including body surface or body weight scaling of daily exposures, have often
 14      been recommended (U.S. EPA, 1992, 1996).
 15             Given the data available on dioxin and related compounds, dose can be expressed in a
 16      multitude of metrics (DeVito et al., 1995) such as daily intake (ng/kg/d), current body burden
 17      (ng/kg), average body burden over a given period of time, plasma concentration, etc. Examples
 18      of other dose metrics of relevance for TCDD  and related compounds can be found in the
 19      literature including concentration of occupied AhR (Jusko, 1995), induced CYP1A2 (Andersen
 20      et al., 1997; Kohn et al., 1993) and reduced epidermal growth factor receptor (EGFR) (Portier
 21      and Kohn, 1996).  Considering the variety of endpoints seen with TCDD and expected with other
 22      dioxin-like chemicals in different species, it is unlikely that a single dose metric will be adequate
 23      for interspecies extrapolation for all of these endpoints. The issue of an appropriate dose metric
 24      for developmental effects considering the potential for a narrow time window of sensitivity, for
 25      instance, has been discussed in a number of places in this document.  Furthermore, the use of
 26      different dose metrics with respect to the same endpoint may lead to widely diverse conclusions.
 27      This latter point is discussed in more detail in Part II, Chapter 8.  Nevertheless, it is possible to
 28      express dose in a form that allows for comparison of responses for selected endpoints and
 29      species. This can be done by choosing a given exposure  and comparing responses (e.g., URL) or
 30      choosing a particular response level and comparing the associated exposures (e.g., ED).
 31             As discussed above, dose can be expressed in a number of ways.  For TCDD and other
 32      dioxin-like compounds, attention has focused on the consideration of dose expressed as daily
 33      intake (ng/kg/day), body burden (ng/kg), or AUC (DeVito et al, 1995; Aylward et al, 1996). The
 34      concept of physiological time (lifetime of an animal) complicates the extrapolation, as  the
 35      appropriate scaling factor is uncertain for toxic endpoints. Because  body burden incorporates
36      differences between species in TCDD half-life (these differences are large between rodent

        6/8/00                                     66        DRAFT—DO NOT CITE OR QUOTE

-------
 1      species and humans [Table 8.2], this dose metric appears to be the most practical for this class of
 2      compounds (DeVito et al, 1995). Average lifetime body burden is best suited for steady-state
 3      conditions, with difficulties arising when this dose metric is applied to evaluation of acute
 4      exposures, such as those occurring in the 1976 accidental exposure of some people living in
 5      Seveso, Italy (Bertazzi and di Domenico, 1994).  In cases such as this, increased body burden
 6      associated with the acute exposure event is expected to decline (half-life for TCDD is
 7      approximately 7 years) until it begins to approach a steady-state level associated with the much
 8      smaller daily background intake. However, this issue of acute exposure is not a major factor in
 9      the current analyses. In general, daily excursions in human exposure are relatively  small and
10     have minor impact on average body burden. Instead, PBPK models suggest that human body
11      burdens increase over time and begin to approach steady-state after approximately 25 years with
12     typical background doses. Occupational exposures represent the middle ground where daily
13     excursions during the working years can significantly exceed daily background intakes for a
14     number of years, resulting in elevated body burdens. This is illustrated in Table 5-1.  Estimation
15     of the range and mean or median of "attained" body burden in accidentally  or occupationally
16     exposed cohorts is presented and compared with  body burdens based on background exposures.
17     These data are presented  graphically in Figure 5-1.
18            Table 5-1 and Figure 5-1 summarize literature on levels of dioxin TEQs in the
19     background human population and in commonly cited epidemiological cohorts.  Table 5-1
20     collates data on tissue lipid levels (ppt lipid adjusted) in populations, principally from serum,
21      tabulating either current levels for the background population or back calculated levels for the
22     exposed cohorts. Figure  5-1 graphs the estimated range and central tendency of the total TEQDFP
23     body burden (ng/kg whole body), combining the  range of measured 2,3,7,8-TCDD  values with
24     the estimate of the background non-2,3,7,8-TCDD TEQ level from the U.S. population in the late
25     1980s/early 1990s. TEQ levels are calculated for PCDD, PCDF, and PCBs, based on
26     TEQDFP-WHO98 values, and  assume a constant 25% body fat ratio when converting from serum
27     lipid ppt to ng/kg body burden.  Total TEQ values for the Hamburg cohort women were
28     calculated by the authors, and for this cohort the  TCDD graph includes non-TCDD TEQ. Seveso
29     values reported by Needham et al. (1999) are  based on stored  serum samples from subjects
30     undergoing medical examinations contemporaneous with the exposure, and were not back-
31      calculated.
32            For the background U.S. populations (CDC; USA ~1990s), the bars represent the range of
33     total TEQ measured in the population.  The lower shaded portion represents the variability from
34     non-2,3,7,8-TCDD derived TEQs, the upper shaded portion the variability in the 2,3,7,8-TCDD.
35     Note, that the respective bar sizes do not represent the total non-2,3,7,8-TCDD TEQ or
36     2,3,7,8-TCDD contributions, because a portion of each of these contributions is contained within
       6/8/00                                     67       DRAFT—DO NOT CITE OR QUOTE

-------
  1      the region between the x-axis and bottom of the bar, namely the minimum estimated body
  2      burden. For each of the back-calculated epidemiological cohort exposures, the bar was estimated
  3      based on the combination of two distributions: the 2,3,7,8-TCDD levels measured in the
  4      respective cohort plus the estimated range of background non-2,3,7,8-TCDD derived TEQs from
  5      the U.S. population.  The lower estimate is the combination of the lower 2,3,7,8-TCDD and
  6      lower non-2,3,7,8-TCDD TEQ contributions; the shading junction represents the variability in
  7      background U.S. population non-2,3,7,8-TCDD levels that have been added to this bar; the
  8      mean/median/geometric mean indicators represent the addition of the measured 2,3,7,8-TCDD
  9      central estimate with the mean background US population non-2,3,7,8-TCDD TEQ level
 10      (-47.6 ppt lipid, 11.9 ng/kg body burden at 25% body fat); and the upper limit is the combination
 11      of the upper 2,3,7,8-TCDD and upper non-2,3,7,8-TCDD TEQs.
 12             As discussed earlier, using background of total body burden (TEQDFP-WHO98) as a point
 13      of comparison, these often- termed "highly exposed" populations have maximum body burdens
 14      that are relatively close to general population backgrounds at the time. When compared to
 15      background body burdens of the late 1980s, many of the median values and some of the mean
 16      values fall within a range of one order of magnitude (factor of 10) and all fall within a range of
 17      two orders of magnitude (factor of 100). General population backgrounds at the time are likely
 18      to have been higher.  As these are attained body burdens, measured at the time of the Seveso
 19      accident or back-calculated to the time of last known elevated exposure, being compared to
 20      backgound, average lifetime body burdens in these cohorts will be even closer to lifetime average
 21      background levels. This will  be important if, as demonstrated for some chronic effects in
 22      animals and as assumed when relying on average body burden as a dose metric, cancer and other
 23      noncancer effects are a consequence of average tissue levels over a lifetime.  Body burdens begin
 24      to decline slowly soon after elevated exposure ceases. Some data in humans and animals suggest
 25      that elimination half-lives for dioxin and related compounds may be dose dependent, with high
 26      doses being eliminated more rapidly than lower doses. Nonetheless, the use of an approximately
 27      7-year half-life of elimination presents a reasonable approach for evaluating both back-calculated
 28      and average lifetime levels, because for most cohorts the exposure is primarily to TCDD.
 29             The ability to detect effects in epidemiologic study is dependent on a sufficient difference
 30      between control and exposed populations. The relatively small difference (<10-100 fold)
 31      between exposed and controls in these studies makes exposure characterization in the studies a
 32      particularly serious issue. This point also strengthens the importance of measured blood or tissue
 33      levels in the epidemiologic analyses, despite the uncertainties associated with calculations
34      extending the distribution of measured values to the entire cohort and assumptions involved in
35      back-calculations.


        6/8/00                                     68       DRAFT—DO NOT CITE OR QUOTE

-------
 1             Characterization of the risk of exposure of humans today remains focused on the levels
 2      of exposure that occur in the general population, with particular attention given to special
 3      populations (see Part I).  For evaluation of multiple endpoints and considering the large
 4      differences in half-lives for TCDD across multiple species, it is generally best to use body burden
 5      rather than daily intake as the dose metric for comparison unless data to the contrary are
 6      presented. Further discussion of this point, which provides the rationale for this science-based
 7      policy choice, is presented in Part II, Chapters 1 and 8.
 8
 9      5.1.1. Calculations of Effective Dose (ED)
10             Comparisons across multiple endpoints, multiple species, and multiple experimental
11      protocols are too complicated to be made on the basis of the full dose-response curve. As
12      discussed above, comparisons of this sort can be made by either choosing a given exposure and
13      comparing the responses, or choosing a particular response level and comparing the associated
14      exposures. In the analyses contained in Chapter 8 and elsewhere in the reassessment, comparison
15      of responses is made using estimated exposures associated with a given level of excess response
16     or risk. To avoid large extrapolations, this common level of excess risk was chosen such that for
17     most studies, the estimated exposure is in or near the range of the exposures seen in the studies
18     being compared, with extra weight given to the human data. A common metric for comparison is
19     the effective dose or ED, which is the exposure dose resulting in an excess response over
20     background in the studied population.  EPA has suggested this approach in calculating
21      benchmark doses (BMD) (Allen et al, 1994) and in its proposed approaches to quantifying
22     cancer risk (U.S. EPA, 1996). Although effective dose evaluation at the 10% response level
23     (ED10 or lower bound on ED10 [LED10]) is somewhat the norm, given the power of most chronic
24     toxicology studies to detect an effect, this level is actually higher than those typically observed in
25     the exposed groups in studies of TCDD impacts on humans. To illustrate,  lung cancer mortality
26     has a background lifetime risk of approximately 4% (smokers and nonsmokers combined), so
27     that even a relative risk of 2.0 (2 times the background lifetime risk) represents approximately a
28     4% increased lifetime risk. Based upon this observation and recognizing that many of the
29     TCDD-induced endpoints studied in the laboratory include 1% effect levels in the experimental
30     range, Chapter 8 presents effective doses of 1% or ED01. The use of ED values below 10% is
31      consistent with the Agency's guidance on the use of  mode of action in assessing risk, as
32     described in the evaluation framework discussed in Section 3.3, in that the observed range for
33     many "key events" extends down to or near the 1% response level. Determining the dose at
34     which key events for dioxin toxicity begin to be seen in a heterogeneous human population
35      provides important information for decisions regarding risk and safety.
36

       6/8/00                                     69        DRAFT—DO NOT CITE OR QUOTE

-------
  1      5.2.  EMPIRICAL MODELING OF INDIVIDUAL DATA SETS
  2             As described in Chapter 8, empirical models have advantages and disadvantages relative
  3     to more ambitious mechanism-based models.  Empirical models provide a simple mathematical
  4     model that adequately describes the pattern of response for a particular data set and can also
  5     provide the means for hypothesis testing and interpolation between data points. In addition, they
  6     can provide qualitative insights into underlying mechanisms.  However, the major disadvantage
  7     is their inability to quantitatively link data sets in a mechanistically meaningful manner. Data
  8     available for several biochemical and toxicological effects of TCDD. and on the mechanism of
  9     action of this chemical, indicate that there is good qualitative concordance between responses in
 10     laboratory animals and humans (see Table 1). For example, human data on exposure and cancer
 11      response appear to be qualitatively consistent with animal-based risk estimates derived from
 12     carcinogenicity bioassays (see Part II, Chapter 8). These and other data presented throughout this
 13     reassessment would suggest that animal models are generally an appropriate basis for estimating
 14     human responses. Nevertheless, there are clearly differences in exposures and responses between
 15     animals and humans, and recognition of these is essential when using animal data to estimate
 16     human risk. The level of confidence in any prediction of human risk depends on the degree to
 17     which the prediction is based on an accurate description of these interspecies extrapolation
 18     factors. See Chapter 8 for a further discussion of this point.
 19            Almost all data are consistent with the hypothesis that the binding of TCDD to the AhR is
 20     the first step in a series of biochemical, cellular, and tissue changes that ultimately lead to toxic
 21      responses observed in both experimental animals and humans (see Part II, Chapter 2). As such,
 22     an analysis of dose-response data and models should use, whenever possible, information on the
 23     quantitative relationships among ligand (i.e., TCDD) concentration, receptor occupancy, and
 24     biological response. However, it is clear that multiple dose-response relationships are possible
 25     when considering ligand-receptor mediated events. For example, dose-response relationships for
 26     relatively simple responses, such as enzyme induction, may not accurately predict dose-response
 27     relationships for complex responses such  as developmental effects and cancer. Cell- or
 28      tissue-specific  factors may determine the quantitative relationship between receptor occupancy
 29      and the ultimate response. Indeed, for TCDD there are much experimental data from studies
 30      using animal and human tissues to indicate that this is the case. This serves as a note of caution,
 31      as empirical data on TCDD are interpreted in the broader context of complex exposures to
 32      mixtures of dioxin-like compounds as well as to non-dioxin-like toxicants.
 33             As for other chemical mechanisms where high biological potency is directed through the
 34      specific and high-affinity interaction between chemical and critical cellular target, the
 35      supposition of  a response threshold for receptor-mediated effects is a subject for scientific
36      debate.  The basis of this controversy has  been ret jntly summarized (Sewall and Lucier, 1995).

        6/8/00                                    70       DRAFT—DO NOT CITE OR QUOTE

-------
 1             Based on classic receptor theory, the occupancy assumption states that the magnitude of
 2     biological response is proportional to the occupancy of receptors by drug molecules. The
 3     "typical" dose-response curve for such a receptor-mediated response is sigmoidal when plotted
 4     on a semilog graph or hyperbolic if plotted on a arithmetic plot. Implicit in this relationship is
 5     low-dose linearity (0-10% fractional response) through the origin.  Although the law of mass
 6     action predicts that a single molecule of ligand can interact with a receptor, thereby inducing a
 7     response, it is also stated that there must be some dose that is so low that receptor occupancy is
 8     trivial and therefore no perceptible response is obtainable.
 9            Therefore, the same receptor occupancy assumption of the classic receptor theory is
10     interpreted by different parties as support for and against the existence of a threshold. It has been
11     stated that the occupancy assumption cannot be accepted or rejected on experimental or
12     theoretical grounds (Goldstein et al., 1974).  To determine the relevance of receptor interaction
13     for TCDD-mediated responses, one must consider (1) alternatives as well as limitations of the
14     occupancy theory; (2) molecular factors contributing to measured endpoints; (3) limitations of
15     experimental methods; (4) contribution of measured effect to a relevant biological/toxic
16     endpoint; and (5) background exposure.
17            Throughout this reasssessment, each of these considerations has been explored within the
18     current context of the understanding of the mechanism of a action of TCDD, of the
19     methods for analysis of dose-response for cancer and noncancer endpoints, and of the available
20     data sets of TCDD dose and effect for several rodent species, as well as humans that were
21     occupationally exposed to TCDD at  levels exceeding the exposure of the general population.
22
23     5.2.1.  Cancer
24            As described in Section 2.2.1.4, TCDD has been classified as a human carcinogen, and is
25     a carcinogen in all species and strains of laboratory animals tested. The epidemiological
26     database for TCDD, described in detail in Part II, Chapter 7a, suggests that exposure may be
27     associated with increases in all cancers combined, in respiratory rumors and, perhaps, in soft-
28     tissue sarcoma. Although there  are sufficient data in animal cancer studies to model dose-
29     response for a number of tumor  sites, as with many chemicals, it is generally difficult to find
30     human data with sufficient information to model dose-response relationships. For TCDD, there
31     exist three studies of human occupational exposure with enough information to perform a
32     quantitative dose-response analysis. These are the NIOSH study (Fingerhut et al., 199la), the
33     Hamburg cohort study (Manz et al., 1991), and the BASF cohort study (Zober et al., 1990).  In
34     Part II, Chapter 8, simple empirical models were applied to these studies for which
35     exposure-response data for TCDD are available in human populations.

       6/8/00                                      71       DRAFT—DO NOT CITE OR QUOTE

-------
  1             Modeling cancer in humans uses slightly different approaches from those used in
  2     modeling animal studies.  The modeling approach used in the analysis of the human
  3     epidemiology data for all cancers combined and lung cancer involves applying estimated human
  4     body burden to cancer response and estimating parameters in a linear risk model for each data
  5     set. A linear risk model was used because the number of exposure groups available for analyses
  6     was too small to support more complicated models.  Because of this, evaluating the shape of the
  7     dose-response data for the human studies was not done. Access to the raw data may make it
  8     possible to use more complicated mathematical forms that allow for the evaluation of shape. In
  9     the one  case in which this has been done, the dose-response shape suggested a response that was
 10     less than linear (dose raised to a power <1) (Becher et al.,  1998). For these studies, there are
 11      several assumptions and uncertainties involved in modeling the data, including extrapolation of
 12     dosage, both in back-calculation and in elimination kinetics, and the type of extrapolation model
 13     employed.
 14            As described in Part II, Chapter 8, the data used  in the analyses are from Aylward et al.
 15     (1996) for the NIOSH study, Flesch-Janys et al. (1998) for the Hamburg cohort, and Ott and
 16     Zober (1996a,b) for the BASF cohort. The limited information available from these studies is in
 17     the form of standard mortality ratios (SMRs) and/or risk ratios by exposure subgroups with some
 18     estimate of cumulative subgroup exposures. Exposure subgroups were defined either by  number
 19     of years of exposure to dioxin-yielding processes or by extrapolated  TCDD levels.  No study
 20     sampled TCDD blood serum levels for more than a fraction of its cohort, and these samples were
 21      generally taken decades after last known exposure. In each study, serum fat or body fat levels of
 22     TCDD were back calculated using a first-order model.  The assumed half-life of TCDD used in
 23     the model varied from study to study. Aylward et al. used the average  TCDD levels of those
 24     sampled in an exposure subgroup to represent the entire subgroup. Flesch-Janys et al. and Ott
 25     and Zober performed additional calculations, using regression procedures with data on time spent
 26     at various occupational tasks, to estimate TCDD levels for all members of their respective
 27     cohorts. They then divided the cohorts into exposure groups based on the estimated TCDD
 28     levels.  The information presented in the literature cited above was used to calculate estimated
 29      average  TCDD dose levels in Chapter 8.
 30           To provide ED01 estimates for comparison in Chapter 8,  Poisson regression (Breslow and
 31      Day, 1987) was used to fit a linear model to the data described above.  Analysis of animal cancer
 32      data suggests a mixture of linear and nonlinear responses with linear shape parameters
 33      predominating; complex responses to TCDD, both cancer and noncancer, are more often than not
 34      nonlinear.  Besides the issue of use of a linear model, several other important uncertainties
 35      discussed in Chapter 8 are the representativeness and precision of the dose estimates that were
36      used, the choice of half-life and whether it is dose dependent, and potential interactions between

        6/8/00                                     72       DRAFT—DO NOT CITE OR QUOTE

-------
 1      TCDD and smoking or other toxicants.  Nevertheless, with these qualifications, it is possible to
 2     apply simple empirical models to studies in which exposure data for TCDD are available in
 3     human populations.
 4           The analysis of these three epidemiological studies of occupationally exposed individuals
 5     suggest an effect of TCDD on all cancers, and on lung cancers in the adult human male. The
 6     EDOIs based upon average excess body burden of TCDD ranged from 6 ng TCDD/kg to 161  ng
 7     TCDD/kg in humans. The lower bounds on these doses (based on a modeled 95% C.I.) range
 8     from 3.5 ng TCDD/kg to 77 ng TCDD/kg. For the effect of TCDD on lung cancers, the only
 9     tumor site increased in both rodents and humans, the human ED01s ranged from 24 ng/kg to 161
10     ng/kg. The lower bounds on these doses (based on a modeled 95% C.I.) range from 10.5 ng
11     TCDD/kg to 77 ng TCDD/kg.  These estimates of ED01s are compared to animal estimates later
12     in this discussion.
13           Both empirical and mechanistic models were used to examine cancer dose-response in
14     animals.  Portier et al. (1984) used a simple multistage model of carcinogenesis with up to two
15     mutation stages affected by exposure to model the five tumor types observed to be increased  in
16     the 2-year feed study of Kociba et al. (Sprague-Dawley rats, 1978) and the eight tumor types
17     observed to be increased  in the 2-year gavage cancer study conducted by the National Toxicology
18     Program (Osborne-Mendel rats and B6C3F, mice, 1982a). The findings from this analysis,
19     which examined cancer dose-response within the range of observation are presented in their
20     Table 8.3.2., which is reproduced with slight modifications as Table 5-2.  All but one of the
21     estimated ED01s are above the lowest dose used in the experiment (approximately 1 ng
22     TCDD/kg/day in both studies) and are thus interpolations rather than extrapolations. The
23     exception, liver cancer in female rats from the Kociba study, is very near the lowest dose used in
24     this study and is only a small extrapolation (from 1 ng TCDD/kg/day to 0.77 ng TCDD/kg/day).
25     Steady-state body burden calculations were also used to derive doses for comparison across
26     species.  Absorption was assumed to be 50% for the Kociba et al. study (feed experiment) and
27     100% for the NTP study  (gavage experiment). Also presented in Table 5-2 are the shapes of the
28     dose-response curves as determined by Portier et al. (1984).
29           The predominant  shape of the dose-response curve in the experimental region is linear;
30     this does not imply that a nonlinear model such as the quadratic or cubic would not fit these data.
31     In fact, it is unlikely that  in any one case, a linear model or a quadratic model could be rejected
32     statistically for these cases. These studies had only three experimental dose groups, hence these
33     shape calculations are not based upon sufficient doses to guarantee a consistent estimate; they
34     should be viewed with caution. The ED0! steady-state body burdens range from a low value  of
35     14 ng/kg based upon the linear model associated with liver tumors in female rats to as high as
36     1,190 ng/kg based upon a cubic model associated with thyroid follicular cell adenomas in female

       6/8/00                                     73       DRAFT—DO NOT CITE OR QUOTE

-------
  1      rats. Lower bounds on the steady-state body burdens in the animals range from 10 ng TCDD/kg
  2      to 224 ng/kg.  The corresponding estimates of daily intake level at the ED01 obtained from an
  3      empirical linear model range from 0.8 to 43 ng TCDD/kg body weight/day depending on the
  4      tumor site, species, and sex of the animals investigated.  Lower confidence bounds on the
  5      estimates of daily intake level at the ED01 in the animals range from 0.6 to 14 ng TCDD/kg body
  6      weight/day. In addition, using a mechanistic approach to modeling, Portier and Kohn (1996)
  7      combined the biochemical response model of Kohn et al. (1993) with a single initiated phenotype
  8      two-stage model of carcinogenesis to estimate liver tumor incidence in female Sprague-Dawley
  9      rats from the 2-year cancer bioassay of Kociba et al. (1978).  By way of comparison, the ED0,
10      estimate obtained from this linear mechanistic model was 0.15 ng TCDD/kg body weight/day
11      based on intake, which is equivalent to 2.7 ng TCDD/kg steady-state body burden. No lower
12      bound on this modeled estimate of steady-state body burden was provided.
13            As discussed in Part II, Chapter 8, different dose metrics can lead to widely diverse
14      conclusions.  For example, as described  in Chapter 8, the ED0, intake for the animal  tumor sites
15      presented above ranges from less than 1  to tens of ng/kg/day, and the lowest dose with an
16      increased tumorigenic response (thyroid tumors) in a rat is 1.4 ng/kg/day (NTP, 1982a). The
17      daily intake of TCDD in humans  is estimated to be 0.14 to 0.4 pg TCDD/kg/day.  This implies
18      that humans are exposed to doses 3,500 to 10,000 times lower than the lowest tumorigenic daily
19      dose in rat thyroid. However, 1.4 ng/kg/d in the rat leads to a steady-state body burden of
20      approximately 25 ng/kg, assuming a half-life of TCDD of 23 days and absorption from feed of
21      50%2. If the body burden of TCDD in humans is approximately 5 ng TCDD/kg lipid or 1.25
22      ng/kg body weight (assuming about 25% of body weight is lipid), humans are exposed to about
23      20 times less TCDD than the minimal carcinogenic  dose for the rat. If total TEQ is considered
24      the difference is even less, approaching only a factor of 2 difference. The difference between
25      these two estimates is entirely due to the approximately 100-fold difference in the half-life
26      between humans and rats. At least for this comparison, if cancer is a function of average levels
27      in the body, the most appropriate metric for comparison is the average or steady-state body-
28      burden, since the large differences in animal to human half-life are accounted for.  Comparisons
29      of human and animal ED01s from Part II, Chapter 8, for cancer response on a body-burden basis
30      show approximately equal potential for the carcinogenic effects of TCDD. In humans, restricting
31      the analysis to log-linear models in Part II, Chapter 8, resulted in cancer ED01s ranging from 6
32      ng/kg to 161 ng/kg. This was similar to  the empirical modeling estimates from the animal
        • 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.
        6/8/00                                     74       DRAFT—DO NOT CITE OR QUOTE

-------
 1      studies, which ranged from 14 ng/kg to 1,190 ng/kg  (most estimates were in the range from 14
 2      to 500 ng/kg). The lower bounds on the human body-burdens at the ED01s (based on a modeled
 3      95% C.I.) range from 3.5 ng TCDD/kg to 77 ng TCDD/kg.  Lower bounds on the steady-state
 4      body burdens in the animals range from 10 ng TCDD/kg to 224 ng/kg. The estimate for the
 5      single mechanism-based model presented earlier (2.7 ng/kg) was approximately 2 times lower
 6      than the lower end of the range of human ED01 estimates and less than the lower bound on the
 7      LED01. The same value was approximately 5 times lower than the lower end of the range of
 8      animal ED01 estimates and less than 4 times less than the LED0).
 9            Using human and animal cancer ED01s, their lower bound estimates, and the value of
10     2.7 ng TCDD/kg from the single mechanism-based model, slope factors and comparable risk
11      estimates for a human background body burden of approximately 5 ng TEQ/kg (20 ng TEQ/kg
12      lipid) can be calculated using the following equations:
13           Slope factor (per pg TEQ/kgBW/day) = risk at ED01 / intake (pg TEQ/kgBW/day)
14     associated with human equivalent steady-state body burden at ED01 where:
15     Risk at ED01 = 0.01; and
16     Intake (pgTEQ/kgBW/day) = [body burden at ED., fag TEO/kg)*half-life Cdavs)] * f      (5-1)
17                                                Ln(2)
18     half-life = 2,593 days in humans and 25 days in rats (see Table 8.1 in Part II, Chapter 8)
19     f = fraction of dose absorbed; assumed to be  50% for absorption from food (Kociba et al., 1976)
20     and 100% for other routes.
21
22           Upper bound  on excess risk at human background body burden = (human          (5-2)
23           background body burden (ng/kg))(risk at ED01)/lower bound on human equivalent
24           steady-state body burden (ng/kg) at EDO 1, where:
25
26     Risk at ED01 = 0.01
27
28           Use of these approaches reflects methodologies being developed within the context of the
29     revised draft Cancer Risk Assessment  Guidelines.  Slopes are estimated by a simple proportional
30     method at the "point  of departure" (LED01) at the low end of the range of experimental
31      observation.  As discussed below, these methods can be compared to previous approaches using
32     the linearized multistage (LMS) procedure to determine if the chosen approach has significantly
33     changed the estimation of slope.  The estimates of ED01/LED01 represent the human-equivalent
34     body burden for 1% excess cancer risk based on exposure to TCDD and are assumed for
35      purposes of this analysis to be equal for TCDD equivalents  (total TEQ). This assumption is
36      based on the toxicity  equivalence concept discussed throughout this report and in detail in Part II,
       6/8/00                                    75        DRAFT—DO NOT CITE OR QUOTE

-------
  1      Chapter 9. All cancer slope factors can be compared to the Agency's previous slope factor of
  2      1.6x 10'4perpgTCDD/kgBW/day(orl.6x 105permgTCDD/kgBW/day) (U.S. EPA, 1985).
  3
  4      5.2.1.1. Estimates of Slope Factors and Risk at Current Background Body Burdens Based on
  5      Human Data
  6             Estimates of upper bound slope factors (per pg TCDD/kgB W/day) calculated from the
  7      human ED01s presented in Table 8.3.1 range from 5.3 *  10'3, if the LED01 for all cancer deaths in
  8      the Hamburg cohort is used, to 2.4 x 10'4 if the ED0] for lung cancer deaths in the smaller BASF
  9      cohort is used.  All of the other slope factors for all cancer deaths or lung cancer deaths in the
 10      three cohorts would fall within this range. LED01s for all cancer deaths span approximately an
 11      order of magnitude and would generate slope factors in the range of 5 x 10~3 to 5  x 10"4. Slightly
 12      smaller slope factors are generated when LED0,s for lung cancer are used. The largest slope
 13      factors based on LED01s come from the Hamburg cohort (5.3 x 10"3 and 1.8 x 10"3 respectively
 14      for all cancer deaths and lung cancer deaths.)  These estimates compare well with the estimates
 15      of risk associated with TCDD exposure in the Hamburg cohort published by Becher et al. (1998).
 16      The risk estimates of Becher et al. derived from data on TCDD exposure to male workers with a
 17      10-year latency and taking greater caution over other factors affecting risk including choice of
 18      model, latency, job category, dose metric, and concurrent exposures. These estimates range from
 19      1.3 x 10'3 to 5.6 x 10~3 per pg TCDD/kgB W/day. In this analysis all excess cancers are
 20      attributed to TCDD exposure, despite significant levels of other dioxin-like compounds in blood
 21      measurements of this cohort (see Table 5-1). Although risk estimates using TCDD alone in this
 22      cohort might suggest an overestimate of risk, no evidence for this emerged from the  analysis and,
 23      assuming that TCDD will still dominate total TEQ in this population, differences in  slope factor
 24      estimates are likely to be less than a factor of 2 and may not be discemable. Taking into account
 25      different sources of variation, Becher et al. (1998) suggest a range of 10"3 to 10"2  for additional
 26      lifetime cancer risk for a daily intake of 1 pg TCDD/kg BW/day.  By inference, that range could
 27      also apply to total TEQ intake. As described in Section 4.4.2, current estimates of intake in the
 28      United States are estimated to be approximately 1 pg TEQ/kg BW/day.  Using Equation 5-2, the
 29      upper bound range of risks estimated from current human body burdens of 5  ng TEQ/kgBW
 30      (which equates to a serum level of 20 pg/g lipid [see Table 4.7]) based on all cancer  deaths in the
 31      three cohorts ranged from 1.4 x 10"2 to 1.3 x 10"3; based on lung cancer deaths, the upper bound
 32      on the estimates of excess risk extended to 6 x 10'4. The range of these estimates provides
 33      further support for the perspective on risk provided by Becher et al. (1998). Uncertainties
 34      associated with these estimates from human studies are discussed in Part II, Chapter 8, and in
35      Becher etal. (1998).
36

        6/8/00                                     76       DRAFT—DO NOT CITE OR QUOTE

-------
 1      5.2.1.2.  Estimates of Slope Factors and Risk at Current Background Body Burdens Based on
 2     Animal Data
 3           Upper bound slope factors (per pg TCDD/kgBW/day) for human cancer risk calculated
 4     from lower bounds in ED0,s (LEDs01) for the animal cancers presented in Table 5-2 range from
 5     1.9 x 10'3 to 8.4 x 10'5. This spans a range from being 12 times greater than the previous upper
 6     bound estimate on cancer slope (1.6 x 1Q-4 [U.S. EPA, 1985]) to 2 times less. The largest slope
 7     factor is derived from the same study as the  1985 estimate; that is, the slope factor derived from
 8     the female liver cancer in the Kociba et al. (1978) study continues to give the largest slope factor.
 9     In attempting these comparisons, two issues became apparent. First, the body burden and the
10     intake at the ED0, from Portier et al. (1984) does not result in the same slope factor as U.S. EPA
11     (1985). Despite the use of the same study results, a slope factor of 1.8 x Q"5 per pg
12     TCDD/kgBW/day results using the LMS approach. This is a factor of approximately 10 lower
13     than the EPA (1985) estimate of the slope. The differences are attributable to the aims of the
14     respective calculations at the time. Portier et al. (1984) calculated "virtually safe doses"
15     assuming that rodent and human doses scaled on a mg/kg basis, and he used the original tumor
16     counts from the study.  EPA (1985), on the other hand, used (BW)2/3 to arrive at a human
17     equivalent dose and used the pathology results from a reread of the original Kociba study (U.S.
18     EPA, 1980).  In addition, tumor counts were adjusted for early mortality in the study. The factor
19     to adjust for (BW)3/4-scaling in the rat is 5.8. The  correction for early mortality can be accounted
20     for with a factor of 1.6 (this is the ratio of the intake values at the ED01 with and without the
21     early mortality correction). If the Portier et al. slope factor (1.8 x 10~5 per pg TCDD/kgBW/day)
22     is multiplied by these two factors, a slope of 1.7 x 10~4per pg TCDD/kgBW/day is calculated.
23     This is equivalent to the U.S. EPA (1985) estimate of 1.6 x 10'4per pg TCDD/kgBW/day.
24     Reconciling these issues  is important to ensure appropriate comparisons of slope factor
25     estimates.
26           More important is the calculation of slope factor estimates using current methods of
27     analysis that recognize the importance of the dose metric and the differences in half-life of
28     dioxins in the bodies of laboratory animals and humans (see Part II, Chapter 8, for detailed
29     discussion). The major difference between the approaches  used to calculate risks in the mid-
30  .   1980s (Portier et  al., 1984; U.S. EPA, 1985) and the current approach is the use of body burden
31     as the dose metric for animal-to-human dose equivalence. All things being equal, the use of body
32     burden accounts for the approximately 100-fold difference  between half-lives of TCDD in
33     humans and rats (2,593 days versus 25 days [see Part II, Table 8.1]). Use of Equation 5-1 results
34     in an estimated body burden at the LED01  of 6.1 ng TEQ/kg to be derived from the EPA (1985)
35     Kociba tumor counts. This compares favorably with the Portier estimate of 10 ng TEQ/kg found
36     in Table 5-2.  The difference is entirely accounted for by the early deaths adjustment by EPA

       6/8/00                                     77       DRAFT—DO NOT CITE OR QUOTE

-------
  1      (1985). Use of these body burdens at the LEDOI results in slope factor estimates of 1.9 * 10"3 per
  2      pg TCDD/kgBW/day and 3.1 * 1O'3 per pg TCDD/kgBW/day for the Chapter 8 and the newly
  3      derived body burden, respectively. Again, the difference is due solely to the adjustment for early
  4      mortality and EPA believes this provides a better estimate of upper bound lifetime risk than does
  5      the unadjusted. EPA's new slope factor (3.1 x 10'3per pgTCDD/kgBW/day) is 19 times greater
  6      than the slope factor from 1985.
  7            A second issue with the modeling of the Kociba data relates to the appropriate tumor
  8      counts to use.  As mentioned in Section 2, Goodman and Sauer (1992) reported a second
  9      re-evaluation of the female rat liver tumors in the Kociba study using the latest pathology criteria
 10      for such lesions. Results of this review are discussed in more detail in Part II, Chapter 6. The
 11      review confirmed only approximately one-third of the tumors of the previous review (U.S. EPA,
 12      1980). Although this finding did not change the determination of carcinogenic hazard because
 13      TCDD induced tumors in multiple sites in this study, it does have an effect on evaluation of
 14      dose-response and on estimates of risk. Because neither the original EPA (1985) slope factor
 15      estimate nor that of Portier et al. (1984) reflect this reread, it is important to factor these results
 16      into the estimate of the ED01  and slope  factor. Using the LMS procedure used by EPA in 1985
 17      and the tumor counts as reported in Part II, Chapter 6, Table 6.2, the revised slope factor is
 18      reduced by approximately 3.6-fold to yield a slope factor of 4.4 x 10"5 per pg TCDD/kgB W/day.
 19      However, because the original estimates used a (BW)3/4 scaling, this must be adjusted to use body
 20      burden and obtain an appropriate result. When dose is adjusted and Equation 5-1 is used, an
 21      LED01 of 22.2 ng TEQ/kg and a slope factor of 8.3 x l O'4 per pg TCDD/kgB W/day are derived.
 22      This represents EPA's most current upper bound estimate of human cancer risk based on animal
 23      data.  It is 5.2 times larger than the slope factor calculated in U.S. EPA (1985). This number
 24      reflects the increase in slope  factor based on use of the body burden dose metric (19 times
 25      greater) and the use of the Goodman and Sauer (1992) pathology (3.6 times less).
 26
 27      5.2.1.3. Estimates of Slope Factors and Risk at Current Background Body Burdens Based on
 28      a Mechanistic Model
 29            As discussed above, Portier and Kohn (1996) combined the biochemical response model
 30      of Kohn et al.  (1993) with a single initiated-phenotype two-stage model  of carcinogenesis to
 31      estimate liver tumor incidence in female Sprague-Dawley rats from the Kociba et al. (1978)
 32      bioassay. The model is described in more detail in Part II, Chapter 8. This model adequately fit
 33      the tumor data, although it overestimated the the observed tumor response at the lowest dose in
 34      the Kociba study. The  shape of the dose-response curve was approximately linear and the
35      estimated ED01 value for this model was 1.3 ng/kg/day. The corresponding body burden giving a
36      1% increased effect was 2.7 ng/kg. The model authors believe that the use of CYP1A2 as a dose

        6/8/00                                    78       DRAFT—DO NOT CITE OR QUOTE

-------
 1     metric for the first mutation rate is consistent with its role as the major TCDD-inducible estradiol
 2     hydrolase in liver and with its hypothesized role in the production of estrogen metabolites leading
 3     to increased oxidative DNA damage and increased mutation (Yager and Liehr, 1996; Hayes et
 4     al., 1996; Barman et al., 1986; Roy et al., 1992). Although no lower bound estimate of the ED01
 5     is calculated, a maximum likelihood estimate of the slope factor can be calculated. It is
 6     7.1 x 10° per pg TCDD/kgBW/day. This estimate represents an example of the type of modeling,
 7     based on key events in a mode of action for carcinogenesis, which is consistent with future
 8     directions in dose-response modeling described in EPA's revised proposed cancer risk
 9     assessment guidelines (U.S. EPA, 1999). Although a number of uncertainties remain regarding
10     structure and parameters of the model, the slope estimate is  consistent with those derived from
11     humans and animals.  More details on this model can be found in Part II, Chapter 8.
12
13      5.2.2. Noncancer Endpoints
14            At this point, sufficient data are not available to model noncancer endpoints in humans.
15      Many studies are available to estimate ED0] values for noncancer endpoints in animals. However,
16    there are a number of difficulties and uncertainties that should be considered when comparing the
17     same or different endpoints across species.  Some of these include differences in sensitivity of
18     endpoints, times of exposure, exposure routes, species and strains, use of multiple or single
19     doses, and variability between studies even for the same response. The estimated ED01s may be
20     influenced by experimental design, suggesting that caution should be used in comparing values
21      from different designs. In addition, caution should be used when comparing studies that
22     extrapolate ED01s outside the experimental  range. Furthermore, it may be difficult to compare
23     values across endpoints. For example, the human health risk for a 1% change of body weight
24     may not be equivalent to a 1% change in enzyme activity. Finally, background exposures are not
25     often considered in these calculations simply because they were not known. Nevertheless, given
26     these considerations, several general trends were observed and discussed in Part II, Chapter 8.
27     The lowest ED0]s tended to be for biochemical effects, followed by hepatic responses, immune
28     responses, and responses in tissue weight. An analysis of shape parameters implies that many
29     dose-response curves are consistent with linearity over the range of doses tested. This analysis
30    does not imply that the curves would be linear outside this range of doses, but it does inform the
31     choices for extrapolation. This is particularly true when body burdens or exposures at the lower
32    end of the observed range are close to body burdens or exposures of interest for humans, which is
33    the case with dioxin-like chemicals.
34            Overall, shape parameter data suggest that biochemical responses to TCDD are more
35     likely to be linear within the experimental dose range, while the more  complex responses are


      6/8/00                                      79       DRAFT—DO NOT CITE OR QUOTE

-------
  1      more likely to assume a nonlinear shape. However, a large number (> 40%) of the more complex
  2      responses have shape parameters that are more consistent with linearity than nonlinearity.
  3             The tissue weight changes seen for animals (using only data sets with good or moderate
  4      empirical fits to the model) yielded a median ED01 at average body burdens of 510 ng/kg in the
  5      multidose studies (range; 11 to 28000 ng/kg) and a median ED01 of 160 ng/kg (range 0.0001 to
  6      9,700 ng/kg) in the single dose studies. Toxicity endpoints from the single dose studies resulted
  7      in a median value at average body burdens of 4,300 ng/kg (range 1.3 to 1,000,000 ng/kg). For
  8      tissue weight changes, 43% of the dose-response curves exhibited linear response. In contrast, the
  9      toxicity endpoints from the single-dose studies exhibited predominantly nonlinear responses
 10      (80%). All multidose studies demonstrated a greater degree of linear response (41%) than did
 11      single-dose studies (37%), especially for tissue weight changes and toxicity endpoints (50%
 12      linear for multidose versus 34% for single dose). In general, it is not possible to dissociate the
 13      differences between cancer and noncancer dose-response as being due to differences in endpoint
 14      response or simply to differences in the length of dosing and exposure. Also, a greater
 15      percentage of the noncancer ED01s were extrapolations below the lower range of the data (42%)
 16      than was the case for the cancer endpoints (8% in animals and no extrapolations in humans).
 17
 18      5.3. MODE-OF-ACTION BASED DOSE-RESPONSE MODELING
 19             As described in Chapter 8, mechanism-based modeling can be a powerful tool for
 20      understanding and combining information on complex biological systems.  Use of a truly
 21      mechanism-based approach can, in theory, enable reliable and scientifically sound extrapolations
 22      to lower doses and between species. However, any  scientific uncertainty about the mechanisms
 23      that the models describe is inevitably reflected in uncertainty about the predictions of the models.
 24      The assumptions and uncertainties involved in the mechanistic modeling described in Chapter 8
 25      are discussed at length in that chapter and in cited publications.
 26             The development and continued refinement of PBPK models of the tissue dosimetry of
 27      dioxin have provided important information concerning the relationships between administered
 28      does and dose to tissue compartments (section 8.2).  Aspects of these models have been validated
 29      in the observable response range for multiple tissue compartments, species, and class of
 30      chemical. These models will continue to provide important new information for future revisions
 31      of this health assessment document.  Such information will likely include improved estimates of
 32      tissue dose for liver and other organs where toxicity has been observed, improved estimates of
 33      tissue dose(s) in humans, and improved estimates of tissue dose for dioxin related compounds.
 34             As a part of this reassessment, the development of biologically based dose-response
35      (pharmacodynamic) models for dioxin and related compounds has lead to considerable and
36      valuable insights regarding both mechanisms of dioxin action and dose-response relationships for

        6/8/00                                     80       DRAFT—DO NOT CITE OR QUOTE

-------
 1      dioxin effects. These efforts, described in some detail in Chapter 8, have provided additional
 2     perspectives on traditional methods such as the linearized multistage procedure for estimating
 3     cancer potency or the uncertainty factor approach for estimating levels below which noncancer
 4     effects are unlikely to occur.  These methods have also provided a biologically based rationale
 5     for what had been primarily statistical approaches. The development of models like those in
 6     Chapter 8 allows for an iterative process of data development, hypotheses testing and model
 7     development.
 8
 9     5.4.  SUMMARY DOSE-RESPONSE CHARACTERIZATION
10            All humans tested contain detectable body burdens of TCDD and  other dioxin-like
11     compounds that are likely to act through the same mode of action. It is possible that any
12     additional exposure above current background body burdens will be additive to ongoing
13     responses. The magnitude of the additional response will be a function of the toxicity
14     equivalence of the incremental exposure. This observation, the relatively small margin of
15     exposure for "key events," and the high percentage of observed linear responses suggest that a
16     proportional model should be used when extrapolating beyond the range of the experimental
17     data. Short of extrapolating to estimate risk in the face of uncertainties described above, a simple
18     margin-of-exposure approach may be useful to decision-makers when discussing risk
19     management goals. However, this decision would have to be based upon a policy choice because
20     this analysis does not strongly support either choice.
21            Because human data for cancer dose-response analysis were available and because of a
22     strong desire to stay within the range of responses estimated by these data, the risk chosen for
23     determining a point of departure was the 1% excess risk. Doses and exposures associated with
24     this risk (the ED0]s) were estimated from the available data using both mechanistic and empirical
25     models. Comparisons were made on the basis of body burdens to account for differences in
26     half-life across the numerous species studied.
27            In humans, restricting the analysis to log-linear models resulted in cancer ED0,s ranging
28     from 6 ng/kg to 161 ng/kg. This was similar to the estimates, from empirical modeling, from the
29     animal studies which ranged from 14 ng/kg to 1,190 ng/kg (most estimates were  in the range
30     from 14 to 500 ng/kg), and 2.7 ng/kg for the single mechanism-based model. Lower bounds on
31     these ED01 estimates were used to calculate upper bound slope factors and risk estimates for
32     average background body burdens. These estimates are presented above.  Upper bound slope
33     factors allow the calculation of the probability of cancer risk for the highly vulnerable in the
34     population (estimated to be the top 5% or greater). Even though there may be individuals in the
35     population who might experience a higher cancer risk on the basis of genetic factors or other
36     determinants of cancer risk not accounted for in epidemiologic data or animal studies, the vast

       6/8/00                                     81       DRAFT—DO NOT CITE OR QUOTE

-------
  1      majority of the population is expected to have less risk per unit of exposure and some may have
  2      zero risk. Based on these slope factor estimates (per pg TEQ/kgBW/day), average current
  3      background body burdens  (5 ng/kgBW) that result from average intakes of approximately 3
  4      pgTEQ/kgBW/day are in the range of 10~3 to 10~2.  A very small percentage of the population
  5      (< 1 %) may experience risks that are 2-3 times higher than this if they are among both the most
  6      vulnerable and the most highly exposed (among the top 5%) based on dietary intake of dioxin
  7      and related compounds.  This range of upper bound risk for the general population has increased
  8      an order of magnitude from the risk described at background exposure levels based on EPA's
  9      draft of this reassessment (10-4-10-3) (U.S. EPA, 1994).
10             Estimates for noncancer endpoints showed much greater variability, ranging over 10
11      orders of magnitude. In general, the noncancer endpoints displayed lower ED0]s for short-term
12      exposures versus longer term exposures, and for simple biochemical endpoints versus more
13      complex endpoints such as tissue weight changes or toxicity. In addition, the noncancer
14      endpoints generally displayed higher estimated ED01s than the cancer endpoints, with most
15      estimates ranging from 100 ng/kg to 100,000 ng/kg.  The mechanism-based models for
16      noncancer endpoints gave  a lower range of ED01s (0.17 to 105 ng/kg). Although most of these
17      estimates were based upon a single model the estimate from the hepatic zonal induction model
18      gave an ED0, for CYP1A2 induction of 51 ng/kg and hence was within the same range.
19             These estimates, although highly variable, suggest that any choice of body burden, as a
20      point of departure, above 100 ng/kg would likely yield >1% excess risk for some endpoint in
21      humans. Also, choosing of a point of departure below  1 ng/kg would likely be an extrapolation
22      below the range of these data and would likely represent a risk of <1%. Any choice in the middle
23      range of 1 ng/kg to 100 ng/kg would be supported by the  analyses, although the data provide the
24      greatest support in the range of 10 ng/kg to 50 ng/kg.
25
26
                                  6. RISK CHARACTERIZATION

  1             Characterizing risks from  dioxin and related compounds requires the integration of
 2      complex data sets and the use of science-based inferences regarding hazard, mode of action, dose
 3      response, and exposure.  It also requires consideration of incremental exposures in the context of
 4      an existing background exposure that is, for the most part, independent of local sources and
 5      dominated by exposure through the food supply. Finally,  this characterization must consider risks
 6      to special populations and developmental stages (subsistence fishers, children, etc.) as well as the
 7      general population.  It is important that this characterization convey the current understanding of
 8      the scientific community regarding these issues, highlight uncertainties in this understanding, and

        6/8/00                                     82        DRAFT—DO NOT CITE OR QUOTE

-------
 1      specify where assumptions or inferences have been used in the absence of data.  Although
 2      characterization of risk is inherently a scientific exercise, by its nature it must go beyond
 3      empirical observations and draw conclusions in untested areas.  In some cases, these conclusions
 4      are, in fact, untestable given the current capabilities in analytical chemistry, toxicology, and
 5      epidemiology.  This situation should not detract from our confidence in a well structured and
 6      documented characterization of risk, but should serve to confirm the importance of considering
 7      risk assessment as an iterative process that benefits from evolving methods and data collection.
 8
 9      Dioxin and related compounds can produce a wide variety of effects in animals and might
10      produce many of the same effects in humans.
11             There is adequate evidence based on all available information discussed in Parts I and II
12      of this reassessment, as well as that discussed in this Integrated Summary, to  support the
13      inference that humans are likely to respond with a broad spectrum of effects from exposure to
14     dioxin and related compounds.  These effects will likely range from biochemical changes at or
15      near background levels of exposure to adverse effects with increasing severity as body burdens
16     increase above background levels. Enzyme induction, changes in hormone levels, and indicators
17      of altered cellular function seen in humans and laboratory animals represent effects of unknown
18      clinical significance but that may be early indicators of toxic response. Induction of
19      activating/metabolizing enzymes at or near background levels, for instance, may be adaptive, and
20     in some cases,  beneficial, or may be considered adverse. Induction may lead  to more rapid
21      metabolism and elimination of potentially toxic compounds, or may lead to increases in reactive
22     intermediates and may potentiate toxic effects. Demonstration of examples of both of these
23      situations is available in the published literature and events of this type formed the basis for a
24     biologically based model discussed in Section 5.  Subtle effects, such as the impacts on
25      neurobehavioral outcomes, thyroid function, and liver enzymes (AST and ALT) seen in the
26      Dutch children exposed to background levels of dioxin and related compounds, or changes in
27      circulating reproductive hormones in men exposed to TCDD, illustrate the types of responses
28      that support the finding of arguably adverse effects at or near background body burdens. Clearly
29      adverse effects including, perhaps, cancer may not be detectable until exposures contribute to
30      body burdens that exceed background by one or two orders of magnitude (10 or 100 times). The
31      mechanistic relationships of biochemical and cellular changes seen at or near background body
32      burden levels to production of adverse effects detectible at higher levels remains uncertain, but
33      data are accumulating to suggest mode of action hypotheses for further testing.
34             It is well known that individual species vary in their sensitivity to any particular dioxin
35      effect.  However, the evidence available to date indicates that humans most likely fall in the
36      middle of the range of sensitivity for individual effects among animals rather than at either

       6/8/00                                      83       DRAFT—DO NOT CITE OR QUOTE

-------
  1      extreme. In other words, evaluation of the available data suggests that humans, in general, are
  2      neither extremely sensitive nor insensitive to the individual effects of dioxin-like compounds.
  3      Human data provide direct or indirect support for evaluation of likely effect levels for several of
  4      the endpoints discussed in the reassessment, although the influence of variability among humans
  5      remains difficult to assess.  Discussions have highlighted certain prominent, biologically
  6      significant effects of TCDD and related compounds. In TCDD-exposed men, subtle changes in
  7      biochemistry  and physiology such as enzyme induction, altered levels of circulating reproductive
  8      hormones, or reduced glucose tolerance and, perhaps, diabetes, have been detected in a limited
  9      number of epidemiologic studies. These findings, coupled with knowledge derived from animal
 10      experiments,  suggest the potential for adverse impacts on human metabolism, and developmental
 11      and/or reproductive biology, and, perhaps, other effects in the range of current human exposures.
 12      These biochemical, cellular, and organ-level endpoints have been shown to be affected by
 13      TCDD, but specific data on these endpoints do not generally exist for other congeners. Despite
 14      this lack of congener-specific data, there is reason to infer that these effects may occur for all
 15      dioxin-like compounds, based on the concept of toxicity equivalence.
 16             In this volume, dioxin and related compounds are characterized as carcinogenic,
 17      developmental, reproductive, immunological, and endocrinological hazards. The deduction that
 18      humans are likely to respond with noncancer effects from exposure to dioxin-like compounds is
 19      based on the fundamental level at that these compounds impact cellular regulation and the broad
 20      range of species that have proven to respond with adverse effects. For example, because
 21      developmental toxicity following exposure to TCDD-like congeners occurs in fish, birds, and
 22      mammals, it is likely to occur at some level in humans. It is not currently possible to state
 23      exactly how or at what levels individuals will respond with specific adverse impacts on
 24      development or reproductive function, but analysis of the Dutch cohort data and laboratory
 25      animal studies suggests that some effects may occur at or near background levels.  Fortunately,
 26      there have been few human cohorts identified with TCDD exposures high enough to raise body
 27      burdens significantly over background levels (see Table 5-1 and Figure 5-1 in Section 5), and
 28      when these cohorts have been examined, relatively few clinically  significant effects were
 29      detected. The lack of exposure gradients and adequate human information and the focus of most
 30      currently available epidemiologic studies on occupationally TCDD-exposed adult males makes
 31      evaluation of the inference  that noncancer effects associated with  exposure to dioxin-like
 32      compounds may be occurring, difficult. It is important to note, however, that when exposures to
 33      very high levels of dioxin-like compounds have been studied, such as in the Yusho and Yu-
 34      Cheng cohorts, a spectrum  of adverse effects have been detected in men, women, and children.
35      Some have argued that to deduce that a spectrum of noncancer effects will occur in humans in
36      the absence of better human data overstates the science; most scientists involved in the

        6/8/00                                     84       DRAFT—DO NOT CITE OR QUOTE

-------
 1      reassessment as authors and reviewers have indicated that such inference is reasonable given the
 2     weight-of-the-evidence from available data. As presented, this logical conclusion represents a
 3     testable hypothesis which may be evaluated by further data collection. EPA, its Federal
 4     colleagues, and others in the general scientific community are continuing to fill critical data gaps
 5     that will reduce our uncertainty regarding both hazard and risk characterization for dioxin and
 6     related compounds.
 7
 8     Dioxin and related compounds are structurally related and elicit their effects through a
 9     common mode of action.
10           The scientific community has identified and described a series of common biological
11     steps that are necessary for most, if not all, of the observed effects of dioxin and related
12     compounds in vertebrates including humans.  Binding of dioxin-like compounds to a cellular
13     protein called the AhR represents the first step in a series of events attributable to exposure to
14     dioxin-like compounds including biochemical, cellular, and tissue-level changes in normal
15     biological processes.  Binding to the AhR appears to be necessary for all well-studied effects of
16     dioxin but is not sufficient, in and of itself, to elicit these responses. There remains some
17     uncertainty as to whether every dioxin response is AhR-mediated. Sensitive biological tools such
18     as aryl hydrocarbon receptor deficient (AhR~;~) mice indicate a small residual of effects to
19     exposure to TCDD that does not allow us to rule out receptor-independent alternative pathways.
20     The well-documented effects elicited by exposure of animals and, in some cases, humans, to
21     2,3,7,8-TCDD are shared by other chemicals with similar structure and AhR binding
22     characteristics. In the past  5 years, significant data has accumulated that support the concept of
23     toxicity equivalence, that is at the heart of risk assessment for the complex mixtures of dioxin
24     and related compounds encountered in the environment. These data have been analyzed and
25     summarized in Part II, Chapter 9. This chapter has been added to EPA's dioxin reassessment to
26     address questions raised by the Agency's Science Advisory Board (SAB) in 1995. The SAB
27     suggested that, because the TEQ approach was a critical component of risk assessment for dioxin
28     and related compounds, the Agency should be explicit in its description of the history and
29     application of the process and go beyond reliance on the Agency's published reference
30     documents on the subject (U.S. EPA, 1987, 1989).
31
32     EPA and the international scientific community have adopted toxicty equivalence of dioxin
33     and related compounds as prudent science policy.
34           Dioxin and related compounds always exist in nature as complex mixtures. As discussed
35     in the Exposure Document, these complex mixtures can be characterized through analytic
36     methods to determine concentrations of individual congeners. Dioxin and related compounds

       6/8/00                                      85       DRAFT—DO NOT CITE OR QUOTE

-------
  1      can be quantified and biological activity of the mixture can be estimated using relative potency
  2      values and an assumption of dose additivity. Such an approach has evolved over time to form
  3      the basis for the use of TEQ in risk assessment for this group of compounds. Although such an
  4      approach is dependent on critical assumptions and scientific judgement, it has been characterized
  5      as a "useful, interim" way to deal with the complex mixture problem and has been accepted by
  6      numerous countries and several international organizations.  Alternative approaches, including
  7      the assumption that all congeners carry the toxicity equivalence of 2,3,7,8-TCDD, or that all
  8      congeners other than 2,3,7,8-TCDD can be ignored, have been generally rejected as inadequate
  9      for risk assessment purposes.
 10             Significant additional literature is now available on the subject of toxicity equivalence of
 11      dioxin and related compounds, and Chapter 9 provides the reader with a summary that is up to
 12      date through 1999.  A recent international evaluation of all of the available data (van den Berg et
 13      al., 1998) has reaffirmed the TEQ approach and has provided the scientific community with the
 14      latest values for TEFs for PCDDs, PCDFs, and dioxin-like PCBs. Consequently, we can infer
 15      with greater confidence that humans will respond to the cumulative exposure of AhR-mediated
 16      chemicals. The position taken in this reassessment is that these 1998 TEFs should be adopted for
 17      use by the Agency. Future  research will be needed to address remaining uncertainties inherent in
 18      the current approach.  The WHO has suggested that the TEQ scheme be reevaluated on a
 19      periodic basis and that TEFs and their application  to risk assessment be reanalyzed to account for
 20      emerging scientific information.
 21
 22      Complex mixtures of dioxin and related compounds are highly potent, "likely"
 23      carcinogens.
 24            With regard to carcinogenicity, a weight-of-the-evidence evaluation suggests that
 25      mixtures of dioxin and related compounds (CDDs, CDFs, and dioxin-like PCBs) are strong
 26      cancer promoters and weak direct or indirect initiators and likely to present a cancer hazard to
 27      humans. Because dioxin and related compounds always occur in the environment and in humans
 28      as complex mixtures of individual congeners, it is  appropriate that the characterization apply to
 29      the mixture. According to the Agency's revised draft Cancer Guidelines, the descriptor likely is
 30      appropriate when the available tumor effects and other key data are adequate to demonstrate
 31      carcinogenic potential to humans. Adequate data are recognized to span a wide range. The data
 32      for complex mixtures of dioxin and related compounds represents a case that, according to the
 33      draft Guidelines, would approach the strong-evidence end of the adequate-data spectrum.
 34      Epidemiologic observations of an association between exposure and cancer responses (TCDD);
 35      unequivocal positive responses in both sexes, multiple species, and different routes in lifetime
36      bioassays or initiation-promotion protocols or other shorter-term in vivo  systems such as
37      transgenic models (TCDD plus numerous PCDDs, PCDFs, dioxin-like PCBs); and mechanistic

        6/8/00                                     86       DRAFT—DO NOT CITE OR QUOTE

-------
 1     or mode-of action data that are assumed to be relevant to human carcinogenicity (PCDDs,
 2     PCDFs, dioxin-like PCBs) all support the description of complex mixtures of dioxin and related
 3     compounds as likely human carcinogens.
 4            Even though the database from cancer epidemiologic studies remains controversial, it is
 5     the view of this reassessment that this body of evidence is supported by the laboratory data
 6     indicating that TCDD probably increases cancer mortality of several types. Although not all
 7     confounders were ruled out in any one study, positive associations between surrogates of dioxin
 8     exposure, either length of occupational exposure or proximity to a known source combined with
 9     some information based on measured blood levels, and cancer have been reported.  These data
 10     suggest a role for dioxin exposure to contribute to a carcinogenic response but do not confirm a
 11     causal relationship between exposure to dioxin and increased cancer incidence.  Available human
 12     studies alone cannot demonstrate whether a cause-and-effect relationship between dioxin
 13     exposure and increased incidence of cancer exists. Therefore, evaluation of cancer hazard in
 14     humans must include an evaluation of all of the available animal and in vitro data as well as the
 15     data from exposed human populations.
 16            As discussed earlier in Section 2.2.1.4, under EPA's current approach, individual
 17     congeners can also be characterized as to their carcinogenic hazard. TCDD is best characterized
 18     as "carcinogenic to humans."  This means that, based on the weight of all of the evidence
 19     (human, animal, mode of action), TCDD meets the criteria that allows U.S. EPA and the
 20     scientific community to accept a  causal relationship between TCDD exposure and cancer hazard.
 21     The guidance suggests that "carcinogenic to humans" is an appropriate descriptor of human
 22     carcinogenic potential when there is an absence of conclusive epidemiologic evidence to clearly
 23     establish a cause-and-effect relationship between human exposure and cancer, but there is
 24     compelling carcinogenicity in animals and mechanistic information in animals and humans
 25     demonstrating similar modes of carcinogenic action. The "carcinogenic to humans" descriptor is
 26     suggested for TCDD because all of the following conditions are met:
 27         •  There is evidence from occupational epidemiologic studies for an association between
 28            TCDD exposure and increases in cancer at all sites, in lung cancer and, perhaps, at other
 29            sites, but the data are insufficient on their own to demonstrate a causal association.
 30         •  There is extensive carcinogenicity in both sexes of multiple species at multiple sites.
 31         •  There is general agreement that the mode of TCDD' s carcinogenicity is AhR dependent
 32            and proceeds through modification of the action  of a number of receptor and hormone
 33            systems involved in cell growth and differentiation, such as the epidermal growth factor
 34            receptor and estrogen receptor.
 35         •   Key events such as equivalent body burdens in animals and in human populations
 36            expressing an association  between exposure to TCDD and cancer, and the determination
37            of active AhR and dioxin  responsive elements in the general human population. There is

       6/8/00                                     87       DRAFT—DO NOT CITE OR QUOTE

-------
  1             no reason to believe that these events would not occur in the occupational cohorts
  2             studied.
  3             Other individual dioxin-like compounds are characterized as "likely" human carcinogens
  4      primarily because of the lack of epidemiological evidence associated with their carcinogenicity,
  5      although the inference based on toxiciry equivalence is strong that they would behave in humans
  6      as TCDD does. Other factors, such as the lack of congener-specific chronic bioassays, also
  7      support this characterization.  For each congener, the degree of certainty is dependent on the
  8      available congener-specific data and their consistency with the generalized mode of action that
  9      underpins toxiciry equivalence for TCDD and related compounds. On the basis of this logic,
 10      complex environmental mixtures of TCDD and dioxin-like compounds should be characterized
 11      as "likely" carcinogens, with the degree of certainty of the characterization being dependent on
 12      the constituents of the mixture, when known. For instance, the hazard potential, although
 13      "likely," would be characterized differently for a mixture whose TEQ was dominated by OCDD
 14      as compared with one which was dominated by pentaCDF.
 15             Although uncertainties remain regarding quantitative estimates of upper bound cancer
 16      risk from dioxin and related compounds, efforts of this reassessment to bring more data into the
 17      evaluation of cancer potency have resulted in evaluation of the slope of the dose-response curve
 18      at the low end of the observed range (using the LED01) using a simple proportional (linear) model
 19      and a calculation of both upper bound risk and margin  of exposure (MOE) based on human
 20      equivalent background exposures and associated body  burdens.  Evaluation of shape parameters
 21      (used to estimate degree of linearity or nonlinearity of dose-response within the range of
 22      observation) for biochemical effects indicates that many of these biochemical effects can be
 23      hypothesized to be to key events in a generalized dioxin mode-of action model. These analyses
 24      do not argue for significant departures from linearity below a calculated ED0, for endpoints
 25      potentially related to cancer response, for at least one to two orders of magnitude lower exposure.
 26             Risk estimates for intakes associated with background body burdens or incremental
 27      exposures based on this slope factor represent a plausible upper bound on risk based on the
 28      evaluation of animal and human data. The slope factors based on the most sensitive  cancer
 29      responses, both animal and human, calculated in Section 5 fall in a range of 5 x 10~3 to 5 * 10~4
 30      per pg TEQ/kgBW/day. The ranges of estimates of upper bound cancer potency calculated from
 31      the human and animal data analyzed in Part II, Chapter 8, overlap. The range above is bounded
 32      on the upper end by the estimate of slope from the Hamburg cohort epidemiology study and on
 33      the lower end by the estimate from the reanalyzed Kociba study. Consequently, the Agency,
 34      although fully recognizing this range and the public health conservative nature of the slope
 35      factors that make up the range, suggests the use of 5 x  10"3 per pg TEQ/kgBW/day as an
 36      estimator of upper bound cancer risk for both background intakes and incremental intakes above
3 7      background.  Slope factors allow the calculation of the probability of cancer risk for the highly
        6/8/00                                     88       DRAFT—DO NOT CITE OR QUOTE

-------
 1     vulnerable in the population (estimated to be the top 5% or greater).  Although there may be
 2     individuals in the population who might experience a higher cancer risk on the basis of genetic
 3     factors or other determinants of cancer risk not accounted for in epidemiologic data or animal
 4     studies, the vast majority of the population is expected to have less risk per unit of exposure and
 5     some may have zero risk.  Based on these slope factor estimates (per pg TEQ/kgBW/day),
 6     average current background body burdens (5 ng/kgBW) resulting from average intakes of
 7     approximately 3  pgTEQ/kgBW/day are in the range of 10'3 to 10"2. A very small percentage of
 8     the population (< 1%) may experience risk that are 2-3 times higher than this if they are among
 9     both the most vulnerable and the most highly exposed (among the top 5%) based on dietary
10     intake of dioxin and related compounds. This range of upper bound risk for the general
11     population has increased an order of magnitude from the risk described at background exposure
12     levels based on EPA's draft of this reassessment (10-4-10-3) (U.S. EPA, 1994).
13            Despite the use of the epidemiology data to describe an upper bound on cancer risk, the
14     Peer Panel that met in September 1993 to review an earlier draft of the cancer epidemiology
15     chapter suggested that the epidemiology data alone were still not adequate to implicate dioxin
16     and related compounds as "known" human carcinogens, but that the results from the human
17     studies were largely consistent with observations from laboratory studies of dioxin-induced
18     cancer and, therefore, should not be dismissed or ignored.  Other scientists, including those who
19     attended the Peer Panel meeting, felt either more or less strongly about the weight of the
20     evidence from cancer epidemiology studies, representing the range of opinion that still exists on
21     the interpretation of these studies. Similar opinions were expressed in the comments documented
22     in the SAB's report in 1995 (U.S. EPA,1995). More recently, the International Agency for
23     Research on Cancer (1997), in its reevaluation of the cancer hazard of dioxin and related
24     compounds, found that whereas the epidemiologic database for 2,3,7,8-TCDD was still
25     "limited," the overall weight of the evidence was sufficient to characterize 2,3,7,8-TCDD as a
26     Category 1 "known" human carcinogen. Other related members of the class of dioxin-like
27     compounds were considered to have "inadequate" epidemiologic data to factor into hazard
28     categorization. A similar classification has been proposed  within the context of the Department
29     of Health and Human Services' Report on Carcinogens (NTP, 2000). They too base their
30     characterization on the broad base of human, animal, and mode-of-action information in humans
31     and animals that supports this conclusion. Therefore, given that 2,3,7,8-TCDD is contained in
32     complex mixtures of dioxin and related compounds, and that the TEQ approach has been adopted
33     as a reasonable approach to assessing risks of these complex mixtures, it is also reasonable to
34     apply estimates of upper bound cancer potency derived from epidemiology studies where 2,3,7,8-
35     TCDD was associated with excess cancer risk to complex mixtures of dioxin and related
36     compounds.

       6/8/00                                     89        DRAFT—DO NOT CITE OR QUOTE

-------
  1             The current evidence suggests that both receptor binding and most early biochemical
  2      events such as enzyme induction are likely to demonstrate low-dose linearity. The mechanistic
  3      relationship of these early events to the complex process of carcinogenesis remains to be
  4      established.  If these findings imply low-dose linearity in biologically based cancer models under
  5      development, then the probability of cancer risk will be linearly related to exposure to TCDD at
  6      low doses. Until the mechanistic relationship between early cellular responses and the
  7      parameters in biologically based cancer models is better understood, the shape of the dose-
  8      response curve for cancer in the below the range of observation can only be inferred with
  9      uncertainty.  Associations between exposure to dioxin and certain types  of cancer have been
 10      noted in occupational cohorts with average body burdens of TCDD approximately 1- 3 orders of
 11      magnitude (10-1,000 times) higher than average TCDD body burdens in the general population.
 12      The average body burden in these occupational cohorts level is within 1-2 orders of magnitude
 13      (10-100 times) of average background body burdens in the general population in terms of TEQ
 14      (see Table 5-1 and Figure 5-1).  Thus, there is no need for large-scale low-dose extrapolations in
 15      order to evaluate background intakes and body burdens, and little if any  data to suggest large
 16      departures from linearity in this somewhat narrow window between the lower end of the range of
 17      observation and the range of general-population background exposures. Nonetheless, the
 18      relationship of apparent increases in cancer mortality in these worker populations to calculations
 19      of general population risk remains a source of uncertainty.
 20             TCDD has been clearly shown to increase malignant tumor incidence in laboratory
 21      animals.  In addition, a number of studies analyzed in this reassessment demonstrate other
 22      biological effects of dioxin related to the process of carcinogenesis. Initial attempts to construct
 23      a biologically based model for certain dioxin effects as described in this  reassessment will need
 24      to be continued and expanded to accommodate more of the available biology and to apply to a
 25      broader range of potential health effects associated with exposure to dioxin-like compounds.
 26
 27      Use a "margin-of-exposure approach" to evaluate risk for noncancer and cancer endpoints.
 28             The likelihood that noncancer effects may be occurring in the human population at
 29      environmental exposure levels is often evaluated using a MOE approach. The Agency has used
 30      this approach for a number of years in its assessment of the safety of pesticides. This concept has
 31      also been incorporated into the revised Cancer Risk Assessment guidelines. A MOE is
 32      calculated by dividing a "point of departure" for extrapolation purposes at the low end of the
 33      range of observation in human or animal studies (the human-equivalent animal LOAEL,
 34      NOAEL, BMD, or effective dose [EDxx]) by the human exposure or body burden level of
 35      interest. Generally speaking, when considering either background exposures or incremental
 36      exposures plus background, MOEs in range of 100-1,000 are considered adequate to rule out the
37      likelihood of significant effects occurring in humans based on sensitive animal responses or

        6/8/00                                     90       DRAFT—DO NOT CITE OR QUOTE

-------
 1     results from epidemiologic studies. The adequacy of the MOE to be protective of health must
 2     take into account the nature of the effect at the "point of departure," the slope of the dose-
 3     response curve, the adequacy of the overall database, interindividual variability in the human
 4     population, and other factors. Considering MOEs based on incremental exposures alone divided
 5     by the human exposure of interest, is not considered to give an accurate portrayal of the
 6     implications of that exposure unless background exposures are insignificant.
 7            One of the difficulties in assessing the potential health risk of dioxins is that background
 8     exposures not be insignificant when based on total TEQ.  The average levels of background
 9     intake and associated body burdens of dioxin-like compounds in terms of TEQs in the general
10     population would be well within a factor of 100 of human-equivalent exposure levels associated
11     with NOELS, LOAELs, BMDs, or EDOI values in laboratory animals exposed to TCDD or
12     TCDD equivalents. In many cases, the MOE compared to background using these endpoints is a
13     factor of 10 or less (see Tables 2-2 and 2-3).  These estimates, although variable, suggest that
14     any choice of body burden, as a point of departure, above 100 ng/kg would likely yield >1%
15     excess risk for some endpoint in humans (see Section II, Chapter 8). Also, choosing of a point of
16     departure below 1 ng/kg would likely be an extrapolation below the range  of these data and
17     would likely represent a risk of < 1 %. Any choice for a point of departure in the middle range of
18     1 ng/kg to 100 ng/kg would be supported by the analyses, although the data provide the greatest
19     support for a point of departure in the range of 10 ng/kg to 50 ng/kg.
20            Because of the relatively high background compared to effect levels, the Agency is not
21     recommending the derivation of an RfD for dioxin and related compounds. Although RfDs are
22     often useful because they represent a health risk goal below which there is likely to be no
23     appreciable risk of noncancer effects over a lifetime of exposure, their primary use is to evaluate
24     increments of exposure from specific sources when background exposures are low and
25     insignificant. Any RfD that the Agency would recommend under the traditional approach for
26     setting an RfD is likely to be 2-3 orders of magnitude (100-1,000) below current background
27     intakes and body burdens. Because exceeding the RfD is not a statement of risk, discussion  of an
28     RfD for an incremental exposure when the RfD has already been exceeded by average
29     background exposures is meaningless.
30            When evaluating incremental exposures associated with specific sources, knowing the
31     increment relative to background  may help to understand the impact of the incremental exposure.
32     For instance, it would be misleading to suggest that an incremental exposure of
33     0.001 pg TEQ/kg/day was below the RfD if "background" exposures were already at or above
34     that level. On the other hand,  as part of the total, the increment represents less than a 0.1%
35     increase over average "background," and we estimate that individuals within the 50%-95% range
36     of exposure within the population may be 2-3 times (200%-300%) higher. This has led us to
37     suggest that perhaps the best information for a decision-maker to have is: (1) a characterization
       6/8/00                                     91        DRAFT—DO NOT CITE OR QUOTE

-------
  1      of average "background" exposures; (2) a characterization of the percent increase over
  2      background of individuals or subpopulations of interest; and (3) a policy statement about when
  3      increases over average "background" become significant for the decision.  This is not easy
  4      because one could argue that, given high "background," any addition, if it is widespread, is too
  5      much.  On the other hand, someone else could argue that a 10% increase in incremental exposure
  6      for a small population around a specific point source would be well within the general population
  7      exposures and would not constitute a disproportionate exposure or risk. In this case, the strategy
  8      might be to bring average "background" exposures down and to focus on large incremental
  9      exposures or highly susceptible populations. This would be a strategy that would parallel the
 10      Agency's lead strategy. Other parallel issues between dioxin-like compounds and lead are under
 11      discussion within the Agency.
 12             ATSDR (1999) set a minimal risk level (MRL), which is defined similarly to the EPA's
 13      RfD, for dioxin and related compounds of 1.0 pg TEQ/kgBW/day. Some of the data regarding
 14      lower bounds on the ED01s from various noncancer effects call that MRL into question.  WHO
 15      (2000) has set a tolerable daily intake of 1-4 pg TEQ/kgBW/day and has indicated that, although
 16      current exposures in that range are "tolerable"  (a risk management decision rather than a risk
 17      assessment), efforts should be made to ultimately reduce intake levels. Findings in this
 18      reassessment appear to be supportive of that recommendation.
 19
 20      Children's risk from exposure to dioxin and related  compounds may be increased, but
 21      more data are needed to fully address this issue.
 22             The issue of children's risk from exposure to dioxin-like compounds has been addressed
 23      in a number of sections throughout this reassessment. Data suggest a sensitivity of response in
 24      both humans and animals during the developmental period, both prenatally and postnatally.
 25      However, data are limited.  Because evaluation of the impacts of early exposures on both
 26      children's health and health later in life is important to a complete characterization of risk,
 27      collection of additional data in this area should be a high priority to reduce uncertainties in future
 28      risk assessments.
 29             Data from the Dutch cohort of children exposed to PCBs and dioxin-like compounds
 30      suggest impacts of exposure to background levels of dioxin and related compounds prenatally
 31      and, perhaps, postnatally on neurobehavioral outcomes, thyroid function, and liver enzymes
 32      (AST and ALT). Although these effects cannot be attributed solely to dioxin and related
 33      compounds, several associations suggest that these are,  in fact, likely to be Ah-mediated effects.
 34      An investigation of background dioxin exposure and tooth development was done in Finnish
 35      children as a result of studies of dental effects in dioxin-exposed rats, mice, and nonhuman
36      primates, and  in PCB-exposed children.  The Finnish investigators examined enamel
37      hypomineralization of permanent first molars in 6-7 year old children. The length of time that

        6/8/00                                     92        DRAFT—DO NOT CITE OR QUOTE

-------
 1     infants breast fed was not significantly associated with either mineralization changes or with
 2     TEQ levels in the breast milk. However, when the levels and length of breast feeding were
 3     combined in an overall score, a statistically significant association was observed (r = 0.3, p =
 4     0.003, regression analysis).
 5            In addition, effects have been seen where significantly elevated exposure occurred. The
 6     incidents at Yusho and Yu-Cheng resulted in increased perinatal mortality and low birthweight in
 7     infants born to women who had been exposed. Rocker bottom heal was observed in Yusho
 8     infants, and functional abnormalities have been reported in Yu-Cheng children. The similarity of
 9     effects observed in human infants prenatally exposed to the complex mixture in Yusho and
10     Yu-Cheng with those reported in adult monkeys exposed only to TCDD suggests that at least
11     some of the effects on children are due to the TCDD-like congeners in the contaminated rice oil
12     ingested by the mothers of these children. The similar responses include a clustering of effects in
13     organs derived from the ectodermal germ layer, referred to as ectodermal dysplasia, including
14     effects on the skin, nails, and Meibomian glands; and developmental and psychomotor delay
15     during developmental and cognitive tests. Some investigators believe that because all of these
16     effects in the Yusho and Yu-Cheng cohorts do not correlate with TEQ, some of the effects are
17     exclusively due to nondioxin-like PCBs or a combination of all the congeners.  In addition, on
18     the basis of these data,  it is still not clear to what extent there is an association between overt
19     maternal toxicity and embryo/fetal toxicity in humans. Further studies in the offspring as well as
20     follow-up to the Seveso incident may shed further light on this issue.  In addition to chloracne
21     and acute responses to TCDD exposure seen in Seveso children, elevated levels of serum GOT
22     have been observed within a year after exposure in some of the more highly exposed Seveso
23     children. Long-term pathologic consequences of elevated GOT have not been illustrated by
24     excess mortality from liver disorders or cancer or in excess morbidity, but further follow-up is
25     needed. It must be recognized that the absence of an effect thus far does not obviate the
26     possibility that the enzyme levels may have increased concurrent to the exposure but declined
27     after cessation. The apparently transient elevations in ALT levels among the Seveso children
28     suggest that hepatic enzyme levels other than GGT may react in this manner to  2,3,7,8-TCDD
29     exposure.
30            Impacts on thyroid hormones provide an example of an effect of elevated postnatal
31     exposure to dioxin and related compounds. Several studies of nursing infants suggest that
32     ingestion of breast milk with a higher dioxin TEQ may alter thyroid function. Thyroid hormones
33     play important roles in the developing nervous system of all vertebrate species, including
34     humans.  In fact, thyroid hormones are considered so important in development that in the United
35     States all infants are tested for hypothyroidism shortly after birth.  Results from the studies
36     mentioned above suggest a possible shift in the population distribution of thyroid hormone
37     levels, particularly T4,  and point out the need for collection of longitudinal data to assess the

       6/8/00                                     93       DRAFT—DO NOT  CITE OR QUOTE

-------
  1      potential for long-term effects associated with developmental exposures. The exact processes
  2     accounting for these observations in humans are unknown, but when put in perspective of animal
  3     responses, the following might apply: dioxin increases the metabolism and excretion of thyroid
  4     hormone, mainly T4, in the liver. Reduced T4 levels stimulate the pituitary to secrete more TSH,
  5     which enhances thyroid hormone production. Early in the disruption process, the body can
  6     overcompensate for the loss of T4, which may result in a small excess of circulating T4 in
  7     response to the increased TSH. In animals, given higher doses of dioxin, the body is unable to
  8     maintain homeostasis, and TSH levels remain elevated and T4 levels decrease.
  9            A. large number of studies in animals have addressed the question of effects of dioxin-like
 10     chemicals after in utero or lactational exposure. These have included both single-congener
 11      studies and exposures to complex mixtures. However, the vast majority of the data are derived
 12     from studies of 2,3,7,8-TCDD, or single congeners (e.g., PCB 77) or commercial mixtures of
 13     PCBs. Exposure patterns have included single doses to the dams as well as dosing on multiple
 14     days during gestation beginning as early as the first day of gestation. These studies are discussed
 15     in detail in Part II, Chapter 5.  The observed toxic effects include developmental toxicity,
 16     neurobehavioral and neurochemical alterations, endocrine effects, and developmental
 17     immunotoxicity. For instance, results of this body of work suggest that 2,3,7,8-TCDD clearly
 18     has the potential to produce alterations in male reproductive  function (rats and hamsters) and
 19     male sexual behavior (rats) after prenatal exposure. In addition, impacts on neuromotor and
 20     cognitive behavior as well as development of the immune system have been indicated in a
 21      number of studies.
 22            No epidemiological data and limited animal data are  available to address the question of
 23     the potential impact of exposure to dioxin-like compounds on childhood cancers or on cancers of
 24     later life.  Given the relative impact of nursing on body burdens (see the discussion of breast milk
 25     exposures and body burdens below), direct impacts of increased early postnatal exposure on the
 26     carcinogenic process are expected to be small.  This conclusion is based on the reasonable
 27     assumptions that cancer risk is a function of average lifetime body burden or that, because dioxin
 28     is a potent cancer promoter rather than a direct initiator of the cancer process, exposures later in
 29      life might be more important than those received earlier.  However, recent studies of Brown et al.
 30      (1998) suggest that prenatal exposure of rats to dioxin and related compounds may indirectly
 31      enhance their sensitivity  as adults to chemical carcinogenesis from other chemical carcinogens.
 32      Further work is needed to evaluate this issue.
 33            In addition to potential vulnerability during development, fetuses, infants, and children
 34      are exposed to dioxins through several routes. The fetus is exposed in utero to levels of dioxin
 35      and related compounds that reflect the body burden of the mother. It is  important to recognize
36      that it is not the individual meals a pregnant woman eats during pregnancy that might affect
37      development, but the consequence of her exposure history over her life, which has the greatest

        6/8/00                                     94       DRAFT—DO NOT CITE OR QUOTE

-------
 1     impact on her body burden. Again, good nutrition, including a diet with appropriate levels of fat,
 2     has consequences on dietary intake and consequent body burdens of dioxin and related
 3     compounds.  Nursing infants represent special cases who, for a limited portion of their lives, may
 4     have elevated exposures on a body-weight basis when compared with non-nursing infants and
 5     adults (see discussion). In addition to breast milk exposures, intakes of CDD/CDFs and dioxin-
 6     like PCBs are more than three times higher for a young child than those of an adult, on a body-
 7     weight basis. Table 4-9 in Section 4 of this document describes the variability in average intake
 8     values as a function of age using age-specific food consumption rates and average food
 9     concentrations, as was done for adult intake estimates. However, as with for the nursing infants,
10     the differences in body burden between children and adults are expected to be much less than the
11     differences in daily intake.  Assuming that body burden is the relevant dose metric for most if not
12     all effects, there is some assurance that these increased intake levels will have limited additional
13     impact on risk as compared with overall lifetime exposure.
14
15     Background exposures to dioxin and related compounds need to be considered when
16     evaluating both hazard and risk.
17            The term "background" exposure has been used throughout this reassessment to describe
18     exposure of the general population, who are not exposed to readily identifiable point sources of
19     dioxin-like compounds. Adult daily intakes of CDD/CDFs and dioxin-like PCBs are estimated
20     to average 45 and 25 pg TEQDFP-WHO98/day, respectively, for a total intake of 70 pg/day
21     TEQDFP-WHO9g. Daily intake is estimated by combining exposure media concentrations (food,
22     soil, air) with contact rates (ingestion, inhalation). Table 4-8 summarizes the intake rates derived
23     by this method. The intake estimate is supported by an extensive database on food consumption
24     rates and food data. PK modeling provides further support for the intake estimates.  Current
25     adult tissue levels reflect intakes from past exposure levels, which are thought to be higher than
26     current levels (see Trends, Section 2.6).
27            CDD/CDF and dioxin-like PCB intakes for the general population may extend to levels at
28     least three times higher than the mean. Variability in general-population exposure is primarily  a
29     result of differences in dietary choices that individuals make. These are differences in both
30     quantity and types of food consumed. A diet that is disproportionately high in animal fats will
31     result in an increased background exposure over the mean. Data on variability of fat
32     consumption indicate that the 95th  percentile is about twice the mean and the  99th percentile is
33     approximately three times the mean.  Additionally, a diet that substitutes meat sources that are
34     low in dioxin (i.e., beef, pork, or poultry) with sources that are high in dioxin (i.e., freshwater
35     fish) could result in exposures elevated more than three times the mean.  This scenario may not
36     represent a significant change in total animal fat consumption, even though it results in an
37     increased dioxin exposure.  Intakes of CDD/Fs and dioxin-like PCBs are over three times higher

       6/8/00                                      95        DRAFT—DO NOT CITE OR QUOTE

-------
  1      for a young child as compared to that of an adult, on a body weight basis. Using age-specific
  2      food consumption rate and average food concentrations, as was done above for adult intake
  3      estimates, Table 4-9 describes the variability in average intake values as a function of age.
  4             The average CDD/CDF tissue level for the general adult United States population appears
  5      to be declining; the best estimate of current (late 1990s) levels is 25 ppt (TEQDFP-WHO9g, lipid
  6      basis). The tissue samples collected in North America in the late 1980s and early 1990s showed
  7      an average TEQDFP-WHO9g level of about 55 pg/g lipid. This finding is supported by a number of
  8      studies, all conducted in North America, that measured dioxin levels in adipose tissue, blood, and
  9      human milk. The number of people in most of these studies, however, is relatively small and the
 10      participants were not statistically selected in ways that assured their representativeness of the
 11      general United States adult population.  One study, the 1987 National Human Adipose Tissue
 12      Survey (NHATS), involved more than 800 individuals and provided broad geographic coverage,
 13      but did not address coplanar PCBs.  Similar tissue levels of these compounds have been
 14      measured in Europe and Japan during similar time periods.
 15             Because dioxin levels in the environment have been declining since the 1970s (see trends
 16      discussion), it is reasonable to expect that levels in food, human intake, and ultimately human
 17      tissue have also declined over this period. The changes in tissue levels are likely to lag the
 18      decline seen in environmental levels, and the changes in tissue levels cannot be assumed to occur
 19      proportionally with declines in environmental levels.  CDC (2000) summarized levels of CDDs,
 20      CDFs, and PCBs in human blood collected during the time period 1995 to 1997.  The individuals
 21      sampled were all U.S. residents with no known exposures to dioxin other than normal
 22      background. The blood was collected in seven different locations from 400 individuals with an
 23      age range of 20 to 70 years.  All TEQ calculations were made assuming nondetects were equal to
 24      half the detection limit. Although these samples were not collected in a manner that can be
 25      considered statistically representative of the national population and lack wide geographic
 26      coverage, they are judged to provide a better indication of current tissue levels in the United
 27      States than the earlier data (see Table 4-7). PCBs 105, 118, and 156 are missing from the blood
 28      data for the comparison populations reported in the Calcasieu study (CDC, 2000). These
 29      congeners account for 62% of the total PCB TEQ estimated in the early 1990s. Assuming that
 30      the missing congeners from the Calcasieu study data contribute the same proportion to the total
 31      PCB TEQ as in earlier data, they would increase our estimate of current body burdens by another
 32      3.7 pg TEQ/g lipid for a total PCB TEQ of 5.9 pg/g lipid and a total DFP TEQ of 25 pg/g lipid.
 33             Past background exposure of about 3 pg TEQ/ kgBW/day leads to body burdens in the
 34      human population that currently average approximately 5 ng/kg (20-30 pg TEQ/g lipid) when all
 35      dioxins, furans  and PCBs are included; body burdens have been higher in the past.  DeVito et al.
 36      (1995) estimated  that body burdens averaged 9-13 ng/kg based on intake values of 4-6 pg
37      TEQ/kg/day and blood levels of 40-60 pgTEQ/g lipid using data from the late 1980s. If the

        6/8/00                                     96       DRAFT—DO NOT CITE OR QUOTE

-------
 1     general population were exposed to dioxins and related compounds at the current level of intake
 2     (approximately 1 pg TEQ/kg/day) for a lifetime, average steady-state body burdens would be <2
 3     ng/kg and blood levels would be 7-8 pg TEQ/g lipid. These estimates are based on the
 4     assumption of 50% absorption of dioxin-like compounds from the diet. Using the same
 5     assumption used for intake values, high-end estimates of body burden of individuals in the
 6     general population (approximately the top 5%) may be more than twice as high as these average
 7     estimates. This calculation is based on data for dietary fat consumption and the assumption that
 8     body burdens of dioxin and related compounds in the general population are associated with fat
 9     consumption.  The top 1 % is likely to be three times higher based on its intake of fat.
10            Characterizing national background levels of dioxins in tissues is uncertain because the
11     current data cannot be considered statistically representative of the general population. The task
12     is also complicated by the fact that tissue levels are a function of both age and birth year.
13     Because intake levels have varied over time, the accumulation of dioxins in a person who turned
14     50 in 1990 is different from that in a person who turned 50 in 2000. Future studies should help
15     address these uncertainties. The National Health and Nutrition Examination Survey (NHANES)
16     began a new national survey  in 1999 that will measure dioxin blood levels in about 1,700 people
17     per year (see http:www.cdc.gov/nchs/nhanes.htm).  The survey is  conducted at 15 different
18     locations per year and is designed to select individuals statistically representative of the civilian
19     U.S. population in terms of age, race, and ethnicity.  These new data should provide a much
20     better basis than the currently available data for estimating national background tissue levels and
21     evaluating trends.
22            As described above, current intake levels from food sources are estimated in this
23     reassessment to be approximately  1 pg TEQ/KgBW/day. Certain  segments of the population
24     may be exposed to additional increments of exposure by being in proximity to point sources or
25     because of dietary practices.  These will be described below.
26
27     Evaluation of exposure of "special" populations and developmental stages is critical to risk
28     characterization.
29            As discussed above, background exposures to dioxin-like compounds may extend to
30     levels at least three times higher than the mean. This upper range is assumed to result from the
31     normal variability of diet and human behaviors. Exposures from local elevated sources or unique
32     diets would be in addition to this background variability. Such elevated exposures may occur in
33     small segments of the population, such as individuals living near discrete local sources, or
34     subsistence or recreational fishers. Nursing infants represent a special case where, for a limited
35     portion of their lives, these individuals may have elevated exposures on a body-weight basis
36     when compared to non-nursing infants and adults. This exposure will be discussed in a separate
37     section.
       6/8/00                                     97        DRAFT—DO NOT CITE OR QUOTE

-------
  1             Dioxin contamination incidents involving the commercial food supply have occurred in
  2      the United States and other countries.  For example, in the United States, contaminated ball clay
  3      was used as an anticaking agent in soybean meal and resulted in elevated dioxin levels in some
  4      poultry and catfish.  This incident involved less than 5% of national poultry production and has
  5      since been eliminated. Elevated dioxin levels have also been observed in a few beef and dairy
  6      animals where the contamination was associated with contact with pentachlorophenol-treated
  7      wood.  This kind of elevated exposure was not detected in the national beef survey.
  8      Consequently, its occurrence is likely to be low, but it has not been determined. These incidents
  9      may have led to small increases in dioxin exposure to the general population. However, it is
 10      unlikely that such incidents have led to disproportionate exposures to populations living near
 11      where these incidents have occurred, because in the United States meat and dairy products are
 12      highly distributed on a national scale.  If contamination events were to occur in foods that are
 13      predominantly distributed on a local or regional scale, then such events could lead to highly
 14      exposed local populations.
 15             Elevated exposures associated with the workplace or industrial accidents have also been
 16      documented.  United States workers in certain segments of the chemical industry had elevated
 17      levels of TCDD exposure, with some tissue measurements in the thousands of ppt TCDD. There
 18      is no clear evidence that elevated exposures are currently occurring among United States
 19      workers. Documented examples of past exposures for other groups include certain Air Force
 20      personnel exposed to Agent Orange during the Vietnam War and people exposed as a result of
 21      industrial accidents in Europe and Asia.
 22             Consumption of unusually high amounts of fish, meat, or dairy products containing
 23      elevated levels of dioxins and dioxin-like  PCBs can lead to elevated exposures in comparison to
 24      the general population. Most people eat some fish from multiple sources, both fresh and salt
 25      water.  The typical dioxin concentrations in these fish and the typical rates of consumption are
 26      included in the mean background calculation of exposure.  People who consume large quantities
 27      of fish at typical contamination levels may have elevated exposures because the concentration of
 28      dioxin-like compounds in fish  is generally higher than in other animal food products. These
 29      kinds of exposures are addressed within the estimates of variability of background and are not
 30      considered to result in highly exposed populations. If high-end consumers obtain their fish from
 31      areas where the concentration of dioxin-like chemicals is elevated, they may constitute a highly
 32      exposed subpopulation. Although this scenario seems reasonable, no supporting data could be
33      found for such a highly exposed subpopulation in the United States. One study measuring
34      dioxin-like compounds in blood of sports  fishers in the Great Lakes area showed elevations over
35      mean background, but within the range of normal variability.  Elevated CDD/CDF  levels in
36      human blood have been measured in Baltic fishermen.  Similarly, elevated levels of coplanar


        6/8/00                                     98        DRAFT—DO NOT CITE OR QUOTE

-------
 1     PCBs have been measured in the blood of fishers on the north shore of the Gulf of the St.
 2     Lawrence River who consume large amounts of seafood.
 3            High exposures to dioxin-like chemicals as a result of consuming meat and dairy products
 4     would occur only in situations where individuals consume large quantities of these foods and the
 5     level of these compounds is elevated. Most people eat meat and dairy products from multiple
 6     sources and, even if large quantities are consumed, they are not likely to have unusually high
 7     exposures. Individuals who raise their own livestock for basic subsistence have the potential for
 8     higher exposures if local levels of dioxin-like compounds are high. One study in the United
 9     States showed elevated levels in chicken eggs near a contaminated soil site. European studies at
10     several sites have shown elevated CDD/CDF levels in milk and other animal products near
11     combustion sources.
12            In summary, in addition to general population exposure, some individuals or groups of
13     individuals may also be exposed to dioxin-like compounds from discrete sources or pathways
14     locally within their environment. Examples of these "special" exposures include contamination
15     incidents, occupational exposures, direct or indirect exposure to local populations from discrete
16     sources, or exposures to subsistence or recreational fishers.
17
18     Breast-feeding infants have higher intakes of dioxin  and related compounds for a short  but
19     developmentally important part of their lives. However, the benefits of breast feeding are
20     widely recognized to  outweigh the risks.
21            Two studies have compared dioxins in infants who have been breast-fed versus those  who
22     have been formula-fed, and both have shown elevations in the concentrations of dioxins in
23     infants being breast-fed.  Formula-fed infants had lipid-based concentrations < 5 ppt TEQDF-
24     WHO98 whereas breast-fed infants had average lipid-based concentrations above 20 ppt TEQDF-
25     WHO9g (maximum of 35 ppt TEQDF-WHO98). The dose to the infant varies as a function of
26     infant body weight, the concentration of dioxins in the mother's milk, and the trend of dioxins in
27     the mother's milk to decline over time.  Doses at birth could exceed 200 pg TEQDFP-
28     WHO9g/kg/day, which would drop to about 20 pg TEQDFP-WHO9g/kg/day after 12 months. The
29     average dose over a year was calculated to be 77 pg TEQDFP-WHO9g/kg/day. Although this
30     average annual infant dose of 77 pg TEQDFP-WHO98/kg/day exceeds the currently estimated adult
31     dose of 1 pg TEQDFP-WHO9g/kg/day, the effect on infant body burdens is expected to be less
32     dramatic, i.e., infant body burdens will not exceed adult body burdens by 77 times. This is due to
33     the rapidly expanding infant body weight and lipid volume, the decrease in concentration of
34     dioxins in the mother's milk over time, and possibly more rapid elimination in infants. A
35     pharmacokinetic exercise comparing a 12-month nursing scenario with formula feeding showed
36     infant lipid concentrations to exceed 40 ppt TEQDFP-WHO98, compared with lipid concentrations
37     less than 10 ppt for the formula-fed infants. The dioxin concentrations in these two hypothetical
       6/8/00                                    99        DRAFT—DO NOT CITE OR QUOTE

-------
  1      children merged at about 10 years of age, at a lipid concentration of about 13 ppt TEQDFP-
  2      WHO98.
  3             The American Academy of Pediatrics (1997) has made a compelling argument for the
  4      diverse advantages of breast-feeding and the use of human milk for infant feeding to infants,
  5      mother, families and society. These include health, nutritional, immunologic, developmental,
  6      psychological, social, economic, and environmental benefits. Breast milk is the point of
  7      comparison for all infant food, and the breast-fed infant is the reference for evaluation of all'
  8      alternative feeding methods.  In addition, increasing the rates of breast-feeding initiation is a
  9      national health objective and one of the goals of the United States Government's Healthy People
 10      2010.  The World Health Organization (1988) maintained that the evidence did not support an
 11      alteration of WHO recommendations that promote and support breast-feeding.  A more recent
 12      consultation in  1998 (WHO, 2000) reiterated these conclusions.  Although it is important that the
 13      recommendations of these groups continue to be reevaluated in light of emerging scientific
 14      information, the Agency does not believe that finding contained in this report provides a
 15      scientific basis for initiating such a reevaluation. This conclusion is based on the fact that
 16      stronger data have been presented that body burden, not intake, is the best dose metric; that many
 17      of the noncancer effects, particularly those seen in children, are more strongly associated with
 18      prenatal exposure and the mother's body burden rather than postnatal exposures and breast milk
 19      levels; and that dioxin-like compounds are strong promoters of carcinogenicity, a mode of action
 20      that depends on late-stage impacts rather than early-stage impacts on the carcinogenic process.
 21
 22      Many dioxin sources have been identified and emissions to the environment are being
 23      reduced.
 24             Current emissions of CDDs/CDFs/PCBs to the United States environment result
 25      principally from anthropogenic activities.  Evidence that supports this finding includes matches
 26      in time of rise of environmental levels with rise in general industrial activity (see trend discussion
 27      in Section 4.6), lack of any identified large natural sources and observations of higher
 28      CDD/CDF/PCB body burdens in industrialized versus less industrialized countries (see
 29      discussion on human tissue levels in Section 4.4).
 30             The principal identified sources of environmental release may be grouped into five major
 31      types: (1) combustion and incineration sources; (2) chemical manufacturing/processing sources;
 32      (3) industrial/municipal processes; (4) biological and photochemical processes; and (5) reservoir
 33      sources. Development of release estimates is difficult because only a few facilities in most
 34      industrial sectors have been tested for CDD/CDF emissions. Thus an extrapolation is needed to
 35      estimate national emissions.  The extrapolation method involves deriving an estimate of
 36      emissions per unit of activity at the tested facilities and multiplying this by the total activity level
37      in the untested facilities. In order to convey the level of uncertainty in both the measure of

        6/8/00                                      100       DRAFT—DO NOT CITE OR QUOTE

-------
 1     activity and the emission factor, U.S. EPA developed a qualitative confidence rating scheme.
 2     The confidence rating scheme, presented in Section 4, Table 4-1, uses qualitative criteria to
 3     assign a high, medium, or low confidence rating to the emission factor and activity level for those
 4     source categories for which emission estimates can be reliably quantified. The dioxin
 5     reassessment has produced an inventory of source releases for the United States (Table 4-2).  The
 6     inventory was developed by considering all sources identified in the published literature and
 7     numerous individual emissions test reports. The inventory is limited to sources whose releases
 8     can be reliably quantified (i.e., those with confidence ratings of A, B, or C as defined above).
 9     Also, it is limited to sources with releases that are created essentially simultaneously with
10     formation.  This means that the reservoir sources are not included. The inventory presents the
11     environmental releases in terms of two reference years: 1987 and 1995. EPA's best estimates of
12     releases of CDD/CDFs to air, water, and land from reasonably quantifiable sources were
13     approximately 2,800 gram (g) (1.3 pounds) TEQDF-WHO9g in 1995 versus 13,500 g (6 pounds)
14     TEQDF-WHO9g in 1987. The decrease in estimated releases of CDD/CDFs between 1987 and
15     1995 (approximately 80%) was due primarily to reductions in air emissions from municipal and
16     medical waste incinerators.
17            The environmental releases of CDD/CDFs in the United States occur from a wide variety
18     of sources, but are dominated by releases to the air from combustion sources. Insufficient data
19     are available to comprehensively  estimate point-source releases of dioxin-like compounds to
20     water. Sound estimates of releases to water are available only for chlorine-bleached pulp and
21     paper mills and manufacture of ethylene dichloride/vinyl chloride monomer.  The contribution of
22     dioxin-like compounds to waterways from nonpoint source reservoirs is likely to be greater than
23     the contributions from point sources. Current data are only sufficient to support preliminary
24     estimates of nonpoint source contributions of dioxin-like compounds to water (i.e., urban storm
25     water runoff and rural soil erosion). These estimates suggest that, on a nationwide basis, total
26     nonpoint releases are significantly larger than point source releases.  Other releases to water
27     bodies that cannot be quantified on the basis of existing data include effluents from POTWs and
28     most industrial/commercial sources.
29            Based on the available information, the inventory includes only  a limited set of activities
30     that result in direct environmental releases to land. The only releases to land quantified in the
31     inventory are land application of sewage sludge and pulp and paper mill wastewater sludges.  Not
32     included in the inventory's definition of an environmental release is the disposal of sludges and
33     ash into approved landfills.  While this inventory is the most comprehensive and well-
34     documented in the world, it is likely to underestimate total releases.  The magnitude of the
35     underestimate is unknown but it is unlikely that noncombustion sources today, other than
36     reservoir sources, play a  dominant role in human exposure. In terms of 1995 releases from
37     reasonably quantifiable sources, this document estimates releases of 2,800 g WHO98TEQDF for
       6/8/00                                    101       DRAFT—DO NOT CITE OR QUOTE

-------
  1      contemporary formation sources and 2,900 g WHO98TEQDF for reservoir sources. In addition,
  2      there remain a number of unquantifiable and poorly quantified sources that are described in
  3      Section 4.
  4             As described above, combustion appears to be the most significant process of formation
  5      of CDDs/CDDFs today. Important factors that can affect the rate of dioxin formation include the
  6      overall combustion efficiency, post-combustion flue gas temperatures and residence times, and
  7      the availability of surface catalytic sites to support dioxin synthesis. Although chlorine is an
  8      essential component for the formation of CDD/CDFs in combustion systems, the empirical
  9      evidence indicates that for commercial-scale incinerators, chlorine levels in feed are not the
 10      dominant controlling factor for rates of CDD/CDF stack emissions.  The conclusion that chlorine
 11      in feed is not a strong determinant of dioxin emissions applies to the overall population of
 12      commercial scale combustors. For any individual commercial-scale combustor, circumstances
 13      may exist in which changes in chlorine content of feed could affect dioxin emissions. For
 14      uncontrolled combustion, such as open burning of household waste, chlorine content of wastes
 15      may play a more significant role in affecting levels of dioxin emissions than observed in
 16      commercial-scale combustors.
 17             No significant release of newly formed dioxin-like PCBs  is occurring in the United
 18      States. Unlike CDD/CDFs, PCBs were intentionally manufactured in the United States in large
 19      quantities from 1929 until production was banned in 1977.  Although it has been demonstrated
 20      that small quantities of coplanar PCBs can be produced during waste combustion, no strong
 21      evidence exists that the dioxin-like PCBs make a significant contribution to TEQ releases during
 22      combustion. The occurrences of dioxin-like PCBs in the U.S. environment most likely  reflects
 23      past releases associated with PCB production, use, and disposal.  Further support of this finding
 24      is based on observations of reductions since 1980s in PCBs in Great Lakes sediment and other
 25      areas.
 26             It is unlikely that the emission rates of CDD/CDFs from known sources correlate
 27      proportionally with general population exposures. Although the emissions inventory shows the
 28      relative contribution of various sources to total emissions, it cannot be assumed that these
 29      sources make the same relative contributions to human exposure. It is quite possible that the
 30      major sources of dioxin in food (see discussion in Section 2.6 indicating that the diet is  the
 31      dominant exposure pathway  for humans) may not be those sources that represent the largest
 32      fractions of total emissions in the United States. The geographic locations of sources relative to
 33      the areas from which much of the beef, pork, milk, and fish come is important to consider. That
 34      is, much of the agricultural areas that produce dietary animal fats are not located near or directly
 35      downwind of the major sources of dioxin and related compounds.
 36             The contribution of reservoir sources to human exposure may be significant.  Several
37      factors support this finding.  First, human exposure to the dioxin-like PCBs is thought to be

        6/8/00                                    102       DRAFT—DO NOT CITE OR QUOTE

-------
 1     derived almost completely from reservoir sources. Because one-third of general population TEQ
 2     exposure is due to PCBs, at least one-third of the overall risk from dioxin-like compounds comes
 3     from reservoir sources. Second, CDD/CDF releases from soil via soil erosion and runoff to
 4     waterways appear to be greater than releases to water from the primary sources included in the
 5     inventory. CDD/CDFs in waterways can bioaccumulate in fish-leading to human exposure via
 6     consumption of fish, which makes up about one-third of the total general population CDD/CDF
 7     TEQ exposure. This suggests that a significant portion of the CDD/CDF TEQ exposure could be
 8     due to releases from the soil reservoir.  Finally, soil reservoirs could have vapor and particulate
 9     releases that deposit on plants and enter the terrestrial food chain. The magnitude of this
10     contribution, however, is unknown.
11            This assessment adopts the hypothesis that the primary mechanism by which dioxin-like
12     compounds enter the terrestrial food chain is via atmospheric deposition. Dioxin and related
13     compounds enter the atmosphere directly through air emissions or indirectly, for example,
14    through volatilization from land or water or from resuspension of particles. Once introduced into
15     the environment, dioxin-like compounds are widely distributed in the environment as a result of
16    a number of physical and biological processes. The dioxin-like compounds are essentially
17    insoluble in water, generally classified as semivolatile, and tend to bioaccumulate in animals.
18    Some evidence has shown that these compounds  can degrade in the environment, but in general
19    they are considered very persistent and relatively immobile in soils and sediments. These
20    compounds are transported through the atmosphere, as vapors or attached to airborne  particulates
21     and can be deposited on soils, plants, or other surfaces (by wet or dry deposition). The dioxin-
22    like compounds enter water bodies primarily via direct deposition from the atmosphere, or by
23    surface runoff and erosion. From soils, these compounds can reenter the atmosphere either as
24    resuspended soil particles or as vapors. In water, they can be resuspended into the water column
25    from sediments, volatilized out of the surface waters into the atmosphere, or become buried in
26    deeper sediments. Immobile sediments appear to serve as permanent sinks for the dioxin-like
27    compounds.  Though not always considered an environmental compartment, these compounds
28    are also found in anthropogenic materials (such as pentachlorophenol) and have the potential to
29    be released from these materials into the broader  environment.
30           The two primary pathways for the dioxin-like compounds to enter the ecological food
31     chains and human diet are  air-to-plant-to-animal and water/sediment-to-fish.  Vegetation receives
32    these compounds via atmospheric deposition in the vapor and particle phases. The compounds
33    are retained on plant surfaces and bioaccumulated in the fatty tissues of animals that feed on
34    these plants.  Vapor-phase transfers onto vegetation have been experimentally shown to dominate
35     the air-to-plant pathway for the dioxin-like compounds, particularly for  the lower chlorinated
36     congeners. In the aquatic food chain, dioxins enter water systems via direct discharge or
37     deposition and runoff from watersheds. Fish accumulate these compounds through direct contact
      6/g/OO                                    103       DRAFT—DO NOT CITE OR QUOTE

-------
  1      with water, suspended particles, and bottom sediments and through the consumption of aquatic
  2      organisms. Although these two pathways are thought to normally dominate contribution to the
  3      commercial food supply, others can also be important. Elevated dioxin levels in cattle resulting
  4      from animal contact with pentacholorophenol-treated wood have been documented by the
  5      USD A.  Animal feed contamination episodes have led to elevations of dioxins in poultry in the
  6      United States, milk in Germany, and meat/dairy products in Belgium.
  7             Deposition can occur directly onto soil or onto plant surfaces.  At present, it is unclear
  8      whether atmospheric deposition represents primarily current contributions of dioxin and related
  9      compounds from all media reaching the atmosphere or whether it is past emissions of dioxin and
10      related compounds which persist and recycle in the environment. Understanding the relationship
11      between these two scenarios will be particularly important in understanding the relative
12      contributions of individual point sources of these compounds to the food chain and assessing the
13      effectiveness of control strategies focused on either current or past emissions of dioxins in
14      attempting to reduce the levels in food.
15             As discussed in Section 4.3, estimates for background levels of dioxin-like compounds in
16      environmental media are based on a variety of studies conducted at different locations in North
17      America.  Of the studies available for this compilation, only those conducted in locations
18      representing "background" were selected. The amount and representativeness of the data varies,
19      but in general these data lack the statistical basis to establish true national means. The
20      environmental media concentrations were consistent among the various studies and were
21      consistent with similar studies in Western Europe.  These data are the best available for
22      comparing site-specific values to national background levels. Because of the limited number of
23      locations examined, however, it is not known if these ranges adequately capture the full national
24      variability; if significant regional variability exists, making national means of limited utility; or if
25      elevated levels above this range could still be the result of background contamination processes.
26      As new data are collected, these ranges are likely to be expanded and refined.  The limited data
27      on dioxin-like PCBs in environmental media are summarized in the document (Part I, Volume 3,
28      Chapter 4), but were not judged adequate for estimating background levels.
29             Concentrations of CDDs/CDFs and PCBs in the United States environment were
30      consistently low prior to the 1930s. Then concentrations rose steadily until about 1970. At this
31      time, the trend reversed and concentrations have declined to the present. The most compelling
32      supportive evidence of this trend for CDD/CDFs and PCBs comes from dated sediment core
33      studies. Sediment concentrations in these studies are generally assumed to be an indicator of the
34      rate of atmospheric deposition.  CDD/CDF and PCB concentrations in sediments began to
35      increase around the 1930s and continued to increase until about 1970. Decreases began in  1970
36      and have continued to the time of the most recent sediment samples (about 1990).  Sediment data
37      from 20 United States lakes and rivers from seven separate research efforts consistently support

        6/8/00                                     104       DRAFT—DO NOT CITE OR QUOTE

-------
 1      this trend. Additionally, sediment studies in lakes located in several European countries have
 2     shown similar trends.
 3           It is reasonable to assume that sediment core trends should be driven by a similar trend in
 4     emissions to the environment. The period of increase generally matches the time when a variety
 5     of industrial activities began rising, and the period of decline appears to correspond with growth
 6     in pollution abatement. Many of these abatement efforts should have resulted in decreases in
 7     dioxin emissions, i.e., elimination of most open burning, particulate controls on combustors,
 8     phaseout of leaded gas, and bans on PCBs, 2,4,5-T, hexachlorophene, and restrictions on use of
 9     pentachlorophenol. Also, the national source inventory of this assessment documented a
10     significant decline in emissions from the late 1980s to the mid-1990s. Further evidence of a
11     decline in CDD/CDF levels in recent years is emerging from data, primarily from Europe,
12     showing declines in foods and human tissues.
13           In addition to the congener-specific PCB data discussed earlier, a wealth of data on total
14     PCBs and Aroclor mixtures exist that also supports  these trends.  It is reasonable to assume that
15     the trends for dioxin-like PCBs are similar to those  for PCBs as a class because the predominant
16     source of dioxin-like PCBs is the general production of PCBs in Aroclor mixtures.  PCBs were
17     intentionally manufactured in large quantities from  1929 until production was banned in the
18     United States in 1977.  United States production peaked in 1970, with a volume of 39,000 metric
19     tons.  Further support is derived from data showing declining levels of total PCBs in Great Lakes
20     sediments and biota during the 1970s and 1980s.  These studies indicate, however, that during
21     the 1990s the decline slowed and may be leveling off.
22           Because dioxin-like chemicals are persistent and accumulate in biological tissues,
23     particularly in animals, the major route of human exposure is through ingestion of foods
24     containing minute quantities  (part per trillion or ppt levels) of dioxin-like compounds.  This
25     results in widespread low-level exposure of the general population to dioxin-like compounds.
26     The issue of general population background exposure was discussed earlier.
27
28     Risk Characterization Summary Statement
29           Based on all of the data reviewed in this reassessment and scientific inference, a picture
30     emerges of TCDD and related compounds as potent toxicants in animals with the potential to
31     produce a spectrum of effects. Some of these effects may be occurring in humans at general
32     population background levels and may be resulting  in adverse impacts on human health. The
33     potency  and fundamental level at which these compounds act on biological systems is analogous
34     to several well-studied hormones. Dioxin and related compounds have the ability to alter the
35     pattern of growth and differentiation of a number of cellular targets by initiating a series of
36     biochemical and biological events, resulting in the potential for a spectrum of cancer and

       6/8/00                                    105       DRAFT—DO NOT CITE OR QUOTE

-------
 1      noncancer responses in animals and humans. Despite this potential, there is currently no clear
 2      indication of increased disease in the general population attributable to dioxin-like compounds.
 3      The lack of a clear indication of disease in the general population should not be considered
 4      strong evidence for no effect of exposure to dioxin-like compounds. Rather, lack of a clear
 5      indication of disease may be a result of the inability of current data and scientific tools to directly
 6      detect effects at these levels of human exposure. Several factors suggest a need to further
 7      evaluate the impact of these chemicals on humans at or near current background levels.  These
 8      are the weight of the evidence on exposure and effects, an apparently low margin of exposure for
 9      noncancer effects, potential for significant risks to some portion of the general population, and
10      additivity to background processes related to carcinogenicity in the case of incremental exposures
11      above background.
       6/8/00                                     106       DRAFT—DO NOT CITE OR QUOTE

-------
       Table 1-1. The TEF scheme for I-TEQ
                                             DF
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

aNote 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.
6/8/00
107
DRAFT—DO NOT CITE OR QUOTE

-------
        Table 1-2. The TEF scheme for TEQDFP-WHO
                                                          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-like
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
0.0001
0.0001
0.0005
0.0005
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 equivalence
 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.
6/8/00
108
DRAFT—DO NOT CITE OR QUOTE

-------
       Table 1-3.  The TEF scheme for TEQDFP-WHO,
                                                       98
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
equivalence 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 assigned in the new WHO TEF scheme.
6/8/00
109
DRAFT—DO NOT CITE OR QUOTE

-------
oo
o
o
                Table 2-1.  Effects of TCDD and related compounds in different animal species
Effect
Presence of AhR
Binding of TCDD:
AhR Complex to
the DRI- (enhancer)
Enzyme induction
Acute lethality
Wasting syndrome
Teratogenesis/fetal
toxicity, mortality
Endocrine effects
Immunotoxicity
Carcinogenicity
Neurotoxicity
Chloracnegenic
effects
Porphyria
Ilepatotoxicity
Edema
Testicular atrophy
Bone marrow
hypoplasia
Human
+
+
+
0

+/-
+/-
+/-
+/-
+
+
+
+



Monkey
+

+
+
+
+
+
+

+
+
0
+
+
+
+
Guinea
Pig
0
+
+
+
+
+

+



0
+/-
0
+
+
Rat
+
+
+
+
+
+
+
+
+
+

-t-
-t-
0
+

Mouse
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+/-
Hamster
+
+
+
+
+
-f

+
+


0
+/-
+


Cow
+
+

+
+


+


+

+



Rabbit
+
+
+
+
+
+




+

+



Chicken
+
+
+
+

+

+

+

+
+
+

+
Fish
+
+
+
+
+
+
+
+
+

+

+
+


Avian
wildlife
+

+
+
+
+
+





+



Marine
mammals
+

+
+
+
+
+
+




+



Mink
+

+
+
+
+
+





+



 O
 O

 2
 O
 H
 O
 H-t
 H
 m
+ = observed.

+/- = observed to limited extent, or +/- results.

0 = not observed.

Blank cells = no data.
o
G
O
H

-------
      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.
6/8/00                                         111   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
nonrepresentati ve .
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
6/8/00
112  DRAFT—DO NOT CITE OR QUOTE

-------
Table 4-2. Quantitative inventory of environmental releases of TEQDF-WHO9g in the United States
Emission source category
Releases (g TEC
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 incinerators'
Power/Energy Generation
Vehicle fuel combustion - leaded1"
- 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 air'
Confidence rating*
Reference vear 1995
A
t^mto,/
































B
C
Confidence rating"
Reference vear 1987
A

1250
5.8



14.8






27.6
60.1







2.3


28

<0.5€


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
































B

8877
5



6.1






26.4
50.8







2




<0.5e


1.29


C



0.78
2590
5 S

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
6/8/00
113   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 rating"
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
'Leaded 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.
dlt 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 TEQDF-WHO,S emission estimate.
'Included within estimate for Wood Combustion - Industrial.
6/8/00
114    DRAFT—DO NOT CITE OR QUOTE

-------
  Table 4-3.  Preliminary indication of the potential magnitude of TEQDF-WHO98 releases from
  "unqualified" (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)
Backyard Barrel Burning
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
Water
Water
Water
Preliminary release estimate
(g WHOM-TEQDF/yr)
0.22s
6.0"
39.6"
32.0"
• T
6.6
1,050"
>20"
28.3"
804
6.9"
44.3"
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-WHO,S emissions estimate.
6/8/00
115   DRAFT—DO NOT CITE OR QUOTE

-------
  Table 4-4. Unqualified sources
Category
Combustion sources
Metal smelting and refining
Chemical manufacturing
Biological and photochemical processes
Reservoir sources
Unqualified sources
Uncontrolled combustion of PCBs
Agricultural burning
Primary aluminum
Primary magnesium
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
6/8/00
116   DRAFT—DO NOT CITE OR QUOTE

-------
Table 4-5. Estimates of the range of typical background levels of dioxin-Iike compounds in various
environmental media
Media
Rural soils
Urban soils
Sediments
Rural air
Urban air
TEQnF-WHO98 concentrations
l-6pg/g(ppt)
7-20 pg/g
l-60pg/g
0.002-0.02 pg/m3
0.02-0.2 pg/m3
 6/8/00
                                         117  DRAFT—DO NOT CITE OR QUOTE

-------
 Table 4-6. Estimates of levels of dioxin-like compounds in food
Food type
Beef
Pork
Eggs
Chicken
Milk
Dairy products
Marine fish
Freshwater fish
Marine shellfish
Vegetable fats
Water
CDD/CDFs
(pgTEQDF-WH098/g
fresh weight)
0.2
0.22
0.032
0.11
0.031
0.12
0.36
1.2
0.79
0.056
0.00056 (pg/L)
PCBs
(pgTEQP-WH098/g
fresh weight)
0.094
0.09
0.1
0.044
0.016
0.058
0.25
1.2
0.042
0.037
NA
Total
(pgTEQDFP-WH098/g
fresh weight)
0.29
0.31
0.13
0.15
0.047
0.18
0.61
2.4
0.83
0.093
NA
NA = not available.
6/8/00
118  DRAFT—DO NOT CITE OR QUOTE

-------
 Table 4-7. Background serum levels in the United States 1995 -1997

Median
Mean
95th Percentile
TEQDFPWH098(pg/glipid)
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.
6/8/00
                                            119   DRAFT—DO NOT CITE OR QUOTE

-------
  Table 4-8.  Adult contact rates and background intakes of dioxin-like compounds
Exposure route
Soil ingestion
Freshwater fish
Marine fish
Marine shellfish
Inhalation
Milk
Dairy
Eggs
Beef
Pork
Poultry
Vegetable fat
Water
Contact rate
50 mg/d
6g/d
12.5 g/d
1.6g/d
13.3mVd
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-WHO,8
12 pg/g
1.2pg/g
0.36 pg/g
0.79 pg/g
0.12pg/m3
0.031 pg/g
0.1 2 pg/g
0.032 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«,/ke-d)
0.0085
0.13
0.064
0.018
0.023
0.078
0.094
0.0077
0.13
0.048
0.054
0.014
0.000011
0.65
(45 pg/d)
Dioxin-like PCBS
Concentration
TEQp-WHO,,
NA
1-2 pg/g
0.25 pg/g
0.042 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
(Pg TEQP-
WHO«./ke-d)
NA
0.11
0.045
0.0096
NA
0.040
0.046
0.024
0.063
0.0020
0.022
0.0090
NA
0.35
(25 pg/d)
Total
intake
(pg TEQDFP-
WHO0, /kg-d)
0.0085
0.24
0.11
0.028
0.023
0.12
0.14
0.032
0.19
0.05
0.076
0.023
0.000011
1.0
(70 pg/d)
6/8/00
120  DRAFT—DO NOT CITE OR QUOTE

-------
 Table 4-9. Variability in average daily TEQ intake as a function of age
Age range
l-5yr
6-1 lyr
12-19 yr
Adult
Intake, mass basis
pgTEQnFP-WHOq))/d
54
58
63
70
Intake, body weight basis
PR TEQnFP-WHO,8/kg-d
3.6
1.9
1.1
1
6/8/00
                                        121  DRAFT—DO NOT CITE OR QUOTE

-------
ON
oo
O
O
Table 5-1. Scrum dioxin levels in the background population and epidemiological cohorts (back-calculated)
Cohort
CDC comparison
population, USA 1995 -
97; CDC 2000
Background, Dioxin
Assessment, USA -1990s
No.
316
pooled
results
Total TEQ
ppt lipid
Lower
2a
30
Central
Tend.
25.4 meanb
52.8 mean
55 median
Upper
50"
70
2,3,7,8-TCDD
ppt lipid
Central Tendency
2. 1 mean
1.9 median
(95% UCL = 4.2)
5.2 mean
SD~1.32C
PCBs
Mean
TEQ
5.3 (est.)b
18.8 mean
20 median
Non-2,3,7,8-TCDD
TEQ ppt lipid
Central Tendency
23.3 mean
47.6 mean
Comment

TEQDFP-WHO9g; serum;
missing PCBs 105, 118,
156 estimated
TEQDFP-WHO,8; serum,
adipose, breast milkd
Back-Calculated
Ranch Hand, low;
Ketchumetal. 1999
Ranch Hand, high;
Ketchumetal. 1999
Hamburg cohort women;
Flesch-Janys et al. 1999
N1OSH, Fingerhut et al.
1991b,NTIS
BASF, severe chloracne;
Ottetal. 1993
BASF, moderate
chloracne; Ott et al. 1993
BASF, no chloracne;
Ottetal. 1993
Seveso Zone A;
Landietal. 1998
Seveso Zone A, medical;
Needhametal. 1999
276
283
652.3,7,8
64TEQ
253
56
59
139
7
296


19.3e








811.2 meanc
172.8s median








6789. T






52.3 median
(range 27 - 94)
195.7 median
(range 94 - 3,290)
506.8 mean
125.8 median
(range 2.4 - 6397.4)
2,000 mean
(rangef 2 -32,000)
1008 geom. mean
(range8 20- 13360)
420.8 geom. mean
(range8 2.72 -49 15)
38.4 geom. mean
(range8 2.72 -2981)
230 geom. mean
325.9 median
(range 4 1.2 -399.7)
381 -4 89 median
(range 1.5 -56,000)











304.4 meane






serum
serum
I-TEQs, dioxin and
furan TEQ only; serum
serum
serum
serum
serum
serum
Samples taken 1976, not
back-calculated; serum;
using 1/2 DL
 K)
 s-/

 5
 H
 O
 O
 z;
 o
 H
 O
O
&
O
w

-------
o
o
           Table 5-1. Serum dioxin levels in the background population and epidemiological cohorts (back-calculated) (continued)
Seveso Zone B;
Landietal. 1998
Seveso Zone B, medical;
Needham et al. 1999
Seveso Zone R, medical;
Needham etal. 1999
Seveso NonABR;
Landietal. 1998
Dutch Accident;
Hooiveld et al. 1996
Dutch Main Production;
Hooiveld et al. 1996
51
80
48
52
14
5


















47.5 geom. mean
52.5 median
(range 5.3 - 273)
87- 147 median
(range 1.8 -725)
15 -89 median
(range I - 545)
4.9 geom. mean
5.5 median
(range 1.0- 18.1)
1841.8 arith. mean
1433.8 geom. mean
(range 301 -3683)
608.2 arith. mean
285.9 geom. mean
(range 17-1160)












serum
Samples taken 1 976, not
back-calculated; serum;
using Vz DL
Samples taken 1976; not
back-calculated; serum;
using '/2 DL
serum
serum
serum
to
OJ
       a Estimated from ATSDR 1999 Calcasieu comparison population graph.
       b CDC data scaled upward to adjust for missing data on PCB congeners 105, 1
       c SD approximated from unweighted estimate.
       d Weighted average levels for the subset of serum lipid TEQs were 4.54 ng/kg
       congeners).
       ePCDD and PCDF derived TEQ only, using I-TEFs.
       f Lower interval on current level.
       8 Range estimated from exponential log distribution graph.
18 and 156, by matching to PCB congener ratios measured in the early 1990s.

for 2,3,7,8-TCDD, and 55.4 ng/kg for total TEQ (PCB contribution not adjusted for missing
 U
 O
 2
 o
 H
 O
 §
o
 c:
 o
 H
 tn

-------
Table 5-2. Doses yielding 1% excess risk (95% lower confidence bound) based upon 2-year
animal carcinogenicity studies using simple multistage (Portier et. al, 1984) models3
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
ED
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)
01
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)
' Reprinted with slight modifications from Chapter 8, Table 8.3.2.
6/8/00
124     DRAFT—DO NOT CITE OR QUOTE

-------
     Cl
       2,3,7,8-Tetrachlorodibenzo-p-dioxin
     Cl
       1,2,3,7,8-Pentachlorodibenzo-p-dioxin
         Cl
        3,3',4,4',5,5'-Hexachlorobiphenyl
                                                Cl
2,3,7,8-Tetrachlorodibenzofuran
                                                Cl
2,3,4,7,8-Pentachlorodibenzofuran
                                                                          ci
3,3',4,4',5-Pentachlorobiphenyl
Figure 1-1.  Chemical structure of 2,3,7,8-TCDD and related compounds.
6/8/00
                                              125       DRAFT—DO NOT CITE OR QUOTE

-------
                    phosphorylation/
                    dephospht »ry lation
                    Differentiation
                         and
                     Proliferation
                                                       Altered gene expression
       Changes in protein levels
       (e.g., CYPIA1.IL-1B,...)
                                mRNA
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.
6/8/00
126
DRAFT—DO NOT CITE OR QUOTE

-------
 CYP1A1
 CYP1A2
 CYP1B1
 Glutathione S-Transferase Ya
 Aldehyde-3-Dehydrogenase
 NAD(P)H:Quinone Oxidoreductase
TGF-a
TGF-P
Plasminogen Activator Inhibitor-2
Interleukin-lp
c-fos
jun
Figure 2-2. Some of the genes whose expression is altered by exposure to TCDD.
6/8/00
                                      127
    DRAFT—DO NOT CITE OR QUOTE

-------
                                       Best Estimate of
               Emission Source
              (tested/total units)
  Municipal Solid Waste Inctrveratton (39/13O)

      Medical Wane I nc In »fan on (2O/2.4OO)

          Secondary Cooper Smelling (2/3)


                    Forest  Fires (NA)

    Cement Kilns Burning H02 Wasie (1OY34)

    Residential Wood Burning (7/25.OOO.OOO)

   Utility  / Industrial  Coal Combustion (11/7)

  On-Road Diesel  Fuel Combustion (MA)

      Secondary Aluminum Smelling (6/76)

             Industrial Wood Burning I9/?)

          Iron Ore Sinter Plants  (2/1 1)

Cement Kilns Not Burning Haz Waste (15/178)

        Sewage Sludge Incineration (13/257}

             Manufacture of EDC/VC (?/?)

   UUmy  i Industrial Oil Combustion  (^-2/7)

                   Crematoria (16/1.555)

Hazardous Waste  Incineration (17/162)



  On-Road Leaded Gas Fuel Combustion (?/?]
Emission Factor
 (ng/kg or ng/L)
   Total Annual "Activity"
(thousand metric tons/yr or
      million L/yr)
                                                   o  a o  o  <•   -
                                                   >•  o o  o  ,;
                                                      ,-00°
                                                        -  o
                                     o   o   o
                                     O   0   O
                                         0   O
The figures include sources with annual TEQ emission estimates greater than 5 g TEQDF-WH09B/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 TEQ Emission
 (g TEQDF-WH098/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.
       6/8/00
                         128
               DRAFT—DO NOT CITE OR QUOTE

-------
     Municipal Solid Waste Incineration
           Medical Waste Incineration
         Secondary Copper Smelting
                    Forest Fires
      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 EDC/VC
      Utility/Industrial Oil Combustion
                    Crematoria
        Hazardous Waste Incineration
On-Road Unleaded Gas Fuel Combustion
  On-Road Leaded Gas Fuel Combustion
                                                 10
         100
                1000
10000
                                                     1995 D1987
        Figure 4-2.  Comparison of estimates of annual I-TEQ emissions to air (grams I-TEQ/yr)
        for reference years 1987 and 1995.
        6/8/00
129
DRAFT—DO NOT CITE OR QUOTE

-------
   10000
                                           14000
   1000
JU


"f  100

1
1
, —


Mulian'


Vledia


n







Mean


•«


Mean





\
I



g-Mea

J
2

1
?-Mean
i

»-Mea


|

 g>
I   10
o
o
o
(N
           U
                o
                0
     !

     on
Bti
03
                     O
                    I
     U
     on
03
o
on
15
u
g
E
03
o
>
1
<
o
w

on


"ca
•5
u
E
<
o
VI
on
£
_o
'
•o
§
0
c
a
K
f
IE
I
T3
n
u
^
Q£
**r
cd
QJ
£?
3
E
I
                                                              on
                                                              O
                                          —   ^
                                     u
                                     o

                                     IX
                                     on
                                     <
                                     03
          c
          (LI
         T3
-a
o
E
u.
on
<
03
                                                        u.   •«
                                                        5    =
                                                        ffl   Q
-a
o
£
j=

Q
                          ; 2378-TCDD g Est. Background Non-TCDD TEQs
        Figure 5-1. Dioxin body burden levels in background populations and epidemiological
        cohorts (back-calculated).
        6/8/00
                            130
                            DRAFT—DO NOT CITE OR QUOTE

-------
        REFERENCES FOR RISK CHARACTERIZATION

 1      Abbott, BD; Schmid, JE; Pitt, JA; et al., (1999) Adverse reproductive outcomes in the transgenic AhR-deficient
 2      mouse. Toxicol Appl Pharmacol 155(1):62-70.
 3
 4      Abraham, K; Krowke, R; Neubert, D. (1988) Pharmacokinetics and biological activity of 2,3,7,8-
 5      tetrachlorodibenzo-p-dioxin.  1. Dose-dependent tissue distribution and induction of hepatic ethoxyresorufin 0-
 6      deethylase in rats following a single injection. Arch Toxicol 62:359-368.
 7
 8      Ahlborg, VG; Becking, GC; Birnbaum, LS; et  al. (1994) Toxic equivalency factors for dioxin-like PCBs
 9      Chemosphere 28(6): 1049-1067.
10
11      Alaluusua, S; Lukinmaa, P-L; Torppa, T; et al. (1999) Developing teeth as biomarker of dioxin exposure. Lancet
12      353:206.
13
14      Alaluusua, S; Lukinmaa, P-L; Vartiainen, T; et al. (1996) Polychlorinated dibenzo-p-dioxins and dibenzofurans via
15      mother's milk may cause developmental defects in the child's teeth. Environ Toxicol Pharmacol 1:193-197.
16
17      Allen, BC; Kavlock, RJ; Kimmel, CA; et al. (1994) Dose-response assessment for developmental toxicity. II.
18      Comparison of generic benchmark dose estimates with no observed adverse effect levels. Fundam Appl Toxicol
19      23:487-495.
20
21      Allen, JR.; Lalich, JJ. (1962) Response of chickens to prolonged feeding of crude "toxic fat." Proc Soc Exp Biol
22      Med 109:48-51.
23
24      Allen, JR; Carstens, LA. (1967) Light and electron microscopic observations in Macaco mulatto monkeys fed toxic
25      fat. Am J Vet Res 28:1513-1526.
26
27      Allen, JR; Barsotti, DA; Van Miller, JP; et al. (1977) Morphological changes in monkeys consuming a diet
28      containing low levels of 2,3,7,8-tetrachIorodibenzodioxin. Food Cosmet Toxicol 15:401-410.
29
30      Allen, JR.; Barsotti, DA; Lambrecht, LK; et al. (1979) Reproductive effects of halogenated aromatic hydrocarbons
31      on nonhuman primates. Ann  N Y Acad Sci 320:419-425.
32
33      Alsharif, NZ; Lawson, T; Stohs, SJ. (1994) Oxidative stress induced by 2,3,7,8-tetrachlorodibenzo-p-dioxin is
34      mediated by the aryl hydrocarbon (Ah) receptor complex. Toxicology 92:39-51.
35
36      American Academy of Pediatrics. (1997) Breastfeeding  and the use of human milk. Pediatrics 100 (6): 1035-1039.
37
38      Ambrosone, CB; Freundenheim, JL; Graham, S; et al.(1995) Cytochrome P4501A1 and glutathione-s-transferase
39      (Ml) genetic polymorphisms and post-menopausal breast cancer risk. Cancer Res 55:3483-3484. (Mike D. - Check
40      page numbers)
41
42      Andersen, ME; Birnbaum, LS; Barton, HA; et  al. (1997) Regional hepatic CYP1A1 and CYP1A2 induction with
43      2,3,7,8-tetrachlorodibenzo-p-dioxin evaluated with a multi-compartment geometric model of hepatic zonation.
44      Toxicol Appl Pharmacol 144:145-155.
45
46      Ariens, EJ; van Rossum, JM; Koopman, PC. (1960) Receptor reserve and threshold phenomena. 1. Theory and
47      experiments with autonomic  drugs tested on isolated organs. Arch Int Pharmacodyn 127:459-478.
48
49      Arnold, DL; Nera, EA; Stapley, R; et al. (1996) Prevalence of endometriosis in rhesus (Macaca mulatta) monkeys
50      ingesting PCB (Aroclor 1254): review and evaluation. Fundam Appl Toxicol 31(1 ):42-55.
51
52
        6/8/00
131        DRAFT—DO NOT CITE OR QUOTE

-------
  1       ATSDR. (1999) Toxicological profile for chlorinated dibenzo-p-dioxins. United States Department of Health and
  2       Human Services.
  3
  4       Aylward, LL; Hays, SM; Karch, NJ; et al. (1996) Relative susceptibility of animals and humans to the cancer hazard
  5       posed by 2,3.7,8-tetrachlorodibenzo-p-dioxin using internal measures of dose. Environ Sci Technol 30:3534-3543.
  6
  7       Barsotti, DA; Abrahamson, LJ; Allen, JR. (1979) Hormonal alterations in female rhesus monkeys fed a diet
  8       containing 2,3,7,8-tetrachlorodibenzo-p-dioxin. Bull Environ Contam Toxicol 21:463-469.
  9
10       Becher, H; Flesch-Janys, D; Kauppinen, T; et al. (1996) Cancer mortality in German male workers exposed to
11       phenoxy herbicides and dioxins. Cancer Causes Control 7:312-321.
12
13       Becher, H; Steindorf, K.; Flesch-Janys, D. (1998) Quantitative cancer risk assessment for dioxins using an
14       occupational cohort. Environ Health Perspect 106(2):663-670.
15
16       Beck, H; Eckart, K; Mathar, W; et al. (1989) Levels of PCDD's and PCDF's in adipose tissue of occupationally
17       exposed workers. Chemosphere 18:507-516.
18
19       Bertazzi, PA; di Domenico. (1994) Chemical, environmental, and health aspects of the Seveso, Italy, accident. In:
20       Dioxins and Health. Arnold Schecter, ed. New York: Plenum Press, pp. 587-632.
21
22       Bertazzi, PA; Pesatori, AC; Consonni, D; et al. (1993) Cancer incidence in a population accidentally exposed to
23       2,3,7,8-tetrachlorodibenzo-para-dioxin. Epidemiology 4(5):398-406.
24
25       Bertazzi, PA; Zocchetti, C; Guercilena, S; et al. (1997) Dioxin exposure and cancer risk: a 15-year mortality study
26       after the "Seveso Accident." Epidemiology 8(6):646-652.
27
28       Bertazzi, PA; Bemucci, I; Brambilla, G; et al. (1998) The Seveso studies on early and long-term effects of dioxin
29       exposure: a review. Environ Health Perspect 106(2):625-633.
30
31       Bertazzi, PA; Pesatori, AC; Consonni, D; et al. (1999) Epidemiology of long-term health effects of dioxin exposure
32       in the Seveso population. Organohalogen Compounds 44:337-338.
33
34       Birnbaum, L. (1994a) Evidence for the role of the AhR in responses to dioxin. In: Receptor-mediated biological
35       processes: implications for evaluating carcinogenesis. Progress in Clinical and Biological Research, vol. 387.
36       Spitzer, HL; Slaga, TJ; Greenlee, WF; et al., eds. New York: Wiley-Liss, Inc., pp.  139-154.
37
38       Birnbaum, LS. (1994b) The mechanism of dioxin toxicity: relationship to risk assessment. Environ Health Perspect
39       102 (Supplement 9):157-167.
40
41       Bjerke, DL; Peterson, RE. (1994) Reproductive toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in male rats:
42       different effects of in utero versus lactational exposure. Toxicol Appl Pharmacol 127:241-249.
43
44       Bjerke, DL; Sommer, RJ; Moore, RW; et al. (1994a).  Effects of in utero and lactational 2,3,7,8-tetrachlorodibenzo-
45       p-dioxin exposure  on repsonsiveness of the male rat reproductive system to testosterone stimulation in adulthood.
46       Toxicol Appl  Pharmacol 127:250-257.
47
48       Bjerke, D L; Brown, TJ;  MacLusky, NJ; Hochberg, RB; Peterson, RE. (1994b). Partial demasculinization and
49       feminization of sex behavior in male rats by in utero and lactational exposure to 2,3,7,8-tetrachIorodibenzo-p-
50       dioxin is not associated with alterations in estrogen receptor binding or volumes of sexually differentiated brain
51       nuclei. Toxicol Appl Pharmacol 127(2): 258-67.
52
53
54
         6/8/00                                           132        DRAFT—DO NOT CITE OR QUOTE

-------
 1       Bond, GG; McLaren, EA; Brenner, FE; et al. (1989) Incidence of chloracne among chemical workers potentially
 2      exposed to chlorinated dioxins. J Occup Med 31:771-774.
 3
 4      Bookstaff, RC; Kamel, F; Moore, RW; et al. (1990a) Altered regulation of pituitary gonadotropin-releasing
 5      hormone (GnRH) receptor number and pituitary responsiveness to GnRH in 2,3,7,8-tetrachlorodibenzo-p-dioxin-
 6      treated male rats. Toxicol Appl Pharmacol 105:78-92.
 7
 8      Bookstaff, RC; Moore, RW; Peterson, RE. (1990b) 2,3,7,8-tetrachlorodibenzo-p-dioxin increases the potency of
 9      androgens and estrogens as feedback inhibitors of luteinizing hormone secretion in male rats. Toxicol Appl
10      Pharmacol 104:212-224.
11
12      Boyd, JA; Clark, GC; Walmer, D; et al. (1995) Endometriosis nd the environment: biomarkers of toxin exposure.
13      Conference on Endometriosis 2000, May 15-17.
14
15      Breslow, NE; Day, NE. (1987) Statistical methods in cancer research. Volume II-The design and analysis of cohort
16      studies. lARCSciPubl 82:1-406.
17
18      Brown, NM; Manzolillo, PA; Zhang, JX; et al. (1998) Prenatal TCDD and predisposition to mammary cancer in the
19      rat. Carcinogenesis 19(9): 1623-1629.
20
21      Bruner-Tran, KL; Rier, SE; Eisenberg, E; et al. (1999) The Potential Role of Environmental Toxins in the
22      Pathophysiology of Endometriosis. Gynecol Obstet Invest Oct;48 Suppl SI :45-56.
23
24      Bueno de Mesquita, HB; Doornbos, G; van der Kuip, DM; et al. (1993) Occupational exposure to phenoxy
25      herbicides and chlorophenols and cancer mortality in the Netherlands. Am J Ind Med 23:289-300.
26
27      Calvert, GM; Homung, RW; Sweeney, MH; et al. (1992) Hepatic and gastrointestinal effects in an occupational
28      cohort exposed to 2,3,7,8-tetrachlorodibenzo-para-dioxin. JAMA 267:2209-2214.
29
30      Calvert, GM; Willie, KK; Sweeney, MH; et al. (1996) Evaluation of serum lipid concentrations among U.S.
31      workers exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Arch Environ Health 51(2): 100-107.
32
33      Calvert, GM; Sweeney, MH; Deddens, J; et al. (1999) Evaluation of diabetes mellitus, serum glucose, and thyroid
34      function among United States workers exposed to 2,3,7,8-tetrachlorodi-benzo-/?-dioxin. Occup Environ Med
35      56(4):270-276.
36
37      Caramaschi, F; Del Caino, G; Favaretti, C; et al. (1981) Chloracne following environmental contamination by
38      TCDD in Seveso, Italy. Int J Epidemiol 10:135-143.
39
40      Carver, LA; LaPres, JJ; Jain, S; et al. (1998) Characterization of the AhR-associated protein, ARA9. J Biol Chem
41      273(50):33580-33587.
42
43      CDC (2000) Personal communication from D. Patterson,  CDC, Atlanta, GA to M. Lorber, U.S. EPA, Washington,
44      DC. April, 2000.
45
46      Centers for Disease Control Vietnam Experience Study. (1988) Health status of Vietnam veterans. II. Physical
47      health. JAMA 259:2708-2714.
48
49      Chahoud, I.; Krowke, R.; Schimmel, A.; et al. (1989) Reproductive toxicity and pharmacokinetics of 2,3,7,8-
50      tetrachlorodibenzo-/>-dioxin. I. Effects of high doses on the fertility of male rats. Arch Toxicol 63:432-439.
51
52      Chen, YCJ; Guo, YLL; Hsu, CC. (1992) Cognitive development of children prenatally exposed to polychlorinated
53      biphenyls (Yu-Cheng children) and their siblings. J Formosan Med Assoc 91:704-707.
54
        6/8/00                                           133        DRAFT—DO NOT CITE OR QUOTE

-------
  1       Cheung, MO; Gilbert, EF; Peterson, RE. (1981) Cardiovascular teratogenicity of 2, 3, 7,
  2       8-tetrachlorodibenzo-p-dioxin in the chick embryo. Toxicol Appl Pharmacol 61(2): 197-204.
  3
  4       Clark, GC; Tritscher, A; Maronpot, R; et al. (1991) Tumor promotion by TCDD in female rats. In: Banbury Report
  5       35: biological basis for risk assessment of dioxin and related compounds. Gallo, M; Scheuplein, R; van Der
  6       Heijden, K, eds. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory; pp. 389-404.
  7
  8       Clark, AJ. (1933) The mode of action of drugs on cells. Baltimore: Williams and Wilkins.
  9
 10       Cohen, GM; Bracken, WM; Iyer, RP; et al. (1979) Anticarcinogenic effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin
 11       on benzo(a)pyrene and 7,12-dimethylbenz(a)anthracene tumor initiation and its relationship to DNA binding.
 12       Cancer Res 39:4027-4033.
 13
 14       Courtney, KD; Moore, JA. (1971) Teratology studies with 2,4,5-T and 2,3,7,8-TCDD. Toxicol Appl Pharmacol
 15       20:396-403.
 16
 17       Couture, LA; Abbott, BD; Bimbaum, LS. (1990) A critical review of the developmental toxicity and teratogenicity
 18       of 2,3,7,8-tetrachlorodibenzo-p-dioxin: recent advances toward understanding the mechanism. Teratology 42:619-
 19       627.
 20
 21       Cummings, AM; Metcalf, JL; Birnbaum, L. (1996) Promotion of endometriosis by
 22       2,3,7,8-tetrachlorodibenzo-p-dioxin in rats and mice: time-dose dependence and species comparison. Toxicol Appl
 23       Pharmacol 138(1):131-139.
 24
 25       Dannan, GA; Porubek, DJ; Nelson, SD; et al. (1986) 17 beta-estradiol 2- and 4-hydroxylation catalyzed by rat
 26       hepatic cytochrome P-450: roles of individual forms, inductive effects, developmental patterns, and alterations by
 27       gonadectomy and hormone replacement. Endocrinology 118:1952-1960.
 28
 29       Davis, D; Safe, S. (1988) Immunosuppressive activities of poly chlorinated dibenzofuran congeners: quantitative
 30       structure-activity relationships and interactive effects. Toxicol Appl Pharmacol 94:141-149.
 31
 32       Denison, MS; Phelan, D; Elferink, CJ. (1998) The AhR signal transduction pathway. In: Toxicant-receptor
 33       interactions. Denison, MS; Helferich,  WG, eds. Bristol. PA: Taylor &  Francis, pp. 3-33.
 34
 35       Dertinger, SD; Silverstone, AE; Gasiewicz, TA. (1998) Influence of aromatic hydrocarbon receptor-mediated events
 36       on the genotoxicity of cigarette smoke condensate. Carcinogenesis 19:2037-2042.
 37
 38       DeVito, MJ; Ma, XF; Babish, JG; et al. (1994) Dose-response relationships in mice following subchronic exposure
 39       to 2,3,7,8-tetrachlorodibenzo-p-dioxin: cyplal, cypla2, estrogen-receptor, and protein-tyrosine phosphorylation.
 40       Toxicol Appl Pharmacol  124:82-90.
 41
 42       DeVito, MJ; Birnbaum, LS; Farland, WH; et al. (1995) Comparisons of estimated human-body burdens of
 43       dioxinlike chemicals and TCDD body burdens in experimentally exposed animals. Environ Health Perspect
 44       103:820-831.
 45
 46       DiGiovanni, J.; Berry, DL; Gleason, GL; et al. (1980) Time-dependent inhibition by 2,3,7,8-tetrachlorodibenzo-p-
47       dioxin of skin tumorigenesis with polycyclic hydrocarbons. Cancer Res 40:1580-1587.
48
49       Diliberto, JJ; Akubue, PI; Luebke, RW; et al. (1995) Dose-response relationships of tissue distribution and
 50       induction of CYP1A1 and CYP1A2 enzymatic-activities following acute exposure to 2,3,7,8-tetrachlorodibenzo-p-
 51       dioxin (TCDD) in mice. Toxicol Appl Pharmacol 130:197-208.
 52
53       Doss, M; Saver, H; von Tiepermann, R; et al. (1984) Development of chronic hepatic porphyria (porphyria cutanea
54       tarda) with inherited uroporphyrinogen decarboxylase deficiency under exposure to dioxin. J Biochem 16:369-373.
55

         6/8/00                                           134        DRAFT—DO NOT CITE OR QUOTE

-------
 1       Dragan, YP; Xu, X; Goidsworthy, XL; et al. (1992) Characterization of the promotion of altered hepatic foci by
 2       2,3,7,8-tetrachlorodibenzo-p-dioxin in the female rat. Carcinogenesis 13(8)-1389-1395
 3
 4       DiGiovanni, J; Viaje, A; Berry, DL; et al. (1977) Tumor initiating ability of TCDD and Arochlor 1254 in the two
 5       stage system of mouse skin carcinogenesis. Bull Environ Contam Toxicol 18:552-557.
 6
 7       Dunagin, WG. (1984) Cutaneous signs of systemic toxicity due to dioxins and related chemicals  J Am Acad
 8       Dermatol 10(4):688-700.
 9
10      Dunson, DB; Haseman, JK; van Birgelen, APJM; et al. (2000) Statistical analysis of skin tumor data from Tg.AC
11       mouse bioassays. Toxicol Sci, in press.
12
13      Eastin, WC; Haseman, JK; Mahler, JF; et al. (1998) The National Toxicology Program evaluation of genetically
14      altered mice as predictive models for identifying carcinogens. Toxicol Pathol 26:461 -473.
15
16      Egeland, GM; Sweeney, MH; Fingerhut, MA; et al. (1994) Total serum testosterone and gonadotropins in workers
17      exposed to dioxin. Am J Epidemiol 139:272-281.
18
19      Enan, E; Matsumura, F. (1994)  2,3,7,8-Tetrachlorodibenzo-/?-dioxin (TCDD)-induced changes in glucose
20      transporting activity in guinea pigs, mice, and rats in vivo and in vitro. J Biochem Toxicol 9(2):97-106.
21
22      Enan, E; Matsumura, F. (1996)  Identification of c-Src as the integral component of the cytosolic AhR complex,
23      transducing the signal of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) through the protein phosphorylation pathway.
24      Biochem Pharmacol 52(10):1599-1612.
25
26      Eriksson, M; Hardell, L; Berg, NO; et al. (1981) Soft-tissue sarcomas and exposure to chemical substances: a case-
27      referent study. Br J Ind Med 38:27-33.
28
29    '  Eriksson, M; Hardell, L; Adam, H. (1990) Exposure to dioxins as a risk factor for soft tissue sarcoma: a population-
30      based case-control study. J Natl Cancer Inst 82:486-490.
31
32      Ernst, M; Flesch-Janys, D; Morgenstern, I;  et al. (1998)  Immune cell functions in industrial  workers after exposure
33      to 2,3,7,8-tetrachlorodibenzo-p-dioxin: dissociation of antigen-specific T-cell responses in cultures of diluted whole
34      blood and of isolated peripheral blood mononuclear cells. Environ Health Perspect 106 Suppl 2:701-705.
35
36      Eskenazi, B; Mocarelli, P; Warner, M; et al. (1998) Seveso women's health study: A study  of the effects of TCDD
37      on reproductive health. Orgaonhalogen compounds 38:219-222.
38
39      Esteller, M; Garcia, A; Matinez-Palones, JM; et al. (1997) Germ line polymorphisms in cytochrome P450IA1
40      (C4887 CYP IA1) and methylenetetrahydrofolate reductase (MTHFR) genes and endometrial cancer susceptibility.
41       Carcinogenesis 18:2307-2311.
42
43      Fernandez-Salguero, PM; Hilbert, DM; Rudikoff, S; et al. (1996) Aryl-hydrocarbon receptor-deficient mice are
44      resistant to 2,3,7,8-tetrachlorodibenzo-/>-dioxin-induced toxicity.  Toxicol Appl Pharmacol 140(1):173-179.
45
46      Fingerhut, MA; Halperin, WE;  Marlow, DA. (1991a) Cancer mortality in workers exposed to 2,3,7,8-
47      tetrachlorodibenzo-p-dioxin. New Engl J Med 324:212-218.
48
49       Fingerhut, MA; Halperin, WE;  Marlow, D; et al. (1991b) Mortality among United States workers employed in the
50      production of chemicals contaminated with 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Cincinnati, OH: U.S.
51       Department of Health and Human Services, National Institute for Occupational Safety and Health. NTIS# PB 91 -
52       125971.
53
       6/g/00                                           135        DRAFT—DO NOT CITE OR QUOTE

-------
  1       Flesch-Janys, D; Steindorf, K; Gum, P; et al. (1998) Estimation of the cumulated exposure to polychlorinated
  2       dibenzo-p-dioxins/furans and standardized mortality ratio analysis of cancer mortality by dose in an occupationally
  3       exposed cohort. Environ Health Perspect 106(supplement 2):655-662.
  4
  5       Flesch-Janys, D; Becher, J; Berger, J; et al. (1999) Epidemiological investigation of breast cancer incidence in a
  6       cohort of female workers with high exposure to PCDD/CDF and HCH. Organohalogen Compounds 44:379-382.
  7
  8       Flesch-Janys, D; Berger, J; Gum, P; et al. (1995) Exposure to polychlorinated dioxins and furans (PCDD/CDF) and
  9       mortality in a cohort of workers from a herbicide-producing plant in Hamburg, Federal Republic of Germany. Am J
10       Epidemiol 142:1165-1175.
11
12       Flodstrom, S; Ahlborg, UG. (1992) Relative tumor promoting activity of some polychlorinated dibenzo-p-dioxin-,
13       dibenzofuran-, and biphenyl congeners in female rats. Chemosphere 25:1(2):169-172.
14
15       Gaido,  KW; Maness, SC; Leonard, LS; et al. (1992) 2,3,7,8-Tetrachlorodibenzo-p-dioxin-dependent regulation of
16       transforming growth factors-a and P2 expression in a human keratinocyte cell line involves both transcriptional and
17       post-transcriptional control. J Biol Chem 267:24591 -24595.
18
19       Gasiewicz, TA. (1997) Dioxins and the AhR: probes to uncover processes in neuroendocrine development.
20       Neurotoxicology 18:393-414.
21
22       Gasiewicz, TA; Holscher, MA; Neal, RA. (1980) The effect of total parenteral nutrition on the toxicity of 2,3,7,8-
23       tetrachlorodibenzo-p-dioxin in the rat. Toxicol Appl Pharmacol 54:469-488.
24
25       Gierthy, JF; Bennett, JA; Bradley, LM; et al. (1993) Corrleation of in vitro and in vivo growth suppression of MCF-
26       7 human breast cancer by 2,3,7,8-tetrachlorodibenzo-p-dioxin. Cancer Res 53:3149-3153.
27
28       Gerhard, I; Runnebaum, B; (1992) Grenzen der hormonsubsittution bei schadstoffbelastung und
29       fertilitatsstorungen. ZentBl Gynekol 114:593-602.
30
31       Goldstein, JA; Hickman, P; Jue, DL. (1974) Experimental hepatic porphyria induced by polychlorinated biphenyls.
32       Toxicol Appl Pharmacol 27(2):437-448.
33
34       Goodman, DG; Sauer, RM. (1992) Hepatotoxicity and carcinogenicity in female Sprague-Dawley rats treated with
35       2,3,7,8-tetrachlorordibenzo-p-dioxin (TCDD): a Pathology Working Group reevaluation. Regul Toxicol Pharmacol
36       15:245-252.
37
38       Gorski, JR;  Rozman, K. (1987) Dose-response and time course of hypothyroxinemia and hypoinsulinemia and
39       characterization of insulin hypersensitivity in 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)-treated rats. Toxicology
40       44(3):297-307.
41
42       Gradin, K; McGuire, J; Wenger, RH; et al. (1996) Functional interference between hypoxia and dioxin signal
43       transduction pathways: competition for recruitment of the ARNT transcription factor. Mol Cell Biol
44       16(10):5221-5231.
45
46       Graham, MJ; Lucier, GW; Linko, P; et al. (1988) Increases in cytochrome P-450 mediated 17 beta-estradiol
47       2-hydroxylase activity in rat liver microsomes after both acute administration and subchronic  administration of
48       2,3,7,8-tetrachlorodibenzo-p-dioxin in a two-stage hepatocarcinogenesis model. Carcinogenesis 9:1935-1941.
49
50       Gray, LE, Jr.; Kelce, WR; Monosson, E; et al. (1995a) Exposure to TCDD during development permanantly alters
51       reproductive function in male Long Evans rats and hamsters: reduced ejaculated and epididymal sperm numbers and
52       sex accessory gland weights in offspring with normal androgenic status. Toxicol Appl Pharmacol  131:108-118.
53
         6/8/00                                           136        DRAFT—DO NOT CITE OR QUOTE

-------
 1      Gray, LE, Jr.; Ostby, J; Wolf, C; et al. (1995b) Functional developmental toxicity of low doses of 2,3,7,8-
 2      tetrachlorodibenzo-p-dioxin and a dioxin-like PCB (169) in Long Evans rats and Syrian hamsters: reproductive,
 3      behavioral and thermoregulatory alterations. Organohalogen Compounds 25:33-38.
 4
 5      Gray, LE, Jr.; Ostby, JS. (1995) In utero 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) alters reproductive
 6      morphology and function in female rat offspring. Toxicol Appl Pharmacol 133:285-294.
 7
 8      Gray, LE; Ostby, JS; Kelce, WR. (1997a). A dose-response analysis of the reproductive effects of a single
 9      gestational dose of 2,3,7,8-tetrachlorodibenzo-p-dioxin in male Long Evans Hooded rat offspring Toxicol Appl
10     Pharmacol 146(1):11-20.
11
12      Gray, LE; Wolf, C; Mann, P; Ostby, JS. (1997b). In utero exposure to low  doses of 2,3,7,8-tetrachlorodibenzo-p-
13      dioxin alters reproductive development of female Long Evans hooded rat offspring. Toxicol Appl Pharmacol
14      146(2): 237-44.
15
16     Grubbs, WD; Wolfe, WH; Michalek, JE; et al. (1995) Air Force Health Study: an epidemiologic investigation of
17     health effects in Air Force personnel following exposure to herbicides. Report number AL-TR-920107.
18
19     Gu, Yi-J; Hogenesch, JB; Bradfield, CA. (2000) The PAS Superfamily: Sensors of Environmental and
20     developmental signals. Annnu. Rev Pharmacol. Toxicol 40:519-561.
21
22     Gupta, BN; Vos JG, Moore; JA, Zinkl; et al. (1973) Pathologic effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin in
23     laboratory animals.  Environ Health Perspect 5:125-140.
24
25     Guzelian, PS. (1985) Clinical evaluation of liver structure and function in humans exposed to halogenated
26     hydrocarbons. Environ Health Perspect 60:159-164.
27
28     Hahn, ME. (1998) The aryl hydrocarbon receptor: a comparative perspective. Comp Biochem Physiol 121:23-53.
29
30      Halperin, W; Vogt, R; Sweeney, MH; et al. (1998) Immunological markers among workers exposed to 2,3,7,8-
31     tetrachlorodibenzo-p-dioxin. Occup Environ Med 55:742-749.
32
33     Hankinson, O. (1995) The aryl hydrocarbon receptor complex. Ann Rev Pharmacol Toxicol 35:307-340.
34
35     Hardell, L; Eriksson, M. (1988) The association between STSs and exposure to phenoxyacetic acids: a new case-
36     referent study. Cancer 62:652-656.
37
38     Hardell, L; Sandstrom, A.  (1979) Case-control study: soft-tissue sarcomas and exposure to phenoxyacetic acids or
39     chlorophenols. Br J Cancer 39:711-717.
40
41      Harper, N; Connor,  K; Steinberg, M; et al. (1994) An enzyme-linked immunosorbent assay (ELISA) specific for
42     antibodies to TNP-LPS detects alterations in serum immunoglobulins and isotype switching in C57BL/6 and DBA/2
43     mice exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin and related compounds. Toxicology 92:155-167.
44
45     Haseman, JK, Johnson, FM. (1996) Analysis of National Toxicology Program rodent bioassay data for
46     anticarcinogenic effects. Mutat Res  350(1):131-141.
47
48     Hatch, M. (1984) Reproductive effects of the dioxins. In: Public health risks of the dioxins. Lowrance, WW, ed.
49     California: William Kaufmann; pp.  255-275.
50
51      Hayes, CL; Spink, D; Spink, B; et al. (1996) 17-beta Estradiol hydroxylation catalyzed by human cytochrome P450
52     1B1. Proc Nat Acad Sci 93:9776-9781.
53
       6/8/00                                           137        DRAFT—DO NOT CITE OR QUOTE

-------
  1       Hebert, CD; Harris, MW; Elwell, MR; et al. (1990) Relative toxicity and tumor-promoting ability of
  2       2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), 2,3,4,7,8-pentachlorodibenzofiiran (PCDF), and
  3       1,2,3,4,7,8-hexachlorodibenzofuran (HCDF) in hairless mice. Toxicol Appl Pharmacol 102:362-377.
  4
  5       Hemming, H; Bager, Y; Flodstrom, S; et al. (1995) Liver tumour promoting activity of 3,4,5,3',4'-pentachloro-
  6       biphenyl and its interaction with 2,3,7,8-tetrachlorodibenzo-p-dioxin. Eur J Pharmacol 292:241-249.
  7
  8       Hertzman, C; Teschke, K; Ostry, A; et al. (1997) Mortality and cancer incidence among sawmill workers exposed to
  9       chlorophenate wood preservatives. Am JPubl Health 87(1 ):71-79.
10
11       Hill, AB. (1965) The environment and disease: association or causation. Proc R Soc Med 58:295-300.
12
13       Hooiveld, M; Heederik, D; Bueno de Mesquita, HB. (1996) Preliminary results of the second follow-up of a Dutch
14       cohort occupationally exposed to phenoxy herbicides, chlorophenols, and contaminants. Organohalogen
15       Compounds 30:185-189.
16
17       Hooiveld, M; Heederik, DJJ; Kogevinas, M; et al. (1998) Second follow-up of a Dutch cohort occupationally
18       exposed to phenoxy herbicides, chlorophenols, and contaminants. Am J Epidemiol  147(9):891 -901.
19
20       Hornung, MW; Spitsbergen, JM; Peterson, RE. (1999) 2,3,7,8-Tetrachlorodibenzo-p-dioxin alters cardiovascular
21       and craniofacial development and function in sac fry of rainbow trout (Oncorhynchus mykiss). Toxicol Sci
22       47(1):40-51.
23
24       Huff, JE; Salmon, AG; Hooper, NK; et al. (1991) Long-term carcinogenesis studies on
25       2,3,7,8-tetrachlorodibenzo-p-dioxin and hexachlorodibenzo-p-dioxins. Cell Biol Toxicol 7(l):67-94.
26
27       Huisman, M; Koopman-Esseboom, C; Lanting, CI; et al. (1995a) Neurological condition in 18-month-old children
28       perinatally exposed to polychlorinated biphenyls and dioxins. Early Hum Dev 43:165-176.
29
30       Huisman, M; Koopman-Esseboom, C; Fidler, V; et al. (1995b) Perinatal exposure to polychlorinated biphenyls and
31       dioxins and its effect on neonatal neurological development. Early Hum Dev 41 (2): 111 -127.
32
33       Hurst CH, DeVito MJ, Setzer RW, Birnbaum LS. (2000) Acute administration of 2,3,7,8-tetrachlorodibenzo-
34       p-dioxin (TCDD) in pregnant Long Evans rats: association of measured tissue concentrations with developmental
35       effects. Toxicol Sci 53(2):411-20.
36
37       IARC. (1997) IARC monographs on the evaluation of carcinogenic risks to humans. Volume 69. Polychlorinated
38       dibenzo-para-dioxins and polychlorinated dibenzofurans. Lyon, France: IARC.
39
40       Johnson, RD; Tietge, JE; Botts, S. (1992) Carcinogenicity of 2,3,7,8-TCDD to Medaka. The Toxicologist
41       12(1):138.
42
43       Johnson, L; Wilker, CE; Safe, SH; et al. (1994) 2,3,7,8-tetrachlorodibenzo-p-dioxin reduces the number, size, and
44       organelle content of Leydig cells in adult rat testes. Toxicology 89:49-65.
45
46       Johnson, KL; Cummings, AM; Birnbaum LS. (1997) Promotion of endometriosis in mice by polychlorinated
47       dibenzo-p-dioxins, dibenzofurans, and biphenyls. Environ Health Perspect 105(7):750-755.
48
49       Jung, D; Berg, PA; Edler, L; et al. (1998) Immunologic findings in workers formerly exposed to 2,3,7,8-
50       tetrachlorodibenzo-p-dioxin and its congeners. Environ Health Perspect 106 Suppl 2:689-695.
51
52       Jusko, WJ. (1995) Pharmacokinetics and receptor-mediated pharmacodynamics of corticosteroids. Toxicology
53       102:189-196.
54
         6/8/00                                           138        DRAFT—DO NOT CITE OR QUOTE

-------
 1      Kadlubar, FF; Butler, MA; Kaderlik, RK; et al. (1992) Polymorphisms for aromatic amine metabolism in humans:
 2      relevance for human carcinogenesis. Environ Health Perspect 98:69-74.
 3
 4      Kawajari, K; Nakachi, K; Imai, K; et al. (1993) Germ line polymorphisms of p53 and CYPIA1 genes involved in
 5      human lung cancer. Carcinogenesis  14(6):1085-1089.
 6
 7      Kayajanian, GM. (1997) Dioxin is a promoter blocker, a promoter, and a net anticarcinogen. Regul Toxicol
 8      Pharmacol 26(1):134-137 (Review).
 9
10     Kayajanian, GM. (1999) Dioxin is a systemic promoter blocker, II. Ecotoxicol Environ Saf 42(2): 103-109.
11
12     Ketchum, NS; Michalek, JE; Burton JE. (1999) Serum dioxin and cancer in veterans of Operation Ranch Hand Am
13     J Epidemiol 149(7):630-639.
14
15     Kimmel, GL. (1988) Appendix C. In: A cancer risk-specific dose estimate for 2,3,7,8,-TCDD. U.S. EPA, External
16     Review Draft.
17
18     Kitchin, KT; Woods, JS. (1979) 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) effects on hepatic microsomal
19     cytochrome P-448-mediated enzyme activities. Toxicol Appl Pharmacol 47:537-546.
20
21      Kleeman, JM; Moore, RW; Peterson, RE. (1990) Inhibition of testicular steroidogenesis in 2,3,7,8-
22     tetrachlorodibenzo-p-dioxin-treated rats: evidence that the key lesion occurs prior to or during pregnenolone
23     formation. Toxicol Appl Pharmacol 106:112-125.
24
25     Kociba, RJ; Keeler, PA; Park, GN; et al. (1976) 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD): results of a 13 week
26     oral toxicity study in rats. Toxicol Appl Pharmacol 35:553-574.
27
28     Kociba, RJ; Keyes, DG; Beyer, JE; et al. (1978) Results of a two-year chronic toxicity and oncogenicity study of
29     2,3,7,8-tetrachlorodibenzo-p-dioxin in rats. Toxicol Appl Pharmacol 46:279-303.
30
31     Kogevinas, M; Saracci, R; Winkelmann, R; et al. (1993) Cancer incidence and mortality in women occupationally
32     exposed to chlorophenoxy herbicides, chlorophenols and dioxins. Cancer Causes Control 4:547.
33
34     Kogevinas, M; Becher, H; Benn, T;  et al. (1997)  Cancer mortality in workers exposed to phenoxy herbicides,
35     chlorophenols, and dioxin. An expanded and updated international cohort study. Am J Epidemiol 145(12):1061-
36      1075.
37
38     Kohn, MC; Lucier, GW; Clark, GC; et al. (1993) A mechanistic model of effects of dioxin on gene expression in
39     the rat liver. Toxicol Appl Pharmacol 120:138-154.
40
41      Koninckx, PR; Braet, P; Kennedy, SH; et al. (1994) Dioxin pollution and endometriosis in Belgium. Hum Reprod
42     9(6): 1001-1002.
43
44     Koopman-Esseboom, C; Weisglas-Kuperus, N; de Ridder, MAJ; et al. (1995b) Effects of PCB/dioxin exposure and
45     feeding type on the infant's visual recognition memory. Chapter 7 in dissertation entitled: Effects of perinatal
46     exposure to PCBs and dioxins on early human development. Erasmus Universiteit Rotterdam, pp. 107-121.
47
48     Koopman-Esseboom, C; Weisglas-Kuperus, N; de Ridder, MAJ; et al. (1996) Effects of polychlorinated
49     biphenyl/dioxin exposure and feeding type on the infant's mental and psychomotor development. Pediatrics 97:700-
50     706.
51
52     Koopman-Esseboom, C; Huisman, M; Weisglas-Kuperus, N; et al. (1994a) Dioxin and PCB levels in blood and
53     human milk in relation to living areas in The Netherlands. Chemosphere 29(9-11):2327-2338.
54


       6/8/00                                           139        DRAFT—DO NOT CITE OR QUOTE

-------
  1       Koopman-Esseboom, C; Morse, DC; Weisglas-Kuperus, N; et al. (1994c) Effects of dioxins and polychlorinated
  2       biphenyls on thyroid hormone status of pregnant women and their infants. Pediatr Res 36(4):468-73.
  3
  4       Koopman-Esseboom, C; Huisman, M; Touwen, BCL; et al. (1995a) Effects of PCB/dioxin exposure and feeding
  5       type on the infant's visual recognition memory. Chapter 5 in dissertation entitled: Effects of perinatal exposure to
  6       PCBs and dioxins on early human development. Erasmus Universiteit Rotterdam, pp. 75-86.
  7
  8       Koopman-Esseboom, C; Huisman, M; Weisglas-Kuperus, N; et al. (1994b) PCB and dioxin  levels in plasma and
  9       human milk of 418 Dutch women and their infants. Predictive value of PCB congener levels in maternal plasma for
10       fetal and infant's exposure to PCBs and dioxins. Chemosphere 28:1721-1732.
11
1 2       Kuratsune, M; Ikeda, M; Nakamura, Y; et al. (1988) A cohort study on mortality of Yusho patients: a preliminary
1 3       report. In:  Unusual occurrences as clues to cancer etiology. Miller, RW; et al., eds. Jpn Sci Soc Press: Tokyo/Taylor
14       & Francis, Ltd., pp. 61-68.
15
16       Kuratsune, M. (1989) Yusho, with reference to Yu-Cheng. In: Halogenated biophenyls, terphenyls, naphthalenes,
17       dibenzodioxins and related products. Kimbrough, RD; Jensen, AA, eds. 2nd ed. New York: Elsevier Science
18       Publishers; pp. 381-400.
19
20       Kutz, FW; Barnes, DG; Bretthauer, EW; et al. (1990) The International Toxicity Equivalency Factor (I-TEF)
21       method for estimating risks associated with exposures to complex mixtures of dioxins and related compounds.
22       Toxicol Environ Chem 26:99-109.
23
24       Lahvis, GP; Bradfield, CA; (1998) Ahr null alleles: distinctive or different? Biochem Pharmacol 56(7):781-787.
25
26       Lampi, P; Hakulinen, T; Luostarinen, T; et al. (1992) Cancer incidence following chlorophenol exposure in a
27       community in southern Finland. Arch Environ Health 47(3): 167-175.
28
29       Landi, MT; Consonni, D; Patterson, DG, Jr.; et al. (1998) 2,3,7,8-Tetrachlorodibenzo-p-dioxin plasma levels in
30       Seveso 20  years after the accident.  Environ Health Perspect 106(5):273-277.
31
32       Lathrop, GD; Wolfe, WH; Albanese, RA; et al. (1984) An epidemiologic investigation of health effects in Air Force
33       personnel following exposure to herbicides. Baseline morbidity study results. Brooks Air Force Base, TX: U.S. Air
34       Force School of Aerospace Medicine, Aerospace Medical Division (unpublished).
35
36       Lathrop, GD; Wolfe, WH; Michalek, JE; et al. (1987) An epidemiologic investigation of health effects in Air Force
37       personnel following exposure to herbicides. First follow-up examination results, January 1985-September 1987.
38       Brooks Air Force Base, TX: U.S. Air Force School of Aerospace Medicine, Aerospace Medical  Division
39       (unpublished).
40
41       Lebel, G; Dodin, S; Ayotte, P; et al. (1998) Organochlorine exposure  and the risk of endometriosis. Fertil Steril
42       69(2):221-228.
43
44       Li, X; Johnson, DC; Rozman, KK. (1995a) Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on estrous
45       cyclicity and ovulation in female Sprague-Dawley rats. Toxicol Lett 78:219-222.
46
47       Li, X; Johnson, DC; Rozman, KK. (1995b) Reproductive effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in
48       female rats: ovulation, hormonal regulation and possible mechanism(s). Toxicol Appl Pharmacol 133:321-327
49
50       Limbird, LE; Taylor. P. (1998) Endocrine disrupters signal the need for receptor models and mechanisms to inform
51       policy Cell 93:157-163.
52
53       Longnecker, MP; Michalek, JE. (2000) Serum dioxin level in relation to diabetes mellitus  among Air Force veterans
54       with background levels of exposure. Epidemiology 11:44-48
55

         6/8/00                                           140         DRAFT—DO NOT CITE OR QUOTE

-------
 1       Liu, H; Biegel, L; Narasimhan, TR; et al. (1992)  Inhibition of insulin-like growth factor-I responses in MCF-7 cells
 2      by 2,3,7,8-tetrachlorodibenzo-p-dioxin and related compounds. Mol Cell Endocrinol 87(1-3): 19-28
 3
 4      Lu, YC; Wong, PN. (1984) Dermatological, medical, and laboratory findings of patients in Taiwan and their
 5      treatments. Am JIndMed 5:81-115.
 6
 7      Lucier, GW; Tritscher, A; Goldsworthy, T; et al. (1991) Ovarian hormones enhance TCDD-mediated increases in
 8      cell proliferation and preneoplastic foci in a two stage model for rat hepatocarcinogenesis. Cancer Res 51 • 1391 -
 9      1397.
10
11      Lucier, GW; Lui, EMK; Lamartiniere, CA. (1979) Metabolic activation/deactivation reactions during perinatal
12      development. Environ Health Perspect 29:7-16.
13
14      Lynge, E. (1998) Cancer incidence in Danish phenoxy herbicide workers,  1947-1993. Environ Health Perspect
15      106(Supplement 2): 683-688.
16
17      Mably, TA; Moore, RW; Goy, RW; et al. (1992b) In utero and lactational exposure of male rats to 2,3,7,8-
18      tetrachlorodibenzo-p-dioxin: 2. Effects on sexual behavior and the regulation of luteinizing hormone secretion in
19      adulthood. Toxicol Appl Pharmacol 114:108-117.
20
21      Mably, TA; Bjerke, DL; Moore, RW; et al. (1992c) In utero and lactational exposure of male rats to 2,3,7,8-
22      tetrachlorodibenzo-p-dioxin: 3. Effects on spermatogenesis and reproductive capability. Toxicol Appl Pharmacol
23      114:118-126.
24
25      Mably, TA; Moore, RW; Peterson, RE. (1992a) In utero and lactational exposure of male rats to 2,3,7,8-
26      tetrachlorodibenzo-p-dioxin: 1. Effects on androgenic status. Toxicol Appl Pharmacol 114:97-107.
27
28      Manz, A; Berger, J; Dwyer, JH; et al. (1991) Cancer mortality among workers in chemical plant contaminated with
29      dioxin. Lancet 338:959-964.
30
31      Maronpot, RR; Foley, JF; Takahashi, K; et al. (1993) Dose-response for TCDD promotion of hepatocarcinogenesis
32      in rats initiated with DEN: histologic, biochemical, and cell proliferation endpoints. Environ Health Perspect
33      101:634-642.
34
35      Martin, 3V. (1984) Lipid abnormalities in workers exposed to dioxin. Br J Ind Med 41:254-256.
36
37      Matsumura, F. (1994) How important is the protein phosphorylation pathway in the toxic expression of dioxin-type
38      chemicals? Biochem Pharmacol 48(2):215-224.
39
40      Matzke GR; Frye,  RF; Early JJ; Straka RJ; Carson SW.  (2000)  Evaluation of the influence of diabetes mellitus on
41      antipyrine metabolism and CYP1A2 and CYP2D6 activity. Pharmacotherapy. 20(2): 182-90.
42
43      May, G. (1982) Tetrachlorodibenzodioxin: a survey of subjects ten years after exposure. Br J Ind  Med 39:128-135.
44
45      Mayani, A; Barel,  S; Soback, S; et al. (1997) Dioxin concentrations in women with endometriosis. Hum Reprod
46      12:373-375.
47
48      McConnell, EE; Moore, JA; Haseman, JK; et al. (1978) The comparative toxicity of chlorinated dibenzo-p-dioxins
49      in mice and guinea pigs. Toxicol Appl Pharmacol 44:335-356.
50
51      McNulty, WP. (1977) Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin for rhesus monkeys: brief report. Bull
52      Environ Contam Toxicol 18:108-109.
53
        6/8/00                                            141        DRAFT—DO NOT CITE OR QUOTE

-------
  1       Mebus, CA; Reddy, VR; Piper, WN. (1987) Depression of rat testicular 17-hydroxylase and 17,20-lyase after
  2      administration of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Biochem Pharmacol 36(5):1727-1731.
  3
  4      Michalek, JE; Akhtar, FZ; Kiel, JL. (1999) Serum dioxin, insulin, fasting glucose, and sex hormone-binding
  5      globulin in veterans of Operation  Ranch Hand. J Clin Endocrinol Metab (5): 1540-1543.
  6
  7      Michalek JE; Ketchum NS; Check IJ. (1999) Serum dioxin and immunologic response in veterans of Operation
  8      Ranch Hand. Am J Epidemiol 149:1038-1046.
  9
 10      Michalek, JE; Ketchum, NS; Akhtar, FZ. (1998) Postservice mortality of United States Air Force veterans
 11       occupationally exposed to herbicides in Vietnam: 15-year follow-up.  Am J Epidemiol 148(8):786-792.
 12
 13      Mocarelli, P; Needham, LL; Marocchi, A; et al. (1991) Serum concentrations of 2,3,7,8-tetrachlorodibenzo-p-dioxin
 14      and test results from selected residents of Seveso, Italy. J Toxicol Environ Health 32:357-366.
 15
 16      Mocarelli P; Brambilla P; Gerthoux, PM; et al. (1996) Change in sex ratio with exposure to dioxin [letter]. Lancet
 17      348:409.
 18
 19      Mocarelli, P; Gerthoux, PM; Ferrari, E; et al. (2000) Paternal concentartions of dioxin and sex ratio of offspring.
 20      Lancet, 355:1858-1863.
 21
 22      Mocarelli, P; Marocchi, A; Brambilla, P; et al. (1986) Clinical laboratory manifestations of exposure to dioxin in
 23      children. A six year study of the effects of an environmental disaster near Seveso, Italy. JAMA 256:2687-2695.
 24
 25      Moore, RW; Peterson, RE. (1988) Androgen catabolism and excretion in 2,3,7,8-tetrachlorodibenzo-p-dioxin-
 26      treated rats. Biochem Pharmacol 37:560-562.
 27
 28      Moore, R W; Parsons, J A; Bookstaff, R C; Peterson, RE. (1989) Plasma concentrations of pituitary hormones in
 29      2,3,7,8-tetrachlorodibenzo-p-dioxin-treated male rats. J Biochem Toxicol 4:165-172.
 30
 31       Moore, RW; Bookstaff, RC; Mably, RA; et al. (1991) Differential effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on
 32      responsiveness of male rats to androgens, 17B-estradiol, luteinizing hormone, gonadotropin releasing hormone, and
 33      progesterone. Presented at: Dioxin '91, 11th international symposium on chlorinated dioxins and related
 34      compounds; Research Triangle Park, NC.
 35
 36       Moore, RW; Potter, CL; Theobald, HM; et al. (1985) Androgenic deficiency in male rats treated with 2,3,7,8-
 37       tetrachlorodibenzo-p-dioxin. Toxicol Appl Pharmacol 79:99-111.
 38
 39       Moses, M; Lilis, R; Crow, KD; et al. (1984) Health status of workers with past exposure to
 40      2,3,7,8-tetrachlorodibenzo-p-dioxin in the manufacture of 2,4,5-trichlorophenoxyacetic acid. Comparison of
 41       findings with and without chloracne. Am J Ind Med 5:161-182.
 42
 43       Murray, F J; Smith, F A; Nitschke, K D; Humiston, CG; Kociba, RJ; Schwetz, BA. (1979) Three-generation
 44       reproduction study of rats given 2,3,7,8-tetrachlorodibenzo-/j-dioxin (TCDD) in the diet. Toxicol Appl Pharmacol
 45       50:241-252.
 46
 47       Nagayama, J; Okamura, K; lida, T; et al.  (1998) Postnatal exposure to chlorinated dioxins and related chemicals on
 48       thyroid hormone status in Japanese breast-fed infants. Chemosphere 37(9-12):1789-1793.
49
 50       Nagel, S; Berger, J; Flesch-Janys, D; et al. (1994) Mortality and cancer mortality in a cohort of female workers of a
 51       herbicide producing plant exposed to polychlorinated dibenzo-p-dioxins and furans. Inform. BiometEpidemiol.
52       Med. Biol.,25:32-38.
53
         6/8/00                                            142        DRAFT—DO NOT CITE OR QUOTE

-------
 1      Narasimhan, TR; Craig, A; Arellano, L; et al. (1994) Relative sensitivities of 2,3,7,8-tetrachlorodibenzo-p-dioxin-
 2      induced Cypla-1 and Cypla-2 gene expression and immunotoxicity in female B6C3F1 mice  Fundam Appl Toxicol
 3      23:598-607.
 4
 5      NAS/NRC (Naitonal Academy of Sciences/National Research Council). (1983) Risk assessment in the Federal
 6      Government. Washington, DC: National Academy Press.
 7
 8      NAS/NRC. (1994) Science and Judgment in Risk Assessment. Washington, DC: National Academy Press
 9
10      Needham, LL; Gerthoux, PM; Patterson, DG; et al. (1999) Exposure Assessment: Serum Levels of TCDD in
11      Seveso, Italy. Envir Res (A) 80:S200-S206.
12
13      Neuberger, M; Landvoigt, W; Demt, F. (1991) Blood levels of 2,3,7,8-tetrachlorodibenzo-p-dioxin in chemical
14      workers after chloracne and in comparison groups. Int Arch Occup Environ Health 63:325-327.
15
16      Neubert,  R; Golor, G; Stahlmann, R; Helge, H; Neubert, D. (1992) Polyhalogenated dibenzo-p-dioxins and
17      dibenzofurans and the immune system.  4. Effects of multiple-dose treatment with
18      2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on peripheral lymphocyte subpopulations of a non-human primate
19      (Callithrixjacchus). Arch Toxicol 66:250-259.
20
21      NTP (National Toxicology Program). (1980) Bioassay of a mixture of 1,2,3,6,7,8-hexachlorodibenzo-p-dioxin and
22      1,2,3,7,8,9-hexachlorodibenzo-p-dioxin for possible carcinogenicity (gavage study). Tech. Rept. Ser. No. 198.
23      Research Triangle Park, NC: U.S. DHHS, PHS.
24
25      NTP. (1982a) Bioassay of 2,3,7,8-tetrachlorodibenzo-p-dioxin for possible carcinogenicity (gavage study). Tech.
26      Rept. Ser. No. 201. Research Triangle Park, NC: U.S. DHHS, PHS.
27
28      NTP. (1982b) Bioassay of 2,3,7,8-tetrachlorodibenzo-p-dioxin for possible carcinogenicity (dermal study). Tech.
29      Rept. Ser. No. 201. Research Triangle Park, NC: U.S. DHHS, PHS.
30
31      NTP. (2000) Report on carcinogens, ninth ed: carcinogen profiles 2000. U.S. Department of Health and Human
32      Services, Public Health Service, Research Triangle Park, NC.
33
34      Olsen, H; Enan, E; Matsumura, F. (1994) Regulation of glucose transport in the NIH 3T3 LI preadipocyte cell line
35      by TCDD. Environ Health Perspect 102(5):454-458.
36
37      Olson, JR; Holscher, MA; Neal, RA. (1980) Toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in the Golden Syrian
38      hamster. Toxicol Appl Pharmacol 55:67-78.
39
40      Olson, JR; McGarrigle, BP. (1990) Characterization of the developmental toxicity of 2,3,7,8-TCDD in the Golden
41      Syrian hamster.  Toxicologist 10:313.
42
43      Ott, MG; Zober, A. (1996b) Cause specific mortality and cancer incidence among employees exposed to 2,3,7,8-
44      TCDD  after a 1953 reactor accident.  Occup Environ Med 53:606-612.
45
46      Ott, MG; Zober, A. (1996a) Morbidity study of extruder personnel with potential exposure to brominated dioxins
47      andfurans 2  Results of clinical laboratory studies.  Occup Environ Med 53:844-846.
48
49      Ott, MG; Messerer, P; Zober, A. (1993) Assessment of past occupational exposure to 2,3,7,8-tetrachIorodibenzo-
50      p-dioxin using blood lipid analyses. Int Arch Occup Environ Health 65:1-8.
51
52      Ott, MG; Zober, A; Germann, C. (1994) Laboratory results for selected target organs in 138 individuals
53      occupationally exposed to TCDD. Chemosphere 29:2423-2437.
54


        6/8/00                                           143       DRAFT—DO NOT CITE OR QUOTE

-------
  1       Ouiddir, A; Planes, C; Femandes, I; et al. (1999)  Hypoxia upregulates activity and expression of the glucose
  2      transporter GLUT1 in alveolar epithelial cells. Am J Respir Cell Mol Biol (6):710-718.
  3
  4      Park, J-YK; Shigenaga, MK; Ames, BN. (1996) Induction of cytochrome P4501AI by 2,3,7,8-tetrachlorodibenzo-p-
  5      dioxin or indolo(3,2-b) carbazole is associated with oxidative DNA damage. Proc Nat Acad Sci 93:2322-2327.
  6
  7      Patandin, S; Koopman-Esseboom, C; de Ridder, MA; et al. (1998) Pediatr Res 44(4):538-545.
  8
  9      Patandin, S; Lanting, CI; Mulder, PG; et al. (1999) Effects of environmental exposure to poly chlorinated biphenyls
 10      and dioxins on cognitive abilities in Dutch children at 42 months of age. J Pediatr 134(1):33-41.
 11
 12      Pauwels, A; Cenijn, P; Covaci, A; et al. (1999) Analysis of PCB congeners (by GC-ECD) and dioxin-like toxic
 13      equivalence (by CALUX assay)  in females with endometriosis and other fertility problems.  Organohalogen
 14      Compounds 44:408-412.
 15
 16      Pazderova-Vejlupkova, J; Nemcova, M; Pickova, J; et al. (1981) The development and prognosis of chronic
 17      intoxication by tetrachlorodibenzo-p-dioxin in man. Arch Environ Health 36:5-11.
 18
 19      Pesatori, AC; Zocchetti, C; Guercilena, S; et al. (1998)  Dioxin exposure and non-malignant health effects: a
 20      mortality study. Occup Environ Med 55(2):126-131.
 21
 22      Pesatori, AC; Tironi, A; Consonni, A; et al. (1999) Cancer incidence in the Seveso population, 1977-1991.
 23      Organohalogen Compounds 44:411-412.
 24
 25      Peterson, RE; Theobald, HM; Kimmel, GL. (1993) Developmental and reproductive toxicity of dioxins and related
 26      compounds: cross-species comparisons. Crit Rev Toxicol 23 (3):283-335.
 27
 28      Pluim, HJ; Koppe, JG; Olie, K; et al. (1992) Effects of dioxins on thyroid function in newborn babies. Letter to the
 29      editor. Lancet 339:1303.
 30
 31       Pluim, HJ; de Vijlder, JJM; Olie, K; et al. (1993) Effects of pre- and postnatal exposure to chlorinated dioxins and
 32      furans on human neonatal thyroid hormone concentrations. Environ Health Perspect 101(6):504-508.
 33
 34      Pluim, HJ; Koppe, JG; Olie, K; et al. (1994) Clinical laboratory manifestations of exposure to background levels of
 35      dioxins in the perinatal period. Acta Paediatr 83(6):583-587.
 36
 37      Pohjanvirta, R; Tuomisto, J. (1994)  Short-term toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in laboratory
 38      animals: effects, mechanisms, and animal models. Pharmacol Rev 46(4):483-549.
 39
 40      Poland, AD.  (1996)  Meeting report. Receptor-acting xenobiotics and their risk assessment. Drug Metab Disp
 41       24:1385-1388.
 42
 43      Poland, AD; Knutson, JC. (1982) 2,3,7,8-Tetrachlorodibenzo-p-dioxin and related halogenated aromatic
 44      hydrocarbons: examination of the mechanism of toxicity. Ann Rev Pharmacol Toxicol 22:517-554.
 45
 46      Poland, AD; Palen, D; Glover, E. (1982) Tumor promotion by TCDD in skin of HRS/J mice. Nature
 47      300(5889):271-273.
 48
 49      Portier, CJ; Kohn, MC. (1996) A biologically-based model for the carcinogenic effects of 2,3,7,8-TCDD in female
 50      Sprague-Dawley rats. Organohalogen Compounds 29:222-227.
 51
 52       Portier, C; Hoel, D; van Ryzin, J. (1984) Statistical analysis of the carcinogenesis bioassay data relating to the risks
 53       from exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. In:  Public health risks of the dioxins.  Lowrance, W, ed. Los
 54       Altos, NM: W.  Kaufmann, pp. 99-120.
55


         6/8/00                                           144        DRAFT—DO NOT CITE OR QUOTE

-------
 1       Portier, CJ; Sherman, CD; Kohn, M; et al. (1996) Modeling the number and size of hepatic focal lesions following
 2      exposure to 2,3,7,8-TCDD. Toxicol Appl Pharmacol 138:20-30.
 3
 4      Puga, A; Barnes, SJ; Dalton, TP; et al. (2000a)  Aromatic hydrocarbon receptor interaction with the retinoblastoma
 5      protein potentiates repression of E2F-dependent transcription and cell cycle arrest.  J Biol Chem 275(4)-2943-2950
 6
 7      Puga, A; Barnes, SJ; Chang, C; et al. (2000b) Activation of transcription factors activator protein-1 and nuclear
 8      factor-kappaB by 2,3,7,8-tetrachlorodibenzo-/>-dioxin. Biochem Pharmacol 59(8):997-1005.
 9
10      Rao, MS; Subbarao, V; Prasad, JD; et al. (1988) Carcinogenicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin in the
11      Syrian golden hamster. Carcinogenesis 9(9):1677-1679.
12
13      Rhile, MJ; Nagarkatti, M; Nagarkatti, PS. (1996) Role of Fas apoptosis and MHC genes in 2,3,7,8-
14      tetrachlorodibenzo-p-dioxin (TCDD)-induced immunotoxicity of T cells. Toxicology 110:153-167
15
16      Rier, SE; Martin, DC; Bowman, RE; et al. (1993) Endometriosis in rhesus monkeys (Macaca mulatta) following
17      chronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Fundam Appl Toxicol 21 (4):433-441.
18
19      Roegner, RH; Grubbs, WD; Lustik, MB; et al. (1991) Air Force Health Study: an epidemiologic investigation of
20      health effects in Air Force personnel following exposure to herbicides. Serum dioxin analysis of 1987 examination
21      results. NTIS# AD  A-237-516 through AD A-237-524.
22
23      Rogan, W. (1989) Yu-Cheng. In: Halogenated biphenyls, terphenyls, naphthalenes, dibenzodioxins and related
24      products. Kimbrough, RD; Jensen, AA, eds. 2nd ed. New York: Elsevier Pub.; pp. 401-415.
25
26      Rogan, WJ; Gladen, BC; Hung, K-L; et al. (1988) Congenital poisoning by polychlorinated biphenyls and their
27      contaminants in Taiwan. Science 241:334-338.
28
29      Roman, BL; Sommer, RJ; Shinomiya, K; et al. (1995). In utero and lactational exposure of the male rat to 2,3,7,8-
30      tetrachlorodibenzo-p-dioxin: Impaired prostate growth and development without inhibited androgen production.
31      Toxicol Appl Pharmacol 134:241-250.
32
33      Romkes, N; Safe, S. (1988) Comparative activities of 2,3,7,8-tetrachlorodibenzo-p-dioxin and progesterone as
34      antiestrogens in the female rat uterus. Toxicol Appl Pharmacol 92:368-380.
35
36      Romkes, N; Piskorska-Pliszynska, J; Safe, S. (1987) Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on hepatic and
37      uterine estrogen receptor levels in rats. Toxicol Appl Pharmacol 87:306-314.
38
39      Rowlands, JC; Gustafsson, J-A. (1997) Aryl hydrocarbon receptor-mediated signal transduction. Crit  Rev Toxicol
40      27:109-134.
41
42      Roy, D; Bernhardt, A; Strobel, HW; et al. (1992) Catalysis of the oxidation of steroid and stilbene estrogens to
43      estrogen quinone metabolites by the beta-naphthoflavone-inducible cytochrome P450 IA family. Arch Biochem
44      Biophys 296:450-456.
45
46      Rozman, KK; Lebofsky, M; Pinson, DM. (2000) Anemia and lung cancer in 1,2,3,4,6,7,8-heptachlorodibenzo-p-
47      dioxin (HPCDD)-treated female Sprague-Dawley rats after various single and multiple oral doses. Toxicol Sci
48      54(1):277.
49
50      Ryan, RP; Sunahara, GI; Lucier, GW;  et al. (1989) Decreased ligand binding to the hepatic glucocorticoid and
51      epidermal' growth factor receptors after 2,3,4,7,8-pentachlorodibenzofuran and 1,2,3,4,7,8-hexachlorodibenzofuran
52      treatment of pregnant mice. Toxicol Appl Pharmacol 98(3):454-464.
53
        6/8/00                                            145         DRAFT—DO NOT CITE OR QUOTE

-------
  1       Safe, S. (1995) Modulation of gene expression and endocrine response pathways by 2,3,7,8-tetrachlorodibenzo-p-
  2      dioxin and related compounds. Pharmacol Ther 67(2):247-281.
  3
  4      Saracci, R; Kogevinas, M; Bertazzi, P; et al. (1991) Cancer mortality in workers exposed to chlorophenoxy
  5      herbicides and chlorophenols. Lancet 38(3774): 1027-1032.
  6
  7      Schantz, SL; Bowman, RE. (1989) Learning in monkeys exposed perinatally to 2,3,7,8-tetrachlorodibenzo-p-dioxin
  8      (TCDD). Neurotoxicol Teratol 11:13-19.
  9
 10      Schantz, SL; Barsotti, DA; Allen, JR. (1979) Toxicological effects produced in nonhuman primates chronically
 11       exposed to fifty parts per trillion 2,3,7,8- tetrachlorodibenzo-p-dioxin (TCDD). Toxicol Appl Pharmacol 48(Part 2):
 12      A180.
 13
 14      Schecter, A, ed. (1994) Dioxins and health. New York: Plenum Press.
 15
 16      Schmidt, JV; Bradfield, CA. (1996) AhR signaling pathways. Ann Rev Cell Dev Biol 12:55-89.
 17
 18      Schrenk, D; Buchmann, A; Dietz, K; et al. (1994) Promotion of preneoplastic foci in rat liver with
 19      2,3,7,8-tetrachlorodibenzo-p-dioxin,  1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin and a defined mixture of 49
 20      polychlorinated dibenzo-p-dioxins. Carcinogenesis 15:509-515.
 21
 22      Sewall, CH; Lucier, GW. (1995) Receptor-mediated events and the evaluation of the Environmental Protection
 23      Agency (EPA) of dioxin risks. Mutat Res 333(1-2):! 11-122 (Review).
 24
 25      Sewall, CH; Lucier, GW; Tritscher, AM; et al. (1993) TCDD-mediated changes in hepatic epidermal growth factor
 26      receptor may be a critical event in the hepatocarcinogenic action of TCDD. Carcinogenesis 14:1885-1893.
 27
 28      Shimizu, Y; Nakatsuru, Y; Ichinose, M;  et al. (2000) Benzo[a]pyrene carcinogenicity is lost in mice lacking the aryl
 29      hydrocarbon receptor. Proc Natl Acad Sci USA 97:779-782.
 30
 31       Slezak, BP; Hatch, GE; DeVito, MJ; et al. (2000) Oxidative stress in female B6C3F1 mice following acute and
 32      subchronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Toxicol Sci, in press.
 33
 34      Smialowicz, RJ; Riddle, MM; Williams, WC; et al. (1994) Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)
 35       on humoral immunity and lymphocyte subpopulations: differences between mice and rats. Toxicol Appl Pharmacol
 36       124:248-256.
 37
 38       Spink, DC; Lincoln, DW, II; Dickerman, HW; et al. (1990) 2,3,7,8-Tetrachlorodibenzo-p-dioxin causes an
 39       extensive alteration of 17p-estradiol metabolism in MCF-7 breast tumor cells. Proc Natl Acad Sci USA 87:6917-
 40      6921.
 41
 42       Squire, RA. (1980) Pathologic evaluations of selected tissues from the Dow Chemical TCDD and 2,4,5-T rat
 43       studies. Submitted to Carcinogen Assessment Group, U.S. Environmental Protection Agency on August 15 under
 44      contract no. 68-01-5092.
 45
 46       Steenland, K; Piacitelli, L;  Deddens, J; et al. (1999) Cancer, heart disease, and diabetes in workers exposed to
 47       2,3,7,8-tetrachlorodibenzo-p-dioxin. J Natl Cancer Inst 91(9):779-786.
 48
49       Stephenson, RP. (1956) A modification of receptor theory. Br J Pharmacol  11:379.
 50
 51       Suskind, RR. (1985) Chloracne, the hallmark of dioxin intoxication. Scand  J Work Environ Health 11:165-171.
 52
53       Stohs, SJ. (1990) Oxidative stress induced by 2,3,7,8-tetrachlorodibenzo-p-dioxin  (TCDD). Free Rad Biol Med
54       9:79-90.
         6/8/00                                           146        DRAFT—DO NOT CITE OR QUOTE

-------
 1      Suskind, RR; Hertzberg, VS. (1984) Human health effects of 2,4,5-T and its toxic contaminants JAMA
 2      251:2372-2380.
 3
 4      Sweeney, A. (1994) Reproductive epidemiology of dioxins. In: Dioxins and health. Schecter, A, ed. New York:
 5      Plenum Press, pp. 549-558.
 6
 7      Sweeney, MH; Fingerhut, MA; Connally, LB; et al. (1989) Progress of the NIOSH cross-sectional medical study of
 8      workers occupationally exposed to chemicals contaminated with 2,3,7,8-TCDD. Chemosphere 19:973-977.
 9
10      Sweeney, MH; Calvert, GM; Egeland, GA; et al. (1997-98) Review and update of the results of the NIOSH medical
11      study of workers exposed to chemicals contaminated with 2,3,7,8-tetra-chlorodibenzo-p-dioxin. Teratog Carcinog
12      Mutagen 17(4-5):241-247.
13
14      Taylor, BL; Zhulin, IB. (1999) PAS domains: internal sensors of oxygen, redox potential, and light. Microbiol Mol
15      Biol Rev 63(2):479-506.
16
17      Teeguarden, JG; Dragan, YP; Singh, J; et al. (1999) Quantitative analysis of dose- and time-dependent promotion of
18      four phenotypes of altered hepatic foci by 2,3,7,8-tetrachlorodibenzo-p-dioxin in female Sprague-Dawley rats.
19      ToxicolSci51:211-223.
20
21      Tian, Y; Ke, S; Denison, MS; et al. (1999)  AhR and NF-kappaB interactions, a potential mechanism for dioxin
22      toxicity. J Biol Chem 274(1):510-515.
23
24      Tonn, T; Esser, C; Schneider, EM; et al. (1996)  Persistence of decreased T-helper cell function in industrial workers
25      20 years after exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Environ Health Perspect 104:422-426.
26
27      Tritscher, AM; Goldstein, JA; Portier, CJ; et al.  (1992) Dose-response relationships for chronic exposure to 2,3,7,8-
28      tetrachlorodibenzo-p-dioxin in a rat-tumor promotion model: quantification and immunolocalization of CYPlAland
29      CYP1A2 in the liver. Cancer Res 52:3436-3442.
30
31      Tritscher, AM; Clark, GC; Sewall, C; et al.  (1995) Persistence of TCDD-induced hepatic cell proliferation and
32      growth of enzyme altered foci after chronic exposure followed by cessation of treatment in DEN initiated female
33      rats. Carcinogenesis 16:2807-2811.
34
35      Tritscher, AM; Seacat, AM; Yager, JD; et al. (1996) Increased oxiditative DNA damage in livers of 2,3,7,8-
36      tetrachlorodibenzo-p-dioxin treated intact but not ovariectomized rats. Cancer Lett 98:219-225.
37
38      U.S. EPA. (1980) Risk assessment on (2,4,5-tetrachlorophenoxy) acetic acid [2,4,5-T], (2,4,5-trichlorophenoxy)
39      propionic acid, and 2,3,7,8-tetrachlorodibenzo-^-dioxin [TCDD]. Washington, DC.
40
41      U.S. EPA. (1985) Health  effects assessment document for polychlorinated dibenzo-p-dioxins. Prepared by the
42      Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, OH,
43      for the Office of Emergency and Remedial Response, Washington, DC. EPA/600/8-84/014F.
44
45      U.S. EPA. (1987) Interim procedures for estimating risks associated with exposures to mixtures of chlorinated
46      dibenzo-p-dioxins and -dibenzofurans (CDDs and  CDFs). EPA/625/3-87/012.
47
48      U.S. EPA. (1989a) Interim procedures for estimating risks associated with exposures to mixtures of chlorinated
49      dibenzo-p-dioxins and -dibenzofurans (CDDs and  CDFs) and 1989 update. Washington, DC: Risk Assessment
50      Forum. EPA/625/3-89.016.
51
52      U.S. EPA.  (1989b) Review  of draft documents: a cancer risk-specific dose estimate for 2,3,7,8-TCDD.
53      Washington DC. EPA Science Advisory Board Ad Hoc  Dioxin Panel.
54
        6/8/00                                           147        DRAFT—DO NOT CITE OR QUOTE

-------
  1       U.S. EPA. (1991a) Workshop report on toxicity equivalency factors for polychlorinated biphenyls congeners.
  2       EPA/625'3-91/020.
  3
  4       U.S. EPA. (1991 b) Guidelines for developmental toxicity risk assessment. Federal Register 57:22888-22938.
  5
  6       U.S. EPA. (1992) Draft report: A cross species-scaling factor for carcinogen risk assessment based on equivalence
  7       of mg/kg3/4/day. Federal Register 57(109):24152-24173.
  8
  9       U.S. EPA. (1995) An SAB Report: A second look at dioxin. EPA-SAB-EC-95-021.
10
11       U.S. EPA. (1994) Health assessment document for 2,3,7,5-tetrachlorodibenzo-p-dioxin (TCDD) and related
12       compounds. External review draft. Prepared by the Office of Health and Environmental Assessment, Office of
13       Research and Development, Washington, DC. EPA/600/BP-92/001a, b, c.  Available from NTIS, Springfield, VA
14       PB94-205457.
15
16       U.S. EPA. (1996) Proposed guidelines for carcinogen risk assessment. Federal Register 61:17960-18011.
17
18       U.S. EPA. (1999) Revised proposed guidelines for carcinogen risk assessment.
19
20       van Birgelen, AP; Van der Kolk, J; Fase, KM: et al. (1995) Subchronic dose-response study of 2,3,7,8-
21       tetrachlorodibenzo-p-dioxin in female Sprague-Dawley rats. Toxicol Appl Pharmacol 132:1-13.
22
23       van Birgelen, APJM; Diliberto, Devito, MJ; Bimbaum, LS. (1996) Tissue CYP1A1 activity relfects tissue 2,3,7,8-
24       tetrachlorodibenzo-p-dioxin concentrations. Organohalogen Compounds 29:439-442.
25
26       van Birgelen, APJM; Johnson, JD; Fuciarelli, AF; et al. (1999) Dose and time-response of TCDD in Tg.AC mice
27       after dermal and oral exposure. Dioxin '99:  19th International Symposium on Halogenated Environmental Organic
28       Pollutants and POPs. (ISBN 88-87772-02-9), Venice, Italy. Organohalogen Compounds 42:235-239.
29
30       van den Berg, M; Bimbaum, L; Bosveld, ATC; et al. (1998) Toxic equivalency factors (TEFs) for PCBs, PCDDs,
31       PCDFs for humans and wildlife. Environ Health Perspect 106( 12):775-792.
32
33       van den Heuvel, JP; Clark, GC; Kohn, MC; et al. (1994) Dioxin-responsive genes: examination of dose-response
34       relationships using quantitative reverse transcriptase-polymerase chain reaction. Cancer Res 54:62-68.
35
36       van der Plas, SA; Haag-Gronlund, M; Scheu, G; et al. (1999) Induction of altered hepatic foci by a mixture of
37       dioxin-like compounds with and without 2,2',4,4',5,5'-hexachlorobiphenyl in female Sprague-Dawley rats. Toxicol
38       Appl Pharmacol 156:30-39.
39
40       Vecchi, A; Sironi, M; Canegrati, MA; et al. (1983) Immunosuppressive effects of 2,3,7,8-tetrachiorodibenzo-p-
41       dioxin in strains of mice with different susceptibility to induction of aryl hydrocarbon hydroxylase. Toxicol Appl
42       Pharmacol 68:434-441.
43
44       Vena, J; Boffetta, P; Becher, H; et al. (1998) Exposure to dioxin and nonneoplastic mortality in the expanded IARC
45       international cohort study of phenoxy herbicide and chlorophenol production workers and sprayers.  Environ Health
46       Perspect 106 Suppl 2:645-653.
47
48       Vineis, P; Terracini, B; Ciccone, G; et al. (1986) Phenoxy herbicides and soft-tissue sarcomas in female rice
49       weeders: a population-based case-referent study. Scand J Work Environ Health 13:9-17.
50
51       Vogel, C; Donat, S; Dohr, O; et al. (1997) Effect of subchronic 2,3,7,8-tetrachlorodibenzo-p-dioxin exposure on
52       immune system and target gene responses in mice: calculation of benchmark doses for CYP1 Al and CYP1A2
53       related enzyme activities. Arch Toxicol 71:372-382.
54
         6/8/00                                           148        DRAFT—DO NOT CITE OR QUOTE

-------
 1      Waern, F; Flodstrom, S; Busk, L; et al. (1991) Relative liver tumour promoting activity and toxicity of some
 2      polychlorinated dibenzo-p-dioxin- and dibenzofuran-congeners in female Sprague-Dawley rats. Pharmacol Toxicol
 3      69:450-458.
 4
 5      Walker, NJ, Kim, A, Lucier, G, Tritscher, A. (1998) The use of tissue burden as a dose metric for TCDD-inducible
 6      presponses in rat liver is end point-specific. Organohalogen Compounds 38:337-340.
 7
 8      Walker, NJ; Portier, CJ; Lax, SF; et al. (1999) Characterization of the dose-response of CYP1B1, CYP1A1, and
 9      CYP1A2 in the liver of female Sprague-Dawley rats following chronic exposure to 2,3,7,8-tetrachlorodibeiizo-p-
10      dioxin. Toxicol Appl Pharmacol 154:279-286.
11
12      Walker, NJ; Tritscher, AM; Sills, RC; et al. (2000) Hepatocarcinogenesis in female Sprague-Dawley rats following
13      discontinuous treatment with 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol Sci, in press.
14
15      Webb, KB; Evans, RG; Knudsen, DP; et al. (1989) Medical evaluation of subjects with known body levels of
16      2,3,7,8-tetrachlorodibenzo-p-dioxin. J Toxicol Environ Health 28:183-193.
17
18      Weisglas-Kuperus, N; Sas, TCJ; Koopman-Esseboom, C; et al. (1995) Immunologic effects of background prenatal
19      and postnatal exposure to dioxins and polychlorinated biphenyls in Dutch infants. Pediatr Res 38:404-410.
20
21      WHO. (2000) International Programme on Chemical Safety: harmonization of approaches to the assessment of
22      chemicals. Fact Sheet No.8.
23
24      Wilson, CL; Safe, S. (1998) Mechanisms of ligand-induced aryl hydrocarbon receptor-mediated biochemical and
25      toxic responses. Toxicol Pathol 26:657-671.
26
27      Yager, JD; Liehr, JG. (1996) Molecular mechanisms of estrogen carcinogenesis.  Ann Rev Pharmacol Toxicol
28      36:203-232.
29
30      Yang, JZ; Foster, WG. (1997) Continuous exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin inhibits the growth of
31      surgically induced endometriosis in the ovariectomized mouse treated with high dose estradiol. Toxicol Ind Health
32      13(l):15-25.
33
34      Yang, JZ; Agarwal, S; Foster, WG. Subchronic exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin modulates the
35      pathophysiology of endometriosis in the cynomolgus monkey. Toxicol Sci. In press.
36
37      Zeise, L; Huff, JE; Salmon, AG; et al. (1990) Human risks from 2,3,7,8-tetrachlorodibenzo-p-dioxin and
38      hexachlorodibenzo-p-dioxins. In: Advances in modern environmental toxicology, v. 17. Princeton, NJ: Princeton
39      Scientific Publishing Co., Inc; pp. 293-342.
40
41      Zober, A; Messerer, P; Huber, P. (1990) Thirty-four-year mortality follow-up of BASF employees exposed to
42      2,3,7,8-TCDD after the 1953 accident. Int Arch Occup Environ Health 62:138-157.
43
44      Zober, MA; Ott, MG; Papke, O; et al.  (1992) Morbidity study  of extruder personnel with potential exposure to
45      brominated dioxins and furans. I. Results of blood monitoring  and immunological tests. Br J Ind Med 49:532-544.
46
47      Zober, A; Ott, MG; Messerer, P. (1994) Morbidity follow up study of BASF employees exposed to 2,3,7,8-
48      tetrach'lorodib'enzo-p-dioxin (TCDD) after a 1953 chemical reactor incident. Occup Environ Med 51:479-486.
49
50
51
52
       6/8/00                                           149        DRAFT—DO NOT CITE OR QUOTE

-------