Summary of U.S. EPA Dioxin Workshop, February 18-20, 2009, Cincinnati, Ohio


                               EPA/600/R-09/027
                                   May 2009
Summary of U.S. EPA
   Dioxin Workshop
    February 18-20,2009
         Cincinnati, Ohio
   National Center for Environmental Assessment
      Office of Research and Development
     U.S. Environmental Protection Agency
          Cincinnati, OH 45268

-------
                                     DISCLAIMER
       This document summarizes the discussions presented at the Dioxin Workshop in
February 2009, in Cincinnati, OH, as documented by the Session Co-Chairs. This document is
not all inclusive or binding.  Conclusions and recommendations to the U.S. EPA may not
represent full consensus.  The views expressed in this document are those of the Dioxin
Workshop Panelists and do not necessarily reflect the views and policies of the U.S.
Environmental Protection Agency.  Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
Preferred Citation:
U.S. Environmental Protection Agency (U.S. EPA). (2009) Summary of U.S. EPA Dioxin Workshop:
February 18-20, 2009. U.S. Environmental Protection Agency, National Center for Environmental Assessment,
Cincinnati, OH. EPA/600/R-09/027.
                                            11

-------
                        TABLE OF CONTENTS
DIOXIN WORKSHOP TEAM	iv
ACKNOWLEDGMENTS	iv

INTRODUCTION	1

     REFERENCES	2

SCIENTIFIC WORKSHOP TO INFORM THE TECHNICAL WORK PLAN FOR
   U.S. EPA'S RESPONSE TO NAS COMMENTS ON THE HEALTH EFFECTS
   OF DIOXIN PRESENTED IN U.S. EPA'S DIOXIN REASSESSMENT	3

     SESSION 1: QUANTITATIVE DOSE-RESPONSE MODELING IS SUES	3
     SESSION 2: IMMUNOTOXICITY	6
     SESSION 3 A: DOSE-RESPONSE FOR NEUROTOXICITY AND
         NONREPRODUCTIVE ENDOCRINE EFFECTS	8
     SESSION 3B: DOSE-RESPONSE FOR CARDIOVASCULAR TOXICITY AND
         HEPATOTOXICITY	11
     SESSION 4A: DOSE-RESPONSE FOR CANCER	13
     SESSION 4B: DOSE-RESPONSE FOR REPRODUCTIVE/DEVELOPMENTAL
         TOXICITY	16
     SESSION 5: QUANTITATIVE UNCERTAINTY ANALYSIS OF DOSE-
         RESPONSE	20

APPENDIX A: 2009 U.S. EPA DIOXIN WORKSHOP AGENDA	24

APPENDIX B: 2009 U.S. EPA DIOXIN WORKSHOP QUESTIONS TO GUIDE
   PANEL DISCUSSIONS	31

APPENDIX C: 2009 U.S. EPA DIOXIN WORKSHOP DRAFT SELECTION
   CRITERIA TO IDENTIFY KEY IN VIVO MAMMALIAN STUDIES THAT
   INFORM DOSE-RESPONSE MODELING FOR
   2,3,7,8-TETRACHLORODIBENZO-p-DIOXIN (TCDD)	34
                                 in

-------
                            DIOXIN WORKSHOP TEAM
The Dioxin Workshop Team, under the leadership of Peter W. Preuss, Director, NCEA,
comprised the following members:

National Center for Environmental Assessment, Office of Research and Development,
U.S. Environmental Protection Agency, Cincinnati, OH 45268
   Belinda S. Hawkins
   Janet Hess-Wilson
   Glenn Rice
   Jeff Swartout
   Linda K. Teuschler
   Bette Zwayer

Argonne National Laboratory, Argonne, IL 60439
   Maryka H. Bhattacharyya
   Andrew Davidson
   Mary E. Finster
   Margaret M. MacDonell
   David P. Peterson
ACKNOWLEDGMENTS

The Track Group, Alexandria, VA 22312
   Kara Hennigan
   Alan Minton
   Brandy Quinn

ECFlex, Inc., Fairborn, OH 45324
   Dan Heing
   Heidi Glick
   Amy Prues
   Lana Wood

IntelliTech Systems, Inc., Fairborn, OH 45324
   Cris Broyles
   Luella Kessler
   Stacey Lewis
   Linda Tackett
                                         IV

-------
INTRODUCTION
This document provides a summary of the Scientific Workshop to Inform EPA's
Response to National Academy of Science Comments on the Health Effects of Dioxin in EPA's
2003 Dioxin Reassessment. The U.S. Environmental Protection Agency (U.S. EPA) and
Argonne National Laboratories (ANL), through an inter-Agency agreement with the U.S.
Department of Energy, convened this scientific workshop ("Dioxin Workshop") on February
18-20, 2009, in Cincinnati, Ohio. The goals of the Dioxin Workshop were to identify and
address issues related to the dose-response assessment of 2,3,7,8-tetrachlorodibenzo-p-dioxin
(TCDD). This report summarizes the discussions and conclusions from this workshop.
Previously, at the request of the U.S. EPA, the National Academy of Sciences (NAS) prepared a
report, Health Risks from Dioxin and Related Compounds: Evaluation of the EPA Reassessment
(NAS, 2006), which made a number of recommendations to improve the U.S. EPA's risk
assessment for TCDD (U.S. EPA, 2003). The 3-day Dioxin Workshop was convened
specifically to ensure that the U.S. EPA's response to the NAS recommendations focuses on the
key issues and reflects the most meaningful science.

The Dioxin Workshop included seven scientific sessions:
(1) Session 1: Quantitative Dose-Response Modeling Issues
(2) Session 2: Immunotoxicity
(3) Session 3 A: Dose-Response for Neurotoxicity and Nonreproductive Endocrine Effects
(4) Session 3B: Dose-Response for Cardiovascular Toxicity and Hepatotoxicity
(5) Session 4A: Dose-Response for Cancer
(6) Session 4B: Dose-Response for Reproductive/Developmental Toxicity
(7) Session 5: Quantitative Uncertainty Analysis of Dose-Response

During each session, the U.S. EPA asked a panel of expert scientists to:

• identify and discuss the technical challenges involved in addressing the key NAS
comments on the TCDD dose-response assessment in the U.S. EPA Reassessment
(U.S. EPA, 2003);

• discuss approaches for addressing the key NAS comments; and

• identify important published, independently peer-reviewed literature, particularly studies
describing epidemiologic and in vivo mammalian bioassays, which are expected to be
most useful for informing the U.S. EPA's response.

The sessions were followed by open comment periods during which members of the
audience were invited to address the Panels. At the conclusion of the open comment periods, the
Panel Co-Chairs were asked to summarize and present the results of the panel discussions. The
summaries could include minority opinions stated by panelists. The main points derived from
the session summaries were used to prepare this document. Additionally, this document includes
a list of the session panelists and their affiliations and three appendices. Appendix A presents
the Dioxin Workshop Agenda. Appendix B identifies the charge questions presented to the
Panel. Appendix C describes draft study selection criteria proposed by the Dioxin Workshop
Team for consideration by the workshop panelists.

-------
REFERENCES
NAS (National Academy of Sciences).  2006. Health Risks from Dioxin and Related
Compounds: Evaluation of the EPA Reassessment. National Academies Press, Washington, DC
(July).  Available at http://www.nap.edu/catalog.php7record id= 11688.

U.S. EPA (U.S. Environmental Protection Agency). 2003.  Exposure and Human Health
Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds. NAS
review draft, Volumes 1-3 (EPA/600/P-00/001Cb, Volume 1). U.S. Environmental Protection
Agency, National Center for Environmental Assessment, Washington, DC (December).
Available at http://www.epa.gov/nceawwwl/pdfs/dioxin/nas-review/.

-------
SCIENTIFIC WORKSHOP TO INFORM THE TECHNICAL WORK PLAN FOR U.S.
EPA'S RESPONSE TO NAS COMMENTS ON THE HEALTH EFFECTS OF DIOXIN
PRESENTED IN U.S. EPA'S DIOXIN REASSESSMENT

Dioxin Workshop Co-Chairs: Peter W. Preuss and Glenn Rice

The Dioxin Workshop session summaries were prepared by the session panel Co-Chairs
with input from the panelists, as requested by the U.S. EPA prior to the workshop. The Co-
Chairs subsequently presented these summaries to all of the workshop participants during
designated periods at the workshop. In these summaries, the U.S. EPA asked that the Co-Chairs
summarize the key issues from the panel discussions. Because the sessions were not designed to
achieve consensus among the panelists, the summaries do not necessarily represent consensus
opinions; rather, they reflect the essence of the panel discussions. Some of the specific points
may represent the views of multiple panelists, while others only the views of a single panelist.
Prior to the summarizations, there were opportunities for public comments on the discussion
topics. Some Co-Chairs met with their sessions' panelists after their sessions ended to develop
these summaries, while others developed reports based on their personal notes. Because Session
5 was the last session of the workshop—with little time provided to develop the summary—the
Co-Chairs circulated a draft for comment by the Session 5 panelists after the workshop, prior to
finalizing the session summary. The U.S. EPA collected the session summaries and then
prepared this document. A draft of this document was distributed to all of the session Co-Chairs
to provide them with a final opportunity to comment and make revisions. Finally, it should be
noted that U.S. EPA was not prescriptive to the session Co-Chairs with respect to the format of
the presentation materials and provided no specific instructions, resulting in unique formats
among the session summaries.
SESSION 1: QUANTITATIVE DOSE-RESPONSE MODELING ISSUES
This session discussed the general dose-response modeling issues related to TCDD.
Many of these issues were highlighted by NAS (2006). There was a general introductory
presentation on TCDD kinetics, including information and uncertainties pertaining to the
conversion of administered doses in animals to human body burden (BB) and additivity to
background issues. This presentation was followed by a Panel discussion on the state of the
science regarding dioxin dose-response modeling issues.

Session 1 Panelists (Session Co-Chairs are identified by asterisk)
• Bruce Allen, Bruce Allen Consulting
• Lesa Aylward, Summit Toxicology
• Roger Cooke, Resources for the Future
• Kenny Crump, Louisiana Tech University
• MikeDeVito, U.S. EPA
• Dale Hattis, Clark University
• Rick Hertzberg, Biomath Consulting
• Rob McDowell, U.S. Department of Agriculture
• Jim Olson, State University of New York, University at Buffalo

-------
• *Lorenz Rhomberg, Gradient
• Woody Setzer, U.S. EPA
• * Jeff Swartout, U.S. EPA

Please note that the use of the term "concluded" or "recommended" in this summary does not mean that a consensus
was reached. Session Summaries were written from the material prepared by the non-EPA/ANL Co-Chair and
represent a synopsis of the panel discussions.
Key Study Selection Criteria
The Panel discussed the advantages and disadvantages of using key study criteria
(Appendix C). They concluded that a priori criteria foster transparency and consistency, and
could deflect a posteriori criticism. However, the Panel also acknowledged that having a priori
criteria could introduce the potential for excluding useful data. Although the key study criteria
provided by the U.S. EPA listed studies using TCDD only as a criterion, the Panel posed the
possibility of using closely related dioxin-like compounds (DLCs) as surrogates for TCDD. The
criterion for use of data from mammalian studies only was one criterion that received generalized
support due to the lack of extrapolation protocols for nonmammalian species. The Panel also
discussed the specific exposure-duration criterion and asked if there should be a preference for
longer-term rather than acute studies. The Panel made three suggestions to modify U.S. EPA's
key study selection criteria:
(1) Define more relevant exposure-level (i.e., dose) cut points using tissue concentrations.
(2) Reword statistical criteria to include do-it-yourself analysis.
(3) Reword the response criteria to clarify "outside of normal range."

Dose Metrics
The Panel discussed the relative merits of various measures of dose for modeling TCDD
dose response. One general conclusion was that tissue concentration (TC) is the preferred
metric, especially lipid-adjusted TC, because this measure more closely approximates exposures
close to the target tissue when compared to administered doses. However, the Panel
acknowledged that these data are often unavailable. They further noted that BB, which is
defined as the concentration of TCDD in the body (ng/kg body weight) (U.S. EPA, 2003), might
be useful as a surrogate for TC provided the two measures were proportional.

The Panel suggested that a linear approach to BB estimation, which was utilized by
U.S. EPA (2003), is too simplistic because this approach does not take into account toxicokinetic
issues related to TCDD—e.g., sequestration in the liver and fat, age-dependent elimination, and
changing elimination rates over time. The Panel recommended the use of kinetic/mechanistic
modeling to the extent possible to quantify tissue-based metrics.

The Panel raised the issue of whether the preferred dose metric would be different for
different endpoints and exposure durations. This led to the Panel's comment that the peak
exposure might be a more important metric than average BB for variable exposure scenarios.
Given this discussion about different exposure durations being relevant to a specific endpoint,
the Panel suggested that the U.S. EPA also consider peak measures in dose-response modeling.

-------
The last point raised in this part of the discussion centered on the possibility of dose
errors in experimental studies. The Panel highlighted the need for the U.S. EPA to consider dose
error (i.e., uncertainty in the x-axis of the dose-response curve) when using dose surrogates.

Dose-Response Modeling of Mammalian Bioassays
The Panel considered several issues related to dose-response modeling of mammalian
bioassay data for TCDD: supralinearity and incomplete response data ("anchoring"), defining the
benchmark response (BMR) level with respect to establishing the point of departure (POD), and
the use of threshold modeling—as further explained below.

The Panel discussed the specific issues of supralinearity and anchoring raised by the
U.S. EPA with respect to modeling noncancer endpoints. The panel recognized that, for many of
the most sensitive endpoints, the response at the lowest dose is high (e.g., quantal responses
above 25% and continuous endpoints differ substantially from the mean, often implying 100%
incidence in the treated animals). This lack of response anchoring at the low end of the dose-
response curve (near the BMR) results in the higher responses determining the shape of the
curve.

The Panel asked whether new tools might be needed or whether the current tools could be
applied differently. In the context of developing new tools, the Panel emphasized the need for
collaboration between biologists and mathematicians. When discussing application, the Panel
suggested that the problem with supralinearity might be overcome by simply dropping the
requirement for using the lower bound on the Benchmark Dose. In addition, the Panel posed
several more approaches for further consideration in dose-response modeling by the U.S. EPA:
(1) Combine similar data sets to fill in data gaps.
(2) Use mechanistic approaches to model the data gaps.
(3) Dichotomize continuous data.

Finally, the Panel acknowledged that, in certain situations, there simply may not be enough
information to provide meaningful answers.
The Panel discussed the BMR level for establishing a POD in the context of deriving a
Reference Dose (RfD). The Panel generally agreed that, while the effective dose level (EDoi)
used in the 2003 Reassessment may be useful for comparative analysis across endpoints, the
EDoi estimates developed for all endpoints considered in the Reassessment were not appropriate
for deriving an RfD because they were not based on the effect's adversity. The panel noted that
EDoi also is much lower than typical EPA BMR levels. The Panel recommended that the U.S.
EPA work to define endpoint-specific BMRs based on the consideration of adversity. Given that
the same uncertainty factor framework is applied to all PODs, the Panel emphasized the need for
consistency in BMRs; numerical consistency is needed for quantal BMRs and consistency in the
choice of biological relevance should be applied for continuous BMRs.

The Panel generally discouraged threshold modeling by stating that thresholds are very
difficult to pin down and suggested that the lower bound may always be zero.

-------
Dose-Response Modeling of Epidemiological Studies
The Panel noted that many studies have been published with measured concentrations of
TCDD that could be used for dose reconstruction. In this discussion, the Panel acknowledged
that use of these data would entail dealing with toxicity equivalence (TEQ) issues and
pharmacokinetic (PK) modeling. Pertaining to the use of these data for quantitative risk
assessment by the U.S. EPA, the Panel posed the question, "At what point does indirect or
confounded human data supersede controlled animal bioassay data?", or alternatively, "How
much human data uncertainty can we tolerate?" The Panel suggested, at the least, that the
epidemiologic data could be used to "ground-truth" the animal bioassay modeling results.

Supporting Information
The Panel acknowledged that Ah receptor (AhR) binding affinities are not necessarily
tied to endpoint sensitivity, but they reiterated the need to consider mechanistic modeling to aid
in developing appropriate dose metrics or filling in data gaps in the existing dose-response data.

References

NAS (National Academy of Sciences). 2006. Health Risks from Dioxin and Related
Compounds: Evaluation of the EPA Reassessment. National Academies Press, Washington, DC
(July). Available at http://www.nap.edu/catalog.php?recordid=l 1688.

U.S. EPA (U.S. Environmental Protection Agency). 2003. Exposure and Human Health
Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds. NAS
Review Draft (EPA/600/P-00/001Cb). U.S. Environmental Protection Agency, National Center
for Environmental Assessment, Washington, DC. Available at
http://www.epa.gov/nceawwwl/pdfs/dioxin/nas-review/.
SESSION 2: IMMUNOTOXICITY
The U.S. EPA plans to consider development of a quantitative dose-response assessment
for the immunologic effects associated with TCDD exposure. Such an assessment would be
based on information in U.S. EPA (2003), NAS (2006) and key studies identified in this
workshop. The purpose of this session was to identify and discuss key issues pertaining to dose-
response assessment for dioxin-induced immunologic effects.

Session 2 Panelists (Session Co-Chairs are identified by asterisk)

• Roger Cooke, Resources for the Future
• Rob Goble, Clark University
• *Belinda Hawkins, U.S. EPA
• Nancy Kerkvliet, Oregon State University
• Manolis Kogevinas, Centre for Research in Environmental Epidemiology
• Robert Luebke, U.S. EPA
• Paolo Mocarelli, University of Milan
• * Allen Silverstone, State University of New York, Upstate Medical University

-------
• Courtney Sulentic, Wright State University
• Nigel Walker, National Institute of Environmental Health Sciences

The Panel discussed several studies often considered important for assessing the
immunotoxic effects of TCDD exposure. The Oughton et al. (1995) mouse bioassay was
discussed and, although the study does meet the proposed criteria, it could not be considered a
key study; specifically, the Panel contended that since there were no functional alterations
observed or measured in this bioassay, the changes in cellular phenotypes are only "suggestive"
of immune alterations and cannot be regarded as having immunopathologic significance.

The Panel discussed two additional studies for further consideration by the U.S. EPA:

• Baccarelli et al. (2002). The Panel discussed this as a potentially key human
epidemiological study that should be reviewed and considered further by the U.S. EPA.
It measured the level of IgG, demonstrating a significant decline relative to dioxin body
burdens.

• Smialowicz et al. (2008). The Panel noted that this study identified the antibody response
to sheep red blood cells (SRBCs) as the critical effect, labeling this protocol as a
functional assay. The Panel stated that if modeled, the U.S. EPA could calculate the
BMR for this endpoint as 1 standard deviation from the control mean.

References

Baccarelli, A., P. Mocarelli, D.G. Patterson et al. 2002. Immunologic effects of dioxin: New
results from Seveso and comparison with other studies. Environ. Health Perspect.
NAS (National Academy of Sciences). 2006. Health Risks from Dioxin and Related
Compounds: Evaluation of the EPA Reassessment. National Academies Press, Washington, DC
(July). Available at http://www.nap. edu/catalog.php?record_id=l 1688.

Oughton, J.A., C.B. Pereira, O.K. Dekrey, J.M. Collier, A.A. Frank and N.I. Kerkvliet. 1995.
Phenotypic analysis of spleen, thymus, and peripheral blood cells in aged C57BI/6 mice
following long-term exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol. Sci. 25(l):60-69.

-------
Smialowicz, R.J., MJ. DeVito, W.C. Williams and L.S. Birnbaum. 2008. Relative potency
based on hepatic enzyme induction predicts immunosuppressive effects of a mixture of
PCDDS/PCDFSandPCBS. Toxicol. Appl. Pharmacol. 227(3):477-484.

U.S. EPA (U.S. Environmental Protection Agency). 2003. Exposure and Human Health
Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds. NAS
Review Draft (EPA/600/P-00/001Cb). U.S. Environmental Protection Agency, National Center
for Environmental Assessment, Washington, DC. Available at
http://www.epa.gov/nceawwwl/pdfs/dioxin/nas-review/.
SESSION 3A: DOSE-RESPONSE FOR NEUROTOXICITY AND NONREPRODUCTIVE
ENDOCRINE EFFECTS
The U.S. EPA plans to consider development of a quantitative dose-response assessment
for neurological and/or nonreproductive endocrine effects associated with TCDD exposure.
Such an assessment would be based on information in U.S. EPA (2003), NAS (2006) and key
studies identified in this workshop. The purpose of this session was to identify and discuss key
issues pertaining to dose-response assessment for dioxin-induced neurological and/or
nonreproductive endocrine effects.

Session 3A Panelists (Session Co-Chairs are identified by asterisk)

• *Maryka Bhattacharyya, Argonne National Laboratory
• Mike DeVito, U.S. EPA
• Mary Gilbert, U.S. EPA
• Rob Goble, Clark University
• Nancy Kerkvliet, Oregon State University
• Fumio Matsumura, University of California-Davis
• Paolo Mocarelli, University of Milan
• Chris Portier, National Institute of Environmental Health Sciences
• Lorenz Rhomberg, Gradient
• Allen Silverstone, State University of New York, Upstate Medical University
• Marie Sweeney, National Institute of Occupational Safety and Health
• *Bernie Weiss, University of Rochester

-------
endpoint. The Panel acknowledged that, in some cases, the window of exposure may precede the
window of expression of toxicity.

Epidemiological Study Selection
Developmental Neurotoxicity
The Panel recognized that an unusual feature for this endpoint is that there are sufficient
human data for dose-response modeling (e.g., Dutch children [Huisman et al., 1995; Patandin et
al., 1999] and U.S. children [Jacobson and Jacobson, 1996]) and there is an internal dose metric
(serum concentrations). Additionally, the Panel discussed recent studies that address this
endpoint in humans (from Japan [reference not provided] and Holland [e.g., Koopman-Esseboom
et al., 1996; Vreugdenhil et al., 2002]). For continued investigation into this endpoint, the Panel
raised two issues to the U.S. EPA:

• Conduct an evaluation of whether a modeled effect can be attributed to TCDD and not
some other persistent organic pollutant (POP), although the Panel recognized that it is
unlikely U.S. EPA will be able to distinguish among these exposures because other POPs
are intrinsic confounders in the Dutch study.

• Allow animal data to inform the dose-response modeling of epidemiological data.

Thyroid Dysfunction
The Panel identified the availability of human data for this endpoint (e.g., Calvert et al.,
1999; Koopman-Esseboomet al., 1994). Much of the thyroid dysfunction literature has been
published since the 2003 Reassessment (e.g., Wang et al., 2005; Baccarelli et al., 2008). The
Panel also noted the availability of an internal dose metric (serum concentrations). Additionally,
the Panel discussed the mechanistic studies in animals that link TCDD to thyroid dysfunction.
For continued investigation into this endpoint, the Panel raised three issues for the U.S. EPA to
consider:

• Consider the newly available human data since the Reassessment.

• Investigate and clarify of the role of TCDD-induced thyroid dysfunction in
developmental neurotoxicity.

• Evaluate and determine whether an effect can be attributed to TCDD or other
contaminants.

Diabetes
The Panel discussed that data suggest that diabetes incidence in those under 55 years old
may be associated with exposure to PCBs. They acknowledged that whether this is a dioxin-like
compound (DLC) mediated effect or whether other POPs are responsible is still undetermined.
The Panel also acknowledged that no animal model exists for the investigation of xenobiotic-
induced diabetes, and that separating the injury dose level from the current body burdens would
depend on good pharmacokinetics in humans. For continued investigation into this endpoint, the
Panel listed two issues for the U.S. EPA to consider:

• Results from the Anniston study and the Great Lakes Fishermen study (references not
provided) should be examined for dose metrics (both studies examine human PCB
exposures).

-------
   •   Changes of adipose tissue status need to be considered, given that dieting can cause
       release of lipid-soluble contaminants.

References

Baccarelli, A., S.M. Giacomini, C. Corbetta et al.  2008.  Neonatal thyroid function in Seveso 25
years after maternal exposure to dDioxin. PLoS Med.  5(7):el61.
doi:10.1371/journal.pmed.0050161.

Calvert, G.M., M.H. Sweeney, J. Deddens and O.K. Wall.  1999.  Evaluation of diabetes
mellitus, serum glucose, and thyroid function among United States workers exposed to
2,3,7,8-tetrachlorodibenzo-p-dioxin. Occ. Env. Med.  56:270-276.

Huisman, M., C. Koopman-Esseboom, V. Fidler et al.  1995. Perinatal exposure to
polychlorinated biphenyls and dioxins and its effect on neonatal neurological development.
Early Hum. Devel. 41(2):111-127.

Jacobson, J.L. and S.W. Jacobson.  1996. Intellectual impairment in children exposed to
polychlorinated biphenyls in utero. N. Engl. J. Med. 335:783-789.

Koopman-Esseboom, C., N. Weisglas-Kuperus, M.A.J. deRidder, C.G. Van derPaauw,
L.G.M.Th. Tuinstra and PJJ. Sauer.  1996.  Effects of poly chlorinated biphenyl/dioxin exposure
and feeding type on infants' mental and psychomotor development.  J. Pediatr.  97(5):700-706.

Koopman-Esseboom, C., D.-C. Morse, N. Weisglas-Kuperus et al. 1994. Effects of dioxins and
polychlorinated biphenyls on thyroid hormone status of pregnant women and their infants.
Pediatr. Res.  36:468-473.

Patandin, S., C.I. Lanting, P.G.H. Mulder, E.R. Boersma, PJJ. Sauer and N. Weisglas-Kuperus.
1999. Effects of environmental exposure to polychlorinated biphenyls and dioxins on cognitive
abilities in Dutch children at 42 months of age. J. Pediatr.  134:33-41.

NAS (National Academy of Sciences).  2006.  Health Risks from Dioxin and Related
Compounds: Evaluation of the EPA Reassessment.  National Academies Press, Washington, DC
(July).  Available at http://www.nap.edu/catalog.php?record_id=l 1688.

U.S. EPA (U.S. Environmental Protection Agency). 2003. Exposure and Human Health
Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds.  NAS
Review Draft (EPA/600/P-00/001Cb). U.S. Environmental Protection Agency, National Center
for Environmental Assessment, Washington, DC.  Available at
http://www.epa.gov/nceawwwl/pdfs/dioxin/nas-review/.

Vreugdenhil, H.J., C.I. Lanting, P.G. Mulder, E.R. Boersma and N. Weisglas-Kuperus. 2002.
Effects of prenatal PCB and dioxin background exposure on cognitive and motor abilities in
Dutch children at school age. J. Pediatr. 140:48-56.
                                          10

-------
Wang S.L., P.H. Su, S.B. Jong, Y.L. Guo, W.L. Chou and O. Papke. 2005. In utero exposure to
dioxins and polychlorinated biphenyls and its relations to thyroid function and growth hormone
in newborns. Environ. Health Perspect. 113:1645-1650.
SESSION 3B: DOSE-RESPONSE FOR CARDIOVASCULAR TOXICITY AND
HEPATOTOXICITY

The U.S. EPA plans to consider development of a quantitative dose-response assessment
for cardiovascular and/or hepatic effects associated with TCDD exposure. Such an assessment
would be based on information in U.S. EPA (2003), NAS (2006) and key studies identified in
this workshop. The purpose of this session was to identify and discuss key issues pertaining to
dose-response assessment for dioxin-induced cardiovascular and/or hepatic effects.

Session 3B Panelists (Session Co-Chairs are identified by asterisk)

• Bob Budinksy, Dow Chemical
• Manolis Kogevinas, Centre for Research in Environmental Epidemiology
• Rob McDowell, U.S. Department of Agriculture
• Jim Olson, State University of New York, University at Buffalo
• Marian Pavuk, Agency for Toxic Substances and Disease Registry
• * Jeff Swartout, U.S. EPA
• *Mary Walker, University of New Mexico
• Nigel Walker, National Institute of Environmental Health Sciences

The panel also was concerned that the gavage procedure can increase mouse blood
pressure. Consequently, the panel recommended that gavage studies not be used for the blood
pressure endpoint (i.e., only dietary dosing studies should be considered).

Human Health Endpoints
In relation to the hepatic endpoint, the Panel acknowledged the large body of dose
response information on hepatic effects in rodents and that enzyme (mostly CYP1 Al) induction
was a sensitive effect. However, the Panel cited the lack of linkage of CYP1A1 to downstream
events, which complicates the toxicological interpretation of this endpoint, and concluded that
11

-------
the more important liver effects in rodents are probably on the "road to cancer." The Panel noted
that hepatic effects were not seen in the epidemiological studies, but acknowledged that these
studies were not designed to detect them.

In relation to the cardiovascular endpoint, the Panel identified hypertension and ischemic
heart disease (IHD) as two key endpoints from the epidemiological studies. The Panel
recommended that the U.S. EPA perform a meta-analysis of these data. The Panel also
commented that recent animal studies support the observations linking TCDD exposure to IHD
and hypertension. In particular, the National Toxicology Program (NTP) study shows
inflammatory and structural effects on resistant vascular arterioles (NTP, 2006). Additional
evidence from the study suggests that the vascular effects may be CYP1A1-dependent. The
Panel suggested that the NTP study data might be used as a surrogate for dose-response
modeling of hypertension and that such an approach would be supported by data on the role of
AhR in vascular function and remodeling.

POD Issues
The Panel was not supportive of 1% of maximal response (EDoi), which was utilized in
the 2003 Reassessment. The Panel concluded that the POD should depend on the specific
endpoint and recommended the following to the U.S. EPA:

• For continuous measures, base the BMR on difference from control. Consider the
adversity level—at what point does the endpoint become adverse?

• For incidence data, set the BMR to a fixed-risk level.

Supporting Information
The Panel posed several suggestions to the U.S. EPA for reducing uncertainty and
improving the knowledge base for TCDD toxicity.

• Use in vitro data to define uncertainties, such as the relative sensitivity between rodents
and humans and around the definition of a POD.

• Consider studies on dioxin-like compounds (DLCs).

• Use PK modeling to define the dose metric for hepatic effects.

• Use body burden or serum concentrations for cardiovascular endpoints.
Finally, the Panel recommended that U.S. EPA finish the reassessment quickly and establish a
definitive plan to review and incorporate new data as they become available.

References

-------
NTP (National Toxicology Program). 2006. Toxicology and Carcinogenesis Studies of
2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) (CAS No. 1746-01-6) in Female Harlan Sprague-
Dawley Rats (Gavage Studies). U.S. Department of Health and Human Services. NTP TR 521.
Research Triangle Park, NC (April).

U.S. EPA (U.S. Environmental Protection Agency). 2003. Exposure and Human Health
Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds. NAS
Review Draft (EPA/600/P-00/001Cb). U.S. Environmental Protection Agency, National Center
for Environmental Assessment, Washington, DC. Available at
http://www.epa.gov/nceawwwl/pdfs/dioxin/nas-review/.
SESSION 4A: DOSE-RESPONSE FOR CANCER
The U.S. EPA plans to consider development of a quantitative dose-response assessment
for cancer associated with TCDD exposure. Such an assessment would be based on information
in U.S. EPA (2003), NAS (2006) and key studies identified in this workshop. The purpose of
this session was to identify and discuss key issues pertaining to dose-response assessment for
dioxin-induced cancer.

Session 4A Panelists (Session Co-Chairs are identified by asterisk)

• Lesa Aylward, Summit Toxicology
• Kenny Crump, Louisiana Tech University
• Dale Hattis, Clark University
• * Janet Hess-Wilson, U. S. EPA
• Karen Hogan, U.S. EPA
• Manolis Kogevinas, Centre for Research in Environmental Epidemiology
• Marian Pavuk, Agency for Toxic Substances and Disease Registry
• Chris Portier, National Institute of Environmental Health Sciences
• Lorenz Rhomberg, Gradient
• Jay Silkworm, General Electric
• *Nigel Walker, National Institute of Environmental Health Sciences

Key Study Selection
The Panel discussed both human and rodent studies. In reviewing the epidemiological
data, the Panel agreed the EPA should focus on four cohort studies (Dutch cohort, NIOSH
cohort, BASF accident cohort, and Hamburg cohort) and pointed out that there are numerous
updates and reevaluations of data now in the literature and others will be published soon. The
Panel stated that it is appropriate for the U.S. EPA to consider the increase in total cancers for
modeling human cancer data, however, Non-Hodgkin's lymphoma, and lung tumors are the main
TCDD-related cancer types seen in humans exposed to TCDD. The Panel suggested the U.S.
EPA focus the quantitative dose-response modeling on the human data.
13

-------
In reviewing the rat data, the Panel identified four new NTP rodent cancer bioassays with
liver and lungs as the main target organs. However, they suggested that dose-response modeling
efforts should model "all cancers" from these NTP data sets as well and use tumor incidence—
not individual rats as measures.

Key Study Selection Criteria
The Panel discussed whether data for TCDD only should be used or if PCB126 could be
used to develop a dose-response curve. From this discussion, the Panel reached a general
agreement that limiting the dose-response modeling and cancer assessment to TCDD only would
be the best approach.

Regarding the oral dosing regimens, the Panel discussed the differences in results from
different bioassays. They concluded that there were insufficient data to pick between oral feed
(Kociba et al., 1978) and oral gavage (NTP, 2006) studies, but stated "If all aspects of studies
were equal, an oral feed study is preferred." However, given that current data sets are not equal,
they agreed that U.S. EPA should consider both feed and gavage studies.

The Panel put forth the recommendation that studies that include initiation-promotion
model data and TgAC transgenic model data from oral exposure studies should be excluded from
the primary category in the key study selection criteria (Appendix C lists the draft study selection
criteria distributed prior to the meeting). Studies from both classifications should be moved to
the second tier.

The Panel was also unsupportive of the "response magnitude outside the range of normal
variability" criterion, as they did not believe it was applicable to a cancer endpoint.

Critical Endpoints to Consider
The Panel recognized that the MOA for TCDD includes cell growth/differentiation
dysregulation, that different endpoints (tumor types) across species may be expected, and that
there are differences in tumor sites across species. The Panel further acknowledged that there is
insufficient information to determine if rodent tumor types observed are relevant to humans.
Thus, the Panel suggests the following:

• U.S. EPA should consider all the observed cancer endpoints in its evaluation.

Nonlinear (aka threshold) Versus Linear Dose-Response Modeling
The Panel agreed that NTP bioassays appear to demonstrate nonlinear dose response, but
they expressed concern about using animal data to infer slope and dose response for humans.
The Panel pointed out that there are differences in slopes across different bioassays, and
specifically, that some appear linear while others appear nonlinear. Given the observation of
both nonlinear vs. linear, the Panel concluded that neither could be ruled out for extrapolation
below the POD simply based on the available data. One panelist noted that U.S. EPA Cancer
Guidelines (U.S. EPA, 2005) state that only if one can demonstrate that the MOA has a threshold
dose-response shape, and can exclude all other potential linear MO As, can one use a nonlinear
model. Lastly, the Panel noted that there are data and rationales to support use of both linear and
14

-------
nonlinear response below POD. From this discussion, the Panel raised one possibility to the U.S.
EPA:

• Both linear and nonlinear model functions should be considered in the dose-response
analysis.

Dose Metrics
In considering human data, the Panel expressed a preference for lipid-adjusted serum
levels over body burden (BB), and they expressed concerns over the assumptions used in the
back calculation of the BB in the epidemiologic cohorts. In considering the rat data, the Panel
supported the use of BB—especially lipid-adjusted BB. The Panel, however, did express
concern over the sequestering of TCDD in liver and then the use of liver levels in BB
calculations.

Supporting Information—Biologically-Based Dose-Response (BBDR) Models and MOA
The Panel discussed BBDR. Though once considered an attractive proposition, BBDR
models may mask uncertainty within the models, necessitating them to be used with greater
caution. The Panel suggested two issues for the U.S. EPA to consider:

• If there is a published model, use it if it is valid—do not generate a new model.
• Focus on the actual experimental data to drive the analysis.

References

Kociba, R.J., D.G. Keyes, I.E. Beyer et al. 1978. Results of a two-year chronic toxicity and
oncogenicity study of 2,3,7,8-tetrachlorodibenzo-p-dioxin in rats. Toxicol. Appl. Pharmacol.
46:279-303.

NTP (National Toxicology Program). 2006. Toxicology and Carcinogenesis Studies of
2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) (CAS No. 1746-01-6) in Female Harlan Sprague-
Dawley Rats (Gavage Studies). U.S. Department of Health and Human Services. NTP TR 521.
Research Triangle Park, NC (April).

US EPA (U.S. Environmental Protection Agency). 2005. Guidelines for Carcinogen Risk
Assessment. U.S. Environmental Protection Agency Risk Assessment Forum.
EPA/630/P-03/001F.
15

-------
SESSION 4B: DOSE-RESPONSE FOR REPRODUCTIVE/DEVELOPMENTAL
TOXICITY
The U.S. EPA plans to consider development of a quantitative dose-response assessment
for reproductive and developmental effects associated with TCDD exposure. Such an
assessment would be based on information in U.S. EPA (2003), NAS (2006) and key studies
identified in this workshop. The purpose of this session was to identify and discuss key issues
pertaining to dose-response assessment for dioxin-induced reproductive and developmental
effects.

Session 4B Panelists (Session Co-Chairs are identified by asterisk)

• Barbara Abbott, U. S. EPA
• Bruce Allen, Bruce Allen Consulting
• Roger Cooke, Resources for the Future
• George Daston, Procter & Gamble
• MikeDeVito, U.S. EPA
• Rob Goble, Clark University
• *Fumio Matsumura, University of California-Davis
• Paolo Mocarelli, University of Milan
• Brian Petroff, University of Kansas
• *GlennRice, U.S. EPA
• Marie Sweeney, National Institute of Occupational Safety and Health
• Mary Walker, University of New Mexico
• Bernie Weiss, University of Rochester

Key Study Selection
The Panel reviewed the charge questions (Appendix B), discussed them, and listed two
issues for the U.S. EPA to consider:

• Concerning key study determination, use a stepwise approach that is dependent upon the
information available and needed to address the question.
16

-------
• Concerning the key studies informing the POD and the POD endpoint choice, use the
POD to depart from what is certain and use a high-confidence study that has found
effects at a low enough level at which other effects are protected.

The Panel also developed Table 1, based on the information presented in this session. Table 1
identifies specific reproductive and developmental effects of concern, listing whether an effect
has been observed in test animals and epidemiologic cohorts. It also identifies the EDi0
estimated by the U.S. EPA (2003) for health effects observed in rodent bioassays. If the U.S.
EPA did not report an EDi0 for an effect, the table identifies a study where the effect was
reported and the lowest study dose where the effect was observed. Table 1 also identifies the
epidemiologic cohort where the specific reproductive and developmental effects were observed.

Epidemiological Study Utility
The Panel reviewed the charge questions (Appendix B), discussed them, and made two
suggestions to the U.S. EPA:

• Concerning the ability of epidemiological studies to inform critical effects, start with
concordance across species (including humans) for the spectrum of effects.

• Concerning the ability of epidemiological studies to inform dose-response modeling, start
with the epidemiology and then go to animal data if the dose response has not been well
characterized for an endpoint of interest and compare to animal data as a reality check.

Animal Model Utility
The Panel reviewed and discussed the charge questions (Appendix B). Table 1, which
identifies the effects that occur in animals and also have relevance to humans, summarizes much
of this discussion. Regarding the influence of mode of action (MO A) on animal model choice,
the Panel concluded that by evaluating concordance among health effects reported in
epidemiologic and animal bioassay data, the U.S. EPA could identify a set of plausible
reproductive and developmental effects to consider. Actual animal and human MOA
information is helpful in that it creates comfort with the animal models and in defining the
boundaries of possible effects.
17

-------
TABLE 1
Reproductive/Developmental Effects of Concern for Human Health
Endpoint
Sperm Count/Motility
Sex Ratio
Delayed Puberty Males
Delayed Puberty in Females
Cleft Palate
Premature Senescence
Hormones E2
Low Birth Weight
Reproductive Cycling
(prolongation)
Rodent
(EDiong/kg-d)
Yes (6.2-28;
66-200)
No
Yes (94)
Yes
Yes (6300-6400)
Yes
Yes
Yes (190)
Yes
Human
Yes
Yes, Seveso
Yu-cheng
No in Seveso
No
No, Seveso
Yes, Males —
Seveso
Suggestive
effect in Seveso
in first 8 years
after exposure
Yes, Seveso
Prepubertal
exposure
Notes
ED10 bases Mabley et al. (1992a,b) caudal
sperm count and daily sperm production
range from 6.2-28; Gray et al. (1997)
epididymal sperm count and total testis sperm
counts range from 66-200.

ED10 basis rat male puberty delay Gray et al.
(1997). Need to qualify epidemiology data
because of cohort PCDD/PCDFs exposures.
Gray and Ostby (2002) report delayed
puberty in female offspring of pregnant rats
receiving a single dose of 1 ug TCDD/kg on
GD 15.
ED10 basis Birnbaum et al. (1989).
Franczac et al. (2006) report that rats
prematurely entered reproductive senescence,
after receiving cumulative TCDD doses as
low as 1.7 ug TCDD/kg. They considered
first occurrence of prolonged interestrous
interval (>6 d) as evidence of onset of
reproductive senescence.
Li et al (1995) report serum estradiol-17p
(E2) concentrations induced by equine
Chorionic Gonadotropin injection were
significantly elevated in female rats orally
administered 10 ug/kg TCDD onPND 22.
While E2 decreased dramatically in control
animals during the preovulatory LH surge, it
did not in TCDD-treated rats.
ED10 basis Gray et al. (1997).
Franczac et al (2006) report loss of normal
cyclicity in female rats at 8 months of age
following a cumulative dose of 1.7 ug
TCDD/kg.
18

-------
Supporting Information
       The Panel reviewed the charge questions (Appendix B), discussed them, and made two
suggestions to the U.S. EPA:

   •   Concerning deviation from default approaches for noncancer endpoints, there needs to be
       a careful assessment of the POD and the application of uncertainty factors in light of
       PK/pharmacodynamics (PD), population characteristics  and variability, and MOA
       information.

   •   Concerning the MOA's ability to clarify endpoint and the incorporation of a cascade of
       cellular event into dose-response for noncancer endpoint, any study that helps inform the
       dose response should be considered—including studies not specific to dioxins.
       Complicated mechanistic models need not be developed. Standard dose-response models
       can be applied. One can look at the cascade of events in a stepwise,  simple way.

References

Birnbaum, L.S., M.W. Harris, L.M. Stocking et al.  1989. Retinoic acid and 2,3,7,8-
tetrachlorodibenzo-p-dioxin selectively enhance teratogenesis in C57BL/6N mice.  Toxicol.
Appl. Pharmacol. 98:487-500.

Franczak, A., A. Nynca, K.E. Valdez, K.M. Mizinga and B.K. Petroff.  2006.  Effects of acute
and chronic exposure to the aryl hydrocarbon receptor agonist 2,3,7,8-tetrachlorodibenzo-
p-dioxin on the transition to reproductive senescence in female  Sprague-Dawley rats.  Biol.
Reprod. 74:125-130.

Gray, L.E. and J.S. Ostby. 2002.  In utero 2,3,7,8-tetrachlorodibenzo-^-dioxin (TCDD) alters
reproductive morphology and function in female rat offspring.  Toxicol. Appl. Pharmacol.
133(2):285-294.

Gray, L.E., J.S. Ostby and W.R. Kelce.  1997.  A dose-response analysis of the reproductive
effects of a single gestational dose of 2,3,7,8-tetrachlorodibenzo-p-dioxin in male Long Evans
Hooded rat offspring. Toxicol. Appl. Pharmacol.  146:11-20.

Li, X., D.C. Johnson and K.K. Rozman. 1995. Reproductive effects of 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) in female rats: ovulation, hormonal regulation, and possible
mechanism(s). Toxicol. Appl. Pharmacol.  133:321-327.

Mably, T.A., D.L. Bjerke, R.W. Moore et al. 1992a. In utero and lactational exposure of male
rats to 2,3,7,8-tetrachlorodibenzo-p-dioxin. 3. Effects on spermatogenesis and reproductive
capability. Toxicol. Appl. Pharmacol.  114:118-126.
Mably, T.A., R.W. Moore, R.W. Goy et al. 1992b. In utero and lactational exposure of male
rats to 2,3,7,8-tetrachlorodibenzo-p-dioxin. 2. Effects on sexual behavior and the regulation of
luteinizing hormone secretion in adulthood. Toxicol. Appl. Pharmacol.  114:108-117.
                                           19

-------
NAS (National Academy of Sciences). 2006. Health Risks from Dioxin and Related
Compounds: Evaluation of the EPA Reassessment. National Academies Press, Washington, DC
(July). Available at http://www.nap.edu/catalog.php?record_id= 11688.

U.S. EPA (U.S. Environmental Protection Agency). 2003. Exposure and Human Health
Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds. NAS
Review Draft (EPA/600/P-00/001Cb). U.S. Environmental Protection Agency, National Center
for Environmental Assessment, Washington, DC. Available at
http://www.epa.gov/nceawwwl/pdfs/dioxin/nas-review/.
SESSION 5: QUANTITATIVE UNCERTAINTY ANALYSIS OF DOSE-RESPONSE
This session addressed the uncertainty analysis to be considered for the dose-response
assessments. The session opened with a presentation on current estimates of dioxin exposure
levels. Then it focused on the factors to include in the scope of an uncertainty analysis including
dioxin kinetics.

Session 5 Panelists (Session Co-Chairs are identified by asterisk)
• Bruce Allen, Bruce Allen Consulting
• Lesa Aylward, Summit Toxicology
• Roger Cooke, Resources for the Future
• Kenny Crump, Louisiana Tech University
• MikeDeVito, U.S. EPA
• Dale Hattis, Clark University
• *Rick Hertzberg, Biomath Consulting
• Nancy Kerkvliet, Oregon State University
• Leonid Kopylev, U.S. EPA
• Rob McDowell, U.S. Department of Agriculture
• Lorenz Rhomberg, Gradient
• Woody Setzer, U.S. EPA
• Marie Sweeney, National Institute of Occupational Safety and Health
• *Linda Teuschler, U.S. EPA

The Panel summarized the NAS comments regarding uncertainty. Areas for improvement
include:

• Ensure "transparency, thoroughness, and clarity in quantitative uncertainty analysis."

• Describe and define (quantitatively to the extent possible) the variability and uncertainty
for key assumptions used for each key endpoint-specific risk assessment, including
choices of data set, point of departure, dose-response model, and dose metric.

• Incorporate probabilistic models to represent the range of plausible values.
20

-------
• Assess goodness-of-fit of dose-response models.

• Provide upper and lower bounds on central tendency estimates for all statistical estimates.

• When quantification is not possible, clearly state it, and explain what would be required
to achieve quantification.

Identification of Important Uncertainties
The Panel reviewed the charge questions (Appendix B), discussed them, and listed eight
issues for consideration by the U.S. EPA:

• Concerning species and strain differences in the U.S. EPA's Response to NAS, current
U.S. EPA procedures do not take this into account when selecting one data set for risk
assessment. Issues include "Where are humans in the distribution of potencies that can
be generated? How likely is it that human response is similar to the selected data? Can
we infer inter-individual variability from these differences?"

• Concerning the use of animal data for cross species extrapolation to humans (PK and PD
uncertainties), issues to consider include differences in distribution and responses
following bolus doses from those of subchronic and chronic protocols; uncertainty in
liver doses due to sequestration; differences in receptor binding affinity among
congeners; and age factors (e.g., assumption of a lifetime constant daily dose for a cancer
extrapolation).

• Concerning the description of AhR response, biochemical changes occur at lower doses
than toxicological changes. There should be an effort to identify the biochemical changes
that would mark Ah receptor binding to inform the BMR, and, thus, prevent toxicity.

• Concerning model uncertainty, the mathematical model choice depends on endpoint.
There should be an effort towards determining what is the most sensitive endpoint(s) for
humans and conducting animal studies to model that endpoint(s).

• Concerning exposure and dose response in human studies, ensure enough similarity to
current human exposure profiles (mixture composition) so that a dose-response
assessment can be done. Incorporate new epidemiological studies. Evaluate
concordance with animal data and consistency across studies. Panel-acknowledged
uncertainties include exposure estimates from person to person, shape of human dose-
response curve, healthy worker effect, and age dependence.

• Concerning POD determination, uncertainty factors are inherently mathematically
inconsistent and that should be conveyed in the discussion of uncertainties when
interpreting the POD.

• Concerning dose metric, tissue concentration is preferred. It should be evaluated against
a background of variability in AhR-binding expression. There is uncertainty in what
level of binding should be considered, in different cell types, tissues, life stage
(development). The relationship between dose metric and causation of adverse effects
should be examined.
21

-------
Low-Dose Extrapolation
The Panel reviewed the charge questions and discussed them (Appendix B). The Panel
concluded that curve-fitting uncertainty (for a given dataset, dose metric, and model) can be
characterized and is useful, but, by itself, it is an incomplete characterization of uncertainty. The
Panel acknowledged the difficulty of fully characterizing uncertainty, especially quantitatively.
Some panelists argued that the problem is insurmountable and that no meaningful uncertainty
analysis is likely to be performable. Other panelists contended that, the difficulties
notwithstanding, "good-faith" efforts to do something practical and forthright to characterize
uncertainty in low-dose extrapolation would be useful and important. The Panel clarified "good
faith" as meaning a characterization that is useful and not misleading to decision makers and is
inclusive of approaches that have meaningful support in the scientific community as a whole.
Being in "good faith" is more important than being complete (i.e., addressing every uncertain
element), especially since completeness is not a realistic goal. From this discussion, the Panel
listed four issues for consideration by the U.S. EPA:

• Review alternative data sets, dose metrics, and models to see where consequential
uncertainties and impacts on low-dose implications arise.

• Consider the impacts of choices among plausible alternative data sets, dose metrics,
models, and other more qualitative choices—issues include how much difference the
choices make and also how much relative credence should be put to each alternative as a
way of gauging and describing the landscape of imperfect knowledge
regarding possibilities for the true dose-response.
• Hard to do quantitatively, since the factors are not readily expressed as statistical
distributions, but can describe the rationale for believing/doubting each alternative in
terms of available supporting evidence, contrary evidence, and needed assumptions.
• Expert judgment methods may be helpful in characterizing the relative weights of
scientific credibility among alternatives. The expert judgment process, when
conducted systematically, can be thought of as adding data to the assessment of
credibility of alternatives, rather than as just an opinion poll.
• Information on plausibility of alternative low-dose extrapolation approaches can
come from external considerations of mode of action, and not just from statistical
success at fitting particular (high-dose) data sets.

• Characterizing uncertainty through a variety of approaches could be tried, and their
relative merits and shortcomings discussed, as a way forward.

• Consider the sources of potential error, particularly in epidemiological data (e.g., TEF
uncertainty and variation in congener mixtures) and if possible quantify their impact on
the dose-response assessment.

Considerations for Conducting Uncertainty Analysis
Overall, the Panel was split on whether U.S. EPA should do quantitative uncertainty
analyses. The Panel noted that if done on only some of the uncertainties, then results would be
misleading and could be misused. Ultimately, the Panel listed seven issues for consideration by
the U.S. EPA:
22

-------
• The Panel recapped what some consider as being the first integrated risk assessment, with
structured expert judgment and uncertainty analysis, i.e., the Rasmussen Report
(WASH-1400; U.S. Nuclear Regulatory Commission, 1975). In their discussion of the
report, the Panel noted that in addition to standard event tree/fault tree modeling, this
report also tackled difficult model uncertainty issues involved in accident progression,
dispersion of released pollutants in the atmosphere, environmental transport, exposure,
health, and economic impacts. And though the Panel also recognized that this method
was no longer state-of-the-art, the Panel contended that it represents a good example of a
structured approach and methodology that could be built upon.

• The Panel also discussed TEQs used in epidemiological studies, based on intake, and
recognized that the key uncertainty in what was measured was not just intake but also
involved PK/PD issues. The Panel acknowledged that the TEQ system is regularly used
on a concentration basis, but they expressed concern that the qualification becomes lost.
TEQs ignore pharmacokinetics and the common practice of rounding to orders of
magnitude introduces more error.

• Structure the risk assessment along MOA steps—identify key biochemical measures
(-5-10) common across toxic endpoints and identify the degree of meaningful change in
effect or effect variance. Make a table with all options for data set, model, etc.; make
best estimates/choices and determine which of these choices matter the most to the
answer.

• Use expert panels—expert judgment can be collected scientifically (procedures are
published). But there are known biases; central tendency estimates work much better
than extremes.

• Use supporting studies to fill in critical data gaps—Info filling methods do exist (e.g., PK
modeling). Put short-term studies into the "supporting info" category (unless, of course,
the risk assessment is for acute exposures, such as chemical spills).

• Be creative in the analysis of uncertainty. Intermediate steps between AhR binding and
the end processes can be hypothesized based on data, experiences, and analogies related
to other chemicals.

• The 2003 Reassessment presented potency estimates on wide variety of
endpoints/models; needed to be more transparent in that discussion. Statistical graphics
can be used to convey uncertainties.

Reference

U.S. Nuclear Regulatory Commission. 1975. Reactor Safety Study: An Assessment of Accident
Risks in U.S. Commercial Nuclear Power Plants. WASH-1400 (NUREG-75-014). Washington,
DC.
23

-------
APPENDIX A: 2009 U.S. EPA DIOXIN WORKSHOP AGENDA

SCIENTIFIC WORKSHOP
TO INFORM THE TECHNICAL WORK PLAN FOR U.S. EPA'S RESPONSE TO
NAS COMMENTS ON THE HEALTH EFFECTS OF DIOXIN
PRESENTED IN U.S. EPA'S DIOXIN REASSESSMENT

Cincinnati, OH

Date: February 18-20, 2009

BACKGROUND/WORKSHOP OBJECTIVE
At the request of the U.S. Environmental Protection Agency (U.S. EPA), the National
Academy of Sciences (NAS) prepared a report, Health Risks from Dioxin and Related
Compounds: Evaluation of the EPA Reassessment (NAS, 2006), that made a number of
recommendations to improve the U.S. EPA's risk assessment for 2,3,7,8-tetrachlorodibenzo-
/>-dioxin (TCDD). In response, the U.S. EPA will prepare a technical report that addresses key
comments on the dose-response assessment for TCDD. The U.S. EPA intends to develop its
response through a transparent process that provides multiple opportunities for input.

To assist in this effort, a Workshop will be held to inform the U.S. EPA's evaluation of
the NAS recommendations. The Workshop will be open to the public. At the Workshop, the
U.S. EPA will solicit input from expert scientists and the public.

The goal of the Workshop is to ensure that the U.S. EPA's response to the NAS
comments focuses on the key issues and reflects the most meaningful science. The three main
objectives of the Workshop are to (1) identify and discuss the technical challenges involved in
addressing the NAS key comments on the TCDD dose-response assessment in the U.S. EPA
Reassessment (U.S. EPA, 2003), (2) discuss approaches for addressing these comments, and
(3) identify key published, independently peer-reviewed literature, particularly studies describing
epidemiologic and in vivo mammalian bioassays, which are expected to be most useful for
informing the U.S. EPA response.

Workshop participants will be encouraged to think broadly about the body of scientific
information that can be used to inform the U.S. EPA's response and to participate in open
dialogue regarding ways in which the science can best be used to address the key dose-response
issues. This Workshop is similar to scientific workshops being conducted under the new review
process for the National Ambient Air Quality Standards (NAAQS)1 that assess health-related
information for criteria pollutants.
1 Please see http://www.epa.gov/ttn/naaqs/ for more information on the new NAAQS review process.

-------
The Workshop discussions are expected to build upon two prior publications:
1. Exposure and Human Health Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin
(TCDD) and Related Compounds (U.S. EPA, 2003). This external review draft
provides a comprehensive reassessment of dioxin exposure and human health effects.
This "dioxin reassessment" was submitted in October 2004 to the National Academy
of Sciences (NAS) for review.
2. Health Risks from Dioxin and Related Compounds: Evaluation of the EPA
Reassessment (NAS, 2006).

Workshop participants are encouraged to review both of these documents and other
relevant materials (e.g., the National Toxicology Program report on TCDD [NTP, 2006]) before
the meeting because they provide important insights into the key questions and challenges.
There are a number of open comment periods that are intended to facilitate a broad discussion of
the issues.

Scientists with significant expertise and experience relevant to the health effects of
TCDD or dioxin-like compounds and associated topics will be asked to serve on "expert panels"
for discussions throughout the Workshop. Workshop panelists will include a wide range of
experts representing many scientific areas needed to assess TCDD dose-response (e.g.,
epidemiology, human and animal toxicology, nuclear receptor biology, dose-response modeling,
risk assessment, and uncertainty analysis). The Workshop panelists will be asked to highlight
significant and emerging research and to make recommendations to the U.S. EPA regarding the
design and scope of the technical response to NAS comments on the dose-response analysis for
TCDD—including, but not limited to, recommendations for evaluating associated uncertainty.
Open comment periods will follow each panel discussion session. Public participation will be
encouraged by way of these designated open comment periods and, also, by participation in the
scientific poster session planned for the second evening (February 19).

U.S. EPA will use the input received during this Workshop as the foundation for its
development of a technical work plan for responding to the NAS comments on the TCDD dose-
response analysis. The work plan will outline the schedule, process, and approaches for
evaluating the relevant scientific information and addressing the key issues. The work plan also
will identify the key literature to be utilized in U.S. EPA's response.

As a follow-on activity to this Workshop, a panel is being established under the Federal
Advisory Committee Act (FACA) to guide and review the U.S. EPA's response to NAS
comments. The FACA panel will be asked to conduct a consultation with the Agency on the
draft technical work plan. At the same time, the public will also have the opportunity to provide
comments to the FACA panel on the work plan. The final technical work plan will guide the
development of the technical report that will constitute the U.S. EPA's response to NAS
comments. During the development of this response, the U.S. EPA will seek advice from the
FACA panel and the public several times. Finally, the FACA panel will be asked to review the
technical report in a public forum.

The preliminary Agenda presented on the following pages may be revised prior to the
Workshop following review by the session Co-Chairs; the dates and general timing of the
25

-------
sessions, however, will not change. A final Agenda and a set of charge questions, intended to
provide general direction for the Workshop discussions, will be posted on the Workshop Internet
site (http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l99923) prior to the meeting.

       A poster session will be held on the evening of the second day (February  19). The
purpose of this poster session is to provide a forum for scientists to present recent studies
relevant to TCDD dose-response assessment and to encourage open discussion about these
presentations.

REFERENCES
NAS (National Academy of Sciences). 2006. Health Risks from Dioxin and Related
Compounds: Evaluation of the EPA Reassessment. National Academies Press, Washington, DC
(July).  Available at http://www.nap.edu/catalog.php?record_id= 11688.

NTP (National Toxicology Program). 2006. Toxicology and Carcinogenesis Studies of
2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) (CAS No. 1746-01-6) in Female Harlan Sprague-
Dawley Rats (Gavage Studies). U.S. Department of Health and Human Services. NTP TR 521.
Research Triangle Park, NC (April).

U.S. EPA (U.S. Environmental Protection Agency). 2003. Exposure and Human Health
Reassessment of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds, NAS
review draft, Volumes 1-3 (EPA/600/P-00/001Cb, Volume 1). U.S. Environmental Protection
Agency, National Center for Environmental Assessment, Washington, DC (December).
Available at http://www.epa.gov/nceawwwl/pdfs/dioxin/nas-review/.
                                          26

-------
WORKSHOP AGENDA


                                   Davl


8:00-9:00          Registration

9:00-9:30          Welcome/Purpose of Meeting/Document Development Process

9:30-9:45          Panel Comments/Questions on Charge


9:45-2:45        Session 1: Quantitative Dose-Response Modeling Issues
                  (Hall of Mirrors)

     9:45-10:10     Background/Introductory Remarks

     10:10-10:35    TCDD Kinetics: Converting Administered Doses in Animals to
                  Human Body Burdens
                  Presenter: Michael Devito

     10:35-11:30    Panel Discussion

     11:30-1:00     Lunch

     1:00-2:00      Panel Discussion cont.

     2:00-2:45      Open Comment Period


2:45-3:05        Break


3:05-5:15        Session 2: Immunotoxicitv (Hall of Mirrors)

     3:05-3:15      Background/Introductory Remarks

     3:15-4:45      Panel Discussion

     4:45-5:15      Open Comment Period
                                      27

-------
8:00-8:30

    8:00-8:15

    8:15-8:30
                 Day 2

Report-Outs for Sessions 1 and 2 (Hall of Mirrors)

Report-Out for 1: Quantitative Dose-Response Modeling Issues

Report-Out for 2: Immunotoxicity
8:30-11:30

8:30-11:30


    8:30-8:45

    8:45-11:00
8:30-11:30


    8:30-8:45

    8:45-11:00
Sessions 3A and 3B (concurrent sessions)

Session 3A: Dose-Response for Neurotoxicity and
Nonreproductive Endocrine Effects (Hall of Mirrors)

Background/Introductory Remarks

Panel Discussion
     11:00-11:30   Open Comment Period
Session 3B: Dose-Response for Cardiovascular Toxicity and
Hepatotoxicitv (Rookwood Room)

Background/Introductory Remarks

Panel Discussion
     11:00-11:30   Open Comment Period

11:30-1:00       Lunch
1:00-2:00
Report-Outs for Sessions 3A and 3B (Hall of Mirrors)
The structure of the session report-outs will include the following:

      •   Summary of session presentation including minority opinion
      •   Public comments
      •   Discussion

     1:00-1:15     Report-Out for 3A: Dose-Response for Neurotoxicity and
                 Nonreproductive Endocrine Effects

     1:15-1:30     Open Comment Period
                                     28

-------
    1:30-1:45     Report-Out for 3B: Dose-Response for Cardiovascular Toxicity and
                 Hepatotoxicity

    1:45-2:00     Open Comment Period
2:00-5:15        Sessions 4A and 4B (concurrent sessions)

2:00-5:15        Session 4A: Dose-Response for Cancer (Hall of Mirrors)

    2:00-2:15     Background/Introductory Remarks

    2:15-4:45     Panel Discussion

    4:45-5:15     Open Comment Period


2:00-5:15        Session 4B: Dose-Response for
                 Reproductive/Developmental Toxicitv (Rookwood Room)

    2:00-2:15     Background/Introductory Remarks

    2:15-4:45     Panel Discussion

    4:45-5:15     Open Comment Period


6:45-8:15        Poster Session (Rosewood Room)



                                  Day 3


8:30-9:30        Report-Outs for Sessions 4A and 4B (Hall of Mirrors)

    8:30-8:45     Report-Out for 4A: Dose-Response for Cancer

    8:45-9:00     Open Comment Period

    9:00-9:15     Report-Out for 4B: Dose-Response for Reproductive/Developmental
                 Toxicity

    9:15-9:30     Open Comment Period
                                    29

-------
9:30-3:30


     9:30-9:40

     9:40-10:10



10:10-10:30

     10:30-11:30

     11:30-1:00

     1:00-2:15

     2:15-2:30

     2:30-3:00

     3:00-3:15
     3:15-3:30
3:30
Session 5: Quantitative Uncertainty Analysis of Dose-
Response (Hall of Mirrors)

Background/Introductory Remarks

Evidence of a Decline in Background Dioxin Exposures in Americans
Between the 1990s and 2000s
Presenter: Matt Lorber

Break

Panel Discussion

Lunch

Panel Discussion cont.

Break

Open Comment Period

Report-Out for 5: Quantitative Uncertainty Analysis of Dose-
Response

Closing Remarks

Adjourn
                                      30

-------
                 APPENDIX B: 2009 U.S. EPA DIOXIN WORKSHOP
                   QUESTIONS TO GUIDE PANEL DISCUSSIONS
SESSION 1

Dose Metric
Considering all of the endpoints or target tissues, and species that U.S. Environmental Protection
Agency (U.S. EPA)'s dose-response modeling might evaluate, what are the best measures of
dose (e.g., ingested, tissue concentrations, body burden, receptor occupancy, other surrogate) and
why?
Developing Dose-Response Models from Mammalian Bioassays
How best can the point of departure (POD) be determined when the response range is
incompletely characterized (i.e., high response at the lowest dose or low response at the highest
dose; observed in several key 2,3,7,8-Tetrachlorodibenzo-p-Dioxin [TCDD] studies)?

If considered to be biologically plausible, how can a threshold be incorporated into a dose-
response function (e.g., for TCDD cancer data)?

How can nonmonotonic responses be incorporated into the dose-response function?
Developing Dose-Response Models from Epidemiological Studies
How can the epidemiological data be utilized best to inform the TCDD exposure-response
modeling? Which epidemiological studies are most relevant?
Supporting Information
For those toxicological endpoints that are Ah receptor-mediated, how would the receptor kinetics
influence the shape of the dose-response curve? How would downstream cellular events affect
the shape of the dose-response curve? How can this cascade of cellular events be incorporated
into a quantitative model of dose-response?

-------
SESSIONS 2, 3A, 3B, 4A, AND 4B

Key Study Selection
For this endpoint, what refinements should be made to the draft criteria for selection of key
studies?

What are the specific effects of concern for human health for this endpoint?

Based on the draft criteria for the selection of key studies, what are the key studies informing the
shape of the dose-response curve above the POD and the choice of the POD for this endpoint?
Epidemiological Study Utility
How and to what extent do the epidemiological data inform the choice of critical effect?

How can the epidemiological data inform the quantitative dose-response modeling?


Animal Model Utility
Are there types of effects observed in animal models that are more relevant to humans than
others?  To what extent does information on mode of action (MOA) influence the choice of
animal model (species,  strain,  sex)?


Supporting Information

Are there studies that establish a sufficient justification for departure from the default procedures
that address the shape of the dose-response  curve below the POD under the cancer guidelines?

Are there studies that establish a sufficient justification for departing from U.S. EPA's default
approaches for noncancer endpoints?

To what extent can MOA information clarify the identification of endpoints of concern and dose-
response metric for this endpoint? How can the cascade of cellular events for this endpoint be
incorporated into a quantitative model of dose response?
                                           32

-------
SESSION 5

For cancer and noncancer TCDD dose-response assessments, U.S. EPA is interested in
developing a quantitative uncertainty analysis addressing both parameter and model uncertainty,
if feasible. Uncertainties will include, among others, choice of endpoint; underlying study
uncertainties; choice of dose metric; interspecies extrapolations such as kinetic uncertainties; and
choice of dose-response model, including threshold models. The U.S. EPA is currently
examining techniques and tools for uncertainty analysis—including Bayesian and frequentist
approaches.
Identification of Important Uncertainties
What are the major uncertainties pertaining to modeling the animal data?
       Consider the dose metric (species or tissue specificity), vehicle of administration,
       exposure frequency, exposure duration, and POD determination (e.g., benchmark
       response selection or no-observed-adverse-effect level/lowest-observed-adverse-effect
       level identification).

What are the major uncertainties pertaining to dose-response modeling below the POD?
       Consider how receptor kinetics and downstream cellular event information might be used
       to bound the uncertainties associated with dose-response modeling below the POD.

What are the major uncertainties in cross-species extrapolation (e.g., half-lives, tissue
distribution, and toxicodynamics)?
       Consider the primary species dosed with  TCDD: mice, hamsters, rats, guinea pigs, and
       monkeys.

What are the major uncertainties pertaining to intrahuman variability?
       Consider what data sets would be useful to represent sensitive subpopulations.

What are other significant sources of uncertainty for the cancer and noncancer assessments?
Considerations for Conducting Uncertainty Analysis
What data sets could be used to quantify uncertainties in cancer and noncancer TCDD dose-
response assessments?
       Consider dioxin-like compound dose-response data.
       Consider MOA information.

What are the appropriate techniques for the TCDD dose-response uncertainty analysis, and what
are their respective strengths and weaknesses of these approaches as applied to TCDD?
                                           33

-------
3 February 2009
     APPENDIX C: 2009 U.S. EPA DIOXIN WORKSHOP DRAFT SELECTION CRITERIA TO IDENTIFY KEY IN VIVO
     MAMMALIAN STUDIES THAT INFORM DOSE-RESPONSE MODELING FOR 2,3,7,8-TETRACHLORODIBENZO-
                                                          jo-DIOXIN (TCDD)a
Study Feature
Chemical, purity,
matrix/medium
Peer review
Study design,
execution, and
reporting
Study subject:
species, strain, and
sensitivity forgiven
endpoint; litter; life
stage; gender
Exposure route
Dose level
Exposure frequency,
duration, and timing
Controls
Response
Statistical evaluation
Selection Rationale
Primary13
TCDD-only doses included, purity specified,
matrix in which TCDD is administered is identified
Independently peer-reviewed, publicly available
Clearly documented and consistent with standard
toxicological principles, testing protocols,
and practice (i.e., endpoint-appropriate,
particularly for negative findings)
Mammalian species
Strain and gender identified
Animal age at beginning of treatment identified
Litter confounders (within/between) accounted for
Oral
Lowest dose <200 ng/kg-d for noncancer
endpoints and <1 ug/kg-d for cancer
Dosing regimen characterized and explained
Appropriate and well characterized
Effect relevant to human health
Magnitude outside range of normal variability
Clearly described and appropriate to the endpoint
and study design (e.g., per error variance,
magnitude of effect)
Secondary0
TCDD purity or matrix not clearly identified
Supplementary materials accompanying
peer-reviewed publication
Testing protocol provides incomplete
coverage of relevant endpoint-specific
measures, particularly for negative
findings
Mammalian species, in vivo, but only
studying an artificially sensitive subject
(e.g., knockout mouse)
Parenteral (e.g., intravenous, intramuscular,
intraperitoneal, subcutaneous)
Lowest dose >200 ng/kg-d for noncancer
endpoints, or >1 .0 ug/kg-d for cancer

Effect reported, but with no negative control
Precursor effects, or adaptive responses
potentially relevant to human health
Limited statistical context
Currently Excluded
Studies of dioxin-like compounds
(DLCs) or mixtures
Not formally peer-reviewed; literature
not publicly available
Studies not meeting standard
principles and practices
Non-mammalian or not in vivo
Inhalation, dermal, ocular

Characterization/explanation missing
or cannot be determined

Lethality

a NAS (2006) commented that the selection of data sets for quantitative dose-response modeling needed to be more transparent. These draft criteria are
  offered for consideration at the kickoff workshop. These criteria would be used to identify candidate studies of non-human mammals that would be used to
  define the point-of-departure (POD). These criteria are not designed for hazard identification or weight-of-evidence determinations. Studies addressing
  data other than direct TCDD dose-response in mammals (including toxicokinetic data on absorption, distribution, metabolism, or elimination; information
  on physiologically-based pharmacokinetic [PBPK] modeling, and mode of action data) will be evaluated separately.
b Presents preliminary draft criteria for evaluating a study being considered for estimating a POD in a TCDD dose-response model.
c Presents preliminary draft criteria that could qualify  a study as primary with support from other lines of evidence (e.g., PBPK modeling), when no study for
  an endpoint meets the "primary" criteria.
                                                                   34

-------