U.S. EPA

1
U.S. EPAs Dioxin ReassessmentThe Current
State Of Scientific Understanding
William H. Farland, Ph.D. Acting Deputy
Assistant Administrator for Science Office of
Research and Development U.S. Environmental
Protection Agency 6/28/04
2
The Concept of Toxicity Equivalence (TEQ)...
Total TEQ Ccongener X TEFcongener

Toxicity Equivalence of
CDDs/CDFs/PCBs

3
Four of 17 Toxic CDD/CDF Congeners and One of
the 11 Toxic PCBs Account for Most of the
Toxicity in Human Tissue Concentrations

• These five compounds make up about 80 of the
  total TEQ in human tissue

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

4
Sources and Pathways to Human Exposures

• Sources
• Combustion
• Metal Smelting, Refining, Processing
• Chemical manufacturing
• Biological and Photochemical Processes
• Reservoir sources
• Others?
• Pathways
• Ingestion of soil, meats, dairy products, fish
• Inhalation of vapors and particulates
• Dermal contact with soils
• For more details, see EPA Inventory

5
Key Findings of the ReassessmentExposure
Document

• Environmental levels have declined since the 70s
• Current US regulatory efforts have addressed most
  of the known large industrial sources (80
  reduction between 87 and 95 further reductions
  anticipated)
• Open burning of household wastes is the biggest
  unaddressed contemporary source identified so
  far.
• There remain many uncharacterized sources that
  could be significant (ag. burning, ceramics,
  forest fires, secondary steel, reservoir sources)
  
• Exposure to general population has declined but
  currently averages 1pg/kg/day

6
Adult Average Daily Intake of CDDs/CDFs/dioxin-lik
e PCBs
2000 Draft Estimate 65 pg TEQDFP-WHO98/day
Soil ingestion
Vegetable fat
Soil dermal contact
Other meats
Poultry
Freshwater fish and
shellfish
6
Pork
5
19
Marine fish and shellfish
Beef
7
14
1
Inhalation
4
16
Eggs
21
Milk
Dairy
7
Key Findings of the Reassessment Exposure
Document (continued)

• General Population Exposure is from animal fats
  in the commercial food supply
• -- Local sources make little contribution to
  most peoples exposure
• -- Environmental levels in meat dairy
  production areas major contributor
• Air deposition onto plants consumed by domestic
  meat and dairy animals is the principal route for
  contamination of commercial food supply
• Reservoir sources are a significant component of
  current exposure and may dominate future exposure
• -- accounts for most coplanar PCB exposure
• -- unknown contribution for D/Fs
• Special populations may be more exposed but
  prevalence is not well substantiated

8
Dioxin Exposure Variability

• Dioxin intakes for general population may reach
  levels at least 3 times the mean
• Support
• ? 99th percentile total animal fat consumption
  is 3 times the mean
• ? 99th percentile blood level is 3 times the
  mean
• Potentially highly exposed populations may exceed
  this range
• ? Nursing infants
• ? Subsistence fishers/farmers in contaminated
  areas
• ? Occupational groups (historical)

9
Dioxin Exposure Trends

• Environmental levels
• Peaked in late 60s/early 70s declined since
  based on sediment data
• Decline also supported by Emissions Inventory
  which shows significant decrease from 1987 to
  1995 (80)
• Human tissue data suggest current levels are
  about half of 1980 levels (55 to 25 pg TEQDFP/g
  lipid)
• Steady state PK modeling of current intake
  levels project tissue levels of about 11 pg
  TEQDFP/g lipid.

10
Effects of Dioxins

• Multiple effects in multiple tissues of both
  sexes of multiple species throughout the
  vertebrate kingdom

11
Key Findings of the Reassessment Health Document

• Toxic equivalents (TEQ) provide the best means
  for evaluating mixtures
• -- Use WHO98 TEFs
• -- Include coplanar PCBs
• Body burden is the best dose metric for
  estimating risk
• Environmental mixtures of dioxin-like compounds
  are likely to be carcinogenic to humans and
  2,3,7,8-TCDD is carcinogenic to humans.
• Dioxins produce a variety of noncancer effects in
  animals humans
• -- Developmental Toxicity
• -- Immunotoxicity
• -- Endocrine Effects
• -- Chloracne
• -- Others

12
Dioxin-like Compounds and Human Carcinogenicity
Complex Environmental ? Likely to be
Mixtures carcinogenic 2,3,7,8-TCDD
? Carcinogenic to humans Other dioxin-like ?
Likely to be compounds carcinogenic
Based on

• Unequivocal animal carcinogen
• Limited human information (epidemiological/other)
  
• Mechanistic plausibility

Cancer potency increasingly focusing on human
studies
Note In February 1997, the International Agency
for Research on Cancer (IARC) classified TCDD
as a Category 1, Known human carcinogen. In
1999, the U.S. DHHS 9th Report on Carcinogens
(ROC) proposed the same, and finalized this
listing in 2001 (9th ROC, as revised).
13
Summary of All Site Cancer ED01s and Slope
Factor Calculations from Published Studies
ED01/LED01 (95 lower bound) ng/kg
Slope factor1, All cancer risk/pg/kg/day
STUDY
Hamburg cohort, Becher et al. 1998 NIOSH
cohort, Steenland et al. 20012 BASF cohort, Ott
and Zober, 1996 Sprague-Dawley rats, Kociba et
al. 1978 Goodman and Sutter, 1992 (pathology)
6.0 32.2 0.9 5.1 x 10-3
18.6 (11.5) 1.5 x10-3 (2.5 x10-3)
50.9 (25.0) 0.57 x10-3 (1.2 x 10-3)
31.9 (22) 3 0.97 x10-3(1.4 x
10-3) BMD dose
38 (27.5) 0.80 x10-3(1.1 x 10-3) BMD dose
Upper bound estimates in parentheses
See Footnotes next slide
14
Cancer Dose Response - Foot Notes
1. Assumes 25 of body weight is lipid 80 of
dioxin dose is absorbed from the normal diet in
humans the TCDD half-life is 7.1 years in
humans. Background all cancer mortality rate
calculated through lifetable analysis to 75
years. Summary results are for male all cancer
risk, because the male lifetime (to 75 years) all
cancer risk is greater than for females, leading
to correspondingly higher cancer slope factors.
As detailed in Chapter 8, RelRisk(ED01) 0.99
0.01/Risk(0 dose). Based on the manner in which
the dose-response data were calculated using Cox
Regression rate ratio analyses, risks are given
as cancer slope factors for 1 pg/kg/day above
background, assumed 5 ppt TCDD in lipid. 2.
Steenland et al. (2001) power model results are
not included as this formula predicts
unreasonably high attributable risks at
background dioxin levels in the community due to
the steep slope of the power curve formula at
very low levels. 3. Modeled using U.S. EPA
benchmark dose software, version 1.2, with either
dose or adipose concentration as the metric. 50
absorption assumed from food pellets in animals.
BMD 0.00176849 ug/kg/day. BMDL 0.00122517
ug/kg/day. Therefore, rat LED01 1.2251 x 25 x
0.5/ln2 22 ng/kg human equivalent LED01 22 x
ln2 x 1000/2593/0.8 7.38 pg/kg/day slope
factor 0.01/7.38 1.4 E-3 risk/pg/kg/day
15
Non-cancer Dose/Response Modeling

• Use body burden (Ng/Kg BW) as dose
  metric
• Accounts for differences in half-life
• Shape of dose/response curve
• Most toxic responses are non-linear
• Most biochemical response and 40 of complex
  responses are linear
• Derive ED01/ED10 within/close to experimental
  data

16
Body Burdens (Ng/Kg) Associated With Effects
MOE I 4 I 0.1 - 8 I 10 I 0.3 - 2 I 4 I
0.6 - 16 I I 0.1 - 7 I 0.4 - 17 I I 2

• Adverse Effects
• Developmental neurotoxicity 22
• Developmental/reproductive toxicity 0.7 - 42
• Developmental immunotoxicity 50
• Adult immunotoxicity 1.6 - 12
• Endometriosis 22
• Cancer 3.3 - 80
• Biochemical Effects
• CYP1A1 Induction 0.6 - 33
• CYP1A2 Induction 2.1 - 83
• Functional Effects
• Oxidative stress 10

MOE effect level / current average U.S.
background body burdens of 5 Ng/Kg
17
US/International Comparisons
BODY BURDEN (LOAEL/ NOAEL) EFFECT SAFETY/ UNCERT. FACTOR GUIDANCE DAILY INTAKE (pg/kg/day)
WHO 1997 10-40 ng/kg Several 10 TDI 1-4
ATSDR1999 32 ng/kg Neuro. Dev. 90 MRL 1
JECFA 2001 13/25 ng/kg Devel. 3.2/9.6 TMI 2.3
Body burden from original publication ATSDR
used intake of 0.12 ng/kg/day Based on TMI
70 pg/kg
18
Similarities/Differences with EPA
Similarities

• Focus on lowest adverse effects
• Use body burden as dose metric (except ATSDR)
• Suggest additional decrease in intake necessary
• Assume cancer will be insignificant at guidance
  level
• Choose safety/uncertainty factor (3.2-90) to
  account for NOAEL/LOAEL, PD, human variability
• Choose safety assessment vs. MOE/QRA

Differences
19
Key Findings of the Reassessment Risk
Characterization

• Cancer slope factor is based primarily on
  recently published analyses of human studies and
  is revised upward by a factor of 6 over the 1985
  EPA value based on 1978 study in rats.
• Cancer risks to the general population may exceed
  10-3 (1 in 1,000) from background (dietary)
  exposure but are likely to be less and may even
  be zero for some individuals
• Adverse non-cancer effects have been observed in
  animal and humans within 10 times background
  exposure. It is likely that part of the general
  population is at, or near, exposure levels where
  adverse effects can be anticipated

20
Summary

• Dioxin science has evolved rapidly more data
  lead to better understanding, but more questions.
• Expanded human data on cancer reinforces our
  previous concern for the potential for human
  health impacts.
• Identification of non-cancer effects in animals
  and human are sufficient to generate a similar
  level of concern to cancer
• Environmental levels and human exposure are
  declining but are still at a level of concern
• Current source characterization is complex with
  uncontrolled burning and reservoir sources
  potentially playing a significant role.

21
SAB Report

• The SAB Report (May 31, 2001)
• Commends the Agency on providing a careful and
  thorough review of the dioxin literature
• Suggests improvements in the following areas
• More focus on non-cancer effects
• Increased emphasis on Mode of Action
• More clarification of uncertainty
• Discusses lack of SAB panel consensus on several
  key dioxin science issues
• Cancer Characterization Carcinogen vs. Likely
  Carcinogen?
• Margin of Exposure (MOE) and/or Reference Dose
  (RfD)?
• Upper Bound Estimate of Cancer Risk?
• Recommends that the Agency expeditiously move
  forward with finalization of EPAs Dioxin
  Reassessment

(On the Internet at www.epa.gov/science1/fiscal01.
htm)
22
Major Issues Identified in SAB/Public Comments
Addressed

• Sparse data to derive national means
• for sources/pathways
• Lack of information on dioxin-like PCBs
• in exposure document
• State of validation of exposure models
• Trends in environmental/body burden levels
• TEFs/TEQs
• Impact of human data on hazard and
• risk characterization
• Significance of enzyme induction and other
• biochemical effects
• Relative roles of data, scientific
  inference, and science policy

23
GAO Report

• In May 2001, U.S. Senators John Breaux and Thad
  Cochran requested that the Congresss General
  Accounting Office (GAO) conduct a study on the
  draft dioxin reassessment that focuses on
• EPAs discussion and characterization of dietary
  exposure risks
• Comparisons between the World Health
  Organizations and EPAs approaches to dioxin
  exposure and estimates of dietary levels and
• Responsiveness to external peer review and other
  comments in the draft reassessment.

24
GAO Report (cont.)

• Major findings in the report
• Positive comments regarding development process,
  quality, and inclusiveness of the Agencys draft
  reassessment
• EPA's work largely reflects the recommendations
  and suggestions provided to the agency by the two
  most recent independent peer review panels.
• GAO commented that the data on dioxin levels in
  food supporting EPAs October 2000 draft
  reassessment were a significant improvement over
  the 1994 draft although citing some aspects and
  limitations of EPAs dietary exposure discussion.
  GAO raised the issue of the currency of the
  dietary survey data referenced in the Agencys
  reassessment.

25
GAO Report (cont.)

• Major findings in the report (cont.)
• Commented that the TEF approach is the
  internationally accepted scientific method for
  risk assessments of dioxins
• Commented that EPA and the scientific community
  conclude that dioxins can cause a variety of
  cancer and noncancer health effects that they
  act in the same way in animals and humans and
  that some effects could occur at or near the
  levels to which the general population is now
  being exposed
• GAO noted that EPA and WHO used different
  approaches to estimating human safety/risk.

26
Status and Next Steps

• Science Advisory Board (SAB) review May 1995
• Report received from the SAB Fall 1995
• Major SAB comments -- revision and re-review of
  Chapter 8 Dose-Response Modeling and Risk
  Characterization add TEF Chapter
• Internal and Inter-Agency Review of TEF Chapter
  and revised Integrated Summary and Risk
  Characterization
• SAB re-review of revised D/R and TEF Chapter and
  Integrated Summary and Risk Characterization
  -- November 1 and 2, 2000
• SAB/Executive Committee review of Nov. meeting
  draft report and letter to Administrator --
  May 31, 2001
• Internal and Inter-Agency Review of D/R and TEF
  Chapter and revised Integrated Summary and Risk
  Characterization -- Oct. 2003
• National Academy of Sciences Review of Selected
  Issues 2004-2005

27
NAS REVIEW

• The National Academies Research Council will
  review EPAs 2004 draft to assess whether EPAs
  risk estimates are scientifically robust and
  whether there is a clear delineation of all
  substantial uncertainties and variability.
• Focus on
• cancer characterization
• modeling assumptions, including use of TEFs
• dose ranges and associated likelihood estimates
  for identified human health outcomes
• selection of studies as the basis of the
  assessment
• quantitative approaches to uncertainty analysis.