Environmental Health Risks: Information on EPA's Draft		 
Reassessment of Dioxins (26-APR-02, GAO-02-515).		 
                                                                 
Dioxins--chemical compounds that share structural and biological 
characteristics--have been linked to human illnesses, including  
cancer. Often the byproducts of combustion and industrial	 
processes, complex mixtures of dioxins enter the food chain and  
human diet through emissions into the air. The Environmental	 
Protection Agency (EPA) and the World Health Organization (WHO)  
noted the potential human health risks of dioxins in the 1970s	 
when animal studies showed them to be among the most potent	 
cancer-causing chemicals.  EPA derived its estimates of human	 
dietary exposure to dioxins in the United States from (1)	 
chemically analyzed samples of 10 food types, (2) toxicity	 
estimates of levels of individual dioxins in these foods, and (3)
estimates of the quantities of these foods consumed by Americans.
To develop more reliable national estimates of dietary exposure, 
EPA incorporated into its analysis some food studies that were	 
nationally representative. Although both EPA and the WHO have	 
assessed the human health risks of dioxins during the last	 
decade, some of their objectives and processes have differed.	 
Nonetheless, the analytical methods used and the conclusions	 
reached have much in common. A major difference in the		 
assessments is whether there are threshold levels below which	 
exposure to dioxins would pose a negligible risk of cancer. EPA  
assumed there is no safe threshold level for cancer effects, but 
the WHO assumed there is. EPA's draft reassessment report	 
reflects the recommendations and suggestions provided to the	 
agency by the two most recent independent peer review panels. The
panels, one consisting of 12 independent reviewers and the other 
convened by EPA's Science Advisory Board, concurred with many key
assumptions and approaches that EPA used.			 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-02-515 					        
    ACCNO:   A03172						        
  TITLE:     Environmental Health Risks: Information on EPA's Draft   
Reassessment of Dioxins 					 
     DATE:   04/26/2002 
  SUBJECT:   Air pollution					 
	     Cancer						 
	     Health hazards					 
	     Chemical and biological agents			 
	     Contaminated foods 				 
	     Toxic substances					 
	     EPA Environmental Radiation Ambient		 
	     Monitoring System					 
                                                                 
	     WHO International Program on Chemical		 
	     Safety						 
                                                                 


******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-02-515
     
Report to Congressional Requesters

United States General Accounting Office

GAO

April 2002 ENVIRONMENTAL HEALTH RISKS

Information on EPA?s Draft Reassessment of Dioxins

GAO- 02- 515

Page i GAO- 02- 515 Draft Reassessment of Dioxins Letter 1

Results in Brief 2 Background 4 EPA Used Data on the Presence of Dioxins,
Toxicity, and Food

Consumption to Estimate Human Dietary Exposure 7 EPA?s and WHO?s Specific
Reassessment Objectives and Processes

Differed, but Their Analytical Methods and Conclusions on Dioxins? Health
Risks Are Similar 23 EPA?s Draft Dioxin Reassessment Report Generally
Reflects the

Views of Recent Peer Reviews 32 Observations 37 Agency Comments and Our
Response 38 Scope and Methodology 38

Appendix I Major Milestones in the EPA and WHO Dioxin Risk Assessment
Efforts 41

Appendix II Comparison of the Major Conclusions from EPA?s and WHO?s Dioxin
Risk Assessments 43

Appendix III Questions EPA Asked Peer Review Panels to Address 48

Appendix IV EPA?s Responses to Peer Review Panels 52

Appendix V Comments from the Environmental Protection Agency 57

Appendix VI GAO Contacts and Staff Acknowledgments 67 Contents

Page ii GAO- 02- 515 Draft Reassessment of Dioxins Tables

Table 1: EPA?s Estimates of the Average U. S. Adult?s Daily Exposure to
Dioxins From Dietary Intake, Picograms per Day 9 Table 2: Numbers, Types,
and Dates of Food Samples EPA Used in

Estimating Dietary Exposure to Dioxins 12 Table 3: Numbers, Types, and Dates
of Food Samples EPA Used in

Estimating Dietary Exposure to PCBs in Four Food Categories 13 Table 4:
EPA?s Estimates of Toxic Concentrations of Dioxins in 10

Food Categories, Picograms per Gram 20 Table 5: Estimated Daily Dietary
Intake of 10 Food Types for an

American Adult Weighing 70 Kilograms (154 pounds) 22 Table 6: Questions for
the July 2000 Panel Review of EPA?s Draft

Dioxin Reassessment 48 Table 7: Questions for the November 2000 Science
Advisory Board

Panel 49 Table 8: EPA?s Responses to July 2000 Panel?s Report 52 Table 9:
EPA?s Responses to Science Advisory Board Panel?s

Comments 54

Abbreviations

CDD polychlorinated dibenzo- p- dioxins CDF polychlorinated dibenzofurans ED
effective dose EPA Environmental Protection Agency IARC International Agency
for Research on Cancer PCB polychlorinated biphenyls TCDD 2,3,7,8-
tetrachlorodibenzo- p- dioxin TEF toxic equivalency factors TEQ toxic
equivalency USDA Department of Agriculture WHO World Health Organization

Page 1 GAO- 02- 515 Draft Reassessment of Dioxins

April 26, 2002 The Honorable John Breaux The Honorable Thad Cochran United
States Senate

Some dioxins, which are chemical compounds that share certain structural and
biological characteristics, have been linked to adverse human health
effects, including cancer. 1 Often the byproducts of combustion and
industrial processes, complex mixtures of dioxins enter the food chain and
human diet through emissions into the air that settle on soil, plants, and
water. The Environmental Protection Agency (EPA) and other entities, such as
the World Health Organization, began assessing the potential human health
risks of dioxins in the 1970s, when animal studies on one of them- 2,3,7,8-
tetrachlorodibenzo- p- dioxin, or TCDD- showed it to be the most potent
cancer- causing chemical studied to date. EPA?s initial assessment of
dioxins was published in 1985. Since that time, there have been major
advances in the scientific understanding of dioxin toxicity and significant
new studies on dioxins? potential adverse health effects. As a result, in
1991 EPA decided to conduct a reassessment of the health risks of exposure
to dioxins. A draft of this reassessment was reviewed by a scientific peer
review panel in 1995, and three panels reviewed key segments of later drafts
in 1997 and 2000.

EPA plans to release its comprehensive reassessment report on the health
risks of dioxins this year. According to EPA officials, the report will
conclude that dioxins may adversely affect human health at lower exposure
levels than previously thought and that most exposure to dioxins occurs from
eating such American dietary staples as meats, fish, and dairy products,
which contain minute traces of dioxins. These foods contain dioxins because
animals eat plants and commercial feed, and drink water, contaminated with
dioxins, which then accumulate in animals? fatty tissue. EPA plans to use
its reassessment of the risks posed by dioxins to develop a risk management
strategy to address the health risks identified and to determine whether the
nation?s current air, water, and hazardous waste cleanup programs need to be
changed to protect the public health. EPA?s

1 In this report, unless otherwise indicated, we use the term ?dioxins? to
refer to the three closely related families of chemical compounds (dioxins,
furans, and polychlorinated biphenyls) that EPA evaluates in its
reassessment of dioxins. In the scientific literature, these compounds may
be referred to as ?dioxins and dioxin- like compounds.?

United States General Accounting Office Washington, DC 20548

Page 2 GAO- 02- 515 Draft Reassessment of Dioxins

reassessment will be considered by the National Academies 2 in an ongoing
study of the implications of dioxin in the food supply, which is examining,
among other things, options to reduce dietary exposure to dioxins.

Of the several hundred known dioxins, 29 are considered toxic to varying
degrees. TCDD is the most widely studied dioxin and one of the most toxic.
EPA?s reassessment report on the human health risks posed by dietary
exposure to dioxins evaluates the health effects of TCDD and the 28 other
compounds with similar structural and biological characteristics and varying
toxic effects. According to EPA, its evaluation of the effects of these
compounds is sufficient to characterize the effects of environmental dioxins
in general.

Concerned about the potentially significant impact that EPA?s dioxin risk
assessment report could have on consumers and on the food and agriculture
industries, you asked us to examine several aspects of EPA?s reassessment of
dioxins. As agreed with your offices, this report describes (1) the data EPA
used to estimate human dietary exposure to dioxins in the United States; (2)
how EPA?s reassessment objectives, processes, analytical methods, and
conclusions on the health risks posed by dioxins compare with those of the
World Health Organization; and (3) the extent to which the draft dioxin
reassessment report reflects the views of independent peer review panels
that reviewed key aspects of the reassessment. Also as agreed with your
offices, our report provides information on the relevant scientific issues
but does not render an opinion on the scientific merits of the reassessment.
This report is based primarily on EPA?s draft reassessment report dated
October 2001, 3 which EPA circulated for internal agency review, and on the
two most recent peer reviews of key segments of the draft reassessment in
2000.

EPA derived its estimates of human dietary exposure to dioxins in the United
States from (1) various studies that chemically analyzed samples of 10 food
types, (2) toxicity estimates of the various levels of the individual

2 The National Academies consist of four organizations: the National Academy
of Sciences, the National Academy of Engineering, the Institute of Medicine,
and the National Research Council.

3 The September 2000 draft reassessment report is posted on EPA?s Web page.
While the October 2001 draft reflects a number of revisions to the September
2000 version on the Web page in response to public and peer review comments,
the drafts are substantially the same. Results in Brief

Page 3 GAO- 02- 515 Draft Reassessment of Dioxins

dioxins in these foods, and (3) estimates of the quantities of these foods
consumed by Americans. To develop more reliable national estimates of
dietary exposure, EPA incorporated into its analysis some food studies that
were designed to be nationally representative. However, as EPA notes in its
draft reassessment report, the food data were limited in several ways. In
some cases, the food sampling methods, or the number of samples collected,
were not sufficient to reliably estimate average, nationally representative
exposures. In other cases, the studies did not analyze the food samples for
the presence of all the dioxins that EPA was assessing. Further, most of the
samples were collected 5 or more years ago; therefore, they may not reflect
current exposures if, as EPA believes, emissions of dioxins have continued
to decline in the United States since 1995 because of air quality
regulations. Nonetheless, EPA believes that its estimate of average dietary
exposure to dioxins is a reasonable characterization of current exposure
because, for example, the emission reductions that have occurred since most
of the food samples were collected are not believed to be as significant as
earlier emission reductions. Regarding toxicity estimates, because
sufficient data are not available on many of the individual dioxins, EPA
used an approach that relies on data developed by the World Health
Organization to estimate the toxicity of the various mixtures of dioxins
identified in the 10 types of foods. Although this approach may overstate or
understate the concentrations of dioxins in the foods, it is the
internationally accepted scientific method for risk assessments of dioxins.

While both EPA and the World Health Organization have taken steps during the
past decade directed at the general objective of assessing the human health
risks of dioxins, some of their specific objectives and processes have
differed. Nonetheless, the analytical methods the organizations used and the
conclusions they reached have much in common. EPA established a long- range
objective of characterizing the potential human health risks posed by
exposure to dioxins using a comprehensive, multiyear review process
resulting in a reassessment report. In contrast, the World Health
Organization conducted a series of individual reviews with more narrowly
focused primary objectives, such as updating the estimated amount of dioxins
to which a person could be exposed daily for a lifetime without appreciable
health consequences. Regarding analytical methods and conclusions, both EPA
and the World Health Organization:

 Examined similar sets of human and animal study data, considered a similar
range of health effects, and applied some analytical concepts that

Page 4 GAO- 02- 515 Draft Reassessment of Dioxins

both entities determined were more appropriate to the assessment of dioxins
than those often used in assessments of other chemicals.

 Concluded that dioxins could adversely affect human health at lower
exposure levels than previously thought and that some adverse noncancer
effects, such as reproductive and developmental impairments, could occur at
or near the levels to which the general population is now being exposed.

A major difference in the organizations? assessments concerns whether there
are threshold levels below which exposure to dioxins would pose a negligible
risk of cancer. While EPA assumed there is no safe threshold level for
cancer effects, the World Health Organization assumed there is.

EPA?s draft reassessment report largely reflects the recommendations and
suggestions provided to the agency by the two most recent independent peer
review panels, although some areas of disagreement on key scientific issues
remain. The panels, one consisting of 12 independent reviewers and the other
convened by EPA?s Science Advisory Board, concurred with many key
assumptions and approaches that EPA used. In addition, the panels made
recommendations on several issues and provided suggestions for EPA to
consider. EPA generally addressed the panels? recommendations and
suggestions by, for example, performing additional analyses or explaining
that the data currently available are not yet sufficient to address the
recommendation or suggestion. If EPA disagreed with the panels?
recommendations or suggestions, it explained its position in the text.
Additional changes are being made as EPA prepares the draft for external
interagency review. Lack of consensus on some scientific issues, such as
whether the weight of evidence supports EPA?s classification of TCCD as a
human carcinogen, reflects uncertainty in areas where data are limited.
Accordingly, the Science Advisory Board views this reassessment report as an
interim evaluation that will need to be updated and peer reviewed in the
future as important data gaps are addressed.

Dioxins persist for a long time in the environment because they do not
dissolve in water and are relatively immobile in soil and sediment. When
animals consume plants, feed, and water contaminated with dioxins, they
accumulate in the animals? fatty tissue. Similarly, when humans consume
these animals, the dioxins then accumulate in human fatty tissue. According
to EPA, because dioxins also persist in the body for years, recent
significant reductions in dioxin emissions into the air are unlikely to
reduce human health risks in the near term. Background

Page 5 GAO- 02- 515 Draft Reassessment of Dioxins

While EPA estimates that most exposure to dioxins occurs from eating
commonly consumed foods, the draft reassessment report also estimates that
limited exposure to dioxins results from breathing air containing trace
amounts of dioxins; inadvertently ingesting soil containing dioxins; and
absorbing through the skin minute levels of dioxins present in the soil.
Some people may experience higher exposure levels than the general
population as a result of food contamination incidents; workplace exposures;
industrial accidents; or consuming unusually high levels of fish, meat, or
dairy products. When calculating human exposures, dioxins are measured in
picograms- that is, trillionths (0.000000000001) of a gram. Highly
sophisticated measurement techniques and technologies are required to test
foods for the presence of the 29 dioxins identified as having toxic effects.

The several hundred known dioxin compounds can be placed in one of three
closely related families: polychlorinated dibenzo- p- dioxins (CDD),
polychlorinated dibenzofurans (CDF), and polychlorinated biphenyls (PCBs).
CDDs and CDFs are byproducts of combustion and some industrial processes.
According to EPA, U. S. emissions of CDDs and CDFs into the environment
declined by 75 percent between 1987 and 1995 primarily as a result of
reductions in emissions from municipal and medical waste incinerators. Some
PCBs share certain characteristics with CDDs and CDFs and therefore are
identified as ?dioxin- like.? PCBs were at one time manufactured for use in
products such as lubricants and industrial transformers but have not been
made in the United States since 1977. However, because dioxins break down so
slowly, past emissions remain in the environment for years- even decades-
before they diminish. Consequently, a large part of humans? current exposure
to dioxins is due to releases of dioxins that were stored in soil and
sediment, and to a lesser extent in vegetation and the atmosphere. These
sources are called ?reservoir sources.? EPA believes that with the reduction
in current emissions from combustion and incineration, these reservoir
sources have taken on more significance.

According to EPA, dioxins always occur in the environment and in humans as
complex mixtures of individual compounds. However, the complex nature of the
dioxin mixtures to which people are exposed (through foods or other sources)
complicates evaluation of the health risks such mixtures might pose.
Scientists therefore developed the concept of toxic equivalency factors
(TEFs) to facilitate risk assessment of exposure to these mixtures. Because
TCDD is the best- understood dioxin, it is used as a frame of reference for
estimating the toxicity of the other dioxins, and its TEF is set at 1.0.
Only 1 of the other 28 dioxins included in

Page 6 GAO- 02- 515 Draft Reassessment of Dioxins

EPA?s reassessment has a TEF of 1.0; most of the others have TEFs of 0.1 or
less, meaning that they are considered less toxic to humans than TCDD.
International experts review and periodically update the TEFs based on new
data. For its reassessment of dioxins, EPA used the latest revisions that
were made at an expert meeting organized by the World Health Organization in
1997. 4

Since 1991, EPA has been updating its initial 1985 report assessing the
health risks of dioxins. The October 2001 draft reassessment report exceeds
3,000 pages. Part I of the draft report provides information on exposure to
dioxins, including chapters on dietary intake; part II addresses health
assessment methodologies and specific health effects; and part III, the
Integrated Summary 5 highlights information in parts I and II on exposure
and health effects and provides a risk characterization- a statement
summarizing EPA?s assessment of the health risks associated with dioxins. In
the reassessment, EPA studied the risks of cancer as well as noncancer
health effects, such as neurological and reproductive impairments.

Founded in 1948, the World Health Organization (WHO) is a specialized agency
of the United Nations, with 191 member states. WHO?s functions include
giving worldwide guidance in the field of health and setting global
standards for health. WHO carries out these functions through a variety of
offices and programs that often collaborate with each other and with other
public health entities of WHO?s member states and nongovernmental
organizations. The principal contributors to the WHO reassessments of dioxin
risks that are discussed in this report have been (1) the International
Agency for Research on Cancer, which coordinates and conducts both
epidemiological and laboratory research into the causes of cancer; (2) the
WHO European Centre for Environment and Health, which coordinates
comprehensive efforts, in collaboration with the International Programme on
Chemical Safety, to evaluate the possible health risks of dioxins as well as
methods of prevention and control of environmental

4 The World Health Organization (WHO) met in Stockholm in June 1997 to
update earlier TEFs on dioxins for human risk assessment. As of 2002, both
EPA and WHO use the TEFs adopted by WHO in 1997, and published by Van den
Berg et al. in 1998.

5 The full title of part III is Integrated Summary and Risk Characterization
for 2,3,7,8- Tetrachlorodibenzo- p- Dioxin (TCDD) and Related Compounds.

Page 7 GAO- 02- 515 Draft Reassessment of Dioxins

exposure of the general population to these chemicals; 6 and (3) the Joint
Expert Committee on Food Additives of the United Nations? Food and
Agriculture Organization and WHO, which provides scientific evaluations as a
basis for the development of food standards by the Codex Alimentarius (food
code) Commission. 7

To estimate dietary exposure to dioxins, EPA obtained and reviewed
information on (1) the dioxins present in 10 types of foods 8 with high fat
content, (2) the toxicity of individual dioxins contained in these food
types, and (3) the quantities of these foods that people in the United
States typically eat. EPA has incorporated new studies following
improvements in analytical capabilities to detect dioxins in food during the
1990s. However, in its draft reassessment report, EPA identified a number of
limitations with the food data used to estimate dietary exposure that add
uncertainty to the agency?s overall estimate of current average daily
dietary exposure to dioxins. For example, in some cases, the studies
available on the presence of dioxins in foods were not designed to estimate
national averages. Further, while EPA used the accepted method for
estimating the toxicity of the dioxins found in the 10 food types, EPA and
others acknowledge that the method has limitations. Finally, EPA estimated
the quantities of these foods consumed using U. S. Department of Agriculture
(USDA) data on U. S. adults? food consumption based on surveys made between
1989 and 1991; however, EPA believes the dietary habits of Americans have
changed very little over the course of the past decade.

6 The International Programme on Chemical Safety, established in 1980, is a
joint program of three cooperating organizations: the United Nations
Environment Programme, the International Labour Organisation, and WHO. The
International Programme on Chemical Safety?s main roles are to establish the
scientific basis for safe use of chemicals and to strengthen national
capabilities and capacities for chemical safety.

7 The Food and Agriculture Organization of the United Nations was founded in
1945 with responsibilities covering nutrition and associated international
food standards. Among its activities, the Organization approves
international standards and helps frame international conventions and
agreements.

8 EPA?s Integrated Summary (table 4- 7) presents information on exposures to
9 foods, while the detailed exposure chapters, chapters 3 and 4, present
information on 10 foods. However, the estimated total exposures from foods
in both parts of the reassessment are almost identical. EPA Used Data on the

Presence of Dioxins, Toxicity, and Food Consumption to Estimate Human
Dietary Exposure

Page 8 GAO- 02- 515 Draft Reassessment of Dioxins

A body of scientific research on foods in Europe, North America, and other
locations indicates that the primary source of human exposure to dioxins is
the dietary intake of foods, especially those containing animal fat.
According to EPA?s October 2001 draft reassessment report, the average adult
in the United States receives about 95 percent of his or her exposure to
dioxins by eating commonly consumed foods, such as beef, pork, and poultry;
fish; and dairy products. (EPA estimated small exposures to dioxins from the
air and soil as well.)

The 10 types of foods EPA analyzed for its reassessment are beef; pork;
poultry; other meats, such as lamb and baloney; eggs; milk; dairy products,
such as cheese and yogurt; freshwater fish and shellfish; marine fish and
shellfish; and vegetable fat, such as corn and olive oils and margarine.
These foods, only one of which is not of animal origin, are believed to be
the major contributors to dietary exposure to dioxins. Even though vegetable
fat products are estimated to contain low levels of dioxins, EPA included
these foods in its analysis because they are high in fat and common in the
American diet. EPA excluded fruits and vegetables from its analysis because
data on dioxins in U. S. fruit and vegetable products, which generally
contain little or no fat, are extremely limited. The existing data indicate
that typically these products contain low levels of dioxins, which generally
stem from residues- deposits on outer layers with little penetration to
inner portions.

Until recently, chemical analyses of dioxins in foods have focused primarily
on two of the families of dioxins, the CDDs and CDFs, with less attention on
identifying and measuring specific PCBs. The draft reassessment report
includes an evaluation of PCB levels in the 10 food types. The draft report
identifies estimated exposures to CDDs and CDFs together and identifies the
estimated exposure to PCBs separately. This approach provides information
that can inform potential regulatory approaches, among other things, because
CDDs and CDFs result primarily from combustion and industrial processes,
whereas PCBs, which persist in the environment from the 1970s and earlier,
are no longer being manufactured.

As shown in table 1, EPA estimated that the average adult in the United
States is exposed daily to about 63 picograms of dioxins through dietary
intake, with the highest exposure coming from beef and freshwater fish and
shellfish. According to EPA, this exposure level is close to the level that
has caused adverse noncancer effects in animals, such as effects on the
development of reproductive systems. It is important to note that EPA?s
dietary exposure estimates are averages, and they do not apply to EPA
Estimates That Most

Human Exposure to Dioxins Occurs from Eating Certain Types of Foods

Page 9 GAO- 02- 515 Draft Reassessment of Dioxins

adults with additional or unusual exposure to dioxins- for example, from
diets unusually high in fat content or diets of foods high in dioxin
content. 9

Table 1: EPA?s Estimates of the Average U. S. Adult?s Daily Exposure to
Dioxins From Dietary Intake, Picograms per Day

Food type Dietary exposure

to CDDs and CDFs

Dietary exposure

to PCBs Total dietary

exposure to dioxins

Beef 9.0 4. 2 13.2 Freshwater fish and shellfish 5. 9 7.1 13.0 Dairy
products (cheese, yogurt, etc.) 6.6 3. 2 9.8 Other meats (lamb, baloney,
etc.) 4. 5 1.0 5. 5 Marine fish and shellfish 2. 5 2.4 4. 9 Milk 3.2 1. 5
4.7 Pork 4.2 0. 2 4.4 Poultry 2. 4 0.9 3. 3 Eggs 1.4 1. 7 3.1 Vegetable fat
(oils, margarine, etc.) 1.0 0. 6 1.6

Total 40.7 22.8 63.5

Notes: The average adult is assumed to weigh 70 kilograms (154 pounds). A
picogram is onetrillionth of a gram.

Source: Derived from U. S. EPA, October 2001 draft dioxin reassessment
report, chapter 4, tables 4- 30 and 4- 31.

To estimate any population?s dietary intake of dioxins, the specific dioxins
present in the various foods must be identified and measured through
chemical analyses of the foods. However, reliable estimates of the average
concentrations of dioxins in specific foods nationwide have only recently
begun to be available. In the past, data were available only from studies of
dioxin concentrations in a specific food product or products in a specific
location or a few locations, and these data were not sufficient to reliably
estimate average national exposure. During the 1990s, as analytical
capabilities to detect dioxins at parts- per- trillion levels were
developed, new studies of foods in the United States, some with broader
scope than the earlier studies, became available. EPA has incorporated new
studies

9 The reassessment report addresses variability in general population
exposure, indicating it results primarily from differences in dietary
choices that individuals make. EPA estimates that dietary intake of dioxins
for the general population may extend to levels two to three times higher
than the mean estimate. The Food Samples EPA

Studied to Identify the Presence of Dioxins Had a Variety of Limitations

Page 10 GAO- 02- 515 Draft Reassessment of Dioxins

into its analysis of dietary exposure to dioxins to try to develop more
reliable national estimates of such exposure. As a result, the estimates
presented in the October 2001 draft reassessment report are based on more
food data than the drafts developed just a few years ago.

Nevertheless, in its October 2001 draft reassessment report, EPA said that
the amount and the representativeness of the food data it used to estimate
the average U. S. adult?s dietary exposure to dioxins vary. Further, EPA
officials acknowledged that some of the available studies were not designed
to estimate national average exposures. As discussed below, the food sample
data are limited in part by the timing of the sampling, variations in the
methods used to collect the samples, and the types of samples collected and
analyzed. In commenting on a draft of this report, EPA officials said that
these food data limitations do not represent major weaknesses in its
estimates of dietary exposure to dioxins.

 As reported in the draft reassessment, most of the food samples were
collected between 5 and 8 years ago. Current samples would be expected to
have lower dioxin levels because emissions containing dioxins declined by
about 75 percent from 1987 to 1995, and EPA believes the downward trend is
continuing. 10 Nevertheless, EPA believes that the exposure estimates based
on food data from the mid- 1990s are representative of current dietary
exposure for several reasons. First, EPA believes that because most of the
food samples the agency used for its reassessment were collected after the
75- percent decrease in emissions, much of the decrease should already be
reflected in the foods? dioxin concentration numbers. Second, EPA said that,
because most municipal and medical waste incinerators are located far from
and downwind of concentrated meat and dairy production areas, the impact of
any emission reductions since 1995 on the commercial food supply should be
proportionately less than on the environment in general. Third, EPA said
that because reservoir sources of dioxins account for half or more of
current exposure, and because some sources of dioxins are unknown, it is
unlikely that emission reductions that occurred after most of the food
samples were taken would significantly affect the current estimate of
general population exposure from the commercial food supply.

10 EPA does not have information on dioxin emissions reductions post- 1995.
However, EPA?s air toxics regulations are expected to result in further
reductions. For example, EPA expects that its 1995 air toxics rule for large
municipal waste combustors- associated with more than 60 percent of total
dioxin emissions- and its 2000 air toxics rule for small municipal waste
combustors will reduce dioxin emissions from these entities to less than 1
percent of 1990 levels.

Page 11 GAO- 02- 515 Draft Reassessment of Dioxins

 According to EPA, while its analyses of some of the foods are based on
national samples collected from food processing or food monitoring
locations, such as federal slaughtering establishments, other analyses are
based on limited ?market basket surveys?- random purchases of selected
products, such as eggs, direct from grocery stores- in a small number of U.
S. and Canadian cities. 11 Depending on their design, national surveys would
generally be more representative of average dietary exposures to dioxins
than limited surveys.

 Some of the analyses of the foods were derived from individual food
samples, while others were from composite samples. Using composites is more
economical than using individual samples, and EPA believes they are
appropriate for use in analyzing dioxin concentrations to establish average,
or mean, exposure estimates. However, EPA acknowledges that data on the
variability or range of results from individual samples typically are not
available from studies analyzing composite samples. As a result, information
that can provide insight into the reliability of the estimates is not
available. Table 2 shows the number( s), type( s), and date( s) of the
samples EPA used for each of the 10 food categories.

11 Some of the market basket surveys cited by EPA were conducted under the
auspices of the Food and Drug Administration.

Page 12 GAO- 02- 515 Draft Reassessment of Dioxins

Table 2: Numbers, Types, and Dates of Food Samples EPA Used in Estimating
Dietary Exposure to Dioxins

Food type Number/ type of sample Year( s) samples

collected

Beef (back fat samples) 63 individual samples 1994 Pork (belly fat samples)
78 individual samples 1995 Poultry (abdominal fat samples) 78 individual
samples 1996 Milk 8 composite samples 1996 and 1997 Dairy products 8
composite samples 1996 and 1997 Eggs a 15 composite samples

(24 eggs each) 1997 Marine fish and shellfish a 158 individual samples 1995
and 1996 Freshwater fish and shellfish a 222 samples (individual

and composite) 1986- 89; 1994; 1996- 99 Vegetable fat (oils, margarine,
etc.) a 30 individual samples 1995 b Other meats c c a EPA used these
samples to analyze CDDs and CDFs only.

b Estimate based on 1996 study publication date. c EPA did not provide this
information for other meats.

Source: U. S. EPA, October 2001 draft dioxin reassessment report.

 According to the draft reassessment report, data from some of the food
studies were sufficient to estimate exposure to total dioxins- the CDDs,
CDFs, and PCBs. However, the report shows that in other cases, the data only
provided support for estimating exposure to CDDs and CDFs. As a result, for
four food categories, different studies are used for estimating exposure to
CDDs and CDFs than those used for PCBs. As reported in the draft
reassessment, these studies analyzed fewer samples, a number of which were
collected 14 or more years ago and therefore provide data on dioxins that
may not reflect current levels. Table 3 shows information on the four foods
for which EPA used different samples to estimate exposure to PCBs than those
used to estimate CDDs and CDFs.

Page 13 GAO- 02- 515 Draft Reassessment of Dioxins

Table 3: Numbers, Types, and Dates of Food Samples EPA Used in Estimating
Dietary Exposure to PCBs in Four Food Categories

Food type Number/ type of sample Year( s) samples collected

Eggs 18 individual samples 6 composite samples a

1995 1986- 88 b Marine fish and shellfish 1 composite of 13

samples 5 composites a

1995 1986- 1988 Freshwater fish and shellfish 1 composite of 10

samples 6 composites a

1995 1986- 88 b Vegetable fat 5 composites a 1986- 88 a The studies used in
the analyses do not specify the number of individual samples in the
composite samples. b The year that one of the composite samples was
collected is not identified. The study was published in 1989. Source: U. S.
EPA, October 2001 draft dioxin reassessment report.

 As EPA acknowledges in the draft report, its analyses of dioxins present
in foods are based on uncooked foods, even though dioxin levels can be
different in cooked and uncooked foods. According to EPA, while many studies
indicate that foods have similar dioxin concentrations whether they are
cooked or uncooked, the studies show that some foods have lower
concentrations of dioxins when they are cooked, while others have higher
levels when they are cooked. These differences reflect, in part, the fact
that different cooking methods (frying, boiling, grilling, etc.) may have
different effects on dioxin levels. On the basis of the available data,
which it believes are not conclusive, EPA states in the draft reassessment
report that uncooked food is a reasonable surrogate to use for identifying
and quantifying dioxin concentrations in cooked food.

 Because the primary focus of EPA?s exposure assessment was on foods
produced and consumed in the United States, EPA?s analysis does not address
imported food products that may vary from domestic sources in dioxin
content.

Despite these limitations, the data on dioxin levels in foods supporting the
October 2001 draft report reflect a significant improvement compared with
the data EPA had available for use in its 1994 draft reassessment report,
which was peer reviewed in 1995 by EPA?s Science Advisory Board.
Specifically, in the 1994 draft, EPA provided estimates of levels of CDDs
and CDFs for seven food types; the October 2001 draft provided estimates for
ten food types. With the exception of an estimate for fish that was

Page 14 GAO- 02- 515 Draft Reassessment of Dioxins

based on 60 samples, the 1994 draft estimates were developed from samples
ranging in number from 2 to 14; as table 2 shows, the number of samples used
for the 2001 draft is greater. In addition, while EPA recognized that PCBs
were being identified in foods, the agency did not have sufficient data at
that time to develop estimates of the levels of specific PCBs in foods; the
2001 draft does include estimates of PCB levels in foods.

The following sections describe in greater detail the samples EPA used to
identify the level of dioxins in 9 of the 10 foods studied- beef, pork, and
poultry; freshwater and marine fish; milk, dairy, and eggs; and vegetable
fat- and any associated limitations or uncertainties. (The draft
reassessment report does not provide any information supporting EPA?s
estimate of the types and amounts of dioxins in other meats, the tenth food
type. 12 In commenting on a draft of our report, EPA said that information
on other meats would be provided in its final report.)

In estimating exposure to dioxins from beef, pork, and poultry, EPA used
data from the first statistically designed national surveys of dioxin levels
in these foods sponsored by EPA and USDA. These surveys were designed to be
representative of all U. S. regions and all classes of animals slaughtered
in federally inspected slaughtering establishments. EPA believes the three
surveys provide reasonable estimates of the average national concentrations
of dioxins in beef, pork, and poultry. Nonetheless, information EPA provided
in the draft reassessment report about these samples identifies some
limitations and uncertainties about these studies.

 The samples are now between 6 and 8 years old and therefore may not
reflect current exposures. To address this data gap, EPA and USDA are
conducting a follow- up study on dioxin levels in beef, pork, and poultry
that will commence in 2002 and provide updated information. However, EPA
officials said the results of this survey will likely not be available for
incorporation into the dioxin reassessment report that EPA plans to publish
this year.

 The animal samples for beef, pork, and poultry were not meat products sold
in grocery stores but rather were cuts of fat generally not consumed- either
back fat, abdominal fat, or belly fat from slaughtering

12 The report?s references to this category of food are limited to several
tables in the chapters on dietary exposure in which the estimates are
provided, while the Integrated Summary cites nine foods, excluding other
meats. Beef, Pork, and Poultry

Page 15 GAO- 02- 515 Draft Reassessment of Dioxins

establishments. Some uncertainty therefore surrounds the accuracy of EPA?s
estimates of dietary intake of dioxins because of comparability concerns.
EPA used this approach because USDA federal inspectors could obtain the
samples with little disruption to the slaughtering establishments and
because the samples? high fat content would enable more accurate measurement
of dioxins, since the analysis would be of highly concentrated fat samples.
However, this approach assumes that edible meat products sold in grocery
stores contain the same types and amounts of dioxins as the fat samples
(adjusted for differences in percentages of fat). According to EPA, this
assumption is supported by a well- developed understanding of the manner in
which dioxins distribute across fat reservoirs in vertebrates. Therefore,
EPA concluded that the fat samples for all three foods were comparable to
the edible meat samples. EPA also based its conclusion on its analysis of
beef samples- comparing five back fat samples with other cattle parts,
including muscle tissue, which could be representative of edible beef
products. 13 For the five samples, the ratios of CDDs and CDFs in muscle fat
to CDDs and CDFs in back fat varied by up to 300 percent, ranging from 0.58
to 1.7; and the ratios for PCBs varied by up to 50 percent, from 1.0 to 1.5.
14 Although some of the variation may result from imprecision inherent in
measuring picograms, this limited analysis indicates that using fat samples
may overstate or understate to some extent the dioxin levels in beef, pork,
and poultry products.

 EPA reported that it excluded 2 of the 80 samples of abdominal fat from
poultry because they had significantly higher concentrations of certain
dioxins than the other samples. EPA, USDA, and the Food and Drug
Administration investigated the cause of these elevated dioxin levels and
determined that it stemmed from contaminated animal feed that had been
distributed to poultry, fish, hog, and cattle producers in several southern
and southwestern states. 15 EPA considered the two poultry fat samples
inappropriate for the dioxins study, which was aimed at identifying typical
exposures to dioxins. However, it is not clear that the poultry samples with
high concentrations of dioxins were anomalies because the incidence

13 U. S. Environmental Protection Agency, Matthew Lorber et al, Distribution
of Dioxins, Furans, and Coplanar PCBs in Different Fat Matrices in Cattle
(Washington, D. C., 1997). 14 EPA also excluded other beef samples from its
analysis on the basis that the source animals had unusually high levels of
dioxin exposure. However, the analysis does not explain why the ratios (and
thus the overall comparability) would vary with levels of exposure.

15 U. S. General Accounting Office, Food Safety: Agencies? Handling of a
Dioxin Incident Caused Hardships for Some Producers and Processors, GAO/
RCED- 98- 104, Washington, D. C.: Apr. 10, 1998).

Page 16 GAO- 02- 515 Draft Reassessment of Dioxins

of dioxin contamination in animal feeds is not known. For example, this
instance of contaminated animal feed was discovered by the first national
poultry sample, which tested only 80 samples nationwide.

 In response to suggestions from a peer review panel, when the data were
sufficient to do so, EPA presented a standard deviation- the typical amount
of variability around the mean- on its estimates of the average levels of
dioxins in the foods, as well as the range of the levels of dioxins
identified in the samples. For beef, pork, and poultry, EPA was able to
provide this information for the CDDs and CDFs. These data indicated
considerable variability in the levels of CDDs and CDFs in the foods. For
example, the estimated level of 0.28 picograms of dioxins in a gram of pork
has a standard deviation of plus or minus 0.28. In other words, the standard
deviation is equal to or greater than the mean. Accordingly, the estimated
dioxin level is subject to a wide range of uncertainty.

 Because EPA did not have sufficient information to develop a standard
deviation for PCBs, the agency could not develop a standard deviation for
total dioxins (the combination of CDDs, CDFs, PCBs) in beef, pork, and
poultry. As a result, EPA could not state with any degree of certainty that
exposure to total dioxins or to PCBs would fall within specified levels. EPA
believes this limitation is a minor one because it considers the average
exposure level, rather than the more limited extreme exposures, to be of
greater public health interest. Nonetheless, this additional analysis, if
available, would enable policymakers, scientific peer reviewers, and other
users to better evaluate the extent to which the data may be representative
of average national exposures.

Though EPA analyzed more fish samples for the current reassessment draft
than for earlier drafts, the current draft report acknowledges that the
levels of dioxins in fish are more uncertain than those in the other foods
for two reasons. First, the data lack the ?geographic coverage and
statistical power? of the other food surveys. That is, while the sample
sizes for CDDs and CDFs in fish are considerably larger than those used for
the analyses of other foods, they do not provide data that are nationally
representative because of the diversity of fish and bodies of water.
Specifically, there are a significant number and variety of freshwater and
marine fish species living in numerous bodies of water that contain
differing types and levels of dioxins. Moreover, fish consumed in the United
States include both farm- raised and wild fish. Second, EPA based its
estimates for levels of PCBs in fish on a much smaller data set than it used
for CDDs and CDFs. EPA used 222 samples to estimate the levels of CDDs and
CDFs in freshwater fish and shellfish and 158 samples for marine fish and
shellfish compared with 7 and 6 composite samples for PCBs for freshwater
and marine fish, respectively. Further, most of the Freshwater and Marine
Fish

and Shellfish

Page 17 GAO- 02- 515 Draft Reassessment of Dioxins

samples for PCBs were from Canadian rather than U. S. cities, and the
analyses of levels of PCBs in them did not evaluate all of the PCBs
identified as being toxic. For example, according to the report, only one of
the composite samples for marine fish and shellfish, collected between 1984
and 1986, was analyzed for the presence of the most common and toxic PCB,
referred to as PCB- 126. For these reasons, EPA acknowledges in the draft
report that the resulting estimates are not representative of the level of
dioxins in fish nationally. We note that the limitations of the data used to
estimate the levels of PCBs in fish are particularly significant because in
the report, EPA estimates that freshwater fish contains the highest levels
of PCBs (and total dioxins) of all the foods studied.

The samples EPA used to estimate the levels of dioxins in fish were derived
from EPA?s National Bioaccumulation Study 16 and three market basket surveys
in the United States and Canada. Samples for the bioaccumulation study were
collected between 1986 and 1989, whereas the samples for the market basket
surveys were collected about a decade later, between 1995 and 1999. Some of
the limitations and uncertainties associated with these samples that EPA
acknowledged in the draft report are highlighted below.

 Most of the fish samples used for the reassessment draft were collected 5
or more years ago; some are between 13 and 16 years old.

 EPA did not have sufficient data to estimate exposure to PCBs from eating
freshwater or marine shellfish.

 Some of the estimates for freshwater fish, such as trout, are based on
samples from the bioaccumulation study that may be more representative of
wild fish (i. e., fish caught in recreational fishing) than fish typically
purchased by the general population at grocery stores, which is largely
farm- raised. Specifically, in cases in which EPA did not have data on
farmraised freshwater fish or fish purchased in grocery stores, the agency
used the concentration of CDDs and CDFs from samples of wild caught fish
from the bioaccumulation study. This use of older data on wild fish
increases the uncertainty about the representativeness of EPA?s exposure
estimate.

 For some fish species, such as mullet and mackerel, estimates were based
entirely on samples collected in the Mississippi area and therefore may not
be representative of levels seen in other locations.

16 EPA?s National Bioaccumulation Study, published in 1992, investigated the
prevalence of selected bioaccumulative pollutants, including dioxins, in
fish.

Page 18 GAO- 02- 515 Draft Reassessment of Dioxins

 EPA did not have sufficient data to estimate a standard deviation for the
average levels of dioxins in freshwater or marine fish. As a result, EPA
cannot state with any degree of certainty what the related dietary exposure
to dioxins is.

The milk samples upon which both the milk and dairy estimates are based came
from a national survey. In this survey, samples were collected during the
four seasons, providing information on seasonal (temporal) variations. The
milk samples were collected from 51 sampling stations, located in a majority
of the states, that support EPA?s Environmental Radiation Ambient Monitoring
System. 17 In contrast, the estimates for CDDs and CDFs in eggs are based on
Food and Drug Administration market basket surveys in 1997 in California,
Georgia, Minnesota, New York, Ohio, Oregon, Pennsylvania, and Wisconsin. The
estimates for PCBs in eggs are based on market basket surveys in San Diego,
California; Atlanta, Georgia; Binghamton, New York; and five major Canadian
cities. EPA used composite samples of milk and eggs to identify and measure
the presence of specific dioxins in milk, dairy, and eggs. Information
provided in the draft report identifies some limitations associated with
these data.

 Most of the milk samples were collected 6 years ago.

 The egg samples used to support the analyses for CDDs and CDFs were
collected 5 or more years ago. The estimates for PCBs in eggs are based, in
part, on samples obtained in five Canadian cities between 14 and 16 years
ago.

 Only one of the six composite samples used to estimate the level of PCBs
in eggs was analyzed for the presence of the most common and toxic PCB.

According to EPA, its estimates of CDDs and CDFs in vegetable fat were
developed from a market basket survey that was not representative of edible
oil consumption in the United States. The 30 samples of various oils, solid
shortening, margarine, and an oil spray were obtained from grocery stores in
nine U. S. cities or metropolitan areas: Chicago, Illinois; Cincinnati,
Ohio; Denver, Colorado; Miami, Florida; Minneapolis, Minnesota; Salt Lake
City, Utah; San Antonio, Texas; San Francisco, California; and the
Washington, D. C., metropolitan area. Although neither the reassessment
report nor the study (published in 1996) states when the

17 EPA?s Environmental Radiation Ambient Monitoring System is a national
network of monitoring stations that regularly collect air, water,
precipitation, and milk samples for analysis of radioactivity. The samples
were collected from 51 stations located in 41 states, Puerto Rico, and
Panama. Milk, Dairy, and Eggs

Vegetable Fat

Page 19 GAO- 02- 515 Draft Reassessment of Dioxins

samples were collected, there is typically at least a 1- year lag between
collection and publication, indicating that the samples were collected 8 or
more years ago. EPA used limited data to estimate the level of PCBs in
vegetable fat. This estimate is derived from five composite samples of
cooking fats and salad oils, each of which was obtained 14 or more years ago
from one of five major (unidentified) Canadian cities. As a result of these
limitations that EPA identified in the draft report, the estimate for
dietary exposure to dioxins from eating vegetable fats is unlikely to
reflect current average dietary exposure in the United States.

After using the chemical analyses discussed previously to identify the types
and quantities of dioxins present, EPA estimated the toxicity of the dioxins
in the 10 types of foods, using measures called toxic equivalency factors.
(As noted earlier, these measures- called TEFs- are used to create a frame
of reference by comparing the potential toxicity of individual dioxins in a
sample with the toxicity of the most toxic and best understood dioxin, TCDD,
which is assigned a TEF of 1.) EPA used the TEFs that were updated by WHO in
1997. For each of the types of foods, EPA multiplied the measured types and
amounts of the dioxins present by the related TEFs to arrive at a ?dioxin
toxic equivalence value? for that particular food category/ dioxin
combination. 18 For each food category, the total dioxin toxic equivalency
is the sum of these products- that is, the sum of the toxic equivalence
values for (1) CDDs and CDFs and (2) PCBs. This provides an indicator of the
relative toxic concentration of dioxins in each food category. As table 4
shows, EPA estimated that freshwater fish and shellfish had the largest per-
gram concentration of dioxins with toxic effects.

18 For beef, pork, poultry, milk, and dairy products, the toxic
concentrations are also based on EPA?s estimates of the average percentage
of fat in these foods-- 17, 19, 9, 1.8, and 12 percent, respectively. Method
EPA Used to

Estimate Toxicity of Dioxins in the 10 Food Types Is Accepted by Experts but
Has Limitations

Page 20 GAO- 02- 515 Draft Reassessment of Dioxins

Table 4: EPA?s Estimates of Toxic Concentrations of Dioxins in 10 Food
Categories, Picograms per Gram

Food category Dioxin toxic

equivalence values for CDDs and CDFs

Dioxin toxic equivalence

values for PCBs

Total dioxin toxic equivalence

values

Freshwater fish and shellfish 1. 00 1. 20 2. 20 Marine fish and shellfish 0.
26 0. 25 0. 51 Pork 0.28 0.01 0.29 Beef 0.18 0.08 0.26 Other meats 0. 18 0.
04 0. 22 Eggs 0.08 0.10 0.18 Dairy 0. 12 0. 06 0. 18 Poultry 0. 07 0. 03 0.
10 Vegetable fats 0.06 0.04 0.10 Milk 0.02 0.01 0.03

Note: The toxic equivalence values are estimated on a whole (wet) weight
basis, as opposed to a dry weight basis.

Source: Derived from U. S. EPA, October 2001 draft dioxin reassessment
report.

The toxic equivalence approach using TEFs has evolved over the last 20 years
and is the internationally accepted scientific approach for risk assessments
of dioxins. This approach has been formally adopted by several countries and
as guidance by international organizations, such as WHO. TEFs are used to
decrease the overall uncertainty in assessing the health risks of dioxins
because they provide a framework for addressing the complex mixtures of
dioxins to which people are most often exposed. Nonetheless, a number of
uncertainties are involved in the use of the TEF concept. As a result of
these uncertainties, estimates of the concentrations of dioxins in foods
based on this approach may be overstated or understated.

The draft reassessment report acknowledges that there are still many
questions about the use of the TEF method and the validity of some of the
underlying assumptions. The report states that many assumptions are
necessary because of lack of data. Specifically, the derivation of TEFs is
limited by the amount of available data on the relative potency of different
dioxins compared with TCDD. For many dioxins, the available data on relative
potency may be limited to only a few experimentally observed effects. Some
of these effects may not be considered toxic by themselves, but they still
might provide evidence that exposure to dioxins led to biological or
chemical effects in experimental subjects. For example, EPA noted that only
TCDD and one mixture of certain dioxins have been tested

Page 21 GAO- 02- 515 Draft Reassessment of Dioxins

for carcinogenicity. Therefore, in order to develop a TEF that estimates the
cancer potency of a mixture including other dioxins, scientists have assumed
that the relative potencies observed for noncancer effects approximate those
for cancer. In other words, once derived, TEFs apply to all effects, not
just those for which relative potency data were available.

Nonetheless, after considering a number of the uncertainties and limitations
of this approach, the international experts who derived the current TEFs
concluded that the TEF concept is still the most plausible and feasible
approach for risk assessment of dioxins. Furthermore, the TEF values for
individual dioxins are reevaluated and updated periodically to reflect the
available evidence. When WHO established the most recent TEFs in 1998, it
suggested that the toxic equivalency scheme be reevaluated every 5 years and
that the TEFs and their application to risk assessment be reanalyzed to
account for emerging scientific information.

To develop its estimate of the daily dietary intake of dioxins by the
average adult in the United States, EPA needed to calculate the amount of
food containing dioxins that Americans typically eat. EPA obtained this
information for the 10 food types from USDA food intake surveys. The USDA
survey data include information on the amounts of specific foods consumed in
a day by an average person weighing 70 kilograms (154 pounds).

USDA obtained its data from detailed food surveys prepared by thousands of
individuals selected from statistical samples. In these surveys, individuals
generally provided detailed information on food consumption for 2 days. The
surveys used statistical sampling to ensure that all seasons, geographic
regions of the United States, and demographic and sociodemographic groups
were represented. EPA?s analysis of these data tabulated intake rates for
the major foods, as well as for individual food items. The total quantity of
each food eaten by the survey population in a survey day was tabulated and
weighted to represent the quantity eaten by the entire U. S. population in a
typical day. For the draft reassessment report, EPA averaged USDA?s data for
three age groups of adults ranging from ages 20 to 70 and over. Table 5
provides EPA?s estimates of the daily dietary intake of 10 food types by
adults in the United States. Daily Dietary Intake

Estimates Are Primarily Based on Food Surveys of U. S. Adults Administered
Between 1989 and 1991

Page 22 GAO- 02- 515 Draft Reassessment of Dioxins

Table 5: Estimated Daily Dietary Intake of 10 Food Types for an American
Adult Weighing 70 Kilograms (154 pounds)

Food Estimated dietary intake (grams per day)

Milk 175.0 Dairy 55.0 Beef 50.0 Poultry 35.0 Other meats 25.0 Eggs 17.0
Vegetable fats 17.0 Pork 1.5 Marine fish and shell fish 9.6 Freshwater fish
and shellfish 5. 9

Source: Derived from U. S. EPA, October 2001 draft dioxin reassessment
report.

While EPA prefers to use USDA food data from the 1989- 91 USDA

Continuing Survey of Food Intake By Individuals because it has conducted a
statistical analysis of these data and includes them in the agency?s
Exposure Factors Handbook, 19 the draft reassessment report uses other data
for fish and does not provide information on the basis for its estimates of
dietary intake of other meats. Specifically, the draft reassessment report
derived its estimates of the daily dietary intake of beef, pork, poultry,
milk, dairy products, vegetable fats, and eggs from USDA?s Continuing Survey
Food Intake By Individuals conducted from 1989 through 1991. In contrast,
the daily dietary intake of freshwater and marine fish and shellfish were
derived from a March 2000 report on the consumption of fish prepared by
EPA?s Office of Water. This report used data from USDA?s Continuing Survey
of Food Intakes By Individuals

conducted in 1994, 1995, and 1996. In this report, EPA weighted its
estimates of exposure to dioxins from fish by the species- specific
concentrations according to species- specific fish consumption rates for the
U. S. population. However, in cases where species- specific concentration
data were not available, EPA used default values. For example, EPA used data
from the bioaccumulation study as the default for

19 EPA?s August 1997 Exposure Factors Handbook provides data on standard
factors needed to calculate human exposure to toxic chemicals, including the
estimated average daily intake of foods that EPA program offices are
encouraged to use in exposure assessment activities.

Page 23 GAO- 02- 515 Draft Reassessment of Dioxins

certain freshwater fish. The use of various default assumptions adds
uncertainty to the exposure estimates.

EPA officials said that EPA did not use more current dietary intake data
from USDA in the October 2001 draft reassessment because EPA has not yet
fully reviewed surveys subsequent to the 1989- 91 surveys that it uses in
its Exposure Factors Handbook. EPA officials told us that they did not
believe it was necessary to use more current data because the dietary habits
of Americans have changed very little over the course of the past decade.
These officials cited data collected in surveys conducted between 1994 and
1996 that show little change in the intake of the 10 foods compared with
surveys conducted between 1989 and 1991.

EPA and WHO have undertaken extensive efforts to reassess the health risks
of exposure to dioxins. EPA?s comprehensive dioxin reassessment objective
has been to characterize the potential human health risks posed by exposure
to dioxins. To do this, EPA used an extensive, multiyear review process. In
contrast, WHO had more narrowly focused primary objectives and conducted its
reassessments of dioxins through a succession of individual reviews and
meetings. Nonetheless, EPA and WHO used very similar analytical methods to
identify the types of potential human health hazards associated with
exposure to dioxins and assess the probability and severity of harm given
different levels of exposure. Moreover, the conclusions EPA and WHO reached
on the basis of their respective reassessments also reflected much
agreement. However, there were some significant issues on which EPA and WHO
differed, such as whether there are threshold doses of dioxins to which
humans could be exposed over a lifetime without significant risk of cancer
and whether dioxins other than TCDD are human carcinogens. EPA?s and WHO?s

Specific Reassessment Objectives and Processes Differed, but Their
Analytical Methods and Conclusions on Dioxins? Health Risks Are Similar

Page 24 GAO- 02- 515 Draft Reassessment of Dioxins

In general, both EPA and WHO focused their evaluations of the health effects
and risks associated with dioxins on TCDD and 28 other related chemical
compounds (including 12 dioxin- like PCBs) for which consensus toxic
equivalency factors had been established through a 1997 meeting organized by
WHO. 20 However, there were important differences in some of the specific
objectives of EPA?s and WHO?s dioxin reassessments and the processes used by
EPA and WHO to develop the reassessments.

EPA?s overall objective has been very broad: to characterize the available
scientific information on the potential health risks posed by exposures to
dioxins. EPA therefore addressed each of the four major components of a
chemical risk assessment: hazard identification, dose- response assessment,
exposure assessment, and risk characterization. 21 The resulting
characterization of risks posed by dioxins can be used to inform risk
management decisions, such as whether and where to set or revise regulatory
standards, but other information and factors would also enter into such
decisions. 22 The process by which EPA has undertaken this task has been a
comprehensive, multiyear review. Moreover, the EPA reassessment has included
multiple independent scientific peer reviews of various draft reports by
EPA?s Science Advisory Board and others. EPA has also solicited public
review and comments on its draft reassessment.

WHO?s reassessment objective also addressed a broad range of data and issues
regarding the potential exposures and health risks associated with dioxins,
but the specific reports and evaluations WHO produced on dioxins generally
had more narrowly focused primary objectives than EPA?s reassessment report.
In addition, rather than a comprehensive,

20 Because the results of this expert meeting were not published until 1998,
EPA refers to this international consensus scheme as the TEQ- WHO 98 update.
See Martin Van den Berg, et al., ?Toxic Equivalency Factors (TEFs) for PCBs,
PCDDs, PCDFs for Humans and Wildlife,?

Environmental Health Perspectives (1998): Vol. 106, No. 12: 775- 792.) 21
See U. S. General Accounting Office, Chemical Risk Assessment: Selected
Federal Agencies? Procedures, Assumptions, and Policies, GAO- 01- 810
(Washington, D. C.: Aug. 6, 2001) for a more detailed description of the
four- step process and other chemical risk assessment procedures that may be
used by EPA.

22 See U. S. General Accounting Office, Environmental Protection Agency: Use
of Precautionary Assumptions in Health Risk Assessments and Benefits
Estimates,

GAO- 01- 55 (Washington, D. C.: Oct. 16, 2000) and GAO- 01- 810. EPA
Undertook a

Comprehensive Assessment, While WHO Used a Succession of Individual Reviews
and Meetings

Page 25 GAO- 02- 515 Draft Reassessment of Dioxins

integrated process such as EPA?s, WHO?s process consisted of individual
evaluations and meetings for each of those particular objectives. 23

 In 1997, the International Agency for Research on Cancer (IARC), a chief
contributor to WHO?s dioxin risk assessments, published monographs covering
TCDD and 16 other dioxins. (This agency publishes the results of its
evaluations of specific chemicals in its series IARC Monographs on the
Evaluation of Carcinogenic Risks to Humans. In the rest of this report, we
call the monographs covering TCDD and other dioxins the 1997 cancer
monographs.) The primary objective leading to these monographs was to
classify TCDD and other specific dioxins under a standard scheme that
identifies whether and under what circumstances substances are human
carcinogens. 24 Essentially, this objective corresponds to the hazard
identification step of EPA?s four- step risk assessment process.

 Under its activities related to the European Centre for Environment and
Health, WHO organized two meetings of experts addressing issues on the
health effects of dioxins. In June 1997, WHO convened experts in Stockholm,
Sweden, to derive consensus toxic equivalency factors for 29 dioxins that
could be used for human, fish, and wildlife risk assessments. In May 1998,
WHO convened 40 experts from 15 countries in Geneva, Switzerland, to
evaluate scientific data on the health risks and exposures of dioxins with
the principal objective of updating the estimated amount of dioxins to which
humans can be exposed daily without appreciable harm. In the rest of this
report, we call these efforts, respectively, the 1997 TEF meeting and the
1998 consultation.

 At the 57th meeting of the Food and Agriculture Organization of the United
Nations/ WHO Joint Expert Committee on Food Additives in June 2001 in Rome,
Italy, the committee for the first time evaluated the risks associated with
the presence of dioxins in food. The participants specifically evaluated
dioxins (among other specific food additives and contaminants), with the
view toward recommending acceptable intakes for dioxins contained in foods.
The committee used the 1998 consultation?s assessment as the starting point
for its evaluation but took

23 EPA experts participated in all of these international meetings and
evaluations convened by WHO on dioxins. 24 Group 1 is the classification for
chemical agents or mixtures that WHO?s reviewers determine are carcinogenic
to humans. Group 2A is the classification for those probably carcinogenic,
and Group 2B for those possibly carcinogenic. Group 3 includes those not
classifiable as to human carcinogenicity, and Group 4 covers those probably
not carcinogenic.

Page 26 GAO- 02- 515 Draft Reassessment of Dioxins

into account newer studies. In the rest of this report, we call this
evaluation the 2001 food additives meeting.

Appendix I highlights some of the major milestones in the EPA and WHO
assessments of dioxin risks, with a particular focus on the reassessment
efforts that both entities began in the 1990s.

Despite differences in some of the specific objectives and processes of
their respective reassessment efforts, EPA and WHO used similar analytical
methods to identify and assess the potential health risks of dioxins.
Through these analyses, EPA and WHO identified the types of potential
hazards that might be associated with exposure to dioxins, the circumstances
under which these substances could cause adverse effects, and the
probability and severity of expected effects given different levels of
exposure to dioxins. Specifically, both EPA and WHO

 reviewed available scientific data from many studies of humans and animals
covering a variety of effects potentially associated with exposure to
dioxins;

 continued to consider cancer risks, as in the original dioxin risk
assessments, but also paid increasing attention to noncancer health effects,
such as changes in reproductive and developmental functions and the immune
and nervous systems, as well as other health problems, such as chloracne (a
chronic and disfiguring skin disease) and alterations in liver enzyme
levels;

 reviewed evidence regarding other biochemical, molecular, or cellular
effects that have been observed in various studies, agreeing that these
effects might be precursors to subsequent adverse effects; and

 considered a range of analytical methods, models, and approaches to assess
the dose- response relationships for exposure to dioxins.

EPA and WHO also used some analytical concepts and methods that they agreed
were more appropriate to the analysis of dioxins than those that are often
used for risk assessments of other chemicals. For example, both entities
used body burden- the concentration of dioxins in the body- instead of other
dose measures, such as daily intake, to compare risks between humans and
animals and determine doses that would be of equivalent risk in humans and
animals. The organizations also concurred that the concept of toxic
equivalency should be used to facilitate risk assessment of dioxins and
complex mixtures of dioxins. Furthermore, in contrast to chemical risk
assessments in general, EPA and WHO often had sufficient data to focus on
the dose level associated with a 1- percent EPA and WHO Used

Similar Analytical Methods

Page 27 GAO- 02- 515 Draft Reassessment of Dioxins

increase in a particular effect (rather than being limited to the level
associated with a 10- percent increase) and seldom had to extrapolate
outside the observed doses or exposures from the studies that they used to
prepare the reassessments.

Much of the scientific data available to EPA and WHO on the potential
effects of exposure to dioxins came from animal studies, mainly studies of
TCDD on a variety of species. (According to WHO, most other dioxins and
dioxin- like compounds are ?relatively poorly studied? compared with TCDD.
25 ) However, EPA?s and WHO?s recent reassessment efforts also benefited
from the increasing quantity and quality of data on the effects of dioxins
in humans that became available during their reassessments. Among the
sources of these human data were studies of occupational exposure of people
who produce and apply herbicides; residents in a contaminated area of
Seveso, Italy (where an accident at a chemical factory had released a cloud
of toxic chemicals, including dioxins, in 1976); and noncancer effects in
infants and children.

The conclusions EPA and WHO reached on the basis of their respective dioxin
reassessments were frequently similar, but some significant differences also
emerged. With respect to the major areas of agreement, both EPA and WHO
concluded that TCDD is a human carcinogen and that

 dioxins can cause a variety of both cancer and noncancer health effects,

 dioxins act in the same way within the body to cause the effects observed
in animals and humans,

 dioxins adversely affect human health at lower exposure levels than
previously thought, and

 some effects could occur at or near the levels to which the general
population is now being exposed.

EPA and WHO not only concurred at the broad level of these conclusions but
also on many of the supporting details. For example, both entities had
similar reasons for concluding that TCDD is a human carcinogen: the
combination of sufficient evidence that TCDD causes cancer in animals,

25 EPA?s reassessment also noted that there is a broad range in the quality
and quantity of data available for individual dioxins. However, EPA pointed
out that five dioxins (including TCDD) contribute approximately 80 percent
of the total toxic equivalence of dioxin in humans and characterized these
five chemicals as ?well studied.? Nevertheless, EPA?s reassessment relied
primarily on TCDD studies. EPA and WHO Had Similar

Overall Conclusions but Differed on Some Important Issues

Page 28 GAO- 02- 515 Draft Reassessment of Dioxins

more limited evidence of carcinogenicity from human data, and strong
evidence that TCDD operates through the same mode, or mechanism, of action
in animals and humans. 26

The major differences of opinion between EPA and WHO concerned whether (1)
there is a threshold below which exposure to dioxins would not be expected
to cause cancer, (2) it is useful to calculate a ?tolerable?

dose of dioxins or estimate a dose without appreciable risk of deleterious
effects to which humans can be exposed over a lifetime, and (3) both
mixtures of dioxins and dioxins other than TCDD are likely human
carcinogens. In addition, EPA quantified the general population?s possible
additional risk of developing cancer from exposure to dioxins, while WHO did
not. Such differences may make it more difficult for interested parties to
compare the results of EPA and WHO dioxin risk assessments. The following
sections provide additional information on each of these differences.

EPA and WHO disagreed about whether there is a threshold below which
exposure to dioxins would not cause cancer. EPA concluded that available
evidence was insufficient for the agency to depart from its default linear
cancer risk assessment approach, which is based on an assumption that no
threshold exists regarding adverse effects (i. e., any exposure to
carcinogenic substances, no matter how small, poses some risk of developing
cancer). 27 In contrast, WHO concluded that there is a threshold for all
adverse effects, including cancer. Specifically, WHO concluded that dioxins
do not initiate cancer through a direct effect on genetic material (that is,
they are non- genotoxic carcinogens) and, therefore, do not warrant a linear
(no threshold) assessment of risk. WHO also concluded that noncancer health
effects occurred at lower body burdens (concentrations) of dioxins than the
body burdens at which cancer occurred in animals. Accordingly, WHO
determined that establishing a tolerable intake based on estimated
thresholds for noncancer effects would also address any cancer risks (that
is, if the intake were set to avoid

26 The National Toxicology Program of the Department of Health and Human
Services also listed TCDD as a known human carcinogen in the 2001 addendum
to its Report on Carcinogens (9th edition) on the basis of a similar
combination of epidemiological (human) and mechanistic information,
supported by experimental animal studies.

27 EPA?s risk assessment guidelines set forth ?default? assumptions- generic
approaches based on general scientific knowledge and policy judgment that
are applied to various elements of the risk assessment process when specific
scientific information is not available. Cancer Threshold

Page 29 GAO- 02- 515 Draft Reassessment of Dioxins

appreciable noncancer health consequences, it should also avoid appreciable
consequences concerning cancer).

WHO programs estimated a tolerable daily intake for dioxins in 1998 and a
tolerable monthly intake in 2001. These measures represent the amounts of
dioxins that the WHO experts believe a human could ingest daily or monthly
for a lifetime without appreciable health consequences. Expressing these
estimates as ?tolerable? intakes generally does not connote that such
intakes are acceptable or risk free, but rather that any health consequences
would be judged to be tolerable while exposure is continuing to be reduced.
28 EPA?s related (but not identical) measure is the reference dose, which
would estimate a daily exposure to the human population, including sensitive
subgroups, that is likely to be without an appreciable risk of deleterious
effects during a lifetime. 29 EPA, however, chose not to calculate a
reference dose for dioxins, as it generally does for noncancer health
assessments of other substances. According to EPA, it did not do so because
any reference dose that it would recommend for dioxins would likely be below
(perhaps considerably below) the current background intake levels and body
burdens of the U. S. population. EPA pointed out that reference doses are
typically calculated to address the risks of incremental exposures over
background exposure. 30 In the case of dioxins, however, background exposure
is a significant component of total exposure. Therefore, in EPA?s opinion, a
reference dose would be uninformative to risk managers for safety
assessment. EPA also noted that, if it were to set a reference dose, its
estimate likely would be more

28 Although the experts participating in WHO?s 1998 consultation established
a tolerable daily intake range of 1 to 4 picograms per kilogram of body
weight, they also stressed that the ultimate goal should be to reduce human
intake levels below that range and recommended that every effort should be
made to reduce exposure to the lowest possible level.

29 The Agency for Toxic Substances and Disease Registry in the Department of
Health and Human Services uses a similar measure known as the minimal risk
level. Minimal risk levels are estimates of the daily human exposure to a
hazardous substance that are likely to be without appreciable risk of
adverse noncancer health effects over a specified duration of exposure.
These substance- specific estimates are intended to serve as screening
levels to identify contaminants and potential health effects that may be of
concern at hazardous waste sites. In 1999, the Agency for Toxic Substances
and Disease Registry set a minimal risk level for dioxins and related
compounds of 1. 0 picogram TEQ per kilogram of body weight per day, but did
not use body burden as a dose metric.

30 Background exposure to chemicals is the exposure that regularly occurs to
members of the general population from media such as food, air, and soil
that have concentrations of these chemicals within normal background range.
Estimating a ?Tolerable? Dose

or One Without Appreciable Risk

Page 30 GAO- 02- 515 Draft Reassessment of Dioxins

stringent than the tolerable intake levels for dioxins proposed by WHO
because EPA?s traditional approach for setting a reference dose gives more
weight to scientific uncertainties than the approach WHO used in setting its
tolerable intake level.

EPA chose instead to use an alternative approach, the margin of exposure, to
characterize noncancer risks. The margin of exposure is a ratio that shows
how far the actual (or estimated) total human exposure to a particular
substance is from levels at which adverse effects have been demonstrated to
occur in human or animal studies. The margin of exposure is an alternative
way of characterizing the likelihood that noncancer effects may be occurring
in the human population at environmental exposure levels. A reference dose,
on the other hand, estimates a level of exposure below which EPA considers
it unlikely that any adverse effects will occur. EPA generally considers
margins of exposure of 100 or more as adequate to rule out the likelihood of
significant effects occurring in humans. However, for the most sensitive
effects identified with dioxins (i. e., those that occurred at the lowest
doses of exposure), the margins of exposure ranged from 15 to less than 1.

EPA and WHO both characterize TCDD as carcinogenic to humans. While EPA
further characterizes other individual dioxins and mixtures of dioxins as
?likely to be human carcinogens,? WHO does not. Specifically, WHO states
that the carcinogenicity of dioxins other than TCDD cannot be determined
because of insufficient data. This difference of opinion largely reflects
the specific objectives and scopes of EPA?s and WHO?s assessments. EPA?s
conclusion reflects a ?weight of the evidence? judgment- that is, it is
based on EPA?s entire reassessment of dioxins (resting, in particular, on
the conclusion that all dioxins share a similar mode of action and using
evidence from both animal and human studies). In contrast, WHO?s cancer
monographs looked only at individual dioxins, focusing on whether they met
specific criteria. Consequently, WHO?s conclusions reflected a narrower data
set and did not address the risks posed by mixtures of dioxins. However,
because most human exposure is to mixtures rather than individual dioxins,
and both EPA and WHO advocate using the same toxic equivalency factors for
assessing the dioxins in such mixtures, any differences in the
carcinogenicity classifications may have little practical impact.
Characterizing Cancer Risks

Posed by TCDD, Other Individual Dioxins, and Mixtures of Dioxins

Page 31 GAO- 02- 515 Draft Reassessment of Dioxins

Quantifying the lifetime cancer risk to the general population from exposure
to dioxins was an important component of EPA?s dioxin reassessment. EPA
estimated that the upper bound on the general population?s lifetime risk for
all cancers from dioxins might be on the order of 1 in 1,000 or more (i. e.,
people might experience a 1 in 1,000 increased chance of developing cancer
over their lifetime because of exposure to dioxins). EPA?s reassessment also
states that the vast majority of the population is expected to have less
risk per unit of exposure and some may have zero risk. WHO did not carry out
such a quantitative assessment of the general population?s cancer risk for
two main reasons. First, calculations of population risk are beyond the
scope of WHO?s IARC cancer monographs, which evaluate whether and under what
circumstances particular substances could pose a cancer risk to humans but
generally do not provide quantitative risk estimates. 31 Second, as noted
previously, WHO?s conclusion about a cancer threshold for dioxins led it to
focus on noncancer effects when deriving tolerable intake levels for
dioxins. However, WHO did explore the calculation, through modeling, of a
cancer ?benchmark dose,? the dose or body burden estimated to result in a 1-
percent increase in cancer mortality. But WHO noted that its estimates for
this benchmark dose ranged quite widely and strongly depended on the
assumptions made during the modeling.

Appendix II provides a more detailed comparison of the EPA and WHO
conclusions regarding a number of major issues covered by the entities?
dioxin risk assessments.

31 WHO?s IARC working groups may do some quantitative evaluations of human
data in the monographs, but without extrapolation beyond the range of data
available. Quantitative extrapolation of cancer risks from experimental
(animal) data to the human situation is not undertaken. Quantifying Cancer
Risks

Page 32 GAO- 02- 515 Draft Reassessment of Dioxins

Two independent peer review panels, including an EPA Science Advisory Board
32 panel, reviewed major sections of EPA?s draft dioxin reassessment report
in 2000. Both panels generally agreed with a number of key assumptions and
approaches that EPA used to develop its updated health risk assessment of
dioxins. Each of the peer review panels had a number of recommendations and
suggestions for EPA to address or consider, most of which focused on the
approaches and methodologies used to depict the health risks associated with
dioxins. EPA made a number of revisions to its draft report in response to
these recommendations and comments. The peer review panels disagreed with
EPA on a few major points, and the Science Advisory Board panel emphasized
the need for additional research to bridge gaps in data.

Both an independent expert peer review panel and one convened by EPA?s
Science Advisory Board reviewed the draft reassessment report on dioxins in
2000. These reviews resulted in part from the Board?s review of an earlier
version of EPA?s draft reassessment report. In 1995, a Board panel had
reviewed the draft reassessment and requested that EPA make substantive
revisions to the chapter on dose- response modeling 33 and to the Integrated
Summary. The Board had also requested that EPA develop a separate chapter on
toxicity equivalence factors and submit the revised dose- response and new
toxicity chapters to external peer review before the next Board review of
these sections. In response, EPA revised the chapter on dose- response
modeling and had it peer reviewed in 1997. Similarly, EPA wrote a chapter on
toxicity equivalence factors and had it peer reviewed as part of the July
2000 review.

In July 2000, EPA organized an independent peer review panel to review the
revised Integrated Summary and the new chapter on toxicity equivalence
factors. To obtain an objective critique, EPA had a contractor select 12
independent individuals with expertise in several technical fields,
including risk characterization and communication; toxicology; epidemiology;
sources of, and population exposure to, dioxins and related

32 EPA?s Science Advisory Board reviews key scientific studies and
methodologies used by the agency in formulating rules to protect the
environment and public health. The Board comprises nongovernment experts and
provides technical advice directly to the EPA administrator primarily on the
basis of its peer reviews- that is, critical evaluations by panels of
independent experts.

33 Dose- response modeling is used to estimate the health risks associated
with various exposure levels (dose). EPA?s Draft Dioxin

Reassessment Report Generally Reflects the Views of Recent Peer Reviews

Two Peer Review Panels Reviewed the Draft Dioxin Reassessment Report in 2000
and Concurred on Many Key Aspects

July 2000 External Peer Review Panel

Page 33 GAO- 02- 515 Draft Reassessment of Dioxins

compounds; mechanisms and mode of action; and toxic equivalency. The panel
addressed 20 questions about the reassessment report regarding exposure to
and the health risks of dioxins. Table I of appendix III lists the questions
the July 2000 panel addressed in its review.

The panel generally agreed with the approaches and methodologies EPA used in
its reassessment, and noted, among other things, the following:

 Body burden- the concentration of dioxins in the body- is an appropriate

?dose metric? (measure) for comparing health risks across species.

 The use of margin of exposure- a ratio that shows how far actual or
estimated human exposure is from levels at which adverse effects have been
demonstrated to occur in human or animal studies- is a more logical approach
to characterizing noncancer risk of dioxins than comparing exposure to a
reference dose.

 The report?s information on noncancer effects in animals and humans was
adequately assembled, and the explanation of why dioxins? effects observed
in animals are of concern to humans was also sufficient.

 The history, rationale, and support for the toxicity equivalence approach,
which is used to assess risks posed by dioxins and complex mixtures of
dioxins on the basis of their toxicity relative to an equivalent dose of
TCDD, were adequately presented.

As discussed further below, the July 2000 panel also provided several
recommendations and suggestions and identified the topics of greatest
concern for finalizing the Integrated Summary.

Once the July 2000 panel published its recommendations and suggestions in
August 2000, EPA addressed them and sent its revised draft to the Science
Advisory Board?s dioxin reassessment review subcommittee panel in September
2000. The panel comprised several professors and directors employed by
medical institutions and representatives of industry- affiliated research
organizations, consulting firms, and state health agencies. The Board panel
met to review the revised sections of the draft reassessment report in
November 2000. The Board agreed to answer 20 questions on the reassessment
report regarding exposure to and the health risks of dioxins. Most of these
questions were similar to those asked of the July 2000 panel. The Board
panel completed its review and published a report in May 2001. Table 2 of
appendix III lists the questions the Board panel addressed in its review.

The Board panel, as the July 2000 panel before it, endorsed several key
aspects of the reassessment, noting that, among other things, EPA had The
Board Panel

Page 34 GAO- 02- 515 Draft Reassessment of Dioxins

 used appropriate dose metrics, such as body burden, to equate risks across
species;

 assembled and distilled a large and diverse body of literature on
noncancer effects into a coherent document;

 properly chosen the margin- of- exposure approach to characterize
noncancer risks;

 used toxicity equivalence factors to effectively address the joint effects
of complex mixtures of dioxins on human health; and

 compiled an outstanding inventory of dioxin sources and effectively
characterized the estimates of background exposure to dioxins using the
available scientific data.

The Board panel stated that, overall, EPA had prepared a thorough and
objective summarization of the data and had addressed the key issues the
Board had set forth in its 1995 review of the draft. The Board panel
concluded that there was no need to submit further revisions of the
reassessment report and that EPA should proceed to complete and release the
document. However, as discussed in the following section, the Board panel
provided several recommendations and suggestions for EPA to improve the
draft document before its release. The Board panel also recognized the need
for additional research to bridge gaps in data that limit EPA?s ability to
determine the magnitude of the health risks associated with dioxins. In
essence, the Board panel viewed this reassessment as an interim assessment,
recognizing that the data gaps are not likely to be addressed in the
foreseeable future.

While the peer review panels generally agreed with the methodologies and
approaches used by EPA, they made a number of recommendations and
suggestions, and the Board asked specifically that the agency either address
them before this reassessment is released in 2002 or in a future assessment
of dioxins. The panels? recommendations generally reflected either a
consensus of the panelists or the opinion of a majority. EPA generally
addressed the panels? recommendations and suggestions by performing
additional analyses, adding or revising text, identifying the
recommendations or suggestions as related to EPA?s long- term research
goals, or indicating that the data currently available are not adequate to
address the recommendation or suggestion. Additional changes are now being
made as EPA prepares the draft for external interagency review. EPA?s Draft
Reassessment

Report Reflects Changes Made in Response to Recommendations and Suggestions
of Peer Reviews

Page 35 GAO- 02- 515 Draft Reassessment of Dioxins

Four of the five recommendations by the July 2000 panel regarded
improvements EPA could make to the section on health risks associated with
dioxins. The July 2000 panel recommended that EPA

 explicitly explain the relationship between body burden and daily intake,
serum levels, and tissue dose;

 include a table in the final reassessment report summarizing the various
noncancer effects observed in animals and humans at low- level exposures;

 improve the methodologies used in determining the cancer risks of dioxins-
such as requesting more detail on exactly how the cancer slope factor 34 for
estimating cancer risks of the general population was derived; and

 reexamine the basis for its estimate of the upper bound cancer risks to
the general population.

The fifth recommendation of the July 2000 panel involved the use of specific
terminology in the exposure section. In addition, this panel had several
suggestions regarding the health risks associated with dioxins, including
that EPA

 provide more detail in the Integrated Summary on the implications of using
the margin- of- exposure approach rather than comparing exposure with
reference doses;

 more clearly describe the significance of the upper bound cancer risks to
the public; and

 add discussion of the uncertainties associated with using various dose
metrics specifically for evaluating childhood risks.

Ten of the 13 recommendations made by the Board panel also focused on the
need to improve the section on health risks associated with dioxin. These
recommendations included that EPA

 calculate a reference dose to evaluate risk in addition to using the
marginof- exposure approach to provide information on the minimum dose that
humans can receive without suffering harm,

 improve its margin- of- exposure approach by more clearly explaining its
choice to use dose levels associated with a 1- percent increase in a

34 A cancer slope factor is an upper bound estimate of the increased cancer
risk from a lifetime of exposure to an agent, generally approximating or
exceeding the 95 percent confidence limit. This estimate is generally
reserved for use in the low- dose region of the dose- response relationship.

Page 36 GAO- 02- 515 Draft Reassessment of Dioxins

particular effect and also by calculating a dose level associated with the
10- percent increase more commonly used in chemical risk assessments, and

 provide better justification for using a specific dose metric and identify
the important data gaps that could affect the results of those choices.

Three of 13 recommendations asked that EPA improve the section on exposure
to dioxins by evaluating the sources that contribute most to dioxins in the
food chain, discussing all ?special population? exposure in more detail, and
extending breast- feeding exposure scenarios beyond 1 year.

EPA made many additions and changes to the draft reassessment in response to
the peer review reports by both panels. For example, in response to
recommendations from both panels, EPA revised and added text in several
places to better explain the variety of dose metrics available and why body
burden is the best choice for assessing dioxins, while acknowledging that
EPA will need to address data gaps on body burden in the future as further
research is completed. Tables 1 and 2 in appendix IV highlight the actions
EPA took to address both panels? recommendations, suggestions, and concerns.

Overall, the peer review panels agreed with EPA?s approach to the
reassessment, and EPA generally addressed the recommendations, suggestions,
and concerns of the peer review panels. In a few cases, EPA disagreed with
the panels? recommendations or suggestions. In these cases, the agency
explained its position in the text and, in the case of the July 2000 panel,
addressed it in a separate written document. For example, although the Board
panel had recommended that EPA calculate a reference dose and add it to the
text, EPA chose to continue to use only the margin- of- exposure approach
and not calculate a reference dose. EPA stated in the revised draft report
that a calculated reference dose would be lower than most people?s daily
exposure and added a more detailed explanation of why it chose to use the
margin- of- exposure approach.

In addition to disagreeing with EPA on a few key scientific issues, the peer
review panels could not agree among themselves in some cases on EPA?s
findings. In such cases, the panels refrained from making recommendations or
suggestions to the agency. For example, members of both peer review panels
did not reach consensus on the strength of evidence used by EPA to support
the classification of TCDD as a human carcinogen and other dioxin compounds
as likely human carcinogens. EPA and Peer Reviewers

Do Not Agree on a Few Scientific Issues, and Uncertainties Remain Because
Data Are Lacking

Page 37 GAO- 02- 515 Draft Reassessment of Dioxins

EPA officials believe that the weight of scientific evidence on human and
animal exposure supports classifying TCDD as a known human carcinogen, a
view also held by WHO and the U. S. Department of Health and Human Services.

Although neither panel specifically recommended that EPA change its
classification of TCDD as a human carcinogen to a lesser category, such as a
likely human carcinogen, for various reasons most of the peer reviewers did
not endorse EPA?s classification. For example, while the July 2000 panel
agreed that TCDD is clearly a potent carcinogen in many species of animals,
most of the panel thought that human epidemiology studies were too limited,
and the results not consistent enough, to serve as a basis for showing
increased cancer mortality. As a result, the majority felt that the
characterization of TCDD as a known human carcinogen was not justified.
Similarly, the Board panel also noted limitations in the scientific data,
questioning the epidemiological data that indicated dioxins are carcinogens
in humans, as well as the data that supported similar modes of action
occurring in both animals and humans. Almost one- half of the Board did not
support classification of TCDD as a known human carcinogen for various
reasons. Those who did support the classification believed that the results
from studies of TCDD- exposed workers were persuasive and that the variety
of studies from researchers in different countries provided limited but
convincing evidence of TCDD?s carcinogenicity in humans.

A decade in the making, EPA?s draft reassessment report on dioxins was both
improved and limited by the passage of time, particularly in estimating the
daily dietary intake of dioxins by the typical American adult. That is, EPA
was able to include new food studies in the reassessment as they became
available. At the same time, however, these and earlier studies that EPA
relied on became less current with the passage of time. Overall, while EPA?s
draft reassessment report has advanced the state of knowledge on dietary
exposure to dioxins in the United States, the extent to which the estimate
accurately reflects current average daily exposure is not known. EPA
acknowledges the need for additional research on dietary intake, identifying
a number of data limitations associated with the estimates it developed in
its October 2001 draft report. Future efforts could eliminate most of the
food data limitations of the reassessment. Such efforts could include
periodic, comprehensive food surveys that analyze samples of the most
commonly eaten food products in each type of food studied, with samples
collected within the same time frames and analyses performed using
standardized Observations

Page 38 GAO- 02- 515 Draft Reassessment of Dioxins

methodologies. Further, when they become available, the results of the
ongoing EPA/ USDA follow- up study on dioxin levels in beef, pork, and
poultry should provide quantitative information on the changes, if any, in
dioxin levels in these foods from the mid- 1990?s to the present.

We provided EPA with a draft of this report for its review and comment and
the draft segment comparing EPA?s and WHO?s assessments of dioxins to WHO.
In commenting on the draft report, EPA?s assistant administrator, Office of
Research and Development, said that the report was well researched and
written and provided a balanced treatment of the information. However, EPA
believed that additional information on some of the data limitations
discussed in the section on EPA?s estimates of the dietary intake of dioxins
would better enable readers to evaluate the impact of the data limitations.
Where appropriate, we revised the report to reflect the views EPA presented
in its comments. For example, we added information concerning the strength
of the food concentration data used in estimating national mean levels of
exposure to dioxins, the sampling of animal fat rather than meat and poultry
products sold in grocery stores, and the likelihood that current dioxin
levels in food have significantly declined since the mid- 1990s. EPA?s
comments and our evaluation of them are provided in appendix V.

In commenting on the draft segment comparing EPA?s and WHO?s analyses, a
senior advisor of health and environment, the Department of Protection of
the Human Environment, World Health Organization, said the report was well
written and accurate.

To describe the types and extent of data EPA used to reassess human dietary
exposure to dioxins in the United States, we reviewed the relevant portions
of the October 2001 draft reassessment, the 1994 and 2000 drafts that were
peer reviewed, and the initial 1985 health risk assessment. We also reviewed
EPA documents and journal articles on the agency?s national sampling of
beef, pork, and poultry samples, and information about the other samples
used for milk, eggs, fish, dairy products, and vegetable fats. We discussed
the samples and methodology issues about them with EPA officials and
contractor staff. We did not validate or verify EPA?s estimates of dietary
exposure to dioxins.

To compare EPA?s objectives, processes, analytical methods, and conclusions
with those of WHO, we analyzed EPA?s October 2001 draft reassessment report
and various WHO publications on its objectives, Agency Comments

and Our Response Scope and Methodology

Page 39 GAO- 02- 515 Draft Reassessment of Dioxins

analyses, and conclusions. We discussed the similarities and differences
with EPA and WHO officials.

To determine the extent to which EPA?s draft dioxin reassessment reflects
the views of two independent peer review panels, we analyzed the
recommendations, suggestions, and concerns in the reports by the EPA Science
Advisory Board?s dioxin reassessment review subcommittee panel- on reviews
performed in 1994 and 2000- and a report from another independent peer
review panel on its July 2000 review. Recommendations of the Board panel
were noted in bold print in the executive summary, and we considered other
statements to be

?suggestions? when they were the consensus opinion of the panelists or the
opinion of a majority or of some of the panelists. We considered the July
2000 panel?s statements to be ?recommendations,? ?suggestions,? or

?concerns,? when those particular words were used in the executive summary
and where the statements reflected either a consensus or the opinion of a
majority or of some of the panelists. We also reviewed EPA documentation to
determine the changes EPA has made to its draft reassessment as a result of
being peer reviewed, including comparing the agency?s previous drafts of the
reassessment with each other and reviewing the written responses to the July
2000 panel?s recommendations and suggestions. We also met with EPA officials
to identify the agency?s responses to the panels? recommendations,
suggestions, and concerns, including discussing those with which it
disagreed.

We conducted our work from July 2001 through March 2002 in accordance with
generally accepted government auditing standards.

We will send copies of this report to the administrator, EPA, and make
copies available to others who request them. This report will also be
available on GAO?s Web site (www. gao. gov).

Page 40 GAO- 02- 515 Draft Reassessment of Dioxins

If you or your staff have questions about this report, please call me on
(202) 512- 3841. Key contributors to this report are listed in appendix VI.

David G. Wood Director, Natural Resources

and Environment

Appendix I: Major Milestones in the EPA and WHO Dioxin Risk Assessment
Efforts

Page 41 GAO- 02- 515 Draft Reassessment of Dioxins

Appendix I: Major Milestones in the EPA and WHO Dioxin Risk Assessment
Efforts

Late 1970s- 1984 1985- 87 1988 1991- 93

EPA efforts (1987)

WHO updates previously released

monographs on carcinogenicity of some dioxins.

(1987) WHO updates

previously released monographs on

carcinogenicity of some dioxins.

WHO efforts

1990

Experts at meeting in

Bilthoven, Netherlands,

establish a tolerable daily

intake for TCDD. Experts at

meeting in Bilthoven, Netherlands,

establish a tolerable daily

intake for TCDD.

1985 1977

WHO publishes initial cancer monograph that includes evaluation

of some dioxins. WHO

publishes initial cancer monograph that includes evaluation

of some dioxins. EPA releases

first comprehensive health assessment

document on dioxins. EPA releases

first comprehensive health assessment

document on dioxins.

EPA reviews and adopts proposal to use

interim toxic equivalency factor (TEF) procedures for estimating risks
associated with

exposures to mixtures of dioxins. EPA reviews and

adopts proposal to use

interim toxic equivalency factor (TEF) procedures for estimating risks

associated with exposures to mixtures of dioxins.

EPA completes external review

draft on estimating cancer risks posed

by dioxins. Science Advisory Board completes review

of external review draft. EPA completes

external review draft on estimating cancer risks posed

by dioxins. Science Advisory Board completes review

of external review draft. EPA performs

initial risk assessments of the dioxin TCDD.

EPA performs initial risk assessments of

the dioxin TCDD. (1991)

EPA begins current

dioxin reassessment.

EPA holds public meetings and

workshops. * 1991- 92 - EPA

holds public meetings. 1992 - EPA convenes two peerreview workshops.

 1993 - EPA holds a third peerreview workshop.

(1991) EPA begins current

dioxin reassessment.

EPA holds public meetings and

workshops.  1991- 92 - EPA holds public

meetings. 1992 - EPA convenes two peerreview workshops.

 1993 - EPA holds a third peerreview workshop.

Appendix I: Major Milestones in the EPA and WHO Dioxin Risk Assessment
Efforts

Page 42 GAO- 02- 515 Draft Reassessment of Dioxins

1994 1995 1997 2000 2001 2002 1998

EPA completes final internal review draft report, which incorporates

revisions in response to Science Advisory

Board and peer review panel recommendations

and comments. EPA completes

final internal review draft report, which incorporates

revisions in response to Science Advisory

Board and peer review panel recommendations

and comments. Science Advisory

Board and an independent panel of peer reviewers review

major segments of EPA?s revised

draft report. Science Advisory

Board and an independent panel of peer reviewers review

major segments of EPA?s revised

draft report. Dose- response

chapter undergoes peer

review. Dose- response

chapter undergoes peer

review. EPA provides

opportunities for public comment on

draft report on human health risks of

dioxins (150day comment period and 11 public meetings).

EPA submits draft report for Science Advisory

Board review. EPA provides

opportunities for public comment on

draft report on human health risks of

dioxins (150day comment period and 11 public meetings).

EPA submits draft report for Science Advisory

Board review. Science Advisory

Board completes review of draft report, approving

most chapters but recommending

revisions to some parts. Science Advisory

Board completes review of draft report, approving

most chapters but recommending

revisions to some parts.

Dissemination of final dioxin reassessment report for

review by other federal

agencies is pending.

Publication of final dioxin reassessment

report is pending. Dissemination

of final dioxin reassessment report for

review by other federal

agencies is pending.

Publication of final dioxin reassessment

report is pending.

WHO publishes new monographs,

based on reassessments, on the

evidence for carcinogencity of

TCDD and several other

dioxins. Experts at meeting in Stockholm, Sweden, derive consensus TEFs

for dioxins for human, fish, and wildlife risk

assessments. WHO publishes

new monographs, based on reassessments, on the evidence for

carcinogencity of TCDD and several other

dioxins. Experts at meeting in Stockholm, Sweden, derive consensus TEFs for
dioxins for

human, fish, and wildlife risk

assessments. WHO publishes the

consensus TEFs from the Stockholm meeting.

Experts meeting in Geneva, Switzerland, reevaluate the risks to human

health from dioxins and revise the tolerable

daily intake for dioxins. WHO publishes the

consensus TEFs from the Stockholm meeting.

Experts meeting in Geneva, Switzerland, reevaluate the risks to human

health from dioxins and revise the tolerable

daily intake for dioxins. Experts at meeting

in Rome, Italy, evaluate risks associated with dioxins in foods and
establish a provisional tolerable

monthly intake for dioxins. Experts at meeting

in Rome, Italy, evaluate risks associated with dioxins in foods and
establish a provisional tolerable

monthly intake for dioxins.

Appendix II: Comparison of the Major Conclusions from EPA?s and WHO?s Dioxin
Risk Assessments

Page 43 GAO- 02- 515 Draft Reassessment of Dioxins

EPA conclusions WHO conclusions Effects associated with exposure to dioxins

Exposure to dioxins can produce a wide variety of effects in animals
(including cancer and noncancer health effects) and might produce many of
the same effects in humans.

EPA characterizes dioxin and related compounds as carcinogenic and
developmental, reproductive, immunological, and endocrinological hazards and
makes the following specific points. a

 Exposure to TCDD leads to an increased risk of generalized cancers at
multiple organ sites, including lung cancer.

 Long- term noncancer consequences of exposure to TCDD in adults include
chloracne, elevated gamma glutamyl transferase levels, and altered
testosterone levels. b Among the possible noncancer consequences of exposure
to TCDD or other dioxin and dioxin- like compounds are dermatological
conditions such as chloracne; liver diseases; and kidney, nervous system,
and lung disorders.

 Although available data suggest an association between TCDD exposure and
other adverse outcomes, further study is required of circulatory and heart
disease, diabetes and glucose metabolism, reproductive and developmental
outcomes, and immunologic disorders.

Exposure to dioxins may be linked to a variety of adverse effects.

 Short- term human exposure to high levels of dioxins may result in skin
lesions (such as chloracne) and altered liver function.

 Long- term exposure is linked to impairment of the immune system, the
developing nervous system, the endocrine system, and reproductive functions.

 Chronic animal exposure to dioxins has resulted in several types of
cancer. Human data from occupational or accidental exposure has produced
evidence of increased risks for all cancers combined, along with less strong
evidence of increased risks for cancers of particular sites.

Mode of action through which exposure to dioxins can lead to adverse effects

Dioxins are structurally related and elicit their effects through a common
mode of action- binding of dioxins to a cellular protein called the aryl
hydrocarbon receptor. Binding to the aryl hydrocarbon receptor appears to be
necessary for all well- studied effects of dioxins but is not sufficient, in
and of itself, to elicit these responses.

TCDD and related compounds have a common mode of action in animals and
humans. Therefore, there is no reason to expect, in general, that humans
would not be similarly affected as animals at some dose.

A broad variety of data has shown the importance of the aryl hydrocarbon
receptor in mediating the biological effects of dioxins. The precise chain
of molecular events by which the receptor elicits these effects is not yet
fully understood. However, alterations in key biochemical and cellular
functions are expected to form the basis for dioxin toxicity.

Experimental data indicate that TCDD and probably other polychlorinated
dibenzo- p- dioxins (CDD) and polychlorinated dibenzofurans (CDF) are not
direct- acting genotoxic agents (i. e., do not directly affect genetic
material).

Dioxins act through the same mode of action in animals and humans.

Appendix II: Comparison of the Major Conclusions from EPA?s and WHO?s Dioxin
Risk Assessments

Appendix II: Comparison of the Major Conclusions from EPA?s and WHO?s Dioxin
Risk Assessments

Page 44 GAO- 02- 515 Draft Reassessment of Dioxins

EPA conclusions WHO conclusions Use of the toxicity equivalency (TEQ)
concept

EPA and the international scientific community have adopted TEQ of dioxins
as prudent science policy.

(EPA recommended that the TEFs derived by WHO in 1997- published in 1998- be
used to assign TEQ to complex environmental mixtures for assessment and
regulatory purposes.)

The complex nature of CDD, CDF, and polychlorinated biphenyls (PCB) mixtures
complicates the risk evaluation for humans. The concept of TEFs has been
developed to facilitate risk assessment and regulatory control of exposure
to these mixtures.

(WHO derived updated consensus TEFs for 29 dioxins in 1997, with the results
of the meeting published in 1998. Subsequent WHO assessments of dioxins used
this updated set of TEFs for their calculations.)

Whether dioxins are human carcinogens

Complex mixtures of dioxins are highly potent, ?likely? human carcinogens.

 A weight- of- the- evidence evaluation suggests that mixtures of dioxins
are strong cancer promoters and weak direct or indirect initiators and are
likely to present a cancer hazard to humans. c

 Because dioxins and related compounds always occur in the environment and
in humans as complex mixtures of individual congeners, it is appropriate
that the characterization apply to the mixture.

Individual congeners can also be characterized as to their carcinogenic
hazards.

 TCDD is best characterized as ?carcinogenic to humans.? Based on the
weight of all evidence (human, animal, and mode of action), TCDD meets the
criteria that allow EPA and the scientific community to accept a causal
relationship between TCDD exposure and cancer hazard.

 Other individual dioxin- like compounds are characterized as

?likely to be human carcinogens? primarily because of the lack of
epidemiological evidence associated with their carcinogenicity, although the
inference based on TEQ is strong that they would behave in humans as TCDD
does. Other factors, such as the lack of compound- specific chronic animal
studies, also support this characterization.

TCDD is a human carcinogen (group 1), considering limited evidence in
humans, sufficient evidence in experimental animals, and evidence of a mode
of action that functions the same way in humans as in experimental animals.
d

Other dioxins are not classifiable as to their carcinogenicity to humans
(group 3). Depending on the specific compound evaluated, the International
Agency for Research on Cancer (IARC) noted that the available data provided
inadequate evidence for carcinogenicity in humans or limited evidence,
inadequate evidence, or evidence suggesting lack of carcinogenicity in
experimental animals.

Appendix II: Comparison of the Major Conclusions from EPA?s and WHO?s Dioxin
Risk Assessments

Page 45 GAO- 02- 515 Draft Reassessment of Dioxins

EPA conclusions WHO conclusions Whether there appears to be a ?threshold? or
safe dose of dioxins that would not cause adverse effects

The supposition of a response threshold for receptor- mediated effects (such
as those associated with dioxins? binding to the aryl hydrocarbon receptor )
is a subject for scientific debate. The same receptor occupancy assumption
of the classic receptor theory is interpreted by different parties as
support for and against the existence of a threshold.

Empirical dose- response data from cancer studies do not provide consistent
or compelling support for threshold models and are insufficient to move from
EPA?s default policy of linear extrapolation (an approach that assumes there
is no threshold of exposure without risk).

Threshold levels of lifetime exposure to dioxins that would cause toxic
noncancer effects may be below the current level of background exposure and
body burdens, and, therefore, the potential exists for noncancer risk at
background exposure.

TCDD does not affect genetic material, and there is a level of exposure
below which cancer risk would be negligible.

Although TCDD is classified by IARC as a human carcinogen, it is not
considered to be a direct acting carcinogen. Therefore, a threshold approach
could be used in the hazard assessment approach.

A tolerable intake can be established for TCDD on the basis of the
assumption that there is a threshold for all effects, including cancer.
Because cancer occurred in animals at higher body burdens than other toxic
effects, establishing a tolerable intake on the basis of noncancer effects
would also address any carcinogenic risk.

Whether it is useful to set a dose or exposure level that the public could
experience for a lifetime without expectation of harm

EPA did not calculate reference dose or reference concentration values in
this reassessment as it generally does for noncancer effects in other
assessments. Instead, EPA chose to characterize the margins of exposure
between estimated actual human exposure and the exposure levels at which
studies indicated various adverse noncancer effects could occur.

The WHO 1998 consultation set daily limits on exposure levels of dioxins for
non- cancer effects, a tolerable daily intake.

The Joint Expert Committee on Food Additives of the United Nation?s Food and
Agriculture Organization and WHO set a provisional tolerable monthly intake
limit on exposure levels to dioxins, again focusing on noncancer effects.
The Committee participants felt that it was more appropriate to express the
tolerable intake on a monthly rather than a daily basis because of the long
half- life of dioxins (i. e., the body?s stored dioxins decline slowly, with
only half of the accumulated dioxins disappearing over about 7 years).

Appendix II: Comparison of the Major Conclusions from EPA?s and WHO?s Dioxin
Risk Assessments

Page 46 GAO- 02- 515 Draft Reassessment of Dioxins

EPA conclusions WHO conclusions Human exposure to dioxins

Human exposure to dioxins has occurred through background exposure,
contamination of foods, occupational exposure, and exposure associated with
industrial accidents. An increased background exposure can result from
either a diet that favors consumption of foods high in dioxin content or a
diet that is disproportionately high overall in animal fats.

Most (more than 95 percent) background exposure results from the presence of
minute amounts of dioxins in dietary fat, primarily from the commercial food
supply.

The average dioxin tissue level for the general U. S. adult population
appears to be declining.

Five compounds account for most (about 80 percent) of the toxicity in human
tissue concentrations.

Human exposure to dioxins may occur through background (environmental)
exposure and accidental and occupational contamination.

Over 90 percent of human background exposure is estimated to occur through
the diet, with food from animal origin being the predominant source.

Recent studies show decreasing levels of dioxins in food and consequently a
significantly lower dietary intake of these compounds.

Risks of adverse health effects at the general public?s current levels of
exposure to dioxins

In general, EPA?s assessments indicated that dioxins pose risks at lower
levels of exposure than previously estimated and that the general public?s
current levels of exposure are at or near those that have been observed to
cause harm.

EPA estimates that the upper bound cancer risk at average current background
body burdens exceeds 10 -3 (i. e., the upper bound on general population
lifetime risk for all cancers might be on the order of 1 in 1,000 or more).
However, this is an upper bound estimate, so the true risks are likely less
than that and may be zero for most people.

 In 1985, EPA?s estimate of the cancer slope factor based on exposure to
TCDD was 1. 6 x 10 -4 per picogram of TCDD per kilogram of body weight per
day (pgTCDD/ kgBW/ day). e

 EPA?s current upper bound slope factor for estimating human cancer risk on
the basis of human data is 1 x 10 -3 per pgTCDD/ kgBW/ day.

 EPA?s current upper bound slope factor for estimating human cancer risk on
the basis of animal data is 1.4 x 10 -3 per pgTCDD/ kgBW/ day.

EPA estimated that U. S. residents are exposed daily to about 1 picogram of
dioxins per kilogram of body weight, which is close to the level that caused
biological changes in animals. EPA noted that the margins of exposure
between estimated actual human exposure and the exposure levels at which
studies indicated adverse noncancer health effects could occur were
?considerably

less than typically seen for environmental contaminants of toxicologic
concern.?

In general, WHO?s assessments also indicated that dioxins pose risks at
lower levels of exposure than previously estimated and that the general
public?s current levels of exposure are at or near those that have been
observed to cause harm.

In 1990, WHO experts had established a tolerable daily intake for TCDD of 10
picograms per kilogram of body weight. In 1998, the WHO consultation
established a tolerable daily intake for dioxins at a range of 1- 4 TEQ
picograms per kilogram of body weight and noted that subtle effects may
already occur in the general population at current background levels of 2 to
6 picograms per kilogram of body weight. The consultation stressed that the
ultimate goal is to reduce human intake levels below 1 picogram TEQ per
kilogram of body weight per day.

In 2001, Joint Expert Committee on Food Additives of the United Nation?s
Food and Agriculture Organization of the United Nations and WHO determined
that a monthly tolerable intake level made more sense than a daily level and
established a provisional tolerable monthly intake of 70 picograms per
kilogram of body weight per month (equivalent to 2.33 picograms per day) for
dioxins.

The various WHO entities did not calculate quantitative cancer risk
estimates for the additional cancer risk that dioxins might pose to the
general population. However, WHO did explore the calculation of a cancer
?benchmark dose? (the dose or body burden estimated to result in a 1-
percent increase in cancer mortality) through various models. On the basis
of data from three industrial exposure studies, WHO estimated that the body

Appendix II: Comparison of the Major Conclusions from EPA?s and WHO?s Dioxin
Risk Assessments

Page 47 GAO- 02- 515 Draft Reassessment of Dioxins

EPA conclusions WHO conclusions

burden of dioxins associated with a 1- percent excess cancer risk over a
lifetime was 3 to 13 nanograms per kilogram of body weight, which is
associated with a daily dose of dioxins in the range of 2 to 7 picograms per
kilogram of body weight per day.

Risks to population subgroups

Children?s risks from dioxins and related compounds may be greater than that
of adults, but more data are needed to fully address the issue.

There may be individuals in the population who might experience a higher
cancer risk on the basis of genetic factors or other determinants of cancer
risk not accounted for in epidemiologic data or animal studies. In
particular, a very small percentage of the population (less than 1 percent)
may experience risks that are 2 to 3 times higher than the general
population estimate if their individual response is at the upper bound and
they are among the most highly exposed based on dietary intake of dioxins.

Certain population subgroups are at greater risk from dioxins. Fetuses are
most sensitive to dioxin exposure, and newborns may also be more vulnerable
to certain effects. Some individuals or groups of individuals may be exposed
to higher levels of dioxins because of their diets or occupations.

a Endocrinological hazards are those related to the system of ductless
glands that secrete hormones directly into the blood stream for distribution
throughout the body, such as the pituitary, thyroid, and adrenal glands. b
Elevated gamma glutamyl transferase levels are among the changes in liver
function and structure that have been observed using human data. c According
to EPA?s revised proposed guidelines for carcinogen risk assessment, the
descriptor

?likely to be a human carcinogen? is appropriate when the available tumor
effects and other key data are adequate to demonstrate carcinogenic
potential to humans, yet not sufficient to infer a cause and effect
relationship. d For additional information on WHO- International Agency for
Research on Cancer?s (IARC) evaluation

categories and the definitions of degrees of evidence, see the Preamble to
the IARC Monographs

available on the IARC Internet site (http:// www. iarc. fr/). e The cancer
slope factor is an upper bound of the probability of cancer risk in the
population. According to EPA, the slope factor generally approximates or
exceeds a 95- percent confidence limit, meaning that there is a greater than
95- percent chance that cancer risks will be less than the upper bound.

Source: GAO review of EPA and WHO documents on dioxin reassessment efforts.

Appendix III: Questions EPA Asked Peer Review Panels to Address

Page 48 GAO- 02- 515 Draft Reassessment of Dioxins

EPA sought expert opinions from both a July 2000 panel of independent peer
reviewers and a November 2000 Science Advisory Board expert panel on several
key questions that pertain to the content of the documents under review. The
questions are classified into 11 general topics. Most of the questions are
the same for both panels. However, according to usual Science Advisory Board
practice, EPA staff, Board staff, and the chair of the Board?s dioxin
reassessment review subcommittee jointly developed additional questions for
the Board?s review. Table 6 and table 7 show the topics and questions
addressed by the July 2000 panel, and Board panel, respectively.

Table 6: Questions for the July 2000 Panel Review of EPA?s Draft Dioxin
Reassessment Topic Question

Body burden 1. Did EPA adequately justify its use of body burden as a dose
metric for inter- species scaling? Should the document present conclusions
based on daily dose? Use of margin- of- exposure approach 2. How might the
rationale be improved for EPA?s decision not to calculate a reference

dose/ reference concentration, and for the recommended margin- of- exposure
approach for conveying risk information? Is a margin- of- exposure approach
appropriate, as compared to the traditional reference dose/ reference
concentration? Should the document present a reference dose/ reference
concentration? 3. Are the calculations of a range of effective dose (ED) 01
body burden for noncancer effects in rodents responsive and clearly
presented? Please comment on the weight of evidence interpretation of the
body burden data associated with a 1 percent response rate for noncancer
effects that is presented in Chapter 8, appendix I and figure 8- 1 (where
EPA considers that the data best support a range estimate for ED 01 body
burdens from 10nh/ kg to 50 ng/ kg). Mechanisms and mode of action 4. How
might the discussion of mode of action of dioxin and related compounds be

improved? 5. Despite the lack of congener- specific data, does the
discussion in the Integrated Summary and Risk Characterization support EPA?s
inference that these effects may occur for all dioxin- like compounds, based
on the concept of toxicity equivalence? TEFs 6. Is the history, rationale,
and support for the TEQ concept, including its limitations and

caveats, laid out by EPA in a clear and balanced way in Chapter 9? Did EPA
clearly describe its rationale for recommending adoption of the 1998 WHO
TEFs? 7. Does EPA establish clear procedures for using, calculating, and
interpreting toxicity equivalence factors? Noncancer effects 8. Have the
available human data been adequately integrated with animal information

in evaluating likely effect levels for the noncancer endpoints discussed in
the reassessment? 9. Do reviewers agree with the characterization of human
developmental, reproductive, immunological, and endocrinological hazard?
What, if any, additional assumptions and uncertainties should EPA embody in
these characterizations to make them more explicit? Cancer effects 10. Do
you agree with the characterization in this document that dioxin and related

compounds are carcinogenic hazards for humans? 11. Does the document clearly
present the evolving approaches to estimating cancer risk (e. g. margin of
exposure and the lower limit on ED 01 as a point of departure) as described
in EPA?s April 1996 ?Proposed Guidelines for Carcinogenic Risk Assessment??
Is this approach equally as valid for dioxin- like compounds?

Appendix III: Questions EPA Asked Peer Review Panels to Address

Appendix III: Questions EPA Asked Peer Review Panels to Address

Page 49 GAO- 02- 515 Draft Reassessment of Dioxins

Topic Question

12. Please comment on the presentation of the range of upper bound risks for
the general population based on this reassessment. What alternative
approaches should be explored to better characterize quantitative aspects of
potential cancer risk? Is the range that is given sufficient, or should more
weight be given to specific data sources? Background and population
exposures 13. Have the estimates of background exposures been clearly and
reasonably

characterized? 14. Has the relationship between estimating exposure from
dietary intake and estimating exposure from body burden been clearly
explained and adequately supported? 15. Have important ?special populations?
and age- specific exposures been identified and appropriately characterized?
Children?s risk 16. Is the characterization of increased or decreased
childhood sensitivity to possible

cancer and noncancer outcomes scientifically supported and reasonable? Is
the weight- of- evidence approach appropriate? Relative risks of breast
feeding 17. Has EPA adequately characterized how nursing affects short- term
and long- term

body burdens of dioxins and related compounds? Risk characterization summary
statement 18. Does the summary and analysis support the conclusion that
enzyme induction,

changes in hormone levels, and indicators of altered cellular function seen
in humans and laboratory animals represent effects of unknown clinical
significance, but they may be early indicators of toxic response? 19. Has
the short summary statement in the risk and hazard characterization on page
107 adequately captured the important conclusions and the areas where
further evaluation is needed? What additional points should be made in this
short statement? Sources 20. Are these sources adequately described and are
the relationships to exposure

adequately explained? Source: EPA.

Table 7: Questions for the November 2000 Science Advisory Board Panel Topic
Question

Body burden 1. Did EPA adequately justify its use of body burden as a dose
metric for inter- species scaling? Should the document present conclusions
based on daily dose? Use of margin- of- exposure approach 2. Has EPA?s
choice of the margin- of- exposure approach to risk assessment adequately

considered that background levels of dioxins have dropped dramatically over
the past decade and are continuing to decline? How might the rationale be
improved for EPA?s decision not to calculate a reference dose/ reference
concentration, and for the recommended margin- of- exposure approach for
conveying risk information? Is a marginof- exposure approach appropriate, as
compared to the traditional reference dose/ reference concentration? Should
the document present a reference dose/ reference concentration? 3. Are the
calculations of a range of ED 01 body burden for noncancer effects in
rodents responsive and clearly presented? Please comment on the weight- of-
evidence interpretation of the body burden data associated with a 1- percent
response rate for noncancer effects that is presented in chapter 8, appendix
I and Figure 8- 1 (where EPA considers that the data best support a range
estimate for ED 01 body burdens from 10nh/ kg to 50 ng/ kg).

Appendix III: Questions EPA Asked Peer Review Panels to Address

Page 50 GAO- 02- 515 Draft Reassessment of Dioxins

Topic Question

Mechanisms and mode of action 4. How might the discussion of mode of action
of dioxin and related compounds be improved? 5. Despite the lack of
congener- specific data, does the discussion in the Integrated Summary and
Risk Characterization support EPA?s inference that these effects may occur
for all dioxin- like compounds, based on the concept of toxicity
equivalence? TEFs 6. Is the history, rationale, and support for the TEQ
concept, including its limitations and

caveats, laid out by EPA in a clear and balanced way in Chapter 9? Did EPA
clearly describe its rationale for recommending adoption of the 1998 WHO
TEFs? 7. Does EPA establish clear procedures for using, calculating, and
interpreting toxicity equivalence factors? Noncancer effects 8. Have the
available human data been adequately integrated with animal information in

evaluating likely effect levels for the noncancer endpoints discussed in the
reassessment? Has EPA appropriately defined noncancer adverse effects and
the body burdens associated with them? Has EPA appropriately reviewed,
characterized, and incorporated the recent epidemiological evidence for
noncancer risk assessment for human population? 9. Do reviewers agree with
the characterization of human developmental, reproductive, immunological,
and endocrinological hazard? What, if any, additional assumptions and
uncertainties should EPA embody in these characterizations to make them more
explicit? Cancer effects 10. Do you agree with the characterization in this
document that dioxins and related

compounds are carcinogenic hazards for humans? Does the weight of the
evidence support EPA?s judgment concerning the listing of environmental
dioxins as a likely human carcinogen? 11. Does the document clearly present
the evolving approaches to estimating cancer risk (e. g., margin of exposure
and the lower limit on ED 01 as a point of departure) as described in EPA?s
1996 ?Proposed Guidelines for Carcinogenic Risk Assessment?? Is this
approach equally as valid for dioxin- like compounds? Has EPA appropriately
reviewed, characterized, and incorporated the recent epidemiological
evidence for cancer risk assessment for human populations? 12. Please
comment on the presentation of the range of upper bound risks for the
general population based on this reassessment. What alternative approaches
should be explored to better characterize quantitative aspects of potential
cancer risk? Is the range that is given sufficient or should more weight be
given to specific data sources? Background and population exposures 13. Have
the estimates of background exposures been clearly and reasonably

characterized? 14. Has the relationship between estimating exposure from
dietary intake and estimating exposure from body burden been clearly
explained and adequately supported? Has EPA adequately considered available
models for the low- dose exposure- response relationships (liner, threshold,
?J? shaped)? 15. Have important ?special populations? and age- specific
exposures been identified and appropriately characterized? Children?s risk
16. Is the characterization of increased or decreased childhood sensitivity
to possible

cancer and noncancer outcomes scientifically supported and reasonable? Is
the weightof- evidence approach appropriate? Relative risks of breast
feeding 17. Has EPA adequately characterized how nursing affects short- term
and long- term

body burdens of dioxins and related compounds?

Appendix III: Questions EPA Asked Peer Review Panels to Address

Page 51 GAO- 02- 515 Draft Reassessment of Dioxins

Topic Question

Risk characterization summary statement 18. Does the summary and analysis
support the conclusion that enzyme induction, changes in hormone levels, and
indicators of altered cellular function seen in humans and laboratory
animals represent effects of unknown clinical significance, but they may be
early indicators of toxic response? 19. Has the short summary statement in
the risk and hazard characterization on page 107 adequately captured the
important conclusions and the areas where further evaluation is needed? What
additional points should be made in this short statement? Sources 20. Are
these sources adequately described and are the relationships to exposure

adequately explained? Source: EPA.

Appendix IV: EPA?s Responses to Peer Review Panels

Page 52 GAO- 02- 515 Draft Reassessment of Dioxins

EPA generally addressed the peer review panels? comments by performing
additional analyses, adding or revising text, or identifying comments as
related to EPA?s long- term research goals. In some instances, EPA thought
that the reassessment already addressed the panel?s comment. The panels
classified their recommendations, suggestions, and concerns, and EPA
responded to each. Tables 8 and 9 show the comments made by the panels and
EPA?s response or action taken.

Table 8: EPA?s Responses to July 2000 Panel?s Report Recommendation EPA
response or action

1. Use terminology such as ?ambient exposures? or ?general

population exposures,? rather than the term ?background

exposure,? which implies normal and acceptable. EPA prefers to use
?background exposure? as it appropriately

recognizes the ubiquitous nature of trace amounts of dioxins in the
environment and food supply even when no sources are identified nearby. EPA
added a definition to its glossary. 2. Present more detail (e. g., sample
calculations) in the Integrated Summary on exactly how the cancer slope
factor was derived. Additional information has been added to Section 5 of
the

Integrated Summary to clearly illustrate how the cancer slope factors were
derived. 3. The panel thought that the upper bound cancer risk of 10 -3 to
10 -2 in the general population, implying an additional 3,000 to 30,000
deaths per year, was alarmist, not warranted and not realistic. Recommended
that EPA should present ?reality check? on the risk estimates relative to
highly exposed past cohorts.

EPA states its estimates were derived from the best data sets available.

4. Include a table to summarize the various noncancer effects observed in
animals and humans at low- level exposures. EPA has added a table (table 2-
2) to the Integrated Summary of

the September 2000 draft. While not extensive, it illustrates the low range
of margins of exposure that is calculated for a variety of effects in
several species, including humans. 5. The panel thought body burden was an
appropriate dose metric. However, the panel recommended that EPA explicitly
explain the relationship among daily intake, serum levels, tissue dose, and
body burden.

Additional discussion of alternative dose metrics has been included in the
Health Assessment of TCDD and Related Compounds (Part II), Chapter 1:
Disposition and Pharmocokinetics. This discussion has also been carried over
to Section 1.3 of the Integrated Summary. The utility, strengths, and
weaknesses of each are presented, and in a number of cases the relationships
of one to another are discussed.

Suggestion EPA response or action

1. Data presented on dioxin levels in food are an improvement over earlier
drafts, but need more specific information on the number of samples
collected, sampling locations, and standard deviations of observed levels
presented in tables 4- 6 and 4- 8.

This additional information was included in Estimating Exposure to Dioxin-
Like Compounds (Part I), Vol. 3, Chapter 4, and has been now added to the
Integrated Summary (see table 4- 5). This table presents dioxin levels in
environmental media and food, along with number of samples, mean, range, and
standard deviation. 2. Revise and expand discussion of dioxin levels in
food, identifying levels of dioxins in other food sources for which data are
available; listing food sources that have not been extensively characterized
(i. e., fish); commenting on changing rate of dioxins in food sources over
the years; and addressing the effects of cooking practices.

Several new paragraphs have been added/ edited in the

Integrated Summary, Section 4, to address these comments.

Appendix IV: EPA?s Responses to Peer Review Panels

Appendix IV: EPA?s Responses to Peer Review Panels

Page 53 GAO- 02- 515 Draft Reassessment of Dioxins

3. Reviewers thought EPA adequately derived approaches to estimate average
daily dose from both dietary intake and body burden. Suggested revisions
included (a) providing a clear definition of body burden and explaining how
body burden relates to tissue levels; (b) presenting equations and sample
calculations in the Integrated Summary to illustrate how average daily dose
can be estimated from dietary intake or from body burden; (c) considering
other sources of data for characterizing trends in body burden levels; and
(d) providing additional detail on the variability in the distribution of
estimated average daily intakes.

(a) EPA added a definition of body burden, and an explanation of how body
burden relates to tissue levels, to the glossary in the

Integrated Summary; (b) equations illustrating how average daily dose can be
estimated from dietary intake or from body burden were in Part I, Vol. III,
Chapter 4, but were not included in the

Integrated Summary in the interest of brevity; (c) although body burden
trends (e. g., differences in age) in the reassessment document are not
statistically based, a current modeling study is underway to more fully
understand body burden trends; (d) section 4.4.3 in the Integrated Summary
on variability in intake levels has been expanded to include key references
and a new discussion on the Center to Disease Control and Prevention blood
study to further support the findings on variability. EPA is also currently
investigating the possibility of using probability methods to further study
variability of dioxin exposure. 4. Reviewers thought EPA identified
important ?special

populations? of highly exposed individuals and suggested that the agency
consider including others, such as people who lose weight rapidly, fetuses,
and people who eat large amounts of potentially contaminated food sources
not explicitly considered in the reassessment (e. g., lamb).

In Part I, Vol. III, Chapter 5, EPA analyzed a large amount of available
data on these special populations: nursing infants (Section 5.2), people who
fish (Section 5. 3), people living near sources of dioxin release (Section
5.4), and cigarette smokers (Section 5.5). Other populations (such as
exposed workers or those living in Seveso) were discussed in detail in Part
II. 5. Reviewers thought EPA may have overstated upper bound risks and
suggested EPA more clearly describe the basis of the current cancer slope
factor and significance of upper bound cancer risks to public.

The text has been revised to put the upper bound estimate of risk in better
perspective. The previous range of upper bound risks was apparently
confusing and has been removed.

6. Most reviewers agreed that developmental, reproductive, immunological,
and endocrinological noncancer effects could be seen in humans, given
sufficient dose. Reviewers suggested EPA improve the justification for the
conclusion that human epidemiological data suggest that noncancer effects
occur at ambient exposures.

Additional discussion has been added to Sections 2, 5, and 6 of the
Integrated Summary to address this issue.

7. The panel thought the Integrated Summary presented a reasonable argument
that cancer risk associated with breastfeeding is likely low and suggested
EPA provide similar argument for noncancer effects.

EPA asserts that the argument that noncancer risk associated with
breastfeeding is also low is already in the report, although it is not
broken out into a separate section. EPA agrees with WHO that on balance, the
benefits of breastfeeding outweighed risks of dioxin exposures. 8. The panel
agreed that the Integrated Summary needs additional discussion on the
uncertainties associated with using various dose metrics specifically for
evaluating childhood risks. Some reviewers continued to have reservations
about EPA?s selection of the body burden dose metric for children,
especially considering that children?s (especially nursing infants?) doses
can be much higher than those of adults, even though their body burdens
often are not.

See recommendation above relating to selection of a dose metric. Figures 4-
4 and 5- 2 were added to the Integrated Summary to help illustrate the
rationale for selecting body burden as the dose metric using a nursing
scenario, and expanded discussion can be found in Section 4. Nonetheless,
uncertainty remains regarding the most appropriate dose metric for any given
effect.

9. Most of the reviewers agreed with the use of margin of exposure to
express exposures rather than comparing exposures with reference dose given
the assumptions made in the assessment, but they suggested the implication
of these assumptions be more clearly defined.

Additional discussion regarding the concept of margin of exposure has been
included in Sections 5 and 6 of the Integrated Summary. A table has been
added to illustrate the concept for several cancer and noncancer endpoints.
Additional details have been added to the discussion regarding the decision
to use a margin of exposure rather than calculate a reference dose.

Appendix IV: EPA?s Responses to Peer Review Panels

Page 54 GAO- 02- 515 Draft Reassessment of Dioxins

10. The reviewers thought the Integrated Summary clearly presented the
entire data set of dose response data that met EPA?s selection criteria, but
that presentation should be improved. Many thought that EPA should attempt
to differentiate effects that are ?frank manifestations of toxicity? from
effects with unknown clinical significance.

Additional discussion has been added to the text to provide this
differentiation, as suggested by the peer reviewers. EPA believes
differentiating effects is inherently difficult since the manifestations of
toxicological response lie along a continuum and biochemical changes may
serve as a biomarker of the potential for frank response. 11. The reviewers
generally agreed that Chapter 9 on TEQs in Part II, presented the history,
rationale and support for the TEQ approach for evaluating dioxin toxicity,
but they were concerned that this approach attributes dioxin toxicity to
compounds for which few toxicologic data are available. Though the reviewers
felt that Chapter 9 establishes clear procedures for using, calculating, and
interpreting TEQs, they stated certain topics needed to be described more
clearly and suggested EPA concisely state why it selected WHO?s 1998 TEFs
over previously used TEFs, present example TEQ calculations as an appendix,
and should stress that risk assessors should characterize fate and transport
of individual dioxins separately.

Chapter 9 on TEQs has been revised in response to peer reviewers? comments.
Additional discussion has been added to the Chapter to focus on 5 compounds
that make up greater than 70 percent of human exposure and body burden on a
TEQ basis. While several of the minor compounds have limited toxicologic
data supporting their TEF values, the major compounds have robust data sets.
This discussion has been carried over to the

Integrated Summary in Section 1.2. Source: EPA.

Table 9: EPA?s Responses to Science Advisory Board Panel?s Comments
Recommendation EPA action or response

1. Carry out additional work on the exposure assessment section to evaluate
sources that make the greatest contribution to dioxins in the food chain,
and make the text consistent with the tables.

This is a long- term research goal of EPA. EPA officials stated that they
interpret the Science Advisory Board recommendation as basically endorsing
what EPA plans to do in the future- linking sources of dioxins with
exposures. The minor issue regarding making the table and text consistent
was resolved by EPA. 2. Include discussion of all ?special population?
exposures in the summary document. EPA added additional information
regarding ?special population?

exposures (i. e., some Native American subsistence fishers could be highly
exposed to dioxins depending on the amount of fish they catch and where).
According to EPA officials, the agency now addresses this issue in the
Integrated Summary -Sections

4 (exposure) and Section 6 (risk characterization) and specifically mentions
Native Americans in the text. 3. Extend breastfeeding exposure scenarios
beyond 1- year. According to EPA officials, they performed additional
analysis

and revised the related text. See Integrated Summary, Section 6. 4. For
human carcinogen designation, better understanding and interpretation of
epidemiological data are needed. Add expected differences in results between
epidemiological studies of genotoxic agents and cancer promoters.

EPA added text in Part II and the Integrated Summary regarding the
expectations for epidemiological studies for strong cancer promoters and
will be including a discussion of new cancer studies in the reassessment to
provide the latest on this issue. 5. Methodology: Agree with use of margin-
of- exposure approach, but in addition calculate a reference dose. EPA
disagreed and chose not to calculate a reference dose, but

explanation of why it did not explained in more detail. See pp. 118- 122 in
the Integrated Summary.

6. Methodology: In future reevaluations develop quantitative estimates of
noncancer risk- similar to those developed for cancer- to the extent methods
become feasible.

EPA?s opinion is that to some extent it is already merging cancer and
noncancer methods using margin- of- exposure analysis for both cancer and
noncancer effects. The text tries to balance the discussion of cancer and
noncancer risks, but noncancer risks cannot be done quantitatively.

Appendix IV: EPA?s Responses to Peer Review Panels

Page 55 GAO- 02- 515 Draft Reassessment of Dioxins

7. Further investigation of noncancer hazards is needed. They receive
insufficient attention on pp. 7 and 11. About half of the panel believes
that the current draft assessment may overestimate the likely cancer hazard.

EPA revised the text to put noncancer effects into better perspective, but
officials acknowledged that the tools used to describe cancer risks are
easier for people to understand. In the text of the report, EPA is providing
more discussion on noncancer effects by providing examples where possible.
8. The panel agreed with using body burden as the dose metric; however,
better justification for using a specific dose metric such as body burden is
needed. Provide more explicit examples of how different dose metrics might
apply to specific toxic endpoints.

EPA officials said that the agency has revised and added text in several
places to better explain the variety of dose metrics available and why body
burden is the best choice for assessing dioxin. EPA revised Part II, Chapter
1 on dose metrics and it also made changes to Chapter 8 and added text in
the Integrated Summary, Sections 1,5, and 6. EPA recognized the need to
better explain that using other dose metrics rather than body burden in
certain situations is also acceptable. 9. EPA should identify important data
gaps on body burden (i. e., how it varies with age and in females depending
on number of offspring) to highlight research opportunities.

EPA officials stated that research opportunities in the future will address
this issue. EPA will be incorporating new studies in future dioxin
assessments, particularly those that look at population dynamics (i. e.,
younger people starting now with lower intake levels /body burdens than in
past) as they become available. Currently, there is major work under way at
the Center for Disease Control and Prevention looking at serum levels
regarding dioxins and other health issues that will give insight /data on
body burdens. First- year data of 3- year study have been collected. 10.
There is some evidence that very low doses of dioxins may result in some
decreases in adverse responses but can produce other adverse effects at the
same or similar doses. Evaluate the totality of the evidence for non-
monotonic dose response as studies become available, particularly evidence
for any ?Ushaped?

dose response curve. EPA officials stated that the agency will continue to
work on the

dose response chapter. The possibility that dioxins are anticarcinogen is
reflected in the Integrated Summary, Section 2 with three Kayajanian
references. However, EPA does not have data on where or if it occurs on the
dose response curve- above or below body burdens.

11. Calculate ED using definitions other than that used for ED 01 and for
comparison purposes present values of ED 10 (since its been applied to other
chemicals by the agency).

EPA has done additional analysis using other effective dose values (e. g.,
ED 10 ). EPA officials stated that it differed from the original calculation
using ED 01 in only a few instances. See pp. 118- 122, in Integrated Summary
and Part II, Chapter 8. 12. Give additional consideration to its
justification of method selection for condensing these effective doses into
a recommended range.

See #11 above. 13. The agency?s description of its calculation of ED 01 was
not sufficiently detailed to permit the calculations to be repeated.
Describe calculation of ED 01 more clearly and completely.

More explanation provided for use of ED 01 . See #11 above.

Suggestion

1. The agency?s calculation of the cancer potency factor is not prominently
featured in the reassessment. Highlighting this calculation would
significantly improve the transparency and accessibility of the
reassessment.

EPA officials disagreed with the comment that cancer potency factor not
prominently featured. EPA officials stated that figure (5- 2) on cancer
potency estimates for animal studies with full page footnote provided in
version reviewed by the Science Advisory Board. Text has been added in Part
II, Section 5 discussing sensitivity of calculations.

Appendix IV: EPA?s Responses to Peer Review Panels

Page 56 GAO- 02- 515 Draft Reassessment of Dioxins

2. The panel suggested that the agency consider making greater and more
systematic use of parametric methods in calculations. This approach would
help readers to develop a better sense of how the results presented depend
upon specific analytical assumptions.

EPA does not plan to perform additional analysis. This is mostly intended as
a recommendation to do further research.

Concern

1. The majority of panel members have concerns about Agency cancer risk
estimates associated with current population exposures and feel that it was
not appropriate for the agency to characterize the risks in such a
quantitative manner without providing a similar quantitative estimate of
uncertainty.

EPA has added text on what EPA can say about quantifying uncertainty. EPA
officials agree with the Science Advisory Board that there needs to be
improvements in methodology (i. e., it will require more/ better data sets).
However, this will require further research in the future. It is a generic
concern, not just regarding dioxins.

Source: EPA.

Appendix V: Comments from the Environmental Protection Agency

Page 57 GAO- 02- 515 Draft Reassessment of Dioxins

Appendix V: Comments from the Environmental Protection Agency

See comment 1. Note: GAO comments

supplementing those in the report text appear at the end of this appendix.

Appendix V: Comments from the Environmental Protection Agency

Page 58 GAO- 02- 515 Draft Reassessment of Dioxins

Now on p. 10. See comment 4. See comment 3.

See comment 2.

Appendix V: Comments from the Environmental Protection Agency

Page 59 GAO- 02- 515 Draft Reassessment of Dioxins

Now on p. 15. See comment 6. Now on p. 14.

See comment 5.

Appendix V: Comments from the Environmental Protection Agency

Page 60 GAO- 02- 515 Draft Reassessment of Dioxins

Now on p. 15. See comment 9. See comment 8. Now on p. 37.

See comment 7.

Appendix V: Comments from the Environmental Protection Agency

Page 61 GAO- 02- 515 Draft Reassessment of Dioxins

See comment 10.

Appendix V: Comments from the Environmental Protection Agency

Page 62 GAO- 02- 515 Draft Reassessment of Dioxins

Now on pp. 20- 21. See comment 13. See comment 11.

Now on p. 6. See comment 12.

Appendix V: Comments from the Environmental Protection Agency

Page 63 GAO- 02- 515 Draft Reassessment of Dioxins

Appendix V: Comments from the Environmental Protection Agency

Page 64 GAO- 02- 515 Draft Reassessment of Dioxins

The following are GAO?s comments on EPA?s letter dated April 17, 2002. 1.
The discrepancies we identify between the Integrated Summary and supporting
chapters appear in the October 2001 reassessment documents that EPA
distributed for internal agency review. We identified them primarily to
inform readers of our report of the source of the information we cite. For
example, a reader of the Integrated Summary would find (outdated)
information on 9 food types, whereas we are citing information on 10 food
types that is provided in the supporting chapters of EPA?s reassessment
documents and that EPA officials told us is correct.

2. Throughout the section of our report on EPA?s estimate of dietary
exposure to dioxins, we attribute the identification of the limitations to
EPA?s draft reassessment report.

3. Our report did not characterize the significance of the limitations EPA
identified in its reassessment documents. We have added to the report EPA?s
opinion that these limitations do not represent major weaknesses in its
estimates of dietary exposure to dioxins.

4. The statement in our report that that the available studies generally
were not designed to estimate national exposures is derived from page 76 of
EPA?s October 2001 Integrated Summary draft. In this document EPA says: ?The
amount and representativeness of the data vary, but in general these data
were derived from studies that were not designed to estimate national
background means.? In its written comments, EPA says that most of the
dietary exposure it estimated was derived from studies specifically designed
to estimate national exposures. In support of this point, EPA says that 66
percent of the estimated exposure to dioxins is from eating beef, pork,
poultry, milk, and dairy products, and that these studies were designed to
estimate national exposures. (We note that these studies cover 5 of the 10
food types on which EPA based its exposure estimates.) Importantly, our
draft report stated that the studies on beef, pork, and poultry were based
on the first statistically designed national surveys of dioxin levels in
these foods and that the milk samples upon which both the milk and dairy
estimates were based came from a national survey with samples collected from
sampling stations in a majority of the states. However, while our review of
EPA?s milk survey design plan indicated the milk samples were intended to
assess the levels of dioxins in the general milk supply of the United
States, the survey design document also stated that (1) the milk would be
collected from dairy plants around the United States that represent
approximately 20 percent of the nation?s milk supply GAO Comments

Appendix V: Comments from the Environmental Protection Agency

Page 65 GAO- 02- 515 Draft Reassessment of Dioxins

and (2) the survey was not designed to be statistically rigorous- that is,
it was not intended to randomly sample milk in such a way that the results
could be generalized to the full milk supply with a known degree of
precision. Thus, we concluded that EPA?s statement in the Integrated
Summary- that the studies covering the 10 food types generally were not
designed to estimate national exposures- was accurate. In light of EPA?s
comments and the fact that the milk samples used to estimate milk and dairy
exposures did have national coverage, we have revised the report to indicate
that EPA acknowledges that some of the available studies were not designed
to estimate national average exposures.

5. We revised the description of the fat samples from ?inedible fat samples?
to cuts of fat, such as back fat on cattle, that generally are not consumed
by the U. S. public.

6. We understand that there is variability associated with measurements at
the picogram level. Nonetheless, we continue to believe that the variability
identified among the five samples studied indicates that using fat samples
not consumed by the public may overstate or understate to some extent dioxin
levels in beef, pork, and poultry products sold to the public.

7. In its comments, EPA stated that it believes that sufficient information
is available to support a conclusion that, in spite of the emission
reduction of the late 1990s, the exposure estimates of the draft
reassessment are a reasonable characterization of contemporary exposure. We
have revised the report to include EPA?s opinion and the reasons it cited in
support of its view that the emission reduction in the late 1990s does not
significantly affect the current estimate of general population exposure.
However, because EPA does not have data on dioxin emissions after 1995, we
cannot evaluate EPA?s conclusion.

8. EPA stated that it plans to delete information on the variability in
dairy concentration data from the reassessment report, and we have therefore
deleted this point from our report.

9. We understand that the contamination of the two samples eliminated from
EPA?s estimate was found to stem from a localized ball clay contamination.
However, we continue to believe that because of the lack of information on
the incidence of dioxin contamination in animal feeds as well as on the
potential sources of such contamination, it is not clear that the poultry
samples with high concentrations of dioxins were anomalies. For example,
this animal feed contamination problem was identified as a

Appendix V: Comments from the Environmental Protection Agency

Page 66 GAO- 02- 515 Draft Reassessment of Dioxins

result of the first national survey of only 80 poultry fat samples. We
acknowledge that a decision to exclude apparently anomalous information
entails professional judgment. However, because the incidence of
contamination of animal feed is unknown, we believe that it is important for
users of the dioxin reassessment to understand the judgments EPA made in
estimating dietary exposure.

10. In the draft report, EPA does not provide information on the assumptions
and analyses used to estimate the average fat percentages for pork and
poultry. However, EPA does provide some information on how it estimated the
fat percentage for beef. The fat percentage estimates affect the exposure
estimates, and we believe this information should be included in the
reassessment report. In its comments to us, EPA stated that the agency is
considering adding information about the pork and poultry estimates to the
report. We are therefore deleting references to this point in our report.

11. We deleted the phrase ?assembled by EPA? to be consistent with
information we provide in the body of the report that the peer review
panelists were selected by an independent contractor.

12. We have revised this statement to reflect the fact that most (rather
than all) of the other dioxins have TEFs of 0.1 or lower.

13. We clarified that TEFs apply to all effects, not just those for which
relative potency data were available.

Appendix VI: GAO Contacts and Staff Acknowledgments

Page 67 GAO- 02- 515 Draft Reassessment of Dioxins

David G. Wood, (202) 512- 3841 Christine Fishkin, (202) 512- 6895

Other key contributors to this report include Timothy Bober, Greg Carroll,
Nancy Crothers, Greg Wilmoth, and Carrie Wheeler. Appendix VI: GAO Contacts
and Staff

Acknowledgments GAO Contacts Staff Acknowledgments

(360096)

The General Accounting Office, the investigative arm of Congress, exists to
support Congress in meeting its constitutional responsibilities and to help
improve the performance and accountability of the federal government for the
American people. GAO examines the use of public funds; evaluates federal
programs and policies; and provides analyses, recommendations, and other
assistance to help Congress make informed oversight, policy, and funding
decisions. GAO?s commitment to good government is reflected in its core
values of accountability, integrity, and reliability.

The fastest and easiest way to obtain copies of GAO documents at no cost is
through the Internet. GAO?s Web site (www. gao. gov) contains abstracts and
fulltext files of current reports and testimony and an expanding archive of
older products. The Web site features a search engine to help you locate
documents using key words and phrases. You can print these documents in
their entirety, including charts and other graphics.

Each day, GAO issues a list of newly released reports, testimony, and
correspondence. GAO posts this list, known as ?Today?s Reports,? on its Web
site daily. The list contains links to the full- text document files. To
have GAO e- mail this list to you every afternoon, go to www. gao. gov and
select ?Subscribe to daily E- mail alert for newly released products? under
the GAO Reports heading.

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U. S. General Accounting Office 441 G Street NW, Room LM Washington, D. C.
20548

To order by Phone: Voice: (202) 512- 6000 TDD: (202) 512- 2537 Fax: (202)
512- 6061

Contact: Web site: www. gao. gov/ fraudnet/ fraudnet. htm E- mail: fraudnet@
gao. gov Automated answering system: (800) 424- 5454 or (202) 512- 7470

Jeff Nelligan, managing director, NelliganJ@ gao. gov (202) 512- 4800 U. S.
General Accounting Office, 441 G Street NW, Room 7149 Washington, D. C.
20548 GAO?s Mission

Obtaining Copies of GAO Reports and Testimony

Order by Mail or Phone To Report Fraud, Waste, and Abuse in Federal Programs

Public Affairs
*** End of document. ***