HUDSON RIVER PCBs REASSESSMENT ROTS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - HUMAN HEALTH RISK ASSESSMENT SCOPE OF WORK
APRIL 1999

&
PRO^°
For
U.S. Environmental Protection Agency
Region 2
and
U.S. Army Corps of Engineers
Kansas City District
Book 1 of 1
TAMS Consultants, Inc.
and
Gradient Corporation

-------
/	—.*c-v
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION 2
290 BROADWAY
NEW YORK, NY 10007-1866
W PRQlt0
April 27, 1999
To All Interested Parties:
The U.S. Environmental Protection Agency (USEPA) is pleased to release the Responsiveness
Summary for the Phase 2 Human Health Risk Assessment Scope of Work (HHRASOW) for the
Hudson River PCBs Reassessment Remedial Investigation/Feasibility Study (Reassessment
RI/FS).
This Responsiveness Summary contains USEPA's responses to the public comments received on
the July 1998 HHRASOW. The HHRASOW presented USEPA's general approach for
conducting the Human Health Risk Assessments for the Upper Hudson River and for the Mid-
Hudson River. The Upper Hudson River Human Health Risk Assessment will be completed in
Summer 1999. The Mid-Hudson River Human Health Risk Assessment will be completed
following USEPA's review of the revised Thomann-Farley model developed for the Hudson
River Foundation.
If you have any questions regarding this Responsiveness Summary or the Reassessment RI/FS in
general, please contact Ann Rychlenski. the Community Relations Coordinator for the site, at
(212) 637-3672.
Sincerely yours,
Richard L. Caspe, Director
Emergency and Remedial Response Division
Internet Address (URL) • http://www.epa.gov
Recycled/Recyclable • Printed with Vegetable Ofl Based Inks on Recycled Paper (Minimum 25% Postconsumer)

-------
HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - HUMAN HEALTH RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
srA/.
Oft)
PRCft
For
U.S. Environmental Protection Agency
Region 2
and
U.S. Army Corps of Engineers
Kansas City District
Book 1 of 1
TAMS Consultants, Inc.
and
Gradient Corporation

-------
HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - HUMAN HEALTH RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
TABLE OF CONTENTS
BOOK 1 OF 1	Page
ACRONYMS
I.	INTRODUCTION AND COMMENT DIRECTORY
1.	Introduction	1
1.1 Recent Developments 	2
2.	Commenting Process	2
2.1	Distribution of HHRASOW	2
2.2	Review Period and Public Availability Meetings 	2
2.3	Receipt of Comments	3
2.4	Distribution of the Responsiveness Summary	3
3.	Organization of HHRASOW Comments and Responsiveness Summary ... 7
3.1	Identification of Comments	7
3.2	Location of Responses to Comments	7
4.	Comment Directory	9
4.1	Guide to Comment Directory	9
4.2	Comment Directory	11
II.	RESPONSES TO COMMENTS ON THE UPPER HUDSON RIVER HUMAN
HEALTH RISK ASSESSMENT SCOPE OF WORK
Responses to General Comments	13
Responses to Specific Comments on the Upper Hudson River Human Health
Risk Assessment Scope of Work Also Applicable to the Mid-Hudson River
Risk Assessment Scope of Work		14
1.	Plan, Synopsis & Objectives	15
2.	Exposure Assessment	16
A.	Concentration of PCBs in Fish 	16
B.	Fish Consumption Rates for the Upper Hudson River 	17
i	TAMS/Gradient

-------
HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - HUMAN HEALTH RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
TABLE OF CONTENTS
BOOK 1 OF 1	Page
D.	Exposure Duration	21
E.	PCB Cooking Losses	23
3.	Toxicity Assessment	24
4.	Risk Characterization 	28
C.	Monte Carlo Analysis 	29
III.	RESPONSES TO SPECIFIC COMMENTS ON MID-HUDSON RIVER HUMAN
HEALTH RISK ASSESSMENT SCOPE OF WORK
1.	Plan, Synopsis & Objectives	30
2.	Exposure Assessment	30
A. Fish Concentration Data 	30
D.	Exposure Duration	31
F.	PCB Concentrations for Deterministic and
Monte Carlo Analyses	31
IV.	ADDITIONAL REFERENCES 	32
V.	COMMENTS ON HUMAN HEALTH RISK ASSESSMENT
SCOPE OF WORK	34
Federal (HF-1)
State (HS-1)
Local (HL-1)
Community Interaction Program (HP-1)
General Electric (HG-1)
ii
TAMS/Gradient

-------
Acronyms
ATSDR
Agency for Toxic Substances and Disease Registry
CDI
Chronic Daily Intake
CERCLA
Comprehensive Environmental Response. Compensation, and Liability Act
CIP
Community Interaction Program
CSF
Carcinogenic Slope Factor
FS
Feasibility Study
GE
General Electric Company
HI
Hazard Index
HHRA
Human Health Risk Assessment
HHRASOW
Human HeaLth Risk Assessment Scope of Work
HROC
Hudson River PCBs Oversight Committee
HQ
Hazard Quotient
IRIS
Integrated Risk Information System
LOAEL
Lowest Observed Adverse Effect Level
NCP
National Oil and Hazardous Substances Pollution Contingency Plan
NPL
National Priorities List
NO A A
National Oceanic and Atmospheric Administration
NYSDEC
New York State Department of Environmental Conservation
NYSDOH
New York State Department of Health
OSWER
Office of Solid Waste and Emergency Response
PCB
Polychlorinated Biphenyl
RfC
Inhalation Reference Concentrations
RfD
References Dose
RI
Remedial Investigation
RI/FS
Remedial Investigation/Feasibility Study
ROD
Record of Decision
RM
River Mile
RME
Reasonably Maximally Exposed
RI/FS
Remedial Investigation/Feasibility Study
SARA
Superfund Amendments and Reauthorization Act of 1986
SCEMC
Saratoga County Environmental Management Council
SOW
Scope of Work
STC
Science and Technical Committee
TAG.V1
Technical and Administrative Guidance Memorandum
TCDD
2,3,7,8-Tetrachlorodibenzo-p-dioxin
TEF
Toxicity Equivalency Factor
TSCA
Toxic Substances Control Act
UCL
Upper Confidence Limit
USDOI
United States Department of Interior
USEPA
United States Environmental Protection Agency
L'SFWS
US Fish and Wildlife Service
TAMS/Gradient

-------
Introduction

-------
HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY
PHASE 2 - HUMAN HEALTH RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
I. INTRODUCTION AND COMMENT DIRECTORY
1. Introduction
The U.S. Environmental Protection Agency (USEPA) has prepared this Responsiveness
Summary to address comments received during the public comment period on the Phase 2 Human
Health Risk Assessment Scope of Work (HHRASOW) for the Hudson River PCBs Reassessment
Remedial Investigation/Feasibility Study (Reassessment RI/FS), dated July 1998.
For the Hudson River PCBs Reassessment RI/FS, USEPA has established a Community
Interaction Program (CIP) to elicit on-going feedback through regular meetings and discussion and
to facilitate review of and comment upon work plans and reports prepared during all phases of the
Reassessment RI/FS.
Because of the large number of CIP participants and associated costs of reproduction, the
HHRASOW is incorporated by reference and is not reproduced herein. No revised HHRASOW will
be published. The comment responses and revisions noted herein are considered to amend the
HHRASOW. For complete coverage, the HHRASOW and this Responsiveness Summary must be
used together.
The first part of this five-part Responsiveness Summary is entitled, "Introduction and
Comment Directory." It describes the HHRASOW review and commenting process, explains the
organization and format of comments and responses, and contains a comment directory.
The second part, entitled. "Responses to Comments on the Upper Hudson and Mid-Hudson
River Human Health Risk Assessment Scope of Work," contains USEPA's responses to all
significant comments that are relevant to both risk assessments. Responses are grouped according
to the section number of the HHRASOW to which they refer. For example, responses to comments
on Section II.2.A. of the HHRASOW are found in Section II.2.A. of the Responsiveness Summary.
Additional information about how to locate responses to comments is contained in the Comment
Directory.
The third part, entitled, "Responses to Specific Comments on the Mid-Hudson
River Human Health Risk Assessment Scope of Work," contains USEPA's responses to all
significant comments on the Mid-Hudson that are not addressed in the second part. Responses are
grouped according to the section number of the HHRASOW to which they refer. For example,
responses to comments on Section III.2.A. of the HHRASOW are found in Section III.2.A. of the
1
TAMS/Gradicnt

-------
Responsiveness Summary. Additional information about how to locate responses to comments is
contained in the Comment Directory.
The fourth part, entitled, "Additional References," contains a listing of references cited in
this Responsiveness Summary that are not listed in the HHRASOW.
The fifth part, entitled, "Comments on the HHRASOW," contains copies of the comments
submitted to USEPA. The comments are identified by commenter and comment number, as further
explained in the Comment Directory.
1.1 Recent Developments
USEPA received the revised Thomann-Farley model on April 27,1999. USEPA will review
the model to determine its appropriateness for use in performing the Mid-Hudson HHRA. In the
HHRASOW (p. 1), USEPA noted that the Upper Hudson and Mid-Hudson HHRAs may be
developed at different times.
2. Commenting Process
This section documents and explains the commenting process and the organization of
comments and responses in this document. Readers interested in finding responses to their
comments may skip this section and go directly to the tab labeled "Comment Directory."
2.1	Distribution of HHRASOW
The HHRASOW, issued in July 1998, was distributed to federal and state agencies and
officials, participants in the CIP and General Electric Company (GE), as shown in Table 1.
Distribution was made to approximately 100 agencies, groups, and individuals. Copies of the
HHRASOW were also made available for public review in 17 Information Repositories, as shown
in Table 2 and on the USEPA Region 2 internet webpage, entitled "Hudson River PCBs Superfund
Site Reassessment," at www.epa.gov/hudson.
2.2	Review Period and Public Availability Meetings
Review of and comment on the HHRASOW occurred from July 23. 1998 to August 31,
1998. During this period, USEPA held a Joint Liaison Group meeting at the Marriott Hotel in
Albany, New York on July 23, 1998. Subsequently, USEPA sponsored an availability session to
answer questions on August 19, 1998 at the Holiday Inn at Latham, New York and on August 20,
1998 at Marist College in Poughkeepsie, New York. These meetings were conducted in accordance
with L'SEPA's "Community Relations in Superfund: Handbook, Interim Version" (1988). Minutes
of the Joint Liaison Group meeting are av ailable for public review at the Information Repositories
listed in Table 2.
?
TAMS/Gradient

-------
As stated in USEPA's letter transmitting the HHRASOW, all citizens were urged to
participate in the Reassessment process and to join one of the Liaison Groups formed as part of the
C1P.
2.3	Receipt of Comments
Comments on the HHRASOW were received in two ways: letters to USEPA and oral
statements made at the Joint Liaison Group meetings on July 23, 1998. USEPA's responses to
comments made at the Joint Liaison Group meeting are provided in the meeting minutes.
All significant written comments received on the HHSOW are addressed in this
Responsiveness Summary. Comments were received from five commenters. Total comments
numbered over 100.
2.4	Distribution of Responsiveness Summary
This Responsiveness Summary will be distributed to the Liaison Group Chairs and Co-
Chairs and interested public officials. This Responsiveness Summary will also be placed in the
seventeen Information Repositories and is part of the Administrative Record.
3
TAMS/Gradienl

-------
TABLE 1
DISTRIBUTION OF REPORTS
HUDSON RIVER PCBs OVERSIGHT COMMITTEE MEMBERS
USEPA ERRD Deputy Division Director (Chair)
USEPA Project Managers
USEPA Community Relations Coordirtator, Chair of the Steering Committee
NYSDEC Division of Hazardous Waste Management representative
NYSDEC Division of Construction Management representative
National Oceanic and Atmospheric Administration (NOAA) representative
Agency for Toxic Substances and Disease Registry (ATSDR) representative
US Army Corps of Engineers representative
New York State Thruwav Authority (Department of Canals) representative
USDOI (US Fish and Wildlife Service) representative
NYSDOH representative
GE representative
Liaison Group Chairpeople
Scientific and Technical Committee representative
SCIENTIFIC AND TECHNICAL COMMITTEE MEMBERS
The members of the Science and Technical Committee (STC) are scientists and technical researchers
who provide technical input by evaluating the scientific data collected on the Reassessment RI/FS,
identifying additional sources of information and on-going research relevant to the Reassessment RI/FS,
and commenting on USEPA documents. Members of the STC are familiar with the site, PCBs, modeling,
toxicology, and other relevant disciplines.
Dr. Daniel Abramowicz
Dr. Donald Aulenbach
Dr. James Bonner, Texas A&M University
Dr. Richard Bopp, Rensselaer Polytechnic Institute
Dr. Brian Bush, New York State Dept. of Health
Dr. Lenore Clesceri, Rensselaer Polytechnic Institute
Mr. Kenneth Darmer
Mr. John Davis, New York State Dept. of Law
Dr. Robert Dexter, EVS Consultants. Inc.
Dr. Kevin Farley, Manhattan College
Mr. Jay Field, National Oceanic and Atmospheric Administration
Dr. Ken PearsalL U.S. Geological Survey
Dr. John Herbich, Texas A&M University
Dr. Behrus Jahan-Parwar. SUNY - Albany
Dr. Nancy Kim, New York State Dept. of Health
Dr. William Nicholson. Mt. Sinai Medical Center
Dr. George Putman, SUNY - Albany
Dr. G-Yull Rhee. New York State Dept. of Health
Dr. Francis Reilly, Jr., The Reilly Group
4
TAMS.'Gradient

-------
Dr. John Sanders
Ms. Anne Secord, U.S. Fish and Wildlife Service
Dr. Ronald Sloan, New York State Dept. of Environmental Conservation
USEPA Community Relations Coordinator (Chair)
Governmental Liaison Group Chair and two Co-chairs
Citizen Liaison Group Chair and two Co-chairs
Agricultural Liaison Group Chair and two Co-chairs
Environmental Liaison Group Chair and two Co-chairs
USEPA Project Managers
NYSDEC Technical representative
NYSDEC Community Affairs representative
STEERING COMMITTEE MEMBERS
FEDERAL AND STATE REPRESENTATIVES
Copies of the HHRASOW were sent to relevant federal and state representatives who have been
involved with this project. These include, in part, the following:
The Hon. Daniel P. Moynihan
The Hon. Charles E. Schumer
The Hon. John E. Sweeny
The Hon. Michael McNulty
The Hon. Sue Kelly
The Hon. Nita Lowey
The Hon. Maurice Hinchey
The Hon. Ronald B. Stafford
The Hon. Benjamin Gilman
The Hon. Richard Brodskv
The Hon. Bobby D'Andrea
17 INFORMATION REPOSITORIES (see Table 2).
5
TAMS/Gradient

-------
TABLE 2
INFORMATION REPOSITORIES
Adriance Memorial Library
93 Market Street
Poughkeepsie, NY 12601
Catskill Public Library
1 Franklin Street
Catskill, NY 12414
A Cornell Cooperative Extension
Sea Grant Office
74 John Street
Kingston. NY 12401
Crandall Library
City Park
Glens Falls, NY 12801
County Clerk's Office
Washington County Office Building
Upper Broadway
Fort Edward, NY 12828
*	A Marist College Library
Marist College
290 North Road
Poughkeepsie, NY 12601
*	New York State Library
CEC Empire State Plaza
Albany, NY 12230
New York State Department
of Environmental Conservation
Division of Hazardous Waste Remediation
50 Wolf Road, Room 212
Albany, NY 12233
*	A R. G. Folsom Library
Rensselaer Polytechnic Institute
Troy, NY 12180-3590
Saratoga County EMC
50 West High Street
Ballston Spa, NY 12020
*	Saratoga Springs Public Library
49 Henry Street
Saratoga Springs, NY 12866
*	A SUNY at Albany Library
1400 Washington Avenue
Albany, NY 12222
*	A Sojourner Truth Library
SUNY at New Paltz
New Paltz, NY 12561
Troy Public Library
100 Second Street
Troy, NY 12180
United States Environmental Protection
Agency
290 Broadway
New York, NY 10007
*	A United States Military Academy Library
Building 757
West Point, NY 10996
White Plains Public Library
100 Martine Avenue
White Plains, NY 12601
Repositories with Database Report
CD-ROM (as of 10/98)
Repositories without Project
Documents Binder (as of 10/98)
6
TAMS/Gradient

-------
3. Organization of HHRASOW Comments and Responsiveness Summary
3.1	Identification of Comments
Each submission commenting on the HHRASOW was assigned "H" for HHRASOW and one of the
following letter codes:
F - Federal agencies and officials;
S - State agencies and officials;
L - Local agencies and officials;
P - Public Interest Groups and Individuals; and
G - GE.
The letter codes were assigned for the convenience of readers and to assist in the organization of this
document. Priority or special treatment was neither intended nor given in the responses to
comments.
Once a letter code was assigned, each submission was then assigned a number, in the order that it
was received and processed, such as F1. Each different comment within a submission was assigned
its separate subnumber. Thus, if a federal agency submission contained three different comments,
the comments would be designated HF1-1, HF1-2, and HF1-3.
Comment letters have been reprinted following the fourth tab of this document. The exception to
this is the more than 300 pages of appendices to the comment letter submitted by GE. These
appendices are materials that do not contain specific comments on the HHRASOW and have not
been reprinted in this Responsiveness Summary because of the volume of paper. GE's comment
letter containing 95 pages of comments specific to the HHRASOW is reprinted herein.
The alphanumeric code associated with each reprinted comment letter is marked at the top right
corner of the first page of the comment letter and the subnumbers designating individual comments
are marked in the margin next to the comment. Comment submissions are reprinted by letter code
in the following order: HF, HS, HL, HP and HG.
3.2	Location of Responses to Comments
The Comment Directory, following this text, contains a complete listing of all commenters and
comments. This directory allows readers to find responses to comments and provides several items
of information.
The first column lists the names of commenters. Comments are grouped first by: HF
(Federal), HS (State), HL (Local), HP (Public Interest Group or Individual) or HG (GE).
The second column identifies the alphanumeric comment code (for example, HF1-1)
assigned to each comment.
The third column identifies the location of the response by the HHRASOW Section number.
For example, comments raised in Section II.2.A of the HHRASOW can be found in the
7
TAMS/Gradient

-------
corresponding Section 11.2. A of the Responsiveness Summary, following the third tab of this
document.
The fourth, fifth, and sixth columns list key words that describe the subject matter of each
comment. Readers will find these key words helpful as a means to identify subjects of
interest and related comments.
8
TAMS/Grailicfi!

-------
4. COMMENT DIRECTORY
4.1 Guide to Comment Directory
This section contains a diagram illustrating how to find responses to comments. The
Comment Directory follows. As stated in the Introduction, this document does not reproduce the
HHRASOW. Readers are urged to utilize this Responsiveness Summary in conjunction with the
HHRASOW.
Step 1 | Step 2
Step 3
Find the commenter or the key
words of interest in the
Comment Directory.
Obtain the alphanumeric
comment codes and the
corresponding HHRASOW
section.
Find the responses following the
Responses tab. Use the Table of
Contents to locate the page of the
Responsiveness Summary for the
HHRASOW section.
Key to Comment Codes:
Comment codes are in the format HXl-a
H=HHRASOW
Y=Commenter group (F=Federal, S=State, L=Local, P=Public Interest Group or Individual, G=GE)
a=Submittal number within the commenter group
b=Numbered comment
Example:
Comment Response Assignment for the HHRASOW
AGENCY/ !
Comment
REPORT
KEY WORDS
1 Name j
CODE
SECTION
1
2
3
NOAA /Rosman	HF1-1	1.1	Fish Tissue
9
TAMS'Gradient

-------
THIS PAGE LEFT BLANK INTENTIONALLY
10	TAMS/Gradient

-------
Comment Directory

-------
4.2 COMMENT DIRECTORY
AGENCY/
NAME
COMMENT
CODE
REPORT
SECTION
Key W ords



1
2
3
NOAA/Rosman
HE1-1-
II General
Fish Tissue
PCB Data
Data Sources
NOAA/Rosman
HF1-2
II (ieneral
Human Health
Upper Hudson
Lower Hudson
NOAA/Rosman
HF1-3
II 1
Ingestion
Food
PCBs
NOAA/Rosman
HE 1-4
II.2.A
Fishing
Fish Advisories
PCBs
NOAA/Rosman
HF1-5
IE2.D
Exposure Duration
Fishing
Mid-Hudson River
NOAA/Rosman
HFI-6
II.2.E
Exposure
Cooking
PCBs
NOAA/Rosman
HE 1-7
II.2.E
Exposure Assessment
Cooking
Head/rail
NOAA/Rosman
HE 1-8
111
Exposure Assessment
Meat /Vegetables
Floodplains
NOAA/Rosman
IIF 1-9
114
Exposure Duration
Latent Risks
Exposure Levels
NOAA/Rosman
HF1-10
III.2D
Fishing I.imitations
-
-
NOAA/Rosman
HFI-11
III 2.F
Trend
Fish Tissue
PCBs Concentrations
NYSDEC/Ports
HS1-1
11.2 A
Fishing Ban
Fish Ingestion
Health Advisory
NYSDEC/Ports
HS1-2
II.2.D
Exposure Duration
Fish Ingestion
Lifetime Exposure
NYSDEC/Ports
HS1-3
11.3
Reference Dose
Cancer Risk
Non-Cancer Risk
NYSDEC/Ports
HSI-4
113
Reference Dose
Inhalation
Ingestion
SCEMC/Balet
HI.1-1
II 3
Cancer Slope Factors
Congener Specific
-
SCEMC/Balet
HI. 1-2
11.2.B
Fishing
IJpper Hudson River
Angler Survey
SCLMC/Balct
HI.1-3
11.2.E
Cooking Fish
Inhalation
Assumptions
SCLMC/Balct
HI.1-4
III 1
Information Repositories
-
-
Scenic Hudson
HPI-1
II.2 B
Subsistence Fishing
Creel Count
Fish Consumption
Scenic Hudson
HP 1-2
II 2 B
Fish Consumption
Fisherman Families
Exposure
Scenic Hudson
HP 1-3
II 2 8
Exposed Populations
Survev of Families
Exposure Assessment
Scenic Hudson
HP 1-4
II.3
Toxicity Values
Reference Dose
PCB Mixtures
Scenic Hudson
IIP 1 -5
II 1
Fish Consumption
Ingestion Pathway
Multiple Exposures
Scenic Hudson
HP 1-6
11.2 A
Fish Consumption
Fishing Bans
Recreational Fishing
Scenic Hudson
HPI-7
II 2 D
Angler Surveys
Biased Sampling
Deterministic Methods
Scenic Hudson
HP1-8
II 2 D
Exposed Populations
Exposure Duration
-
Scenic Hudson
HP 1-9
II 4
Historical Exposure
Latent PCB Effect
Fish Consumption
gi:
HG1-1
II General
Monte Carlo Modeling
Fish Consumption
Risk Analysis
(it
HG1-2
II General
Qualitative
Vague
Proposed Approach
GE
HG1-3
II 2 B
PCB Cancer Risks
Conservative Approach
Extrapolations
GE
HG1-4
II.3
Risk Estimates
Decision Making
-
GF.
HG1-5
II General
Federal Dam
Mid-Hudson River
Site Size
GE
HG1-6
II.2.A, II 4
Fish Consumption
Catch and Release
Exposure Limitation
gi:
HG1-7
II 2 B
Exposure Assessment
Point Estimates
-
GF
HG1-8
II.4.C
Monte Carlo Modeling
Uncertainty
Variability
GE
HGI-9
II 2 B
Fish Consumption
Exposure Factors
Fish Advisories
GE
HG1-10
II 2 B
Subsistence Anglers
Not Relevant
Hudson River Anglers
GE
HG1-11
II 2 B
Monte Carlo Modeling
Micro-Exposure
Input Parameters
GE
HG1-12
II General,II 2 A
Bioaccumulation
Time Variable
PCB Concentrations
GE
HG1-I3
II 2 F
Cooking
PCB Losses
-
GE
HG1-I4
11.3
Aroclor Based Toxicity
Toxicity Assessment
-
GE
1IG1-15
11.3
Cancer Slope Factors
Extrapolation
-
GE
HG1-16
113
Reference Dose
IRIS
Conservative
GE
HG1-17
II.3
Effects Assessment
TEQ
Aroclor Toxicity
GE
HG1-18
11.3
Endocrine Disruption
-
-
GE
HG1-19
II.3
Background
Risk Management
-
GE
I1G1-20
11.4
RMF.
Probabilistic Analysis
Deterministic Analysis
GE
HG1-21
11.4
HHRA
No Action
Remedial Decision
GE
HG1-22
II General
Monte Carlo Modeling
Fish Consumption
Agreement
GF.
HG1-23
11 2 A
Qualitative
Vague
Proposed Approach
GE
IICil-24
11 3
Toxicity Studies
Uncertainty
-
GE
HGI-25
II (ieneral
Troy Dam
the Site
Lower Hudson River
GE
HG1-26
II (ieneral
Data Sources
Additional Data
I Usability
11

-------
AGENCY/
NAME
COMMENT
CODE
REPORT
SECTION
Key Words
GE
HG1-27
III 1
Thomann-Farley Model
Peer Review-
-
GE
HG1-28
11.4
Hypolhetical
Administrative Control
Approach Outline
GE
HG1-29
112
Angler Exposure
Fish Consumption
Exposure Limitation
GE
HG1-30
II 2 B
Exposure Assessment
Worst Case
Default Values
GE
HG1-31
II General
SOW
-
-
GE
HG1-32
II 2 A
Fish Concentrations
Fish Consumption
Predatory Fish
GE
HG1-33
II.2.B
Conceptual Model
Model Inputs
Model Structure
GE
HG1-34
II.4.C
Monte Carlo Modeling
Micro-Exposure
Time Vary ing Models
GE
HG1-35
II 3
Uncertainty Analysis
Uncertainty Sources
-
GE
UG1-36
II.4.C
Monte Carlo Modeling
Input Values
-
GE
HG1-37
11.4.C
Variability
Ingestion Rates
Exposure Limitation
GE
HG1-38
11.4.C
Angler Exposure Model
Uncertainty
Micro-Exposure
GF.
HG1-39
III.2.A
Risk Management
Background Exposure
Dosages
GE
HG1-40
11.3
Toxicokinetic Models
Human Body Burdens
Background Exposure
GE
HG1-41
11.2.B
Fish Ingestion Rates
PCB Exposure
Site Specific
GE
HG1-42
II.2.B
Fishing Restriction
Exposure Limitations
Exposure Dosages
GF.
HG1-43
II.2.B
Angler Population
Consumption Rates
-
GE
UG1-44
11.2.B
Angler Subpopulation
Fish Consumption
Ingestion Rates
GE
HG1-45
II.2.B
Species Specific
Ingestion Rates
-
GE
HG1-46
11.2.B
Surrogate Fish Species
-
-
GE
FIG 1-47
II General
Bioaccumulation Models
Average Concentration
RMF.
GE
HGI-48
II General
Food Chain Model
BSM
Fish Tissue Levels
GE
HG1-49
III.2.A
High End Exposure
Migratory Fish
Mid-Hudson River
GE
HG1-50
II.2.D
Exposure Duration
Receptor Population
-
GE
HG1-51
II 2 E
PCB Losses
Cooking
Fish Advisories
GE
UG1-52
II 2.E
Inhalation
Air Data
PCB Concentrations
GE
HG1-53
II.3
Aroclor Based Toxicity
Toxicity Assessment
1'1-Q
GE
FIG 1-54
II.3
Cancer Slope Factors
Dose Response
IRIS
GF.
HG1-55
11.3
Epidemiological Studies
Human Exposure
PCB Toxicity
GE
HG1-56
II 3
Epidemiological Studies
PCB Induced Cancer
PCB Exposure
GE
HGI-57
II 3
Epidemiological Studies
PCB Induced Cancer
PCB Exposure
GF
HG 1 -58
II 3
Epidemiological Studies
PCB Induced Cancer
PCB Exposure
GE
FIG 1-59
II 3
Non-Negative Effects
PCB Induced Cancer
-
GE
HGI-60
II.3
Adverse Effects
Weight of Evidence
Conservative
GE
UG1-61
II.3
Cancer Slope Factors
Epidemiological Studic
Extrapolations
GE
HG1-62
II.3
Cancer Slope Factors
Cancer Toxicity
Epidemiological Studies
GE
FIG 1-63
II 3
Non-Cancer Toxicity
Reference Dose
Uncertainty Factors
GE
HGI-64
II 3
Epidemiological Studies
Reference Dose
-
GE
HG1-65
II 3
Aroclor Based Toxicity
Aroclor 1254
Reference Dose
GE
HG 1 -66
II 3
Epidemiological Studies
Rhesus Monkey
Clinical Relevance
GE
HG1-67
II 3
Epidemiological Studies
Rhesus Monkey
PCB Metabolism
GE
HG1-68
11.3
Epidemiological Studies
Rhesus Monkev
Dioxins/Furans
GE
HG1-69
II.3
Toxicity Studies
Toxicity Assessment
Toxicological Endpoints
GF
HGI-70
II.3
Study Duration
Uncertainty Factors
Sub-Chronic Exposure
GF
HGI-71
II.3
Endocrine Disruption
Endocrine Effects
Qualitative
GE
HGI-72
11.3
Uncertainty
Toxicity Assessment
Toxicological Criteria
GE
HG 1-73
II 3
Averaging Time
Non-Carcinogenic
Exposure Duration
GE
HG1-74
II.3
Background
PCB Concentrations
Exposure Assessment
GE
HG1-75
II.3
Fish Advisories
Exposure Duration
Fish Consumption
GE
HG1-76
II 4
Central Tendency
High End Exposure
Risk Analysis
GE
HG1-77
11.4
Deterministic Exposure
Probabilistic Exposure
Risk Analysis
GF.
HG1-78
114
Risk Reduction
Risk Management
No Action
GE
UCJ1-79
114
Angler Risk
Hot Spot
Risk Interpretation
GE
HGI-80
II 4
Start Date
Exposure Duration
Risk Analysis
12

-------
Responses

-------
II. RESPONSES TO COMMENTS ON THE UPPER HUDSON RIVER HUMAN
HEALTH RISK ASSESSMENT SCOPE OF WORK
Responses to General Comments
Response to HG1-1 and HG1-22
USEPA acknowledges the comments supporting many aspects of the HHRASOW, including
the Agency's decision to use Monte Carlo modeling and its determination that the greatest risk of
exposure to PCBs in the Hudson River is likely to result from fish consumption. However, USEPA
disagrees with the comment that other possible exposure pathways are not significant or have little
bearing on potential risk; such statements are premature until the results of the human health risk
assessment are known.
Response to HG1 -2 and HG1-31
The level of detail in the HHRASOW is appropriate for a scope of work document. The
HHRASOW provides an outline of the procedures that will be used by USEPA in developing the
baseline Phase 2 Human Health Risk Assessment. USEPA acknowledges that there are decision
points in the process of conducting the risk assessment. The decisions made will be supported in
the HHRA.
Response to HG1-26
The only data sets not previously described in the HHRASOW contain air data. The air data
will update information presented in the Phase 1 risk assessment and address stakeholders' concerns
regarding potential inhalation of PCBs. As part of the evaluation of scientific data for this pathway,
the following studies are being reviewed as background for understanding volatilization of PCBs:
1.	Buckley, E.H. and T.J. Tofflemire, 1983. Uptake of Airborne PCBs by Terrestrial Plants
Near the Tailwater of a Dam. Proc. Natl. Conf. on Environ. Eng., ASCE Specialty
Conference, July 6-8, pp. 662-669.
2.	Nelson, E.D., L.L. McConnell, and J.E. Baker, 1998. Diffusive Exchange of Gaseous
Polycyclic Aromatic Hydrocarbons and Polychlorinated Biphenyls Across the Air-Water
Interface of the Chesapeake Bay. Environ. Sci. Technol. 32:912-919.
3.	Hornbuckle, K.C., J.D. Jeremiason, C.W. Sweet, and S.J. Eisenreich, 1994. Seasonal
Variations in Air-Water Exchange of Polychlorinated Biphenyls in Lake Superior. Environ.
Sci. Technol. 28:1491-1501.
13
TAMS'Gradient

-------
4.	Achman, D.R., K.C. Hombuckle, and S.J. Eisenreich, 1993. Volatilization of
Polychlorinated Biphenyls from Green Bay, Lake Michigan. Environ. Sci. Technol.
27(1 ):75-84.
5.	Hornbuckle. K.C., D.R. Achman, and S.J. Eisenreich, 1993. Over-water and Over-land
Polychlorinated Biphenyls in Green Bay, Lake Michigan. Environ. Sci. Technol.
27(l):87-98.
Decisions regarding the use of the air data and their use in the HHRA will be supported and
described in the appropriate sections of the HHRA.
Responses to Specific Comments on the Upper Hudson River Human Health Risk
Assessment Scope of Work Also Applicable to the Mid-Hudson River Human Health
Risk Assessment Scope of Work
Response to HF1 -1. HG1 -12. HG1 -47. and HG1 -48
Details of the modeling effort are presented in the Phase 2 Report - Further Site
Characterization and Analysis Volume 2B - Preliminary Model Calibration Report, Hudson River
PCBs Reassessment RI/FS, and will be presented in the Baseline Modeling Report (due May 1999).
As described in Section 3.4 of the Phase 2 Report, all modeling work will utilize the extensive
database that was created to support the Hudson River PCBs Reassessment RI/FS (Further Site
Characterization and Analysis Database Report, 1995). The database contains measurements for
sediments, fish and aquatic biota, surface water flow and surface water quality from the USEPA, the
New York State Department of Environmental Conservation (NYSDEC) and General Electric. The
database includes a total of approximately 750,000 records. Almost 350,000 of these records contain
data acquired as part of the USEPA Phase 2 Work Plan and Sampling Plan. The remaining records
contain data from a large number of historical and ongoing monitoring efforts in the Hudson River.
Response to HF1-2
USEPA will conduct HHRAs for the Upper and Mid-Hudson River. USEPA will not
conduct a HHRA for the portion of the Hudson River site between Poughkeepsie and the Battery in
New York City. USEPA's approach for conducting the baseline HHRA is protective of human
health because the risk for individuals who consume fish caught closer to the PCB-contaminated
sediments in the Upper Hudson River is expected to be higher than the risk to individuals who
consume fish caught farther away.
Response to HG1-5 and HG1-25
USEPA has consistently defined the site to include the Lower Hudson River since at least
April 1984, when the Agency issued its FS for the site and before the site was listed on the National
14
TAMS/Gradient

-------
Priorities List (codified at 40 CFR Part 300, App. B). In its September 25, 1984 Record of Decision,
USEPA defines the site by reference to three figures which, together, depict the site as the entire
200-mile stretch of the River from Hudson Falls to the Battery in New York City, plus the remnant
deposits. In addition, during the Reassessment RI/FS, USEPA has consistently defined the site as
including the Upper and Lower River (e.g.. the Scope of Work for Hudson River Reassessment
Ri/FS (December 1990) and the Phase 1 Report for the Reassessment RJ/FS (August 1991)). The
comment regarding USEPA's use of the results of the HHRA (including the Mid-Hudson HHRA)
in evaluating remedial alternatives is a risk management issue, and therefore beyond the scope of
the HHRA. USEPA decision-makers will consider risk management following completion of the
HHRA.
1. Plan, Synopsis & Objectives
Response to HP1-5
USEPA acknowledges this comment supporting evaluation of multiple pathways.
Response to HF1-3 and HF1-8
Routes of exposure not quantified in the Phase 1 HHRA will be reviewed in the Phase 2
HHRA to determine if adequate current data exist to support a quantitative assessment (HHRASOW,
p. 13). One such route of exposure, noted in the comment, is potential ingestion of vegetables, meat,
eggs, and cow's milk from farms located on the flood plains of the Upper Hudson River. USEPA
has determined not to evaluate this route of exposure based on existing data on milk samples and
forage crops ingested by cattle, as described below.
For the past 25 years, the New York State Department of Agriculture and Markets has
analyzed more than 18,200 samples of cow's milk within the state and has not found any detection
of PCBs above the detection limit of 0.6 parts per million (ppm) (lipid normalized) (Rudnick, 1999,
personal communication). This detection limit is significantly less than the U.S. FDA limit of 1.5
ppm (lipid normalized). Moreover, in the 1980s, Dr. Buckley from the Boyce Thompson Institute
at Cornell University collected data on PCBs in forage crops (corn and hay) grown in an area with
PCB-contaminated soil along the Upper Hudson River (Buckley and Tofflemire, 1983). The levels
of PCBs on these crops were below the U.S. Department of Agriculture regulatory level of 0.2 mg/kg
for forage crops. A back-calculation of the level of PCBs in air required to fall on forage crops to
attain the U.S. Department of Agriculture level would be in the range of approximately 40 to 60
nanograms/cubic meter. Based a review of air sampling data collected by GE in 1991 on the Hudson
River, air concentrations at these levels are typically much lower. Based on this information,
USEPA has determined that collection of additional PCB data from vegetables, meat, eggs and milk
is not warranted and that the risk via air deposition and ingestion from foods other than fish are
likely to be minimal. As stated in the HHRASOW (pp. 14 and 21). the HHRA will present total
risks and hazards added together across pathways.
15
TAMS/Gradient

-------
2. Exposure Assessment
Response to HG1-29
USEPA acknowledges the commenter's agreement with the use of site-specific data and
appropriate assumptions to develop estimates of exposure, as well as the specific aspects of
USEPA's approach listed by the commenter.
A. Concentration of PCBs in Fish
Response to HG1-23
As described in the HHRASOW (p. 7), the HHRA will use estimated future PCB
concentrations in fish based on the analytical data and results presented in the Database Report
(USEPA, 1995a), the Preliminary Model Calibration Report (USEPA, 1996a), the Data Evaluation
and Interpretation Report (USEPA, 1997d), and the Baseline Modeling Report (due May 1999).
Specifically, the HHRA will consider annual fish concentrations for Total PCBs as a function of
river location and fish species. Concentrations will be adjusted to reflect standard fillet portions.
Any distributional information provided by the modelers (reflecting uncertainty and/or variability)
will be incorporated into the Monte Carlo analysis.
Response to HF1-4. HP 1-6. and HG1-6
As described in the HHRASOW (p. 1) and consistent with USEPA guidance (USEPA. 1989),
the HHRA will present an analysis of the potential adverse health effects (current or future) caused
by hazardous substance releases from the site in the absence of any remedial actions or institutional
controls. The existing health advisories on fish consumption are institutional controls that control
or mitigate the exposure to PCB-contaminated fish from the river. As such, the HHRA will discuss
the existence of the health advisories, but will assess the risk posed by the site in the absence of the
health advisories.
Response to HS1-1
The clarification is acknowledged and will be incorporated into future reports, as applicable.
Response to HG1-12
USEPA will use fate and transport models to describe the distribution of PCBs in the Hudson
River. These modeled concentrations of PCBs in sediment and water will be used as initial
concentrations in several bioaccumulation models, including a Gobas-type time-variable mechanistic
model. Using the fish concentration output from the Gobas model, USEPA will model human
ingestion of PCBs in fish. The HHRA will be peer reviewed (see also, response to HG1-39 and
HG1-49).
16
TAMS. Gradient

-------
Response to HG1-32
As discussed in the HHRASOW, exposures and risks for the high-end angler will be
calculated using the 95,h percentile of the fish ingestion rate distribution (HHRASOW, p. 7), the high
end of the distribution for exposure duration (HHRASOW, p. 11), and mean body weights
consistent with appropriate toxicity assessment assumptions. USHPA is still assessing the most
representative value for cooking losses (HHRASOW, pp. 11 and 12). Consistent with USEPA
guidance (USEPA, 1989 and 1992c), for the high end exposure point concentration of PCBs in fish,
USEPA will use the mean PCB concentration, weighted by fish species preferences and averaged
over time and location. The possibility that individual anglers may preferentially fish in the most
contaminated stretches of the river, or select only the most contaminated species, will be considered
in the uncertainty analysis.
B. Fish Consumption Rates for the Upper Hudson River
Response to HL1-2
Anglers in the Upper Hudson River are required to have fishing licenses and are subject to
the "catch and release" fishing advisory. Consistent with USEPA guidance (USEPA. 1989), the
expected fish consumption rates for the Upper Hudson River will be evaluated in the absence of
institutional controls. Specifically, fish consumption rates from similar rivers that also require
fishing licenses but are not subject to fishing advisories or bans will be evaluated. The HHRASOW
(pp. 7 and 8) identified the Connelly surveys (Connelly el aL, 1990, 1992, and 1996) as a source of
comprehensive information about thousands of anglers within New York State, including anglers
from similar rivers with similar types of fish. A survey of anglers within 50 miles of the Upper
Hudson River, as suggested by the commenter, would include anglers from water bodies that are not
similar to the Upper Hudson River and who may consume different fish species. Therefore, the
proposed approach is not appropriate for use in the HHRA.
Response to HP 1-1. HP 1-7. HG1-10
The distribution of fish consumption rates for the Mid-Hudson and Upper Hudson River will
include data from the Connelly surveys (Connelly gt aL, 1990, 1992, and 1996), the Clearwater
surveys (Barclay, 1993) and, to a lesser extent, ChemRisk (1991). The distribution of fish ingestion
rates will represent the full range of the Hudson River angler population, from those who may eat
as little as one fish meal from the Hudson River per year to very frequent anglers, for whom Hudson
fish would account for a significant portion of their protein intake, regardless of whether these
anglers fish for recreation or sustenance (HHRASOW, p. 9). Uncertainties associated with
underreporting of subsistence anglers will be addressed in the uncertainty section of the HHRA.
Consistent with USEPA guidance, both "high end" and "central tendency" risks and hazards
(point estimates) will be evaluated in the HHRA. In addition, the Monte Carlo analysis will
consider the full distribution of risk and hazards for Hudson River anglers. The point estimate risk
17
TAMS'Gradient

-------
assessment will provide calculated high end and central tendency risks and hazard estimates, while
the Monte Carlo analysis will provide a distribution of risks and hazards for all anglers.
Response to HP 1-2
As stated in the HHRASOW (p. 9), "a hypothetical study population will be defined as any
individual who would consume self-caught fish from the Hudson River at least once per year in the
absence of a fishing ban." Based on the available fish consumption studies, the HHRA will evaluate
risks and hazards to a high-end fish consumer based on a range of body weights that do not
specifically identify gender. The fish consumption rates will incorporate information on males and
females, and adults and children, to the extent that each group is represented in the existing datasets
of angler populations.
Another factor that must be considered is the toxicity of the chemical. The toxicity values
that will be used in the HHRA are protective of both males and females. For cancer health effects,
the cancer slope factors (CSFs) that will be used are based on an increased incidence of liver tumors
in female rats, reflecting the potential greater sensitivity of this gender. The CSFs generated based
on female rats are higher than those generated for tumors found in male rats. Because risk is a
function of exposure and hazard, use of the higher CSF based on data from the female rats is more
protective of human health than one based on data from male rats.
For non-cancer health effects, the toxicity value is expressed as a Reference Dose (RfD). The
RfDs for PCBs are based on several studies in monkeys where females were exposed through
ingestion perinatally and as adults. The studies found reduced birth weights in offspring of the
perinatally-exposed monkeys and immune effects in adult female monkeys exposed for longer
periods of time. The RfDs that will be used have been adjusted lower to account for several factors,
including sensitive human subpopulations, such as children and the elderly. The use of these RfDs
in assessing potential non-cancer health effects is protective for the populations identified in the
comment.
Response to HP1-3. HG1-3. HG1-7. and HG1-30
The HHRA will evaluate current and potential risks and hazards to the reasonably maximally
exposed (RME) individual. The calculated risks and hazards for the RME individual are compared
to specific criteria of 10"4 to 10"6 for cancer health effects and a Hazard Index (HI) of 1 for non-
cancer health effects.
Consistent with USEPA's guidance which requires evaluation of the risk to an individual
(USEPA. 1989), the HHRA will not estimate the total number of anglers that consume their catch
or the total number of women of child-bearing age exposed through consumption of fish. In
addition, for the current exposure scenario, it would be difficult to identify the number of anglers
who are consuming fish in the presence of fishing bans and health advisories, because of the
potential for underreporting and the threat of fines for anglers keeping fish from the Upper Hudson.
18
TAMS'Gradient

-------
For the future scenario, it is not possible to project with any certainty the number of potential anglers
within various stretches of the river who would consume fish if bans and or health advisories were
lifted.
The comments that USEPA intends to use unrealistically high or worst case exposure
assumptions are not accurate. Consistent with USEPA guidance (USEPA, 1989 [e.g., see p. 14]),
the HHRA will calculate point estimates for central tendency (average) and high-end (RME)
exposures (HHRASOW, pp. 14, 15, 21, and 22).
Response to HG1-9 and HG1-42
The HHRA will describe all fishing bans and restrictions currently in place. For example,
the New York State Department of Health (NYSDOH) issues numerous health advisories on eating
sportfish from New York State for rivers, lakes and streams. The NYSDEC's fishing regulations
guide, which incorporates fishery guidelines set out by the Atlantic Marine Fisheries Commission,
outlines general angling regulations, such as daily limits, minimum lengths of fish, and dates of open
season. These general regulations are not health based, but presumably are established to prevent
depletion of fisheries.
However, fish advisories are not 100% effective in preventing or limiting fish consumption,
as described in numerous studies of compliance with fish consumption including Connelly's studies
(1990, 1992, and 1996). This assumption is supported by EPA's preliminary analysis of the 1992
Connelly data, which showed no significant difference in the mean number of freshwater fish meals
eaten when comparing NY water bodies with full, partial, or no advisories, despite the expectation
that the fishing advisories would likely suppress fish ingestion rates to some degree.
Therefore, to be protective of public health, the HHRA will evaluate Hudson River fish
ingestion rates in the absence of any Hudson-specific fishing bans. The effect of general, non-
specific NYSDEC and NYSDOH fishing regulations that would be in effect regardless of PCB
contamination levels in the Hudson River inherently will be taken into account through use of fish
ingestion rates from the Connelly gt al (1992) data.
The Maine study (Ebert £i al„ 1993) cited by the commenter will be considered in the HHRA
(HHRASOW, p. 8). However, this study is not the best source for angler consumption rates for the
Hudson River. In addition to the differences between New York and Maine anglers due to
differences in climate, fish species present, general fishing regulations etc., the Maine data set is
further limited in that survey respondents were asked only about total fish consumption from all
flowing water bodies, and not from individual water bodies separately. As a result, it was not
possible to screen to Maine data for more "Hudson-like" rivers and streams, as was possible with
the Connelly data.
19
TAMS-'Gradient

-------
Response to HG1-11 and HG1-33
The HHRA will not use the ChemRisk (1995) approach for assessing species-specific
consumption, as requested by the commenter. As described in the HHRASOW (pp.7- 9), the HHRA
will identify species preferences based primarily on the Clearwater studies (Barclay, 1993) and
supplemented by the Connelly data (Connelly ei al-, 1990, 1992, and 1996). Although the
ChemRisk (1995) approach is based on an analysis of fish species preferences from the Connelly
data, several of the assumptions are flawed. For example, more than 40% of the fish species
identified in the ChemRisk (1995) original distribution were salmon and trout species, which are not
present in the Hudson River; these fish species were arbitrarily assigned to be bass. In addition, the
authors arbitrarily apportioned the almost 20% of fish that fell into an "other" category evenly
among seven species found in the Hudson but not listed in the Connelly survey. Considered
together, more than 60% of the ChemRisk (1995) distribution was inconsistent with the data set
upon which it was reportedly based.
Furthermore, for the reasons noted in the response to HP1 -1, HP 1-7, and HG1 -1, USEPA will
use a probabilistic risk analysis as outlined in the HHRASOW (p. 15) and not conduct the micro-
exposure event analysis described by ChemRisk (1995).
Response to HG1-41
USEPA acknowledges the commenter's agreement with its approach regarding fish ingestion
rates for recreational anglers.
Response to HG1-43
USEPA has chosen to define the angler population as those individuals who consume self-
caught fish from the Hudson at least once per year to reflect the RME individual. While some
anglers may consume fish at frequencies less than once per year and some friends or family members
of anglers may consume "gift fish" at infrequent intervals, there are no data to quantify the risks or
hazards to these individuals. Moreover, USEPA's approach is protective of human health because
the exposure to these individuals would be lower than those calculated for the angler population.
USEPA will address the risk to these less exposed individuals in the uncertainty section of the
HHRA.
The commenter also suggested that Mid-Hudson fish ingestion rates consider short-lived
fishing activities, such as fish tournaments or short fish runs. However, there are no data to quantify
consumption rates due solely to these events. Nevertheless, to the extent that such fishing events
take place in various NY flowing water bodies, they are already accounted for in the Connelly data
sets and will be reflected in the fish ingestion rates used in the HHRA.
20
TAMS/Gradiem

-------
Response to HG1-44
USEPA acknowledges the commenter's agreement with USEPA's approach not to
distinguish subpopulations of highly exposed or lesser exposed anglers. As stated in the
HHRASOW (pp. 9 and 18), these subpopulations will be represented in the distributions of risk
generated in the Monte Carlo analysis for the Upper Hudson River HHRA and in the uncertainty
analysis, Monte Carlo analysis or in species-specific consumption rates for the Mid-Hudson River
HHRA. However, USEPA notes that the HHRASOW does not specify that the angler ingesting self-
caught fish is a recreational angler; he or she could be a subsistence angler.
Response to HG1 -45
The HHRA will identify species preferences based primarily on the Clearwater studies
(Barclay, 1993) and supplemented by the Connelly data (Connelly et gl. 1990, 1992, and 1996)
(HHRASOW, pp. 9-10 and 18). After determining average species preferences, these proportions
will be used to weight PCB concentration data for the point estimate analyses. The fact that not all
anglers consume the same species in the same proportions will be addressed in the uncertainty
analysis by evaluating the potential for anglers to consume only the most contaminated fish species.
The HHRA will clearly describe the exposure assumptions used for each population to aid the risk
management decision.
Response to HQ 1-46
As described in the HHRASOW (p. ] 8), the risk assessment intends to characterize species-
specific intake rates for anglers fishing in the Mid-Hudson River using an approach similar to that
used for the Upper Hudson. There is no intention to combine surrogate fish species preferences for
the Mid-Hudson with those of the Upper Hudson. Clearly, there are different fish species present
in the Upper and Mid-Hudson River, and therefore the potential for different angler preferences
exists.
D. Exposure Duration
Response to HF1-5
The concentrations of PCBs found in fish and sediment of the Upper Hudson are higher than
those found in the Mid-Hudson. USEPA will evaluate the risk to anglers in the Upper Hudson and
to anglers in the Mid-Hudson, An evaluation of the potential risks to the anglers who consume fish
from both the Upper and Mid-Hudson River would fall somewhere between the estimates for the
Upper and Mid-Hudson River and can be estimated from the proportion of fish consumed from each
location. Therefore, explicit quantification of this third exposure scenario is not necessary.
21
TAMS/Gradieni

-------
Response to HS1-2 and HP 1-8
The HHRASOW (pp. 10-11 and 18-19) addresses the potential for individuals to move from
one county to another and continue to fish sections of the Hudson River. As indicated in the
discussion, the HHRA will evaluate the distance anglers are willing to travel to fish. Data includes
information from the U. S. Bureau of Census on In- and Out-Migration Special Project Files and the
national census. The combination of datasets will enable a full evaluation of the potential exposures
of these anglers.
Use of a lifetime exposure (e.g., 70 years), as suggested by the commenters, is inconsistent
with the USEPA guidance (USEPA, 1989) and is more representative of a maximum exposure than
an RME scenario.
Response to HP 1-6
The effect of Hudson River fishing bans and advisories on fish consumption rates will be
addressed by using data from angler surveys, such as the Connelly database for New York State
(Connelly gt al., 1990, 1992, and 1996) to determine fish consumption rates in the absence of site-
specific fishing restrictions. As stated in the HHRASOW (p. 11), to the extent possible, the HHRA
will attempt to incorporate the likelihood that an angler may voluntarily choose to stop fishing, based
on an analysis of the percent of licensed anglers in the general New York State population in each
age group. The presence of a fishing ban would not be considered a voluntary decision.
Response to HG1-50
USEPA acknowledges the commenter's agreement with USEPA's decisions to characterize
exposure duration rather than use default values and to consider angler mobility between counties
(HHRASOW, pp. 10-11 and 18-19).
USEPA will consider the risk to anglers who move from one county adjacent to the river to
another county adjacent to the river (HHRASOW, pp. 10 and 18). However, the HHRA will not
consider the risk to an angler who moves from one specific reach of the Upper or Mid-Hudson to
another. Such an approach would involve too many possible combinations and would not be
protective of human health.
USEPA agrees with the comment that cessation of angling should be a factor considered, to
the extent possible, in exposure duration. USEPA will use a number of factors to fully evaluate
angler population activities over the exposure period (HHRASOW, pp. 11 and 19). The lifespan of
the population of anglers at the time of the "start date' will be addressed inherently through a
distribution of ages at which anglers stop fishing, whether due to voluntary fishing cessation, as
described above, or other reasons.
22
TAMS/Gradient

-------
E. PCB Cooking Losses
Response to HL1-3. HG1-13. HG1-51. and HG1-52
USEPA acknowledges the comment agreeing with USEPA's decision to evaluate the
available literature on the losses of PCBs during cooking in the Upper Hudson and Mid-Hudson
HHRAs (HHRASOW, pp. 11-12 and 19). As stated inTthe HHRASOW (pp. 11-12). USEPA's
determination whether these losses should be assessed qualitatively or quantitatively will be based
on a review the available literature (see response to HF1-6, below). As described in the HHRASOW
(p. 12). the preliminary recommendation is to assume no loss from cooking when calculating point
estimates, which is protective of human health, and to use a uniform distribution of the range of
observed cooking losses described in the published studies to represent cooking losses in the
uncertainty analysis (the second phase) of the Monte Carlo analysis.
In addition, inhalation of volatilized PCBs will be evaluated for the Upper Hudson and Mid-
Hudson HHRAs using data collected from near the Hudson River, where available (HHRASOW,
pp. 13 and 21). All data sets evaluated for the inhalation exposure pathway will be identified in the
HHRA. Based on the low vapor pressure of PCBs, the low concentration levels found near the
Hudson River, and a comparison of the CSFs for inhalation and oral exposure, volatilization is not
expected to be a major exposure route compared to fish ingestion.
Response to HF1-6
The commenter asked if PCBs lost in cooking increase exposure through volatilization.
Since issuance of the HHRASOW, USEPA reviewed the scientific literature available on National
Library of Medicine's Database MedLine and failed to find any scientific studies that have measured
the levels of contaminants volatilized during cooking. In the absence of any scientific studies in this
area, it is not possible to quantify the potential risks or hazards from this exposure route. Based on
a qualitative assessment of the cooking frequency for fish, the temperatures used in the cooking, the
various cooking practices used, and the relative toxicity of inhalation versus ingestion of PCB-
contaminated fish, the risks from inhalation while cooking are not considered to be a major exposure
route compared to the ingestion of fish.
Response to HF1-7
Although the Toxicological Profile for PCBs (ATSDR. December 1998) indicated that PCBs
are found in the organ meats and heads of PCB-contaminated fish. USEPA's review of the available
literature did not identify any studies that provide concentrations of PCBs in soup made from
contaminated fish heads. In the absence of information on the concentration of PCBs in soup
associated with levels found in fish, it is not possible to address this quantitatively. Therefore.
USEPA will address the exposure to PCBs from consumption of fish soup qualitatively in the
uncertainty section of the HHRA.
23
TAMS.Gradienl

-------
3. Toxicity Assessment
Response to HS1-3. HL1-1 and HP1-4
As stated in the HHRASOW (p. 14), in the point estimate portion of the HHRA, USEPA will
use the Aroclor-specific RfDs and the Total PCBs CSFs that have been established by USEPA to
assess non-cancer and cancer health effects, respectively (USEPA, 1996c, 1998a-c). Currently, there
are individual RfDs for Aroclors 1016 and 1254 (but none for 1242 or 1260) and three upper bound
and central estimate CSFs for Total PCBs. For samples containing a mixture of PCB congeners,
USEPA will evaluate the analytical data and use the CSFs and Aroclor-specific RfDs that are most
similar to the sample and most appropriate for the exposure scenario being modeled. USEPA will
not use distributions of toxicity values in the probabilistic (Monte Carlo) portion of the HHRA
(HHRASOW, pp. 14 and 21).
The tiered approach for partitioning and bioaccumulation mentioned by one of the
commenters is specific to congener data only, and therefore is not appropriate for the non-cancer
assessment of risk in the HHRA. Instead, USEPA will use its current guidance for evaluating cancer
risks associated with PCB exposure (USEPA, 1996c) to identify the most appropriate CSF. This
guidance is similar to the tiered approach for congeners, in that the highest applicable CSF is applied
for those PCBs bioaccumulated in the environment, including food chain exposures, sediment or soil
ingestion, dust or aerosol inhalation, and the potential for early lifetime exposure. For fish
consumption, a CSF of 2 (mg/kg-day)'1 will be used. For inhalation of volatilized PCBs, a CSF of
0.4 (mg/kg-day)'1 will be used. Because more than 1/2% of the total PCBs sampled have more than
4 chlorines and these higher chlorinated PCBs may be accumulated in the food chain, the use of the
less protective CSF of 0.07 (mg/kg-day)"1 is inappropriate.
Response to HS1-4
USEPA's guidance on development of inhalation reference concentrations (RfCs) does not
support the use of oral RfDs to calculate inhalation RfCs (USEPA, 1990). Such route-to-route
conversion is inappropriate for PCBs because the partitioning, transformation, and bioaccumulation
of PCBs are route-specific and there is no evidence to suggest that they would be the same for
inhalation as they are for ingestion.
Response to HG1-14 and HG1-53
USEPA acknowledges the comment supporting its use of Aroclor-based toxicity values
presented in the USEPA Integrated Risk Information System (IRIS) (USEPA, 1998a-c). USEPA
guidance (USEPA, 1996c) recommends that when congener concentrations are available, the
mixture-based approach (i.e., selection of one of the tiers of PCB slope factors based on the type of
environmental mixture) can be supplemented by analysis ofdioxin Toxicity Equivalency Factors
(TEFs) to evaluate dioxin-like toxicity. Consistent with this guidance, the carcinogenicity of dioxin-
24
TAMS/Gradient

-------
like PCBs will be presented as an uncertainty and discussed in the risk characterization section of
the HHRA.
Response to HG1-4. HG1-15. HG1-24. HG1-S4 through HG1-62
USEPA acknowledges the comments supporting its current CSFs for PCBs (USEPA, 1996c)
based on rat studies. USEPA's policies for reviewing human and animal data in evaluating potential
cancer health effects associated with exposure to carcinogens are provided in its current and
proposed guidelines for cancer assessment (USEPA, 1986, 1996b). These guidelines address
approaches to extrapolate animal data to humans. USEPA's evaluation of potential cancer health
effects includes evaluation of existing positive and negative epidemiological studies. The
epidemiological studies cited by the commenter were considered by USEPA in developing its
current CSFs for PCBs (USEPA, 1996c). Moreover, in May 1996, USEPA held a peer review
workshop, at which its current CSFs for PCBs were reviewed by independent experts on the
carcinogenicity of PCBs (USEPA, 1996d). The expert panel approved USEPA's approach and
recommended that CSFs for PCBs be based on the comprehensive rat study of Aroclors 1016, 1242,
1254, and 1260 (as described in USEPA. 1996c). The expert panel did not recommend that the
epidemiological studies be used to derive CSFs for PCBs, as proposed by the commenter. The
expert panel noted inadequacies in the epidemiological data with regard to limited cohort size,
problems in exposure assessments, lack of data on confounding factors, and the fact that
occupational exposures may be to different congener mixtures than those found in environmental
exposures, as well as other limitations and complications associated with interpreting human data
(USEPA. 1996d, p. 9).
The commenter proposed that USEPA use the lowest CSF submitted in comments on the
Water Quality Guidance for the Great Lakes System (TERRA, 1993). TERRA'S approach combined
estimates of PCB intake by capacitor workers with estimates of cancer mortality by Brown (1987)
or Taylor (1988). The commenter proposed a CSF calculated from the Taylor (1988) study, stating
it is "the largest epidemiological study performed to date and is highly relevant to the Hudson
River." A CSF based on this approach is flawed for the following reasons:
1.	Although the Taylor (1988) study is large, it is not necessarily better for the Hudson River
PCBs site than other studies of PCB exposure. Taylor's cohort included employees who
worked in offices and manufacturing areas where PCBs were not used.
2.	Although the Taylor (1988) study was conducted in New York State, occupational exposures
of employees in capacitor plants are not highly relevant to those of residents living near the
Hudson River. TERRA (1993) estimated that occupational exposure was largely by
inhalation, while USEPA expects residential exposures for the Hudson River PCBs site to
be through ingestion of contaminated fish, soil, or sediment. This is a critical difference, as
inhalation involves mostly the more volatile congeners of low chlorine content, while
contaminated fish, soil, and sediment contain more persistent and less volatile congeners of
high chlorine content. Thus workers and residents are exposed to different fractions of the
PCB mixtures.
25
TAMS'Gradient

-------
3.	Not only are workers and residents exposed to different fractions of the PCB mixtures, but
inhaled PCBs are likely to be less persistent and less carcinogenic than PCB congeners
ingested through contaminated fish, soil, or sediment. USEPA's current CSF (USEPA,
1996c) explicitly recognizes the large differences in persistence and toxicity for the different
mixture fractions encountered by different exposure pathways.
4.	TERRA (1993) used six studies to estimate occupational PCB exposure. Only one, however,
involved the same set of PCB mixtures (Aroclor 1254, then 1242, then 1016) used in the
plants studied by Brown and Jones (1981), Brown (1987) and Taylor (1988). There is great
uncertainty in combining mortality in the plants studied by Brown and Taylor with exposures
to other mixtures in other manufacturing plants, some of these from other countries.
5.	TERRA (1993) also used a pharmacokinetic approach to estimate occupational PCB
exposure. It relied, however, on so-called half-life estimates for different Aroclor mixtures.
In presenting its new CSFs, USEPA (1996c) discussed the fallacy of ascribing a single half-
life to a mixture as variable as PCBs. Further, these half-life estimates tend to be
underestimates; this tends to overstate occupational exposures and understate the CSFs.
Therefore, for the HHRA for the Hudson River PCBs site, USEPA will use its current CSFs,
including the current oral CSF of 2 (mg/kg-day)"1 for ingestion and will not use the less protective
oral CSF proposed by the commenter.
Response to HG1-16 and HQ 1-63 through HG1-70
In the HHRASOW (pp. 14 and 21), USEPA stated that it would use the non-cancer RfDs
established for PCBs in IRIS (USEPA, 1998a-c), which are the Agency's consensus toxicity values.
The commenter objected to the USEPA's oral RfD for Aroclor 1254, asserting that: (1) the clinical
relevance of immunotoxicity end points has not been demonstrated, (2) the ocular, dermal, and
nailbed changes observed in the rhesus monkey study do not correlate to such changes in humans,
(3) there are differences in metabolism of PCBs in monkeys and humans, and (4) the values for
uncertainty factors that were applied in extrapolating from the rhesus monkey studies to humans
were overly conservative. With respect to each of these points raised by the commenter:
1. The oral RfD for Aroclor 1254 developed by USEPA is appropriate for use in the
Hudson River PCBs Site HHRA. USEPA disagrees with the commenter's statement that the clinical
relevance of the immunologic changes used as one of the critical adverse effects has not been
demonstrated. Today, tests very similar to those employed by Tryphonas £l al- (1989, 1991 a.b) to
determine the levels of serum IgG and IgM (produced by the body's immune system to fight
infection and disease) in rhesus monkeys are widely used in hospitals and clinical laboratories to
diagnose immune deficiencies in suspected immuno-compromised patients (Bakerman, 1994, ABC's
of Interpretive Laboratory Data. 3rd edition, Interpretive Laboratory Data. New York). In addition,
the toxicology research community, as evidenced by presentations and audience attendance at
immunotoxicology sessions of the annual Society for Toxicology meetings, has expanded its
presentation and acceptance of immunotoxicology papers that use similar methods from a wide
variety of animal research studies (e.g.. Proceedings of the Society of Toxicology Meeting, New
26
TAMS/Gradicnt

-------
Orleans, LA, March 1999). Animal or human IgG and IgM antibody responses to sheep red blood
cells or similar multi-antigens systems are routinely and widely used in defining
immunocompromised diseases.
2.	The commenter stated that the rhesus monkey is not an appropriate model for the
dermal, ocular, and nailbed effects of PCBs in humans based on a study by Gillis and Price (1996),
who reported lower doses for observed dermal, ocular, and nailbed effects in the rhesus monkey than
in workers exposed to PCBs. However, the effects of PCBs exposures in rhesus monkeys reported
by Gillis and Price (1996) were the result of precise, consistent daily dosing that was monitored and
very well characterized, while the effects reported for workers exposed to PCBs were based on
sporadic doses, may not have been monitored, and were very poorly characterized. Because of these
differences in dosing, monitoring, and reporting, one cannot relv on the Gills and Price (1996) study
to conclude that the rhesus monkey is an inappropriate model for the effects of PCBs on humans.
The data from the rhesus monkey study result from well-controlled dosing, which is more desirable
for a quantitative assessment than exposure that is not well characterized.
3.	With regard to metabolism of PCBs in rhesus monkeys and humans, USEPA notes
that, while slight differences in metabolic processes have been observed by one research group,
differences in the critical adverse effects have not been demonstrated by other research groups.
4.	In developing the consensus RfD for Aroclor 1254, which is contained in the IRIS
database, USEPA used an uncertainty factor of 300 applied to the Arnold and Tryphonas studies,
which includes an uncertainty factor of three out of 10 to account for greater sensitivity in humans
than in rhesus monkeys, an uncertainty factor of three out of 10 to account for the longer life of
humans when compared to the rhesus monkey and the persistence of PCBs, and an uncertainty factor
of 10 out of 10 to account for sensitive humans, such as children and the elderly, and a partial factor
to account for the use of a minimal LOAEL (Lowest Observed Adverse Effect Level) because the
changes in the periocular (around the eye) tissues and nail beds seen at the 0.05 mg/kg-day dosage
are not considered to be of marked severity (USEPA, 1998b). The total uncertainty factor of 300
was developed consistent with USEPA guidelines and current risk assessment practices.
Response to HG1-17 and HG1-71
The commenter stated that a qualitative assessment of potential endocrine effects is
unwarranted. The USEPA disagrees with the comment. Estrogenic effects as well as effects upon
the thyroid gland and resultant metabolic changes has been observed and documented in many
mammalian species, including humans (USEPA, 1997b, c). In addition, the expert panel that peer-
reviewed the USEPA's CSFs for PCBs (USEPA, 1996d) recommended that USEPA discuss
endocrine effects. Therefore, consistent with the USEPA policies for assessment of endocrine
effects (USEPA, 1997b, c), the comprehensive summary of the relevant scientific literature on the
PCB toxicity contained in the HHRA will include the endocrine effects of PCBs (HHRASOW, pp.
14 and 21). A very extensive database is available from animal studies regarding the potential
endocrine effects that may be attributed to PCBs and the number of these reports increases every
27
TAMS/Gradiem

-------
year. In May 1998, the National Institutes of Environmental Health Sciences convened an
international panel of experts to discuss the available literature on endocrine effects of PCBs. The
results of this meeting will be presented in a future issue of the journal Environmental Health
Perspectives. The expert panel concluded that the endocrine effects of PCBs are important and
require additional research.
Response to HG1-18. HG1-35. and HG1-72
The USEPA agrees with the comment that its current policy for conducting probabilistic risk
analysis (USEPA, 1997a) and the associated guiding principles do not apply to dose response
evaluations for human health risk assessments. Because this issue is still under evaluation by
USEPA, the HHRA will not evaluate the uncertainty associated with toxicological data as proposed
by the commenter.
Response to HG1-19. HG1-40. HG1-73. HG1-74. and HG1-75
The commenter proposed using an averaging time based on the half-life of PCBs in humans,
to distinguish between background exposure of PCBs in angler body burdens and the increase in
body burdens related to the site. The evaluation of internal exposure based on half-life of PCBs in
humans was specifically addressed by the expert panel that reviewed USEPA's current CSFs for
PCBs. The expert panel did not support adjusting for internal dose at this time because the data are
not yet available to determine the appropriate dosimetric for PCB carcinogenicity (USEPA, 1996d,
p. 11). Consistent with USEPA's current CSFs for PCBs (USEPA, 1996c), the HHRA will be
conservatively protective of human health and no adjustments to exposure duration or averaging time
will be made in the HHRA based on consideration of the half-life of PCBs in the body.
4. Risk Characterization
Response to HG1-6 and HG1-28
As described in the HHRASOW (pp. 14 and 22), the risk characterization section of the
HHRA will present the current and future risks posed by the site and will explain the assumptions
used in the calculation of the risks. These assumptions include no institutional controls such as the
"catch and release" program for fishing in the Upper Hudson River. The risk characterization will
also address the uncertainties that are inherent in the various components of a risk assessment.
However. USEPA notes that NYSDEC's issuance of tickets for violation of the fishing restriction
indicates that any risks presented in the HHRA may not be entirely hypothetical, as suggested by the
commenter.
Response to HG1-20 and HG1-77
As stated in the HHRASOW (p. 6), the HHRA will contain both point estimate and
probabilistic (Monte Carlo) risk analyses. The purpose of providing both the point estimate and the
28
TAMS'Gradicnt

-------
probabilistic analyses is to provide clarity to the information presented (USEPA, 1992b, 1995b).
It also provides a means to compare the results of the two approaches.
Response to HG1-21. HG1-76. HG1-78. and HG1-79
As stated in the HHRASOW (p. 2), the HHRA is a tool to characterize the site contaminants,
evaluate the toxicity of the chemicals, assess the potential ways in which an individual may be
exposed to the contaminants, and characterize the risks posed by the site (OSWER Directive 9355.0-
30, Memorandum from Don Clay to Hazardous Waste Management Directors, entitled "Role of the
Baseline Risk Assessment in Superfund Remedy Selection Decisions"). The HHRA will calculate
both point estimate and probabilistic (Monte Carlo) estimates of risk (HHRASOW, p. 6).
USEPA will prepare an FS that evaluates remedial alternatives, solicit public comment on
a Proposed Plan that presents its preferred alternative, and select a remedy based on the nine criteria
set forth in the NCP (§300.430(e)(9)). The use of the HHRA by risk managers in the remedy
selection process is beyond the scope of the HHRA and therefore is not described in the risk
characterization section of the HHRASOW, as proposed by the commenter.
Response to HF1-9. HG1-80. and HP 1-9
As noted in the HHRASOW (p. 11), the start date for the exposure of anglers will be 1999.
This is appropriate because the HHRA evaluates current and future risk, and 1999 is the year in
which the HHRA will be completed. The suggestion that USEPA conduct a separate risk assessment
to evaluate the risks that might remain following implementation of various remedial alternatives
and, consequently, to change the start date to no earlier than 2002, is inconsistent with the purpose
of the HHRA, which is to evaluate current and future risk in the absence of remediation. Similarly,
use of a start date before 1999 would not be consistent with USEPA risk assessment policies and
guidance (USEPA, 1989).
C. Monte Carlo Analysis
Response to HG1-8. HG1-34. and HG1-38
The Monte Carlo portion of the HHRA will model variability and uncertainty separately, in
two stages (HHRASOW, p. 15). Details of the Monte Carlo modeling are not contained in the
HHRASOW, which is a general overview of USEPA's approach, but will be presented in the HHRA
itself. USEPA has reviewed the commenter's submittals in developing the Monte Carlo modeling
approach for the HHRA.
Response to HG1-36
The comment addresses the use of the 90th percentile in Monte Carlo modeling (HHRASOW.
p. 15). The >90,h percentile mentioned on p. 15 of the HHRASOW is part of the point estimate
29
TAMS/Gradient

-------
calculation of risk and is not part of the Monte Carlo analysis. The 90th percentile mentioned on pp.
15-16 of the HHRASOW is a typographical error. Consistent with Monte Carlo procedures, all
exposure parameters for which distributions are available will be randomly selected from the
distribution.
Response to HQ 1-3 7
The commenter agrees with USEPA that there are no data on actual year-to-year fish
ingestion rates for anglers in the published literature (HHRASOW, p. 16). To address this
uncertainty, the commenter suggests varying the intake rates within a fixed range of percentiles.
USEPA disagrees with this approach because there is no basis for selecting the range and because
it is reasonable to assume that an avid angler would remain an avid angler for his or her entire
exposure duration. USEPA's approach is protective of human health. For these reasons, the Monte
Carlo modeling will assume that each angler's fish ingestion rate remains constant from year to year.
III. RESPONSES TO SPECIFIC COMMENTS ON THE MID-HUDSON RIVER
HUMAN HEALTH RISK ASSESSMENT SCOPE OF WORK
Many of the issues raised in the comments are applicable to both the Mid- and Upper Hudson
River Risk Assessments. These issues were addressed in Sections II of this Responsiveness
Summary. The responses to issues specific to the Mid-Hudson River are presented below.
1. Plan, Synopsis & Objectives
Response to HL1-4 and HG1-27
As stated in the HHRASOW (p. 17), the model for food-chain uptake in the Mid-Hudson
River is being developed by Drs. Thomann and Farley for the Hudson River Foundation. These
researchers will determine how the model will be peer-reviewed and made available in the public
domain. The USEPA's use of the Thomann-Farley model will be presented in the Mid-Hudson River
HHRA. The Mid-Hudson HHRA will be made available for public comment and USEPA expects
that the Mid-Hudson HHRA will undergo peer review.
4. Exposure Assessment
A. Fish Concentration Data
Response to HG1-39 and HG1-49
The USEPA has consistently defined the site to include the entire Hudson River from Hudson
Falls to the Battery in New York City. It is not appropriate to divide the total dose to a Mid-Hudson
30
TAMS/Gradient

-------
River angler ingesting migratory fish into three separate doses (one each for the Upper Hudson, Mid-
Hudson, and Lower Hudson), as proposed by the commenter. The risk to an angler eating migratory
fish would not be affected by whether the uptake of PCBs into these fish occurred in the Upper
Hudson, the Mid-Hudson, or the Lower Hudson River. Further, evaluating the reduction of risk to
human health from implementing various remedial alternatives in the Upper Hudson is part of risk
management in the remedy selection process, which is beyond the scope of the HHRA.
D. Exposure Duration
Response to HF1-10
The fishing limitations that were in effect in the Mid-Hudson River at the time of the
Connelly surveys (HHRASOW, p. 18) are comparable to the fishing advisory currently in effect.
The NYSDOH advisories are available at its website at " n u .iH-.Uili.M.itc.nv.»v n \Mlnh 11 vh or by
contacting NYSDOH.
F. PCB Concentrations for Deterministic and Monte Carlo Analyses
Response to HF1-11
USEPA agrees with the comment. The intent of the sentence is to convey that if PCB levels
are declining with time, these individuals will have lower risk than the original exposed population.
31
TAMS'Gradient

-------
IV REFERENCES
Achman, D.R., K.C. Hornbuckle, and S.J. Eisenreich, 1993. Volatilization of Polychlorinated
Biphenyls from Green Bay, Lake Michigan. Environ. Sci. Technol. 27(1 ):75-84.
Bakerman. 1994. ABC's of Interpretive Laboratory Data. Third Edition. Interpretive Laboratory
Data, New York.
Brown, D.P. 1987. Mortality of Workers Exposed to Polychlorinated Biphenyls: An Update. Arch.
Environ. Health. 42(6):333-339.
Brown, D.P. and M. Jones. 1981. Mortality and Industrial Hygiene Study of Workers Exposed to
Polychlorinated Biphenyls. Arch. Environ. Health. 36(3): 120-129.
Buckley, E.H. and T.J. Tofflemire, 1983. Uptake of Airborne PCBs by Terrestrial Plants Near the
Tailwater of a Dam. Proc. Natl. Conf. on Environ. Eng., ASCE Specialty Conference, July 6-8, pp.
662-669.
ChemRisk. 1995. Determining the Intake of Upper Hudson River by Species. McLaren/Hart-
ChemRisk, Portland, ME. January.
Ebert, E.S., N.W. Harrington, K.J. Boyle, J.W. Knight and R.E. Keenan. 1993. Estimating
consumption of freshwater fish among Maine anglers. N.Am. J. Fish. Mgt. 13:737-745.
Gillis, C. and P. Price. 1996. Comparison of the Non-Carcinogenic Effects and PCB Body Burdens
in Rhesus Monkeys and Humans: Implications for Risk Assessment. Toxicologist 30(1): 748.
Hornbuckle. K.C., J.D. Jeremiason, C.W. Sweet, and S.J. Eisenreich, 1994. Seasonal Variations in
Air-Water Exchange of Polychlorinated Biphenyls in Lake Superior. Environ. Sci. Technol.
28:1491-1501.
Hornbuckle, K.C., D.R. Achman, and S.J. Eisenreich, 1993. Over-water and Over-land
Polychlorinated Biphenyls in Green Bay, Lake Michigan. Environ. Sci. Technol. 27(l):87-98.
Nelson, E.D., L.L. McConnell, and J.E. Baker, 1998. Diffusive Exchange of Gaseous Polycyclic
Aromatic Hydrocarbons and Polychlorinated Biphenyls Across the Air-Water Interface of the
Chesapeake Bay. Environ. Sci. Technol. 32:912-919.
Taylor, P.R. 1988. The Health Effects of Polychlorinated Biphenyls. Harvard School of Public
Health, Boston, M.A.
TERRA, Inc. 1993. Comments on the Water Quality Guidance for the Great Lakes System.
32
TAMS/Gradient

-------
Typhonas, H., M.I. Luster, G. Schiffman, L.L. Dawson, M. Hodgen, D. Germolec, S. Hayward, F.
Bryce, J.C.K. Loo, F. Mandy and D.L. Arnold. 1991a. Effects of Chronic Exposure of PCB
(Aroclor 1254) on Specific and Non-Specific Immune Parameters in the Rhesus (Macaca mulatto)
Monkey.
Tryphonas, H., M.I. Luster, K.L. White, P.H. Naylor, M.R. Erdos, G.R. Burleson, D. Germolec, M.
Hodgen, S. Hayward and D.L. Aronold. 1991b. Effects of PCB (Aroclor 1254) on Non-Specific
Immune Parameters in Rhesus {Macaca mulatto) Monkeys. Int. J. Immunoph. 13(6):639-648.
Typhonas, H., S. Hayward, L. O'Grady, J.C.K. Loo, D.L. Arnold, F.Bryce nad Z.Z. Zawidzka.
1989. Immunotoxicity Studies of PCB (Aroclor 1254) in the Adult Rhesus (Macaca mulatta)
Monkey. Preliminary Report. Int. J. Immunoph. 11(2): 199-206.
33
TAMS/Gradient

-------
Federal

-------
U.S. DEPARTMENT OF COMMERCE n
National Oceanic and Atmospheric
Administration
National Ocaan Saivica
Office of Ocaan Resources Conservation and Assessment
Hazardous Materials Response and Assessment Division
Coastal Rasourcas Coordination Branch
290 Broadway, Rm 1831
Naw York. New York 10007
August 28,1998
Doug Tomchuk
U.S. EPA
Emergency and Remedial Response Division
Sediment Projects/Caribbean Team
290 Broadway
New York, NY 10007
Dear Doug;
Thank you for the opportunity to review the July 1998 Hudson River PCBs Reassessment
RI/FS Phase 2 Human Health Risk Assessment Scope of Work. The following comments art
submitted by the National Oceanic and Atmospheric Administration (NOAA).
Summary
The July 1998 Hudson River Human Health Risk Assessment (HHRA) Scope of Work
(SOW) indicates that a separate assessment will be performed for the Upper Hudson River
(from Hudson Falls to the Federal Dam in Troy) and the Mid Hudson River (from Albany to
Poughkeepsie). The assessment will evaluate carcinogenic and noncarcinogenic health effects.
Total risks and hazards from PCBs will be developed for the average exposed and reasonably
maximally exposed individual.
Comments
Pages 1 and 17: Data included in the risk assessment: all relevant fish tissue PCB data should
be used, including the most current NYSDEC fish tissue database, data collected by EPA and
NOAA as part of the 1993 Ecological Risk Assessment sampling, and fish tissue data collected 1
by NOAA in 1995. It is still not clear from Section 1.1. Paragraph 4 what the fish tissue
dataset is. The description seems to also ignore fish tissue data collected between the early
1980's and 1990.
A HHRA will be performed for the Upper and Mid Hudson River but not for the Lower
Hudson between Poughkeepsie and the Battery. Since the Phase I HHRA concluded that
ingestion of fish in the Lower Hudson River would produce similar risks to those determined j
for the Upper Hudson River, it is unclear why the scope of the proposed activities does not
include either a quantitative or qualitative assessment far the Poughkeepsie to Battery stretch of
the Hudson River.
Page 4: Risks from ingestion of vegetables, meat, etc. were not quantified during the Phase I
Ride Assessment due to insufficient data. TTiese risks should be evaluated in the proposed ^
HHRA. Numerous farms are located within the floodplain of the Upper Hudson River, hence
exposure from vegetables, meat, eggs, and milk (including dairy products) raised on
potentially PCB-contaminated soils should be assessed. It there is insufficient data to evaluate
this risk, then samples should be collected to aid in the determination.

-------
NOAA comments on July 1998 Hudson River Human Health Risk Assessment SOW
(8/28/98)
Pages 7 and 18: A decrease in fishing activity or fish consumption due to consumption
advisories/bans should not influence an assessment of human health risk. Advisories are
issued based on NYS Department of Health recommendations due to contaminant
concentrations and exceedaoces of FDA, EPA or DOH limits.
Pages 10 and 18: Exposure duration assumes a mean distance of 34 miles travelled to fish.
Neither the Upper or Lower Hudson HHRA account for the possibility that someone living
near the Federal Dam could fish both the Upper and Mid Hudson River and represent a third
exposure scenario.
" " 1: Do potential losses of PCBs in fish during cooking increase exposure to volatilized
Page 12: Modeled FOB concentrations for fish fillets will be used in the deterministic and
Monte Carlo analyses. The models should also take into account the potential for using fish
headsAailsinsoups.
Page 13: Other exposure pathways should be subpopulation consumption of meat, vegetables,
eggs, and daily raised in potentially PCB-contammated fioodplains of the Upper Hudson.
Page 15: If the starting point for exposure is 1999 and it is assumed that no exposure occurred
prior to that date, then die model could underestimate risk. Baseline should therefore be set at
various exposure levels due to potential long-tens exposure prior to 1999.
Page 18: In paragraph 4, it is stated that fishing limitations were in effect during the Connelly
surveys. "Fishing limitations" should have been defined and compared to present conditions.
Page 19: In paragraph 3, it is stated that "Since PCB levels in fish seem to be declining with
time—" yel on page 11 it is stated that "If PCB levels in fish decline with time...**. It appears
that EPA has concluded that PCB fish levels are declining with time in the Mid Hudson River
but not the Upper Hudson River. The trend in fish tissue PCB concentrations can only be
resolved with continued monitoring. The Reassessment RI/FS was jrompted in part by the
lack of significant declines in fish PCB concentrations that were predicted during the original
RI/FS. lie sentence on page 19 should be replaced with the one on page 11: "If PCB levels in
fish decline with time.- .
Thank yon for your continual efforts in keeping NOAA apprised of the progress at this site.
Please contact me at (212) 637-3259 or Jay Held at 206-526-6404 should you have any
questions or would like further assistance.
4
5
6
7
8
9
10
11
Lisa Hessian
NOAA Cbastal Resource Coordinator
cc Michael Qemetson, DESA/HWSB
Robert Hargrove, DEPP/SPMM
Anne Secord, USFWS
Anton P. Giedt, NOAA
Gina Ferxeira, ERRD/SPB
Charles Merckel, USFWS
William Ports, NYSDEC
2

-------
KA
ST
<¦*
to

-------
HS-1
New York State Department of Environmental Conservation
Division of Environmental Remediation
Bureau of Centra! Remedial Action, Room 228
50 Wolf Road, Albany, New York 12233-7010
Phone: {518) 457-1741 FAX: (518) 457-7925
John P. Cahill
Commissioner
August 31,1998
Mr. Douglas Tomchuk
United States Environmental Protection Agency
Region II
290 Broadway - 20* Floor
New York, NY 10007-1866
Dear Mr. Tomchuk:
Re: Hudson River PCBs Reassessment RI/FS
Site No.: 5-46-031
The New York State Department of Environmental Conservation (NYSDEC) and the New York
State Department of Health (NYSDOH) have reviewed the July 1998 Hudson River PCBs Reassessment
RI/FS reports entitled "Volume 2C-A Low Resolution Sediment Coring Report Addendum to the Data
Evaluation and Interpretation Report,'* and "Phase 2 Human Health Risk Assessment Scope of Work."
This letter provides the State's comments on the two documents.
The Low Resolution Sediment Coring Report (LRSCR) presents four major findings. Following
are the State's general comments corresponding to each of these findings.
Finding 1
"There was little evidence found of widespread burial of PCB-contaminated sediments fay clean
sediment in the Thompson Island Pool. Burial was seen at some locations, but more core sites
showed loss of PCB inventory than showed PCB gain or burial.'* [Page ES-3J
State Comment
The State agrees that, based on the data contained In the LRSCR, much of the PCB-contaminated
sediments in the Thompson Island Pool are not being buried with significant amounts of clean
sediment
Finding 2
"From 1984 to 1994, there has been a net loss of approximately 40 percent of the PCB inventory
from the highly contaminated sediment in the Thompson Island Pool." [Page ES-4]

-------
State Comment
The State agrees that, based on the data contained in the LRSCR, there has been an identifiable
PCB inventory loss from the sediments of the Thompson Island Pool. However, based on the
data contained in the report, it is difficult to closely quantify the degree of sediment losses. It
may be more appropriate for the report to present a range of estimates rather than a single
number. This same concern was discussed at the Scientific and Technical Committee meeting on
August 18, 1998.
Finding 3
^Trom 1976-1978 to 1994, between the Thompson Island Dam and the Federal Dam at Troy,
there has been a net loss of PCB inventory in hot spot sediments sampled in the low resolution
coring program." [Page ES-4]
State Comment
The State agrees that, based on the data contained in the LRSCR, there has been an identifiable
PCB inventory loss from the hot spots between the Thompson Island Dam and the Federal Dam
at Troy.
Finding 4
The PCB inventory for Hot Spot 28 calculated from the low resolution coring data is
considerably greater than previous estimates. This apparent "gain" in inventory is attributed to
significant underestimates in previous studies rather than actual deposition of PCBs in Hot Spot
28." [PageES-4]
State Comment
The State agrees with this finding based on the data contained in the LRSCR. This inaccuracy in
past data gathering efforts may also be present in the PCB inventory estimates in other areas
•where the core depths were not sufficient in the past. However, NYSDEC believes the USEPA
evaluation of sediment PCB inventory gain or loss is valid, and not impacted by the earlier data
gathering efforts.
Hie State also has the following specific comments regarding two other findings of the LRSCR:
1.	Page ES-5 and Section 4.1.4 second paragraph — The finding that areas within the Thompson
Island Pool (TIP), outside the known hot spot areas of the TIP, have exhibited a large net gain in
PCB inventory (up to a 100% increase) is significant because the PCBs are more readily
available to fish and other biota.
2.	Section 4.4.3 The revised, sediment PCB concentration estimates for the near shore areas are
noteworthy. This portion of the river environment has not been well characterized in past
investigations, and this information will be useful to both the ecological and human health risk
assessments for the site.
The following are the State's comments, including the NYSDOH, on the Phase 2 Human
Health Ride Assessment Scope of Work:
1. The first sentence of the first full paragraph of page 9 refers to a hypothetical study population
being defined as any individual who would consume self-caught fish from the Hudson River uin
Page 2.

-------
the absence of a fishing ban." This passage should be revised for accuracy to read "...in the
absence of a fish possession ban and health advisory."
2.	The number of years that a person may eat contaminated fish from the Hudson River is estimated
in Section 11,2.D entitled "Risk Characterization from the Consumption of Fish." Data on how
long people live in a county along the river before moving are used to estimate the number of *
years a person may eat contaminated fish. A significant number of people are likely to move
from one county along the river to another county along the river, thus increasing their length of
exposure. The number of years that a person may eat contaminated fish from the Hudson River
will be underestimated if this possibility is not considered in estimating exposure. Furthermore,
a lifetime exposure should be considered in the exposure distribution.
3.	In evaluating risks, both cancer and non-cancer, the reference dose or cancer potency factor for
the Aroclor (e.g. Aroclor 1016, Aroclor 1260, etc.) that is most similar to the PCB mixture in the 3
environmental samples should be used. This approach is more scientifically defensible than
automatically using default values as suggested in the Integrated Risk Information System
guidance.
4.	Non-cancer risks are evaluated by comparing exposures to reference doses (ingestion exposure)
or reference concentrations (inhalation exposure).Since reference concentrations are not
available for the Aroclors, inhalation exposures should be evaluated using reference doses. The 4
risk characterization section of a risk assessment includes a discussion of the uncertainties and
limitations of the risk assessment and the uncertainties and limitations, if any, of using reference
doses instead of reference concentrations should be included in that section.
As additional information becomes available to the parties, the State would welcome the
opportunity to provide comments. The State views the completion of the LRSCR and the Risk
Assessment Scope of Work as important Hudson River Reassessment milestones, and is pleased that
USEPA is adhering to its Reassessment schedule.
Sincere
\ ' r
William T. Ports
Remedial Section A
Bureau of Central Remedial Action
cc: John Davis, NYSDOL
Robert Montione, NYSDOH
Jay Fields, NOAA
Lisa Rosrnan, NOAA
Anne Secord, USF&WS
Page 3.

-------
e
o

-------
HL-1
SARATOGA COUNTY
ENVIRONMENTAL MANAGEMENT COUNCIL
PETER BALET	OiORGE HODGSON
CHAIRMAN	DiMCTOB
August 28, 1998
Mr. Douglas Tomchuk, Project Manager
Hudson River PCB Reassessment
USEPA, Region 2
290 Broadway, 20th Floor
New YOrk, N.Y. 10007- 1866
Attn: HHRA SOW Comments
Dear Mr. Tomchuk:
Below you will find comments from the Saratoga County Environmental M^iagement
Council relative to the Hudson River PCB Reassessment Phase 2, "Human Health Risk
Assessment Scope of Work, July 1998:
COMMENTS ON PHASE 2 HUMAN HEALTH RISK ASSESSMENT SCOPE OF
WORK, JULY 1998
Prepared for the Saratoga County Environmental Management Council
byD.D. Adams, Member-at-Large
1. P. 3.: Are any data available which indicate a difference in the cancer slope factor
(CSF) for lower chlorinated PCBs vs higher chlorinated PCBs? If such data are available, 1
does EPA intend to include different CSF values in the Risk Assessment? The SCEMC
believes this difference could be significant and should be considered by EPA.
2. P. 8. Rather than rely on a survey of anglers from areas outside the Upper Hudson
River Valley, a survey should be made of anglers in the Upper Hudson River Valley only.
To circumvent the problem created by the ban on retaining fish caught in the Upper
Hudson River, the survey could look primarily at fishing in areas within 50 miles of the
Hudson River and consumption of these fish and at catch and release fishing in the Hudson
River. It could be assumed that fish caught & released would likely be eaten, if allowed, if
fish caught in other areas were eaten. Since the amount of fish consumption is a key part
of the Human Health Risk Assessment, it is not acceptable for EPA to say there is no time
for doing such a survey or that doing such a survey would delay the final Reassessment
decision. It should be paramount in EPA's mind to be sure that the final Reassessment
decision is predicated upon the soundest scientific basis possible. The potential
*0 WIST HIOM STRUT
•ALIBTON l»A, N.Y. 12020
(Bit) BS4-4778

-------
consequences of an unsound final decision are too great for EPA to proceed on any other
basis.
3.	P. 12&13: In the EPA Availability Meeting held on August 19, 1998, EPA stated that
loss of PCBs during cooking as well as inhalation of PCBs are not expected to be
significant parameters in EPA's Health Risk Assessment. If these assumptions turn out to
be incorrect, more precise data on these parameters should be obtained for the same
reasons as stated in Comment 1. (Note: comment on previous Phase 2 reports have
questioned the applicability of volatilization data from areas outside the Hudson River
Valley)
4.	P. 17: EPA should provide the public with information on where & when the work on
the Lower Hudson River Model is to be reported. Will comments be accepted on this
work & will it be subjected to peer review?
Thank you for the opportunity to comment.
Sincerely,
Peter M. Balet Gfi
Chairman
PB/gh
cc: Ms. Carol Browner, Administrator, USEPA
Ms. Jeanne Fox, Regional Administrator, Region 2, USEPA
Mr. Richard Caspe, Director, ERRD, Region 2, USEPA
Mr. William McCabe, Deputy Director, ERRD, Region 2, USEPA
Ms. Ann Rychlenski, Public Affairs Specialist, Region 2, USEPA
The Honorable Gerald Solomon
The Honorable Alphonse D'Amato
The Honorable Daniel Moynihan
The Honorable George Pataki
Mr. John Cahill, Commissioner, NYSDEC
Mr. Stuart Buchanan, Region 5 Director, NYSDEC
The Saratoga County Board of Supervisors
Mr. David Wickerham, Administrator, Saratoga County
Hudson River PCB Liaison Group Chairs
SCEMC members & staff

-------
Public Interest Groups

-------
SCENIC
HUDSON
HP-1
INC
9 VASSAR STREET • POUGHKEEPSIE, NY 12601 • (914) 473-4440 • FAX (914) 473-2648
Scenic Hudson's Comments on
US EPA's Scope of Work for
Phase Two Human Health Risk Assessment -
Hudson River PCBs Reassessment RI/FS
Full characterization of the health effects associated with exposure to Hudson River PCBs is a
critical part of the EPA's Reassessment of the Hudson River. While we were pleased to see that
the health risk assessment will now quantify both carcinogenic and non-carcinogenic health
effects from exposure to PCBs for the Upper and Mid-Hudson River, we have the following
concerns:
1.	Accounting for Subsistence Fishing - It appears that exposure due to subsistence fishing will
not be adequately sampled and accounted for in the analysis proposed. Although subsistence
fishing is briefly mentioned on page 9, it is not clear that the creel count and license approach \
planned will give adequate representation of exposure for subsistence anglers and their families.
The existing literature on fishing and fish consumption on the Hudson indicates strongly that
subsistence fishing is common. Therefore it is imporram that this segment of the population be
more carefully accounted for.
2.	Risk to Family Members - As proposed, the health risk assessment does not address the
issue of risk for family members of anglers. Clearwater's angler survey indicated that 58 % of ^
fishermen questioned gave fish to their families for consumption. Risk to these family members
- women and children - must be thoroughly addressed.
3.	Potentially Exposed Populations - Somewhere in the assessment, either in the exposure
assessment or risk characterization, a characterization of the population or populations exposed
to PCB contamination should be included For example, EPA should estimate the number of
anglers thatconsumetheir catch. The assessment should acknowledge that the population	3
exposed is larger than the population of anglers because many anglers share their catch. Further,
because some PCB health effects are associated with pre-natal exposure, the assessment should
estimate the number of women of child-bearing age that consume fish. The characterization of
this potentially affected population would contribute to the risk characterization and qualitative
discussion of endocrine disruption. In other words, it would be desirable to characterize the size
of the total potentially affected population.
Although die study area extends only as far south as Poughkeepsie, the Superfund site extends to
New Yoik City. A sizeable segment of the potentially affected population is outside the study
email: scenichu@mhv.net

-------
Continued
Page 2
area for the risk assessment. Although individual risks may be greatest for anglers within the
study area, EPA can and should acknowledge that the population exposed to PCB contamination
from the site is very large. Unlike most Superfund sites, this site has a very large exposed
population and the "maximally exposed individual" should not be the sole measure of exposure
or determination of acceptable risk.
4.	Toxicity Assessment - The risk assessment proposes to use toxicity values of PCB arochlors
1016 and/or 1254 because the 99reference doses and cancer slope factors have been established
for these two arochlor compounds. It is unclear which value will be used for comparison with
PCB mixtures as they occur in the Hudson River or in Hudson River fish. Choosing either 1016
or 1254 may not adequately capture the toxicological profile of 1242, the arochlor discharged in
greatest volume to the Hudson and present in Hudson River fish.
In a recent article in Environmental Health Perspectives, (Volume 106, Number 6, June, 1998,
article attached) entitled "Assessing the Cancer Risk From Environmental PCBs" by James
Cogliano (Volume 6, Number 6, June, 1998, article attached), it is indicated that EPA in fact has
a method for PCB risk assessment that would address this problem (page 321). According to the
article, EPA is using a tiered approach that considers how partitioning and bioaccumulation
affect each exposure pathway. Using this approach, the high risk slope factor is applied in
situations where: environmental processes tend to increase risk; there is food chain exposure;
there is sediment or soil ingestion; there is dust or aerosol inhalation; there is exposure to dioxin
like, tumor-promoting, or persistent congeners, and there is early life exposure (all pathways and
mixtures). Under this approach, the Hudson River would fall into the "high risk and persistent"
category, the highest risk tier. We question why EPA is instead planning to use comparative
toxicity values instead of it's own current procedure that would require use of the highest risk
tier and attendant slope factors.
5.	Other Exposure Scenarios - The scope (page 7) states that an objective of the Phase n risk
modeling is to confirm the Phase I conclusion that fish ingestion "outweighs" other exposure
scenarios. Fish ingestion is likely to dominate risks from other exposure pathways. However, it
is important to acknowledge that exposed individuals are likely to be exposed to multiple
pathways. We support the proposal (section U(2)(HX1)) to present total risks and hazards added
together across pathways. Pathways with risks above 10-6 should not be excluded because they
are relative to fish ingestion.
Recent studies from the Chicago area document an "urban plume" of PCB contamination. Based
on these findings and other evidence that PCBs are transported by air deposition, EPA should
reevaluate the feasibility of estimating food chain pathway risks via air deposition, and milk and
beef ingestion. If data to evaluate this pathway are unavailable, EPA should at least include a
discussion of the relevant literature related to this exposure pathway.
6.	Fishing Bans - The impact of fishing bans on fish consumption rates (discussed on page 7,
section 2B) should not be included for baseline risk results. Fishing bans could be factored in as
part of die "institutional controls" in a remedial scenario. It is not clear that this is EPA's

-------
Continued
Page 3
proposed approach. People who voluntarily stop fishing (discussed on page 11) because of the
fishing advisories should not be included in the baseline risk assessment There are many
average recreational anglers that are not aware of the PCB contamination, its potential risks, or
the health advisories. The assessment should address these anglers and more highly exposed
subpopulations.
7.	Use of Angler Surveys - The Clearwater angler survey and the Chemrisk survey from Maine
should be evaluated based on availability and relevance. The Maine survey is less relevant to the
assessment than the Clearwater angler survey, which addresses Hudson River anglers. On page 9
in the discussion on angler surveying it should be noted that interviewers may avoid areas where
subsistence fishing may be the greatest due to socio-economic conditions. Also, subsistence
anglers may be under-represented in surveys because of language barriers. Mail-in surveys
under-represent illiterate anglers. In the same discussion it states that "attempts will not be made
to distinguish between subpopulations of highly exposed or lesser exposed...". Does this refer
only to the Monte Carlo approach? Highly exposed anglers should definitely be distinguished for
the deterministic risk model.
8.	Exposure duration - Exposure duration will be estimated assuming that anglers fish from the
Hudson only as long as they live in a county along the Hudson. To estimate the time of
exposure, EPA will analyze "information such as the length of time people live in a single
residence". This may underestimate exposure duration because some anglers will move to a new
residence within the same county or to a residence in another county along the river. For the
deterministic model, high end exposure duration should be based on an angler that lives and
fishes along the Hudson all of his/her lifetime. This is not an overly-conservative scenario.
Later (page 12), the scope says that a high-end estimate of exposure duration will be used for the
deterministic model. What will that estimate be? Will it be lifetime or some high-end estimate
based on county-level mobility data? The simple and more conservative lifetime assumption is
more defensible that a complicated and highly uncertain estimate based on mobility data.
9.	Start Date - The EPA proposes that the risk model will address 1999 onward. It would be
more appropriate to select an earlier start date. Most of the population of anglers in 1999 has
been fishing from the Hudson in earlier years. As proposed, will pre-1999 exposures be ignored?
Some of the 1999 anglers were probably fishing in the 1970s. Fish contamination data are
available in great abundance from at least the early 1980s and should be used. Also, the EPA
reports so far to have focussed on a modeling period in the early 1990s. Why wouldn't the risk
assessment include this period also for consistency and comparison of results?
10/6/98

-------
General Electric

-------
Sidley & Austin
HG-1
A FARTNftJMIF INCLUDING PROFESSIONAL CORPORATIONS
CHICAGO
DALLAS
1722 Eve Street, N.w.
Washington, D C. 20006
Telephone 202 736 8000
Facsimile 202 736 87 1 1
NEW YORK
LONDON
LOS ANGELES
SINGAPORE
Founded 1866
TOKYO
WRITER'S DIRECT NUMBER
(202)736-8161
August 31,1998
Douglas Tomchuk
ISEPA- Region 2
290 Broadway
20th Floor
New York, NY 10007-1866
Enclosed please find General Electric Company's ("GE") comments on the
"Hudson River PCBs Reassessment RI/FS Phase 2 Human Health Risk Assessment Scope of
Work." These comments provide a detailed critique of the Scope of Work, and I will not repeat
that discussion here. One specific issue, however, deserves emphasis.
One of GE's primary concerns is the Scope of Work's vague, muddled, and at
times inconsistent description of the Agency's proposed baseline human health risk assessment.
This has increased the difficulty of assessing and commenting on the Scope of Work.
Consequently, in addition to pointing out the portions of the Scope of Work with which GE
agrees or disagrees, GE's comments provide specific recommendations on how the Agency
should conduct the baseline risk assessment. GE also urges the Agency to reissue the Scope of
Work to provide a more coherent description of the risk assessment and to respond to the issues
raised in GE's comments.
GE welcomes the opportunity to discuss its comments with EPA in greater detail.
Re: HHRA SOW Comments
Dear Mr. Tomchuk:
Sincerely,
Thomas G. Echikson

-------
Sidley & Austin
Washington, D.C.
Douglas Tomchuk
August 31,1998
Page 2
cc: Richard Caspe
William McCabe
Melvin Hauptman
Douglas Fischer, Esq.
Marianne Olson
John Cahill
Frank Bifera, Esq.
Walter Demick
D:\GENbudU2\HHRAeomm
-------
COMMENTS OF GENERAL ELECTRIC COMPANY
ON
Hudson River Reassessment RI/FS
Phase 2 Human Health Risk Assessment
Scope of Work
July 1998
August 31, 1998
Mclvin B. Schweigcr
John G. Haggard
General Electric Company
Corporate Environmental Programs
1 Computer Drive South
Albany, NY 12205
Russell Keenan, PhD
Paul Price
Ogden Environment and Energy Services
15 Franklin Street
Portland, ME 04101
-------
TABLE OF CONTENTS
Page
Executive Summary.....	1
SECTION I. INTRODUCTION	6
SECTION H. GENERAL COMMENTS	7
A.	The site does not extend below the Troy Dam...		 7
B.	The Agency must provide advanced notice of its intent to use additional data	8
C.	The Agency must emphasize that the risk estimates presented in the
HHRA are hypothetical	9
SECTION IE. EXPOSURE ASSESSMENT			11
A.	The proposed approach for determining fish tissue concentrations for
the high-end angler is flawed		 12
B.	Monte Carlo Modeling				 13
B.l. The SOW does not provide an adequate description of the model	13
B.2. The SOW does not explain whether and how the Agency intends to model
variation and uncertainty	14
B.3. Other issues concerning Monte Carlo modeling						 16
B.4.	Recommended approach			17
C.	Fish consumption rates							19
C.l.	Development of a fish consumption rate distribution for recreational
anglers					19
C.2. Angler Subpopulations	26
C.3.	Species-specific fish ingestion rates					30
D.	Determination of future PCB concentrations in fish	31
D.l.	Use of model results						 31
D.2. Selection of Mid Hudson fish species	33
E.	Duration of exposure											33
F.	Cooking loss	35
G.	Inhalation exposures...							35
-------
SECTION IV. TOXICITY ISSUES	36
A.	GE supports the use of Aroclor-based toxicity criteria	36
B.	Cancer dose response	38
B.l. The rationale for using epidemiological studies to establish environmental
standards	38
B.2. Epidemiological data	43
B.3. Derivation of a cancer slope factor for PCBs from the epidemiological
studies	55
B.4. Summary for cancer toxicity assessment	57
C.	Noncancer toxicity values	58
D.	Endocrine disruption	62
E.	Uncertainty in toxicological criteria	63
F.	Averaging time			65
SECTION V. RISK CHARACTERIZATION ISSUES	67
A.	Consideration of background sources of PCBs	67
B.	Development of the central and high-end exposure risk estimates	68
C.	The baseline assessment cannot be used to select remedial options	69
D.	Use of risk finding for small number of anglers fishing hot spots	69
E.	Choice of "start date"	69
REFERENCES	71
-------
FIGURE
Figure 1. Reprint of Figure 9-12 from the EPA Preliminary Model Calibration Report
-------
EXECUTIVE SUMMARY
EPA's Scope of Work for the Human Health Risk Assessment ("SOW") for the
Hudson River PCBs Superfund Site ("Site") sets out the Agency's proposed approach for
conducting and preparing the baseline human health risk assessment ("HHRA") for the
Site. The SOW proposes two risk assessments - one for the Upper Hudson and one for the
Mid Hudson - using deterministic and probabilistic methods and standard, IRIS-derived
toxicity values. This information will then be used to present a hypothetical statement of
the risks associated with the Site against which the risk reduction achieved by various
remedial alternatives can be measured.
There is much in the SOW that General Electric Company ("GE") supports. For
instance, GE agrees with the Agency's proposal to use Monte Carlo modeling to develop a
probabilistic risk analysis capturing both variability and uncertainty associated with the
risk estimate. GE also agrees with the Agency's assessment that the primary route of
exposure to PCBs is through fish consumption and that other possible exposure routes have
little bearing on potential risks.
Despite GE's agreement with much of the general approach described in the SOW,
GE has a number of concerns about the Agency's proposal. Many portions of the SOW are
vague, internally inconsistent and simply fail to provide enough information to permit one
to ascertain the Agency's proposed approach. From the information available, it appears
that the Agency intends, in many instances, to use unrealistically high exposure
assumptions. Further, the Agency apparently intends to ignore the vast body of
epidemiological data that do not show that PCBs cause cancer in humans, instead relying
entirely on the uncertain extrapolations derived from laboratory studies of animals. GE's
comments point out these and other problems, along with recommendations on how the
Agency should conduct the HHRA, convey the risk estimates to the public, and use the
results in its remedial decision-making.
1
-------
Beyond the specific concerns summarized below, there are some general points
worth emphasizing. First, because the Site does not extend below the Federal Dam at Troy
and in light of the numerous sources of PCBs to the lower River and lower River fish, EPA
cannot reasonably rely on the results of the Mid Hudson risk assessment to justify remedial
actions in the Upper Hudson. Second, the Agency must emphasize that its risk estimates
are hypothetical. The existing catch-and-release requirements in the Upper Hudson, and
the consumption advisories in the lower River all significantly limit fish consumption now ^
and into the foreseeable future. If the Agency does not take these facts into account in the
baseline HHRA, it must recognize and clearly state that its estimates of risk are completely
hypothetical.
1. Exposure issues
•	The SOW's method for selecting point estimates for incorporation into the exposure
models is vaguely described and implies that the Agency intends to use unrealistic, 7
worst case exposure assumptions. EPA must use more realistic, site-specific and
relevant assumptions.
•	The SOW's explanation of the Monte Carlo modeling does not clearly explain whether
and how the Agency intends to model uncertainty and variability separately. Doing so
8
is necessary to understand the estimated range of exposures as well as the uncertainty
in those estimates. EPA should adopt and incorporate the approach set out in GE's
previous submissions to the Agency.
•	At a minimum, the baseline fish consumption rates must account for consumption
limitations posed by factors other than PCBs, including the statewide consumption
advisories and the conservation-based fishing restrictions imposed by New York State 9
and the Atlantic Marine States Fisheries Commission. Furthermore, given the lack of
good consumption rate data from the Hudson or other comparable New York waters,
the best source for recreational angler consumption rates is the assessment of such rates
in Maine contained in Ebert, et al. (1993).
2
-------
•	EPA should not consider hypothetical subpopulations of "subsistence" anglers. The
available information demonstrates that income level, ethnic background, and 10
commercial and recreational angler status are not relevant for deriving a subpopulation
of highly-exposed Hudson River anglers.
•	The SOW's proposed approach for assessing species-specific consumption should be ^
replaced with the approach set out in ChemRisk (1995), which provides the appropriate
input parameters for the microexposure Monte Carlo analysis.
•	EPA should use a mechanistic, time-variable bioaccumulation model to compute
average future PCB concentrations in fish. Variability in fish PCB levels should be 12
estimated directly from the NYSDEC database. In addition, for the Mid Hudson, EPA
must account for sources of PCBs other than the Upper Hudson.
•	GE agrees with the proposal to account for cooking loss of PCBs, which, contrary to 13
the implication of the SOW, is well-supported in the peer-reviewed literature.
2. Toxicity issues
•	GE supports the proposal to use Aroclor-based toxicity criteria in light of the more
reliable and complete toxicological, epidemiological and analytical databases for 14
Aroclors. Other methods, including congener-based analysis, have critical scientific
rely on inconsistencies and inappropriate assumptions.
•	The Agency's reassessment of the cancer slope response of PCBs, which is based
entirely on rat feeding studies is an important advance. In light of the difficulties and ^
uncertainties associated with extrapolation from rats to humans, GE believes that it is
feasible and more appropriate to use a cancer slope factor which takes into account the
vast epidemiological database which does not support the proposition that PCBs cause
3
-------
cancer in humans. GE 's comments present a method for deriving a human-based
cancer slope factor that should be used in the HHRA.
The IRIS-derived reference dose ("RfD") cited in the SOW is also flawed and overly
conservative. It is based on a single study of rhesus monkeys and incorporates
numerous uncertainty factors when extrapolating to humans. GE's comments present a
more defensible and realistic RfD that EPA should use in the HHRA.
EPA should not proceed with the SOW's proposal to present a qualitative assessment
of endocrine disruption in light of the lack of evidence that PCBs have any effects on
the human endocrine system.
EPA should incorporate an analysis of the uncertainty associated with the PCB
toxicological criteria incorporated into the risk assessment, just as it intends to do with
the exposure assumptions.
Risk characterization issues
Because one goal of the risk management is to determine how exposures for a
particular source relate to background, the HHRA must recognize the background
levels of PCBs in all individuals.
The assessment of risks to the "reasonably maximally exposed individual" and the
average individual should be based only on the findings of the probabilistic analysis,
which is much more powerful than the proposed deterministic analysis. EPA should
abandon the proposed deterministic analysis, which would lead to overly conservative
risk estimates.
The baseline HHRA cannot and should not be used to select a remedial decision. The
results of the baseline HHRA can only be used to assess the hypothetical risks
associated with the "no-action" alternative and, in the context of remedial decision-
-------
making, must be measured against the risks estimated to result after implementation of
different remedial options.
5
-------
SECTION I
INTRODUCTION
General Electric Company ("GE") is pleased to submit these comments on the July
1998 "Hudson River PCBs Reassessment RI/FS Phase 2 Human Health Risk Assessment
Scope of Work" ("SOW"). GE supports many aspects of the SOW. For example, GE
generally supports the use of Monte Carlo modeling to assess exposure in the Human
Health Risk Assessment ("HHRA"). GE also agrees with the Agency's conclusion that the
greatest risk of exposure to PCBs in the Hudson River is likely to result from fish
consumption and that other exposure routes are not significant.
Nevertheless, the SOW is inadequate in several respects. The SOW fails to explain
in sufficient detail the methodology EPA intends to use to complete many of the identified
tasks. The SOW also contains significant gaps and inconsistencies, as well as confusing
statements and terminology. The SOW provides only a vague description of the
relationship between the exposure assessment and the Agency's effort to model future
levels of PCBs in fish and the Agency's intended use of Monte Carlo modeling. The SOW
proposes to rely entirely on toxicity estimates based on animal studies, ignoring the
extensive data from epidemiological studies that do not show that PCBs cause cancer in
humans. GE's comments focus on these problems and provide recommendations on how
EPA should complete the human health risk assessment ("HHRA") for the Site.
-------
SECTION II
GENERAL COMMENTS
Several broad issues raised by the SOW deserve comment.
A. The site does Dot extend below the Troy Dam.
First, as GE has previously raised with EPA,1 the Company disagrees with the
Agency's description of the Site as including all 200 miles of the Hudson River between
Hudson Falls and the Battery. The documents in the administrative record for the addition
of the Site to the CERCLA National Priorities List explicitly limit the reach of the Site to
the area above the Federal Dam at Troy, and EPA's post-rulemaking comments to the
contrary cannot change this fact. GE's disagreement with EPA on the scope of the Site is
particularly important in the context of the HHRA, in light of EPA's proposal to conduct a
separate analysis of human health risks from PCBs in the fresh water portion of the lower
River, a portion of the River that is not properly considered part of the Hudson River PCBs
Superfund Site. From previous correspondence and statements, GE understands that EPA
is limiting its analysis to potential remedial actions in the Upper River. Assessing human
health risks in the Lower River suggests that the Agency may be attempting to justify a
remedial action on the basis of benefits to the Lower River.
Justifying any remedial action in the Upper Hudson River on the basis of benefits
to the Lower River would have serious consequences to the scope of EPA's present
reassessment. In such circumstances, EPA would be obligated to investigate and evaluate
remedial alternatives, such as source control in the lower River; consider the greatly
increased number of sources of PCBs (and other contaminants) to fish in the Lower
Hudson; and identify the much wider group of parties who rightfully should be classified
as PRPs. The presence of other dischargers of PCBs in the Lower River is well known to
' See Nov. 6,1997, lener from Angus Macbeth to Richard Caspe; May 5,1998, letter from Angus Macbeth
to Douglas Fischer.
7
-------
EPA; the Agency has conducted recent studies of PCB discharges into New York Harbor,
including sampling outfalls, and of comparative contributions of PCBs into the Harbor.
The Agency made the importance of other contaminants plain in its 1984 ROD, concluding
"that detectable levels of dioxin, dibenzofurans, mercury and chlordane (from known and
unknown sources) have also been identified in Hudson River fish, and that even if PCBs
decrease to an acceptable level, the fishing bans would continue on the basis of these other
types of contaminants."
EPA cannot have it both ways. The Agency cannot describe the Site as
encompassing the 150 miles from Troy to the Battery and then address only one
contaminant and one or two PRPs outside that 150 miles as the sole subjects for remedial
consideration. The scope of EPA's Superfimd activity at the Site is circumscribed by the
characterization and definition of the site which EPA promulgated in its rule-making many
years ago.
B. The Agency must provide advanced notice of its intent to use additional data.
The Agency points out in the introduction of each section of the SOW (SOW, at 1,6, 7)
that individual components of the proposed approach may be revised if additional data are 26
identified in the course of preparing the risk assessment. GE agrees that appropriate
additional data should be included if they would lead to a more accurate baseline risk
assessment. We note, however, that this statement appears to be inconsistent with the
Agency's claim that it already possesses all the necessary data to complete the
reassessment and, indeed, that no new data will be considered. We trust that the Agency
will not ignore new and relevant data in its remedial decision making. Regardless, the
Agency should notify the public of any new data it intends to use, indicate how it proposes
to incorporate the data in the risk assessment, and, when appropriate, make such data
available for review and comment by the public. By releasing this information prior to its
use, the Agency will comply with its mandate to include the public in its decision-making
process.
8
-------
In that light, the SOW states that the risk assessment for the Mid River will be based on the
ongoing modeling by Drs. Thomann and Farley (SOW, at 17). This work is currently not 27
available for external review and comment. If the Agency intends to use the
Thomann/Farley model, it must provide the public an opportunity to review and comment
on it. The Thomann/Farley model should also be subject to external peer review consistent
with the peer review of EPA's own modeling effort for the Upper Hudson.
C. The Agency must emphasize that the risk estimates presented in the HHRA
are hypothetical.
By EPA policy, baseline risk assessments of Superfund sites generally do not
consider the effects of administrative and other types of existing controls on exposure
(EPA, 1989). To the extent that the assessments provide a starting point for the decision of
28
what remedies (including administrative controls) are necessary at a site, this approach is
understandable. However, baseline risk assessments also communicate to the public an
understanding of the fundamental nature of the risks that currently exist at Superfund sites.
In the case of the Upper Hudson River, the risks that would occur without fishing
restrictions are different from the risks that actually exist today and into the foreseeable
future. NYSDEC has established and enforces a ban on keeping of fish in the Upper
Hudson.2 This ban is well publicized and enforced by a conservation officer who patrols
the Upper Hudson. This officer interviews each angler he meets, including anglers fishing
from shore and from boats. Over a recent three-year period (August 31,1995 through July
31, 1998), the conservation officer on the Hudson River has checked 1,437 anglers and
issued only nine tickets and three warnings for keeping fish (NYSDEC, 1998, attached).
This finding confirms that the current ban on keeping fish is extremely effective in
controlling exposure to PCBs.
2 As discussed in Section 3.3 below, consumption of fish from the entire Hudson River is also subject to a
general restriction in consumption that is independent of the PCBs in the fish. In addition, stiff,
conservation-based restrictions on fishing are in place on the lower River. The advisory and fishing
restrictions affect fish consumption by truncating the distribution of consumption rates. Since the advisory
and restriction are not functions of PCB contamination, they must be taken into account when assessing fish
consumption in the baseline risk assessment.
9
-------
As to the Lower Hudson, NYSDEC and the Atlantic States Marine Fishery
Commission have imposed conservation-based restrictions on keeping fish of many
species. These restrictions are imposed to assist in the maintenance of fish stocks, not for
reducing exposures to PCBs. Nevertheless, they have the effect of reducing such exposure
dramatically. Under the baseline assessment described in the SOW, estimates of fish
consumption will be developed under the assumption that there is no ban or restriction on
keeping and consuming fish. The findings of an assessment that assumes that all anglers
are free to keep and consume as many fish as they desire would not be a fair description of
the actual risks facing anglers using the Hudson River. Therefore, the HHRA should
clearly state that (1) the fishing restrictions effectively eliminate current and future
exposures, thereby eliminating the risks (if any) from the consumption of fish; and (2) the
risk estimates produced in the Phase 2 assessment are hypothetical risks that would occur
in the absence of the current restrictions on keeping and consuming fish.
10
-------
SECTION III
EXPOSURE ASSESSMENT
For the exposure assessment, the SOW proposes to develop an exposure scenario
using a variety of inputs based both on site-specific data and default assumptions. These
will then be used to develop both deterministic and probabilistic estimates of exposure.
GE concurs with many aspects of this approach. For instance, the SOW's proposal for the
following aspects of the exposure scenerio are generally sound:
•	The definition of the exposed populat on as those anglers who begin fishing
at a specific "start date" ;
•	The assumption that an angler fishes from multiple locations along the
Upper Hudson River (or Mid Hudson);
•	The recognition that PCB levels in fish should vary as a function of date,
location and species of fish;
•	The assumption that an angler catches and consumes a number of different
species of fish;
•	The consideration of cooking loss;
•	The determination of duration of exposure based on site-specific data; and
•	Basing fish consumption rates on site-specific information.
The SOW's failure to include details on how the Agency intends to implement
these principles is troubling and suggests that the Agency may intend to default to "worst
case" exposure assumptions (for example, SOW, at 13).
The SOW's failure to explain the proposed approach for the HHRA adequately also
affects other important aspects of the exposure assessment. In particular, the SOW's
description of the proposed Monte Carlo modeling is muddled and confusing. We point
out these problems below and make recommendations about how the probabilistic
modeling should be implemented. We also present recommendations concerning other
11
-------
aspects of the proposed exposure assessment, including estimation of fish consumption
rates, future PCB concentrations in fish, duration of exposure and cooking loss.
A. The proposed approach for determining fish tissue concentrations for the
high-end angler is flawed.
The SOW's proposal for selecting fish tissue concentrations for the high-end angler 32
is invalid. As indicated in EPA's Phase 1 Report, fish tissue concentrations increase with
the trophic level of the fish. As a result, the fish with the highest levels of PCBs tend to be
predator fish, such as the northern pike. These fish species make up a relatively small
fraction of the edible fish in the Upper Hudson. It is highly in.plausible that an angler
could catch a sufficient number of these fish to support high levels of fish consumption. It
is far more likely that the anglers with high fish consumption rates will consume the more
readily available species and will consume multiple species. Thus, the high level of intake
will be associated with the average fish concentrations of the more available species.
The HHRA should follow the Guidelines for Exposure Assessment's approach for
selecting inputs for the determination of the high-end exposed individual (EPA, 1992).
This guidance recommends that values be selected for one or two inputs based on the 95th
percentile and that the remaining values be assigned values for typical individuals. Use of
extended worst case or even reasonable worst case values for all inputs of a risk assessment
results in implausible results (McKone and Bogen, 1991).
Therefore, GE recommends that the inputs to the high-end angler be revised as
follows:
•	The values for fish intake and duration should be set at the 95th percentile values.
•	The median value of body weight (age adjusted) should be used.
•	Cooking loss should be based on the most likely estimate of reduction.
12
-------
* The value for fish concentration should be based on a weighted-average of the best
estimates (not UCL) of the mean concentrations of PCBs in the most available
species.
B. Monte Carlo Modeling
The SOW provides limited information on EPA's proposed use of Monte Carlo
modeling of exposure, GE supports the Agency's proposal to use Monte Carlo modeling,
but the SOW's failure to present a thorough and transparent description of the modeling
that the Agency intends to use inhibits a detailed critique. Consequently, we present GE's
recommendations on how the Agency should proceed.
B.l The SOW does not provide an adequate description of the model.
The portion of the SOW that discusses the modeling is limited to a few scattered
paragraphs and sentences. These fragments contain inconsistent and confusing
terminology. Moreover, the limited information provided in the SOW concerning the
proposed Monte Carlo modeling suggests that a number of fundamental decisions on the
model structure and the inputs to the model have been made, but the details of these
decisions have been not been disclosed.
The lack of coherent description is surprising in light of the extensive discussions
between EPA and GE concerning this topic and the information that GE has previously
provided to the Agency. GE has developed an advanced Monte Carlo model of exposure
to PCBs from tiie consumption of fish in the Upper Hudson River (CbemRisk, 1995e). In
October 1995, GE met with EPA and provided a conceptual description of the model, a
printout of the computer code, copies of ancillary materials, and electronic copies of the
working model. This material outlined the conceptual issues and provided technical
material to assist in the evaluation of essential scientific questions in modeling. A peer-
reviewed article based on this work has been published on the topic of modeling of
exposure to PCBs from the consumption of fish (Keenan et al., 1996). GE urges EPA to
13
-------
consider, and where appropriate, incorporate these materials into its Monte Carlo modeling
effort.
B.2 The SOW does not explain whether and how the Agency intends to mode!
variation and uncertainty.
Monte Carlo modeling is fundamentally the same for any type of equation, 34
involving the repeated use of an equation to produce a range of answers. However, the
process of Monte Carlo modeling becomes more complex when models separate
uncertainty from variability (Frey, 1993; Hoffman and Hammonds, 1994), and when
models consider time-varying exposures such as microexposure event models (Price et al.,
1996).
Several statements in the SOW suggest that such complex approaches will be used
in the Hudson River assessment. However, the discussion is so scattered and incomplete
that the approach the Agency intends to follow is not clear. For example, the SOW states
that the exposure assessment portion of the risk assessment will consist of two parts: the
first, a standard exposure assessment, and the second, a Monte Carlo analysis (SOW, at 6).
The discussion continues that the probabilistic (Monte Carlo) analysis will attempt to
capture information on uncertainty and variation. Yet, the discussion only includes a brief
description of the steps needed to develop a Monte Carlo model of exposure variation
across individuals; it contains no discussion of uncertainty. From this discussion, it is
unclear whether EPA intends to assess uncertainty in the Monte Carlo modeling.
Similarly, there are various references in the SOW that allude to "two tier" (SOW,
at 10,11,12) and "two-stage" (SOW, at 13) Monte Carlo analysis. It is not clear whether
the Agency intends these statements to refer to separate analyses of uncertainty and
variability or some other sort of analysis. In contrast, in the discussion of risk
characterization (SOW, at 15), the SOW states that "[a]n enhanced Monte Carlo analysis
will be performed to evaluate variability and uncertainty in exposure parameters, using two
phases to distinguish the impacts of variability and uncertainty, where appropriate."
14
-------
Although this implies separate analyses of uncertainty and variability, the description of
the Monte Carlo modeling is insufficient to allow an understanding of how many of the
elements will be performed.
GE urges EPA to analyze variability and uncertainty separately. EPA guidance has
emphasized the value of separating variability and uncertainty in probabilistic analysis
(EPA, 1997a,b). The Agency must recognize, however that this level of analysis poses
significant challenges for the Agency. First, such analyses require the use of nested loops
in programming (Hoffman and Hammonds, 1994). Since the current modeling
recommended by GE requires the use of nested loops (one for each year, and inside of each
year a loop for each fish consumed) (ChemRisk, 1995e), the addition of another layer of
nested loops poses a significant computational challenge.
Second, all inputs with variability must jointly characterize uncertainty and
variability. This may be relatively minor for inputs such as body weights but will pose a
significant challenge for factors such as fish consumption, duration, and fish concentration.
Third, the Agency should not arbitrarily reject the consideration of certain sources of
uncertainty. If the Agency includes quantitative analysis of uncertainty, then it should
strive to include all sources of uncertainty. These sources include:
•	uncertainty in the dose response measurements;
•	uncertainty in intake rates that results from the use of data on the consumption of
fish from multiple bodies of water,
•	uncertainty in the stability of fish consumption rates over time;
•	uncertainty in modeling fish tissue concentrations; and
•	uncertainty in duration of exposure when the effects of cessation are not
considered.
15
-------
BJ Other issues concerning Monte Carlo modeling
There are several other issues concerning Monte Carlo modeling that require
comment.
First, the SOW suggests that an individual's exposure will take into consideration
the time-varying nature of exposure concentrations (SOW, at 13). GE supports the
consideration of time-varying exposures.
Second, the statement in the SOW that the "90th percentile will be always used"
(SOW, at 15) is contradictory to the fundamental nature of Monte Carlo modeling. Monte ^
Carlo modeling requires a random selection of input values from a distribution. Limiting
the selection to a single percentile is clearly wrong.
Third, the SOW correctly states that there are no data on an angler's year-to-year
variation in fish intake rates in the published literature (SOW, at 15). This raises an
important issue on how to account for such potential variation. The SOW correctly states
that intake rates over time will not be perfectly correlated because variations in weather,
productivity of a fishery, vacation choices, and other factors will influence anglers annual 37
fishing rates on a yearly basis. GE does not agree, however, with EPA's rationale for not
allowing the model to vary the ingestion rate from year to year -- namely that such an
approach "would assume that there is no correlation between yearly ingestion rates and
effectively average the high-end consumers out of the analysis" (SOW, at 16). There are a
number of methods by which an angler's year-to-year variation in intake rates can be
modeled without assuming that anglers' annual intakes are not correlated. For example,
intake can be correlated by allowing the intake rates to vary within a fixed range of
percentiles (Price et al., 1997). The uncertainty in annual intake variation is a legitimate
issue that should be addressed in modeling long-term dose rates. The Agency's adoption
of a modeling approach where the intake rate is fixed as a worst case assumption
unnecessarily adds conservatism and fails to use the strengths of Monte Carlo techniques.
16
-------
Finally, the SOW states that the modeling of fish tissue concentrations may or may
not include information on the variation in PCBs in fish of the same species, date, and
location (SOW, at 13). This inter-fish variation can be shown to average out in anglers
with high levels of fish intake. That is, this source of variation averages out and does not
contribute to the variation in doses in anglers consuming large amounts of fish. As a
result, all that may be necessary for modeling of fish tissue concentration is an estimate
(with uncertainty) of the mean. This factor should be taken into consideration in the
development of data on fish tissue levels.
B.4 Recommended approach
As noted above, GE provided EPA with detailed information on a modeling
approach for characterizing exposures to anglers (ChemRisk, 1995 a-e) (attached). Based
on discussions with the Agency, it was apparent at that time that EPA contractors generally
agreed with the approach. GE still believes that this modeling approach is fundamentally
conect and should be applied to the Hudson River. This modeling approach includes:
•	modeling each fish consumed by each angler;
•	defining the angler population of interest as those anglers who would begin using
the river at a certain "start date";
•	accounting for the temporal changes in the concentration of PCBs in fish;
•	modeling exposure duration as a function of the demographics of the angler
population;
•	accounting for cooking loss;
•	modeling the species preference;
•	developing estimates of the average daily dose using a PCB specific averaging
time; and
•	accounting for temporal changes in angler body weight and behavior.
In addition, GE proposes the following additional comments on modeling angler
exposures. First, the modeling approach should separate uncertainty from variability. This
17
-------
should be done using a nested loop approach (Hoffman and Hammonds, 1994; Price et al.,
1995). In the outer loop, values are selected from distributions that characterize the
uncertainty in the inputs. As discussed above, this uncertainty should include all sources
of uncertainty.
Within this outer loop, the model should consist of a microexposure model of
variation in the anglers, as described by ChemRisk (1995c). EPA should investigate
whether the model can be simplified by replacing distributions of inter-fish variation in
concentrations (species, location and date specific) with the mean concentration.
In the case of the Mid Hudson, EPA should revise the model to discriminate between
the exposures that hypothetically occur as a result of eating fish that have accumulated
PCBs from the Upper Hudson River and those that have accumulated PCBs from other
sources (in the Mid River and elsewhere). It will be critical for the Agency to track these
two doses separately. This is necessary in order to provide risk managers with a clear
understanding of the relationship between PCBs in the river and potential exposures in
order to assess the potential effectiveness of remedial options in the Upper River.
EPA should also extend the models of dose to track how exposures from fish affect
the human body burdens of PCBs. This analysis will require the incorporation of simple
toxicokinetic models of PCB intake and retention into the microexposure Monte Carlo
model (Keenan et al., 1997; Avantagio et al., 1998) and data on the background levels of
PCBs in anglers. The goal of this analysis is to determine whether the consumption of fish
from the Hudson River will change the body burdens of anglers (see Section 5).
-------
C. Fish consumption rates
C.l Development of a fish consumption rate distribution for recreational anglers
The amount of fish that anglers consume is an important parameter in the estimate
of exposure to PCBs from Hudson River sediments. GE agrees with the SOW's
conclusion that:
Fish ingestion rates are waterbody specific and depend on a number of factors
including weather, available fish species, angler (man, woman or child who fishes),
preference for specific species, impact of fishing bans, and distance of the angler
from the water body. (SOW, at 7)
This recognition is a significant improvement over the flawed approach taken by
EPA (1991) in the Phase 1 Report, where a marine fisheries value of 30 g/day, subject to
avidity bias problems (Price et al., 1994), was adopted as an estimate of fish consumption
for Hudson River anglers. As GE has explained, the amount of fish consumed by a
population of anglers depends on the numbers and types of waterbodies fished and the
characteristics of the angler population (ChemRisk, 1995c; 1995e, attached). Fish
consumption also depends on factors such as climate, fish species present, productivity,
access, and the size of the angler population.
Consumption of fish is constrained by New York State fishing restrictions.
The most important factors affecting consumption of fish from the Hudson River
are New York State's strict and effective restrictions on fish consumption in the Hudson 42
River, which result in a significantly lower consumption rate than might be assumed under
default exposure scenarios. It can be argued that the baseline HHRA should consider these
restrictions in order to provide a site-specific and realistic estimate of exposure. The
following paragraphs elaborate on this topic in light of EPA science policy and guidance.
19
-------
Even apart from the Hudson-specific restrictions and advisories, New York State
has issued a statewide general health advisory for eating sportfish from any of New York's
freshwaters, "to protect against eating large amounts of fish that haven't been tested or
[that] may contain unidentified contaminants" (NYSDOH, 1997). This advisory urges
individuals to consume no more than one meal per week (32 g/day) of sportfish taken from
any of the state's freshwaters (NYSDOH, 1997). This statewide general health advisory is
different from the consumption restrictions placed on the Upper Hudson River. In
addition, NYSDEC and the Atlantic States Marine Fisheries Commission impose strict,
conseivation-based fishing restrictions for various species, including striped bass. Even if
EPA disregards the effect of the Upper River consumption ban in developing its rate of fish
consumption for use in the baseline HHRA, EPA must address the statewide consumption
advisory and the conservation-based fishing restrictions in this context. EPA must do so in
light of the fact that these advisories and restrictions were issued independent of the
presence of PCBs in the Hudson River. Specifically, any distribution of fish consumption
rates selected for use in the baseline Hudson River HHRA must be truncated at 32 g/day to
reflect the maximum consumption rate allowed by New York's advisory.
It is appropriate and consistent with EPA guidance for the exposure assumptions in
the baseline risk assessment to reflect real-world, current conditions - especially those like
the statewide general health advisory that are unrelated to the risk management of the
Hudson River PCB Superfund Site. As a general rule, EPA's Risk Assessment Guidance
for Superfimd ("RAGS") favors the use of site-specific information instead of generic,
standardized assumptions. See Risk Assessment Guidance for Superfund, Human Health
Evaluation Manual Part A (Interim Final) (July 1989). For example, RAGS recommends
examining actual and potential "land use" when characterizing the potentially exposed
populations, id. at 6-7, and evaluating a number of site-specific factors that affect the
ingestion of chemicals through consumption of fish. Id. at 6-43. Other Superfund
guidance provides that, when available, site-specific exposure information should be used
in the baseline risk assessment in lieu of standardized, default exposure assumptions.
OSWER Directive 9285.6-03, RAGS Volume I: HHEM Supplemental Guidance "Standard
Default Exposure Factors" (Interim Final) (March 25, 1991)(Standard exposure
20
-------
assumptions are to be used only where site-specific values arenot available.). Similarly,
the EPA Science Advisory Board ("SAB") recommends the use of site-specific information
in Superfund risk assessments where site-specific conditions may be unique, as on the
Hudson, and limiting the use of default information to circumstances where exposure
parameters are unlikely to vary significantly from site to site. An SAB Report: Superfund
Site Health Risk Assessment Guidelines (February 1993, at 2).
Indeed, the whole tenor of EPA's risk assessment approach, as articulated in the
above references, is to rely on site-specific information to the greatest extent possible.
EPA's Assessing Human Health Risks from Chemically Contaminated Fish and Shellfish:
A Guidance Manual (September 1989) (EPA-503/8-89-002) recommends "that local or
regional assessments of fishery consumption be performed whenever possible to avoid
errors inherent in extrapolating standard values for the U.S. population to distinct
subpopulations." (Id., at 54). EPA's Exposure Assessment Guidelines recommend the use
of distributional methods of exposure analysis, such as Monte Carlo analysis, because they
rely on site-specific data to provide a more precise understanding of the range of actual
exposures to an affected population. (57 Fed. Reg. 22889, 22922, May 29, 1992). The
SAB similarly recommends that distributional exposure approaches be used in Superfund
risk assessments in order to reflect the actual behavior and exposures of those who visit or
live near a site because it is "more consistent with the exposure assessment guidelines, and
...in the spirit of the Exposure Factors Guidelines." (Id at 17). In fact, the Commission on
Risk Assessment and Risk Management established under the Clean Air Act Amendments
of 1990 advocates the use of distributional exposure assessment methods because they
incorporate population-specific exposure (receptor-based) information, and not merely
assumptions about exposure derived from sources and models. "Risk Assessment and Risk
Management in Regulatory Decisionmaking: Volume 2H (1997, at 774-75). The
Commission noted that source-based exposure information can be "seriously misleading"
as compared to personal measurement results, where individuals in a particular population
are monitored for exposure. (Id., at 191).
21
-------
The distinction between source-based exposure and receptor-based exposure is an
important one for EPA to consider, where the pathway from a contamination source (i.e.,
fish) to a potentially affected population is markedly reduced from that of the generic
consumption assumptions on account of New York's statewide health advisory and the
conservation-based fishing restriction. Unless the advisory's maximum daily rate of fish
consumption (32 g/day) is used to truncate the regional fish consumption distribution
selected for use in the Monte Carlo analysis, the risk assessment will fail to reflect the
actual receptor-based exposure levels of the potentially affected fishing population.
Using standard fish consumption assumptions would also be contrary to Agency
policy. EPA's 1992 Exposure Assessment Guidelines state that, for fish tissue, the
following site-specific data are required to characterize exposure: relationship of samples
to food supply for individuals or population of interest, consumption habits, and
preparation habits (57 Fed. Reg., at 22910). This information is unknown for the Hudson
River, largely because of the consumption ban (and subsequent enforcement) on the Upper
Hudson, the consumption advisory on the Lower Hudson, the conservation-based fishing
restrictions on the lower Hudson, and the statewide general health advisory pertinent to all
freshwaters of the state. As it is reasonable to assume that the latter two factors will remain
in place for the foreseeable future, regardless of what remedial decision is made for the
Hudson River, their effects must be incorporated into exposure information used in the
baseline HHRA.
The data from Ebert et al. (1993) should be used to calculate hypothetical fish consumption
rates for the Upper Hudson.
If, in spite of the Upper River fish consumption ban and the Lower River
advisories, the Agency nevertheless assumes that fish consumption is occurring, then
significant changes need to be made to the SOWs proposed approach for estimating
consumption rates for recreational anglers. Most importantly, EPA should use the data
from Ebert et al. (1993) to calculate hypothetical consumption rates for the Hudson River.
Furthermore, the consumption rate distribution derived from Ebert et al. should be
22
-------
truncated at a maximum value of 32 g/day to reflect the New York statewide health
advisory level for freshwater fish consumption (NYSDOH, 1997).
GE disagrees with the SOW's assessment of the surveys that might be used for
estimating fish consumption in the Hudson River in the absence of the current restrictions.
Of the three surveys of angler behavior on the Hudson River, two are mail surveys of New
York anglers in general (NYSDEC, 1990; Connelly et al., 1992), and the Clearwater creel
survey (Barclay, 1993) was performed on Hudson River anglers. None of these surveys
focused exclusively on fish consumption from the Hudson River. NYSDEC (1990)
evaluated fish consumption from all recreational and commercial sources, including self-
caught fish from the Hudson. Connelly et al. (1992) evaluated self-caught fish
consumption but did not estimate consumption from individual waterbodies. Barclay
(1993) collected data on the frequency of self-caught fishmeals but did not calculate a fish
consumption rate. In addition, this survey does not contain sufficient information to allow
the calculation of a meaningful fish consumption rate for the Upper Hudson River.3
Because the available surveys are flawed and cannot be used to assess hypothetical
consumption rates in the absence of restrictions, the SOW should base the Hudson River
estimates on data from similar bodies of water or from regional data. The selection of a
surrogate study depends on the characteristics of the population under consideration and
the type of waterbody being evaluated. Specifically, it is critical that the study be focused
on the consumption rates of self-caught, freshwater fish over long periods of time. These
criteria must be met to ensure that the fish consumption rate closely approximates
hypothetical consumption from the Hudson. It would be preferable to use a study that
evaluated consumption from a single river that was similar to the Hudson. If a specific
waterbody with appropriate characteristics cannot be identified, it would be more
appropriate to use estimates generated for flowing waters only. The selected study should
1 The SOW implies that the Agency intends to create a distribution of fish coosumption for the Hudson
River by merging data from multiple studies. Such a process is complex, with important logical and
statistical issues that must be addressed before proceeding. It is not clear from the SOW that EPA intends to
do so. Moreover, the rationale for using data from these angler surveys appears to be founded on the idea
that if their results are consistent, EPA can have greater confidence in selecting the Connelly et al. data as its
a priori favorite (SOW, 8).
23
-------
have collected data from regionally appropriate waterbodies. In addition, there should be a
metric that demonstrates the appropriateness of selection and the uncertainties associated
with it.
There are a limited number of studies available in the New York/New England area
that provide information on consumption of sport-caught fish from freshwater rivers and
streams. The Ebert et al. (1993) and Connelly et al. (1992) studies most closely
approximate hypothetical consumption from the Hudson River.4 Both of these studies
evaluated consumption of self-caught freshwater fish by recreational anglers using a mail
recall survey. Given these similarities, it is not surprising that both studies reported very
similar fish consumption rates. The results of Connelly et al. (1992) indicated that the
average New York angler consumes 11 meals per year of self-caught fish from New
York's freshwater fisheries. If it is assumed that each meal is 227 grams in size (1/2
pound) (West et al., 1989; NYSDEC, 1990), it can be estimated that the average New York
angler consumes self-caught freshwater fish at a rate of 7 g/day. This estimate is very
similar to the mean rate of freshwater fish consumption by Maine anglers of 6.4 g/day from
all waters reported by Ebert et al. (1993).
Although the Connelly et al. (1992) study is specific to New York State, there are
several factors that would require the analyst to make additional assumptions to use the
data as the basis for the Hudson River assessment. First, Connelly et al. (1992) only
presents a single point estimate value for fish consumption. The use of a distribution of
consumption rates is necessary in order to characterize interindividual variability and
realistically assess the potential risks to recreational anglers. With only an average
consumption rate value, it is not possible to represent the range of recreational anglers
accurately, including those anglers who ingest higher amounts of fish. While it may be
possible to develop a distribution of consumption rates by going back to the original raw
data, additional analysis will be required to complete this task.
4 Gosnelly et al. (1996) surveyed Lake Ontario anglers to evaluate the effect of Lake Ontario health advisory
recommendations and therefore this study may not be directly relevant to fish consumption in the Upper
Hudson.
24
-------
Second, the mean fish consumption rate determined by Connelly et al. (1992)
represents fish eaten from all freshwaters in the State (i.e., lakes, ponds, rivers, and
streams). As pointed out in Ebert et al. (1993), intake from rivers and streams is only a
fraction of the intake from all freshwaters. In addition, the rate of intake from multiple
waterbodies is higher than that from a single water system (Ebert et al., 1994). Given these
factors, it is highly likely that the fish consumption rate in Connelly et al. (1992)
overestimates the hypothetical fish consumption rate on a single portion of the Upper
Hudson River.
Finally, the purpose of the Connelly et al. (1992) study was not to identify a
consumption rate for New York anglers. Although questions were asked in the survey
regarding fish consumption behaviors, those questions were aimed at estimating how the
effect of health advisories altered the consumption behavior of recreational anglers.
While the data from Ebert et al. (1993) are not specific to New York State, these
data are readily useable and may provide a more appropriate surrogate for Hudson River
anglers than the Connelly et al. (1992) data. Angler demographics and fishing
opportunities are similar in Maine and New York, and the mean fish consumption rates are
similar for both studies (NYSDEC, 1990; Connelly et al., 1992; Ebert et al., 1993). In
addition, Ebert et al. (1993) provides a complete distribution of fish intake rates for
flowing waters, i.e., streams and rivers. The Ebert et al. (1993) survey also addresses each
of the criteria identified in the SOW (SOW, at 9) to evaluate angler surveys.5 Thus, the
best region-specific data on fish consumption rates are available from Ebert et al. (1993),
and GE urges EPA to use these data in the Hudson River risk assessment.
The selection of the most appropriate fish consumption rite is discussed more fully in the paper entitled
Estimating Fish Consumption Rates for the Upper Hudson River (ChemRisk, 1995c) and in the peer-
reviewed journal articles, The Effect of Sampling Bias on Estimates of Angler Consumption Rates in Creel
Surveys (Price et al., 1994), Selection of Fish Consumption Estimates for Use in the Regulatory Process
(Ebert et al., 1994), and Estimating Consumption of Freshwater Fish among Maine Anglers (Ebert el al.,
1993).
25
-------
The SOWs definition of study population is arbitrary and must be changed.
GE disagrees with the SOWs arbitrary definition of the study population. The
SOW truncates the full distribution of hypothetically exposed anglers by defining the lower 43
end of the study population as those anglers "who would consume self-caught fish from the
Hudson River at least once per year in the absence of a fishing ban." Many anglers,
however, will consume fish at frequencies much less than once per year (ChemRisk,
1991a; Hbert et al., 1993). The SOWs omission of such frequencies thus biases upwards
the resulting risk estimates. A better approach would be to define the population as those
anglers who consume fish once in the "start year" (the first year of exposure) but not
require them to consume fish in every subsequent year. This would allow the model to
include anglers who consume fish at lower frequencies than once per year.
For the Mid Hudson, the HHRA should also evaluate its selection of angler
consumption studies in the context of potentially short-lived fishing activities, such as fish
tournaments or short fish runs (e.g., shad run in the Mid and Lower Hudson River), to
determine the potentially exposed populations. Because of these events, it is possible that
the estimates of fish consumption for the Upper Hudson cannot be applied categorically to
the Mid Hudson. Therefore, surrogate surveys of fish consumption that consider such
events are needed to accurately select a distribution of fish consumption rates for this
section of the River.
G2 Angler Subpopulations
The SOW states that no attempt will be made to distinguish subpopulations of "highly 4*
exposed or lesser exposed anglers." GE agrees with this approach, which is supported by the
available data suggesting that recreational anglers are the appropriate population for the
HHRA.
26
-------
Historically, concerns have been raised over hypothetical subpopulations of anglers who
consume greater amounts of self-caught fish than the general recreational angler population,
due to their reliance on fishing as a major or sole source of dietary protein for their families
(USEPA, 1998); Abraham et al., 1995; Becher et al., 1995; McCormack and Cleverly 1990;
West et al., 1991). The term subsistence anglers has been applied to this population.
The characterization of this population has been extremely ambiguous. In North
America, Native American populations that have subsistence and treaty rights to certain
fisheries (CRITFC, 1994) and Arctic Inuits who, because of tradition and their remote
location, rely heavily on native foods obtained from the sea (Kinloch et al., 1992; Coad,
1994) appear to be high consumers. Beyond these fairly well defined populations, clear
examples of subsistence anglers are difficult to define.
In order for an individual to consume at high rates, that person must have access to large
amounts of the fish and must have either a need or preference to consume locally caught fish
in large quantities. There are several factors that could define such a population including:
•	low income individuals who must rely on fish for their dietary needs,
•	native peoples who have cultural traditions of consuming large quantities of fish,
•	commercial anglers who have ready access to large amounts of fish, and
•	recreational anglers who have a strong preference for fish
The fish consumption habits of these subpopulations, compared with the distributions of
consumption rates for the general recreational angler population, are discussed below.
Income level
Low income, in and of itself, does not lead to high levels of fish consumption. The fish
consumption survey literature indicates that there are no significant differences in fish
consumption rates among different income groups (Ebert et al., 1993; Connelly et al. 1990;
West et al., 1991). Wendt (1986) studied the fish consumption habits of low income families
27
-------
living in New York State to determine how much freshwater fish they consumed from New
York State waters. Based on the reported range of meals and an assumed meal size of 1/2 lb.
(227 g), it can be estimated that these individuals consumed at a mean rate of 11 g/day and a
maximum rate of 60 g/day. This mean is consistent with the means reported in more recent
surveys of New York's recreational anglers (Connelly et al., 1990, 1996) and other
recreational anglers in the Northeastern U.S. (Ebert et al., 1993; 1996), while the maximum
rate is lower. Thus, low-income populations living in New York State do not have higher
rates of fish consumption than recreational anglers in the region.
Ethnic background
There are data indicating that certain localized North American ethnic subpopulations
may have higher rates of consumption than the general angler population. Studies of native
peoples in the Pacific Northwest of the U.S. and Canada indicate that they rely more heavily
on fish as a staple of their diets than does the general population. However, these findings
are of little relevance to the Upper and Mid Hudson Rivers since the counties bordering these
portions of the River do not have such local ethnic populations.
In addition, when individuals from these same ethnic populations reside in a more
heterogeneous and economically developed area, these differences diminish (Wolfe and
Walker 1987). While mean consumption rates reported for native peoples living in closer
proximity to economically developed areas were higher than the mean values reported for the
general recreational populations (NYSDOH, 1993; West et al. 1991; SelikofT et al. 1982;
Hutchison and Kraft, 1994; Peterson et al., 1994), their maximum rates were similar. Other
comparisons of fish consumption by ethnic background have reported no significant
differences among consumption rates for those groups (Ebert et al., 1993; Landolt et al.,
1985; Anderson and Rice, 1993).
28
-------
Commercial Anglers
Since at least 1976, commercial angling has not been practiced on the Upper River and
has been severely restricted for many species on the lower River. Nevertheless, this group
could, in theory, be a population of concern for the Mid River. Individuals who have
commercial fishing licenses have unlimited access to their marketable catch and might be
assumed to consume more fish than the recreational angler population. Limited data on the
fish consumption activities of freshwater commercial anglers show that such anglers do not
eat substantial amounts of the fish that they harvest, due to the fact that the sale of those fish
is critical to their household income and their ability to pay for other foods and living
expenses. For example, Hubert et al., (1975) studied commercial freshwater fishing activities
in Upper East Tennessee during 1973 and reported that, of a total of 94,079 kg of fish
commercially harvested by 29 anglers, 2,665 kg were retained for personal use. If this
amount of fish is divided among the 29 anglers and their families and assumed to have
edible portions of 30 percent, the resulting mean consumption rate is 25 g/day. This mean
rate is very similar to mean rates reported for recreational anglers fishing large bodies of
water (SCCWRP and MBC, 1994; Ebert et al., 1994). Thus, commercial freshwater anglers
do not consume substantially more fish than recreational anglers fishing the same types of
waterbodies.
Recreational Anglers
Fish consumption rates among recreational anglers are highly variable. Based on
available survey data and on a critical review of the relevant literature, high-level fish
consumers in North America are a diverse group that cannot easily be defined or identified by
socioeconomic characteristics. With the exception of certain native peoples who have
continued to promote their cultural dietary traditions, there are no social or economic
characteristics that are associated with the presence of a high fish consuming population.
Consequently, EPA has made the correct decision in not developing a separate exposure
assessment for this population.
29
-------
C3 Species-specific fish ingestion rates
In general, GE agrees with the Agency's attempt to address species-specific fish
ingestion rates in the exposure assessment. Anglers typically prefer to catch certain
desirable species and to reject others. Moreover, many anglers engage in short-lived
fishing activities, as mentioned previously. Since PCB levels in fish vary by species, it is
important to capture this angler preference in the estimates of exposure to PCBs. The
NYSDEC study and the work by Connelly et al. show that New York anglers preferentially
select for certain species in both fishing effort and consumption (NYSDEC, 1990;
Connelly et al., 1992). In many cases, the species selected were those that accumulate
lower levels of PCBs, often because these most desirable species have relatively low lipid
contents as compared to other species present in the Upper Hudson. Since the species of
fish sampled by EPA or NYSDEC for PCB tissue analysis are not necessarily consumed by
recreational anglers in amounts proportional to their sampling frequencies, the risk
assessment for the Upper Hudson should consider both interspecies differences in PCB
concentration and angler preferences.
Information on species preference specific to the Upper Hudson River is
unavailable. However, data on angler preference in freshwater rivers in New York similar
to the Upper Hudson River are available from Connelly et al. (1992).6 Based on these data,
it is possible to identify species preferences among New York anglers that can be used as a
surrogate for Hudson River anglers. Connelly et al. (1992) collected information on
fishing behaviors (e.g., species caught, waterways fished) and fish consuming behaviors
(e.g., species eaten, preparation techniques used) of licensed anglers. In order to use these
data for the Upper Hudson, it is necessary to identify rivers and streams with
characteristics and species similar to the Upper Hudson. Such an analysis results in a list
of fish species likely caught in the Upper Hudson and the probability of how often these
species are eaten. By taking this approach, a probability distribution that accurately
reflects species consumption preferences of Hudson River anglers can be developed. This
* EPA should not include the Connelly et al. (1996) itudy as it was a survey of Lake Ontario fishers.
30
-------
issue is addressed in ChemRisk (1995a), which recommends the appropriate input
parameters for the microexposure Monte Carlo analysis.
GE is concerned about the SOW's proposed approach for considering species-
specific consumption (SOW, at 9-10). The SOW states that species-specific fish ingestion
rates will be developed from data collected from multiple studies of anglers. The result of 45
this analysis appears to be some sort of species weighting factors that will be applied to all
anglers. This use of a single set of factors implicitly assumes that all anglers will consume
the same species and in the same proportions. As the SOW acknowledges (SOW, at 10),
this assumption is implausible. Anglers can be expected to vary the species they consume
based on their choice of fishing location, tackle, and means of fishing. The SOW proposes
to address this uncertainty by running a separate analysis in which anglers will be assumed
to consume the species with the highest level of PCBs (SOW, at 13). The SOW does not
indicate how the results of the two estimates will be used in assessing the baseline analysis.
In any event, this approach for assessing the impact of species choice is invalid because it
provides the risk manager with results from two implausible sets of assumptions.
GE is also concerned with the SOW's proposal to combine sunrogate fish species
preferences for the Upper and Mid Hudson. This approach is unjustified and scientifically ^
invalid as these anglers would be expected to have very different preferences based on the
type of fish present in the respective stretches of the river.
D. Determination of future PCB concentrations in fish
D.l Use of mode! results
The SOW states that projected PCB concentrations in fish will be determined using
47
the EPA bioaccumulation models (SOW, at 12). There are two components to these
predictions: the average concentration and the distribution of concentrations.
31
-------
Average concentration
The Bivariate Statistical Model is the primary model developed by EPA to estimate
mean fish total PCB and Aroclor levels. It is subject to several limitations, as described in
GE's comments concerning the Preliminary Model Calibration Report (GE, 1996). One
particular source of concern is that the BSM overestimates the observed values at low
concentrations (see the enclosed reprint of EPA Figure 9-12). PCB levels in fish are
declining, as shown by the trends since 1993 in the upper river. Therefore, it is important to
be sure that model predictions do not overestimate the true future levels. As discussed in
GE (1996), an alternative modeling methodology must be developed; a time-variable
mechanistic bioaccumulation simulation is the preferred alternative.
In addition, the BSM has practically no predictive power within each river reach.
For example, in EPA Figure 9-12, there is no relationship between predicted and observed 48
largemouth bass total PCB concentrations in Thompson Island Pool (symbol "D" on the
Figure). Because of these problems, the PCB levels predicted by the BSM may not reflect
the decline in concentrations in largemouth bass in Thompson Island Pool. This will
produce an unrealistically high computed human health risk.
The probabilistic food chain model (PFCM) is based upon linear steady state
relationships between sediment, water and fish, as is the BSM. Therefore, the PFCM
should be subject to the same bias and lack of predictive power as the BSM in predicting
average PCB levels.
Variability
The SOW states that the high-end exposure point concentration for the HHRA will
be determined using the 95% Upper Confidence Limit on the mean PCB concentration.
The confidence limits of the mean are dependent upon the variance. Therefore, the
distribution of PCB concentrations within each fish subpopulation must be estimated. One
goal of EPA's modeling efforts was to compute the population variability using the PFCM.
32
-------
As described in GE, 1996, the PFCM:
•	is improperly constructed, calculating variability from uncertainty. This causes
model results to have no physical meaning;
•	requires the answer to solve the problem;
•	incorrectly assumes sediment PCB levels do not change over time; and
•	does not take advantage of the substantial information and data that are
available concerning the mechanisms of bioaccumulation.
An alternative already suggested by GE is to use the extensive database collected
by DEC over the course of 20 years to estimate the shape and parameters of the
distributions of PCB levels in fish.
D.2	Selection of Mid Hudson fish species
The SOW states that high-end exposure point concentrations will be estimated using the
most contaminated species and data from the most contaminated stretch of the Mid Hudson
River (SOW, at 20). Unlike the Upper River, the Mid River contains migratory fish
species (striped bass, eels, shad, etc.) that spend significant time away from the Mid River.
During this time, the fish have the potential to accumulate PCBs from other sources than
the Upper River. EPA must take care interpreting results from such fish when making
remedial decisions related to the Upper Hudson.
E.	Duration of exposure
The approach proposed to characterize the duration of exposure (SOW, at 10-11,
18-19) is a significant improvement over the approach used in the Phase 1 assessment,
which relied on default values. As the SOW notes, mobility is a major consideration in
determining the duration of exposure. The Agency's decision to use county mobility as a
33
-------
surrogate for the probability that an angler will cease using the Hudson because of a new
residence location is also appropriate.
However, this approach assumes that a move from one county bordering the
Hudson to another does not end exposure, which is invalid for model runs in which the
location along the river is considered. Moving from Hudson Falls to Albany will affect the
probability of fishing the Thompson Island Pool. Therefore modeling of specific reaches
should consider inter-Hudson River county moves.
GE supports the effort to take cessation of angling into the assessment of duration.
The SOW, however, does not provide any support for the statment that "generally, anglers
are highly dedicated to their sport, and few voluntarily stop fishing." GE has previously
provided to EPA evidence from the Maine angler survey showing that cessation is an
important factor (ChemRisk, 1995b). GE urges EPA to use these data and, as indicated in
the SOW, to perform similar analyses of the data in Connelly et al. (1992).
The SOW fails to discuss the role of lifespan in limiting exposure duration. The
population of concern is that group of anglers who would use the Hudson River at an
appropriate "start date." Such a population at the time of the "start date" would have an
age structure spanning from teenagers to individuals over 65. Therefore, the HHRA must
take lifespan into account when determining the distribution of duration.
The HHRA also should model duration of exposure using the methodology
described in ChemRisk (1995b). Under this approach the duration is not an input to the
model but is directly based on age-specific estimates of cessation, mobility and lifespan.
The advantage of this approach is that the age structure of the population is handled in a
consistent fashion throughout the model.
34
-------
F. Cooking loss
GE supports the use of the peer reviewed literature to model the loss of PCBs
during cooking, as proposed in the SOW (SOW, at 11). Cooking loss of lipophilic
compounds such as PCBs is a real and verified phenomenon that has been repeatedly
demonstrated in more than 20 publications in the peer reviewed literature (e.g., Sherer and
Price, 1993; Wilson et al. 1998) and forms the basis for fish advisories used by the State of
New York (NYSDOH 1996; NYFW, 1995). As discussed in Wilson et al. (1998), the data
are more than adequate to allow the modeling of cooking loss as a function of cooking
method and in turn the type of fish. GE agrees that the estimates of reduction are subject
to uncertainty, and that this uncertainty may warrant consideration in the modeling of
exposures.
G. Inhalation exposures
The SOW proposes that risks from inhalation of ambient air will be computed
using a deterministic assessment approach (SOW, at 13, 21). As an initial matter, the
Agency should recognize that this route of exposure is insignificant. Studies of PCB blood
levels in individuals near other Superfund sites have consistently revealed that such
individuals do not have excessively high blood levels. GE urges the Agency to abandon
this exposure route.
If the Agency still proceeds, it must specify the source(s) of the data it intends to
use, when and with what method(s) the data were collected, the quality of the data, or how
data will be evaluated with respect to calculating the high-end and central tendency point
estimate concentrations. The SOW vaguely describes data collected from sites near the
Hudson river, but provides no details concerning these data. The Agency must specify the
source(s) of air concentration data and, if not cunently publicly available, should make that
data available for public review and comment
-------
SECTION IV
TOXICITY ISSUES
A. GE supports the use of Aroclor-based toxicity criteria.
GE supports EPA's use of Aroclor PCBs in lieu of PCB congeners for the human
health risk assessment. However, some may claim that an alternative approach should be ^
taken, which we do not support based on critical scientific inconsistencies and
inappropriate assumptions. This alternative for evaluating potential risks from exposure to
PCBs in environmental matrices consists of the following steps, air of which serve to
increase the complexity and the uncertainty of the analysis. First, the concentrations of the
11 "dioxin-like" PCB congeners is converted to 2,3,7,8-tetrachlorodibenzo-p-dioxin
toxicity equivalents (TEQs) through the use of one of several TEQ conversion schemes
(Ahlborg et a!., 1994; EPA, 1989; WHO, 1997) The choice of conversion method is left to
professional judgement and can introduce additional uncertainty into the analysis. The
carcinogenic risks are then calculated for the TEQs by combining the TEQ concentrations
for these congeners with a CSF for 2,3,7,8-tetrachlorodibenzo-p-dioxin of 150,000 (mg/kg-
day)"1 or with one of the more recent and more scientifically appropriate values (Keenan et
aL, 1991). For the non-dioxin-like PCBs, this approach uses the total PCB concentration in
conjunction with the CSF for PCBs of 2 (mg/kg-day)*' to yield the "non-dioxin-like" PCB
risk. It then adds these risks together.
To be logically consistent with this approach, the analyst must subtract out the
concentrations of the dioxin-like congeners from the total PCB concentrations before
making the calculations for the other PCBs. If one fails to do so, then the analysis has
additional flaws due to double-counting of the carcinogenic potential of the dioxin-like
congeners by including those congeners both in the risk calculation for the TEQs and in the
risk calculation for the so-called "non-dioxin-like PCBs." Moreover, even if the analyst
subtracts out the concentrations of the dioxin-like congeners in making the risk
calculations for the remaining PCBs, this approach would still double-count the
carcinogenic potential of the dioxin-like congeners, because those congeners are included
36
-------
in the CSF for PCBs. The CSF for PCBs of 2 (mg/kg-day)'1 was based on toxicological
studies of Arocior mixtures that contained dioxin-like congeners. Indeed, EPA has
attributed much of the so-called carcinogenic potency of PCB mixtures to these congeners
(IRIS, 1998). Thus, the CSF of 2 (mg/kg-day)"1 is much too high to represent the
carcinogenic potential of the non-dioxin-like congeners. Accordingly, even if the PCB
concentrations used for the non-dioxin-like PCBs does not include dioxin-like congeners,
the use of a CSF of 2 (mg/kg-day)"1 to calculate the carcinogenic risk of those PCBs
represents a double-counting of risks. In fact, unless there were a CSF for non-dioxin-like
PCBs, there is no defensible way to use both the TCDD CSF and the PCB CSF in the same
assessment.
Furthermore, the toxicological, epidemiological and analytical databases for Arocior
PCBs are more reliable and complete than those for PCB congeners. In summary, the
following represent the advantages associated with the use of Arocior PCBs in lieu of PCB
congeners:
•	The toxicity studies used to derive the Arocior PCB CSFs and RfDs include both
the coplanar and non-coplanar PCB congeners present in the Arocior mixtures.
•	The concentrations of the coplanar PCB congeners - reputedly the more toxic of the
PCB congeners - is known for most of the Arocior PCBs (e.g., Frame et al,, 1996).
•	There is a paucity of toxicity data for the non-coplanar PCB congeners.
•	The comparability of analytical results can be difficult in PCB congener data since
there are inconsistencies in the analytical methods used to quantify coplanar and
non-coplanar PCB congeners.
•	Arocior PCB results are the most appropriate to use if there have not been
significant changes in the PCB peak patterns.
37
-------
B. Cancer dose response
GE supports EPA's recent efforts in reassessing the cancer risk of PCBs (EPA,
1996, "the Reassessment") and continues to have an important interest in working with
EPA on various issues related to PCB toxicology. Although the Reassessment represents a
positive step in evaluating the PCB cancer risk suggested by animal studies, GE believes
that EPA needs to conduct additional analyses of existing and forthcoming data if it is to
accurately assess and quantify the cancer risk that PCBs pose to humans.
Specifically, GE has previously submitted comments in several rulemakings,
urging EPA to consider the numerous epidemiological studies that have been performed on
populations with extensive workplace exposure to PCBs. Others have asked EPA to use
epidemiological studies to establish a human cancer potency factor. EPA's responses to
these comments, as well as statements in the Reassessment and the IRIS database, have
been sparse and have suggested strongly that EPA has not thoroughly reviewed the
epidemiological studies or considered how they can be used in risk assessment. GE
believes that the SOW presents the Agency with a good opportunity to consider this matter
more thoroughly.
B.1 The rationale for using epidemiological studies to establish environmental
standards
To date, EPA has established cancer slope factors for PCBs based on the results of
rat feeding studies. EPA's most recent effort in this regard is the Reassessment, which
advocates use of a range of cancer slope factors based on the results of rat studies. Risk
managers are to choose a slope factor from within the range based on the regulatory
context and the pathway by which humans are expected to be exposed.
Although EPA has historically viewed all positive findings in animal bioassays as
suggesting equally serious human health hazards, in reality chemical carcinogens may have
tissue-specific effects, and different mechanisms of action and pharmacokinetics.
38
-------
Additionally, chemicals may differentially exhibit carcinogenic effects under specific
animal bioassay conditions that are unrelated to reasonable human exposures. Moreover,
as discussed in the reassessment, studies vary in quality and power.
EPA recognizes the difference in potency of chemical carcinogens tested in animal
bioassays, but does not evaluate the probability that such chemicals may not be human
carcinogens. Many chemicals that have been proven to be carcinogenic at high doses in
animal bioassays have not been shown to be carcinogenic in humans at or near
environmental or occupational exposure levels. As an example, over 50 percent of
approximately 400 to 500 chemicals have tested positive in at least one rodent species at
high doses (Ames 1989). Howtver, only approximately 20 chemicals are known to cause
cancer in humans (Doll 1984; Paustenbach et al., 1990). Even after accounting for the
typical shortcomings of some epidemiology studies (small sample size and poor
quantitative knowledge of relatively small exposures), it is clear that many potent rodent
carcinogens do not pose an equivalent cancer hazard in humans. (Houk 1990; Kimbrough
1990).
There are several difficulties in estimating human cancer risks from rodent
bioassays. Differences in pharmacokinetics and susceptibility to organ toxicity complicate
the issue of interspecies extrapolation (MacDonald et al. 1994). Compounds classified as
tumor promoters are particularly troublesome in this regard, because they often produce
rodent liver tumors in long term bioassays, but are not generally known to cause cancer in
humans (Buttenvorth et al. 1995; Schulte-Hermann 1985). Tumor promoters like PCBs
selectively increase the growth of cancerous cells, but do not interact with cellular DNA to
cause the initial heritable change that begins the multi-stage process of cancer. The drug
Phenobarbital is a classic example of a rodent liver tumor promoter that has not been
shown to cause cancer in humans taking this drug for many years (Butterworth et al. 1995).
39
-------
Another problem caused by the use of animal studies to predict human cancer risk
is the need to use a model for extrapolation from high doses to animals to low doses to
humans. In the Reassessment, EPA estimated the carcinogenic potency of PCBs by using
the linear default method presented in EPA's Proposed Guidelines for Carcinogen Risk
Assessment (EPA 1996). This method is likely to overestimate the low-dose carcinogenic
risk of PCBs because it assumes that there is a direct linear relationship between the dose
of the chemical and a carcinogenic effect. The rationale given by EPA for using a linear
low dose extrapolation in the Reassessment is based on the possibility that PCBs might act
in concert with other exposures and processes leading to a background incidence of cancer
that would be linear at low doses.
Originally, the assumption of linearity was based on an elementary theory of the
mechanism of chemical carcinogenesis, in which a single chemical molecule can form an
adduct to DNA, and thereby result in cancer.7 Tumor promotion, however, is characterized
as a reversible process and the dose response relationship is expected to be nonlinear,
including both a threshold dose level and a maximal response (Pitot and Dragan 1991).
EPA's recent cancer guidelines (EPA 1996) allow for nonlinear low dose extrapolation in
cases where the available data support a nonlinear mode of action (e.g., nongenotoxic
agents).
EPA concedes that there are a number of chemicals which produce a carcinogenic
response by mechanisms that may exhibit a non-linear dose response curve at low doses
(EPA 1996; Butterworth and Slaga 1987). The increased acceptance of the nonlinearity of
dose and effect at low doses is evidenced by a growing consensus among risk assessment
practitioners that the linear model is inappropriate for dioxin, thyroid-type carcinogens,
nitrilotriacetic acid, trimethylpentane and, presumably, similar non-genotoxic chemicals,
(Paynter et al. 1988; Andersen and Alden 1989; Paustenbach 1989; EPA 1992b). Given
7 While genotoxic chemicals are assumed to be better modeled by a linear dose response assumption
(Weisburger and Williams 1987), this is not a proven scientific fact Ottobonni (1984) suggested that
genotoxic agents might also exhibit thresholds at low doses. These thresholds may result from a number of
factors including DNA repair mechanisms, cell death, or lethal mutations. Therefore, there is considerable
uncertainty in the assumption of low dose linearity for carcinogens.
40
-------
the uncertainty in cancer dose response modeling, the Agency should reexamine the
evidence for carcinogenic risk that can be derived from human epidemiology studies. It
has been stated that epidemiologic studies are not as statistically robust as animal studies
and, therefore, not as useful (Silbergeld et al. 1988). Although this can be a legitimate
concern in some cases, in many cases human epidemiology studies can and should be used
to validate, confirm, or set upper bound estimates of carcinogenic potency. In general,
when epidemiology data are available, it is not appropriate to accept only the results of
mathematical models that analyze rodent data without serious consideration given to the
human experience8 (Cook 1982; Dinman and Sussman 1983; Layard and Silvers 1989).
EPA (1996b) appears to recognize this point in its proposed cancer guidelines:
Epidemiologic data are extremely useful in risk assessment because they provide
direct evidence that a substance produces cancer in humans, thereby avoiding the
problem of species to species inference. Thus, when available human data are
extensive and of good quality, they are generally preferable over animal data and
should be given greater weight in hazard characterization and dose response
assessment, although both are utilized.
In the case of PCBs, EPA can no longer ignore the many clinical and
epidemiological studies that do not support the proposition that PCBs cause cancer in
humans. GE realizes that toxicologists must be careful in relying on the results of negative
epidemiological studies. However, when, as in the case of PCBs, several excellent
epidemiological studies have been performed using large numbers of workers heavily
exposed to a chemical over a long period of time, and the results of those studies have been
negative, GE submits that such results must be factored into, or used in, the derivation of a
human cancer potency factor. As discussed below, the epidemiology studies of
8 An example of where an animal study yielded implausible results is ethylene dibromide (EDB). In 1982,
it was claimed that workers exposed for 8 hrs/day for 40 years to the OSHA threshold limit value (TLV) for
EDB of 20 ppm incurred a risk of 999 in 1,000 of developing cancer. However, epidemiological evidence of
actual cancer incidence in these workers did not show an increase in the cancer rate (Cook 1993). Although
the EDB risks suggested by the low-dose animal models may initially seem plausible, the human
epidemiologic evidence makes it clear that these workers are not likely to die prematurely as the model
predicted (Hertz-Piciotto et al. 1988).
41
-------
PCB-exposed cohorts do not indicate that PCB exposure leads to increased mortality,
whether based on overall cancer mortality or deaths due to individual cancer types. These
findings strongly suggest that human health risks from PCB exposure have been
significantly overestimated in current regulations and that EPA should undertake a
thorough reevaluation of the actual risks posed by PCB exposures.
*
In assessing the PCB studies, EPA should use state-of-the-art methodology for
interpreting the results of epidemiological studies. This methodology uses a weight-of-the-
evidence test and applies what has become known as "causation analysis." The
methodology is well recognized within EPA (EPA 1992a; EPA 1996b). At least ten
criteria have been proposed for establishing cause and effect relationships (Hill 1965;
Evans 1976; Hackney and Linn 1979; Doll 1984; Guidotti and Goldsmith 1986; Mausner
and Kramer 1985; Monson 1988; Hemberg 1992). However, as typically applied, the
scientific convention applied in weight-of-the-evidence evaluation of epidemiological
studies requires (a) the observation of a specific cancer endpoint, and (b) the meeting of six
other criteria before a causal relationship between an agent such as PCBs and cancer can be
inferred (Hill 1965; Mausner and Kramer 1985; Rothman 1988; Monson 1988; Hemberg
1992; EPA 1985b; IARC 1987; EPA 1996b). The six fundamental criteria are: strength of
association; consistency of association; temporally correct association; dose-response
relationship; specificity of the association; and coherence with existing information (also
called "biological plausibility"). None of the criteria, with the exception of temporality,
should be considered as necessary to establish causation. Each of the criteria is important,
and causation is established by the weight of the evidence and the degree to which all six
criteria are satisfied by the available data. However, the rejection of the association may be
made with a high degree of confidence when three of the criteria — temporality,
consistency, and biological plausibility - are not met (Rothman 1988; EPA 1996b). In
addition to considering weight of the evidence, it is important to understand that studies
with larger cohorts and numbers of cancer deaths are inherently more important when
considering the weight of the evidence than are studies with smaller cohorts and fewer
cancer deaths.
42
-------
B.2 Epidemiological data
The collected evidence from numerous epidemiological studies over the past 20
years fails to demonstrate that PCBs cause cancer in humans, even in populations with
much greater exposures than those involved here. A review of the epidemiology data for
PCBs is provided below. In general, studies of PCB workers, who were exposed to PCB
levels hundreds or thousands of times higher than current environmental levels, have failed
to demonstrate a causal association between PCB exposure and cancer. Further, in a recent
study by Harvard University researchers, no relationship was found between PCBs and
breast cancer among 240 women (Hunter et al., 1997).
The most celebrated incident in which PCBs became suspected of causing cancer in
humans is the so-called "Yusho" incident. In 1968, about 1500 persons in Japan became ill
after consuming rice oil that was accidentally contaminated with a PCB mixture known as
"Kanechlor 400" (Amano et al. 1984). A similar incident, known as "Yucheng," occurred
in Taiwan in 1979. Typical symptoms were chloracne, swelling of eyelids, eye discharges,
brown pigmentation of the nails and skin, and curling of fingernails and toenails. Signs of
the disease were also observed in some offspring of affected mothers. Although the major
symptoms disappeared over the next sixteen to twenty years, subsequent studies suggested
a possible increase of cancer and adverse developmental and behavioral effects in
offspring.
The cause of the incident was extensively studied and the rice oil was found to
contain high levels of polychlorinated dibenzofurans ("PCDFs"), a chemical that is 100 to
1,000 times more toxic than PCBs. After finding that workers exposed to much higher
levels of PCBs showed minimal adverse health effects, and after performing dose-response
studies on the rice oil mixture, Japanese and Taiwanese scientists concluded that PCDFs
were the prime causal factor in the Yusho and Yucheng incidents (Kashimoto et al. 1986).
ATSDR agrees, finding that "there is inconclusive evidence of cancer in people who were
exposed to heated non-Aroclor PCBs during the Yusho and Yu-Cheng incidents, but
PCDFs were major contaminants" (ATSDR 1997).
-------
In 1985, Dr. Kimbrough and Dr. Goyer of the National Institutes of Health unequivocally
concluded that:
The scientific community assumes now that most of the effects observed in
these two outbreaks were caused by the ingestion of the polychlorinated
dibenzofurans. (Kimbrough et al. 1985)
Likewise, the Halogenated Organics Subcommittee of EPA's Science Advisory Board
reviewed a PCB health advisory from EPA and concluded that:
The health effects section suggests that the short-term human exposure to
Yusho poisoning is [not] representative of polychlorinated biphenyl
toxicosis. Recent studies indicate that the major etiologic agents in Yusho
were polychlorinated dibenzofurans rather than polychlorinated
biphenyls... Thus, a discussion of the human health effects of
polychlorinated biphenyls should not use 'Yusho' as an example. Industrial
exposure data more accurately reflect human health effects. (Doull et al.
1986)
A number of years later in her update of PCB exposure and human health effects,
Kimbrough (1995) emphatically stated that:
In the poisoning outbreaks, the PCDFs, not the PCBs. caused the
adverse human health effects.
Significantly, this scientific reinterpretation of the Yusho and Yucheng incidents is
consistent with data from animal studies that show a relatively low level of acute toxicity
— e.g., LD50s ranging from about 1 to 11 g/kg-body-weight in rats, depending on the
Aroclor mixture. Moreover, this explanation is consistent with the numerous studies
44
-------
(discussed below) that show 110 significant adverse health effects in workers who had been
exposed to average levels of PCBs higher than the Yusho patients were.
Subtracting Yusho from the universe of epidemiological studies of the cancer risk
of PCBs leaves a number of other studies which can be grouped into three categories: (!)
negative studies reporting no statistically significant relationship between exposure to
PCBs and cancer (Taylor 1988; Kimbrough et al. 1997, unpublished; Hunter et al. 1997;
Zack and Musch 1979; Gustavson 1986; Nicholson et al. 1987), (2) studies that were
inconclusive due to small cohort sizes or flaws in study design and data interpretation (e.g.,
Yassi et al. 1994); and (3) studies which have been cited by some as suggesting a
relationship between human exposure to PCBs and cancer (Bahn et al. 1976, 1977;
Bertazzi et al. 1987; Brown 1987; Sinks et al. 1991). As will become clear from the
following discussion, the studies, whether considered individually or assessed using a
weight-of-the-evidence approach, provide virtually no support to the claim that PCBs are
human carcinogens.
Inconclusive Studies
Yassi et al. (1994) examined the mortality of 2222 males employed between 1946
and 1975 at a transformer manufacturing plant in Canada. Although some transformers
were filled with PCB-containing fluids, the vast majority were filled with mineral oils
refined predominantly from naphthenic base crudes (only 85 of 51,000 transformers filled
between 1956 and 1975 contained PCB fluids).
This report concludes that neither overall mortality nor total cancer mortality varied
significantly from the expected. There were eleven deaths due to pancreatic cancer, a
statistically significant excess, but only three of the affected workers worked in transformer
assembly. Five of the affected workers worked at the plant for less than one year, and
another worked just two years. There were no liver cancers in the cohort.
45
-------
This study was heavily criticized by Wong (1995) for methodological flaws.
Moreover, the Manitoba Workers' Compensation Board awarded compensation to widows
of two of the men who died of pancreatic cancer on the basis that the cancers were linked
to mineral oil exposures. ATSDR (1997) concluded that the results of Yassi et al. (1994)
"must be regarded as inconclusive due to limitations such as exposure to other chemicals
and the fact that no medical history of the workers was provided."
Studies Cited as Linking Human PCB Exposure to Cancer
Bahn et al. (1976, 1977) evaluated the incidence of tumors occurring in a New
Jersey petrochemical facility where Aroclor 1254 had been used from 1949 to 1957. A
significantly increased incidence of malignant melanomas was observed among research
and development workers (2 of 31) and refinery personnel (1 of 41). In an update of that
same study, NIOSH (1977) observed eight cancers in the total study population (5.7
expected). Three of these tumors were melanomas and two were pancreatic cancers. The
incidence of these tumor types was reported to be significantly above calculated
expectations, although no data were presented. The results of this study were confounded
by the small cohort size, the fact that the workers in this facility were exposed to numerous
other chemicals, and the fact that the expected cancer rates were based on U.S. population
data rather than on local rates (Bahn et al. 1977; Lawrence 1977). ATSDR (1997) states
that the findings of this study should be regarded as inconclusive.
Bertazzi et al. (1987) conducted a retrospective cancer mortality study of 544 male
and 1,556 female workers who had been employed for at least one week in the manufacture
of PCB-impregnated capacitors in an Italian plant between 1946 and 1978. Mortality was
examined for that cohort from 1946 to 1982 and was compared to both national and local
mortality rates. Mortality due to all cancers (14 observed vs. 5.5 national and 7.6 local)
and due to cancer of the gastrointestinal tract (6 observed vs. 1.7 national and 2.2 local)
was significantly increased among male workers. Death rates from hematologic neoplasms
and from lung cancer were also elevated, but not significantly. Overall mortality was
significantly increased above local rates (34 observed vs. 16.5 local) in the female
46
-------
population. Total cancer deaths (12 observed vs. 5.3 local) and mortality from
hematologic neoplasms (4 observed vs. 1.1 local) were also significantly elevated over
local rates in the female population.
These results are limited by several factors, including the small number of cancer
cases observed, the limited latency period, lack of pattern or trend when data were
analyzed by duration of exposure, and some deaths in males with low potential for direct
PCB exposure (ATSDR 1997; Kimbrough 1987). A major problem in the study design
was the one week minimum period of employment required for inclusion in the study
resulted in the inclusion in the cohort of workers who had no PCB exposure. This makes
it difficult to assume that excess cancer cases are attributable to PCB exposures rather than
to other factors. This study also did not show a dose-response relationship or any direct
relationship between latency and the disease. Moreover, as discussed below, the results of
this very small study are dissimilar from the results of much larger and statistically more
valid studies of similar worker populations in the United States and Canada. ATSDR
(1997) found that the results of this study were inconclusive.
Brown (1987) found an excess risk of cancer of the liver, biliary tract, or gall
bladder in 2,588 workers (1,270 male, 1,318 female) from two capacitor factories. The
workers had worked for at least three months in areas where they received heavy exposure
to PCBs. Exposure was to Aroclors 1254, 1242 and 1016 (Lawton et al. 1981). The
workers were also exposed to other chemicals, including trichloroethylene, toluene, and
methyl isobutyl ketone.
The first evaluation of this cohort (Brown and Jones 1981) found increased cancer
mortality that was not statistically significant. After an additional seven years of
observation (Brown 1987), two additional cancers of the liver, gall bladder or biliary tract
were observed, making the cancer increase in this combined cancer grouping significant.
Among the grouped cancers, four of the five occurred in women from one of the plants.
There was no increase in the number of rectal cancers from the previous study. For the
total cohort, total mortality and cancer mortality were less than expected. Total cancer
47
-------
among the cohort at one of the plants was significantly less than expected (18 observed
versus 31 expected).
According to ATSDR, limitations and confounding factors in Brown (1987)
include the small number of cases and the fact that PCB blood levels were higher in the
plant with the lower incidence of cancer (ATSDR 1997). Moreover, the study failed to
account for several factors particular to the plant where the increased cancer incidence was
noted, including ethnicity (dominant Cape Verde background) and life style (the workers
were from a harbor/fishing town where alcohol consumption and smoking behaviors are
high). Furthermore, of the five liver grouping cancers, four of the workers had worked at
the plant 1.5 years or less and the other worker worked at the plant less than 10 years.
Finally, of the five cancers, only one was a primary liver tumor (the type of tumor
predicted by animal studies) and at least one had metastasized from another site (and was
therefore incorrectly identified as a liver tumor).
Sinks et al. (1991) conducted a retrospective cohort mortality analysis of 3,588
workers who were employed for at least one day at an electric capacitor manufacturing
plant between 1957 and 1977. Aroclor 1242 was used in this plant through 1970, and
Aroclor 1016 was used from 1970 to 1977. Mortality from all causes and from all cancers
was less than expected. A significant increase in mortality rate was observed for skin
cancer (8 observed vs. 2 expected) and death rates from brain and nervous system cancers
were non-significantly elevated over expected rates. No excess deaths were observed from
cancers of the rectum, lung, or liver, biliary tract and gall bladder, or from hematopoietic
malignancies. Based on a cumulative dose estimate, which incorporated information on
job station history, limited PCB environmental sampling data, and serologic data, the
authors were not able to establish a clear relationship between latency or duration of
employment and risk for malignant melanoma. Sinks et al. (1991) point out that the skin
cancer excesses are not consistent with those of similar studies. The authors also point out
that mortality may not be the best index of risk for malignant melanoma, as survival can be
affected by differences in health care quality. In addition, other limitations include the lack
of evaluation of exposures to other chemicals (metals, solvents, etc.), the relatively short
48
-------
latency period, the small number of deaths within the cohort, and possible misclassification
of brain cancer cases. Citing additional deficiencies in the study, ATSDR (1997) found the
results of the study inconclusive.
Negative Studies
By contrast to the inconclusive and confounded results of these studies which are ^
sometimes cited to link PCBs with cancer in humans, the largest study of PCB exposed
workers (Taylor 1988) showed no significant increases in mortality or cancers. Taylor
(1988) involved a cohort of 6,292 persons employed for at least three months during the
period 1946-1976 at the GE Hudson Falls and Ft. Edward facilities. This study showed no
increase in cancer mortality or in overall mortality compared to national averages. Deaths
due to malignant melanoma, lymphopoietic cancers, or the combination of liver,
gallbladder and biliary cancers were not significantly elevated, and brain cancers were well
below the expected value. PCB exposure was shown to be negatively associated with (not
statistically significant to) cancer mortality (all types combined) and lung cancer (the only
cancer outcomes with numbers of cases sufficient to permit a regression analysis). In other
words, as PCB exposure increased, the numbers of overall cancer deaths and lung cancer
deaths decreased.
Recently, in a follow-up to Taylor (1988), a retrospective mortality study was
conducted of the same cohort. (Kimbrough et al. 1997, unpublished). All workers were
-followed through the end of 1993. The cohort of 4062 white males and 3013 white
females contributed 120,811 and 92,032 person years of observation, respectively. There
were 763 (19%) deceased males and 432 (14%) deceased females. Death certificates were
available for 98.5% of the decedents and only 1.3% of the cohort was lost to follow-up.
For comparison, standardized mortality rates (SMRs) were calculated using both U.S. and
local county mortality tables.
Overall mortality for the total cohort was significantly lower than that for the
general population, as was the mortality for all cancers. The significantly lower SMRs in
49
-------
the male workers is partly attributable to the low SMRs in the salaried male workers. The
overall mortality and mortality due to all cancers of hourly workers was also significantly
lower. The dramatically low SMRs in salaried male workers were not as evident in
salaried female workers. However, 71% of the male salaried workers had obtained a
college education reflecting a socioeconomic factor that is well correlated with decreased
mortality (Sortie et al., 1995). Finally, there were no statistically significant increases in
mortality due to any of the a priori cancer types. The study concluded that there was no
evidence that PCB exposure at this plant had resulted in cancer mortality.
Additional recent studies undeimine the often-cited link between PCBs and cancer.
In Pittsfield, Massachusetts, PCBs were used in manufacturing over an extended period.
The Massachusetts Department of Public Health (MDPH) recently issued a registry of the
incidence of cancer mortality in Massachusetts from 1987 through 1994. The registry
showed no statistically significant increases in cancer incidence (at any level of statistical
significance) in Pittsfield or Berkshire County for any of the 23 types of cancers evaluated.
MDPH (1997). In fact, the results for all cancer types showed that total cancer incidence in
Pittsfield was 10 percent lower than expected based on the state-wide average, and also
showed lower-than-expected total cancer rates for other towns in the area.9 Last year, in
what many in the scientific world describe as a definitive result, Hunter et al. (1997)
published in the New England Journal of Medicine a study focused on the interaction of
endocrine disruption and cancer. The study showed no link between PCB exposure and
breast cancer. Similar results were reported by Key and Reeves (1994). As Dr. Steven
Safe noted in an editorial accompanying Hunter et al. (1997), this study and others "should
reassure the public that weakly estrogenic organochlorine compounds such as PCBs, DDT,
and DDE are not a cause of breast cancer." (Safe 1997).
9 Similarly, in a prior letter in 1980, the MDPH had advised the City of Pittsfield that review of the cancer
mortality data for 1969-1978 showed no excess cancer mortality in Pittsfield across all causes of cancer, and
further showed no excess canceT mortality in the Lakewood neighborhood adjacent to the GE facility.
Pailcer (1980).
50
-------
Assessment of the Epidemiological Studies of PCB Exposure
Conservatism in hazard identification is manifested when regulatory agencies place
an emphasis on data that chemicals might pose adverse effects, and little weight on data
that suggest that chemicals fail to cause adverse effects. Emphasizing study data that show
adverse health effects in animals while virtually ignoring studies showing no adverse
effects does not represent a balance of scientific information. (Nichols and Zeckhauser
1988) Frequently, extraordinary confidence is placed on a study that suggests that a
chemical may pose a particular hazard, while only modest consideration is given to the
study's quality.
More recently, the scientific community and some regulators have come to accept
that not all scientific data are equal, and that only data of similar quality should be
compared when drawing conclusions regarding toxic effects based on multiple studies.
This philosophy, known as a "weight of evidence" approach, represents an important
refinement that should be applicable to both hazard identification and dose response
assessment (Sielken 1985; Anderson 1989; Gray et al. 1993). EPA's (1996b) proposed
cancer risk guidelines also embrace this philosophy. The benefit of using a "weight-of-
evidence" approach is that the results of several high quality toxicity studies will not be
disregarded simply because the results of one or two poorly controlled studies have
dissimilar findings.
As is clear from the discussion of the PCB epidemiology studies, none of the cancer
incidence and mortality studies demonstrates a cause-effect relationship between PCB
exposure and cancer.10 Not only do the individual studies fail to show causation, but the
weight of the evidence from the studies taken collectively also fails to establish any such
relationship.
10 It is acknowledged that Rothman et al. (1997) observed a dose-response relationship between PCB serum
levels and non-Hodgkin lymphoma; however, the authors pointed out that their "results should be regarded
as hypothesis generating," that "our findings require replication, and biological plausibility," and that the
matter "needs further investigation." The authors also noted that studies of highly exposed capacitor workers
do not support a relationship between non-Hodgkin lymphoma and PCB exposure.
-------
As discussed previously, the scientific convention applied in weight-of-the-
evidence evaluation of epidemiological studies requires (a) the observation of a specific
cancer endpoint, and (b) the meeting of other criteria (strength of association, consistency
of association, dose-response relationship, temporally correct association, specificity of the
association, and coherence with existing information (biological plausibility)) before a
causal relationship between an agent such as PCBs and cancer can be inferred. None of the
criteria, with the exception of temporality, should be considered as necessary to establish
causation. Each of the criteria is important, and causation is established by the weight of
the evidence and the degree to which all six criteria are satisfied by the available data.
However, the rejection of the association may be made with a high degree of confidence
when three of the criteria - temporality, consistency, and coherence with existing
information - are not met. (Rothman 1988; EPA 1996b) In addition to considering
weight of the evidence, it is important to understand that studies with larger cohorts and
numbers of cancer deaths are inherently more important when considering the weight of
the evidence than are studies with smaller cohorts and fewer cancer deaths.
In the PCB studies, small increases in a variety of cancer endpoints were seen in
different populations with no common thread, and several studied populations showed no
increases at all. The discrepancies can be explained in innumerable ways, including
exposures to other chemicals, population life styles, and even chance. Thus, little evidence
exists that PCBs are human carcinogens, and the weight of the evidence fails to establish a
definitive causal relationship between exposure to PCBs ~ even in high concentrations ~
and the incidence of cancer in humans.
In 1993, TERRA, Inc. submitted comments on the Great Lakes Initiative that
provided a thorough weight-of-the-evidence assessment of what the authors classified as
the four "major" cohorts in the PCB epidemiological studies (the Brown, Nicholson, Sinks
and Taylor cohorts (TERRA, 1993). GE incorporates these comments by reference. The
following tables from that document provide a summary of TERRA'S analysis.
52
-------
Summary of The Major PCB Mortality Studies
Study/Cohort
Were Significant Increases Noted?
SMR for All
Cancer Deaths
Total
Cancers
Malignant
Melanoma
Liver/Biliary
Cancers
The Brown Cohort




Brown and Jones 1981
Mortality study of 163 deaths among 2,567
U.S. capacitor workers
89
No
No
No
Brown 1987
Updated mortality of Brown and Jones 1981
analyzing 29S deaths among 2,588 workers
78
No
No
Yes
(but limited
to females)
Nicholson et al. 1987
Mortality study of 188 deaths among 769
U.S. capacitor workers with '5 years
exposure and *10 years latency since first
exposure
79
No
No
No
Sinks et al. 1992
Mortality study of 192 deaths among 3,588
U.S. capacitor workers
85
No
Yes
(but limited
to males)
No
Taylor et al. 1988
Mortality study of 510 deaths among 6,292
capacitor workers
95
No
No
No
53
-------
Evaluation of the Major Cohorts Using Causation Criteria

Was This Criteria Satisfied?
Causation Criteria
Liver
Skin
1. Strength of the Association
No*
No*
2. Consistency of the Association
No
No
3. Temporal Relationships
No
No
4. Dose-Response Relationships
No
No
5. Specificity of Association
No
No
6. Coherence of Evidence
No
No
* Denotes a statistical observation that is considered weak because it is confounded by a lack of
confirmation in studies of equivalent or greater size, and it lacks confirmation in a study using greater
exposure duration and latency criteria for cohort definition.
From its analysis, TERRA concluded that "the available scientific evidence do not
support the contention that PCBs are carcinogenic in humans."
Other scientists have reached similar conclusions. For example, Chase et al.
(1989), concluded that:
There is insufficient evidence to show a causal relationship between PCB exposure
and the subsequent development of any form of cancer. In light of the long-term
and widespread usage of PCBs in the workplace and, in some cases, the extensive
exposures of workers, it is likely that evidence of carcinogenicity in humans would
have been observed in the various epidemiological studies discussed above if PCBs
were in fact potent carcinogens.
Similarly, Kimbrough 1988 concluded that:
Thus far, no conclusive adverse effects have been demonstrated in people who
carry body burdens of PCBs from environmental exposure to trace amounts of
54
-------
PCBs... Even workers with exposures two orders of magnitude greater than
environmental exposures show no convincing health effects... Thus, despite positive
laboratory animal data and except for chloracne, exposure to PCBs has led to no
convincing, clinically demonstrable, chronic health effects in humans.
In her 1995 update, Dr. Kimbrough reaches a similar conclusion (Kimbrough 1995).
A recent review of the occupational studies by the American Council on Health and
Science also concluded that none of the studies provides evidence that PCB exposure
increases cancer risk in humans (Danse et al. 1997). A recent review of studies seeking to
determine if there was a relationship between environmental exposures to PCBs and any
human health effects, including cancer, found that "none of the 33 studies where exposure
had occurred in the natural environment provided positive or suggestive evidence of an
association with adverse effect." Swanson et al. (1995).
A fair and careful review of the existing PCB occupational studies leads to the
conclusion that there is no credible evidence that PCBs cause cancer in humans, even at
exposures that are orders of magnitude greater than environmental exposures. Therefore,
GE urges EPA to reassess the human carcinogenicity of PCBs in light of the
epidemiological studies.
BJ Derivation of a cancer slope factor for PCBs from the epidemiological studies
Although the weight of the evidence approach results in the conclusion that there is
no credible evidence that PCBs cause cancer in humans, it is still possible to derive a
cancer slope factor from the epidemiological studies. In TERRA (1993), the authors
derived such slope factors using two approaches.
First, the authors assumed that the results of Brown (1987) showing a statistically
significant increase in combined liver and biliary cancers reflected a real measure of cancer
potency. The authors then used the observed increase in cancer incidence, along with a
55
-------
conservative estimate of exposure, to generate a human cancer potency factor. That is, it
was assumed that the results of Brown (1987) were representative of human cancer risk,
even though other studies of comparable or greater size had failed to duplicate the findings
of liver cancer.
Second, the authors used the negative results of the largest study completed at that
time (Taylor 1988), along with a conservative estimate of exposure, to calculate an upper
confidence on the measured zero risk, thereby placing an upper bound on the risk.
The authors' methodology for estimating exposure and calculating cancer slope
factors are described in detail in their paper. It is important to note that the authors
estimated exposure using two different methods: estimating the daily dose capacitor
workers received using reported workplace exposure estimates and known or estimated
absorption; and estimating daily doses received by the capacitor workers using basic
pharmacokinetic principles and reported body burdens and estimated tissue half-lives. In
both methods, the authors used conservative assumptions to assure that there was little
possibility that dose would be overestimated. As one example, when assessing exposure
through the inhalation route, the authors used the geometric mean of work place air
concentrations measured in the mid- to late 1970s, when PCB use was being phased out.
The two methods of estimating exposure arrived at very similar dose estimates.
To be conservative in calculating cancer slope factors from the dose estimates and
the results of Brown (1987) and Taylor (1988), and in recognition that there is some degree
of uncertainty in the incremental risk rates that were calculated, the authors calculated
cancer slope factors using both the measured cancer incidence rate and the 95% upper
confidence limit on the incremental risk rate. The calculated cancer slope factors are as
follows:
56
-------
Cancer Slope Factors
Study/Method
Cancer Slope Factor
(me/ka/day*1)
Brown (1987) - Measured
5.9 xlO'3 1
Brown (1987). 95% UCL
1.9 x 10'2
Taylor (1988) - Measured
7.7 x 10"4
Taylor (1988)-95% UCL
8.9 x 10*3
For the reasons discussed throughout these comments, GE believes that cancer
slope factors calculated from epidemiological studies can be used to establish
environmental standards, including a water quality standard for PCBs. Given that Taylor
(1988) is the largest epidemiological study performed to date and is highly relevant to the
Hudson River, GE recommends using the measured cancer slope factor from this study as
the starting point for establishing environmental standards for PCBs based on cancer risk.
As discussed above, the workers studied in Taylor (1988) were exposed primarily to
Aroclor 1242 and 1254, with minor exposure to Aroclor 1016. Thus, it is conservative to
use the measured cancer slope factor from Taylor (1988) as the cancer slope factor for
Aroclor 1242.
Accordingly, General Electric proposes that the SOW for the Hudson River Phase 2
Risk Assessment use a CSF of 7.7 x 10"4 (mg/kg/day)*1.
B.4 Summary for cancer toxicity assessment
Although the Reassessment was a positive step in reevaluating cancer risk from
exposure to PCBs, GE strongly believes that EPA should use the numerous
epidemiological studies that have been performed to date to further assess the true human
cancer risk of PCBs. EPA can no longer ignore the many clinical and epidemiological
studies that do not support the proposition that PCBs cause cancer in humans. Although
57
-------
toxicologists must be careful in relying on the results of negative epidemiological studies,
when several excellent epidemiological studies have been performed using large numbers
of workers heavily exposed to a chemical over a long period of time, and the results of
those studies have been negative, such results cannot be ignored.
These comments provide a scientifically valid rationale for using epidemiological
studies rather than rodent studies to establish environmental standards for PCBs. They also
provide an assessment of the adequacy of the PCB epidemiological studies for evaluating
the human cancer risk of PCBs and set forth overall conclusions that can be drawn from
those studies using a "weight of the evidence" approach. These comments have further
shown how the studies can be used to derive a conservative cancer slope factor for PCBs.
GE strongly urges the Agency to use the opportunity presented by the HHRA to consider
this matter more thoroughly.
C. Noncancer toxicity values
The noncancer reference dose (RfD) cited in the SOW is flawed and overly
conservative, in that it is based on an inappropriate monkey study and overly conservative
uncertainty factors. In the SOW, EPA plans to base the noncancer risk assessment using
the oral RfD of 2 x 10"5 mg/kg-day (20 ng/kg-day) for Aroclor 1254. This RfD is based on
dermal, ocular and immunologic effects in a series of studies of rhesus monkeys reported
by Arnold et al. (1993a,b) and Tryphonas et al. (1989; 1991a,b).
The Aroclor 1254 reference dose (RfD) is based on the results of a five year feeding ^4
study in rhesus monkeys (Arnold et al., 1993a,b; Tryphonas et al., 1989; Tryphonas et al.,
1991a,b). Groups of 16 adult female monkeys ingested gelatin capsules containing
Aroclor 1254 (in glycerol, com oil vehicle) at daily doses of 0, 5, 20, 40, or 80 ug/kg-day
for over five years. PCB concentrations in the monkeys had achieved steady state
pharmacokinetics by 25 months of exposure, as demonstrated by PCB measurements in
blood and adipose tissue (Tryphonas et al., 1989; Mes et al., 1989). The general health and
clinical pathology findings in the adult female monkeys dosed with Aroclor 1254 were
58
-------
reported by Arnold et al. (1993a,b). Clinical signs of toxicity were limited to eye exudate,
inflammation, and/or prominence of the tarsal (Meibomian) glands, and changes in finger
and toe nails. Significant dose related trends were reported for these clinical signs (Arnold
et al., 1993a). The results of an immunologic assessment of the PCB exposed adult female
monkeys was reported by Tryphonas et al. (1989, 1991a,b). The most significant finding
was a treatment-related decrease in antibody response (IgG, IgM) to sheep red blood cells
(SRBC). The LOAEL for clinical signs and immune system effects was 5 x I0"3 mg/kg-
day. A total UF of 300 was applied (ten for sensitive individuals, three for interspecies
extrapolation, three for minimal effect LOAEL to NOAEL, three for subchronic to chronic
study duration) and a RID of 2 x 10'5 mg/kg-day was calculated.
A number of questions can be raised about EPA's selection and evaluation of the
studies from which the Aroclor 1254 RID was derived. Two critical shortcomings are ^
most important. First, those studies are not appropriate for deriving an RfD because: (a)
the clinical relevance of the immunologic changes reported by the researchers has not been
demonstrated; (b) the rhesus monkey is not an appropriate model for dermal, ocular, and
nail effects of PCBs in humans; and (c) there is compelling evidence to indicate that rhesus
monkeys metabolize PCBs in a significantly different way from humans. Second, EPA
used inappropriate and overly conservative uncertainty factors in extrapolating from this
set of studies to derive an RID for humans. These points are explained in greater detail in a
memorandum prepared by Dr. Russell Keenan and Ms. Carol Gillis, then at ChemRisk,
which was attached as Exhibit 96 to GE's May 1, 1998 comments on EPA's proposal to
list the Housatonic River site on the National Priorities List (GE, 1998) and is incorporated
by reference herein.
Despite changes in immunologic parameters reported by the researchers, clinical
relevance of these changes has not been demonstrated. In addition, the rhesus monkey is
not an appropriate model for dermal, ocular and nail effects of PCBs in humans. A
comparison of effects and body burdens (blood serum levels) seen in workers exposed to
PCBs and the effects and associated body burdens reported in rhesus monkey studies
indicates that PCBs produce nail changes, ocular effects, and dermal effects at much lower
59
-------
doses in rhesus monkeys than in humans (Gillis and Price, 1996). In fact, PCBs alone do
not produce nail changes, chloracne or ocular effects in humans as demonstrated by the fact
that the length of exposures in epidemiological studies would have provided adequate
opportunity to demonstrate these effects if they were occurring. These findings suggest
that rhesus monkeys are significantly more sensitive and respond differpntly to PCBs than
humans
More importantly, there is compelling evidence to indicate that PCBs are 67
metabolized differently in humans than in rhesus monkeys and that the metabolism of
PCBs may be critical to the overall expression of PCB toxicity (Brown, 1994; Brown et al.,
1994). One indication of this difference is the line of evidence suggesting that the patterns
of PCB congeners that accumulate in adipose and hepatic tissues of rhesus monkeys
chronically exposed to Aroclor 1254 differ from patterns of congener retention in humans
exposed to PCBs. Humans produce a retention pattern similar to that observed in in vitro
studies of P4502B enzyme activity, referred to as P4502B-like metabolism (Brown et al.,
1989; 1994). Humans produce a second pattern when exposed to mixtures of PCBs and
furans (Masuda et al., 1978; Kunita et al., 1984). This pattern results from metabolism of
PCBs by a combination of P4502B-like and P4501A-mediated metabolic pathways. It
appears that, in the absence of concurrent exposures to dioxins and furans, PCBs do not
induce the P4501A enzymes in humans. Furthermore, studies of PCB induction of
P4501A in rodents suggest that such induction, if it occurs in humans, would require
exposures of PCBs far higher than have occurred from environmental or historical
occupational exposures (Brown et al., 1991).
In contrast to the metabolism of PCBs in humans (in the absence of concurrent
68
exposures to dioxins and furans in the toxic range), a different pattern is observed in rhesus
monkeys. Metabolism patterns of PCBs in rhesus monkeys indicate that PCBs are
metabolized by means of the P4501A pathway and a second pathway known as the
P450RH pathway, which appears to be unique to the rhesus monkey (Brown, 1994). The
specific enzymes responsible for metabolizing PCBs in the unusual P450RH pattern
observed in monkeys are unclear at this time. However, the differences in enzyme systems
60
-------
support the finding that PCBs are metabolized differently in ihesus monkeys and humans,
and suggest that the rhesus monkey is a poor model for endpoints associated with the
activation of P4501 A.
Several studies have demonstrated that PCB metabolism is critical to the expression
of PCB toxicity in humans (Brown, 1994; Brown et al., 1989; 1991; 1994). For example,
induction of P4501A at low PCB doses is associated with dermal, ocular, and nail effects
in animals (Brown et al., 1994). In humans, Yusho victims, who were exposed to both
PCBs and furans and experienced many of these effects, also displayed P4501A
metabolism. Conversely, metabolism of PCBs under the P4502B-like pathway in
occupationally exposed human populations is not associated with these effects. In
summary, the findings on the metabolism of PCBs suggest that the differences between
rhesus monkeys and humans with respect to PCB toxicology may extend beyond dermal,
ocular, nail and immunological effects.
In addition to EPA's improper selection of the critical study and toxicological
endpoints for establishing the RfD, EPA applied inappropriate uncertainty factors for
quantitatively deriving the RfD. According to EPA, an uncertainty factor of three was
applied to account for interspecies extrapolation due to the similarities in toxic responses
and metabolism of PCBs in monkeys and humans and the general physiologic similarities
between the species. This uncertainty factor implies that humans are three times more
sensitive than the rhesus monkey for the critical effects on which the RfD is based.
However, because the evidence indicates that the rhesus monkey is significantly more
sensitive than humans, EPA's uncertainty factor of three to account for interspecies
sensitivity is inappropriate. An uncertainty factor equal to or less than one would be more
appropriate to address interspecies uncertainty.
In addition, EPA applied an uncertainty factor of three to adjust for study duration.
This factor is intended to account for uncertainties related to less-th an-chronic exposure
and the assumption that a longer exposure duration may result in more pronounced adverse
effects as adverse effects at a low dose. In the case of Aroclor 1254, the monkeys were
-------
dosed for greater than 25 percent of their lifetime and steady-state PCB body burdens were
achieved (Arnold et al., 1993a,b). This suggests that a longer exposure duration would not
result in an increased toxic response. Thus, an uncertainty factor of three is not warranted
and should be reduced to a factor of one. In summary, reduction of the uncertainty factors
for interspecies sensitivity and study duration from a value of three to one results in a total
uncertainty factor of 30, rather than 300. Even if one were to accept EPA's selection of the
toxicological study and agree with EPA's evaluation of the critical effects, the necessary
adjustments to the total uncertainty factor would argue for a revised chronic RfD no more
stringent than 200 ng/kg-day.
D. Endocrine disruption
The SOW currently plans to evaluate "the potential for endocrine effects" in the yj
HHRA. GE disagrees with the need for conducting this evaluation, even if it is limited to
"a qualitative assessment of the currently available information on the potential effects of
PCBs on the endocrine system," as can be inferred from the SOW. We believe this
analysis is unwarranted in light of recent EPA findings on this topic as well as those of
independent scientific researchers. For example, it has been suggested that PCBs can
interfere with normal endocrine function leading to infertility and other hormone related
disorders, although recent reviews suggest that the evidence for these effects is weak and
circumstantial (Danse et al., 1997; Golden et al., 1998). Indeed, EPA (1997a) concluded
that, with few exceptions (for compounds unrelated to PCBs), "an adverse health effect in
humans via endocrine disruption has not been established."
Last year, in what many in the scientific world describe as a definitive result,
Hunter et al. (1997) published in the New England Journal of Medicine a study focused on
the interaction of endocrine disruption and cancer. The study showed no link between
PCB exposure and breast cancer. Similar results were reported by Key and Reeves (1994).
As Dr. Steven Safe noted in an editorial accompanying Hunter et al. (1997), this study and
others "should reassure the public that weakly estrogenic organochlorine compounds such
as PCBs, DDT, and DDE are not a cause of breast cancer." (Safe 1997).
62
-------
GE believes that further qualitative or quantitative evaluation of this topic is not
necessary or appropriate in the Hudson River HHRA. As EPA and independent scientific
researchers have been unable to find a link between PCB exposure and endocrine
disruption, it is not worthwhile to invest the level of effort and resources necessary to
further elucidate this topic in the context of a Superfund site risk assessment. Furthermore,
by raising it as a potential issue in the Hudson River HHRA, which is itself based on a
hypothetical exposure scenario, undue alarm or concern may be generated. This would not
serve the public interest.
E. Uncertainty in toxicologies! criteria
The SOW states the Agency does not intend to evaluate the uncertainty in the dose 72
response criteria (the RID and the CSF) in the human health assessment (SOW, at 14).
The basis for this approach is purportedly consistent "with EPA's policies (EPA,
1997a,b)." These documents are not a statement of the limits of Agency policy. Rather
they are a set of guiding principles for the use and evaluation of Monte Carlo modeling.
The relevant passage in the documents is as follows:
For human health risk assessments, the application of Monte Carlo, and other
probabilistic techniques has been limited to exposure assessments in the
majority of cases. The current policy, Conditions for Acceptance and
associated guiding principles are not intended to apply to dose response
evaluations for human health risk assessments until this application of
probabilistic analysis has been studiedfurther. (EPA 1997a, at 2)
It is clear that the EPA's current policy is limited to the evaluation of exposure
assessments in human health risk assessments, and that the Agency does not provide the
assessor with guidance for the evaluation of Monte Carlo assessments of dose response
(toxicity)." This is not the same as a policy that forbids the consideration of quantitative
" See the identical language in U.S.EPA's Guiding Principals for Monte Carlo Analysis, (EPA, 1997b).
63
-------
modeling of the uncertainty in dose response. It merely is a statement that the current
policy does not address this type of analysis and that the Agency is not offering guidance
on the evaluation of this practice.
Therefore, there is no basis for concluding that the uncertainty in toxicity
measurements can be set aside or ignored in the evaluation of the uncertainty in risk
estimates. GE is unaware of any legal or technical reason to exclude the uncertainty in
toxicity criteria. In fact, the Agency's own technical experts have advocated that the
Agency consider the uncertainty in toxicity on a case-by-case basis (SAP, 1998).
GE believes that it is critical to account for the uncertaiiity in toxicity
measurements. It has long been recognized that the toxicity portion of the risk assessment,
not the exposure assessment, is the greatest source of uncertainty in risk assessment
(McKone and Bogan, 1991). Thus, the decision to exclude consideration of such
uncertainty has grave impacts on the assessment of risks.
Given the importance of this issue, the only basis for excluding consideration of
uncertainty from the HRA is technical infeasibility. However, a number of technical
approaches have been suggested for characterizing uncertainty in toxicity criteria. The
uncertainty in cancer slope factors has been investigated by several authors (Crouch et al.
1996; Evans et al. 1996a,b). Uncertainty in the RfD has also be the subject of a number of
research publications (Slob, 1997; Baird et al. 1996; Price et al., 1997, and Swartout et al.
1998). Swartout et al. (1998) has established a framework for redefining the RfD in
probabilistic terms.
Techniques for the integration of uncertainty in toxicity assessments into
assessment of exposure uncertainty have also been developed (Carlson-Lynch et al. 1997;
Harvey, et al. 1997; Price et al. 1998). In Price et al. (1998), a Monte Carlo model was
constructed of the uncertainty and variation in the PCB dose rates of anglers consuming
fish from the Tennessee and Clinch Rivers. This model was combined with information on
the uncertainty in the toxicity of PCBs to estimate the probability of exceeding the actual
64
-------
dose that is "protective of sensitive individuals". The result is an assessment that fully
discloses the uncertainty in the risk characterization for PCB noncarcinogenic effects and
provides the risk manager with the most appropriate basis for decision making.
As discussed elsewhere, the Agency's approach for Monte Carlo analysis is not
entirely clear, however, at several points the Agency has indicated that uncertainty will be
explicitly modeled (SOW, at 10, 13). Therefore, there appears to be no technical reason
why information on uncertainty on the PCB toxicity criteria cannot be evaluated with the
uncertainty in exposure.
F. Averaging time
The SOW does not discuss the issue of averaging time. Averaging time is a term in 73
the equation to estimate the average daily dose rate that is used to evaluate non-
carcinogenic effects (EPA, 1989). When the dose rate received from a source of
contamination is constant over time, the issue of averaging time does not affect the
estimate of average daily dose. However, when the dose rate changes over time, the choice
of the duration for the averaging time can have a dramatic impact on the estimate of daily
dose (Muir et al. 1998). Since the proposed model will address changes in fish tissue
concentrations over time, the selection of an averaging time is an important issue for the
HHRA.
GE recommends that the averaging time be established based on the half-life of
PCBs in humans. Lipophilic compounds exert their chronic effect as a function of the
long-term body burden of the compounds. Studies of PCBs in test animals determine
doses that are approximately in equilibrium with the body burdens that are associated with
the presence or absence of long tern effects. Therefore, it is important that the averaging
period is sufficiently long that the dose can come into some sort of equilibrium with the
body burdens of the anglers. This suggests that the averaging time be several multiples of
the half-life of PCBs in humans. At a minimum, the averaging time should be greater than
10 years.
65
-------
This Page Was Intentionally Left Blank For Pagination Purposes.
66
-------
SECTION V
RISK CHARACTERIZATION ISSUES
A. Consideration of background sources of PCBs
¦»
PCBs were used in consumer and commercial products from the 1940s through the
1970s. Because of this widespread use and the persistence of the compounds, PCBs still
occur at trace levels in many foods and in many households. As a result, all individuals
cany trace levels of PCBs in their bodies (ATSDR, 1997).
One criteria in risk management is to determine how exposures from a particular
source relate to the background levels and whether exposure from a certain source will
significantly raise an individual's body burdens. For example, if an individual has 2 ppb of
a contaminant in his or her blood because of background sources, and exposure at a
Superfund Site raises the blood level by 0.01 ppb, then there is likely to be little health
benefit to the individual from the control of contamination at the Site. If the exposures at
the site doubles or triples the individual's body burden then the potential for causing
adverse effects is much higher and the control of the site may be warranted.
The impact of a source on background body burdens can be investigated directly by
surveying the blood levels of individuals exposed by certain routes and comparing the 74
results to levels in unexposed populations. Such studies have been performed for exposure
to PCBs in soil (Chase et a!., 1989; ATSDR, 1987) and exposure from the consumption of
fish in the Great Lakes, Tennessee River (ATSDR, 1997b), and in Western Massachusetts
(Housatonic River). In the latter two studies, the individuals who consumed fish did not
have elevated blood levels of PCBs.
Such direct studies cannot be performed for the Hudson River since the ban on
keeping fish has eliminated exposures from fish consumption. However, it is possible to
take the output of the exposure models proposed in the SOW and determine the
incremental change that consumption would cause in background levels of PCB
67
-------
(Avantaggio et al. 1997; Keenan et al. 1997). Such an analysis would provide EPA with
information on whether the control of PCBs from the consumption of fish would have a
significant impact on angler's body burdens.
B. Development of the central and high-end exposure risk estimates
The Agency has indicated that risk management decisions will be made based on
the current and future risks to average exposed individual and to the reasonably maximally
exposed individual (RMEI) (SOW, at 2). These two risk measurements are proposed to be
based on the exposure assessment that consists of a deterministic and probabilistic analysis 76
(SOW, at 6). Thus, EPA is proposing to use two estimates of risk for the average
individual and RMEI. The SOW does not clearly state how the two estimates will be used
in the risk characterization process.
Deterministic assessments have been historically used at Superfund sites to
consider the need to perform remediation and to select a remedial option. At most small
sites the deterministic baseline risk assessment would be representative of the site risks.
For larger and more complex sites with dynamic physical features (like the Hudson River),
a deterministic model is inadequate even to describe "typical" conditions. Probabilistic
models, such as Monte Carlo models, are more appropriate in such cases since they have a
better chance to capture the variability and the uncertainty which is inherent as sites
become more complex.
Consequently, EPA should abandon its proposal to use the findings of the
deterministic exposure assessment As discussed in EPA's Guiding Principles on Monte 77
Carlo (EPA, 1997), the use of probabilistic techniques provides the decision maker with
additional insights that the deterministic methods cannot provide. Therefore, the
assessment of risks to the RMEI and average individual should be based only on the
findings of the probabilistic analysis.
68
-------
C.	The baseline assessment cannot be used to select remedial options.
The baseline risk assessment cannot be used to assess the reduction of risk
associated with the implementation of a remedial strategy. Rather, the baseline HHRA
examines the hypothetical risk associated with the "no action" scenario: These risks must
be measured against the estimated risks that might remain after implementation of various
remedial options in order to assess the appropriateness of such actions. Therefore, a
separate risk reduction analysis should be completed for each remedial option to focus on
the net reduction in risks, if any, associated with the different remedial options.
D.	Use of risk finding for small number of anglers fishing hot spots
The SOW describes the assessments of risks associated with fish caught in hot
spots as one of the analyses that will be performed (SOW, at 13). However, the procedure
to evaluate the results of the risk assessment — i.e., how it will be used in the risk ^9
management and feasibility study remedial alternatives assessment — is not presented in
the SOW. One area of possible misinterpretation is the extreme upper limit risks that will
likely be calculated for the localized areas of elevated contamination. It is important that
these risks be put into proper context in the overall risk. This assessment should include a
discussion of the areal extent of the elevated concentrations, probability of a suitable
fishery, probability of anglers in the same area, angler success, and other relevant factors.
E.	Choice of wstart date"
According to the SOW, the baseline risk assessment will focus on a population of
anglers who begin fishing in 1999 (SOW, at 11). The selection of this date may not be 80
appropriate for the baseline assessment. As discussed above, a separate risk assessment
will be needed to evaluate the risks that might remain after various remedial options.
These risks will not exist until the remediation has been completed. Given the current
schedule for the Hudson River reassessment, this could happen no earlier than 2002 to
2005. The data used in the remedial decision-making process should represent the
69
-------
conditions that best represent the risks beginning at that time. In order to be consistent
with the remedial assessment, the Agency should start the baseline risk assessment clock
beginning in 2002 not 1999.
70
-------
REFERENCES
Abraham, K., L. Alder, H. Beck, W.Mathar, R. Palavinskas, U. Steuerwald, and P. Weihe
(1995). Organochlroine compounds in human milk and pilot whale from Faroe Islands. 15th
International Symposium on Chlorinated Dioxins and Related Compounds 26:63-67.
Edmonton, Canada. August 21-25.
Ahlborg, U.G., G.C. Becking, L.S. Bimbaum, A. Bouwer, H.J.G.M. Derks, M. Feeley, G.
«
Golor, A. Hanberg, J.C. Larsen, A.K.D. Liem, S.H. Safe, C. Schlatter, F. Waem, M.
Younes and E. Yijanheikki. 1994. Toxic equivalency factors for dioxin-like PCBs.
Chemosphere 28(6): 1049-1067.
Amano, M., K. Yagi, H. Nakajima, R. Takehara, H. Sakai, G. Umeda. 1984. Statistical
observations about the causes of death of patients with oil poisoning. Japan Hygiene 39:1 -
5.
Ames, B.N. 1989. Mutagenesis and carcinogenesis: Endogenous and exogenous factors.
Environ. Molec. Mutagen. 1-30.
Andersen, R.L. and C.L. Alden. 1989. Risk assessment for nitrilotriacetic acid (NTA), p.
390-426, In, D.J. Paustenbach, (ed.), The Risk Assessment of Environmental and Human
Health Hazards: A Textbook of Case Studies. John Wiley and Sons, New York.
Anderson, A.C. and J.C. Rice (1993): Survey of fish and shellfish consumption by residents
of the greater New Orleans area. Bull. Environ. Contam. Toxicol. 51:508-514.
Anderson, E.L. 1989. Scientific developments in risk assessment: Legal implications.
Col. J. Environ. Law 14(411):412-425.
Armstrong, B.K. and A. Kricker. 1995. Skin cancer. Dermatol. Clin. 13:583-594.
71
-------
Arnold, D.L., F. Bryce, R. Stapley, P.F. McGuire, D. Bums, J.R. Tanner, and K. Karpinski.
1993. Toxicological Consequences of Aroclor 1254 Ingestion by Female Rhesus (Macaca
mulatto) Monkeys, Part 1A: Prebreeding Phase - Clinical Health Findings. Health and
Welfare Canada, Health Protection Branch, Toxicology Research Division.
t
ATSDR. 1987. Exposure Study of Persons Possibly Exposed to Polychlorinated
Biphenyls in Paoli, Pennsylvania. U.S. Department of Health and Human Services,
Agency for Toxic Substances and Disease Registry. Atlanta, GA. November.
ATSDR. 1997. Toxicological Profile for PCBs (Update). Agency for Toxic Substances
and Disease Registry, Atlanta, GA and U.S. Department of Health and Human Services,
Public Health Service. September.
Avantaggio J.D., P.S. Price, S. Hays, M. Gargas. 1997. Use of microexposure event
modeling to estimate polychlorinated biphenyl (PCB) concentrations in the blood of
anglers who consume contaminated fish. 1997. Fundamental and Applied Toxicology
(Supplement) 36(1 Part 2): 283 Abstract 1441.
Bahn, A.K., I. Rosenwaike, N. Herrmann, P. Grover, J. Stellman, and K. O'Leary. 1976.
Melanoma after exposure to PCBs. N. Engl. J. Med. 295:450.
Bahn, A.K., P. Grover, I. Rosenwaike, et al. 1977. PCB and melanoma. N. Engl. J. Med.
296:108.
Barclay, B. 1993. Hudson River Angler Survey. Hudson River Sloop Clearwater, Inc.,
Poughkeepsie, New York. March. 56 p.
Becher, G., H.R. Johansen, J. Alexander, M.Lcwik, P.I. Gaarder, and W. Gdynia (1995).
Consumption of crabs and body burden of PCDDs/PCDFs: A study of male crab consumers
from a polluted fjord area in Norway. 15th International Symposium on Chlorinated Dioxins
and Related Compounds 26:39-43. Edmonton, Canada. August 21-25.
72
-------
Bcrtazzi, P.A., L. Riboldi, A. Pesatori, L. Radice, and C. Zocchetti. 1987. Cancer
mortality of capacitor manufacturing workers. Am. J. Ind. Med. 11:165-176.
Brown, D.P. 1987. Mortality of workers exposed to polychlorinated biphenyls: An
update. Arch. Environ. Health. 42(6):333-339.
Brown, D.P. and M. Jones. 1981. Mortality and industrial hygiene study of workers
exposed to polychlorinated biphenyls. Arch. Environ. Health. 36(3): 120-129.
Brown, J.F. 1992. Metabolic alterations of PCB residues in aquatic fauna: Distributions
of cytochrome P4501A- and P4502B-like activities. Mar. Environ. Res. 34:261-266.
Brown, J.F. 1994. Unusual congener selection patterns for PCB metabolism and
distibution in the rhesus monkey. Organohalogenated Compounds 21:29-31.
Brown, J.F., R.W. Lawton, and C.B. Morgan. 1994. PCB metabolism, persistence, and
health effects after occupational exposure: Implications for risk assessment.
Chemosphere 29(9-11):2287-2294.
Brown, J.F., R.W. Lawton, M.R. Ross, and J. Feingold. 1991. Assessing the human
health effects ofPCBs. Chemosphere 23:1811-1815.
Brown, J.F., R.W. Lawton, M.R. Ross, J. Feingold, R.E. Wagner, and S.B. Hamilton.
1989. Persistence of PCB congeners in capacitor workers and Yusho patients.
Chemosphere 19:829-834.
Butterworth, B.E. and T. Slaga. 1987. Nongenotoxic Mechanisms in Carcinogenesis.
Banbury Report 25. New York, NY: Cold Spring Harbor Press.
73
-------
Butterworth, B.E., R.B. Conolly and K.T. Morgan. 1995. A strategy for establishing mode
of action of chemical carcinogens as a guide for approaches to risk assessment. Cancer
Lett. 93:129-146.
Carlson-Lynch H., R.E. Keenan, J.C. Swartout, and P.S. Price, M.L. Dourson. 1997.
Application of Quantitative Information on the Uncertainty in the R£D to Noncancer Risk
Assessments. Accepted for Publication
Chase, K.H., J. Doull, S. Friess, J.V. Rodricks, and S.H. Safe. 1989. Evaluation of the
Toxicology of PCBs. Texas Eastern Gas Pipeline Company, Houston, Texas. March 1,
1989.
ChemRisk. 1991a. Consumption of Freshwater Fish from Maine Lakes and Ponds.
ChemRisk® - A Division of McLaren/Hart Environmental Engineering, Portland, ME.
September 6.
ChemRisk. 1991b. Creel Survey for the West Branch of the Penobscot River. ChemRisk®
- A Division of McLaren/Hart Environmental Engineering, Portland, ME. November 22.
ChemRisk, 1995a. Determining the Intake of Upper Hudson River Fish by Species.
McLaren/Hart-ChemRisk, Portland, ME. January.
ChemRisk, 1995b. Estimating Exposure Duration for the Upper Hudson River Risk
Assessment. McLaren/Hart-ChemRisk, Portland, ME. January.
ChemRisk, 1995c. Estimating Fish Consumption Rates for the Upper Hudson River.
McLaren/Hart-ChemRisk, Portland, ME. January.
ChemRisk, 1995d.. Evaluating the Impact of Cooking Processes on the Level of PCBs in
Fish. McLaren/Hart-ChemRisk, Portland, ME. January.
74
-------
ChemRisk, 1995c.. Hudson River Superfund Project: Approach for Performing Human
Health Risk Assessment — Estimating Potential PCB Exposure from Fish Consumption.
McLaren/Hart-ChemRisk, Portland, ME. May.
Coad, S. 1994. Consumption of Fish and Wildlife by Canadian Nature Peoples: A
Quantitative Assessment from the Published and Unpublished Literature. Health and
Welfare Canada.
Columbia River Inter-Tribal Fish Commission (CRITFC). 1994. A fish consumption survey
of the Umatilla, Nez Perce, Yakama, and Warm Springs Tribes of the Columbia River Basin.
Technical Report 94-3. October.
Connelly, N.A., B.A. Knuth andf C.A. Bisogni. 1992. Effects of Health Advisory
Changes on Fishing Habits and Fish Consumption in New York Sport Fisheries. Human
Dimension Research Unit, Department of Natural Resources, New York State College of
Agriculture and Life Sciences, Cornell University, Ithaca, New York. Report prepared for
the New York Sea Grant Institute. September.
Connelly, N.A., B.A. Knuth and T.L. Brown. 1996. Sprtfish consumption patterns of
Lake Ontario anglers and the relationship to health advisories. N. Amer. J. Fish. Mgmt.
16:90.
Connelly, N.A., B.A. Knuth, and T.L. Brown. 1996. Sportfish consumption patterns of Lake
Ontario anglers and the relationship to health advisories. North American Journal of
Fisheries Management 16:90-101.
Connelly, N.A., T.L. Brown and B.A. Knuth. 1990. New York Statewide Angler Survey
J988. New York State Department of Environmental Conservation. Bureau of Fisheries.
Albany, New York.
75
-------
Connelly, N.A., T.L. Brown, and B.A. Knuth. 1990. New York Statewide Angler Survey
1988. New York State Department of Environmental Conservation, Division of Fish and
Wildlife, Albany, NY. April.
Cook, R.R. 1982. The role of epidemiology in risk assessment. Drug. Metab. Review
12(5):913-923.
Cook, R.R. 1993. Internal Dow Report on the Epidemiology of EDB.
«
Cox, C., A. Vaillancourt, and A.F. Johnson. 1987. A Comparison of the Results from the
"Guide to Eating Ontario Sport Fish " Questionnaires. Ontario Ministry of the
Environment, Aquatic Biology Section, Water Resources Branch, Ontario, Canada. ISBN:
0-7729-2359-0. February.
Cox, C., A. Vaillancourt, and A.F. Johnson. 1990. The Results of the 1989 "Guide to
Eating Ontario Sport Fish" Questionnaire. Ontario Ministry of the Environment, Water
Resources Branch, Ontario, Canada. ISBN 0-7729-7084-X. June.
Cox, C., A. Vaillancourt, C. De Ban-os, and A.F. Johnson. 1985. "Guide to Eating
Ontario Sport Fish " Questionnaire Results. Ontario Ministry of the Environment, Aquatic
Contaminants Section, Water Resources Branch, Ontario, Canada. May.
Danse, I.R., R.J. Jaeger, R. Kava, M. Kroger, W.M. London, F.C. Lu, R.P. Maickel, J.J.
McKetta, G.W. Newell, S. Shindell, FJ. Stare, and E.M. Whelan. 1997. Position paper of
the American Council on Science and Health: Public Health Concerns about
Environmental Polychlorinated Biphenyls (PCBs). Ecotoxicol. Environ. Safety 38:71-84.
Dewailly, E., A- Nantel, J.P. Weber and F. Meyer (1989): High levels of PCBs in breast milk
of Inuit women from arctic Quebec. Bull. Environ. Contam. Toxicol. 43:641-646.
76
-------
Dinman, B.D. and N.B. Sussman. 1983. Uncertainty, risk, and the role of epidemiology in
public policy development. Reg. Toxicol. Pharmacol. 7:206-216.
Doll, R. 1984. Occupational cancer: Problems in interpreting human evidence. Ann. Occup.
Hyg. 28(3):291-305.
Doull, J. and S. Abrahamson. 1986. Letter to Richard A. Griesemer, Chair.,
Environmental Health Committee, Science Advisory Board, U.S. Environmental
Protection Agency reporting the Halogenated Organics Subcommittee review of 15 draft
health advisories for drinking water. (SAB-EHC-87-005) 5-27.
Ebert, E.S., N.W. Harrington, K.J. Boyle, J.W. Knight and R.E. Keenan (1993): Estimating
consumption of freshwater fish among Maine anglers. N. Am. J. Fish. Mgt. 13:737-745.
Ebert, E.S., P.S. Price, and R.E. Keenan (1994): Selection of fish consumption estimates for
use in the regulatory process. Journal of Exposure Analysis and Environmental
Epidemiology 4(3)373-393.
Ebert, E.S., S.H. Su, T.J. Bany, M.N. Gray, and N.W. Harrington (1996): Estimated Rates
of fish consumption by anglers participating in the Connecticut Housatonic River creel
survey. N. Am. J. Fish. Mngt. 16:81-89.
EPA. 1985b. Chemical Carcinogens: A Review of the Science and its Associated Principles.
(February). U.S. Environmental Protection Agency, Office of Science and Technology
Policy, Washington, DC. 50 Fed. Reg. 10380 (March 14).
EPA, 1989. Assessing Human Health Risks from Chemically Contaminated Fish and
Shellfish: A Guidance Manual. U.S. Environmental Protection Agency, Washington, D.C.
EPA-503/8-89-002. September.
77
-------
EPA, 1989. Exposure Factors Handbook. U.S. Environmental Protection Agency, Office of
Health and Environmental Assessment, Washington, D.C. EPA/600/8-89/043. July.
EPA, 1991. Phase 1 Review Copy: Interim Characterization and Evaluation Hudson
River PCB Reassessment RI/FS. Prepared by TAMS consultants, Inc. and Gradient Corp.
for U.S. Environmental Protection Agency, Region II, Alternative Remedial Contracting
Strategy (ARCS) for Hazardous Remedial Services, New York, NY. EPA Contract No.
68-S9-2001.
EPA, 1992. Hudson River PCB Reassessment RI/FS. Phase 2 Work Plan and Sampling
Plan. U.S. Environmental Protection Agency, EPA ARCS Region n, New York.
EPA, 1992a. Memo to Assistant Administrators and Regional Administrators from F.H.
Habicht, Deputy Administrator "RE: Guidance on Risk Characterization for Risk
Managers and Risk Assessors." U.S. Environmental Protection Agency, Washington, D.C.
(February 26).
EPA, 1992b. Final Guidelines for Exposure Assessment. U.S. Environmental Protection
Agency, Washington, D.C. 57 FR 104. (May 29).
EPA, 1996. Proposed Guidelines for Carcinogen Risk Assessment. U.S. Environmental
Protection Agency, Office of Research and Development. EPA/600/P-96/003C. April.
EPA, 1997. Special Report on Environmental Endocrine Disruption: An Effects Assessment
and Analysis. U.S. Environmental Protection Agency, Risk Assessment Forum.
Washington, DC. EPA/630/R-96/012.
EPA, 1997a. Policy for use of Probabilistic Analysis in Risk Assessment at the EPA, Office
of Research and Development, Washington D.C., EPA Publication Number: EPA/630/R-
97/001.
78
-------
EPA, 1997b. Guiding Principles for Monte Carlo Analysis
EPA, 1998. Exposure Factors Handbook . U.S. Environmental Protection
Agency,Washington, DC. EPA/600/P-95/002A. June.
Evans, A. 1976. Causation and disease: The Henle-Koch postulates revisited. Yale J. Med.
49:175.
Fein, G.G., J.L. Jacobson, S.W. Jacobson, P.M. Schwartz, and J.K. Dowler. 1984.
Prenatal exposure to polychlorinated biphenyls: Effects on birth size and gestational age.
J. Ped. :315-320.
Fiore, B.J., H.A. Anderson, L.P. Hanrahan, L.J. Olson, and W.C. Sonzogni. 1989. Sport
fish consumption and body burden levels of chlorinated hydrocarbons: A study of
Wisconsin anglers. Arch. Environ. Health 44(2):82-88.
GE, 1996. Comments of the General Electric Company on Phase 2 Report - Review Copy
Further Site Characterization and Analysis Volume 2B - Preliminary Model Calibration
Report Hudson River PCBs Reassessment Rl/FS October J 996.
Gillis, C. and P. Price. 1996. Comparison of the noncarcinogenic effects and PCB body
burdens in rhesus monkeys and humans: Implications for risk assessment. Toxicologist
30(1):748.
Golden, RJ., K.L., Noller, L. Titus-Emstoff, R.H. Kaufman, R. Mittendorf, R. Stillman,
and E. Reese. 1998. Environmental endocrine modulators and human health: An
assessment of the biological evidence. Crit. Rev. Toxicol. 28(2):109-227.
Gray Jr., L.E., J. Ostby, R. Marshall, 1993. Reproductive and thyroid effects of low-level
polychlorinated biphenyl (Aroclor 1254) exposure. Fundam. Appl. Toxicol. 20(3):288-294.
Guidotti, T. and D. Goldsmith. 1986. Occupational Cancer. AFP 34:146.
79
-------
Gustavsson, P. and C. Hogstedt. 1997. A cohort study of Swedish capacitor
manufacturing workers exposed to polychlorinated biphenyls (PCBs). Am. J. Industr.
Med. 32:234-239.
¦>
Gustavsson, P., C. Hogstedt, and C. Rappe. 1986. Short-term mortality and cancer
incidence in capacitor manufacturing workers exposed to polychlorinated biphenyls
(PCBs). Am. J. Ind. Med. 10:341-344.
Hackney, J.D. and W.S. Linn. 1979. Koch's postulates updated: a potentially useful
application to laboratory research and policy analysis in environmental toxicology. Ame
Review of Respir. Disease 119:849-852.
Harvey T., R.E. Keenan, J.C. Swartout, H.L. Carlson-Lynch, C.A. Gillis, and P.S. Price.
1997. Application of probabilistic methods to noncarcinogenic risk assessment: A case
study of hexachloroethane and paraquat. Fundamental and Applied Toxicology
(Supplement) 36(1 Part 2): 208. Abstract 1059.
Hemberg, S. 1992. Introduction to Occupational Epidemiology. Lewis Publishers,
Chelsea, Michigan.
Hertz-Picciotto, I., N. Gravitz, and R. Neutra. 1988. How do cancer risks predicted fron
animal bioassays compare with the epidemiologic evidence? The case of ethylene
bromide. Risk Anal. 8(2):205-214.
Hill, A.B. 1965. The environment and disease: association or causation? Proceedings o:
the Royal Society of Medicine, Section of Occupational Medicine.
Honstead, J.F., T.M. Beetle, and J.K. Soldat. 1971. A Statistical Study of the Habits of
Local Fishermen and its Application to Evaluation of Environmental Dose. A Report to
80
-------
the U.S. Environmental Protection Agency by Battelle Pacific Northwest Laboratories,
Richland, WA.
Houk, V.N. 1990. Testimony of Vernon N. Houk, M.D. before the Subcommittee on
Human Resources and Intergovernmental Relations Committee on Government
Operations.
Hubert, W.A., A.O. Smith, W.T. Morgan, W.P. Mitchell, and R.L. Warden. 1975. Summary
of Commercial Fisherman Surveys 1971-1974. Tennessee Valley Authority, Muscle Shoals,
Alabama.
Hunter, D.J., S.E. Hankinson, F. Laden, G.A. Colditz, J.E. Manson, W.C. Willett, F.E.
Speizer, and M.S. Wolff. 1997. Plasma organochlorine levels and the risk of breast
cancer. N. Eng. J. Med. 337(18):1253-1258.
Hutchison, R. and C.E. Kraft. 1994. Hmong fishing activity and fish consumption. J. Great
Lakes Res. 20(2):471-478.
International Agency for Research on Cancer (IARC). 1987. Monographs on the
evaluation of the carcinogenic risk of chemicals to humans. Supplement 7: update of
IARC Monographs. Vols. 1-42. World Health Organization, Lyon, France.
Jacobson, J.L. and S.W. Jacobson. 1997. Evidence for PCBs as neurodevelopmental
toxicants in humans. Neurotoxicology 18(2):415-424.
Jacobson, J.L., S.W. Jacobson, P.M. Schwartz, G.G. Fein, and J.K. Dowler. 1984.
Prenatal exposure to an environmental toxin: A test of the multiple effects model.
Develop. Psychol. 20(4):523-532.
81
-------
Jacobson, S.W., G.G. Fein, J.L. Jacobson, P.M. Schwartz, and J.K. Dowler. 1985. The
effect of intrauterine PCB exposure on visual recognition memory. Child Develop.
56:853-860.
Javitz, H. 1980. Seafood Consumption Data Analysis; Final Report. Prepared by Statistical
Analysis Department, SRI International for U.S. Environmental Protection Agency,
Washington, D.C. Contract No. 68-01-3887. September 24.
Johnson, T. and J. Capel. 1992. A Monte Carlo Approach to Simulating Residential
Occupancy Periods and Its Application to the General U.S. Population. U.S.
Environmental Protection Agency, Office of Air Quality, Planning, and Standards,
Research Triangle Park, NC. EPA-450/3-92-011. August.
Kashimoto, T., and H. Miyata. 1986. Differences between Yusho and other kinds of
poisoning involving only PCBs. In: PCBs and the Environment, Vol. 3, Waid, J.S. ed.,
Boca Raton: CRC Press, 2-26.
Keenan, R.E., D.J. Paustenbach, R.J. Wenning, and A.H. Parsons. 1991. A pathology re-
evaluation of the Kociba et al. (1978) bioassay of 2,3,7,8-TCDD: Implications for risk
assessment J. Toxicol. Environ. Health 34:279-271.
Keenan, R.E. and N.W. Harrington. 1996. Letter to Kevin Garrahan, Chief, Exposure
Assessment Branch, U.S. EPA Office of Research and Development, re: Results of
Additional Maine Angler Survey Analysis. March 1.
Keenan, R.E., J.D. Avantaggio, and P.S. Price. 1997. Using a combined Microexposure
Event and Toxicokinetic Model to evaluate the need for fish advisories based on a body
burden dosimetric. In: Society for Risk Analysis Proceedings, Annual Meeting and
Exposition. Abstract.
82
-------
Keenan, R.E., P.S. Price, J. McCrodden, and E.S. Ebert. 1996b. Using a microexposure
event analyses to model potential exposures to PCBs through ingestion of fish from the
Upper Hudson River. In: Organohalogen Compounds: Proceedings Dioxin '96-16th
International Symposium on Chlorinated Dioxins and Related Compounds, Amsterdam,
The Netherlands. Organohalogen 30:61-65.
Key, T. and G. Reeves. 1994. Organochlorines in the environment and breast cancer. 308
British Med. J. 6943:1520.
Kimbrough, R.D. 1987. Human health effects of polychlorinated biphenyls (PCBs) and
polybrominated biphenyls (PBBs). Ann. Rev. Pharmacol. Toxicol. 27:87-111.
Kimbrough, R.D. 1990. The Human Health Effects of Polychlorinated Biphenyls. U.S.
Environmental Protection Agency, Washington, D.C.
Kimbrough, R.D. 1995. Polychlorinated biphenyls (PCBs) and human health: An update.
Crit. Rev. Toxicol. 25(2):133-163.
Kimbrough, R.D. 1996. Review of Studies in Children: Polychlorinated Biphenyls,
Dibenzo-p-dioxins, and Dibenzofurans. Institute for Evaluating Health Risks, Washington,
DC. October.
Kimbrough, R.D. 1988. Polychlorinated biphenyls: how do they affect human health?
Health Environ. Digest 2(7): 1.
Kimbrough, R.D. 1990. How toxic is 2,3, 7,8-tetrachIorodibenzodioxin to humans? J.
Toxicol-Environ. Health 30:261-271.
Kimbrough, R.D. and M.L. Doemland. 1997. Neurobehavioral studies in children - The
limitations of epidemiology studies. In: Health Conference, An International Scientific
83
-------
Conference on the Effects of the Environment on Human Health in the Great Lakes and St.
Lawrence River Basins. Montreal, Quebec, Canada. May 12-15.
Kimbrough, R.D., M.L. Domeland, and M.E. LeVois. 1998. Mortality in Capacitor
Workers Exposed to PCBs. Institute for Evaluating Health Risks Report. February.
Kinloch, D., H. Kuhnlein and D.C.G. Muir. 1992. Inuit foods and diet: a preliminary
assessment of benefits and risks. Sci. Tot. Environ. 122:247-278.
Kunita, N.» T. Kashimoto, H. Miyata, S. Fukushima, S. Hon, and H. Obana. 1984. Causal
agents ofYusho. Am. J. Ind. Med. 5:45-58.
Landolt, M., D. Kalman, A. Nevissi, G. van Belle, K. van Ness, and F. Hafer 1987. Final
Report: Potential Toxicant Exposure among Consumers of Recreationally Caught Fish
from Urban Embayments of Puget Sound. NOAA Technical Memorandum NOS OMA 33.
National Oceanic and Atmospheric Administration, Rockville, MD. April.
Landolt, M.L., F.R. Hafer, A. Nevissi, G. van Belle, K. Van Ness and C. Rockwell. 1985.
Potential Toxicant Exposure Among Consumers of Recreationally Caught Fish from Urban
Embayments of Puget Sound. National Oceanic and Atmospheric Administration, Rockville,
MD. Tech. Memo. NOS OMA 23. November.
Lawrence, C. 1977. PCBs and melanoma. N. Engl. J. Med. (January 13).
Lawton, R.W., B.T. Sack, M.R. Ross, J. Feingold. 1981. A progress report: studies of
employees occupationally exposed to PCBs.
Lawton, R.W., M.R. Ross, I. Feingold, and J.F. Brown Jr. 1985. Effects of PCB exposure
on biochemical and hematological findings in capacitor workers. Environ. Health
Perspect. 60:165-182.
84
-------
Layard, M.W. and A. Silvers. 1989. Epidemiology in environmental risk assessment, p.
157-173, In, D.J. Paustenbach (ed.), The Risk Assessment of Environmental and Human
Health Hazards. John Wiley & Sons, Inc., New York.
Loo mis, D., S.R. Browning, A.P. Schenk et al. 1997. Cancer mortality among electrical
workers exposed to polychlorinated biphenyls. Occup. Environ. Med. 54>:720-728.
MacDonald, J.S., G.R. Lanskas, and R.E. Morrissey. 1994. Toxicokinetic and mechanistic
considerations in the interpretation of the rodent bioassay. Toxicol. Path. 22(2): 124-139.
Massachusetts Department of Public Health (MDPH). 1997. Cancer Incidence and
Mortality in Massachusetts 1987-1994. (November).
Masuda, Y., R. Kagawa, H. Kuroki, M. Kuratsune, T. Yoshimura, I. Taki, M. Kusuda, F.
Yamashita, and M. Hayashi. 1978. Transfer of polychlorinated biphenyls from mothers to
foetuses and infants. Food Cosmet. Toxicol. 16:543-546.
Mausner, J.S. and S. Kramer. 1985. Epidemiology - An Introductory Text. Philadelphia:
W. B. Saunders Co..
Mayes, B.A., E.E. McConnell, B.H. Neal, M.J. Brunner, S.B. Hamilton, T.M. Sullivan,
A.C. Peters, M.J. Ryan, J.D. Toft, A.W. Singer, J.F. Brown, Jr., R.G. Menton, and J.A.
Moore. 1998. Comparative carcinogenicity in Sprague-Dawley rats of the polychlorinated
biphenyl mixtures Aroclors 1016,1242,1254, and 1260. Toxicol. Sci. 41:62-76.
McConnack, C. and D. Cleverly. 1990. Analysis of the potential populations at risk from the
consumption of freshwater fish caught near paper mills. U.S. EPA. April 23.
McKone, T.E. and K.T. Bogen. 1991. Predicting the uncertainties in risk assessment.
Environ. Sci. Technol. 25(10): 1674-1681.
85
-------
Mes, J., D.L. Arnold, F. Bryce, D.J. Davies, and K. Karpinski. 1989. The effect of long-
term feeding of Aroclor 1254 to female rhesus monkeys of their polychlorinated biphenyl
tissue levels. Arch. Environ. Contam. Toxicol. 18:858-865.
Monson, R. 1988. Occupational Epidemiology. CRC Press, Boca Raton, Florida.
New York State Department of Fish and Wildlife (NYFW), 1995. Fishing Regulations
Guide, October 1, 1995-September 30,1996, New York Department of Environmental
Conservation
Muir, W.R., Young, J.S., Benes, C.F., Chaisson, D.K., Waylett, M.E., Hawley, C.B.,
Sandusky, Y., Sert, E., DeGraff, P.S.,Price, R.E., Keenan, J.A., Rothrock, N.L., Bonnevie,
J.I., Mccrodden-Hamblen, A Case Study and Presentation of Relevant Issues on Aggregate
Exposure, in ILSI Aggregrate Exposure Workshop, Ed. S. Olin, ILSI Press, Washington
DC., In Press
New York State Department of Health (NYSDOH), 1997. 1997-1998 Health Advisories:
Chemicals in Sportfish and Game. Albany, NY. March. 15 p.
New York State Department of Health (NYSDOH). 1991. Vital Statistics of the United
States: 1990. New York State Department of Health and New York State Data Center,
Albany, N.Y.
Nichols, A.L. and R.J. Zeckhauser. 1988. The perils of prudence: How conservative risk
assessments distort regulation? Reg. Toxicol Pharmacol. 8:61-75.
Nicholson, J.W. 1987. Report of the Special Panel on Occupational PCB Exposure and
Various Cancers: Human Health Effects and Carcinogenic Risk Potential of PCB. Mount
Sinai School of Medicine. (August). Reprinted in the Ontario Gazette in the Matter of
Section 86p of the Workers Compensation Act. (December)
86
-------
NIOSH. 1977. Criteria For A Recommended Standard: Occupational Exposure To
PolychlorinatedBiphenyls (PCBs). NIOSH publ. 77-225. Rockville, MD: U.S.
Department of Health Education and Welfare, Public Health Service, Centers for Disease
Control, National Institute for Occupational Safety and Health.
NYSDEC, 1998. New York State Department of Environmental Conservation Division of
Law Enforcement - Region 5, Memo to Christine Bosy, General Electric, From J.T.
McLean, EnCon - Law Enforcement - Region 5, August 3
NYSDEC. 1990. New York Statewide Angler Survey J988. New York State Department
of Environmental Conservation, Division of Fish and Wildlife, Albany, NY. April.
NYSDOH, 1993. Health Risk Assessment for the Akwesasne Mohawk Population from
Exposure to Chemical Contaminants in Fish and Wildlife from the St. Lawrence River
Drainage on Lands of the Mohawk Nation at Akwesasne and Near the General Motors
Corporation Central Foundry Division at Massena, New York. New York State Department
of Health, Bureau of Toxic Substance Assessment. October.
Parker, G.S. 1980. Letter from MDPH Assistant Commissioner to Peter Arlos, Pittsfield,
Massachusetts City Council.
Paustenbach, D.J. 1989. Important recent advances in the practice of health risk
assessment: Implications for the 1990's. Toxicol. Appl. Pharmacol. 10:204-243.
Paustenbach, D J., J.D. Jemigan, B.L. Finley, S.R. Ripple, and R.E. Keenan. 1990. The
current practice of health risk assessment. J. Air Waste Manag. 40:1620-1630.
Paynter, O.E., GJ. Burin, R.B. Jaeger, and C.A. Gregorio. 1988. Goitrogens and thyroid
follicular cell neoplasia: Evidence for a threshold process. J. Reg. Toxicol. Pharmacol
8:102-119.
87
-------
Peterson, D.E., M.S. Kanarek, M.A. Kuykendall, J.M. Diedrich, H.A. Anderson, P.L.
Remington and T.B. Shefiy. 1994. Fish consumption patterns and blood mercury levels in
Wisconsin Chippewa Indians. Arch. Environ. Health 49(l):53-58.
Pierce, D., D. Noviello, and S. Rogers. 1981. Commencement Bay Seafood Consumption
Study. Tacoma-Pierce County Health Department, Seattle, Washington. December.
Pitot, H.C. and Y.P. Dragan. 1991. Facts and theories concerning the mechanisms of
carcinogenesis. FASEB Journal 5: 2280-2286.
Price, P.S., R.E. Keenan, J.C. Swartout, C.A. Gillis, H. Carlson-Lynch, and M.L. Dourson.
1997. An approach for modeling noncancer dose responses with an emphases on
uncertainty. Risk Anal. Vol 17, No. 4.
Price, P.S., C.L. Curry, P.E. Goodrum, M.N. Gray, J.I. McCrodden, N.W. Harrington, H.
Carlson-Lynch, and R.E. Keenan. 1996. Monte Carlo modeling of time-dependent
exposures using a Microexposure Event approach. Risk Anal. 16(3):339-348.
Price, P.S., S. Su, and M. Gray. 1994. The effect of sampling bias on estimates of angler
consumption rates. J. Expos. Anal. Environ. Epid. 4(3):355-372.
Price, P.S., T.E. Widner, J.I. Hamblin-McCrodden, J.D. Avantaggio, P. Gwinn, C.
Schmidt, N. Bonnaivie, 1998 Oak Ridge Health Studies Oak Ridge Dose Reconstruction,
Task 3 Report, PCBs in the Environment Near the Oak Ridge Reservation, A
Reconstruction of Historical Doses and Health Risks, Submitted to the Tennessee
Department of Health by ChemRisk, August
Puffer, H.W., S.P. Azen, M.J. Duda, and D.R. Young. 1981. Consumption Rates of
Potentially Hazardous Marine Fish Caught in the Metropolitan Los Angeles Area.
88
-------
University of Southern California School of Medicine for Environmental Research
Laboratory, U.S. Environmental Protection Agency. Grant No. R 807 120010.
Richardson, G.M. and D.J, Currie. 1993. Estimating fish consumption rates for Ontario
Amerindians. J. Expos. Anal. Environ. Epi. 3(l):23-37.
Rogan, WJ., B.C. Gladen, J.D. McKinney, N. Carreras, P. Hardy, J. Thullen, J. Tinglestad,
andM. Tully. 1986. Neonatal effects of transplacental exposure to PCBs and DDE. J.
Ped. 109:335-341.
Rothman, K.J. 1988. Causal Inference. Epidemiology Resources, Inc. Chestnut Hill,
Massachusetts.
Rothman, N., K.P. Cantor, A. Blair, D. Bush, J.W. Brock, K. Helzlsouer, S.H. Zahm, L.L.
Needham, G.R. Pearson, R.N. Hoover, G.W. Comstock, and P.T. Strickland. 1997. A
nested case-control study of non-Hodgkin Lymphoma and serum organochlorine residues.
Lancet: 350:240-244.
Rupp, E.M., F.M. Miller, and I.C. Baes. 1980. Some results of recent surveys of fish and
shellfish consumption by age and region of U.S. residents. Health Physics 39:165-175.
Safe, S.H. 1997. Xenoestrogens and Breast Cancer. 337 N. Engl. J. Med. 18:1303
SAP, 1998. Science Advisory Pannel, Federal Insecticide, Fungicide, and Rodenticide
Act, Report of Scientific Advisory Pannel Meeting, March 24
Schmidt C.W., C.A. Gillis, R.E. Keenan, and P.S. Price. 1997. Characterizing inter-
chemical variation in the interspecies uncertainty factor (UFa). Fundamental and Applied
Toxicology (Supplement) 36(1 Part 2): 208. Abstract 1057.
Schulte-Hermann, R. 1985. Tumor promotion in the liver. Arch. Toxicol. 57:147-158.
89
-------
SelikofF, I.J., E.C. Hammond, and S.M. Levin. 1982. Environmental contaminants and the
health of the people of the St. Regis Reserve. Vol. II.
Sherer, R.A. and P.S. Price. 1993. The effect of cooking processes on PCB levels in
edible fish tissue. Qual. Assuran. GoodPrac. Reg. Law 2(4):396-407.
j
Sielken, R.L. 1985. Some issues in the quantitative modeling portion of cancer risk
assessment. Rtgul. Toxicol. Pharmacol. 5:175-181.
Silbergeld, E.K., J. Schwartz, and K. Mahaffey. 1988. Lead and osteoporosis:
Mobilization of lead from bone in postmenopausal women. Environ. Res. 47:79-94.
Sinks, T., G. Steele, A.B. Smith, K. Watkins and R.A. Shults. 1992. Mortality among
workers exposed to polychlorinated biphenyls. Amer. J. Epidem. 136(6):389-398.
Sinks, T., G. Steele, A.B. Smith, R. Rinsky, and K. Watkins. 1991. Westinghouse
Electric, Bloomington Indiana Health Hazard Evaluation and Technical Assistance
Branch DSHEFS, NIOSH; Risk Factors Associated with Excess Mortality Among
Polychlorinated Biphenyl Exposed Workers. HETA 89-116-2094.
Soldat,J.K. 1970. Chapter 35: A Statistical Study of the Habits of Fishermen Utilizing
the Columbia River Below Hanford, p. 302-308, In, W.C. Reinig [ed.], Environmental
Surveillance in the Vicinity of Nuclear Facilities: Proceedings of a Symposium Sponsored
by the Health Physics Society, January 24-26,1968. Charles C. Thomas Publishers,
Springfield, Illinois.
Sortie, P.D., E. Backlund, and J.B. Keller. 1995. U.S. mortality by economic
demographic, and social characteristics; The National Longitudinal Mortality Study. Am.
J. Public Health 85:949-956.
90
-------
Southern California Coastal Water Research Project (SCCWRP) and MBC Applied
Environmental Sciences. 1994. Santa Monica Bay Seafood Consumption Study. Santa
Monica Bay Restoration Project, Monterey Park, CA. June
Swanson, G.M., H.E. RatclifTe, and LJ. Fischer. 1995. Human exposure to
polychlorinated biphenyls (PCBs): A critical assessment of the evidence for adverse health
effects. Reg. Toxicol. Pharma. 21:136-150.
Taylor, P.R. The Health Effects of Polychlorinated Biphenyls. Harvard School of Public
Health, Boston, MA, 1988.
TERRA, Inc. 1993. Comments on the Water Quality Guidance for the Great Lakes System.
Tryphonas, H., M.I. Luster, G. Schiffinan, L.L. Dawson, M. Hodgen, D. Germolec, S.
Hayward, F. Bryce, J.C.K. Loo, F. Mandy and D.L. Arnold. 1991a. Effects of chronic
exposure of PCB (Aroclor 1254) on specific and nonspecific immune parameters in the
rhesus (Macaco mulatto) monkey. Fund. Appl. Toxicol. 16:773-786.
Tiyphonas, H., M.I. Luster, K.L. White, P.H. Naylor, M.R. Erdos, G.R. Burleson, D.
Germolec, M. Hodgen, S. Hayward, and D.L. Arnold. 1991b. Effects of PCB (Aroclor®
1254) on non-specific immune parameters in rhesus (Macaca mulatto) monkeys. Int. J.
Immunoph. 13(6):639-648.
Tryphonas, H., S. Hayward, L. O'Grady, J.C.K. Loo, D.L. Arnold, F. Bryce and Z.Z.
Zawidzka. 1989. Immunotoxicity studies of PCB (Aroclor 1254) in the adult rhesus
(Macaca mulatto) monkey - Preliminary report. Int. J. Immunoph. 11 (2): 199-206.
Turcotte, S. 1983. Memorandum from S. Turcotte, Technical Division, Savannah River
Laboratory to H.P. Olson. RE: Updated "Georgia Fishery Study: Implications for Dose-
Calculations". August 5.
91
-------
USDA. 1980. Food and Nutrient Intakes of Individuals in One Day in the United States:
Spring 1977. Nationwide Food Consumption Survey 1977-1978. Preliminary Report No.
2.
Weisburger, J. and G. Williams. 1987. Chemical carcinogens. In, C. Klaassen, J. Doull,
and M.A. Amdur, (eds.), Casarett and Doull's Toxicology: The Basic Science of Poisons.
Third edition, Macmillan Publishing Company, New York, NY.
Wendt, M.E. 1986. Low income families' consumption of freshwater fish caught from New
York State waters. Masters Thesis. Cornell University. August.
West, P.C., J.M. Fly, F. Larkin, and R. Marans 1991. Minority anglers and toxic fish
consumption: Evidence from a state-wide survey of Michigan. In: Proceedings of the
Michigan conference on race and the incidence of environmental hazards (Bryan and Mohai,
eds.).
West, P.C., J.M. Fly, R. Marans and F. Larkin 1989. Michigan Sport Anglers Fish
Consumption Survey. University of Michigan, School of Natural Resources, Ann Arbor, MI.
Technical Report No. 1. May.
WHO. 1997. Draft Report on the Meeting on the Derivation of Toxic Equivalency Factors
(TEFs) for PCBs, PCDDs, PCDFs and Other Dioxin-like Compounds for Humans and
Wildlife. World Health Organization. June 15-18. Stockholm. 55p.
Willes, R.F., E.R. Nestmann, P.A. Miller, J.C. Orr, and I.C. Munro. 1993. Scientific
principles for evaluating the potential for adverse effects from chlorinated organic
chemicals in the environment. Reg. Toxicol. Pharmacol. 18:313-356.
Wilson, N.D., N.M. Shear, D.J. Paustenbach, and P.S. Price. 1998. The Effect of Cooking
Practices on the Concentration of DDT and PCB Compounds in the Edible Tissue of Fish.
J. of Exp. Anal, and Envior. Epid. Vol. 8 pp. 423-440.
92
-------
Wolfe, R.J. and R.J. Walker 1987. Subsistence economies in Alaska: Productivity,
geography, and development impacts. Arctic Anthropology 24(2):56-81.
Wong, 0. 1995. Risk of acute myeloid leukaemia and multiple myeloma in workers
exposed to benzene. Occup. Environ. Med. 52:380-384.
Yassi, A., R. Tate, and D. Fish. 1994. Cancer mortality in workers employed at a
transformer manufacturing plant. Am. J. Ind. Med. 25:425-437.
Zack, T.A. and D.C. Musch. 1979. Mortality of PCB workers at the Monsanto plant in
Sauget, Illinois. Monsanto Internal Report.
93
-------
Figure 9-12
Comparison of Observed and Predicted Aroclor 1254 Concentrations in
Hudson River Largemouth Bass (Corrected to NYSDEC 1983 Quantitation Basis)
1,200
1.000
2
% 800
O)
3
§ 600
T3
O)
Cl.
400
200
0	200	400	600	800	1,000	1,200
Observed (pg/g-lipid)
Source: TAMS/Gradient Database. Release 3.1
Legend:
A RM 142-1
B RM 160
C RM 175
D RM 189-1!










53









)3

C
D













n










DO
D

c°
IT
D







/






C

C
c £












C







B
f


A
A
A
B





A

A








Figure 9-12 from EPA, 1996. Preliminary Model Calibration Report.
-------
NYSDEC RIVER ENFORCEMENT SUMMARY
OF THE CATCH AND RELEASE FISHING PROGRAM
(8/31/95 - 7/31/98)
VIOLATION TYPE
TICKETS
ISSUED
WARNINGS
ISSUED
TOTAL
KEEPING FISH
9
3
12
BAIT FISH/TIP-UPS
5
74
79
NO LICENSE
72
93
165
NAVIGATIONAL/
OTHER
40
28
68
TOTALS
126
198
324
TOTAL FISHERMEN I
CHECKED TO DATE
: - ; 1437.


08/10/98
-------